Dirhodium (II) carboxylate complexes as building blocks. cis-Chelating dicarboxylic acids designed to bridge the dinuclear core

Article abstract of DOI:10.1039/b310008a

The syntheses and crystal structures of a series of dicarboxylic acids of general structure HO₂CCR₂OArOCR₂CO ₂H (LH₂) designed to span a pair of cis sites on the dirhodium(II) tetracarboxylate framewo

Full text of DOI:10.1039/b310008a

View Article Online / Journal Homepage / Table of Contents for this issue
P A P E R
Dirhodium (II) carboxylate complexes as building blocks.
cis-Chelating dicarboxylic acids designed to bridge
the dinuclear corey
Jamie Bickley, Richard Bonar-Law,* Thomas McGrath, Nirmal Singh and Alexander Steiner
Department of Chemistry, University of Liverpool, Liverpool L69 7ZD, UK
E-mail: bonarlaw@liv.ac.uk
Received (in Durham, UK) 20th August 2003, Accepted 1st December 2003
First published as an Advance Article on the web 3rd February 2004
The syntheses and crystal structures of a series of dicarboxylic acids of general structure
HO2CCR2OArOCR2CO2H (LH2) designed to span a pair of cis sites on the dirhodium(II) tetracarboxylate
framework are described. Monochelate complexes LRh2(OAc)2 and bischelate complexes L2Rh2 were
prepared by heating the diacids with Rh2(OAc)4 in N,N-dimethylaniline. Chelates were formed under kinetic
control and the best yields were obtained from diacids with methyl groups on the ether side arms and bulky
substituents on the aromatic ring next to the ether links. Effective molarities for chelation were measured under
1
thermodynamic control by H NMR in tetrachloroethane. The stability of chelates and the conformational
properties of diacids were investigated by molecular mechanics and quantum chemical techniques.
otherwise insoluble polymers are formed. Our starting point
Introduction
was the structurally characterised cis complex (Ph3CCO2)2-
Rh2(OAc)2 3b, originally prepared by heating Rh2(OAc)4 in
molten triphenylacetic acid.5 We found this complex was more
conveniently prepared by heating Rh2(OAc)4 with two equiva-
lents of triphenylacetic acid in N,N-dimethylaniline, providing
3b as the major species (57%), along with some of the trans
isomer 3c (9%), and the monosubstituted compound
(Ph3CCO2)Rh2(OAc)3 3a (7%), and trisubstituted species
(Ph3CCO2)3Rh2(OAc) 3d (11%). It is worth mentioning that
N,N-dimethylaniline (NNDMA) is a particularly effective sol-
vent for carboxylate exchanges of dirhodium complexes since
the reactions are complete within a few hours at 130–140 ꢀC
(driven by evaporation of acetic acid), reaction mixtures are
usually homogeneous, and solvent is easily removed during
work-up by an acid wash.
Metal complexes are versatile building blocks in supramolecu-
lar chemistry because metal ligand bonds may be formed rever-
sibly, a selection of well defined geometries is available, and
metal centres confer a variety of useful chemical and physical
properties on the final assembly.1 Much use has been made of
single metal complexes to organise ligands with one donor
atom, such as amines, into a variety of rings, cages and other
objects.1 However mononuclear complexes have limitations
when binding ligands which can coordinate in a two point
fashion, and here dinuclear complexes offer some advantages,
being able to anchor ligands more firmly using two bonds. The
robust ‘lantern’ geometry of dinuclear carboxylates2 has
recently been used to connect molecules bearing carboxylic
acid groups at 90 or 180ꢀ to give linear or cyclic polymers.3,4
In our work a dirhodium complex 1 with a pair of cis sites
blocked off with a chelating dicarboxylate ligand, ꢁO2C–X–
ꢁ
CO2
, was condensed with different dicarboxylic acids
HO2C–Y–CO2H to form macrocycles 2.4 This paper reports
the design, synthesis and crystal structures of some of the
chelating diacid ligands, and studies on the stability of their
dirhodium complexes.
To test the stability of the cis arrangement of triphenylace-
tates in 3b, the complex was heated with an excess of benzoic
acid, looking for selective exchange of the acetate groups by
benzoate. However some triphenylacetate exchange occurred,
so a kinetically more inert cis complex was sought. Reasoning
that a bidentate ligand should be displaced less readily due to
the chelate effect, aliphatic a,o-diacids were reacted with
Rh2(OAc)4 to give bridged cis-dicarboxylate complexes. The
nine carbon diacid was the shortest chain to produce a chelate
(O2C(CH2)7CO2)Rh2(OAc)2 4a (15%), accompanied by some
of the bischelate (O2C(CH2)7CO2)2Rh2 4b (6%) (we use the
terms monochelate and bischelate here to refer to dinuclear
complexes containing one or two bridges between neighbour-
ing carboxylate ligands). The yield of 4a was too low to be
synthetically useful, so on the basis of modelling studies we
next prepared a series of more preorganised diacids 5–12,
which turned out to be effective chelates. A related diacid 13
Results
Preliminary studies and chelate design
The current method of macrocycle synthesis requires a com-
plex with a pair of firmly attached cis-related carboxylates,
y Electronic supplementary information (ESI) available: Tables S1
and S2. See http://www.rsc.org/suppdata/nj/b3/b310008a/
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 4 2 5 – 4 3 3
425

View Article Online
based on the binaphthyl skeleton was also screened, but gave
only traces of a monochelate complex.
prepared by reaction of diacids 16 with Rh2(OAc)4 (the poly-
mer is attached either ortho or meta to the phenolic link).7
We prepared diacid 15 as a chemical model for 16, but did
not isolate a significant amount of chelate on reaction with
Rh2(OAc)4 . Gallagher et al. have described a dirhodium com-
plex with two cis-chelated eleven atom dicarboxylate ligands.8
More recently several other more sophisticated cis-spanning
diacids have been used to prepare dirhodium-based catalysts
for carbene transfer reactions.9,10
At this point we became aware of several other cis-chelated
dirhodium complexes. Taber et al. prepared the first such
example from diacid acid 14 and Rh2(CF3CO2)4 .6 We have
confirmed that 14 is a moderately effective ligand, vide infra.
Andersen et al. have reported polymer-bound chelates
Synthesis of diacids
Diacids 5, 6, 8, 11, and 12 were prepared in straightforward
fashion by O-alkylation of resorcinol or a resorcinol deriva-
tive with methyl bromoacetate or ethyl 2-bromo-2-methyl-
propanoate, followed by ester hydrolysis, illustrated in
Scheme 1 for diacid 8. Benzene-1,3-dithiol was the starting
material for the sulfur analogue 11. The use of Cs2CO3 in
acetonitrile11 was found to be efficient for making the more
hindered tertiary phenolic ethers. On a large scale most of
the Cs2CO3 is replaceable by K2CO3 , and 8 could be pre-
pared this way without the need for chromatography.
Attempts to introduce substituents larger than methyl next
to the acid group (e.g. ethyl, phenyl) by alkylation with
the appropriate a-bromoester gave low yields. In the case
of bis-tert-butyl compound 9, O-alkylation of 4,6-bis-tert-
butylresorcinol was slow, and the tert-butyl substituents were
introduced after ether bond formation by silica gel-catalysed
C-alkylation12 of the diethyl ester of 8. The bromo substitu-
ents in 7 and 10 were also conveniently introduced after
O-alkylation by direct bromination.13
Synthesis of chelates
Chelates were prepared by heating a 1:1 mixture of diacid and
Rh2(OAc)4 (both 45 mM) in NNDMA at 140 ꢀC. A mixture of
unreacted Rh2(OAc)4 , monochelate and bischelate was gener-
ally produced and separated by chromatography. The yield of
monochelate (Table 1) varies from 14% for complex 5a with-
out geminal methyl groups or ortho substituents to 79% for
complex 10a with both geminal methyls and ortho bromines.
Scheme 1 Synthesis of diacid 8.
Table 1 Yields and equilibrium constants for chelate formation
Diacid Monochelate Yielda (%) Bischelate Yield (%) K1b /M
c
5
6
5a
6a
14
47
53
63
69
79
65
65
33
—
—
—
—
—
—
14
<10
560 (ꢂ20)
d
7
7a
8a
8
8b
370 (ꢂ120)
1890 (ꢂ190)
810 (ꢂ50)
—
9
9a
9b
10b
—
4.5
10
11
12
14
10a
11a
12a
14a
6
—
14
4
12b
14b
—
80 (ꢂ10)
a
Diacid (45 mM) and Rh2(OAc)4 (45 mM) heated at 140 ꢀC for 1–4 h
b
in N,N-dimethylaniline.
Equilibrium constant for formation of
c
d
monochelate at 100 ꢀC in C2D2Cl4 . Not isolated. Not measured
due to low solubility of 7a.
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
426
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 4 2 5 – 4 3 3

View Article Online
The synthesis of 5a is certainly not clean, with several unchar-
acterised species formed. By-products from some of the other
ligands were characterised, an example being the dimeric com-
pound 8c, a minor component from the reaction of diacid 8
with Rh2(OAc)4 at higher concentration (90 mM).
Complexes with geminal methyl groups are more soluble in
non-polar solvents than those without. As expected, tert-butyl
and dodecyloxy substituents on the aromatic ring also increase
complex solubility, although ortho-bromine substituents
decreased it. The sulfur containing complex 11a was found
to be almost insoluble in any solvent, possibly due to intermo-
lecular S–Rh coordination, and had to be isolated as its more
readily handled tert-butylpyridine adduct.
Reaction of a 1:1 mixture of diacid and Rh2(OAc)4 should
give a statistical 1:2:1 ratio of unreacted Rh2(OAc)4 : mono-
chelate: bischelate, assuming there are no kinetic or thermo-
dynamic features favouring particular species, and that
oligomers are not formed. Heating a monochelate or bischelate
alone in NNDMA does not result in interconversion, but the
reaction can be partly or completely reversed if enough acetic
acid is added. Further heating, allowing the excess acetic acid
allowed to evaporate, gives the same product distribution as
when starting from Rh2(OAc)4 and diacid. This suggests that
monochelates are formed under kinetic control in NNDMA.
The greater than statistical ( > 50%) yields for 7a–12a must
then arise because the chelate ligand slows down substitution
of the remaining two acetates. Qualitative tlc experiments sup-
port this conclusion; when monochelates 8a, 9a and 10a were
reacted in parallel with excess benzoic acid, the acetates in 9a
and 10a were substituted more slowly than those in 8a.
To compare the relative stabilities of different chelates, con-
ditions were sought under which ring formation was reversible.
It was found that chelation under thermodynamic control
could be achieved by heating the reactants at relatively high
dilution in tetrachloroethane in the presence of acetic acid.
Equilibrium constants K1 (Table 1) were measured by 1H
NMR in both the forward direction (LH2 þ Rh2(OAc)4) and
the reverse direction (LRh2(OAc)2 þ AcOH), eqn.
(LH2 ¼ diacid, LRh2(OAc)2 ¼ monochelate).
1
ð1Þ
Fig. 1 Crystal structures of diacids (red ¼ oxygen, green ¼ bromine,
yellow ¼ sulfur).
Chelate stabilities generally increase in line with isolated
yields except for diacids 9 and 10 where the tert-butyl-substi-
tuted chelate 9a is more stable than the bromo-substituted che-
late 10a. The equilibrium constant K2 between monochelate
and bischelate, eqn. 2, was measured in once instance, for
the 8a, 8b pair, giving K2 ¼ 63 (ꢂ14) M. This value is some-
what lower than expected on statistical grounds (K1/4 ¼ 93
M) implying that the first chelate destabilises the second ring
slightly.
adopt a more convergent orientation in the sequence 5, 6, 7,
and diacid 10 is also more convergent than diacid 8. Most of
the diacids pack as expected, forming hydrogen bonded chains
or dimers.14 There are numerous C–Hꢃ ꢃ ꢃO, C–Hꢃ ꢃ ꢃp and in
some cases pꢃ ꢃ ꢃp interactions which in combination with the
stronger diacid hydrogen bonds presumably determine the
structural details.15 The simplest diacid 5 is the only one for
which the carboxylic acids are not all paired. This crystallised
as stacks of slightly ruffled 2D sheets with one of the acids
linked by a single, unsymmetrically bifurcated hydrogen bond.
Closer examination reveals a network of C–Hꢃ ꢃ ꢃO interactions
both within and between the sheets, with particularly short
distances to the singly hydrogen bonded acid.
ð2Þ
Crystal structures of diacids
In an attempt to correlate structure with stability, X-ray struc-
tures of diacid ligands 5, 6, 8, 10 and 11 were obtained, Fig. 1.
Suitable crystals of diacids 7 and 9 proved elusive, but the
dimethyl ester of 7 was solved, and is included. Compared to
the unsubstituted diacid 5, there is a general trend for the ether
arms to be oriented out of the plane of the aromatic ring on
introducing ortho substituents. Thus the carboxyl groups
Modelling studies
The preferred conformations of dimethyl esters of ligands 5–15
were explored with molecular mechanics using the MMFF94
force field,16 Fig. 2. Esters rather than diacids were modelled
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 4 2 5 – 4 3 3
427

View Article Online
since diacids tend to end up intramolecularly hydrogen bonded
in gas phase minimisations, whereas in reality any hydrogen
bonding is likely to be to the solvent. The pKa of NNDMA
in water17 (pKa ¼ 5.15) is higher than typical phenoxyacetic
acids (pKa ꢄ 3) so the acids may in fact be partly dissociated
in NNDMA; we have no experimental evidence on this point.
Stable conformations are arranged in Fig. 2 according the dis-
tance between the carbonyl carbons, with the population of a
particular conformation represented by the height of its bar. A
accessible conformations of suitable geometry. Overall, there
is a qualitative correlation between the synthetic yields and
the degree of preorganisation as estimated from molecular
mechanics.
For ligands 5 and 7, the conformations in the crystal
(marked with x in Fig. 2) are similar to the expected major
solution conformations. However for ligand 8 the solid state
conformation is predicted to be one the minor high energy
˚
ones in solution (COꢃ ꢃ ꢃCO ¼ 8.3 A, 1% occupied), and for 6
˚
pair of cis carbonyl carbons is separated by ꢄ3.5 A in a typical
none of the calculated conformations are good matches for
the solid state structure (the conformation marked (x) is the
closest).
dirhodium chelate complex,18 so ligands which have most of
their low energy conformations approaching this value might
be expected to be good ones.
To further investigate the thermodynamics of complexation,
empirical and quantum mechanical methods were used to
model the isodesmic exchange reaction between ligand
dimethyl esters (LMe2) and Rh2(OAc)4 (eqn. 3).
Comparing ligands 5, 6 and 7, it is clear that ortho substitu-
ents are predicted to bring the carbonyls closer together. Thus
the major conformation (52% occupied) of the unsubstituted
˚
ligand 5 has a COꢃ ꢃ ꢃCO distance of 8.8 A whereas the major
conformation (59% occupied) of the bisbromo ligand 7 has a
˚
ð3Þ
COꢃ ꢃ ꢃCO distance of 4.5 A. Comparing ligands 5 and 8, gem-
inal dimethyls are also predicted to preorganise the ligand for
chelation. Comparing ligands 8, 9 and 10, the main effect of
introducing additional aromatic substituents is to reduce the
number of extended conformations. Of the ligands 13, 14
and 15, only the all-carbon ligand 14 is predicted to have
As a semi-quantitative indication of strain induced on forming
the chelate, we subtracted the lowest energy conformation of a
ligand diester from the lowest energy conformation of its
monochelate, Table 2. For ease of comparison energy differ-
ences are quoted relative to the simplest ligand 5. The overall
trend for the resorcinol-based ligands is that chelation becomes
more favourable with increasing substitution of ligand 5, as
observed experimentally. According to these calculations
ligand 14 forms the least strained complex, although experi-
mentally it is one of the poorer ligands. Significantly, both
types of calculation predict that ligands 13 and 15, which are
ineffective chelates, should form much less stable rings.
Discussion
Why is chelate synthesis in NNDMA under kinetic control?
The mechanism of carboxylate exchange in dirhodium com-
plexes is not known, but may require prior coordination of
the incoming carboxylic acid to the Lewis acidic Rh2 centre,
followed by a thermal rate determining step.19 A basic solvent
like NNDMA would compete with incoming acids for Rh,
slowing the kinetics. This assumption is supported by the find-
ing that carboxylate exchange in tetrachloroethane at 100 ꢀC,
which is reversible, can be inhibited by donor solvents like
1,4-dioxane and NNDMA. Nevertheless, the acetates of
monochelate 8a are exchangeable in NNDMA under equili-
brating conditions4 suggesting that the thermodynamic stabi-
lity of the chelates also contributes to the slow kinetics.
A useful measure of the effectiveness of a chelating ligand is
its thermodynamic effective molarity (EM),20 defined for a
bidentate ligand as the concentration of a monodentate
Table 2 Calculated energies for chelate formation
Ligand
DDE/kJmolꢁ1 a UFFb
DDE/kJmolꢁ1 DFTc
5
6
0
ꢁ1
0
ꢁ18
ꢁ11
ꢁ20
ꢁ17
ꢁ22
þ64
ꢁ26
þ19
7
ꢁ6
8
ꢁ28
ꢁ23
ꢁ18
þ204
ꢁ31
þ69
9
10
13
14
15
a
DDE ¼ Echelate ꢁ Eligand(Me) ꢁ (E5a ꢁ E5(Me)), where Eligand(Me) is the
b
Fig. 2 Calculated conformations of ligand dimethyl esters, plotted as
a function of the distance between carbonyl carbons. The population
of each conformation is represented by the height of its bar. Confor-
mations marked x are similar to the crystal structure.
energy of the ligand dimethyl ester.
c
Universal force field as
implemented in Cerius.2 Density functional method pBP/DN* as
implemented in Spartan.
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
428
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 4 2 5 – 4 3 3

View Article Online
analogue required to give the same degree of reaction. For for-
mation of a monochelate, eqn. 1, the simplest definition is
EM1 ¼ K1/K2, where K is the equilibrium constant for reac-
tion of one end of the bidentate ligand with Rh2(OAc)4 .
Assuming that the acid groups in a diacid are equivalent
to acetic acid, then K ꢄ 1, and EM1 ꢄ K1 for the diacids
(Table 1).21 There appears to be little thermodynamic data
on large ring chelates in the literature with which to compare
these numbers, but the polydentate ligands used as sidero-
phores have EMs in the same range.22 For our purposes the
important finding is that the EMs for chelating diacids such
as 8 are much higher than the concentrations at which macro-
cycle self-assembly reactions are run, so corner protection
remains effective.
Turning to the conformational properties of the diacids,
these ligands incorporate several features familiar from the
study of organic reactivity, including steric buttressing by
ortho substituents, and the ‘gem dimethyl effect’.23 In the pre-
sent study conformational analysis was found to be a useful
and rapid way of rationalising the effectiveness of different
ligand designs. The approach is similar to the ‘near attack con-
formation’ analysis of rates of intramolecular reactions.24
Ligands which are not preorganised for chelation such as 5,
13 and 15 (see Fig. 2) have a greater tendency to react intermo-
lecularly, giving low yields of chelated complexes. The failure
to isolate any complex from ligand 15 suggests that the com-
plexes derived from polymer-bound ligand 16 may not in fact
be chelated.7
Melting points for metal complexes were > 240 ꢀC unless
otherwise indicated, and are uncorrected. Diacid 5 was
obtained from Lancaster.
(3-Carboxymethoxy-2,5-di-tert-butylphenoxy)ethanoic acid (6)
A mixture of 1,3-dihydroxy-2,5-di-tert-butylbenzene (0.50 g,
2.25 mmol), methyl bromoacetate (1.5 g, 10 mmol), and
K2CO3 (1.5 g, 11 mmol) in DMF (2 ml) was stirred at 60 ꢀC
for 2 h. The resulting slurry was partitioned between diethyl
ether and water. The organic layer was washed several times
with water, dried, and evaporated to an oil which solidified
on standing under vacuum. The crude dimethyl ester thus
obtained was refluxed in a mixture of MeOH (20 ml) and aqu-
eous NaOH (10 M; 1 ml) for 3 h. Most of the MeOH was eva-
porated under reduced pressure, and the residue acidified with
aqueous HCl. The suspension was filtered, washing with water
to give 6 as a white solid (0.70 g, 92%), mp 203–205 ꢀC (Found:
C, 63.8; H, 7.8. C18H26O6 requires C, 63.89; H, 7.81); dH(300
MHz, d6-acetone) 1.4 (18 H, s), 4.74 (4 H, s), 6.59 (1 H, s),
7.23 (1 H, s); dC(75 MHz, d6-acetone) 29.80, 34.36, 65.11,
99.25, 124.88, 129.70, 155.60, 169.55. m/z (FAB) 338.2 (Mþ).
(3-Carboxymethoxy-2,5-dibromophenoxy)ethanoic acid (7)
A mixture of 1,3-dihydroxybenzene (1.1 g, 10 mmol), methyl
bromoacetate (3.4 g, 22 mmol), and K2CO3 (3.0 g, 10 mmol)
in DMF (3 ml) was stirred at 70 ꢀC for 5 h. The resulting slurry
was partitioned between EtOAc–diethyl ether and water. The
organic layer was washed several times with water, dried,
and evaporated to an oil (2.4 g) which solidified on standing
under vacuum. The dimethyl ester of 5 thus produced was pure
enough for use in subsequent reactions. dH(200 MHz, CDCl3)
3.80 (6 H, s), 4.60 (4 H, s), 6.5 (3 H, m), 7.2 (1 H, m). Bromine
(200 ml, 3.90 mmol) was added dropwise to a stirred solution of
above dimethyl ester (0.50 g, 1.97 mmol) in DCM (10 ml) at
ice-bath temperature. After 30 min the reaction mixture was
diluted with DCM, washed with aqueous sodium bisulfite (1
M), and the organic layer dried and evaporated to a white
solid. dH(200 MHz, CDCl3) 3.80 (6 H, s), 4.66 (4 H, s), 6.44
(1 H, s), 7.72 (1 H, s). The crude dibromo diester thus obtained
was heated in a mixture of EtOH (20 ml) and aqueous NaOH
(1 M; 8 ml) at 80 ꢀC for 30 min. Most of the EtOH was evapo-
rated under reduced pressure, and the residue acidified with
aqueous HCl. The suspension was filtered, washing with a
small quantity of cold water to give 7 as a white solid (0.60
g, 79%), mp > 240 ꢀC (Found: C, 31.2; H, 2.1. C10H8Br2O6
requires C, 31.28; H, 2.10); dH(400 MHz, d6-acetone) 4.86 (4
H, s), 6.90 (1 H, s), 7.73 (1 H, s); m/z (FAB) 383.2 (Mþ);
dC(100 MHz, d6-acetone) 66.99, 102.35, 104.46, 137.00,
156.32, 169.69.
There is only limited correlation between the conformation
of a ligand in the crystal and the conformation calculated in
solution, e.g. for diacid 8 the acid groups are pointing in oppo-
site directions in the crystal, Fig. 1, but are calculated to be
more convergent in solution, Fig. 2. The large number of inter-
molecular interactions for 5 is presumably the reason why both
ends do not form a hydrogen bonded acid dimer, which is
otherwise a fairly persistent supramolecular synthon.14
In summary, we have prepared a family of cis-spanning
4þ
ligands for the Rh2 core; the ligands are easy to make, and
stability and solubility can be adjusted by adding substituents.
The picture that emerges from experiment and modelling is
that as more substituents are added to the resorcinol O,O0-dia-
cetic acid skeleton 5, the energy of the initial state of the chela-
tion process (eqn. 1) is raised more than the final state,
increasing the thermodynamic driving force and the rate of
the forward reaction. The present work appears to be one of
the few instances in which ligands have been designed to
chelate dinuclear complexes,6,9 and polynuclear species in
general.25,26
Experimental
Diacid ligands and rhodium complexes were prepared under
nitrogen. Weakly bound solvent ligands were removed from
dirhodium complexes by heating under vacuum overnight. Sol-
vent was the internal reference for NMR spectra in d6-acetone
and d2-tetrachloroethane, and TMS was the reference in
CDCl3 or mixed solvents containing mostly CDCl3 . Coupling
constants are in Hz. Fast atom bombardment (FAB) mass
spectra were obtained on a VG7070E instrument using m-
nitrobenzyl alcohol as matrix, chemical ionisation (CI) spectra
were obtained on a TRIO1000 instrument, and electrospray
(ES) spectra were obtained on a Micromass LCT instrument.
Masses are quoted for the most intense isotopomer peak in
the molecular ion envelope. Silica gel of 230–400 mesh was
used for column chromatography. Organic extracts were dried
over Na2SO4 . Solvents dimethylformamide (DMF), dichloro-
methane (DCM), ethyl acetate (EtOAc), ethanol (EtOH), acet-
onitrile (MeCN) and methanol (MeOH) were reagent grade,
and were dried when necessary using standard procedures.
2-(3-(1-Carboxy-1-methylethoxy)phenoxy)-2-methylpropanoic
acid (8)
K2CO3 (27.6 g, 200 mmol) and Cs2CO3 (16.3 g, 50 mmol) were
added to a stirred solution of 1,3-dihydroxybenzene (5.5 g, 50
mmol) and ethyl 2-bromo-2-methylpropanoate (48.75 g, 250
mmol) in dry MeCN (125 ml). After stirring at 80 ꢀC for 24
h, the reaction mixture was partitioned between ether and
water. The yellow organic layer was washed consecutively with
hydrochloric acid (2 M), brine, pH 7 buffer solution (1 M),
then dried and evaporated to a yellow oil which was left under
vacuum (1 mmHg) at 50 ꢀC overnight to remove excess ethyl 2-
bromo-2-methylpropanoate. The crude diethyl ester of 8 thus
produced (15.9 g, 94%) was pure enough to be used in the next
step. dH(200 MHz, CDCl3) 1.25 (6 H, t, J 7), 1.56 (12 H, s),
4.23 (4 H, q, J 7), 6.38 (1 H, t, J 2.5), 6.48 (2 H, dd, J 8,
2.5), 7.07 (1 H, t, J 8); m/z (FAB) 338 (Mþ). Aqueous NaOH
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 4 2 5 – 4 3 3
429

View Article Online
(1 M; 50 ml) was added to a solution of the above diethyl ester
(7 g, 20.7 mmol) in EtOH (50 ml) and stirred at 60 ꢀC over-
night. The volatiles were removed and the residue was parti-
tioned between EtOAc (50 ml) and aqueous HCl (2 M; 50
ml). The organic layer was dried and evaporated to an oil that
crystallised slowly on standing affording 8 (5 g, 86%), mp 94–
96 ꢀC (Found: C, 59.72; H, 6.47. C14H18O6 requires C, 59.57;
H, 6.43); dH(400 MHz, d6-acetone) 1.57 (12 H, s), 6.52 (1 H,
t, J 2.3), 6.57 (2 H, dd, J 8.2, 2.3), 7.12 (1 H, t, J 8.2);
dC(100 MHz, d6-acetone) 26.05, 80.09, 111.82, 114.08,
130.32, 157.82, 175.73; m/z (FAB) 282.2 (Mþ), 305.2
(M þ Naþ).
2-(3-(1-Carboxy-1-methyl(ethylsulfanyl))-phenylsulfanyl)-2-
methylpropanoic acid (11)
A mixture of benzene-1,3-dithiol (159 mg, 1.12 mmol), ethyl 2-
bromo-2-methylpropanoate (585 mg, 3 mmol), Cs2CO3 (1.0 g,
3 mmol) in dry MeCN (5 ml) was stirred at RT for 12 h. After
partitioning between diethyl ether and water the yellow
organic phase was washed twice with water, dried, and evapo-
rated to an oil. The dimethyl ester of 11 thus produced was
pure enough for use in subsequent reactions. dH(300 MHz,
CDCl3) 1.24 (6 H, t, J 7), 1.49 (12 H, s), 4.12 (4 H, q, J 7),
7.28 (1 H, m), 7.48 (2 H, dd, J 7.5, 1.8), 7.63 (1 H, t, J 1.8).
The above crude product was refluxed in a mixture of MeOH
(12 ml) and aqueous NaOH (1 M; 5 ml) for 5 h. Most of the
EtOH was evaporated under reduced pressure, the residue
acidified with aqueous HCl and then partitioned between
EtOAc and brine. The organic phase was washed with brine,
dried, and evaporated to an oil which crystallized on standing
to give 11 (0.310 g, 88%), mp 129–131 ꢀC (Found: C, 53.4; H,
5.7. C14H18O4S2 requires C, 53.48; H, 5.77); dH(200 MHz, d6-
acetone) 1.46 (12 H, s), 7.37 (1 H, m), 7.56 (2 H, m), 7.70 (1 H,
t, J 1.8); m/z (FAB) 314.1 (Mþ); dC(75 MHz, d6-acetone)
25.49, 50.76, 129.11, 132.64, 137.56, 144.51, 174.48.
2-(3-(1-Carboxy-1-methylethoxy)-2,5-di-tert-butylphenoxy)-2-
methylpropanoic acid (9)
tert-Butyl bromide (7.67 g, 56 mmol) was added to a stirred
solution of the diethyl ester of 8 (2.1 g, 6.2 mmol, prepared
as described above), Na2CO3 (5.95 g, 56 mmol) and silica gel
(7 g) in dry tetrachloromethane (50 ml). After stirring at
70 ꢀC for 24 h, the reaction mixture was filtered and the silica
washed thoroughly with diethyl ether. The washings were com-
bined with the filtrate and the solvent evaporated to give a pale
yellow oil, which was purified by column chromatography (0
to 10% EtOAc in hexane) to give the diethyl ester of 9 as a
crystalline solid (1.89 g, 68%) which was used without further
purification; (Found: C, 69.4; H, 9.5. C26H42O6 requires C,
69.30; H, 9.39); dH(200 MHz, CDCl3) 1.22 (6 H, t, J 7), 1.36
(18 H, s), 1.61 (12 H, s), 4.23 (4 H, q, J 7), 5.74 (1 H, s),
7.17 (1 H, s); m/z (FAB) 450.3 (Mþ). Aqueous NaOH (1 M;
5 ml) was added to a solution of the above diethyl ester (0.9
g, 2 mmol) in EtOH (15 ml) and stirred at 60 ꢀC overnight.
The volatiles were removed and the residue was partitioned
between EtOAc and aqueous HCl (2 M). The organic layer
was dried and evaporated to give 9 as white crystalline solid
(0.72 g, 91%), mp 194–196 ꢀC (Found: C, 67.0; H, 8.7.
C22H34O6 requires C, 66.98; H, 8.69); dH(400 MHz, d6-acet-
one) 1.33 (18 H, s), 1.66 (12 H, s), 6.11 (1 H, s), 7.21 (1 H,
s); dC(100 MHz, d6-acetone) 26.42, 30.97, 35.37, 79.09,
106.04, 125.89, 130.82, 153.21, 175.73; m/z (FAB) 394.3
(Mþ), 417.2 (M þ Naþ).
2-(3-(1-Carboxy-1-methylethoxy)-5-dodecyloxyphenoxy)-2-
methylpropanoic acid (12)
A mixture of 1,3,5-trihydroxybenzene (4.0 g, 32 mmol), 1-bro-
mododecane (2.49 g, 10 mmol) and K2CO3 (1.5 g, 11 mmol) in
DMF (10 ml) was stirred at RT for 8 h, then at 75 ꢀC for 4 h.
The resulting reddish slurry was partitioned between diethyl
ether and aqueous HCl (1 M). The organic phase was washed
twice with water, dried, and evaporated to a yellow solid.
Chromatography (30% hexane in EtOAc) provided 1,3-dihy-
droxy-5-dodecyloxybenzene as beige flakes (0.985 g, 33%) pure
enough for use in subsequent reactions. dH(200 MHz, d6-acetone)
0.87 (3 H, br t), 1.1–1.5 (18 H, br m), 1.6–1.8 (2 H, m),
3.86 (2 H, t, J 6.6), 5.95 (3 H, m), 8.14 (2 H, s, OH); m/z
(FAB) 295.2 (MHþ). A mixture of 1,3-dihydroxy-5-dodecyloxy-
benzene (0.35 g, 1.2 mmol), ethyl 2-bromo-2-methylpropano-
ate (2.0 g, 10 mmol), and Cs2CO3 (3.34 g, 10 mmol) in MeCN
(3 ml) was stirred at 50 ꢀC for 12 h. The reaction mixture was
partitioned between diethyl ether and aqueous HCl (1 M). The
organic phase was washed twice with water, dried, and evapo-
rated to a yellow solid. Chromatography (10% EtOAc in hex-
ane) provided the diethyl ester of 12 as a colourless oil (0.410 g,
69%); dH(200 MHz, CDCl3) 0.88 (3 H, br t), 1.1–1.5 (24 H, br
m), 1.56 (12 H, s), 1.6–1.8 (2 H, m), 3.83 (2 H, t, J 6.6), 4.23 (4
H, q, J 7), 5.92 (1 H, t, J 2), 6.08 (2H, d, J 2). The above diethyl
ester was then refluxed in a mixture of EtOH (15 ml) and aqu-
eous NaOH (4 ml; 1 M) for 12 h. Most of the EtOH was evapo-
rated under reduced pressure, and the residue partitioned
between EtOAc and brine. The organic phase was washed with
brine, dried, and evaporated to an oil which crystallized on
standing to give 12 (0.363 g, 94% from diester), mp 71–73 ꢀC
(Found: C, 66.7; H, 9.1. C26H42O7 requires C, 66.93; H,
9.07); dH(300 MHz, d6-acetone) 0.88 (3 H, br t), 1.1–1.5 (18
H, br m), 1.56 (12 H, s), 1.73 (2 H, m), 3.89 (2 H, t, J 6.6),
6.08 (1 H, t, J 2), 6.14 (2H, d, J 2); m/z (FAB) 467.3
(MHþ), 489.2 (M þ Naþ); dC(75 MHz, d6-acetone) 13.56,
22.55, 24.91, 26.09, 31.87, 67.91, 79.02, 99.86, 102.76, 157.32,
159.86, 174.68.
2-(3-(1-Carboxy-1-methylethoxy)-2,5-dibromophenoxy)-2-
methylpropanoic acid (10)
Bromine (77 ml, 1.47 mmol) was added dropwise to a stirred
solution of the diethyl ester of 8 (0.25 g, 0.73 mmol, prepared
as described above) in DCM (2 ml) at ice-bath temperature.
After 10 min the reaction mixture was diluted with DCM,
washed with aqueous sodium bisulfite (1 M), and the organic
layer dried and evaporated to a light yellow oil (0.358 g) which
was pure enough to use in the next step. dH(200 MHz, CDCl3)
1.29 (4 H, t, J 7), 1.58 (12 H, s), 4.25 (4 H, q, J 7), 6.56 (1 H, s),
7.69 (1 H, s). The crude dibromoester thus obtained was
heated in a mixture of EtOH (6 ml) and aqueous NaOH (1
M; 3 ml) at 80 ꢀC for 12 h. Most of the EtOH was evaporated
under reduced pressure, the residue acidified with aqueous HCl
and partitioned between EtOAc and brine. The organic phase
was washed with brine, dried, and evaporated to an oil which
crystallized on standing to give 10 (0.311 g, 95%), mp 174–
176 ꢀC (Found: C, 38.3; H, 3.7. C14H16Br2O6 requires C,
38.21; H, 3.66); dH(300 MHz, d6-acetone) 1.61 (12 H, s), 6.85
(1 H, s), 7.74 (1 H, s); dC(75 MHz, d6-acetone) 24.81, 81.03,
107.81, 110.77, 135.62, 152.55, 173.89. m/z (FAB) 439.9
(Mþ). Bromoacid 10 was also prepared in 97% yield by direct
bromination of diacid 8, following the same bromination
method as above (a beige precipitate formed shortly after
addition of bromine).
2-(20-(1-Carboxy-1-methylethoxy)-[1,10]-binaphthalenyl-2-
yloxy)-2-methylpropanoic acid (13)
Caesium carbonate (5.85 g, 18.0 mmol) was added to a stirred
solution of (ꢂ)-1,10-binaphthalene-2,20-diol (1.0 g, 3.5 mmol)
and ethyl 2-bromo-2-methylpropanoate (3.5 g, 17.9 mmol) in
MeCN (8.0 ml). After heating at 80 ꢀC for 2 hours, the mixture
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
430
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 4 2 5 – 4 3 3

View Article Online
was partitioned between EtOAc and aqueous HCl. The pale
yellow organic layer was washed twice with aqueous HCl, then
dried and evaporated to afford an oily yellow solid (0.99 g).
Aqueous NaOH (2 M, 5 ml) was added to a solution of the yel-
low solid in hot EtOH (15 ml) and stirred at 50 ꢀC overnight.
The volatiles were removed and the solution was acidified with
aqueous HCl causing a white solid to precipitate. This aqueous
suspension was heated to near boiling (solid dissolved) and
then left to cool. Filtration gave 13 as a white solid (0.57 g,
36%), mp 80–82 ꢀC (Found: C, 73.3; H, 5.7. C28H26O6 requires
C, 73.35; H, 5.72); dH(400 MHz, d6-acetone) 1.07 (6 H, s), 1.11
(6 H, s), 6.91 (1 H, d, J 8), 7.04 (1 H, t, J 8), 7.15 (1 H, t, J 8),
7.20 (1 H, d, J 9), 7.71 (1 H, d, J 8), 7.76 (1 H, d, J 9); dC(100
MHz, d6-acetone) 25.2, 26.09, 80.42, 120.36, 124.61, 125.20,
126.91, 127.25, 129.70, 131.02, 135.31, 152.54, 176.43; m/z
(FAB) 458.3 (Mþ); m/z (ES, MeOH) 457.1640 [(M–H)ꢁ
requires 457.1651].
yields in (Table 1). The physical data and crystal structures
of dirhodium complexes 8a, 8b, 9a, 9b, 10a, 10b have been
described previously.18
(Ph3CCO2)nRh2(OAc)4ꢁn (3a–d)
A solution of triphenylacetic acid (306 mg, 1.06 mmol) and
Rh2(OAc)4 (235 mg, 0.53 mmol) in N,N-dimethylaniline (8
ml) was stirred at 125 ꢀC for 6.5 h. The cooled reaction mixture
was diluted with chloroform (40 ml) and washed with aqueous
HCl (2 M) three times. The green solution was dried and eva-
porated to a green solid. Chromatography (0 to 25% MeCN in
DCM) afforded four green compounds 3d (63 mg, 11%) eluting
in neat DCM, 3c (45 mg, 9%) eluting in 5% MeCN, 3b (270
mg, 57%) eluting in 10 to 15% MeCN and 3a (23 mg, 7%) elut-
ing in 20 to 25% MeCN. Analytical data for (Ph3CCO2)3R-
h2(OAc) 3a: Found: C, 65.70; H, 4.20. C62H48O8Rh2 requires
C, 66.08; H, 4.23; dH (200 MHz, CDCl3) 2.16 (3 H, s), 6.80–
7.20 (45 H, Ar); m/z (FAB) 1126.3 (Mþ). Analytical data for
cis-(Ph3CCO2)2Rh2(OAc)2 3b: Found: C, 58.8; H, 4.1.
C44H36O8Rh2 requires C, 58.81; H, 4.04; dH(200 MHz, CDCl3)
1.87 (6 H, s), 6.92–6.95 (10 H, Ar), 7.10–7.20 (20 H, Ar); m/z
(FAB) 898.1 (Mþ). Analytical data for trans-(Ph3CCO2)2R-
h2(OAc)23c: Found: C, 59.0; H, 4.1. C44H36O8Rh2 requires
C, 58.81; H, 4.04; dH(200 MHz, CDCl3) 1.60 (6 H, s), 6.80–
6.85 (10 H, Ar), 6.90–7.20 (20 H, Ar); m/z (FAB) 897.9
(Mþ). Analytical data for (Ph3CCO2)Rh2(OAc)3 3d: Found:
C, 46.8; H, 3.6. C26H24O8Rh2 requires C, 46.59; H, 3.61;
dH(200 MHz, CDCl3) 1.90 (9 H, s), 6.90–7.20 (15 H, ar);
m/z (FAB) 670.0 (Mþ).
3-(3-(2-Carboxyethyl)phenyl)propanoic acid (14)
This diacid was prepared using a minor modification of the lit-
erature method.27 A mixture of malonic acid (8.5 g, 82 mmol)
and benzene-1,3-dicarboxaldehyde (3.62 g, 27 mmol) in pyri-
dine (10 ml) was stirred at 50 ꢀC for 2 h and 100 ꢀC for 2.5
h. After cooling the mixture was poured into aqueous sulfuric
acid (50 ml, 1 M) and the white precipitate filtered and dried to
give 3-(3-(2-carboxyvinyl)phenyl)prop-2-enoic acid as a white
powder (5.7 g, 97%). A sample of this crude diacid (436 mg,
2 mmol) was stirred in a mixture of aqueous NaOH (0.5 ml,
6.25 M), water (3 ml) and palladium on carbon (30 mg, 10%
w/w) under hydrogen (65 psi) for 24 h. The catalyst was fil-
tered off, and the reaction mixture acidified with conc. aqueous
HCl, producing a white precipitate. Acetic acid (0.7 ml) was
added and the mixture stirred at 80 ꢀC for 20 min to dissolve
the precipitate. On cooling a precipitate was formed again
which was filtered off, washing with water, to give diacid 14
as white crystals (130 mg, 29%) mp 134–136 ꢀC (lit 148–
150.5 ꢀC) (Found: C, 64.6; H, 6.3. C12H14O4 requires C,
64.85; H, 6.35); dH(400 MHz, d6-acetone) 2.62 (4 H, t, J 8),
2.91 (4 H, t, J 8), 7.10 (2 H, d, J 7), 7.20 (2 H, m); dC(100
MHz, d6-acetone) 31.91, 36.30, 127.28, 129.61, 129.64,
142.46, 174.28; m/z (CI, NH3) 240 (M þ NH4þ).
(O2C(CH2)7CO2)Rh2(OAc)2 (4a) and (O2C(CH2)7CO2)2Rh2
(4b)
Rh2(OAc)4 (200 mg, 0.45 mmol) and nonanedioic (75 mg, 0.40
mmol) in N,N-dimethylaniline (45 ml) was stirred at 140 ꢀC.
After 48 h, the reaction mixture was diluted with chloroform
(30 ml) and MeCN (1 ml) and washed with hydrochloric acid
(2 M) three times. The solution was dried and evaporated to
give a green solid. Chromatography (10 to 35% MeCN in
DCM) afforded two green compounds 4b (15 mg, 6%) eluting
at 15 to 20% MeCN and 4a (35 mg, 15%) eluting at 25 to 35%
MeCN. Analytical data for 4a: dH(200 MHz, 5% v/v d3-
MeCN in CDCl3) 0.99 (2 H, m), 1.20 (4 H, m), 1.62 (4 H,
m), 1.96 (6 H, s), 2.21 (4 H, m); m/z (FAB) 509.8 (Mþ); m/z
(ES, MeOH) 532.9144 [(M þ MeOH þ Na)þ requires
532.9166]. 4a.(4-tert-butylpyridine)2 : dH(200 MHz, CDCl3)
1.12 (2 H, m), 1.25 (4 H, m), 1.40 (18 H, s), 1.60 (4 H, m),
1.90 (6 H, s), 2.12 (4 H, m), 7.65 (4 H, brd), 9.20 (4 H, brs);
Analytical data for 4b: dH(200 MHz, 40% v/v d3-MeCN in
CDCl3) 0.95 (4 H, m), 1.25 (8 H, m), 1.60 (8 H, m), 2.10 (8
H, m); m/z (FAB) 578 (Mþ); m/z (ES, MeOH) 600.9800
[(M þ Na)þ requires 600.9729]. 4b.(4-tert-butylpyridine)2 :
dH(200 MHz, CDCl3) 1.15 (4 H, m), 1.29 (8 H, m), 1.40 (18
H, s), 1.61 (8 H, m), 1.90 (6 H, s), 2.15 (8 H, m), 7.65 (4 H,
brd), 9.20 (4 H, brs);
Analytical data for 5a. Found: C, 31.1; H, 2.6.
C14H14O10Rh2 requires C, 30.68; H, 2.58; dH(200 MHz, d6-
acetone) 1.77 (6 H, s), 4.44 (4 H, s), 5.44 (1 H, t, J 2.4), 6.40
(2 H, dd, J 8, 2.4), 7.07 (1 H, t, J 8); m/z (FAB) 548.3 (Mþ).
Analytical data for 6a: Found: C, 40.1; H, 4.6.
C22H30O10Rh2 requires C, 40.02; H, 4.58; dH(200 MHz, 5%
v/v d3-MeCN in CDCl3) 1.31 (18 H, s), 1.94 (6 H, s), 4.44 (4
H, s), 5.47 (1 H, s), 7.12 (1 H, s); m/z (FAB) 659.9971 [Mþ
required 659.9949].
(3-Carbomethoxyphenyl)ethanoic acid (15)
Acetyl chloride (0.5 ml) was added to a solution of 3-hydroxy-
phenylethanoic acid (1.0 g, 6.6 mmol) in dry MeOH (50 ml)
and stirred at 50 ꢀC for 1.5 h. Volatiles were evaporated and
a solution of the residue in DMF (5 ml) stirred with methyl
bromoacetate (1.6 g, 10.5 mmol) and K2CO3 (1.6 g, 11.6
mmol) at 80 ꢀC for 3 h. The reaction mixture was diluted with
EtOAc, washed three times with water followed by brine, and
the organic layer dried and evaporated to a pale yellow oil.
Aqueous NaOH (15 ml, 2.5 M) was added to a solution of
the preceding oil in MeOH (50 ml) and stirred at 50 ꢀC for 1
h. The volatiles were evaporated and the residue partitioned
between EtOAc and aqueous HCl. The organic layer was dried
and evaporated to give diacid 15 as a white solid (0.78 g, 57%),
mp 130–132 ꢀC (Found: C, 57.2; H, 4.8. C10H10O5 requires C,
57.15; H, 4.80); dH(400 MHz, d6-acetone) 3.62 (2 H, s), 4.72 (2
H, s), 6.86 (1 H, d, J 7.5), 6.96 (2 H, m), 7.26 (1 H, t, J 7.5);
dC(100 MHz, d6-acetone) 41.64, 65.82, 113.94, 117.26,
123.63, 130.55, 137.76, 159.55, 170.64, 173.08; m/z (CI,
NH3) 228 (M þ NH4þ).
Dirhodium complexes. The preparation of dirhodium com-
plexes is illustrated below for 3a–d. Monochelates and bische-
lates, green solids unless otherwise indicated, were prepared
using the same method but at reactant concentrations 45
mM, following the reactions by tlc (reaction times 0.5 to 4 h,
Analytical data for 7a: Found: C, 23.9; H, 1.7.
C18H20Br2O10Rh2 requires C, 23.82; H, 1.71; dH(200 MHz,
5% v/v d3-MeCN in CDCl3) 1.94 (6 H, s), 4.57 (4 H, s), 5.60
(1 H, s), 7.64 (1 H, s); m/z (FAB) 705.7 (Mþ).
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 4 2 5 – 4 3 3
431

View Article Online
Analytical data for 8c: Found: C, 39.9; H, 4.1.
C45H54O22Rh4 requires C, 40.31; H, 3.97; dH(200 MHz, 5%
v/v d4-MeOH in CDCl3) 1.29 (12 H, s),1.36 (12 H, s), 1.39
(12 H, s), 1.91 (6 H, s), 6.05 (5 H, m), 6.50 (4 H, brd), 6.74
(1 H, t, J 8), 7.07 (2 H, t, J 8); m/z (FAB) 1369.9 (Mþ).
Analytical data for 11a.(4-tert-butylpyridine)2 : To keep this
complex in solution 4-tert-butylpyridine was added during
workup and to the chromatography solvent (0 to 2% MeOH
in DCM). Red solid, mp 190 ꢀC (dec); dH(200 MHz, CDCl3)
1.31 (12 H, s), 1.39 (18 H, s), 1.87 (6 H, s), 7.1–7.4 (4 H, m),
7.56 (4 H, brd), 9.0 (4 H, brd); m/z (FAB) 770.9928 [M.ButPyþ
requires 770.9949].
Analytical data for 12a: Found: C, 45.9; H, 5.9.
C30H46O11Rh2 requires C, 45.70; H, 5.88; dH(200 MHz,
CDCl3) 0.87 (3 H, brt), 1.25 (18 H, brs), 1.42 (12 H, s), 1.70
(2 H, m), 2.01 (6 H, s), 3.81 (2 H, brt), 5.58 (1 H, brt), 6.1
(2H, brd); m/z (FAB) 789.1 (Mþ).
Analytical data for 12b: Green glass, mp 72–75 ꢀC (Found:
C, 55.4; H, 7.1. C52H80O14Rh2 requires C, 55.03; H, 7.10);
dH(200 MHz, 5% v/v d4-MeOH in CDCl3) 0.88 (6 H, brt),
1.26 (36 H, s), 1.33 (24 H, br s), 1.72 (4 H, m), 3.83 (4 H,
brt), 5.57 (2 H, d, J 2), 6.08 (2H, d, J 2). m/z (FAB) 1153.3
(Mþ).
Analytical data for 14a. Found: C, 35.3; H, 3.4.
C16H18O8Rh2 requires C, 35.32; H, 3.33; dH(200 MHz, 5%
v/v CD3CN in CDCl3) 1.95 (6H, s), 2.31–2.42 (4H, m),
2.74–2.85 (4H, m), 6.87 (2 H, dd, J 7.5, 1.6), 7.00 (1H, brt),
7.11 (1 H, t, J 7.5); m/z (FAB) 543.9 (Mþ)
direction started from an approximately equimolar mixture
of Rh2(OAc)4 and diacid 8 (ca 20 mM each) with various
amounts of acetic acid, letting the system come to equilibrium
and integrating as above. Because of the difference in magni-
tude of K1 and K2 a negligible amount of bischelate 8b is pre-
sent under these conditions. The 8a, 8b equilibrium constant
K2 was measured starting from 8b with the addition of small
amounts of acetic acid (5–20 ml).
Modelling. MacSpartan Pro 1.0.3 (Wavefunction Inc.) was
used for conformational analysis of diacid ligands, modelled
as dimethyl esters (Fig. 2). Rotatable bonds were varied in
30ꢀ, 60ꢀ or 120ꢀ increments using a simulated annealing Monto
Carlo-type search and the MMFF94 force field (the default for
molecules of this size in MacSpartan), with a further cycle of
energy minimisation to ensure convergence. For each ligand
several searches starting from different starting geometries
were performed until no new conformations less than 10 kJ
molꢁ1 above the global minimum were produced. In some
cases systematic searches were also run. Conformer popula-
tions were calculated using population ¼ s.exp(ꢁDE/RT ),
where DE ¼ (energy of conformation) ꢁ (energy of global
minimum), and s is a number depending on the symmetry of
the conformation: for conformations of C1 symmetry, s ¼ 4,
for C2 or CS , s ¼ 2, for C2V , s ¼ 1 (Table S1y). Cerius2
(Accelrys Inc) was used for empirical estimates of the differ-
ence in energy between diacid ligands (as dimethyl esters)
and monochelate complexes. Low energy conformations were
obtained from multiple runs of systematic and Boltzmann
jump searches, using the UNIVERSAL 1.02 force field with
no charges and no constraints.
Analytical data for 14b. Found: C, 44.5; H, 3.9.
C24H24O8Rh2 requires C, 44.61; H, 3.74; dH(300 MHz, 1%
v/v d4-MeOH in CDCl3) 2.30–2.41 (8H, m), 2.70–2.81 (8H,
m), 6.81–6.92 (6H, m), 7.08–7.17 (2H, m); m/z (FAB) 646.1
(Mþ)
PC Spartan Pro 1.0.7 was used for pBP/DN* (dft) calcula-
tions (Table S2y). The lowest energy conformation of a ligand
(dimethyl ester) according to dft was obtained by reminimising
the ten lowest energy conformations previously found by mole-
cular mechanics (MMFF94). The lowest energy conformation
of the corresponding dirhodium complex was obtained by first
minimising a chelate of arbitrary conformation using dft. The
Rh2(OAc)2 part of the molecule was then fixed, conformations
of the chelate ring were generated using MMFF94, and then
reminimised using dft with no constraints. Known crystal
structures of monochelates have axial ligands (usually pyri-
dines), but these could not be included in dft calculations
due to the size of the system. However calculated bond lengths
and angles for the dirhodium core were quite close to experi-
mental values,18 despite the absence of axial ligands. For 8a,
Chelate stability constants. The method is illustrated for
reaction of monochelate 8a with acetic acid. A solution of 8a
(4.8 mg) and d4-acetic acid (50 ml) in d2-tetrachloroethane
(600 ml) was heated in a septum capped NMR tube at 100 ꢀC
1
until the H NMR spectrum no longer changed (ꢄ24 h). The
geminal methyl resonances for 8a at 1.33 ppm and free diacid
8 at 1.54 ppm were integrated to give a ratio I8a:I8 ¼ 1.28. The
equilibrium
constant
was
then
calculated
using
K1 ¼ I8a(I8a þ I8)[AcOH]2/(I82[8a]o) ¼ 392 M, where [8a]o is
the initial concentration of monochelate. The experiment was
repeated using different concentrations of 8a and acetic acid,
adjusting the amounts so that the peaks to be integrated were
not too different in intensity. Experiments in the forward
˚
calculated: Rh–Rh ¼ 2.403 A, and Rh–O (average) ¼ 2.085
Table 3 Crystal data and collection parameters
5
6
7 dimethyl ester
8
10
11
Empirical formula
Formula weight
T/K
C
10H10O6
C18H26O6
338.39
213(2)
C12H12Br2O6
412.04
213(2)
C28H36O12
564.57
293(2)
C14H16Br2O6
440.09
293(2)
C14H18O4S2
314.40
213(2)
226.18
213(2)
Triclinic
Crystal system
Space group
Monoclinic
C2/c
Triclinic
¯
P1
Monoclinic
C2/c
Monoclinic
P21n
Monoclinic
P21c
¯
P1
˚
a/A
6.7348(2)
7.0518(2)
11.256(2)
75.30(3)
73.98(3)
71.04(3)
477.94(2)
2
19.461(3)
5.7894(7)
32.412(6)
90
8.694(2)
9.240(3)
9.899(3)
110.10(3)
98.04(3)
98.76(3)
722.0(3)
2
18.591(2)
10.996(2)
21.767(3)
90
11.2340(1)
7.4699(1)
20.1840(2)
90
10.5028(2)
12.5327(1)
12.1758(2)
90
˚
b/A
˚
c/A
a
b
g
94.33(2)
90
100.15(1)
90
96.541(1)
90
92.081(2)
90
3
˚
V/A
3641.3(1)
8
0.092
4379.8(1)
6
0.101
1682.8(4)
4
4.841
1601.6(4)
4
0.341
Z
m(Mo-Ka)/mmꢁ1
Data/parameters
R1 (I > 2s(I), wR2
R1,R2 (all data)
0.132
5.634
1399/147
0.0311, 0.0763
0.0390, 0.0790
2273/219
0.0535, 0.1245
0.0750, 0.1393
2155/186
0.0580, 0.1413
0.0715, 0.1481
2855/257
0.0697, 0.2529
0.1291, 0.3050
2491/205
0.0281, 0.0598
0.0424, 0.0615
2453/185
0.0351, 0.0894
0.0513, 0.0959
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
432
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 4 2 5 – 4 3 3

View Article Online
˚
53; (c) J. K. Brea, B. W. Smucker, R. A. Walton and K. R.
Dunbar, Chem. Commun., 2001, 2562; (d ) K.-B. Shiu, H.-C.
Lee, G.-H. Lee and Y. Wang, Organometallics, 2002, 21, 4013;
(e) S. M. Kuang, P. E. Fanwick and R. A. Walton, Inorg. Chem.,
2002, 41, 1036; ( f ) S. L. Schiavo, G. Pocsfalvi, S. Serroni, P.
Cadiano and P. Piraino, Eur. J. Inorg. Chem., 2000, 1371.
J. F. Bickley, R. P. Bonar-Law, C. Femoni, E. J. MacLean,
A. Steiner and S. J. Teat, J. Chem. Soc., Dalton Trans., 2000,
4025.
F. A. Cotton, T. R. Felthouse and J. L. Thompson, Inorg. Chim.
Acta, 1984, 81, 193.
D. F. Taber, R. P. Meagley, J. P. Louey and A. L. Rheingold,
Inorg. Chim. Acta, 1995, 239, 25.
(ꢂ0.006) A. From the crystal structure of 8a.(4-tert-butylpyri-
˚
dine)2 : Rh–Rh ¼ 2.408 A, and Rh–O (average) ¼ 2.047
18
˚
(ꢂ0.008) A.
Crystallography. Suitable crystals of diacids were obtained
from the following solvents: 5, acetone; 6, diethyl ether; 7
dimethyl ester, MeOH; 8, acetone, 10, acetone–hexane; 11,
DCM. None of the crystals contained solvent in the lattice
(Table 3). Crystallographic data were collected on a STOE-
IPDS image plate diffractometer using graphite monochro-
4
5
6
7
8
9
˚
mated MoKa radiation (l ¼ 0.71073 A).z Structure solution
by Direct Methods and structure refinement by full-matrix
least-squares was based on all data using F2 (G. M. Scheldrick,
SHELX-97, Programs for Crystal Structure Solution and
Refinement, University of Go¨ttingen, Germany, 1997). All
non-hydrogen atoms were refined anisotropically, with the
exception of disordered atoms, which were refined isotropi-
cally. Hydrogen positions were placed geometrically. The fol-
lowing were disordered. Diacid 7: two disordered methyl
groups. Disordered atomic positions were split and refined
without restraints and one occupancy parameter per disor-
dered group. Diacid 8: side groups disordered. Disordered
atomic positions were split and refined using similar distance
and one occupancy parameter per disordered group. Diacid
10: double bonded oxygen disordered. Disordered atomic
position was split and refined without restraints.
J. M. Andersen, N. Karodia, D. J. Miller, D. Stones and D. Gani,
Tetrahedron Lett., 1998, 39, 7815.
J. F. Gallagher, G. Ferguson and A. J. McAlees, Acta Cryst. Sect.
C, 1997, 53, 576.
H. M. L. Davies and S. A. Panaro, Tetraheron Lett., 1999, 5287.
10 J. Seitz and G. Maas, Chem. Commun., 2002, 338.
11 J. C. Lee, J. Y. Yuk and S. H. Cho, Synth. Commun., 1995, 25,
1367.
12 Y. Kamitori, M. Hojo, R. Masuda, T. Izumi and S. Tsukamoto,
J. Org. Chem., 1984, 49, 4161.
13 M. Kohn, J. Am. Chem. Soc., 1952, 74, 1601.
14 F. H. Allen, W. D. S. Motherwell, P. R. Raithby, G. P. Shields
and R. Taylor, New J. Chem., 1999, 25.
15 I. Dance, New J. Chem., 2003, 22.
16 T. A. Halgren, J. Comput. Chem., 1999, 20, 730. Conformational
analysis of diacids 5–7 with two other force fields (MMX, MM3)
gave a similar trend in preferred conformations.
17 Handbook of Chemistry and Physics, 74th edn., ed. D. R. Lide,
CRC Press, 1993.
18 R. P. Bonar-Law, T. D. McGrath, N. Singh, J. F. Bickley, C.
Femoni and A. Steiner, J. Chem. Soc., Dalton Trans., 2000, 4343.
19 (a) M. H. Chisholm, J. C. Huffman and S. S. Iyer, J. Chem. Soc.,
Dalton Trans., 2000, 1483; (b) S. Garcia-Granda, P. Lahuerta, J.
Latorre, M. Martinez, E. Peris, M. Sanau and M. Angeles Ubeda,
J. Chem. Soc., Dalton Trans., 1994, 539.
Acknowledgements
We thank the EPSRC for a studentship (NS), and Johnson
Matthey PLC for generous loans of rhodium.
20 (a) L. Mandolini, Adv. Phys. Org. Chem., 1986, 22, 1; (b) A. J.
Kirby, Adv. Phys. Org. Chem., 1980, 17, 183.
21 Taking statistical factors into account EM1 ¼ K1/16K2 and
References
EM2 ¼ K2/4K2.
1
2
3
(a) B. J. Holliday and C. A. Mirkin, Angew. Chem. Int. Ed., 2001,
40, 2022; (b) M. Fujita, K. Umemoto, M. Yoshizawa, N. Fujita,
T. Kusukawa and K. Biradha, J. Chem. Soc., Chem. Commun.,
2001, 509.
(a) F. A. Cotton and R. A. Walton, Multiple Bonds between Metal
Atoms, Clarendon Press, Oxford, 1993, p. 431; (b) F. H. Jardine
and P. S. Sheridan, in Comprehensive Coordination Chemistry,
Pergamon, 1987, vol. 4, p. 933.
22 Z. D. Liu and R. C. Hider, Coord. Chem. Rev., 2002, 232, 151.
23 P. G. Sammes and D. J. Weller, Synthesis, 1995, 1205.
24 T. C. Bruice and F. C. Lightstone, Acc. Chem. Res., 1999, 32, 127.
25 Ligands 8–10 can also chelate dicopper carboxylates: J. Bickley,
R. P. Bonar-Law, M. A. Borrero Martinez and A. Steiner, Inorg.
Chim. Acta, in press.
26 For diacid ligands that support biomimetic iron clusters see:
J. Du Bois, T. J. Mizoguchi and S. J. Lippard, Coord. Chem.
Rev., 2000, 200, 443.
(a) F. A. Cotton, C. Lin and C. A. Murillo, Acc. Chem. Res.,
2001, 34, 759; (b) M. H. Chisholm, Acc. Chem. Res., 2000, 33,
27 C. W. Schimelpfenig, J. Org. Chem., 1975, 40, 1493.
y CCDC reference numbers 220194–220199. See http://www.rsc.org/
suppdata/nj/b3/b310008a/ for crystallographic data in .cif or other
electronic format.
T h i s j o u r n a l i s Q T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 4
N e w . J . C h e m . , 2 0 0 4 , 2 8 , 4 2 5 – 4 3 3
433