V O L U M E 2 4 , NO. 8, A U G U S T 1 9 5 2
1327
Stearne Inc., and Elmer J. Lawson, Sterling-Winthrop Research Institute, for their helpful criticism and suggestions in the preparation of this paper. LITERATURE CITED
(1) Anderson, R. C.. and Chen. K. K.. J . A m . Pharm. Asaoc.. Sci. Ed., 35,353 (1946) (2) Baljet, H., Schweiz. Apoth.-Ztg., 56, 71. 89 (1918). (3) Bell, F. K.. and Krantz, J. C., Jr., J . A m . Pharm. Assoc., Sci. Ed., 37,297 (1948). (4) Bell, F. K., and Krantz. J. C.. .Jr.. b. Pharmucol. Etptl. Therap., 83.213 (1945). (5) Ibid.;87,198 (1946). (6) Ibid., 88, 14 (1946). (7) Canback, T.. J . Pharm. Phormacol., 1, 201 (1949).
(8)Canback, T.. Suensk Farm. Tid., 54, 201 (1950). (9) Kedde, “Bijdrage tot ket chemisch Onderzock van Digitalispreparaten,” dissertation, Leiden, 1946. (10) McChesney, E. R., et al., J . Am. Pharm. Assoc., Sei. Ed., 37, 364 (1 . 948). . (11) Mulliken, R. S., J. Am. Chem. Soc., 74,811 (1952). (12) Raymond, W. D., Analyst. 63, 478 (1938). (13) Rice, W., Chairman, Contact Committee on Digitoxin, Eli Lilly & Co., Indianapolis, Ind., private communication. (14) Tansey, R . P., and Cross, J. JI., J . A m . Pharm Assoc., Sci. Ed., 39, 660 (1950). (15) United States Pharmacopeia, XIVth ed., p. 180, 1950. (16) Warren, 4.T., Hon-land,F. 0..and Green, L. W., .I. A m . Phnrm. Assoc., Sci. Ed., 37, 186 (1948). (17) Winthrop-Stearns Inc., unpublished reports. RECEIVED for revieiv l l a r c h 6, 1952.
Accepted June 7 1952.
Classifying Butyl Rubber with Respect t o Modulus A Chemical Method L. L. CURRIE Butyl Control Laboratory, Polymer Corp., Ltd., Sarnia, Canada The stress-strain method of determining the modulus of Butyl rubber possesses severe shortcomings when applied to plant control, the most serious being the time lag in reporting reliable results. A chemical method is described which is less time-consuming and at the same time more reproducible. The basic iodine-mercuric acetate method for Butyl unsaturation has been improved by the standardization of reaction conditions, and a correlation between unsaturation and stress strain values in the National Bureau of Standards formula has been developed. This comprehensive study has covered commercial grades of Butyl rubber over an 18-month period. The application of the method to plant control is described, and improved product uniformity is indicated. Reliable results suitable for control purposes are available in 3 to 4 hours.
I
N THE commercial production of Butyl rubber, the deter-
mination of the modulus is of paramount importance. In addition to being the most important product specification, this curing characteristic is of primary concern to both the plant operatoi and the rubber processor. Since the inauguration of the synthetic rubber program, the modulus level has been determined by the stress-strain method of compounding, curing, and testing sample specimens of the finished polymer. In spite of numerous improvements ( 8 ) , the stress-etrain procedure is time-consuming and unreliable. Mill roll and press temperatures, humidity conditions, and chemicals are variables that require constant attention. The most serious shortcoming is the 24-hour time lag between the time of sampling and the reporting of a result. Consequently, stress-strain testing has been of little value in the control of Butyl rubber production. For maximum product uniformity, a rapid reliable method for evaluating the effect of plant variables is required. This paper deals with the development of a chemical method for predicting modulus and its adoption for plant control purposes. This study has been based on the testing of the raw isoprene Butylpolymers listed in Table I, and its application to vulcanized stocks would require further work. Although all the Butyl polymers listed in Table I possess low unsaturation compared to natural rubber, i t has been shown (9) that each possesses specific properties depending on the amount of diolefin present. The properties chiefly affected are rate of vulcanization and the characteristics of the stress-strain curve. In developing a chemical method for predicting the rate of vulcanization, it has been necessary to improve the reproduGibility of the method for determining the unsaturation content of
Butyl rubber and to correlate the unsaturation values with the modulus values as measured by stress-strain testing. HISTORY
Iodine chloride ( 3 , 7 ) has been the principal reagent used for the determination of the unsaturation of Butyl polymers. Rehner (6) developed a satisfactory but time-consuming method based on the limiting viscosity of the polymer solution after degradation by ozone. Gallo, Wiese, and Selson ( 1 ) introduced a procedure involving a reaction between the pol-mer and iodine in the presence of mercuric acetate and trichloroacetic acid. This method is a rapid, reproducible test for determining the relative unsaturation of raw isoprene-isobutylene polymers and with slight modification was used in this study. PROCEDURE
A 5.OC-gram sample of the polymer weighed on an analytical balance is cut into small pieces and placed in 500 to 600 ml. of Table I.
Commercial Butyl Grades
hlooney Range 40-Minute Feed (ML 8 Minutes Modulus Range. Commercial Stock” at 212’ F.) Lb./Sq. Inch Grades B-1. O 40-50 ,.... P B . 100, GRI-R-2 B-2.0 40-50 876-1 125 P B . 2 0 0 , GRI-50 B-3.0 40-50 1125-1375 P B . 300, GRI-Y-15 B-2.5 70-80 1125-1375 PB.301, GRI-18 B-4.0 40-50 1326-1575 P B . 400, GRI-Y-25 a B-1.0 represents a mixture of 99% isobutylene a n d 1% isoprene, B-2.0 represents a mixture of 98% isobutylene a n d 2% isoprene, eto.
ANALYTICAL CHEMISTRY
1328 C.P. carbon tetrachloride contained in a quart reaction bottle. T h e bottle is placed in a mechanical shaker and allowed to remain until the sample is completely dissolved (1 to 2 hours). The polvmer-carbon tetrachloride solution is then transferred to a 1000-ml. volumetric flask and the reaction bottle is thoroughly rinsed with carbon tetrachloride. T h e rinsings are added to the flask, and the contents are made u to volume with carbon tetrachloride and thoroughly mixed by sgaking.
160C
1400
4
0:
cn
3
g20c 0
0
DISCUSSION OF PROCEDURE
The carbon tetrachloride-iodine solution is unstable in the presence of light. The solution must be blanketed with nitrogen during storage and thoroughly mixed before using. Although i t has been noted ( 1 ) that the reaction with iodinemercuric acetate is complete in about 0.5 hour, the reaction can, under routine test conditions, continue for 8 to 10 hours. However, by carefully controlling the reaction time a t 30 minutes, reproducible results can be obtained. This empirical method yields relative unsaturation values suitable for routine, control purposes. For research work where absolute unsaturation values are required, the iodine monochloride method developed by Lee, Kolthoff, and Johnson ( 4 )appears to be ideally suited. Sunlight must be avoided in this test, as extremely high results are obtained when the prepared reaction flasks are inadvertently placed in direct sunlight. T o appreciable difference was noted when samples were testcd in total darkness and in ordinar: laboratory light. It has been found advisable to teat a standard polymer along with regular samples, so that any discrepancies arising from techniques, solutions, or reaction conditions may be observed.
5
DEVELOPMENT OF CORRELATION W
As Butyl rubber is usually compounded and vulcanized and the physical properties of the vulcanizates determine the utility of the finished products, the unsaturation values obtained by the method outlined above were correlated with modulus values determined on the vulcanizates compounded in the carbon black test recipe shown in Table 11. Production composites from the commercial unit were tested by both methods during the course of this study, which extended from mid-1950 until the end of 1951. .4 wealth of data covering all the commercial Butyl grades listed in Table I was accumulated in this manner. The stress-strain testing $vas carried out accord-
I-
~rooo 5’
* 0
80C
J I
60C
1.5 2.0 25 MOL. PERCENT UNSATURATION Figure 1. Relationship betw-een Modulus and Unsaturation for Low 1Iooney Butyl Rubber I
)
Aliquots (100 ml.) of the polymer-carbon tetrachloride solution from the volumetric flask are pipetted into clean, dry 500ml. iodine flasks. I n order, 5 ml. of trichloroacetic acid solution (20 grams of C.P. trichloroacetic acid made up to 100 ml. with C.P. carbon tetrachloride), 25 ml. oi iodine solution (12.5 grams of resublimed iodine dissolved in 1 liter of carbon tetrachloride), and 25 ml. of mercuric acetate solution (30 grams of C.P. mercuric acetate in 1 liter of glacial acetic acid) are pipetted into each flask. T h e flasks are shaken and allowed to stand for exactly 30 minutes in ordinary light. hfter reaction, 75 ml. of 7.5% aqueous potassium iodide solution are added to each flask and the contents are titrated with 0.1 AV sodium thiosulfate. The iodine in the aqueous layer is titrated until the solution has but a faint yellow tinge. A small amount of 0.5% starch solution is added and the titration is continued dropwise with vigorous shaking of the stoppered flask until the colorless end point is reached. A blank of carbon tetrachloride is run n i t h the sample. Mole yo unsaturation =
( B - S)(’v)1’87 - (0.575 X weight % stabilizer) wt. of sample where B = titration for blank, nil S = titration for sample, ml. S = normality of thiosulfate 1.87 = factor derived from the equivalent weight of the structural unit (isobutylene), 3 atoms of iodine per double bond being consumed for tertiary olefins As phenyl-2-naphthylamine and -4geRite Stalite, the common stabilizers used in Butyl rubber, react with iodine, a correction (determined experimentally to be 0.575 X weight % of stabilizer) must be applied to each result.
- _ _
I .o 1.5 2.0 MOL. PERCENT UNSATURATION
2
Figure 2. Relationship between Modulus and Unsaturation for High Mooney Butyl Rubber
1329
V O L U M E 24, NO. 8, A U G U S T 1 9 5 2 EFFECT O F \IET4LLIC STE4RATES
Table 11. Carbon Black Test Recipe Parts by Weight Butyl polymer XBS standard Channel black Zinc oxide Stearic acid Sulfur Xercaptobenzothiazole Tetramethylthiurarn disulfide_ Vulcanization, 40 minutes at 30r F
200.0 100.0
10.0 6.0
4.0 1.0 2.0
ing to the specifications for government synthetic rubbers issued by the Reconstruction Finance Corp. ( 5 ) . Figure 1 shows the unsaturation values for the low hIooney gradeq, corrected for the stabilizer content, plotted against the modulus values. This graph include8 the B-1.0, B-2.0, B-3.0, and B-4.0 polymers with a Mooney viscosity of 40-50 (ML 8 minutes at 212" F.) and indicates that a straight-line relationship exists between unsaturation and modulus for each grade and for all grades having the same average Mooney viscosity. Figure 2 shows the correlation derived by plotting the same values for the high ?*looney grade.
1
1600 7 1
-
uj,1400
a:
v--y o 00
P
7
The effect of nonpolymeric sub-tances on the correlation has been investigated in this study. .-\I1 the samples used in this program contained 1.0 to 1.5 weight % stearate and the correlation established in Figures 1 and 2 is valid only within this range. The effect of stearate has been studied by adding zinc stearate in the compounding recipe (Table 11) of a low stearate polymer. Table 111 shows t h a t the determined modulus using the revised recipe agrees with t h a t predicted by unsaturation values based on polymers of high stearate type. It is concluded that the stearate content of the polymers decreases the modulus as measured by stress-strain testing but has no noticeable effect on the unsaturation values. EFFECT OF hIOOYEY SCORCH
The initial cure time of Butyl rubber as measured by the Mooney scorch value obtained using the gum test recipe (5) (Table IV) has been suggested for predicting the curing properties of the polymer and i t has been shown ( 2 ) t h a t nonpolymeric constituents-i.e., aluminum chloride, sodium stearate, etc.markedly affect the ?*looney scorch values using this recipe. Butyl samples having Mooney scorch times from 8 to 26 minutes have been studied and there is no indication that theinitial curing time in the gum recipe affects the accuracy of this unsaturationmodulus correlation. The normal Mooney scorch range is much narrower than the range covered in this study and it is concluded that Mooney scorch can be ignored in the application of this correlation.
/
Table 111. Effect of Stearate on Rlodulus Values I
Sample 40-Minute modulus, lb./sq. inch Predicted from unsaturation B y stress-strain ( 0 . 4 % stearate in polymer) By stress-strain (stearate increased to 1 . 5 % )
1235
101:
1285
I070 ~-
Table I V .
Gum Test Recipe
Polymer XBS standard Zinc oxide Tetraniethylthiurani disulfide Sulfur
3
1032
1225
-
10-50 MOON
I1
-
Gram200 0 4.0 1.2 4.0 209.2
REPRODUCIBILITY OF METHOD
600Figure 3.
3 2 i MOL. PERCENT UNSATURATION Effect of \Iooney on Alodulus-Lnsaturation Relationship of Butyl Rubber
Table V s h o w the modulus values (as predicted by un;qaturation) on a single sample of standard Butyl polymer. These values n-ere obtained on a routine basis over a 1-month period, one sample being tested each day by a single operator. The standard deviation calculated for this data is 12 pounds per square inch compared with the standard deviation of 20 pounds per square inch for the stress-strain values on a standard Butyl polymer shown in Table 1.1 The F value for the data shown in
Table V. 40-Minute Modulus (as Predicted by Unsaturation) on Standard Butyl Polymer h
Equ:itions t o fit the experimental d a t a represented by Figures 1 and 2 have been calculated by the method of least squares and are as follon s:
High Mooney grade (B-2.5) 40-minute modulus (Ib. /sq. inch) = 478.68 X mole % unsaturation 484.82 Low L1ooney grades (B-1.0, B-2.0, B-3.0, and B-4.0) 40-minute modulus (Ib./sq. inch) = 536.65 X mole % unsaturation 268.87
1226. 1235 1225 1225 1230 1225 1230 1255 1230 1215 1235 1225, 1230: 1240: 1215: 1240: 1225: 1255: 1235: 1210: 1200: 1130: 1215, 1225. 1210, 1235, 1225, 1225, 1215, 1220, 1210 Average value. lb./sq. inch 1225 31 S o . of tests Standard deviation. Ib./sq. inch 12 .-
+
(1)
+
(2)
Figure 3 combines the experimental correlations established in Figures 1 and 2 and shows t h a t the higher Mooney polymers possess a lower degree of unsaturation for the same modulus.
Table VI.
40-3linute Modulus (as Determined by StressStrain) on Standard Butyl Polymer B
1295, 1315, 1315, 1285, 1275, 1295, 1265, 1265, 1265, 1295. 1275. 1315. 1305, 1305, 1315, 1265, 1305. 1315, 1295. 1325, 1303. 1315. 1315. 1325. 1355, 1315. 1285, 1285. 1295. 1315 Averaze value. lb./sa. inch 1300 No. of-tests 30 Standard deviation, lb./sq. inch 20
1330
ANALYTICAL CHEMISTRY ~~
SPECIFICATION
I IMlT
1350
21300
0
cn
3 -I
2 l25C 0
I
i 5 120c 0
t I150 S PEClFlCAT ION
LIMIT ,
I
2
3
4
5 6
PRODUCTION
Figure 4.
Table VII. Sample 1
2 3
? 6
7
8 9 10 11 12 13 14
I
O
!
7 8 9 IO I I 12 13 14 15 16 17 18 19 DAY
Typical Quality Control Chart for Butyl Rubber Production
Data Obtained by Unsaturation and StressStrain 40-Minute IIodulus, Lb./Sq. Inch By unsaturation By stress-strain Difference 1325 1300 +25 1315 1350 - 35 1280 1280 ... 1250 1240 '0 1250 1255 1275 1250 -2; 1265 1260 + a 1283 1245 40 1270 1260 10 1265 1290 -23 1240 1265 -25 1255 1260 - 5 1245 1220 i s 5 1320 1295 - 25
+
+-
may be analyzed by the procedure outlined and reliable results reported in 3 t o 4 hours after the samples are prepared. The availability of this information has made the commercial production of Butyl rubber ideally suited for quality control. Using normal control charts n i t h inner control limits estab!ished within the specification band, a marked improvement in product uniformity has been achieved. Figuie 4, a reproduction of one of these charts, illustrates the excellent control v hich can be attained using modulus by unsaturation values. At the same time stress-strain testing, for plant control purposes, ran be drastically reduced. Table VI1 s h o w a comparison of modulus values obtained by both unsaturation and stress-strain on fourteen consecutive production composites. On a routine basis the values by unsaturation and by stress-strain testing agree xithin the esperimental error of the latter test. As the average difference and the standard deviation of the differences are very close to the same value, it can be concluded that the same value i s being obtained by each method. .4CKYOWLEDG>IENT
The author is indebted to the Polymer Corp., Ltd., for permission t o puklish this work, to C. A4 Finigan and H. C. Harlrnese for valuable assistance during the course of the study, to L. P. Gelinas who provided the stress-strain data, and to H. P. Berk, J. S Frayne, and J. E. Zatylny who performed the experimental unsaturation work. LITERATURE CITED
4PPLICATION OF CORREL4TIONS
(1) Gallo, S. G., Wiese, H. K., and Nelson, J. F., I n d . Eng. Chem.. 40, 1277 (1948). (2) Harkxiess, H. C.,Polymer CorD.. Ltd.. unDublished data. (3) Kemp, A. R., and Peters, H., IND.ESG. CHEM.,ANAL.ED.,15, 453 (1943). (4) Lee, T. S.,Kolthoff, I. II.,and Johnson, E., . ~ S A L , CHBM.,22, 995 (1950). (5) Reconstruction Finance Corp., Office of Rubber Reserve, Specifications for Government Synthetic Rubbers. (6) Rehner, J., Jr., Ind. Eng. Chem., 36, 118 (1944). (7) Rehner, J., Jr., and Gray, P., IND.ESG.CHEM...$SAL. ED., 17, 367 (1945). ( 8 ) Stiehler, R. D., and Hackett, R. W., d s a ~ CHEY . , 20, 292 (1948). (9) Thomas, R. &I., Lightbown, I. E., Sparks, IT-. J., Frolich, P. K. and Murphree, E. V.,Iiid. Eng. Chem., 32, 1283 (1940).
The unsaturation method of predicting the modulus level of Butyl rubber based on the correlations established in Figures 1 and 2 can be u L Afor plant control. Composite and spot samples
RECEIVED for review February 29, 1952. Accepted May 29, 1952 Presented before a joint meeting of t h e Ontarlo and Kitchener Rubber Groups, Chemical Institute of Canada, Kitchener, Ontario, November 13 1951.
Average difference 19 Standard deviation of difTerences 23
Tables V and VI is 2.8, while the critical value from the F tables is 2.38. Thus it can be concluded that the standard deviation of the unsaturation prediction is significantly lower than the standard deviation of thc stress-strain data.
