PHYSICAL REVIEW B VOLUME 27, NUMBER 12 15 JUNE 1983 Study of gap states in hydrogenated amorphous silicon by transient and steady-state photoconductivity measurements C. - Y. Huang* The Standard Oil Company (Ohio), Research Laboratories, 3092 Broadway Avenue, Cleveland, Ohio 44115 S. Guha and S. J. Hudgens Energy Conversion Devices, Inc. , 1675 O'. Maple Road, Troy, Michigan 48084 (Received 29 November 1982) The energy distribution of gap states in amorphous hydrogenated silicon has been investi- gated by transient photoconductivity (TPC) and steady-state photoconductivity (SPC) mea- surements. Both TPC and SPC measurements show that the shallow states decrease ex- ponentially with energy away from the conduction-band edge with a characteristic tempera- ture of 300 K whereas the deep states decrease with a characteristic temperature of about 1000 K. The transition energy is located around 0. 3 eV from the conduction-band edge. The density of states at the Fermi level was obtained from frequency dependence of the capacitance (C-co curve) of Schottky diodes. The derived density-of-states distribution as ob- tained from TPC, SPC, and C-co measurements agrees with results obtained from field- effect and C- V measurements. I. INTRODUCTION It is now well established that amorphous hydro- genated silicon (a-Si:H) films are characterized by a distribution of states in the mobility gap. These states which originate from the tailing of the band edges due to the lack of long-range order' and also from dangling bonds, defects, and impurities in the material have pronounced effect on the photovolta- ic properties of the films. Study of these gap states has therefore received a great deal of attention in re- cent years. The most widely used method for determining gap-state distribution is based on the measurement of field effect. The results show a density of around 10' 10' cm eV ' near the midgap in good quality material; the density increases as one approaches the band edges. Estimates of density of states (DOS) have also been obtained from voltage dependence of capacitance (C-V) and temperature and frequency dependence of capacitance (C T co)-- of Schottky diodes and metal-oxide semiconductor (MOS) structures. C- V results show a midgap den- sity of around 2)&10' cm eV '; the density in- creases as one approaches the conduction-band edge. C-T-co studies " which provide only DOS at the Fermi level g(EF) give a midgap density close to 5X10' cm eV '. g(E+) has been found to be higher in films where the Fermi level is closer to the conduction-band edge again indicating that DOS in- creases as one moves away from the midgap. Deep-level transient-capacitance spectroscopy (DLTS) has been used extensively for studying deep traps in crystals. ' This technique which relies on the measurement of the time dependence of capaci- tance of Schottky diodes or p-n junctions as traps empty out has recently been used to obtain DOS in a-Si:H. In striking contrast to the results discussed earlier, DI. TS studies' ' show a minimum in DOS at around 0. 4 eV from the conduction-band edge with the gap states rising as one approaches the midgap. The minimum value obtained from DLTS is also much lower less than 10' cm eV '. It has been argued'" that the higher values of DOS ob- tained from field-effect, C-V, or C-T-co measure- ments, may be due to the influence of surface states. On the other hand, interpretation of DLTS data in amorphous semiconductors with a continuous distri- bution of gap states is quite complex and is limited in its application orily to samples which have been doped to increase their conductivity. Clearly there is a need for measurement of DOS by independent techniques to resolve this problem. The study of transient photoconductivity of amorphous semiconductors provides a useful tool for obtaining information about the states in the gap. When electron-hole pairs are created by pulsed optical excitation, the photocurrent is found to de- crease with time even when the carrier density remains a constant. This can be explained' by the