The enhanced asymmetric hydrogenation of unsymmetrical benzils to hydrobenzoin catalyzed by organosoluble zirconium phosphonate-immobilized ruthenium catalyst

Article abstract of DOI:10.1016/j.apcata.2014.03.039

In this article, a successful design on translating a heterogeneous catalysis of chiral zirconium phosphonate-supported ruthenium catalyst into a homogeneous system was developed by the covalent attachment of chiral (R,R)-1,2-diphenyl-ethylenediamine [(R,R)-DPEN] into the backbone of zirconium phosphonate with the different arm lengths (n = 2, 4, 6) and immobilization of [RuCl₂(p-cymene)]₂. Their catalytic performances in the homogeneous asymmetric transfer hydrogenation of unsymmetrical benzils with m- and p-substituents were enhanced, and the excellent activities (>97.1% conv.), diastereomeric and enantiomeric purities (syn/anti = 11.0-29.4, 91.8-99.2%ee syn) were achieved. These homogeneous supported Ru catalysts could be quantitatively and readily recovered by addition of ethyl acetate and centrifugation by solid/liquid separation, and be reused for five times without significant loss of their catalytic performances with 97.0% conv.; 95.6% syn and syn/anti = 20.7.

Full text of DOI:10.1016/j.apcata.2014.03.039

Applied Catalysis A: General 479 (2014) 49–58
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
The enhanced asymmetric hydrogenation of unsymmetrical benzils to
hydrobenzoin catalyzed by organosoluble zirconium
phosphonate-immobilized ruthenium catalyst
Yu Du, Dandan Feng, Jingwei Wan, Xuebing Ma∗
College of Chemistry and Chemical Engineering, Southwest University, Chongqing 400715, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
In this article,
a successful design on translating a heterogeneous catalysis of chiral zirconium
Received 18 December 2013
Received in revised form 24 March 2014
Accepted 29 March 2014
phosphonate-supported ruthenium catalyst into a homogeneous system was developed by the covalent
attachment of chiral (R,R)-1,2-diphenyl-ethylenediamine [(R,R)-DPEN] into the backbone of zirconium
phosphonate with the different arm lengths (n = 2, 4, 6) and immobilization of [RuCl2(p-cymene)]2. Their
catalytic performances in the homogeneous asymmetric transfer hydrogenation of unsymmetrical ben-
zils with m- and p-substituents were enhanced, and the excellent activities (>97.1% conv.), diastereomeric
and enantiomeric purities (syn/anti = 11.0–29.4, 91.8–99.2%ee syn) were achieved. These homogeneous
supported Ru catalysts could be quantitatively and readily recovered by addition of ethyl acetate and
centrifugation by solid/liquid separation, and be reused for five times without significant loss of their
catalytic performances with 97.0% conv., 95.6% syn and syn/anti = 20.7.
Available online 12 April 2014
Keywords:
Zirconium phosphonate
Homogeneous catalysis
Asymmetric hydrogenation
Unsymmetrical benzil
Ruthenium
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
and challenging catalyst possessing chemical and physical stability,
homogeneous-like behavior and porous structure are highly desir-
Asymmetric metal catalysis is one of three modern cataly-
ses (metal catalysis, biocatalysis and organocatalysis) originating
thousands of published works every year [1–4]. However, the appli-
cation of such methodology in homogeneous catalytic system is
rather limited in chemical industry due to the high cost of chiral
ligand and noble metal, although the stereochemical and chemical
efficiencies of a certain transformation in homogeneous catalysis
are in principle better reproduced and predicted than that in het-
erogeneous catalysis [5]. Therefore, the choice of heterogeneous
supports for asymmetric metal catalysis is proved to be a crucial
and successful decision [6–8]. In heterogeneous catalysis, the cat-
alyst can be easily recovered using simple filtration or extraction
techniques that are impossible to apply in homogeneous catalysis
[9–13]. Generally, in heterogeneous catalysis, molecular diffusion
inside a matrix is a sum of two contributions: diffusion of molecules
inside the solvent, which is also present in homogeneous catalysis,
and diffusion of molecules inside supporting matrix, which resulted
in inferior catalytic performance [5]. To combine the advantages of
both homogeneous catalysis and heterogeneous catalysis, an ideal
able [14–16]. Unfortunately, inorganic supports such as MCM-41,
MCM-48 and MSNs, which can provide stable porous structure, are
generally insoluble in organic mediums [17–19].
This study initiated a novel concept of transforming inor-
ganic material-supported metal complex from heterogeneous into
homogeneous catalytic system, although soluble organic poly-
mers used as catalyst supports had been developed to meet the
need of homogeneous catalysis [20]. After the first layered zir-
conium phosphate Zr(HPO4)2·H2O (␣-ZrP) [21] and the second
Zr(PO4)(H2PO4)·H2O (␥-ZrP) [22] were prepared and elucidated for
the first time by A. Clearfield, their inorganic–organic functional
derivatives had been extensively studied owing to their high struc-
tural flexibilities and applications in many fields such as catalysis
[23–27]. Especially, by the right choice of organic pendant groups
bonded to the backbone of zirconium phosphate, the organosoluble
and filiform zirconium phosphonates with layered and meso-
porous backbone could be prepared by the co-precipitation of
functional phosphonic acid with ZrOCl2·8H2O [28]. In this paper,
the novel organosoluble chiral zirconium phosphonate and their
supported Ru catalyst were developed by the artful choice of
chiral (R,R)-1,2-diphenyl-ethylenediamine [(R,R)-DPEN] bonded
to the backbone of organosoluble zirconium phosphonate and
then the immobilization of [RuCl2(p-cymene)]2 (Fig. 1). Fortun-
ately, in the homogeneous asymmetric transfer hydrogenation of
∗
Corresponding author. Tel.: +86 23 68252360/+86 23 68253237;
fax: +86 23 68264000/+86 23 68254000.
E-mail address: zcj123@swu.edu.cn (X. Ma).
http://dx.doi.org/10.1016/j.apcata.2014.03.039
0926-860X/© 2014 Elsevier B.V. All rights reserved.

50
Y. Du et al. / Applied Catalysis A: General 479 (2014) 49–58
Fig. 1. The synthetic route and schematic diagram of organosoluble chiral zirconium phosphonate-supported ruthenium catalysts.
unsymmetrical benzils with p- and m-substituents in a mixture of
triethylamine and formic acid, the high catalytic activities (>97.1%
conv.) and excellent enantiomeric (91.8–99.2%ee syn) and diaste-
reomeric (syn/anti = 11.0–29.4) purities were achieved.
(ꢀ=254 nm). Fourier transform infrared spectra were recorded
on a Perkin–Elmer Model GX Spectrometer using a KBr pellet
method with polystyrene as a standard. Thermogravimetric analy-
sis (TGA) was performed on a SBTQ 600 Thermal Analyzer (USA)
with a heating rate of 20 ◦C min−1 over a temperature range
of 40–800 ◦C under flowing compressed N2 (100 mL min−1). 1H,
13C and 31P NMR were performed on a Bruker AV-300 NMR
instrument at ambient temperature at 300, 75, and 121 MHz,
respectively. All chemical shifts were reported downfield in ppm
relative to the hydrogen, carbon, and phosphorus resonance of TMS,
chloroform-d1, and H3PO4 (85%), respectively. Elemental analysis
was obtained from an EATM 1112 automatic elemental analyzer
instrument (Thermo, USA). The morphologies of as-synthesized
samples were determined by a Hitachi model H-800 transmission
electron microscope (TEM) and Tecnai G2 F20 (HRTEM) operated
2. Experimental
2.1. General remarks
All chemicals were purchased and used without any further
purification. The starting materials HS(CH2)2S(CH2)nP(O)(OC2H5)2
(n = 2–6) were synthesized according to the reference [29] and
ascertained by 1H and 13C NMR.
TLC was performed on pre-coated aluminum-backed plates,
and spots were made visible by quenching ultraviolet fluorescence

Y. Du et al. / Applied Catalysis A: General 479 (2014) 49–58
51
at 200 kV. N2 adsorption–desorption analysis was carried out at
77 K on an Autosorb-1 apparatus (Quantachrome) and specific
surface areas and pore diameters were calculated by the B.E.T.
method and BJH model, respectively. The anti/syn ratios of 3-
methoxyhydrobenzoin were determined by 1H NMR of crude
products in CDCl3, whose chemical shifts of CHOH were 4.65
and 4.78 ppm, respectively. The enantiomeric excesses (%ee) of
two pairs of enantiomers were determined on Agilent LC-1200
HPLC with a Lux 5u Amylose-2 or Daicel Chiralcel OJ-H chiral col-
umn: eluents, (n-hexane/2-propanol = 80/20); 20 ◦C; UV detector,
254 nm; flow rate, 0.5 mL/min. A mixture containing 1-phenyl-
2-(3-methoxyphenyl)ethanedione (tR = 39.945 min), 2-hydroxy-2-
(m-methoxyphenyl) acetophenone (HmMPA) [tR = 54.111 min (S)
and 68.884 min (R)], 2-hydroxy-2-phenyl-m-methoxy acetophe-
none (HPMA) [tR = 65.063 min (S) and 76.590 min (R)], and
3-methoxyhydrobenzoin [tR = 47.257 min (R,R), 57.126 min (S,S),
61.603 min (R,S), and 62.851 min (S,R)] was completely isolated and
identified by using HPLC on an OJ-H column: eluents, n-hexane/i-
PrOH (90:10); flow rate, 1.0 mL min−1; UV detector, ꢀ = 220 (see Fig.
S4 in ESI†).
Fig. 2. The IR spectra of zirconium phosphonates 3a–c and recovered 3a–Ru.
2.2. General synthetic procedure for 1a–c
In
25 mL
of
round-bottomed
flask
was
charged
HS(CH2)2S(CH2)nP(O)(OC2H5)2 (n = 2,
4
and 6) (20.0 mmol),
Ts-DPEN (0.38 g, 1.0 mmol) and CeCl3 (0.025 g, 0.1 mmol), suc-
cessively. The reaction mixture was stirred at 35 ◦C for 72 h and
evaporated under reduced pressure. The residues were subjected
to flash column chromatography eluting with CHCl3/MeOH
(v/v = 100/1 → 80/1 → 60/1) to remove excess ethyl phosphonate
and then CHCl3/MeOH (v/v = 50/1) to obtain yellow oily liquids
1a–c.
2.3. General procedure for phosphonic acids 2a–c
In 25 mL of round-bottomed flask was added ethyl phospho-
nates 1a–c (1.57 mmol), 0.6 mL of CH2Cl2 and trimethylbromosi-
lane (2 mL, 15.7 mmol) under the protection of argon atmosphere,
stirred at room temperature for 6 h, and concentrated under
reduced pressure. To the residues were added 5 mL of 95% MeOH,
stirred at room temperature for 4 h, and concentrated under
reduced pressure to afford viscous solid. The crude product was
added 2 mL of water, crushed by ultrasonic, filtered, washed with
water and pentane (3 × 5 mL), and dried to afford the yellow solids
2a–c.
1a: 0.48 g, 75%, yellow oily liquid. 1H NMR (300 MHz, CDCl3):
ı = 7.36–6.98 (m, 14 H, Ar H), 4.14 (d, 3J = 3.0 Hz, 1 H, CHN),
4.12–4.10 (m, 5 H, OCH2 and CHN), 2.85–2.73 (m, 12 H, CH2 and
NH2), 2.08–1.99 (m, 3 H, CH2P and NH), 1.33–1.29 (m, 6 H, CH3). 13
C
NMR (75.0 MHz, CDCl3): ı = 144.19, 141.08, 138.75, 138.10, 128.51,
128.25, 128.16, 128.07, 127.86, 127.18, 127.01, 126.77, 126.73,
126.62, 126.32, 126.16 (Ph), 63.09 (CHN), 61.56 (d, 2Jc-p = 6.8 Hz,
OCH2), 60.23 (CHN), 35.63, 32.96, 31.85, 31.07 (CH2S), 26.52 (d,
1Jc-p = 135.6 Hz, CH2P), 24.83 (CH2), 16.14 (d, 3Jc-p = 5.90 Hz, CH3).
31P NMR (75.0 MHz, CDCl3): ı = 29.77 (s).
2a: 0.82 g, 90%, yellow solid. 1H NMR (300 MHz, DMSO):
ı = 7.36–6.78 (m, 14 H, Ar H), 4.62 (d, 3J = 6.9 Hz, CHN, 1 H), 4.43
(d, 3J = 6.9 Hz, CHN, 1 H), 2.70 (m, 10 H, CH2 and CH2S), 1.83–1.72
(m, 3 H, CH2P). 13C NMR (75.0 MHz, DMSO): ı = 145.15, 138.88,
135.84, 134.62, 129.11, 129.00, 128.79, 128.68, 128.10, 128.05,
127.77, 127.69, 127.15, 126.86, 126.81, 126.52, 126.42, 126.37 (Ph),
61.90 (CHN), 58.83 (CHN), 35.60, 32.50, 31.63 (CH2), 29.65 (d,
1Jc-p = 129.8 Hz, CH2P), 25.74, 22.17 (CH2). 31P NMR (75.0 MHz):
ı = 23.28 (s).
1b: 0.46 g, 73%, yellow oily liquid. 1H NMR (300 MHz, CDCl3):
ı = 7.50–7.11 (m, 14 H, Ar H) 4.42 (d, 3J = 3.0 Hz, CHN, 1 H),
4.18–4.07 (m, 5 H, OCH2 and CHN), 2.85–2.56 (m, 12 H, CH2S,
PhCH2 and NH2), 1.90–1.84 (m, 7 H, CH2CH2CH2P and NH), 1.32
(t, 3J = 6.8 Hz, 6 H, CH3). 13C NMR (75.0 MHz, CDCl3): ı = 144.18,
140.57, 138.49, 138.08, 128.76, 128.19, 128.03, 127.78, 127.68,
127.22, 126.94, 126.81, 126.74, 126.55, 126.46, 126.21 (Ph), 63.11
(CHN), 61.18 (d, 2Jc-p = 6.5 Hz, OCH2), 60.13 (CHN), 38.32, 35.63,
32.93, 32.08, 31.17 (CH2), 29.99 (d, 2Jc-p = 16.4 Hz, CH2), 24.93 (d,
1Jc-p = 140.2 Hz, CH2P), 21.34 (d, 3Jc-p = 5.0 Hz, CH2), 16.13 (d, 3Jc-
p = 5.9 Hz, CH3). 31P NMR (75.0 MHz, CDCl3): ı = 33.00 (s).
2b: 91%, yellow solid. 1H NMR (300 MHz, DMSO): ı = 7.36–6.77
(m, 14 H, Ph H), 4.63 (d, 3J = 9.6 Hz, 1 H, CHN), 4.53 (d, 3J = 9.6 Hz,
1 H, CHN), 2.69–2.51 (m, 10 H, CH2 and CH2S), 1.43–1.61 (m, 6
H, CH2). 13C NMR (75.0 MHz, DMSO): ı = 144.95, 138.47, 135.42,
134.15, 129.75, 129.69, 128.86, 128.71, 128.52, 128.43, 127.84,
127.75, 127.48, 127.40, 126.87, 126.55, 126.26, 126.09 (Ph), 61.58
(CHN), 58.58 (CHN), 35.31, 32.24, 31.52, 30.84, 30.30, 30.09 (CH2),
27.25 (d, 1Jc-p = 135.8 Hz, CH2P), 22.16 (d, 2Jc-p = 4.5 Hz, CH2). 31P
NMR (75.0 MHz, CDCl3): ı = 27.50 (s).
1c: 0.52 g, 76%, yellow oily liquid. 1H NMR (300 MHz, CDCl3):
ı = 7.38–6.97 (m, 14 H, Ar H), 4.43 (d, 3J = 3.0 Hz, CHN, 1 H),
4.20–4.04 (m, 5 H, OCH2 and CHN), 2.85–2.73 (m, 8 H, CH2S and
CH2), 2.54 (t, 3J = 5.0 Hz, 3 H, CH2S and NH), 1.78–1.60 (m, 8 H, CH2
and NH2), 1.47 (d, 3J = 18.0 Hz, 4 H, CH2CH2P), 1.29 (t, 3J = 9.0 Hz,
6 H, CH3). 13C NMR (75.0 MHz, CDCl3): ı = 144.23, 140.74, 138.56,
138.14, 128.46, 128.19, 128.12, 128.05, 127.95, 127.80, 127.21,
126.95, 126.81, 126.77, 126.71, 126.43 (Ph), 63.11 (CHN), 61.11 (d,
2Jc-p = 6.6 Hz, OCH2), 60.15 (CHN), 35.65, 32.93, 32.10, 31.92, 29.91,
29.69, 29.0, 27.97 (CH2S and CH2), 25.29 (d, 1Jc-p = 140.3 Hz, CH2P),
22.00 (d, 2Jc-p = 5.2 Hz, CH2), 16.14 (d, 3Jc-p = 6.0 Hz, CH3). 31P NMR
(75.0 MHz, CDCl3): ı = 29.52 (s).
2c: 89%, yellow solid. 1H NMR (300 MHz, DMSO): ı = 7.32–6.77
(m, 14 H, Ph H), 4.61 (d, 3J = 6.0 Hz, 1 H, CHN), 4.44 (d, 3J = 6.0 Hz,
2 H, CHN), 2.68–2.50 (m, 10 H, CH2), 1.50–1.33 (m, 10 H, CH2). 13
C
NMR (75.0 MHz, DMSO): ı = 144.90, 138.51, 135.44, 134.19, 129.73,
129.67, 129.17, 128.83, 128.67, 128.49, 128.41, 127.82, 127.72,
127.43, 127.36, 126.52, 126.23, 126.04 (Ph), 61.58 (CHN), 58.55
(CHN), 35.30, 32.24 (CH2), 31.53 (d, 3Jc-p = 4.5 Hz, CH2P), 31.17
(CH2), 29.79 (d, 1Jc-p = 16.0 Hz, CH2P), 29.15, 28.58, 28.00, 26.77
(CH2), 22.83 (d, 2Jc-p = 4.5 Hz, CH2). 31P NMR (75.0 MHz, CDCl3):
ı = 27.80 (s).

52
Y. Du et al. / Applied Catalysis A: General 479 (2014) 49–58
Fig. 3. The SEM images of zirconium phosphonates 3a prepared in various mediums.
2.4. Preparation of zirconium phosphonates 3a–c
2.6. Asymmetric transfer hydrogenation
A mixture of phosphonates 2a (250 mg, 0.43 mmol), acetic acid
(0.3 mL), concentrated hydrochloric acid (1.0 mL, 36%) and 2 mL of
DMF was stirred at room temperature until the solids were com-
pletely dissolved. Then 0.2 mL of aqueous ZrOCl2 solution (275 mg,
0.86 mmol) was added dropwise with stirring and aged at room
temperature for 1 h. The white solids were filtered, washed by DMF
(3 × 2 mL) and dried to afford zirconium phosphonate 3a (0.41 g).
According to the same procedure, 0.37 g of 3b and 0.32 g of 3c were
also obtained.
The reaction mixture of Ru catalyst 3a–Ru (1.7 wt%, 4.0 mg,
6.7 × 10−4 mmol), unsymmetrical benzyl (0.084 mmol) and 0.3 mL
of solvent was stirred to be dissolved completely, added 0.6 mL
of HCO2H/Et3N (v/v = 1/2), and stirred at 50 ◦C for another 40 h.
In reaction mixture, 3a–Ru catalyst was recovered in quantitative
yield by addition of ethyl acetate and centrifugation. The organic
phase was concentrated under reduced pressure, and the crude
residues were purified by flash column chromatography eluting
with petroleum ether/ethyl acetate (v/v = 20:1 → 10:1 → 2:1) to
obtain pure hydrobenzoin product.
2.5. General preparation of supported ruthenium catalysts
3a–c–Ru
3. Results and discussion
In 25 mL of round-bottomed flask was charged zirconium
phosphonate 3a (30 mg), 1.5 mL of acetonitrile and 0.4 mL of
triethylamine, stirred to be dissolved fully, added 1 mL of acetoni-
trile solution containing [RuCl2(p-cymene)]2 complexes (10 mg,
0.016 mmol), and then stirred at 82 ◦C for 6 h. To the reaction
mixture was added 3 mL of ethyl acetate to precipitate zirconium
phosphonate-supported ruthenium catalyst 3a–Ru. The 1.7 wt% Ru
content of 3a–Ru was determined by ICP method.
3.1. Covalent attachment of (R,R)-TsDPEN
The covalent attachment of (R,R)-TsDPEN on the frame of zirco-
nium phosphonates 3a–c was successfully verified by IR spectra
in the range of 4000–400 cm−1 (Fig. 2). The broad and strong
absorption band extending from 3700 to 2500 cm−1 and centered
at 3420 cm−1 was assigned to the O H stretching vibration, which
illustrated the presence of hydroxyl group or adsorbed water on the

Y. Du et al. / Applied Catalysis A: General 479 (2014) 49–58
53
Fig. 4. The HRTEM images of zirconium phosphonate-supported Ru catalyst 3a–Ru.
internal and external surfaces. The moderate absorption bands at
1631, 1590, and 700 cm−1 were attributed to the stretching vibra-
tion mode and flexural vibration of phenyl group, respectively. The
stretching vibration of C H in CH2 group emerged at 2922 cm−1
was strengthened with the increase of the arm chain lengths (n)
from 2 to 6. The strong and wide absorption band centered at about
calculated to be y = 1.13, 1.43 and 0.96. Accompanied by the
endothermic peaks in the DSC curves, the sharp and slow weight
losses of the appended organic moieties in 3a–e in the tem-
perature range of 150–380 ◦C and 380–800 ◦C corresponded to
the initial and deep decomposition of appended organic frag-
ments, respectively. Based on the 63.5, 63.4 and 64.3% total weight
losses of appended organic moieties in 150–800 ◦C, the x values
were calculated to be 0.85, 0.83 and 0.78, which were in the
good agreement with the chemical analysis results and elemen-
tal analysis. Based on the data obtained from chemical analysis
and TGA, the pendant amounts of chiral ligand (R,R)-TsDPEN
(x) decreased with the increase of arm chain lengths (n) from
2 to 6.
1030 cm−1 was the characteristic of Zr
[30].
O P stretching vibration
3.2. Chemical composition of zirconium phosphonates 3a–c
F− is a good ligand for transitional metals such as zirconium(IV)
with a very high oxidation state. Upon adding 2 mL of HF (30%) and
2 mL of deionized water to zirconium phosphonates 3a–c (50 mg)
and stirring for 6 h at room temperature in a sealed plastic con-
tainer, 3a–c could be decomposed to the water-soluble mixture
3.3. Solubility in organic solvents
2−
of complex ZrF6
and phosphonic acids 2a–c. Then the con-
tent of zirconium(IV) and phosphonic acids 2a–c in the resulting
Generally, the overwhelming majority of inorganic materials
such as silica, zeolite and zirconium phosphonate used as catalyst
supports were difficult to be dissolved in organic solvents. Fortun-
ately, by means of the covalent attachment of chiral (R,R)-DPEN
to zirconium phosphonate, the organosoluble inorganic–organic
hybrid zirconium phosphonates 3a–c (n = 2, 4 and 6) with the
solubilities of 50–100 mg mL−1 in CH2Cl2, CH3CN and acetone
were prepared through the free radical addition of Ts-DPEN
with HS(CH2)nPO(OC2H5)2 (n = 2, 4 and 6), hydrolysis of ethyl
phosphonate, and precipitation with ZrOCl2. Furthermore, the
organosolubility of zirconium phosphonate-supported ruthenium
solution were determined by quantitative colorimetric assay and
1
solution 31P { H} NMR to afford the molar ratios (x) of Zr/P, respec-
tively. The obtained empirical formulae [Zr(OH)4 − 2x(O3PR)x·yH2O,
R = pendant (R,R)–DPEN] (3a: x = 0.88; 3b: x = 0.85 and 3c: x = 0.81)
were in good agreement with the detected results of TGA and ele-
mental analysis. From the DTG and TG curves of 3a–e (see ESI†),
it could be deduced from the weight loss below 150 ◦C that 3.9,
3.2 and 2.7% of surface-bound or intercalated water was absorbed
in the internal pores, and the process was endothermic from the
DSC curve. Then the absorbed waters in empirical formulae were

54
Y. Du et al. / Applied Catalysis A: General 479 (2014) 49–58
Table 1
Asymmetric hydrogenation of unsymmetrical benzils under various conditionsa
.
Entry
R
Cat.
Solvent Ab
Solvent Bc
Conv. (%)d
Yield (%)e
%ee (syn)f
Dr (syn/anti)g
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
m-OCH3
m-OCH3
m-OCH3
m-OCH3
m-OCH3
m-OCH3
m-OCH3
m-OCH3
m-OCH3
m-OCH3
m-OCH3
m-OCH3
m-OCH3
m-OCH3
m-F
m-CF3
p-Cl
p-Br
p-CH3
o-F
o-Cl
o-Br
o-CF3
o-OCH3
o-CH3
3a–Ru
3b–Ru
3c–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
3a–Ru
1a–Ru
2a–Ru
DMF
DMF
DMF
DMF
DMF
DMF
DMF
DMF
DMF
H2O
CH3OH
DMSO
2-PrOH
Acetone
DMF
DMF
DMF
DMF
DMF
DMF
DMF
DMF
DMF
DMF
DMF
DMF
DMF
CH3CN
CH3CN
CH3CN
Acetone
CH2Cl2
CH3OH
DMF
DMSO
H2O
H2O
>99
>99
>99
>99
98.1
71.0
72.7
83.8
>99
>99
>99
>99
98.3
95.6
>99
98.0
>99
>99
97.1
98.5
86.0
81.7
–
95.2
96.6
95.8
96.2
95.2
67.8
66.7
77.9
94.2
94.4
95.8
95.1
93.6
92.2
96.2
96.0
97.4
95.9
92.5
91.1
81.4
75.6
–
99.2
98.7
98.6
98.0
99.3
97.9
98.9
98.8
97.7
94.9
96.5
96.2
96.3
96.8
98.2
97.6
94.8
97.3
91.8
17.0
60.4
93.1
–
29.0
28.2
25.0
27.5
29.4
17.1
12.3
11.7
6.9
6.0
5.4
5.0
5.2
H2O
H2O
H2O
H2O
5.2
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
12.3
14.0
11.0
14.5
25.4
1.8
1.3
21.0
–
1.5
5.1
13.9
16.8
67.3
73.6
>99
>99
58.9
64.8
95.6
94.8
35.5
84.8
96.8
96.7
m-OCH3
m-OCH3
a
0.8% mol Cat., 0.6 mL HCOOH/Et3N (v/v = 1:2), 0.3 mL solvent, 50 ◦C, 40 h.
Medium in the preparation of 3a.
Medium in catalytic reaction.
Determined by HPLC.
Isolated yield.
b
c
d
e
f
Determined by HPLC analysis using a Daicel Chiralcel OJ-H column or Phenomenex Lux 5u Amylose-2 chiral column.
g
Determined by 1H NMR.
catalyst 3a–Ru increased from 65 to 132 mg mL−1 in CH3CN upon
the immobilization of [RuCl2(p-cymene)]2.
regard to the more delicate surface morphologies, the irregular
ribbings with the 0.5–1.5 nm distances were observed by the
HRTEM images of 3a (Fig. 4).
3.4. Porous structures and surface morphologies of 3a–c
3.5. Characterization of supported Ru catalyst 3a–Ru
The porous structures and surface morphologies of
zirconium phosphonates 3a–c were elaborated by N2
adsorption–desorption isotherms, SEM and HRTEM images.
Based on N2 adsorption–desorption isotherms (see ESI†), per-
formed at 77 K, zirconium phosphonates 3a–c had the low BET
surface areas of 1.9, 3.7 and 14.3 m2 g−1, uniform mesopores of 1.3,
Upon the immobilization of [RuCl2(p-cymene)]2 complexes,
some nanopores below 10 nm in 3a–Ru, especially at about 2–4
and 10 nm, disappeared (see ESI†), which demonstrated that
[RuCl2(p-cymene)]2 complexes occupied those nanopores. Unex-
pectedly, the surface area, uniform mesopore and pore volume
of 1a increased from 1.9 to 4.9 m2 g−1, 1.3 to 2.0 nm and 0.6 to
2.5 × 10−3 cm3 g−1 owing to the supporting effect of [RuCl2(p-
cymene)]2. Compared XPS spectra of 3a–Ru with its support 3a,
the N 1s binding energy decreased from 399.5 to 399.1 with
a 0.4 eV decrement. However, there were only 0.1 eV change
for the other binding energies of C 1s, O 1s, P 2p and S 2p.
Furthermore, the catalyst 3a–Ru produced the excellent steros-
electivities (syn/anti = 11.0–29.4 and 91.8–99.2%ee syn). Based on
the above-mentioned results, [RuCl2(p-cymene)]2 complexes were
considered to be exclusively anchored at the definite anchorage of
amino functional groups ( NH2 and NH ). On the other hand, the
cluster formation of [RuCl2(p-cymene)]2 complexes could not be
observed from the HRTEM images, which illustrated that [RuCl2(p-
cymene)]2 complexes were well-dispersed in the framework of
zirconium phosphonate (Fig. 4). The ruthenium content (1.7 wt%)
in 3a–Ru was determined using ICP method and also ascertained
by XPS.
1.1 and 1.0 nm and pore volumes of 0.6, 1.1 and 3.9 × 10−3 cm3 g−1
,
respectively. From the pore size distributions (PSDs), three rep-
resentative sharp peaks at 2.0, 4.0 and 9.5 nm were observed for
zirconium phosphonate 3a. However, the mesopores in 3b and
c were non-uniformly distributed throughout the whole range
of 1.5–10 nm, which implied that the supporting of pendant
(R,R)–DPEN organic moieties declined owing to the strenghened
flexibilities with the increase of arm chain lengths (n) from 2 to
6.
The surface morphology of zirconium phosphonate 3a as an
example was well clarified by SEM and HRTEM images. From
the SEM images, the morphologies and diameters of zirconium
phosphonate 3a were greatly related to the reaction mediums. In
DMF, DMSO and methanol, the more bulkier diameters of 3a with
3–5 ␮m were obtained. However, in water, acetone, acetonitrile
and 2-propanol mediums, the smaller aggregates of 3a with the
300–600 nm diameters were prepared (Fig. 3), which was proba-
bly the reason that 3a could be dissolved in organic solvents. With

Y. Du et al. / Applied Catalysis A: General 479 (2014) 49–58
55
Table 2
The stereoselectivities and catalytic activities of the intermediates and products in the whole asymmetric hydrogenation of 1-phenyl-2-(3-methoxyphenyl)ethanedionea
.
Entry
T (h)
Benzil (%)
88.7
HmMPA % (%eec)b
HPMA % (%eec)b
(S, S)-hydrobenzoin % (%eec)b
1
2
3
4
5
6
7
8
9
10
11
12
13
0.5
1.0
1.5
2.0
2.5
3
4.3 (−21.3)
7.3 (−14.6)
9.0 (−9.5)
12.3 (−6.3)
13.8 (−3.1)
16.3 (−1.7)
18.2 (2.7)
16.8 (4.6)
14.2 (7.8)
10.5 (−4.0)
5.4 (−7.3)
–
6.9 (−34.4)
11.4 (−29.1)
14.2 (−24.2)
19.4 (−19.1)
21.6 (−15.7)
24.9 (−10.6)
29.7 (1.9)
27.2 (5.1)
20.8 (9.3)
17.5 (−3.1)
9.9 (−6.5)
1.1 (−)
0
80.1
75.5
64.3
58.4
50.0
27.6
7.1
4.4
0
1.2 (99.3)
1.3 (99.3)
4.0 (99.3)
6.1 (99.3)
8.8 (99.3)
24.6 (99.3)
49.0 (99.3)
60.5 (99.3)
72.0 (99.2)
85.0 (99.2)
98.9 (99.2)
100(99.2)
8
13
18
23
29
35
40
0
0
0
–
0
a
0.8% mol Cat. 3a–Ru, 0.6 mL HCOOH/Et3N (v/v = 1:2), 0.3 mL CH3CN, 50 ◦C, 40 h.
Determined by chiral HPLC using Daicel Chiralpak OJ-H chiral column.
%ee = [(R) − (S)]/[(R) + (S)] × 100.
b
c
3.6. Asymmetric hydrogenation of unsymmetrical benzils
their supported Ru catalysts 3a–Ru all gave the good performances
(>95.6% conv. and >94.9%ee syn) in the aqueous heterogeneous
asymmetric hydrogenation of unsymmetrical benzil (R = m-OCH3)
owing to being well-dispersed in water medium. However,
their diasteroselectivities (syn/anti = 5.0–6.9) were not satisfac-
tory (entries 9–14). Fortunately, in the hydrogenation reaction in
organic mediums such as DMF, acetone and acetonitrile, the stere-
oselectivities (97.0–99.2%ee syn and syn/anti = 11.7–29.4) were
significantly improved (entries 1–8). It was worthwhile to note that
the homogeneous/heterogeneous catalytic hydrogenation reaction
resulted in an effect on the conversion and diasteroselectivities
of 1-phenyl-2-(3-methoxyphenyl)ethanedione. Due to the insol-
uble properties of 3a–Ru in CH3OH, DMF and DMSO (entries
6–8), the disappointed conversions (71.0–83.8%) and relatively
low diasteroselectivities (syn/anti = 11.7–17.1) were observed in
the heterogeneous catalytic system. After the supported Ru cat-
alysts 3a–c–Ru with the different arm chain lengths (n = 2, 4 and
6) were homogenized in CH2Cl2, acetone and CH3CN mediums, the
excellent activities (>98.1% conv.) and stereoselectivities (>98.0%ee
syn and syn/anti = 25.0–29.4) were achieved owing to the partial
elimination of mass transfer limitation from the surface to the inte-
rior at the catalyst interface and inside the catalyst (entries 1–5).
Furthermore, due to the confinement effect of porous structure
Chiral 1,2-diols, especially hydrobenzoins, are useful building
materials for the synthesis of biologically active compounds, as well
as chiral ligands and auxiliaries in stereoselective organic synthe-
sis [31–33]. Unfortunately, until now, seldom report is available on
the asymmetric transfer hydrogenation of unsymmetrical benzils
to optically active hydrobenzoins, with regard to unavailable raw
materials, separation of four stereo-isomers and absolute configu-
ration assignment [34,35].
3.6.1. The enhanced catalytic properties in homogeneous system
In view of the different surface morphologies of support 3a
and organosolubilities of catalyst 3a–Ru in various mediums,
3a–Ru with different surface morphologies was used to study the
asymmetric hydrogenation of unsymmetrical benzils in different
solvents. From Table 1, it was found that the employed solvents, in
both the preparation of catalyst support 3a and catalytic reaction,
played important roles in catalytic performances including stere-
oselectivities and activities. Although zirconium phosphonates
3a, prepared in the various solvents such as DMF, H2O, DMSO,
CH3OH and acetone, possessed the different diameters with the
sizes of 3–5 ␮m for large particles and 300–600 nm for small ones,
Fig. 5. The reactive and steroselective profiles of 1-phenyl-2-(3-methoxyphenyl)ethanedione plotted versus time during catalytic reaction.

56
Y. Du et al. / Applied Catalysis A: General 479 (2014) 49–58
Fig. 6. The hydrogenation mechanism of 1-phenyl-2-(3-methoxyphenyl)ethanedione.
in zirconium phosphonate backbone, 3a–Ru displayed the bet-
ter stereoselectivities than its corresponding ruthenium catalysts
1a–Ru (>96.8%ee syn and syn/anti = 13.9) and 2a–Ru (>96.7%ee syn
and syn/anti = 16.8) (entries 26 and 27).
substituents, whatever the electrondonating (R = CH3 and OCH3)
and electron-withdrawing (R = F, Cl, Br and CF3) substituents, were
in the ortho-position at the phenyl group, the catalytic activities
decreased from 91.1 to 0% yield with the increased sizes of the
substituents, and the disappointed stereoselectivities (17.0–84.8%
syn and syn/anti = 1.3–5.1) were also obtained except for o-Br sub-
stituent (entries 20–25).
3.6.2. The catalytic performances of various unsymmetrical
benzils
The optimized protocol was expanded to a wide variety of
unsymmetrical benzils bearing o, m and p-substituents at one of
the aromatic rings to investigate the catalytic performances of
3a–Ru. From Table 1, the unsymmetrical benzils bearing m, p-
substituents, both the electrondonating (R = CH3 and OCH3) and
electron-withdrawing (R = F, Cl, Br and CF3) substituents, were
able to be stereoselectively reduced to optically active hydroben-
zoins with the good to excellent enantioselectivities (91.8–99.2%ee
syn) and diastereoselectivities (syn/anti = 11.0–29.7) in 92.5–97.4%
yields with a S/C molar ratio of 120 in HCOOH/N(C2H5)3 = 1:2
(v/v) at 50 ◦C for 40 h (entries 15–19). Disappointedly, when the
3.7. Mechanism of asymmetric hydrogenation of
1-phenyl-2-(3-methoxyphenyl)ethanedione
To elucidate the asymmetric hydrogenation mecha-
nism of unsymmetrical benzils, the intermediates and their
catalytic stereoselectivities and activities of 1-phenyl-2-(3-
methoxyphenyl)ethanedione as an example were monitored by
HPLC in the entire catalytic process. From Table 2 and Fig. 5,
it was found that the regioselectivities of intermediates 2-
hydroxy-2-(m-methoxyphenyl) acetophenone (HmMPA) and

Y. Du et al. / Applied Catalysis A: General 479 (2014) 49–58
57
Fig. 7. The TEM and HRTEM images of recovered 3a–Ru after the fifth run.
2-hydroxy-2-phenyl-m-methoxy acetophenone (HPMA) were
very poor with the molar ratio of HPMA/HmMPA <1.5, although
two carbonyl groups possessed the different cloud densities
owing to the electron effect of electron-donating m-OCH3 group.
The conversion of 1-phenyl-2-(3-methoxyphenyl)ethanedione to
HPMA/HmMPA and (S,S)-3-methoxyhydrobenzoin was completed
within 40 h at 40 ◦C. Simultaneously, the poor enantioselectivities
with syn/anti = 29.0/1 and 99.2%ee syn,
a
plausible catalytic
mechanism of 1-phenyl-2-(3-methoxyphenyl)ethanedione in
asymmetric transfer hydrogenation was proposed in Fig. 6.
From Fig. 6, 1-phenyl-2-(3-methoxyphenyl) ethanedione was
hydrogenated to (R),(S)-HmMPA and (R),(S)-HPMA (Fig. 5). And
then (S)-HmMPA and (S)-HPMA were stereoselectively reduced
to (S,S)-3-methoxyhydrobenzoin under the control of 3a–Ru.
The remaining (R)-HmMPA and (R)-HPMA were racemized to
(S)-HmMPA and (S)-HPMA through keto–enol tautomerism
equilibrium. The reaction was ongoing until the completion of
asymmetric hydrogenation.
of
intermediates
HMmPA
(−34.4–9.3%ee)
and
HPMA
(−21.3–7.8%ee) were obtained and changed frequently between
the negative and positive values. However, the excellent enantios-
electivities of (S,S)-3-methoxyhydrobenzoin (syn/anti = 29.7,
99.2%ee syn) remained stable were not influenced by the
enantioselectivities of HmMPA and HPMA during the whole
catalytic process, even though the %ee values of intermedi-
ates HmMPA and HPMA changed in the action of vibrating.
Theoretically, it is assumed that (R)-HmMPA and (R)-HPMA
could be hydrogenated to two possible stereoisomers (R,R) and
(R,S)-3-methoxyhydrobenzoin. Actually, the very few (R,R) and
(R,S)-3-methoxyhydrobenzoin in resulting products were unfor-
tunately monitored by HPLC. The events that occurred with
regard to intermediates (R)-HmMPA and (R)-HPMA need to be
discussed.
3.8. The recovery and reuse of 3a–Ru catalyst
The homogeneous supported Ru catalyst 3a–Ru could be quan-
titatively and readily recovered by addition of ethyl acetate
and centrifugation by solid/liquid separation. After reused five
times, the product (S,S)-3-methoxyhydrobenzoin was obtained
with 97.0% conv., 95.6% syn and syn/anti = 20.7 without the signifi-
cant loss of catalytic performances.
After ruthenium catalyst 3a–Ru was precipitated out by adding
poor ethyl acetate and centrifugated from the reaction mixture,
the leaching amount of Ru(II) in the resulting centrifugate was
monitored by ICP–AES. It was found that no leaching of Ru(II) was
detected in the first two runs. However, there was a little leaching
of Ru(II) in the following runs, and 99.1% of the loading Ru(II) was
still anchored after the fifth run, which stated clearly that the Ru(II)
complexes could be strongly bound to the anchor points of amino
groups. Furthermore, TEM, HRTEM and IR were used to character-
ize the recovered 3a–Ru catalyst reused for five cycles. From the
images of 3a–Ru shown in Fig. 7, the phenomenon of particle aggre-
gation was not observed. IR spectra showed that nothing much had
changed for the stretching vibration mode and flexural vibration
To investigate the highly steroselective hydrogenation
of ␣-hydroxy ketones HmMPA and HPMA to (S,S)-3-
methoxyhydrobenzoin, the racemic HmMPA and HPMA were
hydrogenated independently under the same catalytic reac-
tion conditions. Unexpectedly, (R) and (S)-HPMA were found
in the separate hydrogenation of racemic HmMPA. Likewise,
(R) and (S)-HmMPA were also observed in the separate hydro-
genation of racemic HmPMA (Table S5 and S6 shown in ESI†).
Therefore, it was indicated that there was a keto–enol tau-
tomerism equilibrium among (R),(S)-HmMPA and (R),(S)-HPMA
in the catalytic system. Based on the (S,S)-configuration of
final product 3-methoxyhydrobenzoin, it was conjectured that
(S)-HmMPA and (S)-HPMA, induced by 3a–Ru with the chiral
centers of (R,R)-configurations, continued to be asymmetri-
cally hydrogenated to produce (S,S)-3-methoxyhydrobenzoin.
The remaining (R)-HmMPA and (R)-HPMA isomers with the
less hydrogenation reactivity, hindered by 3a–Ru, underwent
of phenyl, CH2 and Zr
O P groups (Fig. 2). However, compared
with fresh 3a–Ru with the ribbings at the 0.5–1.5 nm distance, the
surface morphology of reused 3a–Ru became more obscure and
irregular.
4. Conclusions
a
rapid racemization to (S)-HmMPA and (S)-HPMA isomers
through keto–enol tautomerism equilibrium. Otherwise, the
abundant (R,S)-3-methoxyhydrobenzoin could be theoret-
ically formed by the direct hydrogenation of (R)-HmMPA
and (R)-HPMA. Based on the above-mentioned results and
final product (S,S)-3-methoxyhydrobenzoin in 95.2% yield
This study initiated a novel concept of transforming inorganic
backbone material-supported metal complex from heterogeneous
into homogeneous catalytic system. After being homogenized in
CH2Cl2, acetone and CH3CN mediums, these supported ruthe-
nium catalysts produced better activities and stereoselectivities in

58
Y. Du et al. / Applied Catalysis A: General 479 (2014) 49–58
homogeneous than heterogeneous catalytic system. Upon addition
of poor solvents such as ethyl acetate, these supported Ru catalysts
could be quantitatively and readily recovered by solid/liquid sep-
aration and reused for five times without significant loss of their
catalytic performances.
[11] K. Akagawa, T. Fujiwara, S. Sakamoto, K. Kudo, Chem. Commun. 46 (2010)
8040–8042.
[12] R. Wang, J. Wan, X. Ma, X. Xu, L. Liu, Dalton Trans. 42 (2013) 6513–6522.
[13] Y. Ji, X. Ma, X. Wu, Q. Wang, Appl. Catal., A: Gen. 332 (2007) 247–256.
[14] Z. Li, W. Xue, B. Guan, F. Shi, Z. Shi, H. Jiang, C. Yuan, Nanoscale 5 (2013)
1213–1220.
[15] T. Chen, X. Ma, X. Wang, Q. Wang, J. Zhou, Q. Tang, Dalton Trans. 40 (2011)
3325–3335.
Acknowledgments
[16] Y. Ma, X. Ma, Q. Wang, J. Zhou, Catal. Sci. Technol. 2 (2012) 1879–1885.
[17] A. Cormaa, H. Garcia, Adv. Synth. Catal. 348 (2006) 1391–1412.
[18] V. Polshettiwar, R. Luque, A. Fihri, H.B. Zhu, M. Bouhrara, J.M. Basset, Chem. Rev.
111 (2011) 3036–3075.
[19] W. Wang, X. Ma, J. Wan, J. Cao, Q. Tang, Dalton Trans. 41 (2012) 5715–5726.
[20] D.E. Bergbreiter, J. Tian, C. Hongfa, Chem. Rev. 109 (2009) 530–582.
[21] A. Clearfield, J.A. Stynes, J. Inorg. Nucl. Chem. 26 (1964) 117–213.
[22] A. Clearfield, R.H. Blessing, J.A. Stynes, J. Inorg. Nucl. Chem. 30 (1968)
2249–2258.
The work was supported by the Natural Science Foundation of
China (grants. 21071116) and Science Foundation Project of CQ,
China (CSTC, 2010BB4126).
Appendix A. Supplementary data
[23] D. Lanari, F. Montanari, F. Marmottini, O. Piermatti, M. Orrù, L. Vaccaro, J. Catal.
277 (2011) 80–87.
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.apcata.
2014.03.039.
[24] X.C. Zou, X.K. Fu, Y.D. Li, X.B. Tu, S.D. Fu, Y.F. Luo, X.J. Wu, Adv. Synth. Catal. 352
(2010) 163–170.
[25] B. Gong, X. Fu, J. Chen, Y. Li, X. Zou, X. Tu, P. Ding, L. Ma, J. Catal. 262 (2009)
9–17.
[26] X. Wang, X. Ma, T. Chen, X. Qin, Q. Tang, Catal. Commun. 12 (2011) 583–588.
[27] S. Calogero, D. Lanari, M. Orrù, O. Piermatti, F. Pizzo, L. Vaccaro, J. Catal. 282
(2011) 112–119.
References
[28] X. Ma, J. Liu, L.S. Xiao, R. Chen, J. Zhou, X. Fu, J. Mater. Chem. 19 (2009)
1098–1104.
[29] X. Ma, Y. Wang, W. Wang, J. Cao, Catal. Commun. 11 (2010) 401–407.
[30] X. Wu, X. Ma, Y. Ji, Q. Wang, X. Jia, X. Fu, J. Mol. Catal. A: Chem. 265 (2007)
316–322.
[31] L. Bagnoli, C. Scarponi, M.G. Rossi, L. Testaferri, M. Tiecco, Chem. Eur. J. 17 (2011)
993–999.
[32] C. Zhu, Y. Shi, M. Xu, G. Lin, Org. Lett. 10 (2008) 1243–1246.
[33] R.S. Reddy, I.N.C. Kirana, A. Sudalai, Org. Biomol. Chem. 10 (2012) 3655–3661.
[34] K. Murata, K. Okano, M. Miyagi, H. Iwane, R. Noyori, T. Ikariya, Org. Lett. 1 (1999)
1119–1121.
[1] G. Chelucci, Coord. Chem. Rev. 257 (2013) 1887–1932.
[2] M. Yoon, R. Ssirambalaji, K. Kim, Chem. Rev. 112 (2012) 1196–1231.
[3] P. Barbaro, V. Dal santo, F. Liguori, Dalton Trans. 39 (2010) 8319–8402.
[4] A.C. Laungani, J.M. Slattery, I. Krossing, B. Breit, Chem. Eur. J. 14 (2008)
4488–4502.
[5] A.F. Trindade, P.M.P. Gois, C.A.M. Afonso, Chem. Rev. 109 (2009) 418–514.
[6] B. Pierluigi, L. Francesca, Chem. Rev. 109 (2009) 515–529.
[7] M.F. José, I.G. José, A.M. José, Chem. Rev. 109 (2009) 360–417.
[8] J. Lu, H.T. Patrick, Chem. Rev. 109 (2009) 815–838.
[9] T. Tanaka, K. Akagawa, M. Mitsuda, K. Kudo, Adv. Synth. Catal. 335 (2013)
294–296.
[10] Q. Gao, Y. Liu, S. Lu, J. Li, C. Li, Green Chem. 13 (2011) 1983–1985.
[35] H. Zhang, D. Feng, H. Sheng, X. Ma, J. Wan, Qian Tang, RSC Adv. 4 (2014)
6417–6423.