Effects of alkali promotion of TiO2 on the chemisorptive properties and water-gas shift activity of supported noble metal catalysts

Article abstract of DOI:10.1016/j.jcat.2009.07.014

The effects of alkali promotion of TiO₂ on the chemisorptive properties and water-gas shift (WGS) activity of dispersed noble metal catalysts (NM = Pt, Ru, Pd) have been investigated over NM/X-TiO₂ samples of variable promoter type (

Full text of DOI:10.1016/j.jcat.2009.07.014

Journal of Catalysis 267 (2009) 57–66
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Effects of alkali promotion of TiO2 on the chemisorptive properties and water–gas
shift activity of supported noble metal catalysts
*
Paraskevi Panagiotopoulou, Dimitris I. Kondarides
Department of Chemical Engineering, University of Patras, GR-26504 Patras, Greece
a r t i c l e i n f o
a b s t r a c t
Article history:
The effects of alkali promotion of TiO2 on the chemisorptive properties and water–gas shift (WGS) activ-
ity of dispersed noble metal catalysts (NM = Pt, Ru, Pd) have been investigated over NM/X–TiO2 samples
of variable promoter type (X = Li, Na, K, Cs) and loading (0.0 to 0.68 wt.%). Results of H2-TPD experiments
show that addition of alkalis does not affect appreciably the population and chemisorption strength of
hydrogen adsorbed on the surface of dispersed metal crystallites indicating that these sites are not influ-
enced by the presence of the promoter. In contrast, the desorption temperature of hydrogen adsorbed on
sites located at the metal-support interface shifts monotonically toward lower temperatures with
increasing alkali content. This has been attributed to alkali-induced reduction of Ti4+ surface species
and creation of a new type of sites at the perimeter of the dispersed metal crystallites, which are in con-
tact with the support. These sites are of the form of NM–hs–Ti3+, where hs denotes an oxygen defect
vacancy, and provide the necessary dual-function sites required for the WGS reaction to proceed. Cata-
lytic performance tests and kinetic measurements show that activity depends appreciably on the nature
and loading of added alkali, whereas the apparent activation energy (Ea) of the reaction does not vary to a
large extent. For Pt/X–TiO2 catalysts, a volcano-type dependence of intrinsic reaction rate on the chemi-
sorption strength of Pt–hs–Ti3+ sites toward hydrogen has been found to exist. Optimized catalysts are
about three times more active than unpromoted Pt/TiO2. Results are discussed with respect to the
alkali-induced modifications of the physicochemical properties of the support on the population, chem-
isorptive properties, and catalytic activity of sites located at the metal-support interface.
Received 9 April 2009
Revised 26 July 2009
Accepted 28 July 2009
Available online 2 September 2009
Keywords:
Water–gas shift
Noble metal
Titanium dioxide
Oxygen vacancy
Alkali
Promotion
TPD
Kinetic measurements
Metal-support interactions
SMSI effect
Ó 2009 Elsevier Inc. All rights reserved.
1. Introduction
metal catalysts, which is often reflected in enhanced activity, high-
er selectivity, suppression of undesirable reactions, and/or
Chemical [1–11] or electrochemical [12–17] promotion of cata-
lytic materials by alkalis has been the subject of several recent
investigations. The origin of the continued interest in this topic is
both of fundamental and applied nature. Alkali additives are
known to enhance the rate of many industrially important catalytic
reactions, such as ammonia and Fisher–Tropsch syntheses [18],
and can be also used to study the effects of promoters on the chem-
isorptive and catalytic properties of materials. Examples of alkali-
promoted catalytic reactions investigated in the past few years in-
clude selective catalytic reduction of NO by CO or hydrocarbons in
the absence [1,2] or in the presence [6,13,14] of oxygen, oxidation
of NO [4], ethylene epoxidation [15], deep oxidation of ethanol [5]
and hydrocarbons [16], oxidation [17] and preferential oxidation
[6,7] of carbon monoxide, CO hydrogenation [8,9] and water–gas
shift reaction [10,11]. Results of these studies clearly show that
alkalis induce a strong promotional effect on the performance of
improvement of catalyst stability [1–17]. The ability of alkali addi-
tives to modify catalytic performance can be understood by consid-
ering the geometric and electronic-type effects induced by the
promoting species on the chemisorptive properties of metal sur-
faces toward reactive molecules [19,20]. Regarding the latter type
of interactions, alkali metals are electron donors and, when present
on the surface of a metal catalyst, may act by enhancing chemi-
sorption of electron acceptor species, such as carbon monoxide
and oxygen, and/or by suppressing chemisorption of electron do-
nors, such as olefins and hydrogen [12].
Water–gas shift (CO + H2O M CO2 + H2) is a suitable reaction for
studying the effects of promoters on catalytic performance because
it involves simple molecules, namely, carbon monoxide and hydro-
gen, which can be easily probed with a variety of techniques. In
addition, the reaction is of significant practical importance, mainly
due to the recent efforts for developing fuel processors capable of
converting carbonaceous fuels into hydrogen for fuel cell applica-
tions [21].
In our previous studies [22–26], we have investigated in detail
the WGS activity of noble metal catalysts (Pt, Rh, Ru, Pd) dispersed
* Corresponding author. Fax: +30 2610 991527.
E-mail address: dimi@chemeng.upatras.gr (D.I. Kondarides).
0021-9517/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.07.014

58
P. Panagiotopoulou, D.I. Kondarides / Journal of Catalysis 267 (2009) 57–66
on a variety of metal oxide supports. It has been found that Pt is
generally much more active than Pd, with Rh and Ru exhibiting
an intermediate performance [22]. It was also shown that noble
metals supported on reducible metal oxides, such as TiO2 [22–
24] and CeO2 [23–25], are characterized by higher activity for the
WGS reaction than catalysts supported on irreducible oxides, such
as Al2O3 or SiO2 [23,24], and that activity generally increases with
decreasing primary particle size of the support [22,24,26]. In our
recent investigation [27], it was shown that addition of Na or Cs
on Pt/TiO2 catalyst results in the creation of new sites with in-
creased electron density, proposed to be located at the metal-sup-
port interface. The adsorption strength of these sites toward CO
was found to increase with increase of alkali content, whereas
the opposite was found to be true for adsorption of hydrogen
[27]. The aim of the present study is to investigate the promoting
effects of alkalis (Li, Na, K, Cs) on the WGS activity of NM/TiO2 cat-
alysts (NM = Pt, Ru, Pd) and to correlate catalytic performance with
the alkali-induced alterations of the physicochemical properties of
the support and of the chemisorptive properties of dispersed noble
metal crystallites.
60 cm3 minꢀ1. The catalyst was then heated at 500 °C under He
flow, left at that temperature for 15 min, and finally conditioned
by exposure to the reaction mixture at 450 °C for 1 h. Concentra-
tions of reactants and products were determined at steady-state
conditions using the analysis system described above. Similar mea-
surements were obtained following a stepwise lowering of temper-
ature, until conversion of CO dropped close to zero. All
experiments were performed at near atmospheric pressure.
Measurements of intrinsic reaction rates were obtained in sep-
arate experiments under differential reaction conditions, where
the conversion of reactants was kept below 10%. Results were used
to determine the turnover frequency (TOF) of carbon monoxide,
defined as moles of CO converted per surface noble metal atom
per second. Details on the methods and procedures employed
can be found elsewhere [22].
3. Results
3.1. Catalyst characterization
In Table 1 are summarized results of physicochemical charac-
terization of the synthesized materials. It is observed that specific
surface area (SSA) is practically the same for all catalysts investi-
gated (25 to 30 m2 gꢀ1), indicating that this parameter is not influ-
enced by addition of small amounts of alkalis on the support. The
same is true for the primary crystallite size of TiO2, determined by
X-ray line broadening, which was found to vary in the range of 20
to 26 nm. Regarding the anatase content of the support, it is gener-
ally higher for alkali-promoted samples than bare TiO2 (Table 1). It
should be noted, however, that both SSA and anatase content are
substantially lower than those of the parent TiO2 material
(48 m2 gꢀ1, 75% anatase). This is because synthesis included heat
treatment at elevated temperature (600 °C for 3 h), which is known
to result in transformation of TiO2 to its rutile form, particle
growth, and decrease in specific surface area [27].
2. Experimental
2.1. Catalyst preparation and characterization
Alkali-promoted TiO2 supports, denoted in the following as X–
TiO2 (X = Li, Na, K, Cs), were prepared by impregnation of titanium
dioxide powder (Degussa P25) with an aqueous solution contain-
ing the appropriate amount of Li2CO3, NaNO3, KNO3, or CsNO3. This
was followed by drying at 110 °C overnight and calcination in air at
600 °C for 3 h [27]. Dispersed noble metal catalysts (NM = Pt, Ru,
Pd) were prepared employing the wet impregnation method with
the use of (NH3)2Pt(NO2)2, Ru(NO)(NO3)3, or (NH3)2Pd(NO2)2 (Alfa)
as metal precursor salts and the above X–TiO2 powders as supports
[22,27]. The solid residue was dried at 110 °C for 24 h and then re-
duced at 300 °C (400 °C for Ru catalysts) in H2 flow for 2 h. The
nominal metal loading of all catalysts thus prepared was 0.5 wt%.
Specific surface area (SSA) of catalytic materials was measured
employing nitrogen physisorption at the temperature of liquid
nitrogen (B.E.T. method). Noble metal dispersion (DNM) and mean
crystallite size (dNM) were estimated by selective chemisorption
of CO (H2 for Ru catalysts). The anatase-to-rutile content of TiO2
was determined with the use of X-ray diffraction (XRD). Details
on the apparatuses, methods, and procedures used for catalyst
characterization can be found elsewhere [22].
Regarding metal dispersion, results presented in Table 1 show
that platinum is well dispersed for all Pt/X–TiO2 samples investi-
gated, with the average size of Pt crystallites ranging between
1.0 and 1.5 nm. Dispersion of Ru and Pd catalysts is relatively low-
er, and the corresponding mean crystallite sizes vary in the ranges
of 1.4 to 2.3 nm and 1.7 to 2.4 nm, respectively (Table 1).
3.2. Temperature-programmed desorption of H2
3.2.1. Pt catalysts
Temperature-programmed desorption (TPD) experiments were
carried out over freshly prepared catalysts as described in detail
elsewhere [27]. Briefly, an amount of catalyst (200 mg) was re-
duced in situ with hydrogen at 300 °C, purged with He at 500 °C
to remove adsorbed species from the catalyst surface, and then
cooled down to 25 °C under He flow. This was followed by adsorp-
tion of H2 at 25 °C for 15 min, purging with He for 10 min, and lin-
ear increase of temperature (b = 30 °C minꢀ1) to 650 °C under He
flow (40 cm3 minꢀ1). A mass spectrometer (Omnistar, Pfeiffer Vac-
uum) was used for on-line monitoring of the TPD patterns.
Results of H2-TPD experiments obtained over bare and alkali-
promoted Pt/TiO2 catalysts of the same dopant loading (alkali:Pt
atomic ratio = 1:1) are presented in Fig. 1. It is observed that hydro-
gen desorbs from the unpromoted sample (trace a) exhibiting
three peaks centered at ca. 115, 285, and 350 °C. As has been dis-
cussed in detail in our previous study [27], the low temperature
(LT) peak is due to hydrogen chemisorbed on the surface of Pt crys-
tallites [28–30], whereas the high temperature (HT) feature can be
attributed to spillover hydrogen associated with the support
[29,31]. The medium temperature (MT) peak has been assigned
to hydrogen adsorbed on sites located at the periphery of Pt crys-
tallites, which are in contact with the TiO2 support [27]. This
assignment is in agreement with previous work obtained over Pt/
LTL zeolite catalysts [31] and over Rh catalysts supported on
Al2O3 or CeO2 [32,33], which indicated the presence of new adsorp-
tion sites at the metal-support interface.
2.2. Catalytic performance tests and kinetic measurements
The catalytic performance of the synthesized materials for the
WGS reaction was investigated in the temperature range of 100
to 500 °C, using a feed stream consisting of 3% CO and 10% H2O
(balance He). The mass of catalyst used in these experiments was
typically 100 mg (particle size: 0.18 < dp < 0.25 mm) and the total
flow rate was 200 cm3 minꢀ1. Prior to each experiment the catalyst
was reduced in situ at 300 °C for 1 h under a hydrogen flow of
Addition of alkalis (traces b–e) does not affect appreciably the
intensity or position of the LT peak, indicating that adsorption sites
located on the surface of Pt crystallites are not influenced by the
presence of the promoters. This finding, which is supported by

P. Panagiotopoulou, D.I. Kondarides / Journal of Catalysis 267 (2009) 57–66
59
Table 1
Physicochemical characteristics of synthesized NM/X–TiO2 catalysts and representative results of kinetic measurements.
Dispersed metal
catalyst (0.5 wt.%) nominal loading (wt.%)
Promoter type and
Alkali:NM atomic SSA (m2 gꢀ1
ratio (nominal)
)
Anatase
content (%) dispersion
Noble metal dNM (Å) Rate at 250 °C
mol sꢀ1 gꢀ1
Ea (kJ molꢀ1
)
(
l
)
TOF (sꢀ1
)
Pt
None
Li
Na
0
0.0
1.0
0.3
1.0
2.0
3.4
1.0
0.5
1.0
2.0
29
25
30
27
28
28
30
30
28
28
48
36
52
56
50
56
55
50
48
59
0.87
0.89
0.94
1.01
0.66
0.81
0.71
0.78
0.77
0.74
12
11
11
10
15
13
14
13
13
14
10.3
28.9
18.1
38.4
23.4
22.3
26.7
16.9
23.6
0.51
1.27
0.76
1.58
1.35
1.07
1.47
0.86
1.19
0.40
66
76
63
71
69
72
63
69
68
68
0.018
0.017
0.06
0.12
0.20
0.10
0.17
0.34
0.68
K
Cs
7.75
Ru
Pd
None
Na
0
0.06
0.20
0.0
0.5
1.8
29
27
28
48
56
56
0.46
0.70
0.42
21
14
23
1.56
5.87
6.40
0.07
0.17
0.30
66
52
49
None
Cs
0
0.34
0.68
0.0
0.5
1.1
29
28
28
48
48
59
0.67
0.47
0.54
17
24
21
0.82
2.29
1.30
0.03
0.10
0.05
72
62
60
results of our recent FTIR experiments using CO as probe molecule
[27], implies that alkali atoms are mainly located on the TiO2
support. It should be noted that the synthesis method employed
here involves addition of alkali on TiO2 prior to dispersion of noble
metals. As will be discussed below, the added alkalis interact
strongly with the support and this probably hinders diffusion of
promoter atoms to the surface of dispersed metal crystallites.
The HT shoulder decreases in intensity in the presence of alkalis
as a result of alkali-induced dehydroxylation of the support, which
is known to suppress spillover of hydrogen [7,27,31]. Regarding the
MT peak, it shifts from 285 °C for the unpromoted catalyst to ca.
245 °C for samples promoted with Li, K, or Cs and to 260 °C for
the catalyst promoted with Na (Fig. 1). This implies that the
adsorption strength of the corresponding sites toward hydrogen
decreases significantly in the presence of the promoters. Thus, it
seems that the effect of alkali on the chemisorptive properties of
platinum is of short-range, i.e., it is restricted to metal sites located
in the close vicinity of the alkali-promoted TiO2 support.
Similar H2-TPD experiments conducted over catalysts pro-
moted with variable amounts of Na (0 to 0.20 wt.%) or Cs (0 to
0.68 wt.%) showed that the downward shift of the MT peak in-
creases with increasing alkali content [27]. Results obtained for
all Pt/X–TiO2 catalyst samples investigated in our previous [27]
and in the present study are summarized in Fig. 2, where the
temperature at which the maximum of the MT peak is observed
(Tmax) is plotted as function of alkali:Pt atomic ratio. It is ob-
served that Tmax decreases monotonically with increasing alkali
content, and that the effect is more pronounced for Cs-doped
than Na-doped samples of the same alkali:Pt ratio. As will be dis-
cussed below, the presence of electropositive alkali atoms in the
vicinity Pt sites located at the metal-support interface results in
the creation of sites of higher electron-donating properties and,
therefore, of weaker Pt–H bond strength. It should be noted that
the opposite is true for adsorption of CO, which is an electron
acceptor molecule. In fact, our recent FTIR results have shown
Promoter
(e)
(d)
Cs
K
Na
(c)
(b)
Li
that addition of alkali results in a red shift of the m(CO) stretching
frequency of species adsorbed on these sites, which implies
strengthening of the Pt–CO bond with increasing alkali content
[27].
none
(a)
3.2.2. Pd and Ru catalysts
In order to investigate if the effect of alkali promotion on hydro-
gen adsorption/desorption characteristics is operable for noble me-
tal catalysts other than platinum, H2-TPD experiments were also
conducted for Pd and Ru dispersed on representative X–TiO2 sup-
ports. Results obtained over Pd/Cs–TiO2 and Ru/Na–TiO2 catalysts
are depicted in Fig. 3. It is observed that hydrogen desorbs from
unpromoted Pd/TiO2 (Fig. 3A, trace a) exhibiting a LT peak centered
100
200
300
400
Temperature (oC)
Fig. 1. Temperature-programmed desorption profiles obtained over Pt/TiO2 and Pt/
X–TiO2 (X = Li, Na, K, Cs) catalysts of the same alkali:Pt atomic ratio (X:Pt = 1:1)
following adsorption of hydrogen at 25 °C for 15 min.

60
P. Panagiotopoulou, D.I. Kondarides / Journal of Catalysis 267 (2009) 57–66
(A)
Pd/Cs-TiO2
Promoter
280
260
240
220
200
none
Li
Cs:Pd ratio
(c) 1.1
Na
K
Cs
(b) 0.5
(a) 0.0
0
1
2
3
100
200
300
400
500
Temperature (oC)
Alkali:Pt atomic ratio
Fig. 2. Effect of alkali:Pt atomic ratio on the desorption temperature (Tmax) of the
medium temperature (MT) hydrogen peak observed in H2-TPD profiles of Pt/X–TiO2
catalysts (X = Li, Na, K, Cs).
(B) Ru/Na-TiO2
Na:Ru ratio
(c) 1.8
at ca. 100 °C, a MT peak located at 275 °C, and a HT shoulder of
much lower intensity above 300 °C. The LT feature is assigned to
hydrogen adsorbed on Pd surface [34–36] and may contain a con-
tribution from palladium hydride decomposition [36,37]. The ori-
gin of H2-TPD peaks appearing in the temperature range of 150
to 300 °C is unclear in the literature. Peaks in this region have been
often attributed to hydrogen bound to different Pd surface sites,
i.e., top, bridge, and multiple coordinated forms on a given surface
plane, and/or to hydrogen adsorbed on different Pd planes [34–36].
Following the same reasoning used for Pt/TiO2 catalysts, we tenta-
tively assign the MT peak located at 275 °C to hydrogen adsorbed
at the Pd–TiO2 interface and the weak HT feature located above
300 °C to spillover hydrogen.
Addition of Cs does not affect, practically, the position and
intensity of the LT peak (traces b and c). In contrast, the MT peak
exhibits a significant shift toward lower temperatures with in-
crease of Cs content, i.e., from 275 °C over the unpromoted cata-
lyst to 215 and 195 °C for samples with Cs:Pd ratios of 0.5 and
1.1, respectively. The peak intensity increases significantly for
low Cs content (trace b) and decreases again for higher dopant
loadings (trace c). Qualitatively similar results have been reported
by Kazi et al. over Li-promoted Pd/SiO2 catalysts [8]. The authors
found that addition of Li (Li:Pd ratio = 1) resulted in a significant
shift of the major H2 TPD peak from 235 to 160 °C [8]. It was con-
cluded that promotion by Li decreases the strength of hydrogen
adsorption and increases the strength of CO adsorption on Pd/
SiO2 [8].
(b) 0.5
(a) 0.0
400 500
100
200
300
Temperature (oC)
Fig. 3. Temperature-programmed desorption profiles obtained from (A) Cs-pro-
moted Pd/TiO2 and (B) Na-promoted Ru/TiO2 catalysts of the indicated alkali
loadings.
as SiO2. Regarding the HT peak, which is attributed to spillover
hydrogen, it is not discernible over Ru/TiO2. This is in agreement
with results of ESR studies obtained over NM/TiO2 catalysts [39],
which showed that spillover of hydrogen from the metal to the
support takes place readily on catalysts containing Pt, Pd, and Rh,
but not over Ru/TiO2.
The TPD pattern obtained for the Ru/TiO2 catalyst is character-
ized by two peaks located at ca. 110 and 260 °C (Fig. 3B, trace a). As
in the case of Pt and Pd catalysts, the LT peak can be assigned to
hydrogen desorption from Ru metal, and the MT peak to hydrogen
originating from the metal-support interface. Qualitatively similar
TPD patterns have been reported by Lin and Chen [38] over Ru cat-
alysts dispersed on TiO2, Al2O3, and SiO2 where, depending on the
nature of the support, the LT peak was centered at 85 to 125 °C. The
authors did not provide an assignment for the MT peak located at
around 325 °C for Ru/Al2O3 and around 200 °C for Ru/TiO2. It is of
interest to note that the latter peak did not appear over Ru/SiO2
[38]. As will be discussed below, this is probably expected because
creation of sites at the metal-support interface should be less
favorable for noble metals dispersed on irreducible supports, such
Addition of 0.06 wt.% Na (Na:Ru = 0.5) to the TiO2 support re-
sults in a significant shift of the MT peak from 245 to 225 °C (trace
b), whereas increase of Na loading to 0.20 wt.% (Na:Ru = 1.8) re-
sults in a further shift of Tmax to 215 °C (trace c). This is accompa-
nied by a substantial decrease in the intensity of the MT peak.
However, the position of the LT peak is not influenced, practically,
by the presence of alkali.
3.3. Catalytic performance tests and kinetic measurements
Resent results obtained over supported platinum catalysts
[40–42] showed that freshly prepared Pt/TiO2 deactivates with
time-on-stream under WGS reaction conditions due to loss of Pt

P. Panagiotopoulou, D.I. Kondarides / Journal of Catalysis 267 (2009) 57–66
61
surface area [40]. However, Zhu et al. [42] showed that stability of
100
80
60
40
20
0
Pt/TiO2 can be improved by addition of Na, which was shown to in-
hibit sintering of Pt crystallites. In the present study, NM/X–TiO2
catalysts were pre-conditioned in situ following the procedure de-
scribed in Section 2.2 (15 min in He flow at 500 °C followed by
exposure to the reaction mixture at 450 °C for 1 h). This treatment
resulted in materials, which were very stable under the present
reaction conditions, with no signs of deactivation.
(A)
3.3.1. Pt catalysts
Promoter
The effects of addition of alkalis on the catalytic performance of
Pt/TiO2 have been investigated over samples promoted with vari-
able amounts of Li, Na, K, or Cs. Results obtained for catalysts of
the same alkali:Pt atomic ratio (X:Pt = 1:1) are shown in Fig. 4A,
where the conversion of CO (XCO) is plotted as a function of reac-
tion temperature. The equilibrium conversion, predicted by ther-
modynamics, is also shown for comparison (dashed line). For the
unpromoted catalyst, it is observed that XCO becomes measurable
at temperatures around 175 °C and increases with increasing tem-
perature until it reaches equilibrium at ca. 400 °C. In all cases, addi-
tion of alkalis results in a significant shift of the conversion curve
toward lower reaction temperatures. This is more pronounced for
the Na-promoted sample than samples doped with Li, K, or Cs,
which exhibit a similar performance. The Na-promoted catalyst
is able to reach equilibrium CO conversion at temperatures around
325 °C, which is about 75 °C lower than that obtained for the
unpromoted catalyst. Comparison with the H2-TPD results ob-
tained over this set of catalysts (Fig. 1) shows that a relation may
exist between catalytic performance and the temperature at which
the MT peak evolves. In particular, Li-, K-, and Cs-doped catalysts,
which exhibit a similar WGS activity (Fig. 4A), are characterized by
the same Tmax for the MT TPD peak (Fig. 1). This issue will be dis-
cussed in detail below.
none
Li
K
Na
Cs
Temperature (oC)
1E-4
1E-5
1E-6
Promoter
(B)
none
Li
K
Na
Cs
Results of measurements of intrinsic reaction rates, obtained by
using low-conversion data exclusively, are summarized in the
Arrhenius plots of Fig. 4B. It is observed that the rate (per gram
of catalyst) depends on the nature of the promoter and decreases
in the order of Na > Li ꢁ K ꢁ Cs > TiO2(unpromoted). The same
trend is observed for turnover frequencies of CO conversion. Repre-
sentative results listed in Table 1 show that TOF at 250 °C is more
than three times higher for the Na-containing sample than unpro-
moted Pt/TiO2. The apparent activation energy (Ea) of the WGS
reaction was calculated from the slopes of the fitted lines shown
in Fig. 4B and results are summarized in Table 1. It is observed that
Ea varies to some extent for this set of catalysts, taking values be-
1.8
2.0
2.2
2.4
1000/T(K-1)
Fig. 4. Effect of alkali promotion on the catalytic performance of Pt/X–TiO2 catalysts
of the same alkali:Pt atomic ratio (X:Pt = 1:1). (A) Conversion of CO as function of
reaction temperature. Equilibrium CO conversion is shown for comparison (dashed
line). (B) Arrhenius plots of reaction rates obtained under differential reaction
conditions.
tween 66 and 76 kJ molꢀ1
.
The effect of alkali loading on catalytic performance has been
investigated over Pt/X–TiO2 catalysts of variable Na or Cs content,
and results obtained are summarized in Fig. 5. For Pt/Na–TiO2 cat-
alysts, it is observed that increasing Na:Pt ratio from 0.0 to 1.0 re-
sults in a progressive shift of CO conversion curve toward lower
temperatures (Fig. 5A). Further increase of Na content has the
opposite effect, as evidenced by the shift of the conversion curve
toward higher temperatures. Qualitatively similar results were
obtained over Pt/Cs–TiO2 catalysts (Fig. 5B). In particular, increas-
ing Cs:Pt ratio from zero to 1.0 results in a shift of CO conversion
curve toward lower temperatures, whereas further increasing Cs
content results in catalytic performance similar to that of the
unpromoted catalyst. Comparison of results shown in Fig. 5A
and B shows that Na catalysts exhibit higher CO conversions at
a given temperature than Cs samples of the same alkali:Pt atomic
ratio.
creases with increasing Na or Cs content up to a certain value
and then decreases upon further increasing alkali loading. Regard-
ing turnover frequencies, representative results listed in Table 1
show that the highest TOF is obtained for the sample doped with
Na:Pt = 1.0. The apparent activation energy of the reaction does
not depend strongly on the nature or concentration of the pro-
moter and takes values in the range of 63 to 76 kJ molꢀ1 (Table 1).
3.3.2. Ru and Pd catalysts
In our previous study of NM/TiO2 catalysts [22], it was shown
that the specific reaction rate (TOF) for the title reaction decreases
in the order of Pt > Rh > Ru > Pd, with Pt being about 20 times more
active than Pd. In order to investigate the effect of alkali promotion
of TiO2 on the WGS activity of dispersed noble metal catalysts
which are less active than Pt, representative catalytic performance
tests and kinetic measurements were conducted over selected Ru/
Na–TiO2 and Pd/Cs–TiO2 catalysts. In Fig. 7 are shown results ob-
tained for Ru/Na–TiO2 catalysts of variable Na content. It is ob-
Results of reaction rate measurements obtained over the same
sets of Na- and Cs-promoted Pt/TiO2 catalysts are shown in the
Arrhenius-type plots of Fig. 6A and B, respectively. It is observed
that, in both cases, the reaction rate (per gram of catalyst) in-

62
P. Panagiotopoulou, D.I. Kondarides / Journal of Catalysis 267 (2009) 57–66
1E-4
100
80
60
40
20
0
(A)
(A)
Na:Pt ratio
0.0
0.3
1.0
2.0
3.4
1E-5
Na:Pt ratio
0.0
0.3
1.0
1E-6
2.0
3.4
400
1.8
2.0
1000/T (K -1
2.2
2.4
200
250
300
350
)
Temperature (oC)
1E-4
1E-5
1E-6
100
80
60
40
20
0
(B)
Cs:Pt ratio
0.0
(B)
0.5
1.0
2.0
Cs:Pt ratio
0.0
0.5
1.0
2.0
1.8
2.0
1000/T (K -1
2.2
2.4
)
200
250
300
350
400
Temperature (oC)
Fig. 6. Arrhenius plots of reaction rates obtained over (A) sodium- and (B) cesium-
promoted Pt/X–TiO2 catalysts.
Fig. 5. Effects of (A) sodium and (B) cesium loading, expressed as alkali:Pt atomic
ratio, on the catalytic performance of Pt/X–TiO2 catalysts. Equilibrium CO conver-
sion is shown for comparison (dashed line).
the apparent activation energy (Ea) of the reaction which, for Pt/
Na–TiO2 catalysts, is generally higher than unpromoted Pt/TiO2
(Table 1). Qualitatively similar results have been obtained by Yen-
tekakis et al. [43], who investigated reduction of NO by propene
over Na-promoted Pd/YSZ catalysts. The authors reported that
optimal results were obtained for catalyst promoted with
0.068 wt% Na (Na:Pd = 0.67), the activity of which was about 10
times higher than that of unpromoted Pd/YSZ. This was accompa-
nied by a significant increase in Ea from 84 to 160 kJ molꢀ1 with a
concomitant pronounced increase in the pre-exponential factor
[43]. Thus, alkali promoters seem to be able to affect both the
pre-exponential rate term (frequency factor) and the activation en-
ergy of the reaction through the number and strength, respectively,
of catalytically active sites.
served that addition of 0.06 wt.% Na (Na:Ru = 0.5) on the TiO2 sup-
port results in a shift of the CO conversion curve toward lower
temperatures by ca. 60 °C (Fig. 7A) and that the reaction rate (per
gram of catalyst) at 250 °C increases by a factor of 4 (Fig. 7B), com-
pared to unpromoted Ru/TiO2. Further increase of Na:Ru ratio to
1.8 improves catalytic performance slightly. The same trend is ob-
served for TOF (Table 1).
Regarding the least active Pd/TiO2 catalyst, addition of 0.34 wt.%
Cs (Cs:Pd = 0.5) results in a shift of the CO conversion curve toward
lower temperatures by ca. 45 °C (Fig. 8A) and in an increase in reac-
tion rate by a factor of 3 (Fig. 8B, Table 1). Further increasing Cs:Pd
atomic ratio to 1.1 has the opposite effect.
Results of H2-TPD experiments presented in Figs. 1–3 indicate
that adsorption sites related to the MT peak are those that are most
strongly influenced by the presence of the promoters. It is there-
fore reasonable to suggest that these sites, proposed to be located
at the metal-support interface, are the active sites for the WGS
reaction. In the following, an attempt is made to explain the pres-
ent results based on the influence of the promoters on the number
and chemisorption strength of these sites.
4. Discussion
Results of catalytic performance tests and kinetic measure-
ments (Figs. 4–8 and Table 1) show that the WGS activity of NM/
TiO2 catalysts depends strongly on the type and loading of alkali
promoters. A similar, but less pronounced, effect is observed for

P. Panagiotopoulou, D.I. Kondarides / Journal of Catalysis 267 (2009) 57–66
63
100
80
60
40
20
0
100
(A)
(A)
80
60
Cs:Pd ratio
0.0
0.5
40
1.1
Na:Ru ratio
0.0
0.5
1.8
20
0
200
250
300
350
400
450
200
250
300
350
400
450
Temperature (oC)
Temperature (oC)
1E-4
1E-5
1E-6
1E-4
1E-5
1E-6
(B)
(B)
Na:Ru ratio
0.0
0.5
1.8
Cs:Pd ratio
0.0
0.5
1.1
1.8
2.0
1000/T (K-1
2.2
2.4
1.8
2.0
1000/T (K-1)
2.2
2.4
)
Fig. 8. Effects of Cs loading on (A) the catalytic performance and (B) the intrinsic
Fig. 7. Effects of Na loading on (A) the catalytic performance and (B) the intrinsic
reaction rate of Pd/Cs–TiO2 catalysts.
reaction rate of Ru/Na–TiO2 catalysts.
on TiO2 are able to reduce the substrate and themselves become
oxidized according to the reaction:
4.1. The nature of catalytically active sites
Results of our previous studies clearly showed that the WGS
activity of dispersed noble metal catalysts is closely related to
the reducibility of the metal oxide support [22–26]. For instance,
activity of Pt and Ru catalysts was found to be 1–2 orders of mag-
nitude higher when supported on ‘‘reducible” metal oxides, such as
TiO2 and CeO2 than ‘‘irreducible” metal oxides, such as Al2O3 and
SiO2. Regarding the nature of catalytically active sites, results of
DRIFTS experiments obtained over Pt/TiO2 catalyst with the use
of CO as probe molecule indicated the existence of a special kind
of sites with exceptional electron-donating properties, located at
the metal-support interface [26,27]. Based on the results of 18O-
isotopic transient experiments, it has been proposed that these
sites are of the form of Pt–hs–Ti3+, where hs denotes an oxygen de-
fect vacancy [44]. It is therefore of interest to examine the way that
these sites can be created over NM/TiO2 catalysts and the possible
effects of alkali on their number and chemisorption strength.
A prerequisite for the creation of a new type of sites at the
metal-support interface (such as Pt–hs–Ti3+) is interaction of
dispersed metal crystallites with the metal oxide support. As dis-
cussed in detail by Diebold [45], certain metal atoms (M) deposited
M þ TiO2 ! MOx þ TiO2ꢀx
ð1Þ
Thermodynamic considerations imply that reaction (1) may take
place only when the standard free energy of the oxide (MOx) forma-
tion per mol of oxygen,
D D
Hof , is more negative than Hfo of TiO2
reduction to Ti2O3 (or TiO):
1
2
2TiO2 ! 2Ti2O3 þ O2ðgÞ
D
Hof ¼ 364 kJ molꢀ1
ð2Þ
The above requirement is satisfied for transition metals on the left
side of the periodic table, and also for alkali and alkaline earth met-
als. In contrast, noble metals such as Pt, Ru, and Pd are not expected
to interact strongly with stoichiometric TiO2 [45].
On the other hand, it is well known that dispersed noble metal
crystallites may interact with partly reduced titania support. For
instance, results of XPS and UPS experiments obtained over Pt sup-
ported on TiO2(110) showed that a localized electronic charge
transfer occurs from Ti3+ states to Pt clusters in the presence of sur-
face defects [46]. Thus, the presence of Ti3+ species in the close
neighborhood of dispersed noble metal crystallites seems to be a

64
P. Panagiotopoulou, D.I. Kondarides / Journal of Catalysis 267 (2009) 57–66
necessary requirement for the creation of a new type of sites at the
metal-support interface and for the onset of electronic-type inter-
actions between the metal and the support. Such sites can be cre-
ated in several ways on the surface of TiO2, including (i) partial
reduction with hydrogen at elevated temperatures, which is re-
lated to strong metal-support interactions (SMSI effect), (ii) heat
treatment at sufficiently high temperature to remove oxygen from
the support, (iii) photochemical reduction, and (iv) deposition of an
active metal, such as alkali or alkaline earth metal, according to
reaction (1).
tents, the population of hydrogen adsorbed at the metal-support
interface increases due to the creation of a higher amount of hs–
Ti3+ defects and, concomitantly, of a higher amount of NM–hs–
Ti3+ adsorption sites. However, for higher alkali loadings the
amount of adsorbed hydrogen at the metal-support interface de-
creases due to substantial weakening of the bond strength, which
is the case for all NM–X combinations investigated (Fig. 3, and
Ref. [27]).
Creation of new active sites at the metal-support interface has
been often used to account for the relatively high activity of
TiO2-supported metal catalysts [49,50]. For instance, Bracey and
Burch [49], who investigated the activity of Pd/TiO2 and Pd/SiO2
catalysts for the CO/H2 reaction, concluded that the high activity
of Pd/TiO2 is due to the creation of new active sites at the interface
between the metal and the support. The authors proposed that the
active site is a Ti3+ cation exposed in the surface of the support,
adjacent to a normal metal particle. It was proposed that these
new sites have the capacity to adsorb or assist in the adsorption
of CO. Similar conclusions were obtained for Ni/TiO2 catalysts
[50]. Similarly, Marcelin et al. [51], who investigated adsorption
of hydrogen over Cs- and Li-promoted Rh/TiO2 catalysts, proposed
that addition of alkalis results either in changes in the energetics of
the active sites or in the formation of new sites on the support or at
the rhodium–titania interface [51].
Alkali promotion of TiO2 may be also discussed in terms of
strong metal-support interactions (SMSI), first reported by Tauster
et al. [52]. The SMSI effect can be induced by high temperature
reduction treatment of TiO2-supported catalysts and is believed
to result in (a) creation of exposed Ti3+ species, which interact
strongly with dispersed metal crystallites, and/or (b) migration of
reduced suboxide species on the surface of metal crystallites
[52–54]. At low reduction temperatures, the SMSI effect may be
considered to be purely electronic [55], i.e., it is not accompanied
by decoration of noble metal particles by TiOx suboxides. In this re-
spect, promotion by alkalis induces the same effects on the proper-
ties of NM/TiO2 catalysts as does reduction temperature in SMSI,
i.e., creation of hs–Ti3+ defects. The number of these sites may be
controlled by varying either dopant content (alkali promotion) or
reduction temperature (SMSI). Thus, the observed effect of alkali
on the chemisorptive and catalytic properties of NM/TiO2 catalysts
may be viewed as a ‘‘permanent” SMSI effect, which can be used to
manipulate the number and chemisorption strength of sites lo-
cated at the metal-support interface.
Based on the above discussion, the effects of alkali on the chem-
isorptive and catalytic properties of the present NM/X–TiO2 cata-
lysts may be understood, at least in part, by considering the
reactivity of the alkali promoters toward oxygen of the TiO2 sup-
port. In particular, it is proposed that added alkali interacts
strongly with titania thereby leading to partial reduction of TiO2
(according to reaction (1)) and creation of hs–Ti3+ sites. This argu-
ment is supported, for example, by results of Onishi et al. [47] who
reported that Na deposited on TiO2(110) interacts strongly with
surface oxygen atoms thereby causing charge transfer to the sub-
strate and reduction of Ti4+ to Ti3+. If this is the case, then noble
metal atoms located at the perimeter of dispersed metal crystal-
lites are expected to be influenced indirectly by the presence of
the promoter via the creation of NM–hs–Ti3+ sites at the metal-
support interface. The observation that added alkali do not affect,
practically, the characteristics of the LT H2-TPD peak (Figs. 1 and
3) implies that the effect is of short-range and does not influence
adsorption on the surface of noble metal crystallites, at least in
the range of alkali loadings used in the present study. This can
be explained by considering that, because of the high relative
dielectric constant of TiO2 (ꢂ130), the charge located at the inter-
face will be about 130 times larger than the charge at the free me-
tal surface [48]. If one takes into account the attractive
contribution of the positively charged depletion region in the semi-
conductor, it may be concluded that the charge transfer at the free
metal surface will be negligible [48]. This argument is supported
by FTIR results of our previous study [27], which showed that
the stretching frequency and the relative intensity of bands as-
signed to linearly adsorbed CO on Pt crystallites (bands at 2089
and 2062 cmꢀ1) are not affected by the presence of added alkali
on the TiO2 support.
4.2. Effects of alkali on the population and chemisorption strength of
active sites
4.3. The volcano-type dependence of WGS activity on alkali loading
The presence of oxygen vacancies on TiO2 is expected to result
in electron transfer from the support to noble metal atoms located
in their close vicinity, i.e., at the metal-support interface. This
should result in weakening of the adsorption strength of these sites
toward electron donors, such as hydrogen, and in strengthening of
their adsorption strength toward electron acceptors, such as CO
[12]. Results of our present (Figs. 1–3) and previous [27] studies
clearly show that this is the case for all NM/X–TiO2 catalysts
investigated.
It is of interest to note here that weakening of the bond strength
of hydrogen adsorbed at the metal-support interface with increas-
ing alkali content should be accompanied by a decrease in the pop-
ulation of this species. However, results obtained for Pd/Cs–TiO2
(Fig. 3A), Ru/Na–TiO2 (Fig. 3B), and Pt/Cs–TiO2 [27] catalysts show
that this is not the case for low alkali loadings. In particular, the
intensity of the MT H2-TPD peak initially increases with increase
of alkali loading, goes through a maximum at a certain X:NM ratio,
and then decreases for higher promoter contents (Fig. 3). This pro-
vides evidence that the presence of alkali promoters on the support
results in the creation of additional adsorption sites at the metal-
support interface. It may then be argued that, for low alkali con-
In practically all noble metal–alkali combinations investigated
here, the reaction rate initially increases with increase of promoter
loading, goes through a maximum at a certain dopant level, and
then decreases for higher alkali contents (Figs. 5–8 and Table 1).
A similar dependence of catalytic activity on alkali loading has
been reported in several studies [8,9,43]. For instance, Kazi et al.
[8] reported that the rate of CO hydrogenation over Li-doped Pd/
SiO2 catalysts increases by a factor of 2 for samples with Li:Pd ratio
of 1, relative to the unpromoted catalyst, but higher levels of Li
(Li:Pd P 2) give reduced activities [8]. Huang et al. [9], who inves-
tigated CO methanation over alkali-promoted Ni/SiO2-Al2O3 cata-
lysts, found that TOF increases with increase of Na content and
passes through a maximum at 0.2% Na with an enhancement by
a factor of 6, compared to the sodium-free catalyst. The suppres-
sion of the promoting effect at higher Na levels has been attributed
to a kind of ‘‘poisoning effect” induced by the alkali, either directly
via the metal or metal-support interface, or indirectly via the sup-
port itself [9]. A volcano-type dependence of the reaction rate on
sodium loading has also been reported by Yentekakis et al. [43],
who investigated reduction of NO by propene over Pd/YSZ cata-

P. Panagiotopoulou, D.I. Kondarides / Journal of Catalysis 267 (2009) 57–66
65
lysts. The presence of a maximum in the reaction rate has been
attributed to poisoning due to site-blocking above some optimum
alkali loading (0.068 wt% Na) [43].
difference is that Grenoble et al. altered the strength of CO chemi-
sorption by varying the nature of the dispersed metallic phase,
whereas in the present study this has been achieved by varying
the nature and/or loading of alkali promoters on the TiO2 support.
The volcano-type dependence of reaction rate on alkali loading,
which is often observed for chemically and electrochemically pro-
moted catalytic materials, can be understood by taking into ac-
count the effects of the promoter on the chemisorptive bonds of
electron donor and electron acceptor adsorbed species [56].
Regarding the WGS reaction, Grenoble et al. [57] found that a vol-
cano-type relation exists between activity of Al2O3-supported me-
tal catalysts and their respective CO heats of adsorption. In this
respect, it is tempting to examine if such a correlation exists be-
tween the adsorption strength of NM–hs–Ti3+ sites toward CO
and the WGS activity of NM/X–TiO2 catalysts. For this purpose,
one may use the desorption temperature of the MT H2-TPD peak,
which provides a measure of the adsorption strength of active sites
toward hydrogen. It should be reminded here that, although
chemisorption strength toward hydrogen decreases with increase
of alkali loading (Figs. 1–3), the opposite is true for CO adsorption
[27].
4.4. Mechanistic implications
The mechanistic schemes proposed so far for the WGS reaction
over metal oxide-supported NM catalysts can be divided into two
broad categories, namely, redox and associative [58]. Redox mech-
anisms involve production of CO2 via oxidation of CO adsorbed at
the metal-support interface with an oxygen atom of the support,
which is subsequently reoxidized by H2O to liberate H2 [59–61].
Associative mechanisms involve steps where CO adsorbed on the
noble metal interacts with hydroxyl groups of the support to form
an intermediate species (e.g. formate) which further decomposes
to CO2 and H2 [62–64]. A hybrid mechanism, the so called ‘‘associa-
tive formate route with redox regeneration” of the oxide support,
has also been proposed [65–67], in which conversion of surface
formate to H2 and CO2 occurs using oxygen from the support and
not from water.
In Fig. 9 is plotted the turnover frequency of CO conversion at
250 °C (data taken from Table 1) as a function of Tmax (data taken
from Fig. 2) for Pt/X–TiO2 catalysts (X = Li, Na, K, Cs) of variable al-
kali loading. It is observed that reaction rate exhibits a volcano-
type dependence on the chemisorption strength of Pt–hs–Ti3+ sites
toward hydrogen. In particular, addition of small amounts of alkali,
which results in weakening of hydrogen adsorption strength (and
therefore strengthening of the CO adsorption strength), results in
an increase of the WGS reaction rate. For higher amounts of alkali
promoters, the chemisorption strength of Pt–hs–Ti3+ sites for
hydrogen (CO) is further decreased (increased), but this results in
catalysts with lower activity. As a result, specific reaction rate goes
through a maximum for catalysts containing 0.06 wt.% Na or
0.10 wt.% K, which exhibit an intermediate strength for hydrogen
(and CO) adsorption. In both cases, the alkali:Pt atomic ratio is
equal to 1 (Table 1). It is of interest to note that data obtained
for all different types and loadings of alkali promoters fall on the
same curve, implying that the rate is determined to a large extent
by the chemisorptive properties of sites at the metal-support inter-
face. Thus, results of Fig. 9 are qualitatively similar to those re-
ported by Grenoble et al. [57] over M/Al2O3 catalysts. The
It is important to note that all mechanistic schemes involve
bifunctional catalysis, in the sense that the noble metal activates
CO whereas the metal oxide adsorbs and activates H2O [58–67].
In this respect, it is of interest to note that NM–hs–Ti3+ sites fulfill
the requirement for catalyzing this dual site process, since they
may be involved in adsorption/activation of both CO and H2O.
Regarding water, the presence of surface defects on the TiO2 sup-
port seems to be a prerequisite for its activation via defect-induced
water dissociation [68–70]. For instance, results of STM experi-
ments and DFT calculations [68,69] showed that, at low coverage,
water dissociation on the rutile (110) surface occurs exclusively
on oxygen vacancies in the surface layer. These vacancies have
been shown to dissociate H2O through the transfer of one proton
to a nearby oxygen atom, forming two hydroxyl groups for every
vacancy. Similar results have been reported for the most stable
anatase (101) surface [70], for which first-principles molecular-
dynamic simulations showed the existence of a mechanism for
thermodynamically favored spontaneous dissociation of water at
low coverages of oxygen vacancies. It was concluded that it is likely
that all water molecules near vacancy sites are dissociated at low
coverages [70]. The amount of water dissociation has been found
to be limited by the density of oxygen vacancies present on the
clean surface [69]. Thus, alkali-induced creation of hs–Ti3+ sites
should enhance the capacity of titania toward dissociative adsorp-
tion of water. The presence of oxygen vacancies adjacent to a metal
particle provides the necessary dual-function NM–hs–Ti3+ sites re-
quired for the WGS reaction to proceed.
Promoter
none
1.5
Li
Na
K
It should be noted that NM–hs–Ti3+ sites can also be formed in
the absence of alkali promoters because, under the reducing envi-
ronment of the WGS reaction, dispersed noble metals may reduce
the surface of TiO2 effectively. The same is true for other reducible
metal oxides supports. For instance, recent studies conducted over
Pt/CeO2 and Au/CeO2 catalysts [71–73] indicate that the role of the
dispersed metallic phase is not only restricted in providing sites for
CO adsorption, but also in modifying the properties of the support
by creating new active sites at the metal-support interface, namely,
oxygen vacancies within the support or partially reduced sites at
the metal-support interface. Jacobs et al. [64] recognized that, for
Pt/CeO2 catalysts, partial reduction of ceria is a necessary step in
both the redox and the formate mechanisms. In the formate mech-
anism, the reduction of ceria surface shell is necessary to generate
the bridging OH group active sites, whereas in the redox mecha-
nism, reduction of ceria is directly involved in the reaction mech-
anism [64]. Finally, Lefferts et al. [65–67] showed that the WGS
reaction proceeds over Pt/TiO2 catalysts via both the redox route
Cs
1.0
0.5
200
220
240
260
280
Tmax(oC)
Fig. 9. Volcano-type dependence of turnover frequency at 250 °C on the desorption
temperature (Tmax) of the medium temperature (MT) peak observed in H2-TPD
profiles of Pt/X–TiO2 catalysts (X = Li, K, Na, Cs).

66
P. Panagiotopoulou, D.I. Kondarides / Journal of Catalysis 267 (2009) 57–66
[18] H.P. Bonzel, A.M. Bradshaw, G. Ertl (Eds.), Physics and Chemistry of Alkali
Metal Adsorption, Elsevier, Amsterdam, 1989.
[19] D. Heskett, Surf. Sci. 199 (1988) 67.
[20] H.P. Bonzel, Surf. Sci. Rep. 8 (1987) 43.
[21] D.L. Trimm, Z.I. Önsan, Catal. Rev. 43 (2001) 31.
[22] P. Panagiotopoulou, D.I. Kondarides, J. Catal. 225 (2004) 327.
[23] P. Panagiotopoulou, D.I. Kondarides, Catal. Today 112 (2006) 49.
[24] P. Panagiotopoulou, D.I. Kondarides, Catal. Today 127 (2007) 319.
[25] P. Panagiotopoulou, J. Papavasiliou, G. Avgouropoulos, T. Ioannides, D.I.
Kondarides, Chem. Eng. J. 134 (2007) 16.
and the associative formate route with redox regeneration, both of
which require the presence of oxygen vacancies on the support.
That is, water activation can occur by both hydroxyl formation
and redox water splitting on oxygen vacancies present on the re-
duced TiO2 surface. The role of water is related to regeneration of
hydroxyl groups and simultaneous re-oxidation of TiO2 [65–67].
5. Conclusions
[26] P. Panagiotopoulou, A. Christodoulakis, D.I. Kondarides, S. Boghosian, J. Catal.
240 (2006) 114.
[27] P. Panagiotopoulou, D.I. Kondarides, J. Catal. 260 (2008) 141.
[28] K. Foger, J.R. Anderson, J. Catal. 54 (1978) 318.
[29] P.G. Menon, G.F. Froment, J. Catal. 59 (1979) 138.
[30] O. Alexeev, D.-W. Kim, G.W. Graham, M. Shelef, B.C. Gates, J. Catal. 185 (1999)
170.
[31] J.T. Miller, B.L. Meyers, F.S. Modica, G.S. Lane, M. Vaarkamp, D.C.
Koningsberger, J. Catal. 143 (1993) 395.
[32] F. Benseradj, F. Sadi, M. Chater, Appl. Catal. B 228 (2002) 135.
[33] D.I. Kondarides, X.E. Verykios, J. Catal. 174 (1998) 52.
[34] J. Rieck, A. Bell, J. Catal. 103 (1987) 46.
The chemisorptive properties and WGS-activity of NM/TiO2 cat-
alysts (NM = Pt, Ru, Pd) can be modified appreciably by addition of
small amounts of alkalis (X = Li, Na, K, Cs) on the TiO2 support prior
to dispersion of the noble metal. Added alkalis interact strongly
with the TiO2 surface and result in the creation of hs–Ti3+ defects
which, in turn, affect the chemisorptive properties of NM atoms lo-
cated at the metal-support interface. The so formed NM–hs–Ti3+
sites are proposed to be the catalytically active sites for the WGS
reaction, since they are capable of adsorption and activation of
both CO and H2O. Increasing alkali loading results in an increase
of the number of these sites and in a monotonic decrease (increase)
of their adsorption strength toward hydrogen (carbon monoxide).
A volcano-type relation exists between the turnover frequency of
CO and the chemisorption strength of NM–hs–Ti3+ sites toward
hydrogen (or CO). Optimal results are obtained for Na-promoted
Pt/TiO2 catalysts with Na:Pt = 1:1, the specific activity (TOF) of
which is about three times higher, compared to that of the unpro-
moted catalyst. The effects of alkali promotion on the chemisorp-
tive and catalytic properties of NM/TiO2 catalysts may be
described as a ‘‘permanent” SMSI effect.
[35] B. Ngamsom, N. Bogdanchikova, M.A. Borja, P. Praserthdam, Catal. Commun. 5
(2004) 243.
[36] V.H. Sandoval, C.E. Gigola, Appl. Catal. A 148 (1996) 81.
[37] C.-C. Hong, C.-T. Yeh, Mater. Chem. Phys. 20 (1988) 471.
[38] H.-Y. Lin, Y.-W. Chen, Thermochim. Acta 419 (2004) 283.
[39] J.C. Conesa, J. Soria, J. Phys. Chem. 86 (1982) 1392.
[40] K.G. Azzam, I.V. Babich, K. Seshan, L. Lefferts, Appl. Catal. A 338 (2008) 66.
[41] Y.T. Kim, E.D. Park, H.C. Lee, D. Lee, K.H. Lee, Appl. Catal. B 90 (2009) 45.
[42] X. Zhu, T. Hoang, L.L. Lobban, R.G. Mallinson, Catal. Lett. 129 (2009) 135.
[43] I.V. Yentekakis, R.M. Lambert, M.S. Tikhov, M. Konsolakis, V. Kiousis, J. Catal.
176 (1998) 82.
[44] C.M. Kalamaras, P. Panagiotopoulou, D.I. Kondarides, A.M. Efstathiou, J. Catal.
264 (2009) 117.
[45] U. Diebold, Surf. Sci. Rep. 48 (2003) 53.
[46] K.D. Schierbaum, S. Fischer, M.C. Torquemada, J.L. de Segovia, E. Román, J.A.
Martín-Gago, Surf. Sci. 345 (1996) 261.
[47] H. Onishi, T. Aruga, C. Egawa, Y. Iwasawa, Surf. Sci. 199 (1988) 54.
[48] T. Ioannides, X.E. Verykios, J. Catal. 161 (1996) 560.
[49] J.D. Bracey, R. Burch, J. Catal. 86 (1984) 384.
References
[50] R. Burch, A.R. Flambard, J. Catal. 85 (1984) 16.
[1] M. Konsolakis, I.V. Yentekakis, Appl. Catal. B. 29 (2001) 103.
[2] M. Konsolakis, I.V. Yentekakis, A. Palermo, R.M. Lambert, Appl. Catal. B 33
(2001) 293.
[3] G. Goula, P. Katzourakis, N. Vakakis, T. Papadam, M. Konsolakis, M. Tikhov, I.V.
Yentekakis, Catal. Today 127 (2007) 199.
[4] S.S. Mulla, N. Chen, L. Cumaranatunge, W.N. Delgass, W.S. Epling, F.H. Ribeiro,
Catal. Today 114 (2006) 57.
[5] G. Avgouropoulos, E. Oikonomopoulos, D. Kanistras, T. Ioannides, Appl. Catal. B
65 (2006) 62.
[51] G. Marcelin, J.E. Lester, S.F. Mitchell, J. Catal. 102 (1986) 240.
[52] S.J. Tauster, S.C. Fung, R.L. Garten, J. Am. Chem. Soc. 100 (1978) 170.
[53] S. Bernal, J.J. Calvino, M.A. Cuqui, J.M. Catica, C. Larese, J.A. Pérez Omil, J.M.
Pintado, Catal. Today 50 (1999) 175.
[54] G.L. Haller, D.E. Resasco, Adv. Catal. 36 (1989) 173.
[55] J.P. Belzunegui, J. Sanz, J.M. Rojo, J. Am. Chem. Soc. 112 (1990) 4066.
[56] C.G. Vayenas, S. Brosda, C. Pliangos, J. Catal. 203 (2001) 329.
[57] D.C. Grenoble, M.M. Estadt, D.F. Ollis, J. Catal. 87 (1981) 90.
[58] J. Barbier Jr., D. Duprez, Appl. Catal. B 3 (1993) 61.
[59] H. Cordatos, T. Bunluesin, J. Stubenrauch, J.M. Vohs, R.J. Gorte, J. Phys. Chem.
100 (1996) 785.
[6] H. Tanaka, S.-I. Ito, S. Kameoka, K. Tomishige, K. Kunimori, Appl. Catal. A 250
(2003) 255.
[7] C. Pedrero, T. Waku, E. Iglesia, J. Catal. 233 (2005) 242.
[8] A.M. Kazi, B. Chen, J.G. Goddwin Jr., G. Marcelin, N. Rodriguez, R.T.K. Baker, J.
Catal. 157 (1995) 1.
[60] S. Hilaire, X. Wang, T. Luo, R.J. Gorte, J. Wagner, Appl. Catal. A 215 (2001) 271.
[61] Y. Li, Q. Fu, M. Flytzani-Stephanopoulos, Appl. Catal. B 27 (2000) 179.
[62] T. Shido, Y. Iwasawa, J. Catal. 141 (1993) 71.
[9] C.P. Huang, J.T. Richardson, J. Catal. 51 (1978) 1.
[10] H.N. Evin, G. Jacobs, J. Ruiz-Martinez, G.A. Thomas, B.H. Davis, Catal. Lett. 120
(2008) 166.
[11] J.M. Pigos, C.J. Brooks, G. Jacobs, B.H. Davis, Appl. Catal. A 328 (2007) 14.
[12] C.G. Vayenas, S. Bebelis, C. Pliangos, S. Brosda, D. Tsiplakides, Electrochemical
Activation of Catalysis, Kluwer Academic Publishers/Plenum Press, New York,
2001.
[13] F.J. Williams, M.S. Tikhov, A. Palermo, N. Macleod, R.M. Lambert, J. Phys. Chem.
B 105 (2002) 2800.
[14] F. Dorado, A. de Lucas-Consuegra, P. Vernoux, J.L. Valverde, Appl. Catal. B 73
(2007) 42.
[63] G. Jacobs, L. Williams, U. Graham, G.A. Thomas, D.E. Sparks, B.H. Davis, Appl.
Catal. A 252 (2003) 107.
[64] G. Jacobs, P.M. Patterson, U.M. Graham, A.C. Crawford, B.H. Davis, Int. J.
Hydrogen Energy 30 (2005) 1265.
[65] K.G. Azzam, I.V. Babich, K. Seshan, L. Lefferts, J. Catal. 251 (2007) 153.
[66] K.G. Azzam, I.V. Babich, K. Seshan, L. Lefferts, Appl. Catal. B 80 (2008) 129.
[67] P.O. Graf, D.J.M. de Vlieger, B.L. Mojet, L. Lefferts, J. Catal. 262 (2009) 181.
[68] I.M. Brookes, C.A. Muryn, G. Thornton, Phys. Rev. Lett. 87 (2001) 266103.
[69] R. Schaub, P. Thostrup, N. Lopez, E. Laegsgaard, I. Stensgaard, J.K. Norskov, F.
Besenbacher, Phys. Rev. Lett. 87 (2001) 266104.
[70] A. Tilocca, A. Selloni, J. Chem. Phys. 119 (2003) 7445.
[71] Q. Fu, H. Saltsburg, M. Flytzani-Stephanopoulos, Science 301 (2003) 935.
[72] Q. Fu, W. Deng, H. Saltsburg, M. Flytzani-Stephanopoulos, Appl. Catal. B 56
(2005) 57.
[15] Ch. Karavasilis, S. Bebelis, C.G. Vayenas, J. Catal. 160 (1996) 205.
[16] A. de Lucas-Consuegra, F. Dorado, J.L. Valverde, R. Karoum, P. Vernoux, J. Catal.
251 (2007) 474.
[17] I.V. Yentekakis, G. Moggridge, C.G. Vayenas, R.M. Lambert, J. Catal. 146 (1994)
292.
[73] A. Goguet, F. Meunier, J.P. Breen, R. Burch, M.I. Petch, A.F. Chenciu, J. Catal. 226
(2004) 382.