Ultra-selective cycloaddition of dimethylfuran for renewable p-xylene with H-BEA

Article abstract of DOI:10.1039/c3gc40740c

p-Xylene, the precursor for PET bottles, was synthesized at 90% yield by [4 + 2] cycloaddition of biomass-derived ethylene and dimethylfuran followed by subsequent dehydration with Beta zeolite.

Full text of DOI:10.1039/c3gc40740c

Green Chemistry
View Article Online
View Journal
COMMUNICATION
Ultra-selective cycloaddition of dimethylfuran for
renewable p-xylene with H-BEA†
Cite this: DOI: 10.1039/c3gc40740c
Received 18th April 2013,
Accepted 21st May 2013
Chun-Chih Chang,‡ Sara K. Green,‡ C. Luke Williams, Paul J. Dauenhauer* and
Wei Fan*
DOI: 10.1039/c3gc40740c
www.rsc.org/greenchem
p-Xylene, the precursor for PET bottles, was synthesized at 90% process parameters highlights the need for improved selecti-
yield by [4 + 2] cycloaddition of biomass-derived ethylene and vity to p-xylene in the conversion of DMF.17
dimethylfuran followed by subsequent dehydration with Beta
zeolite.
Our research on the reaction of DMF and ethylene has
revealed the fundamental catalytic chemistry, but the optimal
catalytic system converted at least 25% of feedstock carbon
into unusable side products at great cost to overall process
viability.11 The reaction occurs by symmetry-allowed [4 + 2]
Diels–Alder cycloaddition of ethylene and subsequent aromati-
zation by acid-catalyzed dehydration in multiple elementary
steps to p-xylene (Scheme 1). Addition of ethylene to DMF pro-
duces an oxa-norbornene cycloadduct intermediate in a single
uncatalyzed step with a barrier of 24.7 kcal mol−1, determined
by gas-phase DFT.18 In the absence of an acid catalyst, de-
hydration is inhibited by a large barrier (58.0 kcal mol−1);
however, acid-catalyzed dehydration of 1,4-dimethyl-7-oxabi-
cyclo[2.2.1]hept-2-ene occurs readily (19 kcal mol−1), indicating
the need for a Brønsted-acid catalyst. While a previous work
has observed the cycloadduct intermediate with GC–MS,19 it
was not measured in this work. Elucidation of the reaction
pathways by separation of the reaction mixture and characteriz-
ation by NMR/GC–MS has revealed three competing side reac-
tions19 to the main pathway to p-xylene: (1) hydrolysis of DMF
The need for sustainable production of everyday materials in
addition to market volatility of petroleum-based feedstocks
has motivated research into the production of renewable aro-
matic chemicals from biomass.1–6 Cellulose and hemicellulose
comprise a significant fraction of lignocellulosic biomass and
provide six-carbon (e.g. glucose) and five-carbon (e.g. xylose)
sugars as primary feedstocks for the production of renewable
chemicals with six-carbon aromatic ring structures.7 Specific
chemicals of interest include p-xylene, the feedstock for poly-
ethylene terephthalate (PET), and toluene, an important
monomer for polyurethane products.8–10 In this work, we
reveal a highly selective method (>90% product selectivity and
90% yield for p-xylene formation at 99% conversion of
dimethylfuran) for the production of basic monomers from
biomass-derived furan feedstocks by combined catalytic cyclo-
addition/dehydration reactions using H-BEA zeolite.
We have proposed a renewable method of producing renew-
able p-xylene by cycloaddition of biomass-derived dimethyl-
furan (DMF) and ethylene,11 which serves as the last step in a
complete process for producing p-xylene from cellulose.
Glucose, produced by saccharification of cellulose, can be
isomerized with Sn-BEA12 to fructose, and subsequent acid-
catalyzed dehydration produces 5-hydroxymethylfurfural
(HMF).13,14 Hydrogenolysis of HMF over CuRu/C produces
DMF,15 and the final step adds ethylene (which can be derived
from ethanol dehydration)16 to yield p-xylene. Further investi-
gation into the design of this process indicates potential for
economic viability, and a sensitivity analysis of the considered
Department of Chemical Engineering, University of Massachusetts Amherst, 686
N. Pleasant St., 159 Goessmann Lab, Amherst, MA 01003, USA.
E-mail: wfan@ecs.umass.edu, dauenhauer@ecs.umass.edu; Tel: +1 413 545 2507
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c3gc40740c
Scheme 1 Diels–Alder cycloaddition of dimethylfuran [1] and ethylene pro-
duces an oxa-norbornene cycloadduct [2] which dehydrates to p-xylene [3].
Water hydrolyzes dimethylfuran [1] to 2,5-hexanedione [4] in equilibrium.
‡These authors contributed equally.
This journal is © The Royal Society of Chemistry 2013
Green Chem.

View Article Online
Communication
Green Chemistry
to 2,5-hexanedione, (2) secondary addition of DMF to produce both experimentally and by DFT.18,20 With this in mind, the
dimers, and (3) secondary reactions of ethylene to alkylated comparison of the activity of H-BEA (RDMF, the initial reaction
aromatics. From this analysis, further improvement of selecti- rate of DMF is equal to 47 3 mmol-pX (g-cat h)−1) with those
vity to p-xylene can be specifically defined as: (a) elimination of other non-porous materials such as WOx–ZrO2 (RDMF
=
of the hydrolysis side reaction or (b) reduction of secondary 50–55 mmol-pX (g-cat h)−1) shows comparable reaction activi-
addition of either the dienophile (e.g. ethylene) or the diene ties at the same temperature (250 °C).21
(e.g. DMF).
The potential of H-BEA for combined cycloaddition/de-
In this work we show that selectivity to p-xylene from DMF hydration chemistry relative to other materials is further revealed
and ethylene can be significantly enhanced by using the by long-time experiments examining high conversion of DMF.
H-BEA catalyst. The reaction details and analysis protocols can As shown in Fig. 2, most catalytic materials at 250 °C in
be found in the ESI.† As shown in Fig. 1, initial catalytic reac- heptane solvent exhibit a similar behavior with significant
tion rates reveal the activity of several solid-acid catalysts at deactivation after ∼10 hours. Niobic acid, γ-Al2O3, H-Y zeolite,
250 °C in heptane solvent at low conversion (XDMF, the conver- and H-ZSM-5 all exhibit only ∼50–60% conversion after
sion of DMF, <20%). While previous studies using pure DMF 30 hours. In contrast, H-BEA exhibits a significantly different
reactant at 300 °C have indicated that Y-zeolite is more active catalytic behavior and rapidly achieves a high conversion of
than H-BEA,11 the reaction with H-BEA (47 mmol (g-cat h)−1
in the presence of heptane solvent at 250 °C is at least 50%
)
∼90% in 10 hours under the same conditions.
The benefit of the high activity of H-BEA relative to other
more active than H-Y zeolite. Additionally, several other cata- considered materials is a higher selectivity to p-xylene of ∼90%
lysts including H-ZSM-5, niobic acid, and γ-Al2O3 are nearly an on a carbon basis. As shown in Fig. 3, selectivity to p-xylene
order of magnitude less active than H-BEA under these increases with DMF conversion. At XDMF = 20%, selectivity to
conditions.
p-xylene is only 60%, and the hydrolysis product (2,5-hexane-
The initial kinetics of solid-acid catalyzed cycloaddition in dione), alkylated products and oligomers comprise the
heptane solvent support the previously proposed mechanism remaining 40%, within experimental error. However, as XDMF
of p-xylene production, and highlight the importance of increases to >99%, selectivity to p-xylene exceeds 90% (concen-
H-BEA. As shown in Fig. 1, the initial reaction rate of p-xylene tration profiles for other catalysts can be found in Fig. S1–S4†).
is independent of the Si/Al ratios within given materials such This counter-intuitive behavior can be partially explained by
as BEA, which indicates that the overall reaction rate (at low the concentration of the hydrolysis side-product, 2,5-hexane-
conversion) is independent of the total number of Brønsted dione, as revealed in Fig. 3(a). For XDMF < 60%, the con-
acid sites for the considered reaction conditions. However, it centration of all side products increase in step with DMF
was shown previously that materials require a Brønsted acid conversion. However, while no change in concentration is
site to exceed the rate of cycloaddition/dehydration occurring observed for the alkylated products and oligomers after XDMF
>
within the non-catalyzed system.11 The necessity for a Brønsted 80%, the concentration of 2,5-hexanedione decreases from
acid site for which site-density does not correlate with activity 0.08 M for XDMF > 60% and reduces to nearly zero concen-
(at high catalyst loading) is consistent with the proposed tration for XDMF > 99%. This behavior is consistent with the
mechanism of rate-limiting, non-catalyzed Diels–Alder cyclo- hydrolysis reaction equilibrium between DMF–water and 2,5-
addition, described previously.11 H+ has been shown to exhibit hexanedione;22 as DMF reacts with ethylene, 2,5-hexanedione
no potential for decrease in the cycloaddition reaction barrier,
Fig. 1 Initial reaction rates of p-xylene formation for the reaction of DMF in
heptane with ethylene (62 bar) at 250 °C with various catalysts. The Si/Al ratios
of catalysts are shown in parentheses.
Fig. 2 Conversion of DMF as a function of time for the reaction of DMF in
heptane with ethylene (62 bar) at 250 °C with various catalysts. The Si/Al ratios
of catalysts are shown in parentheses.
Green Chem.
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Communication
■
Fig. 4 Product selectivity to ( ) p-xylene, toluene, and benzene; ( ) alkylated
compounds; and ( ) oligomer compounds at high conversion (>99% DMF and
MF, 70% furan). Reaction conditions: 1 M furanic feedstock in heptane with
ethylene (62 bar) at 250 °C with 0.45 0.05 g of catalyst. Error bars were less
than 2% for each compound.
The high activity of H-BEA is more pronounced for the con-
version of DMF to p-xylene than for the reaction of 2-methyl-
furan (MF)–ethylene to toluene or the reaction of furan–
ethylene to benzene. For the reaction of MF and ethylene, the
conversion of MF was 93% with H-BEA and 71% with H-Y after
24 hours of reaction under identical conditions to the experi-
ments in Fig. 2. Similarly, the conversion of furan with ethyl-
ene was 54% with H-BEA and 50% with H-Y after 24 hours (see
Fig. S6–S9†). Fig. 4 compares the product selectivities for con-
Fig. 3 (a) Selectivity of p-xylene and the concentration of 2,5-hexanedione and
(b) the concentration of other products as a function of DMF conversion for the
reaction of DMF in heptane with ethylene (62 bar) at 250 °C with H-BEA (Si/Al =
12.5).
version of DMF and MF at XDMF,MF > 90% and furan at Xfuran
>
70% at 250 °C. While 90% selectivity to p-xylene is achieved
for conversion of DMF, only 46% selectivity to toluene and
35% selectivity to benzene are observed. A significant loss of
carbon occurs in the reaction of MF and furan due to
increased reactions between the furans forming dimers (e.g.
benzofuran) and larger oligomers.21 The increase in these reac-
tions is likely due to the absence of methyl groups at the furan
α-carbon positions, which inhibit these side reactions.
reforms DMF to maintain equilibrium. p-Xylene formation is
highly favourable thermodynamically,11,18 and all reactants
including side products in equilibrium will yield p-xylene at
high conversion. This was confirmed by a separate experiment,
described in the ESI,† which started with 2,5-hexandione and
water in heptane and proceeded to form DMF and eventually
p-xylene in the reaction with ethylene at 250 °C (Fig. S5†).
By the DMF hydrolysis reaction, carbon, which appears lost
Conclusions
to the side product 2,5-hexanedione at low conversion, is recov- The discovery that H-BEA in heptane exhibits high selectivity
ered and results in high selectivity to p-xylene at high conver- with high catalytic reaction rate significantly improves the
sion of DMF. The superior performance of H-BEA for selective potential for lignocellulosic biomass to serve as a renewable
production of p-xylene can therefore be attributed to: (a) its feedstock for aromatic chemicals. Techno-economic analysis
resistance to deactivation, allowing for high conversion of of the process to convert glucose into HMF and eventually
DMF (Fig. 2), (b) its superior activity relative to other solid-acid p-xylene indicates that the use of H-BEA addresses two of the
materials (Fig. 1), and (c) its ability to catalyze dehydration of three most important design parameters for improved econ-
the Diels–Alder cycloadduct without catalyzing important omics.17,23 First, the high selectivity to p-xylene reduces the
side reactions. p-Xylene, once formed from DMF, does loss of feedstock carbon, which has been identified as
not readily isomerize to o- or m-xylene under the considered the largest process cost. Second, the ability to utilize the side
reaction conditions, which provides a significant process product, 2,5-hexanedione, by dehydration to DMF allows for
advantage by eliminating the expensive separation of xylene the recovery of carbon within the process and raises the value
isomers.
of the ring-opened side product. Future research addressing
This journal is © The Royal Society of Chemistry 2013
Green Chem.

View Article Online
Communication
Green Chemistry
the design of a continuous flow reaction system (based on a
standard engineered slurry reactor and separator) which allows
for continuous catalyst replacement and separation of p-xylene
from 2,5-hexanedione–DMF will allow for the hydrolysis side-
7 A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107,
2411–2502.
8 W. Partenheimer, Catal. Today, 1995, 23, 69–158.
9 M. C. Raes and O. J. Proulx, US Pat., 4506040A, 1985.
product to be recycled until it is completely consumed (as 10 P. Raghavendrachar and S. Ramachandran, Ind. Eng. Chem.
shown in the process diagram in the ESI, Scheme S1†), thus Res., 1992, 31, 453–462.
matching the selective performance of the high conversion 11 C. L. Williams, C.-C. Chang, P. Do, N. Nikbin,
batch reactor described here.
S. Caratzoulas, D. G. Vlachos, R. F. Lobo, W. Fan and
P. J. Dauenhauer, ACS Catal., 2012, 2, 935–939.
12 M. Moliner, Y. Román-Leshkov and M. E. Davis, Proc. Natl.
Acad. Sci. U. S. A., 2010, 107, 6164–6168.
13 E. Nikolla, Y. Román-Leshkov, M. Moliner and M. E. Davis,
ACS Catal., 2011, 1, 408–410.
Acknowledgements
This work was supported as part of the Catalysis Center for
Energy Innovation, an Energy Frontier Research Center funded 14 Y. Román-Leshkov, J. N. Chheda and J. A. Dumesic, Science,
by the U.S. Department of Energy, Office of Science, Office of
2006, 312, 1933–1937.
Basic Energy Sciences under award number DE-SC0001004.
15 Y. Román-Leshkov, C. J. Barrett, Z. Y. Liu and
J. A. Dumesic, Nature, 2007, 447, 982–985.
16 M. Stewart, H. Clark, D. Coffey, L. Luo and A. E. Schweizer
Jr., WO Pat., 2011002699A2, 2011.
17 Z. Lin, M. Ierapetritou and V. Nikolakis, AIChE J., 2013, 59,
2079–2087.
18 N. Nikbin, P. T. Do, S. Caratzoulas, R. F. Lobo,
P. J. Dauenhauer and D. G. Vlachos, J. Catal., 2013, 297,
35–43.
Notes and references
1 M. J. Climent, A. Corma and S. Iborra, Green Chem., 2011,
13, 520–540.
2 G. W. Huber, S. Iborra and A. Corma, Chem. Rev., 2006,
106, 4044–4098.
19 P. T. M. Do, J. R. McAtee, D. A. Watson and R. F. Lobo, ACS
Catal., 2013, 3, 41–46.
20 S. Bischoff and F. Kasper, J. Prakt. Chem., 1986, 328, 449–
453.
21 D. Wang, C. M. Osmundsen, E. Taarning and
J. A. Dumesic, ChemCatChem, 2013, 5, 1–8.
22 R. M. Dessau, Zeolites, 1990, 10, 205–206.
23 A. I. Torres, P. Daoutidis and M. Tsapatsis, Energy Environ.
Sci., 2010, 3, 1560–1572.
3 C. H. Christensen, J. Rass-Hansen, C. C. Marsden,
E. Taarning and K. Egeblad, ChemSusChem, 2008, 1, 283–
289.
4 T. R. Carlson, Y.-T. Cheng, J. Jae and G. W. Huber, Energy
Environ. Sci., 2011, 4, 145–161.
5 Y.-T. Cheng, J. Jae, J. Shi, W. Fan and G. W. Huber, Angew.
Chem., Int. Ed., 2012, 51, 1387–1390.
6 T. Carlson, G. Tompsett, W. Conner and G. Huber, Top.
Catal., 2009, 52, 241–252.
Green Chem.
This journal is © The Royal Society of Chemistry 2013