Preferential crystallographic orientation via α-CN stereo-directional effect for superior perovskite indoor photodetectors

Abstract

This study introduces a novel approach to enhance perovskite crystallization for efficient indoor energy harvesting through the addition of a-chloronaphthalene (a-CN) as an anti-solvent additive. We demonstrate that a-CN induces a “Stereo-directional Effect” during crystallization, where its bulky naphthalene structure facilitates preferential crystal growth along specific crystallographic planes. X-ray diffraction analysis reveals that a-CN treatment significantly enhances crystallinity, with the (110) plane showing a 245.47% increase in crystallite size compared to 150.45% for the control sample. Morphological characterization confirms larger grain formation and more uniform film topography in a-CN treated samples. These structural improvements translate to superior electronic properties, including reduced defect state energy (56.5 meV vs. 71.7 meV) and improved ideality factor (1.97 vs. 2.16). Under indoor lighting conditions, a-CN treated devices demonstrate remarkable performance enhancement, achieving a power conversion efficiency of 32.23% at 1200 lux compared to 28.04% for the control device. The detectivity at 720 nm reaches 9.17 × 10¹² Jones, representing a 65.8% improvement over the control. This work establishes a direct correlation between the stereo-directional effect of a-CN and enhanced device performance, particularly under indoor low-light conditions, providing a simple yet effective strategy for developing high-efficiency indoor energy harvesting systems based on perovskite materials.

---

1. Introduction

The rapid proliferation of Internet of Things (IoT) devices has led to increasing demands for self-powered systems in indoor environments, drawing significant attention to the development of efficient indoor optoelectronic devices.¹ These devices must operate optimally under low-light conditions typically ranging from 200 to 1000 lux, a challenging operational regime that requires materials with exceptional photosensitivity and charge transport properties.² In recent years, perovskite-based optoelectronic devices have emerged as promising candidates for indoor energy harvesting and photodetection applications, owing to their excellent absorption coefficients (>10⁵ cm⁻¹), long charge diffusion lengths (>1 μm), and cost-effective solution processability.^3,4 The remarkable spectral response of perovskite materials, particularly in the 400–800 nm range, aligns well with the emission spectra of common indoor light sources, making them ideal for indoor applications.² Furthermore, their tunable bandgap, high carrier mobility, and compatibility with flexible substrates offer additional advantages for next-generation indoor electronic systems.⁵

Recent advancements in anti-solvent engineering have revealed critical molecular interactions governing perovskite crystallization. Zheng et al. demonstrated that anti-solvent donor numbers modulate nucleation kinetics through dipole–dipole interactions with PbX₆⁴⁻ octahedra.⁶ Subsequent work by the same group showed that anti-solvent vapor pressure controls intermediate phase evolution during film formation.⁷ Wang et al. established quantitative relationships between solvent parameters and preferential crystallographic orientation,⁸ while Li et al. optimized anti-solvent application timing to minimize grain boundary defects.⁹ Our work extends these principles by introducing α-CN's steric naphthalene structure to induce directional crystallization.

Despite these promising attributes, significant challenges remain in optimizing perovskite materials specifically for indoor low-light conditions. Previous studies on perovskite optoelectronic devices have primarily relied on low-boiling-point anti-solvents such as chlorobenzene (132 °C) and diethyl ether (34.6 °C) for crystallization control during film formation.⁷ While these conventional approaches have yielded reasonable results under standard testing conditions (1 sun, AM 1.5G), they often result in the formation of small crystallites and non-uniform film quality due to rapid crystallization kinetics.¹⁰ This issue becomes particularly problematic in low-light indoor environments, where efficient light absorption and charge transport become more crucial due to the limited photon flux.² Conventional crystallization methods have struggled to effectively control interface defects and charge recombination losses under these conditions, resulting in suboptimal performance metrics including limited responsivity and high dark currents.¹

Several strategies have been attempted to address these limitations, including compositional engineering, interfacial modification, and processing optimization.¹¹ However, most existing approaches face significant trade-offs between crystallinity, morphology, and electronic properties.¹² Recent advances in crystallization control strategies have explored various approaches to overcome these limitations. For instance, Zhang et al. demonstrated that modulating nucleation dynamics significantly enhance perovskite film quality and device performance through precise control of crystallization kinetics.¹³ Similarly, innovative interface engineering approaches have shown promising results in regulating crystal growth direction and grain boundary formation.¹⁴

For instance, some additive-based approaches that improve crystallinity often introduce new interfacial defects, while solvent engineering techniques that enhance grain size may compromise film uniformity.^11,15 Moreover, the majority of previous research has focused on optimizing perovskite devices for standard solar illumination conditions, with limited attention to the specific requirements of indoor applications where light intensity is typically 100–1000 times lower.⁵ This has resulted in a knowledge gap regarding the precise control mechanisms needed for developing high-performance indoor perovskite devices, particularly in controlling the crystallization process to simultaneously enhance light absorption efficiency in the 400–500 nm region and suppress interfacial charge recombination, which are key factors determining device performance in low-light environments.^3,10,16

In this study, we present an innovative approach to address these challenges by introducing α-chloronaphthalene (α-CN), a high-boiling-point (263 °C) solvent with optimal compatibility with the GBL/DMSO solvent system, to precisely control the crystallization kinetics of CH₃NH₃PbI₃ perovskite films through what we define as the “Stereo-directional Effect.”^7,17,18 Utilizing the bulky naphthalene ring structure of α-CN for strong steric hindrance and its soft Lewis-base Cl⁻ functionality for selective Pb²⁺ coordination, we achieve unprecedented control over the crystallization process through modulated adduct phase formation.^11,19 The stereo-directional effect induced by α-CN fundamentally alters the crystallization dynamics, transforming unaligned nucleation into directional selective nucleation with preferential grain orientation along specific crystallographic planes (particularly the (110), (220), and (222) planes).^10,20 This approach differs significantly from conventional anti-solvent treatments by focusing on the stereo-specific interaction between the anti-solvent additive and the perovskite precursors, rather than merely altering solvent evaporation rates.¹¹ Through optimized α-CN incorporation and enhanced solvent compatibility with the GBL/DMSO system, we achieved uniform grain formation with dimensions of 191.28 ± 28.29 nm and remarkably low defect state energy of 56.5 meV compared to 71.7 meV for control devices. X-ray diffraction analysis confirms the effectiveness of our approach, with α-CN treated films showing a 215.77% increase in (110) peak intensity, a 55.26% decrease in FWHM, and a 245.47% increase in crystallite size. This controlled crystallization process leads to enhanced electron mobility and improved film morphology, resulting in remarkable performances of 32.23% power conversion efficiency and 69% fill factor under 1200 lux indoor lighting, while achieving an impressively low dark current of 8.20 × 10⁻⁹ A cm⁻² and outstanding detectivity of 9.17 × 10¹² Jones. Our approach provides a new paradigm for perovskite crystallization control that specifically addresses the challenges of indoor optoelectronic applications, offering insights into the molecular-level interactions that govern film formation and subsequent device performance.

2. Experimental

Detailed information regarding materials, device fabrication procedures, and characterization techniques would be found in the ESI.†

3. Results and discussion

3.1 Crystallization mechanism of perovskite films with α-CN ligand addition

The crystallization mechanism of perovskite films with and without α-chloronaphthalene (α-CN) as an additive to the anti-solvent is schematically illustrated in Fig. 1. In conventional perovskite film fabrication using only chlorobenzene (CB) as an anti-solvent, the coordination environment is characterized by deficient bonding, leading to unaligned nucleation during the crystallization process.⁷ This results in non-directional crystal growth and random grain orientation, which ultimately affects the optoelectronic properties of the final device.⁷ The absence of directional control during nucleation causes inconsistent grain formation with higher boundary densities and smaller crystallite sizes.

Fig. 1 Scheme for a mechanism of grain formation according to α-CN introduction and measurement environment of device performance under indoor light condition.

In contrast, the introduction of α-CN in the anti-solvent induces a multi-directional coordination environment that fundamentally changes the nucleation dynamics through what we define as the “Stereo-directional Effect.” To further illustrate the stereo-directional effect mechanism, we provide a comprehensive step-by-step visualization in Fig. S1 (ESI†). This comparative illustration demonstrates how α-CN transforms the crystallization process through four key stages: (1) from coordination bond deficiency to multi-directional coordination with Pb²⁺ defects; (2) from unaligned nucleation to directional selective nucleation; (3) from non-directional crystal growth to preferential growth along specific crystallographic planes; and (4) from random grain orientation to aligned grain structure. The figure clearly shows how the bulky naphthalene structure of α-CN guides preferential crystal growth along the (110) plane, leading to the aligned grain orientation observed in our XRD analysis (Fig. 2 and Table 1). This stereo-directional crystallization contributes to the enhanced morphological properties discussed in subsequent sections.

Fig. 2 (a) X-ray diffraction spectroscopy (XRD) full spectra of CH₃NH₃PbI₃ films at adduct phase and perovskite phase; (b) comparative analysis of (110) plane peak profiles without α-CN; (c) enhanced (110) plane peak profiles with α-CN incorporation demonstrating stereodirectional crystallization effect, characterized by FWHM reduction and crystallite size increase. The bulky naphthalene structure of α-CN facilitates preferential crystal growth along specific crystallographic directions, resulting in highly ordered crystal domains.

Table 1 Rate of change of peak intensity, full width at half maximum (FWHM) values, and crystallite size derived from (110) plane of XRD spectra (from adduct to perovskite phase) shown in Fig. 2

(110)	ΔIntensity (%)	ΔFWHM (%)	ΔCrystallite size (%)
W/O α-CN	+205.08	−33.52	+150.45
W/α-CN	+215.77	−55.26	+245.47

---

In contrast, the introduction of α-CN in the anti-solvent induces a multi-directional coordination environment that fundamentally changes the nucleation dynamics through what we define as the “Stereo-directional Effect.”²¹ This phenomenon refers to how the sterically bulky structure of the additive directly influences the directionality of crystal growth along specific crystallographic planes (particularly the (110), (220), and (222) planes).²⁰ As detailed in Table S1 (ESI†), α-CN (C₁₀H₇Cl) possesses distinct physicochemical properties compared to chlorobenzene (C₆H₅Cl), including a significantly higher boiling point (263 °C versus 132 °C) and greater density (1.194 g ml⁻¹versus 1.106 g ml⁻¹).¹² These properties contribute to the altered interaction between the anti-solvent and the perovskite precursor solution during the spin-coating process. The higher boiling point of α-CN allows for a slower evaporation rate, providing extended time for ordered crystallization, while its greater molecular weight and bulky naphthalene ring structure introduce stereo-directional effects that influence the nucleation process.²²

The fabrication process of perovskite films is detailed in Fig. S1 (ESI†), where the anti-solvent treatment plays a critical role in controlling the crystallization kinetics.⁶ During spin-coating of the CH₃NH₃PbI₃ precursor solution, the anti-solvent (CB + α-CN) is dropped approximately 10 seconds before the end of the spinning process, followed by annealing for perovskite grain formation. The steric hindrance effect of α-CN is particularly significant in this context. As illustrated in Fig. S1 (ESI†), the spatial arrangement of atoms in the α-CN molecule induces a multi-directional force field that delays molecular reactivity due to its bulky naphthalene structure.²¹ This steric hindrance effect contributes to inducing a stable nucleation environment during grain formation by moderating the crystallization rate and allowing for more ordered growth along preferential crystallographic orientations.

The stereo-directional effect is a key mechanism in this study, where the sterically bulky naphthalene ring structure of α-CN directly controls the growth direction and quality of perovskite crystals.⁶ The intermolecular interactions between α-CN and the perovskite precursor components facilitate stereo-directional crystallization, resulting in preferential grain orientation as shown in Fig. 1.¹¹ This controlled nucleation process leads to larger perovskite grain formation with improved crystallinity, which is confirmed by the optical analysis presented in Fig. S2 (ESI†).

The UV-visible absorbance spectra (Fig. S2a, ESI†) show enhanced absorption in the 400–500 nm range for samples with α-CN treatment compared to the control samples.²³ This region corresponds to the PbI₂ band edge, providing direct evidence of interaction between Pb²⁺ and α-CN. This enhanced absorption would be attributed to the stereo-directional effect, where the preferential crystal growth along specific planes leads to better light harvesting properties. This observation is consistent with previous research on perovskite absorption characteristics (as referenced in Fig. S2, ESI†), where the absorption band edge corresponds to the optical properties of the CH₃NH₃PbI₃ phase.

Further evidence of enhanced crystallinity is provided by the steady-state photoluminescence measurements (Fig. S2b, ESI†), which demonstrate increased emission intensity for α-CN treated samples. The sharper and more intense PL peak centered at approximately 780 nm for the α-CN treated sample indicates reduced non-radiative recombination pathways, which would be attributed to the improved crystal quality and reduced defect density in the film.²⁴ This enhancement in photoluminescence intensity further confirms the effectiveness of the stereo-directional effect in improving the optoelectronic properties of the perovskite layer.

The crystallization mechanism proposed in Fig. 1 clearly demonstrates how the α-CN additive induces a more favorable environment for controlled perovskite crystal growth through the stereo-directional effect. By introducing multi-directional coordination through steric effects, α-CN guides the nucleation process toward directional selective nucleation rather than random growth. This results in stereo-directional crystallization with preferential grain orientation along the (110) plane as illustrated in Fig. 1, which is beneficial for charge transport in the final device structure.²³ This concept of directional crystal growth aligns with recent findings by Zhang et al.,¹³ who demonstrated that controlled coordination environments significantly influence perovskite crystallization pathways. However, our approach using α-CN introduces a unique steric component to this process, providing an additional control mechanism that differs from previously reported sterically-influenced processes.¹⁴ Additionally, the specific interactions between α-CN and the CH₃NH₃PbI₃ precursor components may contribute to passivating defect sites at grain boundaries, further enhancing the optoelectronic properties of the resulting films.²³

The stark contrast between the conventional (W/O α-CN) and modified (W/α-CN) crystallization processes is clearly depicted in Fig. 1, where the non-directional crystal growth is transformed into a stereo-directional crystallization effect with preferential grain orientation.²⁵ This fundamental difference in crystallization behavior underlies the significant improvements in device performance observed under various lighting conditions, particularly in indoor low-light environments, as the stereo-directional effect enhances the structural quality and optoelectronic properties of the perovskite active layer.²⁶

3.2 XRD analysis and crystalline properties of CH₃NH₃PbI₃ films

To substantiate the proposed stereo-directional effect mechanism, we conducted comprehensive X-ray diffraction (XRD) analyses of CH₃NH₃PbI₃ films with and without α-CN treatment. Fig. 2 presents the full XRD spectra of these films at both the adduct phase and the final perovskite phase, with particular focus on the (110) crystallographic plane. This analysis provides direct evidence for the impact of α-CN on the crystallization process and resultant film quality.

As shown in Fig. 2(a), both samples exhibit characteristic diffraction peaks corresponding to the tetragonal perovskite structure.²⁰ However, the α-CN treated sample displays significantly enhanced peak intensities across the spectrum, particularly for the primary (110) reflection at around 14.1° (2θ).²⁰ The transition from the adduct phase to the perovskite phase reveals crucial differences between the control and α-CN treated samples.⁷ For the control sample without α-CN (Fig. 2(b)), the (110) reflection increases in intensity during the phase transition, but the peak remains relatively broad. In contrast, the α-CN treated sample (Fig. 2(c)) shows a dramatically sharper and more intense (110) reflection in the perovskite phase, indicating superior crystallinity and preferred orientation.²³

A notable feature observed in the XRD patterns (Fig. 2(b), (c) and Fig. S4, S5, ESI†) is the leftward shift (∼0.5°) of the main diffraction peaks ((110), (220), and (222)) in the α-CN treated samples compared to the control. This shift toward lower 2θ values indicates increased interplanar spacing according to Bragg's law (nλ = 2d·sin θ), suggesting lattice expansion in the α-CN treated films. This phenomenon would be attributed to three primary mechanisms. First, the bulky naphthalene structure of α-CN introduces steric hindrance during crystallization, moderating growth kinetics and enabling a more ordered crystal assembly with slightly expanded unit cells. Second, α-CN promotes preferential crystal growth along specific crystallographic planes, as evidenced by the enhanced intensities and reduced FWHM values (Table 1 and Tables S2, S3, ESI†), which helps relieve microstrain within the lattice. Third, the chlorine functionality in α-CN may weakly coordinate with Pb²⁺ at grain boundaries, subtly altering bond lengths within PbI₆ octahedra. These combined effects result in a slight expansion of the perovskite lattice parameters, manifested as the observed peak shifts in the XRD patterns.

The quantitative analysis of these XRD patterns, presented in Table 1, provides compelling evidence for the stereo-directional effect. The rate of change in peak intensity from the adduct to perovskite phase for the (110) plane is notably higher for the α-CN treated sample (215.77%) compared to the control (205.08%). More significantly, the full width at half maximum (FWHM) decrease is more pronounced in the α-CN treated sample (−55.26%) than in the control (−33.52%), indicating a substantial improvement in crystalline order.²⁰ The most striking difference is observed in the crystallite size change, where the α-CN treated sample shows a remarkable 245.47% increase compared to only 150.45% for the control. These quantitative metrics confirm that α-CN profoundly enhances the crystallization process, resulting in larger, more ordered crystallites with preferential orientation along the (110) direction.

The Scherrer equation, presented in Fig. 2, was employed to calculate the crystallite sizes from the XRD data. The significant reduction in FWHM values for the α-CN treated samples translates directly to larger calculated crystallite sizes, confirming the effectiveness of α-CN in promoting crystal growth.

The preferential orientation along specific crystallographic planes is further supported by the analysis of other major diffraction peaks. Fig. S3 and S4 (ESI†) present the XRD data for the (220) and (222) planes, respectively.²⁵ The (220) plane analysis (Fig. S3, ESI†) reveals a similar pattern of enhanced crystallinity with α-CN treatment. As quantified in Table S2 (ESI†), the α-CN treated sample shows a lower increase in peak intensity (+147.31% versus +188.49% for the control) but a more significant decrease in FWHM (−36.59% versus −29.54%), resulting in a larger crystallite size increase (+157.69% versus +141.95%). This suggests that while the overall volume of crystals oriented along the (220) plane may be smaller in the α-CN sample, those crystals are more perfectly formed and larger.

The analysis of the (222) plane, presented in Fig. S4 and Table S3 (ESI†), provides additional evidence for the stereo-directional effect. Here, the α-CN treated sample demonstrates substantially higher increases in peak intensity (+245.87% versus +194.45%), more significant decreases in FWHM (−46.11% versus −23.60%), and larger crystallite size increases (+185.49% versus +130.90%) compared to the control. These results indicate that α-CN not only enhances crystallinity along the primary (110) direction but also improves crystal quality across multiple crystallographic orientations. This preferential crystallographic orientation effect is consistent with observations by Li et al.,²⁷ who reported that controlling growth along specific crystallographic planes would dramatically improve charge transport properties in perovskite films. Moreover, our findings on crystallite size enhancement (245.47%) for the (110) plane exceed previously reported improvements using alternative anti-solvent engineering approaches,²⁸ highlighting the exceptional efficacy of our stereo-directional effect strategy.

The comparative analysis across all three major diffraction planes ((110), (220), and (222)) reveals an important pattern: the α-CN treatment consistently produces larger crystallites with narrower diffraction peaks, indicating higher crystalline perfection. The most dramatic improvements are observed in the (110) and (222) planes, suggesting that the stereo-directional effect particularly enhances growth along these crystallographic directions. This preferential growth would be attributed to the specific interactions between the bulky naphthalene structure of α-CN and the perovskite precursor components during nucleation.

The relationship between these XRD findings and the proposed mechanism in Fig. 1 is particularly significant. The multi-directional coordination environment induced by α-CN during the anti-solvent treatment facilitates more ordered nucleation, which manifests as preferential growth along specific crystallographic directions.¹¹ The steric hindrance effect of α-CN, illustrated in Fig. S1 (ESI†), moderates the crystallization kinetics, allowing for more perfect crystal formation.²¹ This results in the observed enhancement in diffraction peak intensities, reduction in peak broadening, and substantial increase in crystallite sizes.

These XRD results provide compelling quantitative support for the stereo-directional effect mechanism, demonstrating how the introduction of α-CN fundamentally alters the crystallization dynamics of CH₃NH₃PbI₃ films. The larger crystallite sizes, reduced FWHM values, and enhanced peak intensities all point to superior crystalline quality, which directly impacts the optoelectronic properties of the resulting films.²³ The preferential orientation along specific crystallographic planes, particularly the (110) direction, suggests improved charge transport pathways that would enhance device performance, especially under low-light conditions where efficient charge collection is crucial.²⁹

3.3 Morphological analysis of CH₃NH₃PbI₃ films

Building upon the crystallographic evidence from XRD analysis, we conducted detailed morphological investigations to further validate the impact of the stereo-directional effect on perovskite film formation. Fig. 3 presents a comprehensive morphological analysis of CH₃NH₃PbI₃ films with and without α-CN treatment using complementary microscopy techniques.

Fig. 3 Top-view field emission scanning electron microscopy images of CH₃NH₃PbI₃ films (a) without α-CN and (c) with α- (FE-SEM) CN. Size distribution histograms and Gaussian fitting curves of CH₃NH₃PbI₃ film diameters (b) without α-CN and (d) with α-CN. Atomic force microscopy (AFM) images and height distribution of CH₃NH₃PbI₃ films (e), (f) without α-CN and (g), (h) with α-CN. The height distribution was measured along the diagonal line shown in AFM images.

The field emission scanning electron microscopy (FE-SEM) top-view images in Fig. 3(a) and (c) reveal distinct differences in grain morphology between the control and α-CN treated samples. The control sample without α-CN (Fig. 3(a)) displays relatively small grains with irregular shapes and less defined grain boundaries. In contrast, the α-CN treated sample (Fig. 3(c)) exhibits significantly larger grains with well-defined boundaries and more uniform morphology.²⁶ This observation directly correlates with the XRD results presented in Fig. 2 and Table 1, where the α-CN treated sample showed larger crystallite sizes and enhanced crystallinity along the (110), (220), and (222) planes. The larger grain size observed in the SEM images is a physical manifestation of the enhanced crystallization process facilitated by the stereo-directional effect.

To quantify these morphological differences, size distribution histograms were generated from the FE-SEM images, as shown in Fig. 3(b) and (d). The control sample (Fig. 3(b)) exhibits a size distribution centered around 150–200 nm with a relatively narrow range. In contrast, the α-CN treated sample (Fig. 3(d)) shows a distribution peaked at approximately 150–180 nm but extended to larger sizes up to 350 nm. The Gaussian fitting curves applied to these distributions confirm the broader size range and larger average grain size for the α-CN treated sample.³⁰ This quantitative analysis supports the qualitative observation that α-CN treatment leads to larger perovskite grains, consistent with the crystallite size increases calculated from XRD data.

To complement the SEM analysis and gain further insights into the surface topography, atomic force microscopy (AFM) was employed, as shown in Fig. 3(e)–(h). The AFM images of the control sample (Fig. 3(e)) and α-CN treated sample (Fig. 3(g)) both reveal granular surface morphology, but with notable differences in texture and height distribution.³¹ The root mean square (RMS) roughness values, reported as 12.829 nm for the control and 13.568 nm for the α-CN treated sample, indicate a slightly rougher surface for the latter. This increased roughness would be attributed to the larger grain sizes and more pronounced grain boundaries in the α-CN treated film.³⁰

The height distribution histograms derived from the AFM data, presented in Fig. 3(f) and (h), provide additional quantitative insights. The control sample (Fig. 3(f)) shows a broad height distribution spanning from approximately −6 nm to +6 nm, indicating significant variation in surface height. In contrast, the α-CN treated sample (Fig. 3(h)) exhibits a narrower and more concentrated height distribution centered around 2–3 nm. This suggests that while the α-CN treated film has slightly higher overall roughness, it possesses a more uniform and less fragmented surface topology, with most features clustered around a similar height.²⁰ The more homogeneous height distribution in the α-CN treated film indicates better film quality with fewer defects and voids, which would be beneficial for device performance.

Further analysis of film morphology was conducted using whole-area atomic force microscopy (AFM) measurements over larger scan areas (5 × 5 μm²), as shown in Fig. S6 (ESI†). The 3D topography images reveal distinct surface features for both control (Fig. S6a, ESI†) and α-CN treated (Fig. S6c, ESI†) perovskite films. Quantitative analysis of height distribution histograms (Fig. S6b, d, ESI†) demonstrates that α-CN treatment slightly increases the RMS roughness from 12.802 nm to 13.209 nm, with a concurrent increase in peak-to-valley height from 92.942 nm to 97.547 nm. Notably, the surface skewness shifts from negative (−0.185) in the control film to positive (0.089) in the α-CN treated film, indicating a transition from valley-dominated to peak-dominated topography. This morphological transformation, coupled with a reduction in kurtosis from 3.193 to 2.989, suggests more uniform height distribution in the α-CN treated film. These surface modifications enhance light trapping capability and reduce random scattering losses, contributing to the improved photodetector performance under low-light conditions observed in Fig. 4.

Fig. 4 Current–voltage characteristics showing (a) dark J–V curves, (b) ideality factor extraction and (c) trap-assisted tunneling (TAT) current analysis with and without α-CN treatment. (d) external quantum efficiency (EQE) spectra, (e) responsivity (0 V), and (f) detectivity (0 V) of CH₃NH₃PbI₃-based devices without or with α-CN.

The diagonal line profiles shown in the AFM images (Fig. 3(e) and (g)) and further analyzed in Fig. S6 (ESI†) provide additional evidence for the morphological differences. Fig. S6a (ESI†) shows the height profile across the control sample, revealing multiple small peaks and valleys with moderate height differences. In contrast, Fig. S6b (ESI†) displays the height profile for the α-CN treated sample, showing fewer but more pronounced peaks with greater height differences. The α-CN treated sample exhibits a profile with peak heights reaching approximately 15–17 nm, compared to about 8–10 nm for the control sample. This confirms that the α-CN treatment leads to larger and more distinct grains with more defined boundaries, consistent with the SEM observations.

The cross-sectional FE-SEM images presented in Fig. S5 (ESI†) offer a complementary perspective on the film morphology and device architecture. Fig. S5a (ESI†) shows the control device structure while Fig. S5b (ESI†) displays the α-CN treated device structure, both featuring the complete stack of ITO/PEDOT:PSS/CH₃NH₃PbI₃/PC₇₀BM/Al layers. The perovskite layer in the α-CN treated device (Fig. S5b, ESI†) appears to have larger grains that span more of the film thickness compared to the control (Fig. S5a, ESI†), where the grains are smaller and more numerous.³⁰ This observation supports the top-view SEM and AFM findings, confirming that the effect of α-CN treatment persists throughout the entire film thickness.

The morphological analysis results are consistent with and further elaborate on the crystallization mechanism proposed in Fig. 1. The stereo-directional effect, facilitated by the bulky naphthalene structure of α-CN, promotes preferential crystal growth and results in larger, more uniform grains with well-defined boundaries. This improved grain structure directly corresponds to the enhanced crystallinity observed in the XRD analysis (Fig. 2 and Table 1), where the α-CN treated samples showed significant increases in crystallite size and preferential orientation along specific crystallographic planes.

The larger grain size and reduced grain boundary density in the α-CN treated films have significant implications for device performance.³² Larger grains generally result in fewer grain boundaries, which act as recombination sites for charge carriers. The reduction in these interfaces would lead to decreased charge carrier recombination, longer carrier lifetimes, and improved charge transport.^26,33 Additionally, the more uniform height distribution observed in the AFM analysis suggests better film quality with fewer defects, which would further enhance charge carrier dynamics.^26,34 These morphological improvements, driven by the stereo-directional effect of α-CN, provide a physical basis for enhanced optoelectronic performance, particularly under low-light conditions where efficient charge collection is crucial.³⁵

The combination of XRD analysis (Fig. 2) and morphological characterization (Fig. 3) provides comprehensive evidence for the structural and textural improvements achieved through α-CN treatment. The larger crystallite sizes calculated from XRD data are physically manifested as larger grains observed in SEM and AFM, confirming that the stereo-directional effect enhances both the microscopic crystalline order and the macroscopic film morphology. This integrated understanding of how α-CN influences perovskite film formation at multiple length scales provides a solid foundation for interpreting the subsequent device performance data.

3.4 Electrical characteristics and optoelectronic performance of CH₃NH₃PbI₃ devices

Having established the enhanced crystallinity and improved morphology of α-CN treated films, we next investigated how these structural improvements translate to electrical and optoelectronic performance. Fig. 4 presents a comprehensive analysis of the current–voltage characteristics and spectral response of CH₃NH₃PbI₃-based devices with and without α-CN treatment.

The dark current–voltage (J–V) characteristics, shown in Fig. 4(a), reveal a significant reduction in leakage current for the α-CN treated device compared to the control.³⁶ This reduction is particularly pronounced in the low-voltage region, where the α-CN treated device exhibits a dark current density of 8.20 × 10⁻⁹ A cm⁻² compared to 2.01 × 10⁻⁸ A cm⁻² for the control device, as quantified in Table 2. This 59.2% reduction in dark current indicates superior diode characteristics and fewer leakage pathways in the α-CN treated device, consistent with the improved film morphology observed in Fig. 3. The larger grain size and reduced grain boundary density revealed by FE-SEM and AFM analyses directly contribute to this reduction in leakage current by minimizing recombination sites and charge trapping centers.

Table 2 Device performance parameters: values of dark J, ideality factor and defect state energiesresponsivity and detectivity of CH₃NH₃PbI₃-based devices without or with α-CN at 720 nm under 0 V bias

Condition	Dark J (A cm^−2)	Ideality factor	Defect state [meV]	Responsivity (A W^−1)	Detectivity (Jones)
W/O α-CN	2.01 × 10^−8	2.16	71.7	0.443	5.53 × 10^12
W/α-CN	8.20 × 10^−9	1.97	56.5	0.470	9.17 × 10^12

---

To gain deeper insights into the charge transport and recombination mechanisms, we extracted the ideality factor from the dark J–V curves, as shown in Fig. 4(b). The α-CN treated device exhibits an ideality factor of 1.97, closer to the ideal value of 1, compared to 2.16 for the control device (Table 2). This improvement in ideality factor indicates reduced non-radiative recombination in the α-CN treated device, consistent with the enhanced crystallinity and preferential orientation observed in XRD analysis (Fig. 2). The stereo-directional effect induced by α-CN not only improves the structural properties of the perovskite film but also enhances its electronic quality by reducing defect states that contribute to recombination.

Further evidence of improved electronic quality is provided by the trap-assisted tunneling (TAT) current analysis presented in Fig. 4(c). The theoretical fitting of the TAT current reveals defect state energies of 56.5 meV for the α-CN treated device compared to 71.7 meV for the control (Table 2).³⁶ This significant reduction in defect state energy indicates a lower density of electronic trap states in the α-CN treated device, which would be directly attributed to the improved crystalline quality and larger grain size observed in XRD and SEM analyses (Fig. 2 and 3). The stereo-directional effect, by promoting preferential crystal growth along specific crystallographic planes, results in a more electronically homogeneous material with fewer energetic disorders.

To further elucidate the energy level modifications induced by α-CN treatment, we conducted comprehensive cyclic voltammetry (CV) measurements on various layer structures (Fig. S9, ESI†). The CV data reveals that α-CN treatment modifies the HOMO energy levels in the perovskite layer, inducing more favorable band alignment with adjacent transport layers. Specifically, in the full device structure, α-CN shifts the HOMO level from −5.13 eV to −5.03 eV, improving hole extraction at the PEDOT:PSS interface. This energy level optimization works synergistically with the reduced defect state energy and improved crystallinity to enhance charge separation and transport efficiency.

The enhanced electronic quality of the α-CN treated device manifests in superior optoelectronic performance, as demonstrated by the external quantum efficiency (EQE) spectra in Fig. 4(d). The α-CN treated device shows higher EQE values across the entire visible spectrum, with particularly significant enhancement in the 400–500 nm range, corresponding to the spectral region where enhanced absorption was observed in UV-visible spectroscopy (Fig. S2a, ESI†). This correlation between improved absorption and enhanced EQE confirms that the stereo-directional effect not only enhances optical properties but also improves photogenerated charge collection efficiency.

The responsivity spectra at 0 V bias, presented in Fig. 4(e), further validates the superior optoelectronic performance of the α-CN treated device. At the peak wavelength of 720 nm, the α-CN treated device exhibits a responsivity of 0.470 A W⁻¹ compared to 0.443 A W⁻¹ for the control device (Table 2), representing a 6.1% improvement. This enhancement in responsivity indicates more efficient conversion of incident photons to electrical current, which would be attributed to the combined effects of improved absorption, reduced recombination, and enhanced charge transport in the α-CN treated device.

The most significant performance enhancement is observed in the detectivity spectra (Fig. 4(f)), where the α-CN treated device demonstrates a peak detectivity of 9.17 × 10¹² Jones at 720 nm compared to 5.53 × 10¹² Jones for the control device (Table 2), representing a remarkable 65.8% improvement. This substantial enhancement in detectivity stems from the combined benefits of increased responsivity and reduced dark current, both of which are direct consequences of the improved structural and electronic properties facilitated by the stereo-directional effect of α-CN.

The relationship between device structure and optoelectronic performance is further elucidated in Fig. S7 (ESI†), which presents the perovskite device architecture (Fig. S7a, ESI†) and the corresponding energy band diagram (Fig. S7b, ESI†). The energy band diagram illustrates the electron and hole flow mechanisms, highlighting the role of α-CN in modifying the electronic properties of the CH₃NH₃PbI₃ layer.³⁷ The interface between CH₃NH₃PbI₃ and PC₇₀BM facilitates efficient electron extraction, while the PEDOT:PSS/CH₃NH₃PbI₃ interface enables hole extraction. The α-CN treatment, by improving the crystalline quality and reducing defect states, enhances charge separation and transport through these interfaces, resulting in the observed improvements in device performance.

Statistical analyses of device performance parameters, presented in Fig. S8 (ESI†), confirm the reproducibility and reliability of the observed enhancements. Fig. S8a (ESI†) shows the statistical distribution of dark current density for multiple devices, with the α-CN treated devices consistently exhibiting lower values compared to the control devices. Similarly, Fig. S8b (ESI†) demonstrates the statistical distribution of responsivity, with α-CN treated devices showing consistently higher values. These statistical data validate that the performance improvements are not isolated to individual devices but represent a systematic enhancement attributable to the stereo-directional effect of α-CN.

The electrical and optoelectronic performance improvements observed in Fig. 4 and quantified in Table 2 establish a clear correlation with the structural and morphological enhancements demonstrated in Fig. 2 and 3. The stereo-directional effect, which promotes preferential crystal growth and results in larger, more uniform grains with enhanced crystallinity, directly translates to superior electronic properties with reduced defect states, lower dark current, improved ideality factor, and enhanced optoelectronic performance metrics including responsivity and detectivity.

These comprehensive characterizations of electrical and optoelectronic properties provide compelling evidence for the efficacy of α-CN treatment in enhancing CH₃NH₃PbI₃ device performance through the stereo-directional effect. The systematic improvements observed across multiple performance metrics, their correlation with structural and morphological enhancements, and the statistical validation of these improvements establish a robust foundation for understanding how the molecular structure of α-CN influences perovskite crystallization and ultimately device performance.³⁸ Long-term stability measurements conducted over a 40-day period demonstrate superior performance retention in α-CN treated devices (Fig. S12, ESI†). These devices maintained 95–98% of initial responsivity throughout the testing period (Fig. S12b, ESI†), compared to control devices which degraded to approximately 88–90% (Fig. S12a, ESI†). The enhanced stability correlates with the stereo-directional effect, where preferential crystallographic orientation along the (110) plane provides a robust structural framework against degradation. The larger grains and reduced grain boundaries (Fig. 3) minimize ion migration pathways, while lower trap state energy (56.5 meV vs. 71.7 meV, Fig. 4(c)) reduces light-induced defect formation. Notably, α-CN treated devices maintain consistent spectral response across the 400–750 nm range (Fig. S12d, ESI†), while control devices exhibit wavelength-dependent degradation (Fig. S12c, ESI†), particularly at 400–500 nm.

3.5 Performance of CH₃NH₃PbI₃ devices under various light conditions

Having established the superior electrical characteristics and optoelectronic performance of α-CN treated devices under standard measurement conditions, we next investigated their performance under various lighting conditions, with particular emphasis on indoor low-light environments. Fig. 5 presents a comprehensive analysis of device performance metrics across different illumination intensities, highlighting the advantages of α-CN treatment for indoor energy harvesting applications.

Fig. 5 J–V curves under (a) 1 sun, (b) 1200 lux, (c) 917 lux, and (d) 336 lux of CH₃NH₃PbI₃-based devices without or with α-CN. R–V curves under (e) 1 sun, (f) 1200 lux, (h) 917 lux, and (h) 336 lux of CH₃NH₃PbI₃-based devices without or with α-CN.

The responsivity–voltage (R–V) characteristics under 1 sun illumination, presented in Fig. 5(a), show consistently higher responsivity for the α-CN treated device compared to the control across the entire voltage range. This enhanced responsivity under standard AM 1.5G illumination is complemented by superior detectivity, as shown in Fig. 5(b), where the α-CN treated device exhibits significantly higher detectivity values, particularly at lower bias voltages. These improvements under standard solar illumination establish a baseline for the performance enhancements achieved through the stereo-directional effect of α-CN.

More remarkable performance improvements are observed under indoor lighting conditions. Fig. 5(c) shows the R–V characteristics under 1200 lux illumination, representative of typical indoor office lighting. The α-CN treated device maintains its superior responsivity across the voltage range, with particularly enhanced performance at lower voltages.²⁶ Similarly, the detectivity under 1200 lux (Fig. 5(d)) is substantially higher for the α-CN treated device, indicating excellent sensitivity under indoor lighting conditions. This enhanced performance would be directly linked to the improved crystallinity (Fig. 2) and morphology (Fig. 3) of the α-CN treated films, which result in more efficient charge generation and collection even under low-light intensity.³⁹

As the illumination intensity decreases further to 917 lux (Fig. 5(e) and (f)) and 336 lux (Fig. 5(g) and (h)), the α-CN treated device maintains its superior performance over the control. The responsivity values at 0 V bias across these different light intensities demonstrate the consistent advantage of α-CN treatment, with the performance gap between treated and control devices often widening at lower light intensities. This trend suggests that the structural and electronic improvements facilitated by the stereo-directional effect become even more significant under low-light conditions, where efficient charge generation and collection are particularly challenging.

The current–voltage (J–V) characteristics under various light conditions, presented in Fig. S9 (ESI†), provide complementary insights into device performance. Fig. S9a (ESI†) shows the J–V curves under 1 sun illumination, with the α-CN treated device exhibiting a higher short-circuit current density (J_sc) of 20.90 mA cm⁻² compared to 20.12 mA cm⁻² for the control, as detailed in Table S4 (ESI†). More significantly, the fill factor (FF) increases from 72% for the control to 76% for the α-CN treated device, resulting in an improved power conversion efficiency (PCE) of 14.59% compared to 13.53%. This enhancement in FF suggests better charge extraction and reduced series resistance in the α-CN treated device, consistent with the improved morphology and reduced defect density established in our earlier analyses.

The performance enhancement becomes more pronounced under indoor lighting conditions. Under 1200 lux illumination (Fig. S9b, ESI†), the α-CN treated device achieves a remarkable PCE of 32.23% compared to 28.04% for the control (Table S4, ESI†), representing a 15% improvement. Similarly impressive enhancements are observed at 917 lux (Fig. S9c, ESI†) and 336 lux (Fig. S9d, ESI†), with the α-CN treated device achieving PCEs of 29.58% and 25.06% compared to 24.97% and 21.93% for the control, respectively. These substantial efficiency improvements under indoor lighting conditions highlight the particular suitability of α-CN treated perovskite devices for indoor energy harvesting applications.

The exceptional performance of our α-CN treated devices in the context of the broader literature is illustrated in Fig. S10 (ESI†), which presents a performance comparison chart plotting PCE against light intensity for various perovskite-based indoor photovoltaic devices. Our devices (marked as “This Work”) demonstrate competitive PCE values across various light intensities, particularly at higher lux levels, positioning them among the top-performing devices in the field. This comparison validates the significance of the stereo-directional effect in enhancing indoor light harvesting capabilities.

The linear dependence of device performance parameters on light intensity is further analyzed in Fig. S11 (ESI†). The linear dynamic range (LDR) analysis in Fig. S11a (ESI†) demonstrates extended linearity for the α-CN treated device, with an LDR of 128.13 dB compared to 120.01 dB for the control. This enhanced LDR indicates superior sensitivity across a wider range of light intensities, an important characteristic for practical indoor energy harvesting applications.²⁶ The PCE versus light intensity plot (Fig. S11b, ESI†) and detectivity versus light intensity plot (Fig. S11c, ESI†) both show consistent enhancement for the α-CN treated device across all measured light intensities, with the performance advantage often becoming more pronounced at lower illumination levels.

Table S4 (ESI†) provides a comprehensive summary of device performance parameters under various light conditions. Particularly noteworthy is the consistent improvement in fill factor for the α-CN treated devices across all light intensities. Under 1200 lux, the FF increases from 60% for the control to 69% for the α-CN treated device, and similar enhancements are observed at 917 lux (57% to 68%) and 336 lux (48% to 61%). This consistent improvement in FF, particularly pronounced at lower light intensities, highlights the effectiveness of α-CN treatment in enhancing charge extraction efficiency under indoor lighting conditions.³⁹

The performance of our α-CN treated devices in the context of the state-of-the-art is further validated in Table S5 (ESI†), which presents a comprehensive comparison with various perovskite-based indoor photovoltaic devices reported in the literature.²⁶ Our devices demonstrate competitive performance metrics across different indoor lighting conditions, with PCEs comparable to or exceeding those of devices employing more complex architectural strategies or compositional modifications. This favorable comparison underscores the efficacy of our relatively simple approach of incorporating α-CN as an anti-solvent additive to enhance device performance through the stereo-directional effect. Notably, our achieved power conversion efficiency of 32.23% at 1200 lux represents a significant advancement over previously reported efficiencies for similar device architectures,⁴⁰ where conventional anti-solvent approaches typically yielded efficiencies below 30% under comparable lighting conditions. This performance enhancement would be attributed to the superior crystalline quality facilitated by our α-CN-induced stereo-directional effect.

The superior performance of α-CN treated devices under indoor lighting conditions would be directly correlated with the structural and electronic improvements established in our earlier analyses. The enhanced crystallinity and preferential orientation along specific crystallographic planes (Fig. 2) result in improved charge transport pathways. The larger grain size and reduced grain boundary density (Fig. 3) minimize recombination sites and enhance charge carrier lifetime. The reduced defect state energy and improved ideality factor (Fig. 4) indicate a lower density of electronic trap states. Collectively, these improvements enable more efficient charge generation, separation, and collection, particularly under the low-intensity illumination characteristic of indoor environments.

The stereo-directional effect of α-CN, initially proposed as a mechanism for enhanced perovskite crystallization (Fig. 1), is thus validated through comprehensive characterization from crystalline structure to device performance. By promoting preferential crystal growth along specific crystallographic planes, α-CN treatment results in perovskite films with superior structural, morphological, and electronic properties, which translate directly to enhanced optoelectronic performance, particularly under indoor lighting conditions. This integrated understanding of structure–property–performance relationships establishes a solid foundation for the rational design of efficient indoor energy harvesting devices based on perovskite materials.

4. Conclusions

In this work, we have demonstrated that incorporating α-chloronaphthalene (α-CN) as an anti-solvent additive significantly enhances the crystallization and performance of CH₃NH₃PbI₃-based devices for indoor energy harvesting. The stereo-directional effect induced by α-CN's bulky naphthalene structure promotes preferential crystal growth, resulting in superior structural and electronic properties. XRD analysis confirmed enhanced crystallinity in α-CN treated films, with a 245.47% increase in crystallite size along the (110) plane compared to the control. This improved crystallinity manifested as larger grains with well-defined boundaries in morphological characterization. The enhanced structural properties directly translated to superior electronic characteristics, including a 59.2% reduction in dark current density, improved ideality factor (1.97 vs. 2.16), and reduced defect state energy (56.5 meV vs. 71.7 meV). Most significantly, under indoor lighting conditions of 1200 lux, α-CN treated devices achieved a power conversion efficiency of 32.23% and a detectivity of 9.17 × 10¹² Jones, representing improvements of 15% and 65.8% respectively over the control. This comprehensive study establishes a clear correlation between the stereo-directional effect of α-CN and enhanced optoelectronic performance, offering a simple yet effective strategy for developing high-efficiency perovskite-based indoor energy harvesting systems without requiring complex compositional modifications or architectural designs.

Author contributions

Byung Gi Kim: conceptualisation, investigation, formal analysis, writing – original draft, and visualisation; Goda, Emad S.: investigation, formal analysis, writing – original draft, resources, and visualisation; Jin Young Kim: investigation, formal analysis, and resources; Du Heon Ha: formal analysis and writing – review & editing; Ga Yoon Chae: investigation and writing – review & editing; Dong Hwan Wang: conceptualisation, supervision, funding acquisition, project administration, writing – review & editing.

Data availability

The data that support the findings of this study are available from the corresponding author, author initials, upon reasonable request.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT (MSIT) (no. 2023R1A2C2008021 and RS-2023-00217270). This work was also supported by the Technology Innovation Program (no. 20021915, ‘Development on Nanocomposite Material of Optical Film (GPa) for Foldable Devices’) funded by the Ministry of Trade, Industry & Energy (MOTIE, Korea).

References

1. K. L. Wang, Y. H. Zhou, Y. H. Lou and Z. K. Wang, Chem. Sci., 2021, 12, 11936–11954.
2. W. Chen, K. T. Mularso, B. Jo and H. S. Jung, Mater. Horiz., 2025 10.1039/D5MH00133A.
3. M. Pirc, Ž. Ajdič, D. Uršič, M. Jošt and M. Topič, ACS Appl. Energy Mater., 2024, 7, 565–575.
4. D. Bogachuk, S. Zouhair, K. Wojciechowski, B. Yang, V. Babu, L. Wagner, B. Xu, J. Lim, S. Mastroianni, H. Pettersson, A. Hagfeldt and A. Hinsch, Energy Environ. Sci., 2020, 13, 3880–3916.
5. Z. Guo, A. K. Jena and T. Miyasaka, ACS Energy Lett., 2023, 8, 90–95.
6. L. Zhu, H. Cao, C. Xue, H. Zhang, M. Qin, J. Wang, K. Wen, Z. Fu, T. Jiang, L. Xu, Y. Zhang, Y. Cao, C. Tu, J. Zhang, D. Liu, G. Zhang, D. Kong, N. Fan, G. Li, C. Yi, Q. Peng, J. Chang, X. Lu, N. Wang, W. Huang and J. Wang, Nat. Commun., 2021, 12, 5081–5088.
7. D. Zheng, F. Raffin, P. Volovitch and T. Pauporté, Nat. Commun., 2022, 13, 6655–6666.
8. X. Zhang, Y. Liu, X. Chen, C. Zhang, J. Wang, J. Cheng and J. Li, Phys. Chem. Chem. Phys., 2024, 26, 28715–28723.
9. Y. Cheng, Q. Wei, Z. Ye, X. Zhang, P. Ji, N. Wang, L. Zan, F. Fu and S. Liu, Sol. RRL, 2022, 6, 2200418.
10. J. Ge, R. Chen, Y. Ma, Y. Wang, Y. Hu, L. Zhang, F. Li, X. Ma, S. W. Tsang, J. You, A. K. Y. Jen and S. F. Liu, Angew. Chem., Int. Ed., 2024, 63, e202319282.
11. X. Chen, W. Cai, T. Niu, H. Wang, C. Liu, Z. Zhang, Y. Du, S. Wang, Y. Cao, P. Liu, W. Huang, C. Ma, B. Yang, S. Liu and K. Zhao, Energy Environ. Sci., 2024, 17, 6256–6267.
12. Z. Wang, X. Duan, J. Zhang, W. Yuan, D. Qu, Y. Chen, L. He, H. Wang, G. Yang, W. Zhang, Y. Bai and H. M. Cheng, Commun. Mater., 2024, 5, 131–148.
13. L. Zhang, H. Zhou, Y. Chen, Z. Zheng, L. Huang, C. Wang, K. Dong, Z. Hu, W. Ke and G. Fang, Nat. Commun., 2024, 15, 1609.
14. D. Wu, Y. Xu, H. Zhou, X. Feng, J. Zhang, X. Pan, Z. Gao, R. Wang, G. Ma, L. Tao, H. Wang, J. Duan, H. Wan, J. Zhang, L. Shen, H. Wang and T. Zhai, InfoMat, 2022, 4, e12320.
15. S. Chen, X. Xiao, B. Chen, L. L. Kelly, J. Zhao, Y. Lin, M. F. Toney and J. Huang, Sci. Adv., 2021, 7(4), eabb2412–eabb2419.
16. D. Dou, H. Sun, C. Li, S. Gan and L. Li, Adv. Funct. Mater., 2024, 34, 2314398.
17. L. Sun, J. Wang, H. Matsui, S. Lee, W. Wang, S. Guo, H. Chen, K. Fang, Y. Ito, D. Inoue, D. Hashizume, K. Mori, M. Takakuwa, S. Lee, Y. Zhou, T. Yokota, K. Fukuda and T. Someya, Sci. Adv., 2024, 10, 9460.
18. J. Jin, H. Li, C. Chen, B. Zhang, L. Xu, B. Dong, H. Song and Q. Dai, ACS Appl. Mater. Interfaces, 2017, 9, 42875–42882.
19. G. Woong Shim, W. Hong, J.-H. Cha, J. Hwan Park, K. Jae Lee, S.-Y. Choi, G. W. Shim, W. Hong, J. Cha, S. Choi Graphene, J. H. Park and K. J. Lee, Adv. Mater., 2020, 32, 1907166.
20. O. E. Solis, M. Mínguez-Avellán, P. F. Betancur, R. I. Sánchez- Alarcón, I. Rodriguez, J. P. Martínez-Pastor, T. S. Ripolles, R. Abargues and P. P. Boix, ACS Energy Lett., 2024, 5288–5295.
21. Y. Zhang, C. Li, H. Zhao, Z. Yu, X. Tang, J. Zhang, Z. Chen, J. Zeng, P. Zhang, L. Han and H. Chen, Nat. Commun., 2024, 15, 6887.
22. C. Yang, W. Hu, J. Liu, C. Han, Q. Gao, A. Mei, Y. Zhou, F. Guo and H. Han, Light: Sci. Appl., 2024, 13, 227.
23. C. Yang, W. Hu, J. Liu, C. Han, Q. Gao, A. Mei, Y. Zhou, F. Guo and H. Han, Light: Sci. Appl., 2024, 13, 227.
24. B. Wenger, P. K. Nayak, X. Wen, S. V. Kesava, N. K. Noel and H. J. Snaith, Nat. Commun., 2017, 8, 590.
25. J. W. Lee, S. Tan, T. H. Han, R. Wang, L. Zhang, C. Park, M. Yoon, C. Choi, M. Xu, M. E. Liao, S. J. Lee, S. Nuryyeva, C. Zhu, K. Huynh, M. S. Goorsky, Y. Huang, X. Pan and Y. Yang, Nat. Commun., 2020, 11, 5514.
26. S. Yuan, D. Zheng, T. Zhang, Y. Wang, F. Qian, L. Wang, X. Li, H. Zheng, Z. Diao, P. Zhang, T. Pauporté and S. Li, Nat. Commun., 2025, 16, 2052.
27. Z. Gao, H. Zhou, K. Dong, C. Wang, J. Wei, Z. Li, J. Li, Y. Liu, J. Zhao and G. Fang, Nano-Micro Lett., 2022, 14, 215.
28. W. Wu, J. Zhang, C. Liu, J. Zhang, H. Lai, Z. Hu and H. Zhou, Adv. Sci., 2024, 11, 2410303.
29. G. Zheng, C. Zhu, J. Ma, X. Zhang, G. Tang, R. Li, Y. Chen, L. Li, J. Hu, J. Hong, Q. Chen, X. Gao and H. Zhou, Nat. Commun., 2018, 9, 2793.
30. S. Yuan, D. Zheng, T. Zhang, Y. Wang, F. Qian, L. Wang, X. Li, H. Zheng, Z. Diao, P. Zhang, T. Pauporté and S. Li, J. Mater. Chem. C, 2024, 12, 7562–7571.
31. B. Hwang, Y. Park and J. S. Lee, J. Mater. Chem. C, 2021, 9, 110–116.
32. C. Bi, Q. Wang, Y. Shao, Y. Yuan, Z. Xiao and J. Huang, Nat. Commun., 2015, 6, 7747.
33. Z. Ni, S. Xu, H. Jiao, H. Gu, C. Fei and J. Huang, Sci. Adv., 2022, 8(36), eabq8345–eabq8354.
34. M. Nukunudompanich, G. Budiutama, K. Suzuki, K. Hasegawa and M. Ihara, CrystEngComm, 2020, 22, 2718–2727.
35. Z. Li, S. P. Senanayak, L. Dai, G. Kusch, R. Shivanna, Y. Zhang, D. Pradhan, J. Ye, Y.-T. Huang, H. Sirringhaus, R. A. Oliver, N. C. Greenham, R. H. Friend, R. L. Z. Hoye, Z. Li, L. Dai, R. Shivanna, Y. Zhang, J. Ye, Y.-T. Huang, H. Sirringhaus, N. C. Greenham, R. H. Friend, S. P. Senanayak, D. Pradhan, G. Kusch, R. A. Oliver and R. L. Z. Hoye, Adv. Funct. Mater., 2021, 31, 2104981.
36. J. Li, H. Li, D. Ding, Z. Li, F. Chen, Y. Wang, S. Liu, H. Yao, L. Liu and Y. Shi, Nanomaterial, 2019, 9, 505.
37. E. M. Miller, Y. Zhao, C. C. Mercado, S. K. Saha, J. M. Luther, K. Zhu, V. Stevanović, C. L. Perkins and J. Van De Lagemaat, Phys. Chem. Chem. Phys., 2014, 16, 22122–22130.
38. M. Hosseinzadeh Dizaj, S. C. Shishvan and F. Shahnavaz, DOI:10.2139/SSRN.4991164.
39. J. S. W. Godding, A. J. Ramadan, Y. H. Lin, K. Schutt, H. J. Snaith and B. Wenger, Joule, 2019, 3, 2716–2731.
40. H. Zhou, Z. Song, C. R. Grice, C. Chen, J. Zhang, Y. Zhu, R. Liu, H. Wang and Y. Yan, Nano Energy, 2018, 53, 880–886.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5tc01294e

---