Visible light photocatalytic degradation of nitric oxides on PtOx-modified TiO2 via sol-gel and impregnation method

Article abstract of DOI:10.1016/j.molcata.2009.10.015

The visible light active catalysts, PtO_x-doped TiO₂ (PtO_x-TiO₂) and PtO_x-loaded TiO₂ (PtO_x/TiO₂), were successfully synthesized by the acid-catalyzed sol-gel process and the impregnation method. Pt(NH₃)₄(NO₃)₂ or H₂Pt(OH)₆ was employed as the PtO_x precursor. By comparing the results of De-NO_x, the modified photocatalysts exhibited a higher visible-light-responsive activity, and a lower NO₂ selectivity than the unmodified ones. The FE-SEM images suggested that the particle size was unchanged after modification. The XRD patterns showed that the crystal structure still remained as anatase phase. Nitrogen adsorption revealed no significant change in surface areas for all samples. The UV-vis spectra indicated that PtO_x promoted the absorption of visible light. Furthermore, the XPS spectra evidenced that the mixed valence states of PtO-PtO₂ coexisted on the surface of TiO₂. The adding of PtO_x on TiO₂ not only promoted the visible-light-responsive activity of converting NO to NO₂ but also increased the consecutive reaction rate of NO₂ to NO₃^-.

Full text of DOI:10.1016/j.molcata.2009.10.015

Journal of Molecular Catalysis A: Chemical 316 (2010) 163–170
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Visible light photocatalytic degradation of nitric oxides on PtOx-modified TiO2 via
sol–gel and impregnation method
Chun-Hung Huanga, I-Kai Wanga,∗, Yu-Ming Linb,∗∗, Yao-Hsuan Tsengc, Chun-Mei Lud
a Department of Chemical Engineering, National Tsing-Hua University, Hsinchu 30043, Taiwan, ROC
b ITRI South, Industrial Technology Research Institute, Rm. 602, Bldg. 3, 31 Gongye 2nd Rd., Annan District, Tainan 70955, Taiwan, ROC
c Department of Chemical Engineering, National Taiwan University of Science and Technology, Taipei 10607, Taiwan, ROC
d Department of Chemical and Materials Engineering, National Chin-Yi University of Technology, Taichung 41101, Taiwan, ROC
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 3 August 2009
Received in revised form 6 October 2009
Accepted 9 October 2009
Available online 17 October 2009
The visible light active catalysts, PtOx-doped TiO2 (PtOx–TiO2) and PtOx-loaded TiO2 (PtOx/TiO2),
were successfully synthesized by the acid-catalyzed sol–gel process and the impregnation method.
Pt(NH3)4(NO3)2 or H2Pt(OH)6 was employed as the PtOx precursor. By comparing the results of De-
NOx, the modified photocatalysts exhibited a higher visible-light-responsive activity, and a lower NO2
selectivity than the unmodified ones. The FE-SEM images suggested that the particle size was unchanged
after modification. The XRD patterns showed that the crystal structure still remained as anatase phase.
Nitrogen adsorption revealed no significant change in surface areas for all samples. The UV–vis spectra
indicated that PtOx promoted the absorption of visible light. Furthermore, the XPS spectra evidenced
that the mixed valence states of PtO–PtO2 coexisted on the surface of TiO2. The adding of PtOx on TiO2
Keywords:
Platinum oxide
Titanium dioxide
Visible light
Photocatalysis
Nitrogen oxides
not only promoted the visible-light-responsive activity of converting NO to NO2 but also increased the
−
consecutive reaction rate of NO2 to NO3
.
© 2009 Elsevier B.V. All rights reserved.
1. Introduction
One approach to achieve the goal is non-metal doping. Sato [7]
discovered the red-shift phenomena of a N-promoted TiO2. After-
ward, Asahi et al. [8] reported that a small amount of substitutional
N atoms onto the TiO2 structure by the sputtering method would
induce the Ti3+ sites on the TiO2 surface and significantly narrow
the band gap. Nakamura et al. [9] found that TiO2 treated with
hydrogen plasma can generate oxygen vacancies and create sub-
band gap to promote visible light responses, despite the expensive
appliances. Organic dyes as the light sensitization agents were also
considered to be added on the TiO2 surface to improve the light effi-
ciency [10,11]. However, the consumption of dye molecules during
the reaction process obstructed the application.
Another approach is doping transitional metals into TiO2 struc-
ture, such as Cr, V, Mn, Fe, Cu, or others [3,12–15]. However,
the doping metal increases carrier-recombination rates on TiO2
and reduces the thermal stability. Thus, metal-doped TiO2 usu-
ally shows an increased visible light activity at the expense of UV
activity. Recently, noble metals have been paid more attention as
dopants for TiO2 modification. Ohno et al. [16] demonstrated that
the photo-oxidation of water takes place on the Ru-loaded TiO2
powder under visible light. The stronger ability of electron cap-
ture by noble metals is postulated to extend the lifetime of the
electron–hole pairs, hinder the recombination of them, and hence
indirectly enhance the photocatalytic activity. Thus, the metallic
Pt-loaded TiO2 prepared by photodeposition has an improved UV
activity in spite of no visible response [17,18]. However, Vorontsov
Since Fujishima and Honda’s [1] discovery of the photo-induced
water splitting on TiO2 electrodes, photocatalysis based on pow-
dered semiconductors has received much attention for the effective
utilization of solar energy. Among all semiconductors, TiO2 is the
most widely studied photocatalyst, due to its strong redox abil-
ity, chemical stability, non-toxicity, and low cost. Recently, TiO2
is extensively applied to environmental pollutant control, such as
De-NOx, De-VOCs, and water purification [2–6]. Unfortunately, the
band gap of TiO2 is 3.2 eV, corresponding to wavelengths of shorter
than 388 nm. This fatal drawback limits the photocatalytic practice
in the UV light region, which only counts for 4% of the solar spec-
trum, while the main part (45%) belongs to the visible light region.
To overcome this shortcoming, many researchers have made much
effort on the modification of TiO2 to utilize sunlight more effectively
and to further expand the indoor applications.
∗
Corresponding author at: Department of Chemical Engineering, National Tsing-
Hua University, #101, Sec. 2, Kung-Fu Rd., Hsinchu 30043, Taiwan, ROC. Tel.: +886
3 5713763; fax: +886 3 5724725.
∗∗
Corresponding author. Tel.: +886 6 3847432; fax: +886 6 3847288.
E-mail addresses: ikwang@che.nthu.edu.tw (I.-K. Wang), YuMingLin@itri.org.tw
(Y.-M. Lin).
1381-1169/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2009.10.015

164
C.-H. Huang et al. / Journal of Molecular Catalysis A: Chemical 316 (2010) 163–170
Fig. 1. Scheme of synthesis procedure.
et al. [19] reported that the photocatalytic activities depend on
Pt states. Kisch and Macyk [20] have also found that the Pt com-
plex doped TiO2 promotes the visible activities. The oxidized state
of Pt was confirmed to facilitate visible response and was more
thermally stable during heat treatment process than metallic Pt.
Based on the economic consideration, our previous research [21]
demonstrated that the visible-responsive photocatalyst, Pt/TiO2,
can be easily synthesized by sol–gel process or by impregnation
method to load PtOx on TiO2. In this study, we planned to further
improve the visible activities of photocatalysts and to investigate
the relationship of the chemical states of PtOx and the photo-
activities.
For the purpose of doping Pt complex into the TiO2 matrix, Pt2
or Pt4 with a desired atomic ratio of Pt/Ti = 1/100 was added during
the acid-catalyzed sol–gel process. The synthesized photocatalysts
were labeled as Pt2-TiO2 and Pt4-TiO2. On the other hand, impreg-
nation method was employed to load PtOx on the surface of TiO2.
TiO2 powder (either UV100 or home made TiO2) was added into
a Pt2 or Pt4 solution with an atomic ratio of Pt/Ti = 1/100. After
stirring for 1 h, all the samples were dried at 110 ◦C for 4 h and
then calcined at 200 ◦C for 5 h. Four modified photocatalysts by
the impregnation method were labeled as Pt2/UV100, Pt4/UV100,
Pt2/TiO2 and Pt4/TiO2. In addition, the Pt2/UV100 was reduced in a
hydrogen flow with a flow rate of 50 mL/min at 200 ◦C for 3 h, and
the resulting catalyst was labeled as Pt0/UV100.
The above-mentioned procedures are depicted in Fig. 1. Samples
were divided into two groups, the UV100 series (UV100, Pt2/UV100,
Pt4/UV100 and Pt0/UV100) and the TiO2 series (TiO2, Pt2-TiO2,
Pt2/TiO2, Pt4-TiO2 and Pt4/TiO2). The symbols of “-” and “/” repre-
sent the modification via the sol–gel method and the impregnation
method.
2. Experimental
2.1. Preparation of photocatalysts
Two kinds of titania (TiO2) were used as the raw materials
for modification. One was commercially available TiO2, Hombikat
UV100 (Sachtleben Chemie, 100% anatase), and the other was syn-
thesized by acid-catalyzed sol–gel process as described elsewhere
[21]. Two types of platinum-containing compounds, tetra-amine
platinum nitrate (Pt(NH3)4(NO3)2, symbolized by Pt2) and hydro-
gen hexa-hydroxy platinate (H2Pt(OH)6, symbolized by Pt4), were
used to prepare 1% PtOx/TiO2.
2.2. Degradation of NOx on photocatalyst
Photocatalytic degradation of nitrogen oxides (De-NOx) was
carried out by the setup consisting of a continuous flow reactor
connected to the gas suppliers and the analytic system, as shown
Fig. 2. Schematic diagram of De-NOx system.

C.-H. Huang et al. / Journal of Molecular Catalysis A: Chemical 316 (2010) 163–170
165
radicals, was proposed as follows [2]:
NO + OH• ꢀ HNO2(H+ + NO−2 )
HNO2 + OH• ꢀ NO2 + H2O
(1)
(2)
(3)
NO2 + OH• ꢀ HNO3(H+ + NO−3 )
In which, HNO2 is a minor species and can be reversed to NO or
oxidized to NO2. Consequently, NO oxidation can be simplified in a
−
serial mechanism of NO → NO2 → NO3 over photocatalyst under
illumination. The effluent concentrations of NO and NO2 from the
reactor were measured by a NOx analyzer (Eco Physics, CLD 700AL).
The NO3− ion was adsorbed on the photocatalyst surface, and could
be washed out by water. In data analysis, the total NOx concentra-
tion ([NOx]), the NO conversion, the NO2 selectivity, and the total
NOx removal are defined as follows:
Fig. 3. Photon energy distribution of UV lamp and LEDs.
[NOx] = [NO] + [NO2]
(4)
(5)
[NO]in − [NO]out
NO conversion =
NO2 selectivity =
[NO]in
in Fig. 2. To simulate atmospheric environment and reach the opti-
mal value, the relative humidity was fixed at 50% by adjusting
a humidified air stream and a dry air stream [2]. Then, a nitric
oxide (NO) gas stream, provided by a cylinder containing 100 ppmv
NO (N2 balance), was added to the air stream to reach a desired
NO concentration of 1 ppmv and a total volumetric flow rate of
1 L/min. The photocatalyst (0.5 g) was added into an appropriate
amount of D.I. water contained in a 50 mL flask, followed by well
stirring. The mixed slurry was then slowly spread on a disk-like
frosted glass (diameter = 12 cm). After drying at 110 ◦C for 30 min
to remove water and remain sample powder on, the frosted glass
coated with the photocatalyst was loaded into the reactor. A black
light lamp over the reactor provided the ultraviolet illumination in
the wavelength range of 330–420 nm. Light-emitting diodes (LEDs)
provided the visible light illumination including blue LEDs (BLED,
430–530 nm), green LEDs (GLED, 470–570 nm), and red LEDs (RLED,
590–680 nm). The photon energy distribution profiles of the four
irradiation sources are shown in Fig. 3. The intensity of the light
was kept at 1 mW/cm2, measured by a spectrophotometer (Ocean
Optics, USB2000). Blank test was carried out under the same condi-
tion excluding the photocatalyst powder adding. The result ensured
all the materials being inactive on photocatalytic De-NOx.
[NO2]out
(6)
(7)
[NO]in − [NO]out
[NO]in − [NOx]out
total NOx removal =
[NO]in
Since NO2 is more toxic than NO [22,23], the photocatalyst per-
formance should be evaluated by the total NOx removal, or the NO
conversion conjunctive with the NO2 selectivity.
2.3. Characterization and analysis
The particle sizes and the morphology of the synthesized
samples were determined by a field-emission scanning electron
microscopy (FE-SEM, Hitachi S4800-1).
The crystal structures were analyzed by an X-ray powder diffrac-
tometry (XRD, Rigaku RU-H3R) in the reflection mode with Cu K␣
radiation. The angular domain was 20–80◦ (2ꢀ). In addition, the
grain sizes (d) were calculated by Scherrer’s equation, as follows:
0.9ꢁ
ˇ cosꢀ
d =
(8)
Photocatalytic oxidation of NO on TiO2 was investigated and
a reaction pathway based on consecutive reactions with hydroxyl
where ꢁ is the X-ray source wavelength of 1.5405 Å and ˇ is the full
width at half maximum of main peak.
Fig. 4. FE-SEM images of the photocatalysts with (a) UV100, (b) Pt2/UV100, (c) Pt4/UV100, (d) TiO2, (e) Pt2-TiO2, and (f) Pt2/TiO2.

166
C.-H. Huang et al. / Journal of Molecular Catalysis A: Chemical 316 (2010) 163–170
Table 1
The surface areas, atomic compositions and calculated atomic ratios for all samples.a.
Samples
SBET
(m2/g)
Atomic composition (%)
Atomic ratio
Ti
O
C
N
Pt
O/Ti
Pt/Ti
UV100
249
250
260
272
254
244
256
226
217
15.7
16.3
17.0
17.1
12.9
15.0
16.3
17.0
20.2
58.9
55.9
57.8
54.6
54.7
52.3
52.6
49.1
47.1
24.0
25.6
22.4
26.5
31.0
31.0
29.9
33.5
31.9
1.4
2.0
2.6
1.5
1.4
1.4
0.9
0.2
0.4
–
3.75
3.42
3.40
3.19
4.24
3.49
3.23
2.89
2.33
–
Pt0/UV100
Pt2/UV100
Pt4/UV100
TiO2
Pt2-TiO2
Pt2/TiO2
Pt4-TiO2
Pt4/TiO2
0.2
0.2
0.3
–
0.3
0.3
0.2
0.4
0.01
0.01
0.02
–
0.02
0.02
0.01
0.02
a
All samples were calcined at 200 ◦C.
(around 3–10 nm) and more dispersive, except for Pt2-TiO2, which
implied the morphology tended to be more agglomerative by
sol–gel method than that by impregnation method. This could be
rationalized that the platinic precursor was added during hydrol-
ysis of titanium alkoxide prior to crystalline phase formation. On
the other hand, the impregnation method, platinic precursor was
loaded on the well-formed crystalline TiO2 particles.
Fig. 5. XRD patterns of the photocatalysts.
The specific surface areas were measured by nitrogen adsorp-
tion at 77 K using a Micrometrics model ASAP 2000 instrument. All
the samples were outgassed to a vacuum of <10−4 Torr at 200 ◦C for
2 h to remove any adsorbed impurities prior to each measurement.
Surface areas were calculated by BET method.
3.1.2. Crystal structure
The UV–vis absorption spectra of all samples were measured by
a powder UV–vis spectrophotometer (UV-VIS, Shimadzu UV-2450).
The obtained reflectance data (R) were converted to the absorbance
values, F(R), based on the Kubelka–Munk theory, as follows:
The XRD patterns are shown in Fig. 5. All the crystal structures
of samples were almost anatase, attributed to several main peaks
(2ꢀ = 25.3◦ with [1 0 1], 37.8◦ with [0 0 4], 48.1◦ with [2 0 0], 54.0◦
with [1 0 5], 62.7◦ with [2 0 4]). Since all samples were calcined at
a relatively low temperature of 200 ◦C and only a small amount of
PtOx (1%) was added in, there was no significant difference among
the samples. The PtOx was interpreted as well-dispersed by the
evidence of the absence of Pt signal.
According to Scherrer’s equation by introducing the full width at
half maximum of main peak with 2ꢀ = 25.3◦ for anatase phase, the
grain size is larger with the sharper main peak. The calculated result
revealed that the grain size of UV100 series was around 9–12 nm,
and that of TiO2 series around 5–6 nm. The results were consistent
with the FE-SEM images. Besides, the FE-SEM results also suggested
that there is no significant change in particle size before and after
PtOx loading.
(1 − R)2
F(R) =
(9)
2R
The surface atomic composition of the catalysts and the
chemical states of Pt were examined by an X-ray photoelectron
spectroscopy (XPS, Physical Electronics ESCA PHI 1600) with Mg
K␣ radiation. The charging effects were corrected by adjusting the
C 1s peak to a position of 284.5 eV.
3. Results and discussion
3.1. Characterization
3.1.1. Morphology
Fig. 4 shows the FE-SEM images of various photocatalysts. The
particle sizes of UV100 series (UV100, Pt2/UV100 and Pt4/UV100)
were similar, around 7–12 nm. It suggested that the primary par-
ticle sizes were unchanged after impregnating PtOx. The particle
sizes of TiO2 series (TiO2, Pt2-TiO2 and Pt2/TiO2) were smaller
3.1.3. Surface area
The surface areas of all samples, around 217–272 m2/g, are
listed in Table 1. For UV100 series, the three modified samples
of Pt0/UV100, Pt2/UV100 and Pt4/UV100 showed no significant
change or only a slight increase with 250, 260 and 272 m2/g, respec-
Fig. 6. (A) UV–vis absorption spectra and (B) plot of transformed Kubelka–Munk function versus the energy of the light absorbed for (a) UV100, (b) Pt0/UV100, (c) Pt2/UV100,
(d) Pt4/UV100, (e) TiO2, (f) Pt2-TiO2, (g)Pt2/TiO2, (h)Pt4-TiO2, and (i)Pt4/TiO2, with the direct band gap of (a) 3.26, (b) 3.18, (c) 3.20, (d) 3.15, (e) 2.65, (f) 2.87, (g) 2.82, (h)
2.75, and (i) 2.70 eV.

C.-H. Huang et al. / Journal of Molecular Catalysis A: Chemical 316 (2010) 163–170
167
tively, compared to the sample of UV100 (249 m2/g). This was
reasonable that the small amount metallic Pt or PtOx could be well-
dispersive on crystalline TiO2 powder by impregnation method.
The home made TiO2 possessed the surface area of 254 m2/g, similar
to UV100. This result was also corresponding to their similar parti-
cle sizes. After modification, the four samples of Pt2-TiO2, Pt2/TiO2,
Pt4-TiO2 and Pt4/TiO2 showed unchanged or some decrease in sur-
face areas with 244, 256, 226 and 217 m2/g, respectively.
3.1.4. UV–vis absorption
The UV–vis absorbance spectra are presented in Fig. 6A, used to
determine the light absorption initially. For UV100 series ((a)–(d)),
UV100 displayed no visible absorption due to its single and wide
band gap. After the treatment of loading metallic Pt or PtOx, each
of Pt0/UV100, Pt2/UV100 and Pt4/UV100 exhibited small raise of
the visible absorption. On the other hand, the home made TiO2 had
evident absorption in the visible light region, which was attributed
to the carbonaceous residues during sol–gel process [24]. After
PtOx modification either by sol–gel process or by impregnation
method, the four PtOx-modified TiO2 (Pt2-TiO2, Pt2/TiO2, Pt4-TiO2
and Pt4/TiO2) showed decrease in visible absorption band. It is pos-
sible that some carbonaceous species in the PtOx-modified TiO2
samples were decomposed again during secondary calcination.
Despite of that, all the modified samples and the home made TiO2
showed significant visible light absorbance.
For semiconductor materials, the direct band gap can be
obtained by constructing Tauc Plot—(F(R) × hꢂ)1/2 versus hꢂ [25].
As shown in Fig. 6B, the interception (represented by the dotted
lines), which is the extension of the linear region until the base
line, gives the so-called Tauc optical band gap. It was obtained
that the band gap of UV100 (3.26 eV) was corresponding to the
wavelengths shorter than 380 nm with only UV light absorbance.
The modified samples of Pt0/UV100, Pt2/UV100 and Pt4/UV100
showed some red shift to 3.18, 3.20 and 3.15 eV, respectively.
The red shift qualitatively qualified the catalyst to be activated
under visible light illumination. For the TiO2 series, unmodified
TiO2 had the relatively lowest band gap of 2.65 eV. PtOx-modified
TiO2 showed a slight blue shift, but still located in the visible
light region. To summarize, the direct band gaps of all samples
were in the order of TiO2 < Pt4/TiO2 < Pt4-TiO2 < Pt2/TiO2 < Pt2-
TiO2 < Pt4/UV100 < Pt0/UV100 < Pt2/UV100 < UV100.
Fig. 7. C 1s XPS spectra of all samples.
titania. The decreasing hydrophilic property was referred to the
existence of PtOx on the surface.
The ratios of Pt to Ti were higher than or equal to the antici-
pated ratio (1%). It was due to the surface sensitive property of XPS.
The result also implied that most PtOx tended to be on the titania
surface rather than in structure, whether synthesized by sol–gel or
impregnation method.
3.2.2. Analysis of chemical states
The chemical states of elements in materials could be provided
by XPS spectra. The Ti 2p XPS spectra of all samples represented
insignificant variation (almost Ti4+ signal), and the Ti3+ signal was
not found. The C 1s XPS spectra of all samples are demonstrated in
Fig. 7. It can be seen that UV100 series had only one carbon state
with the peak located at 284.5 eV corresponding to adventitious
carbon, while TiO2 series had two more shoulder peaks located at
287 and 288.5 eV, corresponding to C O and O C–O, respectively.
Sakthivel and Kisch [27] reported that the two kinds of carbonate
species could be observed via hydrolysis process followed by cal-
cinations. Li et al. [28] also mentioned that the carbon, in the form
of carbonate instead of substituting form, may narrow the band
gap and induce visible response to oxidation of gaseous benzene.
This also provides an explanation why the visible light absorbance
of TiO2 series was higher than that of UV100 series. And it could
be anticipated that home made TiO2 would exhibit more visible
activity than UV100 due to the carbonate residues.
3.2. XPS analysis of chemical elements and electronic structure
3.2.1. Atomic composition
XPS analysis revealed the surface constitution of the samples.
The atomic compositions and calculated atomic ratios of O/Ti and
Pt/Ti for all samples are listed in Table 1. The carbon content was
unexpectedly high, about 24–33.5%. It was confirmed that the most
carbon arose from adventitious element carbon corresponding to
the peak at 284.5 eV [26]. Noteworthily, the carbon percentages
of TiO2 series were all higher than those of UV100 series. The exis-
tence of carbonaceous species of home made TiO2 might be derived
from alkyl groups during sol–gel process and were not completely
removed out at a calcining temperature of 200 ◦C.
The small amount of nitrogen might come from the residual
impurity of UV100, the nitric acid for sol–gel process, and the Pt2
precursor (Pt(NH3)4(NO3)2) for modification. As the data revealed,
the samples modified by Pt4 had lower nitrogen amounts than
those modified by Pt2. Moreover, samples using home made TiO2
as raw materials also showed less nitrogen than those using UV100.
The atomic ratios of O to Ti for all samples were higher than
2, attributed by the hydrophilic property of TiO2, causing H2O
molecules to be chemisorbed on. It was also found that all the O/Ti
values of the modified samples were lower than those of the bare
Fig. 8. Pt 4f XPS spectra of various Pt loaded onto UV100.

168
C.-H. Huang et al. / Journal of Molecular Catalysis A: Chemical 316 (2010) 163–170
Fig. 10. Time dependency profile of NO, NO2 and NOx concentrations as a function of
illumination time on Pt4/TiO2 under UV, BLED, GLED and RLED illumination: catalyst
loading, 0.5 g; irradiation intensity, 1 mW/cm2; feed concentration of NO, 1 ppmv;
inlet flow rate, 1 L/min; relative humidity, 50%; and reaction temperature, 25 ◦C.
Fig. 9. The deconvolution of Pt 4f XPS spectra for Pt4/UV100.
The Pt 4f XPS spectra of Pt0/UV100, Pt2/UV100 and Pt4/UV100
are shown in Fig. 8. The binding energy at 71.2 eV indicated the
reduced state of Pt0 of Pt0/UV100. The nearly single Pt0 state was
only found on Pt0/UV100, attributed to the treatment of hydro-
gen reduction. For a single oxide valence state of Pt, the integral
intensity ratio of 4f7/2 to 4f5/2 should be 4:3 in theory. However,
our results showed mixed valence states of Pt on Pt2/UV100 and
Pt4/UV100 evidenced by the discordant ratio of integral intensity.
To determine them, peak deconvolution was employed by using
the binding energy of 4f7/2 for Pt0, Pt2+ and Pt4+ as 71.2, 72.5 and
74.8 eV, respectively. The deconvolution of Pt 4f XPS spectra for
Pt4/UV100 is presented as a representative example in Fig. 9, and
the results for the three samples are listed in Table 2. Pt0/UV100
contained mainly Pt0 (94.4%) and some Pt2+ (5.6%), indicating a
mainly metallic state. The Pt states of Pt2/UV100 were similar to
those of Pt4/UV100, major Pt2+ and minor Pt4+. It inferred that
PtOx was in the form of coexistent PtO–PtO2 on the TiO2 surface,
either using Pt2+ (from Pt(NH3)4(NO3)2) or Pt4+ (from H2Pt(OH)6)
ion as the precursor. It was reasonable as both of them were cal-
cined at the same temperature. Kisch and Macyk [20] reported
the behavior of TiO2 modified with PtCl4 in the bulk or on the
surface. The excited platinum complex underwent bond cleav-
age to afford electron transfer to TiO2 and resulted in the shift of
flatband potential. It also increases the driving force of electron
transfer from Pt3+ intermediate and therefore improves the activity
[20]. Consequently, the mixed valence states of PtOx at the sur-
face of TiO2 were regarded as a sensitizer responding to the visible
light.
This also implied that the residual species from Pt(NH3)4(NO3)2
did not form the substitutional N in the TiO2 structure as expected
and could not further cause visible light response.
3.3. Photocatalytic activity
To illustrate the behavior of photocatalytic De-NOx, 0.5 g powder
photocatalyst was used under 1 mW/cm2 intensity of various light
illumination (UV, BLED, GLED and RLED). The typical time depen-
dency profile of NO, NO2 and NOx concentrations is shown in Fig. 10.
NO with a concentration of 1 ppmv was introduced into the reac-
tor under dark condition. After a few minutes, as the adsorption
was reached to equilibrium, the light was turned on. The durability
of the catalyst was confirmed by repeating above procedures with
different light source as shown in Fig. 10.
It was reported that the oxidation reaction can−only go as
far as NO2 once the catalyst is saturated with NO3 [2]. Ohko
et al. [29] studied the reaction of NO on a HNO3 pre-deposited
TiO2 under UV light illumination and proposed a side reaction of
NO + NO3 → 2NO2. However, this side reaction may only occur as
the support was saturated with NO3−. This phenomenon was not
found in our case due to the NO2 concentration approximately
maintained a stable value, and the large surface area of the cat-
alyst (217–272 m2/g) was unable to be saturated with NO3− during
a few hours operation.
By Eqs. (5)–(7), the NO conversion (NOC), NO2 selectivity (NO2S)
and total NOx removal (NOxR) over various photocatalysts are
shown in Figs. 11 and 12. The NO conversion represents the oxi-
dation percentage of NO to NO2 and NO3−. The total NOx removal
expresses the total reaction percentage of producing NO3−, con-
sidered as the total oxidation ability. A well-active photocatalyst
should basically provide a high NO conversion as well as a low NO2
selectivity, resulting in a high total NOx removal. In other words,
with the same NO conversion, the one with lower NO2 selectivity
The N 1s XPS spectra of Pt2/UV100, not shown here, presented
two types of peaks. The major peak located at 400.5 eV was consid-
ered as nitrogen impurity originated from the bulk, and the minor
peak located at 407.4 eV was considered as nitrates attributed to the
Pt2 precursor (Pt2(NH3)4(NO3)2). It was reasonable that nitrates
could not be totally removed out after low temperature calcination.
Some research groups reported that the peak at 396 eV, derived
from the substitutional N (␤-N), was considered as the visible light
source [8]. However, ␤-N peak was not found in our all samples.
−
could catalyze the oxidation of NO2 to NO3 more completely.
The performances of the two bare TiO2 (UV100 (Fig. 11A) and
home made TiO2 (Fig. 12A)) were compared with each other under
UV light illumination. Although UV100 exhibited a higher NO con-
version (UV100 = 71%; TiO2 = 61%), the total NOx removal was less
than home made TiO2 (UV100 = 51%; TiO2 = 57%) due to the varia-
tions in NO2 selectivity. It implied the latter one had better ability
of total oxidation. For visible light activity, the home made TiO2
showed not only a higher NO conversion but also a higher total
NOx removal than UV100, especially in GLED and RLED. The better
visible activity of home made TiO2 was attributed to the carbonate
Table 2
The results of deconvolution from various Pt loaded onto UV100.
(%)
Pt 4f7/2
Pt0% (71.2 eV)
Pt2+ % (72.5 eV)
Pt4+ % (74.8 eV)
Pt0/UV100
Pt2/UV100
Pt4/UV100
94.4
–
–
5.6
82.7
88.2
–
17.3
11.8

C.-H. Huang et al. / Journal of Molecular Catalysis A: Chemical 316 (2010) 163–170
169
NO2 selectivity was also abated by the two samples compared to
bare UV100, even though the lowest NO2 selectivity was found on
Pt0/UV100. Overall, the result proved that PtO–PtO2 loaded TiO2
not only exhibited the ability of NO2 abatement, but also pro-
vided sufficient visible light activity. Pt2/UV100 and Pt4/UV100
possessed the similar Pt states (PtO–PtO2), but the former was less
active under visible light, caused by the different Pt precursors.
XPS spectra revealed that no any ␤-N peak was found. In other
words, the residual nitrates from Pt(NH3)4(NO3)2 precursor could
not be totally removed out and did not form the substitutional N in
the TiO2 structure as expected after low temperature calcination.
Therefore, the nitrates were helpless to visible light activity and
just acted as impurities. Compared to this, H2Pt(OH)6 was used as
the Pt precursor to prepare Pt4/UV100 and subsequently formed
PtO–PtOx on TiO2 surface without any other impurities. It may be
the reason why Pt4/UV100 showed the best De-NOx activity among
UV100 series whether under UV or visible illumination.
When using the home made TiO2 as raw material, as shown
in Fig. 12, the four modified samples (Pt2-TiO2 (b), Pt4-TiO2 (c),
Fig. 11. De-NOx activities on UV100 series consisting of (A) NO conversion, (B) NO2
selectivity and (C) total NOx removal for (a) UV100, (b) Pt0/UV100, (c) Pt2/UV100
and (d) Pt4/UV100 under different light illumination.
species residual from the alkyl oxide precursor and alcohol in the
sol–gel process, as evidenced in XPS result.
Fig. 11 gives the results of De-NOx over various photocata-
lysts with UV100 as the modifiable material. Pt0/UV100 (b), with
metallic Pt0 state, exhibited a higher UV total NOx removal than
UV100, but no significant improvement under visible light illumi-
nation. However, the low NO2 selectivity confirmed that Pt0/UV100
improved the total oxidation of NO to NO3−. Vorontsov et al. [19]
also reported that the photocatalytic activity of Pt/TiO2 depends
on the Pt states. Deconvolution of the XPS spectrum for Pt0/UV100
showed of the occurrence of mainly Pt0 state. The reduced state
of Pt functioned not only as the electron trap center but also as
the adsorption center of O2 to facilitate the total oxidation [30,31].
However, Pt0 state was helpless to visible light activity. On the
other hand, evident improvement of visible light activity on both
Pt2/UV100 and Pt4/UV100 (c and d) was observed. Additionally,
Fig. 12. De-NOx activities on TiO2 series consisting of (A) NO conversion, (B) NO2
selectivity and (C) total NOx removal for (a) TiO2, (b) Pt2-TiO2, (c) Pt4-TiO2, (d)
Pt2/TiO2 and (e) Pt4/TiO2 under different light illumination.

170
C.-H. Huang et al. / Journal of Molecular Catalysis A: Chemical 316 (2010) 163–170
Pt2/TiO2 (d) and Pt4/TiO2 (e)) offered around the same UV activ-
ity as the bare TiO2 (a). Under the illumination of visible light,
four modified samples seemed not to present noticeably higher
NO conversion than bare TiO2. It was reasonable that the home
made TiO2 performed well visible light activity resulted from the
carbonate species residual, as described in XPS result. However, the
four modified samples, except for Pt2-TiO2 under RLED, performed
higher total NOx removal under visible light illumination, due to
the decrease of NO2 selectivity. The results confirmed again that
the existence of PtOx would promote the total oxidation of NO to
NO3−. For Pt2-TiO2, even in the presence of PtOx, hydrolysis process
may cause the PtOx and unnecessary nitrate residues to be doped
in TiO2 matrix, and therefore reduce the visible efficiency.
in photocatalysis. It was evidenced that the mixed valence states
of PtO–PtO2 coexisted on the surface of TiO2. The PtO–PtO2 was
regarded not only as a sensitizer responding t−o the visible light but
also to facilitate the oxidation of NO2 to NO3 by suppression the
NO2 selectivity.
References
[1] A. Fujishima, K. Honda, Nature 238 (1972) 37.
[2] S. Devahasdin, C. Fan Jr., K. Li, D.H. Chen, J. Photochem. Photobiol. A: Chem. 156
(2003) 161.
[3] M.R. Hoffmann, S.T. Martin, W. Choi, D.W. Bahnemannt, Chem. Rev. 95 (1995)
69.
[4] A. Fujishima, T.N. Rao, D.A. Tryk, J. Photochem. Photobiol. C: Photochem. Rev.
1 (2000) 1.
In addition, Pt2/TiO2 and Pt4/TiO2 showed more activity than
Pt2-TiO2 and Pt4-TiO2 in visible region, suggesting that PtOx on the
TiO2 surface was more beneficial for visible response than in the
matrix. On the other hand, by using either UV100 or home made
TiO2 as the raw material, the relative conversion increased in the
order of unmodified one < Pt2 < Pt4. It suggested the platinic precur-
sor of H2Pt(OH)6 was more useful than Pt(NH3)4(NO3)2, because of
the latter would arise some nitrate impurities. It could be conclude
that the visible sensitizer of mixed PtO–PtO2 would function better
on TiO2 surface via impregnation method than in TiO2 matrix via
sol–gel method. And using H2Pt(OH)6 as the precursor was benefi-
cial to form PtO–PtO2 without other impurities.
To summarize, for UV100 series, photocatalytic activities
were in the order of UV100 < Pt2/UV100 < Pt4/UV100. For
TiO2 series, they were in the order of TiO2 < Pt2-TiO2 < Pt4-
TiO2 < Pt2/TiO2 < Pt4/TiO2. The Pt4/TiO2, prepared by the sol–gel
process and followed by impregnation with H2Pt(OH)6 as a PtOx
dopant, exhibited the highest visible activity among all. Even under
red light illumination, the total NOx removal could reach a level of
33%. The visible sensitizer of PtOx loaded on the TiO2 surface was
better than that doped in TiO2. Moreover, the particle size, surface
area and the crystal structure remained roughly unchanged with
adding only 1% PtOx.
[5] J.C.S. Wu, Y.T. Cheng, J. Catal. 237 (2006) 393.
[6] Y. Ohko, Y. Nakamura, A. Fukuda, S. Matsuzawa, K. Takeuchi, J. Phys. Chem. C
112 (2008) 10502.
[7] S. Sato, Chem. Phys. Lett. 123 (1986) 126.
[8] R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Science 293 (2001) 269.
[9] I. Nakamura, N. Negishi, S. Kutsuna, T. Ihara, S. Sugihara, K. Takeuchi, J. Mol.
Catal. A: Chem. 161 (2000) 205.
[10] C. Nasr, K. Vinodgopal, L. Fisher, S. Hotchandani, A.K. Chattopadhyay, P.V.
Kamat, J. Phys. Chem. 100 (1996) 8436.
[11] T. Wu, G. Liu, J. Zhao, H. Hidaka, N. Serpone, J. Phys. Chem. B 103 (1999) 4862.
[12] M. Anpo, M. Takeuchi, K. Ikeue, S. Dohshi, Curr. Opin. Solid State Mater. Sci. 6
(2002) 381.
[13] H. Yamashita, M. Harada, J. Misaka, M. Takeuchi, K. Ikeue, M. Anpo, J. Pho-
tochem. Photobiol. A: Chem. 148 (2002) 257.
[14] S. Klosek, D. Raftery, J. Phys. Chem. B 105 (2001) 2815.
[15] Y. Sakata, T. Yamamoto, T. Okazaki, H. Imamura, S. Tsuchiya, Chem. Lett. 27
(1998) 1253.
[16] T. Ohno, F. Tanigawa, K. Fujihara, S. Izumi, M. Matsumura, J. Photochem. Pho-
tobiol. A: Chem. 127 (1999) 107.
[17] K.C. Cho, K.C. Hwang, T. Sano, J. Photochem. Photobiol. A: Chem. 161 (2004)
155.
[18] S.H. Chien, M.C. Kuo, C.H. Lu, Catal. Today 97 (2004) 121.
[19] A.V. Vorontsov, E.N. Savinov, J. Zhensheng, J. Photochem. Photobiol. A: Chem.
125 (1999) 113.
[20] H. Kisch, W. Macyk, ChemPhysChem 3 (2002) 399.
[21] Y.M. Lin, Y.H. Tseng, J.H. Huang, C.C. Chao, C.C. Chen, I.K. Wang, Environ. Sci.
Technol. 40 (2006) 1616.
[22] O.I. Miller, D.S. Celermajer, J.E. Deanfield, D.J. Macrae, Arch. Dis. Child 70 (1994)
47.
[23] R.S. Wade, C.E. Castro, Chem. Res. Toxicol. 9 (1996) 1382.
[24] C. Lettmann, K. Hildenbrand, H. Kisch, W. Macyk, W.F. Maier, Appl. Catal. B:
Environ. 32 (2001) 215.
[25] Tauc, R. Grigorovici, A. Vancu, Phys. Stat. Sol. 15 (1966) 627.
[26] E. Papirer, R. Lacroix, J.B. Donnet, G. Nanse, P. Fioux, Carbon 33 (1995) 63.
[27] S. Sakthivel, H. Kisch, Angew. Chem. Int. Ed. 42 (2003) 4908.
[28] Y. Li, D.S. Hwang, N.H. Lee, S.J. Kim, Chem. Phys. Lett. 404 (2005) 25.
[29] Y. Ohko, Y. Nakamura, N. Negishi, S. Matsuzawa, K. Takeuchi, J. Photochem.
Photobiol. A: Chem. 205 (2009) 28.
4. Conclusions
In this study, the Pt4/TiO2 photocatalyst, which was synthe-
sized by sol–gel process followed by impregnation with H2Pt(OH)6,
exhibited a notable visible activity for NOx degradation, compared
to bare TiO2. The results of characterization revealed that the parti-
cle size, the crystal structure and the surface area were unchanged
after modification with 1% PtOx and undergoing 200 ◦C calcination.
The chemical state of platinum species played an important role
[30] J. Yang, D. Li, Z. Zhang, Q. Li, H. Wang, J. Photochem. Photobiol. A: Chem. 137
(2000) 197.
[31] A. Sclafani, J.M. Herrmann, J. Photochem. Photobiol. A: Chem. 113 (1998)
181.