1954
J. Phys. Chem. 1985,89, 1954-1958
Equation A9 is solved for A-, for given values of the experimental variables I / V p and C*, and appropriate values of R and B. In the experiments here the light intensity, and hence Z/Vp, was held constant. Thus by inputting values for CC we can observe changes in the value of A- as a function of the kinetic parameters. A- is, of course,related to is by eq 10 and is equivalent to is/nFVp. For convenience we have divided C+ and the resulting values of
A- by the constant term I / Vp. This results in two dimensionless terms i,/nFI and pc*V/I. As we mentioned earlier expressing the current this way is advantageous, since it is an expression of the efficiency of our system. Registry No. MVZt, 1910-42-5;BQ, 106-51-4;FeB(OH)2t, 4717602-3;Fe(CN)63-, 13408-62-3; R U ( N H ~ ) ~18943-33-4; +, CdS, 1306-23-6; Ru02. 12036-10-1; Pt, 7440-06-4;KN03, 7757-79-1.
Fourler Transform Numerkal Analysls of the Long-Range Proton Hyperflne Coupllng In Nitroxide Radicals P.Trousson and Y. Lion*+ Department of Physical Chemistry, The University, Leicester LE1 7RH, U.K. (Received: July 17, 1984; In Final Form: October 30, 1984)
A study of long-range proton coupling in nitroxide radicals has been performed with a numerical analysis program using the Fourier transform technique. The present method provides a means for further identificationof radicals which is particularly useful for species that are otherwise indistinguishable. The superhyperfine structure of piperidine and pyrrolidine-1-oxy1 derivatives, showing y-nuclei coupling constants as small as 0.2 G, has been brought out. The results are in good agreement with those obtained by other resolution-enhancement methods.
Introduction This study presents clear evidence of the potential of a postexperimental resolution-enhancement method. The technique, detailed in ref 1, is based on Fourier transform properties and consists of summing higher derivatives of the ESR signal and narrowing the line by changing the line shape function from Lorentzian to Gaussian. This method, which already proved to be helpfu12v3in solid matrices, is now applied to nitroxide radicals obtained by the spin-trapping method. The spin traps are extensively used to convert short-lived radicals into stable spin adducts by addition. This stabilization allows the detection and identification of the short-lived radicals if the structure of the spectrum is sufficiently resolved. It is unfortunately not the case when nitrone spin traps are used, depending on the radical trapped, the nitrone spectra exhibit splitting constants slightly different in magnitude, but the general structure of the spectra remains a six-line pattern. The information due to the superhyperfine (shf) interactions arising from nuclei three or four u bonds away from the unpaired electron must be extracted from the line width of the observed lines. We demonstrate in this work that the postexperimental data processing enables the observation of long-range nuclei interactions, with splitting constants as small as 0.2 G. Study of such shf interactions on 4-R-2,2,6,6-tetramethylpiperidine- 1-oxy1 and 2-R-5,5-dimethyl-l-pyrroline N-oxide has already been performed by Mossoba et a1.4 with another resolution-enhancement method. Both results are in good agreement, although, in some cases, our spectra exhibit even higher resolution, and our hyperfine structure (cepstral) analysis provides, in every case, precise hyperfine splitting constants. Experimental Section 4-R-2,2,6,6-Tetramethylpipidine 1-oxy1 and 5,5-dimethyl- 1pyrroline N-oxide (DMPO) were purchased from Aldrich Chemical Co. and N-tert-butylphenylnitrone(PBN) from Eastman Kodak Co. Purification of DMPO was carried out according to the procedure of Buettner and O b e r l e ~ . ~M e 2 S 0 and D 2 0 (99.75%) were obtained from Merck Co., and H 2 0 2(30 vol.%) was acquired from UCB (Belgium). Dilute aqueous solutions of
'
Institut de Physique (B5), Universitt Libge, Sart-Tilman, 4000 Liege, Belgium.
H
0'
N
II 0'
PIPERIDINE-
1 -0XYL
-
PY RROLIDINE 1 OXYL
-
N-BUTY L
-
PHENYLNITRONE
TEMPO R:H R=CH3 ,OH,OD,COOH TEMWL R =OH TEMPAMINE R = NH2 TEMPONE R = O
lo-" M of the stable nitroxide radicals were used, and dissolved oxygen was removed by bubbling nitrogen gas prior to the EPR measurements. The methyl adduct of DMPO was prepared by the UV photolysis of 30% H,Ozin an aqueous solution of DMPO (0.1 M) containing MezSO. COOH-DMPO was formed by the UV photolysis of an aqueous solution containing DMPO (0.1 M), sodium formate (1 M), and H202The OH and OD adducts were generated by the photolysis of H20and D 2 0 solutions containing DMPO and H202. (1) Trousson, P.;Rinne, M. Reu. Sci. Instrum., in press. (2) Symons, M. C.R.; Trousson, P. Radiat. Phys. Chem. 1984,23,127. (3) Maj, S. P.;Symons, M. C. R.; Trousson, P.J . Chem. SOC.,Chem. Commun., in press. (4) Mossoba, M. M.; Makino, K.; Riesz. P.; Perkins Jr., R. C . J . Phys. Chem. 1984,88,4717. ( 5 ) Bucttner, G . R.; Oberley, L. W. Biochem. Biophys. Res. Commun. 1978,83,69. (6) Kirmse, D.W. J . Magn. Reson. 1973,ZI, 1. (7) Windle, J. J. J . Magn. Reson. 1981,45, 432. (8) Brii?re, R.; Lemaire, H.; Rassat, A.; Dunand, J.-J. Bull. SOC.Chim. Fr. 1967,4479. (9) Kreilick, R. W. J. Chem. Phys. 1967,46, 4260. (10)Bordeaux, D.;Lajzerowicz, J. Acta Crystallogr., Sect. B 1974,830, 790. (1 1) Heller, C.; McConnell, H. M. J . Chem. Phys. 1960,32, 1535. (12) Lajzerowicz, J. Acta Crystallogr., Sect. B 1968,824,196. (13) Berliner, L. J. Acta Crystallogr., Sect. B 1970,826,1198. (14) Mao, S.W.; Kevan, L. Chem. Phys. Lett. 1974,24, 505. (15)Harbour, J. R.;Chow, V.;Bolton, J. R. Can. J. Chem. 1974,52,3549. (16)Lai, C. S.; Piette, L. H. Arch. Biochem. Biophys. 1978,190, 27. (17)Kasai, P. H.;McLeod Jr., D. J . Phys. Chem. 1978,82, 619.
0022-365418512089-1954%01.50/0 0 1985 American Chemical Society
Proton Hyperfine Coupling in Nitroxide Radicals
The Journal of Physical Chemistry, VoZ.89, No. 10, 1985
1955
Photolysis was carried out in situ in the EPR quartz cell placed in the EPR cavity using a high-pressure mercury lamp (Osram HB0500). PBN adducts were formed by y-radiolysis)C o@ '( of 0.1 M of PBN in the same solvents as above. All spectra were recorded on a Varian E-109 X-band (9.5 GHz) EPR spectrometer. Spectra were digitized by using the interface and the sampling program described in ref 1.
Fourier Transform Procedure The increased resolution is obtained by performing, in the Fourier domain, the differentiation of the EPR signal, by the addition of high derivatives, and by modifying the line width of the line shape function. The equations of the EPR theory are very well-known, and hence their conversion in the Fourier domain could be performed. The main interest in working in the Fourier domain is due to the convolution theorem stating that the Fourier transform of a convolution product of functions is the arithmetical product of their Fourier transforms. An EPR spectrum is the convolution of two functions (positions of the peaks and line shape function). Henceforth its representation in the Fourier domain consists of the product of the Fourier transforms of these two functions. Hence, it is easier in this domain to modify the spectrum compared with a deconvolution process which is always difficult to compute. High derivative calculations are, for similar reasons, more quickly performed. In this study, every spectrum has been processed in two different ways. Firstly, the second and fourth derivatives were calculated and added. The result is then compared with the corresponding enhanced spectrum obtained by Mossoba et al.4 Secondly, the first (i.e., experimental spectrum), third, and fifth derivatives of the spectrum were added, and the line width of this resulting sum was modified. This second manipulatioii generally shows additional features, compared with the second harmonic type, and hence allows easier computation of the splitting constants. Furthermore, a superhyperfine structure analysis was performed on each spectrum to reveal the precise superhyperfine splitting constants of each radical. Further details of the method are available and will be published elsewhere. The precision of our hyperfine structure analysis (cepstral analysis) depends on the number of points used to digitize the spectra. With 512 points, the precision on the shf splitting constant is fO.O1 G (cf. Tables 1-111). Results and Discussion Piperidine-1-oxy1 Derivatives. Figure l a shows the expansion of the M I= +1 line of the triplet due to 14N( I = 1) in Tempo. Figure 1, parts b and c, displays considerable superhyperfine structure due to further interactions of the unpaired electron with y-protons. Figure l b is obtained by adding the second and fourth derivatives of spectrum l a and by modifying the line width from a Lorentzian of 0.12 G to a Gaussian of 0.08 G. Figure I C is obtained by the addition of the first, third, and fifth derivatives and by a line width narrowing from 0.12 to 0.08 G. Figure Id is the superhyperfine analysis display, called cepstrum? of the MI = +1 line. The abscissa of the peaks give the shf splitting constant values and their multiples. The succession of positive and negative peaks allows one to know the spin value of the nucleus. Two sets of nuclei could be extracted from Figure Id. A first set of lines starting with the peak at abscissa 0.19 G corresponds to the four methyl groups (C2, C6). The second set, with a splitting constant of 0.38 G, is due to the four methylene protons (C3, CJ. Those results are consistent with the fact that the first peak (due to 12 equivalent hydrogens) is 3 times more intense than the second (due to 4 equivalent hydrogens). A stimulation of the MI = +1 line has been performed with these data (cf. Table I) and is displayed in Figure le. The simulated spectrum in in good agreement with the results obtained in Figure IC, d. Similar computations have been performed on every spectrum of the compounds under investigation. A selection of these results is presented in Figures 2 and 3 for Tempol, Tempamine, and Tempone. The left column presents the addition of the second
. 1.6 G.
d)
Figure 1. (a) M I= +1 line from Tempo spectrum: the time constant
is 0.128 s, the microwave power is 1 mW, and the modulation amplitude is 0.05 G . (b) Addition of the second and the fourth derivatives of (a) together with a line width narrowing from 0.12 to 0.08 G. (c) Addition of the first, third, and fifth derivatives of (a) together with a line width narrowing from 0.12 to 0.08 G . (d) Hyperfine structure (cepstral) analysis display of the resolved line in (c). Two sets of equivalent protons (12 and 14, respectively) appear with splitting constants of 0.19 and 0.38 G., respectively. (e) Synthetic M I= +1 line, simulated with splittings in Table I and line width of 0.08 G .
.
TEMPAMINE
..
16 G
Figure 2. Left-hand column displays resolved M,= +1 line in the second derivative form (addition of the second and the fourth derivatives plus a line width narrowing) for Tempol, Tempamine, and Tempone, respectively. The right-hand column displays the same lines in the first derivative form (addition of the first, the third, and the fifth derivatives). Splitting constants can be found in Table I. and fourth derivatives and the right column, the sum of the first, third, and fifth derivatives. Our results, summarized in Table I,
1956 The Journal of Physical Chemistry, Vol. 89, No. 10, 1985
Trousson and Lion
TABLE I: Hyperfii Splitting Constants of Piperidine-1-oxy1 Derivatives shf structure 'H NMR
I4N
radicals Tempo (R = H)
splitting, G
involved nuclei
our results'
Mossoba'sb
datac
16.8
4 CH3 (C29 c6) 4 H (C3, Cs) 2 H (C4)
0.19 0.38 d
0.195 0.375 0.175
-0.23 -0.39 +O. 18
Tempo1 (R = OH)
16.6
2 ax CH3 (c2, c6) 2 eq CH3 (C2, c6) 2 ax H (C3, C5) 2 eq H (C39 C,) ax -H (C4)
0.45 d 0.33 0.50 d
0.45 d 0.33 0.50 0.09
-0.45 -0.02 -0.3 1 -0.48 +0.07
Tempamine (R = NH2)
16.5
2 ax CH3 (C2, c6) eq CH3 (c29 c6) 2 ax H (C3, C5) 2 eq H (C39 C5) ax -H (C4) ax -H (C4)
0.48 d 0.33 0.50 0.07 0.11
0.48 d 0.33 O.5Oc 0.1 1 0.14
Tempone (R = 0)
15.6
4 CH3 e 2 7 C,) 4 H ( G 3 C5)
(
