Introduction: The role of a P Cat-doped layer in the passivation of a c-Si surface . 63

63

Chapter 4 Improvement in the passivation quality of a

64

Although P concentration in a P Cat-doped layer is high, donor activation ratio is less than 2% [4]. The sheet carrier density of the P Cat-doped layer varies from 5×10¹¹ to 5×10¹² cm^-2 depending on T_S-dope and t_dope. The shallow P doped layer with suitable sheet carrier density can therefore induce field-effect passivation, which can significantly suppress the recombination of minority carriers on a c-Si surface. Although the shallow doping can also be realized by other techniques such as plasma doping, atomic layer deposition of dopants, and molecular beam epitaxy [7-11],Cat-doping can significantly avoid damage onto a c-Si surface induced by energetic ions since gas molecules are decomposed on a heated wire by catalytic reaction. It has been demonstrated that there is no serious damage from generated radicals onto a c-Si surface at T_cat of 1300 ^oC, at which P doped layer is formed with high sheet carrier density [2]. This advantage makes Cat-doping become a favorable method for the formation of field-effect passivation layers for c-Si. Regarding the application of P Cat-doped layers to passivation technique, it has been already reported that the addition of P Cat-doped layers can reduce the SRVmax of n-type c-Si passivated with an a-c-Si film from 5 to 3 cm/s [12]. In the previous chapter, I have optimized the preparation conditions of SiNx passivation films with a refractive index of

~2.0 at a wavelength of 630 nm prepared by Cat-CVD for n-type c-Si wafers. The highest τeff of 3 ms, corresponding to a low SRVmax of 5 cm/s, can be obtained for n-type c-Si passivated with Cat-CVD SiNx films deposited at a low substrate temperature (~100 °C) and post annealing. The use of SiNx films, whose refractive indexes are adjusted to be 2.0 even after decreasing the substrate temperatures, can avoid optical loss due to parasitic absorption in a-Si for SiNx/a-Si stacked passivation system. The Cat-CVD SiNx films with high passivation quality and high transparency are suitable for application to c-Si solar cells. As I mentioned above, a P Cat-doped layer can induce field-effect passivation, which is effective in suppressing surface recombination by sending electrons away from the c-Si surface. In this chapter, in order to obtain even lower SRVmax on a c-Si surface passivated with a SiN_x film, I attempt to apply P Cat-doping for field-effect passivation. The structure of a SiN_x/P Cat-doped layer is shown in Figure 4.2. The effects of sheet carrier density and annealing as well as H etching on the passivation quality of SiN_x/P Cat-doped layers on a c-Si surface are also investigated.

65

Figure 4.2 Cross-sectional schematic of a c-Si passivated with SiNx/P Cat-doped layers.

4.2 Experimental procedure

Table 4.1 Sample preparation conditions of P Cat-doping, a-Si, and SiNx films.

Sample preparation conditions are summarized in Table 4.1. All c-Si wafers were first cleaned in diluted (5%) HF solution for 10 s to remove native oxide. P Cat-doped layers, SiN_x films and a-Si films were prepared in separate Cat-CVD chambers. 2.25%

helium-diluted PH₃ was used as a gas source for doping process. A tungsten wire (W) with

Doping a-Si SiNx

Gas sources PH3 20 sccm SiH4: 10 sccm SiH4: 8 sccm NH3: 150 sccm Substrate Temperature (Ts-dope) 80-300 ^oC 90 ^oC 100 ^oC

Pressure (P) 1.0 Pa 0.55 Pa 10 Pa

Catalyzer temperature (T_cat) 1300 ^oC 1800 ^oC 1800 ^oC

Catalyzer-substrate distance 12 cm 8 cm 8 cm

Time (tdope) 30-120 s 30 s 184 s

66

a diameter of 0.5 mm and a length of 210 cm was used as a catalyzer in the P doping chamber. Distance between catalyzer and substrate was 12 cm to reduce heat radiation from the heated catalyzer. The schematic of a P Cat-doping chamber is described in Figure 4.3.

Figure 4.3 Schematic of a P Cat-doping chamber.

In this chapter, the properties of P Cat-doped layers were varied by changing doping substrate temperature (Ts-dope) and doping time (tdope). The deposition condition of SiNx films and annealing conditions for the samples after depositing the SiNx films were the same as the optimized conditions, under which high τ_eff of 3 ms can be obtained for a SiNx/c-Si structure, as reported in Chapter 3. 290-µm-thick n-type (100) FZ Si wafers with a resistivity of 2.5 Ωcm and a bulk minority carrier lifetime of >10 ms were used for the investigation of the passivation quality. The structure for τeff measurement is shown in Figure 4.2. In order to investigate the effect of annealing on passivation quality, the P Cat-doped samples were annealed before and after depositing SiNx films. In this paper, I refer them to “annealing A” and “annealing B”, respectively. Annealing A and B were both conducted in a horizontal tubular furnace in nitrogen atmosphere. Two samples were prepared under the same P doping condition at the same batch; one is for a sample with only annealing B, and the other is for a sample with both annealing A and B. The samples

67

without annealing A were passivated with SiNx films immediately after P Cat-doping without air break, while the samples with annealing A were taken out from the P doping chamber, followed by furnace annealing and then SiNx deposition without any additional cleaning prior to the deposition. All the SiN_x-deposited samples were finally annealed at 350 ^oC for 30 min (annealing B). The τ_eff of the samples was measured by μ-PCD using a 904 nm wavelength pulse laser with a photon density of 5×10¹³ cm^-2, as described in Chapter 2. I also measured excess-carrier-density- (Δn-) dependent τeff by quasi-steady-state photoconductance (QSSPC) (WCT-120, Sinton Instruments).

Figure 4.4. Cross-sectional schematic of a sample for the Hall effect measurement.

The Hall effect measurement and SIMS were employed to evaluate the sheet carrier density (N_D) of P Cat-doped samples and P concentration, respectively. The properties of c-Si wafers used for SIMS measurement are the same as those used for τ_eff measurement.

The SIMS measurement was performed from the back side of the samples after removing most of Si wafers in order to avoid the effect of knock-on and resulting unintentional broadening of P profiles. I used 2900 Ωcm p-type FZ c-Si wafers for the Hall effect measurement. High resistivity samples are used for preventing leakage current from a P Cat-doped layer to a c-Si substrate. The reason of choosing p-type wafers here is to form a depletion layer, which is also to avoid current to a c-Si substrate. The capture of carriers at defects on a c-Si surface and oxidization may affect significantly the results of the Hall effect measurement [4,13,14]. In order to prevent these effects, a 10-nm-thick a-Si film was deposited on c-Si immediately after P Cat-doping without air exposure. In order to know the effect of annealing on ND, samples were annealed at 350 ^oC before and after depositing a-Si films. Four Al electrodes were formed by evaporation through a metal hard

68

mask to form the van der Pauw configuration. The samples were annealed at 350 ^oC for 1 min to make Ohmic contact between Al electrodes and a P Cat-doped layer. Figure 4.4 shows the cross-sectional schematic of a sample for the Hall effect measurement. The details of the measurement have been described in Ref. 3. More details of the Hall effect measurement is also summarized in Appendix A-6. The effect of H etching on the morphology of c-Si surface after P Cat-doping was evaluated by atomic force microscopy (AFM).