Solving the clogging problem: Precipitate-forming reactions in flow

Article abstract of DOI:10.1002/anie.200503925

Solids go with the flow: A monodisperse flow in a microreactor provides an efficient method for keeping solid products away from channel walls. The use of a carrier phase, such as mineral oil, hexane, or toluene, enables solids to be synthesized without clogging of the reactor channels. Further injection points can be added to the microreactor to perform multistep syntheses (see picture). (Figure Presented)

Full text of DOI:10.1002/anie.200503925

Angewandte
Chemie
DOI: 10.1002/anie.200503925
Communications
A major drawback to the synthesis of solid products within micro-
reactors is the clogging of the reactor channels. In their communica-
tion on the following pages, D.T. McQuade and co-workers report a
solution to this problem by using a monodisperse droplet flow to
isolate the solid particles from the walls of the reactor tubing.
1544
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 1544 – 1548

Angewandte
Chemie
Microreactors
channels that occurs upon precipitate formation.[9] The
handling and processing of solids make up a significant
proportion of laboratory and industrial processes, and current
microreactor technology is not yet ready to handle these
syntheses efficiently.[10] There has recently been an increased
effort to deal with this problem, although many industrial
solutions come at a large expense. A less expensive approach
incorporates a periodic purging step to flush out solids that
have formed on the channel walls.[11] However, this method is
not effective if significant solid buildup occurs before purging.
A more attractive alternative involves performing reactions
in droplets that travel through the microreactor channels
inside a carrier phase.[12]
Fluid fields generated in microfluidic devices can control
reagent mixing[13] and allow the formation of an emulsion
upon the collision of two immiscible liquids. We have recently
reported a simple microfluidic device that can replicate these
flow phenomena.[14] Our device is composed of syringe
pumps, syringes, needles, and ordinary laboratory tubing—
all of which are relatively inexpensive and commercially
available (Figure 1). Development of our microreactor to
DOI: 10.1002/anie.200503925
Solving the Clogging Problem: Precipitate-
Forming Reactions in Flow**
Sarah L. Poe, Meredith A. Cummings, Michael P. Haaf,
and D. Tyler McQuade*
Techniques developed in recent decades have done little to
change the fundamental processes of chemistry. Reaction
vessels of a century ago continue to be the standard reactors
of today. Recently, however, increasing attention has been
paid to chemical reactions performed in microreactors.[1]
Reactions are performed in these microfluidic devices by
flowing reactants through channels that generally range in
size from 10 to 500 mm. On account of its small proportions, a
flat microchannel with a width of 100 mm has a specific
surface-area-to-volume ratio that is 200 times larger than that
of a 100-mL flask and over 3000 times larger than a tank that
occupies a cubic meter.[2] This increased surface-area-to-
volume ratio allows for better molecular-diffusion and heat-
transfer properties, which allow faster and more-selective
chemical reactions.[3]
From an industrial standpoint, microreactors are advanta-
geous because they eliminate the need to scale up a reaction.
Whereas in traditional process chemistry bench-top syntheses
must be redesigned for industrial compatibility, a method
known as numbering-up involves the addition of micro-
reactors to achieve the desired throughput.[4] As every reactor
is identical to the pilot reactor, there is no need to change
dimensions or conditions. Other advantages include increased
safety,[5] lower costs, and more environmentally friendly
chemistry owing to efficient reactions that require less
solvent.
Despite the numerous advantages of microreactors, they
are not without their drawbacks.[6] Researchers in an aca-
demic setting have been slow to embrace these systems
because of their cost and inflexibility.[7] The manufacture of a
single microreactor can be a very time- and cost-intensive
process, and once a microreactor has been developed, there is
rarely any opportunity to make variations to the device.[8]
Another commonly cited concern is the clogging of the
Figure 1. Basic design of our microfluidic reactor. The top-left syringe
pump contains the carrier phase, the right pump contains the first
disperse phase, and the bottom-left pump contains the second
disperse phase. Reagents are injected through a 30-gauge blunt-edge
needle (see inset).
facilitate chemical syntheses would potentially ameliorate
some of the problems still plaguing microreactors, namely
their cost and channel clogging. By utilizing disperse-phase
droplets as individual reactors, we can confine the solid
products to these droplets, thus keeping them away from the
tubing walls and avoiding clogged channels. Herein, we
present the results of the first chemical syntheses performed
in our microreactor and show that our device is practical and
efficient for the production of solids in a microfluidic
device.[15]
By nature of its design, our microfluidic device is versatile,
bearing the essential features of the reactor illustrated in
Figure 1. Additional fluid junctions may be added as needed
simply by inserting a needle anywhere along the tubing.
To demonstrate the suitability of our device for the
synthesis of solid particles, we chose a simple system in which
aqueous reagents combine to form a solid precipitate. Having
successfully demonstrated the ability to synthesize solids by
[*] S. L. Poe, Prof. D. T. McQuade
Department of Chemistry andChemical Biology
Cornell University
Ithaca, NY 14853 (USA)
Fax: (+1)607-255-4137
E-mail: dtm25@cornell.edu
M. A. Cummings, Prof. M. P. Haaf
Chemistry Department
Center for Natural Sciences
Ithaca College
Ithaca, NY 14850 (USA)
[**] This work was supportedby NSF Sensors (CTS-0329899), ARO
MAP-MURI, Cornell Center for Materials Science, andthe Beckman
Foundation.
Angew. Chem. Int. Ed. 2006, 45, 1544 –1548
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1545

Communications
interfacial polymerization in a two-flow system,[14] we took
because the product stains the poly(vinyl chloride) (PVC)
tubing upon contact, which provides an easy means of
determining the effectiveness of our method for isolating
the solid from the channel walls. We found that a mineral oil
flow rate of 3 mLminÀ1 produces sufficiently small droplets
that are confined to the center of the tubing. Figure 3
illustrates the differences observed when indigo is synthesized
in the presence and absence of a carrier phase.
advantage of the versatility of our device and added a third
fluid flow. In each case, an inert carrier fluid was employed as
the continuous phase, and disperse-phase reagents were
injected into the tubing through separate syringe pumps
located downstream from the carrier-phase source. Reagent
mixing occurred in one of two ways. If both disperse phases
are immiscible with the carrier phase, the mixing of the
reagents is initiated by the collision of two different reagent-
phase droplets (Figure 2a). This type of mixing was observed
Figure 2. Reagents are mixed inside the tubing by droplet–droplet
collision (a) or infusion followedby diffusion (b). Mixing can be
enhanced by chaotic advection induced by passing the reaction stream
through winding tubing (c).
Figure 3. Demonstration of the effectiveness of the carrier phase in the
formation of solids. Comparison of the tubing during (left tube) and
after (right tube) the synthesis of indigo in the presence (a) and
absence (b) of a mineral oil carrier phase.
when mineral oil was used as the carrier phase. On the other
hand, reagents that are miscible with the carrier phase are
injected coaxially as the final reagent. In this case, the mixing
is caused both by infusion of the second reagent into the first
and by diffusion from the carrier phase into the disperse
phase (Figure 2b). In both of these cases, chaotic advection
could be induced by passing the fluid stream through winding
tubing (Figure 2c).[16] We used 30-gauge (0.15mm i.d.) blunt-
edge needles for reagent introduction to obtain spherical
droplets (Figure 1, inset); beveled needles do not result in
clean snap-off of the droplets.
The reagent collision shown in Figure 2a and b induces
some mixing of the reagents. It has been shown that a further
enhancement is observed when the droplets are flowed
through a winding channel, which causes mixing by chaotic
advection.[16] We observed a qualitatively similar phenom-
enon when we wound our tubing through a series of parallel
horizontal bars. When the indigo synthesis was performed
inside this tubing, the indigo formation (observed by a color
change) occurred more rapidly than it did in straight tubing.
Characterization was not performed for this reaction due to
well-established purification issues.[18]
A number of research groups have employed fluorinated
solvents as carrier phases for microfluidic processes to
minimize the possibility of side reactions.[17] These solvents,
however, can be expensive and are rarely used as common
laboratory reagents. Instead, we chose mineral oil, hexane
(mixture of isomers), and toluene as relatively inert and
readily available carrier phases.
The first reaction we performed was the synthesis of
indigo (1), which involves a base-catalyzed aldol condensa-
tion between acetone and 2-nitrobenzaldehyde (2;
Scheme 1). The synthesis of this dye was appealing not only
because it results in the precipitation of a solid, but also
This method of producing solids in microreactors by using
a monodisperse droplet flow can be extended to carrier
phases other than mineral oil. The reaction of glyoxal (3) with
cyclohexylamine (4) results in the precipitation of N,N’-
dicyclohexylethylenediimine (5; Scheme 2). When mineral oil
is used as the carrier phase, the droplet–droplet phenomenon
shown in Figure 2a is observed. However, as 5 is soluble in
mineral oil, product recovery is difficult. In contrast, the use
of hexane (mixture of isomers) as the carrier phase allows
both the formation of solid in the monodisperse flow as well
as easier extraction of the solid. Owing to the decreased
viscosity and density of the hexane carrier phase relative to
mineral oil, higher flow rates are required to achieve the
Scheme 1. Synthesis of indigo (1).
Scheme 2. Synthesis of N,N’-dicyclohexylethylenediimine (5).
1546
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 1544 –1548

Angewandte
Chemie
desired flow type. We found that a flow rate of at least
12 mLminÀ1 for the hexane phase yields desirable conditions
for the formation of solid without channel blockage, although
a rate as low as 5mLmin À1 acts as an efficient purging system
by keeping the channel walls free of solids.
ease of use, the widespread availability of many of its
components, and its versatility provide further benefits.
Future work with our microfluidic device includes temper-
ature-controlled experiments as well as multistep syntheses in
a single device.
The final microfluidic reaction we studied was the
conversion of 4-chlorobenzoyl chloride (6) and methylamine
(7) into 4-chloro-N-methylbenzamide (8; Scheme 3). This Experimental Section
1: Mineral oil (15mL, 3 mLmin À1) was used as the carrier phase in
0.0625inch (1.59 mm) internal diameter (i.d.) PVC tubing. NaOH
(1m in water, 3 mL, 0.6 mLminÀ1) was injected into the center of the
carrier phase. 2 (0.66m in acetone, 3 mL, 0.6 mLminÀ1) was intro-
duced into the tubing further downstream. The pumps were allowed
to run for 5minutes while the product was collected over an ice–water
bath.
5: Hexane (mixture of isomers, 60 mL, 6 mLminÀ1) was used as
the carrier phase in 0.066 inch (1.68 mm) i.d. polyethylene (PE)
tubing. 3 (0.40m in water, 12 mL, 1.2 mLminÀ1) was injected into the
Scheme 3. Synthesis of 4-chloro-N-methylbenzamide (8).
center of the carrier phase. 4 (4.368m in water, 2.4 mL, 0.24 mLminÀ1
)
was introduced into the tubing further downstream. The pumps were
allowed to run for 10 minutes while the product was collected at room
temperature. Evaporation of the solvent yielded a white solid.
8: Toluene (70 mL, 7 mLminÀ1) was used as the carrier phase in
0.066 inch (1.68 mm) i.d. PE tubing. 7 (1.44m in water, 3 mL,
0.3 mLminÀ1) was injected into the center of the carrier phase. 6
(1.0 mL, 0.1 mLminÀ1) was introduced into the tubing further down-
stream. The pumps were allowed to run for 10 minutes while the
product was collected at room temperature. Evaporation of the
solvent and recrystallization from MeOH/H2O afforded needles of
white solid.
highly exothermic reaction was performed in our micro-
reactor with no safety concerns. The small dimensions of the
device not only mitigated the violence with which the reaction
took place, but they have also increased the yield of 8
(Table 1). The use of toluene as the carrier phase for this
reaction further demonstrates the versatility of our system.
Table 1: Synthesis of solids in our microfluidic device.
Product
System
Yield [%][a]
Purity [%][b]
STY[c]
Rel. STY
Received: November 5, 2005
Published online: January 13, 2006
5
5
8
8
flow
batch
flow
97.0
99.4
87.6
76.9
94.5
96.6
>99
>99
11.44
0.4478
9.150
25.55
1.00
20.43
1.59
batch
0.7140
Keywords: aldol reaction · flow reactions · microreactors ·
.
synthesis design
[a] Crude yields are reported for 5; purifiedyields are reportedfor
8.
[b] The purities were determined from 13C satellites in the 1H NMR
spectra. The purities of the crude product are reported for 5; the purities
of the purifiedproduct are reportedfor
8. [c] Space–time yields are
reportedin molm À3 minÀ1
.
[1] a) T. Schwalbe, A. Kursawe, J. Sommer, Chem. Eng. Technol.
2005, 28, 408; b) T. Wu, Y. Mei, J. T. Cabral, C. Xu, K. L. Beers, J.
Am. Chem. Soc. 2004, 126, 9880; P. Watts, S. J. Haswell, Chem.
Eng. Technol. 2005, 28, 290; c) E. Comer, M. G. Organ, J. Am.
Chem. Soc. 2005, 127, 8160; d) X. Zhang, S. Stefanick, F. J.
Villani, Org. Process Res. Dev. 2004, 8, 455; e) J. Yoshida, Chem.
Commun. 2005, 4509.
[2] S. Taghavi-Moghadam, A. Kleemann, K. G. Golbig, K, Org.
Process Res. Dev. 2001, 5, 652.
[3] a) T. Bayer, K. Himmler, Chem. Eng. Technol. 2005, 28, 285; b) J.
Yoshida, A. Nagaki, T. Iwasaki, S. Suga, Chem. Eng. Technol.
2005, 28, 259.
[4] W. Ehrfeld, V. Hessel, H. Lowe, Microreactors: New Technology
for Modern Chemistry, Wiley-VCH, Weinheim, 2000, p. 9.
[5] R. D. Chambers, M. A. Fox, D. Holling, T. Nakano, T. Okazoe,
G. Sanford, Chem. Eng. Technol. 2005, 28, 344.
[6] a) E. R. Delsman, M. H. J. M. de Croon, G. D. Elzinga, P. D.
Cobden, G. J. Kramer, J. C. Schouten, Chem. Eng. Technol. 2005,
28, 367; b) Y. Kikutani, T. Kitamori, Macromol. Rapid Commun.
2004, 25, 158.
[7] K. Shah, W. C. Shin, R. S. Besser, Sens. Actuators B 2004, 157.
[8] R. D. Chambers, M. A. Fox, D. Holling, T. Nakano, T. Okazoe,
G. Sanford, Lab Chip 2005, 5, 191.
[9] C. Boswell, Chem. Mark. Rep. 2004, 266(11), 8.
[10] D. M. Roberge, L. Ducry, N. Bieler, P. Cretton, B. Zimmerman,
Chem. Eng. Technol. 2005, 28, 318.
Batch syntheses were carried out as a control by stirring
the reagents for the same amount of time as the microreactor
experiments were allowed to run. As indicated in Table 1, the
yields for the microreactor products are comparable to—if
not greater than—those for the batch reactions. Also, space–
time yields (STY) for the solids formed by microscale flow
were much higher than those for solids formed in the
macroscale batch reactions. High yields were obtained in
the microfluidic system even when one reagent was miscible
with the carrier phase. Although we expected lower yields as a
result of the miscible reactant flowing into both the carrier
and reactant phases, this was not observed, which suggests
that diffusion is fast enough in our system to ensure reagent
mixing.
In summary, we have reported a practical method for
producing solids in microreactors. As demonstrated by the
indigo synthesis, by performing these reactions in a mono-
disperse droplet flow, the solid particles are effectively
isolated from the walls of the tubing. Our device not only
allows the practical synthesis of solids in microfluidic devices,
it also retains the advantages of traditional microreactors. Its
[11] A. M. Thayer, Chem. Eng. News 2005, 83(22), 43.
Angew. Chem. Int. Ed. 2006, 45, 1544 –1548
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1547

Communications
[12] I. Shestopalov, J. D. Tice, R. F. Ismagilov, Lab Chip 2004, 4, 316.
[13] B. Zheng, J. D. Tice, R. F. Ismagilov, Anal. Chem. 2004, 76, 4977.
[14] E. Quevedo, J. Steinbacher, D. T. McQuade, J. Am. Chem. Soc.
2005, 127, 10498.
[15] Z. Chang, G. Liu, Y. Tian, Z. Zhang, Mater. Lett. 2004 58, 522.
[16] D. L. Chen, C. J. Gerdts, R. F. Ismagilov, J. Am. Chem. Soc. 2005,
127, 9672.
[17] H. Song, J. D. Tice, R. F. Ismagilov, Angew. Chem. 2003, 115,
792; Angew. Chem. Int. Ed. 2003, 42, 768.
[18] R. Kohlhaupt, U. Bergmann, U.S. Patent No. 5,424,453, 1995.
1548
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 1544 –1548