C. SIVERTZ
34
Vol. 63
extensive, systematic studies will be needed to in difference spectra of the shape of those in Fig. 1, secure the interpretation of even the main positive the prominent negative peak a t 680 mp definitely and negative difference bands; but we feel that indicates reversible bleaching of chlorophyll.
STUDIES OF .THE PHOTOINITIATED ADDITION OF MERCAPTANS TO OLEFINS. IV. GENERAL COMMENTS ON THE KINETICS OF MERCAPTAN ADDITION REACTIONS TO OLEFINS INCLUDING CIS-TRANS FORMS BY C. SIVERTZ The Univarsity of Western Ontario, London, Canada Received July 11, 1068
Some aspects of previous work are drawn together to illustrate some common problems in free radics.1kinetics with particular reference to termination. It is shown how useful information can be elicited by the study (a) of the attack of a common radical on various substrates and (b) of different radicals on a common substrate. Reference is made t o a mechanism for "buried" radicals. Finally, a preliminary report is made of a new technique which employs cis and trans forms t o help penetrate the micro kinetics of a radical attack on ?r bonds.
Introduction Since the advent of a quantitative approach to free radical chemistry in terms of the meas'urement of absolute rate constants, much effort and ingenuity has been spent on the "isolation" of a given radical so that its termination fate and action on functional groups may be unambiguously interpreted. This neat arrangement cannot easily be achieved nor is it always desirable, so that one is usually involved with two primary radicals a t least. These in turn may attack several functional groups yielding other radicals in side reactions which can soon produce intractable equations. I n the work to be described, the object has been primarily to learn about radicals, not to elucidate a particular mechanism, no matter how complex. The reactions chosen involve the free radical addition of mercaptans to various olefins, for example, CH3SH CH2=CH2 --+ CH3SCH2CH3. Several reasons influenced this choice after some necessary preliminary facts were known. First the propagation steps, attack and transfer, are ordinarily very fast for mercaptan addition so that a long kinetic chain producing a single main product is ensured. A long chain permits neglect of termination reaction products compared to the main chain products. Secondly, we hoped to make studies of the same basic process in both liquid and gas phase. Even CH3S radicals terminate without the ministrations of a third body in the gaseous phase thus avoiding a termination complication in that phase. For example, this is not true for the gaseous addition of HBr to o1efins.l Finally one can be assured that the field of bio-kinetics will welcome any and all knowledge about the chemistry of mercaptans. If one is to draw conclusions from rate measurements concerning the specific activity of radicals and relate the structures involved, i t is essential that absolute constants be measured since the overall rates of such reactions always involve a t least a termination constant in addition to propagation. Consequently in the work described a dilatometer-
+
(1) K. Graham, Thesis, University of Western Ontario, London, Canada.
sector technique was employed, descriptions of which may be found elsewhere.2 Basic Equations A set of basic equations for a given reaction cannot be given a priori. Even the termination of methyl radicals in the gas phase required hundreds of man hours of research time before it was possible to assert that the mechanism is CHI CH3+ C2H6*. Some are not yet convinced. Preliminary work involving product analysis, but not termination products, with the system under discussion, suggested the following were the main reactions. We will not complicate the preliminary kinetics with another reaction reversing (2) which was discovered in gas phass work. In the liquid phase, reactions were initiated by the photolysis of azoisobutyronitrile (AIN), and of mercaptan in the gas phase. The rate of initiation kI was determined as in ref. 3 and by NO in the gas phase.4
+
Photoinitiation AIN Attack
Propagation Termination
]GI
kl
RS)
+RSM (= X) kd X + RSH +Product + RS
RS -I- M
Displacement
+ h.v +R (- RSH k,
X,+ R.I +Xr+l
(1)
(2)
(3a) (3b)
k4
2RS +RSSR RS
+ X +RSX
(4)
k6
ks
2% --+ 3
(5) (6)
Steady States We will write the steady states for the general case of a monomer such as styrene, capable of propagation and hence the production of a family of X radicals whose total concentration will be designated Z X = XI X, . . . . . A m
+ +
(2) J. P. Flory, "Principles of Polymer Chemistry," Cornell Univ. Press, Ithaca, N. Y.,1953,p. 148. (3) R. Back and C. Sivertz, Can. J . Chem., 84, IOGl (1954). (4) W. Andrews, Theais, University of Western Ontario, London, Canada.
PHOTOINITIATED ADDITION OF MERCAPTANS TO OLEFINS
Jan., 1959 (I) (RS) kI
35
+ kd(RSH)[ZX] = kj(RS)(M) + 2k'4(RS)* j- kj(RS)(ZX)
+
+
(11) (X,) k l ( ~ ~ ) =( k~p)( X 1 ) ( ~ ) ~&)(RsH) 2k16(x1)(XX) k6(Rb)(k:1) (111)
(22)
(I\{)
?ir
kp(%)(M) = kp(X.)(hl)
+
+ kd(Xz)(RSH) 4+
2 k ' 6 ( k % ) ( Z X ) ks(RS)(X,) k(Xr-i)(M) = k,(Sr)(M) kd(Xr)(RSH)
+
2k'g (X,) ( Z X )
+
+ kdRS)(Xr)
N.
2
h
+ Partial'terminations
(7)
which permits us to express (RS) in terms of (ZX) provided the terminations (kl) are relatively negligible. The Termination Problem Summing the steady-state equations
+
kI = 2l~'4(RS)~ 2ks(RS)(ZX)
+ 2I$'g(ZX)'
-
+ 0.8
4
6 0.7 -
v
2
0.6
This statement assumes as usual that k, and ka are independent of radical size. In summing I1 to 00 we have a useful relation kl(RS')(M) = kd(RSH) (ZX)
0.9 .
0.5
-2.0
-1.0 0.0 1.0 Log light flash (sec.). Fig. 1.-Cyclohexene n-BUSHreaction: half-life determination.
(8)
This can be expressed more usefully hy defining 21~'= ~ k4, 2kt6 = ks, kg = 9 dlcrks, (9 3 0) and u2 = kJks. Also through (7) (RS) = [kd(RSH)/k l ( M ) ] ( Z X ) . We will henceforth write ZX = X , and we then have .-
(9)
Any theoretical anticipation of 9 values will iiivolve the same fundamental principles of quantum chemistry and kinetic theory as the much treated problem of "crossed" interaction of molecules through van der Waals forces where, for example, we find the geometric mean, corresponding to 4 = 1, applies to all London-force molecules. Where considerable diff erence in polarity or resonance stabilization exists this would not be true. 4 values as high as 150 have been observed,6 but most reported values are much less.6 There should also exist differences in r$ between gas and liquid phase, and steric effects could play a big role. These considerations suggest two ways of dealing with (9). (1) Measure # and p.--Tsolate the X radical by arranging (M) ) ) ) (RSH). Then (X) = ( k l / k s ) ' / l and the half-life sector measurements evaluate k 6 ; similarly with RS and k d . The magnitude of pr may then be evaluated from observations of rates where both radicals are present. This procedure was first applied by Melville, Robb and Tutton.' It can be followed readily in copolymer studies. I n systems such as ours the success in isolating a radical depends on the ratio ukd/kl. If one has some approximate knowledge of this ratio, the isolation can perhaps be achieved a t least for one of the pair. ( 5 ) C. Walling, "Free Radicals in Solution," John Wiley and Sons, Inc., New York, N. Y.,1957, p. 323. (6) L. Eateman, Quart. Rev., 8 , 147 (1954). (7) H. W. Melville, J. C. Robb and R. C. Tutton. Disc. Faraday S O C . , 10, 6595 (1951).
0.8
1.6 2.4 3.2 4.0 Concn., nioles/l. Fig. 2.-l-Pentene-n-butyl mercaptan; rates us. concentration. 0
8
I
/
6 d
U
2 4 2
tI
'
0 0.4
1.2
2.0 2.8 Concn., rnoles/l. Fig. 3.
3.6
4.4
(2) Assume the Geometric Mean fl = 1 (which one may regard as the "ideal" law of termination). -A good approximate value of u may be employed based on a half-life for a large excess of M which gives k6 approximately and we may employ an estimated value of 5 X 10" for normal thiyl radicals.8 This permits us to explore the important concentration efects implicit in (9). We will see that certain important and firm conclusions can be drawn. Meanwhile, work is proceeding on a new ( 8 ) M. Onyszchuk and C. Sirertz, Can. J . Chem.. 33, 1034 (1954).
3G
C. SIVERTZ
Vol. 63
styrene with 1-pentene, when the former is added to the 1-pentene butyl mercaptan reaction) we conclude tha't kl styrene is about 180 X kl for 1pentene or ca. 1.2 X 109 I./moles ~ e c . Even ~ when we estimate u to be about 30 this gives for ukd/kl = 3 X 10". Consequently the rate can be expressed
as
R
(kd(RSH) f k p ( M ) ) ~
(12)
From the gentle rise of rate with M at constant RSH, we calculate k, = 230 when (RSH)/(M) = 1, This propagation will be due chiefly to Xl a t these concentrations. It is much larger than for the macro styryl radical, viz., ca. 30 a t 25". Note that the termination constant ka = 5 X lo8 is also 0 0.8 1.6 2.4 3.2 4.0 several hundred times greater than for the macro Concn., rnoles/l. Fig. 4.-Cyclohexene n-butyl mercaptan; rates us. concen- styryl radical which is about 2 X lo6 at 25". This is to be expected from previous worklo here. tration. Likewise, when RSH is varied we obtain a linear method for k4 referred to below. With this pro- dependence over a wide range with an intercept which yields a value of IC, = 94 much nearer the cedure we have from (9) macro styryl value. Abnormal Reactions. (c) Figure 4. n-Butyl Mercaptan and Cycl~hexene.~-In this case k, = 0 and inspection shows a t once that as in the case where of styrene Ukd/kl must be quite small. From ka for cyclohexene we estimate u = 20 in spite of we = dkm6 which again Ukd (( ICl. The sharp plateau in Fig. 4 Rates in the Liquid Phase.-Benzene was em- with M varied, hence requires that the displacement ployed as solvent. From (10) the rate reaction is possibly several hundred times less than the attack. Moreover, the variation in rate with RSH shows an opposite behavior from 1-pentene; accelerating a t first with approximately the second power of the mercaptan. This has a reasonable Figure 1 shows a typical determination of the explanation when we make these assumptions: half-life h for radicals in the system cyclohexene (a) the radical attacks the double bond in n-butylmercaptan with a considerable excess of the such a manner as to introduce steric hindrance olefin. From the relation ks = l / k I h 2 we find to the displacement reaction. The first work of ka = 1.3 X logl./moles sec. Goering, et ai., on the addition of HBr to bromoSomewhat arbitrarily we can define addition cyclohexene, supports this view"; (b) the low reactions which can be interpreted by equation value of kd for the mercaptan transfer of hydrogen 11 as normal. Certain reactions described below makes the ordinary a-hydrogen abstraction of require additional basic reactions and hence may relative importance; and (c) finally it follows from be called abnormal. All velocity constants are considerations discussed below for thiophenol, expressed as liters/moles sec. that a reverse step to basic equation 2 must also Normal Reactions. (a) Figure 2. n-Butyl Mer- become important. (See equation 14 below.) captan and 1-Pentene.-In this case These and other details will be discussed in a kD = 0 and Ukd/kl = 0.5 for a best fit to both curve6 paper devoted to the cyclohexene r e a ~ t i o n . ~ k6 = 5 X 1010 and k ; ~estimated to be 5 X 10" and hence Meanwhile, it will suffice to say that when assumpu = 3 tions (b) and (c) are included in the basic equations, 1.4 x lo8 and ki = 7 x 10' kd we have for the rate Since the term in the denominator of (10) (ukd/kl = 0.5) is of the order of unity, consequently, on increasing monomer with RSH constant, we pass from regions where the rate is monomer dependent to a plateau where monomer has no further sig- in which u = .\/k.l/ks, kg is the termination constant nificant influence and similarly with the variation for the allylic radicals resulting from a-dehydroof RSH with monomer fixed. This interpretation is one way of concluding that one is indeed dealing genation of cyclohexene, viz., - . le= is the with consecutive reactions 2 and 3a involving velocity constant for such dehydrogenation and k'd is independent or "free" radicals. the rather low mercaptan hydrogen transfer constant (b) Figure 3. n-Butyl Mercaptan and S t ~ r e n e . ~for the allylic radical. Now provided the last two -In this case we find k6 = 5 X lo8, k d = 1.24 X terms of the denominator dominate, the results 108. By analysis of the competition for butyl (10) J. Longfield, R. Jones and C. Sivsrtz, Can. J . Res., B28, 373 mercaptan radicals (between small amounts of (1950).
c1>
(9) A. Harrison, Thesis, t o be published, University of Western Ontario, London, Canada.
(11)
.
H.L.Goering, P. I. Abelland B. F. Aycock, J . Am. Chem. Soc.,
74, 3588
(1952).
PHOTOINITIATED ADDITIONOF MERCAPTANS TO OLEFINS
Jan., 1959
shown in Fig. 4 follow. Pursuit of the implications of this in other related systems promises to throw special light on the micro process of double bond kinetics. (d) Figure 5. Thiophenol-l-Pentene.12-In this case the rate is first order in both monomer and mercaptan. This mechanism in turn did not become clear until the gas phase kinetics, discussed below, showed the existence of the reverse
s: 32 X d
P
+
24 $ -
M. The thiyl radical step in (2) XI k_l RS CaH5S now is resonance stabilized and hence again results in a weak composite radical X. Indeed the reaction CeH5S M 3 X should be endothermic when the composite radical represents a structure for the electron localized on the number two carbon. When this is incorporated in the steady-state equations, without propagation, we find in place of (7)
16
+
k l y ~ S(M,) )
+ ~ ( R S(MJ ) = I~(RSH) (XI+ P-i(X) + kT-l(X) + Terminations
Q4-
8
0.8
1.6 2.4 3.2 Concn., moles/l. Fig. 5.-l-Pentene-thiophenol; rates VB. concentration.
0
(13)
in which is shown the two alternative paths from the intermediate state, one 1eadi.ng to cis and the other to trans, when isomeric forms are relevant. We now have in place of (11)
37
h
a\
0.9 .
’*
0.8 .
and for the attack of thiophenyl radicals on 1-octene we conclude that uk-l/kl(M)>ukd(RSH)/kl(M)>l and the rate is first order in both M and RSH as found kd(kl/k-1)
(M) (RSH)wa/u
.r(
Y
2 0.7 -
(15)
I n other studies of cis-trans inversion in the liquid state (paper in preparation),l3 it has been shown that when cis-2-butene is employed instead of 1-octene, a trans form is soon formed. This was also found for the case of butyl mercaptyl radical, but for this strong thiyl radical, the relative rates involved are such that k d ( R S H ) > k - l , and hence we do not observe the effects of the term k-l(X) in the primary kinetics.12 Nevertheless it is a very active process and the determination of kl without considering k-1 will be too low (see below) as pointed out by Walling6 in reference to our work and observed independently in this Laborat~ry.’~J~ In the 1-pentene-thiophenol reaction half-lives have not been measured, and from (15) and Fig. 5 one can a t present do no more than extract a value for kdkl/k-I. A study also has been made of the addition of thiophenol to styrene12in which the system returns to a “normal” behavior similar to butyl mercaptanstyrene. This is to be expected since now in spite of the “weak” thiyl radical, kl > k d due to resonance stabilization in the composite radical, C6H5’& CH2-CH-CeH5.
Preliminary studies, some previously reportedI4 of the gas phase addition of methyl mercaptan to a series of olefins, have indicated that the kinetics responds to one additional basic equation which, as in the case of thiophenol, requires us to postulate an active dissociation of the composite radical X whose steady state now requires equation 14 instead of (7), and the rate for a l-olefin is found to be represented by equation 15 and for cis-%butene by (14). The experimental evidence for this is as given: (a) a rate first order in both mercaptan and olefin16 similar to Fig. 5 ; (b) an over-all negative activation energyI4; (c) the appearance of hms-Bbutene from cis as the addition reaction p r o ~ e e d s ~ *(see J ~ Fig. ~ ~ ~6) ; (d) preliminary measurements of the half-life with a semi-square wave character indicated a very high value for the termination constant ca. 10l2I./moles s ~ c .If~ this is confirmed, it suggests CHSS termi-
(12) R. Pallen and C. Siverte, Can. J. Chsm., 81, 723 (1957). (13) R. Townshend, Thesis, t o be published, University of Western Ontario, London, Canada. (14) C. Siverte. W. Andrews, W. Elsdon and K. Graham, J . Polymer Sci., 19, 587 (1956).
(16) P. Pallen, Thesis, to be published, University of Western Ontario, London, Canada. (16) R. Townshend, R. Pallen and C. Siverta, “The Mechaniems of Radical Induced cilrtrans Inversion,” paper preeented a t N. Y. meeting A.C.S., Sept. 13, 1957.
20 30 Time, min. Fig. 6.-(a) Addition of CHaSH to cis- and tran.s-2butene; (b) simultaneous conversion of cis to trans.
10
Gas Phase
C. SIVERTZ
38
nation predominates as a consequence of a strong dissociation of the composite radical X. It should be noted that equation 15 has the form of a third-order process when one recalls that U,!U = u4, represents the concentration of the thiyl radical. There seems, however, little doubt that it concerns, so far as the half-life of the X radical is concerned, a rather sticky collision, defined by k-1. It should be recalled14 from data on the over-all negative activa,tion energy, that the exothermic enthalpy change for the process CH3S kl CzH4+X is ca. 12 kcal. or roughly about half the AH for the complete addition. On the other hand, for 2-butene, cis +. trans AH t 1 0 0 0 cal. (T = 25'). We therefore have the following energetics in mind for the cis-trans radical induced inversion
+
RS
kP
klT
k"-1
kT-1
+ cis ~ri
trans
+ RS
Vol. 63
for the entire conversion of cis to trans may then be set up based on (13). This involves some quite complex kinetics to be discussed later. Meanwhile, two simplifying assumptions appear j ustified by the analysis, namely, that the principal termination is due to the CH3S radical (4,and X = 1. While the latter is obviously not true, from the magnitude of the equilibrium constant for cistrans a t 25" (K = 0.76) and the energetics of the reaction, X could not be far from unity. The net rate of disappearance of cis may then be written
-do = k p (Mc)(RS) - k-p(Xj df
(16)
and if we define (Mc) = ( M C o a )substitution , for (X) made from (13), (RS) = w4 and h = 1 we have K = equilibrium constant for cis
trans
01
=
conversion
The rate of initiation kl was determined by NO +( - 12 kcal.) (-11 ked.)+ inhibition and kd by the sector t e ~ h n i q u e . ~The The equilibrium constant for cis trans then is plot of data from Fig. 6 in the form of 17 yielded a line from whose slope a value of 6 X 107 l./ K = - kl0kT-~/klTkc-l which can be expressed straight moles see. for klT was calculated.16 This is about K = X ICicIkiT. Determination of the True Magnitude of the 8 to 9 times faster than that reported above for Attack Constant kl.-As pointed out above, ob- butyl mercaptan and l-pentene. It is intended to carry through a program deservations of the inversion of cis to trans for the addition of butyl mercaptan to cis-2-butene in the signed to observe such cis-trans kinetics in both liquid phase13 prove that the attack constants liquid and gas phase, induced by an array of radideduced above for, e.g., l-pentene are low. We cals and also with the same radicals in media of varying properties as to, for example, dielectric and now attempt to measure the true" value for IC,. We assume as implied in equation 13 that one viscosity. Special interest attaches to the use of cis and trans form only of the composite radical (X) is involved dideuterated ethylene as a basic structure in prein both the dissociation reaction k-,(X) and the liminary work now underway by Dr. D. Graham displacement reaction kd(RSH) (X), and also that in this Laboratory. Here, complications due to a n attack on cis or trans results in the same inter- monomer dehydrogenations should be eliminated mediate. While these assumptions may not be a t normal temperatures. After such studies it correct, it follows from the energetics of cis, trans will be instructive to observe a series of substithat any values thus derived cannot be much in tuted cis-trans structures attacked by a single or selected radical from which conclusions may be error. Figure 6 represents the experimental results drawn with regard to steric and inductive effects. Finally, when cis-trans rates are sectored in these from preliminary observations made by Pallen16 of the simultaneous rate of addition and formation systems under conditions where little addition of trans from cis-2-butene (T = 25") the latter occurs, one should be in a position to derive the determined by stopping the reaction and making termination constants for relatively isolated radicals gas chromatographic analysis. Details of this in derived from, for example, the photolyiss of Rgas and liquid phase will be published. In 20 S-S-R. Acknowledgment is made of student workers minutes while the cis-trans inversion has nearly reached equilibrium, the addition reaction has con- contributions besides those mentioned in reference sumed only about 3y0 of the mercaptan. If one works, namely, A. Harrison, R. Dyble, R. Townmeasures the ratio of these rates a t zero time ex- shend, Y . Ebisuzaki and J. McIntyre. Also we pressed in proper units one evaluates kT-, (X) are pleased to acknowledge some financial support from' the National Research Council of Canada, kd(RsH)x; similarly by starting with the trans and scholarships held by A. Harrison, R. Townform ko-1 X/kd(RSH)(i) and hence the value for shend and Y. Ebisuzaki granted by the Ontario X could be determined. An integrated expression Research Foundation. + .
-
r
*
