PHYSICAL REVIEW B VOLUME 29, NUMBER 10 15 MAY 1984 Energy distribution of light-induced gap states in hydrogenated amorphous-silicon alloys S. Guha, * C.-Y. Huang, and S. J. Hudgens Energy Conversion Devices, Incorporated, 1675 8'est Maple Road, Troy, Michigan 48084 (Received 23 August 1983; revised manuscript received 5 December 1983) We have studied the effect of prolonged light exposure on the density and energy distribution of gap states in undoped hydrogenated amorphous-silicon alloys. The gap-state distribution in the upper half of the mobility gap has been obtained from measurements of steady-state and transient photoconductivity. The density of states at the Fermi level has been determined from the frequency dependence of capaci- tance of Schottky diodes. The results suggest that light exposure gives rise to new states in the upper half of the mobility gap. Staebler and Wronski' showed that prolonged light expo- sure changes the dark conductivity and the photoconductivi- ty of hydrogenated amorhpous silicon (a-Si:H) alloys. Since then, the effect of light exposure on the properties of a- Si:H alloys has received considerable attention. Several models have been proposed to explain the light-induced effects; a variety of experimental techniques '6 has also been used to study the phenomenon. The experimental results show that light exposure creates new states in the mobility gap. The exact energy distribution of these states, however, is still a subject of controversy. Capacitance vol- tage ' and transient current" measurements on undoped and phosphorus-doped a-Si:H Schottky diodes show that new states are created in the upper half of the mobility gap. Field-effect' " measurements on undoped samples also agrees with these observations. Deep-level transient spec- troscopy (DLTS) t~ experiments on P-doped samples, on the other hand, show that the distribution of density of states (DOS) in the upper half of the gap remains the same after light exposure. However, there is an increase in DOS below the midgap. Similar conclusions have been reached' from the study of temperature dependence of photoconductivity in undoped samples. We have shown elsewhere' that the distribution of gap states can be determined by a combination of measurements of transient photoconductivity (TPC), steady-state photo- conductivity (SPC), and the temperature and frequency dependence of the capacitance of a Schottky diode. In order to resolve the controversy, we have applied these techniques to determine the gap-state distribution of an undoped a- Si:H sample before and after light exposure. Undoped a-Si:H samples were prepared on Corning 7059 glass substrates by rf glow-discharge decomposition of pure silane. The film thickness ranged from 6000 A to 1 p, m. Typical deposition conditions were as follows: substrate temperature 260'C, pressure 0.2 Torr, and rf power 300 mW/cm . Schottky diodes, for use in the C-T-co measure- ments, were prepared on one section of the glass substrate. The bottom Ohmic contacts were obtained by predepositing a NiCr film followed by a 500-A n+layer. The top metal contacts were made by evaporating Pd dots of 50'/0 transmission. Coplanar structures were used for TPC and SPC experi- ments. Both the dark current and the photocurrent were found to be Ohmic up to 10 kV cm ', the highest field used in the experiments. All measurements were carried out in a vacuum system with 10 ' Torr base pressure. To obtain the heat dry state (state A), the samples were annealed at 200'C for 2 h in the vacuum system. For the light- degraded state (state B), the coplanar samples were exposed to 150 mWcm of white light for 2 h at room temperature. For the Schottky diodes the light intensity was increased by a factor of 2 to account for the 50'/o transmission loss through the top Pd contact. For the TPC experiments, a pulsed dye laser (X=640 nm) of 100-ps duration and 10-Hz repetition rate was used for the light source. Intensity of the laser was suitably at- tenuated to avoid bimolecular recombination. The signals after amplification were processed by Biomation Transient Digitizer and Tracor Signal Averager, and finally analyzed by computer. For the SPC experiments, a quartz-halogen lamp with a set of neutral density filters was used to obtain the intensity dependence of photocurrent. In C- T-cu measurements, ' a PAR lock-in amplifier operating in the frequency range of 5-200 Hz was used. Measurements were carried out between 80 and 120'C which is below the temperature at which saturation of capacitance takes place. The Schottky diodes used for the experiments had ideality factors better than 1. 2. In Fig. 1, we present the transient photoconductivity data for an undoped a-Si:H sample in both the heat-dry state and light-degraded state. Below room temperature, one ob- serves that currents decay much quicker for the degraded sample. The steady-state photoconductivity results for the same undoped a-Si:H sample in both states 3 and 8 are shown in Fig. 2. One finds that the light intensity depen- dence of photocurrent has changed substantially after light degradation. At room temperature and above, the slope of the log photocurrent versus log intensity plot in state 8 in- creases with intensity instead of being a constant as in state A. As the temperature is lowered, in state 8 the plot is linear, while in state A the slope reduces at higher intensi- ties. At temperatures lower than 210 K, the plots in both states 3 and 8 are linear, with the slope in state 8 much higher than in state A. The experimental results for TPC can be explained by considering the dispersive motion of charge carriers in the presence of traps. In dispersive transport, ' at a given time t after the initial light pulse, the carrier packet probes the lo- 29 599S O1984 The American Physical Society