HAL Id: jpa-00223059
https://hal.archives-ouvertes.fr/jpa-00223059
Submitted on 1 Jan 1983
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.
TRANSIENT ELECTRICAL PROPERTIES OF LOW ANGLE TILT BOUNDARIES IN GROWN
GERMANIUM BICRYSTALS
A. Broniatowski
To cite this version:
A. Broniatowski. TRANSIENT ELECTRICAL PROPERTIES OF LOW ANGLE TILT BOUND- ARIES IN GROWN GERMANIUM BICRYSTALS. Journal de Physique Colloques, 1983, 44 (C4), pp.C4-339-C4-343. �10.1051/jphyscol:1983440�. �jpa-00223059�
A. Broniatowski
C.N.R.S., Labovatoive P.M.T.M., Universite Paris-Nord, Avenue J.B. Clement, 93430 Villetaneuse, France
Résumé : Les e f f e t s é l e c t r i q u e s t r a n s i t o i r e s associés à l a capture e t à l'émission de porteurs aux niveaux d ' i n t e r f a c e dans des b i c r i s t a u x de germanium sont d i s c u t é s , e t comparés à ceux observés pour l e s d i s l o c a t i o n s dans l e s semiconducteurs déformés plastiquement.
Abstract : Transient electrical effects associated with carrier trapping and emission at dislocation walls in grown.germanium bicrystals are pre- sented, and compared to those obtained for dislocations in plastically deformed semiconductors.
I. INTRODUCTION
This paper reports on the transient electrical effects associated with carrier trapping and emission at dislocation walls in grown Ge bicrystals. Trapping effects have been studied, using direct recordings of capacitance and current transients.
The capture transients exhibit a time dependence similar to that observed for the decay of the photoconductivity in plastically deformed materials III, and explained on the same basis, namely the limitation of the capture processes due to the elec- trostatic repulsion of free carriers by the charged dislocation lines. Recent results obtained from a Deep Level Transient Spectroscopy (DLTS) study of the emission transients will be presented, including a discussion of the energy resolu- tion of this technique as applied to interface states measurements, and the plausi- ble observation of a field enhanced emission effect.
II. CAPTURE TRANSIENTS
The samples are germanium bicrystals, grown by the Czoehralski pulling technique from two seeds set in proper relative orientation / 2 / . The starting material is
13 -3
Hoboken germanium, phosphorous doped to 2.10 cm . A (111) symmetrical low angle tilt boundary is obtained, with a tilt axis [llOJ and a misorientation angle of ~ 3.5 degrees. Electron microscope studies /3/ have shown this type of boundary to consist of [ill] edge dislocations, split into Frank partials 1/3 [ill] with a spacing of ~ 5 0 A. Additional 1/2 <110> and 1/2 <211> dislocations are also observed at places. The density of bulk dislocations estimated by etch pit counting, is about
A -2
10 cm . Because of the small spacing of the boundary dislocations, the screening cylinders overlap so that a continuous (although probably not uniform) potential barrier extends across the bicrystal. Rodlike samples are cut with their long dimension normal to the boundary, and ohmic contacts are electrodeposited on both sides of the dislocation wall. The electrical measurements are of two different types : (i) d.c. measurements of the current across the boundary and (ii) high frequency (1 MHz) measurements of the boundary capacitance.
Fig. la and b shows the current (j) and capacitance (C) transients observed after a voltage step (~ 9 V) was applied from t = 0 across the boundary. The points to be considered are (i) the slow time variation of both the current and the capacitance and (ii) the relative magnitude of the effects, the current change being ~ 3 0 % while in the same time, the capacitance varies by ~ 1 % only.
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1983440
JOURNAL DE PHYSIQUE
C.B.
0 200 LOO 600 800 tisl
Fig. 1 - The current (a) and capacitance Fig. 2 - The bending of the energy (b) transients after a voltage step bands near the boundary under applied (9,4 V) was applied across the boundary. voltage : band bending at t = 0 T = 87 K. Full lines : experiment ; (full line) and at t > 0 (dashed dashed lines : transients computed from line), showing the increase in the Eqs. (4) and (6) . barrier height 6V1.
The explanation of these properties is similar to that of the decay of the photocon- ductivity in plastically deformed semiconductors /I/. From the moment a voltage bias is applied to the specimen, an increasing number of electrons become trapped into vacant boundary states. As a result, the barrier height V1 (fig. 2) also starts increasing ; however, j is strongly dependent upon Vl (as exp(-qVl/k~)). Thus, the current is expected have a large decrease with time, a s compared with a smooth variation in the boundary capacitance (proportional to the change in the boundary charge).
Let us suppose that (i) the electron flow is large enough to make capture strongly predominant over emission and (ii) that the variation in the boundary charge (6Q) remains small compared with the initial value (Qo), SO that the change in the barrier height (bV ) is linear with 6Q :
1
Let j(t) be the current across the sample at time t, and jo its initial value for t = 0. Then, from the thermal emission model for the electron flow over a potential barrier,
where t = kT/qABjo is a time parameter, the same for both the current and the capacitance transients. The capacitance exhibits a time dependence similar to that of 6 Q and given by :
kT t + to c = c (l - qv Log
-)
0
where Co represents the boundary capacitance for t = 0 and V' is a parameter related to A, and estimated by solving Poisson's equation for the shape of the potential barrier to be on the order of 0.3 V in the present case. The dashed lines in Fig. la and b represent the time dependence of j/j and C/C as given by Eqs. (4) and (6) ; a value of t = 1100 s has been taken to fit the experimental data close to the origin. The deviation from the computed curves for increasing t might be explained by the approach to the steady state, where carrier emission can no longer be neglec- ted against capture processes.
111. EMISSION TRANSIENTS
A review of DLTS experiments on the dislocation states in plastically deformed materials will be found in the papers by E. Weber and W. SchrGter (Journal de Physique, this issue). This technique has been recently used to investigate the boundary states in polycrystalline materials as well as in grown bicrystals /4,5/.
In the latter case, filling pulses may be applied across the boundary, and the emission of trapped carriers detected from the change in the boundary capacitance itself 16,'. Both the density and the capture cross-section of the interface states can be obtained from a DLTS analysis of the emission transients, as discussed in refs. /7/ and 181. The energy resolution of this technique will be discussed in relation to the bias dependence of DLT spectra, and the question of field enhanced emission at interface states will be adressed.
111.1. Bias dependence of interface states DLT spectra. - Fig. 3 shows two spectra obtained with the same low angle tilt boundary as in section 11. The spec- trum (a) was recorded under zero applied bias (0.2 V filling pulsesl. The large peak relates to the majority carrier trap El in ref. /8/ (activation energy 0.42 eV). The shoulder on the left side of the peak reveals an additional trap level ; however, the contribution from this level can hardly be distinguished from that of El under these experimental conditions. The spectrum (b) was obtained under 0.4 V biasing (0.6 V filling pulses). The contribution from El is now cancelled, and a well-defined peak (trap E3, activation energy 0.29 eV in ref. 181) appears in the place of the shoulder in the spectrum (a).
JOURNAL DE PHYSIQUE
Fermi
level a
A".
B.Fig. 3 - DLT spectra of a low angle tilt Fig. 4 - The bending of the energy boundary in germanium. (a) 0 V bias, bands at equilibrium (a) and under bias 0.2 V filling pulse ; (b) 0.4 V bias, (b), showing the change in the occupan- 0.6 V filling pulse. Emission rate : cy of the interface levels. EFo : bulk 5,7 s-' ; pulse duration : 100 ms. Fermi level on the negatively biased side of the boundary ; EFt : interface quasi-Fermi level.
To explain these observations, let us consider the voltage dependence of the trap occupancy (fig. 4). Under steady state conditions, a quasi-Fermi level EFt may be defined for the interface states, and shown to be practically coincident with the Fermi level EFo on the negatively biased side of the boundary / 9 / . EFt is shifted upwards as the voltage bias increases, so that the trap level El, initially empty (fig. 4a), is eventually filled with electrons (fig. 4b). Thus, both El and E3 should contribute to the zero-bias spectrum (fig. 3 a ) , whereas the contribution from El should be removed from the spectrum if the applied bias is large enough to raise EFt higher than El, as shown by fig. 3b. This procedure can be followed step by step to describe the density of interface states, starting from the deeper lying levels.
The energy resolution thus obtained should be on the order of the width of the Fermi distribution. i.e. 6-7 kT.
111.2. Field enhanced emission from the interface states. - DLTS provides the possibility to detect field enhanced emission from localized traps, as the emission rate can be measured under variable biasing conditions /lo/. The following experi- ment suggests this effect is actually taking place at the interface states. The El peak was recorded for several biases (fig. 5), 5 V filling pulses being used in all cases to saturate the traps. As the bias increases from 0 V to 0.8 V, the amplitude of the peak is gradually reduced and its position is shifted to lower temperatures.
Both effects can be partly explained on account of the above discussion, the El peak being gradually removed while shallower traps are left to contribute in the emission spectrum. On this basis, the higher bias spectra should be included within those for lower biasing conditions. In fact, much larger temperature shifts are observed than expected, as the successive spectra broadly intersect each other. This observation is consistent with an additional field enhanced emission effect concerning the El trap centre. The magnitude of the field effect can be estimated from the temperature shifts of the peak. Signatures performed under various biasing conditions show a decrease of up to -0.1 eV in the activation energy of the trap, as the applied bias is varied from 0 V to 0.4 V. This estimate is considered indicative only, as shallo- wer traps with an emission rate close to that of El cannot be ignored in the spec-
tra.
REFERENCES
1. FIGIELSKI T., Solid State Electr. 1 (1978) 1403.
2. The samples were provided by the Crystal Growth Division of CENG (Grenoble, France).
3. BOURRET A. and DESSEAUX J., Philos. Mag. A 2 (1979) 405.
4. "Grain Boundaries in Semiconductors", Proceedings of the M.R.S. Meeting 1981, edited by Leamy H.J., Pike G.E. and Seager C.H., North Holland, New York (1982).
5. International Colloquium on Polycrystalline Semiconductors (Perpignan 1982), Journal de Physique 43 (1982) Suppl. C1.
6. SPENCER M., STALL R., EASTMAN L.F. and WOOD C.E.C., J. Appl. Phys. 50 (1979).
8006.
7. BRONIATOWSKI A. and BOURGOIN J., Phys. Rev. Lett. 48 (1982) 424.
8. BRONIATOWSKI A.. Journal de Physique 43 (1982) Suppl. C1, p. 63.
9. STRATTON R., Proc. Phys. Soc. (1956) 513.
10. LANG D.V., J. Appl. Phys. 45 (1974) 3014.
