Laser induced heating of Si nanocrystals

Journal of Non-Crystalline Solids 356 (2010) 1948–1950 Contents lists available at ScienceDirect Journal of Non-Crystalline Solids j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j n o n c r y s o l Laser induced heating of Si nanocrystals S. Gibilisco a,b,⁎, G. Faraci a,b, A.R. Pennisi a,b, A. Irrera b a b Dipartimento di Fisica e Astronomia, Universitá di Catania, Via Santa Sofia 64, 95123 Catania, Italy CNR-MATIS-Istituto Nazionale di Fisica della Materia, Via Santa Sofia 64, 95123 Catania, Italy a r t i c l e i n f o Article history: Received 18 May 2010 Available online 16 June 2010 Keywords: Silicon; Nanocrystals; Semiconductors; Quantum confinement; Raman spectroscopy a b s t r a c t Si nanocrystals were synthesized and deposited on HOPG substrates. Using micro-Raman spectroscopy, we demonstrate that significant laser heating may be induced on Si nanocrystals, leading to an increase in the local temperature. As a consequence, the Raman peak position and linewidth are modified with respect to bulk Si. We also discuss the possibility that a high temperature increase may favor observation of surface phonon modes. © 2010 Elsevier B.V. All rights reserved. 1. Introduction Since the first experimental evidence of strong room temperature Photoluminescence from porous Silicon [1], much research efforts have been devoted to Si based nanostructures, motivated by fundamental reasons and potential technological applications [2–5]. It is widely accepted that the optical properties of silicon nanocrystals are strongly dependent on both quantum confinement effects and structural and chemical surface characteristics [6–9] which have a dominant role in a nanoscale regime. As a consequence of the low dimensionality, the fundamental physical properties may be quite different from those of a bulk crystal. In particular, vibrational modes in confined systems are expected to undergo substantial modifications. Raman spectroscopy can be a very useful technique to investigate the vibrational properties of Silicon nanocrystals. In fact, in a perfect crystal, with a full translational symmetry, in the firstorder Raman scattering, it is only phonons from near the center of the Brillouin zone (q = 0) that are involved because of the momentum conservation law. However, this “selection rule” is relaxed at a nanoscale, where quantum confinement effects become dominant. It has been reported in various experiments [10–12] that quantum confinement effects in Silicon nanocrystals determine a size related Raman peak shift and broadening of the transversal optical (TO) phonon mode, situated for bulk Silicon at 521 cm− 1. These results are supported by theoretical quantum confinement models [13–16], although a full satisfactory agreement between experiments and theory is still lacking. However, it is worth noting that the peak shift and broadening may also be caused by high temperature Raman ⁎ Corresponding author. Dipartimento di Fisica e Astronomia, Universitá di Catania, Via Santa Sofia 64, 95123 Catania, Italy. E-mail address: santo.gibilisco@ct.infn.it (S. Gibilisco). 0022-3093/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jnoncrysol.2010.05.035 measurements [17,18]. The above considerations raise fundamental questions about the role played by possible local heating effects during micro-Raman spectroscopy measurements on Silicon nanocrystals. In fact, in nano-sized particles, heat dissipation is expected to be quite different from macroscopic samples. Moreover, in a typical micro-Raman experiment a laser power of several mW is focused on a spot of few microns of the diameter, and the actual local temperature may be significantly higher than the ambient one. In order to clarify the contribution of laser induced local heating to the Raman spectra of Silicon nanocrystals, we present a study as a function of laser power and local temperature, showing that the laser beam may induce a large increase in the temperature. The surface phonon properties are also taken into consideration to explain a double peak structure seen in some spectra. 2. Experiment Silicon nanocrystals were produced by gas phase cluster beam deposition, using the apparatus described in more detail elsewhere [19]. A beam of Helium, inseminated with silicon atoms, produced silicon nanocrystals, expanding in a supersonic configuration in a vacuum chamber at a pressure of 10−8 torr. After a high vacuum synthesis, the clean Si clusters were deposited, without a mass selection, on a Highly Oriented Pyrolitic Graphite (HOPG) substrate. Several layers were accumulated on the substrate, obtaining a low density, highly porous thin film. The passivation of the surface dangling bonds was obtained by the presence of a low oxygen exposure during cluster evaporation. After deposition, the samples were taken out of the vacuum chamber, and ex-situ Raman spectroscopy was applied for detecting the transversal optical vibrational peak, situated for bulk Si at 521 cm− 1, at room temperature. Both the emitted and absorbed phonon (Stokes and S. Gibilisco et al. / Journal of Non-Crystalline Solids 356 (2010) 1948–1950 1949 anti-Stokes lines) peaks were recorded on several spots of the samples. Micro Raman spectra were taken in backscattering geometries with a HORIBA Jobin-Yvon system, equipped with an Olympus BX41 optical microscope. The He–Ne laser radiation at a wavelength of 632.8 nm was focused to a spot size of about 3 μm by a 100× objective. The incident laser power on the sample was varied from 0.5 to 6.2 mW, and the scattered light was dispersed by a 550 mm focal length spectrometer with 1800 lines/mm grating and detected by a thermoelectrically cooled CCD. The structural characterization was performed by using a Zeiss Supra 25 field emission scanning electron microscope. 3. Results and discussion Scanning Electron Microscopy (Fig. 1) shows, at low magnification, the presence of micro-sized islands on top of a flat, nanogranular, deposition. Higher magnification reveals silicon nanograins of about 17 nm in diameter. It is important to note that, according to the theoretical quantum confinement models, confinement effects should not be significant in this size regime, and it is only a very slight peak shift and broadening that should be eventually observed. The Stokes and anti-Stokes Raman spectra of a typical sample, measured using different laser powers, are shown in Fig. 2a, together with bulk Si. The Raman peak position is always shifted with respect to bulk Si. Moreover, it is clearly evident that, with the increasing laser power, the Raman peak of Si nanocrystals shifts to lower wavenumbers and broadens. In Fig. 2 we also show both the Raman peak position and linewidth for silicon nanocrystals and bulk Si as a function of the laser power. As can be seen in the figure, bulk Si does not exhibit any Fig. 2. (a)(Colour on line) Stokes and anti-Stokes Raman spectra of Si nanocrystals, at different laser power. Raman spectrum of bulk crystalline Silicon c-Si) is also reported for comparison. (b) and (c) Raman peak position and linewidth as a function of laser power, showing the relevant peak shift and broadening for nc-Si, but not for bulk Si. Fig. 1. (a) Scanning Electron Microscopy of Silicon nanoclusters deposited on HOPG, showing a micro-sized island on top of a nanogranular film (b) High magnification of the nanogranular film, showing silicon nanograins. significant variation of the peak position and linewidth, in the range of the power used. On the contrary, for nanocrystal samples, a variation of the laser power determines low-shifted and broadened Raman peaks. These results suggest that, in this power regime, laser induced local heating takes place only in low sized particles. It is important to note that, with the decreasing laser power, the Raman line returns to 1950 S. Gibilisco et al. / Journal of Non-Crystalline Solids 356 (2010) 1948–1950 Fig. 3. Raman peak position as a function of temperature. The solid line is a guide to the eye. The uncertainty for the temperature is about 25 K. its initial position and linewidth, exhibiting reversible behavior. The Stokes/anti-Stokes intensity ratio, IS / Ia = exp(hv / kT), being v the Raman frequency may be used to obtain an evaluation of the actual local temperature T. The Raman peak position as a function of temperature is shown in Fig. 3. We observe that the Raman peak shifts to low wavenumbers with the increasing temperature. The local temperature can reach values up to about 800 K, corresponding to a peak position of 510 cm− 1. On the other hand, low temperatures (about 300 K) correspond to a peak position very close to the bulk one. The above findings strongly suggest that, at these sizes, the low shifted and widened Raman peak is mainly due to laser induced local heating effects, causing a dramatic change in the temperature of the nanograins. It is only a marginal contribution, if any, that can be ascribed to the quantum confinement effects. The high temperatures measured on nanocrystals samples, but not on bulk Silicon, may be explained considering the film morphology and the lack of a long range order. Indeed, the grain boundaries limit the phonon mean free path, and the boundary scattering in nanogranular Si thin films may become very effective. This results in a significant reduction of the thermal conductivity which makes heat dissipation difficult, leading to an increase in the local temperature. We also observe that the insulating oxide layer surrounding each nanograin may favor radiation trapping, with a further increase in temperature. Local heating is invoked to explain another effect observed in our experiments. When taken on microislands, the Raman peak splits in two components, both shifted and broadened with respect to bulk Si. The two components are indicated as low-shifted (L peak) and highshifted (H peak), taking as reference the unshifted peak of bulk Si at 521 cm− 1. We can attribute this double peak behavior to surface phonon modes, which are enhanced for low dimensional objects, with a high surface to volume ratio. Fig. 4 shows Raman spectra taken on a microisland as a function of laser power. At maximum power density, two well defined Raman peaks (L and H) are clearly visible. When reducing the laser power the H-peak gradually weakens until it disappears under the main L-peak. Surface modes have been calculated for reconstructed Si surfaces [20].A surface phonon mode at 64 meV was attributed to the dimer-bond stretching mode, very close to the 65 meV TO phonon of bulk Si. It is important to observe that the double peak appears at high power density only. This is probably due to the fact that the H and L peaks are very close at room temperature, and they cannot be easily resolved unless a significant temperature increase is induced. The calculated Fig. 4. (Colour on line) Raman spectra at different laser powers taken on a micro-sized island. The double peaks are obtained at higher powers, whereas only a single L peak is observed at 2 mW. The H-peak temperatures are 1130 K (4.5 mW) and 1260 K (6 mW). temperatures may be very high for the H peak from 900 K up to 2000 K, exceeding the melting point of bulk silicon. It is possible that the oxide layer surrounding the nanocrystals may favor radiation trapping. This makes the surface layer superheated with respect to the nanocrystal body, giving two well defined Raman features. In conclusion, we prepared Si nanocrystals thin films deposited on HOPG, demonstrating that the observed Raman peak shift and broadening should be mostly ascribed to laser induced local heating effects rather than to quantum confinement. Phonon scattering in nanocrystals can strongly limit heat dissipation leading to a significant increase in temperature. Moreover, we observed a double peak feature in the Raman spectra, vanishing at low laser power, which was attributed to surface phonon modes. References [1] L.T. Canham, Appl. Phys. Lett. 57 (1990) 1046. [2] M.V. Wolking, J. Jorne, P.M. Fauchet, G. Allan, C. Delerue, Phys. Rev. Lett. 82 (1999) 197. [3] L. Pavesi, L. Dal Negro, C. Mazzoleni, G. Franz, F. Priolo, Nature 408 (2000) 440. [4] I. Sychugov, R. Juhasz, J. Valenta, J. Linnros, Phys. Rev. Lett. 94 (2005) 087405. [5] H. Rong, R. Jones, A. Liu, O. Cohen, D. Hak, A. Fang, M. Paniccia, Nature 433 (2005) 725. [6] G. Ledoux, J. Gong, F. Huisken, O. Guillois, C. Reynaud, Appl. Phys. Lett. 80 (2002) 4834. [7] D.-Q. Yang, Jean-Numa Gillet, M. Meunier, E. Sacher, J. Appl. Phys. 97 (2005) 024303. [8] A. Podhorodecki, G. Zatryb, J. Misiewicz, J. Wojcik, P. Mascher, J. Appl. Phys. 102 (2007) 043104. [9] G. Faraci, S. Gibilisco, A.R. Pennisi, G. Franz, S. La Rosa, L. Lozzi, Phys. Rev. B 78 (2008) 245425. [10] M. Ehbrecht, B. Kohn, F. Huisken, M.A. Laguna, V. Paillard, Phys. Rev. B 56 (1997) 6958. [11] G.H. Li, K. Ding, Y. Chen, H.X. Han, Z.P. Wang, J. Appl. Phys. 88 (2000) 1439. [12] A.A. Sirenko, J.R. Fox, I.A. Akimov, X.X. Xi, S. Rumirov, Z. Liliental-Weber, Solid State Commun. 113 (2000) 553. [13] H. Richter, Z.P. Wang, L. Ley, Solid State Commun. 39 (1981) 625. [14] J.H. Campbell, P.M. Fauchet, Solid State Commun. 58 (1986) 739. [15] J. Zi, K. Zhang, X. Xie, Phys. Rev. B 55 (1997) 9263. [16] G. Faraci, S. Gibilisco, P. Russo, A.R. Pennisi, S. La Rosa, Phys. Rev. B 73 (2006) 033307. [17] M. Balkanski, R.F. Wallis, E. Haro, Phys. Rev. B 28 (1983) 1928. [18] M.J. Konstantinovic´, S. Bersier, X. Wang, M. Hayne, P. Lievens, R.E. Silverans, V.V. Moshchalkov, Phys. Rev. B 66 (2002) 161311(R). [19] A.R. Pennisi, P. Russo, S. Gibilisco, G. Compagnini, S. Battiato, R. Puglisi, S. La Rosa, G. Faraci, in: G. Mondio, L. Silipigni (Eds.), Progress in Condensed Matter Physics, Conference Proc., vol. 84, SIF, Bologna, 2003, p. 183; Eur. Phys. J. B 46 (2005) 457. [20] N. Takagi, S. Shimonaka, T. Aruga, M. Nishijima, Phys. Rev. B 60 (1999) 10919.