Catalytic, Asymmetric α-Chlorination of Acid Halides

Article abstract of DOI:10.1021/ja039046t

We present a full account of a tandem catalytic, asymmetric chlorination/esterification process that produces highly optically enriched α-chloroesters from inexpensive, commercially available acid halides using cinchona alkaloid derivatives as catalysts and polychlorinated quinones as halogenating agents. We have performed kinetics and control experiments to investigate the reaction mechanism and establish conditions under which the reactions can be best performed. We have developed NaH and NaHCO₃ shuttle base systems as the easiest and most cost-effective ways of conducting the reactions, rendering the methodology economically competitive with known chiral halogenation procedures. We have also demonstrated the utility of our reactions by converting the products to synthetically useful derivatives.

Full text of DOI:10.1021/ja039046t

Published on Web 03/13/2004
Catalytic, Asymmetric r-Chlorination of Acid Halides
Stefan France, Harald Wack, Andrew E. Taggi, Ahmed M. Hafez, Ty R. Wagerle,
Meha H. Shah, Crystal L. Dusich, and Thomas Lectka*
Contribution from the Department of Chemistry, New Chemistry Building, Johns Hopkins
UniVersity, 3400 North Charles Street, Baltimore, Maryland 21218
Received October 14, 2003; E-mail: lectka@jhu.edu
Abstract: We present a full account of a tandem catalytic, asymmetric chlorination/esterification process
that produces highly optically enriched R-chloroesters from inexpensive, commercially available acid halides
using cinchona alkaloid derivatives as catalysts and polychlorinated quinones as halogenating agents. We
have performed kinetics and control experiments to investigate the reaction mechanism and establish
conditions under which the reactions can be best performed. We have developed NaH and NaHCO3 shuttle
base systems as the easiest and most cost-effective ways of conducting the reactions, rendering the
methodology economically competitive with known chiral halogenation procedures. We have also
demonstrated the utility of our reactions by converting the products to synthetically useful derivatives.
Introduction
Halogenations of organic molecules constitute a synthetically
important and historically significant class of chemical reactions.
Every student of organic chemistry learns from the earliest stages
of his chemical education that diatomic halides are among the
most ubiquitous (and reactive) of halogenating agents. As far
as many textbooks are concerned, the story seems to stop there.
Diatomic halides are known to be highly reactive and generally
nonselective species, proving to be incompatible with a variety
of functionality within complex molecules, as well as being
unsuitable reagents for enantioselective reactions. Fortunately,
in recent years, a vast number of milder, alternative halogenating
agents have appeared in the literature,1 making the study of
stereoselective, catalyzed halogenations more feasible. Recently,
we have become interested in the possibility of catalytic,
enantioselective halogenation reactions. Specifically, R-halo-
genations of carbonyl compounds,2 which historically have
employed diatomic halides as an electrophilic source of halogen,
posed an appealing challenge. For example, enantioselective
R-halogenation could result in compounds that serve as inter-
mediates in the construction of valuable functionalized mol-
ecules.3 In this full paper, we report our work on catalytic,
enantioselective R-chlorination reactions (eq 1) that utilize
inexpensive acid chlorides as starting materials and chiral
nucleophiles as catalysts, as well as employing polyhalogenated
quinones as finely electronically tuned chlorinating agents.
Much recent work in the field of selective R-halogenation of
carbonyl compounds has centered on the development of mild
sources of electrophilic halogenating agents. Electrophilic
R-fluorination has been the most intensely developed, with
reagents (including Selectfluor) attaining widespread use.4 Many
stereoselective R-halogenations have been achieved using chiral
auxiliary-based methods. For example, Evans has developed an
auxiliary-based route to R-bromoimides (eq 2).5
A diastereoselective chlorination using diacetone glucose as
the chiral auxiliary has also been published;6 similarly, a
diastereoselective bromination was reported using chiral gly-
cosides.7 Also, while not technically an R-halogenation, Durst
(1) (a) Rappe, C. Acta Chem. Scand. 1968, 22, 219-230. (b) Larock, R. C.
ComprehensiVe Organic Transformations, 2nd ed.; Wiley-VCH: New York,
1999. (c) Patai, S. The Chemistry of the Carbon-Halogen Bond, Part 1;
Wiley: New York, 1988.
(2) (a) House, H. Modern Synthetic Reactions, 2nd ed.; W. A. Benjamin: New
York, 1972; pp 459-478. (b) De Kimpe, N.; Verche, R. The Chemistry of
(4) (a) Mun˜iz, K. Angew. Chem., Int. Ed. 2001, 40, 1653-1656. (b) Kim, D.
Y.; Park, E. J. Org. Lett. 2002, 4, 545-547.
R-Haloketones, R-Haloaldehydes, and R-Haloimines; John Wiley
&
(5) Evans, D. A.; Ellman, J. A.; Dorow, R. L. Tetrahedron Lett. 1987, 28,
1123-1126.
Sons: New York, 1988.
(3) (a) March, J. AdVanced Organic Chemistry: Reactions, Mechanisms, and
Structure, 4th ed.; John Wiley & Sons: New York, 1992. (b) Carey, F.
A.; Sundberg, R. J. AdVanced Organic Chemistry, 3rd ed.; Plenum: New
York, 1990.
(6) (a) Duhamel, L.; Angibaud, P.; Desmurs, J. R.; Valnot, J. Y. Synlett 1991,
807-808. (b) Angibaud, P.; Chaumette, J. L.; Desmurs, J. R.; Duhamel,
L.; Ple, G.; Valnot, J. Y.; Duhamel, P. Tetrahedron: Asymmetry 1995, 6,
1919-1932. (c) Duhamel, P. Bull. Soc. Chim. Fr. 1996, 133, 457-459.
9
10.1021/ja039046t CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 4245-4255
4245

A R T I C L E S
France et al.
Scheme 1. Tandem Catalytic Asymmetric Chlorination/
Esterification
These would in turn react with a mild electrophilic halogen
source in a tandem halogenation/esterification reaction (Scheme
1). The electrophilic halogen would add to the R-position of
the enolate, producing an acylammonium salt that subsequently
undergoes transacylation with the leaving group of the electro-
phile. This process produces a series of chiral R-chloroesters
with two new points of potential functionalization (the R-halogen
atom and the active ester), that could serve as intermediates for
the conversion to optically active amines, amides, ethers,
alcohols, esters, and sulfides. Through the use of mild haloge-
nating reagents, we could introduce enantioselectivity into our
products. The challenge was multifold: to establish a new chlori-
nation reaction, to observe catalysis, and to impart enantio-
selectivity.
Results and Discussion
has developed an important diastereoselective synthesis of a
series of R-bromo- and R-iodoesters from in situ generated
R-halogenated ketenes (eq 3).8
Catalytic Asymmetric Chlorination Using Proton Sponge
as the Stoichiometric Base. The initial strategy was influenced
by our earlier, successful development of a catalytic, asymmetric
synthesis of â-lactams through the cycloaddition of ketenes and
imines14 employing various chiral nucleophiles, including
cinchona alkaloid derivatives15 such as benzoylquinine (BQ)
3a, as dual catalysts and “shuttle” bases, and proton sponge as
the initial dehydrohalogenating agent.16 In the subsequent
improvement of the â-lactam reaction,17 we found that small
amounts of metal cocatalysts, including salts of In(III) and
Zn(II), could dramatically improve the yields of these reactions,
and also other, less expensive bases, such as NaH, could be
employed. One key aspect of the â-lactam reactions concerns
the putative formation of a chiral, zwitterionic enolate from in
situ generated ketenes (or sometimes more directly from the
acid chlorides themselves) and chiral nucleophiles. In the case
of the â-lactam chemistry, the enolate reacts with imines, but
at that time there was reason to believe that these species could
also condense with other electrophiles of appropriate reactivity,
such as chlorinating agents. In the beginning, the most important
question concerned the choice of chlorinating agent. Diatomic
chlorine was deemed too reactive, whereas most other agents
proved to be completely unreactive. Surprisingly, sources of
halogen that are too mild pose the biggest problem, as they are
unable to react with the weakly nucleophilic zwitterionic enolate.
It was clear to us that only a very narrow window of reactivity
existed in which a desirable chlorinating agent could operate
and that reaction development would be more difficult than
originally anticipated.
Recent attention has focused on the development of catalytic
asymmetric methods for generating halogenated compounds.
Ultimately, the design of a catalytic, enantioselective R-halo-
genation reaction would effectively utilize a very mild source
of halogen that could be easily manipulated. The very first step
in this nascent field was reported by Togni, who has developed
a titanium-taddol-catalyzed halogenation procedure (eq 4).9 This
protocol is effective for the halogenation of highly enolizable
R-alkyl-1,3-dicarbonyl compounds, which can readily form
titanium-based enolates in situ.10 A similar protocol has been
recently developed by Sodeoka et al., who used a Pd(II)-
BINAP complex to achieve the asymmetric fluorination of
R-keto esters.11
As Togni’s initial report appeared, we had begun work on
our own catalytic, asymmetric halogenation (particularly chlo-
rination) reactions. We envisioned a process in which in situ
generated ketenes12 would react with chiral nucleophiles such
as cinchona alkaloid derivatives to form zwitterionic enolates.13
(7) Belucci, G.; Chiappe, C.; D’Andrea, F. Tetrahedron: Asymmetry 1995, 6,
221-230.
(8) Durst, T.; Koh, K. Tetrahedron Lett. 1992, 33, 6799-6802.
(9) (a) Hintermann, L.; Togni, A. Angew. Chem., Int. Ed. 2000, 39, 4359-
4362. (b) Hintermann, L.; Togni, A. HelV. Chim. Acta 2000, 83, 2425-
2453.
(10) It is useful to distinguish between asymmetric processes in which halogen
adds as either an electrophile or a nucleophile. The latter category includes
the enantioselective opening of meso epoxides: Jacobsen, E. N. Acc. Chem.
Res. 2000, 33, 421-431.
(11) Hamashima, Y.; Yagi, K.; Takano, H.; Tama´s, L.; Sodeoka, M. J. Am.
Chem. Soc. 2002, 124, 14530-14531.
(12) (a) Tidwell, T. T. Ketenes; John Wiley & Sons: New York, 1995. (b)
Palomo, C.; Aizpurua, J. M.; Ganboa, I.; Oiarbide, M. Eur. J. Org. Chem.
1999, 3223-3235. (c) Lynch, J. E.; Riseman, S. M.; Laswell, W. L.;
Tschaen, D. M.; Volante, R. P.; Smith, G. B.; Shinkai, I. J. Org. Chem.
1989, 54, 3792-3796.
(13) Chiral ammonium enolates have been shown to be the reactive intermediates
in a number of asymmetric transformations. For examples, see: (a)
Wynberg, H.; Staring, E. G. J. Am. Chem. Soc. 1982, 104, 166-168. (b)
Calter, M. A. J. Org. Chem. 1996, 61, 8006-8007.
Nevertheless, we were encouraged by the results of the
â-lactam reaction and began screening sources of electrophilic
9
4246 J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004

Catalytic, Asymmetric R-Chlorination of Acid Halides
Scheme 2. Chlorination Using t-BuOCl
A R T I C L E S
by an undesired side reaction, the chlorination of electron-rich
proton sponge by the quinone 5a. These results constituted our
first (and very ominous) encounter with byproduct halogenations
that ultimately result in undesired ketene phenolysis and
prompted us to investigate intensively other “clean” methods
of ketene generation that would result in minimal exposure of
the halogenating agents to other substrates (Vide infra).
halogen, such as alkylhypochlorites,18 N-chlorosuccinimide
(NCS), and various N-chloroamides with in situ generated
phenylketene using BQ 3a as a catalyst. Phenylacetyl chloride
1a was added to a solution of 10 mol % catalyst and 1.1 equiv
of proton sponge 4a at -78 °C to generate the zwitterionic
enolate (eq 5). In most cases, N-halosuccinimides and N-
chloroamides were unsuccessful, yielding no detectible products,
as were a multitude of other candidates, such as chlorinated
pyridones, iodanes, and sulfonamides. tert-Butylhypochlorite,
on the other hand, proved to be too reactive, chlorinating almost
anything in the reaction mixture, including solvent (Scheme 2).
After much effort, we finally turned our attention to the
polyhaloquinone-derived reagents, including 5a-5d.19 Polyha-
logenated quinones have a long history as often unanticipated
byproducts of aromatic halogenation reactions. Hundreds of
them are known in the literature, and they are often easy to
make (or in certain cases, purchase). For example, pentachlo-
rophenol reacts readily with tert-butylhypochlorite to produce
quinone 5a in quantitative yield;20 this substance is available
commercially. We have also undertaken the development of an
enantioselective R-bromination reaction, facets of which are
different enough to be reported separately.21
Halogen transfers involving polyhaloquinones are expected
to release stabilized aromatic phenolate anions in a thermody-
namically favorable process; we envisaged that the phenolate
could then react with the resulting acylammonium salt (Scheme
1) and regenerate the catalyst to yield the final product. In our
first attempt using the perchlorinated quinone 5a and pheny-
lacetyl chloride, with 10 mol % BQ as catalyst and 1.1 equiv
of proton sponge, R-chloroester 6a was formed in moderate yield
(40%) but with high enantioselectivity (95% ee). Also isolated
from the reaction mixture, however, was a fair amount of the
achiral ester 7a, the product resulting from the formal alcoholysis
of phenylketene by pentachlorophenol (eq 6). Further investiga-
tion revealed that pentachlorophenol was being generated in situ
We started by using various chlorinated proton sponge
derivatives 4b-d22 as stoichiometric bases, reasoning that they
would be resistant to further halogenation. However, they proved
to be deactivated as bases, affording products in low yield,
although the amount of undesired byproduct also dropped
considerably. At this point, we deemed it necessary to investigate
entirely different classes of more effective stoichiometric bases
for the reaction.
Ketene Formation I: Generation Using BEMP. Our first
attempt at “clean” ketene generation using an alternative base
involved the use of the powerful, resin-bound, phosphazene base
BEMP, which could be filtered off after reaction to leave behind
a solution of pure ketene. We found that when a solution of
phenylacetyl chloride in THF is passed through an addition
funnel containing at least 1 equiv of BEMP at -78 °C, phenyl-
ketene is produced quantitatively.23 The ketene solution was
allowed to drip slowly into a flask (-78 °C) containing 3a (10
mol %), and to this was added a solution of 5a. After this solu-
tion had stirred at -78 °C for 4 h, quenching and column
chromatography yielded the product (S)-6a24 in 80% yield and
99% ee (eq 7).25 A number of other acid chlorides were screened
using this procedure with similar results as summarized in Table
1. As can be seen, a wide range of acid chlorides was success-
fully employed, including those that possess either aliphatic or
aromatic substituents, to afford products in high enantioselec-
tivity and moderate to good chemical yields. Whereas BQ (3a)
forms one series of enantiomers in reliably stereoregular fashion,
benzoylquinidine, or BQd (3b, the “pseudoenantiomer of BQ”),
consistently affords the opposite set (for example, products (R)-
6a and (R)-6b). Notable is the compatibility of acid chloride
1b, possessing an electron-rich (and “halogenatable”) phenoxy
(14) (a) Taggi, A. E.; Hafez, A. M.; Wack, H.; Young, B.; Ferraris, D.; Lectka,
T. J. Am. Chem. Soc. 2002, 124, 6626-6635. (b) Taggi, A. E.; Hafez, A.
M.; Wack, H.; Young, B.; Drury, W. J., III; Lectka, T. J. Am. Chem. Soc.
2000, 122, 7831-7832.
(15) For other timely uses of cinchona alkaloids in catalytic asymmetric synthesis
and others contained therein: France, S.; Guerin, D. J.; Miller, S. J.; Lectka,
T. Chem. ReV. 2003, 103, 2985-3012.
(16) Cinchona alkaloid derivatives have recently been used as stoichiometric
reagents for asymmetric halogenation: (a) Cahard, D.; Audouard, C.;
Plaquevent, J.-C.; Roques, N. Org. Lett. 2000, 2, 3699-3701. (b) Shibata,
N.; Suzuki, E.; Takeuchi, Y. J. Am. Chem. Soc. 2000, 122, 10728-10729.
(17) France, S.; Wack, H.; Taggi, A. E.; Hafez, A. M.; Witsil, D. R.; Lectka, T.
Org. Lett. 2002, 4, 1603-1605.
(22) (a) Ozeryanskii, V. A.; Pozharskii, A. F.; Vistorobskii, N. V. Russ. J. Org.
Chem. 1997, 33, 251-256. (b) Glowiak, T.; Majerz, I.; Malarski, Z.;
Sobczyk, L.; Pozharskii, A. F.; Ozeryanskii, V. A.; Grech, E. J. Phys. Org.
Chem. 1999, 12, 895-900.
(23) (a) Hafez, A. M.; Taggi, A. E.; Dudding, T.; Lectka, T. J. Am. Chem. Soc.
2001, 123, 10853-10859. (b) Hafez, A. M.; Taggi, A. E.; Wack, H.; Drury,
W. J.; Lectka, T. Org. Lett. 2000, 2, 3963-3965.
(24) Proofs of absolute configuration were determined on the basis of conversion
to the known methyl ester and a known R-thio derivative. Stereoregularity
was inferred for other products on the basis of these proofs as well as on
the basis of computational models. See Supporting Information.
(25) Wack, H.; Taggi, A. E.; Hafez, A. M.; Drury, W. J.; Lectka, T. J. Am.
Chem. Soc. 2001, 123, 1531-1532.
(18) Walling, C.; Padwa, A. J. Org. Chem. 1963, 28, 2976-2977.
(19) Denivelle, L. Bull. Soc. Chim. Fr. 1957, 724-728.
(20) Guy, A.; Lemaire, M.; Guette, J.-P. Synthesis 1982, 12, 1018-1020.
(21) Hafez, A. M.; Taggi, A. E.; Wack, H.; Esterbrook, J.; Lectka, T. Org. Lett.
2001, 3, 2049-2051.
9
J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004 4247

A R T I C L E S
France et al.
Scheme 3. Sodium Hydride Shuttle Deprotonation System
Table 1. Alkaloid-Catalyzed Reactions of Acyl Halides 1 Using
BEMP as a Dehydrohalogenating Agent
Table 2. Alkaloid-Catalyzed Reactions of Acyl Halides 1 Using
NaH
a Reactions run with 10 mol % catalyst, 0.13 mmol of ketene, 0.065
mmol of 5a at -78 to 25 °C for 4 h in THF. Yield based on 5a after
chromatography.
group, with the chlorinating agent. Also noteworthy is the use
of R-bromoacetylbromide 1f to afford the R-bromo-R-chloro
ester 6f as the product in 51% yield and 97% ee, a molecule in
which two different halogen atoms are attached to one chiral
center. Finally, with BEMP as base, the thiophen-2-yl-acetyl
chloride 1g can be converted smoothly into its corresponding
R-chloroester (S)-6g in 66% yield and 80% ee, an uncharac-
teristically low degree of optical induction for this series of
reactions.
a Reactions run with 10 mol % catalyst, 2.7 mmol of ketene, 1.3 mmol
of 5a at -78 to 25 °C for 4 h in THF. Yield based on 5a after
chromatography.
In this case, ketene is generated through a shuttle deprotonation
system that employs BQ and 15-crown-5 as a phase transfer
cocatalyst to help solubilize NaH (Scheme 3). Phenylacetyl
chloride was added to a stirred suspension of NaH and catalytic
amounts of 15-crown-5 and BQ in THF at -78 °C. A solution
of 5a in THF was added slowly over 3 h by syringe pump, and
the reaction was warmed to room temperature over 4 h (eq 8).26
Workup and chromatography yielded the R-chlorinated product
(S)-6a in 63% yield and 95% ee. We proceeded to examine the
system on other acid chloride substrates (Table 2) with results
that are comparable to BEMP (ee’s 90-99%; yields 58-79%),
but at a fraction of the cost. Most importantly, we were able to
Ketene Formation II: “Shuttle” Deprotonation Using the
Sodium Hydride/Crown Ether System. In an attempt to
develop a more cost-effective way to generate ketenes (5 g of
BEMP costs about 800 times more than NaH), we began to
test hydride salts that could act as stoichiometric bases in the
dehydrohalogenation of acid chlorides. We also sought a pro-
cedure that is amenable to larger scale halogenations (in large
amounts, BEMP, besides being costly, is difficult to handle and
“gums” up reaction flasks preventing smooth stirring); we
examined the use of the inexpensive and low molecular weight
sodium hydride as a stoichiometric base for ketene generation.
(26) Taggi, A. E.; Wack, H.; Hafez, A. M.; France, S.; Lectka, T. Org. Lett.
2002, 4, 627-629.
9
4248 J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004

Catalytic, Asymmetric R-Chlorination of Acid Halides
A R T I C L E S
Scheme 4. Generation of Zwitterionic Enolates Using Sodium
Bicarbonate
Table 3. Alkaloid-Catalyzed Reactions of Acyl Halides 1 Using
NaHCO3
a Reactions run with 10 mol % catalyst, 2.7 mmol of acid chloride, 1.3
mmol of 5a at -35 to 25 °C for 5 h in chlorobenzene. Yield based on 5a
after chromatography.
directly. The hydrochloride salt of BQ then transfers its proton
to the bicarbonate, thus regenerating the catalyst.
To support our hypothesis, we conducted the experiment at
room temperature (where monosubstituted ketenes are usually
unstable) and isolated product in 65% yield, with albeit lowered
enantioselectivity (56% ee). When we attempted to perform the
corresponding reaction with NaH as stoichiometric base at room
temperature, no desired products were observed. During opti-
mization, we discovered that an excess of sodium bicarbonate
(>10 equiv), in the presence of catalytic amounts of BQ and
15-crown-5 cocatalyst at -35 °C in chlorobenzene, afforded
R-haloester (S)-6a in 68% yield with 90% ee (eq 9). Although
we have not screened a full range of acid chloride candidates
using this protocol, we have reason to believe it should be as
applicable as the earlier methodology, albeit with slightly lower
ee’s at the present time (Table 3).
prepare compound 6a on a larger, multigram scale using this
methodology.
Ketene/Enolate Formation III: Chlorinations Using So-
dium Bicarbonate. Although cost-effective, NaH presents some
disadvantages as a stoichiometric base. For example, NaH
requires a ketene “preformation” step that can be difficult to
manage. High concentrations of ketene can be prone to dimer-
ization and other side reactions; so a method that would involve
a very slow, “time release” ketene or zwitterionic enolate gener-
ation might be more appealing. Along those lines, we have found
that we can effect halogenation of acid chlorides using an excess
of the mild base sodium bicarbonate as a dehydrohalogenating
reagent along with catalytic amounts of BQ and 15-crown-5.27
Sodium bicarbonate is extremely cheap, readily available, very
mild, and air stable. It is potentially more advantageous than
the use of potassium carbonate because racemization due to
excess base or byproduct sodium salts is much less likely,
allowing for a one-pot procedure and eliminating the need for
a filtration step. Furthermore, we presumed that proton transfer
from protonated BQ (or an acylammonium salt) to the bicarbon-
ate should be slower, due to the mild basicity of bicarbonate,
thus ensuring a proportionately slower formation of ketene (or,
perhaps more plausibly, the zwitterionic enolate). In contrast
to our proposed ketene generation with sodium hydride, it is
less likely that free ketene is being generated with sodium
bicarbonate, as the preformation step has been obviated. In fact,
when we attempt (even lengthy) preformation steps with
bicarbonate as stoichiometric base and BQ as the shuttle, no
free ketene (monitored by ReactIR) is formed (Vide infra). A
plausible mechanism is illustrated in Scheme 4. After initial
transacylation by one molecule of BQ, a second molecule of
catalyst abstracts the R-proton to form the zwitterionic enolate
Scale-Up Reactions. In an attempt to probe the utility of
our methodology, we focused our attention on the scale-up of
the halogenation/esterification using either sodium hydride or
sodium bicarbonate. Using 2 g of 5a, we obtained R-chloroester
6a in 65% yield and in 91% ee for NaH and 60% yield with
90% ee for NaHCO3.
Direct Ketene Formation IV: Generation via Wolff
Rearrangement. Another common method for generating free
ketene is the Wolff rearrangement.28 When diazoketone 8 is
photolyzed, it produces phenylketene through a standard Wolff
rearrangement (eq 10). We photolyzed 8 in the presence of
quinone 5a and 10 mol % BQ in THF and afforded R-chlo-
roester 6a in 45% yield and 85% ee, demonstrating that
alternative sources of ketenes are viable in the reaction.
(27) Shah, M. H.; France, S.; Lectka, T. Synlett 2003, 12, 1937-1939.
9
J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004 4249

A R T I C L E S
France et al.
Scheme 5. Proposed Transition States of Chlorination Using 5a
versus 5b
Figure 1. PM3 transition state assembly of the reaction of 5a with the
BQ‚phenylketene derived enolate.
results in convergence to a linear geometry or else dissociation
to the methyl anion and chloromethane upon optimization.
Species 9 can serve as a model for the electrophilic chlorine
transfer process of our system, and it would seem to disfavor
the concerted option on this basis.
Mechanistic Considerations. It occurred to us that it is
possible that the asymmetric chlorinations are operating through
either of two reactive assemblies. One can imagine a process
by which R-chlorination and transacylation occur simultaneously
through a six-membered transition state, or else a mechanism
in which these two steps are separated (Scheme 5). The ortho
relationship between the carbonyl oxygen and the electrophilic
chlorine atom in 5a makes the concerted option a possibility.
To test this theory, we converted 5a to its corresponding para-
substituted isomer 5b under AlCl3 catalysis (eq 11).29 A six-
membered transition state is not possible for reaction of 5b with
an enolate for steric reasons. When 5b was substituted for 5a
in the chlorination of phenylacetyl chloride using NaH as the
base, R-chloroester 6a was formed in 68% yield and only 56%
ee. This result can be rationalized through two possibilities: one
based on transition state ordering, and the other on thermody-
namics. With 5a, the six-membered transition state allows for
facile transfer of the chlorine with simultaneous ester formation
in a rigid transition state assembly. Because this assembly is
not geometrically possible for 5b, only a less ordered stepwise
process is involved, and the selectivity might be diminished.
Alternatively, 5b is lower in energy than 5a, and perhaps less
reactive as a halogenating agent. Consequently, the reaction rate
is slowed and the selectivity similarly is decreased as higher
temperatures are needed to ensure reaction. It is difficult to
imagine a practical experiment to differentiate the two pos-
sibilities, so we turned to theoretical calculations to shed light
on the question.
Stereochemical Models and Transition States. To deter-
mine good starting geometries for molecular orbital-based
transition state calculations, we first turned to molecular
mechanics (MM) given the size and the “all organic” nature of
the putative reactive enolate complex. The MM model that we
propose to account for the sense of induction that we observe
in our cinchona alkaloid-catalyzed halogenation reactions is also
similar to the model proposed for the asymmetric â-lactam
forming process. For example, MM calculations using modified
MMFF or AMBER force fields31 on model ketene-3a (the
complex derived from reaction of BQ with phenylketene) reveal
a low energy structure in which the re face is exposed to the
approach of the halogenating agent, whereas the si-face approach
is 2.5-7 kcal/mol higher in energy, depending on the force field
used in the calculation.32
Next, we calculated transition states at the semiempirical PM3
level of the complex derived from phenylketene and BQ during
its reaction with 5a (Figure 1) using the low energy MM
conformation as a starting geometry. While admittedly crude,
this level of theory was deemed necessary given the size of the
reactive assembly. No stable concerted TS’s were found in this
case; in contrast, the chlorine transfer process occurs nearly
linearly (bond angle 162°) in a “late” TS (distance enolate/Cl
) 1.84 Å; distance quinone/Cl ) 2.40 Å) in a fashion that favors
a stepwise mechanism. When the geometry is constrained to a
cyclic array, it is evident that substantial steric hindrance
between the enolate and 5a would be expected. In the linear,
stepwise transition state, however, an interaction between the
carbonyl/phenolic oxygen on 5a and the hydrogen atoms alpha
to the bridgehead nitrogen on the BQ catalyst constitutes an
apparent stabilizing (rigidifying) Lewis acid/Lewis base (LA-
LB) interaction.33 Thus, at this point, we favor the thermody-
namic explanation for the difference in selectivity between 5a
and 5b (with perhaps a minor LA-LB interaction to stiffen
For example, the transition state of the chlorine transfer step
in the concerted mechanism must be highly bent. However, the
bridged µ-chloro anion 9 shows a strong preference for a linear
geometry at B3LYP/6-31G*;30 a bent starting geometry (10)
(30) Calculations were performed using the Titan Program, version 2.0. On the
other hand, the µ-bridged chloro cation [H3C-Cl-CH3]+ shows a strong
preference for a bent geometry.
(31) We used the program Macromodel (version 7.0) for the calculations
(Schrodinger, Inc.).
(32) Taggi, A. E.; Hafez, A. M.; Dudding, T.; Lectka, T. Tetrahedron 2002,
58, 8351-8356.
(28) Chiang, Y.; Kresge, A. J.; Popik, V. V. J. Am. Chem. Soc. 1999, 121,
5930-5932.
(29) Denivelle, L. Bull. Soc. Chim. Fr. 1956, 1834-1836.
9
4250 J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004

Catalytic, Asymmetric R-Chlorination of Acid Halides
A R T I C L E S
To confirm the pivotal role of BQ with regard to the ketene
forming reaction, several control reactions were also conducted
and monitored by ReactIR. At -35 °C in chlorobenzene, sodium
hydride alone does not produce any visible amount of ketene
(eq 12). Even the addition of 2 equiv of NaH has no visible
effect on the phenylacetyl chloride carbonyl stretch. Similarly,
the addition of 10 mol % 15-crown-5 to a mixture of 1.0 equiv
of NaH and phenylacetyl chloride in chlorobenzene at -35 °C
produces no detectible free ketene. It is not until both of these
reactions warm to near 0 °C that the free ketene signal becomes
apparent, with the 15-crown-5 reaction producing ketene earlier
and to a greater extent, but not nearly to the extent to which
the BQ catalyst induces ketene formation. When BQ is added
to the NaH solution at lowered temperatures, the phenylketene
stretch at 2328 cm-1 then grows in over time (eq 13 and Figure
2).
Kinetics of the NaH and the HCO3- Promoted Reactions.
If the use of NaH as stoichiometric base involves a ketene
preformation step, whereas the use of bicarbonate does not, it
stands to reason that the dependence of the reaction rate of each
shuttle base system on various reaction components could be
different. To our surprise, we determined that the reaction rates
of both the NaH and the bicarbonate promoted halogenation
reactions are proportional to the concentration of halogenating
agent for the standard reaction of phenylacetyl chloride 1a with
5a in THF at -78 °C.35 This result is in marked contrast to the
â-lactam cycloaddition reaction, whose rate is dependent on the
concentration of acid chloride and catalyst, but not imine. These
facts run counter to our intuition that in the case of the
bicarbonate shuttle, slow enolate formation should be the rate-
determining step (RDS), not reaction of enolate with chlorinating
agent, as our results imply. Still, our intuition can be reconciled
with the kinetics if we assume that the rate of enolate formation
in the bicarbonate shuttle is comparable to the rate of reaction
with zwitterionic enolate, or else that enolate formation is
reversible, so that its concentration does not build up over the
course of time (eq 14). Given the basicity of bicarbonate, this
scenario is very plausible, and, as stated, reversible enolate
formation is observed in the â-lactam cycloaddition reaction
with proton sponge as base. Interestingly, we observed no direct
dependence of the catalyst concentration on the rates of product
formation above 10 mol % loading. This is due to the apparent
fact that with higher catalyst concentrations, there is a dramatic
observed increase in competitive phenol formation in the early
stages of reaction, an event which necessarily leads to formation
of nonhalogenated byproduct 7.
Figure 2. Growth of the phenylketene stretch at 2328 cm-1 over time.
the transition state) and a stepwise mechanism in which a tight
ion pair recombines in a fast acylation step.
Spectroscopic Evidence of Ketene Formation. We con-
ducted a series of spectroscopic experiments to provide evidence
for the putative “preformation” of ketene when sodium hydride
is used as the stoichiometric base. Our previous investigations
led us to similar conclusions about the irreversible BEMP
method of ketene generation.34 Our initial shuttle deprotonation
system utilized proton sponge as the stoichiometric base and
did not require the generation of free ketene to proceed; the
zwitterionic enolate was also found to form reversibly under
these conditions. Similarly, the use of bicarbonate bases does
not necessarily mandate the preformation of ketenes.
To shed light on this issue, we examined the formation of
phenylketene from sodium hydride and BQ catalyst by a
ReactIR experiment. With the addition of 10 mol % BQ to the
acid chloride, sodium hydride, and 10 mol % 15-crown-5, free
ketene is gradually formed, as witnessed by the observed IR
frequency for the ketene carbonyl stretch at 2328 cm-1 (Figure
2). The reaction of BQ with phenylacetyl chloride itself is very
fast, with a small, but noticeable, ketene hump (2328 cm-1
)
forming immediately after addition. The broad phenylketene
stretch then grows in fairly rapidly before it levels off and begins
to diminish as the BQ present slowly catalyzes the formation
of ketene dimer. Interestingly enough, the phenylacetyl chloride
carbonyl stretch (1799 cm-1) never completely goes away, thus
explaining why an excess of acid chloride may be needed. As
-
mentioned, ketene formation is not observed with HCO3 as
base, even in the presence of BQ.
The postulate of reversible enolate formation is consistent
with the observation that R,R-dideuteriophenylacetyl chloride
(1a-d2) undergoes fast R-proton exchange in the presence of
BQ catalyst and bicarbonate in THF solvent, resulting in
formation of R-chloro-R-protio product 6a (eq 15).
(33) Short C-HsO distances have been identified in organic compounds and
also in nucleic acid molecules. These interactions have been studied in
detail, with an emphasis on the application of crystal correlation studies.
Due to their influence in crystal packing, studies are underway to ascertain
the extent to which these interactions affect C-H approach to an oxygen
atom. For mechanistic and theorical discussions of C-HsO interactions,
see: (a) Calhorda, M. J. Chem. Commun. 2000, 801-809. (b) Desiraju,
G. R. Acc. Chem. Res. 1991, 24, 290-296. (c) Gu, Y.; Kar, T.; Scheiner,
S. J. Am. Chem. Soc. 1999, 121, 9411-9422. (d) Steiner, T. Chem.
Commun. 1997, 727-734.
(35) Kinetics experiments were performed by varying the concentration of
chlorinating agent 5a, BQ catalyst, and acid chloride substrate, and
measuring the percent conversion at various times (see Supporting
Information).
(34) Taggi, A. E.; Hafez, A. M.; Wack, H.; Young, B.; Ferraris, D.; Lectka, T.
J. Am. Chem. Soc. 2002, 124, 6626-6635.
9
J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004 4251

A R T I C L E S
France et al.
Mechanistic and Practical Implications of Nonhalogenated
Byproducts. Throughout the development of these reactions,
we were intermittently plagued by the capricious formation of
undesired nonhalogenated achiral ester byproducts 7 in quantities
varying from trace amounts to >50%. The byproducts can be
thought of in a formal sense as resulting from the reaction of
pentachlorophenol with in situ generated ketenes. Pentachlo-
rophenol, for example, must arise from the reduction of quinone
5a by species within the reaction mixture that contain active
hydrogen. As explained above, high catalyst loadings paradoxi-
cally promote the formation of byproduct through quinone
reduction. Other possibilities for reductants include: (1) the
solvent, (2) the stoichiometric base, (3) the catalysts and
cocatalysts, (4) the substrate (which can conceivably react with
the quinone through a remote position, as well as the desired
R-position), (5) impurities such as water, and (6) light (which
may induce radical pathways). To minimize (or eliminate)
byproduct formation, each of these possibilities was investigated
individually. Because both the acid chloride and the chlorinating
agent are incorporated into the byproduct, its suppression could
potentially increase our yields. The mechanism of the formation
of this product is also of interest; one could imagine that
pentachlorophenol is being formed in situ and this reacts with
either a free ketene or the acid chloride to give the ester.
Our initial investigation centered on the potential formation
of free pentachlorophenol involving each of the aforementioned
scenarios:
(1) Solvent. This seemed one of the more likely culprits due
to its overwhelming presence in the reaction. Additionally, THF,
a solvent of choice, is known to react readily with certain active
halogenating agents.36 We stirred quinone 5a in THF both in
the presence and in the absence of BQ at temperatures ranging
from -78 °C to reflux (eq 16). Only at room temperature and
above was formation of phenol observed in quantities sufficient
to account for production of the byproduct, and BQ had only a
small effect on promoting phenol formation under these
conditions. The THF is presumably converted into the R-chlo-
ride, as is precedented in published work on THF chlorinations.37
(2) The Stoichiometric Base. We have observed byproducts
with the use of several different stoichiometric bases, including
proton sponge, NaHCO3, and NaH. As we stated earlier, proton
sponge reacts readily with quinone 5a to liberate pentachlo-
rophenol. Potassium carbonate affords much reduced quantities
of byproduct relative to proton sponge, but for practical reasons
carbonate bases are inferior to NaH, as a filtration step is needed.
When prolonged times are employed in the reaction of NaH
and acid halides to produce putative ketenes, ensuring consump-
tion of all NaH, the relative quantities of byproduct are reduced,
suggesting that residual NaH can react with quinone 5a.
However, when NaH is mixed with 5a in CCl4 in the presence
of 15-crown-5 and BQ, only trace amounts of phenol were
observed after 24 h at room temperature, suggesting that quinone
reduction by NaH is probably not a major contributor in the
production of the achiral ester 7 (eq 18).
(3) The Catalysts and Cocatalysts: N-Chloro Species 3c.
In our initial report, we demonstrated that N-chlorination of
resin-bound BQ deactivated the catalyst, although N-chlorination
seemed to be less of a problem using soluble catalysts. Upon
recharacterization of the quinine moiety, we discovered deriva-
tive 3c. We set up a test reaction to see if N-chlorination could
be plausibly occurring in solution under any circumstances, in
light of our catalyst concentration experiments. We employed
2-(N,N-dibenzylaminomethyl)naphthalene 11 to catalyze the
halogenation reaction in place of BQ, assuming that if N-
chlorination were to occur, dehydrohalogenation would follow
yielding an iminium salt 13.
Hydrolysis of 13 should yield naphthaldehyde 14, an easily
identifiable marker (Scheme 6) that provides a convenient test
for N-chlorination. In the event, naphthaldehyde was formed
in 30% yield and R-haloester 6a was formed in 30% yield,
confirming that under select conditions N-chlorination can play
a deleterious role in the reaction, although fortunately BQ seems
to be fairly resistant to N-chlorination under normal reaction
conditions, except perhaps in high concentrations.
Finally, even the role that the cocatalyst 15-crown-5 may play
in byproduct formation was investigated. Quinone 5a was treated
with 15-crown-5 in the inert solvent CCl4 and monitored from
-78 °C to room temperature. No phenol formation was observed
in this case.
(4) Substrates. Under no conditions have we observed remote
chlorination of a substrate that would account for phenol
formation.
(5) Impurities. The most likely impurity in the reaction is
water, especially with the use of hygroscopic solvents such as
THF. To elucidate the role that water could be playing, we have
conducted our reactions under conditions ranging from scru-
pulously dry to wet (1 equiv or more of water). A definite trend
could be discerned: in reactions with little or no water,
Note that the reaction temperatures for phenol formation are
higher than those for desired product formation, suggesting that
careful temperature control could alleviate problems. However,
putative phenol formation occurs at much lower temperatures
in the actual reactions. When the more easily oxidized solvent
2,5-dimethyl-THF was employed in the standard reaction, an
enhanced yield of phenol was observed, implicating the solvent
as a potential substrate (eq 17), and suggesting the use of an
appropriate proton free solvent in some cases.
(36) (a) Koole, G. A.; Nelson, W. L.; Gioannini, T. L.; Angel, L.; Simon, E. J.
J. Med. Chem. 1984, 27, 1718-1723. (b) Crombie, L.; Wyvill, R. D. J.
Chem. Soc., Perkin Trans. 1 1985, 1971-1982.
(37) Gross, H. Angew. Chem. 1960, 72, 268-269.
9
4252 J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004

Catalytic, Asymmetric R-Chlorination of Acid Halides
A R T I C L E S
Scheme 6. Deactivation of Catalyst Due to N-Chlorination
Table 4. Solvent Effects on the Asymmetric Chlorination of
Phenylacetyl Chloride Using NaH under Standard Conditions
solvent
% ee
% yielda
% byproduct 7a
THF
92
92
92
-
65
93
85
65
65
65
0
40
70
45
31
27
25
90
55
-
toluene
CH2Cl2
benzotrifluoride
CFCl3
chlorobenzeneb
c
CCl4
25
a Reactions run with 10 mol % catalyst, 0.13 mmol of ketene, 0.065
mmol of 5a at -78 °C to room temperature for 4 h in THF. Yield based on
5a after chromatography. b Reaction preformed at -35 °C. c Reaction
preformed at -0 °C.
Trying the normal reaction in the presence of a radical scavenger
(cumene) afforded product in 58% yield with 20% of the
unwanted byproduct. Therefore, we can conclude that light is
not significantly affecting the amount of desired product,
although it seems to have an effect on the amount of byproduct
in some cases.
drastically reduced amounts of byproduct were obtained,
whereas in “high” water reactions, substantial amounts of
byproduct were isolated. Under the conditions of the reaction,
it is possible that water could react with quinone 5a to form
hypochlorite and phenol, thus accounting for the observed
results, although this would seem, based on prior observations,
to be contrathermodynamic.
(7) Alternative Solvents. We screened a number of solvents
under the normal conditions, ranging from those with no protons
(CCl4, CFCl3) to those with protons present on deactivated
aromatic rings (chlorobenzene, benzotrifluoride), and compared
the results to those obtained using common solvents such as
THF and toluene (Table 4). Chlorobenzene proved to be the
most compatible with our method, although the reaction
temperature had to be raised to -35 °C to avoid freezing. Upon
the replacement of THF in the published NaH procedure with
chlorobenzene, chloroester 6a was formed in 70% yield with
93% ee, with no detected byproduct. Benzotrifluoride results
in a spectacular failure (>90% byproduct) for unknown reasons.
It must be pointed out that the formation of byproduct otherwise
had little correlation to the presumed reactivity (especially
oxidizability) of the solvent, aside from the case of THF and
dimethyl-THF.
In summary, it is unlikely that any one factor can completely
account for the formation of the nonhalogenated byproduct.
Instead, a complex interplay of potential reasons can conspire
to produce large amounts of byproduct, rendering the reaction
of reduced utility. To minimize its formation, we propose a
number of precautions, including the use of alternative proton
free solvents, NaH as an inexpensive stoichiometric base in
which ketene formation occurs under more extended (>2 h)
time periods, or else NaHCO3 as another, “slow release base”
alternative, and scrupulously water-free conditions. Also, catalyst
loadings should be kept to 5-50 mol %.
A control experiment in which benzoyl chloride was substi-
tuted for phenylacetyl chloride under normal reaction conditions
also revealed that ketene formation is not necessarily a prelude
to the byproduct (eq 19). Benzoyl chloride has no R-protons,
yet was found to undergo transacylation by pentachlorophenol
giving the ester 16.38 This suggests that the acid chloride
activates phenol formation in some manner, perhaps by transient
formation of acyloxonium ion 15, which we expect to be a much
more reactive halogenating species toward not only the substrate,
but the solvent and other reaction components as well. We view
this to be a very significant control experiment, suggesting that
acid chloride itself can play a deleterious role.
(6) Light. Another plausible explanation for the reduction
of 5a could be the formation of chlorine atoms from the
homolytic cleavage of C-Cl bond due to light. To determine
the validity of this hypothesis, we conducted a series of
experiments in the presence and absence of light. First, the
normal reaction (NaH) was performed under a 60-W lamp over
a period of 5 h. R-Chloroester 6a was obtained in 58% yield
with 35% of the nonhalogenated byproduct 7a (eq 20). In the
presence of light and AIBN, a radical initiator, R-chloroester
was formed in 55% yield with 30% byproduct.
Other Acid Halides. We have also used acid fluorides and
acid bromides with varying degrees of success in the chlorination
reaction (eq 21). Under no circumstances is bromine or fluorine
When the reaction was performed purely in the dark, product
was obtained in 58% yield with only 10% of the byproduct.
(38) Igolen, J.; Morin, C. J. Org. Chem. 1980, 45, 4802-4804.
9
J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004 4253

A R T I C L E S
France et al.
Scheme 7. Proposed Mechanism of the Formation of 19
practice, amidation of R-chloropentachlorophenyl esters pro-
ceeds well with primary amines, affording R-chloroamide 21
through room-temperature reactions that retain their optical
purity (eq 23).
Similarly, when alcohols are stirred with our activated
R-chloroesters, transesterification occurs readily without race-
mization to afford the R-chloro methyl ester 22 (eq 24).
incorporated into the R-position instead of chlorine, facts that
confirm the origin of the chlorine atom to be in the polyhalo-
quinone. Whereas acid bromides work as well as acid chlorides,
acid fluorides, on the other hand, generally afford low yields
of product. We attribute this result to the sluggish acylation of
the chiral nucleophilic catalyst by the acid fluoride, which is a
less reactive acylating agent.
A Mysterious Byproduct. To gain insight into competition
between halogenation and â-lactam formation, reactions pre-
sumed to follow similar mechanisms, we conducted an experi-
ment in which the chlorinating agent 5a and R-imino ester 17
were added in equal amounts to a preformed solution of
phenylketene and a stoichiometric amount of BQ at -78 °C
(eq 22). After the mixture was warmed to room temperature,
four different products were observed and isolated from the
reaction mixture. In addition to the R-chlorinated ester 6a,
â-lactam 18, and the corresponding nonhalogenated ester 7a,
variable amounts (from 5% to 50%) of the chlorinated com-
pound 19 were observed. Characterization of 19 led us to the
following proposed mechanism, involving cycloaddition of
phenylketene (or its enolate) with chlorinating agent 5a to form
lactone 20 (Scheme 7).
However, more vigorous conditions are required for amidation
with secondary amines and anilines; racemization may result
in these cases. At this point, we sought to develop new quinone-
based halogenating agents in which the active ester product
would contain a binding site that could be activated toward
amidation by a metal cocatalyst (eq 25). The halogenating agent
itself could be made more reactive by metal binding, perhaps
enhancing chemical yields in the asymmetric process. The
trichloroquinolinone 5d was easily prepared from 8-hydrox-
yquinoline and 3 equiv of tert-butylhypochlorite. When this
reagent was added to a solution of phenylacetyl chloride (1a),
sodium hydride, and catalytic amounts of 15-crown-5, BQ, and
In(OTf)3 in THF at -78 °C, the R-chloroester 23 was observed
in only 22% yield. Whereas the product starts out optically
active, it racemizes rapidly (presumably due to metal coordina-
tion to the active ester), unfortunately rendering this protocol
unviable. Performing the experiment without metal catalyst
yields no product, at least confirming that Lewis acid catalysis
is feasible. Product 23 does undergo very fast transacylations
with a variety of secondary amine substrates under Sc(OTf)3
catalysis, but we decided not to pursue this transformation
intensively, due to the lack of enantioselectivity and yield in
the initial reaction (dashed arrow, eq 25).
Decarboxylation and rearomatization with a concomitant 1,3-
chloro-shift thus provides the product. Note that this mechanism
does not take into account the presence of imine 17, which is
mysteriously necessary to the formation of the byproduct. Thus,
the precise reason this product forms is still not totally
understood.
Conversion to Useful Optically Active Compounds; Al-
ternative Halogenating Agents. The fact that the products of
the asymmetric halogenations are active esters allows facile
derivatization to other esters and amides, at least in theory. In
The second strategy involved making a pentafluorophenyl
ester product through the polyfluorinated chloroquinone 5e.
Chlorine should transfer over fluorine, leading to desired
product. In our hands, pentafluorophenyl esters are an order of
magnitude more reactive toward amidation than the correspond-
ing pentachlorophenyl counterparts. The perfluorinated R-chlo-
roquinone (5e) and the R-chloro-2,4,6-trifluoroquinone (5f) were
easily synthesized from sodium or potassium phenolate and
9
4254 J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004

Catalytic, Asymmetric R-Chlorination of Acid Halides
A R T I C L E S
chlorine gas.39 Regrettably, treatment of either reagent with a
preformed solution of phenylketene failed to yield any product.
We began to realize that we were very lucky indeed to have
chosen quinone 5a, whose reactivity is apparently optimal, for
screening. For example, a simple substitution of a Cl atom by
a phenolate group also results in an inactive agent (witness
quinone 5c).
Conclusion
In summary, we have developed a catalytic, enantioselective
synthesis of R-chloroesters using inexpensive acid chlorides as
starting materials and cinchona alkaloid derivatives as catalysts.
We discovered that polychlorinated orthoquinones are superior
chlorinating agents for our reactions, whereas most other mild
chlorinating agents fail completely. We have also improved upon
our originally published methodology by reducing the amounts
of nonhalogenated byproducts and have developed the use of
cost-effective stoichiometric bases to achieve an overall process
possessing considerable synthetic utility, and one that by any
measure is at least competitive with existing chiral auxiliary
methodology.
Synthesis of Optically Pure 2-Chloro Alcohols: Reduction
by NaBH4. It has been demonstrated in the literature that
optically pure R-chloroacids and esters can be reduced by
various means to the corresponding chloro alcohols without loss
of optical induction.40 Intrigued by this knowledge, we converted
R-chloroester 6a to the alcohol 24. Using NaBH4/CaCl2 as our
reductant, we successfully obtained the (S)-chloro alcohol 24
with no loss of selectivity (eq 26). Furthermore, conversion to
the known chiral sulfide 25 could be accomplished by treatment
with thiophenol with complete inversion.
Experimental Section
General Procedure for the Synthesis of r-Chloroesters 6 Using
Sodium Hydride. To a suspension of NaH (2.7 mmol), BQ 3a (0.13
mmol), and 15-crown-5 (0.13 mmol) in 20 mL of THF at -78 °C was
added phenylacetyl chloride 1a (2.7 mmol) in THF (1 mL). A solution
of the chlorinating agent 5a (1.35 mmol) in 20 mL of THF was added
via syringe pump over 3 h, and the reaction was subsequently allowed
to warm to room temperature. It was then quenched with 150 mL of
water, and the aqueous layer was extracted twice with 25 mL of diethyl
ether and once with 20 mL of CH2Cl2. The organic layers were
combined and dried with MgSO4, filtered, and concentrated. The residue
was taken up in chloroform and adsorbed on silica gel before
undergoing flash column chromatography (100% hexanes), affording
(S)-6a in 63% yield and 95% ee.
Displacement of Chlorine Promoted by Silver(I) Salts.
D’Angeli and co-workers have published a report on the
stereoselective nucleophilic substitution of 2-bromoamides by
amines in the presence of Ag+ or Ag2O.41 The authors observed
inversion of configuration in the presence of AgOTf or
powdered Ag2O as a promoter under sonication (eq 27). We
applied a similar procedure using our R-chloroamides as
substrates. For example, when optically enriched amide 21 is
mixed with AgOTf and an amine such as piperidine, the
corresponding R-amino amide 26 is produced in high yield with
net inversion (although about 10% of the optical purity was
lost).
Acknowledgment. T.L. thanks the NIH (GM 064559), the
Sloan and Dreyfus Foundations, and Merck & Co. for generous
support. S.F. thanks the UNCF, Merck, and Pfizer for Graduate
Fellowships. A.E.T. thanks the ACS Division of Organic
Chemistry for a Graduate Fellowship sponsored by Organic
Reactions, Inc. H.W. thanks Johns Hopkins for Whittaker
Chambers and Sonneborn Fellowships.
Supporting Information Available: General procedures for
using proton sponge, BEMP, NaH, and NaHCO3 as the
stoichiometric base, compound characterization, kinetic, and
control experiment data (PDF). This material is available free
of charge via the Internet at http://pubs.acs.org.
(39) Kobrina, L. S.; Bogachev, A. A. J. Fluorine Chem. 1993, 62, 243-258.
(40) (a) Amouroux, R.; Gerin, B.; Chastrette, M. Tetrahedron 1985, 41, 5321-
5324. (b) Kunec, E. K.; Robins, D. J. J. Chem. Soc., Perkin Trans. 1 1987,
1089-1093.
(41) D’Angeli, F.; Marchetti, P.; Bertolasi, V. J. Org. Chem. 1995, 60, 4013-
4016.
JA039046T
9
J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004 4255