Modulating Reactivity and Selectivity of 2-Pyrone-Derived Bicyclic Lactones through Choice of Catalyst and Solvent

Article abstract of DOI:10.1021/acscatal.7b04311

2-Pyrones, such as coumalic acid, are promising biobased molecules that through Diels-Alder reactions can provide access to a wide range of biobased chemicals, including molecules with functionality that are not easily accessible via conventional petrochemical routes. A complete reaction network and kinetic parameters for three individual diversification routes that start from a single bicyclic lactone produced via the Diels-Alder cycloaddition of coumalic acid and ethylene were examined experimentally and probed through complementary first-principle density functional theory (DFT) calculations, in situ nuclear magnetic resonance (NMR) spectroscopy, and thin film solid-state NMR spectroscopy. These experiments provide insights into the routes for several molecular structures from bicyclic lactones by leveraging Lewis or Br?nsted acid catalysts to selectively alter the reaction pathway. The bicyclic lactone bridge can be decarboxylated to access dihydrobenzenes at a substantially reduced activation barrier using γ-Al₂O₃ as the catalyst or selectively ring-opened via Br?nsted acids to yield 1,3-diacid six membered rings. DFT computations and microkinetic modeling in combination with experimental results provide molecular insights into the catalytically active sites on γ-Al₂O₃ and provide a general mechanism for the catalyzed bicyclic lactone decarboxylation in polar aprotic solvents, which involves CO₂ extrusion as the kinetically relevant step. Solid-state NMR spectroscopy provides direct evidence of strong binding of the bicyclic lactone to the γ-Al₂O₃ surface, fully consistent with DFT simulation results and experimental reaction kinetics. In addition, the role of the solvent was examined and found to be an additional means to improve reaction rates and selectively produce alternative structures from the bicyclic intermediate. The rate of the decarboxylation reaction was increased dramatically by using water as the solvent whereas methanol acted as a nucleophile and selectively induced ring-opening, showing that both pathways are operative in the absence of catalyst. Taken together, the results demonstrate an approach for selective diversification of the coumalate platform to a range of molecules.

Full text of DOI:10.1021/acscatal.7b04311

Subscriber access provided by UNIVERSITY OF THE SUNSHINE COAST
Article
Modulating Reactivity and Selectivity of 2-Pyrone-Derived
Bicyclic Lactones through Choice of Catalyst and Solvent
Toni Pfennig, Ashwin Chemburkar, Robert L. Johnson, Matthew
Ryan, Aaron J Rossini, Matthew Neurock, and Brent H. Shanks
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.7b04311 • Publication Date (Web): 12 Feb 2018
Downloaded from http://pubs.acs.org on February 14, 2018
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 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 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 15
ACS Catalysis
1
2
3
4
5
6
Modulating Reactivity and Selectivity of 2-Pyrone-Derived Bicyclic
Lactones through Choice of Catalyst and Solvent
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,d
c,d
a,d,†
b
b
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
Toni Pfennig, Ashwin Chemburkar, Robert L. Johnson, Matthew J. Ryan, Aaron J. Rossini,
c,d
,a,d
Matthew Neurock, Brent H. Shanks*
a
Department of Chemical and Biological Engineering, Iowa State University, Ames, IA, 50011, USA
Department of Chemistry, Iowa State University, Ames, IA, 50011, USA
b
c
d
Department of Chemical Engineering and Materials Science, University of Minnesota, Minneapolis, MN 55455, USA
NSF Engineering Research Center for Biorenewable Chemicals (CBiRC), Ames, IA 50011, USA
Coumalic acid diversification, renewable dihydrobenzenes, renewable aromatics, benzoic acid, isophthalic acid, decarboxyꢀ
lation, DielsꢀAlder
2
ꢀPyrones, such as coumalic acid, are promising bioꢀbased molecules that through DielsꢀAlder reactions can provide access to a
wide range of bioꢀbased chemicals, including molecules with functionality that are not easily accessible via conventional petroꢀ
chemical routes. A complete reaction network and kinetic parameters for three individual diversification routes that start from a
single bicyclic lactone produced via the DielsꢀAlder cycloaddition of coumalic acid and ethylene were examined experimentally
and probed through complementary firstꢀprinciple density functional theory (DFT) calculations, in situ nuclear magnetic resonance
(NMR) spectroscopy and thin film solidꢀstate NMR spectroscopy. These experiments provide insights into the routes for several
molecular structures from bicyclic lactones by leveraging Lewis or Brønsted acid catalysts to selectively alter the reaction pathway.
The bicyclic lactone bridge can be decarboxylated to access dihydrobenzenes at a substantially reduced activation barrier using γꢀ
Al O as the catalyst or selectively ringꢀopened via Brønsted acids to yield 1,3ꢀdiacid six membered rings. DFT computations and
2
3
microkinetic modeling in combination with experimental results provide molecular insights into the catalytically active sites on γꢀ
Al O and provide a general mechanism for the catalyzed bicyclic lactone decarboxylation in polar aprotic solvents, which involves
2
3
CO extrusion as the kinetically relevant step. Solidꢀstate NMR spectroscopy provides direct evidence of strong binding of the biꢀ
2
cyclic lactone to the γꢀAl O surface, fully consistent with DFT simulation results and experimental reaction kinetics. In addition,
2
3
the role of the solvent was examined and found to be an additional means to improve reaction rates and selectively produce alternaꢀ
tive structures from the bicyclic intermediate. The rate of the decarboxylation reaction was increased dramatically by using water as
the solvent whereas methanol acted as a nucleophile and selectively induced ringꢀopening, showing that both pathways are operaꢀ
tive in the absence of catalyst. Taken together, the results demonstrate an approach for selective diversification of the coumalate
platform to a range of molecules.
tions. Renewable CMA can be synthesized from malic acid via
acid catalyzed dimerization, with malic acid being readily
Introduction
6
New technologies to extract nonꢀrenewable carbon feedstocks,
such as shale gas, exert significant economic pressure on the
potential to manufacture bioꢀbased commodity chemicals.
This challenge creates the need to develop flexible chemical
conversion platforms that can be used to pivot from biomassꢀ
derived bulk chemicals to higher value chemicals. Leveraging
novel transformation pathways from biomass feedstocks via
bioprivileged molecules provides an attractive means to access
higher value chemicals such as nextꢀgeneration nutraceuticals,
antimicrobials, pharmaceuticals, consumer goods, specialty
chemicals, etc., in a manner that is cost competitive compared
produced from biomass via fermentation with genetically
7ꢀ9
modified microorganisms. Given that malic acid has been
cited as one of the 12 most promising bioꢀbased platform
10
chemicals, considerable effort has already been invested into
the biochemical conversion of glucose to malate. A particularꢀ
ly interesting feature of 2ꢀpyrone DielsꢀAlder reactions is the
formation of stable bicyclic lactones that are highly functionalꢀ
ized, stereochemically rich building blocks which can be used
1
1ꢀ22
as versatile synthetic intermediates.
these structures leads to synthetically valuable dihydrobenꢀ
zenes
Decarboxylation of
20,23ꢀ25
(DIH) with potential applications in the manufacꢀ
1
to traditional petrochemical routes.
ture of medically useful products (Scheme 1). DIH can then be
utilized for chemical diversification, granting access to polyꢀ
As a potential bioprivileged molecule platform, DielsꢀAlder
conversion of 2ꢀpyrone coumalic acid (CMA) provides a highꢀ
ly versatile pathway to the synthesis of a wide range of bioꢀ
based chemicals through the utilization of a variety of dienoꢀ
11,18,19,21,24,26,27
2ꢀ6,11,18,22,25,26
cyclic systems
and aromatics
with
wideꢀranging applications. Although bicyclic compounds can
be accessed from 2ꢀpyrones in high yield, their selective transꢀ
formation via thermal decarboxylation can be challenging due
2ꢀ6
philes. Particularly important are those dienophiles that are
less expensive and lead to more efficient downstream separaꢀ
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 15
acid catalyzed dehydration leading to cyclohexaꢀ1,3ꢀdieneꢀ
1,3ꢀdicarboxylic acid in high selectivity.
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
The impact of solvents in steering the reaction was also exꢀ
plored. Using solvents that can serve as weak nucleophiles
(e.g. methanol) provides an alternative path for transforming
(1) to (5)ꢀ(methoxycarbonyl)cyclohexaꢀ1,5ꢀdieneꢀ1ꢀcarboxylic
acid with high selectivity (>99 mol%). Conversely, polar proꢀ
tic water leads to enhanced decarboxylation of (1) to (2) as
compared to the polar aprotic 1,4ꢀdioxane. These routes offer
catalystꢀfree pathways with the potential to be broadly applied
to bicyclic lactone molecules. Lastly, we applied these findꢀ
ings to synthesize benzoic acid (3) by using a bifunctional 1
wt% Pd/γꢀAl O catalyst to activate the rate limiting decarꢀ
Scheme 1. DielsꢀAlder conversion of 2ꢀpyrones to form bicyꢀ
clic lactones followed by decarboxylation to dihydrobenzenes.
23
4ꢀ6,11,18,25
to degradation or in situ aromatization
, which limit
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
the selective access of interesting and potentially valueꢀadded
diene structures in high purity.
2
3
Inꢀsitu aromatization of these conjugated diene species is esꢀ
pecially favored when the desired diene contains a leaving
group (e.g. alkoxy), demonstrated by reactions of methyl
coumalate in combination with ketals, acetals, captodative
dienophiles or vinyl ethers, which lead to high aromatic yields
boxylation step, reduce the reaction temperature, and show the
full conversion of (2) to (3) with a selectivity > 99 mol%.
Results and Discussion
Decarboxylation in the presence of catalyst
4ꢀ6
without the requirement of a dehydrogenation catalyst. We
recently demonstrated that the catalytic production of benzoic
acid from CMA and ethylene proceeds via a sequence of reacꢀ
tions including DielsꢀAlder, decarboxylation, and palladium
catalyzed dehydrogenation. The rate for this reaction is limited
by the decarboxylation of the bicyclic lactone intermediate to
the respective DIH resulting in an activation energy of 142
To overcome the high activation barrier required to decarboxꢀ
ylate (1) to form the DIH (2), a catalyst that can selectively
activate the decarboxylation step is required. Given that
Brønsted acids and Lewis acids facilitate DielsꢀAlder reacꢀ
13,18,20,21,27,29ꢀ31
tions
, we hypothesized that these acid catalysts
could also enable retroꢀDielsꢀAlder extrusion of CO from
2
28
bicyclic lactones to generate DIH.
kJ/mol. Given the high activation barrier of decarboxylation
32
Starting with γꢀAl O as a Lewis acid catalyst, we found that
of the bicyclic lactone molecules, an effective catalyst that
activates the decarboxylation step is of great interest to allow
2
3
(1) can be effectively decarboxylated to (2) at 120 °C (Table 1,
Entry 2). The 47 mol% conversion of (1) resulted in a yield of
for CO extrusion under milder reaction conditions. This could
2
4
3 mol% of (2) and small amounts of (3) (Scheme 3). This is a
preserve the diene functionality and prevent in situ aromatizaꢀ
tion and formation of other decomposition products.
significant improvement over the control experiment which
gave only 9 mol% conversion of (1) at identical reaction conꢀ
ditions but in the absence of a catalyst (Table 1, Entry 1). It is
evident from these experiments that γꢀAl O can be effectively
Herein, we report a highly selective technology platform to
synthesize DIHs through the use of Lewis and Brønsted acid
catalysts that alter the reaction path of the bicyclic intermediꢀ
ate that forms. This approach can be applied to other bicyclic
lactone systems, opening the door to the production of many
different specialty chemicals without the requirement of nuꢀ
merous dienophiles. We demonstrate this approach using a
bicyclic lactone (1) derived from CMA in conjunction with
inexpensive and easily separable ethylene (Scheme 2). Deꢀ
tailed kinetic experiments, density functional theory calculaꢀ
tions, and thin layer solidꢀstate high resolution magic angle
spinning (HRꢀMAS) NMR spectroscopy which were carried
out show that bicyclic intermediate (1) derived from the Dielsꢀ
Alder addition of 2ꢀpyrone and ethylene undergoes a direct
decarboxylation on the surface of γꢀAl O to form cyclohexaꢀ
2
3
used as a highly active decarboxylation catalyst for bicyclic
lactones. Additional experiments were performed with Lewis
acid metals supported on γꢀAl O including 5 wt.ꢀ% Zn/γꢀ
2
3
Al O , Cu/γꢀAl O , Fe/γꢀAl O and Ce/γꢀAl O . The Cu/γꢀ
2
3
2
3
2
3
2
3
Al O showed a slight improvement in the decarboxylation
2
3
activity with a 58 mol% yield at 59 mol% conversion (Table 1,
Entry 2). Zn, Fe, and Ce on γꢀAl O showed either similar or
2
3
lower decarboxylation activity (Table 1, Entry 3ꢀ5). Cu, Zn,
and Ce were chosen due to their known Lewis acid
29ꢀ31
activities.
A more comprehensive correlation between
Lewis acidity of the synthesized catalysts and the decarboxylaꢀ
tion activity will be subject of future experimental investigaꢀ
tion.
2
3
1
,5ꢀdieneꢀ1ꢀcarboxylic acid (2) and show that the coordinaꢀ
Through this approach, we showed that γꢀAl O is an excellent
2
3
tively saturated Alꢀspecies selectively drives the decarboxylaꢀ
tion reaction to completion.
decarboxylation catalyst for bicyclic lactones to access novel
diene species (e.g. (2)) in high selectivity at significantly reꢀ
duced temperatures. Given that some bicyclic lactones and
diene species are susceptible to degradation and in situ aromaꢀ
tization at elevated temperatures, γꢀAl O can be generally
1
Through in situ HꢀNMR spectroscopy experiments, we furꢀ
ther demonstrate that the presence of Brønsted acids result in
the ringꢀopening of the bicyclic intermediate followed by an
2
3
applied to overcome this challenge.
Reaction kinetics for γ-Al O catalyzed decarboxylation
2
3
To gain a more detailed understanding of the role of the γꢀ
Al O in catalyzing decarboxylation, an inꢀdepth kinetic invesꢀ
2
3
tigation of the conversion of (1) to (2) over γꢀAl O was perꢀ
2
3
formed using temperatures between 130ꢀ160 °C in 1,4ꢀ
dioxane. The WeiszꢀPrater number was calculated to be < 1
for the reaction conditions used which verified the absence
Scheme 2. γꢀAl O catalyzed decarboxylation.
2
3
ACS Paragon Plus Environment

Page 3 of 15
1
ACS Catalysis
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
Table 1. Bicyclic lactone (1) conversion in the presence of Lewis acids
Entry
Catalyst
(1) Conv. [mol%]
(2) Yield [mol%]
(3) Yield [mol%]
Byꢀproducts [mol%]
Control
ꢀ
9
7
<2
< 4
< 2
< 3
< 3
< 3
ꢀ
ꢀ
1
2
3
4
5
γꢀAl2O3
47
59
49
47
47
43
58
36
45
34
5 wt% CuO/γꢀAl2O3
5 wt% ZnO/γꢀAl2O3
5 wt% Ce O /γꢀAl2O3
<1
<11
ꢀ
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
3
5 wt% Fe O /γꢀAl2O3
<11
2
3
Conditions: Reaction time 2 h, temperature 120 °C, stirring rate 500 rpm, initial concentration 71.38 μmol/mL, reaction volume 4 mL, 1,4ꢀ
dioxane, 25 mg catalyst.
of mass transfer limitations (see Supporting Information).
the 110 surface are covered by hydroxyls and water molecules.
We find that the coverage of H O* is 40% and that of OH* is
Assuming a firstꢀorder reaction with respect to (1), we obꢀ
tained linear trends of ln([1] /[1] ) with respect to time (Figure
S1) from which the rate constants (Table S1) were calculated
to generate the Arrhenius plot shown in Figure S1. The resultꢀ
ing activation barrier was found to be 108 kJ/mol, which is
significantly lower than nonꢀcatalyzed decarboxylation (142
2
6
0%. On the other hand, on the 100 surface, the OH coverage
0
t
2
of 13 OH/nm results in an OH* coverage of 25%, H O* covꢀ
erage of 50% and empty Al site (*) coverage of 25%. Given
that the γꢀAl O catalyst was not heat treated (dehydroxylated)
prior to the decarboxylation experiments and that the solvent
used in our experiments contained residual water (0.028
2
2
3
28
kJ/mol ) and provides evidence that γꢀAl O strongly impacts
the decarboxylation step.
2
3
mol%), the Al O surface was likely highly hydroxylated.
2
3
Therefore, in order to determine appropriate hydroxyl surface
coverage and to assess its impact on the adsorption behavior of
species (1), we first calculated the Gibbs free energy of adꢀ
sorption of water and (1) on the various sites on surfaces with
different levels of OH coverage. An abꢀinitio thermodynamic
analysis was then performed to predict the OH coverage of the
most abundant surface under these conditions. Further, deꢀ
tailed energetics were calculated to deduce activation barriers
of the surface catalyzed decarboxylation which were then
compared to results obtained from microkinetic modeling.
Computational analysis of the γ-Al O catalyzed decar-
boxylation
2
3
DFT calculations and microkinetic analyses were carried out
to gain further insights into the nature of the active sites on γꢀ
Al O and the surfaceꢀcatalyzed decarboxylation mechanism.
2
3
As discussed below, HRꢀMAS NMR spectroscopy provides
experimental evidence of a strongly adsorbed bicyclic lactone
species (1) on the γꢀAl O surface as well as high surface hyꢀ
droxylation. This information was used to guide DFT compuꢀ
tations.
2
3
Adsorption of water and (1) on different γ-Al O sur-
faces
2
3
Different surfaces such as 110, 100 and 111 faces can exist on
γꢀAl O particles under reaction conditions. The 111 surface
2
3
The standard gas phase Gibbs free energy of adsorption for the
bicyclic lactone (1) and water were carried out at 140 C for
the 110 and 100 Al O surfaces at different surface coverages
of Al(*), H O* and OH*. The results are reported in Tables S2
typically exposes only oxygen atoms and is extremely difficult
to dehydrate under realistic conditions, while the other surfacꢀ
es tend to exist with an appropriate OH coverage and exposed
Al sites that can be catalytically active. Therefore, in the
present work, we explored different Al sites on the 110 and
o
2
3
3
3
2
and S4 and the relevant structures are given in Figure S2. The
adsorption of (1) was found to be more favorable than water
on the 110 surfaces with high OH coverages (Table S2, entries
2 and 3). This is likely the result of the relatively weak molecꢀ
ular adsorption of water at high OH coverages and the strong
adsorption of (1) on Lewis acidic surface Alꢀsites promoted by
hydrogen bonding interactions between (1) and surface OH
groups. The high coverage of the strongly bound bicyclic lacꢀ
tone species (1) on the hydroxylated Al O surface was conꢀ
firmed via HRꢀMAS NMR (vide infra). At lower OH coveragꢀ
es (Table S2, Entry 4), water was found to adsorb dissociativeꢀ
ly, making water adsorption very strong and more favorable
than the adsorption of (1). On the 100 surface, the limited hyꢀ
drogen bonding interactions between surface OH groups and
(1) result in slightly less favorable adsorption of (1) compared
to water at different surface coverages examined in this work
(Table S4). Therefore, on the 110 surface, we find that the
most energetically favorable configuration has an OH* coverꢀ
1
00 surfaces with different levels of hydroxylation.
The fully dehydrated 110 surface exposes AlIII and AlIV sites
while the dehydrated 100 surface exposes different types of
34
Al sites. However, all of these sites are typically covered by
V
OH groups, leading to stabilization of these terminations. OH
groups preferentially occupy the sites with high Lewis acidity
(AlIII > AlIV > Al ). On the 110 surface, the AlIII sites are conꢀ
V
2
3
verted to tetragonal AlIV sites and are potentially blocked by
OH groups. Alternatively, the AlIV sites are converted to AlV
sites that are still catalytically active. However, Al sites can
further be hydroxylated to AlVI sites, rendering them fully
saturated and possibly catalytically inactive. This is predicted
to occur at the Al sites on the 100 surface, generating saturatꢀ
V
V
ed AlVI sites. Results by Digne et al. suggest that under reactꢀ
ing temperatures (120 C – 150 C), the OH coverage can be
o
o
2
2
as high as 15 OH/nm and 13 OH/nm on the 110 and 100 surꢀ
faces respectively, which can expose both catalytically active
Al sites and catalytically inactive tetrahedral Al and octaheꢀ
age of 60%, H O* coverage of 20% and (1*) coverage of 20%
2
V
IV
34
2
dral AlVI sites. At OH coverage of 15 OH/nm , all Al sites on
ACS Paragon Plus Environment

ACS Catalysis
Table S3, Entry 3), while on the 100 surface, the most enerꢀ getically favorable surface has an OH* coverage of 25%,
Page 4 of 15
(
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 1. (A) The individual Gibbs free energy changes for the elementary steps in the mechanism proposed in Scheme 3 (filled
rectangles represent stationary states and unfilled rectangles represent transition states). (B) Decarboxylation reactant (1*), transiꢀ
tion (≠) and product ((2)ꢀCO *) states 110 surface. (C) Decarboxylation reactant (1*), transition (≠) and product states ((2)ꢀCO *)
2
2
on *) on the 100 surface. (D) Concentration of (1) as a function of time: experiments vs the microkinetic model. (E) ln(k) vs time,
where k is the intrinsic rate constant, k , associated with the scission of the CꢀC bond and elimination of CO
2
2
H O* coverage of 50% and Al(*) coverage of 25% (Table S5,
vant structures on both surfaces. All of the reported energies
are reference to an empty active site (*) and (1) in the gas
phase. After adsorption, (1*) undergoes decarboxylation,
forming the product diene (2), which is weakly physisorbed to
the bound CO ((2)—CO *). The surface bound CO * then
desorbs, allowing (2) to adsorb on the surface. The bound surꢀ
face species (2*) then desorbs regenerating the active site and
completing the catalytic cycle.
2
Entry 1). Now, assuming adsorption and desorption of (1) and
water is quasiꢀequilibrated and that different configurations in
Tables S3 and S4 are also in equilibrium with each other, we
can calculate the most favorable surface configuration under
reaction conditions wherein we have 0.028% water and the
balance of (1) (Table S3ꢀS5 with calculations in Supporting
Information). We predict that the most abundant 110 configuꢀ
2
2
2
ration has an 60%, H O* coverage of 20% and (1*) coverage
of 20% (relative abundance of this configuration = 1). The
most abundant 100 configuration was found to have an OH*
2
The free energy diagram reported in Figure 1A indicate that it
is much more favorable for the products that form as result of
decarboxylation to desorb from surface than to react back and
form (1), as the barrier for the microscopic reverse of decarꢀ
boxylation is significantly higher than the energy for desorpꢀ
tion. Therefore, the products preferentially desorb, making the
decarboxylation of (1) irreversible and very likely the kinetꢀ
ically relevant step.
coverage of 25%, H O* coverage of 50% and Al(*) coverage
2
of 25%. Such high abundance (94%) of a slightly energetically
unfavorable configuration was likely due to the very high conꢀ
centration of (1) compared to water. The active sites on both
of these surfaces are pentaꢀcoordinate as discussed previously.
Decarboxylation on γ-Al O : mechanism and microki-
netic modeling
2
3
Assuming that the decarboxylation of (1*) is rate determining
step and all others are quasiꢀequilibrated, we can derive Equaꢀ
tion 1 (see Supporting Information):
The decarboxylation of (1) on the 110 and 100 surfaces obꢀ
tained from thermodynamic analysis was examined by carryꢀ
ing out DFT calculations. Figure 1 shows the Gibbs free enerꢀ
gy change for the proposed mechanism (Scheme 3) with releꢀ
K k ꢀꢁ1ꢂꢃ
1
2
r=
ꢀ
(2)ꢃꢀCO2ꢃ ꢀ(2)ꢃ
K K
(1)
1
+K ꢀꢁ1ꢂꢃ+
+
1
4
3
K4
ACS Paragon Plus Environment

Page 5 of 15
ACS Catalysis
film (2 nm) of a solution of (1) in 1,4ꢀdioxaneꢀd was perꢀ
where k , K , K and K refer to the intrinsic rate constant for
2
1
3
4
8
37
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
decarboxylation, the adsorption equilibrium constant for (1*),
the equilibrium between the weakly physisorbed product diene
(2*) and CO and the equilibrium desorption constant for (2*)
formed. This new and unique NMR technique allows the
study of the interactions of probe molecules on solid catalyst
surfaces (i.e. (1) on alumina) at the liquidꢀsurface interface
which provides additional information to guide DFT computaꢀ
tions. As described below, the NMR experiments provide diꢀ
rect evidence that the surface is highly hydroxylated and that
2
Based on the free energy diagram (Figure 1), high surface
coverages of (1*) and (2*) are expected due to the high Gibbs
free energies of adsorption. Therefore, Equation 1 can be simꢀ
plified as follows:
(
1) is strongly adsorbed on the surface of γꢀAl O .
2
3
Aꢀꢁ1ꢂꢃ
Samples were prepared by first drying γꢀAl O at 120 °C to
2 3
r=
(2)
remove excess water. The dry γꢀAl O was then impregnated
2
3
1
+Bꢀꢁ1ꢂꢃ+C[(2)]
ꢀ
1
in a glovebox with liquid loadings of 1 μL of solution per mg
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
where, A = K k , B = K and C = K . This rate expression
1
2
1
4
of γꢀAl O and the concentrations of (1) were 200 or 132
2
3
was used to model the concentration profiles at different temꢀ
peratures and resulted in a nearly perfect fit as shown in Figꢀ
ure 1D. The values of parameters A, B and C are reported in
the Table S6
mg/mL in dioxaneꢀd . A control sample impregnated with pure
8
dioxaneꢀd was also prepared. All samples were packed into
8
2
.5 mm zirconia rotors under the inert atmosphere. The results
of the HRꢀMAS NMR spectroscopy experiments on the variꢀ
ous impregnated alumina samples are summarized in Figure 2.
The initial activation barrier was calculated by assuming a
high concentration of (1). As such the observed rate constant,
k K [(1)]/K [A(1))], is equivalent to the intrinsic rate constant
1
Rotorꢀsynchronized H spin echo solidꢀstate NMR spectra of
2
1
1
the different impregnated alumina samples are shown in Figꢀ
for decarboxylation, k . An Arrhenius plot (Figure 1E) yields
2
1
ure 2AꢀC. The H solidꢀstate NMR spectrum of the alumina
an initial activation barrier of 87 kJ/mol, which is in good
agreement with the activation barrier determined from experꢀ
imental data (108 kJ/mol, see Table S1). On the free energy
diagram, this barrier corresponds to the intrinsic barrier for the
decarboxylation of (1*).
impregnated with dioxaneꢀd shows a broad underlying signal
8
that covers a shift range of ca. 0 ppm to 10 ppm and three
1
sharp H NMR signals (Figure 2C). The sharp signals are atꢀ
tributed to the residual protons of dioxaneꢀd (δ = 3.6 ppm)
8
and to mobile or dissolved water molecules and/or isolatꢀ
ed/mobile alumina hydroxyl groups (δ = 1.3 ppm and 0.9
27
1
ppm). Proton detected 2D Alꢀ H DꢀRINEPT HETCOR specꢀ
tra confirm that the broad NMR signal arises from the hydroxꢀ
yl groups and/or water molecules that are immobilized on the
surface of the alumina (see below).
1
The H NMR spectrum of alumina impregnated with the 200
mg/mL solution of (1) in dioxaneꢀd shows several additional
8
1
1
sharp H NMR signals at the expected H chemical shifts for
1) (Figure 2A). These narrow signals are attributed to (1)
(
which is either dissolved in the dioxaneꢀd or is very weakly
8
bound to the alumina surface and remains highly mobile. The
Scheme 3. Proposed elementary steps and mechanism for
the decarboxylation of (1).
1
narrow H NMR signals arising from (1) are absent from the
1
H NMR spectrum of alumina impregnated with the solution
1
However, since only temperature was varied, this barrier needs
to be compared with the enthalpy of activation. The enthalpic
barriers for (1*) decarboxylation were calculated to be 96
kJ/mol and 85 kJ/mol on the 110 and 100 surfaces respectiveꢀ
ly, which are in good agreement with the initial barrier resultꢀ
ing from microkinetic modeling (87 kJ/mol). Therefore, deꢀ
carboxylation of (1) can be catalyzed via the retro DielsꢀAlder
of (1) at 132 mg/mL in dioxaneꢀd
8
. Comparison of the H
NMR spectra of the alumina impregnated with the 132 mg/mL
1
solution to the H NMR spectrum of alumina impregnated
with dioxaneꢀd alone shows that there are additional signals at
8
higher chemical shifts of ca. 7.4 ppm and 5.6 ppm. The addiꢀ
tional signals at higher chemical shifts and the absence of
1
sharp H NMR signals suggests that in the sample impregnated
mechanism on Al sites on both 110 and 100 surfaces with the
with the 132 mg/mL solution, all of the molecules of (1) are
immobilized on the alumina surface.
V
reaction being limited by the decarboxylation of surface bound
(1*).
To confirm that (1) was immobilized on alumina, proton deꢀ
1
13
The experimental and theoretical results presented here on the
thermal and Lewis acid catalyzed decarboxylation of 2 are
very similar to those reported for the ring opening of triacetic
tected H{ C} crossꢀpolarization CPꢀHETCOR spectra were
obtained. In this case, proton detection was employed to boost
13
sensitivity and obtain the C solidꢀstate NMR spectra in a
3
8,39
3
5
36
reasonable experimental time (less than 16 hours).
A 2D
acid lactone (TAL) and γꢀvalerolactone (GVL) that readily
undergo acid catalyzed decarboxylation as a result of the unꢀ
saturated bond at the C3ꢀC4 position similar to the unsaturated
C3=C4 bond in 2 which facilitates a direct retro DielsꢀAlder
1
13
H{ C} CPꢀHETCOR spectrum of alumina impregnated with
1
3
a 200 mg/mL solution is shown in Figure 2F. The C dimenꢀ
sion shows NMR signals at all of the chemical shifts expected
1
for (1) and the H dimension shows broad signals between 8
ppm and 3 ppm, consistent with the expected H chemical
elimination of the CO dienophile.
2
1
Characterization of (1) immobilized on γ-Al O by HR-
2
3
shifts of (1). A CP experiment acts as a mobility filter since an
MAS NMR spectroscopy
NMR signal is only observable if the molecules are rigid. Adꢀ
1
To probe for interactions between the bicyclic lactone with the
ditionally, broad H NMR signals observed in the CPꢀ
1
13
27
γꢀAl O surface in the condensed phase, H, C and Al HRꢀ
HETCOR spectrum imply that the molecules of (1) are immoꢀ
2
3
1
13
MAS NMR spectroscopy of γꢀAl O impregnated with a thin
bilized. Therefore, the 2D H{ C} CPꢀHETCOR spectrum
2
3
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 15
1
27
1
confirms that a large fraction of (1) is bound to the alumina
The H NMR spectra obtained from 2D Al→ H DꢀRINEPT
experiments acquired with long and short dipolar recoupling
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
surface, which is consistent with the DFT results. Additional
1
13
27
2D H{ C} CPꢀHETCOR spectra are shown in Figure S3.
times are shown in Figure 2D and 2E, respectively. The Al
filtered H NMR spectrum obtained with a short recoupling
time selectively probes the H spins that are in close proximity
to aluminum. This spectrum shows a very broad signal, which
1
2
7
1
Finally, proton detected 2D Al→ H DꢀRINEPT HETCOR
1
40,41
27
experiments
were used to obtain surface selective Al
solidꢀ state NMR spectra (Figure S4).
1
must represent the H nuclei of the surface hydroxyl and adꢀ
sorbed water molecules, consistent with previous surface seꢀ
27
42,43
lective Al NMR experiments on hydrated alumina.
This
information of a highly hydroxylated γꢀAl O surface provide
2
3
important insights to help guide the DFT calculations and
27
1
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
benchmark their results. The Al filtered H NMR spectrum
obtained with a longer recoupling time shows similar intensiꢀ
13
1
ties and chemical shifts to those observed in the C filtered H
NMR spectrum. Therefore, the spectrum obtained with a longꢀ
1
er dipolar recoupling time suggests that the H nuclei of (1) are
27
proximate to Al spins, consistent with adsorption of (1) to
the alumina surface. At longer recoupling times, the surface
bound species (1) strongly correlates to Al and AlVI sites of
IV
1
27
the alumina surface, while Hꢀ Al correlation to Al sites is
V
suppressed (see full 2D spectra in Figure S4). The key
knowledge of a strong surface bound species (1) existing is
important to further support the DFT results. Additional solidꢀ
state NMR experiments are summarized in the Supporting
Information.
In summary, HRꢀMAS and solidꢀstate NMR spectroscopy is
an excellent technique to investigate the γꢀAl O catalyst surꢀ
2
3
face in the presence of probe molecules (e.g. bicyclic lactone
(1)) and small amounts of water in the condensed phase (e.g.
1,4ꢀdioxaneꢀd ). This unique technique provides key inforꢀ
8
mation about the degree of surface hydroxylation, the moleꢀ
culeꢀsurface interaction and information about the catalytically
active Alꢀspecies on the alumina surface. The results provide
additional means to guide and compare computational studies
with high level experimental detail of reactant/product moleꢀ
cules behavior at the solventꢀcatalyst interphase.
The impact of Brønsted acids on conversion
The impact of Brønsted acidity on the bicyclic lactone (1) was
examined through reactions using a variety of Brønsted acidꢀ
containing catalytic materials. The use of Davisil silica gel
44
(
150 Å), a weakly acidic material , at 120 °C for 2 h, gave 16
mol% conversion of (1) to (2) (Table 2, Entry 1) which was
roughly a twofold increase in conversion compared to the conꢀ
trol experiment without catalyst (Table 1, Control). The use of
the more acidic ZSMꢀ5 zeolite with a Si:Al ratio of 50:1 gave
similar results with only minimal decarboxylation activity
(Table 2, Entry 2). In fact, increasing the acid strength of
ZSMꢀ5 (Si:Al, 23:1) showed no improvement in the decarꢀ
boxylation yield (Table 2, Entry 3). Interestingly, the mesopoꢀ
rous Yꢀzeolite (Si:Al, 30:1), demonstrated enhanced reactivity
and consumed all the bicyclic lactone (1) yielding 53 mol%
DIH (2), 35 mol% isophthalic acid intermediate (5) and
13mol% isophthalic acid (6) (Table 2, Entry 4). Given this
result, it appeared that the larger pores of the Yꢀzeolite strucꢀ
ture provide enhanced accessibility of the bulky lactone (1).
46
1
Figure 2. MAS H solidꢀstate NMR spectra of γꢀAl O imꢀ
2
3
pregnated with dioxaneꢀd solutions with different concentraꢀ
tions of (1). The sample is indicated next to each spectrum.
AꢀC) H spin echo solidꢀstate NMR spectra. (D) and (E) Al
filtered H NMR spectra obtained from the positive projection
of a 2D Al→ H DꢀRINEPT HETCOR spectra obtained with
short and long dipolar recoupling times, respectively. (F)
filtered H NMR spectrum obtained from the positive projecꢀ
tion of the 2D H{ C} HETCOR spectrum shown below. The
C chemical shifts expected for (1) are indicated with red
dashed lines. The blue dashed lines illustrate the similarity of
the solution and solidꢀstate H chemical shifts associated with
1). All experiments were performed with a 25 kHz MAS freꢀ
8
1
27
(
1
27
1
13
C
1
45
1
13
Since Yꢀzeolite contains both Lewis and Brønsted acid
1
3
sites, it was unclear as to which site was responsible for ringꢀ
opening or if a synergistic effect was operative.
1
To decouple the roles of Lewis and Brønsted acid sites, a seꢀ
ries of experiments were performed using the heterogeneous
(
47
Brønsted acidꢀonly catalyst, Amberlyst 45. The experimental
results shown in Table 2, Entries 5 and 6 demonstrate that
quency.
ACS Paragon Plus Environment

Page 7 of 15
ACS Catalysis
Table 2. Bicyclic lactone (1) conversion in the presence of Brønsted acids
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) Conv.
(2) Yield
[mol%]
(3) Yield
[mol%]
(4) Yield
[mol%]
(5) Yield
[mol%]
(6) Yield
[mol%]
Byꢀproducts
[mol%]
Entry
Catalyst
[
mol%]
1
2
3
4
5
6
Davisil (150 Å)
16
15
8
<2
<3
<2
<2
<2
<2
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
<2
<1
ꢀ
ZSMꢀ5 (CBV5524G)
ZSMꢀ5 (CBV2314)
YꢀZeolite (CBV720)
Amberlyst 45
12
7
5
ꢀ
ꢀ
ꢀ
100
51
53
13
14
ꢀ
35
13
39
13
ꢀ
12
32
<11
<2
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
a
Amberlyst 45
90
<4
Conditions: Reaction time 2 h, reaction temperature 120 °C, stirring rate 500 rpm, starting concentration 71.38 μmol/mL, reaction volume
a
4
mL, solvent 1,4ꢀdioxane, 25 mg catalyst. 50 mg catalyst+40 uL D O.
2
Figure 4. Selectivity trend of intermediates and products from
acid catalyzed ringꢀopening/dehydration of bicyclic lactone
(1).
other bicyclic lactone systems. In this set of experiments, 5 ꢁL
(0.089 ꢁmol) of D SO was added to 300 ꢁL of a 71.38
ꢁmol/mL solution of (1) in fully deuterated dioxaneꢀd , transꢀ
ferred to high pressure NMR tubes and heated to the desired
reaction temperature (80ꢀ110 °C) inside the NMR spectromeꢀ
ter (Figure S7).
Figure 3. (A) Reaction scheme of the proposed acid catalyzed
ringꢀopening/dehydration reaction sequence. (B) Concentraꢀ
tion profile of the conversion of bicyclic lactone (1) to (4)
followed by conversion of (4) to (5). (Sꢀ1) and (Sꢀ2) are
formed throughout the reaction and are of unknown origin.
2
4
8
The experimental results revealed that the rate of consumption
of (1) did not match the rate of formation of (5) at all temperaꢀ
tures tested. The reason for this observation was the formation
of the reactive intermediate (4) and species (Sꢀ1) and (Sꢀ2) on
the pathway to (5). Species (4) is likely formed through acidꢀ
catalyzed hydrolysis of the lactone bridge of (1) (Figure 3A).
The concentration profile in Figure 3B further suggests that
the initial formation of (4) was then followed by acidꢀ
catalyzed dehydration of (4) to (5) as evident from the increasꢀ
ing concentration of (5) accompanied with a decrease in (4).
Clearly, the mechanism involved in the formation of (5) from
Brønsted acid sites promote ringꢀopening of the lactone bridge
forming the intermediate 6ꢀhydroxycyclohexꢀ1ꢀeneꢀ1,3ꢀ dicarꢀ
boxylic acid (4) followed by acid catalyzed dehydration to an
isophthalic acid intermediate (5) (Figure 3A). Doubling the
amount of catalyst and adding small amounts of water to the
polar aprotic solvent 1,4ꢀdioxane dramatically enhanced the
acid catalyzed ringꢀopening with increased conversion and
yield towards intermediate (4) and product (5) (Table 2, Entry
3
5
6
). This is consistent with results reported by Chia et al.
which show that the addition of Brønsted acids and water faꢀ
cilitate the ring opening, hydrolysis and dehydration of triꢀ
acetic acid lactone.
(1) is more complicated given the formation of unknowns (Sꢀ
1
) and (Sꢀ2).
1
A series of in situ H solution NMR experiments were perꢀ
With increased reaction temperature (4), (Sꢀ1) and (Sꢀ2) vanꢀ
ished while the concentration of (5) increased, so it appears
that (Sꢀ1) and (Sꢀ2) are intermediates on the pathway from (1)
to (5). This selectivity relationship is apparent in Figure 4
since the formation of (5) increases linearly as a function of
reaction temperature, whereas a linear downward trend is seen
for (4), (Sꢀ1), and (Sꢀ2). After reaction for 4 h and 35 min at
formed to gain a better understanding as to why Brønsted acid
sites led to the formation of (4) and (5). In the experiments, the
bicyclic lactone (1) was exposed to the fully deuterated sulfuꢀ
ric acidꢀd2 and heat using dioxaneꢀd8 to mediate the reaction.
These experiments serve as an ideal model to understand the
reaction mechanism of the Brønsted acid catalyzed ringꢀ
opening reaction that can be potentially further applied to
1
10 °C there was a significant amount of (5) (87 mol%), with
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 15
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 5. (A) DFT calculated enthalpy diagram for the bicyclic lactone decarboxylation reaction using polar protic and aprotic solꢀ
vents. (B) Experimental Arrhenius plot for the water mediated catalystꢀfree decarboxylation of the bicyclic lactone sodium salt (1).
1
(
C) In situ HꢀNMR concentration profile of the consumption of sodium salt of (1) and the formation of (2) as a function of time.
D) Bicyclic lactone diversification in the absence of catalyst using the polar protic solvents water for decarboxylation and methaꢀ
(
nol for ringꢀopening. (E) DFT calculated enthalpy diagram for the methanol mediated ringꢀopening of (1). (F) Experimental Arrheꢀ
1
nius plot for the methanol mediated catalystꢀfree ringꢀopening of (1). (G) In situ HꢀNMR concentration profile of the consumption
of (1) and the formation of (7) as a function of time.
only 2 mol% of intermediate (4) and less than 5 mol% of (Sꢀ1)
and (Sꢀ2) remaining (Table S8). From NMR analysis of the
product solution, (Sꢀ1) and (Sꢀ2) are likely isomers of (4) givꢀ
en the similar proton shifts and NMR correlations. It is known
ce provides general guidance for the conversion of (1) and can
potentially be applied to similar bicyclic lactone systems. In
this particular case, we were able to access the isophthalic acid
intermediate (5) in high selectivity, which cannot be easily
obtained via conventional petroleum routes. Catalytic dehyꢀ
drogenation of this molecule provides isophthalic acid, a bulk
chemical currently manufactured in the petrochemical industry
and used in applications ranging from unsaturated polyester
that lactones are able to undergo different hydrolysis pathways
48
(
e.g., unimolecular and bimolecular) in acidic media, which
could explain the observation of species (Sꢀ1) and (Sꢀ2).
Structural identification of (Sꢀ1) and (Sꢀ2), however, is beyond
the scope of this work and will be a subject of further investiꢀ
gation.
49
and specialty resins to metal organic polyhedral crystals
(MOPs) which are potentially useful for drugꢀdelivery, cataꢀ
50
lysts, sensing and gas storage.
The Brønsted acid catalyzed ringꢀopening/dehydration sequenꢀ
ACS Paragon Plus Environment

Page 9 of 15
ACS Catalysis
Solvent impact on the conversion of (1)
Lactones such as CMA degrade in the presence of water due
to nucleophilic attack and consequent ringꢀopening, which
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
28,51
renders the diene ineffective for DielsꢀAlder chemistry.
This degradation was also observed by Chia et al. showing
that the molecular integrity of the 2ꢀpyrone triacetic acid lacꢀ
tone (TAL) was significantly compromised in the presence of
water but was stable in polar aprotic solvents such as tetrahyꢀ
35
drofuran (THF).
To investigate whether the bicyclic lactone (1) can be ringꢀ
opened to (5) via catalystꢀfree, waterꢀmediated hydrolysis of
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
the lactone bridge, experiments were performed in D O. For
2
the reactions, the sodium salt of the bicyclic intermediate (1)
was heated in D O, and the reaction progress monitored via in
2
1
situ HꢀNMR. The sodium salt of (1) was required to overꢀ
come solubility limitations in D O and to improve the resoluꢀ
2
1
tion of the HꢀNMR spectra for subsequent analysis.
1
Figure 6. Selected HꢀNMR spectra of the bicyclic lactone (1)
Surprisingly, when the consumption of (1) was monitored over
time, decarboxylation occurred rather than ringꢀopening. Adꢀ
ditional kinetic experiments were carried out at temperatures
in the range of 50 to 90 °C along with DFT calculations to
obtain further insights into the waterꢀmediated decarboxylaꢀ
tion mechanism.
ringꢀopening in methanolꢀd4.
Anticipating broad applicability of polar protic solvents to
enhance bicyclic lactone decarboxylation, we also performed
reactions of (1) in methanol. Methanol is an ideal model solꢀ
vent as it provides excellent solubility of (1) through which the
neutralization step of (1) to the sodium salt can be prevented.
Figure 5C shows an example of the concentration profiles for
the waterꢀmediated decarboxylation of (1) to (2). Plots of
ln([1] /[1] ) over time for all temperatures tested displayed
Experimental results showed that (1) in the presence of methꢀ
anol afforded ringꢀopening of the lactone bridge to give the
novel species (7) in high selectivity as confirmed by 2D NMR
experiments in combination with UPLCꢀQDa mass analysis
0
t
linear relationships, (Figure S8) indicating a first order reacꢀ
tion with respect to (1). Since (2) was obtained in high yield
and selectivity with no formation of CMA, the retro Dielsꢀ
Alder reaction of (1) to CMA was absent, meaning the obꢀ
served decarboxylation rate constants kobs (Table S10) are
solely for the conversion of (1) to (2). The activation energy of
the waterꢀmediated decarboxylation was calculated using the
Arrhenius plot shown in Figure 5B. The experimentally obꢀ
tained activation barrier was 107.2 kJ/mol. This value is sigꢀ
nificantly lower than for decarboxylation in the polar aprotic
1
(see Supporting Information). Several selected HꢀNMR specꢀ
tra of the inꢀsitu conversion of (1) to (7) are displayed in Figꢀ
ure 6. These fully resolved spectra make the identification and
quantification of the reactant and product unambiguous with
carbon balances >99 mol%.
This interesting and unexpected result was further investigated
through in situ NMR kinetic investigation and DFT calculaꢀ
tions to obtain kinetic parameters and mechanistic insight as to
why methanol caused (1) to enter this new pathway.
2
8
solvent dioxaneꢀd8 (142 kJ/mol), which explains the enꢀ
hanced decarboxylation activity at moderate temperatures.
NMR kinetics were carried out within the temperature range
of 90ꢀ110 °C. The reactions were conducted in neat methanolꢀ
DFT predicts an enthalpic decarboxylation barrier of the bicyꢀ
clic lactone (1) sodium salt of 119 kJ/mol in close agreement
with experimental results. Additionally, the Gibbs free energy
barrier for decarboxylation was calculated to be 116 kJ/mol,
which is favored over the water mediated ringꢀopening of the
lactone bridge which has a barrier of 121 kJ/mol. Moreover,
the decarboxylation product was found to be thermodynamiꢀ
cally favored over the ringꢀopened product because the system
d4. The consumption of (1) followed a pseudoꢀfirst order reacꢀ
tion. A typical concentration profile of (1) in methanolꢀd is
displayed in Figure 5G. Plots of ln([1] /[1] ) with respect to
4
0
t
time show linear relationships for all temperatures tested (Figꢀ
ure S10), validating our assumption of a pseudoꢀfirst order
reaction. These plots were then used to extract the observed
rate constants (Table S11) and calculate the apparent activaꢀ
tion barrier for the methanol induced ringꢀopening (Figure
5F).
gains entropy through the loss of CO in the former case. This
2
makes decarboxylation both thermodynamically and kineticalꢀ
ly favored over ringꢀopening in water, which explains the exꢀ
perimental results in water (see Figure S9 for the Gibbs free
energy profiles). The strong hydrogen bonding with water
molecules in the bulk makes water a weak nucleophile, limitꢀ
ing the addition of water to the lactone which prevents ring
opening. Although the acid form of (1) resulted in a slightly
higher calculated enthalpic activation barrier of 130 kJ/mol,
the decarboxylation of (1) to (2) remains favored in water as
compared to the polar aprotic solvent 1,4ꢀdioxane in the abꢀ
sence of catalyst (Figure 5A). This rate enhancement in water
is due to stabilization of the charge separated transition state
on the pathway from (1) to (2), which is preferentially stabiꢀ
lized by polar protic solvents such as water.
An activation barrier of 80.8 kJ/mol was calculated using the
Arrhenius plot in Figure 5F. The experimental data and strucꢀ
tures were used to compare with DFT results for the methꢀ
anolysis of the lactone bridge of (1). The DFTꢀcalculated enꢀ
thalpic activation barrier of 89 kJ/mol agrees rather well with
the experimentally measured activation barrier of 80.8 kJ/mol.
The barrier for the ringꢀopening and methanolysis of (1) is
approximately 40 kJ/mol lower than that for direct decarboxyꢀ
lation of (1) (129 kJ/mol) which explains the preference for
ringꢀopening and methanolysis over decarboxylation for reacꢀ
tions carried out in methanol (Figure 5A).
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 15
cies on the surface provides a driving force for its catalytic
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
decarboxylation. The immobilization of (1) was predicted with
DFT computations and experimentally confirmed with MAS
solidꢀstate NMR experiments.
Table 3. CMA conversion to benzoic acid
CMA conv.
mol%]
(1) Sel.
[mol%]
(2) Sel.
[mol%]
(3) Sel.
[mol%]
Entry
The inꢀdepth catalytic investigation presented here provides
[
guidance to tailor a bifunctional 1 wt% Pd/γꢀAl O catalyst
2
3
a
d
1
89.6
100
ꢀ
ꢀ
ꢀ
99.7
46.1
that enables access to the aromatic dropꢀin replacement benzoꢀ
ic acid (BA) with significantly improved selectivity starting
from the 2ꢀpyrone coumalic acid reaction with ethylene. Given
previous literature reports of aromatics synthesized from 2ꢀ
pyrones in an analogous fashion, we believe that this catalyst
can be broadly applied to these systems allowing the producꢀ
tion of aromatics at optimized reaction conditions.
b
c
d
2
53.9
Conditions: CMA 300 mg, 1,4ꢀdioxane 30 mL, temperature
a
b
1
40 °C, time 4 h, catalyst 200 mg 1wt% Pd/γꢀAl O , 20 mg
2 3
c
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
10wt% Pd/C. No other byꢀproducts detected via UPLCꢀQDa
(ESI). Quantified via UPLCꢀQDa (ESI).
d
We further investigated the impact of Brønsted acid catalysts
on the conversion of (1) and showed through kinetic studies
that a ringꢀopening/dehydrogenation sequence yielded a novel
isophthalic acid precursor (5) in high selectivity. Since
isophthalic acid is a bulk chemical in the chemical industry,
this route provides a renewable alternative.
The computational results further suggest that methanol is a
better nucleophile than water, likely because water forms a
strong hydrogen bonding network and results in decarboxylaꢀ
tion rather than hydrolysis of (1). Similar to water, methanol
stabilizes the transition state and promotes proton shuttling,
providing a low energy path from (1) to (7). This pathway
grants access to the selective formation of the new novel speꢀ
cies (7) (> 95 mol%) using low temperatures and a low boiling
solvent in the absence of a catalyst, which provides an attracꢀ
tive and environmentally benign diversification path of the
renewable coumalate platform.
Finally, the impact of polar protic solvents on the conversion
of (1) was explored, through which highly selective, catalyst
free pathways of (1) to novel species (2) and (5) were identiꢀ
fied. The decarboxylation of the sodium salt of species (1) was
mediated in the polar protic solvent, revealing a significantly
reduced decarboxylation barrier compared to that of 1,4ꢀ
dioxane. The nucleophilic solvent methanol induced ringꢀ
opening of the lactone bridge through methanolysis providing
an additional diversification path of (1) to a new novel species
Broader Implications
DIH (2) is a precursor of the bulk chemical benzoic acid
which is currently manufactured via conventional petroleum
routes based on toluene. We recently presented a new route to
benzoic acid with ~90 mol% selectivity at 100 mol% CMA
(7) in high selectivity.
The results presented herein outline a diversification platform
starting from renewable coumalic acid in conjunction with a
single dienophile (ethylene) that allows highly selective access
to a variety of novel molecules with unique functionality
which are of potential interest for new materials. As such, this
technological approach can be leveraged to other bicyclic lacꢀ
tones to form a plethora of new and interesting molecules that
cannot be easily accessed via conventional petroleumꢀbased
routes.
28
conversion. The 10 mol% loss in selectivity was attributed to
byꢀproduct formation from the high reaction temperature (180
°C) needed to overcome the rate limiting decarboxylation step.
The catalyst used to convert (2) to (3) was a commercial 10
wt% Pd/C catalyst. With the goal of further increasing the
selectivity and improving the overall process, we utilized a
bifunctional 1 wt% Pd/γꢀAl O catalyst to perform the reaction
2
3
of CMA and ethylene to benzoic acid (BA) at milder reaction
conditions. This lower temperature was possible because of
the increased decarboxylation activity of bicyclic lactones in
the presence of γꢀAl O .
Methods and Materials
Reagents and Materials
2
3
Pd(II)ꢀ, Cu(II)ꢀ, Ce(III)ꢀ, Zn(II)ꢀacetate (>99.9 %), coumalic
Experimental results are provided in Table 3, with Entry 1
showing that the benzoic acid selectivity was further increased
to > 99 mol% at reduced temperatures (140 °C) with a reaction
time of 4 h. The base Pd/C catalyst, on the contrary, only gave
a selectivity of 46.1 mol% towards (3) at 100 mol% converꢀ
sion (Table 3, Entry 2). The remaining 53.9 mol% were atꢀ
tributed to (1) as no other byꢀproducts were detected via
UPLCꢀQDa analysis. Given the high selectivity of BA, this
concept might be broadly applicable when aromatics from 2ꢀ
pryrones in conjunction with an array of dienophiles are enviꢀ
sioned.
acid (>97 %), 10 wt% Pd/C, sulfuric acid–d (96ꢀ98 wt%, 99.5
2
atom % D), D O (99.9 atom % D) and Davisil (Silica Gel, 150
2
Å, >99 %) were obtained from Sigma Aldrich. DMSOꢀd (99.9
6
%
), dioxaneꢀd (99 %), methanolꢀd (99.8 %) were obtained
8
4
from Cambridge Isotope Laboratories Inc. Dimethylformamid
99.9 %), 1,4ꢀdioxane (99.9 %) and sodium bicarbonate (>
9.7 %) were obtained from Fisher Scientific. Methyl coumalꢀ
(
9
ate (Acros Organics, 98 %). Ethylene (99.5 %, Matheson).
Zeolites ZSMꢀ5 (CBV5524G), YꢀZeolite (CBV 720), and
ZSMꢀ5 (CBV2314), were obtained from Zeolyst International
in very high purity (100 wt%), γꢀAl O (Strem Chemicals,
2
3
Conclusions
Inc., >99.8 %), Amberlyst 45 (Dow >98%). All chemicals
were used without further purification.
The combination of experimental kinetic analyses, DFT calcuꢀ
lations and MAS solidꢀstate NMR provided mechanistic inꢀ
sights into the decarboxylation of the 2ꢀpyrone derived bicyꢀ
Catalyst Preparation
The 5 wt% metal catalysts were prepared via incipient wetness
clic molecule (1). Lewis acid sites on the γꢀAl O act as cataꢀ
2
3
impregnation using γꢀAl O , and the respective metal acetate
2
3
lytically active sites that significantly lower the decarboxylaꢀ
tion activation barrier, providing access to DIH (2) in high
selectivity. The strong adsorption of the bicyclic lactone speꢀ
salts of Cu, Fe, Zn, and Ce, dried at 65 °C (24 h) and calcined
at 450 °C (6 h) in a steady flow (200 mL/min) of air.
ACS Paragon Plus Environment

Page 11 of 15
ACS Catalysis
%
). Through careful evaporation of the solvent using dry air,
1
2
the reaction product (1) was obtained and subsequently disꢀ
solved in the respective deuterated solvent. The solution was
then transferred into the highꢀpressure NMR tubes. The tubes
were placed into the Bruker Avance III 600 MHz spectrometer
and heated to the respective temperature to initiate the reacꢀ
tion. The progress of the reaction was subsequently monitored
over time.
Batch Reaction Experiments
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
The DielsꢀAlder product (1) was obtained from batch reactions
at 110 °C for 16 h using a 50 mL 4590 Parr reactor setup. Meꢀ
chanical agitation was maintained at 400 rpm using a magnetꢀ
driven stirrer. 300 mg of CMA was placed into the reactor
vessel and 30 mL of 1,4ꢀdioxane was added to mediate the
reaction, resulting in a starting concentration of 71.38
μmol/mL. After the reactor was sealed, the reactor was purged
five times to replace residual air with nitrogen. Next, the reacꢀ
tor was charged for 30 min with 500 psig ethylene to ensure
saturation of ethylene in the solvent. Under these conditions
the reactions were performed in large excess of ethylene. After
charging the vessel with ethylene, the reactor was heated to
the desired reaction temperature using a controlled heating
ramp of 10 K/min. After completion of the reaction, the reacꢀ
tor vessel was purged to replace unreacted ethylene with niꢀ
trogen. A sample was withdrawn, and the solvent was carefulꢀ
ly evaporated with dry air before 600 μL fully deuterated solꢀ
The waterꢀmediated decarboxylation of the sodium salt of the
bicyclic lactone (1) was performed in D O at temperatures in
the range of 50ꢀ90 °C. 50 mg of the reactant (1) was dissolved
in 1 mL of D O (263.16 μmol/mL). The sodium salt of (1) was
2
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
synthesized via neutralization using sodium bicarbonate.
The methanolꢀd ꢀmediated reactions were performed in a temꢀ
4
perature range of 90ꢀ110 °C using 71.38 μmol of (1) that was
dissolved in 300 μL (237.93 μmol/mL) of methanolꢀd4.
The ringꢀopening reactions in 1,4ꢀdioxane using D SO as a
2
4
catalyst were performed in a temperature range of 70ꢀ110 °C.
1.38 μmol of (1) was dissolved in 300 μL (237.93 μmol/mL)
dioxaneꢀd and 5 μL of D SO were added to catalyze the reꢀ
7
8
2
4
vent (DMSOꢀd ) and 2.5 μL internal standard (DMF) were
6
action.
1
added. Finally, the samples were analyzed via HꢀNMR.
To elucidate the reaction network, all compounds were veriꢀ
fied via NMR structural assignments performed using differꢀ
ent techniques including HꢀNMR, Hꢀ H COSY, Cꢀ H
HSQC and validated using UPLCꢀQDa (ESI) mass analysis.
Batch reactions for the catalyst screening experiments were
performed in 10 mL thickꢀwalled glass vial reactors (Alltech)
equipped with magnetic stir bars. Agitation and heat were
provided using a temperatureꢀcontrolled oilꢀbath that was regꢀ
ulated with an Isotemp Digital Plate Stirrer from Fisher Scienꢀ
tific. The glass vial reactor was loaded with 4 mL of a 71.38
μmol/mL solution of (1) in 1,4ꢀdioxane. 25 mg of catalyst was
added before the vial was sealed and placed into the hot oil
bath to initiate the reaction. After the reaction was completed,
the product was filtered through a 0.2 micron syringe filter to
remove the catalyst from the products. A sample was then
taken and the solvent was removed via careful evaporation.
1
1
1
13
1
Analytical methods
The batch reaction solution products were analyzed via NMR
using a Bruker spectrometer equipped with a 14.1 Tesla superꢀ
conducting magnet. The data were acquired and processed
using TOPSPIN (version 3.0) and Mestre Nova (version
1
0.0.1ꢀ14719), respectively. NMR samples were prepared
using fully deuterated solvents, to both reduce the solvent
background and to perform field calibration. Additionally, a
2
.5 uL DMF internal standard was added for quantification.
Fully deuterated DMSOꢀd (600 μL) and 2.5 uL of internal
6
1
All H spectra were acquired using a recycle delay of 3.0 sec.
and 30° H excitation pulse lengths. Hꢀ H 2D NMR spectra
were acquired using a COSY pulse sequence, and Cꢀ H 2D
NMR spectra were acquired using a HSQC pulse sequence.
standard DMF (Dimethylformamid) were added to the crude
1
1
1
1
material for quantification via HꢀNMR.
13 1
The γꢀAl O catalyzed decarboxylation reactions were carried
2
3
out in a 50 ml 4590 Parr reactor. The vessel was charged with
0 mL of a 71.36 μmol/mL solution containing (1). 200 mg of
the γꢀAl O catalyst (~ 53 μm particle size) was added before
Ultraꢀpressure liquid chromatography (UPLC) was also apꢀ
plied to analyze known species in the reaction products and to
obtain the amount of each compound. A Waters Acquity Hꢀ
Class System equipped with a Photodiode Array (PDA) and a
QDa mass detector was used to perform the analysis. UPLC
separation was carried out on a Waters BEH Phenyl column
3
2
3
the reactor was sealed and purged 5 times with nitrogen to
remove residual air while agitation was applied (400 rpm).
Next, the respective reaction temperature was obtained using a
controlled heat ramp of 10 K/min. After the reactor reached
the desired temperature a sample was withdrawn as a referꢀ
ence using a highꢀpressure sampling dip tube. Further samples
were withdrawn periodically to obtain a concentration profile
over time. After the reaction was completed, all samples were
filtered via a 0.2 micron syringe filter and the solvent was
carefully evaporated with dry air before 600 μL deuterated
(2.1x100 mm, 1.7 ꢁm particles).
The samples for HRꢀMAS/solidꢀstate NMR experiments were
prepared by first drying γꢀAl O overnight at 120 C in a conꢀ
vection oven to remove residual adsorbed water. Next, 30 mg
γꢀAl O was weighed into a 2 mL polypropylene screwcap vial
and then transferred into the glovebox. The vials were allowed
to sit with the cap removed in the glovebox for 48 h prior to
impregnation. Samples were impregnated with a liquid loading
of 1 μL/mg and allowed to equilibrate overnight with the cap
on. Following impregnation, the samples were then packed
into 2.5 mm rotors inside the glove box to prevent excess
moisture from being introduced.
o
2
3
2
3
solvent (DMSOꢀd ) and 2.5 μL internal standard DMF were
6
1
administered. The samples were then analyzed via HꢀNMR.
In situ NMR Experiments
The general procedure for the in situ kinetic measurements
was performed using highꢀpressure NMR tubes from Wilmadꢀ
Labglass. The solvents used for the decarboxylation and ringꢀ
opening experiments were D O, dioxaneꢀd , and methanolꢀd ,
All solidꢀstate NMR experiments were performed on a 9.4 T
2
8
4
(400 MHz) Bruker Avance III HD NMR spectrometer with a
respectively. As aforementioned, reactant (1) for this study
was synthesized via DielsꢀAlder reaction of CMA and ethꢀ
2
.5 mm HXY triple resonance probe. A 25 kHz MAS frequenꢀ
1
cy was used for all experiments. H solidꢀstate NMR spectra
o
ylene in 1,4ꢀdioxane at 110 C and 16 h, with a high yield (>98
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 15
were obtained with a rotor synchronized spin echo pulse seꢀ al component. Since, the transition state for decarboxylation
2
7
1
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
quence. Proton detected 2D Al→ H DꢀRINEPT HETCOR
spectra were obtained using the previously described Dꢀ
was in close resemblance to the reactant state, the entropy of
activation was assumed to be negligible.
4
0,41
RINEPT pulse sequence.
The symmetry based recoupling
Molecular density functional theory
2
sequence supercycled (S) R4 was used for dipolar reꢀ
1
1
Molecular density functional theory calculations were perꢀ
52
13
coupling. Proton detected 2D H{ C} CPꢀHETCOR spectra
were obtained with the previously described pulse
67
formed using the M062X hybrid functional as implemented
68
69
in Gaussian 09. 6ꢀ311G+(d,p) basis set on an ultrafine grid
and tight convergence criterion for force (RMS force within
38,39
1
13
sequence.
The contact time for the H→ C forwards CP
ꢀ
5
ꢀ5
transfer was 2000 ꢀs and different 2D spectra were obtained
with contact times of 200 ꢀs, 500 ꢀs and 2000 ꢀs for the
10 Hartree/Å, maximum force within 1.5 x 10 Hartree/Å,
ꢀ
5
and maximum displacement within 6 x 10 Å) was used for all
stationary and saddle point calculations. The SMD solvation
1
3
1
13
C→ H backwards CP transfer. CP was performed with a
C
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
69
spin lock pulse with an RF field of ca. 91 kHz and a H spin
lock pulse with a nominal RF field of ca. 68 kHz. The ampliꢀ
tude of the H spin lock pulses was linearly ramped from 85%
to 100% RF field to broaden the CP match conditions. In
both the C and Al NMR experiments SPINALꢀ64 H hetꢀ
model was used to implicitly model the effect of solvation.
In cases where the solvent directly participated in reaction,
explicit solvent molecules were used in conjunction with imꢀ
plicit solvation. Enthalpies and Gibbs free energies were calꢀ
culated at standard state and 298.15 K within Gaussian 09. A
correction factor of RT ln(24.46) was added to the free enerꢀ
gies of species having standard concentration of 1 M. While
for bulk solvents, (water and methanol), corrections correꢀ
sponding to 55.5 M for water and 24.9 M for methanol were
applied.
1
53
13
27
1
5
4
eronuclear decoupling was performed during the indirect
dimension evolution period with a 100 kHz RF field. Details
on acquisition parameters for the 2D spectra (number of scans,
indirect dimension points, etc.) are described in the Figure
captions.
Computational
ASSOCIATED CONTENT
Supporting information
Periodic density functional theory
All of the calculations carried out on model γꢀAl O surfaces
2
3
Experimental data of the kinetic investigation of the catalytic
decarboxylation and ringꢀopening of (1). Experimental data of the
waterꢀmediated decarboxylation and the methanolꢀmediated ringꢀ
opening reaction of (1). Experimental data of catalyst screening
for decarboxylation of (1). Calculation of mass transfer limitaꢀ
tions. 1D and 2D NMR analysis of (2), (4), (5) and (7), including
m/z values from UPLCꢀQDa (MSꢀESI). Computational DFT and
microkinetic analysis data. HRꢀMAS NMR data of the (1) imꢀ
pregnated γꢀ Al O . This material is free of charge via the Internet
were carried out using periodic planeꢀwave density functional
theory methods as implemented in the Vienna ab initio simulaꢀ
5
5ꢀ58
59
tion package (VASP).
calculate exchange and correlation energies. The D3 method
The PBE functional was used to
60
was used to estimate the dispersive interactions. A planeꢀwave
energy cutoff of 396 eV was used for all periodic calculations
reported herein. Optimizations were performed in two steps.
ꢀ
4
First, waveꢀfunctions were converged to within 10 eV, until
the maximum force upon each atom was less than 0.1 eV/Å. In
2
3
at http://pubs.acs.org.
ꢀ
6
step two, waveꢀfunctions were converged to within 10 eV,
with maximum allowable force upon each atom less than 0.05
eV/Å. Step two calculations were carried out spinꢀpolarized
and monopole/dipole moments were calculated to correct for
energies. Transition states were obtained using the nudged
AUTHOR INFORMATION
Corresponding Author
*
bshanks@iastate.edu
6
1,62
60
elastic band method
and the dimer method.
Calculations were performed on 1x1 unit cells of 110 and 2x1
Present Addresses
unit cells of 100 surfaces of γꢀ Al O . The bulk structure of γꢀ
2
3
6
3
† Robert L. Johnson, Department of Chemistry, King Fahd University
of Petroleum & Minerals, Dhahran 31261, Saudi Arabia.
Al O reported elsewhere was used. For hydroxylated surfacꢀ
2
3
33,64
es, structures found by Sautet and coꢀworkers were used.
For 110 surfaces, the periodic slab models consisted of 8 layꢀ
ers (Al:O = 3:2) separated by vacuum with bottom two layers
frozen to bulk position. While for the 100 surfaces, 4 layers
Notes
The authors declare no competing financial interest.
(
Al:O = 3:2) were modeled with the bottom two layers frozen
to bulk positions. Optimizations on the 1x1 unit cells of 110
ACKNOWLEDGMENTS
6
5
surfaces were performed on a 3x3x1 k point mesh while a k
point mesh of 2x2x1 was used for the 2x1 unit cells of 100
surfaces.
We gratefully acknowledge funding from the National Science
Foundation under Award EECꢀ0813570, the Iowa State Universiꢀ
ty Chemical Instrument Facility staff members, computational
support from the Minnesota Supercomputing Institute (MSI) at
the University of Minnesota and the Molecular Science Compuꢀ
ting Facility (MSCF) in the William R. Wiley Environmental
Molecular Sciences Laboratory, a national scientific user facility
sponsored by the U.S. Department of Energy, Office of Biological
and Environmental Research at the Pacific Northwest National
Laboratory. Furthermore, we would like to acknowledge coꢀ
workers in CBiRC for their support.
The entropies of adsorbed molecules were estimated by equaꢀ
tions 3 and 4, where S (T) is the entropy of adsorbed moleꢀ
cule and Sgas (T) is the ideal gas entropy of that molecule.
These values roughly correspond to the loss of 1/3 entropy of
o
ad
o
rd
66
the molecules in the gas phase.
o
o
o
Sad (T) = 0.70 × Sgas (T) − 3.3R, for Sgas (T) < 60 R …
(3)
(4)
o
o
o
Sad (T) = 0.99 × Sgas (T) – 20.7R, for Sgas (T) > 60 R …
As an approximation, the entropy of a weakly physisorbed
molecule was estimated by subtracting just the 1/3 transitionꢀ
REFERENCES
rd
ACS Paragon Plus Environment

Page 13 of 15
ACS Catalysis
(
1)
Creating Value from Biomass. Green Chem. 2017, 19, 3177ꢀ3185.
2) Kraus, G. A.; Riley, S.; Cordes, T. Aromatics from
Shanks, B. H.; Keeling, P. L. Bioprivileged Molecules:
(20) Markó, I. E.; Evans, G. R. Catalytic, Enantioselective,
Inverse ElectronꢀDemand DielsꢀAlder (IEDDA) Reactions of 3ꢀ
Carbomethoxyꢀ2ꢀPyrone (3ꢀCMP). Tetrahedron Lett. 1994, 35, 2771ꢀ
2774.
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
(
Pyrones: ParaꢀSubstituted Alkyl Benzoates from Alkenes, Coumalic
acid and Methyl Coumalate. Green Chem. 2011, 13, 2734ꢀ2736.
21
Markó, I. E.; Evans, G. R.; Seres, P.; Chellé, I.; Janousek,
(
3)
Kraus, G. A.; Pollock Iii, G. R.; Beck, C. L.; Palmer, K.;
Z. Catalytic, Enantioselective, Inverse ElectronꢀDemand DielsꢀAlder
Reactions of 2ꢀPyrone Derivatives. Pure Appl. Chem., 1996; 68, 113.
(22) Kranjc, K.; Kočevar, M. Diels–Alder Reaction of Highly
Substituted 2HꢀPyranꢀ2ꢀOnes with Alkynes: Reactivity and
Regioselectivity. New J. Chem. 2005, 29, 1027ꢀ1034.
(23) Abdullahi, M. H.; Thompson, L. M.; Bearpark, M. J.;
Vinader, V.; Afarinkia, K. The Role of Substituents in Retro Diels–
2
Alder Extrusion of CO from 2(H)ꢀPyrone Cycloadducts. Tetrahedron
2016, 72, 6021ꢀ6024.
(24) Imagawa, T.; Kawanisi, M.; Sisido, K. Double Diels–Alder
Reactions of Coumalic Acid with Butaꢀ1,3ꢀDienes: A SingleꢀStep
Synthesis of Tricyclo[3,2,1,0 ]Octꢀ3ꢀEne Derivatives. J. Chem. Soc.
D., 1971, 20, 1292ꢀ1293.
(25) Juranovič, A.; Kranjc, K.; Perdih, F.; Polanc, S.; Kočevar,
M. Comparison of the Reaction Pathways and Intermediate Products
of a MicrowaveꢀAssisted and HighꢀPressureꢀPromoted Cycloaddition
of VinylꢀMoietyꢀContaining Dienophiles on 2HꢀPyranꢀ2ꢀones.
Tetrahedron 2011, 67, 3490ꢀ3500.
(26) Reed, J. A.; Schilling Jr, C. L.; Tarvin, R. F.; Rettig, T. A.;
Stille, J. K. DielsꢀAlder Reactions of 2ꢀPyrones. Direction of the
Addition Reaction with Acetylenes. J. Org. Chem. 1969, 34, 2188ꢀ
2192.
(27) Markó, I. E.; Evans, G. R.; Declercq, J.ꢀP. Catalytic
Asymmetric DielsꢀAlder Reactions of 2ꢀPyrone Derivatives.
Tetrahedron 1994, 50, 4557ꢀ4574.
(28) Pfennig, T.; Carraher, J. M.; Chemburkar, A.; Johnson, R.
L.; Anderson, A.: Tessonnier, J.ꢀP.; Neurock, M.; Shanks, B. H. A
New Selective Route towards Benzoic Acid and Derivatives from
BiomassꢀDerived Coumalic Acid. Green Chem. 2017, 19, 4879ꢀ4888.
(29) Ipaktschi, J. Z. DielsꢀAlder Reaction in the Presence of
Zeolite. Naturforsch. B Chem. Sci. 1986; 41, 496.
(30) Narayana Murthy, Y. V. S.; Pillai, C. N. DielsꢀAlder
Reactions Catalyzed by Zeolites. Syn. Commun. 1991, 21, 783ꢀ791.
(31) Otto, S.; Bertoncin, F.; Engberts, J. B. F. N. Lewis Acid
Catalysis of a Diels−Alder Reaction in Water. J. Am. Chem. Soc.
1996, 118, 7702ꢀ7707.
(32) Majors, P. D.; Raidy, T. E.; Ellis, P. D. A Multinuclear
SolidꢀState NMR Investigation of the Chemisorption of Ammonia on
.Gamma.ꢀAlumina. J. Am. Chem. Soc. 1986, 108, 8123ꢀ8129.
(33) Wischert, R.; Laurent, P.; Copéret, C.; Delbecq, F.; Sautet,
P. γꢀAlumina: The Essential and Unexpected Role of Water for the
Structure, Stability, and Reactivity of “Defect” Sites. J. Am. Chem.
Soc. 2012, 134, 14430ꢀ14449.
Winter, A. H. Aromatics from Pyrones: Esters of Terephthalic Acid
and Isophthalic Acid from Methyl Coumalate. RSC Adv. 2013, 3,
1
2721ꢀ12725.
4) Lee, J. J.; Kraus, G. A. Divergent Diels–Alder
(
Methodology from Methyl Coumalate toward Functionalized
Aromatics. Tetrahedron Lett. 2013, 54, 2366ꢀ2368.
(
5)
Lee, J. J.; Kraus, G. A. OneꢀPot Formal Synthesis of
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
Biorenewable Terephthalic Acid from Methyl Coumalate and Methyl
Pyruvate. Green Chem. 2014, 16, 2111ꢀ2116.
(
6)
Lee, J. J.; Pollock Iii, G. R.; Mitchell, D.; Kasuga, L.;
2,7
Kraus, G. A. Upgrading Malic Acid to BioꢀBased Benzoates via a
Diels–AlderꢀInitiated Sequence with the Methyl Coumalate Platform.
RSC Adv. 2014, 4, 45657ꢀ45664.
(
7)
Zelle, R. M.; de Hulster, E.; van Winden, W. A.; de Waard,
P.; Dijkema, C.; Winkler, A. A.; Geertman, J. M.; van Dijken, J. P.;
Pronk, J. T.; van Maris, A. J. Malic Acid Production by
Saccharomyces Cerevisiae: Engineering of Pyruvate Carboxylation,
Oxaloacetate Reduction, and Malate Export. Appl. Environ.
Microbiol. 2008, 74, 2766ꢀ77.
(
8)
Zhang, X.; Wang, X.; Shanmugam, K. T.; Ingram, L. O. Lꢀ
Malate Production by Metabolically Engineered Escherichia Coli.
Appl. Environ. Microbiol. 2011, 77, 427ꢀ434.
(9)
Brown, S. H.; Bashkirova, L.; Berka, R.; Chandler, T.;
Doty, T.; McCall, K.; McCulloch, M.; McFarland, S.; Thompson, S.;
Yaver, D. Metabolic Engineering of Aspergillus Oryzae NRRL 3488
for Increased Production of LꢀMalic Acid. Appl. Microbiol.
Biotechnol. 2013, 97, 8903ꢀ8912.
(10) Werpy, T.; Petersen, G.; Aden, A.; Bozell, J.; Holladay, J.;
White, J.; Manheim, A.; Eliot, D.; Lasure, L.; Jones, S. Top value
added chemicals from biomass. Volume 1ꢀResults of screening for
potential candidates from sugars and synthesis gas; DTIC Document:
2
004.
(11) Shusherina, N. P. Diene Synthesis with 2ꢀPyrones and 2ꢀ
Pyridones . Russ. Chem. Rev. 1974, 43, 851ꢀ861.
12) Woodard, B. T.; Posner, G. H. Advances in Cycloaddition,
Vol. 5, Elsevier, 1999.
13) Markó, I. E.; Evans, G. R. Diastereoselective, Lanthanideꢀ
(
(
Catalysed, Inverse ElectronꢀDemand DielsꢀAlder (IEDDA) Reactions
of 3ꢀCarbomethoxyꢀ2ꢀPyrone (3ꢀCMP) Derivatives. Tetrahedron Lett.
1
994, 35, 2767ꢀ2770.
14) Posner, G. H.; Ishihara, Y. Lewis AcidꢀCatalyzed, High
(
Pressure, Stereospecific, Regiospecific, DielsꢀAlder Cycloaddition of
Unsubstituted 2ꢀPyrone: Short Synthesis of a Racemic AꢀRing
Precursor to Physiologically Active 1ꢀHydroxyvitamin D3 Steroids.
Tetrahedron Lett. 1994, 35, 7545ꢀ7548.
(34) Digne, M.; Sautet, P.; Raybaud, P.; Euzen, P.; Toulhoat, H.
Hydroxyl Groups on γꢀAlumina Surfaces: A DFT Study. J. Catal.
2002, 211, 1ꢀ5.
(
15) Posner, G. H.; Carry, J.ꢀC.; Kyoo Lee, J.; Bull, D. S.; Dai,
(35) Chia, M.; Haider, M. A.; Pollock, G.; Kraus, G. A.;
Neurock, M.; Dumesic, J. A. Mechanistic Insights into RingꢀOpening
and Decarboxylation of 2ꢀPyrones in Liquid Water and
Tetrahydrofuran. J. Am. Chem. Soc. 2013, 135, 5699ꢀ5708.
(36) Bond, J. Q.; Wang, D. A.; Alonso, D. M.; Dumesic, J. A.
Interconversion between γꢀValerolactone and Pentenoic Acid Comꢀ
bined with Decarboxylation to form Butene over Silica/Alumina. J.
Catal. 2011, 281, 2, 290ꢀ299.
(37) Johnson, R. L.; Hanrahan, M. P.; Mellmer, M.; Dumesic, J.
A.; Rossini, A. J.; Shanks, B. H. Solvent–Solid Interface of Acid
Catalysts Studied by High Resolution MAS NMR. J. Phys. Chem. C
2017, 121, 17226ꢀ17234.
(38) Wiench, J. W.; Bronnimann, C. E.; Lin, V. S. Y.; Pruski,
M. Chemical Shift Correlation NMR Spectroscopy with Indirect
Detection in Fast Rotating Solids:ꢂ Studies of Organically
Functionalized Mesoporous Silicas. J. Am. Chem. Soc 2007, 129,
12076ꢀ12077.
(39) Ishii, Y.; Tycko, R. Sensitivity Enhancement in Solid State
(15)N NMR by Indirect Detection with HighꢀSpeed Magic Angle
Spinning. J. Magn. Reson. 2000, 142, 199ꢀ204.
H. Mild, Asymmetric DielsꢀAlder Cycloadditions of Electronically
Matched 2ꢀPyrones and Vinyl Ethers. Tetrahedron Lett. 1994, 35,
1321ꢀ1324.
(16) Posner, G. H.; Dai, H.; Bull, D. S.; Lee, J.ꢀK.; Eydoux, F.;
Ishihara, Y.; Welsh, W.; Pryor, N.; Petr, S. Lewis AcidꢀPromoted,
Stereocontrolled, Gram Scale, Diels−Alder Cycloadditions of
Electronically Matched 2ꢀPyrones and Vinyl Ethers:ꢂ The Critical
Importance of Molecular Sieves and the Temperature of Titanium
Coordination with the Pyrone. J. Org. Chem. 1996, 61, 671ꢀ676.
(
17) Cho, C.ꢀG.; Kim, Y.ꢀW.; Lim, Y.ꢀK.; Park, J.ꢀS.; Lee, H.;
Koo, S. Diels−Alder Cycloadditions of 3,5ꢀDibromoꢀ2ꢀpyrone:ꢂ A
New Ambident Diene. J. Org. Chem. 2002, 67, 290ꢀ293.
(
18) Afarinkia, K.; Vinader, V.; Nelson, T. D.; Posner, G. H.
DielsꢀAlder Cycloadditions of 2ꢀPyrones and 2ꢀPyridones.
Tetrahedron 1992, 48, 9111ꢀ9171.
(
19) Imagawa, T.; Sueda, N.; Kawanisi, M. DielsꢀAlder
Reaction of Methyl Coumalate with 1,3ꢀDienes. Tetrahedron 1974,
0, 2227ꢀ2231.
3
ACS Paragon Plus Environment

ACS Catalysis
Page 14 of 15
(
40) Amoureux, J.ꢀP.; Trébosc, J.; Wiench, J.; Pruski, M.
(57) Kresse, G.; Furthmüller, J. Efficiency of AbꢀInitio Total
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
HMQC and RefocusedꢀINEPT Experiments involving HalfꢀInteger
Quadrupolar Nuclei in Solids. J. Magn. Reson. 2007, 184, 1ꢀ14.
Energy Calculations for Metals and Semiconductors using a Planeꢀ
Wave Basis Set. J. Comput. Mater. Sci. 1996, 6, 15ꢀ50.
(58) Kresse, G.; Furthmüller, J. Efficient Iterative Schemes for
Ab Initio TotalꢀEnergy Calculations using a PlaneꢀWave Basis Set. J.
Phys. Rev. B 1996, 54, 11169ꢀ11186.
(59) Perdew, J. P.; Ernzerhof, M.; Burke, K. Rationale for
Mixing Exact Exchange with Density Functional Approximations. J.
Chem. Phys. 1996, 105, 9982ꢀ9985.
(60) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. A Consistent
and Accurate Ab Initio Parametrization of Density Functional
Dispersion Correction (DFTꢀD) for the 94 Elements HꢀPu. J. Chem.
Phys. 2010, 132, 154104.
(61) Henkelman, G.; Jónsson, H. Improved Tangent Estimate in
the Nudged Elastic Band Method gor Finding Minimum Energy Paths
and Saddle Points. J. Chem. Phys. 2000, 113, 9978ꢀ9985.
(62) Henkelman, G.; Uberuaga, B. P.; Jónsson, H. A Climbing
Image Nudged Elastic Band Method for Finding Saddle Points and
Minimum Energy Paths. J. Chem. Phys. 2000, 113, 9901ꢀ9904.
(63) Henkelman, G.; Jónsson, H. A Dimer Method for Finding
Saddle Points on High Dimensional Potential Surfaces using Only
First Derivatives. J. Chem. Phys. 1999, 111, 7010ꢀ7022.
(64) Wischert, R.; Copéret, C.; Delbecq, F.; Sautet, P. Optimal
Water Coverage on Alumina: A Key to Generate Lewis Acid–Base
Pairs that are Reactive Towards the CꢀH Bond Activation of Methane.
Ang. Chem. Int. Ed. 2011, 50, 3202ꢀ3205.
(65) Monkhorst, H. J.; Pack, J. D. Special Points for Brillouinꢀ
Zone Integrations. Phys. Rev. B 1976, 13, 5188ꢀ5192.
(
41) Venkatesh, A.; Hanrahan, M. P.; Rossini, A. Proton
Detection of MAS SolidꢀState NMR Spectra of HalfꢀInteger
Quadrupolar Nuclei. J. Solid State Nucl. Magn. Reson. 2017, 84, 171ꢀ
181.
(
42) Valla, M.; Rossini, A. J.; Caillot, M.; Chizallet, C. l.;
Raybaud, P.; Digne, M.; Chaumonnot, A.; Lesage, A.; Emsley, L.;
Van Bokhoven, J. A.; Copéret, C. Atomic Description of the Interface
between Silica and Alumina in Aluminosilicates through Dynamic
Nuclear Polarization SurfaceꢀEnhanced NMR Spectroscopy and Firstꢀ
Principles Calculations. J. Am. Chem. Soc. 2015, 137, 10710ꢀ10719.
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
(
43) Perras, F. A.; Padmos, J. D.; Johnson, R. L.; Wang, L.ꢀL.;
Schwartz, T. J.; Kobayashi, T.; Horton, J. H.; Dumesic, J. A.; Shanks,
B. H.; Johnson, D. D.; Pruski, M. Characterizing Substrate–Surface
Interactions on AluminaꢀSupported Metal Catalysts by Dynamic
Nuclear
Spectroscopy. J. Am. Chem. Soc. 2017, 139, 2702ꢀ2709.
44) Stanley, B. J.; Guiochon, G. Calculation of Adsorption
PolarizationꢀEnhanced
DoubleꢀResonance
NMR
(
Energy Distributions of Silica Samples Using Nonlinear
Chromatography. Langmuir 1995, 11, 1735ꢀ1743.
(45) Lunsford, J. H.; Rothwell, W. P.; Shen, W. Acid Sites in
Zeolite Y:
Trimethylphosphine as a Probe Molecule. J. Am. Chem. Soc. 1985,
07, 1540ꢀ1547.
46) Wouters, B. H.; Chen, T. H.; Grobet, P. J. Reversible
A SolidꢀState NMR and Infrared Study using
1
(
Tetrahedral−Octahedral Framework Aluminum Transformation in
Zeolite Y. J. Am. Chem. Soc. 1998, 120, 11419ꢀ11425.
(66) Campbell, C. T.; Sellers, J. R. V. The Entropies of
Adsorbed Molecules. J. Am. Chem. Soc. 2012, 134, 18109ꢀ18115.
(67) Zhao, Y.; Truhlar, D. G. A New Local Density Functional
for MainꢀGroup Thermochemistry, Transition Metal Bonding,
Thermochemical Kinetics, and Noncovalent Interactions. J. Chem.
Phys. 2006, 125, 194101.
TM
(
47) Amberlyst 45 Resin High Temperature Strongly Acidic
Catalyst; Form No. 177ꢀ02338ꢀ0315, Rev. 0 [Online]; The Dow
Chemical Company: Midland, MI.
http://msdssearch.dow.com/PublishedLiteratureDOWCOM/dh_092d/
0901b8038092d737.pdf?filepath=liquidseps/pdfs/noreg/177ꢀ
0
2338.pdf&fromPage=GetDoc (accessed Aug 10, 2017)
48) GómezꢀBombarelli, R.; Calle, E.; Casado, J. Mechanisms
of Lactone Hydrolysis in Acidic Conditions. J. Org. Chem. 2013, 78,
(68) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G.
E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.;
Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.;
Hratchian; P., H.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg,
J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.;
Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven,
T.; Montgomery, J. A., Jr.; Peralta; E., J.; Ogliaro, F.; Bearpark, M.;
Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi,
R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.;
Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene,
M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.;
Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi,
R.; Pomelli; C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.;
Zakrzewski, V. G.; Voth, G.; A.; Salvador, P.; Dannenberg, J. J.;
Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.;
Cioslowski, J.; Fox, D. J., Gaussian 09, Revision A.02, Gaussian,
Inc., Wallingford CT, 2016.
(69) Frisch, M. J.; Pople, J. A.; Binkley, J. S. Self‐Consistent
Molecular Orbital Methods 25. Supplementary Functions for
Gaussian Basis Sets. J. Chem. Phys.1984, 80, 3265ꢀ3269.
(70) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. Universal
Solvation Model Based on Solute Electron Density and on a
Continuum Model of the Solvent Defined by the Bulk Dielectric
Constant and Atomic Surface Tensions. J. Chem. Phys. B 2009, 113,
6378ꢀ6396.
(
6
880ꢀ6889.
(49) J. Sheehan, R. Terephthalic Acid, Dimethyl Terephthalate,
and Isophthalic Acid. Ullmann's Encyclopedia of Industrial
Chemistry, WileyꢀVCH Verlag GmbH & Co. KGaA: 2000.
(
50) Chen, T.ꢀH.; Wang, L.; Trueblood, J. V.; Grassian, V. H.;
Cohen, S. M. Poly(isophthalic acid)(ethylene oxide) as
a
Macromolecular Modulator for Metal–Organic Polyhedra. J. Am.
Chem. Soc. 2016, 138, 9646ꢀ9654.
(
51) Pfennig, T.; Johnson, R. L.; Shanks, B. H. The Formation
of pꢀToluic Acid from Coumalic Acid: A Reaction Network Analysis.
Green Chem. 2017, 19, 3263ꢀ3271
(
52) Brinkmann, A.; Kentgens, A. P. M. ProtonꢀSelective
17O−1H Distance Measurements in Fast MagicꢀAngleꢀSpinning
SolidꢀState NMR Spectroscopy for the Determination of Hydrogen
Bond Lengths. J. Am. Chem. Soc. 2006, 128, 14758ꢀ14759.
(
53) Metz, G.; Wu, X. L.; Smith, S. O. RampedꢀAmplitude
Cross Polarization in MagicꢀAngleꢀSpinning NMR. J. Magn. Reson.,
Ser A 1994, 110, 219ꢀ227.
(54) Fung, B. M.; Khitrin, A. K.; Ermolaev, K. An Improved
Broadband Decoupling Sequence for Liquid Crystals and Solids. J.
Magn. Reson. 2000, 142, 97ꢀ101.
(
55) Kresse, G.; Hafner, J. Ab Initio Molecular Dynamics for
Liquid Metals. J. Phys. Rev. B 1993, 47, 558ꢀ561.
(56) Kresse, G.; Hafner, J. Ab Initio MolecularꢀDynamics
Simulation
of
the
LiquidꢀMetal–AmorphousꢀSemiconductor
Transition in Germanium. J. Phys. Rev. B 1994, 49, 14251ꢀ14269.
ACS Paragon Plus Environment

Page 15 of 15
ACS Catalysis
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
ACS Paragon Plus Environment