Nalazite se na CroRIS probnoj okolini. Ovdje evidentirani podaci neće biti pohranjeni u Informacijskom sustavu znanosti RH. Ako je ovo greška, CroRIS produkcijskoj okolini moguće je pristupi putem poveznice www.croris.hr
izvor podataka: crosbi !

Study of microscopic mechanism(s) of spontaneous electrical compensation of donors in CdS by fast diffusers (Cu, Ag, Li and Au) (CROSBI ID 752483)

Druge vrste radova | izvještaj

Desnica Uroš ; Stötzler, Arno ; Desnica-Franković, Ida Dunja ; Deicher, Manfred Study of microscopic mechanism(s) of spontaneous electrical compensation of donors in CdS by fast diffusers (Cu, Ag, Li and Au) // Annual Report for 1998 "Solid State and Cluster Physics", Konstanz Universität; pp 51-52. 1999.

Podaci o odgovornosti

Desnica Uroš ; Stötzler, Arno ; Desnica-Franković, Ida Dunja ; Deicher, Manfred

engleski

Study of microscopic mechanism(s) of spontaneous electrical compensation of donors in CdS by fast diffusers (Cu, Ag, Li and Au)

The goal of this study was to identify microscopic causes of dramatic changes of electrical properties observed in CdS crystal1,2) when doped (or unintentionally contaminated) with some of fast-diffusing (FD) species (FD = Cu, Ag, Li or Au). 1) Influence of FD on electrical properties As the FD atoms penetrate into n-CdS from the surface - provoked by thermal treatment at selected annealing temperatures, Ta- they gradually compensate donors and finally form a strongly compensated, semi-insulating, SI, CdS layer. In order to study in-diffusion quantitatively, a a capacitor structure was created on n-CdS, in which one 'plate' (and one contact) was the evaporated FD dot, while the other 'plate' was the rest of still conductive n-CdS, with alloyed In as a second contact. By measuring capacitance (C), the thickness of the SI-layer, W, reflecting the penetration of FD, was determined after each annealing step. (Furthermore, the C-V measurement enabled the profiling of effective, uncompensated, donor concentrations, n = ND,eff. ş ND-NA = (2/[(q×A2 e)×d(1/C2)/dV]), as well.. By measuring C(V) at various Ta as a function of time spent at that temperature, the penetration rate, and then diffusion coefficients, of each FD can be determined with great precision and spatial resolution (Fig. 1). While the classical methods for determination of diffusioncoefficient are limited to higher T, here we were able to measure well-resolved changes of C-V curves down to 100o C for Cu, and down to 350o C for Au in CdS - i.e. temperature ranges 100-150oC lower than previously reported2). Figure 1 also shows that it was possible to distinguish a faster and a slower-diffusing components of FD: The faster component lowers effective donor concentration deeper inside the CdS (right side of each curve), while the slower diffusing component causes the compensation and formation of SI layer (left). The position of SI/n-CdS transition shifts deeper with the diffusion time. Fastest diffusing FD was found to be Cu, followed by Ag, Li and Au. From the standpoint of the influence on electrical properties, most important fact is that all of the studied FD act effectively as acceptors in CdS causing complete compensation of donors (n in the 1016/cm3 range) in the surface region of n-CdS. For example, for diffusion of Cu at 145oC less than 1 hour was needed to fully compensate 1 micron depth of CdS, while several tens of hours necessitated in the 130-115oC range, and above 100 h at 100oC. 2) Changes at the atomic level due to in-diffusion of FD Perturbed angular Correlation spectroscopy, PAC, was used in order to explore processes at the microscopic level that might cause electrical compensation due to the in-diffusion of FD. For that purpose CdS samples were first pre-doped with 111In probe atoms in a controlled way elaborated earlier3, 4) in a desired lattice sites. Concentration of 111In was very small, below the concentration level of residual impurities, so that pre-doping practically does not change chemical content of the original CdS material. The depth-distribution of probe atoms was evaluated by performing gradual etching and determining concentration profile by measuring the remaining weight and activity of the sample after each etching step. Two types of pre-doped CdS samples were prepared: a) Samples in which all 111In atoms were incorporated at Cd sites (InCd) 3,4), resulting in only 7.8 MHz frequency in PAC spectra (Fig. 2a). In that case all probe atoms are donors, and hence, potentially, a trap for some diffusing acceptors. b) Samples in which only about half 111In probe atoms were incorporated as InCd donors, while the other half was paired with Cd vacancies3,4). The so formed (InCd+-VCd2-)- pairs are acceptors, hence pose a potential trap for some diffusing donors. This situation is characterized by the presence of characteristic 72&78 MHz frequencies (besides 7.8 MHz) which reveal the existence of comparable fraction of (InCd+-VCd2-)-pairs. On both samples the FD layer has then been evaporated. After each annealing step PAC spectra were taken, in order to monitor changes at and around probe atoms. Annealing was performed either under S or Cd partial pressure, in the temperature interval from RT up to 700oC. Thermal treatment of CdS:FD brought considerable changes in the microscopic surrounding of 111In probe atoms already at annealing temperatures as low as 200-450oC, primarily the increase of concentration of (InCd+-VCd2-)- pairs (Figs. 2b abd 2c). This was not the case with control samples (i.e. samples treated without presence of FD), where changes occur only at considerably higher annealing temperatures (above 500oC). 3) Correlation of microscopic and macroscopic data A comparison of C-V and PAC measurements enables establishing the connection between microscopic and macroscopic properties in FD-doped CdS crystals. In particular: a) The formation of neutral donor-acceptor pairs of the type (InCd-FDCd) is not the mechanism of compensation (passivation) of substitutional donors (like 111InCd) in CdS. Only in Cu doped CdS this type of direct passivation of donors cannot be fully excluded, since a new, Cu-related PAC frequency at 19 MHz, was observed in CdS:Cu. The microscopic configuration responsible for this frequency has not been as jet identified. Still, even if one should prove that the 19 MHz PAC signal belongs to (InCd -CuCd), its fraction never exceeded more than ź of InCd donors, indicating that most donors were passivated in some other way. b) In CdS samples on which the layer of either Ag, Li or Au was evaporated, cadmium vacancies were detected (through their pairing with 111InCd) already at temperatures from 200 to 450oC. However, in undoped CdS samples temperatures above 500oC were needed for thermal creation of VCd (and that only for thermal treatment under S pressure). The increase of concentration of VCd (estimated from PAC) seems to follow the rate of in-diffusion of FD species into CdS (as determined from C-V measurements. c) All of these results suggest a complex microscopic mechanism by which FD in-diffusion into CdS creates highly compensated (semi-insulated material): FD atoms diffuse fast into CdS by provoking simultaneous in-diffusion of Cd vacancies into the crystal. Some of these vacancies become trapped at donors passivating them electrically, while the others serve as substitutional sites for diffusing FD atoms. In Au and Li doped CdS practically all donors are compensated directly with VCd trapped at the nearest Cd site, whereas in Ag or Cu-doped CdS most donors are electrically compensated from the distance, by FDCd acceptors. d) In samples prepared to have (InCd+-VCd2-)- pairs, small but distinctive change of PAC parameters belonging to these pairs was observed in the case of Ag and Cu (decrease of PAC frequencies of pairs for 5-. 10% for Ag and even less for Cu). The change is compatible with a notion that diffusing interstitial Cu or Ag atoms get trapped at pairs: Since FD interstitals are donors, Cui+ (or Agi+) are expected to be attracted to the VCd-- whereas at the same time repulsed from the donor InCd+. Hence, the interstitial lattice sites farthest from InCd+, are energetically most favorable. Hence FDi (Ag and Cu interstitials) should be 2.5 places away from probe atoms so that the small perturbation of pair frequencies is understendable. The complex should have a microscopic form (InCd+-VCd2- -FDi+)o and should be electrically neutral. No indication of such complexes was observed in Li or Au doped CdS. (1) J.L. Sullivan, J. Phys D: Appl. Phys. 6, 552 (1973) (2) Landolt-Börnstein New Series III, Diffusion in Semi-conductors, Vol.23A, Ed. D.L. Beke, Springer 1998. (3)R. Magerle, M. Deicher, U. Desnica, R. Keller, W.Pfeiffer, F. Pleiter, H. Skudlik and Th. Wichert, Appl. Surf. Sci. 50, 169 (1991), (4) U.V. Desnica, I.D. Desnica-Frankovic, R. Magerle, A. Burchard and M. Deicher, J. Cryst. Growth (1999), in press.

fast diffusers; II-VI; compensation; Cu; Li; Au

nije evidentirano

nije evidentirano

nije evidentirano

nije evidentirano

nije evidentirano

nije evidentirano

Podaci o izdanju

Annual Report for 1998 "Solid State and Cluster Physics", Konstanz Universität; pp 51-52

1999.

nije evidentirano

objavljeno

Povezanost rada

Fizika