All-silicon solid films with highly efficient and tunable full-color photoluminescence

Available online at www.sciencedirect.com

ScienceDirect Scripta Materialia 76 (2014) 17–20 www.elsevier.com/locate/scriptamat

All-silicon solid films with highly efficient and tunable full-color photoluminescence Jing Wang,a,b Xinglong Wu,b,⇑ Tinghui Lib,c and Paul K. Chud a

School of Electronic Science and Engineering, Nanjing University of Posts & Telecommunications, Nanjing 210003, People’s Republic of China b Key Laboratory of Modern Acoustics, MOE, Institute of Acoustics, Department of Physics, Nanjing University, Nanjing 210093, People’s Republic of China c College of Electronic Engineering, Guangxi Normal University, Guilin 541004, People’s Republic of China d Department of Physics and Materials Science, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong, China Received 23 October 2013; accepted 8 December 2013 Available online 12 December 2013

Silicon nanocrystals with the most probable diameter of 2 nm are surface passivated by glycerol and deposited on porous silicon (PS) substrate with intrinsic red photoluminescence. Tunable full-color emission is observed from this composite film and the quantum efficiency varies from 17% to 28% under different excitation wavelengths. Spectral analysis reveals that the full-color emission arises from combined quantum confinement in the silicon and PS nanocrystals with glycerol passivation. The nanostructured allsilicon film is robust and compatible with microelectronic processing. Ó 2013 Published by Elsevier Ltd. on behalf of Acta Materialia Inc. Keywords: Si nanocrystals; Glycerol passivation; Quantum confinement effect; Tunable photoluminescence

Low-dimensional silicon nanomaterials, including nanowires, nanosheets, nanoribbons and nanocrystals (NCs), have been extensively investigated, and silicon NCs in particular have attracted much interest. For instance, their optical characteristics, which are based on their electronic states, have been studied and found to be potentially useful for light emitting devices [1,2], bioprobes [3,4], fluorescence detectors [5] and spintronic devices [6]. According to previous studies on silicon NCs fabricated by laser irradiation [7], plasmonenhanced chemical vapor deposition [8], the non-thermal plasmas approach [1,9], electrochemical etching [10], chemical dissolution [11], annealing [12] and scanning transmission electron microscopic lithography [6], the silicon NC surface is chemically active and often bonded to oxygen atoms [7,10], hydrogen atoms [8,11,12], hydroxyl groups [8] or other groups [7,11]. Surface-modified silicon NCs can exhibit different optoelectronic properties due to changes in the band structure, suggesting large potential in different applications [6,8,13]. Various types of hybrid silicon NCs have been

⇑ Corresponding author. Tel.: +86 2583686303; e-mail: hkxlwu@nju. edu.cn

produced, such as hydrogenated silicon NCs [8,11,14], oxygen-passivated Si NCs [1,12], organic group-terminated NCs [1,3,7,12,15], and silicon NCs embedded in Si-based matrices such as SiO2 and silicon nitride [12,13]. With regard to the photonic properties, light emission has been observed from silicon NCs due to various mechanisms, such as surface states [16,17], defects [12], the quantum confinement effect [8,12] and interface localized states [13]. Light emission from silicon nanostructures in the infrared [2,8,9,13,15], red [1,2,7,10,12], yellow [7,11], green [7,18] and blue [10,16] ranges have recently been discovered and, as a result, a number of silicon-based light emitting devices [1,10,12,19] have been proposed [1,2,19]. The fabrication of silicon nanostructures is compatible with that practiced by the microelectronic and biomedical [3] industries, and can be readily integrated with other silicon-based devices. Consequently, an all-silicon structure capable of producing tunable and full-color emission has immense potential. In this work, surface-modified silicon NCs are embedded in porous silicon (PS) to produce a solid film that yields tunable full-color photoluminescence (PL) with a quantum efficiency of better than 17%. Our investigation elucidates that the tunable multi-color emission

1359-6462/$ - see front matter Ó 2013 Published by Elsevier Ltd. on behalf of Acta Materialia Inc. http://dx.doi.org/10.1016/j.scriptamat.2013.12.004

80 40 0

1

2

3

4

D ia m e te r (nm )

340 360 370 390 400 420

360 480 600 720 Wavelength (nm) (c) 460 440 490 420 610 400 380

PL Intensity (a.u.)

(a) 320

PL Intensity (a.u.)

Fig. 1. (a) TEM image of the silicon NCs. (b) Two typical highresolution TEM images of the silicon NCs. (c) Optical images of the silicon NC colloid excited at 370 and 420 nm, corresponding to emission wavelengths of 450 and 520 nm. (d) Size distribution of the silicon NCs calculated from TEM images. The most probable size of 2 nm is calculated according to Gaussian fitting.

PL Intensity (a.u.)

arises from the quantum confinement on silicon and PS NCs with glycerol passivation. The approach of fabricating silicon NC colloid is similar to that used in fabricating 3C–SiC NCs [20,21]. Two major differences are the proportions of HF and HNO3 in the etching solution, which are HF (40%):HNO3 (65%) = 20:1 in volume, and the quantities of the chemicals. A 20 ml volume of the etching liquid was added to silicon powders weighing 8 g. The silicon powders were dissolved and ultrasonically treated to obtain silicon NCs dispersed in water. Glycerol was then added to the water suspension containing the silicon NCs to fabricate a surface glycerol-passivated NC film as described previously, because direct dry treatment of the NC suspension will lead to disappearance of the strong PL due to various surface non-radiative defects [20]. The PS was fabricated by electrochemical etching of a piece of ptype Si wafer (<100 > and 10 X cm) in a mixture of HF (40% in volume) and ethanol (volume ratio of 1 to 2) at a current of 40 mA for 20 min [22]. The glycerolpassivated NCs were subsequently added on the PS substrate. The NCs were characterized using a JEOL JEM-2100 transmission electron microscope. The sample was prepared by putting a drop of the water suspension on a copper grid. The PL and PL excitation (PLE) spectra were acquired on an FLS920 fluorescence spectrophotometer (Edinburgh Instruments) equipped with a 450 W Xe lamp with a resolution of 1 nm. All the spectra were corrected for the response of the measurement system. X-ray photoelectronic spectra (XPS) was carried out on a Thermo Fisher Scientific photoelectron spectrometer. The narrow-scan spectra were obtained under ultrahigh vacuum conditions using monochromatic Al Ka X-ray radiation. Figure 1(a) displays a transmission electron microscopy (TEM) image of the silicon NCs taken at an accelerating voltage of 200 kV. Most of them have a diameter of less than 2 nm and some are even smaller than 1 nm. Figure 1(b) depicts two high-resolution TEM images of the silicon NCs. They are highly crystalline, with lattice fringes corresponding to the {2 1 0} and {2 2 0} planes of Si. The size distribution in Figure 1(d) shows that the diameter of the silicon NCs varies between 1 and 4 nm, and there are no NCs larger than 4 nm. The Gaussian fit in Figure 1(d) shows that the most probable diameter is 2 nm. Since NCs smaller than 1 nm may be missed due to the TEM resolution, the actual most probable size may be less than 2 nm. The PL/PLE spectra acquired from the silicon NC colloid, silicon NCs/glycerol film and silicon NCs/glycerol/PS film are shown in Figure 2. As shown in the PL spectra of the silicon NC colloid formed at pH 4 in Figure 2(a), as the excitation wavelength is increased from 320 to 420 nm, the PL peak changes from 400 to 520 nm. There is thus a red shift in the emitted light with increasing excitation wavelength, and the PL peak at 410 nm reaches the maximum intensity at an excitation wavelength of 340 nm. No obvious shift is observed when the excitation wavelength is longer than 440 nm because the intensity of the PL spectra decreases abruptly. Although the band gap of bulk silicon is 1.12 eV (1107 nm), no obvious signal is observed beyond

Counts

J. Wang et al. / Scripta Materialia 76 (2014) 17–20

PLE Intensity (a.u.)

18

350 330 320

370 390

(b)

410 450

360 480 600 720 Wavelength (nm) 370 (d) 350 380 330 400 430 320

240 320 400 480 300 400 500 600 700 800 Excitation wavelength (nm) Emission wavelength (nm)

(e)

Fig. 2. (a) PL spectra of the silicon NC colloid. (b) PL spectra of the silicon NCs/glycerol solid film. (c) PLE spectra of the silicon NCs/ glycerol solid film. (d) PL spectra of the silicon NCs/glycerol/PS solid film. The excitation and emission wavelengths are labeled. (e) Four optical photographs obtained from the silicon NCs/glycerol/PS solid film excited at 350, 380, 400 and 430 nm, corresponding to four different colors: blue (450 nm), green (520 nm), yellow (570 nm) and red (650 nm). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

550 nm, unlike that observed from PS [21]. This can be explained by the absence of NCs bigger than 4 nm in diameter and most of the NCs having a diameter of

J. Wang et al. / Scripta Materialia 76 (2014) 17–20

around 2 nm. It should also be mentioned that the pH used to produce the colloid should be about 7 in order to maximize the PL intensity (Fig. SI-1 in the Supporting information). Between 320 and 340 nm, silicon NCs with the probable size can always be excited and the PL intensity increases with excitation wavelength. However, as the excitation wavelength is increased from 350 to 420 nm, the number of silicon NCs which can be excited decreases, resulting in a drop in the PL intensity and a small water Raman peak. According to the model proposed by Wolkin et al. [22]to describe the electronic states in silicon NCs as a function of size and surface passivation, the most probable diameter can be estimated from the wavelength of the PL peak maximum and is calculated to be 1.8 nm, which is consistent with our results, considering that NCs smaller than 1 nm may not be detected here. This is much smaller than the exciton Bohr radius of bulk silicon of around 4.9 nm [10], implying that the red shift observed with increasing excitation wavelength can be attributed to quantum confinement in the silicon NCs. The image in Figure 1(c) includes two light-emitting photograohs corresponding to excitation wavelengths of 370 and 420 nm (left to right) and emission wavelengths of 450 nm (blue) and 520 nm (green). The optical photographs also show that the emitted light is intense enough for the naked eye to see. The PL spectra acquired from the NCs/glycerol solid film are depicted in Figure 2(b). No effective PL signal can be observed from pure glycerol and, similar to the silicon NC colloid, the PL peak red-shifts with increasing excitation wavelength. The PLE peak wavelength increases monotonically with excitation wavelength as well, as shown in Figure 2(c). The PLE peak shift shows that different numbers of NCs with different sizes can be excited by a single wavelength. This is consistent with the PL spectral results. The PLE spectra span a wider wavelength range than those acquired from the corresponding silicon NC colloid (Fig. SI-2 in the Supporting information). This indicates that glycerol passivation widens the silicon NC band gaps. The PL and PLE spectra provide evidence that the tunable PL from the silicon NCs in the violet to blue-green range originates from quantum confinement in the silicon NCs [21]. No PL degradation occurred when the glycerol-passivated NC film was stored in air for more than 6 months, indicating its stability. We can explain the tunable PL origin theoretically. For the Si NCs without glycerol absorption, the quantized levels of the conduction band are separated due to the quantum size effect [20]. Electrons relax from high to low quantized levels, then radiative recombination with holes in the valence band provide the PL emission with a fixed energy position. Thus, it is hard to tune the wavelength of the emitted photons by changing the frequency of the excitation light in a solid film. By gradually increasing the number of adsorbed glycerol molecules, the quantized levels of the conduction band gradually transform to a continuum. With increasing excitation energy, electrons can be pumped to higher levels in the upper quasi-continuous band and then relax to a higher level in the lower quasi-continuous band as

19

the energy interval of the relaxation is almost unchanged. Consequently, glycerol absorption on ultrathin Si NCs leads to the possibility of tuning the wavelengths of the emitted photons by changing the excitation frequency. The tunable violet to blue-green emission currently observed thus stems from the mutual effects of Si NC size and glycerol bonding. Based on this idea, other ligands that can effectively passivate the Si surfaces without producing nonradiative defect states should also be available, as is the case for 3C-SiC NC films [20]. The light emitted from the PS sample is strong, stable and tunable from 580 to 660 nm when the excitation wavelength is increased from 350 to 480 nm. The most intense PLE peak appears at an excitation wavelength of 370 nm. Since the original silicon NC colloid is fabricated at a pH of around 4 and the surface glycerol layer is more than one monolayer thick [20], it is possible that hydroxyl/alkoxide passivation replaces hydrogen bonding on ordinary PS NC surfaces after the glycerol-passivated silicon NCs have been incorporated into the PS. It thus avoids oxidation of the PS NCs and leads to quantum-confined PL from the PS. Figure 2(d) shows the PL spectra of the silicon NCs/ glycerol/PS solid film. The tunable PL in the violet to blue range (390–450 nm) arises from the quantum confinement effect in silicon NCs. Similar to the PL spectra of the silicon NC solid film, the peak wavelength increases with excitation wavelength, with the maximum intensity being observed at an excitation wavelength of 370 nm. In the blue-green range (450–550 nm) there are two peaks, suggesting that the PL in this range originates from both the silicon NCs and PS. In the orangered range (590–660 nm), the PL is attributed to the PS. The light emitted from the silicon NCs/glycerol/PS solid film is strong enough to be observed by both the naked eye and a digital camera. Figure 2(e) shows four photographs taken at four different excitation wavelengths, corresponding respectively to blue (450 nm), green (520 nm), yellow (570 nm), and red (650 nm). In our experiments, different filters are used to avoid a scattered light background from the Xe lamp. PL quantum yields were obtained in a calibrated integrated sphere using an FLS 920 spectrometer. Depending on the excitation wavelengths, the PL efficiency observed from the silicon NCs/glycerol/PS film varies from 17% to 28%, with the most probable one at about 22%. The Si2p, C1s and O1s core level XPS spectra acquired from the silicon NCs/glycerol/PS solid film are presented in Figure 3. In the Si2p spectrum in Figure 3(a), the strongest peak at 99.7 eV corresponds to silicon, whereas the shoulder at 100.4 eV can be attributed to the Si1+ state, which corresponds to one silicon atom bonding to one oxygen atom [23,24]. It comes from the –OH and –OR (R is dehydrogenated glycerol component) groups on the silicon and PS NCs due to water dissociation and glycerol bonding [23]. The weakest peak at 102.0 eV arises from Si–O bond and is associated with the Si–O–Si bridging bond [25,26]. This weak Si–O–Si peak appears in the Si2p spectra of both the silicon NC solid film (Fig. 3b) and the silicon NCs/glycerol solid film (Fig. SI-3 in the Supporting information), implying that it forms on the silicon NC

20

J. Wang et al. / Scripta Materialia 76 (2014) 17–20 (a) Si 2p

Si 2p

Si 2p

(b) Si 2p

Intensity (a.u.)

Si-OR

Si-O

Si-O

100

102

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/ 10.1016/j.scriptamat.2013.12.004.

104

98

(d) C 1s O-CH3 CHn

Si-O

100 102 104 106 (c) O 1s Si-OR

282 284 286 288 290 531

532

533

534

535

Binding energy (eV) Fig. 3. (a) Si2p, (c) O1s and (d) C1s XPS spectra of the NCs/glycerol/ PS solid film. (b) Si2p XPS spectrum of the silicon NC solid film.

and/or PS surface possibly due to the short storage in air prior to glycerol passivation [22]. The 532.4 eV peak in the O1s spectrum of Figure 3(c) corresponds to Si–O bonding and is consistent with the Si2p spectrum. Another peak at 533.1 eV corresponds to Si–OR bonding and is in agreement with the Si1+ state. There are two peaks in the C1s spectrum in Figure 3(d). The peak at 286.6 eV is consistent with the O–CH3 bond in the glycerol molecule [27] and the other peak, at 284.8 eV, is associated with the CHn group of glycerol. The XPS spectra provide information about the silicon NCs/PS surface, and reveal that the surface silicon atoms are bonded to –OH and –OR groups, together with a few naturally oxidized Si–O–Si bonds [22]. Hence, the silicon atoms are well passivated by hydroxyl/alkoxide groups, and glycerol prevents further oxidation of the silicon NC and PS surface effectively. In conclusion, tunable photoluminescence from violet to blue-green is achieved from silicon NCs with the surfaces passivated by glycerol and, after incorporating the nanostructured materials into PS, intense and tunable full-colored PL is observed from the composite materials. The full-color emission originates from quantum confinement in both the silicon NCs and PS with the surfaces passivated by glycerol. This all-silicon structure, which is robust and compatible with microelectronic processing, has immense potential in display technology. This work was supported by National Basic Research Programs of China under Grants Nos. 2011CB922102 and 2014CB339800 and National Natural Science Foundation (Nos. 21203098 and 11374141). Partial support was also from PAPD and Guangdong – Hong Kong Technology Cooperation Funding Scheme (TCFS) GHP/015/12SZ.

[1] K. Cheng, R. Anthony, U. Kortshagen, R. Holmes, Nano Lett. 11 (2011) 1952. [2] D. Puzzo, E. Henderson, M. Helander, Z. Wang, G. Ozin, Z. Lu, Nano Lett. 11 (2011) 1585. [3] B. Tian, T. Cohen-Kami, Q. Qing, X. Duan, P. Xie, C. Lieber, Science 329 (2010) 830. [4] X. Zhao, R. Tapec-Dytioco, W. Tan, J. Am. Chem. Soc. 125 (2003) 11474. [5] F. Erogbogbo, K.T. Yong, R. Hu, W.C. Law, H. Ding, C.W. Chang, P.N. Prasad, M.T. Swihart, ACS Nano 4 (2010) 5131. [6] M. Fuechsle, S. Mahapatra, F. Zwanenburg, M. Friesen, M. Eriksson, M. Simmons, Nat. Nanotechnol. 5 (2010) 502. [7] K. Kusova, O. Cibulka, K. Dohnalova, I. Pelant, J. Valenta, A. Fucikova, K. Zidek, J. Lang, J. Englich, P. Matejka, P. Stepanek, S. Bakardjieva, ACS Nano 4 (2010) 4495. [8] L. Bagolini, A. Mattoni, G. Fugallo, L. Colomb, Phys. Rev. Lett. 104 (2010) 176803. [9] M. Beard, K.P. Knutsen, P.R. Yu, J.M. Luther, Q. Song, W.K. Metzger, R.J. Ellingson, A.J. Nozik, Nano Lett. 7 (2007) 2506. [10] C.C. Tu, L. Tang, J.D. Huang, A. Voutsas, L.Y. Lin, Appl. Phys. Lett. 98 (2011) 213102. [11] A. Gupta, M. Swihart, H. Wiggers, Adv. Funct. Mater. 19 (2009) 696. [12] S. Godefroo, M. Hayne, M. Jivanescu, A. Stesmans, M. Zacharias, O. Lebedev, G. Van Tendeloo, V. Moshchalkov, Nat. Nanotechnol. 3 (2008) 174. [13] L. Pavesi, L. Dal Negro, C. Mazzoleni, G. Franzo, F. Priolo, Nature 408 (2000) 440. [14] V. Svrek, D. Mariotti, M. Kondo, Appl. Phys. Lett. 97 (2010) 161502. [15] Y. Chao, L. Siller, S. Krishnamurthy, P. Coxon, U. Bangert, M. Gass, L. Kjeldgaard, S. Patole, L. Lie, T. Alsop, A. Houlton, B. Horrocks, Nat. Nanotechnol. 2 (2007) 486. [16] A. Heintz, M. Fink, B. Mitchell, Adv. Mater. 19 (2007) 3984. [17] A. Carvalho, J. Coutinho, M. Barroso, E.L. Silva, S. Oberg, M. Rayson, P.R. Briddon, Phys. Rev. B 84 (2011) 125437. [18] D. Kovalev, M. Fujii, Adv. Mater. 17 (2005) 2531. [19] K. Cheng, R. Anthony, U. Kortshagen, R. Holmes, Nano Lett. 10 (2010) 1154. [20] J. Wang, S.J. Xiong, X.L. Wu, T.H. Li, P.K. Chu, Nano Lett. 10 (2010) 1466. [21] J. Zhu, Z. Liu, X.L. Wu, L.L. Xu, W.C. Zhang, P.K. Chu, Nanotechnology 18 (2007) 365603. [22] M.V. Wolkin, J. Jorne, P.M. Fauchet, Phys. Rev. Lett. 82 (1999) 197. [23] X.L. Wu, S.J. Xiong, J. Zhu, J. Wang, J.C. Shen, P.K. Chu, Nano Lett. 9 (2009) 4053. [24] G. Hollinger, F. Himpsel, Appl. Phys. Lett. 44 (1984) 93. [25] Y. Kanemitsu, N. Shimizu, T. Komoda, P.L.F. Hemment, B.J. Sealy, Phys. Rev. B 54 (1996) 14329. [26] G. Hadjisawas, P. Kelires, Phys. Rev. Lett. 93 (2004) 226104. [27] M. Rosso, A. Arafat, K. Schroen, M. Giesbers, C.S. Roper, R. Maboudian, H. Zuilhof, Langmuir 24 (2008) 4007.