Kinetic study of the reaction of [Rh6(CO)16] with NO2-: Insertion of the nitrogen atom into a Rh6 cluster core

Article abstract of DOI:10.1039/b808666d

The cluster [Rh₆(CO)₁₆] reacts with (PPN)NO ₂ [PPN = N(PPh₃)₂] in dichloromethane to form the known cluster [Rh₆N(CO)₁₅]^- in which the nitrogen atom is encapsulated within a trigonal prismatic Rh₆ cluster. A kinetic study shows that the reactions proceed in six kinetically distinguishable steps followed by another four steps that are kinetically unidentifiable but that can nevertheless be identified speculatively on the basis of analogies to other reactions of carbonyl clusters. Stopped-flow studies show that the initial step involves bimolecular nucleophilic attack on a carbon atom of a CO ligand by an oxygen atom in the NO₂^- ion. This is followed by CO loss (via a [CO]-dependent and a [CO]-independent path) with formation of [Rh₆(CO)₁₄(OCONO)]^-. Activation parameters have been obtained for all three steps. [Rh ₆(CO)₁₄(OCONO)]^- has been characterized spectroscopically and has the C and N atoms coordinated to one Rh atom. Reaction of this cluster with CO leads to the fairly rapid establishment of an equilibrium with the adduct [Rh₆(CO)_n(OCONO)]^-, where n is most probably 15. Both participants in this equilibrium react more slowly with loss of CO₂, and these reactions show negligible dependence on temperature and are therefore largely entropy controlled. The overall reaction is very solvent dependent and numerous side reactions can be identified if the conditions of reaction are not closely controlled.

Full text of DOI:10.1039/b808666d

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
-
Kinetic study of the reaction of [Rh (CO) ] with NO : insertion of
6
16
2
the nitrogen atom into a Rh cluster core†
6
a
a
a
b
Claudia Babij, David H. Farrar, Anthony J. Po e¨ * and Sergey P. Tunik
Received 22nd May 2008, Accepted 16th July 2008
First published as an Advance Article on the web 15th September 2008
DOI: 10.1039/b808666d
The cluster [Rh
6
(CO)16] reacts with (PPN)NO
2
[PPN = N(PPh
3
) ] in dichloromethane to form the
2
-
known cluster [Rh
Rh
6
N(CO)15] in which the nitrogen atom is encapsulated within a trigonal prismatic
6
cluster. A kinetic study shows that the reactions proceed in six kinetically distinguishable steps
followed by another four steps that are kinetically unidentifiable but that can nevertheless be identified
speculatively on the basis of analogies to other reactions of carbonyl clusters. Stopped-flow studies
show that the initial step involves bimolecular nucleophilic attack on a carbon atom of a CO ligand by
-
an oxygen atom in the NO
[
2
ion. This is followed by CO loss (via a [CO]-dependent and a
-
CO]-independent path) with formation of [Rh
obtained for all three steps. [Rh
the C and N atoms coordinated to one Rh atom. Reaction of this cluster with CO leads to the fairly
rapid establishment of an equilibrium with the adduct [Rh
6
(CO)14(OCONO)] . Activation parameters have been
(CO)14(OCONO)] has been characterized spectroscopically and has
¯
¯
-
6
¯
¯
-
6
(CO)
n
(OCONO)] , where n is most probably
¯
¯
1
5. Both participants in this equilibrium react more slowly with loss of CO
2
, and these reactions show
negligible dependence on temperature and are therefore largely entropy controlled. The overall reaction
is very solvent dependent and numerous side reactions can be identified if the conditions of reaction are
not closely controlled.
1
2
Introduction
isotopic labeling to determine the origins of the “encapsulate”.
13
One kinetic study of such reactions has been reported and this
Metal carbonyl clusters have received extensive synthetic, struc-
tural and spectroscopic study for many years, much of it
stimulated by the “cluster-surface analogy” and their potential as
homogeneous catalysts. Although these hopes have not been real-
ized on as great a scale as was originally hoped they have not been
totally unfulfilled, and the study of this large class of compounds
remains of considerable academic importance and interest. Less
attention has been paid to the kinetic investigation of their reaction
mechanisms and energetics, without which intelligent design of
catalysts remains largely a hit-and-miss, intuitive affair. Specific
2-
established that the 86-electron octahedral cluster [Ru (CO)18]
6
1
loses CO reversibly to form the 84-electron, bi-capped tetrahedral
2
2-
[
Ru
loss of CO
cluster [Ru
6
(CO)17] . This is believed to activate a coordinated CO to
2
and eventual formation of the 86-electron octahedral
C(CO)16] . Activation parameters were obtained only
for the initial CO-dissociative reaction. Formation of a partially
2-
6
-
encapsulated N atom from the second order reaction of NO
2
with
-
14
the cluster [FeRu
3
(CO)12NO] has also been studied kinetically.
In view of the paucity of such studies we chose to study the
-
formation of the 90-electron, trigonal prismatic [Rh
6
N(CO)15
]
3
reviews of classic kinetic studies are scarce but considerable
attention has been paid to fluxionality and substitution, and
15
-
cluster via reaction of NO
2
with [Rh
6
(CO)16]. The latter has
the structure shown below as I, and the reaction involves the
insertion of the N atom from the nitrite into a rearranged Rh
4
5
16
6
7
some to fragmentation reactions, oxidative additions, adduct
6
8
9
formation, and reactions with organic molecules.
core to form II.
One group of clusters that has hardly been studied at all by
kineticists comprises those with main group elements encapsulated
within the metal core of the clusters. A large number of these
10
is known, with a very wide variety of encapsulated elements,
and the question of the kinetic effects of these elements is of
considerable interest. Apart from the reactivity of these clusters
11
when formed, the mechanisms of their formation directly from
carbonyl clusters themselves has received some study using
(
The CO ligands in I and II have been omitted for simplicity
but I has two terminal CO ligands on each Rh atom and four
triply bridging ones on alternate Rh faces. II has two terminal
CO ligands on each Rh in a triangular face, and three doubly
bridging CO ligands joining the two Rh triangles.)
a
Lash Miller Chemistry Laboratories, Department of Chemistry, University
of Toronto, 80 St. George Street, Toronto, Ontario, Canada M5S 3H6.
E-mail: apoe@chem.utoronto.ca
3
b
Department of Chemistry, St. Petersburg State University, Universitetskii
3
pr., 26 St. Petersburg, 198504, Russian Federation. E-mail: stunik@inbox.ru
This reaction had been shown to proceed in THF via a
†
Electronic supplementary information (ESI) available: FTIR spectra,
series of spectroscopically detectable intermediates but no detailed
time-dependent absorbance changes, crystallographic structural details.
See DOI: 10.1039/b808666d
17
mechanistic study was undertaken. THF did not turn out to be
5
922 | Dalton Trans., 2008, 5922–5929
This journal is © The Royal Society of Chemistry 2008

View Article Online
a suitable solvent for kinetic studies of this reaction and it was
found necessary to use dichloromethane.
with hexanes. The crystallographic data successfully resolved
an O
h
Rh cluster core but the carbonyl and nitrosyl ligands
6
were disordered in the solid state and the structure could not
be fully resolved. However, the product was characterizable
Experimental section
as PPN[Rh
6
(CO)15(NO)] by dissolving the crystals used in the
General methods
X-ray diffraction study in dichloromethane and obtaining the
-
1
corresponding IR spectrum. IR (CH
2
Cl
2
, cm ): n(CO) at 2072w,
Unless otherwise specified, all manipulations were carried out
2
033vs, 1999w, and 1755m,br 1850m; n(NO) 1711w. and this
under an atmosphere of N
techniques. Dichloromethane was dried over CaCl
from P or CaH under N , and bubbled with CO (100% CO or
0 : 60 CO–N ), N or O as necessary. [Rh (CO)16] was prepared
by refluxing a hexane solution of [Rh (CO)12] (synthesized by the
published method ) or purchased from Strem Chemicals and used
as received. [Rh (CO)15(NCMe)] was synthesized by the published
2
or CO using standard Schlenk
25
spectrum is typical of monosubstituted Rh carbonyl clusters.
6
2
and distilled
2
O
5
2
2
PPN[Rh
0.0047 mmol) of [Rh
PPN(NO ), and degassed and back-filled with N
Alternatively 4.8 mg of [Rh (CO)15(NCMe)] and 2.8 mg of
) were used.) Degassed dichloromethane (5 mL) was
(CO)14NO]. A Schlenk flask was charged with 5 mg
(CO)16] and 3.5 mg (0.0065 mmol)
three times.
6
4
2
2
2
6
(
6
4
2
2
18
(
6
6
PPN(NO
2
19
20
15
method as was PPN(NO
2
) [PPN = N(PPh
) (Aldrich) and (PPN)Cl.
N NMR spectra were recorded using a Bruker AM 500
instrument and referenced to NH NO . IR spectra were recorded
3
)
2
]. PPN( NO ) was
2
added by gas-tight syringe. The resulting brown reaction mixture
was stirred for 3 d and monitored periodically by IR spectroscopy.
1
5
prepared from Na( NO
2
1
5
Approximately 1 equivalent of CO was released during the course
2
4
3
of the reaction. The main product was air-stable and was charac-
using a Nicolet Magna IR 550 FTIR spectrophotometer and
analyzed using OMNIC 1.2 software. A Wilmad temperature
controlled IR cell (± 0.1 C) with NaCl plates, or a Wilmad
-1
terized in situ as PPN[Rh
at 2022vs and 1850m,br. N NMR (CD
6
(CO)14(NO)]. IR (CH
2
Cl
2
, cm ): n(CO)
15
2
Cl , Ar atmosphere): d
2
◦
7
29 ppm (t, JRh-N = 17 Hz), typical of a nitrogen atom coordinated
amalgamated NaCl IR cell with a water-thermostated jacket
to two metal atoms. Attempts to isolate crystalline samples led
to the growth of brown crystals but these were later identified
◦
(
± 0.1 C), were used in kinetic studies. Stopped-flow measure-
21
ments were made as described elsewhere. Crystallographic data
were collected on a Nonius Kappa CCD diffractometer using
by FTIR as PPN[Rh (CO)15(NO)] formed by reaction with CO
6
released during partial decomposition (see above).
˚
graphite monochromated Mo-Ka radiation (l = 0.71073 A) at
22
1
00 K. The X-ray crystal structure was solved by Dr Alan Lough,
using the SHELXTL/PC V5.03 package, and structural details are
given in Table S1.†
Results and discussion
1.
Fast formation of the “2042”
-
intermediate—[Rh (CO)14(OCONO)]
6
Syntheses
¯
¯
The 2-step reaction. The initial reaction of PPN(NO )
2
PPN[Rh
0.0049 mmol) of [Rh
PPN(NO ), and degassed and back-filled with CO three times.
Rh (CO)15(NCMe)] can also be used as the starting material. CO
6
N(CO)15]. A Schlenk flask was charged with 5.2 mg
1
5
or PPN( NO
Rh (CO)15(NCMe)] in dichloromethane is a very rapid formation
of an orange-brown species characterized by a very strong IR
2
)
with one equivalent of [Rh
6
(CO)16] or
(
6
(CO)16] and 3.5 mg (0.0065 mmol)
[
6
2
[
6
-
1
band at 2042 cm . This is formed both in the presence or
absence of CO, without formation of CO or any other apparent
side products, and is very air-sensitive. However, it could be
saturated dichloromethane (5 mL) was added by gas-tight syringe.
The resulting orange-brown reaction mixture was monitored
by FTIR spectroscopy and stirred overnight. A colour change
from orange-brown to yellow was observed and approximately
2
-
1
characterized spectroscopically in situ. IR (CH
2
Cl , cm ): nCO at
2
15
-
1
2081m, 2042vs, 2010m and 1769m,br; nNO at 1698w (n( NO) at
2
2
equivalents of CO
2
(identified by n(CO
2
) at 2338 cm ; e =
1
5
500 cm-1 M ), were released during the course of the reaction.
-1
23
1664w). N NMR (CD
2
Cl
2
): d 471 ppm (d, JRh-N = 60 Hz). FAB
MS (m/z): 1083 (parent ion) and successive formation of ions
with m/z = 1071, 1059, 1047, etc. This species is fairly unstable,
Yellow crystals of the air-stable product, suitable for single
crystal X-ray diffraction study, were grown by reducing the volume
in vacuo to ca. 0.5 mL and layering the solution with hexanes.
Details of the solid state structure (Table S1† ) are very close
even in the complete absence of O , so it was not possible to
2
precipitate it or to grow crystals. However, the spectra, coupled
with kinetic studies of its formation and its reaction with CO to
24
to the ones previously reported, but more precise. The FTIR
spectrum was obtained by dissolving the crystal used in the X-ray
-
form [Rh
[Rh
6
(CO)15(NO)] and CO
2
, lead to the conclusion that it is
-
-1
-
1
6
(CO)14(OCONO)] . Thus, the IR bands at 1698 cm , and at
diffraction study in dichloromethane. IR (CH
at 2034s, 1874s, and 1843br,sh.
2
Cl , cm ): n(CO)
2
664 cm- for the N-labelled form, are characteristic of an N=O
1
¯
¯
15
1
1
5
bond and the doublet in the N NMR spectrum shows that the
nitrogen is coordinated to only one Rh atom. The fact that the
PPN[Rh
uct of the slow reaction of PPN(NO
Rh (CO)15(NCMe)]) in the absence of CO, and was prepared and
6
(CO)15NO]. PPN[Rh
6
(CO)14(NO)] was the prod-
-
2
) with [Rh (CO)16] (or
6
nitrogen and carbon atoms in the O=C–O–N=O ligand are both
¯
¯
[
6
coordinated to the same Rh atom, as in III, rather than bridging
partially characterized as described below. A dichloromethane
solution was bubbled with CO whereupon immediate and com-
plete reaction occurred. Brown, air-stable crystals suitable for
single crystal X-ray crystallography were grown by reducing
the volume in vacuo to ca. 0.5 mL and layering the solution
two of them, is indicated by the fact that the IR spectrum closely
-
26
resembles that of [Rh
6
(CO)14(C
3
5
H )] , IV. Both spectra show
only one strong band in the nCO region whereas a more complex
-
spectrum would be expected if the [Rh
6
(CO)14(OCONO)] cluster
-
2
¯
7
¯
contained a bridging bidentate OCONO ligand.
¯
¯
This journal is © The Royal Society of Chemistry 2008
Dalton Trans., 2008, 5922–5929 | 5923

View Article Online
6
Table 1 Rate constants for step 1 of reactions of [Rh (CO)16] with
PPN(NO )
2
◦
a
k
2
b,c/M-1 s-1
T/ C
N
Under N
2
2
2
1
1
5.0
0.0
5.0
0.0
24
12
15
15
22
31.9 ± 0.6
22.4 ± 0.5
16.1 ± 0.4
11.1 ± 0.3
7.71 ± 0.2
5.0
Under 1 atm CO
The kinetics of formation of this “2042” species were too rapid
for normal FTIR monitoring but there were sufficient changes in
the UV-vis absorbance for the reactions to be monitored by the
stopped-flow method at 375 to 475 nm. The spectra were poorly
defined, and no isosbestic points were evident, but the intensity
changes occurred in two well-separated stages, the first showing
an increase in absorbance and the second a decrease, as shown in
Fig. S1 and S2.†
2
5.0
5.0
4
5
32.8 ± 1.0
8.2 ± 0.3
Under 1 atm O
2
2
0.0
5.0
4
5
19.8 ± 0.8
6.7 ± 0.2
a
b
Number of independent values of k
2
.
Uncertainties correspond to a
. This was derived
from the scatter of 88 independent determinations around the Eyring
In typical runs, a solution of [Rh
6
(CO)16] in dichloromethane
ca. 10- M) was combined with a solution of PPN(NO
) in
dichloromethane (1–3 mM at higher temperatures and 1–8 mM at
lower ones). The [Rh (CO)16] solutions were filtered if necessary
and degassed prior to use, and concentrations were estimated
standard deviation of ± 8.2% for each value of k
2
4
(
2
c
plot: [PPN(NO
2
)] varied from ca. 1 to 3 mM at higher temperatures and
1
–8 mM at lower temperatures. Reactions were monitored separately at
6
l = 375, 400, 425, 450 and 475 nm.
3
-1
-1
-1
using e = (29 ± 4) ¥ 10 M cm (obtained at 2077 cm in
dichloromethane). To protect against decomposition, PPN(NO
2
)
6
Table 2 Rate constants for step 2 of reactions of [Rh (CO)16] with
PPN(NO2)
28
solutions were used within 10 min of being prepared. Reactions
were monitored at five wavelengths in the range 375–475 nm, and
reaction times ranged from 0.2 s (step 1) to 2 s (step 2). Rate
constants, kobs, were obtained by fitting absorbance changes for
each step separately to single exponential functions (Fig. S1 and
S2†). The average of three or more independent determinations of
◦
a
obs /s-1
(kN2 - kCO)/s-1
26.8 ± 2.8
16.5 ± 3.1
6.6 ± 1.80
6.2 ± 1.3
3.55 ± 0.32
—
Atmos.
T/ C [PPN(NO
2
)]/mM
k
N2
CO
25.0
25.0
20.0
20.0
15.0
15.0
10.0
10.0
5.0
1.04–2.87
1.22–2.86
1.66
1.75
1.71
1.18
1.80
1.81
50.7 ± 2.5
23.9 ± 1.3
26.9 ± 2.9
10.4 ± 1.2
14.2 ± 1.6
7.60 ± 0.84
10.7 ± 1.2
4.45 ± 0.49
5.8 ± 0.3
N
2
CO
kobs at different wavelengths were reproducible to within 5–10%.
N
2
Values for the first step were obtained over as wide a range
CO
N
2
of [PPN(NO
was the necessity to maintain at least a 10-fold excess of PPN(NO
over [Rh (CO)16], and at [PPN(NO )] > 8 mM the reaction was
2
)] as possible. The limiting factor at low [PPN(NO )]
2
CO
2
)
N
2
1.16–2.22
1.16–4.79
1.18–2.24
6
2
CO
1 : 1 N
5.0
5.0
2.25 ± 0.11
2.9 ± 0.2
too fast to be monitored by our stopped flow equipment, even
at lower temperatures. The rate constants were found to increase
2
–CO
a
Uncertainties are based on the value of the standard deviation of an
individual measurement obtained from the Eyring analysis of the 12 points
used to determine the activation parameters for the reactions under CO.
linearly with increasing [PPN(NO
2
)] in accord with eqn (1)
k
obs = a + k
2
[PPN(NO )]
2
(1)
where k is a second order rate constant and a is a constant. Values
of a were not significant with respect to the standard deviations
provided by the suitably weighted least-squares analysis, and were
2
a “non-dissociative” path. The data should therefore fit to an
equation such as kobs = kCO + (kN2 - kCO)/{1 + x[CO]}, where kCO
is the lower limit of the rate constant at high [CO] and the second
term describes the rate constants for the “CO-dissociative” path.
The best fit was obtained using the Solver function in Excel and
the line in Fig. 1 was drawn with the parameters obtained, viz:
frequently slightly negative. Values of k can therefore be obtained
2
from kobs/[nitrite] and are given in Table 1 where it can be seen
that the values did not vary with [nitrite], showing that there are
no effects due to ion-pairing. The latter is likely to be significant
4
-1
4
-1
1
0 kCO = 1.3 s , 10 (kN2 - kCO) = 4.4 s and x = 3.7. The results
3
-1
in view of the known association constant of 1.5 ¥ 10 M in
indicate that rate constants under 1 atm CO are close to the lower
limiting value and an Eyring analysis of the values of kobs for
reaction under 100% CO provides good values for the activation
29
CH
2
Cl
2
.
Step 1 was also found not to be significantly affected
by CO or O
2
. The temperature dependence of the k values was
2
also studied and Eyring analysis leads to the activation parameters
‡
-1
‡
parameters DH (non diss) = 77.5 ± 2.3 kJ mol and DS (non diss) = 41.4 ±
‡
-1
‡
-1
-1
DH
2
= 46.8 ± 0.6 kJ mol and DS
2
= 17.3 ± 2.6 J K mol .
-1
-1
8.2 J K mol for the non-dissociative path. A similar analysis of
Values of kobs for step 2 are listed in Table 2. They are
‡
the less precise values of (kobs - kCO) leads to values of DH (diss)
~
independent of [PPN(NO
2
)] and decrease with increasing [CO]
-1
‡
-1
-1
6
7 kJ mol and DS (diss) ~ 8 J K mol for the dissociative path.
as shown in Fig. 1. This decrease suggests an approach to a
lower limiting rate at [CO] ≥ 1 atm, and this in turn suggests that
two paths are being followed, one a CO-dissociative and another
Discussion of the 2-step reaction. The second order nature of
-
the first step immediately suggests nucleophilic attack by the NO
2
5
924 | Dalton Trans., 2008, 5922–5929
This journal is © The Royal Society of Chemistry 2008

View Article Online
one expected from an intramolecular S
N
2 reaction. We suggest
the possibility that both reactions are essentially CO-dissociative
but that in one case (the non-dissociative path) the N atom of the
-
[
Rh-(OCONO)] moiety is in a suitable position, as in V, to take
immediate advantage of the vacant coordination site generated by
the CO dissociation to form a Rh–N bond so that the free CO
present is not able to compete for the vacant site. In the other case
(
the dissociative path) the N atom is remote from the vacant site,
as in VI, and the CO can compete for it. The rather similar values
‡
‡
of DH (non diss) and DH (diss) tend to support this.
II. Formation of [Rh
6
N(CO)15]- from
◦
2
-
Fig. 1 Dependence on [CO] of kobs at 5 C for “step 2” expressed as a
mole fraction in CO–N mixtures.
[Rh
6
(CO)14(j C,N-OCONO)]
¯
¯
2
Time resolved infrared spectroscopy studies. In a typical run,
a Schlenk flask was charged with ca. 4 to 7 mg of [Rh (CO)16] or
[Rh (CO)15(NCMe)] and ca. 3 to 5 mg PPN(NO ) and degassed
and back-filled with N or CO. Degassed or CO saturated
ion (either free or ion-paired) on the carbonyl cluster, specifically
by a hard oxygen atom of the nitrite ion at an electrophillic C
atom of one of the terminal CO ligands. This type of reaction
has been proposed for the second order reactions of nitrite with
6
6
2
2
dichloromethane (2–5 mL) was added by gas-tight syringe. The
resulting solution was stirred and transferred by gas-tight syringe
to an IR cell, and spectra were scanned repetitively in the carbonyl
stretching region. Alternatively, solutions of each reactant were
prepared individually and then mixed. In order to minimize inter-
fering side reactions and maximize nitrido cluster formation the
reaction conditions described above (Syntheses) were employed.
30
31
-
[
Fe(CO)
5
]
and [Cr(CO)
6
] , eventually to form [Fe(CO) (NO)]
3
-
and [Cr(CO)
such as [Fe(CO)
proposed, the former because of the second order kinetics, and
the latter by analogy with earlier proposals. These reactions, and
those of [M
those of [Rh
4
(NO)] , respectively. Involvement of intermediates
-
-
4
CO(ONO)] and [Fe(CO) (OCONO)] were
3
¯
¯
32
28
3
(CO)12] (M = Ru or Os), are much slower than
◦
6
(CO)16] and the latter reacts even more rapidly with
The concentration of CO in a saturated solution at 25 C was
3
-1 -1
◦
33
34
hydroxide (kOH = 15 ¥ 10 M
s
at 25 C in CH
2
Cl
2
).
estimated to be ca. 4 mM and in all cases was well in excess of
These reactions of nitrite with [Rh
6
(CO)16] demonstrate the tran-
the Rh cluster concentration and what was needed in view of the
stoichiometry (eqn (1)).
sitory existence of the adduct formed immediately by nucleophilic
-
attack of NO
2
at a coordinated CO ligand, and define it by its
A typical set of spectra is shown in Fig. 2. It shows clearly the loss
-
1
-
rate of further reaction but with no spectroscopic characterization.
The spectroscopic characterization of the product of further
of bands at 2042 and 1774 cm due to the [Rh (CO)14(OCONO)]
6
-
1
¯
¯
cluster, the growth of bands at 2034 and 1874 cm due to the
nitrido product, and the growth of a band at 2338 cm due to
2
-
-1
reaction, [Rh
6
(CO)14(k C,N-OCONO)] , is the first of its kind,
2
¯
¯
although analogous k C,N-(OCONO) compounds have been
the CO
2
released as in eqn (1). However, a clear and significant
30,31
¯
¯
-1
postulated.
additional feature is the growth and decay of a band at 2013 cm
The activation entropy for the first step is unusual for a second
order reaction in being quite positive. No activation parameters
as shown in Fig. 3. When the absorbance of this species is
-
at a maximum, spectra due to both the [Rh
6
(CO)14(OCONO)]
30,31
-
-
¯
¯
were reported
for NO
2
reactions with [Fe(CO)
5
] or [Cr(CO)
6
]
reactant cluster and the [Rh
6
N(CO)15] product can be subtracted
so it is not clear if this feature is unique. Several factors might
account for it. Thus the 12 terminal CO ligands susceptible to
from that of the mixture to give the complete spectrum of this
new species as shown in Fig. S3.† In addition to the band at
-
1
-1
nucleophilic attack in the Rh
6
cluster would contribute ca. +20 J
2013 cm there is a strong band at 2048 cm , a shoulder at ca.
-
1
-1
‡
-1
-1
K
mol to DS
2
, compared with attack at only one susceptible
2000 cm and weaker bands at 1838 and 1778 cm . The two latter
bands can be most probably assigned to edge- and face-bridging
CO ligands, both being very typical for multinuclear carbonyl
CO. Also the loss of translational entropy that must accompany
the reaction could also be offset to a considerable extent by
1
-1
the large vibrational entropy expected of the flexible [Rh{k C-
clusters. There is no band at 1874 cm characteristic of the doubly
-
CO(ONO)}] moiety formed. Desolvation of the nucleophilic
bridging CO ligands present in the final product. The cleanness
of the reactions can be estimated in several ways. Thus the ratio
oxygen ion would also contribute.
-
1
The nature of the second step also shows some unusual features.
The CO-inhibited, and presumably dissociative, contribution to
the rates shows a moderately positive activation entropy but the
alternative “non-dissociative” displacement of the CO shows a
substantially positive activation entropy rather than a negative
of the intensities of the product bands at 2034 and 1874 cm is
-
1.4 : 1 and when the spectrum of pure [Rh
6
N(CO)15] is subtracted
from a product solution showing this ratio only a weak band at
-
1
2058 cm is apparent. However, when the ratio of the 2034 and
-
1
1874 cm bands is ca. 2 : 1, quite strong bands at 2048, 2033 and
This journal is © The Royal Society of Chemistry 2008
Dalton Trans., 2008, 5922–5929 | 5925

View Article Online
Fig. 2 Typical repetitive spectral scans during reaction of the “2042” cluster with CO.
a slower one. The intensity changes for the slower one can be
studied separately, after the faster one is completed, when it is
found that better fits are still obtained when double exponentials
are used although the contributions of these third stages to the
overall rates are small. This leads us to regard them as being due
only to minor departures of the kinetics from the overall two-step
first-order reactions.
The rate constants obtained in this way are given in Table 3
and they do fall into two distinct groups, corresponding to faster
-
1
and slower steps. The rise and fall of absorbance at 2013 cm
provide clearly distinct values for the slower and faster steps (see
-
1
above) and the decrease in absorbance at 2042 cm occurs in
-
1
two stages, as does the increase in absorbance at 2223 cm , thus
again providing rate constants for the slower and faster steps. The
data are compatible with the faster steps being an approach to
Fig. 3 Absorbance change at 2013 cm- during reaction of the “2042”
cluster with CO.
1
2
-
an equilibrium between the [Rh (CO)14(k C,N-OCONO)] cluster
6
-
1
¯
¯
and the “2042 cm intermediate”, and this equilibrium mixture
then reacts relatively slowly with CO. This slower step must be
followed by several faster ones to form the final product and
a possible set of reactions is shown in Scheme 1. When the
equilibrium between (1) and (2a) has been reached the pseudo
first order rate constant for reaction of the equilibrium mixture
2
015 cm- are evident, as well as a broad band at 1772 cm-1 due
to triply bridging CO. And when the reaction is carried out under
an atmosphere consisting of only 40% CO a multiplicity of “side
products” is seen with only minor amounts of nitrido product.
1
The yield of CO
2
, which should be 2 molar equivalents, can also
is given by k = {k
a
+ k
b
K[CO]}[CO]/{1 + K[CO]}, where k
a
and
be used to estimate the cleanness of reactions. Interestingly, as
will be seen below, these variations show very little affect on the
kinetics of the reactions so the side reactions seem to occur only
after the slow and rate determining steps have been passed.
Rate constants were obtained by fitting the time-dependent
intensities of the various bands shown in Fig. 2 to double
k
b
are rate constants for the loss of (1) and (2a), respectively, and
K is the equilibrium constant for the formation of (2a) from (1).
Unfortunately the major changes in the nature of the reactions as
the concentration of CO is lowered, coupled with the inability to
study the reactions at higher pressures of CO, preclude verification
of this rate equation.
exponential functions. This is shown very clearly by the data for the
2
falls. However, quite good fits were also obtained when intensities
increased or decreased steadily throughout the reactions, all this
demonstrating the presence of an initial fast reaction followed by
Two major ideas are involved in this scheme. Firstly that
nucleophilic attack on carbonyl clusters usually occurs to form
adducts concurrently with breaking one of the metal–metal bonds.
This has been established through kinetic and thermodynamic
013 cm- band in Fig. 3, where the absorbance first rises and then
1
8,35
36
measurements and by structural studies. Secondly the eventual
5
926 | Dalton Trans., 2008, 5922–5929
This journal is © The Royal Society of Chemistry 2008

View Article Online
4
-
Table 3 Rate constants (10 k/s) for reactions of [Rh
6
(CO)14(ONOCO)] under CO
¯
¯
◦
a
b
T/ C
2013↑
2013↓
2042↓
2042↓
2080↓
2034↑
1874↑
1774↓
2338↑
CO
2223↑
(N O)
2223↑
(N O)
R
CO
2
(
2
)
2
2
35.1
30.0
30.0
30.0
30.0
29.5
25.4
25.4
25.4
25.4
25.0
20.1
20.0
20.0
15.1
14.8
—
[27]
16
14
—
[38]
8
—
23
25
25
11
[43]
24
8
10
18
—
9
3.9
3.3
2.5
1.4
3.3
3.7
—
—
2.0
—
3.4
1.1
3.3
3.1
6.3
5.7
26
22
19
[6]
[50]
25
9
4.3
1.0
—
4.1
3.9
2.4
1.8
4.6
4.0
2.3
2.9
1.5
1.9
2.1
3.9
2.9
2.7
3.3
2.8
—
—
—
—
4.6
—
5.4
7.7
—
4.2
—
2.6
—
—
—
—
[13]
[9]
19
26
25
12
32
25
20
2.8
3.0
2.8
1.5
3.4
3.2
4.8
0.8
1.6
1.4
1.5
1.4
1.8
1.4
1.9
2.1
1.9
1.5
1.4
1.4
1.5
1.5
1.5
1.5
—
1.4
1.7
2.0
1.8
1.2
1.4
1.3
2.3
1.6
2.3
1.6
2.0
1.7
2.1
3.0
2.3
—
c
1.1
1.1
1.3
—
2.2
1.7
[3.2]
—
—
0.82
—
0.81
0.77
1.7
0.81
1.23
2.8
3.0
6.7
6.1
6.8
5.4
1.0
2.2
5.8
2.0
3.3
1.1
1.6
1.6
0.6
1.5
1.5
1.4
1.4
9
9.3
—
7.5
—
—
—
17
—
13
15
13
12
1
±
—
—
7.3
10
20
15
18
12
3.7
2.7
1.4
4.7
3.1
2.5
7.1
9.4
18.3
20
18
14
11
17
± 2
1.0
0.5
1.2
3.2
1.9
0.5
18
23
22
20
23
9d.5
—
—
d
d
d
d
d
d
d
d
d
d
3.0
1.0
18
3.3
± 0.4
15
2.0
± 0.4
2.9
± 0.2
4.9
± 0.8
4.1
± 0.5
2.2
± 0.15
± 1.7
± 2
± 0.3
—
—
a
d
Ratio of product nitrido bands at 2034 and 1874 cm-1. b Molar equivalents of CO
Averages with their standard deviations.
c
2
produced. Values in square brackets were omitted from averaging.
Scheme 1
conversion of the octahedral Rh
nitrido cluster is accompanied by a decrease in the number of
Rh–Rh connectivities from 12 to 9. In other words 3 Rh–Rh
6
cluster to the trigonal prismatic
These reactions involve associative attack by CO, concurrently
with Rh–Rh bond breaking. Otherwise the rates of these reactions
would not be dependent on [CO]. Unfortunately, as was said
above, decreasing [CO] leads to major changes in the products
and we were unable to study reactions at elevated pressures of
bonds have to be broken, and this involves overall addition
2
of 3 CO ligands. Bonds in the octahedral [Rh
6
(CO)14(k C,N-
-
OCONO)] cluster that are candidates for this breakage are
CO. Attempts to study the effect of added CO
inconclusive.
2
also proved
¯
¯
indicated by dotted connections in the structures in Scheme 1.
The initial fast reactions with CO involve breaking one Rh–
Rh bond, and the slower reactions involve CO addition and
As drawn, the product 2b is an isomer of the cluster
-
[Rh
6
(CO)15(NO)] which was shown above to be too stable to be an
¯
displacement of CO
unreacted [Rh
2
from either the CO adduct (2) or the
intermediate in the formation of the nitrido product. If formation
of 2b does occur then addition of CO must compete successfully
with isomerization to the closo form of the cluster. Direct reaction
-
6
(CO)14(OCONO)] cluster. This CO
2
comes from
¯
¯
the OCONO ligands, which are therefore converted to NO ligands.
¯
¯
This journal is © The Royal Society of Chemistry 2008
Dalton Trans., 2008, 5922–5929 | 5927

View Article Online
of 1 to form 2b is suggested by the fact that formation of the nitrido
product begins immediately the 2042 cm species is formed, and
before the equilibrium between 1 and 2a is established.
a scholarship (to C. B.), to the Natural Science and Engineering
Research Council, Ottawa, for additional support, and to Dr Alan
Lough for the crystal structure determination.
-
1
When the 3 CO ligands have been added in the remaining three
steps, and 3 Rh–Rh bonds have been broken, the two eclipsed
Rh
Rh
3
triangles can rotate to form the trigonal prismatic cluster
(CO)17(NO)] . It is this cluster that undergoes reduction of the
References
-
[
6
1
M. G. Richmond, Coord. Chem. Rev., 2006, 249, 2763 (and previous
surveys by this author in the same journal);; Transition Metal Carbonyl
Cluster Chemistry, ed. P. J. Dyson and J. S. McIndoe, Gordon and
Breach, Amsterdam, 2000; Metal Clusters in Chemistry, ed. P. Braun-
stein, L. A. Oro and P. R. Raithby, Wiley-VCH, Weinheim, 1999,
vol. 1–3; The Chemistry of Metal Cluster Complexes, ed. D. F. Shriver,
H. D. Kaesz and R. D. Adams, VCH Publishers, Inc. New York, 1990;
Metal Clusters, ed. M. Moskovits, Wiley-Interscience, New York, 1986.
NO ligand to the “nitrido ion”, with concurrent loss of CO
2
and
penetration of the N atom, as a nitrido ion, into the centre of a Rh
6
cluster from the periphery of the cluster. This can therefore claim
to be the most interesting step in this whole series of reactions
but, unfortunately, it is clearly speculative since the slow, rate
determining steps are much earlier in the series. All that can be
-
2 E. L. Muetterties, Bull. Soc. Chim. Belg., 1975, 84, 959; E. L.
Muetterties, T. N. Rhodin, E. Band, C. Brucker and H. Pretzer, Chem.
Rev., 1979, 79, 91.
said is that these later postulated steps from [Rh
6
(CO)15(NO)]
-
to the final [Rh
6
N(CO)15] are entirely reasonable in the light of
the kinetic behavior of clusters in general, and that they must all
be faster than the slow steps defined by the smaller, but not very
small, rate constants in Table 3.
3 D. J. Darensbourg, 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. 4; A. J. Po e¨ , in Metal Clusters, ed. M. Moskovits,
Wiley-Interscience, New York, 1986, ch. 4.
The rate constants listed in Table 3 show rather poor repro-
ducibility (with standard deviations of individual determinations
of a rate constant in the range of 20–40%). This is hardly
surprising in view of the great sensitivity of the reactions to the
experimental conditions, and the inherent problems of double
exponential analyses which involve initial assumptions of molar
4 B. T. Heaton, C. Jacob, I. S. Podkorytov and S. P. Tunik, Inorg.
Chim. Acta, 2006, 359, 3557; S. A. Waterman and M. G. Humphrey,
Organometallics, 1999, 18, 3116; L. J. Farrugia, J. Chem. Soc., Dalton
Trans., 1997, 1783; B. E. Mann, J. Chem. Soc., Dalton Trans., 1997,
1
457; D. Lentz and R. Marschall, Organometallics, 1991, 10, 1487;
B. F. G. Johnson and A. Bott, J. Chem. Soc., Dalton Trans., 1990, 2437.
C. Babij, H. Chen, L. Chen and A. J. Po e¨ , Dalton Trans., 2003, 3184
and references therein.
K. A. Bunten, C. Moreno and A. J. Po e¨ , Dalton Trans., 2005, 1416 and
references therein.
5
6
8b
absorbances etc. In addition, progressive increase of the CO
2
concentration might retard the reactions gradually, leading to
an apparent slower stage of reaction. Although the temperatures
7 G. Kramer and A. J. Po e¨ , Inorg. Chem., 1981, 20, 1362.
8 (a) D. H. Farrar, A. J. Po e¨ and Y. Zheng, Inorg. Chim. Acta, 2000, 300–
◦
cover a range of 20 C this does not seem to contribute to the
3
02, 668; (b) A. J. Po e¨ and C. Moreno, Organometallics, 1999, 18, 5518.
irreproducibility, and this implies that the entropies of activation
(
Note: the first step in Scheme 1 in this reference should be reversible).
9 A. J. Po e¨ , C. N. Sampson and R. T. Smith, J. Am. Chem. Soc., 1986,
-
1
-1
are ca. -200 or -300 J K mol for the faster and slower
reactions, respectively; i.e. they are both enthalpically very facile
but entropically very unfavourable. The enthalpies of bond making
and bond breaking must therefore compensate each other but the
molecular motions involved must be very restricted and therefore
improbable. The groups of fast and slow reactions show reasonable
1
08, 5459.
1
0 B. F. G. Johnson and C. M. Martin, in Metal Clusters in Chemistry, ed.
P. Braunstein, L. A. Oro and P. R. Raithby, Wiley-VCH, Weinheim,
1
999, vol 2, ch. 2.11, p. 877.
11 J. Hao and A. J. Po e¨ , Transition Met. Chem., 1998, 23, 739 and
references therein.
1
2 P. F. Jackson, B. F. G. Johnson, J. Lewis, W. J. H. Nelson and M.
internal consistency. The growth of the very minor product N
2
O,
and of the band at 2034 cm , show poorer reproducibility (60 and
0%, respectively) than the others. The former is perhaps because
McPartlin, J. Chem. Soc., Dalton Trans., 1982, 2099.
-
1
13 S. A. Roth and J. R. Shapley, J. Coord. Chem., 1994, 22, 163.
7
14 D. E. Fjare and W. L. Gladfelter, J. Am. Chem. Soc., 1984, 106,
4
799.
it is a very minor product, and the latter because it is very close
to other major and contrary changes in wavenumber. The average
rate constants are, however, in good agreement with the others.
1
5 S. Martinengo, G. Ciani, A. Sironi, B. T. Heaton and J. Mason, J. Am.
Chem. Soc., 1979, 101, 7095.
16 E. R. Corey, L. F. Dahl and W. Beck, J. Am. Chem. Soc., 1963, 85, 1202.
7 R. E. Stevens, P. C. C. Liu and W. L. Gladfelter, J. Organomet. Chem.,
985, 287, 133.
18 S. Martinengo, P. Chini and G. L. Giordano, J. Organomet. Chem.,
971, 27, 389.
19 S. P. Tunik, A. V. Vlasov and V. V. Kryvikh, Inorg. Synth., 1996, 31,
39.
0 R. E. Stevens, D. J. Yanta and W. L. Gladfelter, Inorg. Synth., 1983, 22,
63.
21 K. A. Bunten, D. H. Farrar and A. J. Po e¨ , Organometallics, 2003, 22,
448.
1
1
Conclusions
1
These results demonstrate that spectroscopic and kinetic studies
of complex reactions of metal carbonyl clusters are capable of
providing clear mechanistic information about many of the steps
involved. Rational speculation regarding the other steps—based
on mechanistic information obtained on analogous systems—can
lead to a convincing and coherent overall picture. In this case
2
2
1
3
2
2 D. Farrar, C. Krywiak (n e´ e Babij), A. Po e¨ and A. Lough, private
communication to the Cambridge Structural Data Base, deposition
number CCDC 683996.
reaction involves the effective reduction of the nitrogen atom in
-
NO
2
from plus three to minus three, with its concomitant insertion
cluster.
23 K. A. Bunten, D. H. Farrar, A. J. Po e¨ and A. Lough, Organometallics,
2
002, 21, 3344.
into the centre of a Rh
6
2
2
2
2
4 R. Bonfichi, G. Ciani, A. Sironi and S. Martinengo, J. Chem. Soc.,
Dalton Trans., 1983, 253.
5 D. H. Farrar, E. V. Grachova, A. Lough, C. Patirana, A. J. Po e¨ and S. P.
Tunik, J. Chem. Soc., Dalton Trans., 2001, 2015.
Acknowledgements
6 G. Ciani, A. Sironi, P. Chini, A. Ceriotti and S. Martinengo,
J. Organomet. Chem., 1980, 192, C39.
7 S. P. Tunik, I. O. Koshevoy, A. J. Po e¨ , D. H. Farrar, E. Nordlander, M.
Haukka and T. A. Pakkanen, Dalton Trans., 2003, 2457.
We are grateful to NATO for support through a Collaborative
Research Grant (OUTR CRG 951482), to the Ontario Graduate
Scholarships in Science and Technology Program for the award of
5
928 | Dalton Trans., 2008, 5922–5929
This journal is © The Royal Society of Chemistry 2008

View Article Online
2
2
8 J. C. Fanning and L. K. Keefer, J. Chem. Soc., Chem. Commun., 1987,
55.
9 I. Svorstoel, H. Hoeiland and J. Songstad, Acta Chem. Scand., Ser. B,
984, 38, 885.
33 C. Babij, D. H. Farrar, I. O. Koshevoy, A. J. Po e¨ and S. P. Tunik, Dalton
Trans., 2005, 116.
34 A. J. Po e¨ , C. N. Sampson, R. T. Smith and Y. Zheng, J. Am. Chem.
Soc., 1993, 115, 3174.
9
1
3
3
0 R. E. Stevens and W. L. Gladfelter, Inorg. Chem., 1983, 22, 2034.
1 D. R. Mantell and W. L. Gladfelter, J. Organomet. Chem., 1988, 347,
35 D. H. Farrar, A. J. Po e¨ and Y. Zheng, J. Am. Chem. Soc., 1994, 116,
6252.
36 B. F. G. Johnson, J. Lewis, J. N. Nicholls, J. Puga, P. R. Raithby, M. J.
Rosales, M. McPartlin and W. Clegg, J. Chem. Soc., Dalton Trans.,
1983, 277.
3
33.
2 D. T. Doughty, G. Gordon and R. P. Stewart, Jr., J. Am. Chem. Soc.,
979, 101, 2645.
3
1
This journal is © The Royal Society of Chemistry 2008
Dalton Trans., 2008, 5922–5929 | 5929