Development of an extremely efficient oxidative chlorination reaction: The value of routine data collection

Article abstract of DOI:10.1021/op900241e

This contribution describes the development of an extremely efficient process for the oxidative chlorination of a benzyl, alkyl thioether to yield an alkylsulfonyl chloride. This process was subsequently run on >100 kg scale. The product alkylsulfonyl chloride was required as an intermediate, being used by several drug projects, to prepare sulfonamides. Routine data collection and reaction profiling has led to understanding, which has allowed an alternative reaction pathway to be exploited for the development of a two-step, oxidation-chlorination process. The scope of this new two-step process was briefly examined. The results of this study have allowed us to propose an empirical method for predicting the course of these oxidative chlorination reactions. During these studies we have developed a simple laboratory rig, constructed from inexpensive, readily available equipment, which allows the controlled accurate delivery of known volumes (100s of milliliters) of chlorine gas at a given rate. In our laboratories, this has made the use of gaseous chlorine a considerably less onerous task. This work is testimony to the fruit which may be borne from attempts to gain process understanding, even of an already high-yielding reaction.

Full text of DOI:10.1021/op900241e

Organic Process Research & Development 2010, 14, 278–288
Development of an Extremely Efficient Oxidative Chlorination Reaction: The Value of
Routine Data Collection
Neil Barnwell,* Philip Cornwall, Daniel Horner, Jason Knott, and John Liddon
Process R&D, AstraZeneca R&D Charnwood, Bakewell Road, Loughborough, Leicestershire, England LE11 5RH
Abstract:
Various methods for the preparation of sulfonyl chlorides
are known.3 In the absence of sensitive functionality, the
oxidative chlorination of a sulfur(II) species using gaseous
chlorine3t,u appears to be the most widely adopted and general
method. These oxidative chlorinations, or chloroxidations, are
typically run in aqueous acidic media or, when solubility or
hydrolytic stability is problematic, in two-phase systems
consisting of an inert solvent (usually DCM, CHCl3, or CCl4)
and water or an aqueous acid. There are examples of acetic
acid/water mixtures,4 and more recently (while our work was
in progress), an efficient large-scale formic acid/water system
was described.3u Navigation through the wealth of literature
relating to the oxidation of sulfur species with chlorine is an
arduous task with general trends difficult to come by. However,
it is clear that injudicious choice of either solvent or sacrificial
sulfur substituent can have a profound effect on reaction
selectivity. Sulfone formation and chlorodesulfurisation can
compete with sulfonyl chloride formation; water also plays an
important, but not well-defined, role in the course of these
reactions. Depending on the reactivity of the sulfonyl chloride,
its thermal and hydrolytic stability can also hamper efforts to
isolate clean product. Although reaction selectivity and solvent
system must largely still be determined from empirical observa-
tion, our brief study of different sacrificial sulfur substituents
has led to the proposal that there are four distinct pathways by
This contribution describes the development of an extremely
efficient process for the oxidative chlorination of a benzyl, alkyl
thioether to yield an alkylsulfonyl chloride. This process was
subsequently run on >100 kg scale. The product alkylsulfonyl
chloride was required as an intermediate, being used by several
drug projects, to prepare sulfonamides. Routine data collection
and reaction profiling has led to understanding, which has allowed
an alternative reaction pathway to be exploited for the develop-
ment of a two-step, oxidation-chlorination process. The scope of
this new two-step process was briefly examined. The results of this
study have allowed us to propose an empirical method for
predicting the course of these oxidative chlorination reactions.
During these studies we have developed a simple laboratory rig,
constructed from inexpensive, readily available equipment, which
allows the controlled accurate delivery of known volumes (100s
of milliliters) of chlorine gas at a given rate. In our laboratories,
this has made the use of gaseous chlorine a considerably less
onerous task. This work is testimony to the fruit which may be
borne from attempts to gain process understanding, even of an
already high-yielding reaction.
1. Introduction
(3) For reviews see (a) Taylor; P. C. ComprehensiVe Organic Functional
Group Transformations; Pergamon: Elsevier Science Ltd., 1995; Vol.
2, pp 674, 717. (b) Hudlicky, M. Oxidations In Organic Chemistry;
ACS Monograph 186; American Chemical Society: Washington DC,
1990. For leading references see: (c) Nishiguchi, A.; Maeda, K.; Miki,
S. Synthesis 2006, 4131. (d) Hanagan, M. A. EP 0237292, 1987. (e)
Prakash, G. K. S.; Mathew, T.; Panja, C.; Olah, G. A. J. Org. Chem.
2007, 72, 5847. (f) Sohmiya, H.; Kimura, T.; Fujita, M.; Ando, T.
Tetrahedron 1998, 54, 13737. (g) Park, Y. J.; Shin, H. H.; Kim, Y. H.
Chem. Lett. 1992, 1483. (h) Ruano, J. L. G.; Parra, A.; Yuste, F.;
Mastranzo, V. M. Synthesis 2008, 2, 311. (I) Brownbridge, P.; Jowett,
I. C. Synthesis 1988, 3, 252. (j) Monnee, C. F. M.; Marijne, M. F.;
Brouwer, A. J.; Rob, M. J.; Liskamp, R. M. J. Tetrahedron Lett. 2000,
41, 7991. (k) Seto, N.; Kamio, T. JP 11060977, 1999. (l) Chantaras-
riwong, O.; Jang, D. O.; Chavasiri, W. Tetrahedron Lett. 2006, 47,
7489. (m) Barco, A.; Benetti, S.; Pollini, G. P.; Taddia, R. Synthesis
1974, 877. (n) Kataoka, T.; Iwama, T.; Setta, T.; Takagi, A. Synthesis
1998, 4, 423. (o) Trost, B. M., Ed; ComprehensiVe Organic Chemistry;
Pergamon: Oxford, 1991; Vol 3, pp 179, 339. (p) Wilden, J. D.;
Geldeard, L.; Lee, C. C.; Judd, D. B.; Caddick, S. Chem Commun.
2007, 1074. (q) Pandya, R.; Murashima, T.; Tedeschi, L.; Barrett,
A. G. M. J. Org. Chem. 2003, 68, 8274. (r) Gareau, Y.; Pellicelli, J.;
Laliberte, S.; Gauvreau, D. Tetrahedron Lett. 2003, 44 (42), 7821. (s)
Frost, C. G.; Hartley, J. P.; Griffin, D. Synlett 2002, 1928 For use of
gaseous chlorine see. (t) Zincke, T.; Frohneberg, W. Ber. 1909, 42,
2721. (u) Wang, C.; Hamilton, C.; Meister, P.; Menning, C. Org.
Process Res. DeV. 2007, 11, 52, and refs 2-5 therein.
Sulfonyl chlorides are extremely important intermediates
exploited by a host of 21st century industries. They are of
particular importance to the agrochemical and pharmaceutical
industries where they are often used to prepare biologically
active sulfonamides.1 As an indication of their pharmaceutical
importance, a recent cross-pharma survey found that 9% of
candidate drugs contained a sulfonamide. Furthermore, of the
top 200 branded and generic drugs, those containing a sulfon-
amide moiety accounted for ∼$10 billion and $2 billion
worldwide sales, respectively, in 2007.2
* Author to whom correspondence may be sent. E-mail: neil.barnwell@
astrazeneca.com.
(1) (a) Labat, Y. Phosphorus, Sulfur Silicon Relat. Elem. 1993, 74 (1-
4), 173. (b) Lyga, J. W., Theodoridis, G., Eds.; Synthesis and Chemistry
of Agrochemicals VII; American Chemical Society: Washington DC,
2007. (c) Kleemann, A.; Engel, J., Kutscher, B., Reichert, D., Eds.;
Pharmaceutical Substances, Synthesis, Patents, Applications; Thieme:
Stuttgart, 1999. (d) Casini, A.; Scozzafava, A.; Mastrolorenzo, A.;
Supuran, C. T. Curr. Cancer Drug Targets 2002, 2, 55. (e) Hughes,
D. T. D. Sulfonamides, Antibiotic and Chemotherapy, 7th ed.; Churchill
Livingstone: Edinburgh, 1997; pp 460-468.
(2) (a) Carey, J. S.; Laffan, D.; Thomson, C.; Williams, M. T. Org. Biomol.
Chem. 2006, 4 (12), 2337. (b) Information compiled by the Njardarson
group at Cornell University. http://www.chem.cornell.edu/jn96/out-
reach.html.
(4) (a) Mosher, C. W.; Silverstein, R. M.; Crews, O. P.; Baker, J. R.;
Baker, B. R. J. Org. Chem. 1958, 23, 1257. (b) Langler, R. F. Can.
J. Chem. 1976, 54, 498. (c) Kwart, H.; Miller, R. K. J. Am. Chem.
Soc. 1956, 78, 5008. (d) Hardstaff, W. R.; Langler, R. F.; Leahy, J.
Can. J. Chem. 1975, 53, 2664.
278
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development
10.1021/op900241e  2010 American Chemical Society
Published on Web 01/15/2010

which these substrates react (under a standard set of conditions
used by us). Only one of these mechanistic manifolds has been
described before.4b-d This has allowed us to draw some general
conclusions about reactivity which should aid choice of
substrate.
A brief survey of the literature surrounding sulfonyl chloride
preparations revealed that Mosher et al. had prepared the des-
methyl analogue of the sulfonyl chloride 2 Via oxidative
chlorination of the disulfide cystine hydantoin in aqueous acetic
acid.4a Langler et al. had also shown that use of the benzyl group
to mask sulfur led to high yields of a range of alkylsulfonyl
chlorides (from alkyl benzyl thioethers) Via oxidative chlorina-
tion in acetic acid in the presence of water.4b Forearmed with
this information, a minimal development led to a process for
the preparation of 2 which was run in our kilo lab; see First-
Generation Oxidative Chlorination Procedure in the Experi-
mental Section for details. This reaction was conveniently found
to be self-indicating; on completion, the reaction temperature
dropped, green (chlorine) colouration could be seen and product
started to crystallize. At this stage of development the workup
and isolation involved two tedious toluene put-and-takes and
an isohexane drown out. Neverthelesss, the product was isolated
in an average of 93% yield, 99.6% purity over three batches.
Figure 1. Atmospheric pressure chlorine gas delivery system.
Valve A is corrosive gas regulator, B is a three-way polypro-
pylene tap, C is a universal variable tubing connector, D is a
4-port/2-way diagonal flow valve, E and F are gas-tight syringes
mounted on syringe pumps connected by Luer-locking syringe
connectors.8
required this capability. Although there are mass flow devices,
which are compatible with corrosive gases (chlorine and more
particularly damp HCl), their cost was considered prohibitive
for a short study such as this. Instead we set about constructing
a lab rig from inexpensive, readily available equipment. A
diagram showing the setup that we developed is given in Figure
1.
2.1.1. Description of Operation. The chlorine cylinder is
opened and the regulator adjusted so that the gas bubbles slowly
through the first scrubber. Valve D is turned so that syringe E
is connected to the chlorine supply. Syringe E is filled manually.
Valve D is switched so that syringe F is now connected to the
chlorine supply, and syringe E is connected to the reaction
vessel. The contents of syringe E are delivered to the reaction
vessel at a specified rate using syringe pump E while syringe
F is filled manually with chlorine. This process is repeated until
the required volume of chlorine has been delivered to the
reaction.8
2.2. Understanding the Workup and Isolation. It was
clear that the two, 5-volume put-and-takes of toluene followed
by an isohexane drown out were the time-limiting part of this
process. It had previously been assumed that these toluene put-
and-takes were necessary to remove acetic acid. Some detailed
examination of the reaction mixture and distillate solvent
composition during these distillations was extremely enlighten-
ing. We found, perhaps not surprisingly, that after the toluene
put-and-takes the acetic acid content was still fairly high but
that the water content was extremely low. It appeared that the
Clearly, this reaction had performed extremely well on scale-
up. However, we had little or no understanding of the process
and so were able to apply for minimum resource to support
further work based around process understandingswhich in our
experience is never wasted.
We also had concerns around the production of benzyl
chloride5 as a byproduct. This material is a suspect carcinogen;
as such, it is a potential genotoxic impurity (PGI),6 and its
environmental emission limits are extremely low.7 As we had
limited information on how benzyl chloride was partitioned
throughout our process, we also decided to examine the use of
alternative sacrificial sulfur substituents; with tert-butyl being
the most favored.
2. Results and Discussion
2.1. Handling Gaseous Chlorine in the Lab. We had
found, during early development work, that obtaining mass
balance information was difficult due to our inability to
accurately dispense small quantities of chlorine. In order to
conveniently obtain process understanding information we
(5) The production of benzyl chloride is accompanied by the concomitant
production of benzyl acetate in this reaction in a ratio of ∼ 3:1. Trace
amounts of benzyl alcohol, benzaldehyde, and benzoic acid are also
produced.
(8) Equipment list: valve A, corrosive gas regulator, ideally with cross
purge (Sigma-Aldrich Z40,605-8), alternative without cross purge
(Z148512-1EA), or basic control valve (Sigma-Aldrich Z146978-
1EA)scaution check lecture bottle thread size before ordering;
valve B, 3-way polypropylene tap (Thermo Fisher ADF-890-050R);
connector C, universal variable connector (Radleys 991008); valve
D, 4-port diagonal flow valve (Radleys 991114); tubing, 5 mm id
portex (Thermo Fisher TWT-200-064A), 1.6 mm id Teflon (Radleys
993011); Luer locking syringe connectors (Sigma-Aldrich Z117366-
1EA); gas-tight syringes, 100 mL Hamilton (www.esslab.com, cat.
no. 86020); syringe pumps, two-directional Harvard Apparatus
(model 70-2211).
(6) (a) International Conference on Harmonization Guideline Q3A(R);
Impurities in New Drug Substances; February 2002. (b) European
Medicines Agency, Committee for Medicinal Products for Human Use;
Guideline on the limits of genotoxic impurities, CPMP/SWP/5199/
02; London, U.K., June 2006. PGIs present as impurities in drug
substance are limited to levels “below the threshold of toxicological
concern” which is currently <1.5 µg day-1
.
(7) Carries an R-45 (may cause cancer) risk phrase, and as such the EU
solvent emissions directive limits emissions to <10 g h-1 or 2 mg
m-3
.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
279

2.3. Examination of Alternative Sacrificial Sulfur Sub-
stituents. While running our oxidative chlorinations we rou-
tinely compiled reaction profile plots using LC/MS data,11 and
recorded temperature profile12 information. This routine data
collection led to an unexpected mechanistic observation in the
reaction of substrate 1a. Langler proposed, in 1976, that this
reaction type most likely proceeded Via initial loss of the benzyl
cation and concomitant formation of the sulfenyl chloride.4b
Scheme 1 shows a possible mechanistic pathway which may
follow initial sulfenyl chloride formation proposed by Langler
(extrapolated to our substrate).4b We have designated this as
path A, and putative intermediates are shown in blue. We found
that, under our conditions and with the substrate 2, this reaction
proceeded Via initial formation of the sulfoxide 8a and not the
proposed sulfenyl chloride 3. Scheme 2 shows the reaction
pathway observed by us, path B, and again putative intermedi-
ates are shown in blue, whereas observed intermediates are
shown in black.
Figure 2. Solubility of 2 in isohexane and isohexane/acetic acid
mixtures. 2 was slurried in the relevant solvent or solvent
mixture (ratios are given in vol:vol) and solution concentrations
were measured after an equilibration period. The solution
concentration in mg/mL has also been related to loss of 2 to
liquor as a percentage of yield.
As previously mentioned, our favored alternative sacrificial
sulfur substituent was tert-butyl, mainly because t-butyl chloride
is more hydrolytically labile and so is not a potential carcinogen
or genotoxin;13 It also has the advantage of being more atom
efficient than a benzyl group.
large toluene distillation volume, while removing some acetic
acid, was mainly removing water.9
This led to the hypothesis that it was residual water that
solublised 2, and not acetic acid. We ran two sets of solubility
experiments to test this. The first, shown in Figure 2, was to
determine the solubility of 2 in various solvent mixtures
representative of workup compositions. The second, shown in
Figure 3, was to determine the effect of water on the solubility
of 2 in our actual workup mixture. We found that the presence
of acetic acid did indeed increase the solubility of 2, but this
effect was minor compared with that of water, which had a
dramatic effect on solubility. Spiking our workup with 5 and
10 mol equiv of water increased losses to liquor by 12% and
21%, respectively.
Conventional wisdom suggests that this reaction required 2
mol equiv of water,4d while our process was using 1 vol equiv
(∼14 mol equiv) of water. If our process would run with less
water, we would have less to remove after completion of the
reaction. We found that when using 4 mol equiv of water, the
process ran identically; whilst with 2 mol equiv, the theoretical
requirement, yield and purity were compromised. Running the
reaction with less water (4 mol equiv) completely removed the
need for the toluene put-and-takes. Overall this resulted in a
reduction in cycle time of 17% and total material usage by
33%,10 sillustrating the value of gaining process understanding
on an already high-yielding reaction. Figure 4 shows the
advantage of our new process from an operational point of view.
The final process conditions are described under Second-
Generation Oxidative Chlorination Procedure in the Experi-
mental Section. This process has been successfully run on 115
kg of starting hydantoin 1a, yielding 99.95 kg (92% yield) of
sulfonyl chloride 2.
When we subjected the tert-butyl masked hydantoin 1b to
our standard oxidative chlorination conditions, we were initially
irritated to find that the product was routinely contaminated with
5-10% of sulfonic acid 10.14 Subjecting sulfonyl chloride 2 to
our reaction conditions did not increase the level of sulfonic
acid showing that this was not due to hydrolysis. Examination
of the reaction profile led to the discovery that 1b was reacting
Via two competing reaction pathways, see Scheme 3. The first,
path A, is the same as that inferred by Langler,4b the second,
path C, to the best of our knowledge has not previously been
described in the literature. It seems that while the monomeric
manifold, path A, leads to clean sulfonyl chloride, the dimeric
manifold, path C, gives rise to the concomitant production of
sulfonic acid 10. We also examined the oxidative chlorination
of thiol 1c and disulfide 1d. To our surprise both of these
substrates also led to the production of sulfonic acid 10 in
5-10% yield.16
We investigated two methods whereby this reaction could
be diverted to proceed Via only path B and so eliminate sulfonic
(11) It is important to realize the limitations of the analytical techniques
being used. Using LC/MS we were aware that we would not see
hydrolytically unstable intermediates. In the LC traces we would not
see compounds without a UV chromophore; in the MS traces we were
restricted to an ES+ ionization source, and so only molecules able to
protonate would ionize.
(12) We have a proprietary in-house system which allows remote control
and monitoring of reactions.
(13) At least in part, as a result of the hydrolytic lability of tert-butyl
chloride, it carries only the risk phrase R-11 (highly flammable).
(14) A post-isolation re-work procedure, which involved slurrying the
product in water, did remove the sulfonic acid; however, this resulted
in a much-reduced yield.
(9) At the same time we modeled the ternary solvent system, and model
data confirmed our findings. The software used was Aspen BatchSep,
using physical properties modeled from Aspen Properties. The property
method employed was the NRTL method with the Hayden-O’Connell
equation of state for the vapor phase to account for any dimerisation
of the acetic acid.
(10) The process was modeled using Aspen BatchPlus with an internal
AstraZeneca “standard” plant configuration consisting of vessels up
to 6.3 m3.
280
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Figure 3. Effect of water on the solubility of 2 in isohexane/acetic acid workup mixture. After our standard reaction workup the
solution concentration of 2 and the water content were measured. The mixture was subsequently spiked with 5 mol equiv of water,
and after an equilibration period the same measurements were taken. This process was repeated once more.
Figure 4. Visio diagram showing workup unit operations pre- and post- process development.
acid formation. In the first instance we attempted to force the
tert-butyl masked substrate 1b to behave in the same manner
as 1a by partially oxidizing to the equivalent sulfoxide 8b prior
to chlorine treatment. This was achieved using hydrogen
peroxide and was extremely successful; product being isolated
in quantitative yield in the absence of sulfonic acid; see Scheme
4. This was also developed into a one-pot process which we
are confident can compete with our oxidative chlorination of
1a. This two-step one-pot process has the added advantage of
requiring less chlorine; the transformation from sulfoxide to
sulfonyl chloride requires only 2 mol equiv (compared to 3 mol
equiv with the original process).15,16
In the second instance we reasoned that limiting the substrate
concentration would reduce formation of disulfide 1d and so
preclude formation of sulfonic acid 10. This was achieved by
slow addition of 1b to a chlorine-saturated aqueous acetic acid
(15) Cornwall, P.; Horner, D. WO/2007/106021, 2007.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
281

Scheme 1. Possible mechanistic pathway following initial sulfenyl chloride formation; structures in black have been identified,
and structures in blue are proposed intermediates and byproducts
product.17 However, both processes remain as viable, economi-
cally attractive alternatives to our current method.
Scheme 2. Mechanistic pathway observed by us in the
oxidative chlorination of benzyl alkyl sulfide 1a
2.4. Understanding Sacrificial Sulfur Substituents. Al-
though there is a large body of literature concerning the use of
chlorine in oxidations of sulfur(II) species, the majority of
mechanistic information is directly related to the so-called
sulfohaloform reaction;18a whereby an alkyl group R-to sulfur
is exhaustively chlorinated prior to cleavage and oxidation of
the resulting sulfenyl chloride to a sulfonyl chloride, see Scheme
5.18f A large portion of this work is focused around explaining
the regiochemistry of chlorine attack on variously substituted
dialkyl thioethers.18a–e On close examination of this work we
found that the results which are reported did not correspond to
our findings for 1a-d. For instance, it is asserted that chlorina-
solution. Again, this hypothesis proved successful as no sulfonic
acid was observed when 1b was added to the chlorination
tion of thioethers proceeds Via initial formation of a sulfoxide
only when there is a strongly electron withdrawing group R-to
sulfur,18a whereas we had found that the benzyl masked substrate
1a proceeded initially through a sulfoxide. Also, as the benzylic
methylene is not exhaustively chlorinated, benzyl chloride being
the byproduct, we felt that this substrate did not fit the criteria
at all for the sulfohaloform reaction; the literature puts forward
no explanation for this.18b Assessment of the relative electrone-
gativity (Xp) of the groups on either side of sulfur has been
solution over a 1 h period. Sulfonyl chloride 2 being isolated
in 96% yield and 99.8% purity. The chlorine usage was not
measured in this experiment, so cannot be compared to our
standard chlorination of 1a, or our two-step oxidation-chlorina-
tion of 1b.
Neither of the two alternative processes described above was
progressed past the proof of concept stage, as by this time we
had data to show that we could control benzyl chloride, both
with respect to environmental emissions and contamination of
(17) Due to the relatively low vapor pressure of benzyl chloride we were
able to demonstrate that the potential air emissions could be success-
fully controlled either by manipulating the distillation conditions or
by utilizing an efficient condenser.
(18) (a) Baum, J. C.; Hardstaff, W. R.; Langler, R. F.; Makkinje, A. Can.
J. Chem. 1984, 62, 1687. (b) Langler, R. F.; Marini, Z. A.; Spalding,
E. S. Can. J. Chem. 1979, 57, 3193. (c) Ahern, T. P.; Kay, D. G.;
Langler, R. F. Can. J. Chem. 1978, 56, 2422. (d) Potvin, M.; Albrecht,
L.; Darvesh, K. V.; Langler, R. F. Aust. J. Chem. 2005, 58, 143. (e)
Ginsburg, L. G.; Darvesh, K. V.; Axworthy, P.; Langler, R. F. Aust.
J. Chem. 1997, 50, 517. (f) Grossert, J. S.; Langler, R. F. Can. J. Chem.
1977, 55, 407.
(16) This was not intended to be a rigorous mechanistic study. The
mechanistic information was gained as a result of routine data
collection. We do not profess to have fully elucidated these mecha-
nisms, and questions still remain around sulfonic and sulfinic acid
formation. Notably: (i) only a small proportion of the material that
progresses via path C, the dimer path, becomes sulfonic acid; (ii) in
all cases sulfonic acid is formed towards the end of the reaction when
the thiosulfonate concentration tends to be at its highest; (iii) we see
no cross-over from path C to path A with substrate 1d even though
we might expect to.
282
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Scheme 3. Mechanistic pathway observed by us in the oxidative chlorination of tert-butyl alkyl thioether 1b
electronegativity.18c This breaks down when sulfur is masked
by a group capable of stabilizing a carbocation. Putting a
theoretical basis to this empirical observation has not proven
to be trivial.18d,e Also, these treatments, while extremely
interesting, cover only a narrow substrate scope, which is not
immediately helpful from the point of view of predicting
preparatively useful trends for sulfonyl chloride formation.
In order to try and gain some understanding of the observed
behaviour of sulfides 1a-d and place them into a broader, and
preparatively useful, reactivity context, we prepared a series of
substrates with various sacrificial sulfur substituents 1e-l, and
also prepared the sulfones 15a-b. We subjected each of the
Scheme 4. Two-step oxidation-chlorination process
Scheme 5. Postulated mechanistic pathway for the
sulfohaloform reaction
substrates in turn, first to our oxidative chlorination conditions,
and then to our two-step peroxide/chlorine oxidation-chlorina-
tion. We followed the course of each reaction generating profiles
from LC/MS data, observing whether the reaction proceeded
Via paths A, B, C, D or multiple manifolds. The results are
summarized in Table 1. See Supporting Information for LC
and LC/MS traces, MS assignments, along with full data for
isolated intermediates and products.
Whilst these reactions are far from fully characterized, it is
apparent that the course of the reaction is strongly influenced
by the groups masking sulfur.16 The reactivity trend observed
with varying sacrificial substituents was found to mirror the
relative carbocation stabilizing ability of these substituents. i.e.
Those sacrificial sulfur substituents that form, relatively, the
most stable carbocations (acyl, 1e, 1f, 1g and tert-butyl 1b) react
Via path A/C. Those groups which may form carbocations of
used successfully to predict the regiochemistry of chlorine
attack; chlorine attacking the side with the largest relative
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
283

Table 1. Observed reaction pathways for 1a-l, 15a-b
has been partly described in the literature previously. We saw
no evidence for the sulfohaloform reaction, path D, under these
conditions. We suggest that the manifold by which a substrate
reacts is empirically predictable from the carbocation stabilizing
ability of the sacrificial sulfur substituent. This order of reactivity
is consistent with the relative carbocation stabilities shown in
Table 2.19
In theory our alkyl-alky- and alkyl-aryl-substituted thioethers
reacting Via the sulfoxide-sulfohaloform reaction, path E,
consume 4 mol equiv of chlorine, i.e. are less efficient in terms
of chlorine usage.
We have LC/MS evidence only for the proposed mono- and
dichlorosulfoxide intermediates, 16 and 17, shown in Scheme
6. We have isolated and characterized the trichloromethylhy-
dantoin byproduct 19 from this reaction. The regiochemistry
of the reaction with 1l is unambiguous, 14l being identical to
an authentic sample of phenylsulfonyl chloride. Whereas, we
cannot rule out mixtures of regioisomers with 1j and 1k
although no sulfonyl chloride 2 was detected, suggesting that
these reactions are also regioselctive (it is unlikely that we would
observe 1,1,1-trichloroethane or trichloromethylcyclohexane by
LC/MS).
two-step
oxidation-chlorination
substrate
one-step chlorination
1a
1b
1c
1d
1e
1f
1g
1h
1i
1j
1k
1l
15a
15b
path Ba
not done
paths A and Cb
paths A and Cb
path Cb
path Ba
not done
not done
path Cd
paths A and Cb,c
paths A and Cb
paths A and Cb
path Ba
path Cd
path Cd
path Ba
path Ba
Path Ba
path Ee
path Ee
path Ee
path Ee
path Ee
path Ee
no reactione
no reactionf
no reactionf
no reactionf
a See Scheme 2 for proposed reaction path. b See Scheme 3 for proposed
reaction path. c Substrate 1e was a mixture of the thioester and thiol. Acyl groups
of 1e, 1f, and 1g, were slowly hydrolysed in acetic acid (cf.: reactivity of
thioesters is much greater than carboxylic esters). d 1e, 1f and 1g were oxidised to
disulfide 1d and not sulfoxides under the reaction conditions. e These reactions
proceeded Via the sulfoxide-sulfohaloform pathway, see Scheme 6 for proposed
reaction path. f The sulfones, 15a and 15b, were stable to the reaction conditions,
however some monochlorination was seen after stirring overnight. We did not
determine the position of chlorination.
Scheme 6. Observed path for sulfoxide-sulfohaloform
reaction
3. Conclusions
We hope that this piece of work highlights the value of
seeking process understanding as a stand-alone aim and that
routine data collection is a tool to this end. We have described
a low-tech solution to the problems associated with chlorine
delivery, which has reduced the inertia towards the use of
gaseous chlorine in our laboratories.
This work has implications for choice of sulfur substituents
when preparing sulfonyl chlorides using gaseous chlorine under
aqueous conditions.
In order to maximize yield and limit loss to sulfonic acid
formation, extremely labile, carbocation stabilizing groups,
should probably be avoided (although, we have developed a
two-step one-pot procedure which precludes sulfonic acid
formation with these substrates). With aryl-alkyl sulfides,
primary alkyl groups will perform the masking function
adequately, although chlorine usage would be minimized with
a secondary alkyl group. The isopropyl group would appear to
be the sacrificial group of choice, from the point of view of
atom efficiency and chlorine usage. From an environmental and
hazard point of view there is little difference between isopropyl
and tert-butyl sacrificial groups (the byproduct being isopro-
pylchloride and tert-butylchloride). Use of the ethyl group to
mask sulfur is much less attractive, the byproduct being 1,1,1-
trichloroethane which is highly toxic to humans, the environ-
ment and a potential carcinogen.
intermediate stability (benzyl 1a, isopropyl 1h, and crotyl 1i)
react Via path B. Those groups which form the least stable
carbocations (ethyl 1j, cyclohexylmethyl 1k, and phenyl 1l)
appear to react Via an entirely different manifold, path E, which
we have designated the sulfoxide-sulfohaloform reaction, see
Scheme 6. In sharp contrast to literature precedent, in all cases,
except 1e, 1f, 1g and 1b where the sacrificial group is cleaved
immediately, we see sulfoxide formation as the first step, prior
to R-chlorination.18a We herein propose four mechanistic
manifolds for the reaction of sulfides under our aqueous
chlorination conditions, paths A, B, C, and E, only one of which
Of course, a range of considerations must be taken into
account in order to successfully “design” the ideal substrate for
an oxidative chlorination. For instance, stability of product
sulfonyl chloride (can it be isolated); how the product may be
isolated (distillation, crystallization); will byproduct interfere
with down-steam chemistry, etc.? Although the above study is
(19) (a) Schultz, J. C.; Houle, F. A.; Beauchamp, J. L. J. Am. Chem. Soc.
1984, 106, 3917. (b) Lossing, F. P.; Holmes, J. L. J. Am. Chem. Soc.
1984, 106, 6917.
284
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Table 2. Gas-phase dissociation energies for R-H f R+ + H-, in kJ mol-1
a No literature data available for gas-phase dissociation energies. b No data available, allyl cation ) 1070 kJ mol-1
.
not exhaustive, we hope that it will help guide chemists towards
an informed choice of sulfur substituents, which will give
byproduct with acceptable properties, when preparing sulfonyl
chlorides.
(33 L, 0.2 vol equiv, 4 mol equiv), at ambient temperature. The
solution was warmed to 33 °C. Chlorine gas (110.44 kg, 3.3
mol equiv) was blown into the reaction mixture over ∼28 h,
maintaining Ti at 32-37 °C resulting in an intense yellow
suspension which was stirred at 32-37 °C for a further 10 h.
The mixture was purged with nitrogen for ∼3 h, then degassed
in Vacuo at 230 mbar, Ti ) 35-25 °C for 28 h. The reaction
mixture was distilled (770 L solvent mixture removed (acetic
acid, water, benzyl chloride)) in Vacuo (230 f 19 mbar, Ti )
37.5-40 °C). The contents were cooled to 30 °C and stirred
for 8 h. Heptane (114 kg, 1 vol equiv) was charged, and the
product was collected in portions by means of a centrifuge. Each
portion was washed with warm (40 °C) heptane. The isolated
material was dried under vacuum at 3 bar, 45 °C to constant
weight, yielding the product, 2, 95.95 kg (91.8% yield, 98.7%
4. Experimental Section
4.1. General. NMR spectra were run on Varian Unity Inova
spectrometers running at proton frequency of 300 or 400 MHz,
coupling constants are quoted in hertz. HPLC samples were
run on HP 1100 series with binary pump and diode array
detector, using a Metachem Polaris C18 (3 µm × 150 mm ×
3 mm) column and mobile phases water (containing 0.05%
TFA) and acetonitrile (containing 0.04% TFA), running with
gradient elution. LC/MS spectra were run using the same HPLC
system described above connected to an Agilent 1100 series
SL LC/MSD, using APESI +ve ionization mode. Accurate
masses were obtained using a Waters LCT time-of-flight MS
(lockspray internal calibration) with loop injection and ESI +ve
ionization mode. FTIR spectra were obtained using a Perkin-
Elmer Spectrum One spectrometer; all samples were run neat.
Melting points were obtained using a Buchi melting point
B-545.
4.2. First-Generation Oxidative Chlorination Procedure
(2). Glacial acetic acid (22 L, 8 vol equiv) was charged to the
reactor followed by 1a (2.75 kg). Distilled water (2.75 L, 1 vol
equiv) was added, and the reaction mixture was cooled to 8
°C. Chlorine gas was introduced Via a dip pipe at such a rate
that Tinternal remained around 12 °C. After completion of the
reaction (∼2.5 h) the solution turned green and the reaction
temperature dropped. The gas flow was stopped, the temperature
was raised to 25 °C and nitrogen was passed through the mixture
for 1 h. Acetic acid was removed by reduced pressure distillation
(down to 11 L, 4 vol equiv remaining, Ti <40 °C). Two toluene
put-and-takes were carried out (2 × 13.75 L, 2 × 5 vol equiv).
Isohexane (13.75 L, 5 vol equiv) was added followed by an
18 h room temperature stir out. The product was collected by
filtration, and the filter cake was washed with isohexane (2 ×
11 L, 2 × 4 vol equiv) and dried overnight in a vacuum oven
at 45 °C to give the product, 2.35 kg (94% yield, 99.6% purity
by LC area), as a white crystalline solid.
1
w/w by H NMR), as a white crystalline solid (GC: benzyl
chloride: 9 ppm).
4.4. ((S)-4-Methyl-2,5-dioxo-imidazolidin-4-yl)methane-
sulfonyl Chloride (2).15 1H NMR (400 MHz, d8-THF) δ 1.52
(3H, s), 4.44 (1H, d, J 14.6), 4.53 (1H, d, J 14.6), 7.58 (1H, s),
9.92 (1H, s). 13C NMR (75 MHz, d6-DMSO) δ 22.980, 55.682,
60.017, 156.207, 177.820. Mp ) 185-190 °C. FTIR 1704,
1408, 1287 cm-1. m/z (%) from LC/MS: 127.20 (100), 227.20
(9, [M + H]+).
4.5. (R,S)-5-Methyl-5-phenylmethanesulfinylmethyl-imi-
dazolidine-2,4-dione (8a). To a solution of (rac)-1a (4.10 g,
16.4 mmol) in acetone (100 mL) was added 35% H2O2 (1.62
mL, 1.1 mol equiv). The solution was stirred at ambient
temperature overnight. The solution was tested with a peroxide
strip and found to contain no peroxide. It was then concentrated
to dryness in Vacuo to give the product, 4.4 g (100%), as a
white solid which was not purified further.
1
Ratio of diastereoisomers ) 59:41 by LC area. H NMR
(400 MHz, d6-DMSO) major isomer δ 1.34 (3H, s), 2.75 (1H,
d, J 13.8), 3.2 (1H, d, J 13.8), 4.02 (2H, d, J 12.8), 7.30-7.40
(5H, m), 8.07 (1H, bs), 10.83 (1H, bs); minor isomer δ 1.36
(3H, s), 2.97 (1H, d, J 13.5), 3.15 (1H, d, 13.5), 4.17 (2H, t, J
13.4), 7.30-7.40 (5H, m), 8.31 (1H, bs), 10.67 (1H, bs). 13C
(75 MHz, d6-DMSO) It was not possible to separate major and
minor isomers, so signals are quoted in pairs, except for the
benzylic carbons signals which are coincedent δ 23.53, 24.39;
57.28; 57.74, 58.11; 60.41, 59.32; 127.82, 127.84; 128.519,
128.43; 130.20, 130.35; 130.96, 131.29; 156.09, 155.75; 176.67,
4.3. Second-Generation Oxidative Chlorination Proce-
dure (2). The reactor was charged with 1a (114.9 kg) followed
by glacial acetic acid (964 kg, 8 vol equiv) and purified water
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
285

176.76. Mp ) 208-210 °C. m/z (%) from LC/MS: minor
isomer 267.20 (100, [M + H]+), 533.20 (2, [2M + H]+); major
isomer 267.20 (100, [M + H]+), 533.20 (13, [2M + H]+).
4.6. (S)-5-Methyl-5-(2-methyl-propane-2-sulfinylmeth-
yl)imidazolidine-2,4-dione (8b).15 To a solution of 1b (3.93
g, 18.2 mmol) in acetone (200 mL) was added 35% H2O2 (1.8
mL, 1.1 mol equiv). The solution was stirred at ambient
temperature for 19 h. It was tested with a peroxide strip and
found to contain no peroxide. The solvent was removed in
Vacuo to give the product, 4.403 g (104% yield), as a white
solid.
1H NMR (400 MHz, d6-DMSO) δ 1.34 (3H, s), 3.43 (1H,
d J 15), 3.55 (1H, d, J 15), 4.45 (1H, d, J 13.6), 4.49 (1H, d, J
13.6), 7.36-7.43 (5H, m), 8.18 (1H, s), 10.82 (1H, s); 13C NMR
(75 MHz, d6-DMSO) δ 24.46, 55.96, 59.05, 60.12, 128.02,
128.46, 131.12, 156.23, 176.35. Mp (decomp.) 182-183 °C.
FTIR 1114, 1307, 1706 cm-1. m/z (%) from LC/MS: 283.20
(39, [M + H]+), 300.20 (100, [M + NH4]+), 301.20 (19, [M
+ H2O + H]+), 321.20 (1, [M + K]+), 565.20 (62, [2M +
H]+). Accurate mass, calc. for [M + NH4]+ C12H18N3O4S )
300.1018, found ) 300.1044 (8.7 ppm error).
4.10. (S)-5-Methyl-5-(2-methyl-propane-2-sulfonylmeth-
yl)imidazolidine-2,4-dione (15b). 1b (1.0129 g, 4.68 mmol)
was dissolved in acetone (25 mL) and added dropwise to a
slurry of oxone (9.4027 g; 15.29 mmol) in acetone (20 vol
equiv) with ice cooling. The exotherm was allowed to subside
and the reaction was heated at 40 °C for 4 h. The resulting
slurry was filtered and the filtrate concentrated in Vacuo then
dried to give the product 1 g (64% yield) as a white solid (74%
w/w by 1H NMR assay).
1
Mixture of diastereoisomers, 71:29, by comparison of H
NMR integral. 1H NMR (500 MHz, d6-DMSO) major isomer
δ 1.13 (9H, s), 1.40 (3H, s), 2.39 (1H, d, 14), 3.25 (1H, d, J
14), 8.28 (1H, s), 10.78 (1H, s); minor isomer δ 1.14 (9H, s),
1.38 (3H, s), 2.85 (2H, s), 8.07 (1H, s), 10.78 (1H, s). FTIR of
mixture 1705, 1016 cm-1. 13C NMR (75 MHz, d6-DMSO)
major isomer δ 22.01, 22.28, 60.57, 53.11, 54.93, 155.75,
176.89; minor isomer δ 22.25, 24.64, 59.33, 52.57, 50.22,
156.10, 176.68. Mp ) 166-170 °C. m/z (%) from LC/MS:
major isomer 233.20 (56, [M + H]+), 465.20 (100, [2M + H]+);
minor isomer 177.20 (100), 233.20 (11, [M + H]+), 465.20 (1,
[2M + H]+).
1H NMR (400 MHz, d6-DMSO) δ 1.27 (9H, s), 1.36 (3H,
s), 3.22 (1H, d, J 14.5), 3.71 (1H, d, J 14.5), 7.99 (1H, s), 10.67
(1H, s). 13C NMR (75 MHz, d6-DMSO) δ 22.25, 24.99, 50.23,
58.94, 59.14, 156.03, 176.40. Mp (decomp.) 202-209 °C. FTIR
1109, 1285 1707 cm-1. m/z (%) from LC/MS: 193.20 (100,
[M-C(CH3)3 + H]+), 249.20 (100, [M + H]+), 266.20 (59, [M
+ NH4]+), 267.20 (9, [M + H2O + H]+), 290.20 (8, [M + H
+ MeCN]+), 497.20 (90, [2M + H]+). Accurate mass, calc.
for [M + MeCN + H]+ C11H20N3O4S ) 290.1175, found )
290.1198 (7.9 ppm error).
4.11. (S)-5-Mercaptomethyl-5-methyl-imidazolidine-2,4-
dione (1c). 1b (45 g, 208 mmol) was heated at reflux in
concentrated HCl (720 mL, 16 vol equiv) at 90 °C for 7 h then
allowed to cool to room temperature overnight. Onset of rapid
gas evolution occurred at 55 °C; the exhaust was scrubbed
(water, and 10 M NaOH trap). The reaction mixture was
concentrated in Vacuo then reslurried in toluene. The resulting
free-flowing solid was collected by filtration and dried overnight
in Vacuo to give the product, 33.3 g (100% yield, 90% w/w
purity by 1H NMR).
4.7. ((S)-4-Methyl-2,5-dioxo-imidazolidin-4-yl)methane-
sulfinic Acid (6).20 To a solution of sodium sulfite (16.68 g,
132.37 mmol) in water (85 mL) was added 2 (10 g, 44.12
mmol) in one portion. After 2 h the reaction was acidifed with
HCl to pH 1-2 then concentrated in Vacuo. The resulting solid
was slurried in hot IPA, filtered, and washed twice with hot
IPA. The combined organics were concentrated in Vacuo to give
the product, 2 g (24% yield), as a white solid. NMR analysis
revealed that this material was a mixture of sulfinic acid and
sulfonic acid in a ratio of 5.9:1, 84.4% w/w by 1H NMR.
1H NMR (300 MHz, D2O) δ 1.54 (3H, s), 2.69 (1H, d, J
13.5), 2.92 (1H, d, J 13.6). m/z (%) from LC/MS: 193.20 (100,
[M + H]+), 234.20 (5, [MH + MeCN]+) 256.40 (8, [MH +
Na + MeCN]+).
4.8. ((S)-4-Methyl-2,5-dioxo-imidazolidin-4-yl)methane-
sulfonic Acid (10).15 A slurry of 2 (2 g, 8.82 mmol) in water
(10 mL) was heated at reflux for 24 h. The resulting solution
was allowed to cool and concentrated in Vacuo. The white solid
obtained was dried to constant weight, slurried in hot MeCN
and collected by filtration to give the product 1.808 g (98%
yield) as a white crystalline solid.
1H NMR (300 MHz, d6-DMSO) δ 1.31 (3H, s, J 5.4), 2.24
(1H, dd, J 9.8, 7.6), 2.64 (1H, dd, J 13.8, 7.6), 2.75 (1H, dd, J
13.8, 9.8), 7.85 (1H, bs), 10.69 (1H, bs). 13C NMR (75 MHz;
d6-DMSO) δ 22.73, 30.87, 63.16, 158.57, 177.09. Mp )
150-153 °C. m/z (%) from LC/MS: 116.20 (100), 161.20 (20,
[M + H]+).
4.12. (5S,5′S)-5,5′-Disulfanediylbis(methylene)bis(5-me-
thylimidazolidine-2,4-dione) (1d).21 1b (5 g, 23.12 mmol) was
stirred in acetic acid (15 mL, 3 vol equiv) with DMSO (1.65
mL, 23.12 mmol) at 5-10 °C. Hydrogen bromide (15.17 mL,
134.1 mmol) was added dropwise Via syringe. Gas evolution
and an endotherm were noted. The reaction was allowed to
warm to room temperature. After 5 h crushed ice was added,
and the resulting solid was collected by filtration and dried in
Vacuo to give the product 3.64 g (99% yield) as a white solid
(97% w/w assay by 1H NMR assay).
1H NMR (400 MHz, d8-THF) δ 1.46 (3H, s), 3.37 (1H, d,
J 14.9), 3.45 (1H, d, J 14.6), 7.21 (1H, s), 9.66 (1H, s). 13C
NMR (75 MHz, D2O) δ 24.09, 56.12, 61.76, 158.84, 180.25.
Mp ) 137-139 °C.
4.9. (S)-5-Methyl-5-phenylmethanesulfonylmethyl-imida-
zolidine-2,4-dione (15a). 1a (20.06 g, 72.13 mmol) was slurried
with oxone (150 g, 243.99 mmol) in acetone (200 mL) at 45
°C overnight. Inorganics were removed by filteration and
washed with acetone. The filtrate was concentrated to a white
foam, then reconcentrated from methanol. The resulting material
was dried in Vacuo to give the product 22.74 g (111% yield)
as a white solid (95.6% w/w by 1H NMR assay).
(21) (a) Shibuya, A.; Saito, M. JP 2004043309, 2004. (b) Matsumoto, S.;
Murao, H.; Yamaguchi, T.; Izumida, M.; Ueda, Y. WO/2005/026110,
2005.
(20) Method adapted from that described by King, J. F.; Hillhouse, J. H.
Can. J. Chem. 1983, 61, 1583.
286
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Single crystals were grown from water, and the structure
was confirmed by X-ray analysis.
1H NMR (400 MHz, CDCl3) δ 1.18-1.94 (10H, m), 1.55
(3H, s), 2.53 (1H, tt, J 3.59, 11.28), 3.07 (1H, d, J 14.2), 3.39
(1H, d, J 14.2), 5.45 (1H, s), 7.99 (1H, s). 13C NMR (75 MHz,
CDCl3) δ 23.19, 26.42, 26.48, 26.76, 30.60, 30.90, 35.44, 53.95,
64.42, 158.80, 178.97, 202.30. Mp ) 195-197 °C. FTIR 1702,
2854, 2935 cm-1. m/z (%) from LC/MS: 288.20 (4, [M + H2O
+ H]+). Accurate mass, calc. for [M + MeCN + H]+
C14H22N3O3S ) 312.1382, found ) 312.1402 (6.4 ppm error).
4.17. Thiobenzoic Acid S-((S)-4-Methyl-2,5-dioxo-imi-
dazolidin-4-ylmethyl) Ester (1g). 1g was isolated in 92% yield,
97% w/w by 1H NMR assay.
1H NMR (300 MHz, d6-DMSO) δ 1.30 (6H, s), 3.08 (4H,
s), 7.98 (2H, bs), 10.75 (2H, bs). 13C NMR (75 MHz;
d6-DMSO) δ 23.02, 48.05, 62.34, 156.13, 176.67. Mp )
278-281 °C. m/z (%) from LC/MS: 319.20 (100, [M + H]+).
Accurate mass, calc. for [M + H]+ C10H15N4O4S2 ) 319.0535,
found ) 300.0524 (3.4 ppm error).
4.13. S-((S)-4-Methyl-2,5-dioxoimidazolidin-4-yl)methy((S)-
4-methyl-2,5-dioxoimidazolidin-4-yl)methanesulfonothio-
ate (9). 1b (18.1 g, 83.7 mmol) was charged to a mixture of
acetic acid (145 mL, 8 vol equiv) and water (4.52 mL, 3 mol
equiv). Chlorine gas was bubbled through the reaction mixture
until the reaction mixture turned bright green (Ti kept <10 °C).
The uptake of chlorine to this point was ∼1.70 mol equiv. A
sample of the reaction mixture (11.4 g) was removed and
concentrated in Vacuo to give an oil that solidified on standing.
This material was slurried in isohexane and collected by filtation
to give the product 0.9 g as a white solid.
1H NMR (400 MHz, d6-DMSO) δ 1.40 (3H, s), 3.38 (1H,
d, 13.8), 3.44 (1H, d, J 13.8), 7.55-7.94 (5H, m), 8.08 (1H, s),
10.76 (1H, s). 13C NMR (75 MHz, d6-DMSO) 22.74, 34.85,
61.79, 126.94, 129.15, 134.17, 135.85, 156.08, 176.71, 189.38.
Mp ) 161.4-162 °C. FTIR 1706 cm-1. m/z % from LC/MS:
176.20 (100), 265.20 (11, [M + H]+), 282.20 (11, [M + H2O
+ H]+), 529.00 (10, [2M + H]+). Accurate mass, calc. for [M
+ MeCN + H]+ ) 306.0912, found ) 306.0914 (0.7 ppm
error).
1H NMR (400 MHz, d6-DMSO) δ 1.33 (3H, s), 1.39 (3H,
s), 3.45 (2H, m), 4.03 (2H, m), 8.15 (1H, d, J 1.4), 8.19 (1H,
d, J 1.4), 10.87 (1H, s), 10.93 (1H, s).
4.18. (S)-5-Isopropylsulfanylmethyl-5-methyl-imidazoli-
dine-2,4-dione (1h).22 1h was isolated in 51% yield, 98.8%
w/w by 1H NMR assay.
4.14. General Procedures for the Preparation of Masked
Intermediates. 4.14.1. Acyl-Masked 1e, 1f, and 1g. To a
solution of 1c in MeCN (5-8 vol equiv) and an alkylating agent
(1 mol equiv) at -5 °C was added Et3N (1 mol equiv) dropwise.
When complete, the reaction was filtered, the filtrate concen-
trated in Vacuo and the resulting solid hot slurried in water then
dried to constant weight in Vacuo. Alkylating agents: for 1e
acetic anhydride was used, for 1f cyclohexanoyl bromide was
used, for 1g benzoyl bromide was used. Deviation from general
procedure: Et3NHOAc was separated from 1e by precipitation
with acetone; 1e was then crystallised from toluene.
4.14.2. Alkyl-Masked 1h, 1i, 1j, and 1k. To a solution of 1c
and an alkylating agent (2 mol equiv) in MeOH (10 vol equiv)
at 60 °C was added NaOMe solution (25% w/w in MeOH, 8
mol equiv) over ∼15 min. After the reaction was complete (by
HPLC) it was quenched with HCl (5 M, 9 mol equiv). The
resulting solid was removed by filtration, and the filtrate was
concentrated to dryness. The residue was slurried with water,
collected by filtration, and dried to constant weight. Alkylating
agents: for 1h isopropyl bromide was used, for 1i crotyl bromide
was used, for 1j ethyl bromide was used, and for 1k bromom-
ethylcyclohexane was used.
1H NMR (400 MHz, d6-DMSO) δ 1.16 (6H, t, J 5.7), 1.30
(3H, s), 2.73 (1H, d, J 13.6), 2.78 (1H, d, J 13.6), 2.94 (1H,
sep, J 6.6), 7.88 (1 H, s), 10.63 (1 H, s). 13C (75 MHz, d6-
DMSO) δ 23.36, 35.43 (broad), 37.19, 62.82, 156.34, 177.41.
Mp ) 185.3-186.6 °C (lit. 161-162 °C).22 FTIR: 1705 cm-1.
m/z (%) from LC/MS: 203.20 (100, [M + H]+), 244.20 (33,
[M + MeCN + H]+), 405.20 (10, [2M + H]+). Accurate mass,
calc. for [M + MeCN + H]+ C10H18N3O2S ) 244.1120, found
) 244.1143 (9.4 ppm error).
4.19. (S)-5-[((E)-But-2-enyl)sulfanylmethyl]-5-methyl-
imidazolidine-2,4-dione (1i). 1i was isolated in 62% yield, 95%
1
w/w by H NMR. Product was ∼85:15 mixture of trans:cis
Major isomer only is reported.
1H NMR (400 MHz, d6-DMSO) δ 1.28 (3H, s), 1.65 (3H,
d, J 6.3), 2.61 (1H, d, J 13.9), 2.71 (1H, d, J 13.9), 3.10 (2H,
d, J 7.3), 5.32-5.40 (1H, m), 5.51-5.60 (1H, m), 7.29 (1H, s),
10.66 (1H, s). 13C NMR (75 MHz, d6-DMSO) δ 17.39, 23.12,
34.31, 37.25, 62.94, 127.11, 127.95, 156.38, 177.45. Mp )
132-134 °C. FTIR: 960, 1705 cm-1. m/z (%) from LC/MS:
190.20 (100), 215.20 (9, [M + H]+), 233.20 (1, [M + H2O +
H]+). Accurate mass, calc. for [M + MeCN + H]+
C11H18N3O2S ) 256.112, found ) 256.1128 (3.1 ppm error).
4.20. (S)-5-Ethylsulfanylmethyl-5-methyl-imidazolidine-
2,4-dione (1j).22 1j was isolated in 44.5% yield, 98% w/w by
1H NMR assay.
4.15. Thioacetic Acid S-((S)-4-Methyl-2,5-dioxo-imida-
zolidin-4-ylmethyl) Ester (1e). 1e was isolated in 63% yield,
96.5% w/w by 1H NMR assay.
1H NMR (400 MHz, d6-DMSO) δ 1.32 (3H, s), 2.35 (1H,
s), 3.13 (1H, d, J 13.8), 3.21 (1H, d, J 13.8), 7.99 (1H, s), 10.72
(1H, s). 13C NMR (75 MHz, d6-DMSO) δ 22.63, 30.46, 34.98,
61.59, 156.04, 176.69, 193.62. Mp ) 89-90 °C. FTIR 1699.
m/z (%) from LC/MS: 114.20 (100), 203.20 (27, [M + H]+),
220.20 (5, [M + H2O + H]+), 405.20 (4, [2M + H]+). Accurate
mass, calc. for [M + MeCN + H]+ C9H14N3O3S ) 244.0756,
found ) 244.0734 (9 ppm error).
1H NMR (400 MHz, CD3OD) δ 1.19 (3H, t, J 7.43), 1.39
(3H, s), 2.51-2.64 (2H, m), 2.79 (1H, d, J 14.1), 2.86 (1H, d,
J 14.1). 13C NMR (75 MHz, d6-DMSO) δ 14.83, 23.15, 26.67,
38.33, 63.04, 156.40, 177.47. Mp ) 129.1-129.5 °C (lit.
113-114 °C).22 FTIR: 1704 cm-1. m/z (%) from LC/MS:
189.20 (100, [M + H]+), 377.20 (30, [2M + H]+). Accurate
mass, calc. for [M + MeCN + H]+ C9H16N3O2S ) 230.0963,
found ) 230.0986 (10 ppm error).
4.16. Cyclohexanecarbothioic Acid S-((S)-4-Methyl-2,5-
dioxo-imidazolidin-4-ylmethyl) Ester (1f). 1f was isolated in
85% yield, 91.5% w/w by 1H NMR assay.
(22) Tahara, S.; Obata, Y. Agric. Biol. Chem. 1971, 35 (1), 53.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
287

4.21. (S)-5-Cyclohexylsulfanylmethyl-5-methyl-imidazo-
lidine-2,4-dione (1k). 1k was isolated in 88%, 100% w/w by
1H NMR assay.
1H NMR (400 MHz, d6-DMSO) δ 0.86-1.82 (11H, m), 1.28
(3H, s), 2.43 (2H, d, J 7.0), 2.69 (1H, d, J 13.9), 2.76 (1H, d,
J 13.9). 13C NMR (75 MHz, d6-DMSO) δ 23.19, 25.43, 25.88,
31.95, 37.43, 63.10, 156.36, 177.42. Mp ) 185.3-186.3 °C.
FTIR: 1703, 2852, 2923 cm-1. m/z (%) from LC/MS: 186.20
(100), 257.20 (20, [M + H]+), 298.20 (25, [M + MeCN +
H]+). Accurate mass, calc. for [M + MeCN + H]+
C14H24N3O2S ) 298.1589, found ) 298.1583 (2 ppm error).
4.22. (S)-5-Methyl-5-phenylsulfanylmethylimidazolidine-
2,4-dione23,25 (1l). 1l was prepared using the procedure de-
scribed by Xu et al.24 and purified by column chromatography
(eluant 2:1, EtOAc/iHex) to give the product 1.45 g (16%) as
a white solid.
Figure 5. X-ray crystal structure of disulfide 9 (the Flack’s ×
parameter was refined to -0.01(9)).
1H NMR (400 MHz, d6-DMSO) δ 1.35 (3H, s), 3.25 (2H,
s), 7.20 (1H, t, J 7.2), 7.30 (2H, t, J 7.2), 7.36 (2H, d, J 7.2),
7.97 (1H, s), 10.74 (1H, s). 13C NMR (75 MHz, d6-DMSO) δ
23.18, 40.56, 62.52, 126.06, 128.79, 128.92, 135.84, 156.23,
176.96. Mp ) 147-149 °C (lit. 155-157).25 m/z (%) from LC/
MS: 237.20 (100, [M + H]+).
4.23. General Procedure for One-Step Oxidative Chlo-
rination of 1a-l and 15a,b. Reactions were run on 1-2 g
scale in 25 mL three-neck round-bottom flasks equipped with
temperature probe, condenser and chlorine line. The sulfides
to be oxidized were dissolved in acetic acid/water (9:1 v/v, 9
( 2 vol equiv), with chlorine gas (∼3 mol equiv) bubbled
through with stirring. Temperature profiles were collected
electronically. Chlorine was administered using the delivery
system described earlier in this paper. Samples were taken at 5
min intervals and analyzed by LC/MS. Reaction profiles were
constructed using LC/MS data.
confirm completion, and then chlorine was administered as
described above.
4.25. (S)-5-Methyl-5-trichloromethyl-imidazolidine-2,4-
dione (19). Following the oxidative chlorination of 1k, the
reaction was concentrated to dryness to afford an oily solid. A
small portion of this material was purified by column chroma-
tography (eluant 2:1, isohexane/ethyl acetate). This yielded the
title compound (41 mg) as a white solid.
1H NMR (400 MHz, d6-DMSO) δ 1.72 (3H, s), 8.95 (1H,
s), 11.26 (1H, s). 13C NMR (75 MHz, d6-DMSO) δ 19.32,
72.58, 101.45, 155.99, 171.02. HMBC and HSQC experiments
helped to confirm that the NMR data was consistent with the
proposed structure. Accurate mass, calc. for [M + H]+
C5H6Cl3N2O2 ) 230.9489, found ) 230.9494 (1.9 ppm error).
Acknowledgment
Craig Baldwin, Larry Coe, Lee Griffiths, Zakariya Patel,
Helen Porter, Carl Reens, Simon Watkins.
4.24. General Procedure for Two-Step Oxidation-Chlori-
nation of 1a-l and 15a,b. Solutions of the sulfide substrates
were prepared as above. These solutions were treated with 35%
hydrogen peroxide (1.08 mol equiv) at ambient temperature and
stirred for ∼2 h. The reactions were analysed by LC/MS to
Supporting Information Available
This material is available free of charge Via the Internet at
http://pubs.acs.org.
(23) Oh, C. H. Bull. Korean Chem. Soc. 1988, 9 (4), 231.
(24) Method adapted from that described by Xu, H.-J.; Zhao, X.-Y.; Deng,
J.; Fu, Y.; Feng, Y.-S. Tetrahedron Lett. 2009, 50 (4), 434.
(25) Furuta, T. Nipon Kagaku Zasshi 1969, 90 (9), 936.
Received for review September 15, 2009.
OP900241E
288
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development