J. Phys. Chem. 1987, 91, 119-784
779
DMR-8217227A-01 is also acknowledged. Registry No. Rh, 7440-16-6; Ni, 7440-02-0; Co, 630-08-0; NO, 10102-43-9;Fe, 7439-89-6; CH,OH, 67-56-1; CH,O, 2143-68-2; H2, 1333-74-0;02, 7782-44-7.
Chemistry Grant No. DMR-8413561. The use of facilities/ equipment provided by the Office of Naval Research under Grant No. N00014-81-K-0505 and the Cornell Materials Science Center through the National Science Foundation under Grant No.
ARTICLES Analysis of Resolved ESR Spectra of Neutral Nitroxide Radicals in Ethanol and Pyridine: The Dynamic Behavior in Fast Motion Conditions M. Francesca Ottaviani Department of Chemistry, University of Florence, 501 21 Firenze, Italy (Received: March 24, 1986; I n Final Form: July 29, 1986)
The ESR spectra of ethanol and pyridine solutions of the spin probes 4-hydroxy-2,2,6,6-tetramethylpiperidine1-oxy1 (Tempol) and 3-carbamoyl-2,2,5,5-tetramethylpyrroline1-oxy1 (Tempyo) were analyzed to obtain information on the dynamic behavior of the radicals in solution. In order to evaluate the true line widths, the dependence of the proton and carbon-13 hyperfine coupling constants on temperature and on solvent polarity were analyzed as a function of aN. Tempol and Tempyo showed an opposite dependence of aHon aN. Furthermore the structure of Tempol was affected more than that of Tempyo by variations of solvent polarity. From the line widths the correlation times for motion were evaluated. A weak hydrogen bond was formed by Tempol and Tempyo with ethanol. However, only Tempol in ethanol gave rise to an anisotropic rotational reorientational motion, with the fastest rotation about the x axis. All samples, except for Tempol in pyridine, showed a deviation from the Debye spectral densities. The hydrodynamic rotational radii were about 3 times lower than the effective geometric spherical radii. However, Tempyo exerted slightly greater torques than Tempol during its rotational motion. At the highest temperatures ( T > 298 K), the dynamic behavior was as expected on the basis of the free volume model, whereas at lower temperatures an increase in the H-bonding strength was found between radical and ethanol molecules.
It is well-known that the analysis of the electron spin resonance (ESR) spectra of paramagnetic probes in solution provides deep information on the dynamics of both the probe and the solvent. In particular, nitroxide radicals have been widely used in the past.' The line shape of undeuterated nitroxide radicals in the fast motional region is complicated by unresolved proton hyperfine ("superhyperfine", shf) interactions leading to inhomogeneous b r ~ a d e n i n g ~and - ~ by Heisenberg spin-exchange broadening due to ~ x y g e n . ~ However, ,~ well-resolved spectra, obtained from thoroughly deoxygenated solutions, provide detailed information about the radical and its environment. Computer simulation of line shapes is usually carried out to evaluate the line widths by using the proton coupling constants obtained from N M R experiments.8-'0 It has been demonstrated that the coupling constants are a function of the polarity of the en~ironment""~and
are influenced by changes in temperat~re.~>*,"-'~ Therefore, to perfectly reproduce the experimental ESR spectra, it should be useful to take into account the variations of proton and carbon-13 constants-namely aHand a~~--withtemperature and solvent. The dynamic behavior of the radicals in solution can be analyzed in terms of the correlation times for motion.20q21 If the nitroxide molecules are considered to be axially symmetric rotors, Freed and c o - ~ o r k e r shave ~ ~ ' ~suggested ~ a very powerful method of analyzing the molecular dynamics of predeuteriated 2,2,6,6tetramethyl-4-piperidone-1-oxy1 (PD-Tempone) in different solvents. Furthermore, the evaluation of the Kivelson parameter K'/3,24-27 defined as the ratio between the effective rotational radius and the geometric spherical radius, gives information on the "stickiness" behavior of the radicals in solution. In the present work, we analyzed ethanol and pyridine solutions of two nitroxide radicals with significantly different structures:
(1) Spin Labeling Theory and Application; Berliner, L. J., Ed.; Academic: New York, 1976; 1979, Vols. I and 11. (2) Poggi, G.;Johnson, C. S.J. Magn. Reson. 1970, 3, 436. (3) Plachy, W.; Kivelson, D. J . Chem. Phys. 1967, 47, 3312. (4) Whisnant, C. C.; Ferguson, S.; Chesnut, D. B. J . Phys. Chem. 1974, 78, 1410. ( 5 ) Lang, J. C.; Freed, J. H. J . Chem. Phys. 1972, 56, 4103. (6) Popp, C. A,; Hyde, J. S. J. Magn. Reson. 1981, 43, 249. (7) Hyde, J. S.; Subczynski, W. K. J . Magn. Reson. 1984, 56, 125. (8) Kreilick, R. W. J . Chem. Phys. 1967, 46, 4760. (9) Briere, R.; Lemaire, H.; Rassat, A,; Rey, P.; Rousseau, A. Bull. SOC. Chim. Fr. 1967, 12, 4479. (10) Windle, J. J. J . Magn. Reson. 1981, 45, 432. (1 1) Kamlet, M. J.; Taft, R. W. J. Chem. SOC.,Perkin Trans. 2 1979, 349. (12) Briere, R.; Lemaire, H.; Rassat, A. Bull. SOC.Chim. Fr. 1965, 11, 3216. (13) Mukerjee, P.; Ramachandran, C.; Pyter, R. A. J . Phys. Chem. 1982, 86, 3180. (14) Knauer, B. R.; Napier, J. J. J. Am. Chem. SOC.1976, 98, 4395. (15) Janzen, E. G.; Coulter, G. A. Tetrahedron Letr. 1981, 22, 615.
0022-3654 ,187,I209 1-0779SO1S O IO I
(16) Janzen, E. G.; Coulter, G. A,; Oehler, U. M.; Bergome, J. P. Can. J . Chem. 1982, 60, 2725. (17) Reddoch, A. H.; Konishi, S. J. Chem. Phys. 1979, 70, 2121. (18) Bullock, A. T.; Howard, C. B. J . Chem. SOC.,Faraday Trans. 1 1980, 76, 1296.
(19) Janzen, E. G. Can. J. Chem. 1984, 62, 1653. (20) Goldman, S. A.; Bruno, G. V.; Polnaszek, C. F.; Freed, J. H. J. Chem. Phys. 1972, 56, 716. (21) Wilson, R.; Kivelson, D. J . Chem. Phys. 1966, 44, 4445. (22) Zager, S. A.; Freed, J. H. J . Chem. Phys. 1982, 77, 3344. (23) Hwang, J. S.; Mason, R. P.; Hwang, L. P.; Freed, J. H. J . Phys. Chem. 1975, 79, 489. (24) Dote, J. L.; Kivelson, D.; Schwartz, R. N. J . Phys. Chem. 1981,85, 2169. (25) Kivelson, D.; Madden, P. Annu. Reu. Phys. Chem. 1980, 31, 523. (26) Hoel, D.; Kivelson, D. J. Chem. Phys. 1975,62, 1323; 1975,62,4535. (27) Kowert, B.; Kivelson, D. J . Chem. Phys. 1976, 64, 5206.
0 1987 American Chemical Societv -
780 The Journal of Physical Chemistry, Vol. 91, No. 4, 1987
Ottaviani
d 2G
Figure 1. m N = 0 hyperfine lines of the computed (dashed lines) and the experimental (full lines) ESR spectra at 298 K of 2 (a) Tempol in ethanol and (b) Tempyo in ethanol.
M solutions of
X
4-hydroxy-2,2,6,6-tetramethylpiperidine1-oxy1 (Tempol) and 3-carbamoyl-2,2,5,5-tetramethylpyrroline1-oxy1 (Tempyo). Their
PH
a) oa
I Tempol
rq
rq
I Tempyo
conformations are known from X-ray analysisz8to be respectively in chair form and in planar form. However, both of them contain groups that may participate in hydrogen bonding, Le., N-O, -OH, -NH2. In what follows we first report on the dependence of aHand allc on temperature and solvent polarity. Then, in the analysis of the results we consider how the dynamic behavior of the radicals is affected by (1) the temperature variation, (2) the polarity of the solvents, and (3) the structure of the radicals. This analysis is really useful to correctly investigate more complicated systems in which Tempol, Tempyo, or other nitroxide radicals can act as spin probe^.^^,^^
Experimental Section ESR Measurements. Solutions (2 X lo4 M) of Tempol and Tempyo (Sigma Munchen, used without further purification) in absolute ethanol and pyridine (Merck, freshly distilled under vacuum) were prepared. Then the solutions were deoxygenated by bubbling through them dry and deoxygenated nitrogen for 2 h. These were stored and always handled in a N2 atmosphere. ESR spectra were measured with a Bruker 200D spectrometer operating at X band (-9.5 GHz). The field calibration was achieved by referring to DPPH. The spectrometer was equipped with the Bruker ST 100/700 variable-temperature assembly. For 77 K spectra, the liquid nitrogen cold finger was used. The spectrometer was interfaced to an ASPECT 2000 computer (Bruker) for data acquisition and handling. (28) Lajzerowicz-Bonneteau, J. In Spin Labeling Theory and Applications; Berliner, L. J. Ed.; Academic: New York, 1976; Vol. I. (29) Martini, G . ;Ottaviani, M. F.; Romanelli, M. J . Colloid Inter$ Sci. 1983, 94, 105. (30) Ottaviani, M. F.; Baglioni, P.; Martini, G. J . Phys. Chem. 1983,87. 3146.
Figure 2. 'HNMR spectra of 1.5 M solutions of Tempol in ethanol-d, at (a) 298 and (b) 258 K.
N M R Measurements. Solutions (1.5 M) of Tempol and Tempyo in ethanol-d6 and pyridine-d5 (Merck, used without further purification) were prepared. NMR spectra were recorded with the aid of a Bruker pulse N M R spectrometer CXP with a B-VT 1000 variable-temperature assembly. Tetramethylsilane (Me,Si) was used as internal standard.
Results and Discussion Analysis of aHand Ul3c Coupling Constants. Figure 1 shows the central hf lines of the ESR spectra of Tempol (la) and Tempyo (1b) solutions in ethanol at 298 K (full lines, experimental spectra; dashed lines, computed spectra). A good resolution of the shf structure due to protons was shown either by the I4N lines or by the I3C lines. The computer simulation was carried out for each 14N hyperfine manifold with a double precision Fortran program by summing derivatives of Lorentzian lines with the same widths. The best fits between computed and experimental spectra were obtained by using the aHcoupling constants for the different groups of equivalent protons listed in Table I. The evaluation of the uH constants was accomplished by using the N M R paramagnetic shifts, namely, AH,, as shown in ref 8-10; that is, uH = -4.97 X 10-'AHpT, where Tis the absolute temperature. However, as shown in Figure 2 for Tempol in ethanol-d, at 298 and 258 K, the broad poorly resolved N M R peaks hardly allowed the evaluation of AH,. Indeed, the ESR spectra computed with proton coupling constants directly obtained from N M R spectra did not fit the experimental ESR spectr;. By a trial-and-error process, such u H values were progressively corrected until the required fit was obtained. The computation process used to reproduce the I4N manifolds was also used to compute the I3C hf lines. The carbon-13 coupling constants reported in Table I were directly
ESR Spectra of Neutral Nitroxide Radicals
The Journal of Physical Chemistry, Vol. 91, No. 4, 1987 781
TABLE I: Hyperfine Couplinz Constants and Anisotropic g Components for Tempol and Temovo in Ethanol and Pvridine radical solvent Tempol ethanol
pyridine
Tempyo
aN*
aHCH3,q9
aHCH3,aau,
aHCH~.cq,
aHCH2,nx*
aH?*
a'CH3.a~.
ac&ring,
aN,x*
aNyr
aN,n
G
G
G
G
G
G
G
G
G
G
298 278 258
G 16.09 16.13 16.15
0.452 0.460 0.470
0.020 0.020 0.020
0.478 0.480 0.480
0.320 0.325 0.333
0.070 0.070 0.080
6.82 6.85 6.89
4.22 4.25 4.27
5.8
7.0
35.5 20093
2.0062 2.0027
298 278 258
15.56 15.58 15.60
0.443 0.454 0.460
0.020 0.020 0.020
0.469 0.470 0.473
0.312 0.318 0.322
0.070
6.40 6.42 6.45
4.00 4.10 4.12
5.6
6.7
34.7 2.0095
2.0063
298 278 258
15.01 14.94 14.91
0.232 0.234 0.235
0.473 0.480 0.490
6.19 6.22 6.26
5.0
5.3
34.7
2.0082
2.0075
2.0021
298 278 258
14.55 14.52 14.49
0.234 0.235 0.235
0.485 0.492 0.508
6.07 6.10 6.13
4.2
5.1
34.4
2.0085
2.0077
2.0021
T,K
ethanol
pyridine
evaluated from the experimental spectra as the peak-to-peak distances between the central lines of the shf structure in each I3C absorption. Such values were lower than those reported e l s e ~ h e r e land, ~ - ~as~ usually found, were very different from values obtained via NMR. Finally, the ESR signal including both nitrogen-14 and carbon-13 lines was obtained by adding the different contributions at the suitable intensity ratio. Analysis of aH and allc Dependence on Solvent Polarity. It is well-known that a decrease in solvent polarity results in a decrease in the nitrogen coupling c~nstants."-'~J~ However, the electronic distribution is differently affected by different solvent parameters,11-14~'6-19 and it is correlated with the conformation of the radicals.32 Almost all proton coupling constants shown in Table I change from ethanol to pyridine following the trends, d(aH)/d(aN) > 0 for Tempol and d(aH)/d(aN) < 0 for Tempyo. In contrast, the carbon- 13 coupling constants of both radicals decreased as aN decreased, as perviously f o ~ n d . ~ ' The , ~ ~a 1 3 ~ constants are known34.35to depend on the nitrogen spin density, p N r : aljc = (Bo B , cosz 8)pNr, where 8 is the dihedral angle between the nitrogen p" orbital and the N-C-C plane. By assuming p N r B I 7 and pN*BO 0.8 for Tempol and pNrBl 9 ' 9 ~calculated ~ and pNrBO= 0 for T e m p y 0 , ' ~ * ~we Tempol in ethanol
+
=
0.080 0.080
gx
gy
gz
2.0027
of the different conformation of the radicals.'* All the aH and anc values of Tempol and Tempyo, in the present work as elsewhere,'.* showed negative temperature coefficients (see Table I). As aNin Tempol, this behavior is due to out-of-plane vibrations with a high barrier to inversion between the minima of potential with at least three vibrational levels below the top of the barrier.I8 Again, the trend was d(aH)/d(aN) > 0 for Tempol and d(a,)/ d(aN) < 0 for Tempyo. ESR Line Widths and Correlation Times. The line widths can be expressed as a function of mN and mH values.z1 However, due to the observation of almost symmetrical spectra?7 as clearly shown in Figure 1, and of the small ratio ( c I H ) / ( u N ) = 0.025,38the M H dependence can be neglected and the line-width formulation is
AHmN =A
+ A'+
BmN
+ CmN2
(1)
The terms A , B, and C include contributions from the modulation of g and aN tensors and at the X-band are given byZo A =
=
=
8(aCCH3-ax) = 22O Tempol in pyridine
8(acg-ring)= 45O
8(aCcH3-ax)= 27O Tempyo in ethanol
= 47O B(aCB-ring)
B = 6.149 X lO9[AgAu + 36g6a]rB
3
4(1
+ weZ~B2) (3)
8(accH3)= 34O Tempyo in pyridine 8(accH1) = 34.5O These values were different from those evaluated either via or via crystallographic structure2sand confirmed the NMR12,31*33 expected trend of an increase in the dihedral angle with a decrease in the solvent p ~ l a r i t y . ' ~ . This ' ~ effect, however, was almost negligible for Tempyo. Analysis of aHand anc Dependence on Temperature. The aN dependence on temperature has been widely studied in the The observed trends (Table I), d(aN)/dT < 0 for Tempol and d(u,)/dT > 0 for Tempyo, were expected on the basis (31) Hatch, G. F.; Kreilick, R. W. Chem. Phys.,Lett. 1971, 10, 490; J. Chem. Phys. 1972, 57, 3696. (32) Sysoeva, N. A.; Sheichenko, V. I.; Buchachenko, A. L.; Bystrov, U. F.; Rozantsev, E. G.; Neiman, M. B. Zh. Strukt. Khim. 1967, 8, 1094. Sysoeva, N. A.; Sheichenko, V. I.; Buchachenko, A. L. Zh. Strukt. Khim. 1968, 9, 1083. (33) Briere, R.; Chapelet-Letourneux, G.; Lamaire, A.; Rassat, A. Mol. Phys. 1971, 20, 21 1. (34) Karplus, M.; Fraenkel, G. K. J . Chem. Phys. 1961, 35, 1312. (35) Heller, C.; McConnel, H. M. J . Chem. Phys. 1960, 32, 1532. (36) Griller, D. J . Am. Chem. SOC.1978, 100, 5240.
where AaN = aN$ - ' / 2 ( a ~ $+ aN/), 6% = ' / z ( ~ N , Y- ON/), Ag = g, - I/z(gd + gy), 6g = l/z(gd - gy), and we = 5.97 X 10'O Hz. T ~ T , ~ and , TC are the correlation times for the modulation of the anisotropies. In a Debye-Stokes-Einstein approximation, the correlation time is expressed as24-25 T = 4n7rO3/3kT,where ro is the hydrodynamic radius of the probe and 7 the viscosity of the medium. It was also assumed that the approximation j(wN) T is valid over the range of T under study. A' is essentially a spin-rotational contribution. A A', B, and C can be evaluated from the line widths of the hyperfine lines as
=
+
Therefore, if the line widths are known, the correlation times may be evaluated with an accuracy of hO.01 X lo-'' s. (37) Romanelli, M. J . Phys. Chem. 1984, 88, 1063. (38) Jolicoeur, C.; Friedman, H. L. Ber. Bunsengs. Phys. Chem. 1971, 75, 248.
182
The Journal of Physical Chemistry, Vol. 91, No. 4, 1987
Ottaviani ?X
I
:I,/, :I/, zx 1 0 ' l r c c I
10'lrecl
IO
a)
L
la)
IO
L
l/T-4
IC)
'I/T=4
1 / ~ ~ 1 0IK-' 3
3
3.5
4 VTxlO' IK' 1
3
3.5
4
3
3.5
4 VTxM' IK'
I
3
3.5
4 VTxlO'
)
(K' 1
Figure 4. Correlation times for modulation of anisotropies against reciprocal temperatures: T ~ full , lines; i ~ dashed , lines. (a) Tempol in ethanol; (b) Tempyo in ethanol; (c) Tempol in pyridine; (d) Tempyo in
pyridine. 1x104
5
1XiOQ