Unveiling the Active Surface Sites in Heterogeneous Titanium-Based Silicalite Epoxidation Catalysts: Input of Silanol-Functionalized Polyoxotungstates as Soluble Analogues

Article abstract of DOI:10.1021/acscatal.8b00256

We report on a site-isolated model for Ti(IV) by reacting [Ti(ⁱPrO)₄] with the silanol-functionalized polyoxotungstates [XW₉O_34-x(^tBuSiOH)₃]^3- (X = P, x = 0, 1; X = Sb, x = 1, 2) in tetrahydrofuran. The resulting titanium(IV) complexes [XW₉O_34-x(^tBuSiO)₃Ti(OⁱPr)]^3- (X = P, 3; X = Sb, 4) were obtained in monomeric forms both in solution and in the solid state, as proved by diffusion NMR experiments and by X-ray crystallographic analysis. Anions 3 and 4 represent relevant soluble models for heterogeneous titanium silicalite epoxidation catalysts. The POM scaffolds feature slight conformational differences that influence the chemical behavior of 3 and 4 as demonstrated by their reaction with H₂O. In the case of 3, the hydrolysis reaction of the isopropoxide ligand is only little shifted toward the formation of a monomeric [PW₉O₃₄(^tBuSiO)₃Ti(OH)]^3- (5) species [log K = -1.96], whereas 4 reacted readily with H₂O to form a μ-oxo bridged dimer {[SbW₉O₃₃(^tBuSiO)₃Ti]₂O}^6- (6). The more confined was the coordination site, the more hydrophobic was the metal complex. By studying the reaction of 3 and 4 with hydrogen peroxide using NMR and Raman spectroscopies, we concluded that the reaction leads to the formation of a titanium-hydroperoxide Ti-(η¹-OOH) moiety, which is directly involved in the epoxidation of the allylic alcohol 3-methyl-2-buten-1-ol. The combined use of both spectroscopies also led to understanding that a shift of the acid-base equilibrium toward the formation of Ti(η²-O₂) and H₃O⁺ correlates with the partial hydrolysis of the phosphotungstate scaffold in 3. In that case, the release of protons also catalyzed the oxirane opening of the in situ formed epoxide, leading to an increased selectivity for 1,2,3-butane-triol. In the case of the more stable [SbW₉O₃₃(^tBuSiO)₃Ti(OⁱPr)]^3- (4), the evolution to Ti(η²-O₂) peroxide was not detected by Raman spectroscopy, and we performed reaction progress kinetic analysis by NMR monitoring the 3-methyl-2-buten-1-ol epoxidation to assess the efficiency and integrity of 4 as precatalyst.

Full text of DOI:10.1021/acscatal.8b00256

Subscriber access provided by READING UNIV
Article
Unveiling the Active Surface Sites in Heterogeneous Titanium–
Based Silicalite Epoxidation Catalysts: the Input of Silanol–
Functionalized Polyoxotungstates as Soluble Analogues
Teng Zhang, Louis Mazaud, Lise-Marie Chamoreau, Celine Paris, Anna Proust, and Geoffroy Guillemot
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.8b00256 • Publication Date (Web): 05 Feb 2018
Downloaded from http://pubs.acs.org on February 5, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street
N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 47
ACS Catalysis
1
2
3
4
5
6
7
8
9
Unveiling the Active Surface Sites in Heterogeneous
Titanium–Based Silicalite Epoxidation Catalysts: the
Input of Silanol–Functionalized Polyoxotungstates
as Soluble Analogues
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
Teng Zhang,† Louis Mazaud,† LiseꢀMarie Chamoreau,† Céline Paris,‡ Anna Proust† and
Geoffroy Guillemot†*
† Sorbonne Université, CNRS, Institut Parisien de Chimie Moléculaire, IPCM, 4 place Jussieu,
Fꢀ75005 Paris, France
‡ Sorbonne Université, CNRS, De la Molécule aux Nanoꢀobjets: Réactivité, Interactions et
Spectroscopies, MONARIS, 4 place Jussieu, Fꢀ75005 Paris, France
ABSTRACT.
We report on a site–isolated model for Ti(IV) by reacting [Ti(iPrO)4] with the silanol–
functionalized polyoxotungstates [XW9O34ꢀx(tBuSiOH)3]3ꢀ (X= P, x=0, 1; X= Sb, x=1, 2) in
tetrahydrofuran. The resulting titanium(IV) complexes [XW9O34ꢀx(tBuSiO)3Ti(OiPr)]3ꢀ (X= P, 3;
X= Sb, 4) were obtained in monomeric forms both in solution and in the solid state, as proved by
diffusion NMR experiments and by X–ray crystallographic analysis. Anions 3 and 4 represent
ACS Paragon Plus Environment
1

ACS Catalysis
Page 2 of 47
1
2
3
4
5
6
7
8
9
relevant soluble models for heterogeneous titanium silicalite epoxidation catalysts. The POM
scaffolds feature slight conformational differences that influence the chemical behavior of 3 and
4 as demonstrated by their reaction with H2O. In the case of 3, the hydrolysis reaction of the
isopropoxide ligand is only little shifted towards the formation of a monomeric
[PW9O34(tBuSiO)3Ti(OH)]3ꢀ (5) species [log K= –1.96], whereas 4 reacted readily with H2O to
form a ꢀꢀoxo bridged dimer {[SbW9O33(tBuSiO)3Ti]2O}6ꢀ (6). The more confined the
coordination site, the more hydrophobic the metal complex. By studying the reaction of 3 and 4
with hydrogen peroxide using NMR and Raman spectroscopies, we concluded that the reaction
leads to the formation of a titaniumꢀhydroperoxide Tiꢀ(η1–OOH) moiety, which is directly
involved in the epoxidation of the allylic alcohol 3ꢀmethylꢀ2ꢀbutenꢀ1ꢀol. The combined use of
both spectroscopies also led to understanding that a shift of the acidꢀbase equilibrium towards the
formation of Ti(η2ꢀO2) and H3O+ correlates with the partial hydrolysis of the phosphotungstate
scaffold in 3. In that case, the release of protons also catalyzed the oxirane opening of the in situ
formed epoxide, leading to an increased selectivity for 1,2,3ꢀbutaneꢀtriol. In the case of the more
stable [SbW9O33(tBuSiO)3Ti(OiPr)]3ꢀ (4), the evolution to Ti(η2ꢀO2) peroxide was not detected
by Raman spectroscopy and we performed reaction progress kinetic analysis by NMR
monitoring the 3ꢀmethylꢀ2ꢀbutenꢀ1ꢀol epoxidation in order to assess the efficiency and integrity
of 4 as precatalyst.
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
KEYWORDS. Titanium, hydrogen peroxide, polyoxotungstate, silanol, site–isolated catalysts,
epoxidation
1. INTRODUCTION
ACS Paragon Plus Environment
2

Page 3 of 47
ACS Catalysis
1
2
3
4
5
6
7
8
9
Substitution of Ti for Si in the framework of zeolites, first reported by Taramasso et al.,1
resulted in the emergence of heterogeneous catalysts of practical use for the liquid–phase
oxidation of various organic compounds at industrial scale.2,3 The use of titanium, a first–row
transition metal, and aqueous hydrogen peroxide, as terminal oxidant, make sense in view of the
need to develop more sustainable oxidation processes.4,5,6 The catalytic performances of
microporous and crystalline titanium–silicalites (TS–1, a MFI zeolitic framework) are strictly
associated to the structural peculiarities of the silica lattice.7 First, the hydrophobic nature of the
internal channel system favors the strong physisorption of the apolar substrates while preventing
the hydrolysis of isolated titanium center. Conversely, the hydrophilicity of macroporous or
amorphous Ti/SiO2 more often results in leaching. Second, the metal ions are sufficiently held in
the silica matrix, as a tripodal anchored species, to prevent conversion into inactive peroxide
Ti(η2ꢀO2) species.8,9 Indeed, the peracidꢀlike mechanism is widely accepted for TS–1 catalyzed
epoxidation, in which a hydroperoxo (Ti–OOH) rather than a peroxide species is involved.
Modeling such a system at a molecular scale is rather difficult. One should mention the titaniumꢀ
containing polyhedral oligomeric silsesquioxanes (POSS), probably the most relevant example,
which enable the coordination of Ti in a tripodal siloxy coordination site.10,11,12 Nevertheless the
POSS ligands remain flexible enough to favor the expansion of the coordination shell and to
easily promote the formation of dimers and also the generation of small embedded
polyoxotitanate clusters [TiOH]4 when in contact with water.11 Accordingly the site–isolated
model, as in heterogeneous systems, no longer applies.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A valuable molecular approach to catalytically active surface sites should fulfill the following
criteria: the model system (i) should display a single isolated site with a geometrical and
chemical local environment similar to the active site, (ii) should allow the successful
ACS Paragon Plus Environment
3

ACS Catalysis
Page 4 of 47
1
2
3
4
5
6
7
8
9
achievement of the same kind of transformations through a closely related chemical pathway.
Moreover, a relevant molecular model should help to bring insights on the structure/activity
relationships. The silanol–decorated polyoxotungstates (POTs) that are depicted in Figure 1 have
proved to be capable of modeling surface vicinal silanols of dehydroxylated silica.13 These
organic hybrids of POTs may also best mimic the coordination environment of the tetrahedral
defective openꢀlattice [(≡Si–O)3Ti(OH)] sites14,15 that are present in the microporous TS–1. First,
the polyoxotungstic frameworks display a rigid and geometrically preꢀorganized set of silanol
functionalities. Second, the tꢀbutyl groups at the silicon atoms create a steric protection around
the metal coordination site. The combination of these two factors prevents the formation of
oligomers and allows the metal ion to fit in a well–defined isolated coordination site. We thus
turned our attention to the synthesis of titanium complexes and their use as catalysts for the
oxidation of alkenes by hydrogen peroxide. Indeed, most of the soluble titanium–silsesquioxane
complexes so far reported were unsuitable for the use of aqueous solutions of hydrogen peroxide.
Both irreversible hydrolysis of the SiO–Ti bonds and water inhibition of the active sites have
been put forward to explain this lack of reactivity, which stands as a main drawback considering
the attractive use of aqueous H2O2 solution as benign oxidant. Hence, different methods have
been developed in order to immobilize titanium silsesquioxane complexes in a hydrophobic
environment. Heterogeneous TiꢀPOSS systems have been prepared by grafting on a threeꢀ
dimensional netted polysiloxane or by integrating into a SBAꢀ15 supported polystyrene film by
in situ copolymerization.16,17 Encapsulation of TiꢀPOSS complexes in a dimethylsiloxane
membrane has also been reported.18 Despite the difficulty to control the threeꢀdimensional
structure of the polymer around the titanium complexes by these methods, it was established that
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
4

Page 5 of 47
ACS Catalysis
1
2
3
4
5
6
the hydrophobic network provided by the polymer enables the catalysts to display good H2O2
efficiency in alkene epoxidation and to be highly recyclable.
7
8
9
Polyoxometalates (POMs), which encompass the POTs family, are oxo clusters of early
transition metals in their highest oxidation states (MoVI, WVI, VV). POMs are receiving much
interest due to the great versatility of their molecular structures and properties. Usually
considered as molecular analogues of metal oxides, POMs are robust, oxidatively and thermally.
Among the numerous applications of POMs, catalysis surely occupies the first place: this
concerns mainly acid catalysis due to the superacidity of heteropolyacids19 and oxidation
catalysis,20 based on the intrinsic redox properties of POMs21 or the incorporation of active
transition metals.22 The organic–inorganic hybrids 1 and 2 used in this study are built on two
different types of trivacant Keggin heteropolyoxotungstates: [PW9O34]9ꢀ and [SbW9O33]9ꢀ (see
Figure 1). The first one belongs to the Aꢀtype [X(W3O10)(W2O8)3]nꢀ usually built around a
tetrahedral X heteroatom such as P(+V) and the second one to a Bꢀtype [Y(W3O11)3]9ꢀ commonly
obtained with pyramidal Y heteroatoms such as Sb(+III).23 The organic functionalization of these
species (by organosilanes in our case) is made possible by the enhanced nucleophilicity of the
oxygen atoms which surround the vacancy.24 For all the aboveꢀmentioned considerations,
derivatives 1 and 2 represent good candidates for exploring oxidation catalysis by their
corresponding titanium derivatives. Hereafter, we first focus on the description of the structural
analogies with silicaꢀsupported titanium materials in order to validate our molecular model of
heterogeneous systems. Secondly, we report our investigation on the catalytic activity of the
titanium derivatives in the epoxidation of alkenes and allylic alcohols with H2O2. Our efforts
have mainly focused on the identification of the active Tiꢀhydroperoxide intermediate and on
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
5

ACS Catalysis
Page 6 of 47
1
2
3
4
5
6
assessing the efficiency and integrity of the catalysts by performing reaction progress kinetic
analysis by monitoring the oxidation reaction by NMR.
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. Anions 1 and 2 used in this study, prepared as tetrahexylammonium salts (THA). The
polyoxotungstate frameworks are shown in polyhedral representation: tungsten in blue,
phosphorus in green, antimony in magenta and oxygen atoms bordering the lacuna as red dots.
2. RESULTS AND DISCUSSION
2.1. The titanium complexes.
Synthesis.
The
synthesis
of
[(αꢀAꢀPW9O34)(tBuSiOH)3]3ꢀ
(1)
and
[(αꢀBꢀ
SbW9O33)(tBuSiOH)3]3ꢀ (2) as tetrabutylammonium salts has been already discussed elsewhere.25
We made use of quaternary ammonium cations with longer alkyl chains to enlarge the variety of
organic solvents in which 1 and 2 could be soluble. When tetrahexylammonium (THA) salts
were employed during the synthesis a good balance between increased solubility and ease of
crystallization could be reached. As reported in our previous work, a preliminary heating of
THA-1 and THA-2 under vacuum was needed to remove water molecules that remain
physisorbed onto the POM and the SiOH functions (210°C at 10ꢀ3 mbar, 3h).13 Then, a clean
reaction of titanium tetra(isopropoxide) with the dehydrated silanolꢀfunctionalized POMs in
tetrahydrofuran (THF) took place leading to the titanium derivatives THA-3 and THA-4 (eq 1)
with the release of 3 equiv of 2ꢀpropanol, which can be monitored by NMR spectroscopy. NMR
ACS Paragon Plus Environment
6

Page 7 of 47
ACS Catalysis
1
2
3
4
5
6
7
8
9
analysis of collected THA-4 revealed the presence of a byꢀproduct, in a very low amount (<5%),
identified as a dimeric form, THA-6, which formation will be discussed below.
[XW9O34ꢀx(tBuSiOH)3]3ꢀ + Ti(OiPr)4 → [XW9O34ꢀx(tBuSiO)3Ti(OiPr)]3ꢀ + 3 iPrOH
(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
Structural characterization. Single crystals suitable for Xꢀray crystallographic analysis were
obtained for THA-3 and THA-4. As expected, both complexes are in monomeric forms (a
complete data refinement for anion 4 is hampered by conformational disorder of the ammonium
cations) (Figures 2 and 3). The geometric constraint imposed by the ligand forces the titanium
center to adopt a quasiꢀregular tetrahedral geometry (mean value for the SiO–Ti–OSi angles =
110.5°), four oxygen atoms coordinating the metal center. It is worth emphasizing that the ≡Si–O
and ≡SiO–Ti bond lengths in anion 3 and particularly those in anion 4 are very similar to those in
TSꢀ1 obtained by EXAFS studies.26 Table 1 gathers the bond distances for ≡Si–O and ≡SiO–Ti
and angle values for compounds 3 and 4 along with the corresponding values for TSꢀ1 and two
titaniumꢀsilsesquioxane complexes relevant to this work.
ACS Paragon Plus Environment
7

ACS Catalysis
Page 8 of 47
1
2
3
4
Figure 2. Crystal structure of anion 3.
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. Model of the titanium derivative 4.
Table 1. Principal bond distances for ≡Si–O and ≡SiO–Ti and angle values.
Si–O (Å) a SiO–Ti (Å) a Si–O–Ti (°) a Ref.
26
1
2
3
4
5
Tiꢀsilicalite (TSꢀ1)
1.62
1.815
153.5
167.5
154.4 b
[PW9O34(tBuSiO)3Ti(OiPr)]3ꢀ, 3
1.621(2)
1.804(3)
1.810 b
1.658(6)
1.826(7)
this study
[SbW9O33(tBuSiO)33Ti(OiPr)]3ꢀ, 4 1.604 b
this study
27
[(POSS)Ti(OSiMe3)]
1.625(6)
1.619(8)
28
[(POSS)Ti(MeOH)(µꢀOMe)]2
a
b
Mean value. A complete refinement for 4 was hampered by disorder at the tetrahexyl
ammonium cations.
ACS Paragon Plus Environment
8

Page 9 of 47
ACS Catalysis
1
2
3
4
5
6
7
8
9
Although a 3ꢀfold symmetry is not crystallographically imposed, the NMR studies of 3 and 4
revealed the expected C3 symmetry in solution. Moreover, the tꢀbutyl groups at the silicon atoms
function as NMR probe to confirm the introduction of the electron deficient {Ti–OiPr}3+
functionality as revealed by a slight low field shielding of the corresponding singlet (see
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
31
experimental section). P NMR chemical shift in anion 3 is little affected, thus indicating that
the tungstophosphate core is retained. The crystal structures also confirm that the bulky tꢀbutyl
groups create a sterically protected single metal site (Figure S1). Noteworthy, the Aꢀtype
[PW9O34]9ꢀ structure from 1 provides a lacuna which is slightly larger [mean diameter= 6.991 Å]
than the lacuna provided by the Bꢀtype [SbW9O33]9ꢀ structure from 2 [mean d= 5.887 Å] which is
reflected in the position of the silicon atoms and consequently the orientation of the tꢀbutyl
groups, pointing in a more axial fashion in 1 than in 2 (Figures 2, 3 and Figure S1). A second
consequence is the more obtuse Si–O–Ti angles observed in anion 3 compared to those in 4,
167.5° against 154.4° (Table 1). Considering again the electrophilic nature of the d0–titanium
center, it is reasonable to assume that the more obtuse the angle the more extended the stabilizing
O(pπ) → M(dπ) bonding.29 Overall, we can present anions 3 and 4 like accurate molecular
models for the tripodal, open–lattice type titanium site depicted in siloxide and zeolite in which
the metal atoms are tetraꢀcoordinated with three ≡SiO– framework bonds and a terminal –OH
group.
Reactivity towards water. Hydrophobicity is a key point governing the efficiency of the
titanium systems in liquid–phase oxidation reactions by TS–1 compared to amorphous TiꢀSiO2
systems. NMR monitoring of gradual addition of water to the solutions of 3 or 4 in dry
acetonitrile–d3 showed the progressive release of 2–propanol as a result of the protonolysis of the
remaining isopropoxide. Whereas the alcohol release was rather unfavorable in the case of 3
ACS Paragon Plus Environment
9

ACS Catalysis
Page 10 of 47
1
2
3
4
5
6
7
8
9
[log(K)= –1.96], compound 4 reacted readily with water. DOSY NMR spectroscopy provided
clues to help in the identification of the hydrolysis products, monomeric Ti–hydroxide vs.
dimeric Ti–hydroxide or oxide. Compound 3 and its product of hydrolysis, 5, displayed the same
diffusion coefficients and therefore 5 can be defined as a monomeric titaniumꢀhydroxyde
[PW9O34(tBuSiO)3Ti(OH)]3ꢀ species (eq 2).
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
[PW9O34(tBuSiO)3Ti(OiPr)]3ꢀ (3) + H2O ꢀ [PW9O34(tBuSiO)3Ti(OH)]3ꢀ (5) + iPrOH
(2)
In contrast, the diffusion coefficient of 6, the product of hydrolysis of compound 4, appeared
markedly smaller (Table S2 and Figure S2). The collected data strongly support the formation of
a ꢀꢀoxo bridged dimer {[SbW9O33(tBuSiO)3Ti]2O}6ꢀ according to eq 3. Formation of such dimers
has been reported for heterogeneous systems30 as well as for other Keggin type
polyoxometalates31,32,33 incorporating titanium centers.
i
2 [SbW9O33(tBuSiO)3Ti(OiPr)]3ꢀ (4) + H2O ꢀ {[SbW9O33(tBuSiO)3Ti]2O}6ꢀ (6) + 2 PrOH
(3)
The hydrolysis product THA-6 was also clearly identified as the byꢀproduct in the synthesis of
THA-4. Since we rigorously worked under anhydrous conditions, we should mention that anion
4, as a metal alkoxide, could undergo a nonꢀhydrolytic condensation (referred to as alcohol
dehydration) to give a ꢀꢀoxo bridged binuclear species through the elimination of
diisopropylether or a propene / 2ꢀpropanol mixture (Scheme S1).30,34 Compound 6 can be easily
isolated and elemental analytical and NMR data clearly indicate that no bridging isopropoxide or
isopropanol ligands are present. Indeed, the titanium ion in 6 remains fourꢀcoordinate, and the
formation of the dimer is made possible because of the smaller bulk provided by the tꢀbutyl
groups in 2 (Figure S1 and Scheme S1) [Note: this assignment was supported by a preliminary
ACS Paragon Plus Environment
10

Page 11 of 47
ACS Catalysis
1
2
3
4
5
6
7
8
Xꢀray analysis, which revealed the dimeric structure]. The system described herein contrasts with
the reported titanium silsesquioxane [(cꢀhexyl)7Si4O9(SiO)3Ti(OR)] series where titanium(IV)
can accommodate coordination numbers ranging from 4 in the monomeric form (R= Me3Si)27 to
9
i
a 5 and 6 in alkoxyꢀbridged dimers (R= Pr or Me).28 It is worth noting that the formation of the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢀꢀoxo bridge does not prevent the use of THA-4 as precatalyst since addition of H2O2 or an
excess of alcohol shifts back the equilibrium towards the formation of the monomeric titaniumꢀ
alkoxide (vide infra, ‘NMR spectroscopy’ section).30,35,36 Overall these results tend to indicate
that the formation of Ti–OH bond is not favorable and instead, the formation of a Ti–O–Ti bond
is more likely the primarily driving force of hydrolysis. We aim at using and developing a
system in which the metal center would fit in a confined and rigid coordination site with possibly
a hydrophobic environment. The titanium complexes reported herein meet fairly well these
requirements.
2.2. Applicability for the catalytic olefin epoxidation with H2O2.
When dealing with the use of POMs as catalysts in oxidation reactions with hydrogen
peroxide, one should first bear in mind the solvolytic instability of most [XW12O40]nꢀ Keggin
type structures in aqueous hydrogen peroxide solutions.37,38 Degradation of the oxotungstate core
proceeds through the release of {WO2}2+ species which can further react with H2O2 to form
Mimoun–39 or Venturello–like species,40 both active in olefin epoxidation.41,42,43 In the case of
hybrids of POMs, both baseꢀ and acidꢀcatalyzed hydrolysis may lead to the release of silanol
arms and the generation of cis–W(=O)2 units (Scheme S2).44 The latter have been proven to react
with hydrogen peroxide to afford cisꢀWO(η2ꢀO2) species, which explains the remarkable
catalytic activities of the lacunary [γꢀSiW10O34(H2O)2]4− and [CoW11O39]9−, both related to the
ACS Paragon Plus Environment
11

ACS Catalysis
Page 12 of 47
1
2
3
4
5
6
7
8
9
Keggin structures.45,46 The involvement of lacunary species in POMꢀcatalyzed oxidation
reactions with hydrogen peroxide is thus most commonly accepted. Some rare examples of non–
lacunary heteropolyoxometalate stable towards H2O2 have been reported, among which we can
mention the titanium–substituted heteropolyoxotungstates reported independently by the groups
of Kholdeeva,47,48 Mizuno,49,50 and Nomiya.33,51 Because of the question of solution stability, a
particular attention must be paid to the origin of the catalytic activity observed for 3 and 4, which
means Ti–centered or W–centered. To this purpose we first probed the stability of the precursors
THA–1 and THA–2 in the standard conditions that we use (1 mol% vs olefin) for the oxidation
of 3–methyl–2–buten–1–ol (0.26 M), our substrate of reference, with a 30%wt aqueous solution
of hydrogen peroxide ([H2O2]i = 0.26 M) in acetonitrile. While they are not expected to be
active, monitoring the reaction by NMR clearly indicated the undesired conversion of the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
31
unsaturated alcohol to epoxide over time when using THA-1 (Figure S3). P NMR analysis of
the catalytic mixture revealed the presence of new lines at ꢀ12.3 ppm, ꢀ12.6 ppm and ꢀ14.5 ppm
of small relative intensities compared to the main line at ꢀ16.9 ppm (1). These revealed the
solvolytic instability of 1 and the formation of small quantities of partially hydrolyzed
derivatives (Scheme S2).52 Conversely, anion 2 showed a much greater robustness to hydrolysis:
at room temperature, no significant conversion to epoxide was observed for the first four hours,
and then, conversion started only very smoothly (Figure S3). Over these first four hours, the
H2O2 concentration did not decrease significantly, showing that the precursor THA-2 also did
not catalyze the decomposition of hydrogen peroxide. Hence, in our conditions,
[SbW9O33(tBuSiOH)3]3ꢀ scaffold represent the best precursor and we thus focused our attention
on the catalytic behavior of its titanium derivative (4).
ACS Paragon Plus Environment
12

Page 13 of 47
ACS Catalysis
1
2
3
4
5
6
7
8
9
Catalytic Olefin Epoxidation. Attempts to oxidize unfunctionalized olefins, even an
electron–rich olefin such as 2–methyl–2–pentene, were unsuccessful, conversions remaining
insignificant (Table 2). Concerning allylic alcohols, conversions were observed when
nucleophilic olefins were employed, such as 3–methyl–2–buten–1–ol, whereas oxidation was
almost unsuccessful in the case of 2–propen–1–ol. Monitoring the olefin conversion over time
clearly indicated that the epoxidation starts readily and can reasonably be attributed to a
titanium–centered oxidation process (Figure 4).
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 2. Epoxidation of Olefins Catalyzed by (THA)3[SbW9O33(tBuSiO)3Ti(OiPr)], THAꢀ4, with
a
a 30wt% aqueous solution of H2O2
Selectivity (%) c
r0
(mM min–1)
Entry Olefins
Conv b (%)
Epox
n.a. d
n.a. d
80
Triol
ꢀ
Aldehyde
ꢀ
1
2
3
4
2ꢀmethylꢀ2ꢀpentene
<5
2ꢀpropenꢀ1ꢀol
<5
3ꢀmethylꢀ2ꢀbutenꢀ1ꢀol e
57
8.2
9.3
7.7
1.43
transꢀ2ꢀpentenꢀ1ꢀol
75.5
63.5
28.7
a All reactions were carried out in acetonitrile at room temperature (20°C) with [catalyst]= 2.6
mM, [olefin]0=[H2O2]0= 0.26 M. b Conversion is based on NMR analysis after 21 h of reaction. c
Undefined side products are not listed. d The conversion is too low to ascertain the selectivity in
e
epoxide. When THAꢀ3 was used as preꢀcatalyst initial rate measurement gave: r0= 0.38 mM
min–1.
ACS Paragon Plus Environment
13

ACS Catalysis
Page 14 of 47
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 4. Plot of 3–methyl–2–buten–1–ol conversion as a function of time in the oxidation by
H2O2 in CD3CN at room temperature, catalyzed by 1 mol % (THA)3[SbW9O33(tBuSiOH)3],
THAꢀ2 (ꢀ, dashed line), and its titanium derivative (THA)3[SbW9O33(tBuSiO)3Ti(OiPr)], THAꢀ
4 (ꢁ) ; [olefin]0=[H2O2]0= 0.26 M.
From these observations we can assume that: (i) the species formed from the interaction of
titanium with hydrogen peroxide is not very reactive, i.e. the proximal oxygen atom to the
titanium center is not electrophilic enough, (ii) olefin binding to the oxophilic titanium center
(through the alcohol function) is required to get an efficient conversion and, (iii) the mechanism
proceeds through a peroxidic pathway53,54 in which the olefins, as nucleophiles, react with the
metalꢀbound electrophilic oxygen in a back side approach.55,56 The second point is also
corroborated by the oxidation of geraniol (trans–3,7–dimethyl–2,6–octadien–1–ol), which
affords a selectivity of 95:5 for the 2,3–epoxy geraniol (Scheme 1).
ACS Paragon Plus Environment
14

Page 15 of 47
ACS Catalysis
1
2
3
4
Scheme 1. Epoxidation of geraniol with H2O2 catalyzed by THAꢀ4.
5
6
7
8
9
In the case of trans–2–penten–1–ol a higher conversion was obtained, most probably thanks to
the decrease of the bulkiness at the active olefinic group. However, the selectivity of the
corresponding triol (1,2,3ꢀpentanetriol) was also higher (28.7%). Triol originates from a
conventional two–step dihydroxylation: epoxide formation followed by its hydrolysis. Its
formation has been often explained by a slight acidic behavior of a titaniumꢀhydroperoxide
species that is postulated to form in situ and which promotes an acid–catalyzed oxirane ring
opening.57 It is however noteworthy that the higher triol selectivity observed in the case of trans–
2–penten–1–ol compared to 3–methyl–2–buten–1–ol tends to favor oxirane opening through a
nucleophilic substitution pathway in which a titanium hydroxide as intermediate, [Ti]–OH, could
be also involved. Finally, beside the epoxide formation, oxidation of the alcohol functionality
leading to the formation of α,βꢀunsaturated aldehyde occurs to a negligible extent (probably also
through a heterolytic mechanism). This dehydrogenation is expected to prevail when the
electrophilic character of the proximal oxygen atom diminishes.58
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. Evidence for a [Ti]–hydroperoxide as active species.
It is worth noticing that in aqueous solutions of hydrogen peroxide, titanium salts are known to
evolve into the more stable Ti(η2ꢀO2) sideꢀon peroxide species. Several Tiꢀperoxo complexes
have been structurally characterized by Xꢀray diffraction; however, when isolated they usually
proved to be not active for oxygen atom transfer to olefins (contrary to their tungsten or
molybdenum analogues).59,60 A Ti2–ꢀ– η2 η2ꢀperoxo core embedded in a [γꢀPW10O36]7ꢀ
:
phosphotungstate framework represents the only example of Tiꢀperoxide reported to be active for
ACS Paragon Plus Environment
15

ACS Catalysis
Page 16 of 47
1
2
3
4
5
6
7
8
9
oxidation.50 In the heterogeneous TS–1/H2O2/H2O system, the generation of Ti(η2ꢀO2) species
has been supported by Raman spectroscopy analysis. The appearance of a strong and sharp mode
at 618 cmꢀ1 has been ascribed to a vibration mode of the Tiꢀperoxide complex by comparison
with the structurally characterized (NH4)3[TiF5O2] model compound.61,62 This very characteristic
band is resonance–enhanced when tuning the exciting laser source within the specific LMCT
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–
transition (Ti4+ ← O2 , centered at 385 nm in TSꢀ1). The combined use of XANES/EXAFS
spectroscopy with common techniques (UVꢀvis, Raman, IR) and quantum chemical
methodologies led to agree on the following points: first, the exothermic formation of Ti–(OOH)
is consistent with both η1 or η2 bonding modes;9 secondly, in the presence of water an acid–base
equilibrium results in the formation of the sideꢀon Ti(η2ꢀO2) peroxide species and H3O+, thus
increasing the acidity of the TS–1/H2O2/H2O system.63 Importantly, the titanium hydroperoxo
species has been described as colorless while the titanium peroxo species was yellow (λ max = 385
nm).61,63 Despite these studies, and mostly because of the lack of relevant homogeneous models,
the detailed structure of the active intermediate is still a matter of debate.64,65 We thus decided to
follow by several spectroscopic analysis the evolution of 3 and 4, which represent relevant
homogeneous models, when in contact with H2O2 and 3ꢀmethylꢀ2ꢀbutenꢀ1ꢀol. Some of the results
are summarized in scheme 2 and discussed below.
NMR spectroscopy. In order to gain a deeper understanding of the nature of the active species
involved in the catalytic reaction, we first followed by NMR the effect of the addition of H2O2
(as a 30 wt% aqueous solution) on 5.2 mM solutions of 3 and 4 in deuterated acetonitrile. In the
case of 3, the addition of 3 eq of H2O2 led to almost complete conversion into a single new 3
fold–symmetric species, 7, along with the release of 1 eq of 2–propanol. By adding 4 eq of 3ꢀ
methylꢀ2ꢀbutenꢀ1ꢀol to this solution, the progressive formation of the corresponding epoxide was
ACS Paragon Plus Environment
16

Page 17 of 47
ACS Catalysis
1
2
3
4
5
6
7
8
9
observed. After consumption of the hydrogen peroxide, the complex 7 disappeared in favor of a
mixture of anions 3, [PW9O34(tBuSiO)3Ti(OiPr)]3ꢀ, and 5 [PW9O34(tBuSiO)3Ti(OH)]3ꢀ (Figure
S4). The same trend was observed when using anion 4: upon addition of H2O2, a new species
formed, 8, epoxidation occurred, and after complete consumption of H2O2 anion 8 disappeared in
favor of 6, which is formed preferentially in these conditions (Figure 5 and Figure S5). These
results clearly suggest the formal substitution of the incoming H2O2 molecule for isopropoxide
ligand as a first step, thus leading to the formation of the Ti–hydroperoxide species (either η1 or
η2–bonded) [PW9O34(tBuSiO)3Ti(OOH)]3ꢀ (7) and [SbW9O33(tBuSiO)3Ti(OOH)]3ꢀ (8).66
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. 1H NMR monitoring of the evolution of a 5.2 mM solution of
(THA)3[SbW9O33(tBuSiO)3Ti(OiPr)], THAꢀ4, in acetonitrileꢀd3: (a) THAꢀ4 ( , major) and its
ACS Paragon Plus Environment
17

ACS Catalysis
Page 18 of 47
1
2
3
4
5
6
7
8
9
related dimer, THAꢀ6 ( , minor); peaks reported by a ꢂ mark belong to the THA cations; (b)
formation of THAꢀ8 after addition of aqueous H2O2 (30wt%) to (a) ; (c) 2 min. after addition of
3ꢀmethylꢀ2ꢀbutenꢀ1ꢀol to (b) ; (d) after 2h of reaction, THAꢀ4 ( , minor), THAꢀ6 ( , major), and
(THA)3[SbW9O33(tBuSiO)3Ti(OCH2CH=C(CH3)2)] (ꢀ, minor). The doublet at 1.11 ppm belongs
to free 2ꢀpropanol.
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
UV–Vis Spectroscopy. The UV–Vis spectra of complexes 3 and 4 in acetonitrile are
dominated by a strong absorption in the UV domain due to the W6+ ← O2– ligandꢀtoꢀmetal
charge transfers. Upon addition of aqueous H2O2 (10 eq) the differences are tenuous but
proceeding to the subtraction of the spectra before and after the addition shed light to interesting
modifications. In both cases, a new absorption band was observed, which was centered at 264
nm for 3 and at 262 nm for 4 (Figure 6 and Figure S6). On the basis of the NMR results
discussed above, these bands could tentatively be attributed to the interaction of H2O2 with the
tetrahedral Ti centers in a η1 fashion, as proposed by Bonino et al,63 thus generating the colorless
Tiꢀ(η1ꢀOOH) species. In the case of 3, the shape of the band suggests the presence of a
supplementary absorption around 350 nm. This could reveal the formation of a small amount of
Tiꢀ(η2ꢀO2) peroxide species as the result of an acidꢀbase equilibrium with Tiꢀ(OOH) species. In
order to elucidate the generation of the Tiꢀ(η2ꢀO2) species in our systems we proceeded to
Raman spectroscopy analysis.
ACS Paragon Plus Environment
18

Page 19 of 47
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. Results of subtraction of UVꢀVis spectra after and before addition of 10 eq. of aqueous
H2O2 (a) to (THA)3[PW9O34(tBuSiO)3Ti(OiPr)], THAꢀ3, in acetonitrile, and (b) to
(THA)3[SbW9O33(tBuSiO)3Ti(OiPr)], THAꢀ4, in acetonitrile.
Raman spectroscopy. To obtain a satisfying signalꢀtoꢀnoise ratio we prepared 60 mM
solutions of 3 and 4 in acetonitrile and we used a 458 nm laser source (see the experimental
section). The Raman spectra are also dominated by two Wꢀspecific modes at 970 cmꢀ1 and 991
cmꢀ1. The intrinsically weaker Tiꢀspecific modes are either undetectable or covered by those of
the POM framework. Upon addition of H2O2 two different behaviors were observed. In the case
of THA-3, at this higher concentration, the solution turned yellow (due to the LMCT Ti4+
←
2–
O2 absorption) and the Raman analysis revealed the appearance of a sharp and intense band at
618 cmꢀ1 indicative of the formation of a Ti–peroxide species. Upon subsequent addition of 3ꢀ
methylꢀ2ꢀbutenꢀ1ꢀol this band widened while remaining centered at 620 cmꢀ1 (Figure 7, left).
This slight modification suggests some rearrangement in the coordination sphere, probably due
to the presence of the allylic alcohol. When the order of addition of the reagents was reversed the
wide peak at 620 cmꢀ1 was immediately observed. In both cases, the bands corresponding to the
ACS Paragon Plus Environment
19

ACS Catalysis
Page 20 of 47
1
2
3
4
1
Ti–peroxide species and 3ꢀmethylꢀ2ꢀbutenꢀ1ꢀol collapsed over time. H NMR analysis of the
5
6
resulting solutions revealed that the oxidation reaction afforded a 1:1 mixture of epoxide and
7
8
9
31
triol (the ratio increased to 2:1 when H2O2 was added after the olefin). In addition, P NMR
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
analysis revealed that a small amount of the POM framework was degraded. It is worth noting
that the intensity of the mode at 618 cmꢀ1 cannot be used for quantitative purposes since it is
resonanceꢀenhanced. In the case of THA-4, no color change was observed upon addition of
H2O2 and no clear formation of a band at 618 cmꢀ1 was observed. Only bubbles appeared at the
laser beam spot, a sign of hydrogen peroxide disproportionation.67 Notably, when H2O2 was
added to a mixture of THA–4 and 3ꢀmethylꢀ2ꢀbutenꢀ1ꢀol, no bands were detectable around 620
cmꢀ1 and only the band assigned to the olefin collapsed over time (Figure 7, right). NMR analysis
of this latter solution clearly indicated a higher selectivity for epoxide (ca. 80%) that matched the
one obtained when running catalysis (Table 2).
ACS Paragon Plus Environment
20

Page 21 of 47
ACS Catalysis
1
2
3
4
5
6
7
8
9
Figure 7. Raman spectra of (a) THA–3 in acetonitrile; (b) after addition of 10 eq. of H2O2 to (a);
(c) 2 min. after addition of 12 eq. of 3ꢀmethylꢀ2ꢀbutenꢀ1ꢀol to (b); (d) when reagents addition
was reversed; (e) THA–4 in acetonitrile, (f) after addition of 12 eq. of 3ꢀmethylꢀ2ꢀbutenꢀ1ꢀol to
(e); (g) 2 min. after addition of 10 eq. of H2O2 to (f); (h) after 30 min. Colors: blue for the Wꢀ
specific modes, red and orange for the Tiꢀperoxide and green for specific mode from 3ꢀmethylꢀ2ꢀ
butenꢀ1ꢀol (C=C bond, 1677 cmꢀ1). The spectra have been baseline corrected and normalized to
the acetonitrile vibration mode at 910 cmꢀ1.
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
From these results, it can be assumed that (i) the formation of a Ti(η2ꢀO2) peroxide complex is
not a prerequisite for epoxidation to occur and either a Ti–(OOH) hydroperoxide intermediate
represents the active species, and (ii) in the specific cases of our titaniumꢀpolyoxotungstates
derivatives 3 and 4, the equilibrium between Ti–(OOH) and Ti(η2ꢀO2) complexes is most
probably correlated with the acid–base equilibria that are involved in partial hydrolysis of the
siloxy component and/or the polyoxotungstate framework. Whereas the ability of titanium in
TS–1 to accommodate coordination numbers higher than four is supported by XANES and
EXAFS studies,68 this is very unlikely (at least energetically unfavorable) in 3 and 4 due to the
rigidity imposed by the polyoxotungstate frameworks. Therefore, in the case of the
tungstophosphate system (THA-3) the formation of Ti(η2ꢀO2) peroxide is probably related to the
partial hydrolysis of the siloxy part whereby the breaking of a SiO–Ti bond decreases the
geometrical constraints around the titanium center. On the contrary, the higher stability of the
tungstoantimonate system (THA-4) towards hydrolysis of the WO–Si bonds disfavors the
formation of titanium peroxide and we can assume that the titanium–hydroperoxide is merely the
only active species in the reaction mixture. It should be kept in mind that the absence of a sharp
band at 618 cmꢀ1 in that case does not exclude the possibility to form undetectable amount of
ACS Paragon Plus Environment
21

ACS Catalysis
Page 22 of 47
1
2
3
4
5
6
7
8
9
Ti(η2ꢀO2) peroxide that could be responsible for the activity. However, the observed production
of a much lower amount of triol and the higher catalytic activity observed for 4 compared to 3, as
depicted by initial rate values (r0= 0.38 mM/min for 3 and r0= 1.43 mM/min for 4 when applying
the catalytic conditions reported in Table 2, entry 3), tend to confirm that Ti–(OOH) is the active
species (Scheme 2). In the absence of olefin, the photodissociation of the TiꢀOOH moiety,
probably enhanced by the laser source, generates a Tiꢀhydroxyl function and O2, which escapes
as bubbles.67,69
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 schemes for the reactions of 4 (X= Sb, top) and 3 (X= P, bottom)
with H2O2 and 3ꢀmethylꢀ2ꢀbutenꢀ1ꢀol. Brackets around Ti atom stand for ligand 2 (X= Sb) and
for ligand 1 either complete or partially degraded (X= P).
2.4. Kinetic Studies.
We also monitored the catalytic reaction progress by NMR spectroscopy in order (i) to probe
the stability of the catalyst, and (ii) to gain understanding on the kinetics. Obviously, a
prerequisite for accurate determination of the substrate concentration dependence is that the
catalyst maintains a constant activity over time. Hence, to probe the stability of the catalyst
system we first performed two runs, with different initial substrate concentrations, but the same
ACS Paragon Plus Environment
22

Page 23 of 47
ACS Catalysis
1
2
3
4
5
6
7
8
9
“excess” value. In a reaction exhibiting a 1 to 1 stoichiometry between two reactants, the
concentrations are related to each other through this “excess [xs]” value (expressed in M) that
remains constant over the course of the reaction. The “same excess” protocol is one of the
mechanistic tools from Reaction Progress Kinetic Analysis (RPKA), a methodology which aims
at analyzing complex catalytic reactions from a minimal number of experiments.70,71 The
significance of the same excess protocol is that two reactions carried out at different initial
concentrations, like entries 2 and 3 in Table 3, represent an identical reaction started from
different points. Indeed, when the substrate concentrations for the reaction in entry 2 equal those
of the reaction in entry 3, the two kinetic profiles should fall on top of one other providing that
no extra process other than the intrinsic kinetics influences the reaction rate. Comparison of the
kinetic profiles is made possible by shifting the curve for entry 3 (as indicated by the arrows in
Figure 8a) in order to intercept the one for entry 2 at the point where concentrations are the same.
Figure 8b clearly indicates that from this point onward the two reactions do not exhibit the same
temporal substrate concentrations. At this point, three main features distinguish the reaction of
entry 2 from that of entry 3: (i) the catalyst has already performed several turnovers, (ii) the
products of the reaction, epoxide and water, have accumulated in the reaction medium, and (iii) a
larger quantity of water is present from the beginning due to the higher hydrogen peroxide
concentration used. Indeed, the higher rate of conversion that is observed for reaction 3
compared to that of reaction 2 at the same substrate concentrations correlates to a catalyst
deactivation, which can result either from the formation of an inactive form (or leaching) or from
a product inhibition.
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
23

ACS Catalysis
Page 24 of 47
1
2
3
4
Table 3. Catalytic conditions used, with same [xs] values, for the NMR monitoring of the
5
6
epoxidation of 3ꢀmethylꢀ2ꢀbutenꢀ1ꢀol with H2O2 catalyzed by THAꢀ4. a
7
8
9
Selectivity (%) c
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
[H2O2]0
(mM)
260
[Olefin]0
(mM)
260
[xs]
(mM)
ꢀ
Entry
%Conv b
Epox
Triol
Ald
1
57 (21 h)
91 (12 h)
96 (4 h)
92 (5 h)
98 (4 h)
80
8.2
9.3
7.7
2.8
6.1
4.9
2
270
420
150
150
150
150
82
10.2
7.1
3
160
310
87
4d
160
310
71.7
84.3
9.9
5e
160
310
10.4
a
b
All reactions were carried out in acetonitrile–d3 at 298K with [THA–4]0= 2.6 mM.
Conversion is observed after the time indicated under brackets, based on the concentration of
[olefin]. Undefined side products are not listed. d Reaction carried out with addition of water.
c
e
Reaction carried out with addition of 2,3ꢀepoxy product.
Figure 8. Plots of the concentration of 3–methyl–2–buten–1–ol as a function of the time in the
oxidation by H2O2 in CD3CN at room temperature, catalyzed by
(THA)3[SbW9O33(tBuSiO)3Ti(OiPr)], THA–4 ; (a) reactions of entries 2 and 3 using the « same
excess » protocole; (b) the red profile is time–adjusted to the point at which concentrations in
[olefin]t and [H2O2]t are the same in the two experiments; (c) comparison of the kinetic profiles
for entry 4 ([H2O] adjusted, ꢀ) and for entry 5 ([epoxide] adjusted, ꢁ).
ACS Paragon Plus Environment
24

Page 25 of 47
ACS Catalysis
1
2
3
4
5
6
7
8
Water has been often reported as a main factor of inhibition for titaniumꢀbased oxidation
catalysis. Thus, we carried out two supplementary experiments for which the initial substrate
concentrations and the water concentration (entry 4) or the epoxide concentration (entry 5) are
identical to those of the reaction in entry 2 at the point of intersection. Comparison of the kinetic
profiles (i) shows that accumulated water does not significantly change the kinetic profile at the
beginning of the catalysis, but that a deviation could be observed over time, (ii) indicates that
accumulated epoxide does not change either the kinetic profile in a significant extent (Figure 8c).
Hence, we can conclude that the difference between the reaction profiles of entries 2 and 3
results from a partial catalyst degradation leading to the formation of less or inactive species.
This is consistent with the fact that the polyoxotungstates 1 and 2 undergo partial hydrolysis
when the hydrogen peroxide concentration is too high (as shown in Figure S3 and Scheme S2).
Using an excess of olefin and decreasing [H2O2]i and [H2O]i resulted in almost complete
conversion, up to 96% in 4 h when a 1:60:120 catalyst:H2O2:olefin mixture was used (entry 3).
These conditions also afforded an increased turnover frequency value over the reaction time
(mean TOF= 24 hꢀ1). It is also not surprising to observe that the selectivity towards epoxide
slightly dropped when a larger amount of water is present (entry 4).
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
More interestingly, monitoring the reaction progress by 1H NMR also allows to draw a picture
of the kinetics along the catalysis, i.e. along the in situ reagents concentration. To do so, the
temporal concentration data have first to be converted mathematically in order to obtain the
reaction rate evolution over time (Figure S7). Then, plotting the rate vs substrate concentration
gives a graphical equation of the reaction. Figure 9 displays a graphical rate equation of the data
obtained for entry 3 as rate vs [olefin] and reveals a linear behavior, in the range of 0.31 to 0.2M,
ACS Paragon Plus Environment
25

ACS Catalysis
Page 26 of 47
1
2
3
4
5
6
7
8
9
with a nonzero intercept (below 0.2M, the observed deviation could be either attributed to a
change of the kinetics at lower concentrations or to an increased noise due to the mathematical
manipulations). This result can be rationalized by expressing the rate law as k⋅[H2O2].
Considering the definition of [xs], we can formulate the concentration of H2O2 as [H2O2]=
[olefin] – [xs], and thereof the rate law as k⋅[olefin] – k⋅[xs]. This defines the equation of the
straight line depicted in Figure 9. The slope of the linear portion gives k = 8.79 10ꢀ3 minꢀ1 and the
yꢀintercept gives –1.45 mM/min, which is in good agreement with calculated – k⋅[xs]= –1.32
mM/min (see also Figure S7). This indicates that (i) the reaction is firstꢀorder in [H2O2] and (ii)
the reaction is pseudo zeroꢀorder in [olefin] (saturation kinetics in [olefin]). The rate law is
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
consistent
with
addition
of
H2O2
being
the
turnoverꢀlimiting
step
and
[SbW9O33(tBuSiO)3Ti(OCH2CH=C(CH3)2)]3ꢀ being the catalyst resting state, which has been
clearly demonstrated by NMR monitoring. We can also underline that the constancy of the
kinetic regime tends to indicate that the catalyst does not undergo significant degradation under
the catalytic conditions used for entry 3. In this study, the possibility to monitor the reaction
progress by NMR and the rather high selectivity of the reaction were the prerequisites to make
use of the RPKA methodologies in order to look inside the reaction to get a full picture of the
catalytic mechanism instead of the snapshots provided by conventional kinetic methods such as
measurements of the initial rate.
ACS Paragon Plus Environment
26

Page 27 of 47
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 9. Graphical rate equation (Table 3, entry 3): rate vs 3ꢀmethylꢀ2ꢀbutenꢀ1ꢀol
concentration, [olefin]. The linear behavior with nonꢀzero intercept related to value of [xs]
indicates a firstꢀorder in [H2O2] and a zeroꢀorder in [3ꢀMeꢀ2ꢀbutenꢀ1ꢀol].
These preliminary kinetic studies proved the ability of the titanium–based polyoxotungstate 4
to catalyze the epoxidation of nucleophilic allylic alcohols with aqueous solution of hydrogen
peroxide at room temperature. One requirement lies in the fact that concentration in hydrogen
peroxide should remain low enough to avoid partial hydrolysis and maintain the integrity of the
catalyst over the reaction. We should remind at this stage that one of the factors governing the
catalytic performances of TSꢀ1 lies in the size and hydrophobic nature of the pores of its
crystalline framework, which promote the physisorption of apolar substrates and ensure slow
diffusion of hydrogen peroxide and a limited amount of water inside the pore volume.72,73 No
homogeneous system can compare to that. However, THA-4 is particularly interesting because
the fact that Ti(IV) site is not prone to expand its coordination sphere is probably the reason why
inhibition of the active site by water molecules is minimized. Above all, our studies reveal that
the Tiꢀhydroperoxide species formed is not very reactive for the conversion of nonꢀ
functionalized alkenes under our catalysis conditions. Therefore, efforts must be made to
ACS Paragon Plus Environment
27

ACS Catalysis
Page 28 of 47
1
2
3
4
5
6
7
8
increase the concentration of olefin at the active center and if possible to increase its electrophilic
character. Further catalytic tests, recycling studies, as well as kinetic studies will be undertaken
in order to improve the catalytic performance and the scope of applications of our systems.
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
3. CONCLUSIONS
The silanol–decorated polyoxotungstates [XW9O34ꢀx(tBuSiOH)3]3ꢀ (X= P, x=0, 1; X= Sb, x=1,
2) have been used as original ligands in an approach to modeling active sites of titaniumꢀ
containing heterogeneous epoxidation catalysts. The most notable feature of these ligands lies in
the preꢀorganization of the silanol functionalities, which leads to complexes with constrained
geometry. The titaniumꢀisopropoxide complexes (nꢀHex4N)3[αꢀAꢀPW9O34(tBuSiO)3Ti(OiPr)]
(THA-3) and (nꢀHex4N)3[αꢀBꢀSbW9O33(tBuSiO)3Ti(OiPr)] (THA-4) have been synthesized and
characterized using NMR spectroscopy and singleꢀcrystal Xꢀray diffraction. The complexes
proved relevant models for singleꢀsite silicaꢀsupported titanium catalysts, especially the defective
openꢀlattice sites that are present in the MFI framework of the titaniumꢀsilicalite TSꢀ1.
Monitoring by NMR and Raman spectroscopies the reaction between THA-3 / THA-4 and
aqueous solutions of hydrogen peroxide showed that a titanium metal atom in such constrained
and confined environments activates H2O2 to form a Tiꢀhydroperoxide Tiꢀ(OOH) intermediate.
Importantly, this work shed light to the fact that this latter proved to be the active species
responsible for high epoxide selectivity in allylic alcohols oxidation with H2O2. In our systems,
the evolution to Ti(η2ꢀO2) peroxide species is only made possible by the limited solution stability
of 3 which undergoes partial hydrolysis of the siloxy functionalities and thus releases the
geometric stress around the titanium centre. In that case, formation of Ti(η2ꢀO2) peroxide
correlates to the increase for 1,2,3ꢀbutaneꢀtriol selectivity as a consequence of the concomitant
ACS Paragon Plus Environment
28

Page 29 of 47
ACS Catalysis
1
2
3
4
5
6
7
8
9
release of protons that catalyze the oxirane opening of the in situ formed epoxide. In the case of
the more stable THA-4, the evolution to Ti(η2ꢀO2) peroxide was not detected by Raman
spectroscopy and Ti–(OOH) species reacted with allylic alcohol to selectively produce epoxide
whereas H2O2 decomposition may occur in the absence of olefin. Reaction progress kinetic
analysis methodology was also meaningful to map the catalytic mechanism over the course of
the reaction and to provide evidence for catalyst integrity at low concentration of H2O2. To the
best of our knowledge, THA–4 represents the first example of a soluble titanium complex
structurally modeling the open sites of TSꢀ1 and able to carry out the epoxidation of olefin with
aqueous solutions of H2O2 without the need of embedment into a hydrophobic environment such
as polymers. Whereas classical models of heterogeneous titanium catalysts are reported to
deactivate in contact with water, THA-4 shows a higher stability, most probably due to the
hydrophobicity provided by the surrounding tetrahexylammonium cations and tꢀbutyl groups, but
also to the rigidity of the POM framework, which disfavors the increase of the titanium
coordination number.
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. EXPERIMENTAL SECTION
General Procedures. The lacunary polyoxotungstates K9[αꢀAꢀPW9O34] and Na9[αꢀBꢀ
SbW9O33] were prepared as reported in the literature.74 Their silanol derivatives (nꢀHex4N)3[(αꢀ
AꢀPW9O34)(tBuSiOH)3] (THA–1) and (nꢀHex4N)3[(αꢀBꢀSbW9O33)(tBuSiOH)3] (THA–2) were
prepared according to modified procedures (vide infra).75,76 All manipulations were conducted
under a nitrogen atmosphere using standard Schlenk techniques. Tetrahydrofuran and acetonitrile
were dried using solvent purification systems or by distillation under argon from appropriate
drying agents. All solvents, including deuterated acetonitrile, were then degassed by several
ACS Paragon Plus Environment
29