CHAINTRANSFER OF DISULFIDES A N D PEROXIDES
July 20, 1962
2705
ORGANIC AND BIOLOGICAL CHEMISTRY [CONTRIBUTION FROM
Reactions of Radicals. BY
THE
DEPARTMEXT OF CHEMISTRY, PURDUE UXIVERSITY,LAFAY ETTE, IND. ]
Rates of Chain Transfer of Disulfides and Peroxides with the Polystyryl Radical’) WILLIAN
A.
PRYOR A N D
TIMOTHY L. P I C K E R I N G 3
RECEIVED JANVARY 22, 1902
Previous attempts to determine whether radical displacement reactions occur by frontside or backside attack are discussed, and i t is concluded that they do not provide a general, unarnbiguous answer. Part of the difficulty arises because SHZ reactions6 on carbon atoms are too slow to compete with side reactions. Therefore, we have examined the reactions of radicals with peroxides and disulfides. These reactions are attractive for study since: (1) ,4few examples have been shown to involve s H 2 reactions. ( 2 ) They are very fast. ( 3 ) Disulfides react with nucleophiles in sN2 reactions by backside attack. Therefore, the radical displacement reaction on sulfur may be compared with the S Nreaction ~ on sulfur, and that, in turn, Kith Sx2 reaction on carbon. As an initial approach, the rate of reaction of disulfides with the polystyryl radical was determined. This approach was chosen since the rates can be obtained from chain transfer studies by well developed techniques. It involves the handicap, however, of not allowing proof of the site of attack. The data show that in both the SNZ reaction and in chain transfer, the rate constant smoothly decreases as hindrance around the sulfur atom is increased. However, in the chain transfer reaction, the spread of rate constants is greatly compressed. For example, in the S N re~ action between mercaptide ions and disulfides, the ratio of the rate of reaction of methyl disulfide to t-butyl disulfide is about lo6. I n the chain transter reaction this ratio is 66. Furthermore, a log-log graph of the rate constants for the SNZreaction versus those for chain transfer is not a straight line, The implications of this are discussed. The transfer rate of benzoyl peroxide is 5 times that of benzoyl disulfide; that of di-t-butyl peroxide is 10 times that of t-butyl disulfide.
Introduction Displacement reactions occupy a key position in the theory of chemistry. Ionic nucleophilic reactions. S N reactions, ~ involve backside a t t a ~ k . ~ Recently i t has been found that sE2 reactions frequently involve frontside attacks5 The stereok ; - -\ - H’ chemistry of displacement reactions by radicals, \ ‘ Si12 reactions,6 is unknown, despite elegant and \.I H sophisticated studies. Theoreticians7~* have been able to show that the lowest lying state on the potential energy surface the reaction of chlorine atoms with hydrogen molefor the reaction of hydrogen atoms with hydrogen cules involves a linear transition state, whereas molecules is linear. It involves backside attack, others10-12believe it is triangular. The first experimental study of an S H reaction ~ therefore, eq. 1. is that of Ogg and P01anyi.l~ They found that H * + H-H + [H...H...HJ +H-H + H * (1) optically active sec-butyl iodide racemizes when They conFrontside attack, involving the triangular transi- heated in the gas phase a t 200-280’. tion state in eq. 2A or in eq. 2B, requires more cluded that the mechanism is d-R-I +d,Z-R + I energy. (3) Unfortunately, conclusions from the theoretical I* d-R-I +Z-R-I + I* (4) treatment of more complex reactions are uncertain Equation 4 was proposed to be a backside S H ~ due to the mathematical difficulties involved. displacement. The possibility was not satisfacFor example, some workersg have coiicluded that torily eliminated, however, that a radical attacks (1) Presented in p a r t t o t h e Organic Division of t h e American the alkyl iodide giving a radical which racemizes.14b Chemical Society: Abstracts of Papers, Chicago, Ill., Sept. 3-8, 1961, I * $. d-R-I +d-R* + 11 (5) p. 72-Q. (2) Supported in part b y grant RG-9020 from t h e National Ind-R* f,I-Re (6) stitute of Health, for which we wish t o express our thanks. d,l-R* 1 2 +d,Z-R-I I* (7) (3) Kational Science Foundation Undergraduate Research Participant, Summer, 1961; Undergraduate Research Assistant, 1961Attempts by subsequent workers14 to choose be1909, on grant PU-0623 from the Research Corporation. tween these two possibilities have revealed the diffi(4) A. Streitwieser, Jr., Chem. Revs., 6 6 , 871 (1956). culties involved, and an unambiguous choice has ( 5 ) S. Winstein, T. G. Traylor and C . S . Garner, J. A m . Chem. Soc., 77, 3741 (192.5); H. B. Charman, E. D . Hughes and C . Ingold, J not been made. Herrmann and no ye^,'^^ however,
+
+
Chem. Soc., 2530 (19.59); F. R . Jensen, J . Ana. Chem. Sor., 82, 2469 (1960); D.J Cram and P. Haberfield, ibid.. 83, 2354 (1961). ( 6 ) Eliel has suggested use of t h e terms SH1 and 81x2 for homolytic
~ nucleophilic reactions; reactions, analogously t o s N 1 and S N for E. L. Eliel, in “Steric Effects in Organic Chemistry,” John Wiley and Sons, Inc., New York, N. Y.,1956, p. 149. (7) Conclusions are summarized in S. Classtone, K. J. Laidler and H. Eyring, “Theory of R a t e Processes,” hlcCraw-Hill Book C o , lnc., S e w York. N. Y., 1941, pp. 87-91, 115. (8) (a) H. Eyring and M. Polanyi, Z . physik. Chem., 12B, 279 (1031); (b) I. Shavitt, J . Chem. Phys., 31, 1389 (1959); (c) R. E. Weston, ibid., 31, 892 (1959). (9) K. S. Pitzer, J . Am. Chcm. SOC.,79, 1804 (1957).
+
(lo) Reference 7, pp. 222-228. (11) J. L. Magee, J. Chem. Phys., 8, 677 (1940). (12) For experimental attempts t o resolve this problem, see: J. Bigeleisen and M. Wolfsberg, ibid, 23, 1535 (1955); 1. Bigeleisen, P. S. Klein, R. E. Weston and M. Wolfsberg, ibid., 30, 1340 (1059). (13) R. A. O g g and M. Polanyi, Trans. Faraday SOC,81, 482 (1935). (14) (a) C. C. Price a n d M. Schwarcz, J . A m . Chem. Soc., 62, 289 (1940); D. Clark, H. 0. Pritchard and A. F. Trotman-Dickenson, J. Chcm. Soc., 2633 (1954); (b) R. A. Herrmann and R. M. Noyes, J. A m . Chem. SOL,7 8 , 6764 (1956).
"'706
LvILLIAM
4. P R Y O R
AND ' h l O T H Y
have been able to limit the mechanistic possibilities. Recently Noyes and his students15have reported an elegant determination of the rates of exchange of a series of alkyl iodides with 1 2 1 3 1 . They studied the reactions in the liquid phase, both degassed and in the presence of oxygen. They reason that if oxygen is present, it will compete with iodine for radicals; Le., eq. 7 will compete with eq. 8.
+
R*
0 2
+ROz*
(8)
They find that the rate of exchange is arfected very little by oxygen, and conclude that very little of the exchange occurs by eq. 5, 6 and 7, and most occurs by eq. 9 and 10. I**
I** +21** RI ----f I * R I *
+
(9) (10)
+
They believe eq. 10 is a LValden inversion by iodine atoms. Their data has a disturbing feature, however. They find the following relative rate sequenceL6: methyl 1.0, isopropyl 29, neopentyl 2.4. It is difficult to conceive of attack on carbon in which methyl iodine reacts at the slowest rate.17 Russian workers18 have reported that organometallic compounds containing asymmetric alkyl groups transfer from one metal atom to another with retention, implying frontside attack. There is no surety, however, that radical processes are involved. Wolfgang and his students have studied the reaction of recoil tritium atoms from nuclear reactions with organic compounds. They findIgthat tritium atoms react with optically active 2-butanol to replace the hydrogen atom attached to the asymmetric carbon atom with 91 i 6% retention. The attack is unambiguously from the frontside. The generality of the result is unknown, however, since these "hot" atoms react having 2 to 10 e.v. of kinetic energy.Ig I t is apparent that the stereochemistry of the S H displacement ~ by complex organic radicals is not known. Part of the difficulty appeared to be that radical displacements on carbon do not compete with other possible reactions, e.g., hydrogen or halogen abstraction, Therefore, we have examined the reactions of radicals with disulfides and peroxides. Several features of S-S and 0-0 bonds make them attractive for study. (1) Radicals cleave disulfides a t the S-S bond and by a reaction that could be a direct displacement.
+ RS-SI< +31-SR + RS*
&I*
(11)
(15) J. E. Bujake, XI. W.T. P r a t t and R . 31.Noyes J . A m . Ckcnz. S O L . , 8.3, 1647 (1961). (16) R a t e constants given are those determined in t h e presence of oxygen, which should favor a direct displacement. T h e degassed runs have approximately t h e same rate profile. (17) Se?, e . g , ref. 4, table 5. Furthermore, formulation of t h e alkyl bv of the relative iodide reaction a s in ea. 5 is suonorted .. . comparison . rates obtained by Pioyes with those observed by F. Evans, R . J . Fox and M. Szw-arc, J . A m . Chew Soc., 81,6414 (1960), for t h e reaction
w.
KI
+ CHa.
+R * + CHJ
'l'hry f i t ~ dthat t h r r r l d t y e rate+ arc R r Slt, 1 ; E t , 4 ; i-Pr, 19 3 ; I B!l, :17. .'l K a r p u y E. Y. U g l u i a and V. A . 3lalyanov. (18) 0 A . Reutov, 1'. I z s r s l . A k a d . A'aick SSSK,Oldel. X k i m . ,";auk, 1311 (1960); C. A , , 54, 23637e (1980) (19) R I . Henchman and K. U'ulfgang, J . A m . Ckeln. S o c . , 8 3 , 9991 (l961),
L.
Yol. s4
PICKERING
When polymers are formed in the presence of disulfides, their rate of growth is unchanged, but their molecular weight is lowered.2"f21Furthermore, thc transfer process puts an RS group a t each end of the polymer. 2 2 , 2 3
+
+ RS*
RSM,,* RSSR --f RSM,,SR RS* T I M-+ KSh'I:I,t*
+
(12) (13)
If a cyclic disulfide is used, the rate of polyincrization is nearly unchanged, but up to 18 sulfur atoms are incorporated into the M,,*
+ P-9 R-CH2+;\In-S--CHp-R-S.
%
%
Mn-S-CH2-R-S-Mn-
(15)
This implies that S-S bond cleavage is involved; if hydrogen abstraction were occurring, only 2-4 sulfur atoms would be incorporated per molecule.
s-s C,"IEH]
s-s
'R-GH
$H :I'
+
-)
\
'R-CHChfno
,
pR-cH-M,-H -s)
cs-s \ R-CH--M,*
7 1 M ~
--f
dimer
(16)
(17)
( 2 ) Cleavage of S-S and 0--0bonds by radicals is extremely fast, and therefore competes with other possible radical reactions. (3) In the cleavage of benzoyl peroxide by various radicals, O'*-labeled carbonyl oxygen has been shown to retain its identity.25 O* 11
O*
+ C6Hs-C-O-O-C-CsH:, ll
I
+ O* ~'
LI-O-C-C6Hj
+ CsHjCOs**
(18)
The mechanism probably involves an S H 2 reaction on the peroxide oxygen, and it is reasonable to suppose that other peroxides also react with ~ radicals by an S H mechanism. ~ of disulfides occurs by (4) The S N reaction backside attack analogously to Walden inversion in carbon compounds. Fava, Iliceto and CameraZ6gz7 have convincingly demonstrated that S N ~ reactions on sulfur and carbon decrease in rate by similar amounts as backside hindrance is inincreased. The el'fect of backside hindrance for the radical displacement on sulfur, therefore, may be compared with that obtained for the SNZ reaction on sulfur, and that, in turn, with the S K ~ reaction on carbon. Either of the latter two re(20) C . Walling, "Free Radicals in Solution" John W l e y aura Sons, Inc., S e w York, Pi Y . , 1957. (21) \V. A. Pryor, "Mechanisms of Sulfur Reactions," 3IcGrawHill Book Co., Inc., Xew York, N. Y.,1962. (22) A . 1 '. Tobolsky and B. Baysal, J. Anr. Clzem. Soc., 75, 1757
,."--\
'I$:;"
R , ~ f pierson, , -4 J . costanza and A. H. U'einstein, J . Polylnev Sei,, 17, 221 (19a5) (21) XV. H . Stoukmayer. R. D. Howard and J. T. Clarke, J . A m C h e m Soc , 7 6 , 1736 I I H i X ) . (2.5) 1,:. IT. Ilrew and J C . hlartin, Ckriiiistiy & I n d u s t r y , 9 2 5 (1959); U. H . Deniiey and G. Feig, .I. A m . Cliem. S O L . ,81, 5322 (1959); W. Doering, K . Okamoto and H . Krauch, ;bid., 82, 3279 (1960). (26) A . Fava and A . Iliceto, ibid., 8 0 , 3-178 (19.58) ( 2 7 ) A Fava, A . Iliceto and E. Camera, ibGi., 7 9 , 8:13 [l!t,j7)
CHAINTRANSFER OF DISULFIDES AND PEROXIDES
July 20, 1962
actions may be used as models for backside attack. Experimental Chemicals .-Liquids were purified by repeated fractional distillation until gas phase chromatography showed a single peak using an F and M model 609 instrument with flame ionization detector, pTogrammed a t 18"/min. from 50" to 300", and a 2-ft. Silicon rubber column. Impurities are believed to beless than 5Op.p.m. Methyldisulfide, Eastman Kodak Co., b.p. 111.2-111.5n; propyl disulfide, Eastman Kodak Co., b.p. 90.5' (11 mm.); isopropyl disulfide, Aldrich Chem Co., b.p. 63.0" (11 mm.); butyldisulfide, Eastman Kodak Co., b.p. 118.8' (20 mm.) and 71" (9 mm.); isobutyl disulfide, Eastman Kodak Co., b.p. 105.5' (20 mm.) and 89.0' (9 mm.); sec-butyl disulfide, Columbia Chemical Co., b.p. 106" (9 mm.) (darkens in the light); t-butyl disulfide, Eastman Kodak Co., 86" (20 mm.), butyl sulfide, Eastman Kodak Co., 151'; t-butyl sulfide, Aldrich Chemical Co., 151O ; t-butyl peroxide, Lucidol, 62.5' (119 mm.). Styrene was stored over white Drierite for 24 hours to remove then distilled; b.p. 50" (20 mm.). Benzoyl disulfide colors if heated in solvents. I t is best purified as follows: 25 g. of hldrich Chemical Co. benzoyl disulfide was dissolved in 350 ml. of acetone and filtered. All but 50 ml. solvent was blown off and decanted. The recovered solid was recrystallized first from 150 ml. of ethanol mixed with 150 ml. of ethyl acetate, then from 300 ml. of ethanol, and finally from 120 ml. of ethanolmixed with 120 ml. of ethyl acetate. This yields about 4070 recovery of pure white platelets, m.p. 131.2-132.2'. 2,P-Dimethylbutane was Phillips research grade, 99.99yG. Runs.-Runs were done in ampoules prepared by constricting a neck half-way down a 25 X 200 mm. ignition tube. Materials were weighed into the ampoules, which were then connected to a vacuum line by a rubber stopper. The ampoule was immersed in liquid air, evacuated to 0.2 inm., pressured with nitrogen, and the contents melted to de-air. The cycle was repeated twice more; the ampoules were then frozen once more and sealed off under vacuum. Polymerizations were carried out by immersing ampoules under a light-opaque oil held a t 60.00 =!= 0.001 They were carried to 10-15yG conversions. Ampoules were opened and yo conversion determined by weighing 0.2-0.5 g. of the mixture onto a tared piece of aluminum foil, placing the foil under an air jet on an open hot-plate, and reweighing when all the styrene monomer had evaporated (about 30 min). This procedure proved faster than isolating the entire polymer and just as reproducible. Yield was independent of the aliquot size. Pure styrene gave the correct ratez8 of thermal polymerization, 0.105c/hr., and the method is therefore thought to be accurate. Degree of polymerization, P , was obtained by weighing 30 mg. of the polymer into a 25-m1. volumetric flask, diluting with dried benzene a t 30.00 =k 0.02". Flow times were measured in Cannon-Ubbelohde viscometers and were converted to specific viscosities (kinetic energy corrections were negligible), using qsp = ( t - to)/to. Intrinsic viscosities were calculated from the equation of Gregg and may^.^^
tp .
[?I
+
[(l
-
1)]/0.75C" (19) Arid the inverse of the degree of polymerization from" =
1/P =
1.5tlSP)1/2
(6.2275 X 10-4)/[7]1.37
(20) The values obtained for l/Pofor pure styrene a t 60" averaged 1.16 X which agrees with previously reported values .28,31 Theory of Chain Transfer.-The molecular weight lowering produced by the presence of a transfer agent can be relateda2 to the ratio of the rate a t which the polymeric radical attacks the transfer compound relative to the monomer. The equation, if the transfer agent is not an initiator, is
1/p = 1/po
+ C ((T)/(M))
(21)
(28) P. H. Boundy, R. F. Boyer and S. M . Stoesser, ''Styrene,'' Reinhold Publ. Corp., New York, N. Y., 1952, p. 208. (29) R. A. Gregg and F. R. Mayo, J . A m . Chem. SOL.,7 0 , 2373 (1948). (30) R. A. Gregg and F. R. Mayo, ibid., 76, 6133 (1953). (31) R. A. Gregg a n d F. R. Mayo, U i s c . Faraday S O L ,2 , 328 (1947).
(32) 1'. R. RIayu,,J .
4in.
C h L u i . S L I L6. ,6 , 2324 (1943); ref 20, 11. 130.
2707
Where C = k t r / k , , T is the transfer agent, 1LI i3 the tnononier, and parentheses indicate inoles/liter. The rate constants are for the reactions ktr
+ T +M n -T f T'* fast T'* + IZM---f T'-M,* k, AI,&* + M +
M,,*
M,'+I.
(22)
(23) (24)
X graph of 1/P 285. the mole ratio (T)/(bl)gives c'. Since the propagation rate constant k , is known for styrene,2o the absolute values of kLr mav be calculated; dt 60", k t , = 145C set.-' X-l. Transfer Agents Which Are Also Initiators.-If T acts as an initiator as well as a transfer agent, a more complex treatment is necessary.33 The molecular weight lowering produced by transfer and by initiation must be separated. &Tone of the sulfur compounds studied here increases the rate of the polymerization of styrene a t 60" in the dark.34 The peroxide, di-t-butyl peroxide, is an initiator and data for i t will be published ~eparately.3~Transfer constants, however, are not sensitive to small amounts of initiation. Even benzoyl peroxide, which is a very effective initiator a t 60°, gives C = 0.1 from the limiting slGpe of eq. 21 and 0.055 if initiation is taken into account. I
Data Table I gives the value of C obtained from a least square treatment of eq. 21, the standard deviation, and the number of runs for each compound. Table I also cites previous determinations of C . The lack of agreement is not serious; most of the values previously reported are based on a single run.
Discussion Comparison of SHZ and SNZ Reactions.-Table I1 compares the relative rate constants for the reactions : M n * + R-S-SR +R-S-Mn RS* (25) 1: *
+ SO,' + R-S-SO:J-- +R-S-SOv- + SO8I& + R-S-SK +R-S-$R + KSI** -!- R-CH2-I +R-CH2-*I + 1. 1 ' - + R-CH2-X R-CH2-Y + X I- + K-CHZ-Cl+ R-CHz-I + C1----f
(26)
(27) (28) (29) (30)
In these comparisons, the atom being attacked has been varied while the substituent R is kept con.
4
stant. Thus, methyl disulfide, CH3-S-(SCH~),has a sulfur atom comparably hindered to ethyl
4
halide CHS-CHZjX), where the arrow indicates the site of attack and parentheses enclose the leaving
4
group. Similarly, t-butyl disulfide, t-Bu-S-(SBu-t), is comparably hindered to neopentyl halide,
4
t-Bu-CH2-(X). Table I1 has been constructed accordingly. Comparison of the two columns for ionic reactions in Table I1 illustrates the point made by Fava26r27 the S N reaction ~ has a similar rate profile regardless of whether carbon or sulfur is being attacked. The ratio of rates for R = methyl to (33) F. R. Mayo, R.A . Gregg and RZ. S . Matheson, i b i d . , 75, 1691 (1951). (34) This agrees with t h e findings of a considerable number of workers w,hose work is reviewed in ref. 21 (see in particular, T. Otsu, J . Polymer Sci., 21, 559 (1956)). Styrene, as discussed in ref. 21, is particularly insensitive t o initiation by disulfides. ( 3 3 ) 11'. A . Pryor and E. P. Pultinas, to be published.
WILLIAMA. PRTOR A N D TIMOTHY L. PICKERING
2708
V O l . €74
TABLE I CHAINTRASSFER CONSTANTS AT 60"
cx
Compound
104b
Std dev.c x 10'
C X
lo4
1it.d
Disulfides Methyl 12 94.0 0.203 Propyl 12 23.4 .199 .115 ;-Propyl 12 6.60 Butyl 14 23.78 .488 120,'68' &Butyl 11 20.37 ,302 11 4.38 .496
