A clean, palladium-catalyzed oxidative esterification of aldehydes using benzyl chloride

Article abstract of DOI:10.1016/j.tetlet.2011.08.021

A highly efficient, mild, simple and clean procedure is presented for the one-pot oxidation of aromatic and aliphatic aldehydes to their corresponding ethyl esters using benzyl chloride as the oxidant under palladium-catalyzed conditions. The reaction is complete in just 30 min under microwave irradiation and inert conditions are not required to obtain good to excellent yields (65-93%) of isolated products.

Full text of DOI:10.1016/j.tetlet.2011.08.021

Tetrahedron Letters 52 (2011) 5319–5322
Contents lists available at SciVerse ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
A clean, palladium-catalyzed oxidative esterification of aldehydes using
benzyl chloride
⇑
Georgios A. Heropoulos, Carolina Villalonga-Barber
Institute of Organic and Pharmaceutical Chemistry, National Hellenic Research Foundation, Vas. Constantinou Avenue 48, 11635 Athens, Greece
a r t i c l e i n f o
a b s t r a c t
Article history:
A highly efficient, mild, simple and clean procedure is presented for the one-pot oxidation of aromatic
and aliphatic aldehydes to their corresponding ethyl esters using benzyl chloride as the oxidant under
palladium-catalyzed conditions. The reaction is complete in just 30 min under microwave irradiation
and inert conditions are not required to obtain good to excellent yields (65–93%) of isolated products.
Ó 2011 Elsevier Ltd. All rights reserved.
Received 21 June 2011
Revised 26 July 2011
Accepted 2 August 2011
Available online 8 August 2011
Keywords:
Aldehydes
Benzyl chloride
Oxidation
Esterification
Palladium
The ester functional group is not only widely abundant in nat-
ural products but is also required in many synthetic strategies. Es-
ters are generally prepared by nucleophilic addition of an alcohol
to activated carboxylic acid derivatives such as acid anhydrides
and chlorides.1 The direct conversion of aldehydes into esters is a
valuable alternative, and much effort has been devoted to achieve
this transformation.2 Many methods, however, have restrictive
requirements such as the use of stoichiometric transition metal
oxidants3 or the employment of co-oxidants in large excess,4 and
therefore greener methods have been sought and recently
reported.5
Here we report a simple and effective catalytic procedure to
oxidize aldehydes directly to the corresponding ethyl ester under
low palladium loading, and using benzyl chloride in only slight ex-
cess generating toluene as the only by-product. The procedure is
suitable for both electron-deficient and electron-rich aromatic
aldehydes as well as aliphatic aldehydes, making it a practical
alternative to the already described procedures.
Oxidative esterifications of aldehydes using transition metal
catalysis in general,4,5a,6 and palladium7 in particular, are not
new. Undoubtedly, the most environmentally friendly are those
using molecular oxygen as the oxidizing agent.6b,7a However, this
option could be problematic in cases where high oxygen pres-
sures are required in reactions running in flammable organic sol-
vents, and also for substrates containing oxygen-sensitive
functionalities. Examples of oxidation reactions using halogen-
based reagents as oxidants are less reported. Muzart8 obtained
mixtures of aldehydes, acids and chloroethyl esters when oxidiz-
ing primary alcohols using palladium chloride in refluxing 1,2-
dichloroethane, and similarly using carbon tetrachloride as both
solvent and co-oxidant, Tsuji9 obtained mainly esters from the
oxidation of alcohols. However, to the best of our knowledge,
benzyl halides have not been previously used in the oxidative
esterification of aldehydes. Instead, palladium-catalyzed transfor-
mations involving aryl halides are more frequently reported than
the analogous reactions with benzylic derivatives10 and exam-
ples of palladium-catalyzed oxidations of primary alcohols using
halobenzenes have been described.11 A procedure that takes
advantage of a related hydrodehalogenation is that of Asensio12
where an
a-bromo sulfoxide is used to oxidize alcohols. How-
ever, this method is not compatible with benzylic alcohols con-
taining electron-withdrawing groups and requires handling
bromine in the synthesis of bromomethyl phenyl sulfoxide.
Moreover, after the oxidation reaction, the sulfoxide by-product
must be separated from the crude reaction mixture. Our method,
on the other hand, relies on readily available reagents and a fac-
ile separation of the reduced by-product by simple evaporation
of the crude reaction mixture.
During a dendrimer synthesis we required the cleavage of an
aryl allyl ether and subsequent benzylation without isolation of
the intermediate phenolic compound.13 However, in an attempt
to do a tandem deallylation–Williamson-type etherification,14 4-
allyloxybenzaldehyde (1) was converted into a mixture of the
transetherified compound 2 and the product derived from an oxi-
dative esterification 3 (ratio 1:1) (Scheme 1).
⇑
Corresponding author. Tel.: +30 2107273874; fax: +30 2107273831.
E-mail address: carol@eie.gr (C. Villalonga-Barber).
0040-4039/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2011.08.021

5320
G. A. Heropoulos, C. Villalonga-Barber / Tetrahedron Letters 52 (2011) 5319–5322
O
O
O
Pd(PPh ) (5 mol%)
3 4
OEt
+
K CO (3 equiv.)
2
3
O
BnO
BnO
Δ, BnBr (1 equiv.)
EtOH (20 ml/mmol)
1
2
3
Scheme 1. Attempted tandem deallylation–Williamson-type etherification of 4-allyloxybenzaldehyde.
Intrigued by this outcome, we chose 4-nitrobenzaldehyde as a
but altered the course of the reaction and acetal 7a was obtained
as the major product.17 That this acetal was not an intermediate
of our reaction was supported by the observation that an at-
tempt to obtain the ester from the acetal under the initial reac-
tion conditions failed. The ratio of the ester to benzyl alcohol
could be increased by reducing the volume of ethanol6d (entry
9) and the reaction time could be reduced to 30 min by applying
microwave irradiation (entry 10). Benzyl alcohol formation was
totally suppressed when the volume of ethanol was further re-
duced to 5 equiv, although in this case acetal formation was pre-
dominant (entry 11).18
With optimized reaction conditions available, we proceeded to
examine the scope of the Pd-catalyzed oxidative esterification of
different aldehydes. These conditions led to total (Table 1, entry
10) or almost total conversion of only electron-poor aromatic alde-
hydes19 and saturated aldehydes.20 The conversion of electron-rich
aromatic aldehydes was significantly lower,21 possibly due to a
lowering of the reactivity of the carbonyl group due to resonance
effects of the electron-donating substituents. Notably, however,
their lower reactivity also precluded their reduction and the corre-
sponding benzyl alcohols were not generated.
We next examined conditions to improve the conversion of
electron-rich aromatic aldehydes and chose 4-methoxybenzalde-
hyde (4b) for this purpose.
Doubling the reaction time (Table 2, entry 1) did not improve
significantly either the conversion or yield. Changing the base to
sodium carbonate, cesium carbonate or triethylamine (entries 2–
4) resulted in no reaction and neither did the addition of sodium
iodide (entry 5) as a solid additive.22 Changing the oxidant to ben-
zyl iodide (entry 6) or 4-methoxybenzyl bromide (entry 7) left the
model substrate in order to establish the conditions required to
efficiently oxidize the aldehyde to the ester. An electron-poor aro-
matic aldehyde was chosen as substrate since these were reported
to be more reactive towards a palladium-catalyzed oxidative ester-
ification when a more complex oxidant was used.7c
Under the initial reaction conditions, [Pd(PPh3)4 (5 mol %),
K2CO3 (3 equiv), benzyl bromide (1 equiv) in refluxing ethanol]
4-nitrobenzaldehyde were fully converted into a mixture of ester
5a and benzyl alcohol 6a (Table 1, entry 1). This result resembled
a Cannizzaro-type reaction which involves the redox conversion
of aldehydes into their respective alcohols and carboxylic acids
or esters in the presence of a strong base.15 However, the observa-
tion that in our reaction the ester was produced in a yield above
50% suggested that another type of mechanism was involved.
We then performed two control experiments to prove unequiv-
ocally the role of the reagents/catalysts in the reaction mechanism.
First, we examined the uncatalyzed reaction in which case the
starting material was recovered quantitatively (entry 2). Exclusion
of benzyl bromide from the reaction medium resulted in decompo-
sition of the aldehyde (entry 3). Because of Pd(OAc)2, a Pd(II) cata-
lyst, can act as Lewis acid and catalyses the nucleophilic attack of
alcohols to aldehydes,16 we speculated that oxidative addition of
Pd(0) to benzyl bromide to generate a Lewis acid Pd(II) salt occurs.
With this in mind, other halides that would allow facile insertion of
Pd(0) into the C–X bond were screened (entries 4–7), but all failed
to improve the results.
Addition of triphenylcarbenium tetrafluoroborate as a hydro-
gen acceptor (entry 8), added in order to decrease the amount
of reduced benzyl alcohol, successfully suppressed its formation
Table 1
Optimization of the oxidative esterification of 4-nitrobenzaldehydea
O
O
OEt
OEt
Pd(PPh3)4 (5 mol%)
OH
OEt
+
O2N
+
K2CO3 (3 equiv.)
O2N
O2N
O2N
Δ, RX (1 equiv.)
EtOH (20 ml/mmol)
4a
5a
6a
Conv.b (%)
7a
Entry
RX
Yieldb (%) 5a/6a/7a
1
BnBr
BnBr
—
100
0
75/23/0
—
0/0/0
40/7/0
21/6/0
32/0/0
12/0/0
10/0/90
82/18/0
87/10/0 (70/10/0)
6/0/60
2c
3
100
100
100
100
84
100
100
100
80
4
5
6
7
EtBr
AllBr
PhBr
4-MeOPhI
BnBr
BnBr
BnBr
BnBr
8d
9e
10e,f
11f,g
a
b
c
d
e
f
Aldehyde (1 equiv), RX (1 equiv), K2CO3 (3 equiv), Pd(PPh3)4 (5 mol %), EtOH (20 mL/mmol), 78 °C, 5 h.
Conv. = conversion; conversions and yields were determined by NMR spectroscopy using an internal standard. Yield of isolated product shown in parentheses.
Reaction carried out without Pd catalyst.
Ph3CþBFꢀ4 (0.5 equiv) was added to the reaction mixture.
Two milliliters of EtOH was used per mmol of aldehyde.
Reaction conducted under microwave irradiation for 30 min at 90 °C.
Five equivalents of EtOH were used.
g

G. A. Heropoulos, C. Villalonga-Barber / Tetrahedron Letters 52 (2011) 5319–5322
5321
Table 2
Table 3
Optimization of the oxidative esterification of 4-methoxybenzaldehydea (4b)
Substrate scope in the palladium-catalyzed oxidative esterification with benzyl
chloridea
Entry
RX
Base
Conv.b (%)
Yieldb (5b) (%)
Entry
1
Aldehyde
Product
Yieldb (%) 5/6
1c
2
3
BnBr
BnBr
BnBr
BnBr
K2CO3
Cs2CO3
Na2CO3
Et3N
62
0
0
57
—
—
—
O
5c
77/0 (70/0)
4
0
5d
6
BnBr
BnI
4-MeOBnBr
4-O2NBnBr
BnCl
K2CO3
K2CO3
K2CO3
K2CO3
K2CO3
K2CO3
K2CO3
0
0
0
50
64
100
85
—
—
—
48
64
100 (88)
85
O
2
3
5d
5e
100/0 (93/0)
100/0 (93/0)
7
8
9
MeO
O
10e
11e
BnCl
BnBr
OMe
a
Aldehyde (1 equiv), RX (1 equiv), K2CO3 (3 equiv), Pd(PPh3)4 (5 mol %), EtOH
O
(2 mL/mmol), MW, 90 °C, 30 min.
4c
5
5f
70/0 (70/0)
96/0 (81/0)
100/0 (79/0)
0/0
b
Conv. = conversion; conversions and yields were determined by NMR spec-
Me2N
troscopy using an internal standard. Yield of isolated product shown in parentheses.
c
Reaction time was 60 min.
NaI (1 equiv) was added to the reaction.
Two equivalents of RX were used.
O
d
5g
5h
5i
e
Cl
O
6
aldehyde unreacted and solvolysis of the benzyl halides occurred.23
4-Nitrobenzyl bromide (entry 8) and benzyl chloride (entry 9) gave
similar results to benzyl bromide, but the competing solvolysis
reaction did take place and finally, doubling the equivalents of ben-
zyl chloride (entry 10) led to full conversion to ethyl 4-methoxy-
benzoate (5b) and toluene.24 Using 2 equiv of benzyl bromide
also increased the yield of the product but conversion was still
not quantitative (entry 11) and ethoxymethylbenzene was ob-
tained as a side-product23 thereby complicating the purification
and making the chloride analogue a better choice as co-oxidant.
Having found conditions that converted fully 4-methoxybenzal-
dehyde into its ethyl ester, we then successfully lowered the
amount of catalyst from 5 to 1 mol %,25 replaced absolute ethanol
with 95% commercial ethanol, and with the new lower catalyst
loading the equivalents of benzyl chloride could be reduced to
1.2 without any appreciable loss in yield. Other halogen-based
co-oxidants were also examined but 1,2-dichloroethane8 and chlo-
robenzene,11a both used in palladium- catalyzed oxidations of alco-
hols, and phenyl iodide, employed in a tandem addition-oxidation
of 2-alkynylbenzaldehydes,7d led to recovery of the starting mate-
rial when utilized under our optimized conditions. We also per-
formed a control experiment in deoxygenated ethanol under
argon and the reaction proceeded quantitatively, therefore exclud-
ing any involvement of aerial oxygen as co-oxidant in the reaction
pathway.
Various aldehydes were subjected to the improved reaction
conditions (Table 3). These proved to be effective for electron-rich
aromatic aldehydes and gave the corresponding aromatic esters in
good to excellent yields (entries 1–6), the least reactive being 4-
dimethylaminobenzaldehyde which led to the recovery of starting
material in 27% isolated yield (entry 4). Chloro- or fluoro-substi-
tuted benzaldehydes were compatible with the reaction conditions
(entries 5 and 6) but, as expected,26 reductive dehalogenation oc-
curred for the bromo substituted derivative (entry 7), and ethyl
benzoate was obtained instead.
Notably, 4-nitrobenzaldehyde gave a higher benzyl alcohol to
ester ratio (22:78) (entry 8) than when benzyl bromide was used
(11:89) (Table 1, entry 10). This trend can be generalized to elec-
tron-poor aromatic aldehydes since 3-nitrobenzaldehyde, 4-triflu-
oromethylbenzaldehyde and ethyl 4-formylbenzoate were also
reduced in discernible yields (entries 9–11). This hydride transfer
reaction to the deactivated aldehydes occurs possibly due to the
slower reactivity of benzyl chloride compared to benzyl bromide
which allows for the competing process to take place. Metallocene
derivatives such as ferrocenecarboxaldehyde (entry 12) and
F
O
7d
8
Br
O
5a
76/22 (65/18)
O2N
O
9
5j
80/20 (67/17)
85/15 (79/15)
82/18 (82/18)
NO2
F3C
O
10
11
5k
5l
O
EtO
O
O
12
5m
95/0 (93/0)
Fe
O
O
O
13
5n
5o
5p
100/0 (78/0)
100/0 (80/0)
75/0
14
15e
a
Aldehyde (1 equiv), BnCl (1.2 equiv), K2CO3 (3 equiv), Pd(PPh3)4 (1 mol %), EtOH
(2 mL/mmol), MW, 90 °C, 30 min.
b
Yield determined by NMR spectroscopy using an internal standard. The yields in
parentheses are the isolated yields.
c
Starting material was recovered (27%).
A mixture of ethyl benzoate (65%) and benzaldehyde (8%) was obtained (yields
d
by NMR spectroscopy).
e
Ethyl 3-phenylpropanoate was also formed in 25% yield.
aliphatic aldehydes (entries 13 and 14) also worked well and the
corresponding esters were obtained in good yields. Oxidative
esterification of cinnamaldehyde was satisfactory and no compet-
itive Heck reaction took place,27 but instead, ethyl 3-phenylpro-
panoate, the product from hydrogenation and oxidative
esterification, was formed in 25% yield (entry 15).
A possible mechanism for the oxidative esterification reaction
involves initial oxidative addition of the benzyl derivative to the

5322
G. A. Heropoulos, C. Villalonga-Barber / Tetrahedron Letters 52 (2011) 5319–5322
6. (a) Hashmi, A. S.; Lothschütz, C.; Ackermanm, M.; Doepp, R.; Anantharaman, S.;
Ph
Marchetti, B.; Bertagnolli, H.; Rominger, F. Chem. Eur. J. 2010, 16, 8012–8019;
(b) Rosa, J. N.; Reddy, S.; Candeias, N. R.; Cal, P. M. S. D.; Gois, P. M. P. Org. Lett.
2010, 12, 2686–2689; (c) Yoo, W.-J.; Li, C.-J. Tetrahedron Lett. 2007, 48, 1033–
1035; (d) Kiyooka, S.; Wada, Y.; Ueno, M.; Yokoyama, T.; Yokoyama, R.
Tetrahedron 2007, 63, 12695–12701 (See also citations in Ref. 2).
7. (a) Zhang, M.; Zhang, S.; Zhang, G.; Chen, F.; Cheng, J. Tetrahedron Lett. 2011, 52,
2480–2483; (b) Liu, C.; Wang, J.; Meng, L.; Deng, Y.; Li, Y.; Lei, A. Angew. Chem.,
Int. Ed. 2011, 50, 5144–5148; (c) Lerebours, R.; Wolf, C. J. Am. Chem. Soc. 2006,
128, 13052–13053; (d) Qin, C.; Wu, H.; Chen, J.; Liu, M.; Cheng, J.; Su, W.; Ding,
J. Org. Lett. 2008, 10, 1537–1540; (e) Wei, L.-L.; Wei, L.-M.; Pan, W.-B.; Wu, M.-J.
Synlett 2004, 1497–1502.
8. Aït-Mohand, S.; Hénin, F.; Muzart, J. Tetrahedron Lett. 1995, 36, 2473–2476.
9. Nagashim, H.; Sato, K.; Tsuji, J. Tetrahedron 1985, 41, 5645–5651.
10. Liégault, B.; Renaud, J.-L.; Bruneau, C. Chem. Soc. Rev. 2008, 37, 290–299.
11. (a) Landers, B.; Berini, C.; Want, C.; Navarro, O. J. Org. Chem. 2011, 76, 1390–
1397; (b) Guram, A. S.; Bei, X.; Turner, H. W. Org. Lett. 2003, 5, 2485–2487; (c)
Tamaru, Y.; Yamada, Y.; Inoue, K.; Yamamoto, Y.; Yoshida, Z.-I. J. Org. Chem.
1983, 48, 1286–1292.
LnPd(0)
Cl
Ph
C
LnPd
Ph
Ln(Cl)Pd
Ph
A
H
O
R
O
R
OEt
LnPd
Ph
Ln(Cl)Pd
EtOH
Ph
O
O
H
OEt
B
12. Rodríguez, N.; Medio-Simón, M.; Asensio, G. Adv. Synth. Catal. 2007, 349, 987–
991.
13. See the original procedure where the intermediate is isolated: Vutukuri, D. R.;
Bharathi, P.; Yu, Z.; Rajasekaran, K.; Tran, M.-H.; Thayumanavan, S. J. Org. Chem.
2003, 68, 1146–1149.
R
R
HCl
Scheme 2. Proposed reaction mechanism.
14. For a successful example of this one-pot sequence, see: Barnickel, B.; Schobert,
R. J. Org. Chem. 2010, 75, 6716–6719.
15. Cannizzaro, S. Liebigs Ann. Chem. 1853, 88, 129–130.
16. Asao, N.; Nogami, T.; Takahashi, K.; Yamamoto, Y. J. Am. Chem. Soc. 2002, 124,
764–765.
17. Transition metal catalyzed acetalization of aldehydes and ketones has been
reported, see: (a) Gregg, B. T.; Golden, K. C.; Quinn, J. F. Tetrahedron 2008, 64,
3287–3295; (b) Cataldo, M.; Nieddu, E.; Gavagnin, R.; Pinna, F.; Strukul, G. J.
Mol. Catal. 1999, 142, 305–316.
18. Acetal formation has been reported as a competing reaction when electron-
deficient benzaldehydes were used in a gold-catalyzed oxidative esterification,
see Ref. 6a.
19. Ninety two percent conversion for 3-nitrobenzaldehyde [ester (5j): benzyl
alcohol (6j) 89:11] and 71% for 4-trifluoromethylbenzaldehyde [ester (5k):
benzyl alcohol (6k) 96:4].
palladium(0) species to generate benzylpalladium complex A,
which then undergoes coordination to the aldehyde and allows
the addition of ethanol onto the carbonyl group to produce the
oxypalladium complex B. b-Hydride elimination generates the es-
ter and the palladium hydride species C which through reductive
elimination affords toluene and regenerates the Pd catalyst
(Scheme 2).
In summary, we have demonstrated a facile and mild palla-
dium-catalyzed oxidation of aldehydes into esters. Our methodol-
ogy employs a low loading of palladium and the readily available
co-oxidant is fully transformed into toluene, resulting in an easy
to work-up and purified reaction. Moreover, this method does
not require inert conditions or the use of ligands and works well
for aromatic, saturated and unsaturated aldehydes.28
20. Ninety
three
percent
conversion
for
both
nonanal
and
3-
phenylpropionaldehyde.
21. Sixty six percent conversion for 4-methylbenzaldehyde and 55% for 4-
methoxybenzaldehyde.
22. Bucos, M.; Villalonga-Barber, C.; Micha-Screttas, M.; Steele, B. R.; Screttas, C. G.;
Heropoulos, G. A. Tetrahedron 2010, 66, 2061–2065.
23. Products of solvolysis were detected by GC and NMR.
24. The 1H NMR spectrum of the crude product showed no signal at 4.55 ppm
Acknowledgment
corresponding to BnCl and the appearance of
corresponding to toluene.
a singlet at 2.33 ppm
25. Lowering further the catalyst loading to 0.5 mol % decreased the conversion to
78%.
26. (a) Chen, J.; Zhang, Y.; Yang, L.; Zhang, X.; Liu, J.; Li, L.; Zhang, H. Tetrahedron
2007, 63, 4266–4270; (b) Zask, A.; Helquist, P. J. Org. Chem. 1978, 43, 1619–
1620.
This work was supported by funding provided by the European
Commission for the FP7-REGPOT-2009-1 Project ‘ARCADE’ (Grant
Agreement No. 245866).
27. Heck reactions have been reported on unsaturated substrates when
bromobenzene was used as co-oxidant, see Ref. 11b.
References and notes
28. General procedure: A 10 mL reaction vessel was charged in air with Pd(PPh3)4
(6 mg, 1 mol %), aldehyde (0.5 mmol), K2CO3 (207 mg, 1.5 mmol), benzyl
1. Larock, R. C. Comprehensive Organic Transformations; VCH: New York, 1999.
2. For a recent review, see: Ekoue-Kovi, K.; Wolf, C. Chem. Eur. J. 2008, 14, 6302–
6315.
3. (a) O’Conner, B.; Just, G. Tetrahedron Lett. 1987, 28, 3235–3236; (b) Abiko, A.;
Roberts, J. C.; Takemasa, S.; Masamune, S. Tetrahedron Lett. 1986, 27, 4537–
4540.
4. (a) Malik, P.; Chakraborty, D. Tetrahedron Lett. 2010, 51, 3521–3523; (b) Wu, X.-
F.; Darcel, C. Eur. J. Org. Chem. 2009, 1144–1147; (c) Gopinath, R.; Patel, B. K.
Org. Lett. 2000, 2, 577–579.
5. (a) Sharma, R. K.; Sharma, C. J. Macromol. Sci. A 2011, 48, 155–163; (b) Reddy, K.
R.; Venkateshwar, M.; Maheswari, C. U.; Prashanthi, S. Synth. Commun. 2010,
40, 186–195; (c) Travis, B. R.; Sivakumar, M.; Hollist, O.; Borhan, B. Org. Lett.
2003, 5, 1031–1034.
chloride (70 lL, 0.6 mmol) and EtOH (1 mL). The vessel was sealed and
submitted to microwave irradiation for 30 min at 90 °C, using an initial power
of 30 W. (Microwave reactions were carried out with a CEM Discover 300 W
monomode microwave instrument. The closed vessels used were special glass
tubes with self-sealing septa that controlled pressure with appropriate sensors
on the top (outside the vial). The temperature was monitored through a non-
contact infrared sensor centrally located beneath the cavity floor.) The mixture
was then allowed to cool to room temperature, filtered over a pad of CeliteÒ
and rinsed with EtOH (5 mL). The filtrate was concentrated in vacuo and the
product was purified by flash chromatography on silica gel (CH2Cl2/hexane).