Enhancing CZTSSe Kesterite Solar Absorbers with Radio‑Frequency Plasma Treatment

Abstract

We present a hydrazine‑free method that improves the electro‑optical and structural performance of kesterite light absorbers. By exposing Cu₂ZnSn(S,Se)₄ (CZTSSe) films to weak hydrogen plasma discharges powered at 13.56 MHz, we achieve more uniform samples, lower strain, and a modified band structure. Raman, infrared, and reflectance spectroscopy confirm a reduction in reflectivity and an altered electronic band gap after RF treatment.

Background

Energy generation and storage are critical as conventional fuels decline and economic demands rise. This drives innovation in light‑harvesting technologies, from silicon (Si) to III–V and organic photovoltaics. Thin‑film solar cells (TFSCs) based on the kesterite structure Cu₂ZnSn(S,Se)₄ (CZTSSe) are rapidly advancing due to their earth‑abundant, non‑toxic constituents and a direct band gap of ~1.5 eV with absorption coefficients >10⁴ cm⁻¹ in the visible range. The current record efficiency for a CZTSSe prototype is 12.6 % [6]. Key challenges include non‑stoichiometric composition, intrinsic defect concentration, phase coexistence, and secondary impurities. Conventional X‑ray diffraction (XRD) struggles to distinguish kesterite from stannite or defect‑modified phases because of similar diffraction patterns and the small difference in Cu/Zn scattering cross‑sections. Raman spectroscopy offers a more accessible alternative for detecting secondary phases and assessing structural homogeneity. Previous high‑efficiency CZTSSe devices employed hydrazine‑based growth and post‑growth annealing in N₂ and air, but hydrazine’s toxicity limits scalability. We therefore propose a hydrazine‑free post‑growth treatment using hydrogen plasma generated in an RF electromagnetic field, aiming to improve bulk and multilayered absorber properties.

Methods

RF Treatment on Silicon Cells
Initial tests applied a 13.56 MHz hydrogen plasma to Si solar cells. The test area was 2 cm² and the structure included Al front grid, 50 nm Si₃N₄ antireflection, 30 nm SiO₂, n⁺⁺ and n⁺ diffusion layers, a p‑Si base, p⁺ junction, and backside Al metallization. Samples were divided into reference, indoor, and outdoor groups. Masking prevented etching of antireflection coatings. Treatments used an inert gas medium and varied exposure times (1–15 min) and powers (0.19–2.25 W cm⁻²). The chamber operated at 0.2 Torr H₂ pressure with a substrate voltage of 1900 V at room temperature. Pre‑cleaning employed a 150 W, 50 kHz N₂ plasma (PlasmaEtch PE‑50 XL). Photovoltaic parameters were measured with a Kelvin probe and LabTraser NI software, and analyzed via a double‑diode model [10]. RF Treatment on CZTSSe Absorbers
Optimal RF regimes (0.8 W cm⁻² for 15 min) were then applied to three types of bulk CZTSSe samples: (i) ZnS/CuS/SnS precursors deposited by flash evaporation on glass with Mo bottom layers; (ii) Bridgman‑grown crystals sputtered onto glass (with/without Mo) at varied substrate temperatures; and (iii) electron‑beam evaporated films for device fabrication. Transmission and specular reflectance were measured across 500–5000 cm⁻¹ using FTIR (Infralum FT‑801) with configurations A (ATR), B_s/B_d (specular/diffuse), C (diffuse), and D (normal incidence). Absorption spectra were derived from reflectance via dispersion integrals [11,12]. Micro‑Raman spectroscopy (Horiba T64000, 514.5 nm Ar⁺ laser, 0.1 mW µm⁻²) provided structural insights; spectra were recorded in <1 min per spot with a 50× objective, averaged across multiple sites for uniformity.

Results and Discussion

Impact on Silicon Solar Cells
Figure 1 (not shown) illustrates AM1.5 IU characteristics of Si cells after RF treatments. Baseline efficiency was 11.69 % (FF 0.746). After 95 W exposure, efficiency rose to 12.34 % (FF 0.775); 225 W yielded 12.29 % (FF 0.783); 300 W reduced efficiency to 11.46 % (FF 0.752), likely due to contact cracking. The increase in short‑circuit current and modest decrease in open‑circuit voltage suggest passivation of dangling bonds by reactive H atoms. These results confirm that RF plasma can modify device parameters, though optimization is needed for thin‑film cells. Optical Properties of CZTSSe
Reflectance spectra before and after RF treatment (Figure 2a) show a notable decrease in reflectivity between 1.2–3 eV for multilayered samples and 2.4–3.3 eV for bulk samples. The differences arise from Schottky contacts in layered structures versus free bulk samples. Using normal‑incidence specular reflection (configuration D) provided the most reliable refractive index and extinction coefficient data; ATR and integrating‑sphere setups introduced significant phase‑shift errors. Transmission–reflection measurements on large‑area bulk CZTS films (Figure 2b) revealed maximal reflectivity reduction after 3 min of RF exposure. Derived absorption coefficients (Figure 2c) increased across the band gap, indicating enhanced light‑absorption. Drude analysis yielded a slight increase in plasma frequency from 2.278 eV to 2.294 eV, implying improved carrier concentration. Chemical Composition by FTIR
Figure 3 compares FTIR spectra of bulk CZTS with and without RF treatment. Across 500–4000 cm⁻¹, the treated sample shows reduced absorption of C–N (1250/1600 cm⁻¹), C–C/C=C (1490–1650 cm⁻¹), CH_n (2870–3100 cm⁻¹), CO₂ (2350 cm⁻¹), and water/organic (2700/3600 cm⁻¹). The suppression of sp² hybridized bonds confirms removal of graphite‑like residues by H⁺ ions. Raman Spectroscopy
Bulk CZTS Raman spectra (Figure 4) exhibit dominant peaks at 286 and 335 cm⁻¹ (A and B symmetries) and weaker features at 251, 305, 343, and 356 cm⁻¹. A small peak near 329 cm⁻¹ indicates Zn/Cu anti‑site disorder. The intensity ratio I₃₂₉/I₃₃₅ = 0.11 aligns with previous reports for thin films. Post‑RF spectra (Figure 5a) show a 2 cm⁻¹ shift of the 286 cm⁻¹ band to higher frequency, halved full‑width at half maximum (22 cm⁻¹), and increased intensity, signifying lattice ordering. The 335 cm⁻¹ A‑symmetry peak also shifts by 2 cm⁻¹ without width change. The I₃₃₁/I₃₃₇ ratio decreases to 0.06, reinforcing reduced disorder. These changes persist for at least one month in air. For CZTSe thin films (Figure 5b), the main Raman peaks at 193 and 176 cm⁻¹ shift by +2 cm⁻¹ after RF exposure, indicating a similar reduction in structural defects.

Conclusions

Hydrogen‑based RF plasma treatment (13.56 MHz, 0.8 W cm⁻², 15 min) effectively improves the optical and structural properties of both bulk and thin‑film CZTSSe absorbers. Raman analysis shows a 2 cm⁻¹ shift and narrowing of key vibrational modes, evidencing lattice ordering. FTIR confirms removal of carbon‑based impurities and spⁿ hybridized species. Reflectance–transmittance data converted to absorption spectra reveal enhanced absorption and modest increases in carrier concentration. These results demonstrate that a hydrazine‑free, vacuum‑compatible post‑processing step can reduce internal strain and improve device‑level performance, making it suitable for scalable production of multilayered kesterite solar cells.

Abbreviations

CZTS

Cu₂ZnSnS₄

CZTSe

Cu₂ZnSnSe₄

CZTSSe

Cu₂ZnSn(S,Se)₄

FTIR

Fourier Transform Infrared Spectroscopy

IR

Infrared

RF

Radio Frequency

SCs

Solar Cells

TFSCs

Thin‑Film Solar Cells

XRD

X‑ray Diffraction