| dc.description.abstract | In this final chapter, we shall summarize our important observations and suggest improvements and problems for further investigation. This research has been a learning experience for us as we had no prior experience in the area of point contacts and tunneling in our lab before this research was undertaken. The fact that a simple system like a point contact could find application in a variety of areas has been a revelation to us. We believe that, in our lab, we are now at a stage where experimentation in electron tunneling and allied areas should be extended to a more advanced level in the future.
The important contribution in Chapter 3 was to point out certain anomalous features of Andreev reflection observed in (N-S) point contacts even in conventional superconductors, like the rapid decrease of dI/dV for eV < ? and the 'dip' in conductance occurring at bias voltages corresponding to ?. This was extended in Chapter 4 where we demonstrated that quantum confinement effects play a role in 'smearing' the Andreev feature and causing it to totally disappear when the contact size is of the order of 0.1 ? (the coherence length).
In the context of oxide superconductors, we showed in Chapter 5 that barrierless (N-S) contacts showing Andreev reflection can be obtained in low-resistance point contact junctions of both YBCO and BSCCO single crystals. In our opinion, Andreev reflection has not been subjected to extensive study in the past and would be a worthwhile problem to look into. We suggest more experiments should be done on Andreev reflection with the N-S junctions in a magnetic field and at different temperatures.
In Chapter 4, we dwelt on the mesoscopic aspects of variable point contact junctions. We found that the conductance curves (dI/dV–V) of point contacts change slope (or change sign from negative to positive) signifying a transition from metallic to tunneling behavior before physical contact between tip and sample is lost. The change in sign of d²I/dV² happens for a number of metallic and superconducting junctions investigated. We argued that this signature of metallic-tunneling transition in dI/dV–V curves is due to confinement effects which cut off propagating modes for a < k < 2, resulting in domination of evanescent electron modes leading to an apparent 'tunneling' behavior of the dI/dV–V characteristics (i.e., d²I/dV² is positive). In superconductors, we pointed out the significance of additional length scale (?) and energy scale (?). With a CSTM one can extend the investigations to encompass the vacuum tunneling regime and also look at the problem the other way (viz., transition from vacuum tunneling to contact as the tip-sample distance is reduced). Though considerable advances have been made in theory and experiments on mesoscopic systems (mainly on lithographically patterned devices), we think that CSTM is a unique tool to study quantum confinement effects as it has the flexibility to span the entire range from tunneling to metallic contact, though with somewhat ill-defined junction geometry.
The highlight of our work on oxide superconductors presented in Chapter 5 was the demonstration of microshort to tunneling transition in both YBCO and BSCCO single crystals. The different types of dI/dV–V curves obtained cover most of the curves reported in literature using various geometries. The message of this study was to point out that the junction resistance plays a central role in determining the nature of the dI/dV–V curve and one should be careful in trying to interpret the tunneling spectra with point contact junctions based on measurements at one junction resistance. (For example, dI/dV–V curves of high-resistance junctions on oxide superconductors show no gap features and only a linear increasing background even at 4.2 K. One cannot conclude the system to be gapless as gap features are seen at lower junction resistances and Andreev reflection is seen at the lowest junction resistances). We discussed a set of experiments on BSCCO single crystal and showed the importance of cleaning and cleaving the surface. We established that multiple features in dI/dV–V curves are often not reproducible and associated with a dirty surface layer. In Chapter 5, we also presented a model to fit the tunneling data on high-Tc superconductors obtained by point contact method. Based on an earlier model by BTK, our model introduced a broadening parameter ? to the BCS density of states expression in the tunneling current and could successfully fit the data (particularly dI/dV at the gap voltage). We also presented the first tunneling measurements on Y-doped BSCCO system where the changes by a factor of 2.5 on Y-doping. Our superconducting gap measurements indicated that ? scales linearly for all Y concentrations. Further improvements would be to use CSTM and look for gap anisotropy in the layered high-Tc cuprates. We urge more studies on good quality single crystals of doped oxide superconductors as we feel that a study of the metal-insulator transition in these systems would provide insight into the normal state properties of oxide superconductors. Besides, interesting consequences due to interplay of disorder-induced localization and superconductivity could be probed in doped systems. A major part of our future work would be geared to look at transport and tunneling in several classes of doped oxide superconductors.
Chapter 6 of our study was based on the very important observation that the anomalous "linear" tunneling conductance observed in cuprates is not something specific to superconductivity. A strong correlation-induced feature exists in the tunneling conductance of oxide metals in general with a characteristic singularity at the Fermi level. We presented the first tunneling study on the sodium system close to the metal-insulator transition in Chapter 6, with the density of states due to electron-electron interaction seen in G(V) characteristics of crystals on the metallic side. A new aspect which emerged from our experiments is that the tunneling conductance near the metal-insulator transition is governed by properties in the critical region. We correlated the linear conductance seen in systems undergoing metal-insulator transition with the density of states from electron-electron interaction. This needs continued investigation. We conclude by saying that a number of problems have emerged from our study over the past six years presented in this thesis and hope these provide the impetus for further experimentation. | |
