Direct in situ synthesis of cationic N-heterocyclic carbene iridium and rhodium complexes from neat ionic liquid: Application in catalytic dehydrogenation of cyclooctadiene

Article abstract of DOI:10.1016/j.jorganchem.2008.04.017

A direct synthetic route to cationic N-heterocyclic carbene (NHC) complexes of rhodium and iridium from neat dialkyl-imidazolium ionic liquids (ILs) has been found. The method uses complexes bearing basic anionic ligands, [M(COD)(PPh₃)X], X = OEt, MeCO₂, which react with the inactivated imidazolium cation in the absence of external bases yielding one M-NHC moiety and the free protonated base. This new one-pot synthesis leaving pure, catalytically active IL solutions is faster, cleaner and more efficient than traditional syntheses of such NHC complexes. The observed reactivity also gives insight into NHC incorporation of rhodium and iridium catalyzed reactions performed in common dialkyl-imidazolium ILs. The complexes synthesised in this manner are compared with their bis-phosphine analogues in terms of activity for catalytic dehydrogenation of 1,5-cyclooctadiene and 1,3-cyclooctadiene in neat [BMIM][NTf₂] as solvent. Even at high temperature, no ligand exchange reaction is observed with [(COD)M(PPh₃)₂] [NTf₂] catalysts. As expected, the yields of all the reactions were low, iridium was much more active in C-H activation than rhodium and the NHC ligands were more stable than triphenylphosphine. For all catalysts, the isomerisation of 1,5-cyclooctadiene is the major reaction. However, the phosphine-NHC complex of iridium seems to be more selective for dehydrogenation than its bis-phosphine counterpart, which is more active in transfer-hydrogenation and less stable under the applied conditions. Different reaction conditions were tried in order to optimise selectivity for dehydrogenation over isomerisation and transfer-hydrogenation. Surprisingly, with 1,3-cyclooctadiene as substrate selectivity for dehydrogenation is much higher than with 1,5-cyclooctadiene for all catalysts.

Full text of DOI:10.1016/j.jorganchem.2008.04.017

Journal of Organometallic Chemistry 693 (2008) 2407–2414
Contents lists available at ScienceDirect
Journal of Organometallic Chemistry
journal homepage: www.elsevier.com/locate/jorganchem
Direct in situ synthesis of cationic N-heterocyclic carbene iridium and rhodium
complexes from neat ionic liquid: Application in catalytic dehydrogenation
of cyclooctadiene
Ulrich Hintermair a,b, Thibaut Gutel b, Alexandra M.Z. Slawin a, David J. Cole-Hamilton a,
,
*
b
Catherine C. Santini b, , Yves Chauvin
*
a EaStChem, School of Chemistry, University of St. Andrews, St. Andrews, Fife KY16 9ST, Scotland, United Kingdom
b University of Lyon, Institut de Chimie de Lyon, C2P2, Equipe Chimie Organométallique de Surface, UMR 5265 CNRS-CPE Lyon, Bat. 308,
43 Bd. du 11 Novembre 1918, 69616 Villeurbanne Cedex, France
a r t i c l e i n f o
a b s t r a c t
Article history:
A direct synthetic route to cationic N-heterocyclic carbene (NHC) complexes of rhodium and iridium from
neat dialkyl-imidazolium ionic liquids (ILs) has been found. The method uses complexes bearing basic
anionic ligands, [M(COD)(PPh3)X], X = OEt, MeCO2, which react with the inactivated imidazolium cation
in the absence of external bases yielding one M-NHC moiety and the free protonated base. This new one-
pot synthesis leaving pure, catalytically active IL solutions is faster, cleaner and more efficient than
traditional syntheses of such NHC complexes. The observed reactivity also gives insight into NHC incor-
poration of rhodium and iridium catalyzed reactions performed in common dialkyl-imidazolium ILs.
The complexes synthesised in this manner are compared with their bis-phosphine analogues in terms of
activity for catalytic dehydrogenation of 1,5-cyclooctadiene and 1,3-cyclooctadiene in neat [BMIM][NTf2]
as solvent. Even at high temperature, no ligand exchange reaction is observed with [(COD)M(PPh3)2]
[NTf2] catalysts. As expected, the yields of all the reactions were low, iridium was much more active in
C–H activation than rhodium and the NHC ligands were more stable than triphenylphosphine. For all cat-
alysts, the isomerisation of 1,5-cyclooctadiene is the major reaction. However, the phosphine-NHC com-
plex of iridium seems to be more selective for dehydrogenation than its bis-phosphine counterpart,
which is more active in transfer-hydrogenation and less stable under the applied conditions. Different
reaction conditions were tried in order to optimise selectivity for dehydrogenation over isomerisation
and transfer-hydrogenation. Surprisingly, with 1,3-cyclooctadiene as substrate selectivity for dehydroge-
nation is much higher than with 1,5-cyclooctadiene for all catalysts.
Received 14 March 2008
Received in revised form 10 April 2008
Accepted 10 April 2008
Available online 16 April 2008
Keywords:
Ionic liquid
N-Heterocyclic carbene
Cationic iridium complexes
Cationic rhodium complexes
Catalytic dehydrogenation
Ó 2008 Elsevier B.V. All rights reserved.
1. Introduction
holds the highest potential for dehydrogenation activity [8]. Since
N-heterocyclic carbene complexes of rhodium or iridium are less
Selective C–H activation of saturated hydrocarbon units via
dehydrogenation is still one of the ‘‘dream reactions” of catalysis
[1,2]. The major problem of the dehydrogenation reaction is the
very unfavourable thermodynamics [3,4]. Strategies to overcome
this difficulty include: (i) the use of hydrogen acceptors [5], (ii)
photochemical stimulation via UV–Vis irradiation [6], (iii) continu-
ous removal of dihydrogen from the catalytic phase [7].
In recent years, much work on dehydrogenation has been car-
ried out by the organometallic community; however, there is still
no satisfactory solution. A number of transition metals have been
successfully applied in C–H activation [4], but iridium intrinsically
thermally sensitive than their alkylphosphine analogues [9,10],
their use in catalytic hydrogenation has recently been tested
[11]. [(COD)Ir(PCy3)(NHC)]+ was found to be even superior in terms
of activity as well as stability to Crabtree’s catalyst [(COD)Ir(PCy3)-
(pyridine)]+.
Much of the efforts so far concentrated on the design of more
reactive catalysts, but little attention was paid to the particular dif-
ficulty of reaction engineering. Besides the problem of product sep-
aration and catalyst recovery (the ‘‘classical” drawbacks of
homogeneous catalysis) [12], the lack of suitable solvents is still
a severe limitation for C–H activation catalysts.
Ionic liquids (ILs) have been promoted as a novel class of sol-
vents for synthesis and catalysis [13–15], since they combine a
variety of advantages such as environmental friendliness (virtually
no vapour pressure, easiness of recovery and recycling), physical
* Corresponding authors.
E-mail addresses: djc@st-andrews.ac.uk (D.J. Cole-Hamilton), santini@cpe.fr (C.C.
Santini).
0022-328X/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2008.04.017

2408
U. Hintermair et al. / Journal of Organometallic Chemistry 693 (2008) 2407–2414
robustness within a large temperature range and high solubilisa-
tion potential. In general, a significant advantage of using ionic liq-
uids in homogeneous catalysis is the stabilization of
organometallic catalysts by their weak coordinating effect. How-
ever, gases like dihydrogen show very low solubility in these media
even at high pressure [16,17], and high temperature [18].
strate [27]. The activity, selectivity and stability of the classical
bis-phosphine complexes will be compared with their mono-NHC
analogues.
For the latter, a direct synthetic route from neat dialkyl-imi-
dazolium ILs was found. The idea of the new approach to
[(COD)M(PR3)NHC]+ complexes depicted in Fig. 2 was to try the
reaction of complexes bearing basic ligands with uncoordinated
imidazolium cations in the absence external bases or halides, but
in the presence of 1 equiv. of the respective phosphine instead.
The formation of the carbene ligand would be forced to occur on
the metal and thus avoid the formation of a free carbene. A similar
approach has been developed in the case of palladium, however it
was not an in situ procedure [28]. ILs, being practically non volatile,
open up the possibility of working in solution under reduced pres-
sures even at high temperatures [29]. This property was exploited
in evaporating the protonated base in order to drive the reaction to
completion.
It has been shown that metallation of imidazolium salts can oc-
cur under mild conditions with certain transition metals to gener-
ate N-heterocyclic carbene (NHC) ligands [19–21]. These findings
suggest that ionic liquids may act as reagents and not only as sol-
vents in transition metal catalyzed reactions. Therefore, in several
catalytic systems the difference of the activity in ionic liquids,
compared with organic solvents, has been explained by a possible
in situ formation of a NHC ligand. This ligand is coordinated to the
metal centre and consequently the nature and the activity of the
catalytic system are not only changed by solvent effects [22,23].
In particular, this formation of NHC ligands probably also plays
an important role in the stabilization of iridium nanoparticles syn-
thesised by reduction of [Ir(COD)Cl]2 in imidazolium-ILs [24].
As the cationic complexes of Osborn and Schrock shown in
Fig. 1, are active hydrogenation catalysts in organic as well in ionic
liquid solvents [25,26], they have been tested in catalytic dehydro-
genation of cyclooctadiene (COD) to cyclooctatriene (cotr) and
cyclooctatetraene (cot) in dialkyl-imidazolium ionic liquids.
The facile coordination of the substrate to the catalyst should be
an advantage for the dehydrogenation reaction: activation of pre-
coordinated ligands in the proximity of the metal centre has a
striking entropic advantage over oxidative addition of free sub-
2. Results and discussion
The target structure of rhodium and iridium complexes that is
supposed to react with the imidazolium cation to give the desired
[(COD)M(PR3)(NHC)]+ complexes is [(COD)M(base)(PR3)]. The
bases used for instance were acetate (MeCO2), 2,4-pentanedioate
(acac) and ethoxide (OEt); chloride was also included for the sake
of comparison. Known synthetic pathways to these complexes are
summarised in Fig. 2 [30–32].
While some of the complexes of general formula [(COD)M(ba-
se)(PR3)] were difficult to obtain in acceptable yields from organic
solvents, the respective base bridged dimers could easily be syn-
thesised in high purity for all cases investigated. Therefore, in all
following activation experiments the target complex was formed
in situ by mixing 1 equiv. of the respective dimer with 2 equiv. of
PPh3 in a 40-fold excess of [BMIM][NTf2] as solvent. From the many
possible weakly coordinating anions, the bistriflamide (NTf2À) was
chosen for the preparation of the ionic liquid used. These BTA-
based ILs exhibit low viscosities, high thermal stabilities and offer
synthetic advantages such as hydrophobicity of the resulting salts.
PPh3
M = RhI, IrI
M
diene = cyclooctadiene (COD)
L = PPh3, NHC
L
Fig. 1. Complexes used for catalytic isomerisation/dehydrogenating of
cyclooctadienes.
MCl3x(H2O)
EtOH / H2O
1,5-cod
Cl
Cl
B
+ KB
- KCl
M
M
M
M
B
2PR3
2 PR3
Cl
B
+ KB
- KCl
2
M
2
M
PR3
PR3
Fig. 2. Synthetic routes to [(COD)M(B)(PR3)] for M = RhI, IrI and B = anionic base (30–32).

U. Hintermair et al. / Journal of Organometallic Chemistry 693 (2008) 2407–2414
2409
This is an important feature for organometallic synthesis and catal-
ysis in ILs as the IL can be extracted with deionised water to be-
come essentially free from halides and other ionic impurities.
Moreover, satisfying dryness of BTA-based ILs is easy to achieve
by heating in vacuum.
Activation was tried by heating and under UV–Vis irradiation at
30 °C, both reactions being carried out in vacuo in order irreversibly
to remove any protonated base. In no case did photolysis lead to
the formation of NHC complexes according to the NMR spectra of
the mixtures.
The [(COD)M(PPh3)Cl] complexes showed different reactivity
for rhodium and iridium. [(COD)Ir(PPh3)Cl] was stable in solution
up to 120 °C in vacuo, higher temperatures resulted in decomposi-
tion to a metallic mirror. [(COD)Rh(PPh3)Cl] stays unchanged at
moderate temperatures (up to 80 °C), but above 100 °C an orange
solid precipitated from the IL. NMR analysis revealed that it was
pure [(COD)RhCl]2, and no traces of a NHC complex could be found
in solution.
The corresponding [(COD)M(PPh3)(acac)] complexes did not be-
have in such a clean manner. Although indication of NHC forma-
tion was seen, the reaction did not go to completion upon
further or longer heating. Instead, decomposition was observed.
In all other successful cases the respective NHC complex remained
stable upon heating to at least 140 °C in vacuum.
Fig. 3. X-ray crystal structure of [(COD)Ir(PPh3)NHC][NTf2] (hydrogen atoms are
omitted for clarity).
anes finally resulted in crystallisation of some pure material in the
case of iridium. The X-ray structure confirmed the successful for-
mation of [(COD)Ir(PPh3)NHC][NTf2], Fig. 3.
The metal–carbene bond length of 2.053(6) Å suggests a single
rather than a double bond, consistent with there being little
p-backbonding from the metal. The Ir–C bond lengths to the COD
ligand are not statistically variant while the C–N bond lengths
within the imidazolium ring support a high degree of delocalisa-
tion. In general, the data are consistent with similar complexes
reported in the literature [33] (see Tables 2 and 3).
In the case of the acetate and ethoxide precursors, upon heating,
the initial slurry suddenly became a clear homogeneous solution
with characteristic colouration (deep orange in the case of rho-
dium, purple for iridium) above a certain temperature. Bubbling
was also observed, indicating liberation of an evaporable species.
In the 1H NMR spectrum of these solutions new signals appeared
close to those from the protons in 4/5 position, and of the N-methyl
and the N-methylene group (approx. 0.5 ppm upfield) of the imi-
dazolium salt, indicating the presence of an altered imidazolium
species. Interestingly, the 31P NMR spectra show only one single
compound with a chemical shift very similar to [(COD)M(PPh3)2]+.
In the case of rhodium, a 1JPRh coupling of 156 Hz proves phosphine
coordination to the metal. In the 13C NMR spectra, all peaks from
PPh3 and COD can be observed as well as the carbon signals of
the imidazolium ion. Finally, one new quaternary signal at
176.23 ppm appears as a doublet of doublet (1JCRh = 47.3 Hz,
2JCP = 16 Hz, NCN) and at 173.39 ppm as a doublet (2JCP = 9.8 Hz),
for the rhodium and iridium complexes, respectively. 1H NMR con-
firms the presence of the coordinated diene; all these data are indi-
cate that [(COD)M(PPh3)NHC]+ complexes were formed
quantitatively.
The phosphine–NHC complexes were also analysed by electro-
spray ionisation (ESI) mass spectrometry [34]. All negative spectra
(ESIÀ) only showed one peak at m/z = 280 arising from the bistrifla-
mide counter anion and in (ESI+) [(COD)Ir(PPh3)NHC]+ gave rise to
signals at m/z = 699.24 and 701.24 in the ratio expected for
(
191Ir/193Ir = 37: 63), see supporting information.
An important observation, which led us to investigate the activ-
ity of these complexes in dehydrogenation reactions, was that, in
the observed fragmentation of all [(COD)M(PPh3)L]+ complexes
(L = NHC or PPh3), there is a progressive loss of 2 mass units (up
to a total of 4) on increasing the cone voltage. This could be inter-
preted as dehydrogenation after the phosphine ligand has dissoci-
ated from the complex. Most probably, COD is dehydrogenated to
cyclooctatriene (cotr) and cyclooctatetraene (cot), affording
[M(cot)L]+ (L = NHC or PPh3). This observation has also been made
before for some tungsten and molybdenum carbonyl complexes of
COD and cotr [35]. However, loss of hydride from other ligands,
such as phosphines or NHC is also possible to some extent. Both li-
gands are known to undergo metallation reactions [36].
Table 1 sums up the thermal activation results.
For the sake of complete characterisation, isolation of the suc-
cessfully formed NHC complexes from the IL was tried. Interest-
ingly, it was found that the in situ formed NHC complexes
showed a very high affinity to the parent IL and could not be a
100% separated from it. Extraction or crystallisation with the help
of organic solvents failed. Even column chromatography over silica
gel only yielded a concentrated oil of the respective complex in
some IL. Dilution of the oil with chloroform and layering with hex-
Table 2
Selected bonds lengths (Å) and angles (°) with estimated standard deviations
Table 1
Ir(1)–C(2)
Ir(1)–C(11)
Ir(1)–C(12)
Ir(1)–C(15)
Ir(1)–C(16)
C(2)–N(1)
C(2)–N(3)
N(3)–C(4)
N(1)–C(5)
C(4)–C(5)
Ir(1)–P(1)
2.053(6)
2.198(6)
2.215(6)
2.200(6)
2.222(14) Å
1.345(8)
1.354(8)
1.378(8)
1.384(8)
1.345(9)
2.3177(16)
C(11)–Ir(1)–C(2)–N(3)
C(12)–Ir(1)–C(2)–N(3)
C(11)–Ir(1)–C(2)–N(1)
C(12)–Ir(1)–C(2)–N(1)
C(15)–Ir(1)–C(2)–N(3)
C(16)–Ir(1)–C(2)–N(3)
P(1)–Ir(1)–C(2)–N(3)
À3.5(11)
À148.3(6)
170.8(6)
26.0(10)
À102.8(6)
À66.2(6)
96.5(5)
Activation results of [BMIM][NTf2] for heating with the respective complex in vacuo
Precursor
Reactivity with [BMIM][NTf2]
[(COD)Rh (PPh3)Cl]
[(COD)Rh(PPh3)OAc]
[(COD)Ir(PPh3)OAc]
[(COD)Rh(PPh3)acac]
[(COD)Ir(PPh3)acac]
[(COD)Rh(PPh3)OEt]
[(COD)Ir(PPh3)OEt]
Dimerisation to [Rh(cod)Cl]2 above 100 °C
Quant. NHC-formation above 80 °C
Quant. NHC-formation above 100 °C
Some NHC-formation above 50 °C
Some NHC-formation above 60 °C
Quant. NHC-formation above 50 °C
Quant. NHC-formation above 60 °C

2410
U. Hintermair et al. / Journal of Organometallic Chemistry 693 (2008) 2407–2414
Table 3
Considering that iridium is generally more active in C–H activa-
tion and forms stronger metal–ligand bonds than rhodium, the re-
sults summarised in Table 1 suggest that pre-dissociation of one of
the ligands is needed before the imidazolium is attacked. As no
external base is present in the medium, the formation of the NHC li-
gand must occur on the metal, and no free carbene should be present
during the reaction. The d8 metal centre must be involved directly in
the observed C–H activation by forming an NHC ligand from an
uncoordinated, common dialkyl–imidazolium IL under pH neutral
conditions. Possible routes of formation are described in Fig. 4.
As disclosed in the beginning, pathway (F) over hexacoordinat-
ed 18 electron complexes should give Ir–NHC complexes more
readily than with rhodium. Considering the observed phosphine
dissociation in the case of [(COD)Rh(PPh3)Cl] and high electrostatic
repulsion of the imidazolium ion to a cationic species, pathway (E)
seems to be the most probable. Therefore, the formation of a
neutral 14 electron complex bearing a basic ligand would be
responsible for the formation of a NHC Ligand from a neat dial-
kyl–imidazolium ionic liquid. In this case, the ligand strength of
both the base and the phosphine determine the equilibrium (A)–
(C)–(D). Reaction (B) might also take place, but is not thought to
lead to NHC formation; the pKa values of all bases are too low to
deprotonate the imidazolium ion. The suggested reaction pathway
also explains the reported reaction time of 3 days for the tradi-
tional synthesis of such complexes from [(COD)M(OEt)]2 with imi-
Crystal data and structure refinement for [(COD)Ir(PPh3)NHC][NTf2]
Formula
C36H41F6IrN3O4PS2
981.01
Formula weight
Crystal size (mm)
Crystal system
Space group
Unit cell dimensions
a (Å)
b (Å)
c (Å)
a (°)
b (°)
0.10 Â 0.03 Â 0.01
Triclinic
ꢀ
P1
10.1476(16)
12.644(2)
15.919(3)
85.442(15)
76.826(13)
75.861(12)
1927.9(6)
2
1.690
12357
6825 [0.0926]
R1 = 0.0551, wR2 = 0.1366
R1 = 0.0569, wR2 = 0.1391
1.015
3.684
976
3.96–25.38
c (°)
Volume (Å3)
Z
Dcalc (Mg/m3)
Reflections collected
Independent reflections [Rint
Final R indices [I > 2r(I)]
R indices (all data)
Goodness-of-fit on F2
]
Absorption coefficient (mmÀ1
F(000)
h Range for data collection (°)
Maximim and minimum transmission
Refinement method
)
1.0000 and 0.8656
Full-matrix least-squares on F2
6825/0/445
2.597 and À2.094
Data/restraints/parameters
Largest difference in peak and hole (e ÅÀ3
)
B
L
+ L
(A)
M
M
M
1/2
B
B
(D)
(C)
(B)
+
+
L
L
B-
M
B
M
(F )
(G)
(E)
BMIM NTf2
HB
N
C
N
M
NTf2-
nBu
L
Fig. 4. Potential pathways of NHC formation.

U. Hintermair et al. / Journal of Organometallic Chemistry 693 (2008) 2407–2414
2411
dazolium halides in organic solvents [10,37]. In this case, halide
bridged dimers would reform and thus only with a large excess
of ethoxide the reaction proceeds slowly. When one equivalent of
a neutral donor ligand reversibly creates the 14 electron fragment,
the reaction proceeds within minutes under vacuum.
At both 140 and 160 °C, the reaction with [(COD)Ir(PPh3)2][NTf2]
went to full conversion. [(COD)Ir(PPh3)NHC][NTf2] demonstrated
much lower activity than the bis-phosphine complex. The main
product using either catalyst was 1,3-COD. Although it seems as
if more hydrogenation than dehydrogenation occurred, the num-
ber of dehydrogenated double bonds has to be taken into account.
As no H2 was added to the system, transfer-hydrogenation must
have taken place. In direct comparison, the NHC complex is 20
times less active in terms of transfer-hydrogenation compared to
the bis-phosphine catalyst. The 31P NMR spectrum shows that
the NHC catalyst was more stable, over 80% remained in the final
solution, compared with 70% for the bis-phosphine complex, Table
4, Fig. 5.
All 1,5-COD had been consumed after 2 h at 160 °C with
[Ir(COD)(PPh3)2][NTf2]. Interestingly, less hydrogenation than dehy-
drogenation was found for the first time, indicating some true dehy-
drogenation with liberation of H2. Approximately four turnovers
of transfer-hydrogenation and one turnover of dehydrogenation
occurred in 2 h. It seems as if the reaction with the NHC complex
is less rapid than with the bis-phosphine catalyst, the, selectivity
to dehydrogenation over transfer-hydrogenation looks better:
roughly 1.5 turnovers of dehydrogenation and 0.2 turnovers of
transfer-hydrogenation took place within 2 h. The catalyst is suffi-
ciently stable: 72% of the initial complex remained (31P NMR), and
the 1H and 13C NMR still showed the presence of the NHC ligand.
Comparing the results at 160 °C with the reaction at 140 °C, it
seems as if refluxing the substrate on the IL phase containing the
catalysts does indeed help H2 to evolve from the reaction medium.
Interestingly, the NHC complex of iridium looks more active in
dehydrogenation than in transfer-hydrogenation: the opposite
behaviour to its bis-phosphine analogue. This, on the other hand,
suggests that the stronger donor facilitates dehydrogenation.
In order to test whether dehydrogenation only occurs during
isomerisation, the main isomerisation product 1,3-cyclooctadiene
was employed as substrate under identical conditions. After 2 h,
the iridium bis-phosphine catalyst yielded 6.8% conversion with
an interesting product distribution. Transfer-hydrogenation seems
to be suppressed. The isomerisation back to 1,4- and even 1,5-COD
was observed with a yield of 2%. One new isomerisation product,
bicyclo[3.3.0]oct-2-ene, was found with 4% yield, which remark-
ably requires a p–r C–C bond rearrangement. Dehydrogenation oc-
curred with 1.1% conversion corresponding to 1.5 turnovers within
2 h. The iridium NHC catalyst behaved similarly with slightly lower
activity, Table 5.
3. Catalysis
[(COD)Rh(PPh3)2][NTf2] was found to be poorly soluble in pure,
dry [BMIM][NTf2]. Slow heating gave an orange homogeneous
solution above 80 °C which contained only 34% of the initial com-
plex and numerous side products. Therefore, the Rh–NHC com-
plexes were barely active in dehydrogenation catalysis, and
therefore only the case of iridium is discussed in the following.
The iridium bis-phosphine complex was even less soluble, but
more stable, however. At 100 °C, the complex reversibly dissolved
in neat [BMIM][NTf2] to give a deep purple solution showing un-
changed [(COD)Ir(PPh3)2][NTf2] which slowly recrystallised from
the IL upon cooling to room temperature. The in situ generated
NHC complexes were entirely soluble in the respective IL, and ther-
mally more stable.
Solubility of the substrate determined at 25 °C by 1H NMR
shows 15 mol% of 1,5-COD in saturated [BMIM][NTf2]. This amount
is sufficient for homogeneous catalysis to occur, but still allows bi-
phasic reaction engineering for the sake of product separation and
catalyst recovery.
All values for conversion, selectivity and turnover numbers
were calculated from GC–MS analysis (see Table 4, Fig. 5). All
C8H12 compounds other than the starting material were considered
as isomerisation products, C8H12Àx as dehydrogenation products
and C8H12+x as hydrogenation products. Apart from additional sol-
vent, no other compounds were found in the organic fraction after
catalysis. The main isomerisation occurred from 1,5- through 1,4-
to 1,3-cyclooctadiene, although traces of saturated tricyclic and
other C8H12 hydrocarbons could also be observed (named as
‘‘other” in the following).
Liquid–liquid biphasic catalysis was attempted for dehydroge-
nation reactions at 140 and 160 °C. A clear, colourless organic
phase remained separated on top of the highly coloured IL phase
containing the catalyst. The mixture was stirred vigorously for sev-
eral hours under ambient pressure of nitrogen while a bubbler al-
lowed potential gas evolution. At the end a clear, colourless organic
phase could be recovered, suggesting the highly coloured catalyst
is essentially insoluble in the substrate.
Table 4
Product distribution from 1,5-COD after 5 h at 140° C or 2 h at 160° C in the presence of [(COD)Ir(PPh3)2][NTf2] or [(COD)Ir(PPh3)NHC][NTf2], with a molar ratio COD/Cat 100:1
[(COD)Ir(PPh3)2][NTf2]
[(COD)Ir(PPh3)NHC][NTf2]
[(COD)Ir(PPh3)2][NTf2]
[(COD)Ir(PPh3)NHC][NTf2]
140 °C (5 h)
1,5-COD
140 °C (5 h)
1,5-COD
160 °C (2 h)
1,5-COD
160 °C (2 h)
1,5-COD
Conv. 100%
Conv. 84.5%
Conv. 100%
Conv. 96%
Isomerisation
Conv. 83.9%
Conv. 79.4%
Conv. 91.9%
Conv. 94.4%
Selectivity
3.5% 1,4-COD
71.3% 1,3-COD
9.1% other
67.0% 1,4-COD
26.4% 1,3-COD
6.0% other
10.9% 1,4-COD
62.1% 1,3-COD
18.9% other
11.5% 1,4-COD
86.9% 1,3-COD
0% other
Dehydrogenation
Conv. 5.3%
Conv. 0.3%
Conv. 4.3%
Conv. 1.4%
Selectivity
1% cotr
0.3% cotr
1.3% cotr
1.2% cotr
3.6% xylenes
0.7% styrene or cot
0% xylenes
0% styrene or cot
2.4% xylenes
0.6% styrene or cot
0% xylenes
0.2% styrene or cot
Hydrogenation
Conv. 10.8%
Conv. 0.3%
Conv. 3.8%
Conv. 0.2%

2412
U. Hintermair et al. / Journal of Organometallic Chemistry 693 (2008) 2407–2414
COA
Hydrogenation
Isomerisation
(Major route)
Isomerisation
1,5-COD
1,4-COD
1,3-COD
Dehydrogenation
COTr
Isomerisation
Dehydrogenation
COT
Fig. 5. Possible reaction sequence of 1,5-COD in the presence of [(COD)Ir(PPh3)2][NTf2] or [(COD)Ir(PPh3)NHC][NTf2].
per hour looks reasonable when compared with other systems
Table 5
for catalytic dehydrogenation.
Product distribution from [(COD)Ir(PPh3)2][NTf2] (upper table) and [(COD)Ir(PPh3)-
NHC][NTf2] (lower table) with 1,3-COD after 2 h at 160 °C with a molar ratio COD/Cat
100:1
When one phosphine ligand is replaced by a more highly donat-
ing NHC ligand in the iridium complex, lower isomerisation as well
as lower transfer-hydrogenation activity is observed. However, the
catalyst is more stable and more selective; less side products are
formed. Importantly, the NHC ligand is stable throughout the reac-
tion; no wingtip-metallation or transalkylation was observed. The
IL seems to prevent irreversible loss of the ligand, probably by act-
ing as a reservoir. The fact that typically 70–80% can be recovered
in the NMR regardless of the reaction time might be due to oxida-
tion originating from the transfer of the catalyst solution, and does
not necessarily reflect degradation through catalysis.
Catalyst
Product
Conversion (%)
[(COD)Ir(PPh3)2][NTf2]
Isomerisation
Dehydrogenation
Hydrogenation
6.8
1.1
0
[(COD)Ir(PPh3)](NHC)[NTf2]]
Isomerisation
Dehydrogenation
Hydrogenation
6.4
0.5
0
When the main isomerisation product, 1,3-cyclooctadiene, is
used as substrate, no transfer-hydrogenation occurs but, besides
some isomerisation, dehydrogenation occurs exclusively. Although
overallactivityislow, selectivityto dehydrogenationismuchhigher.
4. Conclusion
The direct route to NHC complexes of rhodium and iridium
from dialkyl-imidazolium ILs allows easy access to the correspond-
ing complexes. This one-pot synthesis does not necessitate any
workup or purification and gives catalytically active IL solutions.
The application of the complexes [(COD)Ir(PPh3)2][NTf2] and
[(COD)Ir(PPh3)NHC][NTf2] in dehydrogenation catalysis was suc-
cessfully demonstrated. Activity in the range of a few turnovers
5. Experimental
All experiments were carried out under nitrogen atmosphere
using Schlenk techniques. The inert gas was dried and deoxygen-

U. Hintermair et al. / Journal of Organometallic Chemistry 693 (2008) 2407–2414
2413
ated through a chromium(II)/silica glass column. Air stable sub-
stances (solid complexes and IL) were weighed quickly in air. All
fittings were equipped with Teflon QuickFitTM seals; grease was
not used in contact with chemicals.
Solvents were dried with suitable drying agents, distilled under
nitrogen and stored over vacuum heated molecular sieves 4 Å. All
chemicals were purchased from commercial sources including Al-
drich, Acros, Strem and ABCR. Unless otherwise specified, the re-
agents were used as received. Gases were supplied from BOC
Gases, UK.
Melting points were measured on a Gallenkamp Melting Point
Apparatus using glass capillaries. Elemental analyses were con-
ducted by the Analytical Service of the School of Chemistry, Uni-
versity of St Andrews on a Carlo Erba CHNS analyser.
Mass spectrometry was performed on a Micromass LCT high
performance orthogonal acceleration reflection TOF mass spec-
trometer using a Z-flow atmospheric pressure ionisation source
for electrosprayed samples. Measurements were conducted by
the Mass Spectrometry Facility of the Biomolecular Science Depart-
ment of the University of St Andrews.
Gas chromatographic analyses were carried out on a Hewlett
Packard 5890 series gas chromatograph equipped with both flame
ionisation detector (GC-FID) for quantitative analysis and a mass
selective detector (GC–MS) for qualitative analysis. A fused silica
column Supelco MDN-35 (35% phenyl- and 65% methyl-polysilox-
ane, 30 m length, 0.25 mm inner diameter, 0.25 lm film thickness)
was used with a flow rate of 1.0 ml/min. The temperature was held
at 40 °C for 8 min and then raised to 200 °C at a heating rate of
20 °C/min.
NMR spectra were recorded on a Bruker Avance 300 or a Bruker
Avance II 400 spectrometer at 300 K unless otherwise stated. 1H
and 13C spectra were referred internally to deuterated solvents;
chemical shifts are relative to Me4Si. 19F spectra were calibrated
with the help of an internal Bruker reference program using proton
solvent signals. 31P spectra used external 85% H3PO4 as standard.
X-ray crystallography was performed at 93 K using a Rigaku
MM007 rotating anode generator/confocal optics and a Mercury
CCD detector Mo Ka radiation (k = 0.71071 Å). A full hemisphere
of data were collected, corrected for Lorentz polarisation and
(2 ml) and the solid was collected on a frit and washed with hex-
anes. Drying in vacuum yielded of an orange microcrystalline solid
(327 mg, 0.322 mmol, 79%). The product is soluble in acetone,
ether and alcohols but insoluble in hexanes and water.
M.p. = 144 °C (dec.).
1H NMR (CDCl3): d = 7.44–7.27 (m, 30H, phenyl), 4.55 (bs, 4H,
3
CODvinyl), 2.53 (m, 4H, CODallyl inner), 2.24 (qua, JHH = 8.9 Hz, 4H,
CODallyl outer). 13C NMR (CDCl3): d = 134.04, 130.51, 128.72 and
128.67 (m, phenyl), 99.18 (s, CODvinyl), 30.64 (s, CODallyl). 19F
NMR (CDCl3): d = À78.55 (s, N(SO2CF3)2). 31P NMR (CDCl3):
1
d = 26.31 (d, JPRh = 145 Hz, PPh3).
ESI+MS: 735.30 ([Rh(COD)(PPh3)]+); 473.16 [Rh(COD)(PPh3)]+;
471.16 [Rh(C8H10)(PPh3)]+; 469.13 [Rh(C8H8)(PPh3)]+. Anal. Calc.
for C46H42F6NO4 P2RhS2 requires C, 54.4; H, 4.1; N, 1.4. Found: C,
54.4; H, 3.6; N, 1.4%.
5.2. g4-1,5-Cyclooctadienbis(triphenylphosphine) iridium(I)
bis{(trifluoromethyl)sulfonyl}amide
The complex was synthesised following the same procedure as
for the rhodium analogue using. di-l-chloro-bis(g4-1,5-cycloocta-
diene)diiridium(I) (100 mg, 0.149 mmol), lithium bis{(trifluoro-
methyl)sulfonyl}amide (100 mg, 0.348 mmol) and triphenylphos-
phine (400 mg, 1.53 mmol) to give the pure product as a purple
powder (270 mg, 0.244 mmol, 84%). The solubility behaviour is
similar to that of the rhodium complex. M.p. = 164 °C (dec.).
1H NMR (CDCl3): d = 7.45–7.27 (m, 30H, phenyl), 4.19 (bs, 4H,
3
CODvinyl), 2.35 (m, 4H, CODallyl inner), 1.97 (qua, JHH = 8.6 Hz, 4H,
CODallyl outer).13C NMR (CDCl3): d = 134.24, 131.37, 129.45 and
128.70 (m, phenyl), 87.17 (s, CODvinyl), 31.09 (s, CODallyl).19F NMR
(CDCl3): d = -78.59 (s, N(SO2CF3)2).31P NMR (CDCl3): d = 17.41 (s,
PPh3).
ESI+MS.
(
193Ir/191Ir = 63: 37): 825.35, 823.36 ([Ir(C8H10)-
(PPh3)2]+); 561.21, 559.23 [Ir(C8H10)(PPh3)]+; 557.19, 555.20
[Ir(C8H8)(PPh3)]+. Found: C 47.6, H 3.3, N 1.5; C46H42F6IrNO4 P2S2
requires C 50.0, H 3.8, N 1.3.
5.3. Synthesis of N-Heterocyclic carbene complexes from neat [BMIM]
[NTf2]
absorption. Structure solution and refinement used SHELXTL
.
Synthesis of 1-nbutyl-3-methylimidazolium bis{(trifluoro-
Pure [BMIM][NTf2] was dried and degassed by heating to 100 °C
in vacuo over night prior to use. Neat [BMIM][NTf2] (1 g,
2.385 mmol) was mixed with the respective dimer (30–40 mg,
0.0596 mmol) to give a BMIM+ to metal ratio of 20, equivalent to
a metal concentration of 0.085 mol dmÀ3. Triphenylphosphine
(31 mg, 0.119 mmol) was added and the mixture was stirred in va-
cuo for several hours while being heated stepwise. As soon as a
homogeneous solution was obtained, it was placed in an NMR tube
together with a fused capillary of C6D6. The solution was analysed
by NMR spectroscopy and heated further in vacuum.
methyl)sulfonyl}amide;
di-l-chloro-bis(g4-1,5-cyclooctadiene)-
dirhodium(I), di-l-chloro-bis(g4-1,5-cyclooctadiene)diiridium(I);
di-l-acetato-bis(g4-1,5-cyclooctadiene)dirhodium(I); di-l-aceta-
to-bis(g4-1,5-cyclooctadiene)- diiridium(I); g4-1,5-cyclooctadi-
ene-l-2,4-pentanedionato-rhodium(I): l-2,4-pentanedionato-g4-
1,5-cyclooctadiene-iridium(I);
bis(g4-1,5-cyclooctadiene)di-l-
ethanoatodirhodium(I); bis(g4-1,5-cyclooctadiene)di-l-ethanoato-
diiridium(I); chloro -g4-1,5-cyclooctadienetriphenylphosphinerho-
dium(I) were carried out as described in the literature [30–32,38].
Experiments attempting photochemical activation employed
the same mixture of the respective complex with two equivalents
of phosphine in neat [BMIM][NTf2] which was transferred into an
EPR tube made of quartz. A fused capillary of C6D6 was added
and the solution was exposed to UV irradiation for 8 h at 30 °C
while being in vacuo and analysed by NMR.
5.1. g4-1,5-Cyclooctadienebis(triphenylphosphine)rhodium(I)
bis{(trifluoromethyl)sulfonyl}amide
The
complex,
di-l-chlorobis(g4-1,5-cyclooctadiene)dirho-
dium(I) (100 mg, 0.203 mmol) was dissolved in dichloromethane
(2 ml) to give an orange solution. A solution of lithium bis{(trifluo-
romethyl)sulfonyl}amide (120 mg, 0.418 mmol) in deionised water
(2 ml) was added. Under vigorous stirring triphenylphosphine
(400 mg, 1.53 mmol) was added to the emulsion, and the mixture
was stirred for 2 h at room temperature. After phase separation,
the water layer was removed with a syringe and the organic phase
washed with water (2 ml  3). The deep orange solution was re-
duced to half the volume with a stream of nitrogen and isopropa-
nol (2 ml) was slowly added until an orange solid began to
precipitate. Crystallisation was completed by addition of hexanes
5.4. 1-nButyl-3-methylimidazolin-2-ylidene-g4-1, 5-
cyclooctadienetriphenylphosphinerhodium(I)
bis{(trifluoromethyl)sulfonyl}amide
1H NMR: d = 7.48–7.35 (m, 15H, phenyl), 7.07 and 6.99 (bs, 2H,
NCHCHN), 4.85 (bd, 2H, CODvinyl trans PPh3), 4.2* (m, CODvinyl 9i
NHC), 4.1* (m, NCH2R), 3.59 (s, 3H, NCH3), 2.66–2.12 (m, 4H,
3
CODallyl), 1.90** (qui, JHH = 7.4 Hz, 2H, NCH2CH2CH2CH3), 1.38**

2414
U. Hintermair et al. / Journal of Organometallic Chemistry 693 (2008) 2407–2414
(se, 3JHH = 7.4 Hz, 2H, NCH2CH2CH2CH3), 0.95** (t, 3JHH = 7.4 Hz, 3H,
NCH2CH2CH2CH3).
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif. Supplementary data associated with this article
can be found, in the online version, at doi:10.1016/
j.jorganchem.2008.04.017.
1
2
13C NMR: d = 176.23 (dd, JCRh = 47.3 Hz, JCP = 16 Hz, NCN),
133.67 (d, 2JCP = 11.4 Hz, ortho-C phenyl), 131.54 (s, para-C phenyl),
131.03 (d, 1JCP = 23.9 Hz, a-C phenyl), 129.05 (d, 3JCP = 8.7 Hz, meta-C
1
phenyl), 124.38 and 121.97 (s, NCHCHN), 120.13** (qua, JCF
=
References
321 Hz, N(SO2CF3)2), 97.06 (vt, 1JCRh = 9.4 Hz, 2JCP = 7.9 Hz, CODvinyl
trans PPh3), 96.04 (vt, 1J CRh = 9.9 Hz, JCP = 8.9 Hz, CODvinyl trans
2
[1] A.J. Bard, G.M. Whitesides, R.N. Zare, F.W. McLafferty, Acc. Chem. Res. 28
(1995) 21.
[2] J.A. Labinger, J.E. Bercaw, Nature 417 (2002) 507.
[3] R.H. Crabtree, Dalton Trans. (2001) 2437.
[4] R.H. Crabtree, J. Organometal. Chem. 689 (2004) 4083.
[5] R.H. Crabtree, M.F. Mellea, J.M. Mihelcic, J.M. Quirk, J. Am. Chem. Soc. 104
(1982) 107.
[6] G.B. Shul’pin, G.V. Nizova, A.E. Shilov, Chem. Commun. 671 (1983).
[7] T. Sakakura, C. Sodeyama, Y. Tokunaga, M. Tanaka, Chem. Lett. 263 (1988).
[8] J.K. Perry, G. Ohanessian, W.A. Goddard I, Organometallics 13 (1994) 187.
[9] C. Koecher, W.A. Herrmann, J. Organometal. Chem. 532 (1997) 261.
[10] W.A. Herrmann, M. Elison, J. Fischer, C. Koecher, G.R.J. Artus, Chem. Eur. J. 2
(1996) 772.
PPh3), 94.60 and 94.01 (d, 1JCRh = 7.4 Hz, CODvinyl trans NHC), 50.68
(s, NCH2R), 37.15 (s, NCH3), 31.91 (s, NCH2CH2CH2CH3), 30.73
(d,1JCRh = 38.5 Hz, CODallyl trans PPh3), 29.20 (d, JCRh = 40.4 Hz,
1
CODallyl trans NHC), 19.94 (s, NCH2CH2CH2CH3), 13.14 (s,
NCH2CH2CH2CH3).
31P NMR: d = 25.37 (d, 1JPRh = 156 Hz, PPh3). [* Identified through
C–H correlation spectra, peaks partly or completely overlapped
with strong solvents signals of BMIM NTf2. **Common signals for
both the solvent and the complex.]
ESI+MS:
611.16
([Rh(COD)(PPh3)(NHC)]+);
349.09
[11] L.D. Vazquez-Serrano, B.T. Ownes, J.M. Buriak, Inorg. Chim. Acta 359 (2006)
2786.
[12] D.J. Cole-Hamilton, Science 299 (2003) 1702.
[13] P. Wasserscheid, T. Welton (Eds.), Ionic liquids in Synthesis, Wiley–VCH,
Weinheim, 2007;
([Rh(COD)(NHC)]+); 239.0 [Rh(NHC)]+.
5.5. 1-nButyl-3-methylimidazolin-2-ylidene- g4-1, 5-
cyclooctadienetriphenylphosphineiridium(I)
bis{(trifluoromethyl)sulfonyl}amide
V.I. Parvulescu, C. Hardacre, Chem. Rev. 107 (2007) 2615;
P. Wasserscheid, in: Ionic Liquids in Synthesis, Wiley–VCH, Weinheim, 2003, p.
213.
[14] T. Welton, Coord. Chem. Rev. 248 (21–24) (2004) 2459.
[15] Z. Conrad Zhang, Adv. Catal. 49 (2006) 153.
[16] A.M. Voutchkova, L.N. Appelhans, A.R. Chianese, R.H. Crabtree, J. Am. Chem.
Soc. 127 (2005) 17624.
1H NMR: d = 7.48–7.31 (m, 15H, phenyl), 7.07 and 6.97 (bs, 2H,
NCHCHN), 4.46 (bs, 2H, CODvinyl trans PPh3), 4.1* (m, NCH2R), 4.0*
[17] P.J. Dyson, G. Laurenczy, C.A. Ohlin, J. Vallance, T. Welton, Chem. Commun.
2418 (2003).
[18] A. Finotello, J.E. Bara, D. Camper, R.D. Noble, Ind. Eng. Chem. Res ACS ASAP.
(2008).
[19] L. Xu, W. Chen, J. Xiao, Organometallics 19 (2000) 1123.
[20] D.S. McGuinness, K.J. Cavell, B.F. Yates, B.W. Skelton, A.H. White, J. Am. Chem.
Soc 123 (2001) 8317.
(m, CODvinyl trans NHC), 3.47 (s, 3H, NCH3), 2.65–2.05 (m, 4H, CO-
3
D
allyl), 1.89** (qui, JHH = 7.4 Hz, 2H, NCH2CH2CH2CH3), 1.37** (se,
3
3JHH = 7.4 Hz, 2H, NCH2CH2CH2CH3), 0.94** (t, JHH = 7.4 Hz, 3H,
NCH2CH2CH2CH3). Anal. Calc. for C36H41F6IrN3O4PS2 requires: C,
46.7; H, 4.2; N, 4.3. Found: C, 44.2; H, 4.1; N, 4.1%.
13C NMR: d = 173.39 (d, JCP = 9.8 Hz, NCN), 133.85 (d,
2
[21] S. Gru¨ndemann, M. Albrecht, A. Kovacevic, J.W. Faller, R.H. Crabtree, Dalton
2JCP = 10.2 Hz, ortho-C phenyl), 131.23 (s, para-C phenyl), 130.56
(d, 1JCP = 50.4 Hz, a-C phenyl), 129.07 (d, 3JCP = 9.1 Hz, meta-C phe-
Trans. 2163 (2002).
[22] (a) C.J. Mathews, P.J. Smith, T. Welton, A.J.P. White, D.J. Williams,
Organometallics 20 (2001) 3848;
1
nyl), 124.01 and 121.72 (s, NCHCHN), 120.13** (qua, JCF = 321 Hz,
(b) L. Magna, Y. Chauvin, G.P. Niccolai, J.-M. Basset, Organometallics 22 (2003)
4418.
[23] H. Olivier-Bourbigou, V. Lecocq, Stud. Surf. Sci. Catal 145 (2003) 55.
[24] J. Dupont, G.S. Fonseca, A.P. Umpierre, P.F.P. Fichtner, S.R. Teixeira, J. Am.
Chem. Soc. 124 (2002) 4228;
L.S. Ott, M.L. Cline, M. Deetlefs, K.R. Seddon, R.G. Finke, J. Am. Chem. Soc. 127
(2005) 5758;
J.D. Scholten, G. Ebeling, J. Dupont, Dalton Trans. (2007) 5554–5560.
[25] J.A. Osborn, R.R. Schrock, J. Am. Chem. Soc. 93 (1971) 3089.
[26] B. Gilbert, Y. Chauvin, H. Olivier, F. Di Marco-Van Tiggelen, J. Chem. Soc.,
Dalton. Trans. 3867 (1995).
[27] R.H. Crabtree, Chem. Rev. 85 (1985) 245.
[28] N. Marion, P. de Frmont, I.M. Puijk, E.C. Ecarnot, D. Amoroso, A. Bell, S.P. Nolan,
Adv. Synth. Catal. 349 (2007) 2380.
[29] (a) M.J. Earle, J.M.S.S. Esperanca, M.A. Gilea, J.N. Canongia Lopes, L.P.N. Rebelo,
J.W. Magee, K.R. Seddon, J.A. Widegren, Nature 439 (2006) 831;
(b) P. Wasserscheid, Nature 439 (2006) 797.
[30] K. Vrieze, H.C. Volger, A.P. Praat, J. Organometal. Chem. 14 (1968) 185.
[31] L.M. Green, D.W. Meek, Organometallics 8 (1989) 659.
[32] F. Sakurai, H. Suzuki, Y. Moro-oka, T. Ikawa, J. Am. Chem. Soc. 80 (1980)
1749.
2
N(SO2CF3)2), 86.82 (d, JCP = 11.7 Hz, CODvinyl trans PPh3), 85.78
(d, JCP = 12.1 Hz, CODvinyl trans PPh3), 80.05 and 79.56 (s, CODvinyl
2
trans NHC), 50.47 (s, NCH2R), 36.87 (s, NCH3), 30.24 (s, CODallyl
trans PPh3), 29.66 (s, CODallyl trans NHC), 31.74 (s,
NCH2CH2CH2CH3), 19.89 (s, NCH2CH2CH2CH3), 13.09 (s,
NCH2CH2CH2CH3).
31P NMR: d = 18.15 (s, PPh3). [*Identified through C–H correla-
tion spectra, peaks partly or completely overlapped with strong
solvents signals of BMIM NTf2. ** Common signals for both the sol-
vent and the complex.]
ESI+MS
(
193Ir/
191Ir = 63:
37)
701.26,
699.26
[Ir(COD)(PPh3)(NHC)]+, 433.14, 435.14 [Ir(C8H10)(NHC)]+; 431.12,
429.10 [Ir(C8H8)(NHC)]+. The complex was fully characterised by
X-ray diffraction studies.
Acknowledgements
[33] L.D. Vazquez-Serrano, B.T. Owens, J.M. Buriak, Chem. Commun. 2518
(2002).
[34] N.B. Cech, C.G. Enke, Practical Implications of some Recent Studies in
Electrospray Ionization Fundamentals, Wiley Periodicals, New York, 2002.
[35] R.B. King, Appl. Spectr. 23 (1969).
[36] C.M. Crudden, D.P. Allan, Coord. Chem. Rev. 248 (2004) 2247.
[37] W.A. Herrmann, Angew. Chem. Int. Ed. 41 (2002) 1290.
[38] D.R. Baghurst, D.M.P. Mingos, M.J. Watson, J. Organometal. Chem. 368 (1989)
C43.
U. H and T.G greatly appreciate financial support of the Ger-
man-French University DFH-UFA, the région Rhône-Alpes, and
the EC through a Network of Excellence Grant, IDECAT (No.
NMP3-CT-2005-011730).
Appendix A. Supplementary material
CCDC 279579 contains the supplementary crystallographic data
for this paper. These data can be obtained free of charge from The