Influence/Effect of Deep-Level Defect of Absorber Layer and n/i Interface on the Performance of Antimony Triselenide Solar Cells by Numerical Simulation

Abstract

The antimony sulphide (AnS) solar cell is a relatively new photovoltaic technology. Because of its attractive material, optical, and electrical qualities, Sb2Se3 is an excellent absorption layer in solar cells, with a conversion efficiency of less than 8%. The purpose of this research is to determine the best parameter for increasing solar cell efficiency. This research focused on the influence of absorber layer defect density and the n/i interface on the performance of antimony trisulfide solar cells. The researchers designed the absorber thickness values with the help of the SCAPS-1D (Solar Cell Capacitance Simulator-1D) simulation programme. For this purpose, they designed the ZnS/Sb₂Se₃/PEDOT: PSS planar p-i-n structure, and then simulated its performance. This result confirms a Power Conversion Efficiency (PCE) of ≥25% at an absorber layer thickness of >300 nm and a defect density of 10¹⁴ cm⁻³, which were within the acceptable range. In this experiment, the researchers hypothesised that the antimony triselenide conduction band possessed a typical energy of ≈0.1 eV and an energetic defect level of ≈0.6 eV. At the n/i interface, every condition generated a similar result. However, the researchers noted a few limitations regarding the relationship between the defect mechanism and the device performance.

1. Introduction

To address the issue of global warming, there has recently been an increased demand for renewable energy sources. There are many renewable energy sources, including wind, geothermal, bio, and solar energy, with solar energy harvesting offering the most potential in a wide range of applications. Furthermore, photovoltaic technology is regarded as an effective technique for directly converting an abundance of solar energy into electricity, thereby allowing environmental sustainability based on the decrease in carbon emissions and the greenhouse effect [1]. Despite high-efficiency solar cells, Si-solar cells have not reached everyday household purposes due to the high cost. Chapin, Fuller and Pearson (1954) at Bell Laboratories first introduced silicon-based solar cells with an efficiency of 4%, now called first-generation solar cells [1]. Up until now, Si-solar cells have achieved more than 20% PV efficiency over the last four decades and have consequently dominated the PV industry [2]. In the past, the thin-film solar cell technology that used Cadmium Telluride (CdTe) and Copper Indium Gallium Selenide (CIGS) displayed a higher Power Conversion Efficiency (PCE), i.e., 22.1% and 23.35%, respectively [2]. Furthermore, even Perovskite Solar Cells (PSCs) displayed a 25.8% efficiency, though they could not be commercialised [3]. The common materials used for manufacturing solar cells include copper (Cu), cadmium (Cd), indium (In), gallium (Ga), selenium (Se), etc. However, the toxicity of Cd, and the expensive nature of Ga and In, has increased the demand for developing cost-effective materials that are both ecologically-friendly and abundant, as they would help in increasing the market for sustainable solar energy [4]. Messina et al. (2009) reported for the first time the use of Sb2Se3 as an absorber layer for a solar cell by chemical bath deposition (CBD) with a 0.66% concentration. Owing to its exceptional light absorption coefficient (>10⁵ cm⁻¹ for sunlight) and an ideal bandgap (1.1–1.2 eV), antimony chalcogenides such as Sb₂Se₃ have garnered a lot of research interest in the past few years [5]. Due to its properties, Sb₂Se₃ does not allow carrier recombination at the grain boundaries of traditional cubic absorbers, as it possesses a ribbon-like Q1D crystalline structure. A lot of research conducted on Sb₂Se₃ improved its PCE from 5% to 9.2%. Kharade et al. noted a PCE of 17.75% when the absorber layer thickness was increased to 700 nm. Many factors were attributed to this improved PCE, such as better absorber growth processes (such as close space sublimation, vapour transport deposition and hydrothermal processes), buffer layer optimisation (i.e., ZnO, CdTe and TiO₂) and novel architecture designs of the device [5,6]. However, the main weakness of the study is the failure to address how the device’s performance can be improved. Moreover, the main weakness of the Sb₂Se₃ cells is the failure to address how the device’s efficiency with suitable materials. The objective of this research is to use SCAPS simulations with the ZnS/Sb₂Se₃/PEDOT:PSS/Au device structure performance to investigate the influence of deep-level defect density on the absorber layer and the n/I interface of antimony triselenide. The main issues addressed in this paper are (a) open-circuit voltage, (b) short-circuit current, (c) fill factor, and (d) PCE in the p-i-n planar structure. A considerable amount of literature has been published on antimony chalcogenide with HTL/Sb₂Se₃/ETL as a basic structure. While holes are extracted from the rear contact after suppressing negative band bending based on numerical simulations [7], it was noted that the antimony chalcogenide (such as Sb₂Se₃) solar cells required a better Hole Transport Layer (HTL) compared to other solar cells. Furthermore, the inorganic PbS colloidal quantum dot film also improved the Sb₂Se₃ efficiency from 5.42 to 6.50% [8]. A few researchers used an inorganic CuSCN-HTL to increase the PCE of the Sb₂Se₃ cells by 7.5%. The common HTL used by the researchers in different solar cells includes 2,2,7,7-tetrakis (N,N-di-p-methoxyphenylamine)-9,9-spirobi-fluorene (Spiro-OMeTAD). It was also used in the Sb₂ (S, Se)₃ solar cells with a PCE value of <10% [9,10,11]. However, owing to their expensive nature and erratic device performance, it is not advisable to use organic HTLs for large-scale production. Hence, for manufacturing efficient photovoltaic (PV) cells, a new stable, non-toxic, and cost-effective material must be designed. Some researchers used technology based on the loss process and HTL modification approach, such as the PEDOT: PSS. This technique is fairly easy to implement and is commonly used for investigating the effectiveness of HTLs in perovskite devices [12]. Furthermore, the PEDOT:PSS is cost-effective and high-quality, which makes it a good option as an HTL for PSC development. Though PEDOT:PSS is effectively used as an HTL for PSCs, it displays a poor performance and stability in comparison to the other HTLs owing to its highly doped nature [13,14]. This can lead to many problems, such as acute interfacial recombination. Hence, the researchers modified and improved the PEDOT:PSS conductivity by additional p-doping as it helps in matching the energetics and improving the performance of the device [15,16].

Moreover, to design TiO₂ crystals, TiO₂ is used as an ETL and is generally annealed at temperatures above 450 °C [17]. A few metal oxides that are treated at a lower temperature were investigated as ETLs in the PSCs to prevent their annealing at high temperatures, such as SnO₂ [18], In₂O₃ [19], WO₃ [20], amorphous TiO_x [21], Zn₂SnO₄ [22], La-doped BaSnO₃ [23], ZnO [23], etc. Out of the above ETLs, ZnO could be regarded as a prospective ETL as it displays ultra-high electron mobility (205–300 cm⁻²/Vs) [24]. Several researchers implemented techniques for improving the semiconductor property of ZnO and, subsequently, the photovoltaic performance of the PSCs by doping and designing ZnO using other metal oxides or elements [25]. ZnS is similar to the ZnO wide bandgap semiconductor and displays similar physical features. In the case of quantum-dot-sensitised solar cells, ZnS presented excellent electron mobility and could function like the ETL and interfacial passivation layer [26]. ZnS possesses a low Conduction Band Minimum (CBM) compared to ZnO, which makes it a more effective match for MAPbI₃-LUMO. However, the photovoltaic performance of the PSCs still needs to be improved by using ZnO or ZnS to allow the transport of electrons from MAPbI₃ to ZnO, although the ZnS-based PSCs showed a poor photovoltaic performance in the past [27]. The theory states that the open-circuit voltage (V_oc) of the PSCs is based on the energy difference between the CBM (ELT) and the Valence Band Maximum (VBM) [26]. Due to this fact, the addition of ZnS to the ZnO-based PSCs would help in increasing their V_oc value [28]. The addition of ZnS on ZnO surfaces acts like an energy barrier that prevents charge recombination between the ZnO and MAPbI₃ [27]. However, this could amplify the short-circuit current of the ZnO-based PSCs. Overall, there seems to be some evidence to indicate that antimony sulphide solar cells have room to improve cell efficiency.

2. Materials and Methods

2.1. Modelling

The researchers modelled and simulated the Sb₂Se₃ absorber-based thin-film solar cells (TFSC), using the SCAPS-1D programme, which is a 1D solar cell device simulator that was previously developed at the University of Gent [29]. The SCAPS-1D simulation helps in solving the Poisson’s equation, carrier continuity equation, and drift-diffusion equations, which are regarded as the basic equations used for developing a semiconductor device [30]. Thereafter, the study investigated the different properties of the PV devices, such as the capacitance–voltage (C-V), current–voltage (I-V) and capacitance–frequency (C-f), along with their EQE and recombination profiles using the SCAPS-1D simulator. This research showed that the best method for this investigation was to numerically use Sb₂Se₃-based solar cells. The results indicated that factors such as cell thickness, electron affinity, doping concentration, operating temperature, bulk defect density, interface defect density, metal work function and resistance affect the device performance. All simulations were carried out under the following conditions: Temperature = 300 K, illumination of 100 mWcm⁻² and AM1.5 G light spectrum. Figure 1 presents the ZnS/Sb₂Se₃/PEDOT:PSS heterojunction device structure, while Figure 1 depicts the TFSC structure.

The structure showed that the p-type Sb₂Se₃ absorber and rear metal contact were connected by the ultra-thin p-type PEDOT: PSS as HTL. In the device, the ETL and TCO layers included the n-type ZnS and fluorine-doped tin oxide (F:SnO₂). Furthermore, the front exposure Glass/FTO and rear Al metal contacts in the TFSCs were constructed using aluminium (Al) [31].

Table 1. A list of various simulation parameters involved in the perovskite planar structure.

Properties	ZnS (ETL)	Sb2Se3	PEDOT: PSS (HTL)	References
Thickness (nm)	70	200–1200	40	[27,32]
Bandgap, Eg (eV)	3.5	1.04	2.2	[27,33,34]
Electron affinity, $x_e$ (eV)	4.5	4.04	2.9	[35,36,37]
Dielectric permittivity, ${\mathsf{ϵ}}_{\mathrm{r}}$ (relative)	10	18	3	[35,38,39]
CB effective density of states, NC (cm−3)	1.5 $\times$ 1018	2.2 $\times$ 1018	2.2 $\times$ 1015	[36,39,40]
VB effective density of states, NV (cm−3)	1.8 $\times$ 1018	1.8 $\times$ 1019	1.8 $\times$ 1018	[36,39,40]
Electron thermal velocity (cm/s)	1 $\times$ 107	1 $\times$ 107	1 $\times$ 107	[31,34,38]
Hole thermal velocity	1 $\times$ 107	1 $\times$ 107	1 $\times$ 107	[31,34,38]
Electron mobility (cm2/Vs)	50	15	10	[36,39,41]
Hole mobility (cm2/Vs)	20	5.1	10	[36,39,41]
Shallow uniform donor density, ND (cm−3)	1 $\times$ 1022	0	0	[36]
Shallow uniform acceptor density, NA (cm−3)	0	1.0 $\times$ 1017	3.17 $\times$ 1014	[31,36]
Defect density Nt (cm−3)	1 $\times$ 1014	6.90 $\times$ 1013−16	1 $\times$ 1016	[31,36]

depicts the physical parameters included in the simulation. The values of material parameters used for the numerical computation were derived from the literature. The holes and electrons travelled at a speed of 1 × 10⁷ cm s⁻¹ [31]. In this study, the SCAPS-1D software was used to depict the absorption coefficients of the ZnS and PEDOT:PSS HTL materials, while the absorption coefficient values for the Sb₂Se₃ absorber and PEDOT:PSS values were taken from the literature. As mentioned earlier, the defects were similar, and the researchers selected their Gaussian distribution of energetics. The energy levels of the defects were set to mid-bandgap. The investigation included the ETL/absorber and absorber thickness/defect densities for determining the heterojunction TFSC interface carrier recombination. All parameters for the interface faults are presented in Table 1.

2.2. Numerical Modelling

SCAPS–1D was used for solving Poisson’s equation for holes and electrons (Equation (1)), as follows [42]:

$$\frac{{\mathrm{d}}^2\Psi }{\mathrm{d}{x}^2}=\frac{\mathrm{e}}{{\in }_0{\in }_r}\left[P\left(x\right)-n\left(x\right)+N_D-N_A+{\rho }_P-{\rho }_n\right]$$

(1)

Here, $x$ = electrostatic potential; $\Psi$ = elementary charge; ${\in }_r$= relative permittivity; ${\in }_0$ = vacuum permittivity; p = concentration of holes and electrons; N_D, N_A refer to the donor and acceptor charges; while ${\rho }_p$ and ${\rho }_n$ refer to the concentration of holes and electrons, respectively.

Based on a theoretical perspective, the carrier lifetime ($\tau$) is described as the period where the charge carrier can move freely and contribute to the electric conduction. The G electron-hole pairs were developed after a uniform simulation of the Sb₂Se₃. The electron and hole densities developed in E_c and E_v can be described using Equations (2) and (3) as follows [43]:

$$\Delta n=G{\tau }_n$$

(2)

$$\Delta p=G{\tau }_p$$

(3)

If the above carriers were trapped in Sb₂S₃ and thermally recited, the time that was spent in traps was not included in the ${\tau }_n$ and ${\tau }_p$. In the case of a steady state, the rate of generation in the PSC was seen to be equal to the rate of trapping, as shown in Equations (3) and (4) below [43]:

$${\tau }_n=\frac{1}{{\sigma }_p.v_{th.}\left(N_t-N_r\right)}$$

(4)

$${\tau }_{\rho }=\frac{1}{{\sigma }_p.v_{th.}{n}_r}$$

(5)

where $N_t, n_r, {\sigma }_n \mathrm{and} v_{th}$ refer to the total (i.e., occupied and unoccupied), occupied defects, hole capture cross-section, and the thermal velocity, respectively. Equations (4) and (5) indicate that the increase in defect density decreased the carrier lifetime and led to a short carrier diffusion length (L) [44]. This L value was based on the perovskite quality. It was noted that if the L value was higher than the perovskite thickness, the device performance could be improved. Based on the L, recombination current (J_o) and V_oc, the following equations were derived:

$$J_0\approx q\frac{D{n}_i^2}{LN}$$

(6)

$$V_{oc}=\frac{KT}{q}1n\left(\frac{J_{sc}}{J_0}+1\right)$$

(7)

Equations (6) and (7) show that a decrease in the defect density decreased the recombination current and increased the V_oc, which was similar to the results in this study. It was noted that the J_sc and Internal Quantum Efficiency (IQE) were directly proportional to one another. If the researchers presume that the PSC are a shallow junction solar cells with a long minority carrier lifetime, the IQE can rely on the minority carrier diffusion lengths, as presented in Equation (8):

$$IQE=1-\alpha t-\frac{B}{\alpha L^2}$$

(8)

where α, t, and B indicated the spectral absorption coefficient, distance in the perovskite material and perovskite thickness, respectively [45]. The Shockley–Read–Hall (SRH) recombination is used for clarifying the recombination components that affected the defect density in the PSC, as follows:

$${\Re }^{SRH}=\frac{\vartheta {\sigma }_n{\sigma }_p{N}_T\left[np-n_i^2\right]}{{\sigma }_p\left[p+p_1\right]+{\sigma }_n\left[n+n_1\right]}$$

(9)

where ϑ = electron thermal velocity, N_T = N_o. of defects per volume, n_i + intrinsic number density, n and p = Concentration of electrons and holes at equilibrium, and n₁ and p₁ = Conc. of electrons and holes in the trap defect and valence band, respectively. Hence, it was noted that the defect density negatively affected the PSC. The researchers noted that this issue can be resolved by increasing the thickness of the absorber layer.

3. Results and Discussion

3.1. Effect of Sb₂Se₃ Absorption Layer Thickness on the PSC Performance

The Sb₂Se₃ layer thickness plays a vital role in the PSC performance. Different researchers noted that the absorber layer thickness affected the performance of the device at wavelengths ranging between 200 and 1200 nm. The V_oc showed the optimal value of 0.75 V at 700 nm, which decreased with an increase in the absorber layer thickness as the recombination rate increased in the thick films. Figure 2b shows that the J_sc increased rapidly when the absorption layer thickness increased from 200 to 600 nm and then increased steadily beyond 700 nm. The minimal value of J_sc was 34.03 mA cm⁻². Owing to the saturated absorption of the Sb₂Se₃ layer, the absorber layer showed a thickness of 200 nm, while the maximal J_sc value was 45.05 mA cm⁻², at a layer thickness of 1200 nm. An increase in the carrier recombination in the thick layers decreased the J_sc when the absorption layer thickness increased beyond the diffusion length. A thick absorption layer decreased the J_sc value if the diffusion length was higher than the absorption layer thickness.

When the absorption layer thickness increased from 200 to 1200 nm, the FF decreased monotonically (Figure 2c), where the final FF was 84.79%. If the solar cell was paired with an optimal load, the FF reflected the maximal power production. There was an increase in the internal resistance with an increase in the absorption layer, which further accelerated the depletion and decreased the FF value [46]. If the Sb₂Se₃ layer thickness was increased, there was an increase in the PCE, which subsequently decreased to 28.25 at 900 nm, as shown in Figure 2d. A few factors such as light absorption and carrier conveyance helped in assessing the cell efficiency during simulation. Light absorption was a vital factor that helped to determine cell efficiency if the absorption layer was <900 nm. As the layer thickness was smaller than the carrier diffusion length, many carriers reached a corresponding electrode, where the efficiency was limited by the carrier transit. The conveyance of carriers determined the cell’s effectiveness if the absorption layer was thick since the light absorption became saturated when it reached a specific critical thickness. Based on these criteria, it was noted that the Sb₂Se₃ solar cell with a thickness of 900 nm showed the best performance. It displayed a PCE of 28.25%, V_OC of 0.7461 V, J_SC of 44.65 mA cm⁻² and the efficacy factor (FF) of 84.81%.

3.2. Effect of Defect Density of the Absorption Layer on the PSC Performance

Figure 1 presents the equilibrium band diagram of the Sb₂Se₃ solar cell having a p-i-n structure. The absorber layer showed a defect energy level of 0.6 eV that was smaller than the conduction band. Figure 3 presents the different performance metrics of the Sb₂Se₃ solar cells, in addition to their absorber defect density and layer thickness. The device’s performance is assessed by investigating the effect of the deep-defect density and absorber layer thickness (ranging between 10¹³ and 10¹⁷ cm⁻³; 200 and 1200 nm, respectively). The V_oc of the device ranged between 0.73 and 0.778 V, where the device’s performance was affected by fault density. The highest V_oc value of 0.79 V was noted at the defect density of 10¹³ cm⁻³ and layer thickness of 200 nm. Furthermore, it was noted that at the defect densities of >10¹³ and an absorber layer thickness of >200 nm, the V_oc decreased rapidly to 0.778, before steadily decreasing to a value of 0.73 V. The maximal J_sc of >44 mA/cm² was noted at a layer thickness of 800 nm and a defect density of 10¹⁴ cm⁻³. The graph shows that the J_sc decreased from 43 mA/cm² to 32 mA/cm² if the thickness value decreased from 600 nm to 200 nm. A similar trend was noted when the defect densities were >10¹⁵ cm⁻³. Factors such as V_oc and recombination play a role in decreasing the value of FF [47]. The results indicated that the FF value was higher than 80% at the flaw density of 10¹⁴ cm⁻³. This was not dependent on the thickness value. Based on PCE = V_oc, J_sc and FF equations, the researchers described efficiency based on the combination of the three output characteristics. The highest PCE of >25% was noted if the thickness value was >900 nm and the defect density was <10¹³ cm⁻³. The efficiency of the Sb₂Se₃ film can be improved if it is grown perpendicular to the p-n junction contact in a [001] direction, based on the findings presented in earlier studies [9,12]. On the other hand, it was noted that a defect density of >10¹⁵ cm⁻³ showed a significant performance loss of ≈10–15%, while the PCE dropped down to 20% when the layer thickness was below 900 nm.

3.3. Effect of the Interface Defect Density on the PSC Performance

Figure 4 shows the effect of different ZnS/Sb₂Se₃ layer interfacial defects on the device’s performance. The energy level at 0.6 eV was lower than the E_c value of the ZnS layer. This factor shows a similar trend to the above. However, the device gradually becomes more vulnerable to defects with time. As shown in the results, the defect density decreased from 10¹⁴ cm⁻³ to 10¹¹ cm⁻³. A value of 0.75 V was noted in a small region when the absorber thickness was 700 nm and the interface defect density was <10¹¹ cm⁻³. Beyond this range, the V_oc decreased to 0.95 V at >10¹⁵ cm⁻³. It was also noted that the J_sc was not sensitive to defect density and increased from 22 to 27 mA/cm² when the absorber thickness increased from 300 to 1000 nm. The researchers observed that the thickness of the absorber layer did not significantly affect the FF value; however, a fault concentration of >10¹³ cm⁻³ could decrease the FF by ≈50%. Finally, a maximal PCE of 28% was noted if the thickness was >700 nm and the defect densities were >10¹¹ cm ⁻³, respectively. Jako S. Eensalu et al., (2019) reported that a 3.5% efficiency by vacuum deposition methods was experimentally achieved [48]. Based on the experiments, it was seen that a higher series of resistance at the ZnS/absorber interface was due to light absorption. This increased the effective and electrical dissociation of the excitons and transport charge carriers, thus improving the performance of the ZnS film-based Sb₂Se₃ solar cells [5]. This was attributed to the diffusion of metal cations in the Sb₂Se₃ during the developmental stage. Owing to the numerous structural problems in the different materials, major interfacial deformations occurred, which led to charge recombination in the solar cell devices.

4. Conclusions

In this investigation, the aim was to assess and optimise the development of the Sb₂Se₃ solar cells with the help of the SCAPS simulation for the p-i-n configuration. Furthermore, the study set out a primary SC based on the TCO/ZnS/Sb₂Se₃/PEDOT: PSS structure. The main goal of the current study was to determine the absorber layer thickness (ranging between 200 and 1200 nm) and the defect density (ranging from 1013 to 1017 cm³) at an energy level of 0.6 eV. The most prominent finding to emerge from this study is that which showed a PCE of >25%, V_oc of 0.75 V, FF of 80%, and J_sc of 44 mA/cm². Therefore, it generally seems that an increase in the efficiency of the solar cell coincides with an increase in the absorber layer thickness. Furthermore, the results suggest that a defect density of 10¹⁴ cm⁻³ and absorber layer thickness of >300 nm was satisfactory. Thereafter, the researchers measured the ZnS/absorber layer thickness and studied its effect on the device performance. Despite varying the thickness of the absorber layer (200 and 1200 nm) and defect densities (10¹⁰ and 10¹⁶ cm⁻³), they noted the same energy level (0.6 eV). The device’s performance was highly sensitive to the defect densities; however, it showed a similar pattern to the absorber layer thickness. When the absorber layer thickness was increased from 300 to 900 nm, it significantly increased the PCE. Unfortunately, the study showed that when the thickness value was increased from 1000 to 1200 nm, there was a slight decrease in the PCE. Although the current study is based on mathematical simulations, the results suggest that device design and the optimization of Sb₂Se₃ solar cells could greatly assist in experimental research.