Short Communication
The optical absorption and hydrogen production by water
splitting of (Si,Fe)-codoped anatase TiO
2
photocatalyst
Yanming Lin
a,b, Zhenyi Jiang
a,*
, Chaoyuan Zhu
b, Xiaoyun Hu
c, Haiyan Zhu
a,
Xiaodong Zhang
a, Jun Fan
d, Sheng Hsien Lin
b,eaInstitute of Modern Physics, Northwest University, Xi’an 710069, PR China
bDepartment of Applied Chemistry, Institute of Molecular Science and Center for Interdisciplinary Molecular Science,
National Chiao-Tung University, Hsinchu 30050, Taiwan
c
Department of Physics, Northwest University, Xi’an 710069, PR China
d
School of Chemical Engineering, Northwest University, Xi’an 710069, PR China
eInstitute of Atomic and Molecular Sciences, Academia Sinica, Taipei 106, Taiwan
a r t i c l e i n f o
Article history:
Received 7 December 2012 Received in revised form 10 February 2013
Accepted 16 February 2013 Available online 16 March 2013
Keywords: TiO2
Codoped
Photocatalytic activity for hydrogen production
Density functional theory
a b s t r a c t
The electronic and optical properties are studied using the density functional theory in (Si,Fe)-codoped anatase TiO2. The calculated results suggest that the synergistic effects of
(Si,Fe) codoping can effectively induce the redshift of optical absorption edge, which leads to higher visible-light photocatalytic activity for hydrogen production by water splitting than pure anatase TiO2. To verify the reliability of our calculated results, nanocrystalline
(Si,Fe)-codoped TiO2is synthesized by a sol-gel-solvothermal method, and excellent
ab-sorption performance and photocatalytic activity for hydrogen production by water split-ting are observed in our experiments.
Copyrightª 2013, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
Among many candidates of semiconductor photocatalyst, titania (TiO2) has become the most investigated one for overall
water splitting for hydrogen production, due to its outstanding chemical stability, low cost, and non-toxicity
[1,2]. However, the photocatalytic water splitting for hydrogen production of TiO2are restricted to ultraviolet
(UV)-light (l < 385 nm) due to the wide band gap of anatase TiO2
(w3.2 eV). Therefore, reducing the band gap of anatase TiO2to
make it photosensitive to visible-light has become one of the
most important goals in photocatalytic water splitting for hydrogen production. It has been suggested that doping with different cations and anions would result in a reduced band gap for TiO2[3e11]. For example, Yang et al. studied
system-atically the nitrogen concentration influence on N-doped anatase TiO2[6]. The results indicate that some localized N 2p
states are formed above the valence band in N-doped anatase TiO2at lower doping levels, which leads to the reduction of the
photon transition energy. And the energy gap has little further
* Corresponding author. Tel.:þ86 29 88303491; fax: þ86 29 88302331.
E-mail addresses:[email protected](Y. Lin),[email protected](Z. Jiang).
Available online at
www.sciencedirect.com
journal hom epa ge: www.elsev ier.com/locate/he
0360-3199/$e see front matter Copyright ª 2013, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
narrowing compared with that at lower doping levels when the doping level rises. A few latest researches indicate that different cations and anions codoping into TiO2can further
narrow its band gap and enhance its photocatalytic activity
[12e20]. For instance, Jia et al. studied the microscopic mechanism for band gap narrowing and the origin of the enhanced visible-light photocatalytic activity in N/S-codoped anatase TiO2 [13]. Li et al. reported that the C/H-codoping
produces significant band gap narrowing, which leads to higher visible-light photocatalytic efficiency than the C-doped anatase TiO2[16]. The research of Su et al. suggested that the
codoping of TiO2with N and Fe leads to the much narrowing of
the band gap and greatly improves the photocatalytic activity under visible irradiation [17]. The results showed that codoping is one of the most effective approaches to extend the absorption edge to the visible-light region in anatase TiO2.
However, photocatalytic activity for hydrogen production by water splitting and optical absorption properties of (Si,Fe)-codoped TiO2has no report on the theory and experiment.
Therefore, the enhanced visible-light absorption efficiency and photocatalytic activity are expected for (Si,Fe)-codoped TiO2.
In this letter, the electronic and optical properties of (Si,Fe)-codoped TiO2are investigated using the density functional
theory (DFT) to reveal the synergistic effects of (Si,Fe) codop-ing on the mechanism of bandgap reduccodop-ing and the origin of enhanced visible-light photocatalytic activity for hydrogen production by water splitting. Nanocrystalline (Si,Fe)-codoped TiO2was synthesized with sol-gel-solvothermal method. It
was found that the (Si,Fe)-codoped TiO2 showed excellent
photocatalytic activity for hydrogen production, which veri-fied the reliability of our calculated results.
All the spin-polarized calculations were performed using the projector augmented wave pseudopotentials as imple-mented in the VASP code[21,22]. The exchange correlation function was treated by the generalized gradient approxima-tion (GGA) with the PerdeweWang parameterization (known as GGA-PW91) [23]. The Brillouin-zone integrations were approximated by using the special k-point sampling of the MonhkorstePack scheme[24]. A cutoff energy of 500 eV and a mesh size of 9 9 9 were used for geometry optimization
and electronic property calculations. Using the block David-son scheme, both the atomic positions and cell parameters were optimized until the residual forces were below 0.01 eV/A. To obtain the band gap that was consistent with the experi-mental result, the GGAþ U method[25]was employed. The Coulombic interaction U and exchange energy J were set to be 10.0 eV and 1.0 eV, respectively. Accordingly, the calculated band gap of pure anatase TiO2was 2.9 eV, which was in good
agreement with the experimental value[26].
The valence electron configurations considered in this study included Ti (3d24s2),O (2s22p4), Si (3s23p2), and Fe (3d64s2).
All the doped systems were constructed from a relaxed (2 2 1) 48-atom anatase TiO2supercell and it is shown in
Fig. 1(a). As the position of Fe in the TiO2lattice was unclear,
variety of positions of Fe atoms in the lattice were considered, such as substitutional Fe at the Ti site (Fe@Ti) and O site (Fe@O). In the Si-doped TiO2, a Ti atom is substituted by a Si
atom (Si@Ti)[27]. Similar substitutions were also considered for codoped systems, as Fe locates at either Ti or O site and Si locates at Ti site, namely, Si@Ti&Fe@Ti and Si@Ti&Fe@O. To study the stabilities of the doped systems, we calculated the defect formation energy (Ef) for the doped and codoped
sys-tems according to the equations
EfðX@YÞ¼ EðX@YÞ EðpureÞ ðmX mYÞ (1)
EfðSi@Y&Fe@YÞ ¼ EðSi@Y&Fe@YÞ EðpureÞ ðmSiþ mFe mY mYÞ (2)
where X¼ Si, Fe; Y ¼ Ti, O; E represents the total energy and m is the chemical potential. The calculated formation energies are listed inTable 1. It shows that Fe impurity is preferred to sub-stitute Ti in lattice because of the smallest formation energy in both Fe-doped and (Si,Fe)-codoped anatase TiO2systems.
To investigate the electronic properties of Si and/or Fe (co) doping anatase TiO2, the total density of states (TDOS) and
partial density of states (PDOS) were plotted inFig. 2. It in-dicates that the valence band (VB) is dominated by O 2p states while the conduction band (CB) consists mainly of Ti 3d states for pure anatase TiO2. In Si-doped TiO2 (Si@Ti), the VB
broadens with the mixing of O 2p and Si 3p states, and the CB bottom has a decline of about 0.15 eV, which can lead to a
Fig. 1e (A) 48-atom supercell model for defective anatase TiO2shows the location of the dopants. The atom doping sites are
marked with Si and Fe. The gray spheres and red spheres represent the Ti and O atoms, respectively. The purple sphere and cyan sphere represent Si and Fe atom, respectively. (b) XRD patterns for the pure and doped TiO2. (For interpretation of the
band gap narrowing. For Fe-doped TiO2(Fe@Ti), it is shown
that the band gap decreases by about 0.6 eV and most Fe 3d states are located in the band gap compared with the pure anatase TiO2, which may be due to stronger interactions
be-tween the Fe 3d and Ti 3d orbitals. For (Si,Fe)-codoped TiO2
system (Si@Ti&Fe@Ti), some impurity states (Si 3p and Fe 3d ) are mixed with the VB and CB edge. The top of the VB has an obvious upward shift while the CB bottom has an obvious downward shift, which results in a band gap narrowing of about 1.0 eV compared with the pure anatase TiO2. Therefore,
synergistic effect of (Si,Fe)-codoped can lead to a decrease of the photon excitation energy and redshift the optical ab-sorption edge to the visible-light range.
According to the obtained electronic structures, we calcu-lated the complex dielectric function x ¼ x1 þ ix2. The
corresponding absorption spectrum was estimated by the following equation IðuÞ ¼ 2u x2 1ðuÞ þ x22ðuÞ 1=2 x1ðuÞ 2 !1=2 (3)
where I is the optical absorption coefficient,u is the angular frequency (E¼ Zu).
The absorption spectra of the pure and doped anatase TiO2
systems are calculated and shown inFig. 3. It is found that pure anatase TiO2can only respond to the UV-light and shows
no absorption activity in the visible-light region. For Si-doped system, it is clear that the narrowed band gap can result in the reduction of the photon transition energy from the VB to the CB, which induces the increasing optical absorption in the UV-light region. For Fe-doped system, there are a series of impu-rity states (Fe 3d orbital) appearing in the forbidden gap and the band gap has an obvious narrowing compared with the pure anatase TiO2. These results indicate that the electrons
are excited easily from the VB to the CB through the Fe 3d states under the visible-light irradiation, which can lead to a good optical absorption for Fe-doped TiO2in the visible-light
region. For (Si,Fe)-codoped TiO2system, synergistic effect of
(Si,Fe) codoping induces a band gap narrowing and appearing Fe 3d states in the forbidden gap, which lead to a decrease of Table 1e Defect formation energies Effor different doped
anatase TiO2systems.
Doped models
Doped Codoped Si@Ti Fe@Ti Fe@O Si@Ti and
Fe@Ti
Si@Ti and Fe@O
Ef(eV) 8.7293 1.5953 12.7225 11.2535 1.7695
Fig. 2e Calculated TDOS and PDOS of different doped TiO2. The top of the valence band of pure anatase TiO2is taken as the
the photon excitation energy in the view of electronic struc-ture. Therefore, the absorption of visible- and UV-light is greatly enhanced in (Si,Fe)-codoped anatase TiO2compared
with the pure, Si- and Fe-doped anatase TiO2, which may be
responsible for the redshift of optical absorption edge and the outstanding activity for hydrogen generation by photo-catalytic water splitting in (Si,Fe)-codoped anatase TiO2.
To confirm the better photocatalytic activity for hydrogen production by water splitting of the (Si,Fe)-codoped TiO2
compared to that of pure TiO2, we further observed the
UVevis absorption spectrum by experiments.
Nanocrystalline pure, Si-, Fe-, and (Si,Fe)-codoped TiO2
were prepared by a sol-gel-solvothermal method. Firstly, a desired amount (0.3367 g) of Fe(NO3)39H2O was dissolved in
39.66 mL of CH3COOH solution under stirring. Then, 5.58 mL of
(C2H5O)4Si was dropwise added into the solution with stirring
for 1 h. Secondly, 28.36 mL of [CH3(CH2)3O]4Ti was also
drop-wise added into the solution with continuous stirring for 2 h, and the solution was heated in an oven and kept at 140C for 14 h. Finally, the precipitate obtained was dried in a vacuum oven at 70C for 48 h. To evaluate the photocatalytic activity of samples, hydrogen generation by photocatalytic water split-ting was performed in a cylindrical quartz photo-reactor with a 1000 mL capacity. A flow of dry N2gas was used to purge
dissolved O2in the reactor for 30 min prior to illumination. A
500 W long-arc xenon lamp surrounded with a water cooling system was fixed in the center of the reaction cell. Photo-catalyst (0.4 g) was suspended in 60 mL ethanol and 540 mL distilled water under stirring magnetically. H2was analyzed
by gas chromatograph (GC) using a Fuli GC-9790II (ZheJiang, China), equipped with a thermal conductivity detector (TCD) and a stainless steel column (2 m) packed with molecular sieves (5 A) at 323 K.
The crystalline phase was identified by X-ray diffraction (XRD) (Rigaku D/MAX-2400). The BrunauereEmmetteTeller (BET) surface area of the samples was measured through ni-trogen adsorption at 77 K (Nova 2000e). The UVevis absorption spectra were obtained on an UVevis spectrophotometer (UV-3600) and using BaSO4as the reference sample.
Fig. 1(b) shows the XRD patterns of the samples of pure, Si-, Fe-, and (Si,Fe)-codoped TiO2. It is found that all of the
diffraction peaks are contributed by the anatase TiO2phase
and no other visible impurity peak can be distinguished in the pattern of pure or doped sample. The BET surface areas for pure, Si-, Fe-, and (Si,Fe)-codoped TiO2are 84.21, 252.42, 226.72
and 308.31 m2/g, respectively. It is shown that the surface area
of TiO2powders is increased to 308.31 m2/g with the
coexis-tence of Si and Fe in TiO2, about four times of that of pure TiO2
powders.
The optical absorption spectra of the pure and doped sys-tems are measured by experiments, and shown in Fig. 3. Compared with the pure TiO2, it is clear that the incorporation
of Si into TiO2lattice induces the enhanced optical absorption
in the UV-light region. For Fe-doped system, it exhibits an excellent absorption activity in the visible-light region. For (Si,Fe)-codoped TiO2, it is obvious that the optical absorption
in the UVevisible region is stronger than that of pure TiO2,
especially in the visible-light region. The enhancement of absorption in the visible-light region can promote the utili-zation of the solar light for the doping TiO2, which enhances
the visible-light photocatalytic activity of TiO2for hydrogen
production by water splitting. However, there are small mis-alignments between the experimental and theoretical results, which may be due to neglect of the anisotropy of the ab-sorption coefficient and the well-known limitation of DFT. But, from the perspective of qualitative analysis, the experi-mental results are consistent with the calculations.
The photocatalytic activity of the pure and doped TiO
2-samples was evaluated by hydrogen production from water splitting, as shown inFig. 4(a). Pure water produces very low amount of hydrogen under the visible-light irradiation without photocatalyst, indicating that the photolysis can be ignored. It is obvious that Si, Fe, and (Si,Fe) (co)doping can improve the photocatalytic activity of hydrogen production in TiO2material. Among them, (Si,Fe)-codoped TiO2sample
ex-hibits the best photocatalytic activity of hydrogen production compared with the pure TiO2, which may be due to a stronger
absorption of solar light in (Si,Fe)-codoped TiO2. In addition,
Fig. 4(b) shows the VB and CB edge potentials of pure and (Si,Fe)-codoped TiO2vs. normal hydrogen electrode (NHE). The
CBM potential of (Si,Fe)-codoped TiO2is0.6 eV, more
nega-tive than Hþ/H2(0 eV). The results show that it has the ability
to reduce Hþ to produce H2, and the reductive ability is
stronger than pure TiO2. Therefore, TiO2exhibits outstanding
Fig. 3e The optical absorption curves of (a, a’) pure, (b, b’) Si-doped, (c, c’) Fe-doped, and (d, d’) (Si,Fe)-codoped TiO2. The left
photocatalytic activity for hydrogen production by water splitting through Si and Fe codoping.
In conclusion, we have investigated the electronic and optical properties of (Si,Fe)-codoped TiO2based on DFT
cal-culations. The synergistic effects of (Si,Fe) codoping may further reduce the electrons excited energy from VB to CB compared with the pure TiO2under the solar light irradiation,
which enhances the photocatalytic activity for hydrogen production by water splitting and induces the redshift of sorption edge. The photocatalytic splitting water and ab-sorption spectra obtained by experiments indicate that (Si,Fe)-codoped TiO2sample has a much stronger absorption of the
solar light and photocatalytic activity for hydrogen production by water splitting than the pure TiO2, which verifies the
reli-ability of the calculation results.
Acknowledgements
Yanming Lin would like to thank Dr. Kesong Yang and Dr. Run Long for helpful discussions. This work was supported by the National Natural Science Foundation of China under Grants (Nos. 10647008, 50971099, and 21176199), the Research Fund for the Doctoral Program of Higher Education (Nos. 20096101110017 and 20096101110013), Key Project of Natural Science Foundation of Shaanxi Province of China (Nos. 2010JZ002 and 2011JM1001), and Graduate’s Innovation Fund of Northwest University of China (No. YZZ12082).
r e f e r e n c e s
[1] Fujishima A, Honda K. Electrochemical photolysis of water at a semiconductor electrode. Nature 1972;238:37e8.
[2] Antony RP, Mathews T, Ramesh C, Murugesan N,
Dasgupta A, Dhara S, et al. Efficient photocatalytic hydrogen generation by Pt modified TiO2nanotubes fabricated by rapid
breakdown anodization. Int J Hydrogen Energy 2012;37:8268e76.
[3] Yang K, Dai Y, Huang B, Whangbo MH. Density functional characterization of the visible-light absorption in
substitutional C-anion and C-cation doped TiO2. J Phys Chem
C 2009;113:2624e9.
[4] Asahi R, Morikawa T, Ohwaki T, Aoki K, Taga Y. Visible-light photocatalysis in nitrogen-doped titanium oxides. Science 2001;293:269e71.
[5] Lin YM, Jiang ZY, Zhu CY, Hu XY, Zhang XD, Fan J. Visible-light photocatalytic activity of Ni-doped TiO2from ab initio
calculations. Mater Chem Phys 2012;133:746e50.
[6] Yang K, Dai Y, Huang B. Study of the nitrogen concentration influence on N-doped TiO2anatase from first-principles
calculations. J Phys Chem C 2007;111:12086e90. [7] Khan SUM, Al-Shahry M, Ingler Jr WB. Efficient
photochemical water splitting by a chemically modified n-TiO2. Science 2002;297:2243e5.
[8] Spadavecchia F, Cappelletti G, Ardizzone S, Ceotto M, Falciola L. Electronic structure of pure and N-doped TiO2
nanocrystals by electrochemical experiments and first principles calculations. J Phys Chem C 2011;115:6381e91.
[9] Lu J, Dai Y, Guo M, Yu L, Lai K, Huang B.
Chemicalandopticalpropertiesofcarbon-dopedTiO2: a
densityefunctional study. Appl Phys Lett 2012;100:102114e7.
[10] Gonza´lez-Borrero PP, Bernabe´ HS, Astrath NGC, Bento AC, Baesso ML, Castro Meira MV, et al. Energy-level and optical properties of nitrogen doped TiO2: an experimental and
theoretical study. Appl Phys Lett 2011;99:221909e11. [11] Zhao D, Huang X, Tian B, Zhou S, Li Y, Du Z. The effect of
electronegative difference on the electronic structure and visible light photocatalytic activity of N-doped anatase TiO2by
first-principles calculations. Appl Phys Lett 2011;98:162107e9. [12] Xing MY, Li WK, Wu YM, Zhang JL, Gong XQ. Formation of
new structures and their synergistic effects in boron and nitrogen codoped TiO2for enhancement of photocatalytic
performance. J Phys Chem C 2011;115:7858e65.
[13] Jia L, Wu C, Li Y, Han S, Li Z, Chi B, et al. Enhanced visible-light photocatalytic activity of anatase TiO2through N and S
codoping. Appl Phys Lett 2011;98:211903e5. [14] Long R, English NJ. Synergistic effects on band
gap-narrowing in titania by codoping from first-principles calculations. Chem Mater 2010;22:1616e23.
[15] Lin Y, Jiang Z, Hu X, Zhang X, Fan J. The electronic and optical properties of Eu/Si-codopedanatase TiO2
photocatalyst. Appl Phys Lett 2012;100:102105e8.
[16] Li N, Yao KL, Li L, Sun ZY, Gao GY, Zhu L. Effect of carbon/ hydrogen species incorporation on electronic structure of anatase-TiO2. J Appl Phys 2011;110:073513e7.
[17] Su Y, Xiao Y, Li Y, Du Y, Zhang Y. Preparation, photocatalytic performance and electronic structures of visible-light-driven FeeN-codopedTiO2nanoparticles. Mater Chem Phys
2011;126:761e8.
(a)
(b)
Fig. 4e (A) Hydrogen production by water splitting as a function of reaction time over the pure and doped TiO2samples, and
[18] Long R, English NJ. Band gap engineering of (N, Si)-codoped TiO2from hybrid density functional theory calculations. New
J Phys 2012;14:053007e17.
[19] Gai Y, Li J, Li SS, Xia JB, Wei SH. Design of narrow-gap TiO2: a
passivated codoping approach for enhanced photoelectrochemical activity. Phys Rev Lett 2009;102:036402e5.
[20] Lin Y, Jiang Z, Zhu C, Hu X, Zhang X, Zhu H, et al. Enhanced optical absorption and photocatalytic activity of anatase TiO2
through (Si, Ni) codoping. Appl Phys Lett 2012;101:062106e10. [21] Kresse G, Hafner J. Ab initio molecular-dynamics simulation
of the liquid-metaleamorphous-semiconductor transition in germanium. Phys Rev B 1994;49:14251e69.
[22] Kresse G, Furthmu¨ller J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys Rev B 1996;54:11169e86.
[23] Perdew JP, Ahevary CJ, Vosko SH, Jackson KA,
Pederson MR, Singh DJ, et al. Atoms, molecules, solids, and surfaces: applications of the generalized gradient approximation for exchange and correlation. Phys Rev B 2004;46:6671e87.
[24] Monkhorst HJ, Pack JD. Special points for Brillouin-zone integrations. Phys Rev B 1976;13:5188e92.
[25] Dudarev SL, Botton GA, Savarsov SY, Humphreys CJ, Sutton AP. Electron-energy-loss spectra and the structural stability of nickel oxide: an LSDAþU study. Phys Rev B 1998;57:1505e9.
[26] Tang H, Le´vy F, Berger H, Schmid PE. Urbach tail of anatase TiO2. Phys Rev B 1995;52:7771e4.
[27] Yang K, Dai Y, Huang B. First-principles calculations for geometrical structures and electronic properties of Si-doped TiO2. Chem Phys Lett 2008;456:71e5.
