Chemoselective Reduction of Complex α,β-Unsaturated Ketones to Allylic Alcohols over Ir-Metal Particles on β Zeolites

Article abstract of DOI:10.1002/anie.200352275

The combination of Ir, which has a strong affinity for carbonyl groups, with an acid H-β zeolite as a support and promotor, provides an effective catalyst for the chemoselective hydrogenation of α,β-unsaturated aldehydes and ketones to the corresponding allylic alcohols. Steroidal enones are suitable substrates for this transformation (see scheme). R = CH (Me)(CH₂)₃CH(Me)₂.

Full text of DOI:10.1002/anie.200352275

Angewandte
Chemie
procedure has been proposed by Noyori and co-workers,
which involves the use of a Ru–phosphane complex, a chiral
diamine, and a base.[7] However, although the observed
chemoselectivity is excellent, recuperation of the homoge-
neous catalyst is still a problem. Moreover, the selective
=
activation of the C O group of an unsaturated ketone on a
simple, supported metal catalyst remains a formidable
scientific challenge. Only two such reactions have been
described: Ketoisophorone can be reduced by Pd/Al2O3 in
MeOH/CH3COOH to the corresponding allylic alcohol,[8] and
4-phenyl-3-buten-2-one undergoes reduction to the unsatu-
rated alcohol in the presence of Au/Fe2O3.[9] In both cases, the
reaction has only been demonstrated for a single substrate.
Thus, a catalyst with wider scope is required. One strategy
toward the discovery of new, highly selective catalysts is to
investigate the formation of metal particles on well-defined,
porous supports.[10]
Chemoselective Hydrogenation
Herein we report that a new heterogeneous Ir catalyst
facilitates the chemoselective reduction of a variety of
unsaturated ketones and aldehydes to allylic alcohols. A
strong-acid zeolite with a large external surface area, for
example H-b, makes an optimal support. Iridium was selected
because it is more oxophilic than other metals of the Pt
group.[3] With a focus on specialty chemicals, seven unsatu-
rated ketones and three unsaturated aldehydes were selec-
tively transformed into allylic alcohols.
The standard catalyst was prepared by impregnation of a
commercial H-b zeolite from PQ (surface area = 740 m2 gÀ1,
average crystallite size = 0.2 mm, Si/Al = 12) with Ir(acac)3
(acac = acetylacetone) as a solution in toluene. After calci-
nation and reduction, the catalyst was tested on a series of
a,b-unsaturated aldehydes. The desired allylic alcohols were
obtained with high selectivity at high aldehyde conversion
(Table 1). In the hydrogenation of citral, the isolated double
Chemoselective Reduction of Complex
a,b-Unsaturated Ketones to Allylic Alcohols
over Ir-Metal Particles on b Zeolites**
Mario De bruyn, Simona Coman, Roxana Bota,
Vasile I. Parvulescu, Dirk E. De Vos, and
Pierre A. Jacobs*
Through chemoselective catalysis one aims to address only
one of several similar functional groups in a molecule. The
hydrogenation of an unsaturated carbonyl compound to an
allylic alcohol in a chemoselective manner is a major
challenge. Allylic alcohols have a wide range of applications,
for example, as fragrances, or in the pharmaceutical indus-
try.[1,2] The reduction of unsaturated alde-
hydes to primary allylic alcohols proceeds
Table 1: Hydrogenation of a,b-unsaturated aldehydes.[a]
well on supported Pt0 or Ru0.[3,4] These
metals may be modified by Lewis acids
(Snn+, Fe3+,…) or by specific supports, such
as zeolites.[5] The chemoselective reduction
of a,b-unsaturated ketones is much more
difficult. Until recently, the most successful
procedure was CeCl3-assisted reduction
with NaBH4.[6] A homogeneous catalytic
Entry
Catalyst
Substrate
P [MPa]
t [h]
Conv. [%]
Sel. [%]
1
2
3
4
Ir/H-b (2%)
Ru/H-b (2%)
Ir/H-b (2%)
Ir/H-b (2%)
cinnamaldehyde
cinnamaldehyde
a-CH3-cinnamaldehyde
citral
3
3
3
2.4
18
35
3
71
81
72
82
52
89
90
2.25
>98
[a] Conditions: substrate (50 mg), isopropyl alcohol (6.5 g), catalyst (25 mg; calcined at 3008C and
reduced at 4508C). Conv.=conversion, Sel.=selectivity.
bond was left intact, and a mixture of the allylic alcohols
geraniol and nerol was obtained. When a Ru-containing
catalyst was used under identical conditions, substantially
lower chemoselectivity was observed in the hydrogenation of
cinnamaldehyde (Table 1, entry 2). This demonstrates that Ir
is essential for high selectivity in the reaction.
To extend the scope of the catalyst to a,b-unsaturated
ketones, the reaction of the complex enone testosterone (1)
was attempted (Scheme 1). We anticipated that the structural
[*] Prof. P. A. Jacobs, M. De bruyn, Prof. D. E. De Vos
Centre for Surface Chemistry and Catalysis
Katholieke Universiteit Leuven
Kasteelpark Arenberg 23, 3001 Leuven (Belgium)
Fax: (+32)16-32-1998
E-mail: pierre.jacobs@agr.kuleuven.ac.be
Dr. S. Coman, R. Bota, Prof. V. I. Parvulescu
Department of Chemical Technology and Catalysis
University of Bucharest
B-dul Regina Elisabeta 4–12, Bucharest 70346 (Romania)
rigidity of this cyclic enone would lead to a decrease in the
[**] This research was supported by the Belgian Federal Government
(IAP Program Supramolecular Catalysis) and by the Flemish
Government (Bilateral Agreement Flanders–Romania). MDb is a
fellow of IWT (Vlaanderen). PAJ is grateful for support in the frame
of a concerted research action (GOA) of the Flemish Government.
=
reactivity of the sterically hindered C C double bond and an
increase in the susceptibility of the carbonyl group to
reduction, relative to enones in more flexible systems. The
reaction was carried out with a 1% loading of Ir on H-b in
Angew. Chem. Int. Ed. 2003, 42, 5333 –5336
DOI: 10.1002/anie.200352275
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5333

Communications
conversion (Table 2, entries 8 and 10).
The use of other solvents, such as
methanol, invariably led to the forma-
tion of the saturated alcohol. In each of
the above-mentioned cases, the allylic
alcohol androst-4-ene-3b,17b-diol was
formed with high diastereoselectivity
(d.r. > 97:3).[11,12]
Next, the reaction of cholestenone
(2, structurally related to 1) was tested.
As the large alkyl group at position 17
was found to affect the selectivity for allylic alcohol formation
(Table 3, entries 1 and 2), slightly different conditions were
required. With a Ir loading on H-b of 3%, the allylic alcohol
was obtained within 1 h in 98% yield (Table 3, entry 3). As
observed for testosterone, the formation of the 3-b-hydroxy-
4-ene allylic alcohol was highly diastereoselective (d.r.
> 97:3). In the reduction of the tricyclic compound (À)-
Scheme 1. Ir/H-b-catalyzed hydrogenation of 1 and 2.
Table 3: Hydrogenation of cholestenone (2) and isolongifolenone (3) with Ir on zeolite H-b.[a]
iPrOH, and the selectivities
observed for the allylic alcohol
were found to depend on the H2
pressure used. Thus, the selectivity
increased with increasing pressure
to reach a maximum of 64% at
~ 2 MPa and ~ 60% conversion
(Table 2, entries 1–5). With other
zeolites, the selectivities observed
were lower, for example, with H-
Entry
Substrate
Catalyst
P [MPa]
t [h]
Conv. [%]
Selectivity [%]
1
2
3
4
5
1
2
2
3
3
Ir/H-b (2%)
Ir/H-b (2%)
Ir/H-b (3%)
Ir/H-b (1%)
Ir/Na-b (1%)
2.4
2.4
2.8
6.0
6.0
0.75
1
0.67
47
184.5
87
85
100
59
76
41
>98
21[b] (52)[c]
100[b]
35
[a] Conditions: as in Table 2, but with cholestenone (66 mg) or isolongifolenone (40 mg) in iPrOH.
[b] Racemic allylic alcohol. [c] Racemic isolongifolenyl ether.
USY
(Si/Al = 5.75;
Table 2,
entry 7). When a H-Mordenite (Mor) with a Si/Al ratio of
11 was used, the selectivity remained relatively high (Table 2,
entry 6). An increase in the Ir loading of H-b led to a dramatic
isolongifolen-9-one (3) with Ir/H-b (1%) in iPrOH, the main
product isolated was isolongifolen-9-yl isopropyl ether (3b;
Table 3, entry 4). Such isolongifolenyl ethers are important
fragrance compounds.[13] This two-step process mediated by a
bifunctional metal/acid zeolite is more straightforward and
more environmentally sound than alternative reduction–
etherification procedures, in which, for example, NaH and
2-bromopropane are used. However, if desired, the ether-
ification activity of the catalyst can be suppressed by
supporting the Ir0 on a Na-b zeolite (Table 3, entry 5). With
this modified catalyst, complete chemoselectivity for the
allylic alcohol is observed, even if the reaction is slow.
Prostaglandins are important drugs and drug precursors.
The Ir/H-b catalyst was used to reduce the two enone
intermediates 4 and 5 in the prostaglandin synthesis,[2,14]
which each contain a linear rather than a cyclic enone. In
contrast with the reactions of the steroidal enones, the
reduction of these prostaglandin intermediates was found to
be more chemoselective when performed in MeOH, rather
Table 2: Hydrogenation of testosterone (1) over Ir/zeolite catalysts.[a]
Entry Catalyst
P [MPa] Solvent t [h]
Conv. [%] Sel. [%]
1
2
3
4
5
6
7
8
9
10
Ir/H-b (1%)
0.8
1.2
1.6
2.0
2.4
iPrOH
iPrOH
iPrOH
iPrOH
iPrOH 10
iPrOH
iPrOH 25
iAmOH
iPrOH
iAmOH
5
5
5
9
60
61
59
54
65
60
67
30
17
22
36
64
57
48
32
100
76
86
Ir/H-b (1%)
Ir/H-b (1%)
Ir/H-b (1%)
Ir/H-b (1%)
Ir/H-Mor (1%) 2.0
Ir/H-USY (1%) 2.0
Ir/H-b (1%)
Ir/H-b (2%)
Ir/H-b (2%)
2
2.0
2.0
2.0
9
0.75 87
1 75
[a] Conditions: testosterone (50 mg), solvent (6.5 g), catalyst (25 mg),
258C. Catalysts were calcined at 3008C and reduced at 4508C. USY=
ultra-stable Y zeolite, iAm=isoamyl.
decrease in reaction times and was also
beneficial in terms of the selectivities
observed, with 76% selectivity at 87% con-
version (Table 2, entries 9–10). When isoamyl
alcohol was used as the solvent, the selectiv-
ities could be further improved, for example,
to 100% chemoselectivity at 30% enone
5334
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 5333 –5336

Angewandte
Chemie
than in iPrOH (Table 4, entries 1 and 2). When 5 was used as
the starting enone, the high selectivity of the reaction (80%)
can be maintained up to very high conversion (95%), which
allylic alcohol is obtained at high conversion (Table 5,
entries 1 and 3). Along with a-ionol (6a), some b-ionol (7a)
is also formed, probably by isomerization to b-ionone,
followed by reduction. With the Na zeolite, observed selec-
tivities for the allylic alcohol were lower, thus providing
further evidence that proton acidity is important for the
chemoselectivity of the hydrogenation (Table 5, entry 2).
Control experiments with a-ionol showed that further hydro-
genation of a-ionol is very slow, which is consistent with the
high chemoselectivity observed for the allylic alcohol. b-
Ionone (7), in which the carbonyl group is conjugated to two
double bonds, reacts much faster than a-ionone (6), and a
selectivity of 65% for the allylic alcohol b-ionol was observed
with Ir/H-b (2%) (Table 5, entry 8). Selectivities were lower
or even much lower in the reactions of both 6 and 7 when
commercial catalysts such as Ir/C, transition-metal-doped Ir/
C (5%, Degussa), and Ir/CaCO3 were used, as well as with Ir/
Al2O3 and Ir/TiO2 (Table 5, entries 4–7 and 9–13). These data
unequivocally demonstrate the essential role of H-b as a
selectivity-promoting support.
Although the exact optimum conditions, for example, of
pressure, solvent, and Ir content, vary from one substrate to
another, it is very clear that the Ir/H-b (1–3 wt%) catalyst has
unprecedented generality for the chemoselective reduction of
unsaturated ketones and aldehydes. Combined CO-chemi-
sorption/TEM/TPR studies have shown that at low Ir content
(0.5–1 wt%), the Ir is finely distributed over the support, with
hardly any TEM (transmittance electron microscopy)-detect-
able metal particles, and with Ir dispersions of
Table 4: Hydrogenation of prostaglandin intermediates 4 and 5 over Ir/
H-b (1%).[a]
Entry Substrate Solvent P [MPa] t [min] Conv. [%] Sel. [%]
1
2
4
4
5
5
iPrOH 1.0
MeOH 1.0
MeOH 0.2
MeOH 0.2
60
120
60
43
59
95
50
46
71
80
55
3
4[b]
15
[a] Conditions: 4 (18.6 mg) or 5 (14 mg), solvent (7 g), Ir/H-b (1%;
25 mg), 258C. Catalyst was calcined at 3008C and reduced at 4508C.
[b] H-b precalcined at 7008C prior to loading with Ir.
implies that the corresponding unsaturated alcohol is a poor
substrate for the Ir-catalyzed hydrogenation (Table 4,
entry 3). Precalcination of the support at 7008C to convert
part of the Brønsted acidity into Lewis acidity led to
decreased chemoselectivity in the formation of the allylic
alcohol, thus showing that the zeolite protons are important
for the chemoselectivity of the catalyst (Table 4, entry 4). In
contrast with the reactions of the steroids 1 and 2, the
reduction of the prostaglandin enones resulted in significant
but low diastereoselectivities, typically about d.r 60:40.
Finally, as representatives of smaller unsaturated ketones,
a-ionone (6) and b-ionone (7) were subjected to the Ir/H-b-
up to 30%. The catalysts with a higher Ir
content, which were more effective for many
substrates, contain Ir0 particles in the range of
2–10 nm, thus showing that extracrystalline
Ir0 particles play an important role. It should
be stressed that the enones 1–7 are unable to
enter the intracrystalline voids of zeolites with
catalyzed reduction (Table 5). As anticipated based on the
experiments with citral, hydrogenation of the isolated double
bond in 6 is only a minor reaction, and 71% selectivity for the
*BEA topology.[15] Based on TPR (temperature-programmed
reduction, it appears that most of the Ir is in the metallic state.
Reduction at high temperature is in any case required to
obtain chemoselective catalysts.
In conclusion, by combining the carbonyl affinity of
metallic iridium with the promotion effect of the H-b zeolite,
which is a strong Brønsted acid, one can reduce a variety of
a,b-unsaturated aldehydes and ketones to allylic alcohols
effectively. Excellent conversions were paired with high
selectivities. Thus, this Ir/H-b catalyst constitutes a break-
through in selective reduction on heterogeneous catalysts,
especially as many of the substrates used are highly relevant
for fine-chemicals production.[14]
Table 5: Hydrogenation of 6 and 7 over Ir/H-b catalysts.[a]
Entry Substrate Catalyst
P [MPa] t [h] Conv. [%] Sel. [%][b]
1
2
3
4
5
6
6
6
6
6
6
6
6
7
7
7
Ir/H-b (2%)
Ir/Na-b (2%)
Ir/H-b (3%)
Ir/C (1%)
Ir/CaCO3 (5%)
Ir/Al2O3 (1%) 2.4
Ir/TiO2 (1%)
Ir/H-b (2%)
Ir/C (1%)
Ir/C (5%),
doped
3.6
3.6
2.8
4.4
2
11
5.8 69
22.2 92
19
2.3 75
4
48
1.5 84
75
71
45
70
50
15
12
47
65
7
97
98
90
7[c]
8
9
10
4.4
3.6
4.4
3.6
2
18
73
94
Experimental Section
10
The catalyst was prepared by impregnation of zeolite H-b with
Ir(acac)3 as a solution in toluene (1.3 mm; 2 mLgÀ1 zeolite). The
catalyst was dried at 608C for 2 h, after which it was calcined at 3008C
for 4 h. The reduction of the unsaturated ketone or aldehyde in the
presence of the catalyst (25 mg) was performed in the solvent
indicated (6.5 g) under a stream of H2 at 4508C for 6 h. Commercial
catalysts were purchased from Alfa Aesar (Ir/C (1%), Ir/
CaCO3 (5%)) or were obtained as a generous gift from Degussa
11
12
13
7
7
7
Ir/CaCO3 (5%)
Ir/Al2O3 (1%) 2.4
Ir/TiO2 (1%) 4.4
2
0.8 70
2
4
14
4
86
77
23
[a] Conditions: substrate (70 mg), isopropanol (6.5 g), 258C, catalyst
(25 mg). Catalysts calcined at 3008C and reduced at 4508C (except in
entries 4, 9, and 10). [b] Selectivity for ionols. [c] Catalyst: 50 mg,
contains water (50%).
Angew. Chem. Int. Ed. 2003, 42, 5333 –5336
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5335

Communications
(transition-metal-doped Ir/C (5%)). The reactions were performed at
room temperature. Analysis of the reaction and products was
performed by GC(MS), LC, and 1H and 13C NMR.
Received: July 2, 2003 [Z52275]
Keywords: allylic alcohols · heterogeneous catalysis ·
.
hydrogenation · iridium · zeolites
[1]K. Bauer, D. Garbe, H. Surburg, Common Fragrance and Flavor
Materials, Wiley-VCH, Weinheim, 1997.
[2] Ullmann's Encyclopedia of Technical Chemistry, 6th edition,
electronic release, Wiley-VCH, 2000.
[3]P. Gallezot in Handbook of Heterogeneous Catalysis, Vol. 5
(Eds.: G. Ertl, H. Knꢀzinger, J. Weitkamp), VCH, Weinheim,
1997, pp. 2209 – 2221.
[4]P. Reyes, M. C. Aguirre, I. Melian-Cabrera, M. Lopez-Granados,
J. L. G. Fierro, J. Catal. 2002, 208, 229.
[5]D. G. Blackmond, R. Oukaci, B. Blanc, P. Gallezot, J. Catal.
1991, 131, 401.
[6]J.-L. Luche, L. Rodriguez-Hahn, P. Crabbꢁ, J. Chem. Soc. Chem.
Commun. 1978, 601.
[7]T. Ohkuma, H. Ikehira, T. Ikariya, R. Noyori, Synlett 1997, 467.
[8]M. von Arx, T. Mallat, A. Baiker, J. Mol. Catal. A 1999, 148, 275.
[9]C. Milone, R. Ingoglia, M. L. Tropeano, G. Neri, S. Galvagno,
Chem. Commun. 2003, 868.
[10]A. Corma, A. Martinez, V. Martinez-Soria, J. Catal. 1997, 169,
480.
[11]R. C. Cambie, P. S. Rutledge, D. W. Scott, P. D. Woodgate, Aust.
J. Chem. 1979, 32, 695.
[12]A. S. Campos Neves, M. L. Melo, M. J. Moreno, E. J. Tavares
da Silva, J. A. R. Salvador, S. P. da Costa, R. M. Martins, Tetra-
hedron 1999, 55, 3255.
[13]W. Pickenhagen, D. Schatkowski, DE 10038544, 2002 [Chem.
Abstr. 2002, 136, 134933].
[14]For the relevance of allylic alcohols derived from a) steroidal
enones, see: S. Ring, G. Teichmꢂller, G. Weber, S. Schwarz, B.
Erhart, B. Undeutsch, H. Raethe, P. Mꢀllmann, C. Pfeiffer, H.-J.
Palme, US 5831104, 1998 [Chem. Abstr. 1995, 123, 33509]; G.
Humphrey, R. A. Miller, US 5696266, 1997 [Chem. Abstr. 1998,
128, 75585]; b) isolongifolenone, see reference [13]; c) prosta-
glandin precursors, see: M. Ohtani, T. Matsuura, Y. Hamada, I.
Hamada, T. Sakata, K. Takahashi, M. Kishi, US 5585501, 1996
[Chem. Abstr. 1993, 119, 249777]; A. Gutman, G. Nisnevich, M.
Etinger, I. Zaltzman, L. Yudovich, B. Pertsikov, WO PCT 01/
55101, 2001 [Chem. Abstr. 2001, 135, 152659]; d) ionones, see: W.
Kuhn, H. Surburg, US Appl. 2002/0082457, 2002 [Chem. Abstr.
2002, 137, 33431].
[15]http://www.iza-structure.org/
5336
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 5333 –5336