Accurate Reconstruction of the Density of States in a-Si:H by Photothermal and Photoconductive Spectra.
- PDF / 374,615 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 99 Downloads / 168 Views
ACCURATE RECONSTRUCTION OF THE DENSITY OF STATES IN a-Si:H BY PHOTOTHERMAL AND PHOTOCONDUCTIVE SPECTRA. G. Amato*, F. Giorgis* and C. Manfredotti** *National Electrotechnical Institute G.Ferraris, Strada delle Cacce 91, 10135, Torino, Italy "**Experimental Physics Dept., University of Torino, via P. Giuria 1, 10125, Torino, Italy ABSTRACT The distribution of occupied states in a-Si:H has been inferred by applying a new selfconsisting iterative method to the absorption coefficient spectra. This procedure does not require any assumption about the localized states below the Fermi level, and provides a more accurate insight with respect to the simple derivative method. Numerical simulations have been made in order to probe the reliability of our method. The optical spectra have been obtained by means of Photothermal Deflection Spectroscopy (PDS) and Constant Photocurrent Method (CPM); the comparison between the results as obtained by the two techniques suggests that different sensitivities to electronic transitions are involved; this can be used to infer information about the unoccupied defects. INTRODUCTION The knowledge of the Density of States (DOS) in the pseudogap of amorphous semiconductors is a key step in their characterization: in fact, the localized states are responsible for trapping-detrapping and recombination processes, thus limiting the electronic quality of the material. Optical methods have been widely used in the recent past to infer information about the defect density in a-Si:H films. Among them, PDS [1] and CPM [2] techniques have been the subject of considerable study, due to ease of implementing them and to the quite straightforward results. These techniques allow to obtain the absorption coefficient spectrum a(hv). In order to infer the DOS by the a(hv) spectra, some assumptions a-priori are made. A very common assumption concerns the shape of the conduction onset which is approximated by a step function. This allows reconstructing the DOS by means of a straightforward derivative procedure. Aim of this work is a critical investigation of the effect of a realistic shape of the conduction band on the valence band tail and on the distribution of deep states in the gap, as deconvoluted by absorption spectra. THEORY The main differences between PDS and CPM consist in a different sensitivity to defectsband transitions and to surface states. In PDS thecnique, all the electronic transitions are detected, while in the CPM case, the occupied defects are the only detectable localized states; moreover, PDS spectrum is influenced by surface states, as it can be shown in the major interference fringes enhancement in comparison with the CPM one [3]. The absorption coefficient a(hv) can be written as the convolution integral of initial and final states [4]:
a(hv)=_Af
N(E)f(E)N(E+hvYhvdE
(1)
where the optical-transition matrix element is considered weakly energy-dependent; hv is the photon energy, f(E) is the Fermi distribution for the occupied states and 1-f(E+hv) can be omitted in CPM-case because the fina
Data Loading...
