Insight into Electronic and Structural Reorganizations for Defect

ACS2GO © 2018. ← → → ←. loading. To add this web app to the home screen open the browser option menu and tap on Add to homescreen...
0 downloads 0 Views 2MB Size
Letter pubs.acs.org/JPCL

Insight into Electronic and Structural Reorganizations for DefectInduced VO2 Metal−Insulator Transition Xijun Wang,†,§ Zhaowu Wang,‡,§ Guozhen Zhang,† and Jun Jiang*,† †

Hefei National Laboratory for Physical Sciences at the Microscale, iChEM (Collaborative Innovation Center of Chemistry for Energy Materials), CAS Key Laboratory of Mechanical Behavior and Design of Materials, School of Chemistry and Materials Science, University of Science and Technology of China, Hefei, Anhui 230026, China ‡ School of Physics and Engineering, Henan University of Science and Technology, Luoyang City, Henan Province 471023, China S Supporting Information *

ABSTRACT: An oxygen vacancy defect in monoclinic VO2 has been shown to modulate the metal−insulator transition (MIT) at room temperature. However, as the electronic and structural reorganizations occur simultaneously, the origin of MIT is still unclear. Here we performed firstprinciples calculations to examine electronic variations separately from structural reorganizations during MIT. It was found that the oxygen defect induces electronic reorganization by creating polarized 3d orbitial electrons, while structure reorganization makes the conduction band edge states available for occupation. The conduction band states thus hold polarized charges that delocalize over space, bestowing metallic property on the originally insulated VO2. A linear relationship for the number of polarized electrons and the defect concentration is revealed, which would lead to costeffective control of VO2 MIT behavior by defect engineering.

T

high-temperature rutile metal phase VO2 to a low-temperature monoclinic insulator phase VO2.13,22 The defect-induced MIT in VO2 is of particular interest, which was found to be effective in lowering the critical MIT temperature. Recent experimental studies have reported that oxygen vacancy (OV) defect doping to monoclinic VO2 could stabilize the metallic state at room temperature,11,12,23,24 while the microscopic mechanism is still unclear. The problem is that geometric and electronic structure variations occur simultaneously during MIT, bringing huge difficulties to single out their individual contributions to MIT. Xie et al. have identified the key effect of geometric structure rearrangement in driving VO2 MIT,12 while Ruan et al. have found that MIT could be caused by an electronic structure transition originating from charge doping without discernible lattice variation.23 Meanwhile, our recent work revealed that polarized charges from the OV defect can bestow metallic property to the originally insulated monoclinic VO2 by occupying edge states of its conduction band (CB).24 To gain a clear understanding on the MIT mechanism, we performed density functional theory (DFT) calculations to explore electronic and structural reorganizations associated with OV-induced MIT in monoclinic VO2 (namely, VO2−x). Meanwhile, we have assumed a series of “fixed” VO2−x systems, which allow only electronic variations by excluding geometric changes. Both the structure reorganized and “fixed” VO2−x have defect-induced polarized electrons with very similar linear

he metal−insulator transition (MIT) behavior of correlated material has attracted widespread attention owing to its great potential for important applications including optical/electric switching devices, smart materials, light detectors, and temperature sensors.1−3 Various techniques have been developed to modulate the MIT behavior in complex transition-metal oxides, such as chemical doping,4,5 interfacial strain engineering,6−8 electric field modulation,9,10 and defect engineering.2,11 However, understanding the nature of abrupt changes of electric conductance and optical transmission during MIT remains a grand challenge, mainly due to the difficulty dissecting the contributions from structural and electronic reorganizations.12 The proposed mechanisms include electroncorrelation-driven Mott transition,13 structure-driven Peierls transition,14 and the cooperation of those two.12,15 The cooperation mechanism seems to be more convincing because structural reorganizations induced by strain or temperature modulation unavoidably induce changes of the electronic structure,16,17 and electronic variations owing to charge/atom doping or an external field are often accompanied by atomic rearrangements.18 Nevertheless, it is still pressing to reveal the actual cooperation process of structural and electronic rearrangements, as well as their respective contributions to the material metallization process, so as to develop costeffective controls on the MIT behavior. To date, many MIT materials have been found such as MoS2,19 Fe3O4,20 and many early transition-metal oxides.21 Among these, in this study, we focus on vanadium dioxide (VO2), one of the best-known first-order MIT materials with a critical transition temperature at approximately 341 K from a © XXXX American Chemical Society

Received: May 24, 2017 Accepted: June 21, 2017 Published: June 21, 2017 3129

DOI: 10.1021/acs.jpclett.7b01300 J. Phys. Chem. Lett. 2017, 8, 3129−3132

Letter

The Journal of Physical Chemistry Letters

structural rearrangements induced by OV, we built the hypothetical VO2−x models with “fixed” atomic coordinates even after OV creation (namely, fixed VO2−x), based on which electronic structures were computed. Charge distribution analysis was performed with the Bader charge scheme.31 The known energy band structures of defect-free monoclinic VO2 and VO2−x are displayed schematically in Figure 1b, which were further confirmed by our simulated results in Figure 1c,d. It is clear that VO2 holds a direct band gap of ∼0.65 eV with the Fermi level (Ef) located right at the VB top edge (Figure 1c), in line with previous reports.15,32 However, for VO2−x systems with different OV concentrations, their Ef levels were promoted from 4.31 to ∼5.00 eV and went beyond the CB edge (Figures 1d and S3). It suggests that electrons partially filled the bottom edge states of the CB, bestowing metallic property to VO2−x. It should be noted that the VO2−x electronic band features are entirely different from those of metallic rutile VO2 (Figures S4 and S5). Because VO2 and VO2−x have nearly the same energy band structures, the occupation of CB edge states is mainly caused by the electronic reorganization. The geometric rearrangements caused by OV mainly occurred at the three V atoms aside the defective site. Originally, the V−V chains alternate between the long and short V−V pairs with distances of 3.13 and 2.51 Å, respectively, in defect-free VO2, as illustrated in Figure S1. Compared to the fixed VO2−x in Figure 2a, OV mainly leads to two types of

dependence on the OV concentration, while only the reorganized systems offer metallic properties because of the partial occupation of CB edge states by those polarized electrons. In the fixed VO2−x without structural reorganizations, the polarized electrons are trapped at defect sites, contributing to defective states inside of the CB−valence band (VB) gap. Therefore, we speculated that the cooperation mechanism of OV-induced MIT of VO2 involves a couple of connected phenomena. First, the OV defect induces extra defect states, and electronic reorganization creates polarized electrons. Then, the structural reorganization relaxes defect states, which allows polarized electrons to partially occupy CB edge states, and metallic property is obtained with CB edge electrons delocalizing all over the crystal. The linear dependence of the polarized electron number on OV concentration thus provides a path of gradually and precisely controlling VO2 conductance and the MIT process. Our simulations were performed at the DFT level implemented by the Vienna ab initio simulation package (VASP) (see details in the Supporting Information).25 The allelectron projector augmented wave (PAW)26 model and Perdew−Burke−Ernzerhof (PBE) exchange−correlation functional were employed.27 The GGA+U method was applied to describe partially filled 3d orbitals in V.28,29 The first-principlesbased stochastic surface walking (SSW) package HOWTOs program was chosen to search the transition state.30 The modeling was based on a supercell with 2 × 2 × 2 primitive unit cells for the monoclinic vacancy-free VO2 (Figure S1), which contains 96 atoms (32 V atoms and 64 O atoms if not counting the oxygen vacancy). On the basis of the 2 × 2 × 2 VO2 supercell, there are two types of possible oxygen vacancy sites (Figure S2). We focused on the one with the lowest total energy after geometry optimization (Figure 1a). The VO2−x models were built by introducing 1−8 OV defects (distributed as uniformly as possible) in the 2 × 2 × 2 cell. Similarly, a model system with even lower defect concentration was created with only one OV in a 3 × 3 × 3 supercell crystal (324 atoms: 108 V and 216 O). In order to decouple the electronic and

Figure 2. (a) Computed charge distribution of OV-induced electrons (yellow bubbles) on the “fixed” defective monoclinic VO2−x 2 × 2 × 2 cell with one defect. Here V21, V23, and V24 mark three V atoms neighboring the OV defect. (b) Numbers of defective state electrons on each atom in the “fixed” and reorganized VO2−x systems.

geometric deformations: one is that the longer V−V pairs are shortened to 3.06 Å, and the other is that the shorter pairs are stretched to 2.60 Å. Meanwhile, the V−O bond lengths aside the OV site just experience a minor change ranging from −0.07 to 0.10 Å (Table S1). Moreover, this structural reorganization is predicted to be an exothermic process with no barrier according to our simulations. Intriguingly, one OV defect breaks three V− O bonds, which naturally causes polarized negative charges. For the fixed VO2−x, these extra polarized charges are all trapped at the defective site (Figure 2a). It is clear to see from Figure 2b that the computed defective electrons are mostly localized at those three V atoms neighboring OV. This would not help improve electric conductance. In contrast, the defective electrons in the reorganized VO2−x are distributed almost evenly on all atoms in the cell (Figure 2b), suggesting the typical feature of charge delocalization. Now that we understand the space distribution of polarized electrons affected by structural reorganization, the energy states have also been investigated. The simulated density of states (DOS) of the defect-free VO2, fixed VO2−x, and relaxed VO2−x are displayed in Figure 3. Obviously, the basic DOS features are not affected much by the defect, except that the fixed VO2−x

Figure 1. (a) Model structure of the defective monoclinic VO2−x 2 × 2 × 2 cell. Here V21, V23, and V24 mark three V atoms neighboring the OV defect. (b) Schematic band structures of defect-free VO2 and defective VO2−x showing that OV promotes the Fermi level (Ef) and makes the CB edge states d∥* partially filled, as confirmed by computed electronic band structures of the VO2 (c) and the VO2−x cell with one OV defect (d). The red dashed lines are Fermi levels. 3130

DOI: 10.1021/acs.jpclett.7b01300 J. Phys. Chem. Lett. 2017, 8, 3129−3132

Letter

The Journal of Physical Chemistry Letters

Figure 3. Computed electronic DOS of the defect-free VO2, fixed VO2−x, and reorganized VO2−x with one defect, together with the electron distributions (yellow bubbles) of their CB edge states (from the CB bottom to 0.17 eV above, as indicated by dashed boxes). Figure 4. Computed electronic DOS of the fixed VO2−x (a) and reorganized VO2−x (b) with various OV concentrations. The defectinduced electron number (computed by integrating the shadow area in corresponding DOS curves) as a function of OV numbers per unit (defect concentration) in the fixed VO2−x (c) and reorganized VO2−x (d).

exhibits a small defective state in the middle of the CB−VB energy gap of the VO2. This defective state accommodates the OV-induced polarized charges, shifting up the Fermi level in the semiconductor band structure (Figure S6). Nevertheless, the defective state is still ∼0.64 eV lower than the CB edge, representing typical semiconductor property. In contrast, the structure-reorganized VO2−x has no such defective states below the CB, suggesting that the polarized charges have to take the CB edge states, as seen in Figure 3. The resulting Fermi level is about 0.17 eV above the CB bottom edge. Importantly, the CB edge states in VO2 and VO2−x crystals mainly consist of V 3d orbitials, which are distributed evenly all over the lattice. Obviously it favors charge transport and consequently bestows metallic property to the semiconductor. Because polarized charges are induced by defects, it is natural to question whether the conductivity can be controlled by the defect concentration. We examined the relationship between the number of polarized charges and the concentration of oxygen defect. A series of VO2−x configurations with OV concentration ranging from 1/216 (0.5%) to 8/64 (12.5%) were studied. From the simulated DOS features (Figure 4a), one can see that all fixed VO2−x hold fully occupied defective states inside of the CB−VB gap. With increasing defect concentration, the Ef level never goes beyond the CB edge, and CB states remain empty. However, in the reorganized VO2−x systems, all of their CB edge states are partially filled with electrons (Figure 4b). The more that OV defects were introduced, the more that occupied CB states appeared. We have computed the number of polarized charges by integrating the occupied defective states in the fixed VO2−x or the occupied CB edge states in the reorganized VO2−x. In Figure 4c,d, the number of polarized charges increases linearly with the increase of defect concentration in all VO2−x systems. The slope/ratio between defect and polarized charge numbers per one crystal unit cell is nearly the same in both fixed and relaxed VO2−x, implying that each OV defect could induce ∼1.7 units of charges, which could act as a free charge carrier to improve semiconductor conductance. Importantly, this suggests that manipulating the defect concentration is a feasible way to control VO2 MIT behavior and conductivity. In summary, we have revealed the oxygen defect-induced MIT mechanism of monoclinic VO2 and proposed the manipulation of defect concentration to control semiconductor

conductivity. That is, OV defects induce electronic reorganization through the polarization of 3d orbitals; in turn, structural reorganization allows polarized charges to partially occupy the CB edge. The CB thus holds free electrons distributed all over the lattice, favoring charge transport and making the monoclinic VO2−x metallic. The number of polarized charges increases linearly along with the increase of the OV defect concentration, providing a convenient knob to control VO2−x conductivity. These new insights from the case of VO2 can be helpful for studying MIT of other oxygen-deficient correlated oxides



ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpclett.7b01300. Computational details, the optimized monoclinic VO2 2 × 2 × 2 supercell structure, structure of the less stable VO2−x 2 × 2 × 2 cell model with one OV defect, computed electronic band structures of the VO2−x 2 × 2 × 2 cell with two and four defects, the optimized rutile VO2(R) 2 × 2 × 4 supercell structure, computed electronic band of defect-free rutile VO2(R), computed electronic band of fixed geometry monoclinic VO2−x 2 × 2 × 2 cell with one defect, and computed key atomic distances (Å) for the defect-free monoclinic VO2 and VO2−x 2 × 2 × 2 cell with one defect (PDF)



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Guozhen Zhang: 0000-0003-0125-9666 Jun Jiang: 0000-0002-6116-5605 3131

DOI: 10.1021/acs.jpclett.7b01300 J. Phys. Chem. Lett. 2017, 8, 3129−3132

Letter

The Journal of Physical Chemistry Letters Author Contributions

infrared spectroscopy and nano-imaging. Science 2007, 318, 1750− 1753. (14) Booth, J. M.; Casey, P. S. Anisotropic structure deformation in the VO2 metal-insulator transition. Phys. Rev. Lett. 2009, 103, 086402. (15) Koethe, T. C.; Hu, Z.; Haverkort, M. W.; Schüßler-Langeheine, C.; Venturini, F.; Brookes, N. B.; Tjernberg, O.; Reichelt, W.; Hsieh, H. H.; Lin, H. J.; Chen, C. T.; Tjeng, L. H. Transfer of spectral weight and symmetry across the metal-insulator transition in VO2. Phys. Rev. Lett. 2006, 97, 116402. (16) Wu, C. Z.; Wei, H.; Ning, B.; Xie, Y. New vanadium oxide nanostructures: controlled synthesis and their smart electrical switching properties. Adv. Mater. 2010, 22, 1972. (17) Li, Y. F.; Zhu, S. C.; Liu, Z. P. Reaction Network of Layer-toTunnel Transition of MnO2. J. Am. Chem. Soc. 2016, 138, 5371. (18) Yang, J. J.; Pickett, M. D.; Li, X.; Ohlberg, D. A.; Stewart, D. R.; Williams, R. S. Memristive switching mechanism for metal/oxide/ metal nanodevices. Nat. Nanotechnol. 2008, 3, 429−433. (19) Radisavljevic, B.; Kis, A. Mobility engineering and a metal− insulator transition in monolayer MoS2. Nat. Mater. 2013, 12 (9), 815−820. (20) Lee, J.; Kwon, S. G.; Park, J. G.; Hyeon, T. Size Dependence of Metal−Insulator Transition in Stoichiometric Fe3O4 Nanocrystals. Nano Lett. 2015, 15 (7), 4337−4342. (21) Dang, H. T.; Ai, X.; Millis, A. J.; Marianetti, C. A. Density functional plus dynamical mean-field theory of the metal-insulator transition in early transition-metal oxides. Phys. Rev. B: Condens. Matter Mater. Phys. 2014, 90 (12), 125114. (22) Wentzcovitch, R. M.; Schulz, W. W.; Allen, P. B. VO2: Peierls or Mott-Hubbard? A view from band theory. Phys. Rev. Lett. 1994, 72, 3389. (23) Tao, Z. S.; Han, T. T.; Mahanti, S. D.; Duxbury, P. M.; Yuan, F.; Ruan, C. Y. Decoupling of Structural and Electronic Phase Transitions in VO2. Phys. Rev. Lett. 2012, 109, 166406. (24) Chen, S.; Wang, X. J.; Fan, L. L.; Liao, G. M.; Chen, Y. L.; Chu, W. S.; Song, L.; Jiang, J.; Zou, C. W. The Dynamic Phase Transition Modulation of Ion-Liquid Gating VO2 Thin Film: Formation, Diffusion, and Recovery of Oxygen Vacancies. Adv. Funct. Mater. 2016, 26, 3532. (25) Kresse, G.; Furthmüller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 54 (16), 11169. (26) Blöchl, P. E. Projector augmented-wave method. Phys. Rev. B: Condens. Matter Mater. Phys. 1994, 50 (24), 17953. (27) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865−3868. (28) Anisimov, V. I.; Aryasetiawan, F.; Lichtenstein, A. I. Efficient iterative schemes for ab initio total-energy calculations using a planewave basis set. J. Phys.: Condens. Matter 1997, 9, 767. (29) Zheng, X.; Li, C.; Zhang, D. D.; Yang, W. T. Scaling correction approaches for reducing delocalization error in density functional approximations. Sci. China: Chem. 2015, 58, 1825. (30) Shang, C.; Zhang, X. J.; Liu, Z. P. SSW Package HOWTOs; Fudan University: China, 2013. (31) Tang, W.; Sanville, E.; Henkelman, G. A grid-based Bader analysis algorithm without lattice bias. J. Phys.: Condens. Matter 2009, 21, 084204. (32) Cui, Y. Y.; Shi, S. Q.; Chen, L. L.; Luo, H. J.; Gao, Y. F. Hydrogen-doping induced reduction in the phase transition temperature of VO2: a first-principles study. Phys. Chem. Chem. Phys. 2015, 17, 20998.

§

X.W. and Z.W. contributed equally. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was financially supported by the MOST 973 Program (No. 2014CB848900), NSFC (No. 21633006, 21473166, 11404095), and CAS Strategic Priority Research Program B (No. XDB01020000). Numerical calculations were done on the supercomputing system in the Supercomputing Center of USTC and the National Supercomputing Center in Changsha (NSCC-CS).



REFERENCES

(1) Wu, C. Z.; Feng, F.; Xie, Y. Design of vanadium oxide structures with controllable electrical properties for energy applications. Chem. Soc. Rev. 2013, 42, 5157. (2) Jeong, J.; Aetukuri, N.; Graf, T.; Schladt, T. D.; Sa-mant, M. G.; Parkin, S. S. Suppression of metal-insulator transition in VO2 by electric field−induced oxygen vacancy formation. Science 2013, 339, 1402. (3) Yoon, H.; Choi, M.; Lim, T. W.; Kwon, H.; Ihm, K.; Kim, J. K.; Choi, S. Y.; Son, J. Reversible phase modulation and hydrogen storage in multivalent VO2 epitaxial thin films. Nat. Mater. 2016, 15, 1113. (4) Mlyuka, N. R.; Niklasson, G. A.; Granqvist, C. G. Mg doping of thermochromic VO2 films enhances the optical transmittance and decreases the metal-insulator transition temperature. Appl. Phys. Lett. 2009, 95, 171909. (5) Lee, S.; Cheng, C.; Guo, H.; Hippalgaonkar, K.; Wang, K.; Suh, J.; Liu, K.; Wu, J. Q. Axially engineered metal−insulator phase transition by graded doping VO2 nanowires. J. Am. Chem. Soc. 2013, 135 (12), 4850−4855. (6) Aetukuri, N. B.; Gray, A. X.; Drouard, M.; Cossale, M.; Gao, L.; Reid, A. H.; Kukreja, R.; Ohldag, H.; Jenkins, C. A.; Arenholz, E.; Roche, K. P.; Dürr, H. A.; Samant, M. G.; Parkin, S. S. P. Control of the metal-insulator transition in vanadium dioxide by modifying orbital occupancy. Nat. Phys. 2013, 9, 661. (7) Fan, L. L.; Chen, S.; Luo, Z. L.; Liu, Q. H.; Wu, Y. F.; Song, L.; Ji, D. X.; Wang, P.; Chu, W. S.; Gao, C.; Zou, C. W.; Wu, Z. Y. Strain dynamics of ultrathin VO2 film grown on TiO2 (001) and the associated phase transition modulation. Nano Lett. 2014, 14 (7), 4036. (8) Cao, J.; Ertekin, E.; Srinivasan, V.; Fan, W.; Huang, S.; Zheng, H.; Yim, J. W. L.; Khanal, D. R.; Ogletree, D. F.; Grossman, J. C.; Wu, J. Strain engineering and one-dimensional organization of metalinsulator domains in single-crystal vanadium dioxide beams. Nat. Nanotechnol. 2009, 4, 732−737. (9) Ye, J. T.; Zhang, Y. J.; Akashi, R.; Bahramy, M. S.; Arita, R.; Iwasa, Y. Superconducting dome in a gate-tuned band insulator. Science 2012, 338, 1193. (10) Kumar, S.; Pickett, M. D.; Strachan, J. P.; Gibson, G.; Nishi, Y. S.; Williams, R. S. Local Temperature Redistribution and Structural Transition During Joule-Heating-Driven Conductance Switching in VO2. Adv. Mater. 2013, 25, 6128. (11) Appavoo, K.; Lei, D. Y.; Sonnefraud, Y.; Wang, B.; Pantelides, S. T.; Maier, S. A.; Haglund, R. F., Jr. Role of defects in the phase transition of VO2 nanoparticles probed by plasmon resonance spectroscopy. Nano Lett. 2012, 12 (2), 780−786. (12) Yao, T.; Zhang, X. D.; Sun, Z. H.; Liu, S. J.; Huang, Y. Y.; Xie, Y.; Wu, C. Z.; Yuan, X.; Zhang, W. Q.; Wu, Z. Y.; Pan, G. Q.; Hu, F. C.; Wu, L. H.; Liu, Q. H.; Wei, S. Q. Understanding the nature of the kinetic process in a VO2 metal-insulator transition. Phys. Rev. Lett. 2010, 105, 226405. (13) Qazilbash, M. M.; Brehm, M.; Chae, B. G.; Ho, P. C.; Andreev, G. O.; Kim, B. J.; Yun, S. J.; Balatsky, A. V.; Ma-ple, M. B.; Keilmann, F.; Kim, H. T.; Basov, D. N. Mott transition in VO2 revealed by 3132

DOI: 10.1021/acs.jpclett.7b01300 J. Phys. Chem. Lett. 2017, 8, 3129−3132