Supported Ionic Liquid-Like Phases (SILLPs) for enzymatic processes: Continuous KR and DKR in SILLP-scCO2 systems

Article abstract of DOI:10.1039/c0gc00076k

The immobilisation of Candida antarctica lipase B (CALB) onto SILLPs has been studied in order to obtain the maximum activity, selectivity and stability of the resulting supported biocatalyst for the KR of rac-1-phenylethanol by flow processes using scCO₂. This involves the use of hydrophilic SILLPs (chloride anion) with high loadings of IL moieties and CALB. The combination of the immobilized biocatalyst with an acidic zeolite has allowed us to carry out DKR processes. Excellent results in terms of yield (92%) and enantioselectivity (>99.9% ee) are obtained by the use of a one-pot reactor containing a mixture of both catalyst particles under scCO₂ flow conditions (50 °C, 10 MPa), resulting in an environmentally friendly process with high productivity and enantioselectivity.

Full text of DOI:10.1039/c0gc00076k

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/greenchem | Green Chemistry
Supported Ionic Liquid-Like Phases (SILLPs) for enzymatic processes:
Continuous KR and DKR in SILLP–scCO2 systems†
Pedro Lozano,*a Eduardo Garc´ıa-Verdugo,*b,c Naima Karbass,b Kimberley Montague,a Teresa De Diego,a
M. Isabel Burgueteb and Santiago V. Luis*b
Received 3rd May 2010, Accepted 13th August 2010
DOI: 10.1039/c0gc00076k
The immobilisation of Candida antarctica lipase B (CALB) onto SILLPs has been studied in order
to obtain the maximum activity, selectivity and stability of the resulting supported biocatalyst for
the KR of rac-1-phenylethanol by flow processes using scCO2. This involves the use of hydrophilic
SILLPs (chloride anion) with high loadings of IL moieties and CALB. The combination of the
immobilized biocatalyst with an acidic zeolite has allowed us to carry out DKR processes.
Excellent results in terms of yield (92%) and enantioselectivity (>99.9% ee) are obtained by the
use of a “one-pot” reactor containing a mixture of both catalyst particles under scCO2 flow
conditions (50 ◦C, 10 MPa), resulting in an environmentally friendly process with high
productivity and enantioselectivity.
an IL has proved to be the best approach to protect enzymes
Introduction
against these adverse effects of scCO2.10 The high solubility of
Ionic liquids (ILs) are simply salts with melting points below
the scCO2 phase in ILs and, simultaneously, the low solubility of
the ILs in the in the scCO2 phase, allows the supercritical fluid
to be used as the medium to deliver the products to the enzyme
immobilised in the IL and to extract the obtained products from
them. Therefore, it is easy to combine reaction and separation
protocols, leading to integral greener bioprocesses in non-
aqueous media.11
The aim of this work is to push towards the efforts to design
greener enzymatic processes in IL/scCO2 biphasic systems. The
catalytic system is based on an enzyme (Candida antarctica
Lipase B, CALB) immobilised by non-covalent attachment
onto different polymeric Supported Ionic Liquid-Like Phases
(SILLPs). SILLPs are polymeric matrices modified with IL-
like units.12 They can be regarded as “solid ionic solvents” or
as nanostructured materials with microenvironments of tunable
polarity similar to those provided by bulk ionic liquids.13 The
solid nature of the SILLPs facilitates the separation processes,
allowing the development of continuous flow processes and
avoiding accidental spills. This is important in the light of the
toxicological and ecotoxicological concerns recently raised for
some ILs.14 In addition, it reduces the cost of the process, as
smaller amount of ILs need to be employed.15
We have shown that the adsorption of a lipase onto this
linked IL-like phase provides excellent immobilized biocatalysts.
Indeed, these show enhanced activity and increased operational
stability for the continuous synthesis of citronellyl butyrate in
scCO2, compared with the original strategy based on enzymes
coated with ILs.16 The present study reports on the use of
different CALB–SILLPs as biocatalysts for a stereospecific
reaction model such as the kinetic resolution (KR) of rac-1-
phenylethanol in both batch and continuous process. The bio-
catalysts show good performance in both a conventional organic
solvent (hexanes) and in supercritical carbon dioxide (scCO2).
In addition, the continuous one-pot dynamic kinetic resolution
(DKR) of this substrate in scCO2 by in situ racemization of the
◦
100 C, usually close to room temperature.1 Their interest
as green neoteric solvents resides in their negligible vapour
pressure, their excellent thermal stability, and their high ability
to dissolve a wide range of organic and inorganic compounds,
including gases and their non-flammable nature.2 Thus, ILs
can be used to mitigate the problem of volatile organic sol-
vent emission to the atmosphere and have been implemented
as environmentally friendly alternative solvents for various
processes.3 Thus, ILs have emerged as exceptionally interesting
non-aqueous reaction media for enzymatic transformations,4
mainly for asymmetric synthetic transformations, due to the
growing demand for enantiopure pharmaceuticals.5
In the search for cleaner biocatalytic processes, the use of
supercritical fluids (SCFs) is another interesting alternative.6
However, in some cases, the poor stability exhibited by enzymes
in SCFs, particularly in scCO2, is the main drawback for the
application of biocatalytic processes in SCFs.7
Alternatively, different approaches based on the coating of
biocatalysts by ILs or on the suspension of the enzyme in the
IL have been developed to improve the stability and activate the
enzyme.8,9,10 Thus, the synergistic combination of both SCFs and
ILs for enzyme catalysis can led to greener and highly efficient
processes. Indeed, the coating or suspension of biocatalysts in
aDepartamento de Bioqu´ımica y Biolog´ıa Molecular B e Inmunolog´ıa,
Facultad de Qu´ımica, Universidad de Murcia, P.O. Box 4021E-30100,
Murcia, Spain. E-mail: plozanor@um.es
bDepartamento de Qu´ımica Inorga´nica y Orga´nica, Universidad Jaume I,
Campus del Riu Sec, Avenida Sos Baynant s/n, E-12071, Castello´n,
Spain. E-mail: luiss@qio.uji.es
cInstituto de Cata´lisis y Petroleoqu´ımica, CSIC, Campus de la UAM,
C/Marie Curie, 2; Cantoblanco, Madrid, E-28049, Spain.
E-mail: egverdugo@icp.csic.es; Tel: +34 91585 4800
† Electronic supplementary information (ESI) available: Fig. S1–S3 and
Tables S1 and S2. See DOI: 10.1039/c0gc00076k
This journal is
The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 1803–1810 | 1803
©

View Article Online
undesired enantiomer is also reported with good chemical yield
(>85%) and excellent enatiopurity (>99% ee).
where the subscripts S and P stand for substrate and product,
respectively.18 Enantiomeric excesses for the synthetic products
were calculated as:
eeS = [(%(S)-12 - %(R)-12)]/[%(S)-12 + %(R)-12)] ¥ 100
eeP = [(%(S)-15 - %(R)-15]/[(%(S)-15 - %(R)-15)] ¥ 100
One unit of synthetic activity was defined as the amount of
enzyme that produces 1 mmol of (R)-1-phenylethylpropionate
per min. All the experiments were carried out in duplicate
Experimental
Safety note
Experiments with scCO2 involve high pressure and should only
be carried out in equipment with the appropriate pressure rating.
Kinetic resolution of rac-phenylethanol in scCO2
Materials
The immobilized enzymes (300 mg) were packed into a stainless-
steel column (14 inch external diameter, 10 cm length) and then
introduced into an oven at 50 ◦C. The system was equipped with
a supercritical pump (JASCO, model PU-1580-CO2), needle
valves, filter and a back-pressure regulator (JASCO, model BP
2080). The reactor was run by the continuous pumping of scCO2
at 10 MPa through the packed immobilized enzyme, then the
back-pressure regulator, and finally bubbling through a capillary
tube at 1 mL/min total flow in a controlled amount of n-
hexane placed on ice-bath. The KR process was carried out
by the continuous pumping of an equimolar solution of rac-
1-phenylethanol and vinyl propionate at 21.2 mmol/min mass
flow-rate, by using a HPLC pump (model LC-10AT, Shimadzu
Europe, Duisburg, Germany) for 4 h. Collected samples were
analyzed by GC, the substrates/products mass-balance from
the outlet being consistent with the substrate mass-flow inlet.
The aqueous solution of Candida antarctica lipase B (Novozym
525 L, EC 3.1.1.3) was a gift from Novozymes S.A. Acid zeolites
CP811C (zeolite b, 620 m2/g), CBV720 (zelolite Y, 720 m2/g)
and CP811E (zeolite b, 600 m2/g) were obtained from Zeolysts
International (Valley Forge, PA, USA). Substrates, solvents and
other chemicals were purchased from Sigma-Aldrich-Fluka Co.
(Madrid, Spain), and were of the highest purity available.
Preparation of Supported Ionic Liquid-Like Phases 4–6
(SILLPs)
The protocol for preparation of the supported ionic liquid
phases 4–6, bearing different covalently attached IL-like moi-
eties, has been previously reported in detail.12 All these mate-
rials were characterized by various techniques, including FT-
IR, Raman microspectroscopy, and elemental analysis, which
confirmed their expected structures. The quantitative exchange
of the chloride anion in the metathesis reaction was confirmed
by XPS analysis by the absence in the exchanged SILLPs of the
band corresponding to the chloride group.
Dynamic kinetic resolution of rac-phenylethanol in scCO2
Two fractions (150 mg) of immobilized derivative, and one
fraction of acid zeolite (100 mg of CP811C, CBV-720 or
CP811E coated with 150 mL of [Btma][NTf2]) were packed
in three separated columns. The three columns connected in
series (immobilized lipase–acid catalyst–immobilized lipase)
were operated in a continuous scCO2 flow for 4 h cycles by
using the system and procedure described above.
CALB immobilization onto SILLPs
The immobilized enzyme derivatives were prepared by sim-
ple adsorption of 1 mL of an aqueous solution of CALB
(18.2 mg/mL) onto SILLPs 4–6 (500 mg). The mixture was
shaken for 6 h at room temperature to adsorb the enzyme.
The supernatant was recovered and the support washed with
water to remove non-adsorbed enzyme. The polymer was
frozen at -60 ◦C and lyophilised. The supernatant and washing
fractions were collected and used to quantify the amount of
immobilized protein by Lowry’s modified method.17 All the
supported enzymes were stored under controlled water activity
(Aw) conditions over LiCl (Aw = 0.11) in desiccators for 48 h at
room temperature prior to use.
GC analysis
Analysis was performed with a Shimadzu GC-2010 (Shimadzu
Europe, Duisburg, Germany) equipped with an FID detector.
Samples were analyzed on a Beta DEX-120 column (30 m ¥
0.25 mm ¥ 0.25 mm, Supelco), using He as the carrier gas and
a FID detector, as described previously18 Retention times of
compounds are as follows: vinyl propionate (3.1 min), propionic
acid (5.6 min), butyl butyrate (internal standard, 7.1 min), (R)-
1-phenhylethanol (14.8 min), S-1-phenhylethanol (15.5 min),
(S)-1-phenylethyl propionate (18.2 min) and (R)-1-phenylethyl
propionate (18.6 min).
Batch kinetic resolution of rac-1-phenylethanol in hexane
Reactions were carried out in 1 mL screw-capped vials with
Teflon-lined septa containing 600 mL of substrate solution
(450 mM rac-1-phenylethanol, 950 mM vinylpropionate) in dry
hexanes. The reaction was started by adding the CALB–SILLP
(7–11) (25 mg), and run with magnetic stirring at 50 ◦C (oil-
bath) for 8 h. At regular time intervals, 20 mL aliquots were
taken, suspended in 480 mL hexanes containing 20 mM butyl
butyrate (internal standard), and then analysed by GC.
Results and discussion
Synthesis and characterisation of CALB–SILLPs
The polymers modified with IL-like moieties were synthesized
by grafting of butyl imidazole onto a commercial bead-type
Merrifield resin (1–3) leading to the corresponding SILLPs
4–6 (Scheme 1).12 Alternatively, Merrifield-type polymers with
different loadings (1.1 and 4.3 meq Cl/g) crosslinking degrees
and morphologies (gel: 2% DVB or macroporous: 5.5% DVB)
Profiles of (R)-1-phenylethylpropionate concentration with
respect to time were used to quantify the reaction rates of the
system. Conversion was calculated as %c = eeS/(eeS + eeP) ¥ 100,
1804 | Green Chem., 2010, 12, 1803–1810
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Scheme 1 Synthesis of the CALB–SILLPs.
were employed. Details about the synthesis and characterization
of these compounds have been published by us previously.12,16
The metathesis of chloride by NTf2- afforded more hydrophobic
SILLPs (4a–5a).
The immobilisation of the enzyme (CALB) was carried out
by adsorption from an aqueous solution of the enzyme as
described in the Experimental section. Table 1 summarises the
immobilisation yield reached for the different supports assayed.
As can be seen, all the supports were able to adsorb some
protein, with the percentage of immobilisation ranging from
ca. 7 to 60%. This difference in the immobilisation yield clearly
depends on both the nature of the support and the amount
and type of the IL-like moieties present in the SILLPs. Thus,
hydrophilic SILLPs having chloride as counterion (4a and 5a)
adsorbed less protein than the hydrophobic ones (Table 1, Entry
1 and 3 vs. 2 and 4). In contrast, the macroporous resin (6a)
adsorbed a significantly higher percentage of enzyme than the
gel-type analogue (Table 1, Entry 1 and 3 vs. 5). In the case of the
hydrophobic SILLP (4b and 5b) the adsorption depends of the
loading of the ionic liquid units. Thus, an increase in the loading
of the ILs from ca. 21% to ca. 78% by weight, led to a two-fold
increase in the enzyme loading (from 11 to 21 mg of protein/g
of SILLP). However, a similar trend was not observed when the
SILLPs with Cl as the counterion were used. In this case, the
high content in hydrophobic aminoacids placed at the surface of
CALB could be related to these results: the more hydrophobic
microenvironment for the SILLP, the better the immobilization
yield.19
Scheme 2 Kinetic resolution of rac-1-phenylethanol.
first. In all the cases the reaction was carried out in hexanes
at 50 ◦C using the same weight of polymer independent of
protein loading, and 24 h of reaction time. All the supported
enzymes were able to catalyse the selective formation of (R)-
1-phenylethyl propionate with excellent enantiomeric excess
(99.9% ee). However, the catalytic efficiency of the immobilised
enzymes was clearly dependent on the characteristics of the
SILLPs. In general, the catalysts CALB–SILLP 7 and 8 with
low IL-like loadings led to significantly lower yields (18% and
13% yield, respectively) than the analogues with higher IL
loading (50% for 9 and 36% for 10). The protein adsorbed in
macroporous SILLP (11) also showed an excellent activity and
enantioselectivity. Regarding the activity (U per g support), it
was found that the best immobilised catalyst, for the conditions
assayed, was CALB-SILLP 10 due to its higher protein loading
compared with the other CALB–SILLPs. However, if the specific
activity (U/mg of enzyme) was considered, the picture was quite
different. In this case, the more active catalyst was CALB–
SILLP 9.
In general, some trends can be detected: i) enzymes supported
on hydrophilic supports show higher specific activity than those
on hydrophobic supports,20 ii) higher IL-like loadings led to
significant improvement of the specific activity. Both tendencies
The ability of CALB–SILLPs 7–11 to catalyse the KR of rac-
1-phenyletanol with vinyl propionate (Scheme 2) was examined
Table 1 Immobilisation and catalytic parameters for Candida antarctica lipase B adsorbed onto different supportsa
Loading (meq
IL per g)b
Immobilised Immobilisation
f
Entry
R
X-
CALB–SILLPc % IL proteind
yield (%)e
c (%) eeP (%)
E
Activityg Specific activityh
1
2
3
4
5
6
C4H9 Cl-
0.76
7 (g)
8 (g)
9 (g)
10 (g)
11 (m)
—
21.0 2.6
31.3 11.4
64.5 3.0
78.7 21.2
23.5 17.5
7
32
8
59
48
41
18
13
50
36
50
—
99.9
99.9
99.9
99.9
99.9
—
244 18
230 15
1510 68
346 307
1510 195
7.0
1.3
22.5
14.4
11.2
0.3
-
C4H9 NTf2 0.60
C4H9 Cl-
2.33
C4H9 NTf2 1.51
-
C4H9 Cl-
0.85
—
—
—
—
15.0
—
0.02
a 100 mg of the support were suspended in 200 mL of a solution of CALB (18.2 mg/mL). Supplied protein: 36.3 mg per g support. b Based on
elemental analysis. c Type of resin: (g) = gel, (m) = macroporous. d mg of protein per g support. e % immobilisation yield = (mg immobilised protein
per g support)/(mg supplied protein per g support) ¥ 100, calculated by Lowry’s method. f eeP = [(%(S)-15 - %(R)-15]/[(%(S)-15 - %(R)-15)] ¥ 100.
g Activity = U (mmol product per mL per min) per g support. h U per mg protein.
This journal is
The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 1803–1810 | 1805
©

View Article Online
highlight the importance of the enzyme–IL interactions, as well
as the preservation of the water-shell around proteins, to provide
active and stable biocatalysts suitable for being used in scCO2.
Thus, the SILLPs having NTf2 as the counterion showed an
eleven-fold enhancement in the specific activity for a two-fold
protein loading increase. It is noteworthy that, for the gel-
type hydrophilic SILLPs, a three-fold increase of the specific
activity was observed for almost the same protein loading (ca.
of 3 mg per g support); iii) for catalysts 7 and 11 with similar IL
loadings and different polymeric morphology, the macroporous
resin yielded a slightly higher specific activity; iv) the specific
activity of the hydrophobic resins (8 and 9), when compared
with their protein content, seems to indicate an inactivation of
the enzyme by aggregation or by perturbation of the secondary
protein structure for higher loadings, v) although resin 3 was
able to immobilise ca. 40% of the protein supplied, the specific
activity was very small. The last two observations illustrate the
importance of having a large IL/enzyme ratio to stabilise the
immobilised enzyme.
Indeed, the specific activity of our supported enzyme CALB–
SILLP 11 with respect to a commercial one (Novozyme 435
without coating),21 calculated as U per mg of supported enzyme,
is higher than that found for the commercial catalyst under the
same experimental conditions and using the same amount of
supported biocatalyst (ca. 11 vs. 5 U per mg supported enzyme,
respectively).22
Most likely, a rational explanation for the different results
observed for the CALB–SILLPs can be related to the variation
of the secondary structures of the protein produced by the
immobilisation onto the different supported ionic liquids. The
infrared spectra of the proteins, in particular the amide I region
at 1700–1600 cm-1, can provide some valuable information about
the protein’s secondary structure in terms of a-helix, b-sheets,
b-turns and non-ordered or irregular structures.23 Therefore, the
FT-ATR-IR spectra of CALB–SILLP 9, CALB–SILLP 10 and
Novozyme 435 were recorded and analysed by deconvolution of
the amide I region (see ESI†).
the two first cases the immobilisation is be accompanied by a
loss of both a-helix and b-turn substructures. These structural
changes may well explain the low specific activity of these two
supported biocatalysts compared with CALB–SILLP 9.25 In
contrast, it seems that the protein in the SILLP bearing chloride
as the counterion (CALB–SILLP 9) preserves more efficiently
the active protein structure. The combination of hydrophilic IL-
like moieties and high IL/enzyme ratios clearly contribute to the
stabilization of the structure of the enzyme, as has been observed
in bulk ILs.
A study by differential scanning calorimetry (DSC) of the
CALB immobilised onto SILLPs (5a–b) and the commercial
immobilized CALB (Novozyme 434) was also carried out to
characterise the effect of the supported ILs on the thermal denat-
uration of the protein. The DSC traces showed a significant shift
of the maximum temperature observed from 52 ◦C (Novozyme
435) to 57 ◦C (CALB–SILLP 10, 4.3 meq Cl/g, R = C4H9, X =
NTf2) and 68 ◦C (CALB–SILLP 9, 4.3 meq Cl/g, R = C4H9, X =
Cl). Thus, the nature of the polymer and the anion contribute to
the stabilization of the supported protein. The results from both
FT-ATR-IR and DSC proved a significant effect of the SILLPS
over the stabilization of the secondary structure of the protein,
the support having chloride as the counterion and high IL-like
loading being the most efficient in terms of specific activity and
thermal stability.
Once we had proven that the SILLPs can be used as “solid
ionic solvents” for the immobilisation and stabilisation of the
CALB leading to efficient biocatalysis for the KR of rac-1-
phenylethanol, we focused on developing the system further. In
this regard, we studied several variations of the immobilisation
protocol to achieve the highest yield of immobilisation for
each resin, trying to optimise the activity (U per g support).
For this purpose, we decided to study the highly cross-linked
(macroporous) polymer SILLP 6, which has a higher mechanical
strength than gel-type resins, and accordingly might be more
suitable for flow processes.
Table 2 summarises the efficiency of the immobilization
process onto SILLP 6 as a function of the supplied amount of
protein (from 5.5 to 63 mg of protein per g of support). As can be
seen, by increasing the supplied protein, the enzyme loading in
the resin increased, which led to a corresponding enhancement
of the biocatalytic activity (from 24 to 271 U per g support)
for the KR of rac-12. The same excellent stereoselectivity
was achieved for all the cases (>99.9% ee). It is noteworthy
that the specific activity remained practically constant. Only a
slightly lower enzymatic efficiency was observed for the highest
Both FT-ATR-IR spectra of CALB-SILL-10 and Novozyme
435 showed a single peak at 1658 cm-1. In addition, bands
at ca. 1670 cm-1 (associated with b-turns not associated with
hydrogen bonds), and at 1690 cm-1 (which could be related
to the formation of aggregates), clearly indicate the structural
changes of the lipase. On the other hand, the spectrum of
CALB–SILLP 9 showed a band at 1649 cm-1, which can be
associated with the a-helix structure, while peaks at 1660 and
1667 cm-1 are characteristic of b-turns.24 This indicates that in
Table 2 Optimisation of the immobilisation of CALB onto SILLP 6 (CALB–SILLP 11)a
Entry
Protein suppliedb
Immobilised proteinc
Activity (U per g support)c
Specific activity (U per mg protein)
1
2
3
4
5
5.5
0.9
1.4
5.9
10.0
14.0
24
35
122
220
271
25.1
24.8
20.7
22.0
19.3
10.8
25.4
38.0
63.4
a The KR was carried out in n-hexane at 40 ◦C using a rac-1-phenylethanol–vinyl propionate ratio of 1 : 2. b 100 mg of the support was suspended in
200 mL of an aqueous solution with variable concentrations of CALB supplied (mg per g support). c mg CALB per g support calculated by Lowry’s
method: % immobilisation = (mg immobilised protein per g support)/(mg supplied protein per g support) ¥ 100. d U = mmol product per mL per min.
1806 | Green Chem., 2010, 12, 1803–1810
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
protein loading. Furthermore, CALB–SILLP 11 with a loading
of ca. 10 mg of protein per g support could be reused for
three consecutive batch cycles without any significant loss of
activity or stereoselectivity. When comparing the data gathered
in Tables 1 and 2, it can be seen that the supported biocatalyst
CALB–SILLP 11 prepared by the optimised protocol provides
comparable activities and similar specific activities to the best
gel-type supported enzyme.
In order to develop a more sustainable and efficient system,
we decided to test the ability of CALB–SILLPs to catalyse the
KR of rac-12 on a continuous flow system. The solid nature
of the CALB–SILLP 11 (10 mg of protein per g support)
allows one to easily design a fixed-bed reactor to perform
KR under flow conditions.26 In addition, this set-up facilitates
the substitution of a traditional, relatively toxic, solvent (n-
hexane) by a non-toxic one such as scCO2. Fig. 1 depicts the
profiles of productivity (mmol per g support per min) of (R)-1-
phenylethyl propionate vs. time reached for a continuous KR
of rac-12 in scCO2 at 50 ◦C and 10 MPa. The productivity
can be easily improved by adjusting the mass flow rate of the
organic stream from 3 to 15 mmol min-1, as a consequence
of the high enzyme loading of the support. The yield of ester
product was ca. 50% after reaching the steady state, for all the
conditions assayed. The synthetic product 15 was obtained with
an excellent enantiopurity (>99.9% ee), while (S)-ester product
15 was never detected. The system maintained the efficiency even
for an extended period of time (up to 6 days of continuous use).
Furthermore, the use of CALB–SILLPs reduces significantly
the amount of ionic liquid used as compared with either
bulk homogeneous ionic liquid27 or enzymes coated with IL
layers.10 Thus, for instance, for coated enzyme supported on
different materials, a 1 : 1 ILs–support ratio is required to avoid
deactivation.11d The approach here reported does not require an
additional IL layer to stabilise the enzyme against deactivation,
being stable for a long period (Fig. 1). This leads to a significant
minimisation of cost and hazards related to the use of bulk ILs.
Furthermore, coated supported enzymes with IL layers may
leach by either physico-mechanical removal by convention or by
partial dissolution of the ILs in the liquid reaction phase. These
phenomena are not possible in CALB–SILLPs, as the IL-like
phases are covalently linked to the support.
A maximum theoretical yield of 50% can be obtained for
the KR procedure. However, DKR processes, in which in situ
racemisation of the unwanted enantiomer is coupled with KR,
is an attractive method to overcome this limitation.28 Although
some improvements have been made in this direction, the DKR
of aliphatic alcohols using a catalyst under environmentally
friendly and mild conditions remains a challenge.29 The key
issue is related to the use of simple experimental conditions
compatible for both the biocatalyst and the chemocatalyst.
The use of either an immobilized transition metal catalyst
or a heterogeneous acid catalyst for the alcohol racemization
simplified the processes significantly and minimized the possible
(mutual) deactivation during the DKR process.30 Acid catalysts
(Zeolites, Amberlite, Nafion, etc.) are by far a simpler approach,
being cheap and readily available catalysts that have been used
for the DKR of phenyl ethanol. The main problem in combining
the acid-catalyzed racemisation and the enzymatic resolution of
12 by esterification is that the reactions have to occur under
aqueous and water-free conditions respectively. Hence, to get
good enantioselectivity in DKR systems using acid catalysts
(mainly zeolites), the zeolites and the biocatalysts need to be
physically isolated.31
Taking into the account these considerations, we assayed
the DKR using the configuration shown in Fig. 2. Two fixed-
bed reactors loaded with the immobilized enzyme (biocatalytic
reactor, CALB–SILLP 11, 150 mg) and an additional one with
an acid zeolite (chemical racemisation catalyst, 100 mg) were
assembled with the configuration biocatalyst–chemocatalyst–
biocatalyst. It was decided to use separate columns for each
catalyst to isolate them, avoiding any possible enzyme deacti-
vation, which was observed in previous experiments where the
enzymatic catalyst and the acid catalyst were mixed together.32
In addition, this particular reactor configuration easily allows
us to replace the acid catalyst reactor, facilitating the screening
of different zeolites (CP811C, CBV720 and CP811E), which
were initially selected as chemocatalysts as they have proved to
be good catalysts for the racemisation of enantiopure phenyl
ethanol.33
It should be also noted that, for this particular approach,
the (R)-15 yield may only reach 100% by coupling in series
several enzymatic and acidic catalytic columns, following a
dichotomic progression. Thus, we must bear in mind that only
a maximum yield of 75% can be reached with our actual
experimental configuration. Fig. 3 depicts the results obtained
for the continuous DKR at 50 ◦C and 10 MPa.
Fig. 1 Profiles of productivity and ee of (R)-1-phenylethyl propionate
vs. time on stream for the continuous KR of rac-12 in scCO2 at 50 ◦C
and 10 MPa: a) 3 mmol min-1; b) 15 mmol min-1.
We have developed a stable, efficient and stereospecific
continuous KR of rac-12 immobilising the CALB onto SILLPs
and using scCO2 as vector to deliver the substrates to the active
sites and to extract the products. The nature of the support may
affect the partitioning of water and substrate molecules with the
bulk medium, and so, simultaneously, to favour mass-transfer
phenomena and to preserve the hydration shell of the enzyme.
Thus, for instance, the continuous flow of scCO2 cannot extract
the free water molecules from the CALB–SILLP, reducing in this
way the potential hydrolytic reaction. Indeed, very low levels of
propionic acid (<1%) were observed in all the cases.
The modified polymer with IL-like groups provides the
required microenvironment for the stabilisation of the en-
zyme leading to active, stable and enantioselective biocatalysts.
This journal is
The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 1803–1810 | 1807
©

View Article Online
Once we had proven that the DKR based in the combination
of CALB–SILLPs and zeolite can be achieved with excellent
enantioselectivities, we tried to develop our system further using
a single reactor, in which both bio- and chemocatalyst are mixed
together to perform a “one-pot” DKR in scCO2 (see Fig. 4).
Fig. 2 Experimental set-up for the continuous DKR of rac-12 using
a combination of consecutive biocatalytic–chemocatalytic–biocatalytic
reactors.
Fig. 4 Experimental set-up for the continuous “one-pot” chemo-
enzymatic DKR of rac-12.
The “one-pot” DKR using a combination of CALB–SILLP
11 and zeolite (CP811E-150) in the same reactor yielded an
unselective process (63% ee), both enantiomers of the ester 15
being observed (see Entry 1, Table 3). The low selectivity can
be attributed to direct esterification of rac-12 catalysed by the
zeolite. It has been proved that the IL-coating of H-b zeolites
with a hydrophobic IL significantly reduces its racemisation
activity and the side reaction (acylation).34 This is due to the
fact that the IL monolayer increases the mass-transfer of the
reactants to the active sites and also isolates the acid sites
from the eventual presence of water in the system. Therefore,
several combinations of CP811E-150 coated with two different
ILs ([Btma][NTf2] and [BMIM][PF6]) and CALB–SILLP 11 (see
Table 3) were tested in the DKR. In both cases, good initial yields
of ca. 70% were observed. Additionally, the process took place
with a significant increase of the enantioselectivity. Indeed, only
the R-ester 15 was detected (>99.9% ee). Finally, the flow rate
of the system was reduced in order to increase the yield of the
Fig. 3 Profiles of (R)-1-phenylethyl propionate yield (᭹) and enan-
tiomeric excess of ester products (ꢀ) during the continuous DKR of rac-
1-phenyl ethanol catalyzed by CALB–SILLP 11 and zeolites CP811C
(A), CBV720 (B) and CP811E (C) as acid catalysts in scCO2 at 50 ◦C
and 10 MPa.
The DKR was successful in all cases. The total yield of the
R-ester product was ca. 63% for all the time-course profiles,
which corresponded with ca. 84% of the possible R-ester
product. The biocatalytic reactors loaded with the same sample
of CALB–SILLP were used for all acid catalysts screening
experiments. In addition, the CALB–SILLP maintained its
activity in continuous operation for 19 days. This excellent
operational stability was in agreement with previous results
of enzymes in ionic liquids environments and with enzymes
immobilized onto SILLPs.16
The use of different zeolites led to changes in the enantioselec-
tivity of the process, mainly due to the differing ability of zeolites
to catalyze the non-stereoselective transesterification reaction
that produces the S-ester. In this way, the use of the CP811C
zeolite reduced the ee of the R-ester product to 92%, while
the best results were obtained for both CBV720 and CP811E
zeolites, that reached the maximum ee (>99.9%). Furthermore,
the presence of propionic acid was observed in the scCO2 flow
when an acid catalyst column was included into the reactor
configuration. In spite of the dryness of the scCO2, the residual
water content in the porous structure of the chemocatalyst could
be involved in this undesired activity. This shows the necessity
for strict control of the water content of these materials before
starting the DKR processes.
a
Table 3 Continuous “one-pot” DKR of rac-12 in scCO2
Composition (outlet) (%)
Entry Set-upa
(R)-12 (S)-12 (R)-15 (S)-15 ee (%) of 15
1
2
3
2
I
7
8
6
2
11
18
26
6
48
74
68
92
11
0
0
63
II
III
IIb
99.9
99.9
99.9
0
a I: 600 mg of CALB–SILLP 11 and 400 mg of CP811E-150. II: 600 mg
of CALB–SILLP 11 and 200 mg of CP811E-150 + 200 ml [BtmA][NTf2].
III: 600 mg of CALB–SILLP 11 and 200 mg of CP811E-150 + 200 ml
[BMIM][PF6]. 12 = 1-phenylethanol; 15 = phenylethyl propionate;
[BMIM] = butylmethylimidazolium; [BtmA] = butyltrimethylamonium.
b Half the mass flow rate (10.6 mmol min-1) was used.
1808 | Green Chem., 2010, 12, 1803–1810
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
process. Thus, a yield of the R-ester of up to 92% could be
reached with an excellent enantioselectivity >99.9%
Acknowledgements
This work was partially supported by CICYT (CTQ2008-00877
and CTQ2008-04412), SENECA Foundation (08616/PI/08),
Bancaja-UJI (P1-1B2009-58) and CSIC (200980I092) grants.
Conclusions
The results here presented clearly highlight the great potential
of supported IL-like phases (SILLPs) as “solid solvents” for
the direct immobilization and stabilization of enzymes and for
the development of new biocatalytic systems. The efficiency of
the resulting biocatalysts, in terms of both activity and specific
activity, for the KR of rac-1-phenylethanol using CALB is
affected by several factors, as are the nature and morphology
of the polymeric material, the loading of IL-like moieties, the
loading of the enzyme and the chemical structure of the IL-
like fragments, in particular the nature of the anion (which
is essential to define the hydrophilic/hydrophobic balance of
the corresponding material). All those parameters can be easily
adjusted and optimised. Best results are obtained for hydrophilic
IL-like fragments containing chloride as the anion. In the same
way, high loadings of both IL-like moieties and enzymatic units
are preferred, although keeping a relatively high IL-like/enzyme
ratio favour maintaining the active native secondary structure
of the enzyme. This is revealed by the long-term stability of the
resulting supported biocatalyst and in the appreciable increase
in the protein denaturation temperature.
References
1 Ionic Liquids in Synthesis, ed. P. Wasserscheid and T. Welton, Wiley-
VCH, 2008.
2 (a) T. L. Greaves and C. J. Drummond, Chem. Rev., 2008, 108, 206;
(b) H. Olivier-Bourbigou, L. Magna and D. Morvan, Appl. Catal.,
A, 2010, 373, 1.
3 N. V. Plechkova and K. R. Seddon, Chem. Soc. Rev., 2008, 37, 123.
4 (a) F. van Rantwijk and R. A. Sheldon, Chem. Rev., 2007, 107, 2757;
(b) P. D. de Maria, Angew. Chem., Int. Ed., 2008, 47, 6960; (c) M.
Moniruzzaman, K. Nakashima, N. Kamiya and M. Goto, Biochem.
Eng. J., 2010, 48, 295.
5 (a) A. M. Klibanov, Nature, 2001, 409, 241; (b) V. Gotor, I. Alfonso,
E. Garc´ıa-Urdiales Asymmetric Organic Synthesis with Enzymes,
Wiley-VCH Verlag, 2008; (c) N. R. Patel, Coord. Chem. Rev., 2008,
252, 659.
6 (a) H. R. Hobbs and N. R. Thomas, Chem. Rev., 2007, 107, 2786;
(b) M. Habulin, M. Primozic and Z. Knez, Acta Chim. Slov., 2007,
54, 667.
7 A. J. Mesiano, E. J. Beckman and A. J. Russel, Chem. Rev., 1999, 99,
623; A. T. Fricks, E. G. Oestreicher, L. C. Filho, A. C. Feihrmann, Y.
Cordeiro, C. Dariva and O. A. C. Antunes, J. Supercrit. Fluids, 2009,
50, 162.
8 M. Moniruzzaman, N. Kamiya and M. Goto, Org. Biomol. Chem.,
2010, 8, 2887.
9 (a) T. Itoh, H-S. Han, Y. Matsushita and S. Hayase, Green Chem.,
2004, 6, 437; (b) T. Itoh, Y. Matsushita, Y. Abe, S-H. Han, S. Wada,
S. Hayase, M. Kawatsura, S. Takai, M. Morimoto and Y. Hirose,
Chem. Eur. J., 2006, 12, 9228; (c) Y. Abe, T. Hirakawa, S. Nakajima,
N. Okano, S. Hayase, M. Kawatsura, Y. Hirose and T. Itoh, Adv.
Synth. Catal., 2008, 350, 1954.
The use of a macroporous resin facilitates the preparation
of biocatalytic minireactors for continuous-flow applications
in scCO2 that can be used for more than one week without
any appreciable reduction in their efficiency. Excellent yields
(50%, the maximum attainable for a KR) and enantioselectivities
10 P. Lozano, Green Chem., 2010, 12, 555.
◦
(>99.9% ee) are continuously obtained at 50 C and 10 MPa.
11 (a) P. Lozano, T. de Diego, D. Carrie´, M. Vaultier and J. L. Iborra,
Chem. Commun., 2002, 692; (b) M. T. Reetz, W. Wiesenhofer, G.
Francio and W. Leitner, Chem. Commun., 2002, 992; (c) M. T. Reetz,
W. Wiesenhofer, G. Francio and W. Leitner, Adv. Synth. Catal., 2003,
345, 1221; (d) P. Lozano, T. de Diego, S. Gmouh, M. Vaultier and
J. L. Iborra, Biotechnol. Prog., 2004, 20, 661.
12 (a) N. Karbass, V. Sans, E. Garcia-Verdugo, M. I. Burguete and S. V.
Luis, Chem. Commun., 2006, 3095; (b) M. I. Burguete, H. Erythropel,
E. Garcia-Verdugo, S. V. Luis and V. Sans, Green Chem., 2008, 10,
401; (c) M. I. Burguete, E. Garc´ıa-Verdugo, I. Garcia-Villar, F. Gelat,
P. Licence, S. V. Luis and V. Sans, J. Catal., 2010, 269, 150.
13 M. I. Burguete, F. Galindo, E. Garcia-Verdugo, N. Karbass and S. V.
Luis, Chem. Commun., 2007, 3086.
14 S. Morrissey, B. Pegot, D. Coleman, M. T. Garcia, D. Ferguson, B.
Quilty and N. Gathergood, Green Chem., 2009, 11, 475; J. R. Harjani,
R. D. Singer, M. T. Garcia and P. J. Scammells, Green Chem., 2009,
11, 83.
15 (a) M. H. Valkenberg, C. deCastro and W. H Holderich, Green Chem.,
2002, 4, 88; (b) C. P. Mehnert, Chem. Eur. J., 2004, 11, 50; (c) A.
Riisager, R. Fehrmann, M. Haumann and P. Wasserscheid, Top.
Catal., 2006, 40, 91; (d) Y. Gua and G. Li, Adv. Synth. Catal., 2009,
351, 817.
16 P. Lozano, E. Garc´ıa-Verdugo, R. Piamtongkam, N. Karbass, T.
DeDiego, M. I. Burguete, S. V. Luis and J. L. Iborra, Adv. Synth.
Catal., 2007, 349, 1077.
17 E. F. Hartree, Anal. Biochem., 1972, 48, 422.
18 C. S. Chen, Y. Fujimoto, G. Girdaukas and C. J. Sih, J. Am. Chem.
Soc., 1982, 104, 17294.
19 (a) T. De Diego, P. Lozano, S. Gmouh, M. Vaultier and J. L. Iborra,
Biomacromolecules, 2005, 6, 1457; (b) P. Lozano, T. De Diego, S.
Gmouh, M. Vaultier and J. L. Iborra, Biocatal. Biotransform., 2005,
23, 169.
20 The polarity of the SILLPs can be determined using pyrene as probe;
see ref. 11. The estimated dielectric constant value (e) for different
SILLPs 5a and 5b are 21 and 15 respectively.
The combination of this biocatalytic system with an acid catalyst
has allowed us to develop different DKR processes. The as-
sembly of three columnar reactors (biocatalyst–chemocatalyst–
biocatalyst) produces a system that can work efficiently for more
than 3 weeks without a reduction in performance, while raising
the yields to more than 60% and allowing the identification of
the most efficient acid catalyst able to maintain an excellent
enantioselectivity. A “one-pot” single columnar minireactor
can be assembled by loading the corresponding column with
a mixture of CALB–SILLP 11 and CP11E-150. To achieve a
selective process, the zeolite needs to be coated with a small
amount of an IL. Optimisation of this approach has allowed
us to increase the yields of the desired product up to 92% with
enantioselectivities higher than 99% ee.
The system here reported represents an important improve-
ment compared with previously reported DKRs, either in
conventional or neoteric solvents (see the ESI† for a comparison
of different supported catalysts used for DKR). DKR, when
performed in a conventional solvent using a homogeneous
organometallic catalyst for racemisation, usually requires anhy-
drous conditions, inert atmospheres, the addition of a base, and
batch operation for extended periods of time. Those limitations
are overcome in the present system, which is an evolution of
previous systems that, in a simple way, allows us to obtain a high
productivity in short periods of time and with the concomitant
savings not only in terms of the economical viability, but also
environmental factors.10
This journal is
The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 1803–1810 | 1809
©

View Article Online
21 N. Miletic´, Z. Vukovic´, A. Nastasovic´ and K. Loos, J. Mol. Catal.
B: Enzym., 2009, 56, 196.
22 This supported catalyst was selected for the comparison, as both
polymeric materials are macroporous; other CALB–SILLPs pre-
pared in this work based on gel-type polymers may have additional
factors contributing to their catalytic activity.
27 Around 1 mL of IL is generally used; see for instance: P. Hara, U.
Hanefeld and L. T. Kanerva, Green Chem., 2009, 11, 250.
28 H. Pellissier, Tetrahedron, 2008, 64, 1563.
29 N. J. Turner, Curr. Opin. Chem. Biol., 2010, 14, 115.
30 A. Parvulescu, J. Janssens, J. Vanderleyden and D. De Vos, Top.
Catal., 2010, 53, 93.
31 (a) S. Wuyts, K. De Temmerman, D. De Vos and P. Jacobs, Chem.
Commun., 2003, 1928; (b) S. Wuyts, K. De Temmerman, D. E. De
Vos and P. A. Jacobs, Chem. Eur. J., 2004, 11, 386.
32 P. Lozano, T. de Diego, M. Larnicol, M. Vaultier and J. L. Iborra,
Biotechnol. Lett., 2006, 28, 1559.
33 E. L. Teo, G. K. Chuah, A. R. J. Huguet, S. Jaenicke, G. Pande and
Y. Zhu, Catal. Today, 2004, 97, 263.
23 H. Zhao, C. L. Jones and J. V. Cowins, Green Chem., 2009, 11,
1128.
24 R. M Lau, M. J. Sorgedrager, G. Carrea, F. Van Rantwijk, F Secundo
and R. A. Sheldon, Green Chem., 2004, 6, 483.
25 K. J. Halverson, I. Sucholeiki, T. T. Ashburn and P. T. Lansbury Jr.,
J. Am. Chem. Soc., 1991, 113, 6701.
26 S. V. Luis and E. Garcia-Verdugo, Chemical Reactions and
Processes under Flow Conditions, RSC Green Chemistry Series,
2009.
34 P. Lozano, T. De Diego, C. Mira, K. Montague, M. Vaultier and J. L.
Iborra, Green Chem., 2009, 11, 538.
1810 | Green Chem., 2010, 12, 1803–1810
This journal is
The Royal Society of Chemistry 2010
©