The specific capacitance of several types of Josephson tunnel junctions has been measured by observing resonances in lightly-damped 2-junction interferometers. The capacitance was calculated using the resonance voltage, obtained by analyzing steps in the I-V characteristics, and the interferometer inductance, which was measured directly by the injection of a control current. Using this technique, the specific capacitance C s was determined for tunnel junctions with Pb-In-Au alloy base electrodes and Pb-Bi counterelectrodes, as well as for junctions made on Nb films with Pb-In-Au counterelectrodes. In both cases, barriers were produced by rf plasma oxidation. Junctions with Josephson current densities j 1 between 200 and 5000 A/cm2were investigated. It was found that 1/C s decreased with \log j_{1} , and that at j 1 = 1000 A/cm2, C s was 4.2 ±0.3 μF/cm2for the Pb-alloy junctions and 13.4 ±1 μF/cm2for the Nb junctions. These results are discussed in relation to available data on oxide thickness and dielectric constant.

Specific capacitance of Josephson tunnel junctions

Arrays of 10 or 40 series‐connected small densely packed Josephson tunnel junctions were successfully operated as parametric amplifiers at 10 GHz. Between 50 and 90% of the junctions in an array were observed to work coherently in the amplification process. Signal gain exceeding 24 dB was achieved with a pump power as low as −50 to −70 dBm. A power width (i.e., the range of the pump power for a gain within 3 dB of the peak gain) of 0.25 dB and a bandwidth of 15 MHz were measured for 20‐dB gain. The power width was in accordance with theory but the bandwidth was reduced, probably due to noise saturation effects. The latter effects can be reduced by lowering the input noise temperature or by including an external band limitation. The noise temperature was also increased by saturation; however, with an external bandwidth of 60 MHz, a noise temperature of 30±20 K was measured. The application of a magnetic field has proved extremely useful as a means of tuning the amplifiers to stable high‐gain operating poin...

Arrays of Josephson tunnel junctions as parametric amplifiers

Quantum interference effects between two Josephson tunnel junctions [1] incorporated in a superconducting ring were first observed by Jaklevic et al. [2] in 1964. These workers showed that the critical current of the double junction was an oscillatory function of the magnetic flux threading the ring, the period being the flux quantum, Φo. The implications of this result for instrumentation were quickly realized, and a variety of dc SQUIDS* (Superconducting Quantum Interference Devices) were developed and used. These devices included several designs involving machined pieces of niobium connected with point contact junctions [3–5],and the SLUG [6] (Superconducting Low-Inductance Undulatory Galvanometer), which consisted of a bead of solder frozen onto a niobium wire. In the late 1960’s, the rf SQUID [7–9] appeared. The rf SQUID consists of a single Josephson junction on a superconducting ring, and, presumably because only a single junction is required, has become much more widely used than the dc SQUID. Several commercial versions [10–13] of the rf SQUID, complete with sophisticated readout electronics, are available. Close attention has been paid to the optimum coupling of the rf SQUID to the room temperature electronics. As a result of this research, the present rf SQUIDs have a higher sensitivity than the first generation of dc SQUIDs.

/pdf/superconducting-quantum-interference-devices-for-low-2xt3l9yzqk.pdf

Superconducting Quantum Interference Devices for Low Frequency Measurements

Large dc supercurrents at nonzero voltages are observed in dc current‐biased small Josephson tunnel junctions of the type PbIn‐ (PbIn)xOy‐Pb with critical current densities j0 ranging from 2×102 to 2×105 A/cm2. The dependence of the hysteresis on j0 and the form of the subharmonic energy gap structure are compared with the theoretical results of McDonald et al.

Observation of large dc supercurrents at nonzero voltages in Josephson tunnel junctions

The behavior of simple superconducting circuits in the picosecond regime is described in a comprehensive way, with primary emphasis being given to the step function and pulse responses of these circuits. Topics receiving detailed discussion include Josephson-junction modeling with both the microscopic and shunted-junction models. Limitations of the shunted-junction model are explored by comparing it with experimental results and with the microscopic model. An approximate evaluation is given of the important dynamical properties of junctions made with the dominant fabrication technology (Pb-alloy systems), as a function of tunneling barrier thickness. Rounding out the device aspects of the discussion, we describe in detail the properties of superconducting microstrip transmission lines, with an emphasis on their high-speed behavior. Turning to simple circuits we review experimental results on the measurement of picosecond regime transient signals. The concept of turn-on delay is analyzed anew, providing simplified and extended results. Details of concepts for pulse height and pulsewidth measurements are explored, leading to the conclusion that the time resolution of superconducting circuits is limited to approximately the period of one plasma oscillation. With present Pb-alloy fabrication technology this limit is 2 ps.

Picosecond applications of Josephson junctions

A detailed investigation of tunneling in Pb/PbO/Pb junctions has been carried out, in an attempt to independently measure several of the parameters known to enter the tunneling problem In particular, the thickness of the tunnel barrier (orthorhombic PbO thermally grown in situ in an ultraclean vacuum system) has been measured ellipsometrically, and barrier heights of 115 ± 015 eV at both metal‐oxide interfaces have been deduced from Fowler‐Nordheim plots and logarithmic conductivity measurements The junctions are of high quality (less than 01% nontunneling currents) and cover a Josephson current density range of from 10−4 to 103A/cm2 From the dependence of the current density on the barrier thickness, an effective mass ratio, m*/m of 05 ± 01 has been deduced for electrons in the barrier Comparison of experimental results with the tunneling theory shows that to obtain an agreement between them, image‐force corrections must be neglected Reasonable quantitative agreement has been obtained with a small correction of 4–5 A to the ellipsometric thickness values in the range 19–31 A; no other adjustments are found necessary This discrepancy might be attributed to a difference between ellipsometric and tunneling thickness values

S. Basavaiah

Papers

Tunneling in lead‐lead oxide‐lead junctions