Materials and Structures/Mat&iaux et Constructions,Vol.

32, April1999, pp 163-173

RILEM TC 116-PCD: Permeability of Concrete as a Criterion of its Durability

3 Final Report

Concrete durability- An approach towards performance

testing

Former and present full and corresponding members o f the TC as well as members o f the research consortium:
C. Andrade, Spain; A. Bettencourt-Ribeiro, Portugal; N. R. Buenfeld, UK; M. Carcasses, France; N.J. Carino, USA;
F. Ehrenberg, Germany; C. Ewertson, Sweden; E. Garbocczi, USA; M. Geiker, Denmark; O. E. Gjorv, Norway;
A. F. Goncalves, Portugal; H. Gr~if, Germany; H. Grube, Germany; H. K. Hilsdoff, (chairman 1989-1992), Germany;
R. D. Hooton, Canada; J. Kropp, (secretary 1989-1992, chairman since 1992 and project coordinator), Germany;
S. Modry, Czech Republic; Ch. Molin, Sweden; L. O. Nilsson, Sweden; J. P. O]livier, France; C. L. Page, UK;
L.J. Parrott, UK; P. E. Petersson, Sweden; F. R. Rodriguez, Spain; M. Rodhe, Sweden; M. SaM, Portugal; N. Skalny,
USA; A.M.G. Seneviratne, UK; L. Tang, Sweden; F. Tauscher, Germany; R. Torrent, Argentina; D. Whiting, USA.

FOREWORD In 1989, the RILEM Technical Committee TC 116 -

"Permeability of Concrete as a Criterion of its Durability"
Concrete durability has been a key issue for many took up the task of investigating methods for performance
years due to its high significance in the serviceability of testing of concrete in view of its potential durability during
concrete or reinforced concrete structures as well as to its the service life of a concrete structure. Based on the funda-
tremendous economic impact in the construction sector. mental mechanisms involved in the most important corro-
Accordingly, the subject has been studied in all countries sion reactions not only for the concrete itself but also for
using this material and construction technique, thereby the reinforcement, the transport of species through the
focusing on domestic construction traditions with locally interconnected pore spaces of concrete has been identified
available materials that are subjected to local exposure as a major materials characteristic - and, in many cases, a
conditions. In consequence, national requirements on rate controlling parameter - for corrosion reactions.
durability control emerged from long-term experience Although the volume and continuity of the internal pore
in the respective regions. An increasing international system of concrete are determined by parameters such as
cooperation and exchange of knowledge, technology, concrete composition, degree of hydration and the quality
services and goods, however, demonstrated the need to of the aggregate/matrix interracial transition zone, the per-
substitute the regional empirical approach for a more viousness of concrete can only be evaluated directly by a
fundamental treatment of durability with internationally permeation test. However, various media, e.g. ions, gases
recognized requirements on its control. As an example, and liquids, as well as different transport mechanisms, must
the creation of the Common Market in the European be distinguished.
Union demonstrated the need for supra national stan- TC l l 6 - P C D has compiled an extensive state-of-
dards, and in the Construction Products Directive the-art report on the transport of species through con-
(CPD) of the European Commission, materials specifi- crete, the available methods for an early experimental
cations based on performance criteria are preferred over evaluation of concrete transport properties and existing
descriptive requirements on the composition, which correlations between concrete transport parameters and
often obstruct the introduction of new materials and the corrosion resistance of the respective concretes [1].
technologies onto the market. This review is focused on laboratory testing on sepa-

I359-5997/99 9 ILILEM 163

Materials and Structures/Mat~riaux et Constructions,Vol.32, April1999

rarely cast companion specimens for either pre-testing or in the first work package of the project a preconditioning
quality control purposes in concrete production. method was developed to install a controlled water content
Increasing attention has also been paid to on-site testing with a homogeneous distribution in the test specimens.
of existing concrete or reinforced concrete structures for The effectiveness of preconditioning was evaluated in an
the transport properties of the concrete, in particular the initial intercomparison testing. Subsequently, a selection of
concrete cover. More recent developments of specially- frequently used test methods for laboratory and on-site
designed equipment for in-situ testing and their applica- testing of the gas permeability, the absorption of water and
tion are presented in [2] and [3]. the penetration of chloride ions were applied to a range of
Within the Technical Committee 116-PCD, a sub- concrete grades. This enabled evaluating the suitability and
group of 10 research laboratories continued the work on range of applicability of the methods, proposing improve-
test methods for the characterization of concrete trans- ments and correlating the measured transport parameters
port properties with an experimental research project. with durability characteristics.
The work program and all experimental results and their The Cembureau set-up was then selected as a stan-
evaluation were discussed in the Technical Committee; dard method for gas permeability measurements; for the
thus, the conclusions and recommendations emerging capillary absorption of water, a modified ponding test was
from the research program do represent the position of recommended. For the measurement of chloride ion
the Technical Committee 116-PCD. ingress, a migration test and an immersion test were eval-
Under Contract No. MAT1-CT93-0001, the uated; these tests had been used in the past for pre-testing
Commission of the European Community granted finan- of concrete mixes and in production control during the
cial support to conduct an experimental research work erection of advanced structures. However, both methods
which was aimed at developing standard test procedures for have not yet been recommended as routine tests in qual-
hardened concrete with an emphasis on durability control. ity control. In a final intercomparison testing, the
As contractors and partners of the research consortium, the selected methods showed good agreement among the
following European research and testing laboratories par- participating laboratories. Test instructions, which may
ticipated in different steps of the experimental program serve as drafts for European standards on test methods for
according to the agreed work plan: (1) Institut fiir concrete, are proposed for the preconditioning of test
Massivbau und Baustofftechnologie, University of specimens, for the measurement of gas permeability and
Karlsruhe, Germany; (2) Instituto Eduardo Torroja, for the determination of the capillary absorption of water
Consejo Superior de Investigaciones Scientificas, Madrid, on separately cast companion specimens.
Spain; (3) COWI Consulting Engineers and Planners AS,
subcontracting Ramboll, Hannemann and Hojlund AS,
both in Lyngby, Denmark; (4) Laboratorio Nacional de 2. INTRODUCTION A N D OBJECTIVES
Engenharia Civil, Lisbon, Portugal; (5) Forschungsinstitut
der Deutschen Zementindustrie, Diisseldorf, Germany; In current national and international codes and stan-
(6) Department of Building Materials, Chalmers Technical dards on the production and use of concrete as a structural
University, G6teborg, Sweden; (7) Laboratoire Mat&iaux material, minimum requirements are specified for the qual-
et Durabilitd des Constructions, Institut National des ity of the concrete-making materials and their proportion-
Sciences Appliqu&s, Universitd Paul Sabatier, Toulouse, ing as well as for placing and handling of the flesh concrete.
France; (8) Institut fiir Baustoffe, Massivbau und However, no requirements have been set forth as of yet for
Brandschutz, University of Braunschweig, Germany; the actual performance of concrete under a corrosive attack.
(9) Department of Civil Engineering, Aston University at For the majority of the corrosion mechanisms encountered
Birmingham, Great Britain; and (10) Swedish National in practice, the ingress of aqueous solutions, whether gases
Testing and Research Institute, Boras, Sweden. or ions, from the environment into deeper sections of a
In the summary report hereafter, an overview of the concrete member is required to initiate and support damag-
work completed and the major results are presented. ing reactions. Thus, the perviousness of concrete for delete-
Details on the work plan of the project and the results rious media controls the corrosion rate and transport para-
are given in the interim reports of the individual partners meters which provide quantitative information on the
to the project coordinator (TC chairman, Stiftung penetration of various media into concrete and could serve
Institut ftir Werkstofftechnik, and Hochschule Bremen, as a performance criterion for concrete durability.
Bremen, Germany) and his four Milestone Reports to The objective of the research project lies in developing
the European Commission. standard test methods for the determination of concrete
transport coefficients, i.e. the coefficient of gas permeabil-
ity, the capillary absorption of water and the chloride ion
1. ABSTRACT diffusion. These three transport mechanisms have been
selected because they are either directly involved in corro-
In the experimental research project, standard test sion reactions or they exhibit a close correlation with a
methods are developed for the measurement of transport corrosion rate of concrete [1]. Since the transport capacity
coefficients of concrete for gases, water and chloride ions. of concrete largely depends on its water content [4, 5], the
Since the moisture concentration in concrete is a most test methods must also consider a standardized precondi-
important parameter in affecting the transport parameters, tioning, which must then result in a controlled average

164
TC116-PCD

moisture concentration with a uniform distribution. evaluated. Centrally-manufactured concrete specimens

Corresponding standard preconditioning and testing pro- were cured for 3 and 7 days, respectively, and were then
cedures are mandatory to obtain comparable results from subjected to a preconditioning regime. At the age of 28
these tests as a prerequisite for drawing correlations of the days, an initial testing at random was conducted on a
transport coefficients with concrete durability or derMng a limited number of specimens; subsequently, the whole
classification of concrete with regard to its durability. set of test specimens was shipped to the partners in a
Transport coefficients may be used for different pur- sealed state for testing. At an age of 56 days and 180 days,
poses, e.g. a comparative tool to rank a number of similar respectively, the specimens were tested for their gas per-
concrete mixes, a characteristic to describe the open pore meability and the capillary absorption of water.
system of concrete or its potential durability, an input para-
meter for a numerical modeling of corrosion rates, for ser-
vice life prediction, or for the evaluation of existing struc- 3.3 Work Package 3: Evaluation of test
tures with regard to maintenance policies. This research methods and durability characteristics
project focuses on transport coefficients as criteria for con-
crete durability which are available at an early stage, e.g. In Work Package 3, experience should be gained from
during pre-testing of concrete mixes and routine testing in known test methods that are often employed to measure
production control, as well as on materials coefficients for the gas permeability, the capillary absorption of water and
numerical modeling. Consequently, the test methods the diffusion coefficient of chloride ions in concrete.
being sought should provide an accurate value for a true These transport cases, had been chosen due to their signifi-
materials parameter, be suitable for routine testing at an cance for a wide range of concrete corrosion mechanisms.
early stage and be sensitive enough to distinguish different For each of the selected transport cases, a large number of
concrete grades. different test set-ups, equipment and procedures had been
proposed by different researchers. The scope of WP3 was
to evaluate the suitability of the individual methods and to
3. WORK PROGRAM further develop the procedures, if possible.
For the measurement of gas permeability, the
The experimental program was subdivided into four Cembureau equipment [6], the Sch6nlin laboratory
consecutive work packages with different scopes. method [7], as well as bore hole methods proposed by
Parrott [8], Paulmann [9] and Figg [10], were employed.
The capillary suction of water was measured both accord-
3.1 Work Package 1: Preconditioning ing to a ponding method proposed by Fagerlund [11] and
using the measurement procedure comparable to an earlier
The transport of media through concrete occurs in R I L E M Technical R e c o m m e n d a t i o n , C P C 11.2
the interconnected capillary pore spaces, porous inter- "Absorption of water by concrete by capillarity" [12].
faces and micro-cracks. Depending on the exposure Different specimens' sizes were used in this testing how-
conditions, these spaces may be partially occupied by ever. The ISA test [13] and Figg's water absorption method
water films along the internal surfaces or even saturated [10] were also employed. The ingress of chloride ions was
with water. While a water saturated state provides for the studied by migration tests [14, 15], basically related to the
optimum transport capacity for dissolved ions, the gas AASHTO [16] test, as well as by immersion of the test
flow or the capillary activity vanishes. In contrast, the specimens into chloride solutions with subsequent profile
flow of gases as well as the capillary activity increases analyses [17]. Based on the experience gained from the
with decreasing moisture content while the ionic diffu- comparative study, one single test method should be pro-
sion ceases for low moisture concentrations. Thus, the posed for each of the three transport cases as a standard.
moisture concentration of the specimens exerts a signifi- On ten different concrete test series included in WP 3,
cant influence on the observed flow of media. It is the compressive strength, the abrasion resistance and the
mandatory, therefore, to control the average moisture depth of carbonation under laboratory conditions up to an
concentration of the test specimens, yet at the same age of 1.5 years were measured.
time, a homogeneous distribution of the remaining
water content must be assured. The necessary precondi-
tioning of the test specimens should be completed at a 3.4 Work Package 4: Intercomparison testing II
concrete age of 28 days; thus, the measurement of trans-
port coefficients could be introduced into the usual test- The selected standard methods resulting from WP 3,
ing scheme for concrete. including their proposed amendments, were subjected to
a second intercomparison testing with three different
concrete mixes. All specimens were centrally manufac-
3.2 Work Package 2: Intercomparison testing I tured and preconditioned and then shipped to the part-
ners. Tests were conducted at a concrete age of 63 days
In the first intercomparison testing, the effect of the and 182 days, respectively.
remaining equilibrium moisture concentration in the The intercomparison should reveal the repeatability
test specimens, as a result of the preconditioning, was and the reproducibility of the selected methods.

165
Materials and Structures/Materiaux et Constructions,Vol.32, April1999

3.5 Materials 4. RESULTS

For the experimental investigations, five different con- 4.1 Preconditioning

crete mixes were designed which cover a range of concrete
qualities typically encountered in the structural application 4.1.1 Drying methods
of concrete. Not considered were extreme concrecexom- The aim of the preconditioning consists of a uniform
positions, e.g. low grade concretes with a high porosity, moisture distribution with a well-defined and pre-set aver-
such as concrete with a honeycombed structure, or high age moisture concentration of the test specimen which is
strength concretes with very low water cement ratios. The below the original moisture content after the curing
strength class of the cement was 32.5 N/mm 2. regime. The necessary drying of the concrete must be
In Table 1, characteristics of the concrete mixes are accelerated in order to allow the testing within the 28-day
summarized: period. The drying of concrete is accelerated by an elevated
temperature that increases the diffusion coefficient for
water vapor as well as by a low moisture concentration of
Table 1 - Composition of concrete mixes
the ambient air resulting in a steeper moisture gradient.
Mix Type of cement Cement content Water cement Limitations exist however for the application of these dry-
[kg/m 3] ratio
ing conditions: elevated temperatures under moist condi-
A CEM I 380 0.4 tions will enhance the hydration, but under dry conditions
B CEM I 270 0.7 they may cause internal micro cracking of the concrete and
C CEM III A 380 0.4
the degree of the resulting pre-damage may depend on the
concrete quality, e.g. increased micro cracking can be
C' CEM III A 270 0.7 expected in dense and high-quality concretes. On the other
D 9 CEM II B - V 270 0.7 hand, low ambient moisture conditions may cause large
E CEM II A - D * 380 0.4 variations of the water content across the drying specimen.
Two drying methods have been investigated, both
* (where commercially available, this cement correspondedto CEM 1I 52.5 R).
applying an elevated temperature to accelerate the drying
process, but the applied temperature was limited to only
After casting of the fresh concrete in molds and com- 50~ The parameters of the drying regimes were fixed
paction, the top side was covered with a plastic sheet on the basis of desorption isotherms which were mea-
until &moulding. The curing under wet burlap contin- sured at 20~ and 50~ in the range of 55 to 95 percent
ued up to an age of 3 and 7 days, respectively. Test speci- relative humidity:
mens for migration or immersion tests were not specially Method I: The basic aim of drying method I was to
cured but stored under water until testing. precondition the test specimens to a moisture concentra-
In Work Packages 1 and 3, the participating partners tion which is in equilibrium with a 65 percent relative
produced their own test specimens on the basis of Table humidity at 20~ In order to approach the equilibrium
1, using the locally available cement of the individual condition within a 28-day period, the drying process
types and the locally available aggregates. Thus, rounded must be accelerated but, at the same time, no steep
material as well as crushed rock was employed. Mso, the moisture gradients along the drying path in the speci-
grading of the aggregates followed national tradition, i.e. mens should develop. Therefore, it was attempted to
the maximum aggregate size varied between 16 and maintain a high moisture concentration in the outer-
32 mm for the different member states. In order to most zone of the concrete specimens which did not fall
obtain a sufficient workability of the fresh concrete, below the final equilibrium moisture concentration after
plasticizers were used in some laboratories. the preconditioning. From the measured desorption
isotherms, it followed that 85% relative humidity at
50~ corresponds to the same equilibrium moisture
3.6 Size and shape of the test specimens concentration in the concrete specimens as 65% relative
humidity at 20~ see Figs. 1 and 2.
For the gas permeability and the capillary absorption of Test specimens were kept in ventilated containers at
water, the attempt was made to apply the same geometry 85% relative humidity and 50~ over saturated salt solu-
of the test specimens, i.e. circular discs with a diameter of tions or in ventilated climate chambers, and the drying
150 mm and a thickness of 50 mm. Bore hole techniques process was monitored by periodic weighing of the spec-
and some capillary absorption experiments required con- imens. In Fig. 3, the drying behavior of 4 different test
crete cubes with a side length of either 100 mm or 150 series is presented by the remaining relative water con-
mm. Specimens for migration tests were cut from cylin- tent of the specimens as a function of time.
ders with a diameter d = 100 mm, length 1 = 200 mm w ( t ) - Winf
which were then cut into discs of the appropriate height. W--
W o -- Winf

w(t) = water content at time t

w o = original free water content
Win f = equilibrium water content.

166
TC 116-PCD

rium within a 28-day period, but for low

ratios of w/c = 0.4 the m e t h o d failed
because the specimens were still far from
their equilibrium condition after 28 days,
see Fig. 3.
Method II: The second drying method
included the acceleration of the drying
process both by increasing the diffusion coef-
ficient of water vapor at an elevated tempera-
ture and by a steeper moisture gradient across
the specimen. However, a second step in the
preconditioning regime must then redistrib-
ute the remaining water profiles according to
a homogeneous distribution.
From the desorption isotherms, the nec-
essary liberation of water of the individual
concrete series was determined as the differ-
ence between the original water content and
the equilibrium water content at 55, 65, 75,
85 and 93 percent relative humidity values,
see Figs. 1 and 2. The test specimens were
then dried in ventilated drying chambers at
50~ and approximately 10 percent relative
humidity until the pre-set weight loss was
Fig. 1 - Desorption isotherms for concrete A at 20~ and 50~ respectively. achieved. Depending on the original water
content and the porosity of the specimens, as
well as on the desired degree of drying, the
necessary duration of drying was between a
few hours and approximately 3 weeks.
Due to the low ambient humidity,
moisture profiles occurred at the end of the
drying phase. In a second step, the speci-
mens were then kept in sealed containers
over saturated salt solutions of appropriate
relative humidity in order to redistribute
the moisture profiles in the specimens. At
an age of 28 days, the remaining moisture
profiles were measured in terms of relative
pore humidity in cavities as well as in terms
of absolute water content by crushing small
samples taken at various depths. These tests
showed that the pre-drying should be lim-
ited to an equilibrium with 75 percent rela-
tive humidity in order to prevent excessive
moisture gradients, see Fig. 4.
For an equilibrium with a lower relative
humidity of the ambient air, i.e. a higher
degree of pre-drying, the remaining differ-
ence in water content between a specimen's
surface zone and central portion was higher
than 1 percent by dry mass of concrete.
For all further experiments, the test
Fig. 2 - Desorption isotherms for concrete at 20~ and 50~ respectively. specimens were pre-dried to an equilib-
rium with 75 percent relative humidity
according to drying method II.
At _w= 0, the specimen has achieved the equilibrium
moisture condition. 4.1.2 Effect of water content on transport properties
For very porous concretes with high water cement The effect of the internal moisture concentration on
ratios, the accelerating effect of the elevated temperature transport properties was studied for concretes A and B,
brought the specimens close to their moisture equilib- cured for 3 and 7 days, respectively. The test specimens

167
Materials and Structures/Mat~riaux et Constructions, Vol. 32, April 1999

were preconditioned according to method II

in the range of 55 to 93 percent relative
humidity and then tested at an age of 56 days.
Fig. 5 shows the results of the gas permeabil-
ity tests with the Cembureau method and
Fig. 6 demonstrates the results of the capil-
lary absorption tests. These results reveal the
drastic drop in the transport coefficients with
increasing moisture content, and for elevated
moisture concentrations in equilibrium with
93 percent relative humidity, the tests
revealed that no reliable measurements of the
two transport coefficients were possible any
longer. In a number of tests, no flow could
be measured at all and for all measured values
a high scatter of the results was observed.

4.2 Intercomparison testing I

Fig. 3 - Drying curves of concrete test specimens at 50~ and 85 percent relative The first intercomparison testing concen-
humidity,
w(t)- Winf
w= -
trated on the effectivity of the precondition-
W o -- Winf ing by repeated measurements of the gas per-

Fig. 4a -
Profiles o f
relative
pore
humidity
across the
thickness
o f the test
specimens.

Fig. 4 b -
Profiles o f
relative
pore
humidity
across the
thickness
of the test
specimens.

168
TC 116-PCD

The investigation was limited to concrete mix A

cured for 3 days.
Figure 7 presents the results of the permeability
measurements at an age of 56 days. Due to the
large number of required test specimens for all
partners, the concrete was mixed and poured in 4
consecutive batches. Although the same materials
were used and the same p r o p o r t i o n i n g was
attempted, a statistical analysis of the results of a
pre-testing in the manufacturing laboratory
revealed that the concrete batches exhibited differ-
ences in their air permeability. The results of the
random pre-testing for each concrete mix are indi-
cated in the diagram by the horizontal bars; the
columns represent the results obtained in the par-
ticipating laboratories on other specimens of the
same batch. Considering the magnitude of the
measured value, a satisfactory level of agreement
was observed. In cases of increased deviations, the
respective laboratory carried out further calibration
tests with a porous body of known permeability.
Figure 8 shows the results on the capillary
absorption of water after 1, 4 and 24 hours of
Fig. 5 - Effect o f equilibrium moisture concentration on the gas perme- contact with water, for concrete ages of 56 days
ability of concrete. and 90 days, respectively. Since no differ-
ence in the absorption values was observed,
it may be assumed that the specimens had
effectively reached their moisture equilib-
rium during the preconditioning.

4.3 Evaluation of test methods and

durability characteristics
The different test methods pre-selected
for the comparative study were allocated to
the participants so that each test method was
employed by at least three laboratories; a
comparison of the experiences and the
obtained results could therefore be made. It
was also within the scope of Work Package 3
to report typical values for the observed trans-
port characteristics measured with different
methods. Therefore, the test methods were
applied to the whole range of concrete mixes
A, B, C, C', D and E, with curing under plas-
Fig. 6 - Effect of equilibrium moisture concentration on the absorption o f water. tic sheets for 3 and 7 days, respectively.
The preconditioning of the test speci-
mens for laboratory tests was conducted
meability and the water absorption after 56 days and 180 according to method II. In order to calculate the neces-
days, respectively. The required number of test specimens sary loss of water, one point of the desorption isotherm
were manufactured, cured, preconditioned and subjected had to be measured first, i.e. the equilibrium moisture
to a random pre-testing in one laboratory; then, the speci- content at 75 percent r.h.
mens were shipped in a sealed state to the participants. By In Work Package 3, methods suitable for on-site testing
virtue of the results of Work Package 1, the specimens were also employed. For these methods, additional test series
were preconditioned to an equilibrium moisture concen- were not preconditioned by accelerated drying, but rather
tration with 75 percent relative humidity at 20~ After were kept in a climate-controlled chamber at 75 percent r.h.
the initial testing at an age of 56 days, the specimens were until testing.
stored in climate-controlled chambers at 75 percent rela- Further application of transport parameters in dura-
tive humidity until the second testing at an age of 180 days. bility studies was demonstrated with the help of correla-

169
Materials and Structures/Mat6riaux et Constructions, Vol. 32, April 1999

tions between the measured transport para-

meters and durability characteristics. On
the ten concrete test series, the rate of car-
bonation in a laboratory environment and
the abrasion resistance measured in a grind-
ing wheel test were observed.
The experience gained from the indivi-
dual methods for gas permeability tests
showed that the Cembureau method is
very reliable, easy to handle and exhibits
very good repeatability. This method is rec-
o m m e n d e d therefore as a standard test
method for gas permeability measurements.
All bore hole methods frequently failed due
to leakage of the pressurized or evacuated
bore hole - either due to failure of the seal-
ing or because of pre-damage introduced
Fig. 7 - Results of the intercomparison testing of gas permeability: the figures
on top of the horizontal lines represent the results of pre-testing in the manufac-
while drilling. For laboratory tests, it has
turer's laboratory. been recommended therefore to cast the
cavity instead of drilling [4]. Furthermore,
the flow path of the permeating gas is not
known in most cases; therefore, these
methods may only allow a relative ranking
of concretes based on an index instead of
calculating a true materials parameter.
Nevertheless, these methods may be used
for on-site testing on a comparative basis,
but it must be kept in mind that they are
very vulnerable to moisture profiles in the
near-surface zone, e.g. the concrete cover.
A simultaneous measurement of the mois-
ture concentration, as foreseen in Parrott's
method [4], is mandatory for the interpre-
tation of the results. Furthermore, the
composition of the concrete in the surface
zone is different from the deeper sections;
results on the bore hole techniques do not
represent the properties of the bulk con-
crete, and they do not closely correlate with
Fig. 8 - Results of the intercomparison testing on the capillary absorption of water. results from Cembureau tests.
For the capillary absorption of water, the
modified Fagerlund test is recommended; it
is easy to conduct and shows only very little
scatter of the test results.
For the measurement of chloride ion dif-
fusion, no standard test method which could
meet the set of requirements can be recom-
mended as of yet. The immersion test yields
an apparent diffusion coefficient for chloride
ions since binding of chlorides by the
cement's hydration products is not consid-
ered in the calculation of the diffusion coeffi-
dents from total chloride concentration pro-
files. The method requires too much time
and labor for routine testing in production
control, but it may represent a powerful tool
for research or pre-testing of selected con-
cretes. In the migration tests, concepts have
Fig. 9 - Correlation between carbonation depth (1 year) and coefficient of gas been developed to include further analytical
permeability. steps aside from recording the electric charge

170
TC116-PCD

ratories, a close correlation has been found

for the rate of carbonation not only with the
gas permeability, but also with the water
absorption, e.g. as given in Figs. 9 and 10.

4.4 Second Intercomparison

Testing
4.4.1 Experimental results
The second intercomparison testing con-
centrated on those test methods that
emerged from Work Package 3 as recom-
mended standard test methods. From Work
Packages 1 and 2, pre-conditioning method
II with an equilibrium moisture content at
75 percent r.h. was recommended as the
Fig. 10 - Correlation between carbonation depth (1 year) and water absortion. standard. For the measurement of the gas
permeability, the Cembureau method [2]
was selected with minor modifications in
the test parameters; for the measurement of
the capillary absorption, the recommended
method relates to Fagerlund's method [7],
with amendments in specimen size and test-
ing parameters. It was felt that insufficient
experience is available at the present stage to
recommend one single standard test method
for the determination of the chloride ion
diffusivity that meets all requirements.
However, potential methods, which were
further developed in Work Package 3, have
been recognized. Therefore, corresponding
tests as described in [10, 11, 13] were also
continued in the frame work of intercom-
parison testing. On a limited scale, the
Sch6nlin gas permeability test [3] was also
considered because of its potential advan-
tages in rapid testing.
Three different concrete test series were
included, i.e. mixes A, B and C, each cured
Fig. 11 - Results of the intercomparisontestingof gas permeability.The firstcol- for 3 days. After curing, the specimens for gas
umn in eachblockrepresentsthe resultof pre-testingin the manufacturer'slabora- permeability and water absorption testing
tory; the last columngivesthe meanvalueof the intercomparisontesting.
were preconditioned according to method II.
Specimens for chloride diffusion testing were
passing; thus, the calculation of an apparent or an effective water saturated before their testing.
diffusion coefficient for chloride ions will be possible. After their shipment to the participating laboratories,
Basically, the tests on the chloride ion diffusion exhibited a the specimens were tested at an age of 63 days and 182
low scatter of the results within each laboratory, and they days, respectively.
may serve already for a relative ranking of different con- An initial evaluation of the test results shows that the
crete mixes. Cembureau gas permeability is a very sensitive method for
The compressive strengths of the ten concrete test monitoring variations in the properties of the test speci-
series of the program covered a range from 20 MPa to mens; consequently, a considerable scatter of the results is
approximately 80 MPa and as a general trend, the results encountered due to unavoidable inhomogeneities in the
showed that the transport of species through concrete material. However, the spread of the results obtained on
decreases as the compressive strength increases. This obser- different concrete qualities is also large; thus a clear dis-
vation is explained by the porosity of the concretes which tinction between different concrete qualities is provided.
controls both their strength and permeability. A further analysis showed that the scatter of the results
In order to demonstrate the testing concept, the observed between the participating laboratories is in the
observed durability characteristics were correlated with the same range as the scatter of the results observed on differ-
measured transport characteristics. In all participating labo- ent specimens in one laboratory. Fig. 11 presents the

171
Materials and Structures/Mat~riaux et Constructions, Vol. 32, April 1999

4.4.2 Repeatabilityand reproducibility of

the test methods
The test results from the intercompari-
son testing in the participating laboratories
were subjected to an analysis of the
achieved accuracy of the test methods.
Based on the definitions and principles set
forth in ISO 5725:1994 [14], partner 10
investigated repeatability and reproducibility for
the proposed standard test methods as well
as for the Sch6nlin gas permeability and
methods on the chloride ion penetration
[15]. Migration data from partner 2 were
not included in the data analysis because no
parallel testing took place using exactly the
same procedure.
Fig. 13 shows the repeatability of the
individual test methods. With reference to
the definitions of ISO 5725, repeatability
may be understood as the variability of the
test results w i t h i n one laboratory. In
Fig. 14, the reproducibility is shown for the
Fig. 12 - Results of the intercomparison testing of the capillary absorption of test methods. In contrast to repeatability,
water. The first c o l u m n in each block represents the result of pre-testing in the the reproducibility describes the variability
manufacturer's laboratory; the last c o l u m n gives the m e a n value of the intercom- of the test results between the participating
parison testing.
laboratories.
The analysis shows that the variability of
the test results between the laboratories is
in the same range as the variability of results
obtained in one laboratory for all test meth-
ods considered. The variability of the
results is high for gas permeability measure-
ments because of their high sensitivity to
differences in concrete properties. In this
connection, it must be assumed that no
homogeneous set of test specimens was
available for concrete mix C. The observed
reproducibility of the Cembureau method
in the range of 30% is acceptable in view of
the method's high resolution. The absorp-
tion of water exhibited only a small vari-
ability in the test results which remained in
most cases below 10%.
The observation of the chloride ion dif-
fusion was completed with rather low scat-
tering of the test results using one particular
Fig. 13 - Repeatability of the test methods in W o r k Package IV at an age of the method. A direct comparison of the differ-
concrete test specimens of 63 days.
ent methods is not possible because the
measured parameters include different
results of the gas permeability tests at an age of 63 days. implications. More research is needed in this field before
The resolution of the capillary suction measurements a routine test method may be recommended. A new
is less pronounced; however, this method exhibits only a PdLEM Technical Committee, TC TMC - "Testing and
small scatter of the test, results. Similar to the observa- Modeling the chloride penetration in concrete" took up
tions made for the gas permeability tests, the deviation the challenge to investigate more deeply into mechanisms
between the participating laboratories is in the same of chloride contamination of concrete. An outline of the
range as the scatter of the results within one laboratory. planned work program is presented in [20], and the
Although the method is less sensitive, the very low scat- progress of work may be tracked on the TC home page
ter of the results allows distinguishing between different under: http://www.rilem.ens-cachan.fr/tmc.htm.
concrete qualities. In Fig. 12, the absorption data are
shown for tests at an age of 63 days.

172
TC 116-PCD

Non-destructive Testing and Evaluation International

12 (1995) 53-73.
[3] Torrent, R., Frenzer, G., 'A method for the rapid
determination of the coefficient of permeability
of the covercrete', Proc. Int. Symposium on
Non-Destructive Testing in Civil Engineering,
Berlin, Germany, 1995.
[4] Jacobs, F., 'Permeability to gas of partially satu-
rated concrete', Mag. Concrete Research 50 (2)
(1998) 115-121.
[5] Abbas, A., 't~coulements gazeux dans les b&ons
partiellement satur& : Application fi la mesure de
perm&bilit& Thesis, (in print), Institut National
des Sciences Appliqu&s de Toulouse, France,
D&. 1998.
[6] Kollek, J. J., 'The determination of the perme-
ability of concrete to oxygen by the CEMBU-
REAU method - a recommendation', Mater.
8trnct. 22 (1989).
[7] Sch6nlin, K. F., 'Permeabilitiit als Kennwert der
Dauerhaftigkeit von Beton, Schriftenreihe
Institutes ftir Baustofftechnologie, Universit~it
Karlsruhe, 1989.
Fig. 14 - Reproducibility of the test methods in Work Package IV at an age of [8] Chen Zhang Hong, Parrott, L.J., 'Air permeabi-
the concrete test specimens of 63 days. hty of cover concrete and effect of curing', BCA
Report No. C/5, October 1989.
[9] Paulmann, K. and Rostasy, F. S., 'Praxisnahes
Verfahren zur Beurteilung der Dichtigkeit oberfl~ichennaher
Betonschichten im Hinblick auf die Dauerhaftigkeit', Institut
5. CONCLUSIONS Massivbau, Baustoffe und Brandschutz, TU Braunschweig, 1989.
[10] P-6000, Poroscope operating instruction, NDT James
In an experimental program, different test methods Instruments Inc., Chicago.
were evaluated with regard to their suitability for routine [11] Fagerlund, G., 'The critical degree of saturation method of
assessing the freeze/thaw resistance of concrete', Mater. Struct. 10
testing of concrete transport parameters. The results sup- (58).
port the Cembureau gas permeability method and a mod- [12] RILEM Technical Recommendation CPC 11.2: 'Absorption of
ified Fagerlund method for the testing of the capillary water by concrete by capillarity', Ibid. (June 1982).
absorption of water as routine test methods. [13] BS 1881 - Test methods for hardened concrete other than com-
pressive :strength: Part V - Initial surface Absorption of Water,
Corresponding instructions are developed as RILEM
1970
Technical Recommendations [21, 22]; these may also [14] Andrade, C., 'Calculation of chloride diffusion experiments
serve as a basis of&aft standards for testing hardened con- from ionic ruination experiments', Cement and Concrete Research
crete. The test results strongly depend on the precondi- 23 (1993) 724-742.
tioning of the test specimens. A standard procedure for [15] Tang Luping, 'Chloride transport in concrete - Measurement
preconditioning is therefore also mandatory. A method and prediction', Publication P-96:6 Chalmers University of
Technology, Department of Building Materials, G6teborg,
has been developed and the corresponding instructions Sweden, 1996.
are prepared for a RILEM Technical Recommendation or [16] ASTM-C-1202-91: Standard test method for electrical indica-
draft standard on preconditioning [23]. tion of concrete's ability to resist chloride ion penetration.
The measurement of concrete transport parameters is [17] Nord Test, NT Build 443: Concrete testing, hardened concrete,
very sensitive to the test procedure, in particular to the Chloride penetration.
[18] ISO 5725:1994: Accuracy (trueness and precision) of measure-
water content of the test specimens. Comparable results ment methods and results.
may be obtained only if the same testing procedure is [19] Tang Luping, 'Precision analysis according to ISO 5725:1994:
strictly followed. Therefore, consideration of the present Internal report to the coordinator of MATl-CT93-001',
test instructions is intented to provide a common basis Swedish National Testing and Research Institute, Boras,
for further research in which relations with durability Sweden, 1998, unpublished.
[20] RILEM Annual Report 1997, RILEM, Cachan, France, 1998.
characteristics must be included. On the basis of such [21] RILEM Technical Recommendation: 'Measurement of the gas
correlations, limiting values for concrete transport para- permeability by RILEM - CEMBUREAU method'.
meters may be developed and compliance criteria may [22] RILEM Technical Recommendation: 'Determination of the
be derived. capillary absorption of water of hardened concrete'.
[23] RILEM Technical Recommendation: 'Preconditioning of con-
crete test specimens for the measurement of gas permeability and
capillary absorption'.
6. REFERENCES
[1] Performance Criteria for Concrete Durability, J. Kropp, H. K.
Hilsdorf lEd.], RILEM Technical Report No. 12, F &N Spon,
London, 1995.
[2] Basheer, P. A. M., Long A. E. and Montgomery, F. R., 'Clam'
tests for measuring the in-situ permeation properties of concrete,

173