Hot Electron Effects in Excitation of
Visible Photoluminescence in Si Nanowires
T.V. TORCHYNSKAa), L.Y. KHOMENKOVAb), V.N. ZAKHARCHENKOc),
R.V. ZAKHARCHENKOc,d), J. GONZÁLEZ-HERNÁNDEZd), Y.V. VOROBIEVd)
a)Instituto Politecnico Nacional, U.P.A.L.M., 07738, México D.F., MÉXICO
b)Institute of Semiconductor Physics, National Academy of Sciences, Kiev, UKRAINE
c)National Politechnic Institute KPI, Dept. of Gen. Eksp. Physics, Kiev, UKRAINE
d)CINVESTAV-IPN, Unidad Querétaro, Juriquilla, Querétaro, QRO., MÉXICO
Abstract: - Visible photoluminescence of Si nanowires prepared by electrochemical etching was studied with the use of photoluminescence spectral measurements, Atomic Force Microscopy and Magnetic Field effect. The photoluminescence spectra had no dependence upon the nanoparticle size, in contradiction with the model based on the quantum confinement effect. The magnetic field of 0.52 T has a noticeable influence on the luminescence intensity, but not on the emission spectra. The features observed confirm the hipothesis about the emission excitation by hot electrons created by the light absorption in nanowires which energy band structure remains practically the same as in the bulk Si. The effective mobility of electrons involved in the excitation process estimated from the effect value is around 30000 cm2/V·s, which greatly exceeds the normal electron mobility in Si at room temperature.
Key-Words: - Si nanowires, Visible photoluminescence, Magnetic field effect, Hot electrons excitation
1Introduction
The room temperature strong visible luminescence (PL) of Si wire nanostructures (the most frequently used name for them is porous Si, or PSi) has attracted much attention because of its evident incompatibility with the normal Si band structure, and due to its many potential applications in optoelectronics. The most popular explanation of this emission is based on the quantum confinement effect [1], although there are many other models [2-5] which only number is the reason to say that no one is good enough.
Recently a new approach was introduced by one of the authors (Torchynska [6,7]) considering the luminescence excitation process as theballistic effect caused by the hot electrons produced by optical absorption in the Si with practically unaltered band structure, and the hot electrons of the c-band participating in the process. Here we present the new experimental data confirming this model, and the estimation of average mobility of the electrons exciting the emission, which directly indicates that these are hot electrons.
2Experimental
Investigated PSi samples were prepared by electrochemical etching from (100) oriented p-type silicon wafers with resistivity of 1.0 Ω·cm. The HF-ethanol solution was used (HF:H2O:C2H5OH = 1:1:2), and the anodization current density was 20 mA/cm with the duration between 5 and 60 min. Depending on the process duration, the details of the surface relief varied in size, on average, from 30 to 300 nm. Fig. 1 shows the AFM picture of the sample surface after 15 min of etching, with the details of the orders of 100 nm, which proved to be optimum for the PL intensity.
Fig.1. AFM image of the PSi layer fabricated at anodization current density 20 mA/cm2 during 15 min.
The PL spectra of the samples with different etching time are shown in Fig. 2; it is evident that all three spectra have practically the same shape, without any dependence upon the characteristic size of the details of the nanostructure.
Fig.2. The PL spectra for samples made with 5 (squares), 15 (triangles) and 45 (rhombs) min etching time.
The data presented already show that the quantum confinement effects are not responsible for the emission observed: for the characteristic size of the surface observed, these effects are not well developed (whereas the details of the smaller size, which are able to produce really quantum effects, are oxidized quickly, and for that reason are not observable). On the other hand, for these effects, the dependence of the spectrum shape upon the characteristic size ought to be very strong, which is definitely not the case.
On the contrary, the ballistic model introduced in [6,7] well agrees with the stable emission spectrum, which, from the positions of the model, is determined by the radiative defects at Si/SiOx interface. The excitation of these defects is produced by the hot electrons generated by light-induced transitions from the valence band of Si into the Si c-band [6,7]. In this case, the magnetic field effect on the emission intensity is to be expected, since the high energy (“hot”) electrons have relatively high velocity and therefore the magnetic field ought to have a noticeable influence on their motion, reducing the mean free path and correspondingly the probability to reach the surface where the radiative centers are situated.
Fig. 3. Relative reduction of emission intensity caused by magnetic field.
In the Fig. 3 the results of the emission measurements with the magnetic field are shown. The magnetic field of 0.52 T was applied to the Si sample in such a way that the field direction was parallel to the sample surface (i.e. normal to the axes of the nanopiramides created by etching, see Fig. 1). The luminescence was excited with the Hg arc lamp, and the emission in different spectral intervals was selected with the monochromator MDR, and emission intensity was measured with the photomultiplier. It is seen from the figure 3, that the reduction of luminescence intensity does not depend on the emission wavelength, and is equal to 10 % of the initial intensity (in other words, the PL intensity with the magnetic field applied is 0.9 of the initial PL intensity in a given spectral region).
3 Discussion
It was already mentioned that the spectral measurements favor the idea about the photoluminescence excitation with the hot electrons transferred by light to the upper levels of the Si c-band (for the details, see [7]). The effect of the magnetic field reported here strongly supports this point of view. In case of quantum confined systems, the possible magnetic field influence upon the emission may have a form of the Zeeman effect, which (i) demands larger fields to be well pronounced, and (ii) affects, first of all, the emission spectrum, and to less extent the emission intensity. However, we observe quite noticeable effect caused by relatively small field, and no dependence of the effect on the emission wavelength.
On the basis of the hot electron excitation model, the magnetic field effect is explained by the reduction of the electron mean free path caused by the Lorents force and the corresponding curving of electron’s trajectory. If we denote the mean free path in absence of magnetic field as L (the corresponding time between the scattering processes as ), in the magnetic field this path will correspond to the part of a circular trajectory caused by the Lorents force effect determined by the angle of rotation of the radius R of circular trajectory during the time interval and the effective mean free path in magnetic field with induction B will be
L* = 2 R sin /2 ≈ R (1 /24) (1)
Since L = Rwe obtain the relative reduction of the mean free path as
(L – L*)/L ≈ /24 (2)
We take that this value should define the relative reduction of the emission intensity (the smaller is the free path, the smaller amount of electrons get to the surface and excite the emission). The value of angle is determined by the frequency of rotation of electrons in magnetic field (the cyclotron frequency): = e B/m, where “e” is the electron charge, and “m” – its mass. The angle is equal to
= e B /m = B,(3)
where = e/m is the effective mobility of electron.
Taking (L – L*)/L = 0.1, from (2) we obtain = (2.4)0.5 ≈ 1.55, and from (3) we find an electron effective mobility = 30000 cm2/V·s. This value is more than an order of magnitude higher than the normal electron mobility in Si, which also confirms the concept of the hot electron excitation of the photoluminescence.
3Conclusions
The influence of the magnetic field of 0.5 T upon the photoluminescence of Si nanowires was observed: the luminescence intensity in magnetic field constitutes 90 % of the intensity in absence of field. The effect has no dependence on the emission wavelength.
This effect could not be explained on the basis of the quantum confinement model, and is supporting the hypothesis about the luminescence excitation by hot electrons transferred by the illumination from the v-band to the upper levels of the Si c-band, while the Si band structure for the nanowires size observed is pracrtically the same as for the bulk Si.
From the value of the magnetic field effect, we estimated the effective mobility of the electrons participating in the emission excitation process as 30000 cm2/V·s. This value is much higher than the standard electron mobility in Si, which also support the idea about the ballistic mechanism of the PL excitation, with the hot electrons exciting the surface emission centers.
References:
[1] A.G. Cullis, L.T.Canham, P.D.J.Calcott, The structural and luminescence properties of porous silicon, J.Appl.Phys., Vol. 82, 1997, pp. 909-963
[2]Y.Kanemitsu, H. Uto, Y. Masumoto, T.Matsumoto, H.Mimura and T.Futagi, Microstructure and optical properties of free standing porous silicon films: Size dependence of absorption spectra in Si nanometer-sized crystallites, Phys. Rev. B, v.48, 1993, 2827-2830.
[3] M.Rosenbauer, M.Stutzmann, H.D.Fuchs, S.Finkbeiner, J.Weber, Temparature dependence of luminescence in porous silicon and related materials, J.Luminescence, v.57, 1993, p.153-157.
[4] T.V.Torchynska, M.K.Sheinkman, N.E.Korsunska, L.Yu.Khomenkova, B.M.Bulakh, B.R.Dzhumaev, A.Many, Y.Goldstein, E.Savir, OH related emitting centers in interface layer of porous silicon, Physics B,, 273-274 1999, 955-958.
[5] S.M. Prokes, Light emission in thermally oxidized porous silicon. Evidence for oxide related luminescence, Appl. Phys. Lett. Vol. 62, 1993, pp. 3244-3246.
[6] T.V. Torchynska,M. Morales Rodriguez, F.G.Becerril Espinoza, L. Yu. Khomenkova, N.E.Korsunska, L.V.Scherbina, Phys. Rev. Vol.65, 2002, pp.1153131-1153137.
[7] T.V. Torchynska, Yu.V. Vorobiev, Ballistic effect and new concept of Si wire photoluminescence, Microel. Eng. Vol. 66, 2003, pp.17-25