Kinetic studies of some substituted hexarhodium carbonyl clusters

Article abstract of DOI:10.1039/b408687b

Reactions of the halides X^- (X^- = chloride, bromide or iodide) with the substituted cluster Rh₆(CO)₁₅(PPh ₃) in oxygen-free chloroform lead to [Rh₅(CO) ₁₄(PPh₃)]^-, Rh(CO)₂(PPh ₃)₂X and [Rh(CO)₂X₂]^- in the molar ratios 2 : 1 : ～13, Oxidation by the solvent is assumed to lead to most of the Rh(I) product, and the stoichiometry for reactions with I ^- can be defined as 4Rh₆(CO)₁₅(PPh ₃) + 27I^- + 12CHCl₃ → 2[Rh ₅(CO)₁₄(PPh₃)]^- + Rh(CO) ₂(PPh₃)₂I + 13[Rh(CO)₂I ₂]^- + 6C₂H₂Cl₄ + 4CO + 12Cl^-. This can be rationalized quite simply with the aid of a few generally justifiable assumptions. Rate constants for reactions with bromide increase to a limiting value with increasing [Br^-] in a way that shows that breaking of one Rh-Rh bond, with an unusual closo to nido structural change, is rate determining. This opening of the cluster might be spontaneous or solvent induced. To complete the reaction, the bromide has to compete with the reverse nido to closo change. The same closo to nido change is also a major rate determining step for reactions with P(OPh)₃ in oxygen-free solutions, and for reactions with bromide in oxygenated solutions in the presence of trifluoroacetic and some other acids. The limiting rates increase slightly with increasing basicity of the ligands P(p-XC₆H ₄)₃ along the series X = F₃C, Cl, F, H and MeO. Activation parameters for these reactions are reported.

Full text of DOI:10.1039/b408687b

View Article Online / Journal Homepage / Table of Contents for this issue
F U L L P A P E R
Kinetic studies of some substituted hexarhodium carbonyl clusters†
Claudia Babij,a David H. Farrar,a Igor O. Koshevoy,a,b Anthony J. Poe¨*a and Sergey P. Tunik*b
a Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario,
Canada M5S 3H6. E-mail: apoe@chem.utoronto.ca
b Department of Chemistry, St. Petersburg University, Universitetskii pr., 26, St. Petersburg,
198504, Russian Federation. E-mail: stunik@chem.spbu.ru
Received 9th June 2004, Accepted 29th October 2004
First published as an Advance Article on the web 25th November 2004
Reactions of the halides X− (X− = chloride, bromide or iodide) with the substituted cluster Rh6(CO)15(PPh3) in
oxygen-free chloroform lead to [Rh5(CO)14(PPh3)]−, Rh(CO)2(PPh3)2X and [Rh(CO)2X2]− in the molar ratios
2 : 1 : ∼13. Oxidation by the solvent is assumed to lead to most of the Rh(I) product, and the stoichiometry for
reactions with I− can be defined as 4Rh6(CO)15(PPh3) + 27I− + 12CHCl3 → 2[Rh5(CO)14(PPh3)]− +
Rh(CO)2(PPh3)2I + 13[Rh(CO)2I2]− + 6C2H2Cl4 + 4CO + 12Cl−. This can be rationalized quite simply with the aid
of a few generally justifiable assumptions. Rate constants for reactions with bromide increase to a limiting value with
increasing [Br−] in a way that shows that breaking of one Rh–Rh bond, with an unusual closo to nido structural
change, is rate determining. This opening of the cluster might be spontaneous or solvent induced. To complete the
reaction, the bromide has to compete with the reverse nido to closo change. The same closo to nido change is also a
major rate determining step for reactions with P(OPh)3 in oxygen-free solutions, and for reactions with bromide in
oxygenated solutions in the presence of trifluoroacetic and some other acids. The limiting rates increase slightly with
increasing basicity of the ligands P(p-XC6H4)3 along the series X = F3C, Cl, F, H and MeO. Activation parameters
for these reactions are reported.
Rh6(CO)16 is also susceptible to rapid attack by hydroxide
Introduction
ions. (See the ESI† for a summary of these studies). Although
The carbonyl cluster Rh6(CO)16 was the first binary carbonyl
cluster1 in which each metal atom was found to disobey the
simple 18-electron rule,2 and its two extra electrons appear to
enable it to exhibit a rich redox chemistry that very often involves
destruction or enlargement of the cluster framework.3,4 Applica-
tion of MO theory to M6 clusters of this electron configuration
showed that the extra electron pair can be accommodated in an
[A2g] weakly anti-bonding cluster orbital5 so that the 86-electrons
form a closed shell arrangement. The octahedral closo Rh6 core
is fully consistent with PSEPT rules2,6 for 86-electron clusters.
However, some features of its reactivity are poorly un-
derstood. It is, for instance, extremely resistant to simple
substitution reactions such as CO exchange,7 and no studies
of kinetics of substitution by soft ligands such as P-donors
have been reported. This contrasts with reactions of its smaller
congener Rh4(CO)12 which undergoes substitution exceedingly
rapidly.8,9 Attempts to study such reactions of Rh6(CO)16 are
complicated by their great sensitivity to oxygen and solvent
impurities whereby trace amounts of impurities appear to give
rise to irreproducible behaviour.10
So far the only successful kinetic studies of substitutions at the
Rh6 cluster core have involved the rapid displacement of labile
ligands in such “lightly stabilized” clusters as Rh6(CO)15NCMe,
etc.,11 or the unusual chelation reactions in which a substi-
tuted P-bonded pyrrolyl ligand displaces a CO ligand from a
neighbouring Rh atom.12 Similar chelation reactions of P-donor
ligands that contain more usual donor atoms in their pendant
groups have also been described.13 In an attempt to investigate
the nature of substitution reactions at the Rh6 cluster core more
thoroughly we attempted to study reactions of Rh6(CO)16 with
hard, activating nucleophiles that are known to attack other
carbonyl clusters through carbon atoms of the CO ligands.14
Me3NO attacks Rh6(CO)16 quite effectively and is used in the
synthesis of Rh6(CO)14(NCMe)2.15 Kinetic studies show that
2−
Rh6(CO)15 can be synthesized in nearly quantitative yield by
addition of KOH to a slurry of Rh6(CO)16 in MeOH under
nitrogen,16 the kinetic runs have to be carried out in chlorinated
solvents, for solubility reasons, and the products are not well
defined. Halide ions turn out to be only very slightly reactive
towards Rh6(CO)16 and the high susceptibility to traces of
oxygen10,17 makes kinetic studies unprofitable.
We have therefore examined reactions of some Rh6(CO)15P(p-
XC6H4)3 clusters with halide ions, thereby hoping to make use
of the labilizing effect of the quite large P(p-XC6H4)3 ligands,18
as well as the nucleophilic character of halides towards other
carbonyls.17,19–23
Experimental
General comments
[NBu4]Br, CF3COOH and phosphines (Strem), and [NBu4]I
and [PPN]Cl (Aldrich), were used as received. The com-
plexes Rh6(CO)15(PPh3),24 Rh6(CO)15P(p-XC6H4)3 (X = MeO,
Cl, F and CF3),25 Rh4(CO)12,26 and [Rh5(CO)14(PPh3)][NBu4]27
were prepared according to published procedures. Rh(CO)2-
(PPh3)2I28,29 was prepared by a method to be described else-
where,30 and characterized by IR and NMR spectroscopy, and
by crystallography.30
Chloroform (Fisher) for kinetic experiments was washed
three times with equal volumes of distilled water, dried for
12 h over CaCl2, and distilled over P2O5 under nitrogen. Other
solvents were dried over appropriate reagents and distilled prior
to use. IR spectra were recorded using a Nicolet Magna IR 550
FTIR spectrophotometer. UV-Vis spectra were recorded with
Varian-Cary 2000 and Hewlett-Packard 8452A diode array spec-
trophotometers. Spectrometers were fitted with thermostated
(
∼0.1 ◦C or better) cell holders for kinetic studies. Solutions
for kinetic study were prepared under nitrogen, and the reactions
were monitored by repetitive scanning of samples in the ther-
mostated IR or UV cells. Rate constants were obtained by single
or double exponential fitting of the changes of absorbances with
† Electronic supplementary information (ESI) available: Tables of rate
constants, and a description of the kinetics of reactions of Rh6(CO)16
with hydroxide. See http://www.rsc.org/suppdata/dt/b4/b408687b/
1 1 6
D a l t o n T r a n s . , 2 0 0 5 , 1 1 6 – 1 2 2
T h i s j o u r n a l i s
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 5
©

View Article Online
time, making use of Grafit (v.3.09b) software. The 31P NMR
spectra were recorded on a Varian Gemini 300 MHz instrument,
and referenced to external 85% H3PO4.
the rhodium complexes: a sharp doublet centered at 141.9 ppm
(1J(P–Rh) 239 Hz) and a broad resonance around 109 ppm.
Solvent was removed under vacuum, and the residue dissolved in
CH2Cl2 (2 ml), diluted with hexane (2 ml), and transferred onto a
chromatographic column (2.5 × 7 cm). The following bands were
obtained in the order of elution: (eluent CH2Cl2–hexane (3 : 4)),
a broad red–orange band containing Rh4(CO)8{P(OPh)3}4,
83 mg; and (eluent CH2Cl2–THF (5 : 1)), a narrow yellow–
orange band containing RhCl{P(OPh)3}3, 48 mg.
Results
Reaction of Rh6(CO)15(PPh3) with halide ions under an
oxygen-free atmosphere
Analysis of reaction products and stoichiometry. Reactions
of Rh6(CO)15(PPh3) with halide ions (as [NBu4]I, [NBu4]Br and
[PPN]Cl) in oxygen-free solutions gave three main products:
[Rh(CO)2Hal2]−, [Rh5(CO)14(PPh3)]− and [Rh(CO)2(PPh3)2Hal].
The first type of complex is stable indefinitely under any
conditions, but the others are stable only in the absence of
oxygen and in the presence of CO, the amount released during
the reactions being sufficient. Otherwise, even under oxygen-free
conditions, these complexes readily lose CO.27–30
The IR spectrum of the products of reaction with a large excess
of bromide in CHCl3 displayed strong CO stretching bands at
2069 and 1996 cm−1 of [Rh(CO)2Br2]−, which are consistent with
the previously reported data: 2068 and 1993 cm−1 in CH2Cl2.31
The bands due to this compound were subtracted from the IR
spectrum of the product mixture to give the spectrum shown in
Fig. 1 and the remaining bands, and their relative intensities,
correspond well with those reported for [Rh5(CO)14(PPh3)]− in
THF.27
The Rh4(CO)8{P(OPh)3}4 cluster was identified by its C–O
stretching frequencies (2038w, 2023sh, 2011vs, 1832s; cf. 2036w,
2022sh, 2008vs, 1836s)33 and by its 31P NMR spectra in CDCl3
(broad multiplet at d 109.9 ppm). The latter fits well to dynamic
averaging of the resonances observed in the low temperature
limiting spectrum of this cluster.32,33 The 31P NMR spectral data
of the mononuclear complex Rh{P(OPh)3}3Cl (CDCl3, d 113.1,
dd 1J(P–Rh) 225 Hz, 2J(P–P) 53 Hz; d 119.8, dt, 1J(P–Rh)
285 Hz, 2J(P–P) 53 Hz) are consistent with previously reported
1
2
data (CDCl3, d 112.5, dd, J(P–Rh) 222 Hz, J(P–P) 55 Hz, d
118.9 (dt, 1J(P–Rh) 281 Hz, 2J(P–P) 55 Hz).34 The composition
of the mononuclear complex clearly indicates the oxidation of
the reactive intermediates (Rh(0) species) by the chlorinated
solvent. It is very probable that the interaction of the starting
“Rh6” cluster with the excess of P(OPh)3 results in its splitting
into “Rh4” and “Rh2” fragments, the latter being extremely
reactive to give, for example, Rh{P(OPh3}2(l-Cl)2Rh{POPh)3}2.
An analogous binuclear Rh(I) complex containing P(OMe)3 lig-
ands has been characterized earlier35 by 31P NMR spectroscopy
and displays the spectroscopic characteristics (d 138.1 ppm,
J(Rh–P) 295 Hz) which are quite similar to those found for
the low-field signal presented in the reaction mixture. Thus, we
suppose that this signal is generated by the Rh{P(OPh)3}2(l-
Cl)2Rh{P(OPh)3}2 complex, which transforms into the mononu-
clear Rh{P(OPh)3}3Cl complex on silica in the presence of the
ligand excess.
Reaction kinetics by FTIR monitoring. Rate constants were
determined by monitoring the decreasing absorbance of the IR
band of the starting cluster Rh6(CO)15(PPh3) at 2100 cm−1, and
fitting absorbance data to single-exponential curves. However,
when lower concentrations of bromide were used, and rates
were quite slow, initial rates of absorbance change were chosen
together with A∞ values inferred from other reactions. In a
typical experiment, a solution (ca. 1.5 cm3, ca. 1 mM) of
Rh6(CO)15(PPh3) in chloroform was degassed (three freeze–
pump–thaw cycles) using standard Schlenk techniques, and
added by syringe to a degassed crystalline halide salt (chloride,
bromide or iodide) or to a degassed solution of P(OPh)3.
After dissolution or complete mixing, the reaction mixture was
degassed again by one freeze–pump–thaw cycle and transferred
by syringe to an amalgamated NaCl cell (Wilmad Glass) with
a water-thermostated jacket, and the spectra were scanned
repetitively in the carbonyl stretching region. A typical set of
spectra is shown in Fig. 2, and rate constants for a range of
concentrations of nucleophile, and at various temperatures, are
given in Table 1. The rate constants for reaction with bromide at
Fig. 1 An IR spectrum of the final product of the reaction of a large
excess of bromide with Rh6(CO)15(PPh3) in oxygen-free CHCl3 after
subtraction of the IR spectrum of [Rh(CO)2Br2]−.
The 31P NMR spectrum of the final reaction mixture in CDCl3
displayed two signals of equal intensity at 22.4 ppm [dm, 1J(P–
1
Rh) 183 Hz] and 31.2 ppm [d, J(P–Rh) 95 Hz] when Hal =
I. The high-field signal parameters are identical with the spectral
data found for the independently synthesized cluster anion
[Rh5(CO)14(PPh3)]−,27 while the low-field signal corresponds to
the independently synthesized30 complex Rh(CO)2(PPh3)2I (31.2
[d, 1J(P–Rh) 95 Hz]). Iodide was chosen as the halide for these
experiments because of the higher stability of [Rh(CO)2(PPh3)2I]
compared to its bromo and chloro analogues.
◦
26.5 C clearly increase with [Br−] towards a limiting rate, and
Characterization of products of the reaction of Rh6(CO)15-
(PPh3) with P(OPh)3. For comparative purposes, reac-
tions of Rh6(CO)15(PPh3) with P(OPh)3 in oxygen-free solu-
the conditions for the other reactions are such that limiting rates
are attained. The limiting rates all fit onto a common Eyring
plot, irrespective of the nucleophile, as shown in Fig. 3. On
the other hand, the rate constants for reactions with P(OPh)3
increase linearly with [P(OPh)3] in a way that may suggest some
contribution from an associative path. (See Fig. S1 in the ESI†).
tions were examined and the following products identified:
32,33
Rh4(CO)8{P(OPh)3}4
and free PPh3, together with a Rh2
complex that is possibly Rh{P(OPh3)}2(l-Cl)2Rh{P(OPh)3}2
(see below).
Rh6(CO)15(PPh3) (70 mg, 0.054 mmol) was dissolved in CHCl3
(5 ml). The solution was degassed (three freeze-pump-thaw
cycles) andwas added to adegassedsolution ofP(OPh)3 (260mg,
0.839 mmol) in CHCl3 (3 ml). The reaction mixture was left for
10–12 h until a TLC spot test showed complete consumption
of the starting cluster. By this time the 31P NMR spectrum of
the reaction mixture displayed two signals that correspond to
Reactions of the Rh6(CO)15P(p-XC6H4)3 (X = CF3, Cl, F, H,
MeO) clusters with bromide in the presence of oxygen and acid
Analysis of reaction products and stoichiometry. Reactions
of Rh6(CO)15(PPh3) with bromide in the presence of oxygen
alone are appreciably faster than in its absence, and lead mainly
to [Rh(CO)2Br2]− and CO2, together with small amounts of
D a l t o n T r a n s . , 2 0 0 5 , 1 1 6 – 1 2 2
1 1 7

View Article Online
Fig. 2 Typical IR spectral changes for the oxygen-free reaction of bromide with Rh6(CO)15(PPh3) (T = 32 ◦C). The arrows denote increasing and
decreasing absorbances. Inset: final spectrum of the reaction mixture; asterisks denote bands corresponding to [Rh(CO)2Br2]−.
Table 1 Kinetic data for the reaction of Rh6(CO)15(PPh3) with halide
ions or P(OPh)3 under an oxygen-free atmosphere in CHCl3. Initial
values of [Rh6(CO)15(PPh3)] were ca. 0.3 mM
Reactant
Cl−
T/◦C
[Reactant]/mM
103kobs/s−1
38.3
42
53
67
0.785
0.827
0.824
Br−
26.5
5
10
26
51
86
133
217
0.075
0.104
0.119
0.140
0.149
0.178
0.176
Fig. 3 Eyring plot for the reaction of bromide with Rh6(CO)15(PPh3)
in oxygenated and acidified CHCl3. Data for reactions in oxygen-free
solution are superimposed on the Eyring plot.
32.8
38.3
32
48
77
0.275
0.337
0.347
0.371
kinetic data were obtained. The cluster is not sensitive to
oxygen by itself. However, in the presence of trifluoroacetic
acid (TFA), Rh6(CO)15(PPh3) and its congeners Rh6(CO)15P(p-
XC6H4)3 (X = CF3, Cl, F or MeO), react smoothly with
bromide, under an atmosphere of dry air or oxygen, to give
essentially quantitative yields of [Rh(CO)2Br2]− and free P(p-
XC6H4)3. These subsequently react quite slowly to give some
unidentified mononuclear phosphine-carbonyl complexes of
Rh(I). Thus the presence of [Rh(CO)2Br2]− was confirmed by
its CO stretching bands (see above), and the 31P NMR spectrum
of the final reaction mixture when X = H displays one signal,
characteristic of coordinated PPh3, at 29.7 ppm (CDCl3, d,
1J(Rh–P) 114.2 Hz). The addition of an excess of PPh3 to this
reaction mixture leads to the growth of intensity of this NMR
signal and a decrease in the absorbance of the corresponding
CO-stretching bands in the IR spectrum. These observations
are indicative of CO substitution by PPh3 in the slow reaction
following [Rh(CO)2Br2]− formation.
107
24
38
64
87
126
165
0.735
0.897
0.903
0.860
0.879
0.900
45.0
55
59
89
92
112
2.07
2.03
2.13
2.15
2.24
I−
38.3
25.5
40
86
111
0.72
1.03
1.01
P(OPh)3
22
42
187
388
546
0.131
0.137
0.144
0.170
0.167
Reaction kinetics. Typical sets of time-dependent IR spectra
display quite sharp isosbestic points as shown in Fig. 4, and
rate constants could be obtained by monitoring the absorbance
changes of the most intense band in the IR spectrum of the
Rh6(CO)15(PPh3) cluster.
However, the reactions in acidic solutions saturated with
dry air or oxygen could also be monitored quite simply by
continuous measurement of the absorbance at 420 nm. In a
typical experiment, a solution (0.3 cm3; ca. 10 mM) of the cluster
in chloroform was transferred into a pre-thermostated 10 mm
P(OPh)3/1 atm of CO
26.5
23
0.09
[Rh6(CO)15Br]− and other unidentified products. These reactions
are evidently catalytic and complex, and no easily interpretable
1 1 8
D a l t o n T r a n s . , 2 0 0 5 , 1 1 6 – 1 2 2

View Article Online
Fig. 4 Typical IR spectral changes for the reaction of Rh6(CO)15P(p-XC6H4)3 with bromide in the presence of oxygen and trifluoroacetic acid (X =
Cl). The arrows denote increasing and decreasing absorbances. Inset: final spectrum of the reaction mixture.
4Rh6(CO)15(PPh3) + 27Br− + 12CHCl3
Table 2 Activation parameters for the reactions of the Rh6(CO)15P(p-
→ 2[Rh5(CO)14(PPh3)]− + Rh(CO)2(PPh3)2Br
XC6H4)3 clusters with bromide in the presence of oxygen and acid in
CHCl3
+ 13[Rh(CO)2Br2]− + 6C2H2Cl4 + 4CO + 12Cl−
(1)
ꢀ
a
X
pKa
DH=/kJ mol−1
DS=/J mol−1 K−1
rkobs (%)
This quite complicated equation incorporates the observa-
tion of equal intensities of the 31P bands of the products
[Rh5(CO)14(PPh3)]− and Rh(CO)2(PPh3)2I in the NMR spectra
of the products of reaction with I−. By inference, it is concluded
that the same would be true for reactions with Br−. Eqn. (1) also
takes account of the number of ligands needed to maintain the
appropriate coordination numbers of the complexes, the dispro-
portionation of the cluster, and the amount of CHCl3 needed, in
addition to disproportionation, to produce all the Rh(I) species.
Since Rh(CO)2(PPh3)2Br contains two PPh3 ligands compared
with one in [Rh5(CO)14(PPh3)]−, its molar concentration must
be half that of the former product. Further, the molar yield of
[Rh(CO)2Br2]− can be estimated approximately as follows.
If it is assumed that the yield of [Rh(CO)2Br2]−, after reactions
under oxygen, is quantitative then the molar absorbance of
[Rh(CO)2Br2]− at 1996 cm−1 can be estimated to be ca. 8.5% of
that of Rh6(CO)15(PPh3) at 2065 cm −1 from data similar to those
in Fig. 4. Data from four runs similar to those in Fig. 2, show
that the absorbance of the [Rh(CO)2Br2]− at 1993 cm−1 is ca.
64, 66, 52 and 54% of that expected for quantitative formation,
whereas the stoichiometry in eqn. (1) requires a yield of 100 ×
13/24, i.e. 54%. However, the presence of some of the unstable
minor product [Rh5(CO)14(PPh3)]− would contribute somewhat
to the absorbance at 1996 cm−1 and the two higher estimates
for the yield might well be due to this factor. Thus, within the
precision of the estimates, the observed yield of [Rh(CO)2Br2]− is
consistent with that expected from the proposed stoichiometry.
This stoichiometry can be accounted for by the following
sequence of reactions,
CF3
Cl
F
H
MeO
−1.39
0.87
1.63
3.28
5.13
101.3 0.6
95.0 1.0
105.1 0.9
101.2 0.7
93.6 0.7
21.4 1.9
−3.3 3.8
32.2 3.0
20.6 2.4
−3.5 2.3
2.1
4.9
4.1
3.6
3.6
a The standard error of an individual measurement of kobs
.
UV-vis cell, containing a solution of TFA (ca. 100 mM) and
NBu4Br (ca. 170 mM) in chloroform (3 cm3). Rate constants
were usually determined by fitting absorbance data to double-
exponential curves. Double exponential analysis was needed
because of the further reaction of the initial [Rh(CO)2Br2]−
product mentioned above. Convenient temperature ranges were
selected for each cluster and at least four runs were carried
out over a 20 ◦C temperature range. Rate constants obtained
in this way are given in Table S1 in the ESI,† and activation
parameters are given in Table 2. The data show that the rate
constants for each of these reactions all fall on a good Eyring plot
(Fig. 3), irrespective of the concentrations of acid or bromide,
and have satisfyingly small standard errors of measurement
(Table 2). The rates were also independent of the nature and
concentration of other acids. Thus reactions at 25.6 ◦C in CHCl3
with bromide in the presence of 33.7 mM benzoylpropionic
acid, 43.7 mM benzoic acid, and 86.5 mM trifluoroacetic acid
led to values of 103kobs equal to 0.117, 0.113 and 0.112 s−1,
respectively. IR monitoring at ambient temperatures was carried
out for each Rh6(CO)15P(p-XC6H4)3 cluster to check that the
reaction proceeded to give products analogous to those when
X = H.
2Rh6(CO)15L + 4Br−
→ 2[Rh5(CO)14L]− + [Br(CO)Rh(l-Br)2Rh(CO)Br]2−
(2)
2Rh6(CO)15L + 4Br−
Fig. 3 shows that data for reactions of the halides or P(OPh)3
in deoxygenated solutions, and for reactions with bromide in the
presence of acid and oxygen, all fall on the same Eyring plot, i.e.
their limiting rates are always the same.
→ 2[Rh5(CO)14Br]2− + [L(CO)Rh(l-Br)2Rh(CO)L]
(3)
(4)
[L(CO)Rh(l-Br)2Rh(CO)L] → [L2Rh(l-Br)2Rh(CO)2]
2[Rh5(CO)14Br]2− + 12CHCl3 + 18Br−
→ 10[Rh(CO)2Br2]− + 8CO + 12Cl− + 6C2H2Cl4
(5)
Discussion
The course of the reactions
[Br(CO)Rh(l-Br)2Rh(CO)Br]2− + 2CO → 2[Rh(CO)2Br2]− (6)
L2Rh(l-Br)2Rh(CO)2] + 2CO + Br−
The reactions with bromide in chloroform and in the absence of
oxygen and acid can be discussed in terms of the stoichiometric
equation:
→ [Rh(CO)2Br2]− + [Rh(CO)2L2Br]
(7)
D a l t o n T r a n s . , 2 0 0 5 , 1 1 6 – 1 2 2
1 1 9

View Article Online
where the known reactant cluster and products are shown in bold
type. The presence of chloride in the product mixture suggests
that some chloro carbonyl products might be formed. Some
small spectral shifts, consistent with this, are in fact observed
at low concentrations of bromide but this does not affect the
following discussion.
constant k2 is actually the sum of the individual rate constants
for formation of the two pairs of products enclosed in brackets
in eqn. (8).
The sequence of reactions in eqns. (2)–(7) assumes that:
(i) the original Rh6(CO)15L cluster can be attacked by
bromide in two ways, each of which results in abstraction of
a Rh(I) moiety (i.e. [Rh(CO)Br2]− and [Rh(CO)LBr]) which can
dimerize in the usual way to form bridged complexes. (A mixed
[Br(CO)Rh(l-Br)2Rh(CO)L] complex could also be formed but
this does not affect the following discussion). In one case the
Rh(I) carries a bromide ligand with it (eqn. (2)) and in the
other (eqn. (3)) it carries a PPh3 ligand. These two processes
leave behind a Rh5− cluster, and must evidently occur with equal
probability.
(ii) The [Rh5(CO)14Br]2− is unstable to oxidation by the
solvent (eqn. (5)) by electron transfer to a chlorine atom, with
subsequent dimerisation of the CHCl2 radical.
(iii) [L(CO)Rh(l-Br)2Rh(CO)L] can isomerize quantitatively
to [L2Rh(l-Br)2Rh(CO)2].
Fig.
5
Concentration dependence of kobs for the reaction of
Rh6(CO)15(PPh3) with bromide in oxygen-free CHCl3 at 38.5 ◦C with
widely varying values of [cluster] (see Table S2, ESI†). The line through
the data was drawn using the parameters obtained from the ‘inverse plot’
in Fig. 6.
Assumption (i) is perfectly reasonable and fits with the kinetic
indication of initial attack on an ‘opened cluster’ by bromide (see
below). [Rh5(CO)14L]− is one of the characterized products, and
[Rh5(CO)14Br]2− is a known cluster that is extremely sensitive
to oxidation,27 so supporting assumption (ii). Assumption (ii) is
also in accordance with known oxidations of mononuclear car-
bonyl complexes36 and clusters by chlorinated hydrocarbons,37,38
and (iii) is absolutely essential to the formation of the com-
pletely characterized [Rh(CO)2L2X] product. Such migrations
are known in analogous polynuclear complexes.39–42
Reactions of Rh6(CO)15(PPh3) with bromide in the presence
of oxygen are evidently complex and the sequence of compounds
formed cannot even be speculated about. This makes the
simplifying effect of added TFA difficult to explain as well, but
protonation must evidently intervene at some stage in the cat-
alytic sequence of reactions. However, even in the absence of any
understanding of the detailed processes that occur, these reac-
tions with bromide, oxygen and TFA are still useful. They allow
for kinetic monitoring by UV-vis spectroscopy, something that is
not possible for the reactions in the absence of oxygen where the
UV-vis absorbance changes are not so simple. They lead to rate
constants that are identical with those in the absence of oxygen
and enable more rapid collection of precise kinetic data.
Fig. 6 The dependence of 1/kobs on 1/[Br−] for the reaction of
Rh6(CO)15(PPh3) with bromide in oxygen-free CHCl3 at 38.5 ◦C. The
intercept and gradient give values of 1/k1 and k−1/k1k2, respectively (see
eqn. (9)).
This scheme implies that all the subsequent steps are relatively
fast and the cleanness of the spectroscopic changes, as shown
in Fig. 1, shows that the steps suggested in eqns. (2)–(7) are
sufficient to account for the overall yields without complicating
side reactions.
Kinetics and mechanism of the rate determining steps
The most important feature of the reactions with bromide in
oxygen-free solutions is the rise of kobs to a limiting value with
increasing [Br−] as shown in Fig. 5. When a dissociation constant
of 2.7 × 10−5 M is taken for [NBu4Br] in CHCl343 a linear ‘inverse
plot’ of 1/kobs against 1/[Br−] is obtained as shown in Fig. 6. This
is characteristic of slow formation of a reactive intermediate
which undergoes competitive reactions, either reverting to the
original cluster or reacting with free bromide to proceed further
along the reaction sequence as shown in eqn. (8).
A crucial feature of Fig. 5 is that the rate constants were
obtained using initial concentrations of complex that varied by
a factor of over six (see data in Table S2, ESI†) This precludes
the rate-determining step being CO-dissociation. Although CO
is present as a product during the reactions its concentration
varies from zero to only one equivalent of the initial cluster
concentration as the reaction proceeds. This would lead to
progressive retardation of the rates during the course of the
reactions and, therefore, to non-linear rate plots, if the reverse
of CO dissociation were occuring. In addition, this small average
concentration of CO would vary by a factor of >6 as the
initial concentration of cluster was varied by this factor. This
would inevitably lead to less and less effective retardation as
that concentration decreased, and to less and less curvature
in plots such as that in Fig. 5. Such behaviour has been
observed in substitution reactions that were sufficiently sensitive
to released CO that retardation due to that CO occurred at
higher concentrations of complex. Progressive diminution of
the cluster concentration led eventually to the absence of any
retardation.44 It is also unlikely that the small amounts of
CO released would be enough to compete to this extent with
(8)
kobs = k1k2[Br−]/(k−1 + k2[Br−])
(9)
Rh6(CO)15L* is the reactive intermediate, and only the initial
products of its reaction with bromide are given. The rate
equation corresponding to this sequence of reactions is shown in
eqn. (9) which leads to the linear inverse plot shown in Fig. 6. The
intercept of this inverse plot is 1030 s, and the slope is 0.369 M s,
and these lead to a value of 2.8 × 103 M−1 for k2/k−1. The rate
1 2 0
D a l t o n T r a n s . , 2 0 0 5 , 1 1 6 – 1 2 2

View Article Online
the reaction with bromide. Such retardation of CO-dissociative
reactions are usually studied with applied pressures of CO, with
[CO] constant at ∼10−2 M. Attempts to do this with these
reactions led to complications in the spectroscopic changes and
severe irreproducibility of the rate constants. We attribute this to
direct reactions of CO with the reactive intermediate, and major
changes in the reaction sequence. Although a reaction with
P(OPh)3 under an atmosphere of CO did lead to a diminution of
the rate constant, this is a necessary but not sufficient indication
of a CO dissociative reaction.45
as required by our proposal. This ligand effect is statistically
meaningful and essentially provides an enhanced tendency
for nido cluster formation through a “pre-preparation” of the
ground state of these molecules, thus lowering the energy gap
between the ground and transition states. It is also worthwhile
to note at this point that other distortions found in the solid
state structures of Rh6(CO)15(PR3) clusters correlate well with
the mechanisms of intramolecular CO scrambling reactions,51,52
thus demonstrating again the close relations between ground
state structural parameters of these molecules and their dynamic
properties.
We therefore conclude that reaction does indeed proceed
via the sequence shown in eqn. (8), and we propose that the
reactive intermediate Rh6(CO)15L* is a form of the cluster that
has undergone opening to a ‘nido’ structure with one less Rh–
Rh connectivity. This decreased connectivity involves loss of
a ‘bond’ between the Rh atom to which the PPh3 is attached
and one other Rh atom, i.e. there is a decrease in the number
of connectivities from 12 to 11. Although 86-electron clusters
with six metal atoms and only 11 connectivities are quite well
known as isolated and fully characterised clusters,6 we believe
that a vacant coordination site formed by heterolysis of that Rh–
Rh bond would be filled by a solvent molecule, chlorocarbon
ligands being perfectly capable of stabilising Rh6 centres.46 The
charge separation that occurs on Rh–Rh bond heterolysis can
be dissipated by formation of bridging CO groups between the
three pairs of rhodium atoms involved in an opened Rh4 moiety
in a way that has been invoked before as being possible in
associative reactions of metal carbonyl clusters.9 However, such a
repositioning of the CO ligands does not appear to be generally
necessary as a means of dissipating the charge separation, no
pronounced structural changes of that sort being observed when
an acetonitrile ligand is added to the cluster Ru5(C)(CO)15.47
Indeed the moderately polar nature of the CHCl3 solvent could
well help to stabilise a degree of charge separation that might
remain if formation only of semi-bridging CO ligands occurred.
Without being specific about the exact positions of the CO
ligands the structures below are simple and reasonable descrip-
tions of two possible solvated forms of Rh6(CO)15L* where the
solvent molecule can be coordinated to the Rh atom to which
the P-donor is connected, or to a neighbouring Rh atom.This
The limiting rate constants at high [halide] are a measure of
k1, the rate constant for the spontaneous or solvent induced
cluster opening of the Rh6 cluster in deoxygenated solvent, but
reactions with bromide in oxygenated acidic solution are also
governed by the same limiting rate constant k1. Protonation of
reactive intermediates, to form hydride clusters that are very
unstable to oxidation, can presumably account for this.
The value of 2.8 × 103 M−1 for k2/k−1 indicates that the
reverse, nido to closo, reaction of the cluster is quite difficult in
comparison to attack even by the weakly nucleophilic bromide
ion. In the absence of the structural weakness induced by the
substituent, and in the case of reactions with more nucleophilic
ligands, such as P-donors, one would expect the reactions always
to occur at the limiting rates. No evidence for reversible cluster
opening would be obtained, and a CO dissociative mechanism
might be supposed to be operating. In the same vein, the
structural weakness due to the substituent might encourage a
contribution from direct nucleophilic attack, even by the rather
weak P-donor, P(OPh)3, nucleophilic attack being accompanied
by induced closo to nido cluster opening.
ꢀ
A plot of values of logk1 against pKa , which is a useful
measure of the sigma donicity of the P(p-XC6H4)3 ligands,53
is given in Fig. 7 The gradient of the linear plot is positive and
significant, showing that cluster opening is facilitated by better
sigma electron donors, but the effect is quite small and was not
detected in the Rh–Rh bond lengthening mentioned above.25
Because these ligands all have the same cone angle of 145◦,
no indication of steric effects is available although higher cone
angles would also be expected to facilitate cluster opening. Thus,
reactions of clusters substituted with large, highly donor ligands
such as PCy3 would be expected to undergo cluster opening even
faster.
distinction must be related to the relative rates of formation
of the two pairs of products in eqn. (8). There is no evidence
as to when the CHCl3 molecule (S) enters the coordination
site, i.e. whether it enters after the vacant coordination site
is completely formed, or during the process of homolysis.
Further studies involving P-donors with different electronic and
steric properties might help to answer these questions. These
solvated Rh6(CO)15L* intermediates are capable of reacting with
halides or, indeed, with P(OPh)3 which leads to substitution and
fragmentation instead of redox reactions.
A spontaneous opening of the Rh6 cluster would be rather
unusual, spontaneous breaking of Mn–Fe and Fe–Fe bonds
in Cp(CO)2MnFe2(CO)6(l-PR) and RPFe3(CO)10 clusters, re-
spectively, being two reported examples.48,49 On the other hand,
induced cluster opening by nucleophilic attack is thought to be a
virtually universal mechanism for associative reactions of metal
carbonyl clusters with P-donor and similar ligands.50 Solvent
assisted cluster opening might therefore be occuring in this case.
There is also good structural evidence in favor of the
nido species intermediacy. A detailed study of the structural
parameters in mono-substituted Rh6(CO)16 derivatives25 showed
that there is considerable elongation of the Rh–Rh bonds in the
immediate vicinity of the phosphorus donor ligands, exactly
Fig. 7 Dependence of log(kobs) (kobs = k1) for the reactions of
Rh6(CO)15P(p-XC6H4)3 with bromide in oxygenated and acidified CHCl3
at 25 ◦C on the sigma-donating ability of the phosphine ligands. The
gradient is 0.059 0.004, with r2 = 0.993 and r(logkobs) = 0.2.
The activation parameters in Table 2 show a very rough
isokinetic, or compensation, plot of DH‡ against DS‡, and the
gradient indicates that all the reactions would proceed at similar
rates at ca. 40 ◦C. This is fairly close to the temperature at which
ꢀ
the linear free energy plot of logk1 vs. pKa was plotted and
ꢀ
this will lessen the observed dependence of logk1 on pKa . The
activation parameters do not show any correlation with ligand
D a l t o n T r a n s . , 2 0 0 5 , 1 1 6 – 1 2 2
1 2 1

View Article Online
basicity but all show low-to-moderate values of DS‡ consistent
with rather little, but variable, increases in the “flexibility” of
the transition states as the nido structure is approached, or
to variations in the contributions of Rh–S bond making. The
values of ca. 100 kJ mol−1 for DH‡ relate evidently to the ‘ease
of breaking’ the Rh–Rh bond, but not to its intrinsic (or static)
strength because Rh–Rh bond breaking is not necessarily near to
completion in the transition states, and any contributions from
Rh–S bond making would be important. Changes in other Rh–
Rh and Rh–CO interactions as the nido structure is approached
are to be expected.
15 S. P. Tunik, A. V. Vlasov and V. V. Krivykh, Inorg. Synth., 1997, 31,
239.
16 S. Martinengo and P. Chini, Gazz. Chim. Ital., 1972, 102, 344.
17 P. Chini, S. Martinengo and G. Giordano, Gazz. Chim. Ital., 1972,
102, 330.
18 (a) N. M. J. Brodie and A. J. Poe¨, Can. J. Chem., 1995, 73, 1187; (b) C.
Babij, L. Chen, I. O. Koshevoy and A. J. Poe¨, Dalton Trans., 2004,
833.
19 G. Lavigne and H. D. Kaesz, J. Am. Chem. Soc., 1984, 106, 4647.
20 S. H. Han, G. L. Geoffroy, B. D. Dombek and A. L. Rheingold,
Inorg. Chem., 1988, 27, 4355.
21 P. Chini, G. Ciani, L. Garlaschelli, M. Manassero, S. Martinengo, A.
Sironi and F. Canziani, J. Organomet. Chem., 1978, 152, C35.
22 G. Longoni, S. Campanella, A. Ceriotti, P. Chini, V. G. Albano and
D. Braga, J. Chem. Soc., Dalton Trans., 1980, 1816.
23 R. Della Pergola, L. Garlaschelli, S. Martinengo, F. Demartin, M.
Manassero and N. Masciocchi, J. Chem. Soc., Dalton Trans., 1988,
2307.
24 S. P. Tunik, A. V. Vlasov, A. B. Nikol’skii, V. V. Krivykh and M. I.
Rybinskaya, Metalloorg. Khim., 1990, 3, 387.
25 D. H. Farrar, E. V. Grachova, A. Lough, C. Patirana, A. J. Poe¨ and
S. P. Tunik, J. Chem. Soc., Dalton Trans., 2001, 2015.
26 S. Martinengo, P. Chini and G. Giordano, J. Organomet. Chem.,
1971, 27, 389.
27 A. Fumagalli, S. Martinengo, D. Galli, C. Allevi, G. Ciani and A.
Sironi, Inorg. Chem., 1990, 29, 1408.
28 A. R. Sanger, Can. J. Chem., 1985, 63, 571.
29 E. Rotondo, G. Battaglia, G. Giordano and F. P. Cusmano,
J. Organomet. Chem., 1993, 450, 245.
30 I. O. Koshevoy, A. Lough, A. J. Poe¨, S. P. Tunik and O. V. Sizova, in
preparation.
31 J. Browning, P. L. Goggin, R. J. Goodfellow, M. G. Norton, A. J. M.
Rattray, B. F. Taylor and J. Mink, J. Chem. Soc., Dalton Trans., 1977,
2061.
32 G. Ciani, L. Garlaschelli, M. Manassero, U. Sartorelli and V. G.
Albano, J. Organomet. Chem., 1977, 129, C25.
33 D. T. Brown, T. Eguchi, B. T. Heaton, J. A. Iggo and R. Whyman,
J. Chem. Soc., Dalton Trans., 1991, 677.
34 P. Serp, M. Hernandez, B. Richard and P. Kalck, Eur. J. Inorg. Chem,
2001, 2327.
Summary
These kinetic studies of some substituted Rh6 carbonyl clusters
show that the effects of the substituents on the structures of
the clusters are also made manifest in their reactivity to the
extent that they induce an opening up of the cluster which
might be spontaneous or solvent induced. Usually such opening
up of clusters is the result of photochemical activation, or
is induced by direct nucleophilic attack. The participation of
halide ions in these reactions does not, therefore, reflect any
contributions from attack on the coordinated CO ligands such
as is seen with other clusters’ reactions with hard nucleophiles.
Similar results would therefore have been obtained if only P-
donor nucleophiles had been used, except for the fact that the
weakness of bromide as a nucleophile allows for observation of
the competition between bromide attack and reverse closure
of the nido intermediate. This was the essential feature in
demonstrating that the closo to nido cluster opening is rate
determining. The complex stoichiometry of the reactions with
halides and P(OPh)3 is also illustrative of the very wide variety
of redox, substitution and fragmentation chemistry that these
Rh6 clusters can undergo.
35 M. L. Wu, M. J. Desmond and R. S. Drago, Inorg. Chem, 1979, 18,
Acknowledgements
679.
36 (a) M. A. Biddulph, R. Davis, C. H. J. Wells and F. I. C. Wilson,
J. Chem. Soc., Chem. Commun., 1985, 1287; (b) G. Li, Q. Jiang, L.
Zhang, Zh. Zhou and S. Wang, Huaxue Xuebao, 1989, 47, 449.
37 V. G. Albano, M. Sansoni, P. Chini and S. Martinengo, J. Chem. Soc.,
Dalton Trans., 1973, 651.
38 V. G. Albano, P. Chini, S. Martinengo, M. Sansoni and D. Strumolo,
J. Chem. Soc., Dalton Trans., 1978, 459.
We are grateful to the NATO, for support through a Col-
laborative Research Grant (OUTR.CRG 951482) and for the
award of a postdoctoral fellowship to I. O. K., and to the
Natural Science and Engineering Research Council, Ottawa, for
additional support.
39 A. M. Bradford, M. C. Jennings and R. J. Puddephatt,
Organometallics, 1988, 7, 792.
40 R. Ramachandran, D. S. Yang, N. C. Payne and R. J. Puddephatt,
Inorg. Chem., 1992, 31, 4236.
41 G. Bondietti, G. Laurenczy, R. Ros and R. Roulet, Helv. Chim. Acta,
1994, 77, 1869.
42 G. Laurenczy, G. Bondietti, A. E. Merbach, B. Moullet and R.
Roulet, Helv. Chim. Acta, 1994, 77, 547.
43 I. Svorstoel, H. Hoeiland and J. Songstad, Acta Chim. Scand., Ser.
B, 1984, 38, 885.
44 S. K. Malik and A. Poe¨, Inorg. Chem., 1979, 18, 1241.
45 A. J. Poe¨, Inorg. Chim. Acta, 1987, 129, L17.
46 S. P. Tunik, A. I. Yarmolenko and A. B. Nikol’skii, Inorg. Chim. Acta,
1993, 205, 71.
47 B. F. G. Johnson, J. Lewis, J. N. Nicholls, I. A. Oxton, P. R. Raithby
and M. J. Rosales, J. Chem. Soc., Chem. Commun., 1982, 289.
48 J. Schneider, M. Minelli and G. Huttner, J. Organomet. Chem., 1985,
294, 75 .
References
1 E. Corey, L. F. Dahl and W. Beck, J. Am. Chem. Soc., 1963, 85, 1202.
2 K. Wade, Adv. Inorg. Chem. Radiochem., 1976, 18, 1.
3 R. P. Hughes, in Comprehensive Organometallic Chemistry, ed.
G. Wilkinson, F. G. A. Stone and E. W. Abel, Pergamon, Oxford,
1982, vol. 5, ch. 35, p. 317.
4 G. Fachinetti, T. Funaioli and P. F. Zanazzi, J. Organomet. Chem.,
1993, 460, C34.
5 D. M. P. Mingos, J. Chem. Soc., Dalton Trans, 1974, 133.
6 D. M. P. Mingos and A. S. May, in The Chemistry of Metal Cluster
Complexes, ed. D. F. Shriver, H. D. Kaesz and R. D. Adams, VCH
Publishers, Inc., New York, 1990, ch. 2.
7 K. Nomiya and H. Suzuki, J. Organomet. Chem., 1979, 168, 115.
8 J. R. Kennedy, F. Basolo and W. C. Trogler, Inorg. Chim. Acta, 1988,
146, 75.
9 K. A. Bunten, D. H. Farrar and A. J. Poe¨, Organometallics, 2003, 22,
3448.
10 S. P. Tunik, unpublished studies of the kinetics of reactions of
Rh6(CO)16 with NCMe and halides.
11 A. J. Poe¨ and S. P. Tunik, Inorg. Chim. Acta, 1998, 268, 189.
12 C. Babij, C. S. Browning, D. H. Farrar, I. O. Koshevoy, I. S.
Podkorytov, A. J. Poe¨ and S. P. Tunik, J. Am. Chem. Soc., 2002,
124, 8922.
13 S. P. Tunik, I. O. Koshevoy, A. J. Poe¨, D. H. Farrar, E. Nordlander,
M. Haukka and T. A. Pakkanen, Dalton Trans., 2003, 2457.
14 G. Lavigne, in The Chemistry of Metal Cluster Complexes, ed.
D. F. Shriver, H. D. Kaesz and R. D. Adams, VCH Publishers, Inc.,
New York, 1990, ch. 5.
49 K. Knoll, G. Huttner, L. Zsolnai, I. Jibril and M. Wasiucionek,
J. Organomet. Chem., 1985, 294, 91.
50 C. Babij, H. Chen, L. Chen and A. J. Poe¨, Dalton Trans., 2003, 3184,
and references therein.
51 E. V. Grachova, B. T. Heaton, J. A. Iggo, I. S. Podkorytov, D. J.
Smawfield, S. P. Tunik and R. Whyman, J. Chem. Soc., Dalton Trans.,
2001, 3303.
52 E. V. Grachova, M. Haukka, B. T. Heaton, E. Nordlander, T. A.
Pakkanen, I. S. Podkorytov and S. P. Tunik, Dalton Trans., 2003,
2468.
53 L. Chen and A. J. Poe¨, Coord. Chem. Rev., 1995, 143, 265.
1 2 2
D a l t o n T r a n s . , 2 0 0 5 , 1 1 6 – 1 2 2