1、 Ion migration and hysteresis in perovskite batteries
With the improvement of photovoltaic (PV) power conversion efficiency, metal halide perovskite (MHPs) have attracted great interest in various fields of solid-state (photoelectric) electronics. MHPs are soft ion semiconductors that exhibit extensive ion migration, which directly leads to their degradation and operational defects such as hysteresis in optoelectronic devices. An increasing number of studies have linked the improved stability of the new generation of MHPs to a decrease in halide ion migration. Then a challenge arises, which is that most MHP devices with excellent stability simultaneously exhibit reduced hysteretic behavior. Stability and current voltage hysteresis are the main obstacles to the commercialization of metal halide perovskite. Both of these phenomena are related to ion migration, and there is anecdotal evidence that stabilizing devices can produce low hysteresis. However, the underlying mechanism of complex stability lag linkages remains elusive.
2、 Achievement Introduction
With this in mind, the University of North Carolina's Aram Amasian team proposed a multiscale diffusion framework that describes vacancy mediated diffusion of halides in polycrystalline metal halide perovskite, distinguishing between fast grain boundary (GB) diffusivity and slow volume diffusivity of 2-4 orders of magnitude. The inverse relationship between grain boundary activation energy and volume diffusion has been revealed. For example, stable metal halide perovskite exhibits a smaller volume diffusion coefficient (DV), which is related to a larger grain boundary diffusion coefficient (DGB) and a smaller hysteresis. The clarification of multiscale diffusion of metal halides in metal halide perovskites reveals the complex internal coupling between ion migration and grain boundaries in particle volume, which in turn can predict the stability and hysteresis of metal halide perovskites, providing a clearer path to address outstanding challenges in this field.
Figure 1 Schematic diagram of multiscale diffusion in polycrystalline solids, and quantification and modeling of lateral diffusion profiles
According to the Harrison classification, the typical concentration distributions observed in polycrystalline metals or ionic crystals can be divided into three main types (types A, B, and C) as shown in Figure 1. Among the three diffusion modes, type B and type C are profiles that allow volume diffusion and GB diffusion to be decoupled. B-type and C-type spectra have not been reported in MHPs, and due to the high diffusivity of halide ions in MHPs, they are not easily generated in vertical thin film stacks, while MHPs require samples tens to hundreds of microns thick, making accurate quantification of DV and DGB extremely challenging.
3、 Results and Discussion
Point 1: Quantifying Volume Diffusion
Firstly, the volume diffusion characteristics of alien halides (I − and Br −) in different MHPs were clarified through quantitative analysis of the B-C type cross section. The complementary error function (ERFC) solution for volume diffusion can be used to analyze the volume diffusion profile of external ions. Figures 2a and d show 2D and ID diffusion profiles of Br - in MAPbl3, respectively. The diffusion of foreign halides into the lattice of MHP grains leads to a change in the band gap of the MHP layer itself, which can be achieved by performing micro PL on the lateral diffusion edge( μ PL) scan (Figure 2b, c, e) to further demonstrate the volume diffusion of Br - in MAPbl3 particles. The gradient of emission wavelength extends from the apparent edge to - 25 um, confirming that the high concentration region of the complementary error function (ERFC) profile for volume diffusion is dominated by volume diffusion.
Figure 2 Volume diffusion of metal halide perovskite
Figures 2g and h show temperature dependent diffusion profiles of Br − in MAPbI3, and examples of halide diffusion (I − or Br −) in MAPbBr3, MAPbI3, FAPbBr3, and FACsRbI, respectively. Figure 2i summarizes the DV values of the inverse temperature for all investigated MHP systems. The study found that the DV of halides has an Arrhenius temperature dependence, with the fastest volumetric diffusion rate in MAPbI3. At all temperatures studied, the DV of halides was six to nearly ten times smaller in FA based single cation and mixed cation MHPs. In addition, in these systems, the DV of I − is smaller than the DV of Br −. Importantly, temperature correlation analysis of all systems showed that EV had the smallest Br − diffusion in MAPbI3 (0.61 ± 0.02 eV). On the other hand, the largest EV was obtained in FAPbBr3 (0.74 ± 0.03 eV), followed by FACsRbI (0.72 ± 0.02 eV). These changes explain the differences in the diffusion behavior of I − and Br − and greatly facilitate the diffusion of halides in MA based systems rather than FA based systems.
Point 2: Quantifying GB diffusion
The rapid diffusion of ions has proven to be one of the main criteria for non lagging photovoltaic applications. Although in most cited cases, the diffusion coefficient is not directly related to GB diffusion, as determined by the study, the DGB at room temperature is very consistent with the diffusion coefficient previously reported at room temperature: transient electrical measurements of~10-9 to 10 − 11 cm2 s-1. However, the diffusion amount reported in the literature based on transient measurements is driven by the net mass transfer of ions in the presence of an external electric field. However, the diffusion coefficient of particle volume, which is three to five orders of magnitude smaller than GB (DGB DV), can explain why ion transport in transient electrical measurements should be dominated by GB.
Figure 3. GB diffusion as a function of temperature, halide type, and composition
When plotting room temperature DGB and DV on different MHP systems (Figure 3c), a special inverse relationship occurs. This relationship is transformed into a linear relationship between their respective activation energies. From these relationships, it can be seen that among all investigated MHPs, the fastest volumetric diffusion (smallest EV) and the slowest GB diffusion (smallest DGB and largest EGB) belong to the MAPbI3 system. In contrast, the slowest volume diffusion (the largest EV) and the fastest GB diffusion (the smallest EGB) belong to FAPbBr3.
Point 3: Linking multiscale diffusion to stability and hysteresis
Due to the role of halide vacancies as the reaction medium for superoxide formation and subsequent MHP photodegradation, the rate of photoinduced oxygen degradation can be used as an indirect substitute for halide diffusion, characterizing the oxygen induced photodegradation of different MHP films and corresponding PVs. As shown in Figure 4a, MAPbI3 (60 ° C=8.4 × DV at 10 − 13 cm2 s-1) drops very quickly in the presence of light and dry air( τ = 1.52 h, where τ Is a degradation time constant), while FACsRbI (DV at 60 ° C=1.5 × 1013 cm2 s-1) slow degradation rate( τ = 10.46 h) and partially maintain its dark perovskite phase for more than 24 hours after light irradiation at 60 ° C. The faster degradation rate of MAPbI3 is attributed to the higher vacancy density of the system, which promotes the diffusion of halide ions and oxygen, thereby mediating the formation rate of superoxide species in most MHPs. Figures 4b and c show schematic diagrams of MHPs for fast DV (high degradation rate) and slow DV (low degradation rate).
Figure 4 Perovskite degradation under light and O2
Although the lower DV of FA-based systems explains their improved optical stability, the nearly one order of magnitude difference in DGB between MA and FA based systems at room temperature also explains the decrease in hysteresis behavior of the two MHP devices compared to each other. Compared to MAPbI3, the hysteresis index of PV based on mixed cations calculated in forward voltage scanning is smaller, which also proves that the ion DGB in FA based systems is larger. The results show that the hysteresis index can be further reduced by preparing photovoltaic devices based on mixed halide compositions.
Figure 5 Current density versus voltage curve of MHP photovoltaic, halide exchange in MHP, and GB "intensity" model
In addition, the author has developed a simple photostability and hysteresis pre screening method using mature axial mutual diffusion geometry (two MHPs sandwiched together with different components). In Figure 5b, it can be observed that the homogenization time of the MAPBI3/MAPbBr3 bimolecular layer (large DV) is much faster than that of the FACsRbI/FAPbBr3 (small DV). This indicates that the halide exchange rate is limited by volume diffusion rather than GB diffusion. Therefore, this simple method can help predict the photostability of polycrystalline MHPs relative to MA and FA based MHPs.
Point 4: GB "Strength" model
Based on the above research, the author suggests combining EV and EGB( Δ E) The difference between the two serves as an indicator of the similarity of halide diffusion barriers between the crystal interior and GBs, resulting in hysteresis and optical stability. Smaller Δ E is equivalent to "weaker" GBs (Figure 5c, d), which are more prone to halide exchange between grains and their boundaries. In contrast, "stronger" GBs preferentially promote the movement of halides along the GBs rather than occupying vacancies within the particles (FIGS. 5e, f). Δ E (Figure 3d) is the largest in a FA based system, approaching zero in MAPbl3. This result indicates that in MAPbl3, volume and GBs have similar barrier effects on ion diffusion. The author believes that a decrease in EV (for example, in MAPbl3), coupled with a preference for positive charge vacancies to accumulate in GBs, can lead to a "weakening" of GBs( Δ E→O)。 Therefore, the reverse relationship between EV and EGB, especially Δ The decrease in E from 0.25 eV for FA-based systems to<0.05 eV for MAPbl3 is due to a loss of preference for ions to move along GB vacancies and rapid movement through particle volume.
The large concentration of halide vacancies in MHPs is also a pathway for oxygen diffusion and a catalyst for the formation of superoxide under light, which leads to faster photodegradation of MHPs exhibiting larger DVs and weaker GBs. In contrast, FA-based systems exhibit the strongest GBs, which establishes an energy barrier between GB and particle volume, prevents halide exchange, and reduces the photodegradation rate by limiting the entry of oxygen within the volume. The characteristic of strong GBs is that the contrast ratio of diffusivity is very large (DGB/DV ≈ 104-105), which is conducive to reducing hysteresis. The rapid GB and slow volume diffusion pathways in polycrystalline MHPs, as well as the direct evaluation of the volume diffusion of MAPbBr3 single crystals, should be able to quell the ongoing debate about the role of GBs in the diffusion properties of MHPs.
4、 Summary
In summary, the author's direct observation of multiscale halide ion diffusion associates rapidly moving ions in the device with ions diffusing along GBs, and phenomena such as photoinduced MHPs degradation can be quantitatively described by the slow diffusion of halide ions and oxygen mediated by vacancies in particle volume. The results show that small volume and large GB diffusion coefficient are necessary combinations for achieving stable and non hysteresis MHP devices, which requires strong GBs to limit fast ion motion within GB. The GB strength model shows the material composition and the working principles of passivation methods, and provides guidance for designing powerful GBs to develop ultra stable and non hysteretic MHP devices. The multiscale diffusion framework will also help to elaborate accurate models of MHPs ion migration in a wide range of scenarios, including devices under external stress and different applications.
5、 References
Masoud Ghasemi. et al. A multiscale ion diffusion framework sheds light on the diffusion–stability–hysteresis nexus in metal halide perovskites.
Doi: 10.1038/s41563-023-01488-2(2023)
https://www.nature.com/articles/s41563-023-01488-2