Towards the mechanism of heteroborane isomerisation: 1,2 → 1,2 and 1,2 → 1,7 low-temperature isomerisations from metallations of [5-I-7,8-Ph 2-7,8-nido-C2B9H8]2-

Article abstract of DOI:10.1016/j.ica.2004.12.003

The reaction between [5-I-7,8-Ph₂-7,8-nido-C₂B ₉H₈]^2- and NiCl₂(dppe) affords 1,2-Ph₂-4,4-dppe-12-I-4,1,2-closo-NiC₂B₉H ₈ (1) and 1,8-Ph₂-2,2-dppe-10-I-2,1,8-closo-NiC ₂B₉H₈ (2). Reaction between the same carborane ligand and cis-PtCl₂(PMe₂Ph)₂ yields three species, 1,8-Ph₂-2,2-(PMe₂Ph)₂-10-I-2,1,8- closo-PtC₂B₉H₈ (3), 1,8-Ph₂-2,2- (PMe₂Ph)₂-12-I-2,1,8-closo-PtC₂B ₉H₈ (4), and 1,8-Ph₂-2,2-(PMe ₂Ph)₂-7-I-2,1,8-closo-PtC₂B₉H ₈ (5). Compounds 1-5 have been characterised spectroscopically and crystallographically. The 4,1,2-MC₂B₉ architecture of 1 constitutes a 1,2 → 1,2 cage C atom isomerisation, and the 2,1,8-MC₂B₉ architectures of 2-5 a 1,2 → 1,7 cage C atom isomerisation, relative to the presumed first product of the metallations, 1,2-Ph₂-3,3-L₂-9-I-3,1,2-closo-MC₂B ₉H₈ [M = Ni, L₂ = dppe; M = Pt, L₂ = (PMe₂Ph)₂]. The location of the (iodide) labelled boron vertex in the products allows speculation as to the mechanism of these isomerisations and the possible involvement of triangle face rotation is discussed.

Full text of DOI:10.1016/j.ica.2004.12.003

Inorganica Chimica Acta 358 (2005) 1485–1493
www.elsevier.com/locate/ica
Towards the mechanism of heteroborane isomerisation: 1,2 ! 1,2
and 1,2 ! 1,7 low-temperature isomerisations from metallations
of [5-I-7,8-Ph2-7,8-nido-C2B9H8]2ꢀ q
Susan Robertson, Rhona M. Garrioch, David Ellis, Thomas D. McGrath,
*
Bruce E. Hodson, Georgina M. Rosair, Alan J. Welch
School of Engineering and Physical Sciences, Heriot-Watt University, Edinburgh EH14 4AS, UK
Received 15 July 2004; accepted 2 December 2004
Dedicated to Professor F. Gordon A. Stone in recognition of his outstanding contributions to chemistry and with gratitude for his guidance
in the careers of T.D.M., B.E.H. and A.J.W
Abstract
The reaction between [5-I-7,8-Ph2-7,8-nido-C2B9H8]2ꢀ and NiCl2(dppe) affords 1,2-Ph2-4,4-dppe-12-I-4,1,2-closo-NiC2B9H8 (1)
and 1,8-Ph2-2,2-dppe-10-I-2,1,8-closo-NiC2B9H8 (2). Reaction between the same carborane ligand and cis-PtCl2(PMe2Ph)2 yields
three species, 1,8-Ph2-2,2-(PMe2Ph)2-10-I-2,1,8-closo-PtC2B9H8 (3), 1,8-Ph2-2,2-(PMe2Ph)2-12-I-2,1,8-closo-PtC2B9H8 (4), and
1,8-Ph2-2,2-(PMe2Ph)2-7-I-2,1,8-closo-PtC2B9H8 (5). Compounds 1–5 have been characterised spectroscopically and crystallograph-
ically. The 4,1,2-MC2B9 architecture of 1 constitutes a ‘‘1,2 ! 1,2’’ cage C atom isomerisation, and the 2,1,8-MC2B9 architectures of
2–5 a 1,2 ! 1,7 cage C atom isomerisation, relative to the presumed first product of the metallations, 1,2-Ph2-3,3-L2-9-I-3,1,2-closo-
MC2B9H8 [M = Ni, L2 = dppe; M = Pt, L2 = (PMe2Ph)2]. The location of the (iodide) labelled boron vertex in the products allows
speculation as to the mechanism of these isomerisations and the possible involvement of triangle face rotation is discussed.
ꢀ 2004 Elsevier B.V. All rights reserved.
Keywords: Metallacarborane; Synthesis; NMR spectroscopy; X-ray crystallography; Isomerisation mechanisms; Vertex labelling
1. Introduction
We have found that appropriate metallation of the
11-vertex nido carborane [7,8-Ph2-7,8-nido-C2B9H9]2ꢀ
and its derivatives reliably affords isomerised products
at or near room temperature. Use of a {PtL2}2+ me-
tal–ligand fragment (L = phosphine) tends to give
2,1,8-PtC2B9 products [2], this representing a net
1,2 ! 1,7 C atom isomerisation. On the other hand,
use of analogous {NiL2}2+ fragments tends to yield
4,1,2-NiC2B9 species, by a ‘‘1,2 ! 1,2’’ C atom rear-
rangement [3]. An unusual and significant result came
from our use of the {Mo(g-C3H5)(CO)2}+ fragment,
giving a 2,1,8-MoC2B9 product but via an isolable
non-icosahedral intermediate [4], which affords a sign-
post to the rearrangement mechanism, at least in this
case.
Sterically crowded transition metal metallacarbor-
anes are analogues of carboranes with the potential to
undergo low-temperature isomerisation. Thus, such
compounds can be of value in experimental study of
the mechanism(s) of isomerisation since the integrity
of vertex-label bonds can be reasonably assured under
the mild conditions of such isomerisations [1].
q
Steric effects in heteroboranes. Part 30. For part 29, see [9].
Corresponding author. Tel.: +44 131 451 3217; fax: +44 131 451
*
3180.
E-mail address: a.j.welch@hw.ac.uk (A.J. Welch).
0020-1693/$ - see front matter ꢀ 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.ica.2004.12.003

1486
S. Robertson et al. / Inorganica Chimica Acta 358 (2005) 1485–1493
Developing these early studies by labelling specific
assigned to the labelled B atom. This chemical shift
range is reminiscent of 4,1,2-NiC2B9 species afforded
by the ‘‘1,2 ! 1,2’’ C atom isomerisation of bis(phos-
phine)nickelacarboranes, viz. +12 to ꢀ16 ppm in 1,2-
Ph2-4,4-(PMe2Ph)2-4,1,2-closo-NiC2B9H9 [3], +12 to
ꢀ17 ppm in 1,2-Ph2-4,4-(PEt3)2-4,1,2-closo-NiC2B9H9
[3], +14 to ꢀ15 ppm in 1,2-Ph2-4,4-dppe-4,1,2-closo-
NiC2B9H9 [3] and +13 to ꢀ14 ppm in 1,2-Ph2-4,4-
dppe-6-Et-4,1,2-closo-NiC2B9H8 [8]. In compound 2
are six resonances, 1:3:1:1:2:1 from high to low fre-
quency, spanning the chemical shift range ꢀ1 to ꢀ28
ppm. The absence of a high frequency resonance is
inconsistent with a 4,1,2-NiC2B9 architecture but much
more in keeping with a 2,1,8-heteroatom pattern (arising
from net 1,2 ! 1,7 C atom isomerisation). Except for
the lowest frequency resonance (shown from an 11B
spectrum to be due to the I-labelled B atom) the 11B
chemical shift range in 2 is ꢀ1 to ꢀ17 ppm, exactly
the same as in 1,8-Ph2-2,2-dppe-4-Et-2,1,8-closo-
NiC2B9H8 [8]. That the iodide label causes a pro-
nounced upfield shift in the 11B resonance of the
attached boron atom has been previously observed;
compare the chemical shifts of B9 in 1,2-Ph2-9-I-1,2-clo-
so-C2B10H9 [10] (ꢀ15.2 ppm) and B9/B12 in 1,2-Ph2-
9,12-I2-1,2-closo-C2B10H8 [11] (ꢀ10.0 ppm) with B9/
B12 in 1,2-Ph2-1,2-closo-C2B10H10 [12] (ꢀ2.6 ppm).
Although these NMR studies strongly suggested,
therefore, that 1 and 2 had, respectively, 4,1,2-NiC2B9
and 2,1,8-NiC2B9 geometries, definitive structural char-
acterisations, and the important location of the iodide
labels, were only realised by single crystal X-ray diffrac-
tion studies.
boron vertices in the nido carborane ligand led to inter-
esting results. The use of SMe2 labels at positions 9 [5]
and 10 [6] gave neutral products with {Mo(g-
C3H5)(CO)2}+ (again, via non-icosahedral intermedi-
ates), and analysis of these led us to conclude broad
(but not perfect) support for the sequential diamond–
square–diamond (dsd) mechanism of Wales [7]. Nickela-
tion of carborane labelled with Et at vertex 3 gave two
products, a 4,1,2-NiC2B9 species and (unusually for
nickel) a 2,1,8-NiC2B9 species [8]. A similar 2,1,8 species
was afforded by platination of the same carborane, and
of its analogue with an F label, but also isolated was a
slipped, non-isomerised intermediate in the Et case [9].
The position of the label in the 2,1,8-Ni/PtC2B9 species,
on vertex 4 of the closo icosahedron, does not accord
with that predicted by the sequential dsd process. On
the other hand, all the Ni and Pt products could be
rationalised by a single triangle face rotation (tfr) of
the presumed first product of the metallation, a crowded
3,1,2-MC2B9 species.
Seeking further insight into the mechanism of isom-
erisation of heteroboranes, we now report the results
of both nickelation and platination of a further labelled
carborane, [5-I-7,8-Ph2-7,8-nido-C2B9H8]2ꢀ
.
2. Results and discussion
2.1. Nickelacarboranes
The
reaction
between
[5-I-7,8-Ph2-7,8-nido-
C2B9H8]2ꢀ, prepared by deprotonation of [5-I-7,8-Ph2-
7,8-nido-C2B9H9]ꢀ with n-BuLi under THF (THF =
tetrahydrofuran) reflux, and Ni(dppe)Cl2 [dppe = 1,2-
bis(diphenylphosphino)ethane] in THF affords red 1,2-
Ph2-4,4-dppe-12-I-4,1,2-closo-NiC2B9H8 (1), and purple
1,8-Ph2-2,2-dppe-10-I-2,1,8-closo-NiC2B9H8 (2), in
modest yields (not optimised) following work-up involv-
ing thin layer chromatography. Compounds 1 and 2
were initially characterised by microanalysis, and IR
and NMR spectroscopies, and ultimately by X-ray dif-
fraction studies.
The 1H NMR spectra of both compounds are
relatively uninformative, save confirming the expected
relative integrals of phenyl and methylene protons.
The 31P–{1H} spectra reveal that in both compounds
rotation of the {Ni(dppe)} fragment about the nickel-
cage axis is restricted since two resonances are observed
(with two-bond P–P coupling resolved only for 1). The
11B–{1H} NMR spectra, on the other hand, offer strong
clues as to the structural identity of these molecules. For
1 there are five resonances between +12 and ꢀ17 ppm,
with relative integrals (high to low frequency)
1:1:1:4:2. The middle resonance, ꢀ7.30 ppm, remains a
singlet on retention of proton coupling and is therefore
A perspective view of a single molecule of 1 is shown
in Fig. 1, and Table 1 hosts selected molecular parame-
ters. Although the study is of relatively poor precision (a
consequence of poor crystal quality) the heteroatom
pattern is confirmed as 4,1,2-NiC2 and the iodide label
is attached to B12. The Ni atom is slightly slipped [13]
˚
away from C1 (D = 0.11 A) and the metal-bonded B4C
face is slightly folded (5.9ꢁ) about B3ꢁ ꢁ ꢁB5 into an enve-
lope conformation; these distortions result in Ni4–C1
being the longest Ni–cage distance. The plane of the
{NiP2} fragment is somewhat twisted from perpendicu-
larity to the perpendicular bisector of the metal-bonded
B4C ring (dihedral angle 72.4ꢁ). The cage-bound phenyl
rings are orientated with low hPh values [14] (9.2ꢁ for
C101–C106 and 4.8ꢁ for C201–C206). Similar structural
features to all those above have been noted in 1,2-Ph2-
4,4-L2-4,1,2-closo-NiC2B9 species studied previously
˚
[3,8]. The B–I distance, 2.225(19) A, compares well with
1
that in the precursors.
1
˚
B–I = 2.178(4) A in 1,2-Ph2-9-I-1,2-closo-C2B10H9 [ꢀ10]. Unfortu-
˚
nately, the B–I distance in [5-I-7,8-Ph2-7,8-nido-C2B9H9] , 2.211(7) A,
was omitted from Table 4 of [11].

S. Robertson et al. / Inorganica Chimica Acta 358 (2005) 1485–1493
1487
Fig. 1. Perspective view of 1,2-Ph2-4,4-dppe-12-I-4,1,2-closo-
NiC2B9H8 (1). Atoms are drawn as 50% probability ellipsoids except
for H atoms.
Fig. 2. Perspective view of 1,8-Ph2-2,2-dppe-10-I-2,1,8-closo-
NiC2B9H8 (2). Atoms are drawn as 50% probability ellipsoids except
for H atoms.
Table 1
˚
Selected interatomic distances (A) and interbond angles (ꢁ) in 1
C1–C2
C1–B3
C1–Ni4
C1–B5
C1–B6
C2–B3
C2–B7
C2–B11
C2–B6
B3–B7
B3–B8
B3–Ni4
Ni4–B8
Ni4–B9
Ni4–B5
B5–B9
B5–B10
B5–B6
1.72(2)
1.68(2)
2.23(2)
1.73(3)
1.72(3)
1.78(3)
1.68(2)
1.67(2)
1.67(3)
1.80(2)
1.88(3)
2.098(18)
2.11(2)
2.181(17)
2.076(18)
1.83(3)
1.78(2)
1.78(3)
B6–B11
B6–B10
B7–B8
1.73(3)
1.76(3)
1.85(3)
1.79(3)
1.79(3)
1.80(2)
1.76(2)
1.78(3)
1.74(3)
1.78(3)
1.73(3)
1.77(3)
1.48(2)
1.52(3)
2.194(5)
2.151(4)
2.225(19)
88.05(17)
Table 2
B7–B12
B7–B11
B8–B9
˚
Selected interatomic distances (A) and interbond angles (ꢁ) in 2
C1–Ni2
C1–B3
C1–B4
C1–B5
C1–B6
Ni2–B3
Ni2–B7
Ni2–B11
Ni2–B6
B3–B7
B3–C8
B3–B4
B4–C8
B4–B9
B4–B5
B5–B9
B5–B10
B5–B6
2.295(7)
B6–B11
B6–B10
B7–C8
1.872(12)
1.800(13)
1.715(11)
1.749(12)
1.743(13)
1.722(12)
1.723(13)
1.764(14)
1.782(12)
1.766(12)
1.763(13)
1.743(12)
1.507(10)
1.531(12)
2.183(2)
B8–B12
B9–B10
B9–B12
B10–B11
B10–B12
B11–B12
C1–C101
C2–C201
Ni4–P1
Ni4–P2
B12–I12
P1–Ni4–P2
1.697(12)
1.684(11)
1.698(11)
1.720(12)
2.068(9)
B7–B12
B7–B11
C8–B9
2.097(9)
2.117(9)
2.080(9)
C8–B12
B9–B10
B9–B12
B10–B11
B10–B12
B11–B12
C1–C11
C8–C81
Ni2–P1
Ni2–P2
B10–I10
P1–Ni2–P2
1.864(12)
1.765(11)
1.827(12)
1.715(12)
1.769(13)
1.753(14)
1.762(13)
1.776(13)
1.795(12)
2.196(2)
2.172(9)
87.60(8)
The molecular structure of 2 is shown in Fig. 2, with
Table 2 hosting key molecular parameters. Clearly, now,
the cage carbon atoms have separated and the hetero-
atom pattern is 2,1,8-NiC2, as suggested from the 11B
NMR spectrum. Apart from the positions of their labels
other cage-bound Ph group, being able to achieve an
optimal conformation with respect to the {NiP2}
fragment.
˚
[I is attached to B10 in 2, B10–I10 2.172(9) A], com-
pound 2 is thus comparable with 1,8-Ph2-2,2-dppe-4-
Et-2,1,8-closo-NiC2B9H8 [8], the only other structurally
characterised 2,2-L2-2,1,8-NiC2B9 compound. The Ni
2.2. Platinacarboranes
˚
atom is more slipped (D = 0.19 A) and the metal-bonded
B4C face is more folded (11.9ꢁ) than in 1, but on the
other hand the {NiP2} fragment shows less deviation
from perpendicularity to the plane through C1B9B12
(dihedral angle 88.9ꢁ). Possibly this results from the Ph
ring on C1, now freed from the steric influence of the
When [5-I-7,8-Ph2-7,8-nido-C2B9H8]2ꢀ is treated with
cis-PtCl2(PMe2Ph)2 in THF three products are formed.
Compounds 3 and 4 cannot be separated by chromato-
graphic means but were eventually separated by crystal-
lisation. Their combined yield is ca. 25%, but the third

1488
S. Robertson et al. / Inorganica Chimica Acta 358 (2005) 1485–1493
product, 5, is obtained only in minor amounts (yields
not optimised). Again, initial characterisation was by
microanalytical and spectroscopic means, with ultimate
recourse to crystallography.
1
The H NMR spectra of all three products confirm
the presence of Ph and CH3 groups in the expected ratio.
The Me resonances appear as complex multiplets, as-
cribed to prochirality and two-bond P–H coupling.
The 31P–{1H} spectrum of 3 appears as a singlet (with
Pt satellites) under normal conditions, but reduction of
the line-broadening parameter reveals an AB spectrum
2
with JPP ca. 27 Hz. For 4 and 5 the 31P–{1H} spectra
appear as two clear doublets (with appropriate Pt satel-
2
lites) with JPP ca. 27 and 25 Hz, respectively. The 11B–
{1H} NMR spectra of compounds 3–5 all show
resonances between ꢀ1 and ꢀ30 ppm. Significantly
there is no high frequency resonance around +20 ppm.
This suggests that 3–5 are all thermodynamic isomers
with separated cage C atoms, similar in overall form
to 1,8-Ph2-2,2-(PMe2Ph)2-2,1,8-closo-PtC2B9H9 [2,15]
and 1,8-Ph2-2,2-(PMe2Ph)2-4-Et-2,1,8-closo-PtC2B9H8
[9], but fundamentally different to the C-adjacent, ki-
netic isomer 1,2-Ph2-3,3-(PMe2Ph)2-6-Et-3,1,2-closo-
PtC2B9H8 [9]. For 3 and 4 the pattern of 11B resonances
is 2:2:2:1:1:1 and for 5 1:1:1:2:1:2:1 (high to low
frequency) with (in light of the above discussion) the
integral-2 resonances presumably arising from co-
incidences. Only for compound 3 was it possible to dis-
cern which resonance (the lowest frequency in this case)
was due to the labelled B atom.
Fig. 3. Perspective view of 1,8-Ph2-2,2-(PMe2Ph)2-10-I-2,1,8-closo-
PtC2B9H8 (3). Atoms are drawn as 50% probability ellipsoids except
for H atoms.
Molecular structures of compounds 3, 4 and 5 are
shown in Figs. 3–5. respectively, whilst Table 3 summa-
rises key distances and the P–Pt–P angles. Thus,
compounds 3–5 are shown to be 1,8-Ph2-2,2-(PMe2Ph)2-
10-I-2,1,8-closo-PtC2B9H8, 1,8-Ph2-2,2-(PMe2Ph)2-12-I-2,
1,8-closo-PtC2B9H8, and 1,8-Ph2-2,2-(PMe2Ph)2-7-I-2,1,
8-closo-PtC2B9H8, respectively. All three compounds
do, indeed, share a 2,1,8-PtC2B9 architecture afforded
by net 1,2 ! 1,7 isomerisation of the cage C atoms.
They differ, however, in the location of the I label; in
˚
3 on B10, B10–I10 2.204(12) A; in 4 on B12, B12–I12
˚
˚
2.169(9) A; and in 5 on B7, B7–I7 2.197(7) A. Apart
from the label, the cage geometries of compounds 3–5
are similar. In each case the Pt atom is significantly
˚
slipped away from C1 (D values 0.36, 0.37 and 0.33 A,
respectively) and the metal-bonded B4C face is signifi-
cantly folded about B3ꢁ ꢁ ꢁB6 (by 14.5ꢁ, 15.5ꢁ and 13.2ꢁ,
respectively), causing Pt2–C1 to be by far the longest
Fig. 4. Perspective view of 1,8-Ph2-2,2-(PMe2Ph)2-12-I-2,1,8-closo-
PtC2B9H8 (4). Atoms are drawn as 50% probability ellipsoids except
for H atoms.
˚
of the Pt–cage atom distances, ca. 2.6 A versus ca. 2.2
˚
A otherwise. In all cases the {PtP2} plane stands effec-
tively perpendicular (dihedral angles 89.2ꢁ, 77.3ꢁ and
84.6ꢁ, respectively) to the plane through C1B9B12.
The factors responsible for these conformational prefer-
ences and slipping and folding distortions are well doc-
umented [13,16]. Beyond the position of the iodide label,
one minor difference between the three structures is that
in 3 the {Pt(PMe2Ph)2} fragment has effective Cs sym-
metry with both Ph groups on the same side of the
PtP2 plane, whilst in both 4 and 5 this fragment has
effective C2 symmetry (Ph groups on different sides).

S. Robertson et al. / Inorganica Chimica Acta 358 (2005) 1485–1493
1489
Table 3
Selected distances (A) and angles (ꢁ) in 3, 4 and 5
˚
3
4
5
˚
Distances (A)
C1–Pt2
C1–B3
C1–B4
C1–B5
C1–B6
Pt2–B3
Pt2–B7
Pt2–B11
Pt2–B6
B3–B7
B3–C8
B3–B4
B4–C8
B4–B9
B4–B5
B5–B9
B5–B10
B5–B6
B6–B11
B6–B10
B7–C8
B7–B12
B7–B11
C8–B9
C8–B12
B9–B10
B9–B12
B10–B11
B10–B12
B11–B12
B–I
2.617(9)
2.606(6)
2.587(6)
1.686(9)
1.652(9)
1.671(9)
1.707(9)
2.232(7)
2.207(8)
2.241(7)
2.289(7)
1.858(9)
1.760(9)
1.826(10)
1.752(9)
1.750(10)
1.731(11)
1.745(10)
1.795(10)
1.819(10)
1.886(10)
1.807(10)
1.742(9)
1.783(11)
1.808(11)
1.732(10)
1.734(10)
1.794(10)
1.766(11)
1.795(11)
1.771(12)
1.781(11)
2.197(7)
1.506(8)
1.502(10)
2.3085(18)
2.2934(18)
1.681(13)
1.641(14)
1.670(13)
1.722(14)
2.237(10)
2.210(9)
1.706(13)
1.659(10)
1.679(14)
1.692(13)
2.217(10)
2.222(8)
2.227(10)
2.315(10)
1.862(15)
1.757(14)
1.814(15)
1.718(15)
1.746(17)
1.739(16)
1.775(15)
1.742(16)
1.798(15)
1.887(14)
1.782(16)
1.713(13)
1.754(15)
1.782(15)
1.742(14)
1.702(14)
1.781(16)
1.753(15)
1.770(16)
1.770(16)
1.756(14)
2.204(12)
1.491(13)
1.491(13)
2.289(2)
2.237(9)
2.295(9)
1.861(13)
1.757(12)
1.826(13)
1.734(12)
1.750(14)
1.793(13)
1.768(13)
1.775(14)
1.809(13)
1.867(13)
1.796(13)
1.688(11)
1.765(13)
1.804(12)
1.769(12)
1.720(12)
1.789(13)
1.743(13)
1.768(13)
1.760(13)
1.760(13)
2.169(9)
Fig. 5. Perspective view of 1,8-Ph2-2,2-(PMe2Ph)2-7-I-2,1,8-closo-
PtC2B9H8 (5). Atoms are drawn as 50% probability ellipsoids except
for H atoms.
Although the P–Pt–P angle in 3 is ca. 4ꢁ greater than
that in 4 and 5 consideration of the wider literature
[2,9,15] appears to show no real correlation between
{Pt(PMe2Ph)2} conformation and P–Pt–P angle.
C1–C11
C8–C81
Pt2–P1
Pt2–P2
1.489(13)
1.499(10)
2.294(2)
2.3. Mechanistic implications
2.305(3)
2.270(2)
The unambiguous location of the labelled vertex in
the isomerised products 1–5 allows speculation of the
isomerisation mechanism(s), but only if the integrity of
the B–I bond in the reactions is assured. The precursor
ion [5-I-7,8-Ph2-7,8-nido-C2B9H9]ꢀ is deprotonated un-
der THF reflux for 3 h following which all further reac-
tions and procedures are carried out at or below room
temperature. In contrast, the precursor ion is prepared
(in high yield) from 1,2-Ph2-9-I-1,2-closo-C2B10H9 under
EtOH reflux for 72 h [11] throughout which the B–I
bond remains intact. It is reasonable, therefore, to as-
sume that it similarly remains intact under the relatively
benign conditions used to produce 1–5.
It is further reasonable to assume that the first prod-
uct of both the nickelation and platination of [5-I-7,8-
Ph2-7,8-nido-C2B9H8]2ꢀ with {ML2}2+ fragments is
1,2-Ph2-3,3-L2-9-I-3,1,2-closo-MC2B9H8 (I), but that
this species is too internally crowded to survive and
isomerises to relieve that crowding. We recently [9] re-
ported the first isolation of a 1,2-Ph2-3,3-L2-3,1,2-clo-
so-MC2B9 species and were able to demonstrate that
this subsequently isomerised under mild conditions.
How, then, might I rearrange to give compounds 1–5?
Angles (ꢁ)
P1–Pt2–P2
98.34(10)
94.69(8)
94.18(7)
In compound 1, the cage C atoms retain their adja-
cency but C2 has moved away from the adjacent dppe
ligand and now occupies a position in the ‘‘lower’’ (with
respect to the Ni atom) CB4 pentagon. The formation of
this species can be explained by a single tfr (i), shown in
Scheme 1, in which the highlighted MB2 face is rotated
by 120ꢁ. We recently [8] rationalised the formation of
1,2-Ph2-4,4-dppe-6-Et-4,1,2-closo-NiC2B9H8 by rotation
of the same triangular face. In nickelacarborane 2 the
cage C atoms have become non-adjacent and occupy
vertices 1 and 8 of the icosahedron, with the Ni atom
at 2 and the labelled B vertex at 10. Whilst this precise
heteroatom and labelled vertex pattern is that predicted
by Walesꢀ sequential dsd mechanism [7], it can also be
explained by rotation of a CB2 triangular face (ii). This
is the same tfr we have previously [8,9] employed to
rationalise the formation of 1,8-Ph2-2,2-dppe-4-Et-
2,1,8-closo-NiC2B9H8 and 1,8-Ph2-2,2-(PMe2Ph)2-4-X-

1490
S. Robertson et al. / Inorganica Chimica Acta 358 (2005) 1485–1493
M
C
M
Ph
Ph
Ph
Ph
Ph
Ph
Ph
M
Ph
Ph
C
C
C
C
C
C
C
C
C
i
I
I
I
I
I
I
I
Ph
I
I
I
I
I
1
M
M
M
Ph
Ph
C
ii
I
C
Ph
2, 3
M
Ph
Ph
Ph
Ph
C
C
C
C
C
iii
iii
v
I
Ph
C
4
M
M
M
M
M
Ph
Ph
Ph
I
C
C
iv
vi
I
Ph
C
C
C
Ph
5
5
M
Ph
Ph
I
I
C
C
C
Ph
Scheme 1. Possible formation of metallacarboranes 1–5 by one or two triangle face rotations starting from a common transient species I. In each tfr
the face highlighted in bold is assumed to rotate by 120ꢁ.
2,1,8-closo-PtC2B9H8 (X = Et, F), species which cannot
be understood by the Wales mechanism.
tfr routes. Tfr iii generates compound 4, as noted above,
and a subsequent tfr (iv) involving rotation of a B3 trian-
gle, would produce 5. However, compound 4 is recov-
ered unaltered from heating to reflux a THF solution
for 4 h, suggesting that a mechanism involving 4 ! 5
is not realistic. An alternative route from I to 5 could in-
volve first rotation of a B3 face (v) and then a CB2 face
(vi). It is certainly possible that the intermediate species
in this latter route, having iodide and two phenyl sub-
stituents upward pointing from the metallabonded
C2B3 ligand face, would be prone to further rearrange-
ment through steric crowding.
Platinacarborane 3 is structurally analogous to nick-
elacarborane 2, and so it, too, can be understood by the
CB2 tfr ii. Compound 4, also not predicted by the
sequential dsd rearrangement mechanism, is readily
understood by tfr iii, which, were it not for the I label,
would be the symmetry-equivalent complement to tfr
ii. Note that compounds 1–4 could instead be rationa-
lised by alternate single tfrs, shown in Scheme 2. In a
the C2B face is rotated, whilst in b and c the rotating
face is CB2 involving two B atoms from the lower pen-
tagon. However, whilst recognising that triangle face
rotations i–iii in Scheme 1 are not unique in their ability
to explain the formation of compounds 1–4, we are at-
tracted to them on the basis of precedent, these same tfrs
having previously (and exclusively) been used to ratio-
nalise earlier nickela- and platinacarboranes [8,9].
The platinacarborane isolated in lowest yield, com-
pound 5, cannot be generated by a single tfr from tran-
sient species I, and Scheme 1 shows two possible double
3. Conclusions
This paper has reported spectroscopic and crystallo-
graphic characterisation of two isomerised nickelacar-
boranes and three isomerised platinacarboranes
formed from low-temperature metallations of the la-
belled
carborane
[5-I-7,8-Ph2-7,8-nido-C2B9H8]2ꢀ
.

S. Robertson et al. / Inorganica Chimica Acta 358 (2005) 1485–1493
1491
quent manipulation in the open laboratory. Solvents
were freshly distilled over CaH2(CH2Cl2) or Na wire
(THF, Et2O, 40–60 petroleum ether) and were degassed
(3 · freeze–pump–thaw cycles) before use, or were
M
C
M
Ph
Ph
Ph
Ph
Ph
Ph
Ph
C
C
C
C
C
C
C
a
b
c
I
I
I
I
˚
stored over 4 A molecular sieves (CDCl3). Preparative
thin layer chromatography (TLC) employed 20 · 20
cm Kieselgel 60 F254 glass plates. IR spectra were re-
corded from KBr disks using either a Nicolet Impact
400 or a Perkin–Elmer Spectrum RX FT spectropho-
tometer. NMR spectra at 200.1 or 400.1 MHz (1H),
162.0 MHz (31P) or 128.4 MHz (11B) were recorded on
Bruker AC200 or DPX400 spectrometers from CDCl3
solutions at 291 K. Elemental analyses were determined
by the departmental service. The starting materials
[HNEt3][5-I-7,8-Ph2-7,8-nido-C2B9H9] [11], NiCl2(dppe)
[17] and cis-PtCl2(PMe2Ph)2 [18] were prepared by liter-
ature methods or slight variants thereof. All other re-
agents and solvents were supplied commercially and
were used as received.
Ph
I
I
I
1
M
M
Ph
C
I
C
Ph
2, 3
M
M
Ph
C
4.1.1. 1,2-Ph2-4,4-dppe-12-I-4,1,2-closo-NiC2B9H8 (1)
and 1,8-Ph2-2,2-dppe-10-I-2,1,8-closo-NiC2B9H8 (2)
[HNEt3][5-I-7,8-Ph2-7,8-nido-C2B9H9] (0.200 g, 0.42
mmol) was suspended in 30 mL Et2O and cooled to
0 ꢁC. To this was added, dropwise, n-BuLi (0.33 mL
of 2.5 M solution, 0.85 mmol). The reaction mixture
was heated to reflux for 3 h then evacuated to dryness.
The residue was dissolved in THF (30 mL) and cooled
to ꢀ196 ꢁC. To this frozen solid was added NiCl2(dppe)
(0.220 g, 0.48 mmol) in THF (30 mL). After refreezing
the mixture was allowed to warm slowly to room tem-
perature with stirring, then stirred overnight. The
brown, cloudy, solution was filtered through Celiteꢂ
and purified by TLC eluting with CH2Cl2/40–60 petro-
leum ether (1:1) to yield compounds 1 (Rf = 0.41) and
2 (Rf = 0.50) as red and purple bands, respectively.
1: yield 0.037 g, 10%. Anal. Calc. for C40H42B9IP2Ni:
C, 55.4; H, 4.88. Found: C, 56.1; H, 5.58%. IR: mmax
I
C
Ph
4
Scheme 2. Alternate single triangle face rotation mechanisms to
explain the formation of metallacarboranes 1–4.
Although two of the five species reported (the structur-
ally analogous compounds 2 and 3) can be explained
by the sequential dsd isomerisation mechanism (reason-
ably assuming the initial reaction product has a 1,2-Ph2-
3,3-L2-9-I-3,1,2-closo-MC2B9H8 architecture) the others
cannot. We have noted that four of the five products can
be rationalised by single tfr mechanisms which, more-
over, are the same tfrs we have previously used to ex-
plain isomerised metallacarboranes formed from [3-X-
7,8-Ph2-7,8-nido-C2B9H8]2ꢀ (X = Et, F). The final prod-
uct cannot be understood by a single tfr but is explained
by two sequential tfrs.
Studies designed to understand the precise mecha-
nism(s) by which heteroboranes isomerise will continue
to attract attention. Although we are encouraged that,
with the exception of the final compound reported here-
in, some degree of consistency may be emerging from
our work with differently labelled crowded carborane li-
gands, there is clearly much more work that needs to be
done before a clear picture emerges.
1
2583 (BH) cmꢀ1. NMR: H, d 8.0–6.8 (m, 30H, C6H5)
and 2.4–2.1 (m, 4H, CH2). 11B–{1H}, d 11.14 (1B),
ꢀ4.73 (1B), ꢀ7.30 (1B, B12), ꢀ12.26 (4B) and ꢀ16.22
(2B): 31P–{1H}, d 59.34 (br. s, 1P) and 54.66 (br. s, 1P).
2: yield 0.055 g, 15%. Anal. Calc. for C40H42B9IP2Ni:
C, 55.4; H, 4.88. Found: C, 55.1; H, 5.11%. IR: mmax
1
2579 (BH) cmꢀ1. NMR: H, d 7.8–6.7 (m, 30H, C6H5)
and 2.4–1.9 (m, 4H, CH2). 11B–{1H}, d ꢀ1.81 (1B),
ꢀ5.52 (3B), ꢀ9.91 (1B), ꢀ12.97 (1B), ꢀ16.40 (2B) and
ꢀ27.24 (1B, B10): 31P–{1H}, d 58.44 (d, {2JPP = 32
Hz}, 1P) and 56.98 (d, {2JPP = 32 Hz}, 1P).
4. Experimental
4.1.2. 1,8-Ph2-2,2-(PMe2Ph)2-10-I-2,1,8-closo-PtC2B9H8
(3), 1,8-Ph2-2,2-(PMe2Ph)2-12-I-2,1,8-closo-PtC2B9H8
(4), and 1,8-Ph2-2,2-(PMe2Ph)2-7-I-2,1,8-closo-
PtC2B9H8 (5)
4.1. Synthesis – general
Experiments were performed under dry, oxygen-free
N2 using standard Schlenk techniques, with some subse-
Similarly, [HNEt3][5-I-7,8-Ph2-7,8-nido-C2B9H9]
(0.155 g, 0.30 mmol) in 30 mL Et2O was deprotonated

1492
S. Robertson et al. / Inorganica Chimica Acta 358 (2005) 1485–1493
1
by n-BuLi (0.24 mL of 2.5 M solution, 0.60 mmol) and
then treated with cis-PtCl2(PMe2Ph)2 (0.163 g, 0.30
mmol) in THF (25 mL). After stirring overnight, the
resultant canary yellow solution was filtered through
Celiteꢂ and purified by TLC eluting with CH2Cl2/40–
60 petroleum ether (3:2) to yield products 3 + 4 and 5
as yellow bands, the first of which could not be resolved
in spite of exhaustive further chromatography. Com-
pounds 3 and 4 were eventually separated on crystallisa-
tion, diffusion of a CH2Cl2 solution and 40–60
petroleum ether giving first 3 then 4.
2577 (BH) cmꢀ1. NMR: H, d 7.6–6.9 (m, 20H, C6H5),
1.7–1.5 (m, 6H, CH3) and 1.2–1.0 (m, 6H, CH3). 11B–
{1H}, d ꢀ1.32 (1B), ꢀ5.11 (1B), ꢀ9.32 (1B), ꢀ12.70
(2B), ꢀ14.99 (1B), ꢀ16.86 (2B) and ꢀ21.91 (1B): 31P–
{1H}, d ꢀ16.57 (d, {2JPP = 25, JPtP = 3360 Hz}, 1P)
1
and ꢀ18.21 (d, {2JPP = 25, JPtP = 3263 Hz}, 1P).
1
4.2. Crystallography
Single crystals suitable for X-ray diffraction of all
compounds were grown from diffusion of a CH2Cl2
solution and 40–60 petroleum ether at ꢀ30 ꢁC (crystal-
lisation was used to separate 3 and 4, 3 being somewhat
less soluble). Compound 1 crystallises with 1 and com-
pound 2 with 214 equivalents of CH2Cl2. Data from 1,
2 and 5 were collected on a Bruker AXS P4 diffractom-
eter using x scans and were corrected for absorption by
w scans. Data from 3 and 4 were collected on Bruker
AXS Smart APEX CCD-equipped diffractometers, and
were corrected for absorption semi-empirically from
symmetry-equivalent and repeated reflections. All
diffractometers employed graphite-monochromated
3: yield 0.053 g, 17%. Anal. Calc. for C30H40B9IP2Pt:
C, 40.9; H, 4.57. Found: C, 41.0; H, 4.64%. IR: mmax
1
2575 (BH) cmꢀ1. NMR: H, d 7.5–6.9 (m, 20H, C6H5)
and 1.6–1.4 (m, 12H, CH3). 11B–{1H}, d ꢀ2.07 (2B),
ꢀ4.88 (2B), ꢀ10.88 (2B), ꢀ11.67 (1B), ꢀ16.75 (1B)
and ꢀ29.15 (1B, B10): 31P–{1H}, d ꢀ13.60 (co-incident
1
pair of d, {2JPP = 27, JPtP = 3243 and 3300 Hz}).
4: yield 0.013 g, 8%. Anal. Calc. for C30H40B9IP2Pt:
C, 40.9; H, 4.57. Found: C, 40.7; H, 4.57%. IR: mmax
1
2522 (BH) cmꢀ1. NMR: H, d 7.6–6.9 (m, 20H, C6H5)
and 1.6–1.3 (m, 12H, CH3). 11B–{1H}, d ꢀ1.63 (2B),
ꢀ6.62 (2B), ꢀ9.83 (2B), ꢀ12.15 (1B), ꢀ15.73 (1B) and
ꢀ18.77 (1B): 31P–{1H}, d ꢀ14.99 (d, {2JPP = 27,
1JPtP = 3263 Hz}, 1P) and ꢀ16.07 (d, {2JPP = 27,
1JPtP = 3280 Hz}, 1P).
˚
Mo Ka radiation (k = 0.71069 A) and cooled the sample
with cold N2 gas. Structures were solved by direct
and difference Fourier methods and refined [19] by
full-matrix least-squares against F2 The structure of 4
was successfully refined as a racemic twin but required
isotropic restraints on all B atoms. Refinement was com-
5: yield 0.008 g, 3%. Anal. Calc. for C30H40B9IP2Pt:
C, 40.9; H, 4.57. Found: C, 40.5; H, 4.54%. IR: mmax
Table 4
Crystallographic data for compounds 1–5a
Compound
1
2
3
4
5
Formula
C
952.50
40H42B9INiP2 Æ CH2Cl2 C40H42B9INiP2 ꢁ 214CH2Cl2 C30H40B9IP2Pt
C30H40B9IP2Pt
881.84
C30H40B9IP2Pt
881.84
Molecular weight
Crystal size (mm3)
T (K)
1058.66
0.64. · 0.32 · 0.08
160(2)
881.84
0.26 · 0.20 · 0.04
160(2)
0.28 · 0.15 · 0.05 0.60 · 0.03 · 0.03 0.60 · 0.29 · 0.25
160(2)
100(2)
monoclinic
P21
160(2)
monoclinic
P21/c
Crystal system
Space group
monoclinic
P21
monoclinic
P21/c
orthorhombic
P212121
9.308(5)
˚
a (A)
11.146(3)
18.853(4)
11.166(3)
112.649(14)
2165.4(8)
2
9.710(2)
23.877(3)
21.245(3)
90.034(16)
4925.6(15)
4
10.5389(19)
8.7561(16)
18.228(3)
91.879(3)
1681.2(5)
2
18.095(4)
12.313(3)
16.551(3)
109.08(2)
3483.8(14)
4
˚
b (A)
16.303(5)
22.047(5)
˚
c (A)
b (ꢁ)
3
˚
V (A )
3346(2)
4
1.751
Z
Dcalc (Mg mꢀ3
)
1.461
1.389
1.428
1.360
1.742
5.208
1.681
5.025
l (Mo Ka) (mmꢀ1
)
5.234
h Range for collected data (ꢁ) 1.99–25.03
1.92–25.01
9725
8540
1.55–28.81
29259
8082
1.93–28.27
12913
7345
2.04–24.99
6435
5346
Data measured
Unique data, n
Rint
4856
4173
0.0522
0.1152
0.1878
1.070
0.12(5)
505
0.0567
0.1197
0.0687
0.0610
0.1173
1.131
0.0441
0.0474
0.1084
1.024
0.0605
0.0483
0.1098
1.029
R (all data)
wR2 (all data)
S (all data)
Flack parameter, x
0.2418
1.032
0.032(9)
388
2.723
0.570(7)
389
3.915
Variables,ꢀp3
544
392
˚
Emax (e Aꢀ3
)
1.506
ꢀ2.308
1.646
ꢀ1.472
2 1=2
1.664
ꢀ1.654
are variables and
˚
Emin (e A
)
ꢀ2.447
ꢀ2.291
P
P
P
P
2
a
R = iFo| ꢀ |Fci/ |Fo|, wR2 ¼ ½ ½wðF 2o ꢀ F c2Þ ꢂ= wðF o2Þ ꢂ
,
(wꢀ1 = [r2(Fo)2 + (aP)2 + bP] where
a
and
b
P
2
1=2
P = [0.333(Fo)2 + 0.667(Fc)2]), S ¼ ½ ½wðF o2 ꢀ F 2c Þ ꢂ=ðn ꢀ pÞꢂ (where n is the number of data and p the number of parameters).

S. Robertson et al. / Inorganica Chimica Acta 358 (2005) 1485–1493
1493
pleted with all non-hydrogen atoms assigned anisotropic
References
displacement parameters, except that B11 of compound
2 required the application of isotropic restraints. H
atoms were allowed to ride on their bound atom with
[1] A.J. Welch, in: P. Braunstein, L.A. Oro, P.R. Raithby (Eds.),
Metal Clusters in Chemistry, Wiley/VCH, New York/Weinheim,
1999, p. 69.
˚
B–H set at 1.12 A and C–H = 0.95 (phenyl), 0.98
[2] D.R. Baghurst, R.C.B. Copley, H. Fleischer, D.M.P. Mingos,
G.O. Kyd, L.J. Yellowlees, A.J. Welch, T.R. Spalding, D.
OꢀConnell, J. Organomet. Chem. 447 (1993) C14.
[3] R.M. Garrioch, P. Kuballa, K.S. Low, G.M. Rosair, A.J. Welch,
J. Organomet. Chem. 575 (1999) 57.
˚
(CH3) or 0.99 (CH2) A. All H atoms except methyl H
atoms were assigned displacement parameters set at
1.2 times that of the Ueq of their bound atom. Methyl
H atoms were assigned displacement parameters calcu-
lated as 1.5 times the bound carbon atom Ueq. No H
atoms were included in the 14 molecule of CH2Cl2 of sol-
vation in 2. Table 4 contains details of the crystallo-
graphic studies.
[4] S. Dunn, G.M. Rosair, Rh.Ll. Thomas, A.S. Weller, A.J. Welch,
Angew. Chem. Int. Ed. 36 (1997) 645.
[5] S. Dunn, G.M. Rosair, A.S. Weller, A.J. Welch, Chem. Commun.
(1998) 1065.
[6] S. Dunn, R.M. Garrioch, G.M. Rosair, L. Smith, A.J. Welch,
Collect. Czech. Chem. Commun. 64 (1999) 1013.
[7] D.J. Wales, J. Am. Chem. Soc. 115 (1993) 1557.
[8] S. Robertson, D. Ellis, G.M. Rosair, A.J. Welch, Appl. Organo-
met. Chem. 17 (2003) 516.
5. Supplementary material
[9] S. Robertson, D. Ellis, G.M. Rosair, A.J. Welch, J. Organomet.
Chem. 680 (2003) 286.
Crystallographic data for the structure analyses have
been deposited with the Cambridge Crystallographic
Data Centre, CCDC Nos. 237598–237702 for com-
pounds 1–5, respectively. Copies of this information
may be obtained free of charge from The Director,
CCDC, 12 Union Road, Cambridge, CB2 1EZ, UK
(fax: +44 1223 336 033; e-mail deposit@ccdc.cam.ac.uk
or www.ccdc.cam.ac.uk).
[10] T.D. McGrath, M.A. Fox, A.J. Welch, Acta Crystallogr., Sect. C
56 (2000) 487.
[11] S. Robertson, D. Ellis, T.D. McGrath, G.M. Rosair, A.J. Welch,
Polyhedron 22 (2003) 1293.
[12] U. Venkatasubramanian, D.J. Donohoe, D. Ellis, B.T. Giles,
S.A. Macgregor, S. Robertson, G.M. Rosair, A.J. Welch, A.S.
Batsanov, L.A. Boyd, R.C.B. Copley, M.A. Fox, J.A.K. Howard,
K. Wade, Polyhedron 23 (2004) 629.
[13] D.M.P. Mingos, M.I. Forsyth, A.J. Welch, J. Chem. Soc., Dalton
Trans. (1978) 1363.
Acknowledgements
[14] J. Cowie, B.D. Reid, J.M.S. Watmough, A.J. Welch, J. Organo-
met. Chem. 481 (1994) 283.
We thank the Leverhulme Trust (R.M.G., D.E.), the
EPSRC (B.E.H.) and Heriot-Watt University (S.R.,
T.D.M.) for support, G. Evans and C. Graham for micro-
analysis and Dr. A.S.F. Boyd for NMR spectra. We are
very grateful to Dr. L. Ha¨ming (Bruker AXS, Karlsruhe)
for collecting the diffraction data from compound 4.
[15] G.O. Kyd, Ph.D. Thesis, University of Edinburgh, 1993.
[16] D.M.P. Mingos, M.I. Forsyth, A.J. Welch, J. Chem. Soc., Chem.
Commun. (1977) 602.
[17] G. Booth, J. Chatt, J. Chem. Soc. (1965) 3238.
[18] J.M. Jenkins, B.L. Shaw, J. Chem. Soc. A (1966) 770.
[19] G.M. Sheldrick, SHELXTL
Madison, WI, 1999.
, version 5.1, Bruker AXS, Inc.,