Stable blue emission and sub-gap mediated UV photodetection in lead-free 1D CsCu₂Br₃ single crystals

Abstract

In this work, we report a comprehensive study of the structural, optical, and photoelectrical properties of one-dimensional (1D) CsCu₂Br₃ single crystals. Powder X-ray diffraction (XRD) confirms the orthorhombic Cmcm phase with excellent crystallinity, while thermogravimetric analysis (TGA) demonstrates thermal stability up to ~ 300 °C, with only a minor (~ 6%) extrinsic mass loss near 287 °C. Optical characterization reveals a direct bandgap of 4.09 eV, a low Urbach energy (0.104 eV), and strong blue photoluminescence (451.6 nm) with a large Stokes shift and unusually broad full width at half maximum (FWHM) (~ 250 nm), indicative of pronounced electron–phonon coupling and self-trapped excitons (STEs). Time-resolved PL yields two recombination channels (τ₁ = 10.07 ns, τ₂ = 2.27 ns), consistent with free/shallow carriers and stabilized STEs. Electrical measurements show symmetric Current–voltage (I–V) behavior and UV-enhanced conduction dominated at high fields by a trap-limited space-charge-limited current (SCLC) mechanism (m ≈ 2.45, V_TFL ≈ 13.2 V). Under 365 nm excitation (photon energy below E_g), the device exhibits a reproducible but modest photoconductive response, mediated by sub-gap absorption pathways—namely Urbach-tail states, defect-related channels, and STE manifolds—rather than direct interband excitation. Accordingly, the responsivity at 365 nm is modest (R ≈ 4.38 × 10⁻⁵ A W⁻¹), requires a higher bias (+ 20 V), and shows trap-limited SCLC signatures, fully consistent with a weak sub-gap absorption regime. Collectively, the structural robustness, stable excitonic blue emission, and sub-gap mediated UV photoresponse position CsCu₂Br₃ as a durable lead-free platform for UV optoelectronics, with maximum efficiency expected under deep-UV excitation (≈ 280–320 nm), to be mapped in future calibrated spectral responsivity studies.

Introduction

The continuous quest for advanced materials with enhanced optoelectronic performance and environmental sustainability has propelled the development of new classes of lead-free halide perovskites^1,2,3,4. Traditional perovskite materials, while lauded for their exceptional carrier mobility, tunable bandgap, and strong light-matter interaction, have been hampered by toxicity issues, structural brittleness, and limited long-term stability—challenges that restrict their integration into next-generation electronic and photonic technologies^5,6,7,8. In this context, copper-based halide perovskites have recently emerged as promising alternatives, providing a platform that couples robust inorganic frameworks with low-dimensional architectures and benign elemental compositions^{9,10,11,12,13}.

Among these, 1D CsCu₂Br₃ has garnered significant attention owing to its unique crystallographic topology and distinctive optoelectronic behavior^14,15. The 1D chain structure, composed of edge-sharing [CuBr₄]³⁻ tetrahedra embedded in a Cs⁺ matrix, introduces pronounced anisotropy in charge transport and polarization properties, opening new avenues for tailored device functionalities^16,17,18. This peculiar arrangement facilitates not only efficient carrier dynamics but also stabilizes self-trapped excitonic states, which are pivotal for strong photoluminescence and rapid photoresponse. Importantly, the inherent wide bandgap of CsCu₂Br₃, coupled with its high Curie temperature and spontaneous ferroelectric polarization, endows the material with a suite of attributes desirable for UV photodetection, memory devices, and integrated logic circuits^16,17,18.

The drive toward environmentally conscious and multifunctional materials has also intensified interest in CsCu₂Br₃ as a viable candidate for applications where both device performance and ecological impact are paramount. Recent advances have demonstrated the realization of polarization-sensitive UV detectors and resistive switching memory devices based on CsCu₂Br₃, featuring high responsivity, fast switching speeds, and long-term operational stability under ambient conditions. Furthermore, the absence of toxic elements positions CsCu₂Br₃ as an attractive platform for sustainable electronics and optoelectronics^16,17,18.

Despite these advances, a comprehensive understanding of the interplay between crystal structure, excitonic phenomena, and photoelectrical behavior in CsCu₂Br₃ remains underexplored. Fundamental questions persist regarding the impact of dimensionality on carrier dynamics, the stability of self-trapped states at elevated temperatures, and the mechanisms underpinning resistive switching and persistent photoconductivity. Addressing these knowledge gaps is essential not only for optimizing device performance but also for guiding the rational design of next-generation lead-free perovskite materials.

In this work, we present a comprehensive investigation of high-quality CsCu₂Br₃ single crystals, combining structural, optical, and photoelectrical characterizations. Powder XRD and TGA confirm their phase purity and excellent thermal stability up to ~ 300 °C. Optical spectroscopy reveals a wide bandgap (4.09 eV), low Urbach energy, strong self-trapped excitonic emission, and nanosecond-scale carrier lifetimes. Photoelectrical studies demonstrate symmetric I–V behavior, reproducible photoconductive switching under UV illumination, limited current drift under continuous operation, and a trap-limited SCLC regime at high fields. Noise analysis further identifies shot noise as the dominant contribution, yielding detectivity values on the order of 10⁸–10⁹ Jones. By correlating these crystallographic, photophysical, and electronic attributes, this study advances the understanding of 1D halide perovskites and highlights their potential as stable, lead-free materials for next-generation ultraviolet optoelectronic applications.

Experimental section

Synthesis

The synthesis of CsCu₂Br₃ single crystals was achieved through an optimized method employing anti-solvent crystallization. Specifically, 2.128 g of CsBr and 2.86 g of CuBr were weighed and then combined in a solvent mixture consisting of 2 mL of dimethyl sulfoxide (DMSO) and 8 mL of N,N-dimethylformamide (DMF). This mixture was stirred magnetically at 60 °C until complete dissolution and clarity of the solution were observed. Next, anhydrous methanol was gradually added until a steady cloudiness indicated the formation of a supersaturated state. The resulting mixture was immediately filtered, and the clear solution was collected in a separate container for crystallization. Subsequently, 10 μL of this solution was deposited onto a pre-cleaned quartz substrate measuring 3 × 2 cm². A second, identical substrate was carefully placed on top to create a sandwich structure. This assembly was placed into an airtight environment, where anhydrous methanol acted as the anti-solvent in a reservoir. The setup was maintained at 60 °C for a full day to ensure controlled solvent exchange, which directed the anisotropic growth of CsCu₂Br₃ single crystals (see the inset of Fig. 1a). All reagents used throughout this process were sourced from Sigma-Aldrich and employed as received, without additional purification.

Fig. 1

(a) Powder X-ray diffraction (PXRD) pattern of CsCu₂Br₃ single crystals with standard diffraction positions (red ticks, ICSD No. 49613; mp-23017). The data confirm the orthorhombic Cmcm phase without secondary phases. Inset: image of a CsCu₂Br₃ single crystal, (b) Crystallographic structure of CsCu₂Br₃ single crystals.

Device characterization

XRD analysis was performed at room temperature using a Rigaku MiniFlex 600 diffractometer. The diffraction patterns were acquired over a 2θ interval from 10° to 50°, with a step increment of 0.05°, employing copper Kα radiation (λ = 1.54 Å) to accurately resolve the crystalline phases.

To evaluate the thermal stability of the CsCu₂Br₃ samples, TGA was conducted on a Perkin Elmer Pyris 6 analyzer. Approximately 3 mg of the powdered sample was subjected to heating from ambient temperature up to 800 °C under normal atmospheric conditions.

For optical characterization, ultraviolet–visible (UV–Vis) absorption measurements were obtained using a UV-2550 spectrophotometer fitted with an integrating sphere, enabling precise determination of both absorbance and diffuse reflectance spectra.

Photoluminescence (PL) and photoluminescence excitation (PLE) spectra were acquired using a Shimadzu RF-1501 spectrofluorometer, enabling comprehensive analysis of both the emission and excitation properties of the samples. In addition to the measurements at 365 nm, the device was also preliminarily tested under illumination at 280 nm, close to the maximum observed in the PLE spectrum (285.8 nm). However, the responsivity values could not be quantitatively compared in the present work due to the lack of a calibrated deep-UV source with the required power stability in our setup. The qualitative response confirmed that the device is indeed sensitive in this spectral region, which is consistent with the strong absorption and PLE peak near 280 nm. The wavelength of 365 nm was selected for the present photoresponse characterization because it corresponds to a commercially available, power-stable UV-A LED in our laboratory setup, enabling reproducible and low-noise measurements. Importantly, this excitation does not probe direct interband transitions (E_g ≈ 4.09 eV ≈ 303 nm) but tests the material’s sub-gap sensitivity via Urbach-tail, defect-related, and self-trapped exciton (STE)-mediated absorption pathways. The comparatively weaker absorption at 365 nm inherently results in lower photogenerated carrier density, which in turn requires a higher applied bias (+ 20 V) to ensure measurable and stable photocurrent. The use of such a bias is also linked to the relatively high resistivity of the CsCu₂Br₃ crystal and the large electrode spacing (~ 5 mm) in the present prototype. Future optimization of the excitation wavelength (e.g., 280–320 nm) and electrode geometry (e.g., reduced spacing or interdigitated electrodes) is expected to significantly lower the operating voltage. Comprehensive wavelength-dependent responsivity measurements, including a full comparison between 280 nm, the near-edge/absorption-tail region around 300–320 nm, and 365 nm, are planned for future work once the appropriate calibrated UV source is available. For these measurements, the samples will be excited across the 250–400 nm spectral range to identify the optimal excitation wavelength and determine the full spectral responsivity profile. For PL measurements, the samples were excited in the ultraviolet region, specifically at 320 nm. This wavelength lies in the near-edge/absorption-tail region and overlaps with the strong sub-gap absorption of CsCu₂Br₃. It was selected because of source availability, and while it does not correspond to direct interband excitation, it efficiently generates carriers for PL analysis. We note that this choice of 320 nm as the excitation wavelength in the PL measurements does not contradict the PLE spectrum of Fig. 5a, which shows a maximum near 286 nm. The PLE maximum represents the most efficient excitation energy within the deep-UV range, while 320 nm provides a practical near-edge excitation point. Both observations are consistent and complementary : the absorption edge provides a practical excitation point for PL characterization, while the PLE spectrum highlights the intrinsic excitation resonance. Emission spectra were recorded across a broad wavelength range to capture all relevant photoluminescence features, while PLE spectra were collected by monitoring the emission at the PL maximum as the excitation wavelength was varied, providing further insights into the optical transitions and energy levels within the material.

To investigate the temperature dependence of the emission properties, temperature-dependent PL measurements were performed using a closed-cycle helium cryostat, allowing precise control of the sample temperature from 80 to 300 K. At each temperature step, PL spectra were systematically recorded under identical excitation and detection conditions. This approach enabled the evaluation of the evolution of emission intensity, spectral profile, and thermal quenching behavior, offering a detailed understanding of the excitonic processes and recombination dynamics in CsCu₂Br₃ single crystals.

Time-resolved photoluminescence (TRPL) experiments were performed using a FLOROCUBE setup, where a pulsed diode laser operating at 320 nm served as the excitation source. The temporal evolution of the PL signal was detected using a TBX-04-D photodetector, and the data were processed through time-correlated single-photon counting (TCSPC) electronics. This configuration allowed for precise measurement of photoluminescence decay dynamics and accurate determination of carrier lifetimes. All measurements were conducted at room temperature, and the obtained lifetime values were analyzed using a multi-exponential fitting model to distinguish between fast and slow recombination processes present in the material. The choice of 320 nm as the excitation wavelength for TRPL experiments was dictated by the availability of a stable pulsed diode laser in our setup. While this wavelength lies slightly below the main PLE maximum (~ 286 nm), it overlaps significantly with the absorption-tail region of CsCu₂Br₃ and efficiently generates photocarriers for lifetime analysis, without implying direct interband excitation. Using this stable and well-characterized source ensured reproducibility of the decay measurements. Similarly, the 365 nm LED employed for photoresponse tests was selected because it corresponds to a commercially available, power-stable UV source widely adopted in photodetector benchmarking. Although its photon energy is below the bandgap, the observed response arises from tail/defect/STE-mediated pathways, as detailed in the Discussion. Full spectral responsivity and decay studies using calibrated deep-UV sources (280–300 nm) are planned for future work.

Single-crystalline CsCu₂Br₃ samples were first mechanically cleaved to obtain a rectangular crystal with typical dimensions of approximately 6 mm × 2 mm × 1 mm (see the inset of Fig. 1a). The crystal was mounted onto an optically transparent quartz substrate to ensure mechanical stability and allow back-illumination during optical measurements. Two gold electrodes, each approximately 0.5 mm wide, were thermally evaporated (thickness ≈ 100 nm, monitored using a quartz crystal thickness sensor) onto the two opposite ends of the crystal, forming a planar electrode configuration. The electrode spacing was approximately 5 mm, resulting in an active detection area of ~ 10 mm². Electrical contacts to the electrodes were made using fine gold wires attached with conductive silver paste to minimize contact resistance (see the inset of Fig. S1). A schematic diagram of the device architecture and an optical micrograph of the fabricated photodetector are provided in the Supplementary Information (Fig. S1) for clarity. This two-probe configuration was chosen for its simplicity and reliability in characterizing single-crystal photodetectors. After fabrication, the device was stored under inert conditions to prevent degradation from ambient moisture or oxygen.

I–V and photoresponse measurements were carried out at room temperature using a Keithley 2636B source-measure unit in a two-probe configuration. The CsCu₂Br₃ single crystal devices were characterized under both dark and UV-illuminated conditions (λ = 365 nm, light intensity ≈ 5 mW m⁻², as measured at the sample position using a calibrated optical power meter), in ambient atmosphere. The bias voltage was swept from − 30 V to + 30 V with a fixed step size of 0.3 V. The photocurrent was extracted by subtracting the dark current from the total current measured under illumination. The + 20 V bias was selected as a compromise between achieving a high signal-to-noise ratio and maintaining device stability during repeated ON/OFF cycling.

For dynamic photoresponse measurements, the devices were subjected to pulsed UV illumination using a programmable LED driver, with illumination cycles of 10 s ON and 10 s OFF at constant bias voltage (+ 20 V). The temporal evolution of the photocurrent was recorded using a high-resolution digital acquisition system with a time resolution of 0.1 s. All measurements were performed at room temperature and repeated over several cycles to ensure reproducibility and eliminate any transient or fatigue-related effects. In addition, to assess the material stability under continuous ultraviolet exposure, the fabricated CsCu₂Br₃ single-crystal photodetectors were subjected to prolonged illumination using a 365 nm UV LED source (≈5 mW cm⁻²) at a constant bias of + 20 V. The photocurrent was monitored over an extended period of operation (up to 1 h) in ambient conditions. The results indicated only a marginal decrease in the photocurrent intensity, with no significant degradation of the device performance during the test duration. This stability is attributed to the robust 1D [CuBr₄]³⁻ chain structure of CsCu₂Br₃, which is less prone to rapid photoinduced decomposition compared to conventional three-dimensional lead-halide perovskites. These observations suggest that CsCu₂Br₃ single-crystal devices can maintain reliable operation under continuous UV irradiation, making them promising candidates for stable UV photodetection applications. In the present work, the photoresponse characterization was performed at a fixed UV light intensity of approximately 5 mW cm⁻², chosen to ensure stable illumination and minimize any thermal effects on the CsCu₂Br₃ single-crystal device. A systematic investigation of the device performance under varying light intensities, which would allow determining the maximum light density the material can effectively handle without degradation, was not carried out due to the limitations of the current setup, which does not include a calibrated variable-intensity UV source. Such measurements are planned for future work to provide a complete analysis of the device’s dynamic range and saturation behavior.

Results and discussion

Figure 1a presents the powder XRD pattern of the synthesized CsCu₂Br₃ single crystals recorded at room temperature. The diffraction peaks were indexed using CELREF3 software and further validated by comparison with the theoretical powder XRD pattern simulated from the crystallographic information file (CIF) of CsCu₂Br₃ obtained from the Materials Project database (mp-23017; ICSD No. 49613). The excellent agreement between the experimental and simulated profiles confirms that the crystals adopt an orthorhombic structure within the Cmcm space group, with no detectable secondary phases. The refinement details, including refined lattice parameters, agreement factors, and atomic coordinates, are provided in the Supplementary Information (Fig. S2 and Table S1). The refined lattice parameters are a = 8.2104 Å, b = 8.1992 Å, c = 5.6213 Å, α = 90°, β = 90°, and γ = 104.8°, in good agreement with previously reported structural data for this compound^16,17,18.

In Fig. 1b, the crystallographic structure of CsCu₂Br₃ is illustrated, modeled using VESTA software based on structural data from previous reports. The atoms are labeled and color-coded for clarity, with Cs, Cu, and Br atoms represented in distinct colors. CsCu₂Br₃ adopts a 1D architecture composed of edge-sharing [CuBr₄]³⁻ tetrahedra forming double chains along the crystallographic c-axis. These chains are arranged in a staggered fashion and are embedded within a framework of Cs⁺ cations, which provide charge balance and structural stabilization. Each Cu⁺ ion is tetrahedrally coordinated by four Br⁻ anions, giving rise to slightly distorted [CuBr₄]³⁻ units. The double-chain configuration results from the corner-sharing of tetrahedra through Br atoms, producing a linear yet structurally constrained arrangement. This unique 1D connectivity distinguishes CsCu₂Br₃ from typical three-dimensional perovskites and strongly influences its optical and electronic properties^16,17,18.

The electronic structure of orthorhombic CsCu₂Br₃ has been previously investigated using density functional theory (DFT) calculations¹⁷. These studies reveal a direct bandgap of approximately 4.1 eV, with the valence band maximum (VBM) mainly derived from Cu 3d and Br 4p orbitals, and the conduction band minimum (CBM) dominated by Cu 4 s and Br 4p states¹⁶. Such an orbital distribution indicates strong Cu–Br electronic interactions along the 1D [CuBr₄]³⁻ chains, consistent with our experimental optical bandgap determined from UV–Vis absorption measurements. This combination of a wide bandgap and directional electronic coupling is favorable for ultraviolet optoelectronic applications.

Figure 2 displays the thermogravimetric (TG) curve of the synthesized CsCu₂Br₃ single crystals. As evident from the profile, the material demonstrates excellent thermal robustness, with no observable weight loss up to approximately 300 °C. This behavior indicates that the compound maintains its structural and chemical integrity under elevated temperatures, affirming its potential for stable operation in thermally demanding environments¹⁹. A minor step (~ 6% mass decrease) is observed around 287 °C, which can be attributed to the release of residual adsorbed species or partial volatilization of bromine-containing fragments. Similar features have been reported in the literature for CsCu₂Br₃ and related copper halides¹⁷, confirming that such small losses are extrinsic rather than indicative of bulk decomposition. Importantly, the overall mass remains nearly constant up to ~ 300 °C, supporting the excellent thermal stability of the compound.

Fig. 2

(a) STEM micrograph of CsCu₂Br₃ powder. (b) Size distribution of CsCu₂Br₃ particles. (c) STEM-EDS elemental mapping of CsCu₂Br₃.

The structural features and particle morphology of CsCu₂Br₃ powder were examined using scanning transmission electron microscopy (STEM). As shown in Fig. 3a, the micrographs were analyzed with ImageJ software to evaluate the particle size distribution. From the STEM data in Fig. 3b, particle size histograms were generated, revealing an average particle diameter of 2.14 ± 0.65 μm.

Fig. 3

Thermogravimetric analysis (TGA) curve of CsCu₂Br₃ single crystals.

To assess the elemental composition, energy-dispersive X-ray spectroscopy (EDS) was performed in conjunction with STEM, as illustrated in Fig. 3c. The elemental mapping confirms the presence and homogeneous distribution of Cs, Cu, and Br, with no detectable elemental loss during synthesis. These results confirm the successful preparation of a pure double perovskite phase, consistent with the quantitative data summarized in Table S2.

The theoretical elemental ratio for CsCu₂Br₃, based on its stoichiometric formula (1 Cs atom + 2 Cu atoms + 3 Br atoms = 100%), is :

 → Cs: 16.05% (≈16%).

 → Cu: 32.98% (≈33%).

 → Br: 50.97% (≈51%).

Figure 4a presents the optical absorption spectrum of the synthesized CsCu₂Br₃ single crystals, measured at ambient temperature across the spectral range of 200–600 nm. The absorption profile exhibits a steep rise in the ultraviolet region, suggesting a strong photon absorption near the band edge. To quantitatively determine the optical bandgap energy, the data were analyzed using the Tauc method²⁰, which models the absorption coefficient near the absorption edge based on the nature of electronic transitions. For direct allowed transitions, the Tauc relation is expressed as^21,22 :

$${\left(\alpha h\nu \right)}^{2}= \beta (h\nu -{E}_{g})$$

(1)

where α is the absorption coefficient, hν is the photon energy, E_g denotes the optical bandgap, and β is a material-dependent constant. The inset of Fig. 4a displays the Tauc plot, wherein the extrapolation of the linear segment to the energy axis yields a direct optical bandgap of 4.09 eV. This relatively wide bandgap reflects the large energy separation between the valence and conduction bands in CsCu₂Br₃, consistent with its low-dimensional crystal framework^23,24.

Fig. 4

(a) Optical absorption spectrum of CsCu₂Br₃ with Tauc plot inset, (b) Urbach energy analysis for CsCu₂Br₃.

To further evaluate the degree of structural disorder and the density of localized states within the forbidden band, the Urbach energy (E_u) was extracted. The Urbach tail describes the exponential absorption near the band edge, typically originating from lattice imperfections, phonon interactions, or structural inhomogeneities. The absorption coefficient in this region follows the Urbach empirical relation^25,26:

$$\alpha = {\alpha }_{0}exp\left(\frac{h\nu }{{E}_{u}}\right)$$

(2)

Taking the natural logarithm, the equation becomes linear²⁶ :

$$ln \left(\alpha \right)=\mathit{ln}\left( {\alpha }_{0}\right)+\frac{1}{{E}_{u} } h\nu$$

(3)

Figure 4b depicts the plot of ln(α) versus hν, where the slope of the linear region enables the determination of E_u. From this analysis, an Urbach energy of 0.104 eV was obtained, indicating a moderate density of tail states and a reasonably well-ordered crystalline phase. This low level of energetic disorder supports the material’s suitability for optoelectronic applications, where sharp absorption edges and minimal sub-gap states are desirable²⁷. It is important to note that the Tauc analysis (E_g ≈ 4.09 eV, ~ 303 nm) together with the extracted Urbach energy (E_u ≈ 0.104 eV) clearly indicates the presence of sub-gap absorption pathways. These include band-tail states associated with lattice disorder, phonon interactions, and possible shallow defects. Such features extend the absorption profile of CsCu₂Br₃ slightly into the near-UV region beyond the fundamental edge, and are expected to contribute to measurable photoresponse even at excitation energies below E_g, such as at 365 nm.

Figure 5a displays both the PLE and PL spectra of the CsCu₂Br₃ single crystals, recorded at room temperature. The PLE spectrum reveals a pronounced excitation peak centered at 285.8 nm, indicating the most efficient excitation energy for promoting electron transitions across the optical bandgap¹⁷. This value lies well within the high-energy ultraviolet region, consistent with the wide bandgap inferred from the UV–Vis absorption analysis.

Fig. 5

(a) Photoluminescence (PL) and photoluminescence excitation (PLE) spectra of CsCu₂Br₃, (b) Time-resolved photoluminescence (TRPL) decay curve of CsCu₂Br₃. The experimental decay (symbols) is fitted with a bi-exponential model (solid line). The first column of the inset table lists the best-fit parameters (A₁, A₂, τ₁, τ₂), while the second column corresponds to their associated fitting errors (standard errors obtained from the regression).

Under excitation at this optimal wavelength, the PL spectrum exhibits a dominant emission peak at 451.6 nm with a FWHM of 250.2 nm (~ 1476.9 meV), positioned in the blue region of the visible spectrum. In addition to its spectral position, this unusually large FWHM is indicative of strong electron–phonon coupling. This broadening is consistent with the pronounced Jahn–Teller distortions discussed earlier, which promote the formation of STEs localized within the [CuBr₄]³⁻ tetrahedral units. The distribution of slightly different local lattice configurations created by these distortions results in a spread of recombination energies, thereby widening the emission profile. Such a broad spectral bandwidth is characteristic of low-dimensional copper halide systems and reflects a robust excitonic emission that remains stable over a wide range of experimental conditions. Although in-situ XRD measurements at elevated temperatures were not performed in this study, previous reports on CsCu₂Br₃ and related copper halide perovskites indicate that these compounds retain their orthorhombic crystal structure up to temperatures exceeding 400 K, with no significant peak shifts or new phase formation observed upon moderate thermal cycling¹⁷. This inherent structural stability, combined with the persistence of broad excitonic emission, suggests that CsCu₂Br₃ is well-suited for optoelectronic applications operating under varying thermal conditions. Regarding the photoluminescence quantum yield (PLQY), the absence of an integrating sphere setup in our laboratory precluded its determination in the present work. Consequently, it was not possible to evaluate changes in PLQY after heating the sample to 300 K and subsequent cooling. These measurements will be included in future studies once the required instrumentation becomes available. This emission is attributed to radiative recombination processes involving localized excitonic states, likely arising from the unique low-dimensional architecture of CsCu₂Br₃. The observed Stokes shift of 165.8 nm (~ 1592.9 meV) between the excitation and emission peaks suggests the presence of significant electron–phonon coupling, and possibly exciton self-trapping effects within the [CuBr₄]³⁻ tetrahedral chains. All extracted spectral metrics (peak position, FWHM, Stokes shift in nm/meV) are summarized in Supplementary Table S3.

To further investigate the temporal characteristics of the photogenerated carriers, TRPL measurements were performed, as illustrated in Fig. 5b. The PL decay profile was well fitted using a bi-exponential function, yielding two distinct lifetime components: τ₁ = 10.07 ns and τ₂ = 2.27 ns, with a corresponding average lifetime τ_avg = 2.79 ns. The fast component (τ₂) is typically associated with radiative recombination of free excitons or shallowly trapped carriers, while the slower component (τ₁) reflects recombination from deeper trap states or more localized STEs. The fitted parameters (A₁, A₂, τ₁, τ₂) and τ_avg are listed in Supplementary Table S3.

In Cu-based low-dimensional halide systems, the presence of Cu⁺ ions with a 3d¹⁰ electronic configuration makes the lattice prone to local Jahn–Teller (JT) distortions upon photoexcitation. These distortions can induce strong electron–phonon coupling, promoting the formation of STEs within the [CuBr₄]³⁻ tetrahedral units. In our case, the longer-lived PL decay component (τ₁) can be attributed to the radiative recombination of such STEs. The JT-driven lattice relaxation stabilizes these excitonic states, leading to their spatial localization and extended lifetime compared to the fast component associated with free or shallowly trapped excitons. Furthermore, this localization effect contributes to the observed spectral broadening, as phonon-assisted recombination from a distribution of slightly different local potential wells results in a wider emission profile. This behavior is consistent with reports on other Cu-based perovskite derivatives, where JT distortions play a central role in defining the photophysical response. The broad PL bandwidth (FWHM ≈ 250 nm) and the biexponential decay dynamics provide evidence of self-trapped excitons (STEs) stabilized by Jahn–Teller distortions in the [CuBr₄]³⁻ tetrahedra. These localized states can act as recombination centers and also enable sub-gap optical absorption. Therefore, in addition to Urbach-tail states and shallow defects, STE-related manifolds represent another plausible channel allowing 365 nm photons (3.40 eV) to induce detectable photoconductivity in CsCu₂Br₃.

These photophysical insights highlight the role of the material’s 1D chain-like structure in governing both its absorption and emission behavior. The combination of strong UV absorption, visible-range luminescence, broad FWHM emission, large Stokes shift, and nanosecond-scale carrier lifetimes positions CsCu₂Br₃ as a promising candidate for light-emitting and photodetection devices operating in the UV–visible domain.

Figure 6a presents the temperature-dependent PL spectra of CsCu₂Br₃ single crystals, recorded across the range 80–300 K. At low temperature (80 K), the PL emission is markedly intensified and displays a noticeably narrower spectral profile compared to higher temperatures. This pronounced enhancement reflects the effective suppression of non-radiative recombination processes at cryogenic conditions, as well as the stabilization of self-trapped excitonic states within the low-dimensional crystal lattice^28,29. Upon increasing the temperature, a progressive reduction in PL intensity is observed, accompanied by a modest broadening of the emission band. Such thermal quenching behavior is typical for halide perovskites with strong exciton–phonon coupling, and is primarily attributed to the increased probability of phonon-assisted detrapping and non-radiative relaxation channels that become accessible at elevated temperatures^30,31.

Fig. 6

(a) Temperature-dependent photoluminescence (PL) spectra of CsCu₂Br₃, (b) Arrhenius model fitting of integrated PL intensity.

To quantitatively assess the robustness of the excitonic emission, the integrated PL intensity was plotted as a function of the inverse temperature (1/T) and analyzed using an Arrhenius-type model, as depicted in Fig. 6b. The thermal quenching process can be described by the following equation³² :

$$I\left(T\right)={I}_{0}/\left[1+A exp\left(-{E}_{b}/{k}_{B}T\right)\right]$$

(4)

where I(T) is the PL intensity at temperature T, I₀ represents the theoretical maximum intensity at 0 K, A is a pre-exponential factor reflecting the ratio of non-radiative to radiative processes, E_b is the exciton binding (activation) energy, and k_B is the Boltzmann constant. The fit of the experimental data (performed against 1/T) yields E_b ≈ 34.0 meV, in excellent agreement with typical values reported for low-dimensional Cu-based halides. This moderate but significant energetic stabilization of STEs confirms their resilience against thermal dissociation up to room temperature^33,34.

The overall trend revealed by Fig. 5a,b highlights the delicate interplay between structural dimensionality, excitonic self-trapping, and thermal deactivation mechanisms in CsCu₂Br₃. The substantial binding energy—considerably higher than that found in three-dimensional lead-halide perovskites—suggests that the recombination in CsCu₂Br₃ is dominated by highly localized states that are inherently less vulnerable to thermal and defect-mediated non-radiative losses. This robustness of the PL emission, combined with the persistence of a strong, sharp emission band even at elevated temperatures, underscores the promise of CsCu₂Br₃ for optoelectronic devices that demand operational stability and efficiency across a broad temperature range, such as light-emitting diodes and scintillators. Although the direct bandgap of CsCu₂Br₃ is 4.09 eV (~ 303 nm), our devices exhibit measurable photocurrent under 365 nm illumination (3.40 eV). This apparent contradiction is reconciled by considering sub-gap absorption pathways. The finite Urbach energy (E_u ≈ 0.104 eV) evidences significant band-tail states that extend below the band edge. Together with shallow defect states and Jahn–Teller-induced STE manifolds, these sub-gap states provide absorption channels at 365 nm. In this regime, the measured responsivity is modest (R ≈ 4.38 × 10⁻⁵ A W⁻¹) and requires a relatively high bias (+ 20 V) across ~ 5 mm spacing, fully consistent with trap-limited space-charge-limited conduction (m ≈ 2.45) and a weak sub-gap photoconductive mechanism. Preliminary tests at 280 nm (close to the PLE maximum) showed qualitatively stronger response, confirming that the true interband excitation occurs in the deep-UV region (≈ 280–320 nm), while 365 nm probes the defect-/tail-mediated sensitivity of CsCu₂Br₃.

Figure 7 presents the I–V characteristics of CsCu₂Br₃ single crystals measured under dark conditions and under continuous ultraviolet (UV) illumination. The measurements were conducted in a planar electrode configuration with a voltage sweep ranging from –30 V to + 30 V. The device displays a symmetric and nonlinear I–V behavior in both cases, indicating the absence of rectifying interfaces and consistent carrier transport along the low-dimensional crystal structure³⁵.

Fig. 7

Current–voltage (I–V) characteristics of CsCu₂Br₃ under dark and UV-illuminated conditions.

Under dark conditions, the current density remains in the 10⁻⁹–10⁻⁷ A cm⁻² range, increasing gradually with applied voltage due to thermally activated carrier injection. Upon UV exposure, a clear enhancement of the current density is observed across the entire voltage range, confirming the photoconductive nature of the material. Importantly, both the dark and illuminated curves intersect near the origin, showing negligible offset at 0 V, which aligns with the expected behavior for a purely photoconductive response without photovoltaic contributions³⁵.

This symmetric and voltage-dependent increase in conductivity reflects the role of UV-generated free carriers and supports the hypothesis of efficient charge separation in the 1D [CuBr₄]³⁻ tetrahedral chains. The exponential rise in current under higher fields is consistent with field-assisted detrapping of localized carriers or space-charge-limited transport mechanisms, as commonly observed in low-dimensional halide semiconductors.

To further elucidate the high-bias conduction behavior, the I–V data under UV illumination were replotted on a log–log scale (Fig. S3). The plot exhibits a clear power-law dependence J α V^m, with a slope m ≈ 2.45 in the high-voltage region, consistent with a trap-limited SCLC regime. The trap-filled-limit voltage is estimated to be V_TFL ≈ 13.2 V from the kink in the log–log J–V curve. These quantitative results substantiate the SCLC interpretation proposed in the main text and align with reports on low-dimensional halide perovskites¹⁶.

These observations are fully consistent with the findings of¹⁶, where similar current densities and bipolar switching behavior were reported for Al/CsCu₂Br₃/FTO devices. The strong UV sensitivity, coupled with excellent thermal stability and wide bandgap, makes CsCu₂Br₃ a strong candidate for photodetectors and UV-switchable resistive memory applications.

Figure 8 illustrates the time-resolved photocurrent response of CsCu₂Br₃ single crystals under periodically modulated ultraviolet (UV) illumination. The measurement was conducted at room temperature with a fixed bias of + 20 V, and a pulsed UV light source (λ = 365 nm, ~ 5 mW cm⁻²) applied in cycles of 10 s ON and 10 s OFF. The observed photocurrent dynamics reveal a rapid and reproducible switching behavior, characterized by sharp rises and decays in current density upon UV excitation and cessation, respectively. The current density transitions from a baseline of ~ 10⁻⁹ A cm⁻² under dark conditions to ~ 2 × 10⁻⁷ A cm⁻² under illumination, consistent with values previously reported for similar low-dimensional halide perovskites¹⁶.

Fig. 8

Time-resolved photocurrent response of CsCu₂Br₃ under modulated UV illumination.

The near-instantaneous rise in photocurrent upon light exposure reflects efficient photogeneration and extraction of free carriers, with minimal delay or capacitive lag, indicating favorable interfacial contact and low trap density. The decay kinetics upon light-off exhibit a weak memory effect, whereby the current does not fully return to the original dark baseline, suggesting a degree of persistent photoconductivity (PPC). This behavior is often associated with shallow trap states or slow carrier detrapping, phenomena that are commonly reported in halide-based materials with strong exciton–phonon coupling^36,37.

Notably, the high repeatability of the ON/OFF cycles, combined with low noise levels and negligible signal drift, affirms the photostability and reversibility of the CsCu₂Br₃ device under operational conditions. This temporal robustness complements the spectral properties revealed in earlier Figures, demonstrating that the photoconductive gain is not only spectrally strong but also dynamically stable^37,38. The ability of the material to sustain multiple illumination cycles without significant fatigue underscores its promise for practical optoelectronic applications requiring rapid response and environmental resilience.

Taken together, the data in Fig. 8 strengthen the assertion that CsCu₂Br₃ single crystals exhibit a compelling combination of wide bandgap, structural integrity, fast photoresponse, and repeatable switching dynamics—key attributes for UV photodetectors, logic circuits, and optical memory elements. The presence of a slight PPC effect may even be harnessed for functional memory-type behavior, opening avenues for dual-mode sensing and signal retention functionalities in next-generation halide optoelectronics³⁹.

Noise analysis. To clarify the origin and magnitude of the electrical noise, we estimated the main contributions at + 20 V bias under dark conditions. Using the measured dark current density at + 20 V (≈ 10⁻⁹ A·cm⁻²) and the active area (A ≈ 0.10 cm²), the dark current is I_dark ≈ 1.0 × 10⁻¹⁰ A. The corresponding shot-noise spectral density is⁴⁰ :

$${i}_{shot}=\sqrt{2q{I}_{dark}}\approx 5.7\times {10}^{-15} A/\sqrt{Hz}$$

(5)

The Johnson (thermal) noise, estimated from the differential resistance R_d ≈ V/I_dark ≈ 2 × 10¹¹ Ω at 300 K, is⁴¹:

$${i}_{Johnson}=\sqrt{4{k}_{B}T/{R}_{d}}\approx 9\times {10}^{-16} A/\sqrt{Hz}$$

(6)

Hence, in our biasing and bandwidth conditions, the shot noise dominates over Johnson noise. While low-frequency 1/f (flicker) noise can emerge at very low Fourier frequencies, no pronounced 1/f roll-up is observed in our time-domain traces over the acquisition window used for responsivity and timing analysis. Using the measured responsivity R ≈ 4.38 × 10⁻⁵ A·W⁻¹, the shot-noise-limited noise-equivalent power is⁴⁰:

$$NEP={i}_{shot}/R\approx 1.3\times {10}^{-10} W/\sqrt{Hz}$$

(7)

which corresponds to a specific detectivity

$${D}^{*}=\sqrt{A}/NEP\approx 2.4\times {10}^{9} Jones$$

(8)

consistent with our independently extracted detectivity based on dark current. These results indicate that the device operates close to the shot-noise limit under our measurement conditions.

Figure 9 presents the long-term photocurrent stability of CsCu₂Br₃ single crystals under continuous ultraviolet (UV) illumination (λ = 365 nm, intensity ≈ 5 mW cm⁻²) at room temperature. The device was biased at a constant voltage of + 20 V, and the current density was monitored over a 30-min interval to evaluate the durability and reliability of the photoconductive response. The photocurrent demonstrates an initially rapid increase, followed by a gradual drift toward saturation, stabilizing near ~ 2.1 × 10⁻⁷ A cm⁻². This behavior reflects a typical transient-to-steady-state evolution, attributable to the progressive filling of shallow trap states and the establishment of a dynamic equilibrium between photogeneration and recombination processes^42,43.

Fig. 9

Long-term photocurrent stability of CsCu₂Br₃ under continuous UV illumination.

The variation over 30 min remains within ~ 7.5%, which, although not perfectly flat, still indicates reasonable operational stability for a first-generation prototype device. The observed drift is most likely linked to trap filling/detrapping dynamics rather than irreversible chemical or structural degradation, since no long-term decline was detected after repeated cycling. Such behavior is common in low-dimensional halide perovskites and has been reported in similar systems^17,18. Future optimization of electrode geometry and defect passivation strategies should further suppress this transient drift.

These observations are in line with prior reports on halide-based low-dimensional semiconductors, wherein the combination of 1D transport channels and rigid inorganic frameworks supports stable photoresponse under ambient conditions^17,18. No measurable signal drift or fatigue was observed during the exposure period, emphasizing the suitability of CsCu₂Br₃ for integration into continuous-monitoring UV photodetectors or environmental sensing platforms. Notably, the preservation of photoresponsivity without significant drift or relaxation reinforces the idea that the 1D [CuBr₄]³⁻ chains serve as efficient carrier transport pathways while maintaining resilience to external thermal and optical stresses.

In summary, Fig. 9 consolidates the electronic and optoelectronic merits of CsCu₂Br₃, demonstrating not only a fast and reversible photocurrent switching (as seen in Fig. 8), but also a steady and reliable response over extended operational durations. These attributes position CsCu₂Br₃ as a compelling candidate for deployment in UV optoelectronics, especially in contexts requiring continuous exposure, low power consumption, and high signal fidelity over time. In addition to the qualitative and temporal assessments discussed above, the performance of the CsCu₂Br₃ single-crystal photodetector was quantified. Under 365 nm illumination (intensity ≈ 5 mW·cm⁻²) and + 20 V bias, the responsivity evaluated from the areal form R = J_ph/I_opt is R ≈ 4.38 × 10⁻⁵ A W⁻¹. The corresponding external quantum efficiency is EQE = R × (1240/λ) ≈ 1.49 × 10⁻⁴ (0.0149%). Assuming shot-noise-limited performance, the specific detectivity is ${D}^{*}=R/\sqrt{2q{J}_{dark}}\approx 1.66\times {10}^{8}$ Jones. It is worth noting that this shot-noise-based estimation of D^* (~ 2.4 × 10⁹ Jones) is consistent in order of magnitude with the value independently extracted from the dark-current analysis (~ 1.7 × 10⁸ Jones). The slight difference originates from the different estimation methods but both confirm that the device operates in a shot-noise-limited regime with low intrinsic noise. From the UV ON/OFF transients, the 10–90% rise and fall times are both ~ 1.10 s, and the linear dynamic range is ~ 19.2 dB. The device repeatability, assessed over 50 ON/OFF cycles at 365 nm, shows <  ± 4% variation in photocurrent amplitude with no baseline drift, confirming robust and reproducible operation. A rigorous hysteresis assessment (rate-dependent forward/reverse I–V or cyclic voltammetry at multiple scan rates) was not performed in this study ; therefore, we refrain from claiming hysteresis-free operation and instead report the observed low drift and high repeatability under our test conditions.

A comparative performance analysis is presented in Table S4 to benchmark this work against representative UV photodetectors reported in the literature. While some devices such as CsCu₂I₃ films and MoS₂/GaAs structures demonstrate higher responsivity and detectivity values, our CsCu₂Br₃ single-crystal photodetector excels in operational stability and repeatability, with negligible photocurrent drift (< 4%) over extended cycling. Moreover, the simplicity of our device architecture and the demonstration of stable operation at room temperature under ambient conditions reinforce the practical relevance of the material for durable UV sensing applications.

Overall, the comprehensive structural, optical, and photoelectrical investigations of CsCu₂Br₃ single crystals confirm their remarkable potential for ultraviolet optoelectronic applications. The material exhibits a well-defined 1D crystal structure, a wide direct bandgap of 4.09 eV, low Urbach energy, strong and stable photoluminescence, and a robust, reversible photoresponse under UV illumination. Quantitatively, the photodetector shows a responsivity of ~ 4.38 × 10⁻⁵ A.W⁻¹, a specific detectivity of ~ 1.66 × 10⁸ Jones, and an EQE of ~ 0.0149%, with excellent repeatability (< ± 4% variation) over 50 switching cycles. In addition, the 10–90% rise/fall times are ~ 1.10 s and the linear dynamic range is ~ 19.2 dB, measured under 365 nm at + 20 V.

A comparative analysis (Table S4) with representative UV photodetectors from the literature shows that although certain materials—such as CsCu₂I₃ thin films or Cs₃Cu₂I₅ nanonets—achieve markedly higher responsivity and detectivity values, the CsCu₂Br₃ single-crystal photodetector distinguishes itself in terms of operational stability, reproducibility, and fabrication simplicity. Our device maintains a negligible photocurrent drift (< ± 4% over 50 cycles) under ambient conditions at room temperature, a stability level rarely reported in devices with more complex architectures or those requiring encapsulation. Moreover, while the EQE is lower than some high-sensitivity counterparts, the robustness, low drift and high repeatability, and low-noise photoresponse position CsCu₂Br₃ as a reliable and durable candidate for UV sensing in real-world conditions where long-term stability is often prioritized over peak sensitivity. This combination of intrinsic material stability, ease of synthesis, and reproducible performance underscores its practical relevance for scalable UV optoelectronic applications.

Photocurrent transport in the CsCu₂Br₃ device is primarily governed by the generation of photo-induced carriers within the 1D [CuBr₄]³⁻ chains, followed by their extraction through the metallic contacts. Under 365 nm UV illumination, the observed photocurrent does not originate from direct band-to-band transitions, since the excitation photon energy (3.40 eV) is lower than the optical bandgap (4.09 eV). Instead, it is mediated by absorption through Urbach-tail states, shallow defects, and STE-related manifolds, which extend below the fundamental absorption edge. These channels generate photocarriers that are subsequently extracted under bias, consistent with the modest responsivity, the requirement of + 20 V, and the trap-limited SCLC regime observed in our I–V analysis.

In our planar configuration, the photogenerated carriers drift and diffuse under the applied electric field. The absence of a significant Schottky barrier (as indicated by the symmetric I–V curves) suggests nearly ohmic contact between the electrodes and the crystal, minimizing carrier injection losses. The dominant conduction mechanism is photoconductive : photogenerated carriers reduce the bulk resistivity, leading to a photocurrent proportional to the incident photon flux.

At low bias voltages, carrier mobility and lifetime limit the photocurrent, whereas at higher bias, field-assisted detrapping and, potentially, space-charge-limited conduction (SCLC) contribute to enhanced current flow. The low noise level and negligible baseline drift over multiple ON/OFF cycles confirm that nonradiative recombination and deep trap effects at the interface are minimal, accounting for the remarkable reproducibility and operational stability of the device.

Conclusion

This study provides a detailed assessment of CsCu₂Br₃ single crystals, highlighting how their 1D [CuBr₄]³⁻ chain structure governs both photophysical and transport behavior. The crystals combine a wide bandgap (4.09 eV), low energetic disorder (E_u = 0.104 eV), strong self-trapped excitonic emission (451.6 nm, τ_avg ≈ 2.8 ns), and excellent thermal stability up to ~ 300 °C with only minor extrinsic mass loss. Device characterization demonstrates symmetric I–V behavior, a trap-limited SCLC regime at high bias, and reproducible photoconductive switching with rise/fall times ~ 1.10 s. Under 365 nm illumination (photon energy below E_g), the photocurrent arises from sub-gap absorption via Urbach-tail states, shallow defects, and STE manifolds rather than direct interband transitions. Under continuous 365 nm UV exposure, the devices retain stable photocurrent with limited drift (~ 7.5%) and repeatability within ± 4% over 50 cycles. Noise analysis shows shot noise dominates, with detectivity estimates between ~ 1.7 × 10⁸ and 2.4 × 10⁹ Jones, confirming low intrinsic noise. While the responsivity is modest compared to some benchmark devices, preliminary deep-UV (≈280–320 nm) tests already show stronger response, and full calibrated spectral responsivity will be addressed in future work. Overall, CsCu₂Br₃ excels in robustness, reproducibility, and simplicity of fabrication. These attributes consolidate its role as a reliable, lead-free material for UV optoelectronic applications where long-term stability and durability are prioritized, and where performance can be further enhanced under optimized deep-UV excitation conditions.