I
W. F. FISCHER, R. F. NEU, and R. L. ZAPP Enjay Laboratories, Linden, N. J.
Improved Accelerators for Butyl Rubber Higher
Alkyl or Aryl Salts of Dithiocarbamic Acid Accelerator modifications increase surface cure of butyl vulcanized in open steam
THE
ozone attack of butyl rubber has been studied by Buckley (2). Using stress relaxation techniques, he devised a method of characterizing ozone resistance by a cracking rate constant. For any level of molecular unsaturation of the butyl polymer used, lowest cracking rates occurred a t the highest level of cross linking. I n other words, when the double bond site was “shielded” or “immobilized’’ by a cross-linked bridge, resistance to ozone attack was enhanced. During practical compounding studies it was often noted that strips and arricles of butyl vulcanized in open steam did not have the ozone resistance of similar press-vulcanized samples. O n a quantitative basis, a sample vulcanized in open steam could have a n ozone cracking rate 40 times that of a similar press-cured sample by the Buckley stress relaxation technique. Because ozone resistance is influenced by cure state and steam-cured surfaces show decreased resistance, it would follow that steamcured articles have low surface states of cross linking. O n e explanation (not necessarily the only one) could be based on the hydrolysis of sulfur accelerators, thereby drastically reducing the rate of vulcanization at the surface. Retardation of vulcanization bv steam pertains only to vulcanization with sulfur? as dioxinie cures are not affected by exposure to moisture to the same extent. Structural differences between the most commonly used accelerators for butyl are illustrated :
Class
(Typical) Structure
Example
H O - - N =~ = N - O H
Quinoid
p - Quinonedioxime
Thiuram
Tetramethylthiuram disulfide
CH3
\
/
S
S ti
N- C- S-S-C-
II
CH a
N
\
C Ha
CH3 Dithiocarbamate
/
Tellurium diethyl dithiocarbamate
There appear to be no literature references suggesting that hydrolysis of accelerators can occur upon steam curing, although the molecules themselves would be prone to attack. Therefore, the effect of water vapor on the cure state a t the rubber surface was studied. Thin Film Techniques Work with thicker rubber test specimens suggested that increased ozone attack of steam-cured samples \vas a surface phenomenon. because no difference in the state of cure between it and a press-cured sample could be determined by stress-strain analysis. Therefore, to determine the effect of steam curing on the vulcanization characteristics of butyl surfaces, experiments with samples which were preponderantly surface-namely, thin films-were devised. T o study the effects of various accelerators, a simple formulation was used.
Parts Enjay Butyl 215 Zinc oxide Sulfur Accelerators
100 5 2 equal molar concentrations
1-3,
T o prepare thin films fully conipounded unvulcanized stocks were pressed between glass plates a t approximately 200’ F. Attempts were made to maintain the thickness of the film between 0.004 and 0.006 inch. .After the thin films Tvere squeezed out, tivo types of samples were prepared. First, a thin film was kept benveen the t\vo glass plates and the edges were sealed with polyethylene tape. I n another set of samples the top plate was removed, which allowed a surface to be exposed to steam. I n addition, press cures of thin films were frequently compared with the covered and uncovered steam-cured films. T h e samples to be cured in steam were then placed in a n autoclave and cured 45 minutes VOL. 51, NO. 2
FEBRUARY 1959
205
-
CURE STATE-%
VOLUME SWELL
( V. 6000 -POOR)
CURE CONDITIONS
200
PRESS STEAM (PROTECTED) STEAM (EXPOSED)
DPG
STEAM
IO00
=
STEAM (PROTECTED) S T E A M (EXPOSED)
TDEDC
COMBINED SULFUR
rn
PRESS TMTDS
500
o/o
0.71 0.22
STEAM Figure 1. Press and protected steam-cured vulcanizates show better cure state than steam-exposed vulcanizate
a t 320’ F. in open steam. After vulcanization in this fashion, samples were removed from the autoclave and the extent of cure was determined by volume swell techniques (3, 5 ) . Swelling capacities were determined in cyclohexane a t 25’ C. For thin films, equilibrium swelling values could be obtained in relatively short periods of time not exceeding ‘/z hour. T h e higher the volume swell, the lower the state of cure. Retardation of Vulcanization Some evidence of accelerator hydrolysis is shown in Figure 1, which compares press- and steam-cured butyl vulcanizates cured with tetramethylthiuram disulfide (TMTDS), tellurium diethyl dithiocarbamate (TDEDC), and diphenylguanidine (DPG). Without exception, low volume swell measurements indicating good cure state are noted for the press and protected steam-cured pad as
compared to the steam-exposed sample. This behavior, observed with all grades of butyl, is independent of the accelerator type, which suggests all three curatives are “water-sensitive.” Further support for cure retardation by possible hydrolytic deactivation of accelerators is found in the measurement of combined sulfur in the protected and exposed samples of the vulcanizates. Low per cent combined sulfur correlates well with the high volume swell measurement. Both indicate lower cross linking for the steam-exposed sample, and a high state of cross linking for the protected sample. Combined sulfur was determined by the method of Rehner and Holowchak ( 4 ) . A4n explanation of this wide disparity in cure state under the same temperature conditions of vulcanization is that steam hydrolysis effectively deactivates sufficient accelerator before the second stage of vulcanization, represented by sulfur cross linking, is
CURE STATE-% VOLUME SWELL ALKYL RADICAL
PRESS
STEAM (PROTECTED)
2 0 0 400 800
200 400 600
STEAM (EXPOSED)
200
600
1000
DIMETHYL DlETHY L DIPROPYL
METHYL BUTYL ETHYLBUTYL
rn
H
METHYL OCTYL
m m m
Figure 2. Increase in molecular weight of alkyl radical of TDEDC has little effect on press or protected steam cures
206
INDUSTRIAL AND ENGINEERING CHEMISTRY
reached. Bresler ( 7 ) suggested a comparable mechanism for cure retardation. Retardation of vulcanization apparently is the main problem as opposed to reversion, as in the latter case a high volume swell would also show high combined sulfur content (6). This was not observed. Furthermore, modified accelerators under the same “problem steam conditions” produce vulcanizates of good quality. Similar experiments have been run with SBR and natural rubber. In either case, steam-curing an exposed thin film decreases the state of cure? compared to a press or protected film (Table I). However, the decrease due to exposure to steam is not as great as in the case of butyl. Such a reduction in the state of cure would not affect the ozone resistance of diene polymers to any great extent because of the high residual exposed unsaturation always present after crosslinking these types of rubber.
Table I. Steam Vulcanization of Thin Film of SBR Decreases State of Cure Parts
SBR ZnO
Stearic acid Sulfur BTMHS
PBN
100 5
45 Min. in 80-Lb.Steam Thin Film Volume Swell Exposed Protected
1 1.7 1.25 1
325 326 326
251 253 253
Increasing Molecular Weight of Alkyl Radical I t was postulated that increasing the solubility of the accelerator in the butyl hydrocarbon would protect the accelerator from hydrolysis. A series of modified tellurium salts of dithiocarbamic acid was studied, embracing increasing molecular weight of the alkyl radical. The alkyl derivatives tested are shown in Figure 2. Each accelerator was added to the base compound in an amount equal to the molar equivalent of one part of tellurium diethyl dithiocarbamate. Results of the curing studies as followed by volume swell are shown by the bar graphs adjacent to the accelerator. For the press cures and protected steam cures, little difference between the six accelerators representing increasing molecular weights can be determined. I t is when these various compounds embodying higher alkyl derivatives of the dithiocarbamic acid accelerator are cured while exposed to steam that the effect of accelerator molecular weight is displayed. The lowest molecular homolog (dimethyl) is by far the poorest accelerator for the steam curing of butyl. As molecular weight is increased, the volume swell after curing in open steam decreases markedly. Further improvements can be realized a t very high molecular weight substitu-
B U T Y L RUBBER ACCELERATORS tions such as dimethyl octyl. Here, the degree of cross linking as measured by a decrease in volume swell begins to approach the values for the protected steam cure and the press cure.
Further modification of the basic tellurium diethyl dithiocarbamate is obtained when one of the substituent groups is a cyclic or aryl-type substitution. I n Figure 3, the control compound (the diethyl derivative) is shown at the top of the graph. Again, states of cure as measured by volume swell are displayed by a series of bar graphs adjacent to each accelerator. I n all cases, the compounds of higher molecular weight produce better vulcanized net%forkswhen exposed to open steam than the control. This is shown in the series of far right-hand bar graphs. O n the basis of over-all accelerator efficiency, as measured by press and protected steam cures in conjunction with excellent behavior during exposed steam-cure vulcanization, the ethyl benzyl derivative was considered to be one of the most promising. Bulkier substituents like butylcyclohexyl and dit)-clohexyl apparently improve accelerator resistance to steam hydrolysis but are less active under d q heat cure conditions.
Table 11. Steam Curing Butyl-Carbon Black Compounds with Tellurium Ethyl Benzyl Dithiocarbamate Improves Cure 1
2
100
100
5 50
5 50
2
1
...
2
...
1.5
70voluiiie swell Protected Exposed
RADICAL
PRESS
STEAM (PROTECTED)
2 0 y 4?0 700
205) 4?0 700
DIETHYL ETHYLBENZYL
Alkyl Aryl Derivatives
Enjay Butyl 215 Zinc oxide Philblack A Sulfur TDEDC TEBDC
CURE STATE-% VOLUME SWELL
238
253
349
264
Figure 4.
ETHYL CYCLOHEXYL
STEAM (EXPOSED)
2?0
,
6?0
,
l0,OO
:
METHYL CYCLOHEXYL BUTYL CYCLOHEXYL DICYCLOHEXYL Figure 3. Substitution of cyclic groups or aryl radicals on TDEDC improves vulcanization
For simplification, tellurium diethyl dithiocarbamate has been designated TDEDC and the ethyl benzyl derivative as TEBDC. Evaluation of TDEDC and TEBDC in a vulcanizate containing 50 parts of carbon black also shows the same advantageous response of the modified curative under both dry and wet cure conditions as compared to the conventional accelerator (Table 11). Accelerator Modification and Ozone Resistance Higher molecular weight modification of the dithiocarbamate accelerator improves the state of surface cross linking during vulcanization. Does this improvement impart a more ozone-resistant surface to steam-vulcanized butyl compounds? TEBDC, for example, affords a significant improvement in ozone resistance of a commercially compounded weatherstrip when compared to the best organosulfur accelerator currently avail-
able (Figure 4 ) . Dumbbells from three kulcanizates containing TEBDC, regular TDEDC, and a combination of T M T D S and mercaptobenzothiazole accelerators are compared after 80 hours' exposure to 50 p.p.h.m. ozone at 100' F. and 50% extension. Magnification of 70X raveals severe surface cracks on the T M T D S sample and. to a lesser extent, on the surface of regular TDEDC. At this point, TEBDC shobvs only a grainy inner surface with some fine cracks starting along the edge. Confirmation of these observations is presented in Figure 5, which indicates that TEBDC, alone or in combination Fvith other accelerators, yields considerable improvement in the ozone resistance of both hard and soft extrusion compounds over that found for compounds emploving conventional but) 1 accelerators. Although the differences noted were. in terms of numbers of cracks, sometimes small, they were sufficient to mean the difference between passing and failing the standard test for automotive Treatherstrips.
TEBDC shows greater ozone resistance than TMTDS or TDEDC VOL. 51, NO. 2
FEBRUARY 1959
207
Figure 5.
TEBDC improves ozone resistance of hard and soft extrusion compounds 80 Shore butyl weather strips, 70 hours at 50 p.p.h.m. 03, 100' F.
Ozone resistance is a prime prerequisite for wire insulation. where many rubber wire covers are vulcanized in the presence of steam. I n a typical low voltage !Tire insulation formula the TEBDC derivative shows a marked accelerator improvement over TDEDC in the presence of an antiozonant. The bar graphs of Figure 6 illustrate the degree of improvement. Included in the chart are physical and ozone-resistant properties of press-vulcanized samples. Attention is drawn to the relative lengths of the bar graphs of ozone resistance between a press-cured sample and one exposed to steam. For the regular diethvl derivative, a great discrepancy exists; for the ethyl benzyl derivative. the ozone resistance of the press cure and TDEDC
steam cure are a t a high and equivalent level. Compounds tested had the following general formula : Enjay Butyl 217 Elastopar Stearic acid Zinc oxide Channel carbon SRF carbon Soft clay Spider sulfur Wax Lanair Accelerators as indicated
100 1 1
5 15 25
60 2
5 5
Lanair is a commercial antiozonant; a possible structure is a nickel salt of a n aminophenyl-substituted propionic acid. Thus: the ethyl benzyl derivative, TEBDC, still imparts vast improvement to a butyl steam-cured compound even
TEBDC
7
when the vulcanizate is protected with an antiozonant. Summary These studies have shown that water vapor can reduce the state of cross linking a t the surface of the vulcanizate. By swelling techniques and combined sulfur measurements, a hypothesis based upon accelerator hydrolysis was put forth. T o counteract the effect of moisture, accelerator molecules were used that embodied higher molecular weight alkyl or aryl substituent groups on a dithiocarbamate nucleus. These accelerator modifications increased the surface cure of butyl Lvhen vulcanized in open steam. I t was. therefore, concluded that the substituent groups would impart greater solubility to the accelerator molecule in a hydrocarbon and thus protect it from moisture. Steam-vulcanized butyl articles. when containing these higher molecular weight accelerators, shoived increased ozone resistance, confirming the improved state of vulcanization a t the surface. literature Cited (1) Bresler, S. E., others, Rubber Chem. and Technol. 19, 946 (1956). (2) Buckley, D. J., Robison, S. B., J. Polymer Sci. 19, 145 (1956). (3) Flory, P. J., Rehner, John, Jr., J. Chem. Phys. 10, 521 (1943). (4) Rehner, John, Jr., Holowchak, Joseph, IND.ENG. CHEM., ANAL. ED. 16, 98 (1944). ( 5 ) Zapp, R. L., Decker, R. H., J. Polymer Sci. 6, 331 (1951). (6) Zapp, R. L., Ford, F. P., Ibid., -9, 97 (1952).
Figure 6. The TEBDC derivative shows a marked accelerator improvement over TDEDC in the presence of an antiozonant Cure, 20 minutes a t 320° F. Modulus Ozone
208
Press.
Steam
MP OP
MS
INDUSTRIAL A N D ENGINEERING CHEMISTRY
OS
RECEIVED for review November 26, 1957 ACCEPTEDSeptember 11, 1958 Division of Rubber Chemistry, 132nd Meeting, ACS, New York, N. Y . ,September 1957.