Determination of Alpha-Glycol Content of Epoxy Resins

GEORGE A. STENMARK
Shell Development Co., Emeryville, Calif.

M The aqueous periodate method for chloroform-alcohol mixtures. Because nol-A and of glyceryl phenyl ether were
the determination of a-glycol groups these resins dissolve readily in the latter obtained using the above procedure with
in organic materials is inapplicable solvent, this combination of reagent variations in water concentration, acetic
to epoxy resins because of their in- and solvent appeared favorable. acid concentration, time, and tem-
solubility in the reaction medium. perature as shown in Table I. It is
A satisfactory alternative involves PRELIMINARY STUDIES seen that bisphenol-A interferes con-
chloroform as the solvent and an
A 0.023M quaternary ammonium siderably under anhydrous conditions
alcoholic solution of a quaternary and that this interference is reduced
ammonium periodate as the reagent. periodate reagent was prepared by dis- markedly by addition of a trace of wa-
This method gives accurate results with solving periodic acid and an equivalent ter ; an increase in the amount of added
quantity of trimethyl benzyl ammo-
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

pure glycols and is sensitive to small water above 1% of the reagent volume
concentrations of glycol groups in nium hydroxide in methanol. does not result in a further reduction in
epoxy resins. Other functional groups, Procedure. Place 25 ml. of chloro- interference. A reduction in the acetic
including epoxy and phenolic groups, form and 1.2 ml. of glacial acetic acid acid concentration from 1.2 ml. per 25
do not interfere. in a 250-inl. glass-stoppered bottle. ml. of reagent results in a small reduc-
Downloaded via CORNELL UNIV on March 27, 2024 at 15:28:32 (UTC).

Introduce a sample containing no tion in interference. A reduction in

more than 0.3 meq. of -glycol. Add reaction temperature from 40° to 25°
connection with a recent study of exactly 25 ml. of 0.023M methanolic to 0° C. effects a considerable reduction
1~nthe effect of process variables on the periodate solution, swirl the mixture, in the extent of bisphenol-A reaction.
composition of epoxy resins, it was and let it stand at a controlled tem-
perature for a measured length of time.
With glyceryl phenyl ether the varia-
necessary to determine the a-glycol tion of reaction conditions results in
content of these materials. The resins Add 100 ml. of water, stopper the
under consideration were derived from bottle, and shake it vigorously. Add similar, though less marked, changes in
5 ml. of 10% sulfuric acid and 15 ml. of recoveries. On the basis of these data
epichlorhydrin and 4,4'-isopropylidene- 20% potassium iodide solution. Ti- the conditions chosen for further testing
diphenol [2,2 bis(p hydroxyphenyl)-
- -

trate with 0.1JV thiosulfate solution to and incorporated in the final method
propane or bisphenol-A], Possible the starch end point. wrere those indicated in Table I. Al-
sources of -glycol in the resins are
Application of the procedure at tem- though the recovery of the model glycol
glycerol and glycerol monochlorohydrin, was slightly low, 96% under these con-
resulting from side reactions of epichlo- peratures of 25° and 40° C. to the model
glycol, glyceryl -phenyl ether, indi- ditions, the interference of bisphenol-A
rohydrin, and glyceryl ethers of bis- under more vigorous conditions dictated
phenol-A, resulting from hydration of cated good accuracy. However, unreas-
this choice.
epoxide groups or from condensation onably high values obtained for epoxy
resin samples indicated side reactions A comparison of the values obtained
of monochlorohydrin with bisphenol-A.
with small amounts of phenolic impuri- with several simple compounds by the
As epoxy resins are either insoluble or
ties present in the samples. A study quaternary ammonium periodate
only slightly soluble in water, the possi- method under two sets of conditions
was made of the effect of reaction vari-
bility of applying the periodate reaction and by an aqueous method is shown in
in a nonaqueous medium was investi- ables on the interference of bisphenol-A
and on the completeness of reaction of Table II. The values obtained for sim-
gated. Pohle and Mehlenbacher (S)
used periodic acid in a chloroform-ace- glyceryl phenyl ether. In separate de- ple glycols by the former method at 0° C.
tic acid medium for the analysis of terminations the reactivities of bisphe- were 94 to 98% of the values obtained
water-insoluble glycerides. Attempts
to apply the Pohle and Mehlenbacher
reagent failed because of a side reaction Table I. Effect of Variables on Reaction of Bisphenol-A and Glyceryl Phenyl Ether
with free phenolic groups present in
these resins. Samples of bisphenol-A (Reagent. 0.0234/ trirr.ethylbenzyl ammonium periodate in methanol)
and resins high in phenolic functional Added, Reaction as -GIycol, Mole/100 G,
M1./25 Ml. Reagent Time, Temp., Glyceryl phenyl
groups, employed to test this effect, Water Acetic acid Min °
C. ether1*
showed that the periodate was reduced Bisphenol-A
to iodate under these conditions. Al- 0 1.2 90 40 0.68 0.588, 0.580
0 1.2 90 25 0.35 0.570
though the content of free phenolic 0.25 1.2 90 25 0.10
groups in these resins is generally low, 0.25 1.2 90 0 0.05 0.570
it is of the order of magnitude of the 0.5 1.2 90 25 0.11
1.0 1.2 90 25 0.08
-glycol content; thus the errors in- 0.25 0.8 90 0 0.012 0.534
troduced are large. As it was believed 0.25b 0.81 1506 06 0.017,0.027 0.562, 0.563
that the undesired side reaction was 0.25 0.6 90 0 0.010 0.524
caused by the relatively acidic condi- 0.25 0.6 150 0 0.017 0.553
tions in the Pohle and Mehlenbacher 0.25 0.4 30 0 0.005 0.416
0.25 0.4 90 0 0.006 0.512
procedure, a search was made for a less 0.25 0.4 150 0 0.005 0.545
acidic reagent. “
Value found by aqueous periodate method similar to that of Pohle and Mehlenbacher
Tests with other solvent systems (2), 0.584 equiv./100 g.
showed that a quaternary ammonium b
Conditions chosen for present method.
periodate is soluble in alcohol and in

VOL. 30, NO. 3, MARCH 1958 · 381

 ethers in these materials. Two resins
Table II. Reaction of Glycols and Epoxides with Quaternary Ammonium Periodate which were prepared in the presence of
Quaternary Ammonium Periodate, % unusually high concentrations of water,
150 minutes, 90 minutes, resins 3 and 4, exhibited relatively high
Aqueous 0° C., 0.8 ml. 40° C., 1.2 ml. -glycol values. Further evidence of
Compound Periodate,* % acetic acid acetic acid
undesirable side products in these ma-
Ethylene glycol 97.8, 98.3 93.2, 94.3 100.0, 99.3 terials was indicated by low epoxide
Glycerol 99.5 98.4 100.0. 102.3
98.2 94.4, 94.8 values and high hydroxyl values, re-
Glyceryl -phenyl ether 98.8, 97.6
Glyceryl a-monochlorohydrin 100.0 94.5, 94.3 99.6 98.8 sulting from the presence of chlorohy-
Epichlorohydrin <0.2, <0.2 <0.2 drins, other secondary hydroxyl groups,
Allyl glycidyl ether <0.2, <0.2 0.4, 0.2 and phenolic hydroxyl, as well as gly-
°
See a, Table I. cols. The analysis of resins 5, 6, and 7,
representing a wide range in molecular
weight, indicated the presence of small
by the aqueous method. Reaction with concentrations of ce-glycol configura-
the quaternary ammonium periodate re- Table Analysis of Epoxy Resins tions.
agent under more vigorous conditions, It was of interest to analyze hydrated
Hy-
90 minutes at 40° C. with additional a-Glycol droxyl* Epoxide6 epoxy resins (products in which the
acetic acid, resulted in values that Epoxy Mole/ Equiv./ Equiv./ epoxide groups had been converted to
Resin 100 G. 100 G. 100 G.
agreed well with the aqueous values. -glyeol groups by treatment with di-
Thus it is possible to improve the ac- 1 0.001 0.007 0.569 lute aqueous acid). Three products
2 0.002, 0.037 0.529 studied:
curacy of the method for use with ma- 0.002
were
terials known to be free of phenolic im- 3 0.074 0.479 0.101
purities. A negligible reaction was ob- 4 0.054 0.642 0.056 A. A hydrated resin with a high
served with two epoxy compounds, 5 0.014 0.180 0.362 molecular weight, prepared by heating
6 0.027 0.262 0.184 a dioxane solution of the material with
epichlorohydrin and allyl glycidyl ether. 7 0.015 0.360 0.055 0.5% aqueous sulfuric acid for 17.5
REAGENTS
*
By lithium aluminum hydride (4). hours at 90° C. The product was
6
By hydrochloric acid-dioxane (1). washed with water and dried.
Prepare the benzyl trimethyl am- B. A hydrated resin with a low
monium periodate reagent by dissolving
2.6 grams of periodic acid, HjIOe,
in 475 ml. of methanol. Add enough
benzyl trimethyl ammonium hydroxide Table IV. Analysis of Epoxy Resins and Hydrated Products
to neutralize the acid. Add 15 ml. of
glacial acetic acid and 5 ml. of water. Epoxide, Hydroxyl, a-Glycol,
Equiv./ Equiv./ Mole/ Molecular
Potassium iodide, 20% aqueous. Resin 100 G. 100 G. 100 G. Weight
Starch indicator solution, 0.2%
aqueous. A. High molecular weight resin
Chloroform, c.p. Original 0.047 0.372 2985
Sodium thiosulfate, 0.1V. Hydrated 0.005 0.424 0.052 4441
Sulfuric acid, 10%. Mix 61 ml. of Theory for hydrated resin 0 0.466 0.042» 3010
0.0576
95% sulfuric acid with 800 ml. of cold
water. Dilute to 1 liter. B. Low molecular weight resin
Original 0.529 0.037 0.002 351
PROCEDURE Hydrated 0.002 0.710 0.268 881
Theory for hydrated resin 0 1.000 0.485 385
Add 25 ml. of chloroform to a 250-ml.
glass-stoppered bottle from a graduate. C. Crystalline fraction from
Weigh a 0.3- to 0.5-gram sample which Preparation B 0.501
contains not more than 0.3 mmole of Theory for diglycidyl ether
of bisphenol-A 0 1.064 0.532
-glycol. Place the sample in the
bottle, and after solution, cool the mix- *
This value does not include any -glycol originally present.
ture in an ice bath. With a pipet add 6
Calculated assuming an original -glycol content of 0.015 equiv./100 g.
exactly 25 ml. of periodate reagent,
swirl the mixture, and let it stand in the
ice bath for 2.5 hours. At the end of
the reaction period add 100 ml. of ice where molecular weight, prepared by heating
water. Stopper the bottle and shake it 50 grams of resin with 100 ml. of 1%
B = volume of sodium thiosulfate sulfuric acid for 12.5 hours at 100° C.
vigorously for 30 seconds. Add 5 ml. required to titrate the blank, The water layer was decanted, 100
of 10% sulfuric acid and 15 ml. of 20% ml. ml. of water wras added, and the heating
potassium iodide. Titrate with 0.1.V S volume of sodium thiosulfate was continued for an additional 5 hours.
=

thiosulfate to the starch end point. required to titrate the sam-

As the end point is approached, stopper The water-insoluble product was sepa-
and shake the bottle frequently and ple, ml. rated and dried.
V =
normality of the sodium thio- C. A crystalline product in very
vigorously. Run a blank in an identi- sulfate small yield obtained from the aqueous
cal manner, but omit the sample.
weight of the sample, grams layers "in B.
(The sample titration volume must be
at least 80% of the blank titration
volume in order to ensure an adequate RESULTS AND DISCUSSION Analysis of the original resins and of
excess of reagent.)
the hydrated products are summarized
Calculation. Calculate the a-glycol Table III shows the results obtained in Table IV. Included for comparison
content by the following equation: upon application of the method to are theoretical values based upon quan-
epoxy resins varying widely in composi- titative conversion of the epoxide groups
-glycol content, moles/100 g. =
tion. Low glycol values were obtained to -glycol groups. The epoxide values
(B S){N)
-
for low molecular weight resins 1 and 2, found for the hydrated resins indicate
(20)(IF) showing the virtual absence of glyceryl almost complete conversion of the origi-

382 · ANALYTICAL CHEMISTRY

 ACKNOWLEDGMENT
nal epoxide groups to other structures. groups without a corresponding increase
The hydrated resins showed higher hy- in hydroxyl content, as follows: The author expresses his appreciation
droxyl and -glycol values than the to A. P. Poenisch, who performed a
original resins; unfortunately the orig- R major portion of the experimental work.
inal high molecular weight resin was
O O OH
not available for -glycol determination, /\ l
I
LITERATURE CITED
but it is reasonable to assume an a-gly- —C—C— + ROH —C—C—
col value of 0.01 to 0.02 equiv. per 100 (1) Jungnickel, J. L., Polgar, A., Weiss,
grams in view of the value found for a F. T., Peters, E. D., “Organic
similar material (resin 7, Table III). This reaction is probable even in the Analysis,” Vol. 1, p. 127, Inter-
science, New York, 1953.
Neither the hydroxyl nor the a-glycol presence of water, particularly if the (2) Pohle, W., Mehlenbacher, V., J. Am.
value of the hydrated resins indicates resin is not in solution, as was the case in Oil Chemists’ Soc. 24, 155 (1947).
quantitative conversion to a-glycol preparation B. (3) Ibid., 27, 54 (1950).
groups. Substantial increases in mo- Analysis of the crystalline fraction C (4) Stenmark, G. A., Weiss, F. T., Anal.
Chem. 28, 1784 (1956).
lecular weight were observed with both agreed well with theory for the diglyc-
hydrated resins, indicating polymeriza- eryl ether of bisphenol-A, the expected Received for review July 22, 1957.
tion involving consumption of epoxy compound. Accepted October 1, 1957.

Nature of Bindschedler’s Green

Preparation for Analysis of Hydrazo Compounds
H. J. SHINE, ROBERT L. SNELL, and JOHNNIE C. TRISLER

Department of Chemistry and Chemical Engineering, Texas Technological College, Lubbock, Tex.

Bindschedler’s Green, highly suit- chloride solutions. The method is nitrosodimethylaniline hydrochloride,
able for analytical work, can be pre- very convenient and rapid, even though 10.2 grams (0.084 mole) of dimethyl-
pared by Wieland's (9) general all operations must be carried out in aniline, and 14.4 grams (0.084 mole) of
method. It is stable over long periods the absence of oxygen. A disadvantage sodium dichromate dihydrate, 20.85
of time and generally gives clear, is that occasionally a Bindschedler’s grams of a satisfactory product was ob-
tained (53% theory based on p-nitroso-
colorless end points in titrations with Green preparation may give end points
dimethylaniline hydrochloride). Wash-
titanium trichloride. A simple method which are off-color, either brown or ing was carried out twice with absolute
is given for preparing the red form of yellow, making accurate titration diffi- ethyl alcohol and once with ether, or
Bindschedler’s Green. The green and cult. In an attempt to prevent this once with acidified dichromate (5 grams
red forms are discussed; the struc- variation, the nature of the Bindsched- in 50 ml. of water and 5 ml. of concen-
tures of both forms are given from ler’s Green was studied. trated hydrochloric acid), twice with
analytical data, and, to some extent, ethyl alcohol, and once with ether.
from equivalent weights. Bindschedler’s Green (Red Form).
EXPERIMENTAL To a concentrated solution of the green
dye was added slowly, with stirring, a
Materials. p-Nitrosodimethylani- concentrated solution of fused zinc
Green is used as an line hydrochloride was prepared ac- chloride in water. The red solid sepa-
Bindschedler’s
analytical reagent in following the cording to the method of Hickin- rated immediately as a dense powder.
rate of rearrangement of aromatic bottom (6). When prepared by the It was filtered and washed with wrater,
hydrazo compounds. A method of method of Baeyer and Caro (1) the or ivith zinc chloride solution followed
analysis, first developed by Dewar (4), product was a lighter yellow and less by water, ethyl alcohol, and ether.
rapidly oxidizes the hydrazo compound dense. Titanium trichloride solutions, Determination of Equivalent
to the corresponding azo compound us- 0.0LY, were prepared from the 20% Weights. The titanium trichloride
commercial solution and stored in an solution was standardized against
ing Bindschedler’s Green as the oxidant. automatic buret under nitrogen or acidified O.OlvY potassium dichro-
Immediate oxidation occurs when an carbon dioxide. Potassium dichromate, mate solution (J), using Bindschedler’s
alcoholic or acetone solution of the Baker’s reagent grade, was recrystallized Green as indicator. A solution of
hydrazo compound is added to an excess twice from water and dried at 120° C. Bindschedler’s Green, approximately 3
of an aqueous solution of dye. The for 24 hours. grams per liter, was made and 20-ml.
method works equally well with benzene Bindschedler’s Green. The aliquots were titrated with standardized
or other water-immiscible solvents. method of Wieland (9) was used in titanium trichloride. All titrations
The green Bindschedler’s Green is re- principle. Modifications of the relative were carried out under a stream of
duced to the colorless base, as follows: amounts of reactants, washing of the nitrogen. Equivalent weights were cal-
dye, and drying by continuous pump- culated on a basis of a 2-electron change
ing under vacuum gave a green fluffy per molecule of the green solid and a
solid which was stable over long periods 4-electron change per molecule of the
of time. It gave clear, colorless end red.
points against titanium trichloride.
Washing, here, consists in removing
the dye after filtration to a clean beaker DISCUSSION
Colorless containing wash solution and thoroughly
dispersing the dye in the solution before The method of analysis was used to
The dye is titrated to a color-
excess filtering again through a clean funnel. demonstrate second-order hydrogen ion
less end point with aqueous titanous By using 16 grams (0.085 mole) of p- dependence in the acid-catalyzed re-

VOL. 30, NO. 3, MARCH 1958 · 383