Interstitial and substitutional doping of Mn²⁺ in 2D PEA₂PbBr₄ and BA₂PbBr₄ perovskites

Abstract

Mn²⁺ doping imposes intriguing optoelectronic properties on lead-halide perovskites; however, its impact on their crystal structure remains unclear. This study investigates the consequences of interstitial and substitutional Mn²⁺ doping on the lattice-strain and interplanar spacings of 2D perovskites and correlates the structural changes to their optical properties.

---

Doping with a foreign metal ion is an attractive strategy to improve the properties of semiconductors such as lead-halide perovskites. Due to the flexible nature of perovskite structures, a wide range of metal ions have been explored. Among these, Mn²⁺ has been most extensively studied due to its bright characteristic emission and paramagnetic nature.^1,2 Despite these extensive studies, the nature of Mn²⁺ doping such as interstitial or substitutional, and the role of host lattice on the doping is still ambiguous. There have been several contradicting results reported on the type of Mn²⁺ doping. Some reports showed lattice contraction due to substitutional Mn²⁺ doping, while others demonstrated lattice expansion owing to their interstitial nature. Generally, the lattice expansion and contraction can be identified by the shift in their p-XRD peaks. Recently, Torma et al. observed a shift in the XRD peak to lower angles and demonstrated interstitial Mn²⁺ doping in Cs₂PbI₂Cl₂ using nano-XRD and nano-XRF measurements.³ On the contrary, Dutta et al. and Hou et al. observed a peak shift to the higher angles in XRD after doping with Mn²⁺ in 2D Cs₂PbBr₄ nanoplatelets and 3D CsPb(BrCl)₃ perovskite nanocrystals.^4,5 Interestingly, Nag et al. observed no shift in the XRD peak when Mn²⁺ (<1%) is doped in 2D BA₂PbBr₄ perovskite.⁶ Therefore, a comprehensive understanding of the nature of doping, its dependence on the host lattice structure, and its impact on the crystal structure, lattice strain, and optical properties of perovskites is required.

In this study, we utilize Mn²⁺ doping in two different 2D perovskite hosts: (i) phenylethylammonium lead bromide (PEA₂PbBr₄) and (ii) butylammonium lead bromide (BA₂PbBr₄), to investigate their structural changes and link these changes to their optical properties. These two perovskites are selected by considering the difference in their structural rigidity owing to the difference in their A-site ions, although both adopt similar layered 2D structures of the Ruddlesden–Popper (RP) type.⁷ Consequently, our study reveals that the nature of Mn²⁺ doping is significantly different in both PEA₂PbBr₄ and BA₂PbBr₄. We observed a striking difference in their XRD patterns after Mn²⁺ doping. PEA₂PbBr₄ showed lattice expansion perpendicular to the plane of the 2D inorganic framework (interstitial nature) after doping with Mn²⁺ whereas BA₂PbBr₄ showed a lattice contraction along the 2D plane (substitutional doping). Due to the manifestation of a soft lattice with a layered structure of 2D perovskites, the Mn²⁺ doping was feasible, and the changes in their crystal structure were easily noticeable after Mn²⁺ incorporation. Subsequently, this resulted in a dramatic difference in their emission; orange (Mn²⁺) emission in Mn²⁺:PEA₂PbBr₄ is more intense (∼10× higher) than that of Mn²⁺:BA₂PbBr₄. Upon tracking the exciton diffusion using transient photoluminescence microscopy (TPLM), we found that the exciton transport in PEA₂PbBr₄ is unaffected, whereas a noticeable difference in the exciton transport was observed in BA₂PbBr₄ with Mn²⁺ doping.

To commence our study, undoped and a series of Mn²⁺ doped 2D perovskites were synthesized using acid-initiated precipitation (detailed methodology is available in the ESI†). The ICP-MS analyses reveal that the amount of Mn²⁺ increases with an increase in the concentration of Mn²⁺ salts in the precursors (shown in Table S3, ESI†). However, the maximum feasible doping amount was 4.8% and 2.0% of Mn²⁺ (respective to the molar percentage of Pb²⁺) for PEA₂PbBr₄ and BA₂PbBr₄, respectively. Such low doping levels of Mn²⁺ are attributed to a difference in the ionic radii of Mn²⁺ (0.83 Å) compared to that of Pb²⁺ (1.19 Å).^6,8 The crystal structures of PEA₂PbBr₄ and BA₂PbBr₄ were analyzed by scXRD and shown in Fig. 1a and b (see Tables S1 and S2 for refined data, ESI†). Both perovskites show 2D layered Ruddlesden–Popper (RP) type structures and exhibit a higher degree of preferential orientation along the c-axis. Further, their p-XRD patterns were recorded after doping with various amounts of Mn²⁺ (Fig. 1c and d), which was and used for strain analyses. From p-XRD, we found that both perovskites retain the RP structures even after doping with Mn²⁺.

Fig. 1 Crystal structures of (a) PEA₂PbBr₄ and (b) BA₂PbBr₄. The pXRD patterns of undoped and Mn²⁺ doped (c) PEA₂PbBr₄ and (d) BA₂PbBr₄. The 2D perovskite lattices of PEA₂PbBr₄ and BA₂PbBr₄ are composed of a single (n = 1) inorganic layer of [PbBr₆]⁴⁻ octahedra. The PEA⁺ and BA⁺ cations act as organic spacers between the octahedral layers of PEA₂PbBr₄ and BA₂PbBr₄, respectively.

In the case of PEA₂PbBr₄, both undoped and Mn²⁺ doped perovskites show equally spaced and sharp diffraction peaks in p-XRD (Fig. 1c), denoted as (002), (003), (004), etc., and agree with the simulated pattern of a triclinic phase (lattice parameters a ≠ b ≠ c, α ≠ β ≠ γ) derived from scXRD. All these peaks arise from the same family of planes {00l} where (l = n), indicating the high degree of orientation with stacking layers. The periodic peaks due to these (00l) diffractions imply the spacing between the two layers, since these planes are perpendicular to the stacking axis (c-axis), see Fig. 1a. The periodicity obtained from the diffraction peaks was 1.648 ± 0.007 nm, indicating that each layer is ∼1.65 nm apart, as shown in Fig. 1a. After doping with Mn²⁺, no major changes in the p-XRD pattern were observed; however, a shift in some XRD peaks with slight broadening was observed. These peaks are at ∼26.6° and at ∼37.6°, which correspond to the (005) plane and the (007) plane, respectively, as shown in the magnified region of Fig. 1c and Fig. S1 (ESI†). The magnitude of the shift was Δ2θ₀₀₅ = 0.07° and Δ2θ₀₀₇ = 0.09°. Generally, given the difference in the ionic radii of Pb²⁺ (1.19 Å) and Mn²⁺ (0.83 Å), if Pb²⁺ is replaced by smaller Mn²⁺, a lattice contraction is expected with a shift in the XRD peaks to higher angular scale. On the contrary, in this case, the XRD peaks shifted to lower angles, indicating that there is a lattice expansion after doping with Mn²⁺ ions. This also implies that the Mn²⁺ ions are not replacing Pb²⁺. Since this shift is observed for (005) and (007) peaks, the lattice expansion occurs in an out-of-plane direction (perpendicular to (005) and (007) planes) along the c-axis, as shown in Fig. 1a. This is also evident from an increase in the spacing of (005) planes, as shown in Fig. 2a. Similar trends in p-XRD were observed by Torma et al.³ By using nano-XRD, they demonstrated that Mn²⁺ occupies the interstitial position.³

Fig. 2 The change in d-spacing against Mn²⁺ concentration of (a) the (005) plane in PEA₂PbBr₄ (b) the (111) plane in BA₂PbBr₄ and (c) microstrain vs. Mn²⁺ doping percentage.

In the case of 2D BA₂PbBr₄, the powder XRD patterns of both undoped and Mn²⁺ doped show equally spaced and sharp diffraction peaks indexed as (002), (004), (006), etc (Fig. 1d). This diffraction pattern aligns well with the simulated pattern obtained from the reference and scXRD data, confirming the orthorhombic Ruddlesden–Popper crystal structure (lattice parameters a ≠ b ≠ c, α = β = γ = 90°). All these peaks arise from the same family of planes {00l}, where (l = 2n) indicates highly oriented crystallites. The periodicity, obtained from {00l} family peaks, was found to be 1.366 ± 0.008 nm. Note that this interlayer spacing is smaller than the spacing in PEA₂PbBr₄ perovskite. Unlike PEA₂PbBr₄, neither a peak shift nor broadening of the (00l) planes was observed upon doping, indicating these planes are unaffected after the incorporation of Mn²⁺ in BA₂PbBr₄. Interestingly, in addition to the (00l) family peaks, another peak at ∼15.5°, corresponds to the (111) plane, was observed (Fig. 1d). This (111) plane cuts the organic and inorganic layers diagonally, as shown in Fig. 1b.⁹ Upon a close examination, a slight shift in the ∼15.5° peak to higher angles was observed (magnified region of Fig. 1d). This shift was more prominent at higher (>1%) Mn²⁺ doping, and almost negligible at lower doping levels. The magnitude of this shift was Δ2θ = 0.04°. The shift in this peak to higher angular scale implies there is a lattice contraction along the out-of-plane direction along the ab-plane (perpendicular to the (111) plane and the c-axis), as shown in Fig. 1b. This is further corroborated by the decrease in d-spacing of (111) planes, as shown in Fig. 2b. From the crystal structure in Fig. 1b, this ab plane represents the plane of the inorganic framework; the lattice contraction along this plane implies that there is shrinkage of the inorganic framework. This is possible only when the central atom, Pb²⁺, is replaced by an atom with a smaller ionic radius such as Mn²⁺. Therefore, from these analyses, it is evident that the Mn²⁺ in BA₂PbBr₄ is substituting Pb²⁺, whereas, in the case of PEA₂PbBr₄, the Mn²⁺ is occupying interstitial sites.^6,10 This difference in occupancy, despite the similar layered RP type structure, could be attributed to either the smaller interlayer spacing in BA₂PbBr₄ compared to PEA₂PbBr₄ or a difference in the lattice rigidity owing to difference in their organic spacers (aromatic vs. aliphatic). Perhaps this substitutional nature of Mn²⁺ doping is limiting the amount of Mn²⁺ that can be incorporated into BA₂PbBr₄, since a maximum of only ∼2% doping was achievable synthetically.

Regardless of the crystal phases, type of spacers, and interstitial or substitutional doping, the incorporation of Mn²⁺ into the perovskite host can induce strain in the lattice. Generally, from the changes in the p-XRD peaks, one can estimate the strain in the lattice. A peak shift with a slight broadening in PEA₂PbBr₄ after Mn²⁺ doping may be an indication of a homogenous strain caused by the uniform expansion of the lattice. Whereas a moderate change in peak position in the case of BA₂PbBr₄, suggesting that the magnitude of the lattice strain is less in the case of BA₂PbBr₄.¹¹ To quantitatively compare the microstrain induced by Mn²⁺ incorporation in these two host lattices, we employed the Williamson–Hall method (Fig. S2, ESI†).¹² A Voigt function, which is convoluted Lorentzian and Gaussian functions, was used to fit the p-XRD peaks and derive the FWHM for the strain analyses. From Fig. 2c, a striking difference in the microstrain was observed at higher Mn²⁺ levels. At >2% of Mn²⁺ doping, the strain in the PEA₂PbBr₄ is approximately three times higher than that of BA₂PbBr₄, implying that the interstitial doping can induce more strain in the lattice than the substitutional doping.

Next, the impact of these discrepancies in the Mn²⁺ occupancy on the optical properties of perovskites was investigated. We recorded absorption, photoluminescence (PL), and exciton diffusion in PEA₂PbBr₄ and BA₂PbBr₄ perovskites with and without Mn²⁺ doping. As shown in Fig. S3 (ESI†), both pristine perovskites exhibit sharp excitonic features in the UV-vis absorption spectra with similar band-edge absorption at 2.89 eV (430 nm) for PEA₂PbBr₄ and at 2.85 eV (435 nm) for BA₂PbBr₄ (Fig. S4, ESI†). The absorbance remains unchanged upon Mn²⁺ doping of the pristine 2D perovskites. Supporting this, we observed no change in the excitation (PLE) spectra after Mn²⁺ doping (Fig. S6, ESI†). The PL spectra of both pristine 2D perovskites exhibit two emission features around 410 nm and 430 nm, similar to other studies.¹³ As these are 2D layered perovskites, the lower energy peak at 430 nm is due to the recombination of edge excitons, while the higher energy transition at 410 nm is designated to interior excitons.^10,13 When Mn²⁺ is introduced into both BA₂PbBr₄ and PEA₂PbBr₄, a new broad, intense emission at 600 nm was observed. This 600 nm emission has been assigned to ⁴T_1g → ⁶A_1g d–d transition in Mn²⁺.^6,10 These Mn states are typically sensitized from the host's excitation, and the emission originates from the excitonic energy transfer from the host to the dopant ion (Fig. S5, ESI†).⁶ Interestingly, the intensity of the Mn²⁺ peak (600 nm) with respect to the host peak (∼430 nm) is significantly different in Mn²⁺:BA₂PbBr₄ compared to Mn²⁺:PEA₂PbBr₄ (Fig. S7, ESI†); this is true even in the case of samples with the same amount of Mn²⁺ (Fig. S13, ESI†). For instance, the PL intensity ratio between dopant and host emission in ∼2% Mn²⁺-doped PEA₂PbBr₄ suggests an equal intensity dopant peak with the host. Whereas in BA₂PbBr₄, at the same Mn²⁺ loading, the intensity of the dopant peak was only ∼1/4th of the host peak (Fig. S7, ESI†). Consequently, under the 365 nm excitation, the color of Mn²⁺:PEA₂PbBr₄ crystals appeared significantly different from Mn²⁺:BA₂PbBr₄ (Fig. S9–S11, ESI†). The emission color tends to move towards orange (λ_Em = 600 nm) in Mn²⁺:PEA₂PbBr₄; however, a more prominent blue emission in Mn²⁺:BA₂PbBr₄ was observed. These deviations in the PL spectra of Mn²⁺:BA₂PbBr₄ and Mn²⁺:PEA₂PbBr₄ crystals show a clear distinction in the CIE 1931 (Commission Internationale de l’Eclairage) chromaticity diagram in Fig. S14 (ESI†).

Furthermore, the time-resolved photoluminescence (TRPL) was recorded at 430 nm and 435 nm for both undoped PEA₂PbBr₄ and BA₂PbBr₄ on the order of nanoseconds (Fig. 3c, and f). The average lifetimes decrease with an increase in Mn²⁺ levels, as shown in Tables S4 and S5 (ESI†). Under the same excitation fluence, the average lifetime of undoped PEA₂PbBr₄ (∼8.2 ns) is significantly higher than undoped BA₂PbBr₄ (∼3 ns). With a small amount of Mn²⁺ doping, there is a dramatic decrease in the lifetime of the PEA₂PbBr₄ when compared to BA₂PbBr₄, indicating an efficient host-dopant energy transfer in Mn²⁺:PEA₂PbBr₄. This also complements the intense Mn²⁺ PL in Mn²⁺:PEA₂PbBr₄ compared to Mn²⁺:BA₂PbBr₄. Subsequently, the PLQY of the host is decreased (Fig. 3c) and the Mn²⁺ is increased with an increase in the Mn²⁺ doping (Fig. 3f). Both the Mn²⁺ doped 2D perovskites showed a similar decay profile with a prolonged lifetime on the order of microseconds (Fig. S8, ESI†) regardless of the host structure due to the forbidden nature of Mn²⁺ d–d (⁴T_1g → ⁶A_1g) transitions.¹⁰ As expected, the PLQY_Mn of this Mn²⁺ emission at 600 nm increased with an increase in the Mn²⁺ doping in both PEA₂PbBr₄ and BA₂PbBr₄. However, this increase in PLQY_Mn is more prominent in PEA₂PbBr₄ compared to BA₂PbBr₄, reaching a maximum of 43.8% (Fig. 3f). From all these studies, it is evident that the host-to-dopant ion energy transfer is more efficient in the case of Mn²⁺:PEA₂PbBr₄ compared to Mn²⁺:BA₂PbBr₄. Such differences in optical properties hints towards the important contributions from the nature of doping. In addition, the difference in their crystal phases, and the lattice rigidity due to the difference in their organic spacers may play a critical role in the energy transfer. Regardless, introducing foreign Mn²⁺ ions into the lattice can further perturb the host structure, consequently, the dynamics of exciton transport and the host-to-dopant energy transfer can be affected. After photoexcitation, the excitons generated in the host perovskites diffuse within the inorganic framework before they get trapped in the Mn²⁺ states. Therefore, the nature of doping, the location of the dopant ions, and strain in the inorganic framework play a critical role in exciton diffusion and energy transport dynamics.^14,15 The efficient host-dopant energy transfer despite the greater strain in Mn²⁺:PEA₂PbBr₄ perovskites could be due to either (i) the interstitial doping facilitating the smooth sailing of excitons without affecting the landscape of 2D inorganic framework or (ii) the energy transfer barrier for Mn²⁺ may be lower in Mn²⁺:PEA₂PbBr₄. The weak Mn²⁺ emission in the case of Mn²⁺:BA₂PbBr₄ perhaps due to the strain created within the inorganic framework due to the lattice contraction by substitutional doping. This strain may be affecting the exciton transport by creating non-radiative defects caused by an inhomogeneous landscape. Consequently, these excitons decay non-radiatively before reaching Mn²⁺ states. Also, the difference in contribution towards the non-radiative sites from rigid aromatic PEA cations and labile aliphatic BA cations cannot be ruled out.

Fig. 3 (a) PL spectra of Mn²⁺:PEA₂PbBr₄. (b) TRPL of Mn²⁺:PEA₂PbBr₄. (c) Host PLQY against Mn²⁺ percentage. (d) PL spectra of Mn²⁺:BA₂PbBr₄. (e) TRPL of Mn²⁺:BA₂PbBr₄. (f) Dopant PLQY against Mn²⁺ percentage.

To unveil these dynamics and study the impact of the different doping nature of the two perovskites on exciton transport, we employ TPLM, which allows for direct visualization of in-plane exciton transport. For this, we specifically chose Mn²⁺-doped PEA₂PbBr₄ and BA₂PbBr₄ with comparable Mn²⁺ doping percentages. Fig. 4a shows the resulting diffusion maps of the undoped and Mn²⁺ doped materials. For both undoped perovskites (Fig. 4a, lower panels), fast spatial expansion of the initial exciton population is observed. In contrast, for doped perovskites, the initial fast diffusion is followed by a stagnation of the expansion, consistent with excitons reaching dopant sites and undergoing energy transfer. We can quantify the rate of expansion by fitting the variance σ of the population for each time slice and calculating the mean-square displacement (MSD(t) = σ_t² − σ_0⁻), as shown for the different doping levels in Fig. 4b and c for PEA₂PbBr₄ and BA₂PbBr₄, respectively. From the early time dynamics (t < 1 ns), we can extract the diffusivity of the exciton population in the lattice (MSD = 2D_t), as indicated by the dashed lines in Fig. 4b and c. For the undoped cases, we observe a higher diffusivity for PEA₂PbBr₄ (0.22 cm² s⁻¹) compared to BA₂PbBr₄ (0.13 cm² s⁻¹). This is consistent with earlier reports and can be explained by the difference in the softness of the inorganic lattice and a stronger exciton–phonon coupling in BA₂PbBr₄, which slows down transport.¹⁶ Upon doping, it is interesting to note that the diffusivity PEA₂PbBr₄ remains unaffected (Fig. 4b). In contrast, the expansion is notably slower for the highest (2%) doping concentration in BA₂PbBr₄, where the diffusivity drops to 0.08 cm² s⁻¹. Interestingly the same sample, BA₂PbBr₄ with 2% Mn²⁺ doping, showed more prominent peak shift in the p-XRD owing to the lattice distortions along the inorganic 2D framework, highlighting the important role of lattice distortions in the exciton diffusion. Further, the lower diffusivity in BA₂PbBr₄ can partially explain the lower PLQY of the Mn²⁺ doping in that system, with slower diffusion leading to less efficient energy transfer.

Fig. 4 (a) Diffusion maps of undoped and Mn²⁺ doped PEA₂PbBr₄ and BA₂PbBr₄. The extracted mean-square displacement as a function of time for (b) PEA₂PbBr₄ and (c) BA₂PbBr₄ with different doping levels.

In conclusion, Mn²⁺ doping induces distinct structural changes by occupying different preferential sites in PEA₂PbBr₄ and BA₂PbBr₄ 2D perovskite hosts lattice, interstitial in the former and substitutional in the latter. Interestingly, the resultant crystal shows intense Mn²⁺ emission, which is more efficient in PEA₂PbBr₄ compared to BA₂PbBr₄. The discrepancy in optical properties is most likely due to the nature of doping and its impact on the crystal lattice, including dopant-induced in-plane and out-of-plane lattice strain. Consequently, we observed a striking difference in the exciton diffusion. However, a few aspects like the interstitial site of Mn²⁺ in crystal, the local structure around Mn²⁺ at the interstitial site, and the role of organic spacers require further exploration.

This work was partially funded by the European Union (ERC, EnVision, project number 101125962). FP acknowledges support from the Spanish Ministry of Science and Innovation (CEX2023-001316-M). MG, AS, and UK acknowledge the support from DOE (DE-SC0024214).

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

There are no conflicts to declare.

Notes and references

1. M. D. Smith, et al. , Chem. Rev., 2019, 119, 3104–3139.
2. Y. Li, et al. , ChemComm, 2022, 58, 941–944.
3. A. J. Torma, et al. , ACS Nano, 2021, 15, 20550–20561.
4. S. K. Dutta, et al. , ACS Energy Lett., 2019, 4, 343–351.
5. S. Hou, et al. , Joule, 2018, 2, 2421–2433.
6. T. Sheikh and A. Nag, J. Phys. Chem. C, 2019, 123, 9420–9427.
7. F. C. Grozema, et al. , J. Phys. Chem. C, 2020, 124, 28201–28209.
8. C. Zhou, et al. , Mater. Sci. Eng., R, 2019, 137, 38–65.
9. S. A. Cuthriell, et al. , Adv. Mater., 2022, 34, 2202709.
10. A. Biswas, et al. , Chem. Mater., 2017, 29, 7816–7825.
11. A. Khorsand Zak, et al. , Solid State Sci., 2011, 13, 251–256.
12. R. Kamal Yadav and P. Chauhan, Indian J. Pure Appl. Phys., 2019, 57, 881–890.
13. T. Jin, et al. , Nat. Commun., 2023, 14, 1–9.
14. D. B. Straus and C. R. Kagan, J. Phys. Chem. Lett., 2018, 9, 1434–1447.
15. C. M. Mauck and W. A. Tisdale, Trends Chem., 2019, 1, 380–393.
16. M. Seitz, et al. , Nat. Commun., 2020, 11, 1–8.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4cc04074k

---