Linoleic acid isomerization over mesoporous carbon supported gold catalysts

Article abstract of DOI:10.1016/j.cattod.2009.07.065

Gold catalyst supported on mesoporous carbon and silica were synthesized, characterized by TEM, XRD, XPS and tested in linoleic acid isomerization. Nature of the support affects the selectivity towards isomerization in relation to unwanted hydrogenation.

Full text of DOI:10.1016/j.cattod.2009.07.065

Catalysis Today 150 (2010) 32–36
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Linoleic acid isomerization over mesoporous carbon supported gold catalysts
a,d
b
a
a
Olga A. Simakova , Anne-Riikka Leino , Betiana Campo , P a¨ ivi M a¨ ki-Arvela ,
b
a,c
a,
Kriszti a´ n Kord a´ s , Jyri-Pekka Mikkola , Dmitry Yu. Murzin *
a
b
c
˚
˚
Laboratory of Industrial Chemistry, Process Chemistry Centre, Faculty of Technology, Abo Akademi University, Biskopsgatan 8, FIN-20500 Abo/Turku, Finland
Microelectronics and Materials Physics Laboratories, EMPART Research Group of Infotech Oulu, University of Oulu, FIN-90014 Oulu, Finland
Department of Chemistry, Technical Chemistry, Umea˚ University, SE-901 87 Umea˚, Sweden
d
Boreskov Institute of Catalysis, Novosibirsk, Russia
A R T I C L E I N F O
A B S T R A C T
Article history:
Gold catalyst supported on mesoporous carbon and silica were synthesized, characterized by TEM, XRD,
XPS and tested in linoleic acid isomerization. Nature of the support affects the selectivity towards
isomerization in relation to unwanted hydrogenation. In particular carbon support allowed much higher
selectivity in double bond migration compared to silica. Effect of carbon surface oxidative pre-treatment
on selectivity of catalyst was investigated.
Available online 7 August 2009
Keywords:
Gold
Mesoporous carbon
Linoleic acid
Isomerization
ß 2009 Elsevier B.V. All rights reserved.
1
. Introduction
hydrogenation. In fact, selective isomerization under hydrogen
2
atmosphere has been carried out by using Ag/SiO as a catalyst. The
Linoleic acid (C18:2), which has double bonds located on carbons
and 12, both in cis configuration, can be isomerized through
differences in the selectivity of ruthenium and silver catalysts are
based on the differences in the hydrogen adsorption behavior on
the metal catalysts. It is evident that the weakly bound hydrogen
plays a key role in the selective isomerization of conjugated
double bonds (to CLAs) [8]. From this point of view, gold is a
potential catalyst for the selective isomerization reaction as it is
reluctant to hydrogenate C 55 C double bond. Using mesoporous
carbons as a support for preparing gold catalysts is attractive for
several reasons, namely better mechanical stability is achieved as
well as better accessibility of large organic compounds to the
active sites compared to the microporous supports [9]. However,
the formation of carbon supported gold catalysts is challenging
due to reduction of the metal ions by free electrons of the carbon
matrix taking place immediately after carbon is brought in contact
with the solution. Gold is adsorbed on carbon surface only in the
9
double bond migration to conjugated linoleic acids (CLAs) (see
Scheme 1). These acids are widely used as health promiting agents
in food and pharmaceuticals due to their anticarcinogenic and
antioxidative properties [1], which can be attributed to the
presence of 9-cis,11-trans and the 10-trans,12-cis isomers [2–6].
The CLAs are produced industrially under alkaline conditions,
implying alkali bases, solvents and neutralizing agents being a real
shortcoming from an ecological and economic point of view. In
addition, solvents are limited to the ones that are harmless to
human body. These drawbacks could be circumvented using
heterogeneous catalysis. However, in the isomerization reaction
small quantity of hydrogen is needed to enhance the activity.
Selective isomerization under hydrogen atmosphere is challenging
due to hydrogenation of a C 55 C bond over typical metal catalysts
0
form of Au , followed by aggregation of particles [10–14].
(
e.g. ruthenium) and subsequent formation of stearic acid instead
Therefore, surfactant-protected gold particles are formed to
avoid the direct adsorption of gold complexes on carbon surface.
Preparation of gold catalysts by immobilization of gold sols on
carbon surfaces allows to support gold particles in a highly
dispersed state. The size of supported gold particles can be varied
in the range of 3–15 nm. The systematic study of the gold sols
application as a catalyst precursor for preparation of Au/C
catalysts is presented by the group of Prati and co-workers
of CLAs. In our previous work it was demonstrated that
isomerization of linoleic acid is possible over heterogeneous
catalysts [7]. The reaction was performed in two-steps; first the
catalyst was pre-treated with hydrogen, and linoleic acid was
subsequently converted into CLAs under nitrogen. However, it
would be beneficial to carry out the selective isomerization in one-
step under hydrogen atmosphere without C 55 C double bond
[
14–18].
In this work we aimed to investigate selectivity of mesoporous
*
Corresponding author.
carbon supported gold catalysts in comparison with silica
supported gold catalyst in linoleic acid isomerization.
E-mail address: dmurzin@abo.fi (D.Yu. Murzin).
0920-5861/$ – see front matter ß 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.cattod.2009.07.065

O.A. Simakova et al. / Catalysis Today 150 (2010) 32–36
33
Scheme 1. Isomerization of linoleic acid to biologically active conjugated linoleic acid isomers.
2
. Experimental
Peak 4.1 software. The peak areas were computed by fitting the
experimental spectra to Gaussian/Lorentzian curves after remov-
ing the background (using the Shirley function).
2.1. Catalyst preparation and characterization
Hydrogen tetrachloroaurate (99.9% ABCR, Darmstadt) was
2.3. Catalytic linoleic acid isomerization
diluted with deionized water up to concentration 60
mg/ml.
Maintaining this solution under vigorous stirring, 2 wt.% solution
of PVA was added at Au/PVA weight ratio equal to 1:5. Freshly
32 2
Linoleic acid (C18H O ) of 99% purity was supplied by Sigma–
Aldrich (Germany). The solvent n-decane of 95.5% purity was
purchased from Fluka (Switzerland). The trans-10,cis-12, cis-
9,trans-11, cis-9,cis-11, and trans-9,trans-11 isomers of CLA
prepared 0.1 M solution of NaBH
NaBH weight ratio equal to 1:100. Within a few minutes of sol
generation, the sol was immobilized by adding mesoporous carbon
4
was added dropwise at Au/
4
32 2
(C18H O ) were supplied by Matreya Inc. (USA). The isomerization
2
2
material Sibunit (SBET = 450 m /g, micropore area 37.7 m /g,
experiments were carried out in a 200 ml stirred glass reactor,
which was equipped with a reflux condenser and a heating jacket.
average pore diameter 9 nm). The nominal metal loading was
1
wt.%. Carbon support was used as such or after an oxidative
The catalyst (mcat = 1 g, fractions of Au/C and Au/SiO
2
catalysts
treatment. Gold sols were thus supported on non-oxidized and
pre-oxidized by nitric acid carbon materials in order to study the
influence of the support surface on physical–chemical properties of
catalysts. Powdered carbon material Sibunit was mixed with
below 70 and 40–63 respectively to suppress internal
mm
diffusion) was introduced into the reactor, preactivated under
hydrogen flow (100 ml/min) at 150 8C for 1 h. An amount of
200 mg of linoleic acid was mixed with 70 ml of solvent in a glass
drip funnel. The amounts of the catalyst and the substrate were
chosen following the paper which reported the best results for
linoleic acid isomerization performance over ruthenium on carbon
catalysts [22]. Such conditions correspond to the substrate/metal
molar ratio equal to 14 mol/mol. The air above the reaction
mixture and dissolved in the liquid phase was purged out by a
nitrogen flow of 100 ml/min. The isomerization reaction was
conducted at 150 8C under stirring at 900 rpm to eliminate
5
3
wt.% HNO aqua solution at 373 K for 3 h to introduce oxidic
surface group as described in [19]. Due to the oxidative pre-
treatment of carbon the formation of functional groups (phenolic,
aldehydic, carboxylic, etc.) on the surface was achieved [20]. Such
groups presumably allow stabilization of highly dispersed gold.
After 2 h the slurry was filtered and the catalysts were washed
thoroughly with hot deionized water, followed by drying overnight
at 60 8C in air.
The sample of 1 wt.% Au/SiO
2
was synthesized by deposition–
external mass transfer. Samples (5 ml) were withdrawn from the
precipitation method with urea (DPU) as described in [21] using
reactor at certain intervals through a sampling valve.
2
SiO
2
(Merck, SBET = 480–540 m /g, 40–63
mm fraction), HAuCl
4
(
99.9% ABCR, Darmstadt), and urea (Riedel-de-H a¨ en, 99.5%).
2.4. Analytical procedure
2
.2. Catalyst characterization
The samples from the batch reactor were analyzed by a
temperature-programmed gas chromatograph (GC, Hewlett Pack-
ard 6890 Series) using 50 m HP-5 column (inner diameter:
A Carlo Erba sorptometer 1900 was utilized in the measure-
ments of the surface area and micropore volume by nitrogen
physisorption. BET method was used to calculate the surface areas.
0.20 mm, film layer: 0.11 mm) and a flame ionization detector
(FID) operating at 300 8C. Prior to GC method analysis the samples
were silylated by using N,O-bis(trimethylsilyl)trifluoroacetamide
(BSTFA) and trimethylchlorosilane (TMCS) at 60 8C, both supplied
by Fluka (Switzerland).
X-ray diffraction experiments were performed using Cu K
radiation (Siemens D5000 diffractometer equipped with
graphite monochromator to suppress fluorescent and Cu K
a
a
b
radiation). The average crystallite size of gold catalyst particles
was estimated using Scherrer’s equation from the peak half-
3. Results and discussion
widths of Au(1 1 1) reflection measured at 2
Q
a
ꢀ 38.38. Instru-
2 radiation are
mental broadening and unresolved Cu K
neglected (error less than 2%).
a
1 and
3.1. Gold particle size
Electron microphotographs of the samples were taken by a LEO
12 OMEGA energy-filtered transmission electron microscope
using 120 kV acceleration voltage. Histograms of particle size
distribution were obtained by counting at least 100 particles on the
micrographs for each sample.
The XPS measurements were carried out in a PHI Quantum 2000
Scanning ESCA Microprobe spectrometer with Al anode. The
survey and multi-spectra were collected at 187.75 and 11.75 eV
pass energy respectively. Spectra were processed by means of XPS
Characterization data confirmed that gold can be successfully
deposited on mesoporous carbon from gold sols. The metal particle
size was measured by TEM and XRD techniques. It was shown that
the dispersion of gold particles reached on silica and carbons is
approximately the same (Figs. 1–4). The carbon support seems to
be partially graphitized shown by the C(0 0 2) reflection at
9
2Q
ꢀ 268. In both samples, Au(1 1 1) reflection is missing
indicating very small catalyst size. TEM micrographs show the
support is having a weekly ordered structure consisting of

3
4
O.A. Simakova et al. / Catalysis Today 150 (2010) 32–36
2
Fig. 1. TEM image and gold particle size distribution of 1 wt.% Au/SiO .
2
Fig. 2. X-ray diffraction patterns and particle size distribution of gold particles in 1 wt.% Au/SiO .
Fig. 3. TEM images and X-ray diffraction patterns of 1 wt.% Au/C (non-oxidized Sibunit).
connected ball-shaped carbonaceous features of ca. 100 nm
diameter. The carbon content is (partly) graphitized as indicated
by the strong C(0 0 2) and C(0 0 4) rings in the electron diffraction
patterns. The particles size distribution of Au/C (not shown) is very
narrow, with almost 100% of supported gold particles in the range
of 1–2 nm (some of them are indicated by the arrows in the
micrographs). The lack of diffraction patterns from Au is in
agreement with the small particle size. Metal particles size
few very large gold crystals about 20 nm were presented. This
particle size distribution can be explained by the following
mechanism of the active component formation. Point Zero Charge
(PZC) of applied silica was found to be equal to 4. The value of the
initial solution pH is 2.5. Since pH < PZC and the surface of silica is
positively charged, adsorption of anionic complexes of gold can be
À
performed. First gold is adsorbed as [AuCl
3
(OH)] complexes [23]
followed by hydrolysis of Au–Cl bonds on the support surface. This
À
distribution in 1 wt.% Au/SiO
2
can be considered to be mono-
3
species forms highly dispersed gold particles. Other [AuCl (OH)]
modal with the particles predominantly below 3.5 nm, since
besides a minor amount of particles in the range of 5–6 nm also a
gold complexes are hydrolyzed in the solution and adsorbed on the
surface as larger hydroxocomplexes of gold.

O.A. Simakova et al. / Catalysis Today 150 (2010) 32–36
35
Fig. 4. TEM images and X-ray diffraction patterns of 1 wt.% Au/C (oxidized Sibunit).
3
.2. XPS measurements
the presence of the acidic oxygen-containing groups at the edges of
graphene networks due to acid–base interactions [12]. Never-
theless the metal loading is lower in case of supporting gold on the
oxidized carbon. Formed by PVA stabilization in the solution prior
to immobilization gold sols are negatively charged and thus could
be repulsed by surface oxygen groups due to electrostatic
interactions.
2
The XPS analysis of 1 wt.% Au/SiO shows that Au 4f contains
two contributions. The binding energy (BE) of the metallic gold is
4 eV. The doublet observed at lower BE is 0.9 eV shifted (Fig. 5). In
8
gold catalysts this situation could be a consequence of two
phenomena: electron transfer from the support towards the gold
particles [24], or a dominant effect of the atoms at low coordinated
sites as proposed by Radnik et al. [25]. The first possibility should
be rejected due to the well known weak interactions of silica with
noble metal clusters [26]. The second one should be possible if the
gold particle size is lower than 2 nm. In XPS spectra of carbon
supported gold catalysts the BE for metallic gold is 84 eV, for
surface oxygen BE is taking 285 eV (for C( 55 O)–O–, C 55 O, C–OH, C–
O–C, C–C, C–H surface groups) and 533 eV (for C–OH, C–O–C
groups). The XPS shows that supported gold is in the zero valent
state in the case of non-oxidized and oxidized carbons. Atomic
ratio for Au/C and O/C was calculated from the peak areas ratio. The
results are presented in Table 1. Comparison of the Au surface
concentration in case of 1 wt.% Au/C and 1 wt.% Au/Cox shows
lower concentration of gold on the carbon surface in the case of the
oxidized support. The main difference between Au/C is observed in
the O 1s region. The pre-oxidized support has a new contribution at
lower BE attributed to oxygen from the C 55 O group. The ability of
carbon supports to stabilize dispersed metals is often attributed to
3.3. Catalytic properties
Selectivity of synthesized catalysts towards hydrogenation and
conjugation, especially towards desired products was investigated.
1 wt.% Au/Cox provided the highest conversion of linoleic acid and
the highest yield of hydrogenation products (Table 2), while the
evaluated metal loading is the lowest among the tested catalysts.
2
Although the conversion over 1 wt.% Au/C and 1 wt.% Au/SiO was
the same the yield of CLAs is higher in the case of linoleic acid
isomerization over 1 wt.% Au/C. Moreover, in the latter case there
are no hydrogenation products. Thus, the highest selectivity
towards conjugation was observed over non-oxidized carbon
supported gold catalyst. Despite the fact that three different
samples of gold catalysts have approximately the same gold
particle size, the behavior of these catalysts in linoleic isomeriza-
tion is different. It can be thus concluded that chemical properties
of support play an important role. Different catalytic behavior of
gold catalysts can be most likely related to the difference of their
surface chemistry. Presence of acidic groups on the supports
surface in case of oxidized carbon and silica results in formation of
hydrogenation products in linoleic acid isomerization due to
higher acidity of the supports compared to non-oxidized carbon
resulting in the protons donation. Although silica and non-oxidized
carbon supported gold catalysts show the same activity, at the
same time there are no hydrogenation products in the case of the
carbon supported catalyst. That can be attributed to the mechan-
Table 1
Atomic ratio of the atoms on the surface of 1 wt.%Au/C and
1
wt.%Au/Cox, calculated from XPS analysis.
1
wt.% Au/C
1 wt.% Au/Cox
Au/C
O/C
0.0009
0.1328
0.0005
0.1624
2
Fig. 5. Au (4f) region in XPS spectra of 1 wt.% Au/SiO .

3
6
O.A. Simakova et al. / Catalysis Today 150 (2010) 32–36
Table 2
Catalytic properties of gold catalysts in linoleic acid isomerization. Reaction time 4 h. X = conversion; S
P
i
= selectivity; Hydr. = hydrogenation; CLA = sum of CLAs.
P
P
Sample
X, %
S
9-cis,11-trans, %
S
10-trans,12-cis, %
S9-trans,11-trans, %
S
Hydr., %
0
CLA, %
Hydr./ CLA
1
1
1
wt.% Au/C
4
32
2
23
0
45
6
100
0
11
0.4
a
wt.% Au/Cox
wt.% Au/SiO
60
90
26
8
2
4
34
0
40
74
a
The same values of selectivity were observed at conversion of 4%.
[
[
5] V. Mougios, A. Matsakas, A. Petridou, S. Ring, A. Sagredos, A. Melissopoulou, N.
Tsigilis, M. Nikolaidis, J. Nutr. Biochem. 12 (2001) 585.
6] J.D. Palombo, A. Ganguly, B.R. Bistrian, M.P. Menard, Cancer Lett. 177 (2002) 163.
ism of the reaction involving the proton transfer due to Brønsted
and Lewis acidity (relatively weak in case of silica [27]) and the
subsequent hydride transfer mechanism leading to hydrogenation
[7] A. Bernas, N. Kumar, P. M a¨ ki-Arvela, E. Laine, B. Holmbom, T. Salmi, D.Yu. Murzin,
Chem. Commun. 10 (2002) 1142.
[
28,29]. This mechanism explaining the isomerization kinetics is
most likely operative, while the conventional hydrogen addition
Horiuti–Polanyi mechanism [30]) is much less probable since Au
does not adsorb readily H (H ).
[
[
8] M. Kreich, P. Claus, Angew. Chem., Int. Ed. 44 (2005) 7800.
9] Y.I. Yermakov, V.F. Surovikin, G.V. Plaksin, V.A. Semikolenov, V.A. Likholobov, A.L.
Chuvilin, S.V. Bogdanov, React. Kinet. Catal. Lett. 33 (1987) 435.
(
[
10] P.A. Simonov, A.V. Romanenko, I.P. Prosvirin, G.N. Kryukova, A.L. Chuvilin, S.V.
Bogdanov, E.M. Moroz, V.A. Likholobov, in: G. Pocelet, et al. (Eds.), Stud. Surf. Sci.
Catal., vol. 118, Elsevier, Amsterdam, 1998, p. 15.
2
4
. Conclusions
Gold catalysts supported on silica and mesoporous carbons were
[11] J.B. Hiskey, X.H. Jiang, G. Ramadorai, in: D.M. Hausen, et al. (Eds.), Proc. Gold’90
Symp., Society for Mining, Metallurgy, and Exploration, Inc., Littleton, CO, (1990),
p. 369.
[12] P.A. Simonov, V.A. Likholobov, in: A. Wieckowski, E.R. Savinova, C.G. Vayenas
prepared by deposition–precipitation with urea and from gold sols
respectively resulting in similar gold dispersion. However, when
tested in linoleic acid isomerization, these catalysts exhibited
different selectivity and activity. Different behavior of the catalysts
can be attributed to the support acidity leading to different
mechanism of isomerization reaction in the presence of hydrogen.
Thus, oxidative treatment of the mesoporous carbon, although
resulting in overall higher catalytic activity, promotes higher
selectivity towards hydrogenation in comparison with linoleic acid
isomerization. Selectivity towards isomerization is higher for the
silica supported catalyst, that for Au/Cox, however, hydrogenation is
also taking place. The reaction over the non-oxidized carbon
supported catalyst leads to formation of conjugated linoleic isomers
only with high selectivity towards the desired isomers.
(
Eds.), Physicochemical Aspects of Preparation of Carbon-supported Noble Metal
Catalysts//Catalysis and Electrocatalysis at Nanoparticle Surfaces, Marcel Dekker,
Inc., New York, 2003, pp. 425–426.
[
[
13] R. Garcia, M. Besson, P. Gallezot, Appl. Catal. A: Gen. 127 (1995) 165.
14] L. Prati, G. Marta, Gold Bull. 32 (1999) 96.
[15] F. Porta, L. Prati, M. Rossi, S. Coluccia, G. Martra, Catal. Today 61 (2000) 165–172.
16] S. Coluccia, G. Marta, F. Porta, L. Prati, M. Rossi, Catal. Today 61 (2000) 165.
17] L. Prati, F. Porta, J. Catal. 224 (2004) 397.
18] L. Prati, F. Porta, Appl. Catal. A: Gen. 291 (2005) 199–203.
[
[
[
[19] I.V. Deliy, I.L. Simakova, N. Pavasio, R. Psaro, Appl. Catal. A: Gen. 357 (2009) 170.
[
[
20] I.A. Tarkovskaya, Okislenyi Ugol (Oxidized Carbon), Nauk, Dumka, Kiev, 1981.
21] E.V. Murzina, A.V. Tokarev, K. Kord a´ s, H. Karhu, J.-P. Mikkola, D.Yu. Murzin, Catal.
Today 131 (2008) 385.
[22] A. Bernas, N. Kumar, P. M a¨ ki-Arvela, N.V. Kul’kova, B. Holmbom, T. Salmi, D.Yu.
Murzin, Appl. Catal. A: Gen. 245 (2003) 257.
23] F. Moreau, G.C. Bond, Catal. Today 122 (2007) 260.
24] A. Sanchez, S. Abbet, U. Heiz, W.-D. Schneider, H. Hakkinen, R.N. Barnett, U.
Landman, J. Phys. Chem. A 103 (1999) 9573.
[
[
[25] J. Radnik, C. Mohr, P. Claus, Phys. Chem. Chem. Phys. 5 (2003) 172.
References
[26] G.C. Bond, in: B. Imelik, C. Naccache, G. Courier, H. Praliaud, P. Meriaudeau, P.
Gallezot, G.A. Martin, J.C. Vedrine (Eds.), Metal–Support and Metal Additive
Effects in Catalysis, Elsevier, Amsterdam, 1982, , ch. 1.
[
1] A. Bernas, P. Lukkanen, N. Kumar, P. M a¨ ki-Arvela, J. V a¨ yrynen, E. Laine, B.
Holmbom, T. Salmi, D.Yu. Murzin, J. Catal. 210 (2002) 354.
[27] B. Cornils, W.A. Herrmann, R. Schlogl, C.-H. Wong (Eds.), Catalysis from A to Z—A
Concise Encyclopedia, Wiley-VCH, Weinheim, 2000.
[
2] P.R. O’Quinn, J.L. Nelssen, R.D. Goodband, M.D. Tokach, Anim. Health Res. Rev. 1
[
[
28] M.L. Poutsma, Am. Chem. Soc. Monogr. 171 (1971) 431.
29] D. Mukesh, C. Narasimhan, V.M. Deshpande, K. Ramnarayan, Ind. Eng. Chem. Res.
(2000) 35.
[
[
3] L.D. Whigham, M.E. Cook, R.L. Atkinson, Pharmacol. Res. 42 (2000) 503.
4] S.F. Chin, W. Liu, M. Storkson, Y.L. Ha, M.W. Pariza, J. Food Compos. Anal. 5 (1992)
27 (1988) 409.
[30] R. Touroude, F.G. Gault, J. Catal. 32 (1974) 294.
1
85.