Reactions of isonitriles with [Fe3(CO)12] and [Ru3(CO)12] monitored by electrospray mass spectrometry: Structural characterisation of [Fe3(CO)10(CNPh)2] and [Ru4(CO)11(μ3-η2 -CNPh)2(CNPh)]

Article abstract of DOI:10.1016/j.jorganchem.2004.02.026

The reactions of [Fe₃(CO)₁₂] or [Ru₃(CO)₁₂] with RNC (R=Ph, C₆H₄OMe-p or CH₂SO₂ C₆H₄Me-p) have been investigated using electrospray mass spectrometry. Species arising from substitution of up to six ligands were detected for [Fe₃(CO)₁₂], but the higher-substituted compounds were too unstable to be isolated. The crystal structure of [Fe₃(CO)₁₀(CNPh)₂] was determined at 150 and 298 K to show that both isonitrile ligands were trans to each other on the same Fe atom. For [Ru₃(CO)₁₂] substitution of up to three COs was found, together with the formation of higher-nuclearity clusters. [Ru₄(CO)₁₁(CNPh)₃] was structurally characterised and has a spiked-triangular Ru₄ core with two of the CNPh ligands coordinated in an unusual μ₃-η² mode.

Full text of DOI:10.1016/j.jorganchem.2004.02.026

Journal of Organometallic Chemistry 689 (2004) 1691–1701
www.elsevier.com/locate/jorganchem
Reactions of isonitriles with [Fe3(CO)12] and [Ru3(CO)12]
monitored by electrospray mass spectrometry: structural
characterisation of [Fe3(CO)10(CNPh)2] and
[Ru4(CO)11(l3-g2-CNPh)2(CNPh)]
*
Corry Decker, William Henderson, Brian K. Nicholson
Department of Chemistry, School of Science and Technology, University of Waikato, Private Bag 3105, Hamilton, New Zealand
Received 29 January 2004; accepted 17 February 2004
Abstract
The reactions of [Fe3(CO)12] or [Ru3(CO)12] with RNC (R ¼ Ph, C6H4OMe-p or CH2SO2C6H4Me-p) have been investigated
using electrospray mass spectrometry. Species arising from substitution of up to six ligands were detected for [Fe3(CO)12], but the
higher-substituted compounds were too unstable to be isolated. The crystal structure of [Fe3(CO)10(CNPh)2] was determined at 150
and 298 K to show that both isonitrile ligands were trans to each other on the same Fe atom. For [Ru3(CO)12] substitution of up to
three COs was found, together with the formation of higher-nuclearity clusters. [Ru4(CO)11(CNPh)3] was structurally characterised
and has a spiked-triangular Ru4 core with two of the CNPh ligands coordinated in an unusual l3-g2 mode.
Ó 2004 Elsevier B.V. All rights reserved.
Keywords: Iron carbonyl; Ruthenium carbonyl; Isonitrile; Electrospray mass spectrometry; X-ray crystallography
1. Introduction
RNC therefore allows subtle changes that can provide
extra information concerning the parent species.
[Fe3(CO)12] and [Ru3(CO)12] have played key roles
in the development of cluster chemistry. They have
been important examples in the debates about struc-
tures and fluxional properties, particularly with respect
to Johnson’s ligand polyhedron arguments for the
adoption of an icosahedral arrangement of CO ligands
for the smaller Fe3 triangle and the larger cube-octa-
hedron arrangement for the Ru3 core [1]. This leads to
the bridged C2v structure for the former but a non-
bridged D3h structure for the latter. [Fe3(CO)12] has
also been the archetypal compound for the sometimes
heated arguments concerning the mechanisms for CO-
ligand fluxionality [2].
For the iron case, some derivatives were prepared
many years ago [3]. The first structurally characterised
example was [Fe3(CO)11(CNBut)] where the isonitrile
ligand occupies an axial position on the unique iron atom
of the parent C2v structure [4]. This change is sufficient to
give a completely ordered structure rather than the 50:50
disordered form of [Fe3(CO)12] [5], presumably because
the isonitrile ligand prefers a terminal rather than the
bridging position that the ‘‘star-of-David’’ disorder
model would require. Subsequently, the di-substituted
example [Fe3(CO)10(CNBu)2] was characterised, with
both isonitrile ligands on the unique iron atom, one axial
and one equatorial [6]. Again, the structure is fully or-
dered. So far no tri-substituted examples have been
structurally characterised, though they are known [7], so
the pattern of ligand substitution is undetermined since it
cannot be deduced from infrared spectra, nor from
NMR spectra because of the CO-fluxionality involved.
For [Ru3(CO)12] the situation is reversed. The parent
cluster shows no disorder of the metal core [8], whereas
Of all the ligands that most resemble CO, but are
clearly distinguished from it, are isonitriles RNC. Re-
placement of some of the CO ligands of [M3(CO)12] with
* Corresponding author. Fax: +64-7-838-4219.
E-mail address: b.nicholson@waikato.ac.nz (B.K. Nicholson).
0022-328X/$ - see front matter Ó 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2004.02.026

1692
C. Decker et al. / Journal of Organometallic Chemistry 689 (2004) 1691–1701
the mono- and di-substituted examples with the CNBut
ligand are disordered in the ‘‘star-of-David’’ fashion [9–
11]. For [Ru3(CO)11(CNBut)] the disorder is tempera-
ture-dependent, around 12% at room temperature and
6% at 130 K while at 100 K it becomes fully ordered [9–
11].
For [Ru3(CO)10(CNBut)2] the disorder is 50:50, with
the isonitrile ligands attached to different Ru atoms [9].
More recently, Farrugia and Mertes have systemati-
cally studied the structures of mono- and di-substituted
CNBut derivatives of the mixed-metal [Fe2Ru(CO)12]
and [FeRu2(CO)12] clusters, where the CNBut ligands
prefer axial positions on a ruthenium atom [12].
It is interesting to note that the five structurally
characterised [M3(CO)11(CNBut)] examples (M3 ¼ Fe3,
Ru3, Os3, FeRu2, Fe2Ru) [4,9,12,13] make up an un-
usual set of compounds that are isomorphous, but are
not isostructural, since the Fe3 example is derived from
the C2v structure, the Ru3 example based on a twisted-
D3h with a disordered Ru3 unit (12% at room temper-
ature), and the Os example is twisted-D3h but not
disordered. The FeRu2 compound is also based on a D3h
form, with a slightly disordered triangle at 293 K (3.4%)
as well as scrambling amongst the metal sites, while the
Fe2Ru is C2v with no disorder. This is most readily
understood in terms of Johnson’s model – the almost
constant ligand polyhedron determines the crystal
packing and hence space group, with the varying metal
cores occupying the central hole according to their sizes
and ease of rotation within the cage [1]. The work on the
structures and fluxionality of all three [M3(CO)12]
members of the iron triad has been reviewed [2].
We have now investigated the reactions of [M3(CO)12]
were dissolved or diluted in MeOH (unless otherwise
specified) and injected via a Rheodyne valve fitted with a
10 ll sample loop, with a mobile phase flow rate of 0.02
ml minꢀ1. Skimmer cone voltages were kept low to
minimise fragmentation, 20 V for positive-ion spectra
and 5 V for negative-ion spectra. For the negative-ion
spectra, Na[OMe] was added where indicated to aid
ionisation [15].
2.1. Reactions of [Fe3(CO)12] with isonitriles
The general reaction was carried out as follows:
[Fe3(CO)12] (0.5 g, 0.99 mmol) was dissolved in THF (10
ml). The isonitrile (3.97 mmol) was added by syringe or
as a solid and the reaction stirred at room temperature
for ꢁ2 h. The course of the reaction was monitored by
TLC and aliquots were withdrawn at appropriate time
intervals for examination by ESMS.
When the reactions were complete and preliminary
ESMS studies had been carried out, the solvent was re-
moved in vacuo. The reaction products were redissolved
in a minimum amount of CH2Cl2, and chromatographed
on silica plates with a solvent mixture of petroleum
spirits/CH2Cl2 (CNPh derivatives 1:1, CNC6H4OMe-p
derivatives 1:2, TosMIC derivatives 1:2). Yields of the
complexes were not determined because only parts of the
crude reaction product were used at a time for chroma-
tography, and significant decomposition always accom-
panied attempted separation. The compounds that
survived chromatography were removed from the plates
and recrystallised for further characterisation. Satisfac-
tory elemental analyses were obtained for CNPh and
TosMIC derivatives, but not for any compounds in-
volving the CNC6H4OMe-p ligand. The following de-
rivatives were characterised:
(M3
¼
Fe3 or Ru3) with the isonitrile ligands
CNC6H4OMe-p, CNCH2SO2C6H4Me-p (TosMIC) and
CNPh. The first two of these were chosen because their
potentially protonatable sites were expected to facilitate
chemical ionisation of neutral complexes for electrospray
mass spectrometry (ESMS) by giving [M + H]þ ions (c.f.
[14]). However, the simpler ligand CNPh was also found
to form complexes which generated ions in the ESMS
source, so was included. The emphasis in our studies was
the characterisation of reaction mixtures by ESMS, and
the detection of more highly substituted species.
2.1.1. [Fe3(CO)11(CNPh)]
Rf ¼ 0:68 (green). Found: C, 37.79; H, 1.31; N,
2.35%. C18H5Fe3NO11 requires C, 37.31; H, 0.86; N,
2.42%. IR (CHCl3): m(CN) 2153 (m); m(CO) 2080 (m),
2039 (sh), 2033 (s), 2014 (sh), 1998 (sh) cmꢀ1. ESMS
(+ve ion): [M + H]þ m=z 580 (100%).
2.1.2. [Fe3(CO)10(CNPh)2]
Rf ¼ 0:49 (green). Found: C, 44.65; H, 1.52; N,
4.14%. C24H10Fe3N2O10 requires C, 44.04; H, 1.53; N,
4.28%. IR (pet. sprits): m (CN) 2150 (w), 2118 (m);
(CO) 2056 (s), 2030 (sh), 2019 (s), 1996 (sh), 1988 (sh),
1974 (sh) cmꢀ1. ESMS (+ve ion): [M + H]þ m=z 655
(100%).
2. Experimental
Reactions and manipulations were performed under
nitrogen in standard Schlenk equipment, except for
preparative chromatography which was carried out ex-
peditiously in air using silica plates (Merck, Silica gel
60G). Petroleum spirits refers to a 60–80 °C fraction.
Infrared spectra were recorded on a Digilab FTS40
spectrometer, NMR on a Bruker AC300P machine, and
ESMS on a VG Platform II mass spectrometer. Samples
2.1.3. [Fe3(CO)9(CNPh)3]
Rf ¼ 0:24 (green). Elemental analysis did not give
acceptable values, e.g. Found: C, 53.34; H, 3.98; N,
4.51%. C30H15Fe3N3O9 requires C, 49.40; H, 2.08; N,
5.76%. IR (pet. spirits): m(CN) 2148 (w), 2113 (m);

C. Decker et al. / Journal of Organometallic Chemistry 689 (2004) 1691–1701
1693
m(CO) 2052 (m), 2038 (m), 2016 (sh), 2009 (s), 1990
2.1.10. [Fe3(CO)9(TosMIC)3]
(sh), 1978 (sh) cmꢀ1. ESMS (+ve ion): [M + H]þ m=z
730 (100%).
Rf ¼ 0:39 (green). Found: C, 44.18; H, 4.28; N,
3.43%. C36H27Fe3N3S3O15 requires C, 42.99; H, 2.69; N,
4.18%. IR (CHCl3): m(CN) 2168 (m, br), 2143 (w); m(CO)
2083 (sh), 2074 (sh), 2061 (m), 2030 (s), 2015 (sh), 1991
(sh), 1965 (sh) cmꢀ1. ESMS (+ve ion): [M + H]þ m=z
1006 (100%). ESMS ()ve ion): [M ) H]ꢀ m=z 1004
(100%), [M ) CO ) H]ꢀ m=z 976 (20%).
2.1.4. [Fe3(CO)8(CNPh)4]
Rf ¼ 0:09 (green). IR (CHCl3): m(CN) 2148 (w), 2113
(m); (CO) 2000 (s), 1990 (sh), 1986 (sh), 1965 (sh) cmꢀ1
ESMS (+ve ion): [M + H]þ m=z 806 (100%).
.
2.2. Thermolysis of [Fe3(CO)12 ꢀ n(CNR)n], n ¼ 1,2
2.1.5. [Fe3(CO)11(CNC6H4OMe-p)]
Rf ¼ 0:58 (green). IR (CHCl3): m(CN) 2157 (br);
m(CO) 2081 (s), 2057 (sh), 2034 (s), 2041 (sh), 1996 (sh),
1972 (sh) cmꢀ1. 1H NMR (CDCl3): d 7.44 (2H, s, H20), d
6.95 (2H, s, H30), d 3.86 (3H, s, OCH3). 13C–{1H} NMR
(CDCl3): d 212.7 (s, CO), d 160.1 (s, C40), d 126.9 (s,
C20), d 114.8 (s, C30), d 55.7 (s, OCH3). ESMS (+ve ion):
[M + H]þ m=z 610 (100%).
A small amount of [Fe3(CO)12 ꢀ n(CNR)n], n ¼ 1; 2
(20–50 mg) was gently heated in toluene (5 ml) to 75 °C.
The reaction was kept at this temperature for ꢁ5 min
(higher temperatures or longer reaction times were
found to give lower yields by decomposition of the
products, as monitored by TLC). The solvent was re-
moved in vacuo. The products were redissolved in a
minimum amount of CH2Cl2, and chromatographed on
silica plates with a solvent mixture of petroleum spirits/
CH2Cl2 (1:1). Yields of the isolated complexes were not
determined because only parts of the crude reaction
product were used at a time for chromatography
and significant decomposition always accompanied
separation.
2.1.6. [Fe3(CO)10(CNC6H4OMe-p)2]
Rf ¼ 0:21 (green). Elemental analysis did not give
acceptable values, e.g. C, 47.28; H, 2.96; N 3.67%.
C26H14Fe3N2O12 requires C, 43.70; H, 0.16; N, 3.92%.
IR (CHCl3): m(CN) 2126 (br); m(CO) 2059 (s), 2032 (s),
2014 (sh), 1994 (sh), 1989 (sh), 1964 (sh) cmꢀ1. IR (pet.
spirits): m(CN) 2135 (sh) 2125 (m); m(CO) 2057 (s), 2028
(s), 2024 (sh), 2018 (s), 2013 (sh), 1999 (m), 1987 (sh),
2.2.1. [Fe3(CO)9(l3-g2-CNPh)]
1972 (s) cmꢀ1
.
1H NMR (CDCl3): d 7.39 (4H, d,
Rf ¼ 0:83 (brown). IR (CHCl3): m(CO) 2087 (m), 2055
(sh), 2041 (s), 2033 (s), 2017 (sh), 1996 (sh), 1978 (sh)
cmꢀ1. ESMS (MeOH/MeOꢀ, )ve ion): [M + MeO]ꢀ m=z
554 (100%).
JH2 ;H3 ¼ 2:5 Hz, H20), d 6.91 (4H, d, 3JH3 ;H2 ¼ 5:8 Hz,
H30), d 3.85 (3H, s, OCH3). ESMS (+ve ion): [M + H]þ
m=z 715 (100%).
3
0
0
0
0
Similarly, green [Fe3(CO)10(CNBu)2] was converted
to brown [Fe3(CO)8(l3-g2-CNPh)(CNPh)] after 5 min
at 75 °C in toluene. Chromatography gave [Fe3(CO)8
(l3-g2-CNPh)(CNPh)]: Rf ¼ 0:43 (red). IR (CHCl3):
m(CN) 2146 (m), 2107 (m); m(CO) 2041 (s), 2023 (sh),
2015 (sh) cmꢀ1. ESMS (MeOH/MeOꢀ, )ve ion):
[M + MeO]ꢀ m=z 629 (100%).
2.1.7. [Fe3(CO)9(CNC6H4OMe-p)3]
Rf ¼ 0:06 (green). IR (pet. sprits): m(CN) 2120 (m);
m(CO) 2053 (sh), 2037 (m), 2021 (sh), 2016 (sh), 2006 (s),
2000 (sh), 1991 (sh), 1984 (sh), 1974 (m) cmꢀ1. ESMS
(+ve ion): [M + H]þ m=z 820 (100%).
Following the same procedure, heating solutions of
[Fe3(CO)12 ꢀ n(CNC6H4OMe-p)n], n ¼ 1 or 2, allowed
isolation of the corresponding compounds, respec-
tively:
2.1.8. [Fe3(CO)11(TosMIC)]
Rf ¼ 0:78 (green). Found: C, 35.94; H, 1.35; N,
2.16%. C20H9Fe3NSO13 requires C, 35.77; H, 1.34; N,
2.10%. IR (CHCl3): m(CN) 2173 (m, br); m(CO) 2082
(sh), 2062 (s), 2047 (sh), 2003 (sh), 1974 (s) cmꢀ1. ESMS
(+ve ion): [M + H]þ m=z 672 (100%), [2M + H]þ m=z
1344 (10%). ESMS ()ve ion,): [M ) H]ꢀ m=z 670 (20%),
[M ) CO ) H]ꢀ m=z 642 (100%).
2.2.2. [Fe3(CO)9(l3-g2-CNC6H4OMe-p)]
Rf ¼ 0:73 (brown). IR (CHCl3): m(CO) 2086 (m), 2059
(w), 2040 (vs), 2031 (vs), 2015 (s), 1995 (m), 1975 (w)
cmꢀ1. ESMS (MeOH/MeOꢀ, )ve ion): [M + MeO]ꢀ m=z
584 (100%).
2.1.9. [Fe3(CO)10(TosMIC)2]
Rf ¼ 0:63 (green). Found: C, 39.97; H, 2.13; N,
3.35%. C28H18Fe3N2S2O14 requires C, 40.10; H, 2.15; N,
3.34%. IR (CHCl3): m(CN) 2164 (m, br); m(CO) 2098
(sh), 2080 (s), 2062 (sh), 2042 (s), 2033 (sh), 2020 (sh)
cmꢀ1. ESMS (+ve ion): [M + H]þ m=z 839 (100%),
[2M + H]þ m=z 1677 (5%). ESMS ()ve ion): [M ) H]ꢀ
m=z 837 (57%), [M ) CO ) H]ꢀ m=z 809 (100%).
2.2.3.
Me-p)]
[Fe3(CO)8(l3-g2-CNC6H4OMe-p)(CNC6H4O
Rf ¼ 0:4 (red–brown). IR (CHCl3): m(CN) 2146 (m);
m(CO) 2064 (sh), 2058 (m), 2038 (w), 2019 (vs), 1980 (m)
cmꢀ1. ESMS (MeOH/MeOꢀ, )ve ion): [M + MeO]ꢀ m=z
689 (100%).

1694
C. Decker et al. / Journal of Organometallic Chemistry 689 (2004) 1691–1701
2.3. Reactions of [Ru3(CO)12] with isonitriles
2071 (vw), 2062 (w), 2049 (s), 2041 (vs), 2019 (w), 1999
(m), 1992 (m) cmꢀ1]. ESMS (+ve ion): [M + H]þ m=z 745
(100%), [M ) CO + H]þ m=z 719 (30%). ESMS (MeOH/
MeOꢀ, )ve ion): [M ) 2CO + MeO]ꢀ m=z 719 (100%),
[M ) 3CO + MeO]ꢀ m=z 690 (18%).
The general reaction was carried out as follows:
[Ru3(CO)12] (0.5 g, 0.78 mmol) was dissolved in toluene
(10 ml). The isonitrile (3.13 mmol) was added by syringe
or as a solid and the reaction was gently refluxed for
ꢁ30 min. Samples were withdrawn at appropriate time
intervals and the course of the reaction monitored by
TLC. The solvent was removed in vacuo, the reaction
products redissolved in a minimum amount of CH2Cl2,
and chromatographed on silica plates with a solvent
mixture of petroleum spirits/CH2Cl2 (CNPh derivatives
1:1, CNC6H4OMe-p derivatives 1:2). Yields of the iso-
lated complexes were not determined because only parts
of the crude reaction product were used at a time for
chromatography. The identity of the complexes
[Ru3(CO)10(CNPh)2] and [Ru3(CO)9(CNPh)3] were
confirmed by comparison with the IR frequencies of
previously reported analogues.
2.3.5. [Ru3(CO)10(CNC6H4OMe-p)2]
Rf ¼ 0:78 (yellow). IR (CHCl3): m(CN) 2159 (sh),
2142 (m); m(CO) 2067 (m), 2029 (s), 2004 (sh), 1993 (s),
1980 (sh) cmꢀ1; [Lit [9] m(CN) 2154 (w); m(CO) 2093 (m),
2066 (m), 2048 (s), 2040 (s), 2030 (vs), 2022 (s), 1997 (s),
1990 (s) cmꢀ1]. ESMS (+ve ion): [M + H]þ m=z 851
(100%), [M ) CO + H]þ m=z 822 (5%). ESMS (MeOH/
MeOꢀ, )ve ion): [M + MeO]ꢀ m=z 881 (28%),
[M ) CO + MeO]ꢀ m=z 853 (12%, [M ) 2CO + MeO]ꢀ
m=z 822 (100%).
2.4. Synthesis of [Ru4(CO)11(CNPh)3]
[Ru3(CO)12] (0.5 g, 0.78 mmol) was dissolved in tol-
uene (10 ml). CNPh (0.32 ml, 3.13 mmol) was added by
syringe and the reaction was gently refluxed for ꢁ30
min. The course of the reaction monitored by TLC. The
solvent was removed in vacuo, the reaction products
redissolved in a minimum amount of CH2Cl2, and
chromatographed on silica plates with petroleum spirits/
CH2Cl2 (1:1). [Ru4(CO)11(CNPh)3] was obtained from
the fifth fraction (Rf ¼ 0:65, red) after [Ru3(CO)12],
[Ru3(CO)11(CNPh)], [Ru3(CO)10(CNPh)2] and [Ru3
(CO)9(CNPh)3]. This fraction was a mixture of two
compounds (as two close bands that could not be sep-
arated by chromatography) and so no elemental analysis
was obtained on the solid. IR (CHCl3): m(CN) 2166 (m);
m(CO) 2076 (m), 2050 (s), 2033 (s), 2006 (m), 1985 (sh)
cmꢀ1. ESMS (+ve ion): [M + H]þ m=z 1024 (100%).
ESMS (MeOH/MeOꢀ, )ve ion): [M +MeO]ꢀ m=z 1053
(100%) {also observed: [Ru4(CO)12(CNPh)2 + MeO]ꢀ
m=z 976 (33%), [Ru4(CO)12 (CNPh)2 ) CO + MeO]ꢀ m=z
946 (8%), [Ru4(CO)12(CNPh)2 ) 2CO + MeO]ꢀ m=z 919
(35%)}.
2.3.1. [Ru3(CO)11(CNPh)]
Rf ¼ 0:89 (yellow). IR (CHCl3): m(CN) 2158 (m);
m(CO) 2091 (m), 2049 (s), 2041 (s), 2021 (sh), 2010 (s),
1998 (sh) cmꢀ1; [Lit [16] m(CO) 2080 (m), 2035 (sh), 2020
(s), 2005 (sh), 1992 (s), 1982 (s), 1975 (sh), 1958 (m),
1948 (m), 1843 (m), 1805 (m), 1790 (m) cmꢀ1]. 1H NMR
(CDCl3): d 7.39–7.28 (5H, m, Ph). ESMS (+ve ion):
[M + H]þ m=z 716 (100%). ESMS (MeOH/MeOꢀ, )ve
ion): [M ) 2CO + MeO]ꢀ m=z 690 (100%), [M ) 3CO +
MeO]ꢀ m=z 661 (31%).
2.3.2. [Ru3(CO)10(CNPh)2]
Rf ¼ 0:79 (orange). IR (CHCl3): m(CN) 2157 (sh),
2137 (m); m(CO) 2066 (m), 2030 (s), 2005 (sh), 1994 (s),
1986 (sh) cmꢀ1; [Lit [9] for CNBut analogue m(CO) 2065
(w), 2020 (s), 2007(m), 1996 (s), 1990 (m), 1986 (m)
cmꢀ1].
ESMS (+ve ion): [M + H]þ m=z 791 (100%),
[M ) CO + H]þ m=z 762 (7%). ESMS (MeOH/MeOꢀ,
)ve ion): [M + MeO]ꢀ m=z 821 (50%), [M ) CO + MeO]ꢀ
m=z 792 (8%), [M ) 2CO + MeO]ꢀ m=z 763 (100%),
[M ) 3CO + MeO]ꢀ m=z 734 (8%).
2.5. Synthesis of [Ru4(CO)11(CNC6H4OMe-p)3]
2.3.3. [Ru3(CO)9(CNPh)3]
This reaction was carried out in the same manner as
for the CNPh analogue and the product fraction had
Rf ¼ 0:53 (red). This fraction was again a mixture of two
compounds, which could not be separated by chroma-
tography, and so no elemental analysis could be ob-
tained. IR (CHCl3): (CN) 2165 (m); m(CO) 2075 (m),
2062 (sh), 2058 (sh), 2049 (s), 2032 (vs), 2004 (m), 1981
(w) cmꢀ1. ESMS (MeCN/H2O, +ve ion): [M + H]þ m=z
1114, (100%), [M ) CO + H]þ m=z 1085, (17%),
[M ) CO + MeCN + H]þ m=z 1128, (42%), [M +
MeCN + H]þ m=z 1155, (28%).
Rf ¼ 0:73 (red). IR (CHCl3): m(CN) 2162 (m), 2140
(m), 2123 (sh); m(CO) 2062 (m), 2030 (s), 1997 (m), 1978
(sh) cmꢀ1; [Lit [9] for CNBut analogue: m(CO) 2040 (m),
1998 (s), 1971 (s) cmꢀ1]. ESMS (+ve ion): [M + H]þ m=z
866 (100%), [M ) CO + H]þ m=z 838 (5%). ESMS
(MeOH/MeOꢀ, )ve ion): [M + MeO]ꢀ m=z 896 (100%),
[M ) 2CO + MeO]ꢀ m=z 838 (20%).
2.3.4. [Ru3(CO)11(CNC6H4OMe-p)]
Rf ¼ 0:95 (yellow). IR (CHCl3): m(CN) 2161 (m);
m(CO) 2092 (m), 2048 (s), 2041 (s), 2020 (sh), 2009 (s),
1996 (sh) cmꢀ1; [Lit [9] m(CN) 2155 (w); m(CO) 2092 (w),
ESMS (MeOH/MeOꢀ, )ve ion): [M + MeO]ꢀ m=z
1143, (100%).

C. Decker et al. / Journal of Organometallic Chemistry 689 (2004) 1691–1701
1695
Table 1
Crystal data and refinement details for [Fe3(CO)10(CNPh)2] at 293 and 150 K, and for [Ru4(CO)11(CNPh)3]
[Fe3(CO)10(CNPh)2] at 293 K
[Fe3(CO)10(CNPh)2] at 150 K
[Ru4(CO)11(CNPh)3]
Empirical formula
Formula mass
Crystal system
Space group
C24H10Fe3N2 O10
653.89
C24H10Fe3N2O10
653.89
C32H15N3O11Ru4
1021.75
Monoclinic
P21=c
9.4699(1)
19.6533(1)
18.5984(2)
102.666(1)
3377.20(5)
150(2)
Monoclinic
P21=c
Monoclinic
P21=c
ꢀ
a (A)
15.551(1)
14.068(1)
11.975(1)
100.81(1)
2573.0(3)
293(2)
15.229(1)
13.913(1)
11.829(2)
99.80(1)
2469.7(1)
150(2)
ꢀ
b (A)
ꢀ
c (A)
b (°)
3
ꢀ
Volume (A )
Temperature (K)
Z
4
1.688
1.73
4
1.759
1.80
4
2.010
1.82
Density (g cmꢀ3
)
Absorption coefficient (mmꢀ1
)
Tmax;min
0.875, 0.627
13,634
4527
0.894, 0.711
12,815
4376
0.897, 0.712
18,284
6392
Total reflections
Unique reflections
Rint
0.0357
0.0381
0.0823
0.066
0.0551
0.1376
0.023
0.0232
0.0488
R1ðI > 2rðIÞÞ
wR2 (all data)
2.6. X-ray crystal structure determinations
more important than relative abundance. For the reac-
tions with CNPh it was surprising that ESMS-detectable
ions were found at all, since the site for chemical ioni-
sation through protonation is not obvious. Possibly with
increasing numbers of isonitrile ligands the basicity of
the CO ligands increases (especially the l-CO ones) to
the point where protonation is favoured. An alternative
site for ionisation by Hþ is the M–M bonds, as sug-
gested for [Ru3(CO)9(PPh3)3] which also gives strong
[M + H]þ ions in the ESMS whereas the parent
[Ru3(CO)12] does not [15].
Data were obtained on a Siemens SMART CCD
diffractometer operating under standard conditions.
Data were corrected for absorption and other effects
using an empirical method (SADABS [17]) and the
structures were solved and refined on Fo2 using the
SHELX 97 programs [18] manipulated under WinGx
[19]. Hydrogen atoms were included in calculated posi-
tions. Details are given in Table 1.
Chromatography of the reaction mixture allowed
separation and full characterisation of the first two
members [Fe3(CO)12 ꢀ n(CNPh)n] for n ¼ 1, 2, and
spectroscopic characterisation for the next two with
n ¼ 3, 4. However, the higher derivatives with n ¼ 5 or 6
were too fragile for isolation, so rely solely on the ESMS
results from the crude reaction mixture for evidence of
their existence. These results are consistent with previ-
ous work with CNBut where [Fe3(CO)12 ꢀ n(CNBu)n] for
n ¼ 1–3 were well-characterised, but the n ¼ 4 example,
the highest-substituted [Fe3(CO)12] derivative previously
detected, was too unstable even for 13C NMR studies
[4,6,7]. It is noteworthy that no [M3(CO)9(CNR)3]
compound (M ¼ Fe, Ru) has yet yielded crystals suit-
able for X-ray analysis.
The equivalent reaction between [Fe3(CO)12] and
CNR* (R* ¼ C6H4OMe-p) was more complicated. An
initial mass spectrum of the reaction mixture showed not
only peaks arising from substitution products [Fe3
(CO)12 ꢀ n(CNRꢂ)n] for n ¼ 3–5, but also species resulting
from cluster fragmentation [Fe2(CO)9 ꢀ n(CNRꢂ)n] (n ¼
4,5,6) and [Fe(CO)5 ꢀ n(CNRꢂ)n] (n ¼ 2–5). The differ-
ence between this system and the CNPh one may be that
3. Results and discussion
3.1. Reactions of isonitriles with [Fe3(CO)12]
A solution of [Fe3(CO)12] with four equivalents of
CNPh was stirred at room temperature for 2 h. A
sample was extracted, diluted with MeOH and examined
by ESMS. This showed a series of major ions that could
be readily assigned to the [M + H]þ species derived from
the substituted cluster [Fe3(CO)12 ꢀ n(CNPh)n] for
n ¼ 3–5. There was also a weak peak at m=z 955 which
corresponds to [Fe3(CO)6(CNPh)6]. There were no sig-
nals in the mixture for the mono- or di-substituted
clusters, but this was probably because of poor ability to
ionise under the conditions of the experiment, since they
were subsequently shown to be present by chromatog-
raphy. ESMS signals could be obtained from these
lower-substituted derivatives in their pure, isolated form
but only with concentrated solutions. The ESMS data
are therefore useful for indicating which species are
present in solution, but cannot be interpreted quantita-
tively since the relative ease of chemical ionisation is

1696
C. Decker et al. / Journal of Organometallic Chemistry 689 (2004) 1691–1701
CNR* has an –OMe group which can be protonated in
the di-iron and mono-iron compounds, so that equiva-
lent compounds may have been invisible in the ESMS in
the CNPh experiment. Chromatography of the mixture
yielded [Fe3(CO)12 ꢀ n(CNRꢂÞn] for n ¼ 1–3. These were
characterised by their [M + H]þ ions in the ESMS
spectra, and by comparison of their m(CO) spectra with
analogous compounds since analytically pure crystalline
samples could not be obtained.
enough to be isolated by chromatography; (iii) some
fragmentation to give mono- and di-iron compounds
occurs and (iv) TosMIC reacts in a more complicated
fashion than does CNPh, giving rise to a wider range of
products.
3.2. X-ray crystal structure determinations of
[Fe3(CO)10(CNPh)2]
Equivalent reactions using TosMIC provided differ-
ent results again. The ESMS spectrum of the crude re-
action mixture is shown in Fig. 1. There is a clear family
of peaks derived from [Fe3(CO)12 ꢀ n(TosMIC)n] for
n ¼ 1–5 in this case. There is also another series of peaks
which appear to have arisen from addition of a TosMIC
ligand, with concomitant oxidation to give
[Fe3(CO)12(TosMIC)]þ, and substituted derivatives
thereof. This unusual behaviour was not found with any
of the other isonitriles, and chromatography allowed
isolation of only [Fe3(CO)12 ꢀ n(TosMIC)n] for n ¼ 1–3
so no further information could be obtained concerning
these species.
The positive-ion ESMS for the TosMIC derivatives
showed [M + H]þ peaks arising from protonation as
expected. Less predictably, the negative-ion spectra
showed clean peaks assignable to [M ) H]ꢀ ions formed
in situ. Presumably the CH2 protons of the ligand are
rendered acidic enough by the adjacent NC and SO2
groups to facilitate chemical ionisation in the mass
spectrometer by proton removal. This suggests that
TosMIC is a versatile ligand for investigation of reac-
tions by ESMS.
Crystal structures of the mono- and di-substituted
[Fe3(CO)12] with CNBut have been available since 1982
and 1990, respectively [4,6]. For both, the basic
[Fe3(CO)12] structure is preserved. In [Fe3(CO)11
(CNBut)], the isonitrile ligand occupies an axial site on
the non-bridged iron atom, while in [Fe3(CO)12
(CNBut)2], both isonitrile ligands were found to be on
the unique iron atom, one in an axial and the other in an
equatorial site, as in 1. The only other crystal structure
of the type [Fe3(CO)12 ꢀ n(CNR)n] is for n ¼ 1 and R ¼
CF3, where CNCF3 occupies one of the two bridging
positions of an [Fe3(CO)12] structure [20]. There is no
disorder of the metal triangle for any of these deriva-
tives.
To provide another example, the structure of
[Fe3(CO)10(CNPh)2] was determined at )123 °C, using
crystals obtained from a concentrated CH2Cl2 solution
that was slowly cooled to )20 °C. The full structure is
shown in Fig. 2(a), and selected bond parameters are
provided in the caption to the figure.
In the solid state, [Fe3(CO)10(CNPh)2] preserves the
basic [Fe3(CO)12] geometry. The two CNPh ligands have
replaced the two axial CO ligands on the unique iron
atom as in 2. This contrasts with the only other struc-
turally characterised complex of this type, [Fe3(CO)10
(CNBut)2], where the two isonitrile ligands (although
both coordinated to the unique iron atom) occupy one
axial and one equatorial position (1) [6]. In terms of
electronic preferences, axial positions might be favoured
by CNPh because the isonitrile is a weaker p-acceptor
base than CO. In order to maximise p-bonding to the
In summary, the use of ESMS to monitor reaction
mixtures of [Fe3(CO)12] and isonitriles provides useful
information that is not available by other means. It
shows clearly that: (i) the reactions give mixtures of
substituted derivatives rather than single products based
on the stoichiometry of the reaction; (ii) compounds can
be detected with up to six CNR ligands incorporated,
even although only those with up to three are stable
Fig. 1. The positive-ion ESMS of the crude reaction mixture of [Fe3(CO)12] and TosMIC, recorded in MeOH. The pairs of peaks correspond to
[Fe3(CO)n(TosMIC)m + H]þ and [Fe3(CO)n þ 1(TosMIC)m]þ for (a) n ¼ 11, m ¼ 1; (b) n ¼ 10, m ¼ 2; (c) n ¼ 9, m ¼ 3; (d) n ¼ 8, m ¼ 4; (e) n ¼ 7,
m ¼ 5.

C. Decker et al. / Journal of Organometallic Chemistry 689 (2004) 1691–1701
1697
Fig. 2. (a) The structure of [Fe3(CO)10(CNPh)2]. (b) The 5% disorder of the Fe3 triangle found in the determination carried out at 150 K. Selected
ꢀ
bond lengths (A): Fe(1)–Fe(2) 2.6921(11), Fe(1)–Fe(3) 2.6918(11), Fe(2)–Fe(3) 2.5594(11), Fe(1)–C(1) 1.871(6), Fe(1)–C(2) 1.853(6).
remaining CO ligands, they prefer to be trans to Fe–Fe
bonds, which is only accomplished if both CNPh occupy
axial sites. However the different structures adopted by
the examples [Fe3(CO)10(CNR)2] (R ¼ Bu or Ph) sug-
gest that these preferences are not stronger than crystal
packing effects.
Generally, the effects of the isonitrile ligands of
[Fe3(CO)10(CNPh)2] on the rest of the molecule are
ꢀ
small. The Fe–Fe bond lengths (2.5594(1) A bridged,
order is eliminated. Furthermore, for this disorder a 60°
rotation of the iron triangle is required which leads to a
structure where the axial isonitrile ligands become
bridging, which is normally preferred by the CO rather
than the isonitrile ligands (an exception is CNCF3 [20]).
As noted earlier, a small amount of metal site disorder
has recently been observed for the mixed-metal species
[FeRu2(CO)11(CNBut)] and [FeRu2(CO)10(CNBut)2],
but not for [Fe2Ru(CO)11(CNBut)] and [Fe2Ru-
(CO)10(CNBut)2] [12].
The X-ray structure analysis of [Fe3(CO)10(CNPh)2]
was repeated at room temperature to see if the extent of
disorder varied with temperature, since previous work
with [Ru3(CO)11(CNBut)] showed more disorder at
higher temperatures, proving it to be dynamic [10,11].
The space group and unit cell were identical at 293 K to
those found at low temperature. There were also no
significant geometric differences in the structures be-
tween the two temperatures. A reported variable
temperature X-ray structure analysis of [Fe2Ru(CO)10-
(CNBut)2] indicated that the Fe–Fe bond was longer
and the C–C bond distances in the But group were
shorter at higher temperature, which was attributed to
librational effects [12]. An equivalent effect was not ob-
served for [Fe3(CO)10(CNPh)2].
What was completely unexpected was that the room
temperature analysis showed no metal framework dis-
order. Thermodynamically it is not sensible that there
would be less disorder at higher temperatures for a
dynamic process. The crystals for the two determina-
tions were from two different batches so that a struc-
ture with disorder might have been frozen out by
crystal packing interactions, or some twinning of the
crystals might have occurred, in the first batch. Another
possible explanation is that some [Fe3(CO)12] had
co-crystallised with [Fe3(CO)10(CNPh)2] in the low
ꢀ
2.6921(1) and 2.6918(1) A unbridged) are not signifi-
cantly different from those observed in [Fe3(CO)12]
ꢀ
ꢀ
{2.551(2) A bridged, 2.677(2) and 2.684(2) A unbridged,
recorded at 160 K [5]}. Because the CNPh ligands oc-
cupy both axial positions on Fe(1), the Fe(1)–Fe(2) and
Fe(1)–Fe(3) bond lengths are equivalent. It is notewor-
thy that the same observation was made for
[Fe3(CO)10(CNBut)2], although the Fe(1)–Fe(2) and
Fe(1)–Fe(3) bonds are rendered inequivalent by the
equatorial isonitrile ligand. This was considered to be a
result of the close similarity in bonding interactions for
CO and CNR.
The axial and equatorial Fe–C bond lengths to CO
ligands deviate only marginally in [Fe3(CO)10(CNPh)2].
The same is true for all other reported isonitrile-substi-
tuted derivatives. The l2-CO ligands, unlike in
[Fe3(CO)12], are not significantly asymmetric. The CBN
distances are equivalent for the two ligands [1.143(6)
ꢀ
and 1.155(7) A], as are the C–N–C angles [176.1(6)° and
176.3(1)°].
Towards the end of the refinement, three major re-
sidual peaks formed a triangle rotated 60° from the
major Fe3-triangle. This was clearly a Star-of-David
disorder in the location of the iron atoms, and refine-
ment gave 5% for the second compound, as shown in
Fig. 2(b). This was surprising since in all other reported
isonitrile-substituted derivatives of [Fe3(CO)12] the dis-

1698
C. Decker et al. / Journal of Organometallic Chemistry 689 (2004) 1691–1701
temperature batch. Unfortunately, the crystal used for
the room temperature data decomposed before a low-
temperature data set on the same crystal could be
collected. Examination of a series of [Fe3(CO)10-
(CNPh)2] crystals would be needed to understand this
disorder process. What is clear is that in general study
of temperature-dependent disorder should be carried
out on one single crystal to avoid possible variations
from crystal to crystal.
Evidence for substitution with isonitriles higher than
twofold is still limited. One report briefly mentions the
synthesis of the complexes [Ru3(CO)12 ꢀ n(CNBut)n]
(n ¼ 1–4), but without characterisation data [23]. Bruce
and co-workers have isolated two tri-substituted com-
plexes but no satisfactory microanalytical data could be
obtained [9]. We therefore re-investigated the reactions
using ESMS to monitor progress.
When [Ru3(CO)12] was gently heated in toluene for 30
min with four equivalents of the isonitrile ligand (CNPh
or CNC6H4OMe-p) and the crude reaction mixtures were
injected into the ESMS, none of the expected signals such
as [M + H]þ ions of [Ru3(CO)11(CNC6H4OMe-p)] or
[Ru3(CO)10(CNC6H4OMe-p)2] were observed in positive
ion mode. However, in the negative-ion spectrum these
two compounds gave strong signals associated with their
[M + MeO]ꢀ ions. In addition, a weak signal for
[Ru3(CO)9(CNC6H4OMe-p)3 + MeO]ꢀ at m=z 987 was
observed. A number of signals associated with higher-
nuclearity clusters were also detected. The most intense
peaks had a mass difference of 105 amu, reflecting sub-
stitution of CO by CNC6H4OMe-p. From the available
information, a cluster series of the type [Ru4(CO)12 ꢀ n
(CNC6H4OMe-p)n] (n ¼ 2–4) was tentatively assigned.
Very similar results were obtained for the CNPh de-
rivatives. The most intense ESMS ions were MeOꢀ ad-
ducts derived from [Ru3(CO)11(CNPh)] and [Ru3(CO)10
(CNPh)2]. An equivalent higher-nuclearity series of the
type [Ru4(CO)14 ꢀ n(CNPh)n] (n ¼ 2–4) was also ob-
served.
3.3. Thermolysis of [Fe3(CO)12 ꢀ n(CNPh)n] (n ¼ 1,2)
Previously it has been reported [4] that [Fe3(CO)11
(CNBut)] undergoes smooth thermolysis by loss of two
CO ligands and conversion of the isonitrile to a l3-g2
bonding mode, as in 3a. This is still a rare type of iso-
nitrile linkage, the only other example being a niobium
cluster, [Nb3Cl8(CNBut)4(l3-g2-CNBut)] [21]. The same
reaction was therefore carried out with [Fe3(CO)11
(CNPh)]. Heating to 75 °C converted the green solution
to a brown one which ESMS indicated contained 3b,
from a clean [M + OMe]ꢀ peak. Chromatography al-
lowed isolation of 3b. The ESMS spectrum of the pure
compound in positive-ion mode showed no evidence for
formation of a [M + H]þ ion, but the negative ion re-
sulted in a strong [M + OMe]ꢀ one. This is not surpris-
ing since earlier studies on 3a showed it was susceptible
to nucleophilic but not electrophilic attack at the l3-C
atom [4].
A completely analogous reaction occurred with
[Fe3(CO)10(CNPh)2] to give a substituted version of 3b
with one l3-g2-CNPh and one terminal one. The com-
plexes [Fe3(CO)12 ꢀ n(CNC6H4OMe-p)n] (n ¼ 1 or 2)
behaved similarly.
After separation by chromatography, the products
were again analysed by ESMS. Both [Ru3(CO)11
(CNR)] and [Ru3(CO)10(CNR)2] ionised by methoxide
addition as well as by protonation, even though
[M + H]þ ions had not been observed in the crude re-
action solution. The slower-moving bands were of more
interest as they represented higher-substituted and/or
3.4. Reactions of isonitriles with [Ru3(CO)12]
higher-nuclearity
products.
In
the
case
of
In previous work, reaction of [Ru3(CO)12] with
CNBut gave [Ru3(CO)11(CNBut)] and [Ru3(CO)10
(CNBut)2], where substitution was shown to occur
progressively on different metal atoms on axial sites and
as terminal ligands [9,10]. This differs from phosphine
derivatives of [Ru3(CO)12], which were found to prefer
equatorial sites [22]. The structure of [Ru3(CO)12] is not
disordered, but many of its derivatives show disorder of
the metal framework, e.g. [Ru3(CO)10{P(OMe)3}2],
[Ru3(CO)11(CNBut)] and [Ru3(CO)10(PMe3)2], but not
[Ru3(CO)11{P(C6H11)3}]. This disorder has been ra-
tionalised in terms of a model in which the Ru3-triangle
occupies two symmetry-related positions while the pe-
ripheral atom polyhedron (i.e. the O of the CO ligand
and the P of the phosphine or phosphite ligand) remains
unchanged [22]. For [Ru3(CO)11(CNBut)] and [Ru3
(CO)11(PMe3)], it has also been shown that the disorder
is dynamic in origin [11].
CNC6H4OMe-p, they were however very close together
and could not be separated satisfactorily. Together they
gave a signal associated with [Ru3(CO)9(CNC6H4O
Me-4)3 + H]þ as well as from one or more higher-nu-
clearity clusters such as [Ru4(CO)11(CNC6H4OMe-
p)3 + H]þ, while [Ru4(CO)11(CNC6H4OMe-p)3 + MeO]ꢀ
was the only signal observed in negative-ion mode for
the same sample.
For the system involving CNPh, a better separation
of the bands occurred. Thus, the bands following the
di-substituted complex were isolated and injected into
the mass spectrometer. An intense peak corresponding
to [Ru3(CO)9(CNPh)3 + H]þ was observed in positive-
ion mode, while in the negative-ion mode the corre-
sponding [Ru3(CO)9(CNPh)3 + MeO]ꢀ ion was also
clean.
Isolation of the next-slowest band (which appeared to
be a mixture of two species which did not separate) in

C. Decker et al. / Journal of Organometallic Chemistry 689 (2004) 1691–1701
1699
ꢀ
Ru(1)–Ru(3) bond lengths [2.8341(3) A and 2.8226(3)
ꢀ
A, respectively] are identical. An equivalent bond
ꢀ
length was noted for Ru(1)–Ru(4) [2.8258(3) A] and is
marginally shorter than the unbridged Ru(2)–Ru(3)
the [Ru3(CO)12]/CNPh system, gave a single signal at
m=z 1024 in positive-ion, and a number of ions with the
most intense at m=z 1054 in negative-ion mode. Pre-
suming the ions to be [M + H]þ and [M + MeO]ꢀ re-
ꢀ
bond [2.8895(3) A]. As a consequence, the bond angles
spectively, the presence of
a
cluster of formula
[Ru4(CO)11(CNPh)3] was indicated. Crystals of
[Ru4(CO)11(CNPh)3] (4) were obtained for full charac-
terisation. The other component in the mixture is
probably [Ru4(CO)12(CNPh)2] from the ESMS data of
the mixture, but could not be confirmed.
within the Ru-triangle are not quite 60°. The Ru(4)–
Ru(1) vector makes an angle of 107° with the basal
triangle.
The bonding mode of the two bridging isonitrile li-
gands is of particular interest. The carbon atoms [C(61)
and C(71)] each bridge two metal atoms while the ni-
trogen atoms [N(61) and N(71)] form a single bond to
the unique ruthenium atom Ru(4). This results in a
double rather than triple bond between the C and N
3.5. X-ray crystal structure determination of
[Ru4(CO)11(CNPh)3] (4)
ꢀ
Crystals of the cluster [Ru4(CO)11(CNPh)3] were
obtained from a concentrated CH2Cl2/Et2O solution by
cooling to )20 °C. The structure is shown in Fig. 3 and
selected bond parameters are included in the caption.
The core geometry of [Ru4(CO)11(CNPh)3] is that of a
spiked triangle, a much rarer skeleton for M4 clusters
than tetrahedral or butterfly arrangements. The metal
atoms of the triangle each have three terminal CO li-
gands while the unique ruthenium atom [Ru(4)] is at-
tached to two terminal CO groups and one terminal
isonitrile ligand. The remaining two CNPh ligands each
bridge a metal–metal bond of the triangle through the
isonitrile carbon atom. They also bridge the spike-Ru
with a l2-CN bond, as illustrated in Fig. 3. Because of
the nature of the bridging ligands, the Ru(1)–Ru(2) and
atoms. The C–N bond lengths [average 1.272(3) A] are
consistent with this, being longer than in the terminal
ꢀ
isonitrile ligand [1.159(3) A]. The bonding mode ob-
served can be equated with an imine-type arrangement.
The C–N–C [126.2(2)°] and C(71)–N(71)–C(72)
[123.6(2)°] bond angles [average 124.9(2)°] of the
bridging isonitrile ligands also reflect their similarity to
imines (ꢃ120°) rather than that of isonitriles (ꢃ180°).
The terminal isonitrile ligand has an expected C–N–C
angle of 177.5(3)°.
The formal electron count averages 18 for each
metal atom; 19 for Ru(1) and Ru(4) and 17 for Ru(2)
and Ru(3) atoms. This is calculated assuming the
bridging isonitrile ligands each donate one electron to
each metal from the carbon atom and two electrons to
the unique Ru(4) from the nitrogen atoms. Thus,
Ru(1) and Ru(4) are formally electron-rich (19e)
whereas Ru(2) and Ru(3) atoms are electron-deficient
(17e). This imbalance is partly offset by the l2-C at-
oms lying closer to Ru(2) or Ru(3) [average 1.964(3)
ꢀ
ꢀ
A] than to Ru(1) [2.149(3) A]. The overall cluster-
valence-electron count is 64, as expected for a Ru4
cluster with four M–M bonds.
[Ru4(CO)11(CNPh)3] is the only reported Ru4-isoni-
trile structure to date. The observed l3-g2 imine-type
bonding mode for isonitrile ligands has been reported
before, but only for a small number of other systems.
They include Rh3, Ru3 and Os5 clusters [24].
4. Conclusions
With the ligands used in this study, it has been shown
that higher-substituted products do exist for both
[Fe3(CO)12] (up to sixfold) and [Ru3(CO)12] (up to
threefold). Even though they become less stable with the
degree of substitution, they can still be studied by
ESMS. Reactions always seem to give mixtures of
products which complicates full characterisation when
chromatography leads to decomposition.
ꢀ
Fig. 3. The structure of [Ru4(CO)11(CNPh)3]. Bond lengths (A) in-
clude: Ru(1)–Ru(2) 2.8341(3), Ru(1)–Ru(3) 2.8226(3), Ru(2)–Ru(3)
2.8895(3), Ru(1)–Ru(4) 2.8258(3), Ru(1)–C(61) 2.172(3), Ru(2)–C(61)
1.962(3), Ru(1)–C(71) 2.127(3), Ru(3)–C(71) 1.966(3), Ru(4)–C(51)
1.994(3), C(51)–N(51) 1.159(3), C(61)–N(61) 1.274(3), C(71)–N(71)
1.270(3).
The structure of [Fe3(CO)10(CNPh)2] has been de-
termined and differs from the only other structurally

1700
C. Decker et al. / Journal of Organometallic Chemistry 689 (2004) 1691–1701
characterised di-substituted isonitrile derivative of
[Fe3(CO)12] with respect to the orientations of the iso-
nitrile ligands. The metal framework disorder observed
at low temperature is not dynamic in origin, because a
separate structure determination at room temperature
did not show any disorder.
Pyrolysis reactions of isonitrile derivatives of
[Fe3(CO)12] led to new analogues of the previously ob-
served species [Fe3(CO)9 ꢀ n(l3-CNR)(CNR)m] (m ¼ 0, 1)
as the only major products, extending R ¼ But to R ¼
aryl examples. They ionise by formation of [M + MeO]ꢀ
ions and can be detected readily in the negative-ion mode.
The analogous pyrolysis reaction of derivatives of
[Ru3(CO)12] led to a series of products [Ru4(CO)14 ꢀ n
(CNR)n] (n ¼ 2–4) as the only major higher-nuclearity
product. [Ru4(CO)11(CNPh)3] represents the only fully
characterised isonitrile-substituted ruthenium cluster
with four metal atoms.
References
[1] R.E. Benfield, B.F.G. Johnson, in: B.F.G. Johnson (Ed.), Tran-
sition Metal Clusters, Wiley, London, 1980;
B.F.G. Johnson, R.E. Benfield, Topics Stereochem. 12 (1981) 253.
[2] B.F.G. Johnson, S. Tay, Inorg. Chim. Acta 332 (2002) 201;
B.F.G. Johnson, J. Chem. Soc., Dalton Trans. (1997) 1473;
B.E. Mann, J. Chem. Soc., Dalton Trans. (1997) 1457;
L.J. Farrugia, J. Chem. Soc., Dalton Trans. (1997) 1783;
B.E. Hanson, E.C. Lisic, J.T. Petty, G.A. Iannaconne, Inorg.
Chem. 25 (1986) 4062;
D. Lentz, R. Marshall, Organometallics 10 (1991) 1487;
D. Braga, F. Grepioni, L.J. Farrugia, B.F.G. Johnson, J. Chem.
Soc., Dalton Trans. (1994) 2911;
L.J. Farrugia, A.L. Gillon, D. Braga, F. Grepioni, Organomet-
allics 18 (1999) 5022.
[3] S. Grant, J. Newman, A.R. Manning, J. Organometal. Chem. 96
(1975) C11;
M.O. Albers, N.J. Coville, T.V. Ashworth, E. Singleton, H.E.
Swanepoel, J. Chem. Soc., Chem. Commun. (1980) 489.
[4] M.I. Bruce, T.W. Hambley, B.K. Nicholson, J. Chem. Soc.,
Chem. Commun. (1982) 353;
M.I. Bruce, T.W. Hambley, B.K. Nicholson, J. Chem. Soc.,
Dalton Trans. (1983) 2385.
5. Supplementary material
[5] C.H. Wei, L.F. Dahl, J. Am. Chem. Soc. 88 (1966) 1821;
C.H. Wei, L.F. Dahl, J. Am. Chem. Soc. 91 (1969) 1351;
F.A. Cotton, J.M. Troup, J. Am. Chem. Soc. 96 (1974) 3070;
D. Braga, F. Grepioni, L.J. Farrugia, B.F.G. Johnson, J. Chem.
Soc., Dalton Trans. (1994) 2911;
Full crystallographic data have been deposited with
the Cambridge Crystallographic Data Centre, CCDC
nos 231171 (4), 231172 (2 at 293 K) and 231173 (2 at
150 K).
See also H. Chevreau, C. Martinsky, A. Sevin, C. Minot, B. Silvi,
New J. Chem. 27 (2003) 1049.
[6] J.B. Murray, B.K. Nicholson, A.J. Whitton, J. Organomet. Chem.
385 (1990) 91.
Bu
N
Ph
N
[7] X. Chen, B.E. Mann, Chem. Commun. (1997) 2233.
[8] E.R. Corey, L.F. Dahl, J. Am. Chem. Soc. 83 (1961) 2203;
R. Mason, A.I.M. Rae, J. Chem. Soc. A (1968) 778;
M.R. Churchill, F.J. Hollander, J.P. Hutchinson, Inorg. Chem. 16
(1977) 2655;
(CO)3
Fe
(CO)3
Fe
Bu
N
C
C
CO
CO
CO
CO
C
OC
OC
Fe
CO
Fe
OC
D. Braga, F. Grepioni, E. Tedesco, P.J. Dyson, C.M. Martin,
B.F.G. Johnson, Trans. Metal Chem. 20 (1995) 615.
[9] M.I. Bruce, J.G. Matison, R.C. Wallis, J.M. Patrick, B.W.
Skelton, A.H. White, J. Chem. Soc., Dalton Trans. (1983) 2365.
[10] M.I. Bruce, G.N. Pain, C.A. Hughes, J.M. Patrick, B.K. Skelton,
A.H. White, J. Organomet. Chem. 307 (1986) 343.
[11] L.J. Farrugia, C. Rosenhahn, S. Whitworth, J. Cluster Sci. 9
(1998) 505.
Fe
(CO)3
Fe
(CO)3
C
N
(1)
Ph
(2)
Ph
N
[12] L.J. Farrugia, P. Mertes, J. Cluster Sci. 13 (2002) 199.
[13] M.I. Bruce, G.N. Pain, C.A. Hughes, J.M. Patrick, B.W. Skelton,
A.H. White, J. Organometal. Chem. 307 (1986) 343.
[14] C. Decker, W. Henderson, B.K. Nicholson, J. Chem. Soc., Dalton
Trans. (1999) 3507.
R
C
N
(CO)
Ru
C
2
Ph
Ph
Fe(CO)3
Fe(CO)3
[15] W. Henderson, J.S. McIndoe, B.K. Nicholson, P.J. Dyson, J.
Chem. Soc., Dalton Trans. (1998) 519;
Fe
(CO)3
N
C
N
W. Henderson, B.K. Nicholson, L.J. McCaffrey, Polyhedron 17
(1998) 429.
Ru
C
(CO)3
[16] C.J. Cardin, D.J. Cardin, N.B. Kelly, G.A. Lawless, M.B. Power,
J. Organometal. Chem. 341 (1988) 447.
[17] R.H. Blessing, Acta Cryst. A 51 (1995) 33.
3a R = But
3b R = Ph
Ru
(CO)3
Ru
(CO)3
[18] G.M. Sheldrick, SHELX-97-Programs for X-ray Crystal Structure
Determination, University of Gottingen, 1997.
(4)
[19] WinGX L.J. Farrugia, J. Appl. Cryst. 32 (1999) 837. Available
from <http://www.chem.gla.ac.uk/~louis/wingx>.
Acknowledgements
€
[20] I. Brudgam, H. Hartl, D. Lentz, Z. Naturforsch. 39b (1984) 721.
[21] F.A. Cotton, W.J. Roth, J. Am. Chem. Soc. 105 (1983) 3734.
[22] M.I. Bruce, J.G. Matison, B.W. Skelton, A.H. White, J. Chem.
Soc., Dalton Trans. (1983) 2375.
We thank Associate Professor Cliff Rickard, Univer-
sity of Auckland, for collection of X-ray intensity data.

C. Decker et al. / Journal of Organometallic Chemistry 689 (2004) 1691–1701
1701
[23] J.-M. Bassett, D.E. Berry, G.K. Barker, M. Green, J.A.K.
Howard, F.G.A. Stone, J. Chem. Soc. Dalton Trans. (1979) 1003.
[24] A.L. Balch, L.A. Fossett, M.M. Olmstead, Organometallics 6
(1987) 1827;
H. Song, K. Lee, C.H. Lee, J.T. Park, H.Y. Chang, M.G. Choi,
Angew. Chem., Int. Ed. Eng. 40 (2001) 1500;
H. Song, K. Lee, M.G. Choi, J.T. Park, Organometallics 21 (2002)
1756;
O. bin Shawkataly, S.G. Teoh, H.K. Fun, J. Organometal. Chem.
464 (1994) C29;
H. Song, C.H. Lee, K. Lee, J.T. Park, Organometallics 21 (2002)
2514.