Photocatalytic Transfer Hydrogenolysis of Allylic Alcohols on Pd/TiO2: A Shortcut to (S)-(+)-Lavandulol

Article abstract of DOI:10.1002/chem.201704099

We report herein a regio- and stereoselective photocatalytic hydrogenolysis of allylic alcohols to form unsaturated hydrocarbons employing a palladium(II)-loaded titanium oxide; the reaction proceeds at room temperature under light irradiation without stoichiometric generation of salt wastes. Olefin and saturated alcohol moieties tolerated the reaction conditions. Hydrogen atoms were selectively incorporated into less sterically congested carbons of the allylic functionalities. This protocol allowed a short-step synthesis of (S)-(+)-lavandulol from (R)-(?)-carvone by avoiding otherwise necessary protection/deprotection steps.

Full text of DOI:10.1002/chem.201704099

A Journal of
Accepted Article
Title: Photocatalytic transfer hydrogenolysis of allylic alcohols on Pd/
TiO2: A Shortcut to (S)-(+)-lavandulol
Authors: Susumu Saito, Yuki Takada, Joaquim Caner, Selvam
Kaliyamoorthy, and Hiroshi Naka
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201704099
Link to VoR: http://dx.doi.org/10.1002/chem.201704099
Supported by

10.1002/chem.201704099
Chemistry - A European Journal
FULL PAPER
Photocatalytic transfer hydrogenolysis of allylic alcohols on
Pd/TiO2: A Shortcut to (S)-(+)-lavandulol
Yuki Takada,[a] Joaquim Caner,[b] Selvam Kaliyamoorthy,[a] Hiroshi Naka[a,b] and Susumu Saito* [a,b]
Abstract: We report herein
a
regio- and stereoselective
achieved; in addition, greater than stoichiometric amounts of salt
additives were required. Additional systems that have been used
photocatalytic hydrogenolysis of allylic alcohols to form unsaturated
hydrocarbons employing a palladium(II)-loaded titanium oxide; the
reaction proceeds at room temperature under light irradiation without
stoichiometric generation of salt wastes. Olefin and saturated alcohol
moieties tolerated the reaction conditions. Hydrogen atoms were
selectively incorporated into less sterically congested carbons of the
allylic functionalities. This protocol allowed a short-step synthesis of
(S)-(+)-lavandulol from (R)-(–)-carvone by avoiding otherwise
necessary protection/deprotection steps.
for
hydrogenolysis
include:
catalytic
heteropolyacid
(H3[PW12O40]•nH2O);[11] bismuth triflate[12] with stoichiometric
Et3SiH and stoichiometric low-valent titanocene;[13] and
NaB(CN)H3/BF3•OEt2.[14] Nevertheless, stoichiometric by-
production of silyl-, titanium- or boron wastes is inevitable and
migration of double bonds is not controllable. To date, the
reactivity and selectivity of hydrogenolysis of unsymmetrical
allylic alcohols, as well as the applicable substrate range, have
been unpredictable. In addition, an unmet need persists to
minimize byproducts so that a high standard of scalable
production of chemicals can be met in next-generation industrial
applications (Scheme 1a).
Introduction
Reductive deoxygenation is one of the most important methods
for transformation of biorenewable resources into useful carbon
feedstock.[1] Indeed, catalytic deoxygenation of alcohols via
hydrodeoxygenation (HDO: hydrogenolysis of C–O σ bond) or
deoxydehydration (DODH) followed by hydrogenation (apparent
HDO), is a fundamental approach extensively studied in modern
catalysis and is used to produce hydrocarbons as alternative
biofuels.[1a,2] However, many of these processes require
strenuous (> 150 °C) reaction conditions; in addition, two or
more reaction steps are involved in the apparent HDO.
In contrast, hydrogenolysis of the C–O bonds of allylic
alcohols, namely the HDO, offers direct access to the
corresponding unsaturated hydrocarbons, which are useful
building blocks for organic synthesis.[3] A range of allylic alcohols
is available from many natural products (e.g., terpenoides)[4] and
via catalytic multistep DODH of natural polyols.[5] It is a
significant challenge to cleave the C–O σ-bond of allylic alcohols
in the presence of hydrogen donors without accompanying side
reactions of C=C π-bonds. Possible side reactions include
double bond oxygenation (oxirane formation),[6] double bond
reduction[7] and double bond isomerization.[7b,8]
(a) Earlier work: Stoichiometric salt waste under dark
Stoichiometric or large amount of
reagent/catalyst
or
R
OH
R
R
–Salt waste
Refs 9 and 10
Refs 11–14
• Direct C–OH σ bond cleavage9-14
• Catalytic cobalt,9,10 heteropolyacid,11 or bismuth12 agent
• At room temperature12
• More than stoichiometric amount of salt additives9,10
• Stoichiometric Si–H11,12 or B–H14 agent
• Stoichiometric titanium13 or boron14 agent
(b) Present work: No stoichiometric salt waste under light
Photocatalyst
hν, CH3OH
+
H2O
R
OH
R
–H2C(OCH3)2
• Direct C–OH σ bond cleavage
• Photocatalytic (Pd, TiO2)
• Methanol as hydrogen donor
• At room temperature
• Regio- and stereoselective
Partly to this end, Funabiki et al. reported the hydrogenolysis
–3
of allylic alcohols with cobalt complex HCo(CN)5 under a
Scheme 1. Hydrogenolysis of allylic alcohols.
(a) Earlier work by others. (b) Present work: Direct photocatalytic transfer
hydrogenolysis (PcTH).
hydrogen atmosphere.[9] In this reaction, the regioselectivity, SN2
vs. SN2´ reaction, was altered by changing the ratio of potassium
cyanide to cobalt chloride. The regioselectivity of this cobalt
catalysis was later improved under milder reaction conditions.[10]
In both cases, however, a perfect tolerance of ene groups and
hydrogenolysis of γ-disubstituted allyl alcohols were not
In sharp contrast, our recent focus has been on
photocatalysis at room temperature (rt) using
a
metal
nanoparticle (NP)-loaded heterogeneous materials.[15] A novel
protocol was previously demonstrated employing photocatalytic
transfer hydrogenolysis (PcTH) for reductive deoxygenation of
an allylic alcohol with palladium-loaded titanium(IV) oxide
(Pd/TiO2) in CH3OH.[16,17] The photocatalysis is highly
chemoselective (C–O σ-bond over C=C π-bond of allyl alcohol
cleaved) and redox-selective (hydrogenolysis over oxidation of
C–O of allyl alcohol). To the best of our knowledge, this was the
first example of direct hydrogenolysis of an allylic alcohol to
propylene using light energy. However, it was unclear whether
the protocol could be used for regioselective PcTH of structurally
[a]
Y. Takada, S. Kaliyamoorthy, Dr. H. Naka, Prof. Dr. S. Saito
Graduate School of Science,
Nagoya University,
Chikusa, Nagoya 464-8602, Japan,
E-mail: saito.susumu@f.mbox.nagoya-u.ac.jp
Dr. J. Caner, Dr. H. Naka, Prof. Dr. S. Saito
Research Center for Materials Science,
Nagoya University,
[b]
Chikusa, Nagoya 464-8602, Japan
Supporting information for this article is given via a link at the end of
the document.
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/chem.201704099
Chemistry - A European Journal
FULL PAPER
more intricate allylic alcohols while maintaining chemo- and
redoxselectivity. We report herein for the first time the synthetic
potential and application of PcTH to substituted allyl alcohols
(Scheme 1b).
To begin, cinnamyl alcohol (1a) was used as a model
substrate to optimize the reaction conditions (Table 1). After
prescreening photocatalysts, we identified that PdCl2(CH3CN)2-
loaded TiO2 (P25) (Pd(5 wt %)/TiO2) selectively promoted PcTH
to furnish E-2a (entry 1). Amounts of minor side products (Z-2a,
iso-2a and 3a) were less than the detection limit of a calibration
line. These conditions proved to be superior to our previous
conditions for the PcTH of an allylic alcohol[16,17] and other
Results and Discussion
Table 1. Optimization of reaction conditions of PcTH[a]
hν (λ > 365 nm)
Pd(5 wt %)/TiO2, additive
Ph
+
+
+
Ph
OH
Ph
Ph
Ph
CH3OH, Ar
1a
E-2a
Z-2a
iso-2a
3a
(2.0 mmol)
Yield[b] (%)
E-2a
Reaction
Entry
Pd(5 wt %)/TiO2 (mg)
CH3OH (mL)
T (°C)
Additive (mol %)
Conv.[b] (%)
time t (h)
Z-2a
iso-2a
3a
[c]
[c]
[c]
15
15
15
15
15
15
15
15
15
15
15
15
5
25
25
25
25
25
25
25
25
25
50
0
TsOH•H2O (2.5)
TsOH•H2O (0.5)
TsOH•H2O (0.5)
TsOH•H2O (0.5)
TsOH•H2O (0.5)
none
–
–
–
1
2
3
4
5
6
7
8
9
2
38
36
[c]
[c]
[c]
5
–
–
–
2
46
40
79
93
65
29
51
36
32
77
16
68
[c]
[c]
5
–
–
4
90
3
[c]
[c]
5
–
–
4.5
5
> 99
> 99
28
7
[c]
[c]
5
–
–
28
[c]
[c]
[c]
5
–
–
–
2
[c]
[c]
[c]
5
none
–
–
–
5
60
[c]
[c]
[c]
5
TsONa (2.5)
–
–
–
2
41
[c]
[c]
30
30
5
TsOH•H2O (0.5)
TsOH•H2O (2.5)
TsOH•H2O (0.5)
TsOH•H2O (0.5)
–
–
2
40
5
[c]
–
10
2
2
7
2
5
98
3
11
4
[c]
[c]
[e]
–
–
11
–
[c]
[c]
5
0
–
–
12
13[d]
78
99
< 1
8
[c]
[c]
[c]
15
0
5
5
25
25
TsOH•H2O (0.5)
TsOH•H2O (2.5)
–
–
–
10
[c]
[c]
[c]
[c]
–
–
–
14
–
aLight source: 300 W Xe lamp with a cold mirror (λ > 365 nm). Pd(5 wt %)/TiO2 (PdIICl2(CH3CN)2-loaded TiO2) photocatalyst was made using P25 (Degussa
Aeroxide® P25, from Sigma–Aldrich, anatase-rutile mixture, 21 nm particle size, 35–65 m2/g surface area (BET)). bDetermined by GC-MS (dodecane as an
internal standard). cNot determined because the product was barely or not detected (less than < 2%). dUnder H2 instead of Ar without light irradiation. 3-
Phenylpropan-1-ol was generated in 86% yield. eNot determined.
similar conditions (Table S1, ESI). The Pd loading on TiO2 (P25)
was varied to 1, 5 and 10 wt% (Table S1, entries 5 and 6), but a
5 wt% Pd resulted in a slightly higher yield of E-2a than the other
two possible products. Decreasing the amount of TsOH•H2O to
0.5 mol% resulted in a slight increase in yield (entry 2). Indeed,
the lower amount of TsOH•H2O gave the highest reaction rate,
and E-2a was obtained in the highest yield with t = 4.5 h (entry
4). Therefore, the catalyst system shown in entries 2–5 was
used as the optimized conditions for further studies. The PcTH
proceeded in the absence of TsOH•H2O but with a lower
reaction rate (entries 6 and 7). Using the sodium salt of
TsOH•H2O resulted in only a slight decrease in yield (entry 8 vs.
2). A more dilute CH3OH solution has a slight influence on the
yield of E-2a (entry 9). Although the PcTH at 50 °C showed a
higher reaction rate (entry 10), the yield of 2a, however, was
77% with 11% of 3a, whereas it proceeded more sluggishly at
0 °C without an improvement of selectivity (entries 11 and 12). In
control experiments, hydrogen gas (H2) was introduced to the
reaction vessel with an expectation that it functions as a
hydrogen source for reduction of both 1a and PdIICl2(CH3CN)2
optimized conditions
using CD3OD
Ph
OH
Ph
D
64% D
E-2a-d1
(E-2a-d1:1a:3a-d3 = 3:< 0.02:< 1)
optimized conditions
1a
OH
1b
using CH3OD
Ph
Ph
D
78% D
E-2a-d1
(E-2a-d1:1b:3a-d3 = 12:< 0.08:< 1)
D
D
not detected
Ph
Ph
E-2a-d1'
-
d
iso-2a-d1
1
Scheme 2. PcTH of 1a and 1b using deuterated methanol.
Conditions: 1a or 1b (2 mmol), hν (λ > 365 nm), PdCl2(CH3CN)2-loaded TiO2
(Pd(5 wt%)/TiO2) (15 mg), TsOH•H2O (0.5 mol %), CD3OD or CH3OD (5 mL),
5 h, Ar, 25 °C. The products were analysed by 1H and 13C NMR. Yield (molar
ratio) and deuteration ratio were determined by 1H NMR. Major side products
were diastereomers of 3a-dn (PhCH(D/H)CH(D/H)CH2(D/H)) (n = 0–3)
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/chem.201704099
Chemistry - A European Journal
FULL PAPER
loaded on TiO2 (P25); however, almost no hydrogenolysis of the
C–O bond tookplace even after t = 5 h (entry 13). The major
product was 3-phenylpropan-1-ol, which was generated in
favour of hydrogenation of the C=C π-bond of 1a. Light
irradiation in the absence of catalyst led to minimal conversion of
1a (entry 14).
R
Pd NP
H+
e
OH
H+
e–
–
Considering mechanistic scenario, n (n = an integer) holes
(nh+) and n electrons (ne–) are generated in TiO2 photocatalysts
upon light irradiation.[16] In a formal sense, CH3OH is oxidized,
giving 2e– to holes (2h+) on the surface of TiO2 (Figure 1).
Excited 2e– of TiO2 accumulate on the Pd NP surface and
interact with 2H+ (apparently from CH3OH) to give active
hydrogen atoms adsorbed [2H(ad)] on Pd NP. Under slightly
acidic conditions with TsOH, the OH group of allylic alcohol may
be more readily adsorbed on Pd NP covered by a higher
concentration of H+, which could also activate the OH group
CH2O
e–
–
e
h+
h+
TiO2
Figure 1. H+ and 2e– (in dotted square) would work as H–. CH3OH (small
circles in light blue) loses 2e– (which combine with 2h+) and 2H+ (which move
to a surface of Pd NP and interact with 2e– that came from light-excited sites
(shining yellow star) of TiO2 (a large circle in gray)).
+
more effectively, giving R–OH2 . Separately formed transient
+
H(ad) (H+ and 2e−) on Pd NP seems to react with R–OH2 ,
+
which facilitates attack of H(ad) toward a carbon of R–OH2 via
SN2 or SN2´ manner. Allylic alcohols and CH3OH would make a
self-assembled monolayer-type alignment on Pd NP so that the
alkyl and olefin groups could expose to the space distal from the
NP surface. Indeed, a preceding report indicates that at high
coverage of benzyl alcohol on Pd (111) surface, the alcohol
adopts an upright structure (more likely standing up
perpendicular to the surface, rather than sprawling on the
surface), where only the alcohol function is bound to the Pd
surface and the phenyl ring extends into vacuum.[18a,b] Indeed,
benzyl alcohol underwent the hydrogenolysis of the C–O bond
using Pd/TiO2 under light.[18c,d] In contrast, Shon suggested that
the hydrogenation/isomerization reactions of allyl alcohol in the
presence of H2 on alkanethiolate-capped Pd NP surface likely
occur through a mono-σ-bonded Pd-alkyl intermediate followed
by β-hydride elimination. When there is no H2 present, a π-allyl
Pd hydride mechanism is also suggested.[18e,f] Since
hydrogenation and isomerization of, at least, the trisubstituted
olefins of allylic alcohols E- and Z-1g (vide infra) are slower than
the hydrogenolysis of their C–O bonds in the present PcTH, the
mechanism shown in Figure 1 would be more favourable.
Deuterated methanol (CD3OD) was used instead of CH3OH
under the optimized conditions to get further insight into the
characteristics of the PcTH reaction (Scheme 2). As a result, E-
2a-d1 was obtained from 1a as the major product (E-2a-d1 (64%
1
D):1a:3a-d3 = 3:< 0.02:< 1, t = 5 h). H and 13C NMR analysis
indicated that over-reduced product 3a was tri-deuterated. The
yield of fully saturated hydrocarbons 3a-dn (n = 1–3) was
comparable to that obtained previously (Table 1, entry 5).
Table 2. Substrate scope of PcTH[a]
Yield[b](%)
Reaction
Conv.[b] E + Z
iso
time t
Entry
Substrate
Products
(h)
(ratio)[c]
OH
Ph
61
+
+
[e]
1[d]
2
6
69
99
–
7
Ph
Ph
(61:–[e]
)
Ph
E-2c
Z-2c
iso-2c
1c
Ph
OH
Ph
Ph
+
2
90
1d
2d
iso-2d
OH
67[f]
3
4
5
3
5
E-2a + Z-2a + iso-2a
86[f]
4[f]
(64:3)[f]
Ph
1b
n-C6H13
Z-2e
n-C6H13
E-2e
n-C6H13
85
(71:14)
n-C6H13
n-C6H13
OH
+
+
[e]
> 99
> 99
–
E-1e
iso-2e
82
(43:39)
[e]
5.5
–
E-2e + Z-2e + iso-2e
OH
Z-1e
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/chem.201704099
Chemistry - A European Journal
FULL PAPER
OH
72
(57:15)
[e]
6[g]
5
6
E-2e + Z-2e + iso-2e
> 99
> 99
–
n-C6H13
1f
OH
76
+
+
7
18
(73:3)[h]
E-1g
E-2g
Z-2g
iso-2g
OH
71
8
6
E-2g + Z-2g + iso-2g
95
15
(5:66)[h]
Z-1g
76
9
6
E-2g + Z-2g + iso-2g
> 99
> 99
98f
17
16
HO
1h
(51:25)
[h]
OH
81
+
+
10
11k
5.5
10
(–i)
(83)j
2
2
2
2
1i
iso-2i
E-2i
Z-2i
80f
(2j)
+
isomersl
OH
2j
1j
+ isomers of 2j
OH
[e]
12
6
< 10
–
or
iso-2k
1k
2k
aConditions: allylic alcohol (2 mmol), CH3OH (5 mL), Pd(5 wt%)/TiO2 (PdIICl2(CH3CN)2-loaded TiO2) (15 mg), TsOH•H2O (0.5 mol%), 300 W Xe lamp with a
cold mirror (l > 365 nm), Ar atmosphere (1 atm), 25 °C unless otherwise specified. bDetermined by 1H NMR analysis (1,1,2,2-tetrachloroethane or benzene
as internal standard). cThe number in parentheses is molar ratio of the E/Z product determined by 1H NMR analysis unless otherwise noted. dTsONa (2.5
mol %) was used instead of TsOH•H2O. Negligible amount of 4-phenylbutan-2-ol and (3-methoxybutyl)benzene were detected. eNot determined because the
f
product was barely or not detected. Conversion and yield were determined by GC-MS (dodecane as an internal standard). gNonan-3-one and nonan-3-ol
were also observed in 17 and 14% yield, respectively. hArea ratio of E-2g and Z-2g analysed by GC-MS. iZ-isomer was not detected on 13C NMR analysis.
jOf isolated as a mixture of isomers. k2.5 mol % of TsOH•H2O was used. lThe structures of isomers were not confirmed. Area ratio of 2j:isomers was 40:3 in
GC-MS analysis.
Given the absence of deuteration at the β and γ position of E-
2a and based on the mechanistic scenario proposed in Figure 1,
the PcTH of 1a proceeded via an apparent SN2 pathway giving
E-2a-d1, rather than SN2´ followed by migration of the double
bond.[19,20] In contrast, 1b was converted solely to E-2a-d1 using
CH3OD. Since it is highly unlikely that a π-allyl Pd intermediate
was majorly involved in the catalytic cycle, 1b reacted with an
active hydrogen atom (H(ad)) through an apparent SN2´
pathway in preference to an SN2 pathway. These results
suggest that, irrespective of the position of the OH group of 1a
and 1b, H(ad) (hydride or H+ + 2e–) favourably attacked less
sterically congested carbons; in other words, the terminal
positions of the main alkyl chain of 1a and 1b were consistently
targeted. Such a predictable regioselectivity in hydrogenolysis
of allylic alcohols is unprecedented.[11]
With the optimized PcTH conditions in hand, the substrate
scope was investigated (Table 2). 2° alcohol 1c underwent
hydrogenolysis faster than 1a (entry 1). In the PcTH of 1d, 2d
was produced through the SN2 pathway predominately (entry 2).
As mentioned in the deuterium-labelling experiment (Scheme 2),
PcTH of 1b gave E-2a rather than iso-2a (entry 3). Aliphatic
allylic alcohols in three different isomeric forms were tested
(entries 4–6). E-1e was converted into the corresponding E-2e
in a better yield than Z-1e, and the Z-isomer Z-2e was also
detected. The selectivity of Z- vs. E-2e was low in the case of
Z-1e, probably due to the isomerization of olefins occurring
during the PcTH.[19,20] E-2e was generated as the major product
from 2° alcohol 1f through an apparent SN2´-enriched
hydrogenolysis (entry 6). Geraniol (E-1g) and nerol (Z-1g) were
converted stereospecifically to the E-2g and Z-2g, respectively
(entries 7 and 8). Both reactions were more contaminated with
SN2´ reaction products than when E-1e and Z-1e were used,
providing iso-2g bearing a terminal olefin in a small amount
(~15%). The isomeric 3° alcohol 1h underwent an apparent
SN2´ reaction, giving a mixture of E-2g and Z-2g (3:1) as the
two major products (entry 9). This is a stark contrast to the
cobalt catalysis, in which hydrogenolysis of sterically congested
E-1g and 1h proceeded with minimal conversion.[10] Farnesol
(1i) was chosen because of the higher boiling points of the
corresponding hydrocarbon products, which facilitate easy
separation and isolation by column chromatography (entry 10).
The product distribution of isomeric structures was comparable
to that obtained with E-1g. E-2i was obtained in an isolated
yield of 83%, together with minor isomers Z-2i and iso-2i. PcTH
of (S)-perillyl alcohol (1j) was completed by prolonging t, and an
important fragrance, (S)-limonene (2j), was obtained in 80%
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/chem.201704099
Chemistry - A European Journal
FULL PAPER
carbon–carbon triple bond first took place to give Z-1a as major
product (58% after 5 h). The reaction that retards subsequent
C=C bond reduction reminds us of the Lindlar’s catalyst and
recent Pd variants bearing similar functions at a high pressure
of H2;[21] unlike the Lindlar’s catalyst, however, in this system
under light, poisoning of Pd surface (with Pb(OAc)2, CaCO3,
and/or organic substrates such as quinoline) is not required.
Subsequent PcTH of the allylic C–O bond of Z-1a (and 1a)
similarly gave a mixture of E- and Z-2a (75% after 12.5 h).
Because of possible isomerization of the olefin groups under
near UV-Vis light,[19,20] E/Z selectivity was moderate upon a
longer reaction time (up to E/Z = 31:44, see ESI).
hν (λ > 365 nm)
Pd(5 wt %)/TiO2 (15 mg)
TsOH•H2O (0.5 mol %)
+
Ph
OH
impurity
Ph
CH3OH (5 mL)
Ar, 25 °C, 2 h
1a
E-2a
25–38%
(c.f. 40% without impurity)
(2.0 mmol)
(2.0 mmol)
impurity (recovery of impurities: > 90% in all cases)
PhOH PhOCH3 PhCl PhCOOCH3
PhCONH2
PhCN
Scheme 3. Functional group compatibility test.
Yield and conversion obtained after PcTH were determined by GC-MS
(dodecane as an internal standard).
To expand the synthetic application of PcTH, direct
derivation of a set of olefinic products was performed on the
crude mixture obtained after PcTH (Table 3). After removal of
the photocatalyst by filtration, the filtrate was washed with H2O
and extracted with CHCl3 or hexane. The CHCl3 solution of a
crude mixture of E-2a, Z-2a and iso-2a were subjected to
further reactions (See ESI for experimental details). For
instance, epoxidation of a mixture of ene products by m-
chloroperbenzoic acid (MCPBA) furnished trans-4a in 95% yield
(entry 1).[22] A minimal amount of cis-2-methyl-3-phenyloxirane
and 2-benzyloxirane were also detected by 1H NMR. The
epoxidation took place stereospecifically. Likewise, epoxidation
of crude olefinic compounds obtained by PcTH of 1d and E-1e
yielded epoxides 4d and trans-4e, respectively, which were
obtained as major products (entries 2 and 3). When crude ene
products obtained by PcTH of 1a were treated with Br2 instead
of MCPBA, an 83% yield of trans-5a was obtained (entry 4).
By taking advantage of the double bond compatibility under
the PcTH reactions, the synthetic shortcut to (S)-(+)-lavandulol
(S-7) from precursor diol 6 was demonstrated. Lavandulol is in
an important class of monoterpenes mainly found in lavandula
essential oils; it plays a pivotal role as a sex pheromone in
some mealybugs.[23] In industrial applications, lavandulol is an
important additive used to control the odour balance of
perfumes and flavors.[24] Several racemic and stereoselective
syntheses of lavandulol have been reported because of its
importance.[24–26] These methods, however, require a variety of
starting materials or suffer from low regioselectivity and tedious
multistep reactions involving protection/deprotection steps.
Recently, Koo et al. succeeded in a stereoselective eight-step
synthesis of S-7 from (R)-(–)-carvone,[26] which is frequently
utilized as a chiral building block in asymmetric synthesis.[27] In
Koo’s synthesis, a protective activation of the key intermediate
Table 3. Derivation of ene products after PcTH of allylic alcoholsa
Entry
1
1
1a
Product, yieldb
O
H
Ph
H
trans-4a
95%
2
1d
O
O
Ph
H
Ph
4d
iso-4d
81%
8%
3c
E-1e
O
O
O
H
H
H
n-C6H13
n-C6H13
n-C6H13
H
cis-4e
iso-4e
trans-4e
2%
10%
56%
4
1a
Br
Br
H
H
H
Br
Br
H
Ph
Ph
trans-5a
cis-5a
6%
83%
aPcTH conditions: allylic alcohol (2 mmol), CH3OH (5 mL), PdCl2(CH3CN)2-
loaded TiO2 (Pd(5 wt%)/TiO2) (15 mg), TsOH•H2O (0.5 mol%), 300 W Xe
lamp with a cold mirror (λ > 365 nm), Ar atmosphere (1 atm), 25 °C, 4–6 h.
Epoxidation (or dibromination) procedures: After filtration of the reaction
mixture obtained by PcTH, the filtrate was washed with H2O and extracted
with CHCl3 (or hexane), and the mixture was treated with MCPBA (2.9
mmol) at 0 °C for 3 h (or with Br2 (20 mmol) at 0 °C for 12 h). See ESI for
experimental details. The structures in the table show relative configuration.
bDetermined by 1H NMR (1,1,2,2-tetrachloroethane as an internal standard).
c(3-nonyloxiran-2-yl)methanol generated from starting allylic alcohol was
also obtained in 28% yield.
yield (entry 11). Chiral GC-MS analysis showed that the chiral
center of 1j was retained after the PcTH. The C=C bonds distal
from the reacting carbon also tolerated the PcTH conditions
(entries 7–11). This result is markedly different from
conventional hydrogenation of C=C π-bonds over a Pd/solid
support under H2 in the dark. Sterically congested 1k was not
converted to neither 2k nor iso-2k (entry 12).
6
is a critical step. However, the synthesis mandates
stoichiometric generation of multiple (metallic) wastes (Scheme
4).[26]
Two different hydroxyl groups, allylic and homoallylic
alcohols, are incorporated in 6. Our expectation was that PcTH
would selectively deoxygenate an allylic alcohol over
a
Whether other functional groups could tolerate the PcTH
conditions was also investigated (Scheme 3). As a result,
phenol, ether, chloroaryl, ester, amide, and nitrile groups of
non-allylic alcohols essentially unreacted at least in a short light
irradiation time (2 h).
To further probe functional group compatibility, a carbon–
carbon triple bond was also tested under the PcTH conditions
(Table S2, ESI). A propargylic alcohol (3-phenylprop-2-yn-1-ol)
was used instead of allylic alcohols. Semi-hydrogenation of
homoallylic alcohol. After diol 6 was prepared according to
reported procedures,[26] it was subjected to PcTH under the
optimized conditions except that a larger amount of TsOH (2.5
mol %) was used (Scheme 4). To our delight, S-7 was obtained
in 56% yield (isolated yield: 42%) through selective PcTH of the
allylic alcohol. A minor side product, an isomer of S-7 (5-
methyl-2-(prop-1-en-2-yl)hex-5-en-1-ol), was also detected by
1H and 13C NMR analysis (16% yield). The chirality of S-7 was
retained from racemization, ascertained by chiral GC analysis.
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/chem.201704099
Chemistry - A European Journal
FULL PAPER
Pd(5 wt%)/TiO2 (PdCl2(CH3CN)2-loaded TiO2): Pd(5 wt%)/TiO2
(PdCl2(CH3CN)2-loaded TiO2) catalyst was prepared in a 1000 mL
round-bottom flask covered with aluminum foil by impregnation of 6.06 g
of TiO2 (Degussa Aeroxide® P25) with a solution of Pd(CH3CN)2Cl2
It should be noted that the protection-deprotection strategy
critical for the previous total synthesis of S-7[26] was bypassed
with PcTH, where both the homoallylic C=C bond and alcohol
moieties effectively tolerated the PcTH conditions.
(0.319 g, prepared by refluxing PdCl2 in CH3CN) in
a deionized
H2O/THF (5:1, 300 mL). The slurry was sonicated for 5 min and stirred
using rotary evaporator under atmospheric pressure for 1 h at 50 °C.
Then, solvent was removed under reduced pressure. The remaining
solid was ground in a mortar and dried at 50 °C under oil pump vacuum
for 7 h. The final amount of the solid (PdCl2(CH3CN)2-loaded TiO2) was
5.98 g.
PcTH of 6 to (S)-(+)-lavandulol (S-7)
hν (λ > 365 nm)
Pd(5 wt%)/TiO2 (15 mg)
HO
TsOH•H2O (2.5 mol %)
OH
OH
CH3OH (5 mL),
36 h, Ar, 25 °C
6 (2 mmol)
S-7
56% (42%)†
PcTH of 1a employing Pd(5 wt%)/TiO2 (Table 1, entry 3 (t = 4 h))
Conv. 88%
Cf. Previous synthetic route (ref. 26)
A cylindrical Pyrex glass vessel was charged with a cross-shaped
magnetic stir bar, Pd(5 wt%)/TiO2 (15 mg), cinnamyl alcohol (1a, 268
mg, 2.00 mmol), p-toluenesulfonic acid monohydrate (TsOH•H2O, 1.9
mg, 0.01 mmol), n-dodecane (as internal standard, 50 µL) and
anhydrous CH3OH (5 mL). After sonication of the mixture (to fully
disperse the photo-precatalyst), the reaction vessel was sealed with a
rubber septum and the stirred suspension was immediately deaerated
with a stream of Ar. After the Ar stream was stopped, the mixture of a
closed system was irradiated with near-UV–Vis light (λ > 365 nm). After
reaction time (t) of 4 h, the reaction mixture was sampled by a syringe
and analyzed by GC-MS to determine the amount of 1a, trans-b-methyl
styrene (E-2a), cis-b-methyl styrene (Z-2a), 3-phenyl-1-propene (iso-2a)
and propylbenzene (3a). The yields of 1a (0.20 mmol, 90% conversion),
E-2a (1.57 mmol, 79% yield) and iso-2a (0.052 mmol, 3% yield) were
determined based on the corresponding calibration curves made with
authentic samples and n-dodecane as an internal standard. The yields
of Z-2a and 3a were not determined because the amounts detected by
GC were smaller than detection limit of the calibration curves (less than
2%). Minimal amounts of 3-phenylpropan-1-ol and benzaldehyde
dimethyl acetal were also detected by GC-MS.
1. LiBHEt3, Pd catalyst
2. deprotection
EEO
6
OEE
S-7
68%
83%
EE = (CH2)2OC2H5
Scheme 4. Synthesis of (S)-(+)-lavandulol (S-7) from intermediate 6.
Yield and conversion obtained after PcTH were determined by 1H NMR
(1,1,2,2-tetrachloroethane as an internal standard). †Of isolated, purified
product.
Conclusions
In conclusion, we have developed a selective photocatalytic
transfer hydrogenolysis (PcTH) of allylic alcohols to furnish
unsaturated hydrocarbons using palladium(II)-loaded TiO2(P25)
in CH3OH under near-UV–Vis light at room temperature. The
precatalyst is simply prepared by impregnating TiO2 with a
palladium(II) source; PcTH can immediately be started using
the precatalyst without its calcination. The only reagent used
was CH3OH,[28] which is a cheap and sustainable solvent,
(S)-(+)-Lavandulol (S-7) synthesis through PcTH of lavandulol
precursor 6
A cylindrical Pyrex glass vessel was charged with a cross-shaped
magnetic stir bar, PdCl2(CH3CN)2-loaded TiO2 (15.0 mg), (S,E)-2-
methyl-5-(prop-1-en-2-yl)hex-2-ene-1,6-diol (6, 341 mg, 2.00 mmol), p-
toluenesulfonic acid monohydrate (TsOH•H2O, 9.8 mg, 0.05 mmol), n-
dodecane (50 µL) and dehydrated CH3OH (5 mL). After sonication of
the mixture (to fully disperse the photocatalyst), the reaction vessel was
hydrogen/electron
donor,
and
hole
scavenger
in
photocatalysis.[29] Owing to this efficient and green nature, salt
wastes were not formed. In addition, simple removal of catalyst
and solvent facilitated further derivation of the PcTH products to
synthetically useful, functionalized compounds. It is now
predictable that both the SN2 or SN2´ reactions take place so
that hydrogen atoms of CH3OH and/or allylic alcohols are
incorporated more favourably into the less sterically hindered
carbon of starting materials. Many functional groups including
C=C bonds, phenol, ether, chloroaryl, ester, amide, and nitrile
groups of non-allylic alcohols well tolerate PcTH conditions by
setting an appropriate light irradiation time. Application of PcTH
to the rapid synthesis of (S)-(+)-lavandulol was also
successfully achieved without protection/deprotection steps.
We envisage that PcTH will further provide selective and direct
routes to useful (chiral) platform- and fine chemicals,
capitalizing on the renewability of biomass resources and the
sustainability of light energy.
sealed with
a rubber septum and the stirred suspension was
immediately deaerated with a stream of Ar. After Ar stream was stopped,
the mixture of closed system was irradiated with near-UV–Vis light (λ >
365 nm). After t of 36 h, light irradiation was stopped. The mixture was
analyzed directly by 1H NMR and 13C NMR to determine the yield based
on integral ratios between the products and the internal standard
(1,1,2,2-tetrachloroethane). (S)-(+)-lavandulol (S-7) and its stereoisomer
(5-methyl-2-(prop-1-en-2-yl)hex-5-en-1-ol (8, structure of which was
determined based on 1H NMR and 13C NMR) were obtained in 56%
(1.12 mmol) and 16% (0.32 mmol) yield, respectively, and 12% (0.24
mmol) of starting material 6 was recovered. The mixture of PcTH was
filtered through a Celite pad and the residue remained on the pad was
washed with CH2Cl2. The filtrate was added to H2O (15 mL) in a
separatory funnel and the organic phase was extracted with CH2Cl2 (10
mL × 3). The combined organic phase was washed with aqueous NaCl
and dried over Na2SO4. After filtration, the filtrate was evaporated in
vacuo and the residue was purified by column chromatography on silica
gel (hexane/ethyl acetate (EtOAc) = 25:1 to 10:1) to give (S)-(+)-
lavandulol (S-7) as colorless oil with an isolated yield of 42% (129 mg,
0.82 mmol).
Experimental Section
Catalyst preparation
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/chem.201704099
Chemistry - A European Journal
FULL PAPER
Supplemental results, experimental procedures and spectra data of
compounds were available in ESI.
[11] M. Egi, T. Kawai, M. Umemura, S. Akai, J. Org. Chem. 2012, 77,
7092–7097.
[12] G. G. K. S. N. Kumar, K. K. Laali, Org. Biomol. Chem. 2012, 10,
7347–7355.
[13] a) H. Ledon, I. Tkatchenko, D. Young, Tetrahedron Lett. 1979, 20,
173–176. b) F. Sato, Y. Tomuro, H. Ishikawa, T. Oikawa, M. Sato,
Chem. Lett. 1980, 103–106. c) Y. Xiao, L. Chu, Y. Sanakis, P. Liu, J.
Am. Chem. Soc. 2009, 131, 9931–9933. d) H. R. Diéguez, A. López,
V. Domingo, J. F. Arteaga, J. A. Dobado, M. M. Herrador, J. F. Quílez
del Moral, A. F. Barrero, J. Am. Chem. Soc. 2010, 132, 254–259.
[14] A. Srikrishna, R. Viswajanani, J. A. Sattigeri, C. V. Yelamaggad,
Tetrahedron Lett. 1995, 36, 2347–2350.
Acknowledgements
The authors wish to thank Prof. R. Noyori (NU) and Prof. Kudo
(Tokyo University of Science) for fruitful discussions. We are
also grateful to T. Noda, H. Natsume, and H. Okamoto (NU) for
their support in the production of reaction vessels. This
research was partially supported by Advanced Catalytic
Transformation program for Carbon utilization (ACT-C), JST (to
S.S.), the Asashi Glass Foundation (to S.S.), Grant-in-Aid for
Scientific Research C (General, 26410115, to H.N.) from MEXT
and by funds from Tobe Maki Scholarship Foundation (to H.N.).
Y.T. appreciates the Japan Society for the Promotion of
Science (JSPS) for a Research Fellowship for young scientists.
[15] a) Z. Liu, J. Caner, A. Kudo, H. Naka, S. Saito, Chem. - A Eur. J. 2013,
19, 9452–9456. b) V. N. Tsarev, Y. Morioka, J. Caner, Q. Wang, R.
Ushimaru, A. Kudo, H. Naka, S. Saito, Org. Lett. 2015, 17, 2530–2533.
c) M. Shibata, R. Nagata, S. Saito, H. Naka, Chem. Lett. 2017, 46,
580–582.
[16] J. Caner, Z. Liu, Y. Takada, A. Kudo, H. Naka, S. Saito, Catal. Sci.
Technol. 2014, 4, 4093–4098.
[17] Pd/TiO2 was prepared using PdCl2 and characterised by ICP-OES,
HRTEM and XPS analysis (ref. 16).
Keywords: hydrogenolysis • photocatalysis • allylic alcohols •
[18] a) S. H. Pang, A. M. Román, J. W. Medlin, J. Phys. Chem. C 2012,
116, 13654–13660. b) C.-H. Lien, J. W. Medlin, J. Phys. Chem. C
2014, 118, 23783–23789. c) T. P. Ruberu, N. C. Nelson, I. I. Slowig, J.
Vela, J. Phys. Chem. Lett. 2012, 3, 2798–2802. d) Y. Shiraishi, K.
Fujiwara, Y. Sugano, S. Ichikawa, T. Hirai, ACS Catal. 2013, 3, 312–
320. e) E. Sadeghmoghaddam, H. Gu, Y.-S. Shon, ACS Catal. 2012,
2, 1838–1845. f) D. J. Gavia, M. S. Maung, Y.-S. Shon, ACS Appl.
Mater. Interfaces 2013, 5, 12432–12440.
titanium oxide • palladium
[1]
[2]
Reviews: a) A. M. Ruppert, K. Weinberg, R. Palkovits, Angew. Chem.
2012, 124, 2614–2654; Angew. Chem. Int. Ed. 2012, 51, 2564–2601.
b) J. M. Herrmann, B. König, European J. Org. Chem. 2013, 7017–
7027.
Reviews: a) E. Furimsky, Appl. Catal. A Gen., 2000, 199, 147–190. b)
A. Zacher, M. Olarte, D. Santosa, Green Chem., 2014, 16, 491–515;
c) S. De, B. Saha, R. Luque, Bioresour. Technol., 2015, 178, 108–118.
M. M. Smith. March's Advanced Organic Chemistry 7th ed., John
Wiley & Sons, 2013.
[19] a) K. Hashimoto, Y. Masuda, H. Kominami, ACS Catal. 2013, 3, 1349–
1355. b) Y. Li, C. Ma, H. Yang, Z. Zhang, X. Zhang, N. Qiao, J. Wang,
Z. Hao, Chem. Eng. J. 2016, 299, 1–7. c) A. Elhage, A. E. Lanterna, J.
C. Scaiano, ACS Catal. 2017, 7, 250–255.
[3]
[4]
[5]
[20] At t = 4.5 h, yield of Z-2e (46%) is significantly larger than that of E-2e
(11%), indicating that double bond isomerization of Z-2e gave E-2e.
[21] J. Hori, K. Murata, T. Sugai, H. Shinohara, R. Noyori, N. Arai, N.
Kurono, T. Ohkuma, Adv. Synth. Catal. 2009, 351, 3143–3149.
[22] In table 1, the highest yield of E-2a was obtained at light irradiation
time of 4.5 h. However, during epoxidation experiment, t = 4 h was
better due to a slight difference of the light intensity.
A. A. Newman, Ed. Chemistry of Terpenes and Terpenoids, Academic
Press, 1972.
a) M. Shiramizu, F. D. Toste, Angew. Chem. 2012, 124, 8206–8210;
Angew. Chem. Int. Ed. 2012, 51, 8082–8086. b) A. Konaka, T. Tago, T.
Yoshikawa, H. Shitara, Y. Nakasaka, T. Masuda, Ind. Eng. Chem. Res.
2013, 52, 15509–15515. c) V. Canale, L. Tonucci, M. Bressan, N.
D’Alessandro, Catal. Sci. Technol. 2014, 4, 3697–3704. d) A. Konaka,
T. Tago, T. Yoshikawa, A. Nakamura, T. Masuda, Appl. Catal. B
Environ. 2014, 146, 267–273. e) G. Sánchez, J. Friggieri, A. A.
Adesina, B. Z. Dlugogorski, E. M. Kennedy, M. Stockenhuber, Catal.
Sci. Technol. 2014, 4, 3090. f) C. Boucher-Jacobs, K. M. Nicholas,
Organometallics 2015, 34, 1985–1990. g) S. Raju, M.-E. Moret, R. J.
M. Klein Gebbink, ACS Catal. 2015, 5, 281–300. h) G. Sánchez, B. Z.
Dlugogorski, E. M. Kennedy, M. Stockenhuber, Appl. Catal. A Gen.
2016, 509, 130–142.
[23] a) J. Hadian, A. Sonboli, S. Nejad Ebrahimi, M. H. Mirjalili, Chem. Nat.
Compd. 2006, 42, 175–177. b) H. Y. Ho, C. C. Hung, T. H. Chuang, W.
L. Wang, J. Chem. Ecol. 2007, 33, 1986–1996.
[24] H. Sakauchi, H. Kiyota, S. Takigawa, T. Oritani, S. Kuwahara, Chem.
Biodivers. 2005, 2, 1183–1186.
[25] a) Y. Ueno, S. Aoki, M. Okawara, J. Chem. Soc. Chem. Commun.
1980, 683–684. b) G. Cardillo, A. D’Amico, M. Orena, S. Sandri, J.
Org. Chem. 1988, 53, 2354–2356. c) S. Araki, S. Kambe, K. Kameda,
T. Hirashita, Synthesis (Stuttg). 2003, 751–754. d) G. Mehta, S.
Karmakar, S. K. Chattopadhyay, Tetrahedron 2004, 60, 5013–5017. e)
R. A. Fernandes, A. K. Chowdhury, Eur. J. Org. Chem. 2013, 2013,
5165–5170.
[6]
[7]
a) A. Wróblewska, E. Makuch, React. Kinet. Mech. Catal. 2012, 105,
451–468. b) L. Harvey, E. Kennedy, B. Z. Dlugogorski, M.
Stockenhuber, Appl. Catal. A Gen. 2015, 489, 241–246.
a) E. Sadeghmoghaddam, H. Gu, Y. S. Shon, ACS Catal. 2012, 2,
1838–1845. b) M. Moreno, L. N. Kissell, J. B. Jasinski, F. P. Zamborini,
ACS Catal. 2012, 2, 2602–2613. c) R. Bhandari, D. B. Pacardo, N. M.
Bedford, R. R. Naik, M. R. Knecht, J. Phys. Chem. C 2013, 117,
18053–18062.
[26] H. J. Kim, L. Su, H. Jung, S. Koo, Org. Lett. 2011, 13, 2682–2685.
[27] a) A. V. Malkov, M. Bella, I. G. Stará, P. Kočovský, Tetrahedron Lett.
2001, 42, 3045–3048. b) F. Z. Macaev, A. V. Malkov, Tetrahedron
2006, 62, 9–29. c) J. Kulhánek, F. Bureš, P. Šimon, W. Bernd
Schweizer, Tetrahedron: Asymmetry 2008, 19, 2462–2469.
[28] G. A Olah, Angew. Chem. 2005, 117, 2692–2696; Angew. Chem. Int.
Ed. 2005, 44, 2636–2639.
[8]
[9]
a) P. S. Fordred, D. G. Niyadurupola, R. Wisedale, S. Bull, Adv. Synth.
Catal. 2009, 351, 2310–2314. b) K. Voronova, M. Purgel, A. Udvardy,
A. C. Bényei, Á. Kathó, F. Joó, Organometallics 2013, 32, 4391–4401.
c) D. Zsolnai, P. Mayer, K. Szőri, G. London, Catal. Sci. Technol. 2016,
6, 3814–3820.
[29] a) M. Shen, M. A. Henderson, J. Phys. Chem. Lett. 2011, 2, 2707–
2710. b) Z. Wang, Q. Hao, X. Mao, C. Zhou, D. Dai, X. Yang, Phys.
Chem. Chem. Phys. 2016, 18, 10224–10231.
T. Funabiki, Y. Yamazaki, K. Tarama, J. Chem. Soc. Chem. Commun.
1978, 63.
[10] J.-T. Lee, H. Alper, Tetrahedron Lett. 1990, 31, 4101–4104.
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/chem.201704099
Chemistry - A European Journal
FULL PAPER
FULL PAPER
Yuki Takada, Joaquim Caner, Selvam
Kaliyamoorthy, Hiroshi Naka, Susumu
Saito*
hν (λ > 365 nm)
Pd/TiO2 (15 mg)
•Regio- and stereoselective
direct C–OH σ bond cleavage
without stoichiometric salt waste
HO
TsOH•H2O (2.5 mol %)
OH
OH
CH3OH, 25 °C
Page No. – Page No.
Photocatalytic transfer hydrogenolysis of
allylic alcohols on Pd/TiO2: A Shortcut to
(S)-(+)-lavandulol
We report herein a regio- and stereoselective photocatalytic hydrogenolysis of
allylic alcohols to form unsaturated hydrocarbons employing a palladium(II)-loaded
titanium oxide; the reaction proceeds at room temperature under light irradiation
without stoichiometric generation of salt wastes. This protocol allowed a short-step
synthesis of (S)-(+)-lavandulol from (R)-(–)-carvone by avoiding otherwise
necessary protection/deprotection steps.
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.