A practical synthesis of a [2.2.1] bicyclic chiral sulfide for asymmetric transformations

Article abstract of DOI:10.1016/j.tet.2006.06.044

A substantially improved synthesis of synthetically useful chiral sulfide 1 is described. Starting from (+)-10-camphorsulfonic acid, the chiral sulfide was synthesised on large scale in five steps and 56% overall yield. Significant improvements include the use of Bu₃P in place of Ph₃P for the reduction of the chlorosulfonyl group to the thiol, allowing removal of phosphine oxide by aqueous extraction and improvements in the photochemistry using either a modified batch reactor or a single pass continuous flow reactor.

Full text of DOI:10.1016/j.tet.2006.06.044

Tetrahedron 62 (2006) 11297–11303
A practical synthesis of a [2.2.1] bicyclic chiral sulfide
for asymmetric transformations
*
Varinder K. Aggarwal, Guangyu Fang, Christoforos G. Kokotos,
Jeffery Richardson and Matthew G. Unthank
Department of Chemistry, University of Bristol, Cantock’s Close, Bristol, United Kingdom
Received 4 April 2006; revised 13 June 2006; accepted 14 June 2006
Available online 10 July 2006
This paper is dedicated to Professor David MacMillan, the rising star of our generation
Abstract—A substantially improved synthesis of synthetically useful chiral sulfide 1 is described. Starting from (+)-10-camphorsulfonic
acid, the chiral sulfide was synthesised on large scale in five steps and 56% overall yield. Significant improvements include the use of
Bu3P in place of Ph3P for the reduction of the chlorosulfonyl group to the thiol, allowing removal of phosphine oxide by aqueous extraction
and improvements in the photochemistry using either a modified batch reactor or a single pass continuous flow reactor.
Ó 2006 Elsevier Ltd. All rights reserved.
1. Introduction
by sub-stoichiometric quantities of a chiral sulfide (5 mol %)
and rhodium acetate (0.5 mol %).3 The process shows good
substrate scope but with many catalytic process limitations.
As a result of these limitations, we have developed a stoichio-
metric epoxidation protocol, which is highly efficient and
enantioselective and shows very broad substrate scope.4
Sulfur ylides are extremely useful carbon nucleophiles in
organic synthesis.1 They are mostly used for two types of
reactions: three-membered ring formation (epoxidation,
cyclopropanation or aziridination) and rearrangement reac-
tions.2 In the first class of reactions the sulfur ylide reacts
with an electrophilic carbon atom (of a carbonyl group, im-
ine or an activated carbon–carbon double bond) followed by
elimination of the sulfide to give a strained three-membered
ring. We have developed a new catalytic process in which
tosyl hydrazone salts couple with carbonyl compounds to
give epoxides with control over relative and absolute stereo-
chemistry in high yield (Scheme 1). The reaction is mediated
The catalytic process has been extended to the synthesis of
cyclopropanes and aziridines (Scheme 2).5 The practicality
of these processes has been demonstrated in the synthesis of
a number of biologically important targets (CDP 840,5 pre-
lactone B,6 taxol side chain,7 and the anti-hypertensive agent
(+)-LY3547408) involving each class of three-membered
ring compounds with almost complete control of chirality.
S
R
Na
Rh2(OAc)4
[Rh]=CHR
PTC
R1CHO
RCHN2
N
40 °C
R
N
+
Ts
one pot
O
R
H
N
RCHO
N2
R1
H2N
Ts
S
58 - 82% yield
88:12 - >98:2 d.r.
87 - 94% ee
H
(5 mol%)
1
O
Scheme 1. Catalytic cycle for epoxide formation with in situ generation of diazocompounds.
* Corresponding author. Tel.: +44 117 954 6315; fax: +44 117 929 8611; e-mail: v.aggarwal@bristol.ac.uk
0040–4020/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2006.06.044

11298
V. K. Aggarwal et al. / Tetrahedron 62 (2006) 11297–11303
S
S
( 20 mol %)
Na
O
O
( 5-20 mol %)
R3
N
N
H
R
N
Ts
H
R
R2
R1
R
R3
+
R1
R3
R2
H
R3
R2
Rh2(OAc)4 (1 mol %)
PTC (20 mol %)
N
R2
88-92% ee
R1
84-98% ee
R1
1,4-dioxane, 40 ºC
Scheme 2. Catalytic asymmetric synthesis of cyclopropanes and aziridines with in situ generation of diazocompounds.
New and useful reactions of chiral sulfur ylides with organo-
boranes have also been developed and high levels of enantio-
a)
b)
c)
selectivity have been achieved (Scheme 3).9
O
O
O
O
SO3H
7
SO2Cl
2
SH
3
R
R
B
Ar
H
H
H
OH
R
R
+
H2O2
S
S
Ar
+ BR3
S
Ar
>95% ee
d)
Ph
H
BR2
R
+
Ar
O
O
O
H
NH2
O
S
S
4
NH2OSO3H
Ar
>95% ee
O
R
6
5
S
e)
Scheme 3. Asymmetric synthesis of chiral organoboranes using sulfur
ylides.
O
Having developed these useful asymmetric processes, a fac-
ile large-scale synthesis of sulfide 1 was sought. Although 1
had been previously prepared in four high yielding steps
from camphorsulfonyl chloride (Scheme 4), this process
could only be used to prepare 1 on a 5–10 g scale.10
1
Reagents and conditions:
a) PCl5 (1.4 equiv.), 4 h, 0 °C
b) PBu3 (3 equiv.), 1,4-dioxane/H2O (4:1), reflux, 15 h, 98%.
c) PhCOCH2Cl (1.0 equiv.), NaHCO3 (2 equiv.), DMF, r.t., 5 h,100%.
d) cyclopentadiene (40 equiv.), CH2Cl2, –20 °C, continuous flow reactor.
e) H2, 3% Pd/C, EtOH, r.t., 18 h, 59% (two steps).
r.t., 96%.
Scheme 5. Improved synthesis of chiral sulfide 1.
a)
b)
O
O
2. Results and discussion
O
Ph
SH
3
S
SO2Cl
2
4
O
Initially, the synthesis of the sulfide is started from the
commercially available (+)-10-camphorsulfonyl chloride.
However, this reagent is rather expensive (£196/100 g,
Aldrich) and often contained impurities including (+)-10-
camphorsulfonic acid 7, which had to be removed. We there-
fore considered starting from (+)-10-camphorsulfonic acid
(£33.70/100 g, Aldrich). Following the method reported by
Vandewalle and co-workers,11 (+)-10-camphorsulfonic
acid 7 was efficiently converted to the corresponding sulfo-
nyl chloride 2 by phosphorous pentachloride at 0 ꢀC under
solvent free conditions (Scheme 5). Recrystallisation of
the chloride was necessary before continuing to the next
step as the impure material gave significantly reduced yields
in the next step (70% cf. 90% with pure material).
S
S
d)
c)
O
O
O
S
6
1
5
Reagents and conditions:
a) PPh3 (4 equiv.), 1,4-dioxane/H2O (4:1), reflux, 1 h, 82%.
b) PhCOCH2Cl (1.1 equiv.), K2CO3 (5 equiv.), THF, reflux, 20 h, 82%.
c) cyclopentadiene (20 equiv.), CH2Cl2, 20 °C, 16 h, 70%,
20:1 diastereoselectivity.
d) H2, 1.7% PdS/C, EtOH, r.t., 83%.
Scheme 4. Original protocol for the synthesis of chiral sulfide 1.10
This is a consequence of limitations that arose in two of the
key steps: the reduction of camphorsulfonyl chloride, which
was limited by the large quantities of triphenylphosphine
oxide by-product that must be removed and the photolysis-
Diels–Alder reaction, which was limited by the power output
of the lamp used. Each step, and in particular these two,
was therefore studied in order to find the optimal conditions,
with specific attention directed towards finding reaction
conditions that could be performed easily on a large scale.
The final optimised route is shown in Scheme 5 and is
discussed in detail below.
The next step of the synthesis was the formation of thiol 3
by reduction of camphorsulfonyl chloride with 4 equiv of
triphenylphosphine according to the literature protocol.10,12
While this reaction worked well giving the thiol 3 in good
yield, an unappealing aspect of this procedure was the
removal of a large quantity of triphenylphosphine oxide,
which was often non-trivial. Therefore, an alternative phos-
phine was sought that resulted in a water-soluble phosphine
oxide thus making it easier to remove. Both trimethyl-
phosphine and tri-n-butylphosphine met this requirement
although tri-n-butylphosphine was chosen, not least because

V. K. Aggarwal et al. / Tetrahedron 62 (2006) 11297–11303
11299
trimethylphosphine is a sternutator. Interestingly, the water
coolant in
coolant out
solubility of tri-n-butylphosphine oxide has been shown
to be greater at low temperature and so extraction was
performed using ice cold water.13 Substituting tri-n-butyl-
phosphine for triphenylphosphine under identical conditions
(note: the solution was degassed before addition of phos-
phine) gave the thiol in 85% yield after ice-water extraction
and column chromatography. Since only 3 equiv of the phos-
phine are formally required, we tested the reduction under
the reduced loading and this reaction proved very efficient
(98%). It could be performed on large scale and unlike the
use of 4 equiv, required no chromatographic purification.
In the alkylation of thiol 3 with a-chloroacetophenone, it
was found that the generation of undesired by-products
could be minimised by using DMF in place of the previously
employed THF. The use of DMF allowed the reaction to pro-
ceed smoothly at an ambient temperature over a shorter time
scale and ultimately led to an overall cleaner reaction. The
use of a weaker base in this alkylation step also allowed
the reaction to proceed more smoothly, so the original base
K2CO3 was replaced by NaHCO3. The efficient alkylation
of the thiol was thus achieved using 1 equiv of a-chloro-
acetophenone, followed by a simple purifying work-up
procedure and removal of solvent in vacuo gave material
of sufficient purity for the next transformation (Scheme 5).
reactants
lamp
outside
coolant
stirring bar
Figure 1. The batch photochemical reactor.
of the relatively weak output was able to excite the molecule
into the triplet state. Therefore, a lamp with a higher output
in the desired wavelength range was sought. The use of
a 125 W mercury arc lamp in a Vycor immersion well pro-
vided a suitable alternative (no copper sulfate bath used
due to lamp set up—see Fig. 1). Under these conditions
the reaction was significantly faster and the product was
obtained in similar yield and diastereoselectivity to the orig-
inal conditions. This allowed 18.2 g of 4 to be processed in
a single batch carried out in neat cyclopentadiene (20 equiv).
However, the mercury lamp generated a large amount of heat
and efficient cooling of the reaction vessel was required to
achieve good stereocontrol (Fig. 1). This was carried out
by using a circulating coolant on the inner wall and cryo-
static cooling to the outer wall, with both set at ꢁ10 ꢀC.
2.1. Optimisation of batch photochemical reactor
process
While there are a variety of methods available for the gener-
ation of thioaldehydes,14 the photochemical generation of
thioaldehydes from phenacyl sulfides, as described by
Vedejs and co-workers, is the key step in this synthetic
sequence. This method involves the irradiation of a solution
of phenacyl sulfide using a sun lamp, in the presence of an
excess of cyclopentadiene (Scheme 4).14a A copper sulfate
bath was used to filter out light of wavelengths below
320 nm, which are believed to cause undesirable secondary
photochemical reactions, and the reaction temperature was
maintained at 20 ꢀC with the aid of a crystostat. The reaction
needs to be degassed since oxygen can sequester the radical
intermediates14a and this was achieved by simply repeating
the procedure of putting the cooled solution (ꢁ78 ꢀC) under
vacuum (about 15 mbar) followed by charging the vessel
with nitrogen twice.
1
The reaction was monitored by H NMR and it was found
that at this temperature good diastereoselectivity (w20:1)
was observed. Without internal cooling, the reaction temper-
ature is significantly higher and lower diastereoselectivity
(6:1) was obtained, resulting in more difficult downstream
problems for separating the diastereomers.
2.2. A single pass, continuous flow photochemical
reactor
The use of the batch reactor for this photochemical reaction
had certain limitations associated with it that were difficult
to overcome and made further scale up inefficient. Most
notably, the reaction mixture in the batch reactor is unevenly
irradiated. This problem is emphasised when using large
volumes of reactant solution, as diffusion in the narrow
well is inefficient even with rapid stirring, resulting in
reduced yields and increased by-product formation. This
problem was overcome by the use of a single pass, continu-
ous flow, Vycor reactor as described by Booker-Milburn and
co-workers15 (Fig. 2).
Perhaps the biggest obstacle for the scale up of this reaction
is the output of the lamp, since this does not change when the
amount of material that is to be photolysed is increased.
Ordinarily this would simply require increased reaction
times to allow sufficient light to enter the system and thus
complete the reaction. However, cyclopentadiene slowly
dimerises over time and so prolonged reaction times are
not acceptable. Therefore the efficiency of the lamp was
examined. It was found that the phenacyl sulfide 4 had two
peaks in its UV spectrum (ca. 280 nm and 300 nm). The ab-
sorption at ca. 280 nm was attributed to the p–p* transition
of the carbonyl group on the camphor moiety and the absorp-
tion at ca. 300 nm was assigned to the p–p* transition of the
carbonyl that was to be excited. However the spectral output
of the lamp (Osram Ultra-Vitalux 300 W sun lamp) was
discovered to be very broad and quite weak and therefore
not well suited to this reaction, as only a small percentage
This consists of three layers of UV transparent tubing (fluo-
rinated ethenepropylene (FEP), 2.7 mm i.d.ꢂ3.1 mm o.d.)
wound around a traditional cooled immersion well contain-
ing a 400 W mercury lamp (Fig. 2b). The reaction solution
could then be driven around the reactor through the use of
a common HPLC pump (Fig. 2c) and the irradiation time
could be defined by the flow rate. This system would then
allow large volumes of reaction solution to be processed
continuously. The cooling of the reaction solution was effi-
ciently achieved by passing cooled ethylene glycol/water

11300
V. K. Aggarwal et al. / Tetrahedron 62 (2006) 11297–11303
Figure 2. (a) AVycor/FEP continuous flow chemical reactor. (b) Close-up of FEP tubing (2.7 mm i.d.ꢂ3.1 mm o.d.). (c) Attached to water supply and HPLC
pump.
Figure 3. The single pass, continuous flow photochemical reactor.
(1:1) solution at temperature of ꢁ5 to ꢁ20 ꢀC through the
immersion well. In addition to this, the submersion of the
entire reactor in an immersion bath of cooled ethylene
glycol/water (1:1) solution ensured that the reaction tem-
perature was well maintained. A schematic representation
of the continuous flow photochemical reaction is shown in
Figure 3.
reaction yield and the diastereoselectivity.16 It was discov-
ered that reducing the sulfide concentration from 0.4 to
0.2 M increased the yield from 43 to 59% (Table 1, entries
1 vs 2). Increasing the equivalents of cyclopentadiene was
also shown to have a positive effect on both the yield and
diastereoselectivity (Table 1, entries 2 vs 3). Reducing the
temperature of the immersion bath from ꢁ5 to ꢁ20 ꢀC
was shown to have a positive effect on the yield, though no
change in diastereoselectivity was observed (Table 1, entries
3 vs 4). Finally, a further increase in the amount of cyclo-
pentadiene used (40 equiv) was shown to increase the yield
The optimisation of this process was achieved by exploring
how variations in flow rate, acylsulfide concentration, equiv-
alents of cyclopentadiene and temperature affected both the
Table 1. Optimisation of the continuous flow photolysis of sulfide 4
S
h
O
CH2Cl2
O
Ph
S
O
S
O
4
5
6
Entry
Flow rate (ml/min)
Sulfide concn (M)
Cyclopentadiene (equiv)
Internal temp (ꢀC)
External temp (ꢀC)
Yield (%)a
d.r.b
1
2
3
4
5
6
4
4
4
0.4
0.2
0.2
0.2
0.2
0.12
10
10
20
20
40
40
ꢁ20
ꢁ20
ꢁ20
ꢁ20
ꢁ20
ꢁ20
ꢁ5
ꢁ5
43c
59
67
72
82
75
9:1
10:1
11:1
11:1
11:1
10:1
ꢁ5
4
ꢁ20
ꢁ20
ꢁ20
2d
2
Yield is calculated by 1H NMR.
a
Diastereoselectivity is calculated by HPLC using a DiscoveryÒ HS C18 568523-U column 25 cmꢂ4.6 mm, 5 mm eluting with a flow rate of 1 mL/min with
a gradient of water acetonitrile: 0–25 min from 5 to 95% acetonitrile, 25–27 min 95% acetonitrile, 27–37 min from 95 to 5% acetonitrile, 37–42 min 5%
acetonitrile (trminor: 24.34 min, trmajor: 24.82 min).
b
c
Starting material (25%) was recovered.
A reduced flow rate of 2 mL/min was required as the high viscosity of the solution would not allow the pump to achieve a greater flow rate.
d

V. K. Aggarwal et al. / Tetrahedron 62 (2006) 11297–11303
11301
further (Table 1, entries 4 vs 5). A reduced flow rate of 2 mL/
min was required when using these conditions due to the
high viscosity of the reaction solution, which would not
allow the pump system to achieve a greater flow rate. It
was considered that on a large scale the problem of viscosity
could be circumvented by reducing the concentration of
the sulfide in combination with the reduced flow rate, thus
giving a more practical set of optimised conditions (Table 1,
entry 6). The final large scale (38 g of substrate) optimised
conditions for the photolysis of sulfide 4 are shown in Table
1, entry 6 and gave a reaction yield of 75% with a 10:1 d.r.
under reduced pressure to afford an off-white crystalline
solid. This solid was then recrystallised from petrol to
afford the product as white crystalline flakes (104.1 g,
96%). Mp 66–67 ꢀC (hexane) [lit.17 67–68 ꢀC (heptane)];
Rf 0.42 (20% EtOAc/petrol); 1H NMR (400 MHz,
CDCl3) d 0.93 (3H, s), 1.15 (3H, s), 1.49 (1H, ddd,
J¼13.0, 9.5 and 4.0 Hz), 1.78 (1H, ddd, J¼14.0, 9.5,
and 5.0 Hz), 1.99 (1H, d, J¼18.5 Hz), 2.05–2.15 (1H,
m), 2.17 (1H, t(app.), J¼4.5 Hz), 2.40–2.52 (2H, m), 3.73
(1H, d, J¼14.5 Hz), 4.31 (1H, d, J¼14.5 Hz).
4.2. 7,7-Dimethyl-1-(sulfanylmethyl)bicyclo[2.2.1]-
heptan-2-one (3)10
Throughout this study it proved very difficult to separate the
acetophenone by-product from the cycloadduct following
the photolysis process, so the crude material was filtered
through a short silica plug, and then taken onto the final
hydrogenation step. Although initially the pre-poisoned
catalyst PdS/C was used, it was found that the reactive
double bond could be just as easily reduced using (10%)
Pd/C (3 mol % Pd used). Following hydrogenation, the
solvent and acetophenone were removed under reduced
pressure (30–40 ꢀC, 0.2 mbar), then the residue was passed
through a short silica gel pad and the partially purified prod-
uct recrystallised from pentane, to give the sulfide 1 as
a white crystalline solid of >95% d.r.
4.2.1. Method A: reduction using triphenylphosphine.10
(+)-(10)-Camphorsulfonyl chloride 2 (14.0 g, 0.06 mmol)
and triphenylphosphine (63.3 g, 0.24 mol) were refluxed
in a mixture of H2O (50 mL) and 1,4-dioxane (200 mL)
for 1 h. The reaction mixture was allowed to cool to rt
and was extracted with petrol (200 mL and 4ꢂ100 mL).
The combined organic extracts were washed with H2O
(2ꢂ100 mL), saturated aqueous NaCl (100 mL) and dried
over MgSO4. The solvent was removed under reduced pres-
sure to give a yellow oil, which was purified by flash chroma-
tography (3% EtOAc in petrol) to give thiol 3 as a white
crystalline solid (9.1 g, 82%). Mp 63–64 ꢀC (EtOAc/petrol)
[lit.18 65–66 ꢀC]; Rf 0.22 (5% EtOAc/petrol); 1H NMR
(400 MHz, CDCl3) d 0.91 (3H, s), 1.02 (3H, s), 1.36–1.43
(1H, m), 1.67–1.75 (1H, m), 1.88–2.04 (4H, m), 2.09 (1H,
t(app.), J¼5.0 Hz), 2.29–2.42 (2H, m), 2.87 (1H, dd, J¼13.5
and 7.0 Hz).
3. Conclusions
We have developed a practical five-step synthesis of sulfide 1
in 56% overall yield from camphorsulfonic acid. The acid is
extremely cheap and readily available in both enantiomeric
forms. This practical synthesis of the chiral sulfide not only
allows the organocatalytic reaction, which utilises the sulfide
to be run on a large scale but also allows the stoichiometric
reactions to be run on moderate scale.
4.2.2. Method B: reduction using tri-n-butylphosphine. A
solution of (+)-(10)-camphorsulfonyl chloride 2 (50.0 g,
0.2 mol) in 1,4-dioxane (380 mL) was degassed by bubbling
N2 through it for 10 min. Meanwhile, a mixture of H2O
(190 mL) and 1,4-dioxane (380 mL) in a 2 L 3-necked
round-bottomed flask fitted with two heat resistant septa,
a reflux condenser and a large magnetic stirrer bar was
degassed in the same way. After 10 min tri-n-butylphosphine
(150 mL, 0.6 mol) was added to the H2O/dioxane mixture
via syringe, giving a biphasic solution. This was followed
rapidly by the addition of the degassed camphorsulfonyl
chloride solution via cannula, which resulted in a monopha-
sic solution. The reaction mixture was then stirred and
heated to reflux for 15 h. The reaction mixture was then al-
lowed to cool to rt and transferred to a 5 L separating funnel,
where it was partitioned between petrol (1 L) and ice cold
H2O (1 L). The aqueous layer was removed and extracted
with petrol (3ꢂ500 mL), the total volume of these organic
extracts was then reduced to 300 mL and the colourless
solution was then washed with ice cold H2O (3ꢂ500 mL)
and the solvent evaporated to give a white crystalline solid.
The initial petrol layer was washed with ice cold H2O
(3ꢂ500 mL) and then evaporated to give a yellow crystalline
mass. This was then washed with petrol to give a white crys-
talline solid. The organic washings were then evaporated
to dryness and filtered through a short silica plug (eluent
5% EtOAc in petrol) collecting all fractions that stained
a KMnO4 plate. After evaporation of the appropriate frac-
tions, the white crystalline solid obtained was combined
with the other solids obtained to give 3 as white crystalline
solid (36.0 g, 98%).
4. Experimental
4.1. (D)-10-Camphorsulfonyl chloride (2)11
Based on the procedure of Vandewalle and co-workers,11
(+)-10-Camphorsulfonic acid 7 (100.0 g, 0.43 mol) was
placed in a 1 L round-bottomed flask with a gas outlet
tube leading to a saturated NaHCO3 trap (as HCl is pro-
duced in the reaction). The flask was cooled in a large
ice bath and PCl5 solid (124.0 g, 0.60 mol) was added
portion wise. The mixture was agitated manually using
a glass rod to initiate the liquefication of the mixture,
before being stirred magnetically. After 1 h the cooling
bath was removed and the reaction mixture was allowed
to warm to rt and stirred for 3 h. The reaction mixture
was then slowly poured portion wise into a 2 L separating
funnel that was approximately 50% filled with crushed ice.
After addition of each portion the separating funnel was
thoroughly shaken before further material was added.
Once addition of the reaction mixture to the separating
funnel was complete, CH2Cl2 (500 mL) was added, and
the separating funnel was shaken. CH2Cl2 layer was re-
moved and the aqueous material was re-extracted with
CH2Cl2 (2ꢂ200 mL). The combined organic fractions
were then dried with MgSO4 and the solvent was removed

11302
V. K. Aggarwal et al. / Tetrahedron 62 (2006) 11297–11303
4.3. (D)-7,7-Dimethyl-1-{[(2-oxo-2-phenylethyl)-
sulfanyl]methyl}bicyclo[2.2.1]heptan-2-one (4)10
monitoring by TLC (Rf¼0.45 [product], 0.43 [acetophe-
none], 5% ethyl acetate/petrol) to ensure all of the product
is collected. The solvent was then removed under reduced
pressure to give the partially purified product as a pale
yellow oil (65% yield by 1H NMR, using the stoichiometri-
cally produced acetophenone as an internal standard, and
20:1 d.r.), which was dissolved in EtOH (w40 mL) and
10% Pd/C (3.0% based on the Diels–Alder product) was
added. The mixture was stirred under a hydrogen atmo-
Using an improved procedure of Vedejs and co-workers,19
into a solution of 3 (30.0 g, 0.163 mol) and a-chloroaceto-
phenone (1.0 equiv, 25.2 g, 0.163 mol) in DMF (100 mL)
was added NaHCO3 (2 equiv, 27.4 g, 0.326 mol). After 5 h
1
the reaction was monitored by either HPLC or H NMR
to look for complete consumption of the a-chloroaceto-
phenone. If all was consumed the reaction mixture was
then diluted with diethyl ether, filtered and washed with
diethyl ether. The solution was washed with water
(3ꢂ100 mL) and dried over MgSO4. The solvent was
removed under reduced pressure to afford 4 as a pale yellow
oil (49.3 g, 100%) that was of sufficient purity to use in the
subsequent reaction. If any a-chloroacetophenone remained
(lachrymator!), then commercial aqueous ammonia solution
(35%, 40 mL) was added and the reaction mixture stirred for
1 h, then cooled in an ice bath and carefully acidified by the
portion wise addition of aqueous HCl (1 M). The solution
was then diluted with diethyl ether (500 mL) and extracted
with aqueous HCl (3ꢂ100 mL), water (1ꢂ100 mL) and
dried over MgSO4. Diethyl ether was removed and the
resulting dark orange oil dissolved in ethanol (100 mL)
and treated with activated charcoal (5.0 g), then heated to
50 ꢀC for 1 h, allowed to cool, then filtered through Celite
and the solvent removed under reduced pressure to afford
4 as a pale yellow oil (49.3 g, 100%) that was of sufficient
purity to use in the subsequent reaction. [a]2D1 +6.5 (c 20,
CH2Cl2); Rf 0.37 (15% EtOAc in petrol); 1H NMR
(400 MHz, CDCl3) d 0.83 (3H, s), 1.03 (3H, s), 1.28–1.41
(1H, m), 1.52 (1H, dd, J¼12.5 and 3.0 Hz), 1.85 (1H, d,
J¼18.0 Hz), 1.90–2.01 (2H, m), 2.06 (1H, t(app.), J¼
4.0 Hz), 2.34 (1H, ddd, J¼18.0, 4.7 and 2.3 Hz), 2.57 (1H,
d, J¼13.1 Hz), 2.87 (1H, d, J¼13.1 Hz), 3.87 (2H, s), 7.47
(2H, t, J¼8.0 Hz), 7.57 (1H, t, J¼8.0 Hz), 7.98 (2H, d,
J¼7.0 Hz); 13C NMR (100 MHz, CDCl3) d 19.9, 20.2,
26.8, 26.9, 29.4, 39.6, 43.0, 43.5, 47.8, 61.1, 128.6, 128.7,
133.3, 135.4, 194.6, 217.2.
1
sphere (7.0 bar) for 18 h (monitored by H NMR) and the
Pd/C was removed by filtration through a CeliteÒ pad. The
solvent and acetophenone (the by-product of the photolysis)
were removed under vacuum (30–40 ꢀC, 0.2 mbar), then the
residue was passed through a short silica gel pad eluting with
2% ethyl acetate/petrol, to remove the coloured impurities.
The crude product was then recrystallised from pentane to
afford the sulfide 1 as a white solid (8.21 g, 54%, >95%
d.r.). Mp 60–62 ꢀC (MeOH) [lit.10 60–62 ꢀC (MeOH)];
[a]2D5 +35.9 (c 19.5, CH2Cl2) [lit.10 [a]2D5 +39.3 (c 1.0,
MeOH)]; Rf 0.36 (4% EtOAc in petrol); IR (CDCl3) 3025,
2950, 1760, 1230 cmꢁ1 1H NMR (400 MHz, CDCl3)
;
d 0.89 (3H, s), 0.96 (3H, s), 1.32–2.08 (11H, m), 2.38 (1H,
ddd, J¼18.2, 4.8 and 3.1 Hz), 2.52–2.66 (1H, m), 3.24–3.32
(1H, br), 3.60 (1H, m), 3.64–3.71 (1H, m); 13C NMR
(63 MHz, CDCl3) d 20.6, 21.4, 23.5, 24.4, 27.6, 35.2, 41.3,
42.4, 44.0, 44.7, 45.0, 49.3, 49.9, 62.6, 217.6; MS (EI): m/z
(%): 250 (85) [M+], 209 (52), 194 (49) 181 (100); HRMS
(EI) (m/z) calculated for C15H22SO 250.1391, found 250.1401.
4.5. General procedure for large-scale photolysis using
a continuous flow reactor
To a solution of ketosulfide 4 (38 g, 0.126 mol) in CH2Cl2
(630 mL) at ꢁ20 ꢀC was added freshly distilled cyclopenta-
diene (420 mL, 5.04 mol) and the reaction mixture degassed
for 10 min by the passing of N2 through the solution. The
continuous flow reactor (Fig. 3) was set with the internal
and the external cooling systems at ꢁ20 ꢀC (ethylene
glycol/water 1:1). The 400 W mercury lamp was switched
on and degassed CH2Cl2 (20 mL) was pumped through the
Vycor, three-layer continuous flow reactor at a flow rate of
2 mL/min. The reaction mixture was then pumped through
the reactor at a flow rate of 2 mL/min, with collection of
the product in a large conical flask. Once the whole solution
had passed through the reactor, the tubing was washed with
CH2Cl2 (350 mL) using the same flow rate. The reaction
solution and tube wash were then combined and transferred
to a round-bottomed flask and the solvent and cyclopenta-
diene were removed under reduced pressure (20 mbar).
The sulfide was then isolated by passing the crude reaction
mixture through a short silica plug (250 g of silica gel), elut-
ing first with petrol to remove the cyclopentadiene dimers
until the yellow front line reaches the bottom of the column.
The petrol fractions were then discarded and the product was
eluted from the silica plug using 2.5% ethyl acetate/petrol,
monitoring by TLC (Rf ¼ 0.45 [product], 0.43 [acetophe-
none], 5% ethyl acetate/petrol) to ensure all of the product
is collected. The solvent was then removed under reduced
pressure to give the partially purified product as a pale
yellow oil (75% yield by 1H NMR, using the stoichiometri-
cally produced acetophenone as an internal standard, and
10:1 d.r. by HPLC), which was dissolved in EtOH
(w80 mL) and 10% Pd/C (2.6 g, 2.0 mol % based on sulfide
4.4. General procedure for large-scale photolysis using
a batch reactor
The solution of phenacyl sulfide 4 (18.2 g, 60 mmol) and
freshly distilled cyclopentadiene (80.0 g, 20 equiv) in a
125 mL Vycor immersion well was cooled in dry ice-
acetone. The cooled solution was degassed by repeating
the procedure of putting the solution under vacuum (about
15 mbar) followed by charging the vessel with nitrogen
twice. The degassed solution was photolysed at ꢁ10 ꢀC
with a 125 W medium pressure mercury lamp for 9 h (mon-
1
itored by H NMR) with simultaneous cooling of both the
inside and outside walls of the reaction vessel (Fig. 1). The
reaction mixture was then transferred to a round-bottomed
flask and the excess cyclopentadiene (51.1 g) was collected
under reduced pressure (20 mbar); this could then be
recycled if desired. The sulfide was then isolated by passing
the crude reaction mixture through a short silica plug (150 g
of silica gel), eluting first with petrol to remove the cyclo-
pentadiene dimers until the yellow front line reaches the
bottom of the column. The petrol fractions were then dis-
carded and the product and acetophenone were eluted
from the silica plug using 2.5% ethyl acetate/petrol,

V. K. Aggarwal et al. / Tetrahedron 62 (2006) 11297–11303
11303
5. Aggarwal, V. K.; Alonso, E.; Fang, G.; Ferrara, M.; Hynd, G.;
Porcelloni, M. Angew. Chem., Int. Ed. 2001, 40, 1433–1436.
6. Aggarwal, V. K.; Bae, I.; Lee, H. Y. Tetrahedron 2004, 60,
9725–9733.
7. Aggarwal, V. K.; Vasse, J.-L. Org. Lett. 2003, 5, 3987–3990.
8. Aggarwal, V. K.; Grange, E. Chem.—Eur. J. 2006, 12, 568–575.
9. Aggarwal, V. K.; Fang, G.; Schmidt, A. T. J. Am. Chem. Soc.
2005, 127, 1642–1643.
10. (a) Aggarwal, V. K.; Alonso, E.; Hynd, G.; Lydon, K. M.;
Palmer, M. J.; Porcelloni, M.; Studley, J. R. Angew. Chem.,
Int. Ed. 2001, 40, 1430–1433; (b) Aggarwal, V. K.;
Charmant, J.; Dudin, L.; Porcelloni, M.; Richardson, J. Proc.
Natl. Acad. Sci. U.S.A. 2004, 101, 5467–5471.
4 and 3.0% based on the Diels–Alder product) was added.
The mixture was stirred under a hydrogen atmosphere
(7.0 bar) for 18 h (monitored by H NMR) and the Pd/C
1
was removed by filtration through a CeliteÒ pad. The solvent
and acetophenone (the by-product of the photolysis) were
removed under vacuum (30–40 ꢀC, 0.2 mbar), then the resi-
due was passed through a short silica gel pad eluting with 2%
ethyl acetate/petrol to remove the coloured impurities. The
crude product was then recrystallised from pentane to afford
the sulfide 1 as white solid (18.6 g, 59%, >95% d.r.).
Acknowledgements
11. Vandewalle, M.; Van der Eycken, J.; Oppolzer, W.; Vullioud, C.
Tetrahedron 1986, 42, 4035–4043.
12. Ko, K.-Y.; Kim, K.-I. Bull. Korean Chem. Soc. 1998, 19, 378–
379.
13. Higgins, C. E.; Baldwin, W. H. Anal. Chem. 1960, 32,
233–236.
We thank EPSRC, GlaxoSmithKline (Nigel Hussain), NPIL-
pharma (Robin Fieldhouse), Johnson Matthey (Mark
Hooper), Syngenta (David Ritchie), and the DTI for finan-
cial support. We thank Professor Kevin Booker-Milburn
for advice and access to his photochemical apparatus.
14. (a) Vedejs, E.; Eberlein, T. H.; Mazur, D. J.; McClure, C. K.;
Perry, D. A.; Ruggeri, R.; Schwartz, E.; Stults, J. S.; Varie,
D. L.; Wilde, R. G.; Wittenberger, S. J. Org. Chem. 1986, 51,
1556–1562; (b) Bladon, C. M.; Ferguson, I. E. G.; Kirby,
G. W.; Lochead, A. W.; McDougall, D. C. J. Chem. Soc.,
Chem. Commun. 1983, 423–425; (c) Baldwin, J. E.; Lopez,
R. C. G. Tetrahedron 1983, 39, 1487–1498; (d) Krafft, G. A.;
Meinke, P. T. Tetrahedron Lett. 1985, 26, 1947–1950.
15. Hook, B. D. A.; Dohle, W.; Hirst, P. R.; Pickworth, M.; Berry,
M. B.; Booker-Milburn, K. I. J. Org. Chem. 2005, 70, 7558–
7564.
References and notes
1. (a) Claus, P. K. Organosulfur Compounds; Klamann, D., Ed.;
Georg Thieme: Stuttgart, 1985; Vol. E11, p 1344; (b)
Aggarwal, V. K.; Richardson, J. Science of Synthesis; Padwa,
A., Bellus, D., Eds.; George Thieme: Stuttgart, 2004; Vol. 27,
p 21.
2. Aggarwal, V. K.; Badine, D. M.; Moorthie, V. A. Aziridines and
Epoxides in Asymmetric Synthesis; Yudin, A. K., Ed.; Wiley-
VCH: Weinheim, 2006; p 1.
16. This study was carried out on medium scale (w1.5 g of sulfide
per reaction) due to constraints on the availability of starting
material. Table 1, entries 1–5 represent medium scale trial
reactions. Table 1, entry 6 represent the conditions and results
for the 38 g, large scale reaction.
3. (a) Aggarwal, V. K. Comprehensive Asymmetric Catalysis;
Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.; 1999; Vol.
II, pp 679–693; (b) Aggarwal, V. K.; Winn, C. L. Acc. Chem.
Res. 2004, 37, 611–620; (c) Aggarwal, V. K.; Alonso, E.;
Bae, I.; Hynd, G.; Lydon, K. M.; Palmer, M. J.; Patel, M.;
Porcelloni, M.; Richardson, J.; Stenson, R. A.; Studley, J. R.;
Vasse, J.-L.; Winn, C. L. J. Am. Chem. Soc. 2003, 125,
10926–10940.
´
ꢀ
ꢀ
17. Camman˜o, O.; Erın, A.; Fernandez, F.; Gomez, G.; Uriarte, E.
Heterocycles 1988, 27, 2839.
18. Oae, S.; Togo, H. Bull. Chem. Soc. Jpn. 1983, 56, 3802.
19. Vedejs, E.; Stults, J. S.; Wilde, R. G. J. Am. Chem. Soc. 1988,
110, 5452–5460.
4. Aggarwal, V. K.; Bae, I.; Lee, H. Y.; Richardson, J.; Williams,
D. T. Angew. Chem., Int. Ed. 2003, 42, 3274–3278.