Polyhedral iridaborane chemistry: Elements of the 10-vertex closo-isonido-isocloso continuum.

Article abstract of DOI:10.1016/j.ica.2006.01.038

There is experimental difficulty in the isolation and structural definition of examples of compounds of the isonido 10-vertex cluster structure in order that they may be adduced to the pattern of behaviour associated with the 10-vertex closo-isonido-isocloso structural continuum in boron-containing cluster chemistry. By contrast, the closo and isocloso extremes are well recognised, with several definitive examples. An approach involving the synergic application of crystallography (using single-crystal X-ray diffraction analysis and HYDEX), multi-element NMR spectroscopy, and DFT calculations of structure and nuclear magnetic shielding, has been used to delineate and define two basic variants in the quadrilaterally open-faced 10-vertex 'isonido' {IrB₉} metallaborane system. Type A variants, as in [7,7,7-(PPh₃)₂H-μ-3,7-H-isonido-7-IrB₉ H₈-9-(PPh₃)] (2), have an Ir-H-B bridging hydrogen atom involving a boron atom off the open face, and an Ir-H non-bridging linkage. Type B variants, as in [8-Cl-7-(PPh₃)-μ-7^P,10^C-(Ph₂ P-ortho-C₆H₄)-isonido-7-IrB₉H₆ -9-(PPh₃)] (5), have no cluster-bridging hydrogen atoms and no Ir-H unit. Previously unreported single-crystal X-ray diffraction analyses are given for isonido type-A [7-(PPh₃)-μ-7^P,10^C-(Ph₂P- ortho-C₆H₄)-7-H-μ-3,7-H-isonido-7-IrB₉ H₈-9-(PPh₃)] (10), classical closo [7-(PPh₃)-μ-7^P,10^C-(Ph₂P- ortho-C₆H₄)-7-H-closo-7-IrB₉H₆ -3,9-(PPh₃)₂] (13) and classical nido [6-(PPh₃)-μ-6^P,5^C-(Ph₂P-o rtho-C₆H₄)-6-H-nido-6-IrB₉H₁₂ ] (15). Overall, the work well demonstrates the synergic usage of NMR spectroscopy, single-crystal X-ray crystallography (including HYDEX calculations) and DFT calculations of structure and nuclear shielding, to resolve ambiguous problems of chemical constitution in polyhedral boron-containing cluster chemistry, an approach that should have wider applicability.

Full text of DOI:10.1016/j.ica.2006.01.038

Inorganica Chimica Acta 359 (2006) 3723–3735
www.elsevier.com/locate/ica
Polyhedral iridaborane chemistry: Elements of the 10-vertex
closo–isonido–isocloso continuum
Molecular structures of [(PPh3)2HIrB9H9(PPh3)],
[(PPh3)(Ph2PC6H4)IrB9H7(PPh3)],
[(PPh3)(Ph2PC6H4)HIrB9H6Cl(PPh3)],
[(PPh3)(Ph2PC6H4)HIrB9H6(PPh3)2] and [(PPh3)(Ph2PC6H4)HIrB9H12]q
a
a,
*
Jonathan Bould , William Clegg b,c, John D. Kennedy
a
School of Chemistry, University of Leeds, Leeds LS2 9JT, Northern England, UK
School of Natural Sciences (Chemistry), University of Newcastle, Newcastle upon Tyne NE1 7RU, Northern England, UK
b
c
Daresbury Laboratory, Daresbury, Warrington, Cheshire WA4 4AD, Northern England, UK
Received 25 November 2005; received in revised form 30 January 2006; accepted 30 January 2006
Available online 20 March 2006
It is a pleasure to be able to contribute to this special edition for Mike Mingos in order to recognise his many stimulating experimental and theoretical
contributions to many aspects of chemical science.
Abstract
There is experimental difficulty in the isolation and structural definition of examples of compounds of the isonido 10-vertex cluster
structure in order that they may be adduced to the pattern of behaviour associated with the 10-vertex closo–isonido–isocloso structural
continuum in boron-containing cluster chemistry. By contrast, the closo and isocloso extremes are well recognised, with several definitive
examples. An approach involving the synergic application of crystallography (using single-crystal X-ray diffraction analysis and
HYDEX), multi-element NMR spectroscopy, and DFT calculations of structure and nuclear magnetic shielding, has been used to delin-
eate and define two basic variants in the quadrilaterally open-faced 10-vertex ‘isonido’ {IrB9} metallaborane system. Type A variants, as
in [7,7,7-(PPh3)2H-l-3,7-H-isonido-7-IrB9H8-9-(PPh3)] (2), have an Ir–H–B bridging hydrogen atom involving a boron atom off the open
face, and an Ir–H non-bridging linkage. Type B variants, as in [8-Cl-7-(PPh3)-l-7P,10C-(Ph2P-ortho-C6H4)-isonido-7-IrB9H6-9-(PPh3)]
(5), have no cluster-bridging hydrogen atoms and no Ir–H unit. Previously unreported single-crystal X-ray diffraction analyses are given
for isonido type-A [7-(PPh3)-l-7P,10C-(Ph2P-ortho-C6H4)-7-H-l-3,7-H-isonido-7-IrB9H8-9-(PPh3)] (10), classical closo [7-(PPh3)-l-7P,10C-
(Ph2P-ortho-C6H4)-7-H-closo-7-IrB9H6-3,9-(PPh3)2] (13) and classical nido [6-(PPh3)-l-6P,5C-(Ph2P-ortho-C6H4)-6-H-nido-6-IrB9H12]
(15). Overall, the work well demonstrates the synergic usage of NMR spectroscopy, single-crystal X-ray crystallography (including
HYDEX calculations) and DFT calculations of structure and nuclear shielding, to resolve ambiguous problems of chemical constitution
in polyhedral boron-containing cluster chemistry, an approach that should have wider applicability.
ꢀ 2006 Elsevier B.V. All rights reserved.
Keywords: Borane cluster; X-ray structure; Metallaboranes; DFT calculations; 11B nuclear shielding; 10-vertex isonido cluster types; Polyhedral structural
continuum
q
This article was freely submitted for publication without royalty. By acceptance of this paper, the publisher and/or recipient acknowledges the right of
the authors to retain non-exclusive, royalty-free license in and to any copyright covering this paper, along with the right to reproduce all or part of the
copyrighted paper.
*
Corresponding author.
E-mail address: johnk@chem.leeds.ac.uk (J.D. Kennedy).
0020-1693/$ - see front matter ꢀ 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.ica.2006.01.038

3724
J. Bould et al. / Inorganica Chimica Acta 359 (2006) 3723–3735
1. Introduction
adigm [1–8], and in that they exhibit a cluster opening from
pure deltahedral closo that may be regarded as anomalous
In the early 1970s Mike Mingos and others developed
rules that governed the relationships between the shapes
and sizes of clusters of atoms and the numbers of electrons
associated with the intracluster bonding [1–8]. A key clus-
ter shape was the 12-vertex icosahedron – as represented
by the regular icosahedral {B12} framework of the [closo-
B12H12]2À anion (schematic I) and by the shapes of its clo-
sely related congeners, such as neutral [closo-1,2-C2B10H12]
(schematic II), of which the exact geometry deviates from
regular icosahedral principally because of the smaller sizes
of the carbon versus the boron atoms. Transition-element
centres that can be regarded as making three-orbital
two-electron contributions to cluster bonding that mimic
the contribution of a {BH} cluster vertex, such as {Co
(g5-C5H5)} and {Ru(g6-C6H6)}, can also figure as cluster
components [1–6], and so species such as [(g5-C5H5)
CoC2B9H11] and [(g6-C6H6)RuC2B9H11] also exhibit
triangular-faced icosahedral cluster geometries (schematic
cluster structure III) [9–11]. These deviate from the regular
icosahedral of [B12H12]2À because of the larger sizes of the
metal atoms as well as the smaller sizes of the carbon
atoms. It was however known since the late 1960s that
the incorporation of later-transition-element centres can
produce structures that are more open than that of a trian-
gular-faced icosahedron [12–18]. These deviate from icosa-
hedral in that the metal-to-carbon distances now are much
longer, and can be regarded as being beyond a range asso-
ciated with strong intracluster bonding interaction (sche-
matic cluster structure IV). Such ‘slipped closo’ species
have received detailed analysis from Mike Mingos and
co-workers [16,17]. In simplistic terms the ‘slippage’ of
cluster opening arises because these types of metal centre
have a tendency to bind to the remainder of the cluster
principally by the use of two orbitals rather than the three
utilized by units such as {BH}, {CH}, {Co(g5-C5H5)},
{Ru(g6-C6H6)} etc., and the metal-to-cluster bonding con-
nectivity is correspondingly diminished.
[20].
A different, perhaps more subtle, type of skeletal dis-
obedience has over the years become increasingly manifest
[20]. It is apparent that some deltahedral structural motifs
are quite flexible. In such clusters, changes of cluster con-
stituents and substituents can often change one or more
connectivity distances to such an extent that more open
geometries result, often with a quadrilateral open face if
only one connectivity is stretched. Alternatively, some
clusters may instead be regarded as ‘fragile’ rather than
‘flexible’, in that changes of cluster constituents and sub-
stituents can produce catastrophic changes to give another
distinct geometric and electronic cluster type. Sometimes
the differences between ‘cluster fragility’ and ‘cluster flex-
ibility’ may be difficult to distinguish experimentally. The
origins of these sorts of distortion are not yet as clear as
the origins of the types of distortion induced by transi-
tion-element centres in the ‘slipped closo’ cases mentioned
above. Examples here include the nine-vertex closo system
based on the classical tricapped trigonal prismatic motif
(schematic cluster structure VII). The basic closo
[B9H9]2À cluster structure is characterised by three some-
what longer (and thence probably weaker) interboron
linkages between the three boron atoms constituting each
of the two triangular ends of the underlying trigonal
prism (hatched lines in schematic cluster structure VIII)
[21]. In a number of species, such as, for example,
[(CO)(PPh3)2IrCB7H8] (schematic cluster structure IX)
[22,23], the influence of the heteroatoms is to stretch
one or more of these longer connectivities towards a sep-
aration that can be regarded as non-bonding, giving a
quadrilateral open face in a geometry that has been
dubbed ‘isonido’.
Similar ‘slipped closo’ distortions occur in smaller clus-
ters, and in particular in this context studies have been
made on seven-vertex {PtC2B4} systems [18,19], in which
again there is slippage away from conventional closo (sche-
matic cluster structure V) with longer metal-to carbon
interatomic distances (schematic cluster structure VI).
The cluster openings observed in these 12-vertex and
seven-vertex species are examples of ‘skeletal disobedience’,
in that the geometric structures in question do not conform
to the straightforward closo/nido/arachno geometrical par-
We have long been interested in this last type of stretch-
ing phenomenon in the 10-vertex system [24], particularly
the stretching of the closed 10-vertex D4d bicapped square
antiprismatic geometry X of ostensibly closo {IrB9} species
to give so-called ‘isonido’ 10-vertex geometries XI. A geo-
metrical continuation of this distortion results in a dia-
mond-square-diamond process and the generation of a
so-called ‘isocloso’ architecture as in schematic XII.

J. Bould et al. / Inorganica Chimica Acta 359 (2006) 3723–3735
3725
associated with those environments (schematic cluster
structure XIII). The incidence of these two hydrogen atoms
contrasted to the constitution of [8-Cl-7-(PPh3)-l-7P,10C-
(Ph2P-ortho-C6H4)-isonido-7-IrB9H6-9-(PPh3)] (compound
5), which does not have two such hydrogen atoms (sche-
matic cluster structure XIV). Again, as with compound 2
above, hydrogen atoms were not locatable from the sin-
gle-crystal X-ray diffraction work on what we then thought
to be a crystal of compound 5, but the crystallographically
determined heavier-atom structure did show isonido
geometry, with its characteristic quadrilateral open face
(schematic XI above), and NMR spectroscopy revealed
exo-terminal hydrogen atoms on the six boron atoms that
were not otherwise substituted. The detailed dimensions of
the basic {IrB9} skeletal structures of 2 and 5 were appar-
ently very similar indeed. Now, however, there was no
NMR evidence for any other cluster hydrogen atoms,
bridging or otherwise, in compound 5. The possibility of
any fluxionality among any such hydrogen atoms, resulting
in the broadening of their resonances, and thence making
them difficult to detect at particular temperatures, was dis-
counted by variable-temperature NMR work.
A comprehensive investigation of these latter 10-vertex
systems is severely inhibited by the dearth of intermediate
‘isonido’ examples and the experimental difficulty of gener-
ating such suitable examples. In the original work on the
{IrB9} 10-vertex system, there seemed to be two classes
of species within the intermediate isonido area XI of the
structural continuum [24]. Both had quadrilateral open
faces. One type, variant A, appeared to have two bridging
hydrogen atoms associated with the open face, although
there was some uncertainty about their precise positioning
(schematic cluster structure XIII below). The other type,
variant B, had none (schematic cluster structure XIV
below). In many cases, however, and although it was rea-
sonably certain that the crystallographically determined
structures matched solution-phase NMR measurements,
because of the low yields, possibilities of lack of robustness
in solution, and possibilities of isomerization upon crystal-
lisation, allied with non-ideal crystallographic data sets
that precluded hydrogen-atom location, some doubt
remained. We now present results from further crystallo-
graphic structural work, allied with DFT calculations of
structure and 11B nuclear shielding, which clarify the char-
acteristics of these two 10-vertex isonido variants A and B.
The results presented well exemplify the efficacy of the syn-
ergic use of crystallography, NMR spectroscopy and DFT
calculations of structure and nuclear shielding to elucidate
problems of structure and constitution in polyhedral
boron-containing cluster chemistry, and will have general
applicability.
In view of this difference of two in hydrogen-atom content
between compounds 2 and 5 (structural variants A and B,
respectively), it was of interest that the two compounds
appeared to have such very similar basic cluster geometries.
We have therefore now used DFT calculations of structure,
and of magnetic shielding of nuclei, to establish and confirm
the precise nature of the differences, and thence also to clear
previous uncertainties about the constitutions of [(PPh3)2-
HIrB9H10(PPh3)] (2) and [(PPh3)(Ph2PC6H4)IrB9H6(PPh3)]
(5), and thereby better define the Type A and Type B charac-
teristics. DFT calculations at the B3LYP level of theory were
employed, using the 6-31G* basis-sets for phosphorus, car-
bon, boron and hydrogen, and the LANL2DZ basis-set for
iridium. The overall analysis is assisted by an X-ray crystal-
lographic data collection on a new species, a Type-A variant,
[7-(PPh3)-l-7P,10C-(Ph2P-ortho-C6H4)-7-H-l-3,7-H-isonido-
7-IrB9H8-9-(PPh3)] (10), in which, now, all cluster hydrogen
atoms are located (see Fig. 1, left).
2. Results and discussion
Among the 10-vertex {IrB9} isonido-structured cluster
compounds just mentioned in Section 1 was one for which
we proposed the constitution [7,7,7-(PPh3)2-l-7,10-H-l-
8,9-H-isonido-7-IrB9H8-9-(PPh3)] (compound 2) (note that,
for convenience of comparison, in this present paper we use
the compound numbering of our previous report [24] for
compounds mentioned in both works; compounds newly
mentioned in this present work are numbered from 10
onwards). Compound 2 was of variant Type A. Single-
crystal X-ray diffraction data were sufficient to give the
heavier atom skeletal structure, but were of insufficient
quality for hydrogen-atom location. However, NMR spec-
troscopy showed seven {BH(exo)} proton resonances, and
also two proton resonances for two hydrogen atoms at
higher shielding, d(1H) À4.37 and À8.70 ppm. We pro-
posed that these latter two be associated with the open
face, one lower field in a BHB bridge, and the one at higher
shielding in an IrHB bridge, as they were in typical ranges
To reduce calculation time, [(PPh3)2IrB9H10(PPh3)]
(compound 2) was modelled with PMe3 ligands rather than
PPh3, viz., as [(PMe3)2IrB9H10(PMe3)] (compound 2a). As
with the other DFT structural calculations reported in this
paper, compound 2 was modelled initially using hydrogen

3726
J. Bould et al. / Inorganica Chimica Acta 359 (2006) 3723–3735
atoms rather than phenyl groups on the phosphine ligands
(species 2b), and then, based on the geometries so obtained,
the structural and boron nuclear shielding calculations
were extended to the final P-methyl model 2a. Initial calcu-
lations on 2a did not minimise in accord with our original
supposition of two bridging hydrogen atoms associated
with a quadrilateral open face (schematic cluster structure
XIII). However, a stable geometry was achieved with a
placement of one hydrogen atom in an Ir–H non-bridging
linkage held in a quasi-endo configuration above the open
face and the other hydrogen atom in an IrHB bridging
position, but involving a boron atom B(3) off the
{Ir(7)B(8)B(9)B(10)} open face (Fig. 1, right, and sche-
matic cluster structure XV). General non-organyl dimen-
sions for 2a were very similar to those found
experimentally for 2 (Table 1), indicating that the use of
the P-methyl variant 2a as a model for 2 was satisfactory.
An off-open-face bridging hydrogen atom is very rare in
boron-containing cluster chemistry, particularly in open-
faced species, for which the general incidence for a bridging
Fig. 1. ORTEP-3 drawings [60] of (left-hand diagram) the crystallographically determined molecular structure of [7,7-(PPh3)2-7-H-l-3,7-H-isonido-7-
IrB9H8-9-(PPh3)] (compound 2) and (right-hand diagram) the DFT-calculated molecular structure of its methyl-for-phenyl analogue [7,7-(PMe3)2-7-H-l-
3,7-H-isonido-7-IrB9H8-9-(PMe3)] (2a), used to model compound 2 in DFT calculations. For clarity methyl hydrogen atoms are omitted. Selected
interatomic dimensions are given in Table 1.
Table 1
Selected measured interatomic separations within compounds 2, 10, and 13, together with calculated interatomic separations within the P-methyl models
2a, 5a, 10a and 13a, and within the P-hydryl models (see text) 2b, 12a, 12b, 10b and 13ba
Compounds
2
2a [2b]
10
10a [10b]
{5a}
12b [12c]
13
13a [13b]
Ir(7)–B(2)
Ir(7)–B(3)
Ir(7)–B(4)
Ir(7)–B(8)
Ir(7)–B(10)
Ir(7)–B(9)
B(8)–B(10)
B(8)–B(9)
B(9)–B(10)
B(5)–B(9)
B(6)–B(9)
B(4)–B(8)
B(2)–B(10)
B(1)–B(3)
a
2.359(12)
2.193(11)
2.343(12)
2.320(11)
2.304(11)
3.319
2.158(13)
1.646(16)
1.684(14)
1.678(16)
1.709(15)
1.766(16)
1.775(16)
1.658(17)
2.369 [2.360]
2.255 [2.260]
2.386 [2.360]
2.330 [2.318]
2.341 [2.318]
3.320 [3.268]
2.223 [2.271]
1.693 [1.692]
1.689 [1.692]
1.694 [1.694]
1.697 [1.694]
1.783 [1.780]
1.779 [1.780]
1.694 [1.689]
2.338(3)
2.161(4)
2.363(4)
2.306(4)
2.275(4)
3.265
2.382 [2.367]
2.236 [2.244]
2.378 [2.358]
2.339 [2.334]
2.311 [2.297]
3.240 [3.199]
2.354 [2.384]
1.706 [1.702]
1.710 [1.706]
1.711 [1.706]
1.708 [1.703]
1.763 [1.761]
1.778 [1.780]
1.705 [1.698]
{2.377}
{2.074}
{2.360}
{2.196}
{2.302}
{3.125}
{2.392}
{1.706}
{1.712}
{1.736}
{1.723}
{1.759}
{1.747}
{1.712}
2.374 [2.360]
2.069 [2.073]
2.369 [2.365]
2.239 [2.239]
2.291 [2.277]
3.104 [3.048]
2.442 [2.495]
1.710 [1.708]
1.710 [1.706]
1.729 [1.727]
1.727 [1.726]
1.758 [1.755]
1.745 [1.745]
1.712 [1.712]
2.305(4)
2.099(4)
2.324(4)
2.286(4)
2.307(5)
3.429
1.992(6)
1.712(6)
1.706(6)
1.702(6)
1.701(6)
1.837(6)
1.850(6)
1.689(6)
2.341 [2.371]
2.099 [2.081]
2.379 [2.359]
2.330 [2.334]
2.336 [2.325]
3.414 [3.374]
2.069 [2.114]
1.700 [1.687]
1.690 [1.696]
1.684 [1.679]
1.685 [1.684]
1.810 [1.835]
1.825 [1.821]
1.688 [1.678]
2.266
1.704(5)
1.714(5)
1.709(5)
1.707(5)
1.770(5)
1.775(5)
1.707(5)
In the results of all the calculations it is noted that all the distances from iridium to boron are somewhat longer than those observed experimentally. We
have found this to be a general feature in DFT calculations of the structure of metallaboranes of second and third-row transition elements, using the same
basis-sets as those used here, for which we have reported on ruthenaboranes [61] and platinaboranes [62]. Other calculated intercluster connectivities, e.g.,
interboron distances, much more closely match those determined experimentally.

J. Bould et al. / Inorganica Chimica Acta 359 (2006) 3723–3735
3727
position is on the open face [26]; it is most common in com-
pletely closed metallaborane cluster species, particularly
those with high metal-atom content [27–29], but these also
are rare. Because of this novelty, and because we calculated
for the PMe3 species 2a rather than for the experimentally
established PPh3 species 2, and also in order to confirm
correlation of structure with measured NMR properties,
boron nuclear magnetic shieldings within 2a were calcu-
lated by the DFT-GIAO method (Table 2; see also Fig. 4
below). These DFT-GIAO//B3LYP/6-31G*/LANL2DZ
results for 2a showed an agreement with those found exper-
imentally for 2 that was, in view of the methyl-for-phenyl
approximations, and also in view of the presence of the
third-row transition-element iridium, reasonable and
encouraging. Thus, the deviations between the boron
nuclear shieldings calculated for 2a and those measured
experimentally for 2 were all within ca. 8 ppm. In particular
both measured and calculated values concurred in being
diagnostically different to those for non-bridged 5 of iso-
nido type B. The principal deviations between experimental
and calculated shielding values for 2 versus 2a were for the
{B(8),B(10)} and {B(2),B(4)} positions directly bound to
the iridium atom, and for the B(5) and B(6) positions flank-
ing the B(4)B(8) and B(2)B(10) pairs. These variations
between phenyl experimental 2 and methyl calculated 2a
were similar to the variations between methyl calculated
2a and those calculated for the PH3 analogue [7,7,7-
tallographically unlocatable hydrogen atoms are cleared by
the DFT work; in the original work [24,25] all four possible
locations implied by schematics XIII and XV were not suf-
ficiently differentiated from each other in the HYDEX
work. The DFT calculations allied with the 11B shieldings
now confirm XV.
As part of this work we also describe a previously unre-
ported 10-vertex isonido {IrB9} species, compound 10, a
Type-A variant. This was isolated as a purple crystalline
material in about 3% yield from the reaction between
PPh3 and [6,6-(PPh3)2-6-H-nido-6-IrB9H13] (14, of conven-
tional nido 10-vertex structure) in solution in refluxing 1,2-
dichloromethane. Proton NMR spectroscopy on 10 (Table
3) shows resonances at d(1H) ca. À8.5 and ca. À4.5 ppm
(compare À8.7 and À4.4 ppm for compound 2 above). A
crystal and molecular structure was obtained by single-
crystal X-ray diffraction analysis (Fig. 2, left), making use
Table 3
Measureda and calculated d(11B) chemical shifts for compound 10 and its
related P-methyl and P-hydryl models 10a and 10b
Vertex
Compound 10
Calculated d(11B)
Measured 10a
d(11B)
10a
[10b]
{d(1H)}
3
+27.3
[+31.4]
+32.3
{+6.82b}
{+4.04c}
{+4.23}
8, 10
1
9
5, 6
+9.0, +11.0 [+9.1, +12.6] +7.0, +6.0
À3.8
[À2.9]
À0.8
À16.4
[À24.6]
À10.4d
(PH3)2H-l-3,7-H-isonido-7-IrB9H8-9-(PH3)]
2b,
also
À20.1, À24.1 [À19.3, À21.5] À21.7{À0.16}, À24.0{À0.10},
included in Tables 1 and 2 for comparison. These results
demonstrate that the experimental matching of NMR to
structure was reasonable and self-consistent, again indicat-
ing that the use of the P-methyl variant 2a as a model for P-
phenyl 2 was satisfactory, and also confirming the general
structure of the compound. As mentioned, both measured
and calculated boron magnetic shieldings were quite differ-
ent from those found experimentally for compound 5.
Additionally, ambiguities arising from the use of
HYDEX-type programmes [30,31] for the location of crys-
2, 4
À18.6, À23.7 [À17.5, À22.1] À25.8 (2B)
{À0.54, À2.00}
{À4.04}
Ir–H(3,7)
Ir–H(7)
a
{À8.60e}
This work; CDCl3 at 294–300 K; additionally d(31P) +25.6 (sharp),
+17 (broad) and +11.5 (sharp).
b
Strong coupling to lH(3,7) exhibited in the [1H–1H]-COSY spectrum
and in 1H–{11B(selective)} experiments.
c
Site of ortho-phenylene substituent.
Doublet 2J(31P–11B) ca. 145 Hz.
Unresolved overlapping doublet of doublets of doublets, with all three
d
e
nJ(31P–1H) in region ca. 10–25 Hz.
Table 2
Chemical shifts d(11B) as measureda for Compounds 2 and 5 and as calculated for the P-methyl models 2a and 5a and for the P-hydryl models (see text) 2b,
b
12b and 12c
Vertex Compound 2c
Compound 5
Measured for 5 Calculated for 5a Calculated for:
{P-methyl 12c} P-methyl 12a [P-hydryl 12b]
Models 12c, 12a and 12b
Measured 2 Calculated (mean) for:
P-methyl 2a
[P-hydryl 2b]
3
+25.4
À7.0
+22.7
[+26.5]
[À4.6]
[À4.7]
[À11.2]
+59.7
+29.0, +23.6
+9.5
{+68.7}
{+38.3, +26.0}
{+6.5}
{+58.1}
{+27.9}
{+10.6}
{À5.1}
+67.4
[+71.9]
8, 10
1
9
5, 6
2, 4
a
+1.1, +2.5 (+1.8)
À6.0
+33.5, +20.1 [+31.8, +32.9]
+10.8
À8.2
[+11.0]
À11.1
À32.3
À30.0
À12.7
À5.6
{À12.8}
À10.4
[À17.3]
À26.2, À25.0 (À25.6) [À25.5]
À8.2, À16.9
À19.1, À26.8
{À6.8, À17.7}
{À13.3, À24.9}
{À16.2}
À7.6, À5.2
[À6.9, À11.6]
À25.6, À22.8 (À24.2) [À23.5]
{À17.2, À21.6} À24.0, À12.2 [À12.8, À22.2]
Measured data taken from Ref. [24]; CD2Cl2, 294–300 K.
12a is the non-chlorinated analogue of 5a, viz. [7-(PMe3)-l-7P,10C-(Me2P-ortho-C6H4)-isonido-7-IrB9H7-9-(PMe3)]; 12b is the P-hydryl equivalent[7-
b
(PH3)-l-7P,10C-(H2P-ortho-C6H4)-isonido-7-IrB9H7-9-(PH3)], and 12c is the non-ortho-cycloboronated non-chlorinated ‘parent’ P-methyl species [7,7-
(PMe3)2-isonido-7-IrB9H8-9-(PMe3)].
c
For compound 2 the calculated values from 2a do not quite display mirror-symmetry; for 2, the rocking between the enantiomers across the
{Ir(7)B(9)B(1)B(3)} pseudo-mirror plane will have a very low activation energy, resulting in the observations of mean shieldings within each the 11B(8,10),
11B(5,6) and 11B(2,4) pairs.

3728
J. Bould et al. / Inorganica Chimica Acta 359 (2006) 3723–3735
Fig. 2. ORTEP-3 drawings [60] of (left-hand diagram) the crystallographically determined molecular structure of [7-(PPh3)-l-7P,10C-(Ph2P-ortho-C6H4)-
7-H-l-3,7-H-isonido-7-IrB9H8-9-(PPh3)] (10), and (right-hand diagram) the DFT-calculated molecular structure of [7-(PMe3)-l-7P,10C-(Me2P-ortho-
C6H4)-7-H-l-3,7-H-isonido-7-IrB9H8-9-(PMe3)] (10a), used to model compound 10 in DFT calculations. In the crystallographically determined structure,
the B(8) vertex holds a hydrogen atom (with 90% occupancy, compound 10 proper) or a chlorine atom (with a 10% partial occupancy, as illustrated,
compound 11, see text). For clarity methyl hydrogen atoms are omitted. Selected interatomic dimensions are given in Table 1.
of a crystallographic data set obtained using synchrotron-
generated X-radiation. All heavy atoms were located and,
additionally, centres of electron density corresponding to
(a) an IrH endo-terminal hydrogen atom and (b) an IrHB
bridging hydrogen atom to the B(3) position were apparent
in the difference map; these both refined in reasonable posi-
tions. These locations concurred with those found by an
XHYDEX calculation [30,31], as well as by DFT calcula-
tions (see following paragraph). Compound 10 was thence
shown to be [7-(PPh3)-l-7P,10C-(Ph2P-ortho-C6H4)-7-H-l-
3,7-H-isonido-7-IrB9H7-9-(PPh3)]. The incidence of a ca.
10% partial occupancy of chlorine bound exo to the B(8)
position instead of hydrogen also emerged from the crystal-
lographic analysis. This corresponds to a 10% presence in
the crystal of 10 of its isostructural B-chlorinated derivative
[8-Cl-7-(PPh3)-l-7P,10C-(Ph2P-ortho-C6H4)-7-H-l-3,7-H-
isonido-7-IrB9H6-9-(PPh3)] (11). This last compound is in
fact the species in Fig. 3 of Ref. [24], for which hydrogen
Fig. 3. ORTEP-3 drawing [60] of the DFT-calculated molecular structure
atoms could not be located, and which was at the time
of [8-Cl-7-(PMe3)-l-7P,10C-(Me2P-ortho-C6H4)-isonido-7-IrB9H6-9-(PMe3)]
believed to represent the structure of the Type-B com-
(5a), used to model [8-Cl-7-(PPh3)-l-7P,10C-(Ph2P-ortho-C6H4)-isonido-7-
IrB9H6-9-(PPh3)] (5) in the DFT calculations of structure and nuclear
pound [8-Cl-7-(PPh3)-l-7P,10C-(Ph2P-ortho-C6H4)-isonido-
7-IrB9H6-9-(PPh3)] (5).
shielding. For clarity methyl hydrogen atoms are omitted. Selected
interatomic dimensions are given in Table 1.
[(PPh3)(Ph2PC6H4)HIrB9H7(PPh3)] (compound 10) was
modelled by DFT calculation, again using methyl groups
instead of phenyl on the phosphine ligands for speed of cal-
culation, although the rigid ortho-cycloboronated {C6H4}
feature was retained as this may more substantially affect
cluster structure and boron nuclear shielding, particularly
at and around the B(10) site to which it is bound. For this
model species, viz. [7-(PMe3)-l-7P,10C-(Me2P-ortho-C6H4)-
7-H-l-3,7-H-isonido-7-IrB9H7-9-(PMe3)] (compound 10a),
the experimentally observed quadrilateral open face in
the central 10-vertex cluster of 10 (Fig. 2, left; see also sche-
matic cluster XIII above) was thence reproduced in the
energetically minimised structure (Fig. 2, right), and the
general non-organyl dimensions calculated for 10a were
all generally similar to those found experimentally for 10
(Table 1). This correspondence indicates that the use of
10a as a model for 10 was satisfactory. This was further con-
firmed by the results of DFT-GIAO boron magnetic nuclear
shielding calculations on the minimised structure for 10a,
which, as with 2a above, gave a reasonable match with the
values observed experimentally for 10 (Table 3). This agree-
ment showed that the experimental matching of NMR to

J. Bould et al. / Inorganica Chimica Acta 359 (2006) 3723–3735
3729
structure was in fact self-consistent, and gave additional
confirmation of the general structure of compound 10 as
in Fig. 2, as well as further confirmation of the applicability
of the linked calculational/experimental approach.
ferences, with B(5)–B(9) and B(6)–B(9) longer in Type B
than in Type A, whereas the B(4)–B(8) and B(4)–B(10) dis-
tances are shorter. Of particular significance is the Ir(7)–
B(3) distance, shorter in Type B, where it is unbridged,
than in Type A, where it has a hydrogen bridge. Compared
to Type A species, the iridium atom is held somewhat more
intimately by the pentahapto boron framework in Type-B
species, as discussed below.
There is a close match of the experimental and calcu-
lated geometries of these Type-A species 2, 2a, 10 and
10a with the geometry reported experimentally for the
Type-B compound [8-Cl-7-(PPh3)-l-7P,10C-(Ph2P-ortho-
C6H4)-isonido-7-IrB9H6-9-(PPh3)] (5) in our original work
[24], for which hydrogen atoms were not locatable. These
close similarities suggest that the structure reported for 5
in our early work is in fact that of the Type-A compound
[8-Cl-7-(PPh3)-l-7P,10C-(Ph2P-ortho-C6H4)-7-H-l-3,7-H-
isonido-7-IrB9H7-9-(PPh3)] (11), i.e., with the same formu-
lation and substituent positioning as 5, but now with the
two extra hydrogen atoms that distinguish Type A from
Type B. This supposition is substantiated in the following
paragraphs. An implication is that compound 11 readily
forms from 5 by reaction with solvent upon attempted
recrystallisation. In our experimentation for the present
work we have also observed a specific such Type-B-to-
Type-A ready conversion, although we have not been able
to substantiate the precise nature of the individual species
involved (components GB and Gc, see experimental section
below, of Type-B and Type-A characteristics respectively).
In any event, it is of obvious interest to calculate both
structure and boron nuclear shielding for Type-B species
such as 5, in order to see whether the calculated shieldings
match those observed experimentally, and thence to rea-
sonably presume that a calculated structure that gives the
matching nuclear shieldings is a good model for the true
molecular structure.
These general and specific conclusions about structure
are supported by the calculated NMR shieldings (Table 2
Initial calculations involved the P-methyl non-ortho-
cycloboronated non-chlorinated ‘parent’ model [(PMe3)2-
IrB9H8(PMe3)] 12c, which minimised energetically to give
an unbridged isonido-type geometry XIV (for dimensions
see Table 1) and which then gave a boron nuclear shielding
pattern that mimicked that found experimentally for 5 and
its analogues [24]; in particular the very low nuclear shield-
ing of B(3) was reproduced. Thence, by steps, the two
ortho-cycloboronated species 12b (P-hydryl) and 12a
(P-methyl) were calculated, and finally the 8-chloro ortho-
cycloboronated P-methyl species 5a (Fig. 3), the direct
P-methyl analogue of the experimentally isolated P-phenyl
species 5. As with the type-A compound 2 above, the calcu-
lated boron nuclear shieldings for 5a paralleled those deter-
mined experimentally for 5, with similar deviations as for
2a between experimental and calculated values (Table 2;
see also Fig. 4 below).
Fig. 4. Stick diagrams representing the measured and calculated 11B
nuclear shieldings in 10-vertex {IrB9} metallaboranes, expressed as NMR
chemical shifts d (11B): as calculated and measured for the isonido Type-B
species 5 and 5a (top two traces); as calculated and measured for the
isonido Type-A species 2a and 2 (centre two traces), and as calculated and
measured for the classically closo-structured 13a and 13 (bottom two
traces). There are close similarities between the shieldings for classical
closo and for isonido Type A, suggesting similar electronic structures: Type
A can thus perhaps be regarded as ‘stretched’ closo. Type B, on the other
hand, is more significantly different, suggesting a bigger perturbation of
the overall electronic structure from that of classical closo, although that
the principal changes are associated with just three sites adjacent to the
metal centre, and that there is a similar ordering of relative shieldings,
suggest a relatively localised perturbation.
Selected interatomic distances for the Type-B com-
pounds 5a, 12a, 12b and 12c are given in Table 1 above.
It seems apparent from these calculated dimensions that
the Type-B structure is further along the closo–isonido–
isocloso continuum towards isocloso than are the com-
pounds of Type A, with Ir(7)–B(9) for Type B being shorter
˚
by between 0.1 and 0.2 A, and B(8)–B(10) for Type B
˚
longer by between 0.1 and 0.3 A. There are also other dif-

3730
J. Bould et al. / Inorganica Chimica Acta 359 (2006) 3723–3735
Table 4
and Fig. 4), which mimic those observed experimentally. In
particular, the diagnostically low Type-B shielding of the
B(3) nucleus, which contrasts to the higher shielding of this
position in Type-A species, is reasonably reproduced.
Additionally, Type A and Type B are clearly differentiated
by their overall boron nuclear shielding patterns. In terms
of d(11B) values, type A has B(3) to low field (ca. +20 to
+30 ppm), B(1), B(8), B(9) and B(10) at lower intermediate
field (ca. +10 ppm) and B(2), B(4), B(5) and B(6) at higher
field (ca. À30 ppm); whereas type B has B(8) and B(10) to
low field (ca. +30 ppm) and B(3) at very low field (ca. +60
to +70 ppm), with only B(9) at lower intermediate field (ca.
+10 ppm) and with B(9), B(2), B(4), B(5) and B(6) all at
intermediate to high field (ca. +5 to ca. À30 ppm). These
similarities and differences can be clearly be seen by consid-
eration of Fig. 4.
Measureda and calculated chemical shift data for compound 13 and its
related P-methyl model 13a and P-hydryl model 13b
Assignment d(11B)
d(11B)
d(1H)
Calculated (13a) [13b]
Measured Measured
(13)a
(13)a
3
10
1
8
9
4
2, 6
5
+22.3
+11.4
À4.6
[+7.8]
[+2.2]
+24.2b
+5.9
[À8.9]
À1.5
+2.87c
+3.05
À6.6
[À12.9]
[À23.0]
[À31.6]
À7.5
À5.7
À8.6
À19.0
À21.2
+0.1
À25.9, À21.7 [À34.9, À29.6] À22.1(2)
À0.35, À1.73
À26.9
[À36.1]
À26.2
À0.93
IrH
À4.17d
a
This work; CDCl3, 294–300 K; additionally d (31P) +30.5 (sharp),
+24.2 (sharp) and +17.9 (broad, 2P, two overlapping resonances).
b
Doublet 2J(31P–11B) ca. 155 Hz.
To substantiate further that the overall techniques are via-
ble for these third-row transition-element {MB9} species, we
also report here the results of an all-atom single-crystal
X-ray diffraction analysis for [7-(PPh3)-l-7P,8C-(Ph2P-
ortho-C6H4)-7-H-closo-7-IrB9H6-3,9-(PPh3)2] (13) (Fig. 5),
c
Doublet, probably coupled to P(3), 3J(31P–1H) 14 Hz.
d
Overlapping doublet of doublet of doublets; nJ(31P–1H) 22.0, 25.7 and
30.5 Hz.
1
together with its measured H and 11B NMR properties
and B, for the more problematical 10-vertex isonido-struc-
tured species. Compound 13 was isolated in yields of up to
33% from the reactions of PPh3 with [6,6-(PPh3)2-6-H-
nido-6-IrB9H13] (14).
(Table 4). Compound 13 has also been isolated as part of this
work and has unambiguous classical closo structure (Fig. 5
and Table 1; see also schematic cluster structure X above).
It is also robust in solution, and so there is not any doubt that
the experimentally determined NMR parameters (Table 4)
correspond to the molecular structure (Fig. 5). We thence
carried out DFT calculations of structure and of nuclear
shielding on its methyl-for-phenyl-substituted model [7-
(PMe3)-l-7P,8C-(Me2P-ortho-C6H4)-7-H-closo-7-IrB9H6- 3,
9-(PMe3)2] (13a). For both cluster structure and shielding,
the calculated values for 13a match with those observed by
experiment for 13 (Fig. 5 and Tables 1 and 4), supporting
the soundness of the approach to compounds of this nature,
and in particular giving confidence to the conclusions about
the constitutions and different natures of the two types, A
The reactions of PPh3 with [6,6-(PPh3)2-6-H-nido-6-
IrB9H13] (14) also yielded its known ortho-cycloboronated
variant [6-(PPh3)-l-6P,5C-(Ph2P-ortho-C6H4)-6-H-nido-6-
IrB9H12] (15), in trace quantities. Compound 15 has figured
much as a starting material or as a by-product in irida-
borane reaction chemistry [32–34], but its crystallographic
characterisation has previously proved elusive. As part of
this work we have finally been able to isolate crystals suit-
able for a single-crystal X-ray diffraction analysis (Fig. 6).
The compound is of straightforward classical nido 10-ver-
tex cluster configuration analogous to that of the binary
borane model nido-B10H14. The presence of the cyclised
Fig. 5. ORTEP-3 drawings [60] of (left-hand diagram) the crystallographically determined molecular structure of [7-(PPh3)-l-7P,8C-(Ph2P-ortho-C6H4)-7-
H-closo-7-IrB9H6-3,9-(PPh3)2] (13) and (right-hand diagram) the DFT-calculated molecular structure of [7-(PMe3)-l-7P,8C-(Me2P-ortho-C6H4)-7-H-closo-
7-IrB9H6-3,9-(PMe3)2] (13a), used to model 13 in the DFT calculations. For clarity methyl hydrogen atoms are omitted. Selected interatomic dimensions
are given in Table 1.

J. Bould et al. / Inorganica Chimica Acta 359 (2006) 3723–3735
3731
ortho-phenylene group causes a small elongation of the
B(5)–B(10) nido 10-vertex ‘gunwale’ vector compared to
that [35] in the uncyclised compound 14: 2.052(7) in 15 ver-
5 lacks two hydrogen atoms, and therefore has two elec-
trons fewer than either the closo-shaped cluster of com-
pound 13 or the isonido-shaped Type-A cluster of
compound 2. The larger differences in 11B nuclear shiel-
dings (Table 2 and Fig. 4) suggest more fundamental differ-
ences in electronic structure when 5 is compared to 2 and
13. A classical cluster opening from closo to classical nido
requires two extra cluster electrons. If the opening here
from closo to isonido also requires two electrons, then the
two electrons would formally originate from the iridium
core to give it a d4 iridium(V) configuration. This would
imply a four-orbital contribution to the cluster metal centre
which thence could be invoked to explain the unusual iso-
nido geometry, as opposed to the classical nido which
would have a three-orbital contribution from all cluster
constituents. Conversely, however, the distortion could be
a manifestation of an incipient ‘slippage’ related to that
of the ‘slipped-closo’ seven-vertex and 13-vertex species
mentioned above (schematics IV and VI). This would be
visualisable in terms of essentially square-planar d8 irid-
ium(I) character with a tendency towards a formal two-
orbital involvement with the cluster, rather than the
conventional three-orbital involvement. However, the irid-
ium-to-boron distances when 2 and 5 are compared suggest
a more intimate involvement of the iridium centre with the
borane residue in the Type-B compound 5, rather than a
less intimidate ‘slipped’ configuration: all the iridium-to-
boron distances in 5 are in fact either essentially the same
or shorter than those in compound 2.
˚
sus 1.997(16) A in 14. The other B(7)–B(8) ‘gunwale’ vector
remains essentially unchanged, with values of 2.016(7) for
˚
15 and 1.993(16) A for 14.
Whatever the electronic structures of the two 10-vertex
isonido cluster variants A and B – and it would appear that
these are significantly different, judging by the differences in
11B shielding properties (Fig. 4) which will be critically
dependent upon electronic structure – it would appear that
the formal oxidation state of iridium in the two clusters
may differ by two units. Whether it is best regarded as irid-
ium(I) in A and iridium(III) in B, or two units higher in
each case, i.e. iridium(III) in Type A and iridium(V) in
Type B, is an interesting question. Another possibility is
iridium(I) in Type B and iridium(III) in Type A, as dis-
cussed below. Certainly there is no need to invoke other
than conventional d6 iridium(III) in compound 13 of
straightforward classical closo constitution. The constitu-
ents of the cluster of isonido Type-A compound 2 make it
isoelectronic with classical closo 13. The cluster opening
to isonido observed for 2 could therefore be a ‘flexibility’
distortion associated with the accommodation of two extra
hydrogen atoms, with no catastrophic internal changes in
electronic structure (Fig. 4). This conclusion is perhaps
supported by the similarities in 11B shielding patterns
between straightforward closo (e.g., compound 13) and
isonido Type A (Tables 2–4) that could be taken to indicate
no catastrophic differences in internal electronic structure.
This would imply formal iridium(III) in Type-A compound
2. Compared to 2, the Type-B isonido cluster of compound
3. Conclusions
Although the electronic origins of the 10-vertex isonido
{MB9} structures, and variations among these structures,
remain unresolved, the work reported here now more confi-
dently defines the general structural type and the NMR
properties associated with the type. Within the basic 10-ver-
tex isonido skeletal geometry, two variants are substantiated.
Variant A has a metal-to-boron hydrogen bridge that is not
associated with an open-face boron atom. This characteristic
is very unusual in known metallaborane chemistry. Variant
A also has an endo-terminal iridium-hydride unit (schematic
XV above). Type B, by contrast, does not have these two
hydrogen characteristics (schematic XIV above). Variant A
is now well characterised crystallographically. Type B iso-
nido remains crystallographically uncharacterised within
the {IrB9} system, but DFT calculations show the structural
characteristics that are to be anticipated. As far as we are
aware, the only crystallographically characterised Type B
species remains the {MCB8} metallacarbaborane [3-
(OMe)-7-(PPh3)-l-7P,10C-(Ph2P-ortho-C6H4)-isonido-7,8-
IrCB8H6-10-(OH)], reported a quarter of a century ago [36].
The work presented here highlights the usefulness, in the elu-
cidation of metallaborane systems, of the synergic alliance of
(a) crystallography (including HYDEX and XHYDEX
[30,31] for the location of crystallographically unlocated
hydrogen atoms), (b) multinuclear NMR spectroscopy and
Fig. 6. ORTEP-3 drawing [60] of the crystallographically determined
molecular structure of [6-(PPh3)-l-6P,5C-(Ph2P-ortho-C6H4)-6-H-nido-6-
˚
IrB9H12] (15). Selected interatomic distances (in A) are as follows: from
Ir(6) to P(1) 2.3186(10), P(2) 2.2941(12), H(Ir) 1.68, B(2) 2.249(4), B(5)
2.236(4), B(7) 2.297(4), H(5,6) 1.76(6) and H(6,7) 1.79(5); B(5)–C(216) is
1.588 (5) B(5)B(10) is 2.052(7) and B(7)–B(8) is 2.016(7). Selected angles
(in deg.) are P(1)–Ir(6)–P(2) 97.46(4), P(2)–C(211)–C(216) 115.3(3),
C(216)–B(5)–Ir(6) 116.7(3), B(5)–Ir(6)–P(2) 81.32(12) and C(216)–B(5)–
Ir(6) 116.7(3).

3732
J. Bould et al. / Inorganica Chimica Acta 359 (2006) 3723–3735
(c) DFT calculations for structure and nuclear shielding. In
this regard, the successful application of this linked
approach to metallaboranes of the third-row transition ele-
ment iridium is particularly encouraging, since many of the
most interesting metallaborane systems involve late third-
row transition elements.
of compound 5. Measured NMR data for component C
(CDCl3, 294–297 K) are as follows: d(11B) +23.0 (1B; dou-
2
blet, J(31P–11B) ca. 155 Hz), À5.2 (2B) singlet, À7.5 (2B),
À25.2 (2B) and À26.9 (2B); d (1H) +2.60 (1H), +2.35 (1H),
À0.69 (2H), À1.77 (2H) and À5.40 (1H; doublet of dou-
2
blets, J(31P–1H) 50 and 25 Hz); d(31P) +16.8 (sharp sin-
glet) and ca. +17 (broad). However, attempts to obtain
suitable single crystals of C for X-ray diffraction analysis
were unsuccessful and the compound remains incompletely
characterised.
4. Experimental
4.1. General
In a second variant of the reaction, [6,6-(PPh3)2-6-H-
nido-6-IrB9H13] (14, 84 mg, 107 lmol) and PPh3 (56 mg,
210 lmol) were heated in solution in 1,2-dichloroethane
(5 ml) at reflux for the shorter time of 1 h. TLC separation
using 80/20 CH2Cl2/n-C6H12 as liquid phase afforded three
bands: D, RF 0.7 yellow; E, RF 0.5–0.4 yellow; and F, RF
0.4–0.1, mixed blue and purple. Band D was subjected to
Dried and degassed solvents were used throughout.
Reactions were carried out under dry dinitrogen, although
subsequent compound isolation using preparative chro-
matography was carried out in air. [6,6-(PPh3)2-6-nido-
6-IrB9H13] (14) was prepared as described previously
[32–34]. Thin-layer chromatography (TLC) was carried out
using 1 mm layers of silica-gel G (Fluka, type GF254)
made from water slurries on glass pates of dimensions
20 · 20 cm followed by drying in air at 80 ꢁC. HPLC was
performed on a 21 mm · 25 cm column (Knauer, Licho-
sorb Si60, 7 mm) with detection by change in the UV
absorption of the eluted liquid at k = 254 nm. NMR spec-
troscopy was performed at ca. 5.9 T, 9.4 T and 11.4 T
(fields corresponding to 250, 400 and 500 MHz 1H frequen-
cies respectively) using commercially available instrumen-
tation and established techniques [37–39], with overall
procedures as adequately described and enunciated else-
where [40–43]. Chemical shifts d are given in ppm relative
to N = 100 MHz for d(1H) ( 0.05 ppm) (nominally
TMS), N = 32.083972 MHz for d (11B) ( 0.5 ppm) (nomi-
nally [F3BOEt2] in CDCl3) [44] and N = 40.480730 MHz
for d(31P) ( 0.5 ppm) (nominally 85% aqueous H3PO4).
N is as defined by McFarlane [45].
HPLC separation (40/60 CH2Cl2/n-C6H12, 10 ml minÀ1
)
giving a single component (RT 12.7 min), identified from
its 11B,1H and 31P NMR spectra as [6-(PPh3)-l-6P,5C-
(Ph2P-ortho-C6H4)-6-H-nido-6-IrB9H12] 15. Compound 15
is previously reported [32–35], but not previously substanti-
ated crystallographically. Single-crystals suitable of 15 for
X-ray diffraction analysis were therefore obtained, by diffu-
sion of C6H14 into a solution of the compound in CDCl3.
Band E was identified as compound 13 (173 mg, 11 lmol,
10%). The mixed purple and blue bands F, ranging from
RF ca. 0.1 to ca. 0.4, were significantly stronger in this
experiment, and were subjected to HPLC separation (50/
50 CH2Cl2/n-C6H12, 10 ml minÀ1). Two strong blue-purple
components, G (RT 12.8 min) and H (RT 14.6 min), were
eluted. Repeated HPLC separation of G gave two compo-
nents, GA (blue, RT 16.0 min) and GB (blue, RT 19.6 min).
Attempts at crystallisation of the bands and characterisa-
tion were successful in the case of component GA: thus, a
layered n-C6H12/C6H6/CH2Cl2 diffusion gave blue crystals
of [7-(PPh3)-l-7P,10C-(Ph2P-ortho-C6H4)-7-H-l-3,7-H-iso-
nido-7-IrB9H8-9-(PPh3)] (10) (2.7 mg, 2.2 lmol, 2%).
X-ray diffraction analysis showed that the crystals had a
partial (10%) crystal occupancy by [8-Cl-7-(PPh3)-l-
7P,10C-(Ph2P-ortho-C6H4)-7-H-l-3,7-H-isonido-7-IrB9H7-9-
(PPh3)] (11), i.e., an overall formulation of [(PPh3)(Ph2-
PC6H4)HIrB9H6Cl0.1(PPh3)]. In the case of component GB
4.2. Isolation of [(PPh3)(Ph2PC6H4)HIrB9H6Cl(PPh3)]
(10) and [(PPh3)(Ph2PC6H4)HIrB9H6(PPh3)2] (13)
A solution of [6,6-(PPh3)2-6-H-nido-6-IrB9H13] (14;
106 mg, 128 lmol) and PPh3 (67 mg, 260 lmol) in 1,2-
dichloroethane (5 ml) was heated at reflux for 2 h under a
nitrogen atmosphere. After cooling, the solvent was
removed on a Schlenk line at ambient temperature, the res-
idue dissolved in dichloromethane, filtered through a silica
gel plug, reduced in volume, and applied to a number of
preparative TLC plates, which were then developed using
CH2Cl2/n-C6H12 (80/20) as liquid phase. Three bands were
apparent; A, RF 0.7, colourless (observed under UV illumi-
nation); B, RF 0.4, strong yellow and C, RF 0.35, yellow. A
number of relatively faint blue and purple bands were also
apparent, ranged below band C down to the baseline at RF
zero. Band A was identified as PPh3. Bands B and C were
re-chromatographed, affording, for component B, yellow
[7-(PPh3)-l-7P,8C-(Ph2P-ortho-C6H4)-7-H-closo-7-IrB9H6-
3,9-(PPh3)2] (compound 13) (63 mg, 42 lmol, 33%) and, for
component C, 6 mg of a yellow material. NMR spectros-
copy clearly indicated that component C was an analogue
1
preliminary 11B, H and 31P NMR spectroscopy in CDCl3
solution suggested an isonido iridanonaborane of type B,
similar in constitution to Type-A compound 10, but with
no IrH terminal or IrHB bridging hydrogen atoms; it had
a
characteristic type-B 11B(3) resonance at d(11B)
+64.0 ppm. However, relatively rapid reaction with the sol-
vent occurred, giving a solution containing principally a sin-
gle component GC that appeared to be a variant of
compound 10, now with IrH terminal and IrHB bridging
hydrogen atoms, at d(1H) À4.05 and À8.60 ppm, respec-
tively; also, the diagnostically very low-field 11B(3) reso-
nance had shifted to
a
type-A position at d(11B)
+32.5 ppm. Measured NMR data for component GB
(CDCl3, 294–297 K) are as follows: d(11B) +64.0, +29.6

J. Bould et al. / Inorganica Chimica Acta 359 (2006) 3723–3735
3733
2
(2B), +10.8, À1.9 [doublet, J(31P–11B) ca. 150 Hz], À7.9,
[(PPh3)(Ph2PC6H4)HIrB9H6(PPh3)2] (13): C72H66B9-
Ir1P4; M = 1344.7, yellow shard, 0.20 · 0.12 · 0.08 mm,
À15.4, À16.0 and À25.4; d(1H) +6.82, +5.22, +1.22,
+0.87, À0.31 and À1.09; d(31P) +49.7(sharp), +35.5 (sharp)
and +18 (broad). Measured NMR data for component GC
(CDCl3, 294–297 K) are as follows: d(11B) +32.5, +6.8,
À0.9, À10.5 [doublet, 2J(31P–11B) ca. 147 Hz], À21.7,
À24.0 and À25.9; d(1H) +4.23, +4.03, +0.11, À0.15,
À0.54, À2.00, À4.05 and À8.60 (overlapping doublet of
doublets of doublets structure, all nJ(31P–1H) ca. 20–30 Hz).
monoclinic, a = 41.1175(4), b = 15.0755(2), c = 29.0142(3)
3
˚
˚
˚
A, b = 121.541(1)ꢁ, V = 15327.9(3) A , k = 0.71073 A (Mo
Ka), T = 120(2) K, space group C2/c Z = 8, l(Mo Ka)
1.862 mmÀ1
, R1 = 0.0366 for 14518 reflections with
I > 2r(I), and wR2 = 0.0815 for all 17566 unique reflections.
Crystals from [(PPh3)(Ph2PC6H4)IrB9H12] (15) suitable
for the single-crystal X-ray diffraction analysis were grown
in a 5-mm o.d. glass tube from a concentrated solution in
CDCl3, by liquid-liquid diffusion from an overlayer of n-
hexane through an intermediate layer of benzene. The data
collection was carried out using a (pre-Bruker) Nonius
KappaCCD area-detector diffractometer with Mo Ka radi-
ation. The asymmetric unit contained half a molecule of n-
hexane. All non-hydrogen atoms were refined with aniso-
tropic displacement parameters. Hydrogen atoms were
refined isotropically in calculated positions with the excep-
tion of the terminal and the bridging metal hydrogen atoms
H(6) and H(5,6). These were located using XHYDEX
[30,31]. The bridging hydrogen atom H(5,6) thence refined
satisfactorily, mirroring H(6,7) which was located in the
difference map. Atom H(6) was refined in a fixed position.
Crystal data are as follows.
4.3. Single-crystal X-ray diffraction analysis
Crystals from [(PPh3)(Ph2PC6H4)HIrB9H6Cl0.1(PPh3)]
(10) suitable for the X-ray work were grown in a 5-mm glass
tube from concentrated CH2Cl2 solution by liquid–liquid
diffusion from an overlayer of n-pentane through an inter-
mediate layer of benzene. The data collection was carried
out using a Bruker SMART APEX2 CCD diffractometer
with synchrotron radiation (SRS station 9.8, CCLRC
Daresbury Laboratories, UK) [52–54]. The asymmetric unit
contained half a molecule of benzene and a disordered mol-
ecule of pentane, which was refined isotropically over two
positions. All non-hydrogen atoms were refined with aniso-
tropic displacement parameters. Hydrogen atoms were
refined isotropically in calculated positions with the excep-
tion of the terminal and the bridging metal hydrogen atoms
H(7) and H(3,7). These were located in the difference map
and refined freely. Their positions were confirmed using
XHYDEX [30,31] which also corresponded well to those
calculated using DFT. A 10% partial occupancy of the
B(8)H(exo) site by a BCl(exo) unit emerged from the analy-
sis. Crystal data are as follows. [{(PPh3)(Ph2PC6H4)HIrB9-
H5.9Cl0.1(PPh3)}{C6H12}0.5{C6H6}]: C62H67.9B9Cl0.1IrP3;
M = 1199.1, blue block, 0.05 · 0.04 · 0.02 mm, triclinic,
[{(PPh3)(Ph2PC6H4)IrB9H12)}{C6H14}0.5]: C39H49B9Ir1P2;
M = 869.2, yellow block, 0.43 · 0.24 · 0.10 mm, triclinic,
˚
a = 11.108(2), b = 12.545(3), c = 15.410(3) A, a = 93.87(3),
3
˚
b = 107.67(3), c = 104.82(3)ꢁ, V = 1953.3(7) A , Dcalc
=
1.478 Mg mÀ3; k = 0.71073 A (Mo Ka), T = 150(2) K,
˚
space group P1, Z = 2, l = 3.528 mmÀ1, R1 = 0.0401 for
ꢀ
8510 reflections with I > 2r(I), and wR2 = 0.1105 for all
8905 unique reflections.
4.4. Computational method
˚
a = 10.4823(13), b = 13.2092(17), c = 22.321(3) A, a =
3
˚
88.051(2), b = 89.612(2), c = 68.215(2)ꢁ, V = 2868.2(6) A ,
Calculations carried out in this study were performed
using the GAUSSIAN98 and GAUSSIAN03 packages [55–59].
Structures were initially optimised with the STO-3G*
basis-sets for B, C, P, Cl, H and with the LANL2DZ
basis-set for the Ir atom, using standard ab initio meth-
ods. The final optimisations, including frequency analyses
to confirm the true minima, together with GIAO NMR
nuclear-shielding predictions, were performed using
B3LYP methodology, again with the 6-31G* basis-sets
for H, B, C, P and Cl, and the LANL2DZ basis-set for
the iridium atom. The GIAO NMR nuclear-shielding pre-
dictions were performed on the final optimised geometries.
In each case the compounds were modelled initially using
hydrogen atoms rather than phenyl groups on the phos-
phine ligands, and then, based on the geometries so
obtained, the structural and boron nuclear shielding cal-
culations were extended to the final P-methyl models 2a,
5a, 10a and 13a. For species 2, 10 and 13, the frequency
analyses were carried out on the simpler PH3 variants 2b,
10b, and 13b: time constraints precluded frequency calcu-
lations on the methyl variants 2a, 10a and 13a. For the
experimentally less well established species 5, however,
Dcalc = 1.388 Mg mÀ3; k = 0.6933 A (synchrotron, wig-
˚
gler-generated, silicon-monochromated), T = 120(2) K,
space group P1, Z = 2, l = 2.455 mmÀ1, R1 = 0.0423 for
ꢀ
14229 reflections with I > 2r(I), and wR2 = 0.0797 for all
16997 unique reflections.
Crystals of [(PPh3)(Ph2PC6H4)HIrB9H6(PPh3)2] (13)
suitable for the X-ray analysis were grown in a 5-mm glass
tube from a concentrated solution in CH2Cl2 by liquid-
liquid diffusion from an overlayer of cyclohexane through
an intermediate layer of toluene. The collection of data were
carried out using a Bruker-Nonius KappaCCD area detec-
tor using Mo Ka radiation generated from a Bruker-Nonius
FR591 rotating-anode X-ray source. The structure was
solved using standard methods [46–50]. All non-hydrogen
atoms were refined with anisotropic displacement parame-
ters. Hydrogen atoms were refined isotropically in calculated
positions with the exception of the terminal metal hydride
which was freely refined. XHYDEX calculation confirmed
the location. The unit cell contained disordered solvent mol-
ecules that were included in the model using the PLATON/
SQUEEZE procedure [51]. Crystal data are as follows.

3734
J. Bould et al. / Inorganica Chimica Acta 359 (2006) 3723–3735
[14] N.W. Alcock, J.G. Taylor, M.G.H. Wallbridge, J. Chem. Soc.,
Dalton Trans. (1987) 1805.
[15] See, together with references therein: H.M. Colquhoun, T.J. Greenh-
ough, M.G.H. Wallbridge, J. Chem. Soc., Dalton Trans. (1985) 761.
[16] D.M.P. Mingos, M.I. Forsyth, A.J. Welch, J. Chem. Soc., Chem.
Commun. (1977) 605.
[17] D.M.P. Mingos, M.I. Forsyth, A.J. Welch, J. Chem. Soc., Dalton
Trans. (1978) 1363.
[18] See, for example: G.K. Barker, M. Green, T.P. Onak, F.G.A. Stone,
C.B. Ungermann, A.J. Welch, J. Chem. Soc., Chem. Commun. (1978)
169.
[19] See, for example, together with references therein: J.D. Kennedy,
Main Group Metal Chem. 12 (1989) 149.
[20] J.D. Kennedy, in: Disobedient skeletons, in: J. Casanova (Ed.), The
Borane–Carborane–Carbocation Continuum, Wiley, New York,
1998, p. 85 (Chapter 3).
frequency analyses were carried out on the P-methyl
model 5a. The reasonable match of both structures and
individual 11B nuclear shieldings within each of the group-
ings {2,2a,2b}, {5,5a}, {10,10a,10b} and {13,13a,13b}
supports the validity of this approach. In the results of
all the structural calculations it is noted that all the dis-
tances from iridium to boron are somewhat longer than
those observed experimentally. We have found this to
be a general feature in DFT calculations of the structure
of metallaboranes of second-row and third-row transition
elements using the same basis-sets, for which we have
reported on ruthenaboranes [61] and platinaboranes [62].
Other calculated intercluster connectivities, e.g., interbo-
ron distances, much more closely match those determined
experimentally.
[21] M.E. O’Neill, K. Wade, Polyhedron 2 (1983) 963.
ˇ
´
[22] See, for example, together with references therein: B. Stıbr, J.D.
Kennedy, M. Thornton-Pett, E. Drda´kova´, T. Jel´ınek, J. Plesˇek,
Collect. Czech. Chem. Commun. 57 (1992) 1439.
Acknowledgements
ˇ
´
´
´
[23] B. Stıbr, J.D. Kennedy, E. Drdakova, M. Thornton-Pett, J. Chem.
Soc., Dalton Trans. (1994) 229.
[24] J. Bould, J.D. Kennedy, M. Thornton-Pett, J. Chem. Soc., Dalton
Trans. (1992) 563.
[25] M. Thornton-Pett, personal communication.
[26] See, for example, together with references therein: M.J. Carr, S.D.
A proportion of the work has been personally financed
by JB, who thanks the School of Chemistry of the Univer-
sity of Leeds for facilities and Professors P.J. Kocienski
and M.J. Pilling for their good offices. JB and JDK thank
Colin Kilner for help with crystallographic data collection
and Simon Barrett for help with NMR spectroscopy.
Instrumentation grants from the EPSRC are acknowledged
(J/56929, K/05818, L/49505, M/83360, R/61949). We also
thank CCLRC for the award of synchrotron beam time at
the UK SRS Daresbury laboratory, and the EPSRC UK
National X-ray Service for the collection of a crystallo-
graphic data set.
ˇ
´
´
Perera, T. Jelınek, C.A. Kilner, B. Stıbr, J.D. Kennedy, J. Organomet.
Chem. (EUROBORON 3 Special Edition) 690 (2005) 2857.
[27] J.D. Kennedy, Prog. Inorg. Chem. 32 (1984) 519.
[28] J.D. Kennedy, Prog. Inorg. Chem. 34 (1986) 211.
[29] L. Barton, D.K. Srivastava, in: Metallaboranes, in: G. Wilkinson,
F.G.A. Stone, E. Abel (Eds.), Comprehensive Organometallic
Chemistry II, vol. 1, Pergamon, Elmsford, New York, 1995, p. 275
(Chapter 8).
[30] XHYDEX, A.G. Orpen, J. Chem. Soc., Dalton Trans. (1980) 2509.
[31] As implemented in G.X. Win, L.J. Farrugia, J. Appl. Crystallogr. 32
(1999) 837.
[32] J. Bould, J.E. Crook, N.N. Greenwood, J.D. Kennedy, M. Thorn-
ton-Pett, J. Chem. Soc., Dalton Trans. (1990) 1441.
[33] J. Bould, N.N. Greenwood, J.D. Kennedy, J. Chem. Soc., Dalton
Trans. (1990) 1451.
[34] J. Bould, M. Bown, R.J. Coldicott, E.J. Ditzel, N.N. Greenwood, I.
Macpherson, P. MacKinnon, M. Thornton-Pett, J.D. Kennedy, J.
Organomet. Chem. (EUROBORON 3 Special Edition) 690 (2005)
2701.
[35] S.K. Boocock, J. Bould, N.N. Greenwood, J.D. Kennedy, W.S.
McDonald, J. Chem. Soc., Dalton Trans. (1982) 713.
[36] J.E. Crook, N.N. Greenwood, J.D. Kennedy, W.S. McDonald, J.
Chem. Soc., Chem. Commun. (1981) 933.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.ica.2006.
01.038. Crystallographic data for the structural analysis
are deposited with the Cambridge Crystallographic Data
Centre, CCDC deposition nos. 290672 for 10, 29067 for 3
13 and 290674 for 15.
References
[37] See, for example, together with references therein: J. Schraml, J.M.
Bellama, Two-Dimensional NMR Spectroscopy, Wiley, New York,
1982.
[1] D.M.P. Mingos, Nature (Phys. Sci.) 236 (1972) 99.
[2] D.M.P. Mingos, Pure Appl. Chem. 63 (1991) 807.
[3] R.E. Williams, Inorg. Chem. 1 (1971) 210.
[38] W.C. Hutton, T.L. Venable, R.N. Grimes, J. Am. Chem. Soc. 106
(1984) 29.
[39] X.L.R. Fontaine, J.D. Kennedy, J. Chem. Soc., Dalton Trans. (1987)
1573.
[4] R.E. Williams, Adv. Inorg. Chem. Radiochem. 18 (1976) 67.
[5] K. Wade, J. Chem. Soc., Chem. Commun. (1971) 792.
[6] K. Wade, Adv. Inorg. Chem. Radiochem. 18 (1976) 1.
[7] R.W. Rudolph, W.P. Pretzer, Inorg. Chem. 11 (1972) 1974.
[8] R.W. Rudolph, Acc. Chem. Res. 9 (1976) 446.
ˇ
[40] See, for example: J. Plesˇek, B. St´ıbr, X.L.R. Fontaine, J.D. Kennedy,
S. Herˇma´nek, T. Jel´ınek, Collect. Czech. Chem. Commun. 56 (1991)
1618.
[9] R.M. Wing, J. Am. Chem. Soc. 89 (1967) 5599.
[10] M.F. Hawthorne, D.C. Young, T.D. Andrews, D.V. Howe, R.L.
Pilling, A.D. Pitts, M. Reintjes, L.F. Warren Jr., P.A. Wegner, J.
Am. Chem. Soc. 90 (1968) 879.
[11] See, for example, together with references therein: G.B. Dunks, M.F.
Hawthorne, closo-Heteroboranes exclusive of carboranes, in: E.L.
Muetterties (Ed.), Boron Hydride Chemistry, Academic, New York,
San Francisco, and London, 1975, p. 383 (Chapter 11).
[12] L.F. Warren Jr., M.F. Hawthorne, J. Am. Chem. Soc. 90 (1968) 4823.
[13] R.M. Wing, J. Am. Chem. Soc. 90 (1968) 4828.
[41] G. Ferguson, J.D. Kennedy, X.L.R. Fontaine, Faridoon, T.R.
Spalding, J. Chem. Soc., Dalton Trans. (1988) 2555.
[42] M.A. Beckett, M. Bown, X.L.R. Fontaine, N.N. Greenwood, J.D.
Kennedy, M. Thornton-Pett, J. Chem. Soc., Dalton Trans. (1988) 1969.
[43] See, for example, together with references therein: D. Reed, Chem.
Soc. Rev. 22 (1993) 109.
[44] See, for example, together with references therein: J.D. Kennedy,
Boron, in: J. Mason (Ed.), Multinuclear NMR, Plenum, New York
and London, 1987, p. 221 (Chapter 8).

J. Bould et al. / Inorganica Chimica Acta 359 (2006) 3723–3735
3735
[45] W. McFarlane, Proc. Roy. Soc. (London), Ser. A 306 (1968) 185.
[46] G.M. Sheldrick, Acta Crystallogr., Sect. A 46 (1990) 467.
[47] G.M. Sheldrick, SHELX 97, Program System for X-ray Structure
Refinement, University of Go¨ttingen, 1997.
[48] Z. Otwinowski, W. Minor, Meth. Enzymol. 27 (1997) 307.
[49] COLLECT, Data Collection Strategy Program, Nonius, 1999.
[50] Z. Otwinowski, W. Minor, SMART, SAINT, SADABS, SHELXTL and
APEX2, Bruker AXS Inc., Madison, WI, USA, 1997-2004.
[51] P. van der Sluis, C.A. Spek, Acta Crystallogr., Sect. A 46 (1990) 194.
[52] R.J. Cernik, W. Clegg, C.R.A. Catlow, G. Bushnell-Wye, J.V.
Flaherty, G.N. Greaves, I. Burrows, D.J. Taylor, S.J. Teat, M.
Hamichi, J. Synchrotron Rad. 4 (1997) 279.
[56] U. Meier, C. van Wullen, M. Schindler, J. Comput. Chem. 13 (1992)
¨
551.
[57] J.A. Pople et al., Gaussian 98, Revision A.7, Gaussian, Inc.,
Pittsburgh, PA, 1998, and references therein.
[58] J.A. Pople et al., Gaussian 03, Revision B.05, Gaussian, Inc.,
Pittsburgh, PA, 2003.
[59] See, for example, together with references therein: J.W. Bausch, A.J.
Tebben, Computational studies of nido-C4B7H11 carboranes and
nido-8-vertex boranes and carboranes, in: J. Casanova (Ed.), The
Borane–Carborane–Carbocation Continuum, Wiley, New York, NY,
1998, p. 217 (Chapter 9).
[60] ORTEP3 for Windows, L.J. Farrugia, J. Appl. Crystallogr. 30 (1997)
565.
[61] J. Bould, M. Bown, J.D. Kennedy, Collect. Czech. Chem. Commun.
70 (2005) 410.
[53] J. Bould, W. Clegg, J.D. Kennedy, S.J. Teat, M. Thornton-Pett, J.
Chem. Soc., Dalton Trans. (1997) 2005.
[54] W. Clegg, M.R.J. Elsegood, S.J. Teat, C. Redshaw, V.C. Gibson, J.
Chem. Soc., Dalton Trans. (1998) 3037.
´
[62] J. Bould, T. Jelınek, S.A. Barrett, S.J. Coles, M.J. Hursthouse,
ˇ
´
[55] W. Kutzelnigg, M. Schindler, U. Fleischer, NMR, Basic Principles
and Progress, vol. 23, Springer, Berlin, New York, 1990, p. 165.
M. Thornton-Pett, B. Stıbr, J.D. Kennedy, Dalton Trans. (2005)
1499.