Understanding ketone hydrodeoxygenation for the production of fuels and feedstocks from biomass

Article abstract of DOI:10.1021/cs501965w

Although we can efficiently convert bioderived furans into linear alkanes, the most energy-intensive step in this approach is the hydrodeoxygenation of the intermediate polyketone. To fully understand this process, we have examined the hydrodeoxygenation of a model compound, 3-pentanone, which allows us to follow this process stepwise using Pd/C, H₂ (200 psi), and La(OTf)₃ in acetic acid to remove the oxygen atom at temperatures between 25 and 200 C. We have found that ketone reduction to an alcohol is followed by acetoxylation, which provides a more facile route to C-O bond cleavage relative to the parent alcohol. (Chemical Presented).

Full text of DOI:10.1021/cs501965w

Letter
pubs.acs.org/acscatalysis
Understanding Ketone Hydrodeoxygenation for the Production of
Fuels and Feedstocks From Biomass
Amanda E. King,† Ty J. Brooks,† Yong-Hui Tian,‡,§ Enrique R. Batista,‡ and Andrew D. Sutton*,†
‡
†Chemistry Division (C-IIAC; Inorganic, Isotope and Actinide Chemistry), Theoretical Division (T-1; Physics and Chemistry of
Materials), Los Alamos National Laboratory, Los Alamos, New Mexico 87544, United States
S
* Supporting Information
ABSTRACT: Although we can efficiently convert bioderived
furans into linear alkanes, the most energy-intensive step in
this approach is the hydrodeoxygenation of the intermediate
polyketone. To fully understand this process, we have
examined the hydrodeoxygenation of a model compound, 3-
pentanone, which allows us to follow this process stepwise
using Pd/C, H2 (200 psi), and La(OTf)3 in acetic acid to
remove the oxygen atom at temperatures between 25 and 200
°C. We have found that ketone reduction to an alcohol is
followed by acetoxylation, which provides a more facile route
to C−O bond cleavage relative to the parent alcohol.
KEYWORDS: hydrodeoxygenation, biomass, ketone reduction, dehydration, mechanism
on-food-based carbohydrates are attractive renewable
Nstarting materials for conversion into fuels and feed-
stocks.1 Glucose and xylose, as the main building blocks of
lignocellulosic biomass, are the most abundant monosacchar-
ides on the planet, and their use as precursors for these
applications is a worthwhile target.2 Conversion of these sugars
into platform chemicals such as furfural or 5-hydroxymethyl-
furfural (HMF) provides useful precursor molecules for further
chain extension or derivatization to add value or energy
density.3−8 For these molecules to be converted into direct
drop-in fuel replacements, the abundant functional groups
associated with carbohydrates (i.e., oxygen atoms) need to be
removed. Recently, we developed a route for selective chain
extension of furan aldehydes (furfural and HMF) to produce
isolable higher-order furfuraldehydes containing between 8 and
16 carbon atoms. These can then be selectively converted into
linear alkanes in excellent yields using a standard set of catalysts
and reaction conditions.9 The chain-extended furfuraldehydes
are converted into polyketones via hydrogenation of their
exocyclic double bond, followed by subsequent acid-catalyzed
ring opening.10 However, the following hydrodeoxygenation
(HDO) reaction to convert the polyketones into hydrocarbons
requires more forcing conditions (200 °C and 200 psi H2).
Reducing the energy required for this final step (i.e., via lower
temperatures and pressures) should allow for a more
economically viable process for potential commercialization.
Studying this reaction and understanding the stepwise
mechanism would allow us to improve the HDO reaction by
specifically addressing the highest energy transformations.
Herein, we report the use of a simple model system to probe
a stepwise HDO reaction applicable to bioderived molecules.
Initial mechanistic studies utilizing polyketone substrates
were complicated by the presence of multiple reactive centers
within the substrates. For example, 2,5,8-nonanetrione could
potentially yield 39 possible intermediates. To overcome this
complication, 3-pentanone was chosen as a model substrate
containing a single functional group that was anticipated to
proceed through a pathway that included 3-pentanol and 2-
pentene. This allows for rapid and easy identification of
intermediates and gives us a model system in which we can
effectively screen catalysts and reaction conditions to elicit the
desired transformations. To test this hypothesis, we subjected
each of these compounds to our typical HDO conditions (Pd/
C, La(OTf)3, acetic acid, 200 psi H2, 200 °C, 15 h) and
observed pentane formation in all cases. We also obtained ethyl
acetate as a side product from acetic acid reduction under these
conditions.11
To observe the intermediates, we performed the HDO
reaction of each substrate as a function of temperature. Our
standard reactions involved combining the reactants in a
stainless steel reactor and heating for 15 h at temperatures
between 25 and 200 °C (with 200 psi H2 if required).
Following cooling of the reactor, the crude reaction mixtures
1
were filtered and yields were obtained by H NMR relative to
an internal standard. For the HDO reactions of 3-pentanone, 3-
pentanol, and 2-pentene, the product distributions are shown in
Figure 1.
Received: December 8, 2014
Revised: January 13, 2015
Published: January 20, 2015
© 2015 American Chemical Society
1223
DOI: 10.1021/cs501965w
ACS Catal. 2015, 5, 1223−1226

ACS Catalysis
Letter
Figure 1. Temperature-dependent product distributions observed when 3-pentanone (A), 3-pentanol (B), and 2-pentene (C) are subjected to
standard HDO conditions for 15 h ([substrate] = 400 mM, [La(OTf)3] = 60 mM, 10 wt % Pd/C relative to substrate, 2.5 mL acetic acid, 200 psi
■
H2). Substrate key: 3-pentanone = , 3-pentanol = red solid circle, 3-acetoxypentane = magenta solid circle, 2-acetoxypentane = orange solid circle,
2-pentene = green diamond, pentane = blue triangle.
At temperatures above 175 °C, 3-pentanone forms pentane
as the sole product at up to 85% yield. At temperatures below
175 °C, 3-pentanol and the intermediates 2- and 3-
acetoxypentane are observed (Figure 1A). The acetoxypentanes
prove to be important intermediates in our process, and their
importance will be discussed later. Decreasing the temperature
results in lower conversions of 3-pentanone, and the expected
2-pentene intermediate is not observed under these conditions.
The conversion of 3-pentanol into pentane goes to
completion at temperatures above 175 °C, with both 2- and
3-acetoxypentane observed below this temperature. Below 100
°C, 3-acetoxypentane was the major observed product (Figure
1B), and as the reaction temperature is decreased, lower
conversion of 3-pentanol is seen. The lack of observable 2-
pentene indicates that the alkene intermediate is readily
hydrogenated under these reaction conditions; accordingly,
hydrogenation of 2-pentene proved to be facile at all
temperatures investigated (Figure 1C).
Employing Pd/C and La(OTf)3 independently rather than in
Figure 2. Temperature-dependent product distributions observed
tandem allowed us to isolate intermediates in the HDO
process. Subjecting 3-pentanone to HDO conditions (200 °C)
in the absence of La(OTf)3 yields only 3-acetoxypentane
(Figure 2), confirming that the role of La(OTf)3 is to facilitate
C−O bond cleavage. Following this reaction as a function of
temperatures shows reduction to the alcohol at 100 °C,
followed at higher temperatures by a conversion into the
acetoxypentanes with nearly complete conversion at 180 °C. As
expected, heating 3-pentanol in the presence of La(OTf)3 in
acetic acid results in efficient production of 3-acetoxypentane at
100 °C and 2-pentene at 150 °C; in addition, the 2-
acetoxypentane intermediate is also observed at temperatures
>100 °C (Figure 3).
The individual steps of the net dehydration reaction were
also studied. The acetoxylation of 3-pentanol proceeds to
completion at 100 °C in acetic acid, and heating 3-
acetoxypentane with La(OTf)3 results in nearly complete
conversion into 2-pentene at temperatures >150 °C (Figure 4).
The reverse reaction (i.e., the acetoxylation of 2-pentene) does
not proceed at 150 °C in acetic acid alone, although a small
amount of both 2- and 3-acetoxypentane are observed in the
presence of La(OTf)3 and prolonged heating. 2-Acetoxypen-
during the HDO reaction with La(OTf)3 absent for 15 h ([substrate]
= 1.2 M, 10 wt % Pd/C relative to substrate, 2.5 mL acetic acid, 200
psi H2). Substrate key: 3-pentanone = , 3-pentanol = red solid circle,
3-acetoxypentane = magenta solid circle, 2-acetoxypentane = orange
solid circle.
■
tane can be substituted for 3-acetoxypentane with no change in
reactivity observed.
To determine the importance of acetoxypentane and whether
this is a key intermediate that facilitates facile C−O cleavage,
we attempted the dehydration of 3-pentanol and 3-
acetoxypentane in sulfolane. In the presence of La(OTf)3
excluding acetic acid, 3-pentanol did not undergo C−O
cleavage reactions; however, subjecting 3-acetoxypentane to
identical conditions yielded complete conversion into the
alkene at 150 °C. This result indicates that the acetate moieties
are important in enabling more facile C−O bond cleavage and
subsequent alkene formation.
This reaction pathway is summarized in Scheme 1; palladium
reduces the ketone to the alcohol, which is converted to the
acetoxypentane by acetic acid. The C−O bond cleavage is
facilitated by the La(OTf)3, resulting in the alkene, which is
1224
DOI: 10.1021/cs501965w
ACS Catal. 2015, 5, 1223−1226

ACS Catalysis
Letter
Scheme 1. Reaction Pathway and Intermediates Formed in
Acetoxy-Mediated Hydrodeoxygenation
1
two species are isoenergetic (pentanone + H2 and pentanol);
however, Pd/C lowers the H2 dissociation barrier, thus
accelerating this reaction. Step 2 consists of the acetoxylation
of pentanol. As in all the remaining steps, the reaction free
energy can be affected by side reactions in solution; therefore,
several possibilities were considered, as detailed in the
Supporting Information. The formation of the acetoxypentane
is favored by 0.31 kcal/mol. This step is followed by the release
of HOAc and formation of pentene and is exergonic, releasing
ΔG = 6.9 kcal/mol. Again, depending on the different models
we studied for coordination of HOAc in solution, the reaction
free energy can vary by 0.2 kcal/mol, and in the crudest model,
by 2.6 kcal/mol; however, this does not change the conclusion
of the favorability of this step. Finally, reaction with H2 yields
pentane in a highly exergonic reaction, with a ΔG = −17.5
kcal/mol, which supports the experimental data.
Figure 3. H NMR following the La(OTf)3 catalyzed dehydration of
3-pentanol in acetic acid. Substrate key: 3-pentanol = red solid circle,
3-acetoxypentane = orange solid circle, 2-acetoxypentate = magenta
circle, 2-pentene = green diamond.
The results of this study have important implications for the
design of catalysts for the HDO of biomass derived molecules
into fuels, especially those that facilitate the removal of a ketone
fragment. Our standard HDO reaction conditions permit both
ketone reduction and C−O bond cleavage. The latter
transformation is a multistep process that is aided by the
formation of the corresponding acetoxy compound. Cleavage of
the C−O bond in this intermediate requires lower temperatures
than the parent alcohol under the conditions we employ.
Although a number of both precious and nonprecious metal
catalysts for ketone hydrogenation at temperatures as low as 25
°C are known, to the best of our knowledge very few are stable
under the acidic conditions employed in this study.12−17 To
that end, we are currently developing more-effective ketone
reduction approaches into our HDO process. By breaking
down the HDO reaction into several reaction steps that are
easy to study, we can rapidly screen catalysts and conditions to
develop faster and lower energy routes to bioderived
replacements for fuels and chemical feedstocks. In addition,
by studying this reaction in depth, we have uncovered the
importance of acetate esters as intermediates in the cleavage of
C−O bonds within our HDO process. This can be applied to
other routes that have also adopted our approach of utilizing
acetic acid18 and could give insight into the mechanism
involved in the use of other acid catalysts19,20 and the selection
of acid used for this purpose. We have achieved this by using a
simple commercially available surrogate for this reaction that
can be easily and rapidly studied and can be used effectively to
screen new catalysts and reaction conditions in the future. By
Figure 4. Temperature-dependent product distributions observed
during the C−O bond cleavage reaction with La(OTf)3 in acetic acid;
15 hours ([substrate] = 1.2 M, [La(OTf)3] = 40 mM, 2.5 mL acetic
acid). Substrate key: 3-acetoxypentane = magenta solid circle, pentane
= blue triangle.
readily hydrogenated to the alkane. Using acetic acid as a
solvent also has the benefit of being able to solubilize all the
intermediates, which prevents precipitation and char formation
upon heating.
We also analyzed this reaction pathway via computational
studies applied to reaction free energies for all the intermediate
steps in Scheme 1. The computed free energy landscape
indicates that once the first step (pentanone to pentanol) is
accomplished, the rest of the reaction follows in an exergonic
fashion (i.e., downhill).
Our calculations did not include the modeling of the Pd/C
catalysis within ketone reduction; however, we find this
reaction to be uphill by only ΔG = +0.4 kcal/mol. From the
point of view of the accuracy of density functional theory, these
1225
DOI: 10.1021/cs501965w
ACS Catal. 2015, 5, 1223−1226

ACS Catalysis
Letter
(15) Langer, R.; Leitus, G.; Ben-David, Y.; Milstein, D. Angew. Chem.,
Int. Ed. 2011, 50, 2120−2124.
doing this, we have revealed an additional component of our
initial HDO chemistry using acetic acid for us to consider as we
strive to minimize required energy inputs for the conversion of
biomass-derived substrates into drop-in fuels and feedstocks.
Future work is focused on optimizing performance through the
use of flow reactors to capitalize on the insight presented in this
manuscript and applying this stepwise approach to polyketone
substrates.
(16) Zhang, G.; Scott, B. L.; Hanson, S. K. Angew. Chem., Int. Ed.
2012, 51, 12102−12106.
(17) Zhang, G.; Vasudevan, K. V.; Scott, B. L.; Hanson, S. K. J. Am.
Chem. Soc. 2013, 135, 8668−8681.
(18) Liu, D.; Chen, E. Y. X. ACS Catal. 2014, 4, 1302−1310.
(19) Liu, D.; Chen, E. Y. X. ChemSusChem 2013, 6, 2236−2239.
(20) Wegenhart, B. L.; Yang, L.; Kwan, S. C.; Harris, R.; Kenttamaa,
̈
H. I.; Abu-Omar, M. M. ChemSusChem 2014, 7, 2742−2747.
ASSOCIATED CONTENT
■
S
* Supporting Information
The following file is available free of charge on the ACS
Publications website at DOI: 10.1021/cs501965w
Full experimental and computational details (PDF)
AUTHOR INFORMATION
Corresponding Author
*E-mail: adsutton@lanl.gov.
■
Present Address
§(Yong-Hui Tian) College of Life Science, Sichuan University,
Chengdu 610065, China.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Funding for this research was received from Los Alamos
National Laboratory LDRD-ECR (Laboratory Directed Re-
search and Development − Early Career Research). We thank
Dr. John C. Gordon for useful discussions during the
preparation of this manuscript and congratulate him on being
appointed as a Laboratory Fellow at Los Alamos National
Laboratory. Los Alamos National Laboratory is operated by Los
Alamos National Security, LLC, for the National Nuclear
Security Administration of the U.S. Department of Energy
under Contract DE-AC5206NA25396. LA-UR-14-28593.
REFERENCES
■
(1) Huber, G. W.; Iborra, S.; Corma, A. Chem. Rev. 2006, 106, 4044−
4098.
(2) Huang, Y.-B.; Fu, Y. Green Chem. 2013, 15, 1095−1111.
(3) Corma, A.; de la Torre, O.; Renz, M.; Villandier, N. Angew.
Chem., Int. Ed 2011, 50, 2375−2378.
(4) vom Stein, T.; Grande, P. M.; Leitner, W.; Dominguez de Maria,
P. ChemSusChem 2011, 4, 1592−1594.
(5) Yang, W.; Grochowski, M. R.; Sen, A. ChemSusChem 2012, 5,
1218−1222.
(6) Yang, W.; Sen, A. ChemSusChem 2011, 4, 349−352.
(7) Yang, Y.; Hu, C.-W.; Abu-Omar, M. M. ChemSusChem 2012, 5,
405−410.
(8) Mascal, M.; Nikitin, E. B. Energy Fuels 2010, 24, 2170−2171.
(9) Sutton, A. D.; Waldie, F. D.; Wu, R.; Schlaf, M.; Silks, L. A.;
Gordon, J. C. Nat. Chem. 2013, 5, 428−432.
(10) Waidmann, C. R.; Pierpont, A. W.; Batista, E. R.; Gordon, J. C.;
Martin, R. L.; Silks, L. A. P.; West, R. M.; Wu, R. Catal. Sci. Technol.
2013, 3, 106−115.
(11) Brewster, T. P.; Miller, A. J. M.; Heinekey, D. M.; Goldberg, K.
I. J. Am. Chem. Soc. 2013, 135, 16022−16025.
(12) Di Mondo, D.; Ashok, D.; Waldie, F.; Schrier, N.; Morrison, M.;
Schlaf, M. ACS Catal. 2011, 1, 355−364.
(13) Glinski, M.; Kijenski, J.; Jakubowski, A. Appl. Catal., A 1995,
128, 209−217.
(14) Langer, R.; Iron, M. A.; Konstantinovski, L.; Diskin-Posner, Y.;
Leitus, G.; Ben-David, Y.; Milstein, D. Chem.Eur. J. 2012, 18, 7196−
7209.
1226
DOI: 10.1021/cs501965w
ACS Catal. 2015, 5, 1223−1226