AUGUSTH. ~ I A K AI N D DAVID H. GESKE TABLE IS
CHEMICAL
SHIFTS AND SPIS-SPIN
COUPLINGS I N
OXTBESZESES
Compound
1,2-Dimetlioxybenzene
1,3-Diniethoxybenzene
1,4Dinietliosj-benzene 1,2-Diniethoxy-4inetli;-lbenzene
6c.
Group
p.p.m.
c-1,2 CH-3,6 CH-4,5 OCHI c-1,3 CH-2 CH-4,6 CH-5 OCHj c-1,4 CH-2,3,5,6 OCHa c-1 t c - 2 CI-I-3 CH-6
42.4 80.5 71.9 137.5 31.7 92.0 86.7 62.8 138.0 38.6 78.1 137.2 43.8 80.0
+
[ COSrRIECTIOS FROM
THE
c-4 CH-5 OCHI CHI
DIMETH-
G(obsd.) G(calcd.), J, p.p.m. c./s.
-4.3
3. 5
158
1.9
160 144
1.0 -1.0 0.7 1.8 -1.1 1.1
158 161 160 145
158 146
156
DEPARTMENT
OF
1101. 83 02 71 137 172
3
6
158
4 2
142 120
likely that the small differences observed between the different compounds are real. The order anisole, 1,3-dimethoxybenzene, 1,Zdimethoxybenzene (and the 4-methyl derivative), 1,4dimethoxybenzene is that expected if the interactions take place through the ring and are somewhat altered in the ortho compounds. Spin-Spin Coupling Constants.--111 ring CH coupling constants in the phenols and methoxybenzenes are the same as that in benzene ( l a 9 c./s.) within the experimental uncertainties, and the CH3 couplings are the same as those in the methylbenzenes.'O The OCH, couplings are all within the range 144 =t2 c.,/s.
CHEMISTRY, HARVARD UNIVERSITY, CAMBRIDGE, LIASSACHUSETTS]
Electron Spin Resonance and Polarographic Investigation of Substituted Nitrobenzene Negative Ions BY
-4UGUST
H.
L h K I AND
DAVIDH. GESKE
RECEIVED SEPTEMBER 29, 1960 Elcctroii spin resonance spectra have been obtained for 14 pala substituted nitrobenzene moiionegative ions t h a t i r e r e generated in acetonitrile solution by electrochemical reduction within a microwave cavity. Polarographic d a t a were also obtained for these para substituents and for 8 mete substituents. An excellent linear correlation was found between the half-wave potentials for t h e nzeta substituents and un and u0 values. From gross deviation of p-nitro and p-amino substituents from t h e ??zeta correlation line, pava resonance energies are estimated. T h e W4 coupling constant of the nitro grouji is found t o decrease as t h e ease of reducing the parent molecule is increased by t h e para substitution a n d zice versa. Measurement of ring proton coupling constants leads t o t h e conclusion t h a t the total electron spin magnetization in t h e four central ring x-orbitals is approximately constant. T h e para substituent merely effects a redistribution of the magnetization among these ring positions. Evidence based on hyperfine structure is presented for hindered rotation of t h e aldehyde group of para iiitrobenzaldehyde anion, and a maximum rotational frequency of 2.8 X IO6 C.P.S. is estimated. Comparison of the F19 hyperfine coupling constant of p-fluoronitrobenzene with t h e pare proton coupling constant of nitrobenzene leads to the ratio ~ F / U H = 2.12; from this ratio is cstiinated t h a t QF = -47.5,
Introduction
In recent years numerous electron spin reso-
reduction of the parent inolecule RXriYO,
+e
.RArNOl-
(1)
nance (e.s.r.) studies have been made of organic free radicals, particularly aizion free radicals pre- using acetonitrile as the solvent medium. An pared by one-electron reduction of the neutral earlier study,2 which demonstrated the experiparent molecule.' n'hile a variety of anion mental method, dealt with nitrobenzene anion radicals have been studied, no attempt has been radical, and a subsequent investigation3 was conmade t o examine a related series in which a single cerned with the three dinitrobenzene anion radisubstituent is varied. Such a project is of particu- cals. The experimental simplicity and convenience lar interest in view of the possibility of estimates of the electrogeneration technique have greatly of odd-electron distribution derived from exami- facilitated the present extensive study. Although a polarographic study of the series nation of the hyperfine structure of the e.s.r. of para substituted nitrobenzenes was not the absorption spectrum. We have chosen to examine a series of para primary objective of this study, half-wave POsubstituted nitrobenzene anion free radicals (.RAr- tential and current constant data were obtained. NO?-). The interaction of the odd electron with Availability of both e s r . and polarographic data the nitrogen nucleus is expected t o depend on the provides an opportunity to relate the two quantinature of the substituent. R. The exDerimenta1 tatively. approach is the same as that presented p r & i o u ~ l y ~ * ~ ; Experimental the anion is generated by Reagents,-Sjource and purification of the solverlt, ace(1) (a) J €?. Wertz, Chem. Revs., 66,829 (1955); (b) D. J. E. Ingram, "Free Radicals as Studied by Electron Spin Resonance," Academic Press, I n c . , h-ew I'urk, N. Y . , 1958, pp. 13.5-109. (2) D. H. Geske and A . H. Maki, J . A m . Ch?tn. Soc., 82, 2671 (l9li0). (3) A . 11. Ndki m d I). A . ( k c k c , J . c ' h c w . Z'hy,~., 33, 825 (1960).
tonitrile and supporting electrolyte, tetra-Z-propylalIlmonium perchlorate, were previously given .z Cnless Otherwise noted organic compounds were secured from Distillation Product Industries, Eastman Kodak c o . or E;. &- K . Laboratories (4-nitrobiphenyl and p-iiitroacetophenonc). LIelting points agreed with literature values. Preparatioii
April 20, 1961
INVESTIGATION OF SUBSTITUTED NITROBENZENE loxs
of p-nitrosonitrobenzene was done according t o Bamberger and Hubner.4 Melting point was 118' (lit. value,4 119"). p-Sitrophenylthiocyanate was prepared by the method of Muller6 with a melting point of 131" (lit. value,5 133"). Preparation of methyl p-nitrobenzenesulfonate was carried o u t according t o the procedure of Morgan and Cretcher,6 starting with p-nitrobenzenesulfonyl chloride obtained from K. & K. Laboratories. The crude product was recrystallized from ligroin-benzene and sublimed under reduced pressure. The white, crystalline sublimate melted at 91" (lit. value,7 91-92'), Polarographic Measurements.-Polarographic data were obtained using the same instrumentation described previously.2 Measurements of half-wave potential, El/%, and limiting current constant, I d , were made on approximately 1 miU acetonitrile solutions which were decimolar in tetra-n-propylammonium perchlorate for the supporting electrolyte. An aqueous saturated calomel reference electrode (s.c.e.) was used throughout. Drop-weight, m, and drop-time, t , were measured at a potential on the limiting current plateau. Typical values were m = 1.34 mg. sec.? and t = 4.44 sec. a t -1.1 v . with a mercury column of 50 cm . Maximum, not average, current values were measured, and thus cell resistances for "iR" correction were determined at maximum drop size. \There multiple waves occurred the tabulated I d value refers only t o current resulting from the new wave, and not t o total current as measured from zero. Free Radical Generation.-The technique of electrochemical generation of anion free radicals in a n electrochemical cell placed directly within (intra muros) the resonance cavity of a n ESR spectrometer has been described earlier.*,3 The potential for generation of a particular anion radical (ordinarily 0.1 volt more negative than E l / %was ) chosen on the basis of the polarographic data. An aqueous s.c.e. with a n agar bridge was used rather than the soft glass Pyrex Perley seal previously described.2 Acetonitrile solutions approximately 1 m M in parent molecule and decimolar in tetra-n-propylammonium perchlorate were used in the free radical studies. E.s.r. Spectrometer.-The e.s.r. spectrometer employed has been d e ~ c r i b e d as , ~ ~well ~ as the technique of placing absolute field calibration marks on the e.s.r. spectrum as i t is recorded. The sweep was found t o be linear within the accuracy of measurement of the field calibration marks. E.s.r. Results.-In all para substituted compounds investigated, except for the p-iodo-, p-nitroso- and p-thiocyanato-nitrobenzenes, a n electro spin resonance was observed when a solution of the compound was electroreduced at the appropriate potential within the microwave cavity. The character of the spectra differs widely. Many spectra exhibit a pronounced three-fold symmetry; t h a t is, the same sub-pattern appears at the center of the resonance and also is displaced by a n equal interval toward both high and low fields where i t occurs with intensities equal t o the central sub-pattern. This interval has been assumed to result from the hyperfine interaction of the electron with the N14nuclear moment of the nitro group (nuclear spin, I , of one). The three-fold symmetry of the hyperfine pattern may be obscured if the nitrogen hyperfine interaction is not dominant. I n this case, the nitrogen coupling constant must be derived by fitting the positions and intensities of the observed hyperfine components t o a set of coupling constants with all interacting nuclei of the molecule. For this assignment, advdntage is taken of the equivalence of nuclei which is a consequence of the symmetry of the parent molecule. Isotopic substitution was not undertaken in this study. As a result, the assignment of certain coupling constants t o particular molecular positions is not unambiguous. From a study of the relative intensities and spacings of the hyperfine pattern, however, the number of equivalent nuclei and the spin of each nucleus associated with a particular coupling constant could be determined in each case. Because of the good resolution and large number of lines observed, i t is thought t h a t these assignments are unique. When interaction of the odd-electron with two nuclei of I = 1 occurs, the larger coupling constant is identified with the W4 of the (4) E. Bamberger and R. Hdbner, BEY., 36,3803 (1903). ( 5 ) H . A. Muller, Chem. Z e n l v . , 77, 1588 (1906). (6) M . S. Morgan and L. H . Cretcher, J . A m . C h z m . S O L . 70, , 375 (1948). ( 7 ) R . K u h n and €1. 1%'. Ruelius, C h e m B e y . , 83, 420 (1030).
1S53
nitro group. The coupling constants in Table I were obtained on the basis of this analysis. E.s.r. spectra in Figs. 1-3 are recorded as the derivative of the absorption, dx"/dH, vs. magnetic field, H , which increases t o the right along the horizontal axis. -. _ _ ~ - . b . _
-
d
-
_-
'
I
I
Fig. 1.-E.s.r. spectrum of p-nitrotoluene anion; constructed absorption spectrum calculated from data of Table I. Total width between extreme hyperfine components 41.62 =k 0.05 gauss.
Fig 2 -E s r spectrum of p-fluoronitrobenzene anion; constructed absorption spectrum calculated from data of Table I. Total width between extreme hyperfine components 39 37 i 0.05 gauss.
_
~
_
~
-*.- -
-
-
_ _ _ _d
-----
__-___-a_ 4 -
~
--
Fig. 3.-E.s.r. spectrum of p-nitrobenzaldehyde anion; constructed absorption spectrum calculated from data of Table I. Total width between extreme hyperfine components 19.24 0.05 gauss. p-Aminonitrobenzene .-The spectrum of the negative ion p-aminonitrobenzene exhibits three-fold symmetry in the hyperfine structure. This major splitting of 12.18 gauss is assigned to the W4 nuclear moment of the nitro group. A very good fit of the observed spectrum is obtained by assumal = 1.12 gauss where al,a2 and u3 ing u2 = u3 = u4 = are each a n interaction with two equivalent protons and a4 is an interaction with one spin of 1 nuclear moment. This assignment predicts multiplets of 13 lines with intensity ratios 1:5: 11:16:21:27:30:27:, etc. Observed intensity ratios are 1.4:6.0:11.0:15.4:19.7:26.3:30.0:26.3, etc. p-Methoxynitr0benzene.-The three-fold symmetry in the e.s.r. spectrum of p-methoxynitrobenzene anion again leads to the nitrogen coupling constant which now must refer unequivocally t o the nitro group. The spectrum is easily interpreted since the three other coupling constants are sufficiently different in magnitude to be immediately evident upon inspection. p-Nitrotoluene .-The p-nitrotoluene anion exemplifies a n e.s.r. spectrum, Fig. 1, lacking the immediately obvious three-fold symmetry. Below the experimental spectrum is a n absorption spectrum synthesized from the coupling constants reported in Table I. p-Fluoronitrobenzene .-Fig. 2 shows the spectrum of pfluoronitrobenzene anion with a synthesized absorption spectrum drawn to scale below. The spectrum was based on (if
1854
AUGUSTH. MAKIAND DAVIDH. GESKE
Vol. 83
actions lead t o quintets with intensity ratios 1:3:4:3:1 in agreement with the spectral data. OBSERVEDCOUPLING CONSTANTS(IN GAUSS) FOR THE p-Nitroacetophenone.-The three-fold pattern in the NEGATIVEIONS OF VARIOUS para SUBSTITUTED NITRO- spectrum of p-nitroacetophenone anion leads to the coupling constant of the N" nucleus of the nitro group. The detailed BENZENES structure is consistent with two more coupling constants, a Coupling constant (absolute value) larger one with two equivalent protons and a smaller one Substituent asa a1 a1 a2 a4 with five equivalent protons. The coupling constant of the methyl group protons, consequently, must be nearly the -NHz 12.18 3.36b 1.12* 1.12* 1.12" same as t h a t of one pair of ring protons. 1.11' 0.30d ... -0CHa 11.57 3.43* Methyl p-Nitrobenzenesulfonate.-The hyperfine struc-CHI 10.79 3.9Sd 3.39* l.llb ... ture of the methyl p-nitrobenzenesulfonate anion is qualita--F 10.76 8.41' 3.56b 1.16* ... tively similar to that of the p-chloronitrobenzene anion. The coupling constants are given in Table I . The line -Hf 10.32 3.97" 3.3gb 1.09' ... width is less than t h a t of the p-chloronitrobenzene anion. -Ce" 9.84 3.60b ... ... ... No evidence of hyperfine interaction with the methyl group -c1 9.83 3.46b 1.17b ... ... protons was observed. -Br 9.iO 3.43* 1.15' ... ... p-Nitrobenza1dehyde.-The hyperfine structure of t h e e .s.r. spectrum of the p-nitrobenzaldehyde anion which ap-CONHz 8.37 3.20b 0.98' , .. ... pears in Fig. 3 cannot be explained on the assumption t h a t -COOCHI 7.73 3.11b ... ... ... the ring protons are equivalent in pairs. The synthesized -CN 7.15 3.12b 0.76' 0.76" ... absorption spectrum which appears below the observed -COCHa 7.02 2.95' 0.66* 0.66d ... spectrum was calculated using the coupling constants given in Table I . The coupling constants a l , a2 and a3 each rep-SOsCHa 6.90 3.03* 0.64* ... ... resent a hyperfine interaction with a single proton, and a, -CHO 5.83 3.10e 2.37e 1.23O 0.44h represents a n interaction with two approximately equivalent -NO20 1.74 1.12A ,.. ... ... protons. Results of Polarographic Studies .-Polarographic data for Nitrogen coupling constant of the nitro group. * Insubstituted nitrobenzenes as given in Tables I1 and I11 were Interteraction with two equivalent spins of I = '/z. obtained for only a single concentration (ca. one m M ) for action with one spin of I = 1. Interaction with three each compound. Thus, this study does not represent the equivalent spins of I = 1 / 2 . e Interaction with one spin of usual systematic investigation in which the dependence of I = l/2. D a t a from ref. 2. 0 Data from ref. 3. Inlimiting current and half-wave potential on concentration is teraction with four equivalent spins of I = examined and the dependence of limiting current on height the coupling constants given in Table 1. coupling constant of the mercury column is established. Primary concern represents an interaction with a single nucleus of spin l/z was directed toward measurement (under identical condiOf half-wave potentials (estimated accuracy 415 whereas a2 and a3 each represent interaction with two tions) mv.8) and limiting current constants for the various oneequivalent nuclei of spin I / ~ . promsymmetry arguments al must be identified as the hyperfine coupling constant electron reduction reactions used t o generate the anion free radicals. KOattempt was made t o elucidate electrode reacthe nucleus. tions occurring a t more reducing potentials. 4-Nitrobiphenyl.-The e.s.r. spectrum of the 4-nitrobiA necessary ( b u t not sufficient) condition for the "polarophenyl anion is Characterized by the three-fo1d symmetry allows the identification of the nitrogen coupling con- graphic reversibility" of a n electrode reaction is t h a t the has a Of -" m v *(at 250) for a quantity, Ea/a stant and consists of three broad triplets, each with intensity one-electron reaction.9 Values of this quantity for the ratios 1:2: 1. purther resolution of the spectrum could not probably because of the complex hyperfine in- various reduction waves are given in Table 11. Uncertainty be of measurement is approximatcly =t5mv. With t h e exteraction with the p-phenyl ring. p-Chloronitrobenzene.-The three-fold symmetry in the ception of the iodo- and thiocyanato- compounds (neither of spectrum of the p-chloronitrobenzene anion permits identi- which showed e.s.r.) the first reduction wave for all compounds in Table I1 shows reasonable agreement with the fication of the nitrogen coupling constant. interaction criterion stated above. Thus, half-wave potentials for these with two pairs of equivalent protons is also evident. No structure resulting from the chlorine nuclear moment is compounds represent thermodynamic standard reduction evident, b u t its presence apparently contributes t o the Polarographic behavior of most of the para substituted width of the lines. p-Bromoniwobenzene ,-The spectrumof the p-bromoninitrobenzenes was similar to t h a t of nitrobenzenez in that the rcversible one-electron was followed a t more negative trobenzene anion is qualitatively the Same as t h a t of the p potentials by another reduction wave. The latter wave was chloronitrobenzene anion. frequently distorted by a maximum. The second wave of p-Nitrobenzamide.-The spectrum of the p-nitrobenzthree of the more easily reduced derivatives (-COCHs, amide anion reveals a three-fold symmetry which can be attributed t o a hyperfine interaction with one of the iV4 -CHO and -NOz) appears to represent a second one-electron nuclei of the radical. The spectrum is qualitatively similar reduction. Although no study of the e.s.r. of mete substituted nitrot o t h a t of the p-chloronitrobenzene anion except the lines benzene anion radicals (except for dinitrobenzene3) was are much broader. The broadness of the lines is possibly caused by unresolved hyperfine interaction with the amide undertaken at this time, examination of the polarographic behavior of some members of the meta series (Table 111) group. The resolved hyperfine interactions are interpreted seemed desirable. I n all of the meta compounds examined, a as arising from the W4 nucleus of the nitro group and the four ring protons which, as far as the resolution will allow, reversible one-electron reduction was found, followed a t more negative potentials by further reduction. These data seem to be equivalent in pairs. contradict an earlier report" t h a t various meta Substituted Methyl p-Nitrobenzoate.-Three very broad triplets, each nitrobenzenes in acetonitrile undergo a four-electron reducof the intensity ratio 1 :2: 1, characterize the e.s.r. spectrum of the methyl p-nitrobenzoate anion. The nitrogen coupling tion. This mistaken conclusion was based on comparison constant and t h a t of two approximately equivalent ring protons can be identified. It is unlikely that broadening of ( 8 ) For t h e several cases where replicate determinations of E l / % the lines is caused by the methyl group protons. were made t h e reproducibility was in fact better t h a n f 5 mv. p-Nitrobenzonitri1e.-The large NI4 coupling constant in (9) J. Tomes, Collecfion Czechoslov. Chem. Commun., 9, 12, 81, 150 the e .s.r. spectrum of p-nitrobenzonitrile anion is obtained (1937). from the large three-fold splitting. T h e entire spectrum is (10) This statement assumes only t h a t t h e diffusion coefficients of well explained by assuming the coupling constants given in t h e oxidized and reduced species are equal and t h a t t h e half-wave potential is independent of t h e ionic strenHh of the solution u p t o 0.1 Table I. There are two coupling constants, each representM . If the latter assumption is inadequate, then the half-wave potening a hyperfine interaction with two equivalent ring protons. The smaller of these is approximately equal t o the small N14 tials represent formal potentials. (11) L. Holleck, R . Schindler and 0. Zohr, .\Tnff~rluissenschne~, 46, coupling constant which is assumed t o result from the nitro625 (1959). gen atom of the C N group. These latter hyperfine inter-
TABLE I
.;
INVESTIGATION OF SUBSTITUTED NITROBENZENE IONS
April 20, 1961
1855
TABLE I1 POLAROGRAPHIC DATAFOR SUBSTITUTED NITROBENZENES'' (0.1 M tetra-n-propylammonium perchlorate supporting electrolyte in acetonitrile a t 25.0') f 3 / 4
R
-0CHa -CH3 -H
-F -c,H~
-c1
1.358 2.0d 1.250 2.I d 1.203 2.06 1.147 1.9d 1.128 1.93d 1.103 1.5d 1.063 l.gd
1.050 1.49 1.050
-Br
-I"
3.4 14 4.0 24 4.0 10 4.1 7.5 4.3 9.0 4.4 12 4.2 9 4.0 11 6.2f 9
R
8
so
58
-0.172
-0.38
-CONHz
60
-
- .16
-SCiY:"
-
-COOCH3
E8/2
Id'
-NH2
gara Substituents
-
.111
56 112 56
- .129 0
0
-COCHa
57
0.056
0.17
-CN
.15
-CHO
63 61
.238
.27
-SOsCHz
56
,265
.26
-NOzh
-E
Id'
1.014 1.6d 0.96 1.30 0.947 1.4d 0.925 1.42 0.875 1.4gd 0.863 1.34 0.847 1.42 0.69 .89 ,525 .86
3.5 11
4.2 12 4.2 9 3.9 4.1 4.0 9 3.7 3.5 3.4 3.8 5.0 3.6 3.8 3.5
106 65
0.385
60 55 62
.502
0.40
,674
.63
59 70 58 43 53 58 60 65
.50
.778
.73
-NOe 94 1.8 Half-wave potentials in volts v e n w aqueous saturated calomel electrode. I d represents the quantity id/iiz%t'/eC, where m is the drop-weight in mg. set.-', t is t h e drop-time in sec., id is limiting current in pamp. and C is concentration expressed in millimolar dimensions. Wave distorted Expressed in millivolts; uncertainty of measurement is ca. 5 mv. by a maximum. e No electron spin resonance observed. f Current is not diffusion-controlled. a Wave is badly defined and distorted by a maximum. Data from ref. 3. 4 A "primary" un value, ref. 21; unless otherwise indicated, other un values are "secondary" un values.
*
et al., as contrasted to our use of tetra-n-propylammonium TABLE I11 perchlorate may be of some significance in comparison of the POLAROGRAPHIC DATAFOR SUBSTITUTED NITROBENZENES' two sets of data. (0.1 M tetra-n-propylammonium perchlorate supporting elecDiscussion trolyte in acetonitrile a t 25.0')
Polarography.-The
meta-Substituents R 2"-
-CHs -0CHs
- E1/aR 1.208 1.90 1.180 1.9d 1.147 1.80
-COOCHz
1.044 1.7d 2.3 1.042 1.65d 2.06 1.016 1.6d 1.95 0.938 1.5d 0.898 1.25 2.019
El/r
-
Idb
Ea/,C
a*
d
3.6 12 4.0 11 4.8 15 4.0 10
58
-0.O3gL
-0.14
58
-
-
7
.069"
.07
half-wave potential (for the one-electron reduction) and the nature of the substituent. One such relation is expressed in the HammettI3 "sigmarho" equation log
k
-=
kQ
pa
61
.076'
.06
58
.3211
.36 stants for substituted (meta or para) and unsubsti-
where k and ko represent rate or equilibriuni contuted compounds, respectively. p is a constant characteristic of the reaction series, and u is a con.34 stant peculiar to the substituent. The correlations of electrochemical data by Brockman and Pearson,14 Zuman16 and Fox, et aZ.,lB are typical of those made using the form AEl/Z = E I / ~ R- EI/,H = P U (31 where and EllaHare the half-wave potentials .62 for the substituted and unsubstituted compounds. Note that in eq. 3, pis expressed in volts. The intrin.i 0 sic difficulties of attempting El/, - u correlations
3.8 60 .376"' 10 11 -CHO 3.9 59 .381k 6.3 6.0 4.1 58 CN .613' 11 -NOzh 4.2 61 . 7 10" 3.6 63 10 j Footnotes a s in Table 11. See footnote d, T a b k , I V . A priRecalculated meta u" values, see footnote 22. mary'' urnvalue, ref. 21. -COCHa
polarographic data in Table
I1 can be examined for a possible relation between
of wave heights with p-benzoquinone. For only one of the seven compounds investigated by both Holleck, et al., and in this study is the agreement of half-wave potentials better t h a n 60 mv. The use of sodium iodide or tetraethylammonium perchlorate a s supporting electrolyte'* by Holleck,
(12) Unfortunately the table of half-wave potentials in ref. 11 does not specify t o which of the two supporting electrolytes the data refer. (13) L. P. Hammett, "Physical Organic Chemistry," McGraw-Hill Book Co., Inc., New York, N. Y., 1940, p. 186. (14) R. W.Brockman and D. E. Pearson, J . A m . Chcm. SOC.,74, 4128 (1952). (15) P. Zuman, Chcm. Lisly. 47, 1234 (1953); Collection Csechoslm. Chem. Commvn., 19, 599 (1954). (16) I. R. Fox, R. W. Taft, Jr., and J. M. Schempf, unpublished results as referred t o by R. W. Taft. Jr., and I. C. Lewis, J . A m . Chem. Soc., 81, 5343 (1959).
1S56
AUGUSTH. MIKIAND DAVID H . GESKE
when the electrode reactions are irreversible have been Unfortunately, many organic reductions and oxidations are irreversible and hence half-wave potentials cannot have thermodynamic significance. I n this context, perusing the data in Table I1 is somewhat more encouraging. Except for the para iodo- and para thiocyanato-derivatives all of the electrode reactions for the initial oneelectron reduction fulfill to a reasonable extent the previously stated criterion for reversibility. Observation of a free-radical reduction product, furthermore, establishes the course of the reaction beyond question. Since the half-wave potentials for the one-electron reduction are an excellent approximationlo to the standard reduction potential for the reaction written in eq. l , they are properly valid data for correlation with u values. Examination of a relation between AE?I, and u may be done graphically merely by linear plotting of one against the other using known u value^.'^^^^ Separation of resonance and inductive effects has also been attempted.16 It seemed most satisfying to us to utilize the recent reevaluation of the Hammett relation done by van Bekkum, 'l'erkade and Wepster,*O hereafter referred to as BVW. Inasmuch as the treatment b y these workers may not be generally known a recapitulation here seems desirable. Tn effect, BVLI- presented the case for a continuously variable scale of para u values in those cases in which the substituent undergoes resonance interaction with the reaction center. Hammetti3 had originally suggested a "duality" of u constants in certain instances. BVLV designated the normal constant u values as un and defined a group of ten "primary" un values on the basis of the usual convention t h a t p = 1 for dissociation of benzoic acids in water a t 25'. These primary values,?' established by precision measurements of acid dissociation constants, were used to calculate new p values for the various reaction series in the literature. The recalculated p values were obtained with better correlation coefficients and standard deviations than previously attained in the work of HammettI3 and Jaffe.lg Using the recalculated p values new u values were then calculated for the other substituents in the various reaction series. Inspection of these data led t o the conclusion that there is "little cause to suspect any meta substituent of hxving inconstant u values" .2n.32 On the other hand, for a given para substituent recalculated u values extended over rather wide ranges. BIT\: then selected those reaction series in which the para substituent could not enter into (17) 1'. Berzins and P. Delahay, J . A n i . Chem. Soc., 75, 8716 (19.53).
(18) W,H . Reinmuth, I,. B. Rogers and L. E. I. Hummelstedt, t b i d . , 81, 2947 (1959). fl9) H. H . Jaffe, C h e m Reus., 5 3 , 191 (1953). ( 2 0 ) H. \.an Bekkum, P. E. Verkade and B. M. Wepster, Rev. lrav. Olitn., 7 8 , 81,j (1959). 121) D a t a for two fiova substituents were included (the remainder being inela substituents) and were used in later calculations of p only when reionance interaction "could be considered out of the question." ( 2 2 ) T h e mean of the recalculated c values for each mefa substituent, which was not included in t h e definitive "primary" group, however, ivzs gcnernlly slightly different from the usual value.'h's
Vol.
sx
resonance interaction with the reaction center, and upon finding a relatively narrow range of such u values, designated the mean u value as a "secondary an value." Thus the un is intended to represent the substituent constant when no para resonance interaction occurs between substituent and reaction center. The extent of change of resonance interaction, AAF,, is given quantitatively'O as .LAF,> = -2.303RT p (u - u")
(4;
where u is calculated from eq. 3 using the experimental AEl:, and p is established by the meta substituents. Essentially the same conclusions regarding inconstancy of pala sigma values were reached simultaneously by Taft and Lewis.?3 A scale of uo values, which in principle should be identical with the BL"CT' un quantity, was developedz4; values of u'' are included in Tables I1 and 111. The quantitative expression for AF,,, eq. 4, was given only by BVm'. Data in Table I11 for mcta substituents were used to establish a least-squares correlation between AEl/, and the various sigma values. From the summary of correlations, Table IV, it is clear that there is good correlation between A E I / ~values and all three sets of sigma values. The line in Fig. 4 represents the equation LEl/z = 0 . 3 7 6 ~ "- 0.021
(5)
The standard derivation of un as assigned by BVT;I' is shown in Fig. 4 by the length of the horizontal line. Primary un values are plotted as points. TABLE IV EVALUATION O F CORRELATIOS E Q U A T I O S mete HALF-TAVE POTESTIAL AXD SIGMA I-ALCES LEI/, = m X - b
LEAST-SQUARES FOR
S-
in
b
i1/1
~ . h
i s c
0.991 0.0137 0.876: 0 . 0 2 1 9'' ,997 ,0106 ,019 8" ,357 ,0131 ,009 9 ,991 ,347 u (Harnmett) A-urnber of esperitiieiital points used. * Corrclatiori coefficient. Standard deviation of a single point, sigma value assumed invariant, liecalculated un values for metaaldehyde of 0.464, 0.217,0.409 and 0.416 are included in Table IX, ref. 20, but a mean value was not included in surnmary Table 111 by B\-\V, T h u s we have arbitrarily chosen to erripl(iy the older Ilar~inicttx--aluel3 of 0.381, which is in fact quite close t o tlic iiieitii of tlie f(iur values given in t h e preceding setitencc. ' A41tiehydeomitted from correlation. u'~ ( B \ X 7 ) u' (Taft) '2
772(3/(1
Al> a’ - a”1 we can neglect the last stituted nitrobenzene negative ions by controlled term of eq. 12 for all electron resonance transitions. potential electrolysis, studied the electron spin (Note that the last term is always zero if I’ = I”.) resonance of the ions and also obtained polaroEquation 12 thus leads to the following hyperfine graphic data. Correlation oi the reversible polaroenergies for the allowed electron resonance transi- graphic half-wave potentials with 0” values of van tions Bekkum, Verkade and XTepster2Oas well as with uo values of Taft and L e w P was made for the e = 0, 0 , 5 1/2(a‘ a”) 14 paru substituted nitrobenzenes as well as for is The appearance of the hyperfine structure under meta substituted ones. Deviations of the halfthe condition h ~ , > > ~ u ’ - a ” ~is the same as that of wave potentials of @-dinitrobenzene and p-nitrotwo equivalent protons each with a coupling con- aniline are interpreted as arising from resonance stant 1/2(a’ a”). The effect of increasing effects and resonance energies are estimated. a’ - a”, is t o cause a mixing of the singlet and From a comparison of the nitrogen coupling triplet components of the state with mI = 0, constant of the nitro group with the polarographic which breaks down some of the selection rules used half-wave potential of the pura substituted nitroin the above treatment. The transition between benzenes, a decrease in the nitrogen coupling conthe two extreme types of hyperfine structure would stant was found t o be concomitant with an inoccur when a’ - a” A hv,. crease in the ease of reducing the parent molecule. Observation of three different proton coupling The ortho and meta (to the nitro group) proton constants, each attributable t o the interaction coupling constants of the series of para substituted of the electron with a spin of 1 / 2 nucleus in p - nitrobenzenes were found t o be correlated by the nitrobenzaldchyde negative ion, is a strong argu- relationship ment for the presence of hindered rotation of the no + a,,,= - 2 30 gauss aldehyde group about the carbon-carbon bond. The two largest of these coupling constants, 3.10 when certain assumptions regarding the signs and gauss and 2.37 gauss, are assigned to the two assignment of the coupling constants were made. protons ortho to the nitro group and the remaining This relationship implies that the total spin-magcoupling constant of 1.23 gauss to the aldehyde netization in these 7r-orbitals of para substituted proton. An upper limit of (26lh)AllI can be set nitrobenzenes is independent of the substituent Evidence of hindered rotation of the aldehyde on the rotational frequency, v,, of the aldehyde group, where AH is the separation in gauss of the group of p-nitrobenzaldehj-de anion about the two ortho proton coupling constants. From this carbon-carbon bond is provided by an analysis of the hyperfine pattern of the resonance. An we find upper limit to the rotation frequency is 2.S X r,
