Process development for a large scale stereoselective synthesis of (Z)-(1-bromobut-1-ene-1,2-diyl)dibenzene, a key intermediate of a selective estrogen receptor modulator

Article abstract of DOI:10.1021/op100112r

Two efficient large scale syntheses of (Z)-(1-bromobut-1-ene-1,2-diyl) dibenzene are described. The first is a three-step synthetic sequence from trimethyl(phenylethynyl)silane in 63% overall yield. The key transformations involved the stereospecific carb

Full text of DOI:10.1021/op100112r

Organic Process Research & Development 2010, 14, 1147–1152
Process Development for a Large Scale Stereoselective Synthesis of
(Z)-(1-Bromobut-1-ene-1,2-diyl)dibenzene, a Key Intermediate of a Selective
Estrogen Receptor Modulator
Reginald O. Cann,* Robert E. Waltermire, Jihchin Chung,† Matthew Oberholzer,‡ Jiri Kasparec,§ Yun K. Ye, and
Robert Wethman
Chemical Process Research and DeVelopment, Pharmaceutical DeVelopment R & D, Bristol-Myers Squibb Company, One
Squibb DriVe, New Brunswick, New Jersey 08903-0191, U.S.A.
Abstract:
resistant tumors.4 Moreover, due to the estrogen agonistic
activity of tamoxifen, postmenopausal patients undergoing long-
term treatment with tamoxifen are prone to increased incidence
of endometrial hyperplasia cancer.5 There is, therefore, a
substantial unmet medical need for the development of a
pharmaceutical agent that inhibits estrogen induced metastatic
breast cancer without stimulating uterine endometrial tissue
growth.
Two efficient large scale syntheses of (Z)-(1-bromobut-1-ene-1,2-
diyl)dibenzene are described. The first is a three-step synthetic
sequence from trimethyl(phenylethynyl)silane in 63% overall yield.
The key transformations involved the stereospecific carbometa-
lation reaction of trimethyl(phenylethynyl)silane followed by a
bromination. Subsequent Miyaura-Suzuki coupling with phenyl-
boronic acid and transformation of the vinyltrimethylsilane to a
vinyl bromide afforded the target. In an improved synthesis, a
stereoselective nickel acetylacetonate catalyzed PhZnEt addition
to but-1-ynylbenzene, generated an organozincate intermediate,
which was brominated in 58-62% overall yield. A key feature of
this work was the production of highly regiopure olefin. The
optimization effort that resulted in the utilization of substoichio-
metric amounts of Ph2Zn and the safety precautions taken to
facilitate process scale-up are discussed.
(Z)-3-(4-((Z)-1,2-Diphenylbut-1-enyl)phenyl)acrylic acid (1),
is a SERM that was originally discovered by Glaxo-Wellcome6
and was later shown to be effective against estrogen induced
cancers in animal studies. It was also shown to have a low
potential to induce endometrial cancer growth. Further work
demonstrated that 1 has attractive estrogen agonist and antago-
nist properties desirable for the treatment of breast cancer
patients that have failed tamoxifen treatment.7 Moreover, the
low toxicity and the potent activity makes 1 a promising drug
for the treatment of other hormone induced diseases.
As part of the development of a scalable synthesis of 1,8,9
a
Introduction
key issue was the control of isomeric impurities. The specifica-
tion for the (E)-isomer 2, a known estrogen agonist in rats,10
was set at <0.15% for the drug substance. The Glaxo-Wellcome
synthetic procedure6a relied on diastereomer control during the
synthesis of vinyl bromide 3 to control the level of the isomeric
Breast cancer is the most common malignancy in women
and it is estimated that nearly one in nine women will develop
the disease within their lifetimes.1 The current therapy for these
patients include selective estrogen receptor modulators
(SERMs)2 with tamoxifen being the most widely prescribed
medication for estrogen receptor (ER)-positive breast cancer.3
Tamoxifen is a partial estrogen agonist/antagonist SERM
to which nearly 50% of hormone-induced breast cancer patients
respond, but most patients eventually relapse with tamoxifen-
(4) (a) Fisher, B.; Costantino, J. P.; Redmond, C. K.; Fisher, E. R.;
Wickerham, D. L.; Cronin, W. M. J. Natl. Cancer Inst. 1994, 86, 527–
537. (b) Jordan, V. C.; Assikis, V. J. Clin. Cancer Res. 1995, 1, 467–
473.
(5) (a) Fornander, T.; Rutqvist, L. E.; Cedermark, B.; Glas, U.; Mattsson,
A.; Silfverswald, C.; Skoog, L.; Somell, A.; Theve, T.; Wilking, N.
Lancet 1989, 334, 117–120. (b) van Leeuwen, F. E.; Benraadt, J.;
Coebergh, J. W. W.; Kiemeney, L. A. L. M.; Diepenhorst, F. W.; van
den Belt-Dusebout, A. W.; van Tinteren, H. Lancet 1994, 343, 448–
452.
* To whom correspondence should be addressed. Telephone (732) 227-7964.
E-mail: Reginald.cann@bms.com.
† Current address: Department of Chemical Development, Boehringer Ingel-
heim Pharmaceutical, Inc., 900 Ridgebury Road, Ridgefield, CT 06877.
‡ Current address: DuPont Crop Protection, Stine-Haskell 1090 Elkton Rd.,
Newark, DE 19711.
(6) (a) Wilson, T. M.; Henke, B. R.; Momtahen, T. M.; Charifson, P. S.;
Batchelar, K. W.; Lubahn, D. B.; Moore, L. B.; Oliver, B. B.; Sauls,
H. R.; Triantafilou, J. A.; Wolfe, S. G.; Baer, P. G. J. Med. Chem.
1994, 37, 1550–1552. (b) Wilson, T. M. U.S. Patent 5681835;Chem.
Abstr. 1997, 128, 3547.
(7) Wilson, T. M.; Norris, J. D.; Wagner, B. L.; Asplin, I.; Baer, P.; Brown,
H. R.; Jones, S. A.; Henke, B.; Sauls, H.; Wolfe, S. G.; Morris, D. C.;
McDonnell, D. P. Endocrinology 1997, 138, 3901–3911.
(8) Cain, G. A.; Cann, R. O.; Teleha, C. A.; Murphy, D. K. WO
2001077055;Chem. Abstr. 2001, 135, 303676.
(9) (a) Bignon, E.; Pons, M.; Crates de Paulet, A.; Dore, J.-C.; Gilbert,
J.; Abecassis, J.; Miquel, J.-F.; Ojasoo, T.; Raynaud, J.-P. J. Med.
Chem. 1989, 32, 2092–2103. (b) Rubin, V. N.; Ruenitz, P. C.;
Boudinot, F. D.; Boyd, J. L. Bioorg. Med. Chem. 2001, 9, 1579–1587.
(10) (a) The (E,E)-isomer of Tamoxifen is an estrogen agonist in
rats. Jordan, V. C.; Haldemann, B.; Allen, K. E. Endocrinology 1981,
108, 1353–1361. (b) Harper, M. J.; Walpole, A. L. Nature 1966, 212,
87.
§ Current address: Department of Medicinal Chemistry, Respiratory and
Inflammation Centre of Excellence for Drug Discovery, GlaxoSmithKline
Parmaceuticals, 709 Swedeland Road, King Of Prussia, PA 19406.
(1) (a) DeGregorio, M. W.; Weibe, V. J. Tamoxifen and Breast cancer;
Yale University Press: New Haven, CT, 1994. (b) Cancer Facts and
Figures; American Cancer Society: New York, 1997.
(2) (a) For recent reviews on SERMs: Jordan, V. C. J. Med. Chem. 2003,
46, 883–908. (b) Jordan, V. C. J. Med. Chem. 2003, 46, 1081–1111.
(c) Meegan, M. J.; Lioyd, D. G. Curr. Med. Chem. 2003, 10, 181–
210. (d) Bryant, H. U. Endocrine and Metabolism Disorders 2002, 3,
231–241.
(3) (a) Lerner, H. J.; Band, P. R.; Israel, L.; Leund, B. S. Cancer Treat.
Rep. 1976, 60, 1431–1435. (b) Lerner, L. J.; Jordan, V. C. Cancer
Res. 1990, 50, 4177–4189. (c) Howell, A.; Dodwell, D. J.; Anderson,
H.; Redford, J. Ann. Oncol. 1992, 3, 611–617.
10.1021/op100112r  2010 American Chemical Society
Published on Web 09/02/2010
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1147

Scheme 1. Retrosynthetic analysis
Scheme 2. Discovery chemistry synthesis of vinyl bromide 3
impurity 2.11 Vinyl bromide 3 was then elaborated via a
Miyaura-Suzuki coupling followed by a Horner-Emmons
homologation and hydrolysis to produce 1. This general
approach for assembling the tetrasubstituted core of 1 appeared
to be a scalable process upon our consideration of alternative
chemistry. If the proposed chemistry in Scheme 1 was ultimately
successful as a commercial process, only the vinyl bromide 3
would require synthesis, as the other two synthons: (4-
formylphenyl)boronic acid (4) and methyl 2-(dimethoxyphos-
phoryl)acetate (5) are commercially available. The key to the
synthesis would be to develop an efficient, highly stereoselective
synthesis of 3.
3 with no more than 0.3% of the (E)-isomer 9, as we had
demonstrated that this would result in 1 containing <0.15% of
2, our required specification for the API. This paper describes
the development of two efficient, large scale stereoselective
syntheses of 3, the key intermediate in the synthesis of our new
SERM candidate.
Results and Discussion
Early Process. Preparation of (E)-(1-Bromo-2-phenylbut-
1-enyl)trimethylsilane, 7. We perceived the most critical issue
for the preparation of 7 to be the stability of the diethylaluminum
chloride-titanocene dichloride complex. Extensive yield loss
from decomposition occurred when the organometallic complex
was held >3 h, which simulated the hold times we expected in
the pilot plant implementation. This stability problem could be
circumvented by the formation of the complex in the presence
of the starting material. To accomplish this, the titanocene
dichloride was combined with (trimethylsilyl)phenylacetylene
(6) in methylene chloride and a heptane solution of diethyla-
luminum chloride was added. Under these conditions, the
organometallic complex’s existence was transitory as it rapidly
reacted with the acetylene. It was then necessary to add NBS
at -40 °C to limit decomposition from the heat of reaction as
the calculated adiabatic temperature rise was 72 °C, a rather
exothermic reaction.
A secondary problem was the volume efficiency of the
reaction. The initial procedure had a Vmax of ∼118 L/kg of
starting material, primarily due to the volume of methylene
chloride. This was improved by first quenching the reaction
mass in a slurry of Celite, sodium hydroxide, and sodium
bisulfite solution,16 followed by distillation to remove most of
the solvent. The addition of heptane then precipitated the
inorganic byproduct, which was removed by filtration and the
product-rich organic phase could be used directly in the next
step once the aqueous layer was removed. These operational
changes improved the Vmax to ∼40 L/kg. Incorporating these
improvements along with other optimization of reaction pa-
rameters, we converted 13 kg of the acetylene 6 into a total of
The published11,6 sequence to 3 was modified by our
Discovery chemistry group.12 The synthesis of 3 first entailed
the attachment of an ethyl group to the benzyl position of
trimethyl(phenylethynyl)silane (6) by treatment with a mixture
of the diethylaluminum chloride and titanocene dichloride
complex13 (Scheme 2). The vinyl organometallic group was
regiospecifically replaced by bromine by the reaction with
N-bromosuccinimide (NBS) to produce (E)-(1-bromo-2-phe-
nylbut-1-enyl)trimethylsilane (7) in 85% yield after column
chromatography. The Negishi protocol was replaced by a
Miyaura-Suzuki coupling14 of the vinyl bromide with phenyl-
boronic acid catalyzed by Pd(PPh3)4 to provide (Z)-(1,2-
diphenylbut-1-enyl)trimethylsilane, 8, in 85% yield after column
chromatography. Finally, addition of bromine to the vinyl silane,
8, followed by the elimination of TMSBr with sodium meth-
oxide15 produced 3, which was purified by another column
chromatography in 75% yield.
In addition to the elimination of the repeated column
chromatography purifications, which are not solvent efficient
processes and other nonscalable operations, a key challenge for
the development of a viable chemical process was to produce
(11) Miller, R. B.; Al-Hassan, M. I. J. Org. Chem. 1985, 50, 2121–2123.
(12) Unpublished results of DuPont Pharmaceutical Company Discovery
Chemistry group.
(13) (a) Snider, B. B.; Karras, M. J. Organomet. Chem. 1979, 179, C37–
C41. (b) Miller, R. B.; McGarvey, G. J. Org. Chem. 1984, 43, 4424–
4431.
(14) Miyaura, N.; Suzuki, A. Tet. Lett. 1979, 20, 3437–3440.
(15) (a) Miller, R. B.; Reichenbach, T. Tet. Lett. 1974, 543–546. (b) Miller,
R. B.; McGarvey, G. Synth. Commun. 1977, 7, 475–482. (c) Miller,
R. B.; McGarvey, G. J. Org. Chem. 1978, 43, 4424–4431.
(16) Celite was added to aid the filtration of the inorganic by-products.
1148
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development

Scheme 3. Ni(acac)2 catalyzed Ph2Zn addition to acetylenes
20 kg of 7, as a heptane solution of 42.9 wt/wt % potency in
four batches with an overall solution yield of 90%.
Preparation of (Z)-(1,2-Diphenylbut-1-enyl)trimethylsilane,
8. Fortunately, the Miyaura-Suzuki reaction often allows a
wide variety of reaction conditions. We were able to replace
the suspected teratogen solvent dimethoxyethane with ethyl
acetate. The Pd(PPh3)4 catalyst loading was reduced from 5 to
2 mol %, and the reaction was run under nitrogen pressure
(∼1020 mmHg) to enable the reflux temperature to reach 75
°C and thus drive the reaction to completion in only 3 h. The
product mixture was cleaner as compared to the prior chemistry
and allowed the worked-up heptane solution of 8 to be
telescoped once more into the next step. This chemistry scaled
up uneventfully into the plant, and we produced 17.8 kg of 8
as a heptane solution of 41.8 wt/wt % potency in two batches
with an overall solution yield of 92%.
Preparation of (Z)-(1-Bromobut-1-ene-1,2-diyl)dibenzene, 3.
To minimize operator exposure to bromine, we sought to change
the original addition sequence of bromine addition to the
substrate. Instead, the heptane solution of 8 produced in the
last step was added to a -65 °C solution of bromine in
methylene chloride, followed by elimination of TMSBr using
a 25 wt % solution of sodium methoxide to regenerate the
double bond. The bromide 3 was eventually isolated following
workup by precipitation from 2-propanol. This crude product
was subsequently recrystallized from an aqueous methanol
solution to eliminate the last traces of the (E)-regio-impurity 9
as well as any impurities that were carried along the previous
two telescoped steps. We successfully implemented this pro-
cedure in the pilot plant at 8.6 kg input scale in two batches to
produce 13.2 kg of 3 with a yield of 76%.
reaction of PhLi and ZnCl2 in THF;18 however, the subsequent
reactions gave low yields with poor conversion as compared
to those obtained using sublimed Ph2Zn.17a We postulated the
low reactivity of the in situ generated Ph2Zn might be due to
the residual chloride salts in the reaction mixture reacting with
the Ni(acac)2 to generate a less efficient NiCl2 catalyst. An
alternative protocol to generate the Ph2Zn in situ by the reaction
of triphenylborane and diethylzinc19 also gave low yield. On
the basis of these results, we concluded that a salt free Ph2Zn
reagent would be required and purchased the reagent as a THF
solution.20
As we had little latitude on the use of the expensive
diphenylzinc, we explored the possibility for lowering the
charge. Reduction of the equivalents of diphenylzinc from 4 to
1.5 did not impact the yield or purity. Another means to reduce
the equivalents of diphenylzinc would be to utilize all of the
phenyl ligands. As proposed by Knochel,17a the mechanism of
this reaction suggests that only one phenyl group from Ph2Zn
is transferred to the product with the other phenyl group only
serving as a spectator group. We considered the example of
mixed alkyl-aryl-zinc reagents in asymmetric additions to
aldehydes as these zinc reagents selectively transfer exclusively
the phenyl group.21 In analogy, we hypothesized that a mixture
of Ph2Zn and Et2Zn would be effective in our case if an
integrated species equivalent to the expected disproportionation
product EtZnPh21a formed that would enable useful transfer of
all of the phenyl groups.
Overall this three-reaction sequence with a single isolation
step was sufficient to generate enough 3 in the pilot plant to
meet our initial requirements for API. However, the scale of
this campaign was about the largest that could be efficiently
conducted for this process. For the production of larger
quantities of 3, we set as goals (i) the reduction of the length
of the synthesis (three steps) and process cycle time, (ii) removal
of the cryogenic conditions (steps 1 and 2), and (iii) an
alternative for hazardous elemental bromine.
Second Preparative Process. The chemistry used by
Knochel to prepare a closely related analogue, (Z)-(1-iodobut-
1-ene-1,2-diyl)dibenzene (12)17 (Scheme 3) suggested to us an
alternative route that would allow the addition of both the ethyl
and aryl groups over a single operation. The synthesis consisted
of a Ni(acac)2 catalyzed addition of diphenylzinc to pheny-
lacetylenes in a stereospecific fashion at -35 °C to produce
organozincate (11), which was subsequently quenched with
iodine to produce vinyl iodide (12). This would preclude the
necessity of conducting a separate Miyaura-Suzuki coupling
and would leave functionality that could be exploited to append
the final aryl ring.
We were pleased to observed only <3% of the corresponding
diethyl analogue of 3, when 1.5 equiv of Et2Zn was added to
our optimized reaction with 1.5 equiv of Ph2Zn (Scheme 4).
These encouraging results prompted us to further optimize the
reaction, leading to a ∼50% reduction of the zinc reagents (0.7
equiv Ph2Zn and 0.7 equiv Et2Zn) without loss of yield, thus
significantly reducing our cost of goods.
We identified one challenge with this new reaction sequence;
the subsequent bromination with NBS to produce 3 now
generated a difficult to agitate gelatenous precipitate of the
(18) Von dem Bussche-Huennefeld, J. L.; Seebach, D. Tetrahedron 1992,
48, 5719–5730.
An immediate challenge to the utilization of the reaction on
scale was the expense of diphenylzinc. We sought to reduce
the cost of goods by preparing diphenylzinc in situ by the
(19) Rottlander, M.; Palmer, N.; Knochel, P. Synlett 1996, 573–575.
(20) The solution of the salt free diphenylzinc in tetrahydrofuran was
supplied by Albemarle, Baton Rouge, LA, U.S.A.
(21) (a) Nehl, H.; Schmeidt, W. R. J. Organomet. Chem. 1985, 289, 1–8.
(b) Bolm, C.; Muniz, K. Chem. Commun. 1999, 1295–1296. (c) Bolm,
C.; Hermanns, N.; Hildebrand, J. P.; Muniz, K. Angew. Chem., Int.
Ed. 2000, 39, 3465–3467. (d) Fontes, M.; Verdaguer, X.; Sola, L.;
Pericas, M. A.; Riera, A. J. Org. Chem. 2004, 69, 2532–2543.
(17) (a) Studemann, T.; Knochel, P. Angew. Chem., Int. Ed. Engl. 1997,
36, 93–95. (b) Studemann, T.; Ibrahim-Ouali, M.; Knochel, P.
Tetrahedron 1998, 54, 1299–1316. (c) Houpis, I. N.; Lee, J. Tetra-
hedron 2000, 56, 817–846.
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1149

Scheme 4. Addition of PhZnEt to 1-phenylbutyne
Figure 1. Serm API, 1, and its (E)-isomer 2.
Table 1. Screen of brominating reagents
%
ratio
entry
electrophile
conversion of Z:E
1
2
4
5
6
7
bromine
82%
2%
96%
59%
98%
85%
98: 2
100: 0
97: 3
93: 7
97: 3
97: 3
1,2-dibromo-1,1,2,2-tetrafluoroethane
pyridium tribromide
N-bromo acetamide
N,N′-dibromo-5,5-dimethylhydantoin
N-bromosuccinimide
Scheme 5. Optimized reaction of PhZnEt with
1-phenylbutyne
Figure 2. Heat flow plot for solid N,N′-dibromo-5,5-dimethyl-
hydantoin addition.
succinimide byproduct. Other brominating reagents were screened
to avoid this issue and to further improve the yield of the
bromination (Table 1).
N,N′-dibromo-5,5-dimethylhydantoin, a safe and inexpensive
bromine source, proved to be a superior brominating reagent
for this reaction, avoiding the gelling and also improving the
yield of the reaction to 65-75%. Following the usual optimiza-
tion of key reaction parameters, recrystallization led to 3 in 65%
isolated yield and 99.3% purity with only <0.1% diethyl
analogue and none of the (E)-isomer detected (Scheme 5).
A final process scale up challenge became apparent when
the thermochemical evaluation of the new reaction was per-
formed. The heat of reaction of a solution of N,N′-dibromo-
5,5-dimethylhydantoin in THF/NMP (12 L/kg of 10) with the
organozincate 11 was -846 kJ/mol with a calculated adiabatic
temperature (Tad) rise of 247 °C. Furthermore, the solution of
N,N′-dibromo-5,5-dimethylhydantoin in THF/NMP itself would
decompose with an onset temperature of 25 °C and heat of
reaction of -284 kJ/mol. Subsequent ARC (accelerated reaction
calorimetry) analysis showed the exothermic decomposition
reached a maximum self-heat rate of 220 °C/min with a
maximum rate of pressure increase of 100 psi/min occurring at
65 °C. Clearly these were unscalable reaction conditions. Instead
of adding N,N′-dibromo-5,5-dimethylhydantoin as a solution,
the process was changed whereby the N,N′-dibromo-5,5-
dimethylhydantoin was added as a solid. The thermochemistry
for the addition of solid N,N′-dibromo-5,5-dimethylhydantoin
was only marginally better at -572 kJ/mol of 10 with a
calculated Tad rise of 88 °C; however, the heat flow measure-
ments for this addition (Figure 2) indicated that portionwise
addition did provide a dose controlled heat signature.
Figure 3. Solid N,N′-dibromo-5,5-dimethylhydantoin addition
and conversion.
dimethylhydantoin addition (figure 3). The data generated clearly
showed the anticipated increase in product concentration, thus
confirming that the N,N′-dibromo-5,5-dimethylhydantoin rapidly
reacts under the reaction conditions and does not accumulate.
To safely implement the portionwise addition of the N,N′-
dibromo-5,5-dimethylhydantoin into the plant, the following
engineering controls were implemented: (i) an enclosed solids
charging funnel was used to prevent exposure of N,N′-dibromo-
5,5-dimethylhydantoin to water, which would generate bromine,
known to be incompatible with THF;22 (ii) the N,N′-dibromo-
5,5-dimethylhydantoin was added in portions of 0.3 kg/kg of
10 per charge, which limited the potential heat production; (iii)
HPLC analysis was performed during the addition to verify
complete conversion and that there was no accumulation of
reagent; (iv) the plant reactor was cooled to -10 °C between
additions. The plant protocol began by the formation of the
mixed zincate as described above. After the zincate formation
was judged to be complete, the mixture was treated with toluene
and was cooled to -10 °C. The solid N,N′-dibromo-5,5-dimeth-
ylhydantoin (1.5 equiv) was added via the solids addition funnel
in ten equal portions with hold periods for the conversion to
complete and also to allow the reaction mixture temperature to
In addition, the thermochemical data indicated that no reagent
accumulation occurred. To confirm this, HPLC analyses of the
reaction mixture were performed with each N,N′-dibromo-5,5-
(22) Tayim, H. A.; Absi, M. Chem. Ind. 1973, 347–349.
1150
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development

cool back to -10 °C. At the end of the addition, the reaction
mixture was warmed to 20 °C and stirred for 2 h to consume any
remaining starting material followed by aqueous work up and
crystallization. The process was scaled up in the pilot plant to
produce 70.3 kg of 3 in two batches with an average yield of 58%
and 99.6% purity. The process was later transferred to a vendor to
produce an additional 156 kg of (Z)-(1-bromobut-1-ene-1,2-
diyl)dibenzene, 3, in comparable yield and purity. The production
of multiple batches by different plants provided us with sufficient
confirmation that our process for producing 3 was robust.
NBS (6.5 kg, 36.5 mol, 2 equiv) was added in portions over
1 h at <-35 °C. The reaction mixture was warmed to 0 °C
over 2 h and was stirred for 1 h at 20 °C. The reaction mixture
was transferred into a quench solution, which was prepared by
combining water (22 kg), 30% aqueous sodium hydroxide
solution (12 kg, 86.7 mol, 4.66 equiv), sodium sulfite (2.82 kg,
23.3 mol, 1.2 equiv), and Celite 560 (1.6 kg) at 20 °C (caution:
ethane is evolved). The mixture was heated to 50 °C to remove
methylene chloride by distillation. Heptane (41 kg) was added,
and the mixture was cooled to 20 °C. The slurry was stirred
for 15 min and was filtered. The cake was washed with heptane
(10.4 kg) and the combined filtrates were transferred back to
the reactor. The filtrates were allowed to settle and the bottom
aqueous phase was removed. The product rich organic phase
was concentrated by vacuum distillation at 100 mmHg with
the jacket temperature at 60 °C to produce a heptane solution
of 7 (57.5 kg of solution, 8.87 wt % of 7, 90% solution yield).
Three additional batches were performed, and the heptane
solutions of the four batches were combined. The combined
heptane solution was concentrated by distillation to produce 46.6
kg of a heptane solution containing 20.0 kg of 7. A sample
was concentrated in vacuo for characterization: 1H NMR (400
MHz, CDCl3) δ 0.01 (s, 9H), 1.08 (t, J ) 7.5 Hz, 3H), 2.80 (q,
Conclusion
Two processes have been developed to prepare multikilo-
gram quantities of the vinyl bromide 3, the key intermediate
for the synthesis of 1. The early process involved a three-step
sequence starting with trimethyl(phenylethyl-1-nyl)silane (6).
The stereochemistry of the tetrasubstituted olefin was established
by the stereoselective carbometalation reaction with a mixture
of Cp2TiCl2 and Et2AlCl followed by NBS quench of the
organometallic intermediate to generate (E)-(1-bromo-2-phe-
nylbut-1-enyl)trimethylsilane (7). Subsequent Miyaura-Suzuki
coupling with phenylboronic acid produced (Z)-(1,2-diphenylbut-
1-enyl)trimethylsilane (8). An addition-elimination sequence with
bromine followed by sodium methoxide produced the key inter-
mediate 3. This three-step sequence was later replaced with a one-
step reaction based on the Ni(acac)2 catalyzed substoichiometric
diphenylzinc/diethylzinc addition to but-1-ynylbenzene (10). The
generated organozincate was brominated with N,N′-dibromo-5,5-
dimethylhydantoin under well-defined conditions to ensure safety
to form 3. The two processes supported the generation of sufficient
amounts of 1 to support pharmaceutical development activities and
clinical trials.
J ) 7.6 Hz, 2H), 7.25-7.27 (m, 2H), 7.42-7.48 (m, 3H); 13
C
NMR (100 MHz, CDCl3) δ 0.04, 11.15, 34.32, 127.40, 127.52,
127.99, 128.68, 141.85, 156.66. HRMS Calcd for C13H19SiBr:
282.0439 [M+]. Found: 282.0432 [M+].
(Z)-(1,2-Diphenylbut-1-enyl)trimethylsilane (8). A mixture
of potassium carbonate (15.0 kg, 108.53 mol, 3.07 equiv) and
water (31 kg) was stirred for 30 min to dissolve all the solids.
The heptane solution of 7 (23.3 kg solution containing 10.0 kg
of 7, 35.3 mol, 1 equiv) and ethyl acetate (55.0 kg) were added
to the solution. The mixture was cooled to 5 °C, was first inerted
with nitrogen, and was then inerted with argon. Phenylboronic
acid (6.5 kg, 53.3 mol, 1.51 equiv) and tetrakis(triphenylphos-
phine)palladium(0) (0.84 kg, 0.73 mol, 0.02 equiv) were added;
the reaction mixture was inerted again with argon and was
sparged with nitrogen for 15 min. The reaction mixture was
heated under a pressure of 1020 mmHg to 78 °C and was held
for 3 h. HPLC indicated the reaction was complete. All ethyl
acetate was removed by distillation at 760 mmHg with a jacket
set point of 95 °C. The concentrated mixture was cooled to 50
°C and heptane (55.5 kg), methanol (16.1 kg), and water (20
kg) were sequentially added. The mixture was stirred for 20
min and was filtered through a 24 in. plate filter containing a
2-3 in. bed of Celite (3.0 kg) to remove insoluble residues.
The reactor and the filter cake were rinsed with heptane (17.5
kg), and the rinse was combined with the filtrate. The combined
filtrates were allowed to settle for 20 min, and the bottom
aqueous phase was separated. Water (23 kg) and methanol (41.0
kg) were added to the heptane layer. The mixture was stirred
for 20 min and allowed to settle for 20 min, and the bottom
aqueous layer was separated. The heptane layer was washed
again with water (23 kg) and methanol (41 kg) and was
concentrated by vacuum distillation to produce a heptane
solution of (Z)-(1,2-diphenylbut-1-enyl)trimethylsilane (8), (80.0
kg of solution, 12.13 wt % of 8, 92% solution yield).
Experimental Section
All the chemicals and solvents obtained from commercial
sources were used as received without further purification. All
the reactions were performed under a nitrogen atmosphere. The
reversed-phase HPLC technique was utilized to analyze the
chemical and isomeric purities of the compounds. The HPLC
purity of the compounds was performed using a Zorbax SB-
Phenyl, 150 × 4.6 mm column; the mobile phase was 10 mM
NaH2PO4, pH 3.0 buffer (solvent A) and 100% acetonitrile (solvent
B), flow rate 1 mL/min, temperature 40 °C, and detection at 265
nm. The isomeric purity of the compounds was analyzed with a
Hypersil Hypercarb, 100 × 4.6 mm column; the mobile phase was
80% acetonitrile/20% water/0.1% trifluoroacetic acid solution
(solvent A) and 100% tetrahydrofuran/0.1% trifluoroacetic acid
solution (solvent B), flow rate 1.5 mL/min, temperature 40 °C,
and detection at 265 nm.
(E)-(1-Bromo-2-phenylbut-1-enyl)trimethylsilane (7). A
mixture of methylene chloride (88 kg), bis(cyclopentadienyl)ti-
tanium dichloride (Cp2TiCl2, 7.2 kg, 28.7 mol, 1.5 equiv) and
trimethyl(phenylethynyl)silane (6) (3.25 kg, 18.6 mol, 1 equiv)
was cooled to 15 °C, and a 25 wt % heptane solution of
diethylaluminium chloride (13.85 kg, 18.6 mol, 1.5 equiv) was
added over 1 h at <20 °C. The reaction mixture was warmed
to 20 °C and was stirred for 2 h. HPLC indicated all 6 had
reacted. The reaction mixture was cooled to -40 °C, and solid
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1151

A second batch was performed, and the heptane solutions
were combined and concentrated by distillation to produce 42.6
kg of a heptane solution containing 17.8 kg of 8. A sample
was concentrated in vacuo for characterization: 1H NMR (400
MHz, CDCl3) δ 0.02 (s, 9H), 1.05 (t, J ) 7.5 Hz, 3H), 2.47 (q,
J ) 7.6 Hz, 2H), 7.34-7.36 (m, 2H), 7.49-7.56 (m, 5H)
7.59-7.68 (m, 5H); 13C NMR (100 MHz, CDCl3) δ 0.12, 12.82,
30.14, 125.04, 126.74, 127.69, 127.91, 128.08, 128.99, 140.87,
143.92, 144.59, 155.35. HRMS Calcd for C19H24Si: 280.1647
[M+]. Found: 280.1653 [M+].
drofuran (78 kg) sequentially. The mixture was stirred for 30
min to generate a solution. The solution of 10 and nickel
acetylacetonate was transferred to the cold diphenylzinc/
diethylzinc solution over 30 min at <5 °C. The reaction mixture
was warmed to 20 °C and was stirred over a period of 4 h.
HPLC indicated the reaction was complete. Toluene (458 kg)
was charged, and the reaction mixture was cooled to -10 °C.
N,N′-Dibromo-5,5-dimethylhydantoin (145 kg, 507 mol, 1.5
equiv) was charged in 10 portions over 5 h at <10 °C. The
mixture was warmed to 20 °C and was stirred for 2 h. The
mixture was cooled to -5 °C, and a solution of concentrated
hydrochloric acid (106 kg) and water (220 L) was added over
30 min at <10 °C (caution: ethane is evolved). The mixture
was warmed to 20 °C and stirred for 30 min. The phases were
separated, and the organic phase was cooled to -5 °C. A
solution of sodium sulfite (106 kg) and water (420 L) was added
at <10 °C. The mixture was warmed to 20 °C and stirred for
30 min. The phases were allowed to settle, and the bottom
aqueous phase was separated. The product rich organic phase
was washed with 15% aqueous sodium chloride solution (580
kg). The organic solvent was exchanged to 2-propanol by
distillation and addition of a total of 2-propanol (1030 kg) to a
final volume of 420 L. GC analysis indicated the toluene was
0.8 v/v % and the solution was cooled to 20 °C over 2 h to
effect crystallization. Water (160 L) was added over a period
of 30 min, and the slurry was stirred for 1 h. The slurry was
filtered, and the cake was washed with 45% aqueous 2-propanol
(154 kg). The crude product was filtered, and the cake was dried
for 16 h on a Nutsche filter under vacuum with a nitrogen purge.
The crude 3 was charged back to the reactor, 2-propanol (370
kg) was added, and the mixture was heated to 70 °C to generate
a solution. The solution was cooled to 20 °C over 2 h to effect
crystallization, and water (123 kg) was added over 30 min. The
slurry was stirred for 1 h and was filtered, and the cake was
washed with 55% aqueous 2-propanol (220 kg). The wet cake
was dried under vacuum at 70 °C to produce 3 (57.4 kg, 58%
yield, 99.6 area %, 97.8 wt % and 100% (Z)-isomer). 1H NMR
(400 MHz, CDCl3) δ 0.90 (t, 3H, J ) 7.3 Hz), 2.39 (q, 2H, J
) 7.3 Hz), 7.32-7.38 (m, 5H), 7.41-7.48 (m, 5H); 13C NMR
(100 MHz, CDCl3) δ 13.03, 29.46, 118.62, 127.17, 128.15,
128.29, 128.33, 128.88, 140.80, 142.33, 144.73. HRMS Calcd
for C16H15Br: 286.0357 [M+]. Found: 286.0363 [M+].
(Z)-(1-Bromobut-1-ene-1,2-diyl)dibenzene (3). Process A.
A solution of methylene chloride (95 kg) and bromine (7.68
kg, 48 mol, 1.5 equiv) was cooled to -65 °C, and a heptane
solution of 8 (21.3 kg of solution, 8.6 kg of 9, 32.0 mol, 1.0
equiv) was added over 1 h at <-55 °C. The reaction mixture
was held at -55 °C for 30 min at which time HPLC indicated
the bromination was complete. A 25 wt % solution of sodium
methoxide in methanol (13.89 kg of solution, 3.47 kg of
NaOMe, 64.0 mol, 2 equiv) was added over 2 h at <-55 °C.
The reaction mixture was stirred at -55 °C for 1 h and was
warmed to 0 °C over 30 min. The reaction mixture was diluted
with water (24 kg) over 30 min at <10 °C. The mixture was
warmed to 20 °C over 30 min and was stirred for 1 h. The
mixture was allowed to settle, and the phases were separated.
The product rich organic phase was heated to 60 °C to remove
methylene chloride by distillation and exchange the solvent to
2-propanol. The solution was further diluted with 2-propanol
(141 kg) to a final volume of 87 L. After GC analysis indicated
the solvent exchange was complete, the solution was cooled to
35 °C over 3 h. Water was added over 15 min, and the slurry
was stirred for 2 h at 20 °C. The crude product was filtered,
and the cake was dried for 16 h on a Nutsche filter under
vacuum with a nitrogen purge. The crude product was charged
back to the reactor and methanol (85 kg) was added over 15
min. The mixture was heated to 65 °C over 1 h to dissolve the
solids. Water (12 kg) was added over 30 min at >50 °C, and
the solution was cooled to 30 °C over 2 h to effect crystalliza-
tion. Additional water (15 kg) was added, and the slurry was
cooled to 20 °C and aged for 1 h. The slurry was filtered, and
the cake was washed with 50% aqueous methanol (36 kg). The
wet cake was dried under vacuum at 50 °C to produce 3 (6.8
kg, 76% yield, 99.65 area %, 100.2 wt % and 100% (Z)-isomer).
A second batch was performed on an 8.6 kg input scale to
generate a total of 7.2 kg of 3 in 72% yield and 99.7% HPLC
purity area or wt/wt with no (E)-isomer detected.
Acknowledgment
We thank Drs. Jaan A. Pesti, Joerg Deerberg, and Xihua
Qian for useful suggestions during the preparation of this
manuscript. Gratitude is also due to Mr. Michael Peddicord for
assistance with mass spectrometry.
Process B. A solution of 8.3 wt % THF solution of
diphenylzinc (626.5 kg of solution, 52.0 kg of diphenylzinc,
236.7 mol, 0.7 equiv) was concentrated by distillation at
atmospheric pressure to a final volume of 145 L. NMP (90 kg)
was added to the concentrated solution, and the mixture was
cooled to 15 °C. A 20 wt % toluene solution of diethylzinc
(146 kg of solution, 29.2 kg of diethylzinc, 236.7 mol, 0.7 equiv)
was added. The mixture was warmed to 20 °C, stirred at 20 °C
for 30 min, and cooled to -5 °C. To a separate reactor was
charged nickel acetylacetonate (2.2 kg, 8.45 mol, 0.025 equiv),
but-1-ynylbenzene (10) (44 kg, 338 mol, 1 equiv), and tetrahy-
Supporting Information Available
1H and 13C NMR spectra. HRMS analysis data. This material
is available free of charge via the Internet at http://pubs.acs.org.
Received for review April 21, 2010.
OP100112R
1152
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development