Regarding the latter, for example, the basic issue such as "the identity of optical excitation at room temperature" is still controversial, because there are contradicting evidences for two candidates, free carriers and excitons. [9,10] The key question is the stability of the exciton against thermal ionization of the constituents. The exciton binding energy ϕ X in 3D HPs such as MAPbX 3 (MA = CH 3 NH 3 , X = I, Br, Cl) ranges from 15 meV to 70 meV, [9,11,12] which is indeed comparable to room temperature (≈26 meV), and therefore, an exciton may or may not be stable. However, optical excitation is certainly excitonic at low temperatures, for example, 10 K (≈0.9 meV).Like a hydrogen molecule formed by binding of two hydrogen atoms, two excitons can bind to form an excitonic molecule, typically known as a biexciton. The numerical value for the biexciton binding energy ϕ XX can be roughly estimated by a theoretical model, which predicts its universal dependence on ϕ X and the effective mass ratio σ between the electron and the hole. [13][14][15] According to variational methods, [16,17] the predicted ϕ XX for MAPbBr 3 (ϕ X = 38 meV [9] and σ ≈ 1 [18] ) can range from 1.0 meV to 4.8 meV, which is sufficiently large to ensure the stability of the biexciton at nominal cryogenic temperatures. However, there is no single experimental evidence for their existence in the 3D HPs, which is quite surprising because biexcitons are well established in conventional semiconductors like Si and Ge with similar and even much smaller values of ϕ XX = 1.5 meV [19] and 0.3 meV, [20] respectively.Here, we demonstrate that biexcitons do exist in 3D MAPbBr 3 single crystals at low temperatures (T < 30 K). However, biexcitons are unstable at the crystal exposed to ambient, explaining their elusive nature when probing as-grown crystals with typical photoluminescence (PL) spectroscopy. Therefore, the freshcut pristine quality of the crystal should be established for the stable biexciton formation. The evidence of the biexcitonic phase is very clear from the rapidly growing PL peak having an inverted Boltzmann shape, reflecting the kinetic energy distribution of the biexcitons. [21,22] We show that a cold biexciton can be directly generated using giant resonant two-photon absorption (2PA) with polarization control aided by the nanoscale surface flatness of the crystal with fine-scale photoluminescence Halide perovskites (HPs) are fascinating materials whose optoelectronic properties are arguably excitonic. In the HP family, biexcitons are known to exist only in low dimensions where exciton-exciton binding is strongly enhanced by quantum and dielectric confinements. In this paper, however, it is shown that they indeed do exist in 3D bulk CH 3 NH 3 PbBr 3 (MAPbBr 3 ) single crystals if the pristine crystal quality is ensured for subtle binding of two excitons. The existence of biexcitons is clearly evidenced below 30 K with a binding energy of ≈3.9 ± 0.3 meV according to i) exciton-biexciton population dynamics, ii) giant resonant two-photon ex...

Two-dimensional (2D) semiconductors have emerged as an excellent platform for studying various excitonic matter under strong quantum and dielectric confinements. However, such effects can be seriously overestimated for Coulomb binding of two excitons to form a biexciton by a naive interpretation of the corresponding photoluminescence (PL) spectrum. By using 2D halide perovskite single crystals of [CH3(CH2)3NH3]2Pb1–x Mn x Br4 (x = 0–0.09) as a model system, we investigated both population and relaxation kinetics of biexcitons as a function of excitation density, temperature, polarization, and Mn doping. We show that the biexciton is formed by binding of two dark excitons, which are partially bright, but they radiatively recombine to yield a bright exciton in the final state. This renders the spectral distance between the exciton peak and the biexciton peak as very different from the actual biexciton binding energy (ϕ) because of large bright–dark splitting. We show that Mn doping introduces paramagnetism to our 2D system and improves the biexciton stability as evidenced by increase in ϕ from 18.8 ± 0.7 to 20.0 ± 0.7 meV and the increase of the exciton–exciton capture coefficient C from 2.4 × 10–11 to 4.3 × 10–11cm2/ns within our doping range. The precisely determined ϕ values are significantly smaller than the previously reported ones, but they are consistent with the instability of the biexciton against thermal dissociation at room temperature. Our results demonstrate that electron–hole exchange interaction must be considered for precisely locating the biexciton level; therefore, the ϕ values should be reassessed for other 2D halide perovskites that even do not exhibit any dark exciton PL.