b Laboratoire de Spectroscopie Raman, Facultd des Sciences de Tunis, D@artement de Physique, ... confinement formalism have been proposed by John and.
ELSEVIER
Thin Solid Films 304 (1997) 358-364
Changes in photoluminescence behaviour and structure of porous silicon related to preparation conditions and laser irradiation H. Elhouichet a, B. Bess,/is a, O. Ben younes b, H. Ezzaouia a, M. Oueslati b a lnstitut National de Recherche Scientifique et Technique, US1, Laboratoire de Photot,olta'ique et des Mat&iaux Semiconducteurs. BP 95, 2050 Hammam-Lif Tunisia b Laboratoire de Spectroscopie Raman, Facultd des Sciences de Tunis, D@artement de Physique, 1006 Le Beh,~dbre. Tunis, Tunisia
Received 21 August 1996; accepted 6 February 1997
Abstract Photoluminescence (PL) spectra of porous silicon (PS) are fitted by a theoretical model based on quantum confinement of electrons in Si nanocrystallites having spherical and cylindrical forms. This model permits one to correlate the PL spectra with the PS structure. It was found that the PS structure is almost independent of the porosity of the PS samples when elaborated in tim same HF solution, but it depends on the composition of the electrolytic solution and post-anodisation treatments such as oxidation. The specific surface area (SSA) was estimated and was found to decrease linearly when the porosity increases. It was pointed out that the SSA plays a key role in the PL behaviour within laser irradiation. The effect of laser irradiation on the PL behaviour has been discussed according to the proposed model, and was shown to be dependent on ambient atmosphere. It was shown that the crystallite size decreases throughout photo-oxidation under laser irradiation in air. © 1997 Elsevier Science S.A. Keywords: Laser irradiation; Luminescence; Nanostructures; Silicon
1. Introduction The discovery of visible room temperature photoluminescence (PL) from porous silicon (PS) has attracted much interest owing to its possible application to optoelectronic devices. Although, a large number of works on PS devices have been published, most basic questions concerning the PL origin and the morphology of PS remain unclear. The essentials of mechanisms which have been proposed for the PL are quantum confinement of carriers in Si nanocrystallites [1] and specific molecular luminescence centres, such as potysilanes [2] or sitoxene [3]. One reason for such a variety in PL mechanisms [4] could be the complexity of PS structures which depends specifically on the experimental conditions. The latter may particularly change the form and the size of the crystallites and the porosity which may have an influence on the specific surface area (SSA) of PS, depending on the electrical resistivity [5,6]. In most works, the effect of interaction between the ambient atmosphere and the SSA of PS has been ignored. Indeed, due to the fact that PS has a very reactive surface, its SSA may play an important role in the PL behaviour within laser irradiation. Thereby, the effect of intrinsic PS 0040-6090/97/$17.00 © 1997 Elsevier Science S.A. All rights reserved. PII S0040-6090(97)00091-6
parameters such as porosity, crystallite size, etc., should be measured in vacuum or in an inert ambient. We try in this paper to point out the effect of porosity, electrolytic solution, oxidation and laser irradiation on the PL behaviour and structure of PS. From theoretically calculated PL spectra, based on the quantum confinement formalism, we correlate PL behaviour to PS structure.
2. Experiment The PS layers are prepared by the electrochemical anodisation method at a constant current density in a HF solution. The substrate used is p-type Si (100) of 1-1.5 f~ cm resistivity; the back ohmic contact is made with eutectic I n - G a . The porosity is determined by the gravimetric method [6]. For the elaborated samples, the layer thickness varies from 3 to 10 ~m. PL and Raman spectra are measured using a " D i l o r " triple monochromator, and a photomutiplier with a GaAs photocathode. The 514.5 nm line of a continuous Ar + laser, with a power density of 4 W cm -2, was used as an excitation source.
359
H, Elhouichet at al. / T h i n Solid Films 304 (1997) 3 5 8 - 3 6 4
3. Theoretical a p p r o a c h
3.1. Model of PL Several models have been proposed to explain the visible PL from PS. Koch et al. [7] explain the PL emission by both the "quantum wire" model and the effect of Si-related surface states, They show clearly that the surface states (i.e. electronic states) contribute to the PL emission in the infrared region. This was recently confirmed by Allan et al. [8]. Other models purely based on the quantum confinement formalism have been proposed by John and Singh [9] and Fishman et al. [10]. They used the effective mass approximation and a Gaussian distribution of the crystallite size to explain the PL line shape. Our PL spectra show a visible emission in the 1.5-2 eV energy range. In this range, one can neglect, in a first approximation, the surface state effects (since ~hey emit in the infrared region [7,8]) and explain the PL line shape purely on the basis of a quantum confinement model. As the PL line shape is much too large, there are no crystallites of well-defined shape and size, and since PS is a disordered system where the formation of the Si crystallites is a stochastic process, then it appears reasonable to describe the crystallites in a statistical way. As has been reported elsewhere [9,10], we consider that PS is formed by wires and dots crystallites obeying to a Gaussian distribution of diameters d W and d d centred around means do, ,, and doa respectively. These distributions are given by: 1 Pw=~exp
-
l(dw-d0w)
2]
(1)
tion of the energy gap vs. the crystallite diameters [11]; - the confinement energy is A E = Cl/2/d 139 given by A E = h ~o - Eg - E b where h w is the PL energy, Eg is the bulk silicon gap varying in the range of 1.14-1.17 eV depending on the temperature, and E b ~ 0.15 eV [9] is the exciton binding energy. Thus, one may take Eg - E b -=- 1.00 e V , and h ~o = A E + 1 (eV). The PL line shape vs. the confinement energy AE, is then given by:
P( A E) = xwP,v( A E) + xdPd( A E) where x w and x~ are the proportions of wire and dot crystallites in PS. This model permits the determination of the average diameters dow and dod, their mean root squares and the proportions Xw and x d.
3.2. Specific surface area calculation From the calculated mean diameter sizes d0w and d0d , and the proportions x w and x d of wires and dots, we can estimate the value of the SSA. Considering a PS layer of porosity P with a top surface of 1 cm 2 and a thickness e (Ixm), the mass of this layer may be written as:
m--pe(1-P)
where p is the density of the bulk silicon, Nd and Nw are respectively the number of wire and dot crystallites in the PS layer, m d and m w are respectively dot and wire average mass.As XW
1
Pd = 2g~crdeXp --
(2)
=Ndm d + Nwm ~,
-
N,v -
Nw+Nd
and x d
Nd :V`` +
then 177
By using the same calculations and assumptions made by John and Singh [9], and taking a confinement in d-1.39 [11], the probability distributions of electrons participating in the PL process in wires and dots are: Pw(A E) = K w A E -3"16
I(
×exp -~-
% ] /d0'v
)]
lr/w -t- .~.---~rfld a n d N a -
m
~m w+md
The inside surface of the layer is: S 1 =NwS w + N j d where S w = ~rd0,~e (surface of a wire) and Sd ~ rrd~a (surface of a dot). The SSA is ' then: S=SI~zT-F lo 7 (m z cm-3).
4. Effect o f p r e p a r a t i o n conditions
(3)
Pa( A E) = K zXE-3 88
X exp - ~ - /
Nw =
/ /
°'a ] \ d0a (4)
where K w and K d a r e suitable normalisation constants; - C~ and C a are the quantum confinement constants respectively for wires and dots deduced from the varia-
In this section, we try to correlate PL spectra to PS structures throughout the previously presented PL model. Fig. 1 shows experimental and theoretical PL spectra corresponding to freshly prepared PS layers having different porosities and elaborated in the same aqueous HF solution (C HF --- 20%). The PL spectra were measured, under vacuum, after the stabilisation of the PL intensity. According to Fig. 1, the PL spectra of freshly PS show a blueshift of the PL peak energy with increasing porosity. One can also notice that during laser irradiation exposure
H. Elhouichet et al. / Thin Solid Films 304 (1997) 358-364
360
3.6 ¸
)
.~
3.4
•~
3.2
dots
09
3.0
