Understanding the Regioselectivity of Aromatic Hydroxylation over Divanadium-Substituted γ-Keggin Polyoxotungstate

Article abstract of DOI:10.1021/acscatal.7b02694

The aromatic hydroxylation of pseudocumene (PC) with aqueous hydrogen peroxide catalyzed by the divanadium-substituted γ-Keggin polyoxotungstate TBA₄[γ-PW₁₀O₃₈V₂(μ-O)(μ-OH)] (TBA-1H, TBA = tetrabutylammonium) has been studied using kinetic modeling and DFT calculations. This reaction features high chemoselectivity and unusual regioselectivity, affording 2,4,5-trimethylphenol (TMP) as the main product. Then the computational study was extended to the analysis of the regioselectivity for other alkoxy- and alkylarene substrates. The protonation/deprotonation of TBA-1H in MeCN/tBuOH (1:1) was investigated by ³¹P NMR spectroscopy. Forms with different protonation states, [γ-PV₂W₁₀O₄₀]^5- (1), [γ-HPV₂W₁₀O₄₀]^4- (1H), and [γ-H₂PV₂W₁₀O₄₀]^3- (1H₂), have been identified, and the protonation equilibrium constants were estimated on the basis of the ³¹P NMR data. DFT calculations were used to investigate the oxygen transfer process from hydroperoxo species, [γ-PW₁₀O₃₈V₂(μ-O)(μ-OOH)]^4- (2) and [γ-PW₁₀O₃₈V₂(μ-OH)(μ-OOH)]^3- (2H), and peroxo complex [γ-PW₁₀O₃₈V₂(μ-²:²-O₂)]^3- (3) toward the different positions in the aromatic ring of PC, anisole, and toluene substrates. Product, kinetic, and computational studies on the PC hydroxylation strongly support a mechanism of electrophilic oxygen atom transfer from peroxo complex 3 to the aromatic ring of PC. The kinetic modeling revealed that the contribution of 3 into the initial reaction rate is, on average, about 70%, but it may depend on the reaction conditions. DFT calculations showed that the steric hindrance exerted by peroxo complex 3 is responsible for the origin of the unusual regioselectivity observed in PC hydroxylation, while for anisole and toluene the regioselective para-hydroxylation is due to electronic preference during the oxygen transfer from the active peroxo species 3.

Full text of DOI:10.1021/acscatal.7b02694

Subscriber access provided by READING UNIV
Article
Understanding the regioselectivity of aromatic hydroxylation
over divanadium-substituted #-Keggin polyoxotungstate
Igor Y. Skobelev, Vasiliy Yu Evtushok, Oxana A. Kholdeeva, Nataliya V. Maksimchuk,
Raisa I Maksimovskaya, Josep M. Ricart, Josep M. Poblet, and Jorge J. Carbó
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.7b02694 • Publication Date (Web): 03 Nov 2017
Downloaded from http://pubs.acs.org on November 3, 2017
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 33
ACS Catalysis
1
2
3
4
5
6
7
8
9
Understanding the Regioselectivity of Aromatic
Hydroxylation Over DivanadiumꢀSubstituted γꢀ
Keggin Polyoxotungstate
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
Igor Y. Skobelev,a,b,* Vasiliy Yu. Evtushok,a,b Oxana A. Kholdeeva,a,b Nataliya V. Maksimchuk,a,b
Raisa I. Maksimovskaya,a Josep M. Ricart,c Josep M. Poblet, c Jorge J. Carbóc,*
a Boreskov Institute of Catalysis, Lavrentiev ave. 5, Novosibirsk 630090 (Russia). bNovosibirsk
State University, Pirogova str. 2, Novosibirsk 630090 (Russia). cDepartment de Química Física i
Inorgànica, Universitat Rovira i Virgili, Marcel·lí Domingo 1, 43007 Tarragona (Spain).
KEYWORDS. Aromatic hydroxylation, hydrogen peroxide, homogeneous catalysis, kineticꢀ
modeling, pseudocumene, polyoxometalate, vanadium, DFT
ABSTRACT. The aromatic hydroxylation of pseudocumene (PC) with aqueous hydrogen
peroxide catalyzed by the divanadiumꢀsubstituted γꢀKeggin polyoxotungstate TBA4[γꢀ
PW10O38V2(ꢁꢀO)(ꢁꢀOH)] (TBAꢀ1H, TBA = tetrabutylammonium) has been studied using kinetic
modeling and DFT calculations. This reaction features high chemoselectivity and unusual
regioselectivity, affording 2,4,5ꢀtrimethylphenol (TMP) as the main product. Then the
computational study was extended to the analysis of the regioselectivity for other alkoxyꢀ and
alkylarene substrates. The protonation/deprotonation of TBAꢀ1H in MeCN/tBuOH (1:1) was
ACS Paragon Plus Environment
1

ACS Catalysis
Page 2 of 33
1
2
investigated by the 31P NMR spectroscopy. Forms with different protonation state, [γꢀ
PV2W10O40]5ꢀ (1), [γꢀHPV2W10O40]4ꢀ (1H) and [γꢀH2PV2W10O40]3ꢀ (1H2), have been identified,
and the protonation equilibrium constants were estimated on the basis of the 31P NMR data. DFT
calculations were used to investigate the oxygen transfer process from hydroperoxo species, [γꢀ
PW10O38V2(ꢁꢀO)(ꢁꢀOOH)]4ꢀ (2) and [γꢀPW10O38V2(ꢁꢀOH)(ꢁꢀOOH)]3ꢀ (2H), and peroxo complex
[γꢀPW10O38V2(ꢁꢀη2:η2ꢀO2)]3ꢀ (3) toward the different positions in the aromatic ring of PC, anisole
and toluene substrates. Product, kinetic, and computational studies on the PC hydroxylation
strongly support a mechanism of electrophilic oxygen atom transfer from peroxo complex 3 to
the aromatic ring of PC. The kinetic modeling revealed that contribution of 3 into the initial
reaction rate is, on the average, about 70%, but it may depend on the reaction conditions. DFT
calculations showed that the steric hindrance exerted by peroxo complex 3 is responsible of the
origin of the unusual regioselectivity observed in PC hydroxylation, while for anisole and
toluene the regioselective paraꢀhydroxylation is due to electronic preference during the oxygen
transfer from the active peroxo species 3.
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
INTRODUCTION
The selective oxidation of aromatic rings is an important field of both industrial and synthetic
chemistry. Substituted phenols are important intermediate products in the polymer processing
and production of fine chemicals,1ꢀ3 while quinones are crucial intermediates in the synthesis of
essential vitamins, nutraceuticals, and pharmaceuticals.4ꢀ7 During the past decade, a few efficient
catalyst systems have been elaborated for the selective oxidation of alkylphenols to quinones
using environmentally benign oxidants (O2, H2O2, tBuOOH).8 On the contrary, selective
oxygenation of aromatic rings of alkylbenzenes remains one of the most challenging
ACS Paragon Plus Environment
2

Page 3 of 33
ACS Catalysis
1
2
transformations in the organic synthesis.9ꢀ12 Alkylbenzenes have electronꢀdonating groups that
activate the aromatic nucleus toward electrophilic attack, but on the other hand, alkyl
substituents are prone to oxidation, which results in poor aromatic oxidation selectivity. Nonꢀ
catalytic stoichiometric oxidations of arenes with hazardous and toxic reagents, such as CrO3,
MnO2, OsO4 and others, are still widely employed in fine organic synthesis.1,4,5,13 Such
processes produce large amounts of inorganic waste contaminated by organic tars. The recently
emerged urgency of the economical and environmental sustainability encourages the scientific
community to develop catalytic chemical processes, which reduce the burden on the
environment.14ꢀ16 In particular, the development and implementation of efficient catalytic
systems for the aromatic oxyfunctionalization of alkylsubstituted arenes using hydrogen peroxide
as the green oxidant is a challenging goal of modern catalysis.
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
Aromatic oxidation of 1,2,4ꢀtrimethylbenzene or pseudocumene (PC) to 2,3,5ꢀtrimethylꢀ1,4ꢀ
benzoquinone (TMBQ, vitamin E key intermediate) is considered as in posse industrial
process17. Currently, TMBQ is produced on industrial scale through oxidation of 2,3,6ꢀ and
2,3,5ꢀtrimethylphenols (TMP).8,18ꢀ20 In contrast to the oxidation of TMP, the direct aromatic
oxidation of PC has been scarcely described in the literature.9 Few nonꢀcatalytic procedures with
metaꢀchloroperoxybenzoic acid21 and H2O2/HCOOH mixture22 as oxidants have been reported.
Beller and coꢀworkers suggested use of iron complexes as catalysts for aromatic oxidations with
hydrogen peroxide.18 So far, methylrhenium trioxide seems to be the most efficient catalyst for
H2O2ꢀbased oxidation of PC.17,23,24
Earlyꢀtransitionꢀmetalꢀoxygen anionic clusters or polyoxometalates (POMs for short) are known
as versatile catalysts for a wide range of liquidꢀphase oxidations.25,26,27,28,29 While many POMs
catalyze heterolytic activation of H2O2 and oxygen atom transfer to alkenes and thioethers, thus
ACS Paragon Plus Environment
3

ACS Catalysis
Page 4 of 33
1
2
3
4
5
6
7
8
9
far only one POM, i.e. the divanadium substituted γꢀKeggin phosphotungstate (Bu4N)4[γꢀ
PW10O38V2(ꢁꢀO)(ꢁꢀOH)] (designated as TBAꢀ1H, where TBA stands for tetrabutylammonium)
(see Scheme 1b), is known as an effective catalyst for aromatic hydroxylation of alkylarenes.30
The same divanadiumꢀsubstituted POM had been able to catalyze stereoꢀ and regioselective
hydroxylation of alkanes with H2O2.31 Moreover, some of us have found that TBAꢀ1H catalyzes
effectively oxidation of various alkylated phenols/naphtols32 and methoxyarenes33 to 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
corresponding
pꢀbenzo/naphthoquinones,
outperforming
the
divanadiumꢀsubstituted
silicotungstate (Bu4N)4H2[γꢀSiW10V2O40] and dititaniumꢀsubstituted silicotungstate (Bu4N)8[{γꢀ
SiW10Ti2O36(OH)2}2(ꢀꢀO)2] catalysts. Recently, we have also reported that TBAꢀ1H is able to
catalyze hydroxylation of the aromatic ring of PC.34 The lack of methyl group oxidation products
was consistent with electrophilic hydroxylation mechanism while unusual regioselectivity
toward the formation of 2,4,5ꢀ and 2,3,5ꢀTMP (ca. 7 : 1 : <0.15, see Scheme 1b) indicated that
steric factors may have a strong influence on the oxygen transfer process. In general, it appears
that the divanadium substituted γꢀKeggin phosphotungstate behaves as a privileged catalyst in a
range of selective oxidations with the green oxidant H2O2 under mild conditions, and therefore,
understanding its functioning in detail can contribute to the development of novel effective
oxidation processes.
The group of Mizuno revealed that the presence of acid coꢀcatalyst and MeCN/tBuOH as solvent
is crucial for the catalytic activity of TBAꢀ1H in H2O2ꢀbased oxygenation of alkanes,31 electronꢀ
deficient alkenes,35 and arenes.30 In our previous work, we confirmed that the addition of acid
coꢀcatalyst and tBuOH coꢀsolvent is required for the PC aromatic hydroxylation in the presence
of TBAꢀ1H.34 Without HClO4, the yield of TMP attained only 24% after 60 min whereas it
reached 64% after 10 min if 1 equiv. of HClO4 (relative to the POM catalyst) was added. Mizuno
ACS Paragon Plus Environment
4

Page 5 of 33
ACS Catalysis
1
2
3
4
5
6
7
8
9
suggested that the role of acid in the 1Hꢀcatalyzed oxidations is to form a diprotonated
polyoxoanion, [γꢀH2PV2W10O40]3ꢀ (1H2) that, upon interaction with H2O2, forms hydroperoxo
complex [γꢀPW10O38V2(ꢁꢀOH)(ꢁꢀOOH)]3ꢀ (2H), dehydration of which leads to a highly
electrophilic peroxo species, [γꢀPW10O38V2(ꢁꢀη2:η2ꢀO2)]3ꢀ (3) (Scheme 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
Scheme 1. Molecular structure of species 1H (a), and possible products of the hydroxylation of
pseudocumene with experimental distribution in parenthesis (b).
H2O
H2O
H2O
H
O
O
H+
H2O2
H
O
O
O
3
O
V
V
V
V
V
V
V
V
O
H
O
H
O
H
H+
H2O
H2O2
2H
1H2
1H
Scheme 2. Tentative mechanism of the formation of active peroxo complex 3.
Several computational studies were dedicated to characterization of the catalytic activity of the
siliceous analog of 1H, [γꢀSiW10O38V2(ꢁꢀOH)(ꢁꢀOH)]4ꢀ, in epoxidation of olefins with H2O2.36,37
While Nakagawa and Mizuno considered only the ꢁꢀη2:η2 peroxo complex [γꢀSiW10O38V2(O2)]4ꢀ
(3’) as a possible origin of the catalytic activity,36 Kuznetsov et al. analyzed reactivity of the
ACS Paragon Plus Environment
5

ACS Catalysis
Page 6 of 33
1
2
hydroperoxo species [γꢀSiW10O38V2(ꢁꢀOH)(ꢁꢀOOH)]4ꢀ (2H’) as well and concluded that peroxo
species 3’ is more active than 2H’.37 Kamata et al. showed by DFT that the calculated energy
barrier of the oxidation of transꢀ2ꢀbutene by peroxo complex 3 is higher than the barrier for cisꢀ
2ꢀbutene,35 which pointed out steric hindrance of 3. However, it has not been reported yet any
computational analysis of either the reactivity of the divanadiumꢀsubstituted phosphotungstate
1H/H2O2 catalyst system toward substituted arenes or of the regioselectivity of the oxygen
transfer process.
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
In this work, we present a detailed study of the mechanism of the TBAꢀ1Hꢀcatalyzed aromatic
31
hydroxylation of alkylbenezenes using PC as the model substrate by means of the tools of P
NMR, kinetic modeling, and DFT calculations, as well as the extension of computational
analysis to other alkoxyꢀ and alkylarenes reported in the literature. Our aim was to determine the
mechanism of this reaction and to understand the origin of the observed regioselectivity for
different aromatic substrates.
EXPERIMENTAL
Materials. Pseudocumene (97+%) was obtained from Aldrich. Acetonitrile (Panreac, HPLC
grade) and tBuOH (Acros Organics, 99+%) were dried and stored over activated 4 Å molecular
sieves. All other compounds were of the best available reagent grade and used without further
purification. The concentration of H2O2 (ca. 35 wt.% in water) was determined iodometrically
prior to use.
Catalysts preparation and characterization. The cesium salt of the dilacunary
phosphotungstate Cs7[γꢀPW10O36]·xH2O was synthesized as described in the literature.38 The
tetrabutylammonium salt TBA4[γꢀHPV2W10O40] was prepared according to the general
ACS Paragon Plus Environment
6

Page 7 of 33
ACS Catalysis
1
2
procedure reported by Kamata et al.31 with some modifications.32 The number of acid protons in
the TBA salt was corroborated by potentiometric titration with TBAOH (Figure S1 in Supporting
Information (SI)). The identity and purity of the POM was confirmed by FTIR (0.5–1.0 wt.%
samples in KBr pellets, Agilent 660 FTIR) and multinuclear NMR. IR (cmꢀ1): 1650, 1380, 1100,
1050, 1030, 980, 890, 800, 520, 490 (the corresponding FTIR spectrum is shown in Figure S2 in
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
51
31
31
SI). V NMR (dry MeCN): −581 ppm (Figure S3); P NMR (dry MeCN): −14.2 ppm. P and
51V NMR spectra were recorded on a Brüker AVANCEꢀ400 spectrometer at 161.67 and 105.24
MHz, respectively, using a highꢀresolution multinuclear probe head. Chemical shifts for P and V,
δ, were determined relative to 85% H3PO4 and VOCl3, respectively. For convenience, a 0.2 M
31
solution of H4PVMo11O40 in water was used as a secondary external standard for P (−3.70
ppm).
31P NMR monitoring of protonation/deprotonation of TBAꢀ1H. The interaction of TBAꢀ1H
with HClO4 and TBAOH was studied in tꢀBuOH/MeCN (1:1 v/v) solution at room temperature.
A 0.01 M solution of TBAꢀ1H was placed into a 10 mm o.d. sample tube, and then an
appropriate amount of HClO4 or TBAOH (0.1–2.0 equiv. relative to POM) was gradually added
31
and the chemical shift of the POM signal was measured by P NMR immediately after the
addition and after some time to make sure that the equilibrium had been reached.
Kinetic experiments. Kinetic experiments of PC hydroxylation with H2O2 in the presence of
TBAꢀ1H were carried out in temperatureꢀcontrolled glass vessels at 40 °C under vigorous
stirring (500 rpm). Concentrations of the reactants were varied in the range of [PC] = 0.8–2.4,
[H2O2] = 0.036–0.180, [TBAꢀ1H] = 0.00041–0.0024, [HClO4] = 0.0009–0.0072, and [H2O] =
0.14–1.14 M. Typically, the reactions were initiated by the addition of H2O2 to a mixture
containing aromatic substrate, TBAꢀ1H catalyst and HClO4 in a solvent mixture MeCN/tꢀBuOH
ACS Paragon Plus Environment
7

ACS Catalysis
Page 8 of 33
1
2
3
4
5
6
7
8
9
(1:1 v/v; the total volume of the reaction mixture was 1 mL). Samples were taken periodically
every 8 minutes during the reaction course (average reaction time = 80 minutes) by a syringe,
and accumulation of TMP was determined by GC (Chromos GCꢀ1000 equipped with a flame
ionization detector and a quartz capillary column BPX5 30 m×0.25 mm) using biphenyl as
internal standard. Each experiment was reproduced at least two times. In all experiments, the
conversion of PC was in 2ꢀ4.5% range.
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
Kinetic modeling. Initial PC hydroxylation rates were computed as rates of TMP accumulation.
The rate law of the PC hydroxylation was derived applying a steadyꢀstate approximation to
concentrations of all active species. Procedures for calculation of the initial rate of TMP
formation and fitting the rates with the derived law were similar to those described by some of us
earlier.39 For the detailed description of the kinetic modeling procedure and derivation of the rate
law, see the SI.
Computational details. The DFT analysis of the reaction pathway was carried out with
Gaussian09 rev. C01 software.40 Geometry optimization of reagents, intermediates, transition
states and products was made using B3LYP density functional.41ꢀ43 LANL2DZ pseudopotential44
was used for W and V atoms and 6ꢀ31g(d,p) basis set45ꢀ47 was used for all atoms of PC, water
and hydrogen peroxide as well as for oxygen atoms directly bonded to V. For other atoms, 6ꢀ31g
basis set45 was used. The geometry optimization was full and without any symmetry constrains,
and IEFPCM implicit solvation model included in Gaussian09 (solvent – acetonitrile) was used.
Such a computational level has been successfully employed in the study of the reactivity of
earlyꢀtransitionꢀmetalꢀoxygen clusters including quantitative agreement with Arrhenius
activation energies48 and kinetic modeling.39 It is out of the scope of this study to analyze the
specific role of solvent mixture, which has been shown to be important in reaction yield, and
ACS Paragon Plus Environment
8

Page 9 of 33
ACS Catalysis
1
2
3
4
5
6
7
8
9
countercations. Future computational studies are planned to address these effects using dynamic
simulations analogous to those employed in some of our previous studies.49 Nevertheless, the
effect of the dielectric constant on the regioselectivity was tested by repeating the calculations
using the dielectric constant of tBuOH (12.5), and then, averaging the results with those obtained
for acetonitrile. The relative free energies for the isomeric transition states of heterolytic oxygen
transfer from hydroperoxo species (TS3, see below) vary only in 0.1 kcal·molꢀ1. Thus, the
reported values were obtained using only the free energy in acetonitrile in order not to increase
the computational cost unnecessarily. The reaction under investigation occurs in the liquidꢀ
phase, but Gaussian calculates gasꢀphase free energies. To take into account the liquidꢀphase
character of the reaction, a translation entropy correction50 was incorporated to the computed free
energy.
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
RESULTS AND DISCUSSION
31P NMR monitoring of protonation/deprotonation of TBAꢀ1H and evaluation of
protonation constants.
It has long been known that protonation of αꢀKeggin heteropolyanions [PV2W10O40]5ꢀ (in this
POM, two vanadium atoms are statistically distributed over 12 positions of the αꢀKeggin
31
structure) in acidic media causes a highꢀfrequency shift of the P NMR signals.51ꢀ53 The same
tendency was observed for γꢀKeggin polyanion 1.30,31 In CH3CN solutions, TBAꢀ1H reveals a
31P NMR signal at δ –(14.2±0.1) ppm.30,31,35 In turn, a form containing two protons,TBAꢀ1H2, is
manifested by a 31P NMR signal at δ –13.8.31,33 A mixture of monoꢀ and diprotonated forms 1H
and 1H2 may give two separate signals in dry diluted MeCN solutions.33 However, if some water
ACS Paragon Plus Environment
9

ACS Catalysis
Page 10 of 33
1
2
31
3
4
or tBuOH is added, only one averaged signal is observed due to fast proton exchange on the P
5
6
NMR time scale.
7
8
9
Since the presence of tBuOH as coꢀsolvent and HClO4 as coꢀcatalyst is crucial to accomplish
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
effectively hydroxylation of alkylbenezenes in the presence of TBAꢀ1H,30,34 we employed P
NMR to investigate protonation/deprotonation of TBAꢀ1H in a mixture of MeCN/tBuOH (1:1
v/v) and then used the 31P NMR data (vide infra) to evaluate the corresponding protonation
constants for 1 and 1H (Scheme 3).
K1
+ H+
(1)
1
1H
K2
+ H+
(2)
1H
1H2
Scheme 3. Protonation equilibriums for species 1 and 1H.
While the addition of tBuOH to MeCN produced insignificant changes in the position of the 31P
NMR signal of TBAꢀ1H, additives of acid or base resulted in substantial signal shifting. The
gradual alterations observed in the 31P NMR spectra of TBAꢀ1H upon titration with TBAOH and
HClO4 are shown in Figure 1a and Figure 1b, respectively. The 31P signal moved upfield
reaching δ –15.0 ppm upon addition of 1 equiv. of TBAOH and then remained constant (Figure
1a), indicating the formation of nonprotonated form 1. On the contrary, the addition of acid
31
caused a progressive downfield shift of the P NMR signal, which attained δ –13.83 ppm at ca.
ACS Paragon Plus Environment
10

Page 11 of 33
ACS Catalysis
1
2
1 equiv. of H+ and after that almost did not change (Figure 1b). Note that the chemical shift of –
3
4
5
6
13.8 ppm has been previously assigned to the diprotonated form 1H2.31
7
8
9
Fast exchange on the NMR time scale can be described with a simple formalism54 using values
of chemical shifts of individual forms and protonation equilibrium constants (eq. S1.1 in the SI).
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
Given that, we used the dependences of the P chemical shift of TBAꢀ1H on the amount of
added HClO4 and TBAOH to evaluate the protonation equilibrium constants K1 and K2 in
MeCN/tBuOH (1:1), the solvent mixture that had been identified as optimal for the
hydroxylation of alkylbenzenes. Figure S4 shows fitting of the experimental dependences with
equation S1.1. The estimated values of K1 and K2 equal to 2.5·105 and 8.6·103 Mꢀ1 with standard
deviations of 1.4·104 and 1.5·102 Mꢀ1, respectively. These values will be used in kinetic
modeling of PC oxidation (vide infra).
Figure 1. 31P NMR spectra of TBAꢀ1H in tBuOH/MeCN (1:1 v/v) upon addition of TBAOH (a)
and HClO4 (b).
Kinetics study of PC hydroxylation
ACS Paragon Plus Environment
11

ACS Catalysis
Page 12 of 33
1
2
3
4
5
6
7
8
9
Initial rates of TMP formation (W0) were measured in a wide range of concentrations of all the
reagents. Plots of W0 versus concentrations of water, oxidant, acid, catalyst, and arene are shown
in Figure 2 a1–c1 and Figure 3 a1ꢀb1.
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
12

Page 13 of 33
ACS Catalysis
1
2
3
4
5
6
7
8
9
Figure 2. Experimental dependences of initial rates of TMP formation on initial reagent
concentrations and their fitting with equation S3.7 (a1–c1) and computed contributions of the
active species into the initial reaction rate (a2–c2). Reaction conditions: 0.0081 M TBAꢀ1H (a,
b), 0.036 M H2O2 (a, c), 0.7 M H2O (b, c), 0.009 M HClO4, 0.8 M PC, 40°C, 1 mL of solvent
(MeCN/tBuOH = 1:1 v/v).
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. Experimental dependences of initial rates of TMP formation on initial reagent
concentrations and their fitting with equation S3.7 (a1–b1) and computed contributions of the
active species into the initial reaction rate (a2–b2). Reaction conditions: 0.0162 M TBAꢀ1H (a)
or 0.0081 M TBAꢀ1H (b), 0.036 M H2O2, 0.7 M H2O (a) or 0.48 M H2O (b), 0.009 M HClO4 (b),
0.8 M PC (a), 40°C, 1 mL of solvent (MeCN/tBuOH = 1:1 v/v).
ACS Paragon Plus Environment
13

ACS Catalysis
Page 14 of 33
1
2
3
4
5
6
7
8
9
We attempted to rationalize these complicated dependences suggesting a mechanism similar to
that proposed by Mizuno30 that involves the formation of active peroxo species 3. However, this
mechanism has been expanded with an assumption that hydroperoxo complexes [γꢀ
PW10O38V2(ꢁꢀO)(ꢁꢀOOH)]4ꢀ (2) and [γꢀPW10O38V2(ꢁꢀOH)(ꢁꢀOOH)]3ꢀ (2H) might also be able to
participate in the hydroxylation process. The protonation equilibria between different forms were
also taken into account. The overall reaction mechanism is presented in Scheme 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
K1
k4
k7
k8
k9
+ H+
+ H+
+ H+
+ H2O (4)
(1)
(2)
+ H2O2
(7)
(8)
+ PC
+ PC
+ PC
TMP
TMP
TMP
k-4
1
1H
2
1H
1H2
2H
1H
2
2
K2
k5
k-5
(5)
(6)
+ H2O
+ H2O2
1H2
2H
2H
k6
k-6
K3
(3)
+ H2O
(9)
2H
3
3
Scheme 4. Proposed mechanism of PC hydroxylation with H2O2 in the presence of TBAꢀ1H.
By applying a steadyꢀstate approximation to concentrations of all active species (2, 2H and 3), a
rate law described by eqs S3.2 and 1 has been drawn (see SI for details). To estimate the
corresponding rate and equilibrium constants for the reaction steps depicted in Scheme 4, all the
dependencies of the initial reaction rate on reagent concentrations were fitted simultaneously
with the rate law derived (Figure 2, a1ꢀc1, and 3 Figure 3, a1–b1). Recently39, we employed a
similar approach for studying mechanism of thioether oxidation with H2O2 catalyzed by [(γꢀ
SiTi2W10O36(OH)2)2(ꢁꢀO)2]8ꢀ.
ACS Paragon Plus Environment
14

Page 15 of 33
ACS Catalysis
1
2
3
4
5
6
7
8
9
Kinetic regularities mainly reflect the reactivity and concentration of the active species, while the
protonation equilibria involve both initial forms of the catalyst, 1, 1H and 1H2, and active peroxo
species, 2 and 2H. The kinetic dependencies themselves are scarcely a good source of data for
estimation of protonation equilibrium constants K1 and K2 but, fortunately, the values of K1 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
31
K2 could be determined independently from the P NMR data. To incorporate these values into
the kinetic model, we applied a regularization procedure (see SI for details). After modeling with
the full parameter set, i.e. with all equilibrium and rate constants for the proposed mechanism
(Scheme 4), it was found that concentration of 1 under the experimental conditions employed is
very low, so that reaction 1 in Scheme 4 as well as parameter K1 can be excluded from the
kinetic model. The optimal values of k5 and kꢀ5 turned out significantly lower than those of k4
and kꢀ4, indicating that the formation of 2, 2H and 3 mainly proceeds through interaction of 1H
with H2O2 (reaction 4 in Scheme 4) while reaction of 1H2 with H2O2 (reaction 5 in Scheme 4) is
significantly slower. Finally, the following set of parameters was used in the modeling
procedure: K2, K3, K6, k4, kꢀ4, k7, k8, and k9, where K6 = k6/kꢀ6. To calculate the value of the
initial reaction rate with eq. 1 it is necessary to know some additional quantities (eq. 2) which, in
turn, depend from the rate and equilibrium constants, initial concentration of reagents and from
equilibrium concentration of protons. The latter can be obtained from simultaneous solution of
material balance eqs. 3 and 4. The optimal values of all the parameters of eqs. 1 and 2 are given
in Table 1.
ꢃ
ꢄꢅ
ꢃ
ꢅ
(kꢁ + Kꢂ H (kꢆ + αꢂkꢇ))αꢈ POM ꢃPCꢅꢀ
ꢀ
(1)
(2)
Wꢀ =
Φꢉꢊꢋ
ꢃ
ꢅ ꢃ
ꢅ
kꢌ HꢈO HꢈOꢈ ꢀ
Kꢎ
ꢀ
αꢈ =
αꢂ =
ꢃ
ꢅꢈ
ꢃ
ꢄꢅ
HꢈO
ꢃ
ꢅ
ꢀ
kꢍꢌ HꢈO + kꢇKꢎKꢂ H ꢃPCꢅꢀ
ꢀ
ACS Paragon Plus Environment
15

ACS Catalysis
Page 16 of 33
1
2
3
4
5
6
ꢃ
ꢄꢅꢀ
ꢃ
ꢄꢅ
) )
= H + ꢏꢐ 1 + 2Kꢈ H + αꢈ(1 + 2Kꢂ H 1 + αꢂ )
ꢃ
ꢅ(
ꢃ
ꢄꢅ
ꢃ
ꢄꢅ(
H
(3)
(4)
ꢃ
ꢅ
ꢃ
ꢅ(
ꢃ
ꢄꢅ
ꢃ
ꢄꢅ(
) )
ꢃ
ꢅ
POM = ꢏꢐ 1 + Kꢈ H + αꢈ(1 + Kꢂ H 1 + αꢂ ) = ꢏꢐ Φꢉꢊꢋ
ꢀ
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
Table 1. Optimal parameters of the kinetic model (eqs. 1ꢀ2) and their standard deviations.
Parameter
K2, Mꢀ1
Value
8.9·103
2.0·105
9.4·10ꢀ3
9.2·102
4.0·103
9.3
Standard deviation
1·102
3·104
1·10ꢀ3
90
K3, Mꢀ1
K6, M
k4, (M·min)ꢀ1
kꢀ4, (M·min)ꢀ1
k7, (M·min)ꢀ1
k8, (M·min)ꢀ1
k9, (M·min)ꢀ1
6·102
6·10ꢀ1
5·10ꢀ2
10
4.3·10ꢀ1
92
One may conclude from the data given in Table 1 that peroxo complex 3 is the most active form
of the catalyst since the value of the rate constant k9 (92 (M·min)ꢀ1) is the highest among
constants k7–k9. This fact is consistent with the hypothesis proposed by Mizuno that namely
peroxo species 3 is responsible for the catalytic activity of TBAꢀ1H in aromatic hydroxylation30.
A contribution of each peroxo species into the overall initial rate of TMP formation can be
estimated from the kinetic modeling (eq. S3.8–S3.10, see SI). Although the ratio of all the
peroxo species depends on the reaction conditions (Figure 2, a2ꢀc2, and Figure 3, a2–b2), peroxo
complex 3 usually predominates and its contribution is about 70% in a wide range of reagent
ACS Paragon Plus Environment
16

Page 17 of 33
ACS Catalysis
1
2
3
4
concentrations (Figure 2, b2, and Figure 3, a2–b2). Average contributions of 2H and 2 are 21%
5
6
and 9%, respectively.
7
8
9
While varying the concentration of water, the contribution of peroxo complex 3 changed from
90% ([H2O] = 0.14 M) to 50% ([H2O] = 1.14 M). This is consistent with the proposed scheme of
the formation of 3 via dehydration of 2H (reaction 6 in Scheme 4). Since the assessment of the
rate constants showed that 3 is the most active form, the decrease in its contribution into the
oxidation process leads to the reduction of the overall reaction rate (Figue2, a1). The other
reason of the observed decrease in the reaction rate is that the increase in water concentration
generally disfavors the formation of the active species. Obviously, the contribution of 3 into the
initial reaction rate also decreased with increasing concentration of TBAꢀ1H (Figure 2, c2).
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 complex behavior of the system was observed when the concentration of acid coꢀcatalyst was
varied. The experimental dependence of the initial reaction rate on the concentration of acid went
through a maximum, which was also reproduced by the kinetic model (Figure 3, a1). Such a
behavior may result from a complex equilibrium between all forms of the catalyst. At a low
concentration of acid, the concentration of 3 may decrease because it is generated from the
protonated form 2H. Note that 2 gives a significant contribution to the reaction rate when the
concentration of acid is low (Figure 3, a2). When the concentration of acid increases, the
concentrations of 2H and 3 also enlarge as well as their contribution into the reaction rate, and
then the reaction rate reaches its maximum (Figure 3, a1–a2). However, further increase in the
concentration of acid leads to decreasing concentration of 1H, the main form that participates in
the formation of the active peroxo species upon interaction with H2O2, which, in turn, results in
decreasing concentration of the active species. Consequently, the overall reaction rate decreases
(Figure 3, a1) while the individual contributions of the peroxo species into the reaction rate
ACS Paragon Plus Environment
17

ACS Catalysis
Page 18 of 33
1
2
3
4
5
6
7
8
9
remain constant (Figure 3, a2). The augmentation of the PC concentration leads to the increase in
the reaction rate (Figure 3, b1), which is consistent with the proposed mechanism of the
interaction of the active species with PC (Scheme 4). Therefore, the reaction mechanism
depicted in Scheme 4 has led to the rate law which is fully consistent with all the kinetic
regularities observed experimentally for PC hydroxylation with H2O2 catalyzed by TBAꢀ1H in
the presence of acid coꢀcatalyst.
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
+26.1
TS2
+22.3
+20.4
+20.0
TS1H
TS2H
TS3
+6.6
+5.7
- H+
+ H+
2
1H
+1.9
0.0
-1.0
3
1H2
2H
TMP
Figure 4. Potential energy profile (kcal·molꢀ1) for 1Hꢀcatalyzed hydroxylation of PC to TMP
derived from kinetic modeling.
DFT study of PC hydroxylation
Figure 4 shows the energy profile for PC hydroxylation catalyzed by divanadiumꢀsubstituted γꢀ
Keggin polyoxotungstate derived from the kinetic modeling using the transition state theory.39
The mechanism can be divided in three main steps: (1) H2O2 activation to form the vanadium
ACS Paragon Plus Environment
18

Page 19 of 33
ACS Catalysis
1
2
3
4
5
6
7
8
9
hydroperoxo species 2, which protonates to yield 2H, (2) dehydration of 2H to vanadium peroxo
species 3, which is somewhat endothermic, and (3) electrophilic oxygen transfer from the peroxo
species 3, which is preferred to vanadium hydroperoxo path through species 2H. This proposal is
consistent with the previously DFTꢀcharacterized mechanism of a related process, the
epoxidation of alkenes by [γꢀH2SiV2W10O40]4ꢀ anion.36,37
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. Evaluated from kinetic modeling and calculated freeꢀenergy barriers (kcal·molꢀ1) for
oxygen transfer from 2H and 3 to PC.a
2H
3
ꢁG≠
ꢁG≠
21.4
24.2
19.7
18.1
21.3
16.8
kin. mod.
DFT
ꢁG≠
DFTꢀcorr
a
The ꢁG≠
and ꢁG≠
values represent uncorrected and corrected free energy barriers,
DFTꢀcorr
DFT
respectively.
In order to understand the origin of the unusual regioselectivity in the 1Hꢀcatalyzed aromatic
hydroxylation, we carried out DFT calculations on the electrophilic oxygen transfer from 2H and
3 to PC substrate. We did not consider the reaction path through hydroperoxo species 2 because,
according to the kinetic modeling, it has the highest energy barrier (+26.1 kcal·molꢀ1) and the
lowest contribution into the initial reaction rate (9%, see above). Table 2 collects the computed
and evaluated from kinetic modeling free energy barriers of the lowest energy paths for oxygen
transfer, while Figure 5 shows the geometries of the corresponding transition states. Since
calculated Gibbs free energies overestimate the translational entropy loss in the associative
processes, we corrected them by applying Whitesides’ correction50 (ꢂG≠DFTꢀcorr) and provide both
uncorrected and corrected values in Table 2. It is important to note that the entropy correction is
ACS Paragon Plus Environment
19

ACS Catalysis
Page 20 of 33
1
2
3
4
5
6
7
8
9
less accurate for protic solvents, and especially, for water. However, the combination of
corrected and uncorrected values is informative because it provides a lower and upper limit of
entropy loss in associative processes, such as the intermolecular attack of arene to POM active
intermediates. Importantly, the values of the two experimental energy barriers lay within the
energy range defined by the calculations (see Table 2). Also in agreement with the experimental
values, the computed free energy barrier of oxygen transfer decreases on going from
hydroperoxo 2H to peroxo 3 (2.9 kcal·molꢀ1 by DFT vs. 3.3 kcal·molꢀ1 experimentally). These
results mutually support each other, confirming the characterization of the mechanism by kinetic
modeling and DFT calculations.
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. Representation of the DFT computed transition states for oxygen transfer from
hydroperoxo complex 2H and peroxo complex 3 to PC. Selected, lowestꢀenergy regioisomers.
Distances in Å.
ACS Paragon Plus Environment
20

Page 21 of 33
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. Representation of the DFT computed structures of intermediates obtained after oxygen
transfer for hydroperoxo and peroxo paths, 4 and 5, respectively. Distances in Å.
The transition states TS2H (Figure 5 and S5) involve the transfer of the Oβ–H group to PC from
the bridging vanadium hydroperoxo group, yielding a protonated TMP molecule that remains
hydrogenꢀbonded to the V2(ꢀꢀO) group of the POM framework as shown by structure 4 in Figure
6. Then the Hꢀbonded hydrogen can return to the V2(ꢀꢀO) oxo group to recover catalyst 1H2
while the other proton rearranges to form the final TMP product. In fact the same mechanism
was proposed for direct conversion of benzene to phenol catalyzed by the enzyme cytochrome
P450 on the basis of DFT calculations.55 Alternatively, the transfer of the Oβ–H group to PC
might induce a direct proton release from the aromatic ring to a proton acceptor solvent molecule
via a classical SN2 aromatic substitution mechanism. However, the analysis of the role of explicit
solvent for the minor reaction path is besides of the scope of this study. Analyzing in more detail
the preferred peroxo path, we found that the transition state TS3ꢀ2,3,5 leads to a ꢀꢀalkoxo
divanadium intermediate (5) with the new C–O bond lengthening 1.39 Å (see Figure 6), which
ACS Paragon Plus Environment
21

ACS Catalysis
Page 22 of 33
1
2
freeꢀenergy is 21.3 kcal.molꢀ1 bellow reactants. Thus, calculations indicate that the formation of
3
4
5
6
7
5 is irreversible (reverse freeꢀenergy barrier ꢁG≠
≌ 40 kcal·molꢀ1) and, consequently, that
5
TS3
ꢀ
8
9
the electrophilic oxygen transfer is the selectivityꢀdetermining step. From 5, the freeꢀenergy
barrier associated to proton rearrangement to give the final TMP product is 24.7 kcal·molꢀ1, that
is feasible for a reaction occurring at 40ºC.
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 3. Calculated relative free energies (kcal·molꢀ1) for the isomeric transition states of
heterolytic oxygen transfer from 3 and 2H to PC and relative free energies (kcal·molꢀ1) of the
protonated PC substrate.
ꢁꢁG≠ (TS3)
0.0
ꢁꢁG≠ (TS2H)
ꢁꢁG (PC+H+)
Isomer
+0.3
+3.3
0.0
0.0
2,4,5
2,3,5
2,3,6
+0.6
+3.1
+0.2
+5.7
Table 3 (second column) collects the relative free energies of the different isomeric transition
states leading to 2,4,5ꢀ, 2,3,5ꢀ and 2,3,6ꢀTMP isomers through the preferred peroxo path (TS3).
The results are consistent with the unusual regioselectivity toward 2,4,5ꢀ and 2,3,5ꢀTMP isomers
(ca. 7:1) that was observed in PC hydroxylation in the presence of TBAꢀ1H.34 The computed
lowest energy path corresponds to TS3ꢀ2,4,5 followed at relatively low energy (0.6 kcal.molꢀ1)
by TS3ꢀ2,3,5 path. This energy difference is qualitatively close to that expected from the
observed isomer ratio at 40º C, i.e., 1.2 kcal·molꢀ1. The TS3ꢀ2,3,6 path is significantly higher in
energy (5.7 kcal·molꢀ1) and no reaction is expected to occur through this pathway. Since
hydroperoxo path (TS2H) has a minor contribution to final product distribution (see above), we
can conclude that calculations reproduce qualitatively the preferred formation of 2,4,5ꢀTMP
ACS Paragon Plus Environment
22

Page 23 of 33
ACS Catalysis
1
2
3
4
5
6
product, although the quantitative prediction of the distribution of isomers (specially of the minor
ones) is not very accurate.
7
8
9
Origin of the unusual regioselectivity of PC hydroxylation: steric vs electronic preference
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
In the peroxo path (TS3), the product distribution differs from the hydroperoxo path (TS2H),
which might explain the origin of the unusual regioselectivity (see Table 3). We propose that the
regioselectivity switch from its electronic preference due to the steric factors governing the
interaction between the substrate and the POM framework. In fact, Mizuno and coꢀworkers have
demonstrated that the calculated energy barrier of epoxidation of cisꢀ2ꢀbutene by peroxo
complex 3 is lower than that of transꢀ2ꢀbutene,35 indicating crucial importance of steric factor for
the reactivity of 3. Also, we have shown that other POM structures such as the dititaniumꢀ
substituted sandwich anion determines the selectivity in the epoxidation of alkenes via a similar
steric effect.56,57 To evaluate the electronic preference toward electrophiles of the different
carbon sites of PC, their affinity towards protonation was investigated. The relative free energies
of protonated PC species are shown in Table 3. They follow roughly the same trend as oxygen
transfer from the Oβ oxygen of the hydroperoxo group (TS2H), for which there is no sterical
hindrance. Conversely, in peroxo intermediate 3, the oxygen transfer occurs closer to the POM
framework and steric repulsion with the substrate comes into play. Consequently, the most
hindered position with methyl substituents at the two alpha positions, TS3ꢀ2,3,6, becomes the
highest in energy. The lowest energy isomeric transitions states are TS3ꢀ2,4,5 and TS3ꢀ2,3,5,
which are 5.7 and 5.1 kcal·molꢀ1 lower than TS3ꢀ2,3,6, while the atomic charges, determined
with the CHELPG method, at the C5 and C6 carbons of PC (ꢀ.22 and ꢀ.17 a.u., respectively) are
less negative that that at the C3 carbon (ꢀ.29 a.u.). Therefore, the positions C5 and C6, although
less nucleophilic, have less sterical hindrance with the POM framework and consequently
ACS Paragon Plus Environment
23

ACS Catalysis
Page 24 of 33
1
2
3
4
5
6
7
8
9
become favorable, thereby switching the electronicallyꢀdriven regioselectivity (TMPꢀ2,3,6)
usually observed in the oxidations of PC to the stericallyꢀdriven one (TMPꢀ2,4,5). For
hydroperoxo species 2H, free energies of TS2Hꢀ2,3,6 and TS2Hꢀ2,4,5 (Table 3) are similar,
which implies that the formation of 2,3,6ꢀ and 2,4,5ꢀTMP might be a competitive process.
However, only minor amounts of 2,3,6ꢀTMP were detected in the experimental study of PC
hydroxylation,34 which is consistent with the low (~21%) contribution of 2H into the initial
aromatic hydroxylation rate. Therefore, the DFT study pointed out that the steric hindrance
exerted by peroxo complex 3 during electrophilic oxygen transfer is responsible of the origin of
the unusual regioselectivity observed in PC hydroxylation over TBAꢀ1H.
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
Origin of regioselectivity in hydroxylation of alkoxyꢀ and alkylarene substrates
Once we have characterized the mechanism and identified the regioselectivityꢀdetermining
species for the aromatic hydroxylation using pseudocumene as a model substrate, we analyze
whether this mechanism can be also used to explain the unique regioselectivity exhibited by
divanadiumꢀsubstituted phosphotungstate 1H2 in the direct hydroxylation of various arenes with
H2O2 reported recently by Mizuno et al.30 To this end, we select two experimentally tested
monosubstituted substrates, anisole as representative of alkoxyarenes and toluene as
representative of alkylarenes (Scheme 5). For both substrates, the freeꢀenergy barrier of the
peroxo path (3 + arene → TS3ar: 17.4 and 20.5 kcal·molꢀ1 for anisole and toluene, respectively)
is significantly lower than that of the hydroperoxo path (2H + arene → TS2Har: 22.5 and 28.2
kcal·molꢀ1, respectively). Thus the peroxo complex 3 is the active species mainly responsible of
the oxygen transfer to the substrate.
ACS Paragon Plus Environment
24

Page 25 of 33
ACS Catalysis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 5. Hydroxylation of alkoxyꢀ and alkylarenes with H2O2 catalyzed by divanadiumꢀ
substituted phosphotungstate 1H2.
Table 4. Calculated relative free energies (kcal·molꢀ1) for the isomeric transition states of
heterolytic oxygen transfer from 2H and 3 to anisole and toluene, and comparison between
computational and experimental regioselectivities (%).
ꢁꢁG≠ (TS2Har) ꢁꢁG≠ (TS3ar)
compound isomer
% selec. (calc.) % selec. (exp.)
ortho
meta
para
ortho
meta
para
1.2
5.9
0.0
ꢀ1.6
1.2
0.0
1.0
5.2
0.0
2.8
1.1
0.0
18
<1
82
1
5
anisole
toluene
<1
95
7
16
83
16
77
Table 4 collects the relative freeꢀenergies of the isomeric transition states for oxygen transfer
through hydroperoxo (TS2Har) and peroxo (TS3ar) paths. Importantly, there is an excellent
agreement between the experimental regioselectivities and those predicted from the relative freeꢀ
energies of the peroxo path, TS3ar (see Table 4). Calculations do not only reproduce the
preference for the para isomer but also the relative distribution between the ortho and meta
isomers. This correspondence between experimental and computational regioselectivities further
ACS Paragon Plus Environment
25

ACS Catalysis
Page 26 of 33
1
2
3
4
5
6
7
8
9
supports that the oxygen transfer form peroxo complex 3 is the key regioselectivityꢀdetermining
step. The comparison of the relative freeꢀenergies for peroxo (TS3ar) path with those of
hydroperoxo (TS2Har) can be used to evaluate the steric and electronic effects on the
regioselectivity (see Table 4). In the case of anisole the distribution of transition state energies is
almost identical for TS3ar and TS2Har paths, indicating that the regioselectivity is governed by
the electronic preference of the substrate. We recall that in TS2Har the oxygen transfer occurs
farther from the POM framework than in TS3ar. For toluene the distribution of the minor
regioisomers is slightly different from anisole with observable amounts of the metaꢀsubstituted
product. Comparing peroxo (TS3ar) and hydroperoxo (TS2Har) transition states, we found that
the relative energies of meta/para regioisomers are almost identical indicating an electronic
control in their relative orientation. On the other hand, the ortho isomer is somewhat destabilized
in peroxo path TS3ar compared to the hydroperoxo one TS2Har. This suggests that in ortho path
the alkyl substituent of the arene becomes closer to the POM framework inducing repulsive
steric interaction and decreasing its reactivity.
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
CONCLUSIONS
Three forms of polyanion 1 differing in their protonation state, [γꢀPW10O38V2(ꢁꢀO)2]5ꢀ (1), [γꢀ
PW10O38V2(ꢁꢀO)(ꢁꢀOH)]4ꢀ (1H) and [γꢀPW10O38V2(ꢁꢀOH)2]3ꢀ (1H2), have been identified by
acidꢀbase titration monitored by 31P NMR in MeCN/tBuOH solution of TBAꢀ1H. The
31
protonation equilibrium constants of 1 and 1H estimated from the P NMR data are equal to
2.5·105 and 8.6·103 Mꢀ1, respectively. The kinetic modeling study revealed that hydroperoxo
complexes [γꢀPW10O38V2(ꢁꢀO)(ꢁꢀOOH)]4ꢀ (2) and [γꢀPW10O38V2(ꢁꢀOH)(ꢁꢀOOH)]3ꢀ (2H) give a
relatively small contribution to the initial rate of PC hydroxylation while peroxo complex [γꢀ
PW10V2O38(O2)]3ꢀ (3) is the most active form of the catalyst and its contribution to the initial
ACS Paragon Plus Environment
26

Page 27 of 33
ACS Catalysis
1
2
3
4
5
6
7
8
9
reaction rate is about 70%, although it may depend on the reaction conditions. The DFT
calculations also point out that the most active species in the aromatic hydroxylation is the
peroxo complex 3¸ which in addition can be used to explain the observed regioselectivity further
supporting the mechanistic proposal. The preferred positions of oxygen transfer from 3 in the
pseudocumene molecule are defined rather by steric than electronic factors. Consequently,
unusual regioselectivity toward 2,4,5ꢀ and 2,3,5ꢀtrimethylphenols results from steric hindrance of
the formation of 2,3,6ꢀtrimethylphenol via interaction of pseudocumene and 3. For the studied
alkoxyꢀ and alkylarene substrates, anisole and toluene, the selectivityꢀdetermining step is also the
oxygen transfer from the peroxo complex 3 but, in this case, the observed paraꢀhydroxylation
regioselectivity is governed by the electronic preference of the substrate.
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
AUTHOR INFORMATION
Corresponding Author
*Igor Y. Skobelev (email: skobelev@catalysis.ru) and *Jorge J. Carbó (j.carbo@urv.cat)
Author Contributions
The manuscript was written through contributions of all authors. All authors have given approval
to the final version of the manuscript.
ACKNOWLEDGMENT
The authors thank Dr. G. M. Maksimov for the synthesis of the lacunary precursor Cs7[γꢀ
PW10O36]. This work was conducted within the framework of the budget project No. 0303ꢀ2016ꢀ
0005 for the Boreskov Institute of Catalysis. The research at Universitat Rovira i Virgili was
supported by the Spanish Ministry of Science and Innovation (grant CTQ2014ꢀ52774ꢀP), the
DGR of the Generalitat de Catalunya (grant2014SGR199 and XRQTC). I.Y.S. acknowledges the
ACS Paragon Plus Environment
27

ACS Catalysis
Page 28 of 33
1
2
3
4
5
6
Polyoxometalate Chemistry for Molecular Nanoscience (PoCheMoN) action in the framework of
the European Cooperation in Science and Technology (COST) program for financial support.
7
8
9
ASSOCIATED CONTENT
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
Supporting Information. Additional spectral data, detailed equations of the kinetic modeling,
and Cartesian coordinates and absolute energies of computational structures.
REFERENCES
1. The chemistry of Phenol, Rappoport, Z., Ed., John Wiley Ltd: Chichester, 2003.
2. Lorenc, J. F., Lambeth, G., Scheffer, W. In Kirk-Othmer Encyclopedia of Chemical
Technology, WileyꢀVCH: Weinheim, 2003, Vol. 2, p 203ꢀ233.
3. Fiege, H., Voges, H.ꢀW., Hamamoto, T., Umemura, S., Iwata, T., Miki, H., Fujita, Y., Buysch,
H.ꢀJ., Garbe, D., Paulus, W. In Ullmann's encyclopedia of industrial chemistry, WileyꢀVCH:
Weinheim, 2002.
4. The Chemistry of the Quinonoid Compounds, Patai, S., Rappoport, Z., Eds., Wiley: New York,
1988.
5. In The Vitamins, Schudel, P., Mayer, H., Isler, O., Sebrell, W. H., Harris, R. S., Eds.,
Academic Press: New York, 1972, Vol. 5, p 168ꢀ217.
6. Handbook of Vitamins, Rucker, R. B., Suttie, J. W., McCormick, D. B., Machlin, L. J., Eds.,
Marcel Dekker: New York, 2001.
7. Nicolaou, K. C., Chen, J. S., Edmonds, D. J., Estrada, A. A. Angew. Chem. Int. Ed. 2009, 48,
660ꢀ719.
ACS Paragon Plus Environment
28

Page 29 of 33
ACS Catalysis
1
2
3
4
5
6
7
8
8. O. A. Kholdeeva, O. V. Zalomaeva, Coord. Chem. Rev. 2016, 306, 302ꢀ330.
9. O. A. Kholdeeva, in Arene Chemistry: Reaction Mechanisms and Methods for Aromatic
Compounds, (Ed.: J. Mortier), Wiley, 2016, Ch. 14, pp.365ꢀ398
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
10. Lücke, B., Narayana, K. V., Martin, A., Jähnisch, K. Adv. Synth. Catal. 2004, 346, 1407ꢀ
1424.
11. Alonso, D. A., Nájera, C., Pastor, I. M., Yus, M. Chem. Eur. J. 2010, 16, 5274ꢀ5284.
12. Enthaler, S., Company, A. Chem. Soc. Rev. 2011, 40, 4912ꢀ4924.
13. Haines, A. H. Methods for the Oxidation of Organic Compounds: Alkanes, Alkenes, Alkynes
and Arenes, Academic Press: London, 1985, Vol. 1985.
14. Sheldon, R. A., Arends, I. W. C. E., Hanefeld, U. Green chemistry and catalysis, Wileyꢀ
VCH: Weinheim, 2007.
15. Cavani, F., Centi, G., Perathoner, S., Trifirò, F. Sustainable Industrial Chemistry: Principles,
Tools and Industrial Examples, WileyꢀVCH: Weinheim, 2009.
16. Liquid phase oxidation via heterogeneous catalysis: organic synthesis and industrial
applications, Clerici, M. G., Kholdeeva, O. A., Eds., Wiley: Hoboken, New Jersey, 2013.
17. Carril, M., Altmann, P., Drees, M., Bonrath, W., Netscher, T., Schütz, J., Kühn, F. E. J.
Catal. 2011, 283, 55ꢀ67.
18. Möller, K., Wienhöfer, G., Westerhaus, F., Junge, K., Beller, M. Catalysis Today 2011, 173,
68ꢀ75.
19. Bonrath, W., Netscher, T. Appl. Catal. A: Gen. 2005, 280, 55ꢀ62.
20. Casani, R. In Kirk-Othmer Encyclopedia of Chemical Technology, WileyꢀVCH: Weinheim,
2003, Vol. 25, p 144ꢀ150.
21. Asakawa, Y., Matsuda, R., Tori, M., Sono, M. J. Org. Chem. 1988, 53, 5453ꢀ5457.
ACS Paragon Plus Environment
29