Competing orders in strongly correlated systems studied by transport measurements

Competing orders in strongly correlated systems lead to rich phase diagrams comprising many electronic phases, such as superconductivity, charge/spin density wave, charge order, or bad metallicity. These phases are generically sensitive to a variety of parameters, for example temperature, magnetic field, dimensionality, presence of disorder, geometrical frustration. In this thesis, we employ electronic transport measurements under high pressure on few model compounds to gain insight into the complex physics of strongly correlated compounds. The transport coefficients, resistivity and thermoelectric power, shed light onto conduction processes and the thermodynamics. The pressure is a perfect tool to investigate competition of different ground states: by modifying the lattice parameters, it can tune the interactions without introducing disorder. Several representative compounds were chosen for this study. In the first part, we focus on the transport properties of the quasi-one dimensional BaVS3. The main characteristic of this 3d1 system is the coexistence of a broad one-dimensional dz2 electronic band and a narrow isotropic eg band at the Fermi level. The suppression of the insulating phase by high pressure leads to a non-Fermi liquid phase. We showed that magnetic field does not recover the Fermi liquid behavior, and that the disorder pushes the system further into non-Fermi liquid state. This is at variance with what has been observed in other non-Fermi liquid compounds, and confirms the novelty of the mechanism for non-Fermi liquid behavior in BaVS3. To achieve better understanding of the role of the localized electrons, we investigated systematically the influence of disorder. In addition, we studied the properties of the BaVSe3, which due to the reinforced interchain interactions may be considered as the high-pressure counterpart of BaVS3. The system is a metallic ferromagnet, in which the strong interaction of dz2 and eg electrons dictates the behavior of transport coefficients. In the following part we studied the rich physics of quasi-one dimensional β-vanadium bronzes. In the stoichiometric β-SrV6O15, we followed the pressure dependence of the semiconductor-insulator transition by resistivity and thermopower. We found evidence suggesting that the ground state is charge ordered. Under pressure, the changing character of the transport coefficients implied a competition of different ground states. Moreover, we observed resistive switching in the insulating phase. When strontium doping is decreased, in SrxV6O15 and x < 1, the disorder starts governing the physics of the system. The off-stoichiometric compounds are characterized by the absence of phase transition, absence of resistive switching, and possibly by the presence of polarons. We also found resistive switching in another charge ordered transition-metal oxide, Fe2OBO3. This system shows an interplay of commensurate and incommensurate charge order. The switching is restricted to the incommensurate phase, whose origin probably lies in the geometrical frustration of the interactions between iron atoms. With pressure we enhance the Coulomb repulsion, and the incommensurate phase shrinks in temperature. In the final part, we address the high-pressure transport of a superconductor on a geometrically frustrated pyrochlore lattice, ΚOs2O6. The potassium atoms are enclosed in oversized cages and their rattling motion introduces a localized low-energy mode. The transport coefficients in this compound are highly anomalous: the resistivity shows no saturation at low temperatures, and the scythe-shaped thermoelectric power is reminiscent of the one observed in cuprates. We were able to reproduce the temperature and pressure dependence of the transport coefficients within a simple model of the density of states.

Thermopower as a Unique Tool to Probe Quantum Phase Transitions

Stevan Arsenijevic

In my thesis, transport measurements such as resistivity and, more importantly, thermopower S, were used to explore the phase diagram of bad metals. Bad metals are electronically correlated systems whose ground state lies close to a quantum phase transition. By tuning the control parameters, such as temperature (T ), magnetic field (B), hydrostatic pressure (p) or chemical substitution (x), we can induce phase transitions between the various electronic, magnetic and structural phases. Here, the thermopower is presented as a unique tool for probing quantum phase transition because it is a measure of the entropy of conducting electrons. The main part of the thesis is dedicated to the study of Fe-based superconductors (FeSC) discovered in 2008. Their parent compound has an antiferromagnetic (AF) ground state, where the itinerant electrons form a spin-density wave (SDW), a periodic modulation of spin density. This coincides or is preceded by a structural, tetragonal-to-orthorhombic transition. The nesting between the electron and hole Fermi surface is believed to be the driving mechanism for the SDW state. By changing the structural or chemical properties the AF ground state of FeSC is suppressed, giving way to superconductivity (SC). The remaining antiferromagnetic fluctuations above the transition can provide a glue for SC pairing. Here, the analysis of the thermopower S/T of BaFe1−xCoxAs2 (BFCA) in the x-T phase diagram shows the signatures of the spin fluctuation which have a dome-like dependence and follow the trend of superconducting Tc . The logarithmic increase of S/T upon decreasing T is ascribed to the proximity of the spin-density-wave quantum critical point. It can be understood as an increase of entropy due to the incommensurate AF spin fluctuations. We can ascribe the high values of thermopower in BFCA at intermediate- and room-temperatures to the influence of low-T quantum criticality. To probe the response of the electronic system in FeSC to structural changes, we performed measurements under pressure of the parent compound BaFe2As2 (BFA), the SC electron-doped BFCA, and hole-doped Ba1−xKxFe2As2 (BKFA). In the parent compound pressure suppresses the structural/SDW transition, similar to the effect of doping. For doped systems, in order to describe the behavior of thermopower in the high-T range (above 100K) we used a semi-metallic two-band model which was fitted to the data in order to extract the pressure dependence of the band parameters. In both doping cases the effect of pressure was similar, an increase of the band overlap and of the effective number of charge carriers. With this model we can explain the high-T , x and p dependence of thermopower in both electron- and hole-doped BFA. In a structurally simpler Fe-chalcogenide Fe1+yTe1−xSex compound, the excess of Fe has a Kondo-like influence on the charge carriers which dramatically changes the physics of the normal state. To probe the normal state, pressure, doping, magnetic Fe-excess concentration (y) and temperature were used as control parameters. At low-T a characteristic upturn of resistivity (ρmag ) is observed, followed by an increase of thermopower (Smag ), which we identify as the magnetic contribution caused by the spin-flip scattering events. Increasing the y resulted in an increase of ρmag , and a decrease of Smag , which is in agreement with the behavior of canonical Kondo-systems. Pressure suppresses the magnetic contribution to transport, thus increasing the itinerancy of the system. MnSi is another system in which the sensitivity of thermopower to entropy brings new information related to the complex magnetic structure. Pressure was used to drive the system from a helically ordered, canonical Fermi-liquid (FL) phase with T 2-resistivity to the intrinsically disordered, non-Fermi-liquid (NFL) phase above pc with T3/2-dependence. Our contour plot of S/T demonstrated how powerful the thermopower technique is, by reproducing the whole previously established T -p phase diagram. At the phase transition from the magnetically-ordered FL phase to the disordered NFL, the thermopower is dramatically enhanced. We bring useful information about the mysterious partial order (PO) phase inside the NFL phase, previously detected only by neutron scattering. The fluctuating helices scenario can describe the observed increase of entropy/thermopower in the PO phase. At ambient pressure, close to the helical transition of MnSi, a moderate magnetic field can stabilize the skyrmion lattice - the lattice of topological magnetic whirls, vortices. We observe a signature of the skyrmion lattice as a minute drop in thermopower. It is located exactly in the same region of the T − B phase diagram where an increase in magnetoresistance and Hall effect was reported previously. This feature originates from the additional scattering of conducting electrons on magnetic vortices, while the change in S is dominated by the decrease of entropy as the stable skyrmion lattice is formed. Overall, resistivity was used to confirm the established phase diagram, while thermopower, as an interesting and not sufficiently understood technique, was used to probe the sensitive changes of the charge carriers at the Fermi surface. We explored various phases showing how useful thermopower is to probe the entropy of electronic system on the verge of quantum phase transition.

EPFL2012

Study of novel electronic conductors

Neven Barisic

This thesis presents the results of a concerted effort to understand the complex array of physical properties exhibited by the BaVS3 family of materials. As a 3d1 system, BaVS3 displays a unique collection of correlation-driven phenomena, including a metal-insulator transition driven by spin-density waves and or charge-density waves as well as a pressure-dependent crossover between the non-Fermi-liquid and Fermi-liquid behaviors. In an attempt to better understand these and many other properties, we have undertaken a systematic experimental study of BaVS3 and related compounds. The primary measurements carried out were those of the transport properties, resistivity and thermoelectric power (TEP). Through the construction of specialized measuring apparatuses, we were able to measure simultaneously these transport properties under conditions of variable temperature (from 2 to 600 K), pressure (up to 3 GPa) and magnetic field (up to 12 T). At ambient pressure and in the range of 250 to 600K, BaVS3 shows nearly isotropic but poor metallic behavior with linear temperature dependences of resistivity and TEP and a Curie-like magnetic susceptibility. The nearly isotropic conductivity contrasts with the 1D 2kF fluctuations observed by lowering the temperature below 250 K (at which the first Jahn-Teller structural phase transition occurs) deep in the metallic phase. The fluctuations reveal the 1D aspects of the electronic character, originating from the chain like structure of the material. The salient feature of BaVS3 at ambient pressure is the second order metal-to-insulator (MI) transition at TMI = 69 K, accompanied by the tetramerization (doubling of the 2V unit cell in the chain direction). In addition to the transport measurements, the strong changes in the electrical and magnetic properties of the system around TMI were followed by magnetic susceptibility, angle-resolved photoelectron spectroscopy and frequency-dependent conductivity. By increasing the pressure, the three-dimensionality of BaVS3 is enhanced and the MI phase transition is suppressed to lower temperatures. TEP and magneto TEP measurements in this pressure range revealed the existence of polarons and of spin fluctuations in the metallic phase. The latter can be identified as precursors to the MI transition. Around 1.8 GPa, where TMI ~15 K, the system enters a strongly fluctuating regime, highly sensitive to the magnetic field, the amplitude and the frequency of the measuring current and to a further increase of the pressure. Closely related to these features, the phase boundary collapses and a hysteretic behavior appears in the transport properties and their magnetic-field-dependent counterparts. At a critical pressure of ~2 GPa, a non-Fermi liquid (NFL) state arises (with n~1.5 in the Tn resistivity law between 1 and 15 K) in relation to the proximate Quantum Critical Point. The p-H-T phase diagram of BaVS3 in this region has been explored in some detail with particular emphasis placed upon the relevance of the spin degrees of freedom (on the insulator side) and the role of quantum fluctuations above the critical pressure. Finally, as the pressure is increased further, the conventional Fermi-liquid exponent of n = 2 is obtained. In order to delve further into the subtleties of the MI transition and the NFL behavior, a comparative study was carried out on the sister compounds Bax-1SrxVS3 and BaVS3 and on sulfur-deficient BaVS3. From this study it became apparent that the imposed chemical substitution manifests itself as an additional effective pressure. The inherent disorder of these samples was observed to affect the NFL behavior by reducing the exponent n towards a value of 1, while the apparent ferromagnetic order below 15 K was seemingly independent of pressure. All the observed features are consistent with a relatively simple two-band tight-binding, vanadium-based, model consisting of a wide quasi one-dimensional band hybridized with a rather isotropic narrow band. Within this context, it is concluded that the MI transition is controlled by the Coulomb-interacting electrons of the wide quasi one-dimensional band. Based on this model a new description of the NFL regime has been proposed.

EPFL2005

Normal state transport properties of novel superconductors

Balazs Sipos

The discovery of high temperature superconductivity in the cuprates in 1986 has boosted the research in strongly correlated materials. One strong motivation was and stays the understanding the high-Tc phenomenon with the hope that one can ultimately engineer new materials with even higher Tc. Besides the in-depth investigation of cuprates, there is a strong tendency in the solid state community to find new superconductors, which by themselves are interesting for applications, or by their properties they can contribute to the understanding of the high-Tc phenomenon. The program of my doctoral thesis was three-fold: i) to address one important issue in the cuprate superconductors, that of the role of homogeneity in the underdoped part of the phase diagram; ii) what is the effect of disorder in MgB2 superconductor, which has high potentials for applications; iii) to discover new superconductors in the family of transition metal dichalcogenides. All these materials are in some sense unconventional superconductors. The cuprates by their high Tc and the symmetry of the order parameter, MgB2 by its two-band superconductivity and Tc of 39 K, and the dichalcogenides by the appearance of superconductivity on the background of competing interactions. Measurements of transport properties, such as resistivity and thermoelectric power, were used to get insight in the behavior of these materials. Besides temperature as variable, I applied high pressure, extreme magnetic fields and controlled disorder introduced by fast electron irradiation. In the first part I present the pressure dependent study of two members of the transition metal dichalcogenides having 1T structure, 1T-TiSe2 and 1T-TaS2, where superconductivity was never observed in a pristine sample. 1T-TiSe2 has a CDW phase below 220 K which origin, weather it is driven by an excitonic mechanism or by a Jahn-Teller distortion, is an ongoing question. By applying pressure I showed that the pristine sample is superconducting in the pressure range of 2.0–4.0 GPa. This range remarkably coincides with the short range fluctuating CDW before its disappearance at the upper pressure value. If CDW is due to excitonic interactions than our observations suggest that it can be at the origin of superconductivity, as well. The second dichalcogenide is the 1T-TaS2, where a Mott-insulator phase appears on the top of a commensurate CDW. By applying pressure I was able to melt that Mott-phase, and reveal that the material is superconducting above 2.5 GPa with Tc of 5.9 K. Unexpectedly, superconductivity is born from a nonmetallic phase, and stays remarkably stable up to the highest applied pressure of 24 GPa. In the second part I tried to give my contribution to the field of high-Tc superconductors. I carefully selected few high quality underdoped Bi2Sr2PrxCa1-xCu2O8-δ sample, to address the nature of the low temperature ground state by applying high magnetic field. Although former measurements by other groups showed that at high underdoping, the ground state is an insulator, I found that a sample with as low Tc as 15 K exhibits metallic behavior up to 60 T. Furthermore, I showed that a inhomogeneous distribution of the doping atoms can completely mask the intrinsic normal state of a high-Tc superconductor. In the last part of my thesis I focused on the two-band superconductor MgB2 by studying the scattering between the bands by the means of the Matthiessen's rule. I made a systematic study of the influence of defects created by fast electron irradiation, and found that the the Matthiessen's rule is satisfied for the defect concentration range I induced. I further compare the influence of defects on the critical temperature and the residual resistivity in MgB2 with superconductors with various order parameters, and found that the decrease-rate of Tc in our system is within the range of a response of a superconductor with an s-wave order parameter.

EPFL2008