EPR Study of
p-Nitrophenol Adduct of
The Journal of Physical Chemistry, Vol. 83, No. 11, 1979
Cu[en(sal),]
1391
An Electron Paramagnetic Resonance Study of the p-Nitrophenol Adduct of N ,"-Et hylenebis(salicylideneiminato) copper (II)$ Carl E . Tauber and Harry C. Allen, Jr.*+ Department of Chemistry, Jeppson Laboratory, Clark University, Worcester, Massachusetts 0 16 10 (Received October 2, 1978) Publimtion costs assisted by the U.S. Bureau of Mines
The EPR and optical spectra of the p-nitrophenol adduct of N,N'-ethylenebis(salicylideneiminato)copper(II) have been investigated. The values of the parameters in the spin Hamiltonian are found to be g, = 2.192 f cm-', A, = 36.3 f 3.0 X cm-', A, 0.001., g, = 2.048 f 0.002, g = 2.043 f 0.002, A, = 201.9 f 1.0 X cm-', AYN= 14.2 f 0.6 X cm-l, AxN = 15.3 f 0.6 X cm-l, A ) = 12.8 f 0.6 X = 29.6 f 3.0 X cm-'. The AIH are assumed to be one-half the values of AIN on the basis of the hyperfine structure behavior of the parent complex. The spin-Hamiltonian parameters are very close to the values found for the parent d,, transition complex. The x values seem to reflect the effects of the hydrogen-bonded adduct. The d,z-,z is found at 18 500 cm-l.
-
Introduction During the past several years a series of tetradentate Schiff-base complexes of copper(I1) have been studied in this 1aboratory.lJ The general formulas of some of these complexes are given in I-IX. Several of these complexes
I.
R = R " = CH3
R ' = -CH2-CH2-
Cu en(acac)2
R = C F ~R " = C H ~
R' = -CH~-CH~-
Cu en(tfacac)2
111.
R = C2H5 R " = CH3
R ' = -CH2-CH2-
Cu
IV.
R = C2H5 R" = CF3
R ' = -CH2-CH2-
Cu e n ( t f h e x )
V.
R = C2Hg R " = CF3
R ' = -CH
11.
VI.
-
2 Cu 1 , 2 - p n ( t f h e x ) 2
P!H3
R = H
R ' = phenylene
Cu p h e n ( s a l ) 2
VII.
R = CH
R'
Cu p h e n ( a c e t ) 2
VIII.
R = H
R ' = -CH2-CH
R = CH3
R ' = -CH2-CH2-
IX.
3
= phenylene
2
-
Cu e n ( s a l ) 2
Cu en(acet)
2
form adduct^^,^ with chloroform in which the chloroform is hydrogen bonded to one of the oxygen atoms bonded to the Cu(I1). Two of these adducts have beer1 the subject of EPR studies even though they readily lose chloroform upon exposure to air.3 An analysis of the isotropic hyperfine interaction for the series of thirteen related complexes1 shows that the value of
x
= (4x/S)(~ICG(r,)S,il~) i
(1)
is negative and remarkably constant. However, it was Supported lcly the U.S. Bureau of Mines under Grant G0166203. 0022-3654/79/2083-1391$01 .OO/O
noted that the values of -x for the two chloroform adducts studied, C u [ e n ( a ~ e t )and ~ ] Cu[phen(~al)~] are significantly higher than the average of the -x values for the series. In order to ascertain if this is a real effect on the isotropic hyperfine constant we have undertaken an EPR study of the p-nitrophenol adduct of C~[en(sal)~]. This 1:l adduct is readily formed and is stable in chloroform solution and in air. The p-nitrophenol is hydrogen bonded to the ligand oxygen in the same manner as in the chloroform adduct^.^ The sample studied was a single crystal of the p-nitrophenol adduct of N i [ e n ( ~ a l )into ~ ] which about 1mol % of 63Cu[en(sa1)Z]was doped.
Experimental Section The synthesis of the 63Cu-doped Ni[en(sal),] adduct follows that of previous worker^.^^^ 63Cu0 (98.5%) was obtained from Oak Ridge National Laboratories and converted to the more soluble nitrate before use. Reagent grade nickel acetate tetrahydrate, salicylaldehyde, and ethylenediamine were used without further purification. Reagent grade p-nitophenol was recrystallized from benzene before use. The mole ratio of nickel to copper was about 1OO:l in the reagent mixture. This was found to yield a 63Cudoped adduct of sufficiently high copper concentration to give an acceptable signal-to-noise ratio, without obscuring ligand superhyperfine information. Single crystals of the copper doped adduct were grown by slow evaporation from chloroform solution. This yielded rod-shaped crystals of deep red-brown color. A crystal, whose approximate dimensions were 1mm X 2 mm X 5 mm, was chosen for the single crystal study. Powder spectra were obtained from finely ground single crystals. Instrumentation. EPR spectra were measured a t Xband with a Varian E-9 spectrometer, equipped with a Varian E-231 rectangular cavity and a Varian E-229 single axis goniometer. A Magnion G-502 NMR gaussmeter was used to calibrate the spectra and a Hewlett-Packard Model 5245L frequency counter, with a Model 5255A frequency converter, was used to measure the frequencies of the gaussmeter-oscillator, and the klystron. K-band spectra were obtained with an instrument built a t Clark.7 These spectra were used to aid in the understanding of the hyperfine structure; all numerical values were measured from the X-band spectra. Optical and near-infrared data were obtained from a Cary Model 14 spectrophotometer. Spectra of the pure copper adduct 0 1979 American
Chemical Society
The Journal of Physical Chemistry, Vo/. 83,
1392
No.
C . E. Tauber and H. C . Allen
11, 1979
TABLE I: Spin-Hamiltonian Parameters
TABLE 11: Comparison of MO Coefficients Cu[en(sal),] Cu[en(sal),].pnp
g, gx gY
AZb AX
A AYN ,
4; AYH
2
H
A E , cm-' a
2.192 i 0.002 2.049 i 0.004 2.046 i 0.004 201.0 i 1 . 0 31.3 i 3.0 29.3 i 3.0 12.6 i 0.6 15.7 i 0.6 14.5 i 0.6 6.3 i 0.3 7.9 i 0.3 7.3 i 0.3 17900 f 300
From ref 1 and 3.
2.192 i 0.001 2.048 t 0.002 2.043 i 0.002 201.9 f 1.0 36.3 i 3.0 29.6 i 3.0 12.8 i 0.6 15.3 f 0.6 14.2 i 0.6 6.4 i 0.3 1 . 7 i: 0.3 7.1 i: 0.3 18500 i 300
CYz
PZ nz -X
TABLE 111: Comparison of Hydrogen Bonded Adducts
Results The parent complex crystallizes in centrosymmetric dimers yielding two magnetically nonequivalent sites.12J3 Baker et have shown that the addition of an equimolar amount of p-nitrophenol causes the formation of a oneto-one adduct, in which the phenolic proton of p-nitrophenol is hydrogen bonded to one of the ligand oxygens. The space lattice is triclinic with two molecules per unit cell. The dimeric character of the parent is lost, and the chelate molecules stack plane-to-plane a t 3.5-A intervals
x Values
compd Cu[ en( acet), ] Cu[phen(sal),] Cu[en(sal),],pnp average for 11 other chelates
A values in units of cm-' x l o 4 .
were obtained from KBr pellets. Single crystal data were obtained after the method of Hathaway and Billing.8 The crystal was mounted on a quartz rod with flat faces ground parallel and perpendicular to its axis. To obtain spectra about three orthogonal axes, the rod-shaped crystal was mounted with its long axis successively along the three axis of the quartz rod. The orientational conventions8 inherent in the Geusic and Browng method of data reduction were adhered to. Spectra were recorded in 5" increments over a range of a t least 200' for each of the crystal orientations. The general features3 of the spectra are the same as those observed for Cu[en(sa12)]. Each transition gives rise to four sets of eleven lines. The four sets arise from the coupling of the unpaired electron with the copper nuclear spin ( I = 3/2), and the eleven-line pattern of each set arises from the coupling of the unpaired spin with the nuclear spin of the ligand nitrogen atoms and ligand hydrogen atoms in a now familiar manner.3J0 Data Reduction. The single crystal data were reduced by the method of Geusic and Browngas modified by Billing and Hathaway'l to obtain the principal g values and the direction cosines between the laboratory axes and the magnetic axes of the crystal. Unfortunately the principal values of the copper hyperfine tensor, Ai, could not be obtained in this manner due to second-order effects. T o obtain the A values advantage was taken of the fact that the principal axes of the A tensor seemed to be the same as those for the g tensor, a situation that exists in the other related chelates that have been studied.lS2 Thus the crystal was remounted through the use of the direction cosines of the g tensor so that the principal A values were observed directly. The values of g, and A, obtained in this manner agreed with the values obtained from an analysis of a polycrystalline sample. In addition the g values obtained in this manner agreed with those obtained from the more general method. The hyperfine interactions due to the nitrogen and hydrogen nuclei were measured directly also. Thus the parameters in the spin Hamiltonian were determined. The results are presented in Table I together with the results of the measurement of the optical spectrum.
0.78 0.30 0.13 0.67 3.93
0.78 0.31 0.76 0.65 3.98 in -X
4.00 4.00 3.98 3.90
i
0.04
along the a axis, with p-nitrophenol molecules located between the chain^.^ The effect of adduct formation is, in this case, to reduce both the overall crystal symmetry, and the molecular symmetry. In certain orientations of the crystal, the eleven-line packets distort, and split into twelve or more lines. This seemed to be more severe near the planes of the chelate molecules, and raised the possibility of two magnetically nonequivalent sites. An attempt made to resolve the question by taking spectra at K band (ca. 22 GHz) was not successful. It is very likely that this distortion of the packets arises because the molecular planes are rotated slightly with respect to each other leading to two slightly nonequivalent sites. This nonequivalence will be most important when the field is near the molecular plane. This slight nonequivalence together with the anisotropy of the ligand hyperfine structure has been shown to give rise to patterns such as observed here.14 This slight perturbation did not affect the values of the spin-Hamiltonian parameters within the accuracy quoted in Table I. The molecular orbital coefficients have been calculated using the method outlined in ref 1. The single optical transition observed a t 18500 cm-l is assigned to the E,z+ E,, transition. Other d-d transitions could not be observed due to interference from strong ligand and/or charge transfer transitions. In calculating a value for x it has been assumed that the out-of-plane R bonding is negligible. The values for these parameters are given in Table 11. The molecular orbital coefficients are similar to those found in the parent compound which are included in the table for comparison. The results show p2 somewhat larger for the adduct than for the parent compound indicating somewhat less in-plane T bonding. The coefficients for the u-bonding orbital are essentially the same for both the parent compound and the adduct. The optical transition has been shifted to higher frequency. The value of x is more negative than that for the parent compound.
-
Discussion
Experience suggests that the spin-Hamiltonian parameters are sensitive only to changes in the immediate environment of the Cu(I1). Although there are small variations in the g and A values as well as the optical transitions, the differences are not large unless there are large differences in the geometry about the copper It may be that x is a more sensitive indicator of small differences in the Cu environment. In Table I11 the x values for the three known hydrogen-bonded adducts are compared. One can readily see that for the three adducts whose structure has been confirmed by X-ray studies, -x = 4.00 within any reasonable experimental uncertainty.
Solvent Effect on 'F Chemical Shifts
An average value of --x= 3.90 f 0.04 is found from 11other related chelates that are not adducts. The uncertainty quoted in the average is the standard deviation from the mean. Thus we have more evidence to support the hyDothesis that small differences in the immediate environment of the copper(I1) ion may be reflected in the x value. Hydrogen bonding to the ligand oxygen may not be the only parameter affecting the value of x for there is evidence that x is also affected by the amount of tetrahedral distortion about the Cu(II), an effect that is still under investigation,,15The increase in frequency of the d-d transition would seem to rule out any significant tetrahedral distortion since the shift is in the wrong direction. Acknowledgment. The authors thank Dr. M, 1,Scullane for in the synthesis and crysta1 growth phases Of this work, and Professor R. s. Andersen for the use of his
The Journal of Physical Chemistry, Vol. 83,
No. 11, 1979 1393
K-band equipment. A grant from the Research Coporation is gratefully acknowledged.
References and Notes H. C. Allen, Jr., and M. I. Scullane, J . Coord. Chem., 8, 93 (1978). M. I . Scullane and H. C. Allen, Jr., J . Coord. Chem., 8, 87 (1978). M. I. Scullane and H. C. Allen, Jr., J . Coord. Chem., 4, 255 (1975). E. N. Baker, D. Hall, and T. N. Waters, J. Chem. SOC.A , 406 (1970). (5) E. N. Baker, D. Hall, and T. N. Waters, J. Chem. SOC.A , 406 (1970). (6) M. I. Scullane, Ph.D. Thesis, Clark University, 1976. (7) D. Close, Ph.D. Thesis, Clark University 1973. (8) B.J. Hathaway and D. E. Billing, Coord. Chem. Rev., 5, 143 (1970). (9) J. E. Geusic and L. C. Brown, Phys. Rev., 112, 64 (1958). (10) A. H. Maki and B. R. McGarvey, J . Chem. Phys., 29, 35 (1958). (11) D. E. Billing and B. J. Hathaway, J . Chem. Phys., 50, 2258 (1969). (12) D. Hall and T. N. Waters, J. Chem. Soc., 2664 (1960). (13) L. M. Shkol'nikova, E. M. Yumal, E. A. Shugam, and A. Voblikova, Z. Strukt. Khim., 11, 886 (1970). (14) M. I. Scullane and H. C. Allen, Jr., J . Coord. Chem., in press. (15) H. C. Ailen, Jr., and D. J. Hodgson, unpublished. (1) (2) (3) (4)
Use of Fluorine-19 Chemical Shifts to Measure Deviations from Random Mixing in Binary Solutions Near the Consolute Temperature. The System Hexane/Perfluorohexane Norbert Mullert Department of Chemistry, Purdue University, West Lafayette, Indiana 47907 (Received March 6, 1978)
The 19Fchemical shifts of the trifluoromethyl groups have been measured as a function of composition in hexane/perfluorohexane mixtures at 25,35,45, and 55 "C. The lowest temperature is just above the reported upper consolute temperature of 22.65 "C. Differences between the observed shifts and values anticipated for hypothetical ideal solutions are ascribed to the known volume increase on mixing and to the persistence of short-range order, with 1-1 and 2-2 contacts preferred over 1-2 contacts. Interpreted in this way, the data at each temperature provide a means of evaluating the quantity w/z appearing in Guggenheim's quasi-chemical treatment and equal to half the energy cost of converting a 1-1 pair and a 2-2 pair into two 1-2 pairs. The effects of imperfectly random mixing found at 55 "C are roughly 60% as large as those at 25 "C.
Introduction Although there is as yet no theoretical treatment which yields accurate predictions of solvent effects on fluorine NMR chemical shifts, significant information about interactions between components can sometimes be obtained by studying the shift of a probe molecule as a function of composition in a mixed solvent system. When an inert probe is used and the solvent molecules are magnetically isotropic, the solvent shift is dominated by the bulk susceptibility and van der Waals contributions,l and if the cosolvents mix ideally with one another the observed shift is a linear function of the volume fraction, i.e. where J1, ti2,and 6, are the shifts when neat 1, neat 2, or a mixture is used as the solvent. Since the probe concentration typically is kept very low, the volume fraction 420may be evaluated by using where n, is the number of moles of cosolvent i, having molar volume V,O. Behavior consistent with eq 1was first reported by Filipovich and Tiers2 and has been found for a number of other systems in this laboratory. Thus, though no rigorous (derivation of (1)exists, it now seems appropriate to look for special effects whenever observed 0022-365417912083-1393$01 .OO/O
values of 6, deviate appreciably from those given by (1). Several examples of such deviant behavior have recently been reported and explanations proposed. When one of the cosolvents is an electron donor and the probe is an acceptor, charge transfer interactions3 produce a strong curvature in the dependence of 6, on $2. When the probe is inert, but there is a substantial volume change on mixing the cosolvents, as for waterldioxane, a less pronounced but readily measurable curvature is found.' In the solvent systems water / tetrahydrofuran and water/ tert-butyl alcohol, 6, depends on the composition in a much more complex way, suggesting that water tends to enclathrate cosolvent molecules in the highly aqueous region while a t higher cosolvent concentrations the solutions become microheterogenous, with the trifluorohexanol probes located preferentially in cosolvent-rich d0mains.l Such microheterogeneity is also anticipated for strongly nonideal mixtures of nonaqueous solvents, especially in the neighborhood of the critical solution point. For example, it is well known that two-component systems having an upper consolute temperature show strong opalesence just above the consolute point, which is attributed to large concentration fluctuation^.^ This opalescence dies off rapidly when the temperature is raised, as is to be expected since increasing the molecular kinetic energies should tend to erase any residual order. Very few simple experimental approaches exist that permit one to determine to what (E 1979 American Chemical Society