HAL Id: jpa-00220755
https://hal.archives-ouvertes.fr/jpa-00220755
Submitted on 1 Jan 1981
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
MEASUREMENT AND ANALYSIS OF CURRENT TRANSIENTS IN WELL-CHARACTERIZED a-Si:H
M. Thompson, N. Johnson, R. Street
To cite this version:
M. Thompson, N. Johnson, R. Street. MEASUREMENT AND ANALYSIS OF CURRENT TRAN-
SIENTS IN WELL-CHARACTERIZED a-Si:H. Journal de Physique Colloques, 1981, 42 (C4), pp.C4-
617-C4-620. �10.1051/jphyscol:19814135�. �jpa-00220755�
MEASUREMENT AND ANALYSIS OF CURRENT TRANSIENTS I N NELL-CHARACTERIZED a - S i
:
HM . J . ~ h o r n ~ s o n * , N . X . Johnson and R . A . Street
Xerox PaZo A l t o R e s e a r c h C e n t e r s , PaZo A l t o , CA 94304, U.S.A.
Abstract.. Deep-level current transient spectroscopy (DLTS) and thermally stimulated current measurements have been performed on a-Si:H Schottky barriers over a temperature range of 10
-
300K. Current injection and photoexcitation have been used to accomplish trap filling. The voltage dependence of the transient current in DLTS spectra is presented, and the contribution from interface states was found to be negligible. Deep levels have been studied in a wide range of a-Si:H material with spin densities varying from 1015-3 Xion
cm-3. The contributions to the current transient spectra from carrier transport and thermal emission from deep traps are evaluated.Introduction.- Localized electronic levels in.a-Si:H are continuously distributed in energy and thus pose particular problems for the analysis of deep-level transient spectroscopy (DLTS), which is conventionally applied to discrete deep levels in crystalline semiconductors. and capacitance transient(3) measurenlents have been previously used to study defects in a-Si:H. Cohen et al.(3) measured capacitance DLTS on doped sampies; their technique is not sensitive to shallow levels but was used to study deep levels. randa all(^) studied current transients near room temperature and therefore was primarily sensitive to deep levels. The present study was directed towards understanding the influence of the physical processes such as trap emission and carrier transport on the DLTS spectra in well- characterized a-Si:H.
Sample Preparation and Measurement Techniques.- aSi:H was prepared by the glow discharge decomposition of silane as described elsewhere.(4) A range of undoped materials was produced with spin defect densities from 1015-3 X 10" by varying the substrate.temperature. These materials were incorporated into Schottky barrier diodes fabricated on quartz substrates. Cr, Ni or MO was deposited onto the quartz to form a bottom electrode. In some devices an n+ doped (phosphorous) layer of a-Si:H of thickness 50-600 nm was then deposited onto the metal electrode followed by the deposition of undoped material (of thickness 200-2000 nm). Semitransparent Pd dots of 1 mm diameter and 10 nm thickness were then vacuum-deposited onto the a-Si:H to form the Schottky contact. The n + layer was incorporated to provide a good electron injecting contact for the voltage pulse excitation measurement. Some samples were annealed to 180°C for 15-30 mins. in order to form a Pd silicide interface resulting in ideal Schottky barriers with ideality factors of 1.05.~~) The current transient was measured before and after annealing to assess the contribution of interface states to the DLTS signal.
The measurement technique has been described previously.(2) The non-steady state distribution of trapped charge was accomplished either by photoexcitation through the Pd contact with a 633 nm laser source or by pulsing the diode into forward bias to inject electrons into the a-Si:H. The measurement of the relaxation of the material back to the steady state was achieved by monitoring the current under reverse bias conditions. DLTS spectra were obtained by recording the decay current at a fixed time to after trap filling over a range of temperatures (inset in Fig. 1). The time $ defines the emission rate window, i.e., (l/to) of the current spectrometer. Both electron and hole traps are filled when photoexcitation is used whereas only electrons are injected into the material under voltage excitation.
The advantage of photoexcitation is that the charge excitation density is independent of temperature.
In voltage excitation the injected charge density is strongly temperature dependent due to the activation of the Schottky barrier and bulk a-Si:H.
Experimental Results.- Current transient spectra for samples under voltage excitation is shown in Fig, l(a). The a-Si:H was grown at substrate temperatures of 150 and 230C resulting in spin defect
*Permanent address: Dept. of Elec. Eng., University of Sheffield, England.
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:19814135
JOURNAL DE PHYSIQUE
Time (psecl-
-
l
I
l II
v = 10'7 sec- 1
$ ::
Temperature (K) E (eV)
Fig.1 DLTS of voltage-pulse excited a-Si:H.
(a) current transient spectra for samples deposited at T, = 150 and 230C.
Inset - current versus time after voltage excitation.
(b) trapped-charge distribution for same.
densities of 3 X l 0 I 7 and -3 X 10" cmp3 respectively. Thermally stimulated current (TSC) measurements were also performed on similar samples. Here the samples were subjected to continuous photoexcitation at low temperatures. The samples were then slowly heated and the current monitored as a function of temperature. A peak is observed in the TSC spectra in a similar temperature range to the DLTS peak. However, both these spectra are considerably different from the DLTS spectra obtained with photoexcitation as shown in Fig. 2. It can be seen that there are (1) a large low temperature peak between 50 and 75K. (2) a distinct shoulder at llOK and (3) a peak at 200K. The spectrum decreases in magnitude and shifts to higher temperature as the reverse bias voltage is reduced. It was found that the spectrum was identical before and after annealing. The Schottky barrier characteristics did change on annealing due to growth of crystalline Pd,Si at the interface.(') Thus it appears that interface states do not contribute to the DLTS spectra.
In samples with higher defect density the lower temperature peak is shifted towards higher temperature.
In addition increasing the length of the sample or incorporating n+ layers of high defect density in the Schottky barrier produced a displacement of the low temperature peaks towards higher temperatures.
Discussion of Results.- The DLTS spectra were analyzed by assuming that the limiting processes for current collection was thermal emission of carriers from deep traps to extended states, with an emission time T, given by
where v, is the attempt to escape frequency, E is the trap depth from the extended state band, k is Boltzman's constant and T is the absolute temperature. Thus if the current transient is limited only by trap emission then at a time
6
after the trap filling event the current i(tO) is('s6)i(t& = (qAlkT/2to) nbo (E) (2)
where E = kTln(vtO), q is electronic charge, A is electrode area, and I is the film thickness. Thus the trapped charge can be computed from Eq. (2) and is proportional to the current divided by the temperature. The electron charge distribution corresponding to the voltage excited spectra in Fig. l(a) is shown in Fig. f(b). However, as trap saturation was not achieved in these samples at the end of the voltage excitation pulse, this distribution represents a lower ~imR to the actual trap distribution.
The next issue to address is the difference between the p t a g e excitation and photoexcitation DLTS spectra. In order to explain these differences the carrier transport process in the a-Si:H must be taken
into account. The electron and hole mobilities in a-Si:H are relatively low (<l0 cm2 v-I S-l), even at room temperature. Thus at some low temperature the transit time of the carriers is greater than the sampling time to. Below this temperature the current transient will decrease in amplitude due to reduced collection of the thermally emitted carrier. A further complication is that at <300K the electron and hole transport becomes dispersive resulting in a time dependent mobility(7)
where a is the dispersion parameter. In order to evaluate the effect of dispersive transport on the DLTS spectra we consider the simpler case of transport of a charge sheet. The current due to a sheet of charge transported across the material is of the form illustrated in Fig. 3(a). This is a schematic of data obtained in drift mobility measurements by the time-of-flight technique. The transient current obtained over a range of temperatures is shown. Thus if the current was monitored at a fixed time, to, and the temperature varied, the current-temperature response can be calculated. For transit t h e s 7t greater than to, the DLTS signal is
i(tO) a exp (TAW/kTOTr) (4)
where A W is the activation energy for the mobility, a E T/Tf, and To is the peak in the current- temperature response where rt = r0 exp AW/kT. For 7,
<
toi(to) 0: exp (
-
TAW/kToTi) (5)Thus the current-temperature spectrum for a sheet charge transit is calculated and shown in Fig. 3(b) for the assumed values of A W = O.leV and p. = 10 cm2 V-' S-'. A similar calculation can be made for holes which results in a peak at 210K.
JOURNAL DE PHYSIQUE
10 50 100
Temperature (K)
Fig. 3 (a) Schematic of log current versus log time for time-of-flight measurements on a-Si:H at various temperatures.
(b) Computed current versus temperature scan for to = 100 p.
In the above analysis transport of a sheet charge has been considered and thus for the case of the DLTS spectra where trapped charge is initially uniform in the sample, the distribution will be distorted from that shown in Fig. 3(b). In particular the peak in the curve will be more rounded instead of a cusp as in Fig. (2). For a uniform electric field there is a linear weighting to the current from charge which traverses the complete length of the sample. However, space charge effects would result in a superlinear weighting. Thus the sheet charge approximation provides a qualitatively correct description of the transient response.
Conclusion.- Thus it is concluded that the low and high temperature peaks in the photoexcited DLTS spectra are due t~ electron and hole charge transport effects. The displacement of the low temperature peaks with sample length and defect density is consistent with this model. Thus with photoexcitation DLTS these peaks will always be present and the lowest temperz!ure s a k will always arise from electron transport. The low temperature peak is not observed in the voltage excitation spectra because low injection prevents trap filling at these temperatures. Since the high temperature DLTS peak is due to holes, a corresponding peak would not be expected in the voltage excitation spectra as only electrons are injected with the sample. The TSC is a quasi-steady state measurement and thus the transient carrier transport effects producing the low and high temperature DLTS peaks would not be observed. Therefore the shoulder observed in the DLTS spectra at llOK is the feature which is derived from electron trap emission and is related to the peaks in the TSC and voltage excitation spectra. The distribution of states derived from these data relate to an equilibrium steady state distribution of charge in tail states and is a lower limit on the density of states.
We are pleased to acknowledge mobility data and helpful discussions with M. Rosenblum and express our appreciation to R. Lujan and D. Moyer for technical assistance. This work is supported by SERl contract No. XJO-9079-1.
References
1. CRANDALL, R. S., J. Electron. Mater.
9
(1980) 713.2. JOHNSON, N. M., THOMPSON, M. J., and STREET, R. A., A.I.P. Conf. Proc., in press.
3. COHEN, J. D., LANG, D. V., and HARBISON, J. P., Phys. Rev. Lett.,
3
(1980) 197.4. STREET, R. A., KNIGHTS, J. C., and BIEGELSEN, D. K., Phys. Rev.
818
(1978) 1800.5. THOMPSON, M. J., JOHNSON, N. M., NEMANICH, R. J., and TSAI, C. C., Appl. Phys. Letts. (1981).
6. SIMMQNS, T. G. and WEI, L. S., Solid State Electron. 17 (1974) 117.
7. PFISTER, G. and SCHER, H., Adv. Phys.
