Phase-separable catalysis using room temperature ionic liquids and supercritical carbon dioxide

Article abstract of DOI:10.1039/b009701m

A new phase-separable catalysis concept is demonstrated using supercritical carbon dioxide and the room temperature ionic liquid 1-butyl-3-methylimidazolium hexafluorophosphate for hydrogenation of alkenes and carbon dioxide.

Full text of DOI:10.1039/b009701m

Phase-separable catalysis using room temperature ionic liquids and
supercritical carbon dioxide
Fuchen Liu, Michael B. Abrams, R. Tom Baker* and William Tumas*
Los Alamos Catalysis Initiative, Chemistry Division, MS J514, Los Alamos National Laboratory, Los Alamos, New
Mexico 87545, USA. E-mail: weg@lanl.gov; tumas@lanl.gov
Received (in Cambridge, UK) 4th December 2000, Accepted 29th January 2001
First published as an Advance Article on the web 14th February 2001
A new phase-separable catalysis concept is demonstrated
using supercritical carbon dioxide and the room temperature
ionic liquid 1-butyl-3-methylimidazolium hexafluorophos-
phate for hydrogenation of alkenes and carbon dioxide.
= 220 h21) and 96% conversion measured after 3 h. Product
recovery and catalyst recycling were demonstrated for dec-
1-ene using a variable volume reaction vessel to transfer the
upper scCO2 phase to a separate receiver under high pressure
after reaction.8 The reaction vessel was then recharged with
dec-1-ene, hydrogen and carbon dioxide and allowed to react
for 1 h. This process was repeated up to four times and the
reaction proceeded to ca. 98% conversion each time, thereby
demonstrating efficient catalyst recycling through immobiliza-
tion of the rhodium catalyst in the ionic liquid.
Since many organic compounds are not miscible with
imidazolium-based ionic liquids, many of the reported catalytic
reactions in ILs can be considered to be organic/IL biphasic
systems. Product separation has been demonstrated in a number
of cases.9 In order to benchmark the reactivity of the scCO2/IL
system with organic/IL systems, we also examined the above
hydrogenation reactions using n-hexane in place of carbon
dioxide, under otherwise identical conditions.‡ The reaction of
dec-1-ene to form n-decane proceeded to 99% conversion after
1 h at 50 °C in n-hexane/[BMIM][PF6]. Conversions of
cyclohexene to cyclohexane were similar to those measured in
the scCO2/[BMIM][PF6] experiments: 88% after 2 h, 98% after
3 h at 50 °C. These results suggest that there is no reactivity
advantage for CO2 over n-hexane for simple hydrogenation
reactions.
There is a continuing interest in developing new concepts for
biphasic or phase-separable catalysis where a homogeneous
catalyst is immobilized in one liquid phase and the reactants
and/or products reside largely in another liquid phase.1
Brennecke and Beckman2 have recently reported that the room
temperature ionic liquid (IL) 1-butyl-3-methylimidazolium
hexafluorophosphate, 1, ([BMIM][PF6]), exhibits very inter-
esting phase behavior with supercritical carbon dioxide
(scCO2), where CO2 can dissolve significantly (up to 0.6 mole
fraction) into the lower IL phase, but no polar IL dissolves in the
upper scCO2 phase. They further illustrated that organics such
as naphthalene could be extracted from the IL phase using
scCO2. Kazarian et al.3 recently provided corroborating spec-
troscopic evidence that CO2 dissolves in 1. This intriguing
discovery of phase behavior suggested that a reaction system
comprised of scCO2 and an ionic liquid might offer particular
advantages as a new biphasic catalysis system. Herein, such a
scCO2/IL biphasic system is demonstrated, where a homoge-
neous transition metal catalyst is immobilized in an IL phase
and products can be isolated from a distinct scCO2 phase.4
Room temperature ionic liquids5 and supercritical carbon
dioxide6 have both been extensively studied as alternative
solvents for a wide range of homogeneously catalyzed reactions
including hydrogenation, hydroformylation, selective oxidation
and carbon–carbon bond formation. We chose to examine the
scCO2/IL system through investigations of the hydrogenation of
dec-1-ene and cyclohexene using Wilkinson’s catalyst
RhCl(PPh3)3, 2, which is insoluble in scCO2 but soluble in
water-stable [BMIM][PF6], 1.
In order to capitalize on advantages arising from the polar
nature of the ionic liquid phase, we sought to examine reactions
that proceed through polar intermediates which would be
soluble in the IL phase and not in scCO2 (Scheme 1). We
studied the hydrogenation of carbon dioxide in the presence of
dialkylamines to produce N,N-dialkylformamides. Noyori
et al.10 and Baiker et al.11 have reported elegant studies on the
catalytic hydrogenation of CO2 to formamides in scCO2 using
scCO2-soluble
RuCl2(PMe3)4
and
scCO2-insoluble
All reactions were run in custom-designed, high pressure
stainless steel view cells with a sapphire window using
magnetic stirring at 50 °C, as described previously.7† The
hydrogenation of dec-1-ene proceeded in 98% conversion to n-
decane in 1 h at 48 bar H2 and a total pressure of 207 bar, which
implies a time-averaged reaction turnover frequency (TOF) of
410 h21. The reaction mixture was biphasic throughout the
reaction with a colorless carbon dioxide phase above a yellow
ionic liquid phase. Hydrogenation of cyclohexene proceeded
more slowly under identical conditions with 82% conversion
being observed after 2 h reaction at 50 °C (time-averaged TOF
RuCl2(dppe)2 (dppe = Ph2PCH2CH2PPh2), 3, respectively.
These reactions involve ionic carbamate intermediates, since
dialkylamines react with CO2 reversibly to form dialkylammon-
ium dialkylcarbamates, 4 [eqns (3), (4)].12
[cat]
(1)
CO2 + H2 [| HCO2H
HCO2H + R3N fi [HNR3]+[HCO2]2
HNR2 + CO2 fi R2NCO2H
(2)
(3)
(4)
R2NCO2H + NHR2 fi [NH2R2]+[O2CNR2]2
4
[NH2R2]+[O2CNR2]2 + 2HCO2H ?
2HCON(R)2 + 2H2O + CO2
(5)
Noyori et al. reported that salt 4a (R = CH3) separated from
the scCO2 phase right from the start of the reaction starting with
dimethylamine and rates using 4a were similar to those using
dimethylamine. Very high rates for selective N,N-dime-
thylformamide (DMF) formation were reported using either
liquid 4a,10 or dimethylamine.10,11 Significantly Noyori et al.
found that activity and selectivity (i.e. formic acid vs.
Scheme 1 Pictorial illustration of a scCO2/IL biphasic system R = reactant,
P = product, I = polar intermediate (e.g. carbamate 4).
DOI: 10.1039/b009701m
Chem. Commun., 2001, 433–434
This journal is © The Royal Society of Chemistry 2001
433

Table 1 Production of formamides from amines and CO2 in different solvent systems
Amine
Solvent
Catalyst
T/°C
t/h
Con (%) Sel (%) TONa
TOF (h21 b
)
c
NH(C2H5)2
scCO2
scCO2
scCO2/IL
RuCl2(PMe3)4
RuCl2(PMe3)4
RuCl2(dppe)2
100
100
80
13
5
5
38
18
100
46d
30d
> 99
820
260
110
63
52
> 22
NH2(n-C3H7)c
NH(n-C3H7)2
a TON = mol product/mol ruthenium. b Time-averaged TOF. c Ref. 10. d Other product is formic acid.
formamide production) decreased rapidly with longer chain
Notes and references
amines (R ≠ CH3) which lead to solid, scCO2-insoluble
† Typical experimental procedure for olefin hydrogenation: 20 mg (21
mmol) of 2 were added to 10.5 mmol of olefin (500 equiv. relative to catalyst
2) and 1 mL of ionic liquid 1 in the reaction cell which was then pressurized
with hydrogen (48 bar) and CO2 to a total pressure of 207 bar. All the
experiments were conducted at 50 °C using n-nonane as an internal
standard.
‡ While magnetic stirring may not be optimal for these reactions, our
preliminary experimental conditions allow for a semi-quantitative com-
parison.
§ Typical experimental procedure for CO2 hydrogenation: 0.5 mL (4.3
mmol) of 4a was charged into a 33 mL stainless steel reaction cell with a
solution of 10 mg (10 mmol) of 3 in 1 mL of 1. The cell was heated to 80 °C
after being filled with 55 bar H2, then it was pressurized to a total pressure
of 276 bar with CO2. The start of the reaction was defined as the time of CO2
gas introduction. The yield and identification of formamide were performed
by GC, GC-MS and 1H NMR.10
¶ 0.15 mL (1.1 mmol) of di-n-propylamine and 1 mL of 1 were used. Higher
concentration of amine (e.g. 7.3 mmol di-n-propylamine in 1 mL of 1) led
to precipitation of a white solid (carbamate).
carbamate intermediates.10
In the scCO2/IL system, catalyst 3 and carbamate 4 are both
soluble in the IL phase. Carbamate 4a could be completely
converted to DMF after 4 h at 80 °C using 55 bar hydrogen
under a total pressure of 276 bar.§ While this unoptimized
reactivity is significantly less than that reported by Baiker for
the liquid carbamate,11‡ we found that the reactivity and more
importantly the selectivity is higher for amines other than
dimethylamine in the scCO2/IL biphasic system. Under similar
reaction conditions, di-n-propylamine¶ led to complete amine
conversion and exclusive production of the corresponding N,N-
di-n-propylformamide after only 5 h at 80 °C. The activity and
selectivity are higher than those reported10 for less bulky
diethylamine and n-propylamine in neat scCO2 (Table 1). The
increased selectivity in the scCO2/IL biphasic system likely
arises from the increased solubility of the solid dialkylcarba-
mate intermediates in the IL phase or through rate enhancement
of amidation of formic acid derived from CO2 hydrogenation
[eqn (5)].
The highly polar formamide products appear to be very
soluble in the ionic liquid phase, 1. Preliminary experiments
reveal that they do not partition strongly into the scCO2 phase
after only one reaction cycle. Quantitative data on the
partitioning of organic compounds between ionic liquids and
either organic solvents or scCO2 are just starting to appear in the
literature.13,14 We have demonstrated extraction of DMF from
IL 1 in a separate experiment. After stirring 30 mL scCO2 (Ptot
= 276 bar) over a solution of 1.0 mL ionic liquid 1 and 1.0 mL
DMF for 1 h at 80 °C, the upper CO2 phase was transferred
under high pressure to another vessel. Subsequent pressure let-
down led to 151 mg of isolated DMF (16% recovery). We have
been able to demonstrate effective product recovery for N,N-di-
n-propylformamide after several reaction/recovery cycles. The
recovery yield in the first cycle was poor (less than 5%),
however, the yield in the second cycle improved significantly to
61%. N,N-di-n-propylformamide can be almost quantitatively
recovered in the third and fourth cycle, suggesting the IL phase
becomes saturated with the product in the first two cycles.15
In conclusion, we demonstrate one of the first examples4 of
catalysis in a biphasic system incorporating supercritical carbon
dioxide and ionic liquids and more importantly the first example
involving CO2 reaction chemistry. High selectivity, catalyst
recycling and product recovery were observed for hydro-
genation of CO2 in the presence of dialkylamines, demonstrat-
ing the potential advantages arising from dissolving polar
reaction intermediates in the IL phase. We are in the process of
investigating other reactions, particular those involving polar
intermediates, as well as quantifying the partitioning of
reactants (including gases), products and catalyst between the
two phases.
1 B. Cornils and W. A. Herrmann, in Applied Homogeneous Catalysis
with Organometallic Compounds, ed. B. Cornils and W. A. Herrmann,
Weinheim: New York, 1996, Chapter 4.1, pp. 1167–1197; An entire
issue of Catalysis Today (1998, 42 (2)) was devoted to biphasic
homogeneous catalysis.
2 L. A. Blanchard, D. Hancu, E. J. Beckman and J. F. Brennecke, Nature,
1999, 399, 28.
3 S. G. Kazarian, B. J. Briscoe and T. Welton, Chem. Commun., 2000,
2047.
4 While this manuscript was in preparation we learned that other groups
have been working on asymmetric hydrogenation reactions in a scCO2/
IL biphasic system: P. G. Jessop, private communication; R. A. Brown,
P. Pollett, E. McKoon, C. A. Eckert, C. L. Liotta and P. G. Jessop, J. Am.
Chem. Soc., 2001, 123, 1254.
5 T. Welton, Chem. Rev., 1999, 99, 2071; P. Wasserscheid and W. Keim,
Angew. Chem., Int. Ed., 2000, 39, 3772.
6 P. G. Jessop and W. Leitner, in Chemical Synthesis Using Supercritical
Fluids, ed. P. G. Jessop and W. Leitner, Weinheim, Wiley-VCH, 1999;
P. G. Jessop, T. Ikariya and R. Noyori, Chem. Rev., 1999, 99, 475.
7 D. K. Morita, D. R. Pesiri, S. A. David, W. H. Glaze and W. Tumas,
Chem. Commun., 1998, 1397.
8 G. B. Jacobson, C. T. Lee, K. P. Johnston and W. Tumas, J. Am. Chem.
Soc., 1999, 121, 11902.
9 Y. Chauvin, L. Mussmann and H. Olivier, Angew. Chem., Int. Ed., 1995,
34, 2698; B. Ellis, W. Keim and P. Wasserscheid, Chem. Commun.,
1999, 337; C. J. Mathews, P. J. Smith and T. Welton, Chem. Commun.,
2000, 1249.
10 P. G. Jessop, Y. Hsiao, T. Ikariya and R. Noyori, J. Am. Chem. Soc.,
1994, 116, 8851; P. G. Jessop, Y. Hsiao, T. Ikariya and R. Noyori, J. Am.
Chem. Soc., 1996, 118, 344.
11 O. Kröcher, R. A. Köppel and A. Baiker, Chem. Commun., 1997,
453.
12 H. B. Wright and M. B. Moore, J. Am. Chem. Soc., 1948, 70, 3865; K.
Takeshita and A. Kitamoto, J. Chem. Eng. Jpn., 1988, 221, 411.
13 L. A. Blanchard and J. F. Brennecke, Ind. Eng. Chem. Res., 2001, 40,
287.
14 L. A. Blanchard, Z. Gu and J. F. Brennecke, J. Phys. Chem.,
submitted.
We gratefully acknowledge support of this research by the
U.S. Department of Energy through a Laboratory Directed
Research and Development (LDRD) grant at Los Alamos
National Laboratory.
15 D. E. Bergbreiter, Y. S. Liu and P. L. Osburn, J. Am. Chem. Soc., 1998,
120, 4250.
434
Chem. Commun., 2001, 433–434