J . Phys. Chem. 1990, 94, 8 1 13-8 1 18
8113
Ligand Electron Nuclear Double Resonance Study of Cu( I 1)-Doped a-Glycylglycine Single Crystals at 4.2 K C.A. McDowell,* A. Naito, D. L. Sastry, Yu Cui, and K. Sha Department of Chemistry, University of British Columbia, 2036 Main Mall. Vancouver, British Columbia, Canada V6T 1 Y6 (Received: April 16, 1990)
Electron nuclear double resonance (ENDOR) has been used to study Cu( 11)-doped a-glycylglycine single crystals at 4.2 K. Analysis of the ENDOR spectra yielded the principal values and directions cosines of the g tensor, the hyperfine coupling, nuclear quadruple coupling tensors, and the asymmetry parameters for two different I4N nuclei. The two I4N nuclei were identified as N1, belonging to the NH2 group, and N2 to the NH group, respectively. In addition, we determined the hyperfine interaction tensors for four protons and assigned these to their respective molecules. Moreover, by comparing the experimentally determined components of the proton hyperfine interaction tensors with the corresponding values calculated theoretically from the X-ray and neutron diffraction data, we were able to identify uniquely the lattice positions of the interacting protons which yielded the observed ENDOR transitions. We calculated theoretical values of the components of the g tensor using a method that depends on the simultaneous diagonalizing of the ligand field, electron-electron repulsion, and the spin-orbit coding interactions. The theoretical and experimental values are in excellent agreement.
S and the nuclear spin ti;the third term, H,,, is the nuclear Zeeman interaction arising from the effect of the external magnetic Copper complexes have been the subjects of many spectroscopic field B acting on the nuclear spin Ii; and the last term, Hp,accounts studies. That is not only because of their intrinsic interest, but for the nuclear quadrupole interactions. The energy levels of this because those copper complexes containing amino acids, or related spin-Hamiltonian were calculated correct to second order. To compounds, have been regarded as models for many metallocalculate sufficiently accurate values for the principal values and proteinsI4 such as azurin, plastocyanin, etc. Electron parathe direction cosines of the g, the hyperfine and quadrupole magnetic resonance (EPR) and electron nuclear double resonance coupling tensors, we used a computer program which gives those (ENDOR) spectroscopic techniques have been extensively used values correct to seventh order.35 to provide information about the electronic distributions and molecular structures of many Cu(l1) complexes. From EPR studies on single crystals of the complexes one can evaluate accurately the g and hyperfine interaction tensors. The information ( I ) Malkin, R.; Malmstom, B. G . Ado. Enzymol. 1970, 33, 17. on the hyperfine interaction tensors can be obtained more precisely (2) Spiro, T . G., Ed. Copper Proteins; Wiley: New York, 1981. from ENDOR spectroscopy, but in addition, should any of the (3) Harrison, P. M., Ed. Metalloproteins; Verlag Chemie: Weinheim. 1985; FRG, Parts I and 11. ligand nuclei possess a spin I > 1/2, then it is also possible to (4) Gray, H. B. Chem. Soc. Rev. 1986. 15, 17. determine the quadrupole coupling tensor and, of course, and the (5) Losche, A.; Windsch, W. Phys. Status Solidi 1965, 1 1 , K55. quadrupole coupling constants. Of particular interest in this regard (6) Windsch, W.; Webber, M. 2.Naturforsch. 1967, 22, I . are those EPR and ENDOR studies on Cu(I1)-doped single g I y ~ i n e , ~ ?L-~ J ~ J ~ (7) Takeda, K.; Arata, Y.; Fujiwara, S . J . Chem. Phys. 1970, 53, 854. crystals of triglycine suIfate,5,6,13,15,19-22,24-26 (8) Fujimoto, M.; Janecka, J. J . Chem. Phys. 1971, 55, 1 1 52. a-glycylglycine,12~ - h i s t i d i n e ” setc. ~~ (9) Fujimoto, M.; Tomkiewitcz, Y. J . Chem. Phys. 1972, 56, 749. Though the Cu(1I)-ligand bonding sites were readily identified ( I O ) Tomkiewitcz, Y.; Fujimoto, M. J . Chem. Phys. 1972, 56, 3317. in nearly all of those complexes, there remained considerable (1 1) Hirasawa, R.; Kon, H. J . Chem. Phys. 1972, 56, 4467. ambiguityI2 about the nature of the paramagnetic centres in (12) Fujimoto, M.; Saito, S.; Tomkiewicz, Y. Bioinorg. Chem. 1973, 2, Cu(l1)-doped a-glycylglycine. In this case there are several 341. (13) Stankowski, J.; Wiechowski, A.; Hedewy, S . J . Magn. Reson. 1974, possible Cu(I1) bonding sites because there are two nitrogen atoms 15, 498. with different types of environments in each ligand. Because of (14) Bottcher, R.; Heinhold, D.; Windsch, W. Chem. Phys. Left. 1977,49, our successful elucidation of the structure of the active center in 148. Cu( 11)-doped L-histidine hydrochloride mon~hydrate,~’ we chose (15) Bottcher, R.; Heinhold, D.; Wartewig, S.; Windsch, W. J . Mol. to study, in some detail, the ENDOR spectra of Cu(l1)-doped Struct. 1978, 46, 363. single crystals of a-glycylglycine at 4.2 K . (16) Fujimoto, M.; McDowell, C. A,; Takui, T . J . Chem. Phys. 1979, 70, Introduction
~~~
of the EPR and ENDOR Spectra Cu( 11)-doped a-glycylglycine contains different types of nuclei, most of which possess a nuclear spin. In our experiments, we concentrated on evaluating the magnetic parameters of the I4N nuclei and the protons, since we were particularly interested in the bonding of ligands by the Cu(l1) atom. To analyze the I4N and proton ligand ENDOR spectra we used the following spin Hamiltonian: % = He, + Hhf + H,,+ H , (1)
Analysis
+
8% = PB@ C(Ij-Aj.S - @NgN’B*Ij + Ij.Pj.Ij) (2) The first term, He,, represents the effect of the electronic Zeeman interaction of the external magnetic field B on the electron spins; the second term, Hhr,is the hyperfine interaction between
* Author to whom correspondence should be addressed. 0022-3654/90/2094-8 1 13$02.50/0
~
~~
~~
~~
~~
3694. (17) Calvo, R.; Oseroff, S. B.; Abache, H. C. J . Chem. Phys. 1980, 72, 760. (18) (a) McDowell, C. A.; Naito, A. J. Magn. Reson. 1981,45, 205. (b) Thuomas, K. A,; Lund, A. J. Magn. Reson. 1975, 18, 12. (19) Biittcher, R.; Metz, H.; Windsch, W. Chem. Phys. Lett. 1977,49, 148. (20) Bottcher, R.; Heinhold, D.; Wartewig, S.; Windsch, W. J. Mol. Struct. 1978, 46, 363. (21) Wartewig, S . ; Bottcher, R.; Windsch, W. Chem. Phys. 1981, 58, 21 I . (22) Bottcher, R.; Metz, H.; Windsch, W. J . Mol. Struct. 1982, 83, 31. (23) Bottcher, R.; Metz, H.; Windsch, W. Chem. Phys. 1985, 93, 137. (24) Bottcher, R.; Heinhold, D.; Windsch, W. Chem. Phys. Lett. 1979, 65, 452. (25) Bottcher, R.; Heinhold, D.; Windsch, W. Chem. Phys. 1985.93, 339. (26) Windsch, W.; Welter, M. Z . Naturforsch. 1967, 22, 1 . (27) (a) McDowell, C. A.; Naito, A.; Sha, K.; Sastry, D. L.; Wu, Y . J. Mol. Struct. 1989, 195, 361. (b) McDowell. C. A,; Sastry, D. L. J. Chem. Phys. 1988,89(5), 2697. (c) McDowell, C. A,; Sastry, D. L. J . Mol. Struct. 1989, 198, 8 5 .
0 1990 American Chemical Society
8114
The Journal of Physical Chemistry, Vol. 94, No. 21, 1990
>> Hp the eigenvalues to Since we have He, >> (Hhf Hnz) first order for a state (M,m) are given by E ( M , m ) = g@BM + K,m
+ Y2P~[3m2- [ ( I + 111
McDowell et al. TABLE I: Spin Hamiltonian Parameters of Cu(11)-Doped a-Glycylglycine from ENDOR Single-Crystal Spectra at 4.2 K
(3)
where M and m denote the electronic and nuclear magnetic ~.~~ quantum numbers, respectively. As in our earlier ~ o r k s 'we derived expressions from which the hyperfine interaction tensor, A, and the quadrupole coupling tensor, P, can be calculated, based on the treatment of Thuomas and Lund.lsb The resulting expressions derived from first-order perturbation theory provide a sufficiently accurate background for understanding the main features of the observed spectra as shown in our earlier publication~.'~,~~ To obtain more accurate values for the components of the A and P tensors, we used a least-squares fitting procedure to minimize the differences between the observed frequencies and those calculated utilizing the seventh-order perturbation treatment of the spin-Hamiltonian given in eq 2 following the procedure outlined earlier.L8*27.35 The quadrupole coupling tensor P is diagonalized to yield the principal values (Pzz> PYy> P X J ,and their direction cosines, the corresponding nuclear quadrupole coupling constant e2Qq/h = 21(2/ - l)Pzz9and the asymmetry parameter 7 = (Pxx
tensor
(28) Biswas, A. 8.; Hughes, E. W.; Sharma, B. D.; Wilson. J. N. Acta Crystallogr. 1968, 824, 40. (29) Hughes, E. W. Acta Crystallogr. 1968, 824, 1128. (30) Freeman, H. C.; Paul, G.L. Acta Crystallogr. 1970, 826, 925. (31) Griffin, J. F.; Coppens, P. J. Am. Chem. Soc. 1976, 97, 3496. (32) Kvick. A.; Karaghouli, R. A.; Koetzle, T. F. Acta Crystollogr. 1977. 633, 3796. (33) Kvick, A.; Koetzle, T. F.; Stevens, E. W. J . Chem. Phys. 1979, 7 1 ( 1 ) . 173. (34) Dalal. N . S . ; McDowell, C. A.; Srinivasan, R. Mol. Phys. 1974, 24, 417.
principal values
~
g
g,, ,g,
g,, A(63.6sCu) axx a,,? azl
A(I4Nl)
axx
avv a,,
P(14NI)
P,, PYy
P,, A(I4N2)
P(I4N2)
- PJ/f::. Experimental Section Our experiments were confined to the Cu( 11)-doped a-glycylglycine complexes. Glycylglycine crystallizes in several different forms. The crystal structure of a-glycylglycine was determined by Biswas et al.,28and corrected by Hughes.29 Freeman et al.jO carried out a neutron diffraction study on deuterated a-glycylglycine single crystals. A careful room temperature X-ray diffraction study was reported by Griffin and cop pen^.^' Later reported neutron diffraction studies performed Kvick et al.32-33 at 82 K. The crystals belong to the monoclinic P2,/c space group with four molecules in the unit cell. The cell dimensions at 82 K were r e p ~ r t e d ~to* ,be ~ ~a = 8.025 A, b = 9.543 A, c = 7.788 A, and @ = 106.5'. The a-glycylglycine molecules form a layerlike structure with extensive hydrogen bonding, including interlayer bonding. Single crystals of a-glycylglycine were grown from a solution of the compound in heavy water (D,O) which contained less than ~ . crystals were blue and that they were 0.5% of C U ( N O ~ )The the a-modification was confirmed by X-ray measurements. The crystal faces were also established by the X-ray method which showed that the most prominent were the [IOO] and [ l IO] faces. An orthogonal experimental axes system a*, b, c was chosen. This axis system was based on the lattice parameters of Griffin and cop pen^.^^ whereas Fujimoto et a1.12 chose an axis system based on the lattice parameters of Hughes.29 Both axes systems can, of course, readily be interconverted by using the appropriate transformation matrix. The ESR and ENDOR measurements were carried out using a home-built ENDOR spectrometer operating in the superheterodyne mode. A general description of the spectrometer has been given earlier.16s18.27i34 The ESR spectra were obtained by using a field modulation of 90-100 Hz, whereas for the ENDOR spectra a frequency-modulated rf field of 900-1000 Hz was used. A rotable Varian 9-in. magnet with an accurate circular degree scale was used as the static magnetic field. The strength of the static magnetic field was measured by a home-built proton magnetic resonance gauss meter. All spectra were recorded at 4.2 K with
direction cosines with respect to a* b c 0.613 -0.665 0.426 0.568 0.746 0.348 -0.549 0.029 0.835 0.757 -0.419 0.501 0.333 0.908 0.256 -0.562 -0.028 0.827 0.581 -0.244 0.777 -0.735 -0.252 -0.629 0.389 0.937 0.033 4.205 0.038 0.978 -0.810 0.556 -0.191 0.551 0.083 0.831
________
2.033 2.085 2.236 62.3 MHz 190.2 MHz 531.8 MHz 23.2 M H z 24.4 MHz 35.1 MHz 0.179 MHz 1.385 MHz -1.565 M H z
e2Qq/h = -3.13 MHz, q = 0.77 28.8 MHz -0.615 F.013 27.7 MHz 0.786 0.096 37.5 MHz -0.067 0.995 P,, 0.229 M H z -0.223 0.975 PLY 0.963 M H z -0.506 -0.106 P,, -1.192 MHz 0.833 0.197 axx ayy a,,
0.788 0.61 I -0.069 -0.01 I 0.856 0.517
e 2 Q q / h = -2.384 MHz, q = 0.62 Ni
1
v
MPz
--
I +_i
I
,
'
'4 Figure 1. ENDOR spectrum of a single crystal of Cu(I1)-doped a-glycylglycine at 4.2 K. The magnetic field Bo is in the a*c plane and directed 50' from the a* crystallographic axis. The arrow labeled up represents the position of the free proton transition.
the crystal in a rectangular microwave cavity, operating in the TElo2mode, and attached to a stainless steel waveguide tube immersed in a double-walled Pyrex glass dewar containing liquid helium which was covered by another Pyrex glass dewar containing liquid nitrogen.
Results and Discussion It is well-known that cupric ion can occupy interstitial positions in crystals of amino acids and like compound^^-'^^'^^^^'^^^^ to form complexes that exhibit well-defined EPR spectra. The single crystals of Cu(I1)-doped glycylglycine exhibited well-defined EPR spectra at 4.2 K. In our studies we did not find evidence for more than one type of complex, as reported by earlier workers;12 presumably this was because of differences in the concentrations of the Cu(I1) ions used in forming the complexes. To obtain the data necessary for the accurate determination of the magnetic parameters, the EPR spectra at 4.2 K were recorded at 5' intervals when the single crystal was rotated about the three chosen orthogonal experimental axes. The principal values and respective direction cosines of the 63,65Cug and A tensors were determined by using the least-squares fitting procedure described earlier;27*35 the results are recorded in Table I. We see from the data in Table I that there is good correspondence between the direction cosines of the respective principal value of the g tensor and the Cu hyperfine interaction tensor. It is also to be noted that the direction cosines of the largest principal value of the g tensor corresponds with that of the largest principal value of the A tensor for 63365Cu. This shows that the complex has a square planar ligand field, and the ground-state half-filled orbital of the Cu(I1) ion is mainly d&+; this, as will be seen later, is also supported by our theoretical (35) Byfleet, C. A.; Chong, D. P.; Hebden, J. A.; McDowell, C. A. J . Magn. Reson. 1970. 2, 69.
ENDOR Study of Cu( [[)-Doped a-Glycylglycine
vMHz--
The Journal of Physical Chemistry, Vof.94, No. 21, 1990 8115
-F' NP
Figure 2. ENDOR spectrum of a single crystal of Cu(1I)-doped a-glycylglycine at 4.2 K when the magnetic field Bo parallel to the c crystallographic axis. The arrow labeled up represents the position of the free proton transitions.
calculations of the components of the g tensor. There is satisfactory agreement between our results, and those of Fujimoto et al.,I2 for their b l type complex a t room temperature. This presumably means that there is no dramatic change in the electronic structure of this type of Cu(I1)-a-glycylglycine complex on cooling to 4.2 K. We did not, however, observe the presence of two equivalent I4N nuclei, and as will be seen shortly, our detailed ENDOR studies clearly establish that in the Cu(I1) complex that we studied there are two different 14N nuclei. Since the complex is formed with the Cu(1l) ion located interstitially, one I4N nucleus comes from a glycylglycine molecule in one layer, and a different type of bonded I4N nucleus located in an adjacent layer. In Figure 1 we show the ENDOR spectrum of a single crystal of Cu(l1)-doped a-glycylglycine at 4.2 K recorded for the case when the magnetic field Bo is parallel to the c crystallographic axis. The ENDOR spectrum for the same crystal for the orientation when the magnetic field is in the u*c plane, and oriented at 50° to the a* axis, is shown in Figure 2. It is clear from these spectra first that there are two sets of four ENDOR transitions which belong to two nonequivalent nitrogen nuclei N , and N,. This observation differs from the earlier EPR results which were interpreted12 as showing the presence of two equivalent N nuclei in the complex. I n addition, in the ENDOR spectra of Figure 2 one can clearly identify the presence of four proton transitions which we label PI,P,, P3,and P4.These latter transitions were readily distinguished from the N , and N2 transitions by the relative variations of their intensities as the microwave pumping field was varied as described in our earlier p u b l i ~ a t i o n s . ' ~ J * ~ ~ ~ The angular variations of the ENDOR transitions were recorded in steps of 2' as the crystals were rotated about the three orthogonal experimental axes u*, b, and c. The calculated angular variations are compared with experimental data for the two nitrogen nuclei N , and N,, in Figures 3 and 4. These figures show that for each nucleus N , , and N2, there are two magnetically inequivalent sites. The two sites are seen to become equivalent in the ca* plane, reflecting the monoclinic nature of the crystal for which b is the unique axis. The assignment of the nitrogen nuclei to an amino (NH,), or an amide (NH), group was made using the p/s ratio of the bonding orbitals of the two nitrogen nuclei. This ratio was calculated from the anisotropic and isotropic parts of the nitrogen hyperfine tensors by utilizing the atomic parameters described earlier.36 From the respective hyperfine interaction tensors (Table I) we estimate that N, has a p/s ratio of 4.45 while N2 has a p/s ratio of 3.44. It is to be expected that the p/s ratio of an amino nitrogen is higher than that of a peptide amide nitrogen since the bonding orbitals are mainly sp3 and sp2 for amino and amido nitrogens, respectively. In addition, it has been reported that p/s ratio of amino nitrogen coordinated to Cu(1l) are about 5.0.'8*27Thus we assign N , to be an amino nitrogen, while N, is taken to be a peptide amide nitrogen. This assignment is also supported by the finding that the quadrupole coupling constant s Q q / h for N , is greater than that for N, (see Table I). Moreover, the value we obtain for the quadrupole (36) Morton, J. R.; Preston, K. F. J. Mogn. Reson. 1978, 30, 577.
ROTATION ANGLE (DEGREES) NI NITROGEN
Figure 3. Angular variations of the ENDOR transitions for the I4NI nucleus in Cu(I1)-doped single crystals of cy-glycylglycine at 4.2 K as the crystal is rotated in the a*b, bc, and ca* crystallographic planes. The solid lines through the experimental points are calculated from a computer program based on seventh-order perturbation t h e ~ r y . ' * ~The ' ~ line labeled up is the angular variation of the free proton transitions.
$w1
$
a*
30
60
30
60
30
60
a*
ROTATION ANGLE (DEGREES) N2 NITROGEN
Figure 4. Angular variations of the ENDOR transitions for the I4N2 nucleus in a Cu(I1)-doped single crystal of a-glycylglycine at 4.2 K as the crystal is rotated in the o*b, bc, and co* crystallographic planes. The solid lines through the experimental points are calculated from a computer program based on seventh-order perturbation The line labeled up is the angular variation of the free proton transitions.
coupling constant for N,, namely -3.13 MHz (Table I) is similar to that given by Wartewig et aL2' for a N nucleus with similar bonding in Cu(I1)-doped triglycine sulfate. There is now general agreement about the common geometry assumed by the different Cu(1I)-ligand The location of the Cu(l1) nucleus in the present complex was estimated in the following way. The maximum component of the dipolar coupling tensor a,, (see Table I) of nuclei N , and N 2 indicates the direction in which bonding orbital of these nitrogen nuclei is oriented. The direction cosines of a,, are (0.389,0.937, 0.033) for N , and (-0.067, 0.995, -0.0696) for N2. These show that the lone-pair orbital of these two nitrogen nuclei are directed approximately along the b axis. Various N,-N, vector direction cosines were calculated. The direction cosines of the N,-N, vector in which N , belongs an a-glycylglycine molecule a t the (X,Y, Z ) symmetry location (I), and N2 belongs t o an a-glycylglycine U, I / , - Z ) symmetry location (III), are molecule at ( X , (-0.363, 0.931, -0.0405). These values are in close agreement with the observed direction cosines of a,, for both N , and N2. From this we deduce that the Cu(1I) nucleus is located between the N , nucleus of the molecule at an I symmetry position, and the N2 nucleus of a molecule at a 111 symmetry position. The
+
8116
The Journal of Physical Chemistry, Vol. 94, No. 21, 1990
McDowell et al. r
r
Figure 7. ORTEP diagram showing the orientations of the quadrupole coupling tensor of the I4NI(NH2)and I4N2(NH)nuclei with respect to the molecular frame of Cu(I1)-doped a-glycylglycineat 4.2 K. Note that the principal values of the respective quadrupole coupling tensors had to be taken as positive in calculating the orientations of the tensorial ellipsoids. Note that the quadrupole coupling tensors are from N, and N2 on different molecules. They are shown together here merely to indicate the relative orientations of the two tensors. a*
30
60
b
30
60
30
60
(I*
ROTATION ANGLE (DECCREES) PROTON RESONANCES
Figure 5. Angular variations of the ENDOR transitions for the four
protons PI, P2, P3, and P4 in a Cu(l1)-doped single crystal of a-glycylglycine at 4.2 K as the crystal is rotated in the a*b, bc, and ca* crystallographic planes. The curve labeled vp represents the angular variation of the free proton transitions. 1
TABLE 11: Hyperfine Parameters of Observed Protons in Cu(11)-Doped a-Glycylglycine Single Crystals Studied by Spectroscopy at 4.2 K
principal . . values, MHz Droton PI
MHz aXx
ayy a,,
P2
uXx ayY uZ2
P,
aXx
ayy
a,,
P,
uXx
ayy a,,
\-
Figure 6. ORTEP diagram showing the orientations of the hyperfine interaction tensors of the l4NI(NH2)and I4N2(NH)nuclei with respect to the molecular frame of Cu(I1)-dopeda-glycylglycineat 4.2 K. Note
that the principal values of the respective hyperfine interaction coupling tensors had to be taken as positive in calculating the orientations of the
tensorial ellipsoids. Note that the hyperfine interaction tensors are from N , and N2 on different molecules. They are shown together here merely to indicate the relative orientations of the two tensors.
Cu(ll)-N2 direction cosines are almost parallel to the b axis, whereas the Cu(I1)-N, direction cosines indicate that the Cu(II)-N, vector is perpendicular to the c axis, but slightly deviates from the b axis in the ab plane. These considerations lead us to assume that the position of the Cu(I1) nucleus is between the two glycylglycine molecules situated at type I and Ill crystallographic symmetry locations. Using the coordinates of the Cu(I1) ion thus determined, we find the Cu(lI)-N, distance is 2.3 A and the Cu(II)-.N2 distance is 2.1 A. The Cu(l1) nucleus is located in the middle of a distorted square planar atomic configuration consisting of N I (I), N, (11), O1 ( I ) , and O2 (111). All four atoms do not, however, fall on the same plane. It is interesting to note the orientation of the quadrupole coupling and hyperfine tensors of N , and N, with respect to the molecular geometry. The ORTEP diagrams showing the orientation of these tensors are given in Figures 6 and 7. These representations show how the unpaired electron distribution around these nuclei is related to the total charge distribution. The uzr value of both the N , and N 2 nuclei is directed along the Nl-N2 vector. The quadrupole coupling e2Qqr2of N , is directed along the NI-N2 direction making an angle of 12-3', whereas the quadrupole coupling e2Qqzzof N2 is directed along C2-C3 being nearly coplanar with the C,-N,-C, plane. The
direction cosines with respect to
Amism
3.94 -1.97 4.81 -1.10 3.06 8.97 1.49 -0.97 2.39 -0.07 3.51 1.05 -2.59 -2.71 -3.50 -3.62 6.44 6.32 -3.46 -1.66 4.90 6.73 -6.94 -5.14
a*
b
ENDOR Amism
c
MHz
-0.169 -0.055 0.984 -1.73 0.985 0.008 0.169 -1.73 -0.018 0.998 0.053 3.47 -0.245 -0.883 0.401 -0.72 -0.367 0.467 0.804 -0.72 0.897 0.050 0.439 1.44 0.066 0.996 0.057 -3.82 0.977 -0.053 -0.206 -3.82 0.202 0.069 0.978 7.63 -0.843 0.423 -0.332 -2.81 0.283 0.873 0.395 5.62 -0.457 0.239 -0.857 -2.81
quadrupole coupling e2Qqyyof N , makes an angle of 19.5' with the N I X , direction, whereas, 2QqYyof N2 is directed nearly along the N,-N2 vector and is almost perpendicular to the C2-N2-C3 plane. The quadrupole coupling e2Qqxxof N , is in the direction of the 03-NI vector (making an angle of 13.2'), whereas in the case of N2, it is directed nearly along the bisector of C2-N2-C3 bond angle. Knowing the position of the copper nucleus, and the various Cu(I1)-proton distances, we calculated the direction cosines to help identify the possible protons interacting with the copper complex. There were proton resonances clearly identifiable in the ENDOR spectra. Each proton gives rise to two resonances equidistant on either side of the free proton resonance frequency. A total of four distinct protons each at two magnetically inequivalent sites could be identified (see Figures 1 and 2). Even though all the resonances could not be observed at all angles through which the crystal was rotated, the least-squares orientations fitting procedure could be used to determine the hyperfine parameters. Only one site is shown in the angular variation plot of the proton resonances in Figure 5. The principal values of their hyperfine interaction tensors and the direction cosines were determined; and these proton are designated as P I , P2, P3, and P4 in Figure 5. The data are given in Table I1 together with estimates of values of the principal values of the hyperfine interaction tensors deduced from the crystallographic data. In calculating the theoretical values for the hyperfine interactions from the structural information, the simple point dipole approximation was assumed to be valid. The contribution of orbital angular momentum about the metal nucleus to the hyperfine coupling of the protons should not be significant as the anisotropy of the g tensor is small. This model assumes that the unpaired electron is largely lacalized on the metal atom. This model is thought to be valid since the locations of the interacting protons are quite distant from cupric ion. Otherwise, distribution of the unpaired electrons has to be
ENDOR Study of Cu(I1)-Doped a-Glycylglycine
The Journal of Physical Chemistry, Vol. 94, No. 21, 1990 8117
TABLE Ill: Identification of Protons That Interact with the Cu(1l) Nucleus in ENDOR Spectra of Cu(I1)-Doped a-Glycylglycine Single Crystals
proton H4 H, H7
H,'
crystallogr sym element
X Y X Y X Y X '/2
Z
Z Z
-Y
'/2
+
direction cosines' with respect to assigned a* b c proton 0.237 0.887 -0.397 P4 0.1 14 0.994 -0.004 Pi 0.911 0.373 0.17 P2 P3 Z 0.262 -0.083 0.962
8
"Calculated from the crystallographic data. Note: H4 and H,' refer to protons in different molecules in adjacent layers in the crystal.
considered. This model has been used successfully in the literat ~ r e ~to identify ~ . ~ the ~ +interacting ~ ~ ~ protons in crystals that had been studied by ENDOR. It should be mentioned that Bottcher et al.i4923also detected resonances from four protons in their ENDOR spectra of Cu( 11)-doped single crystals of triglycine sulfate. These were assigned to the methylene protons, but their locations in the crystal lattice were not specifically identified. Bottcher et al.24reported the identification of proton resonances in the ENDOR spectra of V02+-doped triglycine sulfate which they assigned to two protons from each of the amino groups in the complex. Other proton ENDOR resonances were also detected which were thought to arise from the C H 2 groups. The well-known work of Hutchison and M c K first ~ ~described ~ ~ the method used above to locate four water of crystallization protons in single crystals of lanthanum nicotinate dihydrate, doped with Nd(ll1) ions, using data from their impressively detailed analysis of the proton ENDOR spectra. It is of interest to note that those authors point out that it should be possible, with precise crystal orientation and high-resolution ENDOR data for all the crystal orientations, to estimate the proton coordinates with a precision of =O.OOl A. The components of the calculated proton hyperfine interaction tensors A'depend on the model of the Cu(I1) complex and its environmental geometry, the reasonable assumption that the anisotropic part of the electron-proton interaction is dipole-diople and the crystal locations of the particular protons. Bottcher et aLZ4employed essentially the same metod to identify the proton resonances observed in the ENDOR spectra of VOz+-doped triglycine sulfate. A method similar to that employed by Hutchison and M c K was ~ later ~ ~ used ~ by Atherton and Shackleton3**to identify the lattice positions of protons whose interactions were observed in the ENDOR spectra of the VO(H20)!2+ ion. It is of interest to note that Schuff and HaeberledEb have likewise used a similar technique to determine the proton sites in a single crystal of malonic acid. These authors carried out two-dimensional NMR studies on m:.!onic acid single crystals and identified the resonances for different protons in the crystal. They were thus able to determine the dipolar interactions of specific nuclei. Comparisions of the dipolar interactions calculated from crystal structure data with the experimental values enabled proton sites to be located in the crystal lattice. In Table I I and I11 we compare the experimental values of the hyperfine interaction parameters for the four protons PI, P2, P3, and P, with values calculated from crystallographic data, and the expected direction cosines, of certain protons in the crystal (see Figure 8). These data were chosen from similar calculations for IO protons in the Cu complex environment as being those most likely to represent the interacting nuclei for which we observe the proton resonances in our ENDOR spectra. It is clear from comparing these data in Table 111 with those in Table I1 that we can identify the crystal sites of H,, H5,H,,and Hq/ in Figure 8 with the four protons P,, Pi, P,, and P3, respectively, observed in the ENDOR spectra. This identification of four proton lattice sites in the crystal as being those that interact with the Cu(11)-doped a-glycylglycine complex, taken together with the earlier work,i4*37q38 is an interesting demonstration of the unique spatial ( 3 7 ) Hutchison, C. A. Jr.; McKay, D. B. J . Chem. Phys. 1977, 66, 331 1. (38) (a) Atherton, N . M.; Shackleton, J . F. Mol. Phys. 1980, 39, 1471. (b) Schuff, N.; Haeberlen, U.J . Magn. Reson. 1983, S2, 267.
Figure 8. Diagram showing the crystal structure of a-glycylglycine in the a*b plane, the proposed location of the Cu(I1) nucleus, and its spacial relationships to the interacting I4NIand 14N2 nuclei as well as several adjacent protons as exhibited in the ENDOR spectra. The positions of the various nuclei are derived from crystallographic
information that can be obtained from careful detailed ENDOR studies on single crystals. Theoretical Calculation of the g Tensor. Theoretical calculations of the g tensors of Cu(1I) complexes derive from the classical work of Abragam and P r y ~ e . ~The ~ l adaption of that approach to ligand field theory was successful for many d' and d9 systems. Recently, it has been shown42that good results are obtained when the complete energy matrix, including the ligand field, electron-electron repulsion, and the spin-orbit coupling are diagonalized together. In an earlier p ~ b l i c a t i o n we , ~ ~outlined a modification using the angular overlap model (AOM) to evaluate the ligand-field matrix elements. That method was successfully applied to calculate theoretically the principal value of the g tensor of Cu(I1)-doped L-histidine HCI m o n ~ h y d r a t e . ~ ~ The present calculations were made within the complete 2D free-ion terms as a basis, perturbed by spin-orbit coupling and the ligand field. The single-electron spin orbitals denoted by 0, 6 , p , u, and 7 (related to symmetry-adopted read orbitals) were used as basis vector of the 1.s matrix. The spin-orbit matrix 1-s is given in earlier publication^^^*^^ Assuming that the Cu(I1)doped site has D2,,symmetry, the AOM ligand-field matrix elements are
+ 0.5eU(N)
(017fla) = 0.5eU(0)
( c l 7 f l t ) = 1 .5e,(0) (c(17fIp) =
+ 1.5eU(N) + 2e,(N)
( v l 7 f l v ) = 2eAN) (?I7fl7) = 2eAO)
(ww) = v ' 3 / 2 [ e , ( ~ ) - e,(0)1
(4)
In eq 4 the angular overlap parameters written as e,(O) and e,(N) refer to u bonds between Cu and oxygen and nitrogen atoms, while e,(O) and e,(N) stand for a-bonding interactions. The values of e, for N and 0 were varied independently from 4000 to 12000 cm-I, and the ratio e,(O)/e,(O) was taken as 0.5. Values of the free-ion spin-orbit interaction ranging from 550 to 850 cm-I (39) Pryce, M . H. L. Proc. SOC.1950, 63, 25. (40) Abragam, A.; Pryce, M. H. L. Proc. R . SOC.(London), Ser. A 1951, 205, 135. (41) Abragam, A.; Pryce, M. H. L. Proc. R. SOC.(London), Ser. A 1951, 206, 164. (42) Lin, W. C. J . Magn. Reson. 1986, 68, 146.
J . Phys. Chem. 1990, 94, 8 1 18-8124
8118
TABLE IV: Experimental and Calculated Values of the g Tensor of C u ( l I ) - h w d Sinele Crvstals of a-Glycylglycine at 4.2 K E", gYY
gn
The g-tensor principal values were calculated from the equations
experimental
calculated 2.040
2.083 2.236
2.033" 2.085"
2.059b
2.236O
2.273b
2.022b
were assumed. To evaluate the energy matrix the AOM orbital parameters given in refs 43 and 44 were used. The total energy matrix consisting of the ligand field and the spin-orbit coupling (there is no electron-electron repulsion to consider in the dl and d9 cases) was diagonalized to yield eigenvalues and eigenvectors. The spin-orbit interaction matrix is given in Table I of ref 43.
The calculated principal values of the g tensor are compared with our experimental data in Table IV. There is very good agreement between the experimental values and those calculated from the modified Lin method42which we adapted to encompass the AOM approximations in our earlier The coefficients of the eigenfunctions of the Kramers doublet showed that the free electron in the complex is mainly in a d$+ orbital. This agrees with the original assignment proposed by Fujimoto et a1.I2
(43) Lever, B. P. Inorganic Elecfronic Spectroscopy, 2nd ed.; Elsevier: Amsterdam, 1984; pp 53, 67. (44) Gerlock, M.; Slade, R. C. Ligand-Field Parameters; Cambridge University Press: Cambridge, U.K.; 1973.
Acknowledgment. We thank the National Sciences and Engineering Research Council of Canada for grants to C.A.M. to support this research. K.S. and Y.C. thank Jilin University and Nankai University. respectively, for leaves of absence.
"This work.
Fujimoto et a l l 2
Electronic Relaxation in Long Polyenes Sarah A. Cosgrove, Melissa A. Guite, Timothy B. Burnell, and Ronald L. Christensen* Department of Chemistry, Bowdoin College, Brunswick, Maine 0401 1 (Received: April 26, 1990)
A series of carotenols with from 7 to 1 1 conjugated double bonds have been synthesized and purified by using HPLC techniques. Absorption, fluorescence, and fluorescence excitation spectra have been obtained in 77 K glasses. The shorter members of this series exhibit the Stokes-shifted, SI So emissions seen in previous studies of model polyenes. For carotenols with more than eight conjugated double bonds, however, the fluorescence is dominated by "anti-Kasha", S2 So fluorescence. These findings in part can be attributed to a larger S2-SI energy difference and the resultant decrease in S2 SI radiationless decay rates in longer polyenes. The precipitous crossover from S, So to S2 So emission, however, cannot be fully accounted for by thc energy gap law, which predicts only modest changes in radiative and nonradiative decay rates with increasing polyene length. The lack of SI fluorescence prohibits the direct observation of the SI state in the longer carotenols. Nevertheless, trends noted in the shorter polyenes indicate a 5500-6500-cm-' S2-Sl energy difference for polyenes such as @-carotene ( 1 1 conjugated double bonds). The implications of large S2-SIenergy gaps for the spectroscopy and photochemistry of @carotene and other long polyenes also are discussed.
-
-
Introduction The electronic spectroscopy and excited-state dynamics of linear polyenes continue to be topics of considerable interest. Recent experiments range from time-resolved studies of cis-trans isomerization in substituted butadienes,' to transient Raman spectroscopy of carotenoids bound to membranes of photosynthetic bacteriaT2to investigations of the electronic properties of polyacetylenes.' These studies encompass a wide range of experimental techniques and theoretical models, and it is not surprising that the relationships between polyenes of different conjugated lengths largely remain unexplored. A common understanding of linearly conjugated systems remains a goal of active interest. This investigation contributes toward this effort by establishing links between the electronic properties of short, model polyenes and the carotenoids, which play important roles in photobiology. A brief review of previous work on polyenes establishes several distinct areas of interest. Spectroscopic and kinetic investigations of short diphenylpolyenes (from stilbene to diphenyl( 1 ) Lee, M.; Haseltine, J . N.; Smith 111, A. B.; Hochstrasser. R. M. J . Am. Chem. Soc. 1989, 1 1 . 5044. (2) Kuki, M.; Hashimoto, H.; Koyama, Y. Chem. fhys. Lett. 1990, 165, 417-421. (3) Yoshizawa, M.; Kobayashi. T.; Fuiimoto. H.; Tanaka, J.; Shirakawa, H. J. Lumin. 1987. 38, 300-304
-
- -
~ctatetraene)~-l I have exploited their commercial availability, stability, high fluorescence yields and relative ease of placement in supersonic jets. Short, unsubstituted polyenes, on the other hand, are not as stable and most notably are nonfluorescent. This has severely hampered their study, and considerably more attention has been focused on unsubstituted or methyl-substituted polyenes of intermediate l e r ~ g t h . l ~ These - ~ ~ compounds combine at least (4) Shepanski, J. F.; Keelan, B. W.; Zewail, A. H. Chem. fhys. Lett. 1983, 103, 9. (5) Kohler, B. E.; Spiglanin, T. A. J . Chem. Phys. 1984,80, 5465-5471. (6) Kohler, 9. E.; Spiglanin, T. A. J . Chem. f h y s . 1985, 82, 2939. (7) Troe, J.; Amirav, A.; Jortner, J. Chem. fhys. Lett. 1985, 115, 245. (8) Horwitz, J. S.; Kohler, B. E.; Spiglanin, T. A. J . Chem. fhys. 1985, 83, 2186. (9) Horwitz. J. S.; Kohler, B. E.; Spiglanin, T. A. J . Phys. Chem. 1985,
89, 1574-1576. ( I O ) Amirav, A,; Sonnenschein, M.; Jortner, J . J . Chem. Phys. 1986, 102, 305. ( 1 1) Itoh. T.: Kohler. B. E. J . f h v s . Chem. 1988. 92. 1807-1813 (l2j Gavin, Jr., R. M.; Weisman, C ~McVey, ; J . K.; Rice, S . A. J . Chem. f h y s . 1978, 68, 522-529. (13) Granville, M. F.; Holton, G. R.; Kohler, B. E.; Christensen, R. L.; D'Amrco, K. L. J . Chem. fhys. 1979, 70,593. (14) D'Amico, K. L.; Manos, C.; Christensen, R. L. J . Am. Chem. SOC. i 9_ ~_ n_ i n_ 1 ,_1777 _ _ . ., .
(15) Heimbrook, L. A,; Kenny, J . E.; Kohler, B. E.; Scott. G. W. J . Cbem. Phys. 1981, 75, 4338-4342. (16) Hudson, B. S.; Kohler, B. E.; Schulten, K. In Excited Stares; Lim, E.C.. Ed.: Academic Press: New York, 1982; Vol. 6, pp 1-95.
0022-3654/90/2094-81 18$02.50/0 (C I990 American Chemical Society
