Solvent-dependent interconversions between Rhi, RhII, and RhIII complexes of an aryl-monophosphine ligand

Article abstract of DOI:10.1002/chem.200700805

Reaction of the aryl-monophosphine ligand α²- (diisopropylphosphino)isodurene (1) with the Rh^I precursor [Rh(coe)₂(acetone)₂]BF₄ (coe = cyclooctene) in different solvents yielded complexes of all thre

Full text of DOI:10.1002/chem.200700805

FULL PAPER
DOI: 10.1002/chem.200700805
Solvent-Dependent Interconversions between RhI,Rh II,and Rh III Complexes
of an Aryl–Monophosphine Ligand
Michael Montag,[a] Gregory Leitus,[b] Linda J. W. Shimon,[b] Yehoshoa Ben-David,[a] and
David Milstein*[a]
Abstract: Reaction of the aryl-mono-
phosphine ligand a2-(diisopropylphos-
phino)isodurene (1) with the RhI pre-
hydride complex (2) was obtained.
However,when the reaction was car-
ried out in acetone or dichloromethane
a dinuclear h6-arene RhII complex (5)
was obtained in the absence of added
redox reagents. Moreover,when aceto-
nitrile was added to a solution of either
the RhII or RhIII complexes,a new sol-
I
cursor [Rh
G
G
vent-stabilized,noncyclometalated Rh
cyclooctene) in different solvents yield-
ed complexes of all three common oxi-
dation states of rhodium,depending on
the solvent used. When the reaction
complex (6) was obtained. In this
report we describe the different com-
plexes,which were fully characterized,
and probe the processes behind the re-
markable solvent effect observed.
À
Keywords: arene complexes · C H
activation · oxidation state · phos-
phanes · rhodium · solvent effects
was carried out in methanol a cyclome-
III
talated,solvent-stabilized Rh
alkyl-
Introduction
tionation of RhI and RhIII complexes,as reported by several
other groups.[5]
Complexes of divalent rhodium have been studied exten-
sively since the discovery of the prototypical rhodium(II)
carboxylates in the 1960s.[1] Since then,hundreds of dinu-
clear and polynuclear species of RhII,as well as a smaller
number of mononuclear species,have been isolated and
characterized. Many of them,especially dinuclear com-
Herein,we describe a cationic rhodium–monophosphine
system which undergoes facile conversion between mononu-
clear RhI and RhIII species,but,more intriguingly,also
yields a dinuclear RhII species without the addition of an ex-
ternal redox reagent. Moreover,these interconversions were
found to be strongly influenced by the coordinating ability
of the solvent. In fact,we found that it is possible to funnel
the system into just one of the oxidation states simply by
use of the appropriate solvent.
[2]
plexes,have also been utilized for catalytic applications.
RhII complexes are usually accessible by oxidation of RhI
or reduction of RhIII complexes,both processes being ac-
complished by the use of appropriate redox reagents.[3] Only
rarely do RhII species form spontaneously from monovalent
or trivalent rhodium complexes,without the addition of ex-
ternal redox reagents. Such transformations may be accom-
plished by either disproportionation of RhI complexes,as re-
cently reported by Willems and co-workers,[4] or compropor-
The rhodium–monophosphine system is based on the
ligand a2-(diisopropylphosphino)isodurene (ligand 1,
Scheme 1a),which was previously reported by our group in
the context of cyclometalation by neutral PtII complexes.[6,7]
Similar cyclometalation was also found to occur upon reac-
tion of ligand 1 with the cationic RhI precursor [Rh
ACHTREUNG(coe)2-
AHCTRE(UNG acetone)2]BF4 (coe=cyclooctene). However,in contrast to
the PtII system,the present cyclometalation reaction was
found to be highly reversible,resulting in the interesting va-
lence interconversions described in this work.
[a] M. Montag,Y. Ben-David,Prof. D. Milstein
Department of Organic Chemistry
Weizmann Institute of Science
Rehovot 76000 (Israel)
Fax : (+972)8-934-4142
E-mail: david.milstein@weizmann.ac.il
Results and Discussion
[b] Dr. G. Leitus,Dr. L. J. W. Shimon
Unit of Chemical Research Support
Weizmann Institute of Science
Rehovot 76000 (Israel)
Reaction of ligand 1 with [Rh
of solvent on the product: When ligand 1 was allowed to
react with one equivalent of the RhI precursor [Rh
(coe)2-
ACHRTUNEG(coe)2ACHTRE(UGN acetone)2]BF4—effect
ACHTREUNG
Chem. Eur. J. 2007, 13,9043 – 9055
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
9043

significant changes in the NMR
spectra of the solution. In the
31P{1H} NMR spectrum the
signal at d=111.81 ppm (d, 1J-
AHCTRE(GNU Rh,P)=190.6 Hz),attributed
to complex 2b,slowly de-
creased and two new signals ap-
peared (with
ratio): a double doublet at d=
86.69 ppm (1J
(Rh,P)=151.1 Hz,
2J
(Rh,P)=20.7 Hz) and
double triplet at d=56.73 ppm
(1J 2J-
(Rh,P)=161.7 Hz,
(Rh,P)ꢀ J(P,H)=21.7 Hz). In
a 1:1 integral
AHCTREUNG
A
a
AHCTREUNG
2
A
ACHTREUNG
1
the H NMR spectrum,the hy-
dride signal of complex 2b,
which appears in acetone as a
Scheme 1. a) Formation of the solvent stabilized alkylrhodium
nol; 2b/4b: Solv=acetone). b) Formation of dimeric complex 5 (Solv=acetone,CH Cl2).
ACHTRE(UNG III) hydride complexes (2a/4a: Solv=metha-
2
double doublet
at
d=
À
A
À23.99 ppm (1J
N
ACHTRE(UNG P,H)=20.1 Hz),gave
2
way to a new hydride multiplet at d=À12.63 ppm (td, J-
1
3
A
ACHRTUNEG(Rh,H)=26.1 Hz, JACHRTE(UGN P,H)=7.7 Hz). Other changes
included the appearance of four aromatic signals instead of
the original two and redistribution of the signals of the
methylene groups (from d=2.5–2.9 to 2.4–3.9 ppm) with sig-
nificant low-field shifts (ꢀ1 ppm). Furthermore,electro-
spray mass spectrometry showed that the primary product
of the above reaction was a doubly charged species with a
mass of 706.5 g molÀ1,which is exactly twice the mass of
complex 2b,excluding the solvent molecules and counter-
ion. This indicated that the new product was some dimeric
form of complex 2. Indeed,full characterization,including
X-ray crystallography (see below),revealed the product,
complex 5,to be a dimeric structure made up of two units
of complex 2. It must be noted that the rapid formation of
this new complex from 2 is in marked contrast to the behav-
ior of complex 4,which was found to be stable in acetone
for days at room temperature.[8,11] It appears that the pri-
mary reason for the differences between the two systems is
the difference in their steric bulk (see below).
plex 2a exhibits a doublet at d=111.85 ppm (1J
A
1
193.0 Hz) and the H NMR spectrum shows a characteristic
1
hydride signal at d=À24.72 ppm (dd, J
ACHTRE(UNG Rh,H)=29.4 Hz, J-
ACHTREUNG
dride situated trans to a vacant coordination site or to a
loosely coordinated ligand (in this case,methanol). The
methylene group attached to the metal center gives rise to
1
two multiplets in the H NMR spectrum,at d=2.73 (dd, J-
(H,H)=8.8 Hz, 2J
(Rh,H)=3.0 Hz) and 2.56 ppm (unre-
solved),and one multiplet in the 13C{1H} NMR spectrum,at
d=13.65 ppm (dd, 1J(Rh,C)=25.4 Hz, 2J
(P,C)=7.4 Hz).
A
ACHTREUNG
A
N
Molecular structure of dimeric complex 5: The unique fea-
tures of complex 5 became clearly apparent upon X-ray dif-
fraction crystallographic analysis. Suitable crystals of com-
plex 5 were grown from tetrahydrofuran (THF) at room
temperature. The crystal structure parameters are given in
This is consistent with a nonsymmetrical,cyclometalated
structure,as shown in Scheme 1a. The number of coordinat-
ed solvent molecules in 2a could not be determined directly
due to difficulties in its isolation in pure solid form.[9] None-
theless,examination of the crystal structure of its tetrame-
thylethylenediamine adduct (see below),as well as analogy
with complex 4a,[10] leads us to conclude that three solvent
molecules are coordinated.
Table 1. The unit cell contains four asymmetric units,each
À
of which comprises the dicationic complex 5,two BF
4
counterions and one noncoordinated solvent molecule. The
cationic complex itself,the structure of which is shown in
Figure 1,possesses a chiral structure (with two asymmetric
centers at the metal atoms) and crystallizes as a racemic
mixture. The complex is made up of two molecules of ligand
1 that bridge together two RhII centers (the oxidation state
was inferred from a formal electron count). As can be seen
in Figure 1,each ligand is bound to one metal center via the
phosphine group and to the second metal center via the aro-
matic moiety (in an h6 fashion).[12] Furthermore,one of the
When the reaction between
1
and [RhAHCRTE(UNG coe)2-
ACHTREUNG(acetone)2]BF4 was carried out in acetone,a very similar
complex, 2b,which is stabilized with acetone instead of
methanol molecules,was rapidly obtained. However,when
the acetone solution was kept at room temperature for a
few hours,or warmed at 60 8C for a few minutes,a pro-
nounced color change from light yellow to dark blue-violet
was observed. The visible change was also accompanied by
9044
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2007, 13,9043 – 9055

Interconversions of Rhodium Complexes
FULL PAPER
Table 1. Crystallographic data for complexes 5 and 7.
5 (from THF)
The metal-to-metal distance
in the crystal structure is
2.6708(9) Š,which falls within
5 (from acetone)
7
formula
W A
[gmolÀ1
C36H62B2F8OP2Rh2
959.24
C35H60B2F8OP2Rh2
938.21
C22H45BF4N2OPRh
574.29
À
F
C
]
the range of known Rh Rh
single bonds.[14] Support for the
existence of this bond in solu-
tion was provided by both
31P{1H} and 1H NMR experi-
space group
crystal system
a[Š]
b[Š]
c[Š]
a[8]
P21/n (no. 14)
monoclinic
10.174(2)
31.348(6)
12.688(3)
90.0
P21/n (no. 14)
monoclinic
10.217(2)
31.049(6)
12.679(3)
90.0
92.16(3)
90.0
4019(1)
4
P21/c (no. 14)
monoclinic
11.645(2)
11.209(2)
20.688(4)
90.0
102.95(3)
90.0
2631.7(9)
4
2
ments in which a large J cou-
b[8]
g[8]
92.70(3)
90.0
4042(1)
4
pling was observed between the
phosphorus and rhodium nuclei
cell volume[Š3]
(2J
A
Z
1calcd [gcmÀ3
]
1.565
1.550
0.965
1.449
0.755
as coupling of the hydride nu-
cleus to both of the phosphorus
m
A
]
0.961
120(2)
120(2)
MoKa (g=0.71073 Š)
3.92
120(2)
MoKa (g=0.71073 Š)
5.64
nuclei (2J(P,H)=26.1 Hz, 3J-
ACHTREUNG
radiation
R(F)[%]
MoKa (g=0.71073 Š)
5.64
9.42
AHCTREUNG
AHCTREUNG
R
A
8.97
12.38
thermore,the existence of a
À
Rh Rh bond is consistent with
ligands is also bound to the metal center (Rh1) through a
methylene bridge that extends from the aromatic ring. The
second metal atom (Rh2) bears a hydride ligand,[13] consis-
tent with the above NMR data.
the diamagnetic nature of the complex,as judged by the var-
ious NMR spectra of complex 5. As for the exact location of
the hydride ligand in solution,the fact that it is attached
only to one metal center,which is different from the one
binding the methylene bridge,was confirmed by a combina-
À
Figure 1. Top: Crystal structure of complex 5 (isolated from THF). All hydrogen atoms (except for the hydride),two BF
counterions,and one THF
4
molecule were omitted for clarity. Bottom: View of ring C21–C26 along the b axis of the unit cell in crystals isolated from THF (right) and acetone
(left),showing the interactions with the BF counterion and solvent molecule. The THF molecule on the right is partly disordered.
4
Chem. Eur. J. 2007, 13,9043 – 9055
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
9045

D. Milstein et al.
tion of 103Rh–1H and 103Rh–31P correlation spectroscopy ex-
periments on a CD2Cl2 solution of 5. These experiments also
confirmed the existence of two magnetically inequivalent
rhodium atoms (as seen in the crystal structure),giving rise
to two 103Rh NMR signals at d=À205.27 and
À471.87 ppm.[15]
Another interesting observation regarding the crystal
structure of complex 5 was made when the crystal isolated
from THF was compared with another crystal isolated from
acetone (grown at À258C). It was found that although the
cocrystallized solvent is different (acetone versus THF) the
two crystal structures are virtually identical in every other
term,including the unit-cell dimensions (see Table 1) and
the relative positions of counterions and solvent molecules.
In particular,both the acetone and THF molecules are lo-
cated close to the methylene-bridged arene moiety (ring
C21–C26),such that the oxygen atoms of both solvents
point towards that ring (Figure 1b). In fact,the distance be-
tween the oxygen atom and H25 (hydrogen bound to C25)
of ring C21–C26 is practically identical in both cases,at
2.50 Š,which is well within the sum of van der Waals radii
of hydrogen and oxygen atoms (rH +rO =1.20+1.52 Š=
2.72 Š). This implies the existence of a hydrogen bond be-
tween the complex (donor) and the solvent molecule (ac-
ceptor). The existence of such interactions may account for
The distances between the metal centers and their corre-
sponding h6-bound arene rings are very similar,being mea-
sured as 1.85 and 1.82 Š for Rh1 and Rh2,respectively.
These values are comparable to those reported by Dixon
and co-workers for a cationic,mononuclear h6-arene piano-
stool RhII complex.[16] The two aromatic rings of complex 5
À
are slanted towards each other around the Rh Rh bond
axis,such that the dihedral angle defined by the centroid of
each ring and the two Rh atoms (arene-Rh-Rh-arene) is
1348. This geometrical arrangement is probably imposed by
the existence of the methylene bridge (Ar-CH2-Rh),without
which the angle would have been closer to 1808 due to steric
repulsions. The bond length between Rh1 and the benzylic
carbon (C30) is 2.141(6) Š,which is longer than the analo-
gous bond (2.087(7) Š) in the mononuclear RhIII complex of
1
the low-field shift of the aromatic protons in the H NMR
spectrum of complex 5 in [D6]acetone (d=7.61–6.42 ppm)
ligand 1 discussed below (complex 7).[17] The Rh P bonds of
relative to those measured in CD2Cl2 (d=7.27–6.25 ppm).
À
À
À
À
complex 5 (Rh1 P1 2.2746(15) Š,Rh2 P2 2.2657(18) Š)
are slightly longer than the analogous bond of the mononu-
clear RhIII complex (2.242(2) Š),yet shorter than those of
Furthermore,one BF
counterion is positioned near ring
4
C21–C26,such that the distance between one of its fluorine
atoms (F6) and H23 (hydrogen bonded to C23) is about
2.49 Š,regardless of the solvent from which the crystal was
isolated (Figure 1b). This distance is also within the sum of
van der Waals radii of hydrogen and fluorine atoms (rH +
rF =1.20+1.47 Š=2.67 Š),again implying hydrogen bond-
ing.
the aforementioned RhII complex reported by Dixon and
[16]
co-workers (2.3381(7) Š,2.3463(8) Š).
Further data re-
garding bond lengths and angles in complex 5 are given in
Table 2.
Table 2. Selected bond lengths [Š] and angles [8] for complex 5 (from
THF).
Effect of solvent on the dimerization of 2 and its reversibili-
ty: As mentioned above,when complex 2 was prepared and
kept in methanol it was found to be highly stable,even
when warmed to 808C. In acetone,on the other hand,com-
plex 2 underwent conversion to dimer 5 even at room tem-
perature,but this reaction was nonquantitative and eventu-
ally reached a state of equilibrium (e.g.,within 11 h for [ 2]0
ꢀ40 mm,[18,19] at which point the molar ratio was 2:5ꢀ1:1).
Nonetheless,when the weakly coordinating dichlorome-
thane was used as solvent quantitative formation of 5 was
observed. Thus,when ligand 1 was reacted with one equiva-
À
À
Rh1 Rh2
2.6708(9)
2.2657(18)
1.50(7)
2.263(5)
2.343(6)
2.410(6)
2.243(6)
2.334(6)
2.340(6)
1.83
Rh1 P1
2.2746(15)
2.141(6)
2.342(6)
2.281(5)
2.322(6)
1.85
2.210(6)
2.401(6)
2.339(6)
2.48(5)
À
À
Rh2 P2
Rh1 C30
À
À
Rh2 H
Rh1 C15
À
À
Rh1 C16
Rh1 C17
À
À
Rh1 C18
Rh1 C19
À
À
Rh1 C20
Rh1 arene
À
À
Rh2 C21
Rh2 C22
À
À
Rh2 C23
Rh2 C24
À
À
Rh2 C25
Rh2 C26
À
À
Rh2 arene
F6 H23
À
O1 H25
2.53(6)
P1-Rh1-Rh2
P1-Rh1-C30
85.11(5)
83.80(14)
131.86(6)
P2-Rh2-Rh1
96.85(4)
lent of [RhACHTERNGU(coe)2ACHTRE(UGN acetone)2]BF4 (80 mm of each,after
P1-Rh1-Rh2-P2
P2-Rh2-Rh1-C30
143.54(14)
mixing) in dichloromethane complex 5 was observed within
only a few minutes at room temperature (evidenced by the
rapid color change from orange-yellow to brown and then
dark blue-violet) and virtually quantitative yield was ob-
tained within six hours. In fact,we utilized this method for
the synthesis of pure samples of complex 5 (see Experimen-
tal Section). Lastly,it is interesting to note that the impor-
tant role of solvent in the formation of complex 5 was also
highlighted by the observation that the absence of solvent
can also promote this reaction. This was seen during at-
tempts to isolate pure samples of complexes 2a and 2b by
evaporating their respective solutions under vacuum. For ex-
ample,when a freshly prepared solution of complex 2b in
acetone,containing no complex 5,was evaporated to dry-
When one examines a space-filling model of 5 it becomes
clearly evident that the structure of this complex is sterically
congested,as the arene and isopropyl moieties of the two
phosphine ligands are in very close proximity. Had ligand 1
been replaced by 3,the extra methyl group added to each
isopropyl group would probably cause the structure to col-
lapse,either by reductive elimination of the methylene
bridge (Ar-CH2-Rh) to relieve the strain,or by complete de-
composition into monomeric complex 4. This may account
for the fact that no dimerization is observed for 4 after sev-
eral days in acetone at room temperature.
9046
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2007, 13,9043 – 9055

Interconversions of Rhodium Complexes
FULL PAPER
ness for a few minutes and then kept under vacuum at room
In an attempt to probe the relationship between com-
plexes 2c and 6a,we carried out the above experiments at
low temperatures. When a sample of pure 5 was dissolved in
cold (À358C) acetonitrile,and was then examined at
À308C,complexes 2c and 6a were found at nearly equimo-
lar amounts,in contrast to the abovementioned case at
room temperature. Then,as the sample was warmed to
room temperature,the amount of complex 2c quickly de-
creased,as the amount of 6a increased,until the system
reached a state of equilibrium,within 30 minutes,for which
the molar ratio was 2c:6aꢀ1:15. Direct evidence for this
equilibrium was obtained from spin saturation transfer
(SST) experiments which showed that 2c undergoes facile,
temperature for 2.5 h,it was found that nearly 90% of
2
had been converted to 5 in the absence of any solvent (a
similar observation was also made in the case of 2a).
The dimerization of complex 2 into 5 was also found to
be highly reversible,as demonstrated by the reactions of
5
with various solvents. When a sample of pure complex 5 was
dissolved in either methanol or acetone,at room tempera-
ture,it underwent conversion to complex 2a or 2b,respec-
tively. The extent of conversion was highly dependent on
the solvent,such that in methanol complex 5 was completely
converted into 2a (within 19 h for initial concentration
[5]0 =10 mm),whereas in acetone the system was found to
reach a state of equilibrium (within 6 h,for the same initial
concentration of 5),at which point the molar ratio was 2:5
ꢀ1:1 (equivalent to ꢀ30% conversion). However,when
acetonitrile was used to dissolve complex 5 a very different
result was obtained. When this solvent,which is more
strongly coordinating than either acetone or methanol,was
added to a sample of pure 5 at room temperature an imme-
diate color change from blue-violet to yellow took place.
This was also accompanied by the complete disappearance
of 5 and formation of two new complexes: a major product,
which was different from 2,and a minor product ( ꢀ16
mol%),[20] which was identified as the acetonitrile adduct of
2 (2c; see Experimental Section for characterization). NMR
analysis of the major product revealed it to be complex 6a,
À
reversible C H reductive elimination of the methylene
bridge (Ar-CH2-Rh). Because 6a is the fully opened product
[21]
À
of C H reductive elimination from 2c, it is reasonable to
assume that complexes 2c and 6a are in mutual equilibrium.
These observations clearly indicate that complex 2c is not
merely a byproduct of the reaction of 5 with acetonitrile,
but is in fact intimately related to the formation of 6a.[22]
This conclusion was further supported by the observation
that adding excess acetonitrile (ꢀ50 equivalents) to a cold
(À358C) methanol solution of complex 2a ([2a]0 =44 mm)
resulted in 2c as the major product,with a molar ratio of
2c:6aꢀ3:1 (as observed at À308C). In this case,complex 2c
most probably originated directly from 2a by simple solvent
substitution,which explains the initially high ratio of 2c:6a
noncyclometalated monophosphine RhI species (see
(i.e.,solvent substitution is faster than C H reductive elimi-
À
a
Scheme 2),which was also obtained in high yield by addi-
tion of excess acetonitrile to a methanol solution of 2a. The
31P{1H} NMR spectrum of complex 6a in CD2Cl2 exhibits a
nation). As the sample was then warmed to room tempera-
ture,complex 6a became the predominant species,as de-
scribed above for the acetonitrile solution (with a very simi-
lar product ratio).
doublet at d=53.05 ppm (1J
ACHTRE(UNG Rh,P)=175.0 Hz) and its
1H NMR spectrum is generally characteristic of a freely ro-
tating monodentate ligand (for example,the aromatic pro-
tons are magnetically equivalent,as are the methylene pro-
tons; see the Experimental Section for further details).
Effect of chelation on the dimerization of 2: Further evi-
dence to the importance of solvent dissociation in the dime-
rization reaction was obtained when the chelating ligand tet-
ramethylethylenediamine (TMEDA) was added to complex
Scheme 2. Reactions of complexes 2a and 5 with acetonitrile.
Chem. Eur. J. 2007, 13,9043 – 9055
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
9047

D. Milstein et al.
2. When one equivalent of this ligand was added to complex
2 the conversion to dimer 5 was completely prevented,even
in weakly coordinating solvents and/or elevated tempera-
tures (e.g.,over one hour at 80 8C in dioxane). The TMEDA
adduct of 2,complex 7 (Scheme 3),was isolated from either
the nature of this fluxionality came from H/D exchange and
SST experiments. When a solution of complex 7 in CD3OD
was warmed to 508C facile H/D exchange took place be-
tween the solvent,the hydride ligand,and the arene methyl
group which is situated ortho to the phosphine “arm”,as in-
dicated by the disappearance of the corresponding signals in
the 1H NMR spectrum. Moreover,room-temperature SST
experiments (also in methanol) clearly demonstrated direct
chemical exchange between the hydride ligand and the
arene methyl group mentioned above. It is noteworthy that
in both exchange experiments the signals associated with
the bridging methylene (Ar-CH2-Rh) could not be observed
due to overlap with the TMEDA protons and it is therefore
unclear whether this group participates in any exchange.[23]
Nonetheless,the above observations indicate that complex 7
À
undergoes facile C H reductive elimination,even at room
temperature.[24] It was shown that ligand dissociation is a
À
prerequisite for C H reductive elimination from octahedral
alkyl hydride RhIII complexes,[25] and hence it is reasonable
to assume that dissociation of either the solvent molecule or
one of the amine groups of the TMEDA ligand (probably
trans to the methylene bridge,due to its strong trans influ-
ence) contributes to the dynamic process of complex 7.
Crystal structure of 7: Crystals of complex 7 suitable for X-
ray diffraction structural analysis were obtained from a con-
centrated acetone solution cooled to À308C. The crystal
structure (Scheme 3b) was found to contain a coordinated
water molecule,which probably originated from trace
amounts of water in the acetone. It exhibits a rhodium
center situated in a slightly distorted octahedral environ-
ment,with the hydride ligand located trans to the water
molecule,in accordance with the NMR data (hydride trans
Scheme 3. a) Reaction of complex
THF) with TMEDA to afford complex 7 (7a: Solv=acetone; 7b: Solv=
THF). b) Crystal structure of complex 7 (H2O adduct). All hydrogen
atoms (except for the hydride) and the BF4 counterion were omitted for
clarity.
2 (2b: Solv=acetone; 2d: Solv=
À
À
to a labile solvent molecule). The Rh N1 (trans to methyl-
À
ene bridge) bond length (2.260(6) Š) is longer than the Rh
acetone (7a) or THF (7b) and fully characterized by solu-
tion NMR techniques and X-ray crystallography
(Scheme 3b; see the Experimental Section for more details).
A solution of complex 7 in acetone exhibits a doublet at d=
N4 (trans to phosphine) bond (2.220(6) Š),the difference
being due to the higher trans influence of the methylene as
compared with the phosphine donor. The other metal-to-
ligand bond lengths (see Table 3) are similar to those ob-
ACHTREUNG
92.32 ppm (1J(Rh,P)=165.5 Hz) in the 31P{1H} NMR spec-
trum,which is markedly different from the signal attributed
Table 3. Selected bond lengths [Š] and angles [8] for complex 7 (H2O
adduct).
to complex 2 (d=111.81 ppm,doublet, 1J
AHCTREUNG
Furthermore,the 1H NMR spectrum shows a high-field hy-
dride signal at d=À26.65 ppm,which is characteristic of a
hydride trans to a vacant coordination site or to a loosely
coordinated ligand,such as acetone. However,the most in-
triguing attribute of 7 is related to the chelating TMEDA
ligand itself. At room temperature,all protons associated
with this ligand give rise to very broad,overlapping signals
at d=2.50–3.00 ppm in the 1H NMR spectrum,indicating
that a facile dynamic process takes place. The individual sig-
nals become resolved only at low temperatures,for example,
À738C. Similar fluxionality is also observed for the parts of
ligand 1 which are close to the first coordination sphere,
that is,some of the isopropyl protons and all of the methyl-
À
À
Rh1 C1
2.087(7)
2.260(6)
2.315(6)
83.6(2)
81.6(2)
89.0(3)
174(3)
Rh1 P2
2.242(2)
2.220(6)
1.47(8)
103.3(2)
91.2(3)
85(3)
À
À
Rh1 N1
Rh1 N4
À
À
Rh1 O1
Rh1 H
C1-Rh1-P2
P2-Rh1-N1
N4-Rh1-C1
C1-Rh1-H
N1-Rh1-N4
C1-Rh1-O1
O1-Rh1-H
N1-C17-C18-N4
60.3(9)
served for the ketol adduct of complex 4,which closely re-
sembles complex 7 and was previously reported by our
group.[8]
ene protons (Ar-CH2-P,Ar-CH -Rh),the latter being practi-
Side reactions and mechanistic insights related to the forma-
2
cally unobserved at room temperature. Further evidence for
tion of complex 5: When ligand 1 was reacted with one
9048
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2007, 13,9043 – 9055

Interconversions of Rhodium Complexes
FULL PAPER
Scheme 4. a) Reaction of 1 with [Rh
(acetone)2]BF4.
ACHTRE(UGN coe)2ACHERT(UGN acetone)2]BF4 in acetone,leading to complex 2b and side products. (b) Reaction of complex 8 with [RhACHTREUNG(coe)2-
ACHTREUNG
equivalent of [Rh
tion of complex 2b was accompanied by the appearance of
two side products (see Scheme 4a). The first side product
A
G
in the solution (with a molar ratio of 5:2b:6bꢀ2:2:1,after
11 h). In an attempt to clarify this observation we reacted a
pure sample of 8 in acetone with one equivalent of [Rh-
gives rise to
a
ACHRTUNEG(coe)2ACHTREU(GN acetone)2]BF4 (simulating the free precursor found in
189.2 Hz) in the 31P{1H} NMR spectrum and was identified
as the acetone analogue of complex 6a. This acetone adduct
(6b) could not be isolated,due to its facile cyclometalation
reaction to afford complex 2b,but its acetonitrile analogue
was fully characterized,as described above. The second side
product gives rise to two signals in the 31P{1H} NMR spec-
the above reaction mixture; see Scheme 4b). The two com-
plexes fully reacted at room temperature within less than 15
minutes to generate complexes 2b and 6b (in a molar ratio
[27]
of roughly 6:1,respectively).
When the resulting solution,
which was practically void of complex 8,was then warmed
to 608C complex 5 was generated. It may therefore be con-
cluded that complex 8 probably has no significant role in
the formation of dimer 5.
On the other hand,complex 6b was found to be persistent
in the acetone solution,along with complex 2b. Complex 2b
was found by SST experiments to undergo facile,reversible
trum: a double doublet at d=80.48 ppm (2J
1J
(Rh,P)=128.9 Hz) and another double doublet at d=
47.67 ppm (2J
ACHTRE(UNG P,P)=332.9 Hz,
ACHTREUNG
1
(P,P)=332.8 Hz, J
N
more,the 1H NMR spectrum of this complex exhibits a
broad singlet at d=À23.11 ppm,characteristic of a hydride
ligand trans to a vacant coordination site or to a loosely co-
ordinated solvent molecule (acetone). Full NMR analysis of
this complex revealed it to be the new complex 8,an adduct
of complex 2 with ligand 1. This complex was synthesized in-
dependently as a pure substance and fully characterized (see
Experimental Section).[26] It is important to note that under
the reaction conditions discussed here the existence of com-
plex 8 means that the solution must also contain an equiva-
À
C H reductive elimination of the methylene bridge (Ar-
CH2-Rh),in a manner similar to its bulkier analogue,com-
plex 4.[8] It is therefore reasonable to assume that,like com-
plexes 2c and 6a discussed above,complexes 2b and 6b are
also in mutual equilibrium,since 6b is the fully opened
[21]
À
product of C H reductive elimination from 2b. This equi-
librium is strongly influenced by the solvent,as demonstrat-
ed above,such that in methanol only 2a is observed,in ace-
tonitrile 6a is the predominant species,and in acetone 2b is
the predominant species.
lent amount of the unreacted precursor [Rh
ACHTRE(UNG coe)2-
ACHTREUNG
fact that two molecules of ligand 1 can bind to the RhI
center is in marked contrast to the behavior of the bulkier
ligand 3,only one molecule of which can bind to the metal
center.[8] This further demonstrates the effect of ligand bulk-
iness on its reactivity,as was evident in the dimerization re-
action (see above).
Both the existence of the above equilibrium (2Q6) and
the solvent effect can be rationalized on the basis of density
functional theory (DFT) calculations previously reported by
our group regarding complex 4a.[28] These calculations were
based on a simplified model system which had methyl in-
stead of tert-butyl substituents on the phosphine moiety,and
hence they are also relevant to the current case. The DFT
study identified three pathways that may be involved in the
As the mixture of complexes 2b, 6b,and 8 (40,20,and
[19]
5 mm,respectively)
in acetone was allowed to stand at
À
room temperature complex 5 slowly formed,whereas the
amounts of the former complexes decreased. However,as
the system approached equilibrium only complex 8 disap-
peared completely,whereas complexes 2b and 6b remained
reversible C H reductive elimination from the cyclometalat-
ed complex (2a or 4a),all of which include the dissociation
of one solvent molecule from the parent complex as a pre-
requisite for C H elimination. Furthermore,all pathways in-
À
Chem. Eur. J. 2007, 13,9043 – 9055
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
9049

D. Milstein et al.
À
volve C H agostic intermediates,which are crucial for the
stability of the alkyl-hydride complexes by serving as a bar-
rier to decomposition. Nonetheless,one of the pathways
does lead to a noncyclometalated product analogous to com-
plex 6 (with methanol instead of acetonitrile or acetone li-
À1
gands),which was calculated to be about 9 kcalmol
less
stable than the cyclometalated complex. This result accounts
for the fact that the noncyclometalated complexes of either
ligands 1 or 3 have not been observed in methanol.
When methanol is replaced with either acetone or aceto-
nitrile,the relative stabilities of the cyclometalated ( 2) and
noncyclometalated (6) complexes change. Acetone,which is
a “softer” s donor than methanol,is bound more weakly to
the RhIII center of 2 (as shown experimentally in our previ-
ous work[28]),but is expected to bind more strongly to the
“softer” RhI center of 6. Therefore,the energy gap between
2b and 6b is expected to diminish relative to the methanol
case,as was indeed observed experimentally,since both
complexes were found to coexist in acetone. As for the
effect of acetonitrile,it is a “softer” donor than either meth-
anol or acetone,and therefore the changes in relative stabil-
ities should be even more pronounced than in the case of
acetone. Moreover,acetonitrile can also function as a
p
acid,receiving backdonation from the rhodium center,espe-
I
cially if it is relatively electron-rich,as in the case of Rh rel-
ative to RhIII. Consequently, 6a becomes more stable than
2c,such that the former is the predominant species ob-
served when excess acetonitrile is present. It is also impor-
tant to note the effect of solvent on the formation and sta-
À
bility of the C H agostic intermediates mentioned above. In
III
general,the more labile the solvent is with respect to Rh
,
À
the more facile is the C H reductive elimination and forma-
tion of the appropriate agostic intermediate. Furthermore,
because all of the agostic intermediates are RhI species they
should also be stabilized by acetone and acetonitrile relative
to methanol.
The above observations regarding complexes 2 and 6,to-
gether with the very structure of complex 5 (which is essen-
tially an interlocking combination of complexes 2 and 6),
point to a plausible mechanism that can lead to the forma-
tion of the RhII dimer. The general outline of this mecha-
nism is shown in Scheme 5. The initial stage would be a bi-
molecular reaction of RhIII complex 2 with RhI complex 6 to
give a mixed-valence RhI/RhIII species in which the rhodium
center of one complex is coordinated to the arene moiety of
the second complex.[29,30] We have identified two such possi-
ble species,namely 9 and 10,neither of which have been ex-
perimentally observed. We believe that the pathway involv-
ing intermediate 9 is more likely,because the formation of
this species requires the dissociation of only two solvent
molecules from complex 6 (at least from the electronic
point of view),whereas the formation of intermediate 10 re-
quires three solvent molecules to be released. Nevertheless,
both pathways should then lead to the same mixed-valence
RhI/RhIII species 11 (not observed),which has the two rho-
Scheme 5. Putative mechanisms for the formation of complex 5.
comproportionation to yield the dirhodium(II) complex 5,
in a process that might involve hydride-bridged electron
transfer.[31]
The putative mechanism described herein accounts for
the strong solvent effect observed for the formation (and de-
composition) of complex 5. Virtually all of the mechanistic
stages leading to 5 involve solvent dissociation and hence
rely heavily on the coordinating ability of the solvent. In
particular,the very first stage is strongly hindered by sol-
vents such as methanol or acetonitrile that strongly stabilize
either species 2 or 6,respectively,as was shown above. On
the other hand,more weakly coordinating solvents,such as
acetone and dichloromethane,would allow both monomeric
species to coexist,as well as allow the dissociation of solvent
molecules necessary for the coordination of the arene moiet-
ies to the rhodium centers.
À
dium centers closely positioned,but still lacks a true Rh Rh
bond. This complex would then undergo intramolecular
Regarding the reversibility of the overall mechanism (e.g.,
reaction of 5 with acetonitrile to yield 6),our findings are
9050
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2007, 13,9043 – 9055

Interconversions of Rhodium Complexes
FULL PAPER
was prepared according to a literature procedure,with appropriate modi-
fications.[34] All solvents were reagent grade or better. All nondeuterated
solvents were refluxed over sodium/benzophenone ketyl and distilled
under argon atmosphere. Deuterated solvents were used as received. All
the solvents were degassed with argon or nitrogen and kept in the glove-
box over 3 Š or 4 Š molecular sieves (except for acetone,which was
dried with Drierite). Commercially available reagents were used as re-
ceived.
strongly supported by previous works that have shown that
both solvent and arene coordination to cationic rhodium
centers can be highly reversible. For example,Bittersmann
and co-workers showed that complexes of the form [Rh
arene)L2]ClO4 (arene=mesitylene,toluene,benzene; L =P-
(OPh)3) undergo reversible displacement of the arene li-
(h6-
ACHTREUNG
ACHTREUNG
gands by solvent (acetone) molecules,as well as other arene
ligands.[32] Similar reactivity was observed by Singewald and
co-workers in a cationic rhodium complex containing two
monophosphine ligands with flexible arene-containing arms,
which are reminiscent of ligand 1.[33]
Analysis: NMR spectra (1H, 13C, 19F,and 31P) were recorded by using
Brüker Avance-400 or Brüker Avance-500 NMR spectrometers. 103Rh–1H
and 103Rh–31P correlation spectroscopies were carried out on the Brüker
Avance-500 NMR spectrometer,by using a TXI probe (with a dedicated
31P channel). All measurements were performed at 208C,unless noted
otherwise. 1H and 13C NMR chemical shifts are reported in ppm relative
1
to tetramethylsilane. H NMR chemical shifts are referenced to the resid-
ual hydrogen signal of the deuterated solvents,and the 13C NMR chemi-
cal shifts are referenced to the 13C signal(s) of the deuterated solvents.
19F NMR chemical shifts are reported in ppm relative to CFCl3 and refer-
enced to an external solution of C6F6 in CDCl3. 31P NMR chemical shifts
are reported in ppm relative to H3PO4 and referenced to an external
85% solution of phosphoric acid in D2O. Abbreviations used in the de-
scription of NMR data are as follows: Ar,aryl; br,broad; v,virtual; s,
singlet; d,doublet; m,multiplet.
Conclusion
The reaction of ligand 1 with one equivalent of the RhI pre-
cursor [RhACHTREUNG(coe)2ACHTREUNG(acetone)2]BF4 yielded several products,de-
pending on the solvent used. In methanol,a hard coordinat-
III
ing solvent,the aryl-hydride Rh complex 2a was obtained,
which is analogous to the previously reported complex 4a.
However,in acetone and dichloromethane,which are softer
and more weakly coordinating solvents than methanol,the
reaction yielded a dimeric,dinuclear Rh structure,complex
5. A third product,the noncyclometalated Rh complex 6a,
Electrospray (ES) mass spectrometry was performed at the Chemical
Analysis Laboratory (Chemical Research Support Unit),Weizmann Insti-
tute of Science,by using a Micromass Platform LCZ 4000 spectrometer
(Micromass,Manchester,UK) with cone voltage of 43 V,extractor volt-
age of 4 V,and desolvation temperature of 150 8C. Elemental analysis
was performed at the H. Kolbe Mikroanalytisches Laboratorium,Mül-
heim an der Ruhr,Germany.
II
I
was obtained when the soft,strongly-coordinating acetoni-
trile was added to either 2 or 5. Thus,the choice of solvent
could be used to dictate the outcome of the reaction,both
in terms of complex structure and formal oxidation state.
The critical role of solvent coordination was further demon-
strated by showing that the addition of a chelating ligand
(TMEDA) to complex 2 completely prevented its transfor-
mation to complex 5,even when the solvent used was ace-
tone and high temperatures were employed. On the basis of
these observations,as well as previous theoretical work
done by our group,we discussed the behavior of the mono-
phosphine–rhodium system in terms of relative stabilities
and suggested a putative mechanism for the formation of
the unique dirhodium(II) complex 5.
In summary,this work clearly demonstrates the remark-
able role of the solvent in the coordination chemistry of rho-
dium. Thus,the complex structure and the metal oxidation
state can be determined simply by the choice of solvent.
This behavior of the cationic monophosphine–rhodium
system described here stems from stabilization of the metal
center by solvent molecules and from the versatile nature of
ligand 1,which can serve as a s donor (via the phosphine),
Synthesis of a2-(diisopropylphosphino)-isodurene (1): The synthesis of
this ligand has been already reported by our group,[35] but the following
procedure was found to provide
equipped with a magnetic stirring bar was loaded,in the glove-box,with
solution of 2-bromomethyl-1,3,5-trimethylbenzene[36] (12.78 g,
a better yield. A pressure vessel
a
60.0 mmol) in methanol (50 mL). To this was added a solution of diiso-
propylphosphine (8.5 g,71.9 mmol) in methanol (25 mL). The vessel was
then tightly closed,removed from the glove-box and heated at 50 8C,
with stirring,for 48 h. At this stage the vessel was cooled to room tem-
perature,reintroduced into the glove-box and triethylamine (20 mL,
143 mmol) was added. The solution was stirred at room temperature for
30 min and the solvent was then removed under vacuum. The resulting
residue was extracted with ether,filtered and the solvent was removed
again under vacuum. The crude residue was distilled under high vacuum
to afford the pure product as a colorless oil (12.6 g,50.3 mmol,84%
yield). B.p. 1058C,0.01 mmHg; 1H NMR ([D6]acetone): d=6.77 (s,2H;
Ar-H),2.81 (d, 2J
2.17 (s,3H; Ar-C H3),1.80 (m, 3J
PCH
(CH3)2),1.10 (dd, 3J(P,H)=11.9 Hz, 3J
(CH3)2),0.99 ppm (dd, 3J(P,H)=12.3 Hz, 3J
(CH3)2); 13C{1H} NMR ([D6]acetone): d=136.82 (d, J
134.99 (d, 5J(P,C)=2.4 Hz,Ar),134.12 (d, 2J
(P,C)=5.4 Hz,Ar),129.80
(d, 4J(P,C)=1.5 Hz, CAr-H),24.26 (d, 1J
(P,C)=16.4 Hz,P CH(CH3)2),
23.95 (d, 1J(P,C)=22.9 Hz,Ar- CH2-P),21.21 (d, 2J
(P,C)=6.5 Hz,P CH-
(CH3)2),19.91 (d, 3J (CH3)2),19.78 ppm (d, 3J
(P,C)=3.6 Hz,PCH (P,C)=
5.9 Hz,PCH
(CH3)2). Assignment of the 13C{1H} NMR signals was con-
firmed by 13C DEPT 135; 31P{1H} NMR ([D6]acetone): d=5.45 ppm (s).
Reaction of [Rh(coe)2A(acetone)2]BF4 with ligand in CD3OD and
ACHTREUNG
A
ACHTREUNG
A
N
ACHTREUNG
A
G
ACHTREUNG
3
A
ACHTRE(UNG P,C)=3.0 Hz,Ar),
A
ACHTREUNG
A
N
ACHTREUNG
A
ACHTREUNG
A
G
A
ACHTREUNG
À
p donor (via the arene moiety),and substrate for C H acti-
AHCTREUNG
vation.[7] We believe that these findings should be kept in
mind in the design and mechanistic evaluations of reactions
promoted and catalyzed by rhodium complexes.
A
G
1
[D6]acetone—formation of complex 2: The following procedure pertains
to CD3OD,but was also used with [D 6]acetone as solvent. To a solution
of [RhACHTREGNU(coe)2ACHTRE(UGN acetone)2]BF4 (17.0 mg,0.032 mmol) in CD 3OD (0.25 mL)
was added ligand 1 (8.1 mg,0.032 mmol) dissolved in CD OD (0.55 mL).
3
Experimental Section
The color of the solution immediately changed from orange to yellow
and then rapidly to pale yellow. The product could not be isolated in dry
solid form due to its instability under vacuum and the characterization
was therefore accomplished in situ. The yield of the product in methanol
was quantitative,based on 1H and 31P{1H} NMR spectra. In acetone the
All experiments with metal complexes and the phosphine ligand were
carried out under an atmosphere of purified nitrogen in an
MBraun MB 150B-G glove-box. The complex [RhACTHERNGU(coe)2ACHTREU(NG acetone)2]BF4
Chem. Eur. J. 2007, 13,9043 – 9055
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
9051

D. Milstein et al.
2
product underwent slow conversion into complex 5,as described in the
main text.
56.94 ppm (dvt, 1J
(Rh,P)=161.7 Hz, 2J
U
ACHTRE(UNG P,H)=22.5 Hz, P-Rh-
H); 103Rh NMR (CD2Cl2): d=À205.27 (Rh-H), À471.87 ppm (Rh-
CH2Ar); ESI MS (MeOH): m+/z (%): 353 (100) [cationic dimer, z=2],
705 (47) [cationic dimerÀH+, z=1]; mÀ/z (%): 87 (100) [BF4]; elemental
analysis calcd (%) for C32H54B2F8P2Rh2: C 43.67,H 6.18,F 17.27; found:
C 43.45,H 6.25,F 17.37.
2a: 1H NMR (CD3OD): d=6.91 (s,1H; Ar- H),6.71 (s,1H; Ar- H),2.87
(m, 2J
(H,H)=8.8 Hz, 2J
CH2-P and Ar-CH2-Rh),2.27 (s,3H; Ar-C H3),2.27 (m,1H; PC H-
(CH3)2),2.20 (s,3H; Ar-C H3),2.08 (m,1H; PC
(CH3)2),1.34 (m, 3J-
(CH3)2),1.01 (m,6H; PCH-
(Rh,H)=29.4 Hz, 2J
(P,H)=21.9 Hz,1H; Rh-
H); 13C{1H} NMR (CD3OD): d=150.95 (d, J
(P,C)=6.7 Hz,Ar),136.05
(s,Ar),135.65 (d, (P,C)=5.3 Hz,Ar),131.16 (s,Ar),127.75 (s, CAr-H),
125.67 (s, CAr-H),27.73 (dd, 1J
(P,C)=32.4 Hz, J=1.5 Hz,P CH(CH3)2),
26.78 (d, 1J (CH3)2),21.07 (d, 1J
(P,C)=22.5 Hz,P CH (P,C)=31.7 Hz,Ar-
CH2-P),21.01 (s,Ar- CH3),20.34 (s,Ar- CH3),18.79 (PCH (CH3)2),18.52
(PCH(CH3)2),18.35 (PCH (CH3)2),18.23 (PCH
(CH3)2),13.65 ppm (dd, 1J-
ACHTREUNG ACHTREUGN
(H,H)=14.1 Hz, 2J(P,H)=11.3 Hz,1H; Ar-C H2-P),2.73 (dd, 2J-
A
ACHTREUNG(Rh,H)=3.0 Hz,1H; Ar-C H2-Rh),2.56 (m,2H; Ar-
X-ray structural analysis of complex 5 that was grown from tetrahydro-
furan: Complex 5 was crystallized from a tetrahydrofuran solution at
room temperature.
A
HACHTREUNG
3
G
ACHTREUNG(P,H)=13.7 Hz,6H; PCH ACHTRENUG
A
N
ACHTREUNG
AHCTREUNG
Crystal data: C32H54P2Rh2 +2BF4 +C4H8O,brown prism,0.5”0.3”
0.05 mm3,monoclinic, P21/n (no. 14), a=10.174(2), b=31.348(6), c=
12.688(3) Š, b=92.70(3)8,from 20 degrees of data, T=120(2) K, V=
JACHTREUNG
A
ACHTREUNG
4042(1) Š3, Z=4, Fw =959.24 gmolÀ1, 1calcd =1.565 gcmÀ3, m=0.961 mmÀ1
.
A
N
ACHTREUNG
AHCTREUNG
Data collection and processing: Nonius Kappa CCD diffractometer,
MoKa (l=0.71073 Š),graphite monochromator,45774 reflections collect-
ed, À12ꢁhꢁ12,0 ꢁkꢁ37,0 ꢁlꢁ15,frame scan width =1.08,scan speed
0.68 minÀ1,typical peak mosaicity 0.467 8,9315 independent reflections
(Rint =0.153). The data were processed with Denzo-Scalepack.
A
N
ACHTREUNG
ACHTREUNG ACHTREUNG(P,C)=7.4 Hz,Ar- CH2-Rh). Assignment of the
(Rh,C)=25.4 Hz, 2J
13C{1H} NMR signals was confirmed by 13C DEPT 135; 31P{1H} NMR
1
(CD3OD): d=111.85 ppm (d, J
A
1
2b: Selected H NMR ([D6]acetone): d=6.82 (s,1H; Ar- H),6.75 (s,1H;
Ar-H),2.92 (m,1H; Ar-C H2-P),2.82 (m,1H; Ar-C H2-Rh),2.75–2.67 (m,
2H; Ar-CH2-P and Ar-CH2-Rh),2.30 (s,3H; Ar-C H3),2.18 (s,3H; Ar-
Solution and refinement: The structure was solved by direct methods
with SHELXS-97. Full-matrix least-squares refinement based on F2 with
SHELXL-97. 483 parameters with no restraints,final R1 =0.0564 (based
CH3), À23.99 ppm (dd, 1J
(Rh,H)=31.0 Hz, 2J
ACHRTEUNG ACHTRE(UNG P,H)=20.1 Hz,1H; Rh-
on F2) for data with I>2s(I) and R1 =0.1260 for 7282 reflections,good-
1
H); 31P{1H} NMR ([D6]acetone): d=111.81 ppm (dd, J
(Rh,P)=189.2 Hz,
À3
ness-of-fit on F2 =1.016,largest electron density peak =1.421 Š
.
2J
Reaction of [Rh
of complex 5: To
0.181 mmol) in CH2Cl2 (1.5 mL) was added ligand
ACHTREUNG
X-ray structural analysis of complex 5 that was grown from acetone:
Complex 5 was crystallized from an acetone solution at À258C.
A
(acetone)2]BF4 with ligand 1 in CH2Cl2—synthesis
solution of [Rh(coe)2A(acetone)2]BF4 (95.3 mg,
(46.1 mg,
a
A
CHTREUNG
Crystal data: C32H54P2Rh2 +2BF4 +C3H6O,orange prism,0.3”0.3”
0.2 mm3,monoclinic, P21/n (no. 14), a=10.217(2), b=31.049(6), c=
12.679(3) Š, b=92.16(3)8,from 20 degrees of data, T=120(2) K, V=
1
0.184 mmol) dissolved in CH2Cl2 (1.1 mL). The solution was stirred at
room temperature for 16 h during which its color changed from light
orange to dark blue-violet. The solvent was removed under vacuum and
the resulting solid was redissolved in CH2Cl2 (1.0 mL) and added to pen-
tane (17 mL),with stirring,to precipitate the product. The liquid phase
was then decanted and the product was washed with pentane and dried
under vacuum to afford the product as a blue-violet powder (79.5 mg,
0.090 mmol,quantitative yield).
4019(1) Š3, Z=4, Fw =938.21 gmolÀ1, 1calcd =1.550 gcmÀ3, m=0.965 mmÀ1
.
Data collection and processing: Nonius Kappa CCD diffractometer,
MoKa (l=0.71073 Š),graphite monochromator,40579 reflections collect-
ed, À12ꢁhꢁ0, À37ꢁkꢁ37, À15ꢁlꢁ15,frame scan width =0.98,scan
speed 0.438 minÀ1,typical peak mosaicity 0.36 8,15660 independent re-
flections (Rint =0.057). The data were processed with Denzo-Scalepack.
1H NMR (CD2Cl2): d=7.27 (s,1H; Ar- H),7.03 (s,1H; Ar- H),6.77 (s,
Solution and refinement: The structure was solved by direct methods
with SHELXS-97. Full-matrix least-squares refinement based on F2 with
SHELXL-97. 438 parameters with no restraints,final R1 =0.0392 (based
1H; Ar-H),6.25 (s,1H; Ar- H),3.92 (dd, 2J
A
ACHTREUNG
9.1 Hz,1H; Ar-C H2-P,left part of ABX-system),3.87 (m,
7.3 Hz, 2J
ACHTREUNG
on F2) for data with I>2s(I) and R1 =0.0571 for 8200 reflections,good-
ACHTREUNG
ness-of-fit on F2 =1.038,largest electron density peak =0.903 Š
.
À3
A
ACHTREUNG
Reaction of complex 5 with acetonitrile at low temperature—in situ for-
mation of complex 2c: Complex 2c rapidly converts to complex 6a,as
described in the main text,and therefore the following low-temperature
technique was employed in order to facilitate the measurement of NMR
A
ACHTREUNG
AHCTREUNG
ACHTREUNG
A
ACHTREUNG
spectra for 2c. Complex
CD3CN (0.62 mL),which was pre-cooled to
5
(16.7 mg,0.019 mmol) was dissolved in
À358C. The resulting
HACHTREUNG
orange-yellow solution was immediately transferred to the NMR spec-
trometer and the sample was rapidly cooled to À308C before measure-
ment was commenced. Some of the NMR signals of 2c were obscured by
those of 6a,because both complexes coexist in solution (molar ratio of
ꢀ1:1 under the conditions described here; see below for full characteri-
zation of 6a).
JACHTREUNG
H
A
HACHTREUNG
2
A
ACHTRE(UNG P,H)=
A
ACHTREUNG
JACHTREUNG
1H NMR (CD3CN, À308C): d=6.84 (s,1H; Ar- H),6.70 (s,1H; Ar- H),
JACHTREUNG
2.90 (m, 2J(H,H)=14.2 Hz, 2J(P,H)=11.5 Hz, 3J
ACHTREUNG AHCERTUNG ACHERT(UNG Rh,H)=2.1 Hz,1H; Ar-
JACHTREUNG
CH2-P),2.48 (dd, 2J
2.42 (m,2H; Ar-C H2-Rh),2.22 (s,3H; Ar-C H3),2.19 (s,3H; Ar-C H3),
2.18 (m,1H; PC H
(CH3)2,overlaps with Ar-CH signals),1.72 (m, 3J-
(H,H)=7.2 Hz,1H; PC H
laps with signals from 6a),1.14 (dd, 3J
3H; PCH
(CH3)2),1.02 (dd, 3J(P,H)=15.8 Hz, 3J
PCH
(Rh,H)=18.5 Hz, 2J
ACHRTEGNU ACHTREU(GN P,H)=8.9 Hz,1H; Ar-C H2-P),
(H,H)=14.2 Hz, 2J
JACHTREUNG
G
G
ACHTREUNG
U
G
A
ACHTREUNG
AHCTREUNG
3
U
ACHTREUNG
A
N
ACHTREUNG
G
ACHTREUNG
A
ACHTREUNG
G
ACHTREUNG
A
N
ACHTREUNG
U
U
ACHTREUNG
A
G
ACHTREUNG
Rh-H); selected 13C{1H} NMR (CD3CN, À308C): d=150.76 (dd, J
ACHTREUNG(P,C)=
G
U
ACHTREUNG
G
ACHTREUNG
7.2 Hz, J=1.5 Hz,Ar),135.81 (d,
JACHTER(UNG P,C)=5.1 Hz,Ar),127.10 (s, CAr-H),
126.37 (s, CAr-H),20.95 (s,Ar- CH3),20.74 (d, 1J
ACHTRE(UNG P,C)=30.4 Hz,Ar- CH2-
G
G
A
ACHTREUNG
7.7 Hz,Ar- CH2-Rh). Assignment of the 13C{1H} NMR signals was con-
firmed by 13C DEPT 135 and 13C–1H heteronuclear correlation; 19F{1H}
NMR (CD2Cl2): d=À151.28 ppm (s,free BF 4); 31P{1H} NMR (CD2Cl2):
P),20.36 (s,Ar- CH3),11.20 ppm (dd, 1J
ACTHERNGU(Rh,C)=20.7 Hz, JACHTRE(UNG P,C)=6.4 Hz,
2
Ar-CH2-Rh). Assignment of the 13C{1H} NMR signals was confirmed by
13C DEPT 135 and 13C–1H heteronuclear correlation; 31P{1H} NMR
d=86.74 (dd, 1J
(Rh,P)=151.6 Hz, 2J
A
(CD3CN, À308C): d=108.26 ppm (d, J
ACHTRE(UNG Rh,P)=166.7 Hz).
1
9052
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2007, 13,9043 – 9055

Interconversions of Rhodium Complexes
FULL PAPER
Reaction of complex 2a with acetonitrile—synthesis of complex 6a: To a
165.5 Hz); 19F{1H} NMR ([D6]acetone,20 8C): d=À151.96 ppm (s,free
solution of [Rh
G
U
BF4).
1H NMR ([D6]acetone, À738C): d=6.97 (s,1H; Ar- H),6.64 (s,1H; Ar-
H),3.26 (mbr,1H; NC H2),3.04 (mbr,1H; Ar-C H2-P),2.89 (sbr,3H;
NCH3),2.74 (sbr,3H; NC H3),2.58 (mbr,1H; NC H2),2.55 (mbr,1H;
Ar-CH2-Rh),2.49 (sbr,3H; NC H3),2.42 (sbr,3H; NC H3),2.34 (mbr,
(0.53 mL) was added ligand 1 (19.3 mg,0.077 mmol) dissolved in CH OH
3
(0.56 mL). The solution was stirred at room temperature for 10 min and
then CH3CN (0.55 mL,10.5 mmol) was added. Stirring was continued for
20 min and the solution was then added to pentane (18 mL),with stirring,
to precipitate the product. The mixture was allowed to stand at À358C
overnight to facilitate the precipitation. The liquid phase was then de-
canted and the product was washed with diethyl ether and dried under
vacuum to afford the product (36.0 mg,0.064 mmol,83.5% yield) as a
yellow powder.
1H; PCH
2.13 (s,3H; Ar-C H3),1.30 (dd, 3J
(CH3)2),1.24 (mbr, 3J
PCH (H,H)=6.4 Hz,3H; PCH
(CH3)2),0.90 (dd, 3J(P,H)=15.3 Hz, 3J
3H; PCH (H,H)=6.7 Hz,3H;
PCH ACTHERU(NG Rh,H)=31.1 Hz, JACTHEURNG(P,H)=18.1 Hz,1H;
A
HACHTREUNG
A
ACHTREUNG
A
N
ACHTREUNG
A
U
ACHTREUNG
1
2
(CH3)2), À26.01 ppm (dd, J
1H NMR (CD2Cl2): d=6.83 (s,2H; Ar),3.01 (d, 2J
Ar-CH2-P),2.73 (s,6H; Ar-C H3),2.28 (sbr,3H; NCC H3),2.20 (s,3H;
Ar-CH3),1.98 (m,2H; PC
(CH3)2),1.92 (s,6H; NCC H3),1.31 (dd, 3J-
(P,H)=14.4 Hz, 3J (CH3)2),1.26 ppm (dd, 3J-
(H,H)=7.0 Hz,6H; PCH
(P,H)=14.4 Hz, 3J
(H,H)=7.1 Hz,6H; PCH (CH3)2). Assignment of the
1H NMR signals was confirmed by 13C–1H heteronuclear correlation;
13C{1H} NMR (CD2Cl2): d=137.98 (d, J
(P,C)=3.7 Hz,Ar),135.88 (d, J-
(P,C)=2.8 Hz,Ar),130.95 (d, (P,C)=
(P,C)=3.6 Hz,Ar),129.52 (d, 4J
2.3 Hz, CAr-H),122.48 (d, 2J
(Rh,C)=12.2 Hz, CN),118.46 (sbr, CN),
26.27 (d, 1J (CH3)2),24.78 (d, 1J
(P,C)=25.3 Hz,P CH (P,C)=19.7 Hz,Ar-
CH2-P),22.54 (s,Ar- CH3),20.75 (d, J=0.6 Hz,Ar- CH3),19.26 (s,PCH-
AHCTRE(UNG P,H)=11.6 Hz,2H;
Rh-H). Assignment of the H NMR signals was confirmed by 13C–1H het-
1
eronuclear correlation; 13C{1H} NMR ([D6]acetone, À738C): d=153.10
HACHTREUNG
(s,Ar),134.67 (d,
A
N
ACHTREUNG
126.75 (s, CAr-H),124.85 (s, CAr-H),61.93 (s,N CH2),61.39 (s,N CH2),
55.98 (sbr,N CH3),52.77 (sbr,N CH3),51.80 (sbr,N CH3),47.56 (sbr,
A
N
ACHTREUNG
NCH3),27.28 (dbr, 1J
25.8 Hz,P CH
PCH(CH3)2),18.88 (s,PCH
P),18.05 (s,PCH (CH3)2),16.43 (sbr,PCH
(Rh,C)=21.3 Hz,Ar- CH2-Rh). Assignment of the 13C{1H} NMR signals
A
(CH3)2),25.85 (dbr, 1J
ACHRTUNEG ACHTREUNG(P,C)=
AHCTREUNG
AHCTREUNG
A
J
A
ACHTREUNG
A
N
ACHTREUNG
AHCTREUNG
A
ACHTREUNG
A
N
ACHTREUNG
AHCTREUNG
was confirmed by 13C DEPT 135 and 13C–1H heteronuclear correlation;
31P{1H} NMR ([D6]acetone, À738C): d=94.28 ppm (d, 1J
ACHTREUNG(Rh,P)=
ACHTREUNG(CH3)2),18.83 (s,PCH ACHTREUNG(CH3)2),3.84 (s,NC CH3),3.40 ppm (sbr,NC CH3).
Assignment of the 13C{1H} NMR signals was confirmed by 13C DEPT 135
and 13C–1H heteronuclear correlation; 19F{1H} NMR (CD2Cl2): d=
À153.07 ppm (s,free BF 4); 31P{1H} NMR (CD2Cl2): d=53.05 ppm (d, 1J-
165.3 Hz); 19F{1H} NMR ([D6]acetone, À738C): d=À150.21 ppm (s,free
BF4); ESI MS (MeOH): m+/z (%): 469 (52) [complexÀTHF]; mÀ/z (%):
87 (100) [BF4]; elemental analysis calcd (%) for C26H51BF4N2OPRh (one
THF molecule present): C 49.70,H 8.18; found: C 49.01,H 7.96.
ACHTREUNG(Rh,P)=175.0 Hz); elemental analysis calcd (%) for C22H36BF4N3PRh: C
46.92,H 6.44; found: C 46.98,H 6.55.
X-ray structural analysis of complex 7 (H2O adduct): Complex 7 was
crystallized at À308C from a concentrated acetone solution (the coordi-
nated water molecule probably originates from trace amounts of water in
In situ formation of complex 6b: To
a solution of [RhACHTRE(UNG coe)2-
ACHTREUNG(acetone)2]BF4 (48.8 mg,0.093 mmol) in [D 6]acetone (0.45 mL) was
added ligand 1 (23.2 mg,0.093 mmol) dissolved in [D ]acetone (0.39 mL).
6
the acetone).
The primary product of this reaction was complex 2b,as described
above. Complex 6b was identified as a by-product of this reaction.
3
Crystal data: C22H45N2OP2Rh+BF4,colorless plate,0.1”0.1”0.03 mm
,
monoclinic, P21/c (no. 14), a=11.645(2), b=11.209(2), c=20.688(4) Š,
a=90.08, b=102.95(3)8, g=90.08,from 20 degrees of data, T=120(2) K,
Selected 1H NMR ([D6]acetone): d=6.92 (s,2H; Ar),3.05 (d, 2J
11.0 Hz,2H; Ar-C H2-P),2.71 (s,6H; Ar-C H3),2.22 (s,3H; Ar-C H3),
1.32 ppm (m,24H; PCH
(CH3)2); 31P{1H} NMR ([D6]acetone): d=
(Rh,P)=189.2 Hz).
Synthesis of complex 7: To solution of [RhACHRTE(UNG coe)2ACHTREU(GN acetone)2]BF4
ACHTRE(UNG P,H)=
V=2631.7(9) Š3, Z=4, Fw =574.29 gmolÀ1
,
1calcd =1.449 gcmÀ3
,
m=
0.755 mmÀ1
.
AHCTREUNG
1
55.50 ppm (d, JACHTREUNG
Data collection and processing: Nonius Kappa CCD diffractometer,
MoKa (l=0.71073 Š),graphite monochromator,13284 reflections collect-
ed, À11ꢁhꢁ11,0 ꢁkꢁ11,0 ꢁlꢁ21,frame scan width =1.58,scan speed
0.178 minÀ1,typical peak mosaicity 0.50 8,3881 independent reflections
(Rint =0.100). The data were processed with Denzo-Scalepack.
a
(55.0 mg,0.105 mmol) in THF (1.2 mL) was added tetramethylethylene-
diamine (12.2 mg,0.105 mmol) dissolved in THF (0.8 mL). The solution
was stirred at room temperature for 15 min,during which its color
changed from orange to orange-yellow. Then a solution of ligand 1
(26.4 mg,0.105 mmol) in THF (1.0 mL) was added and the resulting solu-
tion was stirred at room temperature for 50 min,during which the color
changed to yellow. The solution was concentrated under vacuum (to
0.8 mL) and then pentane (13 mL) was added,with stirring,to precipitate
the product. The liquid phase was decanted and the solids were washed
with pentane and dried under vacuum to afford the product (60.3 mg,
0.096 mmol,91.8% yield) as a light-beige powder. It is important to note
that while the solid product is 7b (THF adduct; see elemental analysis),
the NMR spectra were measured in acetone and are thus representative
of complex 7a (acetone adduct).
Solution and refinement: The structure was solved by direct methods
with SHELXS. Full-matrix least-squares refinement based on F2 with
SHELXL-97. 308 parameters with no restraints,final R1 =0.0564 (based
on F2) for data with I>2s(I) and R1 =0.0652 for 2960 reflections,good-
À3
ness-of-fit on F2 =1.097,largest electron density peak =1.097 Š
.
H/D exchange experiments for complex
7 in CD3OD: Complex 7
(22.2 mg,0.035 mmol) was dissolved in CD 3OD (0.66 mL). The 1H and
31P NMR spectra of this sample were recorded and the solution was
warmed to 508C for 3.5 h. 1H and 31P NMR spectra of the sample were
then recorded again and compared with the initial spectra. No significant
changes were observed in the 31P NMR spectrum,indicating that no ap-
preciable decomposition of complex 7 had taken place. In the 1H NMR
spectrum,on the other hand,some of the signals had disappeared,which
indicated exchange of the respective protons with deuterons from the sol-
vent.
1H NMR ([D6]acetone,20 8C): d=6.95 (s,1H; Ar- H),6.69 (s,1H; Ar- H),
3.00–2.50 (very broad overlapping signals,16H; N CH2CH2N and 2 N-
A
G
ACHTRE(UNG CH3)2),2.30 (s,3H; Ar-
3
G
ACHTRE(UNG H,H)=6.9 Hz,
A
(CH3)2), À26.65 ppm (sbr,1H;
1
Rh-H); Assignment of the H NMR signals was confirmed by 13C–1H het-
Synthesis of complex 8: A solution of [RhACHTERNGU(coe)2ACHTRE(UGN acetone)2]BF4 (69.7 mg,
eronuclear correlation; 13C{1H} NMR ([D6]acetone,20 8C): d=151.48 (d,
0.132 mmol) in acetone (0.8 mL) was added dropwise to a solution of
ligand 1 (66.3 mg,0.265 mmol) in acetone (1.0 mL) and the resulting so-
lution was stirred at room temperature for 40 min. The solution was then
added to pentane (17 mL),with stirring,to precipitate the product. The
liquid phase was decanted and the solids were washed with pentane and
dried under vacuum to afford the product (94.2 mg,0.126 mmol,95.0%
yield) as a yellow powder. The product was found to exhibit fluxional be-
havior in solution at room temperature and therefore NMR spectra were
recorded both at room temperature and at À208C.
J
A
J
A
JACHTRE(UNG P,C)=
A
ACHTREUNG
A
ACHTREUNG
A
ACHTREUNG
ACHTREUNG
AHCTREUNG
Chem. Eur. J. 2007, 13,9043 – 9055
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
9053

D. Milstein et al.
1H NMR ([D6]acetone,20 8C): d=7.00 (s,1H; Ar,cyclometalated
ligand),6.95 (s,1H; Ar,cyclometalated ligand),6.94 (s,2H; Ar,mono-
dentate ligand),3.72 (sbr,1H; Ar-C H2-P),3.58 (sbr,1H; Ar-C H2-Rh),
3.26 (sbr,1H; Ar-C H2-P),3.20–2.93 (overlapping singlets,3H; Ar-C H2-
Acknowledgements
This research was supported by the Israel Science Foundation,the DIP
program for German–Israeli Cooperation,and the Helen and Martin
Kimmel Center for Molecular Design. D.M. is the holder of the Israel
Matz Professorial Chair. The authors would like to thank the referees of
this manuscript for their valuable comments.
P+PCH
PCH(CH3)2),2.34 (s,3H; Ar-C H3,cyclometalated ligand),2.21 (s,6H;
Ar-CH3,cyclometalated ligand +monodentate ligand),1.93 (sbr,1H;
PCH(CH3)2),1.80 (s,6H; Ar-C H3,monodentate ligand),1.62–1.12 (mbr,
18H; PCH(CH3)2),1.08 (sbr,3H; PCH (CH3)2),0.50 (sbr,3H; PCH-
ACHTREUNG(CH3)2 +Ar-CH2-Rh),2.86 (sbr,1H; Ar-C H2-P),2.52 (sbr,1H;
ACHTREUNG
ACHTREUNG
A
ACHTREUNG
A
[1] For general reviews of rhodium(II) complexes,see: a) T. R. Felt-
house, Prog. Inorg. Chem. 1982, 29,73–166; b) D. G. DeWit, Coord.
Chem. Rev. 1996, 147,209–246; c) H. T. Chifotides,K. R. Dunbar in
Multiple Bonds between Metal Atoms,3rd ed. (Eds.: F. A. Cotton,
C. A. Murillo,R. A. Walton),Springer,Berlin, 2005,pp. 465–589.
[2] For reviews of catalytic applications involving dirhodium(II) com-
plexes,see: a) M. P. Doyle, Topics Organomet. Chem. 2004, 13,203–
222; b) D. J. Timmons,M. P. Doyle in Multiple Bonds Between Metal
Atoms,3rd ed. (Eds.: F. A. Cotton,C. A. Murillo,R. A. Walton),
Springer,Berlin, 2005,pp. 591–632.
[3] For example,alcohols (e.g.,methanol,ethanol) are used as mild re-
ductants in the synthesis of RhII carboxylates. Silver salts are also
commonly used.
[4] S. T. H. Willems,J. C. Russcher,P. H. M. Budzelaar,B. de Bruin,R.
de Gelder,J. M. M. Smits,A. W. Gal, Chem. Commun. 2002,148–
149.
[5] For examples,see the following: a) A. L. Balch,M. M. Olmstead, J.
Am. Chem. Soc. 1976, 98(8),2354–2356; b) X. Fu,L. Basickes,B. B.
Wayland, Chem. Commun. 2003,520–521; c) J.-Q. Wang,S. Cai,G.-
X. Jin,L.-H. Weng,M. Herberhold, Chem. Eur. J. 2005, 11,7342–
7350; d) S. Liu,G.-X. Jin, Dalton Trans. 2007,949–954.
[6] a) M. van der Boom,S.-Y. Liou,L. J. W. Shimon,Y. Ben-David,D.
Milstein, Organometallics 1996, 15,2562–2568; b) M. van der Boom,
J. Ott,D. Milstein, Organometallics 1998, 17,4263–4266.
nals was confirmed by 13C–1H heteronuclear correlation; selected 13C{1H}
NMR ([D6]acetone,20 8C): d=130.46 (s, CAr-H,monodentate ligand),
130.16 (s, CAr-H,cyclometalated ligand),128.74 ppm (s, CAr-H,cyclometa-
lated ligand). Assignment of the 13C{1H} NMR signals was confirmed by
13C DEPT 135 and 13C–1H heteronuclear correlation; 31P{1H} NMR
([D6]acetone,20 8C): d=80.48 (dd, 2J
128.9 Hz),47.67 ppm (dd,
19F{1H} NMR ([D6]acetone,20 8C): d=À151.83 ppm (s,free BF ).
1H NMR ([D6]acetone, À208C): d=6.98 (s,1H; Ar,cyclometalated
ligand),6.95 (s,3H; Ar,overlapping cyclometalated
+monodentate li-
(H,H)=
(H,H)=14.1 Hz,1H; Ar-C H2-P),
(H,H)=14.1 Hz,1H; Ar-C H2-P),3.05 (m,1H; PC (CH3)2,
overlaps with methylene signal),2.97 (m,1H; Ar-C H2-Rh),2.81 (m, 2J-
(H,H)=14.4 Hz,1H; Ar-C H2-P),2.54 (m,1H; PC (CH3)2),2.32 (s,3H;
Ar-CH3,cyclometalated ligand),2.30 (m,1H; PC (CH3)2,overlaps with
methyl signal),2.22 (sbr,3H; Ar-C H3,cyclometalated ligand),2.20 (sbr,
3H; Ar-CH3,monodentate ligand),1.82 (mbr,1H; PC (CH3)2),1.75
(sbr,6H; Ar-C H3,monodentate ligand),1.51 (dd, 3J(P,H)=10.6 Hz, 3J-
(H,H)=6.8 Hz,3H; PCH (H,H)=
(CH3)2),1.45 (dd, 3J(P,H)=10.8 Hz, 3J
7.1 Hz,3H; PCH (CH3)2),1.33 (m,6H; PCH (P,H)=
(CH3)2),1.24 (dd, 3J
16.0 Hz, 3J
(H,H)=7.0 Hz,3H; PCH (CH3)2),1.16 (m,3H; PCH (CH3)2),
0.99 (dd, 3J(P,H)=15.4 Hz, 3J
(H,H)=6.9 Hz,3H; PCH (CH3)2),0.34 (dd,
3J
(P,H)=15.5 Hz, J(H,H)=6.5 Hz,3H; PCH
(CH3)2), À22.98 ppm (m, J-
(Rh,H)=29.8 Hz,1H; Rh- H). Assignment of the 1H NMR signals was
confirmed by 13C–1H heteronuclear correlation; 13C{1H} NMR
([D6]acetone, À208C): d=146.05 (dd, (P,C)=9.3 Hz, (P,C)=2.8 Hz,
Ar),138.07 (d, (P,C)=4.1 Hz,Ar),137.42 (d, (P,C)=4.9 Hz,Ar),
137.16 (d, J(P,C)=2.6 Hz,Ar),136.85 (d, (P,C)=1.8 Hz,Ar),135.79 (d,
(P,C)=2.5 Hz,Ar),131.68 (s,Ar),130.54 (s, CAr-H,monodentate
ligand),130.13 (s, CAr-H,cyclometalated ligand),128.76 (s, CAr-H,cyclo-
metalated ligand),25.70 (d, 1J (CH3)2),25.27 (d, 1J-
(P,C)=25.4 Hz,P CH
(P,C)=23.7 Hz,P CH (Rh,C)=
(CH3)2),22.65 (dd, 1J(P,C)=16.7 Hz, 2J
2.1 Hz,P CH(CH3)2),22.21 (sbr,Ar- CH3,monodentate ligand),21.21 (dd,
1J(P,C)=11.0 Hz, 3J
(Rh,H)=1.3 Hz,Ar- CH2-P),20.95 (s,Ar- CH3,mono-
dentate ligand),20.81 (s,Ar- CH3,cyclometalated ligand),20.75 (d, 1J-
(P,C)=22.9 Hz,Ar- CH2-P),20.52 (d, 2J
(P,C)=5.7 Hz,PCH (CH3)2),20.34
(d, 2J
(P,C)=4.0 Hz,PCH (CH3)2),20.22 (s,Ar- CH3,cyclometalated
ligand),19.49 (s,PCH (CH3)2),19.39 (s,PCH (P,C)=
(CH3)2),18.84 (d, 2J
2.0 Hz,PCH (P,C)=3.1 Hz,PCH (CH3)2),17.85 (m,
(CH3)2),18.08 (d, 2J
PCH (P,C)=3.4 Hz,PCH
(CH3)2),16.13 (d, 2J (CH3)2),14.44 ppm (d, 1J-
(Rh,C)=28.0 Hz,Ar- CH2-Rh). Assignment of the 13C{1H} NMR signals
gands),3.79 (m, 2J
7.1 Hz,1H; Ar-C H2-Rh),3.29 (m, 2J
3.10 (m, 2J
E
H
[7] Ligands of this type have also been described by Werner and co-
workers. For examples,see: a) H. Werner, Dalton Trans. 2003,3829–
ACHTREUNG
3837; b) G. Canepa,C. D. Brandt,K. Ilg,J. Wolf,H. Werner,
Eur. J. 2003, 9,2502–2515; c) G. Canepa,C. D. Brandt,H. Werner,
Organometallics 2004, 23,1140–1152.
Chem.
[8] B. Rybtchinski,L. Konstantinovsky,L. J. W. Shimon,A. Vigalok,D.
Milstein, Chem. Eur. J. 2000, 6,3287–3292.
JACHTREUNG
[9] Determination of the number of coordinated solvent molecules re-
quires the removal of excess solvent. This can be accomplished by
precipitation and/or removal of the solvent under vacuum. Unfortu-
nately,attempts to isolate pure samples of 2a by these methods
were unsuccessful due to the instability of the complex in the ab-
sence of a stabilizing solvent (see the main text for further discus-
sion).
[10] It was shown by NMR and IR spectroscopies,as well as X-ray crys-
tallography,that complex 4b contains three solvent molecules (see
ref. [8]). This conclusion may also be applied to 4a,because metha-
nol is both smaller in size (less steric hindrance) and more strongly-
coordinating than acetone. We may then further extend this conclu-
sion to 2a,because the metal center in this complex is less sterically
hindered than in 4a (ligand 1 is less bulky than 3).
G
G
ACHTREUNG
was confirmed by 13C DEPT 135 and 13C–1H heteronuclear correlation;
31P{1H} NMR ([D6]acetone, À208C): d=80.84 (dd, 2J
[11] Complex 4 may also undergo dimerization,as was observed for simi-
lar systems (unpublished results),but such a process has not been
reported thus far.
A
117.7 Hz); elemental analysis calcd (%) for C35H60BF4OP2Rh (one ace-
[12] To the best of our knowledge,this is the only case reported thus far
of a dinuclear RhII complex with an h6-bound benzenoid ligand. For
examples of dinuclear RhII complexes with benzenoid ligands in
other binding modes,see: a) F. A. Cotton,E. V. Dikarev,S.-E. Stiri-
ba, Organometallics 1999, 18,2724–2726; b) H. Wadepohl,A. Metz,
H. Pritzkow, Chem. Eur. J. 2002, 8,1591–1602; c) M. A. Petrukhina,
tone molecule present): C 56.16,H 8.08; found: C 56.05,H 8.02.
Spin saturation transfer experiments:[37] Saturation transfer experiments
were performed for complexes 2 and 7 by selective saturation of the hy-
1
dride resonance in the H NMR spectrum. The response of the signals of
the appropriate aromatic and aliphatic protons was detected by compari-
son (difference spectrum) with a control experiment in which a signal-
free area was irradiated under the same conditions.
A. S. Filatov,Y. Sevryugina,K. W. Andreini,S. Takamizawa,
Orga-
nometallics 2006, 25,2135–2142.
CCDC-648207 (5 from acetone),CCDC-648208 ( 7),and CCDC-648209
(5 from THF) contain the supplementary crystallographic data for this
paper. These data can be obtained free of charge from the Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
[13] The hydride ligand was explicitly located in the electron density
map.
À
[14] For surveys of Rh Rh bond lengths,see: a) P. D. Harvey,F. Shafiq,
R. Eisenberg, Inorg. Chem. 1994, 33,3424–3426; b) ref. [1c].
9054
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2007, 13,9043 – 9055

Interconversions of Rhodium Complexes
FULL PAPER
[15] It should be noted that 103Rh chemical shifts do not correlate well
with the oxidation state of the metal. They are,however,strongly in-
fluenced by ligand type and complex geometry. For a review of this
topic see: J. M. Ernsting,S. Gaemers,C. J. Elsevier, Magn. Reson.
Chem. 2004, 42,721–736.
[16] a) F. M. Dixon,J. R. Farrell,P. E. Doan,A. Williamson,D. A. Wein-
berger,C. A. Mirkin,C. Stern,C. D. Incarvito,L. M. Liable-Sands,
L. N. Zakharov,A. L. Rheingold, Organometallics 2002, 21,3091–
3093; b) F. M. Dixon,M. S. Masar III,P. E. Doan,J. R. Farrell,F. P.
Arnold Jr.,C. A. Mirkin,C. D. Incarvito,L. N. Zakharov,A. L.
Rheingold, Inorg. Chem. 2003, 42,3245–3255.
(see Experimental Section). The fluxionality of complex 8 and relat-
ed issues will be discussed elsewhere.
[27] The mechanism of this reaction is unclear,but it is unlikely to in-
volve predissociation of ligand 1 from 8,since this complex was
found to be stable in acetone,even at elevated temperatures (e.g.,
90 min at 608C),with no free ligand being observed. Further obser-
vations related to complex 8 will be reported elsewhere.
[28] B. Rybtchinski,R. Cohen,Y. Ben-David,J. M. L. Martin,D. Mil-
stein, J. Am. Chem. Soc. 2003, 125,11041–11050.
[29] Bimetallic RhI/RhIII mixed-valence complexes have been reported,
both as labile intermediates and as stable species. For examples,see:
a) L. A. Oro,D. Carmona,M. P. Lamata,A. Tiripicchio,F. J. Lahoz,
J. Chem. Soc. Dalton Trans. 1986, 1,15–21; b) S. B. Duckett,C. L.
Newell,R. Eisenberg, J. Am. Chem. Soc. 1994, 116,10548–10556;
c) K. Polborn,K. Severin, Eur. J. Inorg. Chem. 1998,1187–1192;
d) C. Tejel,M. A. Ciriano,L. A. Oro,A. Tiripicchio,F. Ugozzoli,
Organometallics 2001, 20,1676–1682.
[17] To the best of our knowledge,no crystal structure of analogous
RhII-benzyl complexes has been reported elsewhere.
[18] The reported initial concentration of complex 2 was obtained by re-
acting [RhACHTREUNG(coe)2ACHTREUNG(acetone)2]BF4 and ligand 1,each at 80 m m concen-
tration (after mixing). The relatively low concentration of 2 ob-
tained in this reaction is due to the formation of other products,as
described in the main text.
[30] h6-arene complexes of both RhI and RhIII are known. For examples,
see: a) J. M. Townsend,J. F. Blount, Inorg. Chem. 1981, 20,269–271;
b) ref. [7c]; c) D. A. Loginov,M. M. Vinogradov,Z. A. Starikova,
P. V. Petrovskii,A. R. Kudinov, Russ. Chem. Bull. 2004, 53,1949–
[19] The concentration of complexes was estimated on the basis of
31P{1H} NMR spectra and the initial concentration of precursors
([RhACHTREUNG(coe)2ACHTREUNG(acetone)2]BF4 and ligand 1).
[20] The amount of minor product decreased even further as the system
approached equilibrium at room temperature. See the main text for
further information.
1953; d) J. S. Siegel,K. K. Baldridge,A. Linden,R. Dorta,
Chem. Soc. 2006, 128,10644–10645.
[31] Such an inner-sphere process is expected to be highly efficient and
also preserve the electronic saturation of the two rhodium centers.
J. Am.
[21] As suggested in our previous publication (ref. [8]),the chemical ex-
À
change observed by SST is due to rapid C H reductive elimination
without complete detachment of the arene moiety from the metal
center (a process which involves agostic intermediates).
[32] K. Bittersmann,E. Hildenbrand,A. Cervilla,P. Lahuerta,
nomet. Chem. 1985, 287,255–263.
[33] E. T. Singewald,X. Shi,C. A. Mirkin,S. J. Schofer,C. L. Stern,
J. Orga-
Or-
[22] The fact that addition of acetonitrile to complex 5 leads to a product
ratio of 2c:6aꢀ1:1 can be rationalized on the basis of the structure
of complex 5 (Figure 1a). This is essentially made up of one unit of
complex 2 and one unit of complex 6 (without the solvent mole-
ganometallics 1996, 15,3062–3069.
[34] B. Windmüller,O. Nürnberg,J. Wolf,H. Werner,
Eur. J. Inorg.
Chem. 1999,613–619. The procedure was modified by using AgBF
instead of AgPF6 and conducting the whole process at room temper-
ature.
4
cules),and hence the solvent-promoted breakup of complex
5
should first lead to an equimolar mixture of 2c and 6a.
[35] S.-Y. Liou,PhD Thesis,Weizmann Institute of Science (Rehovot,
Israel), 1995.
[23] We have previously shown that in complex 4 the bridging methylene
protons are also involved in the exchange. For further details,see
ref. [8].
[24] Similar observations were previously reported by our group for
complex 4,see ref. [8].
[36] Prepared according to a literature procedure. See: A. W. van der
Made,R. H. van der Made, J. Org. Chem. 1993, 58,1262–1263.
[37] For previous examples to the use of spin saturation transfer in stud-
ies of chemical exchange in rhodium systems,see: A. Vigalok,H.-B.
Kraatz,L. Konstantinovski,D. Milstein, Chem. Eur. J. 1997, 3,253–
260; and ref. [8].
[25] a) D. Milstein, J. Am. Chem. Soc. 1982, 104,5227–5228; b) D. Mil-
stein, Acc. Chem. Res. 1984, 17,221–226.
[26] Complex 8 was found to exhibit fluxional behavior in solution at
room temperature. Therefore,characterization of this complex by
NMR spectroscopy was done both at room temperature and À208C
Received: May 26,2007
Revised: August 17,2007
Published online: September 28,2007
Chem. Eur. J. 2007, 13,9043 – 9055
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
9055