Organic synthesis in water/carbon dioxide microemulsions

Article abstract of DOI:10.1021/jo981825i

Nucleophilic substitution reactions were performed in H₂O/CO₂ (w/c) microemulsions formed with an anionic perfluoropolyether ammonium carboxylate (PFPE COO^-NH₄⁺) surfactant. These reactions between hydrophilic nucleophiles and hydrophobic substrates were accomplished in an environmentally benign microemulsion without requiring toxic organic solvents or phase transfer catalysts. For the synthesis of benzyl bromide from benzyl chloride and KBr, the yield was an order of magnitude higher in w/c microemulsions versus conventional water-in-oil (w/o) microemulsions. Benzoyl chloride and p-nitrophenyl chloroformate were hydrolyzed in w/c microemulsions with rate constants an order of magnitude faster than those in w/o microemulsions.

Full text of DOI:10.1021/jo981825i

J . Org. Chem. 1999, 64, 1201-1206
1201
Or ga n ic Syn th esis in Wa ter /Ca r bon Dioxid e Micr oem u lsion s
Gunilla B. J acobson, C. Ted Lee, J r., and Keith P. J ohnston*
Department of Chemical Engineering, University of Texas, Austin, Texas 78712
Received September 8, 1998
Nucleophilic substitution reactions were performed in H2O/CO2 (w/c) microemulsions formed with
an anionic perfluoropolyether ammonium carboxylate (PFPE COO-NH4+) surfactant. These
reactions between hydrophilic nucleophiles and hydrophobic substrates were accomplished in an
environmentally benign microemulsion without requiring toxic organic solvents or phase transfer
catalysts. For the synthesis of benzyl bromide from benzyl chloride and KBr, the yield was an
order of magnitude higher in w/c microemulsions versus conventional water-in-oil (w/o) micro-
emulsions. Benzoyl chloride and p-nitrophenyl chloroformate were hydrolyzed in w/c microemulsions
with rate constants an order of magnitude faster than those in w/o microemulsions.
In tr od u ction
In a microemulsion, high concentrations of both water-
soluble and water-insoluble compounds can be dissolved
simultaneously. Also, the surface area between water and
oil can reach values as high as 105 m2 per liter of micro-
emulsion.9 When both reactants are concentrated inside
a microemulsion droplet, it functions as a nanoreactor,
and the size of the droplet can have a significant influence
on the reaction kinetics. In the case where one reactant
is in the continuous phase and the second in the dis-
persed phase, the partitioning of the reactants in the
interphase is of great importance. The reaction can take
place inside the microemulsion droplet, at the surfactant
interface, or in the continuous phase. Depending upon
the charge of the surfactant headgroup, a reaction may
either be catalyzed or inhibited. Recently, a phase trans-
fer catalyst has been added to both microemulsions10 and
emulsions11 to facilitate transfer across the interface; fur-
thermore, high organic substrate concentrations were
used.
Most studies of reactions in microemulsions have been
performed with reactant concentrations below 0.1 M,
which are somewhat low for conventional synthetic
organic chemistry and especially for industrial applica-
tions. In a few cases, reactant concentrations above 1 M
have been used.4,10-12 In these cases, it was necessary to
compensate for the influence of the reactants on the
stability of the microemulsion by using higher surfactant
concentrations.
When performing synthetic organic chemistry, one is
often faced with the problem of reacting a hydrophobic
compound with one that is hydrophilic, e.g., a salt. One
approach is to add a cosolvent to mix water with an
organic solvent. Although this approach makes the
organic substrate accessible to the water soluble reagent,
it also increases the contact of the substrate with water,
promoting competitive hydrolytic pathways. Another
possibility is to add a phase transfer catalyst (PTC),
typically a quaternary ammonium compound. The cata-
lyst forms a complex with the hydrophilic substrate in
the aqueous phase and shuttles it into the organic phase,
allowing the two substrates to react.1 A disadvantage of
phase transfer catalysis is that the organic solvent and
PTC can be toxic. Recently this technique has been
applied to the interface between a solid suspended salt
and CO2 as a replacement for an organic solvent.2
Another approach is to use a microemulsion, which is
a thermodynamically stable and optically transparent
microheterogeneous system.3 Microemulsions, with drop-
let sizes in the range of 50-500 Å, are formed by adding
an amphiphile to a system containing water and oil
(usually a hydrocarbon). A cosolvent is often needed for
formation of a microemulsion to achieve the proper
balance of attractive and repulsive interactions on the
hydrophobic and hydrophilic sides of the interface. The
cosolvent reduces the interfacial tension between the
droplets and the continuous phase.4 Microemulsions are
good solvents for both hydrophilic and hydrophobic
reactants and have been utilized as media for catalytic5
and enzymatic6 reactions, for enhanced oil recovery,7 and
as a mobile phase in chromatographic solutions.8
Microemulsions can also be formed by using super-
critical fluids as the nonaqueous phase, as reviewed
elsewhere.13 Supercritical fluids have liquid-like densities
(8) (a) Hernandez-Torres, M. A.; Landy, J . S.; Dorsey, J . G. Anal.
Chem. 1986, 58, 744-747. (b) Gale, R. W.; Smith, R. D.; Fulton, J . L.
Anal. Chem. 1987, 59, 1977-1979.
(1) Starks, C. M.; Liotta, C. L.; Halpern, M. Phase-Transfer Cataly-
sis; Chapman and Hall:, New York, 1994.
(2) Dillow, A. K.; Yun, S. L. J .; Suleiman, D.; Boatright, D. L.; Liotta,
C. L.; Eckert, C. A. Ind. Eng. Chem. Res. 1996, 35, 1801-1806.
(3) (3) Microemulsions: Theory and Practice; Prince, L. M., Ed.;
Academic Press: New York, 1974.
(4) Menger, F. M.; Erlington, A. R. J . Am. Chem. Soc. 1991, 113,
9621-9624.
(5) Leong, Y. S.; Candau, F. J . Phys. Chem. 1982, 86, 2269-2271.
(6) Khmelnitsky, Y. L.; Neverova, I. N.; Polyakov, V. I.; Grinberg,
V. Y.; Levashov, A. V.; Martinek, K. Eur. J . Biochem. 1990, 190, 155-
159.
(9) Schwuger, M.-J .; Stickdorn, K.; Schoma¨cker, R. Chem. Rev. 1995,
95, 849-864.
(10) Siswanto, C.; Battal, T.; Schuss, O. E.; Rathman, J . F. Langmuir
1997, 13, 6047-6052.
(11) Battal, T.; Siswanto, C.; Rathman, J . F. Langmuir 1997, 13,
6053-6057.
(12) (a) Schoma¨cker, R. J . Chem. Res., Synop. 1991, 92-93. (b)
Menger, F. M., Erlington, A. R. Actual Chim. 1991, 177.
(13) (a) McFann, G. J .; J ohnston, K. P. Supercritical Microemulsions.
In Microemulsions: Fundamental and Applied Aspects; Kumar, P., Ed.;
1998, Marcel Dekker, in press. (b) Fulton,J . L. Structure and Reactions
in Microemulsions Formed in Near-Critical and Supercritical Fluids.
In Microemulsions: Fundamental and Applied Aspects; Kumar, P., Ed.;
Marcel Dekker, 1998, in press.
(7) Langevin, D. In Reverse Micelles; Luisi, P. L., Straub, B. E., Eds.;
Plenum: New York, 1984; pp 287-303.
10.1021/jo981825i CCC: $18.00 © 1999 American Chemical Society
Published on Web 01/22/1999

1202 J . Org. Chem., Vol. 64, No. 4, 1999
J acobson et al.
and gas-like diffusivities and viscosities.14 The adjust-
able physical properties of supercritical fluids, for ex-
ample, the density, have been utilized in areas such as
chromatography and extraction,15 synthesis,16 and ma-
terials processing. Supercritical and near-critical alkanes,
such as ethane and propane, have been used to form
water/alkane microemulsions.17 Recently, water-in-CO2
to form microemulsions, less surfactant may be required
for CO2 than for alkane continuous phases, since γ
without surfactant is only 18 mNm-1
.
In this paper, we report the first attempt to perform a
synthetic organic reaction in a w/c microemulsion. As a
simple model we chose to study the reaction between
benzyl chloride 1 and potassium bromide to form benzyl
bromide 2. This reaction between a nonaqueous com-
pound soluble in CO2 and a CO2-insoluble salt may be
expected to take place at or near the surfactant interface.
This reaction has previously been performed in both W/O
and O/W microemulsions,23 with triphase catalysis24 and
with PTC in supercritical CO2.2 Our results will be
compared to these studies.
(w/c) microemulsions were formed with a hybrid hydro-
+ 19
carbon-fluorocarbon surfactant18 and PFPE COO-NH4
.
For the latter microemulsion, water pools with proper-
ties approaching bulk water were characterized with
FTIR, fluorescence, UV-vis, and electron paramagnetic
resonance spectroscopy and by neutron scattering.19,20
These microemulsions may be used to perform inter-
facial inorganic reactions, e.g., between CO2-soluble
gases, such as H2S and SO2, and water-soluble inorganic
salts.21
Carbon dioxide is an attractive alternative to organic
solvents as it is environmentally benign, essentially
nontoxic, inexpensive, nonflammable, and has relatively
low critical conditions (Pc ) 73.8 bar, Tc ) 31 °C). Its
limited use in organic synthesis has been due to its weak
van der Waals forces, as characterized by its low polar-
izability per volume, making it unable to solvate many
polar and ionic nonvolatile materials. Solubilities can be
increased by adding small amounts of cosolvent (e.g.,
methanol), although a majority of the hydrophilic com-
pounds such as proteins are still very insoluble.
The possibility of achieving high solubilities of both
hydrophilic and hydrophobic compounds in w/c micro-
emulsions offers a new opportunity for synthetic organic
chemistry where no toxic organic solvents or catalysts
are needed. Another advantage of these w/c microemul-
sions is that only a modest amount of surfactant (<1.5
wt %) is needed to achieve a sufficiently low interfacial
tension to form the microemulsion. The interfacial ten-
sion, γ, between H2O and CO2 is on the order of 18
mNm-1 at pressures above 70 bar,22 much lower than
that between H2O and alkanes (30-50 mNm-1).9 The
lower γ may facilitate reaction between organic sub-
strates and hydrophilic reactants at the interface. This
lower γ reflects the fact that CO2 is more miscible with
water than are alkanes, because of its acidity and
quadrupole moment. To achieve low values of γ required
Carboxylic ester hydrolyses, catalyzed by nucleophiles
and enzymes, are commonly occurring reactions in bio-
logical systems. Therefore the investigation of ester
hydrolyses in w/c microemulsions is of obvious impor-
tance. To further compare the use of supercritical carbon
dioxide to hydrocarbons as a continuous phase in micro-
emulsions for synthetic organic chemistry, we chose to
study the hydrolysis of benzoyl chloride 3 and p-nitro-
phenyl chloroformate 5. Both of these substrates are
water-insoluble and have been studied extensively in
microemulsions.25
The addition of reactants to the microemulsion can
have a large effect on the phase behavior of the system,
especially for the high concentrations of substrates
desired for synthetic purposes. Phase behavior studies
were therefore conducted to ensure that all reactions
were performed in a single homogeneous phase.
(14) (a) Squires, T. G.; Vernier, C. G.; Aida, T. Fluid Phase Equilibria
1983, 10, 261-268. (b) Tiltscher, H.; Hofman, H. Chem. Eng. Sci. 1987,
42, 959-977.
(15) Markides, K. E.; Lee, M. L. Analytical Supercritical Fluid
Chromatography and Extraction; Chromatography Conference, Inc.:
Provo, 1990.
(16) Savage, P. E.; Gopalan, S.; Mizan, T. I.; Martino, C. J .; Brock,
E. E. AIChE J . 1995, 41, 1723-1778.
Exp er im en ta l Section
(17) (a) Gale, R. W.; Fulton, J . L.; Smith, R. D. J . Am. Chem. Soc.
1987, 109, 920-92. (b) J ohnston, K. P.; Penninger, J . M. L., Eds.
Supercritical Fluid Science and Technology, ACS Symposium Series
406; American Chemical Society: Washington, DC, 1989. (c) Fulton,
J . L.; Smith, R. D. J . Phys. Chem. 1988, 92, 2903-2907. (d) Yazdi, P.;
McFann, G. J .; Fox, M. A.; J ohnston, K. P. J . Phys. Chem. 1990, 94,
7224-7232.
Gen er a l. AOT (bis(2-ethylhexyl) sodium sulfosuccinate) was
obtained from Aldrich and used as received after drying under
vacuum at 60 °C for at least 10 h. Benzyl chloride, potassium
bromide, benzoyl chloride, p-nitrophenyl chloroformate, ani-
sole, naphthalene, and n-octane were all obtained from Aldrich
and used without further purification. Instrument grade
carbon dioxide (Praxair Inc.) was passed through an oxygen
(18) Harrison, K.; Goveas, J .; J ohnston, K. P.; O′Rear, E. A.
Langmuir 1994, 10, 3536-3541.
(19) J ohnston, K. P.; Harrison, K. L.; Clarke, M. J .; Howdle, S. M.;
Heitz, M. P.; Bright, F. V.; Carlier, C.; Randolph, T. W. Science 1996,
271, 624-626.
(20) (a) Heitz, M. P.; Carlier, C.; deGrazia, J .; Harrison, K.; J ohnston,
K. P.; Randolph, T. W.; Bright, F. V. J . Phys. Chem. 1997, 101, 6707.
(b) Zielinski, R. G.; Kline, S. R.; Kaler, E. W.; Rosov, N. Langmuir
1997, 13, 3934-3937.
(21) Clarke, M. J .; Harrison, K. L.; J ohnston, K. P.; Howdle, S. M.
J . Am. Chem. Soc. 1997, 119, 6399-6406.
(22) (a) Harrison, K. L., Da Rocha, S., J ohnston, K. P., in prepara-
tion. (b) Harrison, K. L. Ph.D. Dissertation, University of Texas, 1996.
(23) Martin, C. A.; McCrann, P. M.; Angelos, G. H.; J aeger, D. A.
Tetrahedron Lett. 1982, 23, 4651-4654.
(24) (a) Regen, S. L. J . Org. Chem. 1977, 42, 875-879. (b) J o′czyk,
A.; Ludwikow, M.; Makosza, M. Angew. Chem., Int. Ed. Engl. 1978,
17, 62-63.
(25) (a) Al-Lohedan. H.; Bunton, C. A.; Mhala, M. M. J . Am. Chem.
Soc. 1982, 104, 6654-6660. (b) Garc´ıa-R´ıo, L.; Leis, J . R.; Iglesias, E.
J . Phys. Chem. 1995, 99, 12318-12326.

Organic Synthesis in H2O/CO2 Microemulsions
J . Org. Chem., Vol. 64, No. 4, 1999 1203
trap (Oxyclear model no. RGP-31-300, Labclear, Oakland, CA)
before entering an ISCO syringe pump (model 260D) which
was used to load the reaction cell. Nanopure II water (Barn-
stead, Dubuque, IA) was used in all experiments. The PFPE
COO-NH4 was a gift from Alba Chittofrati, synthesized
+
according to published procedures.26 For the benzyl bromide
synthesis the molecular weight of PFPE COO-NH4+ was 740
g/mol, and for the hydrolysis it was 940 g/mol.
Product mixtures were analyzed by GC (Hewlett-Packard
model 5890A) equipped with an fused silica precolumn (i.d. )
0.53 mm, 0.5 m, methylated, Supelco) followed by a SPB-5
column (i.d. ) 0.53 mm, 30 m, Supelco). Cool on-column
injection and FID detection were used. Appropriate response
factors relative to an internal standard (anisole or naphtha-
lene) were determined for each substance analyzed. Nitrogen
was used as carrier gas at 30 mL/min, and the following
temperature gradients were used for analyzing the synthesis
of (A) benzyl bromide, 80 °C, 0.5 min; 15 °C/min; (B) benzoic
acid, 75 °C, 0.5 min; 10 °C/min; and (C) p-nitrophenol, 80 °C,
0.5 min; 20 °C/min.
High -P r essu r e Ap p a r a tu s. The reactions were performed
in a variable-volume view cell described previously.27 The
11
cylindrical stainless steel cell (2 in. o.d. ×
/
in. i.d., 35.2
16
mL total volume) was equipped with a sapphire window (1 in.
diameter by /8 in. thick). A piston inside the cell allowed the
F igu r e 1. Cloud point pressure versus temperature for PFPE
3
COO-NH4 w/c microemulsions with varying Wo and salt
concentrations. The two curves without added salt are taken
from Heitz; et al. J . Phys. Chem. 1997, 101, 6707.
+
pressure and temperature to be varied independently. The
reaction mixture was stirred by the use of a Teflon-coated
magnetic stir bar. A gear pump (1750 rpm, Baldor Electric
Co., Ft. Smith, AR) was used to constantly recirculate the
solution through a 6-port 2-position valve (C6W, Valco Instru-
ments Co., Inc., Houston, TX) equipped with a 100 µL sample
loop, allowing samples to be taken throughout the reaction
process. At the desired sampling time the valve was switched
and the sample was trapped in ca. 500 µL of liquid solvent for
analysis by GC.
The surfactant, water, KBr (when used), internal standard,
and a stir bar were loaded into the reaction cell, which was
then filled with a known volume of CO2. The cell assembly
was placed in a polycarbonate water bath, which also served
as a safety shield, and temperatures were controlled to (0.1
°C. After the microemulsion was formed at the desired
temperature and pressure by stirring the cell contents, the
hydrophobic reactant was injected into the cell from the sample
loop, starting the reaction.
Resu lts
P h a se Beh a vior Stu d y. The phase behavior of the
PFPE COO-NH4 w/c microemulsion was measured to
+
study the effect of added salt on the cloud point pressure.
The goal was to determine the maximum amount of salt
that could be dissolved in the aqueous micellar phase at
a certain temperature and pressure while still forming
a homogeneous microemulsion. The results are shown in
Figure 1, which also contains previously published data
for this system where no salt was added.19 The size of
the water pool in the aqueous core is proportional to the
Wo value, the molar ratio of water to surfactant. At a
given temperature and pressure, a certain amount of
water is soluble in the CO2 phase.28 This amount was
subtracted from the total amount of water added to the
cell, to determine a corrected Wo. At a given temperature
and Wo, the addition of salt lowers the cloud point
pressure. The same behavior was observed for the
propane/AOT/CO2 system. For this system, it has been
demonstrated theoretically and experimentally that the
maximum Wo where a one-phase microemulsion is stable
is governed by micelle-micelle interactions.29 These
interactions cause water and surfactant to precipitate at
a Wo above this maximum. Apparently, the presence of
salt weakens these interactions leading to a larger Wo
at a given pressure. At a given Wo and pressure, a
minimum temperature is required to form a one-phase
microemulsion. Below this minimum temperature, a two-
phase system is observed consisting an upper w/o micro-
emulsion phase and a lower aqueous phase.
Low -P r essu r e Exp er im en ts. The experiments using the
AOT surfactant in H2O/n-octane were performed in a 100 mL
round-bottomed flask at atmospheric pressure. AOT, water,
n-octane, internal standard, and salt was added together with
a stir bar into the flask which was then heated in a water bath
(as described above). After the microemulsion formed, the
hydrophobic reactant was added and the reaction was followed
by analyzing aliquots of the reaction mixture by GC.
P h a se Beh a vior Stu d ies. The same high-pressure experi-
mental setup was used with the exception that the gear pump
and the Valco valve were disconnected from the apparatus and
two ports were plugged. The required amount of PFPE
+
COO-NH4 needed to obtain a 1.4 wt % solution in CO2 was
loaded into the variable volume view cell together with the
desired amount of water. After the cell was closed, carbon
dioxide was added into the cell and also used to pressurize
the cell from the back of the piston. The solution was stirred
for at least 1 h at the desired temperature and at a pressure
to ensure a homogeneous microemulsion. The stirring was
stopped, and the pressure was gradually lowered until a cloudy
solution was seen through the sapphire window. This pressure
was defined as the cloud point pressure and was measured at
least three consecutive times. The mean value at each tem-
perature was reported. The typical uncertainty was <10
psia.
Syn th esis of Ben zyl Br om id e in Micr oem u lsion s.
The reaction of KBr and benzyl chloride in w/c micro-
emulsions was studied with the surfactant PFPE-
COO-NH4+. The effect of Wo, substrate concentration,
(28) (a) Wiebe, R.; Gaddy, V. L. J . Am. Chem. Soc. 1939, 61, 315-
318. (b) Wiebe, R.; Gaddy, V. L. J . Am. Chem. Soc. 1940, 62, 815-
817. (c) Wiebe, R.; Gaddy, V. L. J . Am. Chem. Soc. 1941, 63, 475-477.
(29) (a) Peck, G. P.; J ohnston, K. P. J . Phys. Chem. 1991, 95, 9549-
9556. (b) Kaler, E. W.; Billman, J . F.; Fulton, J . L.; Smith, R. D. J .
Phys. Chem. 1991, 95, 458-462.
(26) Chittofrati, A.; Lenti, D.; Sanguineti, A.; Visca, M.; Gambi, C.
M. C.; Senatra, D.; Zhou, Z. Prog. Colloid Polym. Sci. 1989, 79, 218-
225.
(27) Lemert, R. M.; Fuller, R. A.; J ohnston, K. P. J . Phys. Chem.
1990, 94, 6021-6028.

1204 J . Org. Chem., Vol. 64, No. 4, 1999
J acobson et al.
Ta ble 1. Rea ction of Ben zyl Ch lor id e a n d KBr in P F P E
Ta ble 2. Rever sible, Secon d -Or d er Kin etics for th e
+
COO-NH4 w /c Micr oem u lsion sa
Rea ction of Ben zyl Ch lor id e + KBr
[KBr]
(M)b
[BnCl]
(mM)c
mol KBr/
mol BnCl
yield
(%)
K
kf (M-1 h-1
)
equilibrium convn (%)
Wo
2
3
4
0.097
0.097
0.097
3.8
0.99
4.4
28
60
31
1
2
3
4
5
4.6
9.7
9.7
9.7
9.7
4.7
1.2
4.1
4.1
4.1
1.9
2.2
1.1
6.4
10
3.5
1.4
9.7
1.7
1.0
6 ( 2
9 ( 3
17 ( 2
30 ( 3
28 ( 3
mechanism of the interfacial reaction may have been
somewhat different for the lower [BnCl] and the high
ratio of KBr to BnCl. Increasing the [KBr] from 1.2 to
4.1 M had a fairly large effect on the product yield
(compare runs 2 to 3) as expected. The maximum [KBr]
that could be achieved at a Wo ≈ 10 was 4.1 M; adding
more salt split the homogeneous microemulsion into a
two-phase system.
a
Reaction conditions: 65 °C, 276 bar, 16 h, 1.4 wt % surfactant.
b
The product mixtures were analyzed by GC. Based on the total
volume of water. c Based on the volume of CO2.
Using otherwise the same reaction conditions as in
runs 1-5, these experiments were also conducted at 45
°C which in all cases decreased the product yields to <2%.
Increasing the surfactant concentration to 3 wt % (30
mM) at both 45 and 65 °C did not, on the other hand,
have any significant effect on the yields. The effect of
pressure was only significant if it was lowered below the
cloud point pressure where a two-phase system and lower
yields were observed. No traces of benzyl hydroxide
detectable by GC were observed in any of these reactions,
suggesting that the reaction took place in a nonaqueous
environment, e.g., at the micellar interface.
To place these results in perspective, experiments were
also performed with the commonly used anionic surfac-
tant system AOT/H2O/n-octane. For the same surfactant
concentration, Wo, and reactant concentrations as in run
4, experiments were performed at atmospheric pressure
at both 45 and 65 °C. This allowed the use of supercritical
CO2 as a continuous phase to be compared to the more
commonly used hydrocarbon solvents. With AOT/H2O/n-
octane microemulsions, less than 2% product yield was
obtained in 24 h at both 45 and 65 °C.
F igu r e 2. Yield for the reaction of benzyl chloride and KBr
in PFPE COO-NH4+ w/c microemulsions at 65 °C and 276 bar.
The lines are fits of the data, and the constants are given in
Table 2.
Hyd r olysis Rea ction s. The hydrolysis of benzoyl
chloride 3 and p-nitrophenyl chloroformate 4 in PFPE
+
COO-NH4 w/c and AOT/H2O/n-octane microemulsions
temperature, and pressure on the yield of benzyl bromide
was investigated as shown in Table 1. These reactions
were all run at 65 °C and 276 bar for 16 h with 1.4 wt %
(16 mM) surfactant. From run 1 vs run 3-5, it can be
concluded that lowering the Wo decreases the product
yield. Runs 3-4 show that at constant Wo and [KBr] an
increase in [BnCl] increases the yield. After that adding
more BnCl (run 5) does not increase the yield further.
The [BnCl] was not increased past 10 mM, where the
molar ratio of KBr/BnCl was unity.
were performed to further study the use of supercritical
CO2 as a nonaqueous phase for synthetic chemistry. Both
of these substrates are water-insoluble. In these experi-
ments, the concentration of the hydrophilic reactant,
water, was much higher than the KBr concentration in
the benzyl bromide synthesis. All reactions in CO2 were
performed at 35 °C and 276 bar. A substrate concentra-
tion of 0.05 M was used in order to achieve a great excess
of water to ensure pseudo- first-order kinetics. At a Wo
of 10, the yield was 45% for the hydrolysis of benzoyl
chloride in 25 h and for that of p-nitrophenyl chlorofor-
mate in 9 h.
To elucidate the factors affecting the rates of the
nucleophilic substitution, reactions in water-in-CO2 mi-
croemulsions, runs 2-4, were fit to reversible, second-
order kinetics of the form
The observed pseudo-first-order rate constant kobs
d[BnCl]
kobs ) k[H2O]
-
) kf[BnCl][Br-] - kr[BnBr][Cl-]
dt
was measured for the hydrolysis reactions at Wo values
of 5 and 10 (Table 3). At a Wo of 5 the observed rate
constant for hydrolysis of p-nitrophenyl chloroformate is
slightly higher than that for benzoyl chloride in the
respective microemulsion systems. The reverse was
observed at the higher Wo value of 10, with slightly higher
reaction rates for benzoyl chloride in both cases. In all
cases though, the reaction rates were an order of mag-
nitude lower for the w/octane as compared to the w/c
microemulsions.
where kf and kr are forward and reverse rate constants
and all concentrations are in terms of total volume. It
was assumed that the value of the equilibrium constant
K () kf/kr) was equal in all experiments. Figure 2 shows
the fits of the kinetic model, and Table 2 contains the
regressed equilibrium and forward rate constants, along
with the calculated equilibrium conversions. The rate
constants for runs 2 and 4 were in good agreement; the
reason for the lower value for run 3 was not obvious. The

Organic Synthesis in H2O/CO2 Microemulsions
Ta ble 3. Hyd r olysis Rea ction s in Micr oem u lsion sa
Wo ) 5
J . Org. Chem., Vol. 64, No. 4, 1999 1205
Wo ) 10
substrate
benzoyl chloride
microemulsion
kobs (s-1
)
Rb
kobs (s-1
)
Rb
H2O/PFPE/CO2
H2O/AOT/n-octane
H2O/PFPE/CO2
1.7 × 10-6
3.0 × 10-7
2.9 × 10-6
5.9 × 10-7
0.99
0.94
0.97
0.99
6.4 × 10-5
3.4 × 10-6
1.6 × 10-5
7.6 × 10-7
0.98
0.92
0.94
0.99
p-nitrophenyl chloroformate
H2O/AOT/n-octane
a
Conditions: 35 °C, 276 bar in the case of CO2 and atmospheric pressure for n-octane, [substrate] ) 0.05 M. The product mixtures
b
were analyzed by GC. Correlation coefficient for the least-squares fit of ln c/co vs time.
Discu ssion
efficient. A disadvantage however is that high pressures
are needed.
KBr is insoluble in CO2, and benzyl chloride is in-
soluble in water. As a consequence, the reaction must
either occur at the interphase or some other mechanism
must be present to transfer one reactant into the other
phase. No benzyl hydroxide, detectable by GC, was
observed in any of the reactions, also suggesting that the
reaction takes place in a nonaqueous environment, e.g.,
at the micellar interface.23,24a
The structure of the interfacial water in reverse
micelles has been shown to change with increasing Wo.30
+
In the case of a surfactant such as PFPE COO-NH4
,
added water initially hydrates the anionic headgroups,
first binding to the anionic side of the ion pair and
thereafter hydrating the cation.25b Further added water
is incorporated into the micellar core, and as Wo increases
the micelles swells. Water in the periphery is more
structured and has different physical properties, e.g.,
lower polarity, as compared to bulk water.31 Only in large
water droplets do the physical properties resemble those
of bulk water.
The limiting factor in the synthesis of benzyl bromide
was the amount of salt that could be dissolved in the
aqueous micellar core. Without any added salt the
maximum Wo value that has been reported for PFPE
COO-NH4 w/c microemulsions is 20.18 With high con-
+
The effect of varying physical properties of the inter-
facial water was observed in the change of hydrolysis
rates for benzoyl chloride and p-nitrophenyl chloroform-
ate in the microemulsions. The fact that the hydrolysis
rate of benzoyl chloride was greater at a Wo of 10 as
compared to the rate of p-nitrophenyl chloroformate
hydrolysis and that the reverse is true at a Wo of 5 is
consistent with earlier published results25b and has been
explained by the difference in mechanism between the
two substrates.25 For both compounds the mechanism
involves nucleophilic addition of water at the carbonyl
carbon, giving a tetrahedral intermediate, and cleavage
of the C-Cl bond. In the case of benzoyl chloride the
C-Cl bond breaking has progressed further in the
transition state, as compared to p-nitrophenyl chloro-
formate, due to conjugative electron release by the aryl
group which assists in the bond breaking. Therefore the
reaction is less sensitive to the nucleophilicity of water
and is instead favored by higher water polarity (which
increases with increasing Wo) which stabilizes the cat-
ionic acylium ion and also the nucleofuge. The aryloxy
group in p-nitrophenyl chloroformate assists nucleophilic
attack, due to its inductive electron attraction, and the
reaction is more sensitive to the nucleophilicity of the
water molecules in the micelles. A reduction in the
droplet size (Wo) leads to concentration of the anionic
charge of the surfactant headgroups and stronger hydro-
gen bonding between the surfactant and water, thereby
enhancing the nucleophilicity of the interfacial water.25b
For p-nitrophenyl chloroformate hydrolysis, only a slight
decrease in rate was observed as Wo was decreased from
10 to 5, much less than that for the hydrolysis of benzoyl
chloride. Thus, the different effects of Wo on these two
reactions is consistent with the expected changes in
nucleophilicity and polarity of water.
centrations of added salt (up to 4 M), it was not possible
to obtain a Wo above 10 at temperatures of 65 °C. Higher
salt concentrations could be obtained at lower tempera-
tures although the decrease in rate due to decreased
temperature is far greater than the increase obtained by
higher reactant concentrations.
In previous studies of nucleophilic substitution reac-
tions in microemulsions, it has been shown that cationic
surfactants can increase the reaction rate as compared
to nonionic and especially anionic surfactants. In the
latter case inhibition of the reaction rate has even been
reported. These results are explained by the fact that a
negatively charged transition state is stabilized by a
cationic surfactant headgroup, thereby attracting it to
the interphase where the reaction takes place. To date
cationic surfactants have not been developed which can
form w/c microemulsions. However, our product yields
with the anionic surfactant PFPE COO-NH4 are com-
+
parable to or even greater than those reported by Martin
et al.23 in a w/o microemulsion stabilized by a cationic
surfactant at comparable reaction conditions. Further-
more, in their study, the surfactant concentration was 6
times higher and butanol was needed as a cosolvent.
Higher yields for this reaction have been reported in
o/w microemulsions,23 although in this case the surfactant
concentration was quite high (37 wt %). The byproducts
benzyl hydroxide and benzyl butyl ether were also
observed. Higher yields can also be obtained by adding
a catalyst as in traditional PTC or triphase catalysis.24
Dillow et al. added a PTC catalyst to supercritical CO2,2
with a small amount of acetone as a cosolvent to solu-
bilize the catalyst. The larger KBr to benzyl chloride ratio
compared with that in the pesent study led to a higher
equilibrium conversion.
An advantage of the w/c microemulsions for phase
transfer reactions is that organic solvents and catalysts,
both which may be toxic, are not needed. Only small
amounts of PFPE COO-NH4+, an environmentally be-
nign surfactant, are required. After reaction, the solvent
strength of CO2 may be tuned to separate reactants and
product, and solvent recovery and recycle are particularly
For both hydrolysis reactions the rates were an order
of magnitude higher for the w/c system as compared to
(30) Wong. M.; Thomas, J . K.; Gra¨zel, M. J . J . Am. Chem. Soc. 1976,
98, 2391.
(31) (a) Mukerjee, P. In Solution Chemistry of Surfactants; Mittal,
K. L., Ed.; Plenum Press: New York, 1979; Vol. 1, p 153. (b) Cordes,
E. H.; Gitler, C. Prog. Bioorg. Chem. 1973, 2, 1.

1206 J . Org. Chem., Vol. 64, No. 4, 1999
J acobson et al.
H2O/AOT/n-octane under otherwise similar reaction con-
ditions. In earlier studies of the latter system,25b it was
shown that addition of certain additives (such as sodium
perchlorate and 1,3-dimethyl urea) increased the hy-
drolyses rates of these reactions by decreasing the
viscosity at the micellar interface. An overall lowering
of the microviscosity (ca. 30% less than bulk water)
soluble reactants can be performed in a w/c microemul-
sion containing small (<1.5 wt %) amounts of an envi-
ronmentally benign surfactant, PFPE COO-NH4+. No
organic solvent is needed. Reaction yields and rate con-
stants were larger by an order of magnitude in w/c micro-
emulsions versus H2O/AOT/n-octane microemulsions un-
der similar conditions (except pressure). The enhanced
reaction rates are likely due to the considerably lower
microviscosity in w/c as compared to w/o microemulsions.
The use of a tunable supercritical fluid offers the pos-
sibility of fractionation of products and may facilitate
recovery and recycle of surfactant and CO2. In the future,
it would be desirable to find other microemulsion systems
that could dissolve even larger amounts of water in CO2.
within the PFPE COO-NH4 microemulsion core has
+
been suggested due to presence of CO2.20 This reduction
in microviscosity likely contributes to the higher hydroly-
sis rates relative to w/o microemulsions.
Performing hydrolyses reactions in w/c microemulsions
could be advantageous in cases where a water-insoluble
substrate is involved. High concentrations of certain
organic compounds can be dissolved in carbon dioxide,
and the large surface area and low microviscosity of these
microemulsions facilitates the interfacial reaction.
Ack n ow led gm en t. We acknowledge support from
the Swedish Research Council for Engineering Sciences,
NSF (CTS-9626828), DOE (DE-FG03-96ER14664, and
the Separations Research Program at the University of
Texas.
Con clu sion s
We have demonstrated that organic synthesis between
hydrophilic nucleophiles (Br- and water itself) and CO2-
J O981825I