First-Principles Study of Halide Modulation on Deep-Level Traps in FAPbI₃

Abstract

In this study, we investigate the influence of the halogen elements bromine (Br) and chlorine (Cl) on iodine defect properties primarily in FAPbI₃ through first-principles calculations, aiming to understand the effect of high defect densities on the efficiency of organic–inorganic hybrid perovskite cells. The results indicate that Br and Cl interstitials minimally alter the overall band structure of FAPbI₃ but significantly modify the defect energy levels. Br and Cl interstitials, with defect states closer to the valence band and lower formation energies, effectively convert deep-level traps induced by iodine interstitials (I_i) into shallow-level traps. This conversion enhances carrier transport by reducing non-radiative recombination while preserving light absorption efficiency. Excess Br/Cl co-doping in FAPbI₃ synthesis thereby suppresses non-radiative recombination and mitigates the detrimental effects of iodide-related defects.

1. Introduction

In recent years, solar cells based on halide perovskites, such as XPbI₃ (X = Cs, CH₃NH₃, NH₂CHNH₂, etc.) have received widespread attention because of their advantages such as adjustable bandgap, high light absorption rate, low carrier recombination rate, and high photoelectric conversion efficiency (PCE) [1,2,3], which facilitate efficient charge collection and achieve higher PCE [4]. Due to these favorable optoelectronic properties, the PCE of single-junction perovskite solar cells (PSCs) has surged remarkably from 3.8% to over 26% within the past decade [5,6,7]. Among them, MAPbI₃ (CH₃NH₃PbI₃) and FAPbI₃ (NH₂CHNH₂PbI₃) have been the most studied. Compared with MAPbI₃, FAPbI₃ has a more symmetrical crystal structure and a more optimum theoretical bandgap of 1.48 eV [8], approaching the Shockley–Queisser limit for single-junction cells (approximately 33.7%) [5,6,9]. Moreover, FAPbI₃ has better thermal stability and is not easily degraded [7]. Although the photoactive phase of FAPbI₃ is unstable at operating temperatures and tends to transform into the photoinactive yellow phase [8], it remains a highly promising optoelectronic material because stabilization is achievable through doping with Zn (II) cations and formate anions, along with other strategies [10].

Despite rapid efficiency improvements, FAPbI₃ faces a significant challenge in high intrinsic defect density. The soft lattice of synthesized halide perovskite solutions and films [5], combined with tilted FA cations within the relaxed PbI₆ framework [11,12], induces strong lattice distortion that inevitably generates high defect concentrations [13,14,15]. While shallow defects dominate in halide perovskites [16], recent studies confirm the coexistence of detrimental deep-level traps in FAPbI₃ [12,16,17,18]. The deep energy levels inside the bandgap usually have high ionization energy [19] and act as potent non-radiative recombination centers. By capturing charge carriers, they reduce carrier lifetimes, degrade open-circuit voltage (Voc), and ultimately limit device performance [20,21]. Consequently, even with advanced crystal growth techniques, FAPbI₃ polycrystalline films exhibit high defect densities (9.6 × 10¹⁵~1.37 × 10¹⁶ cm⁻³) [22], exceeding those in classical semiconductors (typically 10⁸–10¹⁵ cm⁻³) [23,24,25,26] by orders of magnitude. This severely suppresses device efficiency. Although FAPbI₃ exhibits some defect tolerance [27,28], strategies to mitigate defects remain crucial for further performance gains. The high defect density issue can currently be addressed through light soaking, chemical passivation, interface engineering, and microstructure modulation methods [29]. Recent studies demonstrate that doping and passivation with homologous halide ions effectively enhances halide perovskite performance. Mu et al. [30] found that 20% chlorine doping increases FAPbI₃ power conversion efficiency. Energy-dispersive X-ray analysis confirms that the film’s chlorine contents match the precursor solution ratio, indicating precise chlorine doping improves FAPbI₃ properties. Multi-halide co-doping shows significant performance optimization effects. Mariotti et al. [31] achieved 1.28 eV open-circuit voltage in single-junction cells and 32.5% certified tandem efficiency using triple-halide (I, Br, and Cl) perovskite with piperazine iodide interface engineering for perovskite–silicon tandem solar cells. Xu et al. [32] resolved phase separation through trihalide incorporation and established Br-mediated I-Cl lattice connections, realizing bulk chlorine doping that doubled carrier mobility and lifetime.

Current understanding remains limited regarding defect property modulation in halide perovskites through halide doping, and quantitative analyses of carrier transport and recombination effects are necessary. This study employs first-principles calculations to investigate intrinsic point defects in FAPbI₃ and interstitial defect properties of bromine, chlorine, and fluorine. The calculations include defect formation energies, transition levels, and migration. The results show bromine and chlorine interstitial defects exert less influence on FAPbI₃ performance than intrinsic iodine interstitials, accompanied by decreased migration rates. These defect property changes provide insights into FAPbI₃ performance enhancement via halide doping.

2. Calculation Methods

This paper employs the Vienna ab initio Simulation Package (VASP 5. 3. 5), a computational software based on density functional theory [33]. The initial structure of FAPbI₃ was constructed using experimentally reported perovskite geometries [34,35]. In examining the defect formation energy (DFE) of point defects in FAPbI₃, a 2 × 2 × 2 supercell is employed (as shown in Figure S1), with a point defect formed into the supercell to facilitate the averaging of defect volume [36]. A plane–wave energy cutoff of 500 eV and a 5 × 5 × 5 k-point grid using the Monkhorst–Pack method were employed in the calculations. Structural relaxations were conducted until forces on atoms fell below 0.01 eV/Å, and energy convergence reached 1 × 10⁻⁶ eV. To account for van der Waals interactions, the rev-vdW-DF2 functional proposed by Hamada [37] was employed, ensuring full optimization of atomic positions and lattice parameters.

The effect of spin–orbit coupling (SOC) has been previously reported in Pb-based perovskites, but the predicted bandgap is significantly lower than the experimental value [38]. To obtain accurate bandgap values for qualitative analysis, we employed the G₀W₀ method with spin–orbit coupling (SOC) effects included [39,40], and the calculated bandgap is 1.50 eV, consistent with experimental values [41]. For the calculation of charged defects, it is assumed that the additional or missing electrons from charged defects come from an external electron reservoir or enter the supercell interior with their energy equivalent to the Fermi energy, using Equation (1) [42]:

$$\mathrm{DFE}=E\left(\mathrm{defect}\right)-E(\mathrm{prefect})+{\displaystyle \sum _i{n}_i{\mu }_i}+q(E_F+E_{VBM})+E_{corr}$$

(1)

where $E\left(\mathrm{defect}\right)$ represents the system containing defects, $E(\mathrm{prefect})$ is the energy of the perfect system without defects, and n and μ are the number of atoms and the corresponding chemical potentials individually. For vacancy defects, n_i > 0, and interstitial defects, n_i < 0. In order to ensure the formation of perovskite, the chemical potentials of FA, Pb, and I satisfy the following relationship:

$${\mu }_{FA}+{\mu }_{Pb}+3{\mu }_I=\Delta H\left({\mathrm{FAPbI}}_3\right)$$

(2)

Here, $\Delta H$ represents the formation energy of perovskite in the cubic phase. We selected the cubic phase of Pb solid, the orthorhombic phase of I₂ solid, and the cubic phase of FA to calculate their chemical potentials, respectively. In order to prevent the formation of FAI and PbI₂ as by-products, it is necessary to ensure that the chemical potential values satisfy the following conditions:

$${\mu }_{FA}+{\mu }_I<\Delta H(\mathrm{FAI})$$

(3)

$${\mu }_{Pb}+2{\mu }_I<\Delta H{(\mathrm{PbI}}_2)$$

(4)

In accordance with Equations (2)–(4), the chemical potential regions under equilibrium growth conditions for the perovskite were calculated. In this research, the moderate chemical potentials of FA, Pb, and I were selected for calculating DFE (details of the chemical potential calculations are provided in Figure S2 and Table S1). In the last term $q(E_F+E_{VBM})+\Delta V$, q represents the charge of the defect; E_VBM represents the energy level at the top of the valence band, and E_F is the Fermi level at the top of its relative valence band. At the top of the valence band $E_F=0$, the Fermi levels are permitted to fluctuate within the bandgap, with the upper limit of the valence band maximum (VBM) aligning with the lower limit of the conduction band minimum (CBM) in a defect-free FAPbI₃ crystal, which serves as the reference point for the bandgap. $E_{corr}$ is a correction term added to avoid spurious interactions of charged defects between supercells so that the potential energy in the defective supercells corresponds to the bulk potential energy (as shown in Supporting Information) [43,44]. Equation (1) shows that the formation energy of charged defects depends linearly on the Fermi level within the bandgap, with the slope determined by the charge state q. Consequently, defects of the same type can exhibit different formation energies for different charge states. The intersection points of these formation energy lines define the defect transition levels $\epsilon (q_1/q_2)$ between charge states [36]. Additionally, distinct charge states may correspond to distinct local atomic configurations [45,46,47]. Defect transition levels determine electronic behavior and are often used as a basis for experimental detection or identification of defects. Therefore, accurate calculation of the defect transition level is essential for the identification and characterization of defects. The specific calculation can be expressed by the following formula:

$$\epsilon (q_1/q_2)={\displaystyle \frac{DF{E}^{q_1}-DF{E}^{q_2}}{q_2-q_1}}$$

(5)

where $DF{E}^{q_n}$ is the DFE for a charge state of $q_n$. Figure 1 illustrates Equation (5). When the Fermi level lies below (+/0), the $q$ = +1 charge state remains stable. When the Fermi level exceeds (+/0), the $q$ = 0 state becomes stable. The (+/0) position defines the defect transition level (trap level). Similar principles apply to (0/−). The trap level position relative to band edges determines whether defects act as deep-level or shallow-level centers. Deep-level defects form when trap levels reside within the bandgap (mid-gap). These deep centers degrade carrier mobility [48].

3. Results and Discussion

Because organic cations in perovskite make little or no contribution to the electronic states at the edges of the energy bands, here, we ignore the effect of point defects associated with FA cations [39]. The defect formation energies of an iodine vacancy (V_I), iodine interstitial (I_i), lead vacancy (V_Pb), and lead interstitial (Pb_i) have been calculated and are shown in Figure 2a. The lowest formation energy defects are Pb_i and I_i, which represent the fact that Pb_i and I_i are more likely to form and stabilize under thermodynamic equilibrium, implying a higher defect density. Stabilized I_i has a lower DFE. Figure 2a shows that I_i exists in neutral (0), negatively charged (−), and positively charged (+) states, with the stable charge state being the transition from + to −, forming a more stable negatively charged state. Similarly, Pb_i is also a relatively stable type of defect, with the stable charge state being 2+. In addition to these stable defects, the stable charge state for V_I within the bandgap is +, and for V_Pb, it is −/2−, with the charge state reaching the lowest DFE being 2−. From the thermodynamic level, point defects are more prone to forming interstitial defects with lower DFEs. This phenomenon may be attributed to the soft lattice characteristics of FAPbI₃, which facilitate the easier occupation of interstitial positions within the lattice by adding additional atoms [5]. While all defects will influence the position of the Fermi level, their effect will diminish exponentially with increasing DFE. The more stable I_i defect, for instance, affects the intrinsic Fermi level of the perovskite. In addition to considering the number of defects, it is also important to take into account their activity. This refers to the migration rate of the defects, as thermodynamically favorable defects may be limited by high migration activation energies, which prevent their migration. It is commonly accepted that interstitial defects are more prone to migration than vacancy defects [47], and defects associated with iodine exhibit a higher migration rate than those linked to lead [49,50]; thus, it can be reasonably deduced that defects related to iodine ions may become the primary migration targets [51,52,53]. Furthermore, defect transition levels (trap levels) critically influence charge carrier dynamics and photovoltaic performance in perovskite materials. Although some defects exhibit low defect formation energy, their impact on carrier transport depends primarily on trap level positions within the bandgap [54]. These trap levels act as non-radiative recombination centers, while charged defects reduce conductivity through carrier scattering. Crucially, halide perovskites demonstrate good defect tolerance because defect-induced trap levels typically reside near band edges, minimizing disruption to carrier transport.

Figure 2a shows the defect formation energy (DFE); under thermodynamic equilibrium, the stable charge states for V_Pb within the bandgap are −/2−, indicating that the defect can further capture holes through the trap levels shown in Figure 2b, releasing energy, hence the lower DFE for the 2− charge state. Since the neutral charge state of V_Pb is not stable under thermodynamic equilibrium, and the 0/2− and 0/− charge state transition levels are within the valence band, no deep-level traps are formed, meaning that the neutral charge state defect of V_Pb can be neglected. Furthermore, the 2− charge state migration activation energy of V_Pb indicates that metallic lead is only likely to nucleate on the surface under high temperatures or under illumination [55], so the impact of V_Pb on FAPbI₃ is not significant. I_i can capture electrons and holes through the +/0 and 0/− trap levels at 0.89 eV below the conduction band (CB) and 0.34 eV above the valence band (VB), respectively, acting as both a donor and an acceptor within the bandgap. The stable trap levels at the thermodynamic level, as illustrated in Figure 2a, are in close proximity to the Fermi level of FAPbI₃. This indicates that the trap level exerts a considerable influence on the electronic states situated in proximity to the Fermi surface. Thus, FAPbI₃ performance degrades significantly because all trap levels reside within the bandgap, and deep-level traps actively capture carriers, reducing carrier lifetime. As shown in Figure 2b, all trap levels of Pb_i and V_I are located inside the CB, and there are no trap levels inside the bandgap, which makes it difficult to form a recombination center and has less effect on the carrier transport. Moreover, the +/0 trap level of V_I and Pb_i show shallow trapping behavior at the bottom of the CB, with some polarization around the defect location, and have a resonance with the CB, so that despite the migration activation energy being low, it can still act as a shallow-level trap [56]. From the above conclusion, I_i defects can be considered as the most likely point defect to exist stably and trap carriers because they have the lowest DFE with deep-level traps.

To further investigate the effect of I_i on FAPbI₃, we calculated the density of states in the case of I_i, and the density of states calculation from Figure 3a shows that the conduction band of FAPbI₃ is mainly contributed by electrons in the 6p orbital of Pb atoms, while the valence band is mainly contributed by electrons in the 5p orbital of I atoms. Other elements introduce deep-level states within the VB and CB. These states do not participate in bonding and contribute negligibly to the density of states near the band edges.

As shown in Figure 3b, the different charge states of I_i form traps inside the bandgap with different energy level positions, ${\mathrm{I}}_{\mathrm{i}}^0$ creates a new trap level within the bandgap that can trap either holes or electrons, but according to the Pauli exclusion principle, these two processes cannot occur at the same time, and since its location is closer to the VB, hole trapping can become a dominant process, and as more holes are trapped, the holes that accumulate in the trap level can trap a large number of photo-excited electrons in the valence band, making it more difficult for valence band electrons to transition to the conduction band, reducing the number of electrons that can be collected by the electron transport layer. It is worth noting that due to the close proximity of the trap level location to the VBM if the defect concentration is too high, Fermi level pinning will form, which further enhances non-radiative recombination of electrons and holes and reduces carrier concentration, which consequently decreases both open-circuit voltage and short-circuit current, and induces more pronounced hysteresis [57,58,59]. By observing the ${\mathrm{I}}_{\mathrm{i}}^0$ DOS of Figure 3c, it is found that the interstitial I atom has the same peak position of state density with its neighboring I atoms b and c, indicating that I ions adjacent to this interstitial atom contribute to the formation of the defect state ${\mathrm{I}}_{\mathrm{i}}^{-}$ does not create new trap levels inside the bandgap. This phenomenon arises because the interstitial iodine anion forms an I-Pb bond with the lead ion, which eliminates the unsaturated suspension bond and moves the trap state into the VB, which helps to eliminate the effect of electron–phonon coupling on the lattice.

In the ${\mathrm{I}}_{\mathrm{i}}^{+}$ system, the trap level formed by positively charged iodine cations is located below the CBM, and this trap state can act as an electron acceptor to capture the electrons of valence band transitions, forming electron traps and accelerating non-radiative recombination. In addition, the positively charged interstitial iodide cations are prone to form I trimers by mutual attraction with the point anions on the lattice, which leads to an increase in the number of trapping sites and will result in the localization of charge density to itself, thus reducing the charge density of the VBM and CBM.

The results from previous molecular dynamics simulations suggest that trap levels can be eliminated and carrier lifetimes extended by oxidizing ${\mathrm{I}}_{\mathrm{i}}^0$ to ${\mathrm{I}}_{\mathrm{i}}^{-}$ [60]. However, ${\mathrm{I}}_{\mathrm{i}}^{-}$ is unstable under light conditions, and holes are easily captured and converted back to ${\mathrm{I}}_{\mathrm{i}}^0$ again, reducing the performance of perovskite [48,61]. The preceding analysis indicates that the impact of I_i on the electronic structure of perovskite is more pronounced among the point defects.

In order to minimize the effect of I_i on FAPbI₃, based on the same model, we investigated the possible effect of substitution of I_i by the halides Br_i and Cl_i on FAPbI₃. From the DFE of Figure 4a, it can be seen that the negative charge states (−) of Br_i and Cl_i, which are stabilized inside the bandgap at the thermodynamic level, are the most stable, which is the same as that of I_i, and the magnitude of the DFE is I_i > Br_i > Cl_i, which represents the fact that Br_i and Cl_i are more likely to form, this may be due to the fact that Cl and Br with smaller ionic radii than I are more likely to form interstitial defects that do not cause excessive distortion stresses on the lattice leading to elastic distortions, and therefore require less energy cost and are more likely to occupy interstitial positions in the lattice, and these smaller ionic radii may be more likely to chemically bond with neighboring atoms or ions, resulting in stable chemical bonding with fewer dangling bond effects on the defect state [62]. Compared with I_i, the trap levels of Br_i and Cl_i demonstrated in Figure 4b become more concentrated, the transition levels of defects between different charge states occur more easily with less energy change required, and the trap level as a whole is significantly more biased towards the valence band, gradually shifting from deep-level traps inside the bandgap to shallow-level traps, implying that electrons or holes captured by the trap state can be easily ionized by smaller ionization energies. This renders it difficult to form recombination centers, which reduces the impact on carrier transport, consistent with the improved photovoltaic performance reported in previous studies [63].

We also calculated defect formation energy for F_i replacing I_i, as shown in Figure S3a. The defect formation energy of F_i is significantly higher than that of I_i, Br_i, and Cl_i. This indicates that F_i formation in FAPbI₃ is energetically unfavorable, making it difficult for F_i to occupy I_i sites. Additionally, fluorine demonstrates inferior trap level modulation capability compared with bromine and iodine in Figure S3b. Fluorine ions likely exhibit minimal impact on regulating deep-level traps in FAPbI₃. Subsequent discussions will focus on Br_i and Cl_i properties.

Given the low formation energies of Br_i and Cl_i and their lack of deep-level traps, we investigated their electronic density of states (DOS). Figure 5a reveals that substituting lattice iodine with bromine or chlorine introduces no in-gap states and only slightly widens the bandgap. This suggests minimal disruption to the fundamental electronic structure of FAPbI₃ and largely preserved electronic transport properties. From Figure 5b, it can be found that at a charge state of 0, compared with ${\mathrm{I}}_{\mathrm{i}}^0$, ${\mathrm{Br}}_{\mathrm{i}}^0$, and ${\mathrm{Cl}}_{\mathrm{i}}^0$ make the positions of VB and CB shifted upwards as a whole, which causes the positions of the trap levels to be progressively closer to the valence band, which is more susceptible to ionization. The trap levels formed by the positive charge defects in Figure 5c are located deep in the bandgap, and due to the +/− charge state transitions, there is a trapping behavior of electrons. The positively charged defects act as electron acceptors, and the trap level of ${\mathrm{I}}_{\mathrm{i}}^{+}$ is close to CB; thus, it is expected that the trapping of the electrons will take place much more rapidly, whereas the trapping of electrons becomes more and more difficult in the case of ${\mathrm{Br}}_{\mathrm{i}}^{+}$ and ${\mathrm{Cl}}_{\mathrm{i}}^{+}$ because of the gradual location of trap levels closer to the VB, it is notable that the positively charged defects, Br_i and Cl_i, are not the lowest energy of formation at the thermodynamic level. Consequently, the density of these defects is not expected to be particularly high.

The reduced formation energies of Br_i and Cl_i suppress the generation of I_i defects. Compared with the deep-level traps formed by I_i, shallow-level traps exert less detrimental effects on carrier transport, which may benefit the performance of FAPbI₃. This suggests that Br/Cl doping effectively inhibits non-radiative recombination and prolongs carrier lifetime in halide perovskites, consistent with recent experimental studies. For instance, Zhang et al. [64] used synchrotron-based XRD and XPS to demonstrate that Cl-terminated ligands optimize halogen distribution in FAPbBr_3-xCl_x nanocrystalline films. This optimization induces a 9 nm blue shift in the PL peak and a 1.6-fold increase in exciton binding energy, confirming reduced deep-level traps. Similarly, Br-rich surfaces in CsPbBr₃ nanosheets significantly enhance the PLQY (from 31.15% to 87.2%) and extend carrier lifetime (to 16.69 ns), as reported by Varnakavi et al. [65]. These improvements further evidence the defect-passivation effect of Br doping.

The aforementioned conclusions lead to the conclusion that Br_i and Cl_i primarily influence defect properties rather than the fundamental electronic structure of FAPbI₃. To further investigate their capacity to modulate I_i defects, we examined the impact of the defects on the nature of light absorption. Figure 6 illustrates the optical absorption coefficient versus wavelength for FAPbI₃ with I_i, Br_i, and Cl_i interstitial defects. These results demonstrate that the intrinsic absorption of FAPbI₃ remains nearly identical regardless of I_i, Br_i, or Cl_i interstitial defects, confirming consistent optical bandgaps. The absorption observed in long-wavelength regions originates from FAPbI₃ containing Ii defects, indicating defect-mediated absorption.

The absorption edge that emerges after adding impurities to the absorption peak suggests that the defect is fully ionized. The impurity absorption position of Br_i is slightly blue-shifted, and the absorption peak is diminished, which suggests that Br_i trap level necessitates less ionization energy. Additionally, Cl_i exhibits negligible impurity absorption because of its shallow trap levels, consistent with its rapid ionization behavior., which corresponds to the previous Figure 5b, the shallow-level trap of Cl_i is readily ionized. The suppression of the trap level by Br and Cl has been demonstrated to effectively reduce the effect of I on the light-absorbing properties of FAPbI₃, as evidenced by the absorption spectra.

To evaluate the dynamic behavior of different interstitial defects (I_i, Br_i, and Cl_i), we performed molecular dynamics (MD) simulations at 300 K using the NVT ensemble with a Langevin thermostat [66,67]. The simulations utilized a time step of 1 fs and spanned 5 ps after equilibration. As shown in Figure 7a, the energy fluctuations of the three supercells are within a narrow range, and no significant atomic rearrangements or phase transitions occur during the simulation. Consequently, we conclude that FAPbI₃ containing halide defects is generally thermally stable [68]. Figure 7b shows the root mean square displacement (RMSD) of interstitial atoms versus time. To determine the RMSD in the equilibrium state, the first 2 ps were excluded as the pre-equilibration period. We can observe that the RMSD follows the trend of I_i > Br_i > Cl_i. This indicates that interstitial iodine atoms exhibit larger displacements, suggesting higher dynamic mobility compared with bromine and chlorine interstitials [69]. This may make interstitial defects more prone to ion migration and defect recombination, limiting the efficiency and stability of FAPbI₃ [70]. For this purpose, we calculated the diffusion coefficients using root mean square displacement, and these values are presented in

Table 1. Calculated diffusion coefficient on different interstitial defects.

Defect	Diffusion Coefficient cm2·s−1
Ii	4.53 × 10−8
Bri	3.24 × 10−8
Cli	2.67 × 10−8

(see Supporting Information for calculation details).

The calculation results we obtained match those reported in previous studies [51,71]. Through the analysis of diffusion coefficients, it has been demonstrated that I_i exhibit significantly higher diffusivity compared with Br_i and Cl_i, indicating that I_i possess the fastest migration rates within the crystal lattice, which exerts a more significant influence on the performance and stability of FAPbI₃ perovskite compared with Br_i and Cl_i.

The above calculations show that Br_i and Cl_i can reduce the impact on the performance of perovskite solar cells by transforming the deep-level traps formed by I_i into shallow-level traps, thus improving the performance of FAPbI₃. In a previous study, Lyu et al. [72] added FAX additives (X = Br or Cl) to a solution of perovskite precursor and characterized the properties of FAPbI₃ after treatment with the additive. It was observed that the perovskite solar cells exhibited an enhancement in the short-circuit current density (J_sc), open-circuit voltage (V_oc), and PCE following the addition of FABr and FACl into the FAPbI₃ precursor solution. This indicates that the incorporation of Cl and Br ions markedly enhances the carrier lifetime, with the Cl ion modifier exhibiting a more pronounced impact, which aligns with the calculated outcomes [73,74].

We further investigated the performance of FAPbI₃ cells using Br and Cl as additives (

Table 2. Performance improvement of FAPbI₃ with Br or Cl as dopant or additive.

Device Architecture	Solution	PCE Before Solution %	PCE After Solution %	Stability Performance	Ref.
FTO/SnO2/FAPbI3/Spiro-MeOTAD/Au	FACl/FABr additive	16.55	22.51/20.08	-	[72]
FTO/TiO2/FAPbI2.8Cl0.2/Spiro-MeOTAD/Au	FACl dopant	15.7	17.0	-	[30]
FTO/TiO2/FAPbI2.9Br0.1/Spiro-MeOTAD/Au	FAPbBr3 dopant	12.98	16.569	-	[75]
FTO/SnO2/FAPbI3/Spiro-MeOTAD/Ag	PACl additive	17.39	21.45	40 °C, 40%RH, AM 1.5 G, 70% PCE after 140 h.	[76]
FTO/SnO2/FAPbI3/Spiro-MeOTAD/Au	PFACl additive	22.8	24.4	60 °C, 25 ± 5%RH, 85% PCE after 500 h.	[77]
ITO/SnO2/FAPI3/Spiro-MeOTAD/Au	FFEACl additive	23.04	25.41	25 °C, 30%RH, AM 1.5 G, 83% PCE after 1000 h.	[78]
ITO/PTAA/FAPbI3/C60/BCP/Ag	FABr additive	11.31	14.38	-	[79]
FTO/TiO2/FAPbI3/Spiro-MeOTAD/Au	FACl + FAAC additive	16.42	20.92	50 °C in N2, 87% PCE after 1628 h.	[80]
ITO/PCBM/FAMAPbI3-XBrX/P3HT/Ag	FACl additive	18.60	21.02	25–35 °C, 10–25%, 90% PCE after 1200 h.	[81]

). These data show that halide additives significantly enhance the photoelectric conversion efficiency (PCE) and improve the long-term stability of FAPbI3. This improvement is consistent with the ability of halides to modulate defects, as identified in this study. This reflects the potential of Br and Cl doping or addition in the preparation of perovskites free of cations, such as MA⁺ and Cs⁺.

4. Conclusions

This study employs first-principles calculations to investigate halide doping effects on intrinsic point defects in FAPbI₃. The results show that the low formation energies of Br_i and Cl_i transform deep-level traps from I_i into shallow-level traps based on Br_i and Cl_i. Molecular dynamics calculations reveal I_i defects diffuse more readily than Br_i or Cl_i. With minimal bandgap changes, halide defects (Br_i or Cl_i) mitigate detrimental effects on perovskite device performance compared with intrinsic FAPbI₃ defects. This explains the mechanism behind FAPbI₃ performance enhancement via halide doping. Future work should systematically study halide doping effects on FAPbI₃ carrier transport properties and develop co-doping strategies.