Wet Chemical Oxidation to Improve Interfacial Properties of Al 2 O 3 / Si and Interface Analysis of Al 2 O 3 / SiO x / Si Structure Using Surface Carrier Lifetime Simulation and Capacitance–Voltage Measurement

: A thin silicon oxide (SiO x ) layer (thickness: 1.5–2.0 nm) formed at an Al 2 O 3 / Si interface can enhance the interface properties. However, it is challenging to control the characteristics of thin SiO x layers because SiO x forms naturally during Al 2 O 3 deposition on Si substrates. In this study, a ~1.5 nm-thick SiO x layer was inserted between Al 2 O 3 and Si substrates by wet chemical oxidation to improve the passivation properties. The acidic solutions used for wet chemical oxidation were HCl:H 2 O 2 :H 2 O, H 2 SO 4 :H 2 O 2 :H 2 O, and HNO 3 . The thicknesses of SiO x layers formed in the acidic solutions were ~1.48, ~1.32, and ~1.50 nm for SiO x-HCl , SiO x-H2SO4 , and SiO x-HNO3 , respectively. The leakage current characteristics of SiO x-HNO3 were better than those of the oxide layers formed in the other acidic solutions. After depositing a ~10 nm-thick Al 2 O 3 on an SiO x-acidic / Si structure, we measured the e ﬀ ective carrier lifetime using quasi steady-state photoconductance and examined the interfacial properties of Al 2 O 3 / SiO x-acidic / Si using surface carrier lifetime simulation and capacitance–voltage measurement. The e ﬀ ective carrier lifetime of Al 2 O 3 / SiO x-HNO3 / Si was relatively high (~400 µ s), resulting from the low surface defect density (2.35–2.88 × 10 10 cm − 2 eV − 1 ). The oxide layer inserted between Al 2 O 3 and Si substrates by wet chemical oxidation helped improve the Al 2 O 3 / Si interface properties.


Introduction
The surface passivation of crystalline silicon (c-Si) solar cells can be improved using various materials such as SiO 2 [1][2][3][4][5][6], SiN x [7][8][9], Al 2 O 3 [10][11][12][13], TiO x [14][15][16], MoO x [17,18], and poly-Si [19][20][21][22]. In particular, Al 2 O 3 thin films are most widely used for boron-doped Si surfaces (or p + emitter surfaces) owing to the low surface recombination velocity (SRV) [10,11]. The interface of Al 2 O 3 films can be characterized by a low interface trap density and a high negative charge density because of the ultrathin SiO x and negatively charged interstitial O ions present at the Al 2 O 3 /Si interface [23,24]. The passivation properties of an Al 2 O 3 /Si structure can be improved by optimizing the Al 2 O 3 thickness and annealing temperature [25,26]. In particular, a thin SiO x layer is formed at the Al 2 O 3 /Si interface during the deposition process and can be activated through annealing. However, it is difficult to control the quality of this layer because it forms spontaneously.
In this work, we considered a method to improve the interfacial properties of an Al 2 O 3 /Si structure by inserting a thin silicon oxide layer between Al 2 O 3 and Si substrates. The widely used methods of growing thin silicon oxides are thermal and wet chemical oxidations. In thermal oxidation, although the quality of the silicon oxide formed is excellent, it is difficult to control the thickness (<1.5 nm) at temperatures above 800 • C. If the silicon oxide layer becomes thicker, the field effect passivation of the Al 2 O 3 /Si interface may deteriorate [27]. Therefore, a thickness of~1.5 nm or less is required to improve the interfacial properties of the Al 2 O 3 /Si structure and ensure a good field effect passivation. Wet chemical oxidation is a promising method of growing thin SiO x , and the quality of thin oxides grown using this method has been verified [28][29][30][31][32]. In this research, a thin oxide layer was utilized to improve the passivation property of an Al 2 O 3 /Si structure. We applied wet chemical oxidation to grow~1.5 nm-thick SiO x and then deposited a~10 nm-thick Al 2 O 3 using plasma-assisted atomic layer deposition (PA-ALD). To analyze the passivation characteristics and interface properties of the Al 2 O 3 /SiO x /Si structure, the effective carrier lifetime was measured using quasi steady state photoconductance (QSSPC), and surface carrier lifetime simulation and capacitance-voltage (C-V) measurements were performed.

Experimental
After the Radio Corporation of America (RCA) cleaning developed by Werner Kern to remove ionic and organic impurities in the polished 4~5 Ω·cm p-type Si(100) substrate [33], we performed wet chemical oxidation using three acidic solutions: (i) HCl:H 2 O 2 :H 2 O = 1:1:5, (ii) H 2 SO 4 :H 2 O 2 :H 2 O = 1:1:5, and (iii) HNO 3 (68%) [28][29][30]. The wet chemical oxide layers formed in the acidic solutions are hereinafter abbreviated as SiO x-HCl , SiO x-H2SO4 , and SiO x-HNO3 , respectively. The process temperature was 85 • C for SiO x-HCl and SiO x-H2SO4 and 121 • C for SiO x-HNO3 . The immersion time was varied from 10 to 60 min. The thickness and refractive index of the silicon oxide (SiO x-acidic ) layers formed on Si surface was measured using spectroscopy ellipsometry (SE) and transmission electron microscopy (TEM). The quality of SiO x-acidic was evaluated in terms of the leakage current density, which was measured by conducting a current-voltage (I-V) analysis with mercury probe and Keithley 238 current source meter under dark conditions. To measure the surface passivation characteristics, an approximately 10 nm-thick Al 2 O 3 film was deposited on a SiO x-acidic /Si substrate using PA-ALD. The reaction sources for Al 2 O 3 deposition were trimethylaluminum (Al(CH 3 ) 3 , TMA) and O 2 (purity 99.999%) gas; the purge gas was Ar (purity 99.999%). The deposition process was performed at a substrate temperature of 250 • C, a process pressure of 1.0 torr, a plasma power of 200 W, an O 2 exposure time of 0.5 s, and a distance of 20 mm between the showerhead and the substrate. Annealing was then conducted in an electric furnace at 425 • C for 15 min to activate the Al 2 O 3 layer. After annealing, the carrier lifetimes of the samples were measured using QSSPC to evaluate the surface passivation characteristics. For a detailed analysis of the interface properties of Al 2 O 3 /SiO x-acidic /Si, a surface carrier lifetime simulation and C-V measurement were performed by mercury probe with Agilent E4980A LCR meter. Figure 1 shows the thickness of the silicon oxide (SiO x-acidic ) formed in the acidic solutions with various immersion times. The film thickness was measured by SE. The measured wavelength range was 300-100 nm and the incident beam angle was 75 • in the SE measurement. With the increase in the immersion time, the thickness increased, saturating in the range of 1.32-1.50 nm. The thicknesses of SiO x are~1.48 nm for SiO x-HCl ,~1.32 nm for SiO x-H2SO4 , and~1.50 nm for SiO x-HNO3 . These results are similar to those reported previously [28][29][30]. The thickness of SiO x-HNO3 was also confirmed by TEM Energies 2020, 13, 1803 3 of 10 measurement, as shown in Figure 2. In the TEM measurements, silicon nitride (SiN x ) was used as the capping layer to avoid additional thin oxide growth in the Al 2 O 3 deposition process. The SiO x-HNO3 thickness was in the range of 1.43-1.54 nm, consistent with SE measurements. As the thicknesses of SiO x formed in each acidic solution saturated after approximately 15 min, the immersion time was fixed at 15 min (Table 1). To evaluate the characteristics of the SiO x-acidic layer, the leakage current densities were measured by I-V curves under dark conditions. As shown in Figure 3, the leakage current density of SiO x-HNO3 was the lowest (~9.1 × 10 −3 A/cm 2 at 1 V forward-bias and~0.92 × 10 −3 A/cm 2 at −1 V reverse-bias ), indicating that the quality of SiO x-HNO3 is relatively better than the other oxides. The leakage current density results are similar to those reported by Kobayashi Asuha et al. [30]. fixed at 15 min. To evaluate the characteristics of the SiOx-acidic layer, the leakage current densities were measured by I-V curves under dark conditions. As shown in Figure 3, the leakage current density of SiOx-HNO3 was the lowest (~9.1 × 10 −3 A/cm 2 at 1 Vforward-bias and ~0.92 × 10 −3 A/cm 2 at −1 Vreverse-bias), indicating that the quality of SiOx-HNO3 is relatively better than the other oxides. The leakage current density results are similar to those reported by Kobayashi Asuha et al. [30].   were measured by I-V curves under dark conditions. As shown in Figure 3, the leakage current density of SiOx-HNO3 was the lowest (~9.1 × 10 −3 A/cm 2 at 1 Vforward-bias and ~0.92 × 10 −3 A/cm 2 at −1 Vreverse-bias), indicating that the quality of SiOx-HNO3 is relatively better than the other oxides. The leakage current density results are similar to those reported by Kobayashi Asuha et al. [30].      To investigate the surface passivation properties of the SiOx-acidic, a ~10 nm-thick Al2O3 layer was deposited on both sides of the SiOx-acidic/Si substrate using PA-ALD, and the effective carrier lifetime was measured by QSSPC. The measured effective carrier lifetimes at 1.0 sun injection level [34] of Al2O3/SiOx-HNO3/Si, Al2O3/SiOx-HCl/Si, and Al2O3/SiOx-H2SO4 /Si are ~400, ~317, and ~332 μs ( Figure 4); notably, the passivation quality of SiOx,HNO3 is excellent. Moreover, the reference sample, which did not form an oxide film intentionally, exhibited a significantly lower effective carrier lifetime (~220 μs) than the samples with wet chemical oxides. This indicates that the quality of the thin SiOx layer formed at the Al2O3/Si interface influences the effective carrier lifetime and that the oxides formed in acidic solutions exhibit better passivation properties than oxide films formed naturally during the deposition process. To investigate the interfacial properties of Al2O3/SiOx-acidic/Si, the surface carrier lifetime simulation and C-V measurement were performed. The surface carrier lifetime simulation is based on the extended Shockley-Read-Hall (SRH) recombination equation. This equation was derived from the SRV, expressed in Equation (1) and Equation (2) [35,36].

Results and Discussions
where S is the surface recombination velocity, Us is the surface recombination rate, and Δns is the excess carrier density at the surface, Sn0 and Sp0 are the surface recombination velocities of electrons and holes, ns and ps are the electron and hole concentrations at the surface, σn and σp are the capture cross-sections for electrons and holes, n1 and p1 are parameter in the SRH recombination equation and Dit is the number of surface states per unit area. However, if band bending exists on the Si surface by fixed charges (such as negative or positive charges), it is difficult to evaluate Δns because the carrier To investigate the surface passivation properties of the SiO x-acidic , a~10 nm-thick Al 2 O 3 layer was deposited on both sides of the SiO x-acidic /Si substrate using PA-ALD, and the effective carrier lifetime was measured by QSSPC. The measured effective carrier lifetimes at 1.0 sun injection level [34] of Al 2 O 3 /SiO x-HNO3 /Si, Al 2 O 3 /SiO x-HCl /Si, and Al 2 O 3 /SiO x-H2SO4 /Si are~400,~317, and~332 µs ( Figure 4); notably, the passivation quality of SiO x,HNO3 is excellent. Moreover, the reference sample, which did not form an oxide film intentionally, exhibited a significantly lower effective carrier lifetime (~220 µs) than the samples with wet chemical oxides. This indicates that the quality of the thin SiO x layer formed at the Al 2 O 3 /Si interface influences the effective carrier lifetime and that the oxides formed in acidic solutions exhibit better passivation properties than oxide films formed naturally during the deposition process. To investigate the interfacial properties of Al 2 O 3 /SiO x-acidic /Si, the surface carrier lifetime simulation and C-V measurement were performed. The surface carrier lifetime simulation is based on the extended Shockley-Read-Hall (SRH) recombination equation. This equation was derived from the SRV, expressed in Equations (1) and (2) [35,36].
where S is the surface recombination velocity, U s is the surface recombination rate, and ∆n s is the excess carrier density at the surface, S n0 and S p0 are the surface recombination velocities of electrons and holes, n s and p s are the electron and hole concentrations at the surface, σ n and σ p are the capture Energies 2020, 13, 1803 5 of 10 cross-sections for electrons and holes, n 1 and p 1 are parameter in the SRH recombination equation and D it is the number of surface states per unit area. However, if band bending exists on the Si surface by fixed charges (such as negative or positive charges), it is difficult to evaluate ∆n s because the carrier lifetime is determined by the carrier density in the bulk. To this end, we considered the surface of the semiconductor region (z = 0) and the space charge region (z < d sc ), as shown in Figure 5.
Energies 2019, 12, x FOR PEER REVIEW 5 of 11 lifetime is determined by the carrier density in the bulk. To this end, we considered the surface of the semiconductor region (z = 0) and the space charge region (z < dsc), as shown in Figure 5.   lifetime is determined by the carrier density in the bulk. To this end, we considered the surface of the semiconductor region (z = 0) and the space charge region (z < dsc), as shown in Figure 5.   Therefore, the effective surface recombination velocity (S eff ) is the sum of the interface recombination velocity and the space charge recombination velocity, as shown in Equations (3)-(9) [35][36][37][38][39].
S it can be expressed in Equation (4) based on Equations (1) and (2).
S sc can be described by Equation (5) as follows by using Equations (6)-(9) for surface potential (Ψ s ) and surface electron and hole concentrations by effective charge density (Q f ) [34,35].
Here, S it is the interface recombination velocity between the Si surface and the dielectric layer (z = 0), S sc is the recombination velocity in the space charge region (z < d sc ), d sc is the distance of the space charge region, and z is the coordinate perpendicular to the semiconductor surface and increases toward the bulk of Si, V th is the thermal carrier velocity, Eg is the energy band gap of Si, Ψ is the surface potential, Q f is the fixed charge density, Φ n and Φ p are quasi-fermi level for electrons and holes, p b and n b are the carrier concentration for electrons and holes, n i is intrinsic carrier concentration. With Equation (3), we can calculate the theoretical effective surface carrier lifetime using Equation (10). 1 where τ eff is the effective surface carrier lifetime, τ it is the carrier lifetime in the interface region between the Si surface and the dielectric layer, τ sc is the carrier lifetime in the space charge region, and W is the thickness of the wafer. The surface carrier lifetime simulation was performed by comparing the effective carrier lifetime (τ measured ) measured using Equation (10). Assuming that the capture cross-section of the electron and hole is σ n = σ p = 1 × 10 15 cm 2 , we analyzed the interface trap density (D it ) and the fixed charge density (Q f ) as variable parameters so that the curves of τ simulated and τ measured matched well. Figure 6 shows the τ simulated and τ measured curves. The surface carrier lifetime analysis results show that the Al2O3/SiOx-HNO3/SiOx sample exhibits a lower Dit (2.35 × 10 10 cm −2 eV −1 ) than the other samples ( Table 2). The effective charge densities are similar regardless of the acidic solution used. Considering the lifetime of the SiOx-HNO3 layer and values of Qf and Dit extracted from the surface carrier lifetime analysis, the passivation properties would be more influenced by Dit than Qf. C-V measurements were also performed to compare the values of Dit and Qf with those obtained from the surface carrier lifetime analysis. The C-V measurement results show that VFB (flat band voltage) of the silicon substrates with wet chemical oxide layers shifts by 1.77-2.08 V toward the positive bias, as shown in Figure 7. The surface carrier lifetime analysis results show that the Al 2 O 3 /SiO x-HNO3 /SiO x sample exhibits a lower D it (2.35 × 10 10 cm −2 eV −1 ) than the other samples ( Table 2). The effective charge densities are similar regardless of the acidic solution used. Considering the lifetime of the SiO x-HNO3 layer and values of Q f and D it extracted from the surface carrier lifetime analysis, the passivation properties would be more influenced by D it than Q f . C-V measurements were also performed to compare the values of D it and Q f with those obtained from the surface carrier lifetime analysis. The C-V measurement results show that V FB (flat band voltage) of the silicon substrates with wet chemical oxide layers shifts by 1.77-2.08 V toward the positive bias, as shown in Figure 7. Table 2. Results of capacitance-voltage (C-V) and surface carrier lifetime analyses.

Sample (Average of Five
Samples)  Q f was calculated using V FB extracted from the measured C-V graph, and D it was obtained using the Terman method (Table 2) [40,41]. The negative Q f values of the silicon oxides formed in the different acidic solutions did not change significantly in the range of 3.03-3.24 × 10 12 cm −2 . However, D it was markedly different depending on the acid solution used (Table 2). Therefore, we confirm that the quality of SiO x at the Al 2 O 3 /Si interface affects the passivation characteristics, and the insertion of a wet Qf was calculated using VFB extracted from the measured C-V graph, and Dit was obtained using the Terman method (Table 2) [40,41]. The negative Qf values of the silicon oxides formed in the different acidic solutions did not change significantly in the range of 3.03-3.24 × 10 12 cm −2 . However, Dit was markedly different depending on the acid solution used (Table 2). Therefore, we confirm that the quality of SiOx at the Al2O3/Si interface affects the passivation characteristics, and the insertion of a wet chemical oxide layer improved the passivation quality, compared with the SiOx layer simultaneously formed during the deposition of Al2O3.

Conclusions
In this study, we improved the interfacial properties of an Al2O3/Si structure by performing wet chemical oxidation on an Si substrate. The thickness of the wet chemical oxides grown on the Si surface in different acidic solutions was in the range of 1.32-1.5 nm, which was similar to that of SiOx naturally formed at the interface of the Al2O3/Si sample during PA-ALD deposition. In particular, SiOx-HNO3 showed a relatively lower leakage current than SiOx-HCl and SiOx-H2SO4. The carrier lifetime of Al2O3/SiOx-HNO3/Si measured by QSSPC to evaluate the passivation characteristics was ~ 400 μs, which was higher than those of the other SiOx-acidic layers (~317 μs for SiOx-HCl and ~332 μs for SiOx-H2SO4). C-V measurement and surface carrier lifetime analysis were performed to analyze the interface

Conclusions
In this study, we improved the interfacial properties of an Al 2 O 3 /Si structure by performing wet chemical oxidation on an Si substrate. The thickness of the wet chemical oxides grown on the Si surface in different acidic solutions was in the range of 1.32-1.5 nm, which was similar to that of SiO x naturally formed at the interface of the Al 2 O 3 /Si sample during PA-ALD deposition. In particular, SiO x-HNO3 showed a relatively lower leakage current than SiO x-HCl and SiO x-H2SO4 . The carrier lifetime of Al 2 O 3 /SiO x-HNO3 /Si measured by QSSPC to evaluate the passivation characteristics was~400 µs, which was higher than those of the other SiO x-acidic layers (~317 µs for SiO x-HCl and~332 µs for SiO x-H2SO4 ). C-V measurement and surface carrier lifetime analysis were performed to analyze the interface characteristics of the Al 2 O 3 /SiO x /Si samples. The fixed charges showed little change with the oxides, and the D it values changed significantly. The wet chemical oxide formed in the HNO 3 solution showed better passivation characteristics than those formed in other acidic solutions, resulting from the low interface trap density of SiO x-HNO3 . We confirmed the improvement in the passivation characteristics of the Al 2 O 3 /Si sample by inserting a wet chemical oxide between Al 2 O 3 and Si substrates. In addition, the values of the parameters associated with the interface properties, such as Q f and D it , obtained from the C-V measurement and surface carrier lifetime simulation were similar. Therefore, this simulation can be a useful tool to analyze interfacial characteristics. Moreover, this study lays a foundation for analyzing the interfacial properties of samples, such as poly-Si/SiO x /Si structures with ultra-thin SiO x , that cannot be analyzed by C-V measurements.