Effect of material structure on photoluminescence spectra from silicon nanocrystals

S. M. Orbons, M. G. Spooner, and R. G. Elliman Citation: J. Appl. Phys. 96, 4650 (2004); doi: 10.1063/1.1790058 View online: http://dx.doi.org/10.1063/1.1790058 View Table of Contents: http://jap.aip.org/resource/1/JAPIAU/v96/i8 Published by the American Institute of Physics.

Related Articles
Upper limit of two-dimensional hole gas mobility in strained Ge/SiGe heterostructures Appl. Phys. Lett. 100, 222102 (2012) Shape evolution in glancing angle deposition of arranged Germanium nanocolumns J. Appl. Phys. 111, 104309 (2012) Dopant effects on the photoluminescence of interstitial-related centers in ion implanted silicon J. Appl. Phys. 111, 094910 (2012) Nano-electron beam induced current and hole charge dynamics through uncapped Ge nanocrystals Appl. Phys. Lett. 100, 163117 (2012) Experimental and theoretical analysis of the temperature dependence of the two-dimensional electron mobility in a strained Si quantum well J. Appl. Phys. 111, 073715 (2012)

Additional information on J. Appl. Phys.

Journal Homepage: http://jap.aip.org/ Journal Information: http://jap.aip.org/about/about_the_journal Top downloads: http://jap.aip.org/features/most_downloaded Information for Authors: http://jap.aip.org/authors

Downloaded 07 Jun 2012 to 59.162.23.73. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

JOURNAL OF APPLIED PHYSICS

VOLUME 96, NUMBER 8

15 OCTOBER 2004

Effect of material structure on photoluminescence spectra from silicon nanocrystals

S. M. Orbons, M. G. Spooner, and R. G. Ellimana)
Electronic Materials Engineering Department, Research School of Physical Sciences and Engineering, Australian National University, Canberra, ACT 0200, Australia

(Received 30 March 2004; accepted 16 July 2004) The thickness of a silicon dioxide layer is shown to have a signicant effect on the measured photoluminescence spectrum from silicon nanocrystals embedded in the layer. This can range from signicant but subtle spectral distortions that are difcult to detect, including changes in intensity, peak position and peak width, to gross distortions that are readily apparent as a modulation of the measured spectrum. These distortions are shown to be a simple consequence of the wavelength dependent reectivity of the sample structure but to have important implications for determining nanocrystal properties from such data. 2004 American Institute of Physics. [DOI: 10.1063/1.1790058] Bulk silicon generally exhibits weak luminescence at room temperature due to its indirect band gap and the dominance of nonradiative recombination processes.1 However, it has been shown2,3 that silicon based nanostructures can exhibit efcient photoluminescence at room temperature. This discovery has sparked considerable debate regarding the mechanism for improved emission efciency and an intrinsic element of this debate has been the comparison of experimentally measured photoluminescence (PL) spectra with theoretical predictions based on various models.46 Such analysis is predicated on the assumption that measured PL spectra reect the properties of the silicon nanostructures. However, the broad spectral emission that is typically observed from Si nanocrystals can be distorted by the spectral response of the measurement system as well as by the optical properties of the sample.79 A broad range of material structures have been employed by researchers studying Si nanocrystals, each with a characteristic, wavelength-dependent reectivity.79 Indeed, the optical properties of such structures are often employed to modify the nanocrystal emission, as exemplied in microcavity structures employing distributed Bragg mirrors.1013 However, in other cases, the impact of the sample structure on the measured nanocrystal emission is unintentional and often misleading. The spectral distortion due to simple material structures is a much more subtle and often misinterpreted variance in the observed nanocrystal emission, and one that is of critical importance when comparing PL spectra from different structures, even when they differ only slightly. In this study, commercially prepared 100 oriented silicon wafers were oxidized to produce SiO2 layers of 5 m, 970 nm, 650 nm, and 103 nm. These, together with a 1 mm thick fused silica plate were implanted at room temperature with 30 keV Si ions to a uence of 2.5 1016 ions/ cm2. Figure 1 shows the Si depth distribution resulting from the implant. Each sample was subsequently annealed at 1050 C
a)

for 1 h in a forming gas 95% N2 , 5 % H2 ambient. By keeping both the implantation and annealing conditions constant for each sample, it is expected that they will have the same Si concentration-depth prole and the same nanocrystal size distribution. The only signicant parameter that differs from sample to sample is the thickness of the oxide layer as illustrated schematically in Fig. 2. PL spectra were collected at room temperature using a 532 nm Rumzing diode pumped solid state laser as the pump source (incident angle 14) and a TRIAX 320 spectrometer with a liquid nitrogen cooled SpectrumOne charge-coupled device as the detection system. The PL emission was collected with f4 optics consisting of two matched plano-convex lens of 50 mm diameter and 200 mm focal length, giving a collection angle of up to 7. Reectivity measurements were performed at an incident angle of 5 using a Shimadzu UV-3101PC spectrophotometer with a specular reectance attachment (P/N 206-14046). The bandwidth for the illumination system was set at 5 nm. Figures 3(a)3(d) show both the normalized PL intensity and the measured reectance for each sample as a function of wavelength. It is immediately apparent that the measured PL spectra vary signicantly as a function of oxide thickness, with variations in peak position, width and structure clearly evident. Inspection of panel 3(a) indicates that the fused silica sample yields an approximately constant reectance over the entire wavelength range of interest. The PL spectrum is therefore not expected to be distorted signicantly by

Author to whom correspondence should be addressed; electronic mail: rob.elliman@anu.edu.au 4650

FIG. 1. Depth distribution of implanted ions as calculated using transport of ions in matter TRIM software.20 2004 American Institute of Physics

0021-8979/2004/96(8)/4650/3/$22.00

Downloaded 07 Jun 2012 to 59.162.23.73. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

J. Appl. Phys., Vol. 96, No. 8, 15 October 2004

Orbons, Spooner, and Elliman

4651

FIG. 2. Schematic representation of typical sample structures under investigation. The implantation and annealing condition are the same for all samples. Only the thickness of the oxide layer is different for each sample.

the material structure, even though the PL intensity will still be affected. The strong correlation between the reectivity of the material structure and the measured PL spectra is most readily apparent for the 5 m thick SiO2 layer, shown in Fig. 3(d). In this case the PL spectrum is clearly modulated by the reectivity of the structure resulting in an obvious distortion of the emission spectrum. Figures 3(b) and 3(c) show much more subtle distortions. For example, when compared to the spectrum from the silica sample, the PL spectrum from the 103 nm SiO2 layer, Fig. 3(b), shows higher peak intensity but a similar full width at half maximum (FWHM) (143 nm, compared to 147 nm observed for silica). On the other hand the spectrum from the 970 nm SiO2 layer, Fig. 3(c), exhibits two obvious peaks as a direct consequence of the spectral distortion caused by the

FIG. 4. Photoluminescence (bold line) and reectivity (thin line) spectra from different sample structures: (a) 103 nm oxide layer, (b) 650 nm oxide layer.

FIG. 3. Photoluminescence (bold line) and reectivity (thin line) spectra from different sample structures: (a) fused silica sample, (b) 103 nm oxide layer, (c) 970 nm oxide layer, (d) 5 m oxide layer.

surrounding material structure, with the main peak being centered at a wavelength of 780 nm compared to 740 nm recorded for the silica sample. Such a peak shift can readily lead to misinterpretation of the mean nanocrystal size.8,14,15 For example, using a simple expression for the relationship between the PL emission energy and nanocrystal diameter16 suggests that PL peaks at 780 nm and 740 nm correspond to nanocrystals with mean diameters of 4.8 or 4.1 nm, respectively. The often subtle distortion of spectra is further highlighted in Fig. 4 which compares PL spectra from the 103 nm and 650 nm thick SiO2 layers. The PL spectrum from the latter shows an increase in peak intensity as well as a considerable reduction in FWHM from 143 nm to 103 nm for the thicker layer. Assuming inhomogeneous broadening of the emission this corresponds to a change in the standard deviation of the nanocrystal size distribution from 0.3 nm for the narrower peak to 0.6 nm for the wider peak.5 It is interesting to note that this difference occurs despite the fact that the nanocrystal preparation conditions are identical, the only difference between the two samples being an additional 547 nm of oxide. Clearly, the assumption that the width of PL spectra from Si nanocrystals results primarily from inhomogeneous broadening due to the size distribution of nanocrystals assumes that due care has been taken to account for effects such as those illustrated in Figs. 3 and 4. The data in Fig. 3 and 4 highlight the relationship between the observed PL and sample reectance, demonstrating that this is the dominant effect leading to spectral distortion. Assuming that the structure is nonabsorbing, the transmission and reection characteristics of a given structure are inversely related (i.e., R + T = 1, where R is the reectivity and T the transmissivity of the structure). Hence, a local maximum in reectivity corresponds to a local minimum in transmissivity and therefore to a minimum in the measured PL emission. A rst-order estimate of the position of reectivity maxima can be made from a simple constructive interference model, with max = 2nt / m + 1 / 2, where max is the wavelength for maximum reectivity, n is the refractive index of the SiO2 layer n 1.46, t is the lm

Downloaded 07 Jun 2012 to 59.162.23.73. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

4652

J. Appl. Phys., Vol. 96, No. 8, 15 October 2004

Orbons, Spooner, and Elliman

thickness, and m is the interference order. (Comparison of the PL and reectivity spectra reveals a small phase shift between the reectance and PL spectra, most obvious in Fig. 3(d), an effect that is attributed to the different illumination conditions employed for each measurement.) It should be noted that it is not always trivial to predict the reectivity of the material structure as this depends not only on the initial material structure but also on the distribution and concentration of implanted silicon. By using the calculated implant prole, together with Maxwell-Garnet effective medium theory17 and the Fresnel equations,18 a reasonable estimate of the reectivity can be achieved.19 However, such analysis conrms the signicant role of the implanted silicon distribution in determining interference effects and highlights its inuence on measured reection and absorption characteristics.19 (In previous measurements19 we have shown that this can lead to misinterpretation of optical absorption data.) Also implicit in the above discussion is the fact that the PL intensity and distribution will depend on the angle of detection, and on the wavelength and angle of incidence of the probe beam. In conclusion, it has been shown that the PL emission from Si nanocrystals can be distorted by the optical properties of the sample structure, even in cases where that structure is relatively simple. These distortions affect the intensity, peak position, and width of the PL spectra and it is vital that such distortion be taken into account both when comparing PL spectra from different structures and when extracting

nanocrystal parameters from observed PL spectra.

D. Kovalev and H. Heckler, Phys. Status Solidi B 215, 871 (1999). H. Takagi, H. Ogawa, Y. Yamazaki, A. Ishizaki, and T. Nakagiri, Appl. Phys. Lett. 56, 2379 (1990). 3 L. T. Canham, Appl. Phys. Lett. 57, 1046 (1990). 4 A. Kux and M. B. Chorin, Phys. Rev. B 51, 17535 (1995). 5 V. Ranjan, V. A. Singh, and G. C. John, Phys. Rev. B 58, 1158 (1998). 6 G. Ledoux, O. Guillois, D. Porterat, C. Reynaud, F. Huisken, B. Kohn, and Paillard, Phys. Rev. B 62, 15942 (2000). 7 F. Iacona, G. Franzo, E. Moreira, D. Pacici, A. Irrera, and F. Priolo, Mater. Sci. Eng., C 19, 377 (2002). 8 D. Lockwood and Z. Lu, Phys. Rev. Lett. 76, 539 (1996). 9 D. Lockwood and B. Sullivan, J. Lumin. 80, 75 (1999). 10 J. Bellessa, S. Rabaste, J. Plenet, J. Dumas, J. Mugnier, and O. Marty, Appl. Phys. Lett. 79, 2142 (2001). 11 L. Pavesi, G. Panzarini, and L. Andreani, Phys. Rev. B 58, 15794 (1998). 12 V. Pellegrini, A. Tredicucci, C. Mazzoleni, and L. Pavesi, Phys. Rev. B 52, 14328 (1995). 13 E. F. Schubert, Appl. Phys. Lett. 61, 1381 (1992). 14 M. Brongersma, A. Polman, K. Min, E. Boer, T. Tambo, and H. Atwater, Appl. Phys. Lett. 72, 2577 (1998). 15 R. Soni, L. Fonseca, O. Resto, M. Buzaianu, and S. Weisz, J. Lumin. 83, 187 (1999). 16 G. Ledoux, O. Guillois, D. Porterat, C. Reynaud, F. Huisken, B. Kohn, and V. Paillard, Phys. Rev. B 62, 15942 (2000). 17 C. Bohren and D. Huffman, Absorption and Scattering of Light by Small Particles (Wiley, New York, 1983). 18 M. Born and E. Wolf, Principles of Optics (Pergamon, New York, 1959). 19 R. Elliman, M. Lederer, and B. Luther-Davies, Appl. Phys. Lett. 80, 1325 (2002). 20 J. F. Ziegler, J. P. Biersack, and U. Littmark, The Stopping and Range of Ions in Solids (Pergamon, New York, 1985).
2 1

Downloaded 07 Jun 2012 to 59.162.23.73. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions