A scalable zinc activation procedure using DIBAL-H in a reformatsky reaction

Article abstract of DOI:10.1021/op900192m

The highly exothermic nature of Reformatsky reagent formation and the reported unpredictability of the induction time for its formation pose challenging problems for scaling up Reformatsky reactions. A zinc-activation procedure using DIBAL-H was developed

Full text of DOI:10.1021/op900192m

Organic Process Research & Development 2009, 13, 1094–1099
A Scalable Zinc Activation Procedure Using DIBAL-H in a Reformatsky Reaction
Michael J. Girgis,* Jessica K. Liang, Zhengming Du, Joel Slade, and Kapa Prasad*
Chemical and Analytical DeVelopment, NoVartis Pharmaceuticals Corporation, East HanoVer, New Jersey 07936, U.S.A.
Abstract:
or highly exothermic decomposition reactions can occur. A
temperature runaway, possibly accompanied by an explosion,
can ultimately result.
In the synthesis of multikilogram quantities of an intermedi-
ate for a developmental drug,6 it was necessary to introduce a
chiral center by reaction of imine 1, substituted with chiral
auxiliary tert-butylsulfinamide, with ethyl bromoacetate 2 to
make amine intermediate 3 (Scheme 1).
The highly exothermic nature of Reformatsky reagent formation
and the reported unpredictability of the induction time for its
formation pose challenging problems for scaling up Reformatsky
reactions. A zinc-activation procedure using DIBAL-H was de-
veloped and investigated using reaction calorimetry along with
subsequent parts of the process. This procedure was shown to have
important advantages for scale-up relative to previous zinc
activation methods, including an immediate start of Reformatsky
reagent formation with addition-controlled reaction. Calorimetric
analysis was especially useful in specifying quickly a suitable
temperature for Reformatsky reagent formation. The process was
scaled up successfully.
A suitable process for scale-up of the zinc activation as well
as for formation of the Reformatsky reagent was thus essential.
For such exothermic reactions, a well-defined, detectable start
of reaction on large scale is ideal, since it is then possible to
continue the process safely. Preferably, the reaction should start
when a reagent is added to a vessel and terminate when reagent
addition stops. Such reactions are termed “addition-controlled”.
Processes involving induction periods during addition should
be especially avoided, as reactions that initiate suddenly after
significant amounts of reagent have been added will lead to
large and sudden heat releases (i.e., heat accumulation) and
hazardous conditions.
Because the reaction between the Reformatsky reagent and
1 required a temperature below -5 °C to obtain 3 with high
optical purity, the zinc activation, typically carried at or above
room temperature, had to be performed separately from the
coupling. Thus, procedures similar to the one devised by
Schekenbeek and Siegel,5 involving the addition of a haloester
to a mixture containing the electrophile and activated zinc above
40 °C, though giving an addition-controlled reaction, could not
be applied here. Compound 3 was not isolated, but was
converted to the corresponding amino ester hydrochloride (4)
for characterization purposes (see Experimental Section).
Introduction
The classic Reformatsky reaction comprises the zinc-
mediated formation of a ꢀ-hydroxyester from an R-haloester
and an aldehyde or ketone. Its scope has been broadened to
other electrophiles such as imines.1 The reaction is typically
carried out by first activating zinc using a chemical treatment
to remove the zinc oxide layer from the metal surface and/or
to obtain a high specific surface area of zinc metal. Various
zinc activation methods have been reported,1-3 including
treatment with ultrasound, iodine, chlorotrimethylsilane, sodium
bis(2-methoxyethoxy)aluminumhydride (Red-Al), 1,2-dibro-
moethane, and cuprous chloride, with the latter two reagents
having been used in large-scale reactions.4,5 Reaction between
the activated zinc and the R-haloester to form a zinc enolate
(i.e., the Reformatsky reagent) then occurs. The latter reacts
with the electrophile, with subsequent hydrolysis affording the
product. The electrophile may also be present during Refor-
matsky reagent formation.5
The formation of the Reformatsky reagent is highly exo-
thermic with unpredictable induction times reported in some
cases.5 The unpredictable start of a highly exothermic reaction
poses severe problems on large scale, as a sudden heat release
will lead to a large batch-temperature excursion if cooling
cannot be applied quickly. If the temperature rise is sufficiently
large, pressure buildup (due to rapid solvent vaporization) and/
In this paper, we report on an effective and scalable zinc
activation procedure utilizing diisobutylaluminum hydride
(DIBAL-H). This reagent has been used previously for activa-
tion of magnesium in Grignard reactions.7 All reactions in the
process, especially zinc activation and Reformatsky reagent
* To whom correspondence may be sent. E-mail: michael.girgis@
novartis.com; prasad.kapa@novartis.com.
(1) Ocampo, R.; Dolbier, W. R. Tetrahedron 2004, 60, 9325–9374.
(2) Repic˘, O.; Vogt, S. Tetrahedron Lett. 1982, 23, 2729–2732.
(3) Repic˘, O.; Lee, P. G.; Giger, U. Org. Prep. Proced. Int. 1984, 16,
25–30.
(6) Machajewski, T. D.; Shafer, C. M.; McBride, C.; Antonios-McCrea,
W.: Doughan, B. M.; Levine, B. H.; Xia, Y.; McKenna, M.; Wang,
X. M.; Mendenhall, K.; Zhou, Y.; Gong, B.; Gu, D.; Dolan, J.;
Tulinsky, J.; Brinner, K.; Gao, Z. WO2007041362, 2007; CAN 146:
422010.
(4) Clark, J. D.; Weisenburger, G. A.; Anderson, D. K.; Colson, J.-P.;
Edney, A. D.; Gallagher, D. J.; Kleine, P.; Knable, S. M.; Lantz, M. K.;
Moore, M. V.; Murphy, J. B.; Rogers, T. E.; Ruminski, P. G.; Shah,
A. S.; Storer, N.; Wise, B. E. Org. Process Res. DeV. 2004, 8, 51–61.
(5) Scherkenbeek, T.; Siegel, K. Org. Process Res. DeV. 2005, 9, 216–
218.
(7) Tilstam, U.; Weinmann, H. Org. Process Res. DeV. 2002, 6, 906–
910.
1094
•
Vol. 13, No. 6, 2009 / Organic Process Research & Development
10.1021/op900192m CCC: $40.75  2009 American Chemical Society
Published on Web 09/23/2009

Scheme 1. Introduction of chiral center by Reformatsky reaction between imine 1 and r-bromoester 2
formation, were investigated by reaction calorimetry, allowing
rapid determination of conditions suitable for scale-up.
The -30 to 100 mesh zinc particles were not suspended
completely in the RC1 experiments. Even at a 600 rpm agitation
rate with a pitched-blade impeller, it was not possible to suspend
all the zinc granules, consistent with calculation of the just-
suspended agitation speed Njs value of >8000 rpm from a
literature correlation.8 Calculation of Njs for the pilot-plant
reactor to be used for the scale-up showed that an agitation
rate >1500 rpm would be needed to suspend all the zinc
particles. As such an agitation rate is greater than achievable
values, complete zinc particle suspension would also not occur
on large scale. As will be seen below, good reactivity was still
obtained despite incomplete suspension of zinc particles.
The first task for scale-up was the selection of suitable
temperatures for both zinc activation and Reformatsky reagent
formation. Suitable temperatures should give immediate and
detectable start-of-reaction (indicated by exothermic behavior),
preferably with minimal heat accumulation. In a first experi-
ment, 5.6 g of a 20 wt % solution of DIBAL-H in toluene (0.008
mol of DIBAL-H) was added at 30 °C to a slurry containing
zinc (53.4 g, 0.82 mol), 5% of the total charge of bromoester
2 (Viz., 0.02 mol), and 350 mL of THF solvent. A sharp but
short-lived heat evolution profile resulted (Figure 1), giving an
adiabatic temperature rise of 10 °C, which would be easily
detectable on large scale.
When addition of the remaining solution of bromoester 2 in
THF was started at a reaction time of 0.18 h and a 30 °C batch
temperature (Figure 1), heat evolution did not occur im-
mediately. Instead, heat evolution started at about 0.27 h, and
reached a constant value of about 9 W/kg. Upon heating the
reaction mixture at 1 °C/min to 40 °C, the heat evolution
attained a constant value during the temperature ramp, giving
the profile of an addition-controlled reaction. Thus, a single
calorimetric experiment indicated that Reformatsky reagent
formation did not immediately start when the remaining
bromoester 2 was added at 30 °C (i.e., a ∼5-min delayed
exotherm occurred), suggesting that this temperature was too
low for immediate reagent formation. This result motivated
examining a higher temperature to see if an immediate reaction
start could be obtained.
Results and Discussion
Background. In the original Research procedure, the
Reformatsky reagent was prepared in THF by activation of zinc
dust at reflux with 1,2-dibromoethane, followed by the addition
of the ethyl bromoacetate (2) while maintaining reflux condi-
tions. The mixture was then cooled to -8 °C prior to the
addition of 1.
On scale-up, these conditions did not work well. The yields
were 50-60%, and several impurities were present according
to HPLC analysis. The use of very dry conditions and activated
zinc did not improve the reaction. However, it was found that
one major factor affecting the reaction was the temperature used
for the formation of the Reformatsky reagent. Lowering the
temperature from 65 °C (reflux) to 45-50 °C afforded excellent
conversion with little or no side-product formation. Carrying
out the addition of 1 at -8 °C gave 3 in 98% yield and 99%
optical purity.
Further study revealed that different batches of zinc gave
different results with regard to yield and purity of 3. The reaction
was not reproducible when 1,2-dibromoethane was used to
activate the newer batches of zinc. Since different lots and types
of zinc did not give reliable results, a more reproducible method
for the zinc activation was investigated. It was found that a
catalytic amount of chlorotrimethylsilane was a viable activator
for any grade of zinc. One drawback to this method of zinc
activation was the fact that a minimum of 12 h was needed to
completely activate the zinc. In order to render the process more
efficient for the pilot plant, several other methods of zinc
activation were investigated. DIBAL-H was found to give the
best results in that the activation was instantaneous, or nearly
so, as discussed below.
Zinc Activation with DIBAL-H. The activation procedure
comprises addition of a 20 wt % DIBAL-H solution in toluene
to a slurry of zinc granules (-30 to 100 mesh particle size)
suspended in THF solvent and 5% of the total charge of
bromoester 2. In preliminary experiments performed in ordinary
laboratory glassware, the presence of at least some bromoester
before DIBAL-H addition was necessary to obtain good zinc
activation (i.e., evidenced by detection of an immediate exo-
therm upon adding the majority of bromoester); when DIBAL-H
addition was carried out in the absence of bromoester, a delayed
exotherm was seen when the latter was added subsequently.
The value of 5% of bromoester present before DIBAL-H
addition was selected because it gave good zinc activation, as
described above.
Addition of the second portion of bromoester 2 at 40 °C
(Figure 2) gave immediate heat evolution upon addition, thus
avoiding heat accumulation. In this experiment, DIBAL-H was
added with the RC1 in jacket temperature control mode, causing
the batch temperature to increase by about 2 °C. DIBAL-H
addition at 30 °C in a large-scale vessel, where the heat transfer
(8) Harnby, N., Edwards, M. F., Nienow, A. W., Eds. Mixing in the
Process Industries, 2nd ed.; Butterworth Heinemann: Woburn, MA,
1992; Chapter 16, p 372.
Vol. 13, No. 6, 2009 / Organic Process Research & Development
•
1095

Figure 1. First experiment calorimetric results, with DIBAL-H addition at 30 °C followed by addition of bromoester initially at
30 °C. (The horizontal arrow pointing rightward signifies that the data at the arrow’s origin are to be read on the right-hand axis.)
Figure 2. Calorimetric profile when adding remaining amount of bromoester 2 at 40 °C. In this experiment, the initial bromoester
portion was added immediately before DIBAL-H (i.e., between 0.02 and 0.04 h).
area per unit volume is smaller relative to the RC1 and where
adiabatic behavior would be approached, would give a tem-
perature rise close to the adiabatic temperature rise of ∼11 °C,
thus allowing facile detection on large scale and giving a
temperature near the target temperature for subsequent bro-
moester 2 addition. The addition of the second bromoester
portion was carried out over a 1.5 h period, giving the desired
addition-control calorimetric profile with a maximum heat
evolution rate under 45 W/kg (Figure 2). The latter figure is
well below the estimated 60 W/kg value for maximum heat
removal capability of large-scale (Viz., 10 m3) vessels having
an overall heat transfer coefficient of 300 W/m2 ·K, a heat
transfer area of 20 m2, and a maximum temperature difference
between jacket and batch of 80 °C.9 The heat evolution during
bromoester 2 addition would also be detected easily at large
scale and would thus give a quick indicator of reaction initiation
and progress. The large 130 °C adiabatic temperature rise further
indicates the importance of having an addition-controlled
reaction.
Comparison of these results with those reported earlier for
zinc activation and Reformatsky reagent formation reveals
several advantages of the current process. First, the zinc
activation occurs at or near room temperature, avoiding the need
for heating to reflux, as was reported for activation with 1,2-
dibromoethane2 or cuprous chloride.5 The process time is thus
shortened by at least 2 h, without exposing any reactant to higher
temperatures. Second, no induction period was observed. Thus,
when the reaction was scaled-up to ∼800 L, a single addition
of bromoester was performed, as shown in Figure 2. This
allowed immediate detection of the start of the reaction, avoiding
the need to add the haloester in portions and waiting between
(9) Stossel, F. Course on Thermal Safety; Swiss Federal Safety Institute:
Summit, NJ, June, 1998.
1096
•
Vol. 13, No. 6, 2009 / Organic Process Research & Development

Figure 3. Reformatsky reagent formation after a 1-h delay between zinc activation and the start of addition of second bromoester
portion.
Figure 4. Calorimetric profile during imine addition to Reformatsky reagent.
each addition to verify that reaction occurred or diluting the
reaction mixture, as was done earlier.4,5
start of bromoester addition and the addition-controlled reaction
profile was maintained, further demonstrating process flexibility.
We caution that, while some delay after zinc activation still
leads to a well-defined start of reaction, the impact of longer
interruptions is unknown.
The addition-controlled reaction indicated in Figure 2 also
provided processing flexibility. Thus, if the reaction is carried
out in smaller pilot-plant vessels which have heat removal
capabilities greater than 60 W/kg,9 a faster addition is possible
without compromising process safety. Indeed, when the process
was scaled up to ∼800 L (see the Experimental Section), the
addition could be carried out in 1 h. On the other hand, because
mixing times typically increase with increasing vessel size,
further scale-up may require either slower addition or addition
in portions to give more time for the bromoester to mix and to
thus avoid high local concentrations of the latter. This can lead
to localized high-temperature regions (i.e., hot spots) that could
impact selectivity adversely.
Calorimetric Analysis: Imine Addition and Quench.
Calorimetric analysis was also useful in obtaining process
insight for subsequent steps of the process. After Reformatsky
reagent formation, the reaction mixture was cooled to -8 °C,
and imine 1 was added over 1 h. Heat evolution results indicated
that the reaction was complete in ∼7 h (confirmed by HPLC
analysis), but was not addition-controlled, giving heat ac-
cumulation greater than 50% (Figure 4). Such heat accumulation
would give an adiabatic temperature rise of 13 °C, which would
adversely impact product selectivity, but not pose a safety
concern. A longer addition time could be explored to decrease
heat accumulation and decrease the risk of selectivity loss if
temperature control problems occur.
An experiment was also conducted in which the second
bromoester portion was added 1 h after zinc activation to
simulate the impact of an interruption on large scale. As shown
in Figure 3, reagent formation was not impacted adversely by
the 1-h delay, as heat evolution began immediately after the
The calorimetric analysis was also useful in optimizing the
quench procedure performed by adding brine solution over a
Vol. 13, No. 6, 2009 / Organic Process Research & Development
•
1097

Figure 5. Calorimetric profile during quench with brine.
Figure 6. Schematic diagram of reaction apparatus.
0.75-h period. It was found that the first portions of brine
addition were much more exothermic than subsequent amounts.
To select suitable rates that would give good control on scale-
up, a procedure was developed in which the first half of the
brine solution was added over 30 min, and the remaining
amount added over 15 min. As shown in Figure 5, this
procedure gave an acceptable heat evolution rate with a process
time of 1 h.
A total of six batches of 3 were completed in the pilot plant.
The reactions were carried out at -8 °C overnight to afford 3
with high diastereoselectivity (>99% de). The combined product
was obtained as a 22% MTBE solution in >99% yield and
>95% purity by HPLC (area normalization). The Plant Proce-
dure is given in Experimental Section below.
Reaction calorimetry, discussed in more detail previously,10
is used to measure the instantaneous heat evolved due to
chemical reaction. When a single chemical reaction takes place,
the calorimeter indicates the instantaneous reaction rate and is
thus very useful for determining the start and end of a reaction
as well as the degree of addition control.
General. NMR spectra were recorded on a Bruker AVANCE
DPX-500 spectrometer (1H NMR at 500 MHz). Analytical high
performance liquid chromatography (HPLC) was carried using
a Waters Alliance 2695 separations module, a Waters 996
photodiode array detector (MaxPlot), and a 3.0 × 15 cm Agilent
Zorbax SB C-18 column (3.5 µm particle size). Acetonitrile
was used as mobile phase A and 0.1% trifluoroacetic acid in
water as mobile phase B. A gradient method (10% A to 90%
A over 15 min) was used, at a 1 mL/min mobile phase flow
rate and 40 °C column temperature; The retention times of 1
and 3 were 9.57 and 9.5 min, respectively. Samples for HPLC
analysis were prepared by diluting 1-2 mg of reaction mixture
in 6 mL of methanol; a 2-µL sample size was used. Reactions
were carried out under an atmosphere of nitrogen. Unless
reported otherwise, all reaction temperatures refer to the
Experimental Section
Calorimetry. Reaction experiments were performed using
a Mettler-Toledo RC1 reaction calorimeter equipped with a
jacketed 1-L MP-10 glass vessel, a Hastelloy pitched-blade
impeller, Hastelloy temperature and calibration probes, and a
dosing station (comprising a ProMinent model G4 diaphragm
pump and an electronic balance). A nitrogen purge (28 cm3/
min) was maintained using a Brooks 5880i mass flow controller.
A schematic diagram of the reaction apparatus is shown in
Figure 6.
(10) Girgis, M. J.; Kiss, K.; Ziletener, C. A.; Prashad, M.; Har, D.;
Yoskowitz, R.; Basso, B.; Repic˘, O.; Blacklock, T. J.; Landau, R. N.
Org. Process Res. DeV. 1997, 1, 339–349.
1098
•
Vol. 13, No. 6, 2009 / Organic Process Research & Development

measured temperature of reaction mixtures not to the cooling-
or heating-bath temperatures.
luminum hydride (25% solution in toluene, 1.0 kg, 1.7 mol)
was added from a cylinder over 5 min using nitrogen pressure.
The addition was slightly exothermic. The contents were heated
to 41 °C over 5 min. Ethyl bromoacetate (13.9 kg, 83.2 mol)
was added over 1 h via diaphragm pump while maintaining
the batch temperature below 45 °C. The lines were rinsed with
THF (3 kg) which was added to the batch. The contents were
stirred at 40 °C for 1 h and then cooled to -10 °C. A solution
of (S)-N-[(2-bromo-4-fluorophenyl)methylene]-tert-butylsulfi-
namide in THF (15 kg, 49 mol in 42.9 kg of solution) was
added over 1 h via diaphragm pump while maintaining the
temperature e-5 °C. The lines were rinsed with THF (1 kg)
which was added to the batch. The reaction mixture was stirred
at -8 °C for 15 h. Sodium chloride solution (25%, 4.3 kg)
was charged to the batch over 1 h while maintaining the
temperature e-5 °C. After all of the sodium chloride solution
had been added, the mixture was stirred for 15 min at -5 °C.
The batch was warmed to 20 °C over 30 min. A slurry of Celite
Hyflo (5 kg) in ethyl acetate (30 kg) was added, and the
heterogeneous mixture was stirred at 20 °C for 15 min. The
contents were filtered using a Nutsche filter, and the filter
cake was washed with ethyl acetate (83 kg). The filtrate and
wash were combined and washed with 17% citric acid solution
(78 kg) and 25% sodium chloride solution (65 kg). The organic
solution was concentrated to 100 L by vacuum distillation
(25-30 °C/400-600 mbar). The concentrate was diluted with
methyl tert-butyl ether (97 kg), and the distillation was
continued under the same conditions to a final volume of 40 L.
This step was repeated until the water content by Karl Fisher
was <0.3%. The solution of (ꢀ-R)-ethyl 2-bromo-ꢀ-[[(S)-(1,1-
dimethylethyl)sulfinyl]amino]-4-fluorobenzene-propanoate (3)
was used as is in the next step.
Laboratory Procedure. (ꢀ-R)-Ethyl 2-Bromo-ꢀ-[[(S)-(1,1-
dimethylethyl)sulfinyl]amino]-4-fluorobenzene-propanoate (3).
The RC1 calorimeter was charged with Zn (granular, -30-100
mesh, 53.4 g, 0.82 mol) and dry THF (311.2 g). The agitator
speed was set to 600 rpm and the internal temperature to 30
°C. Ethyl bromoacetate (2) (3.4 g) and 5.81 g of a 20 wt %
solution of DIBAL-H in toluene were added. The suspension
was warmed to 40 °C, and ethyl bromoacetate (2) (64.55 g,
0.39 mol) was added over 1.5 h. The reaction mixture was
maintained at 40 °C for 50 min. The reaction mixture was
cooled to -8 °C, and a solution of (S)-,N-[(2-bromo-4-
fluorophenyl)methylene]-tert-butylsulfinamide (1) (70 g, 0.23
mol) in THF (133.4 g) was added over 1 h. The reaction mixture
was stirred at -8 °C for 14 h. At -8 °C saturated sodium
chloride solution (20 g) was added over 1 h followed by ethyl
acetate (225 g). The mixture was transferred by vacuum to a
flask, and the reactor was rinsed with ethyl acetate (225 g). A
solution of citric acid (60 g) in water (300 g) was added to the
combined solution, and the mixture was stirred at room
temperature for 20 min. The clear solution was filtered through
Celite, and the filter cake was washed with ethyl acetate (100
g). The combined layers were separated, and the upper organic
layer was washed with saturated sodium chloride (300 g). The
organic layer was diluted with tert-butyl methyl ether (MTBE)
(450 g), and the solvent was removed by distillation at 40 °C/
400 mbar to give ∼500 mL of concentrate. The concentrate
was diluted with MTBE (450 g), and the distillation was
continued to give a solution containing 22% by weight of (ꢀ-
R)-ethyl 2-bromo-ꢀ-[[(S)-(1,1-dimethylethyl)sulfinyl]amino]-4-
fluorobenzenepropanoate (3) (113 g of 3 in 408 g of MTBE,
100%): 1H NMR (500 MHz, CDCl3) δ 7.34 (m, 2H), 7.23 (m,
1H), 6.97 (m, 1H), 5.09 (m, 1H), 4.92 (s, 1H), 4.04 (m, 2H),
2.82 (m, 2H), 1.17 (m, 9H), 1.14 (t, 3H); MS (positive-ion DCI)
394 (MH+).
(ꢀ-R)-Ethyl ꢀ-Amino-2-bromo-4-fluorobenzene-propanoate
Hydrochloride (1:1) (4). A solution of (ꢀ-R)-ethyl 2-bromo-ꢀ-
[[(S)-(1,1-dimethylethyl)sulfinyl]amino]-4-fluorobenzene-pro-
panoate (3) (260 g of solution containing 45 g of 3) was added
to a solution of 5-6 N HCl/isopropanol solution (61.5 g, ∼0.34
mol) over a 30-min period. The reaction mixture was stirred at
20-23 °C for 18 h. The resulting yellow suspension was
filtered, and the filter cake was washed with MTBE (3 × 150 g)
and dried at 35 °C/20 mm to give (ꢀ-R)-ethyl ꢀ-amino-2-bromo-
4-fluorobenzene-propanoate hydrochloride (1:1) (4, 23.2 g,
62%): 1H NMR (500 MHz, DMSO-d6) δ 8.91 (bs, 2H), 7.95
(m, 1H), 7.65 (m, 1H), 7.40 (m, 1H), 4.96 (m, 1H), 4,02 (q,
2H), 3.24 (m, 1H), 3.07 (m, 1H), 1.10 (t, 3H); MS (positive-
ion DCI) 291 (MH+).
Conclusions
A zinc-activation procedure using DIBAL-H was developed
for a Reformatsky reaction. The zinc activation and subsequent
Reformatsky reagent formation were investigated by reaction
calorimetry, which was especially useful in selecting a reaction
temperature giving immediate reaction start. The process was
shown to have key advantages for large-scale production over
previously reported processes, including: (a) room-temperature
and easily detectable zinc activation, (b) the absence of an
induction period for Reformatsky reaction formation, and (c)
processing flexibility in conducting Reformatsky reagent forma-
tion. Calorimetric analyses were also useful for optimizing the
quench procedure.
Acknowledgment
We thank Mr. James Vivelo, Dr. Kurt Konigsberger, and
Mr. Robert DuVal for valuable discussions regarding scale-up
and process safety.
Plant Procedure. (ꢀ-R)-Ethyl 2-Bromo-ꢀ-[[(S)-(1,1-di-
methylethyl)sulfinyl]amino]-4-fluorobenzene-propanoate (3). A
slurry of granular Zn (11.4 kg, 174.9 mol) and THF (67.6 kg)
was warmed to 32 °C and ethyl bromoacetate (0.73 kg, 4.4
mol) was added followed by a THF rinse (2 kg). Diisobutyla-
Received for review July 26, 2009.
OP900192M
Vol. 13, No. 6, 2009 / Organic Process Research & Development
•
1099