Rapid catalyst screening by a continuous-flow microreactor interfaced with ultra-high-pressure liquid chromatography

Article abstract of DOI:10.1021/jo100981e

A high-throughput screening system for homogeneous catalyst discovery has been developed by integrating a continuous-flow capillary-based microreactor with ultra-high-pressure liquid chromatography (UHPLC) for fast online analysis. Reactions are conducted in distinct and stable zones in a flow stream that allows for time and temperature regulation. UHPLC detection at high temperature allows high throughput online determination of substrate, product, and byproduct concentrations. We evaluated the efficacies of a series of soluble acid catalysts for an intramolecular Friedel-Crafts addition into an acyliminium ion intermediate within 1 day and with minimal material investment. The effects of catalyst loading, reaction time, and reaction temperature were also screened. This system exhibited high reproducibility for high-throughput catalyst screening and allowed several acid catalysts for the reaction to be identified. Major side products from the reactions were determined through off-line mass spectrometric detection. Er(OTf)₃, the catalyst that showed optimal efficiency in the screening, was shown to be effective at promoting the cyclization reaction on a preparative scale.

Full text of DOI:10.1021/jo100981e

pubs.acs.org/joc
Rapid Catalyst Screening by a Continuous-Flow Microreactor Interfaced
with Ultra-High-Pressure Liquid Chromatography
Hui Fang, Qing Xiao, Fanghui Wu, Paul E. Floreancig,* and Stephen G. Weber*
Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260
sweber@pitt.edu; florean@pitt.edu
Received May 19, 2010
A high-throughput screening system for homogeneous catalyst discovery has been developed by
integrating a continuous-flow capillary-based microreactor with ultra-high-pressure liquid chroma-
tography (UHPLC) for fast online analysis. Reactions are conducted in distinct and stable zones in a
flow stream that allows for time and temperature regulation. UHPLC detection at high temperature
allows high throughput online determination of substrate, product, and byproduct concentrations.
We evaluated the efficacies of a series of soluble acid catalysts for an intramolecular Friedel-Crafts
addition into an acyliminium ion intermediate within 1 day and with minimal material investment.
The effects of catalyst loading, reaction time, and reaction temperature were also screened.
This system exhibited high reproducibility for high-throughput catalyst screening and allowed
several acid catalysts for the reaction to be identified. Major side products from the reactions
were determined through off-line mass spectrometric detection. Er(OTf)3, the catalyst that showed
optimal efficiency in the screening, was shown to be effective at promoting the cyclization reaction
on a preparative scale.
Introduction
capacity for heat and mass transfer and their safety benefits
in handling potentially explosive or toxic compounds rela-
tive to standard batch reactions.1 While most commonly
applied toward preparative procedures, microreactors can
also be applied to reaction condition or catalyst screening.
Screening protocols are an essential component in the opti-
mization of many chemical processes.2 These efforts generally
proceed through running several reactions under conven-
tional conditions followed by analysis. While undoubtedly
effective, this approach requires substantial investments in
materials and time. Applying microreactors with online detec-
tion capabilities to homogeneous catalyst selection in liquid-
phase reactions to screen reaction parameters has the capacity
Microreactors are attracting increased attention for a
variety of synthesis applications because of their superior
(1) (a) Ahmed-Omer, B.; Brandt, J. C.; Wirth, T. Org. Biomol. Chem.
2007, 5, 733–740. (b) Fletcher, P. D. I.; Haswell, S. J.; Pombo-Villar, E.;
Warrington, B. H.; Watts, P.; Wong, S. Y. F.; Zhang, X. Tetrahedron 2002,
58, 4735–4757. (c) Haswell, S. J.; Middleton, R. J.; O’Sullivan, B.; Skelton,
€
V.; Watts, P.; Styring, P. Chem. Commun. 2001, 391–398. (d) Jahnisch, K.;
Hessel, V.; Lowe, H.; Baerns, M. Angew. Chem., Int. Ed. 2004, 43, 406–446.
€
(e) Anspach, J. A.; Maloney, T. D.; Colon, L. A. J. Sep. Sci. 2007, 30, 1207–
1213. (f) Mason, B. P.; Price, K. E.; Steinbacher, J. L.; Bogdan, A. R.;
McQuade, D. T. Chem. Rev. 2007, 107, 2300–2318. (g) Roberge, D. M.;
Bieler, N.; Thalmann, M. PharmaChem 2006, 5, 14–17. (h) Schwalbe, T.;
Autze, V.; Wille, G. Chimia 2002, 56, 636–646. (i) Seeberger, P. H.; Geyer, K.;
Codee, J. D. C. Ernst Schering Found. Symp. Proc. 2007, 1–19. (j) Watts, P.
Curr. Opin. Drug Discovery Dev. 2004, 7, 807–812. (k) Watts, P.; Haswell,
S. J. Chem. Eng. Technol. 2005, 28, 290–301. (l) Watts, P.; Wiles, C. Org.
Biomol. Chem. 2007, 5, 727–732. (m) Kobayashi, J.; Mori, Y.; Kobayashi, S.
Chem. Asian J. 2006, 1, 22–35. (n) Fukuyama, T.; Rahman, R. T.; Sato, M.;
Ryu, I. Synlett 2008, 151–163. (o) Yoshida, J.-i.; Nagaki, A.; Yamada, T.
Chem.;Eur. J. 2008, 14, 7450–7459.
(2) For reviews, see: (a) Gennari, C.; Piarulli, U. Chem. Rev. 2003, 103,
3071–3100. (b) Dahmen, S.; Brase, S. Synthesis 2001, 1431–1449.
(c) Reetz, M. T. Angew. Chem., Int. Ed. 2001, 40, 284–310. (d) Shimizu,
K. D.; Snapper, M. L.; Hoveyda, A. H. Chem.;Eur. J. 1998, 4, 1885–
1889.
DOI: 10.1021/jo100981e
r
Published on Web 07/28/2010
J. Org. Chem. 2010, 75, 5619–5626 5619
2010 American Chemical Society

Fang et al.
JOCArticle
to streamline the process of optimizing reaction conditions
significantly.3 Yet catalysis in liquid phases is rarely studied
in microreactors.1d
Results and Discussion
Reaction Selection. The reaction that we investigated with
this system is shown in Scheme 1. Acylaminal 1, prepared
through a previously reported multicomponent process,7
undergoes ionization to form acyliminium ion 2 when sub-
jected to a Lewis acid such as TMSOTf and subsequently
engages in an intramolecular Friedel-Crafts reaction to
form indanyl amide 3.8 This facile reaction is ideally suited
for a catalyst screening study because of good reagent
solubility and its well-defined array of side reactions. An-
other motivation for studying this reaction is the growing
interest in reactions that proceed through acyliminium ions,9
particularly when they are generated by chiral acids.10
A schematic diagram of the microreactor that was used in
these studies is shown in Figure 1. Catalyst solutions arriving
from an autosampler are combined with the substrate solu-
tion and fed to the loop of a capillary loop injector. The
contents of the loop are pumped onto the reaction capillary
by a syringe pump creating a reaction zone. Reaction zones
are separated naturally by the solvent in this syringe pump.
The reaction time may be controlled by the length and
diameter of the capillary and the flow rate of the syringe
pump. The reaction time is the time required to travel the
entire length of the reaction capillary. The flow can also be
stopped to increase the reaction time once all of the zones are
loaded into the reaction capillary, provided that all zones fit
in the capillary. Thus the volume of the injected zones
(including the solvent-only region) and the total volume of
the reaction capillary together determine the number of
zones that the reaction capillary can hold. These zones,
whether they flow continuously or stop for a time prior to
flowing into the detection system, are surprisingly stable
toward dispersion4 because the narrow diameter of the
capillary suppresses hydrodynamic dispersion of reagents
by permitting radial concentration gradients to relax rapidly
(Figure 2). This allows the same solvent to be used for zone
separation and reaction, thus permitting multiple parallel
reactions to be conducted simultaneously in a single reactor
capillary where temperature can be controlled externally and
reaction time can be varied by adjusting the flow rate or
A microreactor that can screen multiple transformations to
identify optimal parameters such as solvent, catalyst identity
and loading level, temperature, and time would be extremely
valuable in reaction development. Ideally the microreactor
should be capable of conducting reactions over a range of
time scales to accommodate processes with different rates.
Finally, online analysis is desirable to expedite data through-
put. While microfabricated devices have a number of advan-
tages, they do not currently meet the needs of the typical
synthetic laboratory because of the cost associated with their
development and implemenation and the expertise required
for their operation. We have therefore initiated a program
with the objective of developing a capillary-based microreac-
tor by using common laboratory items and equipment that is
applicable for high-throughput catalyst screening for either
fast or slowreactions. Previously we reported the construction
of an automated flow-through fused silica microreactor with
online GC detection and its application to catalyst screening
for the Stille reaction.4 Throughput in this screening instru-
ment is at least partly defined by the time required for online
analysis. An advantage of GC detection is that separations
with good efficiency can be carried out in a few minutes.
However, the reliance upon GC for detection limits its
applicability to the analysis of volatile and thermally stable
compounds. In principle, this limitation can be addressed
by replacing the GC component with an HPLC system.5
However, routine HPLC is generally not capable of fast
separations, causing throughput problems. This bottleneck
can be addressed by employing UHPLC detection at elevated
temperatures. UHPLC in conjunction with elevated column
temperature allows for rapid separation of a wide range of
nonvolatile organic molecules with adequate peak resolution.
In this paper we report that a Teflon/FEP-capillary-based flow
microreactor coupled to a UHPLC separation/detection sys-
tem is an extremely effective device for screening the capacity
of a range of Brønsted and Lewis acids to effect an intramo-
lecular Friedel-Crafts reaction.6 Off-line mass spectrometry
provides data regarding the identities of reaction byproduct in
addition to information about reaction rates and efficiencies.
(7) (a) Wan, S.; Green, M. E.; Park, J.-H.; Floreancig, P. E. Org. Lett.
2007, 9, 5385–5388. (b) DeBenedetto, M. V.; Green, M. E.; Wan, S.; Park,
J.-H.; Floreancig, P. E. Org. Lett. 2009, 11, 835–838.
(3) (a) De Bellefon, C.; Tanchoux, N.; Caravieilhes, S.; Grenoullet, P.;
Hessel, V. Angew. Chem., Int. Ed. 2000, 39, 3442–3445. (b) Garcia-Egido, E.;
Spikmans, V.; Wong, S. Y. F.; Warrington, B. H. Lab Chip 2003, 3, 73–76.
(c) Greenway, G. M.; Haswell, S. J.; Morgan, D. O.; Skelton, V.; Styring, P.
Sens. Actuators, B 2000, B63, 153–158. (d) Pennemann, H.; Hessel, V.; Kost,
H. J.; Loewe, H.; de Bellefon, C.; Pestre, N.; Lamouille, T.; Grenouillet, P.
Better Processes for Bigger Profits, International Conference on Process
Intensification for the Chemical Industry, 5th, Oct. 13-15, Maastricht,
Netherlands 2003, 137-147. (e) Zech, T.; Klein, J.; Schunk, S. A.; Johann,
T.; Schueth, F.; Kleditzsch, S.; Deutschmann, O. High-Throughput Analysis
2003, 491–523. (f) Suga, S.; Okajima, M.; Fujiwara, K.; Yoshida, J.-i. J. Am.
Chem. Soc. 2001, 123, 7941–7942.
(4) Shi, G.; Hong, F.; Liang, Q.; Fang, H.; Nelson, S.; Weber, S. G. Anal.
Chem. 2006, 78, 1972–1979.
(5) Ratner, D. M.; Murphy, E. R.; Jhunjhunwala, M.; Snyder, D. A.;
Jensen, K. F.; Seeberger, P. H. Chem. Commun. 2005, 578–580.
(6) detector systems, pleas/detector systems, please see: (a) Hatakeyama,
T.; Chen, D. L.; Ismagilov, R. J. Am. Chem. Soc. 2006, 128, 2518–2519. (b)
Wang, J.; Sui, G.; Mocharla, V. P.; Lin, R. J.; Phelps, M. E.; Kolb, H. C.;
Tseng, H.-R. Angew. Chem., Int. Ed. 2006, 45, 5276–5281. (c) Murphy, E. R.;
Martinelli, J. R.;Zaborenko, N.;Buchwald, S. L.; Jensen, K. F.Angew. Chem.,
Int. Ed. 2007, 46, 1734–1737. (d) Knudsen, K. R.; Holden, J.; Ley, S. V.;
Ladlow, M. Adv. Synth. Catal. 2007, 349, 535–538. (e) Goodell, J. R.;
McMullen, J. P.; Zaborenko, N.; Maloney, J. R.; Ho, C.-X.; Jensen, K. F.;
Porco, J. A., Jr.; Beeler, A. B. J. Org. Chem. 2009, 74, 6169–6180.
(8) Xiao, Q.; Floreancig, P. E. Org. Lett. 2008, 10, 1139–1142.
(9) (a) Petrini, M.; Torregiani, E. Synthesis 2007, 159–186. (b) Maryanoff,
B. E.; Zhang, H.-C.; Cohen, J. H.; Turchi, I. J.; Maryanoff, C. A. Chem. Rev.
2004, 104, 1431–1628. (c) Speckamp, W. N.; Moolenaar, M. J. Tetrahedron
2000, 56, 3817–3856.
(10) For representative examples, see: (a) Uraguchi, D.; Terada, M.
J. Am. Chem. Soc. 2004, 126, 5356–5357. (b) Uraguchi, D.; Sorimachi, K.;
Terada, M. J. Am. Chem. Soc. 2004, 126, 11804–11805. (c) Uraguchi, D.;
Sorimachi, K.; Terada, M. J. Am. Chem. Soc. 2005, 127, 9630–9631. (d)
Rowland, G. B.; Zhang, H.; Rowland, E. B.; Chennamadhavuni, S.; Wang,
Y.; Antilla, J. C. J. Am. Chem. Soc. 2005, 127, 15696–15697. (e) Terada, M.;
Sorimachi, K. J. Am. Chem. Soc. 2007, 129, 292–293. (f) Rowland, G. B.;
Rowland, E. B.; Liang, Y.; Perman, J. A.; Antilla, J. C. Org. Lett. 2007, 9,
2609–2611. (g) Li, G.; Rowland, G. B.; Rowland, E. B.; Antilla, J. C. Org.
Lett. 2007, 9, 4065–4068. (h) Terada, M.; Machioka, K.; Sorimachi, K. J.
Am. Chem. Soc. 2007, 129, 10336–10337. (i) Raheem, I. T.; Thiara, P. S.;
Peterson, E. A.; Jacobsen, E. N. J. Am. Chem. Soc. 2007, 129, 13404–13405.
(j) Raheem, I. T.; Thiara, P. S.; Jacobsen, E. N. Org. Lett. 2008, 10, 1577–
1580. (k) Li, G.; Fronczek, F. R.; Antilla, J. C. J. Am. Chem. Soc. 2008, 130,
12216–12217. (l) Li, G.; Antilla, J. C. Org. Lett. 2009, 11, 1075–1078. (m)
Terada, M.; Machioka, K.; Sorimachi, K. Angew. Chem., Int. Ed. 2009, 48,
2553–2556. (n) Terada, M.; Toda, Y. J. Am. Chem. Soc. 2009, 131, 6354–
6355. (o) Muratore, M. E.; Holloway, C. A.; Pilling, A. W.; Storer, R. I.;
Trevitt, G.; Dixon, D. J. J. Am. Chem. Soc. 2009, 131, 10796–10797. (p)
Peterson, E. A.; Jacobsen, E. N. Angew. Chem., Int. Ed. 2009, 48, 6328–6331.
5620 J. Org. Chem. Vol. 75, No. 16, 2010

Fang et al.
JOCArticle
FIGURE 1. Block diagram of the reactor. L1 and L2 are loop injectors.
FIGURE 2. Absorbance response of 54 reaction zones passing through the UV-vis optical detector at a flow rate of 0.9 μL/min for a
residence time of 1 h.
SCHEME 1. Test Reaction
stopping the flow for defined time periods. The same condi-
tions that minimize axial dispersion also permit rapid re-
agent mixing by simple diffusion. A Teflon/FEP capillary,
rather than a more conventional SiO2 capillary, was selected
for this system because of its greater stability toward acids.
Also, preliminary work demonstrated that the well-known
acid/base chemistry of silica led to carryover between adja-
cent, acid-containing, reaction zones.
An optical-fiber-based absorbance detector (220 nm) is
placed at the distal end of the reaction capillary in series with
the UHPLC (Figure 1). This detector measures absorbance
inside the capillary rather than a separate flow cell, as
typically found in HPLC detectors. Thus the zones are
visualized but not destroyed. Figure 2 demonstrates the
excellent zone separation. The absorbance-time trace is also
used to initiate an automated sequence based on the absolute
value of the absorbance resulting in injection of a portion of
the zone into the UHPLC for online analysis.
Examples of the UHPLC chromatograms from reactions
catalyzed by a strong and a weak acid are shown in Figure 3.
The internal standard, starting material, and products from
these experiments were completely resolved by UHPLC
within an approximately 6 min run, consistent with the
arrival frequency of the reaction zones at the injector.
The use of an internal standard, while good practice, may
not be required if the process of injection into the liquid
chromatograph, which is automated, is highly reproducible.
An internal standard is typically used to correct for varia-
bility in the injected sample volume. It can also be used to
ascertain the reproducibility of the retention process. An
J. Org. Chem. Vol. 75, No. 16, 2010 5621

Fang et al.
JOCArticle
FIGURE 3. Top: Series of three chromatograms from an HClO4-catalyzed reaction showing peaks for an unknown, product, internal standard, and
substrate. The black dots indicate injection. Bottom: Series of three chromatograms from a diphenylphosphoric acid-catalyzed reaction.
catalyst concentration. Success in this single reaction would
portend well for the use of the reactor as a general device to
identify catalysts for a range of acid-mediated processes and for
applications to more ambitious objectives such as screening for
asymmetric catalysis. Each reaction zone contains about 2.4 μg
of substrate. In this case, the starting material is not particularly
valuable, so there was no need to reduce the substrate consump-
tion by the reactor. As the pump containing substrate (Figure 1)
ran continuously, the creation of each zone took approximately
0.3 mg of the substrate, more than 99% of which went to waste
and was not recovered. Programming the syringe pump and
reducing the UHPLC column size could each reduce the
material requirement by a factor of 100. Thus, running numer-
ous reactions with nanograms of starting material is feasible if
the substrate is not readily accessible. We used diphenylpho-
FIGURE 4. Internal standard retention time ([) and peak area (O)
sphoric acid as the catalyst in these studies. The data for these
relative to the mean internal standard retention time and peak area
experiments are shown in Figure 5. Each reaction was run in
in chromatograms from all zones in a run.
triplicate. The excellent reproducibility between runs is demon-
internal standard adds to the complexity of the chromato-
gram and can be difficult to select. We reasoned that we
could eliminate the internal standard if its peak area and
retention time are highly reproducible in a series of injections
from a variety of reaction zones. As Figure 4 shows, reten-
tion times were highly reproducible but peak areas were not.
We infer from the poor reproducibility of the peak area
that the amount of each reaction zone that is injected is
slightly variable from zone to zone. The variability arises
because optical peak detection is based on absolute mea-
sured absorbance. When the absorbance in a zone exceeds a
predetermined value then the zone is directed into the loop of
the injector. As a result the trigger occurs after a greater
fraction of the zone has passed through the detector for zones
with lower absorbance relative to zones with higher absor-
bance. Thus an internal standard is necessary to account for
this variability in the injected quantity.
strated by the small error bars in Figure 5. The small standard
errors reflect both reproducible reaction conditions and repro-
ducible analysis of the reaction zones. We note in particular that
the reproducibility of the data is better than the reproducibility
of the internal standard peak areas shown above, thereby
validating the benefits of internal standard incorporation for
reducing errors arising from injection volume variability. These
curves demonstrate that reactions show the expected behavior
in the microreactor, with starting material consumption and
product formation increasing as time, temperature, and catalyst
concentration increase. Thus important information broadly
defining the range of behavior of weak Brønsted acid catalysts
can be obtained in a short time with minimal material invest-
ment, indicating that the microreactor is well-suited for rapid
mechanistic studies of acid-mediated processes.
Our next objective was to use the microreactor to identify
new catalysts to promote the reaction. We chose to use reaction
times of 1 h at a reaction temperature of 0 °C and of 0.5 h at a
reaction temperature of 40 °C for catalyst screening. We chose a
catalyst loading of 0.2 equiv to identify species that turn over.
Reaction Studies. Our initial tests focused on demonstrating
the ability to monitor starting material consumption and
product formation as a function of time, temperature, and
5622 J. Org. Chem. Vol. 75, No. 16, 2010

Fang et al.
JOCArticle
and product formation (Figure 6). Ratios of product formation
to starting material consumption were calculated to determine
the specificity of each catalyst under the prevailing conditions.
The results are shown in Table 1. We drew several conclusions
from the data. Reactions were more efficient at 40 °C than
at 0 °C. Weak Brønsted acids such as Cl3CCO2HandF3CCO2H
do not promote substantial product formation or starting
material consumption. Strong Brønsted acids such as HCl,
HNO3, H2SO4, and HClO4 were more effective with respect to
starting material consumption. The rate of substrate consump-
tion for reactions that were catalyzed by Brønsted acids roughly
correlated with aqueous pKa trends, though exact pKa values
for many Brønsted acids in CH3CN can be elusive.11 However,
except for perchloric acid, selectivity for cyclization product
formation is mediocre. The Lewis acids generally performed
better than the Brønsted acids. The lanthanide triflates are by
far the most promising catalysts, showing complete starting
material consumption and high selectivity for the desired
cyclization product. Er(OTf)3, an oxophilic catalyst that is
attracting increased attention for acid-mediated processes,12
showed the most promising profile.
Byproduct Analysis. An advantage of the microreactor
approach with chromatographic detection is that product
formation patterns can be detected. Identifying byproducts
that appear consistently is important to provide information
regarding the reaction pathway and/or product stability. In
our studies UHPLC analysis showed that three major by-
products were formed in reactions that did not produce the
desired bicyclic amide efficiently. In fact, the byproduct
distribution was distinctly different in reactions catalyzed
by the weaker acids in comparison to the stronger acids. We
switched from online UHPLC analysis to offline GCMS
(electron impact ionization) and LCMS (electrospray
ionization) detection of byproducts in reactions in vials in
an effort to identify the structures of these products (see the
Supporting Information). A major byproduct that was
formed (Scheme 2) when weak acids were used showed an
[M þ Na]þ ion at m/z 332.2. This mass is consistent with
ionization followed by addition of H2O to the acyliminium
ion to form acyl hemiaminal 4. The identity of 4 was
confirmed by preparing it independently from nitrile 5
through a sequence of hydrozirconation, acylation, and
water addition,5a subjecting it to UHPLC analysis, and
observing an identical retention time to the unknown by-
product. The identification of 4 led us to postulate that the
other major byproduct from these reactions was aldehyde 6,
which arises from the breakdown of 4. This structural
assignment was also confirmed by independent synthesis
and UHPLC analysis. While water was not deliberately
added to these reactions, it was not rigorously excluded
and it was used in the solvent for UHPLC detection, leading
to two potential sources for byproduct formation.
FIGURE 5. Variations of yield and substrate conversion as a
function of reaction time, temperature, and catalyst concentration.
Plotted data are mean ( SEM (standard error) (n = 3). Top: 0.01 M
substrate and 0.01 M 4-ethylanisole (internal standard) in CH3CN
with 1 mol equiv of diphenylphosphoric acid at 40 °C for various
times. Middle: 0.01 M substrate and 0.01 M 4-ethylanisole with
1 mol equiv of diphenylphosphoric acid at various temperatures for
1 h. Bottom: 0.01 M substrate and 0.01 M 4-ethylanisole with
varying diphenylphosphoric acid concentrations at 40 °C for 1 h.
Thus, we loaded the autosampler with solutions of several
Brønsted and Lewis acids. The automated microreactor com-
bined them with a solution of substrate, and injected them into
the reaction capillary. Each of the acid catalysts was tested in
triplicate. All reactions could be conducted within 6 h and
required only 17.4 mg (syringe pump 1 (Figure 1) flowing at
15 μL/min for 6 h with 0.01 M substrate) or less than 400 μg per
reaction of the cyclization substrate. The zones were monitored
by UHPLC to determine the extent of substrate consumption
(11) For spectrophotometric determination of pKa values for several
€
€
Brønsted acids in acetonitrile, see: Kutt, A.; Leito, I.; Karljurand, I.; Soovali,
L.; Vlasov, V. M.; Yagupolskii, L. M.; Koppel, I. A. J. Org. Chem. 2006, 71,
2829–2838.
(12) (a) Dalpozzo, R.; De Nino, A.; Maiuolo, L.; Nardi, M.; Procopio,
A.; Tagarelli, A. Synthesis 2004, 496–498. (b) Procopio, A.; Dalpozzo, R.; De
Nino, A.; Maiuolo, L.; Russo, B.; Sindona, G. Adv. Synth. Catal. 2004, 346,
1465–1470. (c) Procopio, A.; Dalpozzo, R.; De Nino, A.; Nardi, M.;
Sindona, G.; Tagarelli, A. Synlett 2004, 2633–2635. (d) Procopio, A.;
Dalpozzo, R.; De Nino, A.; Maiuolo, L.; Nardi, M.; Russo, B. Adv. Synth.
Catal. 2005, 347, 1447–1450. (e) Procopio, A.; Gaspari, M.; Nardi, M.;
Oliverio, M.; Rosati, O. Tetrahedron Lett. 2008, 49, 2289, 2293.
J. Org. Chem. Vol. 75, No. 16, 2010 5623

Fang et al.
JOCArticle
FIGURE 6. Series of chromatograms from the 54 reaction zones. The inset chromatogram shows the peaks in the last six reaction zones
representing two catalyts.
TABLE 1. Starting Material Consumption and Product Formation from Various Catalysts
temperature, 0 °C
temperature, 40 °C
ratio of yield/
conversion (%)b
ratio of yield/
conversion (%)
entry
catalyst
yield (%)a
conversiona (%)
yield (%)
conversion (%)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
H2SO4 (98%)
H3PO4 (85%)
HNO3 (70%)
HClO4 (70%)
HCl (38%)
9.4 ( 1.2
2.2 ( 2.4
0
63.3 ( 0.8
2.6 ( 2.4
4.4 ( 0.8
11.8 ( 0.3
0
79.5 ( 0.6
53.6 ( 4.0
44.8 ( 2.1
98.3 ( 0.2
55.4 ( 4.1
69.9 ( 3.3
78.8 ( 1.1
22.8 ( 6.0
17.3 ( 0.5
96.9 ( 0.2
98.8 ( 0.2
99.0 ( 0.1
99.0 ( 0.1
91.1 ( 4.5
83.9 ( 1.3
82.3 ( 0.8
80.6 ( 0.2
78.1 ( 0.7
11.8
4.1
0
64.3
4.8
6.3
14.9
0
0
50.1
51.6
52.8
57.2
36.1
30.8
28.7
26.1
25.6
61.4 ( 0.3
40.9 ( 0.2
38.7 ( 1.2
80.1 ( 3.0
47.1 ( 3.9
61.4 ( 0.3
61.5 ( 0.1
5.1 ( 2.4
1.3 ( 0.3
78.5 ( 0.3
58.8 ( 0.6
70.0 ( 0.2
65.9 ( 0.3
88.2 ( 5.4
85.5 ( 3.1
93.2 ( 0.3
98.8 ( 0.1
72.1 ( 0.1
71.5 ( 0.5
99.6 ( 0.6
84.8 ( 3.1
97.9 ( 0.1
100
24.7 ( 5.4
8.2 ( 2.9
100
100
100
100
97.2 ( 0.9
91.8 ( 1.2
95.7 ( 0.1
62.1
56.7
54.1
80.4
55.6
62.6
61.5
20.7
16.3
78.5
58.8
70.0
65.9
88.2
93.2
97.4
CH3SO3H
p-TsOH H2O
3
CF3COOH
Cl3CCOOH
SnCl4
0
48.6 ( 1.5
51.1 ( 0.3
52.3 ( 0.4
56.7 ( 0.2
32.8 ( 6.4
25.8 ( 2.8
23.6 ( 1.4
21.0 ( 0.9
20.0 ( 1.0
SbCl5
BF3 OEt2
BF3 THF
3
3
Eu(OTf)3
Tb(OTf)3
Er(OTf)3
Ho(OTf)3
Yb(OTf)3
aYields of product and conversions of substrate are determined by UHPLC. Yield and conversion = mean ( SE (standard error) (n = 3). bPercentage
ratio of yield to conversion.
SCHEME 2. Side Reactions with Weak Acids and Independent Byproduct Synthesis
Strong acids produced an additional byproduct with an
[M þ Na]þ m/z value of 286.0, corresponding to a formal loss
of C2H4 from the cyclization product. This molecular formula
suggested that the product arose (Scheme 3) from the acid-
mediated ionization of 3to form benzylic cation 7followed by a
Ritter reaction13 that proceeds through the sequential addition
of acetonitrile (solvent) and water to yield 8. The identity
of 8 was again confirmed by independent synthesis through
subjecting 5 to hydrozirconation, acylation with AcCl, and
Friedel-Crafts cyclization mediated by ZnCl28 and observing
identical UHPLC behavior to the byproduct.
Identifying the side products of the reaction provides
valuable information regarding the behavior of substrates
and products under reaction conditions. The present study
(13) Ritter, J. J.; Minieri, P. P. J. Am. Chem. Soc. 1948, 70, 4045–4048.
5624 J. Org. Chem. Vol. 75, No. 16, 2010

Fang et al.
JOCArticle
SCHEME 3. Side Reaction with Strong Acids and Independent Byproduct Synthesis
shows that acids must be potent but not too potent for the
reaction to proceed efficiently. Mechanistic hypotheses re-
garding the formation of the side products will provide
guidance for selecting acids for other substrates as the scope
of the process is expanded.
homemade nanoliter-volume tee mixer (Figure 1). The com-
bined fluids pass through and fill a loop (750 nL) in a 6-port
2-position HPLC injector (L1). We program the autosampler to
control the loading time for the catalyst and reagents (which are
determined by sampling volume and loading flow rate). For
instance, when 25 μL of catalyst solution is sampled by the
autosampler and the loading flow rate is 15 μL/min, then the
loading time is about 25/15 = 1.6 min. When the loading time is
reached, the autosampler triggers L1 to switch to the “inject
position” and contents of the loop are then injected into the reaction
capillary by syringe pump 2 at a constant flow rate. This injection
time is also controlled by the autosampler. This time is determined
by the loop volume and flow rate of syringe pump 2. When the
injection is over, the autosampler will trigger L1 to switch back to
“load position” and then the next catalyst/reagent zone will be
loaded and injected following the same procedures. We also added
a “stop time” step to the program in the autosampler. During this
time the autosampler is rinsed in order to prevent carryover in the
autosampler. The periodic loading of catalysts by the autosampler
with the constant flow of syringe pump 1 generates a series of
reaction zones in the reaction capillary which contain different
catalysts with common reagents. These zones are separated by the
carrier solvent CH3CN delivered by syringe pump 2. The online
UHPLC analysis time must be less than the total of the times for
“loading”, “injection”, and “stop time” to ensure continuous
analysis of individual serial zones.
The microreactor capillary is a piece of 100 μm i.d., 1/16 in. o.d.,
6.1 m long Teflon capillary tube. A water/oil bath is used for
heating or cooling the reaction capillary to the required tem-
perature. Syringe pump 2 is constantly delivering carrier solvent
CH3CN to push all reaction zones through the reactor. The
reaction time is controlled by flow rate of CH3CN and the length
of the reaction capillary. When the reaction zones come out of
the reactor, they first go through a flow cell monitored by a
UV-vis fiber optic absorbance detector before entering the 10-
port 2-position double-loop (1.0 μL) valve (L2). L2 is a 15000-
psi-high-pressure valve. It enables the loading of one loop from
the reaction capillary while the contents of the other loop are
analyzed by UHPLC. The UHPLC runs continuously. L2 is
triggered by signals coming from the fiber optic absorbance
detector. Thus a single chromatographic run contains the
chromatograms of all of the reaction zones in sequence.
Application to a Preparative Scale Reaction. The results
from the catalyst screening protocol should be applicable to
preparative scale reactions. When we applied the conditions that
were studied in the microreactor (0.01 M substrate in CH3CN,
0.2 equiv of Er(OTf)3, 40 °C, 30 min), however, minimal
product formation was observed. Extending the reaction time
to 8 h and increasing the catalyst loading to 1 equiv resulted in
the formation of 3 in 91% isolated yield when 150 mg of
substrate was employed. Significantly enhanced reaction selec-
tivity and rate are often found in liquid-phase reactions in
microreactors compared to batch reactions. We postulate that
the rate difference between the reaction in the microreactor and
the preparative scale reaction arise from the highly efficient
mixing in the microreactor. Increasing the mixing rate could
increase the conversion rate of the tight ion pair that forms
initially and reversibly upon acylaminal dissociation to a solvent
separated ion pair that would engage in the cyclization reaction
more readily. Despite the rate differences, this study demon-
strated that unique catalysts for efficient preparative scale
reactions can be identified by screening with this microreactor.
Conclusions
We have reported a new flow-through capillary-based micro-
reactor coupled with UHPLC detection that allows for near-
real-time evaluation of reaction efficiency under a variety of
conditions or in the presence of different catalysts. The micro-
reactor is based upon the establishment of reaction zones that
are stable toward dispersion because of the narrow diameter of
the capillary, the low flow velocity through the capillary, and the
relatively high diffusion coefficients of the catalysts and sub-
strate. Reaction results are highly reproducible and display the
expected behavior upon increasing reaction time, temperature,
and catalyst concentration. The microreactor that was described
in this paper has numerous applications beyond those described
above. The reproducibility and high throughput capacity will
allow for mechanistic studies of acid-catalyzed reactions by
facilitating labor intensive determinations of linear free energy
relationships. Employing a separation system with the capacity
to discriminate between enantiomers will be extremely useful for
identifying asymmetric catalysts for enantioselective reactions.
These objectives are currently being pursued in our laboratories
and the results will be reported in due course.
UHPLC Analysis. The isocratic separation was carried out
with a mobile phase of CH3CN:H2O (24:76 v/v) at a flow rate of
0.35 mL/min and a temperature of 70 °C. The injection volume
is 1.0 μL, which is the size of the loop of the 10-port injector. A
single chromatographic run is set to analyze all reaction zones.
Reactions in the Microreactor. The cyclization reaction was
carried out in a continuous flow stream of CH3CN in the
microreactor at a constant temperature. Solutions of Brønsted
and Lewis acids (20 mol % in CH3CN) were placed in 1.5-mL
glass autosampler vials. The catalyst vials were placed in the tray
of the autosampler. The reagents (0.01 M substrate and 0.01 M
4-ethylanisole (internal standard) in CH3CN) were loaded in a
2.5 mL gastight syringe and constantly driven by syringe pump
1. Syringe pump 2 delivered the carrier solvent (CH3CN) at a
constant flow rate to push all reaction zones through the reactor
followed by online UHPLC analysis.
Experimental Section
Microreactor Setup. Catalysts sampled by the autosampler
(25 μL sampling volume) are combined with reagents pushed by
syringe pump 1 at equal flow rates (15 μL/min) through a
J. Org. Chem. Vol. 75, No. 16, 2010 5625

Fang et al.
JOCArticle
General Synthesis: Experimental Details. H NMR and 13C
NMR spectra were recorded at 300 and 75 MHz, respectively.
The chemical shifts are given in parts per million (ppm) on the
delta (δ) scale. The solvent peak was used as a reference value,
for 1H NMR: CDCl3=7.27 ppm, for 13C NMR: CDCl3=77.23.
Data are reported as follows: s=singlet; d=doublet; t=triplet;
q=quartet; dd=doublet of doublets; dt=doublet of triplets;
br=broad. Samples for IR were prepared as a thin film on a
NaCl plate by dissolving the compound in CH2Cl2 and then
evaporating the CH2Cl2. Methylene chloride was distilled under
(neat) 2963, 2934, 2835, 1721, 1589, 1515, 1262, 1190, 1156,
1029 cm-1; HRMS (EI) m/z calcd for C13H18O3Na (MNaþ)
245.1154, found 245.1157.
1
N-(5,6-Dimethoxy-2,2-dimethyl-2,3-dihydro-1H-inden-1-yl)-
acetamide (8). To a solution of 3-(3,4-dimethoxyphenyl)-2,2-
dimethylpropanenitrile7 (219 mg, 1.0 mmol) in dichloromethane
(10 mL) was added Cp2Zr(H)Cl (319 mg, 1.24 mmol). The
reaction mixture was stirred for 30 min at room temperature. Acetyl
chloride (88 μL, 1.2 mmol) was added and the mixture was stirred
for another 15 min. ZnCl2 (400 μL, 1.0 M solution in diethyl ether,
0.4 mmol) was added and the mixture was stirred for 2 h at room
temperature. The reaction was quenched with NaHCO3 (sat.
aqueous). The mixture was extracted with ethyl acetate and the
extracts were washed with H2O and brine. The organic layer was
dried over Na2SO4. The solvent was removed and the residue was
purified by column chromatography (hexanes:ethyl acetate = 1:1)
to give 8 (250 mg, 95%). 1H NMR (300 MHz, CDCl3) δ 6.73 (s,
1H), 6.72 (s, 1H), 5.47 (d, 1H, J = 10.2 Hz), 5.17(d, 1H, J = 9.6
Hz), 3.87 (s, 3H), 3.86 (s, 3H), 2.69 (s, 2H), 2.10 (s, 3H), 1.24 (s, 3H),
1.01 (s, 3H); 13C NMR (75 MHz, CDCl3) δ 169.8, 148.9, 148.1,
133.9, 133.8, 107.8, 107.2, 62.4, 55.9, 55.8, 45.5, 44.3, 27.5,23.3, 22.7;
IR (neat) 3260, 2958, 2926, 1641, 1549, 1503, 1464, 1369, 1315,
1289, 1114, 750 cm-1; HRMS (EI) m/z calcd for C15H21NO3 (Mþ)
263.1521, found 263.1527.
˚
N2 from CaH2. 1,2-Dichloroethane was dried over 4 A mole-
cular sieves. Analytical TLC was performed on precoated (25
mm) silica gel 60F-254 plates. Visualization was done under UV
˚
(254 nm). Flash chromatography was done with 60 A silica gel.
Reagent grade ethyl acetate, diethyl ether, pentane, and hexanes
(commercial mixture) were used as is for chromatography. All
reactions were performed in oven- or flame-dried glassware
under a positive pressure of N2 with magnetic stirring unless
otherwise noted. The synthesis and purification of reaction
substrate 1 was conducted according to a literature protocol.8
N-(3-(3,4-Dimethoxyphenyl)-1-hydroxy-2,2-dimethylpropyl)-
isobutyramide (4). To a solution of nitrile 5 (219 mg, 1.0 mmol) in
dichloromethane (10 mL) was added Cp2Zr(H)Cl (319 mg, 1.24
mmol). The reaction mixture was stirred for 30 min at room
temperature. Isobutyryl chloride (132 μL, 1.24 mmol) was
added and the mixture was stirred for another 15 min. Water
(15 mL) was added and the mixture was stirred for 30 min at
room temperature. The reaction was extracted with ethyl acetate
and the extracts were washed with brine. The organic layer was
dried over Na2SO4. The solvent was removed and the residue
was purified by column chromatography (hexanes:ethyl acetate:
N-(5,6-Dimethoxy-2,2-dimethyl-2,3-dihydro-1H-inden-1-yl)-
isobutyramide (3) (Preparative Scale). A flame-dried 100-mL
round-bottomed flask charged with 1 (150 mg, 0.46 mmol) and a
stirring bar was pumped and purged with argon three times
before acetonitrile (46 mL) was added via syringe. Er(OTf)3 (285
mg, 0.46 mmol) was added as powder in one portion and the
flask was capped with a glass stopper. The resulting clear
solution was stirred at 40 °C for 20 h. The reaction was quenched
with NaHCO3 (sat. aqueous). The mixture was extracted with
ethyl acetate and the extracts were washed with brine. The
organic layer was dried over Na2SO4. The solvent was removed
and the residue was purified by column chromatography
(hexanes:ethyl acetate=7:3) to give 3 (124 mg, 92%). All spec-
1
triethyl amine = 1:1:0.02) to give 4 (230 mg, 74%). H NMR
(300 MHz, CDCl3) δ 6.81-6.70 (m, 3H), 5.97 (d, 1H, J=8.4 Hz),
5.09 (d, 1H, J=8.4 Hz), 3.86 (s, 6H), 2.68 (d, 1H, J=13.2 Hz),
2.57 (d, 1H, J=13.2 Hz), 2.26 (sept, 1H, J=6.9 Hz), 1.07 (d, 3H,
J=6.9 Hz), 1.06 (d, 3H, J=6.9 Hz), 0.96 (s, 6H); 13C NMR (75
MHz, CDCl3) δ 177.7, 148.2, 147.4, 130.4, 122.6, 113.9, 110.6,
78.8, 55.7, 55.6, 43.3, 38.2, 35.5, 23.8, 22.0, 19.2, 18.9; IR (neat)
3362, 2966, 2934, 2873, 1657, 1515, 1465, 1262, 1236, 1156, 1029,
764 cm-1; HRMS (EI) m/z calcd for C17H27NO4 (Mþ) 309.1940,
found 309.1941.
3-(3,4-Dimethoxyphenyl)-2,2-dimethylpropanal (6). To a solu-
tion of nitrile 5 (110 mg, 0.5 mmol) in dichloromethane (5 mL)
was added Cp2Zr(H)Cl (156 mg, 0.60 mmol).The reaction
mixture was stirred for 30 min at room temperature. H2O
(500 μL, 27.7 mmol) was added dropwise and the mixture was
stirred for another 30 min at room temperature. The reaction
was quenched with NaHCO3 (sat. aqueous). The mixture was
extracted with ethyl acetate and the extracts were washed with
H2O and brine. The organic layer was dried over Na2SO4. The
solvent was removed and the residue was purified by column
chromatography (hexanes:ethyl acetate = 8:2) to give 6 (99 mg,
89%). 1H NMR (300 MHz, CDCl3) δ 9.60 (s, 1H), 6.80-6.76 (d,
1H, J = 8.1 Hz), 6.68-6.61 (m, 2H), 3.869 (s, 3H), 3.867 (s, 3H),
2.75 (s, 2H), 1.07 (s, 6H); 13C NMR (75 MHz, CDCl3) δ 206.2,
148.4, 147.7, 129.4, 122.3, 113.4, 110.9, 55.8, 7.0, 42.9, 21.5; IR
tral data are consistent with previously reported values.8 1
H
NMR (300 MHz, CDCl3) δ 6.72 (s, 1H), 6.70 (s, 1H), 5.46 (d,
1H, J=9.9 Hz), 5.17(d, 1H, J=9.6 Hz), 3.87 (s, 3H), 3.85 (s, 3H),
2.70 (s, 2H), 2.44 (sept, 1H, J=6.9 Hz), 1.24 (s, 3H), 1.23 (d, 6H,
J=6.6 Hz), 1.00 (s, 3H); 13C NMR (125 MHz, CDCl3) δ 177.1,
149.3, 148.5, 134.4, 134.3, 108.2, 107.6, 62.3, 56.24, 56.18, 46.0,
44.7, 36.1, 29.9, 28.0, 23.1, 20.2, 19.9.
Acknowledgment. This work was generously supported
by the NIH through grant P50-GM06082 and the NSF
(PEF) through grant CHE-0848299.
Supporting Information Available: Synthetic protocols, side
reaction analysis by GC-MS and LC-MS, H and 13C NMR
1
spectra of new compounds, sampling program for catalyst
loading, and microreactor parts and specifications. This mate-
rial is available free of charge via the Internet at http://pubs.
acs.org.
5626 J. Org. Chem. Vol. 75, No. 16, 2010