Highly selective condensation of biomass-derived methyl ketones as a source of aviation fuel

Article abstract of DOI:10.1002/cssc.201500002

Aviation fuel (i.e., jet fuel) requires a mixture of C₉-C₁₆ hydrocarbons having both a high energy density and a low freezing point. While jet fuel is currently produced from petroleum, increasing concern with the release of CO₂ into the atmosphere from the combustion of petroleum-based fuels has led to policy changes mandating the inclusion of biomass-based fuels into the fuel pool. Here we report a novel way to produce a mixture of branched cyclohexane derivatives in very high yield (>94 %) that match or exceed many required properties of jet fuel. As starting materials, we use a mixture of n-alkyl methyl ketones and their derivatives obtained from biomass. These synthons are condensed into trimers via base-catalyzed aldol condensation and Michael addition. Hydrodeoxygenation of these products yields mixtures of C₁₂-C₂₁ branched, cyclic alkanes. Using models for predicting the carbon number distribution obtained from a mixture of n-alkyl methyl ketones and for predicting the boiling point distribution of the final mixture of cyclic alkanes, we show that it is possible to define the mixture of synthons that will closely reproduce the distillation curve of traditional jet fuel.

Full text of DOI:10.1002/cssc.201500002

DOI: 10.1002/cssc.201500002
Full Papers
Highly Selective Condensation of Biomass-Derived Methyl
Ketones as a Source of Aviation Fuel
Eric R. Sacia,[a] Madhesan Balakrishnan,[a] Matthew H. Deaner,[a] Konstantinos A. Goulas,[a, b]
F. Dean Toste,[b] and Alexis T. Bell*[a]
Aviation fuel (i.e., jet fuel) requires a mixture of C9–C16 hydro-
carbons having both a high energy density and a low freezing
point. While jet fuel is currently produced from petroleum, in-
creasing concern with the release of CO2 into the atmosphere
from the combustion of petroleum-based fuels has led to
policy changes mandating the inclusion of biomass-based
fuels into the fuel pool. Here we report a novel way to pro-
duce a mixture of branched cyclohexane derivatives in very
high yield (>94%) that match or exceed many required prop-
erties of jet fuel. As starting materials, we use a mixture of n-
alkyl methyl ketones and their derivatives obtained from bio-
mass. These synthons are condensed into trimers via base-cat-
alyzed aldol condensation and Michael addition. Hydrodeoxy-
genation of these products yields mixtures of C12–C21
branched, cyclic alkanes. Using models for predicting the
carbon number distribution obtained from a mixture of n-alkyl
methyl ketones and for predicting the boiling point distribu-
tion of the final mixture of cyclic alkanes, we show that it is
possible to define the mixture of synthons that will closely re-
produce the distillation curve of traditional jet fuel.
Introduction
Aviation fuels must meet a number of stringent specifications,
the most significant of which are high energy-density to maxi-
mize aircraft range and low freezing-point to avoid formation
of crystalline wax particles during high-altitude flight. These re-
quirements are readily met by branched and cyclic hydrocar-
bons, and, consequently, these types of fuels are not likely to
be displaced for air transportation. On the other hand, signifi-
cant growth in the consumption of aviation fuel[1] and con-
cerns about anthropogenic CO2 emissions have led to increas-
ing policy mandates to create biomass-derived hydrocarbons
for the aviation industry.[1,2] For this reason, the International
Air Transport Association aspires to reach 6% blends of bio-
fuels in standard jet fuel by 2020, a target which would have
required 5.0 billion liters of biofuel for the United States alone
in 2011.[2a] Biomass-based aviation fuels could also qualify in
the US for the Renewable Fuels Standard and in the EU for the
Renewable Energy Directive.[2b]
acids (HEFA), Fisher–Tropsch jet fuel (FTJ), and pyrolysis jet fuel
(PJ).[1,3] However, each of the processes used to produce these
fuels has its drawbacks. The use of vegetable oils to produce
HEFA has led to increases in food costs in some parts of the
world and is difficult to produce sustainably with high fuel
yields per acre.[4] Moreover, the linear alkanes formed by decar-
boxylation and hydrogenation of vegetable oils must undergo
hydrocracking and isomerization to reduce their molecular
weight and introduce branching.[1,5] FTJ and PJ technologies
can use non-edible lignocellulosic feedstocks but suffer from
low overall process yields and the high energy demand to
carry out biomass gasification or pyrolysis.[3] It should also be
noted that pyrolysis oil must be hydrodeoxygenated to
comply with regulations and to improve fuel stability.
The preceding discussion suggests that alternative ap-
proaches should be considered for converting biomass to jet
fuel. For example, branched alkanes in the required C9–C16
range can be produced in high selectivity and in the absence
of solvent by furan condensation[6] followed by hydrodeoxyge-
nation.[7] Alternatively, butenes, derived from g-valerolactone,
can be oligomerized to produce C8, C12, and C16 alkenes, which
can then be hydrogenated to yield mostly branched alkanes.[8]
A drawback of this approach is that, at high butene conver-
sions, the selectivity to jet fuel hydrocarbons (C8–C16) cannot
be maintained due to formation of higher oligomers (C16+).[8]
Yet another alternative is the dimerization of angelica lactone,
derived from levulinic acid (LA). This strategy produces C7–C10
intermediates, which after hydrodeoxygenation, yield highly
branched alkanes.[9] However, these products are too volatile
to be used in significant fractions as components of jet fuel.
Since conventional jet fuel contains approximately 20% cyclic
alkanes, the production of such hydrocarbons from biomass is
Several technologies have been evaluated for the produc-
tion of biomass-derived jet fuel.[3] The processes that are clos-
est to commercialization are hydroprocessed esters and fatty
[a] Dr. E. R. Sacia,+ Dr. M. Balakrishnan,+ M. H. Deaner, K. A. Goulas,
Prof. A. T. Bell
Department of Chemical and Biomolecular Engineering
University of California, Berkeley
107 Gilman Hall, Berkeley, CA 94720 (USA)
E-mail: bell@cchem.berkeley.edu
[b] K. A. Goulas, Prof. F. D. Toste
Department of Chemistry
University of California, Berkeley
107 Gilman Hall, Berkeley, CA 94720 (USA)
[+] These authors contributed equally to this work.
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201500002.
ChemSusChem 0000, 00, 0 – 0
1
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
These are not the final page numbers! ÞÞ

Full Papers
Figure 1. Overview of production of methyl ketones from biomass.
especially attractive.[10] Recent reports have shown that such
cyclic alkanes can be produced from biomass using pinenes[11]
and cyclopentanone[12] and that they do possess exceptional
fuel properties.
maximum reported selectivity of each of the above-mentioned
methyl ketone-producing steps is typically in excess of 90%,
the number of processing steps to provide methyl ketones
from biomass should be minimized to limit the number of sep-
arations and diminished process yields associated with addi-
tional process steps.
Here we report a novel approach for producing a mixture of
cyclic alkanes that replicates the boiling point distribution of
jet fuel, exhibits very high energy density, and has excellent
cold flow properties. Our strategy involves the catalytic self-
and cross-condensation of biomass-derived alkyl methyl ke-
tones and subsequent hydrodeoxygenation of the resulting
condensates to create a biomass-derived platform technology
for fuels and specialty chemicals. The starting ketones, includ-
ing 2-butanone, 2-pentanone, 2-hexanone, and 2-heptanone,
can be sourced from biomass-derived C5 and C6 sugars via the
processes shown in Figure 1. Both lignocellulosic and tradition-
al biofuels crops, such as corn and sugarcane, can be used to
produce useful sugars for this process. Fermentation of sugars
to 2,3-butanediol has been known for over 100 years, and
present microbial strains can produce this product from glu-
cose in titers above 150 gLꢀ1.[13] Using hybrid biological-chemi-
cal pathways, 2-butanone can be formed in high selectivity
(>90%) in the aqueous phase by acid-catalyzed dehydration
of 2,3-butanediol,[14] after which it can easily be separated by
extraction or distillation. 2-butanone can also be produced
from sugar-derived levulinic acid via chemical[15] and hybrid
chemical/biological pathways.[16]
While aldol condensation of biomass-derived aldehydes,
such as HMF or FUR, with acetone is well established,[19] the
condensation of biomass-derived alkyl methyl ketones has re-
ceived much less attention. An example of this type of reaction
is the base-catalyzed condensation of acetone to form a cyclic
trimer condensate (isophorone), as well as dimer condensates
(mesityl oxide), larger oligomers, and aromatic products (mesi-
tylene).[20] Formation of cyclic trimer condensates, like isophor-
one, is especially interesting due to the high energy density
and low freezing points of cyclic alkanes. However, production
of cyclic trimer condensates from larger n-alkyl methyl ketones
has not been well studied due to the increased steric hin-
drance provided by larger alkyl substitution. For example, con-
densation of 2-butanone was previously found to yield cyclic
trimer condensates in yields of only 24%.[21] Hence, the main
objective of our work was to identify catalysts and reaction
conditions that would promote a platform jet fuel production
pathway through the selective condensation of methyl ketones
to cyclic trimer condensates (3) or aromatic products (5) while
minimizing formation of dimer condensates (2), acylic trimer
condensates (4), and larger oligomers, as shown in Scheme 1.
Additionally, 2-hexanone can be formed in high selectivity
(98%) by hydrogenation of 2,5-dimethylfuran (DMF),[7b] a plat-
form molecule formed in high yield from 5-hydroxymethylfur-
fural (HMF).[17] 2-pentanone can also be formed in the same se- Results and Discussion
lectivity by hydrogenation of 2-methylfuran (MF), a product of
xylose-derived furfural (FUR) hydrogenation.[7b] Formation of 2-
Catalyst selection
pentanone and 2-heptanone has also been demonstrated by
the mono-alkylation of products obtained from ABE fermenta-
tion (acetone, 1-butanol, and ethanol) (Figure 1).[18] While the
Catalysts for condensation of methyl ketones were screened
using 2-hexanone as the reactant (Table 1). Because of the su-
perior quality of cyclic fuel products, we sought catalysts
&
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
2
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Scheme 1. Typical products of methyl ketone condensation by heterogeneous catalysis.
Table 1. Catalyst screening for the condensation of 2-hexanone.[a]
towards isophorone produc-
tion.[20a,b] While MgO (entry 5)
was highly active, its high basicity
produced higher oligomers and
strongly adsorbed species, as evi-
denced by significant darkening
of the catalyst after reaction.
Entry
Catalyst
T
[8C]
t
[h]
Conv. of 1
[%]
Dimers (2)
[%]
Trimers [%]
cyclic (3) aromatic (5)
Total prod.
[%]
1
K3PO4
180
180
180
150
150
150
150
150
150
150
150
150
150
150
150
150
180
180
12
1
1
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
2
32
55
35
98
100
100
96
77
83
76
99
89
96
67
80
76
82
0
19
7
21
3
0
0
3
37
10
46
0
9
1
55
22
4
0
9
0
1
1
0
0
0
0
0
0
0
29
51
34
72
93
92
89
69
77
68
92
70
85
57
54
62
65
2[b]
3[b]
4
NaOH (1 wt%)
KOH (1 wt%)
KF/Al2O3 (5 mol%)
MgO
Mg-Al-O
Mg-Al-O (100 mg)
Mg-Al-O (50 mg)
Mg-Zr-O
La2O3
HAP
TiO2
ZrO2
Ti-Zr-O
43
13
69
93
92
86
32
63
15
76
42
54
1
5
6
7
8
A routine technique for im-
proving the activity and selectivi-
ty of MgO is to add a dopant,
such as Al or Zr.[24] Incorporation
of Al (a stronger Lewis acid than
Mg) into the MgO framework
has been observed to increase
the number of acid–base pairs
with appropriate strength to
achieve high turnover frequen-
cies for aldol condensation.[24]
Moreover, the Lewis acidic Al
site can help to stabilize nega-
tively charged adsorbed inter-
mediates.[24] We observed that,
in contrast to MgO, the use of
9
10
11
12
13
14
15
16
17
18
4
7
16
19
30
1
SrTiO3
SiO2-Al2O3
Nb2O5
23
5
3
8
53
43
NbOPO4
19
[a] Reagents and conditions: 2-hexanone (2 mmol), toluene (3 mL), 200 mg of catalyst unless otherwise stated.
[b] Run in solvent-free conditions with 1 g of 2-hexanone; Mg-Al-O=Calcined Hydrotalcite (Mg/Al=3); HAP=
Hydroxyapatite.
which could selectively produce cyclic (3) and aromatic (5)
trimer condensates. The results of catalyst screening are given
in Table 1.
hydrotalcite-derived Mg-Al-O (Mg/Al=3) resulted in complete
reactant consumption, even for low catalyst loadings. More im-
portantly, this catalyst exhibited exceptional selectivity to cyclic
trimer condensates under all reaction conditions explored (en-
tries 6–8). Other dopants such as Zr were less effective in pro-
viding high yields of trimers (entry 9). La2O3, another highly
basic material, exhibited high activity (entry 10) despite its low
surface area (Supporting Information); however, similar to
MgO, its selectivity to the targeted trimer products was lower
than that of Mg-Al-O.
While K3PO4 is an effective aldol condensation catalyst for
other systems,[18] entry 1 of Table 1 shows that very little of the
starting material is consumed even for long reaction times and
high temperatures. Strong bases such as solid NaOH and KOH
have been used previously for acetone trimerization.[22] We ob-
served that the reaction of 2-hexanone proceeded readily in
the absence of a solvent (entries 2 and 3), but the catalyst un-
derwent dissolution in the by-product, water, making catalyst
recovery impractical. The higher reactivity of KOH compared to
NaOH is attributed to the relative reactivity of the correspond-
ing enolates (entries 2 and 3).[23]
An increased selectivity to aromatic products was observed
for catalysts containing acid–base pairs with increased acidity,
for example TiO2, ZrO2, and their mixed oxides. Highly acidic
materials, such as amorphous aluminosilicate materials and
niobium(V) oxides and phosphates produced significant frac-
tions of aromatic products (5). Reduced overall product yield
in these reactions (entries 16–18) is attributed to the occur-
rence of further oligomerization, cracking, and other unknown
reaction pathways on the solid acid which may be promoted
more significantly by higher Hammett acidity. Because of its
high relative activity and high selectivity for the formation of
Traditional heterogeneous bases were also screened for con-
densation of 2-hexanone. While KF/Al2O3 (Table 1, entry 4)
showed some activity, low selectivity to the desired conden-
sates and the limited water tolerance of the catalyst make it
a poor choice. Alkaline earth metal oxides such as MgO and
mixed metal oxides are among the most studied catalysts for
self-condensation of acetone due to their significant activity
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
3
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
cyclic trimer condensates, we chose hydrotalcite-derived Mg-
Al-O (Mg/Al=3) for further investigation.
Factors influencing catalytic activity and selectivity
Having selected Mg-Al-O as an exceptional catalyst for the self-
condensation of 2-hexanone, we set out to elucidate the func-
tionality of Mg-Al-O responsible for its uniquely high activity
and selectivity for production of cyclic trimer condensates. To
generate the catalytically active, cubic Mg-Al-O, the brucite-like
structure of hydrotalcite materials must first be calcined at
high temperature. This calcination procedure also increases the
surface area and simultaneously generates Lewis basic sites.[25]
Both the basic and acidic sites of Mg-Al-O have been charac-
terized previously in the literature.[26]
To identify the roles of acidic and basic sites in the self-con-
densation of 2-hexanone, reactions were performed in the
presence of various amounts of titrants such as acetic acid or
pyridine (see Table 2). Acetic acid bound to the basic sites of
Mg-Al-O completely deactivated the catalyst at ratios above
1 mmolgꢀ1 catalyst. Since no significant change occurred in
either activity or selectivity upon pyridine addition, regardless
of concentration, we conclude that the acidic sites of Mg-Al-O
do not significantly participate in the mechanism of cyclic
trimer formation, consistent with previous reports for acetone
on MgO.[20a,b]
Figure 2. XRD pattern of as-received hydrotalcite, Mg-Al-O, and spent Mg-Al-
O (Reaction Conditions: 2-hexanone (2 mmol), toluene (3 mL), 1708C, 2 h,
100 mg of Mg-Al-O).
two cycles to 70% by the sixth cycle; however, recalcination of
the catalyst was shown to completely regenerate the original
catalyst activity, as is consistent with previous reports of aldol
condensations with Mg-Al-O[27] (see Supporting Information).
X-ray diffraction (XRD) patterns of catalyst taken after three
cycles showed little difference compared to that of the cal-
cined hydrotalcite (Mg-Al-O) and did not display the character-
istic peaks of the brucite-like hydrotalcite precursors, as shown
in Figure 2.
We also sought to determine the stability of the catalyst
under reaction conditions and in the presence of water, which
is produced during reaction. Previous studies have shown that
the exposure of water to Mg-Al-O causes a loss of surface area,
regeneration of Brønsted basicity, and reformation of the bru-
cite-like structure of hydrotalcite.[25d] Because water is pro-
duced as a byproduct of the condensation reactions studied in
this work, we were interested in defining the effects of water
on the activity and selectivity of Mg-Al-O for 2-hexanone con-
densation. To this end, Mg-Al-O was subjected to six successive
catalytic cycles in the Q-tube reactors to assess its stability.
This number corresponds to approximately 520 turnovers per
site using the number of basic sites as measured by CO2 tem-
perature-programmed desorption (Supporting Information).
The conversion of 2-hexanone decreased from 99% in the first
The water tolerance of Mg-Al-O at higher water concentra-
tions was also tested by introducing increasing amounts of
water into the initial reaction solution. The data in Figure 3
clearly demonstrate the ability of water to inhibit the activity
of the catalyst for condensation of methyl ketones. The loss in
activity can be attributed to a loss in surface area of Mg-Al-O,
due to changes in the catalyst structure, and to inhibition of
reversible steps in the condensation mechanism, leading to
a shift in equilibrium. In order to mitigate the effects of water
during bulk synthesis, we carried out reactions in a Dean–Stark
apparatus to continuously remove water from the reaction
mixture. This technique proved very effective at maintaining
the catalyst activity during large-
scale syntheses discussed in the
Table 2. Effect of acid/base titration on Mg-Al-O activity during the condensation of 2-hexanone.[a]
Supporting Information, even at
milder temperatures.
Entry
Acetic acid Pyridine Mg-Al-O Conv. of 1 Dimers (2)
Trimers [%]
Higher oligo-
[mmol]
[mmol]
[mg]
[%]
[%]
cyclic (3) acyclic (4) mers [%]
1
2
3
4
5
6
7
8
–
10
50
100
200
–
–
–
–
200
–
–
–
–
–
–
10
50
100
200
–
100
100
100
100
100
100
100
100
100
–
99
98
62
2
1
1
23
1
0
1
1
0
1
0
0
95
93
38
0
2
2
1
0
0
2
2
2
2
0
0
1
2
0
0
0
1
1
2
2
0
0
Reaction pathways with methyl
ketones
The temporal progression of the
reaction over Mg-Al-O was moni-
tored to determine the relative
reactivity of 2-hexanone and
dimer/cyclic trimer condensates
(2/3) (Scheme 2). Figure 4 shows
a plot of the products observed
as a function of time. Over half
0
0
99
98
99
99
0
95
95
95
95
0
9
10
11
200
–
0
0
[a] Reagents and conditions: 2-hexanone (2 mmol), toluene (3 mL), 1708C, 2 h, 100 mg of Mg-Al-O.
&
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
4
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Figure 4. Product evolution during the self-condensation of 2-hexanone.
Conditions: 2-hexanone (2 mmol), toluene (3 mL), 1508C, 200 mg of Mg-Al-
O.
Figure 3. Effect of water addition on the activity of Mg-Al-O. Conditions: 2-
hexanone (2 mmol), toluene (3 mL), water (as shown), 1508C, 3 h, 200 mg of
Mg-Al-O.
of the initial 2-hexanone was converted to a mixture of 2 and
3 within 15 min. The dimer condensate (2) was then rapidly
converted into the cyclic trimer condensate (3). A 94% yield to
cyclic trimer condensates (3) then formed within 2 h and did
not undergo further reaction. The relative reactivity of 2 over
Mg-Al-O shown in Figure 4 is in stark contrast to what was ob-
served over hydroxyapatite and strontium titanate (Table 1, en-
tries 13 and 17), in which cases dimer condensates accumulat-
ed in yields of 46% and 55%, respectively.
basic conditions. Under acidic conditions, the aromatic con-
densates 5a and 5b are obtained exclusively from the acyclic
trimer isomers of 4 via acid-promoted trienol formation fol-
lowed by 6p-cyclization (Scheme 2). Reaction mechanisms for
the formation of dimer and trimer condensates shown in
Scheme 2 can be found in the Supporting Information. Forma-
tion of higher oligomer condensates from 3 is largely prevent-
ed by the significant steric hindrance around the reactive cen-
ters of the molecule.
An overall reaction network showing plausible intermediates
and products formed during acid/base-catalyzed self-conden-
sation of alkyl methyl ketones is presented in Scheme 2. These
products are formed via either enolates or enols of the reac-
tant ketone depending on the nature of the catalyst employed
(base or acid, respectively). Dimer condensates 2a and 2b are
formed through 1,2-addition of enolate or enol intermediates
to a ketone. Cyclic trimer condensates 3a–d could originate
from 1,2- and/or 1,4-addition (Michael addition) of enolate in-
termediates of 1 and dimers 2a and 2b (Scheme 2) under
The isomer distributions for condensates derived from the
self-condensation of 2-butanone through 2-heptanone are
given in Table 3. Although there are several possible intermedi-
ates (I–IV) for formation of 3a–d, the formation of intermedi-
ate I by 1,4-addition (Michael addition) is preferred in the pres-
ence of a base catalyst. This preference is reflected in the
isomer distribution of trimer condensates of 2-hexanone
(Table 3, entry 3). The increased selectivity to 3a (84%) from 2-
hexanone also indicates that the reaction proceeds predomi-
nantly by kinetic enolate intermediates (Scheme 2).
Scheme 2. Overview of plausible pathways to trimer condensates from methyl ketones.
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
5
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
respectively) and notably in-
creased formation of 4, 5, and
higher oligomers. Reactions with
internal ketones, such as 3-pen-
tanone and 4-heptanone also
largely form dimers in yields of
44% and 9%, respectively, after
16 h. Conversely, the self-con-
densation of acetone is rapid,
producing 39% of 3 and 57%
higher oligomers after only 2 h
Table 3. Specific activity and product yield from condensation of C4ꢀC7 alkyl methyl ketones.[a]
Entry Methyl ketone (1) Conv. of 1 Dimers (2) Cyclic trimers (3) [%]
Acyclic trimers (4) Higher oligo-
[%]
[%]
(3a) (3b) (3c) (3d) total [%]
mers [%]
1
2
3
4
2-butanone
2-pentanone
2-hexanone
2-heptanone
98
99
99
95
0
1
1
2
53 33
82 14
84 11
80 10
6
0
0
0
0
0
0
0
92
96
95
90
2
2
2
2
5
1
1
1
[a] Reagents and conditions: Methyl ketone (2 mmol), toluene (3 mL), 1708C, 2 h, 100 mg of Mg-Al-O.
While the reactivity of the shown methyl ketones are nearly
equal, as evidenced by similarly high conversions (95–99%),
the distribution of condensates formed from 2-butanone
(Table 3, entry 1) was different from the rest of the starting ma-
terials in Table 3. Notably, the distribution of cyclic trimer con-
densates 3a–d indicates a significant increase in reactivity of
the thermodynamic enolate formed from 2-butanone com-
pared to that formed from 2-hexanone via intermediate II and/
or IV. This difference in reactivity is due to the steric hindrance
associated with the respective alkyl substituents. For all the
biomass-derived methyl ketones investigated, 2-butanone is
the only reactant for which isomer 3c (Scheme 2) was ob-
served in any significant concentration (6%). Also, the high re-
activity of 2-butanone enolates together with lesser steric hin-
drance around the reactive centers of cyclic trimer condensates
formed from 2-butanone leads to increased formation of
higher oligomer condensates (5%, entry 1). For 2-pentanone
through 2-heptanone (Table 3, entries 2–4), the distribution re-
mains consistent, implying that, over this range, the length of
the n-alkyl group has a limited effect on the reactivity of the
starting ketone.
of reaction. These results demonstrate that steric hindrance
plays a profound role on both reactivity and product selectivi-
ty. It is also obvious that the exclusive formation of dimer con-
densates (2) is very difficult for n-alkyl methyl ketones under
the optimized conditions.
The condensation of n-alkyl methyl ketones can be stopped
at the dimer stage (2) if the dimer condensates can be pre-
vented from undergoing 1,4-conjugate addition with ketone
enolates. This can be achieved by in situ hydrogenation of the
olefinic functionality of 2 immediately after its formation. The
hydrogen required for this process can also be generated in
situ from an alcohol by transfer hydrogenation in the presence
of Pd.[18]
In the selective dimerization approach, 2-hexanol served as
a suitable source for the in situ formation of 2-hexanone and
hydrogen required for the reaction. We found that the conden-
sation predominantly produced dimer condensates as a mixture
of ketones and alcohols in the presence of Pd supported on
Mg-Al-O (77% combined, Scheme 3).
Cross-condensation of methyl ketones
The rate of condensation is greatly diminished, however,
when the alkyl group is branched, especially when the branch-
ing is proximate to the reactive centers of the ketone. In this
case, the selectivity shifts significantly to dimers. For example,
ketones such as 3,3-dimethylbutan-2-one and 4,4-dimethylpen-
tan-2-one form dimers in yields of 64% and 81%, respectively,
after 16 h of reaction (see Supporting Information). Decreased
steric hindrance in 3-methylbutan-2-one and 4-methylpentan-
2-one leads to slightly higher conversion compared to their
more hindered counterparts (89% and 99% vs. 68% and 87%,
As noted earlier, the results presented in Table 3 indicate that
the rates of condensation of biomass-derived C4–C7 n-alkyl
methyl ketones are similar, as are the product distributions.
The next question is whether the rates of cross-condensation
of these ketones are close to those of the rates of self-conden-
sation. To this end, mixtures of ketones were subjected to reac-
tion under conditions similar to those used for the self-conden-
sation studies reported in Table 3. Complete consumption of
all of reactants was observed after 4 h of reaction and the
Scheme 3. Pathway for dimerization of 2-hexanol over a Pd/Mg-Al-O catalyst.
&
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
6
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Figure 5. Product distribution of cross-condensations of mixed ketones. Conditions: methyl ketones (2 mmol), toluene (3 mL), 1708C, 4 h, 100 mg of Mg-Al-O.
a) x4 =0.5, x5 =0.5, b) x4 =0.33, x5 =0.33, x6 =0.33, c) x4 =0.5, x5 =0.3, x6 =0.2, d) x4 =0.83, x5 =0.09, x6 =0.08, e) x4 =0.25, x5 =0.25, x6 =0.25, x7 =0.25,
f) x4 =0.33, x5 =0.33, x7 =0.33, g) x4 =0.4, x5 =0.1, x6 =0.1, x7 =0.4.
solid bars appearing in Figure 5 show the resulting carbon
number distributions. In all of the cases shown in Figure 5, the
conversion of the initial methyl ketones exceeded 99% and all
observed products were trimer condensates, with the excep-
tion of trace amounts of dimers and larger oligomers. It is evi-
dent from this Figure that products containing 12 to 21
carbon atoms can be produced by the combined processes of
self- and cross-condensation of C4–C7 methyl ketones.
none. As can be seen, the predicted and observed carbon
number distributions are in close agreement, demonstrating
that rates of self- and cross-condensation of each of these four
methyl ketones are similar. This result also demonstrates that
many component mixtures, similar to those of true fuels, may
be generated with these reactions.
Hydrodeoxygenation of trimerization products
If the reactivity of each methyl ketone is identical, their com-
bination into trimers should be governed by a statistical distri-
bution. The equation representing the trimer product distribu-
tion for a four component mixture containing 2-butanone
through 2-heptanone is given by the following expression:
The major products produced by the self- and cross-condensa-
tion of C4–C7 methyl ketones (3) contain a cyclohexenone core.
To be suitable for jet fuel, the olefin functionality and the O
atom must be removed from the core structure by hydrodeox-
ygenation. The catalyst chosen for this step was Pt/NbOPO4,
based on a previous report indicating the effectiveness of this
catalyst for the hydrodeoxygenation of biomass-derived com-
pounds.[28] Accordingly, trimer condensates produced from 2-
butanone and 2-hexanone were hydrogenated over Pt/
NbOPO4 at 1608C in the presence of 2.8 MPa of H2 for 5 h.
Using these conditions, cyclic alkanes (6) were produced in
quantitative yields and no evidence was found by GC–MS or
high resolution mass spectrometry (HRMS) for cracking prod-
ucts (see Supporting Information). Based on these findings, Pt/
NbOPO4 was used for all further hydrodeoxygenation.
7
7
7
X X X
Probability ¼ 1 ¼
xixjxk
ð1Þ
k¼4 j¼4 i¼4
Here xi, xj, and xk are the initial moles of alkyl methyl ketone
with i, j, and k carbons, respectively, divided by the total initial
moles of methyl ketones present, and the sum of i, j, and k is
equal to the number of carbons in the cyclic trimer condensate
formed. The sum of terms in the expanded form of Equa-
tion (1) having an identical value of i+j+k (irrespective of the
ordering of i, j, and k) may be grouped together and represent
the probability of obtaining a trimer product with a given total
number of carbon atoms. The shaded bars in Figure 5 give the
predicted carbon number distribution based on Equation (1)
for a wide range of mixtures of 2-butanone through 2-hepta-
The pathways by which hydrodeoxygenation is expected to
occur are illustrated in Scheme 4. The enone functionality is
hydrogenated over platinum to produce the corresponding cy-
clohexanol derivative which then undergoes dehydration over
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
7
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Scheme 4. Plausible reaction pathways during hydrodeoxygenation of 2-butanone trimer condensate.
acid sites on NbOPO4. The intermediate alkene is then hydro-
genated over platinum to form the product alkane. The opti-
mized hydrodeoxygenation conditions were effective for vari-
ous mixtures of condensates shown in Figure 5 and produced
the corresponding cyclic alkanes in quantitative yields.
While mixtures of C12 alkanes from 2-butanone condensa-
tion/hydrodeoxygenation meet or greatly exceed most fuel
specifications, the inherently limited number of components in
the mixture derived from entry 1 of Table 3 cannot replicate
the boiling curve of jet fuel shown in Figure 6a. ASTM D7566
specifies that biomass-derived jet fuel should have at least
a 158C increase between the 10th and 50th percentile of the
boiling curve and a 408C difference between the 10th and 90th
percentile. Furthermore, the 10th percentile must be below
2058C and the final boiling point must be below 3008C.[30b] To
meet the boiling point distribution required for jet fuel, we
also tested the properties of alkane mixtures arising from
cross-condensations of a mixture of ketones as depicted in Fig-
ure 5b (C12–C18) and 5e (C12–C21). These mixtures also had low
freezing points (<ꢀ1008C) and high energy density (38.3 and
38.8 MJLꢀ1, respectively); however, their boiling distributions
still did not meet aviation fuels standards (Figure 6). In fact,
the C12–C21 mixture more closely matched the boiling curve of
commercial diesel (Figure 6b),[31] and, due to its high measured
derived cetane number (48.6), may be more suitable as diesel
fuel.
Properties of hydrodeoxygenated jet fuel products
The fuel properties of cyclic alkanes produced by the sequen-
tial condensation/hydrodeoxygenation of alkyl methyl ketones
were evaluated and compared to those of conventional jet
fuel. The mixture of C12 alkanes produced from 2-butanone
(Table 3, entry 1) remained a clear, free-flowing liquid at
ꢀ808C, indicating that the cloud point and pour point of this
material are <ꢀ808C. The evaluated freezing point of the mix-
ture was determined by ASTM D7153[29] to be <ꢀ1008C,
which far exceeds the requirements for standard Jet A and Jet
A-1 blends, ꢀ408C and ꢀ478C, respectively.[30] The kinematic
viscosity of the C12 alkane mixture at ꢀ208C was measured to
be 5.5 cSt, which is significantly lower than the upper limit
(8.0 cSt) specified for synthetic aviation fuel in ASTM D7566.[30b]
The heat of combustion of the mixture was measured as
38.3 MJLꢀ1 (47.0 MJkgꢀ1), providing a volumetric energy densi-
To match the boiling distribution of jet fuel, we developed
a method for identifying the optimal mixture of C4–C7 ketones
that should be used for methyl ketone condensation/hydro-
deoxygenation to form cyclic alkanes. Details of this procedure
are presented in the Experimental Section and Supporting In-
ty that is 6.1% higher than commercial Jet
(36.1 MJLꢀ1).[10]
A fuel
Figure 6. Simulated distillation of example mixtures. a) Distillation curves of jet fuel, hydrodeoxygenated components from Table 3, entry 1, and hydrodeoxy-
genated components from Figure 5b and 5d. b) Distillation curves of C12–C21 alkanes produced from the product mixture in Figure 5e along with standard
diesel fuel.
&
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
8
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
formation. Utilization of this approach indicates that the opti-
mal fit to the boiling point distribution of conventional jet fuel
can be achieved by condensation of a mixture of C4–C6 ke-
tones for which x4 =0.83, x5 =0.09, x6 =0.08. The resulting dis-
tribution of C12–C18 products is shown in Figure 5d. The simu-
lated distillation curve for the mixture of cyclic alkanes derived
from the optimized product mixture is illustrated in Figure 6a
and compared with that of commercial jet fuel.[32] While this
optimized mixture does meet the specifications for boiling dis-
tributions in ASTM D7566 mentioned previously, an even
better fit to traditional boiling distributions of jet fuel could be
made by incorporating small amounts of C8–C14 alkanes pro-
duced by dimerization of C4–C7 ketones via the strategy shown
in Scheme 3. In this way, more components with boiling points
<2008C could be introduced to create an ideal, biomass-de-
rived jet fuel replacement.
Experimental Section
Materials
All purchased chemicals were used as received without further pu-
rification. Starting materials, including 2-butanone (ꢁ99%), 2-pen-
tanone (ꢁ98%), 2-hexanol (99%) 2-hexanone (98%), and 2-hepta-
none (99%), were acquired from Sigma–Aldrich, USA.
Base catalysts, including pyridine (>99%), KF/Al2O3, NaOH, KOH,
K3PO4, synthetic HAP, and hydrotalcite were purchased from
Sigma–Aldrich, USA. Mg-Al-O was prepared from commercially
available synthetic hydrotalcite (Sigma Aldrich) by calcination in
static air at 7008C for 2 h after a 28Cminꢀ1 temperature ramp. Acid
catalysts, including acetic acid (ꢁ99%) and SiO2–Al2O3 (catalyst
support, grade 135, 6.5% Al content) were also purchased from
Sigma–Aldrich, USA. SiO2–Al2O3 was activated by drying overnight
at 1208C. Niobic acid and NbOPO4 were received in kind from
CBMM, Brazil and were calcined in air prior to use at 3008C for 2 h
after a 28Cminꢀ1 temperature ramp to yield Nb2O5 and NbOPO4.
La2O3, MgO, Mg-Zr-O, TiO2, ZrO2, Ti-Zr-O, and SrTiO3 were produced
by solution phase precipitation and calcination. Specific synthesis
procedures for each can be found in the Supporting Information
along with characterization of all catalyst materials by BET and
XRD.
Conclusions
Our work demonstrates the utility of base-catalyzed methyl
ketone condensation for the selective production of conden-
sates which, after hydrodeoxygenation, meet many specifica-
tions for jet fuel. Once methyl ketones are formed from bio-
mass, which may occur via a broad number of pathways,
yields to fuels are essentially quantitative since the small quan-
tities of dimers and higher oligomeric products are also hydro-
deoxygenated and can be blended into jet fuel at low concen-
trations. The optimal catalyst for condensation was found to
be calcined hydrotalcite, Mg-Al-O. While this catalyst can be
deactivated by water, a byproduct of methyl ketone condensa-
tion, this effect can be minimized by in situ removal of water.
It was observed that if the alkyl group was highly branched,
particularly at the a- or b-carbon, that the formation of dimers,
rather than trimers, was favored. However, linear C4–C7 n-alkyl
methyl ketones uniformly underwent condensation reactions
at similar rates to form cyclic trimer condensates, which could
then be converted quantitatively to substituted cyclohexane
derivatives by hydrodeoxygenation over Pt/NbOPO4.
Supported 2 wt% Pt/NbOPO4 was synthesized by incipient wetness
impregnation using H2PtCl6 hexahydrate from Sigma–Aldrich, USA.
Approximately 265 mg of H2PtCl6 hexahydrate was dissolved in
water and slowly added to 5.0 g of calcined NbOPO4 while mixing
and grinding in a mortar and pestle. The mixture was then heated
at 28Cminꢀ1 to 3008C and reduced in
a constant flow of
50 mLminꢀ1 of 9% H2 in He. Similarly, 2 wt% Pd/Mg-Al-O was syn-
thesized by incipient wetness impregnation of Pd(NO3)2 dihydrate
into Mg-Al-O. The catalyst was then calcined in air by ramping at
18Cminꢀ1 to 5508C and holding at that temperature for 4 h.
Product characterization
Reactant conversion and product yield were determined for reac-
tion mixtures by gas chromatography. After diluting reaction mix-
tures in dichloromethane so that desired products were in a con-
centration of 0.01–1.0 mgmLꢀ1, the mixtures were analyzed with
a Varian CP-3800 gas chromatograph equipped with a flame ioniza-
tion detector (FID) and Varian 320-triple quadrupole mass spec-
trometer. Good separation of products was obtained using a Factor-
Four VF-5 capillary column. Individual product peaks were identi-
fied by mass spectrometry and structures were confirmed with
NMR and FTIR spectroscopy of selected purified compounds (Sup-
porting Information). Quantification was achieved by adding
a known quantity of dodecane or nonane as an internal standard.
Calibration curves were generated for all reactants and the purified
products of reaction of 2-butanone and 2-hexanone. Response fac-
tors for other condensation products were calculated using the ef-
fective carbon number method, which has been shown to predict
FID response factors within 1.7%.[33] For the purposes of this paper,
yield and selectivity are defined by Equations (2) and (3):
Models were developed for predicting both the carbon
number distribution of products formed by mixed condensa-
tion of biomass-derived C4–C7 n-alkyl methyl ketones and the
boiling point distribution of the cyclic alkanes produced by hy-
drodeoxygenation of those products. Using these models it
was possible to predict the mixture of C4–C6 n-alkyl methyl ke-
tones required to produce the C12–C18 cyclic alkanes which ex-
hibit a boiling point distribution curve very close to that of
conventional jet fuel. The cyclic alkanes derived from the se-
quence of condensation and hydrodeoxygenation possess ex-
ceptional properties with regards to freezing point and energy
density. The measured freezing points are well below those
specified for conventional jet fuel, and the volumetric energy
density is 6% higher than that of conventional jet fuel. Our
work has also shown that dimerization of secondary alcohols
through bifunctional metal/base-catalyzed condensation could
also provide an effective method for including condensates of
2-hexanone and 2-heptanone into the fuel supply.
nProduct
nReactant,initial
Yield ¼
ꢂ Stoichiometric Factor ꢂ 100 %
ð2Þ
nProduct
nReactant,initialꢀnReactant
Selectivity ¼
ꢂ Stoichiometric Factor ꢂ 100 %
ð3Þ
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
9
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
in which nProduct =mol of product at a given time; nReactant =mol of
reactant at a given time; nReactant,initial =mol of reactant fed; stoichio-
metric factor=stoichiometric number of reactant mol per mol of
product.
three times with N2 and H2 prior to heating to reaction tempera-
ture. After stirring the reaction at 500 rpm for 5 h, the reactors
were cooled to room temperature and opened. Approximately
1.5 mL of product mixture was then removed and centrifuged at
14000 rpm to separate the catalyst. The supernatant was then di-
luted with dichloromethane and products were analyzed by GC.
Catalyst characterization
The surface area of each catalyst was measured prior to use by
Brunauer–Emmett–Teller (BET) analysis. After degassing at 1808C
for 6 h in a stream of flowing Ar, samples were placed onto a Micro-
meritics TriStar unit and the BET isotherms were measured. X-ray
powder patterns were measured by X-ray diffraction to determine
the crystallographic phases of catalyst materials over a 2q range of
10–80 degrees or 20–60 degrees, with a step size of 0.02 degrees
on a Bruker D8 instrument. Metal dispersion was assessed by CO
pulse chemisorption using a Micromeritics AutoChem II chemisorp-
tion analyzer. Metal nanoparticles were reduced by a mixture of
10% H2 in Ar at 2508C for 30 min prior to CO adsorption. The cu-
mulative quantity of CO adsorbed was then used to determine
metal dispersion and particle size for Pt/NbOPO4 and Pd/Mg-Al-O
(Supporting Information).
Measurement of boiling curves by simulated distillation
Simulated distillation was used to evaluate the suitability of alkane
product distributions for replacement of traditional jet fuel blends
using small quantities of synthesized materials. GC retention times
for known C7–C20 n-alkanes measured during linearly temperature-
ramped gas chromatography were correlated with the boiling
points of these compounds (Supporting Information).[34] Using this
correlation, any alkane mixture containing components with boil-
ing points within this boiling range could be analyzed to deter-
mine the boiling curve of the mixture. Individual product peaks
were integrated and their boiling points were determined by their
elution time. The boiling distribution curve was then generated by
the cumulative sum of the individual components. For the purpos-
es of this work, we have assumed that the alkane components de-
rived from mixed condensations may be approximated to be of
uniform density. This assumption makes it possible to compare the
gravimetric recovery data provided by simulated distillation and
the volumetric recovery distillation curves of standard jet fuels. The
validity of the assumption was established by measuring the bulk
density of the alkane blends shown in Figure 6. These measured
densities varied by less than 0.5% from the mean despite differen-
ces in the composition of the compounds.
Base-catalyzed condensation reactions
In a typical experiment, 2 mmol of methyl ketone substrate,
144 mg of dodecane (internal standard), and 3 mL of toluene was
added to a 12 mL glass Q-tube reactor purchased from QLabtech.
A polytetrafluoroethylene (PTFE)-coated stirbar and 50–200 mg of
Mg-Al-O was then added to the reactor and the reactor was sealed
with a PTFE-coated silicone septum. The reactor was then im-
mersed in a preheated oil bath at 150–1708C and stirred magneti-
cally at 500 rpm for the duration of reaction. Upon completion of
the reaction, the Q-tube reactor was removed and cooled. Approxi-
mately 1.5 mL of product mixture was then removed and centri-
fuged at 14000 rpm to separate the catalyst. The supernatant was
then diluted with dichloromethane and products were analyzed by
GC.
Modeling the boiling curve of methyl ketone mixtures
The following procedure was used to identify the mixture of C4–C7
n-alkyl methyl ketones needed to produce (via ketone condensa-
tion and subsequent hydrodeoxygenation of the resulting product)
a mixture of alkyl substituted cyclohexane derivatives with a speci-
fied boiling point distribution. The alkane boiling regions associat-
ed with each of the cyclic products were identified by GC–MS
(Supporting Information). We then approximated the collection of
individual components in each carbon number range by a Gaussian
probability distribution (standard deviation and means provided in
Supporting Information). The area under each Gaussian curve was
normalized to the mass percent of that product predicted by Equa-
tion (1). In this way, the quantity of products in each carbon
number range was predicted by Equation (1), and the boiling
curve of those products was determined by their respective Gaus-
sian distributions.
The cumulative integration of all Gaussian curves was then used to
predict the boiling distribution curve of mixtures. Using this strat-
egy, a system of equations was generated to predict the boiling
distribution of any set of products produced from any initial mix-
ture of 2-butanone, 2-pentanone, and 2-hexanone. Because x4, x5,
and x6 must sum to 1 for this three methyl ketone mixture, only
two independent variables (x4 and x5) were required to generate
any possible boiling curve for hydrodeoxygenated cyclic trimers of
these materials. The exceptional agreement between this predic-
tive boiling model and the measured simulated distillation curve
can be seen in the Supporting Information. The method described
above was used to determine the optimal mixture of ketones re-
quired to produce a product that minimizes the squared residuals
between the commercial jet fuel and the predicted boiling curve.
A contour plot showing regions of minimal differences between
Metal/base-catalyzed reactions to dimer condensates
Approximately 2 mmol of 2-hexanol, 144 mg of dodecane (internal
standard), and 3 mL of toluene was added to a 12 mL glass Q-tube
reactor. The Pd/Mg-Al-O (2 wt% Pd) catalyst was then added to
the reactor at a loading of 1 mol% of Pd (with respect to 2-hexa-
nol). Under optimized conditions, an additional 100 mg of Mg-Al-O
was further added to increase the basic loading. A PTFE-coated
stir-bar was then added to the reactor and the reactor was sealed
with a PTFE-coated silicone septum. The reactor was then im-
mersed in a preheated oil bath at 1808C and stirred magnetically
at 500 rpm for 15 h. Upon completion of the reaction, the Q-tube
reactor was removed from the oil bath and processed in the same
manner as base-catalyzed condensation reactions.
Hydrodeoxygenation reactions
Hydrodeoxygenation reactions were carried out in an eight-unit
HEL ChemSCAN autoclave system. In
a typical experiment,
0.67 mmol of cyclic trimer condensates, 144 mg of nonane (internal
standard), 3 mL of cyclohexane, and 0.25–0.5 mol% Pt/NbOPO4
(total metal content basis) was added to the HEL Hastelloy auto-
clave reactor. A PTFE-coated stirbar was attached to the reactor
system and the autoclave was sealed. The reactor was purged
&
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
10
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
[15] Y. Gong, L. Lin, J. B. Shi, S. J. Liu, Molecules 2010, 15, 7946.
[16] K. Min, S. Kim, T. Yum, Y. Kim, B.-I. Sang, Y. Um, Appl. Microbiol. Biotech-
nol. 2013, 97, 5627.
jet fuel and the produced mixture of alkanes can be found in the
Supporting Information.
[17] T. Thananatthanachon, T. B. Rauchfuss, Angew. Chem. Int. Ed. 2010, 49,
6616; Angew. Chem. 2010, 122, 6766.
[18] P. Anbarasan, Z. C. Baer, S. Sreekumar, E. Gross, J. B. Binder, H. W. Blanch,
D. S. Clark, F. D. Toste, Nature 2012, 491, 235.
Acknowledgements
This work was supported by the Energy Biosciences Institute. E.R
S. also acknowledges support from a National Science Founda-
tion Graduate Research Fellowship under Grant No. DGE
1106400. The authors would also like to acknowledge the contri-
butions of Alice Yeh towards Dean–Stark experiments in the Sup-
porting Information.
[19] G. W. Huber, J. N. Chheda, C. J. Barrett, J. A. Dumesic, Science 2005, 308,
1446.
[20] a) J. I. Di Cosimo, V. K. Dꢁez, C. R. Apesteguꢁa, Appl. Catal. A 1996, 137,
149; b) J. I. Di Cosimo, C. R. Apesteguꢁa, J. Mol. Catal. A 1998, 130, 177;
c) V. A. Bell, H. S. Gold, J. Catal. 1983, 79, 286; d) C. P. Kelkar, A. A.
Schutz, Appl. Clay Sci. 1998, 13, 417; e) R. M. West, E. L. Kunkes, D. A. Si-
monetti, J. A. Dumesic, Catal. Today 2009, 147, 115.
[21] H. J. Seebald, W. Schunack, Arch. Pharm. 1972, 305, 785.
[22] J. R. Walton, B. Yeomans, (USPTO), US3981918A, 1976.
[23] W. B. Renfrow, A. Renfrow, J. Am. Chem. Soc. 1946, 68, 1801.
[24] J. T. Kozlowski, R. J. Davis, ACS Catal. 2013, 3, 1588.
[25] a) A. Corma, V. Fornꢂs, R. M. Martꢁn-Aranda, F. Rey, J. Catal. 1992, 134,
58; b) A. Corma, V. Fornes, F. Rey, J. Catal. 1994, 148, 205; c) M. J. Cli-
ment, A. Corma, S. Iborra, K. Epping, A. Velty, J. Catal. 2004, 225, 316;
d) A. Corma, S. B. A. Hamid, S. Iborra, A. Velty, J. Catal. 2005, 234, 340.
[26] Y. Liu, E. Lotero, J. G. Goodwin, X. Mo, Appl. Catal. A 2007, 331, 138.
[27] a) S. K. Sharma, P. A. Parikh, R. V. Jasra, J. Mol. Catal. A 2007, 278, 135;
b) E. R. Sacia, M. H. Deaner, Y. L. Louie, A. T. Bell, Green Chem. 2015, 17,
2393.
[28] a) R. M. West, M. H. Tucker, D. J. Braden, J. A. Dumesic, Catal. Commun.
2009, 10, 1743; b) N. Li, G. A. Tompsett, G. W. Huber, ChemSusChem
2010, 3, 1154; c) W. Xu, Q. Xia, Y. Zhang, Y. Guo, Y. Wang, G. Lu, Chem-
SusChem 2011, 4, 1758.
[29] ASTM D7153-05: Standard Test Method for Freezing Point of Aviation
Fuels (Automatic Laser Method), ASTM International, 2010.
[30] a) ASTM D1655-14: Standard Specification for Aviation Turbine Fuels,
ASTM International, 2013; b) ASTM D7566-14a: Standard Specification
for Aviation Turbine Fuel Containing Synthesized Hydrocarbons, ASTM
International, 2013.
[31] C. Vendeuvre, R. Ruiz-Guerrero, F. Bertoncini, L. Duval, D. Thiꢂbaut, M.-C.
Hennion, J. Chromatogr. A 2005, 1086, 21.
[32] G. Hemighaus, T. Boval, J. Bacha, Fred Barnes, M. Franklin, L. Gibbs, N.
Hogue, J. Jones, D. Lesnini, J. Lind, J. Morris, Aviation Fuels Technical
Review, Chevron Global Aviation, 2006.
[33] a) T. Holm, J. Chromatogr. A 1999, 842, 221; b) A. D. Jorgensen, K. C.
Picel, V. C. Stamoudis, Anal. Chem. 1990, 62, 683; c) J. T. Scanlon, D. E.
Willis, J. Chromatogr. Sci. 1985, 23, 333.
[34] ASTM D2887–13: Standard Test Method for Boiling Range Distribution
of Petroleum Fractions by Gas Ghromatography, ASTM International,
2013.
Keywords: biofuels · CꢀC coupling reactions · heterogeneous
catalysis · hydrogenation · jet fuel
[1] J. Han, A. Elgowainy, H. Cai, M. Q. Wang, Bioresour. Technol. 2013, 150,
447.
[2] a) M.-O. P. Fortier, G. W. Roberts, S. M. Stagg-Williams, B. S. M. Sturm,
Appl. Energy. 2014, 122, 73; b) P. Gegg, L. Budd, S. Ison, J. Air Transp.
Manage. 2014, 39, 34.
[3] A. Milbrandt, C. Kinchin, R. McCormick, The Feasibility of Producing and
Using Biomass-Based Diesel and Jet Fuel in the United States, (NREL)
NREL/TP-6A20-58015, 2013.
[4] a) C. Somerville, H. Youngs, C. Taylor, S. C. Davis, S. P. Long, Science
2010, 329, 790; b) J. Janaun, N. Ellis, Renewable Sustainable Energy Rev.
2010, 14, 1312; c) G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006,
106, 4044.
[5] T. Kalnes, T. Marker, D. R. Shonnard, Int. J. Chem. React. Eng. 2007, 5,
A48.
[6] M. Balakrishnan, E. R. Sacia, A. T. Bell, ChemSusChem 2014, 7, 1078.
[7] a) A. Corma, O. de La Torre, M. Renz, N. Villandier, Angew. Chem. Int. Ed.
2011, 50, 2375; Angew. Chem. 2011, 123, 2423; b) A. Corma, O. de La
Torre, M. Renz, Energy Environ. Sci. 2012, 5, 6328; c) G. Li, N. Li, J. Yang,
L. Li, A. Wang, X. Wang, Y. Cong, T. Zhang, Green Chem. 2014, 16, 594.
[8] J. Q. Bond, D. M. Alonso, D. Wang, R. M. West, J. A. Dumesic, Science
2010, 327, 1110.
[9] M. Mascal, S. Dutta, I. Gandarias, Angew. Chem. Int. Ed. 2014, 53, 1854;
Angew. Chem. 2014, 126, 1885.
[10] E. G. Eddings, S. H. Yan, W. Ciro, A. F. Sarofim, Combust. Sci. Technol.
2005, 177, 715.
[11] B. G. Harvey, M. E. Wright, R. L. Quintana, Energy Fuels 2010, 24, 267.
[12] J. Yang, N. Li, G. Li, W. Wang, A. Wang, X. Wang, Y. Cong, T. Zhang,
Chem. Commun. 2014, 50, 2572.
[13] X.-J. Ji, H. Huang, P.-K. Ouyang, Biotechnol. Adv. 2011, 29, 351.
[14] a) A. Multer, N. McGraw, K. Hohn, P. Vadlani, Ind. Eng. Chem. Res. 2013,
52, 56; b) A. V. Tran, R. P. Chambers, Biotechnol. Bioeng. 1987, 29, 343;
c) R. R. Emerson, M. C. Flickinger, G. T. Tsao, Ind. Eng. Chem. Prod. RD
1982, 21, 473.
Received: January 1, 2015
Published online on && &&, 0000
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
11
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

FULL PAPERS
E. R. Sacia, M. Balakrishnan, M. H. Deaner,
K. A. Goulas, F. D. Toste, A. T. Bell*
Cool your jets! Aviation fuel requires
a mixture of hydrocarbons having both
a high energy density and a low freez-
ing point. It is demonstrated that bio-
mass-derived n-alkyl methyl ketones un-
dergo selective, base-catalyzed reactions
to form cyclic trimer condensates in
>94% selectivity. These products are
then quantitatively hydrodeoxygenated
to form aviation fuels having a broad
range of volatility and exceptional freez-
ing points and energy densities when
utilizing multiple starting ketones.
&& – &&
Highly Selective Condensation of
Biomass-Derived Methyl Ketones as
a Source of Aviation Fuel
&
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
12
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!