Effect of rf on subharmonic gap structure in superconducting junctions

2014 The rf-power dependence of the structure in the current-voltage characteristic located at the voltage V = (2 0394 ± nhv)/me has been measured and found very similar to the familiar microwave-assisted tunneling (m = 1). 2 0394 is the superconducting energy gap, v the applied frequency, m = 1, 2, 3, ... and n = 0, 1, 2, ... That is, the variation with power of all these structures is found consistent with the expression J2n(meVrf/hv) where Jn(x) is the ordinary Bessel function and Vrf the microwave voltage across the junction. REVUE DE PHYSIQUE APPLIQUÉE TOME 9, JANVIER 1974, PAGE In addition to the structure in the current-voltage (I-V) characteristic of junctions between identical superconductors at the voltage V = 2 4 1e (2 d is the energy gap), structure is commonly observed also at submultiples of this voltage, i. e., at V = 2 d/me where m is an integer. The origin of this so-called « subharmonic » gap structure (SGS) is not yet resolved although many attempts to do so have been made in the past decade. See, e. g., the paper of Rowell and Feldman [1] ] which discusses the phenomenon at length and contains an extensive bibliography. Mainly two explanations of the structure have been suggested, namely multiparticle tunneling [2] and nonlinear self-coupling of the Josephson radiation [3]. Since the latter explanation invokes the ac-Josephson effect directly whereas the former does not, one might hope that a distinction between the two explanations can be established experimentally by means of an externally applied microwave radiation (rf). Giaever and Zeller [4] found, in fact, that the SGS was strongly dependent on the presence of rf-fields, and Longacre and Shapiro [5] suggested that the rf-response of the SGS in small junctions where spatial variation of the rf-fields is neglected was very similar to the ordinary microwave-assisted tunneling (MAT) [6], [7] associated with the energy gap itself. In order to verify the suggestion of Longacre and Shapiro we have made a careful series of measure(*) Permanent address, Department of Electrical Engineering, University of Rochester, Rochester 14627, New York. REVUE DE PHYSIQUE APPLIQUÉE. T. 9, N° 1, JANVIER 1974 ments of the rf-power dependence of the SGS both for thin-film tunnel junctions (Sn-SnO-Sn) and for point-contact junctions (Nb-Nb). Figure 1 shows the I-V characteristic of a typical point-contact junction (trace a) and also the corresponding derivative, dV/dI, which more clearly exhibits the structure (trace b) with no applied rf. The effect of applied 4 mm radiation (66 GHz) on the junction of figure 1 is illustrated in figure 2 which shows a selection of derivatice traces as functions of bias voltage each trace corresponding to a different level of incoming rfpower. The rf-response demonstrated by this figure is representative for all the junctions tested both thin-film tunnel junctions and point-contact junctions in that they all behaved in full agreement with the following results obtained from figure 2. Considering, first, the energy gap region we find that the structure at V = 2 ~/e decreases as the power increases and that it is replicated on either side of the gap voltage spaced by N 270 jV corresponding to a separation hv/e at 66 GHz. This is recognized as MAT [6] in the limit of high frequencies (this limit is defined by the photon energy, hv/e, being greater than the width of the gap structure). Secondly, we find that the peak at half the gap voltage also decreases with increasing power and at a faster rate than did the energy gap structure. Here, replicas (to be named satellites) grow up on either side spaced by N 135 03BCV, i. e., by hv/2 e. Thirdly, the peak at one third of the gap voltage is seen to decrease still faster with increasing power and at the higher power levels it is seen Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/rphysap:0197400901015300