Pt/H-beta zeolites as productive bifunctional catalysts for the one-step citronellal-to-menthol conversion

Article abstract of DOI:10.1016/j.jcat.2006.06.017

Pt-loaded H-beta zeolite was identified as a highly active catalyst for the bifunctional transformation of citronellal to menthol, with isopulegol as the intermediate. With a 2 wt% Pt-loaded catalyst, citronellal is fully converted within 12 h, with only 2.5 wt% catalyst with respect to citronellal, and with a citronellal to Pt molar ratio of 2500. 1, 4-Dioxane is the best reaction solvent, because it minimizes the unwanted direct hydrogenation of citronellal and promotes its stereoselective cyclization to isopulegol, leading to high menthol yields. The stereoselectivity can be improved moderately by using a Zr-impregnated support and more substantially by performing high-temperature (750 °C) treatment of the calcined and reduced catalyst. This treatment presumably creates extra Lewis acidity on the catalyst and results in 88% stereoselectivity for the desired menthol. Overall, an 85% yield of (-)-menthol was obtained.

Full text of DOI:10.1016/j.jcat.2006.06.017

Journal of Catalysis 243 (2006) 7–13
www.elsevier.com/locate/jcat
Pt/H-beta zeolites as productive bifunctional catalysts for the one-step
citronellal-to-menthol conversion
Pascal Mertens a, Francis Verpoort b, Andrei-Nicolae Parvulescu a, Dirk De Vos a,∗
a
Centre for Surface Chemistry and Catalysis, K.U. Leuven, Kasteelpark Arenberg 23, 3001 Heverlee, Belgium
b
Department of Inorganic and Physical Chemistry, U. Gent, Krijgslaan 281, 9000 Gent, Belgium
Received 24 May 2006; revised 19 June 2006; accepted 21 June 2006
Available online 8 August 2006
Abstract
Pt-loaded H-beta zeolite was identified as a highly active catalyst for the bifunctional transformation of citronellal to menthol, with isopulegol
as the intermediate. With a 2 wt% Pt-loaded catalyst, citronellal is fully converted within 12 h, with only 2.5 wt% catalyst with respect to
citronellal, and with a citronellal to Pt molar ratio of 2500. 1,4-Dioxane is the best reaction solvent, because it minimizes the unwanted direct
hydrogenation of citronellal and promotes its stereoselective cyclization to isopulegol, leading to high menthol yields. The stereoselectivity can
◦
be improved moderately by using a Zr-impregnated support and more substantially by performing high-temperature (750 C) treatment of the
calcined and reduced catalyst. This treatment presumably creates extra Lewis acidity on the catalyst and results in 88% stereoselectivity for the
desired menthol. Overall, an 85% yield of (–)-menthol was obtained.
© 2006 Elsevier Inc. All rights reserved.
Keywords: Bifunctional catalyst; Chemoselective hydrogenation; Pt/H-beta zeolite; Citronellal; Menthol; Citral
1. Introduction
menthol stereoisomers, of which only product 3a, menthol, is
profitable. Less valuable and therefore undesirable isomers are
neomenthol (3b), isomenthol (3c), and neoisomenthol (3d).
Transformation of citronellal to menthol in a one-pot process
may be an attractive option if a good performing and selective
bifunctional catalyst is available. Acid functions on the material
should catalyze the cyclization, while a metal function performs
the hydrogenation. Previous workers have used Cu as the ac-
tive metal on a silica support [3] or metallic Ru and ionic Zn
immobilized together on SiO2 [4,5]. Much of this work has fo-
cused on obtaining a high summed yield of menthol 3a and its
stereoisomers 3b and 3c, or on maximizing the stereoselectiv-
ity, that is, the share of the desired isomer 3a in the total of
3a + 3b + 3c. In practice, hardly any 3d is formed. In these pre-
vious studies, the amount of citronellal converted is quite low,
however (ꢀ1 g of substrate per g of catalyst). Similarly low sub-
strate/catalyst ratios were applied in recent work on the one-pot
transformation of citral 7 to menthol [6,7]. The latter reaction
comprises an additional challenge, because in a supplementary
first step, citral needs to be hydrogenated in a chemoselective
(–)-Menthol is a fine chemical compound often used as a fra-
grance or flavor in the production of pharmaceuticals, food
products, and cosmetics. Established routes for menthol pro-
duction are multistep processes, such as the Haarmann and
Reimer process starting from thymol [1] and the Takasago
process starting from myrcene [2]. The synthesis of menthol
starting from citronellal comprises two consecutive steps: acid-
catalyzed isomerization of citronellal to the cyclic isopulegol
intermediate and metal-catalyzed hydrogenation of the double
bond in isopulegol, leading to menthol. Along with one stere-
ogenic carbon atom, citronellal (1) has two prochiral carbon
atoms that are converted to asymmetric centers in the pulegol
intermediate (Scheme 1). Consequently, four different pulegol
stereoisomers (2a–2d) can be formed from one citronellal enan-
tiomer. The subsequent hydrogenation results in four different
*
Corresponding author. Fax: +32 16 32 1998.
E-mail address: dirk.devos@biw.kuleuven.be (D. De Vos).
0021-9517/$ – see front matter © 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2006.06.017

8
P. Mertens et al. / Journal of Catalysis 243 (2006) 7–13
resulting in 5 or 10 wt% of Zr on the solid. The modified H-
beta support was dried for 4 h at 90 ◦C and then subjected to all
further steps of the impregnation with 2 wt% Pt and the standard
postsynthesis treatment.
The Pt dispersion (DPt) was determined by CO chemisorp-
tion via titration with 5-µl CO pulses and mass spectrometric
detection. Before the measurement, the samples were treated
with H2 at 350 ◦C, with a heating rate of 1 ◦C/min; afterwards,
the samples were cooled to room temperature and maintained
there under He atmosphere. By determining the amount of
chemisorbed CO, and assuming an adsorption of 1 CO mole-
cule per accessible Pt atom, DPt values were obtained. The size
of the Pt particles, dPt,C, was derived from the Pt dispersion
with the assumption of a cubic Pt particle shape according to
White [8]: dPt,C = 0.821(1/DPt) (nm).
For TEM characterization, the samples were mounted on
a copper grid-supported carbon film by depositing droplets of
an ultrasonically dispersed suspension of the powder catalyst
in ethanol on the grid, with subsequent drying at ambient con-
ditions. The TEM images were recorded on a Philips CM20
electron microscope operating at 200 kV. The average size of
the Pt particles, dPt,T, was obtained by calculating the number
average diameter for an ensemble of 50 Pt particles observed in
TEM.
Pt presence represents the molar amount of Pt on 1 g of the
Pt/H-beta catalyst, whereas Pt availability represents the molar
amount of accessible surface Pt on 1 g of Pt/H-beta catalyst.
The Pt availability was calculated by multiplying the Pt pres-
ence value by the dispersion of the Pt/H-beta catalyst.
Scheme 1.
way to citronellal. Nickel on mildly acid, Al-containing MCM-
41 emerged from these studies as the most promising catalyst,
even if decarbonylation of the citral is a nonnegligible side re-
action. In any case, citronellal is somewhat less expensive than
citral, which makes the reaction starting from citral of limited
economic significance.
The present work focuses on identifying highly active and
stereoselective catalysts for the citronellal to menthol conver-
sion, with a rapid and selective acid-catalyzed isomerization to
isopulegol 2a followed by a metal-catalyzed hydrogenation to
menthol 3a. Overall, this should result in increased catalyst pro-
ductivity and menthol stereoselectivity.
2.2. Catalytic experiments and analysis
A standard reaction mixture contained 0.6 ml of solvent,
0.4 ml of citronellal, and 0.01 g of the bifunctional powder cat-
alyst. The stirred mixture was then pressurized to 2.0 MPa H2
and heated to 80 ◦C. All reactions were performed in a multire-
actor with 10 minireactors (4 ml volume) from TOP Industries.
The pressure in all vessels was kept constant by a continuous
supply of reagent gas. Samples were analyzed by gas chro-
matography (CP-Wax 58 column, flame ionization detector),
and retention times were compared with those of commercial
reference compounds.
The selectivity to menthols (Smenthols) is the fraction of men-
thol and its stereoisomers (3) in the sum of all reaction products
in the reaction mixture. A low value of Smenthols corresponds to
a high proportion of 3,7-dimethyl-1-octanol (6), pulegols (2),
etc. in the total of reaction products. The isomer stereoselec-
tivity (Sstereo) is the portion of menthol (3a) in the pool of
menthol stereoisomers (3). Yieldmenthol is the yield of 3a, and
Yieldmenthols is the summed yield of 3a + 3b + 3c.
2. Experimental
2.1. Catalyst preparation and characterization
Pt/H-beta catalysts were prepared by incipient wetness im-
pregnation of 1 g of calcined, commercial H-beta zeolite
(Si/Al = 14, Süd Chemie) with 2 ml of an aqueous K2PtCl4
solution. The concentration of the Pt precursor in the impreg-
nation solution depended on the desired metal loading of the
zeolite support (1, 2, 3, or 5 wt%). Next, the suspension was
dried for 4 h at 90 ◦C, and the dry powder was successively
calcined at 300 ◦C for 4 h under flowing O2 (heating rate
2 ◦C/min; C300) and reduced at 350 ◦C for 6 h under flow-
ing H2 (heating rate 2 ◦C/min; R350).
In variants of this standard procedure, different H-beta zeo-
lites were used, with Si/Al ratios of 12.5, 14, and 15 (all from
Süd Chemie). Alternatively, the standard catalyst preparation
was followed by drying at 750 ◦C in flowing N2 for 6 h.
For the preparation of 2 wt% Pt/Zr-H-beta catalysts, 0.176
or 0.353 g of ZrOCl2·8H2O was dissolved in 2 ml of water, and
this solution was impregnated on 1 g of H-beta (Si/Al = 14),
3. Results and discussion
Although the isomerization of citronellal to isopulegol has
often been conducted on a soluble Lewis acid such as Zn2+
,
Sc3+, or Al3+ [9–11], it is clear that the reaction proceeds at
equally high rates on H-beta zeolite, which may contain both

P. Mertens et al. / Journal of Catalysis 243 (2006) 7–13
9
Brönsted and Lewis acid sites. Compared with H-beta, using
a different acid molecular sieve, like H-mordenite, H-MCM-22,
or H-MCM-41, results in substantial side product forma-
tion [12]. In previous work, we combined H-beta with metallic
Ir for hydrogenation [13]. Because Pt is generally a more active
and selective catalyst for C=C hydrogenation than Ir, one may
wonder whether conditions can be found in which Pt/H-beta
selectively transforms citronellal to menthol. Other metals like
Ni and Pd also may be suitable for hydrogenation of the double
bond in isopulegol, but their preference for C=C hydrogenation
is so high that in practice it is difficult to avoid the direct hy-
drogenation of citronellal to saturated, noncyclized compounds
such as 6. In attempts to obtain high catalyst productivity cou-
pled with high menthol stereoselectivity, significant variables
may include the method of catalyst preparation, the Pt content
of the catalyst, the reaction solvent, the H2 pressure, the re-
action temperature, time, and others. Because these variables
clearly form a multivariable space, we decided to use a high-
throughput catalytic reactor, with 10 parallel reaction positions,
to determine the optimal catalyst composition and operation
conditions.
chemisorbed hydrogen on the Pt surface. This hydrogen was
then used, in a hydrogen transfer reaction, for reduction of pule-
gols to the corresponding menthols. Careful chromatographic
analysis showed that in the reaction of citronellal over Pt/H-beta
under 2.0 MPa N2 atmosphere in cyclohexane, a small amount
of cyclohexene was indeed formed. This hydrogen transfer re-
action was so slow that no products of direct C=C or C=O hy-
drogenation of citronellal were formed; the cyclization clearly
preceded the reduction reaction. As can be expected, the iso-
mer distribution in the pulegols determined the ratio of the
different menthols, resulting in a selectivity for menthol 3a
(Sstereo = 3a/(3a + 3b + 3c)) of around 75%.
3.2. Transformation of citronellal under H2 atmosphere
Preliminary experiments on citronellal transformation under
H2 atmosphere over Pt/H-beta showed that besides pulegol (2)
and menthol (3) isomers, only 3,7-dimethyl-1-octanol (6) was
eventually obtained in significant amounts as a byproduct. In
the direct hydrogenation of citronellal over Pt/H-beta, initially
citronellol (5) rather than 3,7-dimethyloctanal (4) was formed.
This reflects that the trisubstituted double bond in citronellal
was hydrogenated relatively slowly; for rapid hydrogenation
of sterically hindered double bonds on Pt, often a liquid acid
(e.g., CH3COOH or HClO4) must be added to the reaction
solvent [14]. Subsequently, citronellol was converted into 3,7-
dimethyl-1-octanol. Thus the selectivity was determined by the
competition between the ene reaction (1 → 2) and the direct
hydrogenation of citronellal (1 → 4, 5, 6). Obviously, the re-
duction was much faster under H2 atmosphere than under N2,
and products of solvent dehydrogenation were not observed.
3.1. Transformation of citronellal under inert N2 atmosphere
First, the citronellal cyclization to the isopulegol interme-
diate was examined using various Pt/H-beta catalysts, in the
absence of H2 (Table 1). Particularly in nonpolar solvents, iso-
merization proceeded smoothly. An extensive solvent screening
indicated that 1,4-dioxane is an excellent alternative for cyclo-
hexane, which has been used in previous work for the reaction
of citronellal to menthol [13]. Table 1 also shows that the iso-
merization activity of the catalyst increased with Pt loading
(Table 1, entries 3–6). Catalyst pretreatment had a relatively
minor effect on isomerization activity; for instance, practically
identical results are obtained when catalysts are not reduced or
are reduced at 150 or 300 ◦C (data not shown).
In these isomerization experiments in cyclohexane or 1,4-
dioxane, only cyclization products were formed from citronel-
lal. The main products were the desired isopulegol (2a) and
its stereoisomers neoisopulegol (2b) and isoisopulegol (2c), to-
gether with minor amounts of menthols. Typically, the fraction
of isopulegol 2a in the total of pulegol isomers (2a + 2b + 2c)
amounted to 75%. Because some menthols were found in re-
actions under N2 over Pt catalysts but not in the reaction with
H-beta as such, it seems that solvent dehydrogenation provided
3.2.1. Effect of solvent
The most important factor influencing the reaction selectiv-
ity is the solvent applied. Therefore, a broad solvent screening
was performed using 2 wt% Pt/H-beta, considering both polar
and nonpolar solvents (Table 2, entries 1–5). In 1,4-dioxane,
formation of 3,7-dimethyloctanol was minimal, and the highest
menthol yields were obtained. The advantage of using 1,4-
dioxane instead of aliphatic or aromatic hydrocarbon solvents
was quite significant. Polar solvents, such as alcohols, resulted
in relatively high selectivities for products of direct hydrogena-
tion of citronellal. It is indeed known that alcohols are preferred
solvents for the hydrogenation of aldehydes on platinum group
metals [14]. In addition, alcohol solvents cause formation of
Table 1
Yield of pulegols (2) and menthols (3) from citronellal (1) under N atmosphere
2
No.
Catalyst
Solvent
Conversion
(%)
Yield
(%)
Yield
(%)
citronellal
pulegols
menthols
◦
◦
◦
◦
◦
◦
at 40 C
at 80 C
at 40 C
at 80 C
at 40 C
at 80 C
1
2
3
4
5
6
2% Pt/H-beta
2% Pt/H-beta
H-beta
1% Pt/H-beta
2% Pt/H-beta
3% Pt/H-beta
Cyclohexane
1,4-Dioxane
1,4-Dioxane
1,4-Dioxane
1,4-Dioxane
1,4-Dioxane
90.2
94.9
66.8
82.3
95.1
97.9
94.2
98.7
74.5
91.1
98.8
99.9
89.8
94.4
66.8
82.2
94.5
96.5
92.9
96.8
74.5
90.4
97.2
98.1
0.4
0.5
0.0
0.1
0.6
1.4
1.3
1.9
0.0
0.7
1.6
1.8
Reaction conditions: 0.4 g citronellal, 0.01 g catalyst, 1 ml solvent, 2.0 MPa N , 6 h.
2

10
P. Mertens et al. / Journal of Catalysis 243 (2006) 7–13
Table 2
Effect of solvent and Pt loading on the citronellal (1) transformation to menthol (3a)
No.
Catalyst
Solvent
Yield
(%)
Yield
(%)
Yield
(%)
Yield
(%)
S
stereo
(%)
3,7-dimethyloctanol
pulegols
menthols
menthol
1
2
3
4
5
6
7
8
9
2% Pt/H-beta
2% Pt/H-beta
2% Pt/H-beta
2% Pt/H-beta
2% Pt/H-beta
1% Pt/H-beta
2% Pt/H-beta
3% Pt/H-beta
5% Pt/H-beta
Tetrahydrofuran
Toluene
44.6
39.8
34.7
22.9
5.1
2.7
4.6
16.9
39.7
14.1
6.4
9.8
5.4
2.1
16.7
2.0
5.8
6.4
41.3
53.8
55.5
71.7
92.8
80.6
93.4
77.3
53.9
31.1
39.4
40.3
54.6
70.5
61.3
71.3
59.2
40.9
75.4
73.2
72.6
76.2
76.0
76.0
76.3
76.6
75.9
para-Xylene
Cyclohexane
1,4-Dioxane
1,4-Dioxane
1,4-Dioxane
1,4-Dioxane
1,4-Dioxane
◦
Reaction conditions: 0.4 g citronellal, 0.01 g catalyst, 1 ml solvent, 80 C, 2.0 MPa H , conversion >99%, 12 h.
2
Table 3
Effect of temperature and H pressure on the citronellal (1) transformation
2
No.
Temperature
( C)
H
(MPa)
pressure
Yield
(%)
Yield
(%)
Yield
(%)
Yield
(%)
2
3,7-dimethyloctanol
pulegols
menthols
menthol
◦
1
2
3
4
5
6
7
8
25
50
75
2.0
2.0
2.0
2.0
0.2
0.5
1.0
2.0
2.3
3.8
4.2
4.9
0.4
1.9
2.8
4.4
16.9
11.8
3.4
2.2
5.8
4.7
4.3
2.9
80.8
84.4
92.4
92.9
93.8
93.4
92.9
92.7
61.2
64.6
69.5
70.6
70.7
70.8
70.0
69.7
100
80
80
80
80
Reaction conditions: 0.4 g citronellal, 0.01 g 2% Pt/H-beta, 1 ml 1,4-dioxane, conversion >99%, 12 h.
Table 4
Comparison between two reaction procedures in the transformation of citronellal (1)
No.
Procedure
Solvent
Yield
(%)
Yield
(%)
Yield
(%)
Yield
(%)
S
stereo
(%)
3,7-dimethyloctanol
pulegols
menthols
menthol
1
2
3
4
1
1
2
2
Cyclohexane
1,4-Dioxane
Cyclohexane
1,4-Dioxane
0.0
0.0
23.2
4.2
1.6
2.9
4.6
2.2
98.4
97.1
72.2
93.6
76.1
75.9
55.3
72.2
77.3
78.2
76.6
77.1
Reaction conditions: procedure 1: 6 h 2.0 MPa N , 12 h 2.0 MPa H ; procedure 2: 12 h 2.0 MPa H ; 0.4 g citronellal, 0.01 g 2% Pt/H-beta catalyst, 1 ml solvent,
2
2
2
◦
80 C, conversion >99%.
such side products as (hemi-)acetals, especially when small
primary alcohols are used as solvents in the presence of acid
catalysts. But not all cyclic ethers are equally good solvents, as
demonstrated by the low yield of menthols in tetrahydrofuran
(entry 1). Polarity does not seem to be the key variable affect-
ing the selectivity, because cyclohexane and 1,4-dioxane have
similar dielectric constants, ε; however, these solvents’ hydro-
gen bond-accepting properties are obviously different.
not yet complete after 12 h. Therefore, 2 wt% Pt/H-beta seems
to be an optimal loading. Although this Pt loading might seem
relatively high, note that only 2.5 wt% catalyst was used, corre-
sponding to an effective citronellal:Pt molar ratio of 2500:1.
3.2.3. Effect of temperature and H2 pressure
Variation of the reaction temperature shows that at least
75 ◦C was required for a sufficient hydrogenation rate of the
pulegol isomers; at lower temperatures, a sizable amount of
pulegols was recovered after 12 h of reaction (Table 3, en-
tries 1–4). A temperature increase not only raised the menthol
yields, but also produced a small, but significant, increase in
the amount of direct citronellal hydrogenation, from 2.3% at
25 ◦C to 4.9% at 100 ◦C. These observations suggest that the
Pt-catalyzed hydrogenation of either citronellal or isopulegols
had greater apparent activation energy than cyclization. Selec-
tivity differences between reactions at 75 and 100 ◦C were mar-
ginal (Table 4, entries 3 and 4). Although an optimal hydrogen
pressure might also be expected, the effects of pressure on the
3.2.2. Effect of Pt loading
After selection of an optimal solvent, Pt loading is the next
most significant variable. As shown in Table 1, Pt not only func-
tions as a reduction catalyst; but also accelerates cyclization.
Therefore, Pt content should be optimized to provide the proper
relative rates of both sequential reactions (Table 2, entries 6–9).
Raising the Pt loading above 2 wt% resulted in too much direct
citronellal hydrogenation, with 3,7-dimethyl-1-octanol yields
exceeding 15%. This can be avoided using 1 wt% Pt/H-beta;
but in this case, the hydrogenation of the pulegol isomers was

P. Mertens et al. / Journal of Catalysis 243 (2006) 7–13
11
menthol isomer yield was small; very little direct hydrogena-
tion was observed at 0.2 MPa H2, but this effect was balanced
by the less effective hydrogenation of pulegols at this pressure.
with 5 wt% Zr and a Si/Al ratio of 14 (Table 5, entries 1–3).
More substantial improvements were realized by heating the
calcined and reduced catalyst to 750 ◦C. This treatment prob-
ably converts part of the Brönsted sites to Lewis sites [17],
which may affect the isopulegol:neoisopulegol:isoisopulegol
(2a:2b:2c) ratio in the cyclization. After this treatment, the H-
beta zeolite with Si/Al ratio 14 emerged as the most appropriate
isomerization catalyst, with a Sstereo of 88.0%, in comparison to
the 78.9% obtained with the catalyst as such (Table 5, entry 7).
For other H-beta catalysts from several batches, with slightly
different Si/Al ratios, high-temperature treatment also proved
beneficial to increasing the stereoselectivity and menthol yield
(Table 5, entries 5 and 9).
3.2.4. Comparison of reaction procedures
To maximize the citronellal transformation to menthol, two
different reaction procedures were compared. In procedure 1,
the reaction was conducted for 6 h under N2 atmosphere and
subsequently for 12 h under H2; alternatively, in procedure 2,
the H2 atmosphere was maintained for the full reaction duration
of 12 h. The first procedure yielded 97–98% of menthol isomers
whether performed in 1,4-dioxane or in cyclohexane. This cor-
responds to menthol yields of >75% (Table 4, entries 1 and 2).
Clearly, the absence of H2 prevents the unfavorable initial hy-
drogenation of citronellal normally observed in cyclohexane
(Table 4, entry 3). In comparison, the best yield of menthol 3a
under continuous H2 pressure and using the standard 2 wt%
Pt/H-beta catalyst was 72% (Table 4, entry 4).
3.3. Transformation of citral under H2 atmosphere
In a final series of experiments, the optimized 2 wt% Pt/H-
beta catalyst was applied in menthol synthesis starting from
citral. Table 6 shows that in identical conditions, the Smenthols
was considerably lower starting from citral than from citronel-
lal (entries 1 and 2). In this three-step menthol formation, cit-
ral should initially be converted to citronellal. Along with cit-
ronellal, the doubly unsaturated alcohol isomers geraniol and
nerol were also formed initially; 3,7-dimethyl-2-octenal was
not observed. The competition between these initial hydro-
genation steps partially accounts for the lower chemoselectivity
compared with the menthol synthesis starting from citronellal.
Moreover, both citronellal and geraniol or nerol can be consec-
utively hydrogenated, and, as before, this direct citronellal hy-
drogenation to citronellol instead of its cyclization to isopulegol
was an additional cause of chemoselectivity loss. Again, only
3,7-dimethyl-1-octanol was observed as a cyclic hydrogenation
product at the end of the reaction. Ni and Pd are probably more
3.2.5. Modification of reaction stereoselectivity
Given the very high yields of menthol isomers from cit-
ronellal, and with a highly productive catalyst in hand, the only
aspect to be improved is the stereoselectivity to 2a in cycliza-
tion, which eventually determines the stereoselectivity for 3a.
Two approaches were followed. First, Zr was deposited on the
beta zeolite before introduction of the Pt [15,16]; second, the
catalyst was subjected to an extra pretreatment at high tem-
perature [17]. Introduction of additional Lewis acid atoms in
the structure was recently proven successful by Corma and
Renz [18], who reported a 85% stereoselectivity to isopulegol
over a Sn-substituted beta zeolite. With the Zr-loaded Pt/H-
beta catalyst, only modest increases in stereoselectivity were
obtained, for example, from 78 to 80% for a sample loaded
Table 5
Menthol (3a) yield and stereoselectivity as a function of support pre- and post-treatments
No.
Support
Yield
(%)
Yield
(%)
S
stereo
(%)
menthols
menthol
1
2
3
4
5
6
7
8
9
H-beta (Si/Al ≈ 14)
97.2
99.5
99.5
95.1
95.0
98.3
96.2
99.2
98.9
76.1
80.1
78.8
71.3
78.2
77.6
84.7
74.2
79.8
78.3
80.2
78.9
75.0
82.3
78.9
88.0
74.8
80.7
5 wt% Zr/H-beta (Si/Al ≈ 14)
10 wt% Zr/H-beta (Si/Al ≈ 14)
H-beta (Si/Al = 12, 5)
◦
H-beta (Si/Al = 12, 5) 750 C
H-beta (Si/Al = 14)
◦
H-beta (Si/Al = 14) 750 C
H-beta (Si/Al = 15)
◦
H-beta (Si/Al = 15) 750 C
◦
Reaction conditions: 0.4 g citronellal, 0.01 g 2% Pt/support, 1 ml 1,4-dioxane, 25 C, 6 h 2.0 MPa N , 12 h 2.0 MPa H , conversion >99%.
2
2
Table 6
Comparison between transformation of citronellal (1) and citral (7) to menthol (3a)
No.
Solvent
Substrate
Yield
(%)
Yield
(%)
Yield
(%)
Yield
(%)
3,7-dimethyloctanol
pulegols
menthols
menthol
1
2
3
4
Cyclohexane
Cyclohexane
1,4-Dioxane
1,4-Dioxane
Citronellal
Citral
Citronellal
Citral
23.8
39.8
3.8
6.5
6.2
2.4
1.2
69.7
54.0
93.8
81.1
53.3
41.0
72.0
62.5
17.7
◦
Reaction conditions: 0.4 g citronellal or citral, 0.01 g 2% Pt/H-beta catalyst, 1 ml solvent, 80 C, 2.0 MPa H , conversion >99%, 12 h.
2

12
P. Mertens et al. / Journal of Catalysis 243 (2006) 7–13
Table 7
Platinum dispersion, Pt particle size and Pt availability of the various Pt/H-beta catalysts
No.
Pt loading
(wt%)
D
d
d
Pt presence
(10 mol/g)
Pt availability
(10 mol/g)
Pt
Pt,T
Pt,C
−6
−6
(%)
(nm)
(nm)
1
2
3
4
1
2
3
5
12.25
7.25
5.95
5.56
6.7
11.3
13.8
14.8
4.9
6.7
8.5
51.3
102.5
153.8
256.3
6.3
7.5
9.2
12.0
14.2
Fig. 1. Transmission electron micrographs of the various Pt/H-Beta catalysts.
appropriate for menthol synthesis starting from citral, because
they reduce more selectively the C=C bonds, which can result
in improved production of the citronellal intermediate [6].
even after the standard postsynthesis treatment. For the 2 wt%
Pt/H-beta, a dPt,T value of 6.7 nm was determined. TEM char-
acterization of a 2 wt% Pt/H-beta that had also been subjected
to the high-temperature treatment found that the dPt,T was not
affected, indicating that this postsynthesis treatment does not
affect the Pt particles. Moreover, calculations of Pt availability
for the Pt/H-beta catalysts (Table 7) show that a 5 times higher
Pt loading led to only a doubling of Pt availability. The rela-
tively low DPt value for the 5 wt% Pt/H-beta catalyst was due
to the larger diameters of the individual particles and also to
a more pronounced aggregation of the particles on the outer
surface of the zeolite, as shown in Fig. 1d.
3.4. Characterization of Pt/H-beta catalysts
Table 7 reports the values of DPt for the standard Pt/H-beta
catalysts. A decline in Pt dispersion was observed for catalysts
with higher Pt loadings. For the 2 wt% Pt/H-beta, 7.25% of the
Pt atoms present was accessible. Based on the Pt dispersion, the
size of the Pt particles was calculated, and these dPt,C values
were compared with the dPt,T values from the TEM character-
ization. TEM pictures of the Pt/H-beta catalysts with varying
loading are reproduced in Fig. 1. A similar trend of increased
Pt particle size at higher Pt loadings was obtained via both
techniques for particle diameter determination, as reflected in
Table 7. Because the dPt,T value was substantially lower than
the dPt,C value, it seems that a fraction of the Pt surface atoms
remained inaccessible or poisoned (by, e.g., chloride species),
4. General discussion
The bifunctional Pt/H-beta zeolite has proven to be an effi-
cient and highly selective catalyst for menthol synthesis from
citronellal, especially in 1,4-dioxane and with a Pt loading of
2 wt%. Selectivities to menthols exceeding 95% were achiev-

P. Mertens et al. / Journal of Catalysis 243 (2006) 7–13
13
able in the one-step transformation, combined with yields of the
desired menthol up to 76%.
the Lewis acid site is embedded in the pore wall, it may dis-
play unusual stereoselectivity. This has been well documented
for the Meerwein reduction of 4-t-butylcyclohexanone in zeo-
lite beta [23]; similar effects may be invoked here to explain the
increased stereoselectivity to isopulegol 2a. The postsynthesis
catalyst treatment eventually resulted in a menthol stereoselec-
tivity of 88% and a menthol yield of 85%.
Several factors seem to be essential in obtaining this re-
sult. First, the Pt discriminates quite well between disubstituted
C=C double bonds, as in pulegols, and trisubstituted bonds,
as in citronellal, which are only slowly hydrogenated. This is
reminiscent of, for example, the chemoselective hydrogenation
of limonene on Pt/C, in which only the exocyclic disubstituted
bond reacts [14]. A metal like Pd is much less suitable for real-
izing such chemoselectivity, because Pd facilitates double-bond
isomerization. Because the C=C bond in citronellal is hardly
hydrogenated on a 2 wt% Pt catalyst, there is sufficient time
for the ene-like cyclization to proceed. The pulegol isomeriza-
tion products 2a → 2d have a disubstituted double bond and
are much more prone to hydrogenation, leading eventually to
the desired menthol isomers 3.
Acknowledgments
P.M. thanks the IWT for financial support. This work was
performed in the frame of the Belgian Interuniversity Poles of
Attraction (IUAP V-3). D.D.V. is grateful to the FWO for grant
G.0323.04. The authors thank Xingpu Ye for assistance with
the TEM characterization of the catalysts.
References
Second, the catalyst also discriminates well between C=C
hydrogenation in the pulegols and the undesired C=O hydro-
genation in citronellal. This is essential to avoid citronellol
formation from citronellal. Indeed, Pt is known to be a suit-
able catalyst for C=O hydrogenation only when specific com-
pounds, such as Fe3+ or Sn4+, are used in the support ma-
terial or as a promotor [14,19]. Only when using the 5 wt%
Pt catalyst was an unexpectedly pronounced C=O hydrogena-
tion observed (Table 2, entry 9). Although this can be due in
part to the relative amounts of acid and hydrogenating func-
tions, one should also consider that the 5 wt% catalyst contains
the largest and most aggregated Pt particles. This seems to in-
dicate that large particles more readily hydrogenate the C=O
bonds. A similar enhanced selectivity for C=O hydrogenation
on large Pt metal particles (>6 nm) with flat surfaces in the
selective hydrogenation of cinnamaldehyde to cinnamyl alco-
hol has been reported by Gallezot et al. [20]. For the Pt/H-beta
catalysts with lower Pt content and smaller particle size, direct
C=O hydrogenation can be limited, which directs the overall
reaction toward cyclic products.
Third, choosing appropriate solvents helps obtain high se-
lectivities. For instance, alcohols are less suitable because they
are the typical solvents for selective C=O reduction in the pres-
ence of C=C bonds. In 1,4-dioxane, the rates of the consecutive
reactions seem well tuned to give high menthol yields.
Finally, a further stereoselectivity increase was obtained af-
ter high-temperature treatment of the bifunctional beta cat-
alyst under inert atmosphere. There is ample spectroscopic
and physicochemical evidence to indicate that high-temperature
treatment of beta zeolites, possibly in combination with steam,
converts pseudotetrahedral or pseudooctohedral species [e.g.,
(H2O)Al(OSi)3 or (H2O)3Al(OSi)3] to framework Lewis acid
sites [e.g., Al(OSi)3 or Al(OSi)2OH] [21–23]. Temperatures
well above 450 ◦C seem to be required for this process. Because
[1] J. Fleischer, K. Bauer, R. Hopp, DE 2109456, Haarmann & Reimer, 1971.
[2] S. Otsuka, K. Tani, T. Yamagata, S. Akutagawa, H. Kumobayashi,
M. Yagi, EP 68506, Takasago, 1972.
[3] N. Ravasio, N. Poli, R. Psaro, M. Saba, F. Zaccheria, Top. Catal. 213
(2000) 195.
[4] C. Milone, C. Gangemi, R. Ingoglia, G. Neri, S. Galvagno, Appl. Catal.
A 184 (1999) 89.
[5] C. Milone, C. Gangemi, G. Neri, A. Pistone, S. Galvagno, Appl. Catal.
A 199 (2000) 239.
[6] A.F. Trasarti, A.J. Marchi, C.R. Apesteguia, J. Catal. 224 (2004) 484.
[7] P. Mäki-Arvela, N. Kumar, D. Kubicka, A. Nasir, T. Heikkilä, V.-P. Lehto,
R. Sjöholm, T. Salmi, D.Yu. Murzin, J. Mol. Catal. A 240 (2005) 72.
[8] M.G. White, Heterogeneous Catalysis, Prentice–Hall, Englewood Cliffs,
NJ, 1991, 88.
[9] Y. Nakatani, K. Kawashima, Synthesis (1978) 147.
[10] V.K. Aggarwal, G.P. Vennall, P.N. Davey, C. Newman, Tetrahedron
Lett. 39 (1998).
[11] T. Iwata, Y. Okeda, Y. Hori, Eur. Pat. Appl. EP 12251632 A2, 2002.
[12] P. Maki-Arvela, N. Kumar, V. Nieminen, R. Sjöholm, T. Salmi, D.Yu.
Murzin, J. Catal. 225 (2004) 155.
[13] F. Iosif, S. Coman, V. Pârvulescu, P. Grange, S. Delsarte, D. De Vos, P. Ja-
cobs, Chem. Commun. (2004) 1292.
[14] S. Nishimura, Handbook of Heterogeneous Catalytic Hydrogenation for
Organic Synthesis, Wiley, New York, 2001.
[15] G.K. Chuach, S.H. Liu, S. Jaenicke, L.J. Harrison, J. Catal. 200 (2001)
352.
[16] Z. Yongzhong, N. Yuntong, S. Jaenicke, G.K. Chuach, J. Catal. 229 (2005)
404.
[17] T.A. Peters, N.E. Benes, A. Holmen, J.T.F. Keurentjes, Appl. Catal. A 297
(2006) 182.
[18] A. Corma, M. Renz, Chem. Commun. (2004) 550.
[19] V. Ponec, Appl. Catal. A 149 (1997) 27.
[20] A. Giroir-Fendler, D. Richard, P. Gallezot, Catal. Lett. 5 (1990) 175.
[21] G.H. Kuehl, H.K.C. Timken, Micropor. Mesopor. Mater. 35–36 (2000)
521.
[22] J.A. van Bokhoven, D.C. Koningsberger, P. Kunkeler, H. van Bekkum,
J. Catal. 211 (2002) 540.
[23] P.J. Kunkeler, B.J. Zuurdeeg, J.C. van der Waal, J.A. van Bokhoven, D.C.
Koningsberger, H. van Bekkum, J. Catal. 180 (1998) 234.