Exploring the scope for scale-up of organic chemistry using a large batch microwave reactor

Article abstract of DOI:10.1021/op900287j

A new batch microwave reactor has been evaluated in the context of palladium-mediated transformations, condensation reactions, nucleophilic aromatic substitution reactions, and alkylations. Importantly, a linear scaling approach was taken, no changes being made to the protocol when moving from the small, developmental scale to larger scales. In some cases reactions were scaled over 18,000-fold when moving from small (0.1-1 mmol) to large (1-18 mol) runs.

Full text of DOI:10.1021/op900287j

Organic Process Research & Development 2010, 14, 205–214
Exploring the Scope for Scale-Up of Organic Chemistry Using a Large Batch
Microwave Reactor
Jason R. Schmink, Chad M. Kormos, William G. Devine, and Nicholas E. Leadbeater*
Department of Chemistry, UniVersity of Connecticut, 55 North EagleVille Road, Storrs, Connecticut 06269-3060, U.S.A.
Abstract:
ous-flow reactors,5 small-scale batch stop-flow protocols,6,7 or
large-scale, single-batch reactors.8 Recent work in our laboratory
and by others9 has been focused at exploring all three possibili-
ties. There are a number of advantages to continuous-flow
chemistry. It limits the amount of material in the microwave
cavity at any given time, and as a result, the possibility of
catastrophic loss of an entire reaction batch is greatly reduced.
The overall scale becomes essentially limitless, and reactions
can be “scaled-out” not “scaled-up.” However, continuous-flow
processing has some drawbacks. Many reaction mixtures are
heterogeneous, biphasic, or require long reaction times (e.g.,
30-60 min) at elevated temperatures. Continuous-flow technol-
ogy is generally not amenable in these cases, and extensive
reoptimization must be undertaken in order to develop ap-
propriate homogeneous reaction conditions and suitable resi-
dence times. This in itself may require additional solvent and/
or catalyst screening. A stop-flow approach to scale-up has
similar limitations: homogeneous conditions must be maintained
throughout the cycle to avoid clogging issues. In addition, the
majority of small-scale reactions are optimized under batch
conditions. Thus, the development of a batch microwave reactor
that could perform reactions on the kilogram scale would be
highly desirable. Ideally, the scaling of a protocol from the
milligram scale to the kilogram scale should be straightforward
with little need for reoptimization. As recently addressed by
Strauss,10 and due to the overwhelming body of evidence that
A new batch microwave reactor has been evaluated in the context
of palladium-mediated transformations, condensation reactions,
nucleophilic aromatic substitution reactions, and alkylations.
Importantly, a linear scaling approach was taken, no changes being
made to the protocol when moving from the small, developmental
scale to larger scales. In some cases reactions were scaled over
18,000-fold when moving from small (0.1-1 mmol) to large (1-18
mol) runs.
Introduction
Microwave heating is a versatile and widely used tool for
preparative chemistry and continually demonstrates its worth
within the laboratory setting.1 Small-scale monomode micro-
wave units facilitate initial drug discovery and development
processes. They are suitable for performing reactions on small
scales, and reactions times can be dramatically shortened due
to the ready access to elevated temperatures in sealed vessels.
Of great interest to the process chemist is the claim that cleaner
reaction profiles can be obtained when performing chemistry
using microwave heating, due to the mitigation of thermal wall
effects.2 Since microwave irradiation heats the reaction mixture
directly and standard laboratory glassware is essentially trans-
parent when compared with the contents, the vessel walls are
the coolest part of the system. Furthermore, the rapid energy
transfer that is possible when using microwave irradiation means
that a reaction can be heated to the target temperature in a
shorter time than with conventional heating.
(5) For examples of continuous-flow processing see: (a) Moseley, J. D.;
Lawton, S. J. Chem. Today 2007, 25, 16. (b) Khadlikar, B. M.; Madyar,
V. R. Org. Process Res. DeV. 2001, 5, 452. (c) Kazba, K.; Chapados,
B. R.; Gestwicki, J. E.; McGrath, J. L. J. Org. Chem. 2000, 65, 1210.
(d) Esveld, E.; Chemat, F. van Haveren. J. Chem. Eng. Technol. 2000,
23, 429. (e) Marquie´, J.; Salmoria, G.; Poux, M.; Laporterie, A.; Dubac,
J.; Roques, N. Ind. Eng. Chem. Res. 2001, 40, 4485.
While the use of microwave heating for performing reactions
on the millimolar scale in sealed vessels is straightforward, our
group and others have been actively addressing the issues
associated with scale-up.3,4 Possible approaches include continu-
* Author for correspondence. E-mail: nicholas.leadbeater@uconn.edu.
(1) A number of books on microwave-promoted synthesis have been
published recently. For examples see: (a) Loupy, A., Ed. MicrowaVes
in Organic Synthesis, 2nd ed.; Wiley-VCH: Weinheim, 2006. (b)
Kappe, C. O.; Stadler, A. MicrowaVes in Organic and Medicinal
Chemistry; Wiley-VCH: Weinheim, 2005. (c) Lidstro¨m, P., Tierney,
J. P., Eds. MicrowaVe-Assisted Organic Synthesis; Blackwell: Oxford,
2005.
(2) Moseley, J. D.; Lenden, P.; Lockwood, M.; Ruda, K.; Sherlock, J.-P.;
Thomson, A. D.; Gilday, J. P. Org. Process Res. DeV. 2008, 12, 30.
(3) For a discussion of scale-up of microwave-assisted organic synthesis
see: Roberts, B. A., Strauss, C. R., Lindstro¨m, P., Tierney, J. P., Eds.
MicrowaVe-Assisted Organic Synthesis; Blackwell: Oxford, 2005.
(4) For recent reports from our laboratories see: (a) Bowman, M. D.;
Holcomb, J. L.; Kormos, C. M.; Leadbeater, N. E.; Williams, V. A.
Org. Process Res. DeV. 2008, 12, 41. (b) Bowman, M. D.; Schmink,
J. R.; McGowan, C. M.; Kormos, C. M.; Leadbeater, N. E. Org.
Process Res. DeV. 2008, 12, 1078. (c) Iannelli, M.; Bergamelli, F.;
Kormos, C. M.; Paravisi, S.; Leadbeater, N. E. Org. Process Res. DeV.
2009, 13, 634.
(6) Arvela, R. K.; Leadbeater, N. E.; Collins, M. J. Tetrahedron 2005,
61, 9349.
(7) (a) Moseley, J. D.; Woodman, E. K. Org. Process Res. DeV. 2008,
12, 967. (b) Pitts, M. R.; McCormack, P.; Whittall, J. Tetrahedron
2006, 62, 4705. (c) Loones, K. T. J.; Maes, B. U. W.; Rombouts, G.;
Hostyn, S.; Diels, G. Tetrahedron 2005, 61, 10338.
(8) For examples of batch processing using one vessel see: (a) Raner,
K. D.; Strauss, C. R.; Trainor, R. W.; Thorn, J. S. J. Org. Chem. 1995,
60, 2456. (b) Shackelford, S. A.; Anderson, M. B.; Christie, L. C.;
Goetzen, T.; Guzman, M. C.; Hananel, M. A.; Kornreich, W. D.; Li,
H.; Pathak, V. P.; Rabinovich, A. K.; Rajapakse, R. J.; Truesdale,
L. K.; Tsank, S. M.; Vazir, H. N. J. Org. Chem. 2003, 68, 267. (c)
Khadilkar, B. M.; Rebeiro, G. L. Org. Process Res. DeV. 2002, 6,
826. (d) Fraga-Dubreuil, J.; Famelart, M. H.; Bazureau, J. P. Org.
Process Res. DeV. 2002, 6, 374. (e) Cleophax, J.; Liagre, M.; Loupy,
A.; Petit, A. Org. Process Res. DeV. 2000, 4, 498. (f) Perio, B.; Dozias,
M.-J.; Hamelin, J. Org. Process Res. DeV. 1998, 2, 428.
(9) For a recent and comprehensive comparison of apparatus for scale-
up, see ref 2.
(10) Strauss, C. R. Org. Process Res. DeV. 2009, 13, 10.1021/op900194z.
10.1021/op900287j  2010 American Chemical Society
Published on Web 01/15/2010
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
205

now exists refuting any nonthermal microwave effects,11
efficient stirring of reactions should negate any microwave
penetration issues. While a batch microwave reactor may not
be the solution when tons or hundreds of tons of a desired
compound must be synthesized yearly, a batch microwave
reactor capable of processing material at the kilogram scale
would help bridge the gap between a small-scale protocol and
larger kilo- or pilot-plant scale, adding a much needed tool to
the process chemist’s toolbox. However, as Moseley and co-
workers recently pointed out, “there is at present no single
commercially available scale-up reactor capable of meeting the
needs of the pharmaceutical industry for the wide range of
reactions typically required on >1 kg scale.”2
Figure 1. (a) CAD rendering of the prototype reactor used in
the current work. (b) CAD rendering of the unit that will soon
be commercially available.
In the development of new chemical entities (NCEs), the
transition from the medicinal chemistry route to a scale where
enough material can be prepared to carry out initial in ViVo
toxicology studies is often the most difficult.12 Besides the fact
that the initial route only had to be efficient enough to obtain a
few milligrams of the NCE, medicinal chemists often use
chemistry or methodologies, including microwave-mediated
transformations, that are currently difficult to scale to the size
necessary to obtain 1-5 kg of the NCE. Due to the prevalence
of scientific microwave apparatus in medicinal chemistry
laboratories, process chemists often are delivered protocols that
used microwave heating in one or more steps. Thus, those
charged with the duty of the initial scale-up often are faced
with the difficult choice between running multiple reactions for
the microwave-mediated steps to achieve desired throughput
or having to develop modified conditions to avoid microwave
heating. This second option may mean lowering temperatures,
lengthening reaction times and employing different solvents and/
or higher catalyst loadings. In the worst-case scenario, process
chemists have to redesign the entire route to the target
compound in order to avoid the microwave-mediated steps
within a reaction sequence. As the likelihood of one of any
given 20-50 promising NCEs becoming the active pharma-
ceutical ingredient in a new drug is low, companies have a
vested interest in pushing any reoptimization of a route to the
target molecule to the latest possible stage, ideally only after
screening and toxicology tests have thinned the field to a few
promising candidates.12
A batch microwave reactor capable of running reactions from
2000-12000 mL in a single batch has very recently become
available to our research group for experimental use though is
not yet commercially available. Our research group has screened
a number of pharmaceutically relevant synthetic transformations
using the unit, and we present the results here. We searched
the medicinal chemistry literature and selected microwave-
mediated protocols. Then, to the fullest extent possible, we
scaled them up in a linear fashion, often over 10,000 times that
at which they were first developed. The purpose of our study
was three-fold. First, we wanted to probe the operating
parameters of the microwave apparatus with respect to heating
Figure 2. Interchangeable reaction flasks of (a) 5-L, (b) 9-L,
and (c) 13-L capacity with working volumes of 2-4 L, 4-8 L,
and 7-12 L, respectively.
profiles, maximum temperature and pressure, magnetron ef-
ficiency, and overall effectiveness. Second, we aimed to
determine which reactions were directly scalable and then
scrutinize those where difficulty was met. Finally, we wanted
to evaluate the ease of performing reactions in a batch
microwave reactor at the kilogram scale.
Results and Discussion
Equipment. The microwave unit, designed by AccelBeam
Synthesis, allows for reactions to be performed on scales from
2-12 L. Engineering renderings of the prototype unit used in
our trials, together with those of the unit to be commercially
available soon, are shown in Figure 1. Currently, there are three
interchangeable reaction vessels; 5-L, 9-L, and 13-L glass
vessels with working volumes of 2-4 L, 4-8 L, and 7-12 L,
respectively (Figure 2). A universal cover acts as the interface
for peripherals including a stirring paddle that operates at
speeds from 0-125 rpm, a fiber-optic temperature probe,
the reaction ejection tube, and a port for interfacing
spectroscopic tools. In addition, a small port on the cover
allows for last-minute addition of catalyst, reagents, or
solvent (Figure 3). The desired reaction vessel, equipped
with the cover, is placed in a mechanically sealed stainless
steel reaction chamber capable of operating at pressures up
to 350 psi (24.1 bar). To run a reaction, the chamber is
prepressurized to 250-300 psi (17.2-20.7 bar) using
nitrogen gas via a high-pressure cylinder. This allows access
to reaction temperatures above the normal boiling points of
solvents at atmospheric pressure. The reactor employs three
2.45 GHz water-cooled magnetrons rated at 3 kW each, with
an accessible power of 2.5 kW each for a total maximum
allowed output of 7.5 kW. Reaction parameters such as time,
temperature, pressure, and magnetron power are monitored
and collected with the use of the software provided by the
manufacturer. Additionally, this software allows for operation
of all pressure inlet and release valves. Eventually, it will
also control power modulation of the magnetrons. However,
on this prototype unit, microwave power input was operator-
(11) (a) Schmink, J. R.; Leadbeater, N. E. Org. Biomol. Chem. 2009, 7,
3842. (b) Herrero, M. A.; Kremsner, J. M.; Kappe, C. O. J. Org. Chem.
2008, 73, 36. (c) Razzaq, T.; Kremsner, J. M.; Kappe, C. O. J. Org.
Chem. 2008, 73, 6321. (d) Gilday, J. P.; Lenden, P.; Moseley, J. D.;
Cox, B. G. J. Org. Chem. 2008, 73, 3130.
(12) Federsel, H.-J. Acc. Chem. Res. 2009, 42, 671.
206
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Figure 4. Heating profiles for various solvents heated using
the microwave unit. In all cases 4 L of the desired solvent was
heated using 7.5 kW (2.5 kW × 3) of applied microwave power,
and stirring was set at ∼60 rpm.
Figure 3. Last-minute addition of solvent to the reaction vessel
microwave chemistry actually performed relatively poorly; 4
L of water takes over 9 min to reach 150 °C. An often-
overlooked component to solvent selection is the heat capacity
and, as a result, the required amount of energy to heat it. As a
rough estimate, we calculated the amount of energy required
to heat 4 L of water to 120 °C to be approximately 2.01 × 103
kJ.13 The same volume of ethanol requires less than half that at
924 kJ, and all the other solvents require less than 900 kJ (4 L,
∆ ) 120 °C). Also, the vessel is charged with 280 psi (19.3
bar) N2 prior to heating. In contrast, typical small-scale
microwave reactors operating in sealed-vessel mode generate
autonomic pressure. As a result, when heating with our large-
scale unit, energy loss due to solvent vaporization is minimized.
For example, the calculated boiling point for dichloromethane
at 280 psi (19.3 bar) is 160 °C (atmospheric boiling point )
40 °C).14
The combination of heat capacity and minimal energy loss
due to vaporization cannot fully explain the unexpected results
we obtain. Since the solvents are being heated using microwave
irradiation, their relative microwave absorptivities must have
some impact. The loss tangent (tan ∂) is proportional to the
ability of a material to convert electromagnetic energy into heat.
The loss tangent is a quotient defined as: tan ∂ ) ε′′/ε′; where
ε′′ is the dielectric loss which describes the efficiency of which
microwave energy is converted into kinetic energy and ε′ is
the dielectric constant which indicates the polarizability of a
molecule within the microwave field. Thus, the ability of a
solvent to couple with the microwave field, and in turn result
in heating, is generally proportional to its loss tangent as well
as its dielectric loss constant. However, this is only true if the
entire sample’s cross-section can effectively be irradiated. On
the small scale this is the case, which means solvents that are
highly microwave absorbing will heat faster than those that have
low microwave absorptivities. At larger scales, the depth to
which the microwave energy can penetrate the contents of the
vessel will vary. Less absorbent solvents such as DCM or THF
have a larger cross-section that is absorbing the microwave
through the port on the lid.
controlled via the built-in analog control dials of each of
the three magnetron power sources. Upon completion of a
reaction, a valve in the reaction ejection line is opened to
allow the nitrogen pressure in the reactor to force the contents
out. This flow can be directed through a water-cooled
counter-flow heat exchanger, or directly into a receiving
chamber at ambient pressure.
Solvent Heating. To begin our investigations, we chose to
heat various solvents in the unit, as a practical “back-of-the-
envelope” indication of the effectiveness of the magnetrons and
to aid us in the selection of solvents for reactions. When using
microwave heating, at the outset it would seem obvious to select
highly absorbing solvents in order to keep ramp times short
and achieve elevated temperatures easily. Indeed, it has been
our experience using a wide range of dedicated small-scale
microwave reactors that solvents such as ethanol, dimethylfor-
mamide (DMF), or dimethylsulfoxide (DMSO) heat well, while
solvents such as dichloromethane (DCM) or tetrahydrofuran
(THF) that interact poorly with microwave irradiation require
significantly more power input in order to reach elevated
temperatures. On the basis of conventional wisdom when using
monomode microwave apparatus and our small-scale micro-
wave experience, we expected that heating solvents such as
DCM or THF would require significantly more energy input
than more polar solvents such as ethanol. However, there are
two significant differences between typical monomode micro-
wave reactors and the AccelBeam unit, namely much larger
reaction volumes and vessel prepressurization.
We studied seven solvents with a range of microwave absorp-
tivities: ethanol, water, acetonitrile, ethyl acetate, THF, 2-butanone
(MEK), and DCM. In all cases, 4 L of solvent was heated in the
5-L reaction vessel until they reached 150 °C using a constant 7.5
kW power delivery from the magnetrons (2.5 kW × 3) and with
the stirring set to ∼60 rpm. The results are shown in Figure 4.
It took between 4-5 min to heat 4 L of MEK (4 min 13 s),
DCM (4 min 15 s), acetonitrile (4 min 21 s), THF (4 min 28 s),
or ethyl acetate (4 min 58 s) from 30-150 °C. Perhaps
surprisingly, the solvents that performed the best in this study
are ones often regarded as “poor” microwave absorbers. On
the other hand, those generally touted as good solvents for
(13) Using the heat capacities of solvents at 20 or 25 °C. For a
comprehensive table, see the Supporting Information.
(14) (a) Goodman, J. M.; Kirby, P. D.; Haustedt, L. O. Tetrahedron Lett.
2000, 41, 9879. (b) The embedded applet described in ref 14a can be
found at http://www-jmg.ch.cam.ac.uk/tools/magnus/boil.html.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
207

acted as a platform for our preliminary studies. We have studied
this reaction previously across a range of scales and in a number
of commercial microwave units.4b,16 Furthermore, the reaction
is homogeneous throughout its course, and the starting materials
are readily available and inexpensive. We find microwave
heating to be very beneficial in this reaction as it affords short
reaction times that mitigate byproduct formation.17 It also allows
for use of low loadings of the piperidine catalyst and gives the
chemist flexibility in the choice of solvent used. In small-scale
studies we found that 1 mol % of the piperidine catalyst is
sufficient to afford moderate-to-good isolated yields of the
desired 3-acetylcoumarin when the reaction is carried out in
ethanol at 130 °C for 20 min. In our first attempt to scale up
this reaction in the batch unit, the heat exchanger was used upon
completion. Due to the crystalline nature of the product and its
high concentration in ethanol (1.5 M), the product crystallized
in the efficient heat exchanger, resulting in clogging. We decided
to explore a flash cooling option, whereby the contents of the
reaction were ejected into a 5-gal (∼20 L) vented receiving
flask at ambient pressure situated in a fume hood via a stainless-
steel exit line. A significant quantity of the ethanol evaporated,
thereby rapidly cooling the reaction contents from 130 °C to
approximately 80 °C almost instantaneously. Not only did this
solve the clogging issue but also provided us an unexpected
benefit. Since the rapid evaporation of solvent resulted in flash
cooling, a granular crystal form and very even distribution of
particle size was noted. This greatly facilitated the filtration and
isolation of the product. Allowing the reaction mixture to cool
slowly inside the microwave unit led to flatter, plate-like crystal
morphologies and required longer filtration times. The differ-
ences in the crystal morphologies obtained using each cooling
mode can be seen in Figure 6.
We performed the reaction on three scales; 3 mol, 12 mol,
and 18 mol, in each case running the reaction at a concentration
of 1.5 M. It proved to be very scalable, affording isolated yields
of 67%,18 74%, and 81%, respectively, at the three reaction
scales (Table 1, entry 1). Furthermore, as we have studied this
reaction in a number of microwave units, we are able to
compare heating profiles, in particular ramp times, to other
reaction vessels. Intuitively, ramp times will lengthen as scale
increases. At a volume of 2 L, the ramp to 130 °C from room
temperature took 3 min 48 s. At the 8-L scale the ramp time
was 7 min 8 s, and on the 12-L scale, it was 12 min 45 s (Figure
7). This shows that the unit is capable of heating the reaction
mixture to the target temperature in an expeditious manner, even
when using large volumes. For comparison, running the same
reaction in the Biotage Advancer, the ramp stage takes 1 min
20 s for the much smaller volume of 300 mL.
Figure 5. Illustration showing that microwave irradiation is
effective at heating a wide range of solvents on a large scale,
regardless of microwave absorptivity.
energy as compared to more absorbent solvents such as ethanol
and water. Microwave absorptivity and heating cross section
are interlinked and inversely proportional which means that
dichloromethane (tan ∂ ) 0.042) can be heated to 150 °C from
30 °C (∆T ) 120 °C) in less time than it takes to heat ethanol
(tan ∂ ) 0.941) across the same temperature range using the
same applied microwave power. A graphical representation of
this can be seen in Figure 5.
The ramifications of these observations are significant.
First, many researchers intimately involved with the scale-
up of microwave chemistry, including our lab, have
reluctantly viewed microwave-mediated batch scale-up as
limited in scope. We and others have hypothesized that
flow chemistry would be the only suitable approach for
>1-kg scale-up using microwave heating, citing poor
penetration depth as a potential issue. While there are a
number of excellent reasons to adopt flow chemistry for
large-scale synthesis, they hold true regardless of whether
microwave or conventional heating is used. Our results
show that, with efficient stirring and properly sized and
engineered magnetrons, it should be possible to design
batch reactors capable of effectively heating reactions on
significantly larger scales than currently used.
As a note of caution, in light of our success with heating
solvents such as dichloromethane we attempted to heat
toluene. The microwave energy does not couple effectively
with the solvent because it is sufficiently transparent.
Instead, arcing was evidenced by a rapid increase in
pressure due to decomposition of toluene vapors into
molecular hydrogen and elemental carbon black. The run
was abandoned after 15 s of microwave heating. However,
a larger solvent cross section (e.g., 12 L instead of 4) could
feasibly allow for sufficient absorption of microwave energy
by the solvent and reduce the potential for arcing. Current
investigations are underway. However, at this point we note
that it is unadvisible to carry out reactions in toluene using
nonabsorbing reagents without additives to increase the
absorptivity of the load.15
Preparation of 4-Phenyl-3,4-dihydropyrimidine-2(1H)-
one. The Biginelli condensation among a urea or thiourea, an
aldehyde, and a ꢀ-keto carbonyl to afford dihydropyrimidines
(15) (a) See Leadbeater, N. E.; Schmink, J. Tetrahedron 2007, 63, 6764.
(b) Van der Eycken, E.; Appukkuttan, P.; De Borggraeve, W.; Dehaen,
W.; Dallinger, D.; Kappe, C. O. J. Org. Chem. 2002, 67, 7904. (c)
Kremsner, J. M.; Kappe, C. O. J. Org. Chem. 2006, 71, 4651. (d)
Leadbeater, N. E.; Torenius, H. M. J. Org. Chem. 2002, 67, 3145.
(16) (a) Schmink, J. R.; Leadbeater, N. E. Nat. Protoc. 2008, 3, 1. (b)
Schmink, J. R.; Holcomb, J. L.; Leadbeater, N. E. Chem.sEur. J.
2008, 14, 9943.
Preparation of 3-Acetylcoumarin. The condensation of
salicylaldehyde with ethyl acetoacetate to form 3-acetylcoumarin
(17) Michael addition of enolate to coumarin to form a dimer.
(18) Product crystallized in heat exchanger, leading to low isolated yield.
208
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Figure 6. Comparison of 3-acetylcoumarin crystals. The left panel illustrates the granular morphology observed upon flash cooling
from 130 °C to approximately 80 °C. The right panel illustrates the plate-like crystal morphology of the coumarin crystals that is
observed upon slow cooling.
Table 1. Reactions performed in the large-scale batch microwave unita
a For comparison, small-scale reactions and corresponding yields of the protocol are included. When scaling up, reactions were performed using identical conditions of
time, temperature, reaction concentration, and catalyst loadings, differing only in reaction scale and ramp time. All yields represent isolated yields, except where noted.
is widely used in medicinal chemistry. An example of a
pharmaceutically relevant dihydropyrimidine is Monastrol (2a),
and urea as substrates. This reaction has been studied extensively
by Kappe and co-workers using microwave heating.20 Their
recent protocol effects high (86% after column chromatography)
to moderate (76%, isolation by filtration) yields of the desired
products using acetonitrile as a solvent and 10 mol % ytter-
bium(III) triflate as a Lewis acid catalyst.21 However, as the
(19) Mayer, T. U.; Kapoor, T. M.; Haggarty, S. J.; King, W. R.; Schreiber,
S. L.; Mitchison, T. J. Science 1999, 286, 971.
first identified in 1999 by Mayer and co-workers as a potential
lead candidate for new anticancer drugs.19 We chose to examine
the Biginelli reaction using benzaldehyde, ethyl acetoacetate,
(20) (a) Kappe, C. O. Acc. Chem. Res. 2000, 33, 879. (b) Stadler, A.; Kappe,
C. O. J. Comb. Chem. 2001, 3, 624.
(21) Dallinger, D.; Kappe, C. O. Nat. Protoc. 2007, 2, 317.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
209

339 g of 4-methoxycinnamic acid (3), this corresponding to a
95.2% isolated yield (Table 1, entry 3), which compares
favorably with our previous efforts performing the reaction on
the 0.1 mol scale (92% isolated yield).4b We repeated the
reaction a further two times, reducing the TBAB loading from
a full equivalent initially to 0.5 equiv and then 0.25 equiv. In
the case of the former, the isolated yield was comparable to
that in our initial protocol (95% isolated yield). When we
reduced the TBAB loading to 0.25 equiv, the yield suffered
(65% isolated yield). This could be due, at least in part, to the
decreasing solubility of the substrates in the reaction mixture
as the quantity of phase transfer agent is decreased. In addition,
TBAB can stabilize palladium colloids; as a result as we
decrease the quantity used, the palladium species may be
agglomerating, which decreases the catalytic activity.
Figure 7. Heating profiles for the syntheses of 3-acetylcoumarin
on 2-, 8-, and 12-L scales.
Preparation of 4′-Methoxybiphenyl. Another reaction on
which our research group has focused considerable attention is
the palladium-catalyzed Suzuki-Miyaura coupling.25 It is
possible to perform the reaction using very low loadings of
palladium chloride in a water-ethanol solvent mixture. The
power and utility of this protocol becomes increasingly evident
at larger scales, as the use of inexpensive solvents, the
elimination of the need for expensive phosphine ligands, and
the very low palladium loadings make this coupling economi-
cally feasible. Performing the reaction on the 4 mol scale
required a mere 1.6 mg of palladium (0.015 mmol Pd). Heating
the 8 L reaction mixture to 150 °C using 7500 W (2500 W ×
3), took 13.3 min. At this point the magnetron power was
modulated to remain at the desired 150 °C (300 W; 100 W ×
3) for an additional 5 min before ejecting the contents into a
receiving flask. A 91.8% isolated yield of the desired biaryl
(4) was obtained (Table 1, entry 4), corresponding to a catalyst
turnover number of 243,000. Previously, we have performed
this reaction on the 50 mmol scale in a sealed vessel using an
identical protocol.4b Using the larger batch reactor here, we are
able to scale this 40-fold in a linear manner and obtain identical
results.
Performing a Four-Step Reaction Sequence. In order to
simulate a situation where multiple sequential microwave steps
were employed in order to reach a desired target compound,
we developed a sequence of reactions as a medicinal chemist
might at the <10 mmol scale in order to synthesize a drug-like
molecule (Scheme 1). We then scaled up each step in a linear
manner. We began with a base-mediated condensation between
thiourea and ethyl acetoacetate in ethanol to afford 6-methylth-
iouracil using an adaptation of a literature procedure (Table 1,
entry 5).26 On the 1-2 mmol scale we obtained a 70% yield of
the desired product (5). We scaled this up in the batch reactor
initially to 2 mols. The 4-L reaction volume was heated to 125
°C and held for 25 min, a 90.0% yield of 5 being obtained.
We scaled the reaction further to 4 mol (8-L volume), obtaining
an isolated yield of 94.6%. The reaction performed significantly
better at scale when compared to the 1-2 mmol optimization
runs. We attribute this to the fact that, over the course of the
reaction, the sodium salt of the thiouracil product is formed,
scale is increased, the lanthanide catalyst becomes rather cost
prohibitive. We employed 20 mol % HCl in ethanol as an
alternative, similar to the original conditions developed by
Biginelli, obtaining a 55% yield on the 2 mmol scale.22 Working
on the 4 mol scale, heating the reaction to 120 °C from ambient
took just under 6 min at which point the microwave power was
modulated to maintain this temperature for 20 min. After the
reaction time elapsed, the contents of the vessel were ejected
into a receiving flask (without external cooling), and the desired
dihydropyrimidinone 2 was obtained in 56.0% yield without
the need for further purification.
Preparation of 4′-Methoxycinammic Acid. The palladium-
catalyzed Heck reaction represents a powerful methodology for
the formation of carbon-carbon bonds. Furthermore, cross-
coupling reactions have seen an increase in utility by pharma-
ceutical companies over the past 20 years.23 Our group has
examined this transformation extensively at a wide range of
scales in numerous scientific microwave units;4a,b,6,24 thus, we
thought it an excellent choice for study here. Our protocol for
this reaction uses water as the solvent, potassium carbonate as
base, a low loading of ligandless palladium source as the
catalyst, and tetrabutylammonium bromide (TBAB) as a phase
transfer agent. Previous experience has shown us that effective
stirring is critical to the successful outcome of the reaction. Even
at 175 °C the heterogeneous nature of the reaction mixture can
lead to ineffective mass transfer and subsequently low yields.
Working on the 2 mol scale and with a total reaction volume
of 4 L, the microwave unit took approximately 15 min to heat
to the desired temperature of 175 °C. We then aimed to
modulate the microwave power to hold the reaction mixture at
this temperature for 15 min, this taking less than 100 W total
(∼25-30 W × 3). However, as the reaction proceeded, carbon
dioxide was generated due to decomposition of the carbonate
base. The pressure gradually rose and approached the limits of
the vessel after 5 min at 175 °C. The microwave power was
then modulated to stay below the 350 psi (24.13 bar) pressure
limit, and by the end of the 15 min hold time, the reaction
temperature was measured to be 165 °C. Although the tem-
perature was not at optimal for the full 15 min, we isolated
(22) Biginelli, P. Gazz. Chim. Ital. 1893, 23, 360.
(23) Dugger, R. W.; Ragan, J. A.; Ripin, D. H. B. Org. Proc. Res. DeV.
2005, 9, 253.
(25) For a comprehensive discussion on substrate scope of the Suzuki
reaction in aqueous media, see: (a) Leadbeater, N. E.; Williams, V. A.;
Barnard, T. M.; Collins, M. J. Org. Process Res. DeV. 2006, 10, 833.
(b) Leadbeater, N. E. Chem. Commun. 2005, 2881.
(24) Leadbeater, N. E.; Williams, V. A.; Barnard, T. M.; Collins, M. J.
Synlett 2006, 18, 2953–2958.
210
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Scheme 1. Four-step reaction sequencea
a (a) Small-scale: 39% isolated yield over 4 steps. (b) Large-scale: 38% isolated yield over 4 steps (1.82 moles).
and even at 125 °C in ethanol, the product readily precipitates
(Figure 8). On the small scale, a monomode microwave unit
was used, and reactions were stirred using a magnetic stir bar
placed into the vessel. We speculate that this becomes ineffec-
tive as the precipitate forms, thus hindering the reaction and
resulting in lower yields. Conversely, on the larger scale with
mechanical stirring, this problem is mitigated, and an ap-
preciable increase in product yield is achieved. However, when
we attempted to scale up to 6 mols (12 L volume), the quantity
of precipitate proved too significant for the stirring mechanism.
As such, we deemed it prudent to discontinue heating to avoid
the potential of scorching the static load. As such, the reaction
contents reached only 100 °C, where it was held for 30 min
before the contents were ejected, leading to a comparably low
isolated yield of only 74.9%, although still higher than the small-
scale, magnetically stirred reactions.
In the second step of our reaction series, we performed the
selective benzylation of the sulfur functionality on 6-methylth-
iouracil (Table 1, entry 6) to form benzylthiouracil (6). We built
on a protocol reported by Botta and co-workers, where they
used microwave heating (130 °C for 5 min), an equivalent of
either of the halobenzyl compounds of choice, DMF as solvent,
and K2CO3 as a scavenger base. We found that using benzyl
chloride and heating the reaction mixture to 100 °C and holding
for 25 min was optimal in our case. Performing the reaction
on the 8.4 mmol scale using a monomode microwave unit, a
70% isolated yield of 6 was obtained. A comparable yield
(64.0%) was obtained upon scaling up to 3.4 mols using the
batch reactor.
A POCl3/Et3N deoxychlorination protocol was next per-
formed (Table 3, entry 7). On the small scale, a 3.0 M
suspension of 6 in POCl3 was prepared, to which 1 equiv of
triethylamine was added. At this point, the reaction was heated
to 120 °C for 10 min. A careful quench of the reaction with
cold, saturated bicarbonate solution followed by extraction with
ethyl acetate afforded 88% isolated yield of the 3-chloropyri-
midine (7) at the 3.4 mmol scale and 98% isolated yield at the
14.7 mmol scale, both with purities >95%. However, after
calorimetry studies (at the 9.0 mmol scale) indicated an
exotherm of approximately 120 kJ/mol upon addition of the
amine to the reaction mixture and because the chlorination
reaction proceeds smoothly at lower temperatures albeit with
longer reaction times, we opted to carry out the reaction at lower
temperatures in separate batches, using conventional heating.
This allowed us to use the exotherm to heat the reaction initially,
and then heat to 100 °C and hold at this temperature until TLC
indicated reaction completion about 90 min at 100 °C, giving
a total reaction time of 2 h.
The final step of our multistep sequence was an acid-
catalyzed nucleophilic aromatic substitution (SNAr) protocol,
directly following a literature procedure, originally developed
on the 0.1 mmol scale.27 Aniline, 7, and 1 equiv of acetic acid
was heated in dioxane to 150 °C and held for 10 min. On the
1 mmol scale a 91% yield of 2-(benzylthio)-6-methyl-4-
(phenylamino)pyrimidine hydrochloride (8) was obtained. Scal-
ing to 1.84 mols using the batch reactor afforded a 75% yield
of 8, representing a greater than 18,000-fold increase in scale
over the original published procedure but with no changes made
(Table 1, entry 8).
Overall, the sequence employed three microwave steps and
afforded 473 g of 8 in 38% overall yield for four steps from
thiourea and ethyl acetoacetate (Scheme 1b). This is almost
identical to the 39% overall yield obtained on the small scale
(Scheme 1a). Thus, using three identical reaction steps (out of
4) with no adjustment other that quantity of reagents used for
a sequence of four reactions developed at the 1 mmol scale,
comparable results were obtained, making this of great potential
benefit to the process chemist.
Summary
We have explored the scope for scale-up of organic
chemistry using a large batch microwave reactor. The unit is
capable of processing 2-12 L per batch, allowing the process
Figure 8. The sodium salt of 7 precipitates during the course
of the reaction. Upon completion, the solvent is ejected, rapidly
cooling the product which remains in the vessel as a spongy
solid that is easily isolated by filtration.
(26) Mai, A.; Artico, M.; Sbardella, G.; Massa, S.; Novellino, E.; Greco,
G.; Loi, A. G.; Tramontano, E.; Marongiu, M. E.; La Colla, P. J. Med.
Chem. 1999, 42, 619.
(27) Wu, T. Y. H.; Schultz, P. G.; Ding, S. Org. Lett. 2003, 5, 3587.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
211

chemist access to the multimole and >1 kg scale. Initial studies
into heating efficiencies illustrate the ability of the unit to heat
a wide range of solvents effectively. The results also show that
penetration depth was not an issue when heating larger volumes
of solvents. A range of reactions have been scaled up by using
the unit, and these include homogeneous condensations, het-
erogeneous reactions, and phosphine-free palladium-catalyzed
transformations carried out in water. The unit, with a microwave
output of 7.5 kW, allowed us to heat reaction mixtures to the
target temperature in an expeditious manner. Of note is that no
reaction we examined required more than 450 W (150 W × 3)
to keep the reaction at the desired temperature, regardless of
volume, reagents, or solvent. The batch microwave unit
performs on par with its smaller monomode cousins, allowing
us to scale reactions without the need for reoptimization of
conditions. This should prove a boon to process chemists faced
with performing microwave-mediated reactions originally de-
veloped at small scale. In addition, this unit should prove an
effective tool for the development of kilo-scale microwave
chemistry.
approximately 6.5 min. At this point the magnetron power was
modulated to remain at the desired 130 °C (300-600W;
100-200 W × 3) for 20 min. After this time, microwave
heating was stopped, and the contents were ejected without
cooling into a 5-gal (∼20 L) receiving vessel containing 1500
mL ethanol. The product was allowed to cool, and the solid
was collected via vacuum filtration and rinsed with ethanol.
The solid was then allowed to dry overnight in an oven (100
°C) to yield 576 g (55.4%) of 2. 1H NMR (300 MHz, DMSO-
d6) δ: 9.19 (s, 1H), 7.74 (s, 1H), 7.32-7.26 (m, 5H), 5.16 (s,
1H), 3.98 (q, 2H, J ) 7.1 Hz), 2.26 (s, 3H), 1.09 ppm (t, 3H,
J ) 7.1 Hz). 13C NMR (75.5 MHz, DMSO-d6) δ: 165.8, 152.6,
148.8, 145.3, 128.8, 127.7, 126.7, 99.8, 59.6, 54.4, 18.2, 14.5.
Preparation of 4′-Methoxycinammic Acid on the 2 mol
Scale. To the 5 L reaction vessel was added 4-bromoanisole
(250 mL, 2.0 mol), methyl acrylate (360 mL, 4.0 mol), and
tetrabutylammonium bromide (161, 322, or 644 g; 0.5, 1.0, or
2.0 mol; 0.25, 0.5, or 1.0 equiv). A 3.7 M aqueous solution of
potassium carbonate (2 L solution, 415 g, 7.4 mol) was added
to the reaction flask. The lid was placed on the reaction, and
the fiber optic temperature probe was inserted into the reaction
mixture. The stirring paddle was fitted to the motor, and the
ejection tubing was inserted into the reaction vessel. A second
solution was prepared containing 4 mL of a 1.006 mg/mL
palladium stock solution diluted to 1 L with deionized water.
This second solution was added to the reaction vessel, and the
reaction chamber was securely closed and prepressurized to 280
psi (19.3 bar) with nitrogen. The reaction was heated to 175
°C using 7500 W (2500 W × 3), the ramp time taking
approximately 11.5 min. At this point the magnetron power
was modulated to remain at the desired 175 °C (300 W; 100
W × 3) for 15 min. At the end of the reaction, the solvent was
ejected into a receiving flask containing 2 L water, allowed to
cool to ambient temperature, and acidified using concentrated
HCl to a pH ) 2. The resulting solid was filtered under vacuum,
washed with 2 L water, and dried at 100 °C overnight to yield
339 g (95.2%, >95% purity by 1H NMR). The white solid (3)
was recrystallized from ethanol. 1H NMR (300 MHz, CDCl3)
δ: 12.2 (bs, 1H), 7.62 (d, 2H, J ) 9 Hz), 7.59 (d, 1H, J ) 16
Hz), 6.93 (d, 2H, J ) 9 Hz), 6.40 (d, 1H, J ) 16 Hz), 3.77 (s,
3H). 13C NMR (75 MHz, CDCl3) δ: 168.3, 161.4, 144.2, 130.3,
127.3, 117.0, 114.8, 55.7.
Experimental Section
Preparation of 3-Acetylcoumarin on the 12 mol Scale.
To the 9 L reaction flask was added 1.26 L (12.0 mols, 1,470
g) salicylaldehyde and 1.56 L (12.0 mols, 1560 g) ethyl
acetoacetate. This mixture was diluted to 8.00 L with ethanol
(1.50 M). The lid was placed on the reaction and the fiber optic
temperature probe was inserted into the reaction mixture. The
stirring paddle was fitted to the motor and the ejection tubing
was inserted into the reaction vessel. At this point, piperidine
(120 mmol, 10.2 g) was added through a small access port.
The reaction chamber was securely closed and prepressurized
to 280 psi (19.3 bar) with nitrogen. The reaction was heated to
130 °C using 7500 W (2500 W × 3), the ramp time taking
approximately 7 min (4 min for 2 L scale, 13 min for 12 L
scale). At this point the magnetron power was modulated to
remain at the desired 130 °C (300-600 W; 100-200 W × 3)
for 20 min. After this time, microwave heating was stopped,
and the contents were ejected without cooling into a 5-gal (∼20
L) receiving vessel containing 1500 mL ethanol; 1.664 kg
(73.7%) 3-acetylcoumarin (1) was collected via vacuum filtra-
tion, rinsed with 1000 mL ethanol, and allowed to dry at
ambient temperature for 3 days. 1H NMR (300 MHz, CDCl3)
δ: 8.51 (s, 1H), 7.67 (m, 2H), 7.39 (m, 2H), 2.73 ppm (s, 3H).
13C NMR (75.5 MHz, CDCl3) δ: 195.4, 159.2, 155.3, 147.4,
134.4, 130.2, 125.0, 124.5, 118.2, 116.7, 30.5.
Preparation of Dihydropyrimidinone 2 on the 4 mol
Scale. To the 5 L reaction flask was added ethyl acetoacetate
(4.4 mols, 557 mL), urea (4.0 mols, 240 g), and benzaldehyde
(4.0 mols, 404 mL), and the mixture was diluted to 4 L with
ethanol (1.0 M). Without waiting for the urea to dissolve, the
lid was placed on the reaction, and the fiber optic temperature
probe was inserted into the reaction mixture. The stirring paddle
was fitted to the motor, and the ejection tubing was inserted
into the reaction vessel. At this point, 67 mL 12.0 M HCl (aq,
20 mol %) was added via an inlet port on the reaction lid. The
reaction chamber was securely closed and prepressurized to 280
psi (19.3 bar) with nitrogen. The reaction was heated to 120
°C using 7500 W (2500 W × 3), the ramp time taking
Preparation of 4′-Methoxybiphenyl on the 4 mol Scale.
A solution was prepared by adding 4-bromoanisole (502 mL,
4.00 mol) to enough ethanol to make 4.0 L of solution. A second
solution containing phenylboronic acid (536 g, 4.4 mol), sodium
hydroxide (320 g, 8.0 mol), and enough water to make 3.8 L
was prepared. A third solution was prepared by diluting 1600
µL of a commercially available palladium stock solution
(Aldrich 207349, 1.001 mg/mL in 5% aq HCl, 1600 µg of
palladium, 15.0 µmol, 0.0004 mol %) to 200 mL with deionized
water. The first two solutions were combined in the 9-L reaction
flask. The lid was placed on the reaction, and the fiber-optic
temperature probe was inserted into the reaction mixture. The
stirring paddle was fitted to the motor, and the ejection tubing
was inserted into the reaction vessel. The reaction chamber was
securely closed and prepressurized to 280 psi (19.3 bar) with
nitrogen. The reaction was heated to 150 °C using 7500 W
212
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

(2500 W × 3), the ramp time taking approximately 13.3 min.
At this point the magnetron power was modulated to remain at
the desired 150 °C (300 W; 100 W × 3) for an additional 5
min. At the end of the reaction, the reaction contents were
ejected into a receiving flask containing 2.0 kg ice. The white
solid was collected via vacuum filtration, rinsed with 2.0 L
water, and dried for three days at 50 °C to yield 669 g of
at 100 °C overnight to yield 510 g (64%) of 6 as an off-white
solid. 1H NMR (500 MHz, DMSO-d6) δ: 7.49 (d, 2H, J ) 7.6
Hz), 7.34 (t, 2H, J ) 7.3 Hz), 7.28 (t, 1H, J ) 7.3 Hz), 5.99 (s,
1H), 5.70 (s, 1H), 4.47 (s, 2H), 2.27 (s, 3H). 13C NMR (125
MHz, DMSO-d6) δ: 164.5, 164.4, 137.9, 129.5, 128.89, 128.87,
127.7, 107.2, 34.1, 23.6.
Conventional Preparation of 2-(Benzylthio)-4-chloro-6-
methylpyrimidine on the 1 mol Scale. A 1-L, three-necked
round-bottom flask was charged with 232 g (1.00 mol) of
S-benzyl-6-methylthiouracil (6) and placed in a heating mantle.
To the flask was added 333 mL of POCl3 (3.6 mol) and fitted
with an addition funnel, an overhead stirring shaft, and a
thermometer. To the stirred solution, Et3N (102.0 g, 1.00 mol)
was added dropwise at a rate to maintain approximately 80 °C,
taking approximately 30 min. The heating mantle was turned
on, and the reaction was heated to 100 °C and held at this
temperature until the reaction reached completion as indicated
by TLC, approximately an additional 90 min. At this point,
heating was discontinued, and the reaction was allowed to cool
to 50 °C. The reaction contents were slowly added portionwise
to 1500 mL of a saturated aqueous solution of NaHCO3 which
was simultaneously being stirred vigorously with an overhead
stirrer, adding additional NaHCO3 as needed (upon cessation
of CO2 generation). Approximately 2 kg of bicarbonate was
needed to completely quench the reaction contents and bring
the pH ≈ 7. This aqueous layer was decanted from the solid
precipitate into a 4.0-L separatory funnel. The precipitate was
rinsed with ∼300 mL of ethyl acetate. The aqueous layer was
extracted with 900 mL of ethyl acetate in three 300-mL
portions. The organic extracts were combined, washed
sequentially with 300 mL of water and 100 mL of
saturated sodium chloride, and dried over MgSO4; the
solvent was removed under reduced pressure to yield 219.5
g (87.5%) of 7 as a reddish brown oil of a purity greater
than 95% (1H NMR) that was used without further
purification in the next step. 1H NMR (300 MHz, CDCl3)
δ: 7.46 (d, 2H, J ) 6.9 Hz), 7.29 (m, 3H), 6.86 (s, 1H),
4.40 (s, 2H), 2.44 (s, 3H). 13C NMR (75 MHz, CDCl3) δ:
172.1, 169.0, 160.7, 137.2, 129.2, 128.4, 127.3, 115.9,
35.5, 23.8.
Preparation of 2-(Benzylthio)-6-methyl-4-(phenylamino)
Pyrimidine Hydrochloride on the 1.8 mol Scale. The 5-L
reaction vessel was charged with 2-(benzylthio)-4-chloro-6-
methylpyrimidine (7) (461 g, 1.838 mol), aniline (172 g, 1.84
mol), and acetic acid (110 g, 1.84 mol) in dioxane (3.5 L
solution, 0.53 M). The lid was placed on the reaction, and the
fiber-optic temperature probe was inserted into the reaction
mixture. The stirring paddle was fitted to the motor, and the
ejection tubing was inserted into the reaction vessel. The reaction
chamber was securely closed and prepressurized to 280 psi (19.3
bar) with nitrogen. The reaction was heated to 150 °C using
7500 W (2500 W × 3), the ramp time taking approximately
11 min. At this point the magnetron power was modulated to
remain at the desired 150 °C (300-600 W; 100-200 W × 3)
for 10 min. After this time, microwave heating was stopped,
and the solution was ejected into a receiving flask containing 2
L of water, leaving behind a spongy solid in the reaction flask.
The solid was allowed to cool and then filtered under vacuum
1
4-methoxybiphenyl, 4 (91.8%). H NMR (400 MHz, CDCl3)
δ: 7.60 (t, 4H, J ) 8.9 Hz), 7.47 (t, 2H, J ) 7.8 Hz), 7.36 (t,
1H, J ) 7.4 Hz), 7.04 (d, 2H, J ) 8.8 Hz), 3.90 ppm (s, 3H).
13C NMR (100 MHz, CDCl3) δ: 159.4, 141.1, 134.0, 129.0,
128.4, 127.0, 126.9, 114.4, 55.6.
Preparation of 6-Methylthiouracil on the 4 mol Scale.
To the 9-L reaction flask was added ethyl acetoacetate (4.0 mols,
506 mL), a suspension of thiourea (5.2 mols, 400 g) in 1 L of
ethanol, and potassium hydroxide (4.0 mols, 404 mL) solution
in 2.5 L of ethanol and was diluted to 8 L with ethanol (0.5
M). The lid was placed on the reaction, and the fiber-optic
temperature probe was inserted into the reaction mixture. The
stirring paddle was fitted to the motor, and the ejection tubing
was inserted into the reaction vessel. The reaction chamber was
securely closed and prepressurized to 300 psi (20.7 bar) with
nitrogen. The reaction was heated to 125 °C using 7500 W
(2500 W × 3), the ramp time taking approximately 10 min (6
min for the 4-L reaction). At this point the magnetron power
was modulated to remain at the desired 125 °C (300-600 W;
100-200 W × 3) for 25 min. At the end of the reaction, the
solvent was ejected into a receiving flask containing 2 L of
water, leaving behind a spongy white solid in the reaction flask,
which was suspended in the ejection contents. The suspension
was adjusted to pH 6 with HCl (conc.) and allowed to cool
overnight. This precipitate was vacuum filtered, suspended in
2 L of water, refiltered, and dried overnight at 100 °C to yield
514 g of white solid. A second crop was collected from the
mother liquor after being allowed to stand at room temperature
for 1 day. It was filtered, washed with 500 mL of water, and
dried overnight at 100 °C to yield a further 24 g of white solid,
bringing the total isolated yield of 6-methylthiouracil (5) to
538 g (94.6% theoretical). 1H NMR (DMSO-d6) δ: 12.25 (bs,
2H), 5.88 (s, 1H), 2.08 (s, 3H). 13C NMR (DMSO-d6) δ: 176.3
161.4, 153.6, 104.1, 18.5.
Preparation of Benzylthiouracil on the 3.42 mol Scale.
6-Methylthiouracil (486 g, 3.42 mol), potassium carbonate (472
g, 3.42 mol), and 5.8 L of N,N-dimethylformamide were added
to the 9-L reaction flask at which point benzyl chloride (394
mL, 3.42 mol) in 1.0 L of N,N-dimethylformamide was added.
The lid was placed on the reaction, and the fiber-optic
temperature probe was inserted into the reaction mixture. The
stirring paddle was fitted to the motor, and the ejection tubing
was inserted into the reaction vessel. The reaction chamber was
securely closed and prepressurized to 100 psi (6.9 bar) with
nitrogen. The reaction was heated to 100 °C using 7500 W
(2500 W × 3), the ramp time taking approximately 5.5 min.
At this point the magnetron power was modulated to remain at
the desired 100 °C (300 W; 100 W × 3) for 25 min. At the
end of the reaction, the solvent was ejected into a receiving
flask containing 2 L of water and acidified using 12 M HCl.
The resulting solid was filtered using vacuum filtration and dried
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
213

and dried overnight at 100 °C to yield the pale-yellow solid
(433 g, 68.4%). The aqueous solution was extracted using ethyl
acetate (4 × 300 mL). The organic extracts were combined,
washed with brine (100 mL), and dried over MgSO4; the solvent
and residual 1,4-dioxane were evaporated under reduced pres-
sure to yield an additional 40.0 g of pale-yellow solid bringing
the total yield of 2-(benzylthio)-chloro-6-methyl-4-(phenylami-
no)pyrimidine to 473 g (74.7%). A small sample of this solid
was combined with 2.0 M NaOH (aq) and extracted using ethyl
acetate. The organic layer was washed with brine and dried
over MgSO4; the solvent was removed under vacuum to afford
Wagner are thanked for technical input. This work was
funded by the National Science Foundation (CAREER
award CHE-0847262), the American Chemical Society
Petroleum Research Fund (45433-AC1) and the University
of Connecticut. The ACS Division of Organic Chemistry
are thanked for a graduate research fellowship sponsored
by GlaxoSmithKline (J.R.S.). We thank Ms. Tzipporah
Kertez for high-resolution mass spectrometry analysis and
Samuel Frueh and Professor Steven Suib for use of
imaging equipment.
1
the freebase of 8. H NMR (500 MHz, CDCl3, freebase) δ:
7.44 (d, 2H, J ) 7.4 Hz), 7.38 (t, 2H, J ) 7.3 Hz), 7.32 (m,
4H), 7.26 (t, 1H, J ) 7.4 Hz), 7.19 (t, 1H, J ) 7.3 Hz), 6.74 (s,
1H), 6.27 (s, 1H), 4.43 (s, 2H), 2.33 ppm (s, 3H). 13C NMR
(125 MHz, CDCl3) δ: 170.7, 166.6, 160.8, 138.17, 138.13,
129.4, 129.0, 128.4, 127.0, 124.8, 122.5, 98.8, 35.1, 24.1.
HRMS: m/z 308.1227, expected: 308.1216.
Supporting Information Available
1
General experimental details as well as H NMR and 13C
NMR spectra for all compounds. This material is available free
of charge via the Internet at http://pubs.acs.org.
Acknowledgment
AccelBeam Synthesis is thanked for access to the batch
microwave unit, and Drs. Hayden Brownell and Richard
Received for review November 3, 2009.
OP900287J
214
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development