This work has been digitalized and published in 2013 by Verlag Zeitschrift für Naturforschung in cooperation with the Max Planck Society for the Advancement of Science under a Creative Commons Attribution4.0 International License.

Dieses Werk wurde im Jahr 2013 vom Verlag Zeitschrift für Naturforschungin Zusammenarbeit mit der Max-Planck-Gesellschaft zur Förderung derWissenschaften e.V. digitalisiert und unter folgender Lizenz veröffentlicht:Creative Commons Namensnennung 4.0 Lizenz.

NMR Studies on Halogeno-cis-bis(phosphine)-ruc?o-pentaboranyl Derivatives of Nickel, Palladium, and Platinum

John D. Kennedy* and John Staves Department of Inorganic and Structural Chemistry, University of Leeds, Leeds LS 2 9JT, England Z. Naturforsch. 34b, 808-813 (1979); received February 14, 1979 Metalloboranes, Multinuclear NMR, Double Resonance, Relaxation

An examination of the 1H-PB} , U B, "B-{1H}, and 31P-{!H} NMR behaviour of the compounds [M(B5H8)(X)(dppe)] has been made [M = Ni, Pd and Pt; X = halogen; dppe = bis(diphenylphosphino)ethane]. Some relaxation phenomena are rationalised, and it is concluded that all the compounds examined have a static metallo -nido-penta -borane structure in which the metal atom occupies a bridging position between two basal boron atoms. Some temperature and solvent effects are also presented and briefly discussed.

Introduction Recently we have reported a number of halogeno-

cis-bis(phosphine)metal derivatives of pentaborane (9), in which the metal was nickel(II) [1], palla-d i u m ^ ) , or platinum(II) [2], and which were prepared by the metathesis represented by equation (1). NMR evidence and analogy with the known [3] structure of the bis(triphenylphosphine)copper(I) analogue, [Cu(B5Hs)(PPh3)2], led to the conclusion that these had stereochemistry as in structure (I),

[B5H8]- + [M(X)2(PR3)2] [M(B5H8)(X)(PR3)2] + X " (1)

but certain aspects of the NMR behaviour were puzzling. For example, the XH and n B NMR data

(I)

available for the nickel(II) derivatives were inter-preted [1] in terms of a fluxional structure, in which the metal moiety was undergoing rapid exchange with the bridging hydrogen atoms on the penta-boranyl cluster. This was in contrast to the structure of a representative platinum(II) analogue which was found [2] to have a static structure, and it was not clear why the change in the metal atom should

* Reprint requests to Dr. J. D. Kennedy. 0340-5087/79/0600-0808/$ 01.00/0

cause this particular difference. Secondly, the n B NMR resonances for the basal boron atoms B(2)-B(5) in the nickel(II) compounds were appar-ently [1] much sharper than those for the platinum compounds [2] and again it was not clear why this should be so. For these and other reasons we have therefore examined the NMR behaviour of these compounds in greater detail and also at higher fre-quencies. We have chosen compounds with bis(di-phenylphosphino)ethane [dppe, Ph2PCH2CH2PPh2] as ligand since these are known [1, 2] to be relatively stable.

Experimental Compounds were prepared as previously reported

[1, 2] and their purity was checked by elemental analyses and by the monitoring of their iH, n B and 31P NMR spectra for spurious peaks. NMR measure-ments were carried out at the stated temperatures in the pulsed (Fourier transform) mode using a JEOL FX-100 spectrometer equipped for 1H, 1 1B-{1H}, and 3 1P-{ 1H} operation and modified for 1H-{11B} experiments. Low temperatures were often used to minimise decomposition during the course of an experiment, and this often had the advantage of increasing the signal-to-noise ratio somewhat. Chemical shifts are presented in the convention that resonances to high frequency (lower applied field) of the reference standard are positive; the standards used were Me4Si (internal), B F 3 O E t 2 (external) and 85% H 3 P0 4 (external) for m , n B and 31P respectively. Accumulation and data-processing parameters were chosen so that digital resolution and any effects due to artificial line-broadening were well within the error limits quoted. The 31P Ti measurements were made on a degassed sample using the 180°-T-90° pulse sequence and the quoted values were obtained from a semi-logarithmic plot. The 180° and 90° pulse-widths determined for the sample used were 32 and 16 //s, respectively.

J. D. Kennedy-J. Staves • NMR of Metallopentaboranes 809

at half-height for the high-frequency resonance envelopes in the chloro-compounds being ca. 440, 370, and > 700 Hz, respectively, and for the low-frequency resonances ca. 65, 75, and 105 Hz, respectively (saturated solutions in CD2CI2 at 23 °C). On broad-band irradiation at *H frequencies, the high-frequency resonances sharpened somewhat, and the low-frequency doublet became a singlet of similar linewidth to the components of the un-decoupled doublet. In contrast to previous reports for the nickel [1] and palladium [2] compounds, no multiplet structure for the larger, high-frequency resonance was apparent at 32 MHz, but experiments on [Pt(B5H8)(Cl)(dppe)] conducted at 64 MHz [4] distinguished two resonances of about equal inten-sity in this region with short quadrupolar relaxation times Ti which were of the order of 1 ms and which differed by a factor of approximately two (Table I) ; however, at 32 MHz 'partially relaxed' spectro-scopy [5] and line-narrowing failed to separate these satisfactorily. On storage in CDCI3 or CD2CI2 solu-tion at room temperature, samples of the nickel and palladium compounds developed additional U B (and iH) resonances attributable to pentaborane(9).

Table I. iH, n B and sip NMR data for [M(B5H8)(X)(dppe)]; saturated solutions in CD2C12 at the indicated temperatures.

M(X) Ni(Cl) Pd(Cl) Pt(Cl) Pt(Br) Pt(I)

iH spectra; r5(1H)/ppm ;a

H(2),H(3) H(4),H(5) H ( L ) ' H(jm-2,5) andH( / U -3 ,4 ) H(/i-4,5)

— 15°C ± 3.2b

+ 1.4b

— 0.5 — 0.3 — 1.2

+ 2.3° + 2.15c

— 0.1 — 1.95 — 2.2

± 2.8d>e

± 2.0e

± 0.2 — 2.0 — 2.18

± 2.75d

± 1.95 ± 0.25 — 1.9 — 1.78

± 2 . 6 d

± 2.0 ± 0.25 — 1.7 — 1.55«

U B spectra; <S(nB)/ppm;h

B(2), B(3)4

B(4), B(5)i B(l) 1J(UB-1H)/Hz; B(l)

4- 20 °C { ca. + 8i

— 44.7 ± 0 . 5 140 ± 5

— 1 1 . 3 ± 1.0e

— 1 3 . 7 ± 1.0e

— 43.5 ± 0 . 5 160 ± 5

,k,l ,k,m

31P spectra; (5(31P)/ppm;n

P(trans to X ) P(trans to B5H8)

— 20 °C + 65.8 ± 0 . 3 ± 41.2 ± 0.5

±67.1 ± 0 . 3 ±45.8 ± 0 . 5

±44 .2 ± 0 . 3 ± 52.9 ± 0 . 5

±46.4 ± 0 . 3 ± 52.9 ± 0 . 5

±47.1 ± 0 . 3 ±51 .4 ± 0 . 5

1 J(l95Pt-3iP)/Hz; P (trans to X) P{trans to B5H8)

— 50 °C - - 3673 ± IP

2440 ± 3P 3663 ± 1 2419 ± 3

3529 ± 1 2402 ± 3

2j (3ip_3ip)/Hz; — 20 °C 43 ± 1 10.5 ± 0 . 5 4.6 ± 0 . 1 3.5 ± 0 . 1 2.35 ±0 .1 a Measured at 100 MHz in ppm to high frequency (low field) of (CH3)4Si. b The resonance assigned to H(2), H(3)

was the broader. c Relative assignment uncertain. d These resonances had satellites due to 12J(195Pt-B-iH) | = 70 ± 10 Hz. e Interrelationship of XH and U B resonances tentatively established [4] by i H - { n B } experiments at 200 MHz. f Assigned by selective i H - { u B } experiments. & These are shoulders on the resonances ascribed to H(/<-2,5) and H(/7-3,4); at 200 MHz they are readily distinguishable [4]. h Measured at 32 MHz in ppm to high frequency (low field) of BF3 • OEt2. 1 Peaks generally too close and too broad for 1 J ( n B- iH) to be resolved, j Very broad resonances centered (approximately) on this value. k At 64 MHz; resonances indistinguishable at 32 MHz. 1 Ti ca. 2 ms. m Ti ca. 1 ms. n Measured at 40 MHz in ppm to high frequency (low field) of 85% H3PO4. p For the dichloro analogue [Pt(Cl)2(dppe)], iJ(195Pt-31P) = 3682 ± 1 Hz for a saturated solution in CD2C12 at —50 °C.

Results Boron-11 NMR spectra

The 32 MHz n B NMR spectra of each of the compounds examined consisted of a broad resonance, relative area 4, to high frequency (i. e. lower applied field), and a sharper doublet, relative area 1, to low frequency {i.e. higher applied field) (Fig. 1 and Table I). The breadth of the lines in the spectra was generally in the sequence P t ~ P d < N i , the widths

Fig. 1. 32 MHz n B (lower trace) and 11B-{1H(broad band)} (upper trace) NMR spectra of [Pt(B5H8)(Cl)(dppe)] in saturated CD2CI2 solu-tion at 23 °C.

810 J. D. Kennedy-J. Staves • NMR of Metallopentaboranes 810

Proton NMR spectra The 100 MHz (broad band)} NMR

spectra showed well-defined resonances for the bridging hydrogen atoms, but one or other of the terminal boron hydride resonances was always obscured by overlapping resonances due to the methylene protons of the phosphine ligand. How-

Fig. 2. Boron hydride region of a 100 MHz 1H-{UB(broad band)} NMR spectrum of [Pt(B5H8)(Br)(dppe)] in saturated C6Ü6 solu-tion at 24 °C (upper trace), and with the sub-traction of an equivalent undecoupled spectrum (lower trace).

+ 6 +3 0 -3

ever, subtraction of equivalent undecoupled spectra, in which the effects of coupling1 J(1 1B-1H) effectively reduce the terminal boron hydride resonances to a somewhat irregular baseline, removed from the spectrum those proton resonances which did not exhibit coupling to the boron nuclei, and thus permitted an unambiguous identification of the proton resonances due to the BöH8 moiety (Fig. 2 and Table I). Relative intensities and chemical shifts were used to assign the bridging and apical proton resonances, and selective 1 H-{ 1 1 B} experi-ments [2] confirmed the latter. At 100 MHz, the resonance of the proton in the fx-(4, 5) bridging position in the platinum compounds was not wrell-

resolved from that due to the ju-(3, 4) and /u-(2, 5) protons, but these were readily distinguishable at 200 MHz [4]. The basal terminal proton resonances has an intensity ratio of 2 :2 ; for the nickel and palladium compounds the assignment of these between the (2, 3) and the (4, 5) positions is un-certain, but one resonance for the nickel compound was somewhat the broader which may indicate that it originates from the (2, 3) positions nearer to the nickel atom and/or the centre of the molecule. For the platinum compounds one of these two resonances had satellites due to coupling nJ(1 9 5Pt-1H) of ca. 70 Hz; this is a reasonable value for 2J( 1 9 5Pt-E-iH) in platinum(II) compounds where E is a coordina-tively saturated first row element [8], and so these may be ascribed to the protons in the (2,3) positions. Selective 1H-{1 1B} experiments performed at 200 MHz [4] indicated that some preferential sharpening of these (2, 3) resonances occurred upon irradiation of the higher-frequency of the two basal n B resonances in [Pt(B5H8)(Cl)(dppe)], but there may be some ambiguity in this result due to the proximity and consequent overlap of the two boron resonances. All the borane proton resonances showed substantial decreases in shielding [viz. shifts to higher frequencies] Avhen benzene was used as solvent rather than chloromethanes (e.g. Table II ) ; there were concomitant increases in shielding of ca. 0.8 ppm for the methylene protons of the [dppe] ligand.

Phosphorus-31 NMR spectra The 40 MHz sip NMR spectra (Table I) of the

nickel, palladium, and platinum compounds at ambient temperatures consisted of two resonances -a sharp doublet and a broad resonance (e.g. Fig. 3):

Table II. Solvent effects in the *H NMR spectra of [Pt(B5H8)(X)(dppe)]; saturated solutions at + 20 °C; <5(1H) to high frequency of TMS.

H(2),H(3) H(4), H(5) H(l) H(/t-2,5) and H^-4,5) HGu-3,4)

[Pt(B5H8)(Cl)(dppe)] in CD2C12 + 0.05 ppm in CDCI3 ± 0.05 ppm in C6D6 + 0.03 ppm

+ 2.8 + 2.8 + 3.38

+ 2.0 + 2.1 + 2.91

+ 0.2 + 0.35 + 1.33

— 2.0 — 1.85 — 1.03

— 2.1 — 1.9 — 0.81

[Pt(B5H8)(Br)(dppe)] in CD2CI2 ± 0.03 ppm in CDCI3 + 0.03 ppm in C6I>6 ± 0.03 ppm

+ 2.73 + 2.80 + 3.46

+ 1.97 + 2.13 + 3.08

+ 0.23 + 0.37 + 1.42

— 1.89 — 1.78 — 0.79

— 1.70 — 1.58 — 0.63

B5H9 50% v/v in C6D6 ± 0.03 ppm f 2.38 + 0.69 — 2.57

J. D. Kennedy-J. Staves • NMR of Metallopentaboranes 811

200 Hz

Fig. 3. 40 MHz 3ip_{iH(broad band)} NMR spectra of [Pt(B5H8)(Cl)(dppe)] in saturated CD2C12 at 0 °C (lower trace), —55 °C (centre trace) and -—82 °C (upper trace). Each resonance shown is flanked (off-scale to high and low frequency) by satellites due to coupling !J(195Pt-3iP) as summarised in Table I.

for the platinum compounds these were flanked [2] by satellites due to coupling 1J(195Pt-31P). Line-narrowing experiments on the broader main reso-nance for [Pt(B5H8)(Cl)(dppe)] at —10 °C showed no fine structure other than the expected doublet feature with splitting ca. 4.5 Hz [2J(3 1P-M-3 1P)(as)] , and the spin-lattice relaxation times Ti at — 2 2 °C for the sharp and broad resonances were found to be 1.9 and 2.1 s, respectively. At lower temperatures the broad signals became sharper (e.g. Fig. 3) and the coupling 2J(3 1P-3 1P) became resolved. At — 82 °C, Ti for the sharp and the (originally) broader resonances for [Pt(B5H8)(Cl)(dppe)] were found to be 0.52 and 0.68 s respectively. There were small reversible changes in coupling constants and chemi-cal shift as the temperature was varied. The changes in 1J(195Pt-31P) for the three platinum compounds were investigated in some detail and in CD2C12

solution were found to be approximately linear with temperature, taking values of ca. + 0 . 1 2 Hz K _ 1 for the sharper resonance (i.e. that trans to halogen) and ca. —0.12 Hz K - 1 for the resonance trans to BsH8. For [Pt(Cl)2(dppe)] the temperature variation of this coupling was non-linear but monotonic, changing from ca. —0.25 Hz K _ 1 at —70 °C to ca. —0.08 Hz K _ 1 at + 20 °C. For comparison purposes, the phosphorus chemical shift of [Pt(Cl)2(dppe)], together with those for [Pd(Cl)2(dppe)] and [Ni(Cl)2(dppe)], were also measured and found to be + 41.7, +63.7 , and + 5 7 . 2 ppm respectively (±0 .3ppm; saturated solutions in CDC13 at + 23 °C).

Discussion The relative intensities and multiplicities of the

boron hydride proton resonances (Table I and Fig. 2)

are clearly consistent with the static structure (I) for all the representative nickel, palladium, and platinum compounds examined and there is no evidence for fluxional behaviour at ambient or near-ambient temperatures; it seems likely that the previous contrary conclusion [1] made for the nickel compounds arose because of decomposition of the solutions examined to yield pentaborane (9), prob-ably arising at least in part from protolysis by the acidic hydrogen atoms of the chloromethane sol-vents used; a similar phenomenon has been noted [4, 7] for the related copper(I) compound [Cu(B5H8)(PPh3)2].

There is no regular change in the proton shieldings as the metal is varied; this is not unexpected since these will be governed by a balance of local and non-local effects including anisotropic electronic circulation (and, in the case of the nickel compounds, possible paramagnetic delocalisation [8]) associated with the metal atoms which are not well understood. In the platinum compounds, however, there are uni-directional changes with increasing atomic weight of the halogen atom. These merit little comment although it may be pointed out that the chemical shifts of the bridging protons are as anticipated on the basis of structure (I) in which dispersion forces resulting in the restriction of the proton diamagnetic shielding would arise from close non-bonded approach of the polarizable halogen atom X . This clearly would not be the case for the other hypo-thetically possible isomeric structure which can be formally derived from (I) by the interchange of X and Pi but which in any event is unlikely on steric grounds. We have not found any evidence for any isomeric forms, other than that of structure (I).

The solvent effects on dissolution in benzene (Table II) are interesting in that all the borane proton resonances exhibited a considerably de-creased shielding; the accompanying shielding in-crease for the ligand methylene protons may there-fore indicate preferential transient solvation in which the benzene molecules lie above and below, and approximately parallel to, the diphospha-platinacyclopentane ring, and also insert between the geminal phenyl groups of the ligand.

The static structure (I) implies two U B resonances for the basal boron atoms but with the observed line widths these would have to be separated by several ppm to be distinguishable at 32 MHz. In the case of the platinum compounds, which have narrower

812 J. D. Kennedy-J. Staves • NMR of Metallopentaboranes 812

lines, they are just distinguishable at 64 MHz [4]. The short relaxation times which determine the broadness of these lines are of interest; a similar broadness has been noted [9] for the copper(I) com-pound [Cu(B5H8)(PPh3)2], but this is not the case for all metallo-m(Zo-pentaboranes. A survey of the literature for diamagnetic transition metal boranes in general shows that 'broad' or 'not well resolved' lines are generally associated with metals which have large ligands such as the aryl phosphines [10]; smaller groups such as carbonylmetal, or cyclo-pentadienylmetal, species are associated with sharper lines. It is unlikely that these ligand varia-tions will significantly affect the electric field gradients at the boron nuclei within the clusters, and so the shorter U B relaxation times associated with bulkier ligands must arise from the increased molecular reorientation correlation times of the larger molecules [10]; the [dppe] ligand for example has over six times the molecular weight and occupies 5 -6 times the volume of a cyclopentadienyl or nido-pentaboranyl ligand. On this basis, however, the greater broadness and different chemical shifts of the spectra for the nickel compound are not explained; in this case the effects may possibly arise from some equilibration with small amounts of a high-spin tetrahedral species [8], although in the absence of crystallographic evidence the possibility of a metallahexaborane [11] structure cannot be rigorously excluded.

The couplings 2J(3 1P-3 1P) measured from the 31P spectra (Table I) are as may be expected, and the significance of 1J(195Pt-31P) has been discussed elsewhere [2, 12, 13]. The 31P chemical shifts for [M(BsH8)(Cl)(dppe)] do not follow a linear sequence as the metal atom M is changed in the order Ni, Pd, Pt, but the measurements for the dichloro deriva-tives [M(Cl)2(dppe)] show that this is also to be expected.

Line-narrowing experiments indicate that the broadness of the lines due to the phosphorus nuclei trans to the borane cluster is probably not due to

partially resolved or unresolved coupling 2 J ( 3 1 P- n B) although we note that the observed n B relaxation times Ti together with the observed 31P line widths (ca. 35 Hz at 20 °C) for [Pt(B5H8)(Cl)(dppe)] would imply [14] a small but not unreasonable value for 2 J( 3 1 P-Pt - U B) (trans) of not more than ca. 10 Hz if this were the case. It is more likely that the domi-nant contribution to the broadness arises principally from scalar contributions to T2; the observed n B Ti values together with the 31P line-widths would imply a perhaps more reasonable [15] value for 2 J( 3 1 P- U B) of a few tens of Hz if this were so. although it is difficult to assess more precisely the effect on these processes of the complex spin systems involving U B and/or l 0B which result from the BsH8

cluster. The removal of the broadening at lower temperatures is consistent with either mechanism. It may be noted that any scalar contributions to Ti will be negligible for coupling between two dissimilar nuclides such as 31P and n B ; in the present case this was confirmed by Ti measurements on both the main 31P resonances in [Pt(BsH8)(Cl)(dppe)] which showed that both had very similar Ti values at both high and low temperatures, Avhich were within ranges typical [16] for bis(phosphine)platinum(II) compounds of this bulk and stereochemistry.

It was also interesting to observe small reversible changes in the coupling constants 1J(195Pt-31P) with temperature in the platinum compounds examined, but this phenomenon has been little investigated to date [17], and further comment would be speculative at this stage. It is however intriguing in this context that the temperature variations of 1J(195Pt-31P) for the phosphorus atom trans to halogen (positive) and trans to the pentaboranyl ligand (negative) occur in the opposite senses; all such temperature coeffi-cients previously reported [17] have been positive.

We thank the S. R. C. for support, Professor N. N. Greenwood for useful discussions, Dr. T. C. Gibb for assistance with computing, and Mr. A. Hedley for microanalyses.

[1] N. N. Greenwood and J. Staves, J. Chem. Soc. Dalton 1977, 1788.

[2] N. N. Greenwood, J. D. Kennedy, and J. Staves, J. Chem. Soc. Dalton 1978, 1146.

[3] N. N. Greenwood, J. A. Howard, and W. S. McDonald, J. Chem. Soc. Dalton 1978, 37.

[4] J. D. Kennedy and B. Wrackmeyer, unpublished observations.

[5] A. Allerhand, A. O. Clouse, R. R. Rietz, T. Roseberry, and R. Schaeffer, J. Am. Chem. Soc. 94, 2245 (1972); R. R. Rietz and R. Schaeffer. J. Am. Chem. Soc. 95, 4580 (1973).

J. D. Kennedy-J. Staves • NMR of Metallopentaboranes 813

[6] 2J(195Pt-C-1H) typically takes values o f — 6 5 to — 80 Hz in cis-bis(phosphine)methylplatinum(II) compounds; J. D. Kennedy, W. McFarlane, and R. J. Puddephatt, unpublished observations.

[7] S. G. Shore, personal communication. [8] E. de Boer and H. van Willigen, Progr. Nucl.

Magn. Reson. Spectrosc. 2, 111 (1967). [9] V. T. Brice and S. G. Shore, J. Chem. Soc. Dalton

1975, 334. [10] See, for example, E. L. Muetterties, Rev. Pure

Appl. Chem. 29, 585 (1972). [11] N. N. Greenwood, J. D. Kennedy, W. S. McDo-

nald, D. Reed, and J. Staves, J. Chem. Soc. Dalton 1979, 117.

[12] N. N. Greenwood, Pure Appl. Chem. 49, 791 (1977).

[13] J. D. Kennedy, Proc. X I X Internat. Conf. Coord. Chem., Prague, 1, 79 (September 1978).

[14] J. Bacon, R. J. Gillespie, and J. W. Quail, Can. J. Chem. 41, 3063 (1963).

[15] 12j(3ip-pt-i3c) (trans) | in [Pt(CH3)2(PMe2Ph)2] is 104 Hz; A. J. Cheney, B. E. Mann, and B. L. Shaw, Chem. Commun. 1971,431; the coupling will probably be somewhat smaller than this in com-pounds such as [Pt(C2H4)(PPh3)2] which has bonding more analogous to the metal-borane bonding of the compounds discussed here.

[16] J. D. Kennedy, unpublished observations. [17] K. R. Dixon, M. Fakley, and A. Pidcock, Can. J.

Chem. 54, 2733 (1976).