Open access peer-reviewed chapter - ONLINE FIRST

Kesterite Cu2ZnSnS4-xSex Thin Film Solar Cells

By Kaiwen Sun, Fangyang Liu and Xiaojing Hao

Submitted: November 11th 2021Reviewed: November 24th 2021Published: December 31st 2021

DOI: 10.5772/intechopen.101744

Downloaded: 57


Kesterite Cu2ZnSnS4-xSex (CZTS) is a promising thin film photovoltaic (PV) material with low cost and nontoxic constitute as well as decent PV properties, being regarded as a PV technology that is truly compatible with terawatt deployment. The kesterite CZTS thin film solar cell has experienced impressive development since its first report in 1996 with power conversion efficiencies (PCEs) of only 0.66% to current highest value of 13.0%, while the understanding of the material, device physics, and loss mechanism is increasingly demanded. This chapter will review the development history of kesterite technology, present the basic material properties, and summarize the loss mechanism and strategies to tackle these problems to date. This chapter will help researchers have brief background knowledge of kesterite CZTS technology and understand the future direction to further propel this new technology forward.


  • kesterite
  • Cu2ZnSn(S
  • Se)4
  • CZTS
  • thin film solar cells
  • loss mechanism

1. Introduction

Thin film photovoltaic (PV) technologies such as Cu(In,Ga)Se2 (CIGS) and CdTe have already demonstrated more than 20% power conversion efficiency (PEC) [1, 2] and are at their commercial stage. However, considering the Restriction of Hazardous Substances Directive (RoHS) adopted in European Union [3] and the recent classification of critical raw materials (CRM) by the European Commission [4], emerging thin film PV technologies with RoHS-compliant and CRM-free constituents are increasingly desirable. Kesterite copper-zinc-tin-selenosulfide and related quaternary semiconductor represented by chemical formula Cu2ZnSnS4-xSex is generally accepted as one promising option for low-cost and nontoxic thin film PV.

The family of kesterite Cu2ZnSnS4-xSex includes pure sulfide Cu2ZnSnS4 (CZTS), pure selenide Cu2ZnSnSe4 (CZTSe), and selenosulfide Cu2ZnSn(S,Se)4 (CZTSSe) and other related compound semiconductors. Herein we abbreviate all Cu2ZnSnS4-xSex compounds as CZTS. The formation of CZTS is derived from cation mutations (cross-substitutions) in CuInS2 by replacing two In atoms with one Zn and one Sn [5, 6]. CZTS possess high absorption coefficient of over 104 cm−1, tunable band gap that can range from 1.0 to 1.5 eV to favorably match the solar spectrum, intrinsic p-type conductivity, and a three-dimensional symmetry of carrier transport [7, 8, 9]. These decent photovoltaic properties of CZTS have attracted considerable attention and enabled its rapid development in the last decades. The highest power conversion efficiencies (PCEs) of Cu2ZnSnS4, Cu2ZnSnSe4and Cu2ZnSn(S,Se)4 have been set to 11%, 12.5%, and 13%, respectively [10, 11, 12], which represent the best PCE among the emerging RoHS-compliant and CRM-free inorganic thin-film PV technologies.

In this chapter, the development history and current status of CZTS PV technology will be briefly introduced. The basic physical and chemical properties of CZTS thin film will be described to help readers have a better understanding of the material. The device architecture and absorber processing will be touched; finally. The limiting factors as well as the perspective for future development of this technology will also be reviewed and addressed.


2. The development history of CZTS

The CZTS single crystal was first grown by Nitsche, Sargent, and Wild in 1966 when they tried to prepare a serious of AI2BIICIVX4-type quaternary chalcogenides using iodine vapor transport [13]. The photovoltaic effect of CZTS was exhibited for the first time in 1988 by Ito and Nakazawa on a heterojunction diode consisting of cadmium-tin-oxide transparent conductive layer and CZTS thin film on a stainless steel substrate. The open-circuit voltage was measured to be 165 mV under AM1.5 illumination [14]. The open-circuit voltage was increased to 250 mV after annealing the device in air, and short-circuit current of 0.1 mA/cm2 was achieved [15].

The first CZTS solar cell with PCE of 0.66% was reported by Katagiri et al. in 1996 at PVSEC-9, with the device structure of ZnO:Al/CdS/CZTS/Mo/soda lime glass (SLG) substrate [16]. The CZTS thin film was fabricated by vapor-phase sulfurization of E–B-evaporated precursors [17]. In 1997, Friedlmeier et al. reported the CZTS solar cell with PCE of 2.3% and open-circuit voltage of 570 mV based on thermal evaporation [18]. In 1999, the Katagiri group improved the PCE up to 2.63% [19], after that they did a lot of work on optimizing the CZTS thin film and pushed the efficiency to 5.74% until 2007 [19, 20, 21, 22, 23, 24].

The CZTS solar cell started to gain intensive interest from academic and industry community from 2008 when its efficiency was further boosted to 6.7% by Katagiri et al. and when the CIGS solar cell became mature in its commercialization stage [25]. A lot of research institutes and solar cell manufactures such as Toyota, IBM, NREL (National Renewable Energy Laboratory), Solar Frontier, EMPA (Swiss Federal Laboratories for Materials Science and Technology), HZB (Helmholtz-Zentrum Berlin), ZSW (Centre for Solar Energy and Hydrogen Research Baden-Württemberg), UNSW (University of New South Wales), and so on involved in the development of this technology and significant advances have been achieved in the following decade [10, 26, 27, 28, 29, 30, 31, 32, 33]. During the development period, Se incorporation has attracted significant attention and led to impressive progress in PCE [34, 35, 36, 37]. Moreover, various fabrication methods including vacuum deposition process and non-vacuum process such as solution method, electrochemical deposition, etc., have been developed for fabricating the CZTS thin film [29, 37, 38, 39, 40].

An important milestone in the development of CZTS solar cell is the 10% benchmark efficiency breakthrough achieved by IBM Thomas J. Watson Research Center in 2011 [34], which shows substantial commercial promise for the CZTS-based class of thin film PV materials. This breakthrough also established the leading position of IBM in CZTS PV technology research, represented by a serious of world record efficiencies [38, 39, 41]. The highest PCE for CZTS solar cell has been stagnant at 12.6% for more than 6 years since 2014 when IBM last updated their efficiency breakthrough [39]. Thanks to the better understanding of the CZTS material and the loss mechanism of the device, quite a few groups reported CZTS PCE close to the 12.6% record efficiency in recent years [11, 42, 43, 44, 45, 46]. The recently announced NREL Best Research-Cell Efficiency Chart included the newly refreshed CZTS record efficiency of 13% achieved by Xin et al. from NJUPT [12]. This small step breakthrough comes from great efforts in this research area and hopefully will bring more interests and confidence to the CZTS R&D community.


3. The physical and chemical properties of CZTS thin film

The investigation and understanding of nature of the CZTS thin film are crucial for further developing this technology. Intensive studies have been conducted during the rapid development stage; therefore, the physical and chemical properties of CZTS have been well revealed.

3.1 Crystal structure

As we briefly introduced earlier, CZTS is derived from the cation mutations of CuInS2 (CIS), while both are originated from the binary II–VI semiconductors adopting the cubic zinc-blende (or hexagonal wurtzite) structure as shown in Figure 1.

Figure 1.

Schematic illustration of the origin of CZTS structure. Reproduced from [6] with permission from the Royal Society of Chemistry.

Similarly, the structure of CZTS is also derived from ternary I-III-VI2 compounds. In general, as shown in Figure 2, chalcopyrite (CH) and CuAu-like (CA) structures are two fundamental I-III-VI2 structures that obey the octet rule [5, 47, 48]. Therefore, the quaternary CZTS is well known as two principal structures, kesterite (space group I4, Figure 2d) and stannite-type (space group I42m,Figure 2e), which are derived from CH structure and CA structure, respectively. Other primitive mixed CA structure (PMCA) (space group I42m,Figure 2f) derived from CA structure has also been reported in CZTS [5, 47, 48]. As the most common structures discussed in literature are kesterite and stannite-type, we only focus on these two structures in this section. The two structures are closely related with the main difference of cation arrangement. Both structures are composed of a cubic close-packed lattice of S anions, with half of the tetrahedral interstices occupied by cations. Sn atoms occupy the same fixed positions in both structures, but the Cu and Zn atoms are in different position [49, 50]. In kesterite structure, cation layers of CuZn, CuSn, CuZn, CuZSn alternated at z = 0, ¼, ½, and ¾ respectively (Figure 2d), while in stannite structure, ZnSn and Cu2 layers alternate with each other (Figure 2e). The similarities of structure make it difficult to distinguish kesterite and stannite experimentally by employing common X-ray diffraction and Raman spectroscopy techniques [51, 52]. Only advanced techniques such as neutron powder diffraction analysis are capable to tell them apart [50, 53].

Figure 2.

The crystal structure of (a) zinc-blende ZnS, (b) chalcopyrite CuInS2, (c) CuAu-like CuInS2, (d) Kesterite-type Cu2ZnSnS4, (e) stannite-type Cu2ZnSnS4, and (f) PMCA-Cu2ZnSnS4. Reproduced from [47] with permission from the American Physical Society.

CZTS usually exists as kesterite-type structure, which is more stable thermodynamically than the stannite-type [54]. This is in agreement with the experimental observation [21, 50, 55, 56]. Plenty of theoretical studies have also confirmed that the kesterite-type structure is the ground state structure in CZTS [5, 47, 48, 57, 58, 59]. This is also the reason that CZTS is named after Kesterite because it crystallizes as kesterite structure. However, the energy difference between the kesterite and the stannite-type structure is rather small [5, 47, 48, 57, 58, 59]. This indicates that kesterite structure should be formed under equilibrium growth conditions, but both phases may exist, especially when growth method and conditions are changed, it should be relatively easy to grow materials with mixed phases. This may also partly explain the existence of disorder structure with more random distribution of the Cu and Zn on the cation positions [60], which will be discussed in more details in the following sections.

3.2 Electronic band structure

Kesterite CZTS has a similar tetrahedral bonds geometry as traditional group I–V, III–V, and II–VI semiconductors. It obeys Lewis’ octet rule with eight electrons around each anion atom (S or Se); the four bonds of each anion therefore form together a close valence shell [48]. However, there are fundamental differences between the Cu-based quaternary compound and the group I–V, III–V, and II–VI binary semiconductors. First of all, the bonds in CZTS involve Cu-d–anion-p hybridized antibonding states. This weakens the bonds in CZTS. Second, Cu in CZTS has only one valence s-like electron while the group III–V (group II–VI) semiconductors have three (two) valence s-like electrons for their cations.

Several first principle studies have revealed the electronic band structure of CZTS [47, 48, 58, 59, 61, 62]. An example is as shown in Figure 3, the calculated band structures of kesterite Cu2ZnSnS4 and Cu2ZnSnSe4 from different methods are presented. Overall, the kesterite CZTS materials are all direct-gap semiconductors. Different calculation methods generated slightly different values, for example, generalized gradient approximation (GGA) method calculated the band gap energy for Cu2ZnSnS4 and Cu2ZnSnSe4 as 1.56 eV and 1.05 eV, respectively. Corresponding band gap values from hybrid functional calculations (HSE06) are 1.47 and 0.90 eV, while the values from the GW0 calculation are 1.57, and 0.72 eV. These values are in agreement with other calculated data and available experimental measurement, converging to the results that Eg ≈ 1.5 eV in CZTS and Eg ≈ 1.0 eV in CZTSe [14, 24, 28, 64, 65, 66, 67, 68, 69, 70, 71, 72] to vary linearly as a function of the Se content x [73] with Eg = 1.47, 1.30, 1.17, 1.01, and 0.90 eV for x = 0, 1/4, 1/2, 3/4, and 1, using the HSE06 potential. This linear relationship also agrees with experimental results [66, 70, 71].

Figure 3.

The electronic band structure of the kesterite structures of Cu2ZnSnS4 and Cu2ZnSnSe4 along four symmetry directions. The energy refers to the VBM (dashed lines). The spin–orbit interaction is included, but the index of the bands refers to spin-independent bands wherec1represents the lowest CB andv1represents the topmost VB. Different line shapes represent different calculation methods [63]. Reproduced from [59] with permission from American Institute of Physics.

The density of states (DOS) of kesterite Cu2ZnSnS4 and Cu2ZnSnSe4 can also be calculated as illustrated in Figure 4 [63]. It is not surprising to see that Cu2ZnSnS4 and Cu2ZnSnSe4 show comparable DOS as they are in same tetrahedral bond geometries. The DOS of conduction bands in Cu2ZnSnS4 is ~0.5 eV higher than that of Cu2ZnSnSe4 because of large energy gap. It is common that in Cu-based chalcogenides including the quaternary (for example, CIS) and ternary compounds (for example, CZTS), the valence band maximum (VBM) is derived mainly from the hybridization of anion p and Cu d states because Cu has higher d orbital energy than Zn, Ga, In, and Sn [59, 74]. In CZTS, the valence p level of S is lower in energy than Se, thus the VBM of the sulfides is lower than that of the selenides. This difference is reduced by anion p–Cu d overlap (p-d hybridization) because the hybridization is stronger in the shorter Cu–S bond and pushes the antibonding VBM level of the sulfide up relative to that of the selenide. Therefore, the valence band offset between the Cu2ZnSnS4 and Cu2ZnSnSe4 is less than 0.2 eV. The DOS of the CZTS conduction band minimum (CBM) is primarily controlled by the Sn-s and anion p-like states due to the lower s orbital energy of Sn than the other cations [59, 74]. More importantly, in CZTS the lowest CB is separated from the higher energy bands and is therefore more localized in energy. Therefore, the CBM of CZTS is expected to vary depending on the alloying on the Sn-site with other group-IV elements (e.g., Ge) due to this localized CB. The band gap energy can therefore be tailored by cation alloying for an optimized optical efficiency of the materials.

Figure 4.

Atomic and angular momentum resolved DOS of kesterite Cu2ZnSnS4 and Cu2ZnSnSe4 presented with a 70 meV Lorentzian broadening. Reproduced from [59] with permission from American Institute of Physics.

3.3 Composition variation and phase competition

As a quaternary semiconductor, CZTS consists of three metals and up to two chalcogens, which provides a wide range for composition variation and secondary phase formation. However, high-efficiency CZTS thin film solar cells require single-phase kesterite absorber, and some secondary phases have been reported detrimental to the device performance. This section reviews the stable phase region of CZTS and possible secondary phases likely to be formed in the quaternary system.

The 3D CZTS quaternary phase diagram as shown in Figure 5 can be deduced from Cu2S-ZnS-SnS2 pseudo-ternary diagram [77] and CuS-ZnS-SnS pseudo-ternary phase diagram [75] (selenide kesterite system also refers to Cu2SnSe3–SnSe2–ZnSe diagram [78]. In addition, the stability region of the CZTS in the atomic chemical potential space can be calculated as shown in Figure 5 [76, 79]. All these diagrams indicate that volume of the stable CZTS region is small, and the slight deviation (maximum 1–2 at%) outside this space will cause the formation of different secondary phases [5, 80]. The narrow stable window also implies that the composition control and the chemical potential control are crucial for the growth of high-quality and single-phase CZTS absorber. The Zn content is particularly important because the stable region along the μZn axis is much narrower. Moreover, it is well accepted experimentally and theoretically that high-efficiency solar cells need CZTS absorber with Zn-rich and Cu-poor composition [8, 9, 10, 27, 28, 33, 67, 81]. This makes it more difficult to control the secondary phases formation.

Figure 5.

Left: CZTS quaternary phase diagram including the known phases. Reproduced from [75] with permission from American Institute of Physics. Right: The calculated chemical-potential stability diagram of Cu2ZnSnS4 in a 2D Cu-rich plane (the stable 3D region is inset) reproduced from [76] with permission from American Institute of Physics.

According to the phase diagram and literature report, the most common secondary phases in the Cu-Zn-Sn-S/Se system are list in Table 1 [92]. The influence of the secondary phases on the solar cell performance depends on their position in the film as well as their physical properties. For example, Cu2S(Se) phases in the final film may act as shunting path due to both the high conductivity and contact with front and back interfaces. However, Cu2S(Se) is also an important fluxing agent to promote lateral grain growth during the film growth [28]. Generally, If the secondary phase has a lower band gap than the CZTS absorber, it will limit the open-circuit voltage of the solar cell. While secondary phases with higher band gaps than CZTS are less detrimental; however, they can block the transport when present in large amounts [93] or at least increase the series resistance [94]. ZnS(e) is the most likely secondary phase considering the phase diagram and especially when adapting the Zn-rich and Cu-poor condition. Fortunately, ZnS(e) is a large band gap compound and is expected to be rather benign if present in small amounts. It has even been reported that ZnS with similar crystalline structures to CZTS may passivate grain boundaries or heterojunction interface [81, 95, 96] by reducing strain and lowering recombination velocities at the grain interfaces. The tin compounds are unlikely to form because they are usually volatile [97] and will evaporate in most preparation conditions. As shown in Table 1, one unfavorable low band gap secondary phase in both Cu2ZnSnS4 and Cu2ZnSnSe4 system is the ternary Cu2SnS(e)3, which is also one reason for optimal composition range (Zn-rich and Cu-poor) found for the best solar cells. However, the non-homogeneous composition across the CZTS film under normal preparation conditions still provides the possibility to form such detrimental secondary phases [98]. Therefore, how to control the amount and position of the secondary phases in the CZTS absorber needs to be elaborately studied.

CompoundBand gap (eV)Ref.CompoundBand gap (eV)Ref.
SnS1.0 indirect, 1.3 direct[87, 88]SnSe1.3[89]

Table 1.

Most common secondary phases in the Cu-Zn-Sn-S/Se system.

3.4 Lattice defects

The formation and properties of lattice defects are important parameters of semiconductor materials and are crucial to the function of photovoltaic devices, because they directly influence the generation, separation, and recombination of electron–hole pairs. The lattice defects (e.g., vacancies, interstitials, antisites) in kesterite CZTS system are complicated because of the increased number of component elements and the similar cation size as well as small chemical mismatch of Cu+ and Zn2+. In this section, the formation and ionization of the lattice defects in kesterite will be briefly reviewed. More importantly, the underlying mechanism of p-type conductivity, Zn-rich and Cu-poor condition growth condition, as well as limiting factors for device performance in kesterite CZTS solar cells from the perspective of lattice defects will be discussed.

The concentration of the lattice defects is determined by their formation energy. In Figure 6, the calculated formation energies of different defects are plotted as functions of the Fermi energy (0 means that the Fermi energy is at VBM, while 1.5 or 1.0 eV means that Fermi energy is at CBM). It is obvious that in both Cu2ZnSnS4 and Cu2ZnSnSe4, CuZn antisite is the lowest energy defects, which is different from the defect properties of their parent compounds (CuInSe2 or CuGaSe2) where the dominant defect is the Cu vacancy VCu [76, 79, 99]. In addition, the formation energies of most acceptor defects are lower than those of donor defects, explaining the intrinsic p-type conductivity observed in literature [21, 48, 64, 84, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110].

Figure 6.

Calculated defect formation energy as a function of the Fermi energy at the thermodynamic chemical-potential point a (fromFigure 5right) for Cu2SnZnS4 (left) and Cu2SnZnSe4 (right). For each value of the Fermi energy, only the most stable charge state is plotted, with the filled circles (change of slope) representing a change in charge state (transition energy level). Reproduced from [76,79] with permission from American Institute of Physics and American Physical Society.

Another important parameter of lattice defects is their ionization (transition) levels, which determines if they can produce free carriers and contribute to the electrical conductivity. The calculated ionization levels of intrinsic defects in the band gap of Cu2ZnSnS4 and Cu2ZnSnSe4 are shown in Figure 7. First of all, in both cases, the dominant defect CuZn has an acceptor level (0.11 eV and 0.15 eV above VBM in Cu2ZnSnSe4 and in Cu2ZnSnS4, respectively) deeper than that of VCu. The relatively deep level of dominant antisite defect in CZTS is not favorable to the device performance because it will limit the open-circuit voltage. This also partially explains why Zn-rich and Cu-poor condition has normally been found to beneficial to solar cell efficiency because it could decrease the formation energy and enhance the population of shallow VCu. The other acceptor defects (e.g., CuSn, ZnSn, VZn, and VSn) have higher formation energy; therefore, they have negligible contribution to the p-type conductivity. However, they may act as recombination centers especially for those deep transition levels such as (4−/3-) and (3−/2-) in band gap as seen in Figure 7.

Figure 7.

The ionization levels of intrinsic defects in the band gaps of Cu2ZnSnS4 (top) and Cu2ZnSnSe4 (bottom). The red bars show the acceptor levels, and the blue bars show the donor levels, with the initial and final charge states labeled in parentheses. Reproduced from [79] with permission from American Physical Society.

In addition to the point defects, various self-compensated defect clusters can be formed in CZTS due to the large amount of low-energy intrinsic defects. Defect compensation in ternary CIS is well known to have electrically benign character because the intrinsic defects undergo self-passivation through the formation of defect complexes such as [2VCu + InCu2+]. Therefore, it is also interesting to see if the same behavior can be observed in quaternary kesterites. CuZn and ZnCu are the lowest-energy acceptor and donor defects, respectively; therefore, antisite pair [CuZn + ZnCu+] also shows extremely low formation energy. Fortunately, its impact on the electronic structure and optical properties is relatively weak. The detrimental defect clusters are those composed of deep level defects, such as SnZn, SnCu, CuSn, and Zni. For example, clusters with SnZn induce large conduction band edge downshift, which could limit the solar cell performance, because the induced states are deep and may trap photo-generated electrons from the high conduction band. [2CuZn + SnZn] clusters could present in high population if in single-phase CZTS (Cu/(Zn + Sn) and Zn/Sn ratios near 1) chemical potential conditions and be detrimental to the solar cell performance. The Zn-rich and Cu-poor condition could prevent the formation of [2CuZn + SnZn] clusters because its formation energy is very sensitive to the chemical potential of Zn. This again partly explains that the Zn-rich and Cu-poor condition is beneficial to solar cell efficiencies from the defect perspective.


4. The path towards high-efficiency kesterite solar cell

4.1 Device architecture

The typical device architecture (Figure 8) of kesterite-based solar cell inherits from its predecessor chalcopyrite CuInSe2 solar cells due to their similar optical and electronic properties. Metal Mo layer with thickness of 500 nm–1000 nm is usually deposited with sputtering on soda lime glass or other substrates as back contact. Then the kesterite CZTS absorber layer is deposited on top of the Mo layer. The p-n junction is formed by p-type CZTS and the following deposited n-type buffer layer. Typical n-type buffer layer in kesterite and chalcopyrite solar cell is 50 nm–100 nm thick CdS layer usually by chemical bath deposition. Alternative buffer materials such as ZnSnO, Zn(O, S), ZnCdS, etc., have been explored, especially in pure sulfide Cu2ZnSnS4 to tackle the unfavorable band alignment found at p-n junction. Next, a 50 nm–100 nm high-resistive intrinsic ZnO (i-ZnO) is deposited. Subsequently, the device structure is completed by deposition of 200 nm–500 nm thick Al-doped ZnO (AZO) or indium tin oxide (ITO) transparent conducting oxide (TCO) layer as the front contact. Ni/Al metal contacts are deposited on the TCO layer for improved current collection. Antireflection coating such as MgF2 is often deposited on top of the cell to reduce the reflection loss.

Figure 8.

Schematic device structure of a typical kesterite Cu2ZnSnS4 thin film solar cell.

Various deposition methods have been investigated for producing high-quality CZTS layer during the development of this technology. These deposition methods are broadly classified as vacuum-based and non-vacuum-based techniques. The vacuum-based methods are usually considered easy to be expanded to commercial scale because of the precise process control. All physical vapor deposition techniques including thermal evaporation, E-beam evaporation [17], sputtering [10, 33, 81, 111], pulsed laser deposition [112] fall into this category. The non-vacuum-based methods are always regarded as low-cost, high-throughput techniques and feasible in roll-to-roll production. These methods usually involve chemical or solution process, such as electrochemical deposition [40], nanoparticle-based synthesis [113, 114], sol–gel spin coating [44, 115], chemical bath deposition (CBD) [116], successive ion layer adsorption and reaction (SILAR) [117], screen printing [118], etc.

4.2 Loss mechanism and solutions

4.2.1 Open-circuit voltage (VOC)

In kesterite CZTS solar cells, it has been widely accepted that VOC loss accounts for the majority (more than 50% [119]) of the efficiency gap between the best performance device and the theoretical limit (i.e., Shockley–Queisser radiative limit [120]). VOC loss is determined by the recombination path in the device. The dominant recombination can occur in the bulk of the absorber in the quasi-neutral zone as band-to-band recombination, via defects, or in the space charge region. Another important recombination path can be located at the heterojunction interface between the buffer/window layer and absorber.

Therefore, the absorber problems discussed in Section 3such as abundant point defects and defect clusters (i.e., cation disordering), composition variation with narrow stable region are believed to contribute to the recombination. The Cu and Zn substitution with low energy causes a large population of antisite defects such as CuZn and ZnCu and related defects complexes. Consequently, severe electrostatic potential fluctuation and associated band tailing can be observed [121]. In addition, the microinhomogeneities in composition, nonuniform strain, as well as formation of secondary phases lead to band gap fluctuations, which are also detrimental to VOC. Moreover, acceptor-like CuZn defects at the interface will cause Fermi level pinning to a low energy level [122], thus reducing the band bending in the absorber and weakening the electric field. To alleviate the recombination occurring in absorber layer, great efforts have been made over recent years. One solution that has been intensively investigated is cation substitutions because the introduction of larger size cations could result in better cationic ordering, thereby reducing the defects formed because of the Cu and Zn substitution. Using Ag as a substitute for Cu [115], Cd as a substitute for Zn [123, 124], and Ge substitute for Sn [125]are popular choices as they are picked from the same cation groups. As shown in Figure 9, Promising progress has been made in this direction both experimentally and theoretically, demonstrating several remarkable efficiencies and mechanism of the defect emerging with cation substitution. Another possible solution to tackle the cationic disorder and related band tailing, as well as to activate the shallower defect is deliberate control of the synthesis condition, for example, post annealing the CZTS absorber within the critical temperature range of 200250°C and reasonable time range of 14 hours will improve the CZTS lattice ordering and increase the band gap [126, 127]. Control of the precursor fabrication process such as the metal stack order, the valence states of the Sn source, has also been reported effective in obtaining less defective absorber and corresponding high-efficiency device [128, 129]. Modification of the sulfurization or selenization condition could also suppress the formation of detrimental intrinsic defects and defect clusters by creating a desirable local chemical environment [11].

Figure 9.

(a)The band diagrams of CZTSSe and Ag-graded (Cu1 − xAgx)2ZnSn(S,Se)4 (CAZTSSe) solar cells under AM 1.5G 100 mW cm2illumination and theJ–Vcurves of CAZTSSe solar cells with different Ag composition gradients from the CdS interface to the Mo substrate.(b)J–Vcurves of CZTS and champion Cu2ZnxCd1–xSnS4 (CZCTS) devices and absorption coefficients and Urbach energy (EU) obtained from PDS measurement.(c)J–Vcurves for the champion CZTSSe and CZTGeSSe (30% Ge) solar cells. Reproduced from [115,124,125] with permission from Royal Society of Chemistry, American Chemical Society and John Wiley and Sons.

In term of the interface recombination path, it is mainly caused by the unpassivated charge defects at the heterojunction and/or the undesirable band alignment between the CZTS absorber and conventional CdS buffer layer. The band alignment problem is more prominent in high band gap pure sulfide Cu2ZnSnS4 because of its higher conduction band position. A serious of alternative wide band gap buffer layer materials including ZnCdS [33], Zn(O,S) [130, 131] and ZnSnO [132, 133] have been reported effective in mitigating the band alignment issues. The charge defects at the interface can be reduced by introducing ultrathin layer that has better lattice match with CZTS [10, 81] or alloying Ag near the heterojunction interface to from intrinsic or weak n-type (Cu,Ag)2ZnSn(S,Se)4 [115, 134]. Dielectric layers have also been investigated for passivating the interface defects, thereby suppressing the interface recombination [111, 135].

4.2.2 Short-circuit current density (JSC)

While the VOC loss contributes dominant performance loss in kestrite solar cells, JSC also represents a significant limiting factor for efficiency increases. The JSC loss is mainly caused by two reasons: one is the light or photons that is reflected or absorbed by layers above the CZTS (such as the buffer or window layers as shown in Figure 8), which cannot generate electron–hole pairs. The other is the low carrier collection efficiency due to the short diffusion length or narrow depletion width. The first problem can be addressed by introducing antireflection coating and optimizing the optical designs of the top layers above CZTS to increase the fraction of incident light reaching CZTS. The solution has been proven feasible when applying wide band gap buffer layer [111, 133], reducing the thickness and roughness of the top layers [119]. The typical External Quantum Efficiency (EQE) curve of CZTS usually shows relative lack of long wavelength response [9], which might be related to the low carrier lifetime. This low lifetime could also be the results of high defect density in the absorber layer or could simply be a consequence of high recombination loss at the back contact or at the front interface. Therefore, the solutions for suppressing the formation of detrimental defects disused in VOC loss above are also good for the JSC improvement In addition, band gap grading has also been developed in CZTS to increase the EQE by controlling the [S]/[S + Se] ratio at the surface and back surface [136].

4.2.3 Fill factor (FF)

FF is another significant parameter deficit in kesterite solar cell, which is mainly limited by the series resistance (RS) in the device. Careful study has been conducted to identify the origin of the RS, which implies that nonohmic back contact contributes greatly [137]. The nonohmic back contact could be resulted from Schottky barrier formation and/or secondary phase formation at the CZTS/Mo back interface. It is especially severe in high band gap CZTS because of the concentration decrease. Furthermore, the presence of low band gap secondary phases in CZTS could act as shunting pathway, which limits the FF and is therefore undesirable in the device as well. Deliberate investigation and control of the chemical reaction at the back interface have been proposed to modify the back interface microstructure and improve the device performance including FF [138, 139]. Introducing barrier layer at the CZTS/Mo interface has also been attempted in improving the back interface quality [140, 141].


5. Conclusion

Kesterite CZTS material is an emerging and promising PV technology, which could have the opportunity to realize low-cost and high-volume thin film solar cell production. After undergoing a rapid development in last decades, the technology has attracted increasing attention and demonstrated high potential in high efficiency. Experimental and theoretical studies reveal that the quaternary material exits in kesterite structure, similar to the chalcogenide CuInS2. It has a narrow stable region regarding the chemical potential, which makes the formation of secondary phases very easy. The intrinsic defects characteristic in the CZTS is complex where the deep level antisites defects and defect cluster could be prevalent. These unique properties make it struggling in achieving high-efficiency kesterite solar cells. The sensitive defects environment causes undesirable band tailing, electrostatic potential fluctuations, etc., which become recombination path and limit the open-circuit voltage, as well as other PV performance. The formation of secondary phases can lead to high serious resistance, shunting path depends on the band gap of the phase, which will also limit the device performance. In addition, the heterojunction interface and the Mo/CZTS back interface could also contribute to the performance loss because of the unfavorable band alignment, unpassivated defects, or detrimental reaction during high-temperature annealing. The solutions and approaches for tackling these loss mechanisms have been reviewed at the end of the chapter. Finally, in order to build a successful kesterite CZTS technology, combining the advanced approaches summarized in this chapter with further exploring the material synthesis and device physics would pave a path for higher efficiency.



This work was supported by the National Key R&D Program of China (No. 2018YFE0203400), the Science and Technology Innovation Program of Hunan Province (No. 2020RC2005), Australian Renewable Energy Agency (ARENA, 2017/RND006), and International Postdoctoral Exchange Fellowship Program (YJ20200207). X.H. acknowledges Australian Research Council (ARC) Future Fellowship (FT190100756). K.S. acknowledges the support from Australian Government through the Australian Renewable Energy Agency (ARENA) and the Australian Centre of Advanced Photovoltaics (ACAP, Grant No.1-SRI001).


Conflict of interest

The authors declare no conflict of interest.


chapter PDF

© 2021 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution 3.0 License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

How to cite and reference

Link to this chapter Copy to clipboard

Cite this chapter Copy to clipboard

Kaiwen Sun, Fangyang Liu and Xiaojing Hao (December 31st 2021). Kesterite Cu<sub>2</sub>ZnSnS<sub>4-x</sub>Se<sub>x</sub> Thin Film Solar Cells [Online First], IntechOpen, DOI: 10.5772/intechopen.101744. Available from:

chapter statistics

57total chapter downloads

More statistics for editors and authors

Login to your personal dashboard for more detailed statistics on your publications.

Access personal reporting

We are IntechOpen, the world's leading publisher of Open Access books. Built by scientists, for scientists. Our readership spans scientists, professors, researchers, librarians, and students, as well as business professionals. We share our knowledge and peer-reveiwed research papers with libraries, scientific and engineering societies, and also work with corporate R&D departments and government entities.

More About Us