Polysulfide Theory of Accelerator Action during Vulcanization EVIDENCE FOR THE THEORY ROSS E. MORRIS Rubber Laboratory, Navy Yard, Mare Island, Calif. benzene a t 30' C., and evaporated off the solvent a t temperatures ranging from 35' to 50' C. The residue apparently consisted of white zinc diethyl dithiocarbamate crystals and yellow sulfur crystals with no evidence of a polysulfide. The identity of the crystals was confirmed by carrying out a mixed melting point determination with pure zinc diethyl dithiocarbamate in the case of the white crystals, and analyzing for zinc in the case of the yellow crystals. The white crystals did not lower the melting point of zinc diethyl dithiocarbamate, and the yellow crystals did not contain zinc. A similar experiment was performed with sulfur and zinc dibutyl dithiocarbamate with the same results. It is evident that the zinc dithiocarbamates do not react with sulfur to form stable polysulfides. The results of the above experiments do not necessarily preclude the possibility that unstable polysulfides of the zinc dithiocarbamates exist. Langenbeck and Rhiem (3) were unable to isolate a polysulfide from a solution of sulfur and dibenzothiazyl disulfide in carbon disulfide, but they proved the existence of several unstable polysulfides of dibenzothiazyl disulfide by means of a fusion diagram. An attempt was made by the author to apply this method to the present problem. Various proportions of sulfur and zinc dibutyl dithiocarbamate were fused together, the fusion mixtures were allowed to solidify, and the behavior of the mixtures in a melting point apparatus was studied. Unfortunately it was found that melted sulfur and melted zinc dibutyl dithiocarbamate are not miscible in all proportions. Such behavior complicates the analysis of a fusion diagram, so further work in this direction was abandoned. The author then turned to the classical method for the determination of decomposition constants of complex addition compounds discovered by Behrend (1). This method depends on the measurement of solubility alterations in suitable heterogeneous systems. For example, the apparent solubility of compound A in a solvent is increased when the reactive compound B is dissolved in the saturated solution because some of compound A reacts with compound B to form the soluble addition compound AB. Although the logical system to study in this case would be rubber-sulfurzinc dithiocarbamate, such a system would reach equilibrium only very slowly because of the high viscosity of rubber; and accurate solubility measurements would be impossible since rubber undergoes a chemical reaction with sulfur in the presence of the ultra-accelerator. It was decided t o employ benzene and cyclohexane as the solvents for the following reasons: They are hydrocarbons as is rubber, they are liquids which may be readily purified, the molecular condition of dissolved substances may be accurately determined by the cryoscopic method, and information concerning the existence of addition compounds in these solvents may be later employed for studying the vulcanization of cements.
In an attempt to confirm the well-known polysulfide theory of accelerator action during vulcanization, the author has demonstrated the existence of unstable addition compounds between sulfur and the ultra-accelerators, zinc diethyl dithiocarbamate and zinc dibutyl dithiocarbamate, by means of solubility measurements in benzene and cyclohexane at 30' C. These compounds, which exist only in solution, consist of one molecule of the zinc dithiocarbamate combined with one molecule of SI. Equilibrium constants? calculated from the solubility data, show that zinc dibutyl dithiocarbamate forms a more stable addition compound than zinc diethyl dithiocarbamate. It has been established that the stability of the addition compounds is no measure of the accelerating activity of the respective zinc dithiocarbamates.
0
N T H E basis of indirect experimental evidence, Scott and Bedford (6) some years ago proposed a mechanism for the accelerator action of metallic dithiocarbamates known as the polysulfide theory. According to these authors a metallic dithiocarbamate reacts with sulfur during vulcanization to form a polysulfide which subsequently decomposes and liberates active sulfur for reaction with the rubber hydrocarbon. This theory gained support with the observation of Jones and Depew (2) that the solubility of zinc dimethyl dithiocarbamate in rubber is increased when sulfur is added to the mix since such behavior may result from the formation of a soluble polysulfide. More recently, Scott and Sebrell (6) questioned Scott and Bedford's polysulfide theory on the basis that no polysulfides of the metallic dithiocarbamates have been isolated. The author does not intend to enter into the controversial subject of the mode of action of vulcanization accelerators, but rather to present experimental data which confirms the original supposition of Scott and Bedford that sulfur forms unstable addition compounds, or polysulfides, with metallic dithiocarbamates. The metallic dithiocarbamates investigated were zinc diethyl dithiocarbamate and zinc dibutyl dithiocarbamate.
Preliminary Work Inasmuch as the zinc dithiocarbamate accelerators bring about vulcanization of rubber or of rubber cements a t room temperature, polysulfides of zinc dithiocarbamates should be formed a t this temperature if the polysulfide theory is correct. The author dissolved sulfur and zinc diethyl dithiocarbamate, in approximately equimolecular proportion, in 503
Vol. 34, No. 4
INDUSTRIAL AND ENGINEERING CHEMISTRY
504 f? 4.4
ii
#e &a
pq $6
E
q
4.0
2 VLL 0
so LOG 3.8
#'
0.5 SULFUR I GRAMS
I.o
IN
1.5
2.0
,
ZINC DIETHYL DITHIOCARBAMATE CRAMS ,IN 100 GRAMS BENZENE
100 GRAMS BENZENE
FIGURE 1. SOLUBILITY OF ZINC DIETHYL
FIGURE 2. SOLUBILITY OF SULFCR IS SOLUTIONS OF ZINC
DIETHYL DITHIOCARBAMATE! IN BENZENE
DITHIOCARBAMATE IN SOLUTIONS OF SULFUR IN BENZENE
Measurement of Solubility Alterations PREPARATION OF MATERIALS.Baker's c. P. benzene was fractionated and the fraction having a freezing of 5.5" C., or _ point _ above, was used. Eastman's P702 cyclohexane was repeatedly fractionated and the fraction having a freezing point of 6.0" C., or above, was used. Commercial 99.5 per cent sulfur was crystallized once from toluene. The crystals were washed with acetone and dried in air with slight warming. Commercial Ethyl Zimate was used as the source of zinc diethyl dithiocarbamate. The Ethyl Zimate was recrystallized four times from toluene containing a little ethanol. Solution was effected for the first two crystallizations by boiling in the presence of bone charcoal. Material used in the solubility measurements was in the form of white crystals having a melting point of I
177.7" C.
Commercial Butyl Zimate was used as the source of zinc dibutyl dithiocarbamate. The Butyl Zimate was recrystallized three times using the solvents and procedure described for zinc diethyl dithiocarbamate. Material used in the solubility measurements was in the form of white crystals having a melting point of 107.3"C. PROCEDURE. The solubility equilibrium was attained in a glass-stoppered, 250-ml. flask immersed in a water bath maintained at 30' * 0.03" C. Amroximatelv 55 ml. of solvent were used. The solution was agizited by a n electromagnetic stirrer in the flask. At least 48 hours were allowed for reaching saturation. The saturated solution was transferred by means of air pressure to a 50-ml., wide-mouth flask with cork stopper which had been previously weighed with and without the cork stopper. The liquid was filtered through two layers of fine linen during transfer. The flask, closed with the cork stopper, was a ain weighed. The cork stopper was removed and the flask driei to constant weight over a 10-watt electric lamp. The temperature of the liquid during evaporation ranged from 35" to 50" C., depending upon room temperature. It had been previously found that no oxidation of zinc dithiocarbamates or sublimation of sulfur takes place after several weeks of heating in this temperature range. The solubilities of the pure substances in benzene and in cyclohexane were determined before proceeding with the investigation of three-component systems. The solubility data are given in Table I. The solubility of each of the three substances in benzene is greater than it is in cyclohexane. This difference is particularly large in the case of zinc diethyl dithiocarbamate. TABLEI.
80LUBILITY O F
PURE SUBSTANCES AT 30' Grams
Sulfur Zinc diethyl dithiocarbamate Zinc dibutyl dithiocarbamate
Sulfur Zino diethyl dithiocarbamate Zina dibutyl dithiocarbamate
2.517 3 , a72 12.80 1.409 0.103 8.992
convenience and accuracy it was desirable that the solubility of the solid phase should not be too low. Accordingly, the systems in Table I1 were selected for investigation. No solubility determinations were made with zinc dibutyl dithiocarbamate as the solute at saturation because its solubility was considered too high. The results of the solubility determinations are shown graphically in Figures 1t o 4. TABLE11. SYSTEMS SELECTED FOR INVESTIGATION Solvent Solute at Variable Concn. Solute at Satn. Zinc diethyl dithioBenzene Sulfur carbamate Benzene Zinc diethyl dithiocarbamate Sulfur Benzene Zinc dibutyl dithiocarbamate Sulfur Cyclohexane Zinc dibutyl dithiocarbamate Sulfur
Analysis of Data In each of the systems investigated the solubility of the solid phase increases in direct proportion to the concentration of the third component. The fact that the relation is of the h s t order in the system sulfur-zinc diethyl dithiocarbamate benzene, when either of the solutes is at saturation, is particularly significant since it proves that one molecule of sulfur combines with one molecule of zinc diethyl dithiocarbamate to form one molecule of the addition compound. The veracity of the last statement is readily demonstrated. Consider the bimolecular reaction, A + B e A B
to reach equilibrium in a solvent. The equilibrium constant
Let sufficient solid compound A be added so that the solution will remain saturated with respect to A . Then the concentration of A in the solution will remain constant in spite of reaction of A with B, and the equilibrium constant can be written:
c.
100 Grams Solvent Benzene Benzene Benzene Cyclohexane Cyclohexane Cyclohexane
I n order that the thermodynamic activities of the solutes should not differ to any extent from their respective concentrations and thus complicate the analysis of the data, it was necessary that a system be selected in which the solubility of the solid phase mould not be excessive. At the same time, for
K1 or
(AB) (B)
( B ) K I = (AB)
Since, in the experimental method used, the amount of A B in solution is not directly determined as such but is calculated from the apparent increase in solubility of A , the latter value is directly proportional to the concentration of B. I n other words, when the increase in solubility of A per 100 grams of solvent is plotted against the concentration of B per 100 grams of solvent, a straight line will result. The same reasoning is applicable to the case where the solution is saturated with respect to B and the concentration of A is variable.
,
INDUSTRIAL AND ENGINEERING CHEMISTRY
April, 1942
ZINC D l e L l M DITHIOCARBAMATE ,GRAMS ,IN 100 GRAMS BENZENE
FIGURE 3. SOLUBILITY OF SULFUR IN SOLUTIONS OF ZINC DIBUTYL DITHIOCARBAMATE IN BP~NZENE
ZINC DIBUTYL DlMlOCARBAMATE ,GRAMS,
IN 100 GRAMS CYCLOHEXANE
FIGURE 4. SOLUBILITY OF SULFUR IN SOLUTIONS OF ZINC DIBUTYL DITHIOCARBAMATE IN CYCLOEEXANB
On the other hand, if the reaction takes place as follows: 2A B AZB or A 2 B e ABa
carbamate is assumed to be bimolecular for the purpose of this calculation although, as pointed out above, the order of this reaction is not definitely known.
or in any other manner where more than one molecule of A or B is involved, the reasoning given above will show that at least one of the curves, obtained by plotting the apparent increase in solubility of the substance at saturation against the variable concentration of the other substance, is not a straight line. The linear relations between the concentration of zinc dibutyl dithiocarbamate and the amount of sulfur required to saturate the solution in benzene or cyclohexane (Figures 3 and 4) prove that one molecule of zinc dibutyl dithiocarbamate reacts with sulfur to produce one molecule of the addition compound. No information can be gained from these curves as to the number of molecules of sulfur which enter into the reaction. It seems reasonable, however, that this reaction should proceed in the same manner as the reaction between einc diethyl dithiocarbamate and sulfur.
TABLE IV. EQUILIBRIUM CONSTANTS AT 30" C.
+ +
E Bonstant uilibrium Reacting Substances Sa Zn diethyl dithiocarbamate Ss 4-Zn diethyl dithiocarbamate Ss Zn dibutyl dithiocarbamate 58 Zn dibutyl dithiocarbamate
+
+ +
K
Solute a t Satn. Zn diethyl dithiocarbamate Sulfur Sulfur
Solvent Benzene
1.41
Benzene Benzene
3.91
Sulfur
ctexane clo-
1.45 2.57
TABLE V. RUBBER CEMENTS A Benzene Pale ore e Zinc o d e Sulfur Zinc diethyl dithiocarbamate Zinc dibutyl dithiocarbamate
100.0 5.0 0.5 1.0 1.5
.....
B 100.0 5.0
0.6 1.0 ..... 1.965
Calculation of Equilibrium Constants The composition of the addition compounds may be ascertained by determining the molecular weights of sulfur and the zinc dithiocarbamates in the solvents used. These values were determined in benzene by the cryoscopic method. The results (Table 111) prove that sulfur in solution is in the form of Sa molecules and that the zinc dithiocarbamates exist as unassociated molecules. It is concluded, therefore, that the reaction takes place as follows: SS Zn(SCSNR& e Zn(SCSNR&.Sa
The values calculated for the equilibrium constant of the sulfur-zinc diethyl dithiocarbamate reaction in benzene from different solubility data are in good agreement (Table IV). Also, the values calculated for the equilibrium constant of the sulfur-zinc dibutyl dithiocarbamate reaction in benzene and cyclohexane, respectively, are different, the magnitude of the values indicating that the sulfur-zinc dibutyl dithiocarbamate addition compound is more stable in benzene than it is in cyclohexane.
where R is a hydrocarbon radical. The equilibrium constant for this reaction may be written [Zn(SCSNR&. S8] K = [Sel [Zn@CSNR&l
Relative Activity of Accelerators It should be emphasized that the equilibrium data reported here cannot be used as a measure of the relative activity of these compounds as vulcanization accelerators because, in
+
general, there is no connection between the equilibrium constant of a reaction and the rate of the reaction. Although the magnitude of the equilibrium constants given in Table I V indicates that sulfur forms a more stable addition compound with zinc dibutyl dithiocarbamate than it does with zinc diTABLE111. MOLECULAR WEIQHTSIN BENZENE BY CRYOSCOPIC ethyl dithiocarbamate, it does not follow that the former comMETHOD pound is the more active accelerator; in fact, as shown below, the contrary is true. --Mol. Wei htSubstance Found heoretical The relative activities of zinc diethyl and zinc dibutyl di250 Sulfur(& 255 thiocarbamates as vulcanization accelerators were found by 302 Zina diethyl dithiocarbamate 306 474 Zinc dibutyl dithiocarbamate 489 observing the time required at 30" C. for the cements given in Table V to gel. The accelerators in these cements were in equimolecular proportion and within their respective soluUsing the slope of the line in the graph and the solubility bility limits. Cement A gelled in 2 days, whereas cement B of the solid phase from Table I, it is possible to calculate the required 3 days to reach the same consistency. Evidently equilibrium constant, K , for each of the systems investigated. zinc diethyl dithiocarbamate is the more active accelerator, The values of K for the various reactions are given in Table notwithstanding the relatively greater instability of its sulIV. The reaction between sulfur and zinc dibutyl dithiofur addition compound. where each term in the fraction represents the concentration of that ingredient in moles per 1000 grams of solvent.
506
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
Other Evidence for Polysulfide Theory In conclusion, the author acknowledges that, in verifying the existence of an addition compound between sulfur and a zinc dithiocarbamate, he has not fully confirmed the accelerator mechanism proposed by Scott and Bedford (6) because he has presented no evidence to prove that the addition compound is actually involved in the vulcanization reaction. However, support for the argument that the addition compound is involved in the vulcanization reaction may be found in the discussion by Lewis, Squires, and Nutting (4). These authors pointed out that the peculiar shape of the curves for combined sulfur us. time of cure may be readily explained if it is assumed that the accelerator and sulfur react to form a sulfur complex which, in turn, reacts with the double bonds of the rubber to form rubber sulfide and thus regenerates the accelerator.
Vol. 34, No. 4
Literature Cited (1) Behrend, Z. physik. Chem., 9, 405 (1892). (2) Jones, H. C., and Depew, H. A., IND.ENG.CHEM., 23, 1467 (1931). (3) Langenbeck, W., and Rhiem, H. C., Kautschuk, 12, 156 (1936); Rubber Chem. Tech., 10, 158 (1937). ENQ.CHEM., (4) Lewis, W. K., Squires, L., a n d Nutting, R. D., IND. 29, 1135 (1937); Rubber Chem. Tech., 11, 107 (1938). (5) Scott, W., a n d Bedford, C. W., J. IND. ENQ.CHEM., 13,125 (1921). (6) Scott, W., and Sebrell, L. B., i n Davis and Blake's "Chemistry a n d Technology of Rubber", A. C. S. Monograph 74, New York, Reinhold Pub. Corp., 1937. PRESENTED before the Division of Rubber Chemistry at the 102nd Meeting of the AMERICAN CHEMICAL SOCIETY, Atlantic City, N. J., under the t i t l e "Unstable Addition Compounds of Sulfur with Zinc Dithiooarbamates". The opinions or assertions in this paper are those of the author and are not t o be construed as officialor reflecting the views of the Navy Department or the naval service at large.
Sorption of Water Vapor by Vermiculite and Its Silica EQUILIBRIUM MEASUREMENTS ~-
By the use of a McBain-Bakr sorption balance, isotherms were obtained at 25", 45", and 65" C. for the sorption of water vapor by vermiculite, exfoliated vermiculite, and the silicas prepared from these materials. Definite hysteresis was observed in the isotherms for the silicas but was absent in those of unexfoliated and exfoliated vermiculite. The hysteresis effects as well as differences in the shapes of the isotherms for the various materials can be explained on the basis of the cavity concept proposed by McBain. The data for the unexfoliated vermiculite fits the Langmuir equation. Using the Langmuir equation, inversion points are found in isotherms for the exfoliated vermiculite and the two silicas which indicate that in the case of these materials capillary condensation takes place at the higher pressures.
ERMICULITE (3) is a nonmetallic mineral classed as one of the micas whose molecular formulta is 22Mg0.v5AlnOI.Fe20a.22Si02.40H20 (4). The most remarkable property of vermiculite is its great expansion when heated. A more complete description of this mineral is contained in an earlier paper in this series (6). According to Guthrie and Wilbor (6) a silica of excellent sorbing properties can be prepared from the mineral. The silica was prepared from both the exfoliated and the unexfoliated vermiculite by treatment for about 3 hours with hot dilute sulfuric acid (1:3). The silica was filtered in a Biichner funnel, washed, and dried a t 105" C. (6). The silica and the exfoliated vermiculite (which had been exfoliated for several months) were activated by heating to 250' C. for one hour, the optimum condition for activation
L. A. SPITZE AND L. A. HANSEN Rensselaer Polytechnic Institute, Troy, N. Y.
(6). The crude (unexfoliated) vermiculite was dried a t 110' C. for 2 hours. The apparatus used for these measurements consisted of one of the varied f o r m of the McBain-Bakr sorption balance made of Pyrex glass (11). The quartz-fiber spirals, which were prepared similarly t o those described by McBain and Bakr ( l l ) ,were calibrated prior to use in the experiments. The sensitivity of an average spiral was 0.01824 gram per cm. To this spiral was attached a glass bucket ( I ) , weighing about 0.07 gram, in which was placed a sample weighing about 0.14 gram. The elongation of the spiral was measured by a Gaertner cathetometer accurate t o 0.001 om., in conjunction with a reference scale made of Invar tape. The system was evacuated by means of a Cenco mercury diffusion pump packed by a Hyvac oil pump. The water used in the sorption was freshly distilled from a quartz still and then placed in Pyrex chambers of about 200 ml. capacity. The ends of these chambers were constricted, and to these constrictions were sealed No-Lub stopcocks, one of which was then sealed to a tube leading to the bottom of the sorbate tubes. By application of a vacuum to the upper stopcock, air was flushed out of the chambers. Most of the water in these chambers was allowed to flow through the bottom stopcock into the sorbate tubes. These water chambers were then sealed off from the rest of the system. The vapor pressure of the water was controlled by immersing the sorbate tubes in an oil bath, the temperature of which was controlled to *0.05" C. Equilibrium vapor pressure of the water, a t the prevailing temperature, was assured by means of an entirely enclosed solenoid glass stirrer (16). The sorbent tubes, connected by large-bore Pyrex tubing to the sorbate tubes, were placed in an air bath whose temperature was controlled to =tO.l' C. The entire apparatus was isolated by sealing off the system from the vacuum equipment; thus a system was formed capable of determining the sorption process of six different samples. The water in the sorbate tubes was kept a t
