How do physical-chemical parameters influence the catalytic hydrogenation of 1,3-cyclohexadiene in ionic liquids?

Article abstract of DOI:10.1021/jp102941n

The catalytic hydrogenation of 1,3-cyclohexadiene using [Rh(COD)(PPh ₃)₂]NTf₂ (COD = 1,5-cyclooctadiene) was performed in two ionic liquids: 1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide, [C₁C₄Im][NTf ₂], and 1-butyl-2,3-dimethylimidazolium bis(trifluoromethylsulfonyl) imide, [C₁C₁C₄Im][NTf₂]. It is observed that the reaction is twice as fast in [C₁C ₄Im][NTf₂] than in [C₁C₁C ₄Im][NTf₂]. To explain the difference in reactivity, molecular interactions and the microscopic structure of ionic liquid +1,3-cyclohexadiene mixtures were studied by NMR and titration calorimetry experiments, and by molecular simulation in the liquid phase. Diffusivity and viscosity measurements allowed the characterization of mass transport in the reaction media. We could conclude that the diffusivity of 1,3-cyclohexadiene is 1.9 times higher in [C₁C₄Im][NTf₂] than in [C₁C₁C₄Im][NTf₂] and that this difference could explain the lower reactivity observed in [C₁C ₁C₄Im][NTf₂].

Full text of DOI:10.1021/jp102941n

8156
J. Phys. Chem. B 2010, 114, 8156–8165
How do Physical-Chemical Parameters Influence the Catalytic Hydrogenation of
1,3-Cyclohexadiene in Ionic Liquids?
Paul S. Campbell,† Ajda Podgorsˇek,‡ Thibaut Gutel,† Catherine C. Santini,*,†
Ag´ılio A. H. Pa´dua,‡,§ Margarida F. Costa Gomes,*,‡,§ Franc¸ois Bayard,† Bernard Fenet,| and
Yves Chauvin†
UniVersité de Lyon Institut de Chimie de Lyon, UMR 5265 CNRS-UniVersite´ de Lyon 1-ESCPE Lyon,
C2P2, Equipe Chimie Organome´tallique de Surface, ESCPE 43 BouleVard du 11 NoVembre 1918,
F-69616 Villeurbanne, France, Laboratoire de Thermodynamique et Interactions Mole´culaires, CNRS, UMR
6272 24 AVenue des Landais, 63177 Aubie`re, France, Clermont UniVersite´ and UniVersite´ Blaise Pascal,
Clermont-Ferrand, France, and Centre Commun de RMN, UCB Lyon 1 - ESCPE Lyon, 43 BouleVard du 11
NoVembre 1918, F-69626 Villeurbanne Cedex, France
ReceiVed: April 1, 2010; ReVised Manuscript ReceiVed: May 11, 2010
The catalytic hydrogenation of 1,3-cyclohexadiene using [Rh(COD)(PPh3)2]NTf2 (COD ) 1,5-cyclooctadiene)
was performed in two ionic liquids: 1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide,
[C1C4Im][NTf2], and 1-butyl-2,3-dimethylimidazolium bis(trifluoromethylsulfonyl)imide, [C1C1C4Im][NTf2].
It is observed that the reaction is twice as fast in [C1C4Im][NTf2] than in [C1C1C4Im][NTf2]. To explain the
difference in reactivity, molecular interactions and the microscopic structure of ionic liquid +1,3-cyclohexadiene
mixtures were studied by NMR and titration calorimetry experiments, and by molecular simulation in the
liquid phase. Diffusivity and viscosity measurements allowed the characterization of mass transport in the
reaction media. We could conclude that the diffusivity of 1,3-cyclohexadiene is 1.9 times higher in
[C1C4Im][NTf2] than in [C1C1C4Im][NTf2] and that this difference could explain the lower reactivity observed
in [C1C1C4Im][NTf2].
1. Introduction
solutes are solvated in the ionic liquid:14 nonpolar species will
be solvated within the hydrophobic domains whereas polar
substrates tend to interact with the polar network. The strength
of cation-anion association15 can also influence the possibility
of establishing ion-solute specific interactions (e.g., π-cation
interactions), as has been shown in the case of toluene.9 For a
nonaromatic π-system, such as 1,3-cyclohexadiene, solvation
is probably a combination of more subtle interactions with the
ionic liquid. This information can be assessed through the study
of the thermodynamic properties of solution (solubility, enthalpy
of solution) and from the characterization of the mass transfer
through viscosity and diffusivity data.
The aim of this work is to study the influence of the nature
of the ionic liquid on the catalytic hydrogenation of 1,3-
cyclohexadiene (CYD) with [Rh(COD)(PPh3)2]NTf2 (COD )
1,5-cyclooctadiene). First, we have quantified the difference in
reactivity when the reaction is performed in two ionic liquids:
1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide,
[C1C4Im][NTf2], and 1-butyl-2,3-dimethylimidazolium bis(tri-
fluoromethylsulfonyl)imide, [C1C1C4Im][NTf2]. Following this,
we have attempted to rationalize the differences encountered
through NMR characterization of the molecular interactions
between the ionic liquids and the substrate, the study of the
microscopic structure of the IL-substrate mixtures, the measure-
ment of the thermodynamic properties of mixing (such as the
solubility and the heat of mixing), the determination of the
viscosity of the reaction media, and the measurement of
diffusivity of CYD in the ionic liquids.
The application of ionic liquids (ILs) in catalyzed reactions
is of increasing importance,1,2 in particular, they provide
excellent media for conducting catalytic hydrogenations.1-3
Several research groups report differences in the rates and
selectivities in ILs when compared to corresponding reactions
in molecular solvents.2
As reaction media, ILs have specific properties that may have
consequences on the catalytic process. In some catalytic
reactions, the difference between the use of ILs and traditional
solvents has been related to the chemical role of ILs, which can
serve as new ligands for the catalytic metal center, as catalyst
activators, as cocatalysts, or even as catalysts themselves.1,4,5 In
other cases, differences in reactivity have a physical-chemical
origin,1,6 resulting from peculiar solvation phenomena including
specific interactions between the IL and the substrate (H-bonds,
cation-π),7-9 mass transfer factors (viscosity, diffusivity),1 and
effects attributed to the highly structured nature of ILs.10-12
ILs have a high degree of self-organization due to the
coexistence of charged moieties and hydrophobic alkyl chains,
to the importance of both Coulombic and van der Waals
interactions, and frequently to the presence of hydrogen bonds
between the cation and the anion. In butylimidazolium ionic
liquids, the alkyl side chain is sufficiently long to allow the
formation of nonpolar domains that coexist with an ionic
network.13 This heterogeneous structure influences the way
* To whom correspondence should be addressed. E-mail: C.C.S., santini@
cpe.fr; M.F.C.G., margarida.costa-gomes@univ-bpclermont.fr.
† Universite´ de Lyon.
2. Experimental Section
‡ CNRS.
§ Clermont Universite´ and Universite´ Blaise Pascal.
| Centre Commun de RMN.
Materials. 1-Methylimidazole (>99%, Aldrich) and 1,2-
dimethylimidazole (>98%, Aldrich) were distilled prior to use.
10.1021/jp102941n  2010 American Chemical Society
Published on Web 05/26/2010

Catalytic Hydrogenation of 1,3-Cyclohexadiene
J. Phys. Chem. B, Vol. 114, No. 24, 2010 8157
Anhydrous 1,3-cyclohexadiene (99.8%, Aldrich) and 1,3-
cyclohexadiene stabilized (96%, Acros Organics, stabilized with
50 ppm of 2,6-di-tert-butyl-4-methylphenol, BHT) distilled over
NaK alloy and stored on zeolites were used for the reaction
and the physical chemical measurements, respectively. Bis(tri-
fluoromethanesulfonyl)imide lithium salt (>99%, Solvionic) and
[Rh(COD)Cl]2 (>99%, Strem) were used without further puri-
fication. All other reagents and solvents were commercially
available and were used as received. Ionic liquids, synthesized
as previously reported,16 were dried overnight under high
vacuum and stored in a glovebox (Jacomex) to guarantee
rigorously anhydrous products.
in presence of toluene as an internal standard. A HP6890
chromatograph equipped with FID detector and an Al2O3/KCl
column (L ) 50 m, φint ) 0.32 mm, film thickness ) 5 µm)
was used. The injector and detector temperatures were set to
230 °C. Samples were injected in a volume of 1 µL. The
temperature of the column was fixed at 190 °C (see Supporting
Information for calculations).
Solubility and Phase Diagrams. Liquid-liquid phase equi-
libria of the mixture of [C1C4ImNTf2] or [C1C1C4Im][NTf2] and
CYD at atmospheric pressure were determined using a dynamic
method with visual detection of solution turbidity, as already
described.18
Synthesis of the Catalysts. [Rh(COD)(PPh3)2]NTf2. The
procedure followed was adapted from the literature.17 A mixture
of [Rh(COD)Cl]2 (100 mg, 0.2 mmol) dissolved in 2 mL of
dichloromethane and of LiNTf2 (107 mg, 0.37 mmol) dissolved
in 2 mL of water was stirred vigorously while triphenylphos-
phine (405 mg, 1.5 mmol) was added. After 2 h, the dichlo-
romethane layer was removed, washed three times with 2 mL
of water and dried over anhydrous Na2SO4. Ethanol (1 mL) and
diethyl ether (2 mL) were slowly added to complete crystal-
lization. The orange crystals were filtered off and dried under
The mixtures of ionic liquid and CYD at different composi-
tions were prepared gravimetrically in a glass vial equipped with
a stirring bar. First, the ionic liquid was introduced into a glass
vial, then the appropriate amount of CYD was added and the
vial was sealed. To minimize the volume of the vapor phase in
equilibrium with the ionic liquid solution and to reduce the error
in composition due to differential evaporation, the glass vial
was almost completely filled with the mixture. The uncertainty
of the mole fraction is estimated as (0.0001. The cells were
then immersed in a thermostatic water bath whose temperature
was monitored using a platinum resistance thermometer with a
precision of (0.1 K. The temperature of the bath was first
increased slowly until one phase was observed. The clear
homogeneous system was then cooled very slowly (5 K/h) under
continuous stirring. The temperature at which the first sign of
turbidity (first cloudiness) appeared was considered as the
temperature of the liquid-liquid phase transition. The overall
accuracy in the measurement of cloud-point temperatures is
estimated to be (2 K.
1
reduced pressure. Yield: 200 mg (99%). H NMR (300 MHz,
CD2Cl2): δ ) 2.2 (m, 8H), 4.6 (s, 4H), 7.5 (m, 30H). 31P NMR
(121 MHz, CD2Cl2): δ ) 23.6 with JRh-P ) 147 Hz. 103Rh NMR
(500 MHz, CD2Cl2): δ ) -80 ppm with JRh-P ) 147 Hz. Mass
spectrometry: positive mode: m/z ) 473 w Rh(COD)(PPh3)+;
627 w Rh(PPh3)2+. UV-vis spectroscopy (6.6 mg of
[Rh(COD)(PPh3)2]NTf2 dissolved in 2 mL of one of ionic
liquids, [C1C4Im][NTf2] or [C1C1C4Im][NTf2]); spectrum re-
corded at λmax ) 450 nm.
Isothermal Titration Calorimetry. The heat effects resulting
from mixing aliquots of CYD with the ionic liquid were
measured at 303.15 K using an isothermal titration nanocalo-
rimeter equipped with 4 mL glass cells in a Thermal Activity
Monitor TAM III from TA Instruments. An electrical calibration
was done before each experiment, and the instrument was
chemically calibrated 5 times by titration of a 0.01 M aqueous
solution of 18-crown-6 ethers with an 0.2 M aqueous solution
of BaCl2. The enthalpies of binding of Ba2+ ions to 18-crown-6
were found to be slightly higher than those reported in the
literature, 2.5%19 and 1.6%,20,21 respectively. No correction
attributable to these differences was introduced in the raw data.
[Rh(CYD)(PPh3)2]NTf2. [Rh(COD)(PPh3)2]NTf2 (100 mg,
9.84 × 10-5 mol) was dissolved in 1 mL of CYD. After 10
min, the excess of solvent was removed under reduced
pressure. Yield: 100 mg (99%). 1H NMR (300 MHz, CD2Cl2):
δ ) 2.1 (s, 2H), 5.3 (s, 2H), 7.5 (m, 30H). 31P NMR (121
MHz, CD2Cl2): δ ) 25.2 ppm with JRh-P ) 17 Hz. Mass
spectrometry: positive mode: m/z ) 707 w Rh(cyclohex-
adiene)(PPh3)2+; m/z ) 627 w Rh(PPh3)2+. UV-vis spec-
troscopy (6.6 mg of [Rh(CYD)(PPh3)2]NTf2 dissolved in 2 mL of
one of ionic liquids, [C1C4Im][NTf2] or [C1C1C4Im][NTf2]);
spectrum recorded at λmax ) 500 nm.
Ligand Exchange. In [Rh(CYD)(PPh3)2]NTf2 was followed
by UV-vis spectroscopy. CYD (0.15 mL, 1.6 mmol) was added
to a system of [Rh(COD)(PPh3)2]NTf2 (1.6 mg, 1.6 µmol) in
one of the ionic liquids, [C1C4Im][NTf2] or [C1C1C4Im][NTf2]
(1 mL). The UV-visible spectra of 1 mL of the solution were
recorded on the Perkin-Elmer LAMBDA 950 Spectrophotometer
in stirred and closed UV cells at λmax ) 500 nm every second.
Reaction with H2. [Rh(CYD)(PPh3)2]NTf2 (8 mg, 8.1 × 10-6
mmol) was dissolved in one of the ionic liquids, [C1C4Im][NTf2]
or [C1C1C4Im][NTf2] (5 mL) and pressurized under 1.2 bar of
hydrogen. The UV-visible spectra of 1 mL of the solution were
recorded at a given time at λmax ) 500 nm.
Hydrogenation of 1,3-Cyclohexadiene. The hydroge-
nation of CYD was carried out at 1.2 atm of H2 and 30 °C.
CYD (0.15 mL, 1.6 mmol) was dissolved in a system of
[Rh(COD)(PPh3)2]NTf2 (3.2 mg, 3.2 µmol) in one of ionic
liquids, [C1C4Im][NTf2] or [C1C1C4Im][NTf2] (1 mL), under
argon resulting in red homogeneous solutions. The reaction
mixture was kept under a hydrogen atmosphere (1.2 atm,
constant pressure) until 4 mL of acetonitrile was added to the
catalytic solution. The product distribution in the reaction
mixture and the conversion were determined by GC analyses
Approximately 2.75 mL of degassed ionic liquid were
introduced into 4 mL glass measuring and reference cells. The
liquid in the measuring cell was stirred by a turbine stirrer at
160 rpm and volumes of 4 µL of CYD were injected during
180 s using a motor driven pump (Thermometric 3810 Syringe
Pump) equipped with a 100 µL gastight Hamilton syringe. In
all experiments, the intervals between consecutive injections
were 35-40 min, which provided a good thermal stabilization
of the ionic liquid solution and the return to a stable baseline.
To minimize the undesirable effects of diffusion of the ionic
liquid into the canula linking the CYD syringe to the cell, the
canula was immersed in the sample 10 min prior the first
injection.
A peak with an area proportional to the resulting heat effect
Qi translates to the thermal effect due to each injection of CYD.
The integration of peaks from the recorded calorimetric plots
was performed using the TAM III Assistant software. Each
experiment was repeated four times to obtain reproducible values
of Qi at different concentrations within the error bar of (2%.
Density and Viscosity. The mixtures of ionic liquid and CYD
at different compositions were prepared gravimetrically follow-

8158 J. Phys. Chem. B, Vol. 114, No. 24, 2010
Campbell et al.
ing the procedure already described, including the precautions
to minimize vapor headspace.9 The viscosity of the mixture was
measured at 298.15 K (controlled to within (0.005 K and
measured with the accuracy better than (0.05 K) using a rolling
ball viscometer from Anton Paar, model AMVn, equipped with
capillary tubes of 3.0 and 1.8 mm in diameter. Before starting
the measurements, the 3.0 mm diameter capillary tube was
calibrated as a function of temperature and angle of measure-
ment with a standard viscosity oil from Cannon (N35). The 1.8
mm diameter tube was calibrated with water by the manufac-
turer. The overall uncertainty of the viscosity is estimated as
(2.0%.
gradients to improve line shape and trapezoidal gradients (the
pulse sequences shown in Figure S-6 in Supporting Information).
The diffusion evolution time was 100 ms, the constant amplitude
part of the gradient was 3 ms, and the cosine raising and falling
part of gradient were 150 µs. The diffusion space was sampled
by 32 linearly spaced gradients.
Molecular Simulation. The microscopic structures of the IL-
CYD mixtures studied experimentally were also investigated
by molecular simulation, using an atomistic force field that
describes interactions and conformations.23-25 CYD was rep-
resented by the optimized potential for the liquid simulations
force field in its all-atom explicit version (OPLS-AA).26 ILs
were represented by a specifically parametrized force field of
the OPLS-AA family in which particular attention was paid to
the description of electrostatic charge distributions and torsion
energy profiles. The OPLS-AA force field is known to reproduce
H-bonds well; the electron density of aromatic systems is
represented by the values of electrostatic charges on the relevant
atoms that account for the molecular mutipoles, combined with
the Lennard-Jones sites that account for dispersion interactions.
Explicit polarization of electron clouds is not included in the
present model. Although this may be important to correctly
reproduce dynamic properties of ILs, structural features and
thermodynamic quantities have been described to equivalent
levels of accuracy using fixed-charge models.27
Molecular dynamics simulations of condensed-phase CYD-
[C1C4Im][NTf2] and CYD-[C1C1C4Im][NTf2] mixtures were
performed using the DL_POLY program.28 System sizes were
chosen so as to contain about 10 000 atoms, and so the numbers
of cations, anions, and CYD molecules varied according to
composition (e.g., 128 ion pairs and 64 molecules of CYD for
R ) 0.5). Initial low-density configurations, with ions and
molecules placed at random in periodic cubic boxes, were
equilibrated to attain liquid-like densities and structures at 400
K and 1 bar. Temperature and pressure were maintained using
a Nose´-Hoover thermostat and barostat, respectively. Produc-
tion runs then took 500 ps with an explicit cutoff distance of
16 Å for nonbonded interactions, and long-range corrections
applied for repulsive-dispersive interactions. Electrostatic ener-
gies were calculated using the Ewald summation method with
a relative accuracy of 10-4. Structural quantities such as radial
and spatial distribution functions were calculated from configu-
rations generated during the production runs.
The densities of the mixtures, necessary to calculate the
viscosities were measured in an Anton Paar vibrating tube
densimeter model 512 P, at 298.15 K (measured by a calibrated
PRT with an accuracy of (0.02 K). The densimeter was
calibrated using n-heptane, bromobenzene, and 2,4-dichloro-
toluene. The overall uncertainty of the density is estimated as
(0.01%.
1
NMR Spectroscopy. H, 13C, and 31P solution NMR data
were collected at room temperature on a Bruker AC 300 MHz
spectrometer with the resonance frequency at 300.130 MHz for
the 1H nucleus. 103Rh solution NMR was carried out on a Bruker
DRX 500 instrument at 298 K (nominal) with a resonance
frequency at 500.130 MHz. The solvent used (CD2Cl2) was
distilled and kept in a rotaflo with molecular sieves. Chemical
shifts are reported in ppm (singlet ) s, doublet ) d, doublet of
doublet ) dd, and multiplet ) m) and were measured relative
to the residual proton of the solvent to CHDCl2 for 1H, to CD2Cl2
for 13C, and to H3PO4 for 31P spectra.
For 1H 1D NMR spectroscopy, the samples with molar ratio
R ) 0.5 (R is the molar ratio between the amount of substance
of the hydrocarbon and the amount of substance of IL) were
prepared. The mixtures of ionic liquids and hydrocarbon were
prepared in closed vials in a glovebox by adding the appropriate
amount of CYD to each ionic liquid, [C1C4Im][NTf2] or
[C1C1C4Im][NTf2]. The resulting systems were stirred for 24 h
at 303 K, resulting in homogeneous monophasic solutions.
Approximately 0.3 mL of the sample was then introduced into
a 5 mm NMR tube. A stem coaxial capillary tube loaded with
CD2Cl2 was inserted into the 5 mm NMR tube to avoid any
contact between the deuterated solvent and the analyzed mixture.
The deuterium in CD2Cl2 was used for the external lock of the
NMR magnetic field and the residual CHDCl2 in CD2Cl2 was
used as the 1H NMR external reference at 5.32 ppm. When 1H
NMR data are obtained in this way, the reference signal of
CHDCl2 remains constant and is not affected by changes in
sample concentration.
3. Results and Discussion
Hydrogenation of 1,3-Cyclohexadiene. Selective hydroge-
nation of CYD into cyclohexene can be carried out, with good
conversion rates, using Osborn’s complex [Rh(NBD)(PPh3)2]PF6
(NBD ) norbornadiene) in IL media such as [C1C4Im][SbF6]
or [C1C4Im][PF6].29 To avoid impurities such as chloride and
water, ILs based on the bis(trifluoromethylsulfonyl)imide anion,
[C1C4Im][NTf2] and [C1C1C4Im][NTf2], were chosen here since
they are hydrophobic and liquid at room temperature and their
purification is well controlled.16 Anion exchange between the
IL and the catalyst can be circumvented by replacing Osborn’s
complex with [Rh(COD)(PPh3)2]NTf2 (COD ) 1,5-cycloocta-
diene).
In the hydrogenation experiment the appropriate quantity of
CYD was added to the yellow solution of [Rh(COD)(PPh3)2]-
NTf2 in [C1C4Im][NTf2] or [C1C1C4Im][NTf2] to reach a molar
ratio between CYD and IL R ) 0.5, corresponding to a molar
ratio substrate/Rh atom r ) 500. The resulting red solution was
stirred under 1.2 bar of hydrogen at 303 K. The formation of
cyclohexene (CYE) was followed by GC analysis. Note that
For ROESY (nuclear overhauser effect spectroscopy) experi-
ments in the rotating frame, the 2D sequence was built with
the scheme proposed by Bodenhausen (the pulse sequence
shown in Figure S-3 in Supporting Information).22 The mixing
time (200 ms) is split into two parts separated by a p-pulse. At
each side of the spin lock the B1 field is ramped linearly (4.5
ms) to ensure adiabatic conditions for spin lock. During the
first spin lock pulse, the frequency is shifted to O1 + Df whereas
the second pulse frequency is set to O1 - Df; O1 is the offset
frequency and Df set to give a B1 field at the magic angle. In
this way, the ROESY response is roughly constant across the
spectra of interest. 1D sequence PFGSE (pulse field gradient
spin echo for selective excitation) has been used, and spin lock
followed the same scheme as previously.
For the 2D DOSY (diffusion order spectroscopy) experiments
a Bruker sequence ledbpgp2s was implemented for shorter pulse

Catalytic Hydrogenation of 1,3-Cyclohexadiene
J. Phys. Chem. B, Vol. 114, No. 24, 2010 8159
Figure 1. Hydrogenation of CYD at 303 K under 1.2 bar H2 (R ) 0.5
Figure 2. Evolution as a function of time of the ligand exchange
reaction COD-CYD for [Rh(COD)(PPh3)2]NTf2 in [C1C4Im][NTf2]
(full line) and [C1C1C4Im][NTf2] (dashed line) monitored at λ ) 500
nm by UV-vis spectroscopy. UV-visible spectra of [Rh(COD)(PPh3)2]-
NTf2 and [Rh(CYD)(PPh3)2]NTf2 are provided in the Supporting
Information (Figures S-1 and S-2).
and
r ) n(CYD)/n(Rh) ) 500) in [C1C4Im][NTf2] ([) and
[C1C1C4Im][NTf2] (b).
CYE is less soluble than CYD in both IL media so the system
will tend to become biphasic during the course of the reaction.
To avoid errors in the determination of the composition of the
reaction mixture, resulting from heterogeneity of the system,
the conversion was measured using GC analysis of the entire
reaction system after dissolution in a mixture of acetonitrile and
toluene 99:1. Each point, presented in Figure 1, corresponds to
a different experiment.
As can be seen in Figure 1, hydrogenation of CYD in both
ILs by [Rh(COD)(PPh3)2]NTf2 leads quantitatively and selec-
tively to CYE. In both cases, no significant amount (less than
2% at high or complete conversion) of cyclohexane (CYH) was
detected by GC. As described in the literature, the rate of CYD
reduction remains constant until near 100% conversion (1 mol
of H2 absorbed per CYD) and CYE is produced quantitatively
before it is hydrogenated to CYH.17 In [C1C4Im][NTf2] the
conversion is complete in 2 h. This result is similar to that
observed in organic solvents17 and in other ionic liquids, such
as [C1C4Im][PF6] or [C1C4Im][SbF6].29 In [C1C1C4Im][NTf2],
only 50% conversion is reached after 2 h with an initial rate of
hydrogenation ca. half of that in [C1C4Im][NTf2].
In organic solvents, during the hydrogenation of 1,3-cyclo-
hexadiene, the catalyst first undergoes a ligand exchange,
yielding [Rh(CYD)(PPh3)2]+. This occurs more rapidly than the
rate at which it then reacts with H2.30 The two reaction steps
were investigated in both ILs. First, [Rh(CYD)(PPh3)2]NTf2 was
synthesized from [Rh(COD)(PPh3)2]NTf2 and fully character-
ized. As [Rh(COD)(PPh3)2]NTf2 is yellow (λmax ) 450 nm) and
[Rh(CYD)(PPh3)2]NTf2 is red (λmax ) 500 nm), the ligand
exchange COD-CYD could be monitored by UV-vis spectros-
copy (λ ) 500 nm) in both ILs, as shown in Figure 2. This
reaction is also much slower in [C1C1C4Im][NTf2] than in
[C1C4Im][NTf2].
Secondly, the red solutions of [Rh(CYD)(PPh3)2]NTf2 in
[C1C4Im][NTf2] or [C1C1C4Im][NTf2] were exposed to hydrogen
and the color disappeared with the concomitant formation of
CYE. This hydrogenation reaction of CYD in [Rh(CYD)(PPh3)2]-
NTf2 was also monitored by UV-vis spectroscopy in both ILs,
the absorbances being represented in Figure 3. The discoloration
of the medium, indicating the hydrogenation step, is also faster
in [C1C4Im][NTf2] than in [C1C1C4Im][NTf2]. These results
indicate that both ligand exchange and hydrogenation reactions
are faster in [C1C4Im][NTf2]. These processes, however, do not
occur on the same time scale (hours versus seconds) and, as in
Figure 3. Evolution of the absorbance at λ ) 500 nm as a function of
time during the hydrogenation reaction of CYD in the presence of
[Rh(CYD)(PPh3)2]NTf2 in [C1C4Im][NTf2] ([) and [C1C1C4Im][NTf2]
(b).
organic solvents, ligand exchange is faster than hydrogenation.
More importantly, the rate at which the CYE is produced (Figure
1) is similar to the rate of discoloration due to attack of
[Rh(CYD)(PPh3)2] by H2 (Figure 3). Consequently, in ac-
cordance with the literature,30 in the hydrogenation of CYD,
this attack is the rate-determining step.
As solubilities of H2 are similar in both ionic liquids,31 the
same overall rate of reaction would be expected. However, this
is not the case and could be attributed to differences in the
availability of H2 in the different media. Indeed, even though
no literature data are available for H2 in ILs, it is generally found
that the diffusion of gases is inversely dependent on the viscosity
of the medium.32 Because the same difference in the rate of
reaction is noted in the ligand exchange, where H2 is not
involved, the differences in the inherent properties of each IL
must be considered. For instance, studies by molecular simula-
tions and NMR of solvation and molecular structure of similar
mixtures showed that according to the nature of the IL, specific
interactions with the solute differ.7-9 In addition, mass transport
factors (viscosity, diffusivity) have also been related to the

8160 J. Phys. Chem. B, Vol. 114, No. 24, 2010
Campbell et al.
Figure 4. Liquid-liquid equilibrium diagrams for the mixture of CYD
and [C1C4Im][NTf2] (b) or [C1C1C4Im][NTf2] ([) and solubility of
toluene in [C1C4Im][NTf2] (]) or [C1C1C4Im][NTf2] (O).
Figure 5. Partial molar excess enthalpies vs mole fraction of CYD in
the binary mixture with [C1C4Im][NTf2] ([) or [C1C1C4Im][NTf2] (b)
at 303.15 K. The lines represent functions: HECYD ) 1.6285 +
0.2481xCYD (b) and HECYD ) 0.8574 - 0.1518xCYD ([).
reactivity1,33 and therefore are also important parameters that
we will determine and discuss here.
molar enthalpy. We have done so for both systems, CYD-
[C1C4Im][NTf2] and CYD-[C1C1C4Im][NTf2]. The excess
molar enthalpy of mixing ∆HEmix was determined by isothermal
titration calorimetry, from the heat effect involved in injections
of small quantities of CYD into the ionic liquid, QCYD. The
partial molar excess enthalpy of solute, HECYD, was calculated
according to eq 1.
Solvation of 1,3-Cyclohexadiene. Figure 4 shows cloud-point
temperatures as a function of composition for [C1C4Im][NTf2]-
CYD and [C1C1C4Im][NTf2]-CYD.18,34 The observation of the
cloud point with decreasing temperature was difficult, especially
for the [C1C1C4Im][NTf2]-CYD system, because in this case
the mixtures often remained in a metastable state and cloudiness
persisted despite overheating. The observation of the cloud point
while decreasing the temperature was not reproducible during
this experiment, mainly because of the effect of precooling and
kinetics of demixing, as already discussed,33 and consequently
the solubility was determined at only two temperatures. Deter-
mination of the solubility of CYD in ionic liquids by liquid-vapor
equilibria measurements using a static apparatus35,36 was not
possible due to the chemical properties of CYD, such as its
tendency to polymerize37 and its compatibility with o’ring
materials.
The solubility of CYD at 303.15 K, expressed in mole fraction
is 0.42 in [C1C4Im][NTf2] and 0.34 in [C1C1C4Im][NTf2] (see
Figure 4), which could indicate less favorable interactions
between CYD and the [C1C1C4Im] ion in comparison with the
[C1C4Im] ion. In the temperature range studied, both systems
CYD-[C1C4Im][NTf2] and CYD-[C1C1C4Im][NTf2] show a
behavior compatible with the existence of upper critical solution
temperatures and very steep liquid-liquid equilibrium lines.
With this method the solubility of the ionic liquid in CYD could
not be detected, as it lies below the limit of detection. From
measurements and from calculations using COSMO-RS,38 it is
known that the concentration of hydrocarbons in the ionic liquid
rich phase is very low, with mole fractions on the order of 10-4.
In comparison with aromatic molecules in these ILs, CYD is
much less soluble. For instance, the solubility of toluene at
303.15 K expressed in mole fraction is 0.73 in [C1C4Im][NTf2]
and 0.72 in [C1C1C4Im][NTf2],9 the solubility of benzene in
[C1C4Im][NTf2] being even higher (mole fraction of 0.78 at
303.15 K).39 The small difference in solubility of toluene with
respect to benzene can be attributed to the presence of an
additional methyl group, whereas the much lower solubility of
CYD is explained by less favorable interactions between the
solute and the ILs. Disruption of the aromatic system changes
the interactions between solute and ionic liquid and thus
decreases the solubility.
∂∆HEmix
∂nCYD
QCYD
HECYD
)
≈
(1)
(
)
∆nCYD
nIL,p,T
where nCYD and nIL denote the quantity of CYD and IL,
respectively, ∆HEmix is the excess molar enthalpy of the entire
system (enthalpy of mixing) and ∆nCYD is the quantity of solute
per injection. ∆nCYD was calculated from the injected volumes
and the density of CYD was obtained from the literature.40 In
these calculations, heat due to evaporation of the solute from
IL solution is assumed to be negligible. Hence, no correction
for the vapor pressure of the solute was made.
Figure 5 represents partial molar excess enthalpies of CYD
in both ILs at 303.15 K as a function of composition. A larger
dispersion of the values (up to 3%), especially at higher
concentration, was observed. At times, before each injection of
CYD into the IL, an exothermic effect was detected, mainly
originating from insufficient mixing of both components at the
beginning of the injection period. As already discussed in the
literature41 in the case of aromatic hydrocarbons, the rate of
dissolution of nonpolar, low density and low viscosity CYD in
the IL is slow and sometimes leads to formation of a solute-
rich layer on the surface of the solution. This is often
accompanied by partial evaporation or even polymerization of
CYD before the mixing process is complete, hence disturbing
the measurements. A large number of injections were made to
ensure reliability. [The formation of a polymer was observed
and confirmed by NMR analysis.37 This side reaction is
attributed to the fact that CYD has been distilled to eliminate
the stabilizer BHT before the physical chemical measurements.]
The dependence of partial molar excess enthalpy of CYD in
[C1C4Im][NTf2] and [C1C1C4Im][NTf2] on the mole fraction was
approximated by a linear regression, eq 2 and Figure 5. The
The energy involved in the interactions of different solutes
dissolved in ILs can be assessed by measuring the excess
partial excess molar enthalpies of CYD in both ILs at infinite
E,∞
dilution, H , were obtained from eq 2 by setting xCYD ) 0
CYD

Catalytic Hydrogenation of 1,3-Cyclohexadiene
J. Phys. Chem. B, Vol. 114, No. 24, 2010 8161
ane in [C1C2Im][NTf2]. As expected, the values of ∆HEmix for
CYD obtained here are intermediate between the values for
cyclohexene and benzene. Comparison of [C1C4Im][NTf2] and
[C1C1C4Im][NTf2] shows that introduction of an additional
methyl group on the C2 carbon of the imidazolium ring strongly
affects the interaction with molecules of CYD, which becomes
less favorable. To rationalize the differences in the solvation of
CYD in both ILs, and to complement the enthalpic data with
structural information, the microscopic structure of the mixtures
was investigated using NMR and molecular simulation.
Molecular Structure of the Reaction Media. The molecular
structure of the mixture of CYD and ILs, [C1C4Im][NTf2] and
[C1C1C4Im][NTf2], and therefore the sites of specific interactions
1
1
were studied by H NMR and ROESY experiments. The H
NMR chemical shifts of the CYD-IL mixtures at R ) 0.5 were
not significantly different from those of the neat CYD and ILs,
in both cases (Table S-1 in Supporting Information).
Figure 6. Excess molar enthalpies of the systems: CYD-[C1C4Im]-
[NTf2] (s) and CYD-[C1C1C4Im][NTf2] (---) vs the mole fraction of
CYD at 303.15 K.
NOESY experiments exhibited very weak or null cross peak
intensities. This was probably due to the fact that the quantity
ωτc was such that the NOE intensity was close to the null point.
Rotating frame NOE experiments (ROESY) allowed us to obtain
positive NOEs irrespective of the long rotational correlation time
TABLE 1: Excess Molar Enthalpies of Mixing of Organic
Solutes and Ionic Liquids at Mole Fraction of Solute 0.1
∆HEmix (J/mol)
T (K)
1
due to the high viscosity of the system. In H-1H ROESY
ionic liquid
solute
techniques based on space cross relaxations, the selective
irradiation of a proton group affects the intensities of integrals
of all proton groups that are spatially close but not necessarily
connected by chemical bonds. The method is based on the
assumption of short-range intermolecular distances (4-5 Å).43
The strength of the ROE signal is proportional to the inverse
sixth power of the distance between the atoms, I ∝ 1/r.6 In the
liquid state, this relation is possible if the intermolecular
association is tight enough to turn the intermolecular relaxation
into “intramolecular” within the ion pair or ion-molecule
association.41 Highly structured, bulk ionic liquids seem to fulfill
these requirements, thus allowing the use of the intermolecular
ROESY to derive lower limits for interionic distances. Conse-
quently, the intensity of the integrals (ICYD-IL) could be
considered as roughly inversely proportional to the intermo-
lecular distances (Figure S-4 and S-5, Supporting Information).
[C1C4Im][NTf2]
[C1C1C4Im][NTf2] 1,3-cyclohexadiene
1,3-cyclohexadiene
+85 ( 9
+164 ( 10
-189.842
-150.440
+435.640
-112.734
-115.639
+356.534
+394.034
303.15
303.15
363.15
363.15
363.15
323.15
298.15
323.15
323.15
[C1C4Im][NTf2]
[C1C4Im][NTf2]
[C1C4Im][NTf2]
[C1C2Im][NTf2]
[C1C2Im][NTf2]
[C1C2Im][NTf2]
[C1C2Im][NTf2]
benzene
toluene
methylcyclohexane
benzene
toluene
cyclohexene
cyclohexane
and were found to be 0.857 kJ mol-1 for CYD-[C1C4Im][NTf2]
and 1.629 kJ mol-1 for CYD-[C1C1C4Im][NTf2].
HECYD (kJ/mol) ) a + bxCYD
(2)
By integrating eq 1 and taking into account eq 2, we
calculated the excess molar enthalpy of mixing ∆HEmix by
Molecular mechanics calculations have been performed using
the SYBYL software with the TRIPOS force field developed
by Clark et al.44 on isolated pairs of CYD and imidazolium
cations. Intermolecular distances were fixed using results from
NMR ROESY experiments and the energy of both systems,
CYD-[C1C4Im] and CYD-[C1C1C4Im], were minimized. The
average distances between molecules of CYD and the imida-
zolium cations were determined from the geometry at the
potential energy minimum. CYD is mainly located near the butyl
chain of [C1C4Im][NTf2], the π bonds being located closer to
the imidazolium cation, a configuration similar to that previously
observed for toluene (Figure 7).9
nCYD
∆HEmix (kJ/mol)
)
)
∆Hmix
)
HECYD dnCYD/(nCYD + nIL)
∫
0
(a + b)xCYD + b(1 - xCYD)ln(1 - xCYD
)
(3)
From Figure 6, it can be seen that the excess molar enthalpies
of mixing for the binary mixtures of CYD and [C1C4Im][NTf2]
and [C1C1C4Im][NTf2] are positive over the studied range of
compositions. The values of excess molar enthalpy of mixing
of several aromatic and nonaromatic solutes in ILs together with
our results are presented in Table 1. To the best of our
knowledge the energetics of solvation of organic solutes in
[C1C1C4Im][NTf2] had never been studied; therefore, we could
establish comparisons only for 3-alkyl-1-methylimidazolium
bis(trifluoromethylsulfonyl)imide, [C1CnIm][NTf2]. Negative
values of ∆HmE ix for benzene and toluene in [C1C4Im][NTf2]
indicate more favorable interactions between aromatic systems
and imidazolium ILs than for methylcyclohexane, for which
the enthalpy of mixing is positive. A similar effect was observed
in [C1C2Im][NTf2], where the aromatic solutes have more
favorable interactions leading to negative enthalpies of mixing
and higher solubilities. Cyclohexene, due to the presence of a
double bond, also has lower enthalpy of mixing than cyclohex-
Access to the microscopic structure of the mixtures in the
condensed liquid phase is possible using molecular dynamics
simulation with all atoms explicitly present and periodic
boundary conditions to represent a vitually infinite system.
Condensed-phase simulations take into account all the two-body
interactions from the environment of each molecule or ion. The
molecular simulation details are given in the Experimental
Section. In Figure 8 are plotted the site-site radial distribution
functionssthe probability of finding pairs of atoms at a given
distance, compared to the averagesbetween the hydrogens on
Cꢀ of CYD, Hꢀ, and selected atoms of the [C1C4Im] and
[C1C1C4Im] cations at R ) 0.5 (atoms are labeled as indicated
in Scheme 1). In both cases, there is a higher probability (a

8162 J. Phys. Chem. B, Vol. 114, No. 24, 2010
Campbell et al.
NTf2]. The strong hydrogen bond between C2-H and the anion
in [C1C4Im][NTf2] prevents such an interaction with CYD. This
observation is consistent with the distances between the IL
cations and CYD obtained from ROESY experiments and
molecular mechanics calculations on isolated pairs, as seen in
Figure 7.
Detailed structural features are much better perceived in
3-dimensional spatial distribution functions. In Figure 10 is
represented the distribution of ions around a CYD molecule.
In blue is plotted the iso-surface corresponding to a local density
of twice the average density of C2 carbon atoms of the
imidazolium cations. Cation headgroups (the blue regions) are
located above and below the CYD ring, interacting preferentially
with the π-system of CYD. It can be seen that the terminal
carbons of the alkyl side chain, C9 (the gray regions), surround
the CYD molecule. Oxygen atoms of the NTf2- anion (plotted
in red) are found in the plane of CYD interacting with the
hydrogen atoms of the double bonds. Similar results have been
previously obtained for toluene in the same ILs.9 This shows
that although in CYD the π-system is smaller than that of a
fully aromatic system, it still determines the structure of the
solvation shell in ILs with cations positioned above and below
the plane of CYD and anion in the plane.
Figure 7. Representation of molecular positions in saturated solutions
of CYD-[C1C4Im][NTf2] and CYD-[C1C1C4Im][NTf2] from ROESY
NMR extrapolation.
Figure 11 shows spatial distribution functions around the
[C1C4Im] and the [C1C1C4Im] cations. With both cations, CYD
(the white regions) is preferentially located above and below
the plane of the imidazolium ring at distances greater than those
of the closest cation-anion pairs. In red is plotted the iso-surface
corresponding to the local density of 4 times the average density
-
of the oxygen atoms from NTf2 anions. As can be seen, for
-
[C1C4Im][NTf2] the probability of finding the oxygen of NTf2
near the C2-H of the imidazolium cation is higher than the
carbon of CYD, indicating a strong H-bond. The situation is
completely different for [C1C1C4Im][NTf2], where the prob-
ability of finding the carbon of CYD near the C2 of the cation
is higher than that of the oxygen atoms of NTf2-. As already
found in the case of solvation of toluene,9 interactions of the
[C1C1C4Im] cation with the anion are mainly through H4, H5
and also the nitrogen atoms, N1 and N3, whereas in [C1C4Im]
cation-anion interactions are mainly through C2-H bonds.
Mass Transport in Reaction Media. To establish whether
the difference in rate of catalytic hydrogenation of CYD can
be attributed to the specific interactions or only to differences
in thermophysical propreties of the reaction media, which affect
the mobility of molecules, the density and viscosity of the
mixtures of [C1C4Im][NTf2]-CYD and [C1C1C4Im][NTf2]-CYD
at R ) 0.5 were measured at 298.15 K and atmospheric pressure
(Table 2 and Figure S-7 in Supporting Information). The
additional methyl group on the imidazolium carbon C2 in
[C1C1C4Im][NTf2] naturally lowers both the mass and molar
density of the liquid in comparison with [C1C4Im][NTf2]. On
the other hand, the viscosity of pure [C1C1C4Im][NTf2] and also
of its mixture with CYD is roughly twice that of
[C1C4Im][NTf2], as reported in Table 2. Diffusion coefficients
(D) of CYD in both ILs at R ) 0.5 were determined from
extrapolation of DOSY data (Figure S-6 in Supporting Informa-
tion): 187.5 µm2/s in [C1C4Im][NTf2] against 97 µm2/s in
[C1C1C4Im][NTf2], meaning that diffusion of CYD in
[C1C4Im][NTf2] is 1.9 times faster than in [C1C1C4Im][NTf2].
Since, according to the Stokes-Einstein relation, the diffusion
coefficient is inversely proportional to the viscosity of the
medium, the ratio of values of D obtained is in agreement with
the ratio of measured viscosities of both mixtures at R ) 0.5,
which is η(CYD-[C1C1C4Im][NTf2])/η(CYD-[C1C4Im][NTf2])
Figure 8. Radial distribution functions between hydrogen on Cꢀ of
CYD, Hꢀ, and selected sites in the [C1C4Im] cation (top) and
[C1C1C4Im] cation (bottom), for R ) 0.5. Atoms are labeled as indicated
in Scheme 1.
stronger association) of finding all hydrogen atoms of CYD near
the side chain rather than in the vicinity of the aromatic nucleus
region of the imidazolium cation. This means that CYD is
preferentially solvated in the nonpolar domain of the ILs, as
already observed for saturated hydrocarbons and for the methyl
group of toluene. This orientation effect relative to the cation
is less distinct for CYD than for toluene,9 as expected given
the weaker cation-π interactions of the former.
Figure 9 represents the comparison of the radial distribution
functions of the mixtures CYD-[C1C4Im][NTf2] and CYD-
[C1C1C4Im][NTf2], again at R ) 0.5, between hydrogen on Cꢀ
and CR of CYD and C2 of the imidazolium rings. It can be
seen that both these hydrogen atoms are found with a higher
probability closer to C2 in [C1C1C4Im][NTf2] than in [C1C4Im

Catalytic Hydrogenation of 1,3-Cyclohexadiene
J. Phys. Chem. B, Vol. 114, No. 24, 2010 8163
SCHEME 1: Atom Labeling of CYD, 1-Butyl-3-methylimidazolium, and 1-Butyl-2,3-dimethylimidazolium Cations
) 1.9. The availability of H2 for the hydrogenation reaction is
also dependent on the diffusivity of the gas in both ILs. No
experimental data were found on the diffusivity of hydrogen in
ILs, but it is expected that it will also be inversely proportional
to the viscosity of the liquid medium since H2 interacts weakly
with ILs (its solubility is very low45,46). Molecular simulation
results for the diffusivity of H2 in [C6C1Im][NTf2] at different
temperatures have recently been published.47 We have calculated
that the increase of diffusivity of hydrogen and the decrease in
viscosity of [C6C1Im][NTf2] at the different temperatures48 are
comparable within the same order of magnitude. This is
commensurate with the expected precision of the molecular
simulation results.27 All these results indicate that the difference
in reactivity of both systems at R ) 0.5 can be mainly assigned
to the difference in viscosity and therefore in mobility of the
molecules, namely, CYD in the exchange reaction and both H2
and CYD in the hydrogenation reaction.
4. Conclusion
In this work, we studied the impact of two ionic liquid
solvents on the rates of two reaction steps of the catalytic
hydrogenation of 1,3-cyclohexadiene (CYD) with [Rh(COD)-
(PPh3)2]NTf2: the ligand exchange [Rh(COD)(PPh3)2]NTf2 to
[Rh(CYD)(PPh3)2]NTf2, and the catalytic hydrogenation of
CYD itself (COD ) 1,5-cyclooctadiene). It was found that
Figure 10. Spatial distribution functions around the C2 carbon of CYD
in CYD-[C1C4Im][NTf2] (left) and CYD-[C1C1C4Im][NTf2] (right)
at R ) 0.5. In blue is plotted the iso-surface corresponding to a local
density of twice the average density of the C2 carbon of the imidazolium
cations. In gray is plotted the iso-surface corresponding to a local density
of twice the average density of terminal methyl carbons from the butyl
side chain, C9. In red is plotted the iso-surface corresponding to a local
density of twice the average density of oxygen atoms from the NTf2
anion. Oxygen atoms are located in the CYD plane interacting with
hydrogen atoms of double bonds, HR and Hꢀ.
Figure 9. Comparison of site-site radial distribution functions between
chosen atoms in CYD and the cations, in CYD-[C1C4Im][NTf2] (top)
and CYD-[C1C1C4Im][NTf2] (bottom) for R ) 0.5. Atoms are labeled
as indicated in Scheme 1.
-

8164 J. Phys. Chem. B, Vol. 114, No. 24, 2010
Campbell et al.
higher diffusion coefficients in [C1C4Im][NTf2] (187.5 µm2/s) when
compared to those in [C1C1C4Im][NTf2] (97 µm2/s), meaning that
diffusion of CYD in [C1C4Im][NTf2] is 1.9 times faster than in
[C1C1C4Im][NTf2]. Since the diffusion coefficient is inversely
proportional to the viscosity of the medium, the values obtained
are in agreement with the Stokes-Einstein relation, since the
ratio of measured viscosities of both ionic liquids at R ) 0.5
η(CYD-[C1C1C4Im][NTf2])/η(CYD-[C1C4Im][NTf2]) is also
1.9. The diffusion coefficient of gases is also expected to vary
inversely with the viscosity of the medium. These results
indicate that the difference in the rate of both ligand exchange
and hydrogenation of CYD with [Rh(COD)(PPh3)2]NTf2 in
[C1C4Im][NTf2] and [C1C1C4Im][NTf2] can be attributed to the
difference in viscosity of both ionic liquids, hence the mobility
of the molecules in solution.
Acknowledgment. A.P. acknowledges the postdoctoral grant
by the project ANR CALIST and P.C. acknowledges the Ph.D.
grant attributed by the Ministe`re de l’Enseignement Supe´rieur
et de la Recherche, France.
Figure 11. Spatial distribution functions around the C2 carbon of the
cations in CYD-[C1C4Im][NTf2] (left) and CYD-[C1C1C4Im][NTf2]
(right) at R ) 0.5. Above and below are different views of the same
iso-surfaces. In white is plotted the iso-surface corresponding to a local
density of twice the average density of CR of CYD. In red is plotted
the iso-surface corresponding to a local density of 4 times the average
Supporting Information Available: Determination of the
conversion and the composition of the reaction mixture by
GC. UV-vis spectra of [Rh(COD)(PPh3)2]NTf2 and
[Rh(CYD)(PPh3)2]NTf2, ROESY and DOSY NMR pulse se-
-
density of oxygen atoms from the NTf2 anion.
1
quences, H NMR shifts and ROESY NMR spectra of the
TABLE 2: Density and Viscosity, at 298.15 K and
Atmospheric Pressure, of Mixtures of CYD-[C1C4Im][NTf2]
and CYD-[C1C1C4Im][NTf2] at R ) 0.5
mixture of CYD in [C1C4Im][NTf2] and [C1C1C4Im][NTf2]. This
material is available free of charge via the Internet at http://
pubs.acs.org.
xIL
R
F (g cm-3
)
η (mPa s)
CYD-[C1C4Im][NTf2]
References and Notes
1.000
0.667
0.000
0.498
1.4375
1.3597
48.45
24.87
(1) Wasserscheid, P.; Welton, T. Ionic Liquids in Synthesis; Wiley
VCH: Weinheim, Germany, 2008.
CYD-[C1C1C4Im][NTf2]
(2) Parvulescu, V. I.; Hardacre, C. Chem. ReV. 2007, 107, 2615.
(3) Dyson, P. J.; Zhao, D. Hydrogenation. Multiphase Homogeneous
Catalysis; Wiley-VCH: Weinheim, 2005; Vol. 2, pp 494.
(4) Olivier-Bourbigou, H.; Vallee, C. Multiphase Homogeneous Ca-
talysis; Wiley-VCH: Weinheim, 2005; Vol. 2, pp 413-431.
(5) Hintermair, U.; Gutel, T.; Slawin, A. M. Z.; Cole-Hamilton, D. J.;
Santini, C. C.; Chauvin, Y. J. Organomet. Chem. 2008, 693, 2407.
(6) Dupont, J.; Suarez, P. A. Z. Phys. Chem. Chem. Phys. 2006, 8,
2441.
(7) Dupont, J.; Suarez, P. A. Z.; De Souza, R. F.; Burrow, R. A.;
Kintzinger, J.-P. Chem.sEur. J. 2000, 6, 2377.
(8) Lachwa, J.; Bento, I.; Duarte, M. T.; Canongia Lopes, J. N.; Rebelo,
L. P. N. Chem. Commun. 2006, 2445.
(9) Gutel, T.; Santini, C. C.; Padua, A. A. H.; Fenet, B.; Chauvin, Y.;
Canongia Lopes, J. N.; Bayard, F.; Costa Gomes, M. F.; Pensado, A. S. J.
Phys. Chem. B 2009, 113, 170.
(10) Mele, A.; Romano, G.; Giannone, M.; Ragg, E.; Fronza, G.; Raos,
G.; Marcon, V. Angew. Chem., Int. Ed. 2006, 45, 1123.
(11) Padua, A. A. H.; Costa Gomes, M. F.; Canongia Lopes, J. N. A.
Acc. Chem. Res. 2007, 40, 1087.
(12) Triolo, A.; Russina, O.; Bleif, H.-J.; Di Cola, E. J. Phys. Chem. B
2007, 111, 4641.
(13) Canongia Lopes, J. N.; Padua, A. A. H. J. Phys. Chem. B 2006,
110, 3330.
(14) Canongia Lopes, J. N.; Costa Gomes, M. F.; Padua, A. A. H. J.
Phys. Chem. B 2006, 110, 16816.
(15) Hunt, P. A. J. Phys. Chem. B 2007, 111, 4844.
(16) Magna, L.; Chauvin, Y.; Niccolai, G. P.; Basset, J.-M. Organo-
metallics 2003, 22, 4418.
(17) Schrock, R. R.; Osborn, J. A. J. Am. Chem. Soc. 1976, 98, 4450.
(18) Blesic, M.; Canongia Lopes, J. N.; Padua, A. A. H.; Shimizu, K.;
Costa Gomes, M. F.; Rebelo, L. P. N. J. Phys. Chem. B 2009, 113, 7631.
(19) Izatt, R. M.; Terry, R. E.; Haymore, B. L.; Hansen, L. D.; Dalley,
N. K.; Avondet, A. G.; Christensen, J. J. J. Am. Chem. Soc. 1976, 98, 7620.
(20) Liu, Y. F.; Sturtevant, J. M. Protein Sci. 1995, 4, 2559.
(21) Briggner, L. E.; Wadso, I. J. Biochem. Biophys. Methods 1991,
22, 101–118.
1.000
0.662
0.000
0.510
1.4177
1.3442
105.00
47.13
both steps are twice as fast in 1-butyl-3-methylimidazolium
bis(trifluoromethylsulfonyl)imide, [C1C4Im][NTf2], than in 1-bu-
tyl-2,3-dimethylimidazolium bis(trifluoromethylsulfonyl)imide,
[C1C1C4Im][NTf2]. The rate-determining step in both ionic
liquids is the hydrogenation of CYD step.
Molecular dynamics simulations and NMR experiments
indicate that in both ILs, for a molar ratio CYD/IL equal to
0.5, CYD is solvated preferentially in lipophilic regions (in close
proximity to the alkyl side chains of the cations) as already
observed for saturated hydrocarbons and the methyl group of
toluene. In addition, in [C1C1C4Im][NTf2] the probability of
finding CYD near the C2 of the cation is higher than in
[C1C4Im][NTf2], in agreement with the shorter CYD-
[C1C1C4Im] distances determined by ROESY NMR. On the
other hand, a higher solubility of CYD in [C1C4Im][NTf2] and
smaller positive enthalpies of mixing for the CYD-[C1C4Im]-
[NTf2] system in comparison with CYD-[C1C1C4Im][NTf2]
indicate more favorable interactions between CYD and the
[C1C4Im] cation than with the [C1C1C4Im] cation. These
thermodynamic factors concur with the differences in catalytic
activity, although they cannot fully explain the differences
observed. Association between macroscopic thermodynamic
information, thermophysical data, and microscopic structural
information is necessary to fully explain the differences in
reactivity found in both IL media.
(22) Cutting, B.; Ghose, R.; Bodenhausen, G. J. Magn. Reson. 1999,
138, 326.
(23) Canongia Lopes, J. N.; Deschamps, J.; Padua, A. A. H. J. Phys.
Chem. B 2004, 108, 2038.
Pure [C1C1C4Im][NTf2] has a higher viscosity than pure
[C1C4Im][NTf2], and the same relative values are observed in
the mixtures with CYD. These differences in viscosity induce

Catalytic Hydrogenation of 1,3-Cyclohexadiene
J. Phys. Chem. B, Vol. 114, No. 24, 2010 8165
(24) Canongia Lopes, J. N.; Padua, A. A. H. J. Phys. Chem. B 2004,
108, 16893.
(25) Canongia Lopes, J. N.; Shimizu, K.; Padua, A. A. H.; Umebayashi,
Y.; Fukuda, S.; Fujii, K.; Ishiguro, S. I. J. Phys. Chem. B 2008, 112, 9449.
(26) Jorgensen, W. L.; Maxwell, D. S.; Tirado-Rives, J. J. Am. Chem.
Soc. 1996, 118, 11225.
(37) Natori, I.; Imaizumi, K.; Yamagishi, H.; Kazunori, M. J. Polym.
Sci., Part B 1998, 36, 1657.
(38) Domanska, U.; Pobudkowska, A.; Eckert, F. Green Chem. 2006,
8, 268.
(39) Lachwa, J.; Bento, I.; Duarte, M. T.; Canongia Lopes, J. N.; Rebelo,
L. P. N. Chem. Commun. 2006, 2445.
(40) Letcher, T. M.; Marsicano, F. J. Chem. Thermodyn. 1974, 6, 509.
(41) Marczak, W.; Verevkin, S. P.; Heintz, A. J. Solution Chem. 2003,
32, 519.
(42) Nebig, S.; Bolts, R.; Gmehling, J. Fluid Phase Equilib. 2007, 258,
168.
(43) Frezzato, D.; Rastrelli, F.; Bagno, A. J. Phys. Chem.B 2006, 110,
5676.
(44) Clark, M.; Cramer, R. D., III; Van Opdenbosch, N. J. Comput.
Chem. 1989, 10, 982–1012. Sybil program version 7.0, Tripos Inc., 1699
South Hanley Rd., St. Louis, MO 63144, U.S.A.
(45) Jacquemin, J.; Costa Gomes, M. F.; Husson, P.; Majer, V. J. Chem.
Thermodyn. 2006, 38, 490. Jacquemin, J.; Husson, P.; Majer, V.; Costa
Gomes, M. F. Fluid Phase Equilib. 2006, 240, 87.
(46) Costa Gomes, M. F. J. Chem. Eng. Data 2007, 52, 472.
(47) Shi, W.; Sorescu, D. C.; Luebke, D. R.; Keller, M. J.; Wickra-
manyake, S. J. Phys. Chem. B 2010, 114, 6431.
(48) Marsh, K. N.; Brennecke, J. F.; Chirico, R. D.; Frenkel, M.; Heintz,
A.; Magee, J. W.; Peters, C. J.; Rebelo, L. P. N.; Seddon, K. R. Pure Appl.
Chem. 2009, 81, 781.
(27) Maginn, E. J. Acc. Chem. Res. 2007, 40, 1200.
(28) Smith, W.; Forester, T. R.; Todorov, I. T.; The DL_POLY molecular
simulation package, 2.20 ed.; T. D. P. m. s., Ed.; STFC Daresbury
Laboratorys, Warrington, U.K., 2007.
(29) Chauvin, Y.; Mussmann, L.; Olivier, H. Angew. Chem., Int. Ed.
1996, 34, 2698.
(30) Schrock, R. R.; Osborn, J. A. J. Am. Chem. Soc. 1976, 98, 2134.
(31) Costa Gomes, M. F. Unpublished results.
(32) Camper, D.; Becker, C.; Koval, C.; Noble, R. Ind. Eng. Chem. Res.
2006, 45, 445.
(33) Cui, Y.; Biondi, I.; Chaubey, M.; Yang, X.; Fei, Z.; Scopelliti, R.;
Hartinger, C. G.; Li, Y.; Chiappe, C.; Dyson, P. J. Phys. Chem. Chem.
Phys. 2010, 12, 1834.
˙
(34) Doman˜ska, U.; Zołek-Tryznowska, Z.; Kro´likowski, M. J. Chem.
Eng. Data 2007, 52, 1872.
(35) Kato, R.; Krummen, M.; Gmehling, J. Fluid Phase Equilib. 2004,
224, 47.
(36) Husson, P.; Pison, L.; Jacquemin, J.; Costa Gomes, M. F. Fluid
Phase Equilib. 2010, DOI: 10.1016/j.fluid.2010.02.021.
JP102941N