Curing Rate of Rubber EFFECT OF PHOSPHATE BUFFER MIXTURES ONRATEOFCUREOFCREPERUBBER E. B. NEWTON' Malayan Research Laboratories, The B. F. Goodrich Company, Kuala Lumpur, Federated Malay States natural fatty acids: Oenslager (8) in 1906, by vulcanization in the presence of aniline and later thiocarbanilide, w&s able to produce good physical properties in solvent-extracted rubbers, an observation which became the basis of a tremendous advance in rubber technology in consuming countries. But with the increase in planted acreage of Hevea rubber in the East shortly after, systematic investigations were undertaken by workers in that region t o establish plantation rubber as equivalent in quality to Para on a scientific basis. This early work was both successful and comprehensive. Though many workers contributed, the extent of the field covered is shown in the summaries of the classic works of Eaton and co-workers (3)9de Vries ( I I ) , and Whitby (16). More recently Hastings and Rhodes (6) and Sackett (9) extended the research on variability. Variability in rate of vdcanization can arise from the composition of the latex a5 it comes from the tree as well as from the operations employed to convert it into crude rubber. Whitby and de Vries noted that the higher rates of cure are associated with those rubbers which contain the higher quantity of serum substances, but just what portion of the serum solids, aside from the fatty acids and from the nitrogen-containing products of putrefaction which appear t o be true accelerators @),are responsible for this relation has never been definitely stated. De Vries (10) was once concerned over the possible effect of adding sodium bisulfite to latex as an inhibitor of oxidase action in the manufacture of crepe. In those days acetic acid wm generally used as the latex coagulant, and since this combination might be expected to give rise to sodium acetate in the rubber, he tried the addition of this compound directly to latex prior to coagulation. Rubber thus prepared cured somewhat faster in a rubber-sulfur stock. He then soaked freshly machined sheets in a solution of sodium acetate and obtained a greater effect. Morton (7) added crystalline sodium acetate directly t o rubber on the mixing mill and obtained excellent activation in accelerated stocks. Several years ago in the course of work on vulcanization, the use of sodium acetate as an activator was repeated. It was compounded (Table I) both in a rubber-sulfur stock and in the A. C. S.Captax (mercaptobenzothiazole) test recipe;
The hypothesis is advanced that the rate of cure of rubber by sulfur and heat is a function of the acid-base equilibrium within the stock during cure. In a rubber-sulfur stock this equilibrium could develop between the serum material present to a greater or lesser extent i n all commercial grades of crude and any reaction products of the vulcanization reaction. In cornpounded stocks any of the pigments not entirely inert would be expected to contribute to a shifting of this equilibrium. Whole rubbers, made by evaporation of latex and containing high proportions of Berum material, act as though their buffer capacity is greater than washed rubbers such as crepe. Equilibria on the alkaline side would favor and those on the acidic side retard the vulcanization reaction.
T
HE phenomenon of variability in rate of vulcanization of crude rubber by sulfur and heat has been studied by rubber technologists for more than forty years. Whether the early workers recognized differences in any particular commercial grade is difficult t o say, as a result both of the lack of discriminating tests and of the secrecy which was formerly maintained in the industry. However, three developments, all coming near the beginning of this century, combined to make workers more aware of any differences which might exist: Weber's work on the chemistry of rubber which led to the first theory of cure (IS) and to the idea of the sulfur coefficient of vulcanization (14 ) ; the wider availability and use of standardized tensile testing machines, although earlier but limited use of such machines was known (16); and finally, the rise of the plantation industry in the Middle East. The first two of these developments supplied methods of sufficient accuracy to detect differences other than by the empirical comparisons of hand tests. The third gave rise to many investigations designed to ascertain differences between the newer plantation and the standard Para grades. All the early work on variability appears to have been done in consuming countries, and some of it was of primary importance. For example, Weber and Glidden (6) discovered in 1903 or 1904 the activating effect of oleic acid on litharge stocks which were compounded with rubber deficient in I
TABLE I. RECIPES FOR STOCK WITH SODIUMACETATEAS ACTIVATOR Recipe No. Rubber
Sulfur Zino oxide Stearic acid
CRPtRX NaCzHsOz.3H~O
Present address, The B. F. Goodrich Company, Akron, Ohio.
374
Rubber-Sulfur Type 6DK 100 100 10 10 D
... ... ... ...
...
... ... 1.0
AM 100 3.5 6 0.5 0.5
...
Captax Type 8AZ 100 3.5 6
0.5 0.5 1.0
9AE 100 3.5 6 0.5
...
1.0
March, 1942
INDUSTRIAL AND ENGINEERING CHEMISTRY
37s
first latex crepe, a slow-curing type, was used in one case, and a fast whole rubber (latex-sprayed) in the other (Table 11). The crepe was made as follows: Field latex was diluted t o 15 per cent dry rubber content with tap water and coagulated by addition of 2 per cent acetic acid to pH 4.8 *O,l, determined by a glass electrode; the coagulum was creped the following day under water spray, rolled thin (total of fifty passes), and dried in air a t 50' C. (122' F.). The whole rubber was a commercial grade (called L. S. hereafter) made by spray drying ammoniated latex. The testing procedure was as follows: Batches of 650 grams of rubber were mixed on a 6 X 12 inch mill (gear ratio 1:1.4) and sheeted off (not calendered) to approximate mold size. They were press-cured the following day in chromium-plated molds (unless otherwise stated), 6 X 8 X 0.1 inch, without the use of any mold wash or other parting layer. Stress-strain data were obtained on the day following cure in 1 X 1 / 4 inch dumbbell test strips on a standard Scott autographic tensile machine (jaw separation 20 inches per minute), using two observers. Room temperature during mixing and testing was 28-30' C. (82.5-86" F.), average relative humidity 75 per cent. Table I1 shows that the whole rubber cures much faster than does the crepe in both the rubber-sulfur and the Captax stocks, a well known fact. I n recipe 9AE, which contains no organic accelerator and insufficient sulfur for vulcanization in the absence of accelerator, the sodium actetate is incapable of performing the functions of such an accelerator as Captax; but in both the rubber-sulfur and the Captax stocks sodium acetate speeds up the vulcanization whether whole latex rubber or crepe is used, but especially with the latter. I n 1922 Bruce (1) published a set of the few analyses extant of the ash of various rubbers including commercial sheet, commercial crepe, and a whole rubber prepared by evaporation of latex. The principal constituents of each ash were potassium and phosphorus. The ratios of KzO to PzO, in the sheet and crepe were similar (approximately 0.55), whereas the ratio in the whole rubber was nearly 1.8. Bruce did not suggest, nor is it suggested here, that these elements were present in the rubber entirely as inorganic phosphates. Nevertheless, these data, together with those above on sodium acetate, might suggest that the latex-sprayed rubber cures faster than crepe because the serum material-that is, the nonrubber constituents-remaining in the finished crude is more alkaline in the case of L. S. than in crepe. It has been known for many years that, in general, addition of alkaline materials to a rubber-sulfur mix accelerate while acidic materials retard cure. In a given series of rubbers showing naturally different rates of cure it might be assumed that the acid-base equilibrium in the nonrubber constituents of those rubbers would form a gradation from an acidic to a basic condition. There is also the further consideration that the serum material remaining in a rubber like L. S. might have a greater buffer capacity in the presence of free hydrogen sulfide generated during cure than does serum material in a rubberlike crepe. Biological fluids generally, including latex serum, are well buffered to resist changes in pH, and it may be to this portion of the serum solids-namely, the inorganic matter or the salts of organic acids or proteins remaining in the finished crude-that the higher rate of cure is due.
Experimental To ascertain whether this view has any possible basis, a well-washed first latex crepe and a whole rubber (L. S.) were chosen, similar to the ones described in Table I. These rubbers were compounded in both rubber-sulfur and A. C. S. Captax test recipes, but except for two examples, discussion here will be confined t o the former stock as it illustrates the
I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
376
~
~~
~
~~~
~
_
_
_
_
_
_
_
~ ~
Vol. 34, No. 3 ~_
_
_
_
_
~
~
TABLE 111. IKQREDIENTS OF PHOSPHATE-BUFFERED RECIPESPRESS-CURED AT 298 F. O
Reoipe No,
sc;"p,", rubber KHzPO4 NmHP04.12HzO pH of ealts in 1/15 M aqueous soh.
D
D.4
100 10
100 10 0.86 0.12
DB 100 10 0.82 0.24
5.6
5.9
..... .
...
DC DD 100 100 10 10 0.74 0.64 0.48 0.72 6.2
6.4
DE
DF
DG
DH
100 10 0.54 0.96
100 10 0.45 1.20
100 10 0.36 1.44
100 10 0.27 1.68
DI
6.6
6.8
7.0
7.2
100
10
0.18 1.92
7.4
DK
DJ 100 10 0.09 2.16
100 10 0.9
...
7.7
4.5
DL 100 10
.... 2.39
Dl1 100 10
. 1.2Sa ...
9.0
4.0
Potassium mid phthalate in this case.
general trend. Into these batches was mixed a series of phosphate buffers of graded pH. Phosphate buffers were chosen because they were the most stable of a number tried in the presence of zinc oxide. Baker's reagent phosphates were used without further purification and were added to give a 11.1 concentration by weight on the rubber as shown in Table 111, The salts were dissolved in approximately an equal weight of distilled water and were added to the rubber running with a rolling bank on the mixing mill by distributing the solution from a pipet over the back roll. Most of the water appeared to be lost during mixing, and no porosity of the cured sheets was observed. Any caking of the salts on the back roll was worked into the rubber by further addition of small quantities of water. Batches were run with potassium hydrogen phthalate, potassium dihydrogen phosphate, and disodium hydrogen phosphate alone, as well as with the buffer mixtures. Recipe ingredients are shown on a batch weight basis, together with the pH value (to nearest tenth) of the salts in '/I6 M aqueous solution at room temperature. Curing and testing procedures were the same as those described above. The general effect, in the case of first latex crepe, of more rapid cure with increase in alkalinity or reduction in rate of cure with increase in acid component as shown in Table I V is best visualized if rate of cure, in minutes at 298' F., is plotted against pH of the salts in '/I6 M aqueous solution. Such a construction (Figure 1) is to be regarded as indicative of a possibility, not as a measured relation. Figure 1suggests that, whereas the well-washed crepe is susceptible t o treatment
4.0 5.0 6.0 7.0 8.0 9.0 pH OF BUFFER SALTS IN 1/15 M AQUEOUS SOLUTION
FIQURE1. EFFECT OF BUFFERSALTS ON RATE OF CUREOF RECIPES D TO DM, INCLUSIVE
O F INCREASE I N ACIDCOhfPONENT T ~ LIv. E EFFECT
Cure a t 298'
F.,Min.
60 90 120 150
180
D (Control) 700% 90 230 370 750 1220
800% 170 370 750 1620 2500
DE
DA 700% 60 220 400 920 1640
DC
800% 90 400 800 1830
..
700% 100
800% 180 500 1100 2150
280
550 1070 1650
..
700% 100 270 630 1180 1870
DD
800% 180 530 1360 2500
..
DE
800% 250 660 1550
700%
170 380 770 1510 2100
.. ..
DF
Tensile of Crepe Rubber in 700% 800% 700% 800% 200 350 280 490 375 820 700 1490 1400 2590 950 1630 ..
....
..
..
.. ..
Tensile of Crepe Rubber a t Break, in Pounds 60 90 120 150
180 Time, min.b
142
149
157
169
136
129
107
Tensile of Latex-Sprayed Rubber
'
15 30 45 60 90
700% 200 560 1030 1725
800% 500 1225 2200 1725
..
..
700% 280 450 850 1650
800% 470 910 1800
....
*.
800% 490 1110 2150
700% 260 425 1030 1800
.. ..
..
800%
700% 240 450 960 1670
400 900 1880
.. ..
..
800% 470 900 1840
700% 260 420 960 1690
.. ..
..
700% 240 450 860 1660
800% 400 920 1710
....
..
700%
800% 450 1150 2300
270 550 1225 1990
.. ..
..
Tensile of Latex-Sprayed Rubber a t Break, in Pounds 15 30 45 60 90 Time. min.b
810 (880%)
44
47
41
48
49
49
42
March, 1942
INDUSTRIAL A N D ENGINEERING CHEMISTRY
with the buffer salts, the L. S. containing all the serum material does not respond to any extent to the same concentrations of buffer salts; it behaves as though the serum material is of itself already a n effective buffer. If this were true, then one might expect a wide variation in rate of cure of rubber made from identical latex, depending on whether coagulation occurred on the acid or alkaline side and on the degree of effective washing the coagulum subsequently received. This has long been known to be the case. The presence of water in the crepe rubber stock at time of cure has an influence on the rate. This effect is shown in Figure 2, drawn on the same basis and with the same reservations as Figure 1. A sufficient quantity of distilled water was worked into each batch just before sheeting it off the mill. Under the conditions of open-mill mixing and overnight storage of the uncured but covered sheeked stock, accurate control of the average water content of each batch was not obtained, but it was measured. Extremes were 1.00 and 2.04 grams of water per 100 grams of rubber, the majority of the batches being near 1.5. These weights were taken just before the sheeted stock was placed in the curing molds. Curing in this series was a t 316’ F., the object of this higher temperature being merely to speed up the test. The effect of the water was to spread the curve, the slow rubber curing more slowly, the fast rubber more rapidly, with minimum alteration in rate of cure when the added buffers showed approximately pH 6.6 in aqueous solution. It is possible that this effect was due to better or more effective distribution of the added phosphates throughout the rubber mass, but it will be noted that the control stock without any added phosphates was also affected by the presence of water. Application of the treatment as a corrective measure in two supposedly acidic stocks is shown in Tables V and VI. Table V compares air-dried acetic-acid-coagulated sheet, with and M without added phosphate buffer salts (which alone in aqueous solution would show about pH 7.0), with air-dried alum-coagulated sheet, with and without added phosphate buffer salts. The stocks are based on the A. C. S. Captax test recipe.
ON
377
I n the case of the alum-coagulated sheet, the presumed acidity was thought to arise from residual coagulant in the rubber, probably in association with the protein. W t e v e r the cause, alum-coagulated rubber has long been known t o exhibit extremely slow cure. The data for compound AM-174 show this, all the cures being very tender. Addition of phosphate buffer (2AX-174), however, produces a stock curing faster to a given modulus than the acetic control but, as might be expected, not so fast as the acetic control sheet containing the same quantity of added phosphates. Both the alum and acetic sheets were made by coagulation of aliquots of identical latex.
!PH
I
201
I
I
I
I
I
I
I
I
5.0 6.0 7.0 6.0 9.0 OF BUFFER SALTS IN 1/15 M AWEMlS SOLUTION
FIGURE2. EFFECTOF WATERIN CREPE RUBBERON RATEOF CURE
Corrective measures were also tried for the slowing of cure experienced on addition of large volumes of channel gas black to vulcanizable stocks (Table VI). These formulas are based on the A. C. S. Captax test recipe. The trials arose from our observation that an aqueous slurry of channel gas black is acidiq(l6). The data indicate that with these particular stocks the addition of the phosphate mixture (3AD-143) has approximately the same effect on the development of modulus as an
RATEOF CUREOF PHOSPHATE-BUPFERDD RUBBERS DG
DH
DJ
DI
DK
DL
DM
Pounds per Square Inoh 700% 325 800 1700
800% 630 1650
700% 300
1080
800% 610 2190
..
..
..
.. ..
....
.. ..
.. ..
700% 350 1150
800% 820
700% 550 1300
..
..
......
......
800% 1180
..*.
....
....
700% 110 260 410 775 1270
800% 200 420 830 1630 2650
700% 590 1350
800% 1240 2480
.. .. *.
.. ..
..
700% 140 250 370 670 1100
800% 226 450 725 1460 2240
per Square Inch (Elongation in Per Cent) 2120 897 ) 2410 590 / 4 795 5 0 1)
101
89
214V(8800/ 2900” 830 1060 450 136511 656
540 375 ) 470 a40 c) 82
88
490 350 81
174
in Pounds per Square Inch 700% 350 610 1200 2030
..
800% 660 1200 2350
....
700% 300 520 1100 2000
..
800% 560 1210 2230
.. ..
700% 280 525 1130 1960
800% 590 1200 2200
..
..
700% 330 650 1290 2090
800% 570 1300 2400
.. ..
..
700% 240 420 890 1670
800% 410 900 1920 2780
..
..
per Square Inch (Elongation in Per Cent) 1070 8 8 5 7 ) 2250 905 1
0
Jaw break.
b Minutes
800% 290 630 1200 2370
..
..
2775 840 ) 1730 510 [680& 375
2250 17257) 2800 840 ) 590 395%) 44
.. ..
.. *.
700% 170 325 570 1300
%;71
E:; E%!
2925 2540 /745${ 835 960 465 01 42
800% 579 1310 2250
700% 300 580 1160
44
41
of cure to reach a modulus of 2100 pounds per square inch at 800%.
48
42
54
INDUSTRIAL AND ENGINEERING CHEMISTRY
378
-
TABLE
v.
TENSILE STRENGTHS
OF
Vol, 34, No. 3
ACETICACID-AND ALUXI-COAGULATED STOCKS
.lcetic Acid-Coagulated Sheet . -AM-173*-2AX-173br-AhI-174aPress Cu?,e a t 60°F.,I\lin. 600% 700y0 A t break BOO% 700% At break 000% 70076 2230 130 3310 (775%) 90 1160 2230 (8207 ) 1100 570 30 300 2540 3290 (7607 173 1290 1500 2700 (800%) 45 775 420 1310 2580 3400 ( 7 7 0 4 j 230 2900 (800%) 1780 840 BO 2620 010 2940 ( 7 3 0 h ) 1520 280 3000 (780%) 1780 75 890 610 2575 3200 (750%) 310 1310 3100 ( 8 l O 7 ) 1950 960 90 GOO 2620 350 1375 3290 3160 ( 8 0 0 4 ) (750%) 1890 940 120 a Rubber 100 zinc oxide 6, sulfur 3.5 Captax 0.5 stearic acid 0.5. b Rubber 100: zinc oxide 6, sulfur 3.5: Captax 0 5 : stearic acid 0.5, NaH2POcFI?O 0.37, IGHl‘04 0.70
--
- -
TABLE VI. TEKSILE STRENQTHS OF
cIIIIIi?iEL
Slum-Coagulated Sheet c -
At break 300 (830%) 600 (810%) 1090 840%) 975 1800%) 1420 (850%) 1600 (860%)
BLACK STOCKS
GOO% 740 850 980 1020 1030 1010
--
0.5 Captax/100 Rubber 1.0 Captax/100 RubberPress Cure at -.----3AC-l43=-3AD-143b.--3AE-1430-260°F., Min. 400% 500% At break 400% 000% At break 400% 500% A t break 400% 1710 1320 1900 2700(625%) 1030 1590 2680(6707) 2300(71097,) 890 1275 30 2300 3375(630 ) 2220 3275 (640%) 1930 1490 2660 (640%) 1625 1900 1220 45 60 1610 2260 3010(600%) 1840 2630 3575 625%) 1760 2530 3700(650%) 2200 2050 2875 3900 (610%) 2420 90 1990 2760 3475 (590%) 2090 2920 3450 [660%) a Rubber (first latex crepe) 100, zinc oxide 6, sulfur 3.5, Captax 3.5, stearic acid 2 . 5 , channel black 50 b Same as 3AC-143 plus 0.37 NaHzPOI.Hz0 and 0.70 KzHP04. c Rubber (first latex crepe) 100, zinc oxide 6, sulfur 3.5, Captax 1.0, stearic acid 2.5, channel black 50. d Same as 3AE-143 plus 0.37 h-aHzP01.Hz and 0.70 KzHPO4. 7
--
additional 0.5 gram of Captax per 100 grams of rubber (3AE-143). Discussion The hypothesis is suggested that a factor, perhaps a major one, in determining rate of cilre of rubber is the acid-base equilibrium developed during cure; it may be said for brevity (somemay think inaccurately) that the pH of the system and its buffer capacity enable it t o maintain a given reaction and thus condition the rate of cure of a given rubber. This assumed equilibrium could result in rubber-sulfur stocks, for example, from the components (a) nonrubber constituents naturally present in the crude rubber mass and (h) acidic materials, such as hydrogen sulfide (4), generated during vulcanization. The extent to which hydrogen sulfide would influence the equilibrium would depend on its rate of formation and rate of removal-that is, its concentration a t any given time. Even with the same rubber this concentration (and resulting acid-base equilibrium) would be expected to be different in the case of rubber-sulfur stocks (assumed high concentration of hydrogen sulfide) as compared with accelerated stocks (assumed low concentration of hydrogen sulfide). Moreover, it might be expected to vary with the type of accelerator used. Addition of various reinforcing agents or fillers, particularly in large volume, could be expected to affect the acid-base equilibrium; but any such effect, on the whole due to addition of standard pigments, would probably be of less magnitude if the batch were compounded with
2AX-174b700% A t break 2430 ( 7 9 0 7 ) 1460 2650 ( 7 8 0 d ) 1740 2900 7 7 5 7 ) 1970 1990 2990 1775%) 2040 2940 (775%) 1040 3040 (780%)
3AF-143d500% At break 2440 3920(660’7) 4000 (660%) 2770 3090 4340(640%) 3375 4100 (575%)
a whole rubber, such as L. S. made from ammoniated latex, than if it were compounded with well-washed (but unfermented) rubbers such as first latex crepe.
Literature Cited (1) Bruce, A., Trop. Aar. (Ceylon), 59, 267 (1922). (2) Bruni and Levi, in Memmler’s “Science of Rubber”, p. 308 (1934). (3) Eaton, Grantham, and Day, Agr. Bull. Federated .?laZaz/ States, 27 (1918). (4) Fisher, H. L., IPTD. ENG.CHEM., 31, 1381 (1939). (5) Glidden, A. A , , personal communication; Weber, L. E., “Chemistry of Rubber Manufacture”, p. 241 (1926). (6) Hastings and Rhodes, IND. ENG.CHEM.,31, 1455 (1939). (7) Morton, H. A., U. S. Patent 1,893,868 ( J a n . 10, 1933). (8) Oenslager, G., IND.ENG.CHEM.,25, 299 (1933). (9) Sackett, G. A., I b i d . , 26, 535 (1934). (10) Vries, 0. de, Arch. RubbeTcuZtuur, 2, 67 (1918); “Estate Rubber”, p. 101 (1920). (11) Vnes, 0. de, “Estate Rubber”, 1920. (12) Ibid., pp. 463 et seq. (13) Weber, C. O., “Chemistry of India Rubber”, 1902; quoted in Whitby’s “Plantation Rubber”, p. 306 (1920). (14) Weber, C. O., Gummi-Ztg., 17, 898 (1903); quoted in Bedford and Winklemann’s, “Systematic Survey of Rubber Chemistry”, p. 217 (1923). (15) Whitby, G. S., “Plantation Rubber and the Testing of Rubber”, pp.219etsep. (1920). ENG.CHEM.,29, 953 (1937). (16) Wiegand, W. B., IND. PRESENTED before t h e Division of Rubber Chemistry a t the 102nd Meeting of the AMIERICINCEEMICAL SOCIETY, Atlantic City, N. J.
EVALUATION OF THE BUFFER CAPACITY OF CRUDERUBBERS E. B. NEWTON’ AND E. A. WILLSOE’ Malayan Research Laboratories, T h e B. F. Goodrich Company, Kuala Lumpur, Federated Malay States
T
HE hypothesis was advanced in the preceding paper that an important factor in determining rate of cure of rubber is the acid-base equilibrium developed during vulcanization, such an equilibrium being determined to a large extent by the nat’ure and quantity of latex serum material remaining in the crude rubber mass. In effect, this serum material-was regarded as a buffer substance. 1
Present address. T h e B. F. Goodrich Company, Akron, Ohio.
If this were true, it follows that differences in buffer capacities of various crude rubbers should be demonstrable, and in the following discussion two sets of experiments designed with this object in view are described. Experimental In the first series two rubbers were chosen-first latex crepe and a commercial latex-sprayed (made by spray drying am-
