MAGNETIC RESONANCE IN CHEMISTRY, VOL. 31, 537-539 (1993)

Observation of *OiHg NMR Spectra for Solid

Gang Wu and Roderick E. Wasylishen* Department of Chemistry, Dalhousie University, Halifax, Nova Scotia, B3H 453, Canada

The first observation of "'Hg NMR signals in the solid state is reported for potassium tetracyanomercurate(II), K,Hg(CN),. NMR measurements of 'OIHg and '99Hg were carried out at 4.70, 7.05 and 9.40 T. The ratio between the observed NMR frequency for *OIHg and '99Hg yields an accurate value for the ratio of their magneto- gyric ratios, 1y(20'Hg)/y('99Hg) I = 0.369 1389(9). This value agrets well with that obtained from optical pumping experiments on mercury vapor in the gas phase. The *"Hg NMR line shapes observed for solid K,Hg(CN), are Lorentzian owing to efficient relaxation via the quadrupolar mechanism. In contrast, the corre- sponding '99Hg NMR line shapes are Gaussian and much narrower.

KEY WOKDS Mercury-201 NMR Magnetogync ratios Solid-State NMR F'otassium tetracyanomercurate(I1)

INTRODUCTION

There are two stable, magnetically active isotopes of mercury: 19'Hg (Z = 112, natural abundance = 16.84%) and 201Hg (Z = 312, natural abundance = 13.22%). Although '99Hg NMR studies are c ~ m m o n , ' - ~ mercury-201 NMR studies have been scarce in the liter- ature.

Mercury-201 has a very large quadrupole moment, Q = 0.455(40) x m2.4 Nuclear quadrupole reson- ante (NQR), electron spin resonance (ESR) and micro- wave studies have revealed large 'O'Hg nuclear quadrupolar coupling constants, e.g. 720 MHz in HgC12,5 900 MHz in the CH2HgC1 radical6 and 1025 MHz in CH,HgCl.' Mercury-201 NMR signals were first recorded in the gas phase by atomic beam and optical pumping experiments.*-1° The Knight shift of the 'O'Hg NMR signal of liquid metallic mercury was studied by using a continuous-wave (CW) NMR tech- nique." More recently, 201Hg nuclear relaxation was investigated by optical experiments in the gas pha~e.". '~ The authors found that when the vapour- phase atoms are adsorbed on the walls of the container, where a non-zero eiectric field gradient (EFG) exists, the 201Hg nuclear relaxation is mainly due to the quadru- polar mechanism. However, to date there has been no report of 'O'Hg NMR measurements in mercury- containing compounds.

In general, the observation of NMR signals from a nucleus with a very large quadrupole moment is diffi- cult. In solution, efficient relaxation due to a large quadrupole interaction results in severely broadened lines which often make detection impossible. However, in some cases, where the quadrupole nucleus is at a site of high symnietry in the crystal, the EFG approaches zero and relatively sharp lines can be observed. For

* Author to whom correspondence should be addressed

example, most of the alkali metal halides have cubic structures and NMR peaks can be observed readily for al1 of the quadrupoiar alkali metal nuclei in the solid state.14-lu In solution, relatively long Ti values are observed for quadrupolar nuclei which occupy sites of tetrahedreil or octahedral symmetry in a molecule, e.g. 14NH4+ (Ref. 20) and 33SF, . 2 1

The cubic structure of solid K,Hg(CN), 2 2 , 2 3 pro- vides one with an opportunity to detect 201Hg NMR signals, since the EFG vanishes at the center of the Hg(CN), tetrahedron. Here, we report observation of 'O'Hg NMR signals for solid K,Hg(CN), at 4.70, 7.05 and 9.40 T. To the best of our knowledge, this is the first repcrt that 201Hg NMR signals have been observed in the solid state. In addition, '99Hg NMR spectra of the same compound are presented. Compari- son between observed "'Hg and '99Hg N M R fre- quencies yields an accurate value of the ratio of their magnetog yric ratios.

EXPERIMENTAL

Crystallinc samples of K,Hg(CN), were prepared by dissolving 0.020 mol of KCN and 0.010 mol of Hg(CN), in methítnol, followed by evaporation of the ~ o i v e n t . ~ ~ , ~ ~

Mercury NMR spectra of K,Hg(CN), were obtained on Bruker MSL-200, MSL-300 and AMX-400 NMR spectrometers with appiied magnetic fields of 4.70, 7.05 and 9.40 T, respectively, using a Bruker broad-band probe with a solenoid coil. Typically, a sample of 1.5 g was used. Mercury-201 NMR experiments were also performed under the magic angle spinning (MAS) con- dition on a Bruker AMX-400 NMR spectrometer. A \ampie of 0.3 g of K,Hg(CN), was packed into a zir- conia rotor (7 mm 0.d.). A typical spinning frequency was 4.0 kHz. Accurate values of the NMR line positions and Iinewidths were obtained using the GLINFIT program supplied by Bruker.

0749-1 58 1/93/060537-O3 $06.50 0 1993 by John Wiley & Sons, Ltd

Receivrd 14 September 1992 Accepted (revised) 1 F'ehruary 1993

538 G. WU AND R. E. WASYLl SHEN

RESULTS AND DISCUSSION

Mercury NMR spectra of solid K,Hg(CN), are shown in Fig. 1. The observed NMR frequencies and line widths for 19'Hg and 'OIHg at applied fields of 4.70, 7.05 and 9.40 T are summarized in Table 1. Since K,Hg(CN), exhibits a cubic structure at room tem- perature,', the local symmetry at the Hg(CN), tetra- hedron insures that the chemical shielding anisotropy at mercury is zero. The '99Hg NMR linewidth at half- weight (Avljz) was found to be 218, 244 and 280 Hz at 4.70, 7.05 and 9.40 T, respectively. The error in these line widths is estimated to be less than 20 Hz. The observation of essentially field-independent '99Hg NMR h e widths indicates that the '99Hg chemical shielding is isotropic. In fact, the '99Hg NMR line width in solid K,Hg(CN), is dominated by the direct dipolar and the indirect spin-spin ( J ) interactions

'''HcJ

a d b

Hg 201

JL ii,

- - 3.0 O -3.0 4.0 O - 4.0

kHz kHz

Figure 1. '99Hg and "'Hg NMR spectra of solid K,Hg(CN), obtained at (a) 4.70, (b) 7.05 and (c) 9.40 T. Note that the scales are slightly different for the '"Hg and 'O' Hg NMR spectra.

between the 199Hg nucleus and the four adjacent I4N nuclei [RDD('99Hg,'4N) = 43 Hz and zJ(199Hg,14N)iso = 20.6 Hz].,~ Small increases observed for the '99Hg NMR line width as a function of applied magnetic field could be due to the bulk magnetic susceptibility of the ~ a m p l e ~ ~ * ~ * or a distribution of chemical shifts due to imperfections in the crystal; nevertheless, this contribu- tion is very small.

The chemical shielding at a particular '"Hg nucleus is expected to be the same as that at a '99Hg nucleus in the same site in any given compound (the primary isotope effect on chemical shielding is negligible for heavy atoms such as merc~ry). ,~ The ratio of the ,OIHg NMR frequency over the 199Hg NMR frequency observed at the same applied magnetic field strength should thus equal the ratio between their magnetogyric ratios, y(201Hg)/y(199Hg). Since both the ,OIHg and '99Hg NMR line widths are essentially independent of applied magnetic field, less error is expected in the fre- quency ratios at high field strength. It is clear from Table 1 that the observed frequency ratios at three mag- netic fields agree with previous results obtained from optical experiments in the gas phase, 1 y(20'Hg)/

~ ( ' ~ ~ H g ) l =0.36913880(15)9 and Iy(201Hg)/y('99Hg)I = 0.369 138 7(7)." Usually, optical experiments exhibit a higher resolution than do the NMR experiments. In the present case, the accuracy of the ratio determination is mainly dependent on the accuracy of measurement of the ,OIHg NMR frequency, since 'O'Hg NMR signals have a lower frequency and are broader than the corre- sponding I9'Hg NMR signals. If the uncertainty in the observed '99Hg and 'O'Hg NMR frequency is assumed to be less than 10 and 100 Hz, respectively, the error in the 199Hg and "'Hg NMR frequencies at 9.40 T will be 0.14 and 3.8 ppm, respectively. At 14.10 T (600 MHz for 'H), a simple extrapolation yields uncertainties as small as 0.09 and 2.5 ppm for the NMR frequencies of '99Hg and 'O'Hg, respectively. In this sense, NMR experi- ments at high applied magnetic fields can produce data as accurate as those obtained in optical experiments, even for solid materials.

As can be seen in Table 1, the ,OIHg NMR line width is always found to be greater than that of the '99Hg NMR signals. This was also the case when '"Hg NMR signals were observed in metallic mercury." Since the chemical shielding anisotropy is found to be zero from

Y ( ' ~ ~ H ~ ) 1 = 0.369 138 80(15)9 and I Y(""'Hg)/

Table 1. Observed '99Hg and *O'Hg NMR frequency (in MHz) and line width (in Hz) for solid K,Hg(CN), at three different applied magnetic fields

Parameter 4 70 4, ( T ) 7.05 9.40

v(lg9Hg) (MHz) 35.832 61 7 53.731 888 71.634 533 v("'Hg) (MHz) 13.227 283 19.834 578 26.443 096 Y(''' H g ) j ~ ( ' " H g ) ~ 0.369 140 8 (1 7) 0.3691398 (12) 0.369 1 38 9 (9) A v , , ~ ( ' " H ~ ) ( H z ) ~ 218 244 280

'Errorc are estirnated assurning that the uncertainty in observed ls9Hg and "'Hg NMR fre- quency is less than 1 O and 1 O0 Hz, respectively.

Avl,2(20'Hg) (Hz)' 912 81 9 1200

Fitted with a Gaussian line shape. Fitted with a Lorentzian line shape.

'"Hg N M R SPECTRA 539

the '99Hg NMR linewidths, its contribution to the "'Hg NMR line width is also expected to vanish. The

g nucleus has a smaller magnetogyric ratio and therefore both dipolar and J interactions to four adjac- ent 14N nuclei in the molecule are expected to be reduced by a factor of 1y(199Hg)/y(201Hg)) x 2.7, resulting in a contribution of less than 100 Hz to the observed 'OiHg line width. It is therefore clear that the line-broadening mechanisms in the 201Hg NMR spectra are different from those in the '99Hg NMR spectra. The other possible he-broadening mechanism for the 201Hg NMR signal is the quadrupolar interaction. Since K,Hg(CN), has a cubic structure in the solid state, the static EFG is expected to be zero. However, it is also a general observation that the quadrupolar inter- action is operative even in cubic crystals because of the presence of crystal defects and lattice vibration~.~' A close inspection of the line shapes of the 199Hg and 201Hg NMR signals in solid K,Hg(CN), reveals the interesting fact that the 199Hg NMR signals can be characterized as Gaussian line shapes and "'Hg NMR signals as Lorentzians. This was confirmed by the line- shape fitting performed using the GLINFIT program. Inversion-recovery ,OIHg spin-Iattice relaxation mea- surements indicate Ti < 1 ms. Since T2 < TI, the appar- ent Lorentzian line shape and substantial line width must be a result of rapid transverse relaxation (T,) due to the quadrupolar mechanism. This is not surprising considering the fact that extremely short Ti values were also observed for quadrupolar nuclei in many other cubic crystals such as KBr, KI and NH,Br.3'932 Mercury-201 MAS experiments were also performed at

201H

9.40 T and no apparent reduction in line width was observed The line width of the Gaussian '99Hg NMR line shapes observed for K,Hg(CN), results from direct dipolar ,md indirect spin-spin interactions between '99Hg and 14N nuclei.26

Although the tetracyanomercurate(I1) ion, Hg(CN):-, is the predominant species present in aqueous solutions of K,Hg(CN), ,33*34 attempts to observe a 201Hg NMR signal for solution samples have been unsuccessful. Two sohtions were examined; one was a 1 M aqueous solu- tion of K,Hg(CN), and the other was also 1 M but con- tained excess KCN. Presumably, the 201Hg relaxation is so efikieiit that the NMR peaks are too broad to be detected, i.e. in solution the 201Hg Tl is probably less than the probe dead time. Any small dynamic changes around the "'Hg nuclei that arise from rapid ligand exchange or molecular collisions cause fluctuating elec- tric field gradients which result in rapid quadrupole relaxation due to the large quadrupole moment of the 201 Hg niiclei. This argument is certainly reasonable considering the fact that "'Hg T, relaxation due to the quadrupolar mechanism in solid K,Hg(CN), is much less than 1 ms at room temperature.

Acknowledgements

We are griteful to M r Ken Wright for preparing a sampie of K,Hg(CN), and to Dr John Walter of the Atlantic Recearch Labor- atory (ARLi for permitting us to obtain NMR spectra at 7.05 T. This work was financiaiiy supported by the NSERC of Canada. We also thank Profcssor T. S. Cameron and Dr B. M. Borecka for helpfui discussions regarding the crystal structure of K,Hg(CN), .

REFERENCES

1 .

2.

3.

4.

5.

6.

7. 8.

9.

1 o.

11.

12. 13.

14. 15. 16.

17. 18.

R. G. Kidd and R. J. Goodfellow, in NMR and the Perjodic Table, edited by R. K. Harris and B. E. Mann, Chapt. 8, p. 195. Academic Press, New York (1 978). R. J. Goodfellow, in Muttinuclear NMR, edited by J. Mason, Chapt. 21, p. 563. Plenum Press, New York (1 987). B. Wrackmeyer and R. Contreras, Annu. Rep. NMR Spectrosc. 24, 267 (1 992). M. N. McDermott and W. L. Lichten, fhys. Rev. 119, 134 (1960). H. G. Dehmelt, H. G. Robinson and W. Gordy, Phys. Rev. 93, 480 (1 954). C. M. Kerr, J. A. Wargon and F. Williams, J. Chem. Soc., Faraáay Trans. 2 72, 552 (1 976). C. A. Rego and A. P. Cox, Chem. Phys. Letr. 139,595 (1987). W. Faust, M. McDerrnott and W. Lichten, Bull. Am. Phys. Soc. i l 3, 371 (1 958). J-C. Lehmann and R. Barbé, C. R. Acad. Sci. 257, 3152 (1963). R. J. Reirnann and M. N. McDermott, Phys. Rev. C 7, 2065 (1973). W. E. Blurnberg, J. Eisinger and R. G. Shulman. J. f h y s . Chem. Sotids 26, 1 187 (1 965). P. A. Heirnann, Phys. Reu. A 23, 1204 (1 981 ). P. A. Heimann, l . A. Greenwood and J. H. Simpson, Phys. Rev.A 23,1209 (1981). T. Kanda, J . Phys. Soc. Jpn. 10, 85 (1955). K. Yosida and T. Moriya, J. Phys. Soc. Jpn. 11, 33 (1956). J. Kondo and J. Yamashita, J . Phys. Chem. Sotids 10, 245 (1 959). R. Baron, J. Chem. Phys. 38, 173 (1963). D. lkenberry and T. P. Das, Phys. Rev. A 138, 822 (1 965) ; J. Chem. fhys. 43, 2199 (1 965).

19.

20.

21.

22.

23.

24.

25.

26. 27.

28. 29.

30.

31.

32. 33.

34.

A. R. iaase, M. A. Kerber, D. Kessier, J. Kronenbitter, H. Krüger, O. Lutz, M. Müller and A. Noile, 2. Naturforsch., Teil A, 32, !352 (1 977). R. E. Mlasylishen and J. O. Friedrich, J . Chem. fhys. 80, 585 (1 984). R. E. Wíasylishen, C. Connor and J. Friedrich, Can. J . Chem. 62, 981 (1 984). P. N. Gerlach and B. M. Powell. J . Chem. fhys. 85, 6004 (1986). B. M. Powell and P. N. Gerlach, fhys. Rev. 8 40, 2426 (1989). F. Wagenknecht and R. Juza, in Handbook of ffeparative lnorgariic Chemistry, edited by G. Brauer, 2nd ed., Vol. 2, p. 11 22. Academic Press, New York (1 965). A. G. Sharpe, The Chemistry of Cyano Compiexes of the Tran- sition Metals. Chapt. 1 1. Academic Press, New York (1 976). G. Wu iind R. E. Wasylishen, submitted for publication. D. L. LanderHart, W. L. Earl and A. N. Garroway, J . Magn. Reson. 44,361 (1981). M. Alla and E. Lippmaa, Chem. Phys. Lett. 87, 30 (1982). C. J. Jameson, in Multinuclear NMR, edited by J. Mason, Chap. Z ! . Plenum Press, New York (1987). M. H. Cohen and F. Reif, in Solid Staie Physics, edited by F. Seitz a,id D. Turnbull, Vol. 5. p. 321. Academic Press, New York (1 957). B. C. Sanctuary and T. K. Halstead, Adv. Magn. Opt. Reson. 15, 79 [1990). T. C. Ng and E. Tward, J . Magn. Reson. 51. 361 (1983). K. G. A:;hurst, N. P. Finkelstein and L. A. Goold, J. Chem. Soc. A 1899 (1971). R, E. Wasylishen, R. E. Lenkinski and C. Rodger, Can. J. Chem. 160,2113 (1 982).