The series Os4(μx-η2-C 2Ph2)(CO)14-n (n = 0, 1, 2; x = n + 2) - Models for site-specific surface catalysts

Article abstract of DOI:10.1139/V05-239

The cluster Os₄(μ-η²-C₂Ph ₂)(CO)₁₄ (1) has been prepared from the reaction of Os₄(CO)₁₄ and C₂Ph₂ in CH ₂Cl₂ at 25°C. Other minor products include the known clusters Os₃(μ₃-η²-C₂Ph ₂)(CO)₁₀ and Os₃(μ-η⁴-C ₄Ph₄)(CO)₉. The structure of 1 reveals an approximately planar C₂Os₄ skeleton with a dimetallacyclobutene ring (C-C = 1.32(4) A) and a flat butterfly Os ₄ unit (Os-Os range = 2.859(2)-2.916(2) A). The ¹3C{¹H} NMR spectrum of 1 indicates the carbonyl ligands are rigid at room temperature. Stirring 1 in CH₂Cl₂ for 2 days (ambient temperature) afforded Os₄(μ₃- η²-C₂Ph₂)(CO)₁₃ (2). The Os atoms in 2 also have an almost flat butterfly arrangement (Os-Os range = 2.7392(7)-2.8947(6) A) with the alkyne ligand located over one of the Os₃ triangles. The ¹³C NMR data for 2 are consistent with rapid rotation on the NMR timescale of the hinge Os(CO)₃ units at 21°C, but slow rotation at -50°C. Heating 2 at 40°C gave Os ₄(μ₄-η²-C₂Ph ₂)(CO)₁₂ (3) after 2 days. Cluster 3 has the common butterfly arrangement of Os atoms with the C₂Ph₂ bound to all four metal atoms (Os-Os range = 2.7457(5)-2.8742(5) A). The ¹³C(¹H) NMR spectra of 3 at 21 and 90°C indicate there is rapid CO exchange of the carbonyls of the two types of Os(CO)₃ units, but not between the units. The spectrum at -90°C indicates one of the rotations (presumed to be that involving the carbonyls of the wingtip Os(CO)₃ units) is slowed on the NMR timescale. Compounds 1-3 form a unique series of clusters that have an alkyne ligand bound to two, three, and four metal atoms. Compound 1 is a model for a corner, compound 2 for a planar surface, and compound 3 a step site, in site-specific surface catalysts.

Full text of DOI:10.1139/V05-239

176
The series Os4(␮x-␩2-C2Ph2)(CO)14–n (n = 0, 1, 2; x =
n + 2) — Models for site-specific surface
catalysts1
John P. Canal, Michael C. Jennings, Glenn P.A. Yap, and Roland K. Pomeroy
2
Abstract: The cluster Os4(µ-η -C2Ph2)(CO)14 (1) has been prepared from the reaction of Os4(CO)14 and C2Ph2 in
2
4
CH2Cl2 at 25 °C. Other minor products include the known clusters Os3(µ3-η -C2Ph2)(CO)10 and Os3(µ-η -C4Ph4)(CO)9.
The structure of 1 reveals an approximately planar C2Os4 skeleton with a dimetallacyclobutene ring (C—C =
1.32(4) Å) and a flat butterfly Os4 unit (Os—Os range = 2.859(2)–2.916(2) Å). The 13C{1H} NMR spectrum of 1 indi-
cates the carbonyl ligands are rigid at room temperature. Stirring 1 in CH2Cl2 for 2 days (ambient temperature) af-
2
forded Os4(µ3-η -C2Ph2)(CO)13 (2). The Os atoms in 2 also have an almost flat butterfly arrangement (Os—Os range =
2.7392(7)–2.8947(6) Å) with the alkyne ligand located over one of the Os3 triangles. The 13C NMR data for 2 are con-
sistent with rapid rotation on the NMR timescale of the hinge Os(CO)3 units at 21 °C, but slow rotation at –50 °C.
2
Heating 2 at 40 °C gave Os4(µ4-η -C2Ph2)(CO)12 (3) after 2 days. Cluster 3 has the common butterfly arrangement of
Os atoms with the C2Ph2 bound to all four metal atoms (Os—Os range = 2.7457(5)–2.8742(5) Å). The 13C{1H} NMR
spectra of 3 at 21 and 90 °C indicate there is rapid CO exchange of the carbonyls of the two types of Os(CO)3 units,
but not between the units. The spectrum at –90 °C indicates one of the rotations (presumed to be that involving the
carbonyls of the wingtip Os(CO)3 units) is slowed on the NMR timescale. Compounds 1–3 form a unique series of
clusters that have an alkyne ligand bound to two, three, and four metal atoms. Compound 1 is a model for a corner,
compound 2 for a planar surface, and compound 3 a step site, in site-specific surface catalysts.
Key words: osmium cluster, diphenylacetylene, dimetallacyclobutene, carbonyl exchange, surface catalysis.
2
Résumé : On a préparé l’agrégat Os4(µ-η -C2Ph2)(CO)14 (1) par réaction du Os4(CO)14 avec du C2Ph2, dans le CH2Cl2,
2
4
à 25 °C; il était accompagné d’autres produits mineurs, dont les agrégats connus Os3(µ3-η -C2Ph2)(CO)10 et Os3(µ-η -
C4Ph4)(CO)9. La structure de 1 comporte un squelette C2Os4 approximativement plan avec un cycle dimétallacyclobu-
tène (C—C = 1,32(4) Å) et une unité de Os4 en forme de papillon applati (Os—Os = 2,859(2) à 2,916(2) Å). Les
spectres RMN 13C{1H} du composé 1 indiquent que les ligands carbonyles sont rigides à la température ambiante. Par
2
agitation pendant deux jours dans du CH2Cl2, à la température ambiante, le composé 1 conduit au Os4(µ3-η -
C2Ph2)(CO)13 (2). Les atomes d’osmium dans le composé 2 adoptent aussi un arrangement en forme de papillon prati-
quement plat (Os—Os = 2,7392(7) à 2,8947(6) Å) dans lequel le ligand alcyne se trouve au-dessus de l’un des trian-
gles Os3. Les données de la RMN du 13C pour le composé 2 sont en accord avec une rotation rapide, à l’échelle de
temps de la RMN et à 21 °C, des unités charnières Os(CO)3; la rotation est toutefois lente à –50 °C. Par chauffage à
2
40 °C pendant deux jours, le composé 2 conduit à la formation du Os4(µ4-η -C2Ph2)(CO)12 (3). Dans l’agrégat 3, les
atomes d’osmium adoptent l’arrangement habituel en forme de papillon et le C2Ph2 est lié aux quatre atomes métalli-
ques (Os—Os = 2,7457(5) à 2,8742(5) Å). Les spectres RMN 13C{1H} du composé 3 à 21 et 90 °C indiquent qu’il se
produit un échange rapide de CO des carbonyles des deux types d’unités Os(CO)3, mais qu’il n’y a pas d’échange
entre les unités. Le spectre à –90 °C indique qu’une des rotations (probablement celle impliquant les carbonyles des
unités Os(CO)3 au bout des ailes) est ralentie à l’échelle de temps de la RMN. Les composés 1, 2 et 3 forment une
série unique d’agrégats comportant un ligand alcyne lié à deux, trois et quatre atomes métalliques. Les composés 1, 2
et 3 sont respectivement un modèle pour un coin, pour une surface plane et pour un site étagé dans les catalyseurs de
surface à site spécifique.
Mots clés : agrégat d’osmium, diphénylacétylène, dimétallacyclobutène, échange de carbonyle, catalyseur de surface.
[Traduit par la Rédaction] Canal et al. 186
Received 15 June 2005. Published on the NRC Research Press Web site at http://canjchem.nrc.ca on 28 February 2006.
This paper is dedicated to Professor Arthur J. Carty for his contributions to Canadian chemistry.
J.P. Canal and R.K. Pomeroy.2 Department of Chemistry, Simon Fraser University, Burnaby, BC V5A 1S6, Canada.
M.C. Jennings. Department of Chemistry, University of Western Ontario, London, ON N6A 5B7, Canada.
G.P.A. Yap.3 Department of Chemistry, University of Ottawa, Ottawa, ON K1N 6N5, Canada.
1This article is part of a Special Issue dedicated to Professor Arthur Carty.
2Corresponding author (e-mail: pomeroy@sfu.ca).
3Present address: Department of Chemistry and Biochemistry, University of Delaware, Newark, DE 19716, USA.
Can. J. Chem. 84: 176–186 (2006)
doi:10.1139/V05-239
© 2006 NRC Canada

Canal et al.
177
(hexanes, cm–1): ν(CO) 2126 (vw), 2080 (s), 2071 (m), 2065
Introduction
1
(vw), 2044 (vs), 2025 (m), 1995 (m), 1909 (m). H NMR
Alkynes form a plethora of complexes with group 8 metal
carbonyls that have a remarkable structural diversity (1–3).
For example, over 20 types of complexes have been isolated
from the reaction of alkynes with Fe3(CO)12; they possess up
to five metal atoms and three alkyne units (2, 4). Carty and
co-workers (5, 6) have made notable contributions to the
area, particularly alkyne derivatives of carbonyl clusters of
ruthenium. The reaction of alkynes with Os3(CO)12 also pro-
vides novel products (3, 7–10).
(CD2Cl2) δ: 7.15–7.10 (m), 7.01–6.92 (m), 6.85–6.83 (dd).
13C NMR (CD2Cl2, RT) δ: 192.13 (2C), 192.08 (2C), 185.6
(2C), 184.6 (1C), 176.6 (1C), 175.4 (2C), 172.5 (1C), 169.4
(1C), 169.3 (1C), 168.8 (1C), 128.1 (1C, Ph-C), 128.0 (2C,
Ph-C), 126.6 (2C, Ph-C). MS (LI-MS) m/z: 1331.9 [M+]
(calcd. for M+ = 1332 (100%), 1331 (81.5%)). Anal. calcd.
for C28H10O14Os4: C 25.26, H 0.76; found: C 25.22, H 0.86.
Os3(␮3-␩2-C2Ph2)(CO)10
IR (hexanes, cm–1): ν(CO) 2100 (m), 2066 (vs), 2047 (s),
2028 (s), 2011 (s), 1996 (m), 1982 (sh), 1965 (w) (in agree-
ment with the literature values (7a)). 13C NMR (CO region,
RT) δ: 187.4 (2C), 186.8 (2C), 179.0 (1C), 174.5 (2C), 169.4
(1C), 168.0 (1C). MS (LI-MS) m/z: 1179.9 [M+] (calcd. for
M+ = 1180 (100%), 1179 (86.1%)).
In contrast to Os3(CO)12, the common carbonyl cluster of
osmium (11), Os4(CO)14 readily undergoes substitution with
two-electron donor ligands under mild conditions (12, 13).
Here we describe the reaction of the tetraosmium cluster
2
with diphenylacetylene to give the unique series Os4(µx-η -
C2Ph2)(CO)14–n (n = 0, 1, 2; x = n + 2) in which a C2Ph2
molecule is successively bound to two, three, and four metal
atoms. The dimetallacyclobutene cluster (i.e., the µ2 com-
pound) is of particular interest and is proposed as a model of
corner sites of site-specific surface catalysts.
Os3(␮-␩4-C4Ph4)(CO)9
IR (hexanes, cm–1): ν(CO) 2113 (s), 2057 (vs), 2043 (vs),
2036 (sh), 2011 (vs), 1999 (m), 1990 (sh), 1974 (m), 1930
(m) (in agreement with the literature values (8a)). 13C NMR
(RT) δ: 175.7 (br s), 129.3 (2C, Ph-C), 128.1 (2C, Ph-C),
127.4 (1C, Ph-C). 13C NMR (–60 °C) δ: 181.6 (~1C), 178.3
(~1C), 176.6 (~2C), 174.7 (~2C), 173.0 (~2C), 172.5 (~2C).
MS (LI-MS) m/z: 1029.9 [M+] (calcd. for M+ = 1030
(100%), 1029 (83.3%)).
Experimental section
Unless otherwise stated, manipulations of starting materials
and products were carried out under a nitrogen atmosphere
with the use of standard Schlenk techniques. Hydrocarbon
solvents were refluxed over potassium, distilled, and stored
over molecular sieves before use. Dichloromethane was
dried in a similar manner except that P2O5 was employed as
the drying agent. The Os4(CO)14 was prepared by a literature
procedure (13). The reactions were carried out in Carius
tubes of volumes between 25–50 mL; each was fitted with a
Teflon valve. NMR spectra were recorded on a Bruker
AMX400 spectrometer at the appropriate operating frequen-
Preparation of Os4(␮3-␩2-C2Ph2)(CO)13 (2)
Cluster 2 was prepared in the same manner used to pre-
pare 1, except only a slight excess of C2Ph2 was used
(~1:1.1) and the solution was allowed to stir at room temper-
ature for 2 days. Again the solution changed from brown-red
to a red over the reaction period. The desired product was
isolated as described previously in about 60% yield. The an-
alytical sample of 2 was obtained as air-stable, deep red
crystals by recrystallization from CH2Cl2–hexanes. IR (hex-
ν
1
cies for the H and 13C{1H} NMR spectra. Samples of the
complexes used for 13C{1H} NMR spectra were enriched to
~30% 13CO. The samples were prepared from 13CO-
enriched Os4(CO)14, which in turn was prepared from 13CO-
enriched Os3(CO)12 (~40% 13CO) (14).
anes, cm–1): (CO) 2117 (m), 2078 (vs), 2046 (vs), 2040
(vs), 2012 (m), 1998 (m), 1987 (w), 1981 (w), 1963 (w),
1
1956 (w, sh). H NMR (CD2Cl2) : 7.15–7.07 (m), 6.98–
δ
6.96 (m). 13C NMR (CD2Cl2, RT) δ: 185.6 (1C, JCC
=
Preparation of Os4(␮-␩2-C2Ph2)(CO)14 (1)
38.7 Hz), 182.6 (1C, JCC = 37.2 Hz), 180.5 (1C), 176.5 (6C),
173.2 (1C), 169.3 (1C), 130 (2C, Ph-C), 127.8 (2C, Ph-C),
127.6 (1C, Ph-C). 13C NMR (CD2Cl2, –50 °C) δ: 185.6
(1C), 181.8 (1C), 180.2 (1C), 177.6 (2C), 175.8 (2C), 175.2
(2C), 172.8 (2C), 169.0 (2C). MS (LI-MS) m/z: 1303.6 [M+]
(calcd. for M+ = 1304 (100%), 1303 (87.4%)). Anal. calcd.
for C27H10O13Os4: C 24.85, H 0.77; found: C 25.01, H 0.98.
To a solution of Os4(CO)14 (25 mg, 0.017 mmol) in
CH2Cl2 (25 mL) at room temperature was added an excess
of C2Ph2 (~20×) (C2Ph2 = diphenylacetylene). The solution
was stirred for 105 min during which time the color of the
solution changed from a red-brown to a deep red. The sol-
vent was removed on the vacuum line and the remaining
solid subjected to chromatography on a silica gel column
(1.5 cm × 15 cm). Elution with hexanes removed the excess
C2Ph2. Elution with hexanes–CH2Cl2 (90:10 by volume)
gave several colored bands. The first (yellow) band was a
Preparation of Os4(␮4-␩2-C2Ph2)(CO)12 (3)
The same reaction conditions used to produce 2 were em-
ployed except the reaction mixture was heated to 40 °C for
2 days. Over the time period the solution went from brown-
red through dark red to orange-red as the initially produced
1 went to 2 and then 3. The desired product in ~90% yield
was isolated in the normal manner. (In an alternate proce-
dure, 3 was isolated in 90% yield when either 1 or 2 was al-
lowed to stir at room temperature in hexanes for 18 days).
The analytical sample of 3 was obtained as air-stable, spar-
kling red-orange crystals by recrystallization from CH2Cl2.
IR (hexanes, cm–1): ν(CO) 2098 (m), 2071 (vs), 2046 (vs),
2
trace amount of the known cluster Os3(µ3-η -C2Ph2)(CO)10.
(In the preparations described below, the orange band of
2
Os4(µ4-η -C2Ph2)(CO)12 (3) eluted next). This was followed
2
by a red band that afforded Os4(µ3-η -C2Ph2)(CO)13 (2) in
approximately 25% yield, then a purple band of another
4
known cluster, Os3(µ-η -C4Ph4)(CO)9, in <5% yield. Finally,
a deep red band of the desired cluster 1 was obtained in
about 70% yield. The analytical sample of 1 was obtained as
air-stable, dark red (almost black) needles by recrystalli-
zation from CH2Cl2 and dried overnight under vacuum. IR
1
2041 (vs), 2019 (s), 2000 (s), 1982 (w), 1971 (m). H NMR
© 2006 NRC Canada

178
Can. J. Chem. Vol. 84, 2006
Table 1. Details of the crystal structure determinations of 1·CH2Cl2, 2, and 3.
Structure name
Compound
1·CH2Cl2
C28H10O14Os4
2
3
C27H10O13Os4
C26H10O12Os4
Formula weight
Crystal system
T (°C)
1416.09
Monoclinic
–123
1 303.15
Triclinic
–70
1275.14
Monoclinic
–70
Space group
a (Å)
b (Å)
c (Å)
α (°)
P21/n
P-1
P21/c
8.8718(6)
25.138(2)
14.814(1)
90
97.564(2)
90
12.492(1)
15.179(1)
15.923(1)
78.215(1)
85.257(1)
84.797(2)
2 937.0(4), 4
2.947
9.634(1)
16.743(2)
17.127(2)
90
90.390(2)
90
β (°)
γ (°)
Cell volume (Å3), Z
Dcalcd. (Mg m–3)
3275.0(4), 4
2.872
2762.7(5), 4
3.066
Abs. coeff (mm–1)
Unique data (I > 2σ)
Ra
15.701
3053
0.0622
0.1063
17.315
12 503
0.0539
0.1584
18.402
6466
0.0424
0.1101
b
Rw
aR= Σ|(|Fo| – |Fc|)|/Σ|Fo|.
2
2
bRw = [Σ(w(|Fo| – |Fc|)2)/Σ(wFo )]1/2, w = [σ2(Fo)2 + kFo ]–1.
(CD2Cl2) δ: 7.21–7.17 (m), 7.14–7.10 (m). 13C NMR (tolu-
ene-d8, 90 °C) δ: 180.4 (s, 6C), 175.5 (s, 6C). 13C NMR
cule was refined fully anisotropically and the C—Cl dis-
tances were fixed at 1.72 Å. Due to the poor quality of the
crystal, good data was only obtained to 20° on θ. The largest
residue electron density peak (2.02 e/Å3) was associated
with one of the osmium–osmium bonds. Full-matrix least-
squares refinement on F2 gave R1 = 6.22 for 2σ data and
wR2 = 12.64 for all data (GOF = 1.134).
(CD2Cl2, RT) δ: 180.5 (6C, JOsC (
187Os) = 106.9 Hz), 175.5
(6C, JOsC = 117.5 Hz), 132.5 (2C, Ph-C), 129.3 (1C, Ph-C),
127.2 (2C, Ph-C). 13C NMR (CD2Cl2, –90 °C) δ: 180.5
(s, 6C), 175.5 (w/broad, 6C). MS (LI-MS) m/z: 1275.6
[M+] (calcd. for M+ = 1276 (100%), 1275 (87.4%)). Anal.
calcd. for C26H10O12Os4: C 24.45, H 0.78; found: C 24.65,
H 0.88.
The structures of 2 and 3 were determined at the Univer-
sity of Ottawa, Ottawa, Ontario. Suitable crystals were se-
lected, mounted with Paratone oil on glass fibres, and flash-
cooled to the data collection temperature of –70 °C. Unit-
cell parameters were obtained from 60 data frames, 0.3° θ,
from three different sections of the Ewald sphere. No sym-
metry higher than triclinic was observed in the diffraction
data for 2 and solution in the centrosymmetric space group
option, P-1, yielded chemically reasonable and computation-
ally stable results of refinement. The unit-cell parameters
and systematic absences in the diffraction data of 3 were
uniquely consistent with the space group P21/c. The data
sets were corrected for absorption with the Bruker SADABS
program based on redundant data. Two symmetry-unique,
but chemically equivalent molecules were located in the
asymmetric unit of 2. Features remaining in the final differ-
ence Fourier electron density map of 2 were located <0.9 Å
from osmium atoms and were ignored as noise artifacts. All
non-hydrogen atoms were refined with anisotropic displace-
ment parameters. All hydrogen atoms were treated as idealized
contributions. Structure factors are contained in the
SHELXTL 6.12 program library. A summary of the data col-
lection and refinement for each structure are given in Ta-
ble 1.4
X-ray structural determinations
The structure of 1 was determined at the University of
Western Ontario, London, Ontario. Crystals of 1·CH2Cl2
were grown from a concentrated dichloromethane solution.
A red-orange, flat needle was cut and mounted on a glass fi-
ber. Data were collected at low temperature (–123 °C) on a
Nonius Kappa-CCD diffractometer. The unit cell parameters
were calculated and refined from the full data set. Crystal
cell refinement and data reduction were carried out with the
program DENZO and the data were scaled with the program
SCALEPACK. The crystal data and refinement parameters
for 1·CH2Cl2 are listed in Table 1. The reflection data and
systematic absences were consistent with a monoclinic space
group, P2(1)/n. The SHELXTL-NT V6.1 (G. Sheldrick)
suite of programs was used to solve the structure by direct
methods. Subsequent difference Fourier syntheses allowed
the remaining atoms to be located. The osmium atoms were
refined with anisotropic thermal parameters. The remaining
atoms were refined isotropically. The hydrogen atom posi-
tions were calculated geometrically and were included as
riding on their respective carbon atoms. The solvent mole-
4 Supplementary data for this article are available on the journal Web site (http://canjchem.nrc.ca) or may be purchased from the Depository
of Unpublished Data, Document Delivery, CISTI, National Research Council Canada, Ottawa, ON K1A 0R6, Canada. DUD 4089. For more
information on obtaining material refer to http://cisti-icist.nrc-cnrc.gc.ca/irm/unpub_e.shtml. CCDC 279768–279770 contain the crystallo-
graphic data for this manuscript. These data can be obtained, free of charge, via http://www.ccdc.cam.ac.uk/conts/retrieving.html (Or from
the Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax +44 1223 336033; or de-
posit@ccdc.cam.ac.uk).
© 2006 NRC Canada

Canal et al.
179
2
Fig. 1. Molecular structure of Os4(µ-η -C2Ph2)(CO)14 (1).
Table 2. Selected bond lengths (angstroms)
and angles (degrees) for 1.
Bond lengths (Å)
Os(1)—Os(2)
Os(1)—Os(4)
Os(2)—Os(3)
Os(2)—Os(4)
Os(3)—Os(4)
C(51)—C(61)
2.866(2)
2.859(2)
2.880(2)
2.916(2)
2.907(2)
1.32(4)
Os(1)—C(51)
2.13(3)
Os(2)—C(61)
2.28(3)
Os—C(CO) range
Bond angles (°)
Os(1)-Os(2)-Os(3)
Os(1)-Os(2)-Os(4)
Os(1)-Os(4)-Os(2)
Os(1)-Os(4)-Os(3)
Os(2)-Os(1)-Os(4)
Os(2)-Os(3)-Os(4)
Os(2)-Os(4)-Os(3)
Os(3)-Os(2)-Os(4)
Os(1)-Os(2)-C(61)
Os(2)-Os(1)-C(51)
Os(1)-C(51)-C(52)
Os(1)-C(51)-C(61)
Os(2)-C(61)-C(51)
Os(2)-C(61)-C(62)
C(52)-C(51)-C(61)
C(62)-C(61)-C(51)
1.84(3)–1.96(3)
119.44(6)
59.25(5)
59.51(4)
118.77(6)
61.25(5)
60.52(4)
59.28(4)
60.20(4)
65.5(6)
73.5(8)
126(2)
Results and discussion
Os4(␮-␩2-C2Ph2)(CO)14 (1)
Compound 1 was isolated as dark red needles in about
70% yield from the reaction of Os4(CO)14 in CH2Cl2 with an
excess of C2Ph2 at room temperature (eq. [1]).
[1]
Os4(CO)14 + C2Ph2
108(2)
113(2)
119(2)
126(3)
25 °C
2
⎯⎯⎯→ Os4(µ-η -C2Ph2)(CO)14 (1)
CH2Cl2
+ other products
128(3)
(By comparison, the reaction of Os3(CO)12 with C2Ph2 re-
quires temperatures of 160 °C or above to proceed at a mea-
surable rate (9)). An excess of the acetylene ensured that 1
formed quickly as it slowly loses CO to give Os4(µ3-η -
C2Ph2)(CO)13 (2). Besides 2, other products isolated from
Os(1) (17). Interestingly, it is the shortest Os—Os bond in 1.
We have reported numerous complexes that contain unsup-
ported Os—M (M = transition metal) dative bonds (18).
Cluster 1 possesses a dimetallacyclobutene ring (i.e.,
Os(1)-Os(2)-C(61)-C(51)). The geometry about each C atom
of the unit is approximately trigonal planar, although the in-
ternal Os-C-C angles (108(2)° and 113(2)°) are somewhat
compressed from the ideal value of 120°. The C—C bond
length of 1.32(4) Å is consistent with an uncoordinated C=C
double bond. The Os—C bond lengths to the alkyne are
2.13(3) (to Os(1)) and 2.28(3) Å (to Os(2)). Because of the
somewhat poor quality of the data, it is not clear that the dif-
ference is significant, or a genuine electronic effect: Os(1) is
six-coordinate, whereas Os(2) is seven-coordinate. The
lengths may be compared with the Os—C lengths to the car-
bonyl ligands in 1 that are in the range 1.84(4) to 1.95(3) Å.
Dimetallacyclobutene compounds are comparatively rare
and a search of the literature (SciFinder Scholar) of this term
yielded only 11 publications containing compounds with this
bonding motif (1, 19–23). Some recent examples are Ir2(µ-
2
the reaction shown in eq. [1] were the known clusters
2
4
Os3(µ3-η -C2Ph2)(CO)10 (7) and Os3(µ-η -C4Ph4)(CO)9 (8, 9).
The stereochemical nonrigidity of the former cluster has
been investigated by variable-temperature 13C{1H} NMR
spectroscopy (15) and found similar to the nonrigidity exhib-
2
ited by Os3(µ3-η -C2R2)(CO)10 (R = Me, Et) (10). All prod-
ucts are brightly colored and were readily separated by
column chromatography. The derivatives appeared air stable,
although crystals of 1 deteriorated on standing. This can be
attributed to either loss of CO (to give 2) or loss of the
CH2Cl2 of crystallization. (Crystals of 1 suitable for the
structural determination were obtained only after some diffi-
culty).
Compound 1 was characterized by C/H/N analysis, IR,
mass (parent ion), 1H and 13C NMR spectroscopy, and X-ray
crystallography. The structure of 1 (as the CH2Cl2 solvate)
reveals an approximately planar Os4C2 unit (Fig. 1); selected
bond lengths and angles for 1 are given in Table 2. The Os4
unit adopts a flat butterfly arrangement with five Os—Os
bonds that range in length from 2.859(2) to 2.916(2) Å.
These lengths are typical for open Os carbonyl clusters and
may be compared with the average Os—Os bond distance in
Os3(CO)12 of 2.877 Å (16). If it is assumed that the Os—Os
bonds are single nondative covalent bonds, then Os(1) has a
17 electron count and Os(4) a 19 electron count. The
Os(1)—Os(4) can therefore be regarded as a dative metal–
metal bond with the Os(4) atom donating two electrons to
2
5
η -C2R2)(η -C9H7)2(CO)2 (R = Ph, Tol) (20) and Re2Cp*2(µ-
2
2
η -C2(CO2CH3)2(CO)4 (21). Only one compound, Ir4[(µ-η -
2
C2(MeCO2)2]2[(µ4-η -C2(MeCO2)2]2(CO)8, contains more
than two metal atoms, that is, it is a cluster compound (22).
The bond lengths and angles in 1 are comparable to those
2
found in Os2[µ-η -C2(CO2Me)2](CO)8 (C—C bond
=
1.33(1) Å; both Os—C bonds = 2.138(5) Å; Os-C-C angle =
111.2(2)°) (23). The C—C bond length in 1 may be also be
compared with the corresponding length (1.53(3) Å) in the
saturated Os-C-C-Os unit in Os2(␮-␩2-C2H4)(CO)8 (24).
© 2006 NRC Canada

180
Can. J. Chem. Vol. 84, 2006
2
Fig. 2. The 13C{1H} NMR spectrum (CO region) of 1 (13CO en-
riched) in CH2Cl2–CD2Cl2 (4:1) at 21 °C.
Fig. 3. Molecular structure of Os4(µ3-η -C2Ph2)(CO)13 (2).
The 13C{1H} NMR spectrum of 1 (13CO enriched) in
CD2Cl2–CH2Cl2 at ambient temperature exhibits four signals
of intensity 2 and six signals of intensity 1 in the carbonyl
region (Fig. 2). This observation is consistent with the view
that 1 has the same structure in solution as that in the solid
state, and the carbonyls are rigid on the NMR timescale at
this temperature. Typically, resonances due to axial carbonyl
ligands come to low field of those due to equatorial carbon-
yls (14, 25). As can be seen from Fig. 2, one signal of inten-
sity 2 is shifted to higher field and one signal of intensity 1
to lower field. These atypical signals are tentatively assigned
to the carbonyls attached to Os(1) because of the unusual
bonding at this atom.
The 18-electron rule requires that the alkyne fragment do-
nates two electrons to the wingtip Os atom (e.g., Os(2) in
molecule 2A) and one electron to each hinge Os atom (e.g.,
Os(1) and Os(3) in 2A). The Os—C(alkyne) lengths that in-
volve the wingtip Os atoms (e.g., Os(2) and Os(6)) are in the
range 2.27(1)–2.29(1) Å and are somewhat longer than the
other Os—C(alkyne) lengths in 2 (range 2.15(1)–2.18(1) Å).
The bonding may be viewed as a dimetallacyclobutene unit
with the C=C double bond acting as a typical alkene ligand
toward the wingtip Os atom. The length of the C—C bond
(1.42 Å) of the alkyne unit suggests a formal bond order less
than two, as observed in similar compounds (7, 10, 26, 27).
The bonding of the alkyne to the Os3 triangle in 2 is similar
2
Os4(␮3-␩2-C2Ph2)(CO)13 (2)
to that in compounds such as Os3(µ3-η -C2R2)(CO)10 (R = Et
2
(10), R = Ph (7)), Os5(µ3-η -C2H2)(CO)17 (26), and Os7(µ3-
When 1 was stirred in CH2Cl2 at room temperature for
2 days cluster 2 was produced in good yield (eq. [2]). The
product was isolated after chromatography as red crystals
that were characterized by C/H/N analysis, IR, mass (parent
2
η -C2Me2)(CO)19 (27).
The 13C{1H} NMR spectra in the carbonyl region of 2
(13CO enriched) in CD2Cl2–CH2Cl2 at –50 and 21 °C are
shown in Fig. 4. The lower temperature spectrum exhibits
eight signals in an approximate 1:1:1:2:2:2:2:2 (low to high
field). This is consistent with the view that 2 has the same
structure in solution as in the crystalline state, and that the
carbonyls are rigid on the NMR timescale at –50 °C. The
two signals at lowest field (i.e., at δ 185.6 and 181.8) exhibit
satellites and are therefore readily assigned to the C atoms
of the two chemically different axial CO ligands on the
unbridged wingtip Os atom (i.e., the C atoms C(23) and
C(25) in Fig. 3). This is because only these peaks would be
expected to exhibit observable 13C–13C coupling owing to
the mutual trans arrangement of these carbonyls. (Recall the
complex was 13CO enriched.) The remaining signal of inten-
sity 1 (at δ 180.2) is attributed to the axial carbonyl atom on
the bridged wingtip Os atom (i.e., C(17)). The signal at
δ 177.6 is assigned to the two C atoms of the axial carbonyls
on the hinge Os atoms (i.e., C(14) and C(20)). As stated pre-
viously, 13C NMR resonances due to C atoms of axial car-
bonyls usually appear to lower field than the corresponding
signals of equatorial carbonyls. The resonances at δ 175.8
and 175.2 are assigned to the C atoms of the equatorial car-
bonyls on the hinge Os atoms on the basis of their exchange
behavior described in the following paragraph. The remain-
1
ion), H and 13C NMR spectroscopy, and X-ray crystallogra-
phy.
2
[2]
Os4(µ-η -C2Ph2)(CO)14
25 °C
2
⎯⎯⎯→ Os4(µ3-η -C2Ph2)(CO)13 (2) + CO
2 days
The structure of 2 has two independent molecules in the
unit cell that differ in the orientation of the phenyl rings, but
are otherwise identical. A view of one of the molecules
(molecule 2A) is given in Fig. 3; selected bond lengths and
angles for each molecule are collected in Table 3. The struc-
ture of 2 consists of an almost flat butterfly arrangement, as
found in 1, with the alkyne spanning one of the Os3 trian-
gles. The peripheral Os—Os lengths associated with the
alkyne in each molecule (range = 2.7392(7)–2.7740(7) Å)
are significantly shorter than the peripheral Os—Os lengths
that involve Os atoms not directly bound to the C2Ph2 (range
2.8665(6)–2.8947(6) Å). The shorter Os—Os bonds can be
attributed to the bonding requirements of the bridging C2Ph2
ligand. The unbridged Os—Os bonds are typical for open Os
clusters. The hinge Os—Os bonds in each molecule are of
intermediate length (2.8204(6) and 2.8007(7) Å).
© 2006 NRC Canada

Canal et al.
181
Table 3. Selected bond lengths (angstroms) and angles (degrees) for 2.
Molecule A
Molecule B
Bond lengths (Å)
Os(1)—Os(2)
Os(1)—Os(3)
Os(1)—Os(4)
Os(2)—Os(3)
Os(3)—Os(4)
C(33)—C(34)
Os(5)—C(47)
Os(6)—C(47)
2.7586(6)
2.8204(6)
2.8665(6)
2.7454(6)
2.8947(6)
1.42(2)
2.17(1)
2.29(1)
2.29(1)
2.16(1)
Os(5)—Os(6)
Os(5)—Os(7)
Os(5)—Os(8)
Os(6)—Os(7)
Os(7)—Os(8)
C(47)—C(48)
Os(5)—C(47)
Os(6)—C(47)
Os(6)—C(48)
Os(7)—C(48)
2.7392(7)
2.8007(7)
2.8871(7)
2.7740(7)
2.8939(7)
1.42(2)
2.17(1)
2.29(1)
2.29(1)
2.16(1)
Os(6)—C(48)
Os(7)—C(48)
Os—C(CO) range
Bond angles (°)
Os(1)-Os(2)-Os(3)
Os(1)-Os(3)-Os(2)
Os(1)-Os(3)-Os(4)
Os(1)-Os(4)-Os(3)
Os(2)-Os(1)-Os(3)
Os(2)-Os(1)-Os(4)
Os(2)-Os(3)-Os(4)
Os(3)-Os(1)-Os(4)
Os(1)-Os(2)-C(33)
Os(1)-C(33)-Os(2)
Os(2)-Os(1)-C(33)
Os(2)-Os(3)-C(34)
Os(2)-C(34)-Os(3)
Os(3)-Os(2)-C(33)
Os(1)-Os(3)-C(34)
Os(3)-Os(1)-C(33)
Os(1)-C(33)-C(34)
Os(3)-C(34)-C(33)
Os(2)-C(33)-C(34)
Os(2)-C(34)-C(33)
C(33)-Os(2)-C(34)
Os(1)-C(33)-C(32)
Os(2)-C(33)-C(32)
Os(2)-C(34)-C(40)
Os(3)-C(34)-C(40)
C(32)-C(33)-C(34)
C(40)-C(34)-C(33)
1.88(1)–1.99(1)
61.65(2)
59.41(2)
60.19(2)
58.62(2)
58.94(2)
118.26(2)
117.74(2)
61.19(2)
49.2(3)
76.8(3)
54.0(3)
53.5(3)
76.1(4)
70.6(3)
71.1(3)
71.0(3)
110.4(8)
107.5(8)
71.2(6)
72.6(6)
36.1(4)
Os(5)-Os(6)-Os(7)
Os(5)-Os(7)-Os(6)
Os(5)-Os(7)-Os(8)
Os(5)-Os(8)-Os(7)
Os(6)-Os(5)-Os(7)
Os(6)-Os(5)-Os(8)
Os(6)-Os(7)-Os(8)
Os(7)-Os(5)-Os(8)
61.06(2)
58.86(2)
60.90(2)
57.96(2)
60.08(2)
118.54(2)
117.15(2)
61.14(2)
124.2(8)
120.7(8)
125.2(8)
125.4(8)
125(1)
126(1)
Note: Two independent molecules in the unit cell.
ing resonances at δ 172.8 and 169.0 are attributed to the C
atoms of the equatorial carbonyls on the wingtip Os atoms.
The latter signal is attributed to the carbonyls of the
Os(CO)4 grouping since it remains sharp in the spectrum at
ambient temperature (see the following paragraph).
In the spectrum at 21 °C, three of the resonances have co-
alesced to a broad singlet (Fig. 4), indicative of exchange
between the carbonyls that give rise to the signals in the
spectrum at lower temperature. This is attributed to carbonyl
exchange of the hinge Os(CO)3 groupings. Rotation of
Os(CO)3 units in condensed Os carbonyl clusters is com-
mon. Examples are Os6(CO)17(L) (L = CO, PPh3) (28, 29)
and Os7(CO)21 (30). The hinge Os atoms are seven-
coordinate, whereas the wingtip Os atoms are nominally
six-coordinate (if the bond to the C2Ph2 is a conventional
metal–alkene bond). It is well-known that mononuclear six-
coordinate complexes are usually rigid, whereas the corre-
sponding seven-coordinate complexes are invariably
nonrigid in solution.
Although the reason for the difference in activation barriers
to exchange may be based upon electronic considerations, it
can be rationalized by using simple steric arguments. We
have shown that the ground-state and transition-state ener-
gies are both important in determining a barrier to rotation
6
of the arene ring in (η -arene)Fe(CO)2(SiCl3) complexes
(31). In six-coordinate complexes, the trigonal prismatic
transition state for M(CO)3 rotation has three eclipsing inter-
actions as depicted in Scheme 1. Also of importance is that
in the ground state all the substituents are perfectly stag-
gered with their nearest neighbors and hence it is at a low
© 2006 NRC Canada

182
Can. J. Chem. Vol. 84, 2006
Fig. 4. The 13C{1H} NMR spectrum (CO region) of 2 (13CO en-
Scheme 1.
riched) in CH2Cl2–CD2Cl2 (4:1) at 21 and –50 °C.
E
Six-coordinate complex
E
Seven-coordinate complex
Although the Ru analogue of 3 has been known since
1977 (32), the Os compound has not been previously re-
ported (SciFinder Scholar). A view of the structure of 3 is
shown in Fig. 5. The structure is typical of clusters of the
energy. In a seven-coordinate complex, the transition state to
rotation of a trigonal unit only has two close eclipsing inter-
actions in the transition state and therefore might be ex-
pected to be at a lower energy than the transition-state
energy in the six-coordinate case (Scheme 1). What may not
be realized is that in the seven-coordinate configuration the
ground state is destabilized relative to the six-coordinate
configuration because in the higher coordination case there
are still close interactions in the lowest energy conformation.
In the spectrum at 21 °C, the signals attributed to the
wingtip Os(CO)3 group have broadened relative to the sig-
nals owing to the wingtip Os(CO)4 unit. This probably indi-
cates the onset of rotation on the NMR timescale of this
Os(CO)3 group and is relevant to the discussion on the rota-
tion of the Os(CO)3 units in 3 discussed in the following.
The NMR spectrum of 2 at higher temperatures was not in-
vestigated because of its ready decarbonylation to 3.
2
formula, M4(µ4-η -C2RR′)(CO)12 (M = Fe, Ru, Os) (1–3, 6,
32–34). These clusters may be viewed as electron-deficient
butterfly clusters with 60 cluster valence electrons, or as
closo octahedral clusters with seven skeletal electron pairs
according to PSEPT (6). In 3 the peripheral Os—Os bonds
range in length from 2.7457(5) to 2.7652(5) Å (Table 4),
lengths that are typical of Os—Os lengths in condensed Os
clusters. The hinge Os(1)—Os(3) bond is longer at
2.8742(5) Å. The four Os—C lengths of the alkyne unit to
the wingtip Os atoms (Os(2) and Os(4)) are slightly longer
(2.251(7)–2.274(8) Å) than the Os—C lengths to the hinge
Os atoms (2.179(8), 2.201(8) Å). The pattern of the Os—Os
and Os—C lengths closely match those previously found in
2
2
Ru4(µ4-η -C2Ph2)(CO)12 (32) and Os4(µ4-η -C2H2)(CO)12
(34). The C—C bond length (1.49(1) Å) of the alkyne ligand
in 3 is the longest of the three alkyne C—C lengths deter-
mined in this study and is close to that of a C—C single
bond length.
Os4(␮4-␩2-C2Ph2)(CO)12 (3)
When the reaction shown in eq. [2] was carried out at
2
40 °C, Os4(µ4-η -C2Ph2)(CO)12 (3) was isolated in ~90%
Carbon-13 NMR spectra in the carbonyl region of 3 (13CO
enriched) at different temperatures are shown in Fig. 6. If
the carbonyl groups were rigid, each chemically distinct
Os(CO)3 (i.e., wingtip and hinge moieties) unit would give
rise to a 2:1 or 1:1:1 pattern of signals in the NMR spectrum
(32). It is apparent that at 21 and 90 °C there is rapid rota-
tion of the two different types of Os(CO)3 groupings in 3,
but that migration of CO ligands between the wingtip and
hinge Os atoms does not occur on the NMR timescale. This
is also observed in clusters such as Os6(CO)18 (28).
In the spectrum at –90 °C, one of the signals has col-
lapsed almost to the baseline, indicative of slowed exchange
yield (eq. [3]). The cluster, as bright orange-red, air-stable
crystals, was characterized by C/H/N analysis, IR, mass
1
(parent ion), H and 13C NMR spectroscopy, and by X-ray
crystallography. Compound 3 has remarkable thermal stabil-
ity. Overnight pyrolysis of 3 in C6F6 under vacuum at
230 °C had little effect on the cluster; decomposition (in part
to Os3(CO)12) only commenced at 250 °C. Even after 4 days
at 250 °C 3 had not completely decomposed.
2
[3]
Os4(µ3-η -C2Ph2)(CO)13
40 °C
2
⎯⎯⎯→ Os4(µ4-η -C2Ph2)(CO)12 + (3) CO
2 days
© 2006 NRC Canada

Canal et al.
183
2
Fig. 5. Molecular structure of Os4(µ4-η -C2Ph2)(CO)12 (3).
Table 4. Selected bond lengths (ang-
stroms) and angles (degrees) for 3.
Bond lengths (Å)
Os(1)—Os(2)
Os(1)—Os(3)
Os(1)—Os(4)
Os(2)—Os(3)
Os(3)—Os(4)
Os—C(CO) range
C(13)—C(20)
Os(1)—C(13)
Os(2)—C(13)
Os(2)—C(20)
Os(3)—C(20)
Os(4)—C(13)
Os(4)—C(20)
2.7515(5)
2.7592(5)
2.8742(5)
2.7652(5)
2.7457(5)
1.908(9)–1.95(1)
1.49(1)
2.179(8)
2.274(8)
2.265(7)
2.201(7)
2.255(8)
2.251(7)
Bond angles (°)
Os(1)-Os(2)-Os(3)
Os(1)-Os(3)-Os(2)
Os(2)-Os(1)-Os(3)
Os(1)-Os(3)-Os(4)
Os(1)-Os(4)-Os(3)
Os(3)-Os(1)-Os(4)
Os(1)-Os(2)-C(13)
Os(1)-C(13)-Os(2)
Os(2)-Os(1)-C(13)
Os(1)-Os(4)-C(13)
Os(4)-Os(1)-C(13)
Os(1)-C(13)-Os(4)
Os(2)-Os(3)-C(20)
Os(3)-Os(2)-C(20)
Os(2)-C(20)-Os(3)
Os(3)-Os(4)-C(20)
Os(4)-Os(3)-C(20)
Os(3)-C(20)-Os(4)
Os(2)-C(13)-C(20)
Os(2)-C(20)-C(13)
C(13)-Os(2)-C(20)
Os(4)-C(13)-C(20)
Os(4)-C(20)-C(13)
C(13)-Os(4)-C(20)
Os(2)-Os(1)-Os(4)
Os(1)-C(13)-C(19)
Os(3)-C(20)-C(26)
C(19)-C(13)-C(20)
C(13)-C(20)-C(26)
62.80(1)
58.37(1)
58.83(1)
58.76(1)
62.95(1)
58.30(1)
50.3(2)
76.3(3)
53.4(2)
52.7(2)
52.7(2)
76.9(2)
52.8(2)
50.7(2)
76.5(2)
51.1(2)
52.7(2)
76.2(2)
70.5(4)
71.1(4)
38.4(3)
70.5(4)
70.8(4)
38.7(3)
91.67(1)
125.1(6)
125.6(6)
126.3(7)
126.0(7)
of one of the pairs of Os(CO)3 units. This was also observed
2
in the variable-temperature 13C NMR spectra of Ru4(µ4-η -
C2Ph2)(CO)12 (32). Given the fluxional behavior of the
Os(CO)3 units in 2, it is probable that the carbonyls that are
still undergoing rapid exchange in 3 are those of the hinge
Os(CO)3 units.
Clusters 1–3 as models for surface catalysts
As mentioned in the Introduction, 1–3 form a unique se-
ries of clusters in which an alkyne ligand is successively
bound to two, three, and four metal atoms. Indeed, any se-
ries of structurally characterized clusters that involves the
stepwise loss of two carbonyls are rare. One such series,
however, is Co4(CO)x(PNR2)2 (x = 8, 9, or 10) recently re-
ported by Carty and co-workers (35).
Muetterties et al. (36, 37) first pointed out that the struc-
tures of molecular clusters may provide insight to the under-
standing of surface catalysis. Despite its importance and
intense study, the precise nature of the active sites in surface
catalysts is still unknown (38–42). Recent investigations by
physical techniques indicate that the catalytic properties of
small particles depend on their size and imperfections (43,
44). For example, a recent study by Wei and Iglesia (44) of
CH4 reactions on Ru-based catalysts concluded that edge
and corner atoms, with fewer Ru neighbors than those on
terraces, were primarily responsible for C-H activation. We
believe that compounds 1–3 may provide insight to the un-
derstanding of these active sites. In Table 5 the C—C and
Os—C bond lengths for 1–3 are collected together for ready
comparison. Scheme 2 summarizes the preparation of the
clusters from Os4(CO)14.
38), the data in Table 5 indicates that the alkyne is tightly
bound to the Os atoms in 3 and as such would be unreactive.
Consider, for example, the thermal stability of 3. The alkyne
also appears to be tightly bound to the Os4 unit in 2, al-
though less so than in 3
Cluster 1 is a model for an alkyne bound to a corner site
of a metal surface; compound 2 is a model for an alkyne
bound to a planar metal surface; whereas 3, with its butterfly
arrangement of metal atoms, is a model for an alkyne bound
to a step site of a metal surface. Although step sites have
been proposed as the active sites of surface catalysts (6, 37,
On the other hand, the C2Ph2 molecule in 1 is bound to
the Os4 nucleus via two σ bonds, one of which is long
(Table 5). It might be expected that there would also be con-
siderable ring strain in the diosmacyclobutene unit. Further-
more, because of the proximity of the osmium atoms the
frontier orbitals about the C atoms in the bound C2Ph2 will
© 2006 NRC Canada

184
Can. J. Chem. Vol. 84, 2006
Table 5. The C—C and Os—C bond lengths of the alkyne-Os unit in 1–3.
Compound
C—C bond lengths (Å)
Os—C bond lengths (Å) (σ)
Os—C bond lengths (Å) (π)
1
1.32(4)
1.42(2)
1.42(2)
1.49(1)
2.13(3), 2.28(3)
2.15(1), 2.18(1)
2.17(1), 2.16(1)
2.179(8), 2.201(7)
—
2Aa
2Ba
3
2.27(1), 2.29(1)
2.29(1), 2.29(1)
2.251(7), 2.255(8), 2.265(7), 2.274(8)
aTwo independent molecules in the unit cell.
Fig. 6. The 13C{1H} NMR spectrum (CO region) of 3 (13CO en-
riched) in CH2Cl2–CD2Cl2 (4:1) at 21 and –90 °C, and at 90 °C
in C7D8.
Scheme 2.
Ph
C
Ph
Os
Os
Os
C
90 °C
C2Ph2
Os
Os
Os
Os
Os
1
Os4(CO)14
-CO
Ph
Os
Ph
Os
Os
C
C
C
-CO
Os
Ph
Os
Os
Ph
Os
C
Os
21 °C
2
3
Hinshelwood mechanism) or by direct attack by molecular
H2 (the Rideal mechanism) (38, 47).
The intermediacy of an organic substrate bound to two
low-coordinate metal atoms on a surface may have more
general applicability to the understanding of the action of
site-specific surface catalysis. The precursory cluster
Os4(CO)14 (that has a tetrahedral Os4 nucleus) is prepared by
pyrolysis of Os4(CO)15. The latter cluster has a planar struc-
ture and hence is a model for a metal surface. Both
Os4(CO)14 and Os4(CO)15 have unusual bonding that is not
found in higher nuclearity Os clusters (48, 49).
-90 °C
Some corners and imperfect edges of metal surfaces re-
semble these clusters. This is shown in Scheme 3 for a cor-
ner site; the active site is designated a “PT” (planar-
tetrahedral) site. To allow the unique metal atom to hop from
the planar to the tetrahedral site (Scheme 3) probably re-
quires the atom not to be coordinated to a metal atom below
the plane it originates from.
There would be more of these atoms the greater the sur-
face area and in freshly prepared catalysts. It is a general ob-
servation for surface catalysts that they are usually more
active the more finely divided they are and if they are freshly
prepared (38). It may be that after the surface and step sites
are saturated (i.e., analogous to 2 and 3) there must be sur-
face reformation in order for the PT sites to be activated. A
ligand must also dissociate from the dangling atom to allow
formation of the tetrahedral form of the PT site. Both steps
might explain the lag time that some catalysts exhibit before
catalysis begins (38). The model also explains why only cer-
tain metals are active and why bimetallic catalysts can be
more active than catalysts comprised of the individual atoms
(38). The metal atoms of the PT site have few metal–metal
be different from uncoordinated C2Ph2 that may also result
in enhanced reactivity of the coordinated alkyne.
The catalytic half-hydrogenation of alkynes to cis-alkenes
is an important industrial process and is normally carried out
with solid Pd catalysts that have been poisoned (e.g.,
Lindlar’s catalyst) (40, 45). Some homogeneous half-
hydrogenation catalysts are however known (46). The mech-
anism of the hydrogenation is not fully understood, but a
mechanism that involves a dimetallacyclobutene unit appears
to be an attractive possibility. Partial hydrogenation of the
alkyne in the dimetallacyclobutene unit could take place by
reaction with coordinated hydrogen (the Langmuir–
© 2006 NRC Canada

Canal et al.
185
Scheme 3.
L
∆
-L
"PT site"
R
C
C
R
R
C
C
Further reaction
R
ford. 1995. Chap. 8 and 13–16; Some recent examples:
(c) M.I. Bruce, A.C. Meier, B.W. Skelton, A.H. White, and
N.N. Zaitseva. Aust. J. Chem. 54, 319 (2001); (d) R.L.
Mallors, A.J. Blake, P.J. Dyson, B.F.G. Johnson, and S. Par-
sons. Organometallics, 16, 1668 (1997); (e) M. Deng and
W.K. Leong. Organometallics, 21, 1221 (2002); ( f ) S.P.
Tunik, V.D. Khripun, I.A. Balova, E. Nordlander, M. Haukka,
T.A. Pakkanen, and P.R. Raithby. Organometallics, 20, 3854
(2001); (g) J.A. Cabeza, I. del Río, P. García-Álverez, and D.
Miguel. Organometallics, 24, 665 (2005)
contacts and might be expected to have frontier orbitals like
the corresponding metal atoms in mononuclear complexes.
The most active metals for surface catalysts usually contain
metal atoms that are present in the most active homogeneous
mononuclear catalysts (e.g., Rh and Pd) (50). In the model,
the substrate is bound to just two metal atoms and therefore
it will have different susceptibilities to attack depending on
whether it is coordinated to metals that are identical or dif-
ferent. Catalysts are often more effective with oxide promot-
ers (38). If the atoms directly below or adjacent to the PT
site were oxide ions, then the reactivity of the site would be
strongly influenced by the ions.
4. (a) V. Busetti, G. Granozzi, S. Aime, R. Gobetto, and D.
Osella. Organometallics, 3, 1510 (1984); (b) S. Brait, G.
Gervasio, D. Marabello, and E. Sappa. J. Chem. Soc. Dalton
Trans. 989 (2000); (c) E. Gatto, G. Gervasio, D. Marabello,
and E. Sappa. J. Chem. Soc. Dalton Trans. 1448 (2002).
5. Recent examples: (a) P.J. Low, K.A. Udachin, and A.J. Carty.
J. Cluster Sci. 15, 1040 (2005); (b) P.J. Low, T.M. Hayes, K.A.
Udachin, A.E. Goeta, J.A.K. Howard, D. Gary, and A.J. Carty.
J. Chem. Soc. Dalton Trans. 1455 (2002); (c) M.I. Alcalde,
A.J. Carty, Y. Chi, E. Delgado, B. Donnadieu, E. Hernandez,
K. Dallmann, and J. Sanchez-Nieves. J. Chem. Soc. Dalton
Trans. 2502 (2001); (d) P.J. Low, K.A. Udachin, G.D. Enright,
and A.J. Carty. J. Organomet. Chem. 578, 103 (1999); (e) P.
Blenkiron, G.D. Enright, P.J. Low, J.F. Corrigan, N.J. Taylor,
Y. Chin, J.-Y. Saillard, and A.J. Carty. Organometallics, 17,
2447 (1998); ( f ) E. Delgado, Y. Chi, W. Wang, G. Hogarth,
P.J. Low, G.D. Enright, S.-M. Peng, G.-H. Lee, and A.J. Carty.
Organometallics, 17, 2936 (1998); (g) P.J. Low, G.D. Enright,
and A.J. Carty. J. Organomet. Chem. 565, 279 (1998).
6. E. Sappa, A. Tiripicchio, A.J. Carty, and G.E. Toogood. Prog.
Inorg. Chem. 35, 437 (1987).
7. (a) M. Tachikawa, J.R. Shapley, and C.G. Pierpont. J. Am.
Chem. Soc. 97, 7172 (1975); (b) C.G. Pierpont. Inorg. Chem.
16, 636 (1977);
8. (a) O. Gambino, G.A. Vaglio, R.P. Ferrari, and G. Cetini. J.
Organomet. Chem. 30, 381 (1971); (b) G. Ferraris and G.
Gervasio. J. Chem. Soc. Dalton Trans. 1813 (1974); (c) A.J.
Poë, D.H. Farrar, R. Ramachandran, and C. Moreno. Inorg.
Chem. Acta, 274, 82 (1998); (d) A.J. Poë and C. Moreno.
Organometallics, 18, 5518 (1999).
Studies on the reactivity of clusters 1–3 (and derivatives
thereof) with hydrogen and other small molecules are cur-
rently in progress.
Acknowledgements
We thank the Natural Sciences and Engineering Research
Council of Canada (NSERC) for financial support.
References
1. (a) E. Sappa, A. Tiripiccio, and P. Braustein. Chem. Rev. 83,
203 (1983); (b) P.R. Raithby and M.J. Rosales. Adv. Inorg.
Radiochem. 29, 169 (1985); (c) S. Deabate, R. Giordano, and
E. Sappa. J. Cluster Sci. 8, 407 (1997); (d) E. Sappa. J.
Organomet. Chem. 573, 139 (1999); (e) P.J. Low and M.I.
Bruce. Adv. Organomet. Chem. 48, 71 (2001).
2. Fe: (a) G. Wilkinson, F.G.A. Stone, and E.W. Abel (Editors).
Comprehensive organometallic chemistry. Vol. 4. Pergamon,
Oxford. 1982. Chap. 31.3–31.5; (b) G. Wilkinson, F.G.A.
Stone, and E.W. Abel (Editors). Comprehensive organo-
metallic chemistry II. Vol. 7. Pergamon, Oxford. 1995. Chap.
2–4; Some recent examples: (c) P.J. Low, A.J. Carty, K.A.
Udachin, and G.D. Enright. J. Chem. Soc. Chem. Commun.
411 (2001); (d) A.J. Carty, G. Hogarth, G.D. Enright, J.W.
Steed, and D. Georganopoulou. J. Chem. Soc. Chem.
Commun. 1499 (1999).
3. Ru/Os: (a) G. Wilkinson, F.G.A. Stone, and E.W. Abel (Edi-
tors). Comprehensive organometallic chemistry. Vol. 4.
Pergamon, Oxford. 1982. Chap. 32.3–32.6 and 33; (b) G.
Wilkinson, F.G.A. Stone, and E.W. Abel (Editors). Compre-
hensive organometallic chemistry II. Vol. 7. Pergamon, Ox-
9. A.J. Poë, C.N. Sampson, and R.T. Smith. J. Am. Chem. Soc.
108, 5459 (1986).
10. E. Rosenberg, J. Bracker-Novak, R.W. Gellert, S. Aime, R.
Gobetto, and D. Osella. J. Organomet. Chem. 365, 163 (1989).
11. A.J. Deeming. Adv. Organomet. Chem. 26, 1 (1986).
© 2006 NRC Canada

186
Can. J. Chem. Vol. 84, 2006
12. (a) C.-Y. Lu, F.W.B. Einstein, V.J. Johnston, and R.K.
Pomeroy. Inorg. Chem. 28, 4212 (1989); (b) W. Wang, R.J.
Batchelor, F.W.B. Einstein, C.-Y. Lu, and R.K. Pomeroy.
Organometallics, 12, 3598 (1993).
32. B.F.G Johnson, J. Lewis, B.E. Reichert, K.T. Schorpp, and
G.M. Sheldrick. J. Chem. Soc. Dalton Trans. 1417 (1977).
33. (a) H. Bantel, B. Hansert, A.K. Powell, M. Tasi, and H.
Vahrenkamp. Angew. Chem. Int. Ed. Engl. 28, 1059 (1989);
(b) H. Chen, B.F.G. Johnson, J. Lewis, and P.R. Raithby. J.
Chem. Soc. Chem. Commun. 373 (1990); (c) G. Chen, M.
Deng, C.H. Lee, and W.K. Leong. Organometallics, 21, 1227
(2002).
13. J.P. Canal, G.P.A. Yap, and R.K. Pomeroy. Organometallics,
22, 3439 (2003).
14. R.F. Alex and R.K. Pomeroy. Organometallics, 41, 2437 (1987).
15. J.P. Canal. Ph.D. thesis, Simon Fraser University, Burnaby,
British Columbia. 2004.
34. R. Jackson, B.F.G Johnson, J. Lewis, P.R. Raithby, and S.W.
16. M.R. Churchill and B.G. DeBoer. Inorg. Chem. 16, 878 (1977).
17. G.R. John, B.F.G. Johnson, and J. Lewis. J. Organomet. Chem.
181, 143 (1979).
Sankey. J. Organomet. Chem. 193, C1 (1980).
35. J. Sanchez-Nieves, B.T. Sterenberg, K.A. Udachin, and A.J.
Carty. Can. J. Chem. 82, 1507 (2004).
18. For example: (a) H.B. Davis, F.W.B. Einstein, P.G. Glavina, T.
Jones, R.K. Pomeroy, and P. Rushman. Organometallics, 8,
1030 (1989); (b) F. Jiang, J.L. Male, K. Biradha, W.K. Leong,
R.K. Pomeroy, and M.J. Zaworotko. Organometallics, 17,
5810 (1998); (c) F. Jiang, H.A. Jenkins, K. Biradha, H.B. Da-
vis, R.K. Pomeroy, and M.J. Zaworotko. Organometallics, 19,
5049 (2000).
19. (a) D.M. Hoffman and R. Hoffmann. J. Chem. Soc. Dalton
Trans. 1471 (1982); (b) M.R. Gagné and J. Takats. Organo-
metallics 7, 561 (1988); (c) A. Krebs, A. Jacobsen-Bauer, E.
Haupt, M. Veith, and V. Huch. Angew. Chem. 101, 640 (1989);
(d) A.A. Cherkas, N. Hadj-Bagheri, A.J. Carty, E. Sappa, M.A.
Pellinghelli, and A. Tiripicchio. Organometallics, 9, 1887
(1990); (e) H.A. Mirza, J.J. Vittal, and R.J. Puddephatt.
Organometallics, 13, 3063 (1994); ( f ) C.P. Casey, W.T.
Boese, R.S. Carino, and P.C. Ford. Organometallics, 15, 2189
(1996); (g) C.P. Casey, R.S. Carino, H. Sakaba, and R.K.
Hayashi. Organometallics, 15, 2640 (1996); (h) H.A. Mirza,
J.J. Vittal, and R.J. Puddephatt. J. Coord. Chem. 37, 131
(1996); (i) S. Doherty, M. Waugh, T.H. Scanlan, M.R.J.
Elsegood, and W. Glegg. Organometallics, 18, 679 (1999).
20. M.C. Comstock and J.R. Shapley. Organometallics, 16, 4816
(1997).
36. (a) E.L. Muetterties. Bull. Soc. Chim. Belg. 84, 959 (1975);
(b) E.L. Muetterties, T.N. Rhodin, E. Band, C.F. Brucker, and
W.R. Pretzer. Chem. Rev. 79, 91 (1979).
37. E.L. Muetterties, W.R. Pretzer, M.G. Thomas, B.F. Beier, D.L.
Thorn, V.W. Day, and A.B. Anderson. J. Am. Chem. Soc. 100,
2090 (1978).
38. G.A. Somorjai. Introduction to surface chemistry and cataly-
sis. Wiley, New York. 1994.
39. (a) G.A. Somorjai and M. Yang. Top. Catal. 24, 61 (2003);
(b) G.A. Somorjai. J. Phys. Chem. B, 106, 9201 (2002);
(c) P.S. Cremer, X. Su, Y.R. Shen, and G.A. Somorjai. J. Am.
Chem. Soc. 118, 2942 (1996); (d) G.A. Somorjai. J. Phys.
Chem. B, 106, 9201 (2002).
40. G.V. Smith and F. Notheisz. Heterogeneous catalysis in or-
ganic chemistry. Academic Press, San Diego. 1999.
41. C.R. Henry. Surf. Sci. Rep. 31, 231(1998).
42. (a) A.M. Doyle, S.K. Shaikhutdinov, S.D. Jackson, and H.-J.
Freud. H. Angew. Chem. Int. Ed. 42, 5240 (2003); (b) V.
Johánek, S. Schauermann, M. Laurin, J. Libuda, and H.-J.
Freund. Angew. Chem. Int. Ed. 42, 3035 (2003).
43. (a) J. Wei and E. Iglesia. J. Phys. Chem. B, 108, 4094 (2004);
(b) J. Wei and E. Iglesia. J. Catal. 225, 116 (2004).
44. J. Wei and E. Iglesia. J. Phys. Chem. B, 108, 7253 (2004).
45. J.G. Ulan, W.F. Maier, and D.A. Smith. J. Org. Chem. 52,
3132 (1987).
21. C.P. Casey, R.S. Carino, R.K. Hayashi, and K.D. Schladetzky.
J. Am. Chem. Soc. 118, 1617 (1996).
22. P.F. Heveldt, B.F.G. Johnson, J. Lewis, P.R. Raithby, and G.M.
Sheldrick. J. Chem. Soc. Chem. Commun. 340 (1978).
23. M.R. Burke and J. Takats. Organomet. Chem. 302, C25 (1986).
24. O.P. Anderson, B.R. Bender, J.R. Norton, A.C. Larson, and
P.J. Vergamini. Organometallics, 10, 3145 (1991).
46. For example: (a) J.A. Cabeza, I. del Río, J.M. Fernández-
Colinas, and V. Riera. Organometallics, 15, 449 (1996); (b) R.
Giordano, E. Sappa, and S.A. Knox. J. Cluster Sci. 7, 179
(1996).
47. M. Farina, C. Morandi, E. Mantica, and D. Botta. J. Chem.
Soc. Chem. Commun. 816 (1976).
48. (a) F.W.B. Einstein, V.J. Johnston, A.K. Ma, and R.K.
Pomeroy. Can. J. Chem. 73, 1223 (1995); (b) V.J. Johnston,
F.W.B. Einstein, and R.K. Pomeroy. J. Am. Chem. Soc. 109,
7220 (1987).
25. B.E. Mann and B.F. Taylor. 13C NMR data for organometallic
compounds. Academic Press, New York. 1981.
26. B.F.G. Johnson, J Lewis, P.R. Raithby, and M.J. Rosales. J.
Chem. Soc. Dalton Trans. 2645 (1983).
27. D. Braga, F. Grepioni, B.F.G. Johnson, J. Lewis, and J.A.
Lunniss. J. Chem. Soc. Dalton Trans. 1101 (1992).
28. C.R. Eady, W.G. Jackson, B.F.G. Johnson, J. Lewis, and T.W.
Matheson. J. Chem. Soc. Chem. Commun. 958 (1975).
29. B.F.G. Johnson, J. Lewis, M.-A. Pearsall, and L.G. Scott. J.
Organomet. Chem. 413, 337 (1991).
30. C.R. Eady, B.F.G. Johnson, J. Lewis, R. Mason, P.B. Hitch-
cock, and K.M. Thomas. J. Chem. Soc. Chem. Commun. 385
(1977).
49. V.J. Johnston, F.W.B. Einstein, and R.K. Pomeroy. Organo-
metallics, 7, 1867 (1988).
50. (a) D.D. Shriver, H.D. Kaesz, and R.D. Adams (Editors). The
chemistry of metal clusters. VCH, New York. 1990; (b) B.C.
Gates, L. Guczi, and H. Knözinger (Editors). Metal clusters in
catalysis. Elsevier, Amsterdam. 1986.
31. V.M. Hansen, R.J. Batchelor, F.W.B Einstein, J.L. Male, and
R.K. Pomeroy. Organometallics, 16, 4875 (1997).
© 2006 NRC Canada