- Email: [email protected]
Quantum size effect and visible light activity of anatase nanosheet quantum dots
Journal of Photochemistry & Photobiology A: Chemistry 379 (2019) 39–46
Contents lists available at ScienceDirect
Journal of Photochemistry & Photobiology A: Chemistry journal homepage: www.elsevier.com/locate/jphotochem
Quantum size effect and visible light activity of anatase nanosheet quantum dots
T
⁎
Alexander V. Vorontsova, , Héctor Valdésb a b
Altai State University, pr. Lenina 61, Barnaul 656049, Russian Federation Laboratorio de Tecnologías Limpias, Facultad de Ingeniería, Universidad Católica de la Santísima Concepción, Alonso de Ribera 2850, Concepción, Chile
A R T I C LE I N FO
A B S T R A C T
Keywords: Anatase nanosheets scc-dftb pm6 Quantum size effect DFT Photocatalysis
Anatase (001) nanosheets have recently attracted great attention as very active catalysts and photocatalysts. These graphene analogs have very high surface area and unique surface properties. In the present paper, very thin two-layer anatase nanosheets are investigated computationally in the form of quantum dots of various size. Quantum size effect (QSE) was clearly observed for nanosheets with fully hydroxylated edges and size up to 14 nm and the ultimate band gap is around 3.4 eV. Dehydroxylation of nanosheets obscured QSE, decreased band gap and induced visible light absorption. Therefore, contradictory trends reported in experimental studies for anatase QSE can be ascribed to different degree of hydroxylation of the TiO2 samples surface. All anatase nanosheet quantum dots retained their flat graphene-like shape. These findings demonstrate that dehydroxylated anatase nanosheet quantum dots are prospective visible-light active photocatalysts even if their inherent band gap is considerably larger than for bulk anatase.
1. Introduction Titanium dioxide remains one of the most studied and utilized materials for photocatalysis, solar cells, batteries, sensors [1], catalysis [2], biomedical and many other applications [3]. Recent advancements in TiO2 preparation methods enabled studies and applications of TiO2 nanoparticles with many different shapes and surfaces [4]. In our previous publications, we studied properties of anatase nanoparticles of various shapes, sizes and different surface functional groups [5–7]. Such developed computational models of TiO2 nanoparticles were subsequently applied for providing understanding of advanced photocatalytic activities of composites [8–10]. However, many studies demonstrated increased functional properties including much elevated rate of photocatalytic reactions over TiO2 in the form of (001) nanosheets as compared to roundish particles and other nanosheets [11,12]. Therefore, studies on such nanosheets are very promising and, no doubt, will be continued for full discovery of the material useful properties [13]. Among the properties of TiO2, its band gap and bands positions are very important from fundamental point of view and for photocatalytic and photoelectrochemical applications. Quantum size effect (QSE) in titanium dioxide is one of the obscure questions in the studies of this material. Modern literature contains contradictory experimental data on the effect of TiO2 particles size on their electronic and optical ⁎
properties. Luca et al. reports optical band gap change from 3.42 to 3.33 eV for change in anatase nanoparticles size from 3.8–11 nm, while commercial 60 nm anatase had band gap of 3.44 eV [14]. In a later paper [15], they reported irregular dependence of the anatase band gap on particles size with a minimum for 3 nm particles. Monticone et al. reported band gap decrease from about 3.4 to 3.2 eV when the size increased from 1 to 3 nm and concluded that the effective mass decreases when the size increases [16]. Satoh et al. reported that three very small anatase nanoparticles containing 6, 14 and 30 TiO2 units obey the effective mass approximation (EMA) with bulk band gap 3.2 eV and dielectric constant 31 [17]. The previously mentioned studies reported deviation from the EMA. To shed some light on the QSE, several theoretical computational studies have been undertaken. Cho et al. investigated nanoparticles in the range of 0.5–3.2 nm with (CAM-)B3LYP hybrid functional and predicted that the bulk-like behavior can emerge for the size above 6.5 nm [18]. In correction to the original paper, they estimated the bulklike size as 20 nm [19]. QSE for anatase (001) bilayer nanoribbons was studied by Vorontsov and Smirniotis [20]. The band gap (Eg) was essentially stable and equal to approximately 3.5 eV for nanoribbons width above 1.3 nm. Local structure deformations and conformation of surface hydroxyl groups exerted strong influence on the Eg and increased or decreased it by up to 0.7 eV. For decahedral anatase nanoparticles, QSE had more complex behavior: surface groups and
Corresponding author. E-mail addresses: [email protected], [email protected] (A.V. Vorontsov).
https://doi.org/10.1016/j.jphotochem.2019.05.001 Received 12 March 2019; Received in revised form 20 April 2019; Accepted 1 May 2019 Available online 01 May 2019 1010-6030/ © 2019 Elsevier B.V. All rights reserved.
Journal of Photochemistry & Photobiology A: Chemistry 379 (2019) 39–46
A.V. Vorontsov and H. Valdés
method also called SCC-DFTB [39]. DFTB + program [40] was used in combination with Slater-Koster (SK) parameter sets tiorg-0-1 [41] and mio-0-1 [42]. Self-consistent charge was calculated with the tolerance 1·10−7 Ha recommended for good precision calculations. The geometry of nanosheets was optimized till maximal component of forces reached below 4.2·10-5 Ha/Bohr that is equivalent to 0.05 kcal/(mol·Å). All atoms of the nanosheets were allowed to move during geometry optimizations. Additional computations of energy and structure optimization were performed with PM6 semiempirical methodology [43] and, for one case, with pm7 [44]. MOPAC2012 and MOPAC2016 programs were employed for the computations with SCF convergence criterion 1·10−6 kcal/mol and maximum force criterion 0.05 kcal/(mol·Å). Electronic absorption spectra were computed using time dependent dftb method in linear approximation [39]. The number of excitations considered was 4000 and this limited the range of the spectrum in the UV region.
coordinatively unsaturated surface Ti atoms with charge < 4+ influenced the Eg [6]. According to the computational results, bulklike properties are expected for decahedral nanoparticles with the size above 2 nm. Due to the predominance of visible light over UV in the solar spectrum, modern research aims at developing photosensitive materials active at λ > 400 nm. Visible light absorption in TiO2 materials including nanosheets can be caused by the presence of point defects such as Ti3+ and oxygen vacancies as well as by some admixtures [21] and heterojunctions [22]. These defects can appear even if oxygen rich air atmosphere is used for calcination of TiO2 as changes in the visible part of diffuse reflection spectrum showed [23]. Traditionally, introduction of dopant atoms is required to obtain visible light activity of TiO2 including nanosheets [24,25]. However, recent studies showed that intrinsic defects can also cause absorption and reactions under visible light [26,27] as well as advanced sensor properties [28]. Even perfect anatase nanocrystals of various shapes possess surface point defects that can cause visible light absorption and photoactivity [6,29]. Despite the very large interest in 2D TiO2 and its numerous applications in photocatalysis and other fields, theoretical consideration is limited to just few publications [30]. Mogilevsky et al. studied dissociative adsorption of water on delaminated bilayer anatase nanosheets and found small distortion of the nanosheet surface [31]. Vittadini and Casarin studied anatase nanosheets containing from 1 to 6 atom layers using periodic models [32]. They found that 1–3 layer anatase (001) nanosheets undergo reorganization into lepidocrocite and another phase. (101) nanosheets also underwent strong distortions. Among the titanate bilayer structures modeled, H2Ti3O7 was found as the most stable [33]. Liao et al. studied monolayer TiO2 nanosheets of surfaces (010) and (101) and found good charge separation with the attached Cd halcogenide quantum dots [34]. Liu et al. investigated codoped TiO2 nanosheets and found their visible light absorption [35]. Wang et al. studied monolayers of various titanate crystal structures using PBE and HSE06 functionals and obtained their band positions [36]. Oxygen vacancy in lepidocrocite was computationally studied in [37]. In the present work, quantum size effect and optical properties are studied for anatase two-layer nanosheets of approximately square shape. The influence of the rim hydroxyl groups is studied as well. Due to the extremely large specific surface area of anatase nanosheets, this material has very important application area in decontamination of chemical warfare agents and equally toxic chemicals in the dark and under solar irradiation [38]. Enhanced visible light absorption for certain configurations of dehydroxylated nanosheets observed in the present study provides opportunities of desining novel advanced photocatalytic materials.
3. Results and discussion Anatase TiO2 remains the most often used photocatalyst for diverse redox reactions. Its high activity can be further improved via formation of nanosheets with predominant surface (001). The ultimately thinnest anatase nanosheets contain just two layers of atoms since monolayer materials rearrange into lepidocrocite. The present investigation considers such ultrathin nanosheets. Preparation methods usually allow obtaining large-scale nanosheets. Such large nanosheets tend to fold and this decreases the amount of available surface area. Thus, an idea to overcome this disadvantage is to limit the horizontal sizes and obtain nanosheet quantum dots instead of semi-infinite sheets. Such quantum dots can be organized into mesoporous nanostructures for gas phase photocatalytic applications or remain as is for liquid phase suspension reactions. Templates can be selected for obtaining quantum dots with needed size. Previously we studied ultrathin anatase titanium dioxide nanoribbons of various width and found their stability and limited quantum size effect [20]. In the present work, ultrathin nanosheets are of approximately square shape. This was hoped to facilitate observation of QSE and provide more generality for the study on visible-light activity. Nanosheet quantum dots of the present study have fully hydroxylated edges or edges with variable extent of dehydroxylation. Hence, the surface density of OH groups varies in the nanosheet quantum dots wildly between 52 OH/nm2 for the smallest nanoparticle with full edges hydroxylation to 0.6 OH/nm2 for the largest dehydroxylated nanosheet. The large variation seems to be in agreement with literature data for roundish anatase nanoparticles [45]. Fig. 1 shows nanosheet quantum dots with fully hydroxylated edges after they were optimized with the scc-dftb method. This method is known to provide very good agreement of anatase lattice structure with the experimental data. Table S1 summarizes all data on nanosheet quantum dots optimized with scc-dftb. Measurements of the nanosheets size reveal that the originally ideal square shape is distorted with contraction along non-hydroxylated edge direction. The hydroxylated edge direction remained almost intact in length. This phenomenon of contraction along one direction of (001) surface is well known in the literature and was observed for decahedral anatase nanoparticles as well [6]. However, this deviation of the crystal structure in (001) surface does not cause any significant aberrations in the nanosheets. The nanosheets quantum dots remain almost ideally flat up to the largest sizes studied here of 4.6 x 3.8 nm. Previously we found that the flatness depends on the number of layers of atoms in the nanosheets: only nanosheets with even number of atomic layers are flat [20]. This necessitates very careful choice of preparation methods to selectively produce even number of layers nanosheets. When the size of the nanosheets increases from TNS11 to TNS22 (Fig. 1), drastic changes in the atomic structure proceed. The structure of titanium atoms surroundings changes from tetrahedra TiO4 in
2. Models and methods Anatase TiO2 nanosheets quantum dots (TNSQD) were prepared from the bulk TiO2 lattice by removing all atoms except two layers of atoms closest to the (001) facet surface. The series of two layer TNSQD studied in the present work are square flakes with the number of unit cells from 1 × 1–12 × 12. The nanosheet quantum dots are designated as TNSKKLL-MvN, where KK is the length of the hydroxylated side in unit cells, LL is the length of the non-hydroxylated edges in unit cells, M is the number of the nanosheet related to the extent of hydroxylation of edges. Additionally, designation v2 was added that means that all hydroxyl groups along each of the edge are aligned in the same direction, and alignment direction is anti-parallel in the neighbor edges of nanosheet. Coordinatively unsaturated Ti atoms with charge less than 4+ in the edges of the nanosheets were saturated with attached OH groups, or oxygen atoms. In the later case, two Ti atoms contain one O atom, which causes non-uniformity of charge among rim Ti atoms. Energy has been obtained in the course of computations with DFTB2 40
Journal of Photochemistry & Photobiology A: Chemistry 379 (2019) 39–46
A.V. Vorontsov and H. Valdés
Fig. 1. Anatase nanosheet quantum dots of approximately square shape with fully hydroxylated edges optimized with scc-dftb.
TNS11-3 and TNS11-3v2 to distorted square pyramids TiO5 for TNS22 and larger nanosheets. The coordination of Ti with five oxygen atoms is typical for (001) surfaces of anatase. The typical TieO bond length for atoms inside the nanosheets is 1.81–1.99 Å and these bond length is also typical for the edge atoms. The OeTieO angle is 116 – 123° for the internal Ti atoms and 126 – 129° for the corresponding OeTi−OH valent angles measured by taking axial O-Ti bonds. Table S1 lists positions of HOMO (valence band edge), LUMO (conduction band edge) and Eg for the nanosheet quantum dots. Recent experimental measurements of anatase band edges estimated ECB = -3.95 eV, EVB = -7.1 eV and Eg = 3.15 eV [46]. Results of scc-dftb method computations gave bands edges shifted by about + 1.5 eV from these experimental values. The exact band edges positions depend slightly on the size of nanoparticles and strongly on the degree of edges hydroxylation: hydroxylation results in bands edge shift to positive potentials by up to 0.6 eV. Fig. 2 illustrates QSE in the nanosheet quantum dots with fully hydroxylated edges. The value of the Eg decreases upon increase in the diameter of the TiO2 quantum dot in agreement with the theory of QSE. Eg decreases clearly upon increase in size from 0.5 to 2.0 nm. TNSQD of larger diameter (diagonal) show scattering of band gap values due to some small changes in the structure including different conformations of edge hydroxyl groups. Consequently, the tendency is not so clear for the larger TNSQD. Least-square fitting with double-exponential decay shown in Fig. 2 points to a fast decrease of band gap for sizes up to 1 nm
Fig. 2. Quantum size effect for anatase nanosheet quantum dots with fully hydroxylated edges according to scc-dftb method results.
41
Journal of Photochemistry & Photobiology A: Chemistry 379 (2019) 39–46
A.V. Vorontsov and H. Valdés
Fig. 3. Anatase nanosheet quantum dots of approximately square shape with dehydroxylated edges optimized with scc-dftb.
and gradual slow decrease for further increase in size. The final value of Eg = 3.4 eV is predicted to be attained at a diameter of approximately 14 nm. This predicted range of QSE for TNSQD is in line with predictions by Cho et al. for anatase nanoparticles that their ultimate band gap is attained at size well above suggested exciton size for TiO2 of 2 nm [18,19]. Removal of chemisorbed water from the surface of TiO2 can be carried out via moderate temperature treatment or chemical reagents application. Fig. 3 shows structure of the TNS with partially dehydroxylated edges. Properties of the optimized TNS are given in Table S1. Dehydroxylation resulted in a considerable decrease of the length of hydroxylated edges while the length of non-hydroxylated edges changed little. The nanosheets retained their flatness even when almost all OH groups were removed and the edges were saturated with oxygen atoms. The structure of the TNS edges became distorted for a larger extent when more hydroxyl groups were removed. Electronic band gap for all dehydroxylated TNSQD of different diameter is shown in Fig. 4. The band gap changes rather irregularly when the size of TiO2 nanosheet increases. This can at least partially explain why some experimental studies report contradictory dependences of optical properties of TiO2 samples on their size. The
Fig. 4. Electronic band gap plotted as a function of size for dehydroxylated anatase quantum dots according to scc-dftb.
42
Journal of Photochemistry & Photobiology A: Chemistry 379 (2019) 39–46
A.V. Vorontsov and H. Valdés
Fig. 5. Anatase nanosheet quantum dots with hydroxylated edges optimized with pm6.
general tendency is however retained: band gap decreases when going from small nanoparticles with the diameter below 3 nm to larger nanoparticles. The band gap is below 3 eV for six kinds of quantum dots having different degree of dehydroxylation. Visible light activity appear first for TNS44-2 and is present for the nanosheets of size above 3 nm. There are two causes here for the band gap below 3 eV – appearance of edge states and deformation of edges – that result from dehydroxylation. Table S1 shows that besides decrease in Eg, dehydroxylation shifts HOMO and LUMO energies towards more negative values. Introduction of visible light activity in small nanoparticles of anatase nanosheets via dehydroxylation of their edges opens a new door to preparation of TiO2-based visible light photocatalysts. All the computations results presented above were obtained with scc-dftb method that was derived from DFT. The method parameters were specially developed for modeling of TiO2 phases. Independent prove of this method results can be done using pm6 method that was derived independently from completely different theory of HartreeFock. Fig. 5 shows atomic structure of nanosheet quantum dots with completely hydroxylated edges. We can see that all nanosheets have flat graphene-like shape. Similarly to the results of scc-dftb computations, the smallest nanoparticle TNS11-3 has strongly distorted structure. Nanoparticle TNS33-3 was also optimized with more recent pm7 method. It is again confirmed that pm7 method is unsuitable for modeling of TiO2 nanoparticles since structure of the TNS33-3 nanosheet optimized with this method is completely different from results of scc-dftb and pm6 method as well as from experimental anatase lattice
Fig. 6. Quantum size effect in anatase bilayer quantum dots with fully hydroxylated edges according to results of pm6 method.
structure. Fig. 6 shows band gap of hydroxylated nanosheets computed with pm6 method and scaled band gap using the factor of 0.37. This scaling factor was selected by fitting of the resultant band gap values to those 43
Journal of Photochemistry & Photobiology A: Chemistry 379 (2019) 39–46
A.V. Vorontsov and H. Valdés
Fig. 7. Anatase bilayer nanosheets with dehydroxylated edges optimized with pm6.
obtained with the scc-dftb method. Unscaled values of the band gap are too large and vary between 9 and 10 eV. Scaling results in more realistic values of 3.5–3.8 eV. Thus, scaled values of Eg computed with pm6 method could probably be used for the rough estimation of band gap for real nanosheet materials. For different shapes of TiO2 nanoparticles, the scaling factor for the band gap obtained with pm6 method is different. For example, for decahedral anatase nanoparticles Ti22r1 – Ti55r4, the scaling factor is 0.31. However, QSE produced with pm6 is completely unrealistic and shows an increase of band gap with an increase in anatase nanosheet quantum dots size. Therefore, such features as dependence of electronic properties on size of nanoparticles cannot be expected to be correctly described and predicted with pm6 method. Fig. 7 demonstrates chemical structure of TNSQD with dehydroxylated edges obtained using pm6 method. The chemical structure features such as relatively strong deformations in the edges are similar to those from the scc-dftb method. The nanosheets remain of almost ideally flat shape. The structural results produced by pm6 method for anatase nanosheets agree qualitatively with the results obtained with the scc-dftb method. Therefore, pm6 method, which is computationally more robust than scc-dftb in presently available implementations, can be used for the studies on TiO2 nanoparticles structures for diverse sets of shapes provided suitable corrections are applied for the obtained unit cell sizes to bring them in general agreement with the experimental anatase lattice parameters. The values of the band gap for the dehydroxylated TNSQD computed with pm6 are given in Fig. 8. Since the predicted values of the band gap overestimate strongly experimentally available values of TiO2 nanoparticles band gap, a correction should be applied. The best value of the correction factor seems to be 0.37 and this scaling factor was used in Fig. 8 for plotting the scaled band gap. Obtained values of the
Fig. 8. Electronic band gap in dehydroxylated square bilayer anatase nanosheets according to pm6 computations.
corrected band gap are around 3.1 eV for the nanosheets having UV absorption while values around 2.7–2.9 eV are obtained for the nanosheets with visible light absorption. Giving the incorrect predictions of the QSE of the pm6 semiempirical method, it was surprising to find out that the scaled band gap values correctly point to TNSQD having visible light absorption in accord to the scc-dftb method results. The exact values of the band gap deviate from the values obtained with the 44
Journal of Photochemistry & Photobiology A: Chemistry 379 (2019) 39–46
A.V. Vorontsov and H. Valdés
Fig. 9, all absorption curves can be linearly interpolated to the absorption threshold of around 340 nm. This threshold corresponds to the value of the optical band gap of 3.65 eV which agrees very well with the values of electronic band gap of around 3.6 eV as shown in Table S1. Therefore, in the case of TNSQD with UV light activity in the present study, there is a good agreement between the values of electronic and optical band gap. The blue shift of absorption threshold for TiO2 nanosheets seen in Fig. 9 agrees with available experimental data for very thin TiO2 nanosheets [22]. However, preparation of pure bilayer nanosheets still remains a challenge for experimental methods and exact comparison with experimental data is not available yet. Absorption spectra for nanosheets having electronic band gap corresponding to visible light absorption are shown in Fig. 10. As it was discussed above, these TNSQD have dehydroxylated edges and their prominent visible light absorption is related to appearance of admixture states localized predominantly in their edges. It can be generally seen in Fig. 10 that visible light absorption for λ > 400 nm is not so strong for the studied TNSQD. There are irregularities in the UV–vis spectra for these materials that are related to the strong contribution of admixture energy levels, which contrasts with the spectra for the nanosheets having intrinsic absorption only in Fig. 9. Absorption edges for these nanosheets deviate from the wavelengths corresponding to the electronic band gap values. The largest wavelength of absorption is obtained for TNS77-2v2 and TNS99-1 that have small absorption peaks approximately at 520 nm. Interestingly, this agrees with the electronic band gap of the TNS99-1 (525 nm). Thus, computed electronic spectra confirm visible light activity of anatase nanosheets quantum dots with dehydroxylated edges. These materials are prospective extremely high surface area adsorbents and visible light active photocatalysts.
Fig. 9. UV absorption spectra of square bilayer anatase nanosheets having no bands in the visible region as resulted from td-dftb calculations. The short wavelength boundary is limited by the number of excitations produced in calculation.
4. Conclusion The present computational study on approximately square bilayer anatase nanosheets quantum dots with the diameter up to 6 nm has led to the following conclusions. 1. Despite their extremely small thickness below 0.5 nm, bilayer anatase nanosheets with hydroxylated and dehydroxylated edges retain anatase-like chemical structure and are essentially flat. 2. Anatase nanosheets with fully hydroxylated edges demonstrate a clear quantum size effect with electronic band gap fitting relatively well to bi-exponential decay. Ultimate band gap of 3.4 eV is attained for diameter approximately 14 nm. Electronic band gap values for these nanosheets agree well with optical band gap values. 3. Some anatase TNSQD with different extent of dehydroxylation of edges possess electronic band gap below 3 eV. Computed electronic absorption spectra confirm that these nanosheets have significant absorption in the visible light with wavelength up to 530 nm. 4. According to the results of the present study, strong difference between published experimentally obtained band gap values for anatase nanoparticles of variable size can be ascribed, at least partially, to different extent of surface hydroxylation of the samples used. Computational results of the investigation suggests that novel visible-light active TiO2 – based photocatalysts with extremely high surface area can be obtained in the form of several nanometers sized nanosheets with dehydroxylated edges and even number of atomic layers.
Fig. 10. Electronic spectra of approximately square partially dehydroxylated anatase nanosheets showing absorption in the visible light region according to td-dftb computations.
scc-dftb method considerably, which is an expected result, when we take into consideration generally poor ability of pm6 method to predict electronic properties qualitatively. In the above consideration, electronic band gap values were considered for diverse hydroxylated and dehydroxylated anatase bilayer nanosheet quantum dots. The electronic band gap is calculated from the values of valence band (HOMO) and conduction band (LUMO) edges. Experimentally, however, band gap values are often obtained using UV–vis diffuse reflection spectra. The obtained thereby band gap values are related to light absorption and can be quite different from the electronic band gap because optical absorption have strong dependence on the density of states of materials. Therefore, it was important to consider optical band gaps of the TNSQD in the present study. Fig. 9 shows absorbance for the TNSQD that are not visible light active. The smallest nanosheets are not considered since they have deformed structure and would hardly be obtained experimentally and they have strongly different values of the band gap. As we can see in
Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.jphotochem.2019.05. 001. References [1] J. Zhang, J. Liao, F. Yang, M. Xu, S. Lin, Regulation of the electroanalytical
45
Journal of Photochemistry & Photobiology A: Chemistry 379 (2019) 39–46
A.V. Vorontsov and H. Valdés
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24] D. Lu, P. Fang, Y. Liu, Z. Liu, X. Liu, Y. Gao, F. Chen, F. Niu, A facile one-pot synthesis of gadolinium doped TiO2-based nanosheets with efficient visible lightdriven photocatalytic performance, J. Nanoparticle Res. 16 (2014), https://doi.org/ 10.1007/s11051-014-2636-3. [25] L. Jiang, L. Sun, D. Yang, J. Zhang, Y.-J. Li, K. Zou, W.-Q. Deng, Niobium-doped (001)-Dominated anatase TiO 2 nanosheets as photoelectrode for efficient dyesensitized solar cells, ACS Appl. Mater. Interfaces 9 (2017) 9576–9583, https://doi. org/10.1021/acsami.6b14147. [26] J.-Y. Zheng, S.-H. Bao, Y. Guo, P. Jin, Anatase TiO 2 films with dominant {001} facets fabricated by direct-current reactive magnetron sputtering at room temperature: oxygen defects and enhanced visible-light photocatalytic behaviors, ACS Appl. Mater. Interfaces 6 (2014) 5940–5946, https://doi.org/10.1021/am500979j. [27] H. Gao, G. Hu, J. Sui, C. Mu, W. Shangguan, M. Kong, W. Shentu, Scalable preparation of defect-rich free-standing TiO 2 sheets with visible-light photocatalytic activity, Appl. Catal. B 226 (2018) 337–345, https://doi.org/10.1016/j.apcatb. 2017.12.055. [28] Y. Wang, J. Liu, M. Wang, C. Pei, B. Liu, Y. Yuan, S. Liu, H. Yang, Enhancing the sensing properties of TiO 2 nanosheets with exposed {001} facets by a hydrogenation and sensing mechanism, Inorg. Chem. 56 (2017) 1504–1510, https://doi. org/10.1021/acs.inorgchem.6b02603. [29] A.V. Vorontsov, Cluster models of photocatalytic anatase TiO2 nanoparticles and their computational characterization, Catal. Today 252 (2015) 168–176, https:// doi.org/10.1016/j.cattod.2014.10.017. [30] F. De Angelis, C. Di Valentin, S. Fantacci, A. Vittadini, A. Selloni, Theoretical studies on anatase and less common TiO 2 phases: bulk, surfaces, and nanomaterials, Chem. Rev. 114 (2014) 9708–9753, https://doi.org/10.1021/cr500055q. [31] G. Mogilevsky, Q. Chen, H. Kulkarni, A. Kleinhammes, W.M. Mullins, Y. Wu, Layered nanostructures of delaminated anatase: nanosheets and nanotubes, J. Phys. Chem. C 112 (2008) 3239–3246, https://doi.org/10.1021/jp076899p. [32] A. Vittadini, M. Casarin, Ab initio modeling of TiO2 nanosheets, Theor. Chem. Acc. 120 (2008) 551–556, https://doi.org/10.1007/s00214-008-0425-8. [33] M. Casarin, A. Vittadini, A. Selloni, First principles study of Hydrated/Hydroxylated TiO 2 nanolayers: from isolated sheets to stacks and tubes, ACS Nano 3 (2009) 317–324, https://doi.org/10.1021/nn800608n. [34] T. Liao, Z. Sun, S.X. Dou, Theoretically manipulating quantum dots on two-dimensional TiO 2 monolayer for effective visible light absorption, ACS Appl. Mater. Interfaces 9 (2017) 8255–8262, https://doi.org/10.1021/acsami.6b15741. [35] Y. Liu, W. Zhou, P. Wu, The electronic structure and optical properties of donoracceptor codoped TiO2 nanosheets from hybrid functional calculations, Mater. Chem. Phys. 186 (2017) 333–340, https://doi.org/10.1016/j.matchemphys.2016. 11.004. [36] L. Wang, D. Wei, S. Kang, X. Xie, Y. Shi, S. Liu, Two-dimensional titania: structures and properties predicted by first principle calculation, J. Phys. Chem. C 122 (2018) 22911–22919, https://doi.org/10.1021/acs.jpcc.8b05412. [37] L. Wang, G. Zhou, Y. Shan, Z. Huang, L. Liu, Oxygen-defect-dependent ferromagnetism and strain modulation in free-standing two-dimensional TiO 2 monolayers, J. Chem. Soc. Faraday Trans. 20 (2018) 27176–27184, https://doi.org/10. 1039/C8CP05223A. [38] A. Sengele, D. Robert, N. Keller, C. Colbeau-Justin, V. Keller, Sn-doped and porogen-modified TiO2 photocatalyst for solar light elimination of sulfure diethyle as a model for chemical warfare agent, Appl. Catal. B 245 (2019) 279–289, https:// doi.org/10.1016/j.apcatb.2018.12.071. [39] T.A. Niehaus, S. Suhai, F. Della Sala, P. Lugli, M. Elstner, G. Seifert, Th. Frauenheim, Tight-binding approach to time-dependent density-functional response theory, Phys. Rev. B 63 (2001), https://doi.org/10.1103/PhysRevB.63.085108. [40] B. Aradi, B. Hourahine, Th. Frauenheim, DFTB+, a Sparse Matrix-Based Implementation of the DFTB Method†, J. Phys. Chem. A 111 (2007) 5678–5684, https://doi.org/10.1021/jp070186p. [41] G. Dolgonos, B. Aradi, N.H. Moreira, T. Frauenheim, An improved self-consistentCharge density-functional tight-binding (SCC-DFTB) set of parameters for simulation of bulk and molecular systems involving titanium, J. Chem. Theory Comput. 6 (2010) 266–278, https://doi.org/10.1021/ct900422c. [42] M. Elstner, D. Porezag, G. Jungnickel, J. Elsner, M. Haugk, Th. Frauenheim, S. Suhai, G. Seifert, Self-consistent-charge density-functional tight-binding method for simulations of complex materials properties, Phys. Rev. B 58 (1998) 7260–7268, https://doi.org/10.1103/PhysRevB.58.7260. [43] J.J.P. Stewart, Optimization of parameters for semiempirical methods V: modification of NDDO approximations and application to 70 elements, J. Mol. Model. 13 (2007) 1173–1213, https://doi.org/10.1007/s00894-007-0233-4. [44] J.J.P. Stewart, Optimization of parameters for semiempirical methods VI: more modifications to the NDDO approximations and re-optimization of parameters, J. Mol. Model. 19 (2013) 1–32, https://doi.org/10.1007/s00894-012-1667-x. [45] R. Mueller, H.K. Kammler, K. Wegner, S.E. Pratsinis, OH surface density of SiO 2 and TiO 2 by thermogravimetric analysis, Langmuir. 19 (2003) 160–165, https://doi. org/10.1021/la025785w. [46] C. Maheu, L. Cardenas, E. Puzenat, P. Afanasiev, C. Geantet, UPS and UV spectroscopies combined to position the energy levels of TiO 2 anatase and rutile nanopowders, J. Chem. Soc. Faraday Trans. 20 (2018) 25629–25637, https://doi.org/ 10.1039/C8CP04614J.
performance of ultrathin titanium dioxide nanosheets toward lead ions by nonmetal doping, Nanomaterials. 7 (2017) 327, https://doi.org/10.3390/ nano7100327. H. Wang, Y. Song, J. Xiong, J. Bi, L. Li, Y. Yu, S. Liang, L. Wu, Highly selective oxidation of furfuryl alcohol over monolayer titanate nanosheet under visible light irradiation, Appl. Catal. B 224 (2018) 394–403, https://doi.org/10.1016/j.apcatb. 2017.10.069. F. Wu, Z. Wang, X. Li, H. Guo, Simple preparation of petal-like TiO2 nanosheets as anode materials for lithium-ion batteries, Ceram. Int. 40 (2014) 16805–16810, https://doi.org/10.1016/j.ceramint.2014.07.060. G. Liu, H.G. Yang, J. Pan, Y.Q. Yang, G.Q.(Max) Lu, H.-M. Cheng, Titanium dioxide crystals with tailored facets, Chem. Rev. 114 (2014) 9559–9612, https://doi.org/ 10.1021/cr400621z. A.V. Vorontsov, S.V. Tsybulya, Influence of nanoparticles size on XRD patterns for small monodisperse nanoparticles of Cu ° and TiO 2 anatase, Ind. Eng. Chem. Res. 57 (2018) 2526–2536, https://doi.org/10.1021/acs.iecr.7b04480. A.V. Vorontsov, P.G. Smirniotis, Size and surface groups effects in decahedral anatase nanoparticles for photocatalytic applications, J. Photochem. Photobiol. A: Chem. 363 (2018) 51–60, https://doi.org/10.1016/j.jphotochem.2018.05.031. A.V. Vorontsov, Structural and electronic effects in acetone adsorption over TiO2 anatase clusters as the first stage of photocatalytic oxidation, J. Nanoparticle Res. 19 (2017) 326, https://doi.org/10.1007/s11051-017-4025-1. J.O. Olowoyo, M. Kumar, B. Singh, V.O. Oninla, J.O. Babalola, H. Valdés, A.V. Vorontsov, U. Kumar, Self-assembled reduced graphene oxide-TiO2 nanocomposites: synthesis, DFTB+ calculations, and enhanced photocatalytic reduction of CO2 to methanol, Carbon. 147 (2019) 385–397, https://doi.org/10.1016/j. carbon.2019.03.019. J.O. Olowoyo, M. Kumar, S.L. Jain, J.O. Babalola, A.V. Vorontsov, U. Kumar, Insights into reinforced photocatalytic activity of the CNT–TiO 2 nanocomposite for CO 2 reduction and water splitting, J. Phys. Chem. C 123 (2019) 367–378, https:// doi.org/10.1021/acs.jpcc.8b07894. J.O. Olowoyo, M. Kumar, N. Singhal, S.L. Jain, J.O. Babalola, A.V. Vorontsov, U. Kumar, Engineering and modeling the effect of Mg doping in TiO 2 for enhanced photocatalytic reduction of CO 2 to fuels, Catal. Sci. Technol. 8 (2018) 3686–3694, https://doi.org/10.1039/C8CY00987B. C. Hu, X. Zhang, W. Li, Y. Yan, G. Xi, H. Yang, J. Li, H. Bai, Large-scale, ultrathin and (001) facet exposed TiO2 nanosheet superstructures and their applications in photocatalysis, J. Mater. Chem. A 2 (2014) 2040, https://doi.org/10.1039/ c3ta14343k. L. Wang, T. Sasaki, Titanium oxide nanosheets: graphene analogues with versatile functionalities, Chem. Rev. 114 (2014) 9455–9486, https://doi.org/10.1021/ cr400627u. C.P. Sajan, S. Wageh, Ahmed.A. Al-Ghamdi, J. Yu, S. Cao, TiO2 nanosheets with exposed {001} facets for photocatalytic applications, Nano Res. 9 (2016) 3–27, https://doi.org/10.1007/s12274-015-0919-3. V. Luca, S. Djajanti, R.F. Howe, Structural and electronic properties of sol−Gel titanium oxides studied by X-ray absorption spectroscopy, J. Phys. Chem. B 102 (1998) 10650–10657, https://doi.org/10.1021/jp981644k. V. Luca, Comparison of size-dependent structural and electronic properties of anatase and rutile nanoparticles, J. Phys. Chem. C 113 (2009) 6367–6380, https:// doi.org/10.1021/jp808358v. S. Monticone, R. Tufeu, A.V. Kanaev, E. Scolan, C. Sanchez, Quantum size effect in TiO2 nanoparticles: does it exist? Appl. Surf. Sci. 162–163 (2000) 565–570, https:// doi.org/10.1016/S0169-4332(00)00251-8. N. Satoh, T. Nakashima, K. Kamikura, K. Yamamoto, Quantum size effect in TiO2 nanoparticles prepared by finely controlled metal assembly on dendrimer templates, Nat. Nanotechnol. 3 (2008) 106–111, https://doi.org/10.1038/nnano. 2008.2. D. Cho, K.C. Ko, O. Lamiel-García, S.T. Bromley, J.Y. Lee, F. Illas, Effect of size and structure on the ground-state and excited-state electronic structure of TiO 2 nanoparticles, J. Chem. Theory Comput. 12 (2016) 3751–3763, https://doi.org/10. 1021/acs.jctc.6b00519. D. Cho, K.C. Ko, O. Lamiel-García, S.T. Bromley, J.Y. Lee, F. Illas, Correction to effect of size and structure on the ground-state and excited-state electronic structure of TiO 2 nanoparticles, J. Chem. Theory Comput. 13 (2017), https://doi.org/10. 1021/acs.jctc.7b01087 5822–5822. A.V. Vorontsov, P.G. Smirniotis, Structure, electronic and optical properties of bilayer anatase nanoribbons, Comput. Mater. Sci. 155 (2018) 266–281, https://doi. org/10.1016/j.commatsci.2018.08.052. J. Cai, Z. Huang, K. Lv, J. Sun, K. Deng, Ti powder-assisted synthesis of Ti3+ selfdoped TiO2 nanosheets with enhanced visible-light photoactivity, RSC Adv. 4 (2014) 19588, https://doi.org/10.1039/c4ra01496k. Y. Xin, Y. Lu, C. Han, L. Ge, P. Qiu, Y. Li, S. Fang, Novel NiS cocatalyst decorating ultrathin 2D TiO2 nanosheets with enhanced photocatalytic hydrogen evolution activity, Mater. Res. Bull. 87 (2017) 123–129, https://doi.org/10.1016/j. materresbull.2016.11.027. A.V. Vorontsov, A.A. Altynnikov, E.N. Savinov, E.N. Kurkin, Correlation of TiO2 photocatalytic activity and diffuse reflectance spectra, J. Photochem. Photobiol. A: Chem. 144 (2001) 193–196, https://doi.org/10.1016/S1010-6030(01)00462-2.
46
Contents lists available at ScienceDirect
Journal of Photochemistry & Photobiology A: Chemistry journal homepage: www.elsevier.com/locate/jphotochem
Quantum size effect and visible light activity of anatase nanosheet quantum dots
T
⁎
Alexander V. Vorontsova, , Héctor Valdésb a b
Altai State University, pr. Lenina 61, Barnaul 656049, Russian Federation Laboratorio de Tecnologías Limpias, Facultad de Ingeniería, Universidad Católica de la Santísima Concepción, Alonso de Ribera 2850, Concepción, Chile
A R T I C LE I N FO
A B S T R A C T
Keywords: Anatase nanosheets scc-dftb pm6 Quantum size effect DFT Photocatalysis
Anatase (001) nanosheets have recently attracted great attention as very active catalysts and photocatalysts. These graphene analogs have very high surface area and unique surface properties. In the present paper, very thin two-layer anatase nanosheets are investigated computationally in the form of quantum dots of various size. Quantum size effect (QSE) was clearly observed for nanosheets with fully hydroxylated edges and size up to 14 nm and the ultimate band gap is around 3.4 eV. Dehydroxylation of nanosheets obscured QSE, decreased band gap and induced visible light absorption. Therefore, contradictory trends reported in experimental studies for anatase QSE can be ascribed to different degree of hydroxylation of the TiO2 samples surface. All anatase nanosheet quantum dots retained their flat graphene-like shape. These findings demonstrate that dehydroxylated anatase nanosheet quantum dots are prospective visible-light active photocatalysts even if their inherent band gap is considerably larger than for bulk anatase.
1. Introduction Titanium dioxide remains one of the most studied and utilized materials for photocatalysis, solar cells, batteries, sensors [1], catalysis [2], biomedical and many other applications [3]. Recent advancements in TiO2 preparation methods enabled studies and applications of TiO2 nanoparticles with many different shapes and surfaces [4]. In our previous publications, we studied properties of anatase nanoparticles of various shapes, sizes and different surface functional groups [5–7]. Such developed computational models of TiO2 nanoparticles were subsequently applied for providing understanding of advanced photocatalytic activities of composites [8–10]. However, many studies demonstrated increased functional properties including much elevated rate of photocatalytic reactions over TiO2 in the form of (001) nanosheets as compared to roundish particles and other nanosheets [11,12]. Therefore, studies on such nanosheets are very promising and, no doubt, will be continued for full discovery of the material useful properties [13]. Among the properties of TiO2, its band gap and bands positions are very important from fundamental point of view and for photocatalytic and photoelectrochemical applications. Quantum size effect (QSE) in titanium dioxide is one of the obscure questions in the studies of this material. Modern literature contains contradictory experimental data on the effect of TiO2 particles size on their electronic and optical ⁎
properties. Luca et al. reports optical band gap change from 3.42 to 3.33 eV for change in anatase nanoparticles size from 3.8–11 nm, while commercial 60 nm anatase had band gap of 3.44 eV [14]. In a later paper [15], they reported irregular dependence of the anatase band gap on particles size with a minimum for 3 nm particles. Monticone et al. reported band gap decrease from about 3.4 to 3.2 eV when the size increased from 1 to 3 nm and concluded that the effective mass decreases when the size increases [16]. Satoh et al. reported that three very small anatase nanoparticles containing 6, 14 and 30 TiO2 units obey the effective mass approximation (EMA) with bulk band gap 3.2 eV and dielectric constant 31 [17]. The previously mentioned studies reported deviation from the EMA. To shed some light on the QSE, several theoretical computational studies have been undertaken. Cho et al. investigated nanoparticles in the range of 0.5–3.2 nm with (CAM-)B3LYP hybrid functional and predicted that the bulk-like behavior can emerge for the size above 6.5 nm [18]. In correction to the original paper, they estimated the bulklike size as 20 nm [19]. QSE for anatase (001) bilayer nanoribbons was studied by Vorontsov and Smirniotis [20]. The band gap (Eg) was essentially stable and equal to approximately 3.5 eV for nanoribbons width above 1.3 nm. Local structure deformations and conformation of surface hydroxyl groups exerted strong influence on the Eg and increased or decreased it by up to 0.7 eV. For decahedral anatase nanoparticles, QSE had more complex behavior: surface groups and
Corresponding author. E-mail addresses: [email protected], [email protected] (A.V. Vorontsov).
https://doi.org/10.1016/j.jphotochem.2019.05.001 Received 12 March 2019; Received in revised form 20 April 2019; Accepted 1 May 2019 Available online 01 May 2019 1010-6030/ © 2019 Elsevier B.V. All rights reserved.
Journal of Photochemistry & Photobiology A: Chemistry 379 (2019) 39–46
A.V. Vorontsov and H. Valdés
method also called SCC-DFTB [39]. DFTB + program [40] was used in combination with Slater-Koster (SK) parameter sets tiorg-0-1 [41] and mio-0-1 [42]. Self-consistent charge was calculated with the tolerance 1·10−7 Ha recommended for good precision calculations. The geometry of nanosheets was optimized till maximal component of forces reached below 4.2·10-5 Ha/Bohr that is equivalent to 0.05 kcal/(mol·Å). All atoms of the nanosheets were allowed to move during geometry optimizations. Additional computations of energy and structure optimization were performed with PM6 semiempirical methodology [43] and, for one case, with pm7 [44]. MOPAC2012 and MOPAC2016 programs were employed for the computations with SCF convergence criterion 1·10−6 kcal/mol and maximum force criterion 0.05 kcal/(mol·Å). Electronic absorption spectra were computed using time dependent dftb method in linear approximation [39]. The number of excitations considered was 4000 and this limited the range of the spectrum in the UV region.
coordinatively unsaturated surface Ti atoms with charge < 4+ influenced the Eg [6]. According to the computational results, bulklike properties are expected for decahedral nanoparticles with the size above 2 nm. Due to the predominance of visible light over UV in the solar spectrum, modern research aims at developing photosensitive materials active at λ > 400 nm. Visible light absorption in TiO2 materials including nanosheets can be caused by the presence of point defects such as Ti3+ and oxygen vacancies as well as by some admixtures [21] and heterojunctions [22]. These defects can appear even if oxygen rich air atmosphere is used for calcination of TiO2 as changes in the visible part of diffuse reflection spectrum showed [23]. Traditionally, introduction of dopant atoms is required to obtain visible light activity of TiO2 including nanosheets [24,25]. However, recent studies showed that intrinsic defects can also cause absorption and reactions under visible light [26,27] as well as advanced sensor properties [28]. Even perfect anatase nanocrystals of various shapes possess surface point defects that can cause visible light absorption and photoactivity [6,29]. Despite the very large interest in 2D TiO2 and its numerous applications in photocatalysis and other fields, theoretical consideration is limited to just few publications [30]. Mogilevsky et al. studied dissociative adsorption of water on delaminated bilayer anatase nanosheets and found small distortion of the nanosheet surface [31]. Vittadini and Casarin studied anatase nanosheets containing from 1 to 6 atom layers using periodic models [32]. They found that 1–3 layer anatase (001) nanosheets undergo reorganization into lepidocrocite and another phase. (101) nanosheets also underwent strong distortions. Among the titanate bilayer structures modeled, H2Ti3O7 was found as the most stable [33]. Liao et al. studied monolayer TiO2 nanosheets of surfaces (010) and (101) and found good charge separation with the attached Cd halcogenide quantum dots [34]. Liu et al. investigated codoped TiO2 nanosheets and found their visible light absorption [35]. Wang et al. studied monolayers of various titanate crystal structures using PBE and HSE06 functionals and obtained their band positions [36]. Oxygen vacancy in lepidocrocite was computationally studied in [37]. In the present work, quantum size effect and optical properties are studied for anatase two-layer nanosheets of approximately square shape. The influence of the rim hydroxyl groups is studied as well. Due to the extremely large specific surface area of anatase nanosheets, this material has very important application area in decontamination of chemical warfare agents and equally toxic chemicals in the dark and under solar irradiation [38]. Enhanced visible light absorption for certain configurations of dehydroxylated nanosheets observed in the present study provides opportunities of desining novel advanced photocatalytic materials.
3. Results and discussion Anatase TiO2 remains the most often used photocatalyst for diverse redox reactions. Its high activity can be further improved via formation of nanosheets with predominant surface (001). The ultimately thinnest anatase nanosheets contain just two layers of atoms since monolayer materials rearrange into lepidocrocite. The present investigation considers such ultrathin nanosheets. Preparation methods usually allow obtaining large-scale nanosheets. Such large nanosheets tend to fold and this decreases the amount of available surface area. Thus, an idea to overcome this disadvantage is to limit the horizontal sizes and obtain nanosheet quantum dots instead of semi-infinite sheets. Such quantum dots can be organized into mesoporous nanostructures for gas phase photocatalytic applications or remain as is for liquid phase suspension reactions. Templates can be selected for obtaining quantum dots with needed size. Previously we studied ultrathin anatase titanium dioxide nanoribbons of various width and found their stability and limited quantum size effect [20]. In the present work, ultrathin nanosheets are of approximately square shape. This was hoped to facilitate observation of QSE and provide more generality for the study on visible-light activity. Nanosheet quantum dots of the present study have fully hydroxylated edges or edges with variable extent of dehydroxylation. Hence, the surface density of OH groups varies in the nanosheet quantum dots wildly between 52 OH/nm2 for the smallest nanoparticle with full edges hydroxylation to 0.6 OH/nm2 for the largest dehydroxylated nanosheet. The large variation seems to be in agreement with literature data for roundish anatase nanoparticles [45]. Fig. 1 shows nanosheet quantum dots with fully hydroxylated edges after they were optimized with the scc-dftb method. This method is known to provide very good agreement of anatase lattice structure with the experimental data. Table S1 summarizes all data on nanosheet quantum dots optimized with scc-dftb. Measurements of the nanosheets size reveal that the originally ideal square shape is distorted with contraction along non-hydroxylated edge direction. The hydroxylated edge direction remained almost intact in length. This phenomenon of contraction along one direction of (001) surface is well known in the literature and was observed for decahedral anatase nanoparticles as well [6]. However, this deviation of the crystal structure in (001) surface does not cause any significant aberrations in the nanosheets. The nanosheets quantum dots remain almost ideally flat up to the largest sizes studied here of 4.6 x 3.8 nm. Previously we found that the flatness depends on the number of layers of atoms in the nanosheets: only nanosheets with even number of atomic layers are flat [20]. This necessitates very careful choice of preparation methods to selectively produce even number of layers nanosheets. When the size of the nanosheets increases from TNS11 to TNS22 (Fig. 1), drastic changes in the atomic structure proceed. The structure of titanium atoms surroundings changes from tetrahedra TiO4 in
2. Models and methods Anatase TiO2 nanosheets quantum dots (TNSQD) were prepared from the bulk TiO2 lattice by removing all atoms except two layers of atoms closest to the (001) facet surface. The series of two layer TNSQD studied in the present work are square flakes with the number of unit cells from 1 × 1–12 × 12. The nanosheet quantum dots are designated as TNSKKLL-MvN, where KK is the length of the hydroxylated side in unit cells, LL is the length of the non-hydroxylated edges in unit cells, M is the number of the nanosheet related to the extent of hydroxylation of edges. Additionally, designation v2 was added that means that all hydroxyl groups along each of the edge are aligned in the same direction, and alignment direction is anti-parallel in the neighbor edges of nanosheet. Coordinatively unsaturated Ti atoms with charge less than 4+ in the edges of the nanosheets were saturated with attached OH groups, or oxygen atoms. In the later case, two Ti atoms contain one O atom, which causes non-uniformity of charge among rim Ti atoms. Energy has been obtained in the course of computations with DFTB2 40
Journal of Photochemistry & Photobiology A: Chemistry 379 (2019) 39–46
A.V. Vorontsov and H. Valdés
Fig. 1. Anatase nanosheet quantum dots of approximately square shape with fully hydroxylated edges optimized with scc-dftb.
TNS11-3 and TNS11-3v2 to distorted square pyramids TiO5 for TNS22 and larger nanosheets. The coordination of Ti with five oxygen atoms is typical for (001) surfaces of anatase. The typical TieO bond length for atoms inside the nanosheets is 1.81–1.99 Å and these bond length is also typical for the edge atoms. The OeTieO angle is 116 – 123° for the internal Ti atoms and 126 – 129° for the corresponding OeTi−OH valent angles measured by taking axial O-Ti bonds. Table S1 lists positions of HOMO (valence band edge), LUMO (conduction band edge) and Eg for the nanosheet quantum dots. Recent experimental measurements of anatase band edges estimated ECB = -3.95 eV, EVB = -7.1 eV and Eg = 3.15 eV [46]. Results of scc-dftb method computations gave bands edges shifted by about + 1.5 eV from these experimental values. The exact band edges positions depend slightly on the size of nanoparticles and strongly on the degree of edges hydroxylation: hydroxylation results in bands edge shift to positive potentials by up to 0.6 eV. Fig. 2 illustrates QSE in the nanosheet quantum dots with fully hydroxylated edges. The value of the Eg decreases upon increase in the diameter of the TiO2 quantum dot in agreement with the theory of QSE. Eg decreases clearly upon increase in size from 0.5 to 2.0 nm. TNSQD of larger diameter (diagonal) show scattering of band gap values due to some small changes in the structure including different conformations of edge hydroxyl groups. Consequently, the tendency is not so clear for the larger TNSQD. Least-square fitting with double-exponential decay shown in Fig. 2 points to a fast decrease of band gap for sizes up to 1 nm
Fig. 2. Quantum size effect for anatase nanosheet quantum dots with fully hydroxylated edges according to scc-dftb method results.
41
Journal of Photochemistry & Photobiology A: Chemistry 379 (2019) 39–46
A.V. Vorontsov and H. Valdés
Fig. 3. Anatase nanosheet quantum dots of approximately square shape with dehydroxylated edges optimized with scc-dftb.
and gradual slow decrease for further increase in size. The final value of Eg = 3.4 eV is predicted to be attained at a diameter of approximately 14 nm. This predicted range of QSE for TNSQD is in line with predictions by Cho et al. for anatase nanoparticles that their ultimate band gap is attained at size well above suggested exciton size for TiO2 of 2 nm [18,19]. Removal of chemisorbed water from the surface of TiO2 can be carried out via moderate temperature treatment or chemical reagents application. Fig. 3 shows structure of the TNS with partially dehydroxylated edges. Properties of the optimized TNS are given in Table S1. Dehydroxylation resulted in a considerable decrease of the length of hydroxylated edges while the length of non-hydroxylated edges changed little. The nanosheets retained their flatness even when almost all OH groups were removed and the edges were saturated with oxygen atoms. The structure of the TNS edges became distorted for a larger extent when more hydroxyl groups were removed. Electronic band gap for all dehydroxylated TNSQD of different diameter is shown in Fig. 4. The band gap changes rather irregularly when the size of TiO2 nanosheet increases. This can at least partially explain why some experimental studies report contradictory dependences of optical properties of TiO2 samples on their size. The
Fig. 4. Electronic band gap plotted as a function of size for dehydroxylated anatase quantum dots according to scc-dftb.
42
Journal of Photochemistry & Photobiology A: Chemistry 379 (2019) 39–46
A.V. Vorontsov and H. Valdés
Fig. 5. Anatase nanosheet quantum dots with hydroxylated edges optimized with pm6.
general tendency is however retained: band gap decreases when going from small nanoparticles with the diameter below 3 nm to larger nanoparticles. The band gap is below 3 eV for six kinds of quantum dots having different degree of dehydroxylation. Visible light activity appear first for TNS44-2 and is present for the nanosheets of size above 3 nm. There are two causes here for the band gap below 3 eV – appearance of edge states and deformation of edges – that result from dehydroxylation. Table S1 shows that besides decrease in Eg, dehydroxylation shifts HOMO and LUMO energies towards more negative values. Introduction of visible light activity in small nanoparticles of anatase nanosheets via dehydroxylation of their edges opens a new door to preparation of TiO2-based visible light photocatalysts. All the computations results presented above were obtained with scc-dftb method that was derived from DFT. The method parameters were specially developed for modeling of TiO2 phases. Independent prove of this method results can be done using pm6 method that was derived independently from completely different theory of HartreeFock. Fig. 5 shows atomic structure of nanosheet quantum dots with completely hydroxylated edges. We can see that all nanosheets have flat graphene-like shape. Similarly to the results of scc-dftb computations, the smallest nanoparticle TNS11-3 has strongly distorted structure. Nanoparticle TNS33-3 was also optimized with more recent pm7 method. It is again confirmed that pm7 method is unsuitable for modeling of TiO2 nanoparticles since structure of the TNS33-3 nanosheet optimized with this method is completely different from results of scc-dftb and pm6 method as well as from experimental anatase lattice
Fig. 6. Quantum size effect in anatase bilayer quantum dots with fully hydroxylated edges according to results of pm6 method.
structure. Fig. 6 shows band gap of hydroxylated nanosheets computed with pm6 method and scaled band gap using the factor of 0.37. This scaling factor was selected by fitting of the resultant band gap values to those 43
Journal of Photochemistry & Photobiology A: Chemistry 379 (2019) 39–46
A.V. Vorontsov and H. Valdés
Fig. 7. Anatase bilayer nanosheets with dehydroxylated edges optimized with pm6.
obtained with the scc-dftb method. Unscaled values of the band gap are too large and vary between 9 and 10 eV. Scaling results in more realistic values of 3.5–3.8 eV. Thus, scaled values of Eg computed with pm6 method could probably be used for the rough estimation of band gap for real nanosheet materials. For different shapes of TiO2 nanoparticles, the scaling factor for the band gap obtained with pm6 method is different. For example, for decahedral anatase nanoparticles Ti22r1 – Ti55r4, the scaling factor is 0.31. However, QSE produced with pm6 is completely unrealistic and shows an increase of band gap with an increase in anatase nanosheet quantum dots size. Therefore, such features as dependence of electronic properties on size of nanoparticles cannot be expected to be correctly described and predicted with pm6 method. Fig. 7 demonstrates chemical structure of TNSQD with dehydroxylated edges obtained using pm6 method. The chemical structure features such as relatively strong deformations in the edges are similar to those from the scc-dftb method. The nanosheets remain of almost ideally flat shape. The structural results produced by pm6 method for anatase nanosheets agree qualitatively with the results obtained with the scc-dftb method. Therefore, pm6 method, which is computationally more robust than scc-dftb in presently available implementations, can be used for the studies on TiO2 nanoparticles structures for diverse sets of shapes provided suitable corrections are applied for the obtained unit cell sizes to bring them in general agreement with the experimental anatase lattice parameters. The values of the band gap for the dehydroxylated TNSQD computed with pm6 are given in Fig. 8. Since the predicted values of the band gap overestimate strongly experimentally available values of TiO2 nanoparticles band gap, a correction should be applied. The best value of the correction factor seems to be 0.37 and this scaling factor was used in Fig. 8 for plotting the scaled band gap. Obtained values of the
Fig. 8. Electronic band gap in dehydroxylated square bilayer anatase nanosheets according to pm6 computations.
corrected band gap are around 3.1 eV for the nanosheets having UV absorption while values around 2.7–2.9 eV are obtained for the nanosheets with visible light absorption. Giving the incorrect predictions of the QSE of the pm6 semiempirical method, it was surprising to find out that the scaled band gap values correctly point to TNSQD having visible light absorption in accord to the scc-dftb method results. The exact values of the band gap deviate from the values obtained with the 44
Journal of Photochemistry & Photobiology A: Chemistry 379 (2019) 39–46
A.V. Vorontsov and H. Valdés
Fig. 9, all absorption curves can be linearly interpolated to the absorption threshold of around 340 nm. This threshold corresponds to the value of the optical band gap of 3.65 eV which agrees very well with the values of electronic band gap of around 3.6 eV as shown in Table S1. Therefore, in the case of TNSQD with UV light activity in the present study, there is a good agreement between the values of electronic and optical band gap. The blue shift of absorption threshold for TiO2 nanosheets seen in Fig. 9 agrees with available experimental data for very thin TiO2 nanosheets [22]. However, preparation of pure bilayer nanosheets still remains a challenge for experimental methods and exact comparison with experimental data is not available yet. Absorption spectra for nanosheets having electronic band gap corresponding to visible light absorption are shown in Fig. 10. As it was discussed above, these TNSQD have dehydroxylated edges and their prominent visible light absorption is related to appearance of admixture states localized predominantly in their edges. It can be generally seen in Fig. 10 that visible light absorption for λ > 400 nm is not so strong for the studied TNSQD. There are irregularities in the UV–vis spectra for these materials that are related to the strong contribution of admixture energy levels, which contrasts with the spectra for the nanosheets having intrinsic absorption only in Fig. 9. Absorption edges for these nanosheets deviate from the wavelengths corresponding to the electronic band gap values. The largest wavelength of absorption is obtained for TNS77-2v2 and TNS99-1 that have small absorption peaks approximately at 520 nm. Interestingly, this agrees with the electronic band gap of the TNS99-1 (525 nm). Thus, computed electronic spectra confirm visible light activity of anatase nanosheets quantum dots with dehydroxylated edges. These materials are prospective extremely high surface area adsorbents and visible light active photocatalysts.
Fig. 9. UV absorption spectra of square bilayer anatase nanosheets having no bands in the visible region as resulted from td-dftb calculations. The short wavelength boundary is limited by the number of excitations produced in calculation.
4. Conclusion The present computational study on approximately square bilayer anatase nanosheets quantum dots with the diameter up to 6 nm has led to the following conclusions. 1. Despite their extremely small thickness below 0.5 nm, bilayer anatase nanosheets with hydroxylated and dehydroxylated edges retain anatase-like chemical structure and are essentially flat. 2. Anatase nanosheets with fully hydroxylated edges demonstrate a clear quantum size effect with electronic band gap fitting relatively well to bi-exponential decay. Ultimate band gap of 3.4 eV is attained for diameter approximately 14 nm. Electronic band gap values for these nanosheets agree well with optical band gap values. 3. Some anatase TNSQD with different extent of dehydroxylation of edges possess electronic band gap below 3 eV. Computed electronic absorption spectra confirm that these nanosheets have significant absorption in the visible light with wavelength up to 530 nm. 4. According to the results of the present study, strong difference between published experimentally obtained band gap values for anatase nanoparticles of variable size can be ascribed, at least partially, to different extent of surface hydroxylation of the samples used. Computational results of the investigation suggests that novel visible-light active TiO2 – based photocatalysts with extremely high surface area can be obtained in the form of several nanometers sized nanosheets with dehydroxylated edges and even number of atomic layers.
Fig. 10. Electronic spectra of approximately square partially dehydroxylated anatase nanosheets showing absorption in the visible light region according to td-dftb computations.
scc-dftb method considerably, which is an expected result, when we take into consideration generally poor ability of pm6 method to predict electronic properties qualitatively. In the above consideration, electronic band gap values were considered for diverse hydroxylated and dehydroxylated anatase bilayer nanosheet quantum dots. The electronic band gap is calculated from the values of valence band (HOMO) and conduction band (LUMO) edges. Experimentally, however, band gap values are often obtained using UV–vis diffuse reflection spectra. The obtained thereby band gap values are related to light absorption and can be quite different from the electronic band gap because optical absorption have strong dependence on the density of states of materials. Therefore, it was important to consider optical band gaps of the TNSQD in the present study. Fig. 9 shows absorbance for the TNSQD that are not visible light active. The smallest nanosheets are not considered since they have deformed structure and would hardly be obtained experimentally and they have strongly different values of the band gap. As we can see in
Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.jphotochem.2019.05. 001. References [1] J. Zhang, J. Liao, F. Yang, M. Xu, S. Lin, Regulation of the electroanalytical
45
Journal of Photochemistry & Photobiology A: Chemistry 379 (2019) 39–46
A.V. Vorontsov and H. Valdés
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24] D. Lu, P. Fang, Y. Liu, Z. Liu, X. Liu, Y. Gao, F. Chen, F. Niu, A facile one-pot synthesis of gadolinium doped TiO2-based nanosheets with efficient visible lightdriven photocatalytic performance, J. Nanoparticle Res. 16 (2014), https://doi.org/ 10.1007/s11051-014-2636-3. [25] L. Jiang, L. Sun, D. Yang, J. Zhang, Y.-J. Li, K. Zou, W.-Q. Deng, Niobium-doped (001)-Dominated anatase TiO 2 nanosheets as photoelectrode for efficient dyesensitized solar cells, ACS Appl. Mater. Interfaces 9 (2017) 9576–9583, https://doi. org/10.1021/acsami.6b14147. [26] J.-Y. Zheng, S.-H. Bao, Y. Guo, P. Jin, Anatase TiO 2 films with dominant {001} facets fabricated by direct-current reactive magnetron sputtering at room temperature: oxygen defects and enhanced visible-light photocatalytic behaviors, ACS Appl. Mater. Interfaces 6 (2014) 5940–5946, https://doi.org/10.1021/am500979j. [27] H. Gao, G. Hu, J. Sui, C. Mu, W. Shangguan, M. Kong, W. Shentu, Scalable preparation of defect-rich free-standing TiO 2 sheets with visible-light photocatalytic activity, Appl. Catal. B 226 (2018) 337–345, https://doi.org/10.1016/j.apcatb. 2017.12.055. [28] Y. Wang, J. Liu, M. Wang, C. Pei, B. Liu, Y. Yuan, S. Liu, H. Yang, Enhancing the sensing properties of TiO 2 nanosheets with exposed {001} facets by a hydrogenation and sensing mechanism, Inorg. Chem. 56 (2017) 1504–1510, https://doi. org/10.1021/acs.inorgchem.6b02603. [29] A.V. Vorontsov, Cluster models of photocatalytic anatase TiO2 nanoparticles and their computational characterization, Catal. Today 252 (2015) 168–176, https:// doi.org/10.1016/j.cattod.2014.10.017. [30] F. De Angelis, C. Di Valentin, S. Fantacci, A. Vittadini, A. Selloni, Theoretical studies on anatase and less common TiO 2 phases: bulk, surfaces, and nanomaterials, Chem. Rev. 114 (2014) 9708–9753, https://doi.org/10.1021/cr500055q. [31] G. Mogilevsky, Q. Chen, H. Kulkarni, A. Kleinhammes, W.M. Mullins, Y. Wu, Layered nanostructures of delaminated anatase: nanosheets and nanotubes, J. Phys. Chem. C 112 (2008) 3239–3246, https://doi.org/10.1021/jp076899p. [32] A. Vittadini, M. Casarin, Ab initio modeling of TiO2 nanosheets, Theor. Chem. Acc. 120 (2008) 551–556, https://doi.org/10.1007/s00214-008-0425-8. [33] M. Casarin, A. Vittadini, A. Selloni, First principles study of Hydrated/Hydroxylated TiO 2 nanolayers: from isolated sheets to stacks and tubes, ACS Nano 3 (2009) 317–324, https://doi.org/10.1021/nn800608n. [34] T. Liao, Z. Sun, S.X. Dou, Theoretically manipulating quantum dots on two-dimensional TiO 2 monolayer for effective visible light absorption, ACS Appl. Mater. Interfaces 9 (2017) 8255–8262, https://doi.org/10.1021/acsami.6b15741. [35] Y. Liu, W. Zhou, P. Wu, The electronic structure and optical properties of donoracceptor codoped TiO2 nanosheets from hybrid functional calculations, Mater. Chem. Phys. 186 (2017) 333–340, https://doi.org/10.1016/j.matchemphys.2016. 11.004. [36] L. Wang, D. Wei, S. Kang, X. Xie, Y. Shi, S. Liu, Two-dimensional titania: structures and properties predicted by first principle calculation, J. Phys. Chem. C 122 (2018) 22911–22919, https://doi.org/10.1021/acs.jpcc.8b05412. [37] L. Wang, G. Zhou, Y. Shan, Z. Huang, L. Liu, Oxygen-defect-dependent ferromagnetism and strain modulation in free-standing two-dimensional TiO 2 monolayers, J. Chem. Soc. Faraday Trans. 20 (2018) 27176–27184, https://doi.org/10. 1039/C8CP05223A. [38] A. Sengele, D. Robert, N. Keller, C. Colbeau-Justin, V. Keller, Sn-doped and porogen-modified TiO2 photocatalyst for solar light elimination of sulfure diethyle as a model for chemical warfare agent, Appl. Catal. B 245 (2019) 279–289, https:// doi.org/10.1016/j.apcatb.2018.12.071. [39] T.A. Niehaus, S. Suhai, F. Della Sala, P. Lugli, M. Elstner, G. Seifert, Th. Frauenheim, Tight-binding approach to time-dependent density-functional response theory, Phys. Rev. B 63 (2001), https://doi.org/10.1103/PhysRevB.63.085108. [40] B. Aradi, B. Hourahine, Th. Frauenheim, DFTB+, a Sparse Matrix-Based Implementation of the DFTB Method†, J. Phys. Chem. A 111 (2007) 5678–5684, https://doi.org/10.1021/jp070186p. [41] G. Dolgonos, B. Aradi, N.H. Moreira, T. Frauenheim, An improved self-consistentCharge density-functional tight-binding (SCC-DFTB) set of parameters for simulation of bulk and molecular systems involving titanium, J. Chem. Theory Comput. 6 (2010) 266–278, https://doi.org/10.1021/ct900422c. [42] M. Elstner, D. Porezag, G. Jungnickel, J. Elsner, M. Haugk, Th. Frauenheim, S. Suhai, G. Seifert, Self-consistent-charge density-functional tight-binding method for simulations of complex materials properties, Phys. Rev. B 58 (1998) 7260–7268, https://doi.org/10.1103/PhysRevB.58.7260. [43] J.J.P. Stewart, Optimization of parameters for semiempirical methods V: modification of NDDO approximations and application to 70 elements, J. Mol. Model. 13 (2007) 1173–1213, https://doi.org/10.1007/s00894-007-0233-4. [44] J.J.P. Stewart, Optimization of parameters for semiempirical methods VI: more modifications to the NDDO approximations and re-optimization of parameters, J. Mol. Model. 19 (2013) 1–32, https://doi.org/10.1007/s00894-012-1667-x. [45] R. Mueller, H.K. Kammler, K. Wegner, S.E. Pratsinis, OH surface density of SiO 2 and TiO 2 by thermogravimetric analysis, Langmuir. 19 (2003) 160–165, https://doi. org/10.1021/la025785w. [46] C. Maheu, L. Cardenas, E. Puzenat, P. Afanasiev, C. Geantet, UPS and UV spectroscopies combined to position the energy levels of TiO 2 anatase and rutile nanopowders, J. Chem. Soc. Faraday Trans. 20 (2018) 25629–25637, https://doi.org/ 10.1039/C8CP04614J.
performance of ultrathin titanium dioxide nanosheets toward lead ions by nonmetal doping, Nanomaterials. 7 (2017) 327, https://doi.org/10.3390/ nano7100327. H. Wang, Y. Song, J. Xiong, J. Bi, L. Li, Y. Yu, S. Liang, L. Wu, Highly selective oxidation of furfuryl alcohol over monolayer titanate nanosheet under visible light irradiation, Appl. Catal. B 224 (2018) 394–403, https://doi.org/10.1016/j.apcatb. 2017.10.069. F. Wu, Z. Wang, X. Li, H. Guo, Simple preparation of petal-like TiO2 nanosheets as anode materials for lithium-ion batteries, Ceram. Int. 40 (2014) 16805–16810, https://doi.org/10.1016/j.ceramint.2014.07.060. G. Liu, H.G. Yang, J. Pan, Y.Q. Yang, G.Q.(Max) Lu, H.-M. Cheng, Titanium dioxide crystals with tailored facets, Chem. Rev. 114 (2014) 9559–9612, https://doi.org/ 10.1021/cr400621z. A.V. Vorontsov, S.V. Tsybulya, Influence of nanoparticles size on XRD patterns for small monodisperse nanoparticles of Cu ° and TiO 2 anatase, Ind. Eng. Chem. Res. 57 (2018) 2526–2536, https://doi.org/10.1021/acs.iecr.7b04480. A.V. Vorontsov, P.G. Smirniotis, Size and surface groups effects in decahedral anatase nanoparticles for photocatalytic applications, J. Photochem. Photobiol. A: Chem. 363 (2018) 51–60, https://doi.org/10.1016/j.jphotochem.2018.05.031. A.V. Vorontsov, Structural and electronic effects in acetone adsorption over TiO2 anatase clusters as the first stage of photocatalytic oxidation, J. Nanoparticle Res. 19 (2017) 326, https://doi.org/10.1007/s11051-017-4025-1. J.O. Olowoyo, M. Kumar, B. Singh, V.O. Oninla, J.O. Babalola, H. Valdés, A.V. Vorontsov, U. Kumar, Self-assembled reduced graphene oxide-TiO2 nanocomposites: synthesis, DFTB+ calculations, and enhanced photocatalytic reduction of CO2 to methanol, Carbon. 147 (2019) 385–397, https://doi.org/10.1016/j. carbon.2019.03.019. J.O. Olowoyo, M. Kumar, S.L. Jain, J.O. Babalola, A.V. Vorontsov, U. Kumar, Insights into reinforced photocatalytic activity of the CNT–TiO 2 nanocomposite for CO 2 reduction and water splitting, J. Phys. Chem. C 123 (2019) 367–378, https:// doi.org/10.1021/acs.jpcc.8b07894. J.O. Olowoyo, M. Kumar, N. Singhal, S.L. Jain, J.O. Babalola, A.V. Vorontsov, U. Kumar, Engineering and modeling the effect of Mg doping in TiO 2 for enhanced photocatalytic reduction of CO 2 to fuels, Catal. Sci. Technol. 8 (2018) 3686–3694, https://doi.org/10.1039/C8CY00987B. C. Hu, X. Zhang, W. Li, Y. Yan, G. Xi, H. Yang, J. Li, H. Bai, Large-scale, ultrathin and (001) facet exposed TiO2 nanosheet superstructures and their applications in photocatalysis, J. Mater. Chem. A 2 (2014) 2040, https://doi.org/10.1039/ c3ta14343k. L. Wang, T. Sasaki, Titanium oxide nanosheets: graphene analogues with versatile functionalities, Chem. Rev. 114 (2014) 9455–9486, https://doi.org/10.1021/ cr400627u. C.P. Sajan, S. Wageh, Ahmed.A. Al-Ghamdi, J. Yu, S. Cao, TiO2 nanosheets with exposed {001} facets for photocatalytic applications, Nano Res. 9 (2016) 3–27, https://doi.org/10.1007/s12274-015-0919-3. V. Luca, S. Djajanti, R.F. Howe, Structural and electronic properties of sol−Gel titanium oxides studied by X-ray absorption spectroscopy, J. Phys. Chem. B 102 (1998) 10650–10657, https://doi.org/10.1021/jp981644k. V. Luca, Comparison of size-dependent structural and electronic properties of anatase and rutile nanoparticles, J. Phys. Chem. C 113 (2009) 6367–6380, https:// doi.org/10.1021/jp808358v. S. Monticone, R. Tufeu, A.V. Kanaev, E. Scolan, C. Sanchez, Quantum size effect in TiO2 nanoparticles: does it exist? Appl. Surf. Sci. 162–163 (2000) 565–570, https:// doi.org/10.1016/S0169-4332(00)00251-8. N. Satoh, T. Nakashima, K. Kamikura, K. Yamamoto, Quantum size effect in TiO2 nanoparticles prepared by finely controlled metal assembly on dendrimer templates, Nat. Nanotechnol. 3 (2008) 106–111, https://doi.org/10.1038/nnano. 2008.2. D. Cho, K.C. Ko, O. Lamiel-García, S.T. Bromley, J.Y. Lee, F. Illas, Effect of size and structure on the ground-state and excited-state electronic structure of TiO 2 nanoparticles, J. Chem. Theory Comput. 12 (2016) 3751–3763, https://doi.org/10. 1021/acs.jctc.6b00519. D. Cho, K.C. Ko, O. Lamiel-García, S.T. Bromley, J.Y. Lee, F. Illas, Correction to effect of size and structure on the ground-state and excited-state electronic structure of TiO 2 nanoparticles, J. Chem. Theory Comput. 13 (2017), https://doi.org/10. 1021/acs.jctc.7b01087 5822–5822. A.V. Vorontsov, P.G. Smirniotis, Structure, electronic and optical properties of bilayer anatase nanoribbons, Comput. Mater. Sci. 155 (2018) 266–281, https://doi. org/10.1016/j.commatsci.2018.08.052. J. Cai, Z. Huang, K. Lv, J. Sun, K. Deng, Ti powder-assisted synthesis of Ti3+ selfdoped TiO2 nanosheets with enhanced visible-light photoactivity, RSC Adv. 4 (2014) 19588, https://doi.org/10.1039/c4ra01496k. Y. Xin, Y. Lu, C. Han, L. Ge, P. Qiu, Y. Li, S. Fang, Novel NiS cocatalyst decorating ultrathin 2D TiO2 nanosheets with enhanced photocatalytic hydrogen evolution activity, Mater. Res. Bull. 87 (2017) 123–129, https://doi.org/10.1016/j. materresbull.2016.11.027. A.V. Vorontsov, A.A. Altynnikov, E.N. Savinov, E.N. Kurkin, Correlation of TiO2 photocatalytic activity and diffuse reflectance spectra, J. Photochem. Photobiol. A: Chem. 144 (2001) 193–196, https://doi.org/10.1016/S1010-6030(01)00462-2.
46