Revealing the impact of Na–Ag single-site Co-doping in CZTSSe

Abstract

Cation doping/substitution is a widely used strategy to overcome the serious open-circuit voltage deficit caused by unfavorable defects and defect clusters in kesterite materials such as CZTSSe. Single cation substitution has proved effective in improving PV performance. However, single element doping of CZTSSe has its own limitations because dopants can reduce carrier concentrations and introduce unfavorable lattice distortion. Co-doping offers broader possibilities for regulating the optoelectronic properties of CZTSSe materials but has seldom been explored due to the enhanced complexity in such systems. Herein, a single-site, co-doping strategy is proposed using Na and Ag ions, which both target the same copper lattice site in CZTSSe. The Na⁺ ions passivate Cu-related defects while the introduction of Ag⁺ reduces the lattice distortion that the Na ions create. Conversely, Na ions help to compensate for the reduction in carrier concentration caused by Ag⁺ doping. Furthermore, the synergistic doping of Na and Ag also promotes the crystal growth of the absorption layer, resulting in a compact, large-grain absorption layer and enhancing the performance of the device. After optimization, we achieved a V_oc of 559 mV and a fill factor (FF) of 67.3% and an overall power conversion efficiency of 13.2%.

---

Introduction

Kesterite Cu₂ZnSn(S,Se)₄ (CZTSSe) is a leading candidate for thin film photovoltaic devices because of its high light absorption coefficient, earth-abundant constituents, tunable band-gap between 1.0–1.5 eV and high theoretical power conversion efficiency (PCE).^1–4 However, the highest PCE obtained from CZTSSe thin film solar cells to date is 15.8%,⁵ still far below the Shockley–Queisser predicted efficiency limit (>30%) and well below the highest PCE obtained from Cu(In,Ga)Se₂ (CIGS) solar cells (23.64%).^6,7 Recently, breakthroughs in the PCE of CZTS and CZTSSe solar cells have been achieved by cation doping. For example, Ag⁺ substitution of Cu⁺,^4,8,9 Cd²⁺ substitution of Zn²⁺,^10–13 and Ge(IV) substitution of Sn(IV)¹⁴ have all led to increased solar conversion efficiencies. Furthermore, direct doping with alkali metal ions including Li,^15,16 Na,^17,18 and K has also improved device performance. The incorporation of alkali metals has been found to improve the crystal quality,^15,19–21 electrical properties^16,21–24 and defect passivation^21,25–27 of the bulk material. Despite the positive impact of alkali metal ions on kesterite materials, their effectiveness has certain limitations. There are strict limitations on the amount and distribution of the alkali metals that can be doped into the material, and it is difficult to achieve adequate defect passivation after doping.¹⁸

Other cations have also been incorporated into kesterite based solar cells. Yan et al. substituted Zn by Cd through heterojunction annealing to reduce band edge tailing and to increase the minority carrier lifetime.²⁸ Wang et al. introduced a GeO₂ layer on the bottom interface to modify the selenization process, achieving a flatter absorber layer with fewer voids as well as significantly reducing the V_OC loss.¹⁴ Zhao et al. introduced Ga and were able to increase the band bending in grain boundaries and boost both the V_OC and J_SC.²⁹ Among these various dopants, the use of Ag doping in CZTSSe has garnered significant attention due to the recent record power conversion efficiencies (PCEs) of 23.64% for CIGSe and 14.9% for CZTSSe solar cells, achieved through partial substitution of Cu with Ag.^6,30 Silver doping offers potential benefits, including defect passivation and reduced charge carrier recombination, which in turn increases the minority carrier lifetime and hence the minority carrier diffusion length.^4,8 However, it should not be overlooked that alloying with Ag⁺ has a negative impact on the carrier concentration of the absorber layer.³¹ To fully realize the potential of CZTSSe solar cells, more effective approaches are needed to minimize the carrier losses, which can limit both the electrical conductivity of the absorber layer as well as the theoretical fill factor (FF) that can be achieved.

Co-doping opens new possibilities for addressing the challenges mentioned above. Introducing two or more elements into the system enhances the potential to tackle issues that individual elements alone cannot resolve. Unold and Wong demonstrated the effects of double cation substitution in CZTS by partially replacing Cu⁺ with Ag⁺ and Zn²⁺ with Cd²⁺, achieving over 10% power conversion efficiency (PCE) in pure sulfide kesterite devices. Their findings indicate that Cd²⁺ enhances device performance by modifying the defect characteristics of acceptor states near the valence band, while Ag⁺ helps reduce nonradiative bulk recombination.³² Liang and Chen also employed a dual-cation strategy, substituting Cu⁺ with Ag⁺ and Sn⁺ with Ti⁺ in CZTSSe. The Ag⁺ substitution serves two roles: it enhances the film crystallinity and it also reduces the Cu_Zn defect density. Concurrently, Ti⁴⁺ incorporation removes fine grains and suppresses major deep-level defects such as Cu_Sn and [2Cu_Zn + Sn_Zn].³³ Furthermore, Hao's group explored a combination of Cd²⁺ and Ge⁴⁺ ions as dopants and obtained a PCE of 11.6% in CZTSSe solar cells. Ge⁴⁺ ions increase the amount of p-type doping, which improves the conductance and suppresses bulk recombination via enhanced quasi-Fermi level splitting; Cd²⁺ ions passivate junction defects, which in turn widens the depletion region and improves carrier collection.³⁴ While effective, this dual-dopant approach introduces an inherent complexity into the CZTSSe system, presenting a key obstacle to further optimization.³⁴

Rather than having two dopants located at different lattice sites in the CZTSSe crystal, a single-site co-doping strategy is proposed here to simplify the analysis and clarify the underlying mechanisms. This approach leverages the complementary roles of Ag and Na: Ag ions reduce non-radiative recombination, while Na ions compensate for the carrier reduction caused by Ag doping. We postulate that, together, they can effectively mitigate losses due to Cu-related point defects, enabling the formation of better-quality absorbers with fewer defects. Additionally, without sacrifice of carrier density, the carrier lifetime and transport efficiency should be improved in the bulk. Moreover, the Ag ions and Na ions synergistically improve the selenization process and yield a compact, large-grain absorber layer without the usual bottom-level, fine-grain region, which reduces carrier transport near the back contact.^35,36 Indeed we find that with careful optimization kesterite solar cells with Na–Ag co-doping can achieve a PCE of 13.2%, with significant enhancements in open-circuit voltage (V_oc) and fill factor (FF).

Results and discussion

Effects of ion doping on local chemistry and grain growth

To investigate the doping effects of Na⁺ and Ag⁺ ions on absorber properties and device performance, four samples were fabricated using different doping strategies (Fig. 1a–d). Moreover, to eliminate the influence of Na ions in soda-lime glass (SLG), Mo-coated SLG with a SiO_x barrier layer between Mo and SLG was used, which blocks the diffusion of Na from SLG into the CZTS precursor layer. Na-doped precursor films were fabricated by adding NaCl to the CZTS solution. For Ag doping, an Ag-containing CZTS solution was spin-coated onto an undoped precursor film—a method previously shown to improve performance (Fig. S1). Na–Ag co-doping was achieved by spin-coating the Ag-containing CZTS solution onto Na-doped precursor films. Furthermore, we employed X-ray diffraction (XRD) (Fig. 1e and i) and Raman spectroscopy (Fig. S2) to confirm the successful incorporation of Na and Ag ions into the CZTSSe lattice.

Fig. 1 Schematic illustrations showing different doping strategies. (a) undoped control devices, (b) Na doping strategy, (c) Ag doping strategy, (d) Na–Ag co-doping strategy. (e) Full range and (i) high resolution XRD patterns for different, doped CZTSSe films. XPS spectra of (f) Cu 2p, (g) Zn 2p, (h) Sn 3d, (j) Se 3d, (k) Na 1s, (l) Ag 3d for different doped CZTSSe films.

The obvious peak shift to a smaller angle in Fig. 1e is ascribed to lattice expansion due to Na and Ag doping into Cu sites. The intermediate peak position for the Na–Ag co-doped sample reflects competing interactions between the dopants during substitution. The local chemical environment was also probed by X-ray photoelectron spectroscopy (XPS) (Fig. 1f–h and j–l). The Cu 2p spectra (Fig. 1f) show a shift towards higher binding energy and a decrease in peak intensity upon doping. Although the shift in the Cu 2p peaks could be attributed to several factors, such as defect-driven Fermi level movement or local stoichiometry variations, our collective evidence suggests that the substitution of Cu by Na and Ag in the CZTS lattice is the primary cause. This conclusion is supported by the emergence of the Na 1s and Ag 3d peaks (Fig. 1k and l), combined with the corresponding peak shifts observed in the XRD patterns and Raman spectra.^33,37 The observed peak splitting confirms the oxidation states of the constituent elements: Cu(I) (19.9 eV for Cu 2p), Zn(II) (23.2 eV for Zn 2p), and Sn(IV) (8.4 eV for Sn 3d). For the dopants, the Na 1s peak at 1072.0 eV is consistent with Na(I), while the Ag 3d peak with a 6.0 eV splitting is characteristic of Ag(I). Furthermore, the Na 1s signal (Fig. 1k) is stronger in the Na–Ag co-doped film than in the purely Na-doped film, suggesting that Ag co-doping facilitates the incorporation of Na ions into the lattice. Conversely, the Ag signal remains clear in both Ag-doped and co-doped films, indicating that Na doping does not significantly impede Ag substitution.

SEM images (Fig. 2) revealed the influence of Na and Ag doping on grain growth. The undoped reference sample exhibited sharp grain edges and visible pinholes, indicating insufficient grain growth in the absorber layer. Both Na and Ag doping effectively addressed these issues: Na doping increased the average grain size and promoted the formation of a denser upper grain layer, while Ag doping alone enhanced vertical grain growth, but with void formation still evident. The Na–Ag co-doped sample exhibited superior film quality, forming a void-less, highly crystallized, bilayer structure. These results suggest that Na⁺ and Ag⁺ cations enhance both lateral and vertical crystal growth.

Fig. 2 Top-view and cross-sectional scanning electron microscope (SEM) images of (a and e) undoped, (b and f) Na doped, (c and g) Ag doped, (d and h) Na–Ag co-doped CZTS absorber layers.

Fig. 3 illustrates the crystallization and growth processes during selenization of CZTS. Fig. 3a depicts the growth process of an undoped precursor, which forms a thick, fine-grained layer with pinholes at the bottom due to insufficient CZTSSe grain growth and limited Se diffusion. When Na is doped into the CZTS (Fig. 3b), it promotes CZTSSe crystallization, yielding a more uniform absorber layer without pinholes; however, incomplete selenization of the lowermost crystals persists, leaving a residual fine-grained layer. In contrast, surface doped Ag (Fig. 3c) can react with Se vapor during selenization to form a low-melting-point, Ag–Se binary liquid phase (m.p. around 219 °C),³⁸ allowing more Se to diffuse to the bottom of the absorber layer,^39,40 which enhances selenization at the bottom interface and produces a bilayer structure with large grains. However, due to restricted lateral growth, voids remain at the bottom. When Na–Ag co-doping was employed (Fig. 3d), downward-diffusing Ag_xSe promotes the diffusivity of Na ions as well, which helps more Na to enter the CZTSSe lattice as the XPS spectra confirm. Thus, Na–Ag co-doping improves the cation distribution and ensures sufficient Se supply to the bottom interface, while Na ions enhance lateral grain growth. This cooperative process yields a flat, void-less bilayer with large CZTSSe grains, which can reduce bulk recombination and enhance carrier transport toward the back contact,³⁵ as confirmed by the Na/Ag distribution in the EDS spectra (Fig. S3).

Fig. 3 Diagram illustrating CZTSSe crystallization and growth processes in (a) undoped, (b) Na doped, (c) Ag doped, (d) Na–Ag co-doped films.

Photoelectric performance

The photovoltaic performance parameters of undoped and doped CZTSe devices including open-circuit voltage (V_OC), short-circuit current density (J_SC), fill factor (FF), and power conversion efficiency (PCE) are summarized in Fig. 4a–d. Na, Ag, and Na–Ag co-doping significantly improve the average PCE, primarily due to enhanced V_OC and FF. Ag ion doping achieves superior defect passivation, yielding a higher V_OC than Na doping, though the photocurrent reduction partially offsets the efficiency gain. Notably, Na–Ag co-doping boosts the FF substantially, elevating the average PCE from ≈11.8% (Na-doped) to ≈12.7%.

Fig. 4 The photovoltaic performance of CZTSSe solar cells obtained using different doping methods. (a) PCE, (b) FF, (c) V_OC, (d) J_SC. (e) J–V curves and (f) EQE curves of champion devices obtained using different doping methods.

Table 1 compares the key parameters of champion cells (undoped, Na-doped, Ag-doped, and Na–Ag co-doped), while Fig. 4e and f presents their J–V curves and EQE spectra. Bandgap and Urbach energies derived from the EQE plots are shown in Fig. S4. Diode analysis (Fig. S5 and Table 1) reveals decreased shunt conductance (G), reduced series resistance (R_S), lower ideality factor (A), and suppressed reverse saturation current density (J₀) in co-doped devices. Specifically, Na–Ag co-doping increases V_OC from 550 mV (Ag-doped) to 559 mV and FF from 62.8% to 67.3%. These improvements stem from (1) elimination of the fine-grained layer, (2) enhanced crystallinity with optimized defect engineering, and (3) reduced carrier transport losses (lower R_S and shunting (G)). The decline in A and J₀ further confirms suppressed non-radiative recombination at the heterojunction interface and within the bulk absorber, mitigating the V_OC deficit.

Table 1 The photovoltaic and diode parameters of champion devices fabricated using different doping methods. The diode parameters were extracted using Site's method

Device	Voc (mV)	Jsc (mA cm^−2)	FF (%)	PCE (%)	A	J 0 (mA cm^−2)	R s (ohm cm^2)	G (mS cm^−2)
Ref	0.526	35.6	58.9	11.0	2.12	9.04 × 10^−5	1.77	0.293
Na	0.545	36.4	61.8	12.3	1.66	4.54 × 10^−5	1.59	0.253
Ag	0.550	36.0	62.8	12.4	1.64	4.07 × 10^−5	1.66	0.219
Na–Ag	0.559	35.1	67.3	13.2	1.57	9.21 × 10^−6	1.38	0.107

---

In the EQE spectra (Fig. 4f), Na–Ag co-doped devices show a heightened photocurrent response at wavelengths from 450–550 nm, attributed to an improved CZTSSe/CdS heterojunction interface. Suns-V_OC measurements (Fig. S6) corroborate the claim of superior heterojunction quality in co-doped devices, with pronounced reductions being observed in the recombination currents (J₀₁, J₀₂) in both the quasi-neutral and space-charge regions. The pseudo-FF (pFF) rises by >5% versus undoped devices, reflecting enhanced diode quality (lower n_mpp) due to suppressed non-radiative recombination.

Ultraviolet photoelectron spectroscopy (UPS) was used to evaluate the band alignment of CZTSSe under different doping conditions. As shown in Fig. 5c, Na doping caused a slight upward shift of the valence band maximum (VBM) and a downward shift of the Fermi level (E_F) compared to undoped CZTSSe. This suggests that Na doping does not significantly alter the positions of the VBM or conduction band minimum (CBM). The downward shift of E_F is likely due to a reduction in Fermi level pinning, which in turn enhances V_OC and FF, as supported by device performance measurements. In contrast, Ag doping resulted in a clear downward shift of both the CBM and VBM, confirming the incorporation of Ag into the CZTSSe lattice. The associated band gap broadening led to an improvement in V_OC, while the widened gap between E_F and VBM may be attributed to a reduction in carrier density after Ag doping.

Fig. 5 (a) Extracted valence band edges and (b) Fermi level edges of different CZTSSe samples from UPS, (c) schematic band diagram of CZTSSe samples with different doping strategies (bandgaps are obtained from EQE data). E_vac, E_F, E_C, and E_V are energy of vacuum, Fermi, conduction, valence level, respectively.

When Na and Ag were co-doped into CZTSSe, the CBM and VBM shifted downward relative to the undoped sample, though the band gap remained smaller than that observed in the purely Ag-doped sample. This indicates that Na modifies the incorporation of Ag, thereby diminishing its perturbation of the CZTSSe electronic structure. These results suggest a coupled dopant interaction rather than the impact of independent effects. Compared to Ag-only doping, the CBM in the Na–Ag co-doped sample is slightly higher, which reduces the conduction band offset (CBO) at the CZTSSe/CdS interface and lowers the electron transport barrier. Additionally, the smaller gap between E_F and VBM relative to Ag doping implies either a higher carrier density or a further alleviation of Fermi level pinning. These factors collectively contribute to the improved V_OC and FF observed in Na–Ag co-doped devices.

Defect and carrier transport analysis

Capacitance–voltage (C–V) and drive-level capacitance profiling (DLCP) measurements were performed to analyze the carrier density and collection properties of the CZTSSe absorbers, as shown in Fig. 6. The derived charge densities, depletion widths (W_d), and defect densities within the depletion region are summarized in Table 2. Na doping increased the charge density from 6.15 × 10¹⁵ cm⁻³ to 1.10 × 10¹⁶ cm⁻³ while reducing the depletion width from 175 nm to 160 nm. In contrast, Ag doping decreased the charge density to 4.52 × 10¹⁵ cm⁻³ and expanded the depletion width to 190 nm. These results demonstrate that Na doping indeed enhances charge density and narrows the depletion region, whereas Ag doping reduces charge density and broadens the depletion width. Importantly, Na–Ag co-doping achieves both a high charge density (8.27 × 10¹⁵ cm⁻³) and an extended depletion width (223 nm), which leads to superior carrier separation and transport efficiency.

Fig. 6 (a) C–V – DLCP and (d) EIS characteristics of the undoped, Na doped, Ag doped, and Na–Ag co-doped CZTSSe devices. The extracted defect activation energies for the (b) reference, (c) Na doped, (e) Ag doped, and (f) Na–Ag co-doped CZTSSe devices deduced from the Arrhenius plots of the normalized inflection frequencies.

Table 2 Summary of carrier and defects properties derived from C–V, DLCP, EIS, and AS measurements

Device	N CV (cm^−3)	N DL (cm^−3)	Depletion width (nm)	Interface state response (relative values)	Effective lifetime (s)	E a (meV)
Ref	6.15 × 10^15	2.06 × 10^16	175	4.09 × 10^15	2.72 × 10^−4	198
Na	1.10 × 10^16	7.56 × 10^15	160	3.44 × 10^15	6.58 × 10^−4	106
Ag	4.52 × 10^15	1.31 × 10^15	190	3.21 × 10^15	6.10 × 10^−4	108
Na–Ag	8.27 × 10^15	5.73 × 10^15	223	2.53 × 10^15	8.31 × 10^−4	104

---

Furthermore, the density of interface trap states (N_it) decreased significantly from 4.09 × 10¹⁵ cm⁻³ to 2.53 × 10¹⁵ cm⁻³ with Na–Ag co-doping, indicating improved interface quality and suppressed recombination. This observation aligns with the earlier analysis. The effective carrier lifetime (τ), calculated from the relationship τ = R_ct × C_p (where R_ct is recombination resistance and C_p is the capacitance) via EIS measurements,³⁷ also exhibited enhancement. As depicted in Fig. 5b and Table 2, the Na–Ag co-doped samples exhibited a prolonged lifetime (8.3 × 10⁻⁴ s) compared to Na-doped (6.6 × 10⁻⁴ s) and Ag-doped (6.1 × 10⁻⁴ s) samples, reflecting a substantial improvement in carrier transport dynamics.

Admittance spectra (AS) was utilized to investigate the defect properties of CZTSSe devices after Na and Ag doping. The temperature-dependent capacitance–frequency profiles are shown in Fig. S7 and −f × dC/df curves for different doping strategies are shown in Fig. S8. Arrhenius plots of versus are used to extract the defect activation energy (E_a) (Fig. 6b, c, e and f) and the calculated defect properties are summarized in Table 2. E_a is approximately the energy difference between the VBM and the defect state and thus can be considered as the energies for carriers de-trapping from defect states to the valence band region.²⁹ For CZTSSe devices without doping, the E_a value is 198 meV, which significantly decreases to 106 meV when there is Na doping, 108 meV when there is Ag doping, and further decreases to 104 meV when Na–Ag co-doping is employed. The smaller values of E_a further confirm the passivation of Cu related defects after the substitution of Cu by Na and Ag, thus explaining the lower recombination rates and suppressed band tailing, which finally results in the improved V_OC and FF values of doped CZTSSe devices.

To further investigate the carrier properties of Na-doped and Na–Ag co-doped samples, we employed time-resolved emission microscopy (TREM) with 532 nm laser excitation to study the reference (Ref), Na-doped, and Na–Ag co-doped devices (Fig. 7; detailed operation is provided in Fig. S9 and the Experimental section). Like conventional time-resolved photoluminescence measurements (TRPL), TREM captures carrier recombination dynamics but offers additional advantages. Its ultrafast temporal resolution enables precise collection of intricate spatiotemporal information.

Fig. 7 Time-resolved emission microscopy (TREM) images of differently doped devices. The maps of the short decay times (τ₁) and long decay times (τ₂) for the undoped devices (a and e), Na doped devices (b and f), Na–Ag co-doped devices (c and g). Tau histograms of the corresponding lifetimes are depicted for (d) τ₁, and (h) τ₂ with red shadows for the undoped devices, blue shadows for the Na doped devices, and green shadows for the Na–Ag co-doped devices.

The derived carrier lifetimes are summarized in Fig. 7a–c and e–g. The fast decay component (τ₁) is primarily associated with interface recombination at the heterojunction, whereas the slow decay component (τ₂) reflects bulk recombination. The corresponding PL intensities are provided in Fig. S11. Although both Na and Ag doping reduce non-radiative recombination, Ag doping exhibits a significantly stronger effect. Besides, compared to the undoped reference, both the Na-doped and Na–Ag co-doped samples exhibited more uniform and prolonged τ₁ values. This indicates a reduction in interface recombination losses, likely attributable to the beneficial effects of doping on absorber morphology and electronic uniformity. Notably, the Na–Ag co-doped devices showed a significantly extended τ₂ compared to the Na-doped samples, suggesting enhanced bulk carrier transport properties. The prolonged bulk lifetime implies an increase in the carrier diffusion length, which improves charge collection efficiency and mitigates the detrimental influence of local inhomogeneities within the absorber. Such enhancement in both interfacial and bulk electronic properties collectively contributes to the superior power conversion efficiency observed in the co-doped devices.

Conclusion

The combination of Na uniform doping and Ag surface doping leads to better performance than single element doping and improves the performance of CZTSSe solar cells. The surface doped Ag helps Na and Se ions diffuse more rapidly through the absorber layer during selenization and significantly enhances the growth of CZTSSe crystals, resulting in a void-free and highly crystallized, bilayer absorber layer. Compared to single element doping, the high-quality bilayer absorber layer in Na–Ag co-doped devices enhances carrier collection. Moreover, co-doping suppresses nonradiative recombination, increases carrier lifetime (τ₂) and collection efficiency, thus significantly improving the V_OC and FF. This study has unraveled the separate roles of Na and Ag cations in promoting the growth of CZTSSe crystals and confirmed the potential of multi-cation doping to further boost the efficiency of CZTSSe solar cells, without causing excessive lattice distortion or decreases in carrier density. With this mechanistic understanding, the overall performance of CZTSSe solar cells has been enhanced and a PCE efficiency of 13.2% has been obtained.

Author contributions

W. C., W. Y. and Y. J.: conceptualization. W. C. and W. Y.: investigation, methodology. W. C., W. Y. and Y. J.: formal analysis. Y. X. and J. Y.: investigation (TREM characterization), formal analysis. Q. Z. and A. K.: investigation (sample preparation, UPS and XPS measurements), data curation. C. M. and P. Z.: investigation (device fabrication). P. Z. and A. C.: methodology, project administration. Y. J., F. L. and P. M.: writing – original draft, writing – review & editing. All authors were involved in project discussions (supervision/validation) and approved the final manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: additional experimental details, supplementary Fig. S1–S9 and supporting data analyses. See DOI: https://doi.org/10.1039/d5ta06324h.

Acknowledgements

W. C. and W. Y. contributed equally to this work. F. L. thanks funding through the National Key Research and Development Program of China (2024YFB4205000). Y. J. thanks funding through the Australian Government Research Training Program (RTP) Scholarship program. P. Z. thanks the support from the Natural Science Foundation of Xinjiang, China (Grant No. 2022D01D79). And P. M. wants to thank the support from the ARC Centre of Excellence in Exciton Science through ARC Grant CE170100026.

References

1. D. B. Mitzi, et al., The path towards a high-performance solution-processed kesterite solar cell, Sol. Energy Mater. Sol. Cells, 2011, 95(6), 1421–1436.
2. M. Ravindiran and C. Praveenkumar, Status review and the future prospects of CZTS based solar cell – A novel approach on the device structure and material modeling for CZTS based photovoltaic device, Renewable Sustainable Energy Rev., 2018, 94, 317–329.
3. X. Liu, et al., The current status and future prospects of kesterite solar cells: a brief review, Prog. Photovoltaics Res. Appl., 2016, 24(6), 879–898.
4. Y. Gong, et al., Elemental de-mixing-induced epitaxial kesterite/CdS interface enabling 13%-efficiency kesterite solar cells, Nat. Energy, 2022, 7, 966–977.
5. M. A. Green, et al., Solar Cell Efficiency Tables (Version 66), Prog. Photovoltaics Res. Appl., 2025, 33(7), 795–810.
6. J. Keller, et al., High-concentration silver alloying and steep back-contact gallium grading enabling copper indium gallium selenide solar cell with 23.6% efficiency, Nat. Energy, 2024, 9(4), 467–478.
7. W. Shockley and H. J. Queisser, Detailed Balance Limit of Efficiency of p-n Junction Solar Cells, J. Appl. Phys., 1961, 32(3), 510–519.
8. J. Zhou, et al., Control of the phase evolution of kesterite by tuning of the selenium partial pressure for solar cells with 13.8% certified efficiency, Nat. Energy, 2023, 8(5), 526–535.
9. Y. Gong, et al., Ag Incorporation with Controlled Grain Growth Enables 12.5% Efficient Kesterite Solar Cell with Open Circuit Voltage Reached 64.2% Shockley–Queisser Limit, Adv. Funct. Mater., 2021, 31(24), 2101927.
10. Z. Su, et al., Device Postannealing Enabling over 12% Efficient Solution-Processed Cu₂ZnSnS₄ Solar Cells with Cd²⁺ Substitution, Adv. Mater., 2020, 32(32), 2000121.
11. H. Guo, et al., Band-gap-graded Cu₂ZnSn(S,Se)₄ drives highly efficient solar cells, Energy Environ. Sci., 2022, 15(2), 693–704.
12. S. Wang, et al., Defects and Surface Electrical Property Transformation Induced by Elemental Interdiffusion at the p–n Heterojunction via High-Temperature Annealing, ACS Appl. Mater. Interfaces, 2021, 13(10), 12211–12220.
13. S. Hadke, et al., Effect of Cd on cation redistribution and order-disorder transition in Cu2(Zn,Cd)SnS4, J. Mater. Chem. A, 2019, 7(47), 26927–26933.
14. J. Wang, et al., Ge-bidirectional Diffusion to Simultaneously Engineer Back Interface and Bulk Defects in the Absorber for Efficient CZTSSe Solar Cells, Adv. Mater., 2022, 34, 2202858.
15. J. Zhou, et al., Regulating crystal growth via organic lithium salt additive for efficient Kesterite solar cells, Nano Energy, 2021, 89, 106405.
16. Z. Shen, et al., Li2S doping into CZTSe drives the large improvement of VOC of solar cell, J. Energy Chem., 2021, 62, 637–644.
17. X. Chang, et al., Synergistic incorporation of NaF and CsF PDT for high efficiency kesterite solar cells: unveiling of grain interior and grain boundary effects, J. Mater. Chem. A, 2021, 9(1), 413–422.
18. W. Yang, et al., The Multiple Roles of Na Ions in Highly Efficient CZTSSe Solar Cells, Small, 2024, e2307807.
19. Y. Gong, et al., Attaining 15.1% Efficiency in Cu2ZnSnS4 Solar Cells Under Indoor Conditions Through Sodium and Lithium Codoping, Sol. RRL, 2025, 9(4), 2400756.
20. H. Xu, et al., 11.88% Efficient Flexible Ag-Free CZTSSe Solar Cell: Spontaneously Tailoring the Alkali Metal Level, Small, 2024, 20(51), e2408122.
21. B. Duan, et al., Highly efficient solution-processed CZTSSe solar cells based on a convenient sodium-incorporated post-treatment method, J. Energy Chem., 2020, 40, 196–203.
22. K. J. Yang, et al., Sodium Effects on the Diffusion, Phase, and Defect Characteristics of Kesterite Solar Cells and Flexible Cu₂ZnSn(S,Se)₄ with Greater than 11% Efficiency, Adv. Funct. Mater., 2021, 31, 2102238.
23. B. Liu, et al., Effect of Na doping on the performance and the band alignment of CZTS/CdS thin film solar cell, Sol. Energy, 2020, 201, 219–226.
24. S. Grini, et al., Strong Interplay between Sodium and Oxygen in Kesterite Absorbers: Complex Formation, Incorporation, and Tailoring Depth Distributions, Adv. Energy Mater., 2019, 9(27), 1900740.
25. Y. Li, et al., Suppressing Deep-Level Trap Toward Over 13% Efficient Solution-Processed Kesterite Solar Cell, Small, 2024, 20(35), e2401330.
26. M. He, et al., High Efficiency Cu₂ZnSn(S,Se)₄ Solar Cells with Shallow Li_Zn Acceptor Defects Enabled by Solution-Based Li Post-Deposition Treatment, Adv. Energy Mater., 2021, 11(13), 2003783.
27. H. Guo, et al., An efficient Li+-doping strategy to optimize the band alignment of a Cu2ZnSn(S,Se)4/CdS interface by a Se&LiF co-selenization process, J. Mater. Chem. A, 2020, 8(42), 22065–22074.
28. C. Yan, et al., Cu₂ZnSnS₄ solar cells with over 10% power conversion efficiency enabled by heterojunction heat treatment, Nat. Energy, 2018, 3(9), 764–772.
29. X. Zhao, et al., Insights into the Efficiency Improvement for CZTSSe Solar Cells with over 12% Efficiency via Ga Incorporation, Adv. Funct. Mater., 2023, 2301377.
30. Y. Li, et al., Suppressing Element Inhomogeneity Enables 14.9% Efficiency CZTSSe Solar Cells, Adv. Mater., 2024, 36(25), 2400138.
31. T. Gershon, et al., Photovoltaic Materials and Devices Based on the Alloyed Kesterite Absorber (Ag_xCu_1-x)₂ZnSnSe₄, Adv. Energy Mater., 2016, 6(10), 1502468.
32. S. H. Hadke, et al., Synergistic Effects of Double Cation Substitution in Solution-Processed CZTS Solar Cells with over 10% Efficiency, Adv. Energy Mater., 2018, 8(32), 1802540.
33. X.-Y. Chen, et al., Ag, Ti dual-cation substitution in Cu2ZnSn(S,Se)4 induced growth promotion and defect suppression for high-efficiency solar cells, J. Mater. Chem. A, 2022, 10(42), 22791–22802.
34. M. He, et al., Systematic Efficiency Improvement for Cu2ZnSn(S,Se)4 Solar Cells By Double Cation Incorporation with Cd and Ge, Adv. Funct. Mater., 2021, 31(40), 2104528.
35. U. Farooq, et al., Defects passivation by solution-processed titanium doping strategy towards high efficiency kesterite solar cells, Chem. Eng. J., 2023, 451, 139109.
36. M. Abbas, et al., Highest Solar-to-Hydrogen Conversion Efficiency in Cu(2)ZnSnS(4) Photocathodes and Its Directly Unbiased Solar Seawater Splitting, Nanomicro Lett., 2025, 17(1), 257.
37. X. Liang, et al., Effects of Ag Incorporation on the Band Structures and Conductivity Types of (Cu_1-xAg_x)₂ZnSnS₄ Solid Solutions, ChemPhotoChem, 2018, 2(9), 811–817.
38. V. B. Rajkumar and S.-W. Chen, Ag-Se phase diagram calculation associatingab–initiomolecular dynamics simulation, Calphad, 2018, 63, 51–60.
39. S. Quadir, et al., Impact of Cation Substitution in (Ag_xCu_1−x)2ZnSnSe4Absorber-Based Solar Cells toward 10% Efficiency: Experimental and Theoretical Analyses, Sol. RRL, 2021, 5(10), 2100441.
40. Y. Qi, et al., Elemental Precursor Solution Processed (Cu(1-x)Ag(x))(2)ZnSn(S,Se)(4) Photovoltaic Devices with over 10% Efficiency, ACS Appl. Mater. Interfaces, 2017, 9(25), 21243–21250.

---

Footnote

† These authors contributed equally to this work.

---