A solvent-controlled highly efficient Pd-C catalyzed hydrogenolysis of benzaldehydes to methylbenzenes via a novel 'acetal pathway'

Article abstract of DOI:10.1016/j.tet.2007.06.105

Pd-C catalyzed hydrogenolysis of benzaldehydes to methylbenzenes has been described to proceed via a 'benzenemethanol pathway'. In this article, a novel 'acetal pathway' was first revealed by a systematic study when lower alcohols were used as solvents and a solvent-controlled highly efficient procedure was established.

Full text of DOI:10.1016/j.tet.2007.06.105

Tetrahedron 63 (2007) 9382–9386
A solvent-controlled highly efficient Pd–C catalyzed
hydrogenolysis of benzaldehydes to methylbenzenes
via a novel ‘acetal pathway’
*
Lixin Xing, Xinyan Wang, Chuanjie Cheng, Rui Zhu, Bo Liu and Yuefei Hu
Key Laboratory of Bioorganic Phosphorus Chemistry and Chemical Biology (Ministry of Education), Department of Chemistry,
Tsinghua University, Beijing 100084, PR China
Received 18 April 2007; revised 27 June 2007; accepted 29 June 2007
Available online 10 July 2007
Abstract—Pd–C catalyzed hydrogenolysis of benzaldehydes to methylbenzenes has been described to proceed via a ‘benzenemethanol
pathway’. In this article, a novel ‘acetal pathway’ was first revealed by a systematic study when lower alcohols were used as solvents and
a solvent-controlled highly efficient procedure was established.
Ó 2007 Elsevier Ltd. All rights reserved.
1. Introduction
done, which makes possible a prediction of satisfactory
conditions, quite different conditions were used in the liter-
ature. For example, MeOH, EtOH, EtOAc or HOAc was
randomly used as a solvent. The acid catalyst was or was
not employed without purpose. High pressure,2 a large
amount of catalyst,3 or prolonged reaction time4 was often
used to overcome the low efficiency and minimize the resid-
ual intermediate.
Palladium on carbon (Pd–C) is the most often used palla-
dium-based catalyst featured with commercially available,
cheap price, and easy regeneration. Pd–C catalyzed hydro-
genation has been widely employed in numerous organic
transformations in both academic and industry. Pd–C cata-
lyzed hydrogenation of benzaldehyde (1) is a fundamental
transformation to yield benzenemethanol (2) and methyl-
benzene (3) (Scheme 1). The active benzaldehyde (1) prefers
to carry out a hydrogenolysis to give methylbenzene (3) as
a major product.1
Therefore, there is a great need to make a systematic study
on this transformation, by which the function of each reac-
tant could be well recognized and a highly efficient method
could be established. In this article, we have made endeavors
to accomplish this goal.
CHO
CH2OH
+
CH3
Pd-C, H2
R
R
R
2
3
2. Results and discussion
1
Scheme 1.
Several hundred tons per year of 1,2,3-trimethoxy-5-methyl-
benzene (3a) were consumed as a precursor for the synthesis
of coenzyme Q10.6,7 So far, 3a was prepared mainly by re-
ductive deoxygenation of 3,4,5-trimethoxy-benzaldehyde
(1a).2,7,8 Surprisingly, only one protocol using Pd–C cata-
lyzed hydrogenolysis was reported in the literature, in which
3a was obtained in 97% yield in acetic acid under 40 atmo-
spheric hydrogen pressure for 3 days.2
Since alkoxyl-substituted benzaldehydes (1) are economic
and readily available starting materials, Pd–C catalyzed
hydrogenolysis of them has afforded a practical method
for preparation of alkoxyl-substituted methylbenzenes
(3).2–5 This transformation has been ambiguously described
to proceed via a ‘benzenemethanol pathway’ because benz-
enemethanol (2) was often isolated as an intermediate or as
a major by-product.1,4b Since no systematic work has been
During the study for an efficient preparation of 3a by reduc-
tive deoxygenation of 1a, we observed that the products of
Pd–C catalyzed hydrogenation of 1a strongly relied upon
the reaction solvents (Table 1). The desired 3a was obtained
in 98% yield in MeOH or EtOH (entries 1 and 2). In other
solvents, Pd–C catalyst was remarkably retarded with uptake
Keywords: Solvent-controlled; Pd–C catalyst; Hydrogenolysis; Benzalde-
hydes.
* Corresponding author. Tel.: +86 10 62795380; fax: +86 10 62771149;
e-mail: yfh@mail.tsinghua.edu.cn
0040–4020/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2007.06.105

L. Xing et al. / Tetrahedron 63 (2007) 9382–9386
Table 2. The intermediates in the hydrogenation of 1a
9383
Table 1. Effect of solvent on the hydrogenation of 1aa
CHO
CH(OR)2
Pd-C (5 wt%), H2 (1 atm)
aq HCl (3 wt%), MeOH, rt
10% Pd-C (5 wt%), H2 (1 atm)
37% aq HCl (3 wt%), Solvent, rt
1a
3a
+
4a
+
MeO
MeO
OMe
MeO
OMe
OMe
1a
OMe
5a R = Me
CH2OR
CH2OH
Me
Entry Time (min) Conversion of 1a (%) 3a (%) 4a (%) 5a (%)
1
2
3
5
8
90
100
100
100
5
50
98
65
48
0
27
0
0
+
+
MeO
OMe
OMe
OMe MeO
OMe
4
OMe
2a
OMe
3a
Entry Solvent
Time (h) 2ac (%) 3ac (%) 4c (%)
Next, the acetalizations of 1a with different alcohols were
tested. As shown in Table 3, dimethyl acetal 5a was obtained
in 95% yield in MeOH within 30 min (entry 1), while 5b and
5c (entries 2 and 3) were obtained in lower yields in EtOH
and i-PrOH. No 5d was detected at all when t-BuOH was
used as a solvent (entry 4). These results strongly suggested
that the selectivity for two intermediates in the hydrogeno-
lysis of 1a simply determined by the steric hindrance of
the alcohol.
1
2
3
4
5
6
MeOH
EtOH
i-PrOH
1.5
2
0
0
98
98
75
70
66
63
0 (4a, R¼Me)
0 (4b, R¼Et)
6 (4c, R¼i-Pr)
0 (4d, R¼t-Bu)
0
7b
16
26
30
34
t-BuOH 5b
EtOAc
THF
5.5b
3.5b
0
a
b
c
Pd–C catalyst: Aldrich 520888, palladium, 10 wt % on carbon powder,
dry, Engelhard code C3645.
At that time the absorption of hydrogen ceased due to the reason that Pd–C
catalyst lost activity completely.
100% conversion of 1a was observed and isolated yields were obtained.
In addition, we also observed that Pd–C alone did not cata-
lyze the acetalization of 1a (entry 5). However, when it was
used together with aq HCl, the yield of diethyl acetal (5b)
doubled (entry 6). This may result from the fact that Pd–H
intermediate was formed in the presence of aq HCl, which
has proved to be a strong catalyst for the acetalization of
carbonyl compounds.10 Thus, the smallest size and doubly
catalyzed acetalization (HCl and Pd–C) provided a reason-
able explanation why 1a was completely converted into 5a
within less than 5 min in MeOH.
of 1.0 mol of hydrogen (entries 3–6). As a result, the hydro-
genolysis was prevented from reaching completion to give
a mixture of 3,4,5-trimethoxy-benzenemethanol (2a) and
3a. 3,4,5-Trimethoxybenzyl isopropyl ether (4c) was also
isolated in entry 3.
To understand those phenomena, two types of control exper-
iments were performed. (a) After 1.0 mol of hydrogen was
consumed in the hydrogenation of 1a, compound 2a was iso-
lated as a major intermediate in entry 3 or a unique inter-
mediate in entries 4–6. (b) The intermediate 2a was
hydrogenolyzed to give 3a with same efficiency (2 h, 95–
98%) in either MeOH or t-BuOH.
Since the acetalization is a key step for the efficient hydro-
genolysis of 1a due to it prevents from the formation of
2a, the effects of acid catalyst on the acetalization of 1a
were tested. As shown in Table 4, 3a was obtained in excel-
lent chemoselectivity and yield in the presence of a catalytic
amount of strong acid, such as aq HCl, HCl or CF3CO2H
(entries 2–4). However, when the reaction proceeded with-
out acid catalyst (entry 1) or with weak acid HOAc (entry
5), a mixture with 2a as a major product was obtained.
Thus, a strong acid catalyst is essential for an efficient hydro-
genolysis of 1a because it catalyzes an efficient acetalization
of 1a.
Thus, three conclusions can be deduced: (a) benzenemetha-
nol 2a was a common intermediate for the hydrogenations in
entries 3–6, and its formation caused remarkable loss of
activity of Pd–C catalyst (for unclear reasons);9 (b) hydroge-
nolysis of 2a was not affected by solvents; (c) the highly
efficient hydrogenolysis of 1a to 3a in entries 1 and 2 must
be mediated by an ‘unknown intermediate’ instead of benz-
enemethanol (2a).
So far, we believe that a novel ‘acetal pathway’ for Pd–C
catalyzed hydrogenolysis of 1a to 3a was revealed and
To find the possible ‘unknown intermediate’, the hydroge-
nolysis of 1a in MeOH was monitored. As shown in Table
2, two samples obtained from the first 5 and 8 min in the hy-
drogenolysis were tested. The former one showed that 1a has
been already exhausted and 3,4,5-trimethoxybenzaldehyde
dimethyl acetal (5a) was isolated besides expected products
3a and 4a (entry 1). The latter one showed that 5a disap-
peared completely and the proportion of 3a increased largely
(entry 2). Thus, dimethyl acetal (5a) is the ‘unknown inter-
mediate’. This hypothesis was confirmed when 5a was
used as a substrate under the same conditions, in which 4a
was detected as an intermediate and 3a was obtained in
98% yield within 1.5 h.
Table 3. Effect of alcohol structure on the acetalization of 1a
Entry ROH
Catalysta
Time (min) 5c (%)
1
2
3
4
5
6
7
MeOH
EtOH
i-PrOH
aq HCl
aq HCl
aq HCl
30
60
60
60
60
95 (5a, R¼Me)
33 (5b, R¼Et)
8 (5c, R¼i-Pr)
0 (5d, R¼t-Bu)
0
t-BuOH aq HCl
EtOH
EtOH
Pd–Cb
aq HCl+Pd–Cb 60
68 (5b, R¼Et)
t-BuOH aq HCl+Pd–Cb 60
0
a
b
c
37% aq HCl was used.
5% w/w of Pd–C was used.
Isolated yields were obtained.

9384
L. Xing et al. / Tetrahedron 63 (2007) 9382–9386
Table 4. Effect of acid catalyst on the hydrogenolysis of 1a
the substrate 1e was converted into 3e in 98% yield within
45 min in MeOH. However, at least 15 h were required for
the same reaction with 96% conversion in EtOAc in
Ref. 4b. So far, this method cannot be applied to the
inactive benzaldehydes. For example, when fluorene-2-
carboxaldehyde was used as a substrate, a mixture of
2-methylfluorene (43%) and (fluoren-2-yl)methanol (57%)
was obtained.
10% Pd-C (5 wt%), H2 (1 atm)
Acid catalyst (3 wt%), MeOH, rt
1a
2a
3a
4a
+
+
Entry Acid catalyst
Conversion Time 2ab 3ab 4ab
of 1a (%)
(h)
(%) (%) (%)
1
2
3
4
5
None 100
aq HCl (37 wt %) 100
2.0
1.5
2.5
5.0
2.0a
90
0
0
0
75
10
100
100
100
7
0
0
0
0
13
HCl
CF3CO2H
CH3CO2H
100
100
100
3. Conclusion
a
At that time the absorption of hydrogen ceased due to the reason that Pd–C
catalyst lost reactivity completely.
Isolated yield was obtained.
In summary, Pd–C catalyzed hydrogenolysis of active benz-
aldehydes to methylbenzenes was systemically studied. We
observed that a benzylaldehyde acetal intermediate was
produced when hydrogenolysis proceeded in lower alcohol
solvent (MeOH or EtOH) instead of benzylmethanol inter-
mediate recorded in the literature. Thus, a three-stage ‘acetal
pathway’ was proposed, by which the functions of each
reaction component were well recognized. As a result, a sol-
vent-controlled highly efficient Pd–C catalyzed hydrogeno-
lysis of benzaldehydes to methylbenzenes was established.
By using this new procedure, the low efficiency occurring
in the existing procedures was overcome completely.
b
a three-stage process was proposed as shown in Scheme 2. In
the first stage, an efficient acetalization of 1a in MeOH was
doubly catalyzed by aq HCl and Pd–C to yield 5a (within
5 min). Then, a Pd–C catalyzed ‘fast C–O hydrogenolysis’
was carried out to give 4a (within 8 min). Finally, 4a under-
went a Pd–C catalyzed ‘slow C–O hydrogenolysis’ to give
3a (within 1.5 h).
Based on the above analyses, the results in Table 1 can be
well explained. By using MeOH or EtOH as a solvent, acetal
5a or 5b was produced as an intermediate through an ‘acetal
pathway’. Then, 5a or 5b underwent a double hydrogenoly-
sis of C–O bonds to give 3a with high efficiency, since benz-
aldehyde acetal and benzyl ether are extremely labile to
hydrogenolysis. Since 1a cannot be acetalized in t-BuOH
or other non-alcoholic solvents, it was hydrogenated to
give 2a as an intermediate by a ‘benzenemethanol pathway’,
which causes remarkable loss of the activity of Pd–C. There-
fore, the hydrogenation in these solvents associated with low
efficiency and was contaminated by residual intermediate
2a. The hydrogenation in i-PrOH went through both path-
ways because i-PrOH has an intermediate steric hindrance.
Actually, the ‘acetal pathway’ afforded a solvent-controlled
highly efficient Pd–C catalyzed hydrogenolysis of 1a to 3a.
4. Experimental
4.1. A typical procedure of Pd–C catalyzed
hydrogenation of 3,4,5-trimethoxybenzaldehyde (1a)
to 1,2,3-trimethoxy-5-methylbenzene (3a)
A suspension of 3,4,5-trimethoxybenzylaldehyde (1a,
491 mg, 2.5 mmol), Pd–C catalyst (24.5 mg), and aq HCl
(37%, 40 mg) in MeOH (30 mL) was hydrogenated under
room temperature and atmospheric pressure until the absorp-
tion of hydrogen ceased completely (it was carried out on an
atmospheric pressure hydrogenator). Then the catalyst was
filtrated off and the solvent was evaporated. The residue
was diluted with CH2Cl2 (30 mL) and washed with brine
and dried over Na2SO4. Removal of the solvent yielded
product 3a (446 mg, 98%), which is pure enough for satis-
factory analyses. Mp 35–37 ꢀC (lit.2 35–37 ꢀC); IR: n
2994, 2939, 2836, 1589, 1507, 1465, 1414, 1331, 1238,
1182, 1127, 1011, 968, 812, 778 cmꢁ1; 1H NMR: d 6.39 (2H,
s), 3.84 (6H, s), 3.82 (3H, s), 2.31 (3H, s); 13C NMR: d 152.8,
135.5, 133.4, 105.7, 60.6, 55.7, 21.6.
To generalize the application of ‘acetal pathway’, the hydro-
genolysis of different benzaldehydes (1a–j) was tested. As
shown in Table 5, they all gave the corresponding 3a–j in
excellent results. Under our standard conditions, o- and p-
substitutents benefited the reaction efficiency (entries 2–7),
while m-substitutents gave negative effects (entries 1, 8,
and 9). In the case of 1i, double amount of Pd–C catalyst
(10 wt %) has to be used, otherwise the hydrogenation could
not go to completion within 20 h (entry 9). Benzo-fused ring
also reduced the reactivity of substrate and prolonged reac-
tion time was required. The control experiments in entries 4,
5 and 9 clearly showed that ‘acetal pathway’ has greater
advantages than ‘benzylmethanol pathway’. For example,
Similar procedure was used to convert the substrates 1b–j ef-
ficiently to the corresponding products 3b–j. Products 3a–j
are all known compounds and their characterization data
and 1H NMR and 13C NMR spectra are reported in Supple-
mentary data.
CHO
CH(OMe)2
CH2OMe
Me
i
ii
iii
MeO
OMe
MeO
OMe MeO
OMe
MeO
OMe
OMe
OMe
5a
OMe
4a
OMe
3a
1a
Scheme 2. Reagents and conditions: (i) acid and Pd–C catalyzed acetalization; (ii) Pd–C catalyzed fast C–O hydrogenolysis; (iii) Pd–C catalyzed slow C–O
hydrogenolysis.

L. Xing et al. / Tetrahedron 63 (2007) 9382–9386
Table 5. Chemoselective hydrogenation of 1a–j to 3a–ja
9385
10% Pd-C (5 wt%), H2 (1 atm)
aq HCl (3 wt %), MeOH, rt
CHO
CH3
R
R
93-98%
1a-j
3a-j
Entry (no.)
1
3
Solvent
Pd–C (wt %)
5
Time (min)
90
Yieldb (%)
98
MeO
MeO
CHO
CHO
MeO
Me
Me
Me
1 (a)
MeOH
MeOH
MeOH
MeO
MeO
MeO
OMe
OMe
OMe
OMe
2 (b)
3 (c)
5
5
30
30
96
97
MeO
MeO
CHO
CHO
Me
MeOH
EtOAc
EtOAc
5
5
20
30
225
40
93
94
93
O
O
O
O
4 (d)
5 (e)
OMe
OMe
MeOH
EtOAc
EtOAc
5
5
5
45
270
900
98
87c
96d
MeO
CHO
CHO
MeO
Me
Me
MeO
MeO
6 (f)
7 (g)
8 (h)
MeOH
MeOH
MeOH
5
5
5
45
30
90
96
93
95
MeO
EtO
MeO
EtO
CHO
Me
EtO
EtO
O
CHO
O
Me
MeO
MeO
MeO
CHO
CHO
MeO
Me
Me
MeOH
MeOH
EtOAc
EtOAc
5
10
5
1200
120
240
380
19e
96
0f
9 (i)
20
96
OMe
OMe
10 (j)
MeOH
5
300
97
MeO
MeO
a
37% aq HCl was used.
Isolated yields were obtained.
A mixture with 10% of 2,3-dimethoxyphenylmethanol.
See Ref. 4b, the hydrogenation without aq HCl.
A mixture with 80% of 3,5-dimethoxybenzyl methyl ether.
No absorption of hydrogen was observed.
b
c
d
e
f
Acknowledgements
in the supplementary data. Supplementary data associated
with this article can be found in the online version, at
doi:10.1016/j.tet.2007.06.105.
This work was supported by the National Natural Science
Foundation of China (20025204) and the Key Subject Founda-
tion from Beijing Department of Education (XK100030514).
References and notes
Supplementary data
1. (a) Nishimura, S. Handbook of Heterogeneous Catalytic
Hydrogenation for Organic Synthesis; John Wiley & Sons:
New York, NY, 2001; Chapter 5; (b) Rylander, P. N.
Hydrogenation Methods; Academic: London, 1985; Chapter 4.
Experimental procedures, characterization data, and 1H
NMR and 13C NMR spectra for products 3a–j are provided

9386
L. Xing et al. / Tetrahedron 63 (2007) 9382–9386
2. Merz, A.; Rauschel, M. Synthesis 1993, 797–802.
(c) Negishi, E. I.; Liou, S. Y.; Xu, C.; Huo, S. Org. Lett.
2002, 4, 261–264; (d) Lipshutz, B. H.; Mollard, P.; Pfeiffer,
S. S.; Chrisman, W. J. Am. Chem. Soc. 2002, 124, 14282–
14283; (e) Lipshutz, B. H.; Bulow, G.; Fernandez-Lazaro, F.;
Kim, S.-K.; Lowe, R.; Mollard, P.; Stevens, K. L. J. Am.
Chem. Soc. 1999, 121, 11664–11673; (f) Lipshutz, B. H.;
Bulow, G.; Lowe, R. F.; Stevens, K. L. J. Am. Chem. Soc.
1996, 118, 5512–5513; (g) Carpino, L. A.; Triolo, S. A.;
Berglund, R. A. J. Org. Chem. 1989, 54, 3303–3310; (h)
Keinan, E.; Eren, D. J. Org. Chem. 1987, 52, 3872–3875; (i)
Terao, S.; Kato, K.; Shiraishi, M.; Morimoto, H. J. Org.
Chem. 1979, 44, 868–869.
3. (a) Benavides, A.; Peralta, J.; Delgado, F.; Tamariz, J. Synthesis
2004, 2499–2504; (b) Sagar, K. S.; Chang, C.; Wang, W.; Lin,
J.; Lee, S. Bioorg. Med. Chem. 2004, 12, 4045–4054; (c) Chida,
N.; Ohtsuka, M.; Nakazawa, K.; Ogawa, S. J. Org. Chem. 1991,
56, 2976–2983.
4. (a) Wang, Q.; Yang, Y.; Li, Y.; Yu, W.; Hou, Z. Tetrahedron
2006, 62, 6107–6112; (b) Connolly, T. J.; Matchett, M.;
McGarry, P.; Sukhtankar, S.; Zhu, J. Org. Process Res. Dev.
2004, 8, 624–627; (c) Anderson, W. K.; Boehm, T. L.;
Makara, G. M.; Swann, R. T. J. Med. Chem. 1996, 39, 46–55;
(d) Monte, A. P.; Marona-Lewicka, D.; Parker, M. A.;
Wainscott, D. B.; Nelson, D. L.; Nichols, D. E. J. Med.
Chem. 1996, 39, 2953–2961; (e) Syper, L.; Kloc, K.;
Mlochowski, J. Tetrahedron 1980, 36, 123–129.
5. (a) Boyer, F.-D.; Ducrot, P.-H. Tetrahedron Lett. 2005, 46,
5177–5180; (b) Yamaguchi, S.; Muro, S.; Kobayashi, M.;
Miyazawa, M.; Hirai, Y. J. Org. Chem. 2003, 68, 6274–6278;
(c) Lai, C.; Shen, Y.; Wang, M.; Kameswara Rao, N. S.;
Liao, C. J. Org. Chem. 2002, 67, 6493–6502; (d) Clive,
D. L. J.; Tao, Y.; Bo, Y.; Hu, Y.; Selvakumar, N.; Sun, S.;
Daigneault, S.; Wu, Y. Chem. Commun. 2000, 1341–1350;
(e) Tibor, E.; Tibor, T. Synth. Commun. 1990, 20, 3219–3222;
(f) de Paulis, T.; Kumar, Y.; Johansson, L.; Raemsby, S.;
7. (a) Ji, Y.; Xu, W.; Jin, W.; Yue, W. Synth. Commun. 2006, 36,
1961–1965; (b) Lu, L.; Chen, F. Synth. Commun. 2004, 34,
4049–4053; (c) Chida, A. S.; Vani, P. V. S. N.; Chandra-
sekharam, M.; Srinivasan, R.; Singh, A. K. Synth. Commun.
2001, 31, 657–660.
8. (a) Cheng, C.; Song, H.; Chen, Y.; Wang, Y.; Ding, S. Hecheng
Huaxue 2004, 12, 319–322; (b) Iwamura, T.; Kihara, Y.;
Hattori, K. PCT Int. Appl. WO 2003095399, 2003; (c)
Bottke, N.; Fischer, R.; Noebel, T.; Roesch, M. PCT Int.
Appl. WO 2003062174, 2003.
9. An interruption has been observed during the hydrogenation
of benzaldehyde to benzenemethanol without any explanation
in early references (see Ref. 1b). Recently, a detail study on
Pd–C catalyzed hydrogenation of 2,3-dimethoxybenzalde-
hyde (1e) in EtOAc also showed that the conversion of 1e
to 2,3-dimethoxybenzenemethanol (2e) caused Pd–C catalyst
to significantly lose its reactivity for unknown reasons (see
Ref. 4b).
€
Hall, H.; Sallemark, M.; Kristina, A.-M.; Oegren, S. O.
J. Med. Chem. 1986, 29, 61–69; (g) dePaulis, T.; Kumar, Y.;
Johansson, L.; Raemsby, S.; Florvall, L.; Hall, H.; Aengeby-
Moeller, K.; Oegren, S. O. J. Med. Chem. 1985, 28, 1263–
1269; (h) Wenkert, E.; Loeser, E. M.; Mahapatra, S. N.;
Schenker, F.; Wilson, E. M. J. Org. Chem. 1964, 29, 435–439.
6. (a) Lipshutz, B. H.; Lower, A.; Berl, V.; Schein, K.; Wetterich,
F. Org. Lett. 2005, 7, 4095–4097; (b) Min, J.-H.; Lee, J.-S.;
Yang, J.-D.; Koo, S. J. Org. Chem. 2003, 68, 7925–7927;
10. (a) Fujii, Y.; Furugaki, H.; Tamura, S.; Yano, S.; Kita, K. Bull.
Chem. Soc. Jpn. 2005, 78, 456–463; (b) Bethmont, V.; Fache,
F.; Lemaire, M. Tetrahedron Lett. 1995, 36, 4235–4236.