Scalable photonic sources using two-dimensional lead halide perovskite superlattices

The study addresses how to realize scalable, decoupled multi-quantum-well (MQW) superlattices from 2D semiconductors whose bright excitons possess predominantly in-plane (IP) transition dipole moments (TDMs). Interlayer coupling in stacked 2D systems typically induces charge-transfer (CT) and momentum-forbidden dark excitons, introducing out-of-plane (OP) dipole components, reducing the IP dipole ratio (R_IP), and quenching photoluminescence via long-lived nonradiative pathways. A central question is what quantum barrier (QB) thickness and material properties are required to fully decouple adjacent 2D layers, and how this requirement correlates with the 2D layer thickness (d_2D) and the QB thickness (d_QB). Previous systems often needed thick barriers yet still showed coupling. The purpose here is to elucidate principles that enable interlayer decoupling with atomically thin QBs, develop scalable fabrication of high-order superlattices, and establish spectroscopic methods (k-space PL) to quantify TDM orientation and interlayer coupling.

Prior approaches to assemble 2D heterostructures and superlattices include layer-by-layer assembly, van der Waals epitaxy, intercalation, and colloidal chemistry, but high-order MQW superlattices with controlled decoupling at scale remain elusive. In CdSe nanoplatelet stacks (d_2D ~15 Å), strong interlayer coupling persists even with d_QB ~37 Å. In monolayer WSe2/MoS2 (d_2D = 6.2 Å), trilayer h-BN (d_QB = 13 Å) reduces but does not eliminate CT exciton emission. Interlayer coupling typically lowers R_IP (e.g., CdSe NPL monolayer R_IP ~0.95 vs multilayer ~0.67, approaching isotropic orientation) and reduces PL due to Auger and other quenching. Conventional III–V MQWs aim for strong emission without multiexciton quenching; similarly, 2D systems motivate insertion of atomically thin QBs to screen coupling while retaining high-efficiency emission. The need to relate decoupling to both d_QB and d_2D is emphasized.

Materials fabrication: Colloidal quantum wells (CQWs) of lead halide perovskites with formula (RNH3)2[CH3NH3PbBr3]nPbBr4 were synthesized in solution (n = 3 selected; d_2D ≈ 24 Å). Organic ligands of varying alkyl chain length (R) provided low-ε (≈2) quantum barriers, enabling control of d_QB; efforts with R shorter than C5H11 were not stable. CQWs were assembled into MQW superlattice thin films on glass with the c-axis perpendicular to the substrate plane. Film thickness (t) and refractive index dispersion (n_SL(λ)) were obtained by spectroscopic ellipsometry (Sellmeier + Tauc-Lorentz modeling). Structural characterization: Synchrotron GIWAXS (BL13A, NSRRC; 12.16 keV; incidence 0.12°) provided superlattice d-spacings via Laue spots along q_z. Together with d_2D, this yielded d_QB and the number of stacking layers (N). STEM provided CQW and cross-sectional superlattice morphology. Optical k-space spectroscopy and dipole orientation: Polarization- and angle-dependent photoluminescence (PL) was measured using a hemispherical glass prism setup (Fluxim Phelos) to access emission beyond total internal reflection (k/k0 > 1). s- and p-polarized emission patterns were recorded and converted to k-space. A transfer-matrix-based dipole emission model (Setfos) simulated radiation patterns of an ensemble of emitters uniformly distributed through the superlattice. Using measured t and n(λ), the in-plane dipole ratio R_IP was the sole fitting parameter for the p-polarized profiles; s-polarized patterns were then computed for consistency. Absolute PL quantum yields (η_PL) were measured in an integrating sphere. Temperature-dependent PL and TRPL: Steady-state PL from room temperature (RT) to 77 K and time-resolved PL (Hamamatsu Quantaurus-Tau, 365 nm excitation) were used to probe interlayer coupling dynamics. A cryostat enabled low-T measurements; dynamics of A (band-edge) and emergent I (interlayer/CT-related) excitons were analyzed, including power-law kinetics at low T. Electronic structure calculations: DFT (VASP) with GGA-PBE and many-body dispersion corrections modeled bulk and 2D MAPbBr3 (n = 0, 3). Band structures and orbital-resolved contributions to VBM/CBM were computed. Many-body perturbation theory (GW0) and Bethe–Salpeter equation (BSE) for n = 0 provided polarized optical absorption and exciton character (A, B, C). Calculations included vdW interactions; k-point meshes: 8×8×1 (2D) and 8×8×8 (bulk). GW with 180 bands; BSE with 4 occupied/6 unoccupied states. Anion exchange (AE): Iodide doping of OLAm-capped CQWs (d_QB ≈ 25.6 Å) tuned emission (e.g., target wavelengths 534, 557, 571 nm) by varying MAI amount and reaction time. Post-reaction purification ensured colloidal stability. GIWAXS confirmed preserved ordering in mixed-halide superlattices.

- Decoupled MQW superlattices with ultrathin organic QBs achieved from perovskite CQWs (n = 3; d_2D ≈ 24 Å) across a wide range of d_QB, including as small as 6.5 Å (comparable to monolayer h-BN). - k-space PL and optical modeling revealed predominantly in-plane exciton TDMs with high R_IP that are essentially independent of stacking layer number N and d_QB: • For d_QB ≈ 25.6 Å, measured/theory-fitted R_IP for N = 1, 2, 4, 10, 19 were 0.84, 0.85, 0.85, 0.83, 0.81, respectively. • Across d_QB series (6.5–24.2 Å) and varying N, R_IP remained ≳0.8 and insensitive to stacking. - Absolute PL quantum yield (integrating sphere) of superlattices reached up to ~0.85; relative η_PL was nearly independent of N even at the smallest d_QB = 6.5 Å, indicating minimal interlayer quenching and preserved 2D optical properties. - Temperature-dependent PL uncovered interlayer coupling only at low T: an additional emission (I exciton, ~490 nm) emerged below ~120 K in superlattices (absent in dilute solution). Dynamics at 77 K: • A exciton exhibited power-law decay (diffusion-controlled interlayer CT). • I exciton showed ultralong delayed emission with τ ≈ 431 ns. At RT, A exciton decay was monoexponential with τ ≈ 9.5 ns. - Mechanism: Strong ionic dielectric screening in lead halide perovskites (ε_optic ~4.5; ε_ion ~25; effective ~30) reduces interlayer electrostatic interactions; a simple model yields 1s CT exciton binding energy <10 meV (< k_BT at RT), rationalizing decoupling at RT and the onset of coupling when phonons freeze below ~150 K. - DFT/GW-BSE: 2D MAPbBr3 shows a shift of direct bandgap from R (bulk) to M (2D); VBM/CBM in 2D have predominantly IP orbital character (Pb pz and px contributions to CBM largely vanish), leading to strongly IP-polarized optical transitions. BSE reproduced A, B, C excitations only for IP polarization; IP dipole matrix elements exceed OP by ≥ an order of magnitude in 2D, unlike bulk. - Mixed-anion (iodide-doped) superlattices retained high ordering and predominantly IP TDMs (R_IP > 0.8) while enabling continuous emission tuning across the visible (e.g., green to orange; representative peaks at ~534, 557, 571 nm).

The work demonstrates that high-order 2D perovskite MQW superlattices can be fabricated at scale with atomically thin organic barriers that effectively decouple adjacent layers at room temperature. The consistent, high in-plane dipole ratios across stacking numbers and barrier thicknesses indicate that interlayer coupling is strongly suppressed, addressing the central challenge identified in prior 2D superstructures where thick barriers were insufficient to prevent coupling. The improvement in light outcoupling due to predominantly IP TDMs directly benefits photonic source efficiency. The low-temperature emergence of an interlayer-related I exciton, along with diffusion-controlled dynamics, corroborates a temperature-dependent coupling picture governed by the ionic lattice’s dielectric response: at RT, strong phonon-assisted ionic screening diminishes CT binding below thermal energy, preventing CT state population; upon cooling, reduced screening allows interlayer states to manifest. First-principles calculations substantiate the experimental findings by revealing IP-dominant band-edge states and IP-polarized excitonic transitions in 2D MAPbBr3. The ability to tune emission wavelength via anion exchange without sacrificing R_IP further underscores the platform’s versatility for integrated quantum and optoelectronic devices.

This study establishes a scalable materials platform for miniaturized photonic sources using 2D lead halide perovskite MQW superlattices with ultrathin organic quantum barriers. The superlattices exhibit predominantly in-plane transition dipole moments (R_IP ≳ 0.8), high and N-independent PL quantum yields, and enhanced light outcoupling, all while maintaining interlayer decoupling even with barriers as thin as 6.5 Å. Temperature-dependent spectroscopy and many-body calculations elucidate that strong ionic dielectric screening localizes intralayer Wannier-Mott-like excitons at room temperature, with interlayer CT states emerging only at low temperatures. Emission can be continuously tuned via anion exchange without degrading dipole orientation. These advances pave the way for efficient, narrowband, and wavelength-tunable quantum emitters applicable to nanoantennas, LEDs, and other photonic devices. Future work should further resolve the nature of the low-temperature I exciton, explore device integration and electrical excitation, expand to other compositions and barrier chemistries, and investigate ultimate thickness limits for perfectly IP dipole orientation.

- The precise nature of the low-temperature I exciton (e.g., CT versus trap-assisted or trion-related processes) requires further photophysical characterization. - Interlayer coupling is suppressed at room temperature but appears at low temperatures; device performance under cryogenic conditions may differ. - Stability constraints limited the shortest ligand length (R < C5) and thus the minimum achievable d_QB during synthesis. - Most detailed studies focused on n = 3 (d_2D ≈ 24 Å); generality across broader thicknesses and compositions warrants further validation. - The study emphasizes optical characterization; demonstration within completed optoelectronic devices (e.g., electrically driven LEDs) is outside the present scope.