Direct conversion of linoleic acid over silver catalysts in the presence of H2: An unusual way towards conjugated linoleic acids

Article abstract of DOI:10.1002/anie.200501852

(Chemical Equation Presented) A silver lining: The isomerization of conjugated double bonds over silver catalysts in the presence of hydrogen are unusual and hitherto unknown. Silver catalysts control the selectivity of the extremely difficult isomerization of linoleic acid towards the physiologically important 9-cis,11-trans- and 10-trans, 12-cis-unsaturated isomers, which are anticarcinogenic and antiarteriosclerotic agents (see scheme, CLA = conjugated linoleic acids).

Full text of DOI:10.1002/anie.200501852

Communications
Isomerizations
DOI: 10.1002/anie.200501852
Direct Conversion of Linoleic Acid over Silver
Catalysts in the Presence of H2: An Unusual Way
towards Conjugated Linoleic Acids
Markus Kreich and Peter Claus*
Dedicated to Professor Gerhard Zimmermann
on the occasion of his 75th birthday
In the search for anticarcinogenic substances towards the end
of the 1980s, isomers of linoleic acid were discovered in beef
and dairy products and proved to be potential mutagen
[*] Dipl.-Ing. M. Kreich, Prof. Dr. P. Claus
Ernst-Berl-Institut/Technical Chemistry II
Technical University of Darmstadt
64287 Darmstadt (Germany)
Fax : (+49)6151-16-4788
E-mail: claus@ct.chemie.tu-darmstadt.de
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
7800
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7800 –7804

Angewandte
Chemie
inhibitors.[1,2] Owing to the conjugated double bonds present
in these isomers, they were called CLAs (conjugated linoleic
acids). Food that is modified by the addition of CLAs (so-
called functional foods) offers specific benefits and advan-
tages for human health.[3] Although functional food has been
at the center of traditional Chinese medicine for over 2000
years, the amount of research has only increased since the
physiological benefits CLAs were proved. Most of this
research has been in the field of food chemistry and is
particularly anchored in a framework program of the Euro-
pean Union.[4,5] These studies have shown that CLAs exhibit a
variety of positive qualities. Besides their anticarcinogenic
and antioxidative qualities, they also influence the fat and
muscle content in the body and display antiarteriosclerotic
qualities. The latest theories attribute these qualities to the 9-
cis,11-trans and the 10-trans,12-cis isomer.[6–10] The influence
of the other CLA isomers on these qualities are largely
unknown.
The established and industrial approach to the production
of CLA is isomerization under alkaline conditions, which
converts linoleic acids and their alkyl esters with alkali bases
or potassium alkoxides into CLAs.[11] Subsequent neutraliza-
tion with acid—typically phosphoric acid—is necessary. The
alkali bases, solvents (e.g. DMSOand propylene glycol), and
phosphoric acid are disadvantageous from an ecological and
economic point of view. The enzyme-catalyzed production of
CLA is known from biochemistry;[12] in nature, this process
occurs in the rumen of cattle by means of Butyrivibrio
fibrisolvens.[13,14] Homogeneous catalysts such as [{RhCl-
(C8H14)2}2] and [RhCl(PPh3)3] have also been applied in the
isomerization of soybean oil.[15]
other hand, direct CLA production is a very complex and
difficult process. It is easy to find a catalyst that hydrogenates
linoleic acid directly to stearic acid. In contrast, the search for
a heterogeneous catalyst that favors selective isomerization to
the physiologically important 9-cis,11-trans and 10-trans,12-
cis-CLAs has proved to be much more challenging
(Scheme 1).
In experiments described in the literature with heteroge-
neous ruthenium catalysts supported on Al2O3 and carbon,
the catalyst was first covered with hydrogen, and linoleic acid
was subsequently converted into CLA under nitrogen.[16] In
order to produce CLA at all, this two-step reaction is
necessary owing to the catalyst properties and the subsequent
reactions competing with the desired isomerization. On the
one hand, the excellent hydrogenation qualities of ruthenium
catalysts results in the complete and rapid conversion of
linoleic acid under hydrogen to give to stearic acid via oleic
acid. On the other hand, it is clear that only a small quantity of
hydrogen is required for the isomerization of linoleic acid into
CLAs. It is therefore a fine line between the high activity of
the catalyst, but under formation of unwanted hydrogenation
products (oleic acid, stearic acid), and an increased selectivity
towards CLAs controlled by the surface concentration of
hydrogen over the catalyst.
Although we first achieved good results with ruthenium
catalysts (Ru/C, Ru/Al2O3) in agreement with literature
reports,[16] the two-step reaction (chemisorptive preadsorp-
tion of hydrogen onto the Ru catalyst followed by deactiva-
tion and chemical reaction in the semi-batch-mode under N2)
was a constant point of criticism in our considerations. The
use of typical hydrogenation metal catalysts (e.g. Ni) prom-
ised no improvement, as was already discussed amply in the
literature.[16c]
Very little work has been carried out on the production of
CLA under heterogeneous catalysis, in which the problem of
catalyst separation is avoided. This is due to the fact that the
CLA have only received attention in the last few years as a
result of investigations in the pharmaceutical field. On the
We describe herein a completely new and highly selective
method for the synthesis of CLAs over heterogeneous silver
catalysts and in the constant presence of hydrogen. According
Scheme 1. Reaction network for the hydrogenation/isomerization of linoleic acid.
Angew. Chem. Int. Ed. 2005, 44, 7800 –7804
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7801

Communications
to the literature, silver is the metal with the lowest hydrogen-
linoleic acid was carried out without previous contact of the
Ag catalyst with hydrogen; this occurred primarily at the
beginning of the reaction by switching from nitrogen to H2.
These experiments delivered the same degrees of conversion
and selectivity as those with preactivation.
binding energy; experimental results and DFT calculations,
although for Ag single crystals, showed weakly bound hydro-
gen on Ag.[17,18] Moreover, our work in the field of selective
hydrogenation of polyunsaturated organic substances shows a
[19]
=
much lower C C hydrogenation activity over silver. This
To understand the different effects of hydrogen on the Ru
catalyst (hydrogenation to stearic and oleic acid) and the Ag
catalyst (isomerization to 9-cis,11-trans- and 10-trans,12-cis-
CLA), calorimetric experiments of H2 adsorption were
carried out with these catalysts. As seen in Figure 1a, at
first an irreversible hydrogen adsorption occurs at the
ruthenium catalyst. A reversible adsorption can be recog-
nized only beyond the appearance of a saturation point (after
the sixth pulse). The adsorption heat of the hydrogen
determined from the calorimetric experiments has been
reactivity is the exact prerequisite of the reaction network,
which calls for a suppression of the parallel and/or consec-
utive hydrogenations to oleic acid and stearic acid. During
isomerization experiments with silver catalysts carried out in
the same way as those with ruthenium catalysts (i.e. under
nitrogen after preactivation of the catalyst with hydrogen), we
did not observe any conversions, as expected. As the success
of the reaction depends on the surface coverage by hydrogen,
as mentioned above, it seemed reasonable in the case of a
metal such as silver to change the reaction procedure and to
convert linoleic acid in the semi-batch-mode under hydrogen.
At first, the catalyst was preactivated with hydrogen as in the
case of the Ru catalysts.
We achieved surprisingly good results with respect to the
direct synthesis of CLA (Table 1). Despite the simple reaction
procedure under hydrogen, linoleic acid underwent 90%
conversion on average after 90 min over Ag/SiO2 using
different grain fractions. The selectivity towards CLAs
reached values between 60 and 67%. This shows at the
same time that the particle size does not exhibit a noticeable
influence on the catalytic properties. Therefore, there is no
evidence of mass transport limitation in the examined area of
ꢀ 500 mm. The formation of the 9-trans,11-trans isomer was
preferred at high conversions. The linoleic acid conversion
increased with the reaction temperature, thus the physiolog-
ically important 9-cis,11-trans- and 10-trans,12-cis isomers
were always the main components of the products. Further-
more, it was observed that the selectivity towards the
hydrogenation products oleic acid and stearic acid decreased
with decreasing substrate/catalyst mass ratio. Consequently, a
desirable practical result is the increase in isomerization/
hydrogenation ratio. From the selectivity–conversion data,
CLA yields of up to 60% can be calculated, depending on the
reaction conditions applied. Furthermore, the conversion of
estimated to about 90 kJmolÀ1 from which a Ru H bond
À
strength of approximately 260 kJmolÀ1 can be calculated.[20]
On the silver catalyst (Figure 1B) only reversible hydrogen
adsorption appears; an irreversible chemisorption could not
be verified.[21] Therefore, we can distinguish between strongly
adsorbed hydrogen on ruthenium and weakly adsorbed
hydrogen in the case of the silver catalyst.[22] These findings
suggest that the differences in the reactivity/selectivity
patterns (hydrogenation activity on Ru versus isomerization
activity on Ag) are based on differences in the hydrogen
adsorption behavior of the active metal catalyst. It is clear
that the weakly bound hydrogen plays a key role in the
selective isomerization of conjugated double bonds (to CLA).
A very similar situation is present in the case of the control of
the intramolecular selectivity with respect to the hydrogena-
[19]
= À =
tion of conjugated C C C Odouble bonds.
With weakly
chemisorbed hydrogen, a discrimination of the hydrogenation
=
=
of the C O- in contrast to the C C-bond is possible; the latter
is hardly hydrogenated by the weakly bonded hydrogen.[19,23]
For the conversion of linoleic acid in the presence of H2, an
addition–elimination mechanism according to Horiuti–Pola-
nyi[16g,20] is assumed, which is typical for the isomerization of
=
C C-double bonds in alkenes and alkadienes (Scheme 2).
According to this mechanism, the selectivity of the reaction is
decided by the type of reaction of the half-hydrogenated
Table 1: Catalytic properties of Ag/SiO2 in the conversion of linoleic acid.[a]
Conditions
XLS [%]
S9-cis,11-trans [%]
S10-trans,12-cis [%]
S9-trans,11-trans [%]
Shydr.[e] [%]
SNP[f] [%]
SCLA[g] [%]
CLA/Hydr.
Variation of catalyst particle size:[b]
<63 mm
63–200 mm
200–500 mm
90
91
86
15
20
20
13
14
14
32
33
33
19
17
20
21
16
13
60
67
67
3.2
3.9
3.4
Variation of the temperature:[c]
383 K
398 K
438 K
35
44
69
25
29
35
19
19
26
15
13
20
31
33
12
10
6
7
59
61
81
1.9
1.8
6.8
Variation of the substrate/catalyst mass ratio:[d]
2
0.5
0.24
16
38
86
7
29
20
7
20
13
14
24
32
61
27
19
11
–
16
28
73
65
0.5
2.7
3.4
[a] t=90 min; XLS=conversion; Si=selectivity. [b] T=438 K,
m
cat =0.8 g. [c] mcat =0.8 g,
dcat =200–500 mm. [d] T=438 K,
d
cat =200–500 mm.
[e] Selectivity towards hydrogenation products (oleic acid + stearic acid). [f] Unidentified products. [g] Sum of CLAs. Hydr.=hydrogenation.
7802 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2005, 44, 7800 –7804

Angewandte
Chemie
case of catalysts with high affinity to hydrogen chemisorption
(Ru) versus elimination of a hydrogen atom with desired
CLA formation in the case of weakly bonded hydrogen (Ag).
Investigations are currently underway to examine the activa-
tion/adsorption of hydrogen on silver catalysts, which is rather
unusual and is in contrast to the oxygen chemisorption known
from oxidation catalysis, but comparable with gold cata-
lysts.[24] Transient experiments with a Ag/SiO2 catalyst pre-
conditioned with H2 in a TAP (temporal analysis of products)
reactor and subsequent D2 pulses revealed the formation of
H–D and, thus, that H2 dissociation must occur. This is not
noticed over pure silver or SiO2.[25]
The results presented herein show not only a direct route
to CLAs but also a completely new method for the isomer-
ization of (conjugated) double bonds. The prospects for the
isomerization of linoleic acid to CLA in the presence of
hydrogen and silver catalysts are obvious. After optimization
of important catalyst properties (e.g. support material, silver
particle size, and metal loading) the high CLA selectivity and
yield as well as the simple reaction procedure in the presence
of hydrogen permit the assignment to a continuous operation
coupled with simple catalyst separation. Despite the high
CLA selectivity, the product spectrum has to be further
improved with respect to potential use in functional food. In
particular, the fraction of 9-trans,11-trans-CLA must be
decreased. Fundamentally, nothing should stand in the way
of applying the presented principle of direct conversion of
linoleic acid into CLAs over silver catalysts in the presence of
H2 as an alternative for the alkaline isomerization.
Figure 1. Results of calorimetric measurements of H2 adsorption over
Ru/C(a) and Ag/SiO 2 (b) catalysts in the presence of H2 (Tads =293 K
(a), 423 K (b)). The heat flow of the DSC(left ordinate) and the H
2
signal of the mass spectrometer (right ordinate) are shown. DSC=dif-
ferential scanning calorimeter.
Experimental Section
General: The experiments were carried out in a 250-mL four-necked
quartz-glass reactor, which was equipped with a reflux condenser, a
drip funnel, a thermometer, a stirrer bar, and a gas-inlet tube. A
magnetic stirrer with an oil bath was used as for heating and stirring.
The Ag/SiO2 catalyst was synthesized by incipient wetness impreg-
nation from silver lactate (Fluka 85210) and SiO2 (Alfa Aesar, “large
pore”) (Ag-content via ICP-OES: 7.7 wt%). The catalyst was dried at
353 K and reduced at 598 K under H2 flow. A TEM examination
(Philips CM 200 UT) resulted in an average Ag particle size of 14 nm.
The Ru/C catalyst was a commercial product of Fluka (Ru content:
5 wt%). The calorimetric measurements of H2 adsorption were
carried out with a DSC 111 (Setaram) with mass spectrometer
(Pfeiffer Vakuum) coupled to a pulse apparatus.
Example of an isomerization experiment: The catalyst (mcat
=
800 mg, grain fraction dcat = 200–500 mm) was introduced into the
reactor, and the mixture of linoleic acid (Aldrich, purity > 99%) and
n-decane was placed in the dropping funnel (V= 70 mL, cLA
=
0.01 molLÀ1). Nitrogen (100 mLminÀ1) was flushed through the
apparatus for 15 min to create an inert atmosphere and for degassing
the linoleic acid/n-decane mixture. The catalyst was then activated
under hydrogen flow (100 mLminÀ1) for 1.5 h (including heating
time) at 438 K. The linoleic acid/n-decane mixture was then added to
the catalyst; this represents the starting time of the reaction (t = 0).
The reaction mixture was kept at a reaction temperature of 438 K
under permanent stirring (1100 rpm). Hydrogen (100 mLminÀ1) was
flushed through the mixture as reaction gas. Subsequent analysis was
performed by means of temperature-programmed capillary gas
chromatography (with FID) through an Agilent HP-5 column (l =
25 m, di = 0.20 mm, tf = 0.11 mm), carried out according to the
literature.[16b] Heptadecanoic acid was used as an internal GC
standard. The samples were concentrated in a heating block and
Scheme 2. Horiuti–Polanyi mechanism for the hydrogenation/isomer-
ization of linoleic acid: weakly bonded hydrogen (on Ag) causes the
desired CLA formation through an addition/elimination mechanism,
whereas strongly bonded hydrogen (on Ru) causes the formation of
oleic acid through consecutive H addition.
intermediate: addition of a further chemisorbed hydrogen
atom under hydrogenation to give undesired oleic acid in the
Angew. Chem. Int. Ed. 2005, 44, 7800 –7804
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7803

Communications
silylated with bis(trimethylsilyl)trifluoroacetamide (BSTFA) (80 mL)
[12] M. W. Pariza, X.-Y. Yang, US 5 856 149, 1999 [Chem. Abstr. 1999,
130, 94531].
[13] J. R. Kepler, K. P. Hirons, J. J. McNeill, S. B. Tove, J. Biol. Chem.
1966, 241, 1350 – 1354.
and trimethylchlorosilane (TMCS) (40 mL). Details of the analysis as
well as a chromatogram example can be found in the Supporting
Information.
[14] T. Y. Lin, C.-W. Lin, Y.-J. Wang, Food Chem. 2003, 83, 27 – 31.
[15] X. Dong, S. Chung, Ch. K. Reddy, L. E. Ehlers, J. Am. Oil Chem.
Soc. 2001, 78, 447 – 453.
Received: May 27, 2005
Revised: August 3, 2005
[16] a) A. Bernas, N. Kumar, P. Mꢁki-Arvela, E. Laine, B. Holmbom,
T. Salmi, D. Murzin, Chem. Commun. 2002, 10, 1142 – 1143;
b) A. Bernas, P. Laukkanen, N. Kumar, P. Mꢁki-Arvela, J.
Vꢁyrynen, E. Laine, B. Holmbom, T. Salmi, D. Murzin, J. Catal.
2002, 210, 354 – 366; c) A. Bernas, N. Kumar, P. Mꢁki-Arvela,
N. V. Kulꢀkova, B. Holmbom, T. Salmi, D. Murzin, Appl. Catal. A
2003, 245, 257 – 275; d) A. Bernas, P. Mꢁki-Arvela, N. Kumar, B.
Holmbom, T. Salmi, D. Murzin, Ind. Eng. Chem. Res. 2003, 42,
718 – 727; e) A. Bernas, D. Murzin, React. Kinet. Catal. Lett.
2003, 78, 3 – 10; f) A. Bernas, N. Kumar, P. Laukkanen, J.
Vꢁyrynen, T. Salmi, D. Murzin, Appl. Catal. A 2004, 267, 121 –
133; g) A. Bernas, N. Kumar, P. Mꢁki-Arvela, B. Holmbom, T.
Salmi, D. Murzin, Org. Process Res. Dev. 2004, 8, 341 – 352.
[17] J. Greeley, M. Mavrikakis, J. Phys. Chem. B 2005, 109, 3460 –
3471.
Keywords: conjugated linoleic acids · functional foods ·
heterogeneous catalysis · hydrogen · isomerization · silver
.
[1] Y. L. Ha, N. K. Grimm, M. W. Pariza, Carcinogenesis 1987, 8,
1881 – 1887.
[2] Y. L. Ha, N. K. Grimm, M. W. Pariza, J. Agric. Food Chem. 1989,
37, 75 – 81.
[3] Ernährungsumschau 1995, 42, 452.
[4] “Lebensmittelchemie 2000”: J. Fritsche, Nachr. Chem. 2001, 49,
374 – 381.
[5] European Commission, Fifth Framework Program, KA1–
FOOD, NUTRITION & HEALTH EUR 19422: FUNCLA:
Conjugated linoleic acid (CLA) in functional food: A potential
benefit for overweight middle-aged Europeans, http://europa.
eu.int/comm/research/quality-of-life/ka1/volume2/qlk1-1999-
00076.htm
[18] “Hydrogen Sorption on Pure Metal Surfaces”: C. Christmann in
Hydrogen Effects in Catalysis, Fundamentals and Practical
Applications (Eds.: Z. Paꢂl, P. G. Menon), Marcel Dekker,
New York, 1988, p. 12.
[6] P. R. OꢀQuinn, J. L. Nelssen, R. D. Goodband, M. D. Tokach,
Anim. Health Res. Rev. 2000, 1, 35 – 46.
[7] L. D. Whigham, M. E. Cook, R. L. Atkinson, Pharmacol. Res.
2000, 42, 503 – 510.
[8] S. F. Chin, W. Liu, M. Storkson, Y. L. Ha, M. W. Pariza, J. Food
Compos. Anal. 1992, 5, 185 – 197.
[9] V. Mougios, A. Matsakas, A. Petridou, S. Ring, A. Sagredos, A.
Melissopoulou, N. Tsigilis, M. Nikolaidis, J. Nutr. Biochem. 2001,
12, 585 – 594.
[19] P. Claus, Top. Catal. 1998, 5, 51 – 62.
[20] The Microkinetics of Heterogeneous Catalysis (Eds.: J. A.
Dumesic, D. F. Rudd, L. M. Aparicio, J. E. Rekoske, A. A.
Trevino), ACS Professional Reference Book, Washington, DC,
1993.
[21] The detailed quantitative evaluation shows that the areas under
the peaks for adsorption (exothermic signal) and desorption
(endothermic signal) are identical.
[22] This is also confirmed by TPD experiments: Whereas a hydro-
gen desorption peak appears at the ruthenium catalyst at 1258C,
the silver catalyst does not show any corresponding signal.
[23] T. Takeuchi, T. Asano, Z. Phys. Chem. 1963, 36, 118 – 125.
[24] Special Issue “Catalysis by Gold” (Eds.: G. Hutchings, M.
Haruta), P. Claus, Appl. Catal. A 2005, 251, 222 – 229.
[25] M. Bron, E. Kondratenko, A. Trunschke, P. Claus, Z. Phys.
Chem. 2004, 218, 405 – 423.
[10] J. D. Palombo, A. Ganguly, B. R. Bistrian, M. P. Menard, Cancer
Lett. 2002, 177, 163 – 172.
[11] a) S. Busch, L. Zander, W. Albiez, P. Horlacher, A. Westfechtel
(Cognis), DE 102 36 086A1, 2002 [Chem. Abstr. 2004, 140,
183582]; b) P. Horlacher, K.-H. Ruf, F. Timmermann, W. Adams,
R. von Kries, (Cognis), DE 102 59 157A1, 2002 [Chem. Abstr.
2004, 141, 73307].
7804 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7800 –7804