The conversion of 1,8-cineole sourced from renewable Eucalyptus oil to p-cymene over a palladium doped γ-Al2O3 catalyst

Article abstract of DOI:10.1016/j.cattod.2011.05.037

The conversion of 1,8-cineole, sourced from the steam distillation of Eucalyptus waste, to p-cymene has been studied. Both alumina and palladium on alumina have been found to be active catalysts, the latter producing p-cymene in near quantitative conversion. Catalyst testing coupled with catalyst characterisation showed the catalysts to be multifunctional, promoting dehydration, dehydrogenation and isomerisation in a one step process.

Full text of DOI:10.1016/j.cattod.2011.05.037

Catalysis Today 178 (2011) 98–102
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
The conversion of 1,8-cineole sourced from renewable Eucalyptus oil to
p-cymene over a palladium doped ␥-Al O catalyst
2
3
a,∗
b
b
a
a
a
Benjamin A. Leita , Peter Gray , Mike O’Shea , Nick Burke , Ken Chiang , David Trimm
a
CSIRO Earth Science and Resource Engineering, Private Bag 10, Clayton South VIC 3169, Australia
CSIRO Materials Science and Engineering, Private Bag 10, Clayton South VIC 3169, Australia
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 7 April 2011
Received in revised form 24 May 2011
Accepted 30 May 2011
Available online 20 July 2011
The conversion of 1,8-cineole, sourced from the steam distillation of Eucalyptus waste, to p-cymene
has been studied. Both alumina and palladium on alumina have been found to be active catalysts, the
latter producing p-cymene in near quantitative conversion. Catalyst testing coupled with catalyst char-
acterisation showed the catalysts to be multifunctional, promoting dehydration, dehydrogenation and
isomerisation in a one step process.
©
2011 Elsevier B.V. All rights reserved.
Keywords:
Eucalyptus oil
Palladium catalyst
p-Cymene
Hydrogen
Renewable
Green process
1
. Introduction
when pumice was used as a support, to p-cymene with yields of up
to 70%.
Para-cresol is an industrially important chemical used in the
Both the metal and the support were suggested to affect the
reaction sequence and this was supported by Roberge et al. [12]
who studied the conversion of pinenes to p-cymene. They found
that the supported palladium catalyst displayed dual-functionality,
with metal sites responsible for hydrogenation/dehydrogenation
and acid sites on the support responsible for isomerisation [13].
Earlier preliminary studies have shown that palladium on alu-
mina was an excellent catalyst for the conversion of cineole to
p-cymene [3]. The present paper reports more extensive studies
of the efficiency and mechanism of the reaction.
manufacture of pharmaceuticals, dyes and fertilisers [1]. Cresols
are generally produced via a two step process, the Friedel–Crafts
isopropylation of toluene to give mixed cymene isomers, followed
by an oxidation to yield mixed cresols and acetone [2]. Although the
raw materials are inexpensive, mixed cresols are produced, which
both are challenging to separate and of limited value in that only
p-cresol has widespread uses.
Alternative synthetic routes based on naturally occurring chem-
icals such as pinenes, terpenes and cineole have recently found
favour [3,4]. In particular, 1,8-cineole hereafter referred to as cine-
ole, isolated by steam distillation of Eucalyptus oil [5–7], has been
heated in the presence of supported metal catalysts to produce both
terpenes and high yields of p-cymene (Scheme 1).
2. Experimental
2.1. Catalyst preparation and characterisation
Related reactions have been studied previously [8,9]. Matsuura
and Waki [10] found that gamma alumina catalysed the pyroly-
Palladium doped ␥-Al O3 catalysts were prepared by wet
2
◦
sis of cineole above about 220 C, conversion being complete at
impregnation. ␥-Al O pellets (Saint-Gobain NorPro) were first
2
3
◦
3
00 C. Alumina was suggested to promote dehydration of cineole
◦
dried at 90 C and then immersed in an aqueous solution of pal-
ladium nitrate dihydrate. The mixture was stirred and the solvent
removed (60 C for 4 h in vacuo). The Pd-␥-Al O catalysts were
then dried and calcined at 350 C for 12 h. The total metal content
0.38% Pd on ␥-Al O was assessed by ICP analysis (Varian Vista
to dipentene, some of which then reacted further to cymene. Hügel
et al. [11] found that platinum and palladium catalysts supported
on silica promoted hydrogenolysis of cineole to p-menthanes and,
◦
2
3
◦
(
2
3)
ICP-OES). The BET surface area was determined by the adsorption
of N2 (Micromeretics Tristar 3000). Metal surface area determina-
tion and temperature programmed reduction experiments were
∗ Corresponding author. Tel.: +61 3 95458130; fax: +61 3 95458380.
E-mail addresses: benjamin.leita@csiro.au, ben.leita@csiro.au (B.A. Leita).
0
920-5861/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.cattod.2011.05.037

B.A. Leita et al. / Catalysis Today 178 (2011) 98–102
99
O
heat
+
+
H O
2
+
H2 + +
CO CO2
catalyst
O , Ar
2
hydrophilic phase
gaseous phase
cineole
p-cymene
other products
hydrophobic phase
dipentene
+
Scheme 1. General scheme for the catalytic transformation of cineole over metal doped ␥-Al2O3 catalysts [3].
carried out on a Micromeretics AutoChem. Metal surface areas were
obtained by hydrogen and carbon monoxide chemisorption, the
catalysts being reduced for 4 hr in hydrogen before analysis. Prior
to temperature programmed reduction (TPR) the fresh catalyst was
◦
◦
dried in argon at 150 C, cooled to 50 C and reduced in a stream of
◦
1
0% H2 in Ar. The temperature was increased at a rate of 5 C/min
to 600 C, where it was held for 30 min before cooling to ambient
temperature at a rate of 10 C/min.
◦
◦
2.2. Catalytic activity measurements
All reactions were studied using an electrically heated stain-
less steel down flow tubular reactor (13.5 mm internal diameter,
00 mm length) maintained at a pre-set temperature and atmo-
3
spheric pressure. A K-type thermocouple was used to monitor the
temperature of the catalyst bed. The liquid product was collected
Fig. 1. TPR analysis of the 0.38% Pd-␥Al2O3 catalyst; Fresh catalyst, 10% H2 in Ar;
temperature increased at a rate of 5 C/min to 600 C, and held for 30 min; Used
catalyst sample: same conditions.
◦
◦
◦
by a two trap trapping system, the first trap at 40 C collected all the
◦
liquid products and the second trap was at 0 C to insure only the
nation of new and used catalysts produced results summarised in
Table 1.
gaseous products went through into an online gas chromatograph
for analysis.
In a typical experiment, 3 g of catalyst was loaded into a stain-
less steel mesh basket and placed inside the tubular reactor. The
temperature was stabilised for 1 h before injection of feedstock. Liq-
uid cineole was injected into a stream of carrier gas (150 mL/min)
PdO + H → Pd + H O
(1)
2
2
The lower temperature peak corresponded to
a
hydro-
gen/palladium ratio of 0.73, in agreement with the results of Bigey
et al. [14]. The second reduction (peak ˇ1) occurred at temperatures
above those used for the cineole reaction and probably corresponds
to the reduction of Pd(II) centres bound to the alumina. The used
catalyst (see below) gave ˛ and ˇ peaks at higher temperatures of
◦
upstream of the pre-heater (200 C) at a rate of 0.5 mL/min using an
ISCO 500D syringe pump. As the reaction started the furnace tem-
perature was adjusted to maintain the desired bed temperature.
Organic and aqueous fractions were collected together from the
first trap. The aqueous fraction was found to be predominantly
water, with very small traces of cineole via GCMS and 1H NMR.
The organic/oil product (10 ␮L) was dissolved in 1.5 mL of 0.1 v/v%
mesitylene in acetonitrile prior to analysis. Mesitylene was used as
the internal standard.
◦
◦
280 C and 502 C. This indicates that the catalyst has undergone
some changes, possibly structural indicating surface rearrange-
ment and some sintering of the palladium, or more likely deposits
of carbonaceous material during use. The reduction in metal surface
area of the used catalyst also suggests either some rearrangement
of Pd metal clusters, possibly due to sintering, or coverage of metal
sites with carbonaceous deposits during the course of the run.
While Pd sintering has been reported at temperatures as low as
These samples were analysed by gas chromatog-
raphy (Varian CP-3800 equipped with
a
ZB-5 column
(
30 m × 0.25 mm × 0.25 ␮m). and a FID detector). The heating
◦
◦
◦
◦
◦
rate used was: 60 C for 2 min, 10 C/min to 110 C, 40 C/min to
2
300 C [15,16] it is unusual to see sintering of Pd at such low tem-
◦
50 C and remaining there for 2 min. Authentic samples of cineole
peratures [17]. It is more likely that the deactivation seen here is
due to fouling of the surface of the catalyst.
(
Aldrich), p-cymene (Aldrich) and dipentene (Aldrich) were used
to verify the assignment of identified products. The major products
TGA analysis performed on the fresh and used catalyst clearly
showed loss of mass in a dilute oxygen atmosphere for the used
samples, suggesting some build-up of material on the catalyst that
could be removed with oxidation. Discolouration of the catalyst
of each oil sample, in most cases p-cymene, were confirmed using
1
13
1
GCMS and H and C{ H} NMR spectroscopy.
Online analyses of gas products were performed using an online
Shimadzu GC 8A fitted with a 12 m HAYESEP Q column (Alltech)
and a TCD detector.
Table 1
Textural properties of doped ␥-Al2O3 catalysts.
3
. Results and discussion
TPR analysis using a stream of 10% H /Ar over the fresh catalyst
Catalyst
Specific
surface
area/m
Metal
surface
area/m
Pore
Average
pore
width/A
volume/
2
−1
2
−1
3
−1
˚
g
g
cm
g
2
◦
◦
showed reduction peaks at 189 C and 387 C (Fig. 1), the ␣ peak
being assigned to the reduction of PdO by H2 in both cases (Eq. (1))
Fresh
Used (4th
cycle)
265
220
0.906
0.426
0.761
0.611
115
111
[
14]. The metal surface area is taken per gram of catalyst. Exami-

1
00
B.A. Leita et al. / Catalysis Today 178 (2011) 98–102
Fig. 2. Product yield as the reaction temperature increased. (␥-Al2O3 and 0.38% Pd-␥-Al2O3 catalyst, cineole injection 0.5 mL/min, total flow rate of carrier gas 150 mL/min.)
was seen after cycling. This was most obvious at the upstream end
of the catalyst bed, suggesting a progressive fouling of the catalyst
along the reactor bed. Morra et al. have reported a similar progres-
sive fouling effect in their study on dehydrogenation of o-xylene
Initial screening of catalyst activity was carried out using both
the alumina support and the Pd-␥Al O catalyst. The tests involved
2
3
a 0.5 mL/min flow of liquid cineole, which was vaporised into
a stream of 150 mL/min argon and passed over 3 g of catalyst,
◦
[
18]. Attempts to quantify carbon build-up on the catalyst over the
which was progressively heated to 350 C. The observed yield of
course of the reaction were hampered by the heterogeneous nature
of the carbon accumulation on the catalyst. In removing the cata-
lyst from the reactor the catalyst did not maintain its orientation
as it was in the reactor tube.
p-cymene and flow rate of hydrogen are summarised in Fig. 2. Both
alumina and Pd-␥-Al O were found to be catalytically active, con-
version of the cineole occurring above ∼200 C. The production of
p-cymene was lower over alumina with a maximum yield of 47%
2
3
◦
Scheme 2. The proposed reaction sequence for the catalytic transformation of cineole into p-cymene. The initial reaction involves breakage of the weakest C–O bond in
cineole [16] followed by rapid dehydration step followed by a dehydrogenation step to produce a mixture of doubly unsaturated isomers. In the final step, dehydrogenation
and rearrangement of the double bonds occurs to produce p-cymene.

B.A. Leita et al. / Catalysis Today 178 (2011) 98–102
101
Fig. 3. Product yields over a 3 h time period at fixed reaction bed temperature. Palladium doped ␥-Al2O3 catalysts, cineole injection 0.3 mL/min, and total flow rate of carrier
gas 150 mL/min.
◦
being observed at around 285 C. No hydrogen was observed until
achieved. Part of the hydrogen can be expected to reduce PdO in the
original catalyst but this should be minimal in a flow system. What
seems more likely is that hydrogen is used to gasify carbonaceous
deposits. The reaction proceeds via dehydration, dehydrogenation
and isomerisation with both dehydration and isomerisation being
favoured by acidic catalysts, such as alumina [19] (see Scheme 2).
Such catalysts are prone to deactivation by deposition of car-
bonaceous material on the surface [19] which can be minimised
by gasification. On the assumption that the difference between
◦
the catalyst was heated to 400 C. In contrast, p-cymene yields
were higher over the Pd-␥-Al O catalyst but reaction was initi-
2
3
◦
ated again at ca 200 C. Initiation probably resulted from reduction
of palladium oxide (189 C) either by cineole or by hydrogen pro-
duced from the alumina catalysed conversion of cineole. The yield
of p-cymene increased with temperature, reaching a plateau at
about 275 C. Hydrogen yields with Pd-␥-Al O catalyst were sig-
nificantly greater than over alumina alone, reaching a plateau at
◦
◦
2
3
−3
ca 1.1 × 10 mol/min. This is somewhat less than the value of
expected and observed hydrogen yields over Pd-␥-Al O3 is also
2
−
3
2
.75 × 10 mol/(min gcat) expected on the 1:1 stoichiometry of
due to removal of hydrogen by gasification, the smaller amounts
of hydrogen expected to be produced in conversion over alumina
would be totally consumed in the same way. At higher tempera-
tures deposition would become faster and the catalytic activity of
alumina would decrease, as observed, while the Pd can be expected
to catalyse gasification and hence to retain overall catalytic activity.
The conversion sequence suggested by Hügel et al. [11] is similar
to what is proposed in Scheme 2. Acid catalysed ring opening of the
C–O bond leads p-menthen-8-ol which subsequently dehydrates
to form a di-olefin. Subsequent rearrangement/dehydrogenation
leads to a range of terpenes and to p-cymene, and Roberge et al.
[12] have shown that the balance between acid catalysis and
hydrogenolysis, as controlled by the catalytic components, dictates
the final product distribution. Using the conversion of pinenes as
cineole conversion.
It is clear that alumina itself catalyses the conversion of cineole
to p-cymene, but this was not the case supports such as pumice
[
1
11]. Hügel et al. [11] report yields of p-cymene increasing from
◦
◦
3% at 200 C to 68% at 300 C over the pumice supported pal-
ladium system, but the conversion was studied in the presence
of 10% hydrogen. The present yields over Pd-␥-Al O3 are signifi-
2
cantly higher (ca. 95%), despite the fact that the Pd loading was
much lower. Consequently, the high yield observed can in part be
attributed to palladium–alumina interactions, not just the additive
effect of the palladium and alumina components.
The absence of hydrogen from the alumina results was surpris-
ing, particularly in view of the high (47%) conversions that could be
Fig. 4. (A) Left graph shows yields of p-cymene as a function of time. (B) Right graph shows hydrogen production over time (8 h run, sample were taken every hour, fixed
◦
bed temperature of 250 C, injection rate for 1,8-cineole of 0.3 mL/min, 4 g (3 g for re-generated) catalyst, 150 mL/min carrier gas, 7.3% oxygen.) (First cycle was carried out
◦
with fresh catalyst; subsequent cycles were carried out under the same conditions but after calcinations at 350 C for 12 h).

1
02
B.A. Leita et al. / Catalysis Today 178 (2011) 98–102
an example, Pd-␥-Al O was found to be very efficient at producing
as the major hydrophilic liquid product and hydrogen as the major
2
3
p-cymene.
gas products. Both palladium and ␥-Al O3 were found to be bi-
2
Roberge et al. also noted catalytic deactivation in the systems
12] and, to investigate this, a further series of experiments was
functional catalysts but the palladium doped system showed very
[
high activity and selectivity, yielding >99% p-cymene, while pro-
◦
carried out better to quantify the yields of p-cymene and hydro-
ducing large amounts of hydrogen at a bed temperature of ∼250 C.
◦
◦
gen over the present catalyst at reaction conditions below 300 C.
The reaction mechanism is suggested to involve C–O bond fission
in cineole, followed by dehydrogenation/isomerisation to produce
p-cymene. The catalyst was found to be very stable but the possibil-
ity of slow deactivation and regeneration was explored. Sequential
regeneration led to some loss of activity. The relatively low reaction
temperatures, high yields, high selectivity, and minimum waste
products of the process make it an attractive route from bio renew-
able materials to p-cymene.
A series of 3 h runs at fixed bed temperatures of 220, 250 and 280 C
were chosen. The production of hydrogen was monitored continu-
ously against time (Fig. 3) while the liquid phase was collected as
one sample over the 3 h period to enable an accurate mass balance
of the product.
The liquid phase analysis showed distinct increases in the initial
◦
yield of p-cymene as the temperature was raised from 220 C to
◦
2
50 C, but little deactivation. A gradual decrease in p-cymene with
time online was noted and an increase in other terpenes occurred
Acknowledgements
◦
at temperatures greater than ∼250 C. Only very small changes in
the amount of hydrogen produced were observed, consistent with
the suggestion that the reaction runs at equilibrium.
The authors wish to thank the CSIRO OCE PDF funding scheme
for financial support for this project.
Although was apparent that the conversion was very stable,
longer term deactivation would not be unexpected [12]. As a result,
a series of experiments was run in which the effect of regenera-
tion could be studied. Small amounts of oxygen were added to the
inlet gas and, after 6 h operation, the catalyst was “regenerated” by
References
[
1] A.K. Mukhopadhyay, Industrial Chemical Cresols and Downstream Derivatives,
Marcel Dekker, 2005.
[
2] H. Fiege, Ullmann’s Encyclopedia of Industrial Chemistry, vol. A8, VCH, Wein-
heim, 1985, p. 25.
◦
calcination at 350 C for 12 h and the cycle repeated.
The results, shown in Fig. 4, showed a small decrease in p-
cymene production in the first cycle and a very small deactivation
with time online together with, not surprisingly, a larger decrease
in hydrogen. Subsequent cycles led to lower yields and to more
rapid deactivation (Fig. 4). Part of the deactivation was due to the
[3] B.A. Leita, A. Warden, N. Burke, D. Trimm, M. O’Shea, Green Chem. 12 (2010)
0–76.
7
[
4] B. Leita, P. Gray, N. Burke, D. Trimm, provisional patent Filed 16th July 2009,
Australian Provisional Patent Application No. 2009903333.
[5] J.L. Simonsen, L.N. Owen, The Terpenes, vol. 1, 2nd ed, Cambridge University
Press, Cambridge, 1973, p. 423.
[
6] J. Bartle, G. Oslen, D. Cooper, T. Hobbs, Int. J. Global Energy Issues 27 (2) (2007)
15–137.
2
−1
2
−1
reduction of metal surface area from 0.906 m g to 0.426 m g
1
in the used catalyst after all 4 cycles. However TPR analysis showed
the used catalyst to be fully reduced only at ca 275 C, a temperature
[7] G.N. Pain, Eucalyptus Oil Report, Verve Energy, 2006.
[8] R.W. Neuzil, D.H. Rosback, R.H. Jensen, J.R. Teague, A.J. deRosset, Chemtech 8
◦
(
1980) 498.
higher than that used for activity testing. Thus, under process con-
ditions, most of the palladium would be present as the catalytically
inactive oxide and the bulk of the conversion would result from
the activity of the alumina support. More extensive deactivation,
as observed, would not be unexpected.
[
9] M.A. Martín-Luengo, M. Yates, M.J. Martínez Domingo, B. Casal, M. Igle-
sias, M. Esteban, E. Ruiz-Hitzky, Appl. Catal. B: Environ. 81 (2008)
218–224.
[
[
10] Matsuura, T. Waki, J. Sci. Hiroshima Univ. 20 (3) (1957) 177–186.
11] H.M. Hügel, W.R. Jackson, C.D. Kachel, I.D. Rae, Aust. J. Chem. 30 (1977)
1287–1292.
[
12] D.M. Roberge, D. Buhl, J.P.M. Niederer, W.F. Hölderich, Appl. Catal. A. 215 (2001)
It is clear that Pd-␥-Al O is an active and selective catalyst that
2
3
111–124.
deactivates very slowly. Regeneration, when it is necessary, must
ensure removal of carbonaceous deposit and subsequent reduction
to the active palladium metal.
[
13] F. Di Gregorioa, N. Kellera, V. Keller, J. Catal. 256 (2008) 159–171.
[14] C. Bigey, 1L. Hilaire, G. Maire, J. Catal. 198 (2001) 208–222.
[15] M. Cargnello, T. Montini, S. Polizzi, N.L. Wieder, R.J. Gorte, M. Graziani, P. For-
nasiero, Dalton Trans. 39 (2010) 2122.
[
16] A.F. Lee, C.V. Ellis, K. Wilson, N.S. Hondow, Catal. Today 157 (2010) 243.
4
. Conclusions
[17] C.H. Bartholomew, Appl. Catal. A: Gen. 212 (2001) 17.
[
18] G. Morra, D. Farrusseng, E. Guillon, S. Morin, C. Bouchy, P. Duchêne, C. Mirodatos,
Catal. Today 137 (2008) 71.
The conversion of the bio-renewable feedstock, 1,8-cineole, over
[
19] D. Buhl, D.M. Roberge, W.F. Hölderich, Appl. Catal. A 188 (1999) 287.
a palladium doped ␥-Al O3 catalyst is found to produce p-cymene
2