Demonstrating the Critical Role of Solvation in Supported Ti and Nb Epoxidation Catalysts via Vapor-Phase Kinetics

Article abstract of DOI:10.1021/acscatal.9b04906

Catalytic oxidation of hydrocarbons with hydrogen peroxide (H₂O₂) has been of the utmost importance for several decades. The vast majority of studies have been performed in the condensed phase, even though condensed phases introduce complex solvent effects and can promote the leaching of active sites. In response, we have built a custom reactor system to understand H₂O₂ activation and selective oxidation in the vapor-phase. In this report, we study the epoxidation of cyclohexene with H₂O₂ over four Lewis-acidic metal oxide catalysts: Ti and Nb grafted on SiO₂ and on the Zr based metal-organic framework, NU-1000. The M-SiO₂ materials are highly selective to the formation of epoxides and diols, as they can be in the condensed phase, while the NU-1000 based materials are far more prone to overoxidation to CO₂, which appears to be connected to their strong reactant adsorption. Apparent activation energies are calculated for all materials when operating in the same kinetic regime, and the heats of cyclohexene adsorption into their pores are then used to directly compare intrinsic enthalpies of activation in the vapor vs condensed phase for the M-SiO₂ catalysts. Nb-SiO₂ catalysts exhibit similar intrinsic enthalpies of activation in the vapor and condensed phases, whereas the condensed phase transition state in Ti-SiO₂ is 24 kJ/mol lower in energy than that of the same material in the vapor phase. These experiments establish another methodology for understanding the various roles of solvent in selective oxidation reactions and studying these reactions under conditions that differ significantly from the thousands of prior studies in the condensed phase.

Full text of DOI:10.1021/acscatal.9b04906

Subscriber access provided by UNIVERSITY OF TECHNOLOGY SYDNEY
Article
Demonstrating the critical role of solvation in supported
Ti and Nb epoxidation catalysts via vapor-phase kinetics
Sol Ahn, Scott L Nauert, Kenton E Hicks, Matthew Alexander
Ardagh, Neil M. Schweitzer, Omar K. Farha, and Justin M Notestein
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.9b04906 • Publication Date (Web): 21 Jan 2020
Downloaded from pubs.acs.org on January 22, 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 23
ACS Catalysis
1
2
3
4
5
6
7
Demonstrating the critical role of solvation in supported Ti and Nb
epoxidation catalysts via vapor-phase kinetics
8
9
Sol Ahn,† Scott L. Nauert,† Kenton E. Hicks,‡ M. Alexander Ardagh,† Neil M. Schweitzer,†
Omar K. Farha,*†‡ and Justin M. Notestein*†
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
†Department of Chemical and Biological Engineering, and ‡Department of Chemistry, Northwestern University,
2145 Sheridan Road, Evanston, Illinois 60208, United States
■ ABSTRACT
Catalytic oxidation of hydrocarbons with hydrogen peroxide (H2O2) has been of the utmost
importance for several decades. The vast majority of studies have been performed in the
condensed phase, even though condensed phases introduce complex solvent effects and can
promote the leaching of active sites. In response, we have built a custom reactor system to
understand H2O2 activation and selective oxidation in the vapor-phase. In this report, we study
the epoxidation of cyclohexene with H2O2 over four Lewis-acidic metal oxide catalysts: Ti and
Nb grafted on SiO2 and on the Zr based metal–organic framework, NU-1000. The M-SiO2
materials are highly selective to the formation of epoxides and diols, as they can be in the
condensed phase, while the NU-1000 based materials are far more prone to overoxidation to
CO2, which appears to be connected to their strong reactant adsorption. Apparent activation
energies are calculated for all materials when operating in the same kinetic regime, and the
heats of cyclohexene adsorption into their pores are then used to directly compare intrinsic
enthalpies of activation in the vapor vs. condensed phase for the M-SiO2 catalysts. Nb-SiO2
catalysts exhibit similar intrinsic enthalpies of activation in the vapor and condensed phases,
whereas the condensed phase transition state in Ti-SiO2 is 24 kJ/mol lower in energy than that
of the same material in the vapor phase. These experiments establish another methodology for
understanding the various roles of solvent in selective oxidation reactions and studying these
reactions under conditions that differ significantly from the thousands of prior studies in the
condensed phase.
KEYWORDS: epoxidation, hydrogen peroxide, selective oxidation, supported catalyst, metal
oxide catalyst, metal–organic framework
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 23
1
2
3
4
5
■ INTRODUCTION
6
7
8
9
Hydrogen peroxide (H2O2) is considered an attractive oxidant because its reaction byproducts
are water and oxygen, and because its oxygen atom efficiency is higher than that of organic
peroxides.1 H2O2 plays an important role as the terminal oxidant of innumerable oxidation
reactions such as alkene epoxidation, alcohol oxidation, sulfoxidation, and so on.2 Alkene
epoxidation is of particular interest because epoxides serve as building blocks for many resins
and fine chemicals.3 Among them, epoxidation of cyclohexene is a frequent probe reaction
because it provides extensive information about the mechanism of H2O2 activation. In
particular, its many reaction products (epoxide, diol, cyclohexenol, CO2, etc.) can help
distinguish between direct and radical reaction pathways that occur during the activation of
H2O2.4-6
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
The vast majority of previous studies have focused on condensed phase reactions when using
aqueous H2O2 or organic peroxide, despite the following limitations. First and foremost,
reaction solvents can significantly affect reactivity and/or selectivity due to competitive
adsorption among solvents and reactants, effects of solvent acidity, variability solubility of
reactants and products, and so on.7-9 Second, it has been known for decades that some active
sites, particularly those based on V or Mo ions, can leach into reaction solvent.10-11 Conversely,
materials like a UiO-66-supported VOx catalyst showed good time-on-stream stability under
gas-phase oxidative dehydrogenation of cyclohexene without metal ion leaching.12 Despite the
promise of circumventing these issues and of better understanding solvent effects, practical
issues have restricted the study of selective oxidation with vaporized H2O2 under flow
conditions. These issues largely arise from the competition between vaporization and
decomposition of H2O2 at elevated temperature. To the best of our knowledge, there are only
two reported examples of heterolytic activation of vaporized H2O2 in flow reactors.13-14
Motivated by the previous two studies, we present a new vapor-phase peroxide reactor (scheme
1) which is robust and optimized for the purpose of studying H2O2 activation in the vapor-
phase. With this reactor, we compare the kinetics of vapor-phase cyclohexene epoxidation over
Ti and Nb grafted on a conventional support (SiO2) and on a less conventional support
consisting of the Zr based metal–organic framework, NU-1000. We then compare the kinetics
of the reaction and the thermodynamics of reactant binding to those previously collected during
high fidelity kinetics experiments of alkene epoxidation in the condensed phase.15-17
ACS Paragon Plus Environment

Page 3 of 23
ACS Catalysis
1
2
3
4
5
■ EXPERIMENTAL METHODS
6
7
8
9
Synthesis of Catalysts. For the synthesis of M-SiO2, 5 g of Selecto SiO2 (Fisher Chemical, 6
nm pore size, 550 m2/g, 63-200 μm particle size) was dried for 30 min at 120 °C in a glass
oven. Then, the SiO2 was added to an oven dried 250 mL round bottom flask along with 50 mL
of freshly distilled anhydrous toluene (Sigma-Aldrich, ACS Reagent, > 99.5%) and stirred at
200 rpm with an oven dried teflon stir bar at ambient temperature. Metal precursors, 227 mg
of titanocene dichloride (Cp2Ti(IV)Cl2, Strem Chemicals, > 99%) or 269 mg of niobocene
dichloride (Cp2Nb(IV)Cl2, Strem Chemicals, > 99%), were added to the mixture. A high
dispersion (0.2 metal/nm2) was targeted for Ti-SiO2 and Nb-SiO2. The mixtures were stirred
for 1 h at ambient temperature, then the mixtures were vacuum filtered, washed with 200 mL
of toluene and hexanes (Fisher Chemical, Certified ACS, > 98.5%), and dried in air for 5 min.
The as-made Cp2Ti-SiO2 and Cp2Nb-SiO2 were loaded into a muffle furnace and calcined at
550 °C for 6 h with a 10 °C/min ramp rate to yield Ti-SiO2 and Nb-SiO2.
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
The catalyst support, NU-1000, was prepared and activated following methods previously
established.18 For the synthesis of M-NU-1000, 200 mg of NU-1000 (185.1 μmol of the Zr6
node) and 449.3 μL of titanium(IV) isopropoxide (Sterm, min. 98%) or 555.1 μL of (tert-
butylimido)tris(diethylamino)niobium(V) (Strem, min. 98%) were mixed in 10 mL heptane
(Sigma Aldrich, 99%, anhydrous) for Ti-NU-1000 and Nb-NU-1000, respectively. These
loadings correspond to 8 metal equivalents per Zr6 node. The reaction mixtures were stored in
a glove box overnight with occasional swirling. The mixtures were decanted, and 10 mL of
fresh heptane were added. The heptane washing cycle was repeated for four times to remove
unreacted metal precursors. After these heptane washing steps, the materials were transferred
from the glove box, and then washed in acetone four times. Any ligands remaining on the
grafted Ti and Nb precursors were removed when exposed to moist air to yield TiOx and NbOx
on the nodes of NU-1000.
Characterization of Catalysts. M-SiO2 catalysts (ca. 2mg) were digested in a mixture of 1
mL of hydrofluoric acid (Macron, 48 wt.%) [CAUTION: handle HF with extreme care] and 9
mL of nitric acid (Fisher Chemicals, 67-70% w/w, trace metal grade) with mixing on a shaker
plate. M-NU-1000 catalysts (ca. 2 mg) were microwave digested in 2 mL of nitric acid at
150 °C for 15 min. Titanium (323.452, 334.941, 336.121, and 337.280 nm), niobium (309.418,
316.340, 319.498, and 322.548 nm), and zirconium (327.305, 339.198, 343.823, and 349.621
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 23
1
2
3
4
5
6
7
nm) loadings were analyzed using inductively coupled plasma atomic emission spectroscopy
(ICP-AES, iCAP 7600, Thermo Scientific) by comparing calibrated standard solutions.
8
9
Prior to N2 physisorption, catalysts were degassed at 120 °C under dynamic vacuum for 18
h using a Smart VacPrep (Micromeritics). N2 adsorption/desorption isotherms were collected
at 77 K using a Tristar II (Micromeritics). Cyclohexene vapor adsorption isotherms on bare
supports (NU-1000 and SiO2) were collected at 313 K, 353 K, and 383 K using a 3 Flex
(Micromeritics). The isosteric heat of adsorption of cyclohexene on SiO2 was calculated using
a linear portion of the isotherm between 0.5-2 cm3 (STP)/gSiO2, using equation (1).
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
ΔHads
RT2
d
(
(
)
)
ln Pc6
=
(1)
|
|
dT
nads
nads
Catalysis. A reactor system was built by our team based on a design shown in Scheme 1. The
system is equipped with a syringe pump (KDS) to inject liquid-phase H2O2 from to a plastic
syringe (Norm-Ject, 10 mL) through a FEP (fluorinated ethylene propylene) capillary tube
(Valco Instruments, 1/16 inch OD, 0.03 inch ID). This tube material minimizes decomposition
of H2O2, unlike stainless-steel. The reactant vapor (cyclohexene in this study) was introduced
through a quartz bubbler (Wilmad Labglass, 1/4 inch connection with extra coarse frit). The
reactor furnace (Northwestern University Instrument Shop) consists of three aluminum heating
blocks separated by a ceramic insulator to minimize heat conduction between adjacent blocks.
The temperatures of all three aluminum heating blocks were individually controlled by a Quad-
K temperature controller (J-KEM scientific) with K-type thermocouples (Omega) and
stainless-steel sheath cartridge heaters (Omega, 150 Watts). During kinetic measurements, the
temperature of the inlet block was fixed at 110 °C to minimize H2O2 thermal decomposition,
as well as to ensure complete vaporization of H2O2 liquid. Figure S1 (150ºC vaporization
temperature) shows massive fluctuations in product concentration with time-on-stream (TOS),
while Figure S2 shows discontinuities in the product concentrations with TOS that result from
a low vaporization temperature of 80 °C. The temperature of the outlet block was fixed at
115 °C to minimize thermal decomposition of the products. A counter example (150 °C outlet)
is shown in Figure S3.
ACS Paragon Plus Environment

Page 5 of 23
ACS Catalysis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Vapor-phase hydrogen peroxide reactor scheme
Kinetic tests were performed at reaction block temperatures of 110, 120, 130, 140, and
150 °C to test apparent activation barriers of catalysts. M-SiO2 (20 mg) and M-NU-1000 (10
mg) were both diluted in 200 mg of quartz sand, and these catalysts beds were supported on
quartz wool in a quartz tube reactor (Wilmad Labglass, 18 inch long, ¾ inch diameter). In a
typical reaction test, the reactor bed was brought to temperature and cyclohexene (Sigma-
Aldrich, ≥ 99.0%) was introduced by flowing He (5 mL/min, 99.999%, Airgas) through the
bubbler at nominal atmospheric pressure at 25 °C (11.9 kPa). Additional He (20 mL/min) was
added through a second mass flow controller (Brooks Instruments). The cyclohexene partial
pressure was varied between 0.6 to 3 kPa by varying the rates of the two He flows while
maintain a constant total He flow of 25 mL/min. Aqueous H2O2 (30 wt. %) was diluted in
acetonitrile (10 mL H2O2 and 20 mL acetonitrile), and then 7 g of anhydrous magnesium sulfate
was added to remove water, followed by decanting to yield a 4 M H2O2 in acetonitrile
solution.19 This solution was loaded into a plastic syringe, bubbles were removed, and this
solution was injected directly into the quartz tube reactor with a rate of 0.2 mL/h. The solution
was diluted further with acetonitrile to vary its partial pressure (0.8 to 3 kPa), while maintain a
constant injection rate 0.2 mL/h. Here, it was important that the tip of the FEP capillary tube
was inside the inlet heating block zone, otherwise droplets would form and lead to system
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 23
1
2
3
4
5
6
7
8
9
instability (e.g. Figure S4) Concentration profiles of cyclohexene and products were collected
up to 1000 mins time-on-stream (TOS). C6 products were separated by GC-FID (Agilent 7890,
He carrier) with a HP-INNOWAX (50 m length, 0.2 mm diameter, 0.4 μm film) column.
Products detected and quantified were cyclohexene oxide (epoxide), trans-1,2-cyclohexane
diol (diol), 2-cyclohexen-1-one (cyclohexenone), 2-cyclohexen-1-ol (cyclohexenol), and
cyclohexanone (cyclohexanone). H2O, O2, and CO2 were separated by a combination of HP-
PLOT/Q (30 m length, 0.53 mm diameter, 40 μm film) and HP-MOLESIEVE (15 m length,
0.53 mm diameter, 50 μm film) and detected with a TCD in the same GC.
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
Product yields were normalized with respect to the cyclohexene feed, as per equation (2).
Some experiments were carried out at high cyclohexene conversion to verify good carbon
balance (> 90%) using CO2 and the C6 products listed above, but otherwise, kinetic studies
were carried out in the differential conversion regime (< 10%) and thus it is more accurate to
discuss product yields rather than reactant conversion. Product yields at near steady state
(typically > 400 min TOS) were extrapolated linearly to t = 0 min to calculate turnover
frequencies (TOF). The conclusions of this paper are not significantly changed if average rates
at long TOS are used instead. H2O2 conversions in the reactor cannot be directly measured;
GC-TCD confirms that all H2O2 fed is converted to H2O under all conditions, presumably due
in part to decomposition in the stainless-steel transfer lines downstream of the reactor.
Product yield
∑
molproduct
=
(2)
∑
molproduct
molreactant
+
Where the sum of products is given in equation (3).
molCO
6
2
molproduct = molepoxide + moldiol + molenone + molenol + molanone
+
∑
(3)
■ RESULTS AND DISCUSSION
Catalysts Synthesis.
Table 1 summarizes the bulk properties of the catalysts synthesized in this study. All catalysts
were synthesized to possess low surface density (i.e. a high dispersion of MOx), which leads to
a high fraction of active Lewis acid sites for heterolytic H2O2 activation, as shown in our
ACS Paragon Plus Environment

Page 7 of 23
ACS Catalysis
1
2
3
4
5
6
7
8
9
previous works.19-21 In the case of M-NU-1000, we were able to achieve low surface density
with high metal loading because of the spatial isolation of the grafting sites (the Zr6 nodes)
afforded by the ~1 nm linkers. M-SiO2 were calcined at 550 ºC after grafting of metal
precursors to generate MOx sites, while the grafting ligands used to prepare M-NU-1000 were
hydrolyzed by ambient moisture because the organic linkers of NU-1000 do not permit high
temperature calcination.
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
Table 1. Summary of catalyst properties
M loading
[wt. %]
Surface density
[M/nm2]
BET Surface area
Catalyst
[m2/gcatalyst
]
a
Ti-SiO2
0.87
1.66
5.16
9.55
0.2
0.2
0.3
0.3
380
300
1930
1330
a
Nb-SiO2
Ti-NU-1000b
Nb-NU-1000b
a
Bare SiO2 surface area 550 m2/g.
b
Bare NU-1000 surface area 2160 m2/g.
Catalysis.
Vapor-phase cyclohexene epoxidation with H2O2 was examined to study trends in reactivity
and selectivity over M-SiO2 and M-NU-1000. After addressing experimental complexities in
delivering stable amounts of vaporized cyclohexene and H2O2, such as optimizing the
temperature of the H2O2 vaporization region and placement of the H2O2 feed tube (discussed
in experimental section), we can achieve stable yields (Figure 1) and cyclohexene
concentration (Figure S5) profiles over long times-on-stream. All kinetic tests were performed
in the differential conversion regime for cyclohexene, and the reaction was run in the absence
of external transport limits (Figure S6). Blank reactor tests indicate that all oxidation reactions
– the formation of C6 products and CO2 – require both H2O2 and catalyst. We observe some
epoxide hydrolysis to diol in the stainless-steel transfer lines downstream of the reactor, so we
do not include rate laws for diol formation in the kinetic model.
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 23
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Representative time-on-stream yield and C6 product selectivity profiles of
cyclohexene epoxidation with H2O2 at 120 °C, cyclohexene 3 kPa, H2O2 3 kPa over (a) Ti-
SiO2 and (b) Ti-NU-1000. (○ total product yield, (green) epoxide, (orange) diol, (yellow)
cyclohexanone, (blue) cyclohexenol, (gray) cyclohexenone) Dashed lines show how data are
extrapolated back to t=0 min. Only selectivity to C6 products are shown in (b) for clarity. See
Figure 2 for representative selectivities including CO2.
Yield profiles of both types of catalysts follow the same general trend in that they decrease
sharply over the first ~ 200 min, then reach a relatively steady rate with constant selectivity
among C6 species. After the initial period, the activity over Ti-SiO2 or Nb-SiO2 (Figure 1(a)
and Figure S7(a)) is very stable, while the activity over Ti-NU-1000 or Nb-NU-1000 continues
to decrease with time on stream without a change in selectivity, indicating a loss of accessibility
to the active sites due to framework restructuring. In the following, rates are extrapolated back
to zero time, but similar conclusions are reached using rates after a given time on stream.
Cyclohexene conversion turnover frequencies vary by 20-fold, with Ti-SiO2 (1.1 × 10-1 molC6
molTi-1 min-1) being ten times faster than Ti-NU-1000 (1.1 × 10-2 molC6 molTi-1 min-1), and Nb-
SiO2 (5.1 × 10-2 molC6 molNb-1 min-1) being an order of magnitude faster than Nb-NU-1000 (5.9
× 10-3 molC6 molNb-1 min-1) at 120 °C with 3 kPa of both reactant and oxidant. In addition, there
are significant differences in product selectivity between the two catalyst supports. The silica
supported Ti and Nb catalysts (Figure 1(a) and Figure S7(a)) cleanly yield only epoxide and
trans-cyclohexanediol, each at ~50% selectivity under all reaction conditions studied. Neither
allylic oxidation C6 products (e.g. cyclohexenol) nor CO2 were detected at any temperature or
reactant partial pressure studied with the M-SiO2 catalysts. No cis-cyclohexanediol was
ACS Paragon Plus Environment

Page 9 of 23
ACS Catalysis
1
2
3
4
5
6
7
8
9
detected, indicating that all diol was derived from hydrolysis of epoxide, rather than direct cis-
dihydroxylation.22 Separately, oxidizing cis-3-hexene with vaporized H2O2 over M-SiO2
catalysts results in only cis-3,4-epoxyhexane, and the absence of isomerization products is
consistent with epoxide formation by a concerted mechanism. In contrast, NU-1000 supported
Ti and Nb (Figure 1(b) and Figure S7(b)) catalysts produce a variety of C6 products including
cyclohexenone, cyclohexenol, and cyclohexanone, in addition to epoxide and diol. Co-feeding
experiments (Table S1) suggest that trans-cyclohexanediol is dehydrated to cyclohexanone
over M-NU-1000 catalysts. Increasing either PH2O2 or time on stream results in a decrease in
cyclohexanone selectivity that is coupled with increasing CO2 selectivity. It appears that there
is a species on M-NU-1000 that begins as a dehydration site and evolves into a combustion site
after extended exposure to H2O2. Zirconia-based mixed oxides frequently show this type of
behavior.23-25 With the exception of the transient formation of cyclohexanone, the selectivity
to other C6 products remains constant with time on stream for all catalysts.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Selectivity dependence on reaction temperatures over Ti-NU-1000 (cyclohexene 3
kPa, and H2O2 3 kPa) at 400 min TOS. (shaded) CO2/6, (blue) cyclohexenol, (gray)
cyclohexenone, (green) epoxide, (orange) diol.
In addition to C6 products, we observe CO2 only over the M-NU-1000 catalysts. Figure 2
shows selectivity trends (Figure S8 for Nb-NU-1000) with respect to reaction temperature (110
– 130 °C). Data were not collected above 130 °C for M-NU-1000 due to instability in the
reaction rate. Selectivity to CO2 and the radical-derived products cyclohexenol and
cyclohexenone increases from ~40 % to ~60 % with increasing temperatures. However,
selectivity among the other C6 products is largely independent of reaction temperatures over
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 23
1
2
3
4
5
6
7
8
9
the temperature range studied. The ratio between cyclohexenol and cyclohexenone is 1:1
regardless of reaction temperatures, consistent with a common origin through the
decomposition of cyclohexenyl hydroperoxide.4-5, 26 As in the case of M-SiO2, all detected diol
is trans-cyclohexanediol, indicating that it is produced from epoxide hydrolysis.
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 have previously investigated the same reaction over Nb-NU-1000 (indicated as Nb-SIM
in the previous study), but in condensed phase batch reactors operating at 1M cyclohexene in
acetonitrile and 65 °C.19 In the previous study, the product selectivity over Nb-NU-1000 was
consistent with 57-72 % pathway selectivity for heterolytic activation of H2O2 leading directly
to epoxide + diol. Over Nb-NU-1000 in the vapor phase, if we include cyclohexanone and the
diol both as epoxide-derived products, the pathway selectivity is 53-63%, depending on TOS,
and remarkably similar to the results in the condensed phase. Individual selectivities to the
various C6 products are also similar in the liquid and vapor phases, indicating that the higher
temperature and the presence or absence of solvent does not change the relative barriers to
formation of the various C6 products, and that the barriers to heterolytic vs homolytic pathways
must be similar to one another over Nb-NU-1000. The similarity between liquid and gas-phase
operation for Nb-NU-1000 is also consistent with the strong adsorption of cyclohexene in this
framework, corresponding to a relatively liquid-like 0.8 cyclohexene molecules/nm2 at 3 kPa
(Figure 3).
Figure 3. Cyclohexene vapor adsorption isotherms at 110 °C on bare supports: number of
cyclohexene molecule per area of supports basis (○: SiO2, ●: NU-1000).
Gas phase reaction of H2O2 and cyclohexene over M-SiO2 produces only epoxide and diol.
The absence of any detectable radical-derived products (cyclohexenol and cyclohexenone)
ACS Paragon Plus Environment

Page 11 of 23
ACS Catalysis
1
2
3
4
5
6
7
8
9
contrasts with the results in 1M cyclohexene in acetonitrile at 65 °C, where radical pathways
account for 48 and 16% of the products for Ti-SiO2 and Nb-SiO2, respectively,20 but are
comparable to selectivities observed by other studies at more dilute conditions over M-beta (<
50 mM cyclohexene in acetonitrile) where epoxide is the only product.15 Higher selectivity
toward direct pathways (to epoxide and diol) can be attributed to the (1) low reactant partial
pressures and attendant low C6 surface concentrations over SiO2 (Figure 3), making it less
likely that homolytic decomposition of the M-OOH will react with a C6 species and (2)
intrinsically more favorable direct vs. radical pathways at the higher reaction temperatures of
the vapor-phase experiments. In condensed phase experiments, homolytic decomposition of
H2O2 has been seen to have a lower or comparable barrier to epoxidation over Ti-SiO2
catalysts,17 from which one would expect selectivity to improve with temperature.
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 4. Reaction orders at 120 °C with respect to (a) cyclohexene (partial pressure: 0.6 to 3
kPa) with fixed H2O2 partial pressure at 3 kPa, and (b) H2O2 (partial pressure: 0.8 to 3 kPa)
with fixed cyclohexene partial pressure at 3 kPa (○: Ti-SiO2, △: Nb-SiO2, ●: Ti-NU-1000, ▲:
Nb-NU-1000). All C6 products and CO2 are considered when calculating TOF.
Figure 4 shows rates used to determine reaction orders of M-SiO2 and M-NU-1000 with
respect to cyclohexene and H2O2. Orders of reaction with respect to cyclohexene are close to
unity over all catalysts (Ti-SiO2 1.1, Nb-SiO2 1.0, Ti-NU-1000 0.8, Nb-NU-1000 1.3), and
orders with respect to H2O2 are near zero for all catalysts. Reaction orders for individual
products generally follow the same pattern (Figure 5). The zeroth order dependence on H2O2
allows us to be confident of our conclusions about reaction kinetics even without necessarily
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 23
1
2
3
4
5
being certain of the extent of H2O2 decomposition in the reactor.
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. The dependence of product formation rates on reactant and oxidant partial
pressures over Ti-NU-1000 at 120 °C and 400 min TOS. Dependence on (a) cyclohexene
partial pressures with H2O2 3 kPa and (b) H2O2 partial pressures with cyclohexene 3 kPa
(+CO2/6, △epoxide and diol, ×cyclohexanone, ◇cyclohexenol, □cyclohexenone).
Figure 5(a) indicates that the formation rate of all C6 products and CO2 linearly depend on
partial pressure of cyclohexene. Note that only CO2 has a non-zero intercept, presumably
arising from a small contribution from oxidation of acetonitrile (the carrier of the H2O2).
Varying H2O2 partial pressure (Figure 5(b)), has a slightly more complex impact on the reaction
orders of product formation. Within the uncertainties for these slow catalysts, the rates to
epoxide + diol, cyclohexenol and cyclohexenone are relatively stable with PH2O2. In contrast,
cyclohexanone rates decrease with PH2O2 while CO2 rates increase with PH2O2 at a given TOS.
Consulting Figures 1(b) or S7(b), we see that this behavior parallels the TOS behavior:
cyclohexanone decreased with TOS and CO2 rates increased with TOS for a given partial
pressure, while the other species were stable. Nb-NU-1000 (Figure S9) has similar reaction
orders to those of Ti-NU-1000, but even smaller selectivities to cyclohexenol and
cyclohexenone. This is consistent with condensed phase trends for Ti-SiO2 vs. Nb-SiO2.20
ACS Paragon Plus Environment

Page 13 of 23
ACS Catalysis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Arrhenius plot of all catalysts. Cyclohexene 3 kPa, H2O2 3 kPa, and 110-150 °C.
(○: Ti-SiO2, △: Nb-SiO2, ●: Ti-NU-1000, ▲: Nb-NU-1000). All C6 products and CO2 are
considered when calculating TOF.
Figure 6 gives an Arrhenius plot of all catalysts collected at 3 kPa partial pressure of both
cyclohexene and H2O2, corresponding to the 1st order regime in cyclohexene and 0th order in
H2O2 for all catalysts. All four catalysts have similar slopes, corresponding to apparent
activation barriers of 26 ± 4 kJ mol-1 (Table 2), activation enthalpies (23 ± 4 kJ mol-1) and
entropies (-130 ± 20 J mol-1 K-1). Only the data between 110-130 °C were used for M-NU-
1000, because reaction rates become unstable at higher temperatures. Overall, the significant
differences in rates absent significant differences in activation enthalpy or entropy, suggests
that these materials primarily differ in the fraction of each MOx that are kinetically relevant
under vapor phase conditions.
Table 2. Apparent activation barriers, activation enthalpies and entropies of catalysts
Catalyst
Ea [kJ mol-1]
ΔH‡ [kJ mol-1]a
ΔS‡ [J mol-1 K-1]a
Ti-SiO2
Nb-SiO2
Ti-NU-1000
Nb-NU-1000
30
27
22
29
27
24
19
26
-110
-130
-150
-140
a
Reported values are calculated using Eyring equation (Figure S10) and steady TOF.
ACS Paragon Plus Environment

ACS Catalysis
Page 14 of 23
1
2
3
4
5
6
7
To describe the observed behavior, we propose an overall Ely-Rideal type reaction
mechanism shown in Scheme 2 for Ti-SiO2 (left) and Nb-SiO2 (right), which follows the
extensively supported mechanism of Bregante et. al.15, 17
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
Scheme 2. Proposed reaction networks of vapor-phase cyclohexene epoxidation in the steady
conversion regime with H2O2 over Ti-SiO2 or Ti-NU-1000 (left) and Nb-SiO2 or Nb-NU-
1000 (right).
Epoxide hydrolysis is not included in the mechanism, as it may occur in the reactor or
downstream of it. The transient formation of cyclohexanone over M-NU-1000 is also not
included. Olefin binding to the catalyst center is assumed to be weak and not along the reaction
trajectory, and thus not shown. However, its concentration in the pore is expected to be strongly
dependent on the support. Reaction in polar solvent may also include bound solvent in resting
and transition states, which are also not shown. Finally, the nature of the H2O2 activation in
step k3 may vary depending on the reaction environment (solvent vs. vapor phase).
For M-SiO2, radical pathways lead only to the consumption of H2O2, since no CO2,
cyclohexenol or cyclohexenone are observed. For M-NU-1000, the radical pathways lead to
these three products. As shown in Scheme 3, the weak allylic C-H on cyclohexene is expected
to be abstracted first, followed by reaction with O2. Once cyclohexenyl peroxy species form
ACS Paragon Plus Environment

Page 15 of 23
ACS Catalysis
1
2
3
4
5
6
7
(radical or the neutral hydroperoxide), they can undergo bimolecular pathways to cyclohexenol
and cyclohexenone. The CO2 observed also is presumably the result of serial radical reactions
that begin with the formation of cyclohexenyl radicals.
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
Scheme 3. Proposed sequence of radical reactions
The remainder of this discussion uses only the catalytic cycle in Scheme 2 and M-SiO2
catalysts. This mechanism presents quasi-equilibrated adsorption/desorption steps of H2O2 (K1),
epoxide (K4), and water (K6) on MOx sites. Following the work of Bregante et. al.,15, 17 which
observed the same reaction orders, we assume that the M-OOH intermediate is formed upon
irreversible activation of H2O2 (k2) and that it is the most abundant surface intermediate (MASI).
For the purposes of this discussion, the exact structure of the M-OOH intermediate is not
critical. This active intermediate can then form epoxides via reaction with cyclohexene (k3) or
can undergo homolytic decomposition (k5), which contributes only to H2O2 decomposition over
M-SiO2. Here, the rate of epoxide and diol formation (repox) is defined as per equation (4):
#
repox = k3[M - OOH][C6H10 ]
(4)
In the limit that [M-OOH] = [M], the turnover frequency (TOF = repox/M) is simply
#
=
TOF
k3[C6H10 ]
(5)
#
#
The notation [C6H10 ] denotes the intrapore concentration of C6H10. If we let [C6H10 ] =
Kads[C6H10] for adsorption in the Henry’s law regime, then overall TOF is defined by equation
(6):
[
]
[
]
=
=
TOF
k3Kads C6H10
kapp C6H10
(6)
This simple rate expression is in accordance with the observed dependences of H2O2 and
cyclohexene. Finally, we define the measured activation enthalpy as a combination of terms
intrinsic to oxidant activation and O transfer, and a term for adsorption of the olefin:
∆Ha‡pp = ∆Hi‡nt + ∆Hads
7
( )
ACS Paragon Plus Environment

ACS Catalysis
Page 16 of 23
1
2
3
4
5
where ΔH‡int is the intrinsic reaction barrier for O transfer (step 3) and ΔHads is enthalpy change
6
7
associated with adsorption of cyclohexene into the pores near the active site.
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
Scheme 4. Proposed Born-Haber thermochemical cycle of cyclohexene epoxidation in
condensed phase and vapor phase. The vertical heights are approximately to scale. (Unit of
numbers in scheme is kJ/mol.)
In order to quantitatively compare the cyclohexene oxidation kinetics in the vapor and
condensed phases, a Born-Harber thermochemcial cycle of cyclohexene epoxidation is
constructed (Scheme 4). In both the liquid and gas phases, apparent activation energies are
referenced to fluid-phase cyclohexene, and the influence of cyclohexene adsorption in the pores
near the active site can be removed by calculating an intrinsic reaction enthalpy (ΔH‡ ) that is
int
given by equation (7). Strong adsorption can make for an apparently smaller enthalpy of
activation, and vice versa. ΔHads is assumed to be dependent on the pore structure and reaction
environment, but not the active site identity. The low-coverage ΔHads of cyclohexene vapor on
SiO2 was determined to be -45 kJ/mol (Figure S11 and Table S2), giving a ΔH‡int in the vapor-
phase for Ti-SiO2 and Nb-SiO2 of 72 kJ/mol and 69 kJ/mol. The ΔHads of cyclohexene from
acetonitrile solution into zeolite β is only -5 kJ/mol.27 Combined with published apparent
enthalpies of activation for cyclohexene epoxidation,15 this results in a ΔH‡int for Ti- β and Nb-
β of 48 kJ/mol and 77 kJ/mol, respectively, for cyclohexene oxidation in liquid acetonitrile.
Scheme 4 shows the measured ΔH‡ values and the values of ΔH‡ backed out from
app
int
equation (7). All of the experiments in both the vapor- and condensed phases were carried out
in the same kinetic regime of first-order olefin and zero-order H2O2, allowing the rate-limiting
ACS Paragon Plus Environment

Page 17 of 23
ACS Catalysis
1
2
3
4
5
6
7
8
9
O-transfer step to be probed directly. Therefore, the intrinsic activation enthalpies can be
compared between the vapor- and condensed-phase experiments. For Nb-SiO2, the intrinsic
activation enthalpies are similar in the two phases, requiring that the precursor state and the
transition state are stabilized similarly by solvent (ΔHsolv,cat ~ ΔHsolv,TS in Scheme 4). On the
other hand, the Ti-SiO2 transition state in the condensed phase (ΔH‡int,solv = 48 kJ/mol) appears
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 be stabilized by 24 kJ/mol relative to the same transition state in the vapor phase (ΔH‡
int,vap
= 72 kJ/mol), which results in larger ΔHsolv,TS than ΔHsolv,cat. No charge develops in the
transition state, so significant stabilization by a mean field solvation effect is not expected.28
Instead, we assign it to the role of liquid solvent in helping shuttle protons during the O transfer
step and greatly decreasing the barrier. Equivalently, the transition state is more stabilized than
the precursor state because of this direct involvement of the solvent.
In liquid phase reactions, the polar acetonitrile solvent and the water co-fed with H2O2
provide ample opportunities to H-bond to the activated H2O2 species. Prior calculations have
shown that coordinated solvent, water, or silanol (for in-framework, microporous materials)
substantially lower the barrier for epoxidation over Ti-based catalysts, via structures as shown
in Figure 7(a).29-31 At the low partial pressures of the vapor phase experiments, these structures
presumably cannot form (Figure 7(b)), and the mechanism is forced through a higher barrier
pathway. Engineering new catalysts to possess an adjacent H-bonding motif may help lower
barriers in the vapor phase. In contrast, Nb-SiO2 does not appear to be in need of an external
H-shuttle, due to its additional Nb-OH bond (Figure 7(c)). Therefore, the liquid and gas-phase
barriers are similar. Overall, the comparisons between the liquid- and gas-phase operation help
support a decades-old computational prediction about the mechanism of epoxidation by Ti-
SiO2 and suggest new avenues of mechanistic investigation and materials design.
Figure 7. Proposed active sites and calculated ΔH‡int for (a) Ti-SiO2 in solvent (S), (b) Ti-SiO2
in the vapor phase, and (c) Nb-SiO2 for either phase.
ACS Paragon Plus Environment

ACS Catalysis
Page 18 of 23
1
2
3
4
5
6
7
■ CONCLUSIONS
8
9
A unique reactor has been built to study vapor phase H2O2 activation reactions, such as
cyclohexene epoxidation over solid Lewis acid catalysts. M-SiO2 catalysts cleanly give only
epoxide and diol products, in contrast to many reports of allylic oxidation in the condensed
phase. In contrast, the M-NU-1000 catalysts show extensive allylic oxidation and CO2
formation in the vapor-phase. The strong adsorption of reactants into MOFs is frequently an
advantage, but here, the strong adsorption of cyclohexene may be a liability.
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
The reactor design allows for high fidelity kinetics experiments to be carried out with
vaporized H2O2. Running all catalysts under conditions where rates were first order with
respect to cyclohexene and zeroth order with respect to H2O2 allows us to directly compare
activation energies for the rate-limiting O-transfer step. Activation energies did not differ
significantly between the different catalysts in the vapor phase, even though apparent TOFs
varied by more than 20 times, suggesting that the fraction of kinetically relevant MOx active
sites varies between the samples. Secondly, we compare intrinsic activation enthalpies between
the condensed and vapor-phases over M-SiO2. Nb-SiO2 behaves very similarly in the two
phases, while Ti-SiO2 has a much lower activation enthalpy in the condensed phase. These
experimental findings provide evidence for a long postulated direct role for solvent in H2O2
activation over Ti-SiO2. Overall, we expect that the ability to experimentally investigate H2O2
activation in the absence of a liquid solvent and residual species like water will be broadly
useful in understanding the role of solvent in many heterogeneously-catalyzed reactions.
■ AUTHOR INFORMATION
Corresponding Author
* Email: o-farha@northwestern.edu, j-notestein@northwestern.edu
Author Contributions
ACS Paragon Plus Environment

Page 19 of 23
ACS Catalysis
1
2
3
4
5
6
7
8
9
The manuscript was written through contributions of all authors. All authors have
given approval to the final version of the manuscript.
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
Notes
The authors declare no competing financial interest.
■ ASSOCIATED CONTENT
Supporting Information
Additional characterization data and kinetic data (PDF) are available in the Supporting
Information, free of charge on the ACS Publications website at DOI: XXX.
■ ACKNOWLEDGMENTS
Authors gratefully acknowledge the financial support from the Inorganometallic Catalyst
Design Center, an EFRC funded by the DOE, Office of Basic Energy Sciences (DE-
SC0012702). S.L.N. acknowledges the financial support from the National Science Foundation
Graduate Research Fellowship under Grant No. DGE-1324585. S.L.N. and M.A.A.
acknowledge support from the Northwestern University Institute for Catalysis in Energy
Processes (ICEP). ICEP is funded through the U.S. Department of Energy, Office of Basic
Energy Sciences (DE-FG02-03ER15457). Metal analysis was performed at the Northwestern
University Quantitative Bio-element Imaging Center. The REACT core facility acknowledges
funding from Northwestern University’s Core Facility funding program used for the purchase
of the 3 Flex. We thank Dr. Stephanie Kwon (University of California, Berkeley) for very
helpful discussions. Authors thank Daniel T. Bregante and Prof. David W. Flaherty (University
of Illinois at Urbana-Champaign) for helpful discussions and isothermal calorimetry
measurement of cyclohexene from acetonitrile.
ACS Paragon Plus Environment

ACS Catalysis
Page 20 of 23
1
2
3
4
5
■ REFERENCES
6
7
1.
Clerici, M. G.; Kholdeeva, O. A., Liquid Phase Oxidation via Heterogeneous Catalysis:
8
9
Organic Synthesis and Industrial Applications. In Liquid Phase Oxidation via Heterogeneous Catalysis:
Organic Synthesis and Industrial Applications, Wiley: 2013; p 526.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2.
Jones, C. W.; Clark, J. H., Application of hydrogen peroxide for the synthesis of fine
chemicals. In Applications of Hydrogen Peroxide and Derivatives, Jones, C. W.; Clark, J. H., Eds. The
Royal Society of Chemistry: 1999; pp 79-178.
3.
Chemistry, Wiley: 2000; Vol. 13, p 139.
4. Morlanés, N.; Notestein, J. M., Grafted Ta–calixarenes: Tunable, selective catalysts for direct
olefin epoxidation with aqueous hydrogen peroxide. J. Catal. 2010, 275 (2), 191-201.
5. Fraile, J. M.; Garcıa, J. I.; Mayoral, J. A.; Vispe, E., Optimization of cyclohexene epoxidation
Sienel, G.; Rieth, R.; Rowbottom, K. T., Epoxides. In Ullmann's Encyclopedia of Industrial
with dilute hydrogen peroxide and silica-supported titanium catalysts. Appl. Catal., A. 2003, 245 (2),
363-376.
6.
silsesquioxane complexes as models. J. Catal. 2005, 233 (1), 90-99.
7. Gilbert, L.; Mercier, C., Solvent effects in heterogeneous catalysis : Application to the
synthesis of fine chemicals. Stud. Surf. Sci. Catal. 1993, 78, 51-66.
8. Balducci, L.; Bianchi, D.; Bortolo, R.; D'Aloisio, R.; Ricci, M.; Tassinari, R.; Ungarelli, R., Direct
Fraile, J.; Garcia, J.; Mayoral, J.; Vispe, E., Catalytic sites in silica-supported titanium catalysts:
Oxidation of Benzene to Phenol with Hydrogen Peroxide over a Modified Titanium Silicalite. Angew.
Chem. Int. Ed. 2003, 42 (40), 4937-4940.
9.
Murata, K.; Yanyong, R.; Inaba, M., Effects of vanadium supported on ZrO2 and sulfolane
on the synthesis of phenol by hydroxylation of benzene with oxygen and acetic acid on palladium
catalyst. Catal. Lett. 2005, 102 (3), 143-147.
10.
hydroperoxides: I. A comparison of various metal catalysts. J. Catal. 1973, 31 (3), 427-437.
11. Sheldon, R. A., Synthetic and mechanistic aspects of metal-catalysed epoxidations with
hydroperoxides. J. Mol. Catal. 1980, 7 (1), 107-126.
12. Nguyen, H. G. T.; Schweitzer, N. M.; Chang, C.-Y.; Drake, T. L.; So, M. C.; Stair, P. C.; Farha,
Sheldon, R. A.; Van Doorn, J. A., Metal-catalyzed epoxidation of olefins with organic
O. K.; Hupp, J. T.; Nguyen, S. T., Vanadium-Node-Functionalized UiO-66: A Thermally Stable MOF-
Supported Catalyst for the Gas-Phase Oxidative Dehydrogenation of Cyclohexene. ACS Catal. 2014,
4 (8), 2496-2500.
13.
Perez Ferrandez, D. M.; de Croon, M. H. J. M.; Schouten, J. C.; Nijhuis, T. A., Gas-Phase
Epoxidation of Propene with Hydrogen Peroxide Vapor. Ind. Eng. Chem. Res. 2013, 52 (30), 10126-
10132.
14.
cyclohexene epoxidation by H2O2 over mesoporous TS-1. J. Catal. 2015, 326, 107-115.
15. Bregante, D. T.; Flaherty, D. W., Periodic Trends in Olefin Epoxidation over Group IV and V
Kwon, S.; Schweitzer, N. M.; Park, S.; Stair, P. C.; Snurr, R. Q., A kinetic study of vapor-phase
Framework-Substituted Zeolite Catalysts: A Kinetic and Spectroscopic Study. J. Am. Chem. Soc. 2017,
ACS Paragon Plus Environment

Page 21 of 23
ACS Catalysis
1
2
3
4
5
6
7
139 (20), 6888-6898.
16. Bregante, D. T.; Priyadarshini, P.; Flaherty, D. W., Kinetic and spectroscopic evidence for
reaction pathways and intermediates for olefin epoxidation on Nb in *BEA. J. Catal. 2017, 348, 75-
8
89.
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
17.
Bregante, D. T.; Thornburg, N. E.; Notestein, J. M.; Flaherty, D. W., Consequences of
Confinement for Alkene Epoxidation with Hydrogen Peroxide on Highly Dispersed Group 4 and 5
Metal Oxide Catalysts. ACS Catal. 2018, 8 (4), 2995-3010.
18.
Wang, T. C.; Vermeulen, N. A.; Kim, I. S.; Martinson, A. B. F.; Stoddart, J. F.; Hupp, J. T.; Farha,
O. K., Scalable synthesis and post-modification of a mesoporous metal-organic framework called
NU-1000. Nat. Protoc. 2016, 11 (1), 149-162.
19.
Ahn, S.; Thornburg, N. E.; Li, Z.; Wang, T. C.; Gallington, L. C.; Chapman, K. W.; Notestein, J.
M.; Hupp, J. T.; Farha, O. K., Stable Metal–Organic Framework-Supported Niobium Catalysts. Inorg.
Chem. 2016, 55 (22), 11954-11961.
20.
Thornburg, N. E.; Thompson, A. B.; Notestein, J. M., Periodic Trends in Highly Dispersed
Groups IV and V Supported Metal Oxide Catalysts for Alkene Epoxidation with H2O2. ACS Catal.
2015, 5 (9), 5077-5088.
21.
Thornburg, N. E.; Nauert, S. L.; Thompson, A. B.; Notestein, J. M., Synthesis−Structure–
Function Relationships of Silica-Supported Niobium(V) Catalysts for Alkene Epoxidation with H2O2.
ACS Catal. 2016, 6 (9), 6124-6134.
22.
Schoenfeldt, N. J.; Korinda, A. W.; Notestein, J. M., A heterogeneous, selective oxidation
catalyst based on Mn triazacyclononane grafted under reaction conditions. Chem. Commun. 2010,
46 (10), 1640-1642.
23.
butanediol over ZrO2 catalyst. Catal. Commun. 2005, 6 (7), 480-484.
24. Ichikawa, N.; Sato, S.; Takahashi, R.; Sodesawa, T., Catalytic reaction of 1,3-butanediol over
solid acids. J. Mol. Catal. A: Chem. 2006, 256 (1), 106-112.
25. Duan, H.; Sun, D.; Yamada, Y.; Sato, S., Dehydration of 2,3-butanediol into 3-buten-2-ol
catalyzed by ZrO2. Catal. Commun. 2014, 48, 1-4.
26. Fraile, J. M.; García, J. I.; Mayoral, J. A.; Vispe, E., Catalytic sites in silica-supported titanium
catalysts: silsesquioxane complexes as models. J. Catal. 2005, 233 (1), 90-99.
27. Jiang, J.-X.; Su, F.; Trewin, A.; Wood, C. D.; Niu, H.; Jones, J. T. A.; Khimyak, Y. Z.; Cooper, A.
Yamamoto, N.; Sato, S.; Takahashi, R.; Inui, K., Synthesis of homoallyl alcohol from 1,4-
I., Synthetic Control of the Pore Dimension and Surface Area in Conjugated Microporous Polymer
and Copolymer Networks. J. Am. Chem. Soc. 2008, 130 (24), 7710-7720.
28.
Sarazen, M. L.; Iglesia, E., Effects of Charge, Size, and Shape of Transition States, Bound
Intermediates, and Confining Voids in Reactions of Alkenes on Solid Acids. ChemCatChem 2018, 10
(18), 4028-4037.
29.
Neurock, M.; Manzer, L. E., Theoretical insights on the mechanism of alkene epoxidation
by H2O2 with titanium silicalite. Chemical Communications 1996, (10), 1133-1134.
30.
Sever, R. R.; Root, T. W., DFT Study of Solvent Coordination Effects on Titanium-Based
ACS Paragon Plus Environment

ACS Catalysis
Page 22 of 23
1
2
3
4
5
6
7
Epoxidation Catalysts. Part Two:ꢀ Reactivity of Titanium Hydroperoxo Complexes in Ethylene
Epoxidation. J. Phys. Chem. B 2003, 107 (17), 4090-4099.
31.
Lundin, A.; Panas, I.; Ahlberg, E., Quantum chemical modelling of ethene epoxidation with
8
hydrogen peroxide—role of catalytic sites. Physical chemistry chemical physics : PCCP 2007, 9 (45),
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
5997-6003.
ACS Paragon Plus Environment

Page 23 of 23
ACS Catalysis
1
2
3
4
5
TOC graphic
6
7
8
9
O
OH
OH
Direct
O
H2O2 vapor
TiOx, NbOx
Radical
CO2
HO
O
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
O
H
O
H
O
OH
O
O
S
O
OH
Ti
Ti
Nb
O
O
O
O
O
O
O
O
vapor or solvent
69 kJ/mol or 77 kJ/mol
solvent
48 kJ/mol
vapor
72 kJ/mol
ACS Paragon Plus Environment