J. Phys. Ckem. 1993,97, 10585-10588
10585
An Isotopic and EPR Probe into Hyperconjugative and Induction Effects in pXylene Cheryl D. Stevenson,’ Eugene P. Wagner 11, and Richard C. Reiter Department of Chemistry, Illinois State University, Normal, Illinois 61761 Received: March 30, 1993; In Final Form: July 27, 1993“
EPR studies show that the anion radicals of a,a,a-d3-p-xylene and a,a’-ds-p-xylene (CD3-CaHd-CD3’-) exhibit hyperfine coupling constants for the ring protons that are 12 and 23 m G larger than those of the p-xylene anion radical itself. This has been attributed to a further lifting of the degeneracy of the antibonding molecular orbitals due to the replacement of the protons with deuteriums to an extent of 29 J/mol per deuterium. These results seem to support the conclusion that both inductive and hyperconjugative effects are responsible for the lifting of this degeneracy due to the methyl groups. Introduction Thirteen years prior to the turn of the century, Henry Armstrong’ first considered the mechanism for the action of the ortho-para and meta directing substituents on benzene. H e concluded that there is no “direct connection” between (for example) a methyl substituent and the para position “and the production of para derivatives must be regarded as the result of a kind of isomeric or intramolecular change.” About 5 years later, Crum Brown explained the situation in more contemporary terms. The Crum Brown rules are expressed in terms of the oxidizability of the protonated substituent. For example, the nitro group is meta directing since HNO2 is oxidizable to H N 0 3 . It was later realized that this effect is due to a perturbation, mediated by hyperconjugation and induction, upon the electron distribution in the T system due to the methyl moiety. More than a half a century later EPR spectroscopy revealed some details regarding the methyl group’s presence, causing an alteration of the electron distribution in the benzene system,2and these EPR observations of substituted benzene anion radicals yielded a picture that is quite consistent with H M O theory.’-5 A pictorial representation of the antibonding MO’s for a benzene system with two electron-releasing groups in the para positions is shown in Figure 1. The expression for the total wave function of the odd electron in the linear combination of the two doubly degenerate LCAO wave functions is given by eq 1.2.5 The antisymmetric wave function (qa)is unperturbed by the presence of the electron-releasing para substituents due to the presence of a nodal plane at the substituted position. When T-T spin polarization is neglected, Cs2 is related to the magnitude of the splitting between the 9,and 9,wave functions of the degenerate molecular orbitals according to Boltzmann’s distribution andcan be calculated according to eq 2.2 The para protons
being absent, represents the effect of the spin density at the sites of substitution as calculated using the McConnell relationship, aH = Qpe3 Using the Q value for the benzene anion radical (-22.5 G),* aH(para) = (Q - 4a~(”,,))/2. Using the ring proton coupling constant measured in this study, aH(ring) = -5.388, and thecalculated aH(pra) = -0.474 for thep-xylene anion radical, C? = 3(0.474)/[2(0.474) + 4(5.388)] = 1/[1 + This simple calculation leads to a AE of 437 1 J/mol at 195 K, which is very close to the empirical value found by Hobey (421 1 J/mol at 203 K).3 If vibronic interactions are included, as carried out Abstract published in Advunce ACS Absrrucrs. September 15, 1993.
0022-3654/93/2097-10585$04.00/0
Figure 1. Two antibonding degenerate orbitals of benzene and their splitting due to the presence of two electron-releasinggroups in the para positions (indicated by the arrows).
by Alper and Silbey for the toluene anion radical: AE is found to be about 13% smaller. Also, M. T. Jones has pointed out the indirect nature of these determinations and that the lack of an observable thermal dependence of the hyperfine splitting indicates that aH(para) calculated in this way is too large and leads to a value for AE that is too smalL6 His direct measurements of the methyl proton splitting show it to be -0.080 G. Given that Q(cH,)is 34.2 G,6 this corresponds to a ground-state spin density in the para position ( p p ( ~ )of) -2.34 X l e 3 . Given the best calculated value for the excited state ( p p ( ~=) 0.380): he found that AE must be at least 1500 cm-’ (17 900 J/mol) to account for the lack of thermal dependence of the averaged spin density in the ith position (pi), eq 3. This argument, of course, is not true if AE itself is thermally dependent. (3) The sensitivity of the spin distribution to exceedingly small perturbations is so great in the benzene system that a difference intheparaproton(3.41 G) and themetaproton(3.92G)splittings can be observed when one ring hydrogen is replaced by a single deuterium. The AE in CsHsD is 209 f 12.5 J/mol.s In the cases of xylene and toluene, hyperconjugation and the inductive effect (which are well understood in terms of molecular orbital theory)’ provide a mechanism by which the hyperfine splitting from the methyl protons can be observed in an EPR spectrum. While there is some disagreement as to the relative magnitude of these two effects, it does appear that the inclusion of the inductive effect is necessary for obtaining the correct ordering of qaand q8.This is thecase, since perturbation theory indicates that 9sis actually lowered in energy via the hyperconjugative effect.’a However, the EPR spectra clearly show that q,is higher in energy than is qa. 0 1993 American Chemical Society
10586 The Journal of Physical Chemistry, Vol. 97, No. 41, 1993
Hyperconjugation and the inductive effect account for the ring activation with respect to electrophilic attack as well as the observed splitting of the degeneracy of the antibonding MO's. Since hyperconjugation involves an overlap of the methyl group orbitals and those of the ring, it should be mass dependent due to zero-point energy effects, Thus, we were motivated to use a potential perturbation upon A electron distribution caused by deuteriation of the methyl groups to gain some insight into the effects of hyperconjugation and the inductive effect. The extreme sensitivity of the electron distribution in the benzene system to substitution makes the EPR spectra highly responsive to isotopic substitution. Thus, EPR is a more sensitive probe into isotopic substitution effects than are kinetic or thermodynamic ring activation measurements. The fact that any imprecision in a H measurements has a large effect on the calculated A E is a drawback. However, in this study, we are interested in the perturbation upon AE,or AAE, due to isotopic substitution. This is dependent only upon a difference in coupling constants between two different samples. We can obtain this parameter quite precisely with the dual cavity technique.8 There are various methods for obtaining AE from EPR data at various levels of sophistication, such as inclusion of vibronic interactions. However, as our interest is in AAE, we need only utilize the relatively simple calculation outlined above. The current study relates alterations in A electron distribution to isotopic substitutions in a methyl groups and resultant changes in hyperconjugation. However, it should be recalled that the most notable result in the early EPR studies of the solution phase alkyl radicals was that the methyl (0 proton) coupling constants are actually larger than those of the a protons, structures IA and IB. The relative magnitudes of the a and fl couplings are, of course, explained in terms of hyperconj~gation.~bThe anion radical of xylene, on the other hand, exhibits a very small splitting for the methyl protons (-80 mG a t 243 K).6
Stevenson et al.
Figure 2. (upper)EPR spectrum of the potassium anion radical ofp-xylene (CH,-CsH4-CH3'-) in DME containing 18-crown-6at 195 K. (lower) Computer simulationgenerated using a mixture of two anion radicals in equal concentration. Both radicals have coupling constants of 5.388 G (4 H's) and 0.085 G (6 H's), but they havedifferent potassium splittings (0.05 and 0.09 G).
24.68 G
H3C,
,CH3 22.38~ 2 6 . 6 7 ~ ;22.11
IA
G
*CH*-CH3 IB
The methyl groups yield small hyperfine couplings, but their electron-releasing nature significantly reduces the thermodynamic stability of this solvated ion (CH3-CsH4-CH3*-).9 Fortunately, the solvated anion radical can be stabilized via the presence of 18-crown-6 (18C6).'0 The addition of the crown ether is also known to eliminate any extra hyperfine splittings due to tight ion association between the metal cation and the nonpolar hydrocarbon anion radical.]' On the basis of these previous studies with the xylene anion radical, we selected a solvent system consisting of dimethoxyethane (DME) containing 18C6. Potassium ion was chosen as the counterion, as the potassium ion does not lead to observable metal hyperfine splitting with hydrocarbon anion radicals in DME or in tetrahydrofuran.
Results and Discussion When a solution of approximately 100 mg of p-xylene in approximately 4 mL of DME and a slightly deficient molar amount of 18-crown-6 is exposed to a freshly distilled potassium metal mirror under high vacuum at -78 "C,a solution of the anion radical ofp-xylene is formed. Dilution of this sample with more DME yields a solution whose EPR analysis at 195 K surprisingly indicates that the anion radical exists in two different states of ion association, both with observable metal hyperfine splittings (Figure 2). To confirm this analysis the same experiment was carried out with a,a'-d6-p-xylene (CDs-C&-CDs); as can be seen in Figure 3, the complication of the small methyl proton splitting is eliminated, and the two potassium splittings (aK = 0.05 and 0.09 G) are further exposed. These results suggest
Figure 3. (upper) EPR spectrum of potassium CD&H&D3'in DME containing 18-crown-6at195K. (lower)Computer simulation generated using a mixture of two anion radicalsin equal concentration. Both radicals have coupling constants of 5.411 G (4 H's) and 0.013 G (6 Ds), but they have different potassium splittings (0.05 and 0.09 G).
the existence of an equilibrium that is slow on the EPR time scale, reaction 4. This is the only example known to these authors
(4)
of a metal hyperfine splitting in the presence of crown ether. Initially, it was thought that perhaps some of the K+ was not complexed by the crown, and it was this uncomplexed ion that resulted in the metal splitting. However, increasing the 18C6 concentration by an order of magnitude does not alter the results, and carrying out the reduction in the absence of crown results in a single potassium splitting of 0.105 G (Figure 4). Evidently, localization of charge in the xylene anion radical relative to that of benzene (which lacks the electron-releasing groups) induces
Hyperconjugative and Induction Effects in p-Xylene
The Journal of Physical Chemistry, Vol. 97, No. 41, 1993 10587
value used for U H ( ~ ~ When ~ ~ ) . U H ( ~ =~ 0.474, ~ ) the values for AAEare found to beO.O,90, and 176 J/mol for CHs-C&-CH3*-, CD3-C&-CH3*-, and CD3-C&I4-CD3'-, respectively (Table I). An analogous calculation carried out using Jones' model, eq 3, can now be attempted. The fact that the ring proton hyperfine splittings have increased by 23 mG when CH3-C6H4-CH3*- is replaced by CD3-C&-CD3*means that pm has increased has increased by 0.023/22.5 = 0.001 02, and the spin density in the para positions has decreased by twice that value (0.002 04). Thus, pm is 0.25 (0.00204 0.00234)/2 = +0.25219, and that in the para position (pp) is -0.00204 - 0.00234 = -0.00438. Now, the question is, what value for AE would be needed to give the values for pm and pp for CD3-C6H4-CD3*-? No value will work, since neither pm nor pp lies between limiting values of pm(0),p p ( ~and ) p m ( ~ )This . suggests that substitution of deuterium for hydrogen in the methyl groups represents a bigger perturbation than this model can accommodate. Perhaps this is "due to a difference in hyperconjugation and/or induction for hydrogen vis-a-vis deuterium" (a reviewer's wording). In general, as the value used for AE increases ( u H ( ~ decreases), the value for AAE calculated from the data also increases (Table I). However, this large deuterium isotope effect constitutes too large of a perturbation to permit application of Jones' model. Theimportant point, however, is that AAE increases with deuterium substitution on the methyl groups. We must keep in mind that there is a splitting from the counterion and thus some drainage of spin density to the K+ ion. This may have an effect upon the calculated values for AE, but since this ion association is the same for all of the systems, its effect cancels out in the A A E values. As the substituents become more electron releasing in character, the spin density increases in the wave function, due to a greater AE.This trend is apparent from Hobey's) theoretical calculations of the expected AE's for a series of substituted benzenes. Substitution of the methyl hydrogens with deuteriums has an effect upon A E qualitatively similar to the effect of increasing the electron-releasing nature of the substituent(s). If we consider each substitution of a deuterium at an equivalent site to have an equivalent effect as predicted by Bigeleisen,12then each deuterium results in a splitting of 29 Jlmol. When very small values are used for U H ( ~ ~the ~ ~agreement ) , with Bigeleisen's additivity rule is lost. If Hobey's model is correct, the effect of deuteriation of the methyl groups of CH3-C&-CH3*is attenuated by a factor of about 7 compared to the effect of deuterium substitution directly upon the benzene anion radical.
+
Figure 4. EPR spectrum of potassium CD,-C6H4-CD3*- in DME and no 18-crown-6at 195 K. (lower) Computer simulation generated using a single anion radical with coupling constants of 5.411 G (4 H's), 0.013 G (6 Ds), and 0.105 G (1 K+).
TABLE I: Perturbations upon the Ring Proton Coupling Constant and upon the Splitting of the Degeneracy of the Antibonding Molecular Orbitals as a Function of the Assumed Para Proton Coupling Constant AAE (J/mol) Aa~(ring) anion radical (mG) U H ( ~ =~ 0.474 ~ ) G U H ( ~ =~ 0.085 ~ ) G CH3-CsHd-CH30.0 0.0 0.0 90 543 CD~-C~H~-CHB-12 1 176 1273 CD3-C6Hd-CD323 2
**
the K+ ion to effectively protrude from the crown envelope and make intimate contact with CH3-C6H4-CH3'- as depicted in structure 11. H3C+CH3
n
+
II
A sample of CH3-C6H4-CH3*- was placed in the cooled back cavity, and a sample of the perylene cation radical (as a standard) was placed in the front cavity of a dual cavity EPR spectrometer; each spectrum was recorded. The directly observable ring proton coupling constant for CH3-C6H4-CH3*- has a value of 5.388 G, corresponding to a AE of 437 1 J/mol within the context of Hoby's model. The experiment was repeated with CD3-C&h-CD3'- in the back cavity. Thirty such measurements involving several independently prepared samples revealed that the ring proton splitting is 23 f 2 mG larger in the hexadeuteriated system. The same procedure was repeated with the anion radical of cu,a,a-d3-p-xylene (CD3-C6H4-CH3). The ring aH for CD3-CsH4-CH3'- proved to be 12 & 2 mG larger than that observed for CH3-C6H4-CHj*- and 1 1 f 2 mG smaller than that observed for CD3-C&-CD3*(Table I). As in the case of the p-xylene anion radical, the potassium splittings are only partially resolved in CD3-C6H.+-CH3*-. This is due to the fact that they are partially obscured by the methyl proton splittings. Only the ring proton coupling constant is needed to obtain Cs2 and AE, which can be calculated as described in the Introduction. The values for APE are quite dependent upon the
Conclusions The presence of the methyl groups in the xylene anion radical simultaneously localizes the charge density relative to that in the benzene anion radical and splits the degeneracy of the antibonding molecular orbitals. The former effect greatly enhances the Coulombic attraction between the anion radical and the counterion, resulting in metal splittings of 0.05 and 0.09 G, even when crown ether is present in the DMEsolution. Theeffect of "pulling" the cation from the protection of the crown (structure 11) results in metal splittings that are quite similar to those found in the absence of the crown ether.*J3 This localization of charge density and consequent increase in the tendency to form ion pairs due to the presence of electron-releasing substituents has been previously observed in other anion radical systems.14J5 The splitting of the degeneracy of the molecular orbitals is due to a combination of inductive and hyperconjugation effects.'a Deuteriation of the methyl groups further increases this splitting. The hyperconjugative effect is a result of the interaction between the methyl antibonding orbital and the ring, and it lowers the energy of \k, in xylene and t o l ~ e n e .One ~ of the valence bond representations is shown in structure 111. Thus, hyperconjugation
~ ~ )
Stevenson et al.
10588 The Journal of Physical Chemistry, Vol. 97, No. 41, 1993 HH-C-H
Q
itself seems to reduce AE. This is why, as stated in the Introduction, consideration of the inductive effect is necessary for the proper ordering of 9sand Within the confines of the Born-Oppenheimer approximation, methyl group deuteriation should lower the magnitude of the hyperconjugative effect. This is the case since methyl group deuteriation lowers the zero-point energies of the methyl C-H bonds and, in turn, attenuates contributions from the methylring interaction (structure 111). This diminishes the hyperconjugative effect of lowering of 9,and leads to an increase in the relative contribution from the inductive effect, hence theobserved isotopic increase in AE.
Experimental Section Materials. The c~,ar,a-d3-p-xylenewas synthesized as described by VogelI6 using perdeuteriated dimethyl sulfate in place of its perprotiated analogue, reactions 5 and 6. The product was purified via preparatory gas chromatography. The other deuteriated xylenes were purchased from MSD Isotopes and used without further purification. Na
p-CH,.C&Cl-
p-CH,.C,H,Na
(5)
The DME was freshly distilled from pure oil-free potassium metal and kept over oil-free NaKz in an evacuated glass bulb. The purity of the alkali metals is important for avoidance of contamination of the solvent with benzene. EPRExperiments. Identical experiments werecarried out with p-xylene, p-CD3-C6H4-CH3, and p-CD3-C6H4-CD3. Samples of each anion radical were generated and analyzed on the same day. For each experiment a new glass-blown apparatus was constructed (Figure 5 ) . Approximately 100 mg of p-xylene was sealed in capillary tube, which was in turn placed in the apparatus at point A (Figure 5). An equal molar amount of 18-crown-6 was placed in the bottom of tube D, and then the apparatus was attached to the vacuum line. A molar excess of potassium metal was placed in tube B which was sealed a t point C. The apparatus was then evacuated, and the metal was distilled into tube D. Tube B was sealed from the apparatus a t point E. Approximately 4 mL of DME was distilled from NaK2 through the vacuum line into tube D. The apparatus was sealed and separated at point F. The anion radical was generated via exposure of the solution to the metal mirror upon breaking of the capillary tube by agitation. A small portion of this solution was then poured into the EPR sample tube, which was subsequently further cooled, resulting in dilution of the same to a point where good EPR resolution could be obtained. All the EPR spectra were recorded with a Bruker ER-200 EPR spectrometer equipped with an IBM variable temperature unit connected to the back cavity of a dual cavity system at 195 K. EPR experiments were carried out on all three anion radicals consecutively. For example, three spectra were recorded for d3p-xylene anion radical, then for the d6-p-xylene anion radical,
C .--_-_
Tube B +
E/ I
Potassium '18-crown-6
Figure 5. Apparatus usedfor the generation of thep-xyleneanion radicals. The ground glass joint was connected directly to the vacuum line.
and finally for thep-xylene anion radical. This entire procedure was completed five times with no systematic order to the sample recordings. This technique has proven to bevery applicable toward the measurement of very small coupling constant differences between nearly identical samples.* New anion radical samples were then prepared, and the entire set of EPR experiments were repeated. All spectra wererecordedandsavedon an IBMPC with EPRW are data collection-simulation software developed by P. Morse and R. Reiter.1' Using thecomputer displayed spectra, two field indicating cursors were placed on the midpoints of the two outer most packets of peaks. Theseparationof thecursors was displayed in data pixels which were converted to gauss using the known sweep width. Some of the spectra were then simulated to ensure the accuracy of this method of measurement. It is important to note that the differences observed in the ring proton hyperfine coupling constant are not dependent upon any error in the absolute magnitude of aH. All of the comparative measurements (more than 30 for each pair of samples) indicated that aH(Ang)for CD3-C6H4-CD3*- is greater than aHi(ring) for CD&&-CH3*-, which is in turn greater than U H ( ~ for ~ , ,CH&6H4-CH3'-. ~)
Acknowledgment. We thank the National Science Foundation (Grant CHE-9011801) for support of this work and Dr. Otis Rothenberger for helpful discussion. References and Notes (1) Armstrone. H. E. J. Chem. SOC.1987. 51. 258. (2j Bolton, J.'-d.; Carrington, A.; Forman, A.; Orgel, L. E. J . Am. Chem. SOC.1968, 90, 6275. (3) Hobey, W. D. J . Chem. Phys. 1965, 43, 2187. (4) Alper, J. S.; Silbey, R. J . Chem. Phys. 1970,52, 569. (5) . . (a) . . Lawler. R.; Fraenkel. G. K. J. Chem. Phvs. 1968.49. 1126. Ib) Stevenson, G. R.; Reidy, K. A.; Peters, S. J.; Reiter, R.C. J. Am. Chem. S&. 1989, 111, 6578. (6) Jones, M. T. J . Phys. Chem. 1978,82, 1138. (7) (a) Purins, D.; Karplus, M. J . Am. Chem. Soc. 1968,90,6275. (b)
Carrington, A.; McLachlan,A. D. Introduction to Magnetic Resonance;Harper & Row: New York, 1967; pp 83-85. (8) Stevenson, G. R.; Ballard, M. K.; Reiter, R. C. J . Org. Chem. 1991, 56, 4070. (9) Lawler, R. G.; Tabit, C. T. J . Am. Chem. SOC.1969, 91, 5671. (10) Weissman, S. I.; Komarynsky, M . A. J . Am. Chem. SOC.1975,97, 1589. ( 1 1 ) The alkali metal anion radicals of benzene, toluene, and xylene do
not exhibit metal splitting even in hydrocarbon solvents (see ref 10). (12) Bigeleisen, J. J. Chem. Phys. 1955, 23, 2264. (13) de Boer, E.; Colpa, J. P. J . Phys. Chem. 1967, 71, 21. (14) Stevenson, G. R.; Ocasio, I. J . Am. Chem. SOC.1976,98, 1976. (15) Stevenson, G. R.; Echegoyen, L. J . Phys. Chem. 1973, 77, 2339. (16) Vogel, A. Vogel's Textbook of Practical Organic Chemistry, 4th 4.; Longnard Group Ltd.: London, 1978; p IV, 2. (17) Morse 11, P. D.; Reiter, R.C. EWSIM, Scientific Software Services: Bloomington, IL, 1992.
