DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1999, 1917–1919 1917 Observation of triple-quantum eVects in the HMQC spectra of substituted derivatives of Rh6(CO)16 † Brian T. Heaton,a Jonathan A. Iggo,*a Ivan S. Podkorytov,*b Daniel J. Smawfield,a Sergey P. Tunik c and Robin Whymana a Department of Chemistry, University of Liverpool, PO Box 147, Liverpool, UK L69 7ZD. E-mail: iggo@liv.ac.uk b S.V. Lebedev Central Synthetic Rubber Research Institute, Gapsalskaya 1, St. Petersburg, 198035, Russia c Department of Chemistry, St.Petersburg University, Universitetskii pr., 2, St. Petersburg, 198904, Russia Received 15th March 1999, Accepted 5th May 1999 Triple-rhodium quantum effects in the HMQC spectra of [Rh6(CO)15L] [L 5 MeCN, I, PBun 3, P(OPh)3, P(4-XC6H4)3; X 5 H, OMe, F] modulate the intensity and position of the correlations in the rhodium dimension; cross peaks are displaced from the true chemical shift, additional cross peaks are seen, and the intensity of the coherences varies as a function of the mixing time and coupling constant, going to zero at the conventional value of 1/(2J).Inverse detected multiple quantum coherence (HMQC) experiments1 are now routinely used for the detection of insensitive nuclei such as 15N and 13C using 1H as the detector nucleus. More recently there has been a small number of reports using e.g. 31P as the detector nucleus to study metal nuclides such as 57Fe and 183W in inorganic systems.2–4 In all these experiments the low natural abundance of the insensitive nucleus I ensures that only single-quantum transitions of the I spin are important.Extension of the HMQC experiment to metal cluster compounds in which the insensitive (i.e. metal) nucleus has high natural abundance is not straightforward since the detector nucleus may now couple to several I spins. Multiple-quantum transitions, in which the assembly of metal spins acts as a unit, must now be considered.Although the eVect of multiplequantum transitions of the I spins is well understood,5 until now, it has been of little practical significance since, as mentioned above, studies using HMQC have concentrated on nuclei where multiple metal-quantum eVects cannot be observed because of the low natural abundance of the I spin. Ruegger and Moskau have recently presented a set of rules for the interpretation of 31P detected 195Pt–31P HMQC spectra (acquired without phase cycling to allow observation of resonances from multiple-quantum coherences normally suppressed by the phase cycling) of some platinum-phosphine systems in which the detector phosphorus nucleus can couple to two 195Pt.6 We now report that where the detector nucleus can couple to three metal spins, e.g.when a face-bridging carbonyl is used to detect 103Rh in rhodium carbonyl cluster compounds, not only are the correlations displaced from the “true” chemical shift in the rhodium dimension but also may be missing altogether as a result of modulation of the spectral intensity by the multiple metal-quantum transitions.Simple experimental modifications are described to overcome these eVects. The eVect of the HMQC pulse sequence, Fig. 1, on a facebridging carbonyl coupled to three rhodium atoms can be described using the product operator formalism.7,8 At the † Supplementary data available: the HMQC pulse program and phase cycling used in the experiments. For direct electronic access see http:// www.rsc.org/suppdata/dt/1999/1917/, otherwise available from BLDSC (No.SUP 57547, 2 pp.) or the RSC Library. See Instructions for Authors, 1999, Issue 1 (http://www.rsc.org/dalton). beginning of the evolution period t1, just after the first (p/2)– 103Rh pulse, the state of the four-spin face-bridging 13C103Rh3 system is given by eqn. (1) s = (0Rh) 1 (1Rh) 1 (2Rh) 1 (3Rh) (1) where (0Rh) = C(p/2 2 WCd2)cos3pJd2 (2) (1Rh) = 2C(p 2 WCd2) × {Rh1(p/2) 1 Rh2(p/2) 1 Rh3(p/2)}cos2pJd2sinpJd2 (3) (2Rh) = 4C(3p/2 2 WCd2) × (4) {Rh1(p/2)Rh2(p/2) 1 Rh1(p/2)Rh3(p/2) 1 Rh2(p/2)Rh3(p/2)} × cospJd2sin2pJd2 (3Rh) = 8C(2p 2 WCd2) × (5) Rh1(p/2)Rh2(p/2)Rh3(p/2)sin3pJd2 and C(f) and Rhi(f) denote the carbon and rhodium spin operators respectively where, for arbitrary phase f, and S is S(f) = Sxcosf 1 Sysinf (6) either C or Rhi.The chemical shift of carbon is WC, J is the coupling constant and d2 is the delay used in the pulse sequence for polarization transfer, Fig. 1.Spin operators belonging to the three diVerent rhodium atoms are numbered by the subscripts 1, 2, and 3. It is convenient to refer to the operators (0Rh), (1Rh), (2Rh) and (3Rh) as zero-, one-, two-, and three-rhodium spin coherences, respectively since we are interested principally in the order of the coherence with respect to rhodium. Fig. 1 Pulse sequence used for obtaining inverse detected HMQC 2D spectra.1918 J. Chem. Soc., Dalton Trans., 1999, 1917–1919 The zero-rhodium spin coherence (0Rh) contains no rhodium spin operators so cannot be used to correlate the carbon and rhodium chemical shifts.Magnetization arising from this term is eliminated by the phase cycling used in the standard HMQC experiment. The terms (1Rh), (2Rh) and (3Rh) evolve under the action of the rhodium chemical shifts, WRh1, WRh2 and WRh3 and can be used to correlate carbon and rhodium shifts. Cross peaks resulting from (2Rh), through zero- and double-rhodium quantum coherences, are suppressed by the standard phase cycling whilst (1Rh) gives cross peaks at the expected rhodium chemical shifts through single-rhodium quantum transitions.These cross peaks appear as 1:1:1:1 quartets in the carbon dimension, F2. The magnitude of each component of the quartets is given by eqn. (7) |(1/16)cos2pJd2sinpJd2| (7) and reaches a maximum of ÷3/72 when d2 = (1/pJ)arctg(÷2/2) ª 1/(5J) (8) Thus the maximal magnitude of these one-rhodium spin cross peaks is a factor of ca. 10 weaker than the maximal amplitude of those arising from a terminal Rh(CO) moiety (which, as can be shown, is equal to 1/4). Importantly, the maximum does not occur at the conventional mixing delay of 1/(2J), when the intensity of these cross peaks equals zero, but at a delay d2 that is ca. 2.5 times shorter. The evolution of the three-rhodium spin coherence (3Rh) under the action of the rhodium chemical shifts produces four cross peaks in the 2D spectrum, three single-quantum peaks and one triple-quantum.All the cross peaks appear as 1:3:3:1 quartets centred at WC in the carbon dimension, however in the rhodium dimension the three single-quantum peaks occur at 2WRh1 1 WRh2 1 WRh3, WRh1 2 WRh2 1 WRh3 and WRh1 1 WRh2 2 WRh3 whilst the triple quantum peak occurs at 2WRh1 2 WRh2 2 WRh3. The magnitude of the single- and triple-quantum peaks (the height of the inner quartet lines) is |(3/64)sin3pJd2| (9) and reaches a maximum of 3/64 at the conventional value of d2 = 1/(2J).The structure of, and labelling scheme for, [Rh6(CO)15{P(4- FC6H4)3}] are shown in Fig. 2.9 Figs. 3 and 4 show respectively the inverse detected HMQC 13C-{103Rh} spectra of this cluster recorded under standard conditions using mixing times of 1/(2J), J = 28 Hz, a typical value for 103Rh coupling to a facebridging 13CO, and 1/(5J).10 The only “correct” correlation (i.e. single-spin–single-quantum rhodium) seen in Fig. 3 is the correlation C(1)–Rh(B). All other “correct” correlations are too weak to be observed. All the remaining cross peaks seen are due to three-rhodium spin operators (triple-quantum and Fig. 2 Structure of and labelling scheme for [Rh6(CO)15{P(4-FC6- H4)3}]. single-quantum with respect to rhodium). Two of the threespin –single-quantum correlations are accidentally located at the expected coordinates of the ‘correct’ correlations C(2)– Rh(D) and C(3)–Rh(D) because the cluster contains two equivalent atoms Rh(B) and Rh(C).These “unexpected” correlations are reproduced in the simulated spectrum, Fig. 5. By contrast in Fig. 4, using the unconventional 1/(5J) mixing delay, cross peaks are observed at the ‘correct’ places even though the delay is suYciently far removed from the conventional value that no correlations might have been expected to be observed. These single-quantum (with respect to rhodium) transitions are produced by one-rhodium spin operators and are observed due to their intensities reaching a maximum at 1/(5J) whilst the intensity of correlations due to multiple-rhodium spin transitions is close to zero for this delay.Fortuitously, for rhodium Fig. 3 Experimental inverse detected HMQC 13C-{103Rh} NMR spectrum of [Rh6(CO)15{P(4-FC6H4)3}] (bridging region) obtained with a conventional delay d2 = 1/(2J) = 17.9 ms, J = 28 Hz. The “expected” correlations to the face-bridging carbonyls are weak or entirely absent and correlations due to three-rhodium spin coherences are seen.The projections in the rhodium and carbon dimensions and a 1D 13C spectrum are also shown. Fig. 4 Experimental inverse detected HMQC 13C-{103Rh} NMR spectrum of [Rh6(CO)15{P(4-FC6H4)3}] (face-bridging region) obtained with a non-conventional delay d2 = 1/(5J) = 7.14 ms, J = 28 Hz. Strong correlations to the bridging carbonyls are seen at the “correct” rhodium chemical shifts although the delay d2 used is “unconventional”.J.Chem. Soc., Dalton Trans., 1999, 1917–1919 1919 carbonyl clusters, couplings to the terminal carbonyls (70 Hz) are approximately 2.5 times greater than those to face-bridging carbonyls (28 Hz) allowing the one-rhodium spin correlations to both face-bridging and terminal carbonyls to be observed in a single experiment. These eVects are not limited to [Rh3(m-CO)] fragments but are expected whenever the detector nucleus is coupled to several metal spins that can act as a unit.Clearly care must be taken in the application and interpretation of HMQC experiments to the detection of insensitive nuclei in which the coupling of the detector nucleus to several insensitive spins can occur. In Fig. 5 Simulated HMQC spectrum for C(1)O for the case of a conventional delay d2 = 1/(2J) = 17.9 ms. J[Rh(A)–P] = 140 Hz, J[C(1)– Rh(A)] = 20 Hz, J[C(1)–Rh(B)] = 29 Hz, J[C(1)–Rh(C)] = 31 Hz. particular the spin system must be carefully analysed to determine the optimal value of the mixing delay to ensure that single metal spin transitions are observed.Acknowledgements The authors thank INTAS/RFBR, EPSRC and the University of Liverpool Research Development Fund for funding this work. D. J. S. thanks EPSRC for a studentship and B. T. H. thanks the Leverhulme Foundation for the award of a Research Fellowship. References 1 A. Bax, R. H. GriVey and B. L. Hawkins, J. Magn. Reson., 1983, 55, 301. 2 L. Carlton, Magn. Reson. Chem., 1997, 35, 153. 3 F. Lianza, A. Macchioni, P. Pregosin and H. Ruegger, Inorg. Chem., 1994, 33, 4999. 4 S. J. Berners-Price, R. J. Bowen, P. J. Harvey, P. C. Healy and G. A. Koutsantonis, J. Chem. Soc., Dalton Trans., 1998, 1743. 5 D. Nanz and W. von Philipsborn, J. Magn. Reson., 1991, 92, 560. 6 H. Ruegger and D. Moskau, Magn. Reson. Chem., 1991, 29, S 11. 7 O. W. Sorensen, G. W. Eich, M. H. Levitt, G. Bodenhausen and R. R. Ernst, Prog. Nucl. Magn. Reson. Spectrosc., 1983, 16, 163. 8 R. R. Ernst, G. Bodenhausen and A. Wokaun, Principles of Nuclear Magnetic Resonance in One and Two Dimensions, Oxford University Press, London, 1987. 9 S. P. Tunik, I. S. Podkorytov, B. T. Heaton, J. A. Iggo and J. V. Sampanthar, J. Organomet. Chem., 1998, 550, 222; S. P. Tunik, unpublished work. 10 S. Allevi, S. Bordoni, C. P. Clavering, B. T. Heaton, J. A. Iggo, C. Seregni and L. Garlaschelli, Organometallics, 1989, 8, 385. Communication 9/03513C