Ru nanoparticles on TiO2 with various anatase-to-rutile ratios tuned by selective chemical dissolution: Effect of support polymorph composition on selective benzene hydrogenation

Article abstract of DOI:10.1016/j.apcata.2019.02.021

Using P25 TiO₂ as the precursor, phase-pure anatase and rutile TiO₂ as well as TiO₂ with tunable anatase-to-rutile molar ratios (n_A/n_R) were prepared via a facile “selective chemical dissolution” approach. The Ru/TiO₂ catalysts were then synthesized and comprehensively characterized by spectroscopic techniques and transmission electron microscopy. It is confirmed that these catalysts possessed similar physicochemical parameters and hydrophilicity, but different Ru sizes (2.3–4.5 nm). Ru nanoparticles (NPs) supported on TiO₂ with n_A/n_R of 2.66 (Ru/TiO₂-1.5%HF) exhibited moderate size of 3.4 nm and electron-deficient state, which was originated from the unique Ru-on-anatase/rutile junction structure. In selective benzene hydrogenation, the turnover frequencies of benzene of the mixed phase TiO₂-supported Ru catalysts were similar irrespective of the n_A/n_R ratios and ranged between phase-pure anatase and rutile TiO₂-supported Ru catalysts, which can be ascribed to the size effect of Ru. For the cyclohexene selectivity, Ru NPs supported on TiO₂ with n_A/n_R of 0.88–2.66 exhibited higher S₀ (92–95%) than Ru NPs on TiO₂ with other n_A/n_R ratios, in which the Ru/TiO₂-1.5%HF catalyst displayed a maximum of 95%. We demonstrated that in addition to the proper Ru size, the electron-deficiency of Ru on the Ru/TiO₂-1.5%HF catalyst contributed greatly to the highest selectivity since it can promote the desorption of cyclohexene to avoid suffering over-hydrogenation.

Full text of DOI:10.1016/j.apcata.2019.02.021

Accepted Manuscript
Title: Ru nanoparticles on TiO2 with various anatase-to-rutile
ratios tuned by selective chemical dissolution: Effect of
support polymorph composition on selective benzene
hydrogenation
Authors: Gongbing Zhou, Lan Jiang, Daiping He
PII:
S0926-860X(19)30077-8
DOI:
Reference:
https://doi.org/10.1016/j.apcata.2019.02.021
APCATA 16991
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
10 December 2018
20 January 2019
13 February 2019
Please cite this article as: Zhou G, Jiang L, He D, Ru nanoparticles on TiO2 with
various anatase-to-rutile ratios tuned by selective chemical dissolution: Effect of support
polymorph composition on selective benzene hydrogenation, Applied Catalysis A,
General (2019), https://doi.org/10.1016/j.apcata.2019.02.021
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Title: Ru nanoparticles on TiO2 with various anatase-to-rutile ratios
tuned by selective chemical dissolution: Effect of support polymorph
composition on selective benzene hydrogenation
Author names and affiliation:
Gongbing Zhou*, Lan Jiang, Daiping He
Chongqing Key Laboratory of Inorganic Functional Materials, College of
Chemistry, Chongqing Normal University, Chongqing 401331, People's
Republic of China
*Corresponding author:
Gongbing Zhou
E-mail address: gbzhou@cqnu.edu.cn
Graphical abstract
100
90
80
70
60
0
1
2
3
4
Anatase to rutile ratio in TiO2
1

Research highlights
 TiO2 with different anatase-to-rutile molar ratios (nA/nR) were fabricated.
 The performances of Ru/TiO2 catalysts in selective benzene hydrogenation
were tested.
 TiO2 with nA/nR of 0.88~2.66 supported Ru NPs exhibited high S0 of 92~95%.
 The impact of nA/nR on activity and selectivity of Ru/TiO2 catalysts was
unveiled.
Abstract
Using P25 TiO2 as the precursor, phase-pure anatase and rutile TiO2 as well as TiO2
with tunable anatase-to-rutile molar ratios (nA/nR) were prepared via a facile
“selective chemical dissolution” approach. The Ru/TiO2 catalysts were then
synthesized and comprehensively characterized by spectroscopic techniques and
transmission electron microscopy. It is confirmed that these catalysts possessed
similar physicochemical parameters and hydrophilicity, but different Ru sizes
(2.3~4.5 nm). Ru nanoparticles (NPs) supported on TiO2 with nA/nR of 2.66
(Ru/TiO2-1.5%HF) exhibited moderate size of 3.4 nm and electron-deficient state,
which was originated from the unique Ru-on-anatase/rutile junction structure. In
selective benzene hydrogenation, the turnover frequencies of benzene of the mixed
phase TiO2-supported Ru catalysts were similar irrespective of the nA/nR ratios and
ranged between phase-pure anatase and rutile TiO2-supported Ru catalysts, which can
be ascribed to the size effect of Ru. For the cyclohexene selectivity, Ru NPs supported
on TiO2 with nA/nR of 0.88~2.66 exhibited higher S0 (92~95%) than Ru NPs on TiO2
with other nA/nR ratios, in which the Ru/TiO2-1.5%HF catalyst displayed a maximum
of 95%. We demonstrated that in addition to the proper Ru size, the
electron-deficiency of Ru on the Ru/TiO2-1.5%HF catalyst contributed greatly to the
2

highest selectivity since it can promote the desorption of cyclohexene to avoid
suffering over-hydrogenation.
Keywords: Ru; TiO2; anatase-to-rutile ratio; benzene; hydrogenation
3

1. Introduction
Selective hydrogenation of benzene to cyclohexene is an industrially valuable
reaction since the reactive product can be easily converted to multifunctional and
high-valued chemicals including cyclohexanol, caprolactam, and adipic acid via
simple olefin reactions. Unfortunately, the attainment of a high yield of cyclohexene
from this route that featured by operation simplicity, atom economy, and feedstock
abundance is strictly restricted to kinetics and thermodynamics, i.e., the energy barrier
of benzene hydrogenated to cyclohexene is higher than that of cyclohexene
hydrogenated to cyclohexane even on the most selective metal (ruthenium) so far as
evidenced by first-principles calculations [1], and the standard free energy change for
cyclohexene formation (23 kJ mol–1) from benzene is less negative than that of
cyclohexane formation from benzene (98 kJ mol–1) [26]. As a result, engineering
the surface properties of the catalysts to break the kinetics limitation is the most direct
and effective way to promote the selectivity of cyclohexene.
As a catalyst support, TiO2 is widely applied in heterogeneous catalysis owing to its
uniquely electronic feature, eco-friendliness, abundance, and stability [7]. Aside from
the morphology [8], exposed facets [9], grain size [10], crystallinity [11], and dopant
type and concentration [12], tuning the polymorph composition, especially the molar
ratio of the metastable anatase and the stable rutile, is demonstrated to be a facile and
promising route to promote its catalytic potential [1318]. For instance, Kim et al.
prepared Ru/TiO2 catalyst by mixing RuO2/anatase and RuO2/rutile prior to annealing
and reduction, and uncovered a synergy between these two polymorphs in CO2
4

methanation, i. e., the activity of Ru supported on mixed oxides is obviously higher
than the weighted average activity of the two catalysts based on pure polymorphs [14].
They attributed this activity enhancement to the higher Ru dispersion originated from
the migration of RuO2 NPs towards rutile phase and the subsequent stabilization of
the Ru NPs by epitaxy. For selective oxidation of benzyl alcohol over the Pt/TiO2
catalysts, Liu et al. observed a quasi-linear relationship between the activity and the
anatase content in support which consisted of anatase and rutile polymorphs [16]. Rui
et al. found that the CH4 oxidation activity of the Pt/TiO2 catalyst was enhanced when
using anatase–rutile mixture rather than single phase TiO2 as support, owing to the
enhancement of the CH4 chemisorption [17]. For the selective hydrogenation of
benzene to cyclohexene, we previously demonstrated that the Ru/P25 catalyst
displayed superior activity and selectivity than the pure anatase and rutile-supported
Ru catalysts, attributing to the electron-deficient Ru species formed on the
anatase/rutile junction that led to a tightly-bonded benzene species while weakened
the cyclohexene adsorption [18]. However, the specific relationship between the nA/nR
ratios in TiO2 and the catalytic performance of the Ru/TiO2 catalysts in selective
benzene hydrogenation is still unclear, which is believed to have remarkable
significance in uncovering the support effect and thereby rationally designing more
effective support for this reaction.
Simply mixing anatase and rutile is an easy way to obtain mixed-phase TiO2
powders with controllable ratios, while it is hard to guarantee the sufficient contact
between the two polymorphs, and as a result, the synergy is often unobservable [19].
5

To solve this issue, annealing is often carried out according to the fact that
transformation of a certain fraction of anatase into rutile will occur when gradually
elevating the annealing temperature and thereby the connection between anatase and
rutile was enhanced [14,20]. Unfortunately, in this process, sintering is always
accompanied with the variation of the nA/nR ratio, which led to the simultaneous
changes in particle size, surface area, and crystallinity, and consequently, impeding
the achievement of any conclusion about the effect of the anatase-to-rutile ratio alone.
2


To efficiently weaken the sintering degree, additives including SO4 [21], NO3 [22],
and F and Cl [23] were added into the preparation precursor. Nevertheless, this
strategy often introduced undesired impurities. Other approaches, such as
hydrothermal method [24,25], supercritical anti-solvent precipitation [26],
nonaqueous reaction route [27], “low temperature dissolution–reprecipitation process”
and solvethermal method [28], and plasma electrolytic oxidation [29], were also
developed to synthesize mixed-phase TiO2 powders with tailored polymorph
compositions. However, these routes are limited for versatile applications by
complicated preparation process, severe fabrication conditions, and/or unattainability
of TiO2 with various nA/nR ratios but other similar properties.
Recently, Jovic et al. thoroughly isolated anatase and rutile from P25 TiO2 by
“selective chemical dissolution” based on the mechanism that anatase and rutile were
dissolved in HF and H2O2–NH3H2O mixed solutions, respectively [30]. Inspired by
this work, we utilized this approach that featured in high flexibility of polymorph
composition control and mild synthetic conditions to fabricate TiO2 with varied nA/nR
6

ratios but similar physicochemical properties and grain sizes by controlling the types
and concentrations of dissolution solutions. The prepared TiO2 samples were served
as supports for Ru NPs to investigate the effect of support polymorph composition on
selective hydrogenation of benzene to cyclohexene. Correlation of the support
polymorph compositions of the Ru/TiO2 catalysts with the catalytic performances in
selective benzene hydrogenation was realized and discussed.
2. Experimental
2.1. Materials
Benzene, RuCl33H2O, and P25 TiO2 were supplied by Sigma-Aldrich, Adamas,
and Degussa, respectively. Other unspecified chemicals were purchased from
Greagent. All chemicals are analytical grade. All the gases were provided by
Chongqing Ruike Gas Co.
2.2 Preparation of TiO2 with various anatase-to-rutile ratios
TiO2 with various nA/nR ratios as well as pure anatase and rutile TiO2 were
prepared by “selective chemical dissolution” approach according to Jovic et al. with
some modifications [30]. Specifically, to obtain pure anatase TiO2, 10 ml of the 2.5 wt%
aqueous ammonia solution was added dropwise to 300 ml of a chilled H2O2 aqueous
solution (30 wt%). After uniformly mixing, 2.7 g of P25 TiO2 was added and stirred
for 36 h at room temperature. The solids were separated by centrifugation, washed
with deionized water until neutrality, and dried at 373 K for 12 h. The obtained
7

product was labelled TiO2-H2O2. To obtain pure rutile and TiO2 with different nA/nR
ratios, 5 g of P25 TiO2 was added to 167 ml of the 1, 1.5, 2, 2.5, and 3 wt% HF
aqueous solution at room temperature, respectively. After stirring for 24 h, the
remaining solids were collected by centrifugation, washed repeatedly with deionized
water until neutrality, and dried at 373 K for 12 h. The products were denoted as
TiO2-1%HF, TiO2-1.5%HF, TiO2-2%HF, TiO2-2.5%HF, and TiO2-3%HF, respectively,
according to the concentration of HF used.
2.3 Preparation of the Ru/TiO2 catalysts
The catalysts were fabricated by the wetness impregnation-chemical reduction
approach. Namely, after dispersing 0.5 g of the prepared TiO2 in 10 ml of deionized
H2O, 1.5 ml of 0.4 M RuCl3 aqueous solution was added dropwise and stirred for 3.5
h, in which the nominal Ru loading was 10.7 wt%. To ensure completely reducing
RuCl3 to metallic Ru, 1.5 ml of the 1.58 M KBH4 aqueous solution was added
dropwise. The resulting black precipitates were washed repeatedly with deionized
1
H2O until Cl was undetected. The as-prepared catalysts were labelled Ru/TiO2-H2O2,
Ru/TiO2-1%HF, Ru/TiO2-1.5%HF, Ru/TiO2-2%HF, Ru/TiO2-2.5%HF, and
Ru/TiO2-3%HF, respectively, corresponding to the types of TiO2 support.
2.4. Catalytic testing
To evaluate the catalytic performance of the Ru/TiO2 catalysts in selective benzene
hydrogenation, water (50 ml), zinc sulfate (5.0 g), benzene (25 ml), and the catalyst
8

(0.5 g) were successively added to a 250 ml-capacity autoclave equipped with a
mechanical stirrer, and the autoclave was sealed and purged with hydrogen. The
reaction conditions are referenced to the literatures and typical for this reaction, that is,
the hydrogen pressure of 5.0 MPa, the temperature of 413 K, and the stirring speed of
1000 rpm [26]. The reaction process was monitored by ejecting a little mixture at
intervals, and the constituents of the organic phase were quantified by a SC-3000B
gas chromatograph. The catalytic testing was operated at least in duplicate, and the
catalytic results from replicate operations agreed to within 2%. The activity and
selectivity were expressed as the turnover frequency (TOF) of benzene and the initial
selectivity towards cyclohexene (S0), respectively, in which the former was calculated
on the basis of the initial weight-specific activity (r0) according to our previous works
[2,18,31], and the latter was acquired from the intercept of the selectivity~time curve.
3. Results and discussion
3.1. Polymorph composition of TiO2
Fig. S1A displays the powder X-ray diffraction (XRD) patterns of the TiO2 samples.
For the TiO2-H2O2 and TiO2-3%HF samples, only one peak at 2 of 25.3 and 27.4
corresponding to the anatase (101) and rutile (110) diffractions (JCPDS 21-1272,
21-1276) was appeared, respectively, indicating that the H2O2–NH3H2O mixed
solution and 3 wt% HF solution are capable of completely isolating anatase and rutile
TiO2. As the HF concentration increased from 1 to 3 wt%, the intensity of anatase
(101) peak decreased while that of the rutile (110) peak increased gradually. On basis
9

of these two peak areas (IA for anatase and IR for rutile), we calculated the mole
fractions of anatase (nA) and rutile (nR) by following equations [32]:
0.8(IA / IR )
nA 
100%
1 0.8(IA / IR )
nR  (1 nA )100%
As shown in Figs. S1A and B, the anatase content varied in the whole range (0~100%)
and a quasi-linearity emerged between the nA/nR ratios and the HF concentrations,
revealing the high efficiency of the “selective chemical dissolution” approach in
tuning the polymorph composition of TiO2. Moreover, the average crystallite sizes of
anatase and rutile estimated from the integral width of the strongest anatase (101) and
rutile (110) diffractions by the Scherrer equation were ca. 20 and 25 nm, respectively,
for all the six TiO2 samples [33]. The unchanged grain sizes allowed for a fair
comparison of the role of nA/nR ratios in selective benzene hydrogenation over the
Ru/TiO2 catalysts.
3.2. Selective hydrogenation of benzene to cyclohexene
In Table 1, the catalytic results of selective benzene hydrogenation over these TiO2
samples supported Ru catalysts are displayed. The initial weight-specific activity (r0),
i.e., the consumed millimoles of benzene per gram of the catalyst per minute at the
–1
starting time, exhibits the highest value of 42.6 mmolC6H6 gcat min–1 for the
–1
Ru/TiO2-H2O2 catalyst and the lowest value of 16.4 mmolC6H6 gcat min–1 for the
Ru/TiO2-3%HF catalyst. The r0 of the mixed phase TiO2-supported Ru catalysts
ranges between the above two catalysts and remains nearly constant irrespective of
10

the nA/nR ratios. In light of the r0 and the Ru dispersions shown in Table 2, the TOFs
of benzene were acquired. It is obvious that the TOFs display the converse variation
sequence with the r0. For the S0, it is impressive that the mixed-phase TiO2-supported
Ru catalysts were all outperformed the phase-pure anatase and rutile-supported Ru
catalysts, showing the existence of synergy between anatase and rutile in selective
benzene hydrogenation. The correlation between the S0 of the Ru/TiO2 catalysts and
the nA/nR ratios in TiO2 (Fig. 1) revealed that TiO2 with nA/nR of 0.88~2.66 supported
Ru NPs (Ru/TiO2-2.5%HF, Ru/TiO2-2%HF, and Ru/TiO2-1.5%HF) exhibited higher
S0 of 92~95% than other catalysts, in which the highest S0 of 95% was appeared on
the Ru/TiO2-1.5%HF catalyst (nA/nR of 2.66), which was higher than that of the
Ru/P25 and TiO2-supported Ru-based bimetallic catalysts [18,31].
Selective benzene hydrogenation was acknowledged as a consecutive reaction [34],
which can be embodied by the hydrogenation courses, using the Ru/TiO2-H2O2,
Ru/TiO2-1.5%HF, Ru/TiO2-2.5%HF, and Ru/TiO2-3%HF catalysts as four
representatives (Fig. 2). As the reaction proceeded, benzene exponentially decreased
in content and cyclohexane linearly increased in content. For cyclohexene, a
maximum content emerged at a certain time. These variations verified the above
conclusion. For the four catalysts, the Ru/TiO2-H2O2 catalyst displayed the lowest
yield of cyclohexene (17%) while the Ru/TiO2-1.5%HF catalyst exhibited the highest
yield (41%).
3.3. Physicochemical properties and electronic state of the Ru/TiO2 catalysts
11

To uncover the origins of the polymorph composition-dependent catalytic
performances of the Ru/TiO2 catalysts, the four representative catalysts
(Ru/TiO2-H2O2, Ru/TiO2-1.5%HF, Ru/TiO2-2.5%HF, and Ru/TiO2-3%HF) were
further characterized to probe their electronic and structural features.
Fig. 3 displays the XRD patterns of the catalysts. After Ru loading, the
diffractograms and the nA/nR ratios of the TiO2 supports are finely maintained,
manifesting that the moderate impregnation–chemical reduction conditions have no
impact on the crystal structure of TiO2. A careful check of Fig. 3 discloses a weak and
diffuse peak at 2 of about 44, assigning to the (101) diffraction of hcp Ru (JCPDS
06-0663), which reveals the high dispersion and small size of the Ru NPs on the
catalysts.
Table 2 summarizes the physicochemical parameters of the Ru/TiO2 catalysts.
Inductively coupled plasma-atomic emission spectroscopy (ICP–AES) discloses that
the Ru loadings on the catalysts are essentially identical and are close to the nominal
value of 10.7 wt%. The N2 physisorption isotherms of the catalysts (Fig. S2A) all
belong to type IV with an H3 hysteresis loop. The similarities in the N2 physisorption
isotherms as well as the Barrett-Joyner-Halenda pore size distributions (Fig. S2B)
reveal the close resemblance of the textural properties of the catalysts, which can be
validated by the closed specific surface areas (SBET) and pore volumes (Vpore). This
resemblance afforded the rationality in investigating the role of nA/nR ratio of TiO2
alone. The Ru dispersions of the Ru/TiO2-1.5%HF and Ru/TiO2-2.5%HF catalysts,
estimated by CO chemisorption, are virtually identical and fall in between the
12

Ru/TiO2-H2O2 and Ru/TiO2-3%HF catalysts, implying that their particle sizes are in
the sequence of Ru/TiO2-H2O2 < Ru/TiO2-1.5%HF  Ru/TiO2-2.5%HF <
Ru/TiO2-3%HF. In view of the similarities of the Ru/TiO2 catalysts in the Ru loadings,
preparation methods, and textural properties, the discrepancy in the Ru dispersions
may be originated from the different interactions between Ru and TiO2 with different
polymorph compositions [35,36].
Fig. S3 displays the Ru 3p and Ti 2p X-ray photoelectron spectroscopic (XPS)
spectra of the Ru/TiO2 catalysts. Referring to the partial overlap of the Ru 3p3/2 and Ti
2p peaks, the spectra are deconvoluted to ascertain the individual proportions. For the
Ru 3p3/2 binding energy (BE) of Ru and RuOx, although slight discrepancy appeared
on different literatures, it is widely accepted that the Ru 3p3/2 peak at BE as low as
around 461.5 eV was assigned to metallic Ru, and the Ru 3p3/2 peak at around 463 eV
was assigned to RuOx or Ru(OH)x [3739]. Therefore, for the Ru/TiO2-H2O2,
Ru/TiO2-2.5%HF, and Ru/TiO2-3%HF catalysts, the Ru 3p3/2 BE at 461.5 eV and the
3p3/23p1/2 BE splitting at 22.2 eV render the accurate assignment of Ru as metallic
state [40]. With regard to the Ru/TiO2-1.5%HF catalyst, a positive shift of 0.3 eV for
both the Ru 3p1/2 and 3p3/2 peaks appeared, which was reasonable since the Ru
3p3/23p1/2 BE splitting is constant regardless of metallic Ru or electron-deficient Ru
species [37,38]. The shift revealed that the Ru NPs are electron-deficient. This
electron-deficiency may be caused by the electron transfer from the Ru NPs to the
holes of rutile at anatase/rutile junction, which will be elaborated below.
The Ru/Ti surface molar ratios of the Ru/TiO2 catalysts resulted from Fig. S3 are
13

listed in Table S1. All four catalysts show a higher Ru/Ti surface molar ratio than the
corresponding bulk one, indicating the remarkable surface segregation of Ru. It is also
clear that the Ru/Ti surface ratio possesses the following order: Ru/TiO2-H2O2 >
Ru/TiO2-1.5%HF  Ru/TiO2-2.5%HF > Ru/TiO2-3%HF. According to Zhu et al. [41],
the XPS peak intensity is proportionated to the surface-to-bulk molar ratio of the
metal NPs when the loading is fixed. Therefore, the sequence of the Ru/Ti surface
ratios is an explicit implication of the same metal dispersion order of the Ru/TiO2
catalysts given their similar SBET values (Table 2), which agrees with the CO
chemisorption result.
3.4. Morphology and microstructure of the Ru/TiO2 catalysts
Fig. 4 shows the transmission electron microscopy (TEM) images with
corresponding particle size distribution (PSD) histograms of the Ru/TiO2 catalysts.
The small and dark Ru NPs (marked by arrows), also confirmed by energy-dispersive
X-ray (EDX) spectra (Fig. S4) and mapping EDX images (Fig. S5), exhibit PSDs that
ranged in 1~4, 2~5, 2~5, and 3~13 nm with mean sizes centered at 2.3, 3.4, 3.5, and
5.9 nm for the Ru/TiO2-H2O2, Ru/TiO2-1.5%HF, Ru/TiO2-2.5%HF, and
Ru/TiO2-3%HF catalysts, respectively. The order of Ru sizes conforms with the CO
chemisorption results. A deeper investigation gives an interesting observation that
while the Ru NPs on the Ru/TiO2-H2O2, Ru/TiO2-2.5%HF, and Ru/TiO2-3%HF
catalysts show a random distribution (Figs. 4a, c, and d), Figs. 4b and S6a reveals that
a majority of the Ru NPs on the Ru/TiO2-1.5%HF catalyst are situated at the
14

boundaries between the TiO2 crystallites. The fractions of the Ru NPs situated at
crystallite boundaries of TiO2-H2O2, TiO2-1.5%HF, TiO2-2.5%HF, and TiO2-3%HF
are approximate 5%, 90%, 10%, and 6%, respectively. This observation demonstrated
that the appropriate nA/nR ratio is key to form the Ru-on-grain boundary structure. To
confirm this intriguing distribution of Ru, the high angle annular dark field scanning
transmission electron microscopy (HAADF-STEM) images and the corresponding
EDX elemental mapping analysis are displayed in Fig. S5. For each catalyst, the
mapping EDX images indicated that Ru and Ti had a similar distribution. In addition,
the HAADF-STEM images clearly revealed that the small and bright Ru NPs are
evenly distributed over the entire region of TiO2-H2O2, TiO2-2.5%HF, and
TiO2-3%HF, while on TiO2-1.5%HF, the Ru NPs are mainly distributed on TiO2 grain
boundaries.
The microstructures of the Ru/TiO2 catalysts were further investigated by the
high-resolution TEM (HRTEM) images. As shown in Fig. 5, the Ru NPs with the
interplanar spacing of 2.05 Å attributable to (101) planes have the distinct long-range
disordered and short-range ordered atomic arrangement, conforming with the feature
of chemical reduction [42]. The relative sizes of Ru on the four catalysts agree with
the PSD results. Moreover, only one type of lattice fringes with the interplanar
spacing of ca. 3.52 Å ascribable to anantase (101) planes or 3.25 Å ascribable to rutile
(110) planes is visualized on TiO2 supports of the Ru/TiO2-H2O2, Ru/TiO2-2.5%HF,
and Ru/TiO2-3%HF catalysts (Figs. 5a, c, and d), whereas the lattice fringes of both
anantase (101) and rutile (110) planes exist on the Ru/TiO2-1.5%HF catalyst (Fig. 5b),
15

and the Ru NPs are located at the anatase/rutile junction, revealing that the
Ru-on-grain boundary structure is actually the Ru-on-anatase/rutile junction structure.
This structure is further confirmed by the HRTEM images of the Ru/TiO2-1.5%HF
catalyst shown in Figs. S6b~e.
Scanlon et al. [43] and Deꢀk et al. [44] discovered that in a mixed anatase-rutile
system, the electrons are accumulated in anatase while the holes in rutile, attributing
to an energy band alignment that both conduction and valence bands in anatase
located at lower energies than in rutile. Based on this fact, the formation mechanism
of the Ru-on-anatase/rutile junction structure is postulated in Scheme 1. In the process
of wetness impregnation-chemical reduction, the electron-enriched anatase at the
anatase/rutile junction acts as the pre-activator to reduce Ru3+ to Ru0 atoms. As the
standard reduction potential of Ru3+ is as low as 0.386 V, the reducibility of Ru3+ to
Ru0 by borohydride should be lower than other noble metal cations [45]. This means
that when the Ru0 atoms are formed on the anatase/rutile junction, these Ru0 atoms
concurrently functions as the nucleation centers for the Ru NPs growth and the
catalyst for borohydride decomposition to generate the strongly reducing hydrogen
atoms to reduce the subsequently added Ru3+ [46]. Meanwhile, the neighbouring
hole-enriched rutile of the anatase/rutile junction anchors the produced Ru NPs by
attracting the electrons of the Ru NPs. As a result, the Ru NPs are mainly situated on
the anatase/rutile junction and electron-deficient. This assumption is rationalized by
the observation that the Ru NPs on the Ru/TiO2-1.5%HF catalyst are
electron-deficient as probed by XPS. In addition, the absence of the
16

Ru-on-anatase/rutile junction structure on TiO2-2.5%HF may be originated from the
low electron–hole separation efficiency, as Kho et al. reported that the suitable nA/nR
ratio was key to the charge separation efficiency [47].
3.5. Hydrophilicity and acidic characteristics of the Ru/TiO2 catalysts
In consideration of the significant impact of the catalyst hydrophilicity and acidity
on selective benzene hydrogenation [33,48], we explored these two properties of the
Ru/TiO2 catalysts by O 1s spectra, and Fourier transform infrared spectroscopy of
adsorbed pyridine (Py-IR) and temperature-programmed desorption (TPD) of NH3,
respectively.
As displayed in Fig. S7, after deconvoluting the O 1s spectra of the Ru/TiO2-H2O2,
Ru/TiO2-2.5%HF, and Ru/TiO2-3%HF catalysts, two peaks at 531.9 and 530.1 eV
attributable to oxygen in surface hydroxyl (OOH) and lattice oxygen of TiO2 (OL) [49],
respectively, were obtained. In the case of the Ru/TiO2-1.5%HF catalyst, aside from
the above two peaks, a weak peak at 533.1 eV corresponding to oxygen in H2O (OH2O
existed [41]. The resulting OOH/Ti molar ratios shown in Table S1 are similar for the
four catalysts, disclosing the analogous hydrophilicity irrespective of the nA/nR ratios
in TiO2 supports.
)
Pyridine is recognized as a suitable probe molecule for FTIR spectroscopy to
characterize the nature of acid sites owing to its strong adsorbability, high stability
and selectivity, and good sensibility to the strength of acid sites [50]. Meanwhile,
referring to the previous work [31], a variety of vibration bands of pyridine resulting
17

from diverse adsorption sites existed in room temperature Py-IR spectrum of TiO2,
which included Lewis acid sites-coordinated pyridine (L-Py) at 1447 cm−1,
physisorbed pyridine (ph-Py) at 1576 cm−1, and strong L-Py at 1597 and 1608 cm−1.
When the recording temperature was elevated to 473 K, the band of ph-Py will
disappeared. Hence, the Py-IR spectra of the Ru/TiO2-H2O2, Ru/TiO2-1.5%HF,
Ru/TiO2-2.5%HF, and Ru/TiO2-3%HF catalysts were collected at 473 K and showed
in Fig. 6. It is evident that aside from the L-Py and strong L-Py, an extra weak band at
1542 cm−1 assigning to Brønsted acid sites-bonded pyridine (PyH+) arises for all four
catalysts. The peak intensities of both L-Py and PyH+ for the Ru/TiO2-1.5%HF,
Ru/TiO2-2.5%HF, and Ru/TiO2-3%HF catalysts are similar and higher that those of
the Ru/TiO2-H2O2 catalyst, which can be quantitatively confirmed by the amounts of
+
Lewis acid sites (nL-Py) and Brønsted acid sites (nPyH ) as calculated from the integral
intensity of the bands at 1447 and 1542 cm–1. As a result, the total amount of acid
+
sites probed by Py-IR (nPy) exhibited the same tendency as the nL-Py and nPyH
.
NH3-TPD, a powerful technique to probe the acid amount benefitting by the small
kinetic diameter of NH3 (2.9 Å), was employed to further verify the impact of the
nA/nR ratio of TiO2 on the acidity of the Ru/TiO2 catalysts. As shown in Fig. 7, an
analogous two peak-profile at 480 and 573 K ascribing to NH3 adsorbed on weak and
medium-strength acid sites [31], respectively, was emerged for the Ru/TiO2-H2O2,
Ru/TiO2-2.5%HF, and Ru/TiO2-3%HF catalysts, while an additional peak at 819 K
attributing to NH3 adsorbed on strong acid sites was appeared for the
Ru/TiO2-1.5%HF catalyst, suggesting that the nA/nR ratio of TiO2 influenced the
18

strength of acid sites in some degree. The overall amounts of acid sites (nNH3
)
estimated from the integral intensities of the desorption peaks are compared in Table 3.
It is clear that the NH3-TPD and Py-IR techniques agree finely on the amounts of acid
sites, and the invariably higher values of nNH3 than nPy can be explained by the higher
basicity and smaller kinetic diameter of NH3.
3.6. Implications of the TiO2 polymorph composition on the activity and selectivity
The variation of the r0 elaborated before can be interpreted by the same changing
trend of the Ru dispersions, as the higher dispersion means the higher active surface
area of the Ru NPs. On the other hand, benzene hydrogenation is recognized to be
structure sensitive, that is, the intrinsic reaction rate depends on the availability of
catalytic sites with a specific configuration and thereby it changes with the particle
size [51], which guides us to ascribe the variation of TOFs to the Ru sizes. Campbell
et al. reported that the TOF of 1,3-cyclohexadiene hydrogenation was higher on larger
Ru NPs, attributing to: 1) the facilitation of the -bond activation of the conjugated
structure by the adequate neighboring surface sites on the larger Ru NPs; and 2) the
losing of part resonance energy of the reactant so as to be more reactive via this
-bond activation [37]. As demonstrated by Fan et al. [1], benzene adopts the
identical adsorption configuration on Ru. Consequently, an analogous activity–size
correlation may be suitable for the benzene hydrogenation on Ru. This assumption
has been validated on the Ru/ZrO2 catalyst in our previous work [52]. As the size of
Ru on TiO2 with different nA/nR ratios renders the order of Ru/TiO2-H2O2 <
19

Ru/TiO2-1.5%HF  Ru/TiO2-2.5%HF < Ru/TiO2-3%HF, the TOFs of benzene exhibit
the same sequence.
For the S0, the similarities in the textural properties and hydrophilicity of the
Ru/TiO2 catalysts allow us to exclude the possibility that these two factors aroused the
differences in the S0. According to previous findings [37,48,52,53], the smallest Ru
size and the lowest amount of acid sites of the Ru/TiO2-H2O2 catalyst and the largest
Ru size of the Ru/TiO2-3%HF catalyst may be account for their lower selectivities,
whereas this argument cannot be applicable to explain the observation that the
Ru/TiO2-1.5%HF and Ru/TiO2-2.5%HF catalysts still displayed disparate selectivities
in spite of their similar Ru sizes and acid amounts. On the other hand, We previously
identified that the interfacial electron-deficient Ru species on the Ru/P25 catalyst
lowered the adsorption strength of cyclohexene, thus granting the catalyst
unprecedented selectivity [18]. As the XPS technique indicated that the Ru NPs on the
Ru/TiO2-1.5%HF catalyst are electron-deficient, we examined the applicability of the
above finding on the Ru/TiO2 catalysts by evaluating the adsorption capacity of
cyclohexene. Fig. 8 shows the TPD profiles of cyclohexene on the Ru/TiO2-1.5%HF
and Ru/TiO2-2.5%HF catalysts. It is evident that two desorption peaks at 343 () and
496/522 K () were emerged for the two catalysts, attributing to physically adsorbed
cyclohexene and chemically adsorbed cyclohexene, respectively, based on the peak
intensity and position. The negative shift of 26 K of the  peak from the
Ru/TiO2-2.5%HF catalyst to the Ru/TiO2-1.5%HF catalyst is a direct evidence that
the electron-deficient Ru species on the latter weakened the adsorption strength of
20

cyclohexene. This conclusion can be further validated by the desorption barriers
derived from the  peaks. Assuming the cyclohexene desorption kinetics of first-order
and the pre-exponential factor of 1  1013 s–1 [54], the desorption barriers are
estimated by the following equation:
Ed
Ed
2lnT  ln  
 ln
RT
R
where T, , Ed, R, and  are peak temperature, heating rate of cyclohexene TPD,
desorption barrier, ideal gas constant, and pre-exponential factor, respectively. The
results are 142 and 149 kJ mol–1 for the Ru/TiO2-1.5%HF and Ru/TiO2-2.5%HF
catalysts, respectively, which signify that the intermediate cyclohexene, once formed
on the Ru NPs of the Ru/TiO2-1.5%HF catalyst, will be easier to desorb in benzene
hydrogenation. Therefore, the suppressed adsorption of cyclohexene on the  sites by
electron-deficient Ru species originated from the Ru-on-anatase/rutile junction
structure may be another crucial factor for the superior selectivity of the
Ru/TiO2-1.5%HF catalyst.
4. Conclusions
Adopting the facile “selective chemical dissolution” approach, pure anatase and
rutile TiO2 were successfully isolated from P25 TiO2 by using H2O2–NH3H2O and HF
as dissolution solutions, respectively, and the anatase-to-rutile molar ratios in TiO2
were easily and controllably tuned by changing the concentration of HF. The Ru NPs
supported on these TiO2 samples exhibited different sizes and chemical states, in
which TiO2 with nA/nR of 2.66 supported Ru NPs possessed moderate size of 3.4 nm
21

and electron-deficient state, which was acknowledged to be aroused from the specific
Ru-on-anatase/rutile junction structure as indicated by TEM. In selective benzene
hydrogenation, the TOF of the Ru/anantase catalyst was the lowest while that of the
Ru/rutile catalyst was the highest, and the TOFs of the mixed phase TiO2-supported
Ru catalysts ranged between the above two catalysts and remained nearly unchanged
with the varying of the nA/nR ratios, which can be ascribed to the size effect of Ru. As
to the selectivity towards cyclohexene, it displayed higher values of 92~95% on TiO2
with nA/nR of 0.88~2.66 supported Ru NPs than on other catalysts, in which the
highest value of 95% was obtained on the Ru/TiO2-1.5%HF catalyst (nA/nR of 2.66). It
is identified that the electron-deficiency of Ru on this catalyst is critical to the highest
selectivity since it effectively lowered the desorption barrier of cyclohexene and
thereby suppressed the adsorption of cyclohexene. The versatile “selective chemical
dissolution” strategy and the relationship between the catalytic performances and the
anatase-to-rutile ratios of TiO2 demonstrated herein offer new ideas for the
development of more effective catalysts for selective benzene hydrogenation by
engineering the polymorph composition of support beyond TiO2.
Competing interests statement
There are no competing interests to declare.
Acknowledgements
This work was supported by National Natural Science Foundation of China (No.
21703024) and Natural Science Foundation of Chongqing, China (No.
22

cstc2018jcyjAX0622, cstc2016jcyjA0392).
References
[1] C. Fan, Y.A. Zhu, X.G. Zhou, Z.P. Liu, Catal. Today 160 (2011) 234241.
https://doi.org/10.1016/j.cattod.2010.03.075
[2] G.B. Zhou, H. Wang, J. Tian, Y. Pei, K.N. Fan, M.H. Qiao, B. Sun, B.N. Zong,
ChemCatChem 10 (2018) 11841191. http://dx.doi.org/10.1002/cctc.201701696
[3] D.M. Rao, X.G. Xue, G.Q. Cui, S. He, M. Xu, W.H. Bing, S.X. Shi, M. Wei, Catal.
Sci. Technol. 8 (2018) 236243. http://dx.doi.org/10.1039/C7CY01351E
[4] J.L. Liu, L.J. Zhu, Y. Pei, J.H. Zhuang, H. Li, H.X. Li, M.H. Qiao, K.N. Fan, Appl.
Catal. A: Gen. 353 (2009) 282–287. https://doi.org/10.1016/j.apcata.2008.10.056
[5] X.L. Yu, Y. Li, S.M. Xin, P.Q. Yuan, W.K. Yuan, Ind. Eng. Chem. Res. 57 (2018)
19611967. https://doi.org/10.1021/acs.iecr.7b04642
[6] H. Spod, M. Lucas, P. Claus, ChemCatChem 8 (2016) 26592666.
http://dx.doi.org/10.1002/cctc.201600493
[7] K. Bourikas, C. Kordulis, A. Lycourghiotis, Chem. Rev. 114 (2014) 97549823.
https://doi.org/10.1021/cr300230q
[8] H.J. Jung, S.H. Kye, H.J. Kang, H.J. Yang, J.B. Yoo, K.H. Lee, N.H. Hur, Appl.
Catal. A: Gen. 558 (2018) 917. https://doi.org/10.1016/j.apcata.2018.03.009
[9] M. Chen, J.Z. Ma, B. Zhang, F. Wang, Y.B. Li, C.B. Zhang, H. He, Appl. Catal. B:
Environ. 223 (2018) 209215. https://doi.org/10.1016/j.apcatb.2017.04.012
[10] M.X. Xu, Y.H. Wang, J.F. Geng, D.W. Jing, Chem. Eng. J. 307 (2017) 181188.
23

https://doi.org/10.1016/j.cej.2016.08.080
[11] M. Bellardita, A. Di Paola, B. Megna, L. Palmisano, Appl. Catal. B: Environ. 201
(2017) 150158. https://doi.org/10.1016/j.apcatb.2016.08.012
[12] K.A. Ali, A.Z. Abdullah, A.R. Mohamed, Appl. Catal. A: Gen. 537 (2017)
111120. https://doi.org/10.1016/j.apcata.2017.03.022
[13] U. Caudillo-Flores, M.J. Muñoz-Batista, A. Kubacka, J. Zárate-Medina, J.A.
Cortés, M. Fernández-García, Appl. Catal. A: Gen. 550 (2018) 3847.
https://doi.org/10.1016/j.apcata.2017.10.016
[14] A. Kim, D.P. Debecker, F. Devred, V. Dubois, C. Sanchez, C. Sassoye, Appl.
Catal. B: Environ. 220 (2018) 615625. https://doi.org/10.1016/j.apcatb.2017.08.058
[15] R. Kaplan, B. Erjavec, A. Pintar, Appl. Catal. A: Gen. 489 (2015) 5160.
https://doi.org/10.1016/j.apcata.2014.10.018
[16] J.J. Liu, S.H. Zou, L.F. Lu, H.T. Zhao, L.P. Xiao, J. Fan, Catal. Commun. 99
(2017) 69. https://doi.org/10.1016/j.catcom.2017.05.015
[17] Z.B. Rui, S.R. Wu, C. Peng, H.B. Ji, Chem. Eng. J. 243 (2014) 254264.
https://doi.org/10.1016/j.cej.2014.01.010
[18] G.B. Zhou, R.F. Dou, H.Z. Bi, S.H. Xie, Y. Pei, K.N. Fan, M.H. Qiao, B. Sun,
B.N. Zong, J. Catal. 332 (2015) 119126. https://doi.org/10.1016/j.jcat.2015.09.016
[19] A.C. Breeson, G. Sankar, G.K.L. Goh, R.G. Palgrave, Appl. Surf. Sci. 423 (2017)
205209. https://doi.org/10.1016/j.apsusc.2017.06.161
[20] T. Cottineau, H. Cachet, V. Keller, E.M.M. Sutter, Phys. Chem. Chem. Phys. 19
(2017) 3146931478. http://dx.doi.org/10.1039/C7CP06270B
24

[21] Y. Ma, Q. Xu, X. Zong, D.E. Wang, G.P. Wu, X. Wang, C. Li, Energy Environ.
Sci. 5 (2012) 63456351. http://dx.doi.org/10.1039/C1EE02053F
[22] F.Y. Xu, W. Xiao, B. Cheng, J.G. Yu, Int. J. Hydrogen Energy 39 (2014)
1539415402. https://doi.org/10.1016/j.ijhydene.2014.07.166
[23] A. Rao, V.S. Pundir, A. Tiwari, Y. Padarthi, N.V.M. Rao, S. Aich, B. Roy, J. Phys.
Chem. Solids 113 (2018) 164176. https://doi.org/10.1016/j.jpcs.2017.10.030
[24] Q. Wang, Z. Qiao, P. Jiang, J.L. Kuang, W.X. Liu, W.B. Cao, Solid State Sci. 77
(2018) 1419. https://doi.org/10.1016/j.solidstatesciences.2018.01.003
[25] M.A. Mahadik, G.W. An, S. David, S.H. Choi, M. Cho, J.S. Jang, Appl. Surf. Sci.
426 (2017) 833843. https://doi.org/10.1016/j.apsusc.2017.07.179
[26] R.P. Marin, S. Ishikawa, H. Bahruji, G. Shaw, S.A. Kondrat, P.J. Miedziak, D.J.
Morgan, S.H. Taylor, J.K. Bartley, J.K. Edwards, M. Bowker, W. Ueda, G.J.
Hutchings,
Appl.
Catal.
A:
Gen.
504
(2015)
6273.
https://doi.org/10.1016/j.apcata.2015.02.023
[27] Y. Qiu, F. Ouyang, R.S. Zhu, Int. J. Hydrogen Energy 42 (2017) 1136411371.
https://doi.org/10.1016/j.ijhydene.2017.03.047
[28] Y.N. Gao, H. Wang, J. Wu, R.H. Zhao, Y.F. Lu, B.F. Xin, Appl. Surf. Sci. 294
(2014) 3641. https://doi.org/10.1016/j.ijhydene.2017.03.047
[29] R. Su, R. Bechstein, L. Sø, R.T. Vang, M. Sillassen, B. Esbjörnsson, A. Palmqvist,
F.
Besenbacher,
J.
Phys.
Chem.
C
115
(2011)
2428724292.
https://doi.org/10.1021/jp2086768
[30] V. Jovic, W.T. Chen, D. Sun-Waterhouse, M.G. Blackford, H. Idriss, G.I.N.
25

Waterhouse, J. Catal. 305 (2013) 307317. https://doi.org/10.1016/j.jcat.2013.05.031
[31] G.B. Zhou, Y.L. Dong, D.P. He, Appl. Surf. Sci. 456 (2018) 10041013.
https://doi.org/10.1016/j.apsusc.2018.06.206
[32] R.A. Spurr, H. Myers, Anal. Chem. 29 (1957) 760762.
https://doi.org/10.1021/ac60125a006
[33] G.B. Zhou, X.H. Tan, Y. Pei, K.N. Fan, M.H. Qiao, B. Sun, B.N. Zong,
ChemCatChem 5 (2013) 24252435. http://dx.doi.org/10.1002/cctc.201300175
[34] G.B. Zhou, H. Wang, Y. Pei, M.H. Qiao, B. Sun, B.N. Zong, Acta Chim. Sinica
75 (2017) 321328. http://sioc-journal.cn/Jwk_hxxb/CN/10.6023/A16100569
[35] T. Komaya, A.T. Bell, Z. Weng-Sieh, R. Gronsky, F. Engelke, T.S. King, M.
Pruski, J. Catal. 150 (1994) 400406. https://doi.org/10.1006/jcat.1994.1358
[36] Q.Q. Lin, X.Y. Liu, Y. Jiang, Y. Wang, Y.Q. Huang, T. Zhang, Catal. Sci. Technol.
4 (2014) 20582063. http://dx.doi.org/10.1039/C4CY00030G
[37] P.S. Campbell, C.C. Santini, F. Bayard, Y. Chauvin, V. Collière, A. Podgoršek,
M.F.
Costa
Gomes,
J.
Sá,
J.
Catal.
275
(2010)
99–107.
https://doi.org/10.1016/j.jcat.2010.07.018
[38] K. Yamaguchi, T. Koike, J.W. Kim, Y. Ogasawara, N. Mizuno, Chem. Eur. J. 14
(2008) 1148011487. https://doi.org/10.1002/chem.200801655
[39] M.H. Liu, J. Zhang, J.Q. Liu, W.W. Yu, J. Catal. 278 (2011) 17.
https://doi.org/10.1016/j.jcat.2010.11.009
[40] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-ray
Photoelectron Spectroscopy, Perkin-Elmer, Minnesota, 1992.
26

[41] Y. Zhu, L. Tian, Z. Jiang, Y. Pei, S.H. Xie, M.H. Qiao, K.N. Fan, J. Catal. 281
(2011) 106118. https://doi.org/10.1016/j.jcat.2011.04.007
[42] Y. Pei, G.B. Zhou, N. Luan, B.N. Zong, M.H. Qiao, F. Tao, Chem. Soc. Rev. 41
(2012) 81408162. http://dx.doi.org/10.1039/C2CS35182J
[43] D.O. Scanlon, C.W. Dunnill, J. Buckeridge, S.A. Shevlin, A.J. Logsdail, S.M.
Woodley, C.R.A. Catlow, M.J. Powell, R.G. Palgrave, I.P. Parkin, G.W. Watson, T.W.
Keal, P. Sherwood, A. Walsh, A.A. Sokol, Nat. Mater. 12 (2013) 798801.
http://dx.doi.org/10.1038/nmat3697
[44] P. Deꢀk, B. Aradi, T. Frauenheim, J. Phys. Chem. C 115 (2011) 34433446.
http://dx.doi.org/10.1021/jp1115492
[45] D.R. Lide, CRC Handbook of Chemistry and Physics, 89th ed., CRC Press, Boca
Raton, 2009.
[46] G.O. Mallory, J.B. Hajdu, Electroless Plating: Fundamentals and Applications,
American Electroplaters and Surface Finishers Society, Orlando, 1990.
[47] Y.K. Kho, A. Iwase, W.Y. Teoh, L. Mädler, A. Kudo, R. Amal, J. Phys. Chem. C
114 (2010) 28212829. https://doi.org/10.1021/jp910810r
[48] G.B. Zhou, Y. Pei, Z. Jiang, K.N. Fan, M.H. Qiao, B. Sun, B.N. Zong, J. Catal.
311 (2014) 393403. https://doi.org/10.1016/j.jcat.2013.12.022
[49] K. Lan, Y. Liu, W. Zhang, Y. Liu, A. Elzatahry, R.C. Wang, Y.Y. Xia, D.
Al-Dhayan, N.F. Zheng, D.Y. Zhao, J. Am. Chem. Soc. 140 (2018) 41354143.
https://doi.org/10.1021/jacs.8b00909
[50] M.I. Zaki, M.A. Hasan, F.A. Al-Sagheer, L. Pasupulety, Colloids Surf., A 190
27

(2001) 261–274. https://doi.org/10.1016/S0927-7757(01)00690-2
[51] V.V. Pushkarev, K. An, S. Alayoglu, S.K. Beaumont, G.A. Somorjai, J. Catal. 292
(2012) 6472. https://doi.org/10.1016/j.jcat.2012.04.022
[52] G.B. Zhou, X.H. Tan, R.F. Dou, Y. Pei, K.N. Fan, M.H. Qiao, B. Sun, B.N. Zong,
Scientia Sinica Chim. 44 (2014) 121–130. https://doi.org/10.1360/032013-248
[53] J. Bu, J.L. Liu, X.Y. Chen, J.H. Zhuang, S.R. Yan, M.H. Qiao, H.Y. He, K.N. Fan,
Catal. Commun. 9 (2008) 2612–2615. https://doi.org/10.1016/j.catcom.2008.07.024
[54] H. Freund, Elementary steps and mechanisms, in: G. Ertl, H. Knözinger, F.
Schüth, J. Weitkamp (Eds.), Handbook of Heterogeneous Catalysis, second ed.,
Wiley-VCH, Weinheim, 2008, p. 1381.
28

Figure captions
Fig. 1. The relationship between the nA/nR ratios in TiO2 and the S0 of the Ru/TiO2
catalysts.
Fig. 2. The time courses of benzene hydrogenation over the (a) Ru/TiO2-H2O2, (b)
Ru/TiO2-1.5%HF, (c) Ru/TiO2-2.5%HF, and (d) Ru/TiO2-3%HF catalysts. Reaction
conditions: 0.5 g of catalyst, 25 mL of benzene, 50 mL of H2O, 5.0 g of ZnSO47H2O,
temperature of 413 K, H2 pressure of 5.0 MPa, and stirring rate of 1200 rpm. (■)
benzene, (●) cyclohexene, and (▲) cyclohexane.
Fig. 3. Powder XRD patterns of the (a) Ru/TiO2-H2O2, (b) Ru/TiO2-1.5%HF, (c)
Ru/TiO2-2.5%HF, and (d) Ru/TiO2-3%HF catalysts.
Fig. 4. TEM images and the corresponding PSD histograms of the Ru NPs with
Gaussian analysis fittings of the (a) Ru/TiO2-H2O2, (b) Ru/TiO2-1.5%HF, (c)
Ru/TiO2-2.5%HF, and (d) Ru/TiO2-3%HF catalysts.
Fig. 5. HRTEM images of the (a) Ru/TiO2-H2O2, (b) Ru/TiO2-1.5%HF, (c)
Ru/TiO2-2.5%HF, and (d) Ru/TiO2-3%HF catalysts. The dashed lines represent the
anatase and rutile boundaries, and the anatase/rutile junction is marked by arrow.
Scheme 1. Schematic illustration of the formation mechanism of the
electron-deficient Ru NPs-on-anatase/rutile junction structure on the TiO2-1.5%HF
support.
Fig. 6. Py-IR spectra of the (a) Ru/TiO2-H2O2, (b) Ru/TiO2-1.5%HF, (c)
Ru/TiO2-2.5%HF, and (d) Ru/TiO2-3%HF catalysts at 473 K.
Fig. 7. NH3-TPD profiles of the (a) Ru/TiO2-H2O2, (b) Ru/TiO2-1.5%HF, (c)
29