From batch to flow processing: Racemization of N-acetylamino acids under microwave heating

Article abstract of DOI:10.1021/op9001356

The racemization of N-acetylindoline-2-carboxylic acid in P-xylene revealed beneficial rate enhancements due to microwave effects, by comparing conventional and microwave heating. The magnitude of this effect was governed by the degree of heterogeneity of the reaction system. The amount of catalyst, the temperature and the amount of cosolvent played a decisive role. The microwave effect completely vanished when a homogeneous solution was heated. During the microwave-assisted heterogeneous racemization of N-acetylphenylalanine in P-xylene a comparable microwave effect was observed. The microwave effects could be rationalized by adapting selective heating in the phase boundary region of solid and liquid. Additionally, a straightforward translation was achieved from batchwise operation in a stirred reactor to a batch-loop reactor. The (heterogeneous) racemization of N-acetylindoline-2-carboxylic acid retained its microwave effect in the loop reactor. Our results demonstrated that avoiding plugging of the tubular reactor is a severe challenge in scaling up.

Full text of DOI:10.1021/op9001356

Organic Process Research & Development 2009, 13, 888–895
From Batch to Flow Processing: Racemization of N-Acetylamino Acids under
Microwave Heating
†
†
‡
†
,†
Mark H. C. L. Dressen, Bastiaan H. P. van de Kruijs, Jan Meuldijk, Jef A. J. M. Vekemans, and Lumbertus A. Hulshof*
EindhoVen UniVersity of Technology, Laboratory of Macromolecular and Organic Chemistry, Applied Organic Chemistry,
Den Dolech 2, 5612 AZ EindhoVen, The Netherlands
Abstract:
operation. A continuously operated tubular reactor seems to be
promising for application of microwave technology.
The racemization of N-acetylindoline-2-carboxylic acid in p-xylene
revealed beneficial rate enhancements due to microwave effects,
by comparing conventional and microwave heating. The magni-
tude of this effect was governed by the degree of heterogeneity of
the reaction system. The amount of catalyst, the temperature and
the amount of cosolvent played a decisive role. The microwave
effect completely vanished when a homogeneous solution was
heated. During the microwave-assisted heterogeneous racemization
of N-acetylphenylalanine in p-xylene a comparable microwave
effect was observed. The microwave effects could be rationalized
by adapting selective heating in the phase boundary region of solid
and liquid. Additionally, a straightforward translation was achieved
from batchwise operation in a stirred reactor to a batch-loop
reactor. The (heterogeneous) racemization of N-acetylindoline-2-
carboxylic acid retained its microwave effect in the loop reactor.
Our results demonstrated that avoiding plugging of the tubular
reactor is a severe challenge in scaling up.
In our study, we have selected the racemization of N-acetyl-
indoline-2-carboxylic acid to investigate the benefit of micro-
wave heating compared to conventional heating in fine chemical
4
production. In previous studies, a decreasing heterogeneity of
the reaction mixture has demonstrated a diminishing difference
between the reaction rates observed for microwave and con-
ventional heating. This effect is referred to as a vanishing
microwave effect. In addition to the racemization process in a
stirred batch reactor, N-acetylindoline-2-carboxylic acid has been
racemized in a microwave-heated batch-loop reactor to address
the scaling-up potential of microwave heating.
Racemization of N-Acetylindoline-2-carboxylic Acid in a
Stirred Batch Reactor
In the case of a classical resolution of unnatural N-acetyl-
indoline-2-carboxylic acid, the waste stream is enriched in the
5
undesired enantiomer (Scheme 1). After a separate racemization
step, the product/substrate is reintroduced to the process. On
the one hand a fast racemization step is preferred, and on the
other hand, an easy workup. A cosolvent or a good solvent
increases the solubility of the substrate, thus beneficially
contributing to the reaction rate of the racemization. However,
removal of the solvent (e.g., by distillation) is inefficient from
both time and energy perspectives.
Introduction
Chiral compounds are ubiquitous in our lives. They are the
active constituents of many pharmaceuticals, nutrition products,
flavors and fragrances. The essential components are constructed
from chiral building blocks and their synthesis becomes more
complex in the presence of more stereocenters in their backbone.
One route towards enantiomerically pure products involves
For this process, p-xylene is selected in view of its high
boiling point, inertness, and poorly solubilizing power at room
temperature. Indirectly the high boiling point partially com-
pensates for the poor solubility. The product, in the case where
p-xylene is the solvent, is isolated by an energy-efficient
filtration. The actual racemization is catalyzed by acetic
1
classical resolution (see Scheme 1). Racemization as part of
either a dynamic or a classical resolution process can make the
difference between 50% and complete conversion of the
substrate.
The relevance of microwave heating in organic synthesis is
intensively studied in a small-scale research environment. There
6
2
anhydride (see Scheme 2).
Acetic anhydride reacts with the acid functionality, thus
forming a mixed anhydride. Ring closure eliminates acetate and
is increasing interest in the scaling-up potential of microwave
2,3
heating. Because operation in a stirred tank is limited by
the penetration depth of microwaves, the best opportunities for
the fine chemical industry are considered to lie in a continuous
(
4) Dressen, M. H. C. L.; Van de Kruijs, B. H. P.; Meuldijk, J.; Vekemans,
J. A. J. M.; Hulshof, L. A. Org. Process Res. DeV. 2007, 11, 865.
(
5) Kuilman, T.; Hendrickx, A. J. J. Process for the preparation of an
optically active indoline-2-carboxylic acid or derivative thereof. (DSM
NV). Patent number NL10083202, 1998.
*
To whom correspondence should be addressed. E-mail: L.A.Hulshof@tue.nl.
Laboratory of Macromolecular and Organic Chemistry, Applied Organic
†
Chemistry.
(6) (a) Akira, M.; Hiromitsu, F. Racemization of N-acetylindoline-2-
carboxylic acid. (Toray Industries). Patent number JP61083159, 1986.
(b) Chen, F. M. F.; Benoiton, N. L. Int. J. Peptide Protein Res. 1988,
31, 396.
‡
Process Development Group.
(
1) (a) Sheldon, R. A. Chirotechnology: Industrial Synthesis of Optically
ActiVe Compounds; Dekker: New York, 1993. (b) Souvie, J.-C. Novel
method for the synthesis of (S)-indoline-2-carboxylic acid and ap-
plication thereof in the synthesis of perindopril. (Servier Lab). Patent
number EP 1348684, 2003.
(7) (a) Bergmann, M.; Zervas, L. Eur. J. Biochem. 1928, 03, 280. (b)
Knorr, R.; Huisgen, R. Chem. Ber. 1970, 103, 2598. (c) Chenault,
H. K.; Dahmer, J.; Whitesides, G. M. J. Am. Chem. Soc. 1989, 111,
6354. (d) Parvulescu, A.; Van der Eycken, E.; Jacobs, P. A.; De Vos,
D. E. J. Catal. 2008, 255, 20. (e) Hateley, M. J.; Schichl, D. A.;
Fischer, C.; Beller, M. Synlett 2001, 1, 25.
(
(
2) Kappe, C. O. Angew. Chem., Int. Ed. 2004, 43, 6250.
3) N u¨ chter, M.; Ondruschka, B.; Bonrath, W.; Gum, A. Green Chem.
2
004, 6, 128.
8
88
•
Vol. 13, No. 5, 2009 / Organic Process Research & Development
10.1021/op9001356 CCC: $40.75  2009 American Chemical Society
Published on Web 08/27/2009

Scheme 1. Overview of a classical resolution of N-acetylindoline-2-carboxylic acid with a non-racemic aminea
a
The desired diastereomer as precipitate is collected after filtration and desalted to obtain an important intermediate (S-enantiomer) in the synthesis of perindopril.1
The undesired diastereomer as a majority in the filtrate is desalted likewise and is racemized before recycling to the resolution procedure.
Scheme 2. Racemization of (R)-N-acetylindoline-2-carboxylic
acid
Scheme 3. Proposed mechanism of the racemization of
(R)-N-acetylindoline-2-carboxylic acid: [1] formation of
asymmetric anhydride, [2] ring closure, [3] keto-enol
a
tautomerization and [4] proton-shift
Figure 1. Ternary diagram of (R)- and (S)-N-acetylindoline-
-carboxylic acid and p-xylene at 130 °C.
2
approximately 20 wt % of acetic acid, the reaction mixture
becomes homogeneous, and this leads to a complete vanishing
of the microwave effect. Figure 1 shows a ternary phase diagram
of (R)- and (S)-N-acetylindoline-2-carboxylic acid and p-xylene
at 130 °C.
The previously mentioned microwave effect is defined as a
significant difference between the reaction rates of conventional
and microwave heating that cannot be reproduced in a con-
ventional way. The ratio of initial reaction rates (rini) of both
heating techniques enables us to quantify the microwave effect
a
After tautomerization the oxazolium is achiral; the reverse reaction can form
either enantiomer (R or S). Only one enantiomer is depicted to make this scheme
comprehensible.
(fMW), see eq 1.
gives rise to an oxazolium intermediate. At this stage, the chiral
information is lost by keto-enol tautomerization, and due to
the reversible character of this reaction, the substrate racemizes
completely; see Scheme 3. Other options, such as Pd-catalyzed
rini;MW
fMW ) rini;CH
(1)
The Influence of the Amount of Acetic Anhydride. Initial
5
racemization of amino-acids and more general racemizations,
experiments were performed using a scalemic mixture with an
enantiomeric excess (ee) of 67% (R). Additional experimenta-
tion was performed with the (S)-enantiomer (99% ee). This
study surprisingly revealed a complex pattern that required more
detailed investigation. Upon addition of the catalyst, the ee
8
are reviewed by Zwanenburg et al.
The Influence of Cosolvent. The racemization of N-acetyl-
indoline-2-carboxylic acid using acetic acid as a solvent has
9
been reported. In our case, acetic acid was added to p-xylene
as a cosolvent to study the role of heterogeneity in correlation
4
with an observed microwave effect for the racemization. With
(
9) Weerantunga, G.; Wang, Z.-X.; Raheem, M. A. New processes for
the preparation of optically pure indoline-2-carboxylic acid and
N-acetyl-indoline-2-carboxylic acid. (Apotex Pharmachem, Inc.).
Patent number WO2006053440, 2006.
(
8) Ebbers, E. J.; Ariaans, G. J. A.; Houbiers, J. P. M.; Bruggink, A.;
Zwanenburg, B. Tetrahedron 1997, 53, 9417.
Vol. 13, No. 5, 2009 / Organic Process Research & Development
•
889

Figure 2. Racemization of (S)-N-acetylindoline-2-carboxylic acid at 130 °C under conventional heating (a) with various amounts of
catalytic acetic anhydride and (b) at different temperatures.
Scheme 4. Inhibition of racemization: [1] formation of
oxazolium intermediate with acetic anhydride, [2] keto-enol
tautomerization, [3] formation of oxazolium acetate with
acetic anhydride, [4] precipitation of intermediate, [5]
resolving the intermediate by e.g. microwave heating and [6]
addition of water regenerates the substrate, [7] acetate
a
bonds to the intermediate (covalent or ionic)
Figure 3. Racemization of (S)-N-acetylindoline-2-carboxylic
acid at 130 °C under conventional heating with 0.1 equiv of
catalyst (green line); duplicate (red line) and an extra-dried
batch (black line). After 3 h the reaction vessel was transferred
to continue the reaction in the microwave oven (blue line).
initially decreases as expected, but after a certain period of time
the ee levels off under conventional heating. The progress of
the racemization depends both on temperature and on the
amount of acetic anhydride.
a
After tautomerization the oxazolium is achiral; the reverse reaction can form
either enantiomer (R or S). Only one enantiomer is depicted to make this scheme
comprehensible.
Water may influence the course of the reaction. Related to
acetic anhydride, an equimolar amount of water consumes the
catalyst and interrupts the racemization process. An amount
smaller than one equivalent of water doubles the amount of
acetic acid, which acts as a cosolvent that improves the
racemization.
intermediate presumably originates from a (potentially racem-
ized) substrate and two acetic anhydrides (Figure 2a). One
anhydride is used to form the oxazolium intermediate in the
first part, and another anhydride may be used to form an
eventually inactive structure that inhibits the overall process of
racemization (Scheme 4).
Figure 2a shows that the racemization seems to be stopped
when using smaller amounts of acetic anhydride. Duplicates
illustrate the same behavior. However, to our surprise, when
the same reaction mixture was transferred from the oil bath to
the microwave oven, the racemization started up again (see
Figure 3).
This observation suggests that the molecular species that
stops the racemization under conventional heating is actually
an intermediate in the process. This intermediate, which has
temporarily consumed all of the acetic anhydride, is (partly)
dissolved or reactivated by microwave heating. Given the re-
lationship between the enantiomeric excess (e.g., 50%) at the
plateau and the amount of acetic anhydride (e.g., 1 equiv) the
Only an increase in temperature or a larger amount of
cosolvent can shift the equilibrium to allow completion of the
racemization. The solubility does not play a role at lower ee
due to the higher solubility of the racemate (see Figure 1).
Shortly after addition of acetic anhydride the solubility of
N-acetylindoline-2-carboxylic acid temporarily increases. There-
after, two processes of precipitation and simultaneous racem-
ization allow the ee to slightly increase with respect to the
plateau level. The (S)-enantiomer selectively precipitates in favor
of the (R)-enantiomer, due to the presence of a majority of (S)-
enantiomer crystals. After workup with water the insoluble
structure is dissociated into (R)- or (S)-N-acetylindoline-2-
carboxylic acid. By applying solid-state NMR the intermediate
8
90
•
Vol. 13, No. 5, 2009 / Organic Process Research & Development

Figure 4. (a) Racemization of (R)-N-acetylindoline-2-carboxylic acid (66% ee) with addition of cosolvent (wt% acetic acid) at 130
C for conventional (CH) and microwave heating (MW) conditions and (b) correlation of cosolvent addition with the microwave
°
effect.
has been compared to the enantiomerically pure substrate and
the racemic product. The spectrum shows another set of peaks,
but no conclusive evidence can be produced from this analysis.
The increase of ee in time could, however, not be explained
by the method of sampling, which was checked by comparing
representative samples with the direct workup of the whole
batch. An increase of % ee in a dynamic system during a longer
period of days has also been demonstrated by Blackmond et
Scheme 5. Racemization of N-acetylphenylalanine
increases both solubility and reaction rates. Bulk temperature,
however, remains relatively low (and is identical to the
temperature in the experiment where the reaction mixture was
conventionally heated).
10
al. The increase in ee is explained by a slight enantioimbalance
which directs the trend to a single chiral solid state.
The Influence of Temperature. The racemization of
N-acetylindoline-2-carboxylic acid was also conducted below
The Influence of Substrate. To understand the peculiar
mechanistic phenomena observed in the racemization of N-acetyl-
indoline-2-carboxylic acid and to gain more insight into the
related microwave effect, another chiral N-acetylamino acid was
synthesized and racemized. One alternative substrate, in which
formally the NH-Ph band was cleaved, was selected, e.g.,
N-acetylphenylalanine (see Scheme 5).
130 °C to study the influence of temperatures. This resulted
for this case in an incomplete racemization which is shown in
Figure 1b. The influence of temperature on the racemization is
elaborated in more detail in the section of “The Influence of
Substrate”.
The Influence of Microwave Heating. An identical ex-
perimental setup and identical conditions were applied for both
heating methods with proper stirring and internal temperature
control by fiber optics, which enables a direct comparison
between all experiments. All the heterogeneous reactions were
Under the benchmark conditions (130 °C), the racemization
of (S)-N-acetylphenylalanine was completed within 5 min. For
practical reasons, the temperature was lowered to 85 °C (Figure
5). As with N-acetylindoline-2-carboxylic acid, microwave
heating increases also the reaction rates of the heterogeneous
racemization of N-acetylphenylalanine. Under the studied
reaction conditions, the microwave effect is relatively high at
4
accelerated by microwave heating (Figure 4). Additional
heating experiments with microwave irradiation demonstrated
that neither the solid, N-acetylindoline-2-carboxylic acid, nor
the liquid increased in temperature, i.e. selective heating of both
solid and liquid was not observed. Alternatively, experiments
with homogeneous mixtures did not reveal any reaction rate
enhancement by switching from conventional to microwave
heating. In combination with the low energy of a microwave
-5
photon (∼10 eV) a mechanistic change of racemization by
microwaves is highly unlikely.
The reaction takes place in the phase boundary region of
4
solid and liquid. The reactants and the intermediates in this
phase boundary region selectively interact with microwaves due
to their relatively polar character. Note that, the concentrations
of reactants and intermediates in the phase boundary region
differ from bulk. Local heating occurs, which simultaneously
(
10) Noorduin, W. L.; Izumi, T.; Milemaggi, A.; Leeman, M.; Meekes,
H.; Van Enckvort, J. P.; Kellogg, R. M.; Kaptein, B.; Vlieg, E.;
Blackmond, D. G. J. Am. Chem. Soc. 2008, 130, 1158.
Figure 5. Racemization of (S)-N-acetylphenylalanine at various
temperatures.
Vol. 13, No. 5, 2009 / Organic Process Research & Development
•
891

Figure 6. (a) Racemization of (S)-N-acetylphenylalanine at 70 °C and 85 °C for conventional (CH) and microwave heating (MW)
and (b) influence of temperature on the microwave effect.
7
0 °C, based on the absence of full racemization under
reported in the field of organic chemistry, applying conventional
13
conventional heating (see Figure 6), and a factor of nearly six
at 85 °C.
heating: so-called flow chemistry.
Today’s flow chemistry is based on the principle of Merri-
Θ
14
The heat of dissolution (∆
s
H ) of N-acetylphenylalanine as
field’s solid-phase peptide synthesis. Solid-phase chemistry
15
the racemic mixture or in the enantiomerically pure form was
calculated as approximately 85 kJ/mol (see eq 2). The solubility
itself is specifically used under flow conditions, and this type
of work has been extensively studied and promoted by, for
-1
(S; mol ·L ) of the substrate has been measured at a series of
16
example, Ley and Baxendale.
Θ
s
temperatures (T). The large value of ∆ H indicates that heat
or mass transfer (diffusion) is potentially rate-determining for
the racemization, and therefore, these limitations are most likely
to be located at the boundary layer of the solid particles.
Racemization appears to be fast, and hence, measurements in
the liquid phase show only the presence of a racemic mixture
for all substrates.
Initial research into flow chemistry by microwave heating
was done by Strauss and co-workers. The first publication of
his work appeared in 1992 with positive expectations of this
9
17
novel reactor design for the future. The combination of
microwave heating and flow chemistry was also investigated
by groups with expertise primarily in one of these enabling
18-20
techniques.
Attempts to scale up flow chemistry usually
Θ
The heat of dissolution (∆
s
H ) for N-acetylindoline-2-
involve homogeneous systems or heterogeneous systems with
11
carboxylic acid is 75 kJ/mol, also relatively high.
21
immobilized reagents/catalysts. There is, to the best of the
authors knowledge, only one small-scale example of an
Θ
S(T2)
∆ H
s
1
1
T1
22
inorganic solid-liquid continuous flow system. Interestingly,
ln
) -
-
(2)
(
)
(T
)
S(T1)
R
2
(
(
16) (a) Smith, C. D.; Baxendale, I. R.; Tranmer, G. K.; Baumann, M.;
Smith, S. C.; Lewthwaite, R. A.; Ley, S. V. Org. Biomol. Chem. 2007,
5, 1562. (b) Baxendale, I. R.; Deeley, J.; Griffiths-Jones, M.; Ley,
S. V.; Saaby, S.; Tranmer, G. K. Chem. Commun. 2006, 2566.
17) (a) Strauss, C. R. Chem. Aust. 1990, 186. (b) Chemat, F.; Poux, M.;
di Martino, J.-L.; Berlan, J. Chem. Eng. Technol. 1996, 19, 420. (c)
Roberts, B. A.; Strauss, C. R. Acc. Chem. Res. 2005, 38, 653. (d)
Cablewski, T.; Faux, A. F.; Strauss, C. R. J. Org. Chem. 1994, 59,
Transfer from Batch to Microwave-Heated Flow Processing
Only flow processing enables scaling up with the application
of microwave heating, due to the limited penetration depth of
microwaves. Most studies with continuous-flow reactors are
3
408.
(
11) Solubility in p-xylene of (S)-N-acetylphenylalanine is 3.2, 4.1. and
(18) Baxendale, I. R.; Pitts, M. R. Chem. Today 2006, 24, 41.
(19) Bierbaum, R.; Nüchter, M.; Ondruschka, B. Chem. Eng. Technol. 2005,
28, 427.
2
2
7 mmol/L at 85, 100. and 115 °C; of (R,S)-N-acetylphenylalanine is
.3, 5.4, and 21 mmol/L at 85, 100, and 115 °C and of (R,S)-N-
acetylindoline-2-carboxylic acid is 4.1, 8.9, and 24.4 mmol/L at 102,
(20) (a) N u¨ chter, M.; Ondruschka, B.; Weiss, D.; Beckert, R.; Bonrath,
W.; Gum, A. Chem. Eng. Technol. 2005, 28, 871. (b) N u¨ chter, M.;
M u¨ ller, U.; Ondruschka, B.; Tied, A.; Lautenschl a¨ ger, W. Chem.-Ing.-
Tech. 2002, 74, 910. (c) Will, H.; Scholz, P.; Ondruschka, B.;
Burckhardt, W. Chem.-Ing.-Tech. 2002, 74, 1254. (d) Baxendale, I. R.;
Hayward, J. J.; Ley, S. V. Comb. Chem. High Throughput Screening
2007, 10, 802. (e) Saaby, S.; Baxendale, I. R.; Ley, S. V. Org. Biomol.
Chem. 2005, 3, 3365. (f) Baxendale, I. R.; Griffiths-Jones, M.; Ley,
S. V.; Tranmer, G. K. Chem. Eur. J. 2006, 12, 4407.
1
16, and 132 °C, respectively.
(
(
12) Landolt-B o¨ rnstein, H. R. Zahlenwerte und Funktionen aus Physik,
Chemie, Astronomie, Geophysik und Technik; Springer-Verlag: Berlin,
950-1980; Vol. IV/2a.
13) (a) LaPorte, T. L.; Wang, C. Curr. Opin. Drug DiscoVery DeV. 2007,
0, 738. (b) Anderson, N. G. Org. Process Res. DeV. 2001, 5, 613.
c) Tundo, P. Continuous Flow Methods in Organic Chemistry; Ellis
1
1
(
Harwood: Chichester, 1991.
(
14) (a) Life During a Golden Age of Peptide Chemistry: Merrifield, B. In
Profiles, Pathways and Dreams: Autobiographies of Eminent Chemists;
American Chemical Society: Washington DC, 1993. (b) Jas, G.;
Kirschning, A. Chem.sEur. J. 2003, 9, 5708. (c) Tundo, P. Continuous
Flow Methods in Organic Chemistry; Ellis Horwood: Chichester, 1991.
15) Solodenko, W.; Wen, H.; Leue, S.; Stuhlmann, F.; Sourkouni-Argirusi,
G.; Jas, G.; Schonfeld, H.; Kunz, U.; Kirschning, A. Eur. J. Org. Chem.
(21) (a) Khadilkar, B. M.; Madyar, V. R. Org. Process Res. DeV. 2001, 5,
452. (b) Kabza, K. G.; Capados, B. R.; Gestwicki, J. E.; McGrath,
J. L. J. Org. Chem. 2000, 65, 1210. (c) Shieh, W.-C.; Dell, S.; Repi cˇ ,
O. Tetrahedron Lett. 2002, 43, 5607. (d) Baxendale, I. R.; Hayward,
J. J.; Ley, S. V. Comb. Chem. High Throughput Screening 2007, 10,
802.
(
(22) Bonaccorsi, L.; Proverbio, E. Microporous Mesoporous Mater. 2008,
112, 481.
2
004, 3601.
8
92
•
Vol. 13, No. 5, 2009 / Organic Process Research & Development

Figure 7. Batch-loop reactors for the racemization of (S)-N-
1
acetylindoline-2-carboxylic acid: R ) microwave-heated tu-
2
bular reactor; R ) double-walled reactor vessel with overhead
stirrer; P ) (membrane-)pump; C ) cooler with Archimedes
1
screw; T ) temperature sensor inside the upper part of the
2
tubular reactor and T ) temperature sensor just downstream
the cooler.
Table 1. Racemization of N-acetylindoline-2-carboxylic acid
under different reaction conditions (A and B) in continuous-
flow reactor in batch-loop mode
Figure 8. Racemization of experiment A in batch-loop mode
compared to (a) conventional and (b) microwave-heated batch
experiment under comparable conditions.
entry
1
parameters
(S)-N-acetylindoline- 178 g (S)
-carboxylic acid
experiment A
experiment B
126 g (S)
2
Table 2. Reaction rates (interval between 90 and 70% ee) of
the racemization of N-acetyl-indoline-2-carboxylic acid under
conventional and microwave heating conditions
0
.87 mol
0.62 mol
650 g
72 g
14.8 g
0.14 mol
15
2
3
4
p-xylene
acetic acid
acetic anhydride
679 g
a
22.3 g
0
21
a
overall
fractional
corrected
.22 mol
reaction rate residence reaction rate
-1
a
-1
5
6
wt % substrate
entry
conditions
[∆ee·min ]
time
[∆ee·min ]
wt % cosolVent
10
1
2
3
batch (a); CH
batch (b); MW
experiment A
-0.20
-0.34
-0.08
1
1
0.21
-0.20
-0.34
-0.36
(
acetic acid)
1
222 mL ·min-1
182 min
plugging
7
8
9
flow rate
total reaction time
interruption by:
216 mL ·min-
263 min
plugging
a
Fractional residence time in the reactor compared to the total reaction time
1
[
based on the ratio the volume of the microwave-heated tubular reactor (R , 180
mL) per total reaction volume (850 mL); overall rate ) VR
a
1
/Vtot · rate (MW, R )]
1
)
little water present, which will convert acetic anhydride into acetic acid
(
<2 wt %).
difference between these two runs was the amount of cosolvent.
In the first run, the racemization was performed without
cosolvent. In the second run, acetic acid was added to decrease
the heterogeneity, thereby reducing the risk of plugging.
Experiment A. In this experiment the racemization in
the majority (>75%)23 of fine chemical processes involves
heterogeneous reactions.
In general, reactions requiring short times for acceptable
conversions (<30 min) can be performed continuously. How-
ever, for medium long reaction times (<3 h) typically a batch-
loop mode is preferred (Figure 7).
24
p-xylene was performed without cosolvent. The results are
depicted in Figure 8 and Table 2.
Within the racemization of (S)-N-acetylindoline-2-carboxylic
acid, heterogeneity is governed by the solubility of the substrate.
Changes in reaction conditions, such as temperature or the
amount of cosolvent influence the performance of the batch-
loop reactor. Also, fouling may be influenced. Two experiments
under different conditions were conducted using (S)-N-acetyl-
indoline-2-carboxylic acid as the substrate; see Table 1.
Although each run was prematurely stopped due to plugging,
a prediction based on sampling made a comparison with the
performance of the stirred batch reactor possible. The main
Ideally, experiment A should have gone to completion to
demonstrate that complete racemization can take place in the
batch-loop reactor. However, plugging urged preliminary inter-
ruption of experiment A. Experience from earlier racemization
experiments taught us that inhibition is only observed during
4
the initial period, after which the reaction goes to completion.
This observation, combined with the relative reaction rates,
shows that the conversions in the batch-loop reactor closely
resemble those of the microwave-heated small-scale batch
(
25) After addition of acetic anhydride some technical problems occurred.
The reaction mixture was temporarily collected in one vessel at a
temperature below 20 °C. The system was restarted, and two large
aliquots were taken to perform two experiments batchwise by oil-
bath heating at 60 and 122 °C, see Figure 9. As a result of these large
aliquots, the total reaction volume in the microwave-heated batch-
loop system was less than that in experiment A. As a consequence
the fractional residence time () real reaction time) in tubular reactor
R1 (Figure 7) is higher than in experiment A, see entries 3 and 4 in
Table 3. Experiment B was executed at 127 °C in the batch-loop
reactor. The temperature in vessel R2 (Figure 7) was below 60 °C
(manually measured). The low reaction rate at 60 °C (Figure 9) allows
to neglect racemization in vessel R2. Reaction rates based on an
enantiomeric excess decrease from 86 to 34% ee (area between dotted
lines in Figure 9) were compared between three batch experiments
(c, d, e) and experiment B (Table 3).
(
(
23) Roberge, D. M. Org. Process Res. DeV. 2004, 8, 1049.
24) After addition of acetic anhydride to the reaction mixture a fast
decrease of the enantiomeric excess was observed in the small-scale
batch experiments (a and b of Figure 8) in the first period (<60 min).
During the microwave-heated experiment the racemization steadily
continued, while during conventional heating inhibition was tempo-
rarily observed (see also Figure 2). After 120 min the racemization
rate increased as the solubility of the substrate rose, see Figure 1. For
the small-scale batch reaction conventional and microwave heating
gave rise to different reaction rates. The flow experiment behaved
similarly to the microwave-heated batch experiment as illustrated
by the absence of an inhibition period after the addition of the
catalyst. Reaction rates based on enantiomeric excess drop from
9
0 to 70% ee are shown in Table 2 for both the batch experiments
and experiment A.
Vol. 13, No. 5, 2009 / Organic Process Research & Development
•
893

demonstrated in the batch-loop reactor with respect to larger-
scale applications.
Conclusions
For the homogeneous systems, reaction rate enhancements
could not be detected when conventional heating was replaced
by microwave irradiation. However, for heterogeneous mixtures
different results were obtained. Here the rate-determining step
appeared to be crucially influenced. During the racemization
of two types of N-acetylamino acids (N-acetylindoline-2-
carboxylic acid and N-acetylphenylalanine), significant micro-
26
wave effects were indeed observed under various conditions.
Changing the degree of heterogeneity directly influenced the
magnitude of the microwave effects. There is no doubt that there
is a positive correlation between the heterogeneity of the system
and the observed microwave effects. Presumably, the racem-
ization of N-acetylamino acids under the applied reaction
conditions occurs in the phase boundary region near the solid
particles. Neither the (pure) solvent (p-xylene; loss tangent ≈
Figure 9. Racemization of experiment B in batch-loop mode
compared to (c) conventionally heated at 60 °C, (d) convention-
ally heated at 122 °C, and (e) microwave-heated at 129 °C batch
experiments under comparable conditions.
27
Table 3. Reaction rates (interval between 86 and 34% ee) of
racemization of N-acetylindoline-2-carboxylic acid under dif-
ferent reaction conditions
0.0011) nor the solid (N-acetylamino acid) absorb microwaves
effectively. While the existence of hot-spots being unlikely to
occur in this situation, selective heating of the interphase
between solid and liquid is more plausible. Mechanistically
nothing is altered, but the rate-limiting step is influenced
overall
fractional
corrected
reaction rate residence reaction rate
-1
a
-1
entry
1
conditions
[∆ee·min ]
time
[∆ee·min ]
indirectly by microwave heating. In agreement with the high
batch (c); CH
-0.015
1
-0.015
-1.26
-2.05
-2.43
Θ
heat of solution (∆
s
H
) 85 kJ/mol), small changes in
(
60 °C)
batch (d); CH
122 °C)
batch (e); MW
temperature also have a larger effect on the solubility of the
N-acetylamino acid.
2
3
4
-1.26
-2.05
-0.55
1
1
(
In solution only a racemic mixture could be measured
throughout the whole racemization process. The reaction occurs
in the interphase and microwave energy directed to this layer,
where the reaction takes place, can lead to a beneficial effect
on the intrinsic reaction rates.
b
(
129 °C)
experiment B
127 °C)
0.23
(
a
Fractional residence time in the reactor compared to the total reaction time
1
[
based on the ratio the volume of the microwave-heated tubular reactor (R , 180
mL) per total reaction volume (785 mL); overall rate ) V
Reaction rate observed under microwave heating and comparable conditions.
c
1
1
R
/Vtot · rate (MW, R ).
b
c
V
tot
Experimental Section
)
Vpremixed - Vsampling; 785 mL ) 850 mL - 65 mL.
In our commercially available setup, automated multimode
microwave ovens were used, the MicroSynth and FlowSynth
are from Milestone srl (Italy). These ovens operate at 2.45 GHz
and are temperature controlled by means of a fiber-optic sensor
experiment. In this case, racemization can in principle partially
continue outside the tubular reactor, before a temperature is
reached at which the rate of racemization is negligible. However,
continuation of the racemization outside the microwave-heated
tubular reactor R appears to be insufficient to distinguish it
from a measurement error.
(ATC-FO, fluoroptic probe inserted via a Teflon-coated ceramic
well). During our microwave experiments the maximum power
1
was set at 100 W. The average power input was 18 W (50 °C)
1
and 76 W (100 °C). The reaction mixture was identified by H
Experiment B. In experiment B acetic acid was added to
NMR, GC-MS, and GC-FID, and conversion was determined
25
reduce the heterogeneity of the system. The results are depicted
in Figure 9 and Table 3.
1
1
3
by H NMR. H NMR spectra were recorded in CDCl using
a Varian Mercury 200 MHz with the aid of 1,3,5-tritert-
butylbenzene as the internal standard. The proton chemical shifts
were calibrated to tetramethylsilane (TMS).
Racemization of Wet (R)-N-Acetylindoline-2-carboxylic
Acid of 66% ee (MW). A round-bottomed flask was charged
Experiment B allows us to make a fair comparison between
microwave and conventional heating for the racemization
efficiency during a longer period (with respect to experiment
A) on the basis of drops in the enantiomeric excess. Again the
estimated conversion of the microwave-heated experiment
proved to be in line with expectations.
with wet (∼5 wt % H
2
O) N-acetylindoline-2-carboxylic acid
1.78 g, 8.7 mmol, 66% ee R) and p-xylene (5.5 g). This mixture
(
The racemization was effective during a longer period and
in a larger volume outside the microwave-heated tubular reactor
was heated in a Dean-Stark distillation setup to azeotropically
remove water (∼1.5 mL distillate). The total mass of the
1
R , due to the deliberate absence of cooling power in cooler C
(
Figure 7) as compared to that in experiment A.
(
26) (a) Jacob, J.; Chia, L. H. L.; Boey, F. Y. C. J. Mater. Sci. 1995, 30,
5321. (b) Stuerga, D. A. C.; Gaillard, P. J. MicrowaVe Power
Electromagn. Energy 1996, 31, 87. (c) Stuerga, D. A. C.; Gaillard,
P. J. MicrowaVe Power Electromagn. Energy 1996, 31, 101.
In summary, based on the experiments discussed in this
section it can be concluded that the efficacy of the microwave
effect described in the batch experiments has also been
(27) Magee, M. D.; Walker, S. Trans. Faraday Soc. 1966, 62, 1748.
8
94
•
Vol. 13, No. 5, 2009 / Organic Process Research & Development

reaction mixture was adjusted at 7.0 g with solvent, and the
generously dissolved in a mixture of formic acid (3 mL),
isopropanol (25 mL), and n-hexane (72 mL) and analyzed by
HPLC on a chiral column (Daicel, Chiracel OD) with an eluent
composed of formic acid (1 v %), isopropanol (10 v %), and
n-hexane (89 v %).
mixture was heated to 130 °C in the Milestone microwave oven.
(A Weflon bar was added during MW heating at 0 wt %
cosolvent.) Then, acetic anhydride (100 µL, 1.1 mmol) was
added. After 2 h water (100 µL) was added, and via a
Dean-Stark distillation setup p-xylene/water/acetic acid was
distilled from the reaction mixture (∼2.5 mL distillate). After
cooling, the solid was filtered and washed twice with p-xylene
[Experiment B]. A round-bottomed flask was charged with
dry (S)-N-acetylindoline-2-carboxylic acid (126 g, 0.62 mol,
99% ee S) and p-xylene (650 g). The reaction mixture was
circulated through the continuous-flow reactor of Milestone
equipped with a Weflon stirrer via a membrane pump with a
(
2 mL) and dried under reduced pressure at 50 °C. The reaction
was monitored by quenching a sample (∼50 mg) with water
25 µL) to neutralize the anhydride. The samples were gener-
ously dissolved in a mixture of formic acid (3 mL), isopropanol
25 mL), and n-hexane (72 mL) and analyzed by HPLC on a
-1
1
(
flow of 222 mL·min and heated to 130 °C (T ). No cooling
was applied on top of reactor. Subsequently, acetic anhydride
(14.8 g, 0.14 mol) was added to the flask, and the set tempera-
ture was maintained for a period of 182 min. The total reaction
volume was 785 mL. The average power usage during reaction
was calculated to be 280 W. During the reaction, aliquots (∼50
mg) were taken and quenched with water (25 µL) to neutralize
the anhydride. The samples were generously dissolved in a
mixture of formic acid (3 mL), isopropanol (25 mL), and
n-hexane (72 mL) and analyzed by HPLC on a chiral column
(Daicel, Chiracel OD) with an eluent consisting of formic acid
(1 v %), isopropanol (10 v %), and n-hexane (89 v %).
N-Acetyl-(S)-phenylalanine. A round-bottomed flask was
charged with (S)-phenylalanine (19.8 g, 0.12 mol) and aqueous
(30 mL) NaOH (4.8 g, 0.12 mol). To this mixture was added
acetic anhydride (12.24 g, 0.12 mol) dropwise with the simulta-
neous addition of an aqueous solution (15 mL) of NaOH (4.8
g, 0.12 mol) under vigorous stirring and at room temperature.
The temperature was kept below 30 °C, and pH was maintained
at around 10. The addition procedure of acetic anhydride in
combination with NaOH was repeated once after 30 min. This
mixture was extracted with ethyl acetate (100 mL), the pH was
set at 6 with 6 M HCl under vigorous stirring with cooling in
an ice-bath. The extraction procedure was repeated for pH
values set at 5, 4, and 3. The organic layers were combined
and subsequently washed with brine (100 mL) and demi-water
(
chiral column (Daicel, Chiracel OD) with an eluent composed
of formic acid (1 v %), isopropanol (10 v %), and n-hexane
(89 v %). The yield after workup was 1.35 g (76%). The same
procedure was followed for the conventionally heated experi-
ment by substituting microwave irradiation with oil-bath heating,
except for the usage of a Weflon bar at 0 wt % cosolvent. Yield
after workup: 1.62 g (91%).
Racemization of Dry (S)-N-Acetylindoline-2-carboxylic
Acid of 99% ee (MW). A round-bottomed flask was charged
with (S)-N-acetylindoline-2-carboxylic acid (1.78 g, 8.7 mmol,
9
1
9% ee (S)) and p-xylene (5.5 g). The mixture was heated to
30 °C in the Milestone microwave oven. (A Weflon bar was
added during MW heating at 0 wt % cosolvent.) Then, acetic
anhydride (100 µL, 1.1 mmol) was added. After 2 h water (100
µL) was added, and p-xylene/water/acetic acid was distilled
from the reaction mixture. After cooling, the solid was filtered
and washed twice with p-xylene (2 mL) and dried under reduced
pressure at 50 °C. The reaction was monitored by quenching a
sample (∼50 mg) with water (25 µL) to neutralize the
anhydride. The samples were generously dissolved in a mixture
of formic acid (3 mL), isopropanol (25 mL), and n-hexane (72
mL) and analyzed by HPLC on a chiral column (Daicel,
Chiracel OD) with an eluent composed of formic acid (1 v %),
isopropanol (10 v %), and n-hexane (89 v %). The same
procedure was executed for the conventionally heated experi-
ment substituting microwave irradiation with oil-bath heating,
except for the use of a Weflon bar at 0 wt % cosolvent.
Racemization of Dry (S)-N-Acetylindoline-2-carboxylic
Acid of 99% ee (CFR). [Experiment A]. A round-bottomed
flask was charged with dry (S)-N-acetyl indoline-2-carboxylic
acid (178 g, 0.87 mol, 99% ee (S)) and p-xylene (679 g). The
reaction mixture was circulated through the continuous-flow
reactor of Milestone equipped with a Weflon stirrer via a
(50 mL). Organic solvent was removed under reduced pressure.
1
The yield after workup was 21.0 g (84%). H NMR (CD
3
OD,
400 MHz) typical signals δ (ppm) 7.51 (m, 5H, Ar-H), 4.92
(m, 1H, CH-CH ), 3.45 (dd, 1H, CH-CH ), 3.20 (dd, 1H, CH-
CH ), 2.15 (s, 3H, CO-CH ). Measured: [R]
2
2
23
2
3
23
D
+29.7° (c )
+33° (c ) 10, MeOH).
28
10, MeOH), literature [R]
D
Acknowledgment
We are grateful to SenterNovem for funding this research.
We thank DSM Pharma Chemicals, Geleen, The Netherlands,
and Syncom B.V., Groningen, The Netherlands, for their
support.
-1
membrane pump with a flow of 216 mL ·min and heated to
1
1
30 °C (T ). Cooling was applied on top of the reactor.
Subsequently, acetic anhydride (22.3 g, 0.22 mol) was added
to the flask, and the set temperature was maintained for a period
of 263 min. Total reaction volume was 850 mL. Average power
usage during reaction was calculated to be 407 W. During the
reaction, aliquots (∼50 mg) were taken and quenched with
water (25 µL) to neutralize the anhydride. The samples were
Received for review June 2, 2009.
OP9001356
(28) Smith, H. E.; Neergaard, J. R. J. Am. Chem. Soc. 1996, 118, 7694.
Vol. 13, No. 5, 2009 / Organic Process Research & Development
•
895