DOI: 10.1002/chem.

201505133 Full Paper

& Solvent Effects

Solvent-Dependent Structure of Iridium Dihydride Complexes:

Different Geometries at Low and High Dielectricity of the Medium
Alexey V. Polukeev,[a] Roco Marcos,[b] Mrten S. G. Ahlquist,[b] and Ola F. Wendt*[a]

Abstract: The hydride iridium pincer complex [(PCyP)IrH2] Soc. 2006, 128, 17114). Based on the existence of an agostic
(PCyP = cis-1,3-bis[(di-tert-butylphosphino)methyl]cyclohex- bond between a-CH and iridium in 1 in all solvents, we
ane, 1) reveals remarkably solvent-dependent hydride chem- argue that the coordination of solvent can be rejected. DFT
ical shifts, isotope chemical shifts, JHD and T1(min), with rHH calculations revealed that the structures of 1 and [(POCO-
increasing upon moving to more polar medium. The only P)IrH2] depend on the dielectric permittivity of the medium
known example of such behaviour (complex [(POCOP)IrH2], and these compounds adopt trigonal-bipyramidal geome-
POCOP = 2,6-(tBu2PO)2C6H3) was explained by the coordina- tries in non-polar media and square-pyramidal geometries in
tion of a polar solvent molecule to the iridium (J. Am. Chem. polar media.

Introduction a result of a change in the equilibrium between two distinct

chemical entities (i.e., minima on the potential energy surface,
It has long been known that the choice of solvent often has PES) that always co-exist, but in different ratios. A rare example
an important impact on the direction, rate and yield of organic of the same type in coordination chemistry is represented by
and organometallic reactions.[1] Despite all the advances in the some nickel phosphine complexes, which are known to crystal-
field, such choices are still frequently made on the basis of em- lise in square-planar or tetrahedral geometries depending on
pirical knowledge on what is good for related reactions, rather the choice of solvent.[3]
than on a detailed understanding of how the properties of At the same time, it remains an open question to what
molecules change in different media. Therefore a better under- extent non-specific solvation can affect the properties of
standing of solvation effects is desired. These effects can be a single molecule (specific interactions like hydrogen bonds
caused by non-specific electrostatic and polarisation forces, as apparently may be strong enough to affect molecules and will
well as specific by interactions like hydrogen bonding or Lewis not be further discussed). This issue is highly relevant to the
acid/base type interactions.[1, 2] chemistry of push–pull dyes and advanced materials like
Typically, medium-related features are described in terms of switches[4] and non-linear optic[5] and photorefractive[6] sub-
the differential solvation of several components. The differen- stances. Also, compounds sensitive to changes in medium can
tial solvation of reactants and products is responsible for be used as molecular probes for the characterisation of micro-
changes in chemical equilibria, of reactant and transition states environments in various structures including proteins.[2, 7] In
for changes in reaction rates, and of ground and excited states principle, should the medium effect be significant and well un-
for solvatochromism, that is, the change in the colour of a com- derstood, this may open up access to a continuum of struc-
pound upon changing solvent.[1, 2] A simple and well-known ex- tures of a single compound and new possibilities for the fine-
ample of differential solvation in organic chemistry is keto– tuning of its reactivity and properties.
enol tautomerism. Under certain conditions, spectroscopic The number of publications about the influence of non-spe-
methods may reveal an almost exclusive presence of ketone in cific solvation on molecular properties is limited.[4–8] There has
one solvent and enol in another; this does not mean that the been a discussion as to whether the relative contributions of
ketone molecule no longer exists and is affected in such a way zwitterionic and quinoidal forms of push–pull dyes can be
that it becomes an enol. Instead, such observations would be changed depending on the solvent[4, 5] and how such changes
may affect the solvatochromic behaviour. Although this prob-
[a] Dr. A. V. Polukeev, Prof. O. F. Wendt lem still remains somewhat controversial, recent electro-optical
Centre for Analysis and Synthesis, Department of Chemistry absorption studies[4d] unambiguously showed that the dipole
Lund University, PO Box 124, 22100 Lund (Sweden) moment of the ground state of a dye is changed to some
E-mail: [email protected]
extent depending on the medium. Little is known, however,
[b] Dr. R. Marcos, Prof. M. S. G. Ahlquist
about changes in the structures of the molecules, because fea-
Division of Theoretical Chemistry and Biology, School of Biotechnology
KTH Royal Institute of Technology, 106 91 Stockholm (Sweden) tures of the NMR, UV/Vis and vibrational spectra that may be
Supporting information for this article is available on the WWW under attributed to medium effects do not provide such information
http://dx.doi.org/10.1002/chem.201505133. directly, and because it is not easy to separate specific solva-

Chem. Eur. J. 2016, 22, 1 – 10 1  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim & &

These are not the final page numbers! ÞÞ

 Full Paper

tion effects from non-specific. Computational studies at various These geometry changes are likely occurring within a single
levels of theory typically indicate relatively minor structural molecule.
changes of around 0.01–0.02  or less[4a, b, 5a, 8] upon changing
the medium (under certain conditions up to 0.08 [4b, 6]), which
nevertheless are believed to have an important impact on the
Results and Discussion
spectroscopic properties. The iridium pincer complex 1 was prepared according to
Herein we report an extreme case of a non-specific medium Scheme 1. The 31P{1H} NMR spectrum of 1 in deuterated ben-
effect in transition-metal hydride chemistry. We show that zene reveals a singlet at 86.9 ppm. In the 1H NMR spectrum,
even relatively minor changes in polarity (methylcyclohexane the a-CH group undergoes a significant upfield shift com-
to dichloromethane) result in dramatic structural changes in iri- pared with 2 and 3 (Table 1) and appears as a broadened trip-
dium hydride complexes with pincer ligands, which change let of triplets due to splitting by the adjacent CH groups and
their coordination polyhedra from trigonal bipyramidal (TBP)
to square pyramidal. The specific nature of these compounds
allows H–H distances between hydrides to be estimated by
NMR spectroscopy; the values obtained are in good agree-
ment with those predicted by DFT calculations and together
support large structural changes.
Transition-metal hydrides are important intermediates in
Scheme 1. Synthesis of dihydride complex 1.
a number of catalytic processes[9] and have been extensively
studied in the past several decades. It is now recognised that
there is a continuum of compounds with various degrees of
activation of hydrogen, from “non-classical” dihydrogen com- Table 1. Selected NMR parameters for complexes 1–3 (in C6D6), confirm-
plexes, in which the HH bond remains almost intact, to “clas- ing the agostic bonding in 1.
sical” hydrides, in which it is completely cleaved.[10] The charac- 1 2 3
terisation of these two types of complexes is quite straightfor-
dH (a-CH) [ppm] 0.54 2.42 2.02
ward, whereas species in the middle of the continuum, the so- 3
JHH [Hz] 8.4 2.2 3.0
called “elongated dihydrogen complexes” and “compressed di- 1
JCH [Hz] 104 123 122
hydrides”, often reveal unusual properties and are not straight- dC (a-C) [ppm] 84.8 34.7 35.8
forward to distinguish. This class of compounds is character-
ised by fairly flat, and often anharmonic, potential energy sur-
faces with respect to the motion of the hydrogen atoms, and, the two hydrides. The latter interaction demonstrates a remark-
sometimes, minima for both dihydride and dihydrogen struc- able 3JHH of 8.4 Hz, which, together with the downfield shift of
tures co-exist.[11, 12] One of the most important consequences of the a-carbon atom and the reduced one-bond CH coupling,
such a PES shape is a temperature dependence of JHD in deute- is indicative of an agostic bond between the a-CH group and
rium-labelled species, which has been explained by the ther- iridium. Similar spectral features have been observed for agos-
mal population of excited vibrational states.[12, 13] Although tic bonding in related osmium and iridium complexes with the
a number of studies have focused on this and related phenom- (tBu2PCH2)2CH2 ligand.[18] The two hydrides are in rapid ex-
ena, little is known of the effects of solvent on the behaviour change at room temperature and resonate as a broadened
of complexes with elongated dihydrogen and compressed di- doublet of triplets at 22.84 (3JHH = 8.4 Hz, 2JPH = 9.9 Hz) due to
hydrides. coupling with two chemically equivalent phosphorous nuclei
Iridium pincer complexes with tridentate PCP-type ligands and the a-CH group.
have been shown to be effective catalysts for various dehydro- Although prior studies have shown that calculations per-
genation reactions.[14] Previous studies of the arene-based com- formed by using the LYP correlation functional underestimate
plexes [(PCP)IrH2][15] and [(POCOP)IrH2][16] (POCOP = 2,6- agostic interactions,[19] the B3LYP functional predicts very well
(tBu2PO)2C6H3) suggest that for the latter there is a temperature the agostic interaction in complex 1. It was not possible to
and solvent dependence in the bonding of the hydrides, and locate a minimum corresponding to a non-agostic complex.
this was explained by solvent coordination. We have a long- Other functionals were also tested (PBE and PBE-D3) for the
standing interest in aliphatic pincer ligands based on the cyclo- geometry optimisation, with essentially the same results as ob-
hexyl framework,[17] and therefore we decided to synthesise tained with B3LYP. A natural bond order (NBO) analysis sup-
and investigate the dihydride complex [(PCyP)IrH2] with a cyclo- ported the agostic interaction between a-CH and iridium in
hexyl-based ligand in order to compare it with the aromatic complex 1. The data in Table 1 show that there are no agostic
systems. In this paper we present an experimental and theoret- interactions in complexes 2 and 3 (see the Supporting Informa-
ical analysis of the effect of solvent on the structures of these tion for further details).
dihydrides and show that both aromatic and aliphatic iridium A variable-temperature NMR study of complex 1 revealed
pincer dihydrides undergo structural changes that are deter- the existence of two dynamic processes. The first one operates
mined by solvent polarity rather than solvent coordination. above 0 8C and involves the exchange of hydrides with the a-
CH group; this process will not be discussed here in detail.[20]

& & Chem. Eur. J. 2016, 22, 1 – 10 www.chemeurj.org 2  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

 Full Paper

The second dynamic process involves an exchange between may be a result of changes in the structure of a single com-
the hydrides, with de-coalescence of signals observed below pound or a change in the equilibrium between two structures.
80 8C (Table 2). There is a surprisingly large difference be- Observed JHD values can be straightforwardly correlated to
tween the positions of the hydride signals depending on the H–H distances by using Equation (1), initially suggested by Lim-
solvent used (Figure 1). Thus, the spectra recorded at 100 8C bach and co-workers[21] and refined by Heinekey and co-work-
show non-exchanging hydride signals at 23.0 and ers.[12a]
27.9 ppm in [D2]dichloromethane (Dd = 4.9 ppm), at 23.07
and 23.40 in [D8]toluene (Dd = 0.33 ppm) and at around
22.7 and 22.9 (Dd  0.2 ppm) in [D14]methylcyclohexane. rðH  HÞ ¼ 0:74  0:494
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð1Þ
The free energies of activation for the exchange vary from 7.3 16:1447  260:65  4ð1 þ 0:32895JHDÞ
to 8.3 kcal mol1 depending on the solvent. lnð Þ
2

Table 3. H–H distances between hydride ligands in complex 1 according

to T1(min), JHD and DFT calculations.

Solvent [D14]Methyl- [D8]Toluene [D2]Dichloro-

cyclohexane methane
rHH from T1(min) [] 1.54 1.58 1.70
rHH from JHD [] 1.65 1.71 1.86
rHH from DFT [] 1.69 1.69 1.77

Figure 1. Hydride region of the 1H NMR spectrum of complex 1 in

[D8]toluene (top) and [D2]DCM (bottom) at around 100 8C. The resulting distances are presented in Table 3. Interpreta-
tion of the T1(min) data by using the methodology developed
by Halpern and co-workers[22] requires an accurate consider-
Table 2. Chemical shifts, isotope shifts, T1(min), JHD and Gibbs energies of ation of the relaxation in the tert-butyl groups and other
activation for complex [(PCyP)IrH2] (1) in various solvents. ligand atoms. Owing to the absence of XRD data, we decided
Solvent [D14]Methyl- [D8]Toluene [D2]Dichloro- to take interatomic distances (except for rHH) from the DFT-op-
cyclohexane methane timised structures. We believe that this approach accounts for
dIrH2 (25 8C) [ppm] 22.64 22.85 24.13
the contribution of the pincer ligand to the overall relaxation
davgIrH2 (ca. 100 8C)[a] 22.85 23.23 25.50 rate in a reasonable way. For instance, the value of T1(min) for
[ppm] [(PCyP)IrHCl] (3) calculated by this method is in good agree-
Dd[b] [ppm] 0.21 0.38 1.37 ment with the experimentally observed value (341 vs 312 
dIrHaHb[c] [ppm] 22.7 and 23.07 and 23.0 and
22.9 23.40 27.9
5 ms). The H–H distances so obtained are listed in Table 3 and
Dd(HbHa) [ppm] 0.2 0.33 4.9 are systematically shorter than those obtained from JHD values.
DG°[d] (100 8C) [kcal 8.3 8.1 7.3 This may indicate the presence of some dihydrogen species,
mol1] but on the other hand, it has been noted[23, 24] that Equation (1)
T1(min) [ms] 134 (40 8C) 147 (50 8C) 202 (70 8C)
JHD [Hz] 4.7 3.8 2.0
does not provide a very accurate relationship between JHD and
Dd(IrH2IrHD) [ppm] 0.17 0.26 0.36 rHH for small values of JHD, so distances from JHD may be slightly
overestimated (especially large data scattering is observed for
[a] (dIrHa + dIrHb)/2. [b] from 100 8C to 25 8C. [c] According to spin simu-
lations; possible influence of exchange coupling (see below) is neglected. the longest rHH). Thus, the JHD and T1(min) data are both consis-
Values differ from those directly observed at 100 8C by ca. 0.1 ppm. tent with the description of 1 as a compressed dihydride com-
[d] For IrHaHb exchange. plex with a remarkable change of rHH (0.15–0.20 ) upon
moving from [D2]dichloromethane to [D14]methylcyclohexane
solution. Unfortunately, we were not able to measure the tem-
Owing to the well-known difficulties in the determination of perature dependence of JHD, which is an important property of
hydride positions from X-ray diffraction data and the lack of compressed dihydrides because of the two exchange process-
readily accessible neutron diffraction facilities, solution-state es, which are responsible for line broadening at high and low
techniques for determining H–H distance based on spin–lattice temperatures.
relaxation times T1(min) and H–D couplings in deuterium-la- The hydride chemical shifts for complex 1 are solvent- and
belled compounds have been developed.[10a] It should be temperature-dependent (Table 2), moving to a lower field as
noted that if several species are simultaneously present, the the temperature increases. For instance, Dd between 100
observed values of T1(min) and JHD will be population-weighted and 25 8C in CD2Cl2 is 1.37 ppm, but in hydrocarbon solvents
averages, with dihydrogen-like species contributing dispropor- the difference is smaller. The temperature dependence of d
tionally to the T1(min) value because of the T1(min)–(rHH)6 de- was previously explained by the occupation of excited vibra-
pendence.[16] In the case of 1, the situation is further complicat- tional energy levels within a single structure or by a shift of
ed by the significant solvent effects on T1(min) and JHD. This the dihydride/dihydrogen equilibrium.[11, 12] The lack of NMR

Chem. Eur. J. 2016, 22, 1 – 10 www.chemeurj.org 3  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim & &

These are not the final page numbers! ÞÞ

 Full Paper

signals from the dihydrogen form at low temperatures after The remarkable effect of solvent on JHD and T1(min) for com-
de-coalescence and the results of DFT calculations (see below) plex 1 significantly exceeds those typically observed for hy-
argue for a single structure model. dride complexes and deserves some comments. The trend is
Isotope effects on the chemical shift of the hydride reso- that the H–H distance increases on transferring from
nance in 1 are negative (to low field) and vary from 0.17 [D14]methylcyclohexane to [D2]dichloromethane. To the best of
([D14]methylcyclohexane) to 0.36 ppm ([D2]dichloromethane). our knowledge, only the family of phosphinite pincer com-
These values are larger than those associated with typical in- plexes [(p-X-POCOP)IrH2] show comparable behaviour, specifi-
trinsic isotope effects (less than 0.07 ppm) and may be ex- cally [(p-{3,5-(CF3)2C6H4}POCOP)IrH2] (Scheme 3), in which JHD
plained by an isotopic perturbation of equilibrium (IPE).[25] This
happens when there are non-equivalent hydride sites in a mol-
ecule exchanging rapidly on the NMR timescale and deuterium
preferentially occupies one of these sites (Scheme 2).

Scheme 3. Equilibrium between the solvated dihydride and the dihydrogen

form of [(p-X-POCOP)IrH2], as proposed in ref. [16].

Scheme 2. Equilibrium between two mono-deuterated molecules of 1.

was reported to change from 9.0 Hz in pentane to less than
1 Hz in [D2]dichloromethane.[16] For this compound, the exis-
tence of an equilibrium between a dihydride containing a coor-
For a solution in [D2]dichloromethane, application of Equa- dinated molecule of solvent and a dihydrogen complex was
tion (2),[25a] in which dxdy corresponds to the difference in the proposed.[16] The rationale behind the coordinated solvent was
chemical shifts of the hydride resonances, gives K = 1.38, which the large isotope effect observed on the chemical shifts, which
indicates a preference of deuterium to occupy the site trans to need non-equivalent hydride sites; at the same time, based
the agostic bond. upon previous computational studies of [(PCP)IrH2],[15] one
could ascribe a C2v symmetry to [(p-X-POCOP)IrH2] and thus the
ðdx  dyÞð1  KÞ hydrides should be equivalent.
Dd ¼ dHD  dHH ¼ ð2Þ
2ðK þ 1Þ Complex 1 is an 18-electron compound and does not coor-
dinate any solvent molecules, as shown by the presence of the
The situation is less easy to interpret when it comes to hy- agostic coordination of the a-CH bond at all temperatures in-
drocarbon solvents. Thus, in [D8]toluene, dHHdHD (0.26) is vestigated, but demonstrates similar, although somewhat
larger than the theoretical limit (dxdy)/2 (0.33/2  0.17). The smaller changes. This indicates that the possible coordination
value of dHHdHD in [D8]toluene is almost unchanged with tem- of solvent to iridium is not the reason for the observed trends.
perature up to the coalescence of the isotopomer signals. In Another option might be the formation of a so-called dihydro-
the 2D NMR spectrum, the signal of 1-D2 almost coincides with gen bond (DHB) of the MH···HC[28] type between the basic
the signal of 1-H2, whereas the 1-HD resonance is shifted IrH and the somewhat acidic HCHCl2 sites. The main NMR
downfield by the same value that it is shifted upfield in the spectral criteria for DHB bonding are a small upfield shift of
1
H NMR spectrum, well in line with the hypothesis of an iso- the hydride atom signal (ca. 0.2–0.8 ppm) and a decrease in
topic perturbation of the equilibrium. This leads to the conclu- T1(min).[29] The latter, put simply, happens because additional
sion that dxdy is actually somewhat larger than the value of protons in the vicinity of a hydride provide an additional
0.33 ppm, estimated from the spectrum recorded at 100 8C. source of relaxation. Experimentally, it was found that for com-
A possible reason for this is the existence of quantum-mechan- plex 1, T1(min) increases upon moving to dichloromethane
ical exchange coupling[26] with a magnitude of a few hundred from less polar solvents (see Table 2), which argues against the
Hertz. The barrier for the classical exchange of hydrogen presence of DHBs. If, however, DHB formation leads to struc-
atoms is comparable to those for which exchange coupling tural changes, specifically an elongation of the H–H distance,
has been observed previously.[26] In this case, the lines ob- this may to some extent counter-balance the appearance of an
served in the 1H NMR spectrum of 1 at 100 8C in hydrocarbon additional source of relaxation, but even this cannot explain
solvents are not hydride resonances themselves, but a central the longer relaxation time.[30] Furthermore, electron-rich [(PCy-
part of an AB system, with the individual chemical shifts of the P)IrH2] (1) should reveal a stronger DHB and thus larger
hydrides lying somewhat away from the central pattern. Un- changes than observed for [(POCOP)IrH2], which again contra-
fortunately, a lack of reliable information about dxdy does not dicts the experimental data. Also, DFT calculations showed no
allow a more precise evaluation of the exchange coupling. An indication of DHB formation (see the Supporting Information
alternative explanation is that for 1, the substitution of H by D for details). Thus, it can be reasoned that compounds like
may lead to slight changes in geometry (see below) and thus 1 and [(p-X-POCOP)IrH2] are not prone to specific interaction
an intrinsic isotope effect on the chemical shift may be signifi- with the solvents used but instead are rather sensitive to
cant.[27] changes in the polarity of the medium.

& & Chem. Eur. J. 2016, 22, 1 – 10 www.chemeurj.org 4  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

 Full Paper

DFT calculations provide some insight into the origin of the vacuum reveals the presence of two minima, one with C2
behaviour of 1 (and [(p-X-POCOP)IrH2]) in various solvents. In (almost C2v) symmetry with almost equal CIrH angles of
agreement with the NMR data, no minimum was found corre- 148.88 and a H–H distance of 1.66 , and another in which the
sponding to the dihydrogen structure of 1, so all changes ob- structure is slightly de-symmetrised, with CIrH angles of
served are associated with a single dihydride compound; the 153.5 and 143.58 and an H–H distance of 1.67 . Both struc-
coordination of dichloromethane was found to be clearly ener- tures are characterised as minima with no imaginary frequen-
getically unfavourable. The geometrical preferences of cies and are very close in energy. In a non-polar medium the
[L2IrXYY’] molecules have previously been studied by Eisen- H–H distance is slightly increased; thus, for e = 2 and a non-
stein, Pelissier and co-workers,[31] and very recently Goldman, symmetrical Y shape, the CIrH angles are 141.8 and 153.78,
Hasanayn and co-workers[32] specifically investigated the geo- and the H–H distance is 1.70 . Surprisingly, in a polar
metries of PCP–iridium pincer complexes. To summarise, simi- medium, the behaviour of the two structures is dramatically
larly to the d6 ML5 case, for [(PCP)IrH2] molecules a pseudo-D3h different. The non-symmetrical Y shape spontaneously rear-
trigonal-bipyramidal structure (with the angles between equa- ranges to the non-symmetrical T-shaped geometry: for e = 64,
torial ligands around 1208) is unstable and will undergo some the CIrHa angle becomes close to 1808 (172.88) and thus the
Jahn–Teller-type distortion. This will decrease or increase the corresponding hydride moves trans to the ipso carbon and the
HIrH angle to give a distorted TBP/Y shape or a square-pyra- second hydride occupies the apical position (CIrHb = 110.58,
mid/T shape (symmetrical/non-symmetrical), respectively HIrH = 76.78). This is accompanied by an increase in the dis-
(Scheme 4). For molecules similar to the ones studied in this tance between the hydrides from 1.70 to 1.99 . Although the
work, it was found that the Y-shaped structure is the global symmetrical Y shape is still identified as a minimum, the rela-
minimum,[31, 32] whereas the symmetrical T shape is a transition tive energy (SCF + Gsolv) of the T-shaped structure is 1.0 kcal
state connecting non-symmetrical T shapes,[32] which were mol1 lower. At the same time, the IrHa bond becomes longer
“found to be significantly higher in energy than the symmetri- (in agreement with the increased trans influence) and the Ir
cal Y geometry”[32] and were not discussed in detail. Hb bond shorter. Remarkably, the net atomic charge of Ha
changes from + 0.13e to 0.11e, and the charge of Hb remains
almost the same (+ 0.11e and + 0.14e). Thus, the complex [(p-
H-POCOP)IrH2] exists as a mixture of two components, Y-symm.
and Y-non-symm, with the latter dramatically changing geome-
try upon increasing the polarity of the medium, whereas the
geometry of the former shows only very minor changes. In
a polar medium, mainly the T shape is populated, which ex-
plains the small JHD and unusually large isotope effect on the
chemical shifts experimentally observed for [(p-{3,5-
(CF3)2C6H4}POCOP)IrH2]. To see the effect of the tBu groups we
exchanged them for methyl groups and optimised the geome-
tries of the C2v-symmetric Y-shaped structure and the T-shaped
structure. Again, we found that the Y shape is the more stable
structure (by 1.2 kcal mol1) in the non-polar medium, whereas
the T shape becomes the more stable (by 2.5 kcal mol1) in the
polar medium. One difference from the tBu ligand is that the T
shape is also identified as a minimum in the non-polar
Scheme 4. Jahn–Teller-type distortions in [(PCP)IrH2] complexes in accord medium. In vacuum, only the Y-shaped structure is a mini-
with refs. [31, 32]. mum.
Similar, but less pronounced effects are observed for com-
Here we have shown that solvent effects are extremely im- plex 1 (Figure 3), which behaves similarly to the previously de-
portant for this type of molecule and should not be neglected. scribed non-symmetrical Y shape and becomes more T-shape-
Most clearly, these effects are exemplified by the behaviour of like upon increasing the polarity. The distances between hy-
the complex [(p-H-POCOP)IrH2] (Figure 2). For simplicity and to drides obtained in the calculations are in good agreement
emphasise the changes, calculations for vacuum, e = 2 (non- with those derived from JHD and T1(min) (Table 2). These results
polar medium) and e = 64 (polar medium) are presented. We strongly indicate that no specific interactions with solvent are
do not attempt to make an accurate quantitative comparison required to explain the NMR observations. In non-polar sol-
between experimental and computational data for [(p-H-POCO- vents the geometry around iridium is closer to TBP and the dif-
P)IrH2] because, as shown by the temperature-dependent JHD, ference in chemical arrangement between the hydrides is rela-
this compound is likely affected by the occupied vibrational tively small, which leads to small Dd and dHHdHD values. In
states; for [(PCyP)IrH2] (1), calculations involving the experi- polar solvents, the geometry is closer to square pyramidal, and
mental solvents were also carried out and showed good agree- the hydride, which becomes almost trans to the a-carbon
ment of the H–H distances (Table 3) and IR spectra (see below). atom, undergoes a downfield shift, whereas the hydride that
The DFT-based geometry optimisation of [(p-H-POCOP)IrH2] in occupies the apical position undergoes an upfield shift. Thus,

Chem. Eur. J. 2016, 22, 1 – 10 www.chemeurj.org 5  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim & &

These are not the final page numbers! ÞÞ

 Full Paper

Figure 2. Geometry of [(POCOP)IrH2] in a vacuum (left) and in non-polar (PBF[33] solvation model, e = 2, middle) and polar (PBF solvation model, e = 64, right)
media.

Figure 3. Geometry of [(PCyP)IrH2] (1) in a vacuum (left) and in non-polar (PBF solvation model, e = 2, middle) and polar (PBF solvation model, e = 64, right)
media. The a-HIr distance is 2.58–2.59 .

larger values of Dd and dHHdHD are observed. At the same 2012 cm1; a redistribution of the intensities is also observed
time, the distance between the hydrides increases, which re- (2189 and 1978 cm1, 1:6). This feature is fully captured by DFT
sults in a decrease in JHD and increase in T1(min). To the best of calculations (see Figure S8 in the Supporting Information) and
our knowledge, this is the first example of a solvent-depen- is in agreement with the suggested geometry changes.
dent geometry of a transition-metal dihydride. The simplest way to rationalise the geometry changes in
Additional confirmation of the suggested model comes from 1 and [(p-X-POCOP)IrH2] is in terms of trans influence. Both hy-
IR spectroscopy. The IR spectrum of 1 in hexane reveals two drides and pincer backbones (aryl or cyclohexyl) are ligands
broad bands at 2190 and 2012 cm1 (1:4), which correspond to with strong trans influence, which makes their mutual trans ar-
the symmetrical and asymmetrical stretching vibrations of the rangement unfavourable and this destabilises the symmetrical
IrH bonds. In dichloromethane, the span between bands in- (H trans to H) and non-symmetrical (H trans to Cipso) T shapes
creases, mainly due to a redshift of the absorption at relative to the Y shape. That is why the Y shape is a minimum

& & Chem. Eur. J. 2016, 22, 1 – 10 www.chemeurj.org 6  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

 Full Paper

in vacuum. Interactions with the solvent polarise the IrHa why the T-shaped geometry is disfavoured and changes associ-
bond, which results in its elongation and a diminishing of the ated with different solvents are less pronounced than for [(p-X-
trans influence of Ha. These changes lead to the stabilisation of POCOP)IrH2]. In this respect, it is interesting to note that
the non-symmetrical T-shaped structure, in which Ha is trans to a somewhat related trend can be observed in the structures of
Cipso ; at the same time, the IrHb bond becomes shorter, hydrido-chlorido complexes. The chlorine, which is a weak
which, in the case of the symmetrical T shape may counter-bal- s donor, prefers to occupy the position trans to the aromatic
ance the decreased trans influence of Ha. To verify this hypoth- ring, which makes the iridium arrangement square-pyramidal,
esis we calculated the structures of several compounds with with the hydride occupying the apical coordination site. A C
different X groups in the para position of the aromatic ring of IrCl angle close to 1808 is thus observed for [{2,6-
[(p-X-POCOP)IrH2] (s-donating ability[34] decreases in the order (tBu2PO)2C6H3}IrHCl] (179.5(1)8),[17c] [{4-MeO-2,6-
NH2 > OMe > tBu > H > Br > C6F5 > NO2, see Table 4). (tBu2PO)2C6H2}IrHCl] (174.75(15)8) [35] and [{4-NO2-2,6-
In the polar medium, all the structures are best described as (tBu2PCH2)2C6H2}IrHCl] (179.33(14)8).[36] However, when a strong-
T-shaped. Here, the hydride would be a weaker donor than in ly donating aliphatic carbon, boron or silicon is introduced, the
the non-polar medium and the more weakly donating the aryl C(B,Si)IrCl angle decreases towards that of a TBP-like geome-
group is, the larger the H–H distance (upper line in Table 4), try: [(PCyP)IrHCl] (3; 166.6(1)8),[17c] [(PBP)IrHCl] (157.67(17)8)[37]
which indicates a more pronounced T shape. Two weaker and [(PSiP)IrHCl] (130.64(4)8).[38]
donors are more likely to be mutually trans than two stronger To verify our conclusions, we measured the value of JHD for
ones. In a non-polar medium the behaviour is different. All the the unsubstituted complex [(p-H-POCOP)IrH2] in chlorinated
compounds except that with X = NO2 are Y-shaped. As the p- solvents. The data in Table S1 in the Supporting Information
donating ability of the aryl ring decreases, the distance be- clearly show that the less electron-rich complex [(p-{3,5-
tween the hydrides decreases. This effect is likely due to a low- (CF3)2C6H4}POCOP)IrH2] is more affected by changes in polarity
ering of the energy of the d orbital that interacts with the p of the solvent than [(p-H-POCOP)IrH2] and gives a more T-
system and the two hydrides. However, when the electron shaped geometry in polar medium, which is exemplified by
density at Cipso is at its lowest, in the NO2-substituted ligand, its the dramatic decrease in JHD (from 9.0 to < 1 Hz) and increase
trans influence is low enough for the complex to rearrange to in IPE (from 0.2 to more than 1). In agreement with the
a T-shaped structure. In the case of complex 1, the cyclohexyl computational results presented in Table 4, the changes are
moiety is a stronger s donor than an aryl group,[34] and that is smaller for the complex [(p-H-POCOP)IrH2] (from 8.0 to 4.8 Hz
for JHD and from 0.10 to 0.52 for IPE).
It also can be seen from the data in Table S1 in the Support-
Table 4. Calculated Ha–Hb distances for [(p-X-POCOP)IrHaHb] (non-symm. ing Information that changes in IPE and JHD, which reflect the
Y shape) illustrating the effect of solvent and para substituent on the ge- degree of de-symmetrisation of the complex [(p-H-POCO-
ometry of the complex. P)IrH2], correlate with e of the medium. An exception is o-di-
chlorobenzene ([D4]o-DCB), which, despite having a higher
X sp[a] r(Ha–Hb) []
e=2 e = 64 e than DCM, induces smaller changes. This likely happens be-
cause of the higher molecular volume of o-DCB, which ham-
NH2 0.66 1.72 1.83
OMe 0.27 1.70 1.84 pers the polarisation of IrH bonds. Very importantly, DFT cal-
tBu 0.20 1.70 1.90 culations were able to reproduce the smaller effect of o-DCB
H 0.0 1.70 1.99 compared with DCM (see Figure S1). If the observed trends
Br 0.23 1.69 2.07
would be explained by the coordination of the chlorine to iridi-
C6F5 0.27 1.69 2.07
NO2 0.78 2.02 2.12 um, one could expect a similar or even greater effect from o-
DCB, because the flat nature of this molecule may facilitate its
coordination to iridium (in the plane between four tBu groups)
compared with the tetrahedral DCM.
Given the presence of both Y-symm and Y-non-symm mole-
cules of [(POCOP)IrH2]-type compounds, a quantitative analysis
is complicated. Nevertheless, values of JHD derived from calcu-
lated H–H distances by using Equation (1) are reasonably close
to the experimental values determined at low temperatures.
Thus, for [(p-H-POCOP)IrH2] in DCM, the experimental value of
JHD is less than 3 Hz, whereas the value calculated by using rHH
for the non-symmetrical Y shape is 2.2 Hz. At higher tempera-
tures the experimental values of JHD are somewhat higher than
the calculated values. We believe that two factors account for
this underestimation. The first is an increased population of
the symmetrical Y shape, and the second is a possible non-
zero population of the first excited vibrational state, which is
[a] See ref. [34].
believed to have dihydrogen complex character and is

Chem. Eur. J. 2016, 22, 1 – 10 www.chemeurj.org 7  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim & &

These are not the final page numbers! ÞÞ

 Full Paper

known[12, 13] to increase JHD for compressed dihydride com- ence of solvent, leading to, for example, new possibilities to
plexes. fine-tune the reactivity or design molecular switches.
In a very recent paper[39] that appeared after the completion
of this manuscript, a solvation effect on the hydricity (hydride-
donating ability measured as DG of hydride dissociation as H) Experimental Section
of [HNi(dhmpe)2][BF4] (dhmpe = 1,2-bis(dihydroxymethylphos-
General considerations: All manipulations were conducted under
phino)ethane) was reported. The authors found that compared
an inert gas atmosphere using standard Schlenk and glovebox
with CH3CN and DMSO, more polar water considerably facili- techniques unless otherwise stated. All solvents were distilled
tates the transfer of H . Although it is likely that hydrogen under vacuum from Na/benzophenone. Deuterated hydrocarbon
bonds may be important for this case, such reactivity has con- solvents were distilled under vacuum from Na/benzophenone and
siderable similarity with the effects reported in our work and is CD2Cl2 was distilled under vacuum from calcium hydride. NMR
closely related to the polarisation of MH bonds. Thus, calcula- spectra were recorded on Varian Unity INOVA 500 MHz and Bruker
tions clearly indicate a build-up of negative charge on Ha (+ Avance 400 MHz spectrometers. 1H and 13C NMR chemical shifts are
reported in parts per million and are referenced to the signals of
13e to 0.11e) for the complex [(POCOP)IrH2] upon changing
deuterated solvents. 31P NMR chemical shifts are reported relative
e from 2 to 64, which would apparently facilitate the dissocia-
to 85 % solution of phosphoric acid as external reference. The cali-
tion of Ha as an anion in a more polar medium, fully in line bration curve obtained by using neat methanol was used for varia-
with observations made for [HNi(dhmpe)2][BF4]. However, the ble-temperature measurements. Spin–lattice relaxation times (T1)
energy differences induced by solvation that were observed were determined by using a standard inversion recovery pulse se-
for iridium complexes are not that dramatic due the presence quence. IR spectra were recorded on a Bruker Alpha FT-IR spec-
of a second hydride, which to some extent counter-balances trometer. The complexes [(PCyP)IrHCl] (3)[41] and [(POCOP)IrHD][16]
the polarisation of IrHa by geometry rearrangement and the were prepared according to literature procedures.
lack of strong DHB-type interactions. We nevertheless believe Synthesis of tetrahydride 2: A Straus flask was charged with
that under certain conditions even such a subtle energy gap [(PCyP)IrHCl] (3; 0.030 g, 0.048 mmol) and tBuONa (0.0069 g,
0.072 mmol) inside a glovebox in a nitrogen atmosphere followed
between solvated forms could give rise to different reactivity.
by the vacuum-transfer of C6H6 (5 mL). The flask was refilled with
hydrogen and the red solution was heated on an oil bath over-
night at 100 8C. Completion of the reaction was indicated when
the solution turned colourless. After cooling the flask to room tem-
Conclusion
perature, degassed water (5 mL) was added and the mixture stirred
We have presented herein the synthesis and characterisation for 10 min. The organic phase was decanted, passed through
of dihydride complex 1, which was found to have solvent-de- a thin layer of Celite and the Celite was washed with small amount
pendent spectroscopic parameters. The existence of a weak of C6H6. Evaporation of the volatiles afforded 2 as a white powder
in quantitative yield. The NMR spectra are consistent with those
agostic interaction between a-CH and iridium, which makes
previously reported.[17e]
1 an 18-electron complex, allowed us to rule out potential sol-
Synthesis of dihydride 1: A Straus flask was charged with [(PCy-
vent coordination to the metal. DFT calculations provide an ex-
P)IrH4] (2; 0.028 g, 0.048 mmol) inside a glovebox in a nitrogen at-
planation for such behaviour of 1. Thus, in non-polar media mosphere followed by the vacuum-transfer of C6H6 (5 mL). The
1 has a trigonal-bipyramidal geometry, whereas in polar sol- flask was heated on an oil bath at 100 8C for 4 h, being periodically
vents the arrangement of ligands around iridium is closer to evacuated to remove evolving H2. The resulting orange-red solu-
square-pyramidal. It should be noted that the dihydrogen tion was evaporated to give 1 as a red powder in quantitative
structure is not a minimum on a potential energy surface. Even yield.
1
more marked changes are observed for [(p-X-POCOP)IrH2], H NMR (500 MHz, C6D6): d = 2.41–2.36 (m, 4 H; CHAHB-P, 3-HA, 5-
which explain previous NMR spectroscopic observations[16] of HA,), 2.09 (br m, 1 H; 4-HA), 1.94–1.86 (m, 2 H; 2-H, 6-H), 1.61–1.54
this complex. These results complement the previous theoreti- (m, 3 H; CHAHB-P, 4-HB), 1.27 (vt, j 3JPH + 5JPH j = 12.3 Hz, 18 H; 2 tBu),
1.22 (vt, j 3JPH + 5JPH j = 12.4 Hz, 18 H; 2 tBu), 1.09–1.02 (m, 2 H; 3-HB,
cal and experimental studies of [L’L2MH2] complexes[31, 32] and
5-HB), 0.54 (br m, 1 H; 1-H), 22.84 ppm (br dt, 2 H; Ir-H);
for the first time show that the polarity of the medium can 13 1
C{ H} NMR (126 MHz, C6D6): d = 84.82 (s, C-1), 50.60 (vt, j 2JPC +
play a decisive role in determining the geometry of such sys- 3
JPC j = 17.1 Hz, C-2, C-6), 37.77 (vt, j 1JPC + 3JPC j = 22.4 Hz, CH2-P),
tems in solution. 34.73 (vt, j 3JPC + 4JPC j = 17.9 Hz, C-3, C-5), ca. 34.65 (overlapping, vt,
The reason for such behaviour is the polarisation of IrH j 1JPC + 3JPC j = 20 Hz, C(CH3)3), 34.52 (overlapping, vt, j 1JPC + 3JPC j =
bonds by the solvent, which leads to a change in trans influ- 20.6 Hz, C(CH3)3), 29.90 (vt, j 2JPC + 4JPC j = 6.4 Hz, C(CH3)3), 29.79 (vt, j
2
ence of the hydrides and thus their geometrical preferences. JPC + 4JPC j = 6.3 Hz, C(CH3)3), 28.11 ppm (t, 4JPC = 1.5 Hz, 4-C);
31 1
Although compressed dihydrides possess a relatively flat po- P{ H} NMR (202 MHz, C6D6): d = 86.9 ppm (s); IR (hexane): ñ = 2190
(m), 2012 cm1 (s).
tential energy surface with respect to the movement of hy-
drides, which apparently facilitates the observation of the sol- Computational studies: All DFT calculations were carried out by
using the Jaguar 7.6 program pack-
vent effect, this may not be a necessary requirement.[39, 40]
age by Schrçdinger LLC.[42] For geom-
Thus, other methods of polarising chemical bonds, for in- etry optimisation, solvation energy
stance, the application of an external electrical field, might and frequency calculations, Becke’s
also, in principle, result in similar geometrical changes in these three-parameter hybrid functional
and other molecules and even outperform the reported influ- and LYP correlation functional

& & Chem. Eur. J. 2016, 22, 1 – 10 www.chemeurj.org 8  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

 Full Paper

(B3LYP)[43] was used with the LACVP** level core potential and [17] For selected references, see: a) S. Sjçvall, O. F. Wendt, C. Andersson, J.
basis set. For the geometry optimisation in solution, the Poisson– Chem. Soc. Dalton 2002, 1396 – 1400; b) D. Olsson, A. Arunachalampillai,
Boltzmann reactive field (PBF)[42] implemented in Jaguar 7.6 was O. F. Wendt, Dalton Trans. 2007, 5427 – 5433; c) A. Arunachalampillai, D.
Olsson, O. F. Wendt, Dalton Trans. 2009, 8626 – 8630; d) K. J. Jonasson,
used with standard parameters for benzene, dichloromethane and
O. F. Wendt, Chem. Eur J. 2014, 20, 11894 – 11902; e) A. V. Polukeev, R.
cyclohexane. For the geometry optimisations with different dielec-
Marcos, M. S. G. Ahlquist, O. F. Wendt, Chem. Sci. 2015, 6, 2060 – 2067;
tric constants (e = 2 and 64), the calculations were performed with f) K. J. Jonasson, A. V. Polukeev, R. Marcos, M. S. G. Ahlquist, O. F. Wendt,
2.0  as the probe radius. Angew. Chem. Int. Ed. 2015, 54, 9372 – 9375; Angew. Chem. 2015, 127,
9504 – 9507.
[18] a) M. A. McLoughlin, R. J. Flesher, W. C. Kaska, Organometallics 1994, 13,
Acknowledgements 3816 – 3822; b) D. G. Gusev, A. J. Lough, Organometallics 2002, 21,
2601 – 2603.
Financial support from the Swedish Research Council, the Knut [19] D. A. Pantazis, J. E. McGrady, F. Maseras, M. Etienne, J. Chem. Theory
Comput. 2007, 3, 1329 – 1336.
and Alice Wallenberg Foundation and the Crafoord Foundation [20] A. V. Polukeev, R. Marcos, M. S. G. Ahlquist, O. F. Wendt, unpublished re-
is gratefully acknowledged. Computational resources were pro- sults.
vided by the National Supercomputer Centre in Linkçping, [21] S. Grndemann, H.-H. Limbach, G. Buntkowsky, S. Sabo-Etienne, B.
Sweden. Chaudret, J. Phys. Chem. A 1999, 103, 4752 – 4754.
[22] P. J. Desrosiers, L. Cai, Z. Lin, R. Richards, J. Halpern, J. Am. Chem. Soc.
1991, 113, 4173 – 4184.
Keywords: density functional calculations · hydrides · iridium · [23] D. G. Gusev, J. Am. Chem. Soc. 2004, 126, 14249 – 14257.
[24] J. D. Egbert, R. M. Bullock, D. M. Heinekey, Organometallics 2007, 26,
pincer complexes · solvent effects
2291 – 2295.
[25] a) D. M. Heinekey, M. van Roon, J. Am. Chem. Soc. 1996, 118, 12134 –
[1] C. Reichardt, Solvents and Solvent Effects in Organic Chemistry, 3rd ed., 12140; b) W. J. Oldham Jr., A. S. Hinkle, D. M. Heinekey, J. Am. Chem.
Wiley-VCH, Weinheim, 2003. Soc. 1997, 119, 11028 – 11036.
[2] C. Reichardt, Chem. Rev. 1994, 94, 2319 – 2358. [26] S. Sabo-Etienne, B. Chaudret, Chem. Rev. 1998, 98, 2077 – 2092.
[3] R. G. Hayter, F. S. Humiec, Inorg. Chem. 1965, 4, 1701 – 1706, and refer- [27] X. S. Bogle, D. A. Singleton, J. Am. Chem. Soc. 2011, 133, 17172 – 17175.
ences therein. [28] T. B. Richardson, T. F. Koetzle, R. H. Crabtree, Inorg. Chim. Acta 1996, 250,
[4] a) D. G. Patel, M. M. Paquette, R. A. Kopelman, W. Kaminsky, M. J. Fergu- 69 – 73.
son, N. L. Frank, J. Am. Chem. Soc. 2010, 132, 12568 – 12586; b) J. O. [29] L. M. Epstein, E. S. Shubina, Coord. Chem. Rev. 2002, 231, 165 – 181.
Morley, R. M. Morley, R. Docherty, M. H. Charlton, J. Am. Chem. Soc. [30] If one places an additional hydrogen atom 2.0–2.1  from one of the
1997, 119, 10192 – 10202; c) M. E. Reish, A. J. Kay, A. Teshome, I. Assel- hydrides in 1 to model a weak DHB interaction (MH..HC distances
berghs, K. Clays, K. C. Gordon, J. Phys. Chem. A 2012, 116, 5453 – 5463; typically vary from 1.5 to 2.2  in the solid state with a mean distance
d) F. Wrthner, G. Archetti, R. Schmidt, H.-G. Kuball, Angew. Chem. Int. of 1.96 , see ref. [28]), this would lead to a distance between hydrides
Ed. 2008, 47, 4529 – 4532; Angew. Chem. 2008, 120, 4605 – 4608; e) S. R. of approximately 1.9 . Given that the T1(min)-derived d(HH) in hy-
Marder, J. W. Perry, B. G. Tiemann, C. B. German, S. Gilmour, S. L. Biddle, drocarbons is 0.11–0.13  shorter than the JHD-derived d(HH) (Table 3),
G. Bourhill, J. Am. Chem. Soc. 1993, 115, 2524 – 2526. this value is a bit too high compared with the value of 1.86  derived
[5] a) A. Marini, S. Macchi, S. Jurinovich, D. Catalano, B. Mennucci, J. Phys. from JHD in dichloromethane, which makes such a hypothesis unlikely.
Chem. A 2011, 115, 10035 – 10044; b) F. Meyers, S. R. Marder, B. M. [31] J.-F. Riehl, Y. Jean, O. Eisenstein, M. Pelissier, Organometallics 1992, 11,
Pierce, J. L. Bredas, J. Am. Chem. Soc. 1994, 116, 10703 – 10714. 729 – 737.
[6] G. Archetti, A. Abbotto, R. Wortmann, Chem. Eur. J. 2006, 12, 7151 – [32] A. Baroudi, A. El-Hellani, A. A. Bengali, A. S. Goldman, F. Hasanayn, Inorg.
7160. Chem. 2014, 53, 12348 – 12359.
[7] N. A. Murugan, K. Aidas, J. Kongsted, Z. Rinkevicius, H. Agren, Chem. [33] B. Marten, K. Kim, C. Cortis, R. A. Friesner, R. B. Murphy, M. N. Ringnalda,
Eur. J. 2012, 18, 11677 – 11684. D. Sitkoff, B. Honig, J. Phys. Chem. 1996, 100, 11775 – 11788.
[8] a) D. A. Estrin, L. M. Baraldo, L. D. Slep, B. C. Barja, J. A. Olabe, L. Paglieri, [34] C. Hansch, A. Leo, R. W. Taft, Chem. Rev. 1991, 91, 165 – 195.
G. Corongiu, Inorg. Chem. 1996, 35, 3897 – 3903; b) B. Lecea, A. Arrieta, [35] I. Gçttker-Schnetmann, P. White, M. Brookhart, J. Am. Chem. Soc. 2004,
F. P. Cossio, J. Org. Chem. 1997, 62, 6485 – 6492. 126, 1804 – 1811.
[9] a) B. R. James, Homogeneous Hydrogenation, Wiley, New York, 1973; b) [36] J. C. Grimm, C. Nachtigal, H.-G. Mack, W. C. Kaska, H. A. Mayer, Inorg.
M. Peruzzini, R. Poli, Recent Advances in Hydride Chemistry, Elsevier, Am- Chem. Commun. 2000, 3, 511 – 514.
sterdam, 2001. [37] Y. Segawa, M. Yamashita, K. Nozaki, J. Am. Chem. Soc. 2009, 131, 9201 –
[10] a) G. J. Kubas, Metal Dihydrogen and s-Bond Complexes, Kluwer Academ- 9203.
ic/Plenum Publishers, New York, 2001; b) G. J. Kubas, J. Organomet. [38] D. F. MacLean, R. McDonald, M. J. Ferguson, A. J. Caddell, L. Turculet,
Chem. 2009, 694, 2648 – 2653. Chem. Commun. 2008, 5146 – 5148.
[11] C. E. Webster, D. A. Singleton, M. J. Szymanski, M. B. Hall, C. Zhao, G. Jia, [39] C. Tsay, B. N. Livesay, S. Ruelas, J. Y. Yang, J. Am. Chem. Soc. 2015, 137,
Z. Lin, J. Am. Chem. Soc. 2001, 123, 9822 – 9829. 14114 – 14121.
[12] a) R. Gelabert, M. Moreno, J. M. Lluch, A. Lleds, V. Pons, D. M. Heinekey, [40] During the later stages of the preparation of this manuscript, a theoreti-
J. Am. Chem. Soc. 2004, 126, 8813 – 8822; b) R. Gelabert, M. Moreno, cal study appeared describing the solvent effect on the geometry of
J. M. Lluch, A. Lleds, D. M. Heinekey, J. Am. Chem. Soc. 2005, 127, [RuII(SC6H3Me2-2,6-k1S)2(PMe3)3], in which the authors suggested that
5632 – 5640; c) R. Gelabert, M. Moreno, J. M. Lluch, Chem. Eur. J. 2005, the role of non-specific solvation may be significant: A. L. Pitts, M. B.
11, 6315 – 6325. Hall, Organometallics 2015, 34, 3129 – 3140.
[13] a) R. Gelabert, M. Moreno, J. M. Lluch, A. Lleds, J. Am. Chem. Soc. 1997, [41] A. V. Polukeev, R. Gritcenko, K. J. Jonasson, O. F. Wendt, Polyhedron
119, 9840 – 9847; b) R. Gelabert, M. Moreno, J. M. Lluch, A. Lledos, J. Am. 2014, 84, 63 – 66.
Chem. Soc. 1998, 120, 8168 – 8176. [42] Jaguar, version 7.6, Schrçdinger, LLC, New York, NY, 2009, www.schro-
[14] J. Choi, A. H. R. MacArthur, M. Brookhart, A. S. Goldman, Chem. Rev. dinger.com.
2011, 111, 1761 – 1779. [43] a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648 – 5652; b) C. Lee, W. Yang,
[15] a) S. Niu, M. B. Hall, J. Am. Chem. Soc. 1999, 121, 3992 – 3999; b) K. R. G. Parr, Phys. Rev. B 1988, 37, 785 – 789.
Krogh-Jespersen, M. Czerw, M. Kanzelberger, A. S. Goldman, J. Chem. Inf.
Comput. Sci. 2001, 41, 56 – 63.
[16] I. Gçttker-Schnetmann, D. M. Heinekey, M. Brookhart, J. Am. Chem. Soc. Received: December 22, 2015
2006, 128, 17114 – 17119. Published online on && &&, 0000

Chem. Eur. J. 2016, 22, 1 – 10 www.chemeurj.org 9  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim & &

These are not the final page numbers! ÞÞ

 Full Paper

FULL PAPER
& Solvent Effects Controlled geometries: Dramatic non-
specific solvation-induced geometry
A. V. Polukeev, R. Marcos, changes within iridium pincer com-
M. S. G. Ahlquist, O. F. Wendt* plexes are reported. These compounds
&& – && tend to adopt trigonal-bipyramidal geo-
metries in non-polar media and square-
Solvent-Dependent Structure of pyramidal geometries in polar media
Iridium Dihydride Complexes: (see scheme).
Different Geometries at Low and High
Dielectricity of the Medium

& & Chem. Eur. J. 2016, 22, 1 – 10 www.chemeurj.org 10  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!