Addition of secondary amines to α,β-unsaturated carbonyl compounds and nitrites by using microstructured reactors

Article abstract of DOI:10.1021/op0501949

Several additions of amines to α,β-unsaturated carbonyl compounds (Michael additions) were performed in a continuous-flow microstructured reactor rig and compared to the respective batch reaction. Dimethylamine/diethylamine/piperidine and acrylic acid ethyl ester/acrylonitrile were employed as two sets of reactants, giving six reactions. Some of these reactions are highly exothermal. Using the traditional batch procedure the olefin must be added quite slowly to the diluted amine to ensure temperature control and safe operation; especially this is necessary for the addition of dimethylamine (40 mass % aqueous solution) to acrylonitrile. Good yields (>85%) are achieved in this way; however, processing time is very long (17-25 h). To reveal the intrinsic kinetic potential and thus to accelerate these reactions, the reactants were mixed in a continuous-flow microstructured reactor rig which allows rapid mixing and efficient removal of the reaction heat. In this way, reaction time was decreased to a few seconds up to about half an hour, which is a change by 2 orders of magnitude. While the yields achieved with the continuous-flow microstructured reactor rig matched those for the batch procedure, the space-time yields for the microflow processing are much higher, in the best case by a factor of about 650.

Full text of DOI:10.1021/op0501949

Organic Process Research & Development 2006, 10, 1144−1152
Addition of Secondary Amines to r,â-Unsaturated Carbonyl Compounds and
Nitriles by Using Microstructured Reactors
H. Lo¨we,*,†,§ V. Hessel,‡,§ P. Lo¨b,§ and S. Hubbard
Johannes Gutenberg UniVersity Mainz, Institute of Organic Chemistry, Duisbergweg 10-14, 55128 Mainz, Germany,
EindhoVen UniVersity of Technology, Chemical Engineering and Chemistry, 5600 MB EindhoVen, PO Box 513, STW 1,
35, The Netherlands, IMM Institut fu¨r Mikrotechnik Mainz GmbH, Carl Zeiss Strasse 18-20, 55129 Mainz, Germany, and
Europa Fachhochschule Fresenius, Department of Chemistry and Biology, Limburger Strasse 2, 65510 Idstein, Germany
Abstract:
According to the overall high reactivities, amines 1 easily
add to the R,â-unsaturated carbonyl compounds 2 at ambient
temperature even without any catalyst.
Several additions of amines to r,â-unsaturated carbonyl com-
pounds (Michael additions) were performed in a continuous-
flow microstructured reactor rig and compared to the respective
batch reaction. Dimethylamine/diethylamine/piperidine and
acrylic acid ethyl ester/acrylonitrile were employed as two sets
of reactants, giving six reactions. Some of these reactions are
highly exothermal. Using the traditional batch procedure the
olefin must be added quite slowly to the diluted amine to ensure
temperature control and safe operation; especially this is
necessary for the addition of dimethylamine (40 mass %
aqueous solution) to acrylonitrile. Good yields (>85%) are
achieved in this way; however, processing time is very long (17-
25 h). To reveal the intrinsic kinetic potential and thus to
accelerate these reactions, the reactants were mixed in a
continuous-flow microstructured reactor rig which allows rapid
mixing and efficient removal of the reaction heat. In this way,
reaction time was decreased to a few seconds up to about half
an hour, which is a change by 2 orders of magnitude. While
the yields achieved with the continuous-flow microstructured
reactor rig matched those for the batch procedure, the space-
time yields for the microflow processing are much higher, in
the best case by a factor of about 650.
This reaction is reversible at high temperatures above 200
°C. Therefore, the reverse pathway, the thermal cleavage of
a secondary amine from 3 is also exploited as common
procedure to form R,â-unsaturated carbonyl compounds.
With the use of a batch procedure these reactions give
good yields, in some cases more than 85%; however, the
protocols recommend long reaction times of up to 24 h. This
long reaction time is for most reactions not caused by the
intrinsic kinetics but by the high exothermicity that causes
selectivity issues. In order to enable safe removal of the heat
created by the reaction, it is necessary to add quite slowly
the R,â-unsaturated compound to the amine.2
These characteristics indicate that the use of a continuous
flow rig with microstructured reactors may accelerate the
performance of the reaction without losing thermal control
as, obviously, rapid mixing of the reactants as well as
efficient heat removal have to be ensured. Such capability
of this novel approach has been demonstrated many times,
even under demanding processing conditions.3-8
1. Introduction
Additions of amines to R,â-unsaturated carbonyl com-
pounds such as acrylic acids, acrylic acid esters, or acrylo-
nitrile have been well investigated and are thus one of the
constituting organic synthesis pathways for modern chemistry
of today.1 In this variant of the Michael addition, the CdC
bond reacts with nucleophilic compounds, e.g., secondary
amines, due to the transfer of electrons from the carbonyl
group or from carbonyl analogues such as nitriles. As a result,
the nucleophilic amine is added on the CdC double bond.
The activity found for the R,â-unsaturated nitriles is generally
higher than for the R,â-unsaturated acrylic acid esters.
(2) Autorenkollektiv, Organikum, 15th ed.; VEB Deutscher Verlag fu¨r
Grundstoffindustrie: Berlin, 1981.
(3) Chang, C.-H.; Liu, S.-H.; Doneanu, A.; Tennico, Y.; Rundel, J. T.; Remcho,
V. T.; Blackwell, E.; Tseng, T.; Paul, B. K. Progress towards high-
throughput dendrimer synthesis; AIChE Spring National Meeting, Atlanta,
GA, U.S.A., 10-14 April, 2005.
(4) Liu, S.; Chang, C.-H.; Paul, B. K.; Remcho, V. T. High-rate synthesis of
lower generation dendrimers with a continuous flow microreactor. In
preparation.
(5) Schwalbe, T.; Autze, V.; Hohmann, M.; Stirner, W. Novel Innovation
Systems for a Cellular Approach to Continuous Process Chemistry from
Discovery to Market. Org. Process Res. DeV. 2004, 8, 440-454.
(6) Pennemann, H.; Watts, P.; Haswell, S.; Hessel, V.; Lo¨we, H. Benchmarking
of Microreactor Applications. Org. Process Res. DeV. 2004, 8, 422-439.
(7) Pennemann, H.; Hessel, V.; Lo¨we, H.; Forster, S.; Kinkel, J. Improvement
of Dye Properties of the Azo Pigment Yellow 12 Using a Micromixer-
Based Process. Org. Process Res. DeV. 2005, 9, 188-192.
(8) Hessel, V.; Hofmann, C.; Lo¨we, H.; Meudt, A.; Scherer, S.; Scho¨nfeld, F.;
Werner, B. Selectivity Gains and Energy Savings for the Industrial Phenyl
Boronic Acid Process Using Micromixer/Tubular Reactors. Org. Process
Res. DeV. 2004, 8, 511-523.
* To whom correspondence should be addressed. E-mail: loewe@uni-
mainz.de.
† Johannes Gutenberg University Mainz.
‡ Eindhoven University of Technology, Chemical Engineering and Chemistry.
§ IMM Institut fu¨r Mikrotechnik Mainz GmbH.
Europa Fachhochschule Fresenius.
(1) Ranu, B. C.; Dey, S. S.; Hajra, A. Solvent-free, catalyst free Michael-type
addition of amines to electron-deficient alkenes. ARKIVOC 2002, 2, 76-
81.
1144
•
Vol. 10, No. 6, 2006 / Organic Process Research & Development
10.1021/op0501949 CCC: $33.50 © 2006 American Chemical Society
Published on Web 10/10/2006

Table 1. Additions of amines to r,â-unsaturated carbonyl compoundsa
a 1a: pure substance or dissolved as 40 mass % water solution. 1b: pure substance or dissolved as 40 mass % ethanol solution. 1c: pure substance.
In literature, there are several reports on the use of
microstructured reactors for Michael additions (using enolates
of 1,3-dicarbonyl compounds, instead of the amines used
here) to couple to R,â-unsaturated alkene or alkyne-based
derivatives. Among others, the addition of 2,4-pentanedione
enolate to ethyl propiolate and the addition of 2,4-pentanedi-
one enolate to methyl vinyl ketone was investigated.9
These Michael additions were made in smart, glass chip-
type devices under electroosmotic flow conditions. Thus,
flow rates were very low, in the mL/h range, and the
(9) Wiles, C.; Watts, P.; Haswell, S. J.; Pombo-Villar, E. 1,4-Addition of
investigations were analytically oriented. For the first reac-
tion, a conversion of 56% was achieved (batch synthesis:
enolates to R,â-unsaturated ketones within a micro reactor. Lab Chip 2002,
2, 62-64.
Vol. 10, No. 6, 2006 / Organic Process Research & Development
•
1145

89%).9 Using stopped-flow technique (2.5 s with field
applied; 5.0 s with field turned off) to enhance mixing, a
conversion of 95% was determined. In a further study,
substrates with diketone moieties of known different reac-
tivities, such as 2,4-pentadione, benzoyl acetone, and diethyl
malonate, were processed, each with the same acceptor ethyl
propiolate. Also, a reaction with the less alkynic Michael
acceptor methyl vinyl ketone was carried out. The conver-
sions observed followed the sequence of reactivity found by
carrying out batch experiments in advance. For example, only
15% conversion was found for the less reactive reagent
benzoyl acetone in the microreactor experiment, while a 56%
conversion was determined when using the more reactive
2,4-pentadione (batch syntheses: 78% and 89%, respec-
tively).9 Using stopped-flow to enhance mixing, the conver-
sions for both syntheses were increased to 34% and 95%,
respectively. Using a further improved stopped-flow tech-
nique (5.0 s with field applied; 10.0 s with field turned off),
conversion could be further enhanced to 100% for the
benzoyl acetone case. For the two other substrates, diethyl
malonate and methyl vinyl ketone, similar trends were
observed.
In terms of processing time, the microreactor synthesis
turned out to be much faster than the batch synthesis. To
achieve comparable amounts of conversion, 24 h operation
time was needed for batch synthesis, whereas microreactor
operation needed only about 20 min.9
One application for the type of Michael addition using
amines reported here is the synthesis of dendrimers, e.g.,
built up from polyamidoamine (PAMAM). Ethylene diamine
reacts with methyl acrylate to form the first half-generation,
and from there, further branching occurs. By using a
continuous-flow microstructured reactor rig, the reaction time
could be shortened from 96 h (batch) to 13 s (microstructured
reactor) at 99% yield without any detectable side products.3
Table 2. Comparison of yields for the batch reactions of a
stirred 40 mass % aqueous solution of dimethylamine with
acrylonitrile (R1) and with acrylic acid ethyl ester (R4)
reaction R1
reaction R4
yield (GC)
literature
95.6%
96%20
92.5%
87%21
Table 3. Comparison of yields for the batch reactions of a
stirred 40 mass % aqueous solution of diethyl amine/
piperidine with acrylonitrile (R1/R3) and with acrylic acid
ethyl ester (R5/R6)
reaction R2
reaction R3
reaction R5
reaction R6
yield (GC)
literature
97.1%
85%2
99.7%
90%2
99.1%
85%22
99.8%
-
constant at 30 °C. After 16 h/24 h, respectively, the reaction
mixture was analyzed by GC (Table 2).
Reactions R2, R3, R5, and R6. An amount of 1.0 mol
acrylonitrile (2a)/acrylic acid ethyl ester (2b) was added
dropwise under stirring to 1.1 mol amine (1b-c) dissolved
in 150 mL of ethanol. The process parameter and the
analytical procedure were the same as those used for
reactions R1 and R4 (see Table 3). Higher yields were found
for the performed experiments compared to the literature
values, which can be attributed to the fact that conversion
was determined by GC directly from the raw reaction mixture
without purification steps that lead to loss of yield (Table
3).
2.4. Microstructured Reactors. A slit-type interdigital
micromixer (SIMM-V2/Ag40, IMM Institut fu¨r Mikrotechnik
Mainz GmbH, Germany) was chosen as a microreactor. This
device is made of stainless steel and reaches flow rates of
about 2 L h-1 at a pressure drop of 2.8 bar for water (which
is about the limit for the type of pumps used). The mixer is
suited for pressures up to 100 bar (using stronger pumps,
e.g., HPLC) and temperatures of 200 °C. Inside the mixer,
an inlay of 11 mm × 7.5 mm × 3.6 mm size provides
interdigitally grouped microstructured channels (width 40 µm
and depth 300 µm; material: silver on copper) which ensure
the formation of alternate fluid streams. These streams are
guided through a tiny confinement of slit shape where they
are considerably compressed. In this way, the SIMM-V2/
Ag40 mixer thus relies on multilamination and geometric
focusing to speed up diffusion between lamellae being only
a few 10 µm thick (Figure 1). Segregation indices in the
range of about 10-3 to 10-5 are found, depending on the
flow rate.10-13 By analytical calculations, a mixing time can
be derived from this which ranges 0.8-5 ms.
2. Experimental Section
2.1. Materials. As secondary amines, dimethylamine
(“dissolved”, 40 mass % in water, Aldrich), dimethylamine
(“pure”, gas, Air Liquide), diethylamine (purum g 99.0%,
Fluka), piperidine (purum g 99.0%, Fluka) were considered
in this investigation. As R,â-unsaturated compounds, acrylo-
nitrile (purum g 99.0%, Aldrich) and acrylic acid ethyl ester
(purum g 99.0%, Aldrich) were taken.
2.2. Reactions. Table 1.
2.3. Batch Processing. For reasons of comparison, most
of the reactions were performed additionally as batch process.
A common Michael addition protocol was used,2 with some
minor variations as indicated below. All batch processes were
carried out in a 500-mL three-necked flask, equipped with a
magnetic stirrer, reflux cooler, thermometer, and a tap funnel.
The reactions were performed by slowly dripping the acrylic
compound into the dissolved and diluted amines. To ensure
thermal control of the reactions the reaction mixture was
cooled by an external water bath.
(10) Hessel, V.; Lo¨we, H.; Mu¨ller, A.; Kolb, G. Chemical Micro Process
Engineering - Processing and Plants; Wiley-VCH: Weinheim, 2005.
(11) Hessel, V.; Hardt, S.; Lo¨we, H.; Scho¨nfeld, F. Laminar mixing in different
interdigital micromixers - Part I: Experimental characterization. AIChE J.
2003, 49, 566-577.
(12) Hardt, S.; Scho¨nfeld, F. Laminar mixing in different interdigital micromixers
- Part 2: Numerical simulations. AIChE J. 2003, 49, 578-584.
(13) Lo¨b, P.; Drese, K. S.; Hessel, V.; Hardt, S.; Hofmann, C.; Lo¨we, H.; Schenk,
R.; Scho¨nfeld, F.; Werner, B. Steering of liquid mixing speed in interdigital
micromixers - from very fast to deliberately slow mixing. Chem. Eng.
Technol. 2003, 27, 340-345.
Reactions R1 and R4. An amount of 1.0 mol acrylonitrile
(2a)/acrylic acid ethyl ester (2b) was added drop by drop
under vigorous stirring to a 40 mass % solution of 1.1 mol
dimethylamine (1a) in water. The temperature was kept
1146
•
Vol. 10, No. 6, 2006 / Organic Process Research & Development

Figure 1. Mixing principle of slit-interdigital mixer (left) and image of the used device SIMM-V2 (right)
Figure 2. Microreactor heat exchanger: stainless steel internal platelets (left), diffusion bonded device (right).
Figure 3. Flow scheme of the experimental setup for performing a chemical reaction continuously in a microstructured reactor.
B1, B2, supply vessels; P1, P2, syringe pumps; W1, W2, preheating capillaries; R1, microstructured reactor/mixer; B3, tubular
reactor; W4, water bath; V1, valve; P001, P002, P003, pressure gages; W3, cooler/heat exchanger; B4, storage vessel.19
As a microheat exchanger for quenching the reaction
mixture, the device CRMH (IMM Institut fu¨r Mikrotechnik
Mainz GmbH, Germany) was used. It consists of a stack of
stainless steel platelets which guide the coolant and the
reaction mixture in cross-flow mode. These platelets have a
size of 26 mm × 26 mm × 0.2 mm and consist of nine
channels with a length of 12 mm and a width of 700 µm.
The internal fluid hold-up of 0.015 mL is very low (Figure
2). With an overall heat exchange surface area of 12,800 m2
m-3 a heat transfer coefficient of 4000 W(m2 K)-1 for water
at a flow rate of 0.5 L min-1 was determined.14-16
2.5. Experimental Rig. A micromixer/tube reactor rig
for continuous flow operation with tailored process flow was
constructed, based on off-the-shelf components, as given
above. Figure 3 shows the flow-scheme of the experimental
setup. Two supply vessels (B1, B2) store the reactants
(amines and R,â-unsaturated carbonyl compounds) which are
fed by two syringe pumps (P1, P2) (PN 1610 syringe dosing
systems, 2.5 mL syringe volume; Postnova Analytics,
Landsberg, Germany) in separate capillaries (W1, W2), with
a volume of 900 µL, for means of preheating. These hot
(15) Ja¨hnisch, K.; Hessel, V.; Lo¨we, H.; Baerns, M. Chemistry in Microstructured
Reactors. Angew. Chem., Int. Ed. 2004, 43, 406-446.
(16) Ehrfeld, W.; Hessel, V.; Lo¨we, H. Microreactors; Wiley-VCH: Weinheim,
2000.
(14) Hessel, V.; Lo¨we, H. Micro Chemical Engineering: components - plant
concepts - user acceptance: Part II. Chem. Eng. Technol. 2003, 26 (4),
391-408.
Vol. 10, No. 6, 2006 / Organic Process Research & Development
•
1147

Figure 4. Photographic image of the microreactor setup.
flows are then contacted in the mixer/reactor (R1), the
SIMM-V2/Ag40 micromixer. For prolongation of residence
time, a tube reactor was added as delay loop element. The
reaction is thereafter quenched by rapidly cooling the flow
via the heat exchanger CRMH (W3) which is connected to
an external thermostat. The reaction mixture is then stored
in a separate vessel (B4). The preheater capillaries (W1, W2),
the reactor (R1), and the tube reactor (B3) are immersed into
a thermostat (W4). For heating to higher temperatures, water
is replaced by oil. The pressure gauges P001 and P002 in
front of the mixer monitor possible blockage of the mixer
or of the tube reactor. The gauge P003 controls the pressure
of the system which can be manually adjusted by a needle
valve (V1).
aminopropionic acid ethyl ester (3e), all reactants and
products were commercially available. The commercial
samples were used for identification and quantification of
the reaction mixtures by using them as calibration substances
for the GC.
3-Diethyl aminopropionic acid ethyl ester (3e) was
synthesized in a batch process, purified by distillation, and
then used for GC calibration. Identification was performed
by using a GC/MS (GC: 6890 N, MS: 5973 NMSD
Quadrupole 69.9 eV, both Agilent Technologies) procedure.
A (5%) phenyl methoxysiloxane column was used for
separation. The mass spectrum of (3e) is given in Figure 5.
3. Results and Discussion
Two tube reactors of different geometry were used in the
experiments, a tube with 1/4′′ outer diameter/74.5 mL inner
volume and one with 1/8′′ outer diameter/9.8 mL inner
volume. All experiments were performed under pressure,
because the amines (dimethylamine, bp 7 °C, diethylamine,
bp 55 °C, and piperidine bp 104-106 °C) would otherwise
be or become gaseous under the temperatures employed. For
the experiments at not-too-high pressure, all parts of the rig
were connected by PTFE tubes. For the experiments using
liquefied dimethylamine at 20-30 bar, stainless steel tubes
had to be used. The liquified dimethylamine was fed from a
pressurized gas cylinder via an HPLC pump. In some
experiments the use of a tube section was not necessary; the
residence time then only amounts to the time it takes to pass
through the micromixer and the subsequent connectors (inner
volume approximately about 1 mL). In Figure 4 a photo-
graphic image shows the overall experimental setup as well
as a detailed view of the mixer/tubular reactor mounted in
the thermostat.
3.1. Development Target. The figure of merit for the
experiments using the microreactor was to find process
parameters yielding maximal space-time yields. The most
governing parameters were assumed to be temperature,
residence time (by total flow rate or by volume of the
residence time section), and the molar ratio of amine 1 to
R,â-unsaturated compound 2 (in the following, described
briefly as molar ratio). The goal in the first step was to
increase the reaction rate by increasing the temperature with
the two other parameters set large; this was expected to
promote conversion, since reliable kinetic data were not
available. Then, residence time and excess of one of the
reactants ought to be reduced as much as conversion is not
reduced. Pressure was applied to the system to maintain a
single-phase condition, but not with the intention of having
an additional parameter for process optimization.
3.2. Reaction Time. Reaction time was decreased by
orders of magnitude, from about 16-24 h for the batch
processes to, at best, 0.1 min by using microstructured
reactors for reaction R1, which adds dimethylamine (40 mass
% in water) to acrylonitrile. This process probably is even
faster, but this could not be detected with the experiments
performed due to limitations in flow rate. Even for the much
less reactive diethylamine (40 mass % ethanol), the addition
to acrylonitrile was completed within less than 30 min, still
2.6. Reaction Analysis. The reaction mixtures were
analyzed by a GC (Dani - GC 8610) with a Stabilwax
(Crossbound Carbowax-PEG) 30-m column. All samples
were injected at a temperature of 250 °C, and the column
heating was programmed as a ramp starting from 80 °C to
210 °C with a rate of 10 K min-1. As detector, a FID was
used which was operated at 300 °C. Apart from 3-diethyl
1148
•
Vol. 10, No. 6, 2006 / Organic Process Research & Development

Figure 5. Mass spectrum of 3-diethyl aminopropionic acid ethyl ester (3e).
Table 4. Reactivities of amines (1) and r,â-unsaturated carbonyl compounds (2) under similar conditionsa
conversion
acrylonitrile (2a)
[%]
conversion
acrylic acid ethyl ester (2b)
[%]
tres
(min)
tres
(min)
dimethyl amine (1a) 40 mass % in water
dimethyl amine (1a)
diethyl amine (1b) 40 mass % in ethanol
diethyl amine (1b)
98.7b
91.5 c
67.3
40.5
97.2
1.6
3.8
7.2
28.8
7.7
96.4
-
61.1
12.6
98.8
7.2
-
7.4
14.8
7.4
piperidine (1c)
a T ) 50 °C, P ) 3-4 bar, molar ratio (1) to (2) ) 1.1, volume of residence time section: 9.8 mL. Exceptions are indicated. b 60 °C. c 30 bar.
being sped up by more than 1 order of magnitude compared
to the batch process.
3.3. Amine Reactivity. In order to compare reactivities,
the yields of the different reactions studied are compiled in
Table 4 for similar reaction conditions (molar ratio 1.1; all
at 50 °C, with one exception; and at a pressure of 3-4 bar,
with the exception of the pure dimethylamine experiment).
Since the flow rates were not absolutely equal for reasons
of pump pulsation and hysteresis behavior of the pressure
regulation valve and even the same residence time section
of 9.8-mL volume was applied, differences in the reaction
times were given. These, however, are taken into account in
the following considerations.
The following sequence of reactivity of the pure amines
was found in the reaction with both R,â-unsaturated carbonyl
compounds:
Figure 6. Yield as a function of residence time for the reaction
of dimethylamine (9 pure, 20 °C, molar ratio 1.1, 30 bar; b 40
mass % aqueous solution, 80 °C, molar ratio 2.0, 3-4 bar) with
acrylonitrile in the microreactor.
mobility by entrainment into a ring structure. Dimethylamine
(pKs ) 10.732)17 is the strongest base and is sterically
hindered very little.
dimethylamine > piperidine >> diethylamine
Reaction times range from 0.1 min (minimal) for the
addition of dimethylamine (40 mass % in water) to acrylo-
nitrile (R1) (50 °C; molar ratio 1.1) for achieving a yield of
99% (Figure 6).
In turn, for pure diethylamine even after the long residence
time of 28.8 min, and even despite the high temperature of
90 °C (molar ratio of 1.1), only about half of that yield,
40.5%, was achieved (Figure 7). The reaction of pure
piperidine with acrylonitrile (R3) is in between, with 99%
yield approached after about 15 min (50 °C; molar ratio 1.1)
(Figure 8). By replacing pure diethylamine by diethylamine
dissolved in ethanol (40 mass % in ethanol), 30 min reaction
Dimethylamine is much more reactive than its homologue,
diethylamine, for both the dissolved and pure compounds.
Piperidine is less reactive than dimethylamine, but still
exceeds by far the diethylamine. This sequence follows the
basicity and steric accessibility of the amines, for example
diethylamine being the weakest base (pKs ) 10.489),17 and
additionally the nucleophilic attack is sterically hindered by
the two ethyl groups. In the case of piperidine (pKs )
11.123),17 the two alkyl substituents are restricted in their
(17) Lide, D. R. Handbook of Chemistry and Physics, 79th ed.; CRC Press
LLC: Boca Raton, Fl, 1998.
Vol. 10, No. 6, 2006 / Organic Process Research & Development
•
1149

amount of methyl amine (1d) impurities apart from the
reactant dimethylamine which can undergo two addition
reactions.
3.6. Residence Time. Residence time is expected to
increase yield up to a certain value. Indeed, basically all
curves show initially a strong increase of yield with residence
time which then levels off at longer residence times. The
exception hereto is the addition of dimethylamine (40 mass
% solution in water) to acrylonitrile (R1) which gives near
quantitative yields of 98-99% within the whole residence
time interval inspected (Figure 6). It is expected, however,
that this addition follows the same dependency as found for
the other reactions (R2-R6), but at much lower residence
times, which were hardly realizable even with the microre-
actor setup used. Evidence for this is provided by the less
reactive pure dimethylamine showing such trend. The
reduction of residence time from about 15.3 to 1.9 min leads
here to a certain decrease of yield from 98.4% to 95.6%
(Figure 6).
Figure 7. Yield as a function of residence time for the reaction
of diethylamine (9 pure, 90 °C, molar ratio 1.1; b 40 mass %
in ethanol, 50 °C, molar ratio 1.2) with acrylonitrile.
For the still quite reactive piperidine an increase of the
yield above 90% with residence time could be demonstrated
for addition to acrylonitrile and above 96% for addition to
the R,â-unsaturated carbonyl ester.
3.7. Molar Ratio. The molar ratio of amine to the
unsaturated addition compound was varied from 2-fold
excess (2.0) to stochiometric (1.0). Yields were reduced by
no more than 10%. Surprisingly, only for the small change
from a molar ratio of 1.0 to 1.1 were measurable effects
given, and during the increase of the ratio from 1.1 to 2.0,
the performance was constant. The largest impact was found
for reaction R3 (molar ratio, 1.0: 88% as compared to molar
ratio 1.1: 98%; 50 °C; 7.7 min) (Figure 10).
Also for reactions R2 and R6, a notable impact was given.
For reaction R2, decreasing the molar ratio to 1.0 decreased
yield down to 91.0% (90 °C; 14 min). For reaction R6, yields
changed by about only 5% for the same change in molar
ratio (1.0: 98.9% as compared to 1.1: 94.8%; 30 °C, 7.4
min). For reactions R1 and R4, the molar ratio had still less
impact (molar ratio 1.0, 98.0%; molar ratio 1.1: 97.2%; 30
°C; 3.6 min). For reaction R5 only at high molar ratio, i.e.
at large excess of diethylamine, were yields increased (molar
ratio 2.0: 97.9%; 90 °C; 27 min).
Thus, the sensitivity of the yield to the molar ratio seems
to correlate with the reactivity of the amine. Dimethylamine
is so reactive that all molecules are converted under
stochiometric conditions; an excess of reactant cannot be
helpful here anymore. The lower activities of the two other
amines result in sensitivity to excess reactant.
3.8. Solvent-Free with Pure Agents vs Diluted Process-
ing. For reaction R1, the yield is higher when using an
aqueous solution of dimethylamine as compared to pure
dimethylamine (liquefied at 30 bar) to be added to acryloni-
trile (20 °C, molar ratio of 1.1). The same holds for reaction
R2 with diethylamine added to acrylonitrile. The latter is
noteworthy because in the reaction R5 pure diethylamine
reacts very slowly with acrylic acid ethyl ester, as preliminary
experiments showed (not given here). An explanation of this
behavior can be given by the lack of acidic compounds which
Figure 8. Reaction of piperidine with acrylonitrile: yield as
function of residence time for 50 °C.
is needed to result in 90% yield (50 °C, molar ratio 1.2),
thus effectively doubling the reaction speed.
3.4. R,â-Unsaturated Carbonyl Compound Reactivity.
The reactivities of the two R,â-unsaturated carbonyl com-
pounds investigated, acrylonitrile and acrylic acid ethyl ester,
are similar for reactions with dimethylamine or piperidine
(Figure 9, right). If diethylamine is applied, however, the
reactivity turned out to be different, as follows.
acrylonitrile > acrylic acid ethyl ester
After 25 min reaction, the addition of diethylamine (40
mass % in ethanol) to acrylonitrile is beyond 96% yield
(molar ratio 1.2, 50 °C) (Figure 9, left), while it is just about
60% for the ester (molar ratio 1.1, 50 °C). Even if the
reaction temperature is increased to 90 °C, a yield of only
90% could be achieved.
3.5. Yields and Selectivity. Near-quantitative yields
(∼99%) were found for the addition of dimethylamine (40
mass % in water) to acrylonitrile (R1) (50 °C, molar ratio
1.1) (Figure 6) and for the reaction of pure piperidine with
acrylonitrile (R3) (50 °C, molar ratio 1.1). Within the time
frame of about 30 min,, a yield of only 41% was achieved
when using pure diethylamine for (R1) (90 °C, molar ratio
of 1.1). The yield with diethylamine (40 mass % in ethanol)
was at maximum 90% (50 °C, molar ratio 1.2) (Figure 7).
For all reactions investigated, R1-R6, and under all
process conditions, e.g., with either pure amine or dissolved,
no traces of side products were found (via GC analysis);
thus, it is assumed that conversion equals yield. An exception
hereto is reaction R4 showing only traces of a 2-fold addition
product (4) by GC/MS. This is explained by a very small
1150
•
Vol. 10, No. 6, 2006 / Organic Process Research & Development

Figure 9. (Left) Yield as function of residence time for the reaction of diethylamine (40 mass % in ethanol) with acrylonitrile (b,
molar ratio 1.2, 50 °C) in comparison with the reaction with acrylic acid ethyl ester (2; 90 °C, molar ratio 1:1.1) (Right) Yield as
function of residence time for the reaction of piperidine with acrylonitrile (b) (molar ratio 1.1, 50 °C) in comparison with the
reaction with acrylic acid ethyl ester (2, 50 °C, molar ratio 1.1).
Figure 10. Reaction of piperidine with acrylonitrile: depen-
dence of yield on molar ratio (temperature 50 °C, residence
time 7.7 min, 4 bar).
Figure 11. Reaction of diethylamine (b 40 mass % in ethanol,
molar ratio 1.2). Increase of yield as function of reaction
temperature.
can easily protonate the carbonyl compound to enhance
reactivity.18 Therefore, all later microreactor experiments for
R5 were carried out with diethylamine dissolved in ethanol
(40 mass % solution).
Similar trends are found for the other reactions (see Figure
12), albeit with different slopes. Only for the reaction of
piperidine with the acrylic acid ester is such dependency not
found, as almost full conversion is given for all temperatures
considered.
At still higher temperatures, thermodynamics are ruling,
and the start of the reverse reaction can be seen, e.g., as for
reaction R2. The yields dropped here considerably from
100% at 120 °C to 10.9% and 11.8% for temperatures of
150 °C and 200 °C, respectively (Figure 13).
3.10. Space-Time Yields. The productivity of the batch
and continuous processes in the microstructured reactor can
be compared on the basis of space-time yields, which
normalize the obtained yields (g or mol) with regard to
reactor volume (mL) and reaction time (h). In the case of
the batch process the sum of the volumes of the two reactants
was taken as reactor volume, while in the case of the
continuous working microreactor the volume of the residence
time section was used (as the micromixer volume can be
neglected).
3.9. Temperature. The effects of temperature are two-
fold, to kinetically increase reaction speed at lower temper-
atures for the less reactive amines diethylamine and piperi-
dine and to decompose product by thermodynamically
reversing the reaction at high temperatures. Reaction R1 is
always operated at its kinetic limits, i.e., to full conversion.
By a temperature of 50 °C and a pressure of 30 bar, the
dimethylamine is kept in liquid phase, and the yield decreases
from 97% (20 °C) to 92%, probably due to the fact that the
reverse reaction became more noticeable.
For reaction R2, yield rises nearly linearly from 80% to
almost 100% as the temperature is increased from 50 to 90
°C for kinetic reasons (Figure 11).
(18) Moghaddam, F. M.; Mohammadi, M.; Hosseinnia, A. Water promoted
Michael addition of secondary amines to R, â-unsaturated carbonyl
compounds under microwave irradiation. Synth. Commun. 2000, 30, 643-
650.
Throughout all reactions and all parameter variations the
space-time yield for the chemical processing in the microre-
actor is much higher than for the corresponding batch process
(Table 5), given the fact that even nonoptimal process
conditions are included in Table 5. At best, an improvement
of space-time yield by a factor of 652 was achieved.
(19) Hubbard, S. M. Addition von sekunda¨ren Aminen an R,â-ungesa¨ttigte
Carbonylverbindungen und Nitrile im kontinuierlich betriebenen Mikroreak-
tor. Diplomarbeit, Europa Fachhochschule Fresenius, University of Applied
Sciences, Idstein, Germany, 2005.
(20) Nazarov, I.; Shvekhgeimer, G. Zhur. Obshch. Khim. 1954, 24 ((Chem. Abstr.
1954, 49, 3034i)), 163-169.
(21) Fuson, R. J. Am. Chem. Soc. 1928, 50, 1444-1449.
(22) Adams, D. W. J. Chem. Soc. 1949, 144-155.
Vol. 10, No. 6, 2006 / Organic Process Research & Development
•
1151

Table 5. Comparison of space-time yields for batch processing and continuous-flow processing in microstructured reactors
R,â-unsaturated carbonyl compound
acrylonitrile (2a)
acrylic acid ethyl ester (2b)
space-time yield
space-time yield
[g (mL h)-1
]
[g (mL h)-1
]
ratio space-time yield
microreactor/batch
ratio space-time yield
microreactor/batch
batch
0.03
microreactor
17.60
batch
0.04
microreactor
0.30
dimethyl amine (1a)
40 mass % in water
dimethyl amine (1a)
diethyl amine (1b)
40 mass % in ethanol
diethyl amine (1b)
piperidine (1c)
652
8
-
10.10
2.24
-
-
-
-
-
-
0.02
102
2.50
-
0.03
0.6
6.1
-
235
0.02-
0.02
0.40
6.80
21
309
a The space-time yields for the microreactor refer to the corresponding conditions given in Table 4.
Figure 12. (Right) Temperature dependence of yield for the reaction of the diethylamine solution with acrylonitrile [(b) molar
ratio 1.2, residence time 14 min] and with acrylic acid ethyl ester [(2) residence time 27 min, molar ratio 1.1]. (Left) Temperature
dependence of yield for the reaction of piperidine with acrylonitrile [(b) molar ratio 1.1, residence time 7.7 min] in comparison
with the reaction with acrylic acid ethyl ester [(2) 7.4 min, molar ratio 1.1].
microreactor process based on a microstructured mixer/
tubular-reactor setup has been established, optimized, and
compared to the corresponding batch processes. Using a
batch procedure, good yields (>85%) could be achieved;
however, the process takes a very long time, up to 17 to 25
h.
The microreactor process accelerated the reaction by
orders of magnitude via ensuring fast mixing and efficient,
simultaneous removal of reaction heat under improved
thermal control. Yields of more than 90% could be achieved
in this way which is slightly better than the yields of the
batch process. Due to the speed-up of reaction by using
microstructured reactors, high production rates are achiev-
able, e.g., space-time yields are increased by a factor of up
to 650 compared to those of the batch process.
Figure 13. Decrease of yield for the reaction of diethylamine
with acrylonitrile at high temperatures.
4. Conclusions
For the reaction of secondary amines 1a-c with R,â-
unsaturated carbonyl compounds and nitriles 2a-b to tertiary
amines 3 (a variation of the Michael addition) a continuous
Received for review October 10, 2005.
OP0501949
1152
•
Vol. 10, No. 6, 2006 / Organic Process Research & Development