- Nano Express
- Open Access
Influence of Fabrication Processes and Annealing Treatment on the Minority Carrier Lifetime of Silicon Nanowire Films
Nanoscale Research Lettersvolume 12, Article number: 242 (2017)
Surface passivation and bulk carrier lifetime of silicon nanowires (SiNWs) are essential for their application in solar cell devices. The effective minority carrier lifetime of a semiconductor material is influenced by both its surface passivation and bulk carrier lifetime. We found that the effective carrier lifetime of SiNWs passivated with aluminum oxide (Al2O3) was significantly influenced by the fabrication process of SiNWs. We could not measure the effective lifetime of SiNWs fabricated by thermal annealing of amorphous silicon nanowires. Nevertheless, the SiNWs fabricated by metal-assisted chemical etching of polycrystalline silicon displayed an effective lifetime of 2.86 μs. Thermal annealing of SiNWs at 400 °C in a forming gas improved the effective carrier lifetime from 2.86 to 15.9 μs because of the improvement in surface passivation at the interface between the SiNWs and Al2O3 layers.
Silicon nanowires (SiNWs) possess promising qualities for use in thin-film solar cells, i.e., they exhibit a strong optical confinement effect that is essential for light trapping in solar cells [1–5]. SiNWs can be fabricated by metal-assisted chemical etching (MAE), which is a low-cost chemical solution-based process and, therefore, presents a potential for low-cost fabrication of SiNW-based solar cells [6–11]. Carrier recombination can significantly affect the performance of solar cells. Thus, the application of SiNWs in solar cells will require the carrier recombination in SiNWs to be examined. One of the most important measures of carrier recombination is the effective carrier lifetime (τ eff). The value of τ eff is very sensitive to the surface recombination and bulk lifetime, and therefore, τ eff measurements provide information about surface recombination and bulk lifetime at the same time [12–14]. Owing to the difficulties with the fabrication of isolated SiNWs, however, only few studies examining the τ eff of SiNWs have been reported in the literature to date [15, 16]. SiNWs are often fabricated by chemical etching of a crystalline Si (c-Si) substrate [6, 17]. In SiNWs fabricated using the proposed method, the c-Si substrate remains underneath the SiNWs and influences the outcome of τ eff measurements. For this reason, fabrication techniques that do not employ the c-Si wafer are needed to remove the influence of the c-Si substrate on the τ eff measurements of SiNWs.
In our previous work, isolated SiNWs were fabricated from polycrystalline Si (poly-Si) on a quartz substrate [7–9]. Additionally, our previous work also demonstrated that τ eff measurements of isolated SiNWs and their optical characterization are possible .
In this study, we investigated the influence of the fabrication process of isolated SiNWs passivated with aluminum oxide (Al2O3) on τ eff. Isolated SiNWs can be fabricated using two different processes, as shown in Fig. 1. Here, SiNWs were fabricated by thermal annealing of amorphous silicon (a-Si) nanowires, marked in Fig. 1 as process A. The a-Si nanowires were fabricated using the MAE process. Additionally, SiNWs were fabricated by MAE of poly-Si, in a process marked as B in Fig. 1. In this case, the poly-Si layer was fabricated by thermal annealing of an a-Si layer. Finally, the influence of annealing time and temperature on the value of τ eff was investigated.
Isolated SiNWs were fabricated using two different procedures as shown in Fig. 1. A highly doped thin poly-Si layer of thickness 163 nm was fabricated on a quartz substrate by aluminum-induced crystallization (AIC). The details of the AIC process have been reported previously . The AIC-poly-Si layer acts as a seed layer for the crystallization of SiNWs. Subsequently, an intrinsic a-Si (i-a-Si) layer was deposited on the AIC-poly-Si by radio frequency sputtering. In process A, amorphous SiNWs were fabricated by MAE of the i-a-Si layer. During the MAE process, silver particles were deposited on the i-a-Si layer by electroless silver plating. The sample containing silver nanoparticles was immersed in an etching solution comprising HF and H2O2. Following the formation of amorphous SiNWs, the sample was dipped in an HNO3 solution to remove the silver nanoparticles. Finally, the oxide layer formed on the surface of the amorphous SiNWs was removed using an HF solution. The amorphous SiNWs were crystallized by thermal annealing at 800, 900, and 1000 °C for 30 min in a forming gas (FG) under ambient conditions. In process B, the i-a-Si layer was crystallized by solid phase crystallization (SPC) to give the poly-Si prior to nanowire formation through thermal annealing at 900 °C for 30 min in a forming gas at ambient conditions. Finally, the SiNWs were fabricated from the poly-Si layer by MAE.
Following the preparation of SiNWs, an Al2O3 layer was deposited on the SiNWs for surface passivation by atomic layer deposition. As reported in , the effective lifetime of SiNWs lacking a passivation layer cannot be measured because of the extremely high effective surface recombination velocity. Figure 2 depicts the cross-sectional image obtained by using a scanning electron microscope (SEM) and the energy dispersive X-ray spectroscopy (EDS) images of SiNWs containing an Al2O3 layer. The EDS images clearly showed that the signals arising from Al and O are observed uniformly across the entire area of the sample. This result indicates that the fabricated SiNWs were entirely covered with the Al2O3 layer. Thermal annealing treatment was conducted in a forming gas (4% H2 in N2) at 300, 400, and 500 °C. The influence of annealing time on the passivation quality was also investigated.
The structure of the fabricated SiNW samples was characterized by field-emission SEM (FE-SEM) and EDS (JEOL, JSM-7001F). The crystallization of SiNWs was confirmed by Raman spectroscopy (JASCO, NRS-1000, excitation wavelength of 532 nm). The effective carrier lifetime of the samples was measured using a microwave photoconductivity decay (MWPCD) technique and a 904-nm laser (KOBELCO, LTE-1510EP). The interface-state density (D it) and fixed-charge density (Q ss) were measured using a noncontact corona-voltage (C-V) technique (SEMILAB Co. Ltd., PV-2000).
Results and Discussion
Figure 3 depicts the cross-sectional SEM images of SiNWs fabricated using process A and process B. The shape of the wires was found to be very similar for both processes, indicating that similar patterns of silver nanoparticles were formed on both the i-a-Si and poly-Si layers. This result demonstrated that the Ag pattern does not depend on the employed fabrication process. In the next step, the internal structure of SiNWs fabricated using different procedures was investigated by Raman spectroscopy. Figure 4a depicts the Raman spectra of SiNWs fabricated using process A. The Raman spectra of SiNWs crystallized at 800, 900, and 1000 °C are shown in Fig. 4a. Three peaks, attributed to crystalline silicon (peak 1), nanocrystalline silicon (peak 2), and amorphous silicon (peak 3), were observed in all spectra. In contrast, only peaks 1 and 2 were observed in the spectra of SiNWs fabricated using process B (Fig. 4b). These results indicate that the SiNWs fabricated using process A contained a residual a-Si phase. The results of Raman spectroscopy, therefore, suggest that the crystallization temperature of a-SiNWs is higher than that of a-Si. This temperature difference can be explained by the Helmholtz free energy of poly-Si. When a polycrystalline silicon with a radius r is prepared, the change in the Helmholtz free energy can be calculated as shown in Eq. (1).
where ΔF ν is the variation in Helmholtz free energy between a-Si and poly-Si per unit volume and σ is the interface energy. The first term represents the variation in free energy because of phase transformation from a-Si to poly-Si, and therefore, ΔF is expected to become negative as the poly-Si grows. The second term represents the increase in the surface area of a-Si that results in the free energy increase. When the influence of the second term is stronger than that of the first term, the a-Si does not crystallize. The crystal nucleation in a-SiNW is affected easily by the nanowire surface, and therefore, the second term is larger than the first term. As a result, the crystallization temperature of a-SiNW is higher compared with that of a-Si. To date, a number of studies reported in the literature have shown that the crystallization temperature of a-Si nanostructure increases [18–20].
In addition to structural characterization, we investigated the τ eff of SiNWs/Al2O3 fabricated using different procedures. The results showed that the τ eff of SiNWs/Al2O3 fabricated using process A could not be measured, whereas that of SiNWs/Al2O3 fabricated using process B was successfully measured. Most likely, this difference can be attributed to the bulk lifetime of SiNWs since the surface passivation quality of SiNWs fabricated using both processes was essentially identical. The results of Raman spectroscopy confirmed the presence of a residual a-Si phase in the SiNWs fabricated using process A, and this presence significantly affects the bulk lifetime of SiNWs. Overall, these results indicate that process B is more useful for solar cell applications than process A.
The surface passivation quality of the c-Si/Al2O3 interface is known to be strongly influenced by thermal annealing. Therefore, we investigated the effect of thermal annealing on the τ eff of SiNWs/Al2O3 fabricated using process B. Figure 5 depicts the decay curves of the MWPCD signal of as-deposited and annealed SiNWs/Al2O3. Each decay curve can be divided into fast and slow decay components. Although the origin of the slow decay has not been clarified, it is most likely related to the carrier trap in SiNWs. We estimated the τ eff of as-deposited SiNWs/Al2O3 from the fast decay curve as 2.86 μs. Thermal annealing of SiNWs/Al2O3 at 400 °C for 30 min in the forming gas increased the τ eff drastically from 2.86 to 15.9 μs.
Generally, the surface recombination velocity of the Si and Al2O3 interface decreases during thermal annealing because of the formation of a negative fixed charge [21–24]. Therefore, we examined the effect of thermal annealing on the SiNW/Al2O3 samples in more detail. Figure 6a depicts the variation in τ eff observed as a function of annealing time and temperature. For annealing temperatures of 300 and 400 °C, the τ eff increased as the annealing time was increased to 30 min, followed by a region of saturation. Interestingly, the behavior exhibited by τ eff at the annealing temperature of 500 °C was very different from that exhibited by τ eff at other temperatures. The annealing at 500 °C for 1 min increased the τ eff from 3 to 11 μs. However, further increase in annealing time resulted in a monotonic decrease in τ eff. The difference in behavior observed at the low annealing temperatures (300 and 400 °C) and the high temperature (500 °C) can be attributed to the changes in surface passivation. Figure 6b depicts the changes in τ eff of c-Si/Al2O3 determined as a function of annealing time and temperature. This experiment was performed to examine the influence of annealing on the quality of surface passivation. To eliminate the influence of bulk lifetime, we employed a monocrystalline Si substrate (p-type, , 8–12 Ωcm, thickness of 550 μm). The results showed that annealing at 400 °C increased the τ eff monotonically, whereas annealing at 500 °C decreased the τ eff after a long annealing time. Overall, the behavior observed for the τ eff of c-Si/Al2O3 was similar to that determined for SiNWs/Al2O3. The τ eff of c-Si/Al2O3 samples is known to be determined by D it and Q ss. Nevertheless, determining D it and Q ss at the interface of SiNWs/Al2O3 is extremely challenging, and therefore, we estimated these parameters from those of c-Si/Al2O3, under the assumption that similar D it and Q ss can be expected. Figure 6 shows the values of D it and Q ss determined as a function of annealing time. In the absence of thermal annealing, the Q ss at 400 and 500 °C was found to be 5.27 × 1010 and 1.0 × 1010 cm−3, respectively. We used different samples for the annealing experiments at 400 and 500 °C, and therefore, a small difference in the observed values of Q ss was expected. The relationship between the surface recombination velocity and Q ss was established as follows: the results showed that low values of Q ss (below 1011 cm−3) did not reduce the surface recombination velocity . As shown in Fig. 7a, Q ss increased drastically after annealing. The Al2O3 layer fabricated on the Si generates negatively charged Al vacancies in the vicinity of the interface [23, 26]. The formation of Al vacancies represents one of the possible explanations for the origin of the negative fixed charge at the Al2O3/c-Si interface. The samples annealed at 400 and 500 °C exhibited almost identical Q ss. Therefore, the decrease in τ eff cannot be explained by the values of Q ss. Figure 6b illustrates that D it varied significantly with changes in the annealing temperature. For example, while the D it at 400 °C was found to decrease with increasing time, the D it at 500 °C exhibited the opposite trend. At the annealing time at 400 °C, the hydrogen from Al2O3 and FG can connect to the dangling bond. At annealing temperatures above 500 °C, on the other hand, the dangling bonds on the Si surface are generated as a result of bond disconnection between Si and H . The results of our analysis showed that the value of τ eff as a function of time was affected by the time-variable parameter D it. These results indicate that the SiNWs/Al2O3 annealing conditions significantly affect the interface quality.
We investigated the influence of SiNW fabrication and annealing procedures on the effective lifetime of SiNW/Al2O3 samples. Complete crystallization of SiNWs is essential to obtain a high effective lifetime, in particular, because the a-Si phase remaining in the material can result in a decrease in the bulk lifetime of SiNWs. Here, fully crystallized thin-layer SiNWs were obtained by MAE of large-grain poly-Si. Through thermal annealing at 400 °C for 30 min in a forming gas, the effective lifetime of SiNWs/Al2O3 drastically improved from 2.86 to 15.9 μs. By comparing the effective lifetime of SiNWs/Al2O3 and c-Si/Al2O3 samples, we concluded that the thermal annealing conditions, especially temperature, significantly affect the state density at the SiNW/Al2O3 interface.
- Al2O3 :
Metal-assisted chemical etching
Microwave photoconductivity decay
Solid phase crystallization
- τ eff :
Effective carrier lifetime
Kato S, Kurokawa Y, Watanabe Y, Yamada Y, Yamada A, Ohta Y et al (2013) Optical assessment of silicon nanowire arrays fabricated by metal-assisted chemical etching. Nanoscale Res Lett 8: 216
Lin CX, Povinelli ML (2009) Optical absorption enhancement in silicon nanowire arrays with a large lattice constant for photovoltaic applications. Opt Express 17:19371–19381
Kurokawa Y, Kato S, Watanabe Y, Yamada A, Konagai M, Ohta Y et al (2012) Numerical approach to the investigation of performance of silicon nanowire solar cells embedded in a SiO2 matrix. Jpn J Appl Phys 51:11PE12-1 -11PE12-4
Wang X, Peng KQ, Pan XJ, Chen X, Yang Y, Li L et al (2011) High-performance silicon nanowire array photoelectrochemical solar cells through surface passivation and modification. Angew Chem Int Ed 50:9861–9865
Yu P, Wu J, Liu ST, Xiong J, Jagadish C, Wang ZMM (2016) Design and fabrication of silicon nanowires towards efficient solar cells. Nano Today 11:704–737
Zhang ML, Peng KQ, Fan X, Jie JS, Zhang RQ, Lee ST et al (2008) Preparation of large-area uniform silicon nanowires arrays through metal-assisted chemical etching. J Phys Chem C 112:4444–4450
Ishikawa R, Kato S, Yamazaki T, Kurokawa Y, Miyajima S, Konagai M (2014) Solid-phase crystallization of amorphous silicon nanowire array and optical properties. Jpn J Appl Phys 53:02BE09-1 - 02BE09-4
Yamazaki T, Kato S, Miyajima S, Konagai M (2014) Minority carrier lifetime of thin polycrystalline silicon nanowire films on polycrystalline silicon layer prepared by aluminum-induced crystallization. Next Generation Technologies for Solar Energy Conversion V 9178:91780L-1 - 9178L-7
Kato S, Yamazaki T, Miyajima S, Konagai M (2014) Influence of substrates on formation of polycrystalline silicon nanowire films. Next Generation Technologies for Solar Energy Conversion V 9178:91780M-1 - 91780M-6
Brodoceanu D, Alhmoud HZ, Elnathan R, Delalat B, Voelcker NH, Kraus T (2016) Fabrication of silicon nanowire arrays by near-field laser ablation and metal-assisted chemical etching. Nanotechnology 27:075301-1 - 075301-8
Zhang SL, Wang XW, Liu H, Shen WZ (2014) Controllable light-induced conic structures in silicon nanowire arrays by metal-assisted chemical etching. Nanotechnology 25:025602-1 - 025602-11
Brody J, Rohatgi A, Ristow A (2003) Review and comparison of equations relating bulk lifetime and surface recombination velocity to effective lifetime measured under flash lamp illumination. Sol Energy Mater Sol Cells 77:293–301
Lauer K, Laades A, Ubensee H, Metzner H, Lawerenz A (2008) Detailed analysis of the microwave-detected photoconductance decay in crystalline silicon. J Appl Phys 104:104503-1 - 104503-9
Ogita YI (1996) Bulk lifetime and surface recombination velocity measurement method in semiconductor wafers. J Appl Phys 79:6954–6960
Kato S, Kurokawa Y, Miyajima S, Watanabe Y, Yamada A, Ohta Y et al (2013) Improvement of carrier diffusion length in silicon nanowire arrays using atomic layer deposition. Nanoscale Res Lett 8:361
Sivakov V, Andra G, Gawlik A, Berger A, Plentz J, Falk F et al (2009) Silicon nanowire-based solar cells on glass: synthesis, optical properties, and cell parameters. Nano Lett 9:1549–1554
Kato S, Watanabe Y, Kurokawa Y, Yamada A, Ohta Y, Niwa Y et al (2012) Metal-assisted chemical etching using silica nanoparticle for the fabrication of a silicon nanowire array. Jpn J Appl Phys 51:02BP09-1 - 02BP09-4
Hirasawa M, Orii T, Seto T (2006) Size-dependent crystallization of Si nanoparticles. Appl Phys Lett 88:093119-1 - 093119-3
Mishra P, Jain KP (2000) Temperature-dependent Raman scattering studies in nanocrystalline silicon and finite-size effects. Physical Review B 62:14790–14795
Kanzawa Y, Kageyama T, Takeoka S, Fujii M, Hayashi S, Yamamoto K (1997) Size-dependent near-infrared photoluminescence spectra of Si nanocrystals embedded in SiO2 matrices. Solid State Commun 102:533–537
Agostinelli G, Delabie A, Vitanov P, Alexieva Z, Dekkers HFW, De Wolf S, et al (2006) Very low surface recombination velocities on p-type silicon wafers passivated with a dielectric with fixed negative charge. Solar Energy Mater and Solar Cells 90:3438–3443
Hoex B, Schmidt J, Pohl P, van de Sanden MCM, and Kessels WMM (2008) Silicon surface passivation by atomic layer deposited Al(2)O(3). J Appl Phys 104:044903-1 - 044903-12
Hoex B, Gielis JJH, de Sanden MCMV, Kessels WMM (2008) On the c-Si surface passivation mechanism by the negative-charge-dielectric Al2O3. J Appl Phys 104: 113703-1 - 113703-7
Dingemans G, Seguin R, Engelhart P, van de Sanden MCM, Kessels WMM (2010) Silicon surface passivation by ultrathin Al2O3 films synthesized by thermal and plasma atomic layer deposition. Physica Status Solidi-Rapid Res Lett 4: 10–12
Miyajima S (2009) Surface passivation films for crystalline silicon solar cells. The Japan Society of Plasma Science and Nuclear Fusion Research 85:820–824
Naumann V, Otto M, Wehrspohn RB, Werner M, Hagendorf C (2012) Interface and material characterization of thin ALD-Al2O3 layers on crystalline silicon. Proc 2nd Int Conf Crystalline Silicon Photovoltaics 27: 312–318
Cantin JL, von Bardeleben HJ (2002) An electron paramagnetic resonance study of the Si(100)/Al2O3 interface defects. J Non Cryst Solids 303:175–178
This work was supported in part by the Ministry of Education, Culture, Sports, Science and Technology (MEXT) and FUTURE-PV Innovation.
SK carried out the experiment and wrote the initial draft of the manuscript. TY contributed to the sample fabrication and sample measurement. SM supervised the work and finalized the manuscript. YK and MK gave the final approval of the version of the manuscript to be published. All authors read and approved the final manuscript.
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.