Competition between magnetic order and charge localization in Na2IrO3 thin crystal devices
Abstract
Spin orbit assisted Mott insulators such as sodium iridate (Na2IrO3) have been an important subject of study in the recent years. In these materials, the interplay of electronic correlations, spin-orbit coupling, crystal field effects and a honeycomb arrangement of ions bring exciting ground states, predicted in the frame of the Kitaev model. The insulating character of Na2IrO3 has hampered its integration to an electronic device, desirable for applications, such as the manipulation of quasiparticles interesting for topological quantum computing. Here we show through electronic transport measurements supported by Angle Resolved Photoemission Spectroscopy (ARPES) experiments, that electronic transport in Na2IrO3 is ruled by variable range hopping and it is strongly dependent on the magnetic ordering transition known for bulk Na2IrO3, as well as on external electric fields. Electronic transport measurements allow us to deduce a value for the localization length and the density of states in our Na2IrO3 thin crystals devices, offering an alternative approach to study insulating layered materials.
Sodium iridate (Na2IrO3) and other 5d iridates have been widely studied in recent years as they have been identified as experimental realizations of the Kitaev model, an exactly solvable model that describes a set of spin-1/2 moments in a honeycomb lattice with highly anisotropic exchange interaction and exciting ground states such as spin liquids Chun ; Singh ; Winter . In real materials such as Na2IrO3, the existence of Heisenberg and off diagonal interactions benefit the appearance of magnetically ordered ground states over a spin liquid. Recent studies have explored the effect of external pressure and high magnetic fields on the ground states of Na2IrO3 6Proposal ; 7Proposal ; 8Proposal . First principle calculations 6Proposal have found that under high pressure, Na2IrO3 goes through successive structural and magnetic phase transitions, some of them zigzag magnetic ordered and some nonmagnetic, all with low energy excitations that can be well described by jeff=1/2 states and in particular having some phases (with a structure corresponding to space group )6Proposal resembling a gapped spin liquid Kitaev state. Similarly, magnetic torque measurements have found that at magnetic fields up to 60T, long range spin correlations functions decay rapidly pointing to a field-induced quantum spin liquid 8Proposal . In the absence of high pressure or magnetic fields, recent inelastic x-ray scattering measurements have reported a proximate spin liquid regime above the long-range ordering temperature, where fractional excitations are emergent, revealed by spin-spin correlations restricted to nearest neighbor sites, among other factors Ravelli .
Older and newer studies on Na2IrO3 have been limited to probes that require bulk crystals, where the insulating character of Na2IrO3 has hampered its integration to an electronic device, advantageous for the study and manipulation of quasiparticle excitations, of interest in topological quantum computingKitaev ; Kasahara . Here we report electronic transport measurements on thin crystals of Na2IrO3, where electronic transport is ruled by variable range hopping at temperatures above the magnetic ordering transition known for bulk Na2IrO3 ( K) and electric field assisted above a critical applied field. Fits of our data to these two mechanisms allow us to deduce the localization length as well as the density of states at the Fermi level in our Na2IrO3 devices. Our Angle Resolved Photoemission Spectroscopy (ARPES) measurements show a non-vanisihing density of states at the Fermi level, consistent with our electronic transport experiments.
Bulk crystals of Na2IrO3 were grown mixing elemental Ir ( purity, BASF) with Na2CO3 ( purity, Alfa-Aesar) in a molar ratio. The mixture was ground for several minutes and pressed into a pellet at approximately psi. The pellet was subsequently warmed in a furnace to and held at this temperature for 48 hours, before being cooled to over 24 hours and finally furnace-cooled. Single crystals of more than one square millimeter were collected from the surface of the pellet. Crystals were exfoliated onto a Si/SiO2(280 nm) substrate and immediately coated with e-beam resist, minimizing the exposure to air.
Some wafers with exfoliated crystals were left uncoated and promptly analyzed through Raman spectroscopy. Raman scans were made over a range of 56 to 800 . Three Raman active modes were observed at 454 , 484 and 561 (Figure 1a) consistent with the reported first-order Raman modes of bulk sodium iridate Gupta . Exfoliated crystals were about 100 nm thick, measured through atomic force microscopy (Figure 1b and c). Electrodes were patterned on the ebeam resist-coated wafers using standard electron beam lithography, previous to the deposition of Ti(4 nm)/Au(140 nm) using electron beam evaporation.
Current-voltage (IV) characteristics were measured (two probe) in a closed cycle cryostat (Oxford Instruments) from 300 K down to 1.5 K. As shown in Figure 2a, at high temperatures the IV curves show an ohmic-like behavior. As the temperature is lowered, the slope of the IV curves decreases indicating an increase of the device’s resistance. At sufficient low temperatures, the IV characteristics become non-linear (Figure 2b). This behavior is consistent with the known Mott-insulating character of Na2IrO3. Previous numerical works that consider both Coulomb interactions and spin-orbit coupling Comin have been able to explain the observed insulating gap in Na2IrO3Comin ; Sohn considering a Coulomb repulsion U (3eV) that is larger than the effective bandwidth ( 1eV), supporting the idea of Na2IrO3 being a spin-orbit assisted Mott insulator.
Being an accurate measurement of the conductance of the sample at zero bias non-accessible at low temperatures, we tracked the change of the current across the sample at non-zero bias as a function of the temperature in a range of 3K-120K (see Fig. 3), and compared it to the thermal activation theory, characterized by an Arrhenius law,
(1) |
where EA is the activation energy and is Boltzmann constant. Figure 3a shows versus on a semi-log plot for different bias voltages. While the data fits in small range of temperatures to a thermal activated behavior (at around 120K), it follows in a more extended range a three-dimensional variable range hopping (VRH) mechanism Mott (-120K), as shown in Figure 3b. Mott’s variable range hopping (VRH) law considers 3-dimensional hopping between remote sites whose energy levels happen to be close to the Fermi level , as spatial neighboring sites have larger resistances BookES9 .
Typically, the hopping length increases with lowering temperature, therefore the name of variable range hopping BookES9 . In general, Mott’s VRH is described by
(2) |
with for three-dimensional VRH and the energy scale related to localization length of the charge carriers,
(3) |
with the localization length and the density of states at the Fermi level. A fit of the data at low bias (50mV) to three-dimensional VRH yields (see Fig. 3b), similar to the reported values for heteroepitaxial Na2IrO3 thin films Jenderka .
Inspecting more carefully the data in Fig. 3b at different bias voltages, two main effects are visible. First, there is a change in the trend of the curves near (grey line in Figures 3a and b), corresponding to the critical temperature of the long range antiferromagnetically ordered state reported for bulk crystals after magnetic susceptibility and heat capacity measurements Singh2010 ; Liu2011 , neutron and x-ray diffractionYe . The onset of long range magnetic ordering seems here to have an effect on the charge transport mechanism in the thin crystals of Na2IrO3, possibly suppressing VRH below the transition temperature. A similar phenomenon has been observed in the temperature dependence of the resistivity of heteroepitaxially grown thin films of Na2IrO3 Jenderka . Also, non-equilibrium optical measurements have shown signature of a long range ordered state in photoinduced reflectance measurements Hinton . Second, as the bias voltage increases, vs deviates from a linear dependence, indicating also a divergence from a VRH mechanism.
We can understand effect 1 by taking a closer look to VRH. Following the model by Miller and Abrahams BookES4 , the resistivity resulting from hopping over sites ij is given by:
(4) |
where is the spatial distance to the nearest neighboring site and correspond to the difference of the hopping sites energies and lying in a narrow band near the Fermi energy, whose width decreases with temperature and depends on the density of states at the Fermi level go and localization length ,
(5) |
Hopping between sites depends on both the spatial and energetic separation of the sites. If (with the average spatial distance to the nearest neighbor empty size), the conduction mechanism is reduced to nearest neighbor hopping. If on the other hand is of the order or less than unity, the second term in equation 4 contributes the most to the resistance and hops to sites that are further away in space but closer in energy are favored. Hopping between sites is in this case variable range, which corresponds to the transport mechanism in our thin crystals of Na2IrO3 above the magnetic ordering transition. In fact, Mott’s VRH relation (equation 2) is derived by finding the energy from equation 4 at which the resistivity is minimum. In the case of a magnetic ordered state, previous works have introduced an extension of Mott’s VRH model in which the hopping energy has an additional term related to the relative magnetization of the hopping sites Wagner . It is however unclear if associating a cost on the hopping energy based on the spin orientation of the hopping sites is valid, as carriers have tendency to align their spin to the localized spin CommentWagner ; DeGennes . Based on our data, we believe that electronic transport in our Na2IrO3 devices in the magnetic ordered regime goes beyond Mott’s VRH model.
Effect 2 caused by the bias voltage, can be unveiled by adding the effect of an electric field to equation 4, obtaining
(6) |
At high electric field strengths, where , hopping in the direction of the electric field will compensate the difference energy between hopping sites, . Therefore hops between sites that are closer in energy are no longer energetically favorable and the first term in equation 6 dominates. The average distance between hopping sites is expected to satisfy . Taking (where is the density of states at the Fermi level) the average distance between hopping sites becomes giving rise to a field assisted transport mechanism Shklovskii
(7) |
with the characteristic localization field.
(8) |
Above a critical external electric field, carrier hopping is therefore mediated by the electric field and there is no longer a clear VRH mechanism, as observed for the highest fields in Figure 3b. In fact, there is an electric field assisted transport at high fields that can be observed in the data, which is accentuated at lower temperatures as is shown in Figure 4, where IV curves at different temperatures are represented following relation 7. Below a critical electric field, transport is no longer field assisted. A fit of the data represented in Fig.4 to equation 7 allow us to extract . In particular, for 20K we obtain .
From the definitions of (equation 8) and (equation 3), we are able to deduce a localization length and a density of states in our Na2IrO3 thin crystal devices. Localization length is about 10 times larger than the one reported for heteroepitaxial thin films, calculated using a density of states estimated for other iridates Jenderka . Our analysis provides a value for both the localization length and the density of states in Na2IrO3 thin crystals deduced from the same set of data.
Mott’s VRH requires a non-vanishing density of states at the Fermi level, as hopping between sites are concentrated in a band near the Fermi energy with a bandwidth (eqn. 5). Our ARPES measurements confirm this in our samples.
ARPES measurements were performed at Beamline 4.0.3 (MERLIN) at the advanced light source using eV photons. Vacuum was better than Torr. The total energy resolution was meV with angular resolution (). Data was taken at 290K to avoid charging of the sample. A fresh crystal of Na2IrO3 (as those used for the electronic device fabrication) was mounted and cleaved in-situ the ARPES chamber. Figure 5a shows the measured electronic band structure along the high symmetry direction -M with a photon energy of 90eV. A feature near the Fermi energy is visible, as reported in previous ARPES works. Initially, this in-gap feature was identified as a metallic surface state Alidoust but later on it was found through spatially resolved-ARPES measurements Moreschini that it was consequence of quasiparticle formation in one of the possible cleavage planes of Na2IrO3. Indeed, the edge sharing IrO6 octahedra in Na2IrO3 form a layered stacking alternating with pure Na layers, which gives rise to both Na and Ir-O terminated surfaces after cleavage. In the former, the Ir-O octahedra being embedded between two Na layers, create a charge transfer from Na to Ir giving rise to a quasiparticle at the Fermi level Moreschini . This band formed at one of the possible terminations in Na2IrO3 after cleavage (the Na-terminated) is what we believe mediates the VRH mechanism in our Na2IrO3 devices. In the frame of Mott’s VRH, the width of the band responsible for conduction (eqn. 5) at 290K, corresponding to the ARPES measurements, is 90meV, which falls within the quasiparticle observed in Fig. 5. Fig. 5b shows energy distribution curves (EDC) in correspondance of the quasiparticle (in black, at ) and at a larger momentum outisde (in red) that testifies a non-vanishing density of states at the Fermi level, as required by Mott’s VRH.
After cleavage, the Na and O terminated surfaces are approximately 10-40 large, as demonstrated through ARPES Moreschini . Our ARPES data being taken with a spotsize of , captures an average of the two terminations. However, the spectral weight measured in the vicinity of the Fermi level is to be attributed to the Na-terminated surface only, since the Ir-O surface has no states in that energy rangeMoreschini . This does not invalidate any of our conclusions and in fact supports them, since this density of states is also representative of the bulk, which we probe with our transport measurements. The band structure characteristic of the Ir-O cleave that lacks of states near the Fermi energy, is instead an ”anomaly” given the absence of the Na layer which causes a difference charge balance between the two topmost layers of the cleave Moreschini . Therefore, the electronic structure as measured by ARPES, both here and in refs. Alidoust, and Moreschini, , does show a small density of states within a small range from the Fermi level.
In conclusion, we have integrated a thin crystal of Na2IrO3 into an electronic device, presenting the same Raman active modes as for the bulk crystals. We have observed that the main transport mechanism at temperatures above the formation of an antiferromagnetic long range order (15K) is VRH. This observation is supported by ARPES measurements at the high temperature end (290K) that testify a non-vanishing density of states at the Fermi level as required by Mott’s VRH. We have observed that as the system approaches long range order, it deviates from VRH with no evidence of this mechanism below the ordering temperature. Similarly, we have found that in the presence of an external electric field, transport diverges from VRH, becoming field assisted. Contrasting our data to the Mott’s VRH and electric field assisted models allow us to deduce a localization length of and a density of states of /eVm3 in our Na2IrO3 thin crystal devices. Our work constitutes a first approach to integrate an exfoliated thin crystal of Na2IrO3 into an electronic device, where separate ARPES measurements inform the electronic transport experiments. Using these two experimental techniques independently on similarly prepared samples can unveil important properties in other layered materials.






Acknowledgements.
The primary funding for this work was provided by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences under contract DE-SC0018154. The Advanced Light Source is supported by the Director, Office of Science, Office of Basic Energy Sciences, of the U.S. Department of Energy (U.S. DOE-BES) under contract no. DE-AC02-05CH11231. Work by J.G.A., G.L., V.N. and N.P.B. was supported by the Department of Energy Early Career Program, Office of Basic Energy Sciences, Materials Sciences and Engineering Division, under Contract No. DE-AC02-05CH11231 for crystal growth. P. S and H. T. were supported by the National Institute of General Medical Sciences of the National Institutes of Health under Award Numbers UL1GM118979, TL4GM118980, and RL5GM118978 for Raman measurements. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. We would like to acknowledge Luca Moreschini for valuable advice on the ARPES data as well as Jonathan Denlinger for assistance during the ARPES experiments.References
- (1) Sae Hwan Chun, Jong-Woo Kim, Jungho Kim, H. Zheng, Constantinos C. Stoumpos, C. D. Malliakas, J. F. Mitchell, Kavita Mehlawat, Yogesh Singh, Y. Choi, T. Gog, A. Al-Zein, M. Moretti Sala, M. Krisch, J. Chaloupka, G. Jackeli, G. Khaliullin and B. J. Kim, Nat. Phys. 11, 462 (2015).
- (2) Yogesh Singh, S. Manni, J. Reuther, T. Berlijn, R. Thomale, W. Ku, S. Trebst, and P. Gegenwart, Phys. Rev. Lett. 108, 127203 (2012).
- (3) Stephen M Winter, Alexander A Tsirlin , Maria Daghofer, Jeroen van den Brink, Yogesh Singh, Philipp Gegenwart and Roser Valentí, J. Phys.: Condens. Matter 29, 493002 (2017).
- (4) Kaige Hu, Zhimou Zhou, Yi-Wen Wei, Chao-Kai Li, and Ji Feng, Phys. Rev. B 98, 100103(R) (2018).
- (5) Alexei Kitaev Annals of Physics 321, 2 (2006).
- (6) Y. Kasahara, T. Ohnishi, Y. Mizukami, O. Tanaka, Sixiao Ma, K. Sugii, N. Kurita, H. Tanaka, J. Nasu, Y. Motome, T. Shibauchi and Y. Matsuda, Nature 559, 227 (2018).
- (7) Xiaoxiang Xi, Xiangyan Bo, X. S. Xu, P. P. Kong, Z. Liu, X. G. Hong, C. Q. Jin, G. Cao, Xiangang Wan and G. L. Carr Phys. Rev. B 98, 125117 (2018).
- (8) Sitikantha D. Das, Sarbajaya Kundu, Zengwei Zhu, Eundeok Mun, Ross D. McDonald, Gang Li, Luis Balicas, Alix McCollam, Gang Cao, Jeffrey G. Rau, Hae-Young Kee, Vikram Tripathi, and Suchitra E. Sebastian, Phys. Rev. B 99, 081101(R) (2019).
- (9) A. Revelli, M. Moretti Sala, G. Monaco, C. Hickey, P. Becker, F. Freund, A. Jesche, P. Gegenwart, T. Eschmann, F. L. Buessen, S. Trebst, P. H. M. van Loosdrecht, J. van den Brink and M. Gruninger, ArXiv: 1905.13590
- (10) Satyendra Nath Gupta, P. V. Sriluckshmy, Kavita Mehlawat, Ashiwini Balodhi, Dileep K. Mishra, S. R. Hassan, T. V. Ramakrishnan, D. V. S. Muthu, Yogesh Singh and A. K. Sood EPL 114, 47004 (2016).
- (11) R. Comin, G. Levy, B. Ludbrook, Z.-H. Zhu, C. N. Veenstra, J. A. Rosen, Yogesh Singh, P. Gegenwart, D. Stricker, J. N. Hancock, D. van der Marel, I. S. Elfimov and A. Damascelli, Phys. Rev. Lett. 109, 266406 (2012).
- (12) N. F. Mott, Phil. Mag. 8, 835 (2006).
- (13) Yogesh Singh and P. Gegenwart, Phys. Rev. B 82, 064412 (2010).
- (14) X. Liu, T. Berlijn, W.-G. Yin, W. Ku, A. Tsvelik, Young-June Kim, H. Gretarsson, Yogesh Singh, P. Gegenwart and J. P. Hill, Phys. Rev. B 83, 220403R (2011).
- (15) Feng Ye, Songxue Chi, Huibo Cao, Bryan C. Chakoumakos, Jaime A. Fernandez-Baca, Radu Custelcean, T. F. Qi, O. B. Korneta, and G. Cao, Phys. Rev. B 85, 180403R (2012).
- (16) Marcus Jenderka, José Barzola-Quiquia, Zhipeng Zhang, Heiko Frenzel, Marius Grundmann and Michael Lorenz, Phys. Rev. B 88, 045111 (2013).
- (17) J. P. Hinton, Patankar, E. Thewalt, A. Ruiz, G. Lopez, N. Breznay, A. Vishwanath, J. Analytis, J. Orenstein, J. D. Koralek and I. Kimchi, Phys. Rev. B 92, 115154 (2015).
- (18) B.I. Shklovskii and A.L. Efros, ”Electronic Properties of Doped Semiconductors” (Springer-Verlag, Berlin Heidelberg, 1984) Chap. 9, pp. 202-210.
- (19) B.I. Shklovskii and A.L. Efros, ”Electronic Properties of Doped Semiconductors” (Springer-Verlag, Berlin Heidelberg, 1984) Chap. 10, pp. 228-238.
- (20) V.Yu. Butko, J. F. DiTusa, and P.W. Adams, Phys. Rev. B 84, 1543 (2000).
- (21) A. L. Efros and B. I. Shklovskii, J. Phys. C: Solid State Phys. 8, L49 (1975).
- (22) S. K. Choi, R. Coldea, A. N. Kolmogorov, T. Lancaster, I. I. Mazin, S. J. Blundell, P. G. Radaelli, Y. Singh, P. Gegenwart, K. R. Choi, S.-W. Cheong, P. J. Baker, C. Stock, and J. Taylor, Phys. Rev. Lett. 108, 127204 (2012).
- (23) F. Ye, S. Chi, H. Cao, B. C. Chakoumakos, J. A. Fernandez-Baca, R. Custelcean, T. F. Qi, O. B. Korneta, and G. Cao, Phys. Rev. B 85, 180403 (2012).
- (24) B.I. Shklovskii and A.L. Efros, ”Electronic Properties of Doped Semiconductors” (Springer-Verlag, Berlin Heidelberg, 1984) Chap. 4, pp. 74-93.
- (25) P. Wagner, I. Gordon, L. Trappeniers, J. Vanacken, F. Herlach, V. V. Moshchalkov, and Y. Bruynseraede, Phys. Rev. B 81, 3980 (1998).
- (26) Sudhakar Yarlagadda, Phys. Rev. Lett. 84 4017 (2000).
- (27) P.G. de Gennes, Phys. Rev. 118, 141 (1960).
- (28) B.I Shklovskii, Sov. Phys. Semicond. 6, 1964-1967 (1973).
- (29) S. Chikara, O. Korneta, W. P. Crummett, L. E. DeLong, P. Schlottmann and G. Cao, J. Appl. Phys. 107 09D910 (2010).
- (30) Nasser Alidoust, Chang Liu, Su-Yang Xu, Ilya Belopolski, Tongfei Qi, Minggang Zeng, Daniel S. Sanchez, Hao Zheng, Guang Bian, Madhab Neupane, Yu-Tzu Liu, Stephen D. Wilson, Hsin Lin, Arun Bansil, Gang Cao, and M. Zahid Hasan, Phys. Rev. B 93, 245132 (2016).
- (31) L. Moreschini, I. Lo Vecchio, N. P. Breznay, S. Moser, S. Ulstrup, R. Koch, J. Wirjo, C. Jozwiak, K. S. Kim, E. Rotenberg, A. Bostwick, J. G. Analytis, and A. Lanzara Phys. Rev. B 96, 161116(R) (2017).
- (32) C. H. Sohn, H.-S. Kim, T. F. Qi, D. W. Jeong, H. J. Park, H. K. Yoo, H. H. Kim, J.-Y. Kim, T. D. Kang, Deok-Yong Cho, G. Cao, J. Yu, S. J. Moon and T. W. Noh, Phys. Rev. B 88, 085125 (2013).