Document Type: Original Research Paper


1 Molecular Simulation laboratories (MSL) of Azarbaijan Shahid Madani University, Tabriz, Iran

2 Computational Nanomaterials Research Group (CNRG), Azarbaijan Shahid Madani University, Tabriz, Iran

3 Department of Chemistry, Faculty of Basic Sciences, Azarbaijan Shahid Madani University, Tabriz, Iran

4 Department of Physics, Faculty of Basic Sciences, Azarbaijan Shahid Madani University, Tabriz, Iran


First-principles calculations within density functional theory (DFT) have been performed to investigate the interactions of NO2 molecules with TiO2/Gold nanocomposites in order to completely exploit the adsorption properties of these nanostructures. Given the need to further comprehend the behavior of the NO2 molecules positioned between the TiO2 nanoparticle and Au monolayer, we have geometrically optimized the complex systems consisting of the NO2 molecule oriented at appropriate positions between the nanoparticle and Au monolayer. The structural properties such as bond lengths, bond angles, adsorption energies and Mulliken population analysis and the electronic properties including the density of states and molecular orbitals have been also analyzed in detail. The results indicate that the interaction between NO2 and undoped TiO2-N/Gold nanocomposites is stronger than that between gas molecules and N-doped TiO2/Gold nanocomposites, which reveals that the pristine nanocomposite can react with NO2 molecule more efficiently. Therefore, the obtained results also suggest a theoretical basis for the potential applications of TiO2/Gold nanocomposites in gas sensing, which could help in the developing of novel TiO2 based advanced sensor devices.



Titanium dioxide has fascinated exciting interests as the hopeful semiconductor material due to its peculiar properties such as non-toxicity, high catalytic efficiency and extensive band-gap [1]. TiO2 has been broadly employed in numerous fields such as photo-catalysis, gas sensing, organic dye-sensitized solar cells, water-splitting and removal of air pollutants [2, 3, 4, 5]. It has been found in three important crystallographic phases namely anatase, rutile, and brookite [6]. The photocatalytic properties of TiO2 were limited because of its large band gap (3.2 eV) and that’s why the TiO2 only absorbs a minor percentage (3-5 % of the incoming solar light). TiO2 doping with some nonmetal elements such as nitrogen is a suitable method, which would enhance the optical response of TiO2 to the visible area and recover its photocatalytic properties [7, 8]. Newly, oxide supported Au nanoparticles have involved significant attention due to the catalytic activities of these materials at small temperatures [9, 10] for a progressively greater number of oxidation processes. From the experimental viewpoint, several propositions have been considered for the source of the excellent activity of Au nanoparticles. Goodman et al. [11] suggested size and effects of the band structure and measure the onset of the reactivity to look when the Au particles become as tinny as two monolayers at which point they subject to a change from a metallic to a separating state. From the theoretical standpoint, the clarifications of the low-temperature activity of supported Au nanoparticles are also numerous. By now, a number of density functional theory calculations have been published for O2 adsorption and CO+O2 reaction at different Au collections. Hakkinen Landman and co-workers [12] HL) identified in their calculations the charging of the Au particles as an imperative parameter. With the large number of suggestions for essential parameters accounting for the extraordinary high reactivity of supported Au particles, there are many theoretical investigations, which have been considered for more systems. The adsorptions of the O2 and CO2 and reaction at gold nanoparticle models supported by TiO2 surfaces have been investigated by density functional theory study [13]. However, numerous computational studies of N-doped TiO2 anatase and Au nanoparticles have been published separately, explaining some of the important electronic and structural properties of these particles. For instance, Liu et al. [14] studied the adsorption of NO molecule on undoped and N-doped TiO2 anatase nanoparticles. Carrying out first-principles calculations, Liu et al. suggested that the N-doped TiO2 anatase nanoparticles can interact with CO molecule more capably compared to the pristine ones [15]. There are a few computational and experimental studies on the adsorption behaviors of TiO2/Gold nanocomposites. NO2 molecules have been characterized as toxic gases, which are mainly emitted from power plants and vehicle engines. So, controlling the concentrations of these injurious molecules is a vital subject to public health and environmental protection [16]. In this study, the interaction of NO2 molecules with TiO2/Gold nanocomposites has been investigated by DFT computations. We presented here results of calculations of complex systems consisting of NO2 molecule positioned between the TiO2 anatase nanoparticle and Au monolayer. The electronic structures of the adsorption systems have been also analyzed including the projected partial density of states (PDOS) and molecular orbitals. The main aim of this study is to provide a general understanding on the adsorption behaviors of TiO2/Goldnanocomposites as the effective sensor devices.



Density functional theory

Density Functional Theory (DFT) calculations [17, 18] were performed with the Open source Package for Material eXplorer (OPENMX) ver. 3.7 [19], being an efficient software package for nano-scale materials simulations based on DFT, norm-conserving pseudopotantials, and pseudo-atomic localized basis functions [20, 21]. Pseudo atomic orbitals (PAO's) were used as basis sets. In the calculations related to the energetics of the system, the cutoff energy of 150 Rydberg is considered, which is demonstrated to be the most important criterion related to the computational effort and accuracy. The employed PAO's were generated by using the basis sets (tree-s, three-p, one-d) for Ti atom, (three-s, three-p, two-d, one-f) for Au atom, (two-s and two-p) for O and N atoms, within cutoff radii set to the values of 7 for Ti, 9 for Au, 5 for O and N (all in Bohrs). As an exchange-correlation energy functional approximation, the generalized gradient approximation (GGA) in the Perdew–Burke–Ernzerhof form (PBE)has been utilized[22]. The convergence criteria of the self-consistent field (SCF) and energy calculation have been set to the values of 1.0 × 10-6 Hartree and 1.0 × 10-4 Hartree/bohr, respectively. In the course of geometry optimization, ’Opt’ is used as the geometry optimizer, which is a robust and efficient scheme. All atoms of the considered nanocomposites were allowed to relax. The chosen pristine and N-doped nanocomposites are placed in a 20 Ǻ × 20 Ǻ ×30 Ǻ box, which is much bigger than the size of the composites. The thickness of the vacuum spacing is 11.5 Å, which is essential for reducing the interaction between the nearest particles according to the periodic boundary conditions. The open-source program XCrysDen [23] was employed for picturing data such as molecular orbitals and to prepare the Figs. considered in this study. The box considered in these computations contains 86 atoms (14 Au, 48 O and 24 Ti atoms) of undoped or N-doped TiO2 nanoparticle with Au surface. The Gaussian broadening method for evaluating electronic DOS is used. A 21×21×21 ᴦ-centered k-point mesh is used in the calculations associated with the electronic DOS. The adsorption energy of gas molecules adsorbed on the TiO2/Gold nanocompopsite is evaluated using the following formula:

Ead = E (composite + adsorbate) - E composite – E adsorbate

where E( composite + adsorbate ) is the total energy of the adsorption system, E compositeis the energy of the TiO2/Gold nanocomposite, and E adsorbateis the energy of isolated gas molecules. According to this definition, the adsorption energies of stable configurations are negative; the more negative the adsorption energy, the more energy favorable the complex system.


Modeling TiO2/Gold nanocomposites

The considered TiO2 anatase nanoparticles containing 72 atoms were made by setting 3×2×1 numbers of TiO2 unit cells along x, y and z axis, respectively. The unit cell was taken from "American Mineralogists Database" webpage [24] and was reported by Wyckoff [25]. Fig. 1 displays the optimized geometry of a 72 atom TiO2 anatase nanoparticle prepared from 3×2×1 unit cells. Two suitable oxygen atoms of TiO2 nanoparticle were substituted by nitrogen atoms to construct the N-doped particles. In one doping configuration, a nitrogen atom substitutes an oxygen atom in the middle of the particle and the other configuration characterizes the replacement at twofold coordinated oxygen position. The optimized geometries of the N-doped TiO2 anatase nanoparticles have been shown in Fig. 2.Moreover,the coupled structures of N-doped TiO2 and Au nanoparticle have been presented in Fig. 3.The crystal model of TiO2 anatase nanoparticle contains two types of titanium atoms, referred to as five-fold (5f-Ti) and six-fold (6f-Ti), as well as two types of oxygen atoms, specified by three-fold (3f-O) and two-fold (2f-O) O atoms. Generally, the 2f-O and 5f-Ti atoms are more reactive than the 3f-O and 6f-Ti atoms due to the undercoordination in 2f-O and 5f-Ti atoms. Fig. 4 shows the structure of the considered pristine Au monolayer. A gold structure consisting of 14 Au atoms has been inspected in this study and coupled with TiO2 nanoparticle in order to model a TiO2/Gold nanocomposites. The 86 atom TiO2/Gold nanocomposite placed in a large cubic supercell has been represented in Fig. 5.



Fig. 1. Optimized geometry of a 72 atom TiO2 anatase nanoparticle with bond length results prepared from 3×2×1 unit cells.



Fig. 2. Optimized N-doped TiO2 anatase nanoparticles. The larger gray spheres are Ti atoms and the small red and blue ones represent O and N atoms, respectively.



Fig. 3. Optimized geometries of two types of N-doped TiO2/Gold nanocomposites. The yellow, gray, red and blue balls denote gold, titanium, oxygen and nitrogen atoms, respectively.



Fig. 4. Optimized geometry of an Au monolayer structure with bond length results.



Fig. 5. Optimized geometry of the considered TiO2/Au nanocomposites constructed from anatase nanoparticle and Au monolayer.



Bond lengths, angles and adsorption energies

The interaction of NO2 molecule between TiO2/Gold nanocomposites has been also shown in Fig. 6 as named by types A-C adsorption configurations. Adsorption on the Au nanoparticle is found to be more energetic than the adsorption on the TiO2 nanoparticle. The reason can be easily checked using the calculated adsorption energies and molecular orbital results. At the interface of the adsorption complex, the NO2 molecule preferentially interacts with the Au nanoparticle, being mostly attributed to the higher activity of these nanoparticles in adsorption processes. Unlike Au nanoparticle, TiO2 nanoparticle does not interacts with NO2 molecule due to its lower activity in the composite structures during the adsorption process. In configurations A and B, NO2 molecule interacts with nitrogen doped nanocomposite and the other configuration contains the NO2 molecule adsorbed on the pristine nanocomposite. The adsorption of NO2 molecule on the nanocomposite in configuration C leads to the most stable complex from the energetic viewpoint, compared to the other complexes. Table 1 containsthe lengths for N-O bonds of the adsorbed NO2 molecule and the newly-formed Au-O and O-N-O bond angles before and after the adsorption process. The results of this table indicate that the N-O bonds of the adsorbed NO2 molecule are stretched after the adsorption due to the transference of electronic density from the old Au-Au and N-O bonds to the newly-formed Au-O bonds between nanocomposite and NO2 molecule. The structure of NO2 molecule before the adsorption process has been displayed with bond length and bond angle results in Fig. 7. The O-N-O bond angle of the NO2 molecule after the adsorption on the considered nanocomposites was decreased. This decrease could be mostly ascribed to the formation of new bond and therefore transferring the electronic density from the old bonds of nanocomposite and adsorbed NO2 moleculeto the newly-formed bonds, which makes the weakening of the N-O bonds of NO2 molecule [14, 15]. This formation of new bond also increases the p characteristics of bonding molecular orbitals so that the sp2 hybridization of nitrogen in the NO2 molecule becomes near-sp3. The adsorption energy values for nanocomposites have been reported in table 2. Adsorption on the undoped nanocomposite is found to be energetically more favorable than the adsorption on the N-doped ones, which means that the N doping has a deactivating role on the adsorption of NO2 on Au particles at the interface of TiO2/Gold nanocomposites. In other words, pristine nanocomposite can react with NO2 molecule more efficiently. The adsorption energy of configuration C is about -2.32 eV, suggesting the most energy favorable adsorption configuration in comparison with the other configurations.



 Fig. 6. Top view of the optimized geometry configurations of the interaction of NO2 molecule with TiO2/Goldnanocomposites. The NO2 molecule is preferentially adsorbed on the Au particle surface.



Fig. 7. Structure of NO2 molecule before the adsorption process (Non-adsorbed state).



Table 1. Bond lengths (in Ǻ) for and angles (in degrees) of NO2 molecule adsorbed on TiO2/Gold nanocomposites.



 Table 2. Adsorption energy and Mulliken charge values for NO2 molecule adsorbed on TiO2/Gold nanocomposites.



Electronic structures

Fig. 8 presents the total density of states (TDOS) for pristine and two types N-doped TiO2/Gold nanocomposites before and after the adsorption process. This figure shows that a small peak was created in the DOS of N-doped nanocomposite at the energy values close to -12 eV. The Fermi energy is set to zero denoted by orange solid line in Fig. 8. The TDOS of nanocomposites only exhibit small differences in comparison with the isolated non-adsorbed nanocomposites. These differences are included both shifting of energies of the peaks and appearing some peaks in the DOS of the considered systems. However, these changes in energy of the states would affect the electronic transport properties of the nanocomposites and this feature can be useful for designing and manufacturing NO2 sensors in the adsorption process. The projected partial density of states (PDOS) for the interaction of NO2 molecule with TiO2/Goldnanocomposites have been displayed in Fig. 9 (a-f). Panels (a, b) present the PDOS of the oxygen atom of NO2 molecule and the Au atom for configuration A. The large overlap between the PDOS of these two atoms shows that the oxygen atom of NO2 molecule forms a chemical bond with the Au atom of nanocomposite, suggesting the effective interaction of NO2 molecule from this site. The PDOSs for NO2 interaction with nanocomposite in other two configurations have been also presented as panels (c-d and e-f), which indicates a high overlap between the PDOS of oxygen atom of NO2 molecule and the Au atom of nanocomposite. Fig. 10 (a-f) shows the PDOS of the Au atom before and after the adsorption on the considered nanocomposites, which also suggests a shifting of the PDOS of Au atom to the lower energy values. To fully analyze the charge transfer in the course of adsorption process, the isosurfaces of HOMO and LUMO molecular orbitals for NO2 adsorption on the nanocomposites have been shown in Fig. 11, which suggests that the HOMO is mainly localized on the TiO2 particle, whereas the LUMO is strongly localized on the Au particle. For brevity, the isosurfaces for one configuration is only given.




Fig. 8. DOS for the pristine and two types of N-doped TiO2/Gold nanocomposites. a: A complex; b: B complex and c: C complex. The Fermi energy is set to zero denoted by orange solid line.




Fig. 9. PDOS for the adsorption complexes of TiO2/Gold nanocomposites, a: A complex; b: A complex; c: B complex; d: B complex; e: C complex; f: C complex.



Fig. 10. PDOS for the adsorption complexes of TiO2/Gold nanocomposites in before and after the adsorption process, a: A complex; b: A complex; c: B complex; d: B complex; e: C complex; f: C complex.




Fig. 11. The isosurfaces of HOMO and LUMO molecular orbital of the NO2 molecule adsorbed on the TiO2/Gold nanocomposites


Mulliken Charge Analysis

To further explain the charge transfer between NO2 molecule and TiO2/Gold nanocomposites, we report in table 2 the partial charge values based on Mulliken charge approach. The charge difference for the particle i after and before adsorption, was evaluated using the following formula:

          ∆Qj = Qi (in complex) – Qi (in vacuum)                                                         (2)

where Qi is the value of Mulliken charge of the i. Subscript "i" symbolizes the TiO2/Gold nanocomposites or NO2 molecule. The charge difference, ∆Qi, is a measure of the amount of charge transferred to, or, from the considered nanocomposites from, or, to the NO2 molecule. Our charge analysis indicates that NO2 adsorption induces a comparatively considerable charge transfer (about-0.145 e) from nanocomposite to NO2 for configuration C, suggesting that NO2 behaves as an electron acceptor. The calculated Mulliken charge values for complexes A and B (N-doped systems) are about +0.0038 and +0.118 e respectively, which indicate a lower charge transfer from nanocomposite to the NO2 molecule, compared to the pure (undoped) system. This means that the N doping reduces the efficiency of the charge transfer. Also, it is revealed that the charge transfer causes changes on the conductivity of the system. This could be an operative property to help in the design of novel TiO2-based sensor devices for NO2.



Density functional theory calculations have been carried out to investigate the interaction of NO2 molecules with undoped and N-doped TiO2/Goldnanocomposites to completely comprehend the sensing properties of these nanocomposites in adsorption processes. The results indicate that the adsorption of NO2 molecule causes the stretching of the N-O bonds of the adsorbed NO2 molecule. The bond angle of the NO2 molecule after the adsorption process are lower than those in the gas phase state, being attributed to the transference of the electronic density from the Au bonds of nanocomposite and the N-O bonds of NO2 molecule to the newly-formed bond at the interacting point. The results also suggest that the undoped nanocomposites have a higher capability to interact with toxic NO2 molecule, compared to the N-doped ones. The variations in the electronic structure and adsorption energies are found to be responsible for changing the conductivity of the system. Our reported results thus suggest a theoretical basis for the probable application of TiO2/Gold hybrid nanostructures as gas sensors for main air pollutants such as NO2 in the environment.



This work has been supported by Azarbaijan Shahid Madani University.



The authors declare that there are no conflicts of interest regarding the publication of this manuscript.

1. A. Fujishima and K. Honda, 1972. Electrochemical Photolysis of Water at a Semiconductor Electrode. Nature, 238: 37–38.

2. A. Fujishima., K. Hashimoto and T. Watanabe, 1999. TiO2 Photocatalysis: Fundamentals and Applications, Bkc, Tokyo.

3. R. Erdogan., O. Ozbek and I. Onal, 2010. A periodic DFT study of water and ammonia adsorption on anatase TiO2 (001) slab. Journal of Surface science, 604:1029-1033. 

4. M. Mirzaei, H. Ahadi, M. Shariaty-Niassar and M. Akbari, 2015. Fabrication and characterization of Visible Light active Fe-doped Nanocomposites as Nanocatalyst. Int. J. Nanosci. Nanotechnol, 11(4): 289-293.

5. H. Zarei, M. Zeinali, H. Ghourchian and Kh. Eskandari, 2013. Gold Nano-particles as electrochemical signal amplifier for immune-reaction monitoring. Int. J. Nano. Dimens, 4(1): 69-76.

6. M. Batzilla., E. H. Morales and U. Diebold, 2007. Surface studies of nitrogen implanted TiO2, Journal of Chemical Physics, 339:36–43.

7. M. Bellardita., M. Addamo., A. Di Paola and L. Palmisano, 2007. Photocatalytic behaviour of metal loaded TiO2 aqueous dispersions and films, Chemical Physics.339:94–103.

8. O. Zuas, H. Budiman and N. Hamim, 2013. Anatase TiO2 and mixed M-Anatase TiO2 (M=CeO2 or ZrO2) nano powder: Synthesis and characterization. Int. J. Nano. Dimens, 4(1), 7-12.

9. M. K. Otoufi, N. shahtahmasebebi, A. Kompany and E. Goharshadi, 2014. A systematic growth of gold nanoseeds on Silica for Silica@Gold core-shell nanoparticles and investigation of optical properties. Int. J. Nano. Dimens, 5(6): 525-531.

10. M. Haruta, N. Yamada, T. Kobayashi, S. Iijima, 1989. Gold catalysts prepared by coprecipitation for low-temperature oxidation of hydrogen and of carbon monoxide. Journal of catalysis, 115 (2): 301-309.

11. M. Valden, X. Lai, and D. W. Goodman, 1998, Onset of Catalytic Activity of Gold Clusters on Titania with the appearance of Nonmetallic Properties. Science, 281: 1647-1650.

12. A. Sanchez, S. Abbet, U. Heiz, W. D. Schneider, H. Ha¨kkinen, R. N. Barnett, and U. Landman, 1999, When gold is not noble: nanoscale gold catalysts. The Journal of Physical Chemistry A. 103 (48): 9573-9578.

13. L. M. Molina, M. D. Rasmussen, and B. Hammer. 2004, Adsorption of O2 and oxidation of CO at Au nanoparticles supported by TiO2 (110). J. Chemical Physics. 120(16): 7673.

14. J. Liu, Q. Liu, P. Fang, C. Pan and W. Xiao, 2012. First principles study of the adsorption of a NO molecule on N-doped anatase nanoparticles, J. Applied Surface Science. 258:8312-8318.

15. . J. Liu, L. Dong, W. Guo, T. Liang and W. Lai, 2013. CO adsorption and oxidation on N-doped TiO2 nanoparticles, J. Physical Chemistry C. 117:13037-13044.

16. M. Breedon, M. J. S. Spencer and I. Yarovsky, 2009. Adsorption of NO and NO2 on the ZnO surface: A DFT study. Surface Science, 603: 3389.

17. P. Hohenberg and W. Kohn, 1964. Inhomogeneous Electron Gas, J. Physical Review, 136: B864–B871.

18. W. Kohn and L. Sham, 1965. Self-Consistent Equations Including Exchange and Correlation Effects. J. Physical Review. 140: A1133–A1138.

19. The code, OPENMX, pseudoatomic basis functions, and pseudopotentials are available on a web site ''.

20. . T. Ozaki, 2003. Variationally optimized atomic orbitals for large-scale electronic structures. J. Physical Review. B, 67: 155108.

21. T. Ozaki and H. Kino, 2004. Numerical atomic basis orbitals from H to Kr. J. Physical Review. B, 69: 195113.

22. J. P. Perdew, K. Burke and M. Ernzerhof, 1997. Generalized Gradient Approximation Made Simple. J. Physical Review Letters, 78: 1396.

23. A. Koklj, 2003. Computer graphics and graphical user interfaces as tools in simulations of matter at the atomic scale. J. Comput. Mater. Sci, 28: 155−168.

24. Web page at:

25. R. W. G. Wyckoff, crystal structures, 1972. Second edition. Interscience Publishers, USA, New York.