ISOMER DISTRIBUTION IN AROR.I.\TIC ELECTROPHILIC SUBSTITUTION
June 3 , 1CLjt5
benzenesulfonate and 1.35 g. (10% excess) of potassium acetate in 25 ml. of acetic acid was heated under reflux for 48 hours. The products were isolated as in the acetolysis The of endo-norbornyl-2,3-C214 p-bromobenzenesulfonate. yield of the mixture of nortricyclyl and exo-dehydronorbornyl acetate was 1.2 g. (63%). The corresponding alcohol mixture (0.76 9.) obtained after lithium aluminum hydride reduction absorbed 17y0 of a molar equivalent of hydrogen. The hydrogenation product (0.64 8.) was mixed nith 0.70 g. of inactive norborneol and the whole was oxidized with potassium permanganate, giving 0.69 g. (60% based on the norborneol present) of cis-cyclopentane-l,3-dicxboxylic acid. Results of radioactivity analyses of the tiegridation products are given in Table I . Solvolysis of em-Dehydronorbornyl-2 ,3-C2I4 p-Bromobenzenesulfonate in Formic Acid.-A mixture of 7.0 g. (0.021 mole) of mixed p-bromobenzenesulfonates prepared from the isomerized e m - and endo-dehydronorborneol-2,3C2“, 1.45 g. (0.021 mole) of sodium for2ate and 40 ml. of anhydrous formic acid was heated a t 45 for 1 hour. The products were isolated as for the acetolysis experiment. The recovered endo-dehydr0norborny1-2,3-C2’~ p-bromobenzenesulfonate weighed 2.5 g. (36Y0). The yield of nortricyrlyl and rao-drhydronorbornyl formate mixture, b.p.
[CONTRIBUTION FROM
THE
303 7
60” (8mm.), was 1.0 g. (34%). The mixture of formates was converted t o the corresponding alcohols which absorbed 4y0of a molar equivalent of hydrogen. The hydrogenation product and 0.40 g. of norborneol carrier was oxidized to give 0.40 g. of ciscyclopentane-l,3-dicarboxylicacid. Degradation and radioactivity assays were carried out the usual way. eltdo-Dehydronorb0rny1-3-C~~-amine with Nitrous Acid.The experiment using acetic acid as solvent was carried out The crude ester from as with endo-n0rbornyl3-C~~-amine.3 2.5 g. of endo-dehydronorborny1-3-C1*-amine and sodium nitrite in acetic acid was cleaved with lithium aluminum hydride and yielded 0.56 g. of mixed alcohols, m.p. 85-95”. The product absorbed 17y0 of one molar equivalent of hydrogen. The hydrogenated material was diluted with carrier em-norborneol and degraded as before. The aqueous fluoboric acid deamination employed the previously described procedurea and 2.5 g. of amine afforded 0.60 g. of crude alcohol mixture, m.p. 95-103’. On quantitative hydrogenation, 7% of one molar equivalent of hydrogen was absorbed. The reduction product was degraded in the usual way. CAMBRIDGE 39, MASS.
DEPARTMENT O F CHEMISTRY O F PURDEE USIVERSITY]
A Quantitative Treatment of Isomer Distribution in Aromatic Electrophilic Substitution’,2 BY
CIIARLES
w. L f C G A R Y ,
JR.,3
Y.O K A M O T O * AND HERBERT e. B R O W N
RECEIVED DECEMBER 20, 1954 An extension of the Hammett relationship to electrophilic substitution reactions is proposed. I t has been demonstrated that the usual side-chain up substituent constants are not suitable for such electrophilic reactions. New up values are developed which permit the quantitative treatment of directive effects in aromatic substitutions. Good agreement is obtained for para/meta ratios calculated in this way with ratios observed experimentally. Ortho and meta reactivities in electrophilic substitutions generally follow a relationship similar to that previously proposed for pura and meta substitution. Only in the case of mercuration, chloromethylation and isopropylation are serious deviations observed. I t is proposed therefore that the present quantitative treatment can be extended to include ortho substituents in many electrophilic aromatic substitutions where the steric factor is either moderate or negligible. The treatment has been extended mathematically to polysubstituted benzene derivatives. I t is shown to be identical with the relationship demonstrated previously on an empirical basis for the chlorination and mercuration of the methylbenzenes.
I n the development of the theory of aromatic substitution, attention has been directed primarily toward the aromatic component. Thus, orientation and relative rates have been interpreted qualitatively in terms of various electrical and steric effects in the substituent already present6 I n general, the effect of the activity of the substituting agent upon isomer distribution has been largely ignored. In a few cases where steric and polar explanations fail, such as in diazonium coupling,0a such activity has been invoked occasionally to account for ortholpara ratios.Ob A detailed discussion of the importance of the activity of the attacking species in controlling the (1) Directive Effects in Aromatic Substitution. VII. Previous publications in this series: THIS J O U R N A L , 77, 2300 (1955); 77, 2306 (1965): 77, 2310 (1955). (2) Based upon a thesis submitted b y C. W. McGary, Jr., in partial fulfillment of t h e requirements for t h e degree of Doctor of Philosophy. (3) Purdue Research Foundation Fellow, 1962-1963; American Cyanamid Corp. Fellow, 1953-1954. Research assistant on a grant lrom t h e h’ational Science Foundation, 1954. (4) Research assistant on a grant from t h e Petroleum Research F u n d , 1884-1955. ( 5 ) C . I(. Ingold, “Structure and Mechanism in Organic Chemist r y . ” Cornell University Press, Ithaca, K. Y., 1953, Chapt. VI. (ti) (a) A. Lapworth and R. Rohinson, .!+‘em. Proc. . M a n c h r s f u Lit. o n d Phil. SOL.,78, 243 (1928). See p. 264. ref. 4. (h) P. R. 11. de l a Mare, J . Chem. SIJC., 2871 (1949).
para/meta isomer distribution has been presented recently.7,8 Thus, it has been shown that in numerous reactions of aromatic nuclei, including chlorination, chloromethylation, base strength, nitration, mercuration and isopropylation of toluene, the orientation can be correlated with the “activity” or “selectivity” of the attacking species, It has been d e m o n ~ t r a t e dthat ~ ~ ~these reactions obey the relationship log Pf = c log (prlmr) (1) where pf and mf are the partial rate factors for substitution in the para and meta positions of toluene and, presumably, of other monosubstituted benzenes. This expression was developed and tested as an empirical relationship. However, the validity of this relationship argues for the existence of a linear free energy expression for aromatic substitution of the same type as that developed by Hammett for side-chain aromatic derivatives log (k/k,) =
pug
( 2)
The relationship between the empirical expres(7) H. C . Brown and K. L. h-elson, THIS J O U R N A L , 76, 6292 (1963). (8) H. C. Brown and C. U’. McGary, Jr., ibid., 77, 2300, 230F, 2310 (1955). (9) L. P. H a m m e t t , “Physical Organic Chemistry,” McGraw-Hill Book Co.. Inc., New York, N. Y.,1940, Chapt. VII.
3038
CHARLES IT.AICGARY,JR., Y.OKAMOTC AND HERBERT C. BROWN
sion and linear free energv expressions of the Hamniett type may be demonstrated simply. Writing the Hammett equation to apply to the partial rate factors for para and metn substitution, we have log PI = P O , (3) log inf =
pum
Subtracting (4) from (3), we obtain log (prlmr) = d o , - urn) Dividing (3) by ( 5 ) we have
(4) (5)
This equation is identical with (1) with c = (UP
'
- urn).
It appears reasonable, therefore, t h a t a quantitative treatment of aromatic substitution should be possible. Indeed, Hammett has pointed out previously t h a t the relative reactivities in the nitration of monosubstituted benzene derivatives may be qualitatively correlated with the side-chain U values.g However, our attempts to develop a qzrantifatiere treatment based on these U-values failed. I t has been observed that the rates of rearrangement of a series of para substituted acetophenone oximes do not follow the U-values quantitatively and a separate set of values have been proposed for this electrophilic reaction.1° Furthermore, Condon has suggested that the a-values for electrophilic substitutions may differ from those in side-chain reactions, since para and nzeta reactivities in the halogenation and hydrogen fluoride-boron trifluoride reactions are inconsistent with the Hammett u-values.I1 More recently Roberts and his co-workers have looked into the possibility of a quantitative relationship between partial rate factors for nitration and the side-chaiii u-constants.12 However, large deviations were observed in the case of prim substituents. They attributed these deviations to resonance interactions which occur in electrophilic para substitutions, but do not occur to the same extent in the side-chain reactions. The nature of the problem may be clarified by reference to Table I where the data for thirteen reactions with toluene result in a ratio of log p , 'log nzf of 4.24 =k 0.20. The ratio calculated from the side-chain U-values for para and metn methyl groups is 2.46. From the foregoing discussion, then, the conclusion follows that. in order to correlate reactivity data for electrophilic substitution reactions, substituent constants must be taken independently from those for side-chain reactions. This conclusion is supported also by the theoretical consideration of resonance factors operating in electrophilic substitution reactions. l 3 Consequently, we undertook to develop a treatment for electrophilic substitution reactions which was independent of sidechain U-values. Determination of Reaction and Substituent Constants.-It is assumed that aromatic substitu(LO) D. E. Pearsori, J. F Baxter and J . C. Rlartin, J Oug. Chrni , 17, 1511 (1952). 111) F. E. Condon, THISJ O U R N A L , 74, 2.528 (1952). (12) J. D. Roberts, J . IC. Sanford, I;. L.J. Sixma. H. Cerfontain crtid t . ;bid., 76, 452.5 '1454). 1;. 1. 1 S i x m a , R p t - , fvniv c h i i i i , 73, 243 (1954).
PAR.l'I.41,
Vol. 7 7
TABLE I RATE12ACTORS IS TOLUESE SL~TRSTITUIION Partial rate factors
Reaction
Of
Chlorination Chloromethylation Basicity A, HF B, HT-RT; Kitration A , lh' B, 4,:" hlercuration A, HClOd cntalyicd "50
I17f
pr
5 . 0 870 4 . 3 7 430
600 117
&Pi
I,it.
log
ref.
VIE
3 21 4.11
21
7
I
B, Same, ;iOo C, Same, 75" L), uncatalyzed, 50" E, Same, 70" F,Same, 90' Detrixnethy1silyl:ition Sulfonylatioii Isopropylation Partial equilibrium factors calculated froin the data of A I . Kilpatrick and P. E . Luborsky, THISJ O U R S ~ L75, , 377 (1953). D . '4. hlccsulay and A . P. Lien, i b i d . , 73, 2013 (1931). C. I;. Ingold, A . LapTrorth, E. Rotlistein and D . TVard, J . Chem. Soc., 1959 (1931). 1-1. Cohn, E. D . Hughes, M . H. Jones and M . A . Peeling, -Vatwe, 169, 291 (1952). Private communication froin Professor R . A . Benkeser, Dr. H . R . Krysiak and Mr. 0 . H. Thomas. f 'CT'. E. Truce and C. W . Vriesen, THIS JOURNAL, 75, 5032 (1953). 0 S . C. J . Olivier, Rtc. t m v . chim., 33, 1963 (1914). h F FE. . Condon, THISJOURNAL, 70; 2265 (1948); 71, 3544 (1949). Data for nitration :it 18 and isopropylation are not included. iL
tion will obey a linear free energy expression of the Hatnmett type log ki =
pu+
(71
where K f is the partial rate factor, p is the reaction constant and u L is a substituent constant applicable to electrophilic aromatic substitution. lo The data i n Table I show that for such reactions, the following relationship exists between the U+-values for pnvn and meta methyl.
The methyl group is of low polarity and of relatively low susceptibility to resonance interactions, particularly in the meta position. Let us assume that the value of g + r r r - h f e will not differ significantly from u,, .\re ( - 0.069). l 4 Ifre can imniediatelv calculate a value for u+{ hre (-0.'793) and values of p for the various reactions for which data are available (Table I). These p-values are summarized in Table IT. I t should be emphasized t h a t sufficient data are now available to permit the determination of p onlv from substitution data for para and metci to methyl. However, i t mav be seen from the IOTV mean deviations in Table I1 that the agreement is excellent with but two exceptions, isopropylatioii and the metu value for nitration a t 18". These data are the numerical expression of the empirical relationship previously observed (equation 1) .'s9
14) ti €1 TatTe,
Chr,ii
Km(
53, IO1 'lQi'3)
3039
ISOMER DISTRIBUTION IN AROMATIC ELECTROPHILIC SUBSTITUTION TABLE I1 RE.4CTION CONSTANTS FOR ELECTROPHILIC SUBSTITUTION
IN
SUBSTITUTED BENZENE DERIVATIVES Reaction constants
Reaction
Chlorination Chloromethylation Basicity Kitration hfercurntion
Conditions
Pm
PD
P
Clz in HOAc a t 24" C H 2 0 in HOAc a t 60" with HCl and ZnCL A, H F B, HF-BF3 A, Ac0502 in Ac2O a t 18" B, H S O s in 90% HOAc a t 45' A, Hg(OAc)?in HOAc at 2.5' with HClOl B, 50'
-10.13 - 9.28 - 8.06 - 7.12 - 6 81 - 5.77 - 5 10 - 5 . ,i4 - 5.04 - 4.30 - 3.80 - 3.33 - 4.36 - :1 . 6,i
-10.05 - 9.00 - 8.94 - 7 39 - 5.95 - 6.03 - 5.19 - 4.98 - 4.75 - 4.10 - 3.86 - 3.59 - 4.16 - 2.15
-10.09f0.04 - 9 . 1 4 f .14 - 8 . 5 0 z t .49 - 7 . 2 6 i. .14 - 6 . 4 3 ~ 2 .48 ~ - 5.904Z .13 - 5 , 1 5 4 ~.0 - 5 2 7 i .28 - 4 . 9 o i ,15 - 4 . 2 5 3 ~.06 - 3 . 8 3 h .03 - 3 . 4 6 i .13 - 4 . 2 6 i .10 - 2 . 9 0 i .75
c, 75"
D, Hg(0.4c)z in HOAc a t 50"
Detrimethylsilylation Isopropylation
E, 70" F, 90" ArSi( CHI)3 HsO+ in HOAc a t 25" CsHs at 40" with AlCls
+
The pp-value for nitration with acetyl nitrate in acetic anhydride a t 18'l5 agrees closely with the pmand pp-values for nitration with nitric acid in 90% acetic acid a t 45' (Table 11). We therefore may conclude that the p-values for these two reactions will not differ greatly and adopt the average value of -5.92 i 0.10 for these nitration reactions. It is now possible to calculate u + ~ and u+,-values for substituents other than methyl by utilizing the available partial rate data for nitration of monosubstituted benzene derivatives. The o+-values obtained in this way are listed in Table 111 together with the corresponding U-values. l 4
I n Fig. 1 there are plotted the u+-values versus the U-values. The values for the meta constants follow the relationship u + = ~
0.902~~
In view of the large uncertainty in the experimental data from which the u +m constants were derived, it appears reasonable to assume that the u+,-values are identical with the corresponding Hammett values. Ufm
(9)
= urn
On the other hand, there is a significant difference between the u +p and the corresponding up-values (Table 111, Fig. 1).
TABLE I11 COMPARISON OF SUBSTITUENT CONSTANTS FOR ELECTROPHILIC ASD SIDE-CHAIN REACTIONS Substituent m-C(CHa)a m-CHr m-CHzCOnC2HE m-CHnC1 m-I m-COtCsHa
m-Br m-C1 9-C(CHda 9-CHs fi-CHnCOnCzHa fi-CHKI
9-1 9-c1 p-Br p-COnCnHs
Ua
L7+
-0.102 - ,0672 - ,0109 ,144 ,325 ,355 ,507 ,520 - ,317 - ,297 - ,172 ,00372 ,0176 ,149 .167 ,514
u+
-0.120 - ,0690
-
06 u
0.018 ,002 0 4
-
,352 .398 ,391 ,373 ,197 ,170
-
,027 ,043 ,116 ,147
02
- ,120 - ,128 5:
,184 ,276 ,227 ,232
,522 .678(~*)~
- ,180
-
,258 ,078 ,065 ,008 ,164
The U-values are from ref. 14. * Table I. ref. d . c C . K. Ingold and F. R.. Shaw, J . Chem. Soc.: 575I )l .e 'I'hc rnetIi>l:iti~,ii< l a t a .irc from i:n]iul>li\hcd \- cases i t 1s observed t h a t much greater amounts of the m t h o t h a n pura isomers are formed (Table V). This phenomenon is d u e presumably t o coordination oi the electrophilic reagent with the substituent followed h y an enhanced rate of attack of t h e w t h o position, This phenomenon x : i l introduce a n additional fartor to he con' cd in the rlevelopnient 01 , i complctc theoretical t r c a t ~ n c n tul ,lll~,~~tl,t,,l
June 5 , 1935
ISOMER DISTRIBUTION IN AROMATIC ELECTROPHILIC SUBSTITUTION
3043
which vary only in one particular partial rate factor. The criterion as to whether one, three or six straight lines will be obtained is determined by whether two, one or none, respectively, of the following relationships hold.
Application to Polysubstituted Benzenes.-An extension of the Hammett equation to polysubstituted benzene derivatives has been made for side-chain r e a ~ t i 0 n s . l ~In the case of electrophilic substitution, assuming additivity in the substituent constants, the equation may be written log kj = p + Cc'i ( 19) i
where co+iis the summation of substituent coni
stants of the various substituents located on the ring and kj is the rate of substitution in a given position relative to one position in benzene. By recalling the relationships between the partial rate factors and their respective u+i values and by solving for kj and adding the rates of all available positions, we obtain kt = C k j = z e + p cc+i i
0
IP y
I
I
0
I
2 LOG
kj
I 3 (mercuration),
I 4
Fig. 3.-Relationship betiveen the partial relative rates in the basicity and mercuration of methylbenzenes.
where kt is the over-all rate of substitution in the substituted aromatic relative t o one position in benzene and kf is the partial rate or equilibrium factor. I n the case of m-xylene, for example, kt is of2 2o& mf2. Consequently, from a knowledge of the partial rate factors, it is possible to calculate the relative rates of any polysubstituted benzene derivative. Using equations of the above type, Condon21has calculated the relative rates in the chlorination of the methylbenzenes. He also has utilized the reverse procedure in calculating the partial basicityfactors from the data on the reaction of hydrogen fluoride-boron trifluoride with the methylbenzenes." I n the preceding paper of this series we have demonstrated that the relative rates of mercuration of all of the isomeric methylbenzenes can be calculated with considerable precision from partial rate factors.22 It also has been observed empirically that the logarithms of the relative rates of halogenation and the relative basicities (hydrogen fluoride-boron trifluoride) of the methylbenzenes are linearly related.l1rZ3 It is important that these empirical relationships can be shown to follow from the present treatment. Thus, since will be the same for individual
+
+
c~+i i
comparisons, the following relationship exists log kj
= (p/p')
log kj'
(21)
where k, is the relative rate of reaction a t one site, kj' is the same for the comparison reaction, and p and p' are the reaction constants for the two reac( 2 1 ) F E. Condon, THIS JOURNAL,7 0 , 1963 (1948). 11'2) FI. C. Brown and C. W. hlcGary, Jr., ibid., 77, 2310 (1053) (25) H. C. Brown and J . D. Brady, ibid., 74, 3370 (195'2)
tions. For all available positions, however, the equation becomes kt Zki = Z ( k j ' ) p / p ' log Zkj = log Z k j ' p l p '
(22) (23)
T o be consistent with the empirical linear relationship, equation 23 should be in the form log Zkj =
(p/p')
log Zkj'
(24)
The condition for such a relationship to hold is that p / p ' should be nearly unity or that one term in the summation should be large. Apparently one or both of these conditions must be responsible for the excellent quantitative relationship observed in the chlorination and basicity reactions ( p = - 10.09 and - 7.26, respectively). However, this derivation suggests that the empirical linear relationship may not be completely general. Conclusions.-The Hammett equation, with the modifications here introduced, appears capable of correlating the available data on electrophilic aromatic substitutions. The treatment offers promise of providing a truly quantitative approach to isomer distribution in aromatic substitution. great deal of quantitative data on aromatic substitution will be required to test the approach presented in this paper. %-e wish at this time to encourage other workers to undertake research in this area in order that a rigorous test of the utility of the treatment may soon be possible. Acknowledgment.--a'e wish to acknowledge the financial support of the Purdue Research Foundation, the american Cyanamid Corporation, the National Science Foundation and the Petroleum Research Fund which iiiatle this study possible, LAPASETTE,
INDIANA
