Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

View Online / Journal Homepage / Table of Contents for this issue

PAPER www.rsc.org/materials | Journal of Materials Chemistry

Iodine doped anatase TiO2 photocatalyst with ultra-long visible light


response: correlation between geometric/electronic structures and
mechanisms†
Gang Liu,‡ab Chenghua Sun,‡bc Xiaoxia Yan,c Lina Cheng,c Zhigang Chen,a Xuewen Wang,a
Lianzhou Wang,b Sean C. Smith,*c Gao Qing (Max) Lu*b and Hui-Ming Cheng*a
Received 20th November 2008, Accepted 4th February 2009
First published as an Advance Article on the web 11th March 2009
DOI: 10.1039/b820816f
Downloaded by UNIVERSITY OF SOUTH AUSTRALIA on 09 November 2012

We report that surface iodine doped TiO2 with coexisting atomic configurations of iodine dopant, I–O–
Published on 11 March 2009 on http://pubs.rsc.org | doi:10.1039/B820816F

I and I–O–Ti structures, exhibits an extended absorption edge up to 800 nm. Furthermore,
photocatalytic activity investigations confirm the efficient generation of important oxidative species
cOH radicals in photocatalysis oxidation processes beyond 600 nm. It is theoretically found that iodine
atoms prefer to be doped within the surface due to the strong I–O repulsion. I-doping in the surface led
to little change in its intrinsic bandgap, but the distribution and occupation of localized states of iodine
strongly depend on the surface iodine configurations. The latter is responsible for the wider range
visible light response exhibited in anatase TiO2 with the coexistence of I–O–I and I–O–Ti structures.
The coexistence of I–O–I and I–O–Ti structures can distinctly change the surface structure due to the
release of local strain energies.

1. Introduction can also be impacted by both the chemical composition and


specific atomic configurations of the dopant atoms. This point is
Photocatalysis involves the excitation, bulk diffusion and surface effectively demonstrated by a recent work showing that clear
transfer of photo-induced carriers which depend on the elec- synergetic photocatalytic effects of B/N codoping are realized in
tronic structure and surface structure of the semiconductor the presence of B–N bonds in TiO2,19 whereas only a negligible or
photocatalyst.1,2 Doping strategies,3–14 in particular nonmetal limited effect is found without the formation of B–N bonds.14,20
doping,5–14 have been widely employed to modify the electronic It is therefore expected that efficient photocatalysts with favor-
structure of wide-bandgap photocatalysts, particularly TiO2 in able electronic structure and surface structure can be accom-
order to improve the photocatalysis efficiency and accomplish plished by well controlling the atomic configurations of dopants.
visible light photocatalytic activity. In principle, the energy level Extensive efforts have been devoted to nonmetal doping in the
position and abundance of states in the bandgap associated with bulk on the microscopic and electronic structure of TiO2.21–28
the doped atoms are sensitive not only to the chemical compo- Many fruitful theoretical and experimental investigations on
sition of, but also to the atomic configurations (different bond single crystal TiO2 surfaces have been reported in the past
structures), of the dopant atoms. The doping process can also decade, providing background knowledge of the undoped
alter the surface structure of photocatalysts. For example, it is surface as a reference.29,30 A few theoretical works are available
demonstrated both experimentally and theoretically15,16 that N involving N doping on anatase (101) and on rutile (110).31,32 In
doping in TiO2 can introduce oxygen vacancies and trigger contrast to many other nonmetal dopants (N, C, B, S), which
surface reconstruction, which may cause not only limited visible substitute lattice oxygen, the less-studied iodine13a,33–35 is usually
light activity but also impaired UV activity.17,18 Surface structure thought to be able to replace lattice Ti. Very recently, Su et al.36
reported I7+/I 1 co-doped titania, in which the iodine dopant was
a
Shenyang National Laboratory for Materials Science, Institute of Metal considered as a surface-adsorbed species, such as the IO4 group,
Research, Chinese Academy of Sciences, 72 Wenhua Road, Shenyang, and they performed calculations on the effects of surface
110016, China. E-mail: cheng@imr.ac.cn; Fax: +86 24 23903126; Tel: adsorbed I and IO4 groups. Theoretical studies of I-doped bulk
+86 24 23971611
b
ARC Centre of Excellence for Functional Nanomaterials, School of
TiO2 have also been carried out (see, e.g., Long et al.37 and Yang
Engineering and Australian Institute of Bioengineering and et al.38).In addition, we note that the maximum experimentally
Nanotechnology, The University of Queensland, Qld, 4072, Australia. reported absorption edge of I-doped TiO2 is ca. 550 nm. In all
E-mail: maxlu@uq.edu.au; Fax: +61 7 33656074; Tel: +61 7 33653735 these studies, the important effects of different atomic configu-
c
Centre for Computational Molecular Science, Australia Institute for
Bioengineering and Nanotechnology, The University of Queensland, Qld,
rations of the iodine dopant on the electronic structure and
4072, Australia. E-mail: s.smith@uq.edu.au; Fax: +61 7 33463992; Tel: surface structure of TiO2 are not considered.
+61 7 33463949 In this work, we report surface iodine doped anatase TiO2 with
† Electronic supplementary information (ESI) available: XRD patterns, two atomic configurations, I–O–I and I–O–Ti, which has an
XPS spectra, element mapping, photodecomposition activity and
electronic structure for bulk anatase TiO2 and clean (101) surface. See
extended absorption edge up to around 800 nm and can be active
DOI: 10.1039/b820816f in generating photocatalytic activity beyond 600 nm. By
‡ These authors contributed equally to this work. combining the first-principles calculations with experiments, the

2822 | J. Mater. Chem., 2009, 19, 2822–2829 This journal is ª The Royal Society of Chemistry 2009
View Online

origin of the extended absorption edge together with changed Theoretically the doped iodine atoms may replace neighboring
surface structure in doped anatase TiO2 is well investigated in or separated titanium atoms, which represent two doping
terms of both bulk/surface distribution and configurations of configurations of I–O–I and I–O–Ti. Both bulk and surface
iodine dopant (I–O–I and I–O–Ti). Significant insights into the doping sites are possible and should be considered, indicated as
surface doping properties of TiO2, which intrinsically determine bulk doping and surface doping. For the former, a 2  2  1
its photon harvest, redox power and surface transfer of photo- supercell was employed, consisting of 16 units of TiO2, as shown
induced carriers, are thereby achieved. in Fig. S5.† For the latter case, surface (101) was employed for
two reasons: (i) the equilibrium morphology of the single crystal
2. Experimental and computational methods of TiO2 is mainly determined by {101} (the majority); (ii) among
the low-indexed surfaces, (101) represents the most stable surface
2.1 Experimental details of anatase TiO2. For (101), a stoichiometric slab model 4  4 was
Titanium tetrafluoride (TiF4, Aldrich) and HIO4 acid were employed with 64 units of TiO2, as shown in Fig. S6.† The
surface is composed of two- and three-coordinated oxygen and
Downloaded by UNIVERSITY OF SOUTH AUSTRALIA on 09 November 2012

employed as TiO2 precursor and iodine dopant precursor to


synthesize Tia(IO4)bFc intermediate compounds. In a typical five- and six-coordinated titanium for clean (101). To set the
Published on 11 March 2009 on http://pubs.rsc.org | doi:10.1039/B820816F

synthesis procedure, TiF4 powder precursor was dissolved into doped models for both bulk doping and surface doping, titanium
an aqueous solution of 1 M HIO4, followed by hydrothermal atoms are replaced by iodine at different sites, which makes it
treatment in a Teflon-lined autoclave at 180  C for 12 h. The possible to study the difference between I–O–I and I–O–Ti bond
collected samples were calcined in air at 400  C for 2 h to obtain structures.
iodine doped TiO2. The measurements of the amount of cOH
were conducted according to the literature.39 40 mg of photo-
catalyst was suspended in 80 ml of aqueous solution containing 3. Results and discussion
0.01 M NaOH and 3 mM terephthalic acid. Before exposure to
visible light irradiation, the suspension was stirred in the dark for 3.1 Structure and compositions of iodine doped TiO2
30 min. Then, 5 ml of solution was taken out after every 30 min, Fig. 1A shows the SEM image of the intermediate compound
and centrifuged for fluorescence spectrum measurements. (Tia(IO4)bFc (its XRD pattern and compositions are shown in
During the photoreactions, no oxygen was bubbled into Fig. S1-a and S2†) prepared by a hydrothermal process at 180
suspension. The employed excitation light in recording fluores- 
C. The thermal decomposition of the intermediate compound in
cence spectra is 320 nm. The light source was a 300 W Xe lamp air at 400  C led to the formation of iodine doped TiO2. The
(Beijing Trusttech Co. Ltd PLS-SXE-300UV) with cut-off by obtained TiO2 nanoparticles were anatase (Fig. S1-b) and ca. 40
three groups of glass filters with the wavelength of 420–770 nm, nm in size (Figs. 1B,C). The lattice fringes with the distance of
540–770 nm and 600–770 nm, respectively. 3.54 Å in Fig. 1D can be assigned to the most stable and most
X-Ray diffraction (XRD) patterns were recorded on a Rigaku frequently observed anatase TiO2 (101).
diffractometer using Cu irradiation. Scanning electron micros-
copy (SEM) and transmission electron microscopy (TEM)
images were recorded on a JEOL 6300 and a Tecnai F30. A UV-
vis spectrophotometer (JASCOV550) was used to obtain the
optical absorbance spectra of the samples. Raman spectra were
recorded on a LabRam HR 800. Chemical compositions of
derived TiO2 were analyzed using X-ray photoelectron spec-
troscopy (Thermo Escalab 250, a monochromatic Al Ka X-ray
source). All binding energies were referenced to the C 1s peak
(284.6 eV) arising from adventitious carbon.

2.2 Computational methods

To investigate the doping effect at the electronic level, the


geometric and electronic structures of iodine doped TiO2 with
different atomic configurations of iodine were studied in the
framework of density functional theory (DFT) within the
generalized-gradient approximation (GGA).40 All calculations
were carried out using the exchange–correlation functional of
Perdew–Burke–Ernzerhof (PBE),41,42 based on the numerical
double-numerical polarization (DNP) basis set, which has been
implemented in the Dmol3 modules.43,44 More tests and the
discussion regarding the efficiency and reliability of the DNP
basis set can be found in Delley’s work, in which the estimated Fig. 1 Morphology and HRTEM images of samples: A) SEM image of
errors from the PBE functional with the DNP basis set were intermediate compound; B) SEM, C) TEM and D) high resolution TEM
supposed to be lower than those with the hybrid B3LYP/6- images of the iodine doped titania by calcining the intermediate
31G** functional.44 compound of A) at 400  C for 2 hrs in air.

This journal is ª The Royal Society of Chemistry 2009 J. Mater. Chem., 2009, 19, 2822–2829 | 2823
View Online

To investigate the compositions of the obtained TiO2 and 3.2 Raman spectra and optical absorption spectra
chemical states of related elements, we conducted measurements
Raman spectra of iodine doped TiO2 and pure anatase TiO2 are
of F1s, I3d5/2, Ti2p and O1s core levels using the XPS technique.
shown in Fig. 3a. Compared to reference anatase titania
No F species were detected in the TiO2 obtained. There are two
(Aldrich), there is no notable shift for active modes B1g and A1g,
states of iodine with binding energies of 624.7 eV and 627.1 eV,
but an obvious shift to higher wavenumber of 146 cm 1 for the Eg
shown in Fig. 2. So far, two kinds of iodine dopant have been
mode at 144 cm 1 in iodine doped titania, which is consistent
reported. One is 627.5 eV reported by Hong et al.13a and the other
with the previously reported iodine doped titania with I–O–I.33
is 624.5 eV in the previous work,33 both of which were considered
Importantly, two new active modes at 182 cm 1 and 192 cm 1
as substitutional iodine dopant in the titanium lattice. In prin-
appear in iodine doped TiO2, together with the much weakened
ciple, there are two possibilities of iodine doping in the TiO2
Eg mode at 196 cm 1. Considering that no such effects appear in
crystal structure, I–O–I and I–O–Ti. The binding energy for I3d5/2
iodine doped TiO2 with a sole state of I–O–I,33 two new Raman
in HIO4 is ca. 624.8 eV. Therefore, the local structure of iodine
active peaks are therefore related to I–O–Ti, whose force
dopant with the binding energy around 624.8 eV is very close to
Downloaded by UNIVERSITY OF SOUTH AUSTRALIA on 09 November 2012

constant is different from those of Ti–O–Ti and I–O–I. In


the I–O–I configuration as in HIO4. In the sample preparation
Published on 11 March 2009 on http://pubs.rsc.org | doi:10.1039/B820816F

nitrogen doped titania with the N–Ti–O configuration, the new


process, the decomposition of the intermediate compound
Raman active peak with its frequency at ca. 550 cm 1 was
(Tia(IO4)bFc into iodine doped titania breaks most I–O–I bonds
observed,45 which is attributed to the first-order scattering of
in Tia(IO4)bFc and forms some new I–O–Ti bonds. The binding
a non-stoichiometric titanium oxynitride. In our case, I–O–I and
energy of I–O–Ti bonds is determined by the equilibrium
I–O–Ti may be responsible for the shift of the Eg mode at 144
between the much larger electronegativity of I (2.66) than Ti
cm 1 and the generation of the two new active modes at 182 cm 1
(1.54) and the local structure distortion. The iodine with the
and 192 cm 1, respectively, indicating a greatly changed surface
binding energy around 627.1 eV can be ascribed to the I–O–Ti
structure caused by the coexistence of I–O–I and I–O–Ti.
configuration. The formation of I–O–I and I–O–Ti in iodine
The optical response spectrum of iodine doped TiO2 is shown
doped TiO2 was further confirmed by the appearance of two
in Fig. 3b. Compared to the extended absorption edge to ca. 550
additional higher states of oxygen at 531.7 and 533.2 eV and one
nm in the reported iodine doped TiO2 with only I–O–Ti or I–O–
higher oxidation state of titanium at 460.6 eV in the spectra of
I,13a,33 the absorption edge in the iodine doped TiO2 with coex-
O1s and Ti2p. These additional two oxygen states and one tita-
isting I–O–I and I–O–Ti is up to 800 nm. Derived from the plot
nium state are considered to originate from I–O–I and I–O–Ti in
(the inset in Fig. 3b) of the Kubelka–Munk function versus the
iodine doped TiO2 compared to pure TiO2. Considering the
energy of the light, the intrinsic band gap of iodine doped TiO2 is
higher content of I–O–I than I–O–Ti, the oxygen states at 531.7
and 533.2 eV can be attributed to I–O–I and I–O–Ti, respec-
tively. The energy filtered TEM (EFTEM) images (Fig. S3†)
show the elemental distributions of Ti, O and I in iodine doped
TiO2. The obvious signal of elemental iodine in Fig. S3-D†
demonstrates the effective doping of iodine in all particles. The
composition of iodine doped TiO2 was determined to be
Ti0.65I0.26O2 by the XPS technique.

Fig. 3 Raman spectra (a) of anatase TiO2 (A) and iodine doped TiO2 (B)
and UV-visible absorption spectrum (b) of iodine doped TiO2. The inset
Fig. 2 High resolution XPS spectra of I3d5/2, Ti2p and O1s of iodine in (b) is a plot of the transformed Kubelka–Munk function versus the
doped TiO2. energy of the light.

2824 | J. Mater. Chem., 2009, 19, 2822–2829 This journal is ª The Royal Society of Chemistry 2009
View Online

3.14 eV, which is nearly identical to that of pure anatase TiO2 our iodine doped TiO2, whereas no signal was detected with pure
(3.2 eV), suggesting that iodine doping in the case here can not anatase TiO2 under the same conditions. It should be pointed out
decrease the intrinsic band gap. that, besides the dominant route of holes attacking water or
hydroxyl groups, another secondary route to produce hydroxyl
radicals is also possible by the following reactions:
3.3 Photocatalytic activity
O2 + e / O2 (1)
It is vital to demonstrate whether the changed surface structure
and electronic structure of TiO2 caused by the coexistence of I–
2O2 + 2H+ / 2cOH + O2 (2)
O–I and I–O–Ti is favorable for photocatalysis. The holes with
enough oxidation power in the valence band and localized states The key in this route is the involvement of protons in the second
in the band gap after photoexcitation can be directly involved in step. However, the probability of this step may be low due to the
the photocatalytic reaction or generate some active species such employed basic solution. According to the reported procedures,46
Downloaded by UNIVERSITY OF SOUTH AUSTRALIA on 09 November 2012

as important cOH by reacting with surface adsorbed water and further photocatalytic degradation of a model organic pollutant
hydroxyl groups.1,2 It is known that cOH reacts with terephthalic
Published on 11 March 2009 on http://pubs.rsc.org | doi:10.1039/B820816F

azo dye Rhodamine B demonstrated that the iodine doped TiO2


acid (TA) in basic solution and generates 2-hydroxyterephthalic obtained shows much better photocatalytic activity than the
acid (TAOH), which emits unique fluorescence at around 426 benchmark P25 TiO2 under both UV and visible light irradiation
nm.39 As shown in Fig. 4, obvious fluorescence spectra associated (Fig. 5). Furthermore, the prepared iodine doped anatase TiO2
with TAOH are generated upon the visible light irradiation of the photocatalyst has good photocatalytic stability as evidenced by
iodine doped TiO2 suspended in TA solution within 420–770 nm, recycling tests.
540–770 nm and 600–770 nm for different irradiation times,
indicating that the holes generated on the modified electronic
3.4 First-principles calculations of bulk doped anatase TiO2
states with strong enough oxidation power by the photon exci-
with I–O–I and I–O–Ti
tation at wavelengths >600 nm can be easily transferred to
surface adsorbed water and hydroxyl groups on the newly con- To explore the origin of ultra-long visible light response and
structed surface structure. The linear relationship between fluo- changed surface structure caused by the coexisting I–O–I and I–
rescence intensity and irradiation time confirms the stability of O–Ti structures in iodine doped anatase TiO2, we investigated

Fig. 5 Photocatalytic degradation of pollutant Rhodamine B under


Fig. 4 Fluorescence spectra (A) measured for the visible-light (420 nm– various reaction conditions: (A) under the irradiation of UV-visible light;
770 nm) irradiated iodine doped TiO2 suspension in 3 mM terephthalic (B) under the irradiation of visible light (420 nm–770 nm): (a) blank
acid at different irradiation times. (B) Time dependence of the fluores- without any photocatalyst; (b) commercial benchmark P25 titania; (c) the
cence intensity at 426 nm under irradiation at (a) 420–770 nm; (b) 540– iodine doped titania. 50 mg of photocatalyst was dispersed in 100 ml of
770 nm and (c) 600–770 nm. 2  10 5 M Rhodamine B solution in the photodegradation reactions.

This journal is ª The Royal Society of Chemistry 2009 J. Mater. Chem., 2009, 19, 2822–2829 | 2825
View Online
Downloaded by UNIVERSITY OF SOUTH AUSTRALIA on 09 November 2012
Published on 11 March 2009 on http://pubs.rsc.org | doi:10.1039/B820816F

Fig. 6 Structural models and calculated energy band structures for bulk
doping with different doping contents: (A) Ti15O32I, (B) Ti14O32I2 and
(C) Ti13O32I3. Oxygen, titanium and iodine atoms are indicated by red,
grey and blue spheres, respectively. Fig. 7 Structural models and calculated energy band structures for bulk
doping with two configurations, (A) I–O–I; (B) I–O–Ti. Oxygen, titanium
and iodine atoms are indicated by red, grey and blue spheres, respectively.
the first-principles calculations of I–O–I and I–O–Ti structures The numbers in parentheses are the calculated binding energies.
doped in both bulk and surface (101) of TiO2. Although the
systematic undervaluation of the absolute bandgap by DFT is
constructed by O–O and Ti–O. To release the local strain ener-
a well known drawback in first-principles calculations, this will
gies, a strong distortion results, as shown in Fig. 7(B). Although
not impair the validity of bandgap comparisons between undo-
no significant distortion is found in Fig. 7(A), the bond length of
ped and doped TiO2. By way of establishing a comparative
I–O extends to 2.04 Å from the Ti–O bond length of 1.93 Å,
reference with the more commonly studied case of bulk doping,
indicating strong local strain energy. It should be pointed out
our reported calculations began by examining the effects of
that such strain energy associated with the I–O repulsion is not
different bulk doping fractions as well as atomic configurations.
fully released even with the significant structural asymmetry
Utilizing a 2  2  1 supercell, with the same doping configu-
shown in Fig. 7(B) according to the calculated binding energies.
ration of I–O–Ti as studied by Long et al.,37 it is clear that the
Also shown in Fig. 7(A) and (B) are the calculated band
calculated bandgap is sensitive to the extent of doping [Fig. 6
structures for I-doped TiO2, from which two common features
(A)–(C)]. In this supercell, when more than two titanium atoms
for the doping configurations of I–O–I and I–O–Ti can be
are replaced by iodine atoms, an obvious decrease in the bandgap
summarized: (i) the band gap shows dramatic narrowing; and (ii)
can be achieved compared to undoped supercell (Fig. S4†), which
new states between the valence band (VB) and the conduction
indicates that increasing the extent of bulk doping may result in
band (CB) are observed, as indicated by the red lines. Qualita-
the improvement of the visible light response and photocatalytic
tively, the above calculated results agree well with early calcu-
activity of TiO2 due to the decreased bandgap and the internal
lations reported by Long et al.37 These features suggest that
dipole moment. However, from the heavily distorted structures
I-doped TiO2 may present higher photocatalytic activity than
shown in Fig. 6, the strong I–O repulsion may prevent the bulk-
undoped samples. Note, however, that the positions of the new
doping process from being energetically feasible. As a result, the
interband states, the VB and the CB strongly depend on the
doping process may reasonably be conjectured to occur mainly
distribution of iodine atoms. For I–O–I doping, new states are
on the surface.
located below the Fermi level and almost in the middle of VB and
Consideration of the effects of different atomic configurations
CB; however, for I–O–Ti doping, one state is around the Fermi
of the dopant in the bulk is also informative and revealing. The
level and the other one is very close to the CB bottom.
doped iodine atoms may replace either neighboring or separated
titanium atoms, which may be characterized in terms of two
distinct doping configurations: I–O–I and I–O–Ti. These two
3.5 First-principles calculations of surface doped anatase TiO2
atomic doping configurations are illustrated in Fig. 7(A) and (B).
with I–O–I and I–O–Ti
The values in parentheses are binding energies in eV. From the
relaxed structures, the main difference between two different We now consider the effects of surface iodine doping, which for
doping configurations is that only I–O–Ti doping shows signifi- reasons noted above may in fact be a closer representation of the
cant distortion. Normally, the equilibrium structure of bulk experimental situation. Detailed baseline information on the
anatase is determined by the balance between O–O repulsions clean (101) surface structure model and calculated band structure
and attractive Ti–O p interactions.47 Lazzeri et al.48 confirmed (Fig. S5†) is provided in the ESI. For surface doping on the (101)
this view and further pointed out that the O–O repulsions appear surface, the I-doping configurations are complicated since both
to be the most important factor. Since the electronegativity of five- and six-coordinated titanium atoms on the (101) surface
iodine is much stronger than that of titanium, the introduced may be replaced. Basically, there are six possible combinations,
iodine brings I–O repulsions, which destroys the local balance denoted as I6c&I6c, I5c&I6c, I5c&I5c, I6c–O–I6c, I5c–O–I6c, and

2826 | J. Mater. Chem., 2009, 19, 2822–2829 This journal is ª The Royal Society of Chemistry 2009
View Online

To investigate the effect of iodine atoms on the electronic


structures of TiO2, the energy band structures of I-doped (101)
surfaces are presented in Fig. 8(A–F). With respect to the
undoped (101) surface, two important features are apparent.
Firstly, the band gaps of the six doping models investigated here
are similar, ranging from 2.92 eV to 3.11 eV, which are also close
to the undoped case (3.09 eV), suggesting that the surface doping
and different doping configurations have little effect on the
intrinsic band gap of TiO2. Secondly, all I-doped surfaces possess
rich interband states. These new states are located above the
Fermi level for the I–O–I configurations and are unoccupied,
while they lie beneath the Fermi level for the I–O–Ti configura-
tions and are occupied, suggesting that both the charge transfer
Downloaded by UNIVERSITY OF SOUTH AUSTRALIA on 09 November 2012

properties and oxidative power of these states strongly depend on


Published on 11 March 2009 on http://pubs.rsc.org | doi:10.1039/B820816F

doping configurations. With respect to the bulk doping, the first


feature regarding the band gap listed above is the essential
difference, which may serve as an indicator of surface doping.

3.6 Discussion
The ionic radius of Ti4+ is 0.68 Å, and that of I7+ is 0.50 Å. The
covalent radii for Ti and I are 1.32 Å and 1.33 Å. From the
viewpoint of radius matching, it is possible for I to substitute Ti.
Due to the I–O repulsion, the replacement of lattice titanium
with iodine causes distortion of the crystal lattice as shown in the
above calculation models. Then, where is the iodine dopant
located, in the bulk or at the surface of our experimental TiO2?
Fig. 8 Structural models (a–f) and calculated energy band structures Our bulk doping calculation results showed that, when the
(A–F) for doping on the (101) surface with different configurations: (a, A) doping content is up to 2/16 (I/Ti + I), a serious reduction in the
I5c–O–I5c; (b, B) I5c–O–I6c; (c, C) I6c–O–I6c; (d, D) I5c–O–Ti; (e, E) I5c/6c–
band gap results, together with the introduction of abundant
O–Ti; (f, F) I6c–O–Ti. Oxygen, titanium and iodine atoms are indicated
localized states between VB and CB, whilst such a reduction is
by red, grey and blue spheres, respectively. The numbers in parentheses
are the calculated binding energies (eV).
not observed for surface doping atomic configurations, which is
the main difference between bulk doping and surface doping.
Experimentally, I/Ti + I determined from XPS is more than 4/16.
Supposing that the iodine atoms are mainly located in the bulk of
I5c–O–I5c, with the first three for I–O–Ti doping and the other TiO2, two effects are generated. One is the remarkably decreased
three for I–O–I doping, as shown in Fig. 8 (a)–(f). Strong surface intrinsic band gap according to calculated results, and the other
distortions are observed in Fig. 8-(a), (b) and (c), resulting from is that such heavy doping of iodine into the bulk TiO2 will result
the I–O repulsions as elaborated above in relation to the bulk in serious collapse or distortion of the lattice structure. Our UV-
doping. To compare the relative stabilities of these surfaces with visible absorption results showed that the intrinsic band gap of
different doping configurations, the binding energies were iodine doped titania (3.14 eV) is very close to the band gap of
calculated and are given in parentheses in Fig. 8. In contrast to undoped anatase TiO2 of 3.2 eV, and XRD patterns and
the bulk-doping case, the four heavily distorted surfaces (I5c&I6c, HRTEM images (Fig. S1 and 1D†) demonstrate good retention
I6c–O–I6c, I5c–O–I6c, and I5c–O–I5c) present higher stabilities of the anatase TiO2 crystal structure. These results strongly
than two almost undistorted ones (I5c–O–Ti and I6c–O–Ti), indicate: (i) the surface doping is overwhelming, especially on the
which is not surprising because the surface distortion can often {101} facets which often dominate the surface of samples
release local strain energies effectively. For the bulk doping according to the Wulff construction;48 and (ii) the content of bulk
shown in Fig. 7, the supercell in Fig. 7(B) (I–O–Ti doping) with doping is very low. For the former, the I–O repulsion can be fully
heavy distortion is less stable than that in Fig. 7(A) (I–O–I released with various doping atomic configurations, which is
doping), which indicates that such distortion does not play the energetically preferred. Although the bulk doping is also possible
same role in releasing local strain energies (or, not to the same and may result in reduced band gap and introduce internal dipole
extent) in the bulk. It can be concluded that the stabilities of moments associated with the heavy distortion of the TiO6 octa-
various doping configurations in the I-doped TiO2 surface are hedra,37 the average doping content in the bulk should be very
overwhelmingly determined by whether the local strain energies low due to the strong I–O repulsion, otherwise remarkable
resulting from the I–O repulsions can be effectively released. This changes in the intrinsic band gap and the lattice structure should
may serve as a guideline to predict the distribution of iodine be observed in the UV-visible adsorption measurement and
atoms. Furthermore, it should be emphasized that surface XRD patterns. It is therefore reasonable that most iodine atoms
doping appears to be energetically preferable to bulk doping. are located within the surface of TiO2.

This journal is ª The Royal Society of Chemistry 2009 J. Mater. Chem., 2009, 19, 2822–2829 | 2827
View Online

surface properties and consequent photocatalytic activity of


photocatalysts.

Acknowledgements
The financial support from the External Cooperation Program of
Chinese Academy of Sciences (Grant No.GJHZ200815), Major
Basic Research Program, Ministry of Science and Technology
China (No. 2009CB220001) and Australian Research Council
through its Centre’s grant and DP0666345 is gratefully
acknowledged.

Scheme 1 The mechanism of photon excitations via different pathways


References
Downloaded by UNIVERSITY OF SOUTH AUSTRALIA on 09 November 2012

under the irradiation of UV and visible light, and subsequent surface


reactions of photoexcited electrons and holes on iodine doped anatase
Published on 11 March 2009 on http://pubs.rsc.org | doi:10.1039/B820816F

1 (a) A. L. Linsebigler, G. Q. Lu and J. T. Jr. Yates, Chem. Rev., 1995,


TiO2 with I–O–I and I–O–Ti surface structures. CB: conduction band, 95, 735; (b) M. R. Hoffmann, S. T. Martin, W. Choi and
VB: valence band. D. W. Bahnemann, Chem. Rev., 1995, 95, 69.
2 (a) H. Tada, T. Mitsui, T. Kiyonaga, T. Akita and K. Tanaka, Nature
Mater., 2006, 5, 782; (b) T. Kawahara, Y. Konishi, H. Tada,
As a result of surface doping with two atomic configurations of N. Tohge, J. Nishii and S. Ito, Angew. Chem. Int. Ed., 2002, 41, 2811.
3 (a) M. Anpo and M. Takeuchi, J. Catal., 2003, 216, 505; (b)
iodine dopant, I–O–I and I–O–Ti, two new Raman active modes Z. G. Zou, J. H. Ye, K. Sayama and H. Arakawa, Nature, 2001,
at 182 cm 1 and 192 cm 1 and an Eg mode shift from 144 cm 1 to 414, 625; (c) I. Tsuji, H. Kato, H. Kobayashi and A. Kudo, J. Am.
146 cm 1 in iodine doped anatase TiO2 can be explained in terms Chem. Soc., 2004, 126, 13406; (d) J. C. Yu, G. Li, X. Wang, X. Hu,
C. W. Leunga and Z. Zhang, Chem. Commun, 2006, 2717.
of the different force constant of the I–O–Ti structure compared 4 (a) T. Umebayashi, T. Yamaki, H. Itoh and K. Asai, J. Phys. Chem.
to that of Ti–O–Ti and the greatly destroyed symmetry of the Ti– Solids, 2002, 63, 1909; (b) I. Justicia, P. Ordejon, G. Canto,
O–Ti network (Fig. 8a–c) by the I–O–I structure, respectively. J. L. Mozos, J. Fraxeda, G. A. Battiston, R. Gerbasi and
Also importantly, in contrast to the sole I–O–I or I–O–Ti A. Figueras, Adv. Mater., 2002, 14, 1399; (c) M. J. Paek, T. W. Kim
and S. J. Hwang, J. Phys. Chem. Solids, 2008, 69, 1444.
structure, the much lower energy photon excitation pathways 5 T. L. Thompson and J. T. Jr. Yates, Chem. Rev., 2006, 106, 4428.
from the occupied states of the I–O–Ti structure just above the 6 X. Chen and S. S. Mao, Chem. Rev., 2007, 107, 2891.
VB to the unoccupied states of the I–O–I structure below CB in 7 R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki and Y. Taga, Science,
2001, 293, 269.
the band gap are directly responsible for the greatly extended
8 (a) S. U. M. Khan, M. Al-Shahry and W. B. Jr. Ingler, Science, 2002,
light absorption edge up to ca. 800 nm of the iodine doped TiO2 297, 2243; (b) K. Maeda, K. Teramura, D. L. Lu, T. Takata, N. Saito,
with coexisting I–O–I and I–O–Ti structures. Furthermore, the Y. Inoue and K. Domen, Nature, 2006, 440, 7082.
occupied states very close to the VB give rise to the strong 9 (a) C. Burda, Y. B. Lou, X. B. Chen, A. C. S. Samia, J. Stout and
J. L. Gole, Nano lett., 2003, 3, 1049; (b) S. Sakthivel and H. Kisch,
oxidative power of generated holes on them, which, together with ChemPhysChem, 2003, 4, 487; (c) S. Sakthivel, M. Janczarek and
the favorable surface structure for surface transfer of carriers, H. Kisch, J. Phys. Chem. B, 2004, 108, 19384; (d) S. Livraghi,
contributes to the efficient photoactivity of iodine doped TiO2 M. C. Paganini, E. Giamello, A. Selloni, C. Di Valentin and
beyond 600 nm. The mechanism of photon excitations via G. Pacchioni, J. Am. Chem. Soc., 2006, 128, 15666; (e) H. Kisch,
S. Sakthivel, M. Janczarek and D. Mitorraj, J. Phys. Chem. C,
different pathways under the irradiation of UV and visible light, 2007, 111, 11445; (f) X. B. Chen, Y. B. Lou, A. C. S. Samia,
and subsequent surface reactions of photoexcited electrons and C. Burda and J. L. Gole, Adv. Funct. Mater., 2005, 15, 41; (g)
holes on iodine doped anatase TiO2 with I–O–I and I–O–Ti E. Martı́nez-Ferrero, Y. Sakatani, C. Boissiere, D. Grosso,
A. Fuertes, J. Fraxedas and C. Sanchez, Adv. Funct. Mater., 2007,
surface structures, is shown in Scheme 1. 17, 3348; (h) S. S. Soni, M. J. Henderson, J. F. Bardeau and
A. Gibaud, Adv. Mater., 2008, 20, 1493; (i) D. Z. Li, Z. X. Chen,
Y. L. Chen, W. J. Li, H. J. Huang, Y. H. He and X. Z. Fu,
Environ. Sci. Technol., 2008, 42, 2130; (j) J. Zhou, L. Lv, J. Yu,
4. Conclusions H. Li, P. Guo, H. Sun and X. S. Zhao, J. Phys. Chem. C, 2008,
112, 5316; (k) S. Yin, H. Yamaki, M. Komatsu, Q. Zhang,
A highly efficient photocatalyst was prepared by doping TiO2 J. Wang, Q. Tang, F. Saito and T. Sato, J. Mater. Chem., 2003, 13,
2996; (l) T. Sano, N. Negishi, K. Koike, K. Takeuchi and
with surface-dominant I–O–I and I–O–Ti structures. This pho- S. Matsuzawa, J. Mater. Chem., 2004, 14, 380; (m) G. Liu, F. Li,
tocatalyst was found to have an extended absorption edge up to Z. Chen, G. Q. Lu and H.-M. Cheng, J. Solid State Chem., 2006,
around 800 nm and resulted in visible light photocatalytic 179, 331; (n) G. Liu, F. Li, D. Wang, D. Tang, C. Liu, X. Ma,
G. Q. Lu and H.-M. Cheng, Nanotechnology, 2008, 19, 025606; (o)
activity beyond 600 nm. The much lower energy photon excita- G. Liu, X. Wang, Z. Chen, H. M. Cheng and G. Q. Lu, J. Colloid
tion pathways from the occupied states of I–O–Ti structure just Interface Sci., 2009, 329, 331.
above the VB to the unoccupied states of I–O–I structure below 10 (a) S. Sakhivel and H. Kisch, Angew. Chem. Int. Ed., 2003, 42, 4908;
the CB in the band gap are concluded as the cause of the ultra- (b) L. W. Zhan, H. B. Fu and Y. F. Zhu, Adv. Funct. Mater., 2008, 18,
2180.
long light response of iodine doped anatase TiO2. The low levels 11 T. Umebayashi, T. Yamaki, H. Itoh and K. Asai, Appl. Phys. Lett.,
of occupied states just above the valence band together with the 2002, 81, 454.
constructed favorable surface structure are responsible for 12 W. Zhao, W. Ma, C. Chen, J. Zhao and Z. Shuai, J. Am. Chem. Soc.,
2004, 126, 4782.
the photoactivity beyond 600 nm. This work has demonstrated
13 (a) X. T. Hong, Z. P. Wang, W. M. Cai, F. Lu, J. Zhang, Y. Z. Yang,
the decisive role that controlling surface atomic configurations N. Ma and Y. J. Liu, Chem. Mater., 2005, 17, 1548; (b) W. Ho,
of the dopant can play in achieving favorable electronic and J. C. Yu and S. Lee, Chem. Commun., 2006, 1115.

2828 | J. Mater. Chem., 2009, 19, 2822–2829 This journal is ª The Royal Society of Chemistry 2009
View Online

14 S. In, O. Alexander, B. Regina, G. Felipe, P.-J. Sergio, T. S. Mintcho, 


32 J. Graciani, L. J. Alvarez, J. A. Rodriguez and J. F. Sanz, J. Phys.
W. S. Dominic and L. M. Richard, J. Am. Chem. Soc., 2007, 129, Chem. C, 2008, 112, 2624.
13790. 33 G. Liu, Z. Chen, C. Dong, Y. Zhao, F. Li, G. Q. Lu and H. M. Cheng,
15 M. Batzill, E. H. Morales and U. Diebold, Phys. Rev. Lett., 2006, 96, J. Phys. Chem. B, 2006, 110, 20823.
026103. 34 S. Tojo, T. Tachikawa, M. Fujitsuka and Tetsuro Majima, J. Phys.
16 C. D. Valentin, G. Pacchioni, A. Selloni, S. Livraghi and E. Giamello, Chem. C, 2008, 112, 14948.
J. Phys. Chem. B, 2005, 109, 11414. 35 Z. He, X. Xu, S. Song, L. Xie, J. Tu, J. Chen and B. Yan, J. Phys.
17 H. Irie, Y. Watanabe and K. Hashimoto, J. Phys. Chem. B, 2003, 107, Chem. C, 2008, 112, 16431.
5483. 36 W. Y. Su, Y. F. Zhang, Z. H. Li, L. Wu, X. X. Wang, J. Q. Li and
18 T. Lindgren, J. M. Mwabora, E. Avendano, J. Jonsson, A. Hoel, X. Z. Fu, Langmuir, 2008, 24, 3422.
C. G. Granqvist and S. E. Lindquist, J. Phys. Chem. B, 2003, 107, 5709. 37 M. Long, W. Cai, Z. Wang and G. Liu, Chem. Phys. Lett., 2006, 420,
19 G. Liu, Y. Zhao, C. Sun, F. Li, G. Q. Lu and H. M. Cheng, Angew. 71.
Chem. Int. Ed., 2008, 47, 4516. 38 K. Yang, Y. Dai, B. Huang and M.-H. Whangbo, Chem. Mater.,
20 V. Gombac, L. De Rogatis, A. Gasparotto, G. Vicario, T. Montini, 2008, 20, 6528.
D. Barreca, G. Balducci, P. Fornasiero, E. Tondello and 39 T. Hirakawa and Y. Nosaka, Langmuir, 2002, 18, 3247.
M. Graziani, Chem. Phys., 2007, 339, 111. 40 W. Kohn and L. M. Sham, Phys. Rev., 1965, 140, A1133.
Downloaded by UNIVERSITY OF SOUTH AUSTRALIA on 09 November 2012

21 T. Umebayashi, T. Yamaki, S. Yamamoto, A. Miyashita, S. Tanaka, 41 J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77,
T. Sumita and K. Asai, J. Appl. Phys., 2003, 93, 5156. 3865.
Published on 11 March 2009 on http://pubs.rsc.org | doi:10.1039/B820816F

22 C. D. Valentin, G. Pacchioni and A. Selloni, Phys. Rev. B, 2004, 70, 42 G. Kresse and J. Joubert, Phys. Rev. B, 1999, 59, 1758.
085116. 43 B. J. Delley, J. Chem. Phys., 1990, 92, 508.
23 H. Kamisaka, H. Adachi and K. Yamashita, J. Chem. Phys., 2005, 44 B. J. Delley, J. Chem. Phys., 2000, 114, 7756.
123, 084704. 45 S. M. Prokes, J. L. Gole, X. Chen, C. Burda and W. E. Carlos, Adv.
24 C. D. Valentin, G. Pacchioni and A. Selloni, Chem. Mater., 2005, 17, Funct. Mater., 2005, 15, 161.
6656. 46 (a) J. W. Tang, Z. G. Zou, J. Yin and J. Ye, Chem. Phys. Lett., 2003,
25 K. Yang, Y. Dai, B. Huang and S. Han, J. Phys. Chem. B, 2006, 110, 382, 175; (b) Y. Sakatani, D. Grosso, L. Nicole, C. Boissiere, G. Soler-
24011. Illia and C. Sanchez, J. Mater. Chem., 2006, 16, 77; (c) D. Chen and
26 N. Serpone, J. Phys. Chem. B, 2006, 110, 24287. J. Ye, Adv. Funct. Mater., 2008, 18, 1922; (d) I. S. Cho, D. W. Kim,
27 V. Kuznetsov and N. Serpone, J. Phys. Chem. B, 2006, 110, 25203. S. Lee, C. H. Kwak, S. T. Bae, J. H. Noh, S. H. Yoon, D. W. Kim
28 A. V. Emeline, N. V. Sheremetyeva, N. V. Khomchenko, and K. S. Hong, Adv. Funct. Mater., 2008, 18, 2154; (e) F. Gao,
V. K. Ryabchuk and N. Serpone, J. Phys. Chem. C, 2007, 111, 11456. X. Y. Chen, K. B. Yin, S. Dong, Z. F. Ren, F. Yuan, T. Yu,
29 U. Diebold, Surf. Sci. Rep., 2003, 48, 53. Z. Zou and J. M. Liu, Adv. Mater., 2007, 19, 2889.
30 X. Q. Gong, A. Selloni, M. Batzill and U. Diebold, Nature Mater., 47 J. K. Burdett, T. Hughbanks, G. J. Miller, J. J. W. Richardson and
2006, 5, 66. J. V. Smith, J. Am. Chem. Soc., 1987, 109, 3639.
31 E. Finazzi, C. Di Valentin, A. Selloni and G. Pacchioni, J. Phys. 48 M. Lazzeri, A. Vittadini and A. Selloni, Phys. Rev. B, 2001, 63,
Chem. C, 2007, 111, 9275. 155409.

This journal is ª The Royal Society of Chemistry 2009 J. Mater. Chem., 2009, 19, 2822–2829 | 2829

You might also like