Hydrothermal decarboxylation and hydrogenation of fatty acids over Pt/C

Article abstract of DOI:10.1002/cssc.201000370

We report herein on the conversion of saturated and unsaturated fatty acids to alkanes over Pt/C in high-temperature water. The reactions were done with no added H₂. The saturated fatty acids (stearic, palmitic, and lauric acid) gave the corresponding decarboxylation products (n-alkanes) with greater than 90% selectivity, and the formation rates were independent of the fatty acid carbon number. The unsaturated fatty acids (oleic and linoleic acid) exhibited low selectivities to the decarboxylation product. Rather, the main pathway was hydrogenation to from stearic acid, the corresponding saturated fatty acid. This compound then underwent decarboxylation to form heptadecane. On the basis of these results, it appears that this reaction system promotes in situ H ₂ formation. This hydrothermal decarboxylation route represents a new path for using renewable resources to make molecules with value as liquid transportation fuels.

Full text of DOI:10.1002/cssc.201000370

DOI: 10.1002/cssc.201000370
Hydrothermal Decarboxylation and Hydrogenation of
Fatty Acids over Pt/C
[a, b]
[b]
[a]
Jie Fu,
Xiuyang Lu, and Phillip E. Savage*
We report herein on the conversion of saturated and unsatu-
rated fatty acids to alkanes over Pt/C in high-temperature
decarboxylation product. Rather, the main pathway was hydro-
genation to from stearic acid, the corresponding saturated
fatty acid. This compound then underwent decarboxylation to
form heptadecane. On the basis of these results, it appears
water. The reactions were done with no added H . The saturat-
2
ed fatty acids (stearic, palmitic, and lauric acid) gave the corre-
sponding decarboxylation products (n-alkanes) with greater
than 90% selectivity, and the formation rates were independ-
ent of the fatty acid carbon number. The unsaturated fatty
acids (oleic and linoleic acid) exhibited low selectivities to the
that this reaction system promotes in situ H formation. This
2
hydrothermal decarboxylation route represents a new path for
using renewable resources to make molecules with value as
liquid transportation fuels.
Introduction
Natural oils and fats (e.g., plant and seed oils) are complex
mixtures of triglycerides, which consist of a glycerol backbone
with three fatty acid moieties. The fatty acids can be either sa-
turated or unsaturated, and the carbon numbers range from
about C8 to C24, but C12, C16, and C18 fatty acids tend to be
the most abundant. These natural materials can serve as feed-
stocks for the production of renewable liquid transportation
fuels, such as biodiesel (e.g., fatty acid methyl esters) or green
diesel (e.g., hydrocarbons). There is considerable interest in
producing hydrocarbons from the fatty acid moieties in trigly-
cerides because hydrocarbon molecules would, in principle, be
compatible with the existing distribution infrastructure devel-
oped for petroleum-derived liquid transportation fuels.
than H O. Thus, stoichiometrically no H is required for the re-
2
2
action. Moreover, the loss of a CO molecule does not lead to
2
a loss in any of the chemical energy in the molecule. In fact, it
produces a hydrocarbon molecule with an even higher energy
density. This path is an attractive one for making renewable
[3]
hydrocarbon fuels, but aside from our recent report and a set
[18]
of earlier catalyst screening experiments, all previous work
on fatty acid decarboxylation has been done in organic reac-
tion media, such as dodecane or mesitylene. In contrast, the
focus herein is on fatty acid decarboxylation in water. This hy-
drothermal decarboxylation is technologically important, but
the reaction pathways are essentially unexplored.
There are many biofuel production schemes in which fatty
acids are produced in an aqueous phase. One example is
Kusdiana and Saka’s two-step process for handling low-cost tri-
Recent review articles have highlighted the significant re-
search and development (R&D) efforts that have been devoted
to using and adapting petroleum hydrotreatment technology
[19]
glyceride-containing feedstocks. The first step is hydrolysis
of the triglycerides to produce an aqueous stream of fatty
acids. They proposed to recover the fatty acids and then con-
vert them to biodiesel, but one could also, and perhaps even
more simply, perform a hydrothermal catalytic decarboxylation
to produce alkanes. This alternate second step would eliminate
the need to separate the fatty acids from water and it would
produce a hydrocarbon biofuel. Another aqueous phase pro-
[1,2]
to convert triglycerides and fatty acids into hydrocarbons.
These efforts typically involve conventional hydrotreatment
catalysts (e.g., CoMo/Al O3 NiMo/Al O ) and reaction condi-
2
2
3
tions. A high H2 partial pressure is required to remove the
oxygen atoms in the fatty acids as H O molecules. UOP/Eni
2
and Neste Oil have developed hydrotreatment processes for
vegetable and seed oils that can operate on a large scale. The
[20]
high H consumption associated with these processes is their
cess is that of Centia for making biofuel from triglycerides.
This process, now in scaleup by Diversified Energy, uses a hy-
2
main drawback. H is not currently available in large quantities
2
from renewable resources and H costs can be high. Moreover,
2
H is made primarily from steam reforming of natural gas and
2
[
a] Dr. J. Fu, Prof. P. E. Savage
Department of Chemical Engineering
University of Michigan
Ann Arbor, MI, 48109 (USA)
Fax: (+1)(734)763-0459
CO is the byproduct. Thus, a near-term process for the pro-
2
duction of fully renewable biofuels needs to operate without
added H2.
Recognizing the advantages of both a hydrocarbon biofuel
E-mail: psavage@umich.edu
and reducing the amount of H needed to produce it from
2
[
b] Dr. J. Fu, Prof. X. Lu
plant oils, several labs have been exploring decarboxylation
Department of Chemical and Biological Engineering
Zhejiang University
[
3–17]
chemistry as an alternative to hydrodeoxygenation.
Decar-
boxylation removes the O atoms in fatty acids as CO rather
Hangzhou 310027 (PR China)
2
ChemSusChem 2011, 4, 481 – 486
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
481

P. E. Savage et al.
drolysis step to convert triglycerides into free fatty acids. Again
the fatty acids are separated from the water and then decar-
boxylated in an organic solvent. The process can probably be
simplified by omitting the separation step and performing the
decarboxylation in water. There are also recent reports on dif-
ferent hydrothermal processes for converting the lipids in wet
Unless otherwise indicated, each experiment was conducted at
3
308C with 0.176 mmol of fatty acid and 5 mg Pt/C in each reactor.
The amount of water loaded was such that the expanded liquid
aqueous phase occupied 95% of the reactor volume under the re-
[25]
action conditions, based on the saturation densities
of pure
water at the reaction temperatures. After being loaded, each reac-
tor was sealed by attaching and then tightening the reactor cap.
The reactor headspace contained air, but the amount of residual
[
21]
algal biomass to either fatty solids or crude bio-oils with
[
22,23]
very high fatty acid contents.
Rather than separating the
O present was enough to oxidize completely no more than 0.2%
2
water from the solids or bio-crudes, perhaps hydrothermal de-
carboxylation could be employed to produce a hydrocarbon
(by moles) of the fatty acid. No H
any of these experiments.
was added to the reactors in
2
[
24]
Sealed reactors containing fatty acid, Pt/C, and water were placed
in a Techne SBL-2 fluidized sand bath preheated to the desired re-
action temperature, controlled to within 18C. The time required for
these reactors to reach reaction temperature was roughly 2 min.
After the desired reaction time was completed, the reactors were
removed from the sand bath and submerged in a water bath at
ambient temperature to quench the reaction. The reactor contents
were transferred to a volumetric flask, and the reactors were subse-
quently rinsed with repeated acetone washes until the total
volume collected was 10 mL.
fuel. Finally, Li et al. reported a catalytic hydrothermal con-
version process for triglycerides, but again, they separated the
fatty acids produced by hydrolysis from the aqueous phase
and processed them in an organic solvent instead. It is clear
that there are many potential biofuel processes under develop-
ment in which aqueous streams of fatty acids are produced.
Our previous article showed that Pt/C and Pd/C catalysts are
active for decarboxylation of a saturated fatty acid in an aque-
[
3]
ous medium, with Pt/C showing higher activity. Moreover,
The reaction products were identified by fragmentation patterns
from an Agilent 5970 mass spectrometric detector and by match-
ing gas chromatograph retention times with known standards.
Product separation for saturated fatty acids was achieved by using
an Agilent 6890 gas chromatograph equipped with a 50 mꢁ
0.2 mmꢁ0.33 mm capillary column (HP-5 for flame ionization de-
tection (FID) and HP-5MS for MS detection). The injection port tem-
the work showed that no added H was required to do the de-
2
carboxylation. Herein, we provide more detailed information
about the hydrothermal reaction pathways and kinetics for this
catalyzed transformation.
Watanabe and co-workers provide the only other report on
[
18]
hydrothermal decarboxylation of fatty acids.
Some of the
perature was 3258C and the temperature program consisted of a
materials they tested (e.g., NaOH, KOH) on stearic acid in su-
percritical water at 4008C produced the decarboxylation prod-
uct (C17 alkane). Other materials (e.g., CeO , Y O , ZrO ) pro-
À1
5
min soak at 408C followed by a 108Cmin ramp up to a final
temperature of 2908C, which was then held for 15 min. For the un-
saturated fatty acids, product separation was achieved by using an
Agilent 6890 GC equipped with a 30 mꢁ0.32 mmꢁ0.25 mm Nukol
capillary column. The injection port temperature was 2208C and
the temperature program consisted of a 2 min soak at 608C fol-
2
2
3
2
duced the C16 alkene. None of the materials they tested were
as active as Pt/C.
Herein, we report on the catalytic hydrothermal decarboxyla-
tion of five different fatty acids that are prevalent in nature:
stearic (18:0), palmitic (16:0), lauric (12:0), oleic (18:1), and lino-
leic (18:2) acids. The numbers in parentheses indicate, respec-
tively, the number of carbon atoms and number of double
bonds in each of the fatty acids. This suite of compounds al-
lowed us to determine the influence of the fatty acid carbon
chain length and its degree of unsaturation on the Pt/C-cata-
lyzed hydrothermal decarboxylation of fatty acids.
À1
lowed by a 108Cmin ramp up to a final temperature of 2008C,
which was then held for 55 min.
Quantitative analysis was accomplished by generating and using
calibration curves for each compound of interest. The FID response
was linear for each component over the concentration ranges
used. Minor products for which we did not have authentic stand-
ards were quantified by assuming their detector responses were
similar to those of a compound that eluted with a similar retention
time and for which we had an experimental calibration curve.
Product yields were calculated as the amount (moles or mass) of
product recovered divided by the initial amount (moles or mass) of
fatty acid loaded into the reactor. Selectivities were calculated as
the number of moles of product recovered divided by the number
of moles of fatty acid that had reacted (i.e., molar yield/conver-
sion). Uncertainties reported herein are standard deviations, which
were determined by replicating the experiments. Each data point
represents the mean result from at least three independent experi-
ments.
Experimental Section
We conducted reactions in unstirred mini-batch reactors assembled
from 3/8 inch stainless steel Swagelok parts. Port connectors
sealed with a cap on each end gave a reactor volume of about
1
.67 mL. Prior to use in any experiments, reactors were rinsed with
acetone, dried, loaded with water (1.1 mL), and conditioned at
008C for 30 min to remove any residual materials remaining from
3
the manufacture of the metal parts. Stearic, palmitic, lauric, oleic,
and linoleic acids; heptadecane; undecane; and 5% Pt/C catalyst
were purchased from Sigma–Aldrich. Pentadecane was purchased
from Wiley. We have previously reported details about the catalyst
activity maintenance, surface area, pore size, and metal disper-
Results and Discussion
This section presents and interprets the experimental results
and places them in the context of previous work done on the
catalytic decarboxylation of both saturated and unsaturated
fatty acids over Pt/C. Our earlier analysis used the Weisz–Prater
criterion to show that the experimental system we used can
provide access to intrinsic reactivities unfettered by mass trans-
[
3]
sion. Briefly, the catalyst particle diameter was 50 mm, the surface
2
À1
area of the fresh catalyst was 1483 m g , and the metal dispersion
was 39%. HPLC grade acetone was obtained from Fisher Scientific.
All chemicals were used as received.
4
82
www.chemsuschem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2011, 4, 481 – 486

Decarboxylation and Hydrogenation of Fatty Acids
[
3]
port limitations. Moreover, our previous work showed that
the catalyst can be reused at least three times with no appreci-
able loss in activity.
a discernible influence on the decarboxylation rate. The cata-
lytic chemistry occurs at the carboxylic acid end of the mole-
cule and one would expect those catalyst–molecule interac-
tions to be largely insensitive to the addition or removal of a
few carbon atoms from the other end of the fatty acid mole-
cule. Moreover, recent work on fatty acid decarboxylation over
Pd catalysts in organic solvents has also demonstrated the in-
sensitivity of the reaction rate to the fatty acid chain
Decarboxylation of saturated fatty acids
Figures 1 and 2 show the temporal variation of the molar yield
of and the selectivity to the C
alkane that arises from the
NÀ1
[7,17]
decarboxylation of the C saturated fatty acids at 3308C. The
length.
N
Decarboxylation of unsaturated C18 fatty acids
Having found that the decarboxylation of saturated fatty acids
was insensitive to the fatty acid alkyl chain length (at least
within the C12–C18 range), we next explored the influence of
the degree of unsaturation on decarboxylation. Figures 3 and
4
show the temporal variation of the molar yield of and selec-
Figure 1. Temporal variation of alkane molar yields from decarboxylation of
stearic acid (
&
), palmitic acid (
), and lauric acid (~).
&
Figure 3. Temporal variation of heptadecane molar yields from stearic acid
), oleic acid (
), and linoleic acid (~).
(
&
&
Figure 2. Temporal variation of alkane selectivities from decarboxylation of
stearic acid ( ), palmitic acid (
), and lauric acid (~).
&
&
[
3]
data for palmitic acid are from our earlier work, wherein we
used a 0.195 mmol reactant loading. The decarboxylation rates
and selectivities to alkanes are very similar for the three satu-
rated fatty acids, which suggest that the decarboxylation rates
are independent of the carbon number for the saturated fatty
acids we tested. Additionally, the high selectivities (85–95%)
indicated in Figure 2 show that the saturated fatty acids can
be selectively decarboxylated over Pt/C in high-temperature
water. The selectivities were not 100%, however, which indi-
cates that other products were also formed, albeit in much
lower yields. These minor products were primarily C to C n-al-
Figure 4. Temporal variation of heptadecane selectivity from stearic acid (
oleic acid (
), and linoleic acid (~).
&
),
&
tivity to heptadecane, the C
alkane that would form from
NÀ1
the decarboxylation of a C18 fatty acid. Heptadecenes were
also formed in the experiments with the unsaturated fatty
acids, but their yields were always lower than that of heptade-
cane, so we show only the heptadecane data here. Discussion
of the complete product distribution will be deferred until the
next section.
7
N
kanes.
These results extend our previous work by showing that the
C18 fatty acid (stearic) and the C12 fatty acid (lauric) behave
similarly to the C16 acid (palmitic) that we had reported previ-
[
3]
ously. It was expected that the chain length would not have
ChemSusChem 2011, 4, 481 – 486
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
483

P. E. Savage et al.
Figures 3 and 4 show that the degree of unsaturation has a
strong effect on the decarboxylation of fatty acids because the
yields and selectivities from oleic acid (18:1) and linoleic acid
experiments, the H must be formed in situ during the reac-
2
tion. To test this hypothesis of in situ H formation, we con-
2
ducted an experiment wherein oleic acid (0.195 mmol) was re-
acted over Pt/C (10 mg) at 3708C for 3 h in a reactor equipped
with a valve so that we could collect and analyze the gas-
phase reaction products. This experiment revealed a 30% H2
(18:2) to heptadecane are much lower than those from stearic
acid (18:0). Again, note that for both oleic and linoleic acid,
the yield of heptadecane, although low, exceeded the yields of
heptadecenes at all reaction times investigated. Thus, for these
unsaturated fatty acids, it appears that a reaction other than
decarboxylation is the main reaction path. Recall that no H2
was added to the reactor. The next section explores the reac-
tion pathways in more detail.
molar yield, so H is indeed formed in situ during the reaction.
2
CO , CH , and C H were also detected, with molar yields of
2
4
2
6
100, 10, and 5.1%, respectively. Pt catalysts are commonly
used for aqueous-phase reforming of biomass-derived mole-
[
28,29]
cules,
and the gas products we detected are typically ob-
[30]
served during aqueous-phase reforming. Therefore, the Pt/C
catalyst in our system might be playing this role by converting
Oleic acid reaction pathway
a small amount of the oleic acid to H . Of course, a variety of
2
Figure 5 shows the temporal variation of the molar yield of
oleic acid and the major reaction products formed at 3308C.
Stearic acid, the product of oleic acid hydrogenation, is the
reactions such as CÀC bond cleavage and water-gas shift occur
during aqueous-phase reforming to generate H molecules.
2
The rate and selectivity for H production is a function of both
2
the metal catalyst and the support.
Table 1 shows a more complete itemization of the products
identified from the hydrothermal reaction of oleic acid over Pt/
C. In addition to the major products illustrated in Figure 5, the
Table 1. Products from oleic acid reactions over Pt/C at 3308C for 1.5 h
Compound
Yield [wt%]
2
2
-heptanone
-pentanone-4-hydroxy-4-methyl
0.54Æ0.02
0.08Æ0.01
0.05Æ0.01
0.10Æ0.06
0.16Æ0.02
1.3Æ0.5
tetradecane
pentadecane
hexadecane
Figure 5. Temporal variation of product molar yields (stearic acid (
decane (~), and heptadecenes (*)) from oleic acid (^).
&
), hepta-
aromatic derivatives
9
1
3
2
2
-octadecanone
-propene-1-methoxy-2-methyl
-heptadecanone
-heptadecanone
(3H)-furanone-dihydro-5-tetradecyl
0.15Æ0.17
0.08Æ0.02
0.16Æ0.11
0.06Æ0.01
3.9Æ0.1
major product at all times. The yield of stearic acid increases
with time until about 2 h when it appears to plateau or per-
haps even begin to slowly decrease. The molar yield of hepta-
decane increased steadily with reaction time, and it was always
present in higher yields than the heptadecenes (e.g., 8-hepta-
decene, 3-heptadecene), which are the decarboxylation prod-
uct(s) from oleic acid. The trends in Figure 5 suggest that stea-
ric acid is the intermediate product between oleic acid and
heptadecane. That is, oleic acid was hydrogenated first, and
then the stearic acid produced was decarboxylated to hepta-
decane. A similar sequential hydrogenation–decarboxylation
pathway for oleic acid was proposed by Immer et al. for the re-
heptadecane
heptadecenes
stearic acid
oleic acid
9.2Æ2.8
2.6Æ0.5
36.1Æ2.9
31.1Æ8.8
7.0Æ1.2
other
reaction also produced lesser amounts of other products. The
most abundant minor product was a tetradecylfuranone. The
minor products also included ketones; C , C , and C alkanes;
14
15
16
and some aromatic compounds that could not be positively
identified. We believe these are aromatic compounds because
[
9]
action in a dodecane solvent.
Pt is a good hydrogenation catalyst
[
26,27]
and to determine
each contained a strong signal in the mass spectrum at m/z
+
the role of this catalyst on the hydrogenation pathway we car-
ried out experiments with oleic acid at 3308C for 1 h, but with-
out Pt/C. No heptadecane was detected, and the yield to stea-
ric acid was (4.0Æ1.7)%; this is much lower than the yield to
stearic acid in the presence of Pt/C. Nevertheless, this result in-
dicates some capacity for hydrogenation in high-temperature
water alone, perhaps catalyzed by the stainless-steel reactor
walls. Of course, hydrogen is required for hydrogenation to
91, which is indicative of a tropylium ion (C H7 ); a signature
7
signal for alkyl-substituted aromatic compounds. Likewise, the
products we placed in the “other” category of Table 1 are ones
that we were unable to identify. The product amounts in
Table 1 sum to 93 wt%, which is typical of the mass balances
we routinely observed in the oleic acid experiments. Note that
this overall mass balance does not statistically differ from
100% when one considers the uncertainty in the yields of the
individual products.
occur. Since no H was loaded into the reactor in any of these
2
4
84
www.chemsuschem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2011, 4, 481 – 486

Decarboxylation and Hydrogenation of Fatty Acids
Taken collectively, the results in Figure 5 and Table 1 suggest
that aqueous-phase reforming, hydrogenation, and decarboxy-
lation are the main reaction paths for oleic acid over Pt/C.
These combine to convert unsaturated fatty acids to alkanes.
Along with these three reactions, other reactions such as crack-
ing, aromatization, and cyclization occurred over Pt/C in this
hydrothermal medium.
could not positively identify. In doing so, we made the reason-
able assumption that the FID response was proportional to the
mass of carbon atoms in a given compound. Since we do not
have identities for these heavy and light compounds and thus
lack accurate molecular weights, we report the product yields
on a mass rather than molar basis.
Figure 6a shows that oleic acid was always the most abun-
dant product. The heavy products were second in abundance
at reaction times up to 1.5 h. The yield of stearic acid was also
high and it increased with time to become the second most
abundant product after 1.5 h. The yield reached about 25%
after 2.5 h. Figure 6b shows that the yield of heptadecane in-
creased steadily throughout the reaction, reaching about 5%
after 2.5 h. The yield of the light products showed a similar
trend, whereas the yield of heptadecenes was largely stable at
about 1%. Recall that the yield of heptadecenes from oleic
acid showed essentially the same trend.
Linoleic acid reaction pathway
Figure 6 shows the temporal variation of the product yields
from hydrothermal treatment of linoleic acid at 3308C over the
5
% Pt/C catalyst. The conversion of linoleic acid was 97% even
Oleic acid and heavy products being the most abundant
products at short reaction times is consistent with linoleic acid
reacting first to form oleic acid and heavy products. The oleic
acid can then be hydrogenated to form stearic acid, and the
stearic acid undergoes decarboxylation to form heptadecane.
The presence of heptadecenes indicates that decarboxylation
is also a pathway, albeit minor, for linoleic and oleic acids. Ki-
tayama and co-workers found that many kinds of monoenoic
acids, including those with different double-bond positions
and different conformations (trans or cis), formed during linole-
[27]
ic acid hydrogenation over Pt/Al O . The heavy products we
2
3
observed could well correspond to compounds of this type.
Figure 7 provides a schematic depiction of the major overall
Figure 6. a) Temporal variation of mass yields of major products (oleic acid
), stearic acid (~), and heavy products (*)) from linoleic acid. b) Temporal
variation of mass yields of minor products (heptadecane (~), heptadecenes
*), and light products (^)) from linoleic acid.
(
&
(
Figure 7. Hydrothermal catalytic reaction pathways for C18 fatty acids.
at the shortest time investigated, so we did not include the
low yields of remaining linoleic acid in Figure 6. The disappear-
ance of this polyunsaturated fatty acid was much more rapid
than that of the monounsaturated oleic acid. The product
spectrum from this polyunsaturated compound was also more
complex. There were several products that we could not iden-
tify with confidence, even with the assistance of GC-MS analy-
sis. Snꢂre et al. also observed a complicated product spectrum
reaction pathways for hydrothermal decarboxylation of linoleic,
oleic, and stearic acids. The thickness of each arrow in the reac-
tion network is suggestive of the relative importance of that
pathway. The pathways in Figure 7 are the same as those
[8]
noted by Snꢂre et al. for linoleic acid decarboxylation in or-
ganic solvents, so the reaction medium does not appear to in-
fluence the main reaction pathways.
[
8]
when working with linoleic acid. We lumped these products
together as either “heavy” products, which elute from the GC
near the linoleic acid reactant, or “light” products, which elute
near heptadecane. We used the experimentally determined GC
calibrations for oleic acid and heptadecane to quantify the
mass yields of heavy and light products, respectively, that we
The mass balance for the linoleic acid experiments was
about 75%. We suspect that coupling reactions can occur with
this polyunsaturated compound to produce higher-molecular-
weight material that would not elute from the GC and hence
go undetected in our analysis. The literature lends some sup-
port for this hypothesis because Watanabe et al. observed cou-
ChemSusChem 2011, 4, 481 – 486
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
485

P. E. Savage et al.
pling products in their study on fatty acids in supercritical
[2] B. Smith, H. C. Greenwell, A. Whiting, Energy Environ. Sci. 2009, 2, 262–
[
18]
271.
water. Snꢂre et al. also reported the formation of high-mo-
lecular-weight products from deoxygenation of unsaturated
[
[
3] J. Fu, X. Lu, P. E. Savage, Energy Environ. Sci. 2010, 3, 311–317.
4] P. Maki-Arvela, I. Kubickova, M. Snꢂre, K. Eranen, D. Y. Murzin, Energy
Fuels 2007, 21, 30–41.
[
8]
fatty acids.
[
[
5] J. G. Immer, H. H. Lamb, Energy Fuels 2010, 24, 5291–5299.
6] S. Lestari, P. Maki-Arvela, K. Eranen, J. Beltramini, G. Q. Max Lu, D. Y.
Murzin, Catal. Lett. 2010, 134, 250–257.
Conclusion
[
7] S. Lestari, P. Maki-Arvela, I. Simakova, J. Beltramini, G. Q. Max Lu, D. Y.
Murzin, Catal. Lett. 2009, 130, 48–51.
A commercial 5% Pt/C catalyst is active in hot water for the
decarboxylation of saturated fatty acids (stearic, palmitic, and
lauric acid). The selectivity to the decarboxylation product is
[8] M. Snꢂre, I. Kubi cˇ kovꢄ, P. Mꢃki-Arvela, D. Chichova, K. Erꢃnen, D. Yu.
Murzin, Fuel 2008, 87, 933–945.
[
9] J. G. Immer, M. J. Kelly, H. H. Lamb, Appl. Catal. A: Gen. 2010, 375, 134–
39.
about 90%. No added H is required for this reaction. The de-
2
1
carboxylation rates for these three saturated fatty acids are in-
dependent of their carbon numbers. The unsaturated C18
fatty acids underwent decarboxylation much more slowly, and
hydrogenation was the main reaction pathway even though
[10] M. Snꢂre, I. Kubi cˇ kovꢄ, P. Mꢃki-Arvela, K. Erꢃnen, D. Y. Murzin, Ind. Eng.
Chem. Res. 2006, 45, 5708–5715.
[
11] H. Bernas, K. Erꢃnen, I. Simakova, A-R Leino, K. Kordꢄs, J. Myllyoja, P.
Mꢃki-Arvela, T. Salmi, D. Y. Murzin, Fuel 2010, 89, 2033–2039.
12] I. Simakova, O. Simakova, P. Mꢃki-Arvela, A. Simakov, M. Estrada, D. Y.
Murzin, Appl. Catal. A: Gen. 2009, 355, 100–108.
[
no H was added to the reactor. The saturated fatty acid thus
2
formed then underwent decarboxylation. These results suggest
[13] S. Lestari, I. Simakova, A. Tokarev, P. Mꢃki-Arvela, K. Erꢃnen, D. Y. Murzin,
Catal. Lett. 2008, 122, 247–251.
that H is formed in situ during this hydrothermal catalytic pro-
2
[
14] E. W. Ping, R. Wallace, J. Pierson, T. F. Fuller, C. W. Jones, Microporous
Mesoporous Mater. 2010, 132, 174–180.
cess.
Several general features noted herein for Pt/C-catalyzed hy-
drothermal decarboxylation (e.g., rate independent of fatty
[15] P. T. Do, M. Chiappero, L. L. Lobban, D. E. Resasco, Catal. Lett. 2009, 130,
9–18.
[
16] B. Rozmysłowicz, P. Mꢃki-Arvela, S. Lestari, O. A. Simakova, K. Erꢃnen,
I. L. Simakova, D. Y. Murzin, T. O. Salmi, Top. Catal. 2010, 53, 1274–1277.
17] I. Simakova, O. Simakova, P. Mꢃki-Arvela, D. Y. Murzin, Catal. Today 2010,
150, 28–31.
acid chain length, in situ H formation, hydrogenation of unsa-
2
turated fatty acids precedes appreciable decarboxylation, high
selectivity to the decarboxylation product) have also been
noted previously for fatty acid decarboxylation in organic sol-
vents. Thus, the identity or nature of the solvent does not
appear to have a strong influence on these aspects. Other fea-
tures of this reacting system, however, such as the catalyst ac-
[
[18] M. Watanabe, T. Iida, H. Inomata, Energy Convers. Manage. 2006, 47,
344–3350.
3
[
[
19] D. Kusdiana, S. Saka, Appl. Biochem. Biotechnol. 2004, 115, 0781–0791.
20] http://www.diversified-energy.com/index.cfm?s_webAction=centia, ac-
cessed October 2010.
tivity, the deactivation rate, and the ease of in situ H forma-
[21] R. B. Levine, T. Pinnarat, P. E. Savage, Energy Fuels 2010, 24, 5235–5243.
2
[
[
[
22] T. Brown, P. Duan, P. E. Savage, Energy Fuels 2010, 24, 3639–3646.
23] P. Duan, P. E. Savage, Ind. Eng. Chem. Res. 2011, 50, 52–61.
24] L. Li, E. Coppola, J. Rine, J. L. Miller, D. Walker, Energy Fuels 2010, 24,
tion may yet be sensitive to the reaction medium. Additional
research is required to address these issues.
1305–1315.
[
25] L. Haar, NBS/NRC Steam Tables (Eds.: J. S. Gallagher, G. S. Kell), Hemi-
sphere, Washington, 1984, pp. 14–15, 165, 263.
26] C. Yoon, M. X. Yang, G. A. Somorjai, Catal. Lett. 1997, 46, 37–41.
27] Y. Kitayama, M. Muraoka, M. Takahashi, T. Kodama, E. Takahashi, M. Oka-
mura, JAOCS 1997, 74, 526–529.
Acknowledgements
[
[
J.F. thanks the China Scholarship Council for financial support as
a joint PhD student, and we acknowledge financial support from
the U-M College of Engineering and the National Science Foun-
dation (CBET-0755617, EFRI-0937992).
[
28] G. W. Huber, J. W. Shabaker, S. T. Evans, J. A. Dumesic, Appl. Catal. B: En-
vironm. 2006, 62, 226–235.
[29] A. V. Kirilin, A. V. Tokarev, E. V. Murzina, L. M. Kustov, J. Mikkola, D. Y.
Murzin, ChemSusChem 2010, 3, 708–718.
[
30] J. W. Shabaker, R. R. Davda, G. W. Huber, R. D. Cortright, J. A. Dumesic, J.
Catal. 2003, 215, 344–352.
Keywords: biofuels · biomass · green diesel · heterogeneous
catalysis · reaction pathways
Received: October 29, 2010
[1] S. Lestari, P. Mꢃki-Arvela, J. Beltramini, G. Q. M. Lu, D. Y. Murzin, ChemSus
Revised: December 2, 2010
Chem 2009, 2, 1109–1119.
Published online on January 27, 2011
4
86
www.chemsuschem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2011, 4, 481 – 486