Antireflection and passivation property of titanium oxide thin film on silicon nanowire by liquid phase deposition

SCT-22003; No of Pages 7 Surface & Coatings Technology xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Surface & Coatings Technology journal homepage: www.elsevier.com/locate/surfcoat

Antireflection and passivation property of titanium oxide thin film on silicon nanowire by liquid phase deposition Jung-Jie Huang a,⁎, Che-Chun Lin b,1, Dong-Sing Wuu b,1 a b

Department of Electrical Engineering, Da-Yeh University, Changhua, 51591, Taiwan, ROC Department of Materials Science and Engineering, National Chung Hsing University, Taichung 40227, Taiwan, ROC

a r t i c l e

i n f o

Article history: Received 30 August 2016 Revised 3 January 2017 Accepted in revised form 9 January 2017 Available online xxxx Keywords: Antireflection coating Passivation Liquid phase deposition Titanium oxide

a b s t r a c t To improve a high-efficiency silicon nanowires solar cell, using the antireflection coatings and surface passivation technique were very important. In this investigation, titanium oxide antireflection coatings were deposited on silicon nanowires by using liquid phase deposition. The deposition solution of (NH4)2TiF6 and H3BO3 were used for titanium oxide deposition. The concentration of the H3BO3 in the deposition solution play important roles in the formation of Ti–Si1−xOy interface layer between the titanium oxide/silicon nanowires interface and to control the trace amount of hydrofluoric acid in the solution. The titanium oxide films modification decreases and increases the reflectance and effective minority carrier lifetime of the silicon nanowires arrays. Under the optimal condition, the reflectance and effective minority carrier lifetime of liquid phase deposited titanium oxide film were 3.6% and 1.29 μs, respectively. The titanium oxide films were used herein to fabricate antireflection coating and passivation film to ensure low cost, good uniformity, favorable adhesion, mass producibility, and the formation of large-area thin film; thus, the liquid phase deposition-antireflection coating film was highly favorable for silicon-based solar cells. © 2017 Elsevier B.V. All rights reserved.

1. Introduction Recently, one-dimensional nanostructure-based solar cell surfaces have received increasing attention for two reasons: (i) fabrication of mesoporous nanostructures is simple and cost effective and (ii) the one-dimensional array nanostructures effectively decrease reflectance with increasing absorption of sunlight. Several studies have investigated one-dimensional nanostructures. This technology may be produced both by “bottom-up” and “top-down” approaches such as vaporliquid-solid process [1,2], laser-assisted catalytic growth [3], hydrothermal synthesis method [4,5], and electroless etching [6,7]. The mechanism of high temperature vapor–liquid–solid process growth in a furnace requires a long process time and the growth conditions cannot be easily controlled. Laser-assisted catalytic growth requires expensive laser devices to grow the nanowires. Moreover, the hydrothermal synthesis method was preferred to prepare a seed crystal as it has a long process time and produces nanowires of large diameter. The electroless etching process was used for fabricating silicon nanowires (SiNWs) in this study due to its great advantages such as low-cost, simplicity, mass producibility, uniformity, and formation of large-area thin films.

⁎ Corresponding author. E-mail address: [email protected] (J.-J. Huang). 1 These authors contributed equally to this work.

SiNWs with excellent antireflection properties are widely used in high efficiency solar cells [8–10]. However, SiNWs exhibit a surface recombination phenomenon and have uneven pore structures as they possess a high specific surface area and different etching rate, wherein numerous dangling bonds tend to exist, which increases the reflectance [11–13]. At present, antireflective coating (ARC) passivation films are most widely used in crystalline silicon solar cells due to the high effective minority carrier lifetime (τeff) that can be obtained, which prevents carrier recombination behavior in some high recombination regions (e.g., the cell surface, the contact region at the cell surface, and the metal electrode) and minimizes the front reflection, thereby improving the conversion efficiency of the cell. Herein, the TiO2 thin films were applied as the surface passivation layer and ARC for the SiNWs-based solar cell due to the excellent antireflection properties and fluorine doping, which can inhibit the recombination of photogenerated electrons and holes. Titanium oxide thin films are generally formed using vacuum processes such as atomic layer deposition [13] and chemical vapor deposition [14–16]. These methods can be used to produce films with uniform thickness and favorable electrical properties. However, conventional vacuum deposition processes are expensive and unsuitable for continuous mass production, particularly for forming coatings for use in lowcost solar cells. Sol-gel dip coating [17,18], hydrothermal process [19–21], spray pyrolysis [22], and liquid phase deposition (LPD) [23] are some of the non-vacuum processes currently used for depositing

http://dx.doi.org/10.1016/j.surfcoat.2017.01.027 0257-8972/© 2017 Elsevier B.V. All rights reserved.

Please cite this article as: J.-J. Huang, et al., Surf. Coat. Technol. (2017), http://dx.doi.org/10.1016/j.surfcoat.2017.01.027

2

J.-J. Huang et al. / Surface & Coatings Technology xxx (2017) xxx–xxx

Fig. 1. Flow chart of LPD-TiO2 thin film deposition on SiNWs.

TiO2 thin films on a substrate. Among these processes, in this study, LPD has been used for fabricating TiO2 thin films due to its various advantages, including low costs, uniformity, favorable adhesion, mass producibility, fluorine passivation and formation of large-area thin films. LPD mainly includes hydrolysis and a direct deposition reaction of the metal–fluoro complex at low ambient temperature (including room temperature) without any heating process, which has advantages such as high selectivity, large area, simplicity, ease of change of the film composition, and ease of mass production [24–26]. However, there are absence studies of the passivation and antireflection properties of TiO2 thin films on the SiNWs solar cell using LPD. To achieve the optimal passivation and antireflection of TiO2 thin films on SiNWs, the precursor concentration of AgNO3 for controlling the Si substrate etching and H3BO3 for controlling the TiO2 film deposition should be adjusted. In this study, the passivation and antireflection properties of TiO2 thin films on the SiNWs substrate using LPD were investigated.

2. Experiment A boron-doped, p-type (100)-oriented silicon wafer with a resistivity of 0.5–3 Ω·cm was used as the substrate in this study. The Si substrate was degreased in a solvent, chemically etched in a solution (HF:H2O = 1:10) for 30 s, and then rinsed in deionized (DI) water. The etching and deposition system contains (1) a temperaturecontrolled water bath that provides uniform etching and deposition temperature at an accuracy of ±0.1 °C and (2) a Teflon vessel containing the etching and deposition solution. Fig. 1 shows a flow chart of the SiNWs and LPD-TiO2 thin films fabrication process using metal-assisted wet chemical etching and LPD. First, the SiNWs etching solution was prepared using a mixture of silver nitrate and hydrofluoric acid (HF) solution with the combined proportions of 40 mL of 0.04 M silver nitrate and 32 mL of 0.4 M HF. The etching temperature was maintained at 40 °C during the etching process. After the necessary etching process was completed, the silicon substrate was immersed in the nitric acid and HF solution to remove the residual silver particle and native oxide layer, respectively, and the SiNWs structure was thus obtained. Titanium oxide thin films were orderly deposited on a SiNWs substrate using LPD. 20 mL of (NH4)2TiF6 (0.2 M) solution that was saturated with TiO2 powder was mixed with 20 mL H3BO3 (0.3–0.7 M) for depositing the TiO2 thin films. The deposition temperature was maintained at 60 °C during the deposition. After the deposition of the LPD-TiO2 thin films, the substrate was rinsed in DI water and dried using purified nitrogen gas. Finally, post-deposition annealing was performed in a quartz

furnace at a temperature of 500 °C for 30 min under nitrogen ambient to increase the film density, adhesion, and passivation properties. The surface morphologies of the LPD films and SiNWs were analyzed using field-emission scanning electron microscopy (FE-SEM; JEOL JSM7000F) at an accelerating voltage of 15 kV. The reflection spectra of the samples at wavelengths from 400 to 800 nm were obtained using a UV– vis spectrophotometer. Chemical compositions of the LPD films were obtained by X-ray photoelectron spectroscopy (XPS; PHI 5000 VersaProbe) with an Al Kα radiation (photon energy of 1486.6 eV). The energy resolution of this instrument was 0.5 eV full-width at half maximum. The measurement was conducted at a base pressure of 7.4 × 10−7 Pa in an analyzer chamber. A 2 kV argon ion beam with a current density of 100 A/cm2 was used to acquire the depth profiles, and the binding energy of each element was self-calibrated to C 1s (284.5 eV) reference peak states. The τeff of the LPD-TiO2 thin films were measured using a Sinton Instruments WCT-120 system in the QSSPC mode. The τeff of LPD-TiO2 thin films on the Si substrate using different H3BO3 concentration were measured under photo-excitation, which was mainly the recombination time of the sample exciting electrons and holes after illumination. Furthermore, τeff depends on both bulk minority carrier lifetime (τbulk) and surface recombination velocity

Fig. 2. The reflectance spectra of the SiNWs as a function of AgNO3 concentration at wavelengths from 400 to 800 nm.

Please cite this article as: J.-J. Huang, et al., Surf. Coat. Technol. (2017), http://dx.doi.org/10.1016/j.surfcoat.2017.01.027

J.-J. Huang et al. / Surface & Coatings Technology xxx (2017) xxx–xxx Table 1 The average reflectance at wavelengths from 400 to 800 nm of the SiNWs as a function of AgNO3 concentration. Concentration of AgNO3 (M)

0.01

0.02

0.03

0.04

0.05

Bare Si wafer

LPD-TiO2/Si wafer for λ = 4nd

Average reflectance (%)

7.2

4.9

4.3

3.9

8.1

35.4

5.4

(Seff) expressed as follows [27]: 1 1 2S ¼ þ eff τeff τbulk W

ð1Þ

where W is the C-Si wafer thickness. 3. Results and discussion In order to decrease the production cost of the solar cells, this study processed commercial-scale solar-grade silicon wafers into nanostructures by metal-assisted wet chemical etching. First, the influence of the AgNO3 concentration on reflectance was analyzed. Fig. 2 and

Fig. 3. Etching rate and porosity of SiNWs as a function of AgNO3 concentration.

3

Table 1 show the reflectance spectra and average reflectance of the SiNWs as a function of AgNO3 concentration. In general, the bare silicon wafer has high reflectance up to 35.4%, whereas the reflectance of the etched silicon wafer with the wavelength of 400–800 nm was significantly reduced to 3.9% under the AgNO3 concentration of 0.04 M. This was because after being soaked in a mixed etching solution of HF and AgNO3, the silicon wafer would be etched into a one-dimensional array of SiNWs. The size of the SiNWs was smaller than or approached the visible light wavelength; therefore, light will not undergo interference or diffraction while passing through it. Hence, the reflectance of the SiNWs with a wavelength of 400–800 nm was overall effectively reduced. In contrast, the thickness of the traditional single-layer antireflection coating with a one-fourth wavelength was restricted by λ = 4nd; therefore, the reflectance can only reduce the intensity of the light with a specific wavelength to nearly zero (where λ, n, and d denote incident wavelength, refractive index of film, and thickness of film, respectively). The observations reveal that when the concentration of AgNO3 was increased from 0.01 to 0.04 M, the reflectance of SiNWs decreased significantly, which was attributed to the porosity of SiNWs. Fig. 3 shows the etching rate and porosity of SiNWs under different AgNO3 concentrations from 0.01 to 0.05 M. The etching rate and porosity increase as the AgNO3 concentration increases, owing to the metal-assisted wet chemical etching was generated the flow of electrons by continuous reduction reaction of the silver ions. The etching process can be expressed as follows [7]: Si þ 2H2 O→ SiO2 þ4Hþ þ4e−

ð2Þ

Agþ þe− →Ag0

ð3Þ

SiO2 þ 6HF→H2 SiF6 þ2H2 O

ð4Þ

At the initial stage, the etching of the silicon wafer and deposition of silver occurred simultaneously. First, the silver ions accumulated to form a nucleation layer, which then grew into a twig-shaped cluster. Finally, the silver-colored cluster covered the surface of the silicon wafer. The silver ions and the silicon of peripheral would undergo a redox reaction, and the HF in the solvent was dissociated into hydrogen ions and fluoride ions, wherein fluoride ions and silicon would form H2SiF6 (dissolved in the solution). With continuous redox reactions, silicon

Fig. 4. SEM cross-sectional images of SiNWs/Si after immersed in nitric acid and HF solution to remove the residue silver particle and native oxide layer as a function of AgNO3 concentration, (a) bare silicon, (b) 0.01 M, (c) 0.02 M, (d) 0.03 M, (e) 0.04 M, and (f) 0.05 M.

Please cite this article as: J.-J. Huang, et al., Surf. Coat. Technol. (2017), http://dx.doi.org/10.1016/j.surfcoat.2017.01.027

4

J.-J. Huang et al. / Surface & Coatings Technology xxx (2017) xxx–xxx

Fig. 5. SEM cross-sectional images of LPD-TiO2/Si substrate and LPD-TiO2/SiNWs/Si (the SiNWs was determined using AgNO3 0.04 M) at different H3BO3 concentration, (a) LPD-TiO2 (0.5 M)/Si substrate, (b) LPD-TiO2 (0.3 M)/SiNWs/Si, (c) LPD-TiO2 (0.4 M)/SiNWs/Si, (d) LPD-TiO2 (0.5 M)/SiNWs/Si, (e) LPD-TiO2 (0.6 M)/SiNWs/Si, and (f) LPD-TiO2 (0.7 M)/SiNWs/Si.

was continuously oxidized where silver was deposited, causing the silver particles to be etched downward, and finally a regular SiNW array was formed by the metal-assisted wet chemical etching process. Therefore, this study points out that increased AgNO3 concentration can increase the etching rate of the silicon wafer and the porosity of SiNWs, further reducing the reflectance of the SiNWs, as shown in Fig. 2. However, Figs. 2 and 3 indicate that when the AgNO3 concentration was increased to 0.05 M, the reflectance of the SiNWs with a high porosity of 57% was significantly increased. Owing to the porosity of the SiNW layers was determined based on the ratio of porous volume to porous layer total volume. For all experimental conditions, the samples were weighed both before and after the metal-assisted wet chemical etching and after the removal of the porous silicon layer of SiNWs in a 1 N NaOH solution (the SiNWs structures were removed based on the characteristics of different etching rates for silicon and SiNWs in the NaOH solution). Porosity and porous layer total volume (Vp) can be described as follows [28]: Porosity ¼

M1 ‐M2 M1 ‐M3

Vp ¼

M1 ‐M3 ρ

the length of the SiNWs would increases from 0.75 to 2.5 μm. However, when the concentration was increased to 0.05 M, the etching rate also became increased, which caused over etching and a collapse of the SiNWs (as shown in Fig. 4(f)), thus causing a larger portion of light to be reflected from the surface. The abovementioned analysis suggests that the AgNO3 solution with a concentration of 0.04 M could lead to long SiNWs with low reflectance, which could be used as a benchmark for optimizing the follow-up study. The traditional ARC requires an optical design and a one-fourth wavelength thickness of the TiO2 deposited on the surface of silicon substrates (LPD-TiO2/Si substrate) (as shown in Fig. 5(a)) to reduce the intensity of reflected light with a specific wavelength to nearly zero, indicating that the antireflection effect was restricted to a specific wavelength. In contrast, the SiNWs exhibit a broadband antireflection effect in this study and were more suitable to be applied in high-efficiency solar cells. Nevertheless, the surface of the SiNWs has numerous defects, which would increase the number of carriers to the surface for recombination. In order to further reduce the reflectance and the recombination

ð5Þ

M1, M2, M3, and ρ represent the silicon wafer weight before the metal-assisted wet chemical etching, the specimen weight after the metal-assisted wet chemical etching, the specimen weight (after being soaked in 1 N NaOH solution) excluding the SiNWs weight, and the silicon specific gravity (ρ = 2330 kg·m−3), respectively. A higher porosity, an important parameter influencing the SiNWs structure, indicates that more silicon atoms were etched off as well as that the silicon column size between pores is smaller. However, an overly high concentration of AgNO3 solution often causes over etching in both longitudinal and lateral directions and leads to a higher calculated porosity value. Hence, the obtained porosity should be illustrated in conjunction with the SEM image. Fig. 4 shows the SEM cross-sectional images for SiNWs after immersion in the nitric acid and HF solution to remove the residual silver particle and oxide layer as a function of the concentration of AgNO3 solution and bare silicon wafer, which indicates that the surface of the silicon wafers without being treated by the AgNO3 solution was smooth, and that SiNWs would be uniformly formed on the surface of the silicon substrates if the aforementioned silicon wafers had been soaked in the AgNO3 solution with a concentration of 0.01–0.04 M. As the concentration of the AgNO3 solution increases,

Fig. 6. The reflectance spectra of LPD-TiO2 thin film on SiNWs as a function of H3BO3 concentration. The inset shows the average reflectance at wavelengths from 400 to 800 nm.

Please cite this article as: J.-J. Huang, et al., Surf. Coat. Technol. (2017), http://dx.doi.org/10.1016/j.surfcoat.2017.01.027

J.-J. Huang et al. / Surface & Coatings Technology xxx (2017) xxx–xxx Table 2 The average reflectance at wavelengths from 400 to 800 nm of LPD-TiO2 thin film on SiNWs as a function of H3BO3 concentration. Concentration of H3BO3 (M)

0.3

0.4

0.5

0.6

0.7

Average reflectance (%)

8.3

4.6

3.6

3.2

2.8

rate of surface carriers, the characteristics of LPD (the coating was not influenced by the shape of substrates) were used to deposit a TiO2 film satisfying the relationship λ = 4nd on the SiNWs of optimal parameters using HF 0.4 M and AgNO3 0.004 M. Because reflection was minimal when the thickness of the ARC layer, in the case, the thickness and average refractive index (average refractive index of different H3BO3 concentrations of TiO2 thin films for a wavelength of 550 nm) of the LPD-TiO2 layer were 78 nm and 1.76, respectively. As shown in Fig. 5(b)–(f), the SEM images concerning the deposited TiO2 film (corresponding to the boric acid solutions with different concentrations) indicate that the LPD method can be used to perfectly deposit TiO2 film on SiNWs. The TiO2 film remained uniform as the concentrations of boric acid increased from 0.5 to 0.7 M. However, when the boric acid concentration decreased below 0.5 M, the TiO2 film surface became rough, and the pore spaces among the SiNWs increased. The main cause can be inferred using the chemical reaction formula (involving ammonium hexafluorotitanate and boric acid), which can be expressed as [29]: ðNH4 Þ2 TiF6 þ2H2 O

↔

H3 BO3 þ 4HF

BF4 ‐ þ H3 Oþ þ2H2 O

ðTiF6 Þ2− þnH2 O

↔ ↔

TiO2 þnNH4 F þ 4HF

TiF66‐n ðOHÞn


2−

þ nHF

ð6Þ ð7Þ ð8Þ

The abovementioned formula reveals that in the process of generating the TiO2 film, the function of boric acid was to inhibit the increase in HF concentration. Therefore, the deposition rate would increase as the boric acid concentration increases. In contrast, in the LPD-TiO2 reaction process, HF solution with a higher concentration would be produced if the boric acid concentration was low, which enhanced the etching capability to etch the SiNWs and TiO2 film. In other words, the boric acid

5

concentration could be adjusted to generate different surface morphology. Moreover, in the process with a low boric acid concentration, the silicon columns were easily etched, and thus the spaces among SiNWs increased. Fig. 6 and Table 2 show the reflectance spectra and average reflectance of the TiO2 film deposited on SiNWs as a function of boric acid concentration, suggesting that as the boric acid concentration was increased from 0.3 to 0.7 M, the TiO2 film quality was enhanced since the increase of trace HF in the solution was inhibited, and thus the reflectance was decreased, with the lowest reflectance of 2.8% occurring at a concentration of 0.7 M. Fig. 6 also suggests that when the boric acid concentration decreases below 0.5 M, the TiO2 film quality was deteriorated because the increase in the HF concentration cannot be inhibited, and thus the reflectance was significantly increased. In summary, adjusting the boric acid concentration could cause different TiO2 film surface morphology and spaces among the SiNWs, which can further influence the reflectance variation of the SiNWs. The reflectance uniformity and stability of LPD-TiO2/SiNWs/Si samples (the TiO2 film deposited by H3BO3 at 0.5 M) were further investigated; the uniformity of samples detection method was to calculate the average reflectivity of five samples of LPD-TiO2/SiNWs/Si (the average reflectance of the LPD-TiO2/SiNWs/Si samples was measured at five positions). The average reflectances of the five samples were 3.60%, 3.65%, 3.70%, 3.76%, and 3.72%. The uniformity of the LPD-TiO2/SiNWs/ Si samples was calculated as 2.17% by U% = [(max − min)/(max + min)]100%, where max and min are the maximum and minimum average reflectance of the five samples of LPD-TiO2/SiNWs/Si. The stability of samples detection method was to calculate the decay rate of LPD-TiO2/ SiNWs/Si samples after it has been placed in air for 180 days. The reflectance of LPD-TiO2/SiNWs/Si sample was 3.64% with the decay rate of 1.09% after it has been placed for 180 days. The decay rate of the samples was calculated as the formula of [Abs(b − a)/b]100%, where a and b are the average reflectance of samples initially (R = 3.60%) and placed in air for 180 days. Thus, the reflectance of the LPD-TiO2 thin films deposited on SiNWs/Si has better uniformity and stability. In order to expound the reflectance influence of LPD-TiO2 film deposition on the SiNWs, this study proposed a hypothesis concerning the influence of optical behaviors of incident light wavelengths on the LPD-TiO2/SiNWs/Si wafer and SiNWs/Si wafer, wherein four cases

Fig. 7. Schematic diagram of the four representative processes between incident light (λ) and the spaces among SiNWs (d) of LPD-TiO2/SiNWs/Si wafer and SiNWs/Si wafer. (a) λ b b d (b) λ N d (c) λ N N d (d) λ b d.

Please cite this article as: J.-J. Huang, et al., Surf. Coat. Technol. (2017), http://dx.doi.org/10.1016/j.surfcoat.2017.01.027

6

J.-J. Huang et al. / Surface & Coatings Technology xxx (2017) xxx–xxx

were considered, as shown in Fig. 7. Fig. 7(b) indicates that the SiNWs cause a very low reflectance, which was accounted for by the fact that the size of the SiNWs array was smaller than or approaching the wavelength of visible light, and hence light would not be interfered or diffracted (when passing through the SiNWs structure). In other words, the SiNWs array constrained the reflectance. However, if the surface of the etched structure was very rough and had an overly large etched opening, the optical behavior would be similar to that shown in Fig. 7(a) and its reflectance behavior cannot be effectively restricted, which indicates that a larger proportion of light was reflected from the surface, and hence the reflectance was significantly increased. The LPD process provided good coating capability, and local coating deposit would not occur when the LPD-TiO2 film was deposited on the SiNWs, (i.e., the LPD-TiO2 film would be uniformly deposited on the SiNWs, as shown in Fig. 7(c)). As a result, the spaces among the SiNWs were reduced, making the size of the SiNWs array even smaller than the wavelength of visible light (corresponding to the coating deposition condition when the boric acid concentration was higher than 0.5 M). Moreover, due to the uniformly LPD-TiO2 film deposited on the SiNWs, the proportion of light reflected from the top surface of the SiNWs would be reduced. Consequently, the reflectance of the LPDTiO2/SiNWs/Si wafer was lower than that of the SiNWs/Si wafer. When the LPD-TiO2 film deposition process was performed in a lowconcentration boric acid solution, the etching capability of HF to etch the SiNWs and the TiO2 film was enhanced because an increase of trace HF in the solution failed to be inhibited. Hence, more rough film appearances were formed and the spaces among the SiNWs were increased, as illustrated in Fig. 7(d). The results would significantly affect the reflectance of the SiNWs. Because the spaces among the SiNWs were significantly increased, a large proportion of the incident light was reflected from the surface. Following the confirmation of the optical characteristics of the TiO2 film deposited on the SiNWs, this study further delved into the passivation performance of the TiO2 film deposited on the SiNWs/Si. Fig. 8 demonstrates the τeff of the LPD-TiO2/SiNWs/Si as a function of the H3BO3 concentrations, and SiNWs/Si in the HF and AgNO3 solution with a concentration of 0.4 M and 0.04 M, suggesting that the LPDTiO2 film deposited on the SiNWs/Si was helpful in increasing the passivation performance. Because the LPD-TiO2 process was a trace amount of HF relative to the SiNWs of metal-assisted wet chemical etching. The trace amount of HF in the LPD-TiO2 process can be controlled by H3BO3, lead to the deposition rate higher etching rate, and trace amounts of HF will etching the silicon substrate, to form a Ti− Si1−xOy interface layer between the LPD-TiO2/SiNWs interface. Thus, it can passivate the

Fig. 8. Effective minority carrier lifetime of LPD-TiO2 thin film deposition on SiNWs as a function of H3BO3 concentration.

surface of the SiNWs and oxygen vacancies of the TiO2 structure, to enhance the film quality of LPD [26]. Fig. 9(a) shows the Ti 2p3/2 XPS spectra of the interfacial layer of the LPD-TiO2/SiNWs/Si wafer after annealing at 500 °C (the TiO2 film used H3BO3 0.5 M). This can be deconvoluted into three components with binding energies 455.6, 457.7, and 459.3 eV [30,31], which correspond to the TiO, Ti2O3 and TiO2 bonds, respectively, indicating that oxygen within the TiO2 film was deficient (the peak area of TiO, Ti2O3 and TiO2 were 18.38, 30.69 and 50.93, as shown in the inset table in Fig. 9(a)). Fig. 9(b) shows the O 1 s XPS spectra of the interfacial layer of the LPD-TiO2/SiNWs/Si wafer after annealing at 500 °C (the TiO2 film used H3BO3 0.5 M). This can be deconvoluted into three components with binding energies 530.3, 531.1, and 532.2 eV [31–33], which correspond to the TiO2, Ti2O3 and SiO2 bonds, respectively, indicating trace amounts of HF etching the surface of SiNWs, to form of a Ti– Si1−xOy interlayer between the LPD-TiO2/SiNWs interface. Thus, it can passivate the surface of the SiNWs and oxygen vacancies of the TiO2 structure, to enhance the film quality of LPD. However, the XPS analysis revealed that different concentrations of H3BO3 would result in different passivation degrees, the peak areas of SiO2 and Ti2O3 decrease from

Fig. 9. XPS spectra of the interfacial layer of the LPD-TiO2/SiNWs/Si wafer after annealing at 500 °C (the TiO2 film was using H3BO3 0.5 M) (a) the Ti 2p3/2 XPS spectra and the peak area of TiO, Ti2O3 and TiO2 as shown in the inset table; (b) the O 1 s XPS spectra and the peak areas of TiO2, Ti2O3 and SiO2 as shown in the inset table.

Please cite this article as: J.-J. Huang, et al., Surf. Coat. Technol. (2017), http://dx.doi.org/10.1016/j.surfcoat.2017.01.027

J.-J. Huang et al. / Surface & Coatings Technology xxx (2017) xxx–xxx

32.61% to 31.46% and 32.61% to 31.46% as the H3BO3 concentration increases, as shown in the inset table in Fig. 9(b), which displays the peak areas of LPD-TiO2 thin films deposited on SiNWs/Si as a function of H3BO3 concentration from 0.3 to 0.7 M. Owing to the fact that in the process of generating the TiO2 film, boric acid was used to control the trace amount of HF in the solution. Thus, when the boric acid concentration increased (exceeded 0.5 M), the trace amount of HF was inhibited, which resulted in a LPD-TiO2 film for the high degree of oxygen vacancies and lower Ti–Si1−xOy, and thus a low passivation. Although boric acid solutions with a concentration below 0.5 M could enhance the passivation performance of the TiO2 film by the function of fluorine and Ti– Si1−xOy; however, it could not completely inhibit the decrease of HF in the solution and could possibly generate a low quality and rough TiO2 film, thereby influencing the τeff of the TiO2 film. The results indicate that the optimal surface passivation performance occurs when the boric acid concentration is 0.5 M, which resulted in a higher quality TiO2 film and higher passivation performance. 4. Conclusions The TiO2 film was deposited on the silicon nanowires as ARCs and the passivation layer using LPD, which was a low-cost and simple method that could be used at a low deposition temperature and was suitable for coating a large area. Under optimal conditions, the reflection and τeff of the liquid phase deposited TiO2 film were 3.6% and 1.29 μs, respectively, for the H3BO3 of 0.5 M after annealing at 500 °C in N2 atmosphere for 30 min. These results demonstrate that the LPD-TiO2 films exhibit favorable antireflection and passivation properties, which are comparable to other vacuum deposition methods, such as PECVD and ALD; thus, the LPD-TiO2 films were highly favorable for silicon nanowire-based solar cells. Acknowledgement The authors thank the Ministry of Science and Technology of the Republic of China, Taiwan, for financially supporting this research under contract No. MOST 104-2221-E-212-019. The authors also thank the Center for Micro/Nano Science and Technology, National Cheng Kung University for supporting material analysis. References [1] R.S. Wagner, W.C. Ellis, Vapor-liquid-solid mechanism of single crystal growth, Appl. Phys. Lett. 4 (1964) 89. [2] T. Ogino, M. Yamauchi, Y. Yamamoto, K. Shimomura, T. Waho, Preheating temperature and growth temperature dependence of InP nanowires grown by self-catalytic VLS mode on InP substrate, J. Cryst. Growth 414 (2015) 161. [3] X. Duan, C.M. Lieber, Laser-assisted catalytic growth of single crystal GaN nanowires, J. Am. Chem. Soc. 122 (2000) 188. [4] P. Hiralal, C. Chien, N.N. Lal, W. Abeygunasekara, A. Kumar, H. Butt, H. Zhou, H.E. Unalan, J.J. Baumberg, G.A.J., Nanowire-based multifunctional antireflection coatings for solar cells, Nanoscale 6 (2014) 14555. [5] C.H. Chao, C.H. Chan, J.J. Huang, L.S. Chang, H.C. Shih, Manipulated the band gap of 1D ZnO nano-rods array with controlled solution concentration and its application for DSSCs, Curr. Appl. Phys. 11 (2011) S136. [6] G. Jia, J. Westphalen, J. Drexler, J. Plentz, J. Dellith, A. Dellith, G. Andra, F. Falk, Ordered silicon nanowire arrays prepared by an improved nanospheres selfassembly in combination with Ag-assisted wet chemical etching, Photonics Nanostruct. Fundam. Appl. 19 (2016) 64.

7

[7] E.S. Kooij, K. Butter, J.J. Kelly, Silicon etching inHNO3/HF solution: charge balance for the oxidation reaction, Electrochem. Solid-State Lett. 2 (1999) 178. [8] M.D. Kelzenberg, S.W. Boettcher, J.A. Petykiewicz, D.B. Turner-Evans, M.C. Putnam, E.L. Warren, J.M. Spurgeon, R.M. Briggs, N.S. Lewis, H.A. Atwater, Enhanced absorption and carrier collection in Si wire arrays for photovoltaic applications, Nat. Mater. 9 (2010) 239. [9] J.A. Czban, D.A. Thompson, R.R. LaPierre, GaAs core-shell nanowires for photovoltaic applications, Nano Lett. 9 (2009) 148. [10] Y. Cui, J. Wang, S.R. Plissard, A. Cavalli, T.T.T. Vu, R.P.J. van Veldhoven, L. Gao, M. Trainor, M.A. Verheijen, J.E.M. Haverkort, E.P.A.M. Bakkers, Efficiency enhancement of InP nanowire solar cells by surface cleaning, Nano Lett. 13 (2013) 4113. [11] M.V. Fernandez-Serra, C. Adessi, X. Blase, Conductance, surface traps, and passivation in doped silicon nanowires, Nano Lett. 6 (2006) 2674. [12] J.E. Allen, E.R. Hemesath, D.E. Perea, J.L. Lensch-Falk, Z.Y. Li, F. Yin, M.H. Gass, P. Wang, A.L. Bleloch, R.E. Palmer, L.J. Lauhon, High-resolution detection of Au catalyst atoms in Si nanowires, Nat. Nanotechnol. 3 (2008) 168. [13] Y. Dan, K. Seo, K. Takei, J.H. Meza, A. Javey, K.B. Crozier, Dramatic reduction of surface recombination by in situ surface passivation of silicon nanowires, Nano Lett. 8 (2011) 2527. [14] V. Pore, T. Kivela, M. Ritala, M. Leskela, Atomic layer deposition of photocatalytic TiO2 thin films from TiF4 and H2O, Dalton Trans. 6467 (2008). [15] S. Yamauchi, S. Saiki, K. Ishibashi, A. Nakagawa, S. Hatakeyama, Low pressure chemical vapor deposition of Nb and F Co-doped TiO2 layer, J. Cryst. Process Technol. 4 (2014) 79. [16] A. Kafizas, N. Noor, P. Carmichael, D.O. Scanlon, C.J. Carmalt, I.P. Parkin, Combinatorial atmospheric pressure chemical vapor deposition of F:TiO2; the relationship between photocatalysis and transparent conducting oxide properties, Adv. Funct. Mater. 24 (2014) 1758. [17] T. Giannakopoulou, N. Todorova, T. Vaimakis, S. Ladas, C. Trapalis, Study of fluorinedoped TiO2 sol-gel thin coatings, J. Sol. Energy Eng. 130 (2008) 410071. [18] N.S. Leyland, J. Podporska-Carroll, J. Browne, S.J. Hinder, B. Quilty, S.C. Pillai, Highly efficient F, Cu doped TiO2 anti-bacterial visible light active photocatalytic coatings to combat hospital-acquired infections, Sci. Rep. 6 (2016) 24770. [19] J. Yu, Y.L. Yang, R.Q. Fan, L. Li, L.G. Wei, Mechanism of performance enhancement via fluorine doped titanium dioxide nanoparticles in dye sensitized solar cells, J. Fluor. Chem. 176 (2015) 71. [20] G. Ren, Y. Gao, X. Liu, A. Xing, H. Liu, J. Yin, Synthesis of high-activity F-doped TiO2 photocatalyst via a simple one-step hydrothermal process, React. Kinet. Mech. Catal. 100 (2010) 487. [21] W. Ho, J.C. Yu, S. Lee, Synthesis of hierarchical nanoporous F-doped TiO2 spheres with visible light photocatalytic activity, Chem. Commun. 100 (2006) 1115. [22] D. Li, H. Haneda, S. Hishita, N. Ohashi, N.K. Labhsetwar, Fluorine-doped TiO2 powders prepared by spray pyrolysis and their improved photocatalytic activity for decomposition of gas-phase acetaldehyde, J. Fluor. Chem. 126 (2005) 69. [23] H. Maki, Y. Okumura, H. Ikuta, M. Mizuhata, Ionic equilibria for synthesis of TiO2 thin films by the liquid-phase deposition, J. Phys. Chem. C 118 (2014) 11964. [24] C.C. Lin, J.J. Huang, D.S. Wuu, C.N. Chen, Surface passivation property of aluminum oxide thin film on silicon substrate by liquid phase deposition, Thin Solid Films 618 (2016) 118. [25] D.S. Wuu, C.C. Lin, C.N. Chen, H.H. Lee, J.J. Huang, Properties of double-layer Al2O3/ TiO2 antireflection coatings by liquid phase deposition, Thin Solid Films 584 (2015) 248. [26] J.J. Huang, Y.T. Lee, Self-cleaning and antireflection properties of titanium oxide film by liquid phase deposition, Surf. Coat. Technol. 231 (2013) 257. [27] M.J. Kerr, A. Cuevas, Very low bulk and surface recombination in oxidized silicon wafers, Semicond. Sci. Technol. 17 (2002) 35. [28] R. Herino, G. Bomchil, K. Barla, C. Bertrand, Porosity and pore size distributions of porous silicon layers, J. Electrochem. Soc. 134 (1987) 1994. [29] Y. Gao, Y. Masuda, T. Yonezawa, K. Koumoto, Site-selective deposition and micropatterning of SrTiO3 thin film on self-assembled monolayers by the liquid phase deposition method, Chem. Mater. 14 (2002) 5006. [30] M. Hannula, K. Lahtonen, H. Ali-Loytty, A.A. Zakharov, T. Isotalo, J. Saari, M. Valden, Fabrication of topographically microstructured titanium silicide interface for advanced photonic applications, Scr. Mater. 119 (2016) 76. [31] I. Bertoi, M. Mohai, J.L. Sullivan, S.O. Saied, Surface characterisation of plasmanitrided titanium:an XPS study, Appl. Surf. Sci. 84 (1995) 357. [32] Q. Fang, M. Meier, J.J. Yu, Z.M. Wang, J.Y. Zhang, J.X. Wu, A. Kenyon, P. Hoffmann, I.W. Boyd, FTIR and XPS investigation of Er-doped SiO2–TiO2 films, Mater. Sci. Eng. B 105 (2003) 209. [33] C. Anandan, P. Bera, XPS studies on the interaction of CeO2 with silicon in magnetron sputtered CeO2 thin films on Si and Si3N4 substrates, Appl. Surf. Sci. 283 (2013) 297.

Please cite this article as: J.-J. Huang, et al., Surf. Coat. Technol. (2017), http://dx.doi.org/10.1016/j.surfcoat.2017.01.027