Sensitive and selective chromogenic sensing of carbon monoxide via reversible axial CO coordination in binuclear rhodium complexes

Article abstract of DOI:10.1021/ja206251r

The study of probes for CO sensing of a family of binuclear rhodium(II) compounds of general formula [Rh₂{(XC₆H ₃)P(XC₆H₄)}_n(O₂CR) _4-n]·L₂ containing one or two metalated phosphines (in a head-to-tail arrangement) and different axial ligands has been conducted. Chloroform solutions of these complexes underwent rapid color change, from purple to yellow, when air samples containing CO were bubbled through them. The binuclear rhodium complexes were also adsorbed on silica and used as colorimetric probes for "naked ey" CO detection in the gas phase. When the gray-purple colored silica solids containing the rhodium probes were exposed to air containing increasing concentrations of CO, two colors were observed, in agreement with the formation of two different products. The results are consistent with an axial coordination of the CO molecule in one axial position (pink-orange) or in both (yellow). The crystal structure of 3·(CO) ([Rh₂{(C₆H₄)P(C₆H ₅)₂}₂(O₂CCF₃) ₂]·CO) was solved by single X-ray diffraction techniques. In all cases, the binuclear rhodium complexes studied showed a high selective response to CO with a remarkable low detection limit. For instance, compound 5· (CH₃CO₂H)₂ ([Rh₂{(m- CH₃C₆H₃)P(m-CH₃C₆H ₄)₂}₂(O₂CCH₃) ₂]· (CH₃CO₂H)₂) is capable of detection of CO to the "naked ey" at concentrations as low as 0.2 ppm in air. Furthermore, the binding of CO in these rhodium complexes was found to be fully reversible, and release studies of carbon monoxide via thermogravimetric measurements have also been carried out. The importance of the silica support for the maintenance of the CO-displaced L ligands in the vicinity of the probes in a noninnocent manner has been also proved.

Full text of DOI:10.1021/ja206251r

ARTICLE
pubs.acs.org/JACS
Sensitive and Selective Chromogenic Sensing of Carbon Monoxide via
Reversible Axial CO Coordination in Binuclear Rhodium Complexes
†
,‡,§
†,‡
†,‡
,†,‡,§
María E. Moragues,
Julio Esteban, Jos ꢀe Vicente Ros-Lis, Ram ꢀo n Martínez-M ꢀa ~n ez,*
,|| †,‡ †,‡,§
†
,‡,§
M. Dolores Marcos,
Manuel Martínez,* Juan Soto, and F ꢀe lix Sancen ꢀo n
†
Centro de Reconocimiento Molecular y Desarrollo Tecnol ꢀo gico (IDM), Unidad Mixta Universidad Polit ꢀe cnica de Valencia-
Universidad de Valencia, Spain
‡
Departamento de Química, Universidad Polit ꢀe cnica de Valencia, Camino de Vera s/n, E-46022 Valencia, Spain
§
CIBER de Bioingeniería, Biomateriales y Nanomedicina (CIBER-BBN)
Departament de Química Inorg ꢁa nica, Universitat de Barcelona, Martí i Franqu ꢁe s 1-11, E-08028 Barcelona, Spain
S Supporting Information
b
ABSTRACT: The study of probes for CO sensing of a
family of binuclear rhodium(II) compounds of general
formula [Rh {(XC H )P(XC H )} (O CR) ] L contain-
2
6
3
6
4
n
2
4ꢀn
3
2
ing one or two metalated phosphines (in a head-to-tail
arrangement) and different axial ligands has been conducted.
Chloroform solutions of these complexes underwent rapid
color change, from purple to yellow, when air samples contain-
ing CO were bubbled through them. The binuclear rhodium
complexes were also adsorbed on silica and used as colorimetric
probes for “naked eye” CO detection in the gas phase. When the gray-purple colored silica solids containing the rhodium probes
were exposed to air containing increasing concentrations of CO, two colors were observed, in agreement with the formation of two
different products. The results are consistent with an axial coordination of the CO molecule in one axial position (pink-orange) or in
both (yellow). The crystal structure of 3 (CO) ([Rh {(C H )P(C H ) } (O CCF ) ] CO) was solved by single X-ray diffraction
3
2
6
4
6
5 2 2
2
3 2
3
techniques. In all cases, the binuclear rhodium complexes studied showed a high selective response to CO with a remarkable low
detection limit. For instance, compound 5 (CH CO H) ([Rh {(m-CH C H )P(m-CH C H ) } (O CCH ) ] (CH CO H) )
3
3
2
2
2
3
6
3
3
6
4 2 2
2
3 2
3
3
2
2
is capable of detection of CO to the “naked eye” at concentrations as low as 0.2 ppm in air. Furthermore, the binding of CO in these
rhodium complexes was found to be fully reversible, and release studies of carbon monoxide via thermogravimetric measurements
have also been carried out. The importance of the silica support for the maintenance of the CO-displaced L ligands in the vicinity of
the probes in a noninnocent manner has been also proved.
’
INTRODUCTION
over a few hours (a CO level between 70 and 100 ppm) includes
flu-like symptoms, such as headaches, sore eyes, and a runny
nose. Medium exposure (a CO level between 150 and 300 ppm)
will produce dizziness, drowsiness, and vomiting. Extreme ex-
posure (a CO level of 400 ppm and higher) will result in
unconsciousness, brain damage, and death. Unfortunately, it is
quite usual that people with CO poisoning overlook the symp-
toms (e.g., headache, nausea, dizziness, or confusion), and
unintentional CO exposure accounts for an estimated 15 000
emergency department visits and 500 unintentional deaths in the
United States alone each year.
Carbon monoxide is an airborne contaminant difficult to detect
since it is colorless and odorless and toxic at low concentration
levels. Moreover, this compound may present a real danger to
humans in automobiles, airplanes, industrial plants, mines,
homes, and other environments in which persons may be present
for extended periods of time. Typical anthropogenic sources of
this deadly pollutant involve combustion engines and improper
burning of other fuels. Examples include residential furnaces,
gasoline engines, charcoal grills, gas heaters, and others. Carbon
monoxide toxicity lies in the fact that hemoglobin presents about
1
Therefore, there is an increasing interest for the development
of chemical sensor systems capable of detecting the presence of
carbon monoxide in air at low concentrations. Most existing CO
2
00 times higher affinity for carbon monoxide than for oxygen.
Therefore, carbon monoxide can readily displace oxygen on
2
hemoglobin, reducing the amount of the latter that can be
carried to tissues and impairing the mechanisms of cellular
respiration. For this reason, carbon monoxide is toxic to humans
at relatively low concentrations. For healthy adults, CO be-
5
sensors use metal-oxide semiconductors (MOS) that have been
3
widely chosen as gas-sensing devices and applied in monitoring
4
comes toxic when it reaches a level higher than 50 ppm with
continuous exposure over an eight hour period. Mild exposure
Received: July 20, 2011
Published: August 24, 2011
r 2011 American Chemical Society
15762
dx.doi.org/10.1021/ja206251r
|J. Am. Chem. Soc. 2011, 133, 15762–15772

Journal of the American Chemical Society
ARTICLE
quality air control. As an alternative to these methods, the deve-
lopment of chromogenic sensing systems for CO detection at
poisonous concentrations is of great importance. Colorimetric
methods are especially undemanding offering several advantages
over other analytical procedures, such as real-time monitoring
and the use of very simple and inexpensive instrumentation.
Additionally, certain colorimetric changes, even at low concentra-
tion of analytes, can be observed to the “naked eye”, thus making
chromogenic protocols unbeatable for certain applications.
Scheme 1. General Structure for Binuclear Rhodium Com-
pounds Used in This Work
However, colorimetric probes for the detection of this deadly
chemical are still rare. An example involves the use of oxoacetato-
6
bridged triruthenium cluster complexes, which show color
changes in acetonitrile. An electron transfer process with a zinc
tetraphenylporphyrin changes the oxidation state of the ruthe-
nium metal center thus allowing its coordination to carbon
monoxide. Polypyrrole functionalized with iron porphyrin has
also been reported to act as a suitable material for carbon
monoxide detection at low concentrations in water/methanol
7
solutions. In a recent example, a stereospecific binding of CO at
Table 1. Binuclear Rhodium Compounds Used in This
a pentacoordinated diisopropylphosphino diaminopyridine iron
complex when exposed to high concentrations of carbon mon-
oxide (1 atm of gaseous CO) has been reported; the correspond-
ing hexacoordinated derivative is formed, as indicated by the
Work with General Formula [Rh {(XC H )P(XC H ) } -
2
6
3
6
4 2 n
(
O CR) ] L
2 4ꢀn 2
3
compound
n
X
R
L
8
color change from yellow to red. The optical detection of parts-
1
2
2
(CH
3
CO
2
2
H)
H)
2
2
1
2
2
2
2
2
H
H
H
H
H
CH
CH
CH
3
3
3
CH
3
3
CO
2
H
H
3
3
3
per-million levels of carbon monoxide in air by UVꢀvis spec-
troscopy in the transmission mode using a monolayer of bime-
tallic rhodium complexes has also been reported, although with
(CH
(H
3
CO
CH
CO
2
2
O)
2
2
H O
9
3 (CF CO H)
CF3
3
CF CO H
3
2
2
2
relatively poor changes in color. In all reported examples, until
3
3
3
4
5
((CH
3
)
3
CCO
2
H)
2
C(CH
)
3 3
(CH
3
)
3 2
CCO H
now, the modulations in the presence of carbon monoxide show
certain shortcomings typically involving poor color changes. Ad-
ditionally, in most cases colorimetric CO detection was performed
in solution but not in air. Furthermore, relatively large detection
limits hamper their use as reliable sensing systems.
(CH
3
CO H)
2
2
m-CH
3
CH
3
3 2
CH CO H
out that even relatively simple systems such as tetra-μ-carbox-
ylate-dirhodium(II) derivatives show an important π-back-bond-
ing capability of the rhodium(II) atoms. This fact has been
attributed to an extensive mixing of the orbitals with π symmetry
on the two metal centers. This effect causes the rhodium(II)
metal centers to be very effective in π-back-bonding to the axial
ligands which is a very interesting feature for the design of CO
Following our interest in the design of novel chromogenic
1
0
systems, we considered the application of binuclear rhodium
compounds as potential colorimetric probes for the sensing of
CO by exploiting the well-known ability of these complexes to
11
coordinate different bases in their labile axial sites. In a recent
12
17
preliminary communication, we have reported the develop-
ment of a colorimetric probe based on the binuclear rhodium
compound 2 (CH CO H) of formula [Rh {(C H )P(C H ) } -
binding complexes. In addition, when biscyclometalated com-
pounds are compared with dirhodium tetracarboxylate deriva-
tives, a higher ability of the former for π-back-donation to the
axial ligand has been reported. Moreover, it has been observed, in
much of the chemistry associated with dirhodium(II) cyclome-
talated complexes, that the presence of different ligands coordi-
nated to the axial positions of these derivatives produce important
3
3
2
2
2
6
4
6 4 2 2
(
O CCH ) ] (CH CO H) . On the basis of the excellent
2 3 2 3 2 2
3
sensing results obtained for this rhodium complex in terms of
selectivity and sensitivity, we report herein an extended study
using a family of binuclear rhodium(II) complexes of general
18,19
formula [Rh {(XC H )P(XC H ) } (O CR) ] L contain-
changes in the UVꢀvis spectra in solution.
2
6
3
6
4 2 n
2
4ꢀn
3
2
ing one or two (in a head-to-tail arrangement) metalated phos-
phine ligands and different axial ligands. For these compounds,
the color modulations observed in the presence of carbon mon-
oxide, and induced via axial labile coordination of CO groups, have
been studied.
These enhanced π-back-donation properties of dirhodium(II)
cyclometalated complexes and the color changes observed upon
axial coordination with different ligands make these complexes
suitable as potential chromogenic systems for CO detection.
With these design concepts in mind, we have recently reported a
colorimetric probe based on a binuclear cyclometalated rhodium
compound. On the basis of the possible modulation of the
sensing features via changes in the donorꢀacceptor proper-
ties of the π-bringing ligands, we have prepared a family of
’
RESULTS AND DISCUSSION
Design of the Probe Complexes. The design of the chromo-
20
genic probes involves the use of binuclear rhodium(II) cyclo-
metalated complexes and the well-documented labile coordina-
tion ability of these compounds in their axial sites. These have
binuclear rhodium complexes (see Scheme 1) of general formula
[Rh {(XC )P(XC (O containing one
H
H
)
}
CR)4ꢀn] L
2
6
3
6
4
2
n
2
3
2
2
1
(n = 1) or two (n = 2) metalated phosphine ligands in a
head-to-tail arrangement. Two different phosphines (X = H or
1
3,14
been demonstrated to act both in their own reactivity
and as
1
5
16
catalytic centers in a number of transformations. Addition-
m-CH
C(CH
ligands (L = CH
3
)
3 3
) and three different carboxylic acids (R = CH
) as equatorial ligands, as well as four different axial
CO H, CF CO H, (CH CCO H, or H O),
3 3
, CF , or
ally, some coordination studies on adduct formation between
11
dirhodium compounds and different Lewis bases have pointed
3
2
3
2
)
3
3
2
2
1
5763
dx.doi.org/10.1021/ja206251r |J. Am. Chem. Soc. 2011, 133, 15762–15772

Journal of the American Chemical Society
ARTICLE
Scheme 2. Complexes with General Structure
[
#
Rh {(XC H )P(XC H ) } (O CR) ] L , Simplified as
2
6
3
6
4 2 n
2
4ꢀn
3
2
L , and the Corresponding Products, Simplified as
3
3
2
#
(L,CO) and # (CO) , Obtained upon Coordination of
3
2
Carbon Monoxide at Labile Axial Positions
Figure 1. (a) Simplified scheme of the structures of compounds 2 L to
3
2
5
2
L . (b) ORTEP plot of the crystal structure of 3 (CHCl3,CO);
have been used. Table 1 shows the different compounds
studied.
3
3
thermal ellipsoids are set at the 50% probability level. Selected bond
lengths [Å] and angles [°]: Rh1ꢀRh2, 2.5333(9); Rh1ꢀP1,
Reactivity with Carbon Monoxide. Once the family of birho-
dium derivatives was prepared, UVꢀvis spectrophotometric studies
for the CO coordination of all compounds, i.e., 1 (CH CO H) ,
2
2
.2574(23); Rh1ꢀC20, 1.9698(122); Rh2ꢀP2, 2.2132 (21); Rh2ꢀCl3,
.9023(31); C20ꢀO5, 1.1042(152); Rh1ꢀC20ꢀO5, 176.63(1).
3
3
2
2
2
(CH CO H) , 2 (H O) , 3 (CF CO H) , 4 ((CH ) -
3
3 2 2 2 2 3 2 2 3 3
3
3
3
CCO H) , and 5 (CH CO H) , were carried out in chloro-
hindrance caused by the more bulky groups. In this respect, the
observed RhꢀRhꢀO bond angles in the monometalated
2
2
3
3
2
2
form solution. All complexes, # L , underwent a fast and distinct
3
2
axial
color change, from purple to yellow, when CO-containing air
samples were bubbled through their chloroform solutions.
Furthermore, depending on the concentration of CO in the air
samples, two different products were obtained, with two distinct
electronic spectra. For low CO concentrations, the purple
solutions turned pink-orange, whereas at high CO concentration
solutions became yellow. Interestingly, when CO-free air was
bubbled on these final yellow solutions, the color reverted to
purple, thus indicating that the process is reversible. All these
observations agree with the formation of the corresponding
mono-, # (L,CO), and bis-CO-substituted, # (CO) , species
complex 1, 169.0(1)° and 174.4(1)°, deviate only slightly from
linearity, while for the doubly metalated complexes deviations are
more significant (within the 163.6(1)ꢀ166.9(1)° range).
To understand the ability of carbon monoxide to bind these
dirhodium complexes in the axial positions, the comparison
between the previously reported structures and those of the
corresponding CO-substituted complexes should be conducted.
Suitable crystals for single X-ray diffraction were obtained by
slow evaporation of CO-bubbled chloroform solutions of the
desired compounds layered with hexane. The crystal structures
of two different families of compounds have been obtained:
the first one (reported in a recent communication) is a mixture
3
3
2
(
Scheme 2) of the dirhodium(II) starting materials.
Characterization of the Axial Carbon Monoxide Complexes.
The crystal structures of dirhodium complexes containing cyclo-
1
2
of 2 (CH CO H, CO) and 2 (CO) , while the second one,
reported in this paper, corresponds to a 3 (CHCl ,CO)
stoichiometry.
3
3
2
3
2
3
3
20,22,23
24
metalated phosphines have already been extensively reported.
All the molecular structures of dirhodium(II) complexes 2ꢀ5
are very similar and can be rationalized with the single scheme
in Figure 1a. All of them contain two bonded rhodium atoms
that are also bridged by two carboxylate groups and two phos-
phine ligands. In the latter, metalation has occurred at one of the
phenyl rings of each, producing a head-to-tail final conforma-
tion. Compound 1 (CH CO H) is a little bit different as it
contains three carboxylates and one phosphine ligand bridging
the dirhodium(II) lantern core. Additionally, the coordination
spheres of the metal atoms are completed by means of two
other ligands, mainly acetic acid and trifluoroacetic acid, that
occupy the labile axial positions. In this way, a slightly distorted
octahedral coordination exists around each of the metal atoms
in the dirhodium(II) group. The values of the RhꢀRh bond
distances in these bismetalated complexes present a very small
variation (in the 2.499(2)ꢀ2.513(2) Å range), whereas a slightly
shorter distance, 2.430(2) Å, is found for the monometalated
The molecular structure of compound 3 (CHCl CO)
3
3,
(Figure 1b) consists of a rhodium binuclear core with two
metalated phosphine ligands in a head-to-tail conformation and
two carboxylate (trifluoroacetate) ligands in the equatorial posi-
tions, as expected by comparison with the 2 to 5 family of
complexes. In this case, however, the coordination on the axial
labile positions of the rhodium(II) centers is completed with
only one carbon monoxide molecule ligand at a bonding
distance. The second axial position is occupied by a solvent
molecule (chloroform) at a very long distance (2.902(3) Å).
Relevant crystallographic data are collected in Tables S1 and S2
(Supporting Information).
3
3
2
2
The structure reported here, together with the one recently
12
reported, allows the comparison of the series of analogous
[Rh {(C H )P(C H ) } (O CR) ] compounds differing only
2
6
4
6
5 2 2
2
2
in the bridging carboxylate ligands (R = CH or R = CF ).
3
3
Tables 2 and 3 collect the main metalꢀligand distances and
analog. The d
and d
bond distances represent a wider
angles that allow a proper comparison between the two sets of
RhꢀP
RhꢀC
range (2.193(6)ꢀ2.220(3) Å and 1.960(2)ꢀ2.020(2) Å, respec-
data. The difference in the d
groups of the carboxylate ligands change from CH to CF is
bond lengths when the R
RhꢀRh
tively), while the variation in the RhꢀO distances (2.301(4)ꢀ
axial
3 3
2
.412(5) Å) represents the largest margin. It is interesting to note
rather small, 0.009 Å, as has already been found for all 2ꢀ5
known complexes (0.005 Å between 2 (CH CO H) and 3
that the dRh-L and d
significant trend upon variation of either the carboxylate R or the
axial L groups, apart from that derived from the inherent steric
bond distances do not show any
RhꢀRh
3
3
2
2
3
(CF CO H) ). It is clear that the important change in the elec-
3 2 2
tronic character of these carboxylates has only a small influence
1
5764
dx.doi.org/10.1021/ja206251r |J. Am. Chem. Soc. 2011, 133, 15762–15772

Journal of the American Chemical Society
ARTICLE
a
Table 2. Relevant Bond Distances (Å) for Biscyclometalated Binuclear Rhodium Compounds Used in This Work
2
0
12
26
b
2
3 2
(CH CO H)
2
2 (CH
3
CO
2
H,CO)
3 (CF
3 2
CO H)
2
3
3 (CHCl ,CO)
3
3
3
3
d
d
d
d
d
d
d
d
d
d
d
Rh1ꢀRh2
Rh1ꢀOax
Rh1ꢀCOax
Rh2ꢀOax
Rh2ꢀCOax
Rh1ꢀOav
Rh2ꢀOav
Rh1ꢀP
2.508(1)
2.342(5)
-
2.5425(5)
-
2.513(2)
2.340(2)
-
2.5333(9)
-
1.9810(5)
2.3580(18)
-
1.9698(122)
-
2.342(5)
-
2.360(14)
-
-
2.190(4)
2.136(4)
2.210(2)
2.210(1)
1.997(4)
1.996(6)
2.1520(3)
2.1525(3)
2.2384(12)
2.2117(11)
2.0400(4)
2.210(2)
2.210(2)
2.218(7)
2.205(7)
2.020(2)
1.990(2)
2.1832(54)
2.16765(7)
2.2574(23)
2.2132(21)
2.0326(79)
1.9942(82)
Rh2ꢀP
Rh1ꢀC
Rh2ꢀC
2.0220(5)
a
b
Atom labeling corresponds to the nomenclature shown in Figure 1b. Present work.
Table 3. RhꢀRhꢀL Angles (°) for Biscyclometalated Binuclear Rhodium Compounds Used in This Work
2
3 2
(CH CO H)
2
2 (CH
3
CO
2
H,CO)
3 (CF
3 2
CO H)
2
3
3 (CHCl ,CO)
3
3
3
3
Rh
2
ꢀRh
ꢀRh
1
ꢀL
163.6(1)
163.6(1)
171.03(14)
167.40(4)
166.5(4)
166.3(4)
170.71(33)
-
LꢀRh
2
1
on the intermetallic distance. A more important effect is found
when the carboxylic acid in the axial positions is substituted by a
carbon monoxide. In this case, the differences found are 1 order
of magnitude larger (0.0345 Å between 2 (CH CO H) and
3
3
2
2
2
3
(CH CO H, CO); 0.0203 Å between 3 (CF CO H) and
3
3
3 2 3 2 2
3
(CHCl ,CO)). This effect fits very nicely with the π-acceptor
3
character of the carbon monoxide molecule that withdraws
electron density from the metalꢀmetal bond. A combination
of the two above-mentioned facts may be the actual reason for
the high asymmetry of the 3 (CHCl ,CO) complex. The
3
3
complete coordination of the dirhodium(II) core in this complex
can be achieved with only a simple solvent molecule, which
supports the idea that the metal to axial ligand bond could not
11
only be explained considering a pure sigma interaction.
Having a look at the Rh-equatorial ligand distances, a similar
trend is evident. A longer d distance is, nevertheless,
Figure 2. Diffuse reflectance UVꢀvis spectra of a mixture of silica
containing 5 (CH CO H) and the changes observed in the presence
of air containing 8 and 2000 ppm of CO.
3
3
2
2
RhꢀP
observed for metal atoms having carbon monoxide in their
coordination sphere, which is balanced by a shortening in the
corresponding dRh-Oeq bond lengths. Very similar averaged
equatorial distances when changing the carboxylic acid by carbon
monoxide are found in all cases.
weight ratio 2ꢀ10 times) resulted in purple solids. The solids
were separated after five minutes of stirring at room temperature,
and the solvent was removed on a rotary evaporator. The
resulting solid was dried at 343 K for at least 48 h before use.
The colored silica supports containing the rhodium probes
underwent important color changes in seconds when exposed to
air containing different concentrations of carbon monoxide. As
observed in chloroform solutions, a more detailed study of the
titration process, using increasing concentrations of carbon
monoxide, clearly revealed that two consecutive substitution re-
actions occur. As an example, Figure 2 shows the diffuse reflec-
With respect to the bond angles, a general deviation of linearity
of the RhꢀRhꢀL unit, both for R = CH and CF (Table 3), is
3
3
observed, probably due to the presence of the phosphine ligands
in complexes 2 (CH CO H,CO) and 3 (CHCl ,CO). The
3
3
2
3
3
substitution of axial carboxylic acid molecules by carbon mon-
oxide allows a relaxation in the Rh ꢀRh ꢀL angles, producing
2
1
less strained units.
Carbon Monoxide Sensing Behavior in Air. Despite the
interesting spectrophotometric response of the studied dirhodium-
tance UVꢀvis spectra of silica beads containing 5 (CH CO H)
3
3
2
2
and the changes observed in the presence of air containing 8 and
000 ppm of CO. Whereas 5 (CH CO H) shows a band at
(
II) complexes found in chloroform solutions, the detection
2
3
3
2
2
of CO is of main importance in the gas phase. Therefore, we
took a step forward toward the potential use of this kind of
sensing derivatives as probes for the detection of carbon
monoxide in air. Adsorption of compounds 1ꢀ5 on silica beads
550 nm, a large excess of CO (2000 ppm) resulted in the ap-
pearance of a new band at 412 nm, whereas at lower concentra-
tions of CO (8 ppm) a band at 495 nm was observed. These
changes are consistent with an axial coordination by CO and the
formation of the derivatives 5 (CO) and 5 (CH CO H, CO)
(via solution of each binuclear rhodium compound in a mini-
3
2
3
3
2
mum amount of chloroform followed by the addition of silica at a
(Scheme 2).
1
5765
dx.doi.org/10.1021/ja206251r |J. Am. Chem. Soc. 2011, 133, 15762–15772

Journal of the American Chemical Society
ARTICLE
Table 4. UVꢀvis Spectral Data for the Different Binuclear
Rhodium(II) Compounds Studied in Chloroform Solution,
As Well As Their Diffuse Reflectance Spectra Supported on
a
Silica Beads, At Room Temperature
CHCl
3
solution
solid
max (nm) νCO (cm
550
ꢀ1
ꢀ1
ꢀ1
compound
λ
max (nm), (ε (M cm ))
λ
)
1
1
2
2
3
3
4
4
5
5
(CH
3
CO
2
H)
2
2
556, (157)
394, (421)
567, (552)
398, (7040)
571, (711)
400, (6570)
547, (560)
397, (3689)
564, (265)
397, (1247)
3
3
3
3
3
3
3
3
3
3
(CO)
(CH
(CO)
(CF
(CO)
(CH
(CO)
(CH
(CO)
2
400
545
398
537
402
539
400
550
412
2040
2028
2048
2033
2033
3
CO
2
H)
2
3
2 2
CO H)
2
Figure 3. Loss of the intensity of the band centered at 557 nm vs
3
CO
2
H)
2
2
3 2 2
concentration of CO in air for 5 (CH CO H) adsorbed on silica.
3
2
3
CO
2
H)
2
a
The bands assigned to the CO stretch in the IR spectrum are also
included.
Table 5. Detection Limits (ppm) for Compounds # L in the
3
2
a
Presence of CO
detection limits (ppm) of CO
UVꢀvis spectrophotometer naked eye
compound
1
2
2
3
4
5
(CH
3
3
CO
2
2
H)
2
2
0.7
0.6
52
42
35
38
50
3
3
3
3
3
3
(CH
CO
H)
Figure 4. Sensing array of the studied complexes for CO detection in
air. On the right, the increment of the RGB values is shown.
(H
2
O)
2
1.6
(CF
3
CO
2
H)
CCO
CO H)
2
0.8
((CH
3
)
3
2
H)
2
1.7
silica gel with increasing quantities of carbon monoxide in air
allowed determining the detection limits for CO. From a chemo-
sensing point of view, a low detection limit for carbon monoxide of
less than 1.7 ppm was found for all studied complexes using a
conventional UVꢀvis spectrophotometer (see Table 5). Fig-
ure 3 shows the changes in absorbance of the band centered at
557 nm of 5 (CH CO H) adsorbed on silica beads in the
(CH
3
2
2
<0.1
0.2
a
Limits calculated from UVꢀvisible spectral data (first column) and the
to the naked eye (second column).
A very similar change in color and hypsochromic shift of the band
was observed for all dirhodium(II) complexes studied. Table 4
shows the spectroscopic characteristics of the complexes in chloro-
form solutions, as well as adsorbed on silica, together with the data in
the presence of an excess of carbon monoxide. Compounds 1ꢀ5
display an absorption band in the 550ꢀ570 nm range in chloroform
that is not basically modified when adsorbed on silica beads. Upon
addition of an excess of carbon monoxide, new bands centered in the
3
3
2
2
presence of different concentrations of CO in air; the response of
complex 5 (CH CO H) to CO shows a lineal trend between 1
3
3
2
2
and 100 ppm at this wavelength. A similar lineal behavior was also
found for the other rhodium derivatives studied.
Table 5 indicates the estimated detection limit to the “naked eye”
for compounds 1ꢀ5 in the presence of CO, i.e., the minimum
amount of CO necessary to observe a clear color change in the
complexes. All the compounds display a chromogenic modulation
to the naked eye at concentrations at which CO is toxic (i.e., 50ꢀ
60 ppm), and compound 5 (CH CO H) even shows a very
3
94ꢀ400 nm region were found for all complexes both in chloro-
form and on silica. The table also shows that the molar absorptivities
of the corresponding carbon monoxide complexes are in general 1
order of magnitude larger than those of the starting derivatives.
Table 4 also collects the carbon monoxide stretching frequency
for the # (CO) derivatives adsorbed on silica, which range from
3
3
2
2
remarkable color modulation at CO concentrations as low as 0.2 ppm.
This range of visual CO detection values found for all com-
pounds made it possible to design systems for semiquantita-
tive sensing of CO in air at different concentration ranges. For
instance, a sensing array containing compounds 4 ((CH ) -
3
2
ꢀ
1
25
2
028 to 2048 cm . These stretching ν frequencies are, in all
cases, considerably lower than the CO stretching frequency of the
CO
ꢀ
1
free carbon monoxide (2143 cm ), which indicates the presence of
highly acidic axial metal sites with relatively high levels of M f CO
π-back-bonding. This effect is naturally stronger when donor groups
are also present in the bridging carboxylate groups.
Perhaps one of the most remarkable behaviors in relation to the
response of the tested dirhodium(II) complexes is that the chromo-
genic response is observed at relatively low concentrations of carbon
monoxide. Titration studies of the derivatives 1ꢀ5 adsorbed on
3
3 3
CCO H) , 2 (H O) , and 5 (CH CO H) , displaying clear
2
2
3
2
2
3
3
2
2
visual detection limits from CO concentration of 50, 35, and
0.2 ppm, respectively, will allow us to sense the presence of CO in the
ranges, 0.2 < CCO <35ppm, 35<CCO <50ppm, andCCO > 50 ppm.
Figure 4 shows such an array and the color changes for
4 ((CH ) CCO H) , 2 (H O) , and 5 (CH CO H) deriva-
3
3 3
2
2
3
2
2
3
3
2
2
tives for concentrations of CO of 0, 2, 35, and 50 ppm in air.
1
5766
dx.doi.org/10.1021/ja206251r |J. Am. Chem. Soc. 2011, 133, 15762–15772

Journal of the American Chemical Society
ARTICLE
a
Table 6. Summary of the Observed Behavior for VOCs and Gases Tested with All Compounds
VOCs (ppm)
solid
acetone
ACN
chloroform
ethanol
formaldehyde
hexane
toluene
1
2
2
3
4
5
(CH
3
3
CO
2
2
H)
2
2
-
-
-
-
-
-
4600
4600
4600
4600
4600
4600
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
3
3
3
3
3
3
(CH
CO
H)
(H
2
O)
2
(CF
3
CO
2
H)
CCO
CO H)
2
((CH
3
)
3
2 2
H)
(CH
3
2
2
gases (ppm)
solid
water
CO
2
N
2
O
2
Ar
SO
2
NO
NO
2
1
2
2
3
4
5
(CH
3
3
CO
2
2
H)
2
2
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
8140
8140
814
2650
2650
500
3
3
3
3
3
3
(CH
CO
H)
(H
2
O)
2
(CF
3
CO
2
H)
CCO
CO H)
2
4070
4070
8140
2650
2650
5000
((CH
3
)
3
2 2
H)
(CH
3
2
2
a
Both responses are shown: not induced color change (-) and concentrations (in ppm) necessary to induce color changes.
Selectivity. In all cases, the binuclear rhodium complexes
studied showed a remarkable selective response to carbon
monoxide in air. For instance, no reaction was observed in the
presence of CO , N , O , or Ar at very high concentrations (up to
2
2
2
5
0 000 ppm). Similarly, no color changes were observed in the
presence of volatile organic compounds such as acetone, chloro-
form, ethanol, formaldehyde, hexane, or toluene (up to 30 000
ppm in air). Some color changes to yellow were observed in the
presence of acetonitrile vapor, although only at concentrations of
4
600 ppm. Studies with other coordinating oxygen-containing
gaseous species, such as SO , NO, and NO , were also carried
2
2
out. No noticeable color changes were observed on any of the
compounds adsorbed on silica beads in the presence of SO (up
2
to 38 000 ppm in air), but NO and NO produced color changes
2
rather similar to those observed in the presence of CO. This
reactivity, nevertheless, was only observed at very high concen-
trations of nitrogen mono- or dioxide. Additionally, whereas the
Figure 5. Diffuse reflectance UVꢀvis spectral response for the band
centered at λ = 400 nm for silica beads containing 2 (CO) exposed to
an air atmosphere. The cycling of the sensor is checked by letting CO-
containing air in contact with the sample after full recovery (ca. 15 h) of
the starting complex probe.
reaction with NO was reversible, the reaction with NO was
3
2
2
irreversible. Thus, when the yellow compounds obtained by
reaction of the binuclear rhodium derivatives supported on silica
with high concentrations of NO were left in NO-free air the solid
reverted to purple again, which was not observed for NO . All the
2
reactivity and relevant concentration conditions with VOCs and
gases tested are summarized in Table 6 where concentrations in
which color modulations of the complexes 1ꢀ5 were observed
are shown.
scale, the complete reverse transformation in air from # (CO)2
3
to # L takes approximately 15 h at room temperature. Notice-
3
2
ably, the desorption of carbon monoxide from # (CO) occurs
3
2
also in two consecutive steps producing initially # (L,CO) and
the further appearance of the purple dirhodium(II) starting
3
Reversibility of Carbon Monoxide Coordination. Ideal
sensing systems not only need to display selectively and sensing
features at low concentration but also must be able to show
reversible binding with the target analyte. In line with the
successful application of the above-mentioned probes, binding
of CO in the dinuclear rhodium complexes was found fully
reversible. When the yellow solid corresponding to any # (CO)
materials, # L . Figure 5 shows a representative example of the
3
2
changes observed in the intensity of the band at 400 nm from the
diffuse reflectance UVꢀvis spectrum of the probe from complex
2 (H O) upon cyclic contact with CO-containing and CO-free
3
2
2
air. This cyclic process was repeated at least 10 times without
significant degradation of the sensing ability of the material.
Moreover, the release of carbon monoxide from the corre-
sponding # (CO) was followed by thermogravimetry, and in all
3
2
complex embedded in silica was left in an open air atmosphere,
the diffuse reflectance spectra evolve back to the initial corre-
sponding purple complex probe. Although the uptake of CO in
the adsorbed compounds occurs in the seconds/minutes time
3
2
cases the weight changes correspond to 2 equiv of CO per
molecule. Furthermore, no loss of the L ligands from the initial
1
5767
dx.doi.org/10.1021/ja206251r |J. Am. Chem. Soc. 2011, 133, 15762–15772

Journal of the American Chemical Society
L2 probe complexes was observed during the processes, thus
indicating that these axial ligands remain adsorbed on the silica
beads during all the cycles.
ARTICLE
#
Scheme 3. Carbon Monoxide Liberation in 2 (CO) in Two
3
3
2
Consecutive Kinetic Steps, Producing the Regeneration
of 2 L
3
2
In this respect and to determine the role played by the axial
ligand in the reversible CO binding process, the formation and
CO release from 2 (CO) using two different starting probe
3
2
complexes (2 (CH CO H) and 2 (H O) ) was monitored
3
3
2
2
3
2
2
thermogravimetrically. Although in both cases two CO mol-
ecules are adsorbed/released reversibly, remarkable differences
in the release kinetics were found. While for the 2 (CO)
derivative obtained from silica bead-supported 2 (CH CO H)
the loss of CO is complete in some hours, the complete CO release
3
2
2
3
3
2
from the 2 (CO) derivative obtained from the 2 (H O) probe
3
2
3
2
2
requires ca. two weeks at room temperature.
Kinetico-Mechanistic Studies of the Substitution of CO by
L on the Immobilized Compounds. The uptake of carbon
monoxide by the immobilized complexes occurs at room temper-
ature in a time scale of minutes, much longer than the practically
instant process observed for the systems in solution. Even though
the diffusion of gaseous CO into the porous silica beads should
play a crucial role in this process, as already established for other
26,27
immobilized systems,
no further studies have been carried
out for the CO uptake process. The fully reversible character of
the process for all adsorbed compounds indicated in Table 1 was
confirmed by thermogravimetric and UVꢀvis analyses (see
Scheme 2). As indicated above, the results are especially relevant
with respect to the presence of the CO-displaced axial L
ligands of the starting material inside the silica beads, which
creates the need for a further kinetico-mechanistic investiga-
tion of the process.
Figure 6. Eyring plot for k
sponding to CO loss for immobilized compounds 2 (CH
1
(full) and k
2
(empty) rate constants corre-
CO H)
2
3
2
3
The release of CO from the # (CO) derivatives obtained
3
2
2 2
(squares) and 2 (H O) (circles).
3
from silica immobilized samples of # L corresponds to the neat
3
2
weight loss of two CO molecules per dirhodium unit and was
studied in detail from a kinetico-mechanistic perspective. The
time monitoring at a fixed temperature of this weight loss pro-
vided an excellent handle for the measure of the rates of the
process, when fitted to a double exponential decay, as corre-
The temperature dependence of both k and k rate constants
has also been determined by the use of the Eyring equation.
Figure 6 shows the plots for two of the starting immobilized
1
2
3
0
compounds studied, 2 (CH CO H) and 2 (H O) ; the re-
3
3
2
2
3
2
2
sponds to the intermediate formation of the # (CO,L) species.
levant kinetic and thermal activation parameters for these
processes on all studied compounds are collected in Table 7.
The very different steric demands of the possible hydrogen
bonding interactions for the immobilized complexes, as well as
the dramatic change of the σ-donor character of the oxygen
donors, are bound to produce important differences between the
CO by CH CO H, CF CO H, (CH ) CCO H, or H O sub-
3
Effectively, the weight loss associated to each one of fitted set of
two consecutive exponentials is half that of the total decrease, in
excellent agreement with the kinetic detection of the # (CO,L)
3
intermediate complexes (see above and Figure S1 (Supporting
Information) for the 5 (CO) complex). The observed first-
3
2
order character of both consecutive processes can be associated
either with a limiting CO dissociative process or with a unim-
olecular reaction involving the L ligands already attached in an
3
2
3
2
3 3
2
2
stitution. As seen in the data collected in Table 7, the thermal
activation parameters associated with the faster process (k1)
indicate that for all immobilized compounds the values are
surprisingly the same within experimental error, with rather
low enthalpies and very negative entropies of activation.
These indicate a mechanism with an important degree of
ordering during the substitution process with very similar ener-
getics. However, the total recovery of the 2 (H O) system is
28
outer-sphere fashion to the dirhodium(II) probe complex.
Alternatively, instead of the back coordination of the CO-
displaced L ligands, attachment of the available OH groups in
the silica beads can also explain such behavior, the L ligands being
mere spectators of the process. However, the latter possibility
was rejected, as complex 2 (which axial ligands have been
3
2
2
29
extracted via thermal treatment) adsorbed on silica beads was
found to be brownish in color, in deep contrast to the purple
color of the 2 L complexes. In the same line, the recycling of the
much slower than for the rest of the probes, in good agreement
with the differences observed, due to the much lower value of k2
for the recovery of compound 2 (H O) .
3
2
3
2
2
silica beads on CO depletion also reverts to the original spectra
The only way to come to terms with the surprising homo-
for all the adsorbed # L complexes, a clear indication that the re-
geneity within the k set of data between such different com-
3
2
1
entry of the original axially bound ligands, L, on the dirhodium
compounds is occurring. Consequently all the weight-loss/time
traces (Figure S1, Supporting Information) can be fully asso-
ciated with Scheme 3, where L represents the axial ligands
existing in the starting materials.
plexes is by considering that the CO-displaced axial ligand (L)
remains still attached via hydrogen bonding interactions to the
neighboring bridging carboxylato units of the dirhodium com-
plex. Such interactions have already been described in some
31
32
Pd(II) XRD data as well as in DFT and mechanistic studies
1
5768
dx.doi.org/10.1021/ja206251r |J. Am. Chem. Soc. 2011, 133, 15762–15772

Journal of the American Chemical Society
Table 7. Kinetic and Thermal Activation Parameters for the Processes Indicated in Scheme 3 for Immobilized Complexes # L2
ARTICLE
a
3
3
342
ꢀ1
q
1
ꢀ1
q
1
ꢀ1
ꢀ1
4
342
ꢀ1
q
ꢀ1
q
ꢀ1
ꢀ1
compound
10 ꢁ
1
k /s
ΔH
/kJ mol
ΔS
/J K mol
10 ꢁ
k
2
/s
ΔH
2
/kJ mol
2
ΔS /J K mol
1
2
2
3
4
5
(CH
3
3
CO
2
2
H)
2
2
3.6
5.3
3.2
28 ( 3
35 ( 3
30 ( 3
33 ( 3
30 ( 3
30 ( 3
ꢀ210 ( 8
3.3
6.8
60 ( 5
ꢀ145 ( 10
ꢀ20 ( 10
180 ( 10
3
3
3
3
3
3
(CH
CO
H)
ꢀ190 ( 11
ꢀ210 ( 10
ꢀ185 ( 11
ꢀ210 ( 10
ꢀ200 ( 11
100 ( 5
180 ( 5
21 ( 5
(H
2
O)
2
0.078
16
2.2
6.0
(CF
3
CO
2
H)
CCO
CO H)
and k
2
13
ꢀ240 ( 10
ꢀ70 ( 40
ꢀ95 ( 30
((CH
3
)
3
2
H)
2
2.4
3.5
85 ( 15
75 ( 10
(CH
3
2
2
a
k
1
2
were also determined at 334, 352, 361, and 371 K for the evaluation of the thermal activation parameters.
Scheme 4. Mechanism Proposed for CO Liberation in
Compounds with Carboxylic Acid As Axial Ligands
Scheme 5. Possible Outer-Sphere “Solvent-Assisted” Inter-
q
q
Figure 7. ΔH /ΔS compensation plot for all the systems studied. The
action for the 2 (H O) System
3
2
2
values for k cannot be distinguished given its similarity.
1
much more positive for the 2 (H O) system, thus implying a
3
2
2
much more dissociated transition state than for the systems
derived from its analogous 2 (CH CO H) probe. Given the
3
3
2
2
fact that the lack of a statistical factor in the value of the rate
constants is a clear indication of the electronic transmission
via the RhꢀRh bond, already established in equilibrium
1
1
studies, the better the σ-donor character of the already
coordinated first L ligand, the better π-back-donation into
the remaining axially bound CO. As a consequence, the
process becomes slower, with higher enthalpic demands that
promote an important degree of decoordination of CO before
the new ligand L can be coordinated to the fairly electron
enriched rhodium center. This effect should be very pro-
1
3,33
on electrophilic CꢀH bond activation
processes assisted by
acid. By this attachment, the important differences between the L
ligands could be producing, de facto, a rather similar electronics
in all cases. Scheme 4 represents a reasonable sequence for such
processes exemplified for the 2 (CH CO H) complex. The
nounced for L = H O and minimal (or even opposite) for
2
3
3
2
2
L = CF CO H as effectively observed. Figure 7 shows the
3
2
transition state indicated presents an effective high degree of
ordering with little enthalpic demands considering the already
q
q
ΔH /ΔS compensation plot for all the systems studied,
which corroborates the homogeneity of the mechanism oper-
ating for all the reactions.
“
initiated” Rh-carboxylate bonding.
Nevertheless, in the case of decoordinated H O arrangement,
2
2
(H O) , the involvement of the silica OH groups cannot be
3
2 2
disregarded, given the existence of a single oxygen donor in the
’
CONCLUSIONS
water molecule. An outer-sphere “solvent-assisted” interaction
ꢀ
between one of the oxygens of the μ-CH CO2 group and the
A family of binuclear rhodium compounds of general formula
3
H O ligand (OꢀC(CH )ꢀO HꢀO H-O-H) should be
[Rh {(XC H )P(XC H ) } (O CR) ] L containing one or
2
3
3 3
3 3
2 6 3 6 4 2 n 2 4ꢀn 2
3
considered as responsible for the similarity with the carboxylic
two, in a head-to-tail arrangement, metalated phosphine ligands
and different equatorial and axial ligands have been used as
chromogenic chemosensors for CO detection. UVꢀvis studies
were carried out, both in solution and immobilized on silica
beads, resulting in hypsochromic shifts. In the case of compounds
adsorbed on silica beads in the presence of CO, two different
spectra were observed due to the axial coordination of CO
groups: orange (one CO in axial position) and yellow (two
CO). The crystal structure of 3 (CHCl ,CO) was solved by
acid systems (OꢀC(R)ꢀO HꢀOꢀC(R)ꢀO) (Scheme 5).
3
3 3
Examination of the kinetic and thermal activation data for k
2
for the systems studied (Table 7) indicates a general increase of
the activation enthalpy accompanied by a decrease of the
ordering during the process (less negative or positive activa-
tion entropies) with respect to k , the exception being for the
1
3
(CF CO H) complex with very poor σ-donors in its axial
3 2 2
q q
3
positions. Furthermore, the values for ΔH and ΔS are even
3
3
1
5769
dx.doi.org/10.1021/ja206251r |J. Am. Chem. Soc. 2011, 133, 15762–15772

Journal of the American Chemical Society
ARTICLE
single X-ray diffraction techniques. In all cases the binuclear
rhodium complexes studied showed a remarkable low detection
limit for CO and a high selective response. From all VOCs and
gases tested as possible interferences, the only sensing was
compounds were obtained after removal of the solvent under reduced
pressure.
[Rh {(XC
2
6
H
3
)P(XC
6
H
4
)
2
}(O CR)
2
3
] L
2
. [Rh
2
{(C )P(C
6
H
4
6 5 2 2
H ) }(O C-
3
3
CH )
3
] (CH
3
CO H)
2
2
complex, 1 (CH
3
CO H) , was obtained follow-
2
2
3
3
ing the same synthetic procedure as used for compounds with general
observed in the presence of acetonitrile, NO, and NO ; never-
2
formula [Rh {(XC H )P(XC H ) } (O CR) ] L but using a phos-
theless, this only occurs at high concentrations (4600, 4070, and
2
6
3
6
4 2 2
2
2
3
2
phine:dirhodium(II) tetraacetate molar ratio of 0.9:1.
Silica Gel Immobilization of the Dirhodium(II) Complexes.
2
650 ppm, respectively). As a special feature, all the complexes
tested display a very clear and remarkable color change to the
naked eye at concentrations at which CO becomes toxic (i.e.,
Each binuclear rhodium compound was dissolved in a minimum volume
of CHCl
3
. An excess of silica (230ꢀ240 mesh, weight ratio 2ꢀ10-fold)
5
0ꢀ60 ppm). In particular, compound 5 (CH CO H) dis-
3
3
2
2
was added to the colored solution, and the resulting mixture was stirred
at room temperature for five minutes. After removal of the solvent on a
rotary evaporator, the solid was dried in an oven at 343 K during at least
two days prior to its use. A careful weight control of the samples was
carried out, indicating that the effective final weight corresponds to the
mass of silica plus the mass of the # L complex and that no evolution
plays clear color modulations at CO concentrations as low as 0.2
ppm, even allowing the quantification CO in air, based on simple
and visual color changes. Binding of CO in these systems was
found fully reversible, and the release of carbon monoxide can be
monitored by thermogravimetric measurements. The kinetic and
thermal activation data obtained for the depletion of CO indicate
that the outer-sphere coordination of the L ligands, even in the
presence of axially bound CO, is determinant for the easiness and
reproducibility of the reversibility of the process. The silica
3
2
from the silica beads of the axially bound ligands, L, has occurred under
the working conditions.
Instrumentation. Infrared spectra were recorded at room tem-
perature on a JASCO FT-IR 460 Plus spectrometer. UVꢀvis spectra
were recorded using a Jasco V-650 spectrophotometer equipped with a
diffuse reflectance Sphere (model ISV-722 Sphere) for measurements
on solids. In the latter case, measurements were conducted at room
temperature over a wavelength range of 350ꢀ800 nm with a wavelength
step of 1 nm. The reflectance data were transformed using Kubelkaꢀ
support behaves, especially for L = H O, as a noninnocent
2
solvent assisting important interactions that avoid the total removal
of the CO-substituted L groups.
’
EXPERIMENTAL SECTION
37
Munk function [F(R)], where R is the fraction of incident light
reflected by the sample.
General Considerations. Commercially available reagents,
[
Rh
2
(O
2
CCH
3
)
4
], as well as the different phosphines (P(C
6
H
5
)
3
,
2
ð1 ꢀ RÞ
P(m-CH C H ) ) and carboxylic acids (CF CO H, CH CO H,
(
FðRÞ ¼
3
6
4 3
3
2
3
2
2
R
CH
3
)
3
2
CO H), were used as purchased. All solvents were of analy-
tical grade. Column chromatography was carried out on silica gel 60
For kinetic studies, CO loss was monitored via thermogravimetric
analyses carried out on a TGA/SDTA 851e Mettler Toledo balance,
set at a fixed temperature, using nitrogen as purge gas. The fitting of the
weight changes versus time to a double exponential decay with identical
amplitudes was carried out by the standard software packages available.
Carbon monoxide concentrations were measured by an ambient carbon
monoxide analyzer (Testo 315ꢀ2 model 0632 0317), properly validated
with a calibration certificate issued by Spanish Certification Entity (ENAC).
X-ray data collection was performed on a Bruker Kappa CCD
diffractometer using graphite monochromated Mo Kα radiation (λ =
0.71073 Å) at 293 K. Data reduction and cell refinements were
(
230ꢀ240 mesh). Solvent mixtures are volume/volume mixtures,
unless specified otherwise. All reactions were carried out in oven-
dried glassware under an argon atmosphere, although the isolated
solids are air-stable. The solvents were degassed before use. All the
gases used in this work were generated in situ: carbon monoxide by
reaction of sulfuric acid with formic acid; carbon dioxide by adding
chloride acid to sodium carbonate; nitrogen monoxide and nitrogen
dioxide by oxidation of copper with nitric acid; sulfur dioxide by
copper oxidation with sulfuric acid.
Compounds. [Rh
P(XC H ) } (O CCH ) ] (CH CO H) complexes, 2 (CH CO H)
2
2 6 3 6 4 2 2 2 2 2 2 6 3
{(XC H )P(XC H ) } (O CR) ] L . [Rh {(XC H )-
3
23
performed with HKL DENZO and SCALEPACK programs. Crystal
structures were solved by direct methods with the aid of successive
difference Fourier maps and were refined using the SHELXTL 6.1
6
4 2 2
2
3 2
3
3
2
2
3
3
2
and 5 (CH
3
CO
phosphine with [Rh
under argon atmosphere as described in the literature.
2
H)
2
, were obtained by refluxing the corresponding
3
2
(O CCH ] in toluene:acetic acid media (3:1)
2
3 4
)
1
5,34,35
38
For these
software package. All non-hydrogen atoms were refined anisotropi-
compounds with two cyclometalated phosphines, the molar ratio phos-
phine:dirhodium(II) tetraacetate was 2:1. These compounds containing
two ortho-metalated aryl phosphines and two carboxylates as bridging
ligands show fast and quantitative interchange of both carboxylate ligands
cally. High residual electronic density was found around the CF
in the trifluoroacetate ligands, and very anisotropic thermal ellipsoids
were obtained for fluorine atoms. Hence, a disordered model for CF
3
groups
3
groups was applied using a riding model refinement and constraining the
occupancy to one for those groups bonded to the same carbon atom in
both acetate ligands. The hydrogen atoms were assigned to ideal positions
and refined using a riding model. The details of the data collection, cell
dimensions, and structure refinement are given in the Supporting Informa-
tion. Full crystallographic data have been deposited in the Cambridge
Crystallographic Data Centre with the number CCDC 820952.
RGB values were analyzed using Adobe Photoshop Elements 7.0
software.
36
by other bridging ligands, which facilitates the interchange with other
carboxylates. Thus, other derivatives with the same ortho-metalated
dirhodium(II) skeleton can be easily obtained (see below).
Compounds of general formula [Rh {(XC H )P(XC H ) } -
2
6
3
6
4 2 2
(
O
2
CR)
2
] L2, 3 (CF
3
CO
2
H)
2
, and 4 ((CH
3
)
3
CCO
2 2
H) were ob-
3
3
3
tained by stirring the starting product, 2 (CH
3
CO
2
H) , in the presence
2
3
of an excess of the substituting carboxylic acid (trifluoroacetic and pivalic
acid, respectively) in chloroform solution. The mixture was then heated
under reduced pressure to dryness. The procedure was repeated until the
complete interchange was produced (5ꢀ6 times). The excess of the acid
was then eliminated by column chromatography (CH Cl :Hexane 2:1).
’ ASSOCIATED CONTENT
2
2
S
Supporting Information. Crystallographic data and struc-
[
Rh {(C H )P(C H ) } (O CCH ) ] (H O) , 2 (H O) , was
b
2
6
4
6
5 2 2
2
3 2
3
2
2
3
2
2
obtained by stirring 2 (CH
sodium carbonate in acetone. In this procedure, the water contained in
the acetone is the origin of the substituting axial ligand. The resulting
3
CO
2
H)
2
in the presence of an excess of
ture refinement of [Rh
(3 CO). Selected geometric data for [Rh
(O CCH ) ] (CO) (3 CO) at 293 K. Weight loss of a sample of
[(C H )P(C H )
2 6 4 6 5
2
]
2
(O
2
H
CCH
)P(C H ) ] -
6 5 2 2
3 2
) ] (CO)
3
3
[(C
2
3
6
4
2
3 2
3
3
1
5770
dx.doi.org/10.1021/ja206251r |J. Am. Chem. Soc. 2011, 133, 15762–15772

Journal of the American Chemical Society
(CO) with time at 50 °C. This material is available free of
ARTICLE
5
(15) Hirva, P.; Lahuerta, P.; P ꢀe rez-Prieto, J. Theor. Chem. 2005, 113,
3
2
charge via the Internet at http://pubs.acs.org.
63–68.
(
16) (a) Estevan, F.; Lahuerta, P.; P ꢀe rez-Prieto, J.; Pereira, I.; Stiriba,
S. E. Organometallics 1998, 17, 3442–3447. (b) Estevan, F.; Lahuerta, P.;
P ꢀe rez-Prieto, J.; Sana ꢀu , M.; Stiriba, S. E.; Ubeda, M. A. Organometallics
’
AUTHOR INFORMATION
ꢀ
Corresponding Author
rmaez@qim.upv.es; manel.martinez@qi.ub.es.
1997, 16, 880–886. (c) Barberis, M.; Lahuerta, P.; P ꢀe rez-Prieto, J.;
Sana ꢀu , M. Chem. Commun. 2001, 439–440. (d) Barberis, M.; P ꢀe rez-
Prieto, J.; Stiriba, S. E.; Lahuerta, P. Org. Lett. 2001, 3, 3317–3319. (e)
Estevan, F.; Herbst, K.; Lahuerta, P.; Barberis, M.; P ꢀe rez-Prieto, J.
Organometallics 2001, 20, 950–957.
’
ACKNOWLEDGMENT
(
17) Drago, R. S.; Tanner, S. P.; Ritchman, R. M.; Long, J. R. J. Am.
Chem. Soc. 1979, 101, 2897–2903.
18) Gonz ꢀa lez, G.; Martínez, M.; Estevan, F.; García-Herbosa, G.;
The authors wish to express their gratitude to the Spanish
(
Ministerio de Ciencia y e Innovaci ꢀo n (projects MAT2009-
4564-C04-01 and CTQ2009-14443-C02-02) and Generalitat
Lahuerta, P.; Peris, E.; Ubeda, M.; Diaz, M. R.; Garcia-Granda, S.;
Tejerina, B. New J. Chem. 1996, 20, 83–94.
1
Valenciana (project PROMETEO/2009/016) for their support.
MEM is grateful to the Spanish Ministerio de Ciencia e In-
novaci ꢀo n for an FPU grant.
(19) Esteban, J. PhD Thesis 2006, Universitat de Val ꢁe ncia.
(20) Chakravarty, A. R.; Cotton, F. A.; Tocher, D. A.; Tocher, J. H.
Organometallics 1985, 4, 8–13.
(
21) (a) Barcel, F.; Cotton, F. A.; Lahuerta, P.; Llusar, R.; Sana ꢀu , M.;
ꢀ
Schwotzer, W.; Ubeda, M. A. Organometallics 1986, 5, 808–811. (b)
Barcel, F.; Cotton, F. A.; Lahuerta, P.; Sana ꢀu , M.; Schwotzer, W.; Ubeda,
’
REFERENCES
ꢀ
(
1) (a) Wattel, F.; Favory, R.; Lancel, S.; Neviere, R.; Mathieu, D.
M. A. Organometallics 1987, 6, 1105–1110. (c) Cotton, F. A.; Dunbar,
K. R.; Verbruggen, M. G. J. Am. Chem. Soc. 1987, 109, 5498–5506. (d)
Cotton, F. A.; Dunbar, K. R. J. Am. Chem. Soc. 1987, 109, 3142–3143. (e)
Cotton, F. A.; Dunbar, K. R.; Eagle, C. T. Inorg. Chem. 1987,
26, 4127–4130. (f) Lahuerta, P.; Paya, J.; Peris, E.; Pelinghelli, M. A.;
Tiripicchio, A. J. Organomet. Chem. 1989, 373, C5–C7. (g) Morrison,
E. C.; Tocher, D. A. Inorg. Chim. Acta 1989, 156, 99–105. (h) Morrison,
E. C.; Tocher, D. A. Inorg. Chim. Acta 1989, 157, 139–140. (i) Lahuerta,
P.; Pay ꢀa , J.; Peris, E.; Aguirre, A. S.; García-Granda, A.; G ꢀo mez-Beltr ꢀa n,
F. Inorg. Chim. Acta 1992, 192, 43–49. (j) Lahuerta, P.; Pay ꢀa , J.; Solans,
Bull. Acad. Natl. Med. 2006, 190, 1961–1975. (b) McGrath, J. J. Carbon
Monoxide, 2nd ed.; Salem, H., Katz, S. A., Eds.; CRC, Boca Raton, 2006;
p 695. (c) Wilkinson, L. J.; Waring, R. H.; Steventon, G. B.; Mitchell,
S. C. Molecules of death. Carbon monoxide-the silent killer, 2nd ed.;
Imperial College Press, London, 2007.
(2) http://emedicine.medscape.com/article/1009092-overview (accessed
May 25, 2011).
(
3) (a) Claireaux, G.; Thomas, S.; Fievet, B.; Motais, R. Respir.
Physiol. 1988, 74, 91–98. (b) Marsh, M. G.; Marino, G.; Pucci, P.;
Ferranti, P.; Malorni, A.; Kaeda, J.; Marsh, J.; Luzzatto, L. Hemoglobin
ꢀ
X.; Ubeda, M. A. Inorg. Chem. 1992, 31, 385–391. (k) Lahuerta, P.; Peris,
1
991, 15, 43–51.
E. Inorg. Chem. 1992, 31, 4547–4551. (l) Borrachero, M. V.; Estevan, F.;
Lahuerta, P.; Pay ꢀa , J.; Peris, E. Polyhedron 1993, 12, 1715–1717. (m)
Estevan, F.; Lahuerta, P.; Latorre, J.; Peris, E.; García-Granda, S.;
G ꢀo mez-Beltr ꢀa n, F.; Aguirre, A.; Salvad, M. A. J. Chem. Soc., Dalton
(4) Benzon, H. T.; Brunner, E. A. JAMA 1978, 240, 1955–1964.
(5) Wang, B.; Zhao, Y. D.; Hu, L. M.; Cao, J. S.; Gao, F. L.; Liu, Y.;
Wang, L. J. Chin. Sci. Bull. 2010, 55 (3), 228–232.
6) Itou, M.; Araki, Y.; Ito, O.; Kido, H. Inorg. Chem. 2006,
5, 6114–6116.
7) Paul, S.; Amalraj, F.; Radhakrishnana, S. Synt. Math. 2009, 159,
019–1061.
8) Benito-Garagorri, D.; Puchberger, M.; Mereiter, K.; Kirchner, K.
Angew. Chem., Int. Ed. 2008, 47, 9142–9145.
9) Giulino, A.; Gupta, T.; Altman, M.; Lo Schiavo, S.; Mineo, P. G.;
ꢀ
(
Trans. 1993, 1681–1688. (n) Lahuerta, P.; Ubeda, M. A.; Pay ꢀa , J.;
4
García-Granda, S.; G ꢀo mez-Beltr ꢀa n, F.; Anillo, A. Inorg. Chim. Acta 1993,
205, 91. (o) Lahuerta, P.; Latorre, J.; Peris, E.; Sana ꢀu , M.; Ubeda, M. A.;
Garcia-Granda, S. J. Organomet. Chem. 1993, 445, C10–12. (p) Pruch-
nik, F. P.; Starosta, R.; Lis, T.; Lahuerta, P. J. Organomet. Chem. 1998,
568, 177–183. (q) Pruchnik, F. P.; Starosta, R.; Smolenski, P.; Shesta-
kova, E.; Lahuerta, P. Organometallics 1998, 17, 3684–3689.
(
1
(
(
Fragal ꢁa , I. L.; Evmenenko, G.; Dutta, P.; Van der Boom, M. E. Chem.
Commun 2008, 2900–2902.
(22) (a) Taber, D. F.; Malcolm, S. C.; Bieger, K.; Lahuerta, P.; Sana ꢀu ,
M.; Stiriba, S. E.; P ꢀe rez-Prieto, J.; Monge, M. A. J. Am. Chem. Soc. 1999,
121, 860–861. (b) Lahuerta, P.; Martínez-Ma ~n ez, R.; Pay ꢀa , J.; Peris, E.
Inorg. Chim. Acta 1990, 173, 99–105. (c) Lahuerta, P.; Pay ꢀa , J.;
Pellinghelli, M. A.; Tiripicchio, A. Inorg. Chem. 1992, 31, 1224–1232.
(23) Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307–326.
(10) See for instance: (a) Climent, E.; Marcos, M. D.; Martínez-
M ꢀa ~n ez, R.; Sancen ꢀo n, F.; Soto, J.; Rurack, K.; Amor ꢀo s, P. Angew. Chem.,
Int. Ed. 2009, 48, 8519–8522. (b) Climent, E.; Bernardos, A.; Martínez-
M ꢀa ~n ez, R.; Maquieira, A.; Marcos, M. D.; Pastor-Navarro, N.; Puchades,
R.; Sancen ꢀo n, F.; Soto, J.; Amor ꢀo s, P. J. Am. Chem. Soc. 2009,
2 6 4 6 5 2 2 2 3 2
(24) The crystal structure of [Rh [(C H )P(C H ) ] (O CCH ) ]
3
131, 14075–14080. (c) Comes, M.; Aznar, E.; Moragues, M.; Marcos,
(CO) (3 CO) has been deposited at the Cambridge Crystallographic
3
M. D.; Martínez-M ꢀa ~n ez, R.; Sancen ꢀo n, F.; Soto, J.; Villaescusa, L. A.; Gil,
Data Centre and allocated the deposition number CCDC 820952.
(25) Kou, X.; Rasika Dias, H. V. Dalton Trans. 2009, 7529–7536.
(26) Basallote, M. G.; Bozogli ꢀa n, F.; Fernandez-Trujillo, M. J.;
Martínez, M. New J. Chem. 2007, 32, 264–272.
ꢀ
L.; Amor ꢀo s, P. Chem.—Eur. J. 2009, 15, 9024–9033. (d) Abalos, T.;
Royo, S.; Martínez-M ꢀa ~n ez, R.; Sancen ꢀo n, F.; Soto, J.; Costero, A. M.; Gil,
S.; Parra, M. New J. Chem. 2009, 33, 1641–1645. (e) Aznar, E.; Coll, C.;
Marcos, M. D.; Martínez-M ꢀa ~n ez, R.; Sancen ꢀo n, F.; Soto, J.; Amor ꢀo s, P.;
Cano, J.; Ruiz, E. Chem.—Eur. J. 2009, 15, 6877–6888. (f) Climent, E.;
Calero, P.; Marcos, M. D.; Martínez-M ꢀa ~n ez, R.; Sancen ꢀo n, F.; Soto, J.
Chem.—Eur. J. 2009, 15, 1816–1820.
(27) Basallote, M. G.; Blanco, E.; Bl ꢀa zquez, M.; Fern ꢀa ndez-Trujillo,
M. J.; Litr ꢀa n, R.; M ꢀa ~n ez, M. A.; Ramírez del Solar, M. Chem. Mater. 2003,
15, 2525–2532.
(28) Espenson, J. H. Chemical Kinetics and Reaction Mechanisms;
McGraw-Hill: New York, 1981.
(29) A sample of silica containing the complex was heated at a rate of
5° per minute until a weight loss equivalent to the two L ligands is
observed. Alternatively, when stirring a suspension of the prepared silica-
adsorbed sample in chloroform at 70 °C for prolonged periods, followed
by reduced pressure drying, a sample of axially free dirhodium complex
was also produced.
(
11) Hirva, P.; Esteban, J.; Lahuerta, P.; P ꢀe rez-Prieto, J. Inorg. Chem.
007, 46, 2619–2626.
12) Esteban, J.; Ros-Lis, J. V.; Martínez-Ma ~n ez, R.; Marcos, M. D.;
Moragues, M.; Soto, J.; Sancen ꢀo n, F. Angew. Chem., Int. Ed. 2010,
9, 4934–4937.
13) Estevan, F.; Gonz ꢀa lez, G.; Lahuerta, P.; Martínez, M.; Peris, E.;
van Eldik., R. J. Chem. Soc., Dalton Trans. 1996, 1045–1050.
14) Esteban, J.; Hirva, P.; Lahuerta, P.; Martínez, M. Inorg. Chem.
006, 45, 8776–8784.
2
(
4
(
(
(30) Tobe, M. L.; Burgess, J. Inorganic Reaction Mechanisms; Longman:
U.K., 1999.
2
1
5771
dx.doi.org/10.1021/ja206251r |J. Am. Chem. Soc. 2011, 133, 15762–15772

Journal of the American Chemical Society
ARTICLE
(
31) Albert, J.; Andrea, L.; Bautista, J.; Gonzalez, A.; Granell, J.;
Font-Bardia, M.; Calvet, T. Organometallics 2008, 27, 5108–5117.
32) Aull ꢀo n, G.; Chat, R.; Favier, I.; Font-Bardía, M.; G ꢀo mez, M.;
Granell, J.; Martínez, M.; Solans, X. Dalton Trans 2009, 8292–8300.
33) (a) G ꢀo mez, M.; Granell, J.; Martínez, M. Eur. J. Inorg. Chem.
000, 217–224. (b) G ꢀo mez, M.; Granell, J.; Martínez, M. J. Chem. Soc.,
(
(
2
Dalton Trans. 1998, 37–44. (c) Gonz ꢀa lez, G.; Lahuerta, P.; Martínez, M.;
Peris, E.; Sanau, M. J. Chem. Soc., Dalton Trans. 1994, 545–550.
(
34) Esteban, J.; Esteban, F.; Sanau, M. Inorg. Chim. Acta 2009,
62, 1179–1184.
35) Esteban, J.; Esteban, F.; Sanau, M. Inorg. Chim. Acta 2008, 361,
274–1280.
3
(
1
(
(
36) Lahuerta, P.; Peris, E. Inorg. Chem. 1992, 31, 4547–4551.
37) Zinchuk, A. V.; Hancock, B. C.; Shalaev, E. Y.; Reddy, R. D.;
Govindarajan, R.; Novak, E. Eur. J. Pharm. Biopharm. 2005, 61, 158–170.
38) Sheldrick, G. M. SHELXTL, v. 5.1; Bruker AXS, Inc. ed.:
Madison, WI, 2000.
(
1
5772
dx.doi.org/10.1021/ja206251r |J. Am. Chem. Soc. 2011, 133, 15762–15772