HCl/DMF for enhanced chemoselectivity in catalytic hydrogenolysis reactions

Article abstract of DOI:10.1016/j.tetlet.2010.07.142

An improved, chemoselective hydrogenolysis method has been developed. By employing a solvent-acid combination (i.e., DMF-aq HCl) we were able to favor debenzylation rather than aromatic hydrogenation and acid-mediated bond cleavage which are the two main drawbacks of these reactions. The generality of the method, which was primarily developed as a solution to a carbohydrate problem, is shown by the successful hydrogenolysis of 1,8-naphthalide, a previously unsolved problem.

Full text of DOI:10.1016/j.tetlet.2010.07.142

Tetrahedron Letters 51 (2010) 5200–5202
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
HCl/DMF for enhanced chemoselectivity in catalytic hydrogenolysis reactions
Agata Ochocinska, Anna Siegbahn, Ulf Ellervik *
Organic Chemistry, Lund University, PO Box 124, SE 2210 00 Lund, Sweden
a r t i c l e i n f o
a b s t r a c t
Article history:
An improved, chemoselective hydrogenolysis method has been developed. By employing a solvent-acid
combination (i.e., DMF–aq HCl) we were able to favor debenzylation rather than aromatic hydrogenation
and acid-mediated bond cleavage which are the two main drawbacks of these reactions. The generality of
the method, which was primarily developed as a solution to a carbohydrate problem, is shown by the
successful hydrogenolysis of 1,8-naphthalide, a previously unsolved problem.
Received 9 June 2010
Revised 15 July 2010
Accepted 23 July 2010
Available online 3 August 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Due to the ease of formation, the chemical stability and an array
of mild methods for deprotection, the benzyl group is often pre-
ferred for protection of hydroxy groups, amines, and thiols in syn-
thetic organic chemistry. The most common method for selective
debenzylation is metal-catalyzed hydrogenolysis, and palladium
is usually the preferred catalyst due to less probability of satura-
tion of aromatic structures, in contrast to platinum and ruthenium
for hydrogenolysis of compounds containing acid-sensitive groups
as well as aromatic residues prone to saturation.
There is a plethora of palladium reagents recommended for
debenzylation. Unfortunately, most reagents suitable for hydrog-
enolysis of benzyl ethers are also appropriate for aromatic ring
hydrogenation. Since the first choice, that is, Pd/C, was not active
enough, we turned our attention to the more active Pearlman’s cat-
1
5
catalysts.
2
alyst, that is, [Pd(OH) ].
Palladium on activated charcoal (Pd/C) and palladium hydrox-
ide [Pd(OH) ] are the two most widely used types of palladium
2
Catalytic debenzylation of the model compound 1, using palla-
dium hydroxide gave good conversions in several solvents. Unfor-
tunately, all the reactions were accompanied by saturation of the
outer naphthalene ring to yield substantial amounts of compound
3 (Table 1). The formation of 3 persisted regardless of the solvent
or the amount of catalyst. We were unable to separate 2 from 3
using standard purification methods.
for these reactions, and common solvents include ethanol, acetic
acid, tetrahydrofuran, hexanol, and toluene.1 In addition, strong
acids, such as hydrochloric acid or perchloric acid, might be used
to protonate the ether oxygen atoms and thereby facilitate the
2
debenzylation. It has also been shown that chloride anions have
a general rate-accelerating effect on the hydrogenolysis reactions,
probably due to the formation of a highly active palladium oxo-
chloride complex on the surface of the catalyst.3
Since the initial studies using Pearlman’s catalyst resulted in
unacceptable amounts of partly saturated product we then turned
In a recent project aimed at anti-tumor drugs based on naph-
4
thoxylosides we encountered unexpected selectivity problems.
Pd/C,
2
H , THF
Despite the general belief that catalytic debenzylations are simple,
some cases show unusually low chemoselectivity. The straightfor-
ward debenzylation of compound 1 proved to be such an example.
In preliminary investigations we tested palladium on activated
charcoal (Pd/C) from several suppliers in a range of solvents (e.g.,
methanol, THF, ethyl acetate and acetic acid). To our great surprise,
at ambient pressure, compound 1 was inert to these conditions, ex-
cept for partial saturation of the naphthol residue. Moreover, the
reaction in EtOH at increased hydrogen pressure (34 atm) resulted
in complete saturation of the outer ring of the naphthalene unit to
give compound 3 (Scheme 1). Finally, addition of acids, for exam-
ple, hydrochloric or sulfuric acid, mainly resulted in cleavage of
the acid-sensitive glycosidic bond to release naphthol 4. We there-
fore initiated a study to find a new, efficient, and selective method
O
HO
HO
O
O
OH
2
O
BnO
BnO
O
OBn
1
Pd/C,
H (40 bar),
2
EtOH
O
HO
HO
OH
3
Pd/C,
H , THF,
2
H SO
HO
2
4
4
*
Corresponding author. Tel.: +46 46 2228220; fax: +46 46 2228209.
E-mail address: ulf.ellervik@organic.lu.se (U. Ellervik).
Scheme 1. Reaction outcomes from the debenzylation of compound 1.
0
040-4039/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2010.07.142

A. Ochocinska et al. / Tetrahedron Letters 51 (2010) 5200–5202
5201
Table 1
sulted in slow conversion, the reactions were complete only after
9 h and also saturation of the aromatic moiety (3–13%). Acetic
acid proved the least useful additive since these reactions were
not complete even after 29 h. In contrast, addition of hydrochloric
acid (37%) gave good results with respect to conversion. The reac-
tions turned out to be dependent on the equivalents of acid used
Benzyl cleavage using Pearlman’s catalyst
1
Amount of catalyst (wt %)
Solvent
Degree of saturationa (%)
1
2
2
2
0
0
0
0
EtOAc
AcOH
EtOH
THF
20
22
No reaction
12
23
(Fig. 1). However, the conversions were accompanied by some sat-
1
00
THF
uration, 1–2%, and the substoichiometric yields were due to acid-
catalyzed glycosidic bond cleavage yielding naphthol 4. The yields
were also clearly dependent on the work-up method. The reactions
that were simply concentrated showed decreased yields of the
product with increased amounts of HCl used. In contrast, neutral-
ization with triethylamine generally increased the yields signifi-
cantly. To confirm the acid-directed hydrolysis, we performed a
simple experiment where compound 1 was dissolved in acetone-
containing sulfuric acid, which clearly indicated the formation of
naphthol.
a
Calculated from the crude 1H NMR spectra.
to a Degussa type catalyst, which has proven valuable in earlier
hydrogenolysis studies.6 The Degussa type catalyst comprises Pd/
C with 10% Pd loading (dry basis) on a wet support (50% water).
This particular catalyst was used in a series of debenzylation
reactions of 1 under atmospheric pressure of hydrogen, varying
the solvent (Table 2). The resulting reaction mixtures were ana-
lyzed by TLC in 1-h intervals for the first 4 h, which indicated that
the fastest consumption of the starting material took place in DMF,
To moderate the effect of the acid, while keeping the concentra-
tion of the activating chloride anions constant, we changed the sol-
vent from acetone to DMF. The protonated form of acetone has a
followed by acetone and THF. The reactions in CH
toluene, 1,4-dioxane, and CH Cl were not complete after 36 h.
3 2
CN, EtOAc, Et O,
7
2
2
pK of À7, similar to that of HCl. On the contrary, DMF is more ba-
a
Unfortunately, all the reactions were accompanied by partial sat-
uration. Acetone proved to be the best solvent with only 13% satura-
tion. However, it is known that preconditioning of the catalyst by
simply purging the catalyst with hydrogen gas before addition of
the substrate can improve the reaction. Thus, preconditioning of
the Degussa type catalyst reduced substantially the proportion of
saturation to low levels, that is 1% and 2% in acetone and DMF,
respectively. Although these reactions gave clean conversions the
yields obtained were generally low (approximately 50%).
To improve the yields and to further suppress saturation of the
aromatic moiety, we introduced acid additives. Several different
acids were tested using the Degussa catalyst in acetone. Although
sulfuric acid accelerated significantly the reaction rate (1 was con-
sumed after 1 h) it gave a substantial degree of saturation (2–12%)
accompanied by glycosidic bond cleavage. Trifluoroacetic acid re-
sic with a pKa of the conjugated acid that is protonated DMF, of
8
approximately À1. A series of experiments in DMF was then con-
ducted using 10% Pd/C (Degussa type) and different equivalents of
HCl (Fig. 2).
The reactions in DMF showed a clearly different behavior with a
9
maximum 88% isolated yield using 22 equiv of acid. No saturated
compound was found, nor was any glycosidic bond cleavage ob-
served. In the search for alternative sources of chloride anions, LiCl
and CsCl were used in similar reactions in both DMF and acetone.
Generally, LiCl had a higher activating effect compared to CsCl, but
neither gave results comparable with HCl, nor led to completion of
these reactions, even after five days.
Subsequently, the conditions were tested under high pressure
1
(50 bar). According to the H NMR spectrum of the crude reaction
mixture, the deprotection of 1 was successful and no saturation or
glycosidic bond cleavage was observed. This stresses the role of
DMF in the chemoselectivity of these hydrogenolysis reactions,
allowing debenzylation, but preventing saturation and acid
hydrolysis.
Table 2
Benzyl cleavage using the Degussa type catalyst
Solvent
Degree of saturation without
preconditioning (%)
Degree of saturation with
preconditioning (%)
To further evaluate the role of DMF, and to reduce possibly its
amount and thus simplifying laboratory manipulations, we tested
DMF as an additive (Table 3). Although the reaction rates and
yields of the product varied, the principle of using DMF/HCl as
additives for selective debenzylation without the risk of over-satu-
ration or acid-related side reactions was proved.
a
a
DMF
Acetone
THF
17
13
44
2
1
25
a
Calculated from the crude 1H NMR spectra.
Finally, a number of commercially available palladium-based
catalysts were submitted to these conditions, to evaluate the
Figure 1. Debenzylation of compound 1 using the Degussa catalyst in acetone with
the addition of HCl. Simple work-up with evaporation (diamonds) generally gave
lower yields compared to neutralization with triethylamine (circles).
Figure 2. Debenzylation of compound 1 using the Degussa catalyst in DMF with the
addition of HCl. Each data point represents at least two experiments.

5
202
A. Ochocinska et al. / Tetrahedron Letters 51 (2010) 5200–5202
Table 3
tions such as saturation of aromatic residues or decomposition of
The use of DMF as an additive
acid-labile groups. The palladium is activated by the hydrochloric
acid, and DMF buffers the proton activity. Using these conditions
we were able to isolate compound 2 in 88% yield without any sign
of saturation or acid-mediated cleavage of the glycoside. The use-
fulness of the method was further verified in a previously unsolved
Solvent
CH CN
Et
Reaction time (h)
Yield of 2 (%)
3
45
29
46
41
—
66
—
2
O
Toluene
Acetone
Not complete (46 h)
1.5
Not complete (46 h)
13,14
hydrogenolysis problem.
2
CH Cl
2
EtOAc
THF
6
5
50
54
Acknowledgments
General conditions: HCl (22 equiv), DMF (1 mL) in solvent (3 mL).
The Swedish Research Council, Lund University, and The Crafo-
ord Foundation supported this work. We thank Martin Axelsson for
preliminary studies.
Table 4
The use of other palladium catalysts in DMF with the addition of HCl
Supplementary data
Palladium catalyst
Reaction time (h)
Conversiona (2) (%)
1
1
1
1
0% Pd/C (Aldrich)
0% Pd/C, unreduced (Acros)
2.5
8
8
72
71
79
71
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.tetlet.2010.07.142.
0–20% Pd(OH)
2
(Aldrich)
0% Pd/C, dry (AlfaAesar)
8
References and notes
a
Calculated from 1H NMR spectra.
1
2
3
.
.
.
Smith, G.; Notheisz, F. Heterogeneous Catalysis in Organic Chemistry; Elsevier,
Academic Press, 1999.
(a) Hawker, S.; Bhatti, M. A.; Griffin, K. G. Chim. Oggi 1992, 10, 49–51; (b)
Nishimura, S. Bull. Chem. Soc. Jpn. 1959, 32, 1155–1157.
Ronchin, L.; Toniolo, L.; Cavinato, G. Appl. Catal. 1997, 165, 133–145.
O
OH
O
OH
O
O
1
0% Pd/C
+
4. Nilsson, U.; Johnsson, R.; Fransson, L.-Å.; Ellervik, U.; Mani, K. Cancer Res. 2010,
0, 3771–3779.
5. Pearlman, W. M. Tetrahedron Lett. 1967, 8, 1663–1664.
HCl (22 equiv.),
DMF, H2
7
6.
Markus, H.; Mäki-Arvela, P.; Kumar, N.; Kul’Kova, N. V.; Eklund, P.; Sjöholm, R.;
Holmbom, B.; Murzin, D. Y. Catal. Lett. 2005, 103, 125–131.
5
6 (58%)
7 (22%)
7.
Serjeant, E. P.; Dempsey, B.. Ionization Constants of Organic Acids in Solution. In
IUPAC Chemical Data Series No. 23; Pergamon Press: Oxford, UK, 1979.
Combelas, P.; Costes, M.; Garrigou-Lagrange, C. Can. J. Chem. 1975, 53, 442–447.
2-Naphthyl b-D-xylopyranoside (2). 10% Pd/C (Degussa type, 18 mg) was
suspended in DMF (2 mL) and aq HCl (0.040 mL, 22 equiv) was added in DMF
Scheme 2. Chemoselective hydrogenolysis of 1,8-naphthalide.
8
9
.
.
(
1 mL). The flask was evacuated and filled with H
2
while stirring. Compound 1
was re-
importance of the catalyst. The reactions were conducted in DMF
with addition of HCl (aq 37%, 22 equiv.) at ambient pressure (Table
(
35 mg, 0.060 mmol) was then added in DMF (1 mL) after 10 min, H
2
introduced and stirring continued for another 1.5 h. The reaction mixture was
quenched with Et N (0.090 mL) and filtered through Celite. The filtrate was
4).
3
diluted with H
organic layer was washed with brine (40 mL), dried over MgSO
concentrated under vacuum. The resulting crude was chromatographed (SiO
heptane/EtOAc 1:1 to 0:1) to give 2 (88%). Compound data: Fritz, T. A.;
Lugemwa, F. N.; Sarkar, A. K.; Esko, J. D. J. Biol. Chem. 1994, 269, 300–307.
2
O (30 mL) and extracted with EtOAc (2 Â 30 mL). The combined
The new conditions for debenzylation that is aq HCl in DMF,
4
, filtered and
proved very useful since all the tested catalysts gave good conver-
sions without any side reactions. However, the required reaction
times were longer and the yields slightly lowered compared to
the Degussa catalyst.
2
,
1
0. 8-Methyl-1-naphthoic acid (6). 10% Pd/C (Degussa type, 31 mg) was suspended
in DMF (3 mL) at rt and aq HCl (0.033 mL) was immediately added in DMF
To prove the general usefulness of this method, we examined
the hydrogenolysis of 1,8-naphthalide 5 which gave the expected
(
1 mL). The flask was evacuated and filled with H
5 (19 mg, 0.10 mmol) was added in DMF (1 mL) and H
resulting reaction mixture was stirred for 15 h, and then filtered through
Celite. The filtrate was subsequently diluted with H O (30 mL) and extracted
with EtOAc (2 Â 30 mL). The combined organic layer was washed with brine
(40 mL), dried over MgSO and concentrated under vacuum. The resulting
crude was chromatographed (SiO , CH Cl /MeOH 40:1 to 10:1) to give 6
2
while stirring. After 10 min,
2
was re-introduced. The
-methyl-1-naphthoic acid (6) in 58% yield (Scheme 2).10
8
2
Chemoselective hydrogenolysis of 1,8-naphthalide was, until
now, an unsolved problem. Pourahmady et al. originally submitted
,8-naphthalide 5 to different hydrogenolysis conditions, but all at-
4
1
2
2
2
11
(11.9 mg, 62%). Compound data: Fritz, J. M.; Ramos, E. L.; Platz, M. S. J. Org.
Chem. 1985, 50, 3522–3526.
tempts led to a mixture of tetralin products. The problem turned
out to be the energetically favorable release of peri strain of 1,8-
disubstituted naphthalenes under normal palladium-catalyzed
hydrogenolysis conditions. Even though it is possible to obtain
the acid by other means, our results allude to the general useful-
ness of this method for complicated deprotections.
11. Pourahmady, N.; Vickery, E. H.; Eisenbraun, E. J. J. Org. Chem. 1982, 47, 2590–
593.
2
12
12. (a) Månsson, M. Acta Chem. Scand. B 1974, 24, 677–680; (b) Nieuwstad, T. J.;
Klapwijk, P.; van Bekkum, H. J. Catal. 1973, 24, 2269–2272; (c) Kawai, H.;
Nagasu, T.; Takeda, T.; Fujiwara, K.; Tsuji, T.; Ohkita, M.; Nishida, J.-i.; Suzuki, T.
Tetrahedron Lett. 2004, 45, 4553–4558.
1
3
1
3. Brown, R. F. C.; Eastwood, F. W.; Kissler, B. E. Aust. J. Chem. 1989, 42, 1435–
445.
14. Rylander, P. N. Hydrogenation Methods; Academic Press: New York, 1985.
To conclude, we have reported new reaction conditions, that is
1
1
0% Pd/C Degussa type catalyst in DMF using aqueous hydrochloric
acid as an additive, for catalytic debenzylation without side reac-