ESRINVESTIGATION OF METHYLATED ANTHRASEMIQUINONE RADICALS
71
An Electron Spin Resonance Investigation of the Anion Radicals 1,4-Dimethylanthmserniquinoneand 1,4,5,8-Tetramethylanthrasemiquinone1 by Richard H. Schlossel, David H. Geske,2&and Wilson M. Gulick, Jr.2b Department of Chemistry, Cornell University, Ithaca, New York 14860 (Received M a y 29, 1968)
Electron spin resonance spectra have been observed for the radical anions 1,4,5,8-tetramethylanthrasemiquinone and 1,4-diinethylanthrasemiquinone. For the former radical in dimethoxyethane solution we find 12 equivalent methyl protons with UH = 0.329 G and 4 ring protons with UH = 1.153 G. Assignments for spectra of both radicals indicate that the methyl groups are freely rotating on the time scale of the esr experiment. This finding is in contrast to a previous report in which rotation of methyl groups in the 1 position of several anthrasemiquiiione anions was reported to be so strongly hindered that, the protons were no longer magPetically equivalent.
Introduction The chemical literature is replete with studies of semiquinone radicalsJ3 including several papers concerning anthrasemiquinones.4 Interest in the two radicals discussed here, however, was generated by a recent report by Elofson and coworkers5 in which it was asserted that the rotation of methyl groups in the 1 position of several anthrasemiquinorie anions was sufficiently hindered that the methyl protons were magnetically nonequivalent and exhibited different esr hyperfine coupling constants. This assertion was made for radicals in 2-propanol solution at room temperature. We found it remarkable that the rotation of such methyl groups should be so completely hindered, since, to our knowledge, in the host of radicals studied by esr in solution at room temperatures, no similar assertions have previously been possible.e In fact, on the basis of part of the work presented here, these assertions have already been q ~ e s t i o n e d . ~We report here esr data for the P,4,5,8-tetramethylanthrasemiquinoneanion and a reexamination of the esr spectrum of one of the radicals reported5 to exhibit nonequivalent methyl protons, 1,4-dimethylanthrasemiquinone anion. We shall employ throughout the standard numbering scheme given below.
8
0
Experimental Section Materials. Solvents and reagents were commercially available and were purified by established literature methods prior to use. Tetrabutylammonium perchlorate (Southwestern Analytical Chemicals), or the tetraethyl salt (Distillation Products Industries), was
employed as the supporting electrolyte in electrochernical procedures at a concentration of 0.1 A I . 1,4,5,8-Tetramethylanthraquinone (TXAQ) and 1,4-dimethylanthraquinone (DMAQ) were synthesized by modifications8 of the procedures described by Alder and I < ~ t h . Final ~ purification of both compounds 'was effected via vacuum sublimation. Satisfactory elemental analyses and exact agreement with litcratureg melting points were obtained for both quinones. Electyochemist?-y. Polarography was carried out using the Oak Ridge National Laboratory three-electrode polarograph described previously.'" Polarographic data were obtained in acetonitrile using tetraelhylammonium perchlorate supporting electrolyte. Radicals for esr measurements were produced by total electrolytic reduction of TMAQ and DIIAQ in 1,Bdimethoxyethane (DRlE) solution to yield the corresponding semiquinone anions (TMAQ- and DRIA&-). A vacuum electrolysis cell similar to that described by Bolton and Fraenliel" was employed. (1) Abstracted frorn the Ph.D. thesis of R. H. Schlossel, Cornell University, 1968. (2) (a) Deceased, Dec 4, 1967; (b) author to whom correspondence should be addressed at the Florida State University, Tallahassee, Fla., 32306. (3) W. M. Gulick, Jr., and D. H Geske, J . Amer. Chem. Soc., 8 8 , 4119 (1966), and references cited therein. (4) See, for example, G. Vincow and G. K. Fraenkel, J . Chem. Phys., 34, 1333 (1961).
(5) R. M. Elofson, K. F. Schub, B. E. Galbraith, and R. Newton, Can. J . Chem. 43, 1553 (1965). (6) Different coupling constants for the protons of a methyl group were observed for the propyl radical in propane solution in the temperature range - 145 t o - 180' (R. W. Fessenden and R. H. Schuler, J . Chem. Phys., 39, 2147 (1963)). (7) D. H. Geske, Progr. Phys. Org. Chem., 4 , 188 (1967). (8) The authors thank Professor M. J . Goldstein for helpful suggestions with the syntheses. (9) K. Alder and R. Kuth, Ann. Chem., 609, 32, 35 (1957). (10) K. Kuwata and D. H. Geske, J . Amer. Chem. Soc., 86, 2101 (1964). (11) J. R. Bolton and G. K. Fraenkel, J . Chem. Phgs., 40, 3307 (1964).
Volume 73, Number 1 January 1969
R. H.SCHLOSSEL, D. H. GESKE,AKD W. 14. GULICK,JR.
72
Table X : Results CalodO spin density Radical
Position
No. of
-Hyperfine Calcd*
(P",
splittings, G ExptP
(a)
(14
equiv
spins of I = l/a
DMAQ- (ethanol), line width of 0.045 G
1 2 5 6
0.02191 0.03939 0.02071 0.03738
0.767 -1.064 -0.559 -1,009
0.837 1.193 0.433 0.931
6 2 2 2
DILIAQ- (UME), line width of 0.035 G
1 2 5 6
0.01143 0.04167 0.01054 0.04035
0.400 -1,125 -0.285 -1.089
0.512 1.179 0.210 0.902
6 2 2 2
TMAQ- (DME), line width of 0,055 G
1 2
0.01142 0.04161
0.400 -1,123
0.329 1.153
12 4
Anthrasemiquinone (DNE)
1 2
0.01085 0.04041
-0.285 -1,091
0. 271d 0.974d
4 4
Anthrasemiquinone (ethanol-water)
1 2
... ...
0. .i50d 0 . 962d
4 4
.
C..
a.
a Calculated using molecular orbital parameters from M. R. Das and G. K. Fraenkel, J . Chem. Phys., 42, 1350 (1065),and methyl group parameters from C. A. Coulson and V. A. Crawford, J . Chem. Soc., 2052 (1953). Spin densities given include the approximate configuration interaction method of McLachlan (A. D. McLachlan, MoE. Phys., 3, 233 (1960)). * Calculated via U H = -27 pc* and Data from M. R. Das and G. K. Fraenkel, J . Chem. Phys., 42, 1350 U H ( C H ~ ) = 35 pan. ' Uncertainty estimated a t the 1% level.
(1965).
1
a
1.5 0
I
Figure 1. (a) Low-field half of the esr spectrum of 1,4-dimethylanthraquinoneanion in absolute ethanol (small numbered arrows indicate points of correspondence in Figures l a and 2a); (b) computed spectrum based on the assignment and line width given in Table I.
Samples prepared in this way were stable at room temperature for several weeks. DMAQ was also reduced electrolytically in absolute ethanol using a cell designed in these laboratories and described elsewhere.I2 Esr Spectra. The Varian X-band 100-kHz spectrometer system employed has been described previously'o as has the method of magnetic field ~alibration.~ The Varian V-4547 variable-temperature accessory was used to obtain spectra below room temperature. Simulations of esr spectra were computed using a modification of the program devised by Stone and 1\!laki18executed on a Control Data Gorp. 1604 computer. The Journal of Physical Chemistry
Figure 2. (a) Esr spectrum of 1,4-dimethylanthraquinone anion in absolute ethanol, low-field end line region (see Figure la for position relative to the rest of the spectrum); (b) reconstruction of this portion of the spectrum based on the assignment given in Table I.
Results Esr results are summarized in Table I. All spectra (12) W. M. Gulick, Jr., W. E. Geiger, Jr., and D. €1. Geslte, J . Amer. Chem. SOC.,90, 4218 (1968). (13) E.W. Stone and A. H. Maki, J . Chem. Phys., 38, 1999 (1963)
ESRINVESTIGATION
O F RIETHYLATED
ANTHRASEMIQUINONE RADICALS
73
Table I1 : Polarographic Results in Acetonitrile
- EI12" DMAQ TMAQ Anthraquinoned Nitrobenzenee
1st reduction wave
- [&/a - E1/41b
1.082 1.272 0.94 1.147
55 49
... 56
2nd reduction w&ve IO
4.3
3.2 1.5 4.1
-El/r"
1.569 1.727 1.45
...
- [&/a - Ei/alb
I=
52 70
3.8 3.1 1.5
... ...
.
...
'
a Volts us. aqueous saturated colomel electrode. Millivolts; for a reversible electron transfer -56/n mV is predicted (J. Tomes, Coll. Czech. Chem. Commun., 9, 1281 (1937)). I = id/Cm*/st'/6. Data from M. E. Peover, J . Chem. SOC.,4540 (1962). Data from D. H. Geske and A. H. Maki, J . Amer. Chem. Soc., 82,2671 (1960).
Figure 3. (a) Esr spectrum of 1,4,.5,8-tetramethylanthraquinoneanion in 1,Z-dimethoxyethane; (b) computed spectrnm based on the assignment and line width given in Table I.
obtained were assignable only on the basis of magnetic equivalence of all methyl proton^.'^ Figure 1 compares the esr spectrum of DMAQ- in absolute ethanol at 25" with a computed spectrum based on the coupling constants and line width indicated in Table I. I n Figure 2 a detailed assignment of the end line region of this spectrum is shown. Comparable agreement between experimental and computed spectra was obtained for DhIAQ- in DXE. In Figure 3 the experimental spectrum of TMAQ- in DME at 25" is compared with a spectrum computed using the parameters given in Table I. Magnetic equivalence of the methyl protons in TlIAQ- is retained at lower temperatures. At -50" in DlIE the spectrum is readily assigned to 12 methyl protons with CLH = 0.353 G, 4 ring protons with CLH = 1.153 G, and a line width of 0.087 G . Electrochemical results are given in Table 11.
assignments given are correct. Particular attention was paid to eliminating alternative assignments in which one proton from each methyl group had a coupling constant significantly smaller than the other two protons. Such alternative assignments proved completely unacceptable for TRIAQ- and exceedingly unlikely for DllAQ- in either solvent. When possible, assignments of coupling constants to specific molecular positions have been made from the numbers of equivalent protons involved as deduced from the esr splitting patterns. The ambiguities remaining for DMAQ- were resolved with the aid of molecular orbital calculations and by comparison of the effect of changing the solvent medium with the solvent effect reported'5 for anthrasemiquinone. The previous assignment given for DMAQ- in 2propanol5 is six pairs of equivalent protons with coupling
Discussion Careful examination of the spectra presented here, particularly the end line regions, convinces us that the
(14) A detailed discussion of these assignments and consideration of alternative assignments can be found in the Ph.D. thesis of R. H. Schlossel, Cornell University, 1968.
Volume 73, Number 1
January 1069
74
R . H. SCHLOSSEL, D. H. GESKE,AND W. M. GuI;ICKI,JR.
-50" the rotation of methyl groups in these radicals is rapid with respect to any geometrically induced difference in hyperfine frequency. Since the esr spectra of methylated anthrasemiquinones were previously6 assigned by a computerized numerical analysis,lO we wish to emphasize the cautions with which one must approach any type of mathematFigure 4. Computed esr spectrum of 1,4-dimethylanthraquinone ical analysis which attempts to analyze and reproduce anion in 2-propanol based on the assignment of Elofson, observed data. First, given a sufficient number of adet U L . , ~ and assuming a line width of 0.035 G (see the text). justable parameters, it is possible to reproduce any set of data. Second, the confidence one may put in the constants as follows: ring protons: aH = 0.68, a H = conclusions drawn from any such analysis procedure 0.64 and a H = 0.33, and methyl protons: a H = 0.99, must depend on the accuracy and detail of the data a H = 0.96 all U H = 0.05 (all values in gauss). Although which are being approximated. The implications for we were unable to obtain esr spectra of either radical computer analysis of em spectra follow in a straightin 2-propanol] we feel that the ethanol system is suffi- forward manner. cient*lysimilar to permit reasonable comparison. The Acknozoledgment. The authors gratefully aclmowlprevious report6 gave no indication of the experimental edge support for this research from the National line width; however, for the sake of comparison a Science Foundation under Grant GP-4906, as well as computed spectrum based upon the coupling constants through Grant GP-1687 for partial support for purchase given above and a very narrow line width, 0.035 G, of the esr spectrometer. is given in Figure 4. This computed spectrum appears to be better resolved than the previous5 experimental spectrum of DMAQ-, yet it is not certain that unam(15) M. R. Das and G. K. Fraenkel, J . Chem. Phys., 42, 1350 (1965). biguous assignment of even this spectrum is possible. The experimental proton splittings from this reference for anthrasemiquinone in basic ethanol-water are in turn taken from G. It is clear that the previous assignment is wholly Vincow and G. K. Fraenkel, J . Chem. Phys., 34, 1333 (1961). incompatible with the experimental results given here. I n a personal communication, Dr. M. R. Das assures us that substantially identical proton splittings were observed when anthraThus, barring the unlikely circumstance of some unique semiquinone was generated by eloctroreduction in neutral absolute medium effect of basic 2-propanol, we are compelled ethauol. Thus the differences between splittings observed in DLME and ethanol represent a true solvent effect rather than a pH or to reject as untenable the assertion of blocked rotation gegenion effect. of l-methyl groups in anthrasemiquinone radicals. (16) It. Newton, K. F. Schulr, and R. M. Elofson, Cun. J . Chem., We conclude just the opposite, that, down to at least 44, 752 (1966).
The Journal of Physical Chemistry
