Revealing the competing contributions of charge carriers, excitons, and defects to the non-equilibrium optical properties of ZnO

Abstract

Due to its wide band gap and high carrier mobility, ZnO is, among other transparent conductive oxides, an attractive material for light-harvesting and optoelectronic applications. Its functional efficiency, however, is strongly affected by defect-related in-gap states that open up extrinsic decay channels and modify relaxation timescales. As a consequence, almost every sample behaves differently, leading to irreproducible or even contradicting observations. Here, a complementary set of time-resolved spectroscopies is applied to two ZnO samples of different defect density to disentangle the competing contributions of charge carriers, excitons, and defects to the nonequilibrium dynamics after photoexcitation: time-resolved photoluminescence, excited state transmission, and electronic sum-frequency generation. Remarkably, defects affect the transient optical properties of ZnO across more than eight orders of magnitude in time, starting with photodepletion of normally occupied defect states on femtosecond timescales, followed by the competition of free exciton emission and exciton trapping at defect sites within picoseconds, photoluminescence of defect-bound and free excitons on nanosecond timescales, and deeply trapped holes with microsecond lifetimes. These findings not only provide the first comprehensive picture of charge and exciton relaxation pathways in ZnO but also uncover the microscopic origin of previous conflicting observations in this challenging material and thereby offer means of overcoming its difficulties. Noteworthy, a similar competition of intrinsic and defect-related dynamics could likely also be utilized in other oxides with marked defect density as, for instance, TiO 2 or SrTiO 3 .

Ultrafast generation and decay of a surface metal

“…Beyond chemical doping, shallow dopants could result from photoexcitation of deep defects as outlined in the introduction if their lifetime is sufficiently long to be pumped and probed by two subsequent laser pulses provided by our laser system (200 kHz ≙ 5 μs). As the luminescence of defect excitons in ZnO is known to extend to the μs regime 27 , formation of such a photostationary state of long-lived defect excitons would actually be expected. We test this hypothesis by tuning the repetition rate of our laser system from normally 200 kHz down to 5 kHz, which varies the separation of two subsequent laser pulses from 5 to 200 μs.…”

“…Consequently, the surface is charged even more positively, which leads to an increase in surface BB. A resolution-limited upper boundary of 80 ps for this process was recently identified by X-ray absorption spectroscopy 54 and few-ps time constants were determined by optical spectroscopy 27 . However, we note that hole polaron formation 55 may result in similar dynamics.…”

“…Among them is the wide band gap (3.4 eV), intrinsically n-doped ( E F ca. 0.2 eV below the CB) semiconductor ZnO, which has a variety of native deep donor defects leading to a broad photoluminescence signal below the fundamental band gap energy 27 , 28 . For ZnO, a photoinduced, ultrafast control of the conduction properties would be especially appealing, as any application would benefit from the ease of nanostructuring and transparency to visible light of this material 29 , 30 .…”

Rules for Selecting Metal Cocatalyst Based on Charge Transfer and Separation Efficiency between ZnO Nanoparticles and Noble Metal Cocatalyst Ag/ Au/ Pt

“…When Pt is loaded, the Schottky barrier would prevent the backward electron transfer from Pt to TiO 2 and preventing the charge recombination in the TiO 2 phase . Secondly, at room temperature (kT≈25 meV), the photogenerated carriers exist as binding excitons in ZnO (binding energy 60 meV), while almost free carriers for TiO 2 (binding energy 10 meV for anatase and free carriers for rutile) . Therefore the efficiency of electron transfer from ZnO NPs to Pt metal cocatalyst would be significantly lower than that from TiO 2 to Pt.…”

“…A simplified model for carrier relaxation can be described in Figure c and 2d for ZnO and metal‐ZnO respectively (details in Description for the carrier relaxation model of SI). In this model, as the binding energy of exciton in ZnO is 60 meV, the excited carriers are in a form of thermal equilibrium for free electrons and excitons. N 1 refers to the photogenerated carriers which will decay to the shallow trapped carriers at a rate of k 13 , N 2 denotes the emissive excitons at a decay rate of k 2 , N 3 denotes the shallow trapped carriers having a recombination rate of k 3 .…”