Abstract

Grain boundaries (GBs) in polycrystalline perovskite films play a crucial role in determining photogenerated carrier transport and recombination, thereby impacting both efficiency and stability of perovskite solar cells (PSCs). Despite extensive research into grain boundary engineering to optimize PSC performance, the specific mechanisms through which GBs influence carrier dynamics remain unclear and highly debated. Here, we employ high-resolution, intensity-dependent fluorescence lifetime imaging microscopy (FLIM) to systematically investigate the behaviors of different types of GBs in hybrid perovskite films under varying light conditions. Our analysis reveals three distinct categories of GBs: I-type, which remains invisible at low excitation intensities and exerts minimal influence on carrier transport; W-type, characterized by a W-shaped lifetime profile at high light intensities, suggesting significant carrier scattering at the boundary; and V-type, marked by a V-shaped lifetime profile, indicating a more specialized role in regulating carrier dynamics. These findings provide critical insights into how various GBs modulate photogenerated carrier behavior, offering a new framework for understanding their diverse impacts on the optoelectronic properties of perovskite films. Our results underscore the importance of targeted grain boundary engineering strategies to advance the design and optimization of next-generation perovskite-based photonic and optoelectronic devices.

Results and Discussion

Click to copy section linkSection link copied!

---

We synthesized organic–inorganic hybrid perovskite films of (FA_0.78Cs_0.22) Pb(I_0.82B_r0.15Cl_0.03)₃ using the antisolvent method, following established protocols. (32−34) These wide bandgap perovskite films are widely used as absorption layers in high-performance solar cells. (35,36) Before conducting FLIM measurements, we characterized the films using a range of experimental techniques, including PL spectroscopy, ultraviolet–visible (UV–vis) absorption spectroscopy, scanning electron microscopy (SEM), and Fourier transform infrared spectroscopy (FTIR). The PL and UV–vis absorption spectra, shown in Figure 1c, confirm that the fluorescence properties of the samples are ideal, with the bandgap values consistent with typical perovskite materials. SEM images (Figure S1) reveal relatively large grain sizes, indicating well-formed films. FTIR results (Figure S2) show minimal impurities, suggesting a high level of purity in the perovskite films. These combined results demonstrate that our samples are well-prepared and in optimal condition for further experimentation.

Figure 1

High Resolution Image

Download MS PowerPoint Slide

We proceeded with high-resolution, intensity-dependent FLIM on these high-quality perovskite films, as illustrated in Figure 1a. A pulsed laser was focused onto the perovskite surface, and the emitted PL was detected and converted into pulsed electrical signals. These signals were processed using time-correlated single-photon counting (TCSPC) electronics, generating PL decay curves. To perform intensity-dependent experiments, two variable neutral density (ND) filters were utilized: ND filter 1 systematically modulated the laser intensity, while ND filter 2 minimized pile-up effects. (37) By raster scanning the laser across the film, we acquired PL decay curves at each pixel, resulting in two sets of images: PL intensity and PL lifetime. (8,21) PL intensity images depict the spatial distribution of PL emission across the sample, providing insights into the quantum yield of different regions. By contrast, PL lifetime images map the temporal decay of PL, offering detailed information on carrier recombination dynamics and interactions with defects or grain boundaries. Figure 2b presents a typical FLIM measurement of a polycrystalline perovskite film at an excitation intensity of 6000 mW/cm², where the PL intensity and lifetime images are overlaid. The separate PL intensity and lifetime images are shown in Figure S3.

Figure 2

High Resolution Image

Download MS PowerPoint Slide

Figure 2a–c presents intensity-dependent PL intensity mappings. As the excitation light intensity increases, the contrast in fluorescence intensity between different grains decreases significantly. This behavior aligns with previous studies and can be attributed to changes in carrier dynamics at higher excitation intensities. (38) At low light intensity, fluorescence intensity varies significantly between grains due to heterogeneous trap-mediated monomolecular nonradiative recombination. In contrast, at high light intensity, diffusive behavior becomes dominant, and nonradiative recombination pathways reach saturation, reducing fluorescence intensity differences across grains. Furthermore, both PL intensity and lifetime evolve near the grain boundaries (GBs) with varying excitation intensities. By comparing intensity-dependent FLIM images, we observed distinct behaviors at grain boundaries, with three typical examples─GB1, GB2, and GB3─highlighted by white dashed circles in Figure 2c. Unlike GB2 and GB3, GB1 was invisible at low light intensities but became increasingly distinct as the intensity increased. In the corresponding PL lifetime images (Figure 2d–f), GB1 and GB2 exhibited longer lifetimes compared to their neighboring regions, whereas GB3 showed a shorter lifetime.

To further analyze these trends, we performed line profile analyses for the three GBs, as marked by white dashed boxes in Figure 2f. As shown in Figure 3a–c, under low-intensity excitation, the profile of GB1 remained nearly flat near the boundary, while GB2 and GB3 displayed local minima. We categorize GB1 as an I-type GB, which remains invisible at low excitation intensities. In Figure 3d–f, the PL lifetime profiles of GB1 and GB2 developed peaks near the GBs. Notably, GB2’s profile in Figure 3e gradually transformed into a W shape, identifying it as a W-type GB. In contrast, GB3 consistently showed a V-shaped lifetime profile, classifying it as a V-type GB. Interestingly, the local minimum in the lifetime profile of the V-type GB did not align with the intensity profile’s local minimum, indicating an asymmetric impact on carrier dynamics. Through the combination of PL intensity and PL lifetime data under varying light intensities, we identified three distinct types of GBs: I-type, W-type, and V-type. To further validate our findings, we conducted intensity-dependent FLIM on additional areas, as illustrated in Figures S4 and S5, and performed another set of line profile analyses presented in Figure S6. These profiles align with the descriptions of the GB types discussed in Figure 3 and confirm the reliability of our categorization. Among the clearly resolved GBs in Figures 2, S4 and S5, we found that W-type GBs were the most prevalent (66.2%), followed by I-type GBs (26.5%), while V-type GBs were the least common (7.3%) as summarized in Figure S7.

Figure 3

High Resolution Image

Download MS PowerPoint Slide

The intensity-dependent FLIM results suggest that grain boundaries (GBs) may undergo changes due to variations in photogenerated carrier concentration. A plausible explanation for this behavior involves trap filling and the formation of barriers at the GBs, as discussed in previous studies. (11,29,39−41) This process is illustrated in Figure 4a, where photogenerated carriers fill traps at the GB between two grain interior (GI) regions, forming a barrier that induces carrier scattering.

Figure 4

High Resolution Image

Download MS PowerPoint Slide

Two primary sources contribute to these energy barriers: the intrinsic barrier at the GBs and the trap-induced barrier. (16,42) The model proposed for the three types of GBs incorporate both intrinsic and trap-induced barriers, as shown in Figure 4b–d. For I-type GBs, the intrinsic barrier is negligible (Figure 4b), so under low excitation intensity, these GBs have minimal influence on carrier transport, rendering them invisible. As the excitation intensity increases, the barrier induced by trapped photogenerated carriers becomes more pronounced, making I-type GBs visible. Overall, I-type GBs play a neutral role in affecting PSC performance under typical operating conditions.

In the case of W-type GBs, there is a significant intrinsic barrier, making these GBs visible even at low intensities (Figure 4c). As the excitation intensity increases, the trap-induced barrier further enhances carrier scattering. The traps at W-type GBs are charged by photogenerated carriers, increasing the barrier height and scattering effect. This enhanced scattering is indicated by the PL intensity and lifetime profiles under the highest intensity conditions (Figure 3a,d). The M-shaped PL intensity profile and W-shaped PL lifetime profile suggest that carriers scattered back from the GBs enhance radiative recombination near the GBs, leading to increased PL intensity and decreased PL lifetime. W-type GBs could be detrimental for lateral optoelectronic devices due to obstructed lateral transport, but under certain circumstances become beneficial when enhanced vertical transport is preferred.

For V-type GBs, an interesting asymmetry is observed, as the local minima in the PL intensity and lifetime profiles (Figure 3c,f) do not align. By comparing V-type and W-type GBs, we see that the right side of both exhibits similar behavior, but the left side of V-type GBs lacks the local minimum associated with carrier scattering. This suggests that V-type GBs have an asymmetric effect on carrier transport. We propose this asymmetry arises from an energy diagram created by a built-in electric field at V-type GBs (Figure 4d), where the intrinsic barrier predominantly affects carrier transport in one direction. V-type GBs can play a more specialized role in regulating carrier dynamics and may be beneficial or detrimental, depending on the processes involved.

Our results demonstrate that intensity-dependent FLIM effectively differentiates between various types of GBs and reveals their distinct effects on photogenerated carrier transport and dynamics. However, the detailed microscopic structure–property-performance relationship of these GBs remains unclear. Future experiments will combine intensity-dependent FLIM with advanced microscopy techniques, such as scanning transmission electron microscopy (STEM), nano X-ray diffraction (nano-XRD), and multimodal atomic force microscopy (AFM). These correlative microscopy characterizations will help elucidate the unique atomic structures of different GB types, local strain profiles, and their varied photophysical properties, providing deeper insights into their impact on carrier dynamics. (30,43−46) This understanding will enable tailored grain boundary engineering to improve the efficiency and longevity of perovskite-based photonic and optoelectronic devices.