= Giaever gap experiment
{c}
= Giaever superconducting tunnelling experiment
{c}
{synonym}
<Ivar Giaever>'s gap experiment used a metal–insulator–metal sandwich. His lecture describes evaporating <aluminium> onto glass, briefly exposing it to air to form an oxide barrier, then evaporating <lead> cross-strips. Cooling in liquid <helium> made the <lead> superconducting while the <aluminium> stayed normal in the measurements compared. A voltmeter and ammeter supplied the junction <electric potential difference> and <electric current> readouts. Thick oxides initially gave unmeasurably small <electric currents>; the thinner barrier made the measurement practical. Temperature comparisons and a subsequent controlled <magnetic field> test changed the <current-voltage characteristic> with the superconducting state. <Electron Tunneling and Superconductivity by Ivar Giaever>[Giaever, Nobel lecture (12 December 1973), Figs. 2–3 and 6–8; preparation and first successful measurement]{ref}
The 1960 letter compares measurements at $4.2$ and $1.6\,\mathrm K$. An applied parallel <magnetic field> of $2.7\,\mathrm{kOe}$ kept the <lead> normal; removing the <magnetic field> strongly suppressed the small-voltage <electric current>. Figure 2 divides the superconducting $dI/dV$ by the normal-state slope. The letter calls its gap estimates preliminary and notes a temperature-dependent broadened peak. It also flags a limitation of interpreting tunneling directly through the <density of states>; the ideal NIS calculation above states additional assumptions explicitly. <Energy Gap in Superconductors Measured by Electron Tunneling by Ivar Giaever>[Giaever (1960), p. 147, Figs. 1–2; p. 148, concluding paragraph and footnote 2]{ref}
\Image[/giaever-gap-tunnelling-measurement.png]
{title=Functional NIS gap measurement and an original idealized quasiparticle current-voltage curve}
{description=Left: functional schematic of film preparation, cooling and electrical readout in the <Giaever gap experiment>; no historical dimensions or instrument topology are claimed. Right: the derived zero-<temperature> NIS model with a stipulated fixed gap $\Delta$, not digitized data or a fit to <lead>. The normal-state line uses the same $R_N$. Neither curve is a <Josephson effect> pair current.}
The central inference is that the barrier lets <electric current> interrogate available <quasiparticle> states. Suppression at small <electric potential difference> and the rise near the gap edge support a gapped spectrum; tracking the superconducting state helps distinguish that feature from an arbitrary nonlinear barrier. A leakage path through a metallic short, imperfect barrier, or an energy-dependent tunneling matrix element limits a literal <density of states> interpretation. The preparation and state controls therefore matter as much as seeing a nonlinear <current-voltage characteristic>.
Back to article page
