A gated photoconductor structure is proposed for the dual purpose of enhancing the performance of and investigating surface effects in HgCdTe photoconductive infrared detectors. It is verified both theoretically and experimentally that the passivating native oxide which accumulates the HgCdTe surface, thereby quenching surface recombination, also causes excessive surface shunting and an overall reduction in device responsivity. It is found that for x = 0.23 Hg 1−x Cd x Te photoconductive detectors passivated with native oxide/ZnS the optimum surface conditions prevail at a gate bias corresponding to a semiconductor surface potential of 50 mV. Experimentally, it is found that operation at this optimum condition yields approximately a 70% increase in responsivity in comparison to floating gate conditions, which corresponds to a surface potential of 72 mV for the as-grown native oxide/HgCdTe interface. The gated photoconductor is shown to be a powerful diagnostic device structure that can be used to evaluate and optimize surface passivation layers for HgCdTe infrared detectors.

No abstract

The use of MOCVD-grown wider-bandgap Hg 1−x Cd x Te as a capping layer for long-wavelength infrared (LWIR) Hg 1−x Cd x Te photoconductors has been studied using both theoretical and experimental results. A device model is derived which shows that in the presence of a suitable energy barrier between the Hg 1−x Cd x Te infrared absorbing layer and the overlaying passivation layer, the high surface recombination rate which is usually present at the semiconductor/passivant interface is prevented from having a significant effect on device performance. The energy barrier, which repels photogenerated minority carriers from the semiconductor surface, is introduced by employing an n-type Hg 1−x Cd x Te wafer which consists of a wider-bandgap capping layer that is grown in situ by MOCVD on an LWIR absorbing layer. The derived model allows the responsivity to be calculated by taking into account surface recombination at both the front and back interfaces, thickness of capping and absorbing layers, recombination at the heterointerface, and variations in equilibrium electron concentration. Calculations show that for an x = 0.22 Hg 1−x Cd x Te absorbing layer, the optimum capping layer consists of x ≥ 0.25 and a thickness of the order of 0.1 to 0.2 µm.Experimental results are presented for x = 0.22 n-type Hg 1−x Cd x Te conventional single-layer LWIR photoconductors, and for heterostructure photoconductors consisting of an LWIR absorbing layer of x = 0.22 capped by an n-type layer of x = 0.31. The model is used to extract the recombination velocities at the heterointerface and the semiconductor/substrate interface, which are determined to be 250 cm s −1 and 100 cm s −1 respectively. The experimental data clearly indicate that the use of a heterostructure barrier between the overlaying passivation layer and the underlying LWIR absorbing layer produces detectors that exhibit much higher performance and are insensitive to the condition of the semiconductor/passivant interface.

Mercury annealing of reactive ion etching (RIE) induced p- to n-type conversion in extrinsically doped p-type epitaxial layers of HgCdTe (x=0.31) has been used to reconvert n-type regions created during RIE processing. For the RIE processing conditions used (400 mT, CH4/H2, 90 W), p- to n-type conversion was observed using laser beam induced current (LBIC) measurements. After a sealed tube mercury anneal at 200 °C for 17 h, LBIC measurements clearly indicated that no n-type converted region remained. Subsequent Hall measurements confirmed that the material consisted of a uniform p-type layer, with electrical properties equivalent to that of the initial as-grown wafer (NA−ND=2×1016 cm−3, μ=350 cm2 V−1 s−1).

No abstract