Return to search

Transient spectroscopy of II-VI semiconductors

DLTS, ODLTS and DLOS have been used to characterise the main deep level trapping centres in some II-VI semiconductors; these were single crystal CdS, (ZnCd)S, CdSe, CdTe and ZnS, and polycrystalline CdS films. Undoped, single crystal CdS contained four electron traps as detected by DLTS, at 0.29eV, 0.41eV, 0.61eV and 0.74eV below the conduction band (CB). The first two were observed in all samples and were due to native defects. The two states of highest energy were found only in material that had been annealed in S or Cd vapours. The 0.61ev level could be photoinduced by illumination at photon energies greater than about 1eV. It decayed in the dark with an activation energy of 0.25eV. The 0.61eV and 0.74eV centres were associated with electrically active extended defects (subgrain boundaries Such samples had dislocation densities of about 10(^10) cm(^-2). Copper was found to be a residual impurity in CdS. It produced two deep hole traps resulting from a crystal field splitting of the Cu d(^9) state. They were detected by ODLTS and DLOS and were found at 0.35eV and 1.lev above the valence band (VB).Introduction of the isoelectronic impurity tellurium into CdS induced a hole repulsive centre at 0.21eV above the VB. This is thought to be an inportant radiative recombination centre. The main electron trap in CdS at 0.41eV was found to shift to higher energy with incorporation of Zn. Replacement of 20% of the Cd with Zn shifted the energy to 0.63eV. The level appeared fixed to the VB and had a similar functional dependence on composition as the band gap. The activation energies of the copper centres observed in CdS remained unchanged with incorporation of Zn up to the composition (^Zn)0.45 (^cd)0.55(^s) showed that the crystal field splitting was constant and that these levels were also pinned to the VB. During the fabrication process of the (ZnCd)S/Cu(_2)S solar cell, a deep level was induced at about 1.2eV below the CB. This is thought to be a recombination centre and one of the contributory factors to the reduction observed in the current collection efficiency of these devices. Polycrystalline CdS films were prepared by silk screen printing (SP) and evaporation. The SP films were annealed at various times and temperatures to improve the crystallinity of the layers. At 640C for 1hr, deep states at 0.16eV and 0.48eV were detected. The levels disappeared when annealed at 670C-700C and a new level was observed at 0.13eV. CdS/Cu(_2)S heterojunctions were prepared on the material sintered at 670C; this induced a further trapping level at 1.1eV and one that was poorly resolved. Copper diffused into the CdS during the fabrication of the device so the states associated with copper were detected at 0.35eV and 1.1eV, The evaporated CdS layers showed that the defect signature was sensitive to the type of substrate. Using Ag instead of the usual SnO(_x), deep states were induced at 0.48eV and 0.98eV below the CB. These Ag-associated impurity centres prevent the indiffusion of Cu during the optimising heat treatment of the CdS/Cu(_2)S heterojunction. This maintains the stoichicmetry of the Cu(_2)S layer, thereby, preventing degradation of the devices. CdSe and copper doped CdSe were found to contain several important defect centres: a native sensitising centre (0.64eV from the VB), a class I recombination centre (0.9eV from the CB), a copper impurity centre (0.2eV from the CB) and two native defects (0.16eVand 0.45eV from the CB). n-type CdTe grown by the Piper-Polich technique contained6 electron traps at 0.15eV, 0.21eV, 0.40eV, 0.47eV, 0.53eV and 0.63eV. Their presence was shown to be dependent upon the method of growth of the crystal by comparing with material grown by other techniques. One or more of these states were thought to be due to extended defects or Te precipitates. Low resistivity ZnS contained two deep electron traps at 0.25eV and O.50eV as detected by DLTS. In addition DLOS showed the presence of four further states at 1.25eV, 1.37eV, 1.89eV and 2.19eV below the CB. The first two are thought to be the strong luminescence centres observed by other workers.

Identiferoai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:356650
Date January 1985
CreatorsClaybourn, M.
PublisherDurham University
Source SetsEthos UK
Detected LanguageEnglish
TypeElectronic Thesis or Dissertation
Sourcehttp://etheses.dur.ac.uk/9298/

Page generated in 0.0017 seconds