Open Access

A multilayered approach of Si/SiO to promote carrier transport in electroluminescence of Si nanocrystals

  • Ding Li1,
  • Yong-Bin Chen1,
  • Yong Ren1,
  • Jiang Zhu1,
  • You-Yuan Zhao1 and
  • Ming Lu1Email author
Nanoscale Research Letters20127:200

DOI: 10.1186/1556-276X-7-200

Received: 16 November 2011

Accepted: 27 March 2012

Published: 27 March 2012


The electroluminescence (EL) and photoluminescence of Si nanocrystals (Si-nc) from multilayered samples of Si/SiO are investigated. Si-nc are formed within Si and SiO layers after furnace annealing. It is found that the presence of Si interlayers creates extra carrier paths for EL emission. A comparative study is further performed on a multilayered Si/SiO sample and a single-layered one with Si and SiO homogeneously mixed. Both samples have the same ratio of Si to O and the same contents of Si and O. The multilayered sample is found to have higher EL intensity, less turn-on voltage, lower resistance, and higher current efficiency than the single-layered one. The results indicate that Si interlayers in Si/SiO may act as carrier channels, which promote carrier transport and enhance the EL emission of Si-nc.


Development of highly luminescent Si is of vital importance for Si optoelectronics [1]. At present, Si nanocrystals (Si-nc) embedded in SiO2, or SiO2:Si-nc, have been widely accepted as a promising material for Si light emission due to their stable light emission, robust structure, and stimulated emission feature (JZ and ML, unpublished data) [126]. Researchers have been focused on the enhancement of photoluminescence (PL) of Si from the SiO2:Si-nc [3, 6, 7, 11, 20, 24] and that of electroluminescence (EL) of Si [1419, 21, 25]. One of the key problems regarding the enhancement of EL emission is how to promote the carrier (electron or hole) transport [21, 25, 26]. The Si3N4 :Si-nc has once been adopted since the bandgap width of Si3N4 is less than that of SiO2, so carrier transport within Si3N4 would become easier than that within SiO2 [25]. For the same reason, lateral carrier injection has been suggested recently for the sample of SiO2:Si-nc/SiO2 [21]. However, a matrix with a small bandgap width does not favor carrier confinement within Si-nc, which tends to degrade light emission of Si-nc. For instance, it is found that the PL intensity of Si-nc from amorphous Si:Si-nc/SiO2 is one order of magnitude less than that of SiO2:Si-nc/SiO2 (JZ and ML, unpublished data). In this work, we propose another approach to improve carrier transport, that is, an approach of multilayered structure of Si/SiO. After thermal annealing, Si-nc are formed in both Si and SiO layers. Through the investigation of PL and EL emissions from multilayered Si/SiO, it is found that Si interlayers help to facilitate carrier transport. Further comparative studies on a multilayered Si/SiO sample and a single-layered one, with Si and SiO homogeneously mixed, indicate that Si interlayers in Si/SiO act as carrier channels that promote carrier transport and enhance the EL emission of Si-nc.


Preparation of samples

Forty-layered Si/SiO samples were prepared by evaporating Si and SiO alternatively by electron-beam and resistance heating, respectively, onto p+-type Si substrates (0.5 to approximately 1 Ω·cm) in a vacuum chamber with a base pressure less than 5 × 10-5 Pa. An ex-situ furnace annealing in nitrogen atmosphere at 1,100°C was performed for 1 h to induce phase separation (JZ and ML, unpublished data) [3, 4, 7, 11] to form Si-nc. The thickness of SiO remained at 3.75 nm, and that of Si interlayer (dSi) was variable.

For a comparative study, we chose a 40-layered Si/SiO sample with thicknesses of SiO and Si layers being 3.75 and 1.00 nm, respectively, and a single-layered sample prepared by co-evaporating Si and SiO with the same ratio of Si to O and the same contents of Si and O as those of the multilayered sample. For further EL identification, a pure SiO2 sample with a thickness of 150 nm was prepared by electron beam evaporation, followed by furnace annealing at 1,100°C in nitrogen for 1 h.

For EL emission, Al was firstly evaporated on the backside of the p+-type Si substrate of the SiO2:Si-nc; then, the sample was annealed ex situ in nitrogen at 480°C for 10 min. A ring of Al electrode was finally evaporated onto the top surface of sample.

Monitoring, characterization and measurements of samples

Film thickness was monitored by a calibrated microbalance (MDC-360, Maxtek Digicom Limited, Shenzen, China). The PL spectra were recorded on a fluorescence spectrometer (F-4500, Hitachi High-Tech, Minato-ku, Tokyo, Japan) with a xenon lamp as an excitation source. The wavelength of the selected exciting beam was 300 nm. Forward-biased voltage was applied to the EL device. EL emission was also recorded with the spectrometer (Hitachi, F-4500). The biased voltage as well as voltage and current readings were provided by a source meter (2400, Kiethley Instruments Inc., Cleveland, OH, USA). High-resolution transmission electron microscopy (HRTEM) was measured on an electron microscope (G2 F20, Tecnai, Amsterdam, The Netherlands) to examine the formation of Si-nc. Reflectance spectra were measured with a spectrophotometer (UV-3101 PC, Shimadzu Corporation, Nakagyo-ku, Kyoto, Japan). In this study, all the samples refer to those that have been furnace-annealed at 1,100°C for 1 h in nitrogen.

Results and discussion

Figure 1 shows the evolutions of PL and EL intensities of Si-nc from Si/SiO as functions of dSi. In addition to the number of radiative centers, the EL intensity depends on other parameters including applied voltage, temperature change, injected current, and the number of non-radiative centers that are intrinsically present and extrinsically produced during carrier injection. The EL intensity increases at first with the increasing voltage (hence the increasing current), and after a certain voltage, the EL intensity starts to drop down due to the overwhelming effect of non-radiative centers induced by carrier injection [8, 9]. Figure 2 gives an example. The shift of peak position with the voltage has been attributed to the size distribution of Si-nc [9]. The EL spectra thus selected are those that possess the highest intensities during the biased voltage ramping. The following EL intensities will all refer to this type of situation unless otherwise stated. It is seen that with increasing dSi, the PL intensity increases at first, and after reaching a maximum, it decreases. The rise of PL intensity could be attributed to the additional Si-nc formed due to the presence of Si interlayers [24, 27]. With further increase in dSi, the number of 'large' Si crystallite with size greater than the Si Bohr's radius (4.3 nm) [28] increases in both Si and SiO layers due to the ripening process [29], which consumes the excess Si content but contributes little to PL emission, leading to a decrease in PL intensity. Detailed description and identification of the evolution process of PL has been presented in the study of Ren et al. [24]. A similar trend also exists for the EL evolution but with a greater optimized dSi. The inset of Figure 1 gives the turn-on voltage as a function of dSi. The trend indicates that with the increasing dSi, extra conductive percolation paths of carriers for EL are created, which promotes carrier transport. To compare the evolutions of EL and PL intensities with increasing dSi more sensibly, reflection (R) correction is performed for the PL intensity, that is, each PL intensity is multiplied with 1/(1 - R), where R is the measured reflectance at λ = 300 nm, as shown by the dashed curve in Figure 1. As compared to the PL emission, the EL intensity depends not only on the density of Si-nc, but on the carrier transport as well. The difference in the maximal PL and EL intensity in Figure 1 arises from their different excitation processes, sample geometries, and measuring conditions. The larger optimized value of dSi of EL could be due to a compromise between the decreasing density of Si-nc as reflected by the PL intensity and the increasing carrier transport as suggested by the result of the turn-on voltage in the inset of Figure 1.
Figure 1

The PL and EL intensities of Si/SiO samples versus thickness of Si interlayer, d Si . The dashed curves are for PL intensities after R correction. The inset gives the turn-on voltage versus dSi.
Figure 2

The EL spectra with different biased voltages for one sample of Si/SiO.

In the discussion of the trend of EL intensity in Figure 1, it has been implicitly assumed that the EL is from Si-nc. However, the origin of EL is complex. At present, it is generally accepted that the origin of EL is related to oxygen vacancy-related defects within SiO2, the interface states of Si-nc, and Si-nc themselves [8, 1421]. Figure 3 shows the PL and EL spectra of a Si/SiO sample with a dSi of 1.00 nm. With respect to the PL spectrum of Si-nc, the EL has a broad spectral width and a large blueshift, which can be ascribed to the effects of direct electron tunneling [8] and band filling [19], and various defect states. For further identification, in Figure 3, the PL and EL spectra from a pure SiO2 sample with a thickness of 150 nm after annealing at 1,100°C are given. The total thickness of SiO was 150 nm, too. The structure peaking at 655 nm in the PL spectrum is known from the oxygen vacancy-related defects within SiO2 such as non-bridging oxygen hole centers and the neutral oxygen vacancies (see the work of Lin et al. [8] and references therein). The EL of the SiO2 sample differs from that of the Si/SiO, namely, the overall peak has a blueshift of approximately 45 nm, and the intensity is only approximately 33% of Si/SiO. Therefore, it is concluded that the contribution from the defect in SiO2 to the EL of Si/SiO is of minor importance, and those from Si-nc and Si-nc-related defects are dominant in the EL emission. We notice that for SiO2, the EL also has a broad spectral width and blueshift with respect to the PL, indicating that effects of direct electron tunneling and band filling might be at work, too.
Figure 3

The PL and EL spectra from one Si/SiO sample and the sample of SiO 2 .

To further explore the role played by the multilayered structure in promoting carrier transport, two samples are studied. One is a 40-layered sample of Si/SiO termed as 'multi', with thicknesses of SiO being 3.75 nm and dSi 1.00 nm; the other is a single-layered one termed as 'single' with Si and SiO homogeneously mixed, followed by furnace annealing. For both samples, the ratio of Si to O and the contents of Si and O are identical. Figure 4 shows that the PL intensity from multi is only approximately 0.4 times that of single after R correction. Hence, the density of Si-nc is supposed to be greater in single, which is confirmed by the HRTEM results in Figure 5a, b. Two typical images for single and multi are presented here, and the densities of Si-nc with size no greater than 5.0 nm are 9.0 ± 3.4/(46 × 46 nm2) for single and 6.7 ± 1.6/(46 × 46 nm2) for multi as averaged over a dozen of HRTEM images. However, in Figure 4, the EL intensity of multi is approximately 2.1 times that of single. The current efficiency defined as the maximal EL intensity divided by the current density is 597.8 mA cm-2 for multi, and 483.9 mA cm-2 for single. Also, mean resistances of both samples are measured and calculated to be 84.3 ± 10.2 Ω for multi and 117.8 ± 8.7 Ω for single. Furthermore, in the inset of Figure 4, the EL intensity versus the bias voltage is given. The derived turn-on voltage for multi is 5.7 V, and that of single is 7.2 V. All these results indicate that although the two samples possess the same ratios of Si to O and the same contents of Si and O, with the number of Si-nc of single even larger than that of multi, the sample of single with excess Si randomly distributed is less efficient in carrier transport than the sample of multilayered Si/SiO. Let us compare the EL efficiency of the two samples. For multi, the input power is 15 V × 195 mA = 2.9 J; for single, it is 16 V × 150 mA = 2.4 J. The EL efficiency is defined as the EL emission energy divided by the input energy. Considering that the EL intensity of multi is 2.1 times that of single, the EL efficiency of multi is thus 2.1/(2.9 × 2.4) = 1.74 times that of single. In fact, the number of Si-nc in multi is approximately 0.4 times that of single as we estimated from the PL results of Figure 4 and the data of Figure 5; the EL efficiency of Si-nc of multi is therefore 1.74/0.4 = 4.3 times that of single.
Figure 4

The R- corrected PL and EL spectra from 'multi' and 'single'. The inset shows the trends of EL intensity versus bias voltage for the two samples.
Figure 5

HRTEM images for the samples. 'Single' (a) and 'multi' (b). (c) The TEM of multi with low magnification, and (d) a Si-nc image.

It is known that Si-nc are also formed in the Si interlayers after thermal annealing [24, 27], which could be clearly seen for thicker dSi [24]. In Figure 5b, the arrows indicate the positions of Si interlayer. Figure 5c gives a TEM of multi with a lower magnification, and Figure 5d shows a single Si-nc. Considering that the energy levels of various Si crystallites in Si interlayers are close and their bandgap widths are much smaller than that of SiO2, carriers injected into Si interlayers would move more freely within the layers than in SiO2 matrix. With this in mind, a scenario is supposed in Figure 6. Two diagrams of carrier (electron for instance) transport within the single and multi samples are given, respectively, under a biased electric field. Electrons injected from the electrode on the left hand side, for instance, tunnel through the SiO2 barriers and reside on Si-nc. On the other hand, holes are injected from the p+-type Si on the right hand side. In the single sample, further transport of electron would be hindered by sufficiently wide barriers of SiO2 before it reaches the next Si-nc as illustrated in Figure 6a by the dashed arrows. However, when the electron from a Si nanocrystal within SiO2 faces a Si interlayer, it would enter the Si layer readily and then move within the layer either 'upward' or 'downward' until it encounters a Si nanocrystal in the SiO2 layer and jumps onto it via direct tunneling under the bias. This process could continue until the electron meets and recombines a hole that has been injected from the opposite direction at a certain Si-nc to give off EL emission. Thus, the Si interlayers in multi act as a kind of carrier channels, and the carriers may jump from one Si-nc to another efficiently in the manner as illustrated in Figure 6b. Figure 6 explains why the efficient carrier transport in multi is more efficient than in single. In fact, in the work of Berghoff et al. [26], they also studied the charge transport within Si/SiO x , although the EL emission of Si-nc was disregarded there. They found that the charge transport is improved in Si/SiO x as compared with that in Si/SiO2. They attributed this improvement to the narrowing of the barrier width within the SiO x layer because of the formation of Si-nc, a model similar to that in Figure 6.
Figure 6

Diagrams of carrier (electron for example) transport. This carrier transport occurred in the 'single' (a), and 'multi' with Si layers as carrier channels (b).


We report that a multilayered approach of Si/SiO can promote carrier transport for EL emission by establishing carrier channels made of Si interlayers. To promote carrier transport along the normal of the thin film surface within SiO2: Si-nc, one normally tries to align Si-nc closely along the direction of the electric field so that carrier tunneling can become easier. However, how to align the self-organized Si-nc remains an open problem so far. The approach of a multilayered structure of Si/SiO proposed here might, to some degree, circumvent this problem and help to enhance the EL emission of Si-nc.



This work was supported by the National Basic Research Program of China (973 Program) under grant number 2010CB933703 and the National Science Foundation of China under grant numbers 60878044 and 60638010.

Authors’ Affiliations

Department of Optical Science and Engineering, and Shanghai Ultra-Precision Optical Manufacturing Engineering Center, Fudan University


  1. Yuan Z, Anopchenko A, Daldosso N, Guider R, Navarro-Urrios D, Pitanti A, Spano R, Pavesi L: Silicon nanocrystals as an enabling material for silicon photonics. Proceedings of the IEEE 2009, 97: 1250.View Article
  2. Pavesi L, Dal Negro L, Mazzoleni C, Franzo G, Priolo F: Optical gain in silicon nanocrystals. Nature 2000, 408: 440. 10.1038/35044012View Article
  3. Cheylan S, Elliman RG: Effect of particle size on the photoluminescence from hydrogen passivated Si nanocrystals in SiO2. Appl Phys Lett 2001, 78: 1225. 10.1063/1.1338492View Article
  4. Zacharias M, Heitmann J, Scholz R, Kahler U, Schmidt M, Blasing J: Size-controlled highly luminescent silicon nanocrystals: a SiO/SiO2superlattice approach. Appl Phys Lett 2002, 80: 661. 10.1063/1.1433906View Article
  5. Valenta J, Pelant I, Linnros J: Waveguiding effects in the measurement of optical gain in a layer of Si nanocrystals. Appl Phys Lett 2002, 81: 1396. 10.1063/1.1502195View Article
  6. Tetelbaum DI, Trushin SA, Mikhaylov AN, Vasil'ev VK, Kachurin GA, Yanovskaya SG, Gaponova DM: The influence of the annealing conditions on the photoluminescence of ion-implanted SiO2: Si nanosystem at additional phosphorus implantation. Physica E 2003, 16: 410. 10.1016/S1386-9477(02)00614-8View Article
  7. Fang YC, Xie ZQ, Qi LJ, Li WQ, Zhang ZJ, Lu M: The effects of CeF3 doping on the photoluminescence of Si nanocrystals embedded in a SiO2matrix. Nanotechnol 2005, 16: 769. 10.1088/0957-4484/16/6/025View Article
  8. Lin GR, Lin CJ, Lin CK, Chou LJ, Chueh YL: Oxygen defect and Si nanocrystal dependent white-light and near-infrared electroluminescence of Si-implanted and plasma-enhanced chemical-vapor deposition-grown Si-rich SiO2. J Appl Phys 2005, 97: 094306. 10.1063/1.1886274View Article
  9. Jambois O, Rinnert H, Devaux X, Vergnat M: Photoluminescence and electroluminescence of size-controlled silicon nanocrystallites embedded in SiO2thin films. J Appl Phys 2005, 98: 046105. 10.1063/1.2034087View Article
  10. Chen XY, Lu YF, Tang LJ, Wu YH, Cho BJ, Xu XJ, Dong JR, Song WD: Annealing and oxidation of silicon oxide films prepared by plasma-enhanced chemical vapor deposition. J Appl Phys 2005, 97: 014913. 10.1063/1.1829789View Article
  11. Xie ZQ, Chen D, Li ZH, Zhao YY, Lu M: A combined approach to greatly enhancing the photoluminescence of Si nanocrystals embedded in SiO2. Nanotechnol 2007, 18: 115716. 10.1088/0957-4484/18/11/115716View Article
  12. Lin GR, Lin CJ, Chang YC: Optical characterization of CO2-laser-ablated Si-rich SiOx. Appl Phys Lett 2007, 90: 151903. 10.1063/1.2721141View Article
  13. Kim BH, Cho CH, Mun JS, Kwon MK, Park TY, Kim JS, Byeon CC, Lee J, Park SJ: Enhancement of the external quantum efficiency of a silicon quantum dot light-emitting diode by localized surface plasmons. Adv Mater 2008, 20: 3100. 10.1002/adma.200703096View Article
  14. Filip V, Wong H, Wong CK, Nicolaescu D: Electroluminescence of silicon nanoclusters excited by tunneling carrier injection. J Vac Sci Technol 2008, 26: 813. 10.1116/1.2830626View Article
  15. Wang T, Wei DY, Sun HC, Liu Y, Chen DY, Chen GR, Xu J, Li W, Ma ZY, Xu L, Chen KJ: Electrically driven luminescence of nanocrystalline Si/SiO2multilayers on various substrates. Physica E 2009, 41: 923. 10.1016/j.physe.2008.08.001View Article
  16. Ding L, Chen TP, Yang M, Wong J, Cen ZH, Liu Y, Zhu FR, Tseng AA: Relationship between current transport and electroluminescence in Si+-implanted SiO2thin films. IEEE Trans Electron Devices 2009, 56: 2785.View Article
  17. Beyer V, Schmidt B, Heinig KH, Stegemann KH: Light emitting field effect transistor with two self-aligned Si nanocrystal layers. Appl Phys Lett 2009, 95: 193501. 10.1063/1.3242379View Article
  18. Anopchenko A, Marconi A, Moser E, Prezioso S, Wang M, Pavesi L, Pucker G, Bellutti P: Low-voltage onset of electroluminescence in nanocrystalline-Si/SiO2multilayers. J Appl Phys 2009, 106: 033104. 10.1063/1.3194315View Article
  19. Park BY, Lee S, Park K, Bae CH, Park SM: Enhancement of light emission from silicon nanocrystals by post-O2-annealing process. J Appl Phys 2010, 107: 014314. 10.1063/1.3273365View Article
  20. Chen YB, Ren Y, Xiong RL, Zhao YY, Lu M: Modulation of the photoluminescence of Si quantum dots by means of CO2laser pre-annealing. Appl Surf Sci 2010, 256: 5116. 10.1016/j.apsusc.2010.03.080View Article
  21. Ding L, Yu MB, Tu XG, Lo GQ, Tripathy S, Chen TP: Laterally-current-injected light-emitting diodes based on nanocrystalline-Si/SiO2superlattice. Optics Express 2011, 19: 2729. 10.1364/OE.19.002729View Article
  22. Ha NN, Dohnalova K, Gregorkiewicz T, Valenta J: Optical gain of the 1.54 μm emission in MBE-grown Si:Er nanolayers. Phys Rev B 2010, 81: 195206.View Article
  23. Puzzo DP, Henderson EJ, Helander MG, Wang ZB, Ozin GA, Lu ZH: Visible colloidal nanocrystal silicon light-emitting diode. Nano Lett 2011, 11: 1585. 10.1021/nl1044583View Article
  24. Ren Y, Chen YB, Zhao YY, Lu M: Photoluminescence of Si from Si nanocrystaldoped SiO2/Si multilayered sample. Appl Surf Sci 2011, 257: 9578. 10.1016/j.apsusc.2011.06.068View Article
  25. Park NM, Kim TS, Park SJ: Band gap engineering of amorphous silicon quantum dots for light-emitting diodes. Appl Phys Lett 2001, 78: 2575. 10.1063/1.1367277View Article
  26. Berghoff B, Suckow S, Roelver R, Spangenberg B, Kurz H, Sologubenko A, Mayer J: Improved charge transport through Si based multiple quantum wells with substoichiometric SiOxbarrier layers. J Appl Phys 2009, 106: 083706. 10.1063/1.3238294View Article
  27. Lu ZH, Lockwood DJ, Baribeau JM: Quantum confinement and light-emission in SiO2/Si superlattices. Nature (London) 1995, 378: 258. 10.1038/378258a0View Article
  28. Prasad PN: Nanophotonics. New Jersey: John Wiley & Sons; 2004.View Article
  29. Ostwald W: Lehrbruck der Allgemeinen Chemie. Volume 2. Leipzig, Germany; 1896.


© Li et al; licensee Springer. 2012

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://​creativecommons.​org/​licenses/​by/​2.​0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.