Advances on Sb₂Se₃ Solar Cells Fabricated by Physical Vapor Deposition Techniques

Abstract

Sb₂Se₃, as an earth-abundant and low-toxic material, has emerged as one of the most interesting absorbers for clean renewable power generation technologies. Due to its optical properties, especially bandgap and absorption coefficient, the number of papers on Sb₂Se₃-based solar cells has been constantly increasing in the last ten years, and its power conversion efficiency has raised from 1% in 2014 to 10.57% in 2022. In this review, different Sb₂Se₃ solar cells’ fabrication technologies based on physical vapor deposition are described and correlated to the texture coefficient (ribbon orientation). Moreover, recent research works of the most promising solar cell configurations with different electron-transporting layers and hole-transporting layers are analyzed with a special emphasis on photovoltaic performances. Furthermore, different Sb₂Se₃ doping techniques are discussed. All these aspects are considered as new strategies to overcome the Sb₂Se₃ solar cell’s actual limitations.

1. Introduction

Thin-film solar cells are the most promising alternative to crystalline silicon solar cells because they use less material, are less expensive, and can be integrated into flexible devices. Chalcogenide thin-film solar cells, especially those based on CdTe and Cu(In,Ga)Se2, are the most dominant in the thin-film photovoltaic market due to their good stability and efficiency (22.1% for CdTe and 23.35% for CIGS) [1]. Perovskite solar cells have also achieved high efficiencies up to 23.7% in recent years, but they are still not reliable enough for mass production.

Silicon (Si)-based solar cells are the dominant technology in the photovoltaic (PV) market. They have a high power-to-cost ratio due to their high efficiency (up to 26.7%) and the fact that silicon is abundant on Earth. However, the production of single-crystal silicon wafers is energy-intensive, and silicon must be more than 100 μm thick to absorb enough sunlight [2]. Chalcogenide materials, on the other hand, have a high optical absorption coefficient, which allows them to be made into lighter and more flexible panels, and they also have lower material and energy production costs.

Among chalcogenides, antimony selenide (Sb₂Se₃) has attracted attention as a potential absorber material for thin-film solar cells due to its excellent optical (very high absorption coefficient >10⁵ cm⁻¹) and electrical (p-type conductivity between 10⁻⁷ S/cm and 4.6 × 10² S/cm) properties. It is also a binary semiconductor with a low melting point of 608 °C.

Furthermore, the scarcity of indium and the toxicity of cadmium have led to increased interest in this type of material, which is made from abundant elements such as Se, Sb, and S that are not considered to be highly toxic or carcinogenic.

Another interesting feature of Sb₂Se₃ is that there is still a lot of room for improvement in its research. In fact, with its 1.2 eV band gap [3], the Shockley–Queisser theory predicts a theoretical power conversion efficiency (PCE) of 31.7% [4]. The highest experimental efficiency achieved so far is 10.57% [5]. In comparison, silicon solar cells have been developed for over 70 years and have reached their efficiency limit. In contrast, Sb₂Se₃ solar cells have rapidly increased their efficiency from 1% to 10.57% in less than 10 years.

Sb₂Se₃ forms one-dimensional ribbons along the c-axis through Sb–Se covalent bonds [6]. The ribbons are stacked together in the perpendicular directions by weaker van der Waals forces [7].

The quasi-one-dimensional structure of Sb₂Se₃ induces strong anisotropic properties, especially for electrical conductivity. The carrier mobility of the ribbons oriented along the (001) growth direction is higher than along other directions.

This is because surfaces parallel to the (001) direction, such as (110) and (120) surfaces, have no dangling bonds and consequently lead to reduced nonradiative recombination losses [6].

In particular, the hole mobility along the ribbons is high (45 cm²/Vs at 300 K), while the carrier transport perpendicular to the covalent bond directions is much smaller because it is associated with a hopping conduction mechanism [8].

In order to maximize the efficiency of Sb₂Se₃ solar cells, it is important to align the ribbons closer to the growth direction (perpendicular to the substrate), such as the (211) and (221) crystal orientations. It is also important to minimize the grain orientation along the (hk0) directions.

The low experimental efficiency of Sb₂Se₃ solar cells is mainly due to the following factors:

• Short carrier lifetime: The carrier lifetime is the average time that a charge carrier (electron or hole) can exist before it recombines with another carrier. The short carrier lifetime in Sb₂Se₃ is due to the high density of defects, such as Se vacancies and Sb_Se/Se_Sb antisite defects [9].
• Low hole density: The hole density is the number of holes per unit volume in a semiconductor material. The low hole density in intrinsic Sb₂Se₃ is another factor that limits the efficiency of the solar cells [2].
• Lack of suitable hole and electron transport layer materials: The hole transport layer (HTL) and electron transport layer (ETL) are the materials that transport the holes and electrons, respectively, to the respective electrodes of the solar cell. The lack of suitable HTL and ETL materials is also a challenge for the development of high-efficiency Sb₂Se₃ solar cells [10].

In order to reduce interface/surface defects, strategies such as defect passivation have been explored. Defect passivation is the process of covering the defects with a material that will prevent them from recombining with charge carriers. This can be done by selenizing the surface of the Sb₂Se₃ or by covering it with a passivating layer [11].

The band alignment between the electron transport layer (ETL) and hole transport layer (HTL) and Sb₂Se₃ is not ideal, which causes an inefficient extraction and transport of photocarriers at the interface with the absorber.

The commonly used ETLs include CdS, TiO₂, ZnO, and SnO₂. The HTLs that have been reported in the literature are organic, such as spiro-OMeTAD, P3HT, PCDTBT, and CZ-TA, or inorganic, such as NiO_x, CuSCN, and CuSbSe₂ [12].

In the past few years, several studies have been published on the deposition of Sb₂Se₃ thin films for solar cells. One common method is the spin-coating deposition on mesoporous oxides in dye-sensitized solar cells (DSSC) structures [13]. This method is simple and scalable, but it can be difficult to control the thickness and uniformity of the film.

Another common method is to deposit Sb₂Se₃ from the vapor phase. This can be done using a variety of technique: these methods can produce high-quality films with a good control over thickness and uniformity.

Recently, close-spaced sublimation (CSS) and injection vapor deposition (IVD) have demonstrated power conversion efficiencies (PCE) of 9.2% (in 2019) and 10.12% (in 2022), respectively. The record of 10.57% (in 2022) was achieved using additive-assisted chemical bath deposition (CBD) [5]. Although the number of laboratories involved in Sb₂Se₃ research is still small, the improvements in cell efficiency have been even faster than those achieved in CIGS and CZTS technologies in their first 10 years of research (Figure 1) [14].

Although significant progress has been made in increasing the efficiency of Sb₂Se₃ solar cells, many challenges remain, such as surface/interface recombination, effective doping, and optimal HTL/absorber/ETL band alignment.

The photocarrier mobility is strongly dependent on the grain orientation, so controlling the ribbon orientation is one of the most important goals for Sb₂Se₃ absorbers in photovoltaic applications. The texture coefficient is hence a fundamental parameter for evaluating the potential of different deposition methods. However, although many studies have reported their texture coefficients, the relative values are rarely calculated, making it difficult to directly compare the obtained absorbers.

This paper reviews recent advances in Sb₂Se₃ solar cells fabricated in planar configuration using physical deposition techniques (PDT). The different relative texture coefficients (TCs) have been extrapolated from the available literature data, which allows for the comparison of ribbon alignment for different deposition methods and structures, and their correlation with the main solar cell performance. The different ETL/absorber/HTL solar cell architectures are then discussed, especially considering their band alignment strategies for enhancing photovoltaic performance. Finally, the different doping strategies for Sb₂Se₃-based absorbers are summarized, evaluating the achieved carrier transport properties and the reported solar cell efficiencies.

The paper is divided into four sections. Section 2 describes the different physical deposition techniques (PDTs) used to fabricate Sb₂Se₃ solar cells. Section 3.1 correlates the PDTs and cell configurations with the ribbon orientation. Section 3.2 discusses the different HTL/ETL, interfacial engineering, and band alignment strategies that have been reported in the literature. Section 3.3 shows the different doping strategies for Sb₂Se₃-based absorbers. Finally, the paper summarizes the challenges and outlines future perspectives for the development of Sb₂Se₃ photovoltaics, with a special emphasis on increasing conversion efficiency.

The authors believe that comparing and discussing the recently published works, with a special focus on ribbon orientation, ETL/HTL, and doping issues, can support future research to limit interface and defect recombination, especially affecting the open-circuit voltage (V_oc), and improve photocarrier extraction in Sb₂Se₃ solar cells.

2. Physical Vapor Deposition Techniques for Planar-Type Structure Sb₂Se₃ Solar Cells

Planar-configuration Sb₂Se₃ solar cells have generally achieved better performance than mesoporous structure cells, especially due to a higher open circuit voltage, V_oc. This is because in planar cells, defects introduced by chemical solvents, such as pinholes, are avoided, thus achieving a higher crystallinity.

Planar-type structures can be classified into two configurations: substrate and superstrate.

In the substrate architecture, the device configuration generally includes (starting from the top):

• A transparent electrode layer or metal grid (Al or Au)
• A window layer (for example undoped ZnO)
• An ETL (for example CdS)
• An Sb₂Se₃ absorber layer
• An HTL
• A metallic or conductive oxide back-contact
• A glass substrate.

The substrate architecture comes from the CIGS technology, which exhibits main advantages, such as the versatility in choosing different substrates and the possibility of fabricating tandem, bifacial, and eventually flexible cells [10].

Superstrate solar cells are a type of solar cell where the absorber layer is deposited on top of the other layers. This is in contrast to a substrate solar cell, where the absorber layer is deposited on the bottom.

In superstrate solar cells, the device configuration is generally formed by [10]:

• A metal contact
• An HTL (for example Spiro-MeOTAD, P3HT, NiOx)
• An Sb₂Se₃ absorber layer
• An ETL (for example TiO₂ or CdS)
• A transparent conductive oxide
• Glass (substrate).

Superstrate configuration cells take origin from the CdTe technology and generally exhibit the highest PCE values. This is because the interfaces in superstrate cells are less prone to defects.

In this section, we briefly describe the most interesting PDT for the growth of Sb₂Se₃ absorber layers:

• Magnetron sputtering;
• Close-spaced sublimation (CSS);
• Rapid thermal evaporation (RTE);
• Vapor transport deposition (VTD);
• Injected vapor deposition (IVT);
• Pulsed laser deposition (PLD);
• Pulsed electron deposition (PED).

2.1. Magnetron Sputtering

In radiofrequency magnetron sputtering (RF-MS), argon ions are accelerated by a radiofrequency electric field to hit a target. This removes material from the target, which is then deposited on a substrate that can be heated (Figure 2a).

Sb₂Se₃ is obtained by sputtering material from a high-purity, stoichiometric Sb₂Se₃ target at a working pressure of 0.5–1.5 Pa [17,18,19]. The chamber is filled with pure argon gas (99.99%). The RF power is fixed at 30–40 W.

In an alternative method proposed by C. Ma et al. [20], Sb₂Se₃ films can be deposited through co-sputtering of Sb₂Se₃ and Se targets. The pressure is maintained at 0.5 Pa. The RF power on the Sb₂Se₃ target is fixed at 30 W, while the RF power on the Se target is changed during the processing from 0 W to 15 W.

In another method, a thin film of metallic Sb is deposited by RF magnetron sputtering from a high-purity and dense Sb sputtering target. Highly crystalline Sb₂Se₃ thin films are obtained by a post-growth post-selenization heat treatment to induce the in situ combination reaction (from Se highly pure powder) [21].

2.2. Close-Spaced Sublimation

In CSS, the source material and the substrate are close to each other, at 1–5 mm, at the extremities of a graphite box. Deposition occurs in a vacuum atmosphere (or in certain cases in gases such as H₂, argon, nitrogen). The solid-state source (powder) and the substrate are heated at different temperatures: the former is heated to a temperature that favors sublimation (Figure 2b) while the latter is kept at a lower temperature to promote the condensation in thin-film form. The method offers the advantage of high deposition rates, and it is widely used to deposit semiconductor materials such as CdTe [22]. More recently, this method was also used to fabricate Sb₂Se₃–based solar cells [23,24]. Thanks to the low melting point of Sb₂Se₃ (608 °C) and the high saturated vapor pressure (around 100 Pa at 550 °C), CSS is a low-energy-consumption technique for Sb₂Se₃ growth, and it is more likely to form (hkℓ)-oriented grains with ℓ ≠ 0, especially at certain substrate temperatures. An optimal ribbon orientation allowed the CSS-grown Sb₂Se₃ solar cell to reach a high 9.2% efficiency [25].

2.3. Rapid Thermal Evaporation

RTE [6,26,27] is a vacuum process similar to CSS, but in this case, the material melts and evaporates from the liquid phase, not from the solid phase as in a sublimation process.

The deposition method is carried out in a tube furnace, generally maintained at a low vacuum by a mechanical pump (0.5–1 Pa) [6,25]. The distance between source and substrate is within 10 mm. In this process, a rapid increase in temperature and fast deposition are possible, so growth times are generally limited to a few seconds.

For example, H. Guo et al. [26] set a deposition time of 40 s, a preheating temperature of 310 °C for 20 min and the deposition temperature of 570 °C. A PCE of 7% was obtained for a simple Au/Sb₂Se₃/CdS/ITO/glass superstrate solar cell.

2.4. Vapor Transport Deposition

In the VTD method, both the substrate temperature and source-to-substrate distance can be adjusted. A carrier gas is saturated with a vapor from a subliming source (high-purity powder) and conducted towards the substrate at lower temperature, where the film growth occurs (Figure 2c). Sb₂Se₃ absorber-based solar cells fabricated through VTD exhibited an efficiency of 7.6% in a Au/Sb₂Se₃/CdS/ITO/glass solar cell structure [28].

2.5. Injected Vapor Deposition

IVD has been recently developed by Z. Duan et al. [29]. It is a promising technique for the fabrication of Sb₂Se₃ solar cells because it has been demonstrated to achieve one of the highest efficiencies (more than 10%) measured so far.

In an IVD system, the precursor of the Sb₂Se₃ powder is placed into a roller, which is heated at 490 °C. Argon gas is injected from one side of the roller and forms an Sb₂Se₃ vapor which reaches the heated substrate on the opposite side (Figure 2d). In IVD, the injection gas flow is the most crucial parameter affecting the quality of the material and thus the solar cell performance.

2.6. Pulsed Laser Deposition

In PLD, usually in a vacuum chamber, the plume is formed by a high-power pulsed laser beam that hits a target of the material that must be deposited (Figure 2e). The deposition occurs when the material condensates on the substrate, facing the target, which is maintained at an adjustable temperature. An Sb₂Se₃ absorber deposited at 500 °C by PLD reached an efficiency of 4.77% [30,31].

2.7. Pulsed Electron Deposition

PED is a physical deposition technique where a pulsed electron beam is accelerated by a high voltage to ablate a solid material target. The interaction between electrons and target material forms a plasma plume through the fast sublimation of the target which condensates on the surface of a substrate, placed at a few centimeters in front of it (Figure 2f). The nonequilibrium fast solid-vapor transition induces a transfer of material with the same stoichiometry as the target to the substrate. Sb₂Se₃ layers have been grown in an argon atmosphere of 0.3 Pa, setting a discharge voltage of 16 kV, a pulse rate of 9 Hz, with the substrate temperature maintained between 200 and 400 °C. The optimized solar cell based on an Sb₂Se₃ absorber grown by PED (with the substrate structure AZO/ZnO/CdS/Sb₂Se₃/Mo/glass) showed an efficiency of 2.1% [32,33].

Figure 2. Schematized setups of different PVD systems: (a) RF-MS. Reproduced with permission from [19]. © 2021 IOP Publishing. (b) CSS. Reproduced from [34]. © 2020 K. Shen et al. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. (c) VTD. Reproduced from [28]. © 2018, X. Wen et al., http://creativecommons.org/licenses/by/4.0/ (accessed on 15 September 2023). (d) IVD. Reproduced with permission from [29]. © 2022 John Wiley & Sons. (e) PLD. Reproduced with permission from [31]. © The Royal Society of Chemistry 2004. (f) PED. Reproduced with permission from [33]. © 2013 IOP Publishing. (g) Current density–voltage (J–V) curves of the IVD-, CSS-, and CoE (co-evaporated)-Sb₂Se₃ solar cells. Reproduced with permission from [29]. © 2022 John Wiley & Sons.

Figure 2. Schematized setups of different PVD systems: (a) RF-MS. Reproduced with permission from [19]. © 2021 IOP Publishing. (b) CSS. Reproduced from [34]. © 2020 K. Shen et al. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. (c) VTD. Reproduced from [28]. © 2018, X. Wen et al., http://creativecommons.org/licenses/by/4.0/ (accessed on 15 September 2023). (d) IVD. Reproduced with permission from [29]. © 2022 John Wiley & Sons. (e) PLD. Reproduced with permission from [31]. © The Royal Society of Chemistry 2004. (f) PED. Reproduced with permission from [33]. © 2013 IOP Publishing. (g) Current density–voltage (J–V) curves of the IVD-, CSS-, and CoE (co-evaporated)-Sb₂Se₃ solar cells. Reproduced with permission from [29]. © 2022 John Wiley & Sons.

3. The Main Challenges for Sb₂Se₃—Based Solar Cells (1D-Ribbon Alignment, HTL/ETL, Interfacial Engineering and Band Alignment, Doping)

3.1. One-Dimensional-Ribbon Alignment

As discussed in Section 1, Sb₂Se₃ is characterized by a one-dimensional crystal structure, a ribbonlike morphology, which determines strong anisotropic electrical conduction and optical properties (Figure 3). Crystal orientation control is crucial for the deposition of high-quality Sb₂Se₃ films. Sb₂Se₃ has a one-dimensional crystal structure, which gives it strong anisotropic electrical conduction and optical properties. Researchers are working hard to obtain vertically (or quasi-vertically) oriented Sb₂Se₃ ribbons, i.e., parallel to the growth direction. Several factors influence the crystal orientation of Sb₂Se₃, including the deposition technique, the substrate, the postdeposition treatments, and the use of pre-growth seeding methods. It is known that the best transport properties for PV applications can be achieved if the c axis of the crystal structure is normal to the surface (0,0,ℓ), while the worst conditions are obtained when the c axis lies on the surface (h,k,0). It is difficult to compare different results in the literature in terms of grain orientations and their correlation to photovoltaic performance. This is because the relative texture coefficient, TC (hkℓ), is rarely reported. TC (hkℓ) is the ratio between the measured (hkℓ) peak intensity and the intensity of the same peak for a reference, randomly oriented powder, weighted as a percentage of the summation of the same value for all chosen peaks [35]. On the other hand, not all the reflections are always considered. TC (hkℓ) is given by the following equation:

$$TC\left(\mathrm{h}\mathrm{k}\mathcal{l}\right)=\frac{\frac{I\left(\mathrm{h}\mathrm{k}\mathcal{l}\right)}{I_0\left(\mathrm{h}\mathrm{k}\mathcal{l}\right)}}{{\sum }_n\frac{I\left(\mathrm{h}\mathrm{k}\mathcal{l}\right)}{I_0\left(\mathrm{h}\mathrm{k}\mathcal{l}\right)}}\times 100 (\%)$$

where I₀ (hkℓ) is the relative intensity of the reflection with (hkℓ) Miller indices reported in the JCPDS card, and I(hkℓ) is the net intensity measured by the experimental XRD patterns after the background subtraction.

In

Table 1. Extrapolated Σ TC (ℓ ≠ 0) values and reported PCE and J_sc for different deposition methods and underlying layers.

Deposition Method	Substrate/ Underlayer	Σ TC (hkℓ) ℓ ≠ 0	Observations	PCE	Jsc (mA/cm2)	Reference
CSS	TiO2/FTO/glass	86%	With seeding layer	3.8%	21.1	[36]
CSS	TiO2/FTO/glass	65.4%	W/o seeding layer	1.6%	24.9	[36]
CSS	CdS/FTO/glass	89.5%	CdCl2 treatment	6.23%	27.36	[36]
CSS	CdS/FTO/glass	88.1%	//	4.27%	21.74	[21]
CSS	CdS/FTO/glass	97.5%	CdS by CCS	2.82%	19.5	[37]
CSS	TiO2/FTO/glass	~100%	Seeding layer	5.28%	24.6	[37]
CSS	TiO2/CdS/FTO/glass	~100%	//	2.07%	15.6	[37]
CSS	SnO2/FTO/glass	58.3%	CdCl2 treatment	//	//	[38]
CSS	SnO2/FTO/glass	5.4%	No treatments	//	//	[38]
CSS	FTO/glass	92.2%	330 °C	9%	29.8	[39]
VTD	SnO2/ITO/glass	91.4%	//	2.7%	19.3	[40]
VTD	TiO2/ITO/glass	30.9%	//	2%	23.6	[40]
VTD	SnO2/TiO2/ITO/glass	~100%	//	4.8%	25.6	[40]
VTD	CdS/TiO2/ITO/glass	64.7%	CdS-VTD	4.91%	28.6	[41]
VTD	CdS/TiO2/ITO/glass	50.1%	CdS-CBD	4.24%	28.3	[41]
RTE	CdS/FTO/glass	90.4%	w/o KOH treatment	4.8%	28.13	[26]
RTE	CdS/FTO/glass	92.1%	with KOH treatment	7.16%	//	[26]
RF sputtering	CdS/FTO/glass	71.3%	Annealing at 460 °C	6.06%	25.91	[42]
RF sputtering	FTO/glass	60%	//	1.28%	24.83	[19]
RF sputtering	CdS/FTO/glass	91%	//	2.36%	27.06	[19]
RF sputtering	Mo	30%	//	0.24%	5.11	[19]
PED	FTO/glass	60.2%	//	2.1%	20.3	[32]
PED	CdS/FTO/glass	38.4%	//	//	//	[32]
PED	Mo	0.6%	//	//	0.3	[32]
Thermal co-evaporation	Mo	73.5%	Substrate T = 300 °C	2.67–4.51%	19.51	[43]

, we list several works where relative TC (hkℓ) (for the summation of reflections h = 0, k = 0, and ℓ ≠ 0) for Sb₂Se₃ films have been extrapolated from the reported graphs and at least around 10 reflections have been shown. The Σ TC (ℓ ≠ 0) parameter represents the percentage of vertically oriented ribbons. Vertically oriented ribbons do not lie on the substrate surface, and thus, they contribute to the electrical conduction.

In general, the CSS technique and the more recent IVD allow for more vertically oriented ribbons because the material is directly sublimated from the solid, and deposition occurs at high growth rates. In the paper by Krautmann et al., Sb₂Se₃ is deposited on TiO₂ with a seeding layer procedure: a 60 nm seed layer (SL) at 300 °C and then 1.8 micron at 450 °C [36]; the SL growth at low temperature is fundamental to create ribbonlike seeds on which subsequent columnar structures can grow. Vertical ribbons (002) planes show a TC of 1.2 for Sb₂Se₃ with an SL and 0.2 without an SL. The extrapolated Σ TC (hkℓ) ℓ ≠ 0 was 86% with an SL and 65.4% without an SL, respectively. As a consequence, solar cells with a higher TC show an efficiency of 3.8% and solar cells without an SL an efficiency of 1.6%, and an increase in the short circuit current density, J_sc, from 21.1 to 24.9 mA/cm² when the SL is used. In [44], the authors showed that the first grown 750 nm Sb₂Se₃ layer on CdS/FTO was compact, and ribbonlike structures started to grow at a higher growth time on the compact layer; (211), (221), (301), and (311) orientations prevailed when the entire film thickness was between 1.3 and 2.7 µm. A maximum PCE of 6.23% and a J_sc of 28.53 mA/cm² was obtained for a 2.1 µm thick Sb₂Se₃ with a Σ TC (hkℓ) ℓ ≠ 0 of 89.5%. The use of CuInSe₂ sensitizer quantum dots (QDs) to passivate defects of the Sb₂Se₃ nanorod structure allowed a 7.6% PCE.

In [34], the Sb₂Se₃ SL deposited by a CSS process on CdS was 10 nm thick and subsequently an Sb₂Se₃ absorber layer of 500 nm was deposited on the SL. In that case, the SL was grown by setting the source and substrate temperatures at 310 °C for 120 s, raising the source to 510 °C in 60 s (setting the substrate temperature at around 330 °C) and then keeping the source temperature at 510 °C for 10 s. The main absorber layer deposition was performed after a natural cooling to 270° of the substrate. In that case, the film grown after the seed layer showed a more (221) preferred orientation (Figure 4a), a PCE of 7.45%, and a J_sc of almost 30 mA/cm² was achieved for Au/t-Se/Sb₂Se₃/CdS/FTO cells.

L. Guo et al. [23] emphasized the importance of growth temperature and time during CSS growth. In fact, they observed that (211) oriented grains dominated in films thicker than 0.6 mm, and (hk0) orientations were strongly limited at substrate temperatures lower than 300 °C. Σ TC (hkℓ) ℓ ≠ 0 was 88.1% and graphite/Sb₂Se₃/CdS/FTO/glass samples reached a J_sc of 21.74 mA/cm² and a 4.27% PCE, in that case.

Z. Li et al. [25] studied the evolution from a thin-layer film to a nanorod array for Sb₂Se₃ on Mo coated glass substrates deposited by CSS (Figure 5). As in the previous case, deposition time is important for 1D growth: they observed that the Sb₂Se₃ structure varied from a compact thin film to an aligned nanorod array with increasing growth durations (from 60 to 180 s). The nanorod alignment promoted a J_sc of 32.58 mA/cm² leading to a PCE of 9.2%.

C. Guo et al. [45] studied the incorporation of a Ag element into Sb₂Se₃ obtaining quasi-vertical 1D ribbons at 280 °C, which vanished at 420 °C, where the (120) preferential orientation prevailed. Solar cells of AZO/ZnO/(Sb₂Se₃)_x(AgSbSe₂)_1−x/MoSe₂/Mo/glass achieved a J_sc of 25.29 mA/cm² and a 7.8% of PCE when the substrate temperature was set at 320 °C.

In [39], another CSS growth was adopted to obtain Sb₂Se₃, and the authors found that the formation of the Sb₂Se₃ crystal was not strongly dependent on the substrate temperature. However, (hk0) planes, parallel to the substrate, decreased for growth temperature >240 °C (up to 410 °C), while most of the TCs for the (hk1) planes, oblique or perpendicular to the substrate, increased with the growth temperature between 240 °C and 410 °C. Σ TC(hkℓ) ℓ ≠ 0 was 92.2% in this case and AZO/i-ZnO/CdS/Sb₂Se₃/FTO/glass solar cells with a 1350 nm (Figure 4b,c) thick Sb₂Se₃ layer demonstrated a J_sc of 29.5 mA/cm² and a conversion efficiency of 9.0%.

P. Fan et al. [46] investigated Sb₂Se₃ obtained by CSS followed by a selenization process on a Mo/SLG (soda–lime glass) substrate. They studied different temperature ramp for the source during sublimation: from 260 or 290 °C for 2 min to 510 or 520 °C for 2 min. In the selenization process, the Se powder and the as-grown Sb₂Se₃ thin film, placed at a distance of 23 cm, were simultaneously heated at 400 and 420 °C for 15 min, respectively. It was found that the (hk0) grain orientation decreased by increasing the temperature from 260 to 290 °C, while the TC values of the (hk1) and (hk2) increased with the temperature, finding an optimum substrate temperature to stimulate the (hk1) orientation for 290 °C. Moreover, the subsequent post-selenization treatment led to large grain sizes with the preferred (hk1) orientation. In these optimized conditions, the authors obtained a high 509 meV V_oc, with a J_sc of 21.15 mA/cm² reaching a PCE of 4.86% with a Ag/ITO/CdS/Sb₂Se₃/Mo solar cell.

N. Spalatu et al. [37] grew Sb₂Se₃ absorbers by CSS, finding a preferred orientation of grains along the (211) and (221) crystal directions for depositions at 450 °C. For temperatures rising from 300 to 450 °C, the TC values of the (301) and (151) planes increased systematically, and the TC of the (020) and (120) planes decreased. A maximum efficiency of 2.8% was demonstrated for Au/Sb₂Se₃/CdS/FTO/glass solar cells with an Sb₂Se₃ absorber obtained at 450 °C, with a Σ TC (hkℓ) ℓ ≠ 0 of 97.5%. Better performance of PCE around 5.3% was achieved with a superstrate configuration TiO₂/Sb₂Se₃ cell where an Sb₂Se₃ seed layer was implemented at the interface and where Σ TC (hkℓ) ℓ ≠ 0 was around 100%. The SL consisted of a ~60 nm thick Sb₂Se₃ layer deposited at a low T of 300 °C by CSS at the interface between buffer and absorber layers, and it was fundamental in that work to enhance the efficiency from 1.66% (J_sc of 19.8 mA/cm²) to 5.28% (J_sc of 24.6 mA/cm²).

A. Amin et al. [47] fabricated Sb₂Se₃ by CSS on a nanorods (NR)-CdS/nanoparticles (NP)-CdS buffer layer. The nanorod CdS (NR-CdS) buffer layer was synthesized by a hydrothermal method while the NP CdS films were prepared by the CBD method. The authors observed that the NR-CdS and NR-CdS/NP-CdS buffer layers favored the (211) orientation of the Sb₂Se₃ ribbons and the reduction in the (120) orientation. Thanks to the control of the nanoribbon orientation, the Sb₂Se₃ solar cell reached a PCE of 7.16%.

S. Pasini et al. [48] showed the Influence of the window layer on the Sb₂Se₃ ribbon orientation for a superstrate configuration solar cell. A preferential growth of the CSS-grown Sb₂Se₃ along the (001) direction on window layers containing zinc, with a TC for the (002) reflection reaching the maximum value of 27% and a 4.5% efficiency with a ZnCdS (Zn_0.15Cd_0.85S) window layer (with a solar cell structure of Pt/Sb₂Te₂/Sb₂Se₃/ZnCdS/ZnO/ITO/glass) was observed.

In the case of the VTD process, X. Wen et al. [28] obtained Sb₂Se₃ grown on CdS/ITO, studying an evaporation temperature between 480 and 510 °C, reached with a ramp rate of 20 °C/min followed by a plateau of 2 min. When the source temperature was at 510 °C, corresponding to a substrate temperature of 390 °C, the film exhibited the highest crystallinity and XRD intensity of the preferred (221) orientation. Thus, the Au/Sb₂Se₃/CdS/ITO/glass solar cells showed a J_sc of 29.9 mA/cm² leading to a 7.6% PCE.

J. Kim et al. studied Sb₂Se₃ deposition by VTD [49], by co-evaporation from Sb₂Se₃ and Se sources. The chamber pressure was 3 mTorr, with a source temperature of 560 °C and a substrate temperature of 300 °C. The authors observed that Se flux control was very important for the film characteristics. In fact, an extra supply of Se contributed to suppress undesirable vacancies V_Se, while an excessive Se overpressure inhibited the (221) orientation growth, which was detrimental to electrical transport properties. A superstrate configured solar cell (Au/Sb₂Se₃/CdS/ITO/glass) showed a PCE of 5.2% in optimized conditions. S. Dong et al. [50] proposed a modified VTD method, in a double-temperature-zone furnace, to obtain Sb₂Se₃ with Se in excess. A CdS/ITO/glass substrate was placed on the graphite bracket at the extremity of the furnace, the Se powder was kept at a higher temperature than the melting point (>270 °C) to ensure a stable supply of Se was placed at the other end of the tubular furnace above its thermal baffle, and two quartz crucibles containing Sb₂Se₃ powder were put in the center of the two temperature zones. The authors showed that (221) orientations increased with the Se flux; thus, with the Se excess, and a Au/Sb₂Se₃/CdS/ITO/glass structure, the PCE was improved from 5.1% to 6.7%, and J_sc from 27.7 to 30.5 mA/cm² with the absorber grown in Se-rich conditions.

W. Wang et al. [40] implemented VTD-grown Sb₂Se₃ on SnO₂, TiO₂, SnO₂/TiO₂, CdS, and SnO₂/TiO₂/CdS ETL. They observed that SnO₂ or/and SnO₂/TiO₂ ETL underlayers favored the formation of (041), (141), and (061) (Σ TC (hkℓ) ℓ ≠ 0 of 91.4% and almost 100%, respectively) more than in the case of TiO₂ (Σ TC (hkℓ) ℓ ≠ 0 of 30.9%). Since the CdS interlayer was also important to reduce the band offset, the Sb₂Se₃/CdS/TiO₂/SnO₂/ITO/glass solar cell achieved a J_sc of 29.9 mA/cm² and PCE of 7%.

In [41], the authors fabricated a CdS layer by VTD with a better conductivity and higher crystallinity than with CBD. They observed that the CdS-buffer layer induced the growth of Sb₂Se₃ with an (hk1) orientation, such that the extrapolated Σ TC (ℓ ≠ 0) was around 64.7% for Sb₂Se₃ grown on a CdS layer deposited by VTD and 50.1% for the one grown on CdS obtained by CBD. Thus, they obtained a 4.91% PCE for the CdS-VTD-based device, and 4.24% for the CdS-CBD.

In [51], a CdS ETL was also modified with Al₂O3 grown by atomic layer deposition as the interface layer at the CdS/Sb₂Se₃ interface. The Sb₂Se₃ absorber layer was deposited on Al₂O3 by VTD in a dual-temperature-zone tube furnace, showing a strong increase in TCs of (hk1) and decrease in (hk0) in the presence of Al₂O₃. Therefore, the authors obtained a PCE of 6.25% for a carbon/Sb₂Se₃/Al₂O₃/CdS/FTO/glass solar cell configuration.

V. Kumar et al. [52] compared thermally evaporated Sb₂Se₃ subjected to three different processes: just annealing, simultaneous evaporation of Sb₂Se₃/Se during the absorber growth, and post-deposition selenization. They observed that the Sb₂Se₃ layer grew spontaneously with predominant (221) and (321) orientations. Then, post-deposition annealing and co-selenization reduced (321) in favor of the (141) orientation, while in the case of the postselenization, the (221) orientation had a stronger intensity, and the (041) orientation was also present.

Surface treatments can also induce columnar-mode growth as observed from H. Guo et al. [53]. In that case, the RTE method was used to fabricate Sb₂Se₃ in a Au/Al₂O₃/P⁺-Sb₂Se₃/P- Sb₂Se₃/CdS/FTO/glass structure in which the CdS layer was submitted to an oxygen plasma treatment before the Sb₂Se₃ growth. The effect was a decrease in the CdS grain size, a reduction in film roughness, and the introduction of oxygen atoms. The authors argued that the increase in the O content in the CdS surface stimulated a higher crystallinity and the columnar growth of Sb₂Se₃, thus an increase in J_sc from 22.03 mA/cm² to 26.26 mA/cm² and PCE from 2.48 to 5.1% was observed for solar cells without or with plasma treatment, respectively. Moreover, a subsequent Al₂O₃ layer, deposited by atomic layer deposition (ALD) method, increased by two orders of magnitude the hole concentration for Sb₂Se₃ film, leading to a PCE of 6.7%.

In [54], an Sb₂Se₃ nanorod-array was prepared on the Mo substrate at a low temperature by RTE. The authors showed the importance of the SL growth temperature; the SL was performed at 360 °C while 480 °C was the Sb₂Se₃ growth temperature, resulting in a preferred orientation in the (211) direction and a minimum TC value for the (230) diffraction peak. In this manner, they showed an efficiency of 6.3% in a ZnO/CdS/Sb₂Se₃/Mo/SLG solar cell structure.

X. Wen et al. [55] fabricated a highly (001)-oriented Sb₂Se₃ film on a flexible mica where Sb was thermally evaporated, followed by a selenization process into a horizontal quartz tube furnace at 410 °C. In this way, Ag/ITO/CdS/Sb₂Se₃/Mo/mica grid flexible solar cells obtained an efficiency of 8.42%, a J_sc of 31.30 mA/cm2, and a V_oc of 0.47 V.

A KOH treatment also induces preferential grain orientation, as observed in [26], where the TC ratio between the (221) and (211) peaks was higher in the Sb₂Se₃ film exposed to 0.07 M KOH than that in the not-treated films, showing a Σ TC (hkℓ) ℓ ≠ 0 higher than 90%. This concentration was found to be optimal, leading to a 7.16% efficiency and a J_sc of 30.68 mA/cm² in a Au/Sb₂Se₃/CdS/FTO/glass solar cell. At the same time, the KOH solution (0.07 M) treatment played a doping role, through a diffusion of K species; in fact, the hole concentration of the Sb₂Se₃ film increased from 10¹³ cm⁻³ to 10¹⁵ cm⁻³; for higher concentrations (0.15 M), the carrier density decreased, probably due to corrosion.

J. Zhou et al. showed how the SnO₂ underlayer deposition and treatment could influence the Sb₂Se₃ orientation [38]. In that case, Sb₂Se₃ grown by CSS on treated SnO₂, with a CdCl₂ methanol solution, showed a high preferential (211) 1D orientation, Σ TC (hkℓ) ℓ ≠ 0 was 58.3% versus 5.4% for the not-treated surface. The authors claimed that this effect was due to the formation of Cd-[Sb₄S₆]_n ribbon covalent bonds.

Now, considering the sputtering method, G-X. Liang et al. [17] obtained Sb₂Se₃ nanorods with the preferred crystallographic orientation of (221) at a substrate temperature of 375 °C, and a deposition pressure and power of 0.5 Pa and 30 W, respectively. The same conditions at room temperature led to amorphous Sb₂Se₃ films. AZO/ZnO/CdS/Sb₂Se₃ nanorods/Mo solar cells exhibited a 2.1% PCE.

R. Tang et al. [42] reported on a sputtered Sb₂Se₃ cell obtained by a two-step process (growth of amorphous Sb₂Se₃ by RF-MS followed by a selenization process). The selenization step was performed in an evacuated tubular furnace successively filled with pure Ar at a pressure of 5 × 10⁴ Pa. The highly pure Se powder was heated at 400 °C while Sb₂Se₃ was kept at a lower temperature. Temperature is critical during selenization so that the growth of vertical rods is favored when selenization is performed up to 420 °C, then (hk0)-oriented grains start to increase. For the best T, Σ TC (hkℓ) ℓ ≠ 0 was 71.3%, J_sc was 25.91 mA/cm², and the efficiency of a Ag/ITO/CdS/Sb₂Se₃/Mo reached 6.06%. For S. Rijal et al. [56], in Sb₂Se₃ grown by CSS on Mo-coated glass, higher temperatures in selenization enhanced crystallinity, as the average grain size increased with the selenization temperature. The highest TC values of the (211) and (221) planes for the film annealed at 425 °C indicated a preferred Sb₂Se₃ ribbon orientation at this selenization temperature. As a consequence, J_sc increased from 18.6 mA/cm² to 22.9 mA/cm², and the PCE from 1.85% to 6.43%.

C. Yuan et al. obtained [57] an Sb₂Se₃ film by a two-step method: DC-MS of metallic Sb followed by selenization through a rapid thermal process (RTE). The selenization process was conducted with temperatures between 320 and 440 °C for 5 min, under vacuum conditions. The thin films selenized with temperatures below 400 °C exhibited a preferred orientation along the (211) plane and (221) plane, while the morphology quality decreased for higher annealing temperatures with the appearance of holes and a lower carrier mobility. The best achieved efficiency for an Al/Al:ZnO/ZnO/CdS/Sb₂Se₃/Mo solar cell was 3.47%, and J_sc = 16 mA/cm² for selenization performed at 400 °C.

G. Spaggiari et al. [19] observed that the crystal structure and the (Sb₄Se₆)_n ribbon orientation of Sb₂Se₃ obtained by RF sputtering were strongly dependent on the deposition temperature and type of substrate. The authors tested different types of substrate/underlayers (glass, Mo, FTO, CdS, and ZnO) observing a thermodynamical grain alignment along the (001) direction, vertical with respect to the substrate, for depositions at 300 °C on thick samples grown on CdS (Σ TC (hkℓ) (ℓ ≠ 0) = 91%) and ZnO (Σ TC (hkℓ) (ℓ ≠ 0) = 80%). An efficiency of 2.36% was reached, with a J_sc of 27.06 mA/cm² for a Au/Sb₂Se₃/CdS/FTO/glass superstrate configuration solar cell while the PCE was very poor for the absorber grown on Mo substrates.

I. Caño et al. [58] obtained Sb₂Se₃ layers on Mo-sputtered SLG substrates by a selenization of thermally evaporated Sb. The (002) (vertical ribbons) reflection was predominant for Se-rich conditions, while in Sb-rich conditions, the (231), (141), and (221) peaks became dominant. In detail, with 0.36 Se, a slightly poor Se compound, (002) reflections practically disappeared, while Se-rich conditions could also reduce the V_Se-type defects and thus improve the device performance.

Y. Yang et al. [59] fabricated a Au/Sb₂Se₃/CdSe/FTO/glass solar cell by obtaining CdSe and Sb₂Se₃ both by PLD, with the CdSe/Sb₂Se₃ bilayer deposited continuously in the same chamber. The authors studied the CdSe temperature influence on Sb₂Se₃ quality, maintaining a deposition temperature of Sb₂Se₃ at 500 °C and varying the CdSe temperature: when it is increased to 300 °C, the TCs of the (221) and (221) orientations increased, while the (120)- and (230)-orientation TCs decreased, showing a 1D preferential growth for these conditions. Au/Sb₂Se₃/CdSe/FTO/glass solar cells exhibited a J_sc of 20.18 mA/cm² and a PCE of 3.65% for a CdSe deposition temperature of 300 °C.

PLD growth of Sb₂Se₃ films deposited on CdS/FTO/glass was performed by K. Yang et al. [30] at a substrate temperature of 500 °C, a repetition rate of 6 Hz, and a pulse energy of 90 mJ. The samples showed a preferred direction of (221) and a J_sc of 31.68 mA/cm², and an efficiency of 4.77% was achieved.

In the only paper (to date) reporting on PED deposition, the role of the substrate was crucial for the 1D ribbon vertical alignment; in fact, F. Pattini et al. [32] showed that the best ribbon alignment for Sb₂Se₃ grown on FTO/glass (Σ TC (hkℓ) (ℓ ≠ 0) = 60.2%), while for CdS, this value decreased to 38.4% and, differently, ribbons were completely laid on the substrate surface when Sb₂Se₃ was grown on Mo (Σ TC (hkℓ) (ℓ = 0) = 99.4%). As a consequence, a very low J_sc was obtained for Sb₂Se₃ deposited on Mo, while AZO/ZnO/CdS/Sb₂Se₃/FTO/glass solar cells exhibited a J_sc of 20.28 mA/cm² and a PCE of 2.1% under frontal illumination and an equivalent 3.1% conversion efficiency for a bifacial illumination.

S-N. Park et al. [43] applied a co-evaporation method with Sb₂Se₃ and Se sources to obtain an Sb₂Se₃ absorber for solar cells and observed that the fine control of process temperature was extremely crucial; in fact, (hk1)-oriented Sb₂Se₃ ribbons were favored when the substrate was heated in the range between 270 °C and 315 °C; in particular, rod arrays preferentially oriented along (hk1) were obtained for 315 °C, while 330 °C stimulated the growth of hk0 planes. In these conditions, an extrapolated Σ TC (hkℓ) ℓ ≠ 0 was 73.5%. The authors proposed a nucleation and growth model for explaining the influence of the substrate temperature on the thin film formation, in particular, the evolution of Sb₂Se₃ from flat to rods with the increasing substrate temperature, based on a Terrace–Ledge–Kink model. The (001), (211), and (221) surfaces presented higher surface energies than (100), (010), (110), and (120), which had no dangling bonds and tended to form terraces. On the other hand, (hk1) surfaces had large numbers of dangling bonds and presented many kink and ledge sites. Increasing temperatures and consequently the diffusion length stimulated the adatom to migrate to these sites. As the propagation direction was expected to be (001) because of the large number of dangling bonds, a growth of rods along the vertical direction was predicted. The rod array obtained at 315 °C implied a highest PCE of 4.5% and a J_sc of 25.39 mA/cm² for Au/AZO/ZnO/CdS/Sb₂Se₃/Mo solar cells.

A summary of the reviewed deposition techniques, the TC values, different solar cell architectures, and the correlated performances obtained is reported in Table 1. In Figure 6, the extrapolated Σ TC (hkℓ) (ℓ ≠ 0) values are plotted as a function of PCE for different solar cells. Even though the points are slightly scattered, a clear trend between TC and PCE can be observed. Even if the PCE value also depends on other electrical parameters, on the different HTL and ETL materials, and even on the peculiar deposition method, one can argue that high-efficiency cells must have a high degree of vertical ribbon orientation.

3.2. HTL/ETL, Interfacial Engineering, and Band Alignment

V_oc in Sb₂Se₃ solar cells is strongly affected by the interface and SRH recombination. In order to recover V_oc losses, the optimization of the band alignment is fundamental to suppress the interface and defect recombination. In semiconductors, offsets between conduction bands and valence bands at the interfaces of the heterostructures are fundamental for the good working of solar cells. A minimization of the band offset should be achieved, and a “spike-like” configuration is more adequate, where the offset is positive, hence the conduction band minimum of the absorber is lower than the conduction band minimum of the buffer/window layer [60]. The choice of ETLs and HTLs controls the energy band alignments in Sb₂Se₃ solar cells. Different ETLs and HTLs have been explored in the literature to find the optimal solar cell configuration.

3.2.1. ETLs

Among ETLs, CdS is largely used as it is commonly employed as a buffer layer in CIGS, CdTe, and CZTSSe solar cells and it is easily obtained by CBD. The conduction band offset (CBO) at the CdS/Sb₂Se₃ interface [61] is a type-II band alignment, the corresponding calculated CBO and valence band offset (VBO) at the CdS/Sb₂Se₃ interface are 0.24 and 0.94 eV, respectively (Figure 7). However, Cd-free ETLs are objects of study to avoid the toxicity of Cd and other drawbacks associated with CdS, such as a high lattice mismatch (>10%) with Sb₂Se₃, undesired absorption, the need for surface treatments, and the Cd interdiffusion at the Sb₂Se₃/CdS interface which causes further defects [62].

On the other hand, SnO₂, TiO₂, and ZnO, have similar lattice parameters with Sb₂Se₃. These kind of ETLs have been predominantly explored in superstrate configurations. The ZnO/Sb₂Se₃ interface presents a cliff-type CBO of 0.11e V and a VBO of 2.21 eV, while TiO₂/Sb₂Se₃ present a flat interface for CBO and a VBO of 2.33 eV [61]. S. Lu et al. conclude that a promising strategy would be to combine CdS and TiO₂ with Sb₂Se₃ in order to obtain a favorable spikelike band alignment.

Alternative ETLs, investigated in theoretical studies are WS₂ [63], CdZnS [64], CdS-3N, ZnMgO [65], and ZnSe [66].

CdS as an ETL has been investigated by several groups in Sb₂Se₃ solar cells. More recently, modified CdS layers were studied for interface engineering, to modify the CdS/Sb₂Se₃ heterojunction from a “cliff-like” structure to a “spike-like” structure. Y. Luo et al. [67] obtained an Al³⁺ cation doping in the CdS buffer layer (ETL) through spin-coating and postannealing methods. The Al doping allowed the optimization of the band alignment with an ideal CBO, modifying it from a “cliff-like” structure to a “spike-like” one, thus reducing the charge recombination. Sb₂Se₃ was obtained by Sb RF-MS deposition and subsequent selenization while the Al-doped CdS film was obtained via direct current (DC) MS. Solar cells with a Ag/ITO/CdS:Al/Sb₂Se₃/Mo structure achieved a PCE of 8.41%, J_sc = 28.26 mA cm⁻², V_oc = 489 mV, and FF = 60.87%.

In another work, a CdS ETL was also treated by using nontoxic SbCl₃ to form an Sb-CdS/Sb₂Se₃ interface [68], which also produced a “spike”like-type heterojunction. A Au/spiro-OMeTAD/Sb₂Se₃/CdS/FTO/glass superstrate structure delivered a PCE of 6.13%, while with a normal CdS/Sb₂Se₃ interface, the PCE was 3.95%. The authors observed that this SbCl₃ treatment induced the formation of pure hexagonal CdS which stimulated the growth of an Sb₂Se₃ film with a (hk1) preferential orientation.

J. Cheng et al. fabricated a Cd(S,O) ETL [69] through a sol–gel method; the oxygen content was controlled using an anhydrous spin solution. The possibility of changing the Cd/S content ratio permitted the optimization of the CdS layer properties. A Au/Sb₂Se₃/CdS(O)/ITO solar cell structure reached an improved efficiency of 5.76% for a Cs:S ratio of 1:0.85.

An oxygenated CdS (CdS:O) ETL was investigated by Y. Liu et al. [70]. CdS:O was obtained by reactive sputtering with a CdS target and an Ar/O₂ mixed working gas, while the Sb₂Se₃-absorbed layer by CSS with a thickness between 1200 nm and 1500 nm. An AZO/ZnO/CdS:O/CdS/Sb₂Se₃/MoSe₂/Mo/glass solar cell with an O₂ content of 0.5% in the sputtering environment achieved V_oc = 408 mV, J_sc = 28.11 mA cm⁻², FF = 66.98%, and PCE = 7.69%.

L. Guo et al. [71] also studied CdS:O obtained by RF sputtering as an ETL for Sb₂Se₃-based solar cells grown by CSS. Sputtered CdS:O exhibited a better quality than CdS deposited by CBD; the preferred grain orientation was (211) in CdS:O/Sb₂Se₃, while it was (221) in CBD CdS/Sb₂Se₃. Oxygen also played a role in blocking the interdiffusion of Cd into the Sb₂Se₃ film. Thus, the PCE was increased from 6.1 to 7% from a CBD CdS to a sputtered CdS:O in the Ag/graphite/Sb₂Se₃/CdS:O/FTO/glass solar cell configuration.

X. Mao et al. [72] employed a CdS/SnO₂ ETL in a Au/Spiro-OMeTAD/Sb₂(S,Se)₃/CdS/SnO₂/FTO/glass solar cell, achieving a PCE of 8%. In their work, they highlighted the importance of synthesis and specifically the ratio between SnO₂ hydrosol and deionized water during a spin-coating method on the performance of the cells.

G. Li et al. [73] studied a CdZnS ETL, for a ZnO:Al/ZnO/Cd_xZn_1−xS/Sb₂Se₃/Mo/glass solar cell with a Cd_0.75Zn_0.25S buffer layer. They obtained an efficiency of 6.71%, compared to the CdS/Sb₂Se₃-based cell that achieved a PCE of 5.08%. The optimum concentration allowed the formation of a spikelike interface with a CBO of 0.34 eV, which reduced the photocarrier recombination.

SnO₂ has been studied as a valid alternative to CdS thanks to the interesting band alignment with Sb₂Se₃ (spikelike heterojunction) and high electron mobility. J. Zhou et al. [74] studied SnO₂, obtained by a spin-coating method, treated with CdCl₂ as an ETL for Sb₂Se₃ cells. They observed that the CdCl₂ treatment was fundamental to grow vertically oriented Sb₂Se₃ ribbons; in fact, efficiency improved strongly from 0.12 to 4.76% when SnO₂ was treated in a C/P3HT/Sb₂Se₃/SnO₂/FTO/glass solar cell.

W. Wang et al. [40] explored a double interface SnO₂/TiO₂/CdS ETL for VTD-grown Sb₂Se₃ in a superstrate configuration, obtaining a 7% PCE, as already mentioned.

S. Wen et al. [75] also showed the importance of the CdCl₂ treatment on a SnO₂ ETL in a Au/Sb₂Se₃/SnO₂/ITO/glass solar cell measuring efficiency increases from 2.02 to 4.03% when the treatment was performed.

Z. Chen et al. [76] studied La-doped SnO₂ as an ETL, obtaining an efficiency of 3.25% for Au/Sb₂Se₃/SnO₂/FTO/glass solar cells. The authors observed that La doping improved the crystallinity and electrical conductivity of SnO₂ films.

CdCl₂ treatment was also investigated by Y. Wang et al. [77], this time on TiO₂. The importance of this treatment was clear since the Sb₂Se₃ device’s PCE improved from 2.02 to 6.06% with the Au/Spiro-OMeTAD/Sb₂Se₃/TiO2/FTO/glass device structure when CdCl₂ was used on TiO₂. As a consequence of the treatment, the CBO of the Sb₂Se₃/TiO₂ heterojunction reduced from 0.50 to 0.20 eV producing a narrower clifflike CBO of the Sb₂Se₃/TiO₂ heterojunction, which favored a passivated interface recombination and facilitated the electron transportation from Sb₂Se₃ to TiO₂.

X. Wang et al. [78] fabricated Au/Sb₂Se₃/CeO₂/CdS/FTO/glass solar cells, exploring ultrathin cerium dioxide (CeO₂) obtained by evaporation, together with CdS as ETLs. A CeO₂ layer of 2 nm stimulated the (221) orientation in Sb₂Se₃ ribbons and the solar cell exhibited a 5.14% PCE. The authors attributed the increase in PCE to an improvement in crystallinity and a reduction in photocarrier recombination at the interface between Sb₂Se₃ and CdS, due to the larger (3.2 eV) energy gap of CeO₂.

J. Zhang et al. [79] used an n-type MoO₃ layer as the ETL obtaining a 6.3% conversion efficiency in a Au/MoO₃/Sb₂Se₃/CdS/FTO/glass superstrate solar cells, optimizing MoO₃’s thickness and substrate temperature during its deposition (Figure 8).

Alternative Cd-free solar cells have been a strategic goal in order to achieve an environmentally sustainable PV technology. W. Wang et al. obtained TiO₂/SnO₂ double ETLs, treated with SbCl₃, demonstrating a PCE of 5.82% in a Au/Sb₂Se₃/TiO₂/SnO₂/ITO/glass solar cell [80].

Zn_1-xSn_xO layers [81] were obtained via magnetron cosputtering as an ETL. The authors obtained a PCE of 3.44% for a Ag/ITO/ZTO/Sb₂Se₃/Mo/glass substrate solar cell, exploiting a Zn_0.57Sn_0.43O buffer layer.

Y. Dong et al. [82] investigated a ZnCl₂-modified In₂S₃ ETL also for a Cd-free solar cell in a Au/Spiro-OMeTAD/Sb₂Se₃/In₂S₃/FTO solar cell. A ZnCl₂ post-treatment favored a quality improvement of the In₂S₃ film, prepared by CBD, also inducing a longitudinal growth on the Sb₂Se₃ layer, fabricated via VTE. Thus, on InCl₂–In₂S₃, (Sb₄Se₆)_n ribbons exhibited a perpendicular orientation to the substrate surface. The solar cell PCE increased from 2.63% to 5.00% due to the ZnCl₂ treatment.

ZnSe as an alternative ETL was studied theoretically by R. Kumari et al. with SCAPS-1D and the Al/ZnSe/Sb₂Se₃/Mo solar cell efficiency was 24% with optimized parameters [66].

Moreover, an organic ETL was recently studied for Cd-free Sb₂Se₃-based solar cells. For example, C60 was used in [83] for an Al/Alq3(3.0 nm)/C60 (5.0 nm)/Sb₂Se₃ (∼50 nm)/NPB (6.0 nm)/ITO/glass solar cell, in which a PCE of 5.03% was demonstrated. In that paper, the NPB (N,N′-Di(1-naphthyl)-N,N′-diphenyl-(1,1′-biphenyl)-4,4′-diamine) layer was successfully used to facilitate the Sb₂Se₃ crystallization, ribbon orientation, and photocarrier recombination suppression.

D. Rovira et al. [84] presented a polymeric ETL in a CdS-free solar cell, with an AZO/b-PEI/ZnO/Sb₂Se₃/Mo/glass structure. Sb₂Se₃ was obtained by the evaporation of 280nm-thick Sb film on the deposited Mo and a subsequent annealing in a furnace under a selenium atmosphere. The polymeric layer obtained by spin-coating formed interfacial dipoles which contributed to enhance the V_oc from 243 mV to 344 mV, and a PCE of 2.41% was obtained.

In

Table 2. Different materials used as ETL/HTL for Sb₂Se₃ solar cells and their performance.

ETL	HTL	PCE	Absorber Deposition Method	Reference
CdS:Al	/	8.41%	RF-MS + selenization	[67]
Sb-CdS	SpirOmetad	6.13%	RTE	[68]
Cd(S,O)	/	5.76%	Thermal evaporation	[69]
CdS:O	MoSe2	7.69%	CSS	[70]
CdS/SnO2	SpirOmetad	8%	Hydrothermal	[72]
CdZnS	/	6.71%	CSS	[73]
CdS	/	5.08%	CSS	[73]
SnO2	P3HT	4.76%	CSS	[74]
CdS/TiO2/SnO2	/	7%	VTD	[40]
CdS/TiO2	/	5.82%	VTD	[80]
SnO2	/	4.03%	VTD	[75]
La-SnO2	/	3.25%	RTE	[76]
TiO2 (CdCl2 treatment)	SpirOmetad	6.06%	Thermal evaporation	[77]
MoO3	CdS	6.3%	RTE	[79] Figure 8
ZTO	/	3.44%	Magnetron cosputtering	[81]
InCl2–In2S3	SpirOmetad	5.0%	VTE	[82]
C60	NPB	5.03%	VTE	[83]
CdS	CuSCN	7.5%	VTD	[85]
CdS	WO3−x/	7.1%	CSS	[86]
CdS/TiO2	MoSe2	9.2%	CSS	[25] Figure 5b
CdS	MoSe2	10.12%	IVD	[29] Figure 2d
CdS:O	/	7%	CSS	[71]
CeO2	/	5.14%	RTE	[78]
CdS	NiOx	6.5%	VTD	[87]
SnO2/CdS	t-Se	7.45%	CSS	[34]
CdS	MoS3	6.65%	TE	[88]
CdS	MoO2	8.14%	RF-MS + selenization	[89]
CdS	CuSbSe2	5.87%	VTD	[90]
CdS	Mo/PbSe	8.4%	CSS	[91]
TiO2	PCDTBT	6.6%	CSS	[92]
TiO2	P3HT	6.88%	CSS	[93]
CdS	SpirOmetad	7.27%	VTD	[94]

, the ETL, HTL, and their effects on the solar cells’ parameters are summarized.

3.2.2. HTLs

Many hole transport layer (HTL) materials, both inorganic and organic, have been investigated and are still being explored for Sb₂Se₃ solar cells. Inorganic HTLs have the advantage of being stable and generally having higher carrier mobilities, while organic HTLs can be easily processed by printing (inkjet, roll-to-roll) in flexible organic PV.

Y. Cao et al. [95] theoretically studied inorganic hole transport materials such as Cu₂O, CuI, and NiO for Ag/ETL/Sb₂Se₃/HTL/ITO/glass architectures. With the proper absorber thickness, HTL acceptor concentration, and hole mobility, NiO as an HTL and PCBM as an ETL, the authors calculated a theoretical 24.7% efficient device.

A. Sunny et al. [96] proposed an Al/FTO/CdS/Sb₂Se₃/SnS/Mo structure, with a SnS HTL and a CdS ETL, they calculated by 1D (SCAPS-1D) simulation software an interesting PCE of 29.89%.

S. Campbell et al. [97] theoretically studied MoO_x and NiO HTLs in Sb₂Se₃ solar cells. They observed that for CSS-grown superstrate solar cells, a NiO HTL was more adequate than MoO_x since J_sc was enhanced by 40% compared to reference and MoO_x-based devices. In addition, they observed that NiO layers facilitated a vertical ribbon orientation growth.

NiO_x and Ni as an HTL and alternative back contact, respectively, were investigated by J. Zhang et al. [87]. The structure with a 7 nm thick NiO_x exhibited the best PCE of 6.5%.

K. Li et al. [85] investigated CuSCN as an HTL for Sb₂Se₃ solar cells achieving a 7.5% efficiency in CuSCN/Sb₂Se₃/CdS/ITO/glass structure solar cells. In this case, the strong polarity in the Cu–C–N–S chains along the c-axis induced the desired vertical preferential growth. Next, moderate thermal treatments (in particular 85 °C for 10 min) of CuSCN also favored a Cu ion diffusion into Sb₂Se₃ causing a grain boundary inversion and the increase in Cu vacancies in CuSCN, thus enhancing the doping density of CuSCN.

C. Liu et al. [86] investigated spin-coated WO_3−x as an HTL for an Sb₂Se₃ absorber deposited by CSS, and their solar cells reached a PCE of 7.1% in Sb₂Se₃ solar cells as n-i-p devices fabricated with the structure of Au/WO_3−x/Sb₂Se₃/CdS/FTO/glass.

K. Shen et al. [34], as already mentioned, adopted a 2 nm thick trigonal Se (t-Se) film as HTL. They observed a higher open-circuit voltage (V_oc), which was attributed to the band bending at the Sb₂Se₃/t-Se interface. They also observed an increased durability of the solar cells compared to those with a Spiro-OMeTAD HTL (2% vs. 40% efficiency decrease after 40 days). This confirmed the importance of using an inorganic HTL for robust and durable Sb₂Se₃-based solar cells.

Y. Ma et al. [88] investigated the use of MoS₃ as a hole extraction and transport layer. Solar cells with the Au/MoS₃/Sb₂Se₃/CdS/FTO/glass configuration exhibited a PCE of 6.65% versus 3.89% in the absence of HTL.

J. Lin et al. [89] investigated a Mo/MoO₂ interface as HTL in a Ag/ITO/CdS/Sb₂Se₃/MoO₂/Mo/glass structure. In that case, Sb₂Se₃ was obtained by Sb precursor RF sputtering and a post-selenization of the Sb layer. The MoO₂ layer suppressed the formation of the MoSe₂ interface layer, contributing to the interface reaching a very high PCE of 8.14%. The MoO₂ interface layer also induced the growth of (211)-oriented Sb₂Se₃ with large grains.

Q. Cang et al. [90] exploited CuSbSe₂ as an HTL for a Au/CuSbSe₂/Sb₂Se₃/CdS/FTO solar cell, where the CuSbSe₂ layer was obtained by co-sputtering from Cu and Sb₂Se₃ targets; a PCE of 5.87% was obtained in comparison with 4.36% without HTL. CuSbSe₂ contributed to enhancing the carrier concentration and reducing the photocarrier recombination associated with the ultrathin CuSbSe₂ layer.

A modified HTL was studied by et al. [91] with a PbSe layer evaporated at room temperature at the solar cell back contact interface with Mo. The authors observed that PbSe simultaneously facilitated the growth of Sb₂Se₃ crystals obtained by CSS with an enhanced (hk1) orientation and hole extraction at the Sb₂Se₃/Mo interface. The effect was attributed to the reduction in deep-level traps by Pb into the Sb₂Se₃ and to the formation of a blocking barrier by PbSe, contributing to reducing the carrier recombination. An 8.4% efficient Sb₂Se₃ solar cell was achieved for Ag/AZO/ZnO/CdS/Sb₂Se₃/PbSe/Mo fabricated on a flexible polyimide film (PI) substrate.

Pasini et al. [98] investigated a Fe-S-O HTL deposited at RT by RF-MS. The solar cell structure Sb₂Se₃/ZnCdS/glass with contact strips of Fe-S-O covered by Au was reported with a PCE between 5% and 6%.

Alternatives to organic HTLs have been extensively studied for Sb₂Se₃-based solar cells, showing interesting improvements in PV performance in the last years. O. S. Hutter et al. [92] used a TiO₂ layer as ETL and poly[N-9′-heptadecanyl-2,7-carbazole-alt-5,5-(4′,7′-di-2-thienyl-2′,1′,3′- benzothiadiazole)] (PCDTBT), obtained by spin-casting, as HTL achieving an interesting 6.6% efficiency for a Au/PCDTBT/Sb₂Se₃/TiO₂/FTO/glass structure where the Sb₂Se₃ absorber was obtained by CSS and a seed layer method.

Another important organic material, P3HT, has been studied in several works as HTL [74,93]. In the case of C. H. Don et al. [93], it was deposited by spin-coating in a Au/P3HT/Sb₂Se₃/TiO₂/FTO device, achieving a PCE of 6.88%.

Spiro-OMeTAD as HTL has been employed in several works [68,72,77,82,94]. Z. Cao et al. [94] recently investigated Spiro-OMeTAD as HTL and CdS as ETL. In this case the Sb₂Se₃ absorber was fabricated by the VTD method. The authors highlighted the importance of pumping excess oxygen out of the furnace tube, removing oxygen from the CdS/Sb₂Se₃ interface to both decrease the defect concentration and enhance the (hkℓ, ℓ = 1)-oriented ribbon growth. In this way, Au/Spiro-OMeTAD/Sb₂Se₃/CdS/ITO solar cells achieved an efficiency of 7.27%.

3.3. Doping

Undoped Sb₂Se₃ generally presents a low intrinsic p-type charge carrier density (∼10¹³ cm⁻³), which induces a small built-in potential, and thus a limited V_oc. However, it is also difficult to extrinsically dope Sb₂Se₃ because the doping elements often remain located between the conductive ribbons, where they are not active, and not into the lattice [10].

On the other hand, external doping in Sb₂Se₃ has been used to passivate or suppress defects, such as V_Se defects or grain boundaries, especially between the ribbons. This can reduce nonradiative recombination and improve V_oc.

The latter is the case of Na doping that was investigated in Sb₂Se₃-based solar cells by Y. Li et al. [99]. The authors found a negligible influence on the conductivity of the Sb₂Se₃ layer, arguing that Na diffused in the film and remained inert between (Sb₄Se₆)_n ribbons. W-H Li et al. investigated Sb₂Se₃ doping with PbI₂ by thermal diffusion [100]. A 1 nm thin layer of PbI₂ was deposited on an ITO/glass substrate and then an amorphous Sb₂Se₃ film was grown on it. I and Pb interdiffusion was then promoted during annealing. Pb_Sb defects were detected in Sb₂Se₃, and Pb acted as a p-dopant, increasing the hole concentration from 10⁹ to 10¹³ cm⁻³. The authors fabricated ITO/PbI₂-doped Al/Alq3/C60/Sb₂Se₃ solar cells which exhibited a PCE of 4.43%, in comparison with a PCE of 2.87% for the undoped Sb₂Se₃-based cells.

M. Huang et al. [101] also investigated Pb doping in Sb₂Se₃. They reached a hole concentration of 10¹⁶ cm⁻³ also observing that Pb doping induced shallow defects corresponding to Pb_Sb.

Y. Ma et al. [102] studied tellurium-doped Sb₂Se₃ to reduce deep-level defects. In particular, the authors obtained Se-rich Sb₂Se₃ layers, limiting V_Sb and Se_Sb defects. This was also observed in works where the importance of selenization, also able to tailor the Se/Sb ratio to obtain Se-rich films, was clearly highlighted [103].

Cu-doped Sb₂Se₃ was presented by G. Spaggiari et al. [104] for Sb₂Se₃ obtained by LT-PED. Cu-doped Sb₂Se₃ free-carrier concentration was estimated in the order of 10¹⁵ cm⁻³ and successfully increased by about two orders of magnitude with respect to the undoped Sb₂Se₃.

Y. Li et al. [105] investigated Mg and Fe doping in Sb₂Se₃. The authors found that while Mg was predominately inert to the electric properties of Sb₂Se₃, Fe produced n-type doping even if the electron-free density was still low, around 10¹³ cm⁻³. These extrinsic defects were also associated with deep levels, which are detrimental for solar cell applications.

As already mentioned, H. Guo et al. [26] introduced KOH to the surface of Sb₂Se₃ films, improving the doping density from 10¹³ to 10¹⁵ cm⁻³, which contributed to enhance the solar cell performance (PCE of 7.16%).

X. Liu et al. [106] observed that in Sb₂Se₃ material, intrinsic defect acceptors Se_Sb and V_Sb formed preferentially in Se-rich conditions, while deep donors V_Se and Sb_Se were dominant in Se-poor compounds. The authors provided an in situ compensating process with Se vapor supplied during the growth, obtaining Se-rich conditions, and producing a higher film conductivity, shallower defects, and a longer carrier lifetime.

T. D. C. Hobson et al. [107] used MgCl₂ to n-dope Sb₂Se₃ demonstrating that chlorine was a substitutional shallow n-dopant with 10¹⁶–10¹⁷ cm⁻³ of doping density. Solar cells based on this n-type absorber demonstrated a PCE of 7.3%.

T. D. C. Hobson et al. [108] also obtained p doping, with a carrier density of 7.4 × 10¹⁴ cm⁻³, introducing Sn in Sb₂Se₃ layers. Sn substitutionally replaced Sb; however, the introduced Sn atom concentration was 7 × 10¹⁸ cm⁻³, much higher than the net carrier density, probably because of the Fermi level pinning.

Sn doping was also investigated by S. Chen et al. [109] in (Sn_xSb_1−x)₂Se₃ with x ranging from 0 to 0.1. P-type doping was identified, and electrical conductivity was improved by more than three orders of magnitude in comparison with the undoped samples.

Fe-doped Sb₂Se₃ was obtained in electrodeposited samples [110], with a hole density three orders of magnitude higher than the undoped films.

H. Zhou et al. [111] explored S doping of Sb₂Se₃ nanowire arrays with a gradient of its concentration whose characteristics allowed the formation of a cascade band structure, the improvement of the charge carrier transport, and consequently the increase in the solar cell fill factor. In [112], the authors also investigated S-doped Sb₂Se₃ (obtained by an electrodeposition method) increasing the p doping of 2.17 × 10¹⁶ cm⁻³ in the undoped sample to 5.63 × 10¹⁸ cm⁻³ in the S-doped one.

Silver doping has been studied by S. Gautam et al. [113], who observed that Ag doping of Sb₂Se₃ presented a dual nature dependent on the temperature. In particular, at low temperatures, Ag showed an activation energy close to undoped material, while at higher temperatures, the activation energy decreased to 0.26 eV.

Ni doping of Sb₂Se₃ also promotes p-conduction type [114] and its effect in enhancing the broadband in photodetector based on a nanowire array has been investigated.

Bi doping of Sb₂Se₃ permitted to achieve an n-doping density of around 10¹⁹ cm⁻³ in samples obtained by electrodeposition [115].

S. Chen et al. [116] investigated Pt doping of Sb₂Se₃ through a Pt metallic interlayer between Sb₂Se₃ and the CdS layer. Pt generated a p-doping increase of one order of magnitude (to 3.93 × 10¹⁵ cm⁻³) in relation to the undoped sample. The authors also observed a deep-level passivation and an oxygen contamination reduction effect of the formed PtSe₂ layer, which, concurrently with doping, contributed to enhance the PCE from 5.58 to 7.49% in the Ag/ITO/CdS/Sb₂Se₃:Pt/Mo/glass solar structure.

Te-doped Sb₂Se₃ film by a co-post-selenization treatment was investigated by G. Chen et al. [117]. Firstly, an RF-MS deposition method was used to obtain the Sb metallic precursors thin films, then a post-selenization doping treatment was performed in a vacuum tubular furnace with Te and Se powder sources. Te-Sb₂Se₃ thin films with a preferred (hk1) orientation was successfully obtained and the Ag/ITO/CdS/Te- Sb₂Se₃/Mo solar cell achieved a PCE of 7.61% efficiency.

4. Conclusions

In this paper, we summarized the latest research on the grain orientation, ETL and HTL configurations, and doping of planar Sb₂Se₃-based solar cells fabricated using physical deposition techniques. We focused on the relationship between the deposition conditions and the resulting photovoltaic performance.

The efficiency of Sb₂Se₃ solar cells has gradually increased in the last ten years, exceeding 10.5%. This has led to a growing interest in this material as a promising candidate for the thin-film solar cell market. Advances in fabrication methods, control of ribbon orientation, defect suppression, and passivation have certainly contributed to the rapid improvement of photovoltaic performance. However, several issues, such as V_oc losses associated with nonradiative recombination defects and effective external doping, remain under investigation and have not been fully resolved.

Among the deposition methods, high-rate growth processes, such as CSS, RTE, IVD, and VTD, have shown the most promising results (PCE > 9–10%) for Sb₂Se₃ absorber layers in terms of vertical ribbon alignment. This is a fundamental characteristic for obtaining a material with high conductivity. Other physical deposition techniques (PDTs), such as RFMS, which have recently achieved a PCE >8%, also appear to be very promising in terms of results and scalability. For ETLs and HTLs, new strategies in interface band alignments and inorganic (especially Cd-free) ETLs such as CeO₂, TiO₂, and SnO₂ (also combined in bi- or trilayers) and HTLs such as NiO_x, CuSCN, WO_3−x, and MoS₃ appear as very promising candidates for new configurations of solar cells.

Among the possible dopants for Sb₂Se₃ absorber layers, K, Cu, Sn, and Pb have been explored with some success. However, passivation, surface processes, and postdeposition treatments (chemical or thermal) are still open challenges that need to be addressed to further improve the properties of Sb₂Se₃-based devices and their power conversion efficiency (PCE).