Suppressed recombination of electrons and holes and its role on the improvement of photoreactivity of flame-synthesized TiO2 nanopowders

Article abstract of DOI:10.1016/j.cplett.2009.01.065

An in-situ NEXAFS study was performed to directly investigate the change in the hole structures in the conduction band of TiO₂ nanopowders synthesized using a flame method under UV irradiation. The anatase/rutile phase boundary was found to sup

Full text of DOI:10.1016/j.cplett.2009.01.065

Chemical Physics Letters 470 (2009) 269–274
Contents lists available at ScienceDirect
Chemical Physics Letters
journal homepage: www.elsevier.com/locate/cplett
Suppressed recombination of electrons and holes and its role on the improvement
of photoreactivity of flame-synthesized TiO nanopowders
2
a
a
b
c
a,*
HyunSeock Jie , Hoon Park , Keun-Hwa Chae , Masakazu Anpo , Jong-Ku Park
a
Nano-Science Research Division, Korea Institute of Science and Technology, Cheongryang, P.O. Box 131, Seoul 130-650, Republic of Korea
Materials Science and Technology Research Division, Korea Institute of Science and Technology, Cheongryang, P.O. Box 131, Seoul 130-650, Republic of Korea
Department of Applied Chemistry, Graduate School of Engineering, Osaka Prefecture University, 1-1 Gakuen-cho, Sakai Osaka, Japan
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
An in-situ NEXAFS study was performed to directly investigate the change in the hole structures in the
Received 28 October 2008
In final form 20 January 2009
Available online 29 January 2009
conduction band of TiO
anatase/rutile phase boundary was found to suppress the hole–electron recombination, leading to
improved photoreactivity of TiO nanopowders.
2
nanopowders synthesized using a flame method under UV irradiation. The
2
Ó 2009 Elsevier B.V. All rights reserved.
1
. Introduction
lattice defects inside the particles, and the anatase-to-rutile phase
ratios, are determined by the process conditions as well as by the
method of synthesis.
2
Titanium oxide (TiO ) nanopowders with photocatalytic prop-
erties have been fabricated using various methods, including sol–
gel, hydrothermal treatment, and other physical methods [1–3].
Among these methods, the flame synthesis approach is one of
Here we present the results of a direct investigation of the hole
structures in the conduction band of flame-synthesized TiO nano-
2
powders that have been subsequently heat-treated at various tem-
peratures. In addition, we examine the possible link between
changes in the hole structures determined directly under UV irra-
the most popular methods for preparing TiO
high photocatalytic performance [4–6]. In general, the flame syn-
thesis of TiO nanopowders is carried out at high temperature
above 900 °C, affording TiO nanoparticles with very high degrees
2
nanopowders with
2
diation, and the photocatalytic performance of the respective TiO
nanopowders.
2
2
of crystallinity. Therefore, no additional calcination at high temper-
ature is necessary to improve the crystallinity [7–12]. This fabrica-
2
. Experimental
.1. Preparation of photocatalysts
The TiO nanopowders were synthesized by a flame method
tion method results in a low degree of agglomeration of the TiO
2
nanoparticles, which is crucial to obtaining high dispersivity in
liquid media as well as a large specific surface area. Although
TiO nanopowders with an anatase structure have peculiar charac-
2
teristics applicable to the various fields related to photo-induced
electrons, many studies have sought to improve the photoreac-
2
2
using titanium tetra-isopropoxide (TTIP-Aldrich, 97%) as a precur-
sor. The TTIP vapour was vaporized in an oil bath and delivered to
the burner nozzle along with nitrogen gas as a carrier gas. The TTIP
vapour was combined with oxygen gas (an oxidizer) and methane
gas (a fuel) in a burner and the mixture heated to combustion. The
resultant TiO nanoparticles loaded in the flowing product gas
2
were transported to the collection chamber and separated from
the product gas by filtration through a High Airflow Particulate
Air (HAPA) filter. The obtained TiO
ferred to as the as-received TiO nanopowder) was heat treated
at 400, 500, 600, 700, 800 or 900 °C for 1 h in air. Hereafter, the
heat-treated powders are referred to as HTxxx, where xxx indicates
the treatment temperature in degrees Celsius.
tivity of TiO
range to visible light. The characteristics of TiO
2
nanopowders and to expand the light harvesting
nanopowders
2
are known to sensitively depend on the synthesis method used.
Each method has its own advantages and limitations with respect
2
to controlling the properties of the TiO nanopowders. The proper-
ties of a synthesized TiO nanopowder can, however, be altered by
2
post-treatments such as surface treatment in acid solution [13] and
heat treatment at elevated temperature in a controlled atmosphere
2
nanopowder (hereafter re-
2
[
4,14].
Of course, the effects of such post-treatments on the properties
of a TiO nanopowder cannot be completely separated from the
synthesis method used to produce the nanopowder, because the
2
2.2. Characterization of crystalline and electronic structures
The as-received and heat-treated TiO
2
nanopowders were char-
*
Corresponding author. Fax: +82 2 958 5529.
E-mail address: jkpark@kist.re.kr (J.-K. Park).
acterized using X-ray diffraction for determining the constituent
0
009-2614/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.cplett.2009.01.065

2
70
H. Jie et al. / Chemical Physics Letters 470 (2009) 269–274
phase(s) and particle size (XRD; Bruker D8 Advance) [17], and
transmission electron microscopy for determining the shape and
size of the particles (TEM; FEI Technai G .)
surface and the resultant photo-current was recorded. The respec-
tive O K and Ti L edge spectra were recorded for every sample
under normal NEXAFS conditions prior to turning on the UV lamp.
Additional O K and Ti L edge spectra were obtained twice for every
sample under sequential irradiation with UVA followed by UVB.
These NEXAFS spectra are referred to as ‘UVA1 and UVA2’ and
‘UVB1 and UVB2’. Just after turning off the UVB lamp, two addi-
tional scans of the compacted disk (called OFF1 and OFF2) were
performed.
2
2
The hole structures in the conduction band of the TiO nano-
powders were directly investigated under UV irradiation by NEXA-
FS (Near-Edge X-ray Absorption Fine Structure; 7B1 KIST B/L at the
Pohang Accelerating Laboratory (PAL), Korea). The powder samples
were compacted into thin disks without using any polymer binder
in order to prevent surface charging and additional effects associ-
ated with polymer binders. Two kinds of 8 W UV-lamps, radiating
UVA (k = 320–400 nm) and UVB (k = 280–320 nm), were installed
together in the NEXAFS chamber. The compacted disk was
mounted on the specimen holder and positioned vertically beneath
the UV lamp. The UV lamp was located 50 cm from the substrate
surface, forming an angle of 15° from the normal. Prior to data
2.3. Measurements of photocatalytic activities
The photoreactivity of the TiO
from the reaction rates of the oxidative degradation of 2-propanol
into CO and water on the surface of the TiO photocatalysts under
UV light irradiation in the presence of water and oxygen. Briefly,
50 mg of the TiO nanopowder was suspended in a quartz cell
2
nanopowders was determined
2
2
À9
acquisition, the NEXAFS chamber was evacuated to 10 Torr.
The incident beam was irradiated perpendicular to the substrate
2
2
Fig. 1. TEM images of TiO nanopowders: (a) as-synthesized, (b) HT500, (c) HT600, (d) HT700, (e) HT800 and (f) HT900.

H. Jie et al. / Chemical Physics Letters 470 (2009) 269–274
271
À3
containing an aqueous solution of 2-propanol (2.6 Â 10 mol dm
(k > 254 nm) emitted from a 100 W high-pressure Hg lamp.
Aliquots (2 mL) taken from the solution at regular intervals were
centrifuged and filtered through a Millipore filter to separate the
À3
,
25 mL). Prior to UV light irradiation, the suspension was stirred
for 30 min under an oxygen atmosphere in the dark. Then, at 295 K,
the suspension was continuously stirred under an oxygen
atmosphere and simultaneously irradiated with UV light
2
TiO particles from the solution, and then analyzed by gas
chromatography.
3
. Results and discussion
3.1. Catalyst characterization
2
The as-received and heat-treated TiO nanopowders were
observed with TEM and their respective images are shown in
Fig. 1. The particles of the as-received TiO nanopowder are
spherical in shape. These spherical TiO particles were found to
2
2
change into a euhedral shape at 500 °C (Fig. 1c), and then to
octagonal and cuboidal shapes after heat treatment at tempera-
tures of 600 °C and above (Fig. 1d–f). The inset of each image
shows the detailed morphologies of the TiO
magnification. Inter-particle necking is evident above 700 °C.
From the XRD patterns of TiO nanopowders, the average
2
particles under high
2
crystallite size of anatase and rutile and the ratio between two
phases shown in Fig. 2 [15]. The as-received powder was almost
anatase phase containing about 2% of rutile phase. An anatase-
to-rutile phase transition occurred above 700 °C and was com-
2
Fig. 2. Variations of TiO particle size and content of rutile phase in the powders
with heat treatment.
pleted at 900 °C. The content of rutile phase TiO
at 700 °C, 23% at 800 °C, and finally 99% at 900 °C in the powder.
From the X-ray diffraction data, the anatase-phase TiO particles
2
increased to 5%
2
were estimated to grow slightly from 50 nm to 55 nm in diameter
with heat treatment up to 700 °C. On the other hand, the diameter
of rutile-phase TiO
to about 55 nm, very close to that of anatase-phase particles.
Therefore, the morphology change of anatase-phase TiO particles
2
particles in the HT800 powder was estimated
2
below 700 °C is believed not to have a direct relation with the
phase transition, but to be related to surface relaxation through
surface diffusion with help of enhanced atom mobility, because
there is neither anatase-to-rutile phase transition nor evident
particle coarsening.
3.2. Evaluation of photocatalytic activity
Fig. 3 shows the results of the photocatalytic degradation of
2
2
-propanol with the as-received and heat-treated TiO nanopow-
ders. The degradation rate of 2-propanol increased with increasing
heat-treatment temperature up to 800 °C. Interestingly, the
degradation rate increased continuously up to 700 °C, despite no
Fig. 3. Photocatalytic degradation of 2-propanol under UV irradiation by various
kinds of TiO nanopowders (as-synthesized and heat-treated).
2
2
evident differences in the phases of the TiO nanopowder particles
2
Fig. 4. TEM images showing the anatase–rutile phase boundary in the TiO nanopowder treated at 800 °C.

2
72
H. Jie et al. / Chemical Physics Letters 470 (2009) 269–274
treated with temperatures up to 700 °C, as shown in Fig. 2.
Improvement in the photocatalytic property of TiO nanopowders
heat treatment at 500 °C, and then to faceted octagonal and cuboi-
dal shapes after heat treatment at temperatures of 600 °C and
above. Thus, the improvement in photocatalytic activity can be
attributed to partly faceting of the spherical particles.
In order to identify the existence of anatase–rutile phase
boundary in the HT800 powder, several dumbbell-type particles
were observed by TEM. Considering both rutile phase content of
2
with heat treatment up to 700 °C seems to be correlated to the
change of particle shape. Further improvement in the photocata-
lytic property of HT800 is attributed to the contribution of the
anatase/rutile phase boundary [16,17], which was accompanied
by the anatase-to-rutile transition, as shown in Fig. 2. Over
8
00 °C, the degradation rate of 2-propanol decreased drastically,
2
23% and almost same size of anatase and rutile phase TiO parti-
due to the presence of a large fraction of rutile phase with poor
photocatalytic activity [18].
cles, one rutile phase particle may be statistically estimated to con-
tact with four anatase particles in the HT800 powder. Fig. 4 shows
the observation results taken from the HT800 powder dispersed in
the ethanol with ultrasonic treatment. Many clear boundaries were
found between two particles bonded together. Some of them are
anatase–anatase boundaries (grain boundary) and others are ana-
tase–rutile phase boundaries. Low magnification image shows a
clear boundary between two particles, which were identified,
Various factors potentially contribute to the considerable
improvement in the photocatalytic activity afforded by heat treat-
ment with temperatures of up to 700 °C. Changes in the internal
phase and particle size may affect the photocatalytic activity, but
Figs. 1 and 2 show that both the internal phase change and particle
coarsening are negligible during heat treatment up to 700 °C. How-
ever, the TEM images show that the spherical morphology of the
2 2
respectively, as an anatase TiO particle and a rutile TiO particle.
as-prepared TiO
2
particles changes to a euhedral shape following
It is noteworthy from Fig. 4 that as a result of anatase-to-rutile
Fig. 5. Variation of oxygen K-edge spectra of TiO
2
nanopowders under UV irradiation: (a) as-synthesized, (b) HT600, (c) HT700 and (d) HT800.

H. Jie et al. / Chemical Physics Letters 470 (2009) 269–274
273
phase transition, an anatase–rutile phase boundary, strong enough
not to be broken easily during dispersion, is formed, which can
interesting finding in the present analysis was the preferential
filling of holes in the e level in the HT700 and HT800 powders.
Even though the preferential filling of e level seems to be
related to the broadening of e peak of which tail in the low energy
g
play an important role on the photoreactivity of TiO
in the present study.
2
nanopowders
g
g
side is extended toward t2g peak, the variation of the hole
structures with heat treatment at elevated temperature should
be verified further.
3.3. Determination of hole structure under UV irradiation
The photocatalytic behavior of the TiO
related to their electronic structures, especially the hole structures
of the TiO nanopowders, because the photo-excited electrons
2
nanopowders is closely
2
The change in the hole structures of the TiO nanopowders
seems to correspond well with the change in their photocatalytic
properties shown in Fig. 3, which was determined under UV light
irradiation with wavelengths longer than 254 nm. As the lifetime
of the excited electrons in the conduction band increased, the
2
from the valance band fill the holes in the conduction band leaving
holes in the valance band, and both the photo-excited electrons in
the conduction band and the resultant holes in the valance band
play a key role in the photocatalytic reaction. NEXAFS is known
to be an adequate method for determining the hole structures in
conduction bands. In general, the holes in the conduction band
are filled with electrons excited from the valance band, according
to the excitation energy determined by the wavelength of incident
light, which leads to a decrease in the intensity of the holes with
correspondent energy. Fig. 5 shows the in-situ NEXAFS observation
of the O K-edge spectra of the photocatalysts under UV irradiation.
Fig. 5a shows a hole structure in the conduction band of the as-re-
2
photoreactivity of the TiO nanopowders improved. The relation-
ship between Figs. 3 and 5 can be explained in terms of microstruc-
tural aspects, specifically, in terms of the existence of an anatase/
rutile phase boundary, which is known to act as a cationic trap
for electrons excited from the valance band [19,20]. The as-re-
2
ceived TiO nanopowder contained about 10% rutile phase, which
remained almost intact below about 700 °C. The anatase-to-rutile
phase transition proceeded rapidly above 800 °C, and was almost
completed at 900 °C. With the help of surface diffusion causing
nonfaceted–faceted transitions in the particle shape in Fig. 1, the
necks at the particle contacts grow with heat treatment; direct evi-
dence of this is found in the observation of neck growth between
2
ceived TiO nanopowder, which was well matched with that of
TiO with the anatase form. Interestingly, when the samples were
2
irradiated with UV light, serrated curves were obtained irrespec-
tive of the sample condition. However, when the UV light was
turned off, the serrated curves returned to their original smooth
profile.
2
the TiO particles in Fig. 1. Similarly, contact areas between anatase
and rutile particles form anatase/rutile phase boundaries that in-
crease in size as the heat-treatment temperature is increased, even
at temperatures below 700 °C. The anatase phase transforms to the
rutile phase at a fast rate above 800 °C, and the anatase/rutile
phase boundary disappears at 900 °C. This change in the anatase-
to-rutile phase boundary corresponds well with the gradual
improvement in the photocatalytic properties in accordance with
the heat-treatment temperature. As the area of the anatase/rutile
phase boundary increased, more electrons became trapped in the
phase boundary because upward distortion of the conduction band
in the space charge layer of anatase crystal in contact with rutile
crystal blocks the electron transfer from anatase to rutile crystal,
resulting in a long lifetime in the conduction band [20,21].
We determined the peak position, as well as the variation of rel-
ative peak area between the t2g and e
treatment temperature. In order to determine both relative peak
area and peak position of t2g and e in each condition, the NEXAFS
curves in Fig. 5 were deconvoluted, respectively, to separate the t2g
and e peaks. Comparing with the t2g and e peaks obtained from
g
levels, as a function of heat-
g
g
g
the as-received and HT600 powders, those obtained from HT700
and HT800 powders tend to move simultaneously to the low
energy side, 0.8 eV for t2g and 0.4 eV for e
of relatively greater shift of the t2g peak against e
that the e peak broadened and its tail of low energy side expanded
towards t2g peak area below the peak intensity energy position of
2g by 0.4 eV approximately. For comparison, the low energy
minima of e peak of as-received and HT600 powders were found
almost same with the peak intensity energy position of t2g within
the error limit. Each area of t2g and e peaks was calculated and the
relative peak area of [At2g =Aeg ] between t2g and e was determined.
g
. In addition, in spite
g
, it was found
g
The present findings give direct experimental evidence show-
ing, for the first time, that the electrons trapped in the anatase/ru-
tile grain boundary suppress the recombination of electrons and
holes, and in turn, this suppression of recombination contributes
directly to the improvement in the photocatalytic property of
t
g
g
2
flame-synthesized TiO nanopowders.
g
The as-received and HT600 nanopowders showed almost no
change in the [At2g =Aeg ] by about 0.77 during UVA1 irradiation
while the HT700 and HT800 nanopowders showed large values
4. Conclusions
of [At2g =A g
area of e
e
], 1.52 and 1.30, respectively, which means that the peak
decreased significantly compared with that of t2g. In
2
The anatase-phase-rich TiO nanopowders containing very
small volume fraction of rutile phase were fabricated by the flame
method and treated at elevated temperatures to improve their
photoreactivity. Their photoreactivity increased with increasing
heat-treatment temperature within the limit of the anatase-
phase-rich condition, but decreased drastically with formation of
the rutile phase as a result of the anatase-to-rutile phase transition.
g
g
other words, the holes in the e level were filled predominantly
with electrons excited from the valance band under UVA1 irradia-
tion. During subsequent irradiation from UVA2 to UVB2, the
[At2g =A g ] values were kept around 0.70 slightly lower than that
e
of unexposed nanopowder (0.77) and maintained even after turn-
ing off UVB(OFF1), meaning that the holes in the t2g level were
2
The hole structures in the conduction band of TiO nanopowders
filled relatively more that those in the e
t2g =A g ] value was recovered to that of unexposed powder at
OFF2.
Present analyses of t2g and e
band of TiO nanopowders synthesized in the present study moved
g
level. Then, the
were directly determined by in-situ NEXAFS measurements under
UV light irradiation. The NEXAFS results showed that the photoex-
citation of electrons from the valance band became more evident
as the heat-treatment temperature increased up to 800 °C. The
[
A
e
g
peaks show that the conduction
2
improvement in photoreactivity of the TiO nanopowders is attrib-
2
to the low energy side with heat treatment at the elevated
temperature above 700, which leads to easy excitation of valance
electrons and subsequent decrease in hole density at the t2g level.
The electrons filling the t2g level resided in the conduction band as
long as about 30 min after turning off UV irradiation. Another
uted to both the change of particle shape at the low temperature
region and the existence of an anatase/rutile phase boundary,
which traps excited electrons and suppresses the recombination
of electrons and holes, causing the photo-induced electrons to have
long lifetimes.

2
74
H. Jie et al. / Chemical Physics Letters 470 (2009) 269–274
[
[
10] C.K. Chan, J.F. Porter, Y.-G. Li, W. Guo, C.-H. Chan, J. Am. Ceram. Soc. 82 (1999)
66.
11] C.H. Cho, D.K. Kim, D.H. Kim, J. Am. Ceram. Soc. 86 (2003) 1138.
Acknowledgments
5
The NEXAFS measurements at PLS were supported in part by
MOST and POSTECH.
[12] J.F. Porter, Y. Li, C.K. Chan, J. Mater. Sci. 34 (1999) 1523.
[13] T. Taguchi, Y. Saito, K. Sarukawa, T. Ohno, M. Matsumura, New J. Chem. 27
(
2003) 1304.
[
[
[
14] S. Yang, L. Gao, J. Am. Ceram. Soc. 88 (2005) 968.
15] R.A. Spurr, H. Myers, Anal. Chem. Res. 29 (1957) 760.
16] T. Miyagi, M. Kamei, T. Mitsuhashi, T. Ishigaki, A. Yamazaki, Chem. Phys. Lett.
390 (2004) 399.
References
[
[
[
1] C.H. Cho, M.H. Han, D.H. Kim, D.K. Kim, Mater. Chem. Phys. 92 (2005) 104.
2] T. Ohno, K. Sarukawa, M. Matsumura, J. Phys. Chem. B 105 (2001) 2417.
3] K. Nakaso, K. Okuyama, M. Shimada, S. Pratsinis, Chem. Eng. Sci. 58 (2003)
[17] D.C. Hurum, A.G. Agrios, K.A. Gray, T. Rajh, C. Thurnauer, J. Phys. Chem. B 107
(2003) 4545.
3
327.
[18] Yu.V. Kolen, B.R. Churagulv, M. Kunst, L. Mazerolles, C. Colbeau-Justin, Appl.
[
[
[
4] S.U. Khan, M.U.M. Al-Shahry, W.B. Ingler Jr., Science 297 (2002) 2243.
5] B. Neppolian, H.S. Jie, J.P. Ahn, J.K. Park, M. Anpo, Chem. Lett. 33 (2004) 1562.
6] H. Park, B. Neppolian, H.S. Jie, J.P. Ahn, J.K. Park, M. Anpo, D.Y. Lee, Curr. Appl.
Phys. 7 (2007) 118.
Catal. B: Environ. 52 (2004) 51.
[19] G.S. Herman, Z. Dohnalek, N. Ruzycki, U. Diebold, J. Phys. Chem. B 107 (2003)
2788.
[20] Bo Sun, Panagiotis G. Smirniotis, Catal. Today 88 (2003) 49.
[21] Bo Sun, Alexandre V. Vorontsov, Panagiotis G. Smirniotis, Lanmuir 19 (2003)
3151.
[
[
[
7] C.B. Almquist, P. Biswas, J. Catal. 212 (2002) 145.
8] J.-H. Lee, Y.-S. Yang, Mater. Chem. Phys. 93 (2005) 237.
9] H. Wang, Y. Wu, B.-Q. Xu, Appl. Catal. B: Environ. 59 (2005) 139.