Advances in Urethane Science and Technology
Advances in Urethane Science and Technology
Advances in Urethane Science and Technology
in Urethane
Science and
Technology
Editors:
D. Klempner
K.C. Frisch
Advances in
Urethane Science
and Technology
Daniel Klempner
and
Kurt Frisch
A catalogue record for this book is available from the British Library.
ISBN: 1-85957-275-8
i
Advances in Urethane Science and Technology
2.2 Current Liquid Carbon Dioxide Technologies for Flexible Slabstock .........
Polyether Foam Production .................................................................... 88
2.2.1 Machinery ................................................................................... 88
2.2.2 The Foaming Process ................................................................... 90
2.2.3 Additional Tasks of Silicone Surfactants in Flexible Slabstock
Foam Production ......................................................................... 95
2.2.4 Chemistry of a Silicone Surfactant in Flexible Slabstock
Foam Production ......................................................................... 99
2.2.5 A Surfactant Development Example .......................................... 101
ii
Contents
iii
Advances in Urethane Science and Technology
iv
Contents
v
Advances in Urethane Science and Technology
vi
Contents
vii
Advances in Urethane Science and Technology
APPENDIX........................................................................................................ 465
Laboratory Preparation of 2,4-TDI and 4,4´-MDI Prepolymers ................... 465
Laboratory Casting of 4,4´-MDI Prepolymers Cured with BDO .................. 465
Laboratory Casting of One-Shot Elastomers Based on Carbodiimide-
Modified MDI, Polyol, and BDO ................................................................. 465
Laboratory Casting of 2,4-TDI Prepolymers Cured with MBOCA .............. 466
Laboratory Moisture-Curing of 2,4-TDI Prepolymers ................................. 466
Laboratory Preparation of Aqueous Polyurethane/Urea Dispersions
using the Prepolymer Mixing Process ........................................................... 466
viii
Preface
On a more solemn note, one of the editors, Kurt C. Frisch, passed away shortly
before publication. Dr. Frisch, founder of the University of Detroit Mercy’s Polymer
Institute, was one of the pioneers of polyurethanes and was responsible for the
successful introduction of polyether polyurethane flexible foams into commerce in
the mid-1950s. Let us not only mourn the loss of, but also celebrate the life of this
great scholar by continuing to further the frontiers of urethane science and
technology. This volume is a good example of this progress.
I would like to take this opportunity to express my gratitude to the authors who
contributed to this book and to the University of Detroit Mercy for its
encouragement of this effort.
I would also like to thank the staff of Rapra, in particular, Frances Powers, Claire
Griffiths and Steve Barnfield.
1
Advances in Urethane Science and Technology
2
1
Dimensional Stabilising Additives for Flexible
Polyurethane Foams
Gary D. Andrew, Jane G. Kniss, Mark L. Listemann, Lisa A. Mercando,
James D. Tobias and Stephan Wendel
1.1 Introduction
The issues that an automotive seat manufacturer faces when formulating and producing
seats are escalating. Physical properties such as tensile and tear strengths, compression
set and wet set are critical when meeting specific mechanical performance requirements
as defined by the original equipment manufacturer (OEM). As new requirements for
comfort and durability are instituted, tests such as dynamic creep testing, long term
vibration characterisation and repeated compression tests under various atmospheric
and load conditions have been used to characterise foam performance for comfort.
Comfort properties are best controlled by the polyols used to produce the polyurethane
foam cushion. Significant changes in polyol technology to meet these dynamic comfort
properties have had an impact on the processing of polyurethane foam and on physical
properties. Increased tightness of the foam article resulting from changes in these raw
materials has focused more attention by foam producers on crushing methods. Flexible
moulded polyurethane foam requires some type of mechanical crushing to prevent
shrinkage and ultimately maintain part stability.
With recent changes made to polyol technology, mechanical methods of crushing do not
always provide the consistency required to produce a part that is dimensionally stable.
Additionally, producers of polyurethane articles are continually building more complexity
into their seat designs to meet the aesthetic values required by today’s consumers. These
complex seat designs place more emphasis on crushing capability due to the nature of
the designs. With all these changes, additives needed to be developed which provide a
wider processing latitude and increased breathability to the polyurethane article. Wider
processing latitudes should reduce scrap and repair rates on the foam production line
and improve economics for the polyurethane producer [1].
The formation of moulded foam is a complicated chemical process which involves several
reactions occurring simultaneously. There are rapid volume and temperature increases
and the concurrent development of phase separated polymer networks. To understand
how foam properties can be affected by catalyst and surfactant chemistries several
techniques are used to identify key performance benefits and issues. A force to crush
3
Advances in Urethane Science and Technology
(FTC) detection device was used to measure the force required to crush a part to 50% of
its thickness for determination of cell openness. Mass-loss/rate-of-rise was run to
understand rate of rise and height measurements, weight loss from carbon dioxide
generation and temperature profiles. A scanning electron microscope (SEM) was used to
determine differences in cell structure and cell distribution caused by changes in the
catalyst and surfactant chemistries. Physical properties were also tested using ASTM test
methods for flexible cellular polyurethane. A novel chemical reaction foam modelling
technique was also used to determine the selectivity of the catalyst packages, compared
to industrial standard controls [2].
In the past it was thought that the cell structure of polyurethane foam is controlled by
the type and amount of surfactant used. Dabco DC5043 (Air Products and Chemicals,
Inc.) was developed to enhance cell wall drainage to better enable cell opening during
crushing cycles. It was also thought that surfactant technology was the best way to
provide improved crushing techniques; therefore, catalyst technology was ignored [3].
As mentioned earlier, with new polyol technology development more emphasis was placed
on crushing. New additive technology needed to be developed that would open cells
during the foam formation and reduce the requirement and criticality of the crushing
processes. The technology had to go beyond providing easier cell opening at crush to
providing more open cells during the polyurethane formation.
The real challenge in polyurethane foam formation is to control the chemical and
physiochemical processes up to the point where the material finally sets. The sequence
and the rate of the chemical reactions are predominately a function of the catalyst and
the reactivity of the basic raw materials, polyol and isocyanate. The physiochemical
contribution to the overall stability and processability of a system is provided by the
silicone surfactants. Optimum foaming results will be achieved only if the correct
relationship between chemistry and physics exists [4].
4
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Polyurethane foams are prepared from the simultaneous reactions of diisocyanate with
water and with polymeric diols and/or triols to form hydrogen-bonded urea (hard)
segments and polyurethane networks (soft segments). The commercial production of
polyurethanes via isocyanate poly-addition reactions requires the use of one or more
catalysts. Tertiary amines are widely accepted in the industry as versatile polyurethane
catalysts. Amine catalysts are generally stable in the presence of standard polyurethane
formulation components and can have an impact on both the blowing (water-isocyanate)
and gelling (polyol-isocyanate) reactions. Although the use of catalysts in the
manufacturing of polyurethane foam both speeds the production of the foamed article
and, through the judicious choice of catalyst package, allows control of the physical
properties of the product, there are some problems associated with the use of these
additives. A number of commonly used tertiary amine catalysts can volatilise under certain
conditions. Release of tertiary amines during foam processing and from consumer products
is generally undesirable. Therefore, identifying alternatives to standard tertiary amine
catalysts which have no or low volatility, yet exhibit the same type of activity in isocyanate
poly-addition reactions, is desirable.
The non-fugitive catalysts reported in this chapter address the problems associated with
the use of polyurethane catalysts by reducing the odour and volatility of these materials
and by eliminating the ability of these additives to escape from finished foam products.
One strategy has involved functionalising the catalysts to render the species reactive
toward isocyanates, thereby covalently attaching the catalysts to the polymer network.
This strategy not only renders the catalytic material non-fugitive in the final product, but
also reduces the odour and volatility of the catalyst through increases in molecular weight
and polarity. These non-fugitive catalysts also provide equivalent or improved physical
properties when compared to industry standards, whereas conventional reactive amine
catalysts as well as metal catalysts cannot always meet todays ever increasing manufacturer
and consumer performance requirements.
These increasingly evolving requirements have led to the development of both novel
non-fugitive catalysts and new cell-opening non-fugitive catalysts for flexible foam. These
new low emission additives have been developed to meet the challenge of optimised
foaming and result in little or no emissions. Several of the non-fugitive catalysts possess
cell-opening capability. This new technology allows the manufacturer of polyurethane
foam to optimise their system to achieve the best processing latitude for their foam process.
These new additives maintain, or in some cases, improve key physical properties while
providing a more open foam.
5
Advances in Urethane Science and Technology
Data presented herein was derived from a combination of handmix and high pressure
impingement-mixing machine produced foam. Foams were prepared using several general
types of formulations for toluene diisocyanate (TDI) and two general types of formulations
for methylenediphenyl diisocyanate (MDI) which are representative of currently utilised
formulations in the automotive interior component industry. In addition, an all water
blown formulation was used to represent the flexible slabstock industry.
1.2.1 Materials
6
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Table 1 Continued
Trade name Formulation Manufacturer
XF-O11006 Newly developed non-fugitive cell opening APCI
gelling catalyst.
Dabco DC5169 Silicone copolymer surfactant for cold cure systems. APCI
Dabco DC5043 Silicone copolymer surfactant for TDI cold cure APCI
systems.
Dabco DC2585 Silicone copolymer surfactant for MDI cold cure APCI
systems.
Dabco DC2517 Silicone copolymer surfactant. APCI
Dabco DC2525 Silicone copolymer surfactant. APCI
Dabco DC5258 Silicone copolymer surfactant. APCI
XF-N1586 Newly developed silicone copolymer surfactant, APCI
which promotes open cells.
XF-N1587 Newly developed silicone copolymer surfactant, APCI
which promotes open cells.
DEOA-LF Diethanolamine Liquid Form APCI
(Dabco) (85% DEOA: 15% water)
Arcol E848 Conventional polyol with an OH# of 31.5 Lyondell
Chemical
Arcol E851 43% solids copolymer polyol with an OH# of 18.5 Lyondell
Chemical
NC-630 Polyol with an OH# of 31.4 Dow Chemical
NC-700 41% solids copolymer polyol with an OH# of 21.0 Dow Chemical
Voranol 3512 Polyether polyol with an OH# of 48.3 Dow Chemical
Polyol A High functionality triol with an OH# of 32.5 Dow Chemical
Polyol B 41% solids copolymer triol with an OH# of 23.8 Dow Chemical
Polyol C Polyether polyol with an OH# of 28 Dow Chemical
Polyol D Cell opening polyol
PRC-798 Solvent-based release agent Chem-Trend
APCI: Air Products and Chemicals, Inc.
DEOA: Diethanolamine
DPG: Dipropylene glycol
TEDA: Triethylene diamine
7
Advances in Urethane Science and Technology
Handmix experiments were carried out using the following procedure. Formulations
were blended together for approximately 10 minutes using a mechanical mixer
equipped with a 7.6 cm diameter, high shear, mixing blade, rotating at 5000 rpms.
Premixed formulations were maintained at 23 ± 1 °C using a low temperature
incubator. Mondur TD-80 (Bayer; a blend of 2,4-TDI and 2,6-TDI isomers in the
ratio of 4:1) or modified MDI was added to the premix at the correct stoichiometric
amount for the reported index for each foam. The mixture was blended together
with a Premier Mill Corporation Series 2000, Model 89, dispersator for approximately
five seconds. The foaming mixture was transferred to an Imperial Bondware #GDR-
170 food container or ‘chicken’ bucket and allowed to free rise in order to obtain the
processing data.
Handmix experiments were carried out using the following procedure. A premix
consisting of polyol, surfactant and water was prepared by blending the components
in a shaker for approximately 20 minutes. The premix was allowed to stand for 2
hours prior to making the foam to allow for degassing of the mixture. A measured
amount of premix was poured into a 1.9 litre paper cup; the required stoichiometric
amounts of amine and tin catalysts were added to the contents of the cup and mixed
for 20 seconds using a Premier Mill Corporation dispersator equipped with a 5.5 cm
diameter, high shear, mixing blade, rotating at 6,000 rpm. The corresponding amount
of Mondur TD-80 to provide for a 110 index (isocyanate index, which is the amount
of isocyanate used relative to the theoretical equivalent amount [5]) was measured
into a 400 cm3 beaker. Methylene chloride in the correct proportion was added to
the beaker containing the Mondur TD-80; the beaker was carefully swirled for 4 or
5 seconds and the contents poured into the paper cup. The mixture was blended
together for 6-7 seconds and the foaming mixture poured into a paper bucket for up
to 12 seconds and allowed to free rise with the processing data being recorded.
Reactivity profiles were determined from hand-mix foams prepared in 5.68 litre paper
buckets. Foams for physical properties were prepared in 35.6 x 35.6 x 25.4 cm
cardboard boxes. Identical procedures were followed for both reactivity and physical
property experiments.
8
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Machine runs for the TDI flexible moulded foam were carried out on a Hi Tech SureShot
MHR-50 (Hi-Tech Industries, Inc.), cylinder displacement series, high pressure machine.
Fresh premixes, consisting of the appropriate polyols, water, crosslinker, surfactants and
catalysts for each formulation were charged to the machine. Mondur TD-80 was used
throughout the entire study. All chemical temperatures were held at 23 ± 2 °C via the
machine’s internal temperature control units. The foam was poured into an isothermally
controlled, heated aluminium mould maintained at 71 ± 2 °C. The mould was a typical
physical property tool designed with internal dimensions of 40.6 cm x 40.6 cm x 10.2
cm. The mould has five vents, each approximately 1.5 mm in diameter, centred 10.0 cm
from each edge and the geometric centre of the lid. The mould was sprayed with a
solvent-based release agent, Chem-Trend PRC-798, prior to every pour and allowed to
dry for one minute before pouring. The foam premix was puddle poured into the centre
of the mould with a wet chemical charge weight capable of completely filling the mould
and obtaining the desired core density. Minimum fill requirements were established for
each formulation evaluated. The foam article was demoulded at 240 seconds after the
initial pour. After demoulding, the foam was placed through a mechanical crusher, tested
for FTC measurements, or left uncrushed and set aside for 24 hour shrinkage
measurements described in Section 1.2.3.2c.
All foams to be tested in each catalyst set were mechanically crushed 1 minute after
demoulding using a Black Brothers Roller crusher set to a gap of 2.54 cm. Crushing was
carried out three times on each part, rotating the foam 90 degrees after each pass through
the roller. All parts produced for physical testing were allowed to condition for at least
seven days in a constant temperature and humidity room (23 ± 2 °C, 50 ± 2% relative
humidity).
Three to four specimens were produced for any given set of conditions. Four test specimens
were die-cut from each foam pad and evaluated for each physical property listed in
subsequent data tables. All results were included in calculating averages and standard
deviation. Each test was carried out as specified in ASTM D3574 [5].
For each formulation evaluated, duplicate free rise ‘chicken’ buckets were poured at the
same shot size to determine overall reactivities and foam shrinkage. Data recorded were
cream time (the time between the discharge of the foam ingredients from the mixing
head and the beginning of the foam rise [5]), top-of-cup (TOC; the time between the
discharge of the foam ingredients from the mixing head and when the centre of the foam
reaches the same height as the top of the chicken bucket), string gel (the time between
9
Advances in Urethane Science and Technology
pouring of the mixed liquids and the time that strings of viscous material can be pulled
away from the surface of the foam when it is touched with a tool [5]), full rise time and
final height. The free rise buckets were again tested for final heights after 24 hours.
Measurements of height were made using a Mitutoyo height gauge. In addition to all the
standard tests, several more unique tests were performed where indicated, and are
described in Section 1.2.3.2.
1.2.3.2 Tests
A common type of isothermally heated mould was used to determine the flowability of
formulations with each of the catalyst candidates. This maze mould is shown in Figure 1.1.
Machine foam was poured into the mould at the top left corner of the open cavity as
indicated by ‘pour spot’ on the figure. The lid was then closed and clamped tightly. Foam
was allowed to free flow consecutively through each of the five gates for the standard 4
10
Dimensional Stabilising Additives for Flexible Polyurethane Foams
minutes prior to demould. Minimum fill was first determined by completely filling the
cavity with little or no extrusion through the vent at the end of the fifth gate. Mathematical
reduction of the shot size was performed to obtain the first of three systematically scaled
down foam fill weights. This first foam should have a fifth leg (the foam in gate 5 of the
maze flow mould, see Figure 1.1) which barely touches the front cavity wall. The second
reduction in foam fill weight produced a foam that flowed approximately halfway through
the fifth gate. The third reduction in foam fill weight was equivalent to the step change
from foam 1 to foam 2. Shot times were held constant for each of the three foam fill
weights as compared to the control determined standard shot time in any given solids
level formulation. These three foams were weighed for total foam pad and fifth leg weight,
and measured for fifth leg length to obtain a range of flow values for each of the
experimental catalysts compared to the control additives.
Foam dimensional stability is essentially the result of a balance between external and
internal forces. The external forces are defined as the ambient pressure along with any
additional applied loads. The internal forces are the strength of the polymer matrix and
the internal cell pressure [6]. Basically, if the sum of the internal forces is greater than the
external forces, the foam will expand. Consequently, if the sum of the external forces is
greater than the internal forces the foam will shrink. Any expansion or shrinkage will
impact on the internal and/or external forces until an equilibrium is obtained. It is the
internal forces, i.e., cell pressure and strength of the polymer matrix as defined by ‘green
strength’ or cure, which will have an impact on the dimensional stability performance of
the moulded polyurethane.
11
Advances in Urethane Science and Technology
foam will exhibit little or no tendency to shrink after demoulding. Poor dimensional
stability can result in numerous defects of the polyurethane article, such as lack of fit of
a polyurethane piece to the substrate. These defects will ultimately cause loss of revenue
to the polyurethane manufacturers because of increased repair and/or scrap rates.
Additionally the degree of cell openness of polyurethane foam can be measured directly
by the air flow physical properties of the polyurethane part. Higher air flow values
measured for a particular foam would indicate that the foam has less of a tendency to
shrink and therefore be more dimensionally stable as compared to a foam with lower air
flows. Additionally, higher air flows may also indicate that the foam was much easier to
crush-out thereby breaking many of the cell windows. Dimensionally stable foam should
reduce scrap and rework by allowing the foam to conform to near its original moulded
shape or at least return to its original shape after being crushed.
12
Dimensional Stabilising Additives for Flexible Polyurethane Foams
It was designed to measure the average foam pad shrinkage of an uncrushed foam. This
apparatus consists of two 432 mm long x 432 mm wide x 6.35 mm thick, plexiglass
plates, mechanically pinned in each corner with threaded bolts to maintain the plates at
a constant 102 mm spacing. The single uncrushed foam from each of the catalyst sets
was aged 24 hours prior to being placed between the two plates. Both top and bottom
plates each contain 5 mm diameter holes evenly spaced, diagonally from corner to corner,
25 mm apart in an X-shaped pattern. Nineteen holes are contained in each leg of the X,
for a total of 37 holes per plate. Measurements were made with a digital caliper by
inserting the end down through each hole to just touch the foam surface with the indicated
value being recorded. All measurements were normalised to discount the plexiglass plate
thickness and subsequently averaged to a single mould cavity and lid value.
Time Pressure Release (TPR) is the opening of the mould during the curing cycle to
release the internal pressure and then re-closing for the duration of the cure time [8]. The
sudden release of the internally generated pressure bursts the cell windows, thereby
improving the crushability of the foam. The tool is opened only a few millimetres and
for a specific time. TPR can be applied at any time during the curing cycle, however, care
must be taken not to perform the operation too early or too late since surface quality
issues may occur.
A ‘simulated’ TPR process was carried out during this study, whereby the tool lid was
opened approximately 1.5 mm for a three second duration. TPR was applied at various
time intervals throughout the evaluation. Two mechanical clasps affixed to the top and
bottom halves of the tool precisely controlled the gap opening. These clasps were manually
opened and closed at the desired TPR time interval.
Another test was devised to evaluate foam bulk stability in a free-rise mixed foam.
Foams prepared from TDI formulations were poured directly from the machine head
into a large open pail (the pail is a common high density polyethylene plastic with
approximate dimensions of 365 mm high and 290 mm in diameter) at a targeted mass.
Several pours were carried out to ensure equivalent catalyst activity amongst each
formulation. Foams were allowed to stand for 24 hours prior to removal from the pail.
Each foam was weighed to obtain total individual mass. Subsequently, a 25 mm slice
was cut directly through the geometric vertical centre of the foam. Foam slices were
examined for cell structure.
13
Advances in Urethane Science and Technology
Dynamic Fatigue Constant Pounding testing was carried out using standard testing
procedures outlined in ASTM D3574-95 [6]. A 60 minute recovery time and 80,000
cycles were used for each test sample.
Gaseous foam emissions were compared in the standard fogging test procedure outlined
in SAE J1756 [9]. A standard 7.6 cm thick piece of foam was preconditioned for 48
hours, then tested in the Hart fogging apparatus at 100 °C for 3 hours. Glossmeter
readings of the foam were taken after 60 minutes at room temperature.
Foam emissions were also evaluated by cutting equivalent portions, approximately 1 gram,
of foam from the geometric centre of the moulded foam articles 60 minutes after demoulding.
Each sample was inserted into a 20 cm3 Kimble glass crimp top vial with a Teflon seal.
Several vials were sealed without foam to be used as blanks to ensure all emissions had
eluted through the gas chromatograph prior to the injection of a second gas sample. All
vials were loaded into a Tekmar 7000 Headspace Autosampler tray for sequential heating
for 1 hour at 54 °C. After heating and temperature equilibration, the headspace of the vial
was sampled and directly injected within a closed loop system onto a Hewlett Packard
5890 Series II Plus gas chromatograph containing a HP-5 (5%-diphenyl/95%-
dimethylsiloxane copolymer) stationary phase column (30 m, 0.25 mm internal diameter,
1.00 mm film thickness). A standard oven heating profile was used to separate the gas
components for detection with a Hewlett Packard 5972 Series Mass Selective Detector.
Elution peaks were individually identified by comparison to standard libraries.
Machine runs for the MDI flexible moulded foam were conducted on a Krauss-Maffei,
cylinder displacement series, high pressure machine. Fresh premixes, consisting of the
appropriate polyols, water, crosslinker, surfactants and catalysts for each formulation were
charged to the machine. Modified MDI was used throughout the entire study. All chemical
temperatures were held at 25 °C ± 2 °C via the machine’s internal temperature control
units. Foam pours were made into an isothermally controlled heated aluminium mould
maintained at 60 °C ± 2 °C. The mould was a typical physical property tool designed with
14
Dimensional Stabilising Additives for Flexible Polyurethane Foams
internal dimensions of 40.6 cm x 40.6 cm x 10.2 cm. The mould has two vents each
approximately 1.0 mm in diameter centred 10.0 cm from each edge and the geometric
centre of the lid. The mould was sprayed with a solvent-based release agent prior to every
pour and allowed to dry for one minute before pouring. The foam premix was puddle
poured 15 cm away from the geometric centre of the mould with a wet chemical charge
weight capable of completely filling the mould with the appropriate core density. Minimum
fill requirements were established for each formulation evaluated. The foam article was
demoulded at 300 seconds after the initial pour. Upon demoulding, the foam was placed
through a mechanical crusher or tested for FTC measurements.
The foams were mechanically crushed 1 minute after demoulding using a roller crusher set to
a gap of 3.0 cm. Crushing was carried out three times on each part. All parts produced for
physical testing were allowed to condition for at least seven days in a constant temperature
and humidity room (23 °C ± 2 °C, 50% ± 2% relative humidity). Three to four parts were
produced for any given set of conditions. Four test specimens were die-cut from each pad
and evaluated for each physical property listed. All results were included in calculating the
averages and standard deviation. Each test was conducted as specified in ASTM D3574 [5].
For each formulation evaluated, free rise cup foams (see Section 1.2.2.1) were poured to
determine reactivities and foam shrinkage. Data recorded were gel time, full rise time and
final height. The free rise cup foams were tested for final heights and free rise density after 24
hours. Height measurements were carried out using a Mitutoyo height gauge. All experimental
formulations reported in this work were matched by rise profile to each control formulation.
FTC measurements were conducted 90 seconds after demoulding. The foam pad was removed
from the mould, weighed and placed in the FTC apparatus (Instron 4502). The force detection
device is equipped with a 5.0 kN capacity pressure transducer. The actual force is shown on
a digital display. This device mimics the ASTM D3574, Indentation Force Deflection Test
and provides a numerical value of freshly demoulded foams initial hardness or softness. The
pad was compressed to 70 percent of its original thickness at a cross head velocity of 380 mm
per minute with the force necessary to achieve the highest compression cycle recorded in
Newtons. Values are reported as the FTC value for the foam based on the assumption that
the lower the FTC values the better the dimensional stability of the foam.
The automotive industry has placed increased pressures on OEM suppliers to improve
their productivity, quality and cost of the polyurethane articles which they produce.
Styling changes, complex designs and OEM productivity demands for automotive seats
have necessitated the need to produce more open foam in all-water blown TDI- and
15
Advances in Urethane Science and Technology
Flexible moulded polyurethane foam requires mechanical crushing to open foam cells,
which in turn prevents shrinkage and improves overall dimensional stability. Current
mechanical methods for cell opening consist mainly of roller crushing, vacuum rupture
and TPR. However, mechanical methods do not always result in complete or consistent
cell opening and require a flexible moulded foam producer to invest in additional
machinery. Additionally, if the polyurethane article is not crushed properly, dimensional
stability suffers which can cause an increase in repair and scrap rates resulting in a negative
impact on the cost of production. A chemical method for cell opening would be preferred.
When producing flexible high resilience foam, it is important to provide a wider TPR
window to expand processing latitude and at the same time maintain or improve physical
properties. This should result in reduced scrap and/or repair rates, providing improved
economics for the polyurethane producer.
The most commonly used catalyst and surfactant package for all water blown TDI-
based moulded foam production is a blend of Dabco 33-LV and Dabco BL-11, coupled
with Dabco DC5043 silicone surfactant. Additionally, the acid-blocked counterparts of
these two catalysts, Dabco 8154 and Dabco BL-17, can also be used for the production
of high resilience moulded foam. A combination of Dabco 33LV and Dabco BL-17 is
used to facilitate a short delay in the reactivity of the polyurethane foaming process. A
combination of silicone surfactants, Dabco DC5043 and Dabco DC5169, are utilised to
provide good foam stabilisation, improve cell regulation and cell wall drainage. These
combinations of catalysts and surfactants served as the control additives to which the
experimental additives were compared and contrasted.
A newly developed cell opening catalyst, Dabco BL-53, was evaluated to determine its
impact on dimensional stability and general processability. Dabco BL-53 affords all the
benefits of Dabco BL-11 or Dabco BL-17, with the added advantage of cell opening and
slightly delayed initiation times. Dabco BL-53 is not a chemical equivalent for Dabco
BL-11 or Dabco BL-17; however, it will provide similar performance. For rapid
demoulding systems, it is recommend that Dabco BL-53 be used at 0.12 to 0.22 pphp,
with the optimum level at 0.16 to 0.19 pphp, in combination with a Dabco 33-LV level
at 0.30 to 0.32 pphp.
16
Dimensional Stabilising Additives for Flexible Polyurethane Foams
A TDI cushion formulation, with a density of 45 kg/m3 and a TDI back formulation,
with a density of 35 kg/m3 were used in the TDI automotive study. All of these
formulations were modified accordingly with the appropriate crosslinker, water, and
additive levels for the chosen density range. These formulations are shown in
Tables 1.2 and 1.3.
17
Advances in Urethane Science and Technology
1.3.1.1 Reactivity
Formulations used in the mass-loss/rate-of-rise are summarised in Tables 1.2 and 1.3.
Surfactant and catalyst additives were changed according to the formulation being studied.
Foams were run at the optimum index as they were during the machine study. All experiments
were duplicated. Each mixed formulation was poured into a ‘chicken’ bucket equipped with
a thermocouple positioned at the centre of the bucket resting on a Mettler PM 30,000 balance.
The centre height of the rising foam was recorded every second using a DAPS QA Model
#2500 rate-of-rise apparatus. Knowing the foam mass, the rate-of-rise and using the ideal
gas law, it is possible to calculate the carbon dioxide generated or trapped over time [10, 11].
Figures 1.3 and 1.4 show height versus time achieved for the cushion and back formulations
using the control formulations I and V, as well as the experimental formulations II, III, IV,
VI, VII and VIII. The foam height versus time graphs clearly indicate higher rates for
control formulations I and V in both cushion and back formulations. Cushion and back
18
Dimensional Stabilising Additives for Flexible Polyurethane Foams
formulations, which contained the new additives in formulations II, III, IV, VI, VII and
VIII did not achieve the same foam height when compared to the control formulations.
Several things could cause the differences observed in the foam heights. First, reactivity
rates for control formulations might be faster than the experimental catalyst or surfactants.
Second, overall foam stability could be compromised for the experimental catalyst and
surfactants. Lastly, carbon dioxide might be diffusing from the reacting polyurethane foam
19
Advances in Urethane Science and Technology
at an accelerated rate. All of these possibilities were explored. Data generated and
observations will be reported which demonstrate how these new additives promote
dimensional stability through increased carbon dioxide diffusion.
20
Dimensional Stabilising Additives for Flexible Polyurethane Foams
conversion is occurring at the same rate in the control as in the experimental formulations.
The fact that carbon dioxide conversion is occurring at the same rates would not account
for the lower foam heights observed in Figures 1.3 and 1.4.
In the control formulations, I and V, the amount of carbon dioxide diffused was less than the
amount of carbon dioxide diffused using the new additives. Figures 1.7 and 1.8 were generated
using the ideal gas law from data generated with the mass-loss/rate-of-rise apparatus. Diffusion
21
Advances in Urethane Science and Technology
of carbon dioxide from cells as they form in the free rise reaction apparently keeps the
reacting foam from reaching the same foam height when using these experimental additives.
Reactivity profiles for these three new additives are essentially the same. Data discussed
later in Section 9.3.1.1b, Tables 1.4 and 1.5, further supports the fact that there is no
Table 1.4 Machine mix free rise reactivity comparison using cushion formulation
Control BL53 Exp. SSF BL53/Exp. SSF
Formulation I II III IV
Cream (Seconds) 4. 7 4.6 4.9 4.9
Top of Cup (Seconds) 24.1 23.6 23.9 23.4
String Gel (Seconds) 55.4 53.7 54.3 54.2
Full Rise (Seconds) 100.4 103.8 105.3 105.6
Rise: Gel Ratio 1.81 2.01 1.94 1.94
Final Height (mm) 278.4 279.6 280.8 272.6
24 Hour Height (mm) 263.4 265.5 265.5 257.6
% Shrinkage 5.4 5.0 5.4 5.5
Table 1.5 Machine mix free rise reactivity comparison using back formulation
Control BL53 Exp. SSF BL53/Exp. SSF
Formulation V VI VII VIII
Cream (Seconds) 4.8 4.5 4.8 4.8
Top of Cup (Seconds) 23.5 22.9 25.4 26.5
String Gel (Seconds) 53.2 52.1 55.6 55.9
Full Rise (Seconds) 109.5 108.7 111.9 111.2
Rise/Gel Ratio 2.06 2.08 2.02 1.99
Final Height (mm) 273.6 275.7 268.3 270.1
24 Hour Height (mm) 245.6 246.9 236.1 235.9
% Shrinkage 10.2 10.4 12.0 12.7
22
Dimensional Stabilising Additives for Flexible Polyurethane Foams
The machine mix free rise reactivity comparison of all formulations are shown in
Tables 1.4 and 1.5. This experimental data illustrates that the overall free rise foam
reactivity for both the cushion and back formulations remains relatively the same for
the beginning of the reaction. The full rise reactivities in cushion formulations II, III
and IV and back formulations VII and VIII start to deviate slightly from the control
reference formulations I and V.
Percent shrinkage remains fairly consistent within the cushion formulations, I, II, III and
IV. Increased foam shrinkage was observed with back formulations VII and VIII. This
could be attributed to better cell wall drainage efficiency, providing more open foam
and/or an overall increased carbon dioxide diffusion through the polymer network.
When producing polyurethane, manufacturers use some type of mechanical crushing to open
cells and insure the polyurethane article does not lose dimensional stability. Several techniques
can be used to provide the needed mechanical cell opening. Manufacturers will use TPR,
which has been described in Section 1.2.3.2d, along with mechanical roller crushing. Some
producers will rely exclusively on the roller crushing and vacuum crushing techniques to
provide the mechanical cell opening required. In both cases reducing FTC values and improving
foam openness is important for producing polyurethane articles that are dimensionally stable.
If TPR is carried out too soon during the polyurethane moulding cycle, the article will
collapse (blowout) as indicated in Figure 1.9. This is indicative of the foam being
insufficiently cured or lacking enough green strength when TPR was applied. If TPR is
conducted too late in the manufacturing process scalloping (concave surface areas on the
foam article) and tight foam (insufficient number of open cells within the foam article
that causes the hot gas to be trapped and upon cooling forces the entire foam part to
shrink) may also occur. When scalloping occurs the foam article must be repaired or
scrapped. When tight foam occurs dimensional stability will suffer and there will be a
negative impact on physical properties as denoted in Figure 1.10. The foam pad in Figure
1.10 was produced at a 140 second TPR without crushing using formulation V.
23
Advances in Urethane Science and Technology
24
Dimensional Stabilising Additives for Flexible Polyurethane Foams
To understand the benefits of these new additive technologies that provide reduced FTC
values, a TPR range from 70 to 150 seconds was run for each of the formulations in
Tables 1.2 and 1.3. At a 70 second TPR all formulations suffered blowout since the foam
was not sufficiently cured and thus lacked green strength. At an 80 second TPR, no
formulations evaluated experienced blowouts or collapse; however, slight distortions
and imperfections were evident on the foam surfaces to varying degrees of severity. To
complete the TPR window, TPR cycle times were continually ramped up in this fashion
to determine the upper limit at which TPR could be applied for each formulation. The
upper limit is reached for a given formulation when the foam displays the obvious signs
of scalloping and/or ‘dishing’ (concave surface areas of the foam). When this occurs the
foam is usually very tight and cannot be used as a functional part. Additionally, parts
were produced without utilising TPR during the production cycle in order to compare
the difference in foam crushability when TPR is used.
Cushion control formulation I, which is listed in Table 1.2, was evaluated at a 90-100
second TPR. Initial FTC values of 156 N/323 cm2 for a 90 second TPR and 165 N/323
cm2 for an 100 second TPR were observed. These values were acceptable and produced
foam parts of good quality. The new additives in formulations II, III and IV produced
maximum initial FTC values of 160 N/323 cm2 at TPR of 90 to 100 seconds. Foam
produced at an 80 second TPR for the control formulation I and formulations II, III and
IV containing the new additives resulted in minor problems with foam quality, i.e.,
scalloping. At a 70 second TPR, control and experimental formulations failed because of
severe blowout. Figures 1.11 and 1.12 show no significant difference of FTC for all
formulations evaluated.
When TPR values were increased to 120 seconds for control formulation I, initial values
increased to 623 N/323 cm2 and scalloping or foam quality suffered (Figure 1.13).
However, increasing the TPR time for cushion formulations II, III and IV to 120 seconds,
produced maximum initial FTC values of 205 N/323 cm2 (Figure 1.13). Foam surface
quality was very good. Increasing the TPR to 140 seconds increased initial FTC values to
a maximum of 543 N/323 cm2 and 534 N/323 cm2 for formulations II and III, respectively
(Figure 1.14). Foam quality was still very good. Formulation IV, which utilises a
combination of both surfactant and catalyst technologies, achieved a lower initial FTC
value at a 140 second TPR of 191 N/323 cm2. This was only slightly higher than the FTC
values of the control formulation at a 90 second TPR. Figures 1.11, 1.12, 1.13, and 1.14
show the results of FTC through the entire TPR range applied in this study. These figures
clearly demonstrate that use of the new additives can reduce FTC values and maintain
dimensional stability over the applied TPR range.
25
Advances in Urethane Science and Technology
26
Dimensional Stabilising Additives for Flexible Polyurethane Foams
27
Advances in Urethane Science and Technology
Figures 1.15, 1.16, and 1.17 further demonstrate the reduction of FTC values that can
be achieved when the new catalyst and surfactant technologies are utilised. These figures
clearly show a large difference achieved from plotting initial FTC to the final FTC cycle
versus increasing TPR cycle times. Moreover, the lower delta FTC values obtained indicate
an improved crush-out capability. The optimum TPR range for control formulation I is
90-100 seconds, while the optimum range for formulations II and III is increased to 120
Figure 1.16 Delta Difference of 1st to 3rd FTC versus TPR Time - Cushion Formulation
28
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Figure 1.17 Delta Difference of Initial to Final FTC versus TPR Time - Cushion Formulation
Back formulation evaluations were carried out in the same manner as the cushion
formulation study. Because of lower solids and higher water content necessary to obtain
specific densities and physical properties, this system was more sensitive to processing
and TPR range than the cushion formulation.
Back formulations V, VI, VII and VIII found in Table 1.3, were all run at TPR times of
70-150 seconds and no TPR. Collapsed foam was encountered at TPR times of 70 seconds
for all formulations. FTC values for the back control formulation V had an initial FTC
value of 138 N/323 cm2 at a 90 TPR and 160 N/323 cm2 at a 100 second TPR time.
Formulations VI, VII and VIII produced a maximum initial FTC value at a 90 and 100
second TPR of 138 N/323 cm2. Control formulation V and experimental formulations
VI, VII and VIII provided acceptable FTC values at 90 and 100 second TPR times (Figures
18 and 19). No significant difference in FTC was realised with these formulations at
these TPR times. The foam quality at this TPR range was also very good. When TPR
times were increased to 120 seconds for control formulation V, initial FTC increased to
645 N/323 cm2. Experimental formulations VI, VII and VIII maintained acceptable initial
29
Advances in Urethane Science and Technology
FTC values of 334 N/323 cm2, 245 N/323 cm2 and 156 N/323 cm2, respectively, (Figure
1.20). At a 120 second TPR, control formulation V exhibited scalloping while the
experimental formulations VI, VII and VIII continued to produce parts with good foam
quality. Initial FTC values for the control formulation V increased to 1,045 N/323 cm2
at a 140 TPR and severe shrinkage was observed. FTC values for formulations VI and
VII reached maximum values of 649 N/323 cm2, respectively, and 627 N/323 cm2 at a
30
Dimensional Stabilising Additives for Flexible Polyurethane Foams
140 second TPR (Figure 1.21). Foam produced for these two formulations at a 140
second TPR still produced acceptable quality foam. Formulation VIII achieved lower
FTC values than any of the back formulations evaluated at a 140 second TPR range. The
initial FTC value for formulation VIII was 405 N/323 cm2 with foam surface quality
maintained (Figure 1.21). Figures 1.18, 1.19, 1.20, and 1.21 show the results of FTC
values through the entire TPR range utilised in this study.
31
Advances in Urethane Science and Technology
Figures 1.22, 1.23, and 1.24 further demonstrate the reduction of FTC values that can
be achieved in the back formulations when the new catalyst and experimental silicone
surfactant technologies are utilised. At a 90 and 100 second TPR, control formulation V
produces good quality foam. When the TPR is increased above 100 seconds foam tightness
and surface quality issues arise with the control formulation. Formulations VI, VII and
VIII yield good quality foam at all TPR cycle times used in this study. As in the cushion
formulation, these lower FTC delta values demonstrate how these newly developed
additives enhance crush-out capabilities. Furthermore, this effectively illustrates that the
TPR window can be extended as compared to the control formulation.
Figure 1.23 Delta Difference of 1st to 3rd FTC versus TPR Time - Back Formulation
32
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Figure 1.24 Delta Difference of Initial to Final FTC versus TPR Time - Back Formulation
Several polyurethane manufactures do not utilise TPR in their production process. The
formulations evaluated in this study were designed to utilise TPR when producing
polyurethane articles. However, indications of reduced force required to crush a pad
using roller crushing and/or vacuum rupture can be understood from the data generated
when TPR is not applied to the process. The initial FTC values recorded for the control
cushion and back formulations I and V were 1,379 N/323 cm2 and 1,299 N/323 cm2,
respectively. Using formulations II, III, VI and VII, FTC values were reduced to 1,152 N/
323 cm2 and 1,090 N/323 cm2 for the cushion formulation and 1036 N/323 cm2 and
1023 N/323 cm2 for the back formulation. Only minimal shrinkage was observed with
the surface quality being maintained. Using formulations IV and VIII, FTC values could
further be decreased to 1032 N/323 cm2 and 943 N/323 cm2, respectively. Again, foam
surface quality was good with only minimal shrinkage being observed. Figures 1.25 and
1.26 show the FTC values obtained when no TPR was applied during this study.
Physical properties were evaluated for all the formulations found in Tables 1.2 and 1.3
with TPR being applied at various times throughout the moulding cycle as previously
discussed. Additionally, physical properties were evaluated with no TPR being applied.
The 90 second TPR time was chosen to be the minimum time to perform TPR for all
formulations. Physical property pads were produced at this TPR time. Extended TPR
33
Advances in Urethane Science and Technology
times were achieved with the experimental formulations. Physical property pads for
physical testing were produced at these extended times.
Various TPR windows were identified for the formulations outlined in Tables 1.2 and 1.3.
The control formulations I and V both had an upper TPR range of 100 seconds. At a 110
second TPR both formulations I and V had scalloping and surface distortion issues (see
Section 1.3.1.2a). Thus the effective TPR range for formulations I and V is 90 to 100
34
Dimensional Stabilising Additives for Flexible Polyurethane Foams
seconds. This provides a very narrow process range when utilising TPR with these specific
formulations. However, when using the Dabco BL-53 catalyst, or X-N1586/X-N1587
experimental silicone surfactants, both cushion and back formulation TPR times can be
extended well beyond the 100 second TPR upper limit of the control formulations I and V.
Tables 1.6 and 1.7 provide the physical property comparison for all formulations I-VIII
at a 90 second TPR time. The data clearly demonstrates that physical properties are
maintained, and in several cases improved, compared to the control formulations. For
example, airflow can be improved by as much as 20% as compared to both cushion and
back control formulations, when using Dabco BL-53 and experimental silicone surfactants
35
Advances in Urethane Science and Technology
36
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Tables 1.8 and 1.9 illustrate the physical properties for selected formulations at TPR
times of 130 and 150 seconds. Control physical properties were not evaluated at these
extended TPR times since the control formulations I and V had visual surface distortions
and severe scalloping. The data in Tables 1.8 and 1.9 demonstrate that the TPR window
can be extended by using these newly developed additives without negative impact to the
physical properties. In fact, the data indicates that several of the physical properties for
the experimental formulations exceed the control formulation properties at the 90 second
TPR time. For example, airflow measurements are improved when utilising the extended
TPR times. Improvements are greater than 10% with the Dabco BL-53 catalyst and
experimental silicone surfactant combinations. Additional improvements are also observed
with wet set and 50% humid aged compression set values.
Table 1.8 Physical properties at 130 TPR and 150 TPR: cushion formulation
Physical Properties 130 TPR 150 TPR
BL53 Exp. SSF BL53/Exp. SSF
II III IV
ILD (N) AVG SD AVG SD AVG SD
25% 175 1.8 164 3.5 174 1.1
65% 434 2.5 457 1.0 467 4.5
25% Return 138 1.8 143 1.5 152 1.1
Ball Rebound (%) 61 0.5 62 0.5 61 0.4
Airflow (SLM) 43 3.2 46 4.7 53 7.4
3
Density (kg/m ) 43 0.8 42 0.8 43 0.7
Tensile (kPa) 174 4.1 180 5.1 170 3.5
Tear (N/m) 255 7.4 274 10.5 279 12.1
Elongation (%) 180 3.9 180 3.5 173 3.3
Wet Set (%) 21 1.5 19 1.3 19 0.3
50% Compression Set 5 0.5 5 0. 1 4 0 .3
50% Humid Aged 10 0.8 10 0.8 8 0.3
Compression Set
37
Advances in Urethane Science and Technology
Table 1.9 Physical properties at 130 TPR and 150 TPR: back formulation
Physical Properties 130 TPR 150 TPR
BL53 Exp. SSF BL53/ BL53/
Exp. SSF Exp. SSF
VI VII VIII VIII
ILD (N) AVG SD AVG SD AVG SD AVG SD
25% 111 2.1 145 4.2 156 0.7 155 1.2
65% 344 5.1 396 3.5 435 8.0 430 3.0
25% Return 96 2.2 123 2.2 132 0.6 129 2.1
Ball Rebound (%) 59 1.4 59 1.5 62 0.7 60 0.2
Airflow (SLM) 49 5.5 45 2.3 49 1 .8 52 2.1
3
Density (kg/m ) 35 0.8 36 1.0 36 0.8 36 0.2
Tensile (kPa) 136 5.1 148 5.4 152 7.9 147 6.1
Tear (N/m) 254 9.1 261 8.0 248 8.0 239 5.8
Elongation (%) 170 9.3 166 5.3 163 5.6 175 4.9
Wet Set (%) 22 1.1 19 0.8 19 0.5 19 0.8
50% Compression Set 9 0.2 5 0.3 4 0.2 4 0.6
50% Humid Aged 10 0.5 10 0.6 9 0.5 9 0.8
Compression Set
Tables 1.10 and 1.11 illustrate the physical property performance of all formulations I
through VIII without utilising TPR. Given that the specific formulations were designed
to be used with the TPR process, foam pads produced under these conditions were very
tight upon demoulding. Mechanical crushing at 30 seconds after demould was required
instead of the one minute time frame; otherwise, irreversible shrinkage may have occurred
prior to placing the foam through the roller crusher. The data in Tables 1.10 and 1.11
indicate that utilising Dabco BL-53 catalyst and/or the experimental silicone surfactants
X-N1586 and X-N1587, have a positive influence on the physical properties even when
no TPR is applied. Moreover, if the properties of the new additives are compared to the
control formulations I and V at the 90 second TPR time (Tables 1.5 and 1.6), physical
properties are not adversely affected. In some instances the properties are still maintaining
their improvement over the control formulations. Control formulations run at no TPR
exhibited surface quality distortions.
38
Dimensional Stabilising Additives for Flexible Polyurethane Foams
All of the physical property data generated at the higher TPR cycle times illustrates that the
window for TPR can be extended when using Dabco BL-53 catalyst and X-N1586, X-
N1587 experimental silicone surfactants. Physical properties were not adversely affected
when compared to the control samples. By extending the effective TPR window, polyurethane
moulders who utilise this process will enjoy a greater latitude for their processing, providing
more freedom to troubleshoot difficult tooling and moulding operations. Additionally,
polyurethane manufacturers that do not practice the TPR process can also benefit with
improvements realised for physical properties and dimensional stability, as illustrated by
many of the physical properties enhancements found in Tables 1.9 and 1.10.
Finally, the question remains whether the stability, especially shear stability, suffers from the
significantly improved breathability and dimensional stability provided with these new
additives. Examination of cellular structures of the machine mixed free rise ‘chicken’ bucket
39
Advances in Urethane Science and Technology
foams demonstrated no evidence of elongated cells or side shear in the outer edges of the cup.
Furthermore, when the TPR ramp study was being carried out it was determined that all
formulations had similar lower limit TPR times of approximately 90 seconds. Essentially,
this early TPR time would mimic a poorly sealed mould so that the shear forces become
obvious in and around the gap or space created by opening the mould. The foam mixture, if
not sufficiently stabilised or adequately cured, would be accompanied by a pressure relief
which leads to significant mechanical stress, causing collapse. The fact the Dabco BL-53
catalyst, X-N1586 and X-N1587 experimental silicone surfactants did not have blowouts or
shear collapses at the 90 second TPR illustrates shear stability is not compromised.
A scanning electron microscopy study of moulded foam parts was carried out to understand
what affect the additives might have on the morphology of the foam cell structure. It is
clear from the SEM photomicrographs in Figures 1.27, 1.28 and 1.29 that there is no
40
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Figure 1.27
Formulation I, Control
(Reproduced with permission
from APCI)
Figure 1.28
Formulation II, BL-53
(Reproduced with permission
from APCI)
Figure 1.29
Formulation III, Exp. SSF
(Reproduced with permission
from APCI)
41
Advances in Urethane Science and Technology
visual effect on the cell structure of foam produced with the new additives as compared to
the control. The additives did not appear to have an affect on the cell morphology of a
crushed foam. However, it was not possible to determine differences in cell morphology
from an uncrushed part, since the parts were mechanically crushed to prevent normal
distortion from foam shrinkage. Formulations I, II and III were used for the SEM study.
Another emerging need in the automobile market segment is for lower and ultimately no
emissions of polyurethane foam additives into the environment. Current efforts revolve
around automotive passenger compartment air quality improvements by reduction of
the emissions typically from the current migratory additives used to produce interior
foam components. These emissions can cause a variety of problems in the final application
of the foam, ranging from odour to vinyl staining and window fogging. The migratory
nature of current additives is accelerated by sunlight and heat. Thermally induced
dehydrohalogenation of the foam/covering interface is assisted by the migration of fugitive
tertiary amine catalysts causing vinyl staining (colour change). Breakdown of the vinyl
clad covering subsequently releases other chemical compounds, such as plasticisers, which
then contribute to window fogging. Consumers and automakers are becoming more
conscious of automotive interior air quality, interior vinyl degradation and film formation
(windshield fogging) [12].
This section of the chapter reports on the development of non-fugitive gelling and blowing
polyurethane catalysts to address the aforementioned emission issues. These catalysts
chemically bind to the polyurethane foam matrix (contain active hydrogens) rendering
them incapable of migrating back out of the foam after the reaction is complete, while
still producing a quality product. Equivalent or improved physical properties such as
airflow, compression set and crushability, must be maintained as compared to the industry
standards used in TDI-, MDI-based systems, and TDI/MDI blend systems. Presented in
this section is work using these non-fugitive catalysts in a variety of flexible polyurethane
foam applications, including MDI and TDI automotive and flexible slabstock
formulations. The results show that these new additives can give dimensional stability
and processability to the foam, while providing non-fugitive and non-fogging benefits.
Additionally, improvements to the physical properties of the polyurethane article when
using the new additives, occurs.
42
Dimensional Stabilising Additives for Flexible Polyurethane Foams
XF-N1085 is a novel cell opening blowing, non-fugitive catalyst which affords all the
benefits of the well known Dabco BL-11, with the added advantage of cell opening and
no foam emissions. XF-N1085 is not a drop in replacement for Dabco BL-11. For rapid
demoulding of TDI systems, it is recommended that XF-N1085 be used at 0.2 to 0.4
php, with the optimum level at 0.25 to 0.35 pphp, in combination with an XF-O11006
level at 0.6 to 0.9 pphp. XF-O11006 is a novel cell opening gelling non-fugitive catalyst
which gives all the benefits of the well known Dabco 33-LV, with the added advantage of
cell opening, low volatility and no foam emissions. XF-O11006 is not a drop in
replacement for Dabco 33-LV.
Several representative industry cushion and back formulations, listed in Tables 1.12-
1.15, were utilised to compare multiple additives. Dabco NE200 was compared against
43
Advances in Urethane Science and Technology
44
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Dabco BL-11 for equivalent blowing efficency. Dabco NE1060 was compared against
Dabco 33LV for equivalent gelling efficiency. Required blow:gel ratios are listed in each
table. Additionally, XF-N1085 and XF-O11006 were specifically developed as
dimensional stability/cell opening additives and they are also compared here.
1.3.2.1 Reactivity
Tables 1.16 and 1.17 illustrate that the overall free rise foam reactivity for both the cushion
and back formulations remains relatively the same for the entire foaming reaction. This
was achieved by balancing the experimental catalyst levels, blow:gel ratios, and matching
them to the control Dabco BL-11 and Dabco 33LV levels in IX, XIII, XVII and XXI.
45
Advances in Urethane Science and Technology
Table 1.16 Machine mix free rise reactivity comparison using TDI
cushion formulations
Formulation IX X XI XII XVII XVIII XIX XX
Cream 4.9 4.8 4.8 4.5 3.5 3.8 4.5 4.9
(Seconds)
Top of Cup 37 36 39 35 32 35 36 37
(Seconds)
String Gel 56 55 56 53 53 53 55 56
(Seconds)
Full Rise 83 81 84 78 71 75 81 83
(Seconds)
Table 1.17 Machine mix free rise reactivity comparison using TDI
back formulations
Formulation XIII XIV XV XVI XXI XXII XXIII XXIV
Cream 3.8 3.8 3.8 3.4 4.2 3.2 3.2 3.2
(Seconds)
Top of Cup 30 33 32 31 32 37 36 37
(Seconds)
String Gel 58 56 59 57 54 55 57 56
(Seconds)
Full Rise 82 85 94 91 75 86 85 88
(Seconds)
To understand the benefits of these new additive technologies that provide low volatility,
no amine emissions, no fogging and, in some cases, cell opening, all TDI moulded
formulations were run on the Hi Tech machine. Several foams were produced for each
catalyst combination and formulation to obtain physical property pads, FTC pads,
shrinkage pads, and flow evaluations.
46
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Tables 1.18 to 1.21 provide the physical property comparison for the TDI flexible moulded
formulations IX-XXIV. The data clearly demonstrates that physical properties are
maintained, and in several cases improved, compared to the control formulations,
depending on the formulation and the use of experimental cell-opening catalysts. For
example, ILD properties were comparable or improved over the control in both low and
high solids formulations, especially in the Dow polyol system. This is attributed to the
slightly different polymer matrix resulting from non-fugitive catalysis which can produce
higher load properties. In almost all cases, the support factor was comparable to the
control. Tensile values in the Dow system were matched to the control using XF-N1085
in combination with Dabco NE1060, formulations XIX and XXIII. Dry and humid
compression sets (50% dry, 50% humid aged, and 75% humid aged) were, in many
formulations, slightly elevated, with the Dabco NE1060/XF-N1085 catalyst combination
(XIX and XXIII) closest to the control. The compression set data illustrates that this
47
Advances in Urethane Science and Technology
property is very system dependent when using non-fugitive catalysts. For performance
optimisation, one may want to consider reformulation to produce the desired physical
properties. The use of XF-N1085 and XF-O11006 will improve or maintain physical
properties compared to the control for TDI systems. The Japanese wet set data was
determined to be comparable or, in the case of the Lyondell back formulation, slightly
improved when compared to the control foam.
The use of cell opening non-fugitive catalysts was found to improve wet sets in the
Lyondell high solids formulation. In addition, the cell opening catalysts were also found
to decrease hysteresis and increase tensile strength in the back formulations using Lyondell
polyol compared to the control formulation.
48
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Table 1.22 shows the dynamic fatigue test results for Lyondell and Dow cushion foams
tested for 80,000 cycles and 60 minute recovery time. Results indicate that equivalent
dynamic fatigue values are obtainable with use of the new non-fugitive and cell opening
catalysts, based on nominal test error of ± 3.
In both the Lyondell cushion and back formulations IX through XVI listed in Tables
1.11 and 1.12, addition of XF-N1085 and XF-O11006 result in a reduction in FTC
values (Tables 1.22 and 1.23) throughout the initial several repetitions of the crushing
cycle illustrated in Figures 1.30 and 1.31. At the end of the ten complete repetitions, all
three experimental catalyst packages, resulted in the same final FTC value. Without the
49
Advances in Urethane Science and Technology
50
Dimensional Stabilising Additives for Flexible Polyurethane Foams
51
Advances in Urethane Science and Technology
use of these cell opening catalysts, the Dabco NE1060/Dabco NE200 non-fugitive catalyst
package tightens the foam resulting in slightly elevated values during several initial FTC
footplate depressions.
In the cushion Dow formulations XVII through XX, the XF-N1085 and XF-O11006
show similar trends to the Lyondell formulations with reduced FTC (Figure 1.32 and
Table 1.24), thereby producing a more dimensionally stable foam. In contrast, for the
low solids back formulations XXI through XXIV, all catalyst combinations exhibited
no significant difference to the control catalyst (Figure 1.33 and Table 1.25). Inherent
tightness in this particular back formulation did not allow the same degree of catalyst
differentiation to occur in these foams. Adjustment to the catalyst loading levels would
be required to demonstrate the same reduced FTC performance as previously described
in other formulations.
Figures 1.34 and 1.35 further demonstrate the reduction of FTC values that can be
achieved with these new catalyst technologies. The graphs indicate the results achieved
by plotting the delta of the initial and third FTC value for every formulation. Moreover,
lower delta values illustrate an improved crush-out capability.
52
Dimensional Stabilising Additives for Flexible Polyurethane Foams
53
Advances in Urethane Science and Technology
54
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Figure 1.34 Delta Difference Graph of First and Third FTC Values in Lyondell System
55
Advances in Urethane Science and Technology
Figure 1.35 Delta Difference Graph of First and Third FTC Values in Dow System
Data in the Lyondell system, formulations IX through XVI, generated using the
previously described maze flow mould is summarised in Tables 1.27 and 1.28 and
56
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Table 1.27 Maze flow mould data for Lyondell TDI cushion formulations IX - XII
Formulation Identifier Pad wt. (g) Flow past 5th gate (g) % of Total pad wt.
MINIMUM 5th Gate
IX 405 30 8.0
X 404 30 8.0
XI 405 35 9.5
XII 409 30 7.9
MEDIAN 5th Gate
IX 474 101 27.1
X 472 104 28.3
XI 478 109 29.5
XII 479 105 28.1
MAXIMUM 5th Gate
IX 524 150 40.1
X 526 160 43.7
XI 531 161 43.5
XII 528 155 41.6
57
Advances in Urethane Science and Technology
Table 1.28 Maze flow mould data for Lyondell TDI back formulations XIII-XVI
Formulation Identifier Pad wt. (g) Flow past 5th gate (g) % of Total pad wt.
MINIMUM 5th Gate
XIII 619 92 17.5
XIV 617 102 19.8
XV 615 108 21.3
XVI 616 107 21.0
MEDIAN 5th Gate
XIII 667 143 27.3
XIV 666 153 29.8
XV 665 160 31.7
XVI 668 156 30.5
MAXIMUM 5th Gate
XIII 718 193 36.8
XIV 715 202 39.4
XV 716 206 40.4
XVI 718 208 40.8
shown graphically in Figure 1.37. The graph utilises a xy scatter plot with linear
regression trendlines and R2 values (coefficient of determination) reported for each
formulation group. It compares estimated and actual y-values and ranges in value
from zero to one. Since these values are very close to one, there is a near perfect
correlation in the samples. However, if the coefficient of determination approaches
zero, the regression equation is not helpful in predicting a y-value. The data trends
suggest that increasing part weight, regardless of formulation type, shows no significant
penalty in flow compared to the control formulations IX and XIII.
Similar maze flow mould results are also obtained with the Dow cushion and back
formulations summarised in Tables 1.28 and 1.29 and shown graphically in Figure 1.38.
In these Dow formulations, XVII through XXIV, the data suggests slightly improved
linear regression trendlines and R2 values. Again, no significant penalty in flow is observed
as compared to the control formulations XVII and XXI.
58
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Table 1.29 Maze flow mould data for Dow TDI cushion formulations XVII-XX
Formulation Identifier Pad wt. (g) Flow past 5th gate (g) % of Total pad wt.
MINIMUM 5th Gate
XVII 404 40 11.0
XVIII 403 36 9.8
XIX 399 34 9.3
XX 404 25 6.6
MEDIAN 5th Gate
XVII 457 93 25.6
XVIII 457 92 25.2
XIX 458 100 27.9
XX 453 69 18.0
MAXIMUM 5th Gate
XVII 509 149 41.4
XVIII 508 140 38.0
XIX 508 149 41.5
XX 513 136 36.1
59
Advances in Urethane Science and Technology
Table 1.30 Maze flow mould data for Dow TDI back formulations XXI-XXIV
Formulation Identifier Pad wt. (g) Flow past 5th gate (g) % of Total pad wt.
MINIMUM 5th Gate
XXI 568 65 12.9
XXII 566 70 14.1
XXIII 565 64 12.8
XXIV 569 83 17.1
MEDIAN 5th Gate
XXI 617 122 24.7
XXII 616 120 24.2
XXIII 614 112 22.3
XXIV 618 132 27.2
MAXIMUM 5th Gate
XXI 668 168 33.6
XXII 665 172 34.9
XXIII 666 164 32.7
XXIV 670 183 37.6
60
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Shrinkage results displayed in Figures 1.39 and 1.40 were generated with the shrinkage
apparatus previously described in the experimental section. Overall, the cavity shrinkage is
much greater as compared to the lid shrinkage, evidenced in both Lyondell and Dow polyol
systems. The use of Dabco NE1060/Dabco NE200 produced foams with more shrinkage
than the control catalysts. However, utilising non-fugitive cell opening catalysts, XF-N1085
and XF-O11006 resulted in shrinkage comparable to the control and in some cases even less.
As expected, less shrinkage was observed in both Dow and Lyondell cushion formulations
due to the increased solids loading as compared to their respective back formulations.
61
Advances in Urethane Science and Technology
Several machine foam samples were tested in accordance with fogging test method SAE
J1756-94 [9]. Results are listed in Table 1.31. The results obtained in the Lyondell polyol
system showed improved percentage reflectance over the control indicating reduced
emissions from the foam. Comparable results were obtained in all the Dow polyol system
samples regardless of catalyst use.
Headspace analysis in series with gas chromatography and mass spectrometry has
indicated no emissions from the experimental non-fugitive catalysts mentioned in this
paper. Emissions from the fugitive control catalysts were measured and identified. These
results support the fact that these experimental catalysts will result in no VOC emissions
which will ultimately yield no amine fogging and no vinyl staining from the final
polyurethane foam article. Additional volatility measurements have indicated that these
experimental catalysts volatilise 3 to 10 times less than the current industry standards
(Dabco 33LV and Dabco BL-11) in the temperature range of 0 °C to 150 °C.
Vinyl staining tests using commercial formulations and various grades and types of vinyl
have indicated that the experimental catalysts mentioned in this chapter do not contribute
to vinyl staining. Compared to industry standard amine catalysts, use of these experimental
non-fugitive catalysts will result in significant reduction (ΔE of less than 2) in staining
measurements [12]. DE is the numerical total colour difference, using lightness and
chromaticity factors, between a sample and a known colour standard.
62
Dimensional Stabilising Additives for Flexible Polyurethane Foams
The most commonly used catalyst packages for all water blown TDI flexible foam
production are blends of Dabco 33LV and Dabco BL-11 at a typical ratio of approximately
3:1. These new non-fugitive catalysts, Dabco NE1060, Dabco NE200, XF-N1085, and
XF-O11006 are recommended for TDI formulations at various gel:blow ratios depending
upon the selectivity desired. Ratios from 2:1 to 8:1 can be used to tailor the reaction
profile to the preferred selectivity for optimisation of physical properties and adaptation
to various processing conditions. The physical properties obtained with these new non-
fugitive catalysts are similar to the physical properties generated with standard industry
catalysts. However, slight formulation adjustments may be necessary to improve dry and
wet compression set values in certain instances. Improved crush-out capability can also
be achieved when incorporating the new cell-opening non-fugitive catalysts, XF-N1085
and XF-O11006. Complementary methods of emission analysis, headspace and fogging,
both suggest reduced emissions from the foam with these new non-fugitive catalysts.
The current trend toward lower foam densities, faster demould times, and the increased
use of complex metal and plastic insert frames in the construction of automobile interior
components has increased the difficulty of controlling the numerous process and chemical
variations a polyurethane manufacturer encounters on a daily basis. To successfully
produce moulded high resilience polyurethane foam, the manufacturer must maintain a
critical balance between foam over stabilisation, resulting in foam shrinkage, and under
stabilisation, resulting in internal defects, such as basal cell formation and shear collapse.
This often difficult to achieve balance is referred to as processing latitude. Variations on
either side of the processing latitude can result in costly increases for the producer in
scrap and repair rates. To help minimise production losses, polyurethane manufacturers
are essentially forced to over stabilise their foam formulations on a regular basis.
Subsequently, the closed cells which normally result from this over stabilisation are opened
via standard mechanical crushing techniques [13].
New surfactant technologies have been developed which promote improved cell wall
drainage and contribute to the final level of cell openness of the foam product without
causing any of the aforementioned negative attributes typical to a foam process. Due to
this unique cell opening action of these new additive technologies, the foam crushing
portion of the process can be greatly simplified or potentially eliminated.
Figure 1.41 shows a variety of flexible moulded Dabco surfactants which can be utilised
in TDI and/or MDI formulations. Using Figure 1.41, proper selection of the surfactant
can be made, enabling polyurethane formulators to expand their processing latitude, cell
openness, and bulk stability of the foam article in their critical formulations.
63
Advances in Urethane Science and Technology
An increasing North American trend in the moulded polyurethane foam industry today
is the use of MDI. While the advantages and disadvantages of MDI versus TDI are still
being debated, the fact remains that most polyurethane producers are using MDI in at
least part of their daily flexible foam production. The surfactant requirements for MDI-
based formulations can vary dramatically, dependent upon the type of MDI and polyols
being used. Most MDI formulations are inherently quite stable; therefore minimal
stabilising contributions are necessary from the surfactant. Conventional TDI high
resilience moulded surfactants are far too potent for MDI formulations. For example,
utilisation of Dabco DC5043 or Dabco DC5164 in an MDI seating formulation may
result in over stabilisation and significantly contribute to the shrinkage of the foam article.
64
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Data generated in the rate-of-rise comparison tests are summarised in Table 1.33. Foams
were run at an optimum index of 100 and all experiments were duplicated. Each mixed
formulation was poured into a ‘chicken’ bucket equipped with a thermocouple positioned
at the centre of the bucket resting on a Mettler PM 30 balance. The centred height of the
rising foam was recorded in millimetres every second using a DAPS (data acquisition
and plotting system) QA Model #2500 rate-of-rise aparatus. Reactivity profiles for these
new additives are essentially the same. Data discussed next further supports the fact that
there is no significant impact on reactivity or physical properties when comparing the
control to the Dabco DC2517 and Dabco DC5258 formulations.
65
Advances in Urethane Science and Technology
To understand the benefits of these new additive technologies that provide low volatility
and cell opening, all MDI moulded formulations were run in handmix foam to demonstrate
performance. Several foams were produced for each surfactant formulation to obtain
physical property pads, FTC pads, shrinkage pads, and flow evaluations.
Table 1.34 illustrates moulding foams at a density of 47-48 kg/m3; similar physical
properties are obtained for ILD, resilience, airflow, 50% dry sets and extensive properties.
However, slight nominal improvements in 50% humid aged compression sets (HACS)
and wet sets are achieved with Dabco DC2517 and Dabco DC5258.
The FTC graph shown in Figure 1.41 illustrates an approximate 14% decrease in initial
FTC values and continues throughout the remaining FTC cycles. This improvement in
FTC represents foam articles that will shrink less immediately after demoulding when
cooling of the trapped gas begins, allowing more time to complete the post crushing
66
Dimensional Stabilising Additives for Flexible Polyurethane Foams
procedure or potentially eliminating the procedure all together. Several foams were
prepared with each surfactant and allowed to cool for 24 hours in an uncrushed state, to
confirm the relationship between FTC and foam shrinkage. Foams prepared with Dabco
DC2517 and Dabco DC5258 showed significantly less shrinkage, approximately 10-
15% less, as compared to the control and Dabco DC2585. This is readily comparable to
the differences in FTC values shown in Figure 1.42.
The control catalyst combination for the MDI seating is a standard commercial catalyst
package. The balanced non-fugitive catalyst package Dabco NE1060/Dabco NE200 and
the low emission catalyst package Dabco NE1060/Dabco BL-11 use levels were set in
order to match performance to the commercial catalyst control package for the MDI
formulations. The silicone surfactant control was Dabco 2525. Cell opening non-fugitive
catalysts were not needed for this application based on the reported FTC values discussed
later in this chapter.
An MDI formulation, with a density of 40 kg/m3 and an MDI formulation, with a density
of 55 kg/m3 were used in the MDI automotive study with the formulations shown in
Table 1.35. For rapid demoulding of non-fugitive MDI systems, it is recommended that
Dabco NE1060 be used at 0.5 to 2.0 pphp, with the optimum level at 1.0 to 1.5 pphp
with 0.3 to 0.7 pphp of Dabco NE200. For low emission systems the authors recommend
Dabco NE1060 be used at 1.0 to 2.0 pphp, with the optimum level at 1.0 to1.6 pphp
with 0.03 to 0.1 pphp of Dabco BL-11.
67
Advances in Urethane Science and Technology
For complete comparison of these new additive technologies that provide low volatility, no
amine emissions and no fogging, all moulded formulations were run on a Krauss-Maffei
machine. Several foams were produced for each catalyst combination and formulation to
obtain physical property pads and FTC pads, shown in Tables 1.36 to 1.41.
Tables 1.36 and 1.37 provide the physical property comparison for the ~46 kg/m3 MDI
flexible moulded formulations XXV-XXXIV. The data clearly demonstrates that physical
properties are maintained, and in several cases improved, compared to the control
formulations, depending on the formulation and the use of experimental cell-opening
catalysts. For example, the data in Table 1.36 illustrates that at an index of 90, all
physical properties are comparable to the control formulation. When the index is increased
to 105, ball rebound and airflow are improved, with a slight decrease in indentation
force deflection (IFD) properties when the non-fugitive catalyst (XXXIV) is compared to
the control formulation (XXVI). When Dabco NE1060/Dabco NE200 is used (XXIX),
the physical properties are equal to the control at an index of 90. Improved Japanese wet
sets (see Section 1.3.1.2f) and slightly decreased IFD values are observed at an index of
105 compared to the control formulation (XXX).
68
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Table 1.36 Physical properties for 46 kg/m3 MDI moulded formulations at an index of 90
Physical Property XX V XXVI XXVII XXVIII XXIX
ILD (N) AVG SD AVG SD AVG SD AVG SD AVG SD
25% 159 - 159 - 14 9 - 15 5 - 15 5 -
65% 459 - 450 - 43 6 - 45 0 - 44 8 -
25% Return 127 - 126 - 11 9 - 12 3 - 12 3 -
Ball Rebound (%) 51 0.6 51 0.6 51 0.6 51 0.6 51 1.2
Airflow (SLM) 22.4 7.6 24.4 6.5 29 7.4 34 7. 9 17 5.0
Density (kg/m3) 46 3.4 44 1.0 43 2.4 46 3.7 45 0.9
Tensile (kPa) 129 10.6 298 14.4 125 6.7 133 5.9 127 9.9
Tear (N/m) 212 12 201 16 197 12 217 16 205 20
Elongation (%) 114 9.9 118 6.1 112 4.6 120 1.3 118 9.3
Wet Set (%) 13 0.6 16 1.2 10 0.5 13 0. 4 13 1.3
50% Compression 9 1.6 10 1.0 9 1.6 10 0.5 10 0.8
Set (%)
50% Humid Aged 13 0.5 13 2.0 13 1.5 12 1. 3 14 1.0
Compression Set (%)
Table 1.37 Physical properties for 46 kg/m3 MDI moulded formulations at an index of 105
Physical Property XX X XXXI XXXII XXXIII XXXIV
ILD (N) AVG SD AVG SD AVG SD AVG SD AVG SD
25% 275 - 260 - 264 - 261 - 251 -
65% 768 - 731 - 739 - 738 - 71 6 -
25% Return 209 - 197 - 201 - 198 - 190 -
Ball Rebound (%) 43 0.6 45 1.2 45 0.0 45 1. 0 47 1.2
Airflow (SLM) 40.8 8.5 31.4 7.1 38 27.2 33 9.9 46 8.5
Density (kg/m3) 46 1.1 47 4.4 45 0.6 46 3.2 44 1.3
Tensile (kPa) 176 9.8 409 19.2 173 6.8 168 6.3 169 11.7
Tear (N/m 263 8 252 4 25 6 12 279 16 244 12
Elongation (%) 105 4.2 109 4.1 102 6.7 105 4.3 103 6.2
Wet Set (%) 15 0.8 15 0.4 11 0.2 9 0. 7 17 0.2
50% Compression 8 0.6 9 1.3 8 0.3 10 0.9 9 1.0
Set (%)
50% Humid Aged 13 2.0 13 1.8 12 1.0 14 1. 7 14 1.0
Compression Set (%)
69
Advances in Urethane Science and Technology
Table 1.38 Physical properties for 55 kg/m3 MDI moulded formulations at an index of 90
Physical Property XXV XXVI XXVII XXVIII XXIX
ILD (N) AVG SD AVG SD AVG SD AVG SD AVG SD
25% 235 - 238 - 218 - 26 0 - 23 5 -
65% 637 - 636 - 59 7 - 731 - 64 1 -
25% Return 188 - 190 - 17 4 - 197 - 18 7 -
Ball Rebound (%) 46 0.6 44 1.0 46 0.6 45 1.2 42 0.6
Airflow (SLM) 21 4.0 35 11.9 35 5.4 31.4 7.1 27 12.5
Density (kg/m3) 53 3.4 50 3.6 50 3.5 47 4. 4 52 1.9
Tensile (kPa) 148 24.7 164 8.0 161 3.2 149 19.2 162 11.0
Tear (N/m 207 12 220 12 197 4 252 4 20 5 12
Elongation (%) 117 6.0 116 7.9 120 1.7 109 4.1 118 4.1
Wet Set (%) 11 1.4 13 0.9 15 1.3 15 0. 4 10 0.3
50% Compression 8 0.9 7 1.5 9 1. 0 9 1. 3 9 0.4
Set (%)
50% Humid Aged 11 1.5 14 2.1 13 1.3 13 1.8 12 0.9
Compression Set (%)
Table 1.39 Physical properties for 55 kg/m3 MDI moulded formulations at an index of 105
Physical Property XXX XXXI XXXII XXXIII XXXIV
ILD (N) AVG SD AVG SD AVG SD AVG SD AVG SD
25% 398 - 38 0 - 38 8 - 38 3 - 381 -
65% 1066 - 1029 - 1039 - 1042 - 1032 -
25% Return 303 - 288 - 295 - 289 - 289 -
Ball Rebound (%) 46 1.0 43 0.6 44 0.6 43 0.6 45 1.2
Airflow (SLM) 54 17.3 62.9 21.8 48 15.0 67 9.3 48 4.5
Density (kg/m3) 52 3.7 52 1.2 54 0.7 54 0. 4 55 0.5
Tensile (kPa) 217 17.0 218 7.0 207 12.9 217 6.6 223 10.5
Tear (N/m) 299 12 283 24 307 24 29 9 16 315 16
Elongation (%) 105 4.2 108 1.4 102 1.1 111 4.1 109 2.1
Wet Set (%) 14 1.4 12 0.7 13 0.7 11 0. 6 15 0.9
50% Compression 6 1.1 8 0.8 8 0. 6 9 1. 1 9 0.4
Set (%)
50% Humid Aged 14 1.9 14 0.8 13 0.9 12 1. 2 14 0.8
Compression Set (%)
70
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Tables 1.38 and 1.39 provide the physical property comparison for the ~55 kg/m3 MDI flexible
moulded formulations XXV-XXXIV. The data clearly demonstrates that physical properties
are maintained, and in several cases improved, compared to the control formulations, depending
on the formulation. For example, the data in Table 1.38 illustrates that all physical properties
are matched to the control formulation (XXV) at an index of 105 with the exception of a
slightly lower ball rebound for the full non-fugitive catalyst package. At an index of 105, all
physical properties for the experimental non-fugitive catalysts formulations were determined
to be similar to the control (XXX). When Dabco NE1060/BL-11 is used as a low emission
catalyst package (XXXIII), all physical properties at the index of 105 are comparable to the
control foam. At an index of 105 (XXXIII), airflow, elongation and Japanese wet sets and
HACS are improved over the control (XXX) and 50% dry compression sets are slightly elevated.
1.4.2.2 FTC
Tables 1.40 and 1.41 show numerical values which are shown graphically in Figures
1.43 and 1.44. FTC values are dramatically improved versus the control (XXV), shown
71
Advances in Urethane Science and Technology
in Figure 1.43, for the foams with an index of 90. FTC values for all of the formulations
at an index of 105 are very similar to the control (error is ± 6.4 N/323cm2), as illustrated
in Figure 1.44.
72
Dimensional Stabilising Additives for Flexible Polyurethane Foams
The non-fugitive catalyst package Dabco NE1060/Dabco NE200 and the low emission
catalyst combination Dabco NE1060/Dabco BL-11 closely match the entire reactivity
profile demonstrated by the standard control catalysts typically used in MDI flexible
foam. Additionally, the Dabco NE1060/Dabco NE200 combination demonstrates a higher
selectivity towards the blowing reaction as compared to a standard 2.5:1 ratio of Dabco
33LV to Dabco BL-11. This higher blowing selectivity can help improve cell opening and
produce a more dimensionally stable polyurethane article. These non-fugitive catalysts
also provide a reduction in the force necessary to crush freshly demoulded foam without
adversely affecting the remaining physical properties of the foam in MDI formulations
over a range of formulation indices.
Handmix rate of rise comparison tests were performed using the same equipment and
process as described for TDI and MDI hand mix foam. The formulations used are
listed in Table 1.42 and the rate of rise data reported in Table 1.43. Dabco BLV with
Dabco T-9 was used as the industry standard catalyst control for the water blown
flexible slabstock foam formulations. The silicone surfactant used in the reported flexible
slabstock formulations was DC5160. All water blown formulations were made at
densities of 23 kg/m3.
73
Advances in Urethane Science and Technology
1.5.1 Reactivity
Formulations from Table 1.42 were used in rate of rise catalyst comparisons and the
results are summarised in Table 1.43. The data suggests there is no significant change in
reactivity when comparing the control formulation (XXXV) to the experimental
formulations.
Table 1.44 lists the physical property data for the flexible slabstock formulations. The
experimental non-fugitive catalyst Dabco NE500 can be used with a slight increase in
Dabco T-9 catalyst level (shown in Table 1.42, formulation XXXVII). Formulation
XXXVII compares very well to the control (XXXV). Formulation XXXVI illustrates the
use of Dabco NE500 and Dabco NE600 balanced non-fugitive catalyst combination, to
match a Dabco BLV control. In this case, all physical properties were also matched, with
the exception of slightly decreased airflow and improved ILD properties. Overall, either
the non-fugitive catalyst combination Dabco NE500/Dabco NE600 or the use of Dabco
NE500 with increased Dabco T-9 can replace Dabco BLV in all water blown flexible
slabstock formulation.
Dabco NE600 in combination with Dabco NE500 or Dabco NE600 as a sole amine
catalyst used with increased levels of T-9 provides for equal or improved performance
compared to Dabco BLV in all water-blown flexible slabstock foam. In all cases, these
catalyst combinations can completely replace Dabco BLV and eliminate amine emissions
from the flexible slabstock foam.
74
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Details of the TDI and MDI foam model systems have been previously published [2].
The models require the use of mono-functional reactants that are quantitatively analysed
to correlate structure-activity relationships for various classes of catalysts. A realistic
thermal profile is produced through the imposition of an external exotherm. Urethane,
urea, allophanate and biuret reaction products are quantified by liquid chromatographic
analysis of quenched reaction samples. The models effectively account for such non-
ideal conditions as reactant depletion at variable rates, temperature and concentration-
dependent catalyst activity, and catalyst selectivity as a function of isocyanate distribution.
The information below highlights the features characteristic of the TDI and the MDI
flexible moulded models. Table 1.45 illustrates the conversion of the formulations from
Tables 1.12-1.15 into a TDI model system. Table 1.46 illustrates the conversion of the
formulations from Table 1.35 into an MDI model system.
Since the TDI and MDI automotive polyols are highly ethylene oxide (EO) tipped the
reactive hydroxyl group can be represented by a primary alcohol. The electronic effect of
75
Advances in Urethane Science and Technology
76
Dimensional Stabilising Additives for Flexible Polyurethane Foams
the ether two carbons away from the alcohol is accounted for by using an ether alcohol.
Thus both stearic and electronic effects on the reaction kinetics can be accurately reflected
in the model compound. The unreactive polyether backbone of the polyol can be
represented by the dimethyl ether of diethylene glycol. Dipropylene glycol dimethyl ether
would be more representative of the polyether backbone of a typical EO-tipped propylene
oxide (PO) polyol, but the rate and selectivity measurements are minimally affected by
the choice of model polyol backbone. Note the sum of the masses of the model alcohol
and the diglyme slightly exceeds 100 grams, the total polyol mass, since the hydroxyls
from the alkanolamine crosslinkers are included in the model alcohol mass. Nonetheless,
the concentration of reactive groups is comparable to that in the actual foam. Water and
catalysts are the same for both formulation and model. The model system does not
require a surfactant however, since a foam is not actually produced. TDI 80 and Mondur
MR are represented by 40:60 mole% and 80:20 mole% mixtures of phenyl and ortho-
tolyl isocyanate, respectively. Phenyl isocyanate represents the 4-position in 2,4-TDI or
4,4´-MDI, and tolyl isocyanate represents the more sterically hindered 2- and 6-positions
in the 2,4- and 2,6-TDI, or the internal rings of MDI oligomers. It is important to include
both isocyanate types because using only the more reactive phenyl isocyanate significantly
overestimates the reactivity of the system.
Detailed procedures for an individual model run can be found in the literature [14]. For
these runs a masterbatch mixture containing the alcohol, ether, water, catalyst and the
3,3´-dimethylbiphenyl internal standard was prepared in advance. The masterbatch
mixture was charged to a 50 cm3 roundbottom flask equipped with a glass thermocouple
well, septum, and jacketed mechanical stirrer with gas inlet and a septum. For the TDI
model the flask was placed under a slow argon purge in a sand bath capable of raising
the internal temperature from 60 °C to 120 °C in 4 minutes in the absence of any reactions.
The isocyanate was added via a syringe when the internal temperature reached 60 °C.
For the MDI runs the flask was wrapped with insulation but not otherwise heated
externally, and the isocyanate was added at 25 °C. Samples were then withdrawn via a
syringe at 30 second intervals for eight minutes. The samples were quenched with
dibutylamine and analysed by liquid chromatography to determine the yields of the urea
and urethane reaction products as well as the unreacted isocyanates, analysed as the
dibutyl ureas.
The flexible moulded chemical foam model provides a detailed look at the performance
of the industry standard catalyst package Dabco 33LV/Dabco BL-11 compared to the
non-fugitive catalyst packages Dabco NE1060/Dabco NE200 and Dabco NE1060/XF-
N1085. The foam results presented previously can be explained in terms of the overall
catalyst performance. The catalysts were characterised with a model system because
urea, urethane and isocyanate can be quantified as a function of time, so this approach
provides the highest level of detail. However, a model system is only relevant to the
77
Advances in Urethane Science and Technology
extent that it accurately reproduces the environment in an actual foam. The mixing in
the model systems is less energetic, so this may influence the relative rates of catalyst
hydrolysis versus reaction with isocyanate. However, if catalysts are compared under a
consistent set of conditions, the relative performance differences should be meaningful.
The most convenient way to compare catalyst performance is to use blow to gel selectivity.
Selectivity as a function of time (t), is defined next.
Catalyst selectivity is defined as the ratio of the normalised amount of blowing (urea
formation) to the normalised amount of gelling (urethane formation). A selectivity of
1.0 means that the normalised amounts of blowing and gelling are equal at that point in
the reaction. A selectivity substantially below 1.0, for example about 0.3, is indicative of
a strong gelling catalyst. A selectivity greater than 1.0 is indicative of a blowing catalyst.
However, it has been shown that a catalyst with a blow-to-gel selectivity greater than 0.8
can still serve to balance a strong gelling catalyst such as TEDA, and a good quality foam
can be produced [14].
Note the limiting urea and urethane yields are simply the molar equivalents of water and
alcohol, respectively, from Tables 1.45 and 1.46. Selectivities are plotted as a function of
isocyanate conversion in Figures 1.45 and 1.46.
Details on the development and use of foam model systems for flexible moulded foam
have been published [14]. In this section the features characteristic of slabstock foams
will be highlighted. Table 1.47 illustrates the conversion of the formulations from Table
1.42 into a model system.
Since the Voranol 3010 polyol is fully PO tipped, the reactive hydroxyl group can be
represented by a secondary alcohol. The unreactive polyether backbone of the polyol
can be represented by the dimethyl ethers of dipropylene glycol and diethylene glycol.
The diglyme represents the wt% EO incorporation into the polyol. Note the sum of the
masses of the model alcohol and the two dimethyl ethers equals 100 grams, the total
polyol mass. This ensures that the concentration of reactive groups is comparable to that
in the actual foam. Water and catalysts are the same for both formulation and model.
78
Dimensional Stabilising Additives for Flexible Polyurethane Foams
Figure 1.45 TDI Flexible Foam Model Selectivity versus NCO Conversion
Figure 1.46 MDI Flexible Foam Model Selectivity versus NCO Conversion
79
Advances in Urethane Science and Technology
Once again the model system does not require a surfactant, since a foam is not actually
produced. As noted previously, TDI 80 is represented by a 40:60 mole% mixture of
phenyl and ortho-tolyl isocyanate.
Detailed procedures for an individual model run can be found in the literature [14]. For
these runs a masterbatch mixture containing the alcohol, ethers, water, Dabco BLV catalyst
and the 3,3´-dimethylbiphenyl internal standard was prepared in advance. The masterbatch
mixture was charged to a 50 cm3 roundbottom flask equipped with a glass thermocouple
well, septum, and a jacketed mechanical stirrer with gas inlet. The flask was placed under
a slow argon purge in a sand bath capable of raising the internal temperature from 50 °C
to 120 °C in 4 minutes in the absence of any reactions. Isocyanate was added via syringe
when the internal temperature reached 50 °C, and the T-9 was injected from a microliter
syringe immediately afterwards. Samples were then withdrawn via the syringe at 30 second
intervals for eight minutes. The samples were quenched with dibutylamine and analysed
by liquid chromatography to give yields of the urea and urethane reaction products as well
as the unreacted isocyanates, analysed as the dibutyl ureas.
The chemical flexible slabstock foam model provides a detailed look at the performance
of the industry standard catalyst Dabco BLV compared to the non-fugitive catalysts Dabco
NE500 and the Dabco NE500/Dabco NE600 package, keeping the Dabco T-9 level
constant. Selectivity and conversion can be calculated as described previously. Note that
80
Dimensional Stabilising Additives for Flexible Polyurethane Foams
the limiting urea and urethane yields are simply the molar equivalents of water and
alcohol, respectively, from Table 1.47.
Dabco 33LV is considered to be a strong gelling catalyst, but under slabstock conditions
it is actually more of a blowing catalyst, with an initial selectivity near 2. This reflects the
difficulty of catalysing the secondary alcohol-isocyanate reaction in competition with
the water-isocyanate reaction. Tertiary amine catalysts are highly sensitive to stearic
hindrance near the reaction site. Secondary alcohols actually become less reactive than
water, even in the presence of a strong gelling catalyst. This is why tin catalysts are
critical in flexible slabstock formulations because they are less sensitive to stearically
hindered reactions. Dabco BLV contains some bisdimethylaminoethyl ether (Dabco BL-
11), which is a very strong blowing catalyst. Thus the initial selectivity is even higher,
near 4. The non-fugitive catalysts give selectivities comparable to that of Dabco BLV.
Note that the non-fugitive catalyst packages are more blowing selective than both the
TDI and MDI controls. Modern polyol technology tolerates and in some cases even
benefits from higher initial blowing selectivity because of the reduced levels of monol
and diol relative to triol. Thus network formation is more efficient, and correspondingly
less gelling is required to produce a superior urethane network. Higher blowing selectivity
can promote improved cell opening, which is now more critical in the face of improved
network formation. In a sense the control formulations are over gelled, but older
formulations could tolerate high gelling due to weaker urethane networks and the frequent
use of Dabco BL-11, which is a potent blowing catalyst.
Although the selectivity curves are different, the catalyst packages are still rate matched.
Rate matching in foam is usually accomplished by comparing rise profiles, which are
largely determined by the blowing rate. Thus the non-fugitive catalyst packages provide
blowing comparable to the controls, but provide a lower extent of urethane formation
early in the foaming process.
1.7 Conclusions
81
Advances in Urethane Science and Technology
Acknowledgments
The authors would like to thank and acknowledge the efforts and support of the following
people at Air Products and Chemicals, Inc.: Mark A Eckert, Steven E Robbins for physical
property testing of foams generated in this study, and Ilean S Ruhe for preparation and
editing of this work. We would also like to thank Air Products and Chemicals, Inc., for
support of this work and permission to publish it.
References
6. ASTM D3574-95
Standard Test Methods for Flexible Cellular Materials - Slab, Bonded, and
Molded Urethanes Foams.
9. SAE J1756
Test Procedure to Determine the Fogging Characteristics of Interior Automotive
Materials, 1994.
82
Dimensional Stabilising Additives for Flexible Polyurethane Foams
13. D. G Battice and W. I. Lopes, Presented at the SPI Polyurethanes: Exploring New
Horizons Conference, Toronto, Canada, 1986, p.145.
83
Advances in Urethane Science and Technology
84
2
Demands on Surfactants in Polyurethane Foam
Production with Liquid Carbon Dioxide Blowing
Andreas Weier and Georg Burkhart
The history of polyurethane foams and of polyurethanes in general is not that long for a
plastic material. It has been characterised by one special feature: the consistent changes
of the industry.
When Otto Bayer started his research on this new class of plastic material built from the
addition reaction of isocyanates and hydroxyl containing materials in the years 1936/1937
he envisioned a plastic material which could be tailored to many different applications by
the broad variety of acidic hydrogen containing compounds already available.
But there were also setbacks to this concept. The first major drawback was the fact that
the necessary diisocyanates were not available on an industrial scale and large scale
synthesis of these compounds seemed to be a rather high hurdle.
A second major point was an obstacle that Bayer and his group faced again and again.
When they tried to make the polyurethanes they envisioned in solid form, like films, etc.,
they always ended up with gas bubbles in their plastics. That was due to the inefficiency
of the methods available to synthesise pure water free material. After encountering this
effect for nearly four years Bayer made a bold move in line with the concept ‘use what
you cannot prevent’ – in this case the formation of gas bubbles [1, 2]. Thus Bayer’s group
started to make polyurethane foams.
Although these new lightweight porous materials were envisioned as supportive materials
as well as insulative materials in the base patents [3, 4] their market development remained
slow. This can be seen by the fact that even in 1952 polyisocyanates, mainly toluene
diisocyanate (TDI), were available worldwide in quantities of less than 100 tonnes. After
this rather hesitant start of polyurethane history and the first major switch from solid
materials to porous foamed plastics, the industry has been characterised by significant
changes in concept and the resulting industrial application of these switches.
The first major switch resulted from the basic research enabling the technical production
of soft polyurethane foams in the early 1950s. This technology was mostly focused on
polyester polyols as raw materials. About five years later polyether polyols entered the
85
Advances in Urethane Science and Technology
market as yet another major change within the polyurethane foam industry. Since the
early 1950s the foaming machines for the industrial production of polyurethane foams
had been discontinuous machines. With the use of polyalkyleneglycols as polyols, the
importance of continuous foam production grew significantly, for both performance
and commercial reasons. In the early 1950s, the necessary stabilisers for the production
of polyurethane foams had been silicone oils, with the major suppliers being Bayer, Dow
Corning, General Electric and ICI. In the late 1950s the picture changed with the broad
application of the continuous process, first with prepolymers, then the one-shot-process.
For this type of polyurethane foam production a different class of stabilisers was needed:
the new silicone-polyether copolymers.
The larger variability of the available polyetherpolyols laid the foundation for a significant
broadening of the performance range of polyurethanes – and all this with lower costs
than before. The growth of polyurethanes after this chemical change was fast [5]. Even
in 1960 more than 45,000 tons of flexible polyurethane foam were produced. Besides
its major application as insulation materials and comfort products in furniture
applications, new usage areas have been continuously identified and tailor-made
polyurethane materials have been supplied. Nowadays the range covers very high density
contour parts as well as very low density packaging foams, flexible foams as well as
semi-rigid and rigid foams, thermoplastic polyurethanes as well as integral skin foams or
reaction injection moulded (RIM) materials [6]. In the beginning of the 1990s an estimated
500,000 people had jobs in the polyurethane industry, either within production of raw
materials or in the conversion to final products [7].
By definition the production of a foam depends on the formation and stabilisation of gas
bubbles in a liquid. That is true even for polyurethane foam when the additional curing
of the liquid results in an elastic or rigid solid material. One of the central questions is
how the gas bubbles are generated and how long this generation takes.
The basic chemical gas forming process in polyurethane foaming is the very exothermic
reaction of water and isocyanate, resulting in the generation of carbon dioxide, urea and
heat. This method as the sole method to generate the gas bubbles of a polyurethane foam
is limited to the production of a rather small range of density/hardness combinations.
The achievable minimum density is limited by the tolerable heat in a foam bun before
scorching or even self-ignition occurs. Increasing water levels lead to the generation of
increasing amounts of carbon dioxide and therefore to lower densities. Increasing water
levels will also result in higher hardness as more urea is generated. During reaction the
urea appears in a variety of associated forms before it will ultimately separate as a solid
in flexible slabstock polyether foams [8, 9, 10].
86
Demands on Surfactants in Polyurethane Foam Production…
During the history of polyurethane flexible slabstock foam production the accessible density/
hardness range was continuously broadened with a wide variety of techniques. The most
obvious one was the use of a physical blowing agent to generate the gas volume necessary to
expand or generate gas bubbles without generating additional heat during the process. Due
to its inert character and its low biological toxicity, CFC-11 (Freon) was the liquid blowing
agent of choice for many years in Europe, North America and Asia. That changed after
recognition of the fact that CFC-11 had both a high global warming potential (GWP) and a
high ozone depletion potential (ODP). The result was the inception of the Montreal Protocol
and subsequently the exploration of a broad variety of alternative blowing agents (ABA),
including hydrochlorofluorocarbons (HCFC), acetone, cyclopentane and methylene chloride.
With the use of ABA, both the available densities and the available foam hardnesses
were lowered in comparison to the purely water blown foams, as in these cases gas
volume was produced without generation of urea. Softening of foam was also achieved
chemically, either with the use of an additive to change the urea morphology [13] or with
the use of crosslinkers in combination with a low isocyanate index [14].
A few years later machinery modifications enabled foamers to produce low density foams
by the use of increased water levels while reducing traditional blowing agent levels. This
reaction preferentially utilises the reaction generating carbon dioxide as a blowing agent.
When the water in the formulation is increased, the foam bun temperature can become
too high for safe processing. This problem has been tackled by forced cooling equipment
generating an air exchange in the foam bun directly after production and thereby cooling
the bun to prevent scorch or fire problems. There is a variety of technologies for the
forced cooling of foams, including Envirocure [13, 14], Rapid Cure [15], Reeves Brothers
[16] and lateral cooling [17]. For the production of very soft foam grades some physical
blowing agent is still needed. Also there are always some safety concerns over possible
equipment malfunctions when making these hot foam grades.
A safer way to control flexible foam density and hardness is the variation of production
pressure. Making high as well as low pressures available in the foam production
environment can give access to higher density hard grades as well as to low density
foams without the use of physical blowing agents. The two commercial processes
addressing this approach are Variable Pressure Foaming (VPF) and Foam One. As foam
production is done in a closed chamber, volatile materials can be trapped rather efficiently.
This creates the possibility of a clean and safe, as well as ABA free foam production. A
setback is the high capital investment cost, especially for a VPF unit. Therefore this
equipment only becomes cost effective in very high volume production plants.
Review of the many production techniques reveals that the production of a broad range
of flexible foam qualities with reduction of environmental hazards (ODP as well as GWP)
is possible by numerous options. All of these current technologies offer benefits, but
provide no clear economical solution for total ABA elimination.
87
Advances in Urethane Science and Technology
Within all the blowing agents available today, carbon dioxide probably offers the best
combination of zero ODP, low GWP and low price. As it is abundant, it doesn’t have to
be generated. For capturing and transporting only a minor amount of energy is needed.
Merely the point of emission is changed. On top of all that it is also non-flammable,
which minimises safety concerns in the foam plant.
Therefore many attempts have been made to produce polyurethane foams with carbon
dioxide that is not generated by a chemical reaction during the foaming process. The
difficulty is that carbon dioxide is a gas at room temperature under atmospheric pressure.
The only way to utilise carbon dioxide not generated chemically is to mix the carbon
dioxide under pressure or in a liquid state with the chemical components of the foam
formulation. Then the pressure is subsequently released causing the frothing or pre-
expansion. Carbon dioxide as a non-reactive blowing agent in the frothing technique
has been suggested for quite some time already [18, 19]. The pressurised reaction mixture
is ejected at atmospheric pressure causing a turbulent vaporisation of the blowing
agent. This allows the manufacture of a foam with reduced density, but the cell structure
is of very inconsistent quality due to irregular shaped and oversized cells or bubbles
being present.
2.2.1 Machinery
The situation changed when Cannon introduced liquid carbon dioxide foaming technology
for the industrial production of polyurethane flexible foams with its CarDio process. This
process was introduced to the polyurethane foam production industry in 1994. Today foam
manufacturers worldwide consider liquid carbon dioxide as a blowing agent and make this
process available for the industry. The major point of this technology is the control of the
frothing that occurs as soon as the pressurised liquid carbon dioxide as part of the foam
formulation leaves the pressurised equipment and the liquid foam mixture is released into
atmospheric pressure. The controlled release of the mixture into the comparatively low
atmospheric pressure conditions is essential to all the liquid carbon dioxide foaming techniques.
In the CarDio process this is accomplished with the specifically designed lay down device
[20, 21] which is called a ‘Gate Bar’ (see Figure 2.1). The gate bar essentially consists of
a metal bar with a pressure drop slot connecting the outside with a feeding tube inside
the bar. Through a liquid mix injection point formulation components with the pressurised
liquid carbon dioxide enter the gate bar.
88
Demands on Surfactants in Polyurethane Foam Production…
The frothing mixture flows along the frothing cavity and through an outlet aperture
acting as a pressure drop zone onto a substrate, e.g., the moving belt of the foam machine.
A variation of the original gate bar design was introduced with the CarDio 2000 featuring
a pressure adjustable lay down device [22]. It allows adjustment of the slot width in
response to the output and the liquid carbon dioxide percentage (see Figure 2.2).
89
Advances in Urethane Science and Technology
Another patented technology for liquid carbon dioxide foaming is the NovaFlex process
developed by Hennecke and Bayer [23]. The NovaFlex process uses a creamer dispensing
the froth on to the machine. It focuses on the same task of pressure reduction as the
CarDio gate bar, except dispensing the froth occurs at one single spot on the belt and is
not distributed over nearly the whole belt width. The same is true with the third version
of liquid carbon dioxide machinery, the Beamech CO-2 equipment. Obviously, both the
NovaFlex as well as the Beamech CO-2 process use pressure reduction devices different
from the gate bar of the CarDio process as they are said to show a different response to
the use of fillers or generally solid particles within the foam formulations. One modification
of the original NovaFlex process is the MultiStream configuration. It allows the use of
liquid carbon dioxide with a range of polyol types in all different ratios by utilising one
liquid carbon dioxide addition point in one high pressure polyol stream which is then
combined with additional polyols by pass streams [24, 25].
90
Demands on Surfactants in Polyurethane Foam Production…
In standard foam production, a mixture of liquid materials is dispensed and the gas bubbles
within the foam structure are only generated by a rather slow chemical reaction. In the case
of liquid carbon dioxide foaming the mixture is already dispensed in a two-phase process.
No matter which particular pressure drop system is utilised, the pressure drop is very
fast due to the large flow speeds in the system. The fast pressure drop between the mixing
head and the dispensing compresses this time span for bubble formation to fractions of
a second. As all the gas bubbles resulting in foam cells of the final foam have to be
generated in this short instant, clearly the nucleating powers of the formulation ingredients
are challenged severely. This is especially true when looking at the foam stabilisers used
as they are the main formulation ingredients determining the nucleation.
As early as 1969, Kanner and Decker [26] showed by photomicrography that self-nucleation
is essentially absent in a polyurethane foam system. Their results indicate that bubbles are
introduced by the process of mixing and that the presence of a silicone surfactant increases
the volume of air bubbles introduced during the mixing. Nucleation in this context means
the formation of gas bubble cores on which a bubble might grow above its critical bubble
radius. The critical bubble radius is the smallest bubble radius for which the addition of
more gas will result in a net decrease in free energy. A bubble which is smaller than the
critical radius will not spontaneously grow in size because the addition of more gas results
in a net increase in free energy. Obviously a decrease in the surface tension of a foaming
mixture by the silicone surfactant would decrease the energy needed to generate bubbles of
the necessary critical radius. Even under conditions of a supersaturation pressure of 2.03
MPa carbon dioxide in a liquid with a surface tension of 0.025 N/m and a temperature of
25 °C, the critical bubble radius is 2.4 x 10-5 cm. It has been proven that silicone surfactant
can aid nucleation by lowering surface tension [27].
However the surface tension lowering of a silicone surfactant might just be one effect
helping to induce new bubble formation by liquid carbon dioxide. The surfactant is
entirely soluble in the liquid and therefore cannot act as a heterogeneous nucleating
agent. Besides that other surfactants like fluorocarbons can also lower the surface tension
but still don’t act like a silicone surfactant does in polyurethane foam production. A
possible explanation of this difference might be related to the exceptional solubility of
oxygen and nitrogen in silicones as demonstrated by Arkles [28, 29]. It is well known in
the polyurethane industry that insufficient air loading of raw materials like the polyols
results in coarse cell structures even in the conventional foaming process. Obviously a
sufficient amount of nitrogen and/or oxygen is needed for good nucleation. As silicones
show a high permeability and solubility for these gases, it is reasonable to assume that
91
Advances in Urethane Science and Technology
the ease of formation of nuclei for gas bubbles is higher in silicone-like surroundings
provided along the siloxane backbone of a silicone surfactant.
An effect not accessible to Kanner and Decker, and therefore not a subject of their
experimental work, was the high nucleation demand found in the new liquid carbon
dioxide foaming technologies. Their finding that bubbles are introduced during the process
of mixing by the shear energy of the mixer might reflect the observation that the amount
of nuclei generated during mixing has always been sufficient for a standard polyurethane
foaming process. The rather large timescale between mixing and appearance of gas bubbles
large enough to be visually detectable obviously gave carbon dioxide molecules sufficient
time to diffuse to a nucleation site as they were only generated slowly within the foam
formulation. With the new liquid carbon dioxide foaming technologies, a significantly
higher number of nucleation sites is needed for the carbon dioxide to find a gas bubble
or a gas bubble core during the short time frame of the pressure decrease in the pressure
reduction area within the equipment used.
Whereas the former trend in some polyurethane (PU) markets has been to sacrifice more
and more of a surfactant’s nucleating power to gain higher and higher activity, liquid
carbon dioxide foaming has led to a new description of the needed performance profile.
This is recognising the fact that quite a number of the very high active surfactants used at
the time of the introduction of the liquid carbon dioxide processes resulted in unacceptable
foam structures. Irrespective of the liquid carbon dioxide system, the high activity
surfactants generated coarse foams. That is in contrast to the findings of Kanner and
Decker in standard foam systems. Obviously under these conditions the mixing energy
of the machine system has not been high enough to generate sufficient nuclei or gas
bubbles to guarantee a regular fine celled foam structure. Obviously under these conditions
the main quality issue of a surfactant is its nucleation power, and not the stabilisation
activity. Not surprisingly a well established medium active surfactant was used during
the development of the first commercial liquid carbon dioxide process, the CarDio process.
It provided a good balance between nucleation and processing as well as activity. Therefore
this type of surfactant was still the number one choice for processes even four years after
the introduction of commercial liquid carbon dioxide foaming. Still, due to the high
demands in nucleation, the broad variety of foam grades and the flame retardant (FR)
demands of some markets a desire for optimising the different surfactant performance
aspects could clearly be seen. It also provided a focus of interest for the flexible slab
polyurethane community for quite a few years.
92
Demands on Surfactants in Polyurethane Foam Production…
formulation. An obvious example for this performance aspect is the chemical reaction of
water and isocyanate, i.e., the formation of carbon dioxide and urea. The different polarities
of isocyanate and water should normally prevent their reaction with each other in the
short time frame of a PU foam reaction and even the use of polyols as the major formulation
component will not always help to overcome the tendency for phase separation.
Even the polyols used in some special foam types can give rise to a number of problems.
One of the challenges for the emulsification power of silicone surfactants in flexible
slabstock foam production for example is the combination of a standard polyol with a
hypersoft polyol. The consistency of the foaming results and the cell structure obtained
can be severely influenced by the choice of silicone surfactant used. It can be easily
rationalised that the rise profile and physical foam properties like tensile strength and
elongation will be influenced by tendencies for phase separation after the mixing and
dispensing of the formulation ingredients.
Probably the most obvious task of any silicone surfactant in flexible polyurethane foam
production is to stabilise the foam during the foam rise. If the surfactant used is
contaminated or is in too low a concentration, settling of the foam or even foam collapse
will occur. Two possible reasons for the foam instability during the rise time are discussed
in the literature, and lead to different theories on the roles of silicone surfactants used.
The first explanation of the role of silicone surfactants in flexible polyurethane foam is
based on the fact that as soon as the volume fraction of the gas bubbles exceeds 74%, the
spherical bubbles will distort into multi-sided polyhedrals. This means that cell windows
with Gibb’s plateau borders are formed (a polyhedral foam consists of cell windows and
struts. Another term for struts is Gibb’s plateau borders). A pressure difference due to
capillary pressure will cause liquid in the cell window to drain into the struts since the
pressure inside the plateau borders is lower than that in the cell windows. Without adding
silicone surfactants this drainage rate will be very fast, so that film rupture and bubble
coalescence occur rapidly. It has been shown that due to the surface tension gradient
generated by a silicone surfactant the cell window drainage rate is lower [30]. Therefore
different surface structures do have different effects on cell window drainage. That has
been said to result in a distribution of different cell window thicknesses at the time of cell
opening [31]. The critical factor in this picture is the surface elasticity. It is known to be
caused by a surface tension gradient along the cell window upon expansion and it will
retard the cell window drainage rate. This effect is called Gibbs-Marangoni effect and is
reviewed by Scriven and Sternling [32]. Owen and co-workers also studied the dynamic
surface tension of a series of surfactant solutions and stressed the importance of surface
elasticity [33]. Similar work has been performed by Zhang and co-workers [34].
93
Advances in Urethane Science and Technology
A different view point of the mechanism of flexible polyether polyurethane foam stabilisation
by silicone-based surfactants has been published by Rossmy and co-workers [8, 9, 10]. The
authors investigated the urea precipitation phenomenon in polyurethane foams in greater
detail. By adding a defoaming agent to a foam formulation they showed that urea precipitation
turned a clear foam mix opaque. They also noted that the cell rupture always occurred just
after urea precipitation. In light of this observation the hypothesis is that precipitation of
polyurea destabilises the foam mix and leads to cell opening. The surfactant then aids in
stabilising the foam by incorporating the urea precipitate in the foam matrix, adding integrity
to the foam. Today it is not doubted any more that there really are urea domain structures in
polyurethane foams and they have been subject to intensive research [35, 36].
Figure 2.4 shows the remaining fragments of a cell window after cell opening as found in
the fully cured foam. It is very obvious that the opening of a flexible PU foam is not a
94
Demands on Surfactants in Polyurethane Foam Production…
As soon as the foam is produced and cured the silicone surfactants used are not needed
or desired any more. One exception might be the production of hydrophilic or hydrophobic
foams where a surface active material like a silicone will in some cases have a recognisable
influence. But with standard foams as well as with liquid carbon dioxide blown foams,
additional performance aspects besides the stabilisation during the foam rise might be
important for surfactant selection. These criteria therefore have to be addressed as well
and deserve some consideration as they might also affect the number of key additives a
foam operation has to handle. An effect clearly seen in the final foam is the undesirable
effect the silicone surfactant has on the burn performance of the resulting foam. Especially
in liquid carbon dioxide foaming, suitable conventional (or non-FR) surfactants have
been available right from the beginning. But the development of optimised universal or
non-FR-surfactants for liquid carbon dioxide foaming took considerable effort and time.
To understand the effect of the silicone surfactant on burning it is interesting to have a
look at a theory for flame spread development.
The simplified illustration in Figure 2.5 shows that the heat decomposes the organic
material at the surface during an endothermic procedure. Pyrolysis products are created
and because of an exothermic reaction with oxygen at the boundary layer of the flame,
the formation of even higher reactive decomposition products takes place. These aggressive
radicals are responsible for an accelerated degradation of the polymeric surface. As long
as the result of the energy is positive, there is thermal feedback to the endothermic process
95
Advances in Urethane Science and Technology
at the surface and the combustion keeps going. It is believed that the formation of hydrogen
radicals and hydroxyl radicals from the organic material are the key factors initiating
and supporting the combustion phenomena. The working mechanisms of FR address
these phenomena.
The effect of a silicone surfactant on the burning behaviour of the polyurethane foam
can be quite significant even though they normally represent less than 1% of the plastic
material. It is due to the fact that in any case the decomposition of a polyurethane foam
starts at the surface. Because of the surface activity of the foam stabilisers it is easy to
rationalise their enrichment on the surface and it is the surface that is the most influential
part of the polymer regarding flame spread development.
This point can be further highlighted if we look at the influence of surfactants on the
FR-performance of a polyurethane foam with 1 weight percent of surfactant externally
applied (Figure 2.6). To characterise the burn performance in this case, a horizontal
burn test was used and the burn length was measured. Surfactants A, B, C and D are
rigid foam surfactants with different structural parameters in the silicone as well as in
the polyether chains. It can be seen in Figure 2.6 that all the different types of silicone
surfactants used do have a negative influence on the burn performance of the foam.
The reason for this behaviour can be mainly attributed to the structure of the siloxane
backbone within the siliconepolyethercopolymers [37]. To characterise the effect of
the polydimethylsiloxane (PDMS) chain used as the backbone of such a
siliconepolyethercopolymer, interesting tests have been made by external application
of pure PDMS in different weights as well as the use of molecules of different average
chain length (Figure 2.7).
96
Demands on Surfactants in Polyurethane Foam Production…
Not only do increasing amounts of externally applied PDMS lead to increasing burn
length in, but more astonishingly, this effect is obviously correlated to the average
chain length of the applied siloxane. It can be seen that with longer chain length of the
PDMS, the detrimental effect on the FR-performance of the foam increases. This effect
far out balances the pure weight effect that would explain why a higher amount of
PDMS leads to higher burn length value due to the larger amount of materials supporting
the decomposition.
To address the question of whether the different volatility of the different PDMS
chains is the main reason for this effect, an additional test with silica was carried
97
Advances in Urethane Science and Technology
out. It is easy to expect that solid materials, especially if inorganic, will have a positive
effect on the FR-performance of the foam and that they will lead to a decreasing
burn length in a burn test. Depending on the chemical nature they could be either
expected to be a heat sink, a material to catch free radicals, to char the burn front of
the foam or to liberate non-gaseous products to dilute the oxygen at the burn front.
98
Demands on Surfactants in Polyurethane Foam Production…
As outlined previously the main reason for the interest in optimised surfactants for the
liquid carbon dioxide processes lies in their role as nucleation promoters.
99
Advances in Urethane Science and Technology
slabstock surfactants generally have more than one type of polyether attached, this gives
rise to virtually millions of different combinations for surfactant molecules. Also, in most
foaming applications the pendant polyethers have to be unreactive to isocyanates so that
they do not act like crosslinkers which would lead to tight or shrinking foam.
In addition there are two general ways to link the polyethers to the siloxane chains (see
Figure 2.9). This opens the basis for two separate product lines, each having its special
advantages over the other.
The Si-O-C products resulting from the reaction of chlorosiloxanes and hydroxyl
groups of polyethers provide an extremely good processing and superior consistency
combined with being hydrolytically stable under the water-amine conditions found
in polyurethane foaming. The Si-C products offer a more beneficial access to high
activity and lead to an easier production of flame retardant foams. These products
are derived from the addition of an Si-H functionality at the siloxanes to a double
bond, a process called hydrosilylation.
Examples of how this variety of structural parameters affects the development of silicone
surfactants especially for liquid carbon dioxide blown foams, are many [39, 40, 41, 42]
and this latest drive in surfactant development provides good examples of the important
performance issues and how they can be addressed.
100
Demands on Surfactants in Polyurethane Foam Production…
One of the first published attempts to get an indication of the nucleation performance of
a silicone surfactant was the froth test [39, 40, 41]. In this test polyol and surfactant are
stirred in a reproducible fashion and the resulting froth density as well as the time needed
for the foam to collapse is measured. After some experience with industrial surfactants
used in the market these products with known performance in industrial liquid carbon
dioxide foaming were subjected to the froth test. No direct correlation could be found
between the obtained data and the performance of the products on industrial machines
(Figure 2.11).
Figure 2.11 Performance of surfactants (Tegostab B 8228 and B 8220) in the froth test
Comp.: competitive surfactant; EP-H-18: experimental product
101
Advances in Urethane Science and Technology
As can be seen from Figure 2.11, both products Tegostab B 8228 and Tegostab B 8220
show nearly identical performance in this mixing test although industrial experience has
shown that Tegostab B 8220 results in significantly finer cell structure than Tegostab B
8228, even if both of them yield a very regular cell structure.
Even more confusing is the picture resulting from the comparison of a widely used
surfactant and an experimental product called EP-H-18 by Burkhart and co-authors
[42]. Both of the products again resulted in identical values in the froth stability test,
but differed significantly on an industrial production machine. Whereas the competitive
material resulted in a coarse cell structure, the experimental product yielded a fine and
regular foam.
102
Demands on Surfactants in Polyurethane Foam Production…
Because of these results a video microscopy check on the homogeneity of the gas bubbles
in the liquid mixture was performed. For this test the ingredients of a standard foam
formulation were mixed and immediately poured onto a polished metal plate. The
creaming of the foam was then filmed by a video camera with a 10x magnification.
Thirty and 150 seconds after the stirring commenced, a video printout was made and the
number of formed gas bubbles as well as their size homogeneity was characterised (see
Figure 2.12). This procedure is referred to as the video test.
Not taking cyclic molecules into account, the general structures of industrial silicone
surfactants for flexible slabstock foam production can be seen in Figure 2.13. The main
building blocks of these materials are a PDMS backbone and attached polyethers based on
ethylene oxide and propylene oxide addition products. The siloxane backbones can either
be linear or branched and can have their polyether substituents attached in an either pendant
or terminal location. These four general structures are outlined in Figure 2.13).
103
Advances in Urethane Science and Technology
The simplest type of the possible structures is a linear A-B-A-copolymer with a straight
siloxane backbone and polyether chains at the termini. This type of structure is described
as ST (for straight, terminal). With a straight siloxane backbone there is also the possibility
of attaching polyether groups as pendant chains. This combination is denoted SP (straight,
pendant). If the siloxane is branched as well it could have polyether groups attached as
terminal or as pendant groups (BT and BP). These four types of structures gave different
results in the video imaging test (see Figure 2.14).
The ST type of structure resulted in only a small number of gas bubbles being formed in the
early stages of the reaction. Somewhat better was the combination of branched silicone with
pendant polyether groups. The best video imaging results were reported with the branched
siloxane, combined with terminal polyether groups followed closely by the SP type of molecule.
ST SP
– +
BT BP
+(+) o
104
Demands on Surfactants in Polyurethane Foam Production…
These results matched, more or less, with industrial experience. Whereas the ABA type
of copolymer (ST-structure) is known to be a rather poor performing surfactant for
polyurethane production, the SP and the BT type of structures are commonly used. The
structure that performed best in the video imaging test (the BT type structure) is the
standard structural type that can be found in commercial SiOC-PDMS copolymers
obtained by the reaction of a chlorosubstituted PDMS with the hydroxyl group of
polyether (see previously). Although these SiOC products (named after the silicone-oxygen-
carbon bond linking the PDMS and the polyether side chain) provide a very good
combination of broad processing, good nucleation and consistent quality, their main
disadvantage is their very recognisable negative effect on the burning behaviour of the
produced polyurethane foam.
In contrast, the second best type of structure seemed to be the SP type. These structures
are the fundamental means to build SiC-products which are used as universal
surfactants, for example in North America. Although SP structures seemed to be the
obvious choice to build FR or universal type of surfactants (having a less negative
effect on FR-performance) they did not seem to show as good a nucleation as the BT
type of molecules.
Besides the type of branching in the siloxane backbone and the attachment points of
the polyether side groups, another factor characterising the siloxanes is the number
of unmodified PDMS groups for each of the modified methylsiloxane groups in the
molecule. This ratio is often denoted as the P-value and is especially important in
regards to the burn performance of the resulting foams (see Figure 2.10 and discussion
above). As that is another variable that has to be taken into account when
characterising a silicone surfactant for polyurethane foam production, this factor
was screened for as well.
As can be seen from the video test results, there seems to be a slight improvement in the
nucleation efficiency with increasing P-value. However the drawback is that the higher
P-values result in surfactants which will lead to a considerably negative effect on the
burn performance of the resulting foam [43].
The summary of the data at this point seemed to indicate that the combination of good
nucleation efficiency with the BT-structures and the advantage of high P-values seem to
105
Advances in Urethane Science and Technology
P-value
Figure 2.15 Video imaging test; nucleation efficiency with increasing P-value
correlate with the fact that, especially based on SiOC type of chemistry, it is easier to
build a non FR-surfactant structure than to build a universal or true FR-silicone surfactant
for liquid carbon dioxide foaming.
Surprisingly enough in a real foam test these molecules performed rather badly, resulting
in a considerably coarser foam than the reference (see Figure 2.17).
106
Demands on Surfactants in Polyurethane Foam Production…
Surfactant
without aromatic groups with aromatic groups
That observation was rationalised with the following assumption. If indeed aromatic
rings would increase the solubility of carbon dioxide then these types of structures would
be more soluble in a liquid carbon dioxide containing environment. This would result in
a decreased tendency to move to the surface of a such a liquid phase. The obvious result
would be a decrease of surface activity, so the net result could very well be a higher
nucleation efficiency combined with a lower activity during the rise time of the foam,
leading to gas bubble coalescence.
107
Advances in Urethane Science and Technology
If that is true, the polarity and therefore solubility of the molecule would have to be
adjusted to lower values. As explained previously the increase of the P-value was not an
option for the generation of a surfactant with good FR-performance.
The utilisation of the outlined principles and the continuous work with the many possible
variables by all the additive suppliers again and again has brought new high performance
surfactants, supporting the continuously changing demands in the innovative polyurethane
industry and most probably will continue to do so. Therefore this industry has all the
fascination that comes from the interaction of chemistry and physics, industrial application
and theoretical interest, commercial importance and environmental awareness. So even
more than 50 years after its beginning it will remain an area of new challenges and
performance oriented efforts.
108
Demands on Surfactants in Polyurethane Foam Production…
References
11. G. Burkhart, H-H. Schlöns and V. Zellmer, inventors; Th. Goldschmidt AG,
assignee; US Patent 5 132 333, 1992.
109
Advances in Urethane Science and Technology
18. P. Merriman, inventor; Dunlop Rubber Co., Ltd., assignee; US Patent 3,184,419,
1965.
23. R. G. Eiben, Presented at Utech ’96, The Hague, The Netherlands, 1996, Paper
No.31.
28. B. Arkles, Silanes & Silicones catalogue, Petrach Systems, Bartram Road, Bristol,
PA, 19007, USA, 1987, p 87.
110
Demands on Surfactants in Polyurethane Foam Production…
40. S. B. McVey, B. L. Hilker and L. F. Lawler, Presented at Utech 96, The Hague,
Paper No.38.
43. A. Weier, G. Burkhart and V. Zellmer, Presented at the SPI, Polyurethanes ’94
Conference, Boston, MA, 1994, p.202.
45. No inventors; Th. Goldschmidt AG, assignee; German Patent 19, 726, 653, 1999.
111
Advances in Urethane Science and Technology
112
3
Polyurethane Processing: Recent Developments
Max Taverna
This chapter highlights three aspects of manufacturing technology, which have recently brought
significant benefits to the producers of automotive seats, reinforced parts and generic moulded
items. These different manufacturing areas have been grouped here under a single title in
order to cover a wider spectrum of reader’s interests and to illustrate the broad area of
improvement and innovation that can still be found in this steady-growing segment of industry.
Cannon has developed complete systems for the manufacture of moulded polyurethane
automotive seating elements made with varying combinations of several raw materials.
The resultant cushions – although produced in a random sequence on the same moulding
line - are characterised by mechanical properties tailored to the specific application of
each part [1].
• dedicated mixing heads capable of processing six components (all with high-pressure
recirculation), including low-viscosity toluene diisocyanate (TDI) and water-based additives
113
Advances in Urethane Science and Technology
First, the number of suppliers is getting smaller and smaller, due to the extensive merger
and acquisition campaign run in recent years by the two largest players (Lear Co., and
Johnson Controls). This concentration creates an opportunity for making strategic choices
regarding the ‘make or buy’ decision with respect to manufacturing solutions. These
large manufacturers have grown significantly thanks to their own successful development
of both chemical formulations and moulding plants. Providing them with new machinery
sometimes means accepting their request to manufacture a concept that was developed
by their own engineering department.
The second aspect concerns the growing demand for simpler, high-efficiency plants, with
a reduced number of operators and a high degree of operating flexibility. It must be
possible to produce several different parts on the same moulding line so that, in the case
of a sudden change in production plans, complete projects can be switched from one
production line to another within a very short time frame. This increased flexibility
requires careful design of the metering equipment, mix heads and mould carriers because
they must be able to perform very different tasks in sequence or by project. A specific
request involves the potential to process multi-component formulations, where a wide
range of foams can be produced on one machine.
The third aspect, potentially in conflict with the previous concept, involves the development
of families of formulations based on specific chemicals: all diphenylmethane diisocyanate
(MDI), MDI/TDI in various percentages, all TDI and special polyols. This means dealing
with very different demould times and moulding conditions, that render a generic ‘seat
plant’, that was still so useable just a few years ago, obsolete. In this case, specific packages
must be available, which conflicts with the concept of high flexibility expressed above.
The modular design of this unit allows for easy addition of further components to a
formulation. Storage, metering, temperature-time control modules and feed lines are
assembled in modules and can be added as needed. Computerised process control easily
integrates the new chemicals in the formulation.
114
Polyurethane Processing: Recent Developments
Two methods were used to enable fast, precise changing of the output and to ensure
closed-loop control of the machine, which is essential to guarantee continuous, accurate
output of low-viscosity chemicals:
The output control on the closed-loop machine operates continuously, comparing the set
output value of the inverters with the real output as measured by the volumetric flow
transducers. The system enables pouring only when the parameter is within the limits set
by the operator via the keyboard and when it is possible to change the pump speed in less
than half a second.
The unit also includes maintenance and alarm menus. Through the maintenance menu,
it is possible to set limitations on the following parameters:
• number of shots
• material consumption
• working hours
115
Advances in Urethane Science and Technology
Once these set values are reached, the corresponding warning message informs the operator
to perform the required maintenance operation.
The programmable stepping device is very precise - the pump can be set at 256 discrete
positions - and can be easily programmed through the machine programme logic controller
(PLC). Since the switching time between formulations can be as low as 0.6 of a second,
multiple ratios can be set within the same pour program to produce multi-hardness and
multi-density foams with repetitive results.
Cannon has designed a new mixing head, capable of mixing six components, each with
individual recirculation control. The new mixing head was specifically designed to meet the
needs of the automotive seat producers who wished to mould TDI-based flexible foams with
a maximum of flexibility in their formulation. Several pure ingredients are kept separate up
to the point of injection and it is possible to operate each stream on demand, provide high-
pressure recirculation and avoid the contamination of components (see Figure 3.2) [1, 2].
Figure 3.2 The new Cannon Ax Head, specifically designed for multi-component
polyurethane formulations. It can mix up to six chemicals all with a high-pressure
recirculation feature.
116
Polyurethane Processing: Recent Developments
Although this solution is currently available on the market, competitive mix heads are
limited to four streams, their dimension and weight require heavy-duty pour robots while
providing speed and pour pattern limitations during operation and they have complex
pressure set-up and regulation procedures.
The innovative aspect of this mixhead is that TDI - and eventually a TDI-compatible
sixth component - are fed axially into the mixing chamber through the small piston that
cleans the mixing chamber. The other four components, polyols, other catalysts, additives,
flame retardants, are fed radially into the mixing chamber. A seal on the small piston
provides a permanent separation between the polyol and TDI feeding areas so that any
crossover of the low-viscosity components is avoided.
The main advantage of the new mixing head is its reliability. It can operate for millions
of shots because of the remote position of the TDI feeding area. TDI is fed in at a pressure
of only 1 MPa, with perfect mixing efficiency (see Figures 3.3, 3.4 and 3.5).
Figure 3.3 Section of the mixing-chamber’s cleaning piston. Four chemicals are fed radially
in the mixing chamber, while two are fed through a hole drilled in the cleaning piston.
117
Advances in Urethane Science and Technology
Figure 3.4 During the high-pressure recycle phase all six components flow
in the grooves carved in the mixing chamber’s piston and return to their
storage tanks.
Figure 3.5 When the injection signal is received, the mixing chamber’s piston
retracts, the four components fed radially meet the two that are fed axially
through the mixing chamber’s piston. The mixed blend reaches the discharge duct,
which is positioned at an angle of 90° to the mixing chamber and leaves the head
through the pouring hole.
118
Polyurethane Processing: Recent Developments
Defining ‘Formulation #1’ as being composed of polyol #1 and isocyanate and ‘Formulation
#2’ as polyol #2 and isocyanate, an example of a working sequence is as follows:
1. Pour ‘Formulation #1’: polyol #1 and isocyanate recycle through the mixing chamber’s
piston grooves and are poured in the open mould
The maximum response time, which occurs between two consecutive shots with different
formulations, is 0.6 seconds.
The pressure control of the component injectors, which is very important for ensuring the
proper mixing of the different liquid streams, has been achieved via three different approaches:
• closed-loop control of the injection pressure via hydraulic servo valves and feed-
back control from the pressure gauges
The package supplied with the head to achieve pressure control (see Figure 3.6) includes:
• valves for control of the oil flow; each valve has three positions (off, injection pressure
#1 and injection pressure #2)
119
Advances in Urethane Science and Technology
Figure 3.6 Scheme of the circuit used to guarantee a closed-loop control of the pouring
pressure of each chemical.
• one set of high-pressure flexible hoses to connect the proportional valves to the
injector nozzles
• appropriate controls
120
Polyurethane Processing: Recent Developments
3.1.3.4 Advantages
• The pour pattern can be defined very precisely, according to the design of the mould
and the position of any inserts.
• The pouring operation can be executed in an open mould without fear of collision
with the upper mould half.
• The robot carrying the head can achieve high acceleration and speeds without
mechanically stressing the moving elements.
The large number of chemical components that can be handled simultaneously provides
a high degree of flexibility in formulations, allowing for optimum use of the moulding
line. Several different parts can be produced on the same moulding line and different
types of foam can easily be produced in random sequence without forcing the operator
to use pre-defined sequences of moulds.
The basic features of a typical Cannon double-arm pour robot, normally supplied to
carry two heads (see Figure 3.7), are:
• two-axes Cartesian robot with two arms moving independently of each other,
• arm movements (along the ‘X’ and ‘Y’ axes) achieved via a rack-and-pinion system
driven by electronic variable-speed motors,
• racks fitted with position encoders to detect the relevant ‘X’ and ‘Y’ coordinates.
121
Advances in Urethane Science and Technology
The concept of the dedicated flexible foam moulding line relies on a mould carrier having
a simple solid structure, with a minimum number of mechanical parts, which require
maintenance. It consists of two upper and two lower sections linked by means of a
hooking system. The presence of two separate lower platens allows the operator to
compensate for any difference in mould thickness without incurring any problem by
mounting two moulds on the same carrier (see Figure 3.8). Insertion, centring and fixation
of various moulds within the mould carrier are achieved by means of tensioning screws.
Each mould carrier is usually provided with two pneumatic cylinders for opening and
closing. Upon demand, the closing system can be actuated hydraulically or mechanically.
Mould carrier tilting, for optimum evacuation of air during the filling phase, is
accomplished with a cam system. Each mould carrier can have two tilting positions:
either horizontal or frontally inclined by 20 degrees (see Figure 3.9). It is possible to
programme different angles of opening and tilting to correspond with the type of mould
currently being used.
122
Polyurethane Processing: Recent Developments
Figure 3.8 Independently-moving lower platens allow for the use of different thickness
of moulds as a double mould carrier.
Figure 3.9 A different degree of tilting can be obtained for each mould carrier, to
optimise the evacuation of air from the moulds during the filling/
polymerisation phase.
123
Advances in Urethane Science and Technology
for the moulding line are interfaced to the metering system controls in order to co-
ordinate all the working processes.
• mould identification: the control system on the moulding line reads the code coming
from the field
• pour programme selection: inside the mixhead manipulator control unit, a schedule
- assigning a specific pour programme to the different moulds - has to be loaded
• pour programme execution: the manipulator control unit sends the input to the
machine control unit; when the mould under consideration requires multi-hardness
foaming, both mix heads are activated, while in the case of a single-hardness seat,
only one mixhead is activated
As stated previously, the lower half of the mould carrier is composed of two independent
platens. They are fitted with a guiding system to ensure maximum parallelism between
the platens.
In order to ensure equal clamping pressure and complete closure of the moulds over the
entire parting line, pneumatically inflated tubes stroke each platen. They are inflated
after the mould carrier has been closed and deflated just before opening. The total stroke
is a few centimetres. The clamping force afforded by this system is 12,700 kg, using a
working pressure for the air bags of 0.3 MPa.
Individual thermoregulator units are used, one for each press. They are mounted on the
rear of each mould carrier. This solution affords complete autonomy to each press, making
it possible to control the temperature of the moulds prior to being placed on the production
line. Subsequently, when a press is placed on the moulding line, it is ready to commence
production.
The thermoregulators work in a closed circuit; the relevant refilling has to be performed
off-line. The control panel for each thermoregulator is mounted on the operator side for
easier access.
124
Polyurethane Processing: Recent Developments
The mould carrier concept for fixing and centring is designed for quick mould carrier
removal or exchange.
The estimated time to perform this operation is less than five minutes.
Service Station
A dedicated station where service operations on moulds and mould carriers can be
performed, without disturbing the production cycle, is foreseen for each moulding line
(see Figure 3.10).
Figure 3.10 All maintenance and setting operations on moulds and mould-carriers can
be executed off-line, in a dedicated service station that performs all the line’s
movements and operative functions.
125
Advances in Urethane Science and Technology
The service station comes with one complete mould carrier on its freestanding frame. It
is equipped with a dedicated thermoregulator for warming of moulds destined to be
inserted on the moulding line, along with all the required safety fences/light barriers,
controls, and connections for air, water and electrical power.
Cannon provides different mould-carrying systems, each customised to meet the customer’s
needs. The most common solutions are conveyors and turning tables. In the last part of
this section some innovative concepts, which provide a more compact layout, a minimised
investment requirement and a maximum degree of flexibility in a changing manufacturing
scenario, are presented.
Oval Conveyor
The conveyor is made of a central steel frame, which is positioned on the floor. It is
composed of a series of straight-line modules with a curved module at each end and
comes complete with a number of carriages running on an oval steel track (also mounted
on the floor). On each carriage, mainly composed of a rigid steel frame positioned on
four pivoting wheels, a single lid mould-carrier is mounted. Four idle wheels, running
along a central guide plate, maintain the system on its set path (see Figure 3.11).
Continuous movement is achieved by means of one drive system being placed between
every two carriages. Each driving system has two entrainment wheels running along the
guide plate. The entrainment wheels – operated via an AC motor – are maintained in
traction along the guide plate by a spring. An inverter is used to adjust the speed from 4
m to a maximum of 9 m/min, while the AC motor is fitted with a brake so that, in case
of an emergency, the plant can stop within a few centimetres of travel.
This system can easily be expanded to incorporate any new business obtained for that
range of moulded products. Extensions are achieved by fitting an even number of new
carriers to the line and extending the supporting frame accordingly.
Developed in the early 1980s as a revolutionary tool for the production of foamed
refrigerator doors, the well-known concept of the rotary polymerisation system is now
being proposed as a compact and simple mould-carrying system for the production of
automotive seats (see Figure 3.12).
126
Polyurethane Processing: Recent Developments
Figure 3.11 A modular oval conveyor for flexible foam moulding. It can be easily extended,
adding modules of track and pairs of mould carriers, when higher productivity is required.
127
Advances in Urethane Science and Technology
Six or more mould carriers can be fixed on the surfaces of a wheel that rotates vertically –
similar to a Ferris wheel, rather than horizontally, as with a merry-go-round. The opening
and closing movements can be hydraulic or pneumatic. Mould conditioning can be
accomplished using a rotating collector that feeds each mould carrier with water at the
desired temperature. Manual service operations are performed on the lower mould half in
the first station. Foaming is achieved in the subsequent station using a platform-mounted
robot. Polymerisation takes place in the remaining four (or more) elevated stations as they
are passed through a compact suction hood that removes the escaping fumes. Several
advantages can be highlighted for this new version of mould carrying system:
• simplified mechanical construction that does not incorporate wheels which roll
along the floor picking up and transporting scrap foam and/or dirt
• reduced power cost thanks to its vertical layout which minimises the suction area,
requiring only one fan for extraction of the fumes
A number of dedicated moulding lines have been designed for the production of in situ
moulded foams. In situ moulding technology adds a delicate operation to the list of
conventional operations (mould cleaning, release agent spraying, insert positioning,
foaming and demoulding) that must be executed to mould a standard foamed item: the
manual positioning of the textile container into which the foam will be dispensed. This is
a delicate operation that requires some time, yet should not penalise the cycle.
A practical solution consists of a carousel line with a row of service positions where the
operators can work on moulds that have been temporarily taken off-line (see Figure 3.13).
When the press leaves the curing area, it passes in front of the first free operator and is
automatically disengaged by the dragging system. The moulds can be serviced, taking all
the time required, then, when the textile inserts have been positioned, the carrier can be re-
inserted in the first available position in the line.
The availability of multi-component mixheads with variable geometry opens the path to
another interesting option: the addition of natural carbon dioxide for expansion of the
128
Polyurethane Processing: Recent Developments
Figure 3.13 Oval moulding line with 16 double mould carriers, designed for in situ
foaming of textiles.
1 to 4 Service stations, where mould carriers are taken temporarily off-line for
mould service; 5 Spare station; 6 Foaming station; A Pouring robot; B Metering
equipment, on mezzanine; 7 to 16 Polymerisation stations; C Mould carrier
changing station; D Suction hoods
foam with an environmentally friendly blowing agent and reduction of the foam density
by as much as 20-25%. With this new mixhead, it is possible to use two different
approaches to add natural carbon dioxide to the formulations:
• EasyFroth [1, 2, 4, 11, 12, 13,14] for carbon dioxide – a technology that allows the
premixing of given percentages of blowing agent in one of the two components,
usually the isocyanate.
Both methods are currently in industrial production, each having operating and
investment ‘pro’s and con’s’ that must be evaluated according to the production volumes
and flexibility required. For example, a car seat with three different hardnesses can be
produced using two different formulations for the hard and soft parts - (see areas ‘A’
and ‘B’ in Figure 3.14) - and incorporate some natural carbon dioxide to decrease the
seat density below the thighs (area ‘C’).
129
Advances in Urethane Science and Technology
Until now, one of the major limitations in the polyurethane moulding process has been
the necessity to interrupt the working sequence between each moulding, to remove foam
scrap and apply release agent prior to foaming. The introduction of ‘Foam & Film’
technology makes manual intervention unnecessary, removing the one factor that has
always been a major weakness when working with polyurethanes in comparison to other
injected or extruded plastics [15-19].
The main idea behind this new approach consists of thermoforming a thermoplastic
or polyurethane film as part of the moulding sequence. By using a vacuum effect,
this film adheres perfectly and smoothly to the mould cavity, without any creases or
wrinkle formation. The mould is equipped with a dedicated frame device, specifically
designed to hold the film. A heating system ensures that the film reaches the desired
temperature prior to the thermoforming phase and subsequent injection of
polyurethane into the mould.
130
Polyurethane Processing: Recent Developments
Several industrial applications have been found and a number of fully automated plants
currently incorporate this technology. The manufacture of sound deadening parts for
the automotive industry represents one of the most exciting applications for this
innovative technology.
Mould cleaning is required because of the chemical nature of the polyurethane process.
The objective in developing the ‘Foam & Film’ technology was to eliminate the tedious,
time-consuming manual operations, which must be performed on the moulds as part of
a discontinuous polyurethane foaming process.
The need to properly vent the mould, to avoid air entrapment, often results in a thin
flash of polyurethane being formed around the moulded part, which needs to be manually
removed afterwards via a simple trimming operation. These manual operations involve
both cleaning of the mould and spraying of the release agent.
It is necessary to demould parts in the shortest possible time, when the foam is not yet
fully polymerised, because the thin cross-section of the flash makes them very fragile
during the first few minutes following the demould. Often small pieces of flash will
break off when the part is extracted, falling back onto the mould. If they are not removed
from the mould surface, they could mar the surface of the subsequent parts or they could
allow rising foam to leak from the mould if left along the seal. More extensive cleaning
operations are required every few shifts in order to remove deposits of release agent
from the mould surface and there require production to be stopped.
Various technologies have been applied to minimise the required downtime, but so far a
valid solution has not been found. The adhesive force between polyurethane and metal
(either aluminum or steel) requires a release agent to be sprayed onto the moulds every
single cycle (or every few cycles) to enable easy removal of the part. In addition, various
internal mould release technologies are available, but they do not apply to all formulations.
As stated previously, these operations take a long time, are expensive and are typically
executed manually and it is impossible to have a fully automated line without getting rid
of them. To create a fully automated line, both the need to manually clean the mould and
the need to spray release agents on it had to be eliminated.
131
Advances in Urethane Science and Technology
the overall sequence. By using a vacuum, this film adheres perfectly and smoothly to the
mould cavity, without any creases or wrinkle formation.
The system’s key concept is a specially designed and patented frame, integrated into the
mould, which is specifically designed to hold the film in position during this thermoforming
process. This frame can be either two or three dimensional, to better follow the shape
and the cavities of the mould (see Figure 3.15).
The frame receives the film after it is unrolled by pinchers and cut dimensionally. It
then moves it in front of an infrared lamp for several seconds to heat it to the required
temperature. A control system ensures that the film reaches the desired temperature
prior to the thermoforming phase. Special infrared heaters, Cannon’s MVL heaters,
are used to ensure high efficiency and very low thermal inertia. Once the film is heated,
the frame moves into the mould where the film, held in place along its four edges, is
vacuum-formed onto it. The use of the frame during the vacuum forming prevents the
film from folding or wrinkling (see Figure 3.16).
132
Polyurethane Processing: Recent Developments
Figure 3.16 A patented frame holds the releasing film in position so that when
thermoformed over the mould, it is kept under tension and does not form wrinkles.
There are two different types of ‘Foam & Film’ processes available, based on the different
types of film used: adhesive and releasing.
• In the adhesive type, the film sticks to the part and is unloaded with it, granting an
aesthetic finish and a waterproof covering (see Figure 3.17).
• The releasing film, on the contrary, remains vacuum-formed to the mould for several
shots (5-15, dependent upon the process and the materials utilised) and it is then
replaced when it begins to wear.
Obviously, with the adhesive type, the film has a high adhesion coefficient with polyurethane,
while in the releasing version the poorer the adhesion, the better. In both cases, the film
prevents the mould from ever being put in contact with the polyurethane foam. That is
why both cleaning and spraying are no longer necessary and all operations can be automated.
133
Advances in Urethane Science and Technology
Figure 3.17 Two shock absorbing inserts for automotive doors moulded in semi-rigid EA-
foam with the ‘Foam & Film’ technology as they appear immediately after demoulding. The
film protects the foaming against moisture and degradation. No release agents or mould
cleaning operations are needed to produce this part in a highly automated moulding plant.
TPU is a good film for both cold and hot processes, providing good adhesion to the foam
and excellent mechanical properties but the cost is high (around 0.65 US$ per square
metre). PE film can only be used in cold processes, it requires a special treatment for
perfect adhesion and it gives an overall poorer performance but it has the advantage of
being very inexpensive. TP film is only suitable for hot processes, it requires no treatment,
and it guarantees good adhesion at a low cost (less than 0.2 US$ per square metre).
This last type of film is what has been developed in depth, obtaining very good tear and
impact resistance, flexibility, elongation and welding ease. This is the solution that has
been used for most automated Foam & Film plants supplied so far. Using this film,
General Motors makes sound-deadening parts with fully automated equipment that has
a productivity of 8000 parts/day, running 24 hours per day with zero operators [17, 18].
134
Polyurethane Processing: Recent Developments
used: PE and TP. PE, once again, is very common and inexpensive, but gives poor release
from the foam, which minimises the benefits of this technology. TP film provides an easy
process with good release from the foam for several shots at a very reasonable price.
A good example of such a continuous moulding process that incorporates ‘Foam &
Film’ technology is a system developed for the production of industrial vehicle carpets.
These products are usually made out of a sandwich of two or more layers of different
materials. Individually, they provide different features: aesthetics and function (a textile
carpet or a synthetic mat), sound deadening (a layer of polyurethane foam) and protective
(a cheap layer to protect the foam from moisture and degradation).
Figure 3.18 A large full-automated production plant for truck floor-covering mats
using the ‘Foam & Film’ technology. The parts are made of thermoformed PVC (or
thermoplastic elastomer) foam-backed with PU, which is kept separate from the
mould with a film (adhesive type).
135
Advances in Urethane Science and Technology
These excellent results have recently been improved with the introduction of the new
‘Foam & Film’ technology. The technology makes manual intervention unnecessary,
removing the one factor, which has always been a major weakness with polyurethane
processing and added a significant cost to the parts that are perceived to be an economical
component of a vehicle. These components have a surface area of about 3 m2 and consist
of a surface layer of polyvinyl chloride (PVC) or thermoplastic elastomer (referred to as
the ‘heavy layer’), an intermediate sound-deadening layer of flexible, medium-to-low
density polyurethane foam and a lower thermoplastic film. The film is designed to prevent
the formation of flash during the moulding process, eliminate permanent residue usually
left in the mould, and to act as a release agent (once the part is in the vehicle, this film
inhibits water absorption, which is a very frequent problem with industrial vehicles). In
the past, polyurethane films, which were strong mechanically and performed well but
were quite expensive were used. With the new ‘Foam & Film’ technology, it is possible
to replace this polyurethane film with a thinner, less costly thermoplastic one, resulting
in high quality production at a lower cost.
The system is composed of two shuttle-bed clamps served by one metering unit, which
dispenses the pre-heated heavy layer in the mould. Pre-heating is carried out using a
special infra-red heater, which incorporates easily adjustable, special low thermal-inertia
resistances. When a vacuum is applied to the mould, the material adheres to the lower
mould half, taking on its shape and embossed design.
The protective film is automatically unrolled from an overhead source using a vertical
traversing frame with pinchers that pulls an appropriate length of film over the frame
and cuts it to length. The frame is positioned over the edge of the mould. Another bank
of heaters slides laterally in front of the mould and warms the film to the correct forming
temperature; a vacuum is applied at the end of the heating phase to conform the film to
the mould.
The use of two presses, as opposed to one, means that slack periods are eliminated and
use of the cycle time is maximised. While the film and heavy layer are being placed in one
of the moulds and pre-heated, foaming and polymerisation are taking place on the other.
The polyurethane dosing and foaming section of this plant is equipped with a moulding
technology that allows chlorofluorocarbon (CFC) blowing agents to be replaced with
liquid carbon dioxide. This helps reduce the density of the polyurethane considerably
and thus saves on material costs.
The overall production time is just two and a half minutes per part. Being a two-
station plant, this equipment can produce close to 50 finished parts per hour – without
a dedicated operator.
136
Polyurethane Processing: Recent Developments
3.2.5 Applications
This technology has already been used on turnkey plants for sound insulation parts,
carpet back foaming and seat cushions. No limitations for this technology are foreseen
and, actually, ‘Foam & Film’ can be implemented in almost every kind of polyurethane
process needing either release agent or a cover/surface film.
The system requires the presence of an open mould where the film-holding frame can be
inserted to position the film prior to the vacuum-forming phase. Obviously, this technology
is more easily applied to new equipment and new moulds since, most of the time, existing
moulds must be modified to provide the vacuum and hold the frames. When one surface
of the part is covered by an aesthetic or functional layer (carpet, plastics, etc.), obviously
the film is only applied to the opposing side where the foam would be in contact with the
mould (see Figure 3.19).
Figure 3.19 Turnkey installations using the ‘Foam & Film’ technology are producing
numerous parts for the automotive industry.
137
Advances in Urethane Science and Technology
3.2.6 Advantages
‘Foam & Film’ offers significant benefits, such as quality parts, reduced costs, shorter
process times, as well as regular and consistent production cycles.
• polyurethane or TP film can be used, the latter being thinner and cheaper but giving
the same performance.
• films which adhere to the product can be used, becoming an integral part of the finished
component; this can be a very useful feature for non-aesthetic parts that are to be
mounted in hidden positions and will benefit from this extra protection against humidity,
oil, aggressive chemicals and foam-aging agents such as oxygen or other gases.
• films, which adhere to the mould, can be re-used several times as a substitute for
release agent.
• the availability of a wide range of film sizes means no dimensional limitations on the
parts to be moulded.
• eliminates spraying of release agent, saving its cost plus those of all the relevant
dispensing equipment and special fume extraction systems (although regular fume
extraction must be maintained for the polyurethane process),
The combined use of polyurethane and reinforcing agents or fillers has been common
practice for a long time. A recently introduced technology - simultaneously injecting
foam and glass fibre in open moulds - did not give satisfactory results, according to some
of the early users. Cannon has developed an industrial solution - named InterWet - that
138
Polyurethane Processing: Recent Developments
Glass-reinforced polyurethane has been used for many years, utilizing different
technologies: reinforced reaction injection moulding (RRIM), open-mould pouring over
a flat mat (LD-SRIM), closed-mould injection on preformed glassmat-sandwiches (SRIM).
Cannon has for many years dealt with glass-handling polyurethane technologies,
developing in the early 1970s a special RRIM mixing head able to mix formulations
containing high percentages of milled fibre, and launching in the 1980s the HE metering
machines, closed-loop electronically-controlled piston-metering systems able to cope with
the abrasion deriving from glass and mineral charges. Later in the 1980s the Compotec
preformers were introduced, for the production of preformed glass mats required by
SRIM - structural-moulding applications.
A new technology has been introduced by Krauss-Maffei, long fibre injection (LFI) [28],
that simultaneously pours polyurethane and chopped glass roving on the surface of open
moulds; after the shot the press is immediately closed to allow for the expansion of the
foam, which surrounds all the fibres and produces a lightweight, resistant composite
panel. Most recently, Hennecke has introduced FipurTec [29, 30], a technology where
the fibre is chopped outside the head and projected into the flow of reacting chemicals,
where this touches the surface of the mould.
In essence these systems carry out a job similar to that performed for very many years
in the polyester ‘chop-and-spray’ applications, the only differences being that
polyurethane has a different profile of reactivity and viscosity build-up than polyester
and that, at least in one case, the glass roving is conducted through the mixing device
instead of being fed outside of it.
The interest behind this technology lies in the fact that a thin, resistant composite part
can be moulded in only one operation. Using preformers requires more equipment, space,
investments, etc. In addition, glass roving costs roughly half that of glass mat, and this
can represent a considerable saving when producing large, highly reinforced parts.
A few practical problems have been identified in this technology, mainly deriving from
the design of the mixing equipment currently available. In one case the chopped fibre is
139
Advances in Urethane Science and Technology
fed through a pipe that is co-axially positioned in a conventional ‘L-shaped’ head and
then the mixed resin is conveyed around the pipe that feeds the chopped fibre. In the
other case the fibre is projected separately from the polyurethane; in this case the glass
meets the polyurethane only at the end of the head’s final discharge duct. The separation
of the fibre and polyurethane before the exit of the head results in wetting of the fibres
not being optimal; this is visible during moulding, where part of the glass gets stuck
vertically into the base of rising foam, visibly dry.
The main problem comes from the dry chopped glass that - leaving the nose of the head
still not wet - flies everywhere near the mould surface. This results in negative effects on
workers, on cleaning and maintenance operations, and on part quality. This problem
can seriously hinder the application of this technology.
Cannon started looking into this technology in the summer of 1997, at the urging of
some car parts producers who were unsatisfied with the performances of the existing
ones. The main objectives of the project were:
• To supply in the shortest possible time a reliable solution for the injection of glass-
reinforced polyurethane foams:
• To design a multi-purpose solution, not limiting the choice of charge to glass roving only,
but including a wider range of natural and artificial fibres, as well as pulverised fillers,
• To provide maximum flexibility in charge feed, so that products with different content
of charges - in the same moulding or between subsequent moulds - can be produced,
Due to the short delivery time required, it was not realistic to conceive totally new
equipment for this application, and it was decided to concentrate on the use of existing,
proven pieces of hardware. Performance of the foaming section had been optimised in
previous years, and the main obstacle was the proper handling of glass and other fillers.
140
Polyurethane Processing: Recent Developments
The project was split into three parallel lines of development, to speed up results and make
use of the experience of various existing specialists for the specific engineering tasks:
Various qualities of roving are available on the market for different end-uses: the diameter
of the basic threads (bunches of individual glass fibres) and of the final roving (the rope
made with various threads) define the weight and the field of application, while the type
of coating (a thin layer of special resin applied on the fibres after the glass-extrusion
phase) determines the best compatibility of the roving with the polymeric matrix.
A specific type of Owens Corning Fibreglass (OCF) roving was selected as the reference,
having a 2400 Tex specific weight, i.e., the roving weighs 2,400 grams per kilometre, with
high ‘softness’ of the thread. Other types would have been easier to cut (or to be broken,
since a fibre of glass is not shear-cut - it is bent beyond its critical radius until it breaks), but
their individual fibres were more agglomerated in the thread and would have been more
difficult to wet. All parts were designed to cope with a wider range of roving.
The glass-feeding device was conceived to be installed over the polyurethane mixing
head, and its design was made in strict cooperation with the team in charge of designing
the head. It is basically composed of:
• A pulling/cutting device: two opposed rolls that rotate and trap between them one or
more roving, pulling them from their storage. A blade - held in one of the two rolls -
forces the fibres to bend beyond their critic radius until they break;
• A conducting system: a flow of compressed air that pushes the chopped fibre from
the cutting place to the mixing head.
141
Advances in Urethane Science and Technology
• A hydraulic motor, to drive the glass-roving pulling rolls, was chosen to provide
proper speed, torque and easy variation of the driving speed. Its high power/weight
ratio made it preferable to an electric motor;
• A special cutting device was conceived, based on a rotating blade-holder (very easy
to access and maintain) capable of accommodating up to four blades. By fixing more
than one blade in the proper holders one can chop the roving in different lengths, the
longer size being that equal to the circumference of the rotating blade-holder. The
part had to be dimensioned to resist the very high operating speed, up to 5,000 rpm,
without suffering from the presence of glass debris around rotating shafts. This design
proved useful later, in the further development of a variable-length chopping system;
• A glass-guide system using compressed air was designed to take away all the cut
roving from the chopping device and bring it through the mixing head down onto
the mould, with the additional task of providing a final purge of the discharge duct.
Most of the development work was concentrated in designing a mixing head able to wet
all the solid components before they left, to ensure the projection of a very well wetted
mixture rather than a blend of liquid only coupled with a dry, flying filler. The mechanism
which could ensure a thorough wetting of high percentages of solid fillers with a mixture
of polyurethane is to provoke high turbulence in the mixing head’s area where liquid and
solid meet, and then reduce the turbulence to allow for splash-free open-mould distribution
of the blend obtained. It was in fact observed that, with heads where the junction between
liquid and solid was in a turbulence-free zone, the wetting was not optimal.
By using the original mixing chamber for the liquid components and boring a hole through
the main cleaning piston - the plunger - so that a solid component could be fed through it,
a wide range of fillers can be used (see Figure 3.20). When the mixing operation begins, the
142
Polyurethane Processing: Recent Developments
Figure 3.20 The Cannon FPL head used for InterWet incorporates in its upper section
the glass cutting mechanism and relevant safety checks.
main piston is retracted until it clears the discharge duct at the junction of the two cylindrical
chambers. In a controlled sequence, the small piston sealing the mixing chamber retracts
and the two chemical components are fired at high pressure against each other through
injectors in the mixing chamber before leaving it through the discharge duct.
The metering and feeding operation of the solid component from the head’s upper part is
controlled mechanically, assisted by an intense flow of compressed air. Its feed is
synchronised with the arrival of the liquid blend from the mixing chamber. The solid
component meets the liquid blend and together they are co-injected into the mould.
Following co-injection, the small piston seals the mixing chamber, the large piston seals
and cleans the discharge duct and the feed of solid component is interrupted, although
the flow of air is maintained for a short time in order to clean any residual polyurethane
- which could interfere with the next pouring operation - from the nose of the head.
143
Advances in Urethane Science and Technology
Figure 3.21 Open-mould co-injection of PU and glass fibre is executed at high speed
and in a clean working environment, with an InterWet machine.
The innovative concept in this solution is that the solid component meets the liquid
formulation just in front of the mixing chamber, where the turbulent flow of the mixture
is directed through a 90° angle downwards creating a laminar flow. In this way, the
kinetic energy from the pressurised liquids is used to wet the stream of solid component
thoroughly and efficiently at a point just 20-25 mm from the point of impingement. The
blend of polyurethane and filler leaves the head already well blended, and there is no
evidence of ‘flying glass’ out of the head (see Figure 3.21).
This special design and technique ensures that the solid component is thoroughly internally
wetted, hence its name, InterWet.
144
Polyurethane Processing: Recent Developments
Robotics
The development of a proper device to carry the mixing head during the pouring operation
over open moulds required relatively short time. Basic requirements were multi-axes
capacity, speed, precision and easy programming.
The limited weight of the Cannon FPL mixing head allowed for the selection of a medium-
sized model able to carry on the wrist 125 kg of payload at up to 2 m/s, with acceleration
of 3 m/s2. This robot was fitted with two storage boxes for the roving, mounted on the
elbow area (actually this model allows to install there up to 3 glass-roving rolls, for a
maximum 60 kg payload). This is a very convenient glass-storage solution for laboratory
and small production applications. When high quantities of glass are required, a solution
able to reduce rolls-replacement times must be provided. A proper device has been designed
to accommodate the 1,000 kg of glass demanded for industrial heavy-duty tasks. In this
case the glass rolls - that could be cascade-joined one another - must be positioned the
closest possible to the head, so that the way the roving must take to reach the chopper is
not restrained by long guides.
145
Advances in Urethane Science and Technology
First Results
The assembly of the pieces of hardware was completed in less than eight weeks from the
beginning of the project, and immediately trials were scheduled in the laboratory using a flat
test-plate mould.
The glass was carried from its storage to the chopper by means of flexible plastic pipes, and
the path was designed so that there would be no obstructions and friction for the roving.
Glass roving X900A, a 2400 Tex, very soft roving with average attitude to the cut, was
supplied by Owens Corning Fibreglass.
The first tests were run using approximately 30 parts of glass over 100 parts of polyurethane.
Surprisingly, the results were extremely satisfactory from the very beginning (see Figure 3.23).
Figure 2.23 Homogenous wetting of all fibres was obtained since the very first trials
with InterWet, thanks to the interval geometry of the mixing head.
146
Polyurethane Processing: Recent Developments
After fixing minor problems with the cutting control, an optimum distribution of blend
was obtained, and the glass fibres were all wet (this was clearly visible because
polyurethane was pigmented in black, and absolutely no white spot of dry glass was
shown at the end of an open-mould pouring operation). Being all wet, the fibres did not
tend to stick vertically in the rising foam, therefore less possibility remained for the air to
remain trapped in the mixture: the moulded parts were, in fact, totally free of air bubbles.
A number of trials were satisfactorily run using thin, transparent polyethylene film as
liner for both mould halves, so that when the liner was peeled off it was possible to
evaluate the quality of the compound and its surface (see Figure 3.24).
The second series of trials was run on a mould for production of automotive door panels
where a layer of expanded or compact PVC sheet was unrolled on the lower mould-half
and cut to measure, its edges were blocked with a frame and it was vacuum-formed to
perfectly adhere to the mould surface. After this operation the blend of polyurethane
and glass was poured on the rear side of the PVC liner, and the mould carrier closed for
polymerisation. On this occasion - since the first results were very encouraging - the
Figure 3.24 No holes are visible in this InterWet non-presses laboratory sample,
photographed against a strong light. The diffuser-mounted on the InterWet provides a
very distribution of material all over the mould.
147
Advances in Urethane Science and Technology
percentage of glass was increased to 35% of the total blend, i.e., approximately 60 parts
of glass to 100 parts of polyurethane. Again, very nice parts were obtained at the early
stages (see Figure 3.25).
Figure 3.25 Nice parts, free from air entrapments between reinforced PU and outer
skin were obtained at a very early stage of the InterWet development work.
First Improvements
One aspect of the laydown - the width of the pouring path - was worth immediate
attention; since the reactive blend leaving the head covered only a few centimetres of
mould at each pass, a rather high number of passes was required to properly cover the
whole surface which forced use of the robot at very high speed to complete the whole
pattern before creaming started. To avoid the overlapping of two contiguous layers of
mixture, sometimes a narrow gap remained between the layers. This did not create
problems but the distribution of glass in those thin strips was probably different than in
the other areas of the mould.
148
Polyurethane Processing: Recent Developments
Another advantage derived from this pneumatic distributor is that the jet of air pushes
even further down the fibres, which results in a very thin layer of mixture applied on the
mould surface. This means that the male half of the mould - when closing over the
female part - does not ‘wash down’ the vertical walls from the layer of reacting foam and
glass which covers them, since this layer is quite thin and does not constitute an obstacle
for the lowering plug. Parts as thin as 1.5 mm can be moulded with optimum surface
aspect and homogeneous distribution of glass across the whole surface.
The first fibre-cutting device, as stated above, had the possibility of accommodating
more blades. Therefore, the length of cut fibre obtained was equal to the circumference
of the cylindrical blade-holder when only one blade was installed, or half of its
circumference with two blades, or one-quarter of it with four blades. Since this was
foreseen in the original wish list, the option of allowing the user to vary the length of
glass fibre at will, even ‘on-the-fly’ on the same moulding, was soon taken into
consideration. A simple, sturdy mechanical device was designed to allow for this variable
cutting. This mechanism which operates via an impulsion given by the robot’s programme
at a given point of the pouring pattern, cuts the roving in shorter sections; this variable-
length cutter provides different cuts of fibre: from L to L/12, where L is the longest
possible choice and the combinations depend on the use of the 12 blades installed on the
holder.
Operated by a hydraulic motor, the pulling device (two opposed rolls that rotate and
trap between them one or more roving, pulling them from their storage) is able to provide
variable output of glass. By electronically controlling an inverter or a proportional valve
it is quickly possible to vary the motor speed, providing more or less fibre in different
149
Advances in Urethane Science and Technology
parts of the moulding. The motor can be even stopped (thus feeding only polyurethane
foam) or left running after the end of the polyurethane shot (providing dry, unblended
glass fibre over some areas of the mould where an extra reinforcement would be desired).
• Extended safeties
The main concerns in operating this kind of plant derive mostly from the irregular feed
of glass to the head. For this reason the latest InterWet machines have been equipped
with detectors able to command an immediate stop to the machine in case of:
These new developments have resulted in a higher degree of flexibility for the InterWet
machines whose advantages versus the existing and available similar methods can be
summarised as follows:
• Optimum overall mechanical properties, due to an optimum glass fibre wetting with
PU, because of the mix-head design
• Better distribution of the mechanical properties of the moulded part, i.e., no areas
without or with little glass, due to a more homogeneous glass fibre distribution in
the mould, because of the special distribution system.
• PU only
150
Polyurethane Processing: Recent Developments
• Easy operation because it uses a standard (modified) FPL mix-head with obvious
consequences in terms of simplicity of maintenance, reliability in use and availability
of spare parts when required.
• Flexibility in operation because it is designed for more than one type of additive
component: it can easily be converted to process fillers or fibres or chopped scrap
foam. Its range of application includes continuous reinforcing fibres, natural fillers
(mineral or organic), pulverised plastics, etc.
• All the operational advantages deriving from the use of a light mixing head (robot
size, speed, productivity, etc).
When compared with other glass-reinforcing technologies the following advantages must
be added:
• Economics:
in stock (stored preforms can be damaged by sunlight, dust, moisture, and often
must be thrown away)
151
Advances in Urethane Science and Technology
• Quality (foam mixes better when co-injected with glass roving rather than when
injected into a pressed mat preform)
• Flexibility: it’s easy to mount on a carousel various moulds and produce small series
exploiting the flexible properties of the robot.
It is evident that there are applications where both RRIM and mat preforms are still
preferred because of their specific advantages. InterWet has been conceived to fill the
gap between the two technologies, and allow for more extended use of reinforcements
and fillers in applications where neither RRIM nor mat preforms could be used [31].
3.3.1.5 Applications
When this technology was first introduced it immediately appealed - for the above
mentioned reasons – to automotive and transportation end users. Volumes, cycle times
and final properties meet this market segment’s requirements, but until the currently
available solutions address some of the above-illustrated problems this industry would
not select it for mass production.
A market segment where this application would be really interesting is the manufacture
of special sanitary-ware parts, for at least three different purposes. The quickest to
understand for its immediate advantages is the combination of an outer plastic shell -
with either aesthetic or functional purposes - back-foamed with reinforced polyurethane;
examples include whirlpool tubs, bath tubs, shower trays, and vertical elements for shower
cabins integrated with accessories. These products are characterised by medium-small
production series, wide range of models, and large surfaces. The high operating flexibility
allowed by the programmable robot is very adaptable for the production of this type of
moulding: the larger the part, the more advantages can be found in this technology.
Medium-deep-draw shaped parts can be easily thermoformed from large thermoplastic
sheets (even in medium-small series) and immediately reinforced in an integrated
production line including thermoformer, foaming station and curing area. Dimensions,
productivity and investment must be tailored to the project’s requirements.
Another area of interest for this new technology is the re-use of chopped foam scrap,
either flexible or rigid; the design of the solid component’s feeding device - as foreseen in
the project’s objectives - has been made in a way that, using an appropriate volumetric-
feeding mechanism, powders, granulates and recycled foams can be dosed in the mixing
chamber through the same channel bored in the mixing head’s cleaning plunger. It is
possible to add roughly pulverised scrap foam to the virgin formulation directly in the
mixing head. This would solve a recycling problem, as well as lowering the cost of
152
Polyurethane Processing: Recent Developments
Figure 3.26 In addition to glass roving a number of other materials can be co-injected
with PU using the InterWet technology, scrap PU foams, sand, ground cork and other
fillers that would be difficult to add in the component tank.
formulation. Positive trials have been made in a development phase, with filler loading
as high as 10% of the final part’s weight (see Figure 3.26).
Another interesting potential application is the use of continuous fibres, cut in medium-
long size, to reinforce thin pads and cushions moulded with flexible foams. The continuous
development of lighter, thinner parts - where weight-reduction and volume-exploitation
is the object of intense development - requires the use of formulations characterised by
very high mechanical properties that unreinforced foam cannot meet. The use of long
natural or synthetic fibres to reinforce these flexible mouldings had applied for a long
time, but only under the form of textile inserts manually inserted in the mould prior to
the foaming operation. Being able to co-inject foam and long fibres - applying longer or
shorter sizes according to the mechanical properties required - represents an advantage
that comes at a very attractive investment cost.
Note: The Cannon FPL Mixing Head is covered by several international patents.
153
Advances in Urethane Science and Technology
References
9 M. Taverna, Presented at the Utech Asia ’97 Conference, Suntec City, 1997,
Paper No.12.
14 M. Taverna and P. Corradi, Presented at the Utech Asia ’96 Conference, The
Hague, 1996, Paper No.10.
154
Polyurethane Processing: Recent Developments
28 J. Stark and F. Peters, Presented at the Utech ’99 Conference, 1999, Singapore,
Paper No.6.
155
Advances in Urethane Science and Technology
156
4
Recent Developments in Open Cell
Polyurethane-Filled Vacuum Insulated Panels
for Super Insulation Applications
Paolo Manini
4.1 Introduction
Rigid polyurethane (PU) foam is the preferred insulator material in a wide range of
applications encompassing household, industrial and commercial appliances (refrigerators,
freezers, display cases, vending machines), transportation (refrigerated trucks and reefers,
shipping containers), insulation in buildings and in industrial plants.
In applications, the lower insulation efficiency of the CFC-free foams generally leads to
higher energy consumption levels unless specific compensation actions are taken.
This issue is particularly severe for the household appliance industry, which has also
been called upon over the past years to progressively reduce the energy consumption
of their products.
With the energy efficiency targets getting more and more demanding and only marginal
improvement obtainable from the optimisation of conventional technologies, the appliance
manufacturers are seriously looking at alternatives which may provide additional benefit
in terms of energy savings.
The appliance industry is not the only one involved in the debate on energy consumption
reduction. The growing attention to the global warming issue is forcing policy makers to
take actions to cut carbon dioxide emissions and minimise the greenhouse effect. With
this in mind, the Kyoto Conference has posed ambitious targets for energy saving, the
achievement of which calls for improvements in many fields, from transportation, to
refrigeration and building insulation.
The ban on CFC and the demand for a more efficient use of energy, both issues coming from
environmental concerns, are therefore posing serious challenges to the Industry as a whole.
157
Advances in Urethane Science and Technology
Vacuum Insulated Panels (VIP), having a thermal conductivity three to five times lower
than conventional PU foams, allow the achievement of superior insulation performance.
They can be used to partially replace the conventional insulation materials to provide
a more efficient insulation structure, which allows energy saving without the need to
increase the insulation thickness. Alternatively, VIP can be used in those applications
where it is important to reduce the insulation thickness to a minimum value without
loosing thermal performances.
Several fillers have been proposed in the past, either in the form of compressed powders
or fibres, however all have some disadvantages in terms of cost, process complexity and/
or weight. This has prevented VIP technology from finding widespread acceptance in
most applications. The recent development of open cell rigid foams, first introduced into
the market in the early 1990s has sparked off new interest in this technology. However,
in order to fully exploit the unique properties of these insulating materials, the re-
evaluation and re-design of the other key elements of a vacuum panel, i.e., the barrier
film, the gas adsorbent and the manufacturing cycles, has also been necessary. This process,
which is still ongoing and has potential for further improvements, has required the
commitment of several sectors of the industry in the last five to six years.
In this chapter, the present status of the open cell PU foam-filled vacuum panel technology
is discussed and some recent developments in film, getter and processing technologies
are reviewed. Special focus is given to the vacuum issues, which are key for the proper
selection and treatment of the VIP components. Some specific aspects, related to the
manufacturing, characterisation and practical use of vacuum panels, as well as the
achievable benefits, are also considered. Examples of applications where this technology
is finding its place in the market and areas where further work is necessary to make
vacuum panels more cost effective are presented.
Vacuum panels have been studied as a means to improve thermal insulation for a long
time. Several insulating fillers, such as silica and perlite powders [1, 2, 3], fibre glass [4,
5], and aerogels [6, 7, 8] have been proposed as core materials for VIP, each of them
158
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
To overcome some of the drawbacks associated with the use of these materials, new
families of fillers, based on PU [9, 10, 11, 12] and, more recently, polystyrene (PS) [13]
open cell foams, have been developed and proposed to the market. Both open cell PU
and polystyrene foams are now being considered as interesting options due to their
moderate outgassing, good thermal insulation values, low weight, ease of handling
and cost effectiveness.
This chapter will focus on the open cell PU foams, even though many general aspects and
recent developments of the technology, dealing with films, getters and manufacturing
processes, can be also applied, with minor changes, to other micro porous fillers.
The thermal conductivity, or λ factor (mW/m-k), for some of the most popular filler
materials is shown in Figure 4.1 as a function of pressure (Pa). The thermal conductivity
of cyclopentane (CP) blown closed cell PU foam is also given for comparison.
Figure 4.1 Thermal conductivity as a function of pressure for some filler materials
159
Advances in Urethane Science and Technology
Regardless of the type of filler, all curves have a similar behaviour, with a low pressure
region, where the λ factor is constant, followed by a region where the λ factor increases
with increasing pressure.
This common trend can be explained considering that the total apparent thermal
conductivity in micro porous materials is the sum of four physical contributions, i.e., the
thermal conductivity through the solid, λs , the thermal conductivity through the gaseous
phase, λg, the thermal conductivity by thermal radiation, λr and by gas convection, λc:
λ = λs + λr + λg + λc (1)
The last term of Equation (1) can be neglected for core material with cell size smaller than 1 mm.
An extensive treatment of the above physical mechanisms and their role in heat transfer
in low density closed cell foams has been provided by Glicksman [14].
In the case of the open cell PU foams detailed expressions for the first two terms of
Equation (1) have been given by Kodama and co-workers [9].
Both terms are independent of pressure and strictly related to the foam density and
morphology (cell size, structure and degree of anisotropy).
[( ) ]
λ s = 1 − Vg2 / 3 K s / η (2)
To reduce the heat transfer through the solid it is therefore necessary to increase the
cellular anisotropy and to lower the thermal conductivity of the polymer structure.
Following Kodama and co-workers, the heat transfer by radiation is given by:
λ r = 2 / 3 Vg1/ 3 H r d (3)
where d is the cell size and Hr is the coefficient of thermal conduction by radiation which
depends on the emissivity of the polymer and the morphology of the cell. To minimise
the λr contribution, the cell size and Hr have to be reduced. The former can be achieved
by adjusting the formulation of the foam, the latter by reducing the opening size of cell
membranes to make radiation transmission through the cellular windows less efficient.
160
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
The heat transfer by gas conduction, λg, becomes progressively more important as the
pressure increases, as shown by the following set of equations [15]:
( )
λ g = λ goΠ / (1 + 2L / d) (4)
where λgo is the thermal conductivity of air at atmospheric pressure, Π is the porosity of
the foam and L is the mean free path of air, linked with pressure P by:
L = (kT ) / ⎛ 2 π P σ 2 ⎞ (5)
⎝ ⎠
where T is the absolute temperature, k the Boltzmann constant and σ the collision diameter
of air (about 4 x 10-10 m2).
The λg contribution vanishes for L>>d. For currently available foams, having an average
cell size in the 100-200 μm range, this means that pressure has to be kept at 1.0 Pa or
below to completely eliminate the heat transfer by gas conduction.
To ensure this condition the foam has to be properly evacuated and all gas sources
deteriorating the vacuum during the panel life effectively compensated by a suitable
adsorbent, or getter, having sufficient gas capacity and efficiency.
As shown by Equations (1-5), the insulation properties of an open cell foam depend on
a complicated interplay of different parameters such as the PU thermal conductivity,
foam density, cell size and cell size distribution, cell morphology and anisotropy, which
have to be optimised during the foam preparation to obtain the best trade-off among the
three thermal conductivity contributions, λs, λr and λg.
In this process, the mechanical properties of the foam cannot be neglected, being essential
to ensure structural stability of the vacuum panel.
In fact, after sealing, the vacuum panel is exposed to the hydrostatic load of the
atmospheric pressure and has to withstand it for a long time, which can be 15-20 years
or even more depending on the application.
Dimensional stability tests carried out on open cell PU foam-filled vacuum panels showed
that creep problems should not occur for properly prepared panels, provided the open
cell foam preparation has been optimised [15].
The open cell PU cores are produced according to various preparation technologies,
such as lamination and block foaming. In both cases, after the foam has been grown, a
cutting operation is usually necessary to remove the outer skin of the material, which
161
Advances in Urethane Science and Technology
contains a high percentage of closed cells, and to slice the PU into slabs of the proper size
and thickness. The removal of the outer skin may be a labour intensive process which
produces waste, lowers the process yield and increases the overall production costs, this
issue being particularly important in the case of the lamination process. Great attention
is therefore now being paid to design and the open cell foam production process, to
reduce waste and increase productivity.
A different approach to making open cell foam slabs has been proposed recently [11],
which is based on the use of the PU fluff obtained from the recycling of used refrigerators
(Recycled Urethane Fluff, or RUF panel).
During the recycling process, the rigid insulation foam contained in the old refrigerators
or freezers is mechanically ground into a fine powder, to completely separate and recover
the CFC present in the foam. The resulting fluff, composed of completely open cells, is
dried and introduced into a binding machine where it is sprayed with a certain amount
of isocyanate (15-20% by weight) and thoroughly mixed. The blend is then transferred
to a mould, heated at 120 °C and compressed at 0.5 MPa for 10 minutes to consolidate
it and remove residues of the process. After de-moulding and curing, the RUF panel can
be used as core material in a VIP.
The fluff obtained from recycling one single refrigerator can be used to produce enough
vacuum panels to insulate a new appliance, thus generating a virtual recycling loop.
Due to the use of very fine compressed powder, the density of this open cell foam is three
to four times higher than those produced by lamination or block foaming.
Selected physical properties of some open cell PU foams are listed in Table 4.1.
162
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
To fully exploit the insulating performances of the open cell PU foams, pressure in the
panel has to be kept preferably below 1.0 Pa during its life. To achieve this demanding
target, the foam must be 100% open celled and with a very low outgassing rate.
In spite of the increasing use of open cell PU foam, data from the literature on its vacuum
properties are quite scarce.
To estimate the outgassing contribution, specific tests can be carried out using high vacuum
benches equipped with a quadrupole mass spectrometer.
163
Advances in Urethane Science and Technology
Figure 4.2 Scheme of the stainless steel high vacuum bakeable bench equipped with
mass spectrometer for outgassing tests on vacuum components. IG: ionisation gauge.
To quantify the relatively small amounts of gas species released from the sample, it is
necessary to use very clean, bakeable, high or ultra-high vacuum systems made with stainless
steel and/or glass components. This ensures better leak-tightness and a negligible gas emission
from the bench. A sketch of a typical experimental apparatus, having two separate pumping
groups and base pressure at 1 x 10-6 Pa or lower, is shown in Figure 4.2.
The sample to be analysed is mounted in a glass bulb which is connected to the test
bench and evacuated for 10 minutes with a turbo and a rotary pump to a final pressure
of 10-4 Pa. The bulb is then isolated from the pumps and the total pressure increase, due
to the sample outgassing, is monitored with a capacitance manometer for some days.
The use of the capacitance manometer avoids changes in the gas composition which
might occur if a hot filament pressure gauge is used. From time to time a small amount
of gas is sampled from the test volume, through valves V4 and V6, and passed to the
mass spectrometer for partial pressure measurement. Before running the outgassing tests
164
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
Figure 4.3 A stainless steel outgassing bench equipped with a quadrupole mass
spectrometer and an external oven for vacuum bake-out
(Reproduced by courtesy of SAES Getters SpA, Italy).
on the PU sample, a blank run has to be carried out. The result of the blank run is then
subtracted from the outgassing test to get the actual gas emission from the PU sample.
The use of two distinct pumping groups is essential to minimise the system contamination
(mainly water vapour) which occurs when the bulb is opened to the air, to mount the
specimen. In fact, during this operation, valve, V4, is kept closed, reducing the surface
area of the bench which can absorb water and atmospheric gases.
Before being used in a vacuum panel, the open cell PU foam needs a preliminary heat
treatment in air, generally carried out at 120-150 °C for 10-60 minutes to remove water
and other volatile species which otherwise would desorb and rapidly cause the vacuum
to deteriorate. The result of a typical outgassing test carried out at 23 °C on a foam
sample baked at 120 °C for 30 minutes is shown in Figure 4.4 for all desorbed gases but
water. Water is difficult to quantify since it sticks to the walls of the system and only
partially reaches the mass spectrometer. Water can be estimated as the difference between
the total absolute pressure and the sum of the partial pressures of the other gas species,
which can be accurately quantified with the mass spectrometer.
165
Advances in Urethane Science and Technology
Figure 4.4 Outgassing rate for an open cell PU foam. Water is not shown.
For each gas species, the amount released by the sample at a given time is obtained by
simply multiplying the test volume (V) by the gas partial pressure (qi(t) = Pi V, qi(t) is the
outgassing rate (Pa-l/h) at time t, Pi being the measured i-th partial pressure).
The experimentally determined qi value for each gas species can be interpolated over
time according to the semi-empirical law quoted in the literature [16]:
where qi(t) and q0i are the outgassing rate (Pa-l/h) at time t (hour) and the desorbed
amount (Pa-l) after 1 hour of the i-th gas species, respectively. The dimensionless parameter
αI is related to the desorption mechanism of the i-th gas species, its value usually ranging
from 0.5 to 1, depending on the gas species and the material considered [16, 17].
Integration over time of Equation (6), provides an estimate of the PU sample outgassing
load Qi for each gas species after a given time. As an example, the estimated gas released
after 1 to 20 years for the sample of Figure 4.4 and for a 50 x 50 x 2 cm3 panel size is
given in Table 4.2. The main gases, besides water, are carbon dioxide, nitrogen, carbon
monoxide and hydrogen. It is interesting to note that, given the nature of the desorption
166
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
Table 4.2 Outgassing parameters and estimated gas load Qi (Pa-l) after 1, 5
and 20 years for a 50 x 50 x 2 cm VIP
Gas q0i (Pa-l/h) αi Qi (1) (Pa-l) Qi (5) (Pa-l) Qi (20) (Pa-l)
CO2 8 x 10-3 1 22 25 27
CO 2 x 10-3 1 54 64 75
N2 9 x 10-3 1 10.8 12.8 15
-5
H2 5 x 10 0.5 2.3 6.2 13
process (Equation (1)), most of the gas is given off during the initial period of operation
of a VIP.
Results shown in Figure 4.4 and Table 4.2 can vary from sample to sample depending on
the open cell PU foam preparation technique, its microstructure and density and the
pre-treatment.
Since desorption is a thermally activated process [18], the outgassing rates increase as
the temperature increases. The outgassing contribution has therefore to be carefully
evaluated in all those applications where the vacuum panels operate continuously at
temperatures higher than room temperature, e.g., 60-80 °C, or have to withstand high
temperature peaks, for example 100 °C, even for a relatively short period of time. Examples
of such applications are presented and discussed in Section 4.6.
An additional factor which may have an impact on the vacuum properties is the closed
cell content of the PU foam. All the open cell PU foams so far presented in the literature
for VIP applications are quoted to be 100% open cell based on pycnometric measurement.
On the other hand, the accuracy of this test method is generally close to ± 0.2% [15], so
that the possibility of a small fraction of closed cells cannot be completely ruled out and
has to be considered as an additional potential source of pressure build-up in the panel.
Closed cells might be present due to a non optimal foam preparation or to the incomplete
skin removal.
The barrier film plays an important role in vacuum panel technology since it has the task
of minimising air and moisture penetration into the vacuum core. The barrier must be
167
Advances in Urethane Science and Technology
durable and able to resist to puncturing and abrasion. It must be functional over a wide
temperature range and retain its physical properties, such as dimensional stability,
flexibility and sealability for years.
Thermal conductivity of the skin should also be very low to minimise the heat transfer
through the panel edges which can partially reduce the insulation efficiency of the
evacuated panel (‘edge effect’, see Section 3.2.1). These properties have to be coupled
with very low gas permeation and outgassing rates.
Gas permeation through the barrier envelope is one of the most important factors
responsible for the pressure increase in a panel during its life.
Depending on the structure of the barrier film and the materials used, gas ingress can
preferentially take place through the whole surface of the barrier film (permeation
through the surface) or through the heat-sealed plastic VIP flanges (permeation at the
edges), or both.
Several types of barrier materials, having different structures and gas transmission rates,
are commercially available from the food, packaging and electronic industry. Gas barrier
requirements for vacuum panels are however much more demanding.
Composite plastic films, obtained by laminating several polymeric sheets have been
proposed in the past for precipitated silica-filled panels, where pressures as high as 500-
1000 Pa can be tolerated without excessive degradation of the thermal conductivity. In
the case of open cell PU-filled panels, requiring much lower pressure values, this barrier
structure is generally inadequate, if a lifetime exceeding one to two years is needed.
To improve the permeation properties, barrier materials, such as aluminium and silicon
oxide, can be vacuum deposited on the polymer surface in the form of a thin film (0.01-
0.05 μm thick) [19].
However, most commercially available metallised samples do not provide enough barrier
properties for long term applications, mainly due to water transmission through the large
density of pin holes and micro-defects present in the sputtered aluminium layers [20].
168
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
the metal, mainly due to its good workability, which allows production and lamination
of virtually pin hole-free foils having thickness of 6 μm or less [21]. The presence of
the aluminium foil dramatically improves the gas barrier property of the film, well
beyond the sensitivity limits of the analytical techniques commercially available for
the measurement of gas permeation.
In the case of water, which is one of the most important permeating species, two standard
methodologies are at present widely used to measure its transmission rate through a material,
as described by ASTM E96-00 [22]. According to the first procedure (‘the desiccant method’)
the barrier sample is sealed to the open mouth of a test plate, which is cylindrical in shape,
in which the desiccant is placed, and the assembly kept in a humidity and temperature
controlled environment. The permeation of water through the sample is measured by the
weight increase of the assembly. In the second procedure (‘water method’) the test plate
contains water and the transmission rate through the barrier sample into the controlled
environment is measured by the weight decrease of the assembly.
Another very common technique is based on the use of an infrared sensor as described
by ASTM F1249-90 [23]. The barrier sample is sealed between a dry and a wet chamber
kept at known temperature and relative humidity. The two chambers make up a diffusion
cell which is placed in a test station where the dry chamber and the top of the barrier are
flushed with dry air.
Water vapour which penetrates through the barrier film blends with the dry air and is
transported to a pressure-modulated infrared sensor. The infrared radiation is absorbed by
the water molecules and the sensor produces an electrical signal which is proportional to the
concentration of water in the gas phase. The intensity of the signal is then compared with the
signal generated by measurement of a sample having a known water transmission rate. This
allows the determination of the water transmission rate through the test specimen. The
detection limit for water transmission generally quoted for commercially available instruments
based on this principle is in the 0.01 g/m2 day range (at 90% relative humidity and 23 °C).
These techniques are generally more than adequate for many plastic or standard quality
metallised films but can hardly discriminate between samples containing aluminium foil,
which generally have water transmission rates much lower than this value.
To better estimate gas permeation for high quality barriers a novel technique has been
developed [24, 25], which is based on the measurement with a quadrupole mass
spectrometer of the helium transmission rate through the sample.
169
Advances in Urethane Science and Technology
Figure 4.5 Scheme of the experimental bench for the helium permeation test.
The bench is a stainless steel ultra-high vacuum apparatus equipped with rotary, turbo-
molecular and non-evaporable getter (NEG) pumps. The sample, a circular coupon of
about 30 cm2 area, is mounted between two flanges and the helium pressure is applied
(generally in the 0.0001-0.1 MPa range) on one side of the sample, the other side being
in view of the quadrupole mass spectrometer. A variable conductance C (l/s) is mounted
between the mass spectrometer and the evacuation group, so that the helium flow rate F
(Pa-l/s) through the conductance can be measured with the mass spectrometer according
to the equation:
170
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
where P1 and P2 are the pressures before and after the conductance C, respectively.
Under equilibrium conditions, which can be reached after some minutes or some hours,
depending on the nature of the sample, the gas flow, F, provides the helium transmission
rate. A typical graph showing the achievement of steady state conditions is given in
Figure 4.7.
171
Advances in Urethane Science and Technology
Figure 4.7 Achievement of the equilibrium condition in a helium permeation test for a
laminated film composed of 15 μm Nylon, 12 μm PET, 6 μm aluminium, 50 μm high
density polyethylene (HDPE) (or Nylon 15 μm/PET 12 μm/Al 6 μm/HDPE 50 μm).
The use of the mass spectrometer ensures very high sensitivity for helium, better than
10-11 Pa-l/(s m2 Pa), mainly due to the negligible presence of helium as a constituent of
the gas background. Due to its high sensitivity and the relatively short measuring time,
this technique can be used effectively to support the development of improved laminates
and also for quality control in a production environment.
The estimation of the transmission rates for gases other than helium can also be obtained,
provided a preliminary calibration is carried out. This is particularly important for water,
which is the main gas permeating through the barrier skin. To run the calibration
procedure, the helium transmission rate is measured for various samples having different
and known permeation rates for water. The linear correlation between helium and water
transmission rates is then established, as shown in Figure 4.8.
The correlation factor for water was found to be close to 500 for most of the PE/PET/
Nylon-based films analysed by the authors, either metallised or incorporating an
aluminium foil [26]. It was then possible to estimate the order of magnitude of the water
172
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
Figure 4.8 Experimental determination of the correlation factor between helium and
water transmission rates.
transmission rate for an unknown low permeation rate sample by simply measuring the
helium transmission rate and multiplying it by 500.
Table 4.3 provides a comparison between the helium permeation rates in two films, a
laminate incorporating a 6 μm aluminium foil, PET 12 μm/Aluminium 6 μm/HDPE 50 μm
(Film B) and a multi-layered barrier composed of four aluminium-sputtered PET sheets
laminated onto a 50 μm PE (Film A). Tests were run at 24 °C. From the measurement of
173
Advances in Urethane Science and Technology
Table 4.4 Thermal conductivity of some materials used as gas barriers and/or
structural layers in barrier films.
Material Thermal conductivity Typical thickness
(μ /m-K) (μm)
PET 0.15 10-15
HDPE 1.05 40-60
PVDC 0.13 10-20
Nylon 0.43 10-20
Aluminium 273 6-7
PVDC: Polyvinylidene chloride
The metallised barrier here considered cannot be used for long term applications, where
10-15 years or more are targeted, since the amount of desiccant necessary to compensate
for the water transmission would be too large.
A much lower permeation rate is provided by Film B, which is therefore well suited for
long term applications.
In spite of the extremely good gas barrier properties, VIP prepared with such a film
suffers from an intrinsic limitation, i.e., the high thermal conductivity of the aluminium
foil, as shown in Table 4.4.
A fraction of the heat flow is, in fact, transferred from the hot to the cold panel surface by the
aluminium foil through the panel flanges, rather than through the core material (so called
‘edge effect’). As a result, the average insulation value of the panel is less than the expected
value based on the actual insulating properties of the core material (centre of the panel
thermal conductivity), this difference being more remarkable the smaller the panel size.
The edge effect can be evaluated, as a function of the panel size, the laminate structure and
the aluminium thickness, by numerical methods, such as the Finite Element Analysis (FEA).
Results of such an analysis are given in one specific example in Figure 4.9, where the
average thermal conductivity of a vacuum panel using a PET 12 μm/Aluminium 6 μm/
HDPE 50 μm laminate is plotted against the panel area.
174
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
The thermal conductivity and the thickness of the PU core material considered here are
6 mW/m-K and 2 cm, respectively.
Figure 4.9 shows that the panel thermal conductivity approaches the value of the core
material only for a sufficiently large panel area (≈ 1-10 m2). For very small panels, the
insulating properties of the core material are spoiled by the edge effect and the energy
savings in the real application may be minimal.
For this reason, efforts are now being made to improve the aluminium–based laminates
and/or to develop products not containing the aluminium foil but still having sufficiently
good gas barrier properties, so as to achieve an acceptable trade-off between energy
saving performances and acceptable lifetime. This challenging objective can be addressed
by selecting and properly combining together suitable polymers having improved gas
barrier properties.
A typical example of this process, recently reported by Lamb and Zeiler [27], is given in
Table 4.5, which shows how significant improvements can be obtained in non aluminium-
foil containing skins working on the film structure.
175
Advances in Urethane Science and Technology
Table 4.5 Oxygen transmission (OTR) and water (WTR) transmission rates for
some film structures [27]
Film structure OTR x 10-10 m2 day (STP) WTR g/m2 day
PET 34.8 43.4
PET + PVDC 2.3 9.3
PET + PVDC + metallised 0.04 0.62
layer
Mylar 200RSBL300 0.00015 < 0.15
STP: Standard, temperature, pressure
Due to the still developmental nature of most of these barriers, more work is necessary
to assess their usability for long term applications. Extensive characterisation of the
mechanical, sealability and stress resistance properties are also important issues to consider.
The permeation of atmospheric gases through the polymeric sealed flange of a vacuum
panel can be an important contribution to the pressure increase in a VIP for long term
applications. The gas permeation rate depends on the pressure gradient across the vacuum
panel, the type of polymer used as a sealant, the flange geometrical parameters (exposed
surface and width) and temperature. Polyethylene is at present one of the preferred sealing
materials due to the good trade-off achieved among sealability, mechanical properties,
gas permeation, outgassing rates, reliability and cost.
Gas permeability values quoted in the literature for PE are scattered over an order of
magnitude, depending on the actual density of the sample and the gas considered [29,
30, 31]. Typical permeability values corresponding to medium and high density PE are
shown in Table 4.6.
Results of the calculation for argon, nitrogen, oxygen and water and the total pressure,
are given in Figure 4.10 and show the role of permeation in deteriorating the vacuum
level in the panel. The width and height of the PE flange here considered are 10 and 0.1
mm, respectively and the environment is air at 21 °C and 50% relative humidity.
176
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
Table 4.6. Permeation rates for gases (cm3/m2d, STP) and water (g/m2d)
in a 25 μm thick PE sheet
Gas MDPE x 10-10 HDPE x 10-10 Temperature °C
N2 387 96.75 21
O2 1548 425.7 21
Ar 3096 774 21
H2O 10.85 4.65 38, 90% RH
MDPE: medium density PE; RH: relative humidity
Figure 4.10 Pressure increase due to gas permeation at the panel PE flanges.
In the case of panels which are encapsulated, i.e., surrounded by a closed cell PU foam,
as happens for panels in household refrigerators, freezers or vending machines (see Section
4.6), the picture is more complicated. The gas environment surrounding the vacuum
panel will depend on the type of encapsulating foam and will change with time as a
consequence of the gas out-diffusion from the closed cells and the progressive air
penetration from the outside, thus making predictions more difficult. This ageing process
177
Advances in Urethane Science and Technology
will also depend on the temperature and the hermeticity of the case surrounding the
encapsulating foam as shown in various papers [32, 33]. Permeation of blowing agents
through the seals, even though not a major contribution, has also to be considered.
The use of a different PE grade or a different sealant polymer will provide different
pressure build-up curves in the panel both in terms of total and partial gas pressures.
In general, very little data have been published on the outgassing properties of the skins
for vacuum panels, even though they can contribute, in some cases, in a non-negligible
way to the deterioration of the pressure inside the VIP. This can be due to the outgassing
properties of the materials used as barrier layers and/or the lamination process, which
may introduce volatile substances or trap gases in between the various sheets.
The outgassing properties of a barrier sample can be measured using the same experimental
equipment shown in Figure 4.2 and have to be taken into account to estimate the total
gas load in the panel. For the same film having the barrier property in Table 4.3, the
extrapolation of the outgassing experimental data to 20 years in a 50 x 50 x 2 cm3 size
panel is shown in Table 4.7. Water, which is also given off by the barrier, was not quantified
in this test.
Table 4.7 Outgassing parameters and gas load for a barrier incorporating
a 6 μm aluminium foil
Gas q0 (Pa-l/h) α Q (Pa-l)
H2 35 x 10-4 0.5 29
-1
N2 43 x 10 1.3 133
178
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
As discussed in the previous sections, gas desorption from the surfaces exposed to the
vacuum (filler and barrier) and gas permeation through the bag contribute to increase
the pressure in a VIP during its life.
These contributions, as well as the residual gases left in the panel after the exhaust and
seal-off process, have to be taken into account to provide an accurate estimation of the
pressure increase in a VIP as a function of time. Figure 4.11 shows the total pressure
increase, not including water, for some panel sizes, when data from Tables 4.2, 4.3, 4.6
and 4.7 are used.
Figure 4.11 Estimated total pressure increase in some open cell foam-filled VIP as a
function of time. The additional pressure increase due to water vapour is not
considered in this calculation.
179
Advances in Urethane Science and Technology
For given temperature and environmental conditions, the pressure build-up depends on
the specific size of the panel, and in particular on the ratio between the perimeter and the
thickness, the lower the latter the higher the pressure increase. A barrier film incorporating
an aluminium foil and a panel seal-off pressure of 5 Pa have been here considered. In this
example, water has been assumed to be completely absorbed by a proper amount of
desiccant and not to contribute to the pressure increase.
In the case of a typical 50 x 50 x 2 cm3 size VIP, the pressure build-up exceeds 100 Pa
after a few years and even not considering water, the total pressure after 20 years exceeds
250 Pa.
The deterioration of the thermal conductivity for such a panel, a direct consequence of
the increase of the internal pressure, is shown in Figure 4.12, for a PU foam following
the curve of Figure 4.1.
Due to the nature of the thermal conductivity versus pressure curve, the deterioration of
the λ factor is negligible at the beginning but increases steadily with time.
Figure 4.12 Increase of the thermal conductivity in a vacuum following the pressure
build-up as per Figure 4.11. Water is supposed to be absorbed by a desiccant and not
contribute to the pressure rise.
180
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
These examples show that for medium and long term applications (> 2-3 years) a simple
desiccant is not enough and specific getter materials have to be added to sorb the extra
amount of gases generated in the panel.
The wide spectrum of gases present in a panel, including carbon dioxide, carbon monoxide,
hydrogen, nitrogen, oxygen and water, as well as traces of blowing agents, also calls for
an absorption system having high sorption capacities and adequate pumping speed for
all these gases.
Zirconium-based NEG alloys, which are widely used in a variety of vacuum applications
[34, 35, 36, 37] cannot be used in plastic evacuated panels due to their limited sorption
capacity at room temperature and the need to be heat activated at relatively high
temperature (> 350 °C) prior to their use, this process being clearly not compatible with
the panel polymeric components.
Very large area physical adsorbents, like molecular sieves, zeolite or activated charcoal
[38, 39] have very good efficiency for water and some organics but present serious limitations
in sorbing carbon monoxide, hydrogen, nitrogen and oxygen at the temperature and pressure
conditions typically encountered in VIP applications. Therefore, they have beneficial effects
during the very initial life of the VIP without being able to ensure long lifetime, as required,
for example, by the appliance industry. They are also sensitive to the sorption temperature,
i.e., the higher the temperature the lower the sorption performance, which spoils sorption
performances in high temperature applications, e.g., 60-100 °C.
Physical adsorbents have also a second drawback which is related to their treatment,
when mass production quantities have to be handled. To fully take advantage of their
sorption capacity, in fact, a relatively high temperature pre-treatment process which cleans
the surface by promoting gas desorption is required. However, after this treatment, the
adsorbents are exposed again to the ambient environmental conditions during the
subsequent handling and mounting operations. Due to their reactivity, they will rapidly
sorb some gases from the environment and this will reduce their capacity in a poorly
reproducible way, critically dependent on the environmental conditions, temperature
and exposure time. Fluctuations in their sorption performances as well as in the VIP
insulating efficiency could be the final result.
To eliminate the above problems and address the lifetime issues posed by the appliance
industry, a novel getter device, the COMBOGETTER, has been recently proposed [40].
181
Advances in Urethane Science and Technology
At the heart of this device is a barium-lithium alloy, in a 1 to 4 atomic ratio [41], able to
efficiently chemically absorb a large amount of nitrogen at room temperature, up to
more than 2500 Pa-l (N2)/g (alloy) [42].
This very specific feature allows the getter to compensate for the air inlet coming from
permeation, thus ensuring long lifetime requirements.
High efficiency calcium oxide and a metal oxide are also added to barium-lithium to
absorb moisture, hydrogen and some of the most common blowing agents, such as R141
b and cyclopentane which could permeate through the VIP during its life. All these
materials are prepared according to proprietary processes, which confer upon them unique
properties in terms of sorption efficiency and capacity. The active powders are compressed
in a stainless steel cup according to a configuration which allows optimum sorption
performances and ease of use (Figure 4.13).
The getter device does not need to be heat activated or pre-treated before being used in
VIP and can be handled in air for a reasonable period of time (several minutes), during
the panel manufacturing process, without affecting its sorption capacity.
Mounting can be accomplished by inserting the getter in a recess cut in the filling material.
Alternatively, if the filler is sufficiently soft, as is the case for open cell PU, the getter can
be simply pressed into it.
182
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
Gas capacity data for the getter are summarised in Table 4.8.
A typical sorption curve for nitrogen at room temperature is given in Figure 4.14 (sorption
throughput, Pa-l/s, as a function of the sorbed amount, Pa-l).
Figure 4.14 Sorption curve for nitrogen for the COMBOGETTER (at room temperature)
183
Advances in Urethane Science and Technology
This throughput is adequate to compensate for the relatively slow gas inlet rate occurring
during the VIP lifetime. The effect of the temperature on the sorption performance is
shown in Figure 4.15 in the range 20-70 °C.
These data show that the getter is effective also at relatively high temperatures, where
physical adsorbents become less efficient.
The getter in VIP technology has several roles, as demonstrated by the following examples:
a) Life insurance
Pressure in VIP has been recorded by means of a viscous pressure gauge, as explained
in Section 4.4.2.2. The resulting thermal conductivity of gettered panels is therefore
much lower in the long term, as shown in Figure 4.17, which compares the λ factor
in gettered and ungettered panels after almost 3 years of ageing.
184
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
Figure 4.17 Measurement of the thermal conductivity in two panels (with and
without COMBOGETTER) aged at room temperature
185
Advances in Urethane Science and Technology
The ability of the getter to sorb air can be also used to reduce the evacuation time in
panels, this contribution being particularly interesting for medium and large size
panels, e.g., > 0.5 m2, where achieving base pressure in the low 1 Pa can be time
consuming for open cell foams with very small cell openings, which limit the evacuation
efficiency. For each application, a trade-off between the getter capacity needed to
reduce the evacuation time and that necessary to ensure long lasting service
performance has to be found.
A typical example of the getter’s ability to act as a chemical ‘in situ’ pump to sorb
residual gases after the evacuation process is shown in Figure 4.18, which compares
the pressure trend in panels with and without the getter immediately after the
evacuation and seal-off.
Figure 4.18 Pressure decrease in a VIP due to the in situ pumping effect
of the COMBOGETTER.
186
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
It can also compensate for the gases generated in the panel during ageing or heat
treatment, as happens during storage in the warehouse, the foaming process in a
refrigerator cabinet or the high temperature operating conditions required by
several specific applications. Experimental data on vacuum panel performances
under these conditions are presented and discussed in Section 4.6.
For some specific VIP sizes and short term applications (for example one to two
years) a getter for air may not be necessary, since the permeation contribution is
less relevant here. In these cases water is usually the most important gas to be
sorbed and the use of a simple dryer is sufficient to keep the pressure at the
required 10 Pa value. Calcium oxide is one of the preferred desiccants due to its
availability, low cost, good environmental features and large water sorption
capacity. The production process influences the calcium oxide physical structure
(particle size, porosity and morphology) and thus the water sorption efficiency
may be actually an important parameter to consider in short-term applications
(for example in shipping containers, where lifetimes as short as a few months or
less, can be of interest). A comparison between a commercially available product
and a highly efficient calcium oxide, prepared by SAES according to a proprietary
process is given in Figure 4.19.
Also physical absorbents, like molecular sieves and silica gel or activated carbon,
can be used for short-term applications.
187
Advances in Urethane Science and Technology
Figure 4.19 Water sorption test on two calcium oxide samples. Samples were exposed
at 100-200 Pa water partial pressure at 23 °C. Sample A was prepared by SAES
according to a proprietary process to increase material surface area and porosity.
Before being inserted in the barrier bag, the open cell foam needs preliminary pre-
treatment in air to remove water and the residue of the foam production process.
Volatile materials are in fact used to blow the foam chemicals and remain partially
trapped in the foam matrix or condense as a liquid on the foam surface after the cells
have been opened.
Even though part of these volatile substances will desorb at room temperature, the
remaining quantity will generate an unacceptable pressure build-up in the panel [43].
The baking process is also necessary to remove the water absorbed by the PU foam
when exposed to humid air (about 1.8 % by weight, at 23 °C, 50% relative humidity),
this being particularly effective to reduce the evacuation time, as shown by tests of
actual panels [44]. Pre-treatments generally range from 150 °C for 10 minutes to 120
°C for up to a few hours [9, 12, 43], depending on the foam type.
188
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
After the foam has been thoroughly baked, it is important to reduce its exposure time to
the working environment to avoid significant moisture re-absorption, which takes place
when PU foam cools down. As an example, the weight increase due to water re-absorption
during cooling to room temperature is shown in Figure 4.20 for a small PU sample, dried
at 150 °C for 20 minutes and then exposed to 50% relative humidity.
These data indicate the need to design the production process and equipment in such a
way that the foam is exposed to air for a very short time, not exceeding a few minutes. A
limited exposure is also beneficial to reduce the evacuation time of the panel, which
strongly depends on the desorption rate of physisorbed water. To further minimise re-
absorption of water on the foam it is advisable to process the PU in a dry or humidity-
controlled area. The water content picked-up by the foam can be measured in the finished
panel by means of residual gas analysis (RGA) with the mass spectrometer, as described
in Section 4.4.2.2. In general, the proper pre-treatment of the foam followed by adequate
handling ensures better initial insulation performance and shorter evacuation cycles which
also mean higher productivity and lower costs per panel.
The open cell foam slabs are generally easier to handle and process than powder-based
core materials like precipitated silica or perlite. However, precautions have to be adopted
189
Advances in Urethane Science and Technology
during handling of the foam to avoid powder generation which could have an impact on
the quality and reliability of the seals. Care also has to be taken during the production
process to ensure adequate VIP flatness, which is an important factor for the subsequent
proper mounting of the panel in various applications.
Appropriate film handling and processing is also the key to produce reliable and good
quality panels.
Gas barrier properties of films, in fact, mainly depend on the number of defects, such as
cracks and pinholes, present in the layered structure. This number increases as a
consequence of the mechanical stresses the film undergoes during its handling. Moreover,
the high load applied by the atmospheric pressure on the vacuum panel, this value being
particularly high in the corners and along the edges, stretches the layers and locally
increases the defect density.
A preliminary investigation of these effects has been carried out by Sugiyama and co-
authors [21] for some laminated plastic films incorporating a 6 μm aluminium foil.
Films have been analysed as such and after having been twisted to simulate handling
conditions. VIP have also been prepared and samples cut from the flat surface of the
panel, the edges and the corners to check for any local increase of the permeation rate.
Tests were carried out at 23 °C on samples of 2.5 x 10-3 m2 area applying 0.0505 MPa
helium pressure. The structure of the analysed films was Nylon 15 μm/vacuum metallised
PET 12 μm/aluminium 6 μm/HDPE 50 μm. The results, shown in Table 4.9 clearly indicate
that the twisting process is responsible for significantly larger permeation rates. Samples
cut at the corners also show higher permeation rates, which confirms that stress conditions
in these areas increase the local pin-hole density in the aluminium foil.
Table 4.9 Helium permeation rates through barrier films submitted to different
mechanical treatments
Sample number Type of sample He permeation rate
(Pa-l/s) x 10-11
1 Reference film (no treatment) <7
2 Sample cut from the flat surface of 7
the vacuum panel
3 Sample cut from the corner of the 450
vacuum panel
4 Sample mechanically twisted 230
190
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
Even though the total permeation rate is still acceptable, data suggests that suitable
precautions have to be adopted to limit as much as possible stress and film deterioration.
Similar behaviour is also expected for metallised films and has to be addressed accordingly.
The rapid development of the open cell foam-filled vacuum panel technology has required
parallel development and improvements of the analytical techniques necessary to assess
VIP performance and reliability. This latter aspect is key for the widespread adoption of
the technology. The selection of the best components, foam, bag and adsorbent, as well
as the careful control of the manufacturing cycle, minimises the chance of having poorly
performing insulating panels. However, since the potential risk of defective seals or micro-
leaks cannot be completely ruled out, several techniques have been developed to either
support and establish the VIP manufacturing cycle or to assess their quality after
production.
In the following sections, some techniques to check the insulating performance of vacuum
panels are illustrated, a few of them also being able to be used as a tool for Quality
Assurance (QA) and Quality Control (QC) for the manufacturing process.
The direct measurement of the thermal conductivity or λ factor is one of the most common
ways to check the quality of a vacuum panel. Several test methods are available, such as
the heat flow meter [45], guarded hot-plate and guarded-calibrated hot box [46, 47]
procedures, which measure the heat transferred under steady state conditions through
the sample whose surfaces are kept at two different given temperatures.
The last two techniques are absolute methods since they do not require the use of reference
or calibration specimen. However they are relatively expensive and time-consuming and
difficult to adapt to rapid testing.
In the heat flow meter technique, the sample is placed between two plates controlled at
different temperatures. Thermocouples placed just below the plate surfaces measure the
temperature drop across the sample. One (or more) heat flow transducers mounted on
the plate measures the voltage which is proportional to the heat flow through the sample.
Knowing the relationship between the voltage signal of the transducer and the heat flow
through it, which can be obtained by calibrating the transducer to a known thermal
conductivity standard, the thermal conductivity for the specimen can be obtained.
191
Advances in Urethane Science and Technology
Heat flow meter test results generally agree well with those generated by absolute methods.
The technique, which is easier to use and is a universally recognised standard test procedure
[45], is also usable either for high accuracy measurements to support VIP analytical
modelling/development or for fast tests for QA/QC programs.
This latter requirement is becoming more and more important as VIP technology moves
closer to the market. Recent improvements have been based on the use of a novel dual
heat flow transducer design and a tight control of plate temperatures [48].
They allow a pass/fail test on a vacuum panel to be completed within a few minutes, thus
providing the possibility for an extensive quality check during the manufacturing cycle.
This is particularly important not only for VIP manufacturers but also for the final users,
to verify continued panel integrity after handling and transportation and before mounting.
Thanks to the short measuring time (~5 minutes are required between two measurements),
this technique also seems promising as a pass/fail test for QA/QC applications.
Since the thermal performance of the panel is dependent on its internal pressure, an
alternative approach is to measure the pressure evolution as a function of time.
The spinning rotor gauge (SRG) technique [51], which is based on the measurement of
the deceleration rate caused by gas friction on a freely spinning spherical metallic ball, is
particularly well suited to monitor the pressure in sealed-off devices and has been proposed
in the recent years also for vacuum panel applications [5, 40, 52, 53].
The ball, or rotor, is suspended by magnetic forces between two mutually attracting
permanent magnets and additional coils provide the rotating magnetic field to spin it.
192
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
After the rotor has been accelerated and has achieved constant speed, the powering coils
are switched off, allowing the rotor to slow down.
A correlation between the deceleration rate and the pressure can be drawn, provided the
density and the radius of the rotor and the mean velocity of the gas molecules impinging
the rotor surface is known. Positive features of this technique are the wide pressure
range covered, from 10-4 to 105 Pa, the low-cost sensor and the possibility of operating
several sensors with a single electronic unit.
The sensor can be connected to the VIP using a high vacuum compatible resin, as shown
in Figure 4.21.
In the advanced SRG models, such as the SAES Getters SpiroTorr unit [54], the
instrument can operate basically at any spatial orientation on both the sensor and the
gauge head.
Figure 4.21 Diagram of the VIP mounting the spinning rotor gauge. Vacuum tightness
is ensured by a high vacuum epoxy resin.
193
Advances in Urethane Science and Technology
Pressure readings are not very much dependent on the gas type for pressure in the range or
below 10 Pa, so that the precise knowledge of the gas composition is not strictly necessary
to get a fairly accurate pressure measurement. For higher pressures, the ball (which is
suspended by the magnetic field) deceleration rate becomes more dependent on the gas
species and a precise reading requires the knowledge of the main gases involved [55].
Typical results obtained using the SRG are shown in Figure 4.16 and Figure 4.18.
Figure 4.22 Picture of the SAES Getters SpiroTorr SRG connected to a vacuum panel.
194
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
Due to the need to mount the sensor, this technique cannot be used as a quality control
method. Its role is more in the area of basic investigation and technology assessment.
Another way to measure the pressure is by puncturing the panel and analysing the gas
atmosphere with a mass spectrometer. The sample is connected to the analytical bench
of Figure 4.3 by a glass tube which is sealed with a high vacuum epoxy resin, as in
Figure 4.23.
The tubing is then evacuated, the panel punctured, using for example a metallic hammer,
and the internal atmosphere expanded to the mass spectrometer for the residual gas
analysis.
This approach allows measurement of both total and partial pressures in the panel, thus
providing useful information on the gas ratios, this being useful as an R&D tool or to
monitor and improve the manufacturing process. However, since this is a destructive
and relatively time-consuming test it can only be used for one-off samples. A non-
destructive technique to measure total pressure in a vacuum panel has been proposed
recently, which is based on the use of a laser beam source coupled to a detection system.
Figure 4.23 Pictorial view of the method used to connect the analytical bench to the
vacuum panel for RGA analysis.
195
Advances in Urethane Science and Technology
The panel is placed in a vacuum chamber and the laser beam impinging on the panel
surface is reflected by the laminate and recorded by the detector. The chamber is then
slowly pumped down. When the pressure in the chamber is lower than the internal pressure
in the panel, the volume of the bag increases by gas expansion, thus causing the laminate
surface to move and deflect the laser beam in a different direction. This will generate an
electrical feedback signal which is recorded by the equipment as the internal pressure of
the vacuum panel. Present sensitivity of such types of equipment, although not well
established, is around some hundreds of Pa. It is likely that this value can be lowered by
calibrating the laser sensor against an absolute pressure gauge. This measuring technique
is quite fast (about one panel per minute) and can therefore be used as a QC/QA tool in
a manufacturing line.
The actual insulation efficiency of PU-based vacuum panels depends on the intrinsic core
material conductivity, the vacuum level, the type of barrier envelope, the panel size and
arrangement in the insulating structure.
All these aspects have to be optimised to get full benefits from VIP technology. As
previously discussed, laminates incorporating a continuous aluminium foil provide, at
present, the best gas barrier properties for VIP. Their main drawback is the loss of
insulation properties at the panel flanges due to the relatively high thermal conductivity
of the aluminium foil which reduces the overall insulating performance, this effect being
directly related to the panel size and the aluminium foil thickness.
The influence of the edge effect on the insulating performance of the panel in actual
applications can be better appreciated by introducing the ECR (Energy Consumption
Reduction) factor:
where Kpu and Kvip (W/m) are the thermal conductivity of the conventional closed cell
foam and of the vacuum panel, respectively.
The ECR factor provides the percentage increase in the insulation efficiency obtained by
completely replacing the conventional foam with a vacuum panel having given thermal
conductivity Kvip. Calculation of Kvip is not easy, since it is a complicated function of the
size of the panel, its geometry, the thermal conductivity of the core material and the
thermal conductivity of the laminate (which is basically the aluminium foil, due to its
high thermal conductivity). The heat flow through the laminate and through the open
cell foam, both necessary to get Kvip, can be calculated by FEA.
196
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
Figure 4.24 ECR values as a function of the aluminium thickness for VIP of
different sizes.
Once Kvip is known, the ECR factor is obtained by Equation (8). The result of the ECR
calculation is shown in Figure 4.24 as a function of the aluminium thickness for some
different panel sizes.
It is clear from Figure 4.24 that high ECR values can only be obtained by reducing the
aluminium thickness as much as possible and using relatively large vacuum panels, i.e., >
0.5 m per side.
The ECR values for encapsulated panels of different sizes prepared with a laminate
incorporating a 6 μm aluminium foil (Table 4.3, Film B) are given in Table 4.10. The
calculation has been made assuming a foam thermal conductivity of 6 mW/m K and a
thickness of 0.5 cm of additional PU foam (thermal conductivity of 20 mW/m K) for a
total thickness of 0.025 m.
197
Advances in Urethane Science and Technology
The ECR values achievable with the aluminium foil-based laminates are very interesting
for large size panels, where the edge effect is not severe. For small panels, which are also
important, since they can be easily adopted in a variety of applications, from household
appliances to shipping containers, the improvement in the insulation efficiency is not as
remarkable.
The energy consumption reduction in two refrigerator cabinets, each using one panel,
100 x 50 x 2 cm3 large, covering 40% and 60% of the total surface is shown in Table
4.11. The same assumptions used in Table 4.10 have been made for the PU thermal
conductivity and thickness of the barrier foil.
Table 4.11 ECR values for two refrigerators with varying VIP coverage
VIP size (cm) Refrigerator/freezer ECR
surface coverage
100 x 50 x 2 60% 22%
100 x 50 x 2 40% 14.7%
In general, for household appliances, depending on the surface coverage of the cabinet
and the panel thickness, energy savings from 10% to 30% have been reported using
open cell PU foam-filled panels packaged in a 6 μm aluminium foil-based barrier [12,
15, 52, 56].
198
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
The influence of the vacuum panel thickness to the total insulation thickness ratio as
well as the change in performance, by increasing the foam λ value, has been determined
by Hamilton [54]. In general, results can be optimised by adjusting the VIP size and
thickness to provide the most cost effective option.
Refrigerators and freezers account for about 20% of the total electricity consumption of
household appliances. For this reason the appliance industry is under pressure to improve
the energy efficiency of their products to cope with the need to reduce carbon dioxide
emissions, as recently mandated by the Kyoto Conference. This objective has to be achieved
without penalising product performance. Several options to decrease the energy
consumption are under evaluation, ranging from high efficiency compressors to the
adoption of intelligent electronic devices [57, 58, 59].
Improvement of the insulation through the use of vacuum insulated panels filled with
silica powder or glass wool as core materials has been evaluated, and adopted in limited
amounts, by most refrigerator manufacturers [60, 61, 62], cost being the main obstacle
to large sales of these products. The lower cost of the open cell foams, combined with the
low density and good handling properties, has generated renewed interest for this
technology as an environmentally friendly option to energy consumption reduction. A
second driving force for the appliance industry is the possibility of increasing the internal
volume of refrigerators and freezers without increasing the outer dimensions. This aspect
is particularly important in Europe and Japan, where the built-in appliance market is an
important segment and space constraints play a role.
Adoption of PU-based vacuum panels in refrigerators and freezers requires the proper
handling of the foaming process, which may influence the structural and vacuum properties
of the panel.
199
Advances in Urethane Science and Technology
In fact, during the appliance manufacturing cycle, the panels are glued to the cabinet
walls and then foamed in place with a conventional closed cell PU foam, this process
being necessary to further improve the overall structural and thermal insulating properties
of the cabinet. To allow a better flowability of the injected chemicals, the cabinets are
generally pre-heated at about 45 °C for some minutes. The presence of the panels, which
reduces the free flow and expansion of the foam, requires special care to avoid generation
of air-filled void volumes which would deteriorate the insulation performance [56]. To
prevent this, and provide the best integration between VIP and conventional PU foams,
specific systems have been recently developed to be used in conjunction with VIP [10].
As a combined effect of the cabinet pre-heating and the exothermicity of the foaming
reaction, temperatures even higher than 100-120 °C can be reached on the VIP surface,
leading to temperature-enhanced gas desorption. A typical result using a 50 x 50 x 2 cm3
panel is shown in Figure 4.25.
Figure 4.25 Measurement of the temperature increase experienced by the two panel
surfaces during the foaming process.
200
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
This gas contribution has to be evaluated case by case since it depends strictly on the
actual foaming conditions, such as the chemicals used, the process variables, the
refrigerator design and VIP geometry. However, under some circumstances, the pressure
increase can be a measurable fraction of the maximum acceptable level, thus causing a
deterioration of the VIP thermal insulation properties from the very beginning of the
refrigerator life.
As an example, Figure 4.26 shows the effect of a thermal treatment carried out at 50, 70
and 90 °C on a VIP prepared with an open cell foam and an aluminium foil-containing
laminate. The foam was pre-baked at 150 °C for 20 minutes and no absorbent was used.
After the preparation, the panel was put in a oven and kept at the indicated temperature
for 15 to 20 minutes. The panel was then removed and the total pressure measured as a
function of time. To allow the continuous monitoring of the pressure, a SpiroTorr spinning
rotor gauge was mounted in the panel as shown in Figure 4.21.
Desorbed gases after the thermal treatments are only partially reabsorbed by the foam
and generate a measurable pressure build up in the VIP.
201
Advances in Urethane Science and Technology
The effect of the temperature increase resulting from the foaming process, on the vacuum
level in a panel can also be seen in Table 4.12 which shows the result of an RGA carried
out on two test panels encapsulated by an appliance manufacturer. In the getterless panel,
the pressure immediately after the foaming is already exceeding the 10 Pa target value
and close to the maximum acceptable value (generally set at 50 Pa).
The deterioration of the vacuum during the foaming process has been recently addressed
by Kücükpinar and co-workers who ran specific tests aimed at quantifying this effect in
open cell foam-filled VIP with and without getters [53].
As shown in this study, the COMBOGETTER was able to compensate for the extra gas
load generated during the foaming process, thus confirming the results in Table 4.12.
Further specific outgassing studies are necessary to understand the mechanisms of gas
generation during the foaming process, even though this investigation presents several
difficulties because of the large number of parameters involved.
The foaming process may also stress the laminate due to the combined effect of temperature
and applied pressure on critical areas such as the panel edges and the corners. Defects
generated here may in turn increase the gas diffusion inside the panel, reducing the designed
lifetime.
Durability tests on panels are therefore necessary to assess VIP reliability and performance
over time. Figure 4.27 shows the pressure values in some encapsulated panels, as measured
by RGA. Panel A, containing a COMBOGETTER, was aged at room temperature for
some weeks and then kept at 40 °C for more than 30 months. Panels B and C did not
have any adsorbent. The latter was aged at 40 °C for about 3 months, while the former
was unaged.
202
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
Figure 4.27 Residual gas analysis in panels encapsulated and aged for different times.
Panel A contains one COMBOGETTER, panels B and C are without.
The pressure in panel A is well below the 10 Pa target value as compared to panel C
whose pressure is one order of magnitude higher. Pressure in panel B, which was just
encapsulated, was slightly lower than in panel C. For comparison, the calculated pressure
increase in the panel, as predicted based on the model discussed in Section 4.3, is also
shown. This result confirms the role of the getter to compensate for the encapsulation
process and to ensure VIP longevity and also supports the gas load model previously
described.
Still, a better assessment is necessary through longer term tests and, possibly, through
accelerated tests able to provide experimental data within a shorter time. Activity is
ongoing in SAES Getters Laboratories, as well as in various other research centres to
provide such an assessment.
This is very special equipment designed to operate at very low temperatures, e.g., from
-30 °C to -86 °C, to age samples or to store valuable and perishable goods, like organs
and tissues, biological and medical samples or vaccines. Vacuum panels are used mainly
to increase the internal storage volume without increasing the energy consumption.
203
Advances in Urethane Science and Technology
Since the conventional insulation must be very thick to ensure the achievement of such
low temperatures, a partial replacement of the conventional insulator with VIP
contributes significantly to increase the internal volume, from 20% up to 40% or
more in specific models.
Due to the generally higher insulation thickness, the foaming process here may be even
more exothermic than in domestic appliances. Means have therefore to be taken to keep
the VIP temperature as low as possible during the encapsulation.
Ultra-low temperature freezer models using foam filled vacuum panels were successfully
placed on the market by a leading company some years ago and other companies are
expected to follow soon.
Vending machines are especially popular in Japan, Korea and the Far East. It is estimated
that in Japan about 2.5 million of these appliances are in the field with a replacement
market of about 0.4 million pieces/year.
VIP are finding widespread use in vending machines since most manufacturers are
extensively using one or two panels to separate the hot and cold beverage compartments.
Also in this case, the main driving force for VIP adoption has been the possibility to
increase the internal volume for the storage of beverages, rather than the improvement
in energy efficiency.
Vending machines are quite a demanding application for VIP, since the panel operating
temperature is cycled between room temperature and 60-70 °C and the lifetime is 5
years. As already discussed, the high temperature promotes higher diffusion and outgassing
rates and provides additional mechanical stress to the envelope.
Ageing tests have been carried out by vending machines manufacturers to assess VIP
usability and reliability in this application.
Results obtained in SAES Getters Laboratories for 50 x 50 x 2 cm3 panels aged for
several months at 60 °C are shown in Figure 4.28.
204
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
Figure 4.28 Measurement of the thermal conductivity as a function of time for panels
aged at 23 and 60 °C, with and without the COMBOGETTER.
Again, the deterioration of the thermal conductivity, due to the pressure increase, is quite
dramatic in the getterless panels both at 23 and 60 °C. As expected, the thermal conductivity
increase is much more rapid for the panels aged at 60 °C than for the panel kept at room
temperature. Only minimal differences are measured in the panels containing the
COMBOGETTER, which could compensate for the increased gas load at both temperatures.
Large vacuum insulated containers, ships and trucks are under evaluation by various
companies, mainly in Europe and Japan.
The main advantages offered by using VIP are space saving and better insulation, so as
to keep the temperature increase rate as low as possible even in the absence of a power
source. The lifespan is generally from 10 to 15 years.
Since very large VIP, e.g., 2-3 m2 size are considered for this application, the edge effect
(see Section 4.3.2.1) is here a marginal issue.
205
Advances in Urethane Science and Technology
Small commercial shipping containers to store and deliver pharmaceuticals, frozen food
and valuable temperature-sensitive products is also another interesting area for the use
of VIP.
Long lifespan is not a problem, since most of these shipping containers have a projected
life of the order of weeks or months. Therefore, it is even possible to use the VIP more
than once, the specifications on the permeation properties of the barrier film are less
tight and the getter not necessary. A specific dryer, to adsorb water, may be used in most
cases. The small size VIP and the limited projected life suggest the possibility of using
metallised barriers for the skin of the panel in these applications, at least on one side, so
as to overcome the edge effect, which, in this case is very critical.
A variety of other potential applications exist for VIP, from cold stores to insulation in
buildings and in industrial plants, e.g., industrial reactors or liquid natural gas tanks.
Also super insulation at relatively high temperature applications for water heaters, heat
pipes or boiling pots have been considered and are under evaluation. The key issue of the
adoption of VIP in this last family of applications is the possibility to bend the open cell
PU filled VIP into a round or cylindrical shape. The combination of high temperature
and bending provides additional challenges to VIP technology and further improvements
in the material selection and processing might be required.
The open cell PU foam is a very promising filler material for VIP. It shows good vacuum
compatibility, thermal conductivity, light weight, ease of handling and processing and
moderate cost. To allow its use, significant improvements have been made recently in the
barrier and getter technology, as well as in the panel manufacturing and testing procedures.
206
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
The outgassing and permeation data, on both components and finished panels, and the
use of models quoted in the literature, allow the evaluation of the total gas load and the
pressure increase in a vacuum panel as a function of time, size and operating conditions.
This in turn is beneficial for designing the most suitable solution for the absorbent.
A high capacity getter system, the COMBOGETTER, capable of adsorbing air, moisture
and the other gases of interest without the need for any pre-treatment has been specifically
developed for VIP application. Its roles are many since it chemically absorbs residual,
outgassed and permeating species and provides the means to shorten the evacuation
process and increase the manufacturing yield.
Mass production technologies to manufacture high quality open cell foam panels are on
the market. Several techniques are also becoming available for the fast measurement of
the thermal conductivity. This is a very important issue to build a QA/QC system able to
ensure the quality and the reliability of the products, not only after production but also
immediately before use.
As far as the applications are concerned, open cell PU foams are finding their place in the
market for both low and high temperature appliances, from refrigerators to vending
machines. Present production volumes are still limited but near term perspectives are
encouraging, especially in Asia and Europe. The strong push for energy reduction and
the parallel continuous refinement of the technology should provide further motivation
to VIP adoption.
Cost is still the main obstacle to the widespread adoption of this technology. A preliminary
qualitative cost comparison between silica and open cell PU, which includes raw material,
panel manufacture and panel installation costs [10] shows that a cost reduction of about
one-third over silica could be achieved. This has to be improved to bring the cost of the
vacuum panel down further, so as to really interest more segments of the insulation
industry.
Very recent breakthroughs in the foam mass production technologies seem to indicate
that this target is reachable.
Even though far from being complete, the material presented in this chapter shows that
significant technological achievements have been obtained during the last few years.
207
Advances in Urethane Science and Technology
Thanks to this, vacuum panel technology is becoming a technically viable and cost effective
solution to the need to reduce energy consumption in household appliances and in
commercial and industrial applications. To successfully achieve this target, additional
efforts are necessary to further improve the component quality and reliability (foam,
film and adsorbent), to optimise panel production and to reduce costs.
Acknowledgements
The author would like to thank Dr. Paolo della Porta, President and CEO of SAES Getters
SpA, for his continuous support and long lasting commitment to VIP technology. The
author also acknowledges Dr. Bruno Ferrario, Corporate R&D Director, for fruitful
discussions and suggestions.
Special thanks are given to Dr. Roberto Caloi and Dr. Enea Rizzi for their valuable
technical support in the development of theoretical models and in running experimental
measurements on VIP components and finished devices.
References
208
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
11. K. W. Dietrich and D. W. McCullogh, Presented at UTECH 96, The Hague, The
Netherlands, 1996, Paper No.64.
12. W. H. Tao, W. F. Sung and J. Y. Lin, Journal of Cellular Plastics, 1997, 33, 3,
545.
19. P. Mercea, L. Muresan and V. Mecea, Journal of Membrane Science, 1985, 24,
297.
20. T. A. Beu and P. V. Mercea, Materials Chemistry and Physics, 1990, 26, 309.
21. A. Sugiyama, H. Tada and M. Yoshimoto, Vuoto, 1999, 28, 1-2, 51.
22. ASTM Standard E96-00 Standard Test Methods for Water Vapor Transmission of
Materials.
23. ASTM Standard F 1249-90 (1995) Standard Test Method for Water Vapour
Transmission Rate Through Plastic Film and Sheeting Using a Modulated
Infrared Sensor.
209
Advances in Urethane Science and Technology
26. P. Manini and E. Rizzi, Poster presented at the European Workshop on Vacuum
Panel Technology for Super Insulation in Domestic Appliances and Industrial
Applications, Milano, Italy, 1998.
27. W. Lamb and R. Zeiler, Presented at the Vacuum Insulation Panel Symposium,
Baltimore, MD, 1999. Paper available from the SAES Getters web site, http://
www.saesgetters.com.
30. Polymer Handbook, 3rd Edition, Ed., J. Brandrup and E. H. Immergut, John
Wiley and Sons, New York, 1989.
35. C. Boffito, B. Ferrario, P. della Porta and L. Rosai, Journal of Vacuum Science
and Technology, 1981, 18, 1117.
37. C. Boffito, B. Ferrario, L. Rosai and F. Doni, Journal of Vacuum Science and
Technology, 1987, A5, 3442.
210
Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…
41. C. Boffito and A. Schiabel, inventors; Getters SpA, assignee; European Patent
0514348 A1, 1991.
45. ASTM Standard C518-98, Standard Test Method for Steady-State Thermal
Transmission Properties by Means of the Heat Flow Meter Apparatus.
46. ASTM Standard C236-89 (1993) e1, Standard Test Method for Steady-State
Thermal Performance of Building Assemblies by Means of a Guarded Hot Box.
47. ASTM Standard C177-97, Standard Test Method for Steady-State Heat Flux
Measurements and Thermal Transmission Properties by Means of the Guarded-
Hot-Plate Apparatus.
50. C. Dixon and N. Mathis, Presented at the Polyurethanes Expo ‘99, 1999, p.607.
51. J. K. Fremerey, Journal of Vacuum Science and Technology, 1985, A3, 3, 1715.
52. E. Kücükpinar, H. Güclü, A. S. Akkurt and F. Özkadi, Vuoto, 1999, 28, 1-2, 31.
211
Advances in Urethane Science and Technology
212
5
Modelling the Stabilising Behaviour of Silicone
Surfactants During the Processing of
Polyurethane Foam: The Use of Thin Liquid Films
Steven A. Snow, Udo C. Pernisz, Benjamin M. Nugent, Robert E. Stevens,
Richard J. Braun and Shailesh Naire
5.1 Introduction
Foam degradation involves both physical and energetic changes in the foam. Four physical
processes are commonly associated with PU foam degradation:
213
Advances in Urethane Science and Technology
These physical processes involve the release of excess gravitational and surface energy.
Excess gravitational energy is released during the drainage of liquid from the foam. The
liquid can drain from between the bubbles, thinning the liquid membrane known as the
lamella. The liquid can also drain from the plateau borders in the foam. These borders
are the relatively wide channels in the foam created at the intersection point of three
lamellae. Mysels, Frankel and Shinoda [22] estimated that, in aqueous foams, the amount
of gravitational energy released during degradation was five orders of magnitude less
than the amount of surface energy released. It is assumed that this estimation roughly
holds for the degradation of PU foams; therefore, the focus here will be on processes that
reduce the surface energy of PU foams.
Two processes, the diffusion of gas from small to larger bubbles (Ostwald Ripening) [23]
and bubble coalescence, reduce the surface area of the foam and, therefore, the surface
energy. Ostwald ripening involves the diffusive molecular transport of gas from small to
large(r) bubbles. Ostwald ripening in foams leads to the counterintuitive phenomena of
large bubbles in a foam expanding as smaller ones shrink. This process results both in a
reduction in the number of bubbles and in the dispersity of bubble sizes. However,
Owen [12] has demonstrated that Ostwald ripening plays only a minor role in PU foam
stability; therefore, this chapter will focus on the process of bubble coalescence.
Bubble coalescence involves the merging of two distinct bubbles into one, due to the
rupture and recession of the lamella. Early in the PU foaming process, the rupture of a
lamella at the edge of the PU foam will cause bubble collapse as the gas inside is lost to
the surrounding atmosphere. This process will reduce the number of bubbles present
and the foam volume. Lamella rupture within the interior of the foam will cause bubble
coalescence, decreasing the number of bubbles, increasing the average bubble size, but
not necessarily decreasing the foam volume. Later in the foaming process, as the PU gel
network sets up, the lamella may rupture but the two bubbles can retain their separate
identities. This process is known as cell opening. If many of the lamellae have ruptured
in this fashion, the object is more correctly called a sponge instead of a foam. In this
case, air can easily be pushed out of the sponge if it is loaded with weight (as for example
when one sits on PU foam padding on a car seat). Therefore, the mechanical properties
of the PU foam depend greatly on the degree of cell opening present. There is general
agreement that a strong stabilising effect from the surfactant is necessary early in the
214
Modelling the Stabilising Behaviour of Silicone Surfactants…
foam process. The stabilising role, if any, of the surfactant late in the foaming process
remains a controversial issue [2, 19-21, 23-27]. This chapter will focus on the earlier
stages of the foaming process, before the gelling of the liquid PU phase.
The mechanism of rupture depends upon the thickness of the lamella film. Ultra-thin
films, with thicknesses in the 1-20 nm range, are believed to rupture due to the
amplification of thermally-generated surface waves [28-30]. However, studies by Artavia
and Macosko [31] and Akabori and Fujimoto [32] demonstrated that most of the ruptured
lamellae in a PU foam are in the 200-1000 nm thickness range. Therefore, it is assumed
here that the mechanism of rupture of these thicker films is most important when
considering PU foam.
It has been extensively demonstrated that ‘thick’ liquid films rupture due to the
formation (nucleation) of a pinhole in the film which spreads outward radially until
the film is destroyed. The nucleation of the pinhole could result from either the presence
of fluid instabilities in the film (for example, Marangoni, Rayleigh-Taylor or Benard
instabilities) [33] or else by the presence of immiscible microphases which could serve
as nucleation sites.
For PU foams, it has been proposed that film rupture is nucleated by the phase separation
of polyurea segments [19-21, 24-26]. It was postulated that the phase separation process
could be retarded by the presence of silicone surfactants containing a high percentage of
polyoxyethylene. The polyoxyethylene segment of the surfactant could interact strongly
with polyurea via the formation of a hydrogen bonding network. However, in cases
where the silicone contains lesser amounts of polyoxyethylene, such as in many commercial
foam stabilisers, a direct effect of the concentration and structure of the silicone surfactant
on this phase separation process has not been found [2, 26].
Once nucleated, the film hole can spread out radially at a rate inversely proportional to
the bulk viscosity of the film. Spreading can occur until the film is destroyed. For
example, in typical aqueous films, the rate of spreading is so high that high-speed
photography is necessary to capture the event. However, in a gelling PU foam, the rate
215
Advances in Urethane Science and Technology
of spreading can be very low, and, in fact, films can have multiple holes in them without
being destroyed [24-26, 32].
Detailed investigations on the effect of silicone surfactants on the nucleation and growth
of holes in PU films have not been reported. Furthermore, that is not the topic of this
chapter which is specifically concerned with the effect of the silicone surfactant on the
rate of drainage of PU films. However, this topic is highly related to film rupture as it has
been shown that the probability of film rupture is an inverse squared function of the film
thickness [35]. The time-dependent film thickness is a function of the drainage rate of
the film.
Thin film drainage has been investigated in detail both theoretically and experimentally
[22, 28-30, 35]. An extensively applied physical model of a draining thin film is the one
of Reynold’s, a form of which is shown in Equation 1:
where VRe is the velocity of film thinning, F is the external force on the film (causing
drainage), k is the film thickness, μ is the dynamic viscosity, and R is the film radius.
• the surfaces of the film are circular, immovable, parallel plates; surface velocity equals
zero, surface viscosity is infinite.
• the flow inside the film is driven by gravity and/or pressure gradients and is resisted
by the bulk viscosity of the liquid.
The dependencies of drainage (flow) rate on the variables of film thickness, area and
bulk viscosity are quantified in Equation 1. For example, with decreasing film thickness
k, the drainage rate decreases in proportion with k3; furthermore the drainage rate
decreases with increasing viscosity.
216
Modelling the Stabilising Behaviour of Silicone Surfactants…
Previous discussions [34] of PU film drainage rate (and ultimately foam stability and cell
opening) focused on the dependence of rate on bulk viscosity. Per Reynold’s model, factors
which increased the bulk viscosity, such as catalyst concentration and polyol reactivity,
presumably decreased film drainage rate and yielded a more stable, less porous foam.
The Reynold’s model and equation have also been modified to account for surface effects
on film drainage rate [29]. To accomplish this, the assumption (boundary condition) of
a surface velocity of zero must be relaxed. This change also decreases the surface viscosity
from an infinite to a finite value. This process yields Equation 2. Allowing for a finite
surface velocity increases the film drainage rate from that which would be expected
under the Reynold’s conditions.
V / VRe = 1 + 1 / e f (2)
Where V/VRe is the measured velocity of thinning divided by the velocity of thinning
under the conditions of the Reynold’s equation and ef is the surface mobility factor.
The surface mobility factor is analysed in detail in [36]. This analysis demonstrates that
two factors control the degree of surface mobility: the surface viscosity and the presence
of surface tension gradients.
Surface viscosity scales directly with the surfactant surface concentration, the
intermolecular cohesion between the surface molecules and the intermolecular adhesion
between the surfactant molecules and the underlying bulk liquid layer.
Increasing the surface concentration decreases the average distance between molecules,
therefore increasing the sum value of the attractive molecular forces between them. In
217
Advances in Urethane Science and Technology
many cases, high surface concentration yields a low film drainage rate and a highly
stabilised foam.
The intermolecular cohesion is the result of the balance of attractive (van der Waals,
polar bonding) and repulsive (ionic) intermolecular forces. The total cohesive energy is a
product of the total area of cohesion and the cohesive energy per unit area. For example,
the cohesion of the polyether chains of a typical silicone surfactant yields most of the
surface viscosity contribution in PU films and foams [38]. Although it has not been
investigated, one could speculate that the surface viscosity would scale with the length of
the polyether chain. Increasing the chain length increases the potential area of contact
between the chains.
The unit area cohesive energy is a function of the strength of the intermolecular bonds
present. For example, silicone surfactant molecules could possibly cohere via van der
Waals bonding of the methyl groups of the siloxane backbone or else via polar bonding
between the ethylene oxide units in the polyether chain. The van der Waals bonding
between methyl groups is known to be weak; consequently, the surface viscosity of
unsubstituted PDMS is close to zero [15, 38]. However, the polar bonding between the
polyether chains would be much stronger and, therefore, the source of surface viscosity
measured in silicone polyether copolymers.
In order to maximise surface viscosity, and therefore minimise the drainage rate, the
surfactant concentration, the intermolecular cohesion and adhesion should all be high.
Later in the chapter evidence is presented correlating silicone surfactant concentration
and structure (which influences the intermolecular cohesion and adhesion) to surface
viscosity and film drainage rate.
An idealised model of the surface of a thin liquid film is one of a monolayer of evenly-
distributed surfactant molecules. However, a more realistic model is one where the
molecules are not evenly distributed; therefore, the surface concentration depends on
surface position. The result of this heterogeneous distribution is that gradients of surface
concentration, and therefore surface tension, are present. One example of this was pointed
out in Section 5.1 on the effect of Marangoni instabilities on film rupture. Regarding
film drainage, a surface tension gradient exerts a surface stress that can either impede or
218
Modelling the Stabilising Behaviour of Silicone Surfactants…
accelerate the underlying bulk flow (drainage). In thin films, these gradients typically
make the surface rigid, retarding flow and therefore decreasing, drainage rate. The
gradients can be relieved by the diffusion of surfactant from areas of high to low
concentration; the drainage rate of the film is proportionately increased.
Diffusive surfactant fluxes are functions of the intensity of the concentration gradient
and the diffusion coefficient of the surfactant. The intensity of the concentration gradient
depends upon the overall surfactant concentration. At low or high concentrations, these
gradients are relatively weak. At intermediate concentrations, they are quite strong.
The diffusion coefficient of the surfactant is a function of its size and shape. The most
simple and common case to analyse is where the surfactant assumes a spherical shape.
In this case, the diffusion coefficient of the surfactant scales inversely with the solvated
molecular volume (Stoke’s law).
This section begins with a qualitative description of thin liquid PU films. This initial
investigation had five goals in mind: to confirm that stable, vertically-oriented, thin liquid
films could be prepared using mixtures of ingredients designed to model a PU foam, to
study the hydrodynamic phenomena in the films, to compare the physical behaviour of
these films to the behaviour of the more common aqueous soap films, to observe specific
surfactant effects on the properties of these films, and to extrapolate conclusions about
the behaviour of these films to operational PU foam.
After this qualitative description, an accurate measurement of the drainage rates of these
films was sought in order to study the effect of bulk and surface variables on the rate. In
219
Advances in Urethane Science and Technology
Once this method was validated, the effect of silicone surfactant concentration and structure
on the drainage rate of the films was investigated. Generally, the drainage rate displayed a
maximum as a function of surfactant concentration. This maximum was consistent with a
physical model where the two major influences on drainage rate are surface tension gradients
and surface viscosity. As surfactant concentration increased, the retarding effect of gradients
decreased, and the film drainage rate increased. However, above a certain concentration, the
drainage retarding effect of surface viscosity overcame the gradient effect.
The dependence of film drainage rate on silicone surfactant molecular structure was also
systematically investigated. In order to understand this correlation, three physical
parameters of the film affected by surfactant structure must be considered. These
parameters are the surface partition coefficient, the surfactant molecular diffusion
coefficient and the degree of intermolecular cohesion within the surface layer. Specifically,
as the length of the polyether (solvophilic) portion of the surfactant increased, the surface
partition coefficient decreased, the diffusion coefficient decreased, and the degree of
cohesion increased. This resulted, at constant surfactant concentration, in a complex
effect on the film drainage rate.
A quantitative physical model of a draining vertical thin film was developed from first
principles. The starting point was the Navier-Stokes equation. The initial model featured
a fixed-surface, wedge-shaped vertical film, with immobile surfaces. This is essentially
the Reynold’s model modified to this film shape and orientation. Good agreement of the
predictions of this model with experimental data was obtained.
The next model relaxed the condition of fixed film shape (allowed for curvature in the
film surface) and analysed the effects of the menisci on the film drainage. Analysis of the
Navier-Stokes equation was simplified by the application of the lubrication approximation.
The results from this analysis agreed extremely well with experimental values both in
terms of film drainage rate and the changes in film shape with time. The development
and growth of bulges and waves on the bottom of the film were particularly intriguing.
This phenomenon was experimentally observed in these films.
Finally, models were developed where the condition of infinite surface viscosity was relaxed.
This allowed the analysis of surfactant effects on film drainage, in particular surface viscosity
and surface transport. Specifically, the model predicted the decrease in drainage rate as surface
viscosity increases, as expected from the qualitative models and measured experimentally.
The effect of surface transport was significantly less than that of surface viscosity.
220
Modelling the Stabilising Behaviour of Silicone Surfactants…
Unless noted otherwise, the experimental work in this chapter involved a model flexible
slabstock PU foam formulation [39] at the instant of mixing (see below):
• 5.46 g toluene,
For simplicity’s sake, chemical reactions were avoided. This was accomplished by
substituting toluene for toluene diisocyanate (TDI) in the formulation. This formulation
had a nominal viscosity of about 0.8 Pa-s.
The following experimental variables were addressed during the design and construction
of the interferometer:
221
Advances in Urethane Science and Technology
Consideration of these points led to the experimental setup sketched in Figure 5.1.
The film is formed between two narrow blades set by 10 mm apart and supported in a
surrounding frame; a film of 20 mm height could be formed between the blades. The
frame was vertically clamped on an optical bench. A glass cuvette with rectangular
cross section was fabricated with a closely fitting lid through which the clamp rod fits
tightly in order to prevent evaporation of the solvent and at the same time to allow the
cuvette to be lowered and raised to adjust the bulk liquid level or to draw a film. The
cuvette stands on a small pedestal which was driven by a computer-controlled stepper
motor assembly allowing film size control at velocities of the receding or advancing
bulk liquid level up to 25 mm/s. The sample holder assembly was mounted on an
optical bench.
The two nearly plane-parallel surfaces of the film constitute a Fizeau interferometer
which produces interference fringes in the reflected light [40]. The fringes are lines
of constant film thickness. Film drainage data were acquired by generating
monochromatic light by means of an interference filter (λ = 505 nm) or with a HeNe
laser (λ = 632.8 nm).
222
Modelling the Stabilising Behaviour of Silicone Surfactants…
Figure 5.2 Depiction of film formation. The film is formed within the boundaries of
the inner frame by the lowering of the liquid level at a controlled rate. The frame and
liquid are within the transparent walls of the enclosed glass cuvette.
The process of film formation is depicted in Figure 5.2. Films were formed by lowering
the sample liquid from a point where the vertical frame was completely immersed to a
point where it was only partially immersed. In most experiments, the vertical film so
formed was left in direct contact with the bulk liquid. Films were withdrawn from the
bulk liquid at a rate varying from 0.25 to 50 mm/s.
In a typical experiment, stable, vertical, liquid films were formed from the model PU
formulation (see Section 5.3.1.1). These films were stable for two to five minutes. The
series of photographs in Figure 5.3 depicts many of the physical features of these films.
Thirty to sixty seconds after film formation, dark horizontal interference fringes were
observed that initially appeared at the top of the film and steadily moved downward.
Over time, the number of fringes decreased. Flow patterns, including eddy currents,
fingering patterns and swirls, appeared at the bottom and sides of the films and rose
upwards. The rate and amount of these flows decreased as the lifetime of the film increased.
Films were subjected to a rapid raising and lowering of the bulk liquid level. These
motions strongly stimulated the surface flows previously mentioned.
223
Advances in Urethane Science and Technology
Figure 5.3 The visual sequence of a draining film. The first photograph in the
sequence is in the top left hand corner and the sequence of photographs proceeds
clockwise. Photograph 2: one minute after formation. The dark horizontal lines
faintly present are the interference fringes. Photograph 4: 3-5 minutes after film
formation. Fewer fringes are present and span the height of the film. The reduction
in the number of fringes versus Photograph 2 is consistent with a significant amount
of film drainage. Photograph 3: 8-12 minutes after film formation. Photograph 1:
approximately 15 minutes after film formation.
The pattern of horizontal fringes observed on the face of the films suggested a vertical
wedge shape, with the film thickness increasing as one descends the film. The wedge
shape of the vertical PU films has a number of implications for operational PU foam.
Any cell window in PU foam does not have uniform thickness. This non-uniformity has
been reported in at least two studies [24-26, 32]. In addition, it would be expected that
the rupture of the cell window would occur at its thinnest point [24-26, 32].
As seen in Figure 5.3, the interference fringes progress down the face of the film and the
distance between the fringes increases as the film continues to drain. This indicates a
decrease in the film thickness gradient during the drainage of the film. The drainage of
the wedge-shaped film can be visualised in that the sides of the wedge, intersecting at the
224
Modelling the Stabilising Behaviour of Silicone Surfactants…
top of the film, pivot inwards as fluid drains out of the bottom. This is a collapsing
inwards of the wedge.
As the film is draining, interference fringes radiate outwards from the top centre of the
film (photographs 2-4 in Figure 5.3). This radiation suggests that both horizontal and
vertical flow processes are present. These flow processes are driven by both gravity (vertical
flow) and suction into the Plateau borders of the film (horizontal flow). Once the fluid
has entered the Plateau border, it flows down a relatively wide channel into the bulk
liquid. The Plateau borders in these films can be seen in Figure 5.3. They are the dark
regions between the film and the frame. The width of the Plateau channel increases as
one descends the face of the film, giving it roughly a triangular shape. Plateau borders
are also present in operational polyurethane foams. In fact, horizontally-aligned films
within these foams can only drain due to Plateau border suction. The individual Plateau
borders within the PU foam form a network which allows for liquid drainage down the
foam. Ultimately, this network becomes the structural ‘struts’ of the foam.
As reported in the literature, interference fringes in many vertical aqueous films are
horizontal with little curvature. These films also displayed ‘mobile’ surfaces, rapid and
extensive surface flows and rapid rates of drainage [22, 28-30]. These features were all
linked to the presence of a low surface viscosity. Based on our experimental observations,
the PU films stabilised by DABCO DC 198 have low surface viscosities. This conclusion
is supported by reports of measurements of low surface viscosities in polyol solutions
containing silicone polyether surfactants [12, 15, 38].
The phenomena of edge turbulence, fingering and upward flows in aqueous soap films
have been extensively investigated [22, 28-30, 41-46]. These hydrodynamic phenomena
have been defined as marginal regeneration and gravity convection. Marginal regeneration
refers to a process where, simultaneously, thick films are sucked into the Plateau border
and thinner films are pulled out. This exchange results in a net increase of material in
the Plateau border, essentially draining the film. Within the film, after the material
exchange, the new thin spots then migrate upward (gravity convection) until they reach
a height where their thickness equals the film thickness.
225
Advances in Urethane Science and Technology
The surface tension gradients necessary to drive this type of flow can arise from three
different physical phenomena: surface tension gradients are necessary to support the
weight of the vertical film, flows in the film or the channel of the Plateau border impose
stresses on the surface monolayer which are balanced by a surface tension gradient, and
extension or compression of the surface layer can create transient surface tension gradients.
Regarding the cell windows of operational PU foam, it is apparent that surface tension
gradients are constantly present, due to all three mechanisms discussed previously. In
particular, the constant stretching of the bubbles during the process of foam growth acts
as a potent stimuli for these flows.
The distribution of the interference fringes in a liquid film can be viewed as a contour
map of the film thickness, with each successive fringe representing a section of constant
film thickness (isopach). Therefore, the distance between fringes throughout the film
gives the gradient of film thickness. In order to precisely measure PU film drainage rates,
a physical model correlating the time change in the fringe density (ds/dt; the quantity
that can be measured in the laboratory) to the time change in volume (dV/dt, the film
drainage rate) of the film must be applied.
Qualitatively it appears that the drainage of the film collapsed the sides of the film wedge
inwards. During this collapsing process, the two sides of the wedge act as if they were
hinged together at the top of the film. With this geometric model of film shape, the
drainage rate of the film, dV/dt, can be expressed mathematically by Equation 3.
dV c ds
= WL2 (3)
dt 2 dt
Where V is the film volume, t is time, c is the change in film thickness per interference
fringe, W and L are the film width and height, respectively, and s is the fringe density.
If both sides of Equation 3 are divided by the cross-sectional area (WcsL) of the film at
the bottom Equation 4 is obtained:
226
Modelling the Stabilising Behaviour of Silicone Surfactants…
The quantity on the left of Equation 4 is the flux. If it is assumed that the flux is a
hyperbolic function of time, namely,
where b is a drainage constant and t0 is the time where hyperbolic drainage begins.
2b
s = k0 (t + t0 )
m
with m = − (6)
L
As Equation 6 shows, a log-log plot of fringe density versus time should be linear for t >
t0 with a slope of m (<0) and an intercept of logs0 = logk0 + mlogt0. This prediction was
realised experimentally with high degrees of precision and accuracy.
c
dV / dt = WL2 ⋅ mk0 t m −1 = −cWLbk0 t m −1 (7)
2
The drainage rate measurement described in Section 5.3.3.1 was novel and required
experimental verification of its accuracy and precision. This was accomplished by
measurement of the dependence of film drainage rate on the reciprocal of bulk viscosity
(see Figure 5.4). The Reynold’s equation (Equation 1) predicts that this correlation should
be linear. As seen in Figure 5.4, a linear correlation was obtained. It should be noted that
this form of the Reynold’s equation assumed an infinite surface viscosity. This condition
was not able to be achieved in practice. However, it was assumed that the surface viscosity
was constant throughout the range of experiments and that Equation 1 was valid under
this condition.
The excellent correlation achieved between film drainage rate and reciprocal bulk viscosity
motivated the measurement of the correlations between drainage rate and surface variables.
227
Advances in Urethane Science and Technology
Figure 5.4 The experimental dependence of film drainage rate on bulk viscosity. The
bulk viscosity was varied by varying the ratio of toluene to polyol in the formulation
Initially, the drainage rates of PU films stabilised by various commercial silicone surfactants
were measured. These surfactants are applied in the marketplace to stabilise flexible
slabstock PU foam. These data are given in Table 5.1.
The surfactants listed in Table 5.1 span a range of drainage rates of 1.5 to 2. The
narrowness of this range is not surprising because the PU foam stability (in terms of
foam volume and porosity) achievable with these commercial surfactants does not vary
to a great extent. By comparison, as described in Section 5.3.3.3.2, rigid-surfaced films
have been observed which have drainage rates of less than 1% of those listed in Table
5.1. Within this context, the drainage rates in Table 5.1 are rather fast, consistent with
the previously discussed low surface viscosity of the films. This is also consistent with
228
Modelling the Stabilising Behaviour of Silicone Surfactants…
the consideration that these surfactants have been optimised to provide sufficient drainage
rate to produce porous, flexible slabstock PU foam.
Physical models for the potential effect of surfactant concentration on film drainage rate
have been discussed in detail in Section 5.2. It was predicted that surface tension gradients
and surface viscosity affect film drainage rate. It was expected that the drainage rate would
increase with an increase in surfactant concentration due to the lessening of the intensity of
surface tension gradients. However, It was also expected that the drainage rate would
decrease as the surfactant concentration increased due to the increase of surface viscosity.
This correlation was examined for two different silicone surfactants as illustrated in Figure
5.5. These plots show a maximum in film drainage rate as a function of surfactant
concentration, consistent with the concept that the overall rate is the result of two opposing
229
Advances in Urethane Science and Technology
Figure 5.5 The effect of surfactant concentration on film drainage rate: Comparison of
the two surfactants DC 198 (solid line) and DC 5950 (dotted line) 10 mm high films
were drawn from the standard PU film formulation at a rate of 0.25 mm/s
effects. At lower concentrations, the increase in drainage rate observed with increase in
surfactant concentration is consistent with a decrease in the intensity of surface tension
gradients in the film due to increased bulk-to-surface surfactant flux. The decrease in drainage
rate seen at higher surfactant concentrations is consistent with an increase of surface viscosity.
In their studies of three-phase capillary air slug (bubble) flow, Stebe and Maldarelli [47]
correlated surfactant concentration to equivalent flow rate invoking similar reasoning.
At a higher surfactant concentration, flow rate increased with surfactant concentration,
due to the relief of surface tension gradients at the air bubble surface. When using
protein-based surfactants, retardation of flow due to surface viscous effects (flow rate
decreased with an increase in protein concentration) was observed.
230
Modelling the Stabilising Behaviour of Silicone Surfactants…
Following the work of Kanner and co-workers [37], the drainage of films stabilised by a
TCP was examined. As previously mentioned, they observed a significantly low film
drainage rate and attributed it to the high surface viscosity measured for these solutions.
In studies of high TCP concentrations, rigid-surfaced films with very slow drainage (as
detected by the movement of interference fringes) and surface flows were observed. At
low TCP concentrations, the films were mobile-surfaced and fast draining. These
observations are consistent with the concept that the surface viscosity increases with
surface concentration. Further reinforcement of this concept is given by the data in
Table 5.2. Assuming that surface viscosity is proportional to surfactant concentration,
the inverse dependence of film drainage rate on surface viscosity is quite apparent
from this data.
Previously, in the literature, the ‘extremes’ of mobile and rigid surfaces were much
discussed without significant mention of ‘intermediate’ behaviour. Intermediate behaviour
was studied here by simple titration of the appropriate amount of TCP into the film
solutions. In this regime, the film drainage rates were intermediate between those of
mobile and rigid-surfaced films, as were the rate and extent of surface flows.
For both Dabco DC 198 and Dabco DC 5950 (Air Products and Chemicals, Inc.), the
surfactant concentration (approximately 1 pph) leading to maximal drainage rate, is
similar to that optimally used in manufactured PU foam. Presuming that the surfactant
concentration needed for effective film stabilisation is similar to that needed for effective
foam stabilisation, it appears that a maximal drainage rate aids in the optimal production
of flexible slabstock polyurethane foam.
231
Advances in Urethane Science and Technology
Once the effect of surfactant concentration on film drainage rate had been measured, the
effect of molecular structure was investigated. It was hypothesised that the surfactant
structure would affect the intensity of surface tension gradients within the film and the
surface viscosity. Within this framework, the effect of molecular structure on the diffusion
and surface partition coefficients, and on the degree of intermolecular cohesion in the
surface layer was investigated.
By comparing the films prepared from TCP solutions to those from solutions of highly
purified (Me3SiO)8Si8O12, it was deduced that the source of the surface rigidity was the
OH (silanol) groups on TCP. These groups can hydrogen bond to each other, thus forming
a highly cohesive network at the surface and yielding a high surface viscosity. A similar
effect was also noted in aqueous films containing carbohydrate functional organic
surfactants [48]. The source of OH groups in these surfactants are the carbohydrate
functionalities.
These results motivated a more careful look at the specific effects of surfactant structure
on film drainage rate using traditional silicone PU foam stabilisers. In order to accomplish
this, the drainage rates of PU films stabilised by a series of surfactants of the general
molecular structure (Me 3 SiO)(SiMe 2 0) 100 (SiMeRO) 10SiMe 3 (R= -(CH 2) 3 (EO) x
(PO)xOC(O)Me, EO = ethylene oxide, PO= propylene oxide, X= 6, 12, 18, 21, 30) were
investigated. These five surfactants differed only in the length of the alkylene oxide
chains. Therefore, for the rest of this chapter, they will be referred to only by their X
value. Relevant measured physical properties are listed in Table 5.3. Film drainage rate
data is shown in Figure 5.6. Unfortunately, it was not possible to obtain accurate surface
viscosity data for these materials.
Analysis of the data in Table 5.3 and Figure 5.6 begins with four hypotheses regarding
the physical consequences, to a silicone/polyether copolymer, of an increase in polyether
length (X) at constant siloxane chain length. An increase in X leads to a copolymer with:
2) A larger solvated volume and therefore a lower diffusion (through the liquid phase)
coefficient.
232
Modelling the Stabilising Behaviour of Silicone Surfactants…
Figure 5.6 Film drainage rate as a function of surfactant concentration and structure
(X = length of alkylene oxide chain)
233
Advances in Urethane Science and Technology
4) A larger area of polyether chain cohesion in the interfacial layer. This increase in
cohesion area should result in an increase in cohesion energy per molecule and
therefore, at surface saturation, a higher surface viscosity.
Generally, the experimental data in Table 5.3 are consistent with these hypotheses.
Firstly, the surface partition coefficient, Kad, is the ratio of the surface saturation
concentration (Γ∞) to the overall amount of surfactant needed to saturate the surface
(keeping in mind that a significant amount of the surfactant remains solubilised).
Therefore, at relatively constant Γ∞, Kad should decrease as the solubility of the
copolymer increases. This can be seen in Table 5.3. Secondly, the solvated volumes
were not directly measured; however, the diffusion coefficients were measured and
they decreased as X increased. It is generally known that the diffusion coefficient varies
as the inverse of the solvated volume [49]. From this relationship, one would conclude
that the solvated volumes of the copolymers increased with X. Thirdly, the interfacial
areas of the copolymers generally increased as X increased. Finally, we were not able
to measure the surface viscosities of these materials in solution. However, all of them
yielded mobile-surfaced films. This means that their surface viscosities were relatively
low, as expected.
It is unclear why the X = 30 copolymer did not give data consistent with the expected
trends in the data observed for X = 6-18. It would have been helpful to have had more
data on the X = 21 species; perhaps, the deviations observed for the X = 30 species
would have also been recognisable for the X = 21 data. For X = 30, one possible factor
which could have yielded non-compliant data would have been an unexpected
configuration of the polyether groups. After all, they are quite long (total of 60 monomer
units). Perhaps they assumed an unusual coiled or spiral configuration both in bulk
solution and at the interface. This would be consistent with an unexpected solvated
volume, degree of solvation, interfacial area and area of polyether cohesion. Previous
work with silicone polyethers has demonstrated unexpected results with exceptionally
long polyethers [50].
One goal of this study was to correlate the data in Table 5.3 with the experimental
drainage rate data for these surfactants, displayed in Figure 5.6. In order to interpret
these correlations, the following hypothesis were developed.
234
Modelling the Stabilising Behaviour of Silicone Surfactants…
• As X increases, Kad and D should decrease, which decreases the efficiency of relieving
surface tension gradients. Also, the surface viscosity should increase. Both factors
should cause a relative decrease in film drainage rate and furthermore, an increase in
film stability.
The profile for X = 6 is rather complex. Ignoring the points at very low concentrations
(which are difficult to interpret), the profile has only the ‘first half’ of the expected
parabolic shape. At moderately low concentrations, the drainage rate increases with
concentration, consistent with the decrease in the intensity of surface tension gradients.
Incidentally, this decrease in gradient is also reflected in the fact that the film lifetimes
decrease in this range. In fact, above a certain concentration, films are too unstable to
allow for the measurement of the drainage rate. This high degree of (Marangoni)
instability, only observed for the X = 6 species, is consistent with an overly-efficient relief
of surface tension gradients in the film. This efficiency is due to the high bulk-to-surface
flux of this surfactant, which itself is due to its high values of Kad and D. At high surface
concentrations, the films could potentially be stabilised by a high surface viscosity.
However, for X = 6, the surface viscosity would be expected to be low, and this is borne
out by the experimental data, i.e., the films are so unstable that drainage rate measurements
could not be made.
The drainage rate for X = 6 should be the highest of any of the surfactants. This is not
confirmed by experiment. It is possible that the films where this would be demonstrated
(at higher surfactant concentrations) were simply too unstable to be investigated.
The expected parabolic shape of the drainage profile is observed for X = 12, and at a
relatively low concentration, as might be expected. These films were more stable than
those exhibited for X = 6, due to a less efficient relief of surface tension gradients. This
would be expected, as Kad and D should be lower than for X = 6. The decrease in drainage
rate at higher concentrations is due to the build up of surface viscosity. A higher surface
viscosity than for X = 6 would be expected, as the area of polyether cohesion has increased.
The drainage rate profile for X = 18 is at a lower value than either X = 6 or X = 12,
consistent with stronger surface tension gradients and higher surface viscosity. The fact
that film stabilities were higher than for either X = 6 or X = 12 is consistent with this
conclusion.
It is intriguing that, for X = 30, the ‘first half’ of the drainage profile is not observed at
all. Perhaps it could exist at very low surfactant concentrations, but these films prematurely
rupture. A sufficiently high concentration of surfactant is necessary at the surface to
allow for the presence of the surface tension gradients that stabilise films in this regime
235
Advances in Urethane Science and Technology
(in other words, ‘high’ but not ‘too high’). These low gradients would result in a high
drainage rate and, a marginally stable film. As the concentration of this surfactant is
increased, the drainage rate is strongly retarded due to the build up of surface viscosity.
Viscous effects would be expected to be particularly strong in this system due to the long
polyether chains, and the resulting large area of cohesion.
Overall, the limitations of analysing these data in terms of simple, qualitative models
motivated the development of more refined, quantitative models of the draining film.
The development of these models is the topic for the next section.
In quite a few of the experiments, there was an extension region in the upper part of the
film that was essentially two-dimensional. In particular, this occurred in the early part
of the experiments (typically several minutes, at least) and in the upper half to two-thirds
of the film for nearly the entire width. Under these circumstances, the interference fringes
were essentially horizontal and parallel. These were also the conditions under which the
exponent for the drainage rate was measured. This leads us to consider two-dimensional
models for film drainage; they can be considered as a cross-sectional slice down the
centre of the film perpendicular to the surface of the film. All of the theory described in
this section will be for this two-dimensional case.
To develop a quantitative model, the problem was approached in stages. The first
model assumed that the surface of the film was wedge-shaped; the narrowing with
time was predicted in the tangentially-immobile case [51]. The film was then modelled
with a deforming surface; the film still thins, but it takes on more complicated shapes
predicted by the solutions of non-linear partial differential equations. These
generalisations have resulted in a series of models that gradually incorporate more of
the experimental behaviour. A few details and some results from these models will
now be discussed.
The simplest model was a theoretical version of the collapsing film wedge. As discussed
previously (Section 5.3.3.1), the collapsing wedge geometry was imposed on the analysis
of the experimental data to generate instantaneous film drainage rates. Therefore, it
seemed reasonable to use this same shape in the development of the first theoretical
model [51]. This led to the condition that the shape of the film was invariant with
236
Modelling the Stabilising Behaviour of Silicone Surfactants…
time. The other key assumptions applied in the construction of this model were that
all of the flow in the film was downward (parallel flow) and that the surface of the film
was tangentially immobile (infinitely large surface viscosity). Therefore, outside of
surface viscosity, there was no specific surfactant effects on the drainage rate of the
film. In fact, no surfactant transport or concentration gradient effects were incorporated
in the model.
A diagram of the model construction is shown in Figure 5.7. The film shape is represented
by the solid line; the wedge shape that replaces the deforming film shape is sketched with
the dashed line.
Figure 5.7 Schematic representation of the model film studied. The dashed-dotted line
is the centreline of the film; the film is assumed to be symmetric around it. The solid
curve is a sketch of the actual film surface; the dashed line is a wedge approximation
to the film shape. Differences are exaggerated to clarity.
237
Advances in Urethane Science and Technology
Within the wedge, the momentum in the vertical direction is conserved by that component
of the Navier-Stokes equations:
∂w ∂p ∂2 w
ρ = ρg − +μ 2 (8)
∂t ∂z ∂c
To simplify matters, it is assumed that the flow is so slow and gradually changing that
the left hand side of the equation can be neglected. The flat sides of the wedge along
with the slow flow assumption allow the pressure gradient term to be neglected. The
differential equation for the velocity profile in the film is then obtained:
∂2 w ρg
+ =0 (9)
∂x2 μ
The solution to this equation is subject to ∂w/∂x = 0 in the centre of the film and ω = 0 on
the surface of the film, x = k (z, t). The velocity profile is given by
ρg 2
w( x, t ) =
2μ
(k − x2 ) (10)
ρg k2
w= (11)
3μ 4
This equation clearly indicates that drainage of the film is accelerated by increasing
density and inhibited by increasing viscosity.
Further analysis of this model yields a prediction that the slope of the plot of log (fringe
density) against log (time) should be -0.5. These predicted results are in excellent
agreement with experimental results where a rigid-surfaced film was investigated;
experimental data gave a slope of -0.47. For the mobile-surfaced films investigated, the
value of the shape ranged from -0.6 to -0.92; the more negative the value, the faster the
film drainage.
At this point, theory could be developed only for the rigid case; the development of
theory for the mobile cases requires more complicated mathematical models, which are
described next.
238
Modelling the Stabilising Behaviour of Silicone Surfactants…
The mathematical problems in this section are specified in the following way, unless stated
otherwise. The Navier-Stokes equations govern the motion of the incompressible,
Newtonian fluid inside the film [52]. The film is assumed to be symmetric about its vertical
centreline (the z-axis). On the free surface of the fluid, several equations must be satisfied:
the kinematic condition, and the normal and tangential stress conditions [53, 54]. An
equation governing the transport of surfactant in the free surface by both advection (fluid
239
Advances in Urethane Science and Technology
motion) and diffusion also holds [55, 56]. The surface tension, σ (in mks units: N/m), of
the film-air boundary is assumed to vary linearly with the surface concentration, Γ:
∂σ
σ = σm + Γm (Γ − Γm ) (12)
∂Γ
The subscript m denotes the reference values of σ and Γ; these reference values are chosen
for convenience in matching up with experiments. The additional parameters that show
up in the boundary conditions on the free surface are the surface shear viscosity and
surface dilatational viscosity, represented by μs and κs, respectively, and the surface
diffusivity of the surfactant denoted by Ds.
No fluid can enter or exit through the wire frame at the top, and the film surface is
pinned there. At the bottom, the film drains into the bath (or pool) in the cuvette, from
which the film is drawn. The bath is otherwise inert, and it is assumed that the film
shape tends to a static meniscus shape over the bath. This assumption will allow boundary
conditions over the bath to be specified. The details of the mathematical specification of
the problem appear elsewhere [57-60]; in this section, the focus is on simplified models
and comparing results from them with the experimental results discussed previously.
The equations mentioned above will be made dimensionless with the length scales
μV0 σ
d= , D= and l = d1 / 3 D2 / 3 (13)
ρg ρg
d, D and l are the film thickness, the static meniscus radius, and an intermediate scale,
respectively.
The horizontal length scale is d, while the length scale along the film is l . Of particular
importance is that, for the PU solvent of interest here, the ratio of the lengths d/ l is a
small quantity. The ratio
δ 2 ≡ d / l << 1 (14)
is used to reduce the original mathematical fluid dynamics problem to a simpler set of
partial differential equations; the mathematical method is a multiple scale analysis, and
when it is applied to thin fluid layers, it is called lubrication theory. For recent reviews of
this type of approach, see references [29, 61]. We choose
ρgd 2
V0 = (15)
μ
240
Modelling the Stabilising Behaviour of Silicone Surfactants…
as the velocity scale; in mks units, V0 is in m/s. The time scale is l / V0 and the pressure
scale is based on viscous shear, μV0 / ( lδ 4 ). The surface concentration is made
dimensionless with the reference value, Γm.
Several dimensionless groups appear in the non-dimensional equations; these are given
in Table 5.4. These non-dimensional groups serve to highlight the relative importance of
the myriad parameters in the problem; the interpretation of these groups in terms of
ratios of physical forces are also given in Table 5.4. These parameters will be referred to
as appropriate in the following sections.
For the typical experiment of interest, the Reynolds number is very small; this will help
make the simplifications that follow valid. The Péclet number (P) is large; it may be as
large as 104 at the start of some experiments, according to our estimates. In all the
computations presented here, P = 102 will be used. The typical value of δ is 0.1 to 0.2;
the capillary number is then smaller than 10-4. The other quantities will be varied in the
results to follow.
ρV0 l Inertial
Reynolds ℜ=
μ Viscous shear
∂σ
Γm Γm δ 2 Shear from concentration gradients
Marangoni M= ∂Γ
Viscous shear
μV0
V0 l Advective transport
Péclet p=v
Ds Diffusive transport
241
Advances in Urethane Science and Technology
For high concentrations of some TCP in PU, the surfactant develops structure on the
film surface and the surface of the film still deforms but doesn’t move tangentially. This
‘rigid’ case [22] is the slow draining limit of the experiment. In that case, the tangential
stress condition may be replaced by the condition that the tangential velocity component
is zero. We have been able to show mathematically that this case is achieved in the limit
of large surface viscosity [59, 60] or S → ∞; this will be discussed further in the next
section.
When lubrication theory is used in this context, a single, nonlinear partial differential
equation for the film shape k(z,t) is obtained. In one approach, which is called the
‘whole film’ model, the equation obtained is:
⎡k ⎤
kt + ⎢ 3 (κ z + 1)⎥ = 0 (16)
⎣3 ⎦z
which holds over the domain 0 < z < L. The subscripts t and z denote partial
differentiation; for example κz = δk/δz. Here
kzz
κ=
(1 + δ k ) (17)
3 /2
4 2
z
is the curvature of the film. At the top of the film, z = 0, the film is pinned (k(0,t) = 1)
and no fluid leaves or enters, by requiring that
κ z = −1 (18)
At the bottom of the film, we specify the slope, kz(L,t) = C, and the second derivative is
specified via
⎛ 1 ⎞
kzz (L,t ) = 2⎜ 2 −
C
( )
3 /2
⎟ 1+ δ C ≡ s1
4 2
(19)
⎝δ 1+ δ C ⎠
4 2
This approach is based on methods for including the static meniscus in film models that
have been developed by Ruschak [62] and Kheshgi and co-workers [63]; equation (12) is
a first integral of the equation for the static meniscus that is the surface of the bath.
242
Modelling the Stabilising Behaviour of Silicone Surfactants…
The initial condition used for all calculations was a linear shape for k. This was a
convenient shape, and for later times in the computations, the results are rather
insensitive to the initial conditions. This problem is readily solved using numerical
methods described in references [57].
Numerical results are given in Figure 5.9; one can see in the figure that distinct regions of
the film developed: (i) a meniscus at the top of the film (at the wire frame); (ii) a long flat
middle region where effects from the average surface tension were negligible; (iii) a region
where bumps and dips developed near the bottom of the film and (iv) a static meniscus at
the bottom of the film where it joined the bath.
Figure 5.9 Film shapes k(z,t) versus z with L = 37.5, h = 0.05, kz(L) = 5 and δ = 0.2.
The initial slope of the film was 0.23
243
Advances in Urethane Science and Technology
Figure 5.10 Computed thickness for models without and with surface tension, solid and
dashed lines, respectively, at t = 15 on the right and t = 231.9 on the left;
here δ = 0.25. The case with surface tension is computed with kz(L,t) = 0 and neglecting
all terms with d in equations (17) and (using C = 0 in Equation (19)); see [57] for details
A number of interesting possibilities arise. The bumps and dips at the bottom of this
film could be the precursor to complicated three-dimensional behaviour observed at the
bottom of films in experiments. Only further work will verify this conjecture, and we are
currently pursuing that line of inquiry.
The long middle region of the film offers the possibility of still further simplification. If
one rescales the variables in the problem again, a much simpler equation for the (rescaled)
free surface shape K(Z,T) arises, which we shall call the ‘flat film’ model. This model is
244
Modelling the Stabilising Behaviour of Silicone Surfactants…
KT + K 2 K z = 0 (20)
with the single boundary condition specifying K(0,T). Here Z =z/δ and T =t/δ are the
rescaled space and time variables, respectively. Solutions found numerically for the film
shapes from Equation (20) and a more complicated model for the film shape [57] are
given in Figure 5.10. As time increases, the agreement between the approximations
improves. In the middle part of the film, no matter which simplification is used, the
numerical results indicate that the film is flattening out with t-1/2 behaviour at long times.
This simple equation also has an exact solution, which has similarity behaviour for long
times given by K = Z / T [57, 61]. Similar solutions have been found by Mysels and
co-workers [22] and Moriarty and co-workers [64]. In this way, it can be found
analytically that the rate of flattening of the film is proportional to t-1/2 for long times, in
good agreement with the slowest-draining experimental case of t-0.47.
Based on that success, this kind of simplification will be attempted with more complicated
models to see if similar results can be achieved.
When there is finite surface viscosity, the surface of the film is mobile. Lubrication theory
for our vertical draining film gives a system of two partial differential equations describing
the evolution of the film surface k(z,t) and the velocity of the fluid at the surface w(S)(z,t).
These equations are
⎡ k3 ⎤
kt + ⎢ kw( s ) +
3
(1 + κ z )⎥ = 0 (21)
⎣ ⎦z
k
wzz( s ) +
NS
(1 + κ z ) = 0 (22)
Here
1
N≡ (23)
1 + δ 4 kz2
245
Advances in Urethane Science and Technology
Here C is the specified slope and a consistent s1 can be calculated with the first integral
of the equilibrium meniscus shape [62, 63], as in the rigid case discussed above. In
Equations (21) and (24), κ is the curvature of the film given in Equation (27).
Previous work [62, 63] has shown that it is possible to keep the full curvature in
the normal and tangential stress conditions and integrate the partial differential
equation for the free surface k through the matching region (where the film meets
the bath) onto the static meniscus; this model relies on those results. While we do
not rigorously apply matched asymptotic expansions in this work, the equations
contain all of the terms necessary to match the film onto a static meniscus (for the
bath surface) and it has been shown that the terms neglected are uniformly small
for δ<<1 [65].
Some computed results are shown in Figures 5.11 and 5.12. For large S, the film
drains slowly; the downward parabolic shape in the middle region (Figure 5.11; free
surface shapes for varying S) agrees with the typical rigid film profiles given in Chapter
III of Mysels and co-workers [22]. In the limit of S → ∞, w(S) approaches zero along
the whole film and the free surface shapes tend to that of the tangentially-immobile
case [57]. It is also found that for very large S, the film thickness in the middle
region tends to decrease with the power law t -0.5 as before; this is very close to the
experimental rate of t-0.47 obtained at Dow Corning [68] for their most rigid film and
agrees with results from Mysels and co-workers [22]. As S decreases, drainage is
much faster and the film thins more rapidly (Figure 5.11); free surface shapes for
smaller S, take on the typical shapes for mobile films given in Chapter IV of Mysels
and co-workers [22]. In the limit of small S, t-1 thinning in the middle of the film is
found which agrees well with the maximum experimental rate of t-0.92. This model is
able to span the film drainage behaviour from rigid to mobile films.
If the simplification for the flat middle part of the film is pursued. In that case, the
simplified equations become
246
Modelling the Stabilising Behaviour of Silicone Surfactants…
Figure 5.11 Film shape at t = 16 for several values of S (shown in the upper right).
The film is much thinner for small values of S because the free surface is mobile. For
large values of S, the free surface becomes tangentially immobile.
⎡ K3 ⎤
KT + ⎢ KW ( s ) + ⎥ =0 (26)
⎣ 3 ⎦z
( s) K
Wzz + =0 (27)
S
Once again the capitalised variables are rescaled versions of those in Equations (25-29)
[60]. These are subject to the boundary conditions:
247
Advances in Urethane Science and Technology
Figure 5.12 w(s) with t = 16 and kz(L,t) - 10 varying S shown in the upper right. The
surface velocity decreases as S increases
K = W ( s ) = 0, at Z = 0 (28)
and a consistent W(s)(Z,0) is found from its equation. The consistent value of W(S)(Z,0)
is required for the numerical solution of the problem [60].
248
Modelling the Stabilising Behaviour of Silicone Surfactants…
By studying these equations for large and small S, the slow and fast draining limits of the
experiment have been obtained using analytical expressions [60]. For large S, solutions
of the following form are found:
f1 (Z) f (Z)
K≈ 1 /2
, and W ( s ) ≈ 2 1 / 2 (31)
T T
where f1(Z) and f2(Z) satisfy ordinary differential equations. The solutions approximate
computed solutions for the rigid film case very well. In this way the rigid case for large
surface viscosity within lubrication theory is analytically recovered. For small S, we can
find solutions of the form
f3 (Z) f (Z)
K≈ , and W ( s ) ≈ 4 (32)
T T
where f3(Z) and f4(Z) also satisfy ordinary differential equations. In this way, the fast
draining limit of the experiments, where the measured exponents for the fastest thinning
case are -0.92, is analytically recovered.
Surfactant transport will now be included in the model, in an effort to obtain intermediate
thinning rates as observed experimentally.
Surfactant transport is also important in the vertical draining film; surfactant concentration
gradients may develop, which may in turn strongly affect the fluid flow via the Marangoni
effect. When surfactant transport is considered, lubrication theory then gives three
nonlinear partial differential equations for the free surface shape k(z,t), the surface velocity
w(S)(z,t) and the surface concentration of surfactant Γ(z,t). The mathematical problem
to be solved for these dependent variables is:
⎡ k3 ⎤
k2 + ⎢ kw( s ) + (1 + κ z ) ⎥ = 0 (33)
⎣ 3 ⎦z
k M
wzz( s ) + (1 + κ z ) + 2 Γz = 0 (34)
NS N S
249
Advances in Urethane Science and Technology
⎡ 1 ⎤
Γt + N 2 ⎢Γw ( s ) − N 2 Γz ⎥ = 0 (35)
⎣ P ⎦z
where L is the lower end of the computational domain; C is specified and a consistent s1
can be calculated with the first integral of the equilibrium meniscus shape [62, 63], as
discussed previously.
Some normalisation factors in the surface transport equation for G that ordinarily would
become unity in lubrication theory at leading order have also been retained. While we
do not rigorously apply matched asymptotic expansions in this work, the equations
contain all of the terms necessary to match the film onto a static meniscus (for the bath
surface) and it has been shown that the terms neglected are small for δ<<1 [65].
Marangoni effects have a substantial impact for smaller values of S, retarding surface
drainage and enhancing film thickness; Figure 5.13 shows free surface shapes for several
M with S = P = 102 and δ = 0.1. For a small M, the Marangoni effect is too weak to retard
drainage and the film thins rapidly. As M increases, the Marangoni effect become more
prominent, retarding drainage and the film is considerably thicker at the same time in
the computation. As M → ∞, w(S) → 0 all along the film and the surface becomes rigid.
A rigid fluid surface associated with a large Marangoni number was observed in a levelling
problem [69] and in a similar geometry to this work.
Interesting free surface shape features are shown in Figure 5.14, for M = -10, S = 102
with varying t. There are parts of the free surface shape that appear to be draining
rapidly, as if they are mobile, and other parts that drain slowly, as if they were rigid-
surfaced. Concave-out profiles above an upward-propagating wave indicate a mobile
surface while an essentially downward-parabolic profile following the wave indicates a
rigid surface. These assertions are supported by the thinning rates of t-1 above the wave
and t-1/2 below it. The wave that separates the two regions is driven by surface tension
gradients that are shown in Figure 5.15. The localised gradient of the surfactant
concentration on the surface is sufficient to drag fluid up the film; the gradient is preserved
250
Modelling the Stabilising Behaviour of Silicone Surfactants…
Figure 5.13 Film shape at t = 16 with S = 102 fixed for several values of M (shown
in the upper right). The film is much thinner for small values of M. For large values
of M, Marangoni effects are stronger, drainage is retarded and the free surface
becomes tangentially immobile.
because the Péclet number is large, and so it is not smeared out by diffusion. The associated
surface velocity profiles are shown in Figure 5.16; the negative surface velocity clearly
shows that fluid is moving up the film on its surface, as a result of the Marangoni effect.
The wave seen in Figure 5.14 propagates up the entire length of the film. Similar behaviour
is seen in the experiments of Snow and co-workers [70, 71] with PU films, where what
appears to be bumps propagate up the edges of the film. On that basis, it is believed that
the wave observed may be related to those structures observed in the three-dimensional
films. Work is currently being done to verify this conjecture.
251
Advances in Urethane Science and Technology
Figure 5.14 Film shape with S = 102 and M = -10 fixed for several values of time t
(shown in the upper right). A wave driven by surface tension gradients propagates up
the film. Thinning proportional to t-1 and concave-out profiles above the propagating
wave indicate a mobile surface; below the wave, t1/2 thinning and the downward
parabolic shape indicate a rigid surface.
If S < 102, it is found that the film drainage is typically too rapid to allow for much effect
from surface tension gradients (and thus Marangoni stresses). For S > 104, we find that
the films behave, in large part, as if they were rigid [70]. It is clear that the maximum
impact of the Marangoni effect occurs at an intermediate value of the surface viscosity.
These results lend support to the assertion that the greatest foam stability occurs when
the Marangoni and surface viscous effects are at intermediate value, as given in the first
section of this chapter.
252
Modelling the Stabilising Behaviour of Silicone Surfactants…
Figure 5.15 Surfactant distribution with S = 102 and M = -10 fixed for several values
of time t (shown in the upper right). The surface tension gradients driving the wave up
the film are easily seen.
253
Advances in Urethane Science and Technology
Figure 5.16 Surface velocities with S = 102 and M = -10 fixed for several values of
time t (shown in the upper right). Negative values indicate flow up the film, and flow
up the film is caused by the surfactant concentration gradients.
5.5 Summary
The stabilising behaviour of silicone surfactants during the formation of PU foam can be
investigated in detail, both experimentally and theoretically by the use of a foam model
system, specifically vertical, thin liquid PU films. These films were investigated by a
combination of direct observation and interferometric measurements. The formation and
temporal evolution of structures on the surface of the film, including fingering patterns, was
a result of the Marangoni effect. Experimentally, these films demonstrate a wedge shape and
the gravity-driven drainage of these films causes a collapsing inwards of the wedge.
The evolution of structures and the drainage of these films were functions of the
concentration and molecular structures of the surfactants. The two film physical
254
Modelling the Stabilising Behaviour of Silicone Surfactants…
parameters correlated to film structures, film drainage and the properties of the surfactant
are: (1) the strength and lifetime of surface tension gradients, (2) the surface viscosity.
Two extremes of film behaviour were identified. Firstly, a rigid-surfaced film where the
rate of drainage and structural evolution were low. Secondly, a mobile-surfaced film
where the rate of drainage and structural evolution were high. Intermediate cases were
also detected and measured.
The theoretical and numerical model for vertical liquid film drainage that has been
developed reproduces a number of features described in these experiments. These features
include film shapes and thinning rates.
Future work will include adding intermolecular forces that act across the film and
extending our work to three-dimensions in order to make closer comparison with
experiment. Furthermore we endeavor to obtain accurate independent measurements of
surface viscosity and correlate them to both film properties and surfactant parameters.
Acknowledgements
RJB and SN gratefully acknowledge support from Dow Corning Corporation for this work.
RJB is also grateful for the support of the National Science Foundation via grants DMS-
9623092 and DMS-9631287.
SAS and UCP would like to thank Ben Vesper, Mike Owen, Andy Goodwin, John Frey,
Randy Hill, Dave Battice, Mike Stanga, Kathy Goudie, Joan Sudbury-Holtshlag, Mike
Reuter, and Gloria Lyon for many helpful discussions and their support.
References
255
Advances in Urethane Science and Technology
12. M. J. Owen and T. C. Kendrick, Journal of Colloid and Interface Science, 1968,
27, 1, 46.
15. J. Ritter, J. E. Dubjaga and A. B. Komarowa, Plaste und Kautschuk 1979, 26, 11,
624.
18. H. Hamann and J. Ritter, Plaste und Kautschuk, 1983, 30, 7, 364.
256
Modelling the Stabilising Behaviour of Silicone Surfactants…
23. D. F. Evans and H. Wennertröm, The Colloidal Domain, VCH Publishers, New
York, 1994, p.55.
27. S. A. Snow, W. N. Fenton and M. J. Owen, Journal of Cellular Plastics, 1990, 26,
2, 172.
28. J. S. Clunie, J. F. Goodman and B. T Ingram, Surface and Colloid Science, 1971,
3, 167.
29. I. B. Ivanov, in Thin Liquid Films, Surfactant Science Series No.29, Marcel
Dekker Inc., New York, 1988.
30. R. J. Pugh, Advances in Colloid and Interface Science, 1996, 64, 0, 67.
31. L. D. Artavia, A Model for Low Density Foams, University of Minnesota, USA,
1991, [Ph.D Thesis].
257
Advances in Urethane Science and Technology
36. I. B. Ivanov, in Thin Liquid Films, Surfactant Science Series No.29, Marcel
Dekker Inc., New York, 1988, p.419.
42. J. B. M. Hudales and H. N. Stein, Journal of Colloid and Interface Science, 1991,
138, 2, 354.
44. H. N. Stein, Advances in Colloid and Interface Science, 1991, 34, 175.
46. J. B. M. Hudales and H. N. Stein, Journal of Colloid and Interface Science, 1990,
137, 2, 512.
47. K. J. Stebe and C. Maldarelli, Journal of Colloid and Interface Science, 1994,
163, 1, 177.
258
Modelling the Stabilising Behaviour of Silicone Surfactants…
49. Principles of Colloid and Surface Chemistry, 2nd Edition, Ed., P. C. Heimenz,
Marcel Dekker, Inc., New York, 1986, p.65 and p.83.
51. U. C. Pernisz and S. A. Snow, The Drainage of Liquid Polyurethane Films, Dow
Corning Corporation Research Report, 1999, 1999-I000-46182, Midland, MI.
59. S. Naire, R. J. Braun and S. A. Snow, Journal of Colloid and Interface Science,
2000, 230, 1, 91 .
61. A. Oron, S. H. Davis and G. B. Bankoff, Reviews of Modern Physics, 1997, 69,
931.
64. J. A. Moriarty, L.W. Schwartz and E.O. Tuck, Physics of Fluids A, 1991, 3, 5,
733.
259
Advances in Urethane Science and Technology
66. L. W. Schwartz and R. V. Roy, Journal of Colloid and Interface Science, 1999,
218, 309.
68. S. A. Snow, B. M. Nugent and U.C. Pernisz, Dow Corning Corporation Research
Report, 1996, 1996-I0000-42699.
70. S. A. Snow, U. C. Pernisz and R.E. Stevens, Dow Corning Corporation Research
Report 1995, 1995-I0000-40859.
260
6
Synthesis and Characterisation of Aqueous
Hybrid Polyurethane-Urea-Acrylic/Styrene
Polymer Dispersions
Janusz Kozakiewicz, Anita Koncka-Foland, Jan Skarzynski
and Izabella Legocka
6.1 Preface
6.2 Introduction
There are strict ecological limitations relating to allowable volatile organic compounds
(VOC) content in coatings introduced in many regions and countries, e.g., the California
Rule of 1995 [1] or special provisions in the EU, and which force the manufacturers to
search for new, more environment-friendly formulations for particular products which
usually requires replacing of the substantial binder (polymer or resin). Undoubtedly the
products for which the development has been the most significant in the field of coatings
over the last few decades are aqueous dispersion-based and powder coatings. In the
former group, DPUR, which are presently being widely used as binders for coating
formulations have already taken a leading position [2-5] and compete successfully with
acrylic dispersions despite the distinctly higher market prices. Both blends of DPUR and
acrylic dispersions [6], and (recently) hybrid acrylic-urethane dispersions [7-12], obtained
by synthesis, and not simple blending are now available. The latter products are much
261
Advances in Urethane Science and Technology
more interesting from both a scientific and practical point of view since the synthesis of
a hybrid system may lead to new polymer architectures which may result in considerable
improvement of properties. Especially attractive is the possibility of synthesising
interpenetrating polymer network (IPN) systems based on polyurethanes and acrylic
polymers using this principle since it has been shown that in such systems synergistic
effects may occur [12], i.e., the properties of a hybrid may be better than those of a blend
of its components. At the same time, hybrid dispersion particles usually have quite complex
morphology (in most cases a ‘core-shell’ type) which affects properties of dispersions
and films and coatings made of them [13].
If a comparison were made between hybrid dispersion systems taking into account the
kind of dispersion particles they contain, it would appear that the best properties should
generally be observed for systems obtained by synthesis for which the particles have a
uniform structure (they may be called ‘true hybrid’ particles) and the worst – for simple
blends of the two dispersions [14] (see Figure 6.1).
Obviously, both types of dispersion particle morphology presented in Figure 6.1 (‘core-
shell’ and ‘true hybrid’) should be considered as idealised cases. In practice, a variety of
different particle morphologies may be observed. Three of them are shown in Figure 6.2.
The most important factor that affects the morphology of hybrid dispersion particles
seems to be the method of preparation of such hybrids.
262
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
a) Blend of dispersions
⇓
no synergistic effect
• poor stability
• matte and hazy films and coatings
• always two Tgs
• poor mechanical properties of films
b) ‘Core-shell’ c) ‘True-hybrid’
⇓ ⇓
synergistic effect possible synergistic effect probable
• good stability • excellent stability
• usually matte and hazy films and • usually glossy and transparent films
coatings and coatings
• usually two Tgs • usually one Tg
• good mechanical properties of films • excellent mechanical properties of films
Figure 6.1 Comparison of hybrid dispersion systems obtained by (a) blending, (b) and
(c) by synthesis, taking into account particle type.
263
Advances in Urethane Science and Technology
In practice, methods (2) and (3) seem to be the most commonly used. Recently, yet
another method of preparation of hybrid polymer dispersions, called an ‘active blending’
has been reported [18]. The components of such ‘active blends’ or ‘active mixtures’
remain inactive when in the presence of water, and cross-react during film formation.
The other factors that may influence hybrid dispersion particle morphology will be
discussed in detail in Section 6.3.2 of this Chapter.
Aqueous acrylic dispersions have been used as binders for coatings for over 60 years.
Their main advantages include: low cost and excellent tolerance of additives or pigments;
however, they have several drawbacks such as the presence of relatively large amounts of
water-sensitive additives (protective colloids, emulsifiers, etc.) and the need to use
coalescing agents when a continuous film is desired.
Aqueous DPUR appeared on the market much later (in the 1960s), and after only a couple of
years achieved the quality that allowed them to be widely applied as coating binders. Substantial
advantages of DPUR include excellent mechanical properties and solvent resistance. Their
drawbacks are relatively high cost and poor tolerance of additives and pigments.
264
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
(a)
(b)
- polyol chain
- hydrophilic moiety
- hard segment (urethane or urea)
Methods of preparation of aqueous acrylic dispersions are well known [19] and will
not be discussed here. Emulsion polymerisation of acrylic monomers is usually carried
out using a water-soluble initiator or a ‘redox’ system. In this process, relatively high
levels (up to several %) of emulsifiers and protective colloids are used. The dispersed
polymer generally has a linear structure and a high molecular weight (MW). The
properties of dispersions and films are mostly affected by the kind of monomer(s)
used [20].
DPUR can be prepared in many ways [21, 22], which eventually lead to one of the structures
(polyurethane or polyurethane-urea) shown in Figure 6.3.
In practice, the structures shown in Figure 6.3 are partly or fully crosslinked. The two
main methods which lead to these structures are the ‘acetone process’ (historically older)
[23] and the ‘prepolymer-ionomer process’ [24] that is commonly used to manufacture
DPUR at present. A combination of these two methods can also be used.
In the ‘acetone process’, the first step is the synthesis of NCO-terminated urethane
prepolymer which is then dissolved in acetone and reacted with a di-hydroxy or di-
amino compound containing groups which are the precursors of ionic groups, e.g., -SO3,
-COOH, N≡. The next steps are the emulsification of the resulting product in water
combined with the formation of ionic groups and distillation of acetone.
265
Advances in Urethane Science and Technology
In the ‘prepolymer-ionomer process’, the compound which contains the moieties which
are the precursors of ionic groups is incorporated in the polymer chain already at the
stage of urethane prepolymer formation. Then they are neutralised, which results in the
formation of a ‘prepolymer-ionomer’ which is emulsified in water and eventually
crosslinked. In this process, the prepolymer-ionomer is usually dissolved in a small amount
of water-miscible solvent of high boiling point, e.g., N-methylpyrolidone, which plays
the role of coalescing agent in the process of film formation. It is then possible to obtain
DPUR which contain either ‘cationic DPUR’ with a pH of less than 7 (cationic moieties
are attached to the polyurethane or polyurethane-urea chain) or ‘anionic DPUR’ with a
pH of greater than 7 (anionic moieties are attached to the polyurethane or polyurethane-
urea chain). If non ionic hydrophilic moieties are attached to or incorporated in the
polyurethane or polyurethane-urea chain, then a ‘nonionic DPUR’ may be obtained.
Properties of DPUR depend on several factors. The most important are structure of the
polyurethane or polyurethane-urea chain, degree of crosslinking and the amount of
coalescing agent [25].
The basic process was first patented by Inmont, USA, in 1982 [9]. They simply claimed
polymerisation of vinyl polymers in DPUR. In 1986, a patent for another method of
preparation of hybrid acrylic-urethane dispersions (diluting of prepolymer-ionomer with
monomers, emulsifying the resulting solution in water and polymerisation) was granted
to Witco, USA [26].
In subsequent years the number of patents dealing with hybrid acrylic-urethane dispersions
increased rapidly, and approximately 30 patents have been granted so far worldwide,
the majority in Japan.
Despite the large number of patents, at the time acrylic-urethane hybrid dispersions entered
the market (in about 1990), their properties, and specifically their structure/properties
relationships and particle morphology, had not been described in the scientific literature
until 1992, when Gruber from Air Products presented a paper at the ACS Meeting in San
Francisco [27] that dealt with the general characterisation of such dispersions. In 1996,
Hegedus and Kloiber (also from Air Products) published an important paper [28] where
they presented the properties not only of dispersions but also of coatings made on their
basis. However, there was no information on composition and chemical structure of products
266
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
In one of these articles [30] the properties of this kind of dispersion were compared with
the properties of its analogue crosslinked additionally by oxidation of double bonds
present in polyurethane-urea chain (air-drying) and it was concluded that the former
resulted in higher coating hardness. The authors of both publications presented their
schematic drawings showing the ‘core-shell’ morphology of their dispersion particles,
but did not reveal any experimental evidence that would confirm such morphology.
Another interesting paper dealing with hybrid vinyl/acrylic-urethane dispersions was published
in 1996 by Cytec [31]. They prepared hybrid dispersions by swelling DPUR particles with
monomers at a polyurethane:vinyl component ratio of 3:1 (w/w) and then carrying out
emulsion polymerisation. The resulting hybrid dispersions were additionally crosslinked by
the reaction of epoxy groups (when glycydyl methacrylate was used as one of the monomers)
with carboxylic groups attached to the polyurethane-urea chain or by using TMI (unsaturated
aliphatic isocyanate; Cytec industries, USA) as one of the monomers. The authors suggested
that IPN structures were formed in these reactions. The paper contains very good photographs
of dispersion particles, but no information on particle morphology is given.
The effect of type of diisocyanate and polyol used in synthesis of hybrid PU/acrylic polymer
dispersions was recently reported by authors from Lamberti [32].
A recently published patent [33] from Zeneca described in detail the methods of
preparation of hybrid polyurethane/vinyl polymer dispersions but neither discussed the
effect of various factors on the properties of the hybrid dispersions nor provided any
information on their particle morphology.
Based on this brief literature survey, it seems quite clear that not all features of the hybrid
acrylic-urethane dispersions have been studied so far. The purpose of the present study is
to clarify the effect of various factors on the properties of acrylic/styrene-urethane
dispersions. Some results of preliminary investigations of the particle morphology will
also be revealed.
267
Advances in Urethane Science and Technology
It was decided that synthesis of DPUR would be carried out according to the ‘prepolymer-
ionomer’ method generally described in Section 6.2. The scheme of the reactions leading
to DPUR is presented in Figure 6.4.
Step I - polyaddition
Step II - neutralisation
Figure 6.4 Schematic representation of the reactions leading to DPUR using the
‘prepolymer-ionomer process’
For the synthesis of DPUR, a cycloaliphatic diisocyanate, isophorone diisocyanate (IPDI) from
Hüls, was used since unlike aromatic isocyanates, it does not result in yellowing of the coatings
and films. The polyols used were polytetramethylene glycol (PTMG) with a MW of 2000
268
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
supplied by BASF and commercially available polyesterdiols commonly used in coatings (Bester
190 and 195, supplied by Polychimica) with hydroxyl numbers equal to 32 and 122, respectively.
Polyesterdiol containing double bonds (Irpurate 240Tz from InterRokita, Poland, with hydroxyl
numbers equal to 276) was also used in some experiments. Both polyether- and polyester-diols
were used to look at the effect of structure of the polyurethane-urea chain on the properties of
the dispersions, films and coatings. N-methylpyrrolidone (NMP) from Sidaco was used as co-
solvent (coalescing agent) and 2,2´-dimethylolpropionic acid (DMPA) from Angus was used as
the source of pendant carboxylic groups in the polyurethane-urea chain.
Polymerisation of acrylic monomers was carried out according to the standard procedure
using two different initiating systems:
(b) ‘redox’ system – cumene hydroperoxide (CHP) + Rongalit with the structure:
Rongalit
The choice of initiating systems was made by considering their possible effect on particle
morphology (see section 6.3.2). The other additives (emulsifiers, protective colloids) were
standard materials commonly used in emulsion polymerisation of acrylic monomers.
It was assumed that three monomers would be used in this study: methyl methacrylate,
butyl acrylate and styrene. The possiblity of using styrene seemed to be especially interesting
because of its very low cost compared to the other monomers. The other two monomers
were selected for the significant differences in their polarity and solubility in water that
might affect the particle morphology and the Tg of poly(methyl methacrylate) and poly(butyl
acrylate). Difunctional acrylate monomer (1,4-butanediol dimethacrylate) was also used
in some experiments since it encouraged the formation of an IPN structure.
Based on the literature survey (see Section 6.2) and the author’s preliminary experiments
the following methods of synthesis of hybrid dispersions (after this point in the chapter
they will be called modified DPUR (MDPUR)) were selected:
269
Advances in Urethane Science and Technology
It was assumed that the method of MDPUR synthesis would have be the major factor
that would determine particle morphology. In Figures 6.5-6.7, the methods (1) – (3)
applied in this study are presented schematically.
270
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
Figure 6.6 Method 2 (using the monomers as active diluents for prepolymer-ionomer)
271
Advances in Urethane Science and Technology
Before the polymerisation starts, the system will consist of the following components:
(a) DPUR particles that will be only slightly swelled with monomers (in Method 1a –
continuous feeding of the monomers) or fully swelled with monomers (in Method 1b
– swelling of DPUR particles before polymerisation)
~ ~~~
~~~ ~~
~~~~~~~~
~~ ~ ~
272
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
When the process is carried out according to Method 1a it may be assumed that the kind
of initiating system (water-soluble initiator or redox system) does not have any significant
effect on the morphology of hybrid dispersion particles since the polymerisation proceeds
mainly on the surface of DPUR particles and monomer droplets and in the micelles of
emulsifier. In the case of water-soluble initiator (potassium persulphate) the polymerisation
starts in water where the ion-radicals are formed from the monomer molecules dissolved
in water. These ion-radicals can diffuse to DPUR particles, monomer droplets and
emulsifier micelles. After this happens the system will consist of:
(a´) DPUR particles partly swelled with monomer, with adsorbed ion-radicals
~ ~~~~
~ ~~ ~~ ~
~ ~~~ ~
~~ ~ ~
(b´) monomer droplets of the size much smaller than before polymerisation and with the
adsorbed ion-radicals
~ ~
273
Advances in Urethane Science and Technology
After polymerisation is completed both acrylic/styrene polymer particles (a) and the hybrid
particles (b1 and b2) may be formed:
(a)
Hybrid particles of uniform structure (‘true hybrid’) (b1) or hybrid particles of ‘core-
shell’ structure where the core is composed of polyurethane-urea (b2):
(b1) (b2)
Obviously, other morphologies of dispersed particles are also possible (see Section 6.2).
The factor which probably has quite a significant effect on formation of separate particles
of the acrylic/styrene polymer is concentration of initiator. The higher it is, the higher is
the probability that polymerisation would proceed in monomer droplets. If it is relatively
low, this chance is diminished since monomer droplets diffuse into DPUR particles.
Probability of formation of hybrid particles of more uniform structure increases with
increase in solubility of monomer in water since longer-lived ion-radicals may be formed
and their adsorption on DPUR particles is easier. The second important factor affecting
the hybrid particle morphology is solubility of monomer in polyurethane-urea which
constitutes the particles of DPUR. If it is higher, swelling of DPUR particles with monomer
is easier and the probability of formation of hybrid particles of more uniform structure is
higher.
When polymerisation is carried out after initial swelling of DPUR particles in monomers
(Method 1b), then formation of separate particles of acrylic/styrene polymer is also possible,
but much less probable since the amount of monomer present in the system as monomer
droplets is very low in this case (assuming that the amount of monomer used to swell
DPUR particles was not higher than the maximum amount resulting from the equilibrium
swell curve (see Section 6.5.4)). It can be anticipated that in this method of MDPUR synthesis
many more hybrid particles of much more uniform structure will be formed than when the
monomers are continuously added to the reaction system (Method 1a), especially when a
274
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
If this happens, then gelling of the whole reacting system becomes very probable.
In this case it can be assumed that polymerisation will proceed in a similar way as described
previously for Method 1b. The only difference may be that both structure and size of
DPUR particles formed from the prepolymer-ionomer diluted with monomers will differ
from the ‘standard’ case, i.e., when NMP is used as diluent for prepolymer-ionomer.
Moreover, some fraction of the monomer molecules may be so closely ‘trapped’ in the
crosslinked polyurethane-urea structure in DPUR particles that their participation in
polymerisation will be difficult. This may lead to a decrease in average MW of acrylic/
styrene polymer which will be formed inside DPUR particles.
In this case, the course of MDPUR synthesis seems to be much simpler than in Methods
1 and 2 because no polymerisation of monomers will proceed in the system. Before the
process starts, only ASD particles are present in the system:
275
Advances in Urethane Science and Technology
It may be assumed that the factors that would decide the course of the process and the
fraction of hybrid particles would be:
6.4.1 Dispersions
276
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
Chromatopac apparatus. Free monomer content in the dispersions was measured by gas
chromatography (GC) (Giede 18.3 apparatus)
The maximum degree of swelling of DPUR particles with monomers was determined
using the equilibrium swelling method described in Section 6.4.3.
6.4.2 Coatings
Coatings were made from dispersions by applying them onto degreased glass plates with
a 120 μm gap applicator. The resulting coating was dried at room temperature for 72 h
(for hardness and adhesion tests) and at 40 °C (for water resistance test).
Water resistance was determined by examination of the coating after 24 hours immersion.
It was designated by numbers (1-4) and letters (A-D) where the higher number means a
higher number of bubbles and the ‘higher’ letter means a higher dissipation of bubbles
on the plate [35]. Adhesion was determined by a ‘double-cutting knife’ method and
designated by numbers (1-4). Higher numbers mean lower adhesion. For determination
of hardness the pendulum method (Persoz) was applied [36].
The drying time of the coatings was determined according to the relevant Polish standard
(corresponding to DIN 53150 [37]). Times needed to achieve 1° of hardness (touch dry)
and 3° of hardness (hard dry) were measured.
277
Advances in Urethane Science and Technology
6.4.3 Films
Films were obtained from dispersions by casting them onto glass Petri dishes and allowing
them to dry for 14 days at room temperature. Mechanical properties, i.e., tensile strength
(dr), stress at 100% elongation (d100), elongation at break (er) and Shore A hardness were
determined according to ISO 527 [38] (strip type specimens were tested).
Percentage swell in water, methylethylketone (MEK) and xylene was determined for
square specimens (25 x 25 mm) after 2 weeks of immersion at room temperature and
calculated from the following equation:
3
⎛ L⎞
% swelling = 100 ⋅ ⎜ ⎟ − 100
⎝ 25 ⎠
Surface free energy was calculated from experimental values of dynamic contact angle
determined using a tensiometer constructed in the Institute of Chemical Physics of the
Russian Academy of Sciences.
The degree of crosslinking was calculated from the maximum degree of swelling determined
according to the following procedure: samples of film weighing approximately 0.1 g each
were immersed in toluene and the degree of swelling was calculated from the equation:
m(τ) − m0
α(τ) =
m0
After the maximum degree of swelling was reached (determined graphically), the sample
was weighed and dried to constant solids content (mc). Content of the soluble fraction of
the sample (S, %) was calculated according to the equation:
m0 − mc
S=
m0
278
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
1
j=
S+ S
Table 6.1 shows the properties of DPUR used for synthesis of MDPUR (MDPUR-ASD)
together with the properties of prepolymer-ionomers used as intermediates in their
preparation.
Figures 6.8 and 6.9 show the MW distribution for standard prepolymer-ionomers used
in DPUR synthesis, obtained from polyetherdiol (MW = 2000) and a mixture of
polyesterdiols (MW = 920 and 3190), respectively. The weight average MW (Mw) and
number average MW (MN) of these prepolymer-ionomers were (not counting free IPDI):
MN MW MW/MN
Polyetherdiol-based 6083 13604 2,24
Polyesterdiol-based 6619 16010 2,41
MN MW MW/MN
Polyetherdiol-based 5215 11153 2,1
Polyesterdiol-based 6999 16811 2,40
For DPUR the MW distribution could not be determined by GPC as the films did not
dissolve completely.
279
Table 6.1 Properties of DPUR used as starting materials for MDPUR-ASD obtained according to methods 1a and 1b
280
(see Section 6.3.2) and of the prepolymer-ionomers (diluted with NMP or acetone) used in their preparation. Properties of
prepolymer-ionomers diluted with monomers (designated as PPUR) used as intermediates in synthesis of MDPUR-ASD
according to method 2 have also been included in this table.
Properties of prepolymer-
Properties of DPUR
ionomers
Designationa NMP
Average Zeta
of DPUR Polyol Content
NCO, % Mechanical Particle Particle potential MFFT
(PPUR) % Viscosity Solids Viscosity
pH Stability, size size mV °C
mPa-s % mPa-s
minutes distribution
Theor. Pract. nm nm
DPUR 258b PTMG-2000 15.3 2.75 2.15 4,720 36.1 315 7.0-7.3 >90 120-420 161.8 -21.2 <0
PTMG-2000
PPUR 289 3 3.66 3.96 - - - - - - - - -
and Irpurate
DPUR 345 PTMG-2000 17.8 2.74 2.32 38 0 35.9 200 7.6-7.9 >9 0 110-1,200 211.1 -7.9 <0
DPUR 346 PTMG-2000 3 3.24 3.04 7,500 31.5 20 7.6 60 75-770 215.5 -4.5 <0
Bester 190,
DPUR 354 0 4.32 4,72 6,300 39.0 108 7.1-7.4 >90 30-140 56.2 -60.6 <0
195 Irpurate
PTMG-2000
DPUR 356 4 4.53 4,23 3,800 28.4 15.5 7.1-7.4 >9 0 30-850 75.1 -98.8 <0
and Irpurate
Particle size distribution and zeta potential determined for one of the DPUR from Table
6.1 are shown in Figures 6.10 and 6.11, respectively. TEM photographs of the particles
of the same dispersion are shown in Figure 6.12.
Properties of films and coatings made from DPUR included in Table 6.1 are shown in
Tables 6.2 and 6.3, respectively.
DSC thermogram for a typical DPUR from Table 6.1 (DPUR-237) is presented in Figure 6.13.
281
Advances in Urethane Science and Technology
Figure 6.10 Particle size distribution for typical DPUR synthesised in this study (DPUR
268 = DPUR 237 from Table 6.1)
Figure 6.11 Zeta potential for typical DPUR synthesised in this study
(DPUR 268 = DPUR 237 from Table 6.1)
282
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-urea…
Table 6.2 Properties of films from DPUR used as starting materials for synthesis
of MDPUR-ASD according to methods 1a and 1b.
Properties of films
Designation
δ 100 δr εr % swell Tg
of DPUR Appear.
MPa MPa % H2O MEK Xylene °C
DPUR 237 T 3.2 19.0 680 Wh, Pieces Pieces -79.1
(240, 244, 332) Pieces
DPUR 345 T 7.05 7.9 130 Nt1 174.4 130 -71.6
Wh Br Br
DPUR 346 T, OrP Nt1 Nt1 Nt1 Nt1 Nt1 Nt1 -71.1
DPUR 354 T, - - - - - - -15.7
Cracks
DPUR 356 T, El - 16.3 27 12 199 109.7 <0
T, Br T, Br
T - transparent, El - elastic, Br - brittle, Wh - white, OrP - orange peel, Nt1 - not tested because of too
high adhesion to substrate
283
Advances in Urethane Science and Technology
Table 6.3 Properties of coatings from DPUR used as starting materials for
synthesis of MDPUR-ASD according to methods 1a and 1b.
Properties of coatings
Designation
Drying time, min Hardness Water
of DPUR Adhesion
Hardness (1°) Hardness (3°) (Persoz) resistance
DPUR 237 10 15 0.12 4 No bubbles
(240, 244)
DPUR 345 180 340 0.164 1 1D
DPUR 346 25 50 0.178 3 Nt2
DPUR 354 25 25 0.25 None -
DPUR 356 40 60 0.36 4 1C, does not
whiten
Nt2 - not tested because of uneven surface
See section 6.4.2 for an explanation of water resistance codes
Figure 6.13 DSC thermogram for the film obtained from a typical DPUR synthesised
in this study (DPUR 237 from Table 6.1).
284
Table 6.4 Properties of ASD used as starting materials for synthesis of MDPUR according to method 3
Properties of ASD
Design Mech Free Particle Average Zeta
Monomer Initiator Solids, Viscosity MFFT Tg
of ASD pH stability monom., size distr part size potential
% mPa-s °C °C
min % nm nm mV
ASD 15 BA/MM/S K2S2O8 24.2 20 6.2 45 0.04/0.1/ 40-145 39.7 -12.9 >+50 +58. 0
(1/1.32/ 0.1
1.64)
ASD 17 MM K2S2O8 24.9 14 4.0-5.0 45 1.00 35-85 54.0 -43.4 >+50 +123.1
ASD 17 BA K2S2O8 24.9 6 4.0-5.0 45 0.53 30-140 59.7 -41. 3 -88.0
ASD 20 S K2S2O8 23.3 5 5.0-5.3 45 0.24 40-180 79.5 -55.5 >+50 +98.0
ASD 264 BA/MM/S CHP/ 22.5 5 6.2-6.5 >90 Nt Nt Nt Nt >+50 nt
(1/1.32/ Rongalit
1.64)
ASD 266 As above K2S2O8 21.4 6 6.2-6.5 >90 Nt Nt Nt Nt >+50 nt
BA: butyl acrylate
S: styrene
MM: methyl methacrylate
Nt: not tested
285
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
Advances in Urethane Science and Technology
In Figures 6.14, 6.15 and 6.16 the particle size distribution, zeta potential and TEM
micrograph of dispersion particles of typical ASD from Table 6.3 (ASD sample number
15) are presented.
Figure 6.14 Particle size distribution for typical ASD synthesised in this study (ASD 15
from Table 6.3).
286
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
Figure 6.15 Zeta potential for typical ASD synthesised in this study
(ASD 15 from Table 6.3).
Figure 6.16 TEM micrograph of dispersion particles of ASD (see Table 6.3).
Reproduced with permission from Professor I. A. Grickova at Lomonosov
University, Moscow.
287
Advances in Urethane Science and Technology
Tables 6.5, 6.6, 6.7 and 6.8 show the compositions of MDPUR-ASD prepared according
to methods 1a, 1b, 2 and 3, respectively.
288
Table 6.7 Composition of MDPUR-ASD prepared according to method 2 (diluting of the prepolymer-ionomer
in monomers, polymerisation and crosslinking of prepolymer-ionomer). Properties of starting prepolymer-
ionomers are also included.
Designation of Coalescent Properties of starting prepolymer-ionomer
starting (NMP)
Designation of
prepolymer- Polyol Monomer content in NCO, % Viscosity at
MDPUR-ASD
ionomer MDPUR-ASD 25 °C, mPa-s
(PPUR) % Theoret. Pract.
MDPUR-ASD PPUR 288 PTMG-2000 BA,MM,S 3 2.50 2.68 760
97
MDPUR-ASD PPUR 288 PTMG-2000 BA,MM,S 3 2.50 2.68 760
98
MDPUR-ASD PPUR 289 PTMG-2000 BA,MM,S 3 3.66 3.96 730
102
MDPUR-ASD PPUR 333 PTMG-2000 BA,MM,S - 2.75 2.48 Nt
198 Irpurate
Nt – not tested
Table 6.8 Compositions of MDPUR prepared according to method 3 (synthesis of DPUR in ASD).
Properties of starting prepolymer-ionomers are also included
Coalescent Properties of starting prepolymer-ionomer
Designation of Polyol Designation of
(NMP) Content NCO, % Viscosity at
MDPUR starting ASD
in MDPUR, % theoret. Pract. 25 °C, mPa-s
MDPUR 243 PTMG-2000 11.6 ASD 15 2.74 2.64 4000
MDPUR 245 PTMG-2000 11.6 ASD 17 2.74 2.34 4500
MDPUR 247 PTMG-2000 11.6 ASD 18 2.74 2.24 4200
MDPUR 250 PTMG-2000 11.6 ASD 20 2.74 2.51 4000
MDPUR 347 PTMG-2000 3.6 ASD 264 3.12 2.85 5700
MDPUR 348 PTMG-2000 3.6 ASD 266 3.12 2.79 5000
289
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
Advances in Urethane Science and Technology
Particle size distribution and zeta potential of a typical hybrid dispersion obtained using
a water-soluble initiator are presented in Figures 6.17 and 6.18, respectively.
290
Table 6.9 Properties of hybrid dispersions of the same chemical composition prepared according to different
methods using a water-soluble initiator (K2S2O8)
Properties of dispersions
Method Designation
of of hybrid Mechanical Free Average
Viscosity Particle size Zeta pot MFFT
Synthesis dispersion Solids, % pH Stability monomer, part size
MPa-s distribution MV °C
Min BA/MM/S, % nm
MDPUR-
1a 30.7 9 7.1-7.4 15 Nt 30-230 61. 9 -18.4 <0
ASD 268
MDPUR-
1b 34.2 10 6.8-7.1 15 Nt 50-670 9 4. 4 -39.1 <0
ASD 270
MDPUR- 0.7/0.02/
2 31.4 10 6.8-7.1 15 50-120 108.4 -60.7 <0
ASD 97 0.015
MDPUR
3 41.6 200 7.4-7.7 15 Nt 55-520 115.4 -23.0 <0
348
Nt – not tested
Table 6.10 Properties of films made from hybrid dispersions of the same chemical composition prepared
according to different methods using water-soluble initiator (K2S2O8)
Method Designation Properties of films
of of Hybrid δ100 δr εr % swell
Synthesis Dispersion Appear. Tg °C
MPa MPa % H2O MEK Xylene
MDPUR-A-
1a T, El 8.2 10.8 164 13 Wh, El 347 T, Br 198 T, Br -73.4
SD 268
MDPUR-A-
1b T, El 9.2 13.3 180 0 Wh, El 347 T, Br 224 T, Br -73.4
SD 270
MDPUR-A-
2 T, El 6.5 10.0 210 13 Wh 341 T, G 310 T, G -56.6
SD 97
MDPUR
3 M, El 14.1 14.5 117 13 Wh, El 198 T, Br 198 T, Br -75.7
348
T- transparent, El - elastic, M - Mat, Wh - white, Br - brittle, G - gelatinous
291
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
Table 6.11 Properties of coatings made from hybrid dispersions of the same chemical composition prepared
292
according to different methods using a water-soluble initiator (K2S2O8)
Properties of coatings
Method of Designation of hybrid
Drying time, min Hardness
Synthesis Dispersion Adhesion Water resistance
1° 3° (Persoz)
1a MDPUR-ASD 268 30 50 0.22 4 1A whitening disappears after 1 h
1b MDPUR-ASD 270 25 45 0.20 4 2C whitening disappears after 1 h
2 MDPUR-ASD 97 35 60 0.12 4 2B whitening disappears after 1 h
3 MDPUR- 348 20 25 0.20 1 2A whitening disappears after 1 h
See section 6.4.2 for an explanation of water resistance codes
Table 6.12 Properties of hybrid dispersions of the same chemical composition prepared according to different
methods using a ‘redox’ initiator
Advances in Urethane Science and Technology
Properties of dispersions
Method Designation Free
of of hybrid Mechanical Average Zeta
Viscosity monomer, Particle size MFFT
Synthesis dispersion Solids % pH Stability part size potential
MPa-s BA/MM/S distribution °C
Min nm MV
%
MDPUR-
1a 29.7 7 6.8-7.1 15 Nt 35-310 68.0 -39.3 <0
ASD 274
MDPUR-
1b 29.3 10 7.1-7.4 15 Nt 85-650 90.4 -32.6 <0
ASD 272
<0.02/
MDPUR-
2 31.4 8 6.8-7.1 15 <0.008/ 105-165 130.6 -39.0 <0
ASD 98
<0.001
MDPUR
3 32.5 48 7.6-7.9 15 Nt 60-260 141.7 -25.4 <0
347
Nt – not tested
Table 6.13 Properties of films made from hybrid dispersions of the same chemical composition prepared
according to different methods using a ‘redox’ initiator
Properties of films
Method of Designation of Hybrid
δ100 δρ
δρ εr % swell Tg
Synthesis Dispersion Appear.
MPa MPa % H2O MEK Xylene °C
1a MDPUR-ASD 274 T, El 6.3 11.2 250 26 El 445 T, Br 374 T, Br -68.7
1b MDPUR-ASD 272 T, El 7.6 12.0 210 0, Wh, El 280 T, Br 280 T, Br -72.2
2 MDPUR-ASD 98 T, El 5.3 10.6 285 40 Wh, El 374 T, G 409 T, Br -70.6
3 MDPUR 347 M, el 12.5 15.0 187 0 Wh, El 130 T, Br 174 T, Br -77.2
T- transparent, El - elastic, M - Mat, Wh - white, Br - brittle, G - gelatinous
Table 6.14 Properties of coatings made from hybrid dispersions of the same chemical composition prepared
according to different methods using a ‘redox’ initiator.
Properties of coatings
Method of Designation of hybrid
Drying time, min Hardness
Synthesis dispersion Adhesion Water resistance
1° 3° (Persoz)
1a MDPUR-ASD 274 25 45 0.18 3 2D
whitening disappears after 1 h
1b MDPUR-ASD 272 25 50 0.16 4 3C
whitening disappears after 1 h
2 MDPUR-ASD 98 40 120 0.08 1 1A
whitening disappears after 1 h
3 MDPUR 347 30 40 0.19 1 4A
whitening disappears after 1 h
See section 6.4.2 for an explanation of water resistance codes
293
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
Advances in Urethane Science and Technology
294
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
The DSC thermogram of the film obtained from a typical hybrid dispersion obtained
using a water-soluble initiator is presented in Figure 6.19.
Particle size distribution and zeta potential of a typical hybrid dispersion obtained using
a ‘redox’ initiator are presented in Figures 6.20 and 6.21, respectively.
The DSC thermogram of the film obtained from a typical hybrid dispersion obtained
using ‘redox’ initiator is presented in Figure 6.22.
295
Advances in Urethane Science and Technology
296
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
The properties of hybrid dispersions prepared using different methods (1a, 2 and 3 – see
Section 6.3.2) and different levels of coalescent (NMP) as well as of films and coatings
made from these dispersions are presented in Tables 6.15 to 6.17. In all dispersions the
chemical structure of the polyurethane-urea and acrylic/styrene polymer component was
the same (see relevant Tables in Section 6.5.2). All dispersions were prepared using a
water-soluble initiator.
297
298
Table 6.15 Properties of hybrid dispersions of the same chemical composition differing only in the
concentration of coalescent (NMP)
Content Properties of dispersions
Method Designation
of Solids Viscosity pH Mechanical Free monomer Particle size Average Zeta MFFT
of of hybrid
NMP % mPa-s Stability %, BA/MM/S distribution particle size potential °C
synthesis dispersion
% Min nm nm mV
MDPUR- <0.02/0.001/
1a 0 34.3 63 6.5-6.8 45 24-688 134.5 -42. 2 <0
ASD 241 0.008
MDPUR-
1a 2.0 30.7 9 7.1-7.4 15 Nt 30-230 61.9 -18. 4 <0
ASD 268
MDPUR-
1a 11.6 - - - - - - - - -
PUM 259a
Advances in Urethane Science and Technology
MDPUR- <0.02/0.005/
2 0 38.3 21 6.2-6.5 15 30-290 261.3 -63.6 <0
ASD 198 0.008
MDPUR- 0.7/0.02/
2 3.0 31.4 10 6.8-7.1 15 50-120 108.4 -60. 7 <0
ASD 97 0.015
MDPUR
3 3.6 41.6 200 7.4-7.7 15 Nt 55-520 115.4 -23.0 <0
348
MDPUR Nt 0.1/0.17/
3 11.6 35.1 7.6 45 33-133 59.4 -14.4 Nt
243 (settling) 1.64
a
dispersion gelled during synthesis
Nt - not tested
Table 6.16 Properties of films made from hybrid dispersions of the same chemical composition differing only
in the concentration of coalescent (NMP).
Properties of films
Method Designation Content
of of Hybrid of NMP δ100 δr εr Swelling Tg
Synthesis Dispersion % Appearance
MPa MPa % H2O MEK Xylene °C
T, El
MDPUR-A- 0 T, G T, G
1a 0 Partly 9.5 15.8 215 -60.8
SD 24 T Pieces Pieces
cracked
MDPUR-A- 13 347 198
1a 2.0 T, Br 8.2 10.8 164 -73.4
SD 268 Wh, Br T, Br T, Br
MDPUR-A-
1a 11.6 - - - - - - - -
SD 259a
MDPUR-A- 0 280 374
2 0 T, El 5.6 11.2 255 -59.0
SD 198 Wh, E T, G T, G
MDPUR-A- 13 341 310
2 3.0 T, El 6.5 10.0 210 -56.6
SD 97 Wh T, G T, G
MDPUR 13 198 198
3 3.6 M, El 14.1 14.5 117 -75.2
348 Wh, E T, Br T, Br
M, El
MDPUR 12 300 251
3 11.6 Film not Nt Nt Nt -66.3
243 Wh, Br T, Br T, Br
even
a
dispersion gelled during synthesis
T- transparent, El - elastic, M - Matte, Wh - white, Br - brittle, G - gelatinous
Nt - not tested
299
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
300
Table 6.17 Properties of coatings made from hybrid dispersions of the same chemical composition differing
only in the concentration of coalescent (NMP)
Properties of coatings
Content
Method of Designation of Drying time, min
of NMP Hardness
Synthesis Hybrid Dispersion Adhesion Water resistance
% Hardness Hardness (Persoz)
(1°) (3°)
2C
1a MDPUR-ASD 24 - 30 50 0.14 4
whitening disappears after 1 h
1A
1a MDPUR-ASD 268 2.0 30 50 0.22 4
whitening disappears after 1 h
nt
Advances in Urethane Science and Technology
In Tables 6.18 to 6.20 the properties of two hybrid dispersions prepared using the same
method 1a (see Section 3.2) and having the same structure of the acrylic/styrene part of
the hybrid but based on different polyols, as well as of films and coatings made from
them are presented. Neither dispersion contained coalescent (NMP)
(b) Effect of introducing double bonds to the polyurethane-urea part of the hybrid
Properties of hybrid dispersions prepared according to the different methods (1a, 1b and
2 - see Section 6.3.2) and based on the same polyol (PTMG 2000), but differing in the
presence or absence of double bonds in the polyurethane-urea part of the hybrid, as well
as of films and coatings made of them, are presented in Tables 6.21 and 6.22. All
dispersions have a similar low level (2.0-3.3%) of coalescent and have the same structure
of the acrylic/styrene part of the hybrid. ‘Redox’ initiator was used in the synthesis of
dispersions according to the method 2, and in all other dispersions presented in these
tables a water-soluble initiator was applied.
301
Table 6.18 Properties of hybrid dispersions differing only in the structure of the polyurethane-urea
302
part of the hybrid
Properties of dispersions
Designation Solids Viscosity pH Mechanical Free Particle size Average Zeta MFFT
of hybrid Polyol % MPa-s Stability monomer, distribution particle pot °C
Dispersion Min BA/MM/S nm size mV
% nm
MDPUR-ASD Polyetherdiol 34.3 63 6.5-6.8 45 <0.02/0.0- 24-688 134.5 -42.2 <0
24 (PTMG-2000) 01/0.008
MDPUR-ASD Polyesterdiol (Bester
39.5 18 6.5-6.8 15 Nt 78-197 175.6 -62.8 +18
300 190 + Bester 195)
Nt – not tested
Advances in Urethane Science and Technology
Table 6.19 Properties of films made from hybrid dispersions differing only in the structure of
polyurethane-urea part of the hybrid
Properties of films
Designation of
Polyol Appear. δ100 δr εr Swelling Tg
hybrid Dispersion
MPa MPa % H2O MEK Xyl e n e °C
Polyetherdiol T, G,
MDPUR-ASD 24 T, El 9.5 15.8 215 0.9 0.8 pieces -60.8
(PTMG-2000) pieces
Polyesterdiol (Bester
MDPUR-ASD 300 T, pieces - - - - - - -
190 + Bester 195)
T- transparent, El - elastic, M - matte, Wh - white, Br - brittle, G - gelatinous
Table 6.20 Properties of coatings made from hybrid dispersions differing only in the structure of the
polyurethane-urea part of the hybrid
Properties of coatings
Designation of hybrid Drying time, min
Polyol Hardness
Dispersion Adhesion Water resistance
Hardness Hardness (Persoz)
(1°) (3°)
MDPUR-ASD 24 Polyetherdiol (PTMG-2000) 30 50 0.14 4 2C whitening disappears after 1 h
Polyesterdiol
MDPUR-ASD 300 25 25 0.3 2 1B no whitening
(Bester 190 + Bester 195)
See section 6.4.2 for an explanation of water resistance codes
Table 6.21 Properties of hybrid dispersions differing only in the presence of double bonds in the
polyurethane-urea part of the hybrid
Double bonds Properties of dispersions
Method Designation
in Solids Viscosity pH Mechanical Free monom. Particle size Average Zeta MFFT
of of hybrid
polyurethane- % mPa-s Stability %, distribution particle size potential °C
synthesis dispersion
urea Min. BA/MM/S nm nm mV
MDPUR-A-
1a - 30.7 9 7.1-7.4 15 - 43.8-174.4 61.9 -18.4 <0
SD 268
MDPUR-A-
1a + 35.3 24 6.5-6.8 15 - 35-180 81.7 -31.2 <0
SD 307
MDPUR-A-
1b - 34.2 10 6.8-7.1 15 - 66.8-422.1 94.4 -39.1 <0
SD 270
MDPUR-A-
1b + 34.4 18 6.5-6.8 15 - 30-70 32.5 +0.7 <0
SD 304
MDPUR-A- <0.02/<0.008/
2 - 31.4 8 6.8-7.1 15 105-165 130.6 -39.8 <0
SD 98 <0.001
MDPUR-A- <0.02/<0.008/ <0
2 + 33.6 9 5.9-6.2 >9 0 25-190 66.9 -33.1
SD 102 <0.001
303
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
304
Table 6.22 Properties of films made from hybrid dispersions differing only in the presence of double bonds
in the polyurethane-urea part of the hybrid
Properties of films
Method Designation Double bonds in
of of Hybrid polyurethane- δ100 δr εr Swelling Tg
Synthesis Dispersion urea Appearance
MPa MPa % H2O MEK Xylene °C
MDPUR-ASD 13 347 198
1a - T, El 8. 2 10.8 164 -73.4
268 Wh, El T, Br T, Br
MDPUR-ASD
1a + T, Cracks - - - - - - -54.1
307
Advances in Urethane Science and Technology
305
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
Advances in Urethane Science and Technology
TEM photos of the particles of two hybrid dispersions prepared using less hydrophobic
and more hydrophilic monomer prepared according to method 1a are presented in Figures
6.23 and 6.24. A TEM micrograph of the particles of a hybrid dispersion prepared
according to method 3 using more hydrophobic monomer is presented in Figure 6.25.
306
Table 6.24 Properties of hybrid dispersions differing only in the chemical structure of the acrylic/styrene
part of the hybrid
Properties of dispersions
Method Designation
of of hybrid Monomer Solids Viscosity pH Mechanical Free monom. Particle size Average Zeta MFFT
synthesis dispersion % mPa.s Stability %, distribution particle size potential °C
Min. BA/MM/S nm nm mV
1a MDPUR- BA 32.6 75 6.2-6.5 45 0.3/-/- 15-450 93.7 -27.6 <0
ASD 23
1a MDPUR- MM 34.5 70 7.1-7.4 45 -/1.15/- 18-360 83.7 -19.3 <0
ASD 22
1a MDPUR- S 34.0 13 6.2-6.5 45 -/-/0.02 16-525 96.3 -26. 6 <0
ASD 14
1a MDPUR- BA/MM/S 34.3 63 6.5-6.8 45 <0.02/0.001/ 24-688 134.5 -42.2 <0
ASD 24 1/1.32/1.64 0.008
3 MDPUR BA 35.1 40 8.2-8.5 45 0.09/-/- 40-270 73.8 -31.3 <0
247
3 MDPUR MM 35.6 44 8.2 45 -/0.92/- 35-130 61.1 -31.6 nt
245
3 MDPUR S 35.2 36 7.3-7.6 90 -/-/0.11 40-250 87.1 -27.9 <0
250
3 MDPUR BA/MM/S 35.1 nt 7.6 45 0.1/0.17/1.64 33-133 59.4 -14.4 nt
243 1/1.32/1.64 (settling)
nt - not tested
307
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
308
Table 6.25 Properties of films made from hybrid dispersions differing only in the chemical structure of
the acrylic/styrene part of the hybrid
Properties of films
Method Designation
of of Hybrid Monomer δ100 δr εr Swelling Tg
Synthesis Dispersion Appear.
MPa MPa % H2O MEK Xylene °C
MDPUR- 13 347 409 - 57.6
1a BA T, El 3.5 4.2 136
ASD 23 Wh, El T, G T, G +79.6
MDPUR- - 59.8
1a MM T, Pieces nt nt nt nt nt nt
ASD 22 +71.4
MDPUR- - 54.0
1a S Wh Pieces nt nt nt nt nt nt
ASD 14 +72.1
BA/MM/
MDPUR- 0
1a S1/1.32/ T, El 9.5 15.8 215 T, G T, G -60.8
Advances in Urethane Science and Technology
ASD 24 T
1.64
13 483 341 - 77.37
3 MDPUR 247 BA T, El 4.3 6.7 230
Wh T, G T, G - 47.38
- 66.6
3 MDPUR 245 MM T, Pieces nt nt nt nt nt nt
+77.0
26 224 251
3 MDPUR 250 S Wh, El 10.4 14.5 201 - 67.7
Wh, El T, Br T , Br
AB/MM/ M, El
12 310 251
3 MDPUR 243 S1/1.32/ Uneven nt nt nt - 66.3
Wh, Br T, Br T , Br
1.64 film
T - transparent, El - elastic, M - Matte, Wh - white, Br - brittle, G - gelatinous
nt - not tested
Table 6.26 Properties of films made from hybrid dispersions differing only in the chemical structure of
the acrylic/styrene part of the hybrid
Properties of coatings
Method
Designation of Drying time, min
of Monomer Hardness
Hybrid Dispersion Adhesion Water resistance
synthesis Hardness Hardness (Persoz)
(1°) (3°)
1a MDPUR-ASD 23 BA 35 50 0.09 1 2C
1a MDPUR-ASD 22 MM 35 50 0.27 4 1B
1D
1a MDPUR-ASD 14 S 35 50 0.12 1
whitening disappears after 1 h
BA/MM/S 2C
1a MDPUR-ASD 24 30 50 0.14 4
1/1.32/1.64 whitening disappears after 1 h
MDPUR 2B
3 BA 35 60 0.06 4
247 whitening disappears after 1 h
nt
3 MDPUR 245 MM nt nt nt nt
Cracks
1D
3 MDPUR 250 S 40 50 0,14 4
whitening disappears after 1 h
BA/MM/S
3 MDPUR 243 120 16 5 0.14 1 nt
1/1.32/1.64
See section 6.4.2 for an explanation of water resistance codes
nt - not tested
309
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
Advances in Urethane Science and Technology
310
Table 6.27 Properties of hybrid dispersions differing only in the presence or absence of crosslinking of the
acrylic/styrene part of the hybrid
Properties of dispersions
Crosslinking
Designation of Mechanical Particle size Average
of Acrylic/ Solids Viscosity Zeta pot MFFT
hybrid Dispersion pH Stability distribution particle size
styrene % MPa-s mV °C
Min nm nm
MDPUR-ASD 307 - 35.3 24 6.5-6.8 15 35-200 81.7 -31.2 <0
MDPUR-ASD 309 + 35.6 17 6.8-7.1 45 25-150 50.0 -64.5 <0
Table 6.28 Properties of films made from hybrid dispersions differing only in the presence or absence of
crosslinking of the acrylic/styrene part of the hybrid
Properties of films
Designation Crosslinking of
of Hybrid Acrylic/styrene δ100 δr εr Swelling Tg
Dispersion part of hybrid Appear.
MPa MPa % H2O MEK Xylene °C
MDPUR-
- T, cracks - - - - - - - 54. 1
ASD 307
MDPUR- 56 Nt 130
+ T, El - 11.1 53 - 57.5
ASD 309 Wh, Br Pieces T , Br
T - transparent, El - elastic, M - matt, Wh - white, Br - brittle, G - gelatinous
Nt - not tested
311
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
Advances in Urethane Science and Technology
Additional experiments were performed in order to better clarify the mechanism of the process
of formation of hybrid dispersions and the effect of various factors on their properties.
This experiment was designed in order to further clarify the mechanism of hybrid
dispersion particles formation. In this experiment, monomer was continuously added to
DPUR (of the composition corresponding to DPUR-237 – see Section 6.3.1) at a rate of
1 ml/h. After each hour, samples of the dispersion were taken and the following parameters
were determined: particle size, particle size distribution and zeta potential. Changes of
particle size distribution in the course of swelling are shown in Figure 6.26, and changes
of average particle size and zeta potential in Figure 6.27.
312
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
After 5 h
Initial sample
After 2 h
After 1 h
After 9 h
After 3 h
Figure 6.26 Changes of particle size distribution in the course of swelling of DPUR
particles with monomer
313
Advances in Urethane Science and Technology
Taking advantage of the capabilities offered by the Malvern Zeta Sizer apparatus the
computer analysis of total volume of smaller and bigger particles (it may be noticed that
on the curves showing particle size distribution in Figure 6.26 that there are two peaks
visible). Results are:
Figure 6.27 Changes of average particle size and zeta potential during addition of
monomer to DPUR
314
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
This experiment was aimed at clarifying what is the maximum amount of monomer that
would be able to swell polyurethane-urea, i.e., would be able to take part in the formation
of hybrid dispersion particles, assuming that enough time is allowed to achieve equilibrium.
Films made from DPUR 258 were used in this experiment. The following results were
obtained:
The changes of the mass of polyurethane-urea film in the course of swelling with monomers
is shown in Figures 6.28 to 6.30.
Figure 6.28 Change of the mass of polyurethane-urea film (made from DPUR 258) in
the course of swelling with MM
315
Advances in Urethane Science and Technology
Figure 6.29 Change of the mass of polyurethane-urea film (made from DPUR 258) in
the course of swelling with styrene
Figure 6.30 Change of the mass of polyurethane-urea film (made from DPUR 258) in
the course of swelling with MM/S = 1/1 (w/w) mixture
316
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
The degree of crosslinking of the starting dispersions (DPUR and ASD) and the hybrids
was determined using the equilibrium swelling test described in Section 4.3. The following
parameters were calculated based on the results of this experiment:
The morphology of dispersion particles was investigated using the method described
in Section 6.4.1. Examples of different morphologies of particles of hybrid dispersions
synthesised in this study according to methods 1a, 1b and 3 are presented in Figure
6.31 in comparison with particles of the starting dispersion of BA/MM/S copolymer.
The contrast was selected so that in pictures ‘d’ and ‘c’ white colour represents the
polyurethane-urea part of the hybrid and in picture ‘b’ the same colour represents the
acrylic/styrene part of the hybrid.
The surface free energy of films made from hybrid dispersions was determined in order
to look at the values of its non-polar and polar component which could provide some
information on whether polyurethane-urea or acrylic/styrene polymer is on the film surface
(see Table 6.30).
317
Advances in Urethane Science and Technology
318
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
Table 6.30 Surface free energy (γ) and its polar (γp) and non-polar (γd)
components of films made from hybrid dispersions prepared according to
method 3. Results obtained for films made from DPUR and starting
acrylic/styrene dispersions are also shown for comparison
Designation of Acrylic/styrene part of γ γp γd
hybrid dispersion the hybrid mJ/m2 mJ/m2 mJ/m2
MDPUR 243 BA/MM/S 1/1.32/1.64 35.4 4.0 31.4
MDPUR 245 MM 34.7 3.3 31.4
MDPUR 247 BA 31.3 6.3 25.0
MDPUR 250 S 36.7 0.7 36.0
ASD 15 BA/MM/S 1/1.32/1.64 22.5 18.0 4.5
ASD 18 BA 23.0 1.8 21.2
DPUR 258 - 33.7 2.3 31.4
319
Advances in Urethane Science and Technology
6.6 Discussion
• Properties of dispersions
When the properties of hybrid dispersions having the same composition but prepared
according to different methods using water-soluble and ‘redox’ initiator (Tables 6.9 and
6.12) are compared with the properties of starting DPUR and ASD (Tables 6.1 and 6.4)
it can be noted that:
a) The type of initiator does not affect the properties of hybrid dispersions significantly
except for free monomer content which is distinctly lower when a ‘redox’ initiator is
used. This suggests that the mechanism of hybrid particle formation is similar in
both cases. This is strongly supported by the direct evidence that there is no difference
in morphologies of particles of hybrid dispersions obtained according to method 2
(see Section 6.3.2) using water-soluble or ‘redox’ initiator (compare Figures 6.32
and 6.33).
b) Neither method of synthesis nor type of initiator have any influence on mechanical
stability, viscosity or MFFT of hybrid dispersions. However, the following observations
can be made:
• mechanical stability of all hybrid dispersions is slightly lower than that of the
starting dispersions, but nevertheless it is good enough for successful application
of hybrid dispersions in practice (there was no settling of dispersions during
storage). Relatively high negative values of zeta potential observed for all hybrid
dispersions confirm that their stability should be good. For all hybrid dispersions
MFFT is < 0 °C (just as for starting DPUR) which should undoubtedly be
considered as their great advantage when compared to ADS.
• the viscosity of all hybrid dispersions is quite low and similar to the starting DPUR
a) The average particle size of hybrid dispersions depends strongly on the method of
synthesis (regardless of the initiator). The lowest values of average particle size are
observed for MDPUR-ADS prepared according to method 1a, i.e., with continuous
feeding of monomers to DPUR, and the highest for MDPUR prepared according to
method 3, i.e., emulsifying of prepolymer-ionomer in ASD followed by crosslinking
320
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
with polyamine. This result is in good correlation with the assumptions concerning
the mechanism of hybrid particles formation presented in Section 6.3.2. If the
dispersion is prepared according to method 1a, smaller ASD particles can be formed
along with bigger hybrid particles which may have various morphologies. If the
dispersion is prepared according to method 3, only bigger ‘core-shell’ particles can
be formed. Detailed comparison of the size of particles of starting DPUR (DPUR
346) and hybrid dispersions prepared from it according to method 1a (MDPUR-
ASD 268 and MDPUR-ASD 274) confirms this assumption that the average particle
size for starting DPUR is higher than for hybrid dispersions prepared from it.
b) The particle size distribution observed for hybrid dispersions is the lowest for MDPUR-
ASD prepared according to method 2, i.e., diluting prepolymer-ionomer with monomers,
emulsifying this solution in water, crosslinking with polyamine and polymerisation of
monomers. This can be explained by the fact that in this method, formation of hybrid
dispersion particles proceeds ‘in situ’, i.e., both prepolymer-ionomer (precursor of
polyurethane-urea component of the hybrid) and the monomers (precursors of acrylic/
styrene component of the hybrid) are emulsified at the same time, which means that
the chance of formation of hybrid dispersion particles having similar morphology and
size is very high. Higher uniformity of the whole system in the case when the hybrid
dispersions are prepared according to method 2 is also confirmed by the higher negative
values of zeta potential observed for these dispersions.
• Properties of films
From a practical point of view it is very important that, regardless of the type of initiator
and method of synthesis, the films made from hybrid dispersions are transparent (this
proves that their structure is quite uniform), strong, elastic and resistant to water (very
good resistance) and organic solvents (good resistance).
When the properties of films made from hybrid dispersions having the same composition
but prepared according to different methods using water-soluble and ‘redox’ initiator
(Tables 6.10 and 6.13) are compared with the properties of films made from the starting
DPUR (Table 6.2) it can be noted that:
a) Type of initiator practically does not have any influence on mechanical properties,
water resistance or solvent resistance of films made from hybrid dispersions. Slightly
higher values of elongation at break observed for films made from hybrid dispersions
prepared using ‘redox’ initiator (regardless of the method of synthesis) can result
from the fact that less grafting (or crosslinking) proceeds in the course of
polymerisation.
321
Advances in Urethane Science and Technology
b) The effect of method of synthesis on the mechanical properties, water resistance and
solvent resistance of the films can be clearly seen when the results obtained for the
films made from hybrid dispersions prepared according to method 3 and according
to other methods are compared. In the cases where method 3 has been applied, the
films are distinctly stronger, less elastic and better resistant to water and organic
solvents. It is also interesting that their Tgs are lower (closer to the one observed for
DPUR). The explanation of this phenomenon is not clear since taking into account
the assumed hybrid particles formation mechanism presented in Section 6.3.2 one
could rather expect that not only Tg, but also the mechanical properties of the films
made from hybrid dispersions prepared according to method 3 would resemble those
of films made from DPUR (high elasticity and water/solvent resistance) more than
the properties of the films made of dispersions prepared according to the other
methods. Nevertheless, when this problem is more deeply analysed, it can be
anticipated that the reason for this phenomenon is the very high (much higher than
in dispersions prepared according to the other methods) number of particles of ‘pure
core-shell’ structure in dispersions prepared according to method 3. In the film made
of hybrid dispersion prepared according to this method, a very regular heterogeneous
structure is formed during coalescence of particles. In such a structure, acrylic/styrene
polymer (which constitutes the core in dispersion particles) is the dispersed phase
while polyurethane-urea (which constitutes the shell in dispersion particles) is the
matrix. Acrylic/styrene polymer then plays the role of toughener while polyurethane-
urea increases the water and solvent resistance of the film and provides its low Tg.
The second Tg corresponding to acrylic/styrene polymer, should also be detected in
this particular system, but DSC does not have the resolution which would be needed
here. However, when the torsion pendulum method was applied we were able to
spot this second Tg quite easily. We believe that this particular hybrid system is very
interesting and requires further studies to explain the mechanisms of both hybrid
particles and film formation.
• Properties of coatings
Regardless of the initiator and method of synthesis, all coatings made from hybrid
dispersions show very short drying time, reasonable hardness and water resistance. Based
on the results presented in Tables 6.11 and 6.14 it can be concluded that the best properties
of coatings have been achieved with dispersions prepared according to method 3, i.e., by
emulsifying the prepolymer-ionomer in ASD, and the worst – for dispersions prepared
according to method 2, i.e., by diluting the prepolymer-ionomer with monomers. The
properties of coatings made from hybrid dispersions are generally much better than those
of coatings made from starting DPUR (starting ASD dispersions did not form coatings
since they did not contain any coalescent).
322
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
• Properties of dispersions
As can be seen in Table 6.15, concentration of coalescent has a significant effect on the
properties of hybrid dispersions, regardless of the method of synthesis. Very high levels
of coalescent lead to gelling in the course of polymerisation (see MDPUR-ASD prepared
according to method 1a) or partial settling of dispersion during storage (see MDPUR
243 prepared according to method 3). It can also be noted that for dispersions prepared
without coalescent according to methods 1 and 2, average particle size is distinctly higher
and particle size distribution broader than when the process is carried out in the presence
of coalescent. This phenomenon can be explained by the positive influence of small
amounts of coalescent on the process of hybrid particle formation in the course of
polymerisation. It can be assumed that the particles of starting DPUR are partly swelled
by coalescent which helps the monomers to diffuse into them. When hybrid dispersions
have been prepared according to method 3, the reverse effect is observed. Here, increasing
the coalescent concentration results in diminishing both the average particle size and
particle size distribution range. The reason for this may be easier formation of separate
DPUR particles at lower levels of coalescent.
• Properties of films
The effect of coalescent on the properties of films made from hybrid dispersions seems to
be not quite significant (see Table 6.16) which is strange when taking into account the
important role of coalescent in the process of film formation. The only observation that
can be made here is that, as has already been stated in Section 6.6.1.2.1, high levels of
coalescent cannot be recommended in preparation of hybrid dispersions.
• Properties of coatings
Here, the effect of coalescent is clearly visible (see Table 6.17), but only if the coalescent
content is high (MDPUR 243). Then, drying time increases and coating hardness decreases.
However, no differences can be observed between the properties of coatings made from
hybrid dispersions which do not contain any coalescent and those which contain only
small amounts of coalescent.
323
Advances in Urethane Science and Technology
• Properties of dispersions
As can be seen from Table 6.18, the chemical structure of the polyurethane-urea part of
the hybrid does not affect the properties of hybrid dispersions, obviously not counting
the MFFT which is much higher if polyesterdiol has been used as a starting material for
the prepolymer-ionomer synthesis (compare MDPUR-ASD 300 and MDPUR-ASD 24).
• Properties of films
Here, the effect of chemical structure of the polyurethane-urea part of the hybrid is, of
course, substantial (see Table 6.19). For hybrid dispersion synthesised without coalescent
using polyesterdiol as a starting material for the prepolymer-ionomer, the Tg is so high that
films cannot be obtained. On the other hand, if polyetherdiol is applied as a starting material
in the synthesis of the same dispersion, films of very good mechanical properties are obtained.
This was the reason for using polyetherdiol rather than polyesterdiol as the starting material
for synthesis of dispersions in this study.
• Properties of coatings
It can be clearly seen from Table 6.20 that the hybrid dispersion prepared using polyesterdiol
(MDPUR-ASD 300) forms faster-drying coatings of better hardness, adhesion and water
resistance. This information may be important from the practical point of view.
• Properties of dispersions
As can be seen from Table 6.21, introducing double bonds into the polyurethane-urea
component of the hybrid does not affect the properties of hybrid dispersions regardless
of the method used for their synthesis.
• Properties of films
Based on the data contained in Table 6.22 it can be concluded that the effect of introducing
double bonds to polyurethane-urea on the properties of films made from hybrid dispersions
324
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
is exceptionally strong. Regardless of the method of hybrid dispersion synthesis, the presence
of double bonds in polyurethane-urea results in a distinct increase in Tg and a drastic
increase in film stiffness (decrease in film elasticity), and, an increase in resistance to water
and organic solvents. This may mean that grafting or even crosslinking takes place during
polymerisation and/or crosslinking via oxidation of double bonds suggested earlier [29],
proceeds during air-drying of the films.
• Properties of coatings
The presence of double bonds in polyurethane-urea has very significant effects on the
properties of coatings made from hybrid dispersions (see Table 6.23). It leads to drastic
(two-fold) increase in coating hardness and to distinct shortening of drying time. However,
there is no effect of double bonds in polyurethane-urea on adhesion or water resistance
of the coatings.
Generally, based on the results shown above, it can be concluded that introducing double
bonds into polyurethane-urea is undoubtedly one of the ways to obtain coatings of
excellent properties based on hybrid polyurethane-urea-acrylic/styrene dispersions.
• Properties of dispersions
Table 6.24 shows that the chemical structure of the acrylic/styrene part of the hybrid has
no effect on the macroscopic properties of hybrid dispersions, although the appearance
of the dispersion particles may be quite different (compare Figures 6.23 and 6.24).
Difference in the appearance of particles may result from the degree of hydrophobicity
of the monomer (this effect is explained in Section 6.3.2.1).
• Properties of films
The chemical structure of the acrylic/styrene part of the hybrid has a substantial effect on
the properties of films made from hybrid dispersions (see Table 6.24). When the monomers
that form polymers of high Tg (styrene or MM) are used and the synthesis is carried out
according to method 1a, no film is obtained. For the same monomers but a different method
of dispersion synthesis (method 3), a film was obtained only for styrene and only when a
high level of coalescent was applied. In this case, the mechanical properties, water and
solvent resistance of the film were quite good, but the film was not transparent, which
325
Advances in Urethane Science and Technology
indicated that the uniformity of the system, i.e., the compatibility of polystyrene with
polyurethane-urea was rather poor. Based on the results of investigations of the effect of
styrene on the morphology of the model polyurethane-urea-acrylic/styrene hybrid systems
[41] this assumption can be proved.
• Properties of coatings
Using data presented in Table 6.26 it can be concluded that the type of monomer used in
the synthesis of hybrid dispersions has no effect on the drying time of coatings made
from these dispersions, but has a significant effect on coating hardness. The highest
hardness can be achieved with coatings made from dispersions prepared using MM as a
starting monomer, if it is possible to make a coating (compare MDPUR-ASD 22 and
MDPUR 245 in Table 6.26). It is noteworthy that it was not possible to make any
coating from MDPUR 245 (prepared according to method 3), despite the fact that it
contained a high level of coalescent.
Table 6.27 shows that the properties of hybrid dispersions prepared with additional
crosslinking of acrylic/styrene polymer do not differ from the properties of dispersions
of the same composition prepared without additional crosslinking. Also the properties
of films made from both dispersions were very similar (see Table 6.28). However, lack of
positive effects of additional crosslinking of the acrylic/styrene part of the hybrid may
result from improper selection of starting materials for hybrid synthesis. Therefore, we
intend to investigate this effect further in another study along with further studies on the
effect of introducing double bonds to the polyurethane-urea part of the hybrid.
Based on the results of supplementary investigations (see Section 6.5.4) that were carried
out to better understand the mechanism of hybrid particle formation and to determine
particle morphology, the following observations can be made:
326
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
(a) Investigations of the kinetics of the process of swelling the DPUR particles with
monomers (see Section 6.5.4.1) have shown that the number of ‘bigger’ particles
(particles swelled with monomer and monomer droplets) increases gradually while
the number of ‘smaller’ particles (DPUR particles not swelled with monomers)
decreases. After some time (about 5 hours), equilibrium is reached and then a
bimodal particle size distribution occurs. This direct observation of the process of
swelling DPUR particles with monomers is very interesting because it does not
correlate with the course of swelling the film with monomers (compare with Section
6.5.4.2). In the process of swelling the film with monomers, equilibrium is reached
after a much longer time (12 hours) regardless of the monomer used. This happens
at a much higher degree of swelling (approximately 90%) as compared to the process
of swelling DPUR particles (maximum 40% assuming that all monomer diffused
into the DPUR particles). In practice, a significant amount of the monomer remains
in the form of monomer droplets suspended in DPUR.
It is also interesting to note that during the feeding of the monomers, the zeta
potential becomes less negative which indicates that the uniformity and stability of
the whole system is diminishing.
(b) The results of investigations of crosslink density of films made from hybrid
dispersions (see Section 6.5.4.2) show that when the values of the crosslink density
of the films determined by swelling in toluene are compared it can be concluded
that regardless of the method used for synthesis of hybrid dispersions the films are
all crosslinked, or at least they behave as if they were crosslinked. This may mean
either that the polymers in the films are actually crosslinked or very strongly branched
(by the grafting of monomers on polyurethane-urea) and/or that there is strong
domination of ‘core-shell’ structure with a ‘shell’ made of polyurethane-urea. The
latter possibility is supported by the properties of films made from hybrid dispersions
prepared according to method 3 (emulsifying of prepolymer-ionomer in ASD and
then crosslinking with polyamine). It may also be additionally supported by:
327
Advances in Urethane Science and Technology
• the values of the polar and non-polar (dispersed) components of surface free
energy determined for the films made from hybrid dispersions being very similar
to the values found for the films made from DPUR, not ASD (see Table 6.30).
This may be considered as evidence for ‘hiding’ of the acrylic/styrene part of the
hybrid under the film surface, which results presumably from the ‘core-shell’
structure of the coalesced dispersion particles that form the film, where
polyurethane-urea constitutes the shell and acrylic/styrene polymer the core.
(c) The morphology of hybrid dispersion particles shown in Figure 6.31 looks quite
interesting. It can be clearly seen that regardless of the method of synthesis, resulting
in particles of different morphologies, they all have ‘core-shell’-like structures and
differ substantially from the particles of ASD, which have a uniform structure.
Deeper analysis of these structures leads to the conclusion that for MDPUR-ASD
prepared according to Method 1b, the particle morphology resembles that described
as ‘engulfed’ [15] for other types of hybrid dispersions (see Figure 6.1). For our
hybrid dispersion, small round-shaped particles of acrylic/styrene polymer are
‘engulfed’ or ‘englued’ in the bulky core made of polyurethane-urea.
The most interesting result seems to be, however, the specific morphology of the
particles of hybrid dispersions prepared according to method 2 shown in Figures
6.32 and 6.33. It is shown schematically in Figure 6.34.
328
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
Figure 6.35 Schematically presented structure of film made from a hybrid dispersion
of the particle morphology shown in Figure 6.34
329
Advances in Urethane Science and Technology
6.7 Summary
Detailed investigations of coating properties of hybrid dispersions proved that they could
form transparent coatings of reasonable hardness, adhesion and water resistance.
Based on the data in Table 6.31 the following dependencies may be concluded:
• the properties of films made from hybrid dispersions: dependent on method of hybrid
synthesis, chemical structure of the polyurethane-urea part of the hybrid, presence of
double bonds in the polyurethane-urea and the chemical structure of the acrylic/
styrene part of the hybrid
• the properties of coatings made from hybrid dispersions: dependent on chemical structure
of the polyurethane-urea part of the hybrid and presence of double bonds in the
polyurethane-urea
330
Table 6.31 Effect of various factors on the properties of hybrid polyurethane-urea-acrylic/styrene polymer
dispersions synthesised in this study
Properties of dispersions Properties of films Properties of coatings
Water
Stability Particle size
Mechanical and Water
Viscosity and size MFFT Tg Drying Hardness Adhesion
properties solvent resistance
pH distribution
resistance
Method of hybrid
synthesis
- ++ - + + - + + + +
Type of initiator - - - - - - - - - -
Coalescent content - + - - - - + - + -
Chemical structure
of polyurethane-urea - - ++ ++ ++ ++ + + ++ +
part of hybrid
Presence of double
bonds in
polyurethane-urea
- - - ++ ++ ++ ++ - ++ -
part of hybrid
Chemical structure
of acrylic/styrene - - - ++ + ++ - + + +
part of hybrid
Crosslinking of
acrylic/styrene - - - - - - - - - -
polymer
+: some effect was observed
-: no effect was observed
++: distinct effect was observed
331
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
Advances in Urethane Science and Technology
Acknowledgements
References
332
Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…
27. B. Gruber, Presented at the ACS Meeting, San Francisco, CA, 1992, Paper
No.108.
333
Advances in Urethane Science and Technology
35. ISO 1522 Paints and Varnishes - Pendulum Damping Test, 1998.
36. ISO 2812-1 Paints and Varnishes - Determination of Resistance to Liquids - Part
1 - General Methods, 1993.
37. DIN 53150 Testing of Paints, Varnishes and Similar Coating Materials;
Determination of the Drying stage of Coatings (Modified Bandow Wolff
Method), 1995
334
7
Adhesion Behaviour of Urethanes
7.1 Introduction
The adhesion behaviour of urethanes depends not only on their composition but also on
the contribution of physico-chemical factors, in particular on the surface properties of
urethanes. Two specific aspects of surface behaviour are considered in this chapter, which
is arranged in three sections: the first concerns the ability of some polymers, including
PU to adopt a variety of surface structures or orientations, as dictated by preparation
conditions and the medium in contact with the urethane polymer. The ability to undergo
surface restructuring is a factor in the well documented difference between surface and
bulk properties of all polymers. A second aspect of surface behavior, exerting major
influence on the adhesion of urethanes to a variety of plastics or glass, is the acid-base
interaction between the adhesive polymer and the substrate. Acid-base interactions are a
subject of discussion in the second section of this chapter. Finally, the role of silane
modifiers in affecting the adhesion behavior of defined urethane adhesive formulations
is considered.
This study involves surface analyses on formulations typical of PU adhesives. The objective
is to examine the surface properties of such polymers; to investigate the possibility of
surface changes occurring when the adhesive, cured to different degrees, is contacted
with orienting fluids notably water; and to indicate the degree to which surface
restructuring may influence the bonding characteristics of PU adhesives.
335
Advances in Urethane Science and Technology
7.2.1 Experimental
7.2.1.1 Materials
The major portion of this work used as reference a two-part soft-segment PU formulation.
A prepolymer of methylene bis-4-cyclohexylisocyanate (MDI) with a large amount of
2,4-isomer constituted Part A. Part B was a mixture of two tri-functional polyols with
molecular weights of 900 and 500. These were used in weight ratios of 70/30. Part B also
contained about 0.08 wt% of a tin catalyst (UL-28, Witco Chemicals).
The formulations were cured to various degrees. In one cure procedure, the polymer was
first hardened at 100 oC for 1 hour and then exposed to ambient conditions for five
days. In a number of cases, this was followed by further curing at 50 °C/50% relative
humidity. In an alternative procedure, the initial 100 °C/1 hour exposure was followed
by conditioning for varying lengths of time at room temperature and 50% RH. These
alternative post-cure processes were followed to produce a range of crosslink levels and
to vary also the degree of enolisation occurring during cure.
In order to produce PU films, parts A and B were mixed to give ratios of NCO:OH of
1.03, and in some cases, 1.13. These mixtures were cast onto previously degreased
glass microscope slides and drawn down with doctor blades to give final polymer
coating thicknesses of 1.2 mm. Samples were cured by the regimes outlined previously.
Surface energies were obtained from experimental determinations of contact angles
for sessile drops, using a Rame-Hart goniometer. All experiments were conducted at
30 °C. Data were obtained for wetting by dispersion force fluids, represented by a
series of n-alkanes, and by formamide, glycerol, ethylene glycol, and water (liquids
with appreciable non-dispersive forces). In all cases, the dead space around the enclosed
sample was in equilibrium with the vapour of the contacting liquid, and contact angles
at zero contact time were obtained by extrapolation of data at finite times of contact.
This avoided the risk of changes due to any possible liquid-substrate interactions. The
procedures of Gent and Schultz [1] were used for determinations of dispersive and
non-dispersive contributions to the overall surface energies, γsd and γsnd. These were
obtained with experimental uncertainties of less than 6%.
Following initial surface energy evaluations, the PU films were immersed in water at
controlled temperatures and for controlled times. Following immersion, films were
dried under vacuum (room temperature, 24 hours) and surface energies were
redetermined.
336
Adhesion Behaviour of Urethanes
Analyses of chemical structure in cured films, before and after immersion, were carried
out by IR spectroscopy, utilising a Bonem-Michelson-100 FTIR instrument.
Another aspect of surface analysis involved the use of inverse gas chromatographic (IGC)
techniques [2]. Specifically, IGC was used to evaluate the acid-base interaction potential
of PU surfaces and of substrates against which the adhesive was to be bonded. In the
IGC determination, the polymers are the non-volatile phase, while the volatile phase
included n-alkane vapours and vapours defined as acids and bases by contemporary
theories. Such as the theory of Gutmann [3], applied in the present case along with
developments reported separately by Saint Flour and Papirer [4] and by Deng and Schreiber
[5]. In this protocol [5], chloroform represents a pure acid vapour; tetrahydrofuran and
diethylether are strong bases; while acetone may be used as an amphoteric vapour. The
IGC procedure, described in detail elsewhere [2, 4, 5], yields interaction parameters for
the test solids. The parameter acceptor number (AN) is a measure of surface acidity; the
donor number (DN), evaluates the base interaction potential; and an averaged interaction
potential is given by
Ki = DN – AN. (1)
For IGC work with PU, the polymer was cured against Teflon sheet, peeled off, and
powdered by pestle and mortar. About 1.0 g of powder was packed into previously
washed and dried steel columns, 15 cm long and 1.2 cm in diameter. Polymer substrates,
such as PVC and ABS, were deposited onto Chromosorb support (acid-washed, 60/40
mesh diatomaceous earth) following well-established procedures [6]. Small, nanoliter
amounts of the volatile phase were injected into the chromatograph, with helium as the
carrier gas at a flow rate of 10 cm3/min.
7.2.1.3 Adhesion
337
Advances in Urethane Science and Technology
shear data were obtained with an Instron tester at 25 mm/min jaw separation speed. At
least three, and in most cases five to eight, separate determinations were carried out for
each of the systems, with averaged results presented here. The experimental uncertainty
in these determinations ranged from 8 to about 20%.
7.2.2.1.Surface Analyses
338
Adhesion Behaviour of Urethanes
339
Advances in Urethane Science and Technology
agreement with data in reference [5], and the rate of change is strongly temperature-
dependent. Activation energies for an assumed diffusion-dependent process may be
calculated from the results in Figure 7.1. The value obtained, 11.4 kcal/mol, is in
excellent agreement with expected values for diffusion processes of chain molecules
through polymer matrices, reported to be in the 10-12 kcal/mol interval [8]. By this
interpretation, contact with water (γnd = 52 mJ/m2) leads to the diffusion of hard
segments into the surface region of the PU, driven by the thermodynamic need to
minimise the interfacial tension. In the new equilibrium state, the polymer has a
considerably higher total surface energy, since as shown in Figure 7.2, the dispersion-
force component is not greatly affected by contact with water. When in the
conformation dictated by water, the total γs of the PU surface is 42.5 mJ/m2, more
than 10% greater than when this polymer is in its ‘air-conformation’ state.
340
Adhesion Behaviour of Urethanes
While the emphasis above is on structural reconformation as the cause of changes in γsnd,
other mechanisms may be considered. They include:
The former of these may be rejected, since the data in Figure 7.1 were shown to be
independent of the drying time following removal of samples from water. Vacuum-drying
for 5 hours, for the standard 24 hours, and 48 hours produces samples in which γsnd was
unchanged within experimental error. The observed surface energy effects are therefore
unlikely to be caused by the physical incorporation of water in the polymer.
The second mechanism, that of enolisation, may be investigated using IR data, focusing on
the NCO absorption peak at 2300 cm–1 and on the OH signal at 3300 cm–1. Pertinent
results are given in Table 7.2 for specimens prepared under the full range of cure conditions.
341
Advances in Urethane Science and Technology
Comparison of the NCO and OH absorption peak areas shows that chemical changes
continue to take place as the urethane polymer ages. Specimens made by routes C0 and C1
show that an additional 30 minutes at 100oC reduces both signals, and more particularly
that for NCO. Additional reductions are observed for samples cured via routes C2 and C3.
The series C3-C7 shows continued decreases in the OH peak and a slight downward
tendency for the NCO peak. The sequence C8-C12, referring to room temperature exposure
at controlled humidity, follows similar patterns, though the magnitude of the reductions in
peak areas is somewhat diminished. It is concluded that chemical changes do indeed take
place in the curing polymer, and that even after 14 days at 50 °C these changes have not
gone to completion. The speculation on enolisation effects therefore appears to be supported
analytically. The relative contribution of the supposed enolisation effects to observed changes
in surface energy must now be questioned.
One aspect of the question is considered by the results in Figure 7.3. This shows the response
of γnd of the reference polymer to immersion in a non-polar, dispersion-force fluid
342
Adhesion Behaviour of Urethanes
(n-heptane). Prior to immersion in heptane at 60 °C, the polymer had been immersed in
water at the same temperature for 48 hours to bring the surface state into equilibrium with
the orienting fluid. Consequently, the ordinate in Figure 7. 3, namely the difference between
γnd of the polymer in the ‘water’ and the ‘air’ steady states, originates at about 6 mJ/m2. If
this increase, shown in Figure 7.1, were due entirely to surface restructuring, then immersion
in heptane would restore the ‘air’ steady-state and Δγnd = 0. Instead, however, the non-
dispersive surface energy levels out at about 1.8 mJ/m2. Conformational and reversible
restructuring accounts for the major portion of surface energy changes in the partially
cured PU. A smaller, yet significant irreversible portion, however, arises from changes in
the surface chemistry of the polymer when it is allowed to condition in aqueous media.
Another view of the relationship between surface mobility and the level of cure in the PU
is shown in Figure 7.4. This follows the γsnd on immersion in water at 60 °C of samples
cured under procedures C1, C5, and C10 (see Table 7.1). Clearly, in each case there is a
significant rise in γnd, but the magnitude of the increase varies with the cure procedure.
Figure 7.4 Effect of the degree of cure on the surface restructuring of polyurethane on
water immersion, 60 °C. Cure states coded according to Table 7.1.
(● ) C1, (■ ) C5; (●) C10.
Reproduced with permission from A. Sengupta and H. P. Schreiber, Journal of
Adhesion Science and Technology, 1991, 5, 11, 947, Figure 4. ©1991 VSP BV
343
Advances in Urethane Science and Technology
The most drastic cure in this group was applied to sample C5, which produces the lowest
restructuring effect; C10 and C1 represent progressively milder cure strategies and the
restructuring tendency correspondingly increases. Thus, surface chain dynamics of PUs
are a function of the degree of cure, but under each of the procedures used in this work
the restructuring effect remains sufficiently important, potentially, to affect adhesion,
and related properties of the polymer surface.
The lap-shear experiments described previously, clearly show that the adhesive bond
between PU and a polymer substrate is not uniquely defined by the composition of the
materials in the joint. This is shown by the results in Table 7.3 which gives the adhesive
bond strengths of ABS/PU joints, in which the adhesive was cured by the various routes
344
Adhesion Behaviour of Urethanes
described in Table 7.1. A wide variation in bond strengths is noted; the major trend is
toward reduced adhesion with increasing duration of cure, and thus also with increasing
contact time between PU and ABS. In this case, of course, the orienting medium for the
adhesive is the ABS surface. The results, therefore, suggest that the steady-state
conformation of PU relative to ABS is less favourable to strong bonding than is an initial,
transient conformation of the adhesive surface. This somewhat surprising result may be
rationalised with a knowledge of the acid-base characteristics of the polymer.
IGC determinations of AN, DN, and Ki values for ABS, PVC, and PU are summarised in
Table 7.4, all data being at 30 °C.
These results show ABS to be a moderate base and PVC a strong acid. PU, measured here
in its surface conformation relative to air, is mildly basic. There is, therefore, no tendency
for specific non-dispersive forces to act across the ABS/PU interface, the basicity of the
materials being unfavourable to the establishment of strong, short-range forces of this
kind. On prolonged contact, the initial tendency for dispersion-force soft segments to
populate the PU surface preferentially should become more pronounced, accounting for
the data sequence in Table 7.3. Of course, the trends should be reversed with PVC, for in
this case acid-base attraction should promote adhesive bond strength. Only fragmentary
experimental results are available to test the hypothesis, but as shown in Figure 7.5,
these support it. They are also consistent with the recent work of Berger [9], showing the
influence of non-dispersion forces on the lap shear strength of polymer/metal joints.
A matter of much practical importance is the durability of bond strength when joints
are exposed to aggressive environments. A cursory examination of this point was carried
out as follows: joints of PVC/PU and ABS/PU were prepared by cure procedure C2.
Several of these were aged at 50 °C/50% relative humidity for up to 10 days and lap-
345
Advances in Urethane Science and Technology
shear evaluations were carried out as described previously. The results in Figure 7.5,
however, are limited for the most part to single specimen evaluations. A lower degree
of confidence therefore applies to them. The effects seen in the figure, however, are
entirely consistent with the preceding analyses. With ABS, as before, a slight
deterioration takes place in bond strength. The initial bond strength with PVC as
substrate is significantly higher than that with ABS, and, moreover, increases with
ageing, as called for by the acid-base concept. The orienting potential of polymers
including PU therefore may arise from the nature of acid-base interactions at the point
of contact between materials. These effects may have serious consequences on the
performance parameters of materials such as PU adhesives.
346
Adhesion Behaviour of Urethanes
7.2.3 Conclusions
(1) The surface energy of a PU formulation has been shown to vary with the nature of
the medium in contact with the polymer. The variations are attributed to the non-
dispersive part of the overall surface energy. The observed trends follow closely
earlier observations on two-part, linear PU surfaces.
(4) The adhesive bond strength of PU varies with the degree of acid-base interaction
with the substrate and with the conformational state of the adhesive molecule at
the time of testing. Being basic, PU develops stronger bonds with acidic substrates
(PVC in this case) than with basic ones, this is exemplified by ABS.
As noted in the preceding section, IGC is an excellent tool for measurements of surface
properties and of acid base interaction potentials of macromolecular solids (2, 10, 16).
In this section the importance of acid-base interactions relative to the adhesion of PUs
is considered in greater detail. IGC has been applied to a series of PU adhesives and to
selected polymer substrates, allowing quantitative measurements to be made of the
acid/base (electron donor-acceptor) interaction parameters applicable to the surfaces
of these materials. Acid base pair-interaction parameters for substrate/PU combinations
have been calculated. The bond characteristics of polymer/PU combinations have been
measured, in part by conventional lap-shear procedures and in part, by the more recent
constrained blister detachment method [11, 12]. Possible relationships between bond
properties and acid base interactions have been considered, and a comparison of the
two adhesion tests has been made.
347
Advances in Urethane Science and Technology
7.3.1 Experimental
7.3.1.1 Materials
Six PU adhesives, modelled on practical formulations, have been studied. They are coded
PU1 – PU6. All were two-part compounds, denoted as A and B in Table 7.5. In all cases,
part B contained a tin catalyst.
In each of these formulations A and B were thoroughly mixed, applied to the substrate
and bonded in less than 3 minutes following mixing.
The substrates were rigid PVC, ABS impact resin and the GE polyblend Xenoy. All were
commercial samples, and all were injection moulded to form 8.0 cm x 2.5 cm rectangles,
1.5 mm in thickness.
7.3.1.2 Procedures
For IGC work, PU and substrate polymers were coated onto Chromosorb AW support
(60/80 mesh) from solution. From 6.0 – 11 wt% of polymer was supported on the
348
Adhesion Behaviour of Urethanes
Chromosorb. The coated support was packed in stainless steel columns (0.5 cm id, about
0.8 m long) and served as the stationary phase in IGC. The stationary phases were
contacted with selected vapours at very high dilution and described procedures were
followed [6, 11] to measure the net retention volume, Vn, of the vapours probing the
deposited solids. A Perkin-Elmer Sigma-2 chromatograph with hot wire detector was
used. All determinations were in the range 30 – 60 °C, and values of Vn were measured
in at least triplicate, with a reproducibility better than 4%.
The use of IGC data to obtain acid and base interaction parameters for polymer solids has
been described in detail elsewhere [2,13,14]. The procedure calls for a reference relationship
between RTln Vn and the normal boiling point of mobile-phase materials able to interact
through dispersion forces only, represented in this work by n-alkanes (C6-C9) . Vapours
known to be acids and bases were then injected into the columns, again at extreme dilution.
Following the theories of Gutmann [3], diethyl ether (DEE) and chloroform (CHL) were
used as reference base and acid, respectively. As shown in Figure 7.6 for PVC as stationary
Figure 7.6 Retention volume of vapour probes interacting with PVC as function of
normal boiling point of probe liquids. Data are for 40 °C.
Reproduced with permission from PU Congress.
349
Advances in Urethane Science and Technology
phase, the points for DEE and CHL both deviate from the straight line defined by the
alkanes. Thus, PVC interacts with both the acidic and basic probe, but with a pronounced
prevalence of interaction with DEE. The polymer is therefore a net acid. Following the
principles of reference [1], electron acceptor and donor indices, AN and DN for the polymers
are calculated from the behaviour of the DEE and CHL vapours.
Figure 7.7 Schematic of arrangements for adhesion tests. (a) Specimen construction for
lap-shear test. (b) Specimen construction and installation for blister detachment test.
350
Adhesion Behaviour of Urethanes
made in triplicate with an Instron tester, at a separation speed of 50 mm/min. The blister
test arrangement, used with all substrates, is shown schematically in Figure 7.7b. The
support polymer here had a central hole, 6 mm in diameter, which connected to a pumping
system able to channel distilled water onto the polymer/adhesive assembly. During
application of the adhesive this opening was blocked by a Teflon plug, removed after the
urethane was cured. Adhesive films, again 0.35 mm thick, were deposited on the PVC,
ABS and Xenoy surfaces and cured as above. Bond energies, Ea, were obtained by following
protocols described in recent literature [4, 5]. These require measurement of the pressure
generated by the liquid pumped against the joint at which the adhesive debonds from the
substrate. Bond energy data were reproducible to better than 8%; lap-shear determinations
carried a similar uncertainty.
Acid-base interaction parameters for the materials of this research are given in Table 7.6.
As anticipated by the results in Figure 7.7, and in agreement with the results shown in
the preceding section, these show PVC to be predominantly acidic. ABS has a moderately
basic surface and Xenoy is an amphoteric solid with nearly equal AN and DN indices.
All of the PU systems function as both donors and acceptors, with PU5 and PU6 skewed
to acidity, PU1 and PU4 to basicity, while PU3 and (to a lesser degree) PU2 are amphoteric.
351
Advances in Urethane Science and Technology
The acidic adhesives would be expected to interact strongly with ABS, the basic ones
with PVC. The magnitude of acid-base forces is open to question when amphoteric surfaces
such as Xenoy are involved.
Any attempt to relate acid/base interactions between adhesives and substrates to the
bond properties of the relevant pair depend on a quantitative evaluation of a pair
interaction parameter. There are no broadly accepted theoretical guidelines to such an
evaluation, but a pair interaction parameter, Isp, can be defined by
Here subscripts 1 and 2 designate adherent and adhesive, respectively. Values of Isp for
the polymer/adhesive pairs are given in Table 7.7.
A summary of lap-shear and blister detachment test data is presented in Table 7.8. The
six adhesives perform differently with each of the substrates, suggesting a link with specific
interactions at the adherent/adhesive interface. This is considered in the next section.
Although lap-shear evaluations are more familiar, they do not necessarily evaluate the
strength of adherent/adhesive interfaces. Cohesive failure of the weaker component of
an assembly can be a complicating factor, tearing of the PU layer being a possibility in
certain of the present measurements. In principle, the detachment of the adhesive in the
blister test is a pure interfacial event and one that can be quantified provided parameters
352
Adhesion Behaviour of Urethanes
such as adhesive thickness and the geometry of the detaching blister are known [11, 12].
The two test methods may therefore complement each other and the existence of
quantitative correlations between the test data is a point of interest. The theory is tested
by means of Figure 7.8. Although the results for PVC and Xenoy substrates are separated
by a significant margin, in both cases the correlations are very satisfactory. The correlation
coefficient for the more strongly bonded Xenoy-PU system is 0.988, that for the weaker
PVC-PU bonds is 0.935.
Thus it is concluded that for assemblies of the type used here, lap-shear and blister
detachment test methods produce sets of internally consistent data which correlate in a
statistically acceptable manner. The choice of test procedure, therefore, may rest on
convenience and on the proven relevance of test results to field performance.
As already noted, the adhesion test data suggests the existence of significant
contributions to bond strengths from non-dispersion forces acting at the adhesive/
substrate interface. Indeed, quantitative correlations have been confirmed between Isp
and either of the two sets of adhesion performance results. A demonstration is given in
Figure 7.9, using the more complete blister test data. The correlation coefficient for
this linear plot is 0.982.
353
Advances in Urethane Science and Technology
Figure 7.8 Correlation between lap-shear and blister detachment evaluations of bond
strengths in Xenoy/PU and PVC/PU assemblies.
The value is in fact surprisingly large when it is noted that no account has been taken of
contributions made to bond strengths by dispersion (van der Waals) forces. Acid/base
forces appear to play a determinant role when bonding occurs at interfaces of materials
with pronounced electron donor and acceptor tendencies. This alone seems to justify
concern for expressing these tendencies quantitatively, and for favouring the IGC technique
as a convenient route to the objective.
Another noteworthy feature in Figure 7.9 is the inclusion of all polymer-substrate pairs is
the relationship. This encourages speculation that Isp is an objective measure of the
interactions generated by electron donor and acceptor sites in the surfaces of the present
polymers. It also suggests that IGC determinations of Isp may be useful for screening purposes
in designing PU adhesives for optimum performance with specified polymeric substrates.
These suggestions form the origin of further research which is being carried out.
354
Adhesion Behaviour of Urethanes
Figure 7.9 The dependence of equilibrium bond energies for PU/polymer assemblies
on acid/base pair interaction. ● - PU1, ● - PU2, ▲ - PU3, ■ - PU4
One of the most frequent procedures in the formulation of adhesives is the addition of
adhesion promoters. Of these, silanes are by far the most frequently used. The silanes form
a multi-membered family of chemicals. Numerous publications have described aspects of
silane chemistry and the applications of silanes [15-18]. Although many silanes are readily
available, in many industrial formulations of adhesives the most frequently used of these
are the primary amine and diamine versions because of their relatively low cost. However,
these do not necessarily produce optimum benefits in adhesion for all applications.
Guidelines for the selection of silanes best suited for specific applications would therefore
be useful. These should be based on a fundamental understanding of the manner in which
silane compounds affect the interface and interphase in adhesively bonded systems. The
355
Advances in Urethane Science and Technology
last section relates to this objective by describing the effect of various silanes on the adhesion
performance of a model PU adhesive in joints involving selected polymers and glass. The
adhesion and its retention under aggressive ageing conditions is interpreted in terms of
acid-base interactions at the adhesive/substrate interface. The interpretative basis appears
suited for the selection of preferred silanes in the cases under study.
7.4.1 Experimental
7.4.1.1 Materials
The adhesive used was a model semi-structural PU, similar to one described in an earlier
work [7]. The two component formulation consisted of methylene bis-4-cyclohexyl
isocyanate prepolymer and trifunctional polyols with molecular masses in the range 500-
900. A small amount of tin catalyst (UL-28, Witco Chemicals) was also present. The PU
was used as control without additives and was modified by the following silanes:
The silanes were obtained from Dow-Corning Corporation and were added at 0.5 wt%
levels. All formulations were cured by a standard procedure of 1 hour at 100 °C, followed
by 5 days at 50 °C and 50% relative humidity.
As in the preceding section, the polymers to be bonded were PVC, ABS, and the thermoplastic
polyester blend Xenoy. In addition microscope slide glass was also studied. Surfaces of
these materials have been shown to be capable of widely differing acid-base interactions.
The PVC was an unplasticised polymer (MW = 54,000 g/mol) from Synergistic Chemicals,
Inc. It contained 5 phr of Advastab TM-821SP thermal stabiliser. The ABS (Cycolac AR
3501), was a moulding grade, commercial product of GE Plastics, Inc. The same source
supplied the Xenoy blend; the exact composition of this material was not determined.
7.4.1.2 Procedures
356
Adhesion Behaviour of Urethanes
with the polymer [5, 19]. The process involves the diffusion-controlled accretion of polar
moieties from the bulk to the surface layer of the polymer. It has been shown earlier in
this chapter that surface restructuring can be followed by monitoring γnd, the non-
dispersive contribution to the polymer surface energy, when a film, initially dried in air,
is immersed in a polar medium. Static contact angles were measured for n-decane,
methylene iodide and water, on a freshly prepared film specimen of pure PU, and also on
films of PU containing up to 2 wt% APS and ES. The harmonic mean procedure [20]
was used to compute γnd. The measurements were repeated after each of the films had
been exposed for 7 days to water at 90 °C. Preliminary work showed that this was
adequate to allow the PU surface to equilibrate with respect to the aqueous environment.
Whilst the silane-free film restructured in water to increase γnd by about 5 mJ/m2, the
presence of silane sharply reduced the effect. The silane apparently anchors surface-
localised moieties and restricts their ability to adopt new conformations. Of the silanes,
ES appeared to be more successful in inhibiting the restructuring effect, but both ES and
APS, when added at a concentration of 0.5 wt%, were found to greatly reduce
restructuring. This was, therefore, chosen as the concentration level in subsequent work.
For testing bond properties, single lap-shear specimens were assembled from polymer
and glass coupons cut following the practice outlined in reference [19]. The polymers
were moulded to a thickness of 1.5 mm and cut to 8 x 2.5 cm size. No mould release
agents were used in moulding operations. For convenience, the selected overlap area of
4.5 cm2 was bonded with adhesives drawn to a thickness of 0.4 mm. Following
conditioning (24 hours, room temperature), initial bond strengths were measured in
triplicate with an Instron apparatus at a draw speed of 10 mm/min. To ascertain property
retention, certain of the systems were exposed in air ovens to 60 °C, 100% relative
humidity for up to 3 weeks under an applied stress. To generate stress, small holes
were drilled near the ends of the lap-shear coupons, one of the holes being used to
suspend the test pieces from a horizontal beam within the oven. Consistent with certain
industrial procedures, stress was generated by a 200 g weight, suspended from the free
end of the joint. Repeat evaluations of lap-shear bond strengths were carried out after
various ageing times.
357
Advances in Urethane Science and Technology
was in the range 35-65 °C. Net retention volumes, Vn, were obtained from at least triplicate
vapour injections, generally with a reproducibility better than 4%.
-ΔHab/AN = KaDN/AN + Kb
Here AN and DN are Gutmann’s values [3] for the acidity and basicity of the vapour
probes. When Ka and Kb values for substrates and adhesives are known, an acid-base
pair interaction value Isp is calculated:
where subscripts 1 and 2 refer to adhesive and polymer (or glass) substrate, respectively.
In this work, Isp values were primary tools for the generation of correlations between
component interactions and the performance of the bonded systems. In so doing, the
relevance to adhesion of a parameter related to the acid-base contribution to the enthalpy
of component interaction is established.
The acid-base interaction potentials of polymer substrates and of the various adhesive
formulations are given in Table 7.9 as are the relevant values of Isp. The IGC experiments
show the model PU to interact as both mild acid and base, with these tendencies well
balanced. The presence of APS has no pronounced effect on the electron acceptor capacity
of the surface, but shifts the balance to basicity, presumably because of the surface
localisation of the silane’s primary amine groups. Each of the CS, VS and MS additives
produces the opposite trend, the adhesive surface now displaying stronger acidity, but
without significant change in the basic interaction potential. Addition of ES raises both
358
Adhesion Behaviour of Urethanes
Table 7.9 Interaction constants and pair interaction parameters for polyurethane
adhesives and substrates
A: Acid-base interaction constants
Material Ka Kb Ka + Kb
PVC 8.5 2.1 10.6
ABS 3.2 7.0 10.2
Xenoy 6.6 6.8 13.4
Glass 2.6 1.9 4.5
PU 3.1 3.8 6.9
+ APS 3.7 6.0 9.7
+ CS 4.9 4.0 8.9
+ MS 5.0 4.2 9.2
+ VS 4.6 3.9 8.5
+ ES 5.0 6.2 11.2
B: Pair interaction, Isp
PVC ABS Xenoy Glass
PU 39 34 46 16
+ APS 59 45 65 23
+ CS 44 47 60 20
+ MS 46 48 62 20
+ VS 41 45 57 19
+ ES 64 55 74 26
See text for definition of Silane codes
Reproduced with permission from H. P. Schreiber, R. Qin and A. Sengupta, Journal of
Adhesion, 1998, 68, 1-2, 31, Table 1. © 1998, Gordon and Breach Science Publishers.
the acid and base interaction potential with a predominance of basicity. Furthermore,
the total interaction potential, given by the sum Ka + Kb, is most affected by the ES
adhesion modifier. As expected, the PVC surface is strongly acidic, that of ABS is basic,
and Xenoy is amphoteric with excellent balance between donor and acceptor capabilities.
The glass surface is a weak acid. As a result of choosing the Gutmann concept of acidity/
basicity, all of the surfaces retain both acidic and basic interaction potentials.
359
Advances in Urethane Science and Technology
The resulting Isp data in Table 7.9, show ES and APS to be the most effective promoters
of interaction between adhesives and PVC. APS is not an effective interaction modifier
for ABS substrates; all the other surface modifiers exert a greater effect. Indeed, for each
of the four substrates in this work, ES produces the maximum increase of interaction,
although for glass the interaction level with these adhesives remains low when compared
with the polymeric substrates.
360
Adhesion Behaviour of Urethanes
The critical question of interdependence between initial lap-shear bond strengths and the
Isp parameters is addressed in Figure 7.10. Although each of the substrates tends to define
a specific envelope of behaviour (more particularly evident for glass), the first order regression
line drawn through all of the points has a correlation coefficient of 0.915, sufficient to
justify the claim of a true relationship. A more rigorous correlation is probably unrealistic,
since no account is taken here of varying chemical interactions between the silanes and the
substrates, or of variations in the contributions to bonding made by longer range dispersion
forces. Acid-base interactions at the adherent/adhesive interface thus make important
contributions to the initial lap-shear bond strength of these systems. The extrapolation to
Isp = 0 suggests that dispersion forces make a net contribution of about 12 J/m2 to the bond
property. Within experimental error (~10%) this is the initial bond strength for glass joined
by unmodified PU, even though an Isp value of 16 is reported for this pair in Table 7.7.
361
Advances in Urethane Science and Technology
Table 7.11 Effect of ageing (60 °C, 100% relative humidity) on lap-shear bond strength
Age time (h) 0 1.3 3.5 6.9 11 22.4 R
Bond strength (J/m2)
PVC/PU 27 22 17 14 14 9 0.33
+ APS 34 33 30 27 24 22 0.65
+ VS 31 28 23 21 17 14 0.45
+ ES 38 37 37 33 30 28 0.74
ABS/PU 20 18 16 12 8 7 0.35
+ APS 24 21 21 16 17 13 0.54
+ VS 28 25 24 22 20 19 0.68
+ ES 34 34 30 29 27 26 0.76
Xenoy/PU 28 23 19 16 16 12 0.43
+ APS 35 35 32 28 28 25 0.71
+ VS 31 29 27 24 19 17 0.55
+ ES 38 36 35 32 32 29 0.77
Glass/PU 14 12 7 0 0 0 0.00
+ APS 21 20 16 11 11 7 0.33
+ VS 18 17 12 9 7 4 0.22
+ ES 27 25 25 20 17 13 0.48
See text for definition of silane codes
R represents the ratio of final to initial bond strength
Reproduced with permission from H. P. Schreiber, R. Qin and A. Sengupta, Journal of
Adhesion, 1998, 68, 1-2, 31, Table 3. © 1998, Gordon and Breach Science Publishers.
Evidently at this level of Isp, the strength of specific interactions is insufficient to influence
the lap-shear bond strength as measured by the chosen procedure.
Of equal interest to the initial bond strength is the ability of the assembly to withstand
aggressive ageing. Joints using as adhesives the unmodified PU and the PU with 0.5 wt%
APS, VS or ES adhesion promoters were subjected to the ageing protocol of this work.
362
Adhesion Behaviour of Urethanes
Results of the procedure are presented in Table 7.11. Ageing periods are shown as the
square root of exposure hours, in the expectation that property loss would be a diffusion-
controlled process due to the intrusion to the interface of water. In Table 7.11, R is
defined as
where the subscripts f and i represent the bond strength after three weeks of ageing
and the initial value, respectively. Each of the systems is adversely affected, but R
varies significantly among the systems as well as within them. Qualitatively, a
relationship between R and the initial bond strength becomes apparent on inspection
of the data. The ES modified PU, which has the strongest acid-base interaction with
each of the substrates, produces the strongest initial bond, also produces the smallest
property loss.
A clearer view of the apparent correlation is presented in Figure 7.11. This indicates
that below initial bond strengths of about 13-14 J/m2, corresponding to Isp < 18-20,
the selected ageing regime will totally destroy the bonded assembly. The best fit of
the data is obtained from the second order regression curve shown, with a coefficient
of 0.932. However, a first order linear fit to the data (not shown) with a coefficient
of 0.917 is not much inferior and has the attraction of being extrapolatable to R =
1.0. If the linear fit is accepted as realistic, then the extrapolation predicts that an
assembly with an initial lap-shear bond strength > 45 J/m2 would withstand the chosen
ageing exposure indefinitely, without loss of bond strength. As the correlation in
Figure 7.11 makes evident, the degree of acid/base interaction at the bonded interface
also should correlate with the residual bond property. The matter is put to the test
with results shown in Figure 7.12. The resulting linear fit to experimental data has a
correlation coefficient of 0.938, with much of the scatter contributed by the glass
joints, where acid-base interactions are feeble and most subject to uncertain
determination. An extrapolation to R = 1.0, identifies Isp to be greater than or equal
to 85, as the strength of acid-base interactions required to ensure long-term resistance
to the selected ageing procedure. The forecasts should be useful in optimising the
choice of silane (and other) adhesion promoters in the formulation of adhesives such
as the present PU. However, the physico-chemical analyses on which these are based
cannot specify the chemical approaches to be followed. Moreover, the general
applicability of the physico-chemical analyses followed here calls for the study of a
broader range of adhesive/additive/substrate systems.
363
Advances in Urethane Science and Technology
7.4.3 Conclusions
1. A strong correlation has been shown to exist between initial bond strengths displayed
by polymer and glass surfaces bonded by PU adhesives containing a variety of silane
adhesion promoters and the acid-base interaction at substrate/adhesive interfaces.
2. Silane additives change the surface interaction potential of PU in distinct ways; APS
accentuates surface basicity, while CS, MS and VS accentuate surface acidity. The
addition of ES results in significant increases in both surface acidity and basicity.
3. Within the scope of the present work, ES was found to be the most effective adhesion
promoter for each of the substrates. APS was particularly useful with the acidic PVC
substrate, but the CS, MS and VS were preferred additives for the basic ABS substrate.
364
Adhesion Behaviour of Urethanes
4. A correlation has been developed between the residual adhesion of bonded assemblies
following accelerated ageing and the magnitude of acid-base interfacial interactions.
The strength of interactions needed to avoid property loss under the chosen ageing
conditions can be estimated. This capability represents a guideline to the selection of
preferred silane additives for the adhesion of PU adhesives to substrates with known
acid-base interaction potentials.
5. IGC was shown to be a convenient and effective tool for evaluation of acid-base
interactions at substrate and adhesive surfaces.
365
Advances in Urethane Science and Technology
References
10. J. M. Braun and J. E. Guillet, Advances in Polymer Science, Volume 21, Springer
Verlag, Berlin, 1976, p.108.
12. Y-S. Chang, Y-H. Lai and D. A. Dillard, Journal of Adhesion, 1989, 27, 4, 197.
13. C. Saint Flour and E. Papirer, Journal of Colloid and Interface Science, 1983,
91, 63.
16. E. P. Plueddemann, Silane Coupling Agents, Plenum Press, New York, 1982.
366
Adhesion Behaviour of Urethanes
17. D. E. Leyden, Silanes, Surfaces and Interfaces, Gordon and Breach Science, New
York, 1986.
21. J. Schultz, L. Lavielle and C. Martin, Journal of Adhesion, 1987, 23, 1, 45.
367
Advances in Urethane Science and Technology
368
8
HER Materials for Polyurethane Applications
Raj B. Durairaj
8.1 Introduction
Polyurethane elastomers have a wide range of industrial applications due to their unique
combination of valuable physical and mechanical properties. These elastomers, in
general, are composed of phase separated flexible soft segments and crystalline hard
segments. The soft segment is based on polyether or polyester type aliphatic diols and
the hard segment is the result of reaction between aromatic diisocyanate and low
molecular weight diol chain extenders [1]. The soft segments have low polarity and are
longer than the hard segments and provide a soft and flexible matrix. On the other
hand, the hard segments are shorter, highly polar and have a strong tendency to
aggregate. The thermodynamic incompatibility of these two segments leads to phase
separation, which ultimately influences the physical and mechanical properties of
urethane elastomers.
In the development of cast polyurethane elastomers, various chain extenders are used
with toluene diisocyanate (TDI) and 4,4´-diphenylmethane diisocyanate (MDI)
terminated prepolymers based on polyether or polyester polyols. With MDI-terminated
prepolymers, 1,4-butanediol is the most commonly used chain extender. In order to
maintain desired mechanical properties at elevated temperature and to improve hardness
and tear resistance, aromatic diols such as bis-(β-hydroxyethyl) ether of resorcinol
(HER) or hydroquinone (HQEE) are often used. One of the difficulties with HQEE is
its processing problem associated with a high melting point (~102 °C). On the other
hand, the structurally similar isomer HER offers a lower melting point (89 °C), thus
facilitating more forgiving chemistry and processing ease. Previous studies [2-5] have
shown that HER chain extender offers significant processing advantages over HQEE
and the properties of HER extended elastomers are comparable to those of HQEE
cured elastomers.
With the development of technical grade HER (HER TG), the processing temperature
of the HER was further lowered without sacrificing the physical and mechanical
properties of the final cured cast elastomers. The major difference between HER HP
and HER TG chain extenders is the presence of some higher molecular weight (MW)
reactive aromatic diol monomer in the technical grade material. With this unique
369
Advances in Urethane Science and Technology
product, HER TG, there is no need to add plasticisers, diluents or aliphatic diols to
lower the melting points of aromatic diol chain extenders. It is also important to note
that the presence of about 10 weight percent of this higher MW aromatic diol monomer
does not appear to affect the performance of the cast elastomers. With this discovery,
it was decided to investigate different grades of HER materials with varying levels of
higher MW monomer and to study their performance in cast elastomers based on MDI
terminated polyether prepolymer.
1. To develop low melting and highly processable aromatic diol extenders from resorcinol
and ethylene carbonate (HER materials).
3. To investigate how the high MW aromatic diol monomer affects the phase separation
behavior of polyurethane elastomers.
4. To examine the effect of various HER chain extenders on the physical, mechanical
and dynamic mechanical properties of cast polyurethanes made with MDI-terminated
prepolymers.
HER materials used as chain extenders in this work were prepared based on INDSPEC
Chemical Corporation’s patented technology and the product details are given in
Table 8.1.
8.2.2 Prepolymers
Polyether and polyester-based MDI prepolymers obtained from commercial sources were
used. The details are given in Table 8.2
370
Table 8.1 Development of HER materials for polyurethane applications
Extender Type HER-HP HER- HER- HER- HER- HER-LIQ
TG210 TG225 TG250 TG275
(TG-210) (TG-225) (TG-250) (TG-275)
Physical form Crystalline Molten Molten Waxy solid Pasty solid Liquid
solid solid solid
Viscosity (Pa-s, 25 °C) nd nd nd nd nd 20
Capillary melting point (°C) 88-92 78-90 76-85 65-78 63-65 nd
DSC analysis (endotherm)
Onset temperature (°C) 86.7 81.6 71 65.6 3 PEAKS nd
Peak temperature (°C) 89.1 85 80.6 71.7 35-65
HER content (wt. %, by GC) > 99 93 82 68 54 13
HER/H(OCH2CH2)nO-R-O(CH2CH2O)nH 100/0 93/7 83/17 68/32 54/46 0/100
(n = 2 and 3; Theoretical; wt. ratio)
HER/H(OCH2CH2)nO-R-O(CH2CH2O)nH >99 96/4 85/15 69/31 60/40 nd
( n = 2; by NMR; wt. ratio)
Hydroxyl number : Determined 555-560 545-550 525-535 500-520 470-480 379
: Theoretical, calculated 566 554 535 506 469 382
Status COM COM DEV DEV DEV DEV
COM: Commercial
DEV: Developmental
nd: Not determined
HP: High purity
TG: Technical grade
GC: Gas chromatography
HER Materials for Polyurethane Applications
371
Advances in Urethane Science and Technology
Table 8.2
MDI-Prepolymer Supplier % NCO
Polyester
1. Baytec MS-242 Bayer 6.6
2. Baytec MS-090 Bayer 9.0
Polyether
1. VIBRATHANE B 625 Uniroyal Chemical 6.4
2. Baytec ME-050 Bayer 5.8
All elastomers used in this project were prepared by the following standard procedure.
A known quantity of prepolymer was weighed in a glass jar. Separately, in another glass
jar, the chain extender was weighed to achieve 95 or 90% stoichiometry. Both jars were
then placed in a vacuum oven at the desired temperatures and held under vacuum until
the bubbles disappeared indicating that all the dissolved gases and any moisture, if present,
was removed. This process required about 1 mm Hg vacuum and 2.5 to 3.0 hours at
about 80-110 °C depending on the chain extenders (HER materials or HQEE) used for
casting. After degassing, both the prepolymer and extender were mixed thoroughly for
one minute using a stirrer to avoid bubbles. The mix was poured into a heated stainless
steel plate mould pretreated with Teflon or silicone mould release agent. After pouring,
the mould was placed in a programmable oven and cured for 16 hours. The samples,
after curing, were allowed to cool slowly to room temperature. Test specimens were cut
from the cured sheet for determination of tensile, tear strength and dynamic mechanical
analysis (DMA) and then stored for at least 7 days at 22 °C, 50% relative humidity
conditions before testing. Compression set samples were prepared separately, following
the degassing procedure similar to that for plate casting, and aged similarly before testing.
The following test methods were used to determine the properties of cast polyurethane elastomers:
• Pot life determinations were done using a Brookfield viscometer at controlled temperatures.
372
HER Materials for Polyurethane Applications
• Differential scanning calorimetry (DSC) analysis of the cast elastomers was performed
on a Perkin Elmer (DSC-7) instrument at a heating rate of 10 °C/minute under a
nitrogen atmosphere.
High purity (HP) and technical grade HER materials were synthesised by the reaction of
various ratios of resorcinol and ethylene carbonate. Previous studies had demonstrated
that the HER TG has a performance equal to that of the HER HP material [5]. It was
also observed that, in the cast polyurethanes, HER TG could be processed at a lower
temperature than the HER-HP. These interesting experimental facts prompted the
development of various HER-TG materials and an investigation into their effectiveness
as chain extenders to enhance the processing of cast and other polyurethanes. By varying
the molar ratios of resorcinol and ethylene carbonate, the technical grade materials such
as HER TG-210, HER TG-225, HER TG-250, HER TG-275 and HER LIQ were
developed. HER TG-210 is the same as HER-TG reported in prior work and in this
chapter. The major difference between the HP and TG materials is the presence of higher
MW reactive diol monomers present in the TG materials. From the resorcinol and ethylene
carbonate reaction, depending upon the excess of ethylene carbonate, two types of higher
MW diols can be expected:
HOH2CH2CO-R-OCH2CH2OCH2CH2OH
and
HOH2CH2C(OH2CH2C)n-O-R-O(CH2CH2O)nCH2CH2OH
373
Advances in Urethane Science and Technology
On the basis of theoretical calculations and also from the gas chromatography (GC)
analysis, hydroxyl number and nuclear magnetic resonance (NMR) analysis results, it
has been confirmed that the reactive high MW diol present in the technical grade materials
has the structure:
HOH2CH2C(OH2CH2C)nO-R-O(CH2CH2O)nCH2CH2OH
where n = 1 and 2
For HER TG-210, TG-225 and TG-250 materials, the reactive high MW diol has
predominantly n = 1 and is present at approximately 7, 17 and 32 weight percent,
respectively. With higher molar ratios of ethylene carbonate:resorcinol, high MW
diol with n = 1 and 2, having MW of 286 and 374, respectively, are observed to be
present. In HER LIQ, the HMW diol monomers were estimated to be as high as 87
weight percent.
It was suspected that the presence of these reactive diol monomers in the HER
could have several advantages. Since they have primary hydroxyl groups, their
reactivity towards isocyanate groups is similar to that of other reactive diol chain
extenders. The molecular weight of these high MW diols is still low enough that
they act as chain extenders and not as polyols. In addition, they can act as reactive
plasticisers in various polyurethanes and other polymers. Due to the presence of
this high MW material, the melting point of HER is lowered, thereby increasing its
processability in polyurethane applications. Laboratory observations indicated that
HER and other technical grade HER materials have supercooling tendencies. In
fact, HER TG-275 stays liquid at room temperature for a period of 2-3 months.
The TG-250 remains liquid at room temperature indicating that processing could
be easier with this material alone and it would be a good candidate for mixing with
other small chain extenders, such as butane diol, and polyols (polyether and
polyester) at lower temperatures.
The melting points determined by the capillary method and DSC analysis are given
in Table 8.1. The varied melting point determined by the capillary is associated with
the high MW diol impurity. To determine the melting point by DSC, measurements
were done at a heating rate of 5 °C/minute under a nitrogen atmosphere. HER-HP
showed one sharp endothermic peak. With HER TG-210, TG-225 and TG-250
materials, the major endothermic peak shifted to lower temperatures and also showed
one minor peak, indicating one impurity. For HER TG- 275, three endothermic peaks
were observed between 35-65 °C indicating three components. For this material, GC
analysis showed two major peaks in addition to the HER peak.
374
HER Materials for Polyurethane Applications
Before making the castings, pot life determinations were performed to determine the
processing window of the prepolymer and chain extender materials. As a standard practice,
both the prepolymer and chain extender were preheated to the desired temperatures,
degassed thoroughly, mixed at the desired stoichiometry (90 or 95%) and poured into a
thermostatted vial. The viscosity was measured, using a Brookfield viscometer, as a
function of time. The measurements were terminated when either the time of measurement
reached 20 minutes or the viscosity of the mix reached 0.1 Pa-s.
Figure 8.1 shows the viscosity-time plots of VIBRATHANE B 625 extended with HQEE
and various HER materials. As can be seen, the pot life temperatures for HER materials
are considerably lower than HQEE. The shorter pot life for HQEE is due to the higher
processing temperature. If the processing temperature were lower than 110 °C, then
HQEE would crystallise out from the solution. Premature crystallisation is a major
problem encountered by most users of HQEE. But these problems are eliminated with
HER-HP and other HER-TG materials due to their supercooling behaviour.
375
Advances in Urethane Science and Technology
The prepolymer % NCO content and the equivalent weight of the chain extender materials
were determined to calculate the amount of each material to achieve 95% stoichiometry.
After careful degassing and mixing, the material was poured into the stainless steel mould
and cured for 16 hours continuously at 110 °C. Table 8.3 summarises the experimental
details relating to cast elastomer preparation.
HER-HP has a purity of >99%. The other HER-TG materials contained at least one or
two high MW reactive diol monomers. The presence of these reactive diol monomers
can play a major role in the formation of hard segments, hard segment content and
phase separation behavior of the polyurethanes, which will ultimately dictate the final
cured physical and mechanical properties of the cast elastomer materials. A previous
study on HER-LIQ showed that the cast elastomers based on this extender produced
softer material compared to HER-HP [5]. The elastomeric materials developed using
HER-LIQ appeared to be more transparent indicating that the high MW diols present in
this chain extender were responsible for this change. The MW of these high MW diols
are still low enough to produce a hard segment in the polyurethanes, but the nature of
the hard segment can be different from that formed between the isocyanate and HER
due to a longer and more flexible chain.
Based on the extensive study made on HER and HQEE chain extenders, it is well known
that HER can produce hard segments with MDI and other isocyanate monomers, which
are crystalline in character [4]. Due to their high flexibility, the high MW reactive diol
monomers can produce hard segments which may be more amorphous than crystalline
in character. The amorphous hard segment formed can mix with the soft segment of the
polyurethane elastomers and ultimately change the physical and mechanical properties
of the elastomeric materials. Therefore, it was decided to calculate the amount of hard
and soft segment contents of the polyurethanes resulting from the VIBRATHANE B 625
prepolymer and HER materials containing various levels of the high MW reactive diols.
376
Table 8.3 Cast elastomers based on Uniroyal B-625 (polyether) prepolymer elastomers performance
with various HER-TG materials
HER Materials HQEE HER-HP HER- HER- HER- HER- HER-
TG210 TG225 TG250 TG275 LIQ
Prepolymer temperature, (°C) 110 10 0 90 90 90 90 80
Extender temperature, (°C) 110 100 90 90 90 90 80
Mix ratio (Prepolymer/extender) 100/14.6 100/14.5 100/14.9 100/15.5 100/16.1 100/17.2 100/20.7
Stoichiometry, (% theoretical) 95 95 95 95 95 95 90
Cure, (h/110 °C) 16 16 16 16 16 16 16
Tensile Properties (MPa)
100% Modulus 11.35 10.56 10.71 8.84 7.98 9.94 3.99
200% Modulus 13.04 13.71 14.57 11.77 11.35 10.44 6.63
300% Modulus 14.64 17.33 19.08 15.6 16.35 16.79 11.55
Tensile strength 3419 3816 3937 3712 4037 3682 2301
% Elongation 538 559 47 6 462 413 372 321
Energy to break, (kN/m) 67 79 70 53 49 40 20
Tear strength, DIE C, (kN/m) 127 11 5 102 97 86 72 44
Compression set, % 21 15.4 17.4 17 17.5 29.4 9.1
Bashore rebound, % 55 58 60 60 54 46 49
Hardness, Shore-A (D) 96 95 95 (48) 92 (44) 90 (43) 89 (40) 78 (30)
HER Materials for Polyurethane Applications
377
Advances in Urethane Science and Technology
and the high MW diols produce amorphous hard segments. Taking into account the
amount of HER and high MW diols present in the chain extender, for the 95%
stoichiometry, the crystalline hard segment, amorphous hard segment and polyether soft
segment contents were calculated for different cast polyurethanes made using the
VIBRATHANE B 625 prepolymer.
These calculations were based on the reaction scheme (8.1) shown opposite and an
idealised polyurethane structure from this scheme and the details are given in Table 8.4.
As can be seen from Table 8.4, the crystalline hard segment content is decreased from
48.7% to 35.4% when the high MW diol content increases from 0% to 87%.
Concurrently, the amorphous hard segment content increases from 0% to 16.6%. At the
same time, the soft segment content resulting from the polyol was affected only slightly
(changed from 51.3% to 48%). Since the high MW diol changed the concentrations of
the hard and soft segments, it is interesting to see how this will affect the physical and
mechanical properties of the cast elastomers.
In general, the contributions of both the soft and hard segments in the polyurethanes can
be correlated with the properties observed. The soft rubbery block primarily affects
resiliency, wear, tear, compression set and low temperature properties, while the hard
block affects hardness, modulus and tensile properties [11].
The Durometer hardness was measured after the cast elastomers were aged for at least 7
days at 23 °C and 50% relative humidity. The details are given in Table 8.3. From the
results, it is evident that as the crystalline hard segment content of the elastomer is
decreased, the elastomer hardness also decreases. By analysing the soft and hard segments
and their relation to hardness, it appears that the amorphous hard segment behaves like
the polyether soft segment.
The tensile data obtained from the cast elastomers are summarised in Table 8.3, plotted
and shown in Figure 8.2. From this figure, it can be seen that the 100% modulus and %
elongation decreased for the HER-TG materials compared to HQEE and HER-HP chain
extenders. On the other hand, the 200% and 300% modulus values were higher for
HER-HP and HER TG-210 extenders compared to HQEE and the values were lower for
the other technical grade materials. Except for HER-LIQ extender, all HER based
378
Table 8.4 Distribution of hard and soft segments in cast polyurethanes
Polymer-chain Extender Theoretical Calculation
HER Materials Analysis Hard Segment (%) Soft
HER Content H(OCH2CH2)nORO(CH2CH2O)nH Crystalline Amorphous Segment
(%) (%) (%)
OCN-R-NHCOO(Polyether)OOCHN-R-NCO + HO(H2CH2CO)n-R′-(OCH2CH2)n-OH
(Prepolymer) (n = 1-3, chain extender)
O-H2CH2CO-R′-OH2CH2C-OOCHN-R-NHCOO-(Polyether)-OOCHN-R-NHCOO(CH2CH2O)2-3-R′-(OCH2CH2)2-3-O
| Soft | | |
|— Crystalline Hard Segment —| Segment | |— Amorphous (Hard) Segment —|
379
Reaction scheme 8.1
Advances in Urethane Science and Technology
HER Materials for Polyurethane Applications
extenders showed either equal or better tensile strength values than the HQEE extender.
In general, the HER-TG materials showed good tensile properties.
HER-HP had a lower compression set than the HQEE extended elastomer. This value
increases when the concentration of high MW aromatic diols content is increased in the
HER-TG extenders. In fact, HER TG-275 extender with about 46% high MW aromatic
diol showed a high compression set of 29.4%. On the other hand, when this high MW
aromatic diol concentration reached 87% in HER-LIQ, the elastomer based on this chain
extender produced a very low compression set material.
380
HER Materials for Polyurethane Applications
The most important property to look for in choosing materials for applications such as
skate wheels is resilience or rebound [12]. The ether-based elastomers, in general, give
higher rebound resilience. The measured Bashore technique rebound values, show that
HER-HP, HER TG-210 and HER TG-225 extenders have slightly better resilience than
HQEE. The rebound values appeared to be lower for HER TG-250- and HER TG-275-
based elastomers. This value was higher for HER-LIQ (49%) when compared to TG-
275 (46%) extender. Since the rebound is generally measured at room temperature, its
value would relate to tan δ data given in Table 8.5. A comparison of the two indicates a
correlation between tan δ and the rebound values. Obviously, HER-LIQ extender showed
lower tan δ (0.044) compared to HER TG-275 (0.076) material.
The influence of the high MW diols present in the HER chain extenders on the transition
behavior of cast elastomers was studied by DSC analysis. Figure 8.3 compares the DSC
curves of HQEE, HER-HP and different HER-TG extended polyurethanes. The transitions,
which are observed to occur above room temperature, are associated with the melting of
Figure 8.3 DSC curves of cast polyurethanes from Vibrathane B-625-HER materials
381
Table 8.5 Cast elastomers based on Uniroyal B 625 (polyether) prepolymer elastomers performance
382
with various HER-TG materials
Dynamic Mechanical Analysis Results
HER Materials HQEE HER-HP TG210 TG225 TG250 TG275 HER-LIQ
Crystalline HS content (%) 48.7 48 46.9 44.4 43.1 35. 4
Amorphous HS content (%) 0 0.9 2.2 5.1 6.7 16.6
Storage Modulus, G′ (MPa)
at 25 °C 47 30 28 21 14 14 39
at 150 °C 43 22 19 14 9.8 8.4 3 .1
Thermal Stability (Temperature maximum for constant G′)
Temperature (°C) 170 160 160 160 160 155 160
G′, MPa x 108 39.5 19.9 17.4 12.6 8. 8 7.7 2 .9
Loss Modulus, G′′ (MPa)
Peak temperature, Tg (°C) -70.9 -70.8 -68.8 -67 -65.3 -63.5 -59.2
Advances in Urethane Science and Technology
hard segment domains. The melting behaviour of the hard segments can be seen as an
endothermic peak in the DSC curve. The hard segment melting temperatures and heats of
transitions observed from the DSC analysis are given in Table 8.6. From the heats of
transition data (ΔH), it is clear that when the hard segment content is decreased, the energy
associated with the melting of the hard segments is also decreased. In addition to this, the
presence of amorphous hard segments appeared to decrease the melting temperatures of
the crystalline hard segment. Since the temperature at which the first transition takes place
above room temperature in the DSC curve is considered to indicate the limit of the thermal
stability of the polyurethane elastomer, the thermal stability of these elastomers was found
to be decreased when the high MW diol contents were increased. This might be due to the
reduction of hard segment content. Higher hard segment transition in the DSC indicates
higher thermal stability of the resulting polyurethane.
383
Advances in Urethane Science and Technology
segment (polyether) but also the degree of phase separation between the hard and
soft segments. In this work, the DMA method was utilised to understand the effect
of HQEE, HER-HP and various HER-TG materials on the cast elastomers made
using VIBRATHANE B 625 (PTMEG-Polyether) prepolymer. The results are
summarised in Table 8.5.
The storage modulus, G´, values determined by the DMA for the elastomers extended
with HER chain extenders are given in Table 8.5, and the storage modulus curves are
shown in Figure 8.4. Storage modulus quantitatively measures the material’s elastic
properties and also qualitatively determines the elastomer’s stiffness and hardness.
Since the hardness of the elastomer is related to its hard segment content, the DMA
storage modulus was found to decrease as the hardness of the elastomers was decreased
due to the increasing amounts of high MW diols present in the HER-TG materials.
From the DMA curve, it was observed that the G´ was considerably decreased for
HER-LIQ extended elastomer which contained about 87% of the high MW diol
384
HER Materials for Polyurethane Applications
monomers. It has been observed and reported in the literature [13, 15] that elastomers
based on PTMEG polyol exhibit a secondary transition (additional dispersity) which
appears as a step in the temperature—shear modulus function curves. The same kind
of behaviour is observed for elastomers based on HQEE, HER-HP and HER TG-210
chain extenders at temperatures below 0 °C. A sharp transition in the G´ curve may
be associated with good phase separation of the elastomer. In this respect, a gradual
transition occurs with HER TG-225, TG-275 and HER-LIQ compared to HQEE,
HER-HP and HER-TG-210 extended elastomers indicating that poor phase separation
occurrs in these elastomers.
The thermal stability of urethane elastomers can be determined by the DMA method.
The temperature at which the storage modulus (G´) decreases significantly in the
rubbery region is considered to be the limit of thermal stability of the elastomers
[16]. From the G´ values of the elastomers, it was determined that HER extended
materials were stable up to 160 °C whereas HQEE extended elastomer showed a
10 °C higher stability.
In the temperature range between 0 and 160 °C, the loss modulus decreases into a long
plateau and then starts to increase when the temperature reaches beyond 170 °C and
180 °C. This might be due to the melting of the hard segments.
Figure 8.6 shows the plots of storage modulus (G´) and loss modulus (G´´) as a function
of various technical grade HER materials along with HQEE and HER HP measured at
25 and 150 °C. The curves clearly demonstrate that G´ and G´´ decrease as the hard
segment content is decreased in these elastomers.
385
Advances in Urethane Science and Technology
386
HER Materials for Polyurethane Applications
The measure of tan δ is a qualitative tool to determine the hysteresis or heat build up in
an elastomer during dynamic flex conditions. Figure 8.7 shows the plot of tan δ as a
function of temperature and the values obtained from this determination are shown in
Table 8.5. As the hard segment content is decreased, the tan δ peak height is increased
indicating higher hysteresis of the elastomers associated with the higher amounts of high
MW diols. The peak widths are also observed to increase, corresponding to the decrease
in the hard segment content of the elastomers. By comparing the tan δ peak (Tg) to that
of loss modulus peak (Tg) temperatures, the tan δ temperatures appeared to be higher.
The Tg associated with the soft segment were found to shift higher as the amorphous
hard segment content of the elastomers increased. This temperature shift might be
associated with the restriction of mobility of the soft segments due to phase mixing.
Also, the greater the amount of hard segment dissolved in the soft segment phase, the
higher the expected Tg of the soft segment.
387
Advances in Urethane Science and Technology
The Tg of the elastomers increased gradually when the hard amorphous content of
these materials was increased. Since this hard amorphous segment is the result of
reaction between the isocyanate and high MW aromatic diol monomers containing
ether linkages, there is a possibility that this amorphous hard segment can mix
effectively with the soft polyether polyol segments of the elastomers. As the amorphous
hard segment content is increased, it is possible that the effect of phase mixing between
this phase and the soft polyol segment is enhanced, thereby increasing the Tg of the
soft segment of the resulting polyurethane elastomers. The Tg of the HER-LIQ
extended elastomer is increased by as much as 26 °C with 16.6% amorphous hard
segment content. These results indicate that the high MW aromatic diol monomers
present in the HER-TG materials, influence the low temperature properties of the
cast polyether elastomer materials.
Loss compliance quantifies the heat generated during dynamic loading in applications
like tyres, wheels and rollers. Polyurethane elastomers with lower loss compliance
values will experience less heat build up and, consequently, may suffer fewer field
failures [14]. Figure 8.8 shows the graph of loss compliance as a function of
temperature for the cast elastomers made with HER extended materials. The results
are summarised in Table 8.5.
Loss compliance curves show peak maxima at the soft segment Tg. The values appear
to be lower for elastomers containing higher hard segment contents. In the case of
HQEE, HER-HP and HER-TG 210 extended elastomers, loss compliance curves showed
two peaks at the Tg. The two peaks appear to merge into a single peak as the amorphous
hard segment of the elastomers is increased. Based on the loss compliance values, it
appears that HQEE based elastomers show less heat generation than the HER materials
containing higher levels of high MW diols. However, HER-HP and HER TG-210 based
elastomers behave similarly and, therefore, comparable performance can be expected
in their applications.
In order to compare the Tg observed from the peak temperatures of loss modulus (G´´),
tan δ and loss compliance (J´´) curves, the peak temperatures are plotted as a function
of the amorphous hard segment and shown in Figure 8.9. As can be seen in the figure,
the amorphous hard segment produced by the high MW aromatic diol extenders has a
pronounced effect on the soft segment Tg which, in turn, ultimately dictate the low
temperature properties of the elastomers made from them.
388
HER Materials for Polyurethane Applications
389
Advances in Urethane Science and Technology
Polyurethane elastomers containing polyester soft segments are used in applications such
as wheels, rollers, gaskets, seals, sheet goods, belting, cyclone liners and forklift tyres
due to their high tensile strength, tear strength, abrasion resistance, oil resistance and
heat-aging properties.
The experimental details on the pot life determination and elastomer casting conditions
are given in Table 8.7.
Pot lives exceeding 20 minutes can be easily achieved with HER materials due to
their low melting and processing temperatures. Because of their supercooling
behaviour and tendency to stay liquid after cooling to room temperature, further
enhancement in the pot life can be achieved by lowering the mixing temperature of
TG-250 and TG-275 materials.
The data obtained from the tensile measurements are summarised in Table 8.7 and the
graphs are shown in Figure 8.10. From the results, it can be seen that 100% modulus
and %elongation are decreased and the tensile strength increased as the high MW diols
content of the chain extender is increased.
These properties are summarised in Table 8.7 and the values are plotted against elastomer
hardness and shown in Figure 8.11.
390
Table 8.7 Cast elastomers based on Bayer MS-242 (polyester) prepolymer
Effect of HER - HQEE Composition on Elastomer Properties
HER/HQEE Materials HQEE HER-HP HER-TG210 HER-TG225 HER-TG250 HER-TG275
Pot-life Determination
Temperature (°C) 110 100 85 85 85 85
Pot-life (minutes) 7.5 20 >20 >2 0 >20 >2 0
Pot-life viscosity (Pa-s) 10.1 10.8 4.7 6.15 5. 5 5. 3
Casting Conditions
Prepolymer temperature (°C) 110 100 85 85 85 85
Extended temperature (°C) 110 100 85 85 85 85
Mix ratio (prepolymer/extender) 100/14.7 100/15 100/15.1 100/16 100/17 100/17.6
Stoichiometry, % theoretical 95 95 95 95 95 95
Cure, (h/110 °C) 16 16 16 16 16 16
Cast Polyurethane (Calculated)
Hard segment (%) ND 37.3 36.6 35.4 33. 7 31. 8
Soft segment (%) ND 62.7 62.5 62.3 61. 8 61. 4
Amorphous hard segment (%) ND 0 0.9 2.3 4. 5 6.8
Tensile Property (MPa)
100% Modulus 11.53 11.13 10.55 8.9 7.09 5.19
200% Modulus 14.1 14.93 14.49 12.59 10.14 7.48
300% Modulus 16.39 18.8 19.13 17.3 15.62 10.73
Tensile strength 28.42 29.6 34.93 34.49 39.85 34.17
% elongation 726 699 579 499 45 8 43 0
Energy to Break (kN/m)
107 122 100 75 80 40
(area under stress-strain curve)
Tear Strength, Die C (kN/m) 148 144 137 110 10 4 77
Compression set (%) 26.2 30.4 32.4 51.9 55.5 51
Bashore rebound (%) 41 39 34 33 28 15
Hardness, Shore-A (D) 50 D 46 D 44 D 94 ( 40 ) 89 ( 39 ) 84( 34 )
HER Materials for Polyurethane Applications
391
Advances in Urethane Science and Technology
392
HER Materials for Polyurethane Applications
The hardness of cast polyurethanes is decreased as the high MW diols content of the
chain extender is increased. This eventually affects some of the most important elastomer
properties such as the tear strength, 100% tensile modulus and fracture energy. The
presence of high MW diols in the TG materials also increases the amount of amorphous
hard segment which, in turn, interferes with the phase separation of hard and soft
segments of the polyurethane elastomers. The continuous reduction in the rebound
property is associated with the formation of higher amorphous hard segments from
these TG materials.
During the DSC analysis, the elastomer sample is first heated to 225 °C (first heating),
cooled back to room temperature (cool down) and then heated again to 250 °C (second
heating) at 10 °C/minute heating and cooling rates. The endothermic transitions observed
above room temperature during the heating are associated with the melting of hard
segments and exothermic peaks are due to the crystallisation of the hard segment upon
subsequent cooling. DSC studies made on the segmented polyurethanes revealed three
hard segment-associated endotherms at temperatures of 60-80 °C, 120-190 °C and above
200 °C [18-21]. These endothermic transitions are due to short-range, long-range and
microcrystalline ordering of hard segment domains.
The endothermic peak temperature determined from the second heating is taken to be
the thermal stability limit of the elastomer. The hard segment melting and crystallisation
temperatures and their heats of transitions determined from the DSC analysis are given
in Table 8.8. Figure 8.12 shows the DSC curves obtained during the heating and cooling
of cast elastomers based on HER-HP and HER TG-250 chain extenders.
From the second heating endothermic peak temperatures, HER-HP extension resulted
in a thermal stability at 177 °C and the stability decreased to 135 °C for an elastomer
made from TG-275 extender. An interesting observation was also made in the DSC
analysis during cooling. An exothermic peak associated with the crystallisation of hard
segment was seen for the elastomers made from HER-HP, TG-210 and TG-225
extenders. Although this exothermic peak is not seen for TG-250 and TG-275 extenders
during the DSC cooling curve, this exothermic behaviour is seen during the second
heating. This strongly suggests that the higher amorphous hard segments formed from
the isocyanate and high MW diols interfere with the crystallisation of the crystalline
hard segment.
From the endothermic transition energy, HER-HP extension produced a higher hard
segment content compared to the other TG materials.
393
394
Table 8.8 DSC results on cast polyurethanes from MS-242-HER materials
First Heating Cooling Down Second Heating
Endotherm Exotherm Endotherm Exotherm
Peak Energy Peak Energy Peak Energy Peak Energy
HER/HQEE Material Peaks °C J/g °C J/g °C J/g °C J/g
1. HQEE Major 190 28 102 10 188 10 ND ND
Minor 201 218 9
2. HER-HP Major 188 33 83 14 177 13 ND ND
Minor 203, 211
3. HER-TG210 Major 189 23 73 10 173 7 ND ND
Advances in Urethane Science and Technology
Minor ND 140
4. HER-TG225 Major 185 15 66 6 168 6 ND ND
Minor 204 140
5. HER-TG250 Major 170 14 ND ND 143 4 71 6
Minor 215
6. HER-TG275 Major ND ND ND 135 3 74 5
Minor 217 2
ND = Not detected
HER Materials for Polyurethane Applications
The storage modulus (G´), thermal stability, Tg and loss compliance (J´´) properties
obtained from the DMA are given in Table 8.9.
Poly(ester urethanes) based on HER chain extenders behaved in the same way as poly(ether
urethanes). When the high MW diol content of the extender is increased, the storage
modulus is decreased due to the reduction of the hard segment content. As was seen
from thermal analysis, the thermal stability of the HER-HP extended elastomer is slightly
higher than the technical grade materials.
395
Advances in Urethane Science and Technology
396
HER Materials for Polyurethane Applications
Although the cast polyurethanes made using HQEE chain extender produce high
performance materials, processing problems are always encountered with this
material. In order to use HQEE, the processing temperatures must be as high as
120-130 °C. If the temperatures are lower than this, then ‘starring’ (after curing a
‘snowflake’ type material embedded in the continuous matrix, (cast elastomer) can
be seen) occurs due to the crystallisation of HQEE out of the elastomer system. To
overcome this processing problem, HQEE is often blended with various aliphatic
diols such as ethylene glycol, butanediol, and so on [22-24]. The addition of aliphatic
diols into HQEE, therefore, is expected to affect the high temperature physical and
mechanical properties of the resulting elastomers. These problems can be easily
overcome if an aromatic diol chain extender like HER is used in the place of aliphatic
diols. Due to the supercooling behaviour of HER, this extender offers a unique
opportunity to lower the melting point of HQEE without the customary degradation
of properties.
To study the eutectic nature of the HER/HQEE blends, freezing point determinations
were made. HER/HQEE blends were prepared at different weight ratios and melted
at elevated temperatures. Then the melt was allowed to cool slowly under controlled
conditions. The temperature at which the first crystallisation was observed was
considered to be the freezing point of the melt corresponding to that particular HER/
HQEE composition.
The plot of freezing point versus HER/HQEE blend is shown in Figure 8.13. As can
be seen from the plot, HQEE melting can be considerably reduced with the help of
an HER extender.
397
Advances in Urethane Science and Technology
The experimental details on the cast elastomers prepared using Baytec MS-242 and Baytec
ME-050 prepolymers and their physical, mechanical and dynamic mechanical properties
are given in Tables 8.10 and 8.11.
In the case of poly(ester urethanes), HER and HQEE extended elastomers showed
comparable tensile, tear, glass transition temperature (Tg) and hysteresis (tan δ) properties.
Compared to HER, HQEE extension produced elastomer with about 2 Shore-D units
higher hardness. An improvement in the tensile strength value was observed for the
HER/HQEE blend at 25:75 weight ratio. In general, all the physical and mechanical
properties were slightly affected or lowered with HER/HQEE blends compared to either
HER or HQEE extended material.
In the case of poly(ether urethanes), HQEE extended elastomer had slightly higher hardness,
tensile strength and tear strength values than the HER extended material. HER extension
produced elastomer with a lower compression set compared to HQEE extension.
DMA properties of HER and HQEE extended elastomers are similar. All the other physical
and mechanical properties of the elastomers from the HER/HQEE blends were lower
than the HER or HQEE extenders.
398
HER Materials for Polyurethane Applications
399
Advances in Urethane Science and Technology
400
HER Materials for Polyurethane Applications
High hardness polyurethanes are used in applications such as rolls and rollers. When
the hardness of the elastomer is increased, certain other physical and mechanical
properties such as the tensile strength, tear strength, compressive strength and tensile
modulus are also increased. To develop materials with the best possible properties
such as high hardness, solvent, oil and abrasion resistance, MDI-terminated prepolymers
with NCO contents of greater than 9% are often used.
Hard polyurethane elastomers, produced from the higher NCO containing prepolymers,
are generally more hydrolytically stable than the soft ones, since the hard segment is
hydrophobic and is less rapidly attacked [16]. With the high NCO containing
prepolymers, HER chain extended materials are expected to produce harder
polyurethane elastomers due to longer chain length than the conventional 1,4-butanediol
chain extender. On the other hand, the presence of a high MW diol chain extender
impurity in the HER-TG materials would affect the physical and mechanical properties
of the resulting polyurethanes.
The main objective of this work was to investigate and study the properties and
performances of hard cast polyurethane elastomers obtained from the HER chain
extenders.
8.7.1 Cast Elastomers and Their Hard and Soft Segment Contents
Cast elastomers were prepared using Baytec MS-090 prepolymer and HER chain
extenders with 95% stoichiometry. The experimental details of the elastomer castings
and the mix ratios of the prepolymer and chain extender are given in Table 8.12.
Since the cast urethane elastomer properties are greatly influenced by the soft and
hard segment contents and their phase separation behaviour, they are theoretically
calculated from the molar ratios of the prepolymer and chain extenders and are
reported in Table 8.12. Table 8.12 shows the hard segment content of the elastomer
is 45.5% for HER-HP and is reduced to 38.2% for TG-275. This reduction in the
hard segment content is due to the formation of 8.8% amorphous hard segment
from TG-275 chain extender.
401
Advances in Urethane Science and Technology
402
HER Materials for Polyurethane Applications
By correlating the elastomer hardness with that of other physical and mechanical
properties, a clear trend can be seen showing that as the hardness is decreased, the 100%
tensile modulus, fracture energy and tear strength values are also decreased.
On the other hand, a gradual increase in tensile strength and compression set values is
observed for cast elastomers containing more amorphous hard segment contents.
Polyurethane elastomers with higher hard segment content result in more hard segments
mixed into the soft phase [25-27]. Also, polyester soft segments are more compatible
with urethane hard segments. The decrease in Bashore rebound values of TG-225, TG-
250 and TG-275 extended materials is associated with the lower degrees of phase
separation in the polyurethanes.
In order to understand the effect of high MW diols on the formation of amorphous hard
segments and their interaction with the crystalline hard segments, FT-IR analysis of the
cast polyurethanes was performed. Samples were cast as films on a NaCl plate from a
hot dimethyl sulphoxide (DMSO) solution.
Figure 8.14 shows the C=O stretching vibration bands of FT-IR spectra of cast
polyurethanes obtained from the HER materials. The appearance of peaks at 1695 cm-1
and 1705 cm-1 strongly suggests hydrogen bonded C=O groups exist with all the HER
extended materials. The intensities of these two peaks are gradually reduced when the
hard segment content of the elastomer is decreased. The complete disappearance of the
1695 cm-1 peak might be associated with the strong interaction or interference of the
amorphous hard segment in the phase separation of elastomers.
IR analysis also showed a peak at 1736 cm-1 associated with the free C=O groups of the
elastomers. From the IR absorbances of bonded and free C=O groups, the hydrogen-
bonding index, ‘R,’ which is the ratio of Abonded C=O/A nonbonded C=O, can be calculated [30].
403
Advances in Urethane Science and Technology
The greater the value of R indicates an increase in participation of the C=O group in the
hydrogen bonding. The values of R calculated from the absorbance of hydrogen bonded
C=O (1705 cm-1) and free C=O (1736 cm-1) peak values are given in Table 8.13.
The values clearly suggest an increased hydrogen bonding with HER HP extended material
due to high hard segment content. The reason for the higher R value for TG-225 extender
is not known at the present time.
404
HER Materials for Polyurethane Applications
From the endothermic melting temperatures, HER-HP and TG-210 extended elastomers
show a thermal stability of 180 °C. Increase in hard segment content in the elastomer
results in an increase in hydrogen bonding between adjacent hard segments which may
result in higher transition energy. By comparing the energies of endothermic peaks observed
during the first and second heating DSC experiments, HER-HP extended elastomer has
the highest hard segment content. The presence of amorphous hard segment appeared to
affect the crystallisation behaviour, during cooling, and also the thermal stability of the
urethane elastomers.
Due to higher hard segment content, HER-HP and TG-210 extended elastomers had
high storage modulus (G´) compared to the other HER chain extenders. By comparing
the tan δ, Tg and loss compliance properties, it is seen that HER-HP and TG-210 extended
elastomers behaved similarly.
Butanediol chain extender has four aliphatic carbon atoms in the chain. On the other
hand, HER is an aromatic diol chain extender with the following structure:
HOH2CH2CO OCH2CH2OH
405
406
Table 8.14 DSC results on cast polyurethanes from MS-090-HER materials
First Heating Cooling Down Second Heating
Endotherm Exotherm Endotherm Exotherm
Peak Energy Peak Energy Peak Energy Peak Energy
HER/HQEE Material Peaks °C J/ g °C J/ g °C J/ g °C J/g
HER-HP Major 186 31 84 18 180 15 nd nd
Minor nd 144 1
HER-TG210 Major 193,180 25 72 15 181 20 nd nd
Advances in Urethane Science and Technology
Minor 16 4 152
HER-TG225 Major 188 19 58 9 175 8 57 1. 5
Minor 173,163 152
HER-TG250 Major 171 20 nd nd 148 6 71 10
Minor nd
HER-TG275 Major 166 10 nd nd 142 5 81 11
Minor 216 1
nd = Non-detected
HER Materials for Polyurethane Applications
Based on the structures of butanediol and HER, it can be expected that the HER-based
hard segment would have a longer chain length than the butanediol-containing hard
segment. The domains containing these longer hard segments are also expected to show
a higher melting temperature than the shorter ones. Therefore, HER based elastomers
are expected to be more thermally stable than the butanediol extended material.
407
Advances in Urethane Science and Technology
The Durometer hardness measured on the elastomers before and after the heat aging is
given in Table 8.16. Since hardness is the measure of hard segment content of the elastomer,
the unaged HER elastomer showed a 96A (Shore) hardness compared to 88A for the
butanediol elastomer clearly suggesting a higher hard segment content with HER
extension. On heat ageing, the HER elastomer maintained its hardness even after 21 and
14 days of aging at 120 °C and 135 °C, respectively. The butanediol extended elastomer’s
hardness was reduced by 4 Shore A units at all heat-ageing temperatures suggesting
material deterioration during the heat treatment.
Polyurethane elastomers are known for their high elongation, tensile strength and modulus
properties. The combination of these properties provides toughness and durability in
fabricated parts. Cast elastomers extended with butanediol can maintain these tensile
properties when the use temperature is about 80 °C. When these elastomers are subjected
to higher temperatures, reduction in the tensile properties are observed due to the
weakening of physical bonds in the elastomer.
The tensile properties measured on the unaged and heat-aged elastomer materials made
from HER and butanediol are given in Table 8.16.
The 100, 200 and 300% tensile modulus values of unaged elastomers are higher for
HER than for the butanediol elastomer. Analysing the 100, 200 and 300% modulus
values of 100 °C heat-aged samples showed that HER elastomer retained about 95-97%
of its original unaged values compared to 75-76% for butanediol. At 120 °C, there was
93-94% modulus retention for HER whilst 85-88% was observed for butanediol. When
408
Table 8.16 Heat ageing and mechanical properties of cast elastomers from Baytec MS-242
HER-HP 1,4-Butanediol
Heat Ageing Conditions
Temperature (°C) RT 100 120 135 RT 100 120 135
Days at ageing None 28 21 14 None 28 21 14
Hardness, Shore-A 96 94 96 96 88 84 84 84
Tensile property (MPa)
100% Modulus 10.96 10.38 10.35 9.86 6.58 4.93 5.34 5.0
(100) (95) (94) (90) (100) (75) (81) (76)
200% Modulus 14.53 13.75 13.66 12.84 9.51 7.35 8.04 7.37
(100) (95) (94) (88) (100) (77) (85) (77)
300% Modulus 18.46 17.35 17.17 15.97 15.04 11.40 13.24 11.38
(100) (94) (93) (88) (100) (76) (88) (76)
Tensile strength 28.19 27.28 26.15 21.95 46.72 38.80 42.33 35.61
(100) (97) (93) (78) (100) (83) (90) (76)
% Elongation 619 691 635 646 534 642 600 619
Energy to break (kN/m) 110 124 111 104 92 100 98 91
(area under stress-strain curve)
Tear strength, Die C (kN/m) 142 nd nd nd 105 nd nd nd
Note: The values in the parenthesis are the percentages of properties retained after the heat ageing.
nd: Not determined
HER Materials for Polyurethane Applications
409
Advances in Urethane Science and Technology
the aging temperature is increased to 135 °C, HER elastomer showed better modulus
retention property than the butanediol (87% versus 77%).
Although the butadiene elastomer has an unaged tensile strength of 46.72 MPa, this
value drops to 38.80 MPa (83% retention) when this material is heated for 28 days at
100 °C. On the other hand, the HER extended elastomer retains 97% of its original
(unaged) tensile strength under the same temperature heat-aging conditions, indicating
its excellent high temperature stability.
In spite of the severe aging conditions, HER extended elastomer maintained its excellent
(greater than 600%), tensile elongation property. The combination of higher modulus
and good elongation of the HER-based elastomer clearly demonstrates its ability to
withstand higher tensile loads without the risk of failure.
Tensile measurements can also be used to determine the fracture energy of the material.
The area under the stress-strain curve is the measure of fracture energy which is recognised
as a measure of toughness. As can be seen in Table 8.16, HER extended elastomer showed
higher fracture energy before and after heat aging suggesting that it has a higher toughness
than the butanediol elastomer.
Tear strength is a property commonly used to determine the cut growth propagation in
an elastomeric material. The unaged tear strength is much higher for HER than the
butanediol elastomer. The combination of fracture energy and tear strength properties
clearly predicts that the HER elastomer will have better wear resistance than the butanediol
extended elastomer.
From the DSC analysis, the thermal stability of polyurethanes can be determined by the
endothermic transition temperatures associated with the melting of hard segment domains.
The results are summarised in Table 8.17 and also shown in Figure 8.15. The first heating
DSC endothermic peaks for HER extended elastomer are broader and appeared at higher
temperatures than the butanediol elastomer. Higher endothermic peak energy suggests
that larger size crystalline hard segment domains exist within the HER elastomer.
Considering the molecular structures of HER and butanediol chain extenders, there is
no doubt that the higher thermal stability of the HER elastomer is associated with the
formation of large size hard segments from the aromatic isocyanate and extender reaction.
On cooling the heated samples, exothermic peaks, due to the crystallisation of hard
segment domains, were seen in the DSC curves. Again, the higher exothermic peak
temperature suggested the existence of larger size hard segments for HER elastomer.
410
HER Materials for Polyurethane Applications
During the second heating, DSC curves showed that a well-defined endothermic melting
peak appeared at higher temperature confirming the higher HER elastomer thermal
stability. Based on the DSC results, HER extension produced about 20 °C higher hard
segment melting temperature indicating the superior high temperature properties of HER
compared to butanediol elastomer.
411
Advances in Urethane Science and Technology
Figure 8.15 DSC curves of 135 °C heat aged urethane elastomers from Baytec MS-242
DMA measures the ability of the material to store and dissipate mechanical energy. DMA
properties of the elastomeric materials are important because they often correlate to
actual field performance. This method determines the storage modulus (G´, elastic
behaviour), loss modulus (G´´, energy dissipation), tan δ, loss compliance (J´´) and Tg
values. In this work, the DMA method was effectively utilised to understand the
advantages of HER over the butanediol extended elastomer. The results are summarised
in Table 8.18.
Storage modulus quantitatively measures the material’s elastic properties and also
qualitatively determines the elastomer’s stiffness and hardness. The unaged elastomer
storage modulus values are higher for HER at 25, 100 and 150 °C temperatures compared
to butanediol. On heat ageing at 100 °C, the HER elastomer retained about 85-86% of
its original G´ value compared to 60-85% for butanediol. After ageing for 28 days at
412
Table 8.18 DMA of heat aged cast polyurethane from Baytec MS-242 prepolymer
HER-HP 1,4-Butanediol
Ageing Temperature (°C) RT 100 120 135 RT 100 120 135
Storage Modulus, G′ (MPa)
at 25 °C 33.1 28.9 30 29.4 10.9 9.3 9.7 9.1
at 100 °C 21.2 18 18 12.8 7.2 5.1 5.3 3.9
at 150 °C 14.4 12.4 12.7 8.9 6.1 3.7 3.9 2.8
Loss Modulus, G′′ (MPa)
Peak temperature, Tg (°C) -29.8 -29.6 -30 -29.6 -27.4 -25.6 -28 -26
Peak modulus 171 168 167 165 167 117 164 159
Tan Delta
Peak temperature, Tg (°C ) -23.9 -23.7 -21.8 -23.6 -19.3 -16 -17.8 -17.7
Peak tan delta 0.392 0.404 0.401 0.386 0.657 0.628 0.65 0.623
at 25 °C 0.072 0.08 0.083 0.091 0.051 0.087 0.08 0.103
at 100 °C 0.019 0.032 0.033 0.064 0.024 0.081 0.064 0.097
at 150 °C 0.029 0.052 0.051 0.08 0.027 0.075 0.066 0.091
Loss compliance, J′′ (MPa)
Peak temperature, (°C) -11.6 -12.1 -9.4 -13 -9.3 -6 -7.7 -3.6
Peak compliance 2580 3000 2810 2600 1460 1640 1530 15100
at 25 °C 2160 2770 2750 3090 4720 9220 8210 11200
at 100 °C 900 1780 1850 5000 3350 1580 1190 25100
at 150 °C 2030 4210 4040 8940 4450 20000 1690 32000
Peak area: 0 —> 150 °C 22 34 35 67 57 190 160 292
Peak area: -50 —> 150 °C 28 41 41 73 82 216 185 316
HER Materials for Polyurethane Applications
413
Advances in Urethane Science and Technology
100 °C temperature, a very sharp decline in G´ values (60%) for butanediol and a high
retention for HER were observed at 150 °C.
For 120 °C heat-aged samples, 85-90% retention of G´ values for HER compared to 63-
89% for butanediol were seen. The same kind of trend is seen with 135 °C heat-aged
elastomer storage modulus values. Thus the HER extended elastomer has better retention
of storage modulus at higher temperatures than the butanediol elastomer.
Loss modulus measures the energy dissipation in the elastomer. The Tg measured from
the loss modulus curves (Table 8.18) showed lower values for the HER than butanediol
elastomer suggesting good phase separation.
The soft segment Tg is often defined as the maximum in the tan δ versus temperature
curve which determines the low temperature behavior of urethane elastomers. The results
obtained are summarised in Table 8.18. The Tg obtained from the tan δ curves confirms
again that a good phase separation existed with HER elastomer. As a consequence, the
HER-based polyurethane has a higher flexibility at low temperature.
Although the unaged 25 °C tan δ value appeared to be higher for HER, the effect of heat
ageing showed an increase of 26% compared to 101% for BD. Similarly, the 100 °C and 135
°C heat-aged samples showed very high percentages of elevated temperature (100 °C and
150 °C) tan δ value increases in butanediol-based elastomer as compared to the HER material.
These experimentally determined tan δ values from the DMA analysis strongly suggest
that more heat generation problems can be anticipated with butanediol-based elastomer
when they are exposed to temperatures above 80 °C than with HER elastomer.
Figures 8.16 and 8.17 show the plots of loss compliance for butanediol and HER
elastomers, respectively and the results are summarised in Table 8.18.
The loss compliance values of unaged and heat-aged butanediol extended elastomer are
about 3-5 times greater than the HER elastomer.
414
HER Materials for Polyurethane Applications
415
Advances in Urethane Science and Technology
In addition to the loss compliance data at various temperatures, the area under the loss
compliance curve can also be used to predict the tendency of the elastomeric materials to
convert mechanical energy to heat. A smaller area means less mechanical energy is
converted to heat, indicating better dynamic performance of the elastomer over the
temperature range of interest.
By comparing the loss compliance peak areas of heat-aged samples, there is no doubt the
HER extended elastomer is far superior in performance to butadiene elastomer.
Based on the current study, HER extended elastomers has the following room temperature
and high temperature physical and mechanical properties advantages over butanediol-
based elastomers:
8.9 Conclusions
The HER materials developed for polyurethane and other applications have a wide
range of melting characteristics. Lower melting temperatures of these materials can
provide improved processability in the development of cast and thermoplastic
polyurethane elastomers. The high MW diols present in the HER materials not only
reduce the melting point of the HER extender, but also exhibit a profound influence on
the concentration of crystalline hard segments and, as a result, ultimately affect the
physical and mechanical properties of the elastomers. Polyurethanes with varying
properties can be obtained by using HER materials containing different concentrations
of high MW diols. Since the high MW diols can act as plasticisers, the addition of
external plasticisers can be avoided by use of these materials. Cast polyurethanes show
good tensile, tear, compression set and hysteresis properties with HER-based chain
extenders. This results in a wide choice in the selection of a suitable aromatic diol
chain extender from the HER materials family.
416
HER Materials for Polyurethane Applications
Acknowledgements
The author would like to thank Jason Burchianti and Stephen J. Ondrey for preparing
elastomer samples and for conducting pot life measurements, Richard F. George for the
DMA, physical properties and DSC work and Vaughn J. Romell for the FT-IR analysis
work. My special thanks to Fred M. Covelli, V.P. Technology, and Joseph de Almeida,
Marketing Manager of Specialty Products, for their continuous encouragement in this
work.
References
3. J. E. Tiedemann, Presented at the Fall PMA Meeting, Miami, FL, USA, 1993.
6. ASTM D2240-00
Standard Test Method for Rubber Property-Durometer Hardness.
7. ASTM D412-98a
Standard Test Methods for Vulcanized Rubber and Thermoplastic Rubbers and
Thermoplastic Elastomers-Tension.
8. ASTM D624-00e1
Standard Test Method for Tear Strength of Conventional Vulcanized Rubber and
Thermoplastic Elastomers.
9. ASTM D2632-96
Standard Test Method for Rubber Property-Resilience by Vertical Rebound.
417
Advances in Urethane Science and Technology
12. R. Palinkas, Paper Presented at the Fall PMA Meeting, Cincinnati, OH, USA, 1990.
13. J. E. Doyle, Paper Presented at the PMA Meeting, Milwaukee, WI, USA, 1994.
14. S. M. Clift, Paper Presented at the Fall PMA Meeting, Charlotte, NC, USA, 1991.
15. J. R. Lin and L. W. Chen, Journal of Applied Science, 1998, 69, 8, 1575.
21. G. Pompe, A. Pohlers, P. Potschke and J. Pionteck, Polymer, 1998, 39, 21, 5147.
23. M. Palmer, Paper Presented at the Spring PMA Meeting, Kansas City, KY, 1983.
418
HER Materials for Polyurethane Applications
419
Advances in Urethane Science and Technology
420
9
Ultra-Low Monol PPG: High-Performance
Polyether Polyols for Polyurethanes
Stephen D. Seneker, Nigel Barksby and Bruce D. Lawrey
9.1 Introduction
Polyoxypropylene glycols (PPG) have been available since the early 1960s.
Polyurethanes based on these polyols are used in a diverse range of applications
including flexible foams (moulded and slabstock), sealants and adhesives. Until
recently, their use in high-performance applications such as cast and thermoplastic
urethane elastomers has been limited due to their inferior physical properties and
processing characteristics. These problems can be largely attributed to the considerable
amount of monol (monofunctional hydroxyl-containing species) present in these
polyols. Monol acts as a chain terminator, which severely reduces both the molecular
weight (MW) build during polymer formation (processability) and the ultimate
polymer MW (physical properties).
421
Advances in Urethane Science and Technology
Figure 9.1 Reaction of propylene oxide to form polyoxypropylene diol and monol
422
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
In this chapter, the dramatic effect that monol content or polyol functionality has on
processability and properties of polyurethane cast elastomers is discussed. This effect is
shown for elastomers prepared by both the prepolymer and one-shot processes. Further
improvements in elastomer processability and formulating latitude can be achieved by
incorporating oxyethylene moieties into the polyol backbone. For one-shot elastomer
processes, ultra-low monol PPG polyols capped with ethylene oxide have been
commercialised.
423
Advances in Urethane Science and Technology
This chapter also focuses on the effect of the polyol molecular weight distribution (MWD)
or polydispersity on mechanical and dynamic properties of polyurethanes and
polyurethane/ureas. Ultra-low monol PPG polyols are unique in that they have a very
narrow MWD in comparison to other high-performance polyols such as
polytetramethylene ether glycol (PTMEG) or polyester polyols. Narrow polydispersity
results in lower viscosity polyols as well as lower viscosity isocyanate-terminated
prepolymers based on those polyols. However, a not so obvious effect is that polyols
with a narrow MWD result in polyurethanes with significantly different mechanical and
dynamic property profiles. These differences will be highlighted in various polyurethane
and polyurethane/urea systems and explained using dynamic mechanical thermal analysis
(DMTA). By understanding the effect of polyol MWD, we will illustrate how the
polydispersity can be adjusted in order to maximise the mechanical and dynamic properties
obtained from ultra-low monol PPG polyols.
The importance of monol content in 4,4´-MDI/BDO cured elastomer systems was determined
by comparing 4000-MW PPG diols prepared via ultra-low monol technology, DMC and
potassium hydroxide. They are designated as ultra-low monol, low monol and conventional
and have monol contents of 0.005, 0.016 and 0.085 meq/g, respectively. This corresponds
to functionalities of 1.98, 1.94 and 1.71. We prepared 6% NCO 4,4´-MDI prepolymers
and chain extended with BDO at an isocyanate to hydroxyl ratio (NCO:OH) of 1.03 [15].
Table 9.3 summarises the monol effect on elastomer processing characteristics (pot life
and demould time), and physical properties.
The monol content has a major effect on elastomer processability. Lower monol content
results in a faster molecular weight build, which reduces demould time. Elastomers based
on the 4000-MW, ultra-low monol PPG had demould times as short as 20 minutes,
whereas, the low monol PPG gave demould times of over one hour. The elastomer based
on conventional polyol had a demould time greater than three hours.
Lower monol content clearly has a positive effect on all elastomer physical properties.
Dramatic property improvements are even seen when comparing elastomers based on
ultra-low monol and low monol polyol. This is particularly surprising considering the
424
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
425
Advances in Urethane Science and Technology
small difference in functionality (1.98 versus 1.94). For example, Shore A hardness
increased from 68 to 72; rebound from 62 to 70%; tensile strength from 12 to 21
MPa; Die C tear strength from 49 to 67 kN/m; and compression set decreased from 40
to 25%. The load bearing capabilities improved significantly as shown by the
compression deflection results. The stress/strain curves in Figure 9.2 clearly show that
the elastomer based on ultra-low monol PPG exhibits much greater modulus build
with increasing elongation.
426
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
Reducing the monol content also improves the dynamic properties of cast elastomers.
Figure 9.3 shows the DMTA response for the two polymers described above. The polymer
derived from the ultra-low monol polyol has a flatter rubbery plateau region. The higher
modulus in the rubbery plateau is consistent with the polymer’s higher tensile modulus.
The substantial reduction in tan delta (δ) across the entire temperature range should also
be noted. Lower tan delta translates into improved performance in dynamic applications
due to lower heat buildup and improved rebound as noted above.
427
Advances in Urethane Science and Technology
All these improvements indicate that significant changes are occurring in the elastomer
as the monol content is reduced and the polyol functionality increased from 1.94 to
1.98. We believe the benefits are directly attributable to an increase in the polymer
molecular weight. Figure 9.4 shows the relationship between functionality and theoretical
elastomer molecular weight (calculated from the Carothers equation). Raising the
functionality from 1.94 to 1.98 increases the ultimate polymer molecular weight by a
factor of three from 82,000 to 270,000. This increase in MW could easily account for
the observed improvements in processability and properties.
428
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
Ultra-low monol PPG diols are commercially available in molecular weights from 2000
to 12,000. A product summary is given in Table 9.4. To determine the formulating range
of the various diols in MDI/BDO cured elastomers, we prepared MDI prepolymers based
on the 2000, 4000 and 8000 MW diols at a wide range of isocyanate contents. These
prepolymers were chain extended with BDO to produce elastomers with a broad range
of properties. Tables 9.5 to 9.7 summarise the prepolymer compositions and viscosities,
processing characteristics and physical properties of these elastomers.
429
Advances in Urethane Science and Technology
430
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
431
Advances in Urethane Science and Technology
Table 9.7 Elastomers based on 8000-MW ultra-low monol PPG: Effect of 6000-
MW ultra-low monol triol addition (MDI prepolymers; BDO cured)
PREPOLYMER COMPOSITION
% NCO Prepolymer 2. 0 3.0 4.0
NCO/OH 2.9 3.1 4.0 4.2 5. 1 5.5
4,4´-MDI (pbw) 9.9 9.6 13.5 13.2 17.4 17.1
6000-MW PPG triol (pbw) 2.6 — 2.6 — 2.6 —
8000-MW PPG diol (pbw) 97.4 100 97.4 100 97.4 100
Prepolymer Viscosity, Pa-s
20 °C 72 50 40 35 24 22
40 °C 18 12 10 9 6.4 5. 9
60 °C 6.7 4.5 3.7 3. 3 2.4 2.3
80 °C 3.2 1.7 1.6 1. 5 1.1 0.83
PROCESSING Prepolymer @ 65 °C; BDO @ RT; NCO/OH = 1.03; dibutyltin
CHARACTERISTICS dilaurate catalyst; Moulds @ 100 °C; Cured 16 hours @ 100 °C
Prepolymer (pbw) 100 100 10 0 100 100 100
1,4-butanediol, BDO (pbw) 2.1 2.1 3.1 3.1 4.2 4.2
Pot life (minutes) 1-3 1-3 1- 3 1-3 1-3 1-3
Demould Time (minutes) 30-60 >180 20-45 20-45 20-45 20-45
ELASTOMER PROPERTIES Samples conditioned for 4 weeks @ 23 °C/50% humidity
Hardness, Shore A 34A 27A 49A 45A 58A 56A
(32A)* (16A)* (41A)* (53A)*
Pendulum Rebound, % 59 55 68 65 71 69
Ultimate Elongation, % 1030 >1300** 1130 1190 920 960
Tensile Strength, MPa 4.0 2.9 12.7 10.1 13.2 11.8
100% Modulus, MPa 0.5 0.5 1. 2 1.2 1.7 1.6
200% Modulus, MPa 0.7 0.5 1.8 1.7 2.5 2.3
300% Modulus, MPa 0.8 0.6 2.4 2.2 3.4 3.1
400% Modulus, MPa 1.0 0.7 3. 1 2.9 4.4 4.0
Die C Tear, kN/m 16 13 42 42 49 47
Compression Set, %
34 53 22 40 19 21
(22 h @ 70 °C)
Compression Deflection, MPa
5% 0.1 0.1 0. 2 0.2 0.3 0.3
10% 0.3 0.3 0.4 0. 4 0 .6 0.6
15% 0.4 0.4 0.6 0. 6 1.0 0.8
25% 0.7 0.7 1.0 1. 0 1.7 1.5
*
Shore Hardness needle slowly penetrated into the sample to the value indicated in parenthesis
**
Samples did not break during testing
432
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
A review of the elastomer processability and physical property data indicates that there
is an optimum hardness range for each polyol MW. Based on the demould time and the
percent compression set, the following elastomer Shore hardness ranges are recommended.
The lower hardness limits of these ranges can be extended by addition of low levels (less
than 10%) of 6000-MW, ultra-low monol triol (Acclaim Polyol 6300).
The rebound of the elastomers based on 2000-MW, ultra-low monol PPG ranges from
55 to 60%. For high-performance dynamic applications, use of higher MW diols are
recommended, which give rebound values up to 10 units higher, ranging from 65 to
70%. Figure 9.5 shows the DMTA response for the 70 Shore A elastomers based on the
2000, 4000 and 8000 MW diols. The glass transition temperature (Tg) for the elastomers
based on the higher MW diols is about 15 °C lower, which probably accounts for the
higher rebound of these materials.
Figure 9.5 DMTA curves of 70 Shore A elastomers: Effect of polyol molecular weight
(6% NCO MDI prepolymers; BDO cured)
433
Advances in Urethane Science and Technology
The 8000-MW, ultra-low monol PPG is used particularly in the development of soft,
plasticiser-free elastomers [16, 17]. The use of this high MW, low polydispersity polyol
allows for the preparation of low viscosity prepolymers with low isocyanate contents.
Initial evaluations of these very soft elastomers showed lower than expected physical
properties. The lower properties can be attributed to the very low hard-segment content
of these polymers. The lack of hard segment (physical crosslinks) can be compensated
for by the incorporation of low levels of chemical crosslinks (triol) [18]. This is
accomplished by the addition of a 6000-MW, ultra-low monol triol (Acclaim Polyol
6300) into the polymer matrix. Table 9.7 shows the triol effect on prepolymer viscosities,
elastomer processability and physical properties. It should be noted that very low levels
of crosslinking are needed to improve the elastomer properties.
The ultimate goal for MDI/BDO cured elastomers was to design a polyol that would
have a combination of good processability and high rebound in the mainline hardness
range of 80 to 95 Shore A. This was an elusive goal as one can see from the elastomer
data in Tables 9.5 and 9.6. A 90 Shore A elastomer based on the 4000-MW diol had
poor processability but a high rebound, whereas, the 90 Shore A elastomer based on the
2000-MW diol was processable but had a lower rebound.
We attributed the poor processability of the 90 Shore A elastomer based on the 4000-
MW diol to the excessive phase-separation of the MDI/BDO hard segment [19]. To
alleviate the hard segment phase-separation problem, the compatibility of the MDI/BDO
hard segment in the soft-segment phase was increased by incorporating oxyethylene
moieties into the polyol backbone. This technique dramatically enhanced the elastomer
processability. The polyol MW and oxyethylene content were varied to obtain the best
possible combination of processability, high hardness and high rebound. This optimisation
process led to the development of Acclaim Polyol 3205, a 3200-MW diol containing
approximately 20 weight percent oxyethylene moieties. Table 9.8 shows the enhanced
processing latitude of elastomers based on this polyol. In addition to enhanced
processability, elastomer physical properties were improved. Acclaim Polyol 3205 is a
unique material in that it allows the formulation of MDI/BDO cured elastomers from 75
to 95 Shore A with excellent processability and high rebound.
434
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
435
Advances in Urethane Science and Technology
The lower reactivity of secondary hydroxyl groups on the initial ultra-low monol PPG
polyol products limited their use to mainly prepolymer systems. PTMEG and polyester
polyols have primary hydroxyl groups, allowing them to be easily formulated with
curatives such as 1,4-butanediol and reacted with polyisocyanates using a ‘one-shot’ or
‘quasi-prepolymer’ approach. For this reason, high-reactivity versions of ultra-low monol
PPG were developed. These polyols give the performance enhancements expected from
ultra-low monol PPG with the added benefit of being one-shot processable [20, 21]. The
higher reactivity is achieved by capping the propylene-oxide based polyols with ethylene
oxide (EO) resulting in products with a high percentage of primary-hydroxyl groups. A
significant amount of work has been reported on polyurethanes based on EO-capped
low monol PPG [20, 21, 22, 23, 24, 25].
A key factor in determining whether or not a polyol can be used in a one-shot process is
reactivity. As a general rule, the polyol reactivity needs to be nearly equal to or greater
than the reactivity of the curative. For example, most one-shot elastomers based on MDI
use BDO as a curative which contains primary hydroxyl groups. A typical PPG polyol
contains about 95% secondary hydroxyl groups. If one attempts to process a blend of
MDI, BDO and PPG, the MDI will react preferentially with the BDO to form an MDI-
BDO hard segment. The MDI-BDO hard segment is incompatible in the system and
phase separates before the PPG polyol has a chance to react. The end result is a low MW,
two-phase system in which the MDI-BDO hard segment is dispersed in partially reacted
polyol, which results in very poor physical properties. Therefore, successful one-shot
processing requires that the polyol and curative have similar reactivities.
The reactivity of PPG polyols can be increased by reacting with EO, which ring opens to
form a primary hydroxyl group. A 100% primary hydroxyl content cannot be achieved
using this approach since EO has a tendency to form chains instead of distributing evenly
to each secondary hydroxyl group; however, it is well known that a 100% primary
content is not required to achieve one-shot processability. The effect of the primary
hydroxyl content on the processability of a one-shot elastomer system is discussed next.
To determine the effect of the polyol primary content on the processability of a one-shot
system, two 4000-MW, ultra-low monol PPG diols, which were EO-capped to obtain
primary contents of 80 and 87 percent were prepared. These were compared with a
436
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
2000-MW PTMEG-2000, which has 100 percent primary hydroxyl content. These polyols
were evaluated in a one-shot elastomer system consisting of carbodiimide-modified MDI
(29.3% NCO), polyol and BDO formulated to an 80 Shore A hardness. Processability
was assessed by measuring the hardness build up with time. The reaction rate was held
constant by adjusting the dibutyltin dilaurate catalyst concentration to achieve a constant
pot life of 90 seconds. Pot life is defined as the work life of the solution, which is the time
period until the solution becomes too viscous to be poured into a mould. The elastomer
solution was poured into moulds preheated to 100 °C and then placed into a 100 °C
vented oven until the elastomer had developed enough integrity to be demoulded. After
demoulding, the elastomer hardness was measured at regular intervals and placed back
in the oven. The final hardness measurement was made after 16 hours cure at 100 °C.
All the data were normalised to 100% using the final hardness measurement. The
normalised hardness versus time were then plotted.
The hardness build of 80 Shore A, one-shot elastomers based on polyols with primary
contents of 80, 87 and 100 percent are shown in Figure 9.6. As expected, the elastomer
437
Advances in Urethane Science and Technology
based on PTMEG-2000 has the fastest hardness build, achieving 90% of its final
hardness after only 30 minutes. The 87% primary content, 4000-MW, EO-capped diol
gave the next fastest hardness build as it took 60 minutes to achieve 90% of its final
hardness. Surprisingly, there was a large difference between the 80 and 87% primary
content 4000-MW, EO-capped diols as the time to achieve 90% of its final hardness
went from 60 minutes to over 4 hours. Based on this information, it was decided that
the primary-hydroxyl content needs to be at least 85%.
The previous studies led to the development of three products, a 2250-MW diol (Acclaim
Polyol 2220), 4000-MW diol (Acclaim Polyol 4220) and a 6000-MW triol (Acclaim
Polyol 6320). Table 9.4 shows the typical properties of these new higher-reactivity,
ultra-low monol PPG polyols. The behaviour of these products in one-shot systems is
discussed below.
The effect of monol content on the processability and properties of one-shot elastomers
is shown in Table 9.9. We prepared 80 Shore A elastomers based on carbodiimide-
modified MDI and BDO using a 4000-MW, EO-capped, ultra-low monol diol
containing 3% monol content and a conventional EO-capped diol containing 15%
monol. These polyols had comparable primary hydroxyl concentrations of 87 and
89%, respectively.
As expected, the monol content of the EO-capped PPG diols has a significant effect
on the elastomer processability. The Shore hardness of the conventional elastomer
was only 57A after 16 hours cure at 100 °C, whereas the elastomer based on ultra-
low monol PPG had a hardness of 78A. It took three weeks for the conventional
elastomer to achieve its final hardness of 82A.
Virtually all elastomer properties improved with lower monol contents (see Table 9.9). Of
particular note, the rebound increased from 59 to 68 percent, the elongation from 470 to
550%, tensile strength from 10 to 19 MPa, tear strength from 49 to 68 kN/m, compression
set decreased from 64 to 23%, and the Taber abrasion loss (ASTM D4060-95 [26]) decreased
from 220 to 80 mg loss/1000 cycles. The improvement in the stress/strain curves is shown
in Figure 9.7. Lower monol content results in significantly higher ultimate polymer MW,
which results in improved mechanical properties.
438
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
439
Advances in Urethane Science and Technology
The elastomers were prepared using the one-shot process by the hand-casting technique
at a pot life of 1.5 minutes.
The compositions, processability and properties of elastomers prepared using the one-
shot process are given in Tables 9.10 to 9.12. For the 4000-MW, EO-capped diol, the
upper hardness processability limit via hand casting is about 85 Shore A as indicated by
the green strength in Table 9.10. Machine casting using a shorter pot life (30 to 45
440
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
seconds) could extend the hardness limit to the low to mid-90s. The lower hardness limit
as indicated by the increase in compression set is about 60 Shore A. The lower hardness
limit can be extended down to at least 40 Shore A by replacing some of the 4000-MW
441
Advances in Urethane Science and Technology
Pendulum Rebound, % 70 72 73 75 75
Ultimate Elongation, % 880 600 450 320 270
Tensile Strength, MPa 2.5 2. 3 1.9 1.7 1.5
100% Modulus, MPa 0.8 0. 9 0.9 0.9 1.0
200% Modulus, MPa 0.9 1.2 1.2 1.2 1.3
300% Modulus, MPa 1.1 1.4 1.5 1.6 —
400% Modulus, MPa 1.1 1.6 1.7 — —
Die C Tear, kN/m 15 14 12 11 9
Compression Set, % (22 h @ 70 °C) 45 38 20 15 12
Compression Deflection, MPa
5% 0.1 0.1 0.1 0.1 0.1
10% 0.3 0. 3 0.3 0.3 0.3
15% 0.4 0.4 0. 5 0.5 0.5
25% 0.8 0.8 0.9 0. 9 0.9
diol with the 6000-MW, EO-capped, ultra-low monol triol as shown by the reduction in
compression set in Table 9.11. Thus, using combinations of 4000-MW diol and 6000-
MW triol, a broad hardness range from about 40 to 85 Shore A can be achieved.
442
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
The 2250-MW (50 OH#) EO-capped, ultra-low monol diol was designed to enhance
processability at higher Shore hardnesses. The processability and properties of one-shot
elastomers based on this polyol are shown in Table 9.12. One-shot elastomers with
excellent processability and properties can be formulated from an 80 Shore A up to
about a 55 Shore D. Thus, a broad range of Shore hardnesses (40A to 55D) can be
achieved using EO-capped, ultra-low monol polyols.
443
Advances in Urethane Science and Technology
When comparing ultra-low monol PPG polyols with PTMEG it is essential to keep in
mind the fundamental differences between these two high-performance polyether polyols
in order to obtain the maximum benefit in urethane applications. Table 9.13 summarises
these differences in terms of their chemical structure, hydroxyl type (reactivity), MWD
and crystallisability.
Both PPG and PTMEG have polyether backbones, which inherently impart excellent
hydrolytic stability and microbial resistance to polyurethanes. The Tg of PPG
approaches that of PTMEG resulting in polyurethanes that can be used in cold, harsh
environments. PTMEG has primary hydroxyl groups, whereas, ultra-low monol PPG
are available with secondary hydroxyl groups or EO-capped versions for primary
hydroxyl contents in excess of 85%. Ultra-low monol PPG polyols have narrow MWD,
which gives polyols that are an order of magnitude lower in viscosity than PTMEG.
The storage and handling requirements for PPG polyols are minimal since they are
non-crystallising liquids with pour points of about –30 °C. Additionally, polyurethanes
based on non-crystallising polyols will not ‘cold harden’ like some of those based on
crystallisable polyols. A comparison of ultra-low monol PPG versus PTMEG in MDI/
BDO cured elastomers is given below.
444
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
The initial surprise was to find that it takes an 8% NCO MDI/Acclaim 3205 prepolymer to
make a 80 Shore A elastomer, whereas, PTMEG-2000 requires a 6% NCO prepolymer. This
finding will be explained in the section on the effect of polyol MWD (see section 9.5.4).
Note the dramatic difference in prepolymer viscosities in Table 9.14. The prepolymer
based on ultra-low monol polyol has a viscosity almost ten times lower than that of
PTMEG-2000. This is due to a combination of the narrower MWD and the higher %
NCO of the Acclaim 3205 prepolymer.
Lower viscosity prepolymers offer significant advantages in processability. This gives the
polyurethane manufacturer much more latitude in terms of processing conditions such
as temperature and catalyst combinations. Elastomers based on Acclaim Polyol 3205
have demould times and green strengths comparable to those based on PTMEG-2000.
When hand casting elastomers using a 1 to 3 minute pot life, this can give demould times
of significantly less than 30 minutes.
As expected, the fundamental differences in the polyols also have a significant impact on
the elastomer physical properties. Although elastomers based on ultra-low monol PPG
have pendulum rebounds significantly higher than conventional or low monol polyols,
they are about 10 units lower that PTMEG at an equal Shore hardness. It is believed this
is due to fundamental differences in the polyol backbone structure. PPG polyols have a
pendant methyl group which may inherently absorb more energy than the linear methylene
chain of PTMEG.
Elastomers based on Acclaim Polyol 3205 have ultimate elongations that are twice that
of PTMEG. It is believed that this is due to the fact that PTMEG stress crystallises as it is
stretched. Crystallites behave like crosslinks and lower the elongation. To test this theory,
a series of MDI/BDO cured elastomers were prepared using 2000-MW crystallisable
polyesters (polybutylene adipate and polycaprolactone) and non-crystallisable polyesters
(polyethylenebutylene adipate and poly-2-methyl-1,3-propylene adipate). Figures 9.8 and
9.9 compare the stress/strain curves of the crystallisable versus non-crystallisable
polyethers and polyesters, respectively. From these figures the importance of the polyol
stress crystallisability in determining the elastomer stress/strain profile can be seen.
445
Advances in Urethane Science and Technology
446
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
The tensile strengths of ultra-low monol based elastomers are lower than that of PTMEG.
However, if one considers the total energy to break as measured by the area under the
stress/strain curve, the elastomers based on Acclaim Polyol 3205 are significantly tougher
than those based on PTMEG. The tear strengths of the elastomers are comparable.
Previously, it was shown that elastomers based on non-crystallising polyols such as ultra-
low monol PPG have twice the ultimate elongation of crystallising polyols such as PTMEG.
The next experiments were aimed at determining how much ultra-low monol PPG it would
take to eliminate the stress crystallisability of PTMEG [19]. A series of 80 Shore A elastomers
447
Advances in Urethane Science and Technology
were prepared using MDI prepolymers based on 4000-MW, ultra-low monol PPG/
PTMEG-2000 blends chain extended with BDO. The results of the addition of 0, 10, 40
and 100 weight percent 4000-MW PPG based on the total polyol weight are given in
Table 9.15.
It was found that only about 10 weight percent of 4000-MW PPG is required to break
up the stress crystallisability of PTMEG-2000 as shown in Table 9.16 with the elongation
increasing from 510 to over 1000%. Surprisingly, it was found that 4000-MW PPG
addition gave a synergistic response in terms of elongation, tensile and tear strength.
Even more surprising was that the synergy doesn’t reach a maximum until up to 40
weight percent of 4000-MW PPG. This elastomer has a 1270% elongation, 35 MPa
tensile strength and a 77 kN/m tear strength. Additionally, the rebound was only three
units lower than the PTMEG elastomer system and the prepolymer viscosity was
significantly lower which enhances processability.
448
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
So one can maximise the benefit from ultra-low monol PPG polyols in urethane
systems such as MDI/BDO cured elastomers by understanding their fundamental
differences in comparison to other high-performance polyols like PTMEG. A detailed
look at the effect of the polyol MWD on the mechanical and dynamic properties of
polyurethanes follows.
449
Advances in Urethane Science and Technology
Figure 9.10 GPC comparison of 4000-MW ultra-low monol PPG to 2000-MW PTMEG
reduction in polyol and prepolymer viscosities [7]. Additionally, it was found that
polyol MWD has a dramatic effect on the polymer mechanical and dynamic properties,
particularly in polyurethane/urea systems [19, 27].
450
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
elastomer properties were much lower than expected as shown by the hardness (83A
versus 96A); rebound (15 versus 45%); tensile strength (9.2 versus 45 MPa) and tear
strength (39 versus 110 kN/m). After several attempts to improve the properties via
processing conditions, catalyst, etc., it was decided to artificially broaden the MWD of
an ultra-low monol PPG polyol to approximate that of PTMEG-1000. This was achieved
by blending a 4000-MW PPG with a low-MW glycol diethylene glycol (DEG), to an
average MW of 1000. The property profile of this elastomer is shown in the last column
of Table 9.16. Interestingly, the property profile is now closer to the elastomer based on
PTMEG-1000 and is very similar to a commercial PTMEG-based 93A product, Airthane
PET-93A, which is a product of Air Products and Chemicals, Inc.
The above conclusions are supported by the DMTA curves shown in Figure 9.11. The storage
modulus in the rubbery plateau region is significantly higher for the elastomer based on the
PPG/DEG blend, supporting the theory that DEG is contributing to the hard segment. Also,
note the transition at 130-150 °C, which is likely to be the softening point of the DEG/TDI/
MBOCA hard segment. The dramatic lowering of the soft-segment Tg from 0 °C to -42 °C
also supports the view that the PPG polyol and DEG are going into separate phases.
451
Advances in Urethane Science and Technology
Figure 9.11 DMTA curves of TDI prepolymers cured with MBOCA-PPG-1000 versus
4000-MW PPG/DEG blended to 1000 MW
In Figure 9.12, the effect of varying the glycol chain length on the DMTA profiles is
shown by using propylene glycol (PG), dipropylene glycol (DPG), tripropylene glycol
(TPG) and a 400-MW PPG. The short-chain glycols (PG, DPG, TPG) all gave similar
storage-modulus profiles, while the 400-MW PPG gave a dramatically different profile.
This indicates that if the MW of the glycol is too high it becomes miscible in the soft-
segment phase and increases the Tg, thus reducing the elastomer rebound and dynamic
performance.
In some cases it may not be practical or cost effective to perform DMTA analysis on all
samples. It was found that materials can be screened by measuring hardness and rebound
at a series of temperatures (-10 °C, 23 °C and 105 °C) and these results are shown in
Table 9.17. Small changes in properties imply a broad temperature use range for the
polymer. These results are consistent with the DMTA findings.
In Table 9.18 the effect of a higher average MW polyol is shown by comparing 2000-
MW ultra-low monol PPG with a blended product (4000-MW PPG/DEG). Again, the
differences in hardness, rebound and modulus are dramatic and are consistent with the
above trends.
452
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
Figure 9.12 Effect of glycol chain length on DMTA curves of TDI prepolymers
cured with MBOCA
453
Advances in Urethane Science and Technology
To determine if the polyol MWD effect can be seen in other polyurethane/urea systems,
a moisture-cured film was prepared using a 2:1 prepolymer based on 2,4-TDI and PPG-
1000 and this was compared with one based on 4000-MW PPG blended with DEG to a
MW of 1000. Table 9.19 shows the mechanical properties of the films in bold type. The
dramatic increases in 100% modulus (1.3 to 3.6 MPa), tensile strength (14 to 35 MPa)
and tear strength (18 to 58 kN/m) obtained by broadening the MWD of the polyol can
454
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
clearly be seen. The DMTA curves of these materials show the same trends as with the
MBOCA-cured TDI prepolymers as seen in Figure 9.13. A formulation with 4000-MW
PPG decreases the soft-segment Tg from –16 °C to –58 °C greatly enhancing the low
temperature performance of this system.
The versatility of using this approach can be clearly seen by blending 4000-MW PPG
and DEG to a wide range of MW (770 to 3000) as shown in Table 9.19. By simply
blending these two components, one can vary the 100% modulus from 0.4 to 8.2 MPa.
Traditionally, hardness would be adjusted by using various polyol MWs; however, when
using this approach there is a reduction in the low-temperature flexibility as shown by
the DMTA curves in Figure 9.13.
Figure 9.13 DMTA curves of TDI moisture-cured prepolymer films - PPG-1000 versus
4000-MW PPG/DEG blended to 1000 MW
Using the blend approach, the hardness can be varied over a wide range while retaining
the low-temperature flexibility as shown by the DMTA curves in Figure 9.14. Note
that with increasing amounts of DEG, there is a transition beginning at about 100 °C
that becomes more pronounced and shifts to about 80 °C. This transition is likely
455
Advances in Urethane Science and Technology
associated with the Tg of the TDI/DEG/urea hard segment. Similar property profiles
are observed in isophorone diisocyanate-based (IPDI), moisture-cured systems utilised
in deck-coating materials.
456
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
The stress/strain curves in Figure 9.15 show that the film based on the 2000-MW PPG
has a significantly lower hardness than that based on PTMEG-2000. To take into account
the differences in polyol MWD, we then prepared an aqueous dispersion using 4000-
MW PPG and DPG blended to a 2000 MW. Figure 9.15 shows that these stress/strain
curves are now very similar up to 200% elongation. At higher elongations, the film
based on PTMEG-2000 begins to stress crystallise resulting in increased modulus and
reduced ultimate elongation. The DMTA curves in Figure 9.16 show that the film based
on the blend of 4000-MW PPG and DPG has a profile more similar to PTMEG-2000
than to that based on 2000-MW PPG.
457
Advances in Urethane Science and Technology
458
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
Table 9.21 demonstrates the polyol MWD effect by comparing BDO cured 6% NCO
4,4´-MDI prepolymers based on 2000-MW PPG and PTMEG-2000. The elastomer based
on the narrow-MWD polyol has a significantly lower hardness (68A versus 83A) and
rebound (60 versus 75%). This is consistent with PTMEG-2000 containing some short-
chain diol, which phase separates into the hard segment promoting higher hardness.
Also, the soft-segment Tg of the PTMEG-2000 elastomer would be lower since the soft
segment actually consists of PTMEG with MW significantly higher than 2000. Based on
our previous discussion, to best approximate the property profile of a broad-MWD polyol
such as PTMEG, one needs to use a higher MW ultra-low monol PPG polyol with some
additional chain extender (BDO).
Table 9.21 illustrates this effect by increasing the PPG MW from 2000 to 4000 and the
prepolymer % NCO from 6 to 8. This increases the hardness to the same level as the
PTMEG-based elastomer and increases the rebound to 65%. As an aside, increasing the
prepolymer % NCO from 6 to 8 represents a polyol/BDO blend of 2160 MW.
459
Advances in Urethane Science and Technology
460
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
The benefits from using ultra-low monol PPG polyols in polyurethane or polyurethane/
urea systems can be realised by understanding the fundamental differences between them
and other high-performance polyols such as PTMEG. In particular, when considering
the polyol MWD, PTMEG should be regarded as a high MW polyol blended with a low
MW chain extender (or pseudo chain extender). In contrast, ultra-low monol PPG contain
no low-MW fraction, therefore, to mimic the behavior of PTMEG it is necessary to add
additional low MW glycol into the prepolymer or resin components. Because of this
feature, ultra-low monol PPG have a broader formulating latitude, since chain extender
can be added to tailor the property profile to the specific application.
9.6 Conclusions
Ultra-low monol PPG polyols have excellent properties and ease of processing. The
elimination of monol enables polymers to be produced having much higher final MW,
i.e., improved properties, as well as significantly faster MW buildup during polymer
formation, i.e., shorter demould times and excellent green strength.
Another important feature of these polyols is their narrow MWD. Low polydispersity
results in lower viscosities in both the polyol and isocyanate-terminated prepolymers.
Polyol MWD has a significant effect on the mechanical and dynamic properties of
polyurethane and polyurethane/urea systems. Broad-MWD polyols such as PTMEG or
polyester polyols contain significant amounts of hard-segment diol, which limits their
softness. Acclaim polyols contain no hard-segment diol and thus have inherent superior
properties in soft systems. They can be made to approximate broad-MWD polyols such
as PTMEG by incorporation of a low MW glycol.
Acknowledgements
We would like to thank Steve Bailey, Daphne Hale, Beth Lambert, Tony Loveday and
Mike Robinson, who carefully and conscientiously prepared the samples and carried out
the physical property testing.
References
1. D. M. Simons and J. J. Verbanc, Journal of Polymer Science, 1960, 44, 144, 303.
461
Advances in Urethane Science and Technology
5. H. van der Hulst, G. A. Pogany and J. Kuyper, inventors; Shell Oil Company,
assignee; US Patent 4,477,589, 1984.
16. S. D. Seneker and N. Barksby, Presented at the Utech 96 Conference, The Hague,
The Netherlands, March 1996, Paper No.46.
17. N. Barksby, S. D. Seneker and G. L. Allen, Urethanes Technology, 1996, 13, 1, 36.
462
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
26. ASTM D4060-95, Standard Test Method for Abrasion Resistance of Organic
Coatings by the Taber Abraser.
28. S. D. Seneker, N. Barksby and G. B. Ellerbe, Presented at the Utech Asia ‘97
Conference, Suntec City, Singapore, 1997, Paper No.41.
463
Advances in Urethane Science and Technology
464
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
APPENDIX
Isocyanate is added first to the reaction vessel and heated to approximately 50 to 60 °C.
Polyol is then added to the isocyanate, carefully controlling the addition rate to ensure
the reaction temperature does not exceed 85 °C. Higher temperatures can cause harmful
side reactions that can affect prepolymer performance. After reacting at 80 °C for
approximately 4 hours, the extent of reaction was determined by measuring the percent
isocyanate content (% NCO) of the prepolymer. Heating was continued until the
prepolymer was slightly below the theoretical % NCO value.
MDI/BDO cast-elastomer systems were made by mixing and reacting two components,
the MDI-terminated prepolymers (described previously) and BDO. The elastomers in
this chapter were prepared using the hand-mixing procedure. The MDI prepolymer was
preheated to ~65 °C and degassed under vacuum (130 to 665 Pa) until foaming stopped.
Catalyst (25-50 ppm dibutyltin dilaurate) was added to the prepolymer and mixed
thoroughly to give the desired pot life of 1 to 3 minutes.
The BDO was added (isocyanate to hydroxyl ratio (NCO:OH) of 1.03) and mixed thoroughly
using a Jiffy mixer until the blend was homogeneous. This solution was immediately poured
into preheated moulds at 100 °C treated with mould release. After demoulding, the samples
were post-cured at 100 °C for 16 hours. The polymer samples were conditioned at 23 °C and
50% relative humidity for at least four weeks prior to testing.
465
Advances in Urethane Science and Technology
CD-MDI, polyol and BDO were charged to a 250 ml glass jar. Dibutyltin dilaurate catalyst
(25-50 ppm) was added to adjust the pot life to approximately 90 seconds. These
components were mixed thoroughly and poured into preheated moulds at 100 °C, treated
with mould release. After demoulding, the samples were post-cured at 100 °C for 16
hours. The polymer samples were conditioned at 23 °C and 50% relative humidity for at
least four weeks prior to testing.
The TDI/MBOCA cured elastomers were prepared by mixing and reacting two
components: 2,4-TDI prepolymers (described previously) and MBOCA. The elastomers
in this chapter were prepared using the hand-mixing procedure. The 2,4-TDI prepolymer
was preheated to 90 °C and degassed under vacuum (130 to 665 Pa) until foaming
stopped. The MBOCA was melted at 120 °C. The 2,4-TDI prepolymer and MBOCA
were stirred thoroughly using a Jiffy mixer at an NCO:NH ratio of 1.05 without catalyst.
The solution was immediately poured into preheated moulds at 100 °C treated with
mould release. After demoulding, the samples were post-cured at 100 °C for 16 hours.
The polymer samples were conditioned at 23 °C and 50% relative humidity for at least
four weeks prior to testing.
The 2,4-TDI prepolymers were cast as films on glass plates using a 0.762 mm draw-
down bar (a draw-down bar is a metal bar with a gap, which is pulled down a glass plate
leaving a thin film of prepolymer. The prepolymer is then moisture-cured to form the
polymer film). The films were cured/conditioned at 23 °C and 50% relative humidity for
at least four weeks prior to testing.
Charge polyol, DMPA, (stabiliser), NMP, (coalescing aid) in a reaction vessel equipped
with a thermocouple (thermometer) and agitator. Care should be taken to keep
466
Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes
moisture away from the reaction mixture by using a drying tube or a positive pressure
of dry nitrogen.
Heat the mixture to 95 °C and apply vacuum to remove any residual water (DMPA and
polyol stored in the laboratory may contain high water levels).
Cool the mixture to 50-60 °C, then weigh in isocyanate (IPDI, etc.)
React this mixture until the % NCO content is 0.2-0.5 units lower than the theoretical
% NCO. Reaction time is usually 6 to 12 hours depending on the % NCO of the
prepolymer.
Cool the mixture to 50-60 °C, and then add neutraliser, i.e., triethylamine, to the
prepolymer. Mixing is critical during this step. The neutralisation forms the water-
dispersible carboxylate ammonium salt.
Heat the prepolymer amine salt to 60 to 80 °C and slowly pour into a resin kettle charged
with a known amount deionised water at 25 to 50 °C with strong agitation. Add the
prepolymer to give the desired solids content.
Immediately add the chain extender (ethylene diamine), which has been diluted in water
slowly to the aqueous polyurethane/urea dispersion. Typically chain extend only 80-
90% of the actual free NCO groups because some will be lost due to reaction with water.
As an alternative, the amine extender may be added to the water prior to dispersing the
neutralised prepolymer.
Cook out the residual excess isocyanate by heating the aqueous dispersion at 60-80 °C
for two hours. Excess unreacted isocyanate in an aqueous polyurethae/urea dispersion
will lead to carbon dioxide gas and pressure buildup in the container.
467
Advances in Urethane Science and Technology
468
Abbreviations and Acronyms
469
Advances in Urethane Science and Technology
470
Abbreviations and Acronyms
471
Advances in Urethane Science and Technology
QA Quality assurance
QC Quality control
R&D Research and Development
RGA Residual gas analysis
RH Relative humidity
RIM Reaction Injection Moulding
RRIM Reinforced reaction injection moulding
RT Room temperature
RUF Recycled urethane fluff
S Styrene
SAE Society of Automotive Engineers
SD Standard deviation
SEM Scanning electron microscope
SLM Standard litres per minute
SRG Spinning rotor guage
SRIM Structural RIM
SSF Silica surfactant
STP Standard temperature and pressure
TCP Trimethylsilyl-capped polysilicate
TDI Toluene diisocyanate
TEDA Triethylene diamine
TEM Transmission electron microscopy
Tg Glass transition temperature
TG Technical grade
THF Tetrahydrofuran
TOC Top-of-the-cup
TP Thermoplastic
TPG Tri propylene glycol
TPR Time pressure release
TPU Thermoplastic PU
VIP Vacuum insulated plate
VOC Volatile organic compounds
VPF Variable pressure foaming
VS Vinyl trimethyl silane
WTR Water permeation rate
472
Contributors
473
Advances in Urethane Science and Technology
474
Contributors
475
Advances in Urethane Science and Technology
476
Author Index
477
Advances in Urethane Science and Technology
478
Author Index
Gutmann, V. 337, 349, 357, 359, 366 Kheshgi, H.S. 242, 259
Hamann, H. 256 Kim, D. 417
Hamilton, A.J. 199, 212 Kingston, B.M. 110, 256
Harper, S.D. 462 Kistler, S.F. 259
Hartung, M. 332 Klempner, D. 417
Hashimoto, H. 333 Klincke, M. 82, 111
Haworth, G.J. 212 Kloiber, K.A. 266, 334
Hegedus, C.R. 266, 334 Koberstein, J.T. 418
Heimenz, P.C. 259 Koczo, K. 258
Hepburn, C. 418 Koczone, J.K. 258
Hermsen, A. 334 Kodama, K. 160, 209
Herold, R.J. 462 Kollmeier, H.J. 109, 257
Herrington, R.M. 258 Komarova, A.B. 256
Hesketh, T.R. 418 Kopusov, L.I. 256
Hespe, H. 419 Kossmann, H. 333
Higginson, W.C.E. 461 Kovac, J. 211
Hildebrand, J.H. 111 Kozakiewicz, J. 332, 333, 334
Hilker, B.L. 111 Kryszewski, M. 332
Hill, R.M. 255 Kücükpinar, E. 202, 211
Hilyard, N.C. 209 Kulkarni, R.D. 256
Höchtlen, A. 109 Kuropka, R. 333
Hock, K. 258 Kushner, S.A. 82
Hoogendoorn, P. 208 Kuyper, J. 462
Hoppe, P. 109 Lai, Y-H. 366
Hosoi, K. 333 Lamb, W. 175, 210
Hu, C.B. 419 Langenhagen, R. 111
Huber, L.M. 111 Lassila, K.R. 83
Hudales, J.B.M. 258 Lavielle, L. 367
Hutzinger, O. 109 Lawler, L.F. 111
Huybrechts, J.T. 332 Lawrey, B.D. 463
Immergut, E.H. 210 Lean, B. 333
Ingram, B.T. 257 Lee, K.D. 111
Ishida, H. 367 Legocka, I. 334
Ivanov, I.B. 257, 258 Lerner, L.E. 418
Jensen, O.E. 260 Leung, L.M. 418
Juran, R. 210 Lewandowski, L.H. 366
Kanner, B. 91, 92, 231, 256, 258 Leyden, D.E. 367
Kanner, T.G. 110 Lidy, W. 109, 257
Keane, N.W. 209 Lilaonitkul, A. 418
Kendrick, T.C. 110, 256 Lin, J.R. 418
Khemani, K.C. 111 Lin, J.Y. 209
479
Advances in Urethane Science and Technology
480
Author Index
481
Advances in Urethane Science and Technology
482
Main Index
Main Index
483
Advances in Urethane Science and Technology
484
Main Index
485
Advances in Urethane Science and Technology
486
Main Index
487
Advances in Urethane Science and Technology
488
Main Index
489
Advances in Urethane Science and Technology
490
Main Index
491
Advances in Urethane Science and Technology
492
Main Index
U Flange 168
Viscosity 217-218, 220, 227-228, 245,
Urethane elastomers 413 276
Heat aged 413 Bulk 220
Urethanes 335 Bulk dynamic 227
Adhesion behaviour 335 Finite surface 245
Surface 217-218, 220, 225, 227-232,
V 235-237, 242, 255
Surface dilatational 240
Vacuum panel 158, 161, 188, 191 Surface shear 240
Characterisation 188 Volatile organic compounds 4
Insulating performance 191 Voranol 7, 80
Manufacturing process 188
Vacuum panel technology 158
Van der Waals bonding 218 W
Vending machines 177, 204 Water resistance 277
Vibrathane 375-376, 378, 383 Water transmission 176
Video imaging 106 Window fogging 42
Video microscopy 103, 105
Vinyl staining 42, 62
VIP 190, 193, 197, 202, 203
Z
Longevity 203 Zeolite 181
Performance 202 Zeta potential 314
Reliability 202
493
Advances in Urethane Science and Technology
494
ISBN: 1-85957-275-8