Analysis of InP Passivated with Thiourea/Ammonia Solutions and Thin CdS Films
- PDF / 418,091 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 69 Downloads / 166 Views
69 1997 Materials Research Society Mat. Res. Soc. Symp. Proc. Vol. 448 ©
350'C. C-V response was measured at 1 MHz using an HP 4275A multi-frequency LCR meter; quasistatic measurements were made with a Keithley 595 meter. Interface-state densities were calculated by the method of Castagn6 and Vapaille,6 using both high-frequency and quasistatic C- V data.
In order to determine the chemistry of the passivated InP surface following thiourea/ammonia pre-treatment and CdS thin film deposition, XPS analysis was performed on a Physical Electronics PHI 5100 using non-monochromatic Mg K, radiation at 1253.6 eV. The anode is operated at 15 KV with an incident power of 400 W. The diameter of the analysis area is 800 ýtm. A pass energy of 11.75 eV was used for all detail scans. The spectra were corrected for charging effects by referencing the carbon is peak to 284.8 eV. After Shirley background subtraction, Gaussian-Lorentzian peaks were fitted to the spectra for non-linear least squares optimization. The elemental peak areas were corrected using standard sensitivity factors. 7 For
fitting the In 3d5,2 peak, results obtained on vacuum-cleaved InP were used as a guide, with the FWHM for the substrate component being 1.15 eV, and a Gaussian-to-Lorentzian ratio of 2.2. Layer thicknesses, t, were calculated using In
n
Cily,
+
t(
25 x cosO
Csubst-r-e
where Ciaer is the sum of the corrected elemental peak areas in the layer.8 Carbon and oxygen, if detected, were included in this parameter. Csubstrate is the sum of the corrected peak areas of In and P. The polar angle of analysis, 0, was 45'. A mean free path of 25 A was assumed. Following the thiourea/ammonia pre-treatment and, when applicable, the CdS thin film deposition, samples were immediately loaded into the XPS system. Following analysis of the unannealed surface, samples were heated for one hour at 200'C in vacuo, allowed to cool, then re-analyzed. The heating/analysis cycles were repeated at 50'C increments up to 400'C. The base pressure of the system was 1 x 10-10 Torr and rose to - 5 x 10.' Torr during bakeout. An HFetched InP sample was also analyzed as a reference. RESULTS AND DISCUSSION
Figures 1(a) and 1(b) show the 1 MHz and quasistatic C- V responses for HF-etched InP, pretreated InP, and InP with a pre-treatment followed by a thin CdS film (- 30 A) deposition, respectively. For the HF-etched InP, sample A, it is evident that the sample is prevented from becoming accumulated by a high density of fast traps near the conduction band edge, and both high and low frequency response are impeded by a high density of states throughout the band gap. Following thiourea/ammonia pre-treatment, the C-V response is improved, sample B. The sample now goes into accumulation, and the difference in Cox between the 1 MHz and quasistatic curves is significantly reduced. The pre-treated sample still exhibits poor quasistatic response. A higher value of Cmin at 1 MHz also suggests additional states have been added below midgap. This effect has also been reported by Lau et al
Data Loading...
