Tuning Catalytic Sites on Zr6O8 Metal-Organic Framework Nodes via Ligand and Defect Chemistry Probed with tert-Butyl Alcohol Dehydration to Isobutylene

Article abstract of DOI:10.1021/jacs.0c03175

Metal-organic frameworks (MOFs) have drawn wide attention as candidate catalysts, but some essential questions about their nature and performance have barely been addressed. (1) How do OH groups on MOF nodes act as catalytic sites (2) What are the relationships among these groups, node defects, and MOF stability, and how do reaction conditions influence them (3) What are the interplays between catalytic properties and transport limitations To address these questions, we report an experimental and theoretical investigation of the catalytic dehydration of tert-butyl alcohol (TBA) used to probe the activities of OH groups of Zr₆O₈ nodes in the MOFs UiO-66 and MOF-808, which have different densities of vacancy sites and different pore sizes. The results show that (1) terminal node OH groups are formed as formate and/or acetate ligands present initially on the nodes react with TBA to form esters, (2) these OH groups act as catalytic sites for TBA dehydration to isobutylene, and (3) TBA also reacts to break node-linker bonds to form esters and thereby unzip the MOFs. The small pores of UiO-66 limit the access of TBA and the reaction with the formate/acetate ligands bound within the pores, whereas the larger pores of MOF-808 facilitate transport and favor reaction in the MOF interior. However, after removal of the formate and acetate ligands by reaction with methanol to form esters, interior active sites in UiO-66 become accessible for the reaction of TBA, with the activity depending on the density of defect sites with terminal OH groups. The number of vacancies on the nodes is important in determining a tradeoff between the catalytic activity of a MOF and its resistance to unzipping. Computations at the level of density functional theory show how the terminal OH groups on node vacancies act as Br?nsted bases, facilitating TBA dehydration via a carbocation intermediate in an E1 mechanism; the calculations further illuminate the comparable chemistry of the unzipping.

Full text of DOI:10.1021/jacs.0c03175

Subscriber access provided by University of Wollongong Library
Article
Tuning Catalytic Sites on Zr6O8 Metal Organic Framework Nodes via Ligand
and Defect Chemistry probed with t-Butyl Alcohol Dehydration to Isobutylene
Dong Yang, Carlo Alberto Gaggioli, Debmalya Ray, Melike Babucci, Laura Gagliardi, and Bruce C. Gates
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.0c03175 • Publication Date (Web): 05 Apr 2020
Downloaded from pubs.acs.org on April 5, 2020
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.
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
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Tuning Catalytic Sites on Zr6O8 Metal Organic
Framework Nodes via Ligand and Defect Chemistry
probed with t-Butyl Alcohol Dehydration to
Isobutylene
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
Dong Yang,a† Carlo Alberto Gaggioli,b† Debmalya Ray,b Melike Babucci,a Laura Gagliardi,b* and
Bruce C. Gatesa*
a) Department of Chemical Engineering, University of California, Davis, California 95616, United States
b) Department of Chemistry, Chemical Theory Center, and Minnesota Supercomputing Institute, University of
Minnesota, Minneapolis, MN 55455, United States
ABSTRACT: Metal organic frameworks (MOFs) have drawn wide attention as candidate catalysts, but some essential questions
about their nature and performance have barely been addressed: 1. How do OH groups on MOF nodes act as catalytic sites? 2. What
are the relationships between these groups, node defects, and MOF stability, and how do reaction conditions influence them? 3. What
are the interplays between catalytic properties and transport limitations? To address these questions, we report an experimental and
theoretical investigation of the catalytic dehydration of t-butyl alcohol (TBA) used to probe the activities of OH groups of Zr6O8
nodes in the MOFs UiO-66 and MOF-808, which have different densities of vacancy sites and different pore sizes. The results show
that (1) terminal node OH groups form as formate and/or acetate ligands present initially on the nodes react with TBA to form esters;
(2) these OH groups act as catalytic sites for TBA dehydration to isobutylene; and (3) TBA also reacts to break node-linker bonds to
form esters and thereby unzip the MOFs. The small pores of UiO-66 limit the access of TBA and the reaction with the formate/acetate
ligands bound within the pores, whereas the larger pores of MOF-808 facilitate transport and favor reaction in the MOF interior.
However, after removal of the formate and acetate ligands by reaction with alcohols to form esters, interior active sites in UiO-66
become accessible for the reaction of TBA, with the activity depending on the density of defect sites with terminal OH groups. The
number of vacancies on the nodes is important in determining a tradeoff between the catalytic activity of a MOF and its resistance to
unzipping. Computations at the level of density functional theory show how the terminal OH groups on node vacancies act as Brønsted
bases, facilitating TBA dehydration via a carbocation intermediate in an E1 mechanism; the calculations further illuminate the
comparable chemistry of the unzipping.
INTRODUCTION
by X-ray diffraction crystallography, because they are often
present in such low concentrations as to escape detection. Yet
these components, or defects where linkers are missing, may be
crucial to MOF reactivity and catalysis. Ligands on node
vacancies commonly arise (a) from solvents or products of
solvent decomposition during MOF synthesis (e.g., formate,
formed from dimethylformamide, DMF);12-13 (b) from
modulators added to facilitate MOF synthesis (e.g., acetate,
from acetic acid);14-15 or (c) from post-synthesis treatments,
even including washing (e.g., with solvents that leave behind
methoxy, ethoxy, hydroxyl, or sulfate ligands).16-18 In catalysis,
these ligands can act as reaction intermediates (e.g., formate or
acetate for esterification); be active sites themselves (e.g.,
providing proton-donor or acceptor sites); or be reaction
inhibitors by blocking active sites on the nodes (e.g.,
benzoate).4, 13, 19-20 Removal of such ligands may be essential for
creation of catalytic activity, for example, by opening Lewis
acid sites to catalyze reactions such as dehydration,
condensation, and hydrolysis.11, 13, 20
Metal-organic frameworks (MOFs) are an enormous and
rapidly expanding class of crystalline, porous materials formed
by coordination of organic linkers with inorganic nodes.1-3
Some MOFs are good candidate catalysts, offering the
advantages of high densities of accessible active sites, high
degrees of structural tunability, and high thermal stability.4-8
MOFs that are exemplary in these respects prominently include
those with Zr6O8 or Hf6O8 nodes, which offer catalytic sites and
platforms for catalytic sites.9-11 Because they are metal oxide
clusters, these nodes are comparable to catalysts in an extremely
common class, metal oxides, but they are distinct because the
node surfaces differ from bulk metal oxide surfaces.4 An
advantage of these MOFs over metal oxides is the structural
uniformity of the nodes, in contrast to the heterogeneity of the
surfaces of amorphous metal oxides; consequently,
fundamental investigations of structure-catalytic property
relationships of the MOFs offer the prospect of deeper
fundamental understanding than investigations of the metal
oxides—and the further prospect of insights from MOF
chemistry that extend to metal oxides.
In contrast to the aforementioned ligands on MOF nodes, OH
groups on the nodes (like those on metal oxide surfaces) are
often regarded as native functional groups, arising from water
in the synthesis solvents.14, 21-22 Often detected by infrared (IR)
The components on node surfaces—besides the linkers—are
typically not represented in MOF crystal structures determined
ACS Paragon Plus Environment

Journal of the American Chemical Society
spectroscopy,14, 23 these OH groups offer Brønsted acid and base
Page 2 of 15
1
2
3
4
5
6
7
8
sites that may play roles in catalysis—but, surprisingly, in view
of the catalytic importance of such sites on metal oxides—their
subtle chemistry remains to be characterized in depth.4 Our
goals were to better understand this node ligand chemistry,
especially including the OH group chemistry, and relate it to
catalytic properties of the MOFs.
UiO-66 was synthesized as before13 with acetic acid as a
modulator. Analysis of the MOF by H NMR spectroscopy of
1
the solution formed by dissolution of the MOF in NaOH/D2O
(Table 1) showed that a total 1.05 bonding sites per node
(excluding linker bonding sites) had been generated during the
synthesis—with these sites occupied by acetate ligands (0.80
per node, formed from the acetic acid modulator) and by
formate ligands (0.25 per node, formed by decomposition of
DMF used in the synthesis).12-13 IR spectra of the MOF confirm
the presence of these ligands, represented by bands at 2953,
2938, and 2923 cm-1, assigned to νC-H of acetate, and at 2747
cm-1, assigned to νC-O + δC-H of formate,13 matching reported
results.13
Beyond chemistry, understanding of mass transport is important
for understanding catalysis by porous materials such as MOFs,
because it may determine the accessibility of catalytic sites
within the pores.4 This topic is also underexplored, in particular
as it pertains to catalysis by defects on MOFs, and we also
address it here, along with the linked issues of catalyst stability,
which have also been largely overlooked.
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
We chose MOFs incorporating Zr6O8 nodes (UiO-66 and MOF-
808) because of their well-established chemistry and the
opportunities they offer for modifying the node ligands.12, 14, 16
As a catalytic test reaction, we chose t-butyl alcohol (TBA)
dehydration, because it has been well investigated with metal
oxide and other solid catalysts (such as zeolites and cation
exchange resins), often being considered a prototypical
Brønsted acid-catalyzed reaction that is straightforward to
investigate, in part because it usually gives only a single
dehydration product, isobutylene.24-25 Our experimental results,
bolstered by DFT calculations, identify the catalytic sites as
node OH groups; elucidate the catalytic reaction mechanism;
provide evidence of how the MOF pore size affects the transport
of TBA and the accessibility of interior catalytic sites; and show
how the catalysts may be deactivated by unzipping.
Formic acid was used to modulate the synthesis of MOF-808.
Ideally, in the absence of ligands formed from the synthesis
mixture, this MOF would have 6 vacancies per node, as
indicated by its bulk crystal structure (Figure 1).26 However, the
1H NMR data characterizing the product formed by dissolution
of the MOF in NaOH/D2O (Table 1) show that each node, on
average, was bonded to 4.9 formate ligands, a result bolstered
by a strong IR band at 2745 cm-1 representing formate on the
node vacancies (Figure S1 in the Supporting Information, SI).
The IR spectrum of MOF-808 also indicates OH groups
(formed by H2O chemisorption on the nodes), characterized by
a band at 3586 cm-1 (Figure S1, SI), consistent with the
formulation Zr6O5(OH)3(BTC)2(HCOO)5(H2O)2 presented by
Yaghi’s group,16 where BTC represents the linker formed from
the precursor benzene-1,3,5-tricarboxylic acid.
RESULTS
Table 1. Changes of BET surface areas, numbers of nonlinker
ligands per node determined by H NMR spectroscopy of
dissolved sample, and rates of catalytic TBA dehydration
observed for UiO-66 and MOF-808 after various times on
stream in the flow reactor at 150 °C and atmospheric pressure.
1
Ligands Initially Present on Zr6O8 MOF nodes.
MOF
Time on
stream in
the flow
reactor
used for
TBA
BET
surfa
ce
area
of
Numb
er of
forma
te
Numb
er of
acetat
e
Total
number
of
nonlinke
r
carboxyl
ate
ligands
per node
103
rate of
TBA
conversi
on,
×
ligand
ligand
MOF
s
per
s
per
molecule
dehydrati
on, h
(m2
node
node
s node-1 s-
1
g-1)
UiO-
66
0
132
7
0.25
0.80
1.05
0.02a
6
989
0.18
0.16
0.79
0.77
0.97
0.93
0.17
0.20
13
104
7
24
0
988
0.15
4.90
0.76
-
0.91
4.90
0.22
2.12a
MO
F-
159
1
808
6
903
211
104
2.14
1.76
1.28
-
-
-
2.14
1.76
1.28
0.47
0.35
0.23
13
24
aInitial rates were determined by extrapolating differential
conversion data to zero time on stream.
Figure 1. Illustration of crystal structures of (a) UiO-66 and (b)
MOF-808, with details showing (c) missing-linker defects of UiO-
66 and (d) structural vacancies of MOF-808. Ligands on
defects/vacancies are shown in blue.
2
ACS Paragon Plus Environment

Page 3 of 15
Journal of the American Chemical Society
TBA Dehydration Catalyzed by UiO-66 and MOF-808. from the reactor after various times on stream, complemented
by IR spectra of the catalysts and 1H NMR measurements of the
liquid samples formed by dissolution of the catalysts in NaOH
in D2O. The on-line GC data recorded periodically (primarily to
determine the product composition and thereby characterize the
catalytic activity, Table 1) provide complementary information,
showing that ligands initially present on the MOF nodes reacted
with the TBA and desorbed, flowing with the reaction products
and unconverted reactant to the GC. Thus, the GC data indicate
the formation of t-butyl acetate and t-butyl formate, which
formed by the reaction of TBA with the acetate and formate
ligands, respectively, that were initially present on the MOF
nodes. Figure S7 in the SI shows data demonstrating the
formation of these esters from MOF-808 and from UiO-66
during catalysis. The TBA conversion data (measured as the
formation of isobutylene) indicate that MOF-808 catalyst was
initially about 100 times more active than UiO-66 (Table 1), but
it underwent rapid deactivation during operation, with only
11% of the initial activity (measured as rate of reaction
determined from differential conversions, Table 1) remaining
after 24 h of continuous reaction. XRD data characterizing the
catalyst samples that had been used for various times on stream
show that the crystallinity of MOF-808 decreased as it
functioned as a catalyst, with an almost complete loss of
crystallinity after only 13 h onstream; correspondingly, BET
surface area data (Table 1) show that MOF-808 had lost most
of its porosity after this period. Thus, we infer that the
deactivation of MOF-808 resulted from its unzipping,
comparable to the deactivation of UiO-66 used for ethanol
dehydration catalysis.13 This unzipping is explained by the
breaking of node–linker bonds13 as TBA reacted with the BTC
linkers bonded to the nodes, leading to the formation of esters.
Evidence of these esters is provided not only by the
aforementioned GC data but also by IR spectra of the used
catalysts, with bands at 1710 and 1761 cm-1, attributed
respectively to esters and free COOH groups of the unlinked
BTC.27 These IR bands increased in intensity over 24 h of
catalyst operation at 150 °C (Figure S2, SI; details of the band
assignments are presented in the SI). Further, as a result of
reaction with TBA, 74% of the formate ligands initially present
on the node vacancies had been removed after 24 h of operation
of the catalyst in the flow reactor, as shown by the 1H NMR data
(Table 1).
1
Comparison of the MOFs. We compared the activities of UiO-
66 and MOF-808 for TBA dehydration by using a once-through
plug-flow reactor operated at 150 °C and atmospheric pressure,
measuring conversions by gas chromatographic (GC) analysis
of the product stream. Helium carrier gas flowed through a
bubbler in which it was saturated with TBA upstream of the
reactor. Catalyst performance data are shown in Figure 2 and
Table 1. In contrast to ethanol dehydration (for which diethyl
ether was the only dehydration product), the dehydration
product of the TBA reaction catalyzed by these MOFs was an
olefin, isobutylene, with no detectable ether. MOF-808 was
more than 100 times higher in initial activity than UiO-66 under
our conditions in the flow reactor (150 °C; 75 mbar of TBA and
924 mbar of helium; total feed TBA flow rate 5.4
mL(NTP)/min; catalyst mass, 25–100 mg).
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
To understand the contrast between these MOFs, we consider
(a) their different stabilities; (b) the differences in the numbers
of node sites that were not blocked by formate or acetate
(vacancy sites); and (c) the pore sizes, which affect accessibility
of the catalytic sites. MOF-808 has, on average, 6 times more
vacancy sites per Zr6O8 node than UiO-66 (Table 1), and MOF-
808 also has a larger pore aperture diameter than UiO-66 (14
vs. 6 Å). We infer that both of these differences are significant
in influencing the catalytic properties, as summarized below.
12
10
8
6
4
2
0
0
4
8
12
16
20
24
Time on Stream (h)
Figure 2. Conversion of TBA in a flow reactor catalyzed by UiO-
66 (red), MOF-808 (black), γ-Al2O3 (light blue), and ZrO2 (dark
blue). Isobutylene was the only observed product of the catalytic
reaction. The results of blank experiments are as follows: with 2 g
of inert α-Al2O3 but no catalyst in the reactor, the conversion was
essentially negligible, only 0.3 ± 0.1%; bare H2BDC linker was
found to be inactive. Reaction conditions: temperature, 150 °C;
feed partial pressures, 75 mbar of TBA and 924 mbar of helium;
total flow rate 5.4 mL(NTP)/min; catalyst masses: 25 mg of MOF-
808 and 100 mg of each of the following: UiO-66, γ-Al2O3, and
ZrO2.
B
A
4
20
6
8
10
12
14
16
18
20
4
6
8
10
12
14
16
18
Changes in catalysts during operation in a flow reactor. To test
for changes in composition and crystallinity of the MOFs as
they functioned as catalysts, we determined changes in their
structures occurring during the catalytic reaction by making X-
ray diffraction (XRD) measurements of the catalysts removed
2 Theta (o)
2 Theta (o)
Figure 3. XRD patterns of as-synthesized MOF-808 (A) and UiO-
66 AA (B) (black) and the samples after TBA dehydration catalysis
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 15
in the flow reactor at 150 °C for the following times on stream (h):
6 (red), 13 (cyan), and 24 (blue).
formed by the reaction of linkers with TBA, with all of these
being inactive for TBA dehydration at 150 °C.
1
2
3
4
5
6
7
8
9
Role of transport limitations in catalyst pores. In comparing the
initial activities of the two MOFs (those with intact structures
and the formate and acetate ligands remaining from the
syntheses), we need to consider the intrinsic activity, the
number of sites not inhibited by formate and acetate, and also
the influence of transport effects. Rates of catalytic reactions in
porous materials often depend on how fast the reactant
molecules are delivered to the catalytic sites: if there is little
resistance to transport (e.g., in large pores and/or with small
reactant molecules) and the intrinsic activity of the catalyst is
low, reactants may reach all the interior sites equally. In
contrast, when the transport resistance is large or the intrinsic
activity of the catalyst high, the reactants may be largely
converted before they reach the interior space of the catalyst—
and in the limiting case, all the reaction takes place at the
exterior surfaces of the catalyst particles. Because the two
MOFs used in this work had the same nodes and similar sites
on the node surfaces, but differed markedly from each other in
terms of pore geometry and resistance to transport, we infer that
transport was important in determining the differences in their
catalytic properties. The UiO-66 particles, having pore
apertures 6 Å in diameter, nearly matching the critical diameter
of TBA (6.1 × 6.7 × 5.7 Å, calculated by DFT), offered more
resistance to TBA transport than the MOF-808 (14 Å apertures),
so that the initial catalytic activities (Table 1) are consistent
with the inference that the TBA reached more interior sites to
react on MOF-808 than on UiO-66. But the data also show that
most of the defect sites in UiO-66 remained covered with the
formate and acetate ligands that were present initially—not
being replaced by ligands formed from the TBA and not
engaged in the catalysis (Table 1) (that is, there was greater
inhibition of node sites on UiO-66 than on MOF-808). Thus,
the data are not sufficient for a quantitative resolution of the
separate effects of transport and site inhibition in the two MOFs,
although it is clear that both are important.
In contrast to MOF-808, UiO-66 gradually increased in activity
during catalysis in the flow reactor, so that the activities of the
two catalysts became essentially the same after 24 h onstream.
The increase was accompanied by the removal of formate and
acetate ligands, as shown by the GC analyses indicating the
esters that formed by reaction of these ligands with TBA. The
data of Figure S7 in the SI show that the amounts of formate
and acetate desorbed from UiO-66 were much less than those
desorbed from MOF-808, and, consistent with these data, the
1H NMR results confirm that only 13% of these ligands initially
present on UiO-66 had been removed after 24 h of continuous
catalytic reaction at 150 °C. We infer that the observed
activation of the UiO-66 catalyst during operation resulted from
the removal of the inhibitor formate and acetate ligands,
opening catalytic sites on the nodes. After 24 h of continuous
reaction, UiO-66 still had an intact crystal structure with just
slightly decreased XRD peak intensities (Figure 3) and a high
BET surface area of 988 m2/g (vs. 1392 m2/g initially).
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
5
4
3
2
1
0
0
20
40
60
80
100
The removal of inhibitors from the UiO-66 sites in TBA
dehydration is contrasted with that observed for the same MOF
catalyzing the dehydration of ethanol—this smaller alcohol (4.2
× 4.9 × 5.8 Å, calculated by DFT) almost completely removed
these ligands from the nodes within 4 h at 150 °C,13 whereas
only 13% of these ligands on the MOF outer surface were
removed by TBA after 24 h. However, we stress that the
comparison does not afford a quantitative resolution of the
reaction/transport tradeoff, and that TBA dehydration is much
faster than ethanol dehydration on the MOF nodes, as shown by
the lack of measurable catalytic activity of UiO-66 for ethanol
dehydration at our reaction temperature of 150 °C.
Time on Stream (h)
Figure 4. Conversion of TBA to isobutylene catalyzed by UiO-66
in a long-term flow reactor experiment (100 h). Reaction
conditions: temperature, 150 °C; feed partial pressures, 75 mbar of
TBA and 924 mbar of helium; total flow rate 5.4 mL(NTP)/min;
catalyst mass: 100 mg.
To provide further evidence of the stability of UiO-66 under our
TBA dehydration conditions, we extended the period of testing
from about 24 h (Figure 2) to 100 h. The data from the longer
run (Figure 4) show that the activity gradually decreased after a
21-h induction period, indicating that the MOF was being
slowly unzipped during catalysis—thus we infer that
deactivation was occurring as activation was also occurring
initially. The data show that UiO-66 is much more stable than
MOF-808, retaining 84% of its highest activity after 100 h
onstream. This greater stability of UiO-66 is associated with its
far lower density of defect sites per node than MOF-808 (1.05
vs. 6).
Tuning the Ligands on Zr6O8 MOF nodes of UiO-66.
To provide data characterizing the performance of the
undeactivated (not unzipped) catalyst in the absence of the
inhibition by formate and acetate on the nodes of UiO-66, we
used two methods to remove these ligands from the nodes: (1)
treatment of the MOF initially incorporating these ligands with
methanol and water to remove them and (2) treatment of the
MOF at a high temperature (320 °C) to remove them. We
emphasize that UiO-66 maintained its crystallinity under these
treatment conditions, but MOF-808 was not stable.
On the basis of these data, we infer the collapse of the MOFs
led to the formation of aggregated ZrO2 clusters as well as esters
4
ACS Paragon Plus Environment

Page 5 of 15
Journal of the American Chemical Society
µ3-OH groups, characterized by a band at 3674 cm-1, were still
1
2
3
4
present, indicating that the sample had not been fully
dehydrated. Thus, we infer that a mixture of species (nodes with
terminal OH groups and dehydrated nodes) was present in this
sample (Scheme 2).
5
6
7
8
Scheme 1. Topology change resulting from methanol treatment
and water treatment inducing reactions on missing linker site of
UiO-66 nodes. Reaction between methanol and acetate ligands
is suggested to proceed equivalently, with methyl acetate as a
product. The sample after methanol treatment at 200 °C,
designated as UiO-66-OCH3, incorporated a mixture of species,
as shown in the dashed rectangle. The sample after the
subsequent water treatment is denoted as UiO-66-OH.
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
Scheme 2. Topology change of UiO-66 nodes resulting from
treatment at 320 °C under high vacuum. Decomposition of
acetate ligands is suggested to proceed similarly. The resultant
sample is designated as UiO-66-320; the resultant mixture of
species in denoted with the dashed rectangle.
Our previous results have shown that formate ligands on Zr6O8
nodes of UiO-66 react with methanol to form methyl formate
and replace the formate with methoxy.13, 18, 28 To clarify: the
change involves two reactions, first, a methanol molecule
reacting with a formate ligand to form a methyl formate
molecule and a terminal OH group on the node, and, second, a
methanol molecule reacting with a terminal OH group to form
water and a methoxy ligand (Scheme 1).
Ligand Effect on TBA Dehydration Catalyzed by UiO-66
We tested the catalytic activities of UiO-66-OCH3, UiO-66-OH,
and UiO-66-320 for TBA dehydration to gain more insight into
the effects of ligands on node defects and transport limitations.
Similar to the initial UiO-66, UiO-66-OCH3 increased slightly
in activity at the beginning of the flow reactor experiment
(Figure 5), accompanied by the desorption of methanol from the
catalyst, as detected by GC analysis of the product stream
(Figure S8, SI). The data indicate that methoxy ligands on node
defects inhibited the reaction, but the inhibition was much less
pronounced than that attributed to formate and acetate
ligands—because the methoxy ligands are more highly reactive
with TBA than the formate and acetate, and, as a result, the
induction period (around 3 h) is much shorter than that observed
with the as-synthesized MOF (21 h). After the methoxy ligand
removal, the sample was gradually deactivated, as a result of
the slow unzipping of the MOF. The highest conversion in the
reaction with UiO-OCH3 was 24.7%, corresponding to a
reaction rate per node of 1.20 × 10-3 s-1 which is 6-fold higher
than that observed with the initial UiO-66. We infer that the
higher activity of UiO-OCH3 is a consequence of the greater
number of sites opened for catalysis in that sample. On the basis
of the activity data, we estimate that 78% (6 × 13%, a ratio of
defect sites on the nodes of UiO-66 opened for the catalytic
reaction after 24 h) of the defect sites in UiO-66-OCH3
participated the reaction (assuming the number of defect sites
did not change after methanol treatment). However, we
emphasize that UiO-66-OCH3 will have lost some of its defect
sites as a result of some unzipping under the methanol treatment
conditions of 200 °C for 48 h (catalytic methanol dehydration
also took place). Therefore, we infer that the 78% may reflect
the number of defects remaining after the methanol treatment,
and these were accessible for catalysis.
To completely remove the formate ligands, we treated UiO-66
(200 mg) with methanol vapor at 200 °C for 48 h, giving a
sample designated as UiO-66-OCH3. After this treatment, IR
spectra indicated that formate ligands (with a νC−O + δC−H band
at 2745 cm-1) had been replaced by monodentate methoxy
ligands, with νC−H bands at 2929 and 2826 cm-1, and terminal
OH ligands on the nodes, with a νO−H band at 3778 cm-1 (Figure
S4, SI). Thus, a mixture of species formed as a result of the
methanol treatment. We emphasize that it is difficult to prepare
a UiO-66 sample with only methoxy ligands on the nodes under
these conditions, because the catalytic methanol dehydration
reaction takes place, making dimethyl ether and water—and the
water reacts with the methoxy ligands to form methanol and
terminal OH ligands. 1H NMR spectroscopy was used to
determine the number of methoxy ligands after the methanol
treatment. The data (Figure S10, SI) show no evidence of
formate or acetate, confirming that they had been completely
removed. The data (Figure S10, SI) also show that the UiO-66-
OCH3 sample incorporated 0.45 methoxy ligands per node,
which is less than the number per node in the original sample
(1.05) and indicates there were terminal OH groups on the
nodes, a result that is bolstered by the IR data.
To remove methoxy ligands, we further treated the sample of
UiO-66-OCH3 (200 mg) with flowing water vapor at 150 °C for
10 h, giving a sample designated as UiO-OH. As result of the
treatment, all the methoxy ligands were removed, and only
terminal OH ligands (characterized by an IR band at 3780 cm-
1) remained, as shown by both IR (Figure S5, SI) and 1H NMR
data (Figure S11, SI).
Alternatively, formate and acetate ligands on node defect sites
of UiO-66 were removed by a high-temperature treatment,
which instead led to a dehydrated sample. Following the
conditions reported by De Vos et al.,11 we treated UiO-66 at 320
°C under high vacuum for 12 h. Both IR (Figure S6, SI) and 1H
NMR data (Figure S12, SI) show that formate and acetate
ligands were completely removed as a result of the treatment.
However, the IR spectrum (Figure S6, SI) also shows that
terminal OH groups, characterized by a band at 3780 cm-1, and
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Insights into nature of catalytic sites on MOF nodes.
Page 6 of 15
1
2
3
4
5
6
7
8
25
20
15
10
5
4
3
2
1
0
3
2
1
0
B
A
1209
1041
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
3781
1300
1250
1200
1150
1100
1050
1000
3800
3700
3600
3500
Wavenumber (cm-1
)
Wavenumber (cm-1
)
0
0
2
4
6
8
10
12
14
16
Time on Stream (h)
Figure 6. IR spectra charactering as-synthesized UiO-66 (black)
and the sample after TBA dehydration catalysis at 150 oC for 6 h
(red) followed by treatment under vacuum at 150 oC for 12 h (blue).
Figure 5. Conversion of TBA to isobutylene catalyzed by UiO-66-
OCH3 (red), UiO-66-OH (brown), and UiO-66-320 (dark blue);
isobutylene was the only product of the reaction. Conversion in a
blank experiment with 2 g of inert α-Al2O3 but no catalyst in the
reactor was essentially negligible, only 0.3 ± 0.1%. Reaction
conditions: temperature, 150 °C; feed partial pressures, 75 mbar of
TBA and 924 mbar of helium; total flow rate 5.4 mL(NTP)/min;
catalyst mass, 100 mg.
Because the as-synthesized UiO-66 underwent gradual
activation during catalysis at 150 °C, with the removal of
formate and acetate ligands, and because MOF-808 had a much
higher initial activity than UiO-66, we inferred, as stated above,
that the catalytic sites for TBA dehydration were located at
vacancies, with the activity increasing as these opened up
during catalysis. To provide more information about the
changes, we monitored the MOF structure by IR spectroscopy
(Figure 6). After 6 h of catalytic reaction, two new bands had
appeared, at 1041 and 1209 cm-1, which nearly match the bands
of TBA in the liquid phase, at 1016 and 1200 cm-1.31 The
differences in frequencies of these bands indicate a significant
interaction of the alcohol with node vacancies, but the
interaction did not lead to the formation of t-butoxy ligands (as
inferred from the IR data32: zirconium t-butoxide has IR bands
at 1002 and 1186 cm-1).
Because the methoxy ligands had been removed in the treatment
with water, the catalysis data characterizing UiO-66-OH were
not characterized by a clear induction period (Figure 5). The
initial activity of UiO-66-OH was 0.71× 10-3 s-1 (with a
conversion of 14.6%), which is 3.5-fold higher than the initial
activity of UiO-66 in its as-prepared state but lower than that of
UiO-66-OCH3. We infer that the observed decrease in activity
of UiO-66-OCH3 after water treatment (that is the UiO-66-OH
sample) was a consequence of the partial decomposition
(unzipping) of the MOF under hydrothermal conditions at 150
°C, as had been shown before.28
Because all the organic ligands on the defect sites of UiO-66-
320 had been removed, this sample was also characterized by a
lack of an induction period in the flow reactor experiments
(Figure 5). The initial activity of this sample, 0.31 × 10-3 s-1, is
the lowest among the three samples that did not contain formate
and/or acetate ligands, and it slowly decreased during catalysis.
The low activity is not attributed to the decomposition
(unzipping) of the MOF, because the MOF was found by XRD
to still be intact after the 320 °C treatment under high vacuum;
rather, the deactivation is attributed to dehydration of the MOF,
as explained below.
To further clarify the chemistry, we evacuated the UiO-66
sample after 6 h of catalytic reaction, holding it at 150 °C for 12
h. The resulting IR spectrum shows that bands characteristic of
TBA had disappeared, as confirmed by 1H NMR spectra of the
sample dissolved in NaOH/D2O. In contrast, reported results
show that ethoxy ligands formed from ethanol on Zr6O8 nodes
were quite stable under vacuum at 150 °C, and that their
formation was accompanied by the appearance of a new band
in the IR spectrum at 3781 cm-1, assigned to terminal OH groups
on Zr6O8 nodes.13-14 All the data are consistent with the
inference that these terminal OH groups were generated as
formate and acetate ligands were removed, and when the
alcohol was TBA, t-butyl formate (and t-butyl acetate)
desorbed, according to the chemistry shown in Scheme 3.
Comparison of MOFs with Metal Oxides as Catalysts.
We also compared the catalytic activities of the MOFs with
those of metal oxides, namely, γ-Al2O3 and ZrO2. However, by
comparison, each of these oxides had only a low activity at 150
°C (Figure 2). Reported data29-30 show that γ-Al2O3 is active for
TBA dehydration to give isobutylene but not ether at
temperatures higher than 170 °C. The comparison shows that
the catalytic sites on the MOF nodes are significantly different
from and intrinsically more active than those on these metal
oxides.
H
C
(CH3)3COH
150 o
HCOOC(CH3)3
OH
Zr
H
O
O
O
O
H
O
O
Zr
C
Zr
Zr
Scheme 3. Topology change of UiO-66 nodes resulting from
reaction of TBA with missing linker sites. Reaction between
TBA and acetate ligands is suggested to proceed equivalently
with CH3COOC(CH3)3 as a product.
6
ACS Paragon Plus Environment

Page 7 of 15
Journal of the American Chemical Society
Consistent with all the data, we further infer that TBA To provide more insight into the mechanism of the catalytic
1
2
3
4
5
6
7
8
dehydration is catalyzed by these terminal OH groups, with the
strongest basis for this inference being DFT calculations,
described below. This mechanism is contrasted to the
mechanism of ethanol dehydration on the nodes (which
proceeds via ethoxy ligands bonded to Lewis acid vacancies of
node defects13).
reaction, we did experiments to determine the kinetic isotopic
effect (KIE), using TBA-d10 as the reactant and UiO-66-320 as
the catalyst operating under conditions of minimal deactivation.
The conditions were chosen on the basis of vapor pressure
data33 to give the same partial pressures of TBA and TBA-d10
in the comparison experiments. The measured TOFH/TOFD
value is 2.0 (calculated from the respective differential
conversions of 6.3%/3.2%, Figure 5 and Figure S13, SI); below,
this value is compared with the value determined by density
functional theory.
This interpretation is bolstered by the catalytic activities of the
three samples (UiO-OCH3, UiO-OH, and UiO-320) without
formate and acetate ligands. These samples all had terminal OH
groups as initial ligands that were accessible to TBA—and these
account for the high initial activities of these samples.
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
Characterization of TBA Dehydration Mechanism by
Density Functional Theory (DFT).
Recall that UiO-OCH3 incorporated both OH and methoxy
ligands and that the latter are highly reactive and readily
removed with the formation of terminal OH groups by reaction
with water (formed by the dehydration of TBA). Thus, this is
the sample that had the highest density of terminal OH groups
among all those we investigated.
Formate removal from nodes through esterification with TBA.
To gain insight into the mechanisms of the esterification
reaction involving TBA and node-bound formate, we conducted
a DFT mechanistic investigation, modeling UiO-66 and MOF-
808 as described in Figure S15 in the SI.
UiO-66-320 was a highly dehydrated sample, meaning that its
defect sites were mostly dehydrated (Lewis acid) sites, as
shown in Scheme 2 (the species at the far right). Because the
activity of UiO-66-320 was much lower than that of UiO-66-
OCH3, we infer that these Lewis acid sites in the former sample
were inactive or barely active for the dehydration of TBA.
Reported results18 show that the rehydration of dehydrated
Zr6O8 nodes is very slow.
We started the investigation with UiO-66, and analyzed four
pathways for the esterification involving TBA and formate (but
calculations were not done for acetate—as we expect that the
electronic effects of the acetate group are so similar to those of
the formate group that the results presented here can be
extended to the esterification involving TBA and acetate). We
analyzed four separate pathways, involving one or two TBA
molecules in the mechanisms. In the following we show the free
energy diagram for the pathway having the lowest activation
energy among the four analyzed (path 1, Figure 7).
Again, we used in-situ IR spectroscopy, monitoring the changes
in UiO-66-OCH3 and UiO-66-OH during TBA dehydration
catalysis. The data show that the IR bands of terminal OH
groups at 3780 cm-1 and of methoxy ligands at 2826 and 2929
cm-1 all disappeared after contact with TBA at 150 °C;
correspondingly, IR bands characterizing node vacancy-
bonded TBA-derived species appeared at 2975, 1201, and 1037
cm-1 (Figure S4-S5, SI). We infer from the aforementioned data,
including the catalyst performance data, that methoxy ligands
were removed during catalysis and replaced by terminal OH
groups. We emphasize that the TBA-derived species observed
here were not physisorbed TBA, because these species were not
removed by purging with helium at 150 °C and, further, because
the intensity of the 2975 cm-1 band is correlated with the
catalytic activities of UiO-66, UiO-66-OH, and UiO-66-OCH3.
Because TBA did not form t-butoxy ligands, we suggest that it
likely bonds to the open Zr sites shown in Scheme 3 and
interacted with terminal OH groups on neighboring Zr sites,
likely forming hydrogen bonds with them.
The reaction starts (Figure 7) with UiO-66 having a formate
group bonded to Zr atoms and two TBA molecules at infinite
distance. The formate groups are labile ligands, and one Zr
bonding site can be opened up for binding with one TBA
molecule (highlighted in red, structure B). This structure locates
at 3.1 kcal/mol. The proton of the alcohol now becomes part of
a more strongly acidic group (because TBA is interacting with
Zr, a Lewis acid site), and it can protonate the formate to go to
structure C (through TS BC at 30.1 kcal/mol). Then the OR
group (with R being t-butyl) can act as a nucleophile and attack
the C atoms of the formate to produce structure D, with a
tetrahedral coordination around the carbon atom of the formate.
The TS CD is located at 35.7 kcal/mol. We then invoke a
second TBA molecule that binds to Zr (highlighted in red in
structure E), having found that this binding is energetically
favorable, lowering the free energy of structure D by 5.4
kcal/mol. The next step is the removal of the OH group bonded
to carbon by making use of the acidity of the µ3-OH group,
forming water and t-butyl formate (structure F). The transition
state EF lies at 40.3 kcal/mol, and it is characterized by the
highest activation energy of the whole mechanism—that is, the
reaction from E to F is the rate determining step (RDS). The
subsequent removal of t-butyl formate and binding of water to
Zr lower the free energy, forming structure G. Then the
protonation of the µ3-O group (in structure G) by the adjacent
water molecule occurs through TS GH, and structure H is
formed, in which the TBA and an OH group are bonded to the
Zr atoms. The installation of an OH group bonded to Zr is
Because the initial activity of UiO-66 was low (Table 1), we
infer that the four µ3-OH groups that were initially present on
each Zr6O8 node (each node should nominally be expected to be
[Zr6(µ3-O)4(µ3-OH)4]12+) lack significant activity for the TBA
dehydration reaction. These hydroxyl groups are, we infer,
isolated by linkers and lacking neighboring Lewis acid sites, so
that they are not effective by themselves for catalyzing the
dehydration reaction.
Kinetic isotope effect in MOF-catalyzed TBA dehydration
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 15
consistent with the experimental findings, and this is the initial group) are reported. From the PBC optimization of structure E
1
2
3
4
5
6
7
8
structure for the subsequent TBA dehydration catalysis.
we did not notice a major change in lattice parameters (using
UiO-66 with one missing linker as a reference), with a slight
contraction of about 0.5 Å (out of 25.5 Å) along the a and c axes
and a small change (less than one degree) in angles. The slight
contraction of the lattice parameters results in a shrinking of the
unit cell volume by only about 2%, which we infer is probably
a consequence of dispersion interactions between structure E
and the linkers. According to these calculations, we can safely
conclude that structure E is a plausible intermediate for the
esterification mechanism and it can well fit into the pore of the
UiO-66 MOF.
We then questioned whether the steric hindrance of two t-butyl
groups would preclude the formation of structure E. In other
words, we wanted to check whether structure E could fit into
the pore of UiO-66. To achieve this, we recurred to geometry
optimization of structure E using periodic boundary conditions
(PBC, see SI Section S2.2), which takes into account the
periodic structure of the MOF and therefore the dimensions of
the pores. In Table S1 in the SI, the optimized lattice parameters
(lengths, angles, and unit cell volume) for structure E and UiO-
66 with one missing linker (capped with a water and an OH
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 7: Free energy reaction profile for the esterification reaction between formate and TBA for UiO-66, path 1. The last structure contains
a terminal OH group, as inferred from the experimental results. TBA = t-butyl alcohol, TBF = t-butyl formate, R = t-butyl. The alcohol
molecules bonded to Zr and the OH group that is formed after esterification are highlighted with red circles.
After examining the mechanism characterized by the lowest
activation energy, we briefly discuss the other three pathways,
which are characterized by higher activation energies (the free
energy diagrams are shown in the SI, Figure S16–S18). In path
2 (Figure S16) the mechanism is the same as the one shown in
Figure 7, except that only one TBA molecule is invoked in the
reaction. Therefore, structures from A to D are the same as in
Figure 7, but then the protonation of the OH group by the H of
the µ3-OH (TS DE in Figure S16) now lies at 45.2 kcal/mol,
about 5 kcal/mol higher than the RDS activation energy shown
in Figure 7.
proceeds through a RDS transition state similar to the one
shown in Figure S18, in which the H of the alcohol now binds
to the µ3-O group (instead of the oxygen of the formate). This
transition state locates at a very high energy, namely, 49.1
kcal/mol. We therefore have not continued the computation of
this mechanism.
We then computed path 1 for MOF-808, to determine whether
there was any significant difference between the two MOFs,
and the results are shown in Figure S19. From these results we
see that the RDS for MOF-MOF-808 is 39.5 kcal/mol, which is
very similar to what we computed for UiO-66.
The other two analyzed pathways do not involve the binding of
TBA to the Zr atoms. Path 3 is shown in Figure S17 in the SI,
with the RDS transition state in which the oxygen atom of the
alcohol attacks the carbon of the formate and the hydrogen of
the alcohol binds to one of the oxygen atoms of the formate, in
a concerted way (TS BC in figure S17). This transition state
locates at 50.9, about 10 kcal/mol higher than the RDS
activation energy for the path in Figure 7. The pathway then
proceeds with the formation of a tetrahedral structure around
the carbon atom of the formate (structure C), in which then the
C–OH bond easily breaks to form t-butyl formate and an OH
group bonded to Zr, going down in free energy (structure D).
The last analyzed path is shown in Figure S18 in the SI and
In summary, DFT shows that removal of the formate groups via
esterification with TBA is characterized by lower activation
energies if the binding of two TBA molecules to Zr atoms is
invoked. Pathways that do not involve this binding (or only one
TBA molecule binding to Zr) proceed through higher energy
RDSs. The lowest RDS activation energies we could find are
40.4 kcal/mol for UiO-66 and 39.5 kcal/mol for MOF-808,
which are quite high, and corroborate the difficulty in removing
formate groups through esterification with TBA, as was
observed experimentally.
8
ACS Paragon Plus Environment

Page 9 of 15
Journal of the American Chemical Society
first characterizing the formation of the carbocation and the
1
2
3
4
5
6
7
8
second characterizing the C–H bond activation. In the
mechanism we propose in Figure 8, only one transition state has
been located (consisting of a concerted breaking of the C–O and
C–H bonds of the alcohol) and all attempts to stabilize a
carbocation intermediate were not successful (that is, the
structure would optimize either to reactant or to product). This
result is probably an indication that the reaction takes place with
gas-phase reactants, and that no solvent is present to stabilize a
positive charge. Nevertheless, the highly stretched C–O bond of
the alcohol in the TS indicates that a carbocation is formed, and
therefore the mechanism is more consistent with E1 than E2.
9
10
11
12
13
14
We then computed the kinetic isotope effect (KIE) and obtained
a KIE of 2.4, in good agreement with experiment (KIE = 2.0),
further corroborating the mechanism we propose.
Figure 8: Free energy reaction profile for the TBA dehydration
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
We also computed structure A and TS AB for MOF-808 and
found an activation energy of 38.2 kcal/mol, which is 3.5
kcal/mol higher than the value computed for UiO-66. The TS
AB geometry is very similar to what we found for UiO-66, and
the KIE is 2.4 (the same value as for UiO-66). According to
these results, we can conclude that the TBA dehydration
mechanism is the same for the two MOFs—providing further
confirmation that the higher initial activity of MOF-808 is
associated with the higher availability of OH groups and more
rapid transport of the TBA molecules in this MOF.
reaction to isobutylene and water for UiO-66. TBA = t-butyl
alcohol, R = t-butyl. The OH group that acts as a Brønsted base and
the water formed on the node are highlighted with red circles.
TBA dehydration. To elucidate the TBA dehydration
mechanism and the role of the terminal OH group in the
catalysis, we carried out a DFT mechanistic investigation. To
compare our results with the reported computational
investigation of ethanol dehydration on UiO-66, we used the
same model and computational details as before.13 The
description of this model and computational details are reported
in the SI, Figure S20 and Section S2.4, respectively. The
mechanism we found for TBA dehydration reaction for UiO-66
is shown in Figure 8.
DISCUSSION
Comparison of MOFs and metal oxides as catalysts. The MOFs
investigated here and metal oxides such as alumina and zirconia
used as catalysts for TBA dehydration are similar to each other
in presenting OH groups and nearby surface Lewis acid sites
(exposed Al and Zr cations, respectively). The mechanism of
the TBA dehydration reaction on a mixed oxide, silica-alumina,
has been inferred on the basis of experimental reaction kinetics
to proceed via a carbocation intermediate, which is formed in
the reaction between an acidic site with the hydroxyl group of
TBA, further neutralizing a nearby basic site, with the
intermediate subsequently decomposing and yielding
isobutylene,34 consistent with the mechanism shown by DFT to
occur on the MOF nodes.
The catalysis starts with the OH group and TBA bonded to Zr
and an isolated TBA molecule (structure A). The reaction then
proceeds through the transition state (TS AB) located at 34.7
kcal/mol, which produces structure B that lies at 4.7 kcal/mol.
In the latter, water and an OH group (the leaving group from
TBA) are bonded to Zr, and isobutylene is formed, which is still
interacting with the node. Finally, to restore the initial structure
and close the catalytic cycle, the removal of isobutylene and
water and the bonding of a TBA molecule to Zr occur (structure
C). The overall reaction is exothermic, with structure C locating
at -3.6 kcal/mol. An important point is that the barrier for this
reaction is 4.0 kcal/mol lower than the one reported for ethanol
dehydration.13 This result is in agreement with the higher TOF
for the TBA dehydration (see above, experimental results).
We emphasize the differences between the MOFs and the metal
oxides: in contrast to amorphous metal oxide with their smears
of heterogeneous surface sites, the MOFs offer sites in
crystalline environments that can be characterized precisely for
in-depth understanding of reaction mechanism guided by DFT.
In the following, the transition state geometry (TS AB in Figure
8) is shown (Figure 9). The analysis of this geometry can
provide insights into the mechanism of the TBA dehydration.
In Figure 9 it is noted that the carbon–oxygen bond of the
alcohol is highly stretched (2.77 Å), indicating that a tertiary
carbocation is formed. The C–H bond that is being broken (blue
number) is 1.20 Å, which is not highly stretched compared with
another C–H bond of the alcohol that does not participate in the
reaction (1.10 Å, red number). The O–H bond (to form water)
is not yet fully formed (1.53 Å in Figure 9). According to this
geometry, we can conclude that the mechanism of the TBA
dehydration is consistent with an E1 elimination. However, the
E1 mechanism consists of two different transition states, the
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 15
Figure 9: Transition state geometry for the t-butyl alcohol
OH group on a Zr6O8 node is such a weak acid that it plays a
1
2
3
4
5
6
dehydration reaction to isobutylene and water for UiO-66 (TS AB
in Figure 8). All linkers are omitted for clarity. Atom color code:
gray for carbon, red for oxygen, white for hydrogen, light blue for
zirconium. Bond length color code: black for carbon–oxygen bond,
green for oxygen–hydrogen bond, blue for carbon–hydrogen bond
(that is being broken), and red for another carbon–hydrogen bond
(that is not participating in the reaction). All distances in Å.
role best described as that of a Brønsted base in TBA
dehydration. In contrast, strong acids such as sulfonic acid
resins (or p-toluene sulfonic acid) catalyze TBA dehydration at
temperatures much lower than what we used (e.g., 70 °C), with
high rates.24
Transport limitations. We reemphasize the importance of being
aware of possible transport limitations in MOF catalysis. MOF
pores are comparable to those in zeolites, which not only
provide confined environment for reactions, but also limit the
transport of reactions. Transport limitations have been reported
for MOF-catalyzed reactions, for example, for reactions
catalyzed by MOF-encapsulated palladium or platinum
nanoparticles.37-38 The reported results show that reaction rates
decreased markedly as reactant size increased. However, all
these results, like those determined in our work, are qualitative,
falling short of determining quantitative measures of the
transport rates and intrinsic catalytic reaction rates.
7
8
9
Thus, an essential advantage of the MOFs is the opportunities
they offer for fundamental understanding of catalysis on sites
that are representative of a large class of materials—metal
oxides—applied on an enormous scale in catalytic technology.
But we emphasize the subtlety of the MOF catalytic sites and
the need to understand their chemistry beyond the level of bulk
structure and the evidence available from XRD and BET
measurements. There is a need to understand the genesis of the
catalytically active sites and to understand minor components
in the MOFs and specifically the important role of defect sites
on the MOF nodes. Understanding of the chemistry in such
depth requires identification and control of these groups (in
addition to linkers) that may be bonded to the nodes.
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
Our results indicate that the reactions between TBA and formate
and acetate ligands on the nodes to form esters are strongly
limited by the small pore size of UiO-66 (the data are consistent
with the inference that the esters are too large to be transported
readily through the pores of UiO-66—and this transport
limitation minimizes the catalytic activity of UiO-66. However,
after removal of the formate and acetate ligands (either by
reaction with methanol or by high-temperature dehydration),
the number of accessible active sites (defects with terminal OH
groups) increases markedly as they become accessible to TBA
for dehydration to form isobutylene. We infer that the pore
aperture diameter should be regarded as larger than the nominal
(crystallographically determined) value of 6 Å for the ideal
MOF, because of the missing linkers and nodes; consequently,
the limitations of transport of TBA and isobutylene in the pores
of UiO-66 are less than would be expected on just the basis of
the crystallographic data for the ideal MOF. There is much to
learn yet about transport/reaction tradeoffs in MOF catalysis.
The as-synthesized MOFs incorporated formate and acetate
ligands that inhibit TBA dehydration catalysis, and the catalysts
had to be “broken in” by removal of these groups—by reaction
with TBA to make esters, thereby forming Lewis acid sites and
terminal OH groups that are responsible for the dehydration
catalysis. With the benefit of DFT calculations, we see how the
terminal OH groups act as catalytic sites for TBA dehydration
by collaborating with neighboring Lewis acid sites. We
anticipate similar mechanisms on metal oxide catalysts and
posit that understanding of catalytic mechanisms on MOFs,
such as that emerging from this work, will help with
understanding of catalytic mechanisms metal oxides.
Comparing mechanisms of dehydration of ethanol and TBA.
We have reported13 that the active sites of UiO-66 for ethanol
dehydration to diethyl ether are two adjacent defect sites, with
a calculated transition state of 38.7 kcal/mol for the reaction
proceeding through an SN2 mechanism. For TBA dehydration,
in contrast, the calculations show the need for only one defect,
and a transition state energy of 34.7 kcal/mol for reaction
proceeding through an E1 mechanism. All these results are
consistent with experiment, specifically, with the TBA
dehydration reaction (taking place at 150 °C) being faster than
the ethanol dehydration (taking place at 200 °C).13
Stabilities. We further emphasize the importance of being
aware of potential MOF stability limitations in investigations of
MOF catalysis, both because MOF stabilities are important
criteria for potential application and because MOF stabilities
may depend in complex ways on their environments. MOF
stabilities are potentially influenced by numbers of node-linker
bonds, node-linker bond strengths, mechanisms of node-linker
bond breaking, properties of node center elements, etc.39 Our
data provide evidence showing that the number of defects is
also crucial for the stabilities of our MOFs. MOF-808, among
those known that incorporate Zr6O8 nodes, is the one with the
greatest number of node vacancies. This might at first be
considered an advantage, implying high catalytic activity, but it
is also a major disadvantage, implying a low stability compared
with UiO-66, which has a much lower number of defect sites
per node, at least under our alcohol dehydration conditions. One
should be aware of potential tradeoffs related to reactivity and
stability of MOFs.
Because TBA dehydration is catalyzed by a site with both a
Lewis acid site and a terminal OH group, it is highly sensitive
to the density of terminal OH groups. Tuning the ligands on
node defect sites by a range of post-treatment methods has been
shown here to significantly influence the catalytic activity for
TBA dehydration on UiO-66. In contrast, the effect of changing
the ligand has been shown not to be significant for ethanol
dehydration.13 It is clear that these two alcohol dehydration
reactions are valuable probes of the reactivity and catalytic
properties of MOF nodes.
CONCLUSIONS
Acid strength and catalysis. Many results have shown that
hydroxyl groups on MOF nodes are weak acids.35-36 Our data
reaffirm the conclusion, and further demonstrate that a terminal
10
ACS Paragon Plus Environment

Page 11 of 15
Journal of the American Chemical Society
Terminal OH groups on defect sites of Zr6O8 nodes of MOF- at 200 °C and 1 bar for 48 h. The MOF sample (200 mg) was
1
2
3
4
5
6
7
8
808 and UiO-66 catalyze the dehydration of TBA, with the rates
depending on ligands on defect sites and intrapore transport
limitations. Because of their well-defined structures, the MOF
nodes are excellent platforms for fundamental understanding of
catalytic sites and reaction mechanisms related to those on
metal oxides. MOF-808 and UiO-66, which have
approximately 6 and 1 vacancies per Zr6O8 node, respectively,
were compared with alumina and zirconia as catalysts for TBA
dehydration. The greater number of vacancies on MOF-808
nodes accounts for its initially higher activity than that of UiO-
66—but these vacancies also limit the stability of this MOF as
a catalyst. DFT calculations show that the OH groups on UiO-
66 and MOF-808 act as Brønsted basic sites, and they are
generated by the removal of formate/acetate ligands initially
present on the nodes via esterification reactions with the
alcohol. The calculations indicate an E1 elimination reaction,
with an activation barrier for TBA dehydration on UiO-66 that
is 4 kcal/mol lower than that reported for ethanol dehydration
on the same MOF, consistent with higher observed reactivity of
TBA. The computational results further suggest that the binding
of two TBA molecules to node Zr atoms during the
esterification reaction helps in lowering the total activation
energy of this reaction.
loaded into the reactor in an argon-filled glovebox. Methanol
was carried into the reactor in a stream of helium carrier gas that
flowed through a methanol-filled bubbler with the temperature
controlled at 25 °C. The sample is designated as UiO-66-OCH3.
Post-treatment of UiO-66 with water vapor. The UiO-66-
OCH3 sample was further treated with water vapor in the flow
reactor at 150 °C and 1 bar for 10 h. The MOF sample (200 mg)
was loaded into the reactor in an argon-filled glovebox. Water
was carried into the reactor in a stream of helium carrier gas that
flowed through a water-filled bubbler with temperature
controlled at 25 °C. The sample is designated as UiO-66-OH.
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
Post-treatment of UiO-66 at 320 °C. UiO-66 was dehydrated
at 320 °C under high vacuum for 12 h. The sample is designated
as UiO-66-320.
Infrared spectroscopy. A Bruker IFS 66v/S spectrometer with
a spectral resolution of 2 cm−1 was used to collect transmission
IR spectra of powder samples. Approximately 30 mg of solid
sample in an argon-filled glove box was pressed into a thin
wafer and loaded into a cell that served a flow reactor (In-situ
Research Institute, Inc., South Bend, IN). The cell was sealed
and connected to a flow system that allowed recording of
spectra while the reactant gases flowed through the cell at
reaction temperature. Each spectrum is the average of 64 scans.
EXPERIMENTAL AND COMPUTATIONAL METHODS
BET surface area measurements. All N2 isotherms were
measured with a Micromeritics 3Flex Surface Characterization
Analyzer instrument. Measurements were performed at -196 oC,
with the temperature held constant with a liquid nitrogen bath.
Consistency criteria were adapted to choose an appropriate
pressure range for BET surface area calculations.
Synthesis of UiO-66 with acetic acid modulator. ZrCl4 (0.080
g, 0.343 mmol) and 0.7 mL of acetic acid (modulator) were
dissolved in 20 mL of DMF in a 8-dram vial by using ultrasound
for 5 min. The linker precursor benzene-1,4-dicarboxylic acid
(H2BDC, 0.057 g, 0.343 mmol) was then added into the solution
and dissolved by ultrasound for about 5 min. The vials were
kept in an oven at 120 °C under static conditions for 24 h. White
precipitates were produced, and they were isolated by
centrifugation after cooling to room temperature. The solids
were washed with DMF three times to remove unreacted
precursors and with acetone six times to remove DMF. Then the
powder was dried at room temperature and activated at 120 °C
for 12 h under high vacuum.
1H NMR spectroscopy. Samples were prepared by weighing
10 mg of MOF into a 1.5 ml vial. A 600 μL portion of the
digestion medium was then added to the tube. The digestion
medium was 1-M NaOH in D2O. The vials were capped and
inverted 2–3 times before allowing the samples to digest over a
period of 24 h. This OH-based procedure dissolves only the
organic portion of the MOF (linker, modulator, solvent, etc.),
while the inorganic component drops out as ZrO2. After 24 h,
the clear supernatant solution was transferred to an NMR tube.
Synthesis of MOF-808 with formic acid as modulator.
ZrOCl2·8H2O (0.253 g, 0.785 mmol) and 10 mL of formic acid
(88 wt% in water, 0.156 mol) were dissolved in 10 mL of DMF
in an 8 dram vial by using ultrasound for 10 min. The linker
precursor benzene-1,3,5-tricarboxylic acid (H3BTC, 55 mg,
0.262 mmol) was then added into the solution and dissolved by
ultrasound for about 10 min. The vials were kept in an oven at
120 °C under static conditions for 48 h. White precipitates were
produced, and they were isolated by centrifugation after cooling
to room temperature. The solids were washed with DMF three
times to remove unreacted precursors and with acetone six
times to remove DMF. Then the powder was dried at room
temperature and activated at 120 °C for 12 h under high
vacuum.
1
Liquid H NMR spectra were recorded with a Bruker Avance
DPX-500 NMR Spectrometer (500 MHz). The relaxation delay
(d1) was set to 20 s to ensure that reliable integrals were
obtained, allowing for the accurate determination of the relative
concentrations of the molecular components. The number of
scans was 32.
Powder X-ray Diffraction (PXRD). PXRD patterns were
obtained using a Panlytical X-ray Diffractometer Model X’Pert
Pro MRD. Measurements were made over a range of 2° < 2θ <
20° in 0.05 step size at a scanning rate of 1 deg/min.
Catalytic Reaction Experiments. TBA dehydration catalysis
was carried out in a conventional laboratory once-through
tubular plug-flow reactor at °C and 1 bar. Particles of the MOF
catalyst (25–100 mg) were mixed with 2 g of particles of inert,
nonporous α-Al2O3 and loaded into the reactor in an argon-filled
glove box. The feed partial pressures were 75 mbar of TBA and
Post-treatment of UiO-66 with methanol vapor. Post-
treatment of UiO-66 with methanol vapor was carried out in a
conventional laboratory once-through tubular plug-flow reactor
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 15
925 mbar of helium, with a total flow rate of 5.4 mL(NTP)/min
Dong Yang: 0000-0002-3109-0964
1
2
3
4
5
6
7
8
(TBA was carried into the reactor with helium flowing through
a bubbler, and all of the gas lines downstream of the bubbler
were held at 120 °C to avoid condensation of the TBA and
reaction products). Products were analyzed with an on-line
Agilent 6890 gas chromatograph. The conversions to
isobutylene were typically < 10%, and the reactor was well
approximated as differential.
Carlo Alberto Gaggioli: 0000-0001-9105-8731
Melike Babucci: 0000-0001-7785-3755
Laura Gagliardi: 0000-0001-5227-1396
Bruce C. Gates: 0000-0003-0274-4882
REFERENCES
(1) Ferey, G., Hybrid Porous Solids: Past, Present, Future.
Chem. Soc. Rev. 2008, 37, 191-214.
(2) Horike, S.; Shimomura, S.; Kitagawa, S., Soft Porous
Crystals. Nat. Chem. 2009, 1, 695-704.
(3) Furukawa, H.; Cordova, K. E.; O'Keeffe, M.; Yaghi, O. M.,
The Chemistry and Applications of Metal-Organic
Frameworks. Science 2013, 341, 1230444.
(4) Jiang, J. C.; Yaghi, O. M., Bronsted Acidity in Metal-
Organic Frameworks. Chem. Rev. 2015, 115, 6966-6997.
(5) Dhakshinamoorthy, A.; Li, Z.; Garcia, H., Catalysis and
Photocatalysis by Metal Organic Frameworks. Chem. Soc. Rev.
2018, 47, 8134-8172.
(6) Pascanu, V.; Miera, G. G.; Inge, A. K.; Martin-Matute, B.,
Metal-Organic Frameworks as Catalysts for Organic Synthesis:
A Critical Perspective. J. Am. Chem. Soc. 2019, 141, 7223-
7234.
(7) Yang, D.; Gates, B. C., Catalysis by Metal Organic
Frameworks: Perspective and Suggestions for Future Research.
ACS Catal. 2019, 9, 1779-1798.
(8) Wang, Q.; Astruc, D., State of the Art and Prospects in
Metal–Organic Framework (MOF)-Based and MOF-Derived
Nanocatalysis. Chem. Rev. 2020, 120, 1438-1511.
Computational methods. DFT geometry optimizations and
frequency calculations were performed using the M06-L local
density functional40 and the Gaussian 0941 software package.
The def2-SVP basis set was employed for H, C, and O atoms,
and the def2-TZVPP basis set42-43 was employed for the Zr
atoms. Only for the TBA dehydration reaction, in order to
compare the results with a previous study on ethanol
dehydration, same computational details as previously
reported13 were used (see SI, section S2.5). The SDD
pseudopotential44 was employed to describe the core electrons
in Zr. An ultrafine grid was used for performing numerical
integrations. The nature of the stationary points (minima or
transition states) along the potential energy surfaces was
verified through the computation of vibrational frequencies,
which were also employed to calculate enthalpic and entropic
contributions at 150 °C. The frequencies below 50 cm-1 were
replaced by 50 cm-1 when computing all the free energies
reported in this work. Computational details for periodic
boundary condition calculations are reported in Section S2.6 of
the SI.
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
(9) Cavka, J. H.; Jakobsen, S.; Olsbye, U.; Guillou, N.;
Lamberti, C.; Bordiga, S.; Lillerud, K. P., A New Zirconium
Inorganic Building Brick Forming Metal Organic Frameworks
with Exceptional Stability. J. Am. Chem. Soc. 2008, 130,
13850-13851.
ASSOCIATED CONTENT
Supporting Information. Experimental and computational details
1
are reported, primarily IR spectra, catalytic activities, H NMR
spectra, desorption curves, SEM images and Cartesian coordinates
with its associated energies. This material is available free of
charge via the Internet at http://pubs.acs.org.
(10) Mondloch, J. E.; Bury, W.; Fairen-Jimenez, D.; Kwon, S.;
DeMarco, E. J.; Weston, M. H.; Sarjeant, A. A.; Nguyen, S. T.;
Stair, P. C.; Snurr, R. Q.; Farha, O. K.; Hupp, J. T., Vapor-Phase
Metalation by Atomic Layer Deposition in a Metal-Organic
Framework. J. Am. Chem. Soc. 2013, 135, 10294–10297.
(11) Vermoortele, F.; Bueken, B.; Le Bars, G.; Van de Voorde,
B.; Vandichel, M.; Houthoofd, K.; Vimont, A.; Daturi, M.;
Waroquier, M.; Van Speybroeck, V.; Kirschhock, C.; De Vos,
D. E., Synthesis Modulation as a Tool To Increase the Catalytic
Activity of Metal-Organic Frameworks: The Unique Case of
UiO-66(Zr). J. Am. Chem. Soc. 2013, 135, 11465-11468.
(12) Shearer, G. C.; Chavan, S.; Bordiga, S.; Svelle, S.; Olsbye,
U.; Lillerud, K. P., Defect Engineering: Tuning the Porosity and
Composition of the Metal-Organic Framework UiO-66 via
Modulated Synthesis. Chem. Mater. 2016, 28, 3749-3761.
(13) Yang, D.; Ortuño, M. A.; Bernales, V.; Cramer, C. J.;
Gagliardi, L.; Gates, B. C., Structure and Dynamics of Zr6O8
Metal-Organic Framework Node Surfaces Probed with Ethanol
Dehydration as a Catalytic Test Reaction. J. Am. Chem. Soc.
2018, 140, 3751-3759.
AUTHOR INFORMATION
Corresponding Authors
*bcgates@ucdavis.edu (B.C.G.)
*gagliard@umn.edu (L.G.)
Notes
The authors declare no competing financial interest.
Author Contributions
† These authors contributed equally to this work
ACKNOWLEDGMENTS
This work was supported as part of the Inorganometallic Catalyst
Design Center, an Energy Frontier Research Center funded by the
U.S. Department of Energy (DOE), Office of Science, Basic
Energy Sciences (DE-SC0012702). We thank Dr. Ping Yu of the
NMR facility at the University of California, Davis, for help with
the 1H NMR measurements.
(14) Wei, R.; Gaggioli, C. A.; Li, G.; Islamoglu, T.; Zhang, Z.;
Yu, P.; Farha, O. K.; Cramer, C. J.; Gagliardi, L.; Yang, D.;
Gates, B. C., Tuning the Properties of Zr6O8 Nodes in the Metal
Organic Framework UiO-66 by Selection of Node-Bound
Ligands and Linkers. Chem. Mater. 2019, 31, 1655-1663.
(15) Wu, H.; Chua, Y. S.; Krungleviciute, V.; Tyagi, M.; Chen,
P.; Yildirim, T.; Zhou, W., Unusual and Highly Tunable
ORCID
12
ACS Paragon Plus Environment

Page 13 of 15
Journal of the American Chemical Society
Missing-Linker Defects in Zirconium Metal-Organic Metal–Organic Framework hcp UiO-66. ACS Catal. 2020, 10,
1
2
3
4
5
6
7
8
Framework UiO-66 and Their Important Effects on Gas
Adsorption. J. Am. Chem. Soc. 2013, 135, 10525-10532.
(16) Jiang, J. C.; Gandara, F.; Zhang, Y. B.; Na, K.; Yaghi, O.
M.; Klemperer, W. G., Superacidity in Sulfated Metal-Organic
Framework-808. J. Am. Chem. Soc. 2014, 136, 12844-12847.
(17) Trickett, C. A.; Popp, T. M. O.; Su, J.; Yan, C.; Weisberg,
J.; Huq, A.; Urban, P.; Jiang, J.; Kalmutzki, M. J.; Liu, Q.; Baek,
J.; Head-Gordon, M. P.; Somorjai, G. A.; Reimer, J. A.; Yaghi,
O. M., Identification of the Strong Brønsted Acid Site in a
Metal–Organic Framework Solid Acid Catalyst. Nat. Chem.
2019, 11, 170–176,
2906-2914.
(29) Shi, B. C.; Dabbagh, H. A.; Davis, B. H., Catalytic
Dehydration of Alcohols. Kinetic Isotope Effect for the
Dehydration of t-Butanol. Top. Catal. 2002, 18, 259-264.
(30) Jai, J. R.; Pillai, C. N., Catalytic Dehydration of Alcohols
over Alumina: Mechanism of Ether Formation. J. Catal. 1964,
9, 322-330.
(31) Kipkemboi, P. K.; Kiprono, P. C.; Sanga, J. J., Vibrational
Spectra of t-Butyl Alcohol, t-Butylamine and t-Butyl Alcohol
plus t-Butylamine Binary Liquid Mixtures. B. Chem. Soc.
Ethiopia 2003, 17, 211-218.
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
(18) Yang, D.; Bernales, V.; Islamoglu, T.; Farha, O. K.; Hupp,
J. T.; Cramer, C. J.; Gagliardi, L.; Gates, B. C., Tuning the
Surface Chemistry of Metal Organic Framework Nodes: Proton
Topology of the Metal-Oxide-Like Zr6 Nodes of UiO-66 and
NU-1000. J. Am. Chem. Soc. 2016, 138, 15189-15196.
(19) Caratelli, C.; Hajek, J.; Cirujano, F. G.; Waroquier, M.;
Xamena, F. X. L. I.; Van Speybroeck, V., Nature of Active Sites
on UiO-66 and Beneficial Influence of Water in the Catalysis
of Fischer Esterification. J. Catal. 2017, 352, 401-414.
(20) Mondloch, J. E.; Katz, M. J.; Isley, W. C.; Ghosh, P.; Liao,
P. L.; Bury, W.; Wagner, G.; Hall, M. G.; DeCoste, J. B.;
Peterson, G. W.; Snurr, R. Q.; Cramer, C. J.; Hupp, J. T.; Farha,
O. K., Destruction of Chemical Warfare Agents Using Metal-
Organic Frameworks. Nat. Mater. 2015, 14, 512-516.
(21) Hoffman, A. E. J.; Vanduyfhuys, L.; Nevjestic, I.; Wieme,
J.; Rogge, S. M. J.; Depauw, H.; Van der Voort, P.; Vrielinck,
H.; Van Speybroeck, V., Elucidating the Vibrational
Fingerprint of the Flexible Metal-Organic Framework MIL-
(32) Lynch, K, C. T.; Mazdiyasni, S.; Smith, J. S.; Crawford,
W. J., Infrared Spectra of Transition Metal Alkoxides. Anal.
Chem. 1964, 36, 2332-2337.
(33) Kiss, I.; Jákli, G.; Jancsó, G.; Illy, H., Vapor Pressures of
Some Deuterated Alcohols and Amines. J. Chem. Phys. 1967,
47, 4851-4852.
(34) Figueras, F.; Nohl, A.; Mourgueas, L. D.; Trambouze, Y.,
Dehydration of Methanol and tert-Butyl Alcohol on Silica-
Alumina. T. Faraday Soc. 1971, 67, 1155-1163.
(35) Driscoll, D. M.; Troya, D.; Usov, P. M.; Maynes, A. J.;
Morris, A. J.; Morris, J. R., Characterization of
Undercoordinated Zr Defect Sites in UiO-66 with Vibrational
Spectroscopy of Adsorbed CO. J. Phys. Chem. C 2018, 122,
14582-14589.
(36) Volkringer, C.; Leclerc, H.; Lavalley, J. C.; Loiseau, T.;
Ferey, G.; Daturi, M.; Vimont, A., Infrared Spectroscopy
Investigation of the Acid Sites in the Metal-Organic Framework
Aluminum Trimesate MIL-100(Al). J. Phys. Chem. C 2012,
116, 5710-5719.
53(Al) Using
a Combined Experimental/Computational
Approach. J. Phys. Chem. C 2018, 122, 2734-2746.
(22) Benzaqui, M.; Pillai, R. S.; Sabetghadam, A.; Benoit, V.;
Normand, P.; Marrot, J.; Menguy, N.; Montero, D.; Shepard,
W.; Tissot, A.; Martineau-Corcos, C.; Sicard, C.; Mihaylov, M.;
Carn, F.; Beurroies, I.; Llewellyn, P. L.; De Weireld, G.;
Hadjiivanov, K.; Gascon, J.; Kapteijn, F.; Maurin, G.; Steunou,
N.; Serre, C., Revisiting the Aluminum Trimesate-Based MOF
(MIL-96): From Structure Determination to the Processing of
Mixed Matrix Membranes for CO2 Capture. Chem. Mater.
2017, 29, 10326-10338.
(23) Volkringer, C.; Leclerc, H.; Lavalley, J.-C.; Loiseau, T.;
Ferey, G.; Daturi, M.; Vimont, A., Infrared Spectroscopy
Investigation of the Acid Sites in the Metal-Organic Framework
Aluminum Trimesate MIL-100(Al). J. Phys. Chem. C 2012,
116, 5710-5719.
(24) Gates, B.C.; Wisnouskas, J. S.; Heath, H. W., Jr. The
Dehydration of t-Butyl Alcohol Catalyzed by Sulfonic Acid
Resin. J. Catal. 1972, 24, 320-327.
(25) Frilette, V. J.; Mower, E. B.; Rubin, M. K., Kinetics of
Dehydration of tert-Butyl Alcohol Catalyzed by Ion Exchange
Resins. J. Catal. 1964, 3, 25-31.
(26) Furukawa, H.; Gandara, F.; Zhang, Y. B.; Jiang, J. C.;
Queen, W. L.; Hudson, M. R.; Yaghi, O. M., Water Adsorption
in Porous Metal-Organic Frameworks and Related Materials. J.
Am. Chem. Soc. 2014, 136, 4369-4381.
(27) Stepanian, S. G.; Reva, I. D.; Radchenko, E. D.; Sheina, G.
G., Infrared Spectra of Benzoic Acid Monomers and Dimers in
Argon Matrix. Vib. Spectrosc. 1996, 11, 123-133.
(37) Wang, B. Q.; Liu, W. X.; Zhang, W. N.; Liu, J. F.,
Nanoparticles@nanoscale
Metal-Organic
Framework
Composites as Highly Efficient Heterogeneous Catalysts for
Size- and Shape-Selective Reactions. Nano Res. 2017, 10,
3826-3835.
(38) Zhang, W. N.; Lu, G.; Cui, C. L.; Liu, Y. Y.; Li, S. Z.; Yan,
W. J.; Xing, C.; Chi, Y. R.; Yang, Y. H.; Huo, F. W., A Family
of Metal-Organic Frameworks Exhibiting Size-Selective
Catalysis with Encapsulated Noble-Metal Nanoparticles. Adv.
Mater. 2014, 26, 4056-4060.
(39) Burtch, N. C.; Jasuja, H.; Walton, K. S., Water Stability
and Adsorption in Metal-Organic Frameworks. Chem. Rev.
2014, 114, 10575-10612.
(40) Zhao, Y.; Truhlar, D. G., A New Local Density Functional
for Main-Group Thermochemistry, Transition Metal Bonding,
Thermochemical Kinetics, and Noncovalent Interactions. J.
Chem. Phys. 2006, 125, 194101.
(41) 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, H. P.; 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. J. A.; Peralta, J. E.;
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.;
(28) Chen, X.; Lyu, Y. H.; Wang, Z. Y.; Qiao, X.; Gates, B. C.;
Yang, D., Tuning Zr12O22 Node Defects as Catalytic Sites in the
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 15
Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Quality for H to Rn: Design and Assessment of Accuracy. Phys.
1
2
3
4
5
6
7
8
Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;
Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian
09, Revision E.01; Gaussian, Inc.: Wallingford, CT, 2009.
(42) Weigend, F., Accurate Coulomb-Fitting Basis Sets for H
to Rn. Phys. Chem. Chem. Phys. 2006, 8, 1057-1065.
Chem. Chem. Phys. 2005, 7, 3297-3305.
(44) Hay, P. J.; Dunning Jr, T. H., Modern Theoretical
Chemistry, Methods of Electronic Structure Theory. 1976, 3.
(43) Weigend, F.; Ahlrichs, R., Balanced Basis Sets of Split
Valence, Triple Zeta Valence and Quadruple Zeta Valence
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
14
ACS Paragon Plus Environment

Page 15 of 15
Journal of the American Chemical Society
TOC
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
15
ACS Paragon Plus Environment