Anti-Markovnikov hydration of alkenes over platinum-loaded titanium oxide photocatalyst

Article abstract of DOI:10.1039/c3cy00019b

Hydration of various alkenes proceeded over a platinum-loaded titanium oxide photocatalyst. It is notable that, unlike conventional acid-catalysts, the photocatalyst preferentially promoted the hydration according to the anti-Markovnikov rule. The results of some reaction experiments and semiempirical molecular orbital calculations suggested the reaction path of this hydration as follows: an electrophilic surface oxygen radical species formed on the photoexcited titanium oxide attacks a carbon double bond of the alkene to produce an intermediate, followed by the addition of a hydrogen radical formed on the platinum sites to produce an alcohol. The anti-Markovnikov selectivity for each investigated alkene would be determined by the stability of the radical intermediate.

Full text of DOI:10.1039/c3cy00019b

Catalysis
Science&Technology
www.rsc.org/catalysis
Volume 3 | Number 7 | July 2013 | Pages 1649–1856
Themed issue: Photocatalysis
ISSN 2044-4753
COVER ARTICLE
Yoshida et al.
Anti-Markovnikov hydration of alkenes over platinum-loaded
titanium oxide photocatalyst
2044-4753(2013)3:7;1-6

Catalysis
Science & Technology
View Article Online
View Journal | View Issue
PAPER
Anti-Markovnikov hydration of alkenes over
platinum-loaded titanium oxide photocatalyst†
Hayato Yuzawa,a Shoko Yoneyama,a Akihito Yamamoto,a Masanori Aoki,a
Kazuko Otake,a Hideaki Itohb and Hisao Yoshida*a
Cite this: Catal. Sci. Technol., 2013,
3, 1739
Hydration of various alkenes proceeded over a platinum-loaded titanium oxide photocatalyst. It is
notable that, unlike conventional acid-catalysts, the photocatalyst preferentially promoted the hydration
according to the anti-Markovnikov rule. The results of some reaction experiments and semiempirical
molecular orbital calculations suggested the reaction path of this hydration as follows: an electrophilic
surface oxygen radical species formed on the photoexcited titanium oxide attacks a carbon double
bond of the alkene to produce an intermediate, followed by the addition of a hydrogen radical formed
on the platinum sites to produce an alcohol. The anti-Markovnikov selectivity for each investigated
alkene would be determined by the stability of the radical intermediate.
Received 7th January 2013,
Accepted 31st January 2013
DOI: 10.1039/c3cy00019b
www.rsc.org/catalysis
1. Introduction
(1)
Hydration of alkenes (water addition to carbon double bond) is
an important industrial process for the direct synthesis of
alcohols, which are widely used as the chemical intermediates
for resins, agricultural chemicals, surfactants, and so on.
Although the selective hydration to produce branched alcohols
is possible by using homogeneous and heterogeneous acid
catalysts, e.g., sulfonic acid,1a heteropolyacid,1b zeolite,1c ion-
exchanged resin,1d etc., the acid catalysts are unable to produce
primary alcohols (except for ethylene to ethanol) or less sub-
stituted alcohols because of the strict regioselectivity known as
the Markovnikov rule. According to this rule, as shown in
eqn (1),2 the proton (H+) is firstly added to the carbon bonding
to less alkyl substituents in the carbon double bond, followed
by the addition of a functional group (Xꢀ, anion or neutral
molecule) to the carbon with more alkyl substituents. In the
case of the hydration, after the addition of the proton, a water
molecule adds to the carbocation intermediate, followed by the
deprotonation. This selectivity is determined by the stability of
the carbocation intermediate.
for a long time. Hydroboration by borane reagents followed by
oxidation,3 and hydroformylation by transition metal catalysts4
followed by hydrogenation are the very famous routes to give
the excellent anti-Markovnikov selectivity. However, both are
multistep processes and especially the former consumes the
boranes. Recently, one-pot syntheses of primary alcohols have
been attained, e.g., successive hydroformylation/hydrogenation
with the Rh/xantphos and Shvo’s catalyst,5 and Wacker-type
oxidation/hydrolysis/hydrogenation with the PdCl2 and Shvo’s
catalyst.6 In these syntheses, however, strictly defined reaction
conditions were required for each elementary process. So far,
direct hydration processes have been realised only for the
alkenes having an electron withdrawing group, such as photo-
hydration of nitro-substituted styrenes in the presence/absence
of a photosensitiser,7a photohydration of 1,1-diphenylpropene
in the presence of a photosensitising cyclodextrin host,7b and
hydration of such alkenes with phosphine catalysts.8 Thus, the
development of new simple processes that follow the anti-
Markovnikov rule are still required.
On the other hand, various processes to produce anti-
Markovnikov type alcohols from alkenes have been studied
a Department of Applied Chemistry, Graduate School of Engineering,
Nagoya University, Nagoya 464-8603, Japan. E-mail: yoshidah@apchem.nagoya-u.ac.jp;
Fax: +81-52-789-3178
b Division of Environmental Research, EcoTopia Science Institute,
Nagoya University, Nagoya 464-8603, Japan
Photocatalysis has received much attention because of its
unique potential to enable various organic syntheses in mild
conditions by using photoformed hole (h+) and electron (eꢀ),9
and could directly produce alcohols from alkanes (cyclohexane,
etc.) on titanium oxide,10a V2O5/Al2O3,10b Cr–Si binary oxide,10c
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
c3cy00019b
c
This journal is The Royal Society of Chemistry 2013
Catal. Sci. Technol., 2013, 3, 1739--1749 1739

View Article Online
Paper
Catalysis Science & Technology
CuCl2,10d and so on. However, primary alcohols were obtained by evacuation. Then, a gaseous substrate (1-butene or 1,3-
as minor products.10b,d To our knowledge, photocatalytic hydra- butadiene, 1.8 mmol) and liquid water (mainly 10 ml, corre-
tion of alkenes has not been reported yet.
sponding to 560 mmol) were introduced to the cell, where the
In our previous study, we succeeded in the selective aro- substrate should be partly dissolved in the water and the
matic ring hydroxylation by using water with a platinum-loaded photocatalyst also existed in the liquid phase, and the cell
titanium oxide photocatalyst and clarified that the oxygen was irradiated with the 300 W xenon lamp with magnetic
radical species produced from water were electrophilic.11 Thus, stirring of the liquid phase at 314 K. After the irradiation, the
in the present study, we applied this photocatalytic system to gas phase products were collected by a gas tight syringe and
the hydration of alkenes, and found that the photocatalytic analysed by a GC-TCD (Shimadzu, GC-8A). Then, the products
hydration of alkenes proceeded and the regioselectivity prefer- in the liquid phase were analysed by a GC-MS (Shimadzu,
entially followed the anti-Markovnikov rule.
QP-5050A) and a GC-FID (Shimadzu, GC-14B). In the latter
reactor, liquid substrates were investigated. The prepared
Pt(0.1)/TiO2 sample in the quartz cell was irradiated with the
300 W xenon lamp for 1 h in the presence of air. After the air in
the reactor was purged by argon, the substrate (1.0 ml, corre-
g
ꢀ1) and sponding to 8.0–10 mmol) except for thiophene (2.0 ml, corre-
2. Experimental section
2.1. Reagents
Titanium oxide sample, JRC-TIO-8 (anatase, 338 m2
JRC-TIO-6 (rutile, 100 m2 gꢀ1) were supplied by Catalysis Society sponding to 24 mmol) and water (1.0 ml or 10 ml) were
of Japan and the former was used unless stated. Precursors of the introduced to the cell, and the cell was irradiated with the
cocatalyst were chloroplatinic acid hexahydrate (H2PtCl6ꢁ6H2O, 300 W xenon lamp in an argon atmosphere with magnetic
Kishida, 99.9%), palladium chloride (PdCl2, Kishida, 99%), and stirring at 314 K for 60 or 180 min. After the irradiation, the
gold chloride acid tetrahydrate (HAuCl4ꢁ4H2O, Kishida, 99%). gas phase products were analysed by the GC-TCD (Shimadzu,
These reagents were used without further purification. Other GC-8A). Then, methanol was added to the liquid phase and
reagents were 1-butene (TCI, 98%), 1,3-butadiene (TCI, 95%), analysed by the GC-MS (Shimadzu, QP-5050A) and the GC-FID
1-hexene (Aldrich, 99%), cyclohexene (Wako, 97%), 1-methyl- (Shimadzu, GC-14B). In both systems, when the irradiation
1-cyclohexene (TCI, 98%), 2-methyl-1,3-butadiene (TCI, 99%), light wavelength was limited, an optical band pass filter per-
1,5-hexadiene (TCI, 97%), thiophene (Aldrich, 99%), D2O mitting the light transmission of 360 ꢂ 5 nm in wavelength or a
(Acros Organics, 99.95%), and H218O (Euriso-top, 95.4%). These cut-off filter permitting that of longer than 370 nm in wave-
reagents were used after purification by distillation, except for length was used and the reaction temperature was 305 K. As the
the isotopic-labeled water.
control experiments, reaction tests without the photocatalyst or
the photoirradiation were also carried out in the absence of the
photocatalyst or in the dark. These experiments confirmed that
the detected products in this article were produced by the
2.2. Preparation of metal-loaded titanium oxide
photocatalysts
Metal-loaded titanium oxide was prepared by a photodeposition photocatalysis on the M(x)/TiO2 samples.
method. The titanium oxide powder (4 g) was dispersed into an
aqueous methanol solution (400 ml, methanol concentration
25%) containing the metal precursor in a 500 ml beaker with
vigorous stirring, followed by irradiation from the top with a
3. Results and discussion
3.1. Outline of the photocatalytic reaction
300 W xenon lamp, which emitted both UV and visible light, at Table 1 shows the result of the photocatalytic reaction of 1,3-
room temperature for 3 h. Then, the suspension was filtered off butadiene over the M(0.1)/TiO2 samples. We found that some
with suction, washed with water, and dried at 323 K overnight. In kinds of metal-loaded TiO2 photocatalysts were able to promote
this article, x wt% of metal (M)-loaded titanium oxide is hydration of 1,3-butadiene. When the Pt(0.1)/TiO2 sample was
described as M(x)/TiO2. Measurements of ICP-AES with an used under irradiation without the limitation of the wavelength
Optima 3300 DV (Perkin-Elmer Japan) confirmed that the objec- (Table 1, entry 1), the products were 5.6 mmol of primary
tive amount of metal was correctly loaded on the titanium oxide. alcohols (2-butne-1-ol and 3-butene-1-ol), and 1.4 mmol of
The average platinum particle size on the Pt/TiO2 samples was 2-butenal, 3.2 mmol of coupling products and an excess amount
estimated by using a CO-pulse method in a flow system after the (110 mmol) of hydrogen, and thus the alcohol selectivity was low
reductive pretreatment with hydrogen at 423 K for 20 min.
(42%). On the other hand, when the irradiation wavelength was
limited to 360 ꢂ 5 nm, the coupling products and the excess
amount of hydrogen was suppressed and the higher alcohol
2.3. Photocatalytic reaction tests
Photocatalytic reaction tests were carried out by using two selectivity (76%) was obtained (Table 1, entry 2). This would be
kinds of reactors; one was a closed reactor connected with a mainly because the direct photoexcitation of 1,3-butadiene
vacuum line and the other was a closed reactor connected with adsorbed on the titanium oxide resulting in the formation of
an argon gas flow line. In the former system, the photocatalytic byproducts was suppressed.11b It is notable that, although the
reactions of gaseous substrates were investigated. The prepared hydration reaction with various acid catalysts usually follows
M(x)/TiO2 sample in a quartz cell was irradiated with the 300 W the Markovnikov rule,1 the current photocatalytic hydration of
xenon lamp from beneath in an air atmosphere for 1 h followed 1,3-butadiene proceeded according to the anti-Markovnikov
c
1740 Catal. Sci. Technol., 2013, 3, 1739--1749
This journal is The Royal Society of Chemistry 2013

View Article Online
Catalysis Science & Technology
Paper
Table 1 Results of the reaction tests for the photocatalytic hydration of 1,3-butadiene over the M(0.1)/TiO2 samples
Products/mmol
Anti-Markovnikov type
Markovnikov type
Alcohol
Alcohol
Anti-Markovnikov
Entry Loading metal
CPa H2
yieldb (%) selec.c (%) selec.d (%)
1e
2h
3h
4h
Pt
Pt
Pd
Au
5.6 (2.1 + 3.5)f
5.0 (1.5 + 3.5)f
5.2 (2.0 + 3.2)f
1.0 (0.5 + 0.5)f
1.4 (1.4 + n.d.)g n.d.
n.d.
n.d.
0.1
3.2
0.4
2.1
0.8
110
0.31
5.8 0.28
6.1 0.29
5.0 0.056
42
76
50
33
100
100
97
0.8 (0.7 + 0.1)g
0.9 (0.8 + 0.1)g
n.d.
0.1
0.4 (0.4 + n.d.)g n.d.
n.d.
100
a
b
Coupling products, mainly C8 hydrocarbons. The yields of other coupling products were very small. Alcohol yield (%) = 100 ꢄ
c
YAlcohols/(introduced 1,3-butadiene), where Y stands for the yielded amount of product. Alcohol selectivity (%) = 100 ꢄ (YPrimary
+
+
alcohols
d
Y2-Butanol)/(YPrimary alcohols + Y2-Butanol + YAlehydes + Y2-Butanone + 2 ꢄ YCoupling products). Anti-Markovnikov selectivity (%) = 100 ꢄ (YPrimary alcohols
e
YAldehydes)/(YPrimary alcohols + Y2-Butanol + YAldehydes + Y2-Butanone). Catalyst, 0.2 g; 1,3-butadiene, 1.8 mmol; water, 10 ml (560 mmol); reaction time,
180 min; reaction temperature, 314 K; irradiated light wavelength was not limited. Irradiated light intensities were 13, 54, and 148 mW cmꢀ2 when
measured at 254 ꢂ 10, 365 ꢂ 15, and 405 ꢂ 30 nm, respectively. The detection limits of the organic compounds, hydrogen and carbon dioxide were
f
0.01, 0.1 and 1.0 mmol, respectively. The left and right values in the parenthesis correspond to the yield of 3-buten-1-ol and 2-butene-1-ol,
g
respectively. The left and right values in the parenthesis correspond to the yield of 2-butenal and butanal, respectively. n.d. = not detected.
h
Reaction temperature, 305 K; irradiated light wavelength, 360 ꢂ 5 nm; irradiated light intensity was 8.4 mW cmꢀ2 when measured at
365 ꢂ 15 nm. The other conditions were the same as the footnote e.
rule very selectively (almost 100%) regardless of the irradiated having a larger work function would be effective for the
light wavelength (Table 1, entries 1 and 2), i.e., only primary separation of the photoformed hole–electron pairs, and the
alcohols were produced while no secondary alcohols were larger amount of such effective small nanoparticles would
observed. In addition, such high anti-Markovnikov selectivity provide the best performance.12 The size and number of the
was also observed on the samples with the other loading platinum nanoparticles determined by the CO adsorption
metals, such as palladium and gold (97 and 100%, Table 1, method were shown in Fig. 1c. The Pt(0.1)/TiO2 sample had
entries 3 and 4), although they promoted the reaction with less the largest number of platinum nanoparticles with the almost
alcohol selectivity even under the light of limited wavelength.
smallest size among them, which agrees with the above dis-
Table 2 shows the results of the photocatalytic reaction tests cussion. On the other hand, the high anti-Markovnikov selec-
of various alkenes in the optimised conditions for each sub- tivity did not vary with increasing platinum loading amount
strate. From all of the investigated alkenes, the corresponding although the average diameter of the platinum also increased
alcohols were produced, indicating that this photocatalytic (Fig. 1b and c). Thus, the loading amount, i.e., the diameter and
system is available for various alkenes in both gas and liquid the number of the platinum nanoparticles, would not be
states. The alcohols were observed in the liquid phase in the controlling factor for the anti-Markovnikov selectivity, as
each case. Further, the anti-Markovnikov type products were discussed later.
preferentially (65–100%) produced from all of the substrates
Fig. S1 and S2, (ESI†) show the influence of the crystal phase
except for the cyclohexene having a symmetric double bond, and the specific surface area of titanium oxide on the photo-
although the anti-Markovnikov selectivity was different catalytic activity for the hydration of 1-hexene (Fig. S1, ESI†) and
between the listed substrates. This will be discussed later. thiophene (Fig. S2, ESI†). As for the specific surface area of the
In the case of 1-methyl-1-cyclohexene, the alcohol according Pt(0.1)/TiO2(anatase) sample, the total yield increased with
to the anti-Markovnikov rule (1.2 mmol) consisted of cis- and increasing the specific surface area while the anti-Markovnikov
trans-isomers of similar amounts, 0.7 and 0.5 mmol, respectively selectivity was almost constant for both substrates (Fig. S1b
(Table 2, entry 5), which will be also discussed later. For the and S2b, ESI†). This would be because catalytically active sites on
cyclohexene, the alcohol selectivity was especially low (17%) the titanium oxide would increase with increasing the specific
because it is difficult to inhibit the coupling reaction in this surface area as reported in the literature.12,13 As for the crystal
condition. For the other substrates, the alcohol selectivity phase of titanium oxide, the anatase exhibited higher photo-
widely varied (38–83%) because of the yield of the carbonyl catalytic activity than the rutile having almost the same specific
compounds.
surface area (Fig. S1b and S2b, ESI†). Since the potential of the
Fig. 1 shows the effect of the platinum loading amount on the photoformed hole in the rutile (3.04 V vs. NHE)14a is lower than
photocatalytic reaction of 1,3-butadiene. As shown in Fig. 1a, the that in the anatase (3.12 V vs. NHE),14b the rutile would less
highest amount of alcohol was obtained on the Pt(0.1)/TiO2 promote the oxidation of water (e.g., H2O + h+
-
ꢃOH + H+,
sample. In principle, the smaller platinum nanoparticles 2.8 V vs. NHE)14c than the anatase. Thus, anatase titanium
c
This journal is The Royal Society of Chemistry 2013
Catal. Sci. Technol., 2013, 3, 1739--1749 1741

View Article Online
Paper
Catalysis Science & Technology
Table 2 Results of the reaction tests for the photocatalytic hydration of various alkenes over the Pt(0.1)/TiO2 sample
Alcohol Alcohol Anti-
yieldc
selec.d Markovnikov
Entry Substrate
Products/mmol
CPb H2 (10ꢀ2%) (%)
selec.e (%)
1a
3.7
2.6
7.7f 35 3.7
17
—
Anti-Markovnikov type
Markovnikov type
2g,h
1.4 (1.3 + 0.1)i
0.1
3.4
0.3
0.8
0.9f 23 2.6
n.d.k 11 75
0.1l 6.6 0.50
38
73
40
65
72
75
3j
10
3.5
0.1
1.7
4a
0.3
0.1
5a
1.2m [cis-0.7 trans-0.5] 1.3
0.6
n.d.k 2.6 2.1
58
83
81
6n
25.2
0.3
n.d.k
n.d.k
0.1o 62 11
100
Anti-Markovnikov type
7g
1.3 (1.1 + 0.2)i
0.9 (0.8 + 0.1)i
0.1
0.2p 7.4 2.2
81
100
a
Catalyst, 0.2 g; substrate, 1.0 ml (8.0–9.9 mmol); water, 1.0 ml (56 mmol); reaction time, 60 min; reaction temperature, 305 K; irradiated light
wavelength, 4370 nm; irradiated light intensities were 3, 10 and 84 mW cmꢀ2 when measured at 254 ꢂ 10, 365 ꢂ 15 and 405 ꢂ 30 nm, respectively.
The detection limit of the organic compounds was 0.01 mmol, that of hydrogen was 0.1 mmol, and that of carbon dioxide was 1.0 mmol.
b
c
d
CP = coupling products. Alcohol yield (%) = 100 ꢄ YAlcohols/(introduced substrate), where Y stands for the yielded amount of product. Alcohol
e
selectivity (%) = 100 ꢄ YAlcohols/(YAlcohols + YCarbonyl
+ 2 ꢄ YCP). Anti-Markovnikov selectivity (%) = 100 ꢄ (YAnti-Markovnikov
+
type alcohol
products
f
YAnti-Markovnikov type carbonyl product)/(YAlcohols + YCarbonyl products). Dimer (C12) hydrocarbons were produced. The detailed structure was not analysed.
g
Catalyst, 0.2 g; substrate, 1.0 ml (8.4–10 mmol); water, 10 ml (560 mmol); reaction time, 180 min; reaction temperature, 305 K; irradiated light
wavelength, 360 ꢂ 5 nm; irradiated light intensity was 8.4 mW cmꢀ2 when measured at 365 ꢂ 15 nm. Also detected was 10 mmol of carbon
h
i
dioxide. The left and right values in the parenthesis corresponds to the yield of unsaturated and saturated alcohol in the two structural formulas.
j
k
l
1-Butene, 1.8 mmol; water, 10 ml (560 mmol); the other conditions were the same as the footnote g. n.d. = not detected. 5-Methyl undecane.
m
The yields of other coupling products were very small. Cis, trans isomers were produced. The yield of each product is shown in the parenthesis.
n
Thiophene, 2.0 ml (24 mmol); water, 8.0 ml (440 mmol); reaction time, 60 min; reaction temperature, 314 K; irradiated light intensities were 23,
80, and 176 mW cmꢀ2 when measured at 254 ꢂ 10, 365 ꢂ 15 and 405 ꢂ 30 nm, respectively. Also detected was 18 mmol of carbon dioxide.
o
p
Bithiophenes. Dimer (C10) hydrocarbons were produced. The detailed structure was not analysed.
c
This journal is The Royal Society of Chemistry 2013
1742 Catal. Sci. Technol., 2013, 3, 1739--1749

View Article Online
Catalysis Science & Technology
Paper
Fig. 1 The effect of the platinum loading amount on the products yield (a) and selectivity (b) in the photocatalytic hydration of 1,3-butadiene over the Pt(x)/TiO2
samples. The red filled circles, red blank triangles, green filled squares and black filled diamonds in panel a correspond to the total yield of 2-butene-1-ol and 3-butene-
1-ol, that of 2-butenal and butanal, that of the coupling products, and the yield of hydrogen, respectively. The orange filled inverse triangles and orange blank inverse
triangles in panel b correspond to the anti-Markovnikov selectivity and the alcohol selectivity, respectively. The blue blank circles and pink blank squares in panel c
correspond to the average diameter and the number of the platinum nanoparticles, respectively. The average diameter of the platinum nanoparticles was determined
by the CO pulse method. As for the reaction conditions, see the footnote a of Table 1.
Table 3 Results of the isotopic experiments in the photocatalytic hydration of
cyclohexene over the Pt(0.1)/TiO2 samplea
oxide with a large surface area would be most effective for the
hydration.
The magnitude of the influence by the specific surface area
Entry
1b
Water
D2O
Distribution of produced cyclohexanol (%)
was different between the substrates, which would be derived
from the properties of the substrates although the details were
not clarified in the present study.
499
n.d.c
13
3.2. Active species
Table 3 shows the results of the isotopic experiments in the
photocatalytic hydration of cyclohexene. When the reaction was
carried out by using D2O, the mass number (m/z) of the
molecular ion peak for most of the produced cyclohexanol
(499%) increased from 100 to 101 (Table 3, entry 1). In the
present analysis method, the D atom of the hydroxyl group in
the produced alcohol can be easily exchanged with the H atom
of methanol that was added for homogenisation of the liquid
phase after the reaction before the analysis. Thus, the increase
of the mass number would originate from the D atom of D2O,
which was directly added to the carbon double bond. This
result suggests that the alcohol was dominantly produced
through photocatalytic hydration (water addition) of the carbon
double bond as shown in eqn (2).
2d
H218O
87
a
Catalyst, 0.2 g; cyclohexene, 1.0 ml (9.9 mmol); water, 1.0 ml
(56 mmol); reaction time, 60 min; reaction temperature, 305 K; irra-
diated light wavelength, 4370 nm; irradiated light intensities were 3,
10 and 84 mW cmꢀ2 when measured at 254 ꢂ 10, 365 ꢂ 15 and 405 ꢂ
30 nm, respectively. The detection limit of the organic compounds was
0.01 mmol. In the present analysis method, the D atom of the hydroxyl
group in the produced alcohol can be exchanged easily with H of the
methanol that was introduced for analysis after the reaction. n.d. = not
detected. Before the reaction, the photocatalyst was evacuated at 473 K
for 2 h to remove the physisorbed water on the Pt(0.1)/TiO2 sample.
b
c
d
When the reaction was carried out by using H218O, the
produced cyclohexanol contained 87% of 18O atom and 13%
of 16O in the hydroxyl group (Table 3, entry 2). This indicates
that the introduced water dominantly reacts with the alkene,
while the lattice oxygen on the titanium oxide surface also
(2)
c
This journal is The Royal Society of Chemistry 2013
Catal. Sci. Technol., 2013, 3, 1739--1749 1743

View Article Online
Paper
Catalysis Science & Technology
reacts with the alkene. In the literature,15 it was reported that the bond energy of O–H in water (497.10 ꢂ 0.29 kJ molꢀ1 at 298 K)
upon photoirradiation the reaction between water molecules is larger than that of C–H in hydrocarbons (e.g., butane 421.3 and
and the titanium oxide surface could produce an active surface 411.1 ꢂ 2.2 kJ molꢀ1, cyclohexane 416.3 kJ molꢀ1 at 298 K).18
oxygen species and the mechanism in the aqueous solution of
acidic or neutral pH was proposed as follows: a bridging Ti–O–Ti
(5)
site on the surface would trap the photoformed hole, and receive
the nucleophilic attack of water to produce the surface oxygen
radical (eqn (3)). The surface oxygen radical species would have
an electrophilic property and preferably attack the carbon
double bond having high electron density.11b
(6)
(3)
The differences of the final heat of formation for the two
possible isomers of the hydroxyl radical adduct were also listed
(Table 4, A–C). It is notable that the value decreased with
increasing anti-Markovnikov selectivity, meaning that a larger
difference between the two possible radical intermediates
resulted in a higher anti-Markovnikov selectivity. Thus, the
difference of the stability between the intermediates (eqn (5))
would determine the regioselectivity of the hydroxyl group for
each substrate.
On the other hand, the photoformed electron would reduce
the proton in the solution to produce a hydrogen radical at the
platinum sites (eqn (4)).
(4)
Considering eqn (3) and (4), and the results of the isotopic
experiments (Table 3), the hydration of the carbon double bond
would proceed through the addition of the electrophilic surface
oxygen radical and the hydrogen radical. This implies that the
reaction would preferably progress on the titanium oxide sur-
face around the deposited platinum nanoparticles.
3.4. Successive reactions
Table 5 shows the results of the photocatalytic reaction tests for
the successive reactions of 1-butene conversion by using the
alcohols and the carbonyl compounds as the reactants. Butanal
and 2-butanone were converted not to alcohols but to hydrogen
and carbon dioxide (Table 5, entries 1 and 2), where only the
complete oxidation with water occurred. On the other hand,
1-butanol and 2-butanol were oxidised into the carbonyl
compounds (Table 5, entries 3 and 4), that is, the successive
oxidation could occur after the hydration. Thus, the alcohol can
be oxidised to carbonyl compounds as shown in eqn (7), and
further oxidised to the final products, i.e., hydrogen and carbon
dioxide.
3.3. Radical intermediates
To discuss the order of the addition of these two radical species
to the carbon double bond, we calculated the final heat of
formation of the possible intermediate species, i.e., the active
oxygen adduct and the hydrogen radical adduct (Table 4), by
using PM7 method16 in MOPAC 2012.17 In this calculation, the
hydroxyl radical (ꢃOH) was employed instead of the surface
active oxygen species (eqn (3)) in order to simplify the system
because they have similar thermodynamic properties to each
other.11b In the case of the hydroxyl radical adduct, the radical
intermediate in the anti-Markovnikov hydration (Table 4, A)
was more stable than that in the Markovnikov hydration
(Table 4, C) for all the investigated substrates. On the other
hand, in the case of the hydrogen radical adduct, the radical
intermediate in the Markovnikov hydration (Table 4, D) was
more stable than that in the anti-Markovnikov hydration
(7)
Fig. 2 shows the time course of the products yield in the
(Table 4, B). These results clearly indicate that if the hydrogen photocatalytic hydration of 1-butene in the closed reactor.
radical attacks first to the carbon double bond before the Although the total production rate of the alcohols and the
surface oxygen radical does, the product would not follow the carbonyl products decreased once after 1 h, it became constant
anti-Markovnikov rule, which was inconsistent with the fact. at least until 24 h (Fig. 2b, purple filled squares). It is notable
Thus, it is suggested that the surface oxygen radical would be that the carbonyl compounds were produced at a constant rate
first added to the carbon double bond (eqn (5)), followed by the even in the initial stage of the photocatalytic reaction. If the
addition of the hydrogen radical (eqn (6)). In this way, the order carbonyl compounds were produced only through the succes-
of the radicals addition to the carbon double bond, i.e., the sive oxidation of the alcohols (eqn (7)), the production of the
surface oxygen radical and then hydrogen radical, would lead to the carbonyl compounds should be delayed compared to that of
anti-Markovnikov products. Although one may consider another the alcohols, which is inconsistent with the results. Thus, the
possibility instead of eqn (6), that the intermediate species carbonyl compounds are also produced directly and competi-
abstracts a hydrogen atom from water, this could not occur because tively. In the previous study for the aromatic ring hydroxylation,11b
c
1744 Catal. Sci. Technol., 2013, 3, 1739--1749
This journal is The Royal Society of Chemistry 2013

View Article Online
Catalysis Science & Technology
Paper
Table 4 Relationship between anti-Markovnikov selectivity in the photocatalytic hydration and the final heat of formation of the possible radical intermediatesa
Final heat of formation/kJ molꢀ1
Anti-Markovnikov type
ꢃOH adduct (A)
Markovnikov type
Difference
Anti-Markovnikov
selec.b (%)
ꢃH adduct (B)
ꢃOH adduct (C) ꢃH adduct (D)
(A–C)
(B–D)
Entry Substrate
1
ꢀ12.4 27.4
65
72
75
ꢀ89.5
ꢀ163.0
ꢀ205.9
99.0
25.1
ꢀ77.1
ꢀ143.4
ꢀ185.8
71.6
ꢀ1.74
ꢀ44.4
2
3
ꢀ19.6 26.8
ꢀ20.1 27.2
ꢀ17.2
4
ꢀ21.1 12.2
81
ꢀ218.8
ꢀ19.9
ꢀ197.7
ꢀ42.1
5
ꢀ24.0 21.7
ꢀ46.8 56.4
ꢀ50.7 51.3
100
100
100
ꢀ38.4
ꢀ74.5
151.9
142.4
ꢀ14.4
ꢀ27.7
130.2
86.0
6
7c
ꢀ111.2
ꢀ113.5
102.3
121.8
ꢀ60.5
ꢀ56.4
51.0
45.0
8
ꢀ57.1 76.8
100
a
Final heat of formation was calculated by PM7 method16 in MOPAC2012.17 As for the reaction conditions, see the footnotes in Tables 1 and 2.
See the foot note e in Table 2.
b
it was suggested that the hydrogen atom could be abstracted
Considering these reaction routes to produce carbonyl com-
from the intermediate by a radical species. In a similar way, the pounds (eqn (7) and (8)), the regioselectivity of the carbonyl
carbonyl compounds would be produced through the abstrac- group would be the same as that of the hydroxyl group in the
tion of hydrogen from the intermediate species as shown in produced alcohol because of the difference of the stability
eqn (8).
between the intermediates as mentioned above.
3.5. Effect of water quantity
Fig. 3 shows the influence of water quantity on the photo-
catalytic hydration of 1-butene as a gaseous reactant, where it
was a three phase reaction system (gas–liquid–solid). With
increasing the quantity of the liquid water to 10 ml in volume,
(8)
c
This journal is The Royal Society of Chemistry 2013
Catal. Sci. Technol., 2013, 3, 1739--1749 1745

View Article Online
Paper
Catalysis Science & Technology
Table 5 Results of photocatalytic reaction tests of hydration products over the Pt(0.1)/TiO2 samplea
Products/mmol
H2
54
43
CO2
11
Entry
1
Substrate
n.d.
n.d.
4.8
n.d.
n.d.
n.d.
6.8
2
n.d.
8.2
3
4
13.0
n.d.
7.6
n.d.
14.2
n.d.
4.7
54
51
4.9
5.6
n.d.
a
Catalyst, 0.2 g; substrate, 25 mmol in the aqueous solution (10 ml); reaction time, 180 min; reaction temperature, 314 K; light intensities were 13,
54, and 148 mW cmꢀ2 when measured at 254 ꢂ 10, 365 ꢂ 15, and 405 ꢂ 30 nm, respectively. n.d. = not detected.
Fig. 2 The time course of the products yield (a), and the total yield and
Fig. 3 The influence of water quantity on the photocatalytic hydration of
1-butene over the Pt(0.1)/TiO2 sample. The red filled circles, red filled triangles, blue
blank circles, blue blank triangles and black filled diamonds in panel a correspond to
the yields of 1-butanol, butanal, 2-butanol, 2-butanone and hydrogen, respectively.
The orange filled inverse triangles and orange blank inverse triangles in panel b
correspond to the anti-Markovnikov selectivity and the alcohol selectivity, respec-
tively. Catalyst, 0.2 g; 1-butene, 1.8 mmol; reaction time, 180 min; reaction
temperature, 314 K; light intensities were 13, 54, and 148 mW cmꢀ2 when
measured at 254 ꢂ 10, 365 ꢂ 15 and 405 ꢂ 30 nm, respectively.
selectivity (b) in the photocatalytic hydration of 1-butene over the Pt(0.1)/TiO2
sample. The red filled circles, red filled triangles, blue blank circles, blue blank
triangles, black filled diamonds and pink blank diamonds in panel a correspond
to the yields of 1-butanol, butanal, 2-butanol, 2-butanone, hydrogen and carbon
dioxide, respectively. The purple filled squares, orange filled inverse triangles and
orange blank inverse triangles in panel b correspond to the total yield of the
alcohols and the carbonyl products, the anti-Markovnikov selectivity and the
alcohol selectivity, respectively. Catalyst, 0.2 g; 1-butene, 1.8 mmol; water, 10 ml
(560 mmol); reaction temperature, 314 K; irradiated light intensities were 13, 54,
and 148 mW cmꢀ2 when measured at 254 ꢂ 10, 365 ꢂ 15 and 405 ꢂ 30 nm,
respectively.
small (1.0 ml), the color of the catalyst changed, implying that
the yield of the alcohols and carbonyl products increased and undetectable byproducts (e.g., coupling products) would be
that of excess hydrogen decreased (Fig. 3a). In addition, the produced from the 1-butene and the hydration products. This
alcohol selectivity and the anti-Markovnikov selectivity also means that the presence of water would contribute to not only
increased. When the introduced amount of water was very the formation of alcohol and carbonyl products (eqn (9) and (10))
c
1746 Catal. Sci. Technol., 2013, 3, 1739--1749
This journal is The Royal Society of Chemistry 2013

View Article Online
Catalysis Science & Technology
Paper
but also the desorption of the products from the surface as a 3.6. Proposed reaction mechanism
solvent to prohibit the excess adsorption of the products
and side reactions such as successive oxidation and coupling
reactions.
From the above investigations and our previous study, the
proposed reaction mechanism of the photocatalytic hydration
of the carbon double bond is summarised in Fig. 4. In the
current three phase reaction systems (gas–water–solid or oil–
water–solid), the catalyst obviously exists in the aqueous phase.
Thus, water molecules are always adsorbed on the titanium
oxide surface (Fig. 4, i). The titanium oxide is photoexcited to
produce an electron and a hole, and then the electron migrates
to the platinum sites and the hole is trapped by the surface
bridge oxygen sites of titanium oxide as usually proposed
(Fig. 4, ii).15 The trapped hole induces the nucleophilic attack
of the adsorbed water to produce a surface oxygen radical and a
surface hydroxyl group. An alkene molecule would interact with
the hydrogen of the adsorbed water or the surface hydroxyl
group (Fig. 4, iii). The electrophilic surface oxygen radical
attacks the double bond to produce a radical intermediate
(Fig. 4, iv or iv0). It should be noted that the position selectivity
of the surface oxygen radical to the two carbon atoms at the
carbon double bond almost follows the anti-Markovnikov rule
according to the stability of the radical intermediates (Fig. 4, iv).
(9)
(10)
On the other hand, when the volume of water was larger Finally, the electron in the platinum sites reduces a proton to
than 10 ml, the yield of the hydration products decreased produce a hydrogen radical, which adds to the intermediate
although the selectivity was not changed. With an increased species to produce a surface alkoxide species, and it desorbs as
volume of water in the closed system, the decreased concen- an alcohol (Fig. 4, v and v0), which is promoted by water. In this
tration of 1-butene in the aqueous solution might reduce the mechanism, the lattice oxygen atom of the titanium oxide
reaction rate.
surface is introduced to the product while the oxygen atom in
Fig. 4 The proposed reaction mechanism of the photocatalytic hydration of alkene to produce alcohol.
c
This journal is The Royal Society of Chemistry 2013
Catal. Sci. Technol., 2013, 3, 1739--1749 1747

View Article Online
Paper
Catalysis Science & Technology
Fig. 5 The proposed mechanism of the photocatalytic hydration of 1-methyl-1-cyclohexene to produce the cis–trans isomers of 2-methylcyclohexanol. The above
steps correspond to the steps iv and v in Fig. 4.
the water molecule is introduced to the surface in the first cycle, oxide photocatalyst. In all of the investigated alkenes, the
and then the surface oxygen atoms that originated from the hydration products were obtained. Interestingly, the obtained
water molecule are introduced into the product in the succes- regioselectivity of the hydroxyl group preferentially followed the
sive cycles. This is supported by the isotopic experiment with anti-Markovnikov rule, i.e., the primary alcohols or the less
H218O (Table 3, entry 2) and the previous study.11b
substituted alcohols were predominantly produced.
In this hydration reaction, anti-Markovnikov selectivity was
The surface oxygen participates in the present hydration to
not influenced by the platinum loading amount (the average be involved in the products. For the anti-Markovnikov hydra-
size and the number of the platinum nanoparticle, Fig. 1). Since tion, the order of the radical addition is important, i.e., the
the key step of determining the anti-Markovnikov selectivity electrophilic surface oxygen radical, which is produced through
(the addition of the surface oxygen radical) would take place on the reaction between water and the lattice oxygen of the
the titanium oxide surface, this step would not be significantly titanium oxide surface trapping the photoformed hole, should
influenced by the platinum loading amount.
attack the double bond of the alkene to form the radical
From 1-methyl-1-cyclohexene, cis–trans-isomers of 2-methyl- intermediate before the addition of the hydrogen radical, which
cyclohexanol were produced in the present photocatalytic anti- is produced from the proton and the photoformed electron on
Markovnikov hydration. The above hydration mechanism and the platinum sites. At that time, the anti-Markovnikov selecti-
the almost equal yields of cis–trans isomers suggest that the vity would be determined by the stability of the radical inter-
intermediate species can freely rotate in the C–O bond and be mediate, which depends on the nature of substrate.
attacked by the hydrogen radical from both sides as shown
in Fig. 5.
The carbonyl compounds as byproducts were formed by
successive oxidation or by direct photocatalytic oxidation, which
A part of the produced alcohol is further oxidised to produce also followed the anti-Markovnikov rule, since the reaction
carbonyl compounds probably with the formation of hydrogen mechanism would be similar to the hydration.
(eqn (7)). On the other hand, the direct production of carbonyl
The water would also function as a solvent to assist the
compounds would competitively occur. The mechanism is desorption of the products, resulting in the decrease of the
proposed in Fig. S3, (ESI†), in which the steps (iv) and (v) are formation of undesired byproducts.
different from the above. After the formation of the intermediate
(Fig. S3, iii, ESI†), titanium oxide is further photoexcited to form
a hole and an electron (Fig. S3, iv and iv0, ESI†). The hole
Acknowledgements
abstracts the hydrogen atom from the intermediate to produce
an adsorbed enoxy species (Fig. S3, v and v0, ESI†). Finally, the
enol is desorbed from the surface, and the tautomerisation
occurs to produce a carbonyl compound (Fig. S3, vi and vi0,
ESI†) while the electrons in the platinum sites reduce protons to
form a hydrogen molecule (Fig. S3, vii, ESI†).
We thank Dr Kazunari Takahashi and Dr Hidenobu Kajitani
(Mitsubishi Chemical Corporation) for the valuable discussion.
This research was financially supported by Mitsubishi Chemical
Corporation, Grant-in-Aid for JSPS fellows (22–7799) and the
fund for doctoral students in Nagoya University.
Notes and References
4. Conclusion
1 (a) J. L. Jensen, V. Uaprasert and C. R. Fujii, J. Org. Chem.,
1976, 41, 1675; V. J. Nowlan and T. T. Tidwell, Acc. Chem.
Res., 1977, 10, 252; D. J. Nelson, C. Brammer and R. Li,
We investigated the hydration of various alkenes (from C4 to
C7, and thiophene) over the 0.1 wt% platinum-loaded titanium
c
1748 Catal. Sci. Technol., 2013, 3, 1739--1749
This journal is The Royal Society of Chemistry 2013

View Article Online
Catalysis Science & Technology
Paper
Tetrahedron Lett., 2009, 50, 6454; (b) J. Po´zniczek,
2002, 102, 3811; O. Carp, C. L. Huisman and A. Reller, Prog.
Solid State Chem., 2004, 32, 33; G. Palmisano, V. Augugliaro,
M. Pagliaro and L. Palmisano, Chem. Commun., 2007, 3425;
Y. Shiraishi and T. Hirai, J. Photochem. Photobiol., C, 2008,
9, 157.
´
´
A. M. Lubanska, A. M. Ilnicka and A. Bielanski, Appl. Catal.,
ˆ
A, 1999, 176, 101; A. V. Ivanov, E. Zausa, Y. B. Taarit and
`
N. Essayem, Appl. Catal., A, 2003, 256, 225; G. Marcı, E. I. G.
´
Lopez and L. Palmisano, Appl. Catal., A, 2012, 421–422, 70;
(c) D. Nuntasri, P. Wu and T. Tatsumi, Chem. Lett., 2002, 10 (a) P. Du, J. A. Moulijn and G. Mul, J. Catal., 2006, 238, 342;
224; J. Li, Z. Qin, H. Xu, M. Dong, J. Dong and J. Wang, Ind.
Eng. Chem. Res., 2007, 46, 9000; X. Shan, Z. Cheng and Y. Li,
J. Chem. Eng. Data, 2011, 56, 4310; (d) A. Chakrabarti and
M. M. Sharma, React. Polym., 1992, 18, 117; S. Talwalkar,
P. Kumbhar and S. Mahajani, Ind. Eng. Chem. Res., 2006,
45, 8024; X. Shan, Z. Cheng and P. Yuan, Chem. Eng. J., 2011,
175, 423.
2 M. Beller, J. Seayad, A. Tillack and H. Jiao, Angew. Chem., Int.
Ed., 2004, 43, 3368.
3 H. C. Brown and G. Zweifel, J. Am. Chem. Soc., 1961,
(b) K. Teramura, T. Tanaka, T. Hosokawa, T. Ohuchi,
M. Kani and T. Funabiki, Catal. Today, 2004, 96, 205;
(c) Y. Shiraishi, Y. Teshima and T. Hirai, Chem. Commun.,
2005, 4569; (d) K. Takaki, J. Yamamoto, K. Komeyama,
T. Kawabata and K. Takehira, Bull. Chem. Soc. Jpn., 2004,
77, 2251.
11 (a) H. Yoshida, H. Yuzawa, M. Aoki, K. Otake, T. Hattori and
H. Itoh, Chem. Commun., 2008, 4634; (b) H. Yuzawa,
M. Aoki, K. Otake, T. Hattori, H. Itoh and H. Yoshida,
J. Phys. Chem. C, 2012, 116, 25376.
83, 2544; T. C. Morrill, C. A. D’Souza, L. Yang and 12 H. Yuzawa, T. Mori, H. Itoh and H. Yoshida, J. Phys. Chem.
A. J. Sampognaro, J. Org. Chem., 2002, 67, 2481. C, 2012, 116, 4126.
4 F. Ungvary, Coord. Chem. Rev., 2007, 251, 2072; F. Ungvary, 13 H. Kominami, S. Murakami, J. Kato, Y. Kera and B. Ohtani,
´
´
Coord. Chem. Rev., 2007, 251, 2087.
5 K. Takahashi, M. Yamashita, T. Ichihara, K. Nakano and
K. Nozaki, Angew. Chem., Int. Ed., 2010, 49, 4488.
J. Phys. Chem. B, 2002, 106, 10501; L. Cao, Z. Gao, S. L. Suib,
T. N. Obee, S. O. Hay and J. D. Freihaut, J. Catal., 2000,
196, 253.
6 G. Dong, P. Teo, Z. K. Wickens and R. H. Grubbs, Science, 14 (a) A. Shiga, A. Tsujiko, S. Yae and Y. Nakato, Bull. Chem.
¨
2011, 333, 1609; J. Kanand, R. Paciello and M. Roper,
Soc. Jpn., 1998, 71, 2119; (b) D. Duonghong, J. Ramsden and
¨
US 6166265, 2000; L. Hintermann, ChemCatChem, 2012, 4, 321.
7 (a) P. Wan, S. Culshaw and K. Yates, J. Am. Chem. Soc., 1982,
M. Gratzel, J. Am. Chem. Soc., 1982, 104, 2977; (c) C. D. Jaeger
and A. J. Bard, J. Phys. Chem., 1979, 83, 3146.
104, 2509; P. Wan, M. J. Davis and M. A. Teo, J. Org. Chem., 15 A. Imanishi, T. Okamura, N. Ohashi, R. Nakamura and
1989, 54, 1354; (b) G. Fukuhara, T. Mori and Y. Inoue, J. Org.
Chem., 2009, 74, 6714.
Y. Nakato, J. Am. Chem. Soc., 2007, 129, 11569.
16 J. J. P. Stewart, J. Mol. Model., 2013, 19, 1.
8 I. C. Stewart, R. G. Bergman and F. D. Toste, J. Am. Chem. 17 J. J. P. Stewart, MOPAC2012; Stewart Computational Chemistry,
Soc., 2003, 125, 8696. Colorado Springs, CO, 2012, see http://openmopac.net/.
9 M. A. Fox and M. T. Dulay, Chem. Rev., 1993, 93, 341; 18 D. R. Lide, Handbook of Chemistry and Physics, CRC Press,
A. Maldotti, A. Molinari and R. Amadelli, Chem. Rev.,
London, 84th edn, 2003.
c
This journal is The Royal Society of Chemistry 2013
Catal. Sci. Technol., 2013, 3, 1739--1749 1749