The effect of oxidative and reductive treatments of titania-supported metal catalysts on the pairwise hydrogen addition to unsaturated hydrocarbons

Article abstract of DOI:10.1016/j.cattod.2016.02.030

Heterogeneous hydrogenation of unsaturated compounds with parahydrogen is a highly promising technique for boosting the sensitivity of magnetic resonance spectroscopy and imaging by hyperpolarizing reaction products in gaseous and liquid phases, and potentially reaction intermediates as well. This demands an efficient heterogeneous catalyst providing both the high selectivity toward pairwise hydrogen addition (i.e., ability to incorporate both H atoms of H₂molecule in the same product molecule) as well as sufficient overall hydrogenation activity. In this work, we studied the influence of oxidative and reductive treatments of the supported metal catalysts on the NMR signal enhancements provided by parahydrogen-induced polarization (PHIP) effects in hydrogenation of propene, propyne, 1,3-butadiene and 1-butyne. The 5?wt% titania-supported Pt, Pd, Rh and Ir catalysts used here were characterized by X-ray photoelectron spectroscopy (XPS) and high-resolution transmission electron microscopy (HRTEM). Generally, the preliminarily reduced catalysts were found to be more efficient than the oxidized ones. For instance, while the reduced Ir/TiO₂catalyst provided the intense PHIP NMR signals, its oxidized counterpart showed almost no activity in hydrogenation. For the oxidized Pd/TiO₂catalyst, HRTEM revealed the formation of titania pedestals under large (ca. 5–7?nm) PdO nanoparticles. At the same time, the small (ca. 1?nm) partially reduced Pd^δ+particles were observed on the facets of TiO₂support. These changes in catalyst structure led to a significant decrease in pairwise hydrogen addition selectivity.

Full text of DOI:10.1016/j.cattod.2016.02.030

G Model
CATTOD-10060; No. of Pages7
ARTICLE IN PRESS
Catalysis Today xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
The effect of oxidative and reductive treatments of titania-supported
metal catalysts on the pairwise hydrogen addition to unsaturated
hydrocarbons
a,b,∗
, Dudari B. Burueva , Evgeniy Yu. Gerasimovc,b,
a,b
Oleg G. Salnikov
Andrey V. Bukhtiyarov , Alexander K. Khudorozhkov , Igor P. Prosvirinb,c,
c,b
c,b
c,b
a,d
c,b
Larisa M. Kovtunova , Danila A. Barskiy , Valerii I. Bukhtiyarov
,
a,b
a,b,∗
Kirill V. Kovtunov , Igor V. Koptyug
a
Laboratory of Magnetic Resonance Microimaging, International Tomography Center, SB RAS, 3A Institutskaya St., 630090 Novosibirsk, Russia
Novosibirsk State University, 2 Pirogova St., 630090 Novosibirsk, Russia
Laboratory of Surface Science, Boreskov Institute of Catalysis, SB RAS, 5 Acad. Lavrentiev Pr., 630090 Novosibirsk, Russia
Department of Radiology and Radiological Sciences, Institute of Imaging Science, Vanderbilt University, 1161 21st Avenue South, Medical Center North,
b
c
d
AA-1105, Nashville, TN 37232-2310, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Heterogeneous hydrogenation of unsaturated compounds with parahydrogen is a highly promising tech-
nique for boosting the sensitivity of magnetic resonance spectroscopy and imaging by hyperpolarizing
reaction products in gaseous and liquid phases, and potentially reaction intermediates as well. This
demands an efficient heterogeneous catalyst providing both the high selectivity toward pairwise hydro-
gen addition (i.e., ability to incorporate both H atoms of H2 molecule in the same product molecule) as
well as sufficient overall hydrogenation activity. In this work, we studied the influence of oxidative and
reductive treatments of the supported metal catalysts on the NMR signal enhancements provided by
parahydrogen-induced polarization (PHIP) effects in hydrogenation of propene, propyne, 1,3-butadiene
and 1-butyne. The 5 wt% titania-supported Pt, Pd, Rh and Ir catalysts used here were characterized by
X-ray photoelectron spectroscopy (XPS) and high-resolution transmission electron microscopy (HRTEM).
Generally, the preliminarily reduced catalysts were found to be more efficient than the oxidized ones. For
instance, while the reduced Ir/TiO2 catalyst provided the intense PHIP NMR signals, its oxidized counter-
part showed almost no activity in hydrogenation. For the oxidized Pd/TiO2 catalyst, HRTEM revealed the
Received 5 October 2015
Accepted 5 February 2016
Available online xxx
Keywords:
Parahydrogen-induced polarization (PHIP)
Hydrogenation
Heterogeneous catalysis
X-ray photoelectron spectroscopy (XPS)
Transmission electron microscopy (TEM)
formation of titania pedestals under large (ca. 5–7 nm) PdO nanoparticles. At the same time, the small
ca. 1 nm) partially reduced Pd␦ particles were observed on the facets of TiO2 support. These changes in
+
(
catalyst structure led to a significant decrease in pairwise hydrogen addition selectivity.
©
2016 Elsevier B.V. All rights reserved.
1
. Introduction
conducting mechanistic studies of catalytic processes and for char-
acterizing the behavior of operating model catalytic reactors. In
particular, this toolkit allows one to map in situ the distribution of
the liquid phase within the catalyst bed, to directly visualize var-
ious dynamic processes, and to evaluate local reactant-to-product
conversion [6–8]. However, the NMR-based techniques have a sig-
nificant drawback of low intrinsic sensitivity. One of the efficient
ways to overcome this problem is the use of nuclear spin hyper-
polarization techniques [9,10], e.g., dynamic nuclear polarization
Nowadays, methods based on nuclear magnetic resonance
(
NMR), such as magnetic resonance spectroscopy (MRS) and mag-
netic resonance imaging (MRI), are unique analytical tools which
are widely and routinely used in chemistry, medicine, and many
other areas of research and practice. In the context of operando
studies in catalysis [1–5], the MRI/MRS toolkit can be useful for
(DNP) [11] for hyperpolarization of solids, liquids or solutions,
spin-exchange optical pumping (SEOP) [12] for hyperpolarization
of gases, and parahydrogen-induced polarization (PHIP) [13,14]
which can be successfully utilized for hyperpolarization of both
∗ Corresponding authors at: International Tomography Center, SB RAS, 3A Insti-
tutskaya St., 630090 Novosibirsk, Russia.
E-mail addresses: koptyug@tomo.nsc.ru, salnikov@tomo.nsc.ru (I.V. Koptyug).
http://dx.doi.org/10.1016/j.cattod.2016.02.030
0
920-5861/© 2016 Elsevier B.V. All rights reserved.
Please cite this article in press as: O.G. Salnikov, et al., The effect of oxidative and reductive treatments of titania-supported metal catalysts
on the pairwise hydrogen addition to unsaturated hydrocarbons, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.030

G Model
CATTOD-10060; No. of Pages7
ARTICLE IN PRESS
O.G. Salnikov et al. / Catalysis Today xxx (2016) xxx–xxx
2
Fig. 1. (a) The reaction scheme of propene hydrogenation; (b–d) H NMR spectra acquired during propene hydrogenation with p-H2 at 100 C over PdRed/TiO2 and PdOx/TiO2
1
◦
Red
Ox
Red
Ox
catalysts (b), over Rh /TiO2 and Rh /TiO2 catalysts (c); and over Ir /TiO2 and Ir /TiO2 catalysts (d). All spectra were acquired while the gas was flowing (4.1 mL/s) with
signal accumulations and are presented on the same vertical scale.
8
liquids and gases. The basis of PHIP is the exploitation of the singlet
of unsaturated hydrocarbons over supported metal nanoparticles,
meaning that the pairwise addition of hydrogen is possible with
such catalysts [22]. Subsequent studies showed that PHIP effects
can be observed in heterogeneous gas phase hydrogenations of
various hydrocarbons such as propene [23,24], propyne [25], 1,3-
butadiene and 1-butyne [26] as well as ␣,␤-unsaturated carbonyl
compounds [27] over several Pt, Pd, Rh and Ir [28] catalysts sup-
ported on Al O , SiO and TiO2 as well as over bulk metals and
nuclear spin order of a parahydrogen molecule (p-H ). However,
2
because p-H molecule has zero nuclear spin and therefore is NMR
2
silent, the equivalence of the two H atoms of a p-H molecule needs
2
to be broken in order to get an enhanced NMR signal. This can be
achieved via incorporation of p-H2 in a product of a hydrogena-
tion reaction [15]. Two major requirements for such hydrogenation
reaction should be pointed out: (i) the reaction should proceed in a
pairwise manner of hydrogen addition, meaning that two H atoms
2
3
2
bulk metal oxides [29] (Pt, Rh, CaO, Cr O , CeO [30], PtO , PdO and
2
3
2
2
from the same p-H molecule should be added to the same unsatu-
Rh O ). Moreover, polarization was also successfully observed in
2
2 3
rated substrate molecule; and (ii) the two nascent protons of p-H2
molecule should end up in magnetically non-equivalent positions
in the product molecule [16]. If both these requirements are met,
liquid phase heterogeneous hydrogenations over supported metal
nanoparticles [31–33]. Generally, the highest intensities of PHIP
NMR signals were obtained in heterogeneous hydrogenations cat-
1
the corresponding H NMR signals of reaction products or interme-
alyzed by catalysts supported on TiO [26]. The attempts to explain
2
diates can be significantly enhanced (up to 4–5 orders of magnitude
in the case of standard NMR instruments) [16]. Moreover, the NMR
lines get a characteristic line shape which makes it easy to detect
PHIP phenomena and therefore the existence of pairwise hydrogen
addition route [17].
this fact led to an investigation of the influence of strong metal-
support interaction (SMSI) effect on the pairwise hydrogen addition
and PHIP [34]. However, in spite of all these studies the nature of
active sites of supported metal catalysts which, contrary to the dis-
sociative Horiuti-Polanyi mechanism, can add molecular hydrogen
to the unsaturated substrates in a pairwise manner is still unclear
[26]. Despite the fact that the contribution of pairwise hydrogen
addition route is often not very high (ca. 1–3%) [26], finding an
answer to this question is surely of great interest for chemical sci-
ence as well as for possible practical applications. For instance, the
production of hyperpolarized propane gas, which can be used as an
efficient contrast agent for MRI of lungs, by heterogeneous hydro-
genation of propene with p-H2 was recently demonstrated with
the use of titania-supported rhodium catalysts [23,35]. And with-
out doubt, the fundamental understanding of the mechanism of
pairwise hydrogen addition and the nature of active sites of sup-
ported metal catalysts responsible for pairwise hydrogen addition
can help to design an efficient catalytic system with the largest con-
tribution of the pairwise addition route to provide the maximum
NMR signal enhancement suitable for comprehensive MRI studies.
Originally, PHIP effects were observed for the first time in
homogeneous hydrogenations catalyzed by transition metal com-
plexes, e.g. Wilkinson’s catalyst [Rh(PPh ) Cl] [13,18], due to a
3
3
well-defined nature of an isolated catalytic center responsible for
pairwise hydrogen addition. Later on, the PHIP technique was
extensively utilized for mechanistic and kinetic studies of homoge-
neous hydrogenations catalyzed by transition metal complexes and
clusters [17,19]. However, utilization of homogeneous PHIP to pro-
duce hyperpolarized fluids suitable for biomedical MRI studies is
significantly complicated because of the presence of a toxic homo-
geneous catalyst in the reaction mixture, which cannot be easily
removed from the solution after the reaction [20]. At first sight,
this problem can be solved by the use of heterogeneous hydro-
genation. However, the conventional catalysts for heterogeneous
hydrogenation reaction are noble metal nanoparticles supported
on porous oxides or carbon. For this type of catalysts, it is gener-
It was mentioned above that TiO -supported metal catalysts
2
ally assumed that H molecules adsorb dissociatively on the metal
surface according to the established Horiuti-Polanyi mechanism
exhibit higher contribution of pairwise hydrogen addition route
and higher levels of PHIP NMR effects intensities compared to the
metal catalysts on other supports. Therefore, the goal of this paper
is to get some insight into the heterogeneous pairwise hydrogen
addition phenomenon over titania-supported catalysts. The influ-
ence of metal oxidation state on the intensities of PHIP effects was
investigated. For this purpose, a series of preliminary oxidized or
reduced titania-supported Pt, Pd, Rh and Ir catalysts were prepared,
2
[
21]. Therefore, the complete randomization of H atoms involved
in hydrogenation over the catalyst surface is expected, which
should lead to a complete loss of spin correlation between the
parahydrogen-derived protons and, consequently, to the absence
of any PHIP effects in the NMR spectra of the products. Neverthe-
less, PHIP effects were observed in heterogeneous hydrogenation
Please cite this article in press as: O.G. Salnikov, et al., The effect of oxidative and reductive treatments of titania-supported metal catalysts
on the pairwise hydrogen addition to unsaturated hydrocarbons, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.030

G Model
CATTOD-10060; No. of Pages7
ARTICLE IN PRESS
O.G. Salnikov et al. / Catalysis Today xxx (2016) xxx–xxx
3
characterized by XPS and HRTEM and tested in gas phase hydro-
genation of propene, propyne, 1,3-butadiene and 1-butyne with
H ) up to 83% using Bruker parahydrogen generator BPHG 90. The
2
hydrocarbon/H mixtures with 1:10 (in case of propyne) or 1:4 (in
2
p-H . The obtained results allowed us to make a conclusion that a
fully reduced metallic state is generally more efficient in terms of
pairwise hydrogen addition selectivity than the oxidized state.
case of propene, 1,3-butadiene and 1-butyne) ratio were prepared
in a gas cylinder.
The preliminarily reduced or oxidized (see Catalysts prepara-
tion section) Pt, Pd, Rh and Ir catalysts supported on titania were
used for the heterogeneous hydrogenation of unsaturated hydro-
carbons. The catalyst (30 mg) was placed at the bottom of a 10 mm
2
2
. Experimental
2.1. Catalysts preparation
NMR tube located inside an NMR spectrometer and maintained at
◦
1
00 C. The gas mixtures were supplied to the catalyst through a
The Pt/TiO , Pd/TiO , Rh/TiO and Ir/TiO catalysts (with 5 wt%
1/16” OD PTFE capillary. Hydrogenation was performed in a 7.05 T
magnetic field of the NMR spectrometer (the so-called PASADENA
[13] experiment) at atmospheric pressure. The gas flow rate was
varied from 1.4 to 5.7 mL/s using an Aalborg rotameter.
2
2
2
2
metal loadings) were prepared using wet impregnation method.
2
The TiO powder (Hombifine N, SBET = 107 m /g) was calcined at
2
◦
5
00 C for 2 h prior to use. In the case of the supported Pt, Pd and
1
Rh catalysts, the aqueous solutions of corresponding Pt(II), Pd(II)
and Rh(III) nitrates were used for impregnation. The preparation of
nitrates solution is described in detail elsewhere [27]. In the case
H NMR spectra of reaction mixtures in the gas phase were
acquired with 8 signal accumulations on a 300 MHz Bruker AV 300
NMR spectrometer. A single ␲/4 radiofrequency pulse maximizing
the PASADENA signal [16] was used to obtain the NMR spectra.
of the Ir/TiO catalyst, the support was impregnated with aqueous
2
H [IrCl ] solution which was prepared by dissolution of the com-
3
6
mercial IrCl3 (Sigma-Aldrich) in concentrated hydrochloric acid
with subsequent evaporation to dryness and dissolution in water.
The impregnation of TiO2 support with corresponding precursor
solutions was carried out for 1 h at room temperature. Then the sol-
2.5. Calculation of conversion values and percentages of pairwise
hydrogen addition
In the case of propene hydrogenation, two series of experiments
(with p-H2 and with normal H2) were carried out successively with
the same catalysts. It is known that the fast inflow of a fluid into
an NMR instrument results in a substantial loss of an NMR sig-
nal for nominally thermally polarized molecules because of the
insufficient residence time of the molecules in the high magnetic
field to achieve a complete relaxation of nuclear spins to the high
magnetic field equilibrium [26,37]. In the case of 1,3-butadiene
hydrogenation, this problem can be solved by the use of the char-
coal insert placed in the upper part of the NMR tube in the flow
path of the incoming gas, which provides complete relaxation of
1,3-butadiene, 1-butene, 2-butene and butane nuclear spins to the
equilibrium state at the 7.05 T field [26]. However, for propene
hydrogenation this charcoal insert is unable to provide a complete
recovery of propene and propane NMR signals [37]. Thereby, we
used the approximations obtained in the previous study [37] by
varying the flow rate of propene/H and propane/H gas mixtures
vent excess was evaporated and the obtained raw materials were
◦
dried in air at 120 C for 4 h. The obtained M/TiO samples were
2
◦
Ox
either calcined at 400 C in air for 3 h (later denoted as M /TiO2)
or reduced in H flow at 330 C for 3 h (later denoted as M /TiO2)
◦
Red
2
(
M = Pt, Pd, Rh, Ir). Therefore, a total of eight different supported
catalyst samples were prepared.
2.2. XPS experiments
The catalysts were characterized using X-ray photoelectron
spectroscopy (XPS) before and after the hydrogenation of propene,
propyne, 1,3-butadiene and 1-butyne. XPS spectra were recorded
using SPECS spectrometer with PHOIBOS-150-MCD-9 hemispheri-
cal energy analyzer (Al K␣, irradiation, hꢀ = 1486.6 eV, 200 W). The
samples were supported onto the double-sided conducting copper
scotch tape. Binding energy (BE) scale was preliminarily calibrated
by the positions of the peaks of Au 4f7/2 (BE = 84.0 eV) and Cu
2
2
in order to estimate the correct intensities of NMR signals. The
propene conversion values X were calculated from these estimates
of the fully recovered NMR signal intensities.
2
p3/2 (BE = 932.67 eV) core levels. The binding energy of peaks was
calibrated by the position of the Ti 2p peak (BE = 458.8 eV) cor-
responding to the titanium oxide as support. Pt, Pd and Rh foil,
PtO, PdO, Rh O , Ir black and IrO powders were used as reference
The percentages of pairwise H addition were calculated as fol-
2
lows. The thermally polarized NMR spectra were subtracted from
the corresponding spectra acquired with the use of parahydro-
gen. Therefore, the “pure” PASADENA spectra were obtained. The
theoretical PASADENA signal enhancement factors (SEF) ꢁ can be
calculated as [37]
2
3
2
samples. Pt 4f, Pd 3d, Rh 3d and Ir 4f regions peak fittings were
performed using XPSPeak 4.1 software [36].
2.3. HRTEM experiments
4
xp − 1 kT
ꢂꢀB0
The catalysts samples were studied by high-resolution trans-
mission electron microcopy (HRTEM) method using JEM 2010
JEOL, Japan) instrument with the accelerating voltage of 200 kV.
ꢁ =
+ 1 = 16500
3
(
where xp is parahydrogen fraction in H (0.83 in our experiments).
2
Images were recorded using Soft Imaging System (Germany) CCD.
The instrument was equipped with the energy-dispersive X-ray
However, in the case of PASADENA experiments the obtained polar-
ization is reduced because of the partial cancellation of absorption
and emission parts of each NMR signal and is thus dependent on the
NMR line width. This fact was accounted for by comparing the sim-
ulated thermal and PHIP NMR spectra with the corresponding line
width values, which allowed us to obtain the corrected theoretical
SEF values ꢃ depending on the NMR line width. The percentages of
pairwise hydrogen addition ꢄpairwise were calculated as
(
EDX) spectrometer Phoenix (EDAX, USA) with the semiconduc-
tor Si(Li) detector with an energy resolution of 130 eV. The samples
were dispersed in ethanol using an ultrasonic disperser UZD-1UCh2
and deposited on copper or molybdenum grids coated with a car-
bon hole film.
2
.4. Catalytic hydrogenation experiments
Spol
ꢃ
ꢄpairwise
=
Commercially available hydrogen, propene, 1,3-butadiene,
Spol
ꢃ
Stherm
+
n
propyne (Sigma-Aldrich, 98%) and 1-butyne (Sigma-Aldrich, ≥99%)
were used as received. For PHIP experiments, hydrogen gas was
enriched with parahydrogen from 25% (p-H2 content in normal
where Spol is the intensity of the absorption part of the “pure”
PASADENA NMR signal, Stherm is the intensity of thermally polarized
Please cite this article in press as: O.G. Salnikov, et al., The effect of oxidative and reductive treatments of titania-supported metal catalysts
on the pairwise hydrogen addition to unsaturated hydrocarbons, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.030

G Model
CATTOD-10060; No. of Pages7
ARTICLE IN PRESS
O.G. Salnikov et al. / Catalysis Today xxx (2016) xxx–xxx
4
Fig. 2. The histogram of 1H NMR signals intensities for various MRed/TiO2 and MOx/TiO2 catalysts (see the figure legend for substrates designation). NMR signals were
measured for CH2-protons of propane in propene hydrogenation, CH2-protons of butane in 1,3-butadiene hydrogenation, CH-proton of propene in propyne hydrogenation
and CH-proton of 1-butene in 1-butyne hydrogenation. Note that the maximum achieved signal intensity for each substrate was set to 100 a.u., implying that NMR signal
intensities can be compared only in the case of the same substrates. The gas flow rate was 4.1 mL/s in all of these experiments.
Table 1
The mean particle diameters for various MRed/TiO2 and MOx/TiO2 catalysts calculated from HRTEM data.
Catalyst
PtRed/TiO2
1.7
PtOx/TiO2
1.8
PdRed/TiO2
5.0
PdOx/TiO2
3.4a
RhRed/TiO2
1
RhOx/TiO2
0.6
IrRed/TiO2
1
IrOx/TiO2
0.7
Mean particle diameter (nm)
a
The particle size distribution is bimodal (ca. 1 nm and ca. 5 nm nanoparticles are observed).
found that pairwise replacement can be also provided by RhOx/TiO2
NMR signal (corrected for suppression effects caused by the rapid
gas inflow), and n is the number of corresponding protons (n = 2
for propane CH group and n = 6 for propane CH group). It should
Ox
Red
and Pt /TiO as well as by Pd /TiO (Figs. 1 and S1).
2
2
2
3
It should be noted that in the previous heterogeneous PHIP
investigations mostly 1 wt% supported metal catalysts were used
[26,28]. Recently it was shown that in the case of catalytic systems
be noted that these percentages of pairwise hydrogen addition are
largely underestimated because the relaxation of hyperpolariza-
tion as well as hydrogen ortho-para conversion were not taken
into account. However, these two factors should not be significantly
different for different catalysts due to the use of the same exper-
imental conditions and a low concentration of supported metal;
therefore, the estimated values of ꢄpairwise can be compared with
each other.
with 1 wt% metal loading, the Rh/TiO catalyst is more efficient in
2
terms of production of hyperpolarized propane gas than Pt/TiO2
and Pd/TiO catalysts [23]. In another investigation, 1 wt% Pt/TiO2
2
was found to be more efficient than Ir/TiO [28]. However, herein
2
Red
we found that for 5 wt% catalysts Ir /TiO2 allows one to achieve
PHIP NMR signals with nearly the same intensities as the cor-
Red
responding Rh /TiO2 catalyst (Fig. 1d and c). This observation,
Red
without doubt, is very important and the 5 wt% Ir /TiO2 cata-
3
. Results and discussion
lyst may represent a promising alternative for Rh/TiO [23] for the
2
production of hyperpolarized propane gas with the high level of
polarization. Thus, in order to find an optimal catalyst for produc-
tion of hyperpolarized propane gas it is highly important to vary
a many parameters of the catalytic system such as the nature of
the metal and the support, the size of metal nanoparticles and the
metal content.
Previously it was shown that titania-supported catalysts pro-
vide significantly more pronounced PHIP signals than alumina-,
silica-, zirconia- or carbon-supported catalysts [26]. Therefore, in
this study we utilized TiO -supported catalysts. The 5 wt% Pt/TiO ,
2
2
Pd/TiO , Rh/TiO and Ir/TiO catalysts were preliminarily reduced
2
2
2
or oxidized (8 samples in total) and then used in hydrogenation
of four hydrocarbons (propene, propyne, 1,3-butadiene and 1-
butyne) with parahydrogen. The catalysts were characterized by
both XPS and HRTEM before the hydrogenation reaction and by
The percentages of pairwise H addition for propene hydrogena-
2
tion over these catalysts were estimated using both propane CH2
and CH3 groups NMR signals (Table S1). Note that the obtained
values are slightly different because of the different relaxation
properties of CH2 and CH3 protons of propane molecule. It can
XPS after the reaction.
All catalysts except Ir /TiO were found to be active in propene
hydrogenation and most of them (except Ir /TiO and Pd /TiO2)
Ox
2
Red
Ox
Ox
Ox
be seen that Pt /TiO2 and Pt /TiO2 catalysts showed similar
catalytic performance both in terms of PHIP and the overall con-
2
produced pronounced PHIP NMR signals for propane when parahy-
Red
1
version of propene to propane. On the other hand, Pd /TiO2 and
Pd /TiO2 were found to be similar in conversion but very differ-
drogen was utilized (Fig. 1). Moreover, H NMR signals of propene
Ox
(
the reactant) were also polarized. This phenomenon, sometimes
ent in terms of pairwise hydrogen addition. For Rh catalysts, the
referred to as pairwise replacement, was previously observed
for propene hydrogenation over 1 wt% Rh/TiO2 [38], Pt/TiO2 and
Ir/TiO2 [28] catalysts as well as over CeO2 [30]. In this study, we
Red
Rh /TiO2 provided a ca. 3 times higher contribution of pairwise
Ox
hydrogen addition than Rh /TiO , while the latter provided signif-
2
Please cite this article in press as: O.G. Salnikov, et al., The effect of oxidative and reductive treatments of titania-supported metal catalysts
on the pairwise hydrogen addition to unsaturated hydrocarbons, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.030

G Model
CATTOD-10060; No. of Pages7
ARTICLE IN PRESS
O.G. Salnikov et al. / Catalysis Today xxx (2016) xxx–xxx
5
Fig. 3. (a) Ir 4f core-level XPS spectra of IrO2/TiO2 catalysts before and after propene hydrogenation reaction, Ir black and IrO2 were used as references; (b) Pd 3d core-level
XPS spectra of PdO/TiO2 catalysts before and after propene hydrogenation reaction, Pd foil and PdO were used as references; (c) Rh 3d core-level XPS spectra of Rh2O3/TiO2
catalysts before and after propene hydrogenation reaction, Rh foil and Rh2O3 were used as references. Binding energies are correlated to data from the XPS handbook [40].
Fig. 4. HRTEM images of the PdOx/TiO2 sample. In (a) small ca. 1 nm nanoparticles are marked with circles. In (b) 3.4 Å thick titania pedestal with a large PdO particle on top
of it is enclosed in a rectangle. The interplanar distances in PdO particles are presented in the black square. The arrows show small ca. 1 nm PdOx particles.
icantly higher conversion. For Ir catalysts, the IrOx/TiO was found
the IrOx/TiO2 catalyst. Therefore, we can conclude that under our
conditions Ir /TiO2 is practically inactive in the hydrogenation
2
Ox
to be completely inactive in propene hydrogenation. Thereby, it
may be concluded that in the case of heterogeneous hydrogena-
tion of propene to propane over titania-supported Pd, Rh and Ir
nanoparticles the oxidative treatment leads to a decrease in con-
tribution of the pairwise hydrogen addition route (even down to
zero in the case of palladium catalysts). On the contrary, maxi-
mum polarization and therefore the selectivity toward pairwise
reaction.
As was mentioned above, PdOx/TiO catalyst gave no PHIP effects
2
in propene hydrogenation despite the observation of reaction
products. The same was true in propyne hydrogenation (Fig. S2).
However, in the case of 1,3-butadiene and 1-butyne hydrogena-
Ox
tion over the Pd /TiO catalyst, the PHIP effects were observed for
2
◦
hydrogen addition are realized for a fully reduced M state. These
various reaction products though their intensities were found to
be not so high compared to the results obtained with other cata-
lysts (Figs. S3 and S4). Therefore, the nature of the substrate is also
important and may affect the pairwise hydrogen addition selectiv-
ity. In the case of propene and 1,3-butadiene hydrogenation over
observations are in agreement with our recent SMSI investigations
where formation of Pd␦ sites for Pd/TiO was found to completely
+
2
destroy the pairwise hydrogen addition reaction pathway [34].
In the case of propyne, 1,3-butadiene and 1-butyne hydrogena-
tion the PHIP results were generally similar (Figs. S2–S4). As in
the case of propene hydrogenation, most of the catalysts provided
Red
Ox
the Rh /TiO and Rh /TiO catalysts, the formation of methane
2
2
was observed (Figs. 1c and S3). However, there was no methane
in propyne and 1-butyne hydrogenation in spite of the formation
of mostly the same reaction products (Figs. S2 and S4), contrary
to some literature data [39]. These results again show the impor-
tance of the substrate nature for the catalytic performance and the
realization of different reaction pathways.
Ox
PHIP NMR signals for corresponding reaction products. Ir /TiO2
catalyst showed some negligible activity only in propyne hydro-
genation where weak PHIP NMR signals were observed for reaction
product propene (Figs. 2 and S2). In the case of 1,3-butadiene and
1
-butyne hydrogenations, no reaction products were observed for
Please cite this article in press as: O.G. Salnikov, et al., The effect of oxidative and reductive treatments of titania-supported metal catalysts
on the pairwise hydrogen addition to unsaturated hydrocarbons, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.030

G Model
CATTOD-10060; No. of Pages7
ARTICLE IN PRESS
O.G. Salnikov et al. / Catalysis Today xxx (2016) xxx–xxx
6
With the aim to verify the structure of active sites, the cata-
lysts were characterized using XPS and HRTEM methods. X-ray
photoelectron spectra were acquired ex situ before and after hydro-
genation reactions. In most of the preliminarily reduced catalysts
diameter were observed. On the other hand, changes in anatase lat-
tice distances and observation of palladium peaks in EDX spectra
from these areas can be seen, meaning the occurrence of Pd clus-
Ox
ters on the surface or in the bulk of TiO . For Pd /TiO2 sample a
2
Red
Red
Red
(
Pt /TiO , Pd /TiO2 and Rh /TiO ) the corresponding metals
bimodal particle size distribution was obtained. Besides relatively
large nanoparticles 5–7 nm in diameter, small particles with ca.
1 nm size were observed on the anatase surface (Fig. 4a). According
to interplanar distances these particles cannot be attributed to pure
PdO or Pd, probably because of the partial reduction under electron
2
2
◦
◦
were found to be in predominantly metallic states (Pt , Pd and
Rh , respectively). However, in the case of the Ir /TiO2 catalyst
the starting material was found to contain both Ir and Ir in a
2
◦
Red
◦
4+
◦
:1 ratio (Fig. 3a). Upon propene hydrogenation the Ir content
◦
4+
␦+
increased only slightly (the Ir /Ir ratio became 2.4:1).
beam to a Pd state (Fig. 4b). Another interesting fact is the for-
In the case of preliminarily oxidized catalysts the diverse results
mation of something like a pedestal from TiO2 under larger PdO
particles (Fig. 4b). The interplanar distances, both in pedestals and
these PdO nanoparticles, are slightly different from the standard
values.
Ox
were obtained for each system. For the Pd /TiO catalyst the corre-
2
2+
sponding metal was found to be in a pure oxide Pd state (Fig. 3b).
Ox
However, the XPS spectrum of the same Pd /TiO catalyst after
2
◦
As was mentioned above, in the PtRed/TiO and Pt /TiO sam-
Ox
propene hydrogenation clearly exhibited a Pd peak with corre-
sponding 335.2 eV binding energy [40]. The Pd /Pd ratio in this
2 2
◦
2+
ples the metal was found to be in the pure Pt and Pt2+ states,
respectively. However, the metal dispersion and the catalytic per-
formance (including the selectivity toward pairwise hydrogen
addition) of these catalysts were similar. Therefore, in the case of Pt
the metal oxidation state has no significant impact on the catalytic
◦
Ox
material was found to be 1.1:1. Therefore, though the Pd /TiO cat-
2
alyst originally contained pure Pd2+, it is partially reduced to Pd
◦
in the course of the reaction. Similar results were obtained for the
Pt/TiO2 supported catalysts. However, for the RhOx/TiO2 catalyst
the slightly different situation was found (Fig. 3c). According to XPS,
Ox
performance, although the Pt /TiO2 catalyst is partially reduced
◦
3+
the starting material contained both Rh and Rh with the preva-
under hydrogenation conditions. For Rh catalysts the results are
◦
3+
Red
lence of the latter (the Rh /Rh ratio 0.5:1). However, it should be
mentioned that during 2 h of XPS measurements the decrease in
quite similar with those for Pt, although Rh /TiO was found to be
2
slightly more efficient in terms of pairwise hydrogen addition than
3+
◦
Ox
3+
Rh and increase in Rh peaks intensities were observed, implying
that Rh oxide is gradually reduced under X-ray irradiation. Thus, it
could be suggested that despite the use of a fast (just a few scans)
recording of Rh 3d region for the XPS spectrum presented in Fig. 3c,
the metal still undergoes partial reduction. Therefore, it is quite
likely that initially Rh was in the oxide state (Rh3+). Utilization of
the catalyst sample in propene hydrogenation again led to a partial
Rh /TiO . This can be explained by the fact that Rh is significantly
2
◦
3+
reduced to Rh during the reaction (according to XPS data, the Rh
fraction decreased from 67% to 17%). Therefore, it can be concluded
◦
that on the Rh the contribution of pairwise hydrogen addition
3+
route is higher than on Rh . On the other hand, the catalytic per-
formances of Ir /TiO2 and IrOx/TiO2 catalysts in heterogeneous
hydrogenations were found to be dramatically different despite the
fact that no significant alterations in metal nanoparticles size were
observed. Therefore, these differences in activities correlate with
Red
3
+
◦
◦
3+
reduction of Rh to Rh because the Rh /Rh ratio was found to
increase up to 4.9:1. On the other hand, in the Ir /TiO2 material
Ox
4+
Ox
4+
the metal was found to be in a pure Ir state, both before and after
the hydrogenation reaction. This is surely anticipated because as
was mentioned above, this catalyst showed no activity in propene
hydrogenation. Therefore, it is expected that the Ir state remains
unchanged during the exposure to propene/H2 mixture. Based on
XPS data it is easy to verify that partial reduction of metal oxide to
metallic state occurs during hydrogenation reaction but this reduc-
tion is not enough for the formation of pure metal. Therefore, as it
was mentioned above, the presence of oxidized state decreases the
selectivity toward the pairwise hydrogen addition route.
Ir oxidation state. In Ir /TiO sample all metal is in the Ir state.
2
Red
◦
◦
On the other hand, in catalytically active Ir /TiO sample both Ir
2
and Ir4+ states were observed. Therefore, it is possible that only Ir
4+
4+
but not Ir can catalyze hydrogenation of hydrocarbons as well as
be responsible for pairwise hydrogen addition. As for the proper-
ties of Pd /TiO2 and Pd /TiO2 samples, they were found to be
strongly different. First of all, the metal in these two samples was
Red
Ox
◦
2+
initially in different oxidation states (Pd and Pd , respectively),
although partial reduction of Pd2+ to Pd during hydrogenation was
◦
Red
detected. Moreover, while in the case of Pd /TiO2 the particles
size was ca. 5.0 nm, the Pd /TiO2 catalyst contained both large
Ox
Next, the HRTEM investigations were performed. The mean par-
ticles sizes of all catalysts are presented in Table 1. It was found that
for Ir/TiO and Pt/TiO samples the oxidation and reduction proce-
(ca. 5–7 nm) and small (ca. 1 nm) particles located predominantly
at the facets of TiO2 support. Also for the first time the HRTEM
2
2
dures did not lead to significant changes in particle size. On the
revealed the formation of TiO pedestals under Pd particles. As for
2
Red
Ox
other hand, HRTEM revealed that though in Pt /TiO2 all Pt was
the catalytic performance, the Pd /TiO catalyst exhibited signifi-
2
◦
Ox
found to be in Pt state, in Pt /TiO the metal was found to be in
cantly lower efficiency in terms of pairwise hydrogen addition (no
PHIP effects in propene and propyne hydrogenation and less inten-
sive PHIP effects in 1,3-butadiene and 1-butyne hydrogenation).
Our hypothesis is that the reason for this is the formation of Pd␦
sites which are fully inconsistent with pairwise hydrogen addition.
Without doubt, the change in interplanar distances, the formation
of Pd clusters, the modification of electronic structure of PdO par-
2
both metallic and oxide forms, although reduction of Pt under the
electronic beam during HRTEM measurements cannot be excluded.
For Rh/TiO2 catalysts the mean particles diameter was found to
+
slightly decrease from 1 to ca. 0.6 nm upon oxidation, although for
Ox
Rh /TiO the exact estimation of particle size is somewhat compli-
2
cated because of their non-spherical shape. Interestingly, the mean
Red
Red
◦
particle sizes for 5 wt% Rh /TiO2 and Ir /TiO2 catalysts, which
were found to be more efficient for production of hyperpolarized
propane than other samples used in this study, are quite similar to
ticles via pedestals formation allow to block Pd sites which are
required for the pairwise hydrogen addition.
the previously reported 1 wt% Rh/TiO catalyst (ca. 1–1.5 nm) [23].
2
Red
Another interesting fact is that for the Rh /TiO2 sample part of
the active component was found to be in the Rh state accord-
4. Conclusions
3+
ing to interplanar distances. According to HRTEM the TiO surface
was partly disordered, probably because of the interaction between
the active component and the support and/or reduction of the sup-
In this study, titania-supported Pt, Pd, Rh and Ir catalysts were
utilized for hydrogenation of unsaturated hydrocarbons (propene,
propyne, 1,3-butadiene and 1-butyne) with parahydrogen. It was
shown that the oxidation or reduction of the catalysts impact signif-
icantly on the overall catalytic activity and, more importantly, on
2
port surface. For the Pd/TiO2 systems HRTEM allowed to obtain
Red
some unique results. For Pd /TiO the particles with 5 nm mean
2
Please cite this article in press as: O.G. Salnikov, et al., The effect of oxidative and reductive treatments of titania-supported metal catalysts
on the pairwise hydrogen addition to unsaturated hydrocarbons, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.030

G Model
CATTOD-10060; No. of Pages7
ARTICLE IN PRESS
O.G. Salnikov et al. / Catalysis Today xxx (2016) xxx–xxx
7
the selectivity toward pairwise hydrogen addition to a substrate,
i.e., the ability to incorporate both H atoms of an H2 molecule
into the same product molecule. Generally, for M the contribu-
[7] A.A. Lysova, I.V. Koptyug, Chem. Soc. Rev. 39 (2010) 4585–4601.
[
[
8] L.F. Gladden, Curr. Opin. Chem. Eng. 2 (2013) 331–337.
9] P. Nikolaou, B.M. Goodson, E.Y. Chekmenev, Chem. Eur. J. 21 (2015)
◦
3156–3166.
tion of pairwise hydrogen addition route was found to be higher
[10] J.-H. Ardenkjaer-Larsen, G.S. Boebinger, A. Comment, S. Duckett, A.S. Edison, F.
Engelke, C. Griesinger, R.G. Griffin, C. Hilty, H. Maeda, G. Parigi, T. Prisner, E.
Ravera, J. van Bentum, S. Vega, A. Webb, C. Luchinat, H. Schwalbe, L. Frydman,
Angew. Chem. Int. Ed. 54 (2015) 9162–9185.
[11] R.G. Griffin, T.F. Prisner, Phys. Chem. Chem. Phys. 12 (2010) 5737–5740.
[12] B.M. Goodson, J. Magn. Reson. 155 (2002) 157–216.
n+
than for M in the case of titania-supported catalytic systems.
␦
+
For instance, formation of Pd sites or insufficient reduction of
2+
◦
Pd to Pd leads to a decrease in the PHIP signals intensities
and therefore, to a lower contribution of pairwise hydrogen addi-
tion. The formation of titania pedestals with PdO particles on their
top was observed by HRTEM for the first time for Pd/TiO2 cata-
[
[
13] C.R. Bowers, D.P. Weitekamp, J. Am. Chem. Soc. 109 (1987) 5541–5542.
14] R.A. Green, R.W. Adams, S.B. Duckett, R.E. Mewis, D.C. Williamson, G.G.R.
Green, Prog. Nucl. Magn. Reson. Spectrosc. 67 (2012) 1–48.
◦
lyst. The 5 wt% Ir /TiO catalyst was found to demonstrate nearly
[15] C.R. Bowers, D.P. Weitekamp, Phys. Rev. Lett. 57 (1986) 2645–2648.
2
[
16] C.R. Bowers, Sensitivity enhancement utilizing parahydrogen, in: D.M. Grant,
R.K. Harris (Eds.), Encycl. Magn. Reson., John Wiley & Sons, Ltd., Chichester,
2002, pp. 750–770.
the same efficiency for production of hyperpolarized propane as
the corresponding Rh/TiO2 catalyst. Thereby, it is a good alterna-
tive to Rh/TiO , which was earlier shown to be the most efficient
[17] D. Blazina, S.B. Duckett, J.P. Dunne, C. Godard, Dalton Trans. (2004)
601–2609.
2
2
for producing PHIP effects in this reaction. Although the pairwise
hydrogen addition route contributes only a few percent to the
overall hydrogenation reaction mechanism, the improvement of
the catalysts preparation procedure and the identification of the
structure of active sites responsible for pairwise hydrogen addition
can in perspective further improve the selectivity toward pairwise
route and therefore, significantly enhance NMR and MRI sensitivity.
[
18] T.C. Eisenschmid, R.U. Kirss, P.P. Deutsch, S.I. Hommeltoft, R. Eisenberg, J.
Bargon, R.G. Lawler, A.L. Balch, J. Am. Chem. Soc. 109 (1987) 8089–8091.
[19] L. Buljubasich, M.B. Franzoni, K. Munnemann, Parahydrogen induced
polarization by homogeneous catalysis: theory and applications, in: L.T. Kuhn
(
Ed.), Hyperpolarization Methods NMR Spectrosc, Springer-Verlag, Berlin,
013, pp. 33–74.
[20] S. Glöggler, J. Colell, S. Appelt, J. Magn. Reson. 235 (2013) 130–142.
2
[
[
21] I. Horiuti, M. Polanyi, Trans. Faraday Soc. 30 (1934) 1164–1172.
22] K.V. Kovtunov, I.E. Beck, V.I. Bukhtiyarov, I.V. Koptyug, Angew. Chem. Int. Ed.
4
7 (2008) 1492–1495.
Acknowledgments
[23] K.V. Kovtunov, D.A. Barskiy, A.M. Coffey, M.L. Truong, O.G. Salnikov, A.K.
Khudorozhkov, E.A. Inozemtseva, I.P. Prosvirin, V.I. Bukhtiyarov, K.W.
Waddell, E.Y. Chekmenev, I.V. Koptyug, Chem.—A Eur. J. 20 (2014)
OGS, DBB, DAB, KVK, and IVK thank RSCF (grant #14-13-
1
1636–11639.
00445) for supporting the catalytic hydrogenation experiments
[24] S.S. Arzumanov, A.G. Stepanov, J. Phys. Chem. C. 117 (2013) 2888–2892.
[25] R. Zhou, W. Cheng, L.M. Neal, E.W. Zhao, K. Ludden, H.E. Hagelin-Weaver, C.R.
Bowers, Phys. Chem. Chem. Phys. 17 (2015) 26121–26129.
26] K.V. Kovtunov, V.V. Zhivonitko, I.V. Skovpin, D.A. Barskiy, I.V. Koptyug,
Parahydrogen-Induced polarization in heterogeneous catalytic processes, in:
L.T. Kuhn (Ed.), Hyperpolarization Methods NMR Spectrosc., Springer-Verlag,
Berlin, 2013, pp. 123–180.
using parahydrogen and the high-field NMR studies. IPP and EYG
thank RFBR (grants ## 14-03-00374-ɑ and 14-03-93183 МCХ ɑ)
for the support of XPS and HRTEM studies. VIB, AKK, LMK and
AVB thank RSCF (grant #14-23-00146) for supporting the catalysts
preparation. The basic funding of the SB RAS institutes was provided
by FASO.
[
[
27] O.G. Salnikov, K.V. Kovtunov, D.A. Barskiy, A.K. Khudorozhkov, E.A.
Inozemtseva, I.P. Prosvirin, V.I. Bukhtiyarov, I.V. Koptyug, ACS Catal. 4 (2014)
2022–2028.
[
[
[
28] R. Zhou, E.W. Zhao, W. Cheng, L.M. Neal, H. Zheng, R.E. Qui n˜ ones, H.E.
Hagelin-Weaver, C.R. Bowers, J. Am. Chem. Soc. 137 (2015) 1938–1946.
29] K.V. Kovtunov, D.A. Barskiy, O.G. Salnikov, A.K. Khudorozhkov, V.I.
Bukhtiyarov, I.P. Prosvirin, I.V. Koptyug, Chem. Commun. 50 (2014) 875–878.
30] E.W. Zhao, H. Zheng, R. Zhou, H.E. Hagelin-Weaver, C.R. Bowers, Angew.
Chem. Int. Ed. 54 (2015) 14270–14275.
31] A.M. Balu, S.B. Duckett, R. Luque, Dalton Trans. (2009) 5074–5076.
32] O.G. Salnikov, K.V. Kovtunov, I.V. Koptyug, Sci. Rep. 5 (2015) 13930.
33] S. Gloggler, A.M. Grunfeld, Y.N. Ertas, J. McCormick, S. Wagner, P.P.M.
Schleker, L.-S. Bouchard, Angew. Chem. Int. Ed. 54 (2015) 2452–2456.
34] K.V. Kovtunov, D.A. Barskiy, O.G. Salnikov, D.B. Burueva, A.K. Khudorozhkov,
A.V. Bukhtiyarov, I.P. Prosvirin, E.Y. Gerasimov, V.I. Bukhtiyarov, I.V. Koptyug,
ChemCatChem 7 (2015) 2581–2584.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.cattod.2016.02.
030.
[
[
[
References
[
[
[
[
1] B.M. Weckhuysen, Angew. Chem. Int. Ed. 48 (2009) 4910–4943.
2] S. Mitchell, N.-L. Michels, G. Majano, J. Pérez-Ramírez, Curr. Opin. Chem. Eng.
2
(2013) 304–311.
3] J.C.J. Camp, M.D. Mantle, A.P.E. York, J. McGregor, Rev. Sci. Instrum. 85 (2014)
63111.
4] T. Titze, C. Chmelik, J. Kullmann, L. Prager, E. Miersemann, R. Gläser, D. Enke, J.
Weitkamp, J. Kärger, Angew. Chem. Int. Ed. 54 (2015) 5060–5064.
5] C. Sievers, S.R. Bare, E. Stavitski, The 4th international congress on operando
spectroscopy, Catal. Today 205 (2013).
6] I.V. Koptyug, Magnetic resonance imaging methods in heterogeneous
catalysis, in: R. Douthwaite, S. Duckett, J. Yarwood (Eds.), Spectrosc. Prop.
Inorg. Organomet. Compd., Royal Society of Chemistry, Cambridge, 2014, pp.
35] K.V. Kovtunov, M.L. Truong, D.A. Barskiy, I.V. Koptyug, A.M. Coffey, K.W.
Waddell, E.Y. Chekmenev, Chem. —A Eur. J. 20 (2014) 14629–14632.
36] http://xpspeak.software.informer.com/4.1/.
[
[
[
[
0
[
[
37] D.A. Barskiy, O.G. Salnikov, K.V. Kovtunov, I.V. Koptyug, J. Phys. Chem. A 119
(
2015) 996–1006.
[
38] K.V. Kovtunov, M.L. Truong, D.A. Barskiy, O.G. Salnikov, V.I. Bukhtiyarov, A.M.
Coffey, K.W. Waddell, I.V. Koptyug, E.Y. Chekmenev, J. Phys. Chem. C 118
(
2014) 28234–28243.
[
39] B.J. Brandreth, S.D. Jackson, D. Winstanley, J. Catal. 102 (1986) 433–442.
40] J.F. Moulder, W.F. Stickle, P.E. Sobol, K. Bomben, Handbook of X-ray
Photoelectron Spectroscopy, 2nd ed., Perkin-Elmer Corporation, Eden Priarie,
MN, 1992.
[
1
–42.
Please cite this article in press as: O.G. Salnikov, et al., The effect of oxidative and reductive treatments of titania-supported metal catalysts
on the pairwise hydrogen addition to unsaturated hydrocarbons, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.030