Near-unity photoluminescence quantum yield in inorganic perovskite nanocrystals by metal-ion doping

All-inorganic lead-halide perovskite nanocrystals (NCs), typically CsPbX₃ (X = Cl⁻, Br⁻, and I⁻), are being considered as next-generation light emitters due to their desirable photophysical properties, such as high photoluminescence quantum yield (PLQY: e.g., near unity), high color-purity with tunable emission, high defect tolerance, and good thermal and moisture stabilities.^1–9 Because of the ionic nature and low crystallization temperature, colloidal inorganic perovskite NCs can be easily synthesized by low-cost solution-processed methods.⁴ Such superlative properties render CsPbX₃ nanocrystals widely utilizable in optoelectronic applications such as solar cells,^10–13 light-emitting diodes (LEDs),^14–21 lasers,^22–25 photodetectors,^26–28 single-photon sources,^29,30 X-ray scintillators,^31,32 visible-light communications,³³ and high-definition displays.^34,35 Despite the great promise for various applications, several studies indicate that solution-processed perovskite NCs suffer from the existence of a substantial level of trap-mediated nonradiative losses including point defects (excessive Pb²⁺ atoms or halide vacancies) and structural disorders, which act as parasitic nonradiative recombination centers, leading to photoluminescence quenching and poor thermal-stability and photo-stability.^{17,26,36–39} Recently, several promising steps have been taken to break the present bottleneck and tackle these issues, including (i) surface treatment with different ligands or metal ion salts,^17,42–46 (ii) employing halide-rich synthesis conditions or controlling the lead and halogen ratio during the synthesis,^37,46,47 and (iii) doping with targeted metal halide salts.^40,41 In semiconductor nanocrystals, dopant engineering is a promising approach for tailoring and modulating the luminescence efficiency as well as the stability of the NCs.^48,49 In general, the incorporation of dopant states either imparts extra electrons (i.e., n-type doping) or extra holes (i.e., p-type doping), which allows for fine control over the recombination processes and accordingly enhances the optical properties of the host, or induces new emissions from dopants.^50,51 In addition, judicious doping enables introducing new electronic energy levels within the bandgap and corresponding low-energy electronic transitions without changing the crystal structure or the intrinsic optical features of the host materials.^41,52 In the case of the inorganic perovskite system, the intrinsic flexibility of the ionic crystalline scaffold (APbX₃) makes it more convenient and versatile to achieve targeted doping compared to the conventional metal chalcogenide semiconductor nanocrystals, such as PbS and CdSe.⁴¹ Moreover, introducing B-site (i.e., the site of Pb²⁺) dopants could precisely tune the luminescence and optical properties and bandgaps, stabilize crystal structures, create new emission characteristics, and reduce the defect state density.^41,53 In the perovskite NC systems, the doping can be realized by adding the metal-ion salts into the precursor solutions during the hot-injection or room temperature synthesis.^50,54,55 In addition, postsynthetic cation exchange reactions have been shown to be another efficient way for tuning the compositional and optical properties of the NCs.^56,57 Recently, a wide array of metal ions, including transition metals, rare earth elements, and lanthanides, have been successfully incorporated into perovskite NCs and led to a substantial augmentation of the NCs’ optical properties.⁴¹ For instance, the doping of Mn²⁺ and lanthanide ions (e.g., Yb³⁺, Er³⁺, La³⁺, …) into perovskite NCs can generate a high PLQY—approaching near-unity in some cases—of metal ion-related emission as a result of the efficient energy transfer from the perovskite NC host to the dopants.^58–62 However, another variety of monovalent, divalent, and heterovalent metal ion dopants (Ni²⁺, Cd²⁺, Cu²⁺, Ca²⁺, Sr²⁺, Na⁺, Ce³⁺, and Sb³⁺) has led to the improved PLQY of the perovskite NCs without introducing new emission bands.^{50,54,55,63–66} On the other hand, it was also noted that some dopants (i.e., Bi³⁺, Al³⁺, and Ag⁺) tend to introduce trap states in the bandgap that quench the photoluminescence.^67–69 Despite these impressive strides in the area of perovskite doping, the exact influence of dopant ions on the optical properties and their behavior inside the perovskite matrix remains unclear and hard to predict.⁴⁰ In this perspective, we summarize the recent studies on the B-site metal ion doping of CsPbX₃ (X = Cl, Br, and I) NCs that lead to the improved optical characteristics (i.e., a higher PLQY and longer PL lifetime after doping). We present the recent works and progress made on the metal doping of CsPbCl₃, CsPbBr₃, and CsPbI₃ and give an overview of the current understanding of the inherent doping mechanisms that lead to PL enhancement and how dopants control the optical properties of inorganic perovskite materials. Finally, we discuss the present limitations and challenges to be solved to further expand the potential integration of doped perovskite NCs into optoelectronic devices for better efficiency and operational stability.

The CsPbCl₃ NC system is the only system in the Cs-Pb-X perovskite family that can achieve deep-blue emission due to its intrinsically large bandgap (∼3.0 eV). However, point defects (such as Cl vacancies, with the lowest defect formation energy occurring under Cl-poor synthetic conditions²⁶) can introduce local structural distortions and lattice disorders. Such defects can induce intragap defect states and nonradiative recombination centers that degrade the optical properties of the pristine NCs.⁶³ Therefore, several research efforts have been recently devoted to increasing the PLQY of CsPbCl₃ NCs in order to expand their practical optoelectronic applications. Here, we focus on the recent developments related to metal-ion incorporation in pristine CsPbCl₃ NC hosts that improve the optical properties and PLQY without introducing new emission bands from metal ions. In the literature, several dopants have been incorporated into the CsPbCl₃ NC hosts, especially alkaline earth (AE²⁺) metals and transition metals. For instance, due to the comparable ionic radii of AE²⁺ metal ions [i.e., Ba²⁺, Sr²⁺, Ca²⁺, and Mg²⁺ (135, 118, 100, and 72 pm, respectively) relative to Pb²⁺ (119 pm)],⁷⁰ Chen and co-workers have carefully explored the effects of doping AE²⁺ metal ions into CsPbCl₃ NCs (via the hot injection method) on suppressing the structural defects and improving the performance of the LED devices.⁶⁵ As shown in Fig. 1(a), they found that the photophysical properties were correlated with the type and amount of AE metal ions: (i) upon incorporating Sr²⁺, Ca²⁺, and Mg²⁺, a significant increase in the PLQY was observed, from ∼0.8% to 36.6% (Mg²⁺), to 50.7% (Sr²⁺), and to 77% (Ca²⁺); (ii) an extremely low PLQY was observed in the case of Ba²⁺ (0.9%); and (iii) increasing the doping concentration of Ca²⁺ from 2% to 11% shows a remarkable increase in the PL lifetime and PLQY [see the time-resolved PL spectra in Fig. 1(a), lower panel]. In addition, the measured nonradiative decay rates clearly suggest a successful suppression of the surface defect states and/or vacancies.

In the case of doping with transition metal ions, Yong et al. have studied the effect of incorporating Ni ions into CsPbCl₃ NCs using the hot injection method, as Ni can strongly coordinate with halide ions.⁶³ From the steady-state absorption and PL spectra of CsPbCl₃ NCs before and after Ni²⁺-doping [Fig. 1(b)], there is virtually no change in the absorption spectra of CsPbCl₃ NCs after Ni²⁺-doping, but a substantial PL enhancement of CsPbCl₃ NCs was observed. Interestingly, a remarkable increase in the PLQY (up to 96.6%) was achieved, suggesting negligible nonradiative recombination in Ni²⁺-doped NCs when the doping concentration reached 11.9 mol. %. The time-resolved PL measurements indicated that the PL decay of CsPbCl₃ NCs becomes slower after Ni-doping (i.e., an increased PL lifetime), providing another strong piece of experimental evidence for the efficient suppression of the nonradiative recombination channels [Fig. 1(b), left panel]. Importantly, the crystal structural characterizations assisted by both X-ray absorption fine structure (EXAFS) and magic angle spinning (MAS) measurements revealed that Ni²⁺ ions can substantially suppress the structural formation of defects, especially the Cl vacancy, in the CsPbCl₃ NCs. Therefore, the obtained near-unity PLQY of CsPbCl₃ can be attributed to the improved crystal lattice order when the Ni²⁺ ions are introduced into the host NCs.

Samanta and co-workers have recently studied the doping of Cd²⁺ metal ions into CsPbCl₃ NCs using room-temperature postsynthetic cation exchange/treatment.⁶⁴ Similar to the Ni²⁺-doping case,⁶³ the Cd²⁺ ions induced no PL spectral shift despite that the radii of Cd²⁺ ions (93 pm) and Ni²⁺ ions (69 pm) are much smaller than that of Pb²⁺ ions [Fig. 1(c)]. Surprisingly, the PLQY was improved, reaching near-unity (96% ± 2%) after Cd²⁺ doping. The Cd²⁺-doped CsPbCl₃ NCs were found to exhibit improved stability under the continuous illumination in ambient conditions. The time-resolved PL measurements [Fig. 1(c), left panel] suggested that the doped NCs exhibit a much longer lifetime as compared to that of the pristine NCs. In addition, the ultrafast transient absorption and the temperature-dependent PL measurements along with the calculated nonradiative decay rates revealed that the Cd²⁺ ions can introduce a shallow energy level and dramatically suppress the nonradiative channels, which could be one possible reason for the increased PLQY.

Very recently, De et al. have investigated another transition metal ion (i.e., Cu⁺) doped into CsPbCl₃ NCs by using the one-pot hot injection method.⁵⁴ The PLQY increased by almost ∼120 times upon Cu⁺ doping (from ∼0.5% up to 60%). Similar to the previous cases using transition metal ions, no PL shift was observed in Cu⁺-doped CsPbCl₃ NCs. The time-resolved PL measurements showed that the Cu⁺-doped NCs show much longer PL lifetime as compared to undoped NCs; furthermore, the femtosecond transient absorption measurements suggest the suppression of the ultrafast carrier trapping processes in the doped CsPbCl₃ NCs.

Due to the fast ion exchange between perovskite NCs and metal ion dopants, diffusion doping can also be realized via postsynthetic surface treatment. We have recently reported a simultaneous dual-surface doping/passivation using YCl₃ metal ion salt on the surface of CsPbCl₃ NCs.²⁶ We achieved an enhancement in the PLQY (from 1% to 60%) of CsPbCl₃ NCs with improved surface stability [Fig. 1(d)]. From time-resolved PL spectra, we also observed a significant increase in the PL lifetime from 0.4 to 1.7 ns, which suggests that the nonradiative decay rates were substantially reduced after YCl₃ treatment. This was also supported by density functional theory (DFT) calculations, which show that Y³⁺ and Cl⁻ ions efficiently occupy the surface Pb–Cl ion vacancies, enriching the density of states in the conduction band without creating any mid-gap states.

Among Cs-Pb-X compounds, CsPbBr₃ NCs are considered to be the most promising green light emitters for LED applications^17,18 because of their relative thermal stability (compared to CsPbI₃ and CsPbCl₃), low moisture sensitivity, and high PLQY.^1,67,71 However, there is still a need to substantially improve the external quantum efficiencies (EQEs) of the CsPbBr₃ NC-based LED devices before their commercial applications are realized. Hence, doping could also become a key strategy to increase the EQE.⁷² Yao et al. have recently doped CsPbBr₃ NCs with Ce³⁺ ions by using the hot injection method, taking into account that the Ce³⁺ ion is not too dissimilar in ionic radius (103 pm) from the Pb²⁺ ion (119 pm).⁵⁰ They found that CsPbBr₃ NCs when doped with a substantive amount of Ce³⁺ ions remarkably showed a twofold enhancement in the PLQY (reaching up to 89%) as compared to the pristine CsPbBr₃ NCs. The steady-state absorption and PL showed that the Ce³⁺ dopants could induce a slight blue shift in the absorption edge and a large enhancement in the PL intensity [Fig. 2(a)]. Under ambient conditions, the CsPbBr₃ NCs showed significant improvement in stability due to the lattice contraction after Ce³⁺-doping.⁵⁰ Interestingly, through the comparison of the average PL lifetimes and the PLQY of CsPbBr₃ NCs at different dopant concentrations, it was found that the PL enhancement was accompanied by a decrease in the average PL lifetimes as the dopant concentration increased [Fig. 2(a), left side]. They further explained such an unexpected relationship based on the ultrafast transient absorption measurements: the near band-edge states from Ce³⁺ dopants provided more radiative channels for the PL enhancement, and meanwhile, the enhanced couplings between these states as well as between the higher-lying and lowest excitonic states can promote the PL decay. In addition, the fabricated LED devices based on the Ce³⁺-doped CsPbBr₃ NCs showed a 3-fold increase in the EQE from 1.6% to 4.4%.

Very recently, Li et al. have reported a doping strategy of univalent Na⁺ ions into CsPbBr₃ NCs at room-temperature.⁵⁵ The doped CsPbBr₃ NCs showed better color purity, improved stability, and a higher PLQY (∼85%) compared to the pristine NCs. As shown in Fig. 2(b), the steady-state absorption and PL spectra of the colloidal Na⁺-doped CsPbBr₃ NCs with different Na/Pb atom ratios show that upon Na⁺ doping both absorption and emission peaks were gradually blue shifted by increasing the doping concentration, which indicates that such doping can slightly alter the bandgap of CsPbBr₃. The temperature-dependent PL measurements suggest that the emission intensity of the doped NCs decreased by increasing the temperature, which results from the thermally activated nonradiative recombination. Moreover, the time-resolved PL measurements showed that the average PL lifetime became much longer and the calculated nonradiative recombination rates dropped dramatically after Na⁺-doping, highlighting that Na⁺ ions substantially suppressed the formation of charge carrier trapping states in the CsPbBr₃ NCs. Interestingly, the white light-emitting devices (WLEDs) employing the Na⁺-doped NCs as solid-state phosphors exhibited a stable high quality white-light and a notably high power efficiency (67.3 l m/W).

Zhang et al.⁷³ have reported the room-temperature doping of Sb³⁺ ions (92 pm ionic radius) into ultra-small blue emissive CsPbBr₃ NCs. They found that the doping of Sb³⁺ ions could reduce the surface energy, boost the PLQY of blue emission up to 73.8%, and enhance the stability even at elevated temperatures in solution (40–100 °C). The steady-state absorption and PL spectra in Fig. 2(c) demonstrate a slight reduction in the full width at half maximum (FWHM) due to the increased exciton binding energy. The PLQY of the doped NCs [sample C in Fig. 2(c)] increased from a 50% baseline to as high as 73%. The PL decay kinetics suggested that the Sb³⁺ doping could passivate the defect states below the bandgap and increase the probability of radiative recombination. It was noted that the CIE coordinates of the doped CsPbBr₃ NCs were very close to the primary blue color according to the National Television System Committee (NTSC) TV color standard, which bodes well for the applications of these NCs in display.

CsPbI₃ NCs are very promising red-light emitters. Although pristine CsPbI₃ NCs can achieve a high PLQY (>70%), these NCs suffer from severe phase instability. The black phase of CsPbI₃ exhibits an undesirable tolerance factor that leads to poor phase stability, i.e., it undergoes a spontaneous room temperature phase transition to a nonphotoactive yellow phase when exposed to humidity or ambient conditions.^74,75 Thus, tremendous efforts have been devoted for developing various synthetic approaches to stabilize the black phase of CsPbI₃ NCs while preserving a high PLQY.^2,76,77 Due to the similar ionic radii of Sr²⁺ and Pb²⁺ ions (118 pm vs 119 pm), Lu et al.⁶⁶ have studied the doping effects of Sr²⁺ on the photo/electroluminescence efficiency and the phase stability of CsPbI₃ NCs. In their study, the SrCl₂ coprecursor introduced a simultaneous Sr²⁺ ion doping and surface Cl⁻ ion passivation. The steady-state absorption and PL spectra of CsPbI₃ NCs with different amounts of SrCl₂ showed that both the absorption peak and PL peak exhibited a slight blue-spectral shift when increasing the SrCl₂:PbI₂ ratio. Following the SrCl₂ addition, a significant increase in the PLQY of CsPbI₃ NCs from 65% to 84% was observed and the improved stability of the cubic $α$-phase of CsPbI₃ NCs was achieved [Fig. 3(a)]. It is worth noting that the enhanced phase stability of CsPbI₃ NCs can be attributed to the direct increase in the formation of energy and the modifications of the tolerance factor of the CsPbI₃ crystal structure. The time-resolved PL kinetics suggest that the average PL lifetimes dramatically increased when employing SrCl₂, which demonstrates that the surface defects were efficiently reduced and the surface of CsPbI₃ NCs became Cl⁻ enriched—reducing the likelihood of uncoordinated Pb atoms—resulting in the increased PLQY and a much-prolonged charge carrier lifetime. The authors achieved a 13.5% EQE of LED devices by employing the Sr²⁺-doped CsPbI₃ NCs.

Shen et al. have studied the partial substitution of Pb²⁺ ions into CsPbI₃ NCs through Zn²⁺ alloying.⁷⁸ As the ionic radius of Zn²⁺ (74 pm) is much smaller than that of Pb²⁺ (119 pm), the reduced lattice constant of the NCs suggests that Zn²⁺ cations efficiently replace Pb²⁺ cations. The steady-state absorption and PL showed that both the absorption and the PL peaks of the NCs were blue shifted by increasing the amount of Zn²⁺ ions in CsPbI₃ hosts. The authors attributed these spectral shifts to the observations that the lattice contraction and the emission intensity increased at first and then decreased with the increasing concentration of Zn²⁺ in the NCs [see Fig. 3(b)]. Interestingly, the PLQY of the alloyed NCs dramatically increased up to 98.5% when changing the Zn²⁺ concentration. In addition, the radiative decay rates showed a 4-fold increase, while a 10-fold decrease was seen in the nonradiative decay rates, indicating an effective suppression of the defect states in alloyed NCs. The stability of the NCs was also improved as a result of modulating the tolerance factor of the lattice structure. Interestingly, the nanocrystals switched from n-type due to I⁻ vacancies to nearly ambipolar for the alloyed NCs. More importantly, alloying Zn²⁺ into NCs can achieve an enhanced EQE of up to 15.1% in the LED-based device.

Very recently, Bera et al. have found that the incorporation of Sb³⁺ ions into the CsPbI₃ NCs using the hot injection method could stabilize the crystal phase under ambient atmosphere [Fig. 3(c)].⁷⁹ They observed that excess doping changed the shape of the cubic NCs and induced a 2D structure assembly, forming platelet-shaped NCs, and consequently quenched the emission efficiency and reduced the overall stability, as shown in the atomic model of the crystal structure. Since the ionic radius of Sb³⁺ (92 pm) is smaller than that of Pb²⁺ (119 pm), the Sb³⁺ ions can increase the tolerance factor of the crystal (i.e., an indicator for the stability and distortion of perovskite structures, which is defined as $t=rA+r02rB+r0$⁠, where r_A is the radius of the A-site cation, r_B is the radius of the B-site cation, and r₀ is the radius of the anion) and therefore improve the crystal stability. However, replacing Pb (ii) with Sb (iii) ions could lead to severe charge imbalance. The steady-state absorption and PL exhibited an apparent redshift by increasing Sb³⁺ concentration. The time-resolved PL measurements show that the PL lifetime of Sb³⁺-doped CsPbI₃ NCs increased relative to the pristine NCs, but still within the same nanosecond order. The PLQY of the doped CsPbI₃ NCs (doping concentration ∼10%) was 83%, and this value was practically decreased by increasing the Sb³⁺ amount deviated from ∼10%. DFT calculations suggested that upon Sb³⁺ ion incorporation, an increase in the crystal cohesive energy and a decrease in the bandgap were observed. Remarkably, the authors achieved a 9.4% power conversion efficiency (PCE) by employing these doped NCs in solar cell devices.

Mir et al. have recently studied the effect of postsynthetic Mn²⁺ doping on the surface and lattice energy of CsPbI₃ NCs.⁸⁰ They found that Mn²⁺ ions can passivate the NC surface and restrict the growth of the NCs and stabilize the black phase of CsPbI₃ NCs. Therefore, the postsynthetic Mn-doped NCs showed improved PL ambient stability as compared to the undoped NCs [see Fig. 3(d)]. The time-resolved PL decays showed that a radiative lifetime was about the same before and after doping, respectively, suggesting that Mn-doping did not introduce any additional nonradiative decay channels to the CsPbI₃ NC host. Due to the large difference between the Mn–I bond (283 kJ mol⁻¹) and the Pb–I bond (194 kJ mol⁻¹) dissociation energies, the lattice contraction stabilizes the lattice and the surface of Mn-doped NCs relative to the undoped NCs. It is worth highlighting that different from the cases of CsPbCl₃ and CsPbBr₃ NCs, the energy transfer into the d-states of Mn²⁺ from the host would not occur and there is no emission from Mn²⁺ ions in the doped CsPbI₃ NCs because the Mn d-d transient energy level (∼2.12 eV) is higher than the bandgap of CsPbI₃ NCs (∼1.8 eV).

Very recently, Guvenc et al. studied the effects of Gd³⁺ doping on the optical properties and the phase stability of α-CsPbI₃ NCs. Since the ionic radius of the Gd³⁺ ion (93 pm) is smaller than that of the Pb²⁺ ion (119 pm), Gd³⁺ incorporation into the perovskite lattice increased the Goldschmidt tolerance factor of the CsPbI₃ crystal and accordingly led to a more stable perovskite structure.⁸¹ The authors attributed the prolonged phase stability to several factors, such as (i) increased tolerance factor of the perovskite structure, (ii) distorted cubic symmetry, and (iii) decreased defect density in nanocrystals. Interestingly, Gd³⁺ doping (10 mol. %) of α-CsPbI₃ led to an increase in both the PLQY from 70% to 80% and the fluorescence lifetime from 47.4 ns to 64.4 ns, which stem from the promoted radiative recombination due to the reduced defect state density. In addition, similar to the Sb³⁺ doping case,⁷⁹ the DFT calculations revealed that Gd³⁺ doping increases the cohesive energy per atom from 3.22 eV to 3.32 eV for neat and Gd-doped CsPbI₃, respectively, which indicates that the formation of Gd³⁺-doped CsPbI₃ NCs is more favorable than neat CsPbI₃ NCs. Moreover, the electronic band structure of Gd-doped CsPbI₃ crystals showed an n-type semiconductor behavior without any substantial modifications in the band structure.

Several theoretical tools, especially density functional theory (DFT), have been used to understand the doping mechanisms and their relationship to the optical properties in inorganic perovskite materials at the atomic level. For CsPbCl₃, Chen et al. have found that in the case of Ca²⁺ and Sr²⁺doped CsPbCl₃ NCs,⁶⁵ (i) compared to foreign atoms (dopant ions), Pb²⁺ atoms are involved at an early stage of NC formation; (ii) once the NCs were formed, the Pb²⁺ atoms were depleted and uncoordinated surface atoms were generated; and (iii) since the Ca²⁺ ions were abundant in the solution, the Ca²⁺ ions can strongly attach to the surface of NCs because of the large bond dissociation energy of Ca–Cl (409 kJ/mol). The authors also suggested that Ca²⁺ (or Sr²⁺) isovalently occupies the Pb²⁺ crystallographic sites, resulting in the formation of a passivation layer at the NC’s surface, as was concluded from XPS measurements. The formation of such a layer decreases the concentration of point defects caused by uncoordinated constituent atoms at the NC’s surface (e.g., the uncoordinated Pb²⁺ ions) as shown in Fig. 4(a), leading to the improved short-range order of the lattice as suggested by EXAFS measurements. However, for the larger alkaline earth metal ions, such as Ba²⁺, the bond dissociation energy of Ba–Cl (443 kJ/mol) is so large that the Ba²⁺ ions can enter the lattice of NCs even at the early stage of NC growth. Although, due to Ba²⁺ having a larger ionic radius than Pb²⁺ (135 pm vs 119 pm), doping of Ba²⁺ ions can increase the local strain of the lattice, giving rise to a higher probability of forming atomic point defects as suggested by DFT calculations. Such structural defects may further trigger the incorporation of more Ba²⁺ ions into NCs, thus leading to the successful doping of Ba²⁺ in the core region as well as further lattice expansion. Therefore, the PLQY of Ba²⁺-CsPbCl₃ NCs is very low. Different from the case of Ba²⁺ dopants, which can introduce deep trap states that quench the PL of CsPbCl₃ NCs, Modal et al. have shown that both Cd²⁺ and Cl⁻ play a significant role in suppressing the halide vacancies by using CdCl₂ for increasing the PLQY. In addition, De et al. have attributed the high PLQY of Cu⁺-doped CsPbCl₃ NCs to the suppression of carrier trapping and showed that the majority of the carriers follow a longer recombination time as indicated by transient absorption measurements.

Yong et al. achieved a near-unity PLQY by doping Ni²⁺ ions into CsPbCl₃, and their experimental results suggest that the doping of Ni²⁺ ions can substantially suppress the formation of structural defects such as chloride vacancies and results in improved short-range order of the perovskite lattice and more ordered lattice.⁶³ Their DFT calculations revealed that Ni²⁺-doped CsPbCl₃ NCs show a much larger defect formation energy of Pb vacancy (V_Pb) than those of V_Cl and V_Cs, suggesting that V_Cl and V_Cs could be the dominant defects in CsPbCl₃ NCs and Ni doping into CsPbCl₃ favors the suppression of those defects. It is worth mentioning that V_Cl and Ni-V_Cl can introduce deep defect levels, which act as carrier trapping centers for nonradiative recombination [Fig. 4(b)].

For CsPbBr₃ NCs, Li et al. have explained the PL enhancement and the higher PLQY in the Na⁺-doped CsPbBr₃ NCs due to the decrease in the carrier trapping centers.⁵⁵ Their DFT calculations suggested that the formation energy of the antisite Na_Pb defect is the lowest regardless of the position of the Fermi level. Moreover, when the Fermi level is close to the conduction band minimum (CBM), the formation energy of the Na_Pb defect is still the lowest among all the studied defects. The formation energy calculations demonstrate that Na_Pb is the dominant defect in Na⁺-doped orthorhombic-phase CsPbBr₃. To understand the mechanism of the enhanced band-edge emission in the case of Ce³⁺ ions doping, Yao et al. have conducted femtosecond transient absorption measurements to study the ultrafast exciton dynamics of Ce³⁺ doped and undoped samples.⁵⁰ They found that the Ce³⁺ ions can introduce a new band-edge state near the conduction band, leading to the increase in the lowest excitonic state density, facilitating the coupling between the higher-lying excitonic states and the lowest excitonic state as well as the coupling between the lowest excitonic state and the bandgap trap states. In addition, in our DFT calculations, the calculated doping charge transition levels suggest that most of the Ce³⁺ dopant related defects could form the shallow transition levels within the bandgap in both cubic-phase and orthorhombic-phase Ce³⁺-doping CsPbBr₃.⁸² These shallow levels can preserve the bulk electronic structure and do not degrade the optoelectronic properties. From the projected density of states (PDOSs) for cubic-phase CsPbBr₃ after introducing the antisite Ce_Pb and interstitial Ce_i dopants, we found that (i) for pristine CsPbBr₃, the valence band maximum (VBM) and CBM of CsPbBr₃ are mainly composed of Br-4p, Pb-6s, and Pb-6p orbitals and show delocalized electronic distributions; (ii) upon Ce³⁺ doping, the bandgaps of CsPbBr₃ in both phases remain almost unchanged after introducing the Ce³⁺ ions at the Pb²⁺ site, and the charge densities for the CBM retain the delocalized feature; and (iii) the interstitial Ce³⁺ has a small contribution to the conduction band edge in the PDOS of cubic-phase CsPbBr₃, and the same electronic feature can be found in the orthorhombic case with an interstitial Ce³⁺. Therefore, we have confirmed that Ce³⁺ can be doped into CsPbBr₃ NCs without introducing additional trap states and can slightly increase the electron density in the conduction band, which modulates the exciton relaxation and recombination of CsPbBr₃ NCs.

In the case of CsPbI₃ NCs, Bera et al. have highlighted that Sb³⁺ doped CsPbI₃ showed an increase in the cohesive energy along with a decrease in the bandgap due to Sb incorporation based on the DFT calculations.⁷⁹ Lu et al. have also highlighted that the mechanism for the PL enhancement in Sr-doped CsPbI₃ NCs is due to the reduction of surface defects when using SrCl₂ as a coprecursor.⁶⁶ When using SrCl₂ as dopants, the surface of CsPbI₃ NCs becomes enriched with Cl⁻, which eventually passivates the uncoordinated Pb atoms, resulting in the increased PLQY and PL lifetime.

In summary, with the help of the DFT calculations, several aspects of the doping effects in CsPbX₃ NCs can be reasonably explained from atomic level interactions. From the calculated dopant formation energies, we can predict the probable atomic positions of metal ion dopants; the dopants can occupy the Cs, Pb, or halide sites if they have small intersite formation energies; otherwise, they can act as interstitials in the perovskite host. For the former case, we need to further determine the charge transition levels of dopants so that we can predict whether the dopants introduce trap states that quench the PL or increase the conduction band states, thereby enhancing the PL. In the case of interstitial substitution, on the other hand, significant energy transfer from the perovskite host to the dopants can promote the emission from dopants instead of the perovskite host.

Despite the great promise that perovskite NCs hold for light emission applications and optoelectronic devices, optimizing the perovskite material’s quality and stability remains as major impediments. Doping the perovskite NC lattice with metal ions presents a viable approach to overcome those impediments. The near-unity PLQY was achieved in several cases after metal ion doping as a result of the effective suppression of the defect states within the host NCs as well as on their surface. However, several outstanding issues need to be addressed before the full potential of doping strategies is realized, including (i) the precise atomic positions and distribution of metal ion dopants inside the perovskite lattice are still poorly understood. For instance, it has been reported that some dopants can induce deep trap state levels while others merely introduce shallow defect states. In this case, high-level DFT calculations (e.g., the combination of spin-orbit couplings and hybrid functionals) are needed to accurately predict the formation energies and charge transition levels of the dopants in the host; (ii) the energy transfer and energy level alignment between the perovskite nanocrystal host and metal ion guest need to be further explored and deciphered in order to understand the observed PL quenching or enhancement in the same materials when different metal ion dopants are used; and (iii) although a near-unity PLQY in perovskite NCs can be achieved upon metal ion doping in the solution, it is still challenging to obtain spin-coated thin films of perovskite NCs with a high PLQY and practical thermal-stability and photo-stability. The effect of dopants on the morphology of NC thin films should be accorded more attention. Moreover, a proper surface treatment of the doped perovskite NCs may help us to further improve the quality of their thin films and facilitate the fabrication of higher efficiency devices.

The authors acknowledge funding support from KAUST.