Intensified azeotropic distillation: A strategy for optimizing direct amidation

Article abstract of DOI:10.1021/op300058f

The direct synthesis of amides from the corresponding carboxylic acids and amines is shown to operate under varying degrees of mixed kinetic and mass transfer rate control when water is removed by azeotropic distillation. Unless the volumetric heat input rate is reported, it is not possible to make a valid comparison between different catalysts, as the difference in Q_boil alone can be responsible for the apparent difference in observed rate. A systematic approach is developed to quantify the contribution of boil-up rate to conversion rate and so decouple the physical rates from the chemistry. Intensive boiling is used to improve the removal of water during azeotropic distillation and considerably enhance conversion. The results show that some acylations previously thought to be difficult or impossible can be achieved in the absence of coupling agents under green conditions. The use of a cascade of CSTR flow reactors operating under intensified conditions is assessed for scale up of direct amidation reactions and compared to a production scale batch reactor. The findings and conclusions of this work have general applicability to all condensation reactions.

Full text of DOI:10.1021/op300058f

Article
pubs.acs.org/OPRD
Intensified Azeotropic Distillation: A Strategy for Optimizing Direct
Amidation
Christophe Grosjean,*,† Julie Parker,‡ Carl Thirsk,† and Allen R. Wright‡
†LyraChem Ltd, Saint Thomas Street Stables, Saint Thomas Street, Newcastle-upon-Tyne, NE1 4LE, U.K.
‡School of Chemical Engineering and Advanced Materials, Merz Court, Newcastle University, Newcastle-upon-Tyne, NE1 7RU, U.K.
S
* Supporting Information
ABSTRACT: The direct synthesis of amides from the
corresponding carboxylic acids and amines is shown to operate
under varying degrees of mixed kinetic and mass transfer rate
control when water is removed by azeotropic distillation.
Unless the volumetric heat input rate is reported, it is not
possible to make a valid comparison between different
catalysts, as the difference in Qboil alone can be responsible
for the apparent difference in observed rate. A systematic
approach is developed to quantify the contribution of boil-up
rate to conversion rate and so decouple the physical rates from the chemistry. Intensive boiling is used to improve the removal of
water during azeotropic distillation and considerably enhance conversion. The results show that some acylations previously
thought to be difficult or impossible can be achieved in the absence of coupling agents under green conditions. The use of a
cascade of CSTR flow reactors operating under intensified conditions is assessed for scale up of direct amidation reactions and
compared to a production scale batch reactor. The findings and conclusions of this work have general applicability to all
condensation reactions.
high to be achieved on most substrates without using forcing
conditions, which makes this a method with little synthetic
utility. Within the past decade, catalysts have been developed to
lower this activation energy and enable condensation of the
carboxylic acid with amines.7 Arylboronic acids, in particular,
proved efficient catalysts. Additionally, the use of boric acid as a
cheap and effective alternative to more complex boron
derivatives is growing in favor and has been applied successfully
to the large-scale preparation of APIs.8
The mechanism by which the boron species promotes the
formation of amides from carboxylic acids and amines is yet to
be fully determined.9,10 However, even under catalytic
conditions, water removal is usually needed and the importance
of removing water to ensure high conversion levels is regularly
reported.5b,11 Water removal is generally achieved using a
combination of molecular sieves, Soxhlet extraction, and
calcium hydride or using azeotropic distillation with toluene
or xylenes as entrainers together with a Dean−Stark apparatus
(or phase-splitter on plant) to effect separation of the biphasic
mixture.12 The latter is a form of reactive distillation.
1. INTRODUCTION
Direct Amidation. Amide synthesis is undoubtedly one of
the most important bond forming reactions. A recent survey
carried out by three leading pharmaceutical manufacturers
reported that amide bond formation was utilized in the
synthesis of 65% (84 of 128) of the drug candidates surveyed.1
The most common synthetic strategies include activation of the
carboxylic acid, either by forming the acyl chloride (44% of
cases) or mixed anhydride or by using a coupling reagent.2
Although these methods are very well-established in
pharmaceutical manufacturing, they are disadvantaged by the
need for additional process stages and poor atom economy
reagents.3 Indeed, a recent poll from the ACS Green Chemistry
Institute Pharmaceutical Roundtable reported that amide
formation avoiding poor atom economy reagents received the
most number of votes. The poll surveyed and prioritized
common reactions for which member companies felt better
alternatives would bring significant environmental and
economic benefits. These benefits arise mainly through
reductions in waste produced, use of hazardous reagents, and
number of process operations.4
Azeotropic Reactive Distillation and Mixed-Rate
Control Systems. Reactive distillation is a widely used
technique to drive equilibrium-limited reactions such as
condensations to high levels of conversion. Previous
commercial work on a production-scale condensation identified
a positive correlation between boil-up rate and reaction rate,
Because it allows the synthesis of the amide bond in a single
step without the need for any reagents, direct amide synthesis
from the carboxylic acid and the amine is an attractive
alternative. The condensation of carboxylic acids with amines at
elevated temperatures is a known reaction5 that is generally
thought to be the result of “pyrolysis” of the ammonium
carboxylate salt.6 It is commonly assumed that the activation
energy needed for the salt to be converted to the amide is too
Received: March 7, 2012
Published: March 30, 2012
© 2012 American Chemical Society
781
dx.doi.org/10.1021/op300058f | Org. Process Res. Dev. 2012, 16, 781−787

Organic Process Research & Development
Article
which was exploited to bring about significant process
improvements on plant. Yield was increased, and both impurity
levels and batch times were reduced. In the case of
condensation reactions generating water, the position of the
equilibrium can be displaced by removing water to promote
formation of the desired product (Figure 1).
2. RESULTS AND DISCUSSION
In the commercial process (vide supra), water was removed on
plant using azeotropic distillation and a phase-splitter. The boil-
up rate was controlled by varying the rate of heat transfer to the
vessel Qoverall, which itself can be controlled by adjusting the
temperature difference between the heating medium and the
reaction temperature (ΔT). Variation of ΔT experimentally
therefore provides a convenient means to assess the impact of
heat transfer on the efficiency of water removal and hence
conversion.
Volumetric Heat InputQboil. The overall rate of heat
transfer per unit volume of mixture, Qoverall, is a function of the
overall heat transfer coefficient, U, the heat transfer area, A, and
the temperature difference between the heating medium and
the reaction temperature, ΔT.13,14 Under steady state temper-
ature (boiling conditions):
Figure 1. General mixed-rate controlled condensation.
The observed rate of product formation, robs, is a function of
the parallel processes of chemical reaction (r1, r2) and water
removal by evaporation (revap).
VQ overall = UAΔT
(4)
In practice heat losses (Qloss) occur through the top of the
flask and connecting glassware causing vapors to condense on
the walls before reaching the condenser. These losses may be
considerable and vary from experiment to experiment without
removing water from the system. Thus:
robs = f(kinetic rate, mass transfer rate)
(1)
Expressing reaction rates for the species as a power law
function and the boil-up rate for water using Raoult’s Law, eq 1
can be written in the form of a mass balance per unit volume.
Q overall = Qboil + Qloss
−d[A]
(5)
robs
=
=
=
dt
−d[B]
dt
Combining eqs 4 and 5
Qboil = UAΔT/V − Qloss
(6)
It is possible to measure Qboil as a function of ΔT15 by
monitoring the rate of condensation from a boiling solution.
Experimental data for a range of entrainers were collected
(Figures 2 and 3). The lowest values of Qboil used were those
d[C]
dt
= {k1[A]a[B]b − k2[C]c [H2O]d}
(2)
(3)
d[H2O]
= {k1[A]a[B]b − k2[C]c [H2O]d}
dt
Q
γ
χ
ρ
boil H2O H2O H2O
−
∑1n (γχ ρ)
i i
i
where the term on the right-hand side of eq 3 is the rate of
water evaporation from a mixture and Qboil is the heat input per
unit volume of the reaction mixture to provide boiling.
The observed rate of the reaction varies between the two
extremes of kinetic control and mass transfer (evaporation rate)
control. When the rate of water removal and the rate of the
chemical reaction are of similar magnitude, both affect the
overall observed rate of product formation, and the system is
said to be under mixed-rate control. Typically, observed reaction
times of between 0.5 and about 10 h fall into this class.
When water is removed via azeotropic distillation, physical
factors affecting the system boil-up rate, such as Qboil, will retard
or accelerate the water evaporation rate. This has important
implications when assessing system reactivity and when
screening catalysts for mixed-rate control systems. If the effect
of physical factors is not isolated and systematically quantified,
their influence on observed rates may be wrongly attributed to
chemical factors. By intensifying Qboil, reactions will tend
toward kinetic rate control.
Figure 2. Dependence of the volumetric heat input, Qboil, on ΔT: 100
■
●
▲
mL flask; , water; , toluene; , chlorobenzene.
corresponding to the lowest ΔT enabling both nucleate boiling
as well as a constant vapor−liquid circuit inside the Dean−
Stark system. The dependence of Qboil on ΔT was found to be
linear in the region under study (Table 1).
The overall heat transfer coefficient, U, may be calculated
from the slopes and Qloss from the intercepts. In the case where
different entrainers are compared (Figure 2), the same glass
vessel was used; hence, the slopes (i.e. heat transfer coefficient)
are the same but the losses are different due to the differing
boiling points of the solvents. Figure 3 shows the effect of
boiling water in two different sizes of vessel. The difference in
The purpose of this paper is to demonstrate the benefits of
applying a systematic approach to decoupling chemical reaction
rates from the physical rate of evaporation. This is exemplified
by the case of direct amidation reactions carried out under
azeotropic removal of water.
782
dx.doi.org/10.1021/op300058f | Org. Process Res. Dev. 2012, 16, 781−787

Organic Process Research & Development
Article
Scheme 1
Figure 3. Dependence of the volumetric heat input, Qboil, on ΔT:
■
▲
water; , 100 mL flask; , 250 mL flask.
the slopes between the two reactors is due to the combination
of the differences in surface/volume ratios and glass
thicknesses.
In this case Qloss does not vary significantly ( 6%) with the
flask size. This is indicative that most of the heat losses are
caused by the Dean−Stark apparatus and its connection to the
reaction vessel. Indeed, under the conventional design of the
apparatus, the vapor phase is partially condensed by the cooler
condensate phase returning to the vessel before reaching the
condenser.
Effect of Qboil on robs for Direct Amidation. The
acylation of benzylamine (2) by phenylacetic acid (1) was
chosen as a representative system for direct amidation run
under azeotropic distillation conditions (Scheme 1).
When mixed, phenylacetic acid and benzylamine rapidly
form the corresponding ammonium carboxylate salt, 4, in an
exothermic reaction. It is therefore important that the amine is
slowly added to a cold solution of the acid. Alternatively, it was
found that the salt could be isolated and used as starting
material.
◆
■
Figure 4. Conversion of 3 with time; , ΔT = 14.3 °C; , ΔT = 34.3
●
▲
°C; , 5 mol % B(OH)3, ΔT = 14.3 °C; , 5 mol % B(OH)3, ΔT =
34.3 °C.
Table 2. Effects of Varying Qboil on the Yield of the Direct
Amidation of Phenylacetic Acid by Benzylamine
a
yield
b
c
ΔT (°C)
Qboil (kW m−3
)
no catalyst
B(OH)3 (5 mol %)
14.3
34.3
40
17%
37%
34%
84%
482
a
b
Experiments were carried out in toluene using the glassware
equipment described above. The corresponding values for Qboil
at a particular ΔT were obtained from Table 1.
GC yield. Acid/amine, 1:1; toluene; 0.5 M; reflux; Dean−Stark; 2 h.
c
Acid/amine, 1:1; B(OH)3 (5 mol %); toluene; 0.5 M; reflux; Dean−
Stark; 2 h.
In the absence of catalyst, the increase in Qboil from 40 to 482
kW m−3 causes the conversion rate to be doubled (Figure 4
(◆ vs ■) and Table 2). This is a significant finding and in
accordance with the mixed-rate controlled nature of the system.
When operating at a higher rate of water removal, the reaction
tends to be limited by the forward rate for the acylation
reaction. This in turn is limited by the association constant
between the ammonium salt and the amine/carboxylic acid
system. The results show that when operating at a low boil-up
rate the direct acylation between phenylacetic acid and
benzylamine could be regarded as a system of relatively low
reactivity. Increasing Qboil demonstrates that it is in fact not the
case and that high conversion levels can be achieved under
conditions where the reaction is not mass-transfer-limited. The
approach presented here enables the determination of the
optimal volumetric boil-up rate for direct amidation using a
particular vessel and solvent system.
Addition of a catalyst further enhances the rate of conversion
for both low and high Qboil rates. Using a ΔT value of 34 °C
(Qboil = 482 kW m−3) in the presence of boric acid achieves
conversion levels of ca 85% in 2 h. This compares favorably
with previously reported conversions under different sets of
conditions (Table 3).
Table 1. Dependence of Qboil on ΔT for Different Solvents
16
ΔHvap (kJ mol−1
solvent
water
flask (mL)
bp (°C)
d (g cm−3
)
)
linear dependence
100
250
100
100
100.0
100.0
110.6
132.0
1.00
45.8 at 373.75 K
45.8 at 373.75 K
33.3 at 383.8 K
35.2 at 405.2 K
Qboil = (24.2 × ΔT) − 238
Qboil = (13.7 × ΔT) − 210
Qboil = (22.1 × ΔT) − 276
Qboil = (25.0 × ΔT) − 623
water
1.00
toluene
0.867
1.106
chlorobenzene
783
dx.doi.org/10.1021/op300058f | Org. Process Res. Dev. 2012, 16, 781−787

Organic Process Research & Development
Article
Table 3. Yields for the Direct Acylation of Benzylamine by Phenylacetic Acid
ref
solvent
temp (°C)
time (h)
water removal
catalyst/coupling agent
yield
17
18
19
20
21
22
11c
23
neat
140
2
6
sieves
none
none
95%
96%
82%
85%
95%
94%
99%
92%
85%
a
o-xylene
THF
reflux
Dean−Stark
N/A
RT
RT
14
0.5
7
Sn[N(TMS)2]2 (1 equiv)
DCM
CHCl3
DCM
DCM
MeCN
toluene
N/A
2-pyridon-1-yl diphenyl phosphate (1 equiv)
Fe3+-K-10 montmorillonite
ODCT50 4/16/NMM(3 equiv)
o-bromophenylboronic acid (cat.)
B(OMe)3 (2 equiv)
a
reflux
none
RT
RT
80
2
N/A
48
15
2
sieves
none
b
this work
reflux
Dean−Stark
B(OH)3 (cat.)
a
b
Qboil unknown. Qboil = 482 kW m−3.
It is particularly noteworthy from this work that, in the
least one of the reactants is of high or moderate reactivity.
Hence, in the case of benzoic acid and 4-phenylbutylamine,
only a marginal improvement in conversion is observed.
Acylation of secondary amines by carboxylic acids is often
regarded as an even more unfavorable reaction than that of
their primary counterparts. Direct synthesis of tertiary amides
has scarcely been reported in spite of the range of applications.
Useful targets include N-acylpiperidine mosquito repellent
candidates.24 The only reported synthetic procedure to date
involves the intermediacy of the corresponding acylbenzo-
triazole via the acylchloride (Scheme 2).
absence of catalyst, raising the volumetric heat input to 482 kW
m−3 compared to the results obtained with a low rate of Qboil
has similar effects to the addition of 5 mol % boric acid. Thus,
in the absence of a quantified and controlled rate of Qboil or
other means of water removal, the reported performance
ascribed to a catalyst may be in serious error.
Three other sets of substrates were studied, and the catalytic
effect of boric acid on the reactions was assessed under
conditions where the system is not mass-transfer-limited (see
Table 4), using a Qboil of 482 kW m−3.
Table 4. Conversions Recorded for Uncatalyzed and Boric
Acid-Catalyzed Direct Amidations under Non-Mass-
Transfer-Limited Conditions
Scheme 2. General Procedure for the Preparation of N-
Acylpiperidines for Use As Insect Repellents (Bt,
Benzotriazolyl)
One of the most potent of the targets, when compared
against DEET (N,N-diethyl-m-toluamide), was found to be 4-
methyl-1-(1-oxo-10-undecylenyl)piperidine (7), for which a
yield of 76% was obtained after 4 h under the general
procedure in THF at room temperature.25
The direct acylation of 4-methylpiperidine (6) by undecy-
lenic acid (5) in refluxing toluene is a heterogeneous reaction
(Scheme 3). The reaction was initially trialled in the absence of
a
Scheme 3
GC yield; acid/amine, 1:1; toluene; 0.5 M; reflux; Dean−Stark; 482
kW m−3; 5.5 h.
The low reactivity of the benzoic acid-based systems
compared to other aromatic carboxylic acids had already been
reported.10f,11a,b However, this is the first report of conversions
recorded under conditions where boil-up is quantified and
maintained constant. Hence, it is possible to confirm that
benzoic acid is less reactive a carboxylic acid than phenylacetic
acid independently of water removal efficiency. However, it is
also apparent that 4-phenylbutylamine is less reactive than
benzylamine. Consequently, in the cases where the heavier
amine is used in the absence of a catalyst, the difference in
reactivity between phenylacetic acid and benzoic acid (7% vs
2%, respectively) is significantly less than that observed in the
case of benzylamine (69% vs 3%, respectively).
a catalyst with the more elevated Qboil of 482 kWm−3 (ΔT =
34.3 °C). Only 12% conversion was recorded after 8 h (Figure
5; ■). Comparatively, under the conditions used (Qboil = 482
kW m−3 in a 100 mL flask) the conversion to 3 from 1 and 2
was not mass transfer-limited. Therefore, the results indicate
that the undecelynic acid/piperidine system is of low reactivity.
The enhancing effect of the catalyst can also be quantitatively
assessed. In all cases, boric acid has increased the rate of
conversion. The effect is more pronounced in systems where at
784
dx.doi.org/10.1021/op300058f | Org. Process Res. Dev. 2012, 16, 781−787

Organic Process Research & Development
Article
vapor space. Continuous stirred tank reactors (CSTR’s) do
provide a vapor space and are suitable for use with reactive
distillation systems.
A train of CSTR flow reactors with very high volumetric heat
transfer rates could be configured to provide intensified reactive
distillation for condensation reactions. The system is outlined
in Figure 6. The vapor from each vessel would be condensed
■
●
Figure 5. Conversion of 6 with time: , ΔT = 34.3 °C; , 5 mol %
▲
○
B(OH)3, ΔT = 14.3 °C; , 5 mol % B(OH)3, ΔT = 34.3 °C; , 20
mol % B(OH)3, ΔT = 14.3 °C; , 20 mol % B(OH)3, ΔT = 34.3 °C.
△
Higher conversions were attained by using 5 mol % boric
acid; 25% of 7 after ca. 6.5 h (Figure 5; ● and ▲). However,
no further improvement was gained by raising the amount of
heat provided to the system; the reaction is kinetic rate-limited.
By increasing the catalyst loading to 20 mol % boric acid, the
system shifts toward mixed-rate control. At a Qboil rate of 40 kW
m−3 (ΔT = 14.3 °C), only a marginal improvement in
conversion (30% in 6.5 h) was observed whereas when Qboil is
increased to 482 kW m−3, conversion levels of 49% are achieved
in 7 h (Figure 5; ○ and △).
Implications to Process Chemistry and Catalyst
Screening. For both the secondary and tertiary amides
studied in this work, the highest rate of conversion was
achieved through a combination of catalyst and intensified
boiling.
These observations have important consequences on the
screening of catalysts for direct amidation using azeotropic
distillation. Unless the volumetric heat input rate is reported, it
is not possible to make a valid comparison between different
catalysts, as the difference in Qboil alone can be responsible for
the apparent difference in observed rate. Reduction of the heat
losses will also have beneficial effects. Any change in Qloss (i.e.,
better or worse insulation) has to be taken into account when
comparing results.
The underlying principles are not restricted, however, to
direct amidation, but they have wider relevance to other mixed-
rate control systems, which includes the large class of other
condensation reactions. A systematic approach such as the one
developed here could therefore be used to determine which of
these systems would benefit from an optimized intensified
boiling regime, as well as to determine the true efficiency of
condensation catalysts and cocatalysts.
Scale-up Considerations. Although intensified boiling
combined with an inexpensive catalyst can accelerate the rate of
direct amidation to a level where it becomes an attractive
method for production scale, the use of conventional glass-lined
reactors is problematic. For example, a 5 m3 Pfaudler reactor
with U = 350 W m−2 K−1 and jacket heat transfer area of 13 m2,
heated by 2.5 bar steam would achieve Qboil of only 15.5 kW
m−3. Using such equipment intensified boiling is not possible,
and other reactor types must be considered. Flow reactors in
the form of tubes would have a high heat transfer area to
reactor volume ratio; however, they are not suitable for use with
a reactive distillation system, as they do not provide a usable
Figure 6. Schematic of a CSTR train for production.
using a single condenser and fed into a single phase separator.
The outlet from the phase separator would feed into each
reactor simultaneously.
The performance of the system may be calculated using
conventional reaction engineering methods.26 Phenylacetic
acid−benzylamine was chosen as the reference system, and
reaction rates from Figure 4 (▲, 5 mol % B(OH)3, Qboil = 482
kWm−3) were used to calculate yields for each reactor in the
train of CSTR’s. Under the non-mass-transfer-limited con-
ditions, the reaction is pseudo-first-order with a rate constant of
3.5 × 10−4 s−1. The calculations assume a jacketed stainless
steel cylindrical vessel (U = 700 W m−2 K−1) fed with the
reaction mixture in toluene. Each vessel is 5 L in volume, and
the temperature differential between the jacket and reaction
mixture is 20 °C (377 kW m−3). The calculations suggest that
8−10 stages would be required to achieve >97% conversion
with a production rate of 42 tonnes per annum.
The simulated performance of a 1 m3 glass lined batch
production reactor is also shown for comparison again using a
20 °C jacket/reactor temperature differential (Qboil = 32 kW
m−3).27 Batch times are based on the 36 h amidation reaction
time plus 6 h to allow for the charging of the carboxylic acid
and toluene, the highly exothermic addition of the amine to
form the salt, and finally discharging the vessel. This vessel
would also produce 42 tpa. However, the total volume of the
785
dx.doi.org/10.1021/op300058f | Org. Process Res. Dev. 2012, 16, 781−787

Organic Process Research & Development
Article
flow reactor system in this calculation is 20 times smaller than
the batch vessel, and rather surprisingly, the total energy used
for boil up is only 55% that of the batch reactor.27 Although the
batch reactor operates at a low Qboil, its size and long boil-up
time result in this extra energy requirement. If the temperature
differential were raised to 40 °C (Qboil = 754 kW m−3) in the
CSTR’s, only six stages would be required to achieve the same
performance. Although at first sight a volumetric heat input rate
of this order may seem high, it is still much less than the
volumetric heat input rate of a domestic hot water kettle
(>2000 kW m−3) and far beyond the capability of a typical
batch production vessel.
placed in the oil bath. The mixture was stirred and samples
taken at regular intervals. Samples were analyzed by GC-MS
[injection temperature 250 °C; oven temperature; 50 °C for 0.5
min and then 50−80 at 5 °C min−1 and then 80−250 at 30 °C
min−1 and then hold at 250 °C for 8 min].
Amides Isolation. Following on from the general
procedure carried out under catalytic conditions:
The mixture was left to react for a further hour and was then
cooled down and filtered. The salt was washed with toluene (2
× 25 mL) and the combined organic phases concentrated in
vacuo. The residue was dissolved in ethyl acetate (50 mL) and
then washed with 5% HCl (50 mL), brine (50 mL), 5% NaOH
(50 mL), and brine (50 mL). The organic phase was dried over
MgSO4 and the solvent removed in vacuo.
3. CONCLUSIONS
N-Benzylphenylacetamide. The general procedure for
amide isolation was followed. Spectroscopic details were the
same as those reported in the literature.27
N-Benzylbenzamide. The general procedure for amide
isolation was followed. Spectroscopic details were the same as
those reported in the literature.27
The direct synthesis of amides from the corresponding
carboxylic acids and amines is shown to operate under varying
degrees of mixed kinetic and mass transfer rate control when
water is removed by azeotropic distillation. Unless the
volumetric heat input rate is reported, it is not possible to
make a valid comparison between different catalysts, as the
difference in Qboil alone can be responsible for the apparent
difference in observed rate. By characterizing heat transfer and
evaporation/boiling processes and relating these phenomena to
conversion rates, a systematic screening method to reliably
quantify the influence of intensified boiling in direct amidation
reactions has been developed.
N-4-Phenylbutylbenzamide. The general procedure for
amide isolation was followed. Spectroscopic details were the
same as those reported in the literature.9
b
N-4-Phenylbutylphenylacetamide. The title compound
(3.26 g, 50% yield) was isolated as a white solid recrystallized
1
from ethyl acetate/hexane. H NMR (700 MHz, CDCl3) δ
Intensified boiling during azeotropic reactive distillation
combined with an inexpensive catalyst can accelerate the rate of
direct amidation to a level where it becomes an attractive
method for production scale. The use of a cascade of CSTR
flow reactors operating under intensified conditions could
provide the necessary high rates of heat transfer and so offer
considerable advantage over a conventional batch reactor
system.
1.43−1.47 (m, 2H), 1.53−1.58 (m, 2H), 2.58 (t, 2H, J = 7.7
Hz), 3.21−3.24 (m, 2H), 3.56 (s, 2H), 7.11−7.36 (m, 10H);
13C NMR (176 MHz, CDCl3) δ 28.4, 29.0, 35.4, 39.4, 43.9,
125.8, 127.3, 128.3, 128.3, 129.0, 129.4, 135.0, 142.0, 170.8;
HRMS (ES+) [M+] 268.1725 (C18H21NO requires 268.1701);
m/z (ES+) 290 (M + Na, 100%), 268, 196; m/z (EI) 268, 267,
176, 91 (100%); elemental analysis (%): calcd for C18H21NO:
C 80.86, H 7.92, N 5.24, found C 81.00, H 7.81, N 5.05; mp =
61 °C.
4. EXPERIMENTAL SECTION
4-Methyl-1-(1-oxo-10-undecylenyl)piperidine. The general
procedure for amide isolation was followed. Spectroscopic
details were the same as those reported in the literature.25
Materials and Equipment. The reactions were carried out
in either a standard 3-neck 100 mL flask or a 3-neck 250 mL
flask. The Dean−Stark apparatus was a commercially available
piece of apparatus from Sigma-Aldrich.
ASSOCIATED CONTENT
■
All 1H NMR spectra were recorded on either Varian
Mercury-400 or Bruker Avance-400 spectrometers. 13C NMR
spectra were recorded on the Varian Mercury-400 and Bruker
Avance-400 instruments at frequencies of 100 MHz. Chemical
shifts are expressed as parts per million downfield from the
internal standard TMS. Mass spectra for liquid chromatog-
raphy−mass spectrometry (LCMS) were obtained using a
Waters Ltd LCT Premier XE spectrometer running in EI mode
with the positive electrospray method. GC-MS analysis was
carried out on an Agilent 6890N GC equipped with a 5973N
MSD Performance Turbo CI running in EI mode, and an
Anatune Focus Autosampler/liquid handler. Melting points
were recorded using a Stuart Scientific melting point apparatus,
Analogue, SMP 11.
S
* Supporting Information
N-4-Phenylbutyl-phenylacetamide characterization data, includ-
1
ing H NMR spectrum, 13C NMR spectrum, GC chromato-
gram, and electrospray MS results; and productivity calculations
for a train of CSTR’s and a batch reactor. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: christophe.grosjean@lyrachem.com.
Notes
The authors declare no competing financial interest.
Chemical reagents and materials were purchased directly
from Sigma-Aldrich and used without further purification,
unless stated otherwise.
Direct Amidation: General Procedure. The oil bath
temperature was set to 145 °C. In a 3-neck 100 mL round-
bottom flask were placed toluene (50 mL), the acid (0.025
mol), the amine (0.025 mol), and boric acid (5 mol %, 77.3
mg). The flask was fitted with a Dean−Stark apparatus and
ACKNOWLEDGMENTS
■
We thank Dr. Alan Kenwright and the Durham University
Solution State NMR Service for running the 1H and 13C NMR
spectra. We also thank the Elemental Microanalysis and Mass
Spectroscopy services at Durham University for their help with
analyses.
786
dx.doi.org/10.1021/op300058f | Org. Process Res. Dev. 2012, 16, 781−787

Organic Process Research & Development
Article
(22) Rolfe, A.; Probst, D. A.; Volp, K. A.; Omar, I.; Flynn, D. L.;
Hanson, P. R. J. Org. Chem. 2008, 73, 8785.
(23) Starkov, P.; Sheppard, T. D. Org. Biomol. Chem. 2011, 9, 1320.
(24) Katritzky, A. R.; Wang, Z.; Slavov, S.; Tsikolia, M.; Dobchev, D.;
Akhmedov, N. G.; Hall, D. C.; Bernier, G. C.; Linthicum, K. J. Proc.
Natl. Acad. Sci. 2008, 105, 7359.
(25) Katritzky, A. R.; Wang, Z.; Slavov, S.; Tsikolia, M.; Dobchev, D.;
Akhmedov, N. G.; Hall, D. C.; Bernier, G. C.; Linthicum, K. J. Proc.
Natl. Acad. Sci. 2008, 105, 7359 Supporting Information. Whereas
the characterization data report a 76% yield, Table S3 (p 14) lists a
95% yield..
REFERENCES
■
(1) Carey, J. S.; Laffan, D.; Thomson, C.; Williams, M. T. Org.
Biomol. Chem. 2006, 4, 2337.
(2) Montalbetti, C. A. G. N.; Falque, V. Tetrahedron 2005, 61, 10827.
(3) Constable, D. C.; Dunn, P. J.; Hayler, J. D.; Humphrey, G. R.;
Leazer, J. L., Jr.; Linderman, R. J.; Lorenz, K.; Manley, J.; Pearlman, B.
A.; Wells, A.; Zaks, A.; Zhang, T. Y. Green Chem. 2007, 9, 411.
(4) American Chemical Society Green Chemistry Institute
Pharmaceutical Roundtable: http://portal.acs.org/portal/
PublicWebSite/greenchemistry/industriainnovation/roundtable/
CTP_005609.
(26) See the Supporting Information.
(27) Nordstrøm, L. U.; Vogt, H.; Madsen, R. J. Am. Chem. Soc. 2008,
130, 17672.
(5) (a) Mitchell, J.; Reid, E. E. J. Am. Chem. Soc. 1931, 53, 1879 and
references cited therein. (b) Jursie, B. S.; Zdravkovski, Z. Synth.
Commun. 1993, 23, 2761.
(6) Smith, M. B.; March, J. March’s Advanced Organic Chemistry, 5th
ed.; Wiley Interscience: 2001; p 508.
(7) (a) Pattabiraman, V. R.; Bode, J. W. Nature 2011, 480, 471.
(b) Ishihara, K. Tetrahedron 2009, 65, 1085. (c) Allen, C. L.;
Chhatwal, R. A.; Williams, J. M. J. Chem. Commun. 2012, 48, 666.
(d) Lundberg, H.; Tinnis, F.; Adolfsson, H. Chem. Eur. J. 2012 [Online
early access]. DOI: 10.1002/chem.201104055. Published online:
February, 24th 2012.
(8) (a) Tang, P. Org. Synth. 2005, 81, 262. (b) Mylavarapu, R. K.;
GCM, K.; Kolla, N.; Veeramalla, R.; Koilkonda, P.; Batthacharya, A.;
Bandichlor, R. Org. Process Res. Dev. 2007, 11, 1065.
(9) Charville, H.; Jackson, D.; Hodges, G.; Whiting, A. Chem.
Commun. 2010, 46, 1813 and references cited therein.
(10) (a) Sakakura, A.; Ohkubo, T.; Yamashita, R.; Akakura, M.;
Ishihara, K. Org. Lett. 2011, 13, 892. (b) Sakakura, A.; Ohkubo, T.;
Yamashita, R.; Akakura, M.; Ishihara, K. Aust. J. Chem. 2011, 64, 1458.
(c) Charville, H.; Jackson, D.; Hodges, G.; Whiting, A. Eur. J. Org.
Chem. 2011, 30, 5981. (d) Shteinberg, L. Y. Russ. J. Appl. Chem. 2011,
84, 815. (e) Marcelli, T. Angew. Chem., Int. Ed. 2010, 49, 6840.
(f) Arnold, K.; Davies, B.; Giles, R. L.; Grosjean, C.; Smith, G. E.;
Whiting, A. Adv. Synth. Catal. 2006, 348, 813. (g) Ishihara, K.; Ohara,
S.; Yamamoto, H. J. Org. Chem. 1996, 61, 4196. (h) For a review on
boron-catalyzed direct amidation, see: Maki, T.; Ishihara, K.;
Yamamoto, H. Tetrahedron 2007, 63, 8645 and references herein.
(11) (a) Chaudhari, P. S.; Salim, S. D.; Sawant, R. V.; Akamanchi, K.
G. Green Chem. 2010, 12, 1707. (b) Comerford, J. W.; Clark, J. H.;
Macquarrie, D. J.; Breeden, S. W. Chem. Commun. 2009, 2562. (c) Al-
Zoubi, R. M.; Marion, O.; Hall, D. G. Angew. Chem., Int. Ed. 2008, 47,
2876. (d) Shteinberg, L. Y.; Kondratov, S. A.; Shein, S. M.;
Marshalova, V. V. Kinetics Catal. 2007, 48, 632.
(12) For a review on catalytic direct condensations operated under
azeotropic conditions: Ishihara, K. Tetrahedron 2009, 65, 1085.
(13) Coulson, J. M.; Richardson, J. F. Coulson and Richardson’s
Chemical Engineering − Vol. 1, 6th ed.; Elsevier: Oxford, 2000; pp
381−570.
(14) (a) Nukiyama, S. J. Soc. Mech. Eng. (Jpn.) 1934, 37, 367.
(b) Dhir, V. K. Annu. Rev. Fluid Mech. 1998, 30, 365 and references
therein.
(15) On condensation the amount of heat released is the same as that
absorbed during vaporization. Assuming nwater ≪ nent in the mixture,
(ΔnentΔHvap
)
ent
Qboil
=
VΔt
where Δnent is the number of moles of condensed entrainer collected
over time Δt.
(16) Green, D. W.; Perry, R. H. Perry’s Chemical Engineers’ Handbook,
8th ed.; McGraw-Hill: New York, 2008.
(17) Cossy, J.; Pale-Grosdemange, C. Tetrahedron Lett. 1989, 30,
2771.
(18) Agwada, V. J. Chem. Eng. Data 1982, 27, 481.
(19) Burnell-Curty, C.; Roskamp, E. J. Tetrahedron Lett. 1993, 34,
5193.
(20) Kim, S.; Kim, S. S. J. Chem. Soc., Chem. Commun. 1986, 9, 719.
(21) Srinivas, K. V. N. S.; Das, B. J. Org. Chem. 2003, 68, 1165.
787
dx.doi.org/10.1021/op300058f | Org. Process Res. Dev. 2012, 16, 781−787