THERMAL AND OPTICAL SPACE CHARGE SPECTROSCOPY OF GAP STATES IN a-Si:H

(1)

HAL Id: jpa-00220751

https://hal.archives-ouvertes.fr/jpa-00220751

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.

THERMAL AND OPTICAL SPACE CHARGE SPECTROSCOPY OF GAP STATES IN a-Si:H

A. Chenevas-Paule, J. Dijon

To cite this version:

A. Chenevas-Paule, J. Dijon. THERMAL AND OPTICAL SPACE CHARGE SPECTROSCOPY OF GAP STATES IN a-Si:H. Journal de Physique Colloques, 1981, 42 (C4), pp.C4-605-C4-608.

�10.1051/jphyscol:19814132�. �jpa-00220751�

(2)

CoZZoque C4, suppZ4ment au OlO, Tome 42, o c t o b r e 1981 page C4-605

THERMAL AND O P T I C A L SPACE CHARGE SPECTROSCOPY OF GAP S T A T E S I N a - S i : H

A. Chenevas-Paule and J. Dijon

L. E.T. I., Connrrissariat 2 Z ' E n e r g i e A t o m i q u e , 85X

-

3 8 0 4 1 GrenobZe, F r a n c e

Abstract.- In this article we present a new method based on TSC (Ther- mally Stimulated Currents) wich allows the observation of deep centres in amorphous semiconductors. This technique allows to eliminate the experimental contribution of the band tail states in the space charge zone of a Schottky barrier. We discuss the charact6ristics of the deep centres thereby revealed and their connection with the existence of structural microinhomogeneities.

Introduction.- The knowledge of the density of states in the pseudogap of a-Si:H is of primary importance. In crystalline semiconductors, it is accepted that shallow localized states control the mobility (P) of the carriers, while the deep traps determine their lifetime (r), the photoconductivity being proportionnal to the product T p

.

In amorphous semiconductors, the band-tail states created by the disorder can be considered as shallow traps with a very large capture cross-section and they influence strongly the transport properties (dispersive2~;~~;~ft for instance) by limiting the mobility to values lower than 1 cm

Recently, some attempts have been made to measure the total density of states in the forbidden gap of these materials, by the field effect, capacitance studies and DLTS. These methods have the following drawbacks :

Field effect : very sensitive to surface states ; gives an over estima- tion of N(E)

-

^(1).

Capacitance : can only explore the region close to the Fermi level (EF k0.l eV).

DLTS (Deep Level Transient Spectrocopy) : in principle well adapted for measuring the density of states in crystals ; difficult to carry out in the amorphous materials because of the large density of states in the band-tails which tends to swamp the deep centres.

In this paper, we describe a method based on the TSC (Thermally Stimulated Current) which eliminates the contribution of the tail states and hence allows us to study the deep centres, and to measure their principal parameters : T (capture cross section). ET (energy of the traps related to mobility edges), N(E~) (density of states).

Experimental.- The principle of this experiment is to study the thermal relaxation of the traps in a space charge zone (SCZ) (2). The SCZ used here is that of a Schottky barrier in diodes ~t/a-~i:~/n+ a-Si:H, made by sputtering. These diodes are described elsewhere (3). The short-circuit current due to detrapping in the SCZ is measured after the SCZ has been put in a non-equilibriud state.

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:19814132

(3)

JOURNAL DE PHYSIQUE

The different traps are filled at 77K by various methods for obtaining the non- equilibrium state (optical generation of electrons and holes, electric biasing) which constitutes the excitation stage of the experiment.

Experiments of the DLTS type recently reported by other laboratories (4.5) analyse the transient relaxation (capacitance, current) just after excita- tion ; the information obtained is the convolution of components due to the tail states and to the deep centres, and such experiments cannot resolve the different contributions. In our experiment, after excitation at 77K, the shallow centres are allowed to be emptied, leaving only the deep states (61, the thermal relax3tion ff which is then stu led by heating the sample at a constant rate B : 5.10 K.S- < ^fj< _{0.5 K.S}

-P' .

The use of a SCZ also permits us to study the detrapping without bia~&ng the sample. This permits us to observe significant currents as low as 10- A,which is not possible when studying a coplanar structure polarized by an external source, because the effects of band-bending can then be predominant.

Figure 1 shows the successive stages of the experiment, and the different excitation possibilities. The experiment is thoryghly controlled by a HP 9825 computer. The sample is held in a vacuum of P <l0 TORR. The activation energies E are determined by two

complementary methods T : 1. Using the relation ET = 23 k.TM where T

is the value for whi%

d ITSC/dT = 0. This rela- ,at,,e

tion, used for crystals (7), gives correct

results for traps having t ! t

2. Using a method known as "fractional emptying"

the activation energy can be obtained from an Arrhenius plot as in Fig lC, and neighbouring peaks can be deconvolved.

Fig.2 and fig.3 show the results obtained with a conventional a-Si:H diode.

Experimental results and discussion.- In crystalline semiconductors, deep centres generally

constitute discrete levels the energy of which are well-defined with respect to the bands. One might expect a priori that in amorphous semiconductors.

given the disorder, the deep centres would have a distribution of energies with respect to the mobility edges.

Spectrum

r a t w e

Figure 1 :

Chronology of various

t experimental phases.

1 0 " a) Optical excitation.

b) Electrical filling.

c ) Fractional emptying and typical Arrhenius plot.

I : : . . . . . . . .

5 1 ' ' l@

T

(4)

The technique of fractional emptying shows that these bands are in fact Aade of two or three peaks. Assuming a Gaussian distribution of energy of these levels, calculation (J. DIJON unpublished) has shown that the width and shape of the TSC bands require that the half maximum width of the distribution is T= 10 meV. Notice that T = 100 meV.would give TSC spectra exhibiting almost no structure.

Detailed analysis of the spectra gives the following informations : i) The first band includes three peaks corresponding to the following energy

levels (eV) : E E

Tz T3

Arrhenius plot : 0.19 0.22 0.26

2 3 kTM 0.20 0.24 0.28

The _ffQrly2good agreement of these values means that the capture cross- section is = 10 cm

.

Measurements of T as a function of B confirm this result.

M ii) The second band includes also three peaks :

Arrhenius plot : 0.29 0.32 0.5

23 k T, 0.34 0.47 0.53

A A

The 23 kT rule gives erroneous values of _M E indicating that

f g

^thhs

case the capture cross-sections are smaller than the "s?&ndard" value 10- cm

.

-17 2 The analysis of T as a function off, _M gives T~ E T ~ ~ , O cm

.

iii) When the optical excitation is such that the density of absorbed photons is low, the form of the TSC spectra is relatively independent of the excitation wavelength (437 <

X

c656 nm), except the appearence of a negative peak near the second band forX>600 nm, due to the penetration of the excitation beyond the SCZ.

Figure 2 : Figure 3 :

Typical TSC spectrum at low density of Wavelength excitation dependence of absorbed photon and experimental decon- TSC spectrum for optical saturation.

volution.

iv) After saturating the optical excitation. The TSC spectra are shifted according to the wavelength of the excitation (fig.3).

J n 1s-l1 A

1,s.

.

I.ra

l I

, , , , J r KELVIN

-alf W E W

E P

K: B

--Red light e56nm .1sev ..- Blue l i p h t 4 3 l n m

!

.22cV

:\

(5)

JOURNAL DE PHYSIQUE

This could be due to the densities of ET2 and ET5 which are relatively higher at

the surface than in the bulk. a

v) The total density of these deep centres is estimated to be 2.10 17 cm-'. This indicates that given their positions relative to the mobility edges, they cannot be observed by DLTS (capacitance or current), except perhaps E because they

.T6

'

. are located in the regions where the density of band-tail states is high.

In addition, we have demonstrated techniques of electrical filling of traps by biasing the 'schottky diode at low temperature (77K). Peak corresponding to E (0.19 eV) is the only appearing positive or negative according to the sign of t%?6 bias indicating that the current does not come from the relaxation of traps in the SCZ, but is linked to a rearrangement of charges in the structure beyond the SCZ.

Finally, it is remarkable that the levels we find are quasi-discrete (non distributed widely with respect to the mobility edges) in the forbidden gap of this disordered material. This means that the centres are in identical local environments. Consequently, these centres are not homogeneously distributed. The study of the deep centres may therefore be associated to the existence of microinhomogeneities.

Conclusion.- We have not yet identified all the traps the existence of which we have demonstrated. TSCAP (Thermally Stimulated Capacitance) measurements are essential to obain unambiguously the nature of these traps ; these measurements are in progress, but they are much more delicate than the TSC. Moreover, these centres can also act as recombination centres when the material is under illumination (the quasi-Fermi levels are then near the bands) ; thus their study is of great importance.

Finally, we point out that by this means, spectroscopy of deep centres as in crystalline semiconductors should be possible.

The authors wish to thank R. CUCHET and M. EYSSERIC for their technical assistance and D.J. DUNSTAN for many helpful discussions. This work was partly supported by COMES.

References

.-

l) M.J. POWELL, Phil. Mag B43 (1981) 93.

2) BRAUNLICH (Ed), Topics in Applied Physics, Vol 37,

3) L. VIEUX-ROCHAZ, A. CHENEVAS-PAULE, D. JOUSSE and P. VIKTOROVITCH, Proc. of the 1979 Photovaltaic Solar Energy conf., BERLIN (WEST).

4) R.S. CRANDALL, Sol Cell, 2 (19801, 319.

5) J.D. COHEN, D.V. LANG, J.P. HARBISON and J.C. BEAN, Sol Cell, 2 (19801, 331.

6) A. CHENEVAS-PAULE, J. Phys. Soc., JAPAN, 49 (1980). Suppl. A 1205.

7) H. SHADE, J. Appl. Phys. 40, (1969). 2613.