Photovoltaic and Charge Trapping Characteristics of Screen- Printed Monocrystalline Silicon Solar Cells with Molybdenum Oxides as Hole-Selective Layers by H2/Ar Plasma Pretreatment

Photovoltaic characteristics of screen-printedmonocrystalline silicon solar cells (SPSSCs) with molybdenum oxide (MoOx) as holeselective layers (HSLs) were demonstrated. A H2/Ar plasma pretreatment (PPT) was incorporated into a MoOx/p-Si(100) interface, which shows the expected quality in terms of passivation. Moreover, the charge trapping characteristics of the MoOx/p-Si(100) interface were presented. The PPT parameters, including power, treated time, flow ratio of H2/Ar, and temperature, were investigated. The experimental results indicate that the Si-H bond with a relatively high intensity was demonstrated for the H2/Ar PPT. The achievement of a conversion efficiency (CE) improvement of more than 1.2% absolute from 18.3% to 19.5% for SPSSCs with H2/Ar PPT was explored. The promoted mechanism was attributed to the reduction of the interface trap density caused by the large number of Si-H bonds at the silicon substrate and MoOx interface.


Introduction
Passivated emitter and rear cell (PERC) silicon solar cells are the mainstream of the industrial solar cells today [1][2][3]. To increase the capture of photons by solar cells, PERC introduces two additional layers at the rear of the cell compared with SPSSCs [4]. The function of those layers is to improve the passivation of the rear side and increase the reflection of the back light [5]. However, PERC still have a small amount of Al back surface field (Al-BSF), which will cause recombination [6]. Thus, to improve this shortcoming, transition metal oxides, such as MoO 2 , V 2 O 5 , and WO 3 , are proposed as hole-selective layers (HSLs) [7][8][9][10]. The energy band of the p-type silicon substrate was matched with the HSL. Moreover, oxygen vacancy defects in the film can be utilized to transfer holes [11]. Therefore, the laser local opening and Al-BSF can be prevented.
Literatures reported that hydrogen (H 2 ) PPT in fabricating silicon heterojunction solar cells can be adopted to improve surface atomic hydrogen coverage and avoid the generation of defects on the silicon substrate [12]. The H 2 PPT of the Si surface can lead to improvements in minority carrier lifetime [13]. Moreover, the H 2 PPT improves surface passivation compared to classical HF dip [14]. Furthermore, carbon and oxygen contamination can be removed from silicon surfaces by a 30 s H 2 plasma treatment [15]. Thus, to enhance the interface characteristics of silicon and MoO x , a H 2 /Ar PPT was investigated.

Experimental Methods
To demonstrate the effects of H 2 /Ar plasma treatment on photovoltaic characterizations of the SPSSCs, square samples (20 × 20 mm 2 ) of (100)-oriented p-type silicon wafers with 1-3 Ω cm and 190 ± 20 μm were prepared. The texturing process was performed in a solution of 1.73% KOH at 83°C for 10 min. The front emitters were formed by phosphorus diffusion at 840°C for doping and drive-in. After single-side edge isolation and PSG etching processes, the sheet resistances of the front emitters were demonstrated to be approximately 100 ± 10 Ω/sq. For antireflection coating, a standard SiN x film with a thickness of 75 nm was deposited on the n + -emitters using plasma-enhanced chemical vapor deposition at a frequency of 13.56 MHz. Next, an Ag front paste was screen-printed on the front side of the SPSSCs and dried in an infrared belt furnace at 230°C. Then, a 6-zone industrial infrared belt furnace was utilized to fire Ag pastes into the ntype emitters. The peak temperature and the belt speed were set at 790°C and 508 cm/min, respectively. To protect the front metallization of the SPSSCs, the polymer paste was spin-cast onto the front surface of the SPSSCs at 3000 rpm and dried at 150°C for 30 min. Then, the H 2 /Ar PPT was performed on the rear of the cells at power ranged from 50 to 70 W. The treated time was tuned from 0 to 70 s. The H 2 /ðH 2 + ArÞ ratio and temperature were achieved from 22.2 to 66.6% and 200 to 300°C, respectively. The cell without H 2 /Ar PPT was presented as a reference. Next, all HSLs consist of MoO x /Ag films which were achieved by a thermally evaporated technique. Deposition of the approximately 4 nm thick MoO x films was performed by thermal evaporation from granules of MoO 2 (99.9% purity). The thicknesses of evaporated Ag were 500 nm. The CEs of the SPSSCs were measured under standard test conditions (AM 1.5G spectrum, 100 mW/cm 2 , 25°C). The cross-section images and elemental analysis of p-Si(100)/MoO x /Ag were examined by transmission electron microscopy (TEM) and energydispersive X-ray spectroscopy (EDS), respectively. The Si-H bonds of the H 2 /Ar PPT samples were measured using Fourier-transform infrared spectroscopy (FTIR).

Results and Discussion
The cross-section image and elemental analysis of p-Si(100)/MoO x /Ag stacked films were examined by the TEM and EDS with a line drawn as shown in Figure 1. The thicknesses of the MoO x HSLs and SiO x were demonstrated to be approximately 4 and 1.8 nm, respectively. The SiO x interfa-cial layer is present at the MoO x /Si(100) interface during MoO x evaporation [16]. Elemental analysis of the p-Si(100)/SiO x /MoO x /Ag stacked films by EDS is shown in Figure 1(b). The MoO x were thermally evaporated at a base pressure of 5 × 10 −6 Torr. The Mo/O atomic ratio in the MoO x layers was approximately 1/2.3 for the MoO 2 source granules. Literature reported that the Mo +4 , Mo +5 , or Mo +6 states could be presented in MoO x HSL [17]. In this work, the O/Mo atomic ratio of 2.3 in the MoO x HSLs for the MoO 2 granule source was attributed to some oxygen already being in the chamber during evaporation [16]. Figure 2 shows CEs vs. treated time curves of the SPSSCs with and without H 2 /Ar PPT. The H 2 /ðH 2 + ArÞ ratio and   Figure 3. The CE increases as the H 2 /ðH 2 + ArÞ ratio is increased, until an optimum of the H 2 /ðH 2 + ArÞ ratio is reached. Above this optimum condition, the CE decreases as the H 2 /ðH 2 + ArÞ ratio increases. The upgraded mechanism is related to a change in surface hydrogen configuration toward lower hydrides [13]. Excessive H 2 /Ar PPT will cause the surface roughness to further degrade the characteristics of the cells [14]. Achieving CE improvement of more than 0.6% absolute from 18.9% to 19.5% for SPSSCs with a tuning H 2 /ðH 2 + ArÞ ratio was explored. Thus, a H 2 /Ar PPT was incorporated into a MoO x /p-Si(100) interface, which shows the expected qualities in terms of passivation.
For further examining the photovoltaic characteristics of the SPSSCs, Figure 4 shows CEs vs. temperature curves of the SPSSCs. As the duration of temperature exposure increased, the CE increased to 250°C and then decreased. The increase in CE due to the PPT is familiarly considered to result from passivation of surface defects. It has been reported that the 3 International Journal of Photoenergy smaller effective surface recombination velocity can be achieved by the higher deposition temperature [14]. The results indicate that the best result in this work is a CE of 19.5% at deposition temperature ranging from 250 to 300°C.
Because MoO x /silicon contacts induce slightly Fermi level bending on p-type silicon, the interface characteristics were addressed by a MoO x /n-type silicon structure [18]. Figure 5 shows the interface trap density (D it ) of n-Si(100)/SiO x /MoO x /Ag capacitors (a) with and (b) without H 2 /Ar PPT extracted from the conductance method. The conductance (G p ) of n-Si(100)/SiO x /MoO x /Ag capacitors as a function of frequency (ω) was measured and plotted as G p / ωA versus voltages by biasing the Si surface in depletion condition [19]. A was denoted as the area of the capacitor. Figure 5 shows that G p /ωA of the cells without H 2 /Ar PPT has a maximum value and D it is equal to ð2:5/qÞ ðG p /ωAÞ max at that maximum [20]. It is also observed that the D it values of the cells with and without H 2 /Ar PPT were 1:8 × 10 11 and 1:4 × 10 12 eV -1 cm -2 , respectively. Obviously, a H 2 /Ar PPT is beneficial to improve the interface characteristics of MoO x and silicon. To further understand the improvement mechanism, FTIR spectra were addressed.
Infrared spectra of the cells with and without H 2 /Ar PPT were measured using FTIR. The experimental conditions for FTIR were resolution of 1 cm -1 , source of Glowbar which can provide FTIR measurement of wavenumbers ranging from 1000 to 3000 cm -1 , and detector mercury-cadmiumtelluride. Figure 6 illustrates the FTIR spectra of the cells with and without H 2 /Ar PPT. The spectra present an absorption peak at 2170 cm -1 , which is recognized as the Si-H bonds [21]. The absorbance values for nontreatment samples at 2170 cm -1 are smaller than those of the H 2 /Ar PPT ones, demonstrating that the amount of Si-H bonds on the silicon surface is increased by PPT in the H 2 /Ar atmosphere. Figure 7(a) indicates dark and illuminated (AM 1.5G) current density vs. voltage curves of the SPSSCs with and without H 2 /Ar PPT. A CE of SPSSCs with PPT in H 2 /Ar ambient was demonstrated to be around 19.5%. The CE was found to increase after the silicon substrate was exposed to H 2 /Ar PPT for 50 s before the evaporation of MoO x .
Moreover, the open-circuit voltages of 666 and 646 mV for SPSSCs with and without H 2 /Ar PPT, respectively, were achieved as shown in Figure 7(a). This can be proved by the optimization of the interface trap charge, as shown in Figure 5. To understand the improvement mechanism, external quantum efficiencies (EQEs) of the SPSSCs with and without H 2 /Ar PPT are shown in Figure 7(

Conclusions
A H 2 /Ar PPT process was applied to monocrystalline silicon substrate before depositing the MoO x HSLs. The enhanced passivation effects of H 2 /Ar PPT were investigated by using a conductance method of measuring D it , in combination with FTIR spectroscopy. The achievement of a CE improvement of more than 1.2% absolute from 18.3% to 19.5% for SPSSCs with H 2 /Ar PPT was explored. The promoted mechanism could be attributed to the reduction of D it and the increase of the Si-H bonds at MoO x and Si substrate.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare that they have no conflicts of interest.