The hydrodeoxygenation of bioderived furans into alkanes

Article abstract of DOI:10.1038/nchem.1609

The conversion of biomass into fuels and chemical feedstocks is one part of a drive to reduce the world's dependence on crude oil. For transportation fuels in particular, wholesale replacement of a fuel is logistically problematic, not least because of the infrastructure that is already in place. Here, we describe the catalytic defunctionalization of a series of biomass-derived molecules to provide linear alkanes suitable for use as transportation fuels. These biomass-derived molecules contain a variety of functional groups, including olefins, furan rings and carbonyl groups. We describe the removal of these in either a stepwise process or a one-pot process using common reagents and catalysts under mild reaction conditions to provide n-alkanes in good yields and with high selectivities. Our general synthetic approach is applicable to a range of precursors with different carbon content (chain length). This allows the selective generation of linear alkanes with carbon chain lengths between eight and sixteen carbons.

Full text of DOI:10.1038/nchem.1609

ARTICLES
PUBLISHED ONLINE: 7 APRIL 2013 | DOI: 10.1038/NCHEM.1609
The hydrodeoxygenation of bioderived furans
into alkanes
Andrew D. Sutton1 , Fraser D. Waldie2, Ruilian Wu3, Marcel Schlaf2, Louis A. ‘Pete’ Silks III3
*
and John C. Gordon1
*
The conversion of biomass into fuels and chemical feedstocks is one part of a drive to reduce the world’s dependence on
crude oil. For transportation fuels in particular, wholesale replacement of a fuel is logistically problematic, not least
because of the infrastructure that is already in place. Here, we describe the catalytic defunctionalization of a series of
biomass-derived molecules to provide linear alkanes suitable for use as transportation fuels. These biomass-derived
molecules contain a variety of functional groups, including olefins, furan rings and carbonyl groups. We describe the
removal of these in either a stepwise process or a one-pot process using common reagents and catalysts under mild
reaction conditions to provide n-alkanes in good yields and with high selectivities. Our general synthetic approach is
applicable to a range of precursors with different carbon content (chain length). This allows the selective generation of
linear alkanes with carbon chain lengths between eight and sixteen carbons.
he world’s major oil fields have been in production decline for under acidic conditions. This significantly increases the overall dif-
decades1–5. The use of heavy oil, shale oil and oil sands is ficulty in effecting what may be viewed, by some, as trivial reactions.
T
ramping up, but at an insufficient rate to compensate for the The robustness of existing catalysts traditionally applied to fossil-
production losses in the oil fields. Fossil fuel consumption is also derived source upgrading would indeed be put to severe test
projected to increase further, with potentially catastrophic conse- under the constraints listed above. Although this scenario presents
quences for global ecosystems6,7. Sustainable and renewable alterna- some daunting challenges, it also presents some great opportunities
tives to petroleum are therefore critically needed to fill future gaps in to redefine and advance some areas of modern catalysis science.
the supply of transportation fuels and chemical feedstocks8.
Carbohydrates are an attractive carbon source for use as trans-
The conversion of crude oil into fuels and synthetic feedstocks portation fuel synthons10. Glucose and xylose, as the main building
has been investigated for more than a century, resulting in an enor- blocks of ligno-cellulosic biomass, are the most abundant mono-
mous body of knowledge and an understanding of the processes saccharides on the planet and their use as a fuel precursor represents
required to effect these transformations. In contrast, however, a worthwhile target if cost-effective (hemi-)cellulose depolymeriza-
there is only limited scientific and engineering knowhow on how tions and chemical transformations of the resulting monomers
to catalytically convert biomass-derived molecules into hydro- or their dehydrated derivatives such as methylfurfural (FF) or
carbons for fuels and chemical feedstocks. The conversion of 5-hydroxy-methyl-furfural (HMF) can be developed. In addition
fossil fuel-derived molecules into useful petrochemicals, polymers to the overfunctionalization problem alluded to above, the carbon
and so on represents an issue of chemical ‘underfunctionalization’9. chain lengths of these molecules, once hydrodeoxygenated, are in
Research and development in the field of catalysis has traditionally general too short for transportation fuel purposes (that is, they
been aimed at introducing necessary functionality into these mol- have low boiling points and lack sufficient specific energy density
ecules via processes such as hydroformylation, amination, oxi- for this purpose). However, the furan aldehydes derived from
dation, among others. However, biomass-derived molecules such hexoses and pentoses offer pathways for the desired chain
as carbohydrates pose the opposite problem of being overfunctiona- extensions via aldol condensation chemistry (see below)11.
lized by virtue of having an abundance of oxygen atoms.
Moreover, lignocellulose and corn stover have been directly con-
To access species such as high-energy-density hydrocarbons verted into furans for fuels and chemical applications12. Coupling
from these bioavailable sources, efficient dehydration, hydrogen- these aldehydes with other biomass-derived carbon units using
ation and hydrogenolysis reactions are required. By necessity, such aldol condensation chemistry is thus an attractive route towards
conversions will have to be catalytic in nature, but this area rep- fuel precursors of sufficient energy density.
resents a complete paradigm shift from traditional catalysis
Recently, Dumesic and co-workers reported a hydroxide-based
science, which has focused on the development of chemo-, regio- method (that relies on sodium hydroxide in non-catalytic quan-
and enantioselective catalysts and processes. In contrast, the efficient tities) to promote aldol reactions for the elongation of furan alde-
production of transportation fuels from non-food-derived biomass hydes using ketones. This approach indeed promotes carbon
requires the opposite approach: the development of promiscuous, chain extensions, but does so in a non-selective manner as a
non-selective catalysts capable of dehydrating, hydrogenating and result of self- and cross-condensation chemistries under the con-
deoxygenating a wide variety of functionalities (including hydroxyl, ditions studied13. The resulting deoxygenation of these molecules
olefinic and ketonic groups), in as few steps as possible to perform (carried out in the same reaction vessel) provided a complex
reactions in non-traditional media (that is, water or water mixtures) mixture of molecules ranging from C1 to C15 in chain length
1Chemistry Division, MS J582, Los Alamos National Laboratory, Los Alamos, New Mexico 87545, USA, 2Department of Chemistry, University of Guelph,
Guelph, Ontario, Canada N1G 2W1, 3Biosciences Division, MS E529, Los Alamos National Laboratory, Los Alamos, New Mexico 87545, USA.
e-mail: adsutton@lanl.gov; jgordon@lanl.gov
*
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2013 Macmillan Publishers Limited. All rights reserved.
1

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1609
ARTICLES
O
R
Pd, H2
acetic acid/H2O
65 oC, 2 h
O
O
R
O
A R = H
B R = OH
+
O
H
O
O
1
O
:
1
O
A
B
1
O
OH
OH
O
OH
O
O
O
O
2
Pd, H2
acetic acid/H2O
65 oC, 1h
+
H
C
:
1
O
O
OH
OH
O
O
2
Figure 1 | Organocatalysed routes to chain extended furfurals. A range of
different extended FFs can be prepared by varying the components of the
aldol reaction to give derivatives with between eight and sixteen carbon
atoms. In this case, by altering the stoichiometry using acetone, we can
generate C9 (A and B) or C15 (C) precursor molecules.
Figure 2 | Removal of exocyclic unsaturation. The key first step in this
conversion is removal of the carbon–carbon double bond to give species 1
and 2. This can be achieved in a variety of ways, but using a palladium
catalyst with H2 gas with gentle heating proved to be the most
under high H2 pressures and high temperatures. Intrigued by this successful approach.
report, we recently investigated the use of water-soluble and inex-
of H2 in a 50% aqueous acetic acid solution to give 1 (Fig. 2). The
same approach can be applied to B to give 2 in excellent
isolated yield.
pensive organocatalysts for selective chain extension chemistries
of FF and HMF. Organocatalysts can eliminate the need for large
quantities of caustic base or flammable organic solvents, and in
some cases their reactions can be run solventless14,15. These classes
of catalyst also potentially lend themselves to support on solid
matrixes, making the separation of products and reuse of catalyst
potentially more viable.
This general organocatalysed aldol approach has provided us
with access to a range of isolable furfuraldehyde-based derivatives
with between eight and sixteen carbon atoms. The details of this
work have been described elsewhere11,14,15. Examples of C9 and
C15 fragments derived from the organocatalysed aldol reaction
between acetone and either HMF or methylfurfural (FF) are
shown in Fig. 1 (ref. 11). It should be noted that acetone itself
is also available from renewable biomass through fermentation or
via the sequence EtOH ꢀ HOAc ꢀ acetone (the last step being a
catalytic ketonization)16.
While attempting to achieve milder conditions for the removal
of exocyclic unsaturation in B, divergent reaction pathways were
observed that are dependent on the rate of furan hydrogenation
versus furan ring-opening (Fig. 3). Using solvents that promote
higher H2 solubility and faster reaction kinetics (for example,
THF or MeOH; see Supplementary Information), complete hydro-
genation of 2 occurs to give compound 3.
The other pathway occurs under acidic conditions where the
solubility of H2 is lower. For example, the use of dilute aqueous
acetic acid solution allows acid-promoted ring-opening to occur
faster than furan hydrogenation, resulting in 2,5,8-nonanetrione 4
as the sole product. We can completely avoid the formation of 3
from B by preparing 2 in situ and then heating the acidic solution
in air (thereby releasing H2 and preventing further reduction) to
form 4 in one pot in excellent yield (Fig. 3). Product 2 can be
readily isolated and dissolved in acidic solutions and heated to
generate 4, allowing for a one-pot or stepwise synthesis of the poly-
ketone. As a point of note, A and B can both act as precursors for
the formation of triketone 4 due to the facile hydrogenolysis of
the hydroxymethyl group in B to a methyl group under the acidic
conditions and temperatures required for ring-opening.
Results
Compounds accessible by organocatalysed aldol routes still need to
be defunctionalized to alkanes to provide utility as ‘drop-in’ fuel
replacements. This is the subject of the work described in detail in
the following. From our perspective, achieving the synthesis of
hydrocarbons requires three key chemical transformations to be
effected on these classes of molecules: ring opening of furans,
hydrogenation of olefins, and hydrodeoxygenation (HDO) of
ketones. The order in which these are carried out proved to be
crucial to the efficacy of our approach.
The formation of 4 from 2 provides a facile route to generate
hydrocarbons (see below) from aldol condensation products
without having to potentially resort to the more forcing con-
ditions required to ring-open a tetrahydrofuran18,19. The tendency
to fully hydrogenate species such as A or B is well recognized13
but, in the case of these particular biomass-derived constructs,
we have observed that exerting selective control over hydrogen-
ation of the exocyclic unsaturation first, followed by furan ring-
Initial hydrogenation and divergent reaction pathway.
A
significant proportion of our efforts focused on the study of C9
compounds, and, within this class, two particular systems were
studied. As shown in Fig. 1, one is derived from the aldol
condensation between FF (R¼H) and acetone A, and the other is
derived from the corresponding reaction between HMF (R¼OH)
and acetone B. Removal of the exocylic double bond within these
classes of molecules proved to be a key first step for the
opening, rather than attempting to ring-open
a saturated
tetrahydrofuran fragment, provides a milder route to generate
hydrocarbons. Thus the formation of species such as 3 should
be (and can be) avoided.
subsequent HDO chemistry to be successful, as it prevents Hydrogeoxygenation (HDO) reaction and alkane production. To
fragmentation of the substrate by retro-aldol reaction in the facilitate subsequent oxygen atom removal and perform the desired
aqueous reaction medium. This can be accomplished using several HDO chemistry, we envisioned that a Lewis or Brønsted acid could
approaches and we initially used stoichiometric reagents such as aid in the removal of the C¼O bonds in 4 via iterative reduction and
Mg/MeOH (see Supplementary Information). However, a report dehydration, and the resulting unsaturation could then be reduced
from Ram and Spicer indicated that such exocylic unsaturation with H2 and a suitable hydrogenation catalyst. Initial forays in this
could also be removed using catalytic transfer hydrogenolysis with regard revealed that conversion into the corresponding n-alkane
ammonium formate in acetic acid17. This approach also proved was achieved upon the addition of La(OTf)3 to a mixture of
applicable to A and B and we further simplified the method by Pd/C, acetic acid and the polyketone, and heating this mixture at
replacing ammonium formate with H2 gas. This enabled us to 200 8C under H2 pressures up to 3.45 MPa for 12 h (see
readily remove the exocyclic saturation by, for example, heating A Methods). Tetrahydrofuran species such as 3 proved to be
to 65 8C in the presence of palladium (0.16 mol%), under 1 atm completely unreactive under any HDO conditions that could
2
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2013 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1609
ARTICLES
polyketone). All alkane products were authenticated by gas chrom-
atography mass spectrometry (GC–MS) and comparison of their
13C and 1H NMR spectra to those of pure authentic
reference samples.
O
OH
O
1. Pd (1.6 mol%),
H2, acetic acid/H2O,
65 oC, 2 h.
2
Pd (8.0 mol%), H2,
MeOH, 65 oC, 24 h
In the case of 4 (prepared from A), the HDO chemistry to form
n-nonane was followed by H2 uptake measurements (see
Supplementary Information for these experimental details). These
studies revealed that 6 equiv. H2 are consumed in the reaction to
form n-nonane. This observation accounts for the reduction of
the C¼O bonds and the unsaturation following dehydration of
the resulting –OH groups leading to the formation of 3 equiv.
H2O. GC/GC–MS analysis of this reaction also affirmed that no tri-
ketone remained, and quantitative GC analysis of the reaction sol-
utions using calibrated standards of n-nonane indicated that the
alkane is produced in ≥90% yield. The remaining mass balance is
attributed to coking and the formation of small amounts of CO2,
as observed in the headspace of the reaction vessel by micro-GC
against an authentic CO2 gas sample.
2. 100 oC, 3 h
O
O
O
OH
O
O
96% yield
3
4
Pd (14 mol%),
H2, La(OTf)3
Pd (1.6 mol%),
H2, La(OTf)3
glacial acetic acid
225 oC, 24 h
glacial acetic acid
200 oC, 16 h
In a further refinement of this approach, we assumed that La(OTf)3
could be simply acting as a source of triflic acid (TfOH) in solution.
Thus, the dissolution of polyketones 4 or 10 in 1 M aqueous TfOH
in the presence of palladium under the same conditions of tempera-
ture and H2 pressure permits the production of the corresponding
n-alkane without the necessity for a lanthanide metal. Similarly, per-
forming the reaction with 1 M HCl in the absence of a Lewis acid
yields similar results but at higher palladium catalyst loadings
(16 mol%). This general Lewis acid-free approach can also be
applied to other polyketones to afford the corresponding n-alkanes.
Owing to the commonality of substrate functionalities and
reagents used to effect exocyclic double-bond removal, furan ring-
opening, followed by HDO of the resultant polyketones, it became
evident that isolation of each intermediate on the way to hydro-
carbons may not be required. We therefore demonstrated that the
formation of 4 can readily be achieved in a single vessel without
the need for isolation of 2. Having achieved the formation of poly-
ketones without the need for the isolation of intermediates, we
believed that further process refinement could achieve complete
conversion into n-alkanes, directly from the initial aldol constructs
in a single vessel.
R
R
O
5
90% yield
R = OH, H
Figure 3 | Divergent pathway leads to different products. Full
hydrogenation of the furan ring to form 3 proved to be a dead end for the
synthesis of alkanes under mild conditions. However, if the ring is opened to
a polyketone (4) before this can happen, this provides a significantly lower
energy route to generate alkanes (5) selectively in excellent yield.
convert polyketones into hydrocarbons (Fig. 3) emphasizing the
importance of avoiding the formation of such species.
At pressures less than 0.69 MPa, the HDO chemistry appeared
to be sluggish and gave rise to mixtures of intermediate (that is,
partially deoxygenated) products. As the H2 pressure was increased
above 1.03 MPa, the reaction proceeded rapidly to provide a single
alkane product. Similarly, at temperatures less than 180 8C, HDO
did not proceed at any of the pressures tested (up to 2.76 MPa).
The method described above for accessing polyketones and the
subsequent HDO reaction appears to be general for a variety of
aldol products containing different carbon chain lengths (Fig. 4). One-pot synthesis of alkane from bioderived precursor. The
This provides a facile general route to n-alkanes with n-nonane, solubility of the aldol condensation products (A, B, C and D) is
n-dodecane and n-pentadecane being readily isolated in 87%, 76% limited in water alone, and the higher temperatures required for
and 65% isolated yields, respectively (based on the corresponding HDO result in increased substrate decomposition and charring,
R
O
OH
O
OH
O
O
O
O
O
O
O
A R = H
C
D
B R = OH
O
O
O
O
O
O
O
O
OH
O
O
O
O
4
7
10
n-nonane, C9,
n-pentadecane, C15
n-dodecane, C12
87% isolated yield
65% isolated yield
76% isolated yield
Figure 4 | Conversion of carbon chain extended furans into polyketones and n-alkanes. This general synthetic approach to use palladium, H2 and acid to
generate alkanes is applicable to a range of aldol products, all of which share the same furan, alkene and carbonyl functional groups. This allows us to isolate
a range of different linear alkanes.
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2013 Macmillan Publishers Limited. All rights reserved.
3

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1609
ARTICLES
1. Acetic acid, Pd
0.37 MPa H2
100 oC
fuel applications. The very general methodology has been refined
further to simply use palladium in aqueous acids to generate
linear alkanes in excellent yield under conditions similar to those
that have previously been shown to allow for recycling of palladium
OH
O
O
2. La(OTf)3
5
and application in industrial processes22,23
.
74% isolated yield
2.07 MPa H2
200 o
C
B
The key distinguishing feature of our approach is accessing the
intermediacy of furans and their ability to undergo hydrolytic
ring-opening to polyketones. This subtle modification has signifi-
cant consequences in terms of minimizing the energy input required
to transform these molecules to linear alkanes. Although it is well
documented that hydrocarbons can be produced non-selectively
from tetrahydrofuran intermediates, much greater forcing con-
ditions are required than we have described (300 8C, 5.5 MPa H2)13.
In conclusion, we have demonstrated the feasibility of alkane
production using relatively mild conditions, and we are currently
developing this chemistry further through catalyst lifetime and
recycle studies with specific application towards scale-up.
Figure 5 | One-pot synthesis of n-nonane. Isolation of the intermediates is
not necessary, and we can successfully produce alkanes from the aldol
products by control of the reaction conditions.
thus limiting the use of 1 M HOTf or 1 M HCl as a solvent for all
steps. However, the solubility of these compounds in acetic acid
solutions (10% by volume, up to glacial acetic acid) is excellent
and allows for a one-pot sequential synthesis of n-nonane. The
initial hydrogenation and ring-opening steps of A or B can be
achieved in 3 h by heating at 100 8C, under 0.34 MPa H2, in the
presence of 1.6 mol% palladium in glacial acetic acid. Formation
of 4 is confirmed by NMR spectroscopy carried out on an aliquot
of the reaction mixture. Upon complete conversion to 4,
La(OTf)3 can then be added, and the reaction mixture
repressurized to 2.07 MPa H2 and heated at 200 8C for 12 h to
yield the n-alkane in 74% isolated yield (Fig. 5). It should be
noted that when performing this reaction at sufficient scale, phase
separation of n-nonane is observed. La(OTf)3 can also be added
at the initial hydrogenation/ring-opening stage as this does not
impinge upon the initial hydrogenation or ring-opening sequences.
We performed several control experiments to rule out the possi-
bility of stainless steel (SS-316) playing a role in the catalysis20. As
described, the polyketones can be prepared in glass vessels. To
investigate the role of SS-316 in the subsequent HDO steps, we
had several of our reaction vessels and fittings coated with Dursan
by Silcotek, which provides increased resistance to strong acids.
The use of these vessels led to results identical to those in the
untreated vessel. Although a small portion of the reactors still had
exposed stainless-steel surfaces (pressure gauges, pressure relief
devices and so on), this suggested that SS-316 was not playing a
role in the catalysis. To further examine this, we also performed
control experiments to probe the background catalytic activity of
the SS-316 vessels by heating samples of 4 or 10 at 200 8C under
1.38 MPa H2 for 14–24 h in the presence of glacial acetic acid,
1 M HCl or 1 M HOTf, but in the absence of palladium and/or
Lewis acid (see Supplementary Information for experimental
details). In all cases studied, conversion of the polyketone into the
corresponding alkane was not observed, thus ruling out any sig-
nificant contribution to HDO chemistry by SS-316 under the con-
ditions applied. However, other acid-based reactivity was
observed. For example, 4 dehydrated in the presence of 1 M
HOTf acid to ring-close to the exocyclic hydrogenated analogue
of A, which is a reaction pathway we have previously observed in
glass vessels in the absence of SS-31621. Other side reactions result-
ing from reaction of these acids with 4 and 10 are observed, but
alkanes are not produced unless the appropriate combination of
hydrogenation catalyst and/or lanthanide containing Lewis acid is
present in the vessel.
Methods
General experimental. Compounds A, B, C and D were synthesized as described
previously11. Deuterated solvents were purchased from Cambridge Isotope
Laboratories. All chemicals and reagent grade solvents were obtained from Acros and
used as received. 1H and 13C NMR spectra were obtained at room temperature on a
Bruker AV400 MHz spectrometer, with chemical shifts (d) referenced to the residual
solvent signal (1H and 13C). GC–MS analysis was carried out using a Hewlett-Packard
6890 GC system equipped with a Hewlett-Packard 5973 mass selective detector.
Illustrative examples of the conversion of A and B to n-nonane are described in the
following. Additional experimental procedures, including H2 uptake measurements
involved in the conversion of 3 into n-nonane and the synthesis of n-dodecane and
n-pentadecane, are provided in the Supplementary Information.
Conversion of A into 2,5,8-nonanetrione 4. A (563 mg, 3.75 mmol) was dissolved
in a 50% vol/vol solution of acetic acid in H2O (15 ml) and Pd/C added (13 mg
of 5 wt% Pd/C, 0.006 mmol Pd, 0.16 mol% Pd relative to substrate). The solution
was added to a round-bottom flask, placed under an atmosphere of H2 (balloon),
and heated at 65 8C for 2 h to yield a near-colourless solution of 1 in quantitative
yield, as confirmed by GC–MS and NMR on an aliquot. The flask was then equipped
with a condenser, opened to the air, and heated at 100 8C for 3 h to yield a pale
yellow solution. On cooling, the solution was filtered to remove the Pd/C and the
aqueous layer was neutralized with NaHCO3, extracted with dichloromethane
(3 × 5 ml), dried over Na2SO4, filtered and solvent removed in vacuo to yield 3
(0.589 mg, 92% yield). 1H NMR (400 MHz, CDCl3) d 2.92–2.44 (m, 8H), 2.42–1.97
(m, 6 H). 13C NMR (101 MHz, CDCl3) d 207.99, 207.30, 36.99, 36.09, 29.87.
Conversion of B into 2,5,8-nonanetrione 4. B (830 mg, 5.00 mmol) was dissolved
in a 50% vol/vol solution of acetic acid in H2O (15 ml) and Pd/C added (83 mg
of 10 wt% Pd/C, 0.08 mmol Pd, 1.6 mol% Pd relative to substrate). The solution
was added to a round-bottom flask and put under an atmosphere of H2 using a
balloon and heated at 65 8C for 1 h to yield a pale yellow solution of 2. The flask
was then equipped with a condenser, opened to the air and heated at 100 8C for 3 h
to yield a colourless solution. On cooling, the solution was filtered to remove
the Pd/C, the aqueous layer was neutralized with NaHCO3 and extracted with
dichloromethane (3 × 5 ml), dried over MgSO4, filtered and solvent removed
in vacuo to yield 3 (817 mg, 96% yield). 1H NMR (400 MHz, CDCl3) d 2.92–2.44
(m, 8H), 2.42–1.97 (m, 6 H).13C NMR (101 MHz, CDCl3) d 207.99, 207.30, 36.99,
36.09, 29.87.
Synthesis of n-nonane 5. To a 25.00 ml volumetric flask, 851 mg (5 mmol) of 4 was
dissolved in glacial acetic acid together with dimethyl sulfone (DMS) (235.3 mg,
2.5 mmol) as an internal standard for GC–MS to yield an orange solution. The
solution was sonicated for 5 min and then loaded in the autoclave reactor with Pd/C
(170 mg of 5 wt% Pd/C, 0.08 mmol Pd; that is, 1.6 mol% Pd relative to substrate)
and La(OTf)3 (426 mg, 0.727 mmol) and sealed. The reactor was pressurized with
H2 to 3.45 MPa and vented three times to remove any residual oxygen atmosphere.
The reactor was pressurized to 3.45 MPa one final time and heated to the desired
reaction temperature (200 8C). After 16 h the reactor was cooled to room
temperature and the H2 pressure released. The sole reaction product was identified
as n-nonane (90% yield) with no 4 remaining, as confirmed by GC–MS. 1H NMR
(400 MHz, CDCl3) d 1.29 (m, 14H), 0.88 (s, 6 H). 13C NMR (101 MHz, CDCl3)
d 32.14, 29.73, 29.57, 22.81, 14.23.
Discussion
The selective conversion of biomass-derived furans into linear
alkanes under mild conditions represents a significant advance
towards the production of fuels from non-food-based biomass
sources. These conversions can be carried out sequentially in a
‘one-pot’ approach without isolation of the intermediates. A
variety of linear alkanes can be produced selectively in good to
excellent yields from bioderived constructs provided by organoaldol
chemistry. This provides a powerful synthetic and general approach
Received 3 December 2012; accepted 25 February 2013;
for the production of higher linear alkanes CnH2nþ2 with n ≥ 9 for published online 7 April 2013
4
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2013 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1609
ARTICLES
17. Ram, S. & Spicer, L. D. Catalyst type and concentration dependence in catalytic
transfer hydrogenolysis of a,b-unsaturated carbonyls and nitriles via
ammonium formate. Synth. Commun. 22, 2683–2690 (1992).
18. Copes, J. P. & McKinkey, C. Production of 1,4-butanediol from tetrahydrofuran.
US patent 2,686,817 (1954).
19. Watson, J. M. Butane-1,4-diol from hydrolytic reduction of furan. Product R&D
12, 310–311 (1973).
20. Di Mondo, D. et al. Stainless steel as a catalyst for the total deoxygenation
of glycerol and levulinic acid in aqueous acidic medium. ACS Catal. 1,
355–364 (2011).
References
1. Simmons, M. R. Twilight in the Desert: The Coming Saudi Oil Shock and the
World Economy (Wiley, 2006).
2. Armaroli, N. & Balzani, V. The future of energy supply: challenges and
opportunities. Angew. Chem. Int. Ed. 46, 52–66 (2007).
3. Kerr, R. A. Technology is turning US oil around but not the world’s. Science 335,
522–523 (2012).
4. Murray, J. & King, D. Climate policy: oil’s tipping point has passed. Nature 481,
433–435 (2012).
5. Tverberg, G. E. Oil supply limits and the continuing financial crisis. Energy 37,
27–34 (2012).
21. Waidmann, C. R. et al. Catalytic ring opening of furan rings within biomass
derived substrates: an experimental and theoretical study. Catal. Sci. Technol.
http://dx.doi.org/10.1039/C2CY20395B (2012).
6. Exxon Mobil. Outlook for Energy for 2040. http://www.exxonmobil.com/
Corporate/energy_outlook_view.aspx
22. Bissot, T. C. Surface impregnated catalyst. US patent 4,048,096 (1997).
23. Han, Y. F., Kumar, D., Sivadinarayana, C. & Goodman, D. W. Kinetics of
ethylene combustion in the synthesis of vinyl acetate over a Pd/SiO2 catalyst.
J. Catal. 224, 60–68 (2004).
7. IPCC. Contribution of Working Groups I, II and III to the Fourth Assessment
Report of the Intergovernmental Panel on Climate Change. http://www.ipcc.ch/
publications_and_data/ar4/syr/en/contents.html (2007).
8. Xing, R. et al. Production of jet and diesel fuel range alkanes from waste
hemicellulose-derived aqueous solutions. Green Chem. 12, 1933–1946 (2010).
9. Schlaf, M. Selective deoxygenation of sugar polyols to [a],[v]-diols and other
oxygen content reduced materials—a new challenge to homogeneous ionic
hydrogenation and hydrogenolysis catalysis. Dalton Trans. 4645–4653 (2006).
10. Huber, G. W., Iborra, S. & Corma, A. Synthesis of transportation fuels
from biomass: chemistry, catalysts, and engineering. Chem. Rev. 106,
4044–4098 (2006).
11. Keith, J. M. et al. Aqueous organocatalysis for the carbon chain extension of
carbohydrate derivatives: application to the production of transportation fuels.
Curr. Org. Chem. in the press.
12. Mascal, M. & Nikitin, E. B. Co-processing of carbohydrates and lipids in oil
crops to produce a hybrid biodiesel. Energy & Fuels 24, 2170–2171 (2010).
13. Huber, G. W., Chheda, J. N., Barrett, C. J. & Dumesic, J. A. Production of liquid
alkanes by aqueous-phase processing of biomass-derived carbohydrates.
Science 308, 1446–1450 (2005).
Acknowledgements
The authors acknowledge financial support from the Laboratory Research and
Development (LDRD) program at Los Alamos National Laboratory.
Author contributions
A.D.S. made the initial discovery. A.D.S., M.S., L.A.S. and J.C.G. conceived and designed
the experiments. A.D.S. and F.D.W. performed the experiments. R.W. and L.A.S.
contributed materials. A.D.S. and J.C.G. co-wrote the paper. All authors discussed the
results and provided input for the manuscript.
Additional information
Supplementary information and chemical compound information are available in the
online version of the paper. Reprints and permissions information is available online at
www.nature.com/reprints. Correspondence and requests for materials should be addressed to
A.D.S. and J.C.G.
14. Silks, L. A. et al. Method of carbon chain extension using novel aldol reaction.
US patent application 20110040110 (17 August 2009).
15. Sutton, A. D. et al. Compounds and methods for the production of long chain
hydrocarbons from biological sources. International patent application
PCT/US2012/055340 (14 September 2012).
´
´
16. Glinski, M., Kijenski, J. & Jakubowski, A. Ketones from monocarboxylic acids: Competing financial interests
catalytic ketonization over oxide systems. Appl. Catal. A 128, 209–217 (1995).
The authors declare no competing financial interests.
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
5
© 2013 Macmillan Publishers Limited. All rights reserved.