Combining Homogeneous Catalysis with Heterogeneous Separation using Tunable Solvent Systems

Article abstract of DOI:10.1021/jp907325y

Tunable solvent systems couple homogeneous catalytic reactions to heterogeneous separations, thereby combining multiple unit operations into a single step and subsequently reducing waste generation and improving process economics. In addition, tunable solvents can require less energy than traditional separations, such as distillation. We extend the impact of such solvents by reporting on the application of two previously described carbon dioxide tunable solvent systems: polyethylene glycol (PEG)/organic tunable solvents (POTS) and organic/aqueous tunable solvents (OATS). In particular, we studied: (1) the palladium catalyzed carbon-oxygen coupling of 1-bromo-3,5-dimethylbenzene and o-cresol to potassium hydroxide to produce o-tolyl-3,5-xylyl ether and 1-bromo-3,5-di-tert-butylbenzene to potassium hydroxide to produce 3,5-di-tert-butylphenol in PEG400/1,4-dioxane/water and (2) the rhodium-catalyzed hydroformylation of p-methylstyrene in water/ acetonitrile to form 2-(p-tolyl) propanal. In addition, we introduce a novel tunable solvent system based on a modified OATS where propane replaces carbon dioxide. This represents the first use of propane in a tunable solvent system.

Full text of DOI:10.1021/jp907325y

3932
J. Phys. Chem. A 2010, 114, 3932–3938
Combining Homogeneous Catalysis with Heterogeneous Separation using Tunable Solvent
Systems†
Vittoria M. Blasucci,‡,| Zainul A. Husain,‡,| Ali Z. Fadhel,‡,| Megan E. Donaldson,‡,|
Eduardo Vyhmeister,‡,| Pamela Pollet,§,| Charles L. Liotta,*,‡,§,| and Charles A. Eckert*,‡,§,|
Georgia Institute of Technology, School of Chemical & Biomolecular Engineering, 311 Ferst DriVe,
Atlanta, Georgia 30332-0100, Georgia Institute of Technology, School of Chemistry & Biochemistry,
901 Atlantic DriVe, Atlanta, Georgia 30332-0400, and Georgia Institute of Technology, Specialty Separations
Center, Atlanta, Georgia 30332-0100
ReceiVed: July 30, 2009; ReVised Manuscript ReceiVed: September 14, 2009
Tunable solvent systems couple homogeneous catalytic reactions to heterogeneous separations, thereby
combining multiple unit operations into a single step and subsequently reducing waste generation and improving
process economics. In addition, tunable solvents can require less energy than traditional separations, such as
distillation. We extend the impact of such solvents by reporting on the application of two previously described
carbon dioxide tunable solvent systems: polyethylene glycol (PEG)/organic tunable solvents (POTS) and
organic/aqueous tunable solvents (OATS). In particular, we studied: (1) the palladium catalyzed carbon-oxygen
coupling of 1-bromo-3,5-dimethylbenzene and o-cresol to potassium hydroxide to produce o-tolyl-3,5-xylyl
ether and 1-bromo-3,5-di-tert-butylbenzene to potassium hydroxide to produce 3,5-di-tert-butylphenol in
PEG400/1,4-dioxane/water and (2) the rhodium-catalyzed hydroformylation of p-methylstyrene in water/
acetonitrile to form 2-(p-tolyl) propanal. In addition, we introduce a novel tunable solvent system based on
a modified OATS where propane replaces carbon dioxide. This represents the first use of propane in a tunable
solvent system.
1. Introduction
We exploit tunable solvent systems to couple the benefits of
both homogeneous and heterogeneous catalysis. Separation is
more facile in heterogeneous catalysis, but reactivity and product
selectivity suffer and mass transfer limitations may dominate.1
In the investigated tunable systems, gas, either CO2 or propane,
is added to the homogeneous reaction phase to induce a
Figure 1. Tunable solvent system schematic where L1 is the polar
liquid and L2 is the less-polar liquid.
postreaction phase-split between the polar protic component
(water or polyethylene glycol (PEG)) and the relatively nonpolar
is most useful if it occurs under readily attainable conditions
aprotic component (acetonitrile, dioxane, or tetrahydrofuran),
and if it is large. Furthermore, one seeks an area of relatively
as shown in Figure 1. Operationally, this phase split is easily
wide disparity in the compositions of the two liquid phases so
reversed upon gas depressurization, thereby allowing a solvent
that solvent waste is minimized and the solute distribution is
switch between homogeneous and heterogeneous and vice versa.
as skewed as possible to yield efficient separation. Fluids are
Whereas tunable solvents include both supercritical and near-
selected where specific interactions (hydrogen bonds, Lewis
critical fluids, here we focus on those whose properties, such
acid/base interactions, etc.) give preferential distributions.
as solvent power and diffusivity, can be gradually adjusted
Finally, we need to choose fluids that do not interfere with the
through the use of a pressurized gas, most commonly carbon
chemistry of the reaction. Then, such systems may be applied
dioxide (CO2).2 Several such tunable solvent systems have been
to couple homogeneous reaction to heterogeneous separation
published.3-9 Here we expand the use of these systems with
for enhanced reaction rates and minimal environmental impact.
new applications for the pharmaceutical and fine chemical
We report (1) the palladium-catalyzed coupling of 1-bromo-
industries, and we characterize a novel propane-based tunable
3,5-dimethylbenzene and o-cresol to potassium hydroxide to
produce o-tolyl-3,5-xylyl ether and 1-bromo-3,5-di-tert-butyl-
benzene to potassium hydroxide to produce 3,5-di-tert-butylphe-
nol in PEG 400/1,4-dioxane/water, (2) the rhodium-catalyzed
hydroformylation of p-methylstyrene in water/acetonitrile to
form 2-(p-tolyl) propanal, and (3) the phase behavior of a
solvent.
In designing tunable solvent systems, the applicability of the
system to real processes depends strongly on phase equilibria.
There is generally a limited region in pressure-temperature-
composition space where two liquid phases exist. That region
propane-induced tetrahydrofuran (THF)/water tunable solvent
system. Although here we describe the first application of PEG-
based tunable solvents and propane tunable solvents, organic/
aqueous tunable solvent (OATS) systems, using CO2 as a switch,
have been used for enzyme recycle in biocatalytic reactions3,4
and for catalyst recycle in the hydroformylation of 1-octene.9
† Part of the special issue “Green Chemistry in Energy Production
Symposium”.
* Corresponding author.
‡ Georgia Institute of Technology, School of Chemical & Biomolecular
Engineering.
| Georgia Institute of Technology, Specialty Separations Center.
§ Georgia Institute of Technology, School of Chemistry & Biochemistry.
10.1021/jp907325y  2010 American Chemical Society
Published on Web 10/09/2009

Homogeneous Catalysis and Heterogeneous Separation
J. Phys. Chem. A, Vol. 114, No. 11, 2010 3933
For the latter case, the reaction rate was augmented by two
orders of magnitude from the biphasic reaction consistently over
three recycles, and a 99.9% product separation and catalyst
recovery were obtained.
further purified via a Matheson gas purifier and filter cartridge
(model 450B, type 451 filter). The following materials were
used as received from the suppliers: tris(dibenzylideneacetone)-
dipalladium (Pd2dba3, Strem, 21% Pd), 2-di-tert-butylphosphino-
2′,4′,6′-triisopropylbiphenyl (P1, Sigma-Aldrich, 97%), o-tolyl-
3,5-xylyl ether (TCI America, >97%), 1-bromo-3,5-di-tert-
butylbenzene (Alfa Aesar, 99%), HPLC grade tetrahydrofuran
(THF, Sigma-Aldrich, 99.9%), instrument grade propane (Air-
gas, 99.5%), synthesis gas (syngas, 1:1 molar ratio of H2:CO),
o-cresol (Sigma-Aldrich, >99%), and 2,4-di-tert-butylphenol
(Alfa Aesar, 97%). The following solvents were degassed by
the freeze-pump-thaw method: HPLC grade water (Sigma-
Aldrich), HPLC grade acetonitrile (Sigma-Aldrich, >99.9%), and
p-methylstyrene (Alfa Aesar, >98%). Also, the following
materials were used as received and stored in a nitrogen-filled
glovebox: triphenylphosphine-3-sulfonic acid sodium salt
(TPPMS, TCI America, > 90%) and rhodium(I) dicarbonyl
acetylacetonate (Rh(acac), Sigma-Aldrich, 98%). Potassium
hydroxide (KOH, EMD Chemicals) pellets were ground to fine
particles.
PEG offers many advantages as a tunable solvent component.
PEG can increase rates of reactions involving salts because of
its ability to complex cations and activate anions for reaction.10
PEG is attractive because of its low cost, thermal stability,
negligible vapor pressure, biodegradability, and nontoxicity.11
PEG can surpass water as the tunable solvent’s polar phase by
expanding the list of soluble organic cosolvents. We investigate
a polyethylene glycol/organic tunable solvent (POTS) system,
with CO2 as a miscibility switch, for the catalytic production
of phenols or aromatic ethers. Current technologies for their
production suffer from harsh reaction conditions (400 °C and
30 MPa), poor atom economy, and/or hazardous intermediates.12
Recently, Buchwald et al. used a biphasic organic/water solvent
system for the mild palladium-catalyzed route to these molecules
via carbon-oxygen coupling of aryl halides to hydroxide salts.13
The presence of the reactants disrupted the system’s phase
behavior and did not allow for a homogeneous reaction phase
as in OATS. However, the C-O coupling reactions benefit from
a homogeneous system and catalyst recycle, as provided by
POTS. Carbon dioxide serves a dual purpose for phenol
production by not only acting as the separation switch but also
generating in situ carbonic acid, utilizing the equilibrium with
water, for postreaction workup. Postreaction neutralization is
required because the pKa of most phenols is lower than the
reaction media pH. At 0.95 MPa of CO2, the pH of the aqueous
phase of the water/dioxane system was reduced to <3.4
Furthermore, our group has published on the use of this
equilibrium in acid catalysis.14 Any residual carbonic acid is
easily reversed upon venting CO2. The results for both reaction
conversion and catalyst/product partitioning follow.
We applied OATS to the hydroformylation of p-methylsty-
rene, which is an intermediate step in ibuprofen manufacturing.15
For the industrial scale hydroformylation process, OATS can
be implemented with minimal or no modification to existing
facilities because the separation of products from the catalysts
can be achieved with pressures similar to those used during the
reaction. We report conversion and branched product yields as
well as partitioning of both the reactant and desired product
between the two liquid phases as a function of CO2 pressure.
Because of the benefits of using CO2 (such as nonflammability
and nontoxicity) as an OATS antisolvent, only a few reports
have been made on other small gas molecules that could serve
the same purpose.16-18 However, we have found that using
propane as an antisolvent may provide several advantages over
CO2. These include drastically lowering phase split/operating
pressures, elimination of carbonic acid formation from the
equilibrium between CO2 and water for pH sensitive reactions,
avoiding the use of buffers and subsequent solids handling, and
decreased product (organic-rich) phase contamination. These
advantages may or may not substantiate introducing the use of
a flammable gas. The phase diagram for THF/water/propane is
presented herein.
Reaction of 1-Bromo-3,5-dimethylbenzene and o-Cresol
with Potassium Hydroxide to Produce o-Tolyl-3,5-xylyl
Ether in POTS. Solid components including potassium hy-
droxide (1 equiv, 220 mg, 1 M concentration), o-cresol (1 equiv,
420 mg, 1 M concentration), Pd2dba3 (0.02 equiv, 77 mg), and
P1 ligand (0.08 equiv, 136 mg) were added to glass carousel
reaction tubes (radleys carousel 12 plus reaction station)
equipped with magnetic stir bars and degassed. The metal-to-
ligand molar ratio was maintained at 1:2 for all experiments.
PEG 400 (4.0 mL/100 wt %, 2.8 mL/72 wt %, or 2.3 mL/60 wt
%), 1,4-dioxane (0.0 mL/0 wt %, 1.2 mL/28 wt % or 1.0 mL/
24 wt %), and water (0.0 mL/0 wt %, 0.0 mL/0 wt %, or 0.7
mL/16 wt %) were then introduced via an airtight degassed
syringe to keep the total solvent volume at 4 mL. The carousel
was then temperature-controlled at 80 °C. The tops of the tubes
were water-cooled. After 1 h at 80 °C, 1-bromo-3,5-dimethyl-
benzene (1 equiv, 0.5 mL, 1 M concentration) was added to
begin the overnight reaction. Postreaction, the glass tubes were
cooled to room temperature, and the mixture was neutralized
by one of the following techniques: the addition of 8 mL of 1
M hydrochloric acid or bubbling CO2 for 0.5 h. The products
were then extracted into 8 mL of diethyl ether and analyzed
using gas chromatography-mass spectroscopy (GCMS) (Agilent
GC-HP 6890 with a GCMS-HP 5973 detector and HP-5MS
column). All reactions were run in triplicate.
Separation of o-Tolyl-3,5-xylyl Ether using POTS. Parti-
tioning experiments were run in a 60 mL high-pressure Jerguson
cell. The cell temperature was monitored in situ with a
thermocouple (Omega type K) calibrated against a platinum
RTD (Omega PRP-4) with DP251 Precision RTD benchtop
thermometer (DP251 Omega), providing an accuracy of (0.2
K. Pressure in the cell was measured using a Druck pressure
transducer (PDCR 910) and read-out box (DPI 260) calibrated
against a hydraulic piston pressure gauge (Ruska, GE Infra-
structure Sensing) with an uncertainty of (0.1 bar. The cell
was manually shaken by a rotating shaft. The cell was first
evacuated. 1,4-Dioxane (4 mL/60 wt %), PEG 400 (9 mL/26
wt %), water (2 mL/14 wt %), and o-tolyl-3,5-xylyl ether (1.2
mL, 0.5 M concentration) were premixed and added to the cell
via an airtight syringe. CO2 was added to the cell by an ISCO
syringe pump to the desired separation pressure. After equilib-
rium was ascertained, three 0.5 mL samples were taken from
each liquid phase using a sample loop. The samples were rinsed
and diluted in methanol and then analyzed by GCMS (as
2. Experimental Methods
Materials. The following materials were degassed by bub-
bling either nitrogen or argon through the liquids for 0.5 h: PEG
400 (Sigma-Aldrich), 1,4-dioxane (Fischer Scientific, 99.9%),
HPLC grade water (Sigma-Aldrich), and 1-bromo-3,5-dimeth-
ylbenzene (Alfa Aesar, 98%). Carbon dioxide was supercritical
fluid chromatography grade (SFC grade, Air Gas, 99.999%) and

3934 J. Phys. Chem. A, Vol. 114, No. 11, 2010
Blasucci et al.
SCHEME 1: C-O Coupling Reaction among o-Cresol, KOH, and 1-Bromo-3,5-di-methylbenzene
described above). The average value for the three samples is
reported. The direct sampling method compares well with the
synthetic method described by Lazzaroni et al.19,20
and captured in acetone. The samples were analyzed using an
Agilent GC-FID (model 6890) with an Agilent column (model
HP-5MS). External standards of known concentrations were
used to calibrate the FID response.
Separation of PdP12 Using POTS. The partitioning of the
catalyst, PdP12, between the PEG-rich and dioxane-rich phases
was determined using the experimental procedure and Jerguson
apparatus, as previously outlined. P1 (110 mg, 20 mM concen-
tration) and Pd2dba3 (66 mg, 5 mM concentration) were mixed
and heated to 80 °C for 30 min in the ternary solvent system
prior to injection into the cell. Sample concentrations were
determined by ICP-MS and ICP-OES analyses (Columbia
Analytical Services: lower detection limits 2 and 20 ppm,
respectively).
Reaction of 1-Bromo-3,5-di-tert-butylbenzene with Potas-
sium Hydroxide to Produce 3,5-Di-tert-butylphenol in POTS.
Solid components including potassium hydroxide (3 equiv, 660
mg, 3 M concentration), Pd2dba3 (0.02 equiv, 77 mg), P1 ligand
(0.08 equiv, 136 mg), and 1-bromo-3,5-di-tert-butylbenzene (1
equiv, 1.04 g, 1 M concentration) were added to glass carousel
reaction tubes (as previously described) and degassed. PEG 400
(4 mL/100 wt %, 2.8 mL/72 wt %, or 2.3 mL/60 wt %), 1,4-
dioxane (0.0 mL/0 wt %, 1 mL/28 wt %, or 1.2 mL/24 wt %),
and water (0.0 mL/0 wt %, 0.0 mL/0 wt %, or 0.7 mL/16 wt
%) were then introduced via an airtight degassed syringe to keep
the total solvent volume at 4 mL. The carousel was then
temperature controlled at 80 °C for the overnight reaction. Work-
up of the products was performed according to the procedure
outlined for o-tolyl-3,5-xylyl ether production.
Separation of 2,4-Di-tert-butylphenol using POTS. The
partitioning of a structural isomer of 3,5-di-tert-butylphenol, 2,4-
di-tert-butylphenol (650 mg, 0.5 M concentration), between the
PEG-rich and dioxane-rich phases was determined using the
experimental procedure and apparatus, as previously outlined
for the separation of o-tolyl-3,5-xylyl ether using POTS.
p-Methylstyrene Hydroformylation in OATS. p-Methyl-
styrene hydroformylations were carried out in a 300 mL stainless
steel Parr autoclave (Parr Instrument Company, model 4561).
The reaction pressure was monitored with a digital pressure
transducer (Heise, model 901B) providing an accuracy of (0.7
bar. A proportional integral derivative (PID) temperature
controller and tachometer (Parr Instrument Company, model
4842) were used to control the temperature of the reactor to
(2 °C and the stirring speed to (5 rpm. The temperature inside
the reactor was monitored with a type J thermocouple, and heat
was provided by a high-temperature heating mantle. The reactor
was evacuated and flushed with 3.5 bar of syngas. The degassed
p-methylstyrene and catalyst solution (70/30 v/v of acetonitrile/
water) were added using gastight syringes. The total volume of
the reaction mixture was 50 mL with a p-methylstyrene
concentration of 0.15 M. The concentration equivalence of
Rh(acac) and TPPMS is 0.0025 and 0.017, respectively. The
reactor was heated to the desired reaction temperature, stirred
at 300 rpm, and subsequently pressurized with 30 bar of syngas.
After a 1 h reaction period, a liquid phase sample was withdrawn
Separation of p-Methylstyrene and 2-(p-Tolyl) Propanal
in OATS. The partitioning of p-methylstyrene (2.0 mL, 0.15
M concentration) and 2-(p-tolyl) propanal (2.3 mL, 0.15 M
concentration) between the acetonitrile- and water-rich liquid
phases (loaded with 98 mL of 70/30 v/v acetonitrile/water
solution) was determined at room temperature using the Parr
autoclave described above. CO2 was added to the cell until the
desired pressure was reached using an ISCO syringe pump. After
equilibrium, three 0.1 mL samples of the acetonitrile-rich phase
and three 0.5 mL samples of the water-rich phase were taken
using a six-way sample loop and analyzed using the Agilent
GC-FID described above.
Propane OATS Phase Equilibria. Phase equilibria for the
water/propane/THF ternary system were determined by direct
sampling using the fixed volume Jerguson apparatus, as
described above. The Jerguson cell was loaded using an airtight
syringe with a 55 wt % water and 45 wt % THF mixture (12
and 10 g, respectively). Propane was added to the cell using an
ISCO syringe pump. Three 0.5 mL samples were removed from
each liquid phase via a six-way sample loop and bubbled into
and rinsed with methanol using an HPLC pump. These samples
were analyzed for water content via Karl Fischer titration
(Mitsubishi Kasei Corporation CA-20 moisturemeter). The
organic content was measured via GCMS (as described above)
and quantified against a calibration curve. We determined the
propane content by measuring the volume displaced while
expanding three 0.5 mL liquid phase samples into the headspace
of an inverted buret filled with water. The mass of propane was
calculated from the displaced volume via the ideal gas equation
at standard temperature and pressure after allowing the tem-
perature to reach equilibrium.
3. Results and Discussion
Reaction of 1-Bromo-3,5-dimethylbenzene and o-Cresol
with Potassium Hydroxide to Produce o-Tolyl-3,5-xylyl
Ether in POTS. We previously reported the POTS phase
behavior for PEG 400/1,4-dioxane/CO2.3 The room-temperature
phase diagram showed good separation between the two liquid
phases at moderate pressures (i.e., 6 MPa). Less than 1 wt %
PEG 400 was found in the dioxane-rich phase under these
conditions. Therefore, the reaction between 1-bromo-3,5-dim-
ethylbenzene and o-cresol with potassium hydroxide to produce
o-tolyl-3,5-xylyl ether, shown in Scheme 1, was run in three
sets of solvent systems including PEG 400/1,4-dioxane/water
(60, 24, 16 wt %, respectively), PEG 400/1,4-dioxane (72, 28
wt %, respectively), and PEG 400. Dilute acid was used during
the workup procedure to quantify phenol produced as a side
product. The results of these experiments are shown in Table
1. In terms of both conversion and selectivity, the ternary solvent

Homogeneous Catalysis and Heterogeneous Separation
J. Phys. Chem. A, Vol. 114, No. 11, 2010 3935
TABLE 1: Conversion and Selectivity to o-Tolyl-3,5-xylyl
Ether in Three Different Solvent Environments
conversion
(%)
selectivity
(%)
solvent system
PEG 400
PEG 400/1,4-dioxane
PEG 400/1,4-dioxane/water
60 ( 6
64 ( 10
81 ( 9
64 ( 2
62 ( 1
66 ( 6
system is the most efficient. Side products detected were
m-xylene, 3,5-dimethlyphenol, 3,3′-5,5′-tetra-methlbiphenyl, and
1,1′-oxybis[3,5-dimethyl]benzene.
Figure 3. Partition coefficient versus CO2 pressure for the catalyst,
PdP12, between 1,4-dioxane-rich and PEG-rich liquid phases at room
temperature. (Error bars represent root-mean-square deviations.)
The partitioning of the reaction product, o-tolyl-3,5-xylyl
ether, between the dioxane- and PEG-rich phases was deter-
mined at ambient temperature as a function of pressure and
concentration of water. The partition coefficient, K, is defined
as the concentration of the solute in the organic phase divided
by the concentration of the solute in the polar phase. The
partition coefficient can be used as a guide for determining the
efficiency of the separation process. Figure 2 displays these
results for pressures ranging from 4.2 to 5.4 MPa. The addition
of water to the solvent system increases the partitioning of the
ether into the organic-rich phase. This could be due to any or
all of the following effects: (1) an increased polarity of the polar
phase, (2) hydrogen bonding between PEG and water disrupting
the interaction between PEG and the solute molecule, and (3)
2 mL increased volume of the PEG phase from water (dilution
effect). The partition coefficient of o-tolyl-3,5-xylyl ether, in
the case of the ternary solvent system, is highest at lower CO2
pressures. We attribute this to a dilution effect, where adding
more CO2 preferentially expands the volume of the organic
phase, thereby decreasing the concentration of ether in this
phase. The trend reverses with minimal water present because
the volumes of both the PEG and dioxane phase expand by
similar amounts. On the other hand, when water is present in
cosolvent amounts, the PEG-rich phase expands very little when
compared with the dioxane-rich phase because of the limited
solubility of CO2 in water. Optimal results were obtained near
the phase split pressure, 4.2 MPa, and in a three component
solvent system, (PEG 400 (60 wt %)/1,4-dioxane (26 wt %)/
water (14 wt %)), where o-tolyl-3,5-xylyl ether is four times
more concentrated in the organic-rich phase than in the PEG-
rich phase. Therefore, both our reaction and separation are most
efficient in this ternary solvent system.
Figure 3 illustrates that the partition coefficient for the catalyst
versus pressure follows a similar trend as o-tolyl-3,5-xylyl ether.
Increasing CO2 pressure improves catalyst retention in the PEG-
rich phase but decreases product separation into the organic-
rich phase, thereby necessitating an optimization. This result
can also be explained by a dilution effect where as pressure is
increased, the dioxane phase is preferentially volume expanded
over the PEG phase. Figure 3 also shows that at 4.8 MPa, K
equals 0.038, which corresponds to the catalyst being 27 times
more concentrated in the PEG-rich phase than in the product
phase (1/K). This limited solubility of the catalyst in the gas-
expanded dioxane phase could be further decreased by structur-
ally modifying the nonpolar, bulky side groups of the ligand,
P1.
Reaction of 1-Bromo-3,5-di-tert-butylbenzene and Potas-
sium Hydroxide to Produce 3,5-Di-tert-butylphenol in POTS.
The reaction between 1-bromo-3,5-di-tert-butylbenzene and
potassium hydroxide to produce 3,5-di-tert-butylphenol is shown
in Scheme 2. This reaction was chosen for the CO2-philic
character of the phenol, increasing its partitioning potential into
the gas-expanded product phase. Three sets of solvent reaction
environments were investigated including: PEG 400, PEG 400
(72 wt %)/1,4-dioxane (28 wt %), or PEG 400 (60 wt %)/1,4-
dioxane (24 wt %)/water (16 wt %). Conversion and selectivity
to phenol (using aqueous HCl workup) after an overnight
reaction are shown in Table 2. Side products that were formed
were 3,5-di-tert-butylbenzene and 3,3′-5,5′-tetra-tert-butylphe-
nylether. Higher conversions were achieved in the systems
without water as a cosolvent; however, as conversion increased,
selectivity suffered. The use of CO2 as a postreaction neutraliza-
tion medium was examined in the ternary solvent system of
PEG 400/1,4-dioxane/water. This workup method proved to be
efficient because the overnight reaction showed 76% conversion
and 71% selectivity to phenol, values matching those obtained
when using a dilute acid workup. Therefore, the cost and waste
associated with traditional neutralizations is eliminated.
The partitioning of 2,4-di-tert-butylphenol, a regioisomer of
3,5-di-tert-butylbenzene, between the dioxane-rich and PEG-
rich phases was determined at room temperature as a function
of pressure ranging from 4.2 to 4.7 MPa in the ternary solvent
mixture (PEG 400 (60 wt %)/1,4-dioxane (26 wt %)/water (14
wt %) (See Figure 4). The same trend is seen as that with o-tolyl-
3,5-xylyl ether partitioning, where an increased CO2 pressure
decreases the partition coefficient. However, the partition
coefficient reached a maximum of unity slightly above the phase
split pressure at 4.2 MPa.
The separation of the catalyst, PdP12 (where P1 is 2-di-tert-
butylphosphino-2′,4′,6′-triisopropylbiphenyl), from the organic-
rich phase was also determined at room temperature as a
function of CO2 pressures ranging from 4.5 to 4.8 MPa. The
ternary solvent system of PEG 400 (60 wt %)/1,4-dioxane (26
wt %)/water (14 wt %) was used to provide optimal partitioning.
p-Methylstyrene Hydroformylation and Separation in
OATS. We have previously reported the ternary acetonitrile/
aqueous/CO2 phase behavior, which showed 12 mol % water
in the organic-rich phase at 40 °C and 3.1 MPa. Under these
Figure 2. Partition coefficient versus CO2 pressure for o-tolyl-3,5-
xylyl ether between 1,4-dioxane-rich and PEG-rich liquid phases at
various concentrations of water and room temperature. (Error bars
represent root-mean-square deviations.)

3936 J. Phys. Chem. A, Vol. 114, No. 11, 2010
Blasucci et al.
SCHEME 2: Reaction between 1-Bromo-3,5-di-tert-butylbenzene and KOH
TABLE 2: Conversion and Selectivity to
3,5-Di-tert-butylphenol in Three Different Solvent
Environments
Figure 6 shows the partitioning of p-methylstyrene and 2-(p-
tolyl) propanal between the aqueous- and acetonitrile-rich liquid
phases at ambient temperature as a function of CO2 pressure
between 1.0 and 3.5 MPa. The separation of the products and
reactants from the aqueous phase, which contains the catalyst,
is important for simplified downstream processing. We are able
to achieve 99.5% recovery of both the product and reactants at
3 MPa into the gas-expanded organic phase. For both molecules,
an exponential increase in the partition coefficient is seen with
increasing CO2 pressure because of the decreased organic
content in the water phase and decreased water content in the
organic phase.7 Although both p-methylstyrene and 2-(p-tolyl)
propanal distribute primarily into the acetonitrile-rich phase,
p-methylstyrene has a larger partitioning value because of its
decreased polarity when compared with the branched aldehyde
product. This facile CO2-induced separation may offer a green
alternative to more traditional waste-generating and energy-
intensive methods such as liquid-liquid extraction or distillation.
Propane OATS Phase Behavior. Unlike CO2, propane does
not react with water; therefore, no in situ acid is formed.
Reactions sensitive to pH such as enzymatic and base-catalyzed
reactions are made possible by the elimination of carbonic acid
from a CO2 tunable solvent system. Buffers used in traditional/
CO2-induced OATS systems can create processing problems
because the addition of gas pressure may cause the buffer to
salt out, rendering the buffer ineffective and introducing process
conversion
selectivity
(%)
solvent system
PEG 400
PEG 400/1,4-dioxane
PEG 400/1,4-dioxane/water
(%)
100 ( 0
100 ( 0
80 ( 7
60 ( 6
44 ( 4
68 ( 2
conditions, the water-rich phase contains only 7 mol %
acetonitrile. A two-liquid-phase region exists over a large
pressure range, 3.3 MPa span.7 We now report this OATS
system for the hydroformylation of p-methylstyrene, as shown
in Scheme 3. Styrene reacts with CO and H2 in the presence of
rhodium(I) dicarbonyl acetylacetonate (Rh(acac)) and triph-
enylphosphine-3-sulfonic acid sodium salt (TPPMS) to produce
linear and branched aldehydes. The branched aldehyde, 2-(p-
tolyl) propanal, is the desired product because it is used to
produce ibuprofen through oxidation.15 The conversion and
branched product selectivity after 1 h of reaction time are shown
in Figure 5; the selectivity is >95% at 40 °C. Complete
conversion of the starting material occurs in <1 h at 80 °C. The
turnover frequency (TOF, moles reacted per moles catalyst per
hour) of the homogeneous catalyst is 92, 406, and 402 at 40,
80, and 120 °C, respectively. These data show an order of
magnitude increase in the rate of reaction when compared with
heterogeneously catalyzed hydroformylations reported in
literature.21,22 The decrease in branched product yield with
increasing temperature is due to the competing ꢀ-hydride
elimination reaction, which returns the branched product back
to p-methylstyrene.15
Figure 5. Conversion and 2-(p-tolyl)propanal selectivity of p-
methylstyrene hydroformylation after 1 h. (Error bars represent root-
mean-square deviations.)
Figure 4. Partition coefficient versus CO2 pressure for 2,4-di-tert-
butylphenol between 1,4-dioxane- and PEG-rich liquid phases at room
temperature. (Error bars represent root-mean-square deviations.)
SCHEME 3: p-Methylstyrene Hydroformylation
Reaction
Figure 6. Partition coefficient versus pressure for 2-(p-tolyl) propanal
and p-methylstyrene between acetonitrile- and water-rich liquid phases
at room temperature.

Homogeneous Catalysis and Heterogeneous Separation
J. Phys. Chem. A, Vol. 114, No. 11, 2010 3937
TABLE 3: Ternary Phase Behavior of the Water (x1) +
Propane (x2) + Tetrahydrofuran (x3) System at 303 K with
Compositions in Weight Percent
water-rich phase (L1) tetrahydrofuran-rich phase (L2)
P (MPa)
x1
x2
x3
x1
x2
x3
0.43
0.56
0.80
0.92
0.706 0.008 0.286
0.711 0.006 0.283
0.776 0.016 0.208
0.820 0.024 0.156
0.061
0.040
0.031
0.028
0.083
0.175
0.310
0.452
0.856
0.785
0.659
0.520
TABLE 4: Ternary Phase Behavior of the Water (x1) +
Propane (x2) + Tetrahydrofuran (x3) System at 313 K with
Compositions in Weight Percent
water-rich phase (L1) tetrahydrofuran-rich phase (L2)
P (MPa)
x1
x2
x3
x1
x2
x3
0.64
0.85
1.12
1.35
0.752 0.001 0.247
0.769 0.002 0.229
0.831 0.006 0.163
0.836 0.009 0.155
0.055
0.034
0.037
0.022
0.126
0.242
0.290
0.446
0.819
0.724
0.673
0.532
Figure 7. Ternary phase behavior of the water + propane +
tetrahydrofuran system at T ) 303 K with compositions in wt %.
Overall propane reduces the cloud point pressure for OATS by
a factor of four and increases organic phase purity by a factor
of three.
4. Summary
Tunable solvent systems offer facile solutions toward the goal
of coupling homogeneous catalytic reaction with heterogeneous
separation. We have demonstrated two novel applications of
tunable solvent systems with this goal in mind. When producing
o-tolyl-3,5-xylyl ether in the ternary system of PEG/dioxane/
water, we show conversion of 80%, and upon separation, the
product is up to four times more concentrated in the organic-
rich phase and the catalyst is up to nearly 30 times more
concentrated in the PEG-rich phase. The similar trend of the
product and the catalyst partition coefficient with CO2 pressure
requires an overall process optimization to determine the most
effective operating conditions. The reaction to 3,5-dimethylphe-
nol reached up to 100% conversion but was not as favorably
partitioned into the organic phase as the ether product. This is
the first report of the application of PEG-based tunable solvents.
The hydroformylation of p-methylstyrene in OATS showed
elevated TOF (up to 400) and selectivity (up to 95%) at mild
temperatures. Facile product/reactant recovery was obtained at
pressures as low as 3 MPa, which is similar to the pressure
required for the reaction. The phase behavior of propane/water/
THF was characterized. This system improves phase purity,
lowers operating pressures, and eliminates acid formation when
compared with CO2-based OATS. Overall, these examples
illustrate the diverse application of tunable solvent systems once
carefully designed to benefit reaction and separation simulta-
neously. This work shows promise for ongoing experiments,
which consist of validating solvent and catalyst reuse by
optimizing process recycle.
Figure 8. Ternary phase behavior of the water + propane +
tetrahydrofuran system at T ) 313 K with compositions in wt %.
difficulty associated with solids handling. The propane-OATS
system does not require a buffer to control pH.
We report the mass composition of each liquid phase for
water/propane/THF over a representative range of pressures
encompassing the three-phase region. The vapor phase was
assumed to be mostly propane and was therefore not measured.
The cloud points for this system at 303 and 313 K are 0.24 and
0.28 MPa, respectively. These cloud point pressures are the
lowest reported in the literature for any tunable solvent system.
A low-phase split pressure enables reduced energy consumption
and lower operating costs.
Figure 7 shows the mass composition of each liquid phase
in the water/propane/THF system at 303 K and pressures ranging
from 0.43 to 0.92 MPa. Figure 8 shows the same ternary system
at 313 K and pressures ranging from 0.64 to 1.35 MPa. Tables
3 and 4 correspond to Figures 7 and 8, respectively. The phase
equilibria at 303 and 313 K show the same trend. At both
temperatures studied, the THF-rich phase has only 3 wt % water
at pressures greater than 0.8 MPa. The water concentrations in
the organic-rich product phase are a measure of the quality of
separation because the catalyst can be designed to be hydrophilic
and the product is usually hydrophobic. In contrast, the 298 K
water/CO2/THF system described by Lazzaroni et al. has an
organic phase water content of 9 wt % at pressures >4.0 MPa.5
Acknowledgment. OATS work was funded by the NSF
project no. CTS-0553690. Special thanks to two undergraduates
from the Georgia Institute of Technology, Andy Wu and Sarah
Lencenski, for their assistance with OATS experiments.
References and Notes
(1) Dijkstra, H. P.; Van Klink, G. P. M.; Van Koten, G. Acc. Chem.
Res. 2002, 35, 798–810.
(2) Li, H.; Maroncelli, M. J. Phys. Chem. B 2006, 110, 21189–21197.

3938 J. Phys. Chem. A, Vol. 114, No. 11, 2010
Blasucci et al.
(3) Hill, E. M. B.; James, M.; Hallett, J.; Bommarius, A.; Liotta, C.;
Eckert, C. Green Chem. 2007, 9, 888–893.
(14) Weikel, R.; Hallett, J.; Liotta, C.; Eckert, C. Top. Catal. 2006, 37,
75–80.
(4) Broering, J. M. H.; Hill, E.; Hallett, J.; Liotta, C.; Eckert, C.;
Bommarius, A. Angew. Chem. 2006, 45, 4670–4673.
(5) Donaldson, M.; Draucker, L.; Blasucci, V.; Liotta, C.; Eckert, C.
Fluid Phase Equilib. 2009, 277, 81–86.
(6) Draucker, L.; Hallett, J.; Bush, D.; Eckert, C. Fluid Phase Equilib.
2006, 241, 20–24.
(7) Lazzaroni, M. J.; Bush, D.; Jones, R.; Hallett, J. P.; Liotta, C. L.;
Eckert, C. A. Fluid Phase Equilib. 2004, 224, 143–154.
(8) Mellein, B. R.; Brennecke, J. F. J. Phys. Chem. B 2007, 111, 9001–
9009.
(9) Hallett, J. P.; Ford, J. W.; Jones, R. S.; Pollet, P.; Thomas, C. A.;
Liotta, C. L.; Eckert, C. A. J. Ind. Eng. Chem. Res. 2008, 47, 2585–2589.
(10) CodeofFederalRegulations,Title21,Volume3,CITE21CFR172.820;
FDA, 2001.
(15) Nair, V. S.; Mathew, S. P.; Chaudhari, R. V. J. Mol. Catal A: Chem.
1999, 143, 99–110.
(16) Elgin, J. C.; Weinstock, J. J. J. Chem. Eng. Data 1959, 4, 3–12.
(17) Adrian, T.; Wendland, M.; Hasse, H.; Maurer, G. J. Supercrit.
Fluids 1998, 12, 185–221.
(18) Freitag, J.; Sanz Diez, M. T.; Tuma, D.; Ulanova, T. V.; Maurer,
G. J. Supercrit. Fluids 2004, 32, 1–13.
(19) Hill, E. M. Benign Tunable Solvents for Improved Processing of
Pharmaceutically Relevant Products and Catalysts. Ph.D. Dissertation,
Georgia Institute of Technology, August 2007.
(20) Lazzaroni, M. J. Optimizing Solvent Selection for Separation
and Reaction. Ph.D. Dissertation, Georgia Institute of Technology, July
2004.
(21) Hamza, K.; Blum, J. Eur. J. Org. Chem. 2007, 28, 4706–4710.
(22) El, A. B.; Tijani, J.; Fettouhi, M.; El-Faer, M.; Al-Arfaj, A. Appl.
Catal., A 2005, 283, 185–196.
(11) Weissermel, K.; Arpe, H. J. Industrial Organic Chemistry, 3rd ed.;
VCH Publishers: New York, 1997.
(12) Anderson, A.; Ikawa, T.; Tundel, R.; Buchwald, S. J. Am. Chem.
Soc. 2006, 128, 10694–10695.
(13) Starks, C. M.; Liotta, C. L.; Halpern, M. Phase-Transfer Catalysis:
Fundamentals, Applications, and Industrial PerspectiVes; Chapman and
Hall: New York, 1994.
JP907325Y