Highly selective liquid-phase oxidation of cyclohexane to KA oil over Ti-MWW catalyst: Evidence of formation of oxyl radicals

Article abstract of DOI:10.1021/cs400757j

Various types of Ti-containing zeolites, i.e., Ti-MWW, TS-1, Ti-MOR, and Ti-BEA, have been evaluated as candidates for the liquid-phase oxidation of cyclohexane using t-butyl hydroperoxide (TBHP, 7-8 wt %) as model oxidant. Ti-MWW zeolite displayed the hi

Full text of DOI:10.1021/cs400757j

Subscriber access provided by MICHIGAN STATE UNIVERSITY | MSU LIBRARIES
Article
Highly selective liquid-phase oxidation of cyclohexane to KA
oil over Ti-MWW catalyst: evidence of formation of oxyl radicals
Wen-Juan Zhou, Raphael Wischert, Kai Xue, Yuting Zheng, Belen Albela, Laurent
Bonneviot, Jean-Marc Clacens, Floryan De Campo, Marc Pera-Titus, and Peng Wu
ACS Catal., Just Accepted Manuscript • Publication Date (Web): 21 Nov 2013
Downloaded from http://pubs.acs.org on November 28, 2013
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street
N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 38
ACS Catalysis
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
Highly selective liquidꢀphase oxidation of cyclohexane to KA oil
over TiꢀMWW catalyst: evidence of formation of oxyl radicals
1
,3,4
1
4
3
3
WenꢀJuan Zhou,
Raphael Wischert, Kai Xue, YuꢀTing Zheng, Belén Albela,
3
1
1,2
1,*
Laurent Bonneviot, JeanꢀMarc Clacens, Floryan Decampo, Marc PeraꢀTitus, &
4,*
Peng Wu
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
Eco-Efficient Products and Processes Laboratory (E2P2L), UMI 3464 CNRS – Solvay, 3066 Jin Du Road,
Xin Zhuang Ind. Zone, 201108 Shanghai, China
2
Solvay (China) Co., Ltd. 3966 Jin Du Rd., Xin Zhuang Industrial Zone, Shanghai 201108, PR China
3
Laboratoire de Chimie, UMR 5182 CNRS – Ecole Normale Supérieure de Lyon, 46 allée d’Italie, 69364
10
Lyon, France
4
Shanghai Key Laboratory of Green Chemistry and Chemical Processes, Department of Chemistry, East
China Normal University, Shanghai, China
*
Corresponding author: marc.pera-titus-ext@solvay.com, pwu@chem.ecnu.edu.cn
Abstract
Various types of Tiꢀcontaining zeolites, i.e. TiꢀMWW, TSꢀ1, TiꢀMOR and TiꢀBEA, have been evaluatꢀ
ed as candidates for the liquidꢀphase oxidation of cyclohexane using tꢀbutyl hydroperoxide (TBHP, 7ꢀ8
wt.%) as model oxidant. TiꢀMWW zeolite displays the highest activity for cyclohexanol and cyclohexanone
(KA oil) with an overall selectivity higher than 90% at 80 °C, making this catalyst a candidate of choice for
industrial KA oil production by deperoxidation of cyclohexyl hydroperoxide. The effect of the reaction
temperature, reaction time, catalyst amount and catalyst stability on TiꢀMWW were surveyed in detail. The
TiꢀMWW catalyst showed a stable performance and could be recycled at least 4 times without detectable Ti
leaching and loss of structure stability. The active sites for cyclohexane oxidation appear to be located near
external 12ꢀring cups in the TiꢀMWW framework as suggested by a series of positionꢀselective poisoning
tests with tripropylꢀ and triphenylamine, impelling cyclohexane diffusion within the internal 10ꢀring chanꢀ
nels. EPR experiments supported by DFT calculations suggest a coexistence of competitive heterolytic and
homolytic OꢀO cleavage pathways of Ti(IV)(OOꢀtꢀBu) intermediates in TiꢀMWW depending on the reacꢀ
tion conditions. Homolytic pathways are favored during cyclohexane oxidation, resulting most likely in the
formation of peroxyl and oxyl radicals via bimolecular pathways. The catalytic activity is strongly inhibited
in the presence of alkenes, leading to the preferential formation of the epoxidation product with no detectaꢀ
ble formation of radicals. Notably, this is the first time that oxyl species have been detected particularly
with the help of DFT calculations. Predicted differences of g tensors between peroxyl and oxyl species at
various hydration levels in the presence of cyclohexane were consistent with the EPR spectra.
20
30
Keywords: cyclohexane, oxidation, TiꢀMWW, titanosilicate, KA oil, oxyl radical
ACS Paragon Plus Environment
1

ACS Catalysis
Page 2 of 38
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1. Introduction
Cyclohexanol and cyclohexanone (the mixture is known as KA oil, production >200 ktn /
year) are readily oxidized to adipic acid, an important intermediate in the manufacture of condenꢀ
1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
40 sation polymers, in particular polyamides such as Nylon 6.6. KA oil is industrially produced by
the partial oxidation of cyclohexane in a process involving two consecutive steps: (1) nonꢀ
catalytic autoxidation of cyclohexane by oxygen to cyclohexyl hydroperoxide (CHHP), and (2)
deperoxidation of CHHP to KA oil using a homogeneous Co(II) catalyst. However, this process
suffers from significant drawbacks. In the first step, the selectivity for CHHP is only 60ꢀ70%, with
2
,3
a very low cyclohexane conversion of 3ꢀ4%. The need for recycling the unreacted cyclohexane
leads to high operational costs. The second step is seriously flawed by environmental issues, since
the Co(II) catalyst cannot be completely recovered and recycled, as it is easily discharged with
salts leaving the process. In addition, each kg of KA oil requires about 180 kg of NaOH, generatꢀ
ing in turn ca. 1100 kg of exhausted alkaline solution.
5
0
The above stated economical and environmental drawbacks of the KA oil process could be
certainly mitigated by using a heterogeneous catalyst in the deperoxidation step (step 2). Catalytic
formulations based on noble metals incorporated into oxides, zeolites or aluminophosphates have
4
,5,6
shown high activities and selectivities for KA oil.
However, their high cost, combined with the
rapid deactivation by metal leaching act as a deterrent for their implementation. Other formulaꢀ
7,8,9, 10,11,12
tions based on transition metals (e.g., Cr, Cu, Co, Nb, Ta) only afforded modest yields.
The development of a catalyst with optimal selectivity for KA oil at relatively low CHHP concenꢀ
trations is still today an unsolved challenge today.
Tiꢀcontaining zeolites with isomorphously substituted isolated Ti(IV) sites have shown
13
promising potential in the mild oxidation of alkanes and alkenes. In particular, Tiꢀsilicalite (TSꢀ
14
60
1) with mediumꢀsized pore channels has been widely used for the epoxidation of shortꢀchain
alkenes and the hydroxylation of benzene and phenol using hydrogen peroxide as oxidant. In conꢀ
1
5,16,17,18
19
20
trast, Tiꢀloaded zeolites with larger channels (e.g., TiꢀBEA,
TiꢀZSMꢀ12, TiꢀBEC, Tiꢀ
ACS Paragon Plus Environment
2

Page 3 of 38
ACS Catalysis
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
2
1
22,23,24
24
25
MOR ), as well as mesoporous titanosilicates (e.g., TiꢀMCMꢀ41,
TiꢀSBAꢀ15, TiꢀSBAꢀ16 )
2
6,27,28,29
and more recently materials with hierarchical structures,
appear more adapted for the oxiꢀ
dation of bulky aryls and aromatics using either hydrogen peroxide or organic hydroperoxides.
Despite the anticipated benefits of largeꢀpore titanosilicates for CHHP deperoxidation, sigꢀ
nificant drawbacks prevent their industrial use. First, most of these materials suffer from strong
deactivation due to Tiꢀleaching during operation. Second, the dispersion of Ti sites is often more
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
difficult to control than in the case of TSꢀ1, requiring a fine balance of the hydrolysis rates of the
14,30
7
0
Ti and Si precursors during the synthesis,
and a complete removal of Al and other impurities
3
1
during hydrothermal synthesis. Specific postꢀsynthesis techniques for Tiꢀgrafting in mesopores
3
2,33
or dealumination/titanation in some zeolites,
have been proposed. Third, unlike TSꢀ1, these
materials usually exhibit a prominent hydrophilic behavior due to the abundant presence of silanol
groups (or silanol nests) as a result of incomplete condensation during the synthesis, thus reducing
3
4
the catalytic performance.
In this study, we focus on TiꢀMWW, an isostructural analogue of the aluminosilicate MCMꢀ
3
5,36,37
22, as potential candidate for cyclohexane oxidation.
Our choice is justified for several reaꢀ
sons. First, the TiꢀMWW framework features two independent 10ꢀring channel networks (5.3 Å),
one of which is connected to internal 12ꢀring supercages (7.1×18 Å) and side cups located at the
3
8
80
external surface (7.0 Å). TiꢀMWW combines features of TSꢀ1 and TiꢀBEA, leading to a ‘hybrid’
catalytic behavior between both zeolites: good activity in allyl alcohol and 1ꢀhexene epoxidation,
3
9 ,40
as found for TSꢀ1,
and efficient epoxidation of cyclohexene using tꢀbutyl hydroperoxide
41
(
TBHP), as found for TiꢀBEA. Second, Ti leaching during operation is low compared to Tiꢀbeta
and mesoporous titanosilicates. Third, TiꢀMWW is more hydrophobic than TiꢀBEA and TiꢀMCMꢀ
1, the latter showing a high concentration of silanol groups and defect sites at the external surface.
4
Note that the hydrophobicity of the parent TiꢀMWW can be further increased by treatment with
aqueous piperidine (PI) or hexamethyleneimine (HMI) solutions, reducing the number of internal
42
silanol groups and defect sites by about 40% (i.e. Alꢀfree TiꢀMWW).
ACS Paragon Plus Environment
3

ACS Catalysis
Page 4 of 38
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
In light of the versatile oxidation properties offered by TiꢀMWW, this zeolite might show
potential as deperoxidation catalyst for KA oil production. To this aim, we have chosen the oxidaꢀ
tion of cyclohexane with TBHP as model reaction to compare the oxidation ability of different
catalysts. We compare its catalytic performance with TSꢀ1, TiꢀBEA and TiꢀMOR zeolites, all three
considered as benchmarks. The effect of the reaction temperature, time, catalyst amount and cataꢀ
lytic stability of TiꢀMWW has been explored in detail. The location of the active sites was investiꢀ
gated by selective poisoning of the active sites. Electron paramagnetic resonance (EPR) experiꢀ
ments were combined with DFT calculations to study the reaction mechanism and gain insight
into the nature of the active species during cyclohexane oxidation.
9
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2. Experimental details and computational methods
2
.1. Chemicals
1
00 Four different Tiꢀcontaining zeolites with different effective channel sizes were selected to
assess possible shape selectivity effects for cyclohexane oxidation: TSꢀ1 (5.3 Å), TiꢀMWW (5.3
Å, 7.0 Å), TiꢀMOR (6.7 Å), and TiꢀBEA (7.0 Å). TiꢀMWWꢀ1 (Si/Ti=49) was obtained
commercially from Zhejiang TWRD New Materials Co., Ltd, China, while TiꢀMWWꢀ2
(Si/Ti=67), TiꢀBEA (Si/Ti=63), TSꢀ1 (Si/Ti=32), and TiꢀMOR (Si/Ti=84) zeolites were prepared
on our premises following published recipes (see SI). Table 1 lists the textural properties and Ti/Si
ratios (bulk and surface, noted respectively with ‘b’ and ‘s’ subscripts) of the different samples.
Cyclohexane (>99%, Sinopharm), cyclohexanol (>97%, Sinopharm), cyclohexanone
(>99.5%, Sinopharm), tꢀButyl hydroperoxide (TBHP, 70% in water, Aldrich), benzoquinone
(BQ, >98%, Aldrich), tripropylamine (TPA, >98%, Fluka), triphenylamine (TPhA, 97%, Siꢀ
110 nopharm), triphenylphosphine (TPP, >98.5%, Sinopharm), cyclohexene (>99%, Aladdin), cycloꢀ
hexene oxide (>98%, Aladdin), 1ꢀhexene (>99%, Aldrich), and 1,2ꢀepoxyhexane (HO, 97 %, Alꢀ
drich) were used as reagents. Cyclohexane containing TBHP (7.7 wt.%) was prepared from the
extraction of TBHP from H
2
O with cyclohexane (see SI). 1,2ꢀDichlorobenzene (DCB, >99%, Alꢀ
ACS Paragon Plus Environment
4

Page 5 of 38
ACS Catalysis
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
2 2 3
drich) was used as internal standard. Sodium thiosulfate (Na S O , 0.1 M, Sinopharm), potassium
iodide (KI, 99.9%, Sinopharm), and potassium hydrogen carbonate (KHCO , 99.7%, Sinopharm)
3
were used for the iodometric titration of TBHP.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
.2. Catalytic oxidation of cyclohexane using t-butyl hydroperoxide (TBHP)
A cyclohexane solution (4.0 g) containing TBHP (7.7 wt.%, result from titration) was introꢀ
duced into a sealed tube under air together with the catalyst (0.1 g) on a preheated hotplate at
20 80 °C and subjected to mild stirring for 1 h. After cooling down to room temperature and centrifuꢀ
gation at 7200 rpm for 10 min, the reacted solution was removed from the tube and divided into
two volumes for analysis. One volume was used to measure the TBHP conversion after titration,
while a given amount of 1,2ꢀdichlorobenzene (DCB) was added to the second volume together
with TPP to consume the unreacted TBHP for further analysis. The analyses were carried out in a
GC (Agilent 7820A) equipped with a DBꢀWAX column and a flame ionization detector. Mass balꢀ
ances were accurate to within 5% in all the catalytic tests.
1
The KA selectivity was defined as the molar ratio between the produced KA compared to
the converted TBHP by combining the titration and GC analytical methods, whereas the KA yield
was defined as the molar ratio between the produced KA compared to the initial TBHP. The
130 turnover number (TON) was computed as the molar ratio between the KA formed and the number
of Ti sites after 1 h reaction. Two TON numbers were defined (Table 1): TON referred to the
referred to the surface number of Ti sites (XPS
b
total number of Ti sites (ICP values), and TON
values).
s
Some specific cyclohexane oxidation tests were carried out in the presence of TPA and
TPhA for the TiꢀMWW sample to unveil the location of Ti sites (i.e. over the inner 10ꢀring chanꢀ
nels and 12ꢀring cavities or over the external 12ꢀring cups). Additional experiments were carried
out by adding 1ꢀhexene into the reaction mixture (cyclohexane / 1ꢀhexene=6.3) to gain insight into
the formation of nonꢀradical species on the Ti sites.
ACS Paragon Plus Environment
5

ACS Catalysis
Page 6 of 38
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Catalytic stability tests were carried out at 80 °C during 1 h at a solution/catalyst molar ratio
40 of 40. After each run, the catalyst was washed with cyclohexane, dried at 100 °C, and calcined at
1
5
50 °C for 6 h before the next run.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
.3. Physical characterization
Inductively Coupled Plasma Atomic Emission Spectroscopy (ICPꢀAES) was used to quantiꢀ
fy the overall amount of Si and Ti in the different Tiꢀcontaining zeolites, as well as and in the reacꢀ
tion system using a Thermo IRIS Intrepid II XSP atomic emission spectrometer. Before the measꢀ
3
urements, the samples were dissolved using a HNO /HCl/HF solution.
The zeolite catalysts were characterized by Xꢀray powder diffraction (XRD) using a Rigaku
D/MAX 2200 diffractometer (Cu Kα1+2 radiation).
UVꢀVisible spectra were obtained on calcined zeolite powders using a JASCO Vꢀ670 (JASꢀ
50 CO) spectrophotometer operating in diffuse reflectance mode to assess the nature of Ti species in
the different samples. The measurements were performed in the wavelength range 190ꢀ500 nm.
The spectroteflon standard from JASCO with 82% reflectivity was used as a reference.
1
The textural properties of the different zeolites were inspected by N
2
adsorption at 77.4 K
using a Micrometrics Tristar II apparatus. Prior to the tests, the samples were degassed at 473 K
ꢀ4
under vacuum (10 mbar) for 3 h. The BET specific surface was measured in the pressure range
.05 < P/P <0.35, while total pore volume was measured at P/P =0.99. The external surface was
estimated using the tꢀplot method.
0
0
0
Xꢀray Photoelectron Spectra (XPS) were recorded on a Thermo ESCALAB 250 spectromeꢀ
ter with monochromatic Al Kα radiation (hν=1486.6 eV) to measure the surface Si/Ti molar ratio
60 of the different samples.
1
Electron paramagnetic resonance (EPR) spectra were recorded at 118 K using a Brucker
Elexsys e500 Xꢀband spectrometer (9.42 GHz, 20.25 mW) with a standard cavity. The modulation
amplitude and frequencies were 1 G and 100 kHz, respectively. The spectra were measured at 120
ACS Paragon Plus Environment
6

Page 7 of 38
ACS Catalysis
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
K on an aliquot of the reaction media containing both the solution and the catalyst after cyclohexꢀ
ane oxidation at 80 °C for 20 min. For the sake of comparison, EPR spectra were also measured
after TBHP decomposition on the catalyst and cyclohexene epoxidation with TBHP at 80 °C for
2
2
0 min. Before all the measurements, the reaction was quenched in a liquid N bath (77.4 K).
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
.4. DFT simulations
We used a small cageꢀlike cluster as model for Ti sites in TiꢀMWW. The choice of this
70 model is discussed in the SI. Density functional theory (DFT) calculations were performed with
1
4
3
44
SPARTAN ’10.1 for geometry optimization and energies, and ORCA 2.9.1 for the EPR gꢀ
tensor. Optimization was carried out at the DFT (U)B3LYP/6ꢀ31G(d) level, while single point
energies on the optimized structures were calculated at the (U)B3LYP/6ꢀ311+G(d,p) level, using a
4
5,46,47
Wachters allꢀelectron basis set for Ti augmented by fꢀfunctions.
The reported electronic enꢀ
ergies (ꢁE) were corrected for zero point energy (ZPE). ZPE, enthalpic and entropic contributions
were calculated at the (U)B3LYP/6ꢀ31G(d) level, based on harmonic frequencies scaled by a facꢀ
tor of 0.96. All structures were identified as minima on the potential energy surface by the absence
of imaginary frequencies.
The influence of the solvent (cyclohexane) was checked using the SM8 and COSMO solvaꢀ
180 tion models implemented in SPARTAN and ORCA, respectively. Since the effect on structures
and relative stability of species (less than 10 kJ/mol difference) was small, all hereꢀreported values
are for the gasꢀphase. The EPR gꢀtensor of radical species was calculated on optimized structures
at the level used for single point energies using ORCA’s default settings for the gauge origin (cenꢀ
ter of electronic charge). Using the larger EPRꢀIII basis sets for O and accounting for solvent efꢀ
fects had no significant effect on the results.
3
. Results
3
.1. XRD patterns and Ti distribution
Figure 1 shows the XRD patterns of the TiꢀMWW framework after the synthesis, after
ACS Paragon Plus Environment
7

ACS Catalysis
Page 8 of 38
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
calcination, and after four reaction cycles. The characteristic reflections around 3ꢀ7° in the asꢀ
90 synthesized TiꢀMWW sample reveal the formation of a layered structure along the cꢀdirection,
while the other peaks are related to crystalline sheets parallel to abꢀplanes. After calcination, (001)
and (002) reflections disappeared, whereas (h00) and (hk0) reflections remained essentially
unchanged. This observation can be ascribed to the formation of ordered linkages between sheets,
this process occurring simultaneously with the loss of the lamellar structure upon calcination to
result in the 3D MWW framework.
1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The Ti coordination environment of the two TiꢀMWW samples after calcination was inꢀ
spected by UVꢀvisible spectroscopy (Figure 2). Only one band centred at 220 nm could be obꢀ
4+
2ꢀ
3+
ꢀ
served for TiꢀMWWꢀ2 (Si/Ti =67), which can be attributed to a Ti O → Ti O ligandꢀtoꢀmetal
b
charge transfer and is indicative of tetrahedrally coordinated isolated Ti sites in the MWW frameꢀ
48,49
200 work.
b
However, when the Si/Ti ratio increased to 49 (i.e TiꢀMWWꢀ1), another band centered
at 330 nm appeared, reflecting most likely the presence of exꢀframework Ti species (e.g., octaheꢀ
4
2
dral Ti or anatase sites). Similar UVꢀVisible spectra were observed for TSꢀ1 and TiꢀBEA samples,
while the TiꢀMOR sample only presented apparently tetrahedral Ti sites (SI, Figure S1).
3
.2. Screening of Ti-zeolites
The four different Tiꢀcontaining zeolites prepared in this study were tested in the oxidation
of cyclohexane with TBHP (7.7 wt.% in cyclohexane) at 80 °C for 1.0 h. Among the different
samples, TiꢀMWWꢀ1 (Si/Ti =49) showed the highest selectivity and yield for KA oil, with values
b
of 90.1% and 9.64%, respectively (Table 1). A decrease of the bulk Ti content (i.e. TiꢀMWWꢀ2,
Si/Ti = 67) promoted the TBHP conversion at the expense of a decline of the KA selectivity. We
b
210 tentatively attribute such differences to a larger amount of silanol groups in the latter sample (i.e.
more hydrophilic), promoting TBHP decomposition.
The turnover number for KA oil generation was calculated for the different zeolites as a
b b
function of the bulk Ti content (TON , Table 1). The TON increased in the order: TiꢀMOR (1.5)
ACS Paragon Plus Environment
8

Page 9 of 38
ACS Catalysis
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
< TSꢀ1 (2.1) < TiꢀMWWꢀ2 (7.7) < TiꢀBEA (9.7) < TiꢀMWWꢀ1 (10.0). TSꢀ1 showed much lower
activity than TiꢀBEA, which might be attributed to the small pore entrance (10ꢀring) of TSꢀ1
channels, hindering the diffusion of cyclohexane to the Ti sites. In contrast, the two TiꢀMWW
samples, which feature 12ꢀring side cups in addition to 10ꢀring channels, showed an activity
comparable to that of TiꢀBEA. This observation suggests that, in the case of TiꢀMWW, the
catalytic activity towards cyclohexane oxidation with TBHP is not only related to the effective
20 pore size, but also to the local environment of Ti. Based on these results, we further focused on the
TiꢀMWWꢀ1 sample to investigate the effect of different reaction variables.
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
3
.3. Catalytic properties of Ti-MWW-1
3.3.1.Effect of temperature
To explore the effect of temperature on the catalytic properties of TiꢀMWWꢀ1, a series of
experiments were carried out in the range 80ꢀ140 °C. Table 2 collects TBHP conversion, KA
selectivity and yield. Both TBHP conversion and KA yield increased with the reaction
temperature, while the KA selectivity and A/K molar ratio decreased. At 140 °C, TBHP was
almost fully consumed and the KA yield reached a maximum value of 40%, but at the expense of
a much lower KA selectivity (83.7% at 80 °C compared to 39.8% at 140 °C). The decrease of the
30 KA selectivity was probably caused by the selfꢀdecomposition of TBHP on Ti and silanol sites,
2
5
0
this process being promoted by temperature.
3
.3.2.Effect of the reaction time
The effect of reaction time was surveyed from 1 h to 24 h (Table 3). Both cyclohexane conꢀ
version and KA yield increased slightly with time, while KA selectivity decreased from an initial
value of ca. 90.1% after 1 h to 36.9% after 24 h. These results reflect a higher selfꢀdecomposition
of TBHP with reaction time, cyclohexane oxidation mainly occurring during the first 1 h. In this
period, only cyclohexanol was produced in the very beginning of the reaction, whereas cyclohexaꢀ
none was generated after 20 min.
ACS Paragon Plus Environment
9

ACS Catalysis
Page 10 of 38
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3.3.3.Effect of the catalyst weight
2
40
We also explored the effect of the catalyst weight on the reaction (Table 4). The KA yield inꢀ
creased with catalyst weight until 0.2 g with Ti/TBHP molar ratio of 2.52 %, indicating a higher
specific activity of Ti moieties mostly ascribed to a higher TBHP conversion. In contrast, both KA
selectivity and A/K ratio were found to decrease with catalyst amount, probably due to a longer
residence time of reactants on Ti sites.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
.3.4.Catalytic stability
The catalytic stability of the TiꢀMWWꢀ1 sample was also investigated after 1ꢀ4 runs at 80 °C
for 1 h for a solution/catalyst ratio of 40. The catalyst was dried at 100 °C and calcined at 550 °C
for 6 h before the next run. The catalyst showed a stable performance in terms of TBHP converꢀ
sion and KA selectivity (Figure 3), suggesting no deactivation by either pore blockage or potential
2
50 Ti leaching during operation. This observation also suggests an unmodified distribution of active
intermediate species on the Ti sites during operation.
To confirm the integrity of the sample after reaction, the TiꢀMWWꢀ1 sample was
characterized by XRD and ICP before reaction and after the fifth run. From the ICP analyses, the
Ti content was kept almost unchanged after recycling (Table 1, first row), while no Ti was
detected in the solution. These results support the behavior of TiꢀMWW as a truly heterogeneous
catalyst during cyclohexane oxidation. Reused TiꢀMWWꢀ1 showed no detectable change in the
XRD pattern compared to the calcined samples (Figure 1), confirming the stability of the Tiꢀ
MWW framework during reaction.
3
.3.5.Location of the active Ti sites in Ti-MWW: position-selective poisoning
260
Given the promising results on cyclohexane oxidation reported above over TiꢀMWW, we
decided to carry out some additional experiments and characterization tests to gain insight into the
41
reaction mechanism. On the guidance of a previous study, we first used positionꢀselective poiꢀ
soning reagents to unveil the position of the active Ti sites (i.e. in the inner 10ꢀring channels and
ACS Paragon Plus Environment
1
0

Page 11 of 38
ACS Catalysis
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
12ꢀring supercages, or near 12ꢀring side cups located at the external surface of the MWW frameꢀ
work). To address this issue, we selected two molecular probes: (1) tripropylamine (TPA), which
is expected to enter both the inner 10ꢀring channels and 12ꢀring supercages and consequently poiꢀ
son all the Ti sites, and (2) the bulky triphenylamine (TPhA), which is too large to enter the 10ꢀ
ring channels and thus should selectively poison the active Ti sites located at the external 12ꢀring
cups. In both cases, after reaction, the KA yield showed a dramatic yet comparable decrease (Table
70 5) suggesting that most of the Ti sites participating in the oxidation of cyclohexane with TBHP are
located within the 12ꢀring cups at the external surface of the TiꢀMWW particles.
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
Considering then that cyclohexane oxidation with TBHP over TiꢀMWW proceeds mainly on
1
2ꢀring cups (Scheme 1), we can compute the turnover number of the reaction only relying on the
active external Ti sites measured by XPS (i.e. TON , Table 1). A similar approach was considered
s
for TSꢀ1, while for TiꢀBeta and TiꢀMOR, the total Ti loading (measured by ICP) was considered
since in the latter case the inner Ti sites are expected to be active. Among the different zeolites,
the TiꢀMWWꢀ1 sample showed the highest TON values. We also computed the TON for the Tiꢀ
s
s
MWWꢀ2 sample (Si/Ti
b
=67), displaying a value similar to that obtained for TiꢀMWWꢀ1
(
Si/Ti =49) (Table 1). All these results suggest that the higher activity of TiꢀMWW can be
b
280 attributed to its unique framework including 12ꢀring cups, and also to an improved accessibility of
the Ti sites compared to the other zeolites.
3
.3.6.Competition between cyclohexane oxidation and 1-hexene epoxidation
When the catalytic test was carried out with a mixture of 1ꢀhexene and cyclohexane (cycloꢀ
hexane / 1ꢀhexene = 6.3), the production of KA oil was almost shut down. The main product was
now 1,2ꢀepoxyhexane (HO) which originates from 1ꢀhexene epoxidation (Table 5, first row, value
in parentheses). This finding strongly suggests that 1ꢀhexene and cyclohexane compete for the
same Ti centers.
To further confirm this hypothesis, we repeated the same experiment in the presence of the
ACS Paragon Plus Environment
1
1

ACS Catalysis
Page 12 of 38
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
above stated positionꢀpoisoning agents (TPA and TPhA). Both reaction products, KA oil and HO,
90 were found to largely decrease after TPA addition (Table 5, third row, value in parentheses), the
production of both products was again shut down, but a small yet nonꢀnegligible amount of HO
2
was still produced. This result is to be compared with 1ꢀhexene epoxidation with H
2 2
O over Tiꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
MWW and TSꢀ1, where the reaction is strongly inhibited by TPA, whereas TPhA only reduces
41
slightly the activity. In that case, the active Ti sites are believed to reside in the 10ꢀring channels.
However, TBHP is much bulkier than H O , and thus is not expected to diffuse into the 10ꢀring
2
2
channels, which could explain why in our case both poisoning agents exert a similar effect.
Overall, we can conclude that the same sites on 12ꢀring cups are involved in both alkane oxꢀ
idation and alkene epoxidation. The (quasi)ꢀabsence of reaction in the 10ꢀring channels is in both
cases most probably due to the bulkiness of TBHP, preventing prevents the formation of the active
300 sites.
3
.3.7.EPR investigation
50,51
According to previous studies on homogeneous and heterogeneous catalysts,
cyclohexꢀ
ane oxidation by TBHP appears to proceed via a radical chain mechanism. However, from alkene
5
2,53,54
,55,56,57,58
epoxidation over titanosilicates (most often TSꢀ1),
in situ IR and UV Raman
studies, the formation of nonꢀradical hydroperoxo/peroxoꢀtitanium interꢀ
mediates has been suggested when using H or TBHP as oxidant. In the case of TBHP, the forꢀ
mation of Ti(OOtBu) species has been established in previous reports by a combination of EXAFS
and
5
9,60,61,62
computational
2
O
2
6
3
64,65
(Extended Xꢀray Absorption Fine Structure) and DFT calculations.
These species can be seen
as analogues of Ti(OOH) species formed upon reaction with H
2
2
O .
310
To provide direct evidence for the presence of radical species, a series of rationally designed
electron paramagnetic resonance (EPR) experiments was performed. Considering that the reacꢀ
tants, reaction products and Ti(IV) moieties on TiꢀMWW are all diamagnetic and accordingly
EPRꢀsilent, any EPR signal would indicate the presence of species with unpaired electrons. To this
ACS Paragon Plus Environment
1
2

Page 13 of 38
ACS Catalysis
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
aim, after running the reaction at 80 °C for 20 min, an aliquot of the reaction medium containing
both the solution and the catalyst was transferred to an EPR tube. The reaction was quenched in a
2
liquid N bath (77.4 K) and the EPR spectrum was measured at 120 K. In most cases, an EPRꢀ
active species is characterized by three g values corresponding to the trace of the g tensor. These
values can differ for asymmetric species (vide infra).
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 shows the experimental EPR spectra measured for the TiꢀMWW + TBHP + cycloꢀ
20 hexane system after 20 min reaction at 80 °C (Spectrum A), and for TiꢀMWW + TBHP + acetoniꢀ
trile in analogous reaction conditions, i.e. without cyclohexane (Spectrum B). To guide the asꢀ
signment of g values to a given species, it is important to pay attention to the line width of each
individual peak and its relative intensity. Keeping this idea in mind, the signals in the former specꢀ
3
1
1
1
trum were attributed to two different species with g xx, = 2.0028, g yy= 2.0107 and g zz = 2.026,
2
2
2
and g = 2.0046, g = 2.0138 and g = 2.049, respectively. Finally, a very small signal at g =
xx
yy
zz
2
.0000 is likely due to the cavity or residual organic radicals and will not be discussed further. In
contrast, the direct reaction of TiꢀMWW with TBHP in acetonitrile produced a broad and thus
1
1
1
markedly different EPR signal with g ’xx = 2.0028, g ’yy = 2.0107 and g ’zz = 2.020 (Spectrum B).
2
2
2
A second species appears to be observed with g ’xx = 2.0046, g ’yy = 2.0151 and g ’zz = 2.034.
30 More details about the nature of both species and their potential formation pathways are provided
below.
3
Additional experiments were carried out to check whether the reactants and the catalyst (i.e.
TBHP, TiꢀMWW + cyclohexane) could produce an EPR signal. In all cases (not shown), the difꢀ
ferent systems were inactive in EPR, even if dissolved oxygen (paramagnetic) was present in the
samples. Finally, cyclohexene oxidation over TiꢀMWW using TBHP was also monitored by EPR
in the same reaction conditions. As expected for a nonꢀradical reaction, no EPR signal was obꢀ
served, while the yield of cyclohexene oxide kept similar with or without BQ (Table 6), confirmꢀ
ing the consistency of the EPR spectra for cyclohexane oxidation.
4
. Discussion
ACS Paragon Plus Environment
1
3

ACS Catalysis
Page 14 of 38
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
340
From the experimental results reported above, as far as the reaction mechanism is concerned,
two main conclusions emerge: i) cyclohexane oxidation takes place on Ti sites mainly located in
the external 12ꢀring cups, and ii) the reaction proceeds through radical intermediates. The epoxidaꢀ
tion of 1ꢀhexene, which was studied for comparison, takes place on the same external Tiꢀsites, but
it does not involve radical intermediates. Accordingly, radical species appear to coexist with nonꢀ
radical intermediates on Ti sites.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Apart from the aboveꢀmentioned implication of radical species, little is known about the
mechanism of alkane oxidation in Tiꢀcontaining zeolites. To gain insight into the reaction mechaꢀ
nism and assist the assignement of the EPR spectra, we investigated the stability of potential radiꢀ
cal species and calculated their EPRꢀspectroscopic signature by DFT. In particular, the striking
350 differences between the EPR spectra obtained in the presence and in the absence of cyclohexane
triggered such investigation.
4
.1. Formation of Ti-centered radical species: DFT calculations
2
We chose (SiO)
3
Ti(η ꢀOOtBu)(H
2
O) as starting point for the study of potential radical inꢀ
termediates (Scheme 2, reaction i). Its formation is discussed in the SI (Figure S2), since our reꢀ
61
sults are very similar to literature data. Ti is fiveꢀfold coordinated in the starting structure. Such
extension of the coordination sphere by water (which is formed by the reaction of TiOH with
tBuOOH) is generally accepted and both sixꢀfold and more recently fiveꢀfold coordination have
6
6
been suggested. Note that a sixꢀfold coordination does not significantly affect the trends outlined
here (see Scheme S2 and Figure S2).
360
The homolytic fragmentation of Ti(OOtBu) species can yield in principle three species: i)
ꢀ
ꢀ
Ti(III), ii) Ti(IV)ꢀOO (peroxyl, Figure5a), and iii) Ti(IV)ꢀO (oxyl, Figure5b). Ti(III) species on
Tiꢀcontaining zeolites have only been observed in rather reductive media either in the presence of
6
7,68
69
H
2
or CO,
or hydrides. The occurence of peroxyl radicals in Tiꢀcontaining silicalites in conꢀ
70, 71,72,73,74
ꢀ
tact with H
2
O
2
or organic peroxides has been established by EPR.
Ti(IV)ꢀO (oxyl) speꢀ
ACS Paragon Plus Environment
1
4

Page 15 of 38
ACS Catalysis
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
7
5
cies, although already proposed by Clerici in the last decade, have almost never been considered.
Indeed, only recently its formation in the presence of H O was investigated by DFT calculaꢀ
2
2
42
tions. This is somewhat surprising, since Ti(OOR) can be seen as an analogue of HOOR (R = H,
2 2
alkyl) and breaking the OꢀO bond is by far the easiest source of radicals from both H O and
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
tBuOOH (Scheme S1).
ꢀ
370
Ti(III) can be formed by homolytic splitting of the TiꢀO bond, yielding a tBuOO radical
(Scheme 2, reaction ii). However, this process is highly unfavorable (ꢁE = +249 kJ/mol), even
when fully stabilized by water (see SI).
ꢀ
Ti(IV)ꢀOO (peroxyl) can be formed by loss of a tꢀbutyl radical (Scheme 2, reaction iiia), but
ꢀ
this process is likewise unfavorable (ꢁE = +205 kJ/mol). In a similar fashion, Ti(IV)ꢀO can be
ꢀ
formed by the loss of a tBuO radical (Scheme 2, reaction iva). As expected from the weakness of
the OꢀO bond, this process is more favorable, but still high in energy (ꢁE = +161 kJ/mol). Since
the fragmentations are entropically favored, accounting for such contribution leads to a significant
7
6
stabilization of all reactions (ꢁG298 is on average 60 kJ/mol lower than ꢁE). Still, the direct
fragmentations presented so far cannot probably account for the species observed in the experiꢀ
3
80 ments.
77
64
However, as suggested by Clerici, and recently calculated for H O , bimolecular pathꢀ
2
2
ways implying the participation of a second peroxide molecule could explain the formation of the
ꢀ
ꢀ
Ti(IV)ꢀO (oxyl) species. Indeed, such a reaction, yielding a tBuOO radical and tBuOH (as experꢀ
imentally observed), is much lower in energy (reaction ivb in Scheme 2, ꢁE = +54 kJ/mol) and
even favorable in terms of Gibbs free energy (ꢁG298 = –8 kJ/mol). The reaction (ꢁE) energy is
ꢀ
close to the value calculated for Ti(IV)ꢀO formation starting from Ti(OOH) and H O (+40
2
2
64
kJ/mol), being also consistent with the similar OꢀO bond dissociation energy for both peroxides
ꢀ
(Scheme S1). This suggests that Ti(IV)ꢀO radical species are thermodynamically accessible. Note
also that the reaction was carried out experimentally at 80ºC, favoring even more the thermodyꢀ
90 namics of the process. Along these lines we can also propose a bimolecular dissociation to yield
3
ACS Paragon Plus Environment
1
5

ACS Catalysis
Page 16 of 38
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ꢀ
Ti(IV)ꢀOO (peroxyl) species. In stark contrast to the unimolecular fragmentation, this process is
favorable (ꢁE = +15 kJ/mol, ꢁG298 = –63 kJ/mol). In summary, bimolecular pathways can rationꢀ
ꢀ
ꢀ
alize from a thermodynamic viewpoint the formation of Ti(IV)ꢀOO and Ti(IV)ꢀO species.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
.2. Interpretation of the EPR spectra
The calculated EPR gꢀtensor of the aboveꢀdiscussed energetically accessible radical species
is reported together with our experimental and literature data in Table 7. Additional peak data obꢀ
tained after deconvolution can be found in the SI (Figure S4, Table S2). Since hydration affects the
7
8
structure of the Ti site, we expect an effect on the gꢀtensor. Therefore, we modeled different
states of hydration (SI, Figure S3). Indeed, previous experimental studies on the TSꢀ1/H
2
O
2 2
/ H O
ꢀ
4
00 system reported in most cases the formation of two different Tiꢀperoxyl radicals, Ti(IV)ꢀOO , that
70
can be attributed to species with different number of coordinated water molecules, species
7
1
formed in framework and exꢀframework Ti moieties, or species formed in tetrapodal (external)
72,73,74
65
or tripodal (internal) Ti sites.
Furthermore, as pointed out by Wells et al., we do not exꢀ
clude stabilization by surface silanol groups (defects). In our case, the experiments carried out in
the sole presence of TBHP were subjected per se to a much higher hydration level than those in
the presence of cyclohexane, since in the former experiments TBHP was highly diluted in water
(ca. 30 wt.%) to ensure its stability. In the second case, even if water was present due to traces afꢀ
ter TBHP extraction with cyclohexane and also generated during cyclohexane oxidation, the hyꢀ
dration level is expected to be much lower due to the low reaction time used (20 min).
410
The EPR signals referred to as Species 1-A and Species 1-B in the TiꢀMWW + TBHP + cyꢀ
clohexane and TiꢀMWW + TBHP + acetonitrile systems (Figure 4, Spectra A and B, respectively)
1
1
show very similar g and g ’ values within the limits of the experimental error (Table 7), even if
the Ti centers in the second system are expected to show a higher degree of hydration. This obserꢀ
ꢀ
vation matches the trends predicted by DFT for Ti(IV)ꢀOO species, the corresponding g values
being confined in a very narrow range (average over all hydration states: gxx = 2.0029, gyy
=
ACS Paragon Plus Environment
1
6

Page 17 of 38
ACS Catalysis
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
2.0095, gzz = 2.0263 in Table 7) and essentially insensitive to hydration, except for gzz, which varꢀ
ies slightly. Relying on these calculations and the experimental data from the literature (Species 1
ꢀ
and 1’ in Table 7), Species 1ꢀA and Species 1-B can be assigned both to Ti(IV)ꢀOO , but their state
of hydration is difficult to ascertain, since the effect of hydration is weak.
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
4
20
The EPR spectrum of the reaction mixture (Figure 4, Spectrum A) shows an additional sigꢀ
ꢀ
nal (Species 2-A) which cannot be attributed to Ti(IV)ꢀOO or any species reported so far in the
ꢀ
literature. However, we find a very good agreement with the gꢀvalues calculated for Ti(IV)ꢀO
ꢀ
species. All principal values calculated for Ti(IV)ꢀO species are significantly higher than those
ꢀ
found for Ti(IV)ꢀOO (average over all hydration states: gxx = 2.0045, gyy = 2.0132, gzz = 2.0424).
While gxx is almost not affected by hydration and gyy varied only slightly, gzz is strongly affected
ꢀ
by hydration, much more than in the case of Ti(IV)ꢀOO . In light of these results, Species 2-A and
ꢀ
Species 2-B might be tentatively assigned to different hydrated forms of Ti(IV)ꢀO . In spectrum B,
ꢀ
it is more difficult to argue for the presence of such TiꢀO species, most likely due to a masking
ꢀ
ꢀ
effect of water and thus hydration. Indeed, the predicted g values for Ti(IV)ꢀOO and Ti(IV)ꢀO ,
430 most specifically gyy and gzz, are relatively close when 2 or more water molecules are coordinated
ꢀ
to the Ti center (Table 7, Figure 5). This effect of hydration might explain why the Ti(IV)ꢀO speꢀ
cies is hardly detectable in Spectrum B in Figure 4 and also in analogous spectra published in the
literature, typically measured in aqueous or strongly hydrated conditions.
Note finally that Ti(III) is predicted at gꢀvalues (<2.0000) not compatible with experiment
(
SI, Table S1), which is consistent with its unfavorable formation.
5
. Conclusions
We have shown that TiꢀMWW zeolite exhibits a much higher selectivity to KA oil during the
oxidation of cyclohexane with TBHP than other titanium zeolites such as TSꢀ1, TiꢀBeta, and Tiꢀ
MOR. The higher performance of TiꢀMWW is attributed to its unique structure providing accessiꢀ
40 ble Ti sites on external 12ꢀring cups. Unlike the reaction time and the catalyst weight, the temperaꢀ
4
ACS Paragon Plus Environment
1
7

ACS Catalysis
Page 18 of 38
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ture exerted a remarkable effect on the catalytic performance. TiꢀMWW exhibited a high stability
in the reaction and could be reused for at least four runs keeping a similar activity and with no deꢀ
tectable Ti leaching.
In the simultaneous presence of alkenes, competition between radical (alkane oxidation) and
nonꢀradical (alkene epoxidation) mechanisms took place, suggesting that both reactions share the
same precursor of the active site, presumably TiꢀOOtBu species.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In situ EPR experiments combined with DFT calculations suggested the formation of at least
two Tiꢀcentered radical species in our reaction system. One of the signals could be unambiguously
ꢀ
assigned to peroxyl species, Ti(IV)ꢀOO , as frequently reported in the literature. In contrast, the
450 other signal was not compatible with such species, but provided instead an excellent match with
ꢀ
the EPR gꢀtensor values calculated for oxyl radicals, Ti(IV)ꢀO . Although oxyl radicals have been
proposed as active centers for alkane oxidation, to the best of our knowledge they have not been
observed so far. However, opposing to to the published literature, our spectra have been recorded
in anhydrous condition, and indeed our calculations suggest that strong hydration makes the disꢀ
ꢀ
ꢀ
tinction between Ti(IV)ꢀOO and Ti(IV)ꢀO difficult. Both radical intermediates might be generatꢀ
ed via bimolecular pathways as a reminiscence of a homolytic OꢀH and OꢀO cleavage, respectiveꢀ
ly.
We believe that the high catalytic performance of our system should motivate further developꢀ
ꢀ
ment of Tiꢀbased oxidation catalysts. In addition, the possible observation of Ti(IV)ꢀO species
4
60 opens up new mechanistic perspectives for alkane oxidation, which, contrasting to alkene epoxidaꢀ
ꢀ
tion, is still poorly understood. Further investigation of the presence of Ti(IV)ꢀO in other Tiꢀ
containing zeolites is under study.
Acknowledgements
WꢀJ.Z thanks China Postdoctoral Science Foundation (grants 2013M541493). P.W gratefully
acknowledges NSFC (grants 20925310, U1162102, 21373089) and the Ph.D. Program Foundation
ACS Paragon Plus Environment
1
8

Page 19 of 38
ACS Catalysis
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
of the Ministry of Education of China (grant 2012007613000).
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
ACS Paragon Plus Environment
1
9

ACS Catalysis
Page 20 of 38
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
FIGURE CAPTIONS
th
Figure 1. XRD patterns of TiꢀMWW (Si/Ti
70 use.
b
=49): (a) asꢀsynthesized, (b) after calcination, and (c) after 4
4
b b
Figure 2. UVꢀvisible spectra of (a) TiꢀMWW (Si/Ti =67) and (b) TiꢀMWW (Si/Ti =49) after calcination.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Catalytic results of TiꢀMWWꢀ1 in the oxidation of cyclohexane for 4 repeated consecutive runs
with regeneration.
Figure 4. EPR spectra at 120 K of the reaction mixture after 20 min, i.e. TiꢀMWW + cyclohexane + TBHP
(
A), and TiꢀMWW + acetonitrile + TBHP (B). g refers to the top spectrum, while g’ refers to the bottom
spectrum. Color code: blue for species 1 and orange for species 2 in Table 7.
2
Figure 5. Representative radical species as obtained from DFT calculations: a) TiꢀOOꢂ ·1H O, b) TiꢀOꢂ
2
·2H O. Distances are given in Å. Italics are used for Hꢀbonding distances. Nomenclature: O, red ball; H,
light grey; Ti, green ball; Si, grey ball.
4
80 SCHEME CAPTIONS
Scheme 1. Graphic illustration of KA oil formation on 12ꢀring side cups of TiꢀMWW.
2
Scheme 2. i) Formation of Tiꢀ(η ꢀOOtBu)(H
and bimolecular cleavage of Tiꢀη (OOtBu)(H
(and tBuOOH in the case of bimolecular pathways). The full structures are reported in the SI.
2
O). iiꢀiv) Formation of potential radical intermediates via uniꢀ
O). Energies (ꢁE, kJ/mol) are given relative to this species
1
2
TABLE CAPTIONS
*
Table 1. Physicochemical properties and cyclohexane oxidation results of various titanosilicate catalysts
*
Table 2. Effect of reaction temperature on cyclohexane oxidation over TiꢀMWW
*
Table 3. Effect of reaction time on cyclohexane oxidation over TiꢀMWW
*
Table 4. Effect of catalyst amount on cyclohexane oxidation over TiꢀMWW
4
90 Table 5. Positionꢀselective poisoning reagents (PR) over TiꢀMWW for cyclohexane oxidation before and
after introduction of 1ꢀhexene (in parentheses).
*
Table 6. Effect of radical scavengers
Table 7. Comparison of EPR gꢀvalues of different Ti species from experiment and DFT calculations.
ACS Paragon Plus Environment
2
0

Page 21 of 38
ACS Catalysis
Submitted to ACS Catal.
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.
ACS Paragon Plus Environment
2
1

ACS Catalysis
Page 22 of 38
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2.
500
ACS Paragon Plus Environment
2
2

Page 23 of 38
ACS Catalysis
Submitted to ACS Catal.
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 3.
ACS Paragon Plus Environment
2
3

ACS Catalysis
Page 24 of 38
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4.
ACS Paragon Plus Environment
2
4

Page 25 of 38
ACS Catalysis
Submitted to ACS Catal.
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
Required parameters
are missing or incorrect.
Figure 5.
510
ACS Paragon Plus Environment
2
5

ACS Catalysis
Page 26 of 38
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 1.
ACS Paragon Plus Environment
2
6

Page 27 of 38
ACS Catalysis
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
O
H O
2
OH
Ti
O
HOO
-19
(i)
+
Ti
OH2
Ti
+
249
(III)
ii)
+
+
OO
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
H O O O
2
iiia)
+
+205
Ti
+161
+54
H O
O
Ti
2
O
O
iva)
ivb)
H O
2
O
Ti
H O
O
Ti
2
H O O O
2
iiib)
+
+
OO
HO
+
O
Ti
+
15
0
+
HO
HOO
HOO
Scheme 2.
ACS Paragon Plus Environment
2
7

ACS Catalysis
Page 28 of 38
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 1.
2
†
Specific surface (m /g)
Si/Ti ratio
Turnover number (ꢀ)
Pore size
(Å)
TBHP
Conv. % KA % KA %
Sel.
Yield
Catalyst
Si/Tis
A/K
Si/Ti
b
BET
External
TONb
TONs
(
ICP) (XPS)
‡
TiꢀMWWꢀ1 5.3, 7.0
TiꢀMWWꢀ2 5.3, 7.0
435
425
329
557
ꢀ
132
129
193
93
49(51 )
67
53
92
33
52
ꢀ
10.7
14.1
10.1
42.4
0.2
90.1
49.1
28.5
17.6
98.1
9.6
6.9
2.9
7.5
0.2
8.4
15.6
1.0
∞
10.0
7.7
2.1
9.7
1.5
47.4
44.5
4.8
ꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
TSꢀ1
5.3
6.7
7.0
32
TiꢀBEA
TiꢀMOR
63
ꢀ
84
∞
ꢀ
†
‡
TON
TON
b
= KA yield (mol) / total Ti site (mol) in titanosilicate (1 h)
= KA yield (mol) / external Ti site (mol) in titanosilicate (1 h)
s
5
20 After the reaction
*
Reaction conditions: 80 °C, 1.0 h, 0.10 g catalyst, 4 g of cyclohexane containing 7.7 wt.% TBHP
ACS Paragon Plus Environment
2
8

Page 29 of 38
ACS Catalysis
Submitted to ACS Catal.
1
2
3
4
5
6
7
8
9
Table 2.
Cyclohexane
conv. %
TBHP
KA
KA
‡
°
T ( C)
†
A/K
TON
b
conv. % sel. % yield %
80
0.8
1.5
1.9
3.0
11.7
24.8
65.8
96.5
83.7
75.9
36.2
39.8
9.8
6.4
3.9
2.4
2.4
8.4
100
18.8
23.8
38.4
16.3
21.7
33.1
1
1
20
40
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
†
Cyclohexane conversion was calculated based on molar percentage of converted KA oil to
original cyclohexane
‡
b
TON = KA yield (mol) / total Ti site (mol) in TiꢀMWW
*
Reaction conditions: 80 °C, 1.5 h, 0.10 g catalyst, 4 g of cyclohexane containing 7.7 wt.% TBHP
ACS Paragon Plus Environment
2
9