Diels-Alder and Dehydration Reactions of Biomass-Derived Furan and Acrylic Acid for the Synthesis of Benzoic Acid

Article abstract of DOI:10.1021/acscatal.5b01892

Routes to benzoic acid starting from furan-obtained from hemicellulose in high yield-and methyl acrylate are reported. These routes involve Diels-Alder and dehydration reactions of furan and acrylic acid (or methyl acrylate) in a two-step reaction protocol that minimizes side reactions. The Diels-Alder reaction of furan and methyl acrylate (or acrylic acid) was run at 298 K and was catalyzed by Lewis acidic (Hf-, Zr-, and Sn-Beta) zeolite catalysts, and achieving a high turnover frequency (～2 h^-1) and no side reactions were observed. The oxanorbornene product was dehydrated at low temperatures (298 to 353 K) in mixtures of methanesulfonic acid and acetic anhydride in 96% yield. This is compared to an only 1.7% yield of methyl benzoate obtained for the dehydration of the oxanorbornene in neat methanesulfonic acid. The effect of oxanorbornene concentration and stereochemistry was found not to decrease the yield of aromatics, while dehydration of the carboxylic acid form of the oxanorbornene led to a decrease in selectivity to 43% at complete conversion in mixtures of methanesulfonic acid and acetic anhydride. This reaction sequence could be an important entry point for selectively directing high-yield, hemicellulose-derived furans to aromatic products used in the existing chemical process industry.

Full text of DOI:10.1021/acscatal.5b01892

Subscriber access provided by UNIV OF NEBRASKA - LINCOLN
Article
Diels-Alder and Dehydration Reactions of Biomass-Derived
Furan and Acrylic Acid for the Synthesis of Benzoic Acid
Eyas Mahmoud, Jingye Yu, Raymond J. Gorte, and Raul F. Lobo
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.5b01892 • Publication Date (Web): 09 Oct 2015
Downloaded from http://pubs.acs.org on October 11, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street
N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
DielsꢀAlder and Dehydration Reactions of Biomassꢀ
Derived Furan and Acrylic Acid for the Synthesis of
Benzoic Acid
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
b
b
a*
Eyas Mahmoud, Jingye Yu, Raymond J. Gorte, and Raul F. Lobo
a
Catalysis Center for Energy Innovation, Department of Chemical and Biomolecular
Engineering, University of Delaware, Newark, DE 19716, USA.
b
Department of Chemical and Biomolecular Engineering, University of Pennsylvania,
Philadelphia, PA 19104, USA.
*Corresponding author. Eꢀmail: lobo@udel.edu
ABSTRACT: Routes to benzoic acid starting from highꢀyield, hemicelluloseꢀderived furfural
derivative, furan, manufactured from biomass are reported. These routes involve DielsꢀAlder and
dehydration reactions of furan and acrylic acid (or methyl acrylate) in a twoꢀstep reaction
protocol that minimizes polymerization side reactions associated with furan and acrylic acid. The
DielsꢀAlder reaction of furan and methyl acrylate (or acrylic acid) was run at 298 K catalyzed by
Lewis acidic (Hfꢀ, Zrꢀ and SnꢀBeta) zeolite catalysts and relatively high turnover frequency (~2
ꢀ1
h ) and no side reactions were observed. The dehydration of the oxanorbornene was performed
homogeneously at low temperatures (298 to 353 K) in mixtures of methanesulfonic acid and
ACS Paragon Plus Environment
1

ACS Catalysis
Page 2 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
acetic anhydride in 96 % yield. This is compared to only a 1.7 % yield of methyl benzoate
obtained for the dehydration of the oxanorbornene in neat methanesulfonic acid. The effect of
oxanorbornene concentration and stereochemistry were found not to significantly decrease yield
to aromatics while dehydration of the carboxylic acid form of the oxanorbornene led to a drop in
selectivity to 43 % at complete conversion in mixtures of methanesulfonic acid and acetic
anhydride. This reaction sequence could be an important entry point for selectively directing
highꢀyield, hemicelluloseꢀderived furans to aromatic products used in the existing chemical
process industry.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
KEYWORDS: Lewis acid, DielsꢀAlder, zeolite Beta, dehydration, mixed anhydrides
ACS Paragon Plus Environment
2

Page 3 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
INTRODUCTION
The shale gas revolution has created opportunities for the development of efficient processes for
the production of aromatic commodity chemicals from biomass because the shift in cracker
feedstock to natural gas will reduce the availability of benzene, toluene and xylenes (BTX), the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
aromatic building blocks of the chemical industry. One approach to the synthesis of biomassꢀ
derived BTX aromatics is by the DielsꢀAlder cycloaddition of a biomassꢀderived furan and
dienophile followed by dehydration of the oxanorbornene intermediate to form the aromatic
product. Two biomassꢀderived furans obtained from the cellulosic and hemicellulosic fractions
of lignocelluose are 5ꢀhydroxymethylfurfural (HMF) and furfural. The global market of furfural
was estimated to have been around 300,000 tons in 2013. Furfural yields of up to 96 % can be
obtained from cornꢀstover. Thus, finding efficient routes to aromatics from furfural is of interest.
Sequential DielsꢀAlderꢀdehydration routes to aromatics from HMF and furfural have
been investigated, and typically, reactions of furfural derived furans are more unselective. 2,5ꢀ
Disubstituted furans, such as oxidized derivatives of 5ꢀhydroxymethylfurfural (HMF) and 2,5ꢀ
dimethylfuran (DMF), have been reacted by sequential DielsꢀAlderꢀdehydration reactions with
ethylene to produce terephthalic acid intermediates and pꢀxylene with high selectivity (> 81 %
under the best conditions and HMF derivative) using Lewis and Brønsted acidꢀcontaining solid
catalysts. This reaction is particularly selective when using 2,5ꢀdimethylfuran as a reactant where
yields as high as 90 % have been obtained. The high selectivity observed is presumably due to
protecting methyl groups at the alpha and beta carbon positions of the furan ring. In the case of
hemicellulose derived furans, such as 2ꢀmethylfuran and furan, selectivity to the aromatic
products, toluene and benzene, decreases to 46 and 35 % at 90 % conversion due to
polymerization sideꢀreactions of the unsubstituted furan. Although routes for the conversion of
ACS Paragon Plus Environment
3

ACS Catalysis
Page 4 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
unsubstituted furans to aromatics exist, side reactions associated with the polymerization of furan
and the use of high pressure ethylene at high temperatures make these routes more challenging to
commercially implement. Therefore, the discovery of routes to toluene and particularly benzeneꢀ
derived monomers for consumer products and goods that are selective could be useful.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Benzene is transformed to monomers used for the synthesis of Nylon 6, polycarbonates,
phenolic resins, and epoxy resins that are used for the manufacture of everyday consumer goods
such as DVDs and toothbrushes. The yearly production of benzene is about 21 million tons, a
portion of which is alkylated to isopropylbenzene (cumene) and oxidized to produce acetone and
phenol (3.4 million tons produced annually). Benzene is also hydrogenated to cyclohexane which
is converted to caprolactam for Nylon 6 synthesis. Biomassꢀderived benzoic acid can serve as a
dropꢀin replacement to produce many of the consumer products and polymers typically produced
by benzene (Figure 1). Using known industrial processes, benzoic acid can be transformed, into
highꢀvolume commodity chemicals such as phenol, caprolactam, glycol dibenzoates and sodium
3
and potassium benzoate, key intermediates in the synthesis of consumer products and polymers.
ACS Paragon Plus Environment
4

Page 5 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
Figure 1. Conversion of benzoic acid to nylon 6, polycarbonates, phenolic resins, and epoxy
resins.
Here, we report DielsꢀAlder and dehydration routes to selectively produce benzoic acid
from biomass beginning with the furfuralꢀderived molecule, furan. Furan and methyl acrylate (or
acrylic acid) are shown to react in a twoꢀstep DielsꢀAlder and dehydration reaction protocol to
produce methyl benzoate or benzoic acid (Scheme 1). Furan is industrially manufactured from
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
0
biomass by the decarbonylation of furfural in 99 % yield. Acrylic acid could be produced
renewably from biodiesel (glycerol) by catalytic routes or from lignocellulose by chemocatalytic
and biological routes. For example, acrylic acid can be produced from glucose by retroꢀaldol
2
1
condensation to form lactic acid followed by dehydration. Biologically, glucose can be
converted to acrylic acid in 92% yield using E. coli and there are current efforts aimed at
22
commercializing the production of renewable acrylic acid.
Scheme 1. Reaction sequence for the conversion of furan to methyl benzoate.
A biomassꢀbased route to benzoic acid that has previously been considered is the formic
acidꢀmediated deoxygenation of the glucoseꢀderived materials quinic acid and shikimic acid. A
oneꢀstep process achieving ~90 % yield of benzoic acid was reported from both quinic acid and
shikimic acid. Quinic acid can be isolated from natural sources such as cinochona bark, coffee
beans, tobacco or by fermentative production from glucose. Here, we report a route to benzoic
ACS Paragon Plus Environment
5

ACS Catalysis
Page 6 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
acid from biomassꢀderived furfural, which is commercially produced using improvements to the
Quaker Oates batch process originally implemented in the 1920s.
We explored a number of different catalytic materials and conditions for the selective
synthesis of benzoic acid and found that a twoꢀstep DielsꢀAlder and dehydration reaction
procedure is best at this time. Running the reaction in a twoꢀstep procedure is of particular
importance to avoid unwanted side reactions of furan and methyl acrylate (or acrylic acid) in the
presence of Lewis and Brønsted acids required for these DielsꢀAlder and dehydration reactions.
Molecular sieves having the zeolite Beta topology containing Lewis acidic centers in the +4
oxidation state are excellent catalysts for DielsꢀAlder reactions of furan. The absence of strong
Brønsted acidity in these materials is important to avoid unwanted side reactions of furan. The
activity of Hf, Zr and SnꢀBeta Lewis acidic zeolites as compared to other Lewis and Brønsted
acid homogeneous and heterogeneous catalysts for DielsꢀAlder reactions of furan has not been
reported. We also explored a number of different catalytic and nonꢀcatalytic strategies for the
dehydration of the DielsꢀAlder adduct (oxanorbornene carboxylic methyl ester or acid) and
found that dehydration in binary mixtures of methanesulfonic acid (MSA) and acetic anhydride
containing the mixed anhydride, acetyl methanesulfonate is most selective. Acetyl
methanesulfonate is formed by the reaction of MSA and acetic anhydride (Scheme 2). This
compound in known to behave as a strong acylating agent in a variety of other reactions, such as
for ether bond cleavage and esterification. The effect of oxanorbornene concentration,
stereochemistry and functional group on the dehydration of these compounds using mixed
anhydrides has not been reported. Taken together, these findings suggest that a twoꢀstep Dielsꢀ
Alder and dehydration reaction protocol using Lewis acidicꢀzeolites and mixed anhydrides could
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment
6

Page 7 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
be applied to the synthesis of other valuable aromatic chemicals such as benzophenone and
benzaldehyde from furfural or its derivatives in high selectivity.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Formation of mixed sulfonicꢀcarboxylic anhydride from MSA and acetic anhydride.
RESULTS AND DISCUSSION
Materials Characterization
Xꢀray diffraction, nitrogen adsorption and scanning electron microscopy (SEM) were used to
characterize the microporous and mesoporous catalysts tested for DielsꢀAlder and dehydration
reactions for renewable benzoic acid synthesis. Samples of Hfꢀ, Zrꢀ and SnꢀBeta with a Si/M
ratios~100 and 200 (M=Hf, Zr or Sn containing 3.38, 2.01 and 2.45 or 1.57, 1.08 and 1.08 wt%
HfO , ZrO , and SnO respectively) were prepared based on existing protocols (see experimental
2
2
2
2
5
section for details). As a control, samples containing extraꢀframework (EF) Hf, Zr and Sn
oxides, prepared by impregnation over a sample of allꢀsilica zeolite Beta, were also investigated.
The Xꢀray diffraction patterns of the calcined samples are consistent with the structure of zeolite
Beta, with no evidence of impurities and undetectable amounts of amorphous material (Figure
S1). Micropore volumes measured using nitrogen adsorption isotherms show that Hfꢀ, Zrꢀ, Snꢀ
3
and SiꢀBeta all have the high microporous volumes (~0.20 cm /g) characteristic of zeolite Beta
2
4
samples of good quality (Table S1). Nitrogen adsorption experiments of the EFꢀMꢀBeta
M=Hf, Zr, and Sn) indicate that the metal is impregnated on the external surface of the zeolite
(
2
6
particles forming metal oxide (MO ) nanoparticles, as was previously observed. The Xꢀray
2
ACS Paragon Plus Environment
7

ACS Catalysis
Page 8 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
diffraction pattern of SnꢀMCMꢀ41 (Si/Sn=50) was well defined and displayed the characteristic
2
7
(
100), (110) and (200) reflections corresponding to hexagonal array channels (Figure S2). The
nitrogenꢀadsorption isotherm of SnꢀMCMꢀ41 displayed the type IV profile characteristic of
mesoporous MCMꢀ41 materials (Figure S3) with the expected high single point BET surface
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
ꢀ1
27
area of 720 m g (determined at P/P =0.302). SEM images of the synthesized zeolites show the
0
2
8
typical morphology of zeolite Beta prepared in fluoride media having a truncated square
2
9
bipyramidal particle shape and a size between 2ꢀ5 micrometers (Figure 2a).
Spectroscopic investigations of the synthesized zeolite Beta catalysts were performed to
29
verify the effect of metal precursor addition before zeolite crystallization on the catalyst. Si
MAS NMR spectroscopy shows resolved signals at ca.116, 115, 113, 112, and 111 ppm and very
few framework defects (tetrahedralꢀatom vacancies) in all cases, but broader signals for Hf, Sn,
and ZrꢀBeta as compared to SiꢀBeta (Figure S4). This broadening, which increases as metal
content increases, greatly reduces the resolution of the experiments and only three broad
resonances are apparent at ca. ꢀ115, ꢀ113, and ꢀ111 ppm for zeolites with Si/M ratios of around
100. This was previously observed in the case of Ti and Zrꢀheteroatom incorporation in zeolite
3
0
Beta. UVꢀVis spectroscopy shows a progression in the energy of the absorption edge in the
order HfO < EFꢀHfꢀBeta < HfꢀBeta, which has been previously reported for the Sn analogue for
2
3
1
this series. An increase in the absorption edge energy is correlated with a decrease in the metal
cluster size, suggesting smaller clusters of HfO in HfꢀBeta than in EFꢀHfꢀBeta. A progression in
2
the energy of absorption edge was observed for the different metals tested in the order of EFꢀSnꢀ
Beta < EFꢀZrꢀBeta < EFꢀHfꢀBeta (Figure S5), in the samples impregnated with the metal oxides.
ACS Paragon Plus Environment
8

Page 9 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Characterization of zeolite Beta samples by (a) SEM, (b) FTIR and (c) TPDꢀTGA. (a)
SEM images of the synthesized zeolites. Scale bar 6 µm.(b) FTIR spectra of MꢀBeta (i) in pure
He flow at RT immediately after removing CD CN vapor; in He flow after heating to 333 K for
3
(ii) 1 min; (iii) 2 min; (iv) 3 min; (v) 5 min. (c) TPD (lower curves) and TGA (upper curves)
traces of diethyl ether dosed onto MꢀBeta (M=Hf, Sn, Zr, or Si). Diethyl ether was monitored in
the mass spectra by its m/e=31 peak. The Si/M ratios of the zeolites are Si/Hf=219, Si/Sn=230,
and Si/Zr=325 respectively.
Fourierꢀtransform infrared (FTIR) spectra of acetonitrileꢀd adsorbed in MꢀBeta, SiꢀBeta
3
and EFꢀMꢀBeta (Figure 2b) confirm the presence of acetonitrileꢀd interaction with Lewis acid
3
ACS Paragon Plus Environment
9

ACS Catalysis
Page 10 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
sites only in zeolites synthesized with the metal precursors added before zeolite crystallization
(
MꢀBeta zeolites). For example, spectra of HfꢀBeta at room temperature after acetonitrileꢀd3
ꢀ1
exposure show two ν (CꢀN) stretches, centered at 2264 cm corresponding to a bulk acetonitrileꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ1
d stretch and a higher frequency 2302 cm peak corresponding to acetonitrileꢀd (CꢀN) stretch
3
,
3
associated with acetonitrileꢀd interaction with the Lewis acid site. As acetonitrileꢀd is removed
3
3
ꢀ1
from the sample by mild heating and continued flushing with He, the 2264 cm feature
ꢀ1
ꢀ1
decreases, while the feature at 2302 cm narrows and shifts to 2306 cm . The Zrꢀ and SnꢀBeta
samples show similar behavior with the final higher frequency peak shifts being 2301 and 2305
ꢀ1
cm , respectively. The SiꢀBeta sample does not show any features at the higher frequency after
evacuation at higher temperatures confirming the absence of adsorption sites for acetonitrileꢀd3.
FTIR spectra of acetonitrileꢀd on EFꢀMꢀBeta, prepared by impregnation of the metal precursor
3
after zeolite crystallization, were identical to those for SiꢀBeta, a result that is consistent with no
Lewis acids present which interact with the acetonitrileꢀd in these samples (Figure S6). The
3
observed turnover frequencies of the DielsꢀAlder reaction of furan and acrylic acid when
comparing EFꢀMꢀ and MꢀBeta zeolites are consistent with FTIR experiments (see Table 1).
MꢀBeta and EFꢀMꢀBeta zeolites were also characterized by temperature programmed
desorption (TPD) thermogravimetric analysis (TGA) to probe whether the Lewis acids exist as
isolated active sites. It has been shown that diethyl ether forms oneꢀtoꢀone complexes with
framework sites in SnꢀBeta by temperature programmed desorption (TPD) thermogravimetric
31
analysis (TGA). Similar probeꢀmolecule studies using diethyl ether were conducted to
characterize our samples. HfꢀBeta (Hf/Si=219) adsorbs 80 µmol/g of diethyl ether, an amount
that is very close to the Hf concentration of 72 µmol/g as determined by ICPꢀOES (Figure 2c).
Similarly, Zrꢀ and SnꢀBeta samples adsorb 55 and 43 µmol/g of diethyl ether. On the other hand,
ACS Paragon Plus Environment
10

Page 11 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
SiꢀBeta only adsorbed 10 µmol/g. The TPD traces indicate at least a +10 K peak temperature
shift when comparing metalꢀcontaining to SiꢀBeta zeolites with shoulder peaks shifted + 80 K or
greater for Hf and ZrꢀBeta zeolites (Figure 2c).The difference in peak temperature and shape
when comparing Si to MꢀBeta zeolites suggests that there are different adsorption sites for these
two classes of materials. However, because desorption from porous materials is coupled with
diffusion and readsorption, caution must be used in the interpretation of peak. The calculated
activation energies of desorption of these shoulder peaks for Hf and ZrꢀBeta assuming
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
3
ꢀ1
monomolecular adsorption of diethyl ether and a preꢀexponential factor of 10 s are 136 and
133 kJ/mol respectively. Correspondingly, the activation energy of desorption of diethyl ether
from SnꢀBeta is 111 kJ/mol. Increasing the metal concentration in Hfꢀ and ZrꢀBeta zeolites led to
an increase in diethyl ether adsorption to 150 and 120 µmol/g for Si/Hf=127 and Si/Zr=149,
which is in excess of the theoretical adsorption amounts for these materials of 130 and 110
µmol/g as determined by ICPꢀOES. These results are consistent with previous TPDꢀTGA studies
on SnꢀBeta zeolite (Si/Sn=118) showing that while some of the excess diethyl ether is associated
with silanol groups in the zeolite—or unspecific adsorption—, the diethyl ether coverage
approximately corresponds to the calculated concentration of MꢀBeta zeolite, suggesting the
3
1
formation of 1:1 complexes of diethyl ether and the Lewis acid sites. TPDꢀTGA of diethyl
ether dosed on EFꢀMꢀBeta samples, which were prepared by metal impregnation of SiꢀBeta
zeolite (see experimental methods section), were similar to that dosed onto SiꢀBeta (Figure S7).
The TPDꢀTGA results indicate that the Hfꢀ, Zrꢀ and Snꢀ metal cations are present as isolated
metal sites in uniformly dispersed in MꢀBeta samples as supported by UVꢀVis spectroscopy, N2
2
9
adsorption and Si MAS NMR experiments. TPDꢀTGA results also indicate that metal cations in
ACS Paragon Plus Environment
11

ACS Catalysis
Page 12 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
EFꢀMꢀBeta samples exist as metal clusters on the external surface of the zeolite particles as
supported by UVꢀVis and N adsorption experiments.
2
Sequential Diels-Alder-Dehydration Reaction of Furan and Methyl Acrylate
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Furan and methyl acrylate were reacted in heptane to screen for activity in the oneꢀstep
sequential DielsꢀAlderꢀdehydration reaction to methyl benzoate (Scheme 1) using heterogeneous
Brønsted (HꢀBeta) and Lewis acidic (Hfꢀ, Zrꢀ and SnꢀBeta) zeolite catalysts. In all cases for the
heterogeneous catalysts tested for the DielsꢀAlderꢀdehydration reaction (473 K for 7 h with 40
psi of nitrogen), no aromatic product was detected. This was accompanied by fouling of the
catalysts as indicated by the darkening (coking) of the solids. The darkening of the catalyst,
which occurred at low conversions (<3 %) of methyl acrylate, suggests that these catalysts are
deactivated by deposition of byproducts. Doubling the catalyst amount of HfꢀBeta leads to an
increased conversion (11 %) of methyl acrylate but still no aromatics production (423 K for 7 h
with 40 psi nitrogen under solventꢀfree conditions). DielsꢀAlderꢀdehydration reaction has been
run successfully using 2,5ꢀdimethylfuran and ethylene (>99 % conversion 90 % selectivity to
aromatics) at high temperature because those two reactants are much more stable in the presence
1
3ꢀ14, 32
of acids than furan and methyl acrylate.
Aromatics production by DielsꢀAlderꢀdehydration
1
reactions of 2, 5ꢀdisubstituted furans and methyl acrylate conversion to unknown (by H NMR)
side products illustrate two of the major challenges for this DielsꢀAlderꢀdehydration reaction: the
sensitivity of furan to the Brønsted and Lewis acids used and the polymerization of methyl
acrylate at these conditions. To overcome these two challenges, the DielsꢀAlder and dehydration
reactions were run in two separate steps.
Diels-Alder Reaction
ACS Paragon Plus Environment
12

Page 13 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
The DielsꢀAlder reaction of furan and methyl acrylate was run without solvent to minimize
potential costs and the environmental impact of the process. Without a catalyst no reaction was
observed over 24 h at 298 K. To assess the role of thermodynamic equilibrium on yield, the
equilibrium constant was determined for a range of temperatures (see Supporting Information
and Figure S8 for experimental details), and using Van’t Hoff’s relation the equilibrium
conversion as a function of temperature was obtained (See Figure S9). These data indicate that
the reaction is kinetically limited at temperatures below 300 K and equilibrium limited at
temperatures above 330 K. Therefore a highly active catalyst is needed to achieve high
conversion at low temperatures (330 K or below).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
0a
Fraile et al investigated the reaction of methyl acrylate and furan (298 K, 24 h) using
titanium chloride and zinc chloride supported on silica and achieving 75% conversion of methyl
ꢀ1
acrylate. Although conversion is high, turnover frequency was only 0.025 h for TiCl /SiO , the
4
2
most active catalyst. Furthermore, the reacting mixture contained 67% wt/wt catalyst. We
investigated Hfꢀ, Zrꢀ and Snꢀcontaining zeolite Beta catalysts due to their demonstrated ability to
23
24
catalyze the BaeyerꢀVillager and MPV reactions by coordination to the carbonyl groups of
the reactants. For example, using in situ IR spectroscopy SnꢀBeta has been shown to activate the
23
carbonyl group of cyclohexanone for the BaeyerꢀVilliger oxidation. The three materials are
very effective catalysts for the DielsꢀAlder reaction of furan; HfꢀBeta, in particular, catalyzes this
ꢀ1
reaction at a turnover frequency of 1.9 h at 298 K (average over 24 h under solventꢀfree
ꢀ1
conditions). A turnover frequency of 3.7 h is achieved at 308 K after 3 h. A 25% yield of the
oxanorbornene was achieved after 24 h of reaction at 298 K using HfꢀBeta (Si/Hf=219), at a
ꢀ1
ꢀ1
corresponding turnover frequency of 1.9 h (Table S2). Turnover frequencies of 2.4 h and 2.0
ꢀ1
h were achieved using Zr (Si/Zr=325) and SnꢀBeta (Si/Sn=230) zeolites after 24 h of reaction at
ACS Paragon Plus Environment
13

ACS Catalysis
Page 14 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
98 K. The turnover frequencies for Hf, Zr, and SnꢀBeta zeolites are approximately one hundred
10a
times higher than those observed for TiCl /SiO at 298 K. For the Si/M~200 materials which
4
2
were prepared by an unseeded synthesis method, catalyst amount was directly proportional to
product yield (Figure S10). However, Si/M~100 catalysts, prepared using a seeded synthesis
method, have lower turnover frequencies (Table S2) and in these samples there is no 1:1
correlation between catalyst amount and moles of product produced, with a lower product
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ1 ꢀ1
production rate than expected (114.74 vs. the 180.72 µmol product gcat h calculated based on
turnover rates for the ~200 materials). This may be caused by the increasing abundance of silanol
groups in the Si/M~100 materials as compared to the unseeded Si/M~200 zeolites (see Figure
1
29
S11 for Hꢀ Si CPꢀMAS NMR spectra) which leads to the increasing hydrophilic character of
3
0a
the zeolite and possibly catalyst poisoning.
For example, HfꢀBeta (Si/Hf=127) which had nearly twice the amount of silanols, as
determined by CPꢀMAS as compared to ZrꢀBeta (Si/Zr=149), possessed roughly half the
turnover frequency of its counterpart (Table S2). Because the amount of product formed and
catalyst amount were not proportional for these materials, the Si/M~100 zeolite catalysts were
not investigated further. Running the DielsꢀAlder reaction of furan and methyl acrylate using Hꢀ
Beta—that is, a catalyst with Brønsted acid sites—resulted in a much lower turnover frequency
ꢀ1
of the metal site (0.19 h ) as compared to the Lewis acid metalꢀcontaining catalysts at 31% yield
of the oxanorbornene product. It also led to a variety of byproducts due to the reactivity of furans
with Brønsted acids.
Investigations of the DA reaction of furan and acrylic acid, the more reactive dienophile,
for a series of heterogeneous and homogeneous catalysts showed that Hfꢀ, Zrꢀ and SnꢀBeta have
ꢀ1
nearly the same reactivity with an turnover frequency of ~5 h at roughly 50% yield of the
ACS Paragon Plus Environment
14

Page 15 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
oxanorbornene (Table 1). In fact, the allꢀsilica SiꢀBeta (no other metal heteroatoms present)
provided a 13% yield showing that confinement of furan and acrylic acid alone is sufficient to
catalyze the formation of this DielsꢀAlder adduct, although at much lower rates. Hafnium,
zirconium and tin impregnated SiꢀBeta zeolite catalysts (EFꢀHf, EFꢀZr, and EFꢀSnꢀBeta
respectively), that is, materials with similar composition that do not contain Hf, Zr, or Sn present
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
in the synthesis gel, show almost no reactivity achieving only 13% conversion. Bulk HfO , tested
2
ꢀ1
independently, shows reaction rates that are more than ten times slower (0.092 h ) than HfꢀBeta.
When using the Brønsted acid zeolite, HꢀBeta, as a catalyst for the DielsꢀAlder reaction of furan
ꢀ1
and acrylic acid, a much lower turnover frequency of 0.35 h at a 44% yield of the
1
oxanorbornene was detected by H NMR. This was accompanied by darkening of the catalysts
associated with side reactions of furan indicating that Lewis acidꢀcontaining analogs of zeolite
Beta are more effective catalysts for DielsꢀAlder reactions of furan at these conditions. Using
mesoporous tinꢀsilicate SnꢀMCMꢀ41 as a catalyst for the DielsꢀAlder reaction of furan and
acrylic acid resulted in a 15% yield of the corresponding oxanorbornene and a turnover
ꢀ1
frequency of the tin site of 1.4 h after 24 h of reaction. The lower turnover frequency and yields
when comparing SnꢀMCMꢀ41 and SnꢀBeta materials suggest that confinement effects in the
zeolite Beta framework play a role in reactivity. Also, based on adsorption shifts from deuterated
ꢀ1
acetonitrile studies, the more blueꢀshifted hydrolyzed site of SnꢀBeta zeolite (2317 cm ) is a
ꢀ1 33
stronger Lewis acid than the tin site present in SnꢀMCMꢀ41 (2312 cm . These results
demonstrate the superior activity of Hf, Sn, and ZrꢀBeta zeolites compared to bulk metal oxides
and metalꢀimpregnated SiꢀBeta zeolite.
Table 1. Yield and Turnover Frequency of
the DielsꢀAlder Reaction of Furan and Acrylic
Acid Catalyzed by Lewis Acidic Zeolite Beta
ACS Paragon Plus Environment
15

ACS Catalysis
Page 16 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
and other Materials
Turnover
Metal:Hf Yield Frequency
Molar
Catalyst
c
ꢀ1 d
(%)
[h ]
b
Ratio
Hf-Beta
Sn-Beta
Zr-Beta
SiꢀBeta
1.0
1.0
0.68
ꢀ
51
48
31
13
5.1 (3.8)
5.0 (3.6)
4.5 (2.6)
ꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
EFꢀHfꢀBeta
1.1
12
1.1
~0.00)
1.5 (0.27)
1.3
(
EFꢀSnꢀBeta
EFꢀZrꢀBeta
1.1
1.1
16
14
(
0.090)
Hꢀ Beta
12
44
0.35
HfO2
SnꢀMCMꢀ
11
1.1
10
15
0.092
1.4
41
ZnCl2
4.9
11
18
8
0.36
0.076
AlCl *6H
3
2
O
Reactions were run under solvent free conditions at 298 K for
4 h. 6.9 mmoles of furan and 1.79 mmoles of acrylic acid
2
and 0.10 g of zeolite Beta catalyst corresponding to 0.0075
mmol loading of Hf or 0.42 mol% Hf with respect to acrylic
acid (Si/Hf=219). Metal contents of the zeolites in bold were
determined by ICPꢀOES. All reactions were ~100% selective
to the oxanorbornene with respect to the acrylic acid. An
identical mass of SiꢀBeta catalyst was used compared to the
amount of metal containing zeolite catalysts.
Reactions run without a catalyst gave no yield of
oxanorbornene under these reaction conditions.
The molar ratio is defined as the moles of metal, where the
metal can be Hf, Zr, Sn, Al, or Zn in catalyst divided by the
moles of Hf in HfꢀBeta (0.0075 mmoles).
Yield was calculated with respect to acrylic acid conversion.
Turnover frequency calculated by moles of oxanorbornene
produced divided by moles of metal in catalyst. Values in
parenthesis are turnover frequencies calculated after
subtracting the yield of product associated with SiꢀBeta.
a
b
c
d
Because the DielsꢀAlder reaction of furan and methyl acrylate (or acrylic acid) is an
exothermic reversible reaction, an optimal temperature profile can be devised to maximize the
yield at shorter times using a combination of kinetics (Figure S12), thermodynamics, and
ACS Paragon Plus Environment
16

Page 17 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
approximate kinetic rate expressions (see Equations S2 and S3). Equation S3 was used to find a
(optimal) temperature profile using MATLAB. Using the temperature of 4 h at 308 K and 20 h at
298 K, a 30% yield of the oxanorbornene was achieved while minimizing side reactions. Higher
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
temperatures were also applied for the DielsꢀAlder reaction of furan and methyl acrylate. For
example, running the reaction at 308 K for 24 h provided a 46% yield of the DA adduct of
methyl acrylate and furan. However, 16% of the limiting reactant, methyl acrylate, was
1
converted to undetectable sideꢀproducts (by H NMR spectroscopy). Using another temperature
program of 3 h at 313 K, 5 h at 308 K, and 24 h at 298 K, a 67% yield of the DA adduct of
acrylic acid and furan was obtained.
HfꢀBeta zeolite was also shown to be recyclable and maintained similar reactivity over
three runs for the DielsꢀAlder reaction of furan and acrylic acid after washing with methanol and
drying at 393 K overnight; regeneration by calcination did not lead to a significant drop in
activity demonstrating the robustness of this catalyst for this reaction (Figure S13). TGA studies
of the regenerated catalyst indicated that there was no significant leaching of Hf after each
reaction cycle as evidenced by the adsorption of ~90 µmol/g of diethyl ether (experimental error
ꢀ1
of TGA ~20 µmol/g). FTIR spectra also show a 2310 cm adsorption shift associated with
CD CN interaction with Hf in HfꢀBeta. XRD patterns of the regenerated catalyst indicated no
3
significant degradation of the catalyst structure. Finally, the catalytic properties of Hfꢀ, Zrꢀ, and
SnꢀBeta for other smallꢀmolecule DielsꢀAlder reactions were investigated (Table 2). HfꢀBeta
zeolite was able to catalyze the reaction of furans and various alphaꢀBeta unsaturated ketones
ꢀ1
and acids. Turnover frequencies of 2.80 and 3.88 h for the reaction of 2ꢀmethyl and 2,5ꢀ
dimethylfuran with acrylic acid were similar to those observed for the DielsꢀAlder reaction of
ꢀ1
furan and acrylic acid of 5.08 h . The reaction of furan and methyl acrylate possessed a turnover
ACS Paragon Plus Environment
17

ACS Catalysis
Page 18 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ꢀ1
frequency of 2.50 h . For all reactions, endo:exo ratios greater than one were observed,
indicating that the kinetically favored product was made in a larger proportion. Running the
DielsꢀAlder reaction of furan and methyl acrylate at 298 K for five days led to a decrease in the
endo:exo ratio from 1.9 to 1.7 suggesting a greater extent of retroꢀDielsꢀAlder reaction at longer
reaction times as previously seen for the DielsꢀAlder reaction of cyclopentadiene and pꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
benzoquinone catalyzed by Lewis acidic ITQꢀ2. These experiments show that HfꢀBeta is not
only able to selectively catalyze the cycloaddition at higher turnover frequencies, but it is also
reusable.
Table 2. Yield and Turnover Frequency of
the DielsꢀAlder Reaction of Various Furans
and Acrylates Catalyzed by HfꢀBeta Zeolite
Diene
Dienꢀ Yield
ophile (%)
a
Turnover
Frequency [h ]
ꢀ1 b
3
9
3.9
28
2.8
5
1
5.1
1.9
NA
2
5
0
a
Yields reported are for the oxanorbornene products
Turnover frequency calculated by moles of oxanorbornene
produced divided by moles of Hf metal in HfꢀBeta.
Reactions were run under solvent free conditions at 298 K
for 24 h using 0.50 mL of diene and 0.12 mL of dienophile.
Correspondingly, 4.6 mmoles of 2,5ꢀdimethylfuran, 5.6
mmoles of 2ꢀmethylfuran, or 6.9 mmoles of furan and 1.79
mmoles of acrylic acid or 1.36 mmoles of methyl acrylate
were added and 0.0075 mmol loading of Hf (Si/M=219).
b
ACS Paragon Plus Environment
18

Page 19 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
Dehydration of the Diels-Alder Adduct
The dehydration of oxanorbornenes is difficult to accomplish with high selectivity because these
3
4
molecules are sensitive to heat and they can polymerize in the presence of acids. RetroꢀDielsꢀ
Alder, polymerization, and dehydration reactions are all competing channels for the
transformation of the DA adducts and are the principal factors determining product
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
4b, 35
selectivity.
In the case of renewable benzoic acid production, dehydration of oxanorbornene
carboxylic methyl ester catalyzed by the Brønsted acid catalyst, HꢀBeta zeolite, yielded methyl
benzoate in less than 1 % selectivity at 100 % conversion after 7 h of reaction at 423, 473 and
523 K. The oxanorbornene reactant was readily converted by retroꢀDielsꢀAlder reaction to
methyl acrylate and furan. The selectivity to methyl acrylate decreased with increasing
temperature from 35 % at 423 K to 33 and 29 % at 473 and 523 K, presumably due to
decomposition sideꢀreactions. Furan was also detected following reaction though in much lower
selectivity (5.2, 4.1 and 4 %) as compared to methyl acrylate at the temperatures investigated.
The catalyst fouled as indicated by the change in color possibly due to the formation of organic
byproducts from the decomposition of furan, methyl acrylate and/or the oxanorbornene. These
results suggest that the catalytic dehydration of oxanorbornene compounds at higher
temperatures is unselective due to decomposition of the reactant to its retroꢀDielsꢀAlder products
as dictated by thermodynamics (Figure S9) and the decomposition of the retroꢀDielsꢀAlder
products leading to catalyst fouling.
To suppress the retroꢀDielsꢀAlder reaction, which accounts for at least 35% of the
decomposition of the oxanorbornene intermediate at 423 K, the dehydration reaction was run
homogeneously at lower temperatures in the Brønsted acid, MSA. MSA is a nonꢀoxidizing,
36
biodegradable and a thermally stable alternative to sulfuric acid. Running the reaction in neat
ACS Paragon Plus Environment
19

ACS Catalysis
Page 20 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
MSA (at a temperature of 298 K for 2 h followed by a temperature increase to 353 K for 1 h)
produced methyl benzoate but only with low selectivity (1.7%, see Table 3). The reaction
medium was modified to increase selectivity to methyl benzoate by using acetyl
methanesulfonate, a mixed anhydride easily formed from MSA and acetic anhydride, to open the
ether bond of the oxanorbornene. Under the same reaction conditions, the acetyl
methanesulfonate system produced methyl benzoate in 96% selectivity at 100% conversion. The
increase in selectivity to methyl benzoate (from 1.7 to 96 % at 100 % conversion) when
comparing MSA to acetyl methanesulfonate demonstrates the unique properties of the acyl group
to open the ether bond selectively for oxanorbornenes as compared to Brønsted acids. Running
the reaction using oxanorbornene carboxylic acid, the DA adduct of furan and acrylic acid as the
reactant, led to a drop in the selectivity of aromatics from 96% to 43% at 100% conversion under
identical conditions of temperature program, concentration and time. Because of the significantly
lower selectivity, dehydration of the oxanorbornene ester is the preferred route for aromatics
synthesis from oxanorbornene carboxylates as compared to the acid.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 3. Selectivity of the Dehydration of
Oxanorbornene Carboxylic Methyl Ester in Binary
a
Mixtures of MSA and Acetic Anhydride
Reactant
Concentration
Endo:Exo
Ratio
Selectivity
b
(%)
(M)
0
0
.13
.65
0.43
0.43
96
83
ACS Paragon Plus Environment
20

Page 21 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
0
.65
1.89
89
Reactions were run in 31 mmoles of MSA and 5.3 mmoles of acetic
anhydride at 298 K for 2 h followed by temperature increase to 353
K for 1 h.
Reactions run in neat MSA resulted in a 1.7% selectivity to the
product at 100% conversion (0.13 M concentration, 0.43 endo:exo
ratio of oxanorbornene carboxylic methyl ester).
a
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
b
Selectivity at 100% conversion
To identify the reasons for the increase in selectivity when comparing Brønsted to acyl
promoted dehydration of oxanorbornenes, the dehydration reaction was monitored using
quantitative nuclear magnetic resonance spectroscopy (qNMR) to identify stable reaction
intermediates. In the case of Brønsted acid (MSA) promoted dehydration, no reaction
intermediate was detected. In the case of acyl group promoted dehydration (acetyl
methanesulfonate), a stable reaction intermediate was detected. The production of a stable
intermediate decreases side reactions associated with the retroꢀDielsꢀAlder reaction and allows
one to run the reaction temperature programs. The intermediate was isolated by flash
1
chromatography and characterized by H and 2D NMR spectroscopy. The observation of a ringꢀ
opened acetate intermediate suggests the reaction mechanism depicted in Figure 3. The acyl
group is an electrophile and attacks the resonanceꢀstabilized cation formed from ringꢀopening of
the bridging oxygen of the oxanorbornene leading to the more stable ringꢀopened product, 5ꢀ
(methoxycarbonyl)cyclohexꢀ3ꢀeneꢀ1,2ꢀdiyl diacetate. Using the qNMR technique, an optimal
temperature reaction profile for the dehydration of this oxanorbornene was devised by first
running at 298 K for 2 h to minimize the retroꢀDielsꢀAlder reaction and to form a stable
intermediate. The temperature is then increased to 353 K for 1 h to convert the intermediate to
methyl benzoate.
ACS Paragon Plus Environment
21

ACS Catalysis
Page 22 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Proposed reaction mechanism for the dehydration of oxanorbornene carboxylic
methyl ester. The intermediate (5ꢀ(methoxycarbonyl)cyclohexꢀ3ꢀeneꢀ1,2ꢀdiyl diacetate) is
marked with an asterisk. The reaction was run at 298 K for 2 h (blue chemical structures) to
form a stable intermediate then at 353 K for 1 h (red chemical structures) to drive the reaction to
completion.
The dehydration reaction was also run at different concentrations of oxanorbornene,
endo:exo ratios to test the robustness of this dehydration approach (Table 3). Increasing the
concentration fiveꢀfold led to lower (but still acceptable) selectivity of 83% at full conversion.
Running the reaction at the same concentration, a change in the endo:exo ratio of the
oxanorbornene from 0.43 to 1.89 led only to a slight increase in the selectivity from 83 to 89% at
100% conversion. The selectivity for the dehydration of oxanorbornene carboxylic acid
decreased slightly from 43 to 41% selectivity at 100% conversion when running at a higher
ACS Paragon Plus Environment
22

Page 23 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
concentration (0.36 M) at an identical endo:exo ratio and identical reaction conditions. Although
this dehydration process consumes one mole of acetic anhydride for each mole of benzoic acid
produced, the acetic acid produced can be recycled into acetic anhydride using wellꢀknown
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
37
process chemistry. These results indicate that high concentrations of oxanorbornene can be
dehydrated selectively within the stoichiometry of the dehydration reaction with the mixed
anhydride.
Conclusions
We report the discovery of routes to toluene and benzeneꢀderived monomers for consumer
products and goods from a highꢀyield hemicellulose derived furan that are selective. Benzoic
acid is a potential platform to prepare a number of commodity chemicals as it can be transformed
to precursors to nylon 6, polycarbonates, phenolic resins, and epoxy resins, polymers which
make up four of the six major polymers supplied by fossilꢀfuelꢀderived benzene (Figure 1). We
explored a number of different catalytic materials and conditions for the selective synthesis of
benzoic acid and found that a twoꢀstep DielsꢀAlder and dehydration reaction procedure is best at
this time (Figure 2). Running the reaction in a twoꢀstep procedure is of particular importance to
avoid unwanted side reactions of furan and methyl acrylate (or acrylic acid) in the presence of
Lewis and Brønsted acids required for these DielsꢀAlder and dehydration reactions. It has been
shown that Hfꢀ, Zrꢀ, and SnꢀBeta zeolites selectively catalyze the thermodynamically limited
DielsꢀAlder reactions of furans with acrylates with very high activity as compared to other
homogenous and other heterogeneous Lewis acid catalyst with no side reactions observed. It was
also shown that mixtures of MSA and acetic anhydride selectively dehydrate DielsꢀAlder adduct
with yields of up to 96 %, with little effect of the stereochemistry of the DielsꢀAlder adduct and
concentration on the observed selectivity to the aromatic product. Overall, furan can be
ACS Paragon Plus Environment
23

ACS Catalysis
Page 24 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
selectivity transformed into benzoic acid in a 96% combined selectivity of the DielsꢀAlder and
dehydration reactions. These findings suggest that a twoꢀstep DielsꢀAlder and dehydration
reaction protocol using Lewis acidicꢀzeolites and mixed anhydrides could be applied to the
synthesis of other valuable aromatic chemicals such as benzophenone and benzaldehyde from
biomassꢀderived furfural or its derivatives in high selectivity.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Experimental Methods
Materials Synthesis.
Hf, Zr, Sn and SiꢀBeta were prepared by a hydrothermal synthesis method. 15 g of
tetraethylorthosilicate (TEOS, Sigma Aldrich, 99%) was added drop wise to a solution 16.032 g
of tetraethylammonium hydroxide (TEAOH, 35 wt%, Sachem) with constant stirring. After
mixing for 90 min, either HfCl O·8H O (Alfa Aesar, 98+ %), ZrCl O·8H O (Sigma Aldrich,
2
2
2
2
9
8%) or SnCl ·5·H O (Sigma Aldrich, 98%) in 1.943 g of DI water were added. The mixture
4 2
was aged at room temperature until the solution decreased in weight by 13 g due to ethanol
evaporation; additional 1.15 g of water. To the resulting solution, 1.59 g of HF (Acros, 48 wt%)
was added causing the formation of a thick paste. The final gel composition was as follows: 1.0
SiO : 0.005 MO : 0.54 TEAOH: 7.5 H O: 0.54 HF (M=Hf, Zr, and Sn). Si/M~100 samples were
2
2
2
similarly prepared except with the addition of 0.18 g (4 wt. %) dealuminated Beta seed crystals
38
after HF addition. The crystallization was carried out in rotating (40 rpm) Teflonꢀlined stainless
steel autoclaves (Parr) at 413 K for 14 days. The product was collected by filtration and washed
thoroughly with DI water. The zeolites were calcined in air in a twoꢀstep process first at 393 K
for 2 h and then at by 853 K for 3 h using a ramping rate of 3 K/min. HfꢀBeta catalyst was
ACS Paragon Plus Environment
24

Page 25 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
regenerated at identical conditions as calcination. SiꢀBeta was prepared following an identical
procedure except without the addition of a metal source and dealuminated Beta seed crystals.
To determine the effect of Lewis acid metal addition to the zeolite sample, a sample was
prepared by simple impregnation of M (M=Hf, Zr, Sn) into SiꢀBeta (EFꢀMꢀBeta) to compare to
samples where Lewis acid metal addition was done during hydrothermal synthesis. The
appropriate amount of HfCl O·8 H O, ZrCl O·8 H O or SnCl · 5H O were added to 0.3 g of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
2
2
2
4
2
water. The solution was added dropwise to 0.3 g of SiꢀBeta with continuous mixing. The
powder was stirred for 4 h at room temperature then put into a convection oven at 373 K
overnight. The sample was then calcined in air using a ramping rate of 5 K/min up to a
temperature of 853 K and held at this temperature for 4 h.
SnꢀMCMꢀ41 with a Si/Sn ratio of 50 was prepared following a previously reported
27
procedure . HfO (Sigma Aldrich, 98%), ZnCl (Sigma Aldrich), and AlCl ·6H O (Sigma
2
2
3
2
Aldrich, 99%) were used as purchased. HꢀBeta (Guild associates) contained 17.5% clay binder
by weight and a SiO /Al O ratio of 34 (including binder).
2
2
3
Analytical Methods.
Xꢀray powder diffraction (XRD) patterns were recorded on a Phillips X’Pert Xꢀray
diffractometer using Cu Kα radiation. The micropore volumes of samples were determined
using nitrogen adsorption isotherms at 77 K measured using a Micromeritics 3ꢀFlex instrument.
All samples were degassed for 8 h under vacuum at 573 K before adsorption measurements. The
micropore volumes and surface areas of Beta zeolites were determined using the tꢀplot method.
SEM images were taken using a JSMꢀ7400 F high resolution scanning electron microscope.
ACS Paragon Plus Environment
25

ACS Catalysis
Page 26 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Diels-Alder Reactions.
Furan (SigmaꢀAldrich, 99+ %), methyl acrylate (SigmaꢀAldrich, 99%, contains ≤100 ppm
monomethyl ether hydroquinone as inhibitor) and acrylic acid (SigmaꢀAldrich, 99.0+ %) were
stored under an inert atmosphere and used as received. DielsꢀAlder reactions were performed in
20 mL thickꢀwalled glass vials (SigmaꢀAldrich). Reactions were run under solvent free
conditions at 298 K for 24 h. 6.9 mmoles of furan, 1.79 mmoles of acrylic acid and 0.10 g of
zeolite Beta catalyst or the appropriate amount of homogeneous Lewis acid catalyst were loaded
into the reactor. For reactions run using SiꢀBeta catalyst, an identical mass of catalyst was used
as for the Lewis acid metalꢀcontaining zeolites. Reactions run using methyl acrylate were run
using 1.36 mmoles of methyl acrylate. For all reactions, 0.166 mmoles of pꢀxylene (Sigma,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
≥
99%) was used as an internal standard, which was added before addition of the catalyst. The
reactor was flushed with N gas before reaction, sealed and transferred to an optimchem digital
2
balance stirrer temperature controlled at 298 K for reaction.
Yield and selectivity were determined by using the internal standard (pꢀxylene) and the
peaks corresponding to the endo and exo isomers of the cycloadduct (3.18 and 2.49 ppm for
oxanorbornene carboxylic acid, 3.76 and 3.67 ppm for the ester and 2.9 and 2.57 ppm for the
DielsꢀAlder adduct of 2,5ꢀdimethylfuran and acrylic acid). For the DielsꢀAlder reaction of 2ꢀ
methylfuran and acrylic acid, the moles of acrylic acid converted were used to calculated yield
and selectivity since it was known that the mass balance is closed under these conditions of
reaction. NMR spectra were recorded at 298 K with a Bruker AVIII600 NMR spectrometer
1
operating at 600.13 MHz ( H). A test of linearity for acrylic acid using the qNMR parameters
2
13
gave a correlation coefficient R > 98%. C NMR, 2D COSY, 2D HMBC, and 2D HMQC were
run using standard parameters of the Bruker AV400 spectrometer.
ACS Paragon Plus Environment
26

Page 27 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
The oxanorbornenes were isolated by flash chromatography. The solution was filtered
with a 0.2 micron filter (Corning) and rotavapped for 1 h at 298 K to remove furan. The
remaining solution was run through a flash chromatography column containing Versa Flash
spherical flash silica (45ꢀ75 ꢁm) using a 2:1 solution of hexane (Fisher Chemical, 99.9%) to
ethyl acetate (Fisher Chemical, 99.9%) as the mobile phase.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Oxanorbornene
carboxylic
methyl
ester
(endoꢀ
and
exoꢀ2ꢀcarbomethoxyꢀ7ꢀ
3
9
oxabicyclo[2.2.1]heptꢀ5ꢀene) :
1
H NMR (600 MHz, Chloroformꢀd) δ 6.44 (dd, J = 5.9, 1.7 Hz, 1Hendo), 6.39 (dd, J = 5.9, 1.6 Hz,
H ), 6.35 (dd, J = 5.8, 1.7 Hz, 1H ), 6.23 (dd, J = 5.8, 1.5 Hz, 1Hendo), 5.21 – 5.18 (m,
1
exo
exo
1
3
1
H ), 5.18 – 5.14 (m, 1H ), 5.10 – 5.06 (m, 1H ), 5.05 – 5.00 (m, 1H
), 3.73 (s, 3Hexo),
exo endo exo endo
.65 (s, 3Hendo), 3.12 (dt, J = 9.0, 4.2 Hz, 1Hendo), 2.44 (dd, J = 8.6, 3.9 Hz, 1H ), 2.17 (dt, J =
exo
1.6, 4.3 Hz, 1H ), 2.11 (ddd, J = 11.4, 9.3, 4.7 Hz, 1Hendo), 1.62 – 1.57 (m, 1Hendo, 1H ).
exo
exo
3
9
Oxanorbornene carboxylic acid (endoꢀ and exoꢀ2ꢀcarboxylicꢀ7ꢀoxabicyclo[2.2.1]heptꢀ5ꢀene) :
1
H NMR (600 MHz, Chloroformꢀd) δ 6.45 (dd, J = 5.8, 1.7 Hz, 1Hendo), 6.41 (dd, J = 5.8, 1.6 Hz,
1
H ), 6.36 (dd, J = 5.8, 1.7 Hz, 1H ), 6.30 (dd, J = 5.8, 1.6 Hz, 1Hendo), 5.22 (d, J = 1.8 Hz,
exo exo
1Hendo), 5.21 – 5.17 (m, 1H ), 5.12 – 5.08 (m, 1H ), 5.04 (dd, J = 4.9, 1.7 Hz, 1Hendo), 3.17
exo
exo
(
ddd, J = 9.0, 4.8, 3.8 Hz, 1Hendo), 2.47 (dd, J = 8.5, 3.8 Hz, 1H ), 2.20 – 2.07 (m, 1Hendo and
exo
1Hexo ), 1.62 – 1.54 (m, 1Hendo and 1Hexo).
Dehydration Reaction.
ACS Paragon Plus Environment
27

ACS Catalysis
Page 28 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Methanesulfonic acid (99.5%+, Sigma Aldrich) and acetic anhydride (99.4%, Fisher
Scientific) were used as received and stored under N . In a 20 mL disposable scintillation vial,
2
3
1 mmoles of MSA and 7.7 mmoles of MSA or 5.3 mmoles of acetic anhydridein addition to
.208 mmol (0.025 grams) of sulfolane (99%+, Sigma Aldrich) which was used as an internal
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
0
standard. Sulfolane was selected as internal standard because of its high stability against strong
acids as well as good thermal stability. The mixture was then stirred for 5 minutes at room
temperature. In the case of acetic anhydride, MSA and acetic anhydride reacted to form acetyl
1
40
methanesulfonate [ H NMR (400 MHz, CDCl ) δ 2.28 (3H, s, COCH ), 3.33 (3H, s,SO CH )].
3
3
2
3
The oxanorbornene was slowly added while stirring. The vial was then flushed with nitrogen.
Conversion and selectivity were calculated using the internal standard by qNMR.
The intermediate was isolated using a flash chromatography column containing Versa
Flash spherical flash silica (45ꢀ75 ꢁm) using a 3:1 solution of hexane (Fisher Chemical, 99.9%)
to ethyl acetate (Fisher Chemical, 99.9%) as the mobile phase.
4
,5ꢀdiacetoxycyclohexꢀ2ꢀenecarboxylic acid (Note, the methyl group attached to the carboxylic
acid group of 5ꢀ(methoxycarbonyl)cyclohexꢀ3ꢀeneꢀ1,2ꢀdiyl diacetate is hydrolyzed during flash
chromatography as was observed for methyl benzoate):
1
H NMR (600 MHz, Chloroformꢀd) δ 6.31 (ddt, J = 8.7, 6.8, 1.7 Hz, 1H), 5.71 (dt, J = 9.6, 2.3
Hz, 1H), 5.51 (q, J = 2.5 Hz, 1H), 5.00 (dt, J = 5.6, 2.4 Hz, 1H), 3.01 (dd, J = 6.8, 4.7 Hz, 1H),
2.57 (dt, J = 11.4, 5.4 Hz, 1H), 2.24 (d, J = 11.8 Hz, 1H), 2.17 (s, 4H), 2.14 (s, 3H). HRMS (EI)
m/z, calcd for [C H O ]: 201.0763; found: 201.
9
12
5
See Figures S14 and S15 for 2D NMR COSY and HMQC spectra of this compound.
ACS Paragon Plus Environment
28

Page 29 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
TPD-TGA.
TPDꢀTGA experiments were carried out using a CAHN 1000 microbalance mounted
within a high vacuum chamber, as described elsewhere. The base pressure of the system was
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢀ8
1
×10 Torr. During TPDꢀTGA measurements, the sample weight could be recorded
continuously using the microbalance and the desorbing species monitored using a UTI
quadrupole mass spectrometer. The peaks intensities at selected m/e from the mass spectra are
reported here in arbitrary units (a.u.). The sample temperature was measured with a
thermocouple placed near the sample, and the heating rate during desorption was maintained at
1
0 K/min by a temperature controller. Unless otherwise noted, the samples were heated in vacuo
to 773 K, then cooled in vacuo, prior to exposing them to the vapors of the adsorbate of interest.
The solid samples were exposed to diethyl ether at room temperature using 3 Torr of vapor until
there was no further uptake. The system was then evacuated for 1 h before beginning the TPDꢀ
TGA experiment.
UV-Vis Spectroscopy. The samples were analyzed using a UVꢀVvis spectrometer (Jasco Vꢀ550)
equipped with a diffuseꢀreflectance cell.
FTIR Spectroscopy. FTIR spectra were collected on a Mattson Galaxy FTIR with a diffuseꢀ
reflectance attachment (Collector II) purchased from SpectraꢀTech Inc. Spectra were collected at
−
1
2
cm resolution. The samples were initially heated to 773 K in flowing He to remove any
water. After cooling the samples to room temperature, acetonitrileꢀd3 vapors were exposed to the
3
sample, with the excess acetonitrile then flushed from the solid using flowing He at 1 cm /s.
ACS Paragon Plus Environment
29