Metal-insulator transitions in transition metal chalcogenides and oxides investigated by electron spectroscopies

Abstract

In conclusion, we have investigated the electronic structure of hexagonal and millerite phases of NiS, using various electron spectroscopic techniques to elucidate the underlying electronic structure. Using high-resolution photoemission spectroscopy, we establish that the low-temperature phase of hexagonal NiS is an antiferromagnetic metallic state with the first-order phase transition observed in NiS and related compounds arising due to a metal-to-anomalous-metal transition. However, further studies are required to understand the unconventional transport and magnetic properties associated with the low-temperature phase. The understanding of the similarities in the electronic structure of itinerant antiferromagnetic Cr and the low-temperature phase of NiS might be helpful in further understanding the anomalous metallic phase of NiS. Our spectroscopic studies indicate that the substitution of Se in place of S does not show appreciable changes in the electronic structure, indicating similar electronic parameters for the solid solution, NiS???Se?. The band structure results do not describe accurately the details of the spectral features in the paramagnetic phase of the system, while a parameterized many-body multi-band model including electron correlations is found to be successful in describing the core level and valence band spectrum of NiS. The core and valence band spectral calculations of the paramagnetic metallic phase of NiS show that it is a correlated metal with a highly covalent character. The asymmetric line shape as well as parts of the satellite features in the core level spectra have been ascribed to extrinsic loss processes in the system. The electronic parameter strengths obtained for the compounds of this solid solution show that they belong to the pd metallic regime of the ZSA phase diagram and are away from the MIT boundary of the phase diagram. The valence band calculations show that the lower Hubbard band exists well inside the pd metallic regime, as predicted by the DMFT calculations. The millerite phase of NiS has been studied by means of various electron spectroscopic techniques and band structure calculations as well as model Hamiltonian calculations. The band structure calculations were found to be more successful in describing the experimental valence band spectrum in comparison to the case of hexagonal NiS, possibly suggesting a reduced effect of correlation in the millerite phase. This is consistent with the highly conducting ground state of the millerite phase, in contrast to the antiferromagnetic and poorly conducting ground state of hexagonal NiS. However, calculations including the electron correlation effects are found to improve the description of the experimental spectra, indicating the importance of correlation effects even for such a highly metallic system. The various electronic parameter strengths obtained from these calculations indicate that the millerite phase of NiS is a highly covalent metal (pd-metal). Thus, the study of the hexagonal and millerite forms of NiS provides us with the information concerning the evolution of the spectral function in a pd-metal as a function of covalency. We have investigated the electronic structure of the system, NiS???Se?, using various electron spectroscopic techniques to elucidate the underlying electronic structure. Analysis of the S 2p and Se 3d spectra indicate S and Se are in the stoichiometric ratio and form a randomly substituted solid solution even in the surface region. The band structure results do not describe well the details of the spectral features in the paramagnetic phase of the system, while a parameterized many-body multi-band model including electron correlations is found to be successful in describing the core level and valence band spectrum of NiS?. The core and valence band spectral calculations on the paramagnetic phase of NiS? show that it is a strongly correlated system with a highly covalent character. The estimated values of Ni d-S p (pda) hopping interaction strength, the charge transfer energy (A), and multiplet averaged Coulomb interaction strength (Udd) are -1.5 eV, 2.0 eV, and 4.0 eV, respectively. According to this, the compounds in the solid solution NiS???Se? are close to the regime of pd metals with U > A. In Section 4.4, we have reported the evolution of the electronic structure across the insulator-to-metal transition controlled by changes in the bandwidth at a fixed filling in NiS???Se?. Comparison of LDA band structure results with the experimentally obtained electronic structure allows us to identify the various spectral features, establishing the importance of electron correlation effects in determining the spectral features close to EF. We find that the spectral signature of the upper Hubbard band persists for all compositions well into the metallic regime, while the insulator-to-metal transition with decreasing U/W is signaled by a transfer of spectral weight from the upper Hubbard band to states at and near EF. These results are consistent with recent calculations [39] based on the dynamical mean-field treatment of the Hubbard model. In Section 4.5, we have presented the results of our high-resolution UV photoemission experiments probing the surface electronic structure of the bulk insulating compositions. The surfaces of these insulating compositions were found to be metallic at room temperature. At low temperatures, the surfaces of these systems undergo a metal-to-insulator transition as also suggested by the resistivity data. This provides us with an experimental system to study experimentally the 2-dimensional MIT in a dense electronic system. The evolution of the electronic structure across the 2-dimensional MIT is found to be qualitatively different than that from the 3-dimensional case. It is found that a pseudogap is progressively formed in the system, even above the transition temperature. At the lowest temperature, the spectral function is consistent with the formation of a band gap, though the gap region is filled with localized impurity states. The evolution of the electronic structure across the MIT in these 2-dimensional systems is found to be qualitatively different compared to that in the 3-dimensional cases. In this chapter, we have investigated the composition-dependent isostructural metal-insulator transition in BaCo???Ni? using electron spectroscopic techniques. The inability of the band theory to explain the experimental results illustrates the correlated nature of the system. The analysis of core level and valence band spectra has revealed that these compounds are highly covalent in nature. The estimated electronic structure parameters show that BaNiS? is a pd metal, while the decreasing hopping interaction strength, tpd, across the solid solution drives the system insulating for x < 0.2. Our observations are also consistent with the transport measurements under pressure on the insulating compositions of the system. This covalency-driven metal-insulator transition is found to be similar to that observed in the case of LaNiO? and NdNiO?. The disorder effects in this system are important and the possibility of a pseudogap opening up with localized states at EF across the transition cannot be ruled out. However, a detailed spectroscopic study using very high resolution is needed to address these issues.

In this chapter, we investigate the detailed electronic structure of Ce???Sr?TiO? with x = 0.0, 0.1, 0.2, 0.3, 0.4, and 0.7, and CeTiO??? with ? = 0.0, 0.1, and -0.15, as a function of hole and electron dopings, by means of various electron spectroscopic techniques as well as with photoemission calculations. Hole doping in insulating CeTiO? results in a metal-insulator transition (MIT); it is shown that the doped holes reside primarily at the Ti sites. It is observed that even for metallic compositions, the different charge species are not screened and give rise to separate signals in the experimental photoemission core level spectrum. This is found to be a general characteristic of the early transition metal oxides, and our core level photoemission calculations indicate that this effect is possibly due to the large charge transfer energy, A, present in these Mott-Hubbard systems. In electron-doped CeTiO?, the strong disorder present in the system localizes the electronic states close to the Fermi energy, turning the system into an insulator in spite of charge-carrier doping. It is found that the surface of the hole-doped metallic compounds is drastically different compared to the bulk. We extract the bulk and the surface electronic structures of one of the metallic compositions and show that the bulk electronic structure is in better agreement with theoretical predictions than the bare experimental spectrum. It also appears that surfaces of these systems are nearly insulating due to enhanced correlation effects. In this chapter, we describe the fabrication of an inverse photoemission spectrometer in our laboratory. The various electronic and mechanical components that have been fabricated for the spectrometer are described in detail. For the isochromat mode of operation, two different types of band pass detector designs based on the Geiger-Müller tube have been used. The first detector type is based on a CaF? window and helium-iodine mixture. This gives an overall instrumental resolution of 0.65 eV at a mean detection h? of 9.7 eV. It is found that this type of detector is difficult to work with as it corrodes the walls of the detector tube and the various ultra-high vacuum (UHV) components. The second type of detector uses a CaF? window and acetone vapors for detecting photons at about 9.9 eV. The achieved resolution is about 0.4 eV, however, with a lower intensity compared to the He I? detector. Using this detector, the experimental spectra for various metals have been recorded, and the recorded spectra are found to be in agreement with those in the literature. Finally, to demonstrate the applicability of the spectrometer in the study of strongly correlated systems, we have performed experiments on a transition metal oxide system, La???Sr?CoO?, that exhibits an insulator-to-metal transition as a function of hole doping. Across the transition, it is seen that the spectral intensity at E? increases, indicating the closing of the band gap of the insulating system. Future plans for further developments in this direction are also discussed.