A silicon hierarchical structure, silicon nanoporous pillar array (Si-NPA), was prepared by a hydrothermal etching method. The architecture of Si-NPA was characterized to be a regular array of micron-sized, quasi-identical and nanoporous silicon pillars with an additional porous layer beneath the array. The pore walls were proved to be consisted of a SiOx matrix and dispersive silicon nanocrystallites. An integral reflectivity below 4% was achieved in the wavelength range of 240–2400 nm. Three photoluminescence bands, one blue and two red, were observed at room temperature and attributed to the recombination processes through band-to-band transition and luminescent centers, respectively. The structural and physical properties indicate that Si-NPA might be as both a functional silicon nanostructure and a template for assembling silicon-based nanocomposites in fabricating optoelectronic nanodevices.
©2008 Optical Society of America
Silicon nanostructures have attracted much attention in recent years because of their unusual quantum properties and potential applications [1–3], and the exhibited unique structural, optical and electronic properties have made them the most promising material systems in areas as diverse as optoelectronics , single electron devices , sensors , and cold cathodes for field-emission displays . One representative progresses achieved recently was the fabrication of high-performance field-effect transistors (FETs) , where n-type silicon nanowires with controlled phosphorus dopant concentration have been used successfully as building blocks for device constructing. All these achievements indicate that silicon nanostructures might play important roles in fabricating optoelectronic nanodevices in the future. In the past decade, numerous interesting silicon nanostructures, such as porous silicon (PS) [1, 8], the arrays of nanocone , nanopillars , nanorods , and nanowire p-n junction diodes , have been developed by various traditional or newly invented methods, and a lot of unique optical or electrical properties were obtained, but the space remained for constructing and developing novel silicon nanostructures is still tremendous.
In this paper, we report the preparation, structural and optical characterizations of a novel silicon micron/nanometer structural composite system, the silicon nanoporous pillar array (Si-NPA). Here Si-NPA was prepared by a hydrothermal etching method and was proved to be a silicon hierarchical structure with large-area structural regularity. The experiments carried out on the integrated diffuse reflection and photoluminescence (PL) disclosed that Si-NPA could exhibit the properties of very high light absorption and three-band PL. These structural physical properties indicate that Si-NPA might be a good candidate both as a functional silicon nanostructure and as a template for assembling silicon-based nanocomposites in fabricating optoelectronic nanodevices.
The initial silicon wafers used in this experiment were heavily boron-doped, (111) oriented single crystal silicon (sc-Si) wafers with a resistivity of 0.015 Ω cm. The hydrothermal etching process was accomplished in an autoclave. The whole preparation process is divided into five steps. (1) Clean the sc-Si wafers sequentially in ethanol and deionized water by ultrasonic bath method. (2) Prepare the hydrothermal solution by dissolving 0.002 mol analytical grade Fe(NO3)3·9H2O in 55 ml, 13 mol/l hydrofluoric acid aqueous solution. (3) Place the cleaned sc-Si wafers at the bottom of a 65 ml autoclave and then added the prepared hydrothermal solution into the autoclave. (4) Place the sealed autoclave into a furnace and keep the temperature at 130–150 °C for 50 minutes, and then cool down the autoclave naturally to room temperature. (5) Fetch out the samples from the autoclave and wash them with deionized water at suitable temperature, and then dry them with nitrogen blowing. The morphology and microstructure of Si-NPA were characterized by field emission scanning electron microscope (FE-SEM) (JEOL JSM-6700F), transmission electron microscope (TEM) and high resolution TEM (HRTEM) (Philips CM 200 FEG). The elemental compositions as well as the chemical valence states of Si-NPA surface were analyzed by X-ray photoelectron spectroscopy (XPS) (Axis Ultra, Kratos). The diffuse reflectance spectra of both Si-NPA and sc-Si wafers were measured at a wavelength range of 240–2400 nm by a Shimadzu UV-3150 spectrophotometer equipped with an integrating sphere using BaSO4 as reference. The PL and PLE spectra were detected by Hitachi F-4500 fluorescence spectrometer at room temperature.
3. Experimental results and discussion
After the hydrothermal etching process, the color of the sample surface changed from the original silver grey for mirror-polished sc-Si wafer to the final dark black for Si-NPA. Figure 1(a) presents the surface morphology of Si-NPA observed by FE-SEM, where a well-defined pillar array could be clearly observed. Judged from the FE-SEM images taken at different areas, the array of these pillars is fairly uniform all through the sample surface. Therefore it could be concluded that a large area, regular superstructure has been formed on the surface of as-prepared samples. Figure 1(b) shows the cross-sectional FE-SEM image of Si-NPA. It demonstrates that all the silicon pillars are well-separated and perpendicular to the sample surface, with a height of ~4.1 µm and a top-top distance of ~3.6 µm. Furthermore, it also discloses that beneath the pillar layer and above the sc-Si substrate, a ~1.0 µm transitional porous layer with neat and clear up and down boundaries was formed. Figure 1(c) is the amplified FE-SEM image of the transitional porous layer, through which the average pore size was estimated to be ~8.2 nm by statistics. In order to characterize the fine structure of the pillars, the pillar layer was carefully cleaved from the sample and was studied with TEM and HRTEM. Figure 1(d) is the TEM image of an individual silicon pillar, where large quantities of pores were observed. Deduced from the locally magnified TEM images, it could be found that the average size of the pores decreased gradually from top to bottom, ~40 nm at pillar tops and ~15 nm at pillar roots. Together with the size evaluation on the pores of the beneath transitional porous layer, it could be concluded that the average pore size decreased gradually with depth and this might be attributed to the difference contacting time for these sites with etching solution. Figure 1(e) presents a typical HRTEM image of a pore wall, where uniformly distributed crystal zones were observed. These crystal zones were proved to be the cross-sections of silicon nanocrystallites (nc-Si), whose size distribution obtained by statistics over 6 similar regions is depicted in Fig. 1(f). Clearly, the nc-Si sized from ~1.95 to ~4.4 nm, with an average size of ~3.4 nm and a planar density of ~2.3×1012cm-2.
The elemental composition and the chemical valence state of the components were analyzed by the experiment of XPS at room temperature. In the spectrum given in Fig. 2, three XPS peaks were observed, corresponding to the binding energy of Si 2p, Si 2s and O 1s, respectively. This indicates that the elements presented on the surface of Si-NPA are mainly Si and O. For analyzing the valance state of Si atom, the Si 2p peak was re-depicted at the localized horizontal ordinate scale and was found to be an asymmetrically broadened peak ranged from ~98 to ~106 eV, as shown in the inset of Fig. 2. Utilizing Gauss-Newton fitting method, the broad peak could be decomposed into three peaks, which locate at 99.7, 101.7 and 103.1 eV and corresponding to the valence states of Si0, Si2+ and Si4+, respectively . This indicates that the Si atoms presented in Si-NPA are in different valence states, which could be noted as Six+ with x=0, 2 and 4. On the other hand, the sharp O 1s peak locating at ~533.3 eV accurately confirms the O atoms are in the valence state of O2- . Combining with the structural characterizations given in Fig. 1(e), the Si atoms with Si0 valence state should be those constructing nc-Si, while those with Si2+ and Si4+ valence states should be in the states of silicon monoxide and silicon dioxide. Deduced from the area ratio between the peaks corresponding to Si2+ and Si4+ valence states, the quantities of silicon dioxide and silicon monoxide are almost equal. These results strongly indicated that large amount of SiOx (x<=2) has been formed natively in Si-NPA, most probably as SiOx shells encapsulating nc-Si cores. Based on the above experimental results, it could be concluded that Si-NPA is a micron/nanometer structural composite system characterized by hierarchical structure, i.e. a regular array composed of micron-sized, quasi-identical silicon pillars, nanopores densely distributing on the silicon pillars and in the beneath transitional porous layer, and nc-Si constructing the pore walls and encapsulated by SiOx shells.
As is known, antireflection is important for improving the efficiency of silicon-based solar cells, and presently, the construction of periodical surface structures has been proved to be one of the most effective ways . The dark black color and unique structural characteristics of Si-NPA imply that light reflectance suppression might be further enhanced by the densely distributed nanopores imposed on the periodical pillar array of Si-NPA. Figure 3 shows the integral diffuse reflection spectrum of Si-NPA, together with that of sc-Si for comparison. Obviously, all through the measuring wavelength range of 240 – 2400 nm, the reflectance of Si-NPA is greatly reduced, mostly above one order of magnitude compared with the counter part of sc-Si. Even comparing with the traditional sponge-like PS, the reflectance of Si-NPA is rather low. For example, in the wavelength range of 400 – 1000 nm, which is the concerned wavelength range in the field of solar cells, the average reflectivity of Si-NPA is less than 2.0 %, while that for typical PS is generally above ~5.0 % [16, 17]. Because the reflection peak positions are predominantly decided by the lattice structure of crystal silicon, the structure of the reflectance spectra between Si-NPA and sc-Si is very similar, as could be found in Fig. 3. Clearly, the strength of all the reflection peaks of sc-Si is reduced correspondingly for Si-NPA, but their reduction amplitudes are different. An obvious difference is that for Si-NPA, the reflectance strength at short wavelength scale is lower than that at long wavelength scale, which is opposite to what has been observed in sc-Si. For example, the strength of the strong reflection peaks locating at ~270 nm and ~1900 nm reduced from ~74 % to ~2.9 % and ~42 % to ~6.1 %, respectively, while that for the reflection peak locating at ~365 nm seems disappeared completely in Si-NPA. The broad-band, low reflectance of Si-NPA might be directly associated with the formation of the complex hierarchical structure. It is easy to deduce that when a branch of incident light arrives at the surface of Si-NPA, two kinds of reflection processes would occur. One is the multiple reflection amongst the micron-sized porous pillars, and the other is the reciprocating movements of light within the nanopores locating both on the silicon pillars and in the beneath transitional porous layer. It should be strengthened that the multiplicity of the character sizes of the hierarchical structure, ~3.6 µm for top-top distance between pillars, ~15-40 nm for pores on the pillars, and ~8.2 nm for pores in the beneath transitional porous layer, would be very helpful for the capture of the incident photons with different energies (wavelengths). On the other hand, the existence of large quantities of nc-Si in the pore walls as well as their broad size distribution would surely increase the absorption probability of the captured photons. Just as has been proved both experimentally and theoretically , the gap energy of nc-Si would increase correspondingly with decreased crystal size. Therefore an integral increment of light absorption was observed in Si-NPA.
Since the first observation of red emission in PS at room temperature , various silicon nanostructures with visible PL have been intensively investigated, mainly aimed at the potential applications in silicon-based optoelectronics. The room-temperature PL properties of Si-NPA were studied by changing the excitation wavelength (EW) from 340 to 430 nm, with a step of 30 nm. The obtained four PL spectra were depicted in Fig. 4(a). It was found that each PL spectrum has three emission bands, one in blue range and the other two in red range, which construct a triple-band spectral structure. Fitting each spectrum by Gauss-Newton method, the peak positions of the three emission bands could be determined correspondingly. For example, the PL spectrum under the irradiation of 340 nm ultraviolet could be decomposed into three separate emission bands, peaked at ~420 nm, ~640 nm and ~705 nm, respectively. The evolution of the PL peak positions and PL peak intensities with EW were demonstrated in Fig. 4(b), (c) and the insets of them. It is easy to find that in the tested EW scale, the peak intensities of the two red PL bands decrease monotonously with EW, while the peak positions as well as the full width at half maximum (FWHM) of the peaks remain almost unchanged (Fig. 4(b)). For the blue emission band, the PL peak intensity also decreases monotonously with EW, which is similar with the two red PL bands, but the PL peak position increases monotonously with EW (Fig. 4(c)). When EW was changed from 340 to 430 nm, a red shift of ~60 nm of the peak position was observed, from ~420 to ~480 nm.
To clarify the origin of the three emission bands of Si-NPA, the PL excitation (PLE) spectra of Si-NPA were measured by setting the monitoring wavelength at 420 nm, 640 nm and 705 nm, respectively (Fig. 5). It was found that all the three PLE spectra exhibit a peak pinned at ~365 nm, showing no dependence on the selection of the emission wavelength. This indicates that the most effective excitation wavelength for all the three emission bands is the same one and strongly suggests that the excitation process for the three emission bands might be an identical one. Deduced from the microstructure of Si-NPA and the research result on the PL mechanism of various silicon nanostructures [19–21], the PL excitation process in Si-NPA should be attributed to a band-band transition of carriers in nc-Si, as is illustrated in Fig. 6. Therefore the observation of the three separated emission bands could be only attributed to that there exist three different radiative recombination paths for the excited carriers. In addition, we should mention that there also exists a very weak feature at ~465 nm in the PLE spectra of Si-NPA (Fig. 5). It was proved to be caused by the leakage of the 467 nm light of the Xe lamp and has no relation with the PL excitation of Si-NPA, just as what has been often observed in the PLE measurements of PS or 3c-SiC nanocrystallines .
As was found in Fig. 4(c), with EW changed from 340 to 430 nm, the peak position of the blue emission band exhibits a red shift of ~60 nm, from ~420 to ~ 480 nm. Such a strong emission-excitation dependence in nanostructured silicon has been widely accepted as a sign for the band-band radiative recombination under the frame of quantum confinement (QC) effect [18,23]. According to QC model, the emission wavelength of nc-Si depends on its crystal size, i. e., the bigger the size, the longer the emission wavelength. Theoretical calculations based on the model demonstrated that with the crystal size of nc-Si tuned from 0.8 to 4.3 nm, its gap energy would varied from ~5-1.6 eV (~248.5–776.5 nm) [24,25]. As mentioned above, the smallest nc-Si in Si-NPA has been determined to be ~1.95 nm (Fig. 1(f)), the corresponding gap energy was thereafter evaluated to be ~2.95 eV (~421 nm). This gap energy is well in accordance with the peak energy of the blue emission band observed in experiments. It is natural to suppose that when Si-NPA was irradiated by a given wavelength, only the carriers confined in the nc-Si whose size is bigger than a critical size could be effectively excited. These excited carriers with high energy levels would fall down to the excitation states with relatively low energy levels of nc-Si through some nonradiative recombination process (procedure (a) in Fig. 6), and then transit back to the ground state through photon emissions (procedure (b)). This model well explained the phenomenon of the red shift of the blue emission band with increasing EW in Si-NPA. Therefore, we tend to conclude that the blue PL in Si-NPA originates from the QC effect of the carriers confined in variously sized nc-Si.
Different from the evolution behavior of the blue PL band, the peak positions of the two red PL bands remain almost unchanged when irradiated with different EW. This result directly denied the possibility to attribute the origin of the two red PL bands to pure QC model, because emission-excitation dependence is thought to be an essential deduction of QC model. In fact, similar evolutions for the peak positions of red PL bands have been observed and extensively studied in anodized PS and nc-Si-embedded silicon oxide, and it has been demonstrated that the corresponding PL mechanism could be well explained by a QC/luminescence center (QCLC) model [19,20,26]. According to this model, the excitation process also occurred through band-band transition in nc-Si, but the emission process mainly occurred through the luminescence centers (LCs) locating at the surface of nc-Si or in the SiOx layer encapsulating nc-Si [19,20,26]. Among the electrons excited to the high energy levels of nc-Si, excluding the electrons that participate in the procedure of the blue PL, the others would be responsible for the red PL. These carriers would firstly transit to the high-energy excitation states of the LCs locating at the surface of nc-Si or in the SiOx layer encapsulating nc-Si via diffusion or tunneling process (procedure (c) in Fig. 6), then fall down to low-energy excitation states of the LCs through some nonradiative recombination process (procedure (d)), and finally transit back to the ground state accompanied with photon emissions (procedures (e) and (f)). Because the energy levels of the LCs are generally independent of the size of nc-Si, the peak positions of the two red emission bands remain unchanged when irradiated by excited light with different wavelengths.
Based on the above discussion, the mechanisms of the blue emission band and the two red emission bands were attributed to QC model and QCLC model, respectively. In fact, the co-existence of the QC process and the QCLC process has been put forward in studying the PL mechanism of oxidized porous silicon and nc-Si-embedded silicon oxide , where the two processes were suggested to be a pair of competitive photoemission processes and most probably controlled by a critical size Lm. For example, Lm is ~2.0 nm for a nanoscale Si/Si oxide systems with the density of the LCs N LC=1015 cm-3 . In the PL of a silicon nanostructure, when the size of nc-Si is smaller than Lm, the QC process plays the major role; while when the size of nc-Si is bigger than Lm, the QCLC process plays the major role. In the case of Si-NPA, the nc-Si sized in the range ~1.95–4.4 nm and with an average size of ~3.4 nm, which indicates that the majority of nc-Si is bigger than one Lm (Fig. 1(f)). This indicates that the QCLC process would predominate over the PL process of Si-NPA. The experimental fact that the intensity of the two red PL bands is stronger than that of the blue band is therefore well explained.
According the above emission model for the red emissions, there should be two kinds of LCs to explain the origin of the two separate red emission bands (peaked at ~640 nm and 705 nm, respectively) observed in Si-NPA. The research by Carius et al. has disclosed that the PL peak wavelength of a high-quality SiOx thin film could change from visible to near-IR with x variation . For example, for a SiOx film with x=0.45, there will be an emission band peaked at ~745 nm . In the case of Si-NPA, the experimental data of XPS (the inset in Fig. 4) has shown that except the Si atoms in the valence state of Si4+, large quantities of Si atoms are in the valence state of Si2+, and this directly confirmed that the x value of the natively formed SiOx layer seriously deviated from 2. Considering the complexity of the hierarchical structure of Si-NPA, especially the large difference on the average size and depth difference between the nanopores distributing on the silicon pillars and in the transitional porous layer, it is reasonable to deduce that two kinds of SiOx films with different x values might have been formed because of the different oxidation degree. This would lead to the formation of two kinds of LCs and therefore two red emission bands were observed.
In conclusion, a novel silicon nanostructure named as Si-NPA was fabricated by hydrothermally etching sc-Si wafers in hydrofluoric acid containing ferric nitride. It was proved to be a hierarchical structure characterized by a regular array composed of micron-sized, quasi-identical silicon pillars, nanopores densely distributing on the silicon pillars and in the beneath transitional porous layer, and nc-Si constructing the pore walls and encapsulated by SiOx shells. Very high light absorption is realized in a broad wavelength range of 240–2400 nm and is attributed to the complexity of the hierarchical structure as well as its multiplicity of the character sizes. Triple PL bands, one blue band and two red bands, were observed in Si-NPA. Spectral analysis indicates that the excitation processes for the three emission bands are an identical one, which occurs in nc-Si through a carrier band-band transition process. The three emission bands observed in PL spectrum originate from the existence of three different recombination paths in Si-NPA, band-band recombination for blue PL and two kinds of LC-band recombination for red PL. The clarification on the origin of the three emission bands of Si-NPA might be important for controlling both the position and the relative intensity of the PL bands according to future device requirements. Our results indicate that Si-NPA might be a good candidate both as a functional silicon nanostructure and as a template for assembling silicon-based nanocomposites in fabricating optoelectronic nanodevices, such as novel solar cells, nanoheterojunctions, cold-cathode field emitters and nanosensors.
This work was supported by the National Natural Science Foundation of China (No. 10574112).
References and links
1. L. T. Canham, “Silicon quantum wire array fabrication by electrochemical and chemical dissolution of wafers,” Appl. Phys. Lett. 57, 1046–1048 (1990). [CrossRef]
2. M. W. Shao, Y. Y. Shan, N. B. Wong, and S. T. Lee, “Silicon nanowire sensors for bioanalystical applications: glucose and hydrogen peroxide detection,” Adv. Funct. Mater. 15, 1478–1482 (2005). [CrossRef]
3. J. Heitmann, F. Muller, M. Zacharias, and U. Gosele, “Silicon nanocrystals: Size matters,” Adv. Mater. 17, 795–803 (2005). [CrossRef]
4. J. M. Garguilo, F. A. M. Koeck, R. J. Nemanich, X. C. Xiao, J. A. Carlisle, and O. Auciello, “Thermionic field emission from nanocrystalline diamond-coated silicon tip arrays,” Phys. Rev. B 72, 165404 (2005). [CrossRef]
5. S. Fujita, K. Uchida, S. Yasuda, R. Ohba, and T. Tanamoto, “Novel random number generators based on Si nanodevices for mobile communication security systems,” Nanotechnology 3, 309–312 (2003).
6. M. Kanechika, N. Sugimoto, and Y. Mitsushima, “Field-emission characteristics of a silicon tip defined by oxygen precipitate,” J. Appl. Phys. 98, 054907 (2005). [CrossRef]
7. G. Zheng, W. Lu, S. Jin, and C. M. Lieber, “Synthesis and fabrication of high-performance n-type silicon nanowire transistors,” Adv. Mater. 16, 1890–1893 (2004). [CrossRef]
8. W. Sun, J. E. Puzas, T. J. Sheu, X. Liu, and P. M. Fauchet, “Nano- and microscale porous silicon as a cell interface for bone-tissue engineering,” Adv. Mater. 19, 921–924 (2007). [CrossRef]
9. N. G. Shang, F. Y. Meng, F. C. K. Au, Q. Li, C. S. Lee, I. Bello, and S. T. Lee, “Fabrication and field emission of high-density silicon cone arrays,” Adv. Mater. 14, 1308–1311 (2002). [CrossRef]
10. D. V. Scheible and R. H. Blick, “Silicon nanopillars for mechanical single-electron transport,” Appl. Phys. Lett. 84, 4632–4634 (2004). [CrossRef]
11. B. Cluzel, N. Pauc, V. Calvo, T. Charvolin, and E. Hadji, “Nanobox array for silicon-on-insulator luminescence enhancement at room temperature,” Appl. Phys. Lett. 88, 133120 (2006). [CrossRef]
12. K. Peng, Y. Xu, Y. Wu, Y. Yan, S. -T. Lee, and J. Zhu, “Aligned single-crystalline Si nanowire arrays for photovoltaic applications,” Small 1, 1062–1067 (2005). [CrossRef]
13. X. Y. Chen, Y. F. Lu, L. J. Tang, Y. H. Wu, B. J. Cho, X. J. Xu, J. R. Dong, and W. D. Song, “Annealing and oxidation of silicon oxide films prepared by plasma-enhanced chemical vapor deposition,” J. Appl. Lett. 97, 014913 (2005).
14. X. Cheng, Z.-D. Feng, and G.-F. Luo, “Effect of potential steps on porous silicon formation,” Electrochimica Acta. 48, 497–501 (2003). [CrossRef]
15. K. Hadobas, S. Kirsch, A. Carl, M. Acet, and E. F. Wassermann, “Reflection properties of nanostructure-arrayed silicon surfaces,” Nanotechnology 11, 161–164 (2000). [CrossRef]
16. S. Strehlke, S. Bastide, and C. Levy-Clement, “Optimization of porous silicon reflectance for silicon photovoltaic cells,” Sol. Energy Mater. Sol. Cells 58, 399–409 (1999). [CrossRef]
17. M. Lipinski, S. Bastide, P. Panek, and C. Levy-Clement, “Porous silicon antireflection coating by electrochemical and chemical methods for silicon solar cells manufacturing,” Phys. Stat. Sol. A 197, 512–517 (2003). [CrossRef]
18. M. V. Wolkin, J. Jorne, P. M. Faucher, G. Allan, and C. Delerue, “Electronic States and Luminescence in Porous Silicon Quantum Dots: The Role of Oxygen,” Phys. Rev. Lett. 82, 197–200 (1999). [CrossRef]
19. G. G. Qin and Y. J. Li, “Photoluminescence mechanism model for oxidized porous silicon and nanoscale-silicon-particle-embedded silicon oxide,” Phys. Rev. B 68, 085309 (2003). [CrossRef]
20. J. Q. Duan, H. Z. Song, G. Q. Yao, L. Z. Zhang, B. R. Zhang, and G. G. Qin, “Variation of double-peak structure in photoluminescence spectra from porous silicon with excitation wavelength,” Superlattices Microstruct. 16, 55–58 (1994). [CrossRef]
21. P. Li, G. Z. Wang, Y. R. Ma, and R. C. Fang, “Origin of the blue and red photoluminescence from aged porous silicon,” Phys. Rev. B 58, 4057–4065 (1998-I). [CrossRef]
22. X. L. Wu, J. Y. Fan, T. Qiu, X. Yang, G. G. Siu, and P. K. Chu, “Experimental evidence for the quantum confinement effect in 3c-SiC nanocrystallines,” Phys. Rev. Lett. 94, 026102 (2005). [CrossRef] [PubMed]
23. Md. N. Islam and S. Kumar, “Influence of surface states on the photoluminescence from silicon nanostructures,” J. Appl. Phys. 93, 1753–1759 (2003). [CrossRef]
25. Y. Kanemitsu, H. Uto, Y. Masumoto, T. Matsumoto, T. Futagi, and H. Mimura, “Microstructure and optical properties of free-standing porous silicon films: Size dependence of absorption spectra in Si nanometer-sized crystallites,” Phys. Rev. B 48, 2827–2830 (1993-II). [CrossRef]
26. D. W. Cooke, B. L. Bennett, E. H. Famum, W. L. Hults, K. E. Sickafus, J. F. Smith, J. L. Smith, T. N. Taylor, P. Tiwari, and A. M. Porits, “SiOx luminescence from light-emitting porous silicon: Support for the quantum confinement/luminescence center model,” Appl. Phys. Lett. 68, 1663–1665 (1996). [CrossRef]
27. R. Carius, R. Fischer, E. Holzenkampfer, and J. Stuke, “Photoluminescence in the amorphous system SiOx,” J. Appl. Phys. 52, 4241–4243 (1981). [CrossRef]