J. Phys. Chem. 1995,99, 16602-16608
16602
Anisotropy in the 199Hg-31PIndirect Spin-Spin Coupling Tensor of a 1:2 Mercury -Phosphine Complex. A Phosphorus Single-Crystal NMR Study Michael D. LumsdenJ Roderick E. Wasylishen,*?+and James F. Britten' Department of Chemistry, Dalhousie University, HalifLur, Nova Scotia, Canada B3H 4J3, and the Department of Chemistry, McMaster University, Hamilton, Ontario, Canada L8S 4Ml Received: July 13, 1995@
A single crystal of the 1:2 mercury-phosphine complex, Hg(PPh&(N03)2, has been investigated by phosphorus-3 1 NMR spectroscopy. In the solid state, this species possesses two magnetically-distinct phosphorus nuclei which are spin-spin coupled to one another, resulting in 31PNMR spectra which exhibit A2, AB, or AX character depending upon the orientation of the single crystal in the external magnetic field. The three principal components of the phosphorus chemical shift (CS) tensors as well as their orientations in the molecular reference frame have been obtained from analyses of the single-crystal 3'P NMR spectra. The most shielded component of the CS tensor (833 = 15 ppm) lies approximately 14' off the Hg-P bond axis while the least shielded component (81I = 72 ppm) is oriented within the plane containing the smallest HgP-C bond angle. These findings represent the first characterization of a phosphorus CS tensor for a 1:2 mercury-phosphine complex. The absolute sign of the two-bond indirect spin-spin coupling, 2J(31P,31P)is0, has been determined to be positive. Analyses of the orientation dependence of the mercury-199 ( I = I/?, natural abundance = 16.84%), phosphorus-3 1 spin-spin interactions reveals that the 199Hg-3'P J tensor is anisotropic with Jll = 8.2 kHz and JI = 4.2 kHz. This result indicates that indirect spin-spin coupling mechanisms other than Fermi contact are operative. These findings complement an earlier characterization of the 199Hg-3'P J tensor in the 1:l mercury-phosphine complex, HgPCy3(N03)2 (Cy = cyclohexyl). For both the 1:l and 1:2 species, the anisotropy in J is appreciable relative to the isotropic coupling.
Introduction Previous phosphorus-3 1 solution and solid-state nuclear magnetic resonance (NMR) studies have demonstrated the isotropic value of the one-bond 199Hg-31Pindirect spin-spin coupling (J) tensor in Hg(I1) complexes to be an exkemely sensitive function of molecular structure.'%*'J(199Hg,3'P) values, which span a range on the order of 17 W Z , have ~ been correlated with numerous structural parameters such as Hg-P bond lengths, various bond angles, and the number of phosphorus ligands. Analogous empirical relationships between J(M,3'P)i,0 and molecular structure can be found for complexes containing other NMR-active metals such as ' l l " I3Cd and 195pt,1-4
It is well-known that the isotropic component of the J tensor can be mediated by three distinct mechanisms; namely, the Fermi-contact (FC), spin-orbital (SO), and spin-dipolar (SD) However, the possible importance of these latter two interactions has traditionally been neglected when interpreting metal-phosphorus isotropic J couplings. The justification for such an approach appears to be predominantly based upon the results of early semiempirical calculations of J couplings, which utilized simple molecular-orbital (MO) adaptations of Ramsey's theory.* Unfortunately, these methods generally involve numerous gross approximations which make it difficult to assess their reliability. For example, the early calculations of Pople and Santry predict that the SO and SD mechanisms make negligible contributions to J in the absence of multiple bonding. However, later semiempirical calculations provide several examples involving J(I3C,l3C) as well as couplings involving fluorine which contradict this concl~sion.~
* Address correspondence to this author.
' Dalhousie University. @
McMaster University. Abstract published in Advance ACS Abstracts, October 15, 1995.
0022-365419512099- 16602$09.00/0
Although considerable progress has been made more recently in the application of ab initio Hartree-Fock wave functions and beyond to the calculation of J , I o these modern approaches have been predominantly confined to couplings involving firstrow nuclei. Furthermore, neglect of the SO and SD terms has been propagated by the apparent general success of rationalizing isotropic J coupling data in terms of rudimentary concepts, such as the %s character of the intervening bonds between J-coupled nuclei, which stem from the assumption of FC dominance. While these ideas may have some validity in interpreting J(13C,13C),so or J('3C,'H),,, values in simple organic molecules, there is little justification for their application to couplings involving metal nuclei. Finally, the common practice of correlating indirect spin-spin coupling constants measured in solution with structural features measured from X-ray crystallography must be questioned. For Hg(II)-phosphine complexes such as the one considered here, solution dynamics such as phosphine exchange as well as dimerization equilibria can strongly influence 1J(199Hg,31P),ro, making the results of such correlations difficult to evaluate. It is noteworthy that Allman and Lenkinski have measured 1J('99Hg,3'P),,,in a series of 1:2 Hg(I1)-phosphine complexes in the solid state and were unable to find a correlation with predictions based upon a FC-dominated J coupling." It is our basic premise that a more thorough insight into the relationship between molecular structure and any of the fundamental NMR interactions, such as chemical shielding or J coupling, will ultimately be realized by probing the threedimensional or tensorial naturel2.l3of these interactions rather than the orientational average measured in solution. Of the three mechanisms that give rise to J coupling, only the FC mechanism is a scalar and therefore independent of molecular orientation with respect to the external magnetic field.5 Thus, experimental evidence that J is orientation dependent would clearly prove
0 1995 American Chemical Society
Mercury-Phosphine Complex
J. Phys. Chem., Vol. 99, No. 45, I995 16603
the existence of non-Fermi contact mechanisms. Such a result would demonstrate the need to exercise extreme caution when interpreting isotropic J exclusively in terms of the FC mechanism. Given the importance of characterizing the J tensorI4-l6 rather than simply its trace as well as the paucity of reliable experimental data in this field, we have investigated a variety of metal-phosphorus J tensors using solid-state NMR studies of powder samples"-*' and, where possible, single crystals.22 Recently, we have used the 31Psingle-crystal NMR technique to characterize the one-bond 1wHg-31P J tensor in the 1:l mercury-phosphine complex, HgPCy3(N03)2 (Cy = cycloComplexes of the general type Hg(X)2(PR3), have been extensively investigated using 31Psolution NMR and a large amount of data has been compiled for the isotropic value of the 199Hg-31PJ tensor.23 Of particular relevance to the present study is the fact that for a given phosphine and anion, the magnitude of I J(199Hg,31P)iso decreases by 30-50% with increasing n, the number of coordinated phosphine ligand^.^^"-^ Interpretation of this observation generally involves arguments which only consider the FC mechanism. For example, when n = 1, the majority of these complexes exist as dimers in solution and the larger values of 1J(199Hg,31P),so are thought to be the result of a greater electron deficiency at the Hg atom which results in stronger Hg-P o-bonding interactions.23a Stronger o-bonding implies greater "s-character" and hence a FCdominated J coupling is expected to increase. However, given the fact that the 1:1 complex HgPCy3(N03)2 has been found to possess a large orientation-dependent 199Hg-31PJ tensor, with AI = J,,- JI.on the order of 5.4 ~ H zand, with ~ ~similar values having been proposed in several other 1:l complexes,'* such interpretations of these coupling trends must be considered too simplistic. In an effort to complement our earlier characterization of a 199Hg-31PJ tensor in a 1:l complex and to determine how the J tensor changes in the 1:2 complexes, we have used 31Psingle-crystal NMR to characterize the first 199Hg-31PJ tensor in a 1:2 mercury-phosphine complex, Hg(PPh3)2(N03)2.
Scientific. Rotation patterns were obtained by rotating the crystal about each of the three orthogonal cube axes within a hollow, cubic receptacle in the probe goniometer, with the rotation axis oriented perpendicular to the external field direction. Spectra were obtained in 9' increments from 0" to 180°, resulting in the acquisition of 21 spectra for each of the three rotation patterns. For each orientation, typically 80 freeinduction decays were signal averaged using a 5 min recycle delay. In order to obtain accurate peak positions, the 31Psinglecrystal NMR spectra were deconvoluted with Gaussian peaks using the Bruker software package WINFIT. 31PNMR spectra were also obtained at 4.7 T for powdered samples of Hg(PPh3)z(NO& as well as with a Bruker AMX-400 NMR spectrometer (9.4 T) operating at a 31PLarmor frequency of 161.98 MHz. Bruker double-air-bearing MAS probes were employed at both fields with the powdered samples packed into zirconium oxide rotors of outer diameter 7 mm (4.7 T) and 4 mm (9.4 T). All 31PNMR spectra were acquired using the FLIPBACK pulse sequence24under conditions of cross-polarization and highpower proton decoupling. A contact time of 5 ms was used in all cases. Spectra have been referenced with respect to 85% H3P04(aq) which, for the acquisition of spectra for powdered samples, involved setting the isotropic peak observed in the MAS spectrum of NbH2P04 to 0.8 ppm. The following convention is maintained for designation of the three principal components of the 31Pchemical shift (CS) tensors: 811L 8 2 2 2 633, where 611 and 633 are the components of least and greatest shielding, respectively. The span of the CS tensor, Q, is defined as 611 - 833.25 NMR Line Shape Calculations. Calculation of the 31PNMR line shapes obtained for static powder samples of Hg(PPh3)z(NO3)2 at 4.7 and 9.4 T were performed on a 486 personal computer using simulation software developed in this laboratory written in C. Powder averaging was accomplished via the POWDER routine of Alderman et ~ 1 . ~ ~
Experimental Section
31PSingle-Crystal NMR Spectra. The monoclinic unit cell of Hg(PPh3)2(N03)2, space group C2/c (No. 15), contains four crystallographically-equivalentmolecules.23fThe mercury atom of each molecule is positioned on a crystallographic twofold rotation axis relating the two phosphine ligands and the two nitrate groups within each molecule. Although the two phosphorus nuclei within a molecule are crystallographicallyequivalent and therefore possess CS tensors with identical principal components and isotropic chemical shifts, these nuclei are magnetically distinct as the twofold rotation symmetry is not sufficient to constrain the orientations of the phosphorus CS tensors to be ~oincident.~'Consequently, the two phosphorus nuclei constitute a general dipolar-coupled homonuclear AB spin system. Furthermore, consideration of the crystallographic symmetry reveals that these phosphorus homonuclear spin pairs within different molecules are indistinguishable and consequently, for a general orientation of the single crystal of Hg(PPh3)2(N03)2 in an extemal magnetic field, NMR peaks are observed for two distinct phosphorus sites. Based on the X-ray crystallographic study of Buergi et uZ.,*~' the P-P intramolecular separation is 4.474 A in Hg(PPh&(NO3)2 which gives rise to a homonuclear direct dipolar coupling constant, R(31P,31P), of 220 Hz. For this same complex, Wu and Wasylishen have found 2J(31P,31P)iso, the indirect spinspin coupling constant, to be 250 Hz from the analysis of variable-angle-spinning(VAS)28and slow magic-angle-spinning (MAS)29"MR spectra. The fact that the two magneticallydistinct phosphorus nuclei are both directly and indirectly spin-
Compound Preparation. Bis(nitrato)bis(triphenylphosphine)mercury(I1) was prepared according to a previously described synthetic procedure." The large single crystal used in this study was obtained from a dichloromethane solution by slowly evaporating the solvent over the course of a month. X-ray Crystallography. A single crystal of approximate dimensions 4 x 3 x 2 mm was mounted on a hollow, threesided alumina cube with axes 4 mm in length. An NMR cube reference frame was defined by arbitrarily labeling the cube axes X , Y, Z in a right-handed fashion. The orientation matrix describing the location of the monoclinic unit cell axes (space group C2/c) with respect to this reference frame was obtained by placing the NMR sample in its alumina cube holder on an X-ray diffractometer and indexing 29 well-centered reflections. The Euler angles describing the orientation of the orthogonalized monoclinic crystallographic axis system (a, b, c*) with respect to the NMR cube frame were determined to be a = 310.6", ,t3 = 93.8", and y = 5.9". Errors in the position of each axis were estimated to be less than 0.5". The monoclinic unit cell parameters were determined as follows: a = 13.384(8) A, b = 13.994(7) A, c = 17.864(8) A, and /3 = 91.54(4)", which are in good agreement with the unit cell reported in the X-ray crystallographic study of Buergi et aZ.23f Solid-state NMR. Single-crystal 31PNMR spectra were acquired at 81.03 MHz (4.7 T) on a Bruker MSL-200 NMR spectrometer using a single-crystal goniometer probe from Doty
Results and Discussion
Lumsden et al.
16604 J. Phys. Chem., Vol. 99,No. 45, 1995
A
~
"
~
~
l
100
150
'
"
50
~
l
"
'
~
0
l
'
'
~
~
-50
[PPml
Figure 1. Typical 31Psingle-crystal NMR spectra obtained for Hg(PPh3)2(N03)2 at an external magnetic field strength of 4.7 T. The variation in line shape of the central intense peaks is due to changes in the relative sizes of the chemical shift difference between the two phosphorus nuclei and the homonuclear 3'P-3'P spin-spin coupling as a function of the orientation of the single crystal. Thus, these spectra can exhibit predominantly A2 (upper), AB (middle), or AX (bottom) character as the single crystal is rotated. As well, the weaker '99Hg satellite peaks are additionally influenced by the two 199Hg-31Pspinspin coupling tensors.
spin coupled to one another results in some interesting features in the 31Psingle-crystal NMR spectra. Both the CS difference between the two 3'P nuclei and the homonuclear spin-spin coupling interaction change as the single crystal is rotated in the external magnetic field as the chemical shielding and spinspin interactions are anisotropic in nature. Consequently, for a general orientation of the single crystal of Hg(PPh3)2(N03)2, one can, in principle, observe an A2, AB, or Ax pattern in the 3'P NMR spectrum arising from this homonuclear two-spin system (Figure 1). Griffin and co-workers, who examined carbon-13 NMR spectra of a single crystal of diammonium oxalate-I3C2monohydrate, have reported the only other example of this phenomenon in single-crystal NMR spectra.30 For homonuclear spin pairs in static powder samples, where a statistical distribution of molecular orientations relative to the external magnetic field is present, the presence of individual Az, AB, or AX spectra3' has been found to produce complex powder pattern line shapes.32 The expressions governing the peak positions, vi, and intensities, Zi, of the four transitions arising from a homonuclear spin pair are30.31
D A v,=vo+-+2 2
Z,=l--
B D
D v2=vo+--2
A 2
z2=1+-
B D
D A --+2 2
z3=1+-
v =v
O
B D
(3)
-D -A I =I-- B (4) 0 2 2 D where vo = (YA Y B ) / ~YA, and YB are the chemical shifts of the two nuclei (frequency units), and Z V
4
+
D =.JAV~
+ B~
(5)
l
+ R-(3 2
cos2 8 - 1)
(6)
A = J - R(3 cos2 8 - 1)
(7 1
B =J
In the above equations, Av = V A - VB, R is the dipolar coupling constant, and 8 is the angle between the internuclear vector and the external magnetic field. Our single-crystal NMR data provide no experimental evidence for anisotropy in the twobond 31P,31P indirect spin-spin coupling tensor; hence this has not been included in the above equations and will not be considered further. The above equations are similar to those describing an AB spin system in solution NMR, the difference being the incorporation of the homonuclear dipolar interaction into the "mixing term", D, in eq 5. Another interesting feature apparent in the 3'P single-crystal NMR spectra of Hg(PPh3)2(N03)2 is the presence of weak "satellite" peaks flanking the central peaks, which arise from both direct and indirect (J) spin-spin coupling with mercury199 ( I = '/2, natural abundance = 16.84%). These satellite peaks constitute the AB part of an ABX spin system. While the positions of the central peaks due to 31Pnuclei not adjacent to 199Hgnuclei are a function of the two 31PCS tensors and the 31P,31P spin-spin coupling interaction only, the peak positions and hence the splittings involving the '99Hg satellites are additionally influenced by the 199Hg-31Pspin-spin interactions. The fact that the 31PCS tensor information as well as the 31P,31P spin-spin coupling information can be analyzed completely independent of the 199Hg-31P spin-spin coupling tensors simplifies the analysis of the single-crystal NMR spectra. The spectra are also simplified because of the large magnitude of 1J(199Hg,31P),,,,5.5 kHz, which ensures that the two 199Hg subspectra are well separated from the central, uncoupled peaks. Although eqs 1-8 describe the peak positions and intensities arising from 31Pnuclei not coupled to 199Hg,the analogous expressions for the coupled peaks are more complex. From Figure 1, it is clear that the appearance of the 199Hgsubspectra can exhibit a strong dependence on the spin state of the 199Hg nucleus, m = &'/z. The asymmetric behavior is most evident at orientations where the 3'P,3'Phomonuclear spin pair is tightly coupled. The origin of this asymmetry lies in the fact that for a general orientation of the single crystal relative to the external magnetic field, the two 31Pnuclei of the ABX spin system experience a modified CS difference due to spin-spin coupling with 199Hg. This modified CS difference is a function of the spin state of the 199Hgnucleus and, qualitatively, arises due to the fact that the two 199Hg-31Pspin-spin coupling tensors are not coincident. Consequently, for the purposes of calculating the 199Hgsatellite peaks observed in 31PNMR spectra obtained for a single-crystal sample as well as for a static powder sample (vide infra), eqs 1-8 can be employed for each subspectrum provided that the CS difference term, Av, in eq 5 is replaced by a modified expression: (9) Note that this asymmetry in the satellite peaks is a function of the 199Hg-31Pdirect dipolar and anisotropic J coupling only. If the 199Hg-3'P spin-spin coupling interaction consisted of purely isotropic J coupling, identical 199Hgsubspectra would be anticipated for any orientation of the single crystal in the external magnetic field. Finally, asymmetric '99Hg satellite peaks have also been observed in 31PMAS spectra of Hg(PPh3)z-
J. Phys. Chem., Vol. 99, No. 45, 1995 16605
Mercury-Phosphine Complex
x Rotation
."
Y Rotation
TABLE 1: Chemical Shift Tensor Data for the Two Magnetically-Distinct Phosphorus Sites in Hg( PPhJ)Z(NO& Obtained from Analysis of the Single-Crystal NMR Datau
Z Rotation
chemical shift (ppm),,c
71 33 15 40 73 32 14 40
0
30
60
90
30
60
Rotation Angle ['I
90
30
60
90
direction cosines"
-0.645 -0.195 -0.738
-0.721 0.241 0.463 0.865 0.516 -0.440
0.676 -0.722 -0.163 0.140 0.343 -0.928 0.722 0.601 0.333
The convention used for designating the three principal components of the CS tensor is ( 3 1 1 I8 2 2 I633, where 611and 633 are the least shielded and most shielded principal components, respectively. All chemical shifts are referenced with respect to 85% HjPOd(aq) at 0 ppm. Errors in the principal components of the 31PCS tensors are estimated to be 2 ppm. The direction cosines are with respect to the orthogonalized monoclinic crystallographic frame (a, h, c*).
Figure 2. Rotation patterns displaying the variation in 31Pchemical shifts for the two magnetically non-equivalent phosphorus sites (site I , 0, site 2, m) upon rotation about each of the three orthogonal NMR cube axes, X, Y, 2. The results of the three-parameter least-squares fits of the data are indicated by the solid lines.
(NO3)2 at slow spinning speeds, the origin of which is qualitatively analogous to that observed in the present study.33 Phosphorus Chemical Shift Tensors. Rotation patterns displaying the variation in 31Pchemical shifts arising from the two magnetically-distinct phosphorus sites for rotation about each of the three orthogonal NMR cube axes, X,Y,2, are shown in Figure 2 along with the results of the three-parameter leastsquares fits of the data. In extracting the two phosphorus chemical shifts from the single-crystal NMR spectra, an exact analysis based upon eqs 1-7 was employed in order to account for second-order character observed in the single-crystal 31P NMR spectra. The CS tensor single-crystal data has been analyzed in the conventional manner.34 Small phase angles, of magnitude -5" and +lo have been applied to the X and 2 rotation data, respectively. The two 31PCS tensors obtained from this analysis are summarized in Table 1. The errors in the three principal components of the CS tensors are estimated to be 2 ppm. Note that the calculated isotropic chemical shifts are in very good agreement with that obtained from a 31PCP/ MAS spectrum of a powder sample of this complex, 6iso = 40.0 ppm. As well, the magnitudes of the principal components are the same, within experimental error, as those obtained from an analysis of VAS spectra of this complex.28 The direction cosines in Table 1, together with the known crystal structure of Hg(PPh&(N03)2, allow one to construct the orientation of the 31PCS tensors relative to the molecular reference frame. Although there exists, in principle, a twofold ambiguity in assigning the CS tensors to the two magneticallydistinct phosphorus sites, this can be overcome by examining the corresponding orientations of the 1wHg-31P spin-spin coupling tensors.34b Choosing the solution which yields the largest component of the two 1wHg-31Pspin-spin coupling tensors directed approximately along the Hg-P bond results in orientations of the 31PCS tensors as depicted in Figure 3. Examination of the local site symmetry about the two phosphorus nuclei provides further support for this assignment. More specifically, the two Hg-P-C( 1) planes constitute approximate local mirror planes at the phosphorus nuclei. Only for the assignment depicted in Figure 3 is this local symmetry reflected in the orientations of the 31Pchemical shift tensors; both 611 and 6 3 3 lie within this mirror plane while 6 2 2 is directed normal to this plane within experimental error. The angle between the
V
c,
b l l
\
b) Figure 3. Symmetrized orientations of the two 31P CS tensors in Hg(PPh3)2(NO3)2 obtained from the 31P single-crystal NMR study. Hydrogen atoms have been omitted for clarity. (a) View such that the P-Hg-P plane is coincident with the plane of the paper. The two magnetically-distinct phosphorus sites are indicated by closed circles while the ipso carbon atoms are displayed as crossed circles. Labeling of the CS tensor components and ipso carbon atoms has been performed at site 1 only. (b) Projection down the P (site 1) - P (site 2) internuclear vector. Note that for this projection, only the ipso carbon atoms of the phenyl rings are shown.
most shielded direction, 633, and the Hg-P bond axis is 14". Before producing the CS tensor orientations illustrated in Figure 3, the direction cosines in Table 1 were symmetrized such that the two 31PCS tensors were constrained to be related by the crystallographic twofold rotation axis. This exercise enables an estimate of the accuracy of the CS tensor orientations. It was determined that the deviation between eigenvectors in terms of the crystallographic C2 relationship was 8" for both the intermediately and most shielded components while a deviation of 5" was found for the least shielded component. Considering the uncertainty in the magnitudes of the principal components of the 31PCS tensors as well as the differences between these components, errors of this magnitude are not unexpected.3s The present study represents only the second single-crystal NMR investigation of the 31PCS tensor in a mercury-phosphine complex. The orientations of the 31PCS tensors in Hg(PPh&-
Lumsden et al.
16606 J. Phys. Chem., Vol. 99, No. 45, 1995 (NO3)2 are qualitatively analogous to that in the previously studied 1:l complex, HgPC~3(N03)2.~~ In both species, the most shielded component lies closest to the Hg-P bond while the least shielded component lies within the plane containing the smallest Hg-P-C bond angle. However, despite the similarities in Orientation, the magnitudes of the three principal components show marked differences. Each component is more shielded in the 1:2 complex. Although similar trends in the individual components of the 31PCS tensors have been observed for the free phosphine ligands, PPh3 and P C Y ~the , ~observed ~ shielding changes in the mercury complexes, relative to the free ligands, are amplified for both the least and intermediately shielded components. The tensor component least affected by the structural changes present within these two complexes is the most shielded component, 633, oriented closest to the Hg-P bond direction. The magnitude of this component in the 1:2 complex, 15 ppm, falls into the range typical of 31Pchemical shifts along the Hg-P bond, 10-35 ppm.'8.19,22Although the three Hg-P-C bond angles in HgPCy3(N03)2 are similar, with the largest difference about 2", for Hg(PPh3)2(N03)2this is not the case. Specifically, the two Hg-P-C(l) bond angles, 106.2(l)", are much smaller than the remaining Hg-P-C angles in this species, with the greatest difference approximately The larger departure from threefold rotation symmetry about the Hg-P bond in the 1:2 complex may partially justify the apparently larger deviation of the most shielded component from the Hg-P bond axis (14") compared with the 1:l complex (4") as well as the enhanced shielding asymmetry within the plane perpendicular to this component. Finally, the principal components of the 31PCS tensor for the 1:l species, [HgPPh3(N03)2]2, have been previously determined from NMR studies of a powder sample of this complex.I8 Compared with the analogous 1 :2 complex investigated here, both the intermediately and most shielded components have the same magnitudes within experimental error. However, the least shielded component is substantially more shielded in the 1:l complex (611 = 44 ppm). 31P,31PSpin-Spin Coupling. Since metal-phosphine complexes commonly contain more than one phosphine ligand per molecule, there has been considerable interest in both the magnitude and sign of 2J(31P,31P)is0 in these species.'$37This coupling is known to be sensitive to a variety of structural factors such as the metal involved and its coordination number, the P-M-P bond angle, the electronegativity of the substituents, as well as the stereochemistry of the complex. For example, it is a general observation that 2J(31P,31P)trans > 2J(31P,31P)cis. Furthermore, the absolute signs of these couplings often differ with the sign for the trans geometry positive and the cis geometry negative.' Specifically for the mercury-phosphine complexes, 2J(31P,31P)is0 generally exhibits a similar dependence on these structural parameters to that of 1J(199Hg,31P)iso. For example, Allman and Lenkinski have measured 2J(31P,3'P)iso within a series of 1:2 complexes of the type, Hg(PPh3)2(X)2, and have found this coupling constant, which ranged from 110 to 240 Hz, to generally increase with the electron-withdrawing capability of the anion, X." Although 2J(31P,31P)is0 within the nitrate complex investigated here was not reported in their study due to the crystallographical equivalence of the phosphine ligands, this coupling constant has since been measured to be even larger, 250 Hz.2s,29Similar trends have been reported for 1J('99Hg,31P)iso in these same 1:2 complexes.23 We have carefully examined the orientation dependence of the splittings, A, observed in the 31Psingle-crystal NMR spectra due to homonuclear spin-spin coupling between the two phosphorus nuclei (cf. eq 7). Given that typical line widths in the single-crystal NMR spectra were on the order of 250 Hz,
the relatively small magnitude of this coupling precluded its observation at many orientations of the single crystal, making a reliable least-squares analysis of the coupling data impossible. Nevertheless, the sign of 2J(31P,31P)ls0 is available from these data. We observed that the magnitude of the splitting, A, reached a maximum when the P-P internuclear vector was oriented perpendicular to the external magnetic field direction. The largest splitting was observed at the Y( 108") orientation, A = 475 Hz. These results indicate that the relative sign of the dipolar and J coupling between the two 3'P nuclei is the same; thus the absolute sign of 2J(31P,31P)ls0 is positive. A positive sign has also been reported for 2J(31P,31P)lso in [Hg(PMe3)21(NO3)2 (+250 H z ) , ~the~ only other determination of both the sign and magnitude of 2J(31P,31P),s0 in a mercury-phosphine complex of which we are aware. 1wHg-31PSpin-Spin Coupling Tensors. The peak positions of the 199Hgsatellites and hence the separation between these subspectra can be analyzed to obtain information on the sum of the 199Hg-31Pdipolar and J tensors. We have not performed a full three-dimensional analysis of the 199Hg-31P spin-spin coupling tensors due to difficulties in reliably analyzing the satellite peaks obtained for rotation about the cube Z axis. Within the first 60" of this rotation pattern, the maximum chemical shift difference between the two 31Pnuclei is only on the order of 4 ppm. Furthermore, the 31P-31Pdipolar vector is oriented just 14" off the rotation axis, resulting in relatively large 31P-31P spin-spin interactions which varied little with orientation throughout this rotation pattern. Consequently, some of the most tightly-coupled second-order spectra observed in this study were obtained within the first 60" of the Z rotation pattern. The IBHg satellite peaks exhibited relatively little orientation dependence and appeared as featureless, broad humps. With similar spectral features evident near the end of this rotation pattern, a conventional least-squares analysis of the 199Hg-31Pspin-spin coupling data was not possible. Despite these difficulties, reliable information concerning the anisotropy in the 199Hg-31PJ tensors can be deduced from the X and Y rotation patterns. It was observed that when the 199Hg31Pinternuclear vectors were oriented approximately along the external magnetic field direction, maxima were observed in the corresponding 199Hgsatellite splittings. Given the well-known orientation dependence of the dipolar interaction relative to the molecular reference frame, this finding indicates that the largest component of the 199Hg-31PJ tensor is oriented along the Hg-P bond direction within experimental error. Furthermore, combined with the findings of our earlier investigation of H ~ P C Y ~ ( N Oit~is) reasonable ~?~ to assume that the J tensor is axially symmetric. Hence, the splittings between the 199Hg subspectra, h v , can be expressed as
AV = J,,, - ~ , , ( 3 COS* e - 1 )
(10) where R,ff is the effective dipolar coupling constant, R,ff = R - AJl3.
Displayed in Figure 4 is the observed orientation dependence of the 199Hgsatellite splittings obtained for rotation about the cube X axis. Note that the isotropic J coupling constant, 1J(199Hg,3'P)lso = 5550 Hz, obtained from an independent CP/ MAS experiment, has been subtracted from the measured splittings, thereby isolating coupling contributions due to dipolar and anisotropic J coupling only (cf eq 10). As well, the horizontal dashed line in this figure indicates the theoretical maximum value of the effective 199Hg-31Pdipolar coupling due to direct dipolar coupling alone ( i x . , in the absence of an anisotropic J tensor), 2 R ~ ~ -=p 1200 Hz. This result has been derived using a Hg-P bond length of 2.451 A obtained
J. Phys. Chem., Vol. 99, No. 45, 1995 16607
Mercury-Phosphine Complex
X ROTATION
1000
140 120 100 00 60 40 20
0 -20 -40
[PPW i
1
1
1
1
1
0
20
40
60
EO
100
1
1
120 140
1
l
160
180
Rotation Angle ['I Figure 4. Variation in the 1wHg-31PefSective dipolar coupling obtained for rotation about the cube Xaxis (site 1, 0 , site 2, M). The horizontal dashed line indicates the maximum coupling in the absence of an anisotropic 199Hg-31PJ tensor.
in the X-ray diffraction study of Buergi et al.23f The fact that the observed splittings exceed this maximum for site 1 at many orientations immediately indicates that the 199Hg-31PJ tensor is anisotropic. Based upon the observed extrema in the X and Y rotation patterns, a value for Reff can be calculated from eq 10 provided that the orientation of the corresponding Hg-P internuclear vector is known relative to the cube plane within which the external magnetic field probes. For the rotation patterns displayed in Figure 4, the two Hg-P vectors are calculated to project 7" and 30" out of the cube YZ plane for sites 1 and 2, respectively. Examination of the X and Y coupling data in this fashion yields the result, Reff = 725 4~ 150 Hz. Furthermore, the fact that maximum satellite splittings for a particular site were observed when the external magnetic field was oriented approximately parallel with the corresponding Hg-P internuclear vector indicates that the absolute sign of R,ff is opposite to that of Jiso. Given that the sign of 'J(199Hg,31P)iso has previously always been found to be positive,2,6,39the absolute sign of R,ff must be negative. Consequently, this result enables calculation of the anisotropy in the 199Hg-3'PJ tensor, AJ = 4.0 f 0.5 kHz. In order to provide independent support of the single-crystal NMR data, we have obtained 31PNMR spectra for a static powder sample of Hg(PPh&(N03)2 at two different applied magnetic fields. A computer program based on eqs 1- 10 has been written in order to calculate these spectra. The results for the experimental and calculated 3'P powder spectra at both 4.7 and 9.4 T are shown in Figure 5. The agreement between the experimental and calculated spectra i s very good, in support of the single-crystal NMR data. The 31P single-crystal NMR results for both the 1:2 and previously studied 1:1 mercury-phosphine complexes reveal an anisotropy in the 199Hg-31PJ tensor which is comparable to the isotropic coupling constant, being on the order of 70% in both species (cf.Table 2). Therefore, one can conclude that mechanisms other than Fermi contact are operative and make substantial contributions to the transmission of nuclear spin
Figure 5. Experimental and calculated 31PNMR spectra obtained for a static powder sample of Hg(PPh&(N03)2 at two different applied magnetic fields. The calculated spectra have been generated using the data obtained from the single-crystal NMR study. Note that the calculated spectra have been convoluted with a Gaussian line broadening function of magnitude 500 Hz (9.4 T) and 350 Hz (4.7 T).
TABLE 2: Comparison of 1wHg-31PJ Tensors Obtained from 31PSingle-Crystal NMR Experiments for Both a l:la and 1:2bMercury-Phosphine Complex JtXJ' JII Ji AJ Hg(PPWz(NOd2 HgPCydN03)z
5.5 8.2
8.2 11.8
4.2 6.4
4.0 5.4
*
a Reference 22. Present investigation. Obtained from 31PC P M A S spectra of powdered samples.
information between 199Hgand 3'P nuclei in these systems. Such findings are a reminder of the fact that interpretation of 199Hg31Pisotropic coupling constants using concepts based on the Fermi-contact mechanism alone is an extremely oversimplified approach. Although trends in these data may apparently be rationalized using such simplistic ideas, extreme caution must be exercised without knowledge of the behavior of the noncontact mechanisms. The results presented here contribute to a growing body of evidence that metal-phosphorus indirect spin-spin couplings are a n i ~ o t r o p i c . ~ ~Further - * ~ , ~ ~investigations of metal-phosphorus J tensors are currently underway in our laboratory as part of a continuing effort to better understand the dependence of the full J tensor on molecular structure rather than just the trace. Acknowledgment. We thank the Natural Sciences and Engineering Research Council (NSERC) of Canada for financial assistance in the form of equipment and operating grants (REW) as well as a post-graduate scholarship (MDL). NSERC is also acknowledged for financial support of the Atlantic Region Magnetic Resonance Centre (ARMRC), where all spectra in this study were obtained. We are grateful to Drs. Klaus Eichele and Gang Wu for many insightful discussions. References and Notes (1) Pregosin, P. S . ; Kunz, R. W. jiP and "C NMR of Transition Metal Phosuhine Comulexes: NMR Basic Princiules and Prowess: Surincer" Verlag: Berlin, '1979;'Vol. 12. (2) Verkade, J. G.; Mosbo, J. A. In Phosphorus-31 NMR Spectroscopy in Stereochemical Analysis, Organic Compounds and Metal -Complexes;
16608 J. Phys. Chem., Vol. 99, No. 45, 1995 Verkade, J. G., Quin, L. D., Eds.; Methods in Stereochemical Analysis 8; VCH Publishers, Inc.: Deerfield Beach, FL, 1987; pp 425-463. (3) Nixon, J. F.; Pidcock, A. Annu. Rev. NMR Spectrosc. 1969,2, 345421. (4) Pidcock, A. In Catalytic Aspects of Metal Phosphine Complexes; Alyea, E. C., Meek, D. W., Eds.; Advances in Chemistry Series 196; American Chemical Society: Washington, DC, 1982; pp 1-22. (5) Ramsey, N. F. Phys. Rev. 1953, 91, 303-307. (6) Jameson, C. J. In Multinuclear NMR; Mason, J., Ed.; Plenum Press: New York, 1987; pp 89-131. (7) Jameson, C. J. In Phosphorus-31 NMR Spectroscopy in Stereochemical Analysis. Organic Compounds and Metal Complexes; Verkade, J. G., Quin, L. D., Eds.; Methods in Stereochemical Analysis 8; VCH Publishers, Inc.: Deerfield Beach, FL, 1987; pp 205-230. (8) Pople, J. A,; Santry, D. P. Mol. Phys. 1964, 8, 1-18. (9) See for example: (a) Blizzard, A. C.; Santry, D. P. J . Chem. Phys. 1971, 55, 950-963. (b) Schulman, J. M.; Newton, M. D. J . Am, Chem. SOC.1974, 96, 6295-6297. (10) (a) Kowalewski, J. f r o g . Nucl. Magn. Reson. Spectrosc. 1977, 11, 1-78; Annu. Rep. NMR Spectrosc. 1982, 12, 81-176. (b) Contreras, R. H.; Facelli, J. C. Annu. Rep. NMR Spectrosc. 1993,27, 255-356. (c) Fukui, H. In Specialist Periodical Reports on NMR; Webb, G. A,, Ed.; Royal Society of Chemistry: London, 1995; Vol. 24; pp 133-161. (11) Allman, T.; Lenkinski, R. E. Inorg. Chem. 1986, 25, 3202-3204. (12) Mehrinp. M. Hiah Resolution NMR in Solids, 2nd ed.; Springer. Verlag: Berlin,-1983. " (13) Schmidt-Rohr, K.; Spiess, H. W. Multidimensional Solid-State NMR and Polymers; Academic Press; San Diego, CA, 1994. (14) Buckingham, A. D.; Love, I. J . Magn. Reson. 1970, 2, 338-351. (15) Buckingham, A. D.; Pyykko, P.; Robert, J. B.; Wiesenfeld, L. Mol. Phys. 1982, 46, 177-182. (16) Robert, J. B.; Wiesenfeld, L. Phys. Rep. 1982, 86, 363-401 (17) Penner, G. H.; Power, W. P.; Wasylishen, R. E. Can. J . Chem. 1988, 66, 1821-1823. (18) Power, W. P.; Lumsden, M. D.; Wasylishen, R. E. J. Am. Chem. Soc. 1991, 113, 8257-8262. 119) Power. W. P.: Lumsden. M. D.: Wasvlishen. R. E. Inora. Chem. 1991, 30, 2997-3002. (20) Power, W. P.; Wasylishen, R. E. Inorg. Chem. 1992. 31. 217621 81
(21) Wasylishen, R. E.; Wright, K. C.; Eichele, K.; Cameron. T. S. Inorg. Chem. 1994, 33, 407-408. (22) Lumsden, M. D.; Eichele, K.; Wasylishen, R. E.; Cameron, T. S.; Britten, J. F. J . Am. Chem. Soc. 1994, 116, 11129- 11136.
Lumsden et al. (23) See for example: (a) Grim, S. 0.;Lui, P. J.; Keiter, R. L. Inorg. Chem. 1974, 13, 342-345. (b) Alyea, E. C.; Dias, S. A,; Goel, R. G.: Ogini, W. 0.;Pilon, P.; Meek, D. W. Inorg. Chem. 1978, 17, 1697-1700. ( c ) Alyea, E. C.; Dias, S. A. Can. J . Chem. 1979, 57, 83-90. (d) Allman, T.: Goel, R. G.; Pilon, P. Can. J . Chem. 1979, 57, 91-97. (e) Buergi. H. B.; Kunz, R. W.; Pregosin, P. S. Inorg. Chem. 1980, 19, 3707-3712. (0 Buergi, H. B.; Fischer, E.; Kunz. R. W.; Parvez, M.; Pregosin, P. S. Inorg. Chem. 1982, 21. 1246-1256. (24) Tegenfeldt, J.; Haeberlen, U. J . Magn. Reson. 1979,36,453-457. (25) Mason, J. Solid State Nucl. Magn. Reson. 1993, 2, 285-288. (26) Alderman. D. W.; Solum, M. S.; Grant, D. M. J . Chem. Phys. 1986, 84, 3717-3725. (27) Haeberlen, U. In Advances in Magnetic Resonance, Supplement I; Waugh. J. S., Ed.; Academic Press: New York, 1976. (28) Wu, G.; Wasylishen, R. E. J . Chem. Phys. 1994, 100,4828-4834. (29) Wu, G.; Wasylishen, R. E. J . Chem. Phys. 1993, 98,6138-6149. (30) van Willigen, H.; Griffin, R. G.; Haberkorn. R. A. J . Chem. Phys. 1977, 67, 5855-5860. (31) Zilm, K. W.; Grant, D. M. J . Am. Chem. Soc. 1981, 103, 29132922. (32) (a) Lumsden. M. D.; Wu, G.; Wasylishen, R. E.; Curtis, R. D. J . Am. Chem. Soc. 1993, 115, 2825-2832. (b) Curtis, R. D.: Hilbom, J. W.; Wu, G.; Lumsden, M. D.; Wasylishen, R. E.; Pincock, J. A. J . Phys. Chem. 1993, 97, 1856-1861. (33) Wu, G.; Wasylishen, R. E. Mol. Phys. 1994, 83, 539-549. (34) (a) Veeman. W. S. f r o g . Nucl. Magn. Reson. Spectrosc. 1984, 16, 193-235. (b) Kennedy, M. A,: Ellis. P. D: Concepts M a p . Reson. 1989, I , 35-47, 109-129. (35) (a) Eichele, K.; Wu, G.; Wasylishen, R. E.; Britten, J. F. J . Phys. Chem. 1995, 99, 1030- 1037. (b) Harbison, G. S.; Jelinski. L. W.; Stark, R. E.; Torchia. D. A,; Herzfeld, J.: Griffin, R. G. J . Magn. Reson. 1984. 60, 79-82. (36) Penner, G. H.: Wasylishen, R. E. Can. J . Chem. 1989, 67. 19091913. (37) Verkade, J. G. Coord. Chem. Rev. 1972, 9, 1-106. (38) Goodfellow, R. J.: Taylor. B. F. J . Chem. Soc., Dalton Trans. 1974, 1676-1684. (39) (a) McFarlane, W.; Rycroft, D. S. J. Chem. Soc., Faraday Trans. 2 1974, 70, 377-385. (b) Goggin, P. L.; Goodfellow, R. J.; McEwan, D. M.: Griffiths, A. J.; Kessler, K. J . Chem. Res., Miniprint 1979, 23152343. (40) Sears, R. E. J.; Ni, Q. W. J . Chem. Phys. 1994, 100, 4708-4709.
JP95 1971H
