Carbon supported gold and silver: Application in the gas phase hydrogenation of m-dinitrobenzene

Article abstract of DOI:10.1016/j.molcata.2015.07.019

Abstract We have studied the gas phase continuous hydrogenation of m-dinitrobenzene (m-DNB) over acid treated activated carbon (AC) supported Au and Ag prepared by deposition-precipitation. Temperature programmed reduction of a 1%wt. metal loading generat

Full text of DOI:10.1016/j.molcata.2015.07.019

Journal of Molecular Catalysis A: Chemical 408 (2015) 138–146
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Carbon supported gold and silver: Application in the gas phase
hydrogenation of m-dinitrobenzene
Fernando Cárdenas-Lizanaa, Zahara M. De Pedroa,1, Santiago Gómez-Queroa,2
,
Lioubov Kiwi-Minskerb, Mark A. Keanea,∗
a Chemical Engineering, School of Engineering and Physical Sciences, Heriot Watt University, Edinburgh EH14 4AS Scotland, UK
b Group of Catalytic Reaction Engineering, Ecole Polytechnique Fédérale de Lausanne (GGRC-ISIC-EPFL), Lausanne CH-1015, Switzerland
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 10 June 2015
Received in revised form 13 July 2015
Accepted 14 July 2015
Available online 23 July 2015
We have studied the gas phase continuous hydrogenation of m-dinitrobenzene (m-DNB) over acid treated
activated carbon (AC) supported Au and Ag prepared by deposition-precipitation. Temperature pro-
grammed reduction of a 1%wt. metal loading generated a broad distribution of Au (mean = 45 nm) and
Ag (mean = 13 nm) nanoparticles. A decrease in metal content to 0.1%wt. served to increase dispersion
(Au/AC (mean = 4 nm) and Ag/AC (mean = 3 nm)) where both catalysts exhibited an electron enriched
(from XPS) metal phase. m-DNB hydrogenation activity was stable with time on-stream where partial
hydrogenation (to m-nitroaniline (m-NAN)) was favoured at low conversion with a switch to full hydro-
genation (to m-phenylenediamine (m-PDM)) at higher conversions consistent with a stepwise reaction
mechanism. Reaction over 0.1%wt. Au/AC delivered a greater than 5-fold higher turnover frequency (TOF)
relative to Ag/AC that can be attributed to a higher H2 chemisorption capacity under reaction conditions.
Activation energy for m-DNB → m-NAN was the same (110 kJ mol−1) for both catalysts. We have estab-
lished differences in the conversion/selectivity response where m-NAN formation was enhanced over
Ag/AC at the same conversion, which is ascribed to differences in m-DNB adsorption/activation.
© 2015 Elsevier B.V. All rights reserved.
Keywords:
Selective gas phase hydrogenation
m-Dinitrobenzene
Functionalised activated carbon
Au/AC
Ag/AC
1. Introduction
offers a number of advantages over oxide supports, notably cost
effectiveness, stability in acid and base media and the possibility
The selective hydrogenation of nitroarenes to aromatic amines
continues to attract significant research due to its industrial rel-
evance [1]. Exclusive -NO2 group reduction is difficult to achieve
in the presence of other reactive functionalities such as carbonyl,
cyano, alkenyl or halogen groups [2]. The focus has moved from
conventional Fe promoted reduction in acid media (Béchamp pro-
cess) [3] to the application of transition metal (e.g. Pd [4], Pt [5])
catalysts. In previous work, we demonstrated the potential of oxide
supported Au and Ag to promote the chemoselective reduction of
funtionalised nitroarenes [6] and established a nucleophilic reac-
tion mechanism [7]. Oxides have been the most widely used Au
and Ag carriers where support surface chemistry (e.g. Lewis acidity-
basicity) can impact on hydrogenation performance [8,9]. Carbon
of adjusting porosity, specific surface area (SSA > 600 m2 g−1) and
surface chemistry for specific catalytic applications [10]. The appli-
cation of carbon supported Au or Ag catalysts in hydrogen mediated
reactions has been explored to a limited extent. Published work has
primarily dealt with Pd-containing bimetallics where Au [11–13]
and Ag [14] were shown to influence selectivity in liquid [11,12,14]
and gas [13] phase conversion of cinnamaldehyde [11,12], car-
bon tetrachloride [14] and tetrachloromethane [13]. We should
also note literature that has reported enhanced activity and selec-
tivity in hydrogenation (1,3-butadiene [15], cinnamaldehyde [15],
o-, m- and p-chloronitrobenzenes and nitrobenzaldehydes [16]),
hydrodechlorination (carbon tetrachloride [17]) and hydrodeoxy-
genation (vanillin [18]) over (carbon nanotube [15,18] and C60 [16])
supported Au [15,18] and Ag [16,17].
We could not find any publication that has provided a direct
comparison of the catalytic performance of carbon supported Au
and Ag. This may be due, in part, to difficulties in preparing well
dispersed (≤10 nm) Au and Ag particles on carbonaceous carriers
[19], which is critical for significant hydrogenation activity [2,6].
The catalytic response in the production of functionalised anilines
∗
Corresponding author.
E-mail address: M.A.Keane@hw.ac.uk (M.A. Keane).
Current address: Ingeniería Química, Facultad de Ciencias, Universidad
1
Autónoma de Madrid, Cantoblanco, 28049 Madrid, Spain.
2
Current address: Avantium, Zekeringstraat 29, Amsterdam 1014 BV, The
Netherlands.
http://dx.doi.org/10.1016/j.molcata.2015.07.019
1381-1169/© 2015 Elsevier B.V. All rights reserved.

F. Cárdenas-Lizana et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 138–146
139
ance was within 5%. The degree of nitro-group reduction (X
is given by
)
−NO
2
[−NH2]out
[−NO2]in
2 × [m − PDM]out + [m − NAN]out
2 × [m − DNM]in
X
−NO
=
=
(1)
2
where [m-DNB], [m-NAN] and [m-PDM] are, respectively, the
concentrations of m-dinitrobenzene, m-nitroaniline and m-
phenylenediamine; subscripts in and out refer to inlet and outlet
streams. Reactant fractional conversion (Xm-DNB) is defined by
[m − DNB]in − [m − DNB]out
Xm−DNB
=
(2)
[m − DNB]in
and selectivity in terms of (for example) m-NAN (Sm-NAN) is
given by
[m − NAN]out
Sm−NAN
=
(3)
[m − DNB]in − [m − DNB]out
Repeated reactions with the same batch of catalyst delivered
conversion/selectivity values that were reproducibility to better
than 6%.
Scheme 1. Reaction pathways in the partial hydrogenation of m-dinitrobenzene
(m-DNB) to m-nitroaniline (m-NAN, solid line) and full hydrogenation to m-
phenylenediamine (m-PDM, dashed lines).
2.2. Catalyst preparation and activation
The AC support was obtained from NORIT (UK) and subjected
to demineralisation by treatment with HF (100 cm3 g−1). The sup-
port was then refluxed in HNO3 + H2SO4 (1:1 v/v, 20 cm3 g−1) for
1 h to introduce surface oxygen groups [24]. Carbon supported Au
and Ag (1% and 0.1%wt.) were prepared by deposition-precipitation
using urea as basification agent which has been shown to generate
nano-sized (<5 nm) Au [4] and Ag [27]. Aqueous urea (100 times
excess) and HAuCl4 (5 × 10−4 M, 300 cm3) or AgNO3 (5 × 10−4 M,
300 cm3) was added to the AC support (5 g). The pH was con-
tinuously monitored using a crystal-body pH electrode (Hanna
Instruments) coupled to a data collection system (Pico Technology
Ltd.). The suspension was heated to 353 K under constant agitation
(300 rpm) with a pH increase due to the thermal decomposition
of urea [28]. The solid catalyst obtained was washed repeatedly
with deionized water until chlorine free (based on the AgNO3 test)
and dried in He (60 cm3 min−1) at 383 K for 3 h. Samples were
sieved to 75 ␮m average particle diameter (ATM fine test sieves)
and stored at 277 K in the dark. Catalyst activation by temper-
ature programmed reduction (TPR) employed a 2 K min−1 ramp
to 623 K in 60 cm3 min−1 H2. The temperature requirements for
reduction of supported Au and Ag precursors have been discussed
elsewhere [7,29]. Post-activation the samples were cooled to ambi-
ent temperature and passivated in 1% v/v O2/He for 1 h for ex situ
characterisation.
can be controlled by contributions due to metal size and elec-
tronic character which, in turn, are dependent on metal loading
[20], method of synthesis and support interactions [21]. Carbon
carriers are intrinsically hydrophobic and hydrophilic properties
are required to increase surface wettability by polar solvents (e.g.
water) and facilitate homogeneous distribution of supported metal
precursor(s) [22]. Surface oxygen-containing functionalities (i) act
as metal-anchoring sites [22], (ii) promote reduction of the metal
precursor [23] and (iii) lower hydrophobicity and improve support
accessibility during catalyst synthesis [10]. The concentration and
nature of surface groups can be tailored by treatment with oxidis-
ing agents (HNO3, KMnO4, H2SO4/HNO3, (NH4)2S2O8, H2O2, and
O3) [24].
In this work, we compare the performance of activated car-
bon (AC) supported Au and Ag in the gas phase hydrogenation of
m-dinitrobenzene (m-DNB). Products of partial (m-nitroaniline (m-
NAN)) and complete (m-phenylenediamine (m-PDM)) reduction
(see Scheme 1) find multiple use in the manufacture of polymers
and dyes [25,26] where selective formation of m-NAN or m-PDM is
challenging. We address the feasibility of controlling product dis-
tribution by modifying the size and electronic properties of the
supported metal phase through variations in metal loading and
support interactions.
2.3. Support and catalyst characterisation
2. Experimental
The (Au and Ag) metal loading was determined by inductively
coupled plasma-optical emission spectrometry (Vista-PRO, Varian
Inc.) from the diluted extract of aqua regia. Acid/base titrations
were performed by immersing 25 mg of sample in 50 cm3 solution
of 0.1 M NaCl and 0.1 mM oxalic acid, acidified to pH 3 with HCl
(0.1 M) with constant stirring in a He atmosphere. A 0.1 M NaOH
solution was used as titrant, added dropwise (3 cm3 h−1) using
a 100 Kd Scientific microprocessor-controlled infusion pump and
the pH monitored as above. Acidity was quantified as the volume
of NaOH required to neutralise surface groups where pKi ≤ 7 [30].
Micro-Raman spectra of the support were recorded with a Ren-
ishaw Raman Microscope System RM1000 equipped with a Leica
microscope, an electrically refrigerated CCD camera and a diode
laser at 514 nm as excitation source operating at a power level of
3 mW. Total SSA and pore volume were determined by N2 adsorp-
tion at 77 K using the Sorptomatic 1990 (Carlo Erba) system. Prior
2.1. Materials and analytical methods
The m-DNB reactant (Sigma–Aldrich, ≥98%) and solvent (1-
butanol, Riedel-de Häen, ≥99.5%) were used as supplied, without
further purification. All gases (H2, N2, Ar, O2 and He) were of
ultra-high purity (BOC, >99.99%). The composition of the reac-
tion/product mixtures was determined using a PerkinElmer Auto
System XL gas chromatograph (GC) equipped with a programmed
split/splitless injector and a flame ionisation detector, employing a
DB-1 50 m × 0.20 mm i.d., 0.33 ␮m film thickness capillary column
(J&W Scientific). Data acquisition and manipulation were per-
formed using the TotalChrom Workstation chromatography data
system. Reactant and product molar fractions (xi) were obtained
from detailed calibration plots (not shown) and carbon mass bal-

140
F. Cárdenas-Lizana et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 138–146
to analysis, samples were outgassed at 523 K for 2 h under vacuum
(<5 × 10−2 Torr). Total pore volume was obtained from BJH analysis
of the desorption isotherm. Hydrogen chemisorp−ti1on (pulse 10 ␮l
ground glass (75 ␮m). Temperature was continuously monitored
by a thermocouple inserted in a thermowell within the catalyst
bed. m-DNB was delivered as butanolic solutions in a co-current
flow of H2 via a glass/teflon air-tight syringe and teflon line using
a microprocessor controlled infusion pump (Model 100 kd Scien-
tific) at a fixed calibrated flow rate. Molar metal (n) to inlet –NO2
feed rate ratio (F) spanned the range 1 × 10−4–1 × 10−2 h. Hydrogen
was far in excess of the stoichiometric requirements for the produc-
tion of m-NAN and the flow rate was monitored using a Humonics
titration at 523 K) following TPR (in 17 cm3 min
(Brooks mass
flow controlled) 5% v/v H2/N2 to 623 K at 2 K min−1) and temper-
ature programmed desorption (TPD in 65 cm3 min−1 N2 to 873 K
at 50 K min−1) were conducted on the CHEM-BET 3000 (Quan-
tachrome Instrument) unit with data acquisition/manipulation
using the TPR WinTM software [20]. In a blank test, there was
no measurable H2 uptake on the AC support. Combined TCD cal-
ibrations with analysis of the effluent gas using a MICROMASS
PC Residual Gas Analyzer confirmed that the TPD profiles can be
ascribed solely to H2 release. Powder X-ray diffractograms were
recorded on a Bruker/Siemens D500 incident X-ray diffractometer
using Cu K␣ radiation. Samples were scanned at 0.02◦ step−1 and
the diffractograms identified using the JCPDS-ICDD reference stan-
(Model 520) digital flowmeter; GHSV = 2 × 104
h
−1. Passage of m-
DNB in H2 through the empty reactor or over AC did not result in
any detectable conversion.
3. Results and discussion
3.1. Sample characterisation
dards (Au (04-0784) and Ag (04-0783)). Metal particle size (dhkl
)
was estimated using the Scherrer equation:
3.1.1. Activated carbon (AC) support
K × ꢀ
Demineralisation of the AC support (in HF) served to remove
residual metal impurities that could impact on the catalytic
response [31]. Subsequent acid treatment (in HNO3 + H2SO4) was
used to introduce oxygen containing groups as nucleation sites for
the deposition of Au and Ag [22]. A displacement of the acid-base
titration curves for AC pre- and post-acid treatment (not shown)
confirmed an increase in surface acidity due to the generation of
carbonyl (quinone and lactone) and carboxyl (acid and anhydride)
functionalities [32]. Acid-base titration delivered a pH at the equiv-
alency point (pKa) of 6.4, which suggests the absence of strong
carboxylic acid sites (pKa < 5) and the formation of surface quinone,
lactone and anhydride groups [33]. The O 1s XPS signal provides
information on surface functionalisation of carbon substrates [34];
the spectrum for acid treated AC is presented in Fig. 1(I).
After deconvolution, two near equivalent signals with
BE = 531.2 eV and 533.3 eV are in evidence and can be attributed to
carbonyl (quinone (531.1 eV) [35] and lactone (533.3 eV) [35]) and
carboxyl anhydride (531.1 eV) [35] groups. The presence of these
functionalities is consistent with the change in surface acidity
inferred from titration. Activated carbon is characterised by a
random arrangement of “basal planes”, i.e. layers of interlocking
aromatic rings [36]. The scanning electron microscopy images of
acid treated AC (Fig. 1(II)) show heterogeneity in size (10–100 ␮m)
and shape.
dhkl
=
(4)
ˇ × cosꢁ
where K = 0.9,
ꢀ
is the incident radiation wavelength
(1.54056 Å), ␤ is the peak width at half the maximum intensity
and ꢁ represents the diffraction angle (2ꢁ = 38.1◦) corresponding
to the (1 1 1) plane for Au and Ag. Scanning electron microscopy
(SEM) was conducted on a Philips FEI XL30-FEG with an Everhart-
Thornley secondary-electron detector operated at an accelerating
voltage of 10 kV and NORAN System SIX (version 1.6) for data
analysis. The samples were subjected to decontamination using a
plasma-cleaner (EVACTRON). Transmission electron microscopy
(TEM) analysis was performed using a JEOL JEM 2011HRTEM unit
with UTW energy dispersive X-ray detector (Oxford Instruments)
at an accelerating voltage of 200 kV; data acquisition/manipulation
employed Gatan DigitalMicrograph 3.4. Specimens were prepared
by dispersion in acetone and deposited on a holey carbon/Cu grid
(300 Mesh). Up to 1000 individual metal particles were counted
for each catalyst and the surface area-weighted metal diameter (d)
calculated from
n d3
ꢂ
i
i
i
d =
(5)
n d2
ꢂ
i
i
i
where ni is the number of metal particles of diameter di. Analysis
by X-ray photoelectron spectroscopy (XPS) employed an Axis Ultra
instrument (Kratos Analytical) under ultra-high vacuum condition
(<10−8 Torr) with a monochromatic Al K␣ X-ray source (1486.6 eV).
The source power was maintained at 150 W and emitted photo-
electrons were sampled from a 750 × 350 ␮m2 area at a take-off
angle = 90◦. The analyser pass energy was 80 eV for survey spec-
tra (0–1000 eV) and 40 eV for high resolution spectra (O 1s, Au
4f5/2, Au 4f7/2, Ag 3d3/2 and Ag 3d5/2). The adventitious carbon 1s
peak at 284.5 eV was used as internal standard to compensate for
charging effects. Spectra curve fitting and quantification were per-
formed with the CasaXPS software using relative sensitivity factors
provided by Kratos.
Raman spectroscopy measurement generated
a
profile
(Fig. 1(III)) with two peaks attributable to the D- (1354 cm−1
,
defects and/or curvature [37,38]) and G- (1600 cm−1, ordered
structure [39]) bands for carbonaceous materials. The relative
intensity of these bands (ID/IG) serves as an index to evaluate
graphitic character where the measured ID/IG (=0.90) is within the
range (0.73–1.26) reported for AC [40,41], higher than graphite
(0.61 [40]) and consistent with an amorphous structure. Nitrogen
adsorption/desorption isotherms are presented in Fig. 1(IV) where
the associated SSA and total pore volume (Table 1) are in accord
with values in the literature (349–884 m2 g−1; 0.19–0.58 cm3 g−1
)
[40,42]. The sharp increase in N2 adsorption at low relative pres-
sures (P/P0 ≤ 0.15) and lesser uptake at higher P/P0 suggest a type
I IUPAC classification [43]. The distribution of pore diameters is
shown in the inset to Fig. 1(IV)) with a calculated mean of 4.5 nm.
2.4. Catalysis procedure
Reactions were carried out under atmospheric pressure, in situ
immediately after activation, in a fixed bed vertical continuous flow
glass reactor (l = 600 mm, i.d. = 15 mm) over the temperature range
463 K ≤ T ≤ 573 K under conditions of minimal heat/mass transport
limitations. A preheating zone (borosilicate glass beads) ensured
that the m-DNB reactant was vaporised and reached reaction
temperature before contacting the catalyst. Isothermal conditions
3.1.2. Activated carbon supported Au and Ag
The metal loadings (Table 1) coincided with nominal values, a
result that demonstrates the efficiency of the synthesis method-
ology. The TPR profiles (to 623 K, not shown) for 1 and 0.1%wt.
Au/AC and Ag/AC were featureless with no evidence of H2 uptake
or release. This suggests formation of a metallic (Au and Ag)
phase on AC in the as prepared samples, which is in line with
(
1 K) were maintained by thoroughly mixing the catalyst with

F. Cárdenas-Lizana et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 138–146
141
Fig. 1. (I) XPS spectrum over the O 1s region, (II) representative scanning electron microscopy (SEM) images, (III) Raman spectrum and (IV) N2 adsorption (solid sym-
bols)/desorption (open symbols) isotherms for acid treated AC with pore diameter distribution (see inset) from desorption measurements.
Table 1
Specific surface area (SSA), pore volume, metal loading, particle size range and mean (from XRD line broadening (dhkl) and microscopy (d) analysis), H2 chemisorbed (at
523 K)/released (during TPD), XPS binding energies and catalytic results (pseudo-first order rate constant (k) and turnover frequency (TOF)) associated with activated carbon
(AC) support and 1 and 0.1%wt. Au and Ag catalysts.
AC
1%wt. Au/AC
0.1%wt.Au/AC
1%wt. Ag/AC
0.1%wt. Ag/AC
SSA (m2 g−1
)
518
0.34
–
–
–
–
–
–
635
0.48
1
605
0.44
0.1
694
0.57
1
585
0.53
0.1
Pore volume (cm3 g−1
Metal loading (%wt.)
dhkl (nm)a,b
)
12a, 60b
25-60
45
10a, 12b
1-15
4
13a, 30b
3-20
13
10a, 12b
1-9
3
Metal particle size range (nm)b
d (nm)b
−1
b
b
H2 uptake (mmol gmetal
)
)
0.06
1.0
–
1.1
1.3
–
0.04
1.3
–
0.13
0.9
–
TPD H2 released (cm3 g−1
XPS binding energies (eV)b
O 1s
531.2
533.3
Au 4f7/2
Ag 3d5/2
–
–
–
84.0
–
–
83.6
–
126; 210
–
–
368.3
–
368.0
31; 39
k523K (h−1); TOF (h−1
)
a
Samples as prepared.
After thermal treatment to 623 K.
b
reduction of Ag+ (from AgNO3 as precursor) to Ag0. The SSA and
total pore volume were increased after Au and Ag introduction
(Table 1), an effect that was more pronounced at the higher loading
(1%wt.). It has been shown that AC treatment with HNO3 results
in partial pore occlusion [48] where removal of the incorporated
oxygen groups during thermal treatment (up to 573 K) served to
increase surface area and pore volume [49].
The XRD profiles of Au/AC and Ag/AC are presented in Fig. 2 and
show peaks characteristic of metallic Au (2ꢁ = 38.1◦, 44.4◦, 64.7◦ and
77.5◦) or Ag (2ꢁ = 38.1◦, 44.3◦, 64.4◦ and 77.5◦).
reports that have shown (by XPS [44] and EDX [45]) the occur-
rence of zero valent Au and Ag on acid-treated carbon supports.
We should also note reported H2 consumption (at 594 and 423 K)
during TPR of Au/AC [40] and Ag/AC [46] where the AC carrier
was not subjected to acid treatment. In the preparation of Au
on carbon nanofibers Bulushev et al. [23] demonstrated that sur-
face carboxylic groups decomposed during interaction with the
[Au(en)2]Cl3 precursor with concomitant Au3+ reduction to Au0.
Moreover, Jia and Demopoulos [47] proposed that surface hydro-
quinone species on HNO3-treated AC were responsible for the

142
F. Cárdenas-Lizana et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 138–146
in significant metal sintering at the 1%wt. loading which was not
apparent at 0.1%wt.. Published work has shown a wider size range
and formation of larger nanoparticles with increasing metal loading
[20,50]. Metal particle size was further assessed by scanning and
transmission electron microscopy and representative SEM (I) and
TEM (II) images with associated metal particle size distributions
(III) for activated 1%wt. Au (A) and Ag (B) are given in Fig. 3.
The supported metals are present as discrete nanoparticles with
a quasi-spherical shape. Diffractogram patterns for isolated Au
(not shown) and Ag (see inset to image (IIB)) particles confirmed
the metal phase. Supported Au was characterised by a larger size
range (IIIA), suggesting stronger Ag-support interactions that limit
agglomeration (IIIB). Wang et al. [51] and Yang et al. [52] also
reported the formation of larger Au nanocrystals (relative to Ag)
on TiO2 [51] and graphene oxide [52]. The 0.1%wt. loaded samples
exhibit smaller particles, as can be assessed from the TEM images
(I–II) in Fig. 4.
The diffractrogram patterns (III) are consistent with 0.24 and
0.20 nm d-spacings for (1 1 1) Au and (2 0 0) Ag. The TEM-generated
size histograms Fig. 4(IV) establish a narrower size distribution
with mean values that were close for Au (4 nm) and Ag (3 nm).
Smaller particle size at lower loading agrees with the XRD line
broadening measurement. The larger values obtained from the lat-
ter technique can be attributed to limitations of nanoparticle size
measurement using the Scherrer formula where variations in metal
size distribution can result in local modifications in the number of
parallel lattice planes responsible for line broadening, resulting in
over-estimation of size [53].
Fig. 2. XRD patterns associated with (a) as prepared and (b) activated (I) 1%wt. and
(II) 0.1%wt. (A) Au/AC (B) and Ag/AC. JCPDS-ICDD reference diffractograms for Au
(04-0784) and Ag (04-0783) are also included.
The catalytic response in hydrogenation over supported Au
[54] and Ag [55] is influenced by geometric (metal particle size)
and electronic (metal-support interactions) effects that can deter-
mine reactant adsorption/activation. XPS measurements were
conducted to gain some insight into metal site electronic character
and the resulting spectra over the Au 4f (A) and Ag 3d (B) BE regions
associated with 1 (I) and 0.1 (II) %wt. loadings are given in Fig. 5.
An XRD signal for metallic Au and Ag in the as prepared sam-
ples (aIA, aIB, aIIA, aIIB) confirms metal precursor reduction during
synthesis as inferred from TPR analysis. Metal particle size from
standard line broadening (Eq. (4)) was similar (10–13 nm) for the
four as prepared catalysts. Thermal treatment (to 623 K) resulted
Fig. 3. Representative (I) SEM, (II) TEM images and (III) metal particle size distribution for activated 1%wt. (A) Au/AC and (B) Ag/AC. Note: Inset in (IIB) shows the diffractogram
pattern for an isolated Ag particle.

F. Cárdenas-Lizana et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 138–146
143
Fig. 4. Representative (I) medium and (II) high magnification TEM images with (III) diffractogram patterns for isolated metal particles and (IV) size distributions for activated
0.1%wt. (A) Au/AC and (B) Ag/AC.
Fig. 5. XPS spectra over (A) Au 4f and (B) Ag 3d region associated with (I) 1%wt. (solid symbols) and (II) 0.1%wt. (open symbols) Au/AC (᭹,ꢀ) and Ag/AC (ꢀ,ꢁ).
The core level Au 4f7/2 BE (84.0 eV) and Ag 3d5/2 (368.3 eV) for
the higher loading are in good agreement with values reported in
the literature for metallic Au [56] and Ag [57]. A decrease in peak
intensity and increase in line width were observed at the lower
loading as noted elsewhere [58] with a measurable shift to lower
BE values (by 0.3–0.4 eV, see Table 1). Martínez et al. [59] investi-
gating 1%wt. Au/Fe3O4 (9.0 nm, Au 4f7/2 BE = 83.8 eV) and Au/SiO2
(5.7 nm, Au 4f7/2 BE = 82.6 eV) linked the downshift in BE to the for-
mation of Au␦− as a result of electron transfer from the support.
It is known that surface functional groups with different electron-
donating affinities can modify the electron density of supported
metal nanoparticles [60]. Wang et al. demonstrated (by XAS) elec-
tron transfer from HNO3 treated multi-walled carbon nanotubes
(MWCNT) relative to untreated MWCNT and bulk Pt [61]. Such
charge donation via AC oxygenated functionalities appears to apply
to our lower loaded Au and Ag samples. While metal-support elec-
tron transfer can account for differences in the electronic properties
of the supported metal phase in the low vs. high loaded samples,
we should note possible charge transfer effects due to X-ray irra-
diation [62] and a contribution due to differences in metal particle
size [63].
3.2. Catalyst performance
3.2.1. Catalyst activity
The gas phase hydrogenation of m-DNB over Au/C and Ag/AC
at each loading was fully selective in terms of –NO2 reduction
with no evidence of hydrodenitrogenation and/or aromatic ring
reduction. This result is significant as hydrogenolysis of substituted
nitroarenes has been a feature of reaction over oxide supported Pd
[64,65].
Hydrogenation was stable with time on-stream where, under
the same reaction conditions, the lower metal loadings deliv-
ered significantly higher fractional conversion (Fig. 6). In contrast,

144
F. Cárdenas-Lizana et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 138–146
Fig. 6. Variation of –NO2 fractional conversion (X
2 ) with time on-stream over
−NO
1%wt. (solid symbols) and 0.1%wt. (open symbols) Au/AC (᭹,ꢀ) and Ag/AC (ꢀ,ꢁ).
−1
Reaction conditions: T = 523 K, P = 1 atm, n/F = 1 × 10−2 molmetalh mol
.
−NO
2
temporal deactivation in hydrogenation over oxide supported Au
[29,66–69] and Ag [70] has been reported and attributed to coke
formation [29,66], metal sintering [67,70] and reactant poisoning
of active sites [68,69].
Given the higher conversions achieved we focused our atten-
tion on a comparison of 0.1%wt. Au/AC with Ag/AC. Nitro-group
reduction can be quantified in terms of pseudo-first order kinetics
[2]
ꢂ ꢃ
1
n
F
ꢀ
ꢁ
ln[
] = k
(6)
1 − X
−NO
2
where (n/F) has the physical meaning of contact time. The
pseudo-first order rate constants were extracted from the plots
in Fig. 7(I) and are recorded in Table 1. Comparison of intrinsic
activity requires a measure of turnover frequency (TOF, hydrogena-
tion rate per active site obtained from metal dispersion determined
by microscopy [71]) where Au/AC delivered a greater than 5-fold
higher TOF than Ag/AC. Oxide supported Au has been reported to
outperform Ag in the hydrogenation of alkynes (1-hexyne [72]),
esters (dimethyl oxalate [73]) and aldehydes (crotonaldehyde [74]).
Enhanced activity for Au/AC extended over the 463–573 K range
as demonstrated by the Arrhenius plots in Fig. 7(II) with an
apparent activation energy of 110 kJ mol−1 for m-DNB → m-NAN
(Scheme 1) that is somewhat higher than the range of values
(33-74 kJ mol−1) reported for batch liquid phase hydrogenation of
substituted nitroarenes (dinitrobenzene [75], dinitrotoluene [76],
chloronitrobenzene [77] and nitrophenol [78]) over supported Pt
[75,78], Ni [77] and Pd [76].
Hydrogen uptake/activation is a crucial parameter in catalytic
hydrogenation [79]. It has been shown that spillover hydrogen
(i.e. H2 dissociation at a metal site with migration to the sup-
port [80]) can contribute to nitroarene hydrogenation activity
[8,20,81]. Experimental and theoretical work have established that
surface oxygen groups on carbon substrates serve to accommodate
spillover hydrogen [79]. Total surface hydrogen was determined by
TPD and the results are presented in Table 1 and Fig. 8.
Fig. 7. (I) Pseudo-first order kinetic (Reaction conditions: T = 523 K, P = 1 atm) and (II)
Arrhenius plots for the selective hydrogenation of m-DNB to m-NAN over 0.1%wt.
Au/AC (ꢀ) and Ag/AC (ꢁ); (III) Variation of m-NAN (solid symbols) and m-PDM (open
symbols) selectivity (Si) with fractional conversion (Xm-DNB) over 0.1%wt. Au/AC
(᭹,ꢀ) and Ag/AC (ꢀ,ꢁ) (Reaction conditions: T = 548 K, P = 1 atm).
The four samples exhibited H2 desorption over the 645–873 K
temperature range, which is comparable to results (623–873 K)
reported for AC supported Pt [82], Pd [83,84] and Ni [83]. Release
of hydrogen spillover has been shown to require temperatures
in excess of 500 K regardless of the metal or carrier [85]. Total
hydrogen desorbed from the four catalysts was similar (Table 1)
suggesting that H2 uptake during TPR was largely governed by the
surface area and chemistry of the carrier. Chemisorption on sup-
ported Au [71] and Ag [86] is an activated process with increased
uptake at higher temperatures. Hydrogen chemisorption capac-
ity under reaction conditions (523 K) was appreciably higher at
lower loadings (Table 1) and can be linked to facilitated dissocia-
tive chemisorption on edge/corner sites associated with smaller
particles (<5 nm) [87]. The greater uptake on Au relative to Ag at
the lower loading can account for the higher m-DNB hydrogenation
rate.
3.2.2. Catalyst selectivity
Full nitro-group reduction proceeds via stepwise and/or con-
certed routes (Scheme 1). The predominant pathway can be
evaluated from a consideration of product (m-NAN and m-PDM)
selectivity (Si) as a function of m-DNB fractional conversion
(Xm-DNB) and the results over the two low loaded catalysts are
given in Fig. 7(III). Exclusivity in terms of partial −NO2 reduction
to m-NAN at low fractional conversions (<0.06) over both cata-

F. Cárdenas-Lizana et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 138–146
145
are consistent with sequential transformation of m-DNB to m-NAN
at low conversions and a switch to full hydrogenation to m-PDM
at higher conversion. Selectivity to m-NAN was measurably higher
over Ag/AC, which we attribute to differences in m-DNB adsorp-
tion/activation. We have provided an explicit comparison of the
chemoselective hydrogenation response of Au/AC and Ag/AC and
demonstrate that supported Au exhibits a higher intrinsic hydro-
genation rate linked to elevated H2 uptake where selectivity to
m-NAN is greater on Ag.
Acknowledgements
This work was financially supported by the Swiss National
Science Foundation. The authors are grateful to Pedro Benavente-
Donayre and Dr. Antonio Nieto-Márquez for their contributions.
Fig. 8. Hydrogen TPD profiles generated for: (I) 1%wt. and (II) 0.1%wt. (A) Au/AC
(dashed lines) and (B) Ag/AC (solid lines).
References
lysts suggests an analogous reaction mechanism where nitro-group
reduction occurs predominantly in a stepwise fashion (paths (I, II)
in Scheme 1). We should note that catalytic systems reported to
date have promoted the formation of both (m-NAN and m-PDM)
products where high temperatures and/or pressures were identi-
fied as essential to achieve high selectivities to m-NAN. Zhao et al.
[88] reported 97% selectivity to m-NAN over 5%wt. Pt/C (at 160 bar)
[1] P. Lara, K. Philippot, Catal. Sci. Technol. 4 (2014) 2445–2465.
[2] F. Cárdenas-Lizana, M.A. Keane, in: L. Prati, A. Villa (Eds.), Gold Catalysis:
Preparation, Characterization and Applications in the Gas and Liquid Phase,
Pan Stanford Publ, Singapore, 2014.
[3] M. Pietrowski, Curr. Org. Synth. 9 (2012) 470–487.
[4] X. Wang, F. Cárdenas-Lizana, M.A. Keane, ACS Sust. Chem. Eng. 2 (2014)
2781–2789.
[5] F. Cárdenas-Lizana, C. Berguerand, I. Yuranov, L. Kiwi-Minsker, J. Catal. 301
(2013) 103–111.
[6] F. Cárdenas-Lizana, M.A. Keane, J. Mat. Sci. 48 (2013) 543–564.
[7] F. Cárdenas-Lizana, Z. Martínez de Pedro, S. Gómez-Quero, M.A. Keane, J. Mol.
Catal. 326 (2010) 48–54.
[8] F. Cárdenas-Lizana, S. Gómez-Quero, N. Perret, M.A. Keane, Catal. Sci. Technol.
1 (2011) 652–661.
[9] X. Wang, N. Perret, J.J. Delgado, G. Blanco, X. Chen, C.M. Olmos, S. Bernal, M.A.
Keane, J. Phys. Chem. C 117 (2013) 994–1005.
[10] F. Rodríguez-Reinoso, Carbon 36 (1998) 159–175.
[11] T. Szumełda, A. Drelinkiewicz, R. Kosydar, J. Gurgul, Appl. Catal. A: Gen. 487
(2014) 1–15.
[12] H. Gu, X. Xu, A.-A. Chen, P. Ao, X. Yan, Catal. Commun. 41 (2013) 65–69.
[13] M. Bonarowska, Z. Kaszkur, D. Łomot, M. Rawski, Z. Karpin´ ski, Appl. Catal. B:
Environ. 162 (2015) 45–56.
[14] M. Lu, J. Sun, D. Zhang, M. Li, J. Zhu, Y. Shan, Reac. Kinet. Mech. Cat. 100 (2010)
99–103.
[15] X. Zhang, Y.C. Guo, Z.C. Zhang, J.S. Gao, C.M. Xu, J. Catal. 292 (2012) 213–226.
[16] B. Li, H. Li, Z. Xu, J. Phys. Chem. C 113 (2009) 21526–21530.
[17] M. Lu, X. Li, B. Chen, M. Li, H. Xin, L. Song, J. Nanosci. Nanotechnol. 14 (2014)
7315–7318.
[18] X. Yang, Y. Liang, X. Zhao, Y. Song, L. Hu, X. Wang, Z. Wang, J. Qiu, RSC Adv. 4
(2014) 31932–31936.
[19] L. Prati, A. Villa, A.R. Lupini, G.M. Veith, Phys. Chem. Chem. Phys. 14 (2012)
2969–2978.
[20] F. Cárdenas-Lizana, M.A. Keane, Phys. Chem. Chem. Phys. (2015), http://dx.
doi.org/10.1039/C5CP00282F, in press.
[21] F. Cárdenas-Lizana, Y. Hao, M. Crespo-Quesada, I. Yuranov, X. Wang, M.A.
Keane, L. Kiwi-Minsker, ACS Catal. 3 (2013) 1386–1396.
[22] D.S. Cameron, S.J. Cooper, I.L. Dodgson, B. Harrison, J.W. Jenkins, Catal. Today 7
(1990) 113–137.
while selectivities >80% were recorded for reaction (at PH up to
2
45 bar) over Pd/C [25] and Ru/Al2O3 [89]. In this work, full selec-
tivity to m-NAN was attained over Au/C and Au/Ag at ambient
pressure. Selectivity to m-PDM was favoured at increasing con-
version. Consecutive formation of m-NAN and m-PDM has been
reported for batch liquid reaction over Pt/TiO2 [75] and carbon
supported Ni [26] and Ni-Pt [26]. Distinct differences in selectiv-
ity are in evidence for the two metals at higher conversions where
a preferential formation of m-NAN was observed over Ag/AC. This
suggests a deviation in m-DNB adsorption-activation mechanism
for both systems. m-DNB adsorption on Auı− (from XPS) through
the benzene ring results in repulsion between the ꢃ-delocalized
electrons and the charged gold nanoparticles [8]. This generates a
resonance structure with two positive localised charges on the ring
where both –NO2 groups are activated with the formation of m-
PDM via nucleophilic attack. A preferential interaction through the
-NO2 function on supported Ag in functionalised nitroarenes has
been demonstrated by surface-enhanced Raman scattering (SERS)
[90,91]. This results in a resonance structure with a single delo-
calised positive charge on the ring and only one –NO2 group is
activated, favouring m-NAN production.
4. Conclusions
[23] D.A. Bulushev, I. Yuranov, E.I. Suvorova, P.A. Buffat, L. Kiwi-Minsker, J. Catal.
224 (2004) 8–17.
[24] K.A. Wepasnick, B.A. Smith, K.W. Schrote, H.K. Wilson, S.R. Diegelmann, D.H.
Fairbrother, Carbon 49 (2011) 24–36.
[25] V.L. Khilnani, S.B. Chandalia, Org. Proc. Res. Dev. 5 (2001) 263–266.
[26] M.M. Telkar, J.M. Nadgeri, C.V. Rode, R.V. Chaudhari, Appl. Catal. A: Gen. 295
(2005) 23–30.
[27] X. Bokhimi, R. Zanella, V. Maturano, A. Morales, Mater. Chem. Phys. 138
(2013) 490–499.
[28] M. Khoudiakov, M.C. Gupta, S. Deevi, Appl. Catal. A: Gen. 291 (2005) 151–161.
[29] F. Cárdenas-Lizana, X. Wang, D. Lamey, M. Li, M.A. Keane, L. Kiwi-Minsker,
Chem. Eng. J. 255 (2014) 695–704.
[30] C. Amorim, M.A. Keane, J. Chem. Technol. Biotechnol. 83 (2008) 662–672.
[31] G.-M. Choi, T. Ohmi, J. Electrochem. Soc. 148 (2001) G241–G248.
[32] S. Santangelo, G. Messina, G. Faggio, S.H.A. Rahim, C. Milone, J. Raman
Spectrosc. 43 (2012) 1432–1442.
[33] M.K. van der Lee, J. van Dillen, J.H. Bitter, K.P. de Jong, J. Am. Chem. Soc. 127
(2005) 13573–13582.
[34] U. Zielke, K.J. Hüttinger, W.P. Hoffman, Carbon 34 (1996) 983–998.
[35] J.L. Figueiredo, M.F.R. Pereira, M.M.A. Freitas, J.J.M. Órfão, Carbon 37 (1999)
1379–1389.
[36] H. Marsh, F. Rodríguez-Reinoso, Activated Carbon, Elsevier, Amsterdam, 2006.
[37] S. Choi, K.H. Park, S. Lee, K.H. Koh, J. Appl. Phys. 92 (2002) 4007–4011.
Acid (HNO3 + H2SO4) treatment of an amorphous carbon sup-
port incorporated carbonyl and carboxyl anhydride groups (from
acid-base titration and XPS) that facilitate Au and Ag deposition.
Thermal activation of 1%wt. Au/AC and Ag/AC to 623 K generated a
broad metal particle size distribution (25–60 nm Au, mean = 45 nm,
3–20 nm Ag, mean = 13 nm) with XPS BE values close to the refer-
ence metals. A decrease in metal loading (1 → 0.1%wt.) resulted
in a narrower distribution of smaller supported Au (1–15 nm,
mean = 4 nm) and Ag (1–9 nm, mean = 3 nm) nanoparticles with
evidence (from XPS) of charge transfer from the support. The sup-
ported Au and Ag catalysts exhibited stable m-DNB hydrogenation
activity with time on-stream and an equivalent apparent activation
energy (110 kJ mol−1) for selective partial hydrogenation of m-DNB
to m-NAN. Au/AC delivered a (5-fold) higher specific hydrogenation
rate than Ag/AC, which can be attributed to greater H2 chemisorp-
tion under reaction conditions. The conversion/selectivity patterns

146
F. Cárdenas-Lizana et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 138–146
[38] M.S. Dresselhaus, G. Dresselhaus, R. Saito, A. Jorio, Phys. Rep. -Rev. Sec. Phys.
Lett. 409 (2005) 47–99.
[64] T.T. Bovkun, M. Grayevsky, Y. Sasson, J. Blum, J. Mol. Catal. A: Chem. 270
(2007) 171–176.
[39] T. Maiyalagan, B. Viswanathan, Mater. Chem. Phys. 93 (2005) 291–295.
[40] S. Gil, L. Mun˜oz, L. Sánchez-Silva, A. Romero, J.L. Valverde, Chem. Eng. J. 172
(2011) 418–429.
[41] P. Kim, Y.Z. Zheng, S. Agnihotri, Ind. Eng. Chem. Res. 47 (2008) 3170–3178.
[42] A.H. El-Sheikh, A.P. Newman, H.K. Al-Daffaee, S. Phull, N. Cresswell, J. Anal.
Appl. Pyrolysis 71 (2004) 151–164.
[43] M. Thommes, Chem. Ing. Tech. 82 (2010) 1059–1073.
[44] Y. Gu, X. Hou, H. Hu, B. Yu, L. Wang, F. Zhou, Mater. Chem. Phys. 116 (2009)
284–288.
[65] F. Cárdenas-Lizana, S. Gómez-Quero, C. Amorim, M.A. Keane, Appl. Catal. A:
Gen. 473 (2014) 41–50.
[66] Y. Azizi, C. Petit, V. Pitchon, J. Catal. 256 (2008) 338–344.
[67] X. Zhang, H. Shi, B.-Q. Xu, Catal. Today 122 (2007) 330–337.
[68] R. Zanella, C. Louis, S. Giorgio, R. Touroude, J. Catal. 223 (2004) 328–339.
[69] J.A. Lopez-Sanchez, D. Lennon, Appl. Catal. A: Gen. 291 (2005) 230–237.
[70] M. Bron, D. Teschner, A. Knop-Gericke, F.C. Jentoft, J. Kröhnert, J. Hohmeyer, C.
Volckmar, B. Steinhauer, R. Schlögl, P. Claus, Phys. Chem. Chem. Phys. 9 (2007)
3559–3569.
[45] J.S. Hoskins, T. Karanfil, S.M. Serkiz, Environ. Sci.Technol. 36 (2002) 784–789.
[46] L. Chen, D. Ma, X. Bao, J. Phys. Chem. C 111 (2007) 2229–2234.
[47] Y. Jia, G.P. Demopoulos, Ind. Eng. Chem. Res. 42 (2002) 72–79.
[48] J. Li, L. Ma, X. Li, C. Lu, H. Liu, Ind. Eng. Chem. Res. 44 (2005) 5478–5482.
[71] M. Li, X. Wang, N. Perret, M.A. Keane, Catal. Commun. 46 (2014) 187–191.
[72] G. Vilé, J. Pérez-Ramírez, Nanoscale 6 (2014) 13476–13482.
[73] J. Zheng, H. Lin, Y.-N. Wang, X. Zheng, X. Duan, Y. Yuan, J. Catal. 297 (2013)
110–118.
[74] X. Yang, A. Wang, X. Wang, T. Zhang, K. Han, J. Li, J. Phys. Chem. C 113 (2009)
20918–20926.
´
[49] G.S. Szyman´ ski, Z. Karpin´ ski, S. Biniak, A. Swia˛kowski, Carbon 40 (2002)
2627–2639.
[50] S.H. Overbury, V. Schwartz, D.R. Mullins, W. Yan, S. Dai, J. Catal. 241 (2006)
56–65.
[75] H. Rojas, G. Borda, M. Brijaldo, P. Reyes, J. Valencia, Reac. Kinet. Mech. Catal.
105 (2012) 271–284.
[51] Y.M. Wang, G.J. Du, H. Liu, D. Liu, S.B. Qin, J.Y. Wang, X.T. Tao, M.H. Jiang, Z.L.
Wang, J. Nanosci. Nanotechnol. 9 (2009) 2119–2123.
[52] M.-Q. Yang, X. Pan, N. Zhang, Y.-J. Xu, CrystEngComm 15 (2013)
6819–6828.
[53] H. Borchert, E.V. Shevchenko, A. Robert, I. Mekis, A. Kornowski, G. Grübel, H.
Weller, Langmuir 21 (2005) 1931–1936.
[76] G. Neri, M.G. Musolino, C. Milone, S. Galvagno, Ind. Eng. Chem. Res. 34 (1995)
2226–2231.
[77] Y.-Z. Chen, Y.-C. Chen, Appl. Catal. A: Gen. 115 (1994) 45–57.
[78] X. Li, G. Li, W. Zang, L. Wang, X. Zhang, Catal. Sci. Technol. 4 (2014) 3290–3297.
[79] R. Prins, Chem. Rev. 112 (2012) 2714–2738.
[80] R. Prins, V.K. Palfi, M. Reiher, J. Phys. Chem. C 116 (2012) 14274–14283.
[81] F. Cárdenas-Lizana, S. Gómez-Quero, C.J. Baddeley, M.A. Keane, Appl. Catal. A:
Gen. 387 (2010) 155–165.
[54] M. Boronat, P. Concepción, A. Corma, S. González, F. Illas, P. Serna, J. Am.
Chem. Soc. 129 (2007) 16230–16237.
[55] G. Vilé, D. Baudouin, I.N. Remediakis, C. Copéret, N. López, J. Pérez-Ramírez,
ChemCatChem 5 (2013) 3750–3759.
[56] S. Peters, S. Peredkov, M. Neeb, W. Eberhardt, M. Al-Hada, Surf. Sci. 608 (2013)
129–134.
[57] E. Sumesh, M.S. Bootharaju, Anshup, T. Pradeep, J. Hazard. Mater. 189 (2011)
450–457.
[58] W. Grünert, Characterization of Solid Materials and Heterogeneous Catalysts:
From Structure to Surface Reactivity, in: M. Che, J.C. Védrine (Eds.),
Wiley-VCH, Weinheim, 2012.
[82] A. Huidobro, A. Sepúlveda-Escribano, F. Rodríguez-Reinoso, in: A.
Guerrero-Ruíz, I. Rodríguez-Ramos (Eds.), Studies in Surface Science and
Catalysis, Vol. 138, Elsevier, Amsterdam, 2001, pp. 275–283.
[83] C. Amorim, X. Wang, M.A. Keane, Chin. J. Catal. 32 (2011) 746–755.
[84] C. Amorim, G. Yuan, P.M. Patterson, M.A. Keane, J. Catal. 234 (2005) 268–281.
[85] C. Amorim, M.A. Keane, J. Hazard. Mater. 211–212 (2012) 208–217.
[86] J. Hohmeyer, E.V. Kondratenko, M. Bron, J. Kröhnert, F.C. Jentoft, R. Schlögl, P.
Claus, J. Catal. 269 (2010) 5–14.
[87] E. Bus, J.T. Miller, J.A. van Bokhoven, J. Phys. Chem. B 109 (2005) 14581–14587.
[88] F. Zhao, S.-I. Fujita, J. Suna, Y. Ikushima, M. Arai, Catal. Today 98 (2004)
523–528.
[59] J.J. Martínez, H. Rojas, L. Vargas, C. Parra, M.H. Brijaldo, F.B. Passos, J. Mol.
Catal. A: Chem. 383–384 (2014) 31–37.
[60] J. Zhu, A. Holmen, D. Chen, ChemCatChem 5 (2013) 378–401.
[61] X. Wang, N. Li, J.A. Webb, L.D. Pfefferle, G.L. Haller, Appl. Catal. B: Environ. 101
(2010) 21–30.
[89] S. Zhao, H. Liang, Y. Zhou, Catal. Commun. 8 (2007) 1305–1309.
[90] V. Chis¸ , M.M. Venter, N. Leopold, O. Cozar, Vib. Spectrosc. 48 (2008) 210–214.
[91] M. Muniz-Miranda, Appl. Catal. B: Environ. 146 (2014) 147–150.
[62] N. Kruse, S. Chenakin, Appl. Catal. A: Gen. 391 (2011) 367–376.
[63] J.T. Miller, A.J. Kropf, Y. Zha, J.R. Regalbuto, L. Delannoy, C. Louis, E. Bus, J.A.
van Bokhoven, J. Catal. 240 (2006) 222–234.