Synthesis and structure of two tetranuclear osmium carbonyl isotopomers: A crystallographic isotope effect

Article abstract of DOI:10.1039/b711872d

A synthetic route to [Os₄(μ-H)(μ-OH)(μ-CO)(CO) ₁₂] (1) has been devised through the activation of [Os ₄(CO)₁₄] with Me₃NO. The pyrolysis and photolysis of the reactant in the presence of a trace amount of water produces 1 in low yield. The solid-state structure of [Os₄(μ-H)(μ-OH) (μ-CO)(CO)₁₂?H₂O] (1?H₂O) reveals a butterfly Os₄ skeleton with bridging H, OH and CO ligands as well as hydrogen-bonded molecules of water in the crystal lattice. A low-temperature ¹³C{¹H} NMR spectroscopic study revealed a merry-go-round exchange of CO ligands around the Os₃ plane containing the asymmetric bridging CO. The exposure of 1?H₂O to D₂O yielded [Os₄(μ-H)(μ-OD)(μ-CO)(CO)₁₂] (2). Although the solid-state, intramolecular structure of 2 closely matched that of 1?H ₂O, the intermolecular structure did not: its crystal lattice contained no water of crystallization, a previously unreported crystallographic isotope effect. The Royal Society of Chemistry.

Full text of DOI:10.1039/b711872d

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Synthesis and structure of two tetranuclear osmium carbonyl isotopomers:
a crystallographic isotope effect†‡
John P. Canal,*a Michael Jennings,b Glenn P. A. Yapc and Roland K. Pomeroya
Received 17th August 2007, Accepted 13th December 2007
First published as an Advance Article on the web 6th February 2008
DOI: 10.1039/b711872d
A synthetic route to [Os4(l-H)(l-OH)(l-CO)(CO)12] (1) has been devised through the activation of
[Os4(CO)14] with Me3NO. The pyrolysis and photolysis of the reactant in the presence of a trace amount
of water produces 1 in low yield. The solid-state structure of [Os4(l-H)(l-OH)(l-CO)(CO)12·H2O]
(1·H2O) reveals a butterfly Os4 skeleton with bridging H, OH and CO ligands as well as
1
hydrogen-bonded molecules of water in the crystal lattice. A low-temperature 13C{ H} NMR
spectroscopic study revealed a merry-go-round exchange of CO ligands around the Os3 plane
containing the asymmetric bridging CO. The exposure of 1·H2O to D2O yielded
[Os4(l-H)(l-OD)(l-CO)(CO)12] (2). Although the solid-state, intramolecular structure of 2 closely
matched that of 1·H2O, the intermolecular structure did not: its crystal lattice contained no water of
crystallization, a previously unreported crystallographic isotope effect.
1, along with [Os5(CO)19] 11 and [Os3(CO)12].12 Low yields of
Introduction
1 and [Os6(CO)18] 13 with larger amounts of [Os5(CO)16] 14 and
Although the hydrogen/deuterium isotope effect on hydrogen
[Os3(CO)12] were isolated from the pyrolysis of [Os4(CO)14] at
125 ◦C (24 h). The pyrolysis of lower-nuclearity osmium carbonyl
bonds has been the subject of much research, the focus has
largely been on organic compounds.1 The results of these studies
have shown that H/D substitution generally produces only small
changes in the molecular geometry. This is referred to as the
Ubbelohde effect.2 Unlike the system presented here, [(Os4(l-
clusters continues to be the most common method of synthesizing
higher nuclearity clusters.12,15 Preparation of compound 1 in
a higher yield through chemical activation of the tetranuclear
carbonyl cluster allowed for the discovery of its interesting
H)(l-OH)(l-CO)(CO)12·H2O] (1·H2O) and [Os4(l-H)(l-OD)(l-
characteristics.
CO)(CO)12] (2), the molecular structures of H/D isotopomers
Herein we describe the synthesis and characterization of two
isotopomers. The molecular structure of the first contains a water
of crystallization hydrogen-bonded to a carbonyl and a hydroxyl
ligand, while the second does not. The first compound exhibits an
unusual and reversible loss of water of crystallization upon H/D
substitution.
usually do not differ significantly.1e–i,3,4 An exception appears to be
the n(H2O)·m(HF) and n(D2O)·m(DF) (n = 1, m = 1, 2, 4) system
in which an additional deuterated compound [2(D2O)·3(DF)] has
been isolated which bears no protonated analogue.5
The solid-state structure of 1·H2O contains a water molecule
of crystallization hydrogen-bonded to a carbonyl and a hydroxyl
ligand. A previous search of the Cambridge Structural Database
(CSD)6 retrieved only 25 crystalline hydrates with a water molecule
Results and discussion
bound to carbonyl ligands of transition metals.7–9 No studies on
[Os4(l-H)(l-OH)(l-CO)(CO)12] (1)
the effect of substitution of H2O with D2O have been reported for
these compounds. The system presented here appears to be the
first example of a hydrated metal carbonyl complex examined in
both the protonated and the deuterated form.
In the presence of a minute amount of water, the UV photolysis
of [Os4(CO)14] 10 in C6F6 (12 h) yielded a trace amount of
The synthesis of the new cluster [Os4(l-H)(l-OH)(l-CO)(CO)12]
(1) is possible by three different methods. (i) Dropwise addition
of Me3NO (in dry CH2Cl2) to [Os4(CO)14] in dry CH2Cl2 at room
temperature afforded 1 in 40% yield after column chromatography
(eqn (1a)). (ii) The solid-state or solution (dry C6F6) pyrolysis of
[Os4(CO)14] at 125 ◦C (24 h) in vacuo in a flame-dried Carius tube
resulted in the isolation of [Os5(CO)16] 14 and [Os3(CO)12] 12 as well
as a trace amount of 1 (eqn (1b)) after column chromatography.
A trace amount of [Os6(CO)18] 13 was also obtained, though this
was due to a subsequent self-reaction of [Os3(CO)12]. (iii) The
solution (dry C6F6) UV photolysis of [Os4(CO)14] (12h) at room
temperature in vacuo in a flame-dried quartz Carius tube yielded
[Os5(CO)19] 11 and [Os3(CO)12] as well as a trace amount of 1 (eqn
(1c)) after column chromatography.
aDepartment of Chemistry, Simon Fraser University, Burnaby, British
Columbia, Canada V5A 1S6. E-mail: jcanal@sfu.ca; Fax: +1 (1)778
782 3765; Tel: +1 (1)778 782 7661
bDepartment of Chemistry, University of Western Ontario, London, Ontario,
N6A 5B8, Canada
cDepartment of Chemistry and Biochemistry, University of Delaware,
Newark, DE, 19716, USA
† CCDC reference numbers 657923 and 667076. For crystallographic data
in CIF or other electronic format see DOI: 10.1039/b711872d
‡ Electronic supplementary information (ESI) available: Complete
IR(KBr) spectrum of 1·H2O and the 1H NMR of 1 and 2. Tabular form of
the crystallographic details of 1·H2O and 2 are provided, as well as the CIF
file giving crystallographic information for 2. See DOI: 10.1039/b711872d
(1a)
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 1375–1382 | 1375
©

View Article Online
(1b)
(1c)
Although the source of the trace amount of water is not known
(possibly incomplete drying of glassware and solvent, or water
trapped in and on reagent crystals), its presence is crucial to the
formation of 1.§ Under exceedingly dry conditions16 the pyrolysis
of [Os4(CO)14] (reaction (1b)) failed to yield 1, yet the addition
of a drop of water failed to produce 1. (Isolated from the first
reaction were [Os5(CO)19] and [Os3(CO)12] while the latter reaction
gave [Os4H2(CO)13] and [Os3(CO)12].) The effect of the addition of
water on reaction (1a) was also examined, but this proved to be
unproductive due to the interaction of water with Me3NO. Me3NO
is dried prior to its use. The addition of water negated this step
and reduced the effectiveness of Me3NO. The separation of the
products by column chromatography did not affect the formation
of 1. By IR spectroscopy, 1 was observed in the product mixture
prior to its isolation. Also, isolation via column chromatography
under both wet and dry conditions gave identical results.
Compound 1·H2O was isolated as air- and moisture-stable,
yellow crystals by crystallization from dry CH2Cl2 (see later for a
discussion of the crystallization method). The molecular structure
of 1·H2O (Table 1, Fig. 1) has a butterfly arrangement of Os atoms
with the O atom of the OH group bridging the wingtip atoms.
This type of structure is also found in [Os4(l-H)3(l-X)(CO)12]
(X = OH, OPO3H2, NO),17 [Os4(l-H)3(l-I)(CO)12] 18 and [Os4(l-
H)4(l-OH)(CO)12]+.19 The Os–Os bond lengths of 1·H2O range
Fig. 1 The molecular structure of [Os4(l-H)(l-OH)(l-CO)(CO)12·H2O]
(1·H2O). The O(15) is from the H2O of crystallization; H atoms are omitted;
a hydrogen ligand bridges the Os(1)–Os(2) bond.
˚
from 2.8170(4) to 2.8569(4) A (Table 2) if the long Os–Os bond
˚
length of the Os–(l-H)–Os unit (i.e. Os(1)–Os(2) = 2.9701(5) A)
is excluded. The average Os–Os bond length of the short Os–Os
˚
bonds in 1·H2O is 2.839 A which is comparable to the average Os–
20
˚
Os length of 2.877 A in [Os3(CO)12]. The Os–Os bond distances
are usually shorter in condensed Os clusters than in [Os3(CO)12].
Table 2 Bond lengths (A) and angles (◦) for [Os4(l-H)(l-OH)(l-
˚
CO)(CO)12·H2O] (1·H2O) and [Os4(l-H)(l-OD)(l-CO)(CO)12] (2)
Table 1 Details of the crystal structures of [Os4(l-H)(l-OH)(l-
CO)(CO)12·H2O] (1·H2O) and [Os4(l-H)(l-OD)(l-CO)(CO)12] (2)
˚
Bond lengths (A)
1·H2O
2
1·H2O
2
Os(1)–Os(2)
Os(1)–Os(3)
Os(2)–Os(3)
Os(2)–Os(4)
Os(3)–Os(4)
Os(3)–C(13)
Os(4)–C(13)
C(13)–O(13)
Os–COterm
2.9701(5)
2.8343(5)
2.8170(4)
2.8466(4)
2.8569(4)
1.963(8)
2.9482(8)
2.8309(8)
2.8188(8)
2.8547(8)
2.8288(7)
1.99(1)
Empirical formula
Formula weight
Color
C13H4O15Os4
1160.96
Yellow
Monoclinic
C2/c
29.882(2)
9.4247(7)
15.274(1)
90
105.227(1)
90
4146.3(6), 8
3.720
24.512
4932/4932
0.000
0.0358
C13HDO14Os4
1142.95
Yellow
Monoclinic
P2(1)/n
9.2448(2)
14.2439(3)
15.2035(4)
90
97.04(1)
90
1986.93(8), 4
3.821
Cryst. system
Space group
˚
a/A
2.472(7)
1.152(9)
1.864(9)–1.959(8)
1.124(9)–1.15(1)
2.123(5)
2.37(1)
1.18(2)
˚
b/A
˚
c/A
1.86(1)–1.94(2)
1.11(2)–1.17(2)
2.122(9)
2.11(1)
a/◦
b/◦
c /◦
(C–O)term
Os(1)–O(14)
Os(4)–O(14)
Os(1)–Ha
2.132(5)
1.75a
1.80a
3
˚
V/A , Z
1.76a
1.85a
D(calcd)/Mg m−3
Abs. coefficient/mm−1
Reflections collected/unique
R(int)
Os(2)–H
O(14)–O(15)
2.669(8)
25.568
24430/3632
0.121
0.0448
0.1029
Bond angles (◦)
1·H2O
2
a
Os(2)–Os(1)–Os(3)
Os(1)–Os(2)–Os(3)
Os(1)–Os(2)–Os(4)
Os(3)–Os(2)–Os(4)
Os(1)–Os(3)–Os(2)
Os(1)–Os(3)–Os(4)
Os(2)–Os(3)–Os(4)
Os(2)–Os(4)–Os(3)
Os(3)–Os(4)–C(13)
Os(4)–Os(3)–C(13)
Os(3)–C(13)–Os(4)
Os(1)–O(14)–Os(4)
58.01(1)
58.58(1)
72.42(1)
60.58(1)
63.41(1)
74.31(1)
60.22(1)
59.20(1)
42.5(2)
58.34(2)
58.75(2)
72.52(2)
59.81(2)
62.91(2)
74.68(2)
60.72(2)
59.47(2)
43.9(3)
55.7(4)
80.4(5)
108.4(4)
RF
b
RwF
0.0633
b
a RF = R|(|FO| – |FC|)|/R|FO|. RwF = [R(w(|FO| – |FC|)2)/
R(wFO2)]1/2. w = [r2(FO)2 + kFO
]
.
2
−1
58.2(2)
79.3(2)
107.8(2)
§ Please note that this paper discusses two different kinds of water
molecules: (i) the first is the trace amount of water which participates
in the reactions and is split, producing the bridging H and OH groups
found in the cluster; (ii) the other is the molecule of water found as solvent
of crystallization in 1·H2O, but not in the molecular structure of 2.
a Determined with the XHYDEX program.21
1376 | Dalton Trans., 2008, 1375–1382
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Placement of the H ligand across the Os(1)–Os(2) bond gave a
site energy of 2.4 kcal mol−1 by the XHYDEX program.21 This
value is in the range of 0.2–10.6 kcal mol−1 that was previously
found for bridging hydride ligands. The Os(1)–Os(2) vector at
hydrogen bond length to the second hydroxyl ligand on the
˚
neighboring cluster (O(14A)–O(15)) is longer at 2.745(11) A. The
water molecule does not hydrogen-bond to any terminal carbonyls.
This is expected, as there is an increase in the metal-to-CO p back-
donation on going from a terminal to a bridging carbonyl as
evidenced by their respective CO stretching frequencies (discussed
below). Inother words, there is more electrondensity onthe oxygen
atom of a bridging CO, thus a greater preference for H-bonding
relative to a terminal CO.25
˚
2.9701(5) A is by far the longest Os–Os distance in the molecule,
and this is entirely consistent with the view that this bond is H-
˚
bridged. The distance may be compared with 2.985 A, the average
Os–Os bond length for the Os–(l-H)–Os units in [Os4(l-H)3(l-
I)(CO)12] (by neutron diffraction)18 and [Os4(l-H)3(l-X)(CO)12]
(X = OH, OPO3H2, NO).17
1
The H NMR spectrum of 1·H2O in dry CD2Cl2 contains a
The structure of 1·H2O also contains a semi-bridging carbonyl
(CO(13)), which bridges the Os(3)–Os(4) bond (Os(3)–C(13) =
singlet at d 1.5 ppm due to water, as well as singlets at d −2.8
and −13.8 ppm which are assigned to the OH and Os–(l-H)–Os
protons respectively. These resonances are in the normal range for
such groupings.26 After exposure to D2O the resonance at d −2.8
disappeared, which is consistent with the signal being due to the
proton of the OH ligand. As expected, the NMR signal due to the
hydride was not altered, indicating that exchange with deuterium
did not occur.
˚
˚
1.963 (8) A; Os(4)–C(13) = 2.472 (7) A). These distances are
similar to the Os–C bond lengths of the asymmetric bridging CO
in [Os5(CO)18] and [(g -C5Me5)Os4(l-H)(CO)11].22,23
5
The H2O of crystallization in 1·H2O bridges three clusters.
The water molecule is hydrogen-bonded to hydroxyl ligands on
two adjacent clusters and the bridging CO on a third (Fig. 2).
The interaction of water with the bridging CO (O(15)–O(13B))
The 13C{ H} NMR spectrum of 1·H2O at −50 ◦C in dry
1
˚
supports the longest hydrogen bond at 2.944(11) A (Table 3),
CD2Cl2 showed six signals in the approximate intensity ratio
7 : 1 : 1 : 1 : 1 : 2 which is consistent with the solid-state molecular
structure (Fig. 3). The signal of intensity 7 is assigned to the
carbonyls labelled a which are undergoing exchange in the Os(2)–
Os(3)–Os(4) plane (Chart 1). This plane includes the bridging
CO. Clusters often undergo a rapid merry-go-round CO exchange
in the plane containing a bridging CO.27 The signals at d 175.4
(JCH = 12.2 Hz) and 173.1 ppm (JCH = 11.4 Hz), both with
intensity 1, showed coupling to the bridging H atom in the 13C
NMR spectrum. Thus, they were assigned to the carbonyls labelled
b and e. The highest field resonance at d 170.6 ppm with intensity
2 is believed to be the result of the overlap of two signals. One
signal is assigned to d while the other resonance is due to one of c,
˚
which matches the average length (2.94(18) A) reported for these
bonds.8 The shortest hydrogen-bonding interaction in 1·H2O is
˚
between a hydroxyl ligand and water (O(14)–O(15)) at 2.669(8) A.
As seen in other three-centered hydrogen-bonded systems,24 the
˚
Table 3 The O–O bond lengths (A) of the O · · · H–O hydrogen bonds in
1·H2O
˚
Bond lengths (A)
Bond type
O(14)–O(15)
O(15)–O(14A)
O(15)–O(13B)
2.669(8)
2.745(10)
2.944(11)
(l-OH)–(OH2)
(l-OH)–(OH2)
(l-CO)–(OH2)
Fig. 2 Hydrogen bonding interactions in [Os4(l-H)(l-OH)(l-CO)(CO)12·H2O] (1·H2O). Hydrogen atoms are omitted. O(15) and O(15A) are from the
water of crystallization.
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 1375–1382 | 1377
©

View Article Online
carbonyl labelled c is static. The carbonyls attached to an Os atom
with a bridging ligand (other than CO) do not normally undergo
mutual CO exchange.28 The only CO ligand available for exchange
with carbonyls a is carbonyl d. This occurs through a mutual
exchange of the carbonyls on Os(3).
At 50 ◦C all the signals in the 13C{ H} NMR spectrum (dry
1
toluene-d8) had broadened and at 80 ◦C they had essentially
collapsed into the baseline (Fig. 3). The sharp signals in the latter
spectrum are attributed to decomposition products that may be
due to reaction of 1·H2O with the solvent (toluene-d8) at this
temperature. (When the sample was cooled to room temperature
and the NMR spectrum was again obtained, the spectrum was
consistent with both the decomposition product and 1·H2O.)
Other mechanisms of CO exchange are therefore taking place
at elevated temperatures rendering the carbonyls equivalent. This
may involve mutual exchange of the carbonyls of the Os(CO)3
groupings.
[Os4(l-H)(l-OD)(l-CO)(CO)12] (2)
1
Fig. 3 Variable temperature 13C{ H} NMR spectra of [Os4(l-H)(l-
OH)(l-CO)(CO)12·H2O] (1·H2O) (−50 ◦C in dry CD2Cl2; 21 ◦C to 80 ◦C in
dry toluene-d8; the spectra done at 21 ◦C in CD2Cl2, which are very similar
to those done in toluene-d8, are not shown; * decomposition products).
The partially deuterated sample of 1, i.e. [Os4(l-H)(l-OD)(l-
CO)(CO)12] (2), was prepared as follows. A powdered sample of
1·H2O was dried under vacuum for one day. It was then dissolved
in dry CH2Cl2, an excess of D2O was added, and the solution was
stirred for 5 days under nitrogen (eqn (2)).
(2)
The CH2Cl2 solution containing 2 was transferred from the
excess D2O via a pipette to another Schlenk tube and the
solvent was removed on the vacuum line. The resulting solid was
crystallized from wet (D2O) CH2Cl2. CH2Cl2 was dried prior to
its exposure to D2O) to give air-stable, yellow crystals of [Os4(l-
H)(l-OD)(l-CO)(CO)12] (2).
As expected, the molecular structure of 2 (Table 1, Fig. 4) was
essentially identical to that of 1·H2O. Its most remarkable feature,
however, is the absence of a water (D2O) of crystallization. (The
discussion of this finding is left until a later section.) Placement
of the H ligand across the Os(1)–Os(2) bond gave a site energy
1
Chart 1 Labelling in the 13C{ H} NMR spectrum of 1·H2O. (A)
Carbonyl exchange at Os(3). (B) Merry-go-round CO exchange about
the Os(2)–Os(3)–Os(4) triangle of 1·H2O.
f or g. (An explanation for the assignment of d is given in the next
paragraph.) The final two resonances at d 178.2 and 175.4 ppm,
both with intensity 1, are assigned to two of c, f or g.
As the temperature of the sample was raised, the changes in the
spectrum indicated that other exchange mechanisms were taking
place. At room temperature (Fig. 3) the spectrum run in dry
toluene-d8 or CD2Cl2 showed a similar pattern to the spectrum
obtained at −50 ◦C except the intensity ratio had changed to
7 : 1 : 1 : 1 : 1 : 1 (from 7 : 1 : 1 : 1 : 1 : 2) and the lowest field signal
had become broader, indicating that exchange of the carbonyls
labelled a with another carbonyl (or carbonyls) had commenced.
The highest field signal at d 170.6 ppm became sharper but
reduced in intensity from 2 to 1. Since the other signals remained
sharp this indicated that the carbonyls labelled a were now
exchanging with one of the signals originally at d 170.6 ppm.
The resonances assigned to e and b remained sharp at room
temperature indicating that these carbonyls are not undergoing
exchange. Because the carbonyls on Os(1) can only undergo a
mutual exchange and e is stationary, the carbonyls labelled f and g
also do not exchange. Because of the bridging OH group, a mutual
exchange of the carbonyls on Os(4) also does not occur; thus the
Fig. 4 The molecular structure of [Os4(l-H)(l-OD)(l-CO)(CO)12] (2).
1378 | Dalton Trans., 2008, 1375–1382
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
of 2.2 kcal mol−1 (1·H2O: 2.4 kcal mol−1) as calculated by the
XHYDEX program.21 The bond lengths and angles found in the
two structures are similar (Fig. 1 and 4) with average Os–Os
promotes the conversion to 2 while the addition of H2O to 2
produces 1·H2O.
˚
˚
bond lengths of 2.865 A in 1·H2O and 2.856 A in 2 (Table 2).
Examination of the Os–C bond lengths of the bridging carbonyl
in 1·H2O and 2 shows that the bridging CO of 2 is somewhat more
symmetric (i.e. less terminal character) than that of 1·H2O (Os–C
(5)
Comparison of the IR (KBr) spectra of 1·H2O and 2
˚
˚
bond lengths: 1·H2O: 1.963(8), 2.472(7) A; 2: 1.99(1), 2.37(1) A).
The removal of water from the lattice changed the space group
of the sample from C2/c (for 1·H2O) to P2(1)/n (for 2) with an
increase in density (1·H2O: 3.72 Mg m−3; 2: 3.82 Mg m−3).
The IR (KBr) spectra of 1·H2O and 2 are of particular interest
(Fig. 5). The spectrum of 1·H2O contains terminal m(CO) stretches
in the region 2083–1958 cm−1. It also has a band of medium
intensity due to the semi-bridging CO at 1844 cm−1. The terminal
m(CO) stretches are also present in the spectrum of 2, but the
peak due to the semi-bridging CO is not (Fig. 5B). No extra band
attributable to the bridging carbonyl was visible in the spectrum
of 2.
Crystallization results and the effect of H2O–D2O
Suitable crystals of 1·H2O and 2 were grown by slow cooling of
concentrated solutions in CH2Cl2. As in the formation of these
isomers, the presence of water is also critical for the development of
crystals of 1·H2O. All attempts to form a crystal of “dry” 1 failed.
Crystalline material of 1·H2O was only obtained in the presence
of traces of water. The reverse occurred in the crystallization of
2. All attempts to form a crystal of “wet” 2, i.e. a crystal of
2·D2O, failed although crystallization was attempted in both wet
(D2O) and dried CH2Cl2. Both conditions only gave crystals of
2, thus the presence of water did not affect its formation. Other
than to suggest that the thermodynamics favor the formation of
1·H2O and 2 it is not clear why crystalline 1 and 2·D2O were not
obtained.
Interconversion of 1·H2O and 2
The reversibility of the reaction represented by eqn (2) was
examined by the addition of a few drops of H2O to a solution
of 2 in dry CH2Cl2 (eqn (3)). The solution was stirred for one day
before the solvent was removed on the vacuum line.
Fig.
5
IR (KBr, m(CO)) spectra of (A) [Os4(l-H)(l-OH)(l-
CO)(CO)12·H2O] (1·H2O) and (B) [Os4(l-H)(l-OD)(l-CO)(CO)12] (2)
(3)
(* 2040 cm−1, ** 1844 cm−1).
Terminal carbonyl ligands bound to transition metals have CO
stretching frequencies in the 1900–2150 cm−1 region, while bridg-
ing carbonyl stretches appear in the region of 1750–1850 cm−1.30
The m(CO) stretch of a semi-bridging CO ligand can appear in
either the terminal or the bridging region (depending on the degree
of terminal and bridging character).30 In order to predict whether
the CO stretch of 2 should be at a higher or lower frequency
than that of 1·H2O several factors must be considered such as
the degree of bridging character of the carbonyls (i.e. occupancy
of the p* MOs) and the effect of hydrogen bonding on the CO
bond strength. It has been reported31 that hydrogen bonding
interactions with a carbonyl result in a bathochromic shift of the
CO stretching frequency, as observed in the IR spectrum of 1·H2O
when compared to that of 2. The bridging CO stretch in 2 thus
appears to have shifted and overlapped the terminal CO stretches
of 1·H2O. Although the CO bond of the semi-bridging carbonyl
The IR spectrum (KBr disk) of the remaining solid was consistent
with that of 1·H2O. When a solution of 1·H2O in dry CH2Cl2
containing equal amounts of H2O and D2O (‘DHO’) was allowed
to stir for one day (eqn (4)) and the solvent was removed on the
vacuum line, the IR spectrum of the residue (see later discussion)
was consistent with that of 2.
(4)
This result does not negate the possibility that a small amount of
1·H2O or a new mixed isotope product was also present; it means
only that the major product appears to be 2, that is, the exposure
of 1·H2O to H2O–D2O (or HDO) appears to give the form that
does not contain water in the solid-state lattice. This suggests that
HDO exhibited a greater deuterium character.29
The effect of D2O, H2O and a 1 : 1 mixture of both on the
conversion of 1·H2O into 2 (and vice versa) is summarized in
eqn (5). The addition of D2O or H2O–D2O (1 : 1) to 1·H2O
˚
˚
in 2 (1.18(2) A) is longer than that in 1·H2O (1.152(9) A) the
difference is not statistically significant.
The IR spectrum of 1·H2O (in hexanes), which is very similar to
that of 2, does not contain stretching frequencies in the bridging
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 1375–1382 | 1379
©

View Article Online
1
1
carbonyl region. There is also no signal in the 13C{ H} NMR
method.10 Samples used for 13C{ H}/13C NMR spectra were
enriched to ∼25–30% with 13CO and were prepared from 13CO-
enriched [Os4(CO)14]. The 13CO-enriched [Os4(CO)14] was made
from 13CO-enriched [Os3(CO)12].36 Unless explained otherwise
solvents were dried based on literature methods and stored over
molecular sieves.37
spectrum that can be assigned to this carbonyl as it is undergoing
rapid exchange with other (terminal) CO ligands. It may be that,
as previously proposed for [Ru6(l6-C)(l-CO)(CO)16], [Os4(CO)14]
and [Fe3(CO)12],32 the exchange is taking place on the IR time scale
at ambient temperature, and the CO stretch of the semi-bridging
carbonyl is broadened into the baseline.
There is another difference between the IR (KBr) spectra of
1·H2O and 2. The spectrum of 1·H2O contains a sharp m(OH) band
at 3636 cm−1 which is not observed in the spectrum of 2. If this
m(OH) stretch is due to the bridging OH group, a bathochromic
shift (to ∼2570 cm−1) would be expected in the IR spectrum of
the deuterated sample (which contains a bridging OD unit).9h
Since the band did not shift but disappeared, the observed m(OH)
stretch must be due to the water of crystallization in 1·H2O. The
disappearance of the m(OH) band in 2 is consistent with a crystal
structure that does not contain water in the lattice.
Preparation of [Os4(l-H)(l-OH)(l-CO)(CO)12·H2O] (1·H2O)
A solution of [Os4(CO)14] (20 mg, 1.7 × 10−3 mmol) in dry CH2Cl2
(15 mL) was stirred at room temperature while 0.26 mL of a
0.0405 M solution of Me3NO in dry CH2Cl2 was added dropwise
over the course of 5.25 h via a syringe. During this time the
original brown–red solution darkened slightly. The solution was
then evaporated to dryness on the vacuum line and the remaining
solid subjected to chromatography on a silica gel column (1.5 ×
16 cm). The side products ([Os3(CO)12], [Os4(l-H)2(CO)13]) were
first removed by elution with hexanes–CH2Cl2 (50 : 50). Elution
with hexanes–CH2Cl2 (10 : 90) gave a yellow band that contained
the desired product. An analytical sample of 1·H2O (10 mg, 40%)
was obtained as air-stable, yellow crystals by crystallization from
dry CH2Cl2 (as discussed above).
Comparison of 1·H2O and 2: an isotope effect
The difficulty in understanding the 1·H2O/2 system is that it is
unique and consequently it cannot be compared to any other
Os cluster systems. Most osmium carbonyl compounds have soft
ligands with atoms (e.g. P and S) that do not engage in hydrogen
bonding. It has only recently become apparent that Os carbonyl
clusters with hard ligands are stable, unlike their first- and second-
row transition metal analogues.33
Found: C, 13.66; H, 0.20%. Calc. for C13H2O14Os4: C, 13.65;
H, 0.18%. (Elemental analysis was done on a powdered sample
of 1·H2O, kept under an active vacuum for 2 days.). IR (hexanes,
m(CO)/cm−1, baseline 97%):2108 (w), 2078(vs), 2058 (vs), 2053 (vs,
13%), 2021 (m), 2013 (s), 2001 (m); IR (KBr pellet, m(OH)/cm−1,
baseline 60%): 3636 (m, sharp, 33%), ∼3460 (w, vbr) cm−1,
(m(CO)/cm−1) 2083 (sh, w), 2053 (sh), 2040 (s, 8%), 2009 (s),
1991 (s), 1977 (s), 1958 (s), 1844 (br, m) (semi-bridging CO)
(intensities (% transmission relative to the baseline) of a select
number of peaks are also provided); dH (400 MHz, dry CD2Cl2,
ppm) 1.5 (s), −2.8 (s), −13.8 (s) (addition of D2O to the NMR
solution caused the disappearance of the peak at d −2.8 (s)); dC
With this in mind we propose the following explanation. Unlike
most metal carbonyls 1·H2O contains two ligands that can form
strong hydrogen bonds with water. First, 1·H2O has a semi-
bridging CO, which is able to produce stronger H-bonds than
terminal carbonyls. The OH ligand also forms strong H-bonds.
Additional hydrogen bonds usually stabilize a system.34 In 1·H2O
this extra stability is enough for the hydrated form to be favored.
Although the effect of the hydrogen bond strength in the H and D
samples depends on the system, it appears that in 2 hydrogen
bonding interactions are weaker. It has been reported that in
a O · · · H–O hydrogen-bonded system deuterium substitution
expands the O–O distance only if it was initially between ∼2.4
◦
1
(13C/13C{ H}, 400 MHz, dry CD2Cl2, −50 C, ppm) 184.3 (7C),
178.2 (1C), 175.4 (1C), 174.4 (1C, JCH = 12.2 Hz), 173.1 (1C,
JCH = 11.4 Hz), 170.6 (2C); dC (13C{ H}, 400 MHz, dry CD2Cl2,
1
RT, ppm) 183.6 (7C), 177.8 (1C), 175.0 (1C), 174.2 (1C), 173.0
(1C), 170.5 (2C); dC (13C{ H}, 400 MHz, dry toluene-d8, RT) 184.1
1
(7C), 178.6 (1C), 175.8 (1C), 175.2 (1C), 173.9 (1C), 171.2 (2C);
and 2.65 A or greater than ∼ 2.85 A.1d Given the O–O distances of
˚
˚
˚
dC (13C{ H}, 400 MHz, dry toluene-d8, 50 ◦C, ppm) 183.4 (7C,
1
2.669(8), 2.745(10) and 2.944(11) A measured for 1·H2O, it appears
broad), 178.5 (1C), 175.7 (1C), 175.1 (1C), 173.9 (1C), 171.1 (1C);
that two bonds have weakened due to deuterium substitution. In
other words, since the deuterium bonding in 2 is weaker than
the hydrogen bonding in 1·H2O, it is not sufficient to bind the
D2O in the lattice of 2. Because transitional metal clusters are
flexible molecular systems, small changes in the intermolecular
interactions can affect the crystal structure.35 The hydrogen form
(1·H2O) appears to be on the cusp of favoring the hydrated
structure since the small decrease in the H-bond strength of the
deuterated sample (2) is enough to cause loss of D2O (or DHO).
◦
dC (13C{ H}, 400 MHz, dry toluene-d8, 80 C, ppm) 182.0 (weak,
broad), 176.5 (weak, broad); MS (LSIMS) m/z 1143.7 (M+, 100%)
(calcd for M+ = 1144 (100%)).
1
Trace amounts of 1 were also produced by pyrolysis or
photolysis of [Os4(CO)14]: a flame-dried Carius tub◦e with solid
[Os4(CO)14] (25 mg, 0.022 mmol) was heated at 125 C for 24 h.
An IR spectrum of the solid in hexanes showed stretches that
indicated the presence of [Os3(CO)12] and [Os5(CO)16] as well as
traces of 1 and [Os6(CO)18]. (13C NMR analysis of the product
mixture showed that [Os3(CO)12] and [Os5(CO)16] were formed
in a 1 : 1 molar ratio.) The same products were isolated when
the pyrolysis was carried out in 10 mL of dry degassed C6F6.
Upon completion of the reaction the mixture was subjected to
chromatography on a silica gel column (1.5 × 16 cm). Elution with
hexanes gave [Os3(CO)12] (∼5 mg, ∼50%), [Os5(CO)16] (∼8 mg,
∼50%) and [Os6(CO)18] (trace amount). Subsequent elution with
hexanes–CH2Cl2 (10 : 90) also gave traces of product 1.
Experimental
All reactions were carried out under nitrogen using standard
Schlenk techniques. The trimethylamine oxide was purified by
sublimation of Me3NO·2H2O at 70 ◦C under vacuum in the
presence of P2O5 to ensure complete removal of water. The starting
tetranuclear cluster, [Os4(CO)14], was prepared by the literature
1380 | Dalton Trans., 2008, 1375–1382
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
In another experiment [Os4(CO)14] (10 mg, 0.0087 mmol) and
dry C6F6 (10 mL) were added to a flame-dried quartz Carius
tube. The solution was degassed with three freeze–pump–thaw
cycles and the vessel sealed under vacuum. It was subjected to
ultraviolet irradiation overnight (200 W Hanovia lamp; ∼3 cm
between source and edge of reaction vessel). The solvent was
removed on the vacuum line and the remaining solid was subjected
to chromatography on a silica gel column (1.5 × 16 cm). Elution
with hexanes gave [Os3(CO)12] and [Os5(CO)19] (not in a 1 : 1 molar
ratio). Subsequent elution with hexanes–CH2Cl2 (10 : 90) gave a
trace amount of 1.
DENZO.40 The data were scaled using SCALEPACK.40 The
reflection data and systematic absences were consistent with a
monoclinic space group: P2(1)/n.
The SHELXTL-NT V6.138 suite of programs was used to solve
the structure by direct methods. Subsequent difference Fourier
syntheses allowed the remaining atoms to be located. The cluster
was quite well behaved. The heavy atoms were refined with
anisotropic thermal parameters. The carbonyl groups and the
hydroxyl group were refined isotropically. There was one hydrogen
and one deuterium atom in the final structure. The hydride
bridging Os(1) and Os(2) was constrained to have equal Os–H
˚
distances, and that distance was refined to a value of 1.91 A. The
hydrogen attached to the bridging oxygen was constrained to be
along the vector to its nearest neighbour, O(4). The O–H distance
Preparation of [Os4(l-H)(l-OD)(l-CO)(CO)12] (2)
˚
A sample of powdered 1·H2O (25 mg, 0.022 mmol) in a Schlenk
tube was placed on the vacuum line for one day before being
dissolved in thoroughly dried CH2Cl2 (∼15 mL) (dried with P2O5).
D2O (∼5 mL) was added and the mixture was stirred under N2
for 5 days. The CH2Cl2 solution was then transferred by pipette
to another Schlenk tube (leaving behind the D2O) and evaporated
on the vacuum line. The resulting solid was crystallized from dry
CH2Cl2 that had been exposed to D2O, producing air-stable, yellow
crystals of 2 in essentially quantitative yield.
A sample of 2 was also obtained by dissolving 1·H2O (5 mg, 4 ×
10−3 mmol) in dry CH2Cl2 (8 mL) to which equal amounts (1 mL)
of H2O and D2O were added. The solution was allowed to stir for
one day. The sample was dried on the vacuum line to afford 2 as
indicated by its IR (KBr) spectrum.
was constrained to 0.84 A.
−3
˚
The largest residue electron density peak (3.062 e A ) was
associated with one of the osmium atoms (Os(3)). Full-matrix
least-squares refinement on F2 gave R1 = 4.48 for 2r data and
wR2 = 10.90 for all data (GOOF = 1.011).
Conclusions
This work describes the synthesis of two clusters, 1·H2O and 2,
which are interconverted by their exposure to H2O or D2O. These
clusters exhibit a crystallographic isotope effect in which a water of
crystallization was only incorporated in the lattice of 1·H2O, thus
1·H2O contains a hydrogen bonding network that is not present in
2. It is not clear why the isolation of crystalline 1 and 2·D2O was
unsuccessful. The molecular structures of 1·H2O and 2 revealed
a butterfly Os4 skeleton with a bridging hydride, hydroxide and
carbonyl group. These clusters provide a rare example of an isotope
effect on the solid-state lattice of isotopomers.
The IR (KBr) spectrum of 2 was the same as that of 1·H2O
except that the peaks at 3636 (m) and 1844 (m) cm−1 are missing.
IR (KBr pellet, m(OH)/cm−1, baseline 58%) ∼3460 (w, vbr) cm−1;
m(CO)/cm−1 2083 (sh, w), 2053 (sh), 2040 (s, 30%), 2009 (s), 1991
(s), 1977 (s), 1958 (s).
The 13C NMR data of 1·H2O support a merry-go-round
mechanism for CO exchange, a process that is still under debate.
X-Ray analyses of 1·H2O and 2†
Acknowledgements
A suitable crystal of 1·H2O obtained by cooling a saturated
(dry) CH2Cl2 solution was selected, mounted on thin glass fiber
using paraffin oil and cooled to the data collection temperature
(−70 ◦C). Unit cell parameters were determined from 60 data
frames collected at different sections of the Ewald sphere. Data
were collected on a Bruker AXS SMART 1 k CCD diffractometer
using graphite-monochromated Mo Ka radiation with 0.3◦ x-
scans at 0, 90, and 180◦ in φ. The unit cell parameters and
systematic absences in the diffraction data for 1·H2O are consistent
with Cc and C2/c. The data set was treated with absorption
corrections based on redundant data using SADABS.38 Solution
in the centrosymmetric space group option yielded chemically
reasonable and computationally stable results of refinement. All
non-hydrogen atoms were refined with anisotropic displacement
parameters. Structure factors are contained in the SHELXTL 6.12
library.38 Hydride ligands were placed in calculated positions by
use of the XHYDEX program.21
We thank the Natural Sciences and Engineering Research Council
of Canada for financial support, Dr Raymond J. Batchelor (SFU)
for his assistance and Dr Eberhard Kiehlmann for reading the
manuscript.
Notes and references
1 For example: (a) S. Scheiner, Hydrogen Bonding: A Theoretical Perspec-
tive, Oxford University, New York, 1997, pp. 119; (b) M. D. Joesten
and L. J. Schaad, Hydrogen Bonding, Marcel Dekker, New York, 1974,
pp. 183; (c) S. N. Vinogradov and R. H. Linnell, Hydrogen Bonding,
Van Nostrand Reinhold, New York, 1971, pp. 124; (d) G. A. Jeffery, An
Introduction to Hydrogen Bonding, Oxford University, New York, 1997,
pp. 123; (e) H.-C. Kan, M.-C. Tseng and Y.-H Chu, Tetrahedron, 2007,
63, 1644; (f) H.-H. Limbach, M. Pietrzak, H. Benedict, P. M. Tolstoy,
N. S. Golubev and G. S. Denisov, J. Mol. Struct., 2004, 706, 115; (g) R.
Custelcean, C. Afloroaei, M. Vlassa and M. Polverejan, Angew. Chem.,
Int. Ed., 2000, 39, 3094; (h) H. Benedict, H.-H. Limbach, M. Wehlan,
W.-P. Fehlhammer, N. S. Golubev and R. Janoschek, J. Am. Chem.
Soc., 1998, 120, 2939; (i) H. D. Lutz and H. Mo¨ller, J. Mol. Struct.,
1993, 295, 147; (j) A. N. Christensen, Acta Chem. Scand., 1992, 46, 240;
(k) E. Matshushita and T. Matsubara, Prog. Theor. Phys., 1982, 67, 1.
2 (a) M. Tachikawa and M. Shiga, J. Am. Chem. Soc., 2005, 127, 11908;
(b) H. Sugimoto, J. Phys.: Condens. Matter, 1998, 1237; (c) G. Held
and D. Menzel, Phys. Rev. Lett., 1995, 4221; (d) M. I. McMahon, R. J.
Crystals of 2 were grown from a cooled, concentrated (dry)
CH2Cl2 solution. A yellow block was mounted◦on a glass fiber.
Data were collected at low temperature (−123 C) on a Nonius
Kappa-CCD diffractometer with COLLECT.39 The unit cell
parameters were calculated and refined from the full data set.
Crystal cell refinement and data reduction were carried out using
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 1375–1382 | 1381
©

View Article Online
Nelmes, W. F. Kuhs, R. Dorwarth, R. O. Piltz and Z. Tun, Nature,
1990, 348, 317.
in a dried Carius tube subjected to an active vacuum. The C6F6 (15 mL)
was from a new bottle. It was first exposed to molecular sieves, then
dried over P2O5 for 2 days. The reaction was degassed prior to heating.
The products ([Os5(CO)19] and [Os3(CO)12]) were separated via column
chromatography under nitrogen with oven-dried silica gel and dried
solvents.
3 (a) J. A. Stride, U. A. Jayasooriya and J. Eckert, Angew. Chem., Int.
Ed., 1999, 38, 116; (b) M. Ichikawa, T. Gustafsson, I. Olovsson and T.
Tsuchida, J. Phys. Chem. Solids, 1999, 60, 1875.
4 I. Ichikawa, Pol. J. Chem., 1998, 72, 230.
5 W. Poll, M. Lohmeyer and D. Mootz, Z. Naturforsch., B: Chem. Sci.,
1989, 44, 1359.
17 B. F. G. Johnson, J. Lewis, W. J. H. Nelson, J. Puga, K. Henrick and M.
McPartlin, J. Chem. Soc., Dalton Trans., 1983, 1203.
18 B. F. G. Johnson, J. Lewis, P. R. Raithby, G. M. Sheldrick, K. Wong
and M. McPartlin, J. Chem. Soc., Dalton Trans., 1978, 673.
19 B. F. G. Johnson, J. Lewis, P. R. Raithby and C. Zuccaro, J. Chem. Soc.,
Dalton Trans., 1980, 716.
20 M. R. Churchill and B. G. DeBoer, Inorg. Chem., 1977, 16, 878.
21 A. G. Orpen, J. Chem. Soc., Dalton Trans., 1980, 2509. See:
http://www.ccp14.ac.uk/ccp/web-mirrors/bristol/software/xhydex/
index.html.
22 W. Wang, R. J. Batchelor, F. W. B. Einstein, C.-Y. Lu and R. K. Pomeroy,
Organometallics, 1993, 12, 3598.
23 W. Wang, H. B. Davis, F. W. B. Einstein and R. K. Pomeroy,
Organometallics, 1994, 13, 5113.
24 H. D. Lutz and B. Engelen, Trends Appl. Spectrosc., 2002, 4, 355.
25 L. Brammer, in Crystal Design: Structure and Function, ed. G. R.
Desiraju, Wiley, San Francisco, 2003, vol. 7, pp. 14.
6 F. H. Allen, Acta Crystallogr., Sect. B: Struct. Sci., 2002, 58, 380.
7 J. Jiang and S. A. Koch, Chem. Commun., 2002, 16, 1724.
8 D. Braga and F. Grepioni, Acc. Chem. Res., 1997, 30, 81.
9 Some examples of crystalline hydrates interacting with carbonyl
ligands: (a) S. Chan and W.-T. Wong, J. Chem. Soc., Dalton Trans.,
1995, 3987; (b) P. T. Greene and R. T. Bryan, R. T, Inorg. Chem., 1969,
8, 1464; (c) V. Albano, P. Chini and V. Scatturin, Chem. Commun.,
1968, 3, 163; (d) V. Ablano, P. L. Bellon, P. Chini and V. Scatturin,
J. Organomet. Chem., 1969, 16, 461; (e) V. Albano, P. Chini and V.
Scatturin, J. Organomet. Chem., 1968, 15, 423; (f) S. Bhaduri, K.
Sharma, P. G. Jones and C. F. Erdbru¨gger, J. Organomet. Chem., 1987,
326, C46; (g) G. A. Carriedo, V. Riera, M. G. Sanchez, Y. Jeannin and
M. Philoche-Levisalles, J. Organomet. Chem., 1989, 361, 197; (h) G. J.
Kubas, C. J. Burns, G. R. K. Khalsa, L. S. Van Der Sluys, G. Kiss and
C. D. Hoff, Organometallics, 1992, 11, 3390.
10 J. P. Canal, G. P. A. Yap and R. K. Pomeroy, Organometallics, 2003,
22, 3439.
11 (a) D. H. Farrar, B. F. G. Johnson, J. Lewis, J. N. Nicholls, P. R. Raithby
and M. J. Rosales, J. Chem. Soc., Chem. Commun., 1981, 273; (b) J. N.
Nicholls, D. H. Farrar, P. F. Jackson, B. F. G. Johnson and J. Lewis,
J. Chem. Soc., Dalton Trans., 1982, 1395.
12 A. J. Deeming, in Advances in Organometallic Chemistry, ed.
F. G. A. Stone and R. West, Academic Press, New York, 1986,
vol. 26.
26 A. J. Deeming, P. J. Manning, I. P. Rothwell, M. B. Hursthouse and
N. P. C. Walker, J. Chem. Soc., Dalton Trans., 1984, 2039.
27 For example: (a) U. Bodensieck, H. Stoeckli-Evans and G. Su¨ss-Fink,
J. Organomet. Chem., 1992, 433, 149; (b) J. S. Field, R. J. Haines and
D. N. Smit, J. Chem. Soc., Dalton Trans., 1988, 1315.
28 E. G. Bryan, A. Forster, B. F. G. Johnson, J. Lewis and T. W. Matheson,
J. Chem. Soc., Dalton Trans., 1978, 196.
29 N. Dozova, L. Krim, M. E. Alikhani and N. Lacome, J. Phys. Chem.
A, 2005, 109, 10273.
13 (a) J. N. Nicholls and M. D. Vargas, in Inorganic Synthesis, ed.
H. D. Kaesz, Wiley-Interscience, New York, 1989, vol. 26, pp. 295;
(b) An improved preparation has been reported in the PhD thesis
(Simon Fraser University, 2004) of John P. Canal (ref. 32a).
14 (a) C. R. Eady, B. F. G. Johnson, J. Lewis, B. E. Reichert and G. M.
Sheldrick, J. Chem. Soc., Chem. Commun., 1976, 271; (b) C. R. Eady,
B. F. G. Johnson and J. Lewis, J. Chem. Soc., Dalton Trans., 1975, 2606;
(c) W. Wang, F. W. B. Einstein and R. K. Pomeroy, J. Chem. Soc., Chem.
Commun., 1992, 1737; (d) Reported in the PhD thesis (Simon Fraser
University, 2004) of J. P. Canal (ref. 32a) is the 13C NMR spectrum of
[Os5(CO)16]: the analytical sample of Os5(CO)16 was obtained as air-
stable, pink–red crystals by recrystallization from CH2Cl2. 13C NMR
(CD2Cl2, RT) d 176.0; 13C NMR (CD2Cl2, −90 ◦C) d 176.1 (10C), 175.8
(6C); 13C NMR (CD2Cl2–CHFCl2, −125 ◦C) d 176.2, 176.1.
15 (a) E. J. Ditzel, B. F. G. Johnson, J. Lewis, P. R. Raithby and M. J.
Taylor, J. Chem. Soc., Dalton Trans., 1985, 555; (b) J. S.-Y. Wong, Z.-Y.
Lin and W.-T. Wong, Organometallics, 2003, 22, 4798.
30 F. A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry,
Interscience, New York, 3rd edn, 1972, pp. 692.
31 (a) C. O. Area´n and C. M. Carbonell, Vib. Spectrosc., 1995, 8, 411;
(b) J. S. Kristoff and D. F. Shriver, Inorg. Chem., 1974, 13, 499.
32 (a) J. P. Canal, PhD Thesis, Simon Fraser University, 2004; (b) V. J.
Johnston, F. W. B. Einstein and R. K. Pomeroy, Organometallics, 1988,
7, 1867.
33 Comprehensive Organometallic Chemistry III, ed. D. M. P. Mingos and
R. M. Crabtree, Elsevier, New York, 2007, vol. 6.
34 E. M. Boon, S. H. Huang and M. A. Marletta, Nat. Chem. Biol., 2005,
1, 53.
35 D. Braga and F. Grepioni, An. Quim., Int. Ed., 1996, 92, 191.
36 R. F. Alex and R. K. Pomeroy, Organometallics, 1987, 6, 2437.
37 D. D. Perrin, W. L. F. Armarego and D. R. Perrin, Purification of
Laboratory Chemicals, Pergamon, Toronto, 2nd edn, 1980.
38 G. M. Sheldrick, Acta Crystallogr., Sect. A, 2008, 64, 112.
39 COLLECT (Data collection software), Nonius B. V., Delft, The
Netherlands, 1999.
16 Exceedingly dry conditions for eqn (1b): a flame-dried Carius tube was
used. The sample of [Os4(CO)14] (20 mg) was heated overnight at 60 ◦C
40 Z. Otwinowski and W. Minor, Methods Enzymol., 1997, 276, 307.
1382 | Dalton Trans., 2008, 1375–1382
This journal is
The Royal Society of Chemistry 2008
©