betaFeedback

WorksPhysics and Astronomy

Polaronic-Distortion-Driven Enhancement of Excitonic Auger Recombination in ZnO Nanoparticles

Abstract

The surface effect and quantum confinement render nanomaterials the optoelectronic properties more susceptible to nonradiative processes than their bulk counterparts. These nonradiative processes usually contain a series of interwoven and competing sub-processes, which are challenging to disentangle. Here, we investigate the structural origin of Auger recombination in ZnO nanoparticles using transient absorption spectroscopy and ultrafast electron diffraction. The photogenerated hot holes are captured by oxygen vacancies through an Auger mechanism, inducing significant local structural distortions around the oxygen vacancy and its neighboring zinc tetrahedron on a sub-picosecond timescale. The recombination of trapped holes accelerates the lattice thermalization and stabilizes the formed small hole polarons. Subsequently, the recombination of localized polarons forms a confined exciton-polaron complex that may account for the long-lived (>7 ns) visible luminescence observed in ZnO nanoparticles. Our findings are potentially applicable to other transition metal oxide nanomaterials, bringing insights for the optimization of their functional properties.

Zinc oxide nanoparticles (ZnO-NPs) are widely adopted in many applications, for example, as the electron transport layer in perovskite solar cells for their many advantages, [1] including the favorable optoelectronic properties and the potential of large-scale printing production.[2, 3] Undoped ZnO nanomaterials possess an intrinsic n-type conductivity, which is attributed to the abundant defect sites, particularly, to the oxygen vacancies near surface.[4, 5] Although a wide range of defect types have been identified,[6] the underlying mechanisms of defect-mediated carrier relaxation and transport are not yet fully elucidated. Using oxygen vacancy, one of the major species, as an example, the temperature-dependent measurements demonstrated that the luminescence in visible range was associated with the recombination of electrons at the shallow defect levels of doubly ionized oxygen vacancies, VO++V_{\mathrm{O}}^{++}, which were generated through a two-step hole capturing process of VO→VO+→VO++V_{\mathrm{O}} \to V_{\mathrm{O}}^{+} \to V_{\mathrm{O}}^{++};[7, 8] The First-principles calculations revealed the deep donor levels formed by these vacancies,[9] and the local lattice distortion up to 23% of the Zn-O bond length.[10] However, so far no experimental observation on structural evolution testifies these scenarios.

To address the role of oxygen vacancies, pump-probe based techniques have been employed in the investigations of defect-associated transitions in ZnO-NPs. For example, the transient absorption (TA) spectroscopy revealed the hopping carrier diffusion within the energetically continuous defect states,[11] and identified the polaron excited state that injected an additional electron into the conduction band (CB) keywords: ZnO nanoparticles; small polaron; Auger recombination; lattice distortion; ultrafast electron diffraction with prolonged lifetime.[12] The time-resolved X-ray photoelectron spectroscopy confirmed that the hole capture by oxygen vacancies induced a local structural distortion corresponding to 15% of the Zn-O bond length within 80 ps,[13] consistent with the computational simulations.[10] These observations obtained only, however, the carrier-derived information through differential spectroscopy, indirectly inferring the impact of structural responses on the process of defect capturing. Investigations into atomic motions are essential to elucidate the complex carrier–lattice interactions related to the self-trapping and polaron formation in ZnO-NPs. Compared to TA, which probes the changes of the joint density of carrier states[14], ultrafast electron diffraction (UED) captures the lattice evolutions in reciprocal space,[15] e.g., phonon–phonon scattering,[16] anisotropic atomic motions,[17] structural distortions,[18] and so on.[19, 20, 21] The sensitivity of direct structure information makes UED a suitable approach to assess the carrier capture and lattice deformation in ZnO-NPs.

In this work, we investigate the formation of exciton–polarons in ZnO-NPs by combing the complementary probes of TA and UED, to uncover the structural origin underlying the polaronically enhanced excitonic Auger recombination. Upon photoexcitation, the surface oxygen vacancies capture the photogenerated holes within 200 fs, forming the small polarons that subsequently trap the additional electrons in CB to create the long-lived excitonic polarons. This process is accompanied by the local lattice distortion and the compression of Zn-O bonds, which are associated with the oxygen vacancy, enhancing the electron–phonon scattering and promoting the lattice thermalization through an Auger-mediated mechanism. At the end, such a polaron-associated relaxation pathway strongly impacts the overall energy relaxation in ZnO-NPs.

The investigated wurtzite-phase ZnO-NPs were synthesized via a methanolic sol-gel/precipitation method, and characterized by high-resolution transmission electron microscopy (HRTEM) and optical spectroscopies (see Experimental Section and Figure S1 in Supporting Information). The HRTEM images reveal the average particle diameter of 6–7 nm and the hexagonal lattice structure. The steady-state absorption spectrum exhibits a clear absorption edge, from which the bandgap is extracted as ∼3.36\sim3.36 eV. The photoluminescence (PL) spectrum features two discrete emission peaks centered at 3.1 eV (PLUV_{\mathrm{UV}}) and 2.3 eV (PLVIS_{\mathrm{VIS}}), constituting 5.1% and 94.6% of the total intensity, respectively. The former largely arises from the recombination of band-edge excitons[22] or localized excitons within the defect states close to band edge,[11] while the latter is primarily attributed to the luminescence associated with self-trapped excitons and defects.[23]

Sample characterizations: (a) HRTEM image, (b) Fast-Fourier transformation analysis, (c) steady-state absorption spectrum, and (d) photoluminescence spectrum

Figure S1. Sample characterizations. (a) HRTEM image. (b) Fast-Fourier transformation analysis of the selected area in (a). (c) Steady-state absorption spectrum. (d) Photoluminescence spectrum.

We first examine the carrier relaxation upon above-bandgap excitation of 320 nm using TA spectroscopy in transmission geometry. As shown in Figure 1a, the TA spectrum exhibits two photoinduced absorption signals at 2.4–3.0 eV (PIA1) and 3.4 eV (PIA2), and a photo-bleaching (PB) signal at 3.0–3.4 eV. The broadband PIA1 and PB dominate the early few picoseconds, then diminish significantly in ∼50\sim50 ps, leaving only PIA2 that lasts more than 7 ns, as depicted in Figure 1b.

Figure 1.

Image not extracted

Figure 1 (PDF p. 5). Hole capturing revealed by TA measurements. (a) Pseudocolor contour plot of the TA result upon excitation with fluence of 1.23 mJ/cm2^{2}. (b) Spectra extracted from (a) at selected delay times, shown with the PL spectrum for comparison. The blue shift between the PB peak and PLUV_{\mathrm{UV}} may arise from the large exciton binding energy.[24, 25] (c) PIA1 kinetics probed at various energies, extracted from (a). Solid lines, multiexponential fits. Inset, dependences of τ0PIA1\tau_{0}^{\mathrm{PIA1}} and ΔA0PIA1\Delta A_{0}^{\mathrm{PIA1}} on probe energy; Solid lines, linear fits. (d) PIA1 kinetics at 2.4 eV measured with various fluences. Solid lines, multiexponential fits. Inset, dependences of τ0PIA1\tau_{0}^{\mathrm{PIA1}} and ΔA0PIA1\Delta A_{0}^{\mathrm{PIA1}} on excitation fluence; Solid lines, the guides to the eye. See Table S1 and Table S2 in Supporting Information for the results of multiexponential fit.

Probe (eV)2.12.22.32.42.52.62.72.8
τ0PIA1\tau_{0}^{\mathrm{PIA1}} (ps)0.160.180.190.220.240.260.270.32
ΔA1PIA1\Delta A_{1}^{\mathrm{PIA1}} (%)48.743.443.843.640.540.138.432.5
τ1PIA1\tau_{1}^{\mathrm{PIA1}} (ps)5.715.175.996.576.447.438.458.55

Table S1. Fit results of the PIA1 kinetic traces in Fig. 1c in the main text.

Fluence (mJ/cm2^{2})0.120.330.611.231.351.471.842.443.67
ΔA0PIA1\Delta A_{0}^{\mathrm{PIA1}} (mOD)0.461.101.753.354.154.984.947.8811.15
τ0PIA1\tau_{0}^{\mathrm{PIA1}} (ps)0.160.150.140.160.170.150.160.150.14

Table S2. Fit results of the fluence-dependent kinetic traces of PIA1 in Fig. 1d in the main text.

The rise time of PIA1 (τ0PIA1\tau_{0}^{\mathrm{PIA1}}) increases with increasing probe energy while the maximum amplitude (ΔA0PIA1\Delta A_{0}^{\mathrm{PIA1}}) linearly decreases, as depicted in Figure 1c. We fit the PIA1 kinetics probed at various energies, yielding rise times of 160–320 fs, which align with the reported self-trapping times.[23, 26] Such a probe-energy-dependent rise suggests the carrier relaxation from the shallow defect levels to the deep ones after defect trapping,[26] [27] see Supplementary Discussion 1 in Supporting Information. Considering the high mobility of electrons in ZnO-NPs,28,29 holes are expected to be the major species of trapped carriers.30 This picture is corroborated by the fluence-dependent measurements, which show that ΔA0PIA1\Delta A_{0}^{\mathrm{PIA1}}, probed at 2.4 eV, scales linearly with increasing fluence without noticeable saturation, while τ0PIA1\tau_{0}^{\mathrm{PIA1}} decreases dramatically from 330 fs to 42 fs, as depicted in Figure 1d. This expedited rise is interpreted by the increased population of electron–hole pairs, which accelerate the lattice thermalization via enhanced Auger trapping31,32 due to the quantum confinement effects upon strong excitation.33 Moreover, the high exciton density in the initial stage after photoexcitation also induces an excitonic Auger recombination, resulting in the increased rate of hole capturing.[26] [34]

The band-like PB signal extending in 3.0–3.4 eV, which shares the same origin with the Urbach tail in steady-state absorption spectrum (see Figure S1 in Supporting Information), is jointly contributed by the continuous defect states near band edge and the states above CB. The rise of PB band (τ0PB\tau_{0}^{\mathrm{PB}}) occurs in 100–200 fs (see Table S3 in Supporting Information) is similar to τ0PIA1\tau_{0}^{\mathrm{PIA1}}, indicating the simultaneous excitation and capturing of both hot holes and electrons. In contrast, the decay of PB band (τ1PB\tau_{1}^{\mathrm{PB}}) slows with increasing probe energy, as depicted in Figure 2a. This trend further corroborates the picture of defect trapping, in which the defect states closer to band edge initially capture more electrons, resulting in the longer time spans of decay via recombination. [11] [35] As the probe energy approaches the band edge (~3.4 eV), the bleach signal overshoots sharply after the delay time of ~100 ps, giving rise to PIA2.

Bleach signal evolution and the carrier relaxation processes

Figure 2. Bleach signal evolution and the carrier relaxation processes. (a) Dependence of PB kinetics on probe energy. Solid lines, multiexponential fits. Inset, linear dependence of τ1PB\tau_{1}^{\mathrm{PB}} on probe energy.

Probe energy (eV)3.13.153.23.253.33.4
ΔA0PB\Delta A_{0}^{\mathrm{PB}} (mOD)−0.64−1.14−2.05−2.85−3.81−5.09
τ0PB\tau_{0}^{\mathrm{PB}} (ps)0.150.140.110.190.120.17
ΔA1PB\Delta A_{1}^{\mathrm{PB}} (mOD)1.191.672.463.214.084.28
τ1PB\tau_{1}^{\mathrm{PB}} (ps)3.85.66.27.59.110.9
ΔA2PB\Delta A_{2}^{\mathrm{PB}} (mOD)−0.59−0.55−0.45−0.36−0.142.12
τ2PB\tau_{2}^{\mathrm{PB}} (ps)278.5279.4280281.2280279.9

Table S3. Fit results of the energy-dependent kinetic traces of PB band in Figure 2a in the main text.

(b) Dependence of band-edge PB kinetics on excitation fluence. Solid lines, multiexponential fits. Inset, dependences of ΔA03.4 eV\Delta A_0^{3.4\ \mathrm{eV}} and ΔA0PIA2\Delta A_0^{\mathrm{PIA2}} on fluence; Solid lines, the guides to the eye. See Table S3 and S4 in Supporting Information for the results of multiexponential fit. (c) Fitted characteristic parameters of PB kinetics at 3.1 and 3.4 eV as a function of fluence. Solid lines, the guides to the eye. (d) Schematic illustration of carrier relaxation in ZnO-NPs, see text.

The band-edge PB peak induced by the bleach of free excitons,36 demonstrates a saturation of bleach amplitude (ΔA03.4 eV\Delta A_0^{3.4\ \mathrm{eV}}) upon excitation of 2.44 mJ/cm2^2, as shown in Figure 2b. Furthermore, the decay of band-edge PB shows a complex evolution. As depicted in Figure 2c, the first decay amplitude (ΔA13.4 eV\Delta A_1^{3.4\ \mathrm{eV}}) increases with increasing fluence, agreeing with the enhanced Auger recombination in nanoparticles37 and the well-studied non-radiative recombination in ZnO nanomaterials.36,38,39 Upon excitation of low fluence, the first decay time, τ13.4 eV\tau_1^{3.4\ \mathrm{eV}}, is significantly longer than that of PB band at 3.1 eV, τ13.1 eV\tau_1^{3.1\ \mathrm{eV}} (see Figure S2 and Table S5 in Supporting Information for details). τ13.4 eV\tau_1^{3.4\ \mathrm{eV}} is prolonged due probably to the long-lived CB electrons injected from the excited polaronic states with limited recombination channels,[12] and the non-radiative recombination of the free excitons with the holes trapped in deep defects.36,40 As we increase the excitation fluence, τ13.4 eV\tau_1^{3.4\ \mathrm{eV}} and τ13.1 eV\tau_1^{3.1\ \mathrm{eV}} change in opposite ways until the fluence reaches the shared saturation threshold of 2.44 mJ/cm2^2, see Figure 2c. The decreasing τ13.4 eV\tau_1^{3.4\ \mathrm{eV}} also agrees with the picture of enhanced Auger recombination, as the increased carrier density accelerates many-body interactions in confined nanoparticles.33 This trend together with the saturation of ΔA03.4 eV\Delta A_0^{3.4\ \mathrm{eV}} suggest a scenario of exciton absorption with photoinduced transparency, in which the exciton relaxation near band edge probably transitions from the exciton–phonon scattering to the exciton–exciton scattering, or to the electron–hole plasma generation.39 During this process, the constraint of defect-assisted Auger recombination41 results in the increasing τ13.1 eV\tau_1^{3.1\ \mathrm{eV}}.36 Upon excitation with saturation fluence, the recombination rates are governed by the Auger processes, leading to the convergence of τ13.1 eV\tau_1^{3.1\ \mathrm{eV}} and τ13.4 eV\tau_1^{3.4\ \mathrm{eV}}, regardless of the carrier origin of exciton or defect state. The following decay (characterized by τ23.4 eV\tau_2^{3.4\ \mathrm{eV}}) in several tens of picoseconds is assigned to the radiative recombination, and the much slower decay (characterized by τ33.4 eV\tau_3^{3.4\ \mathrm{eV}}) in several hundreds of picoseconds is corresponding to the relaxation into deep defect states, see Supplementary Discussion 2 in Supporting Information.

Fluence-dependent kinetics of the TA signal at 3.1 eV

Figure S2. Fluence-dependent kinetics of the TA signal at 3.1 eV. (a) Kinetic traces upon excitation of various fluences. Solid lines, multiexponential fits (See Table S5 for the fit results). (b) Upper panel: Dependence of the maximum amplitude of PB band at 3.1 eV on excitation fluence, showing a saturation

Fluence (mJ/cm2^2)0.120.330.611.231.351.471.842.443.67
ΔA03.1eV\Delta A_0^{3.1\mathrm{eV}} (mOD)−0.087−0.334−0.599−1.282−1.853−2.411−2.582−3.885−4.509
τ03.1eV\tau_0^{3.1\mathrm{eV}} (ps)0.10.190.10.10.10.10.110.10.15
ΔA13.1eV\Delta A_1^{3.1\mathrm{eV}} (mOD)0.230.6210.9751.7992.2372.7882.9354.2664.913
τ13.1eV\tau_1^{3.1\mathrm{eV}} (ps)34.034.995.696.056.316.656.736.92
ΔA23.1eV\Delta A_2^{3.1\mathrm{eV}} (mOD)−0.14−0.26−0.36−0.532−0.45−0.458−0.433−0.71−0.969
τ23.1eV\tau_2^{3.1\mathrm{eV}} (ps)296.2288.7284.3280.1275.9269.5263.8261254.3

Table S5. Fit results of the fluence-dependent kinetic traces of PB band at 3.1 eV in Figure S2.

In contrast, PIA2 demonstrates a non-monotonic trend of maximum amplitude (ΔA0PIA2\Delta A_0^{\mathrm{PIA2}}) change with increasing fluence, with a saturation at 1.23 mJ/cm2^2 (see the inset of Figure 2b), reflecting the recombination of defect-bound excitons which possess a long-lived lifetime (τ1PIA2\tau_1^{\mathrm{PIA2}}) more than 7 ns,[26] [42] see Supplementary Discussion 3 in Supporting Information.

At this point we are able to summarize the carrier relaxation dominated by defect trapping and Auger processes, as illustrated in Figure 2d. Upon above-bandgap excitation, hot holes are captured by the deep defect states, which are associated with oxygen vacancies, while hot electrons fill the CB edge and shallow defect states, in the similar time spans of τ0PIA1\tau_0^{\mathrm{PIA1}} and τ0PB\tau_0^{\mathrm{PB}}, respectively. Then the Auger recombination processes, with defects involved, occur during τ1PB\tau_1^{\mathrm{PB}} and τ13.4 eV\tau_1^{3.4\ \mathrm{eV}}, followed by the radiative recombination of re-excited extra electrons within τ23.4 eV\tau_2^{3.4\ \mathrm{eV}}. For the remaining electrons in CB, a slow capturing by the deep defects occurs in τ33.4 eV\tau_3^{3.4\ \mathrm{eV}}, resulting in the formation of exciton polarons through binding with the previously localized holes. Subsequently, the recombination of exciton polarons yields the green luminescence with a significantly extended lifetime of τ1PIA2\tau_1^{\mathrm{PIA2}}. Both the polaron formation and the subsequent evolutions are associated to lattice distortions, as discussed next.

The lattice distortion is revealed by UED measurements (see Experimental Section in Supporting Information). Across the probed scattering vector QQ, the radially averaged intensity of all indexed diffraction peaks decreases upon excitation of 343 nm, indicating the lattice thermalization and the increase of root-mean-square atomic displacement,[16] as shown in Figure 3a. Particularly, the (002) and (013) peaks exhibit derivative-like evolutions within 0–20 ps (see Figure S3 in Supporting Information), suggesting a local carrier–lattice interaction,43 which is a precursor phenomenon often linked to polaronic effects.44

Lattice dynamics of global thermalization and local structural distortion

Figure 3. Lattice dynamics of global thermalization and local structural distortion. (a) Differential diffraction intensity at selected delay times, excited with 2.21 mJ/cm22.21\,\mathrm{mJ}/\mathrm{cm}^{2}. Black line, background-subtracted intensity profile. Inset, recorded pattern. (b) Temporal evolutions of relative intensity change of the selected diffraction peaks with high signal-to-noise ratio. I0I_{0}, intensity measured before excitation.

Figure S3. Temporal evolutions of differential diffraction intensity around (002) and (013) peaks. (a) Intensity change at 2.31 Å−1^{-1} reaches its maximum at ~8 ps, indicating a significant left shift of the (002) peak, which intensity remains almost unchanged. The intensity change of the opposite-side shoulder of (002) peak at 2.45 Å−1^{-1} exhibits slow dynamics due to the asymmetric peak profile and the overlapping with adjacent peaks. Such a derivative-like profile change of peak shift suggests the transient modification of local potential landscape, which is probably associated with the presence of short-lived lattice polarizations caused by hot carrier localization.[26] (b) Intensity change at 4.15 Å−1^{-1} reaches its maximum at ~4 ps, accompanied with the slow dynamics of the opposite-side shoulder of (013) peak. Other crystallographic planes with normal vector exhibiting large projection along the cc-axis show similar peak profile changes.

Solid lines, biexponential fits (see Table S6 in Supporting Information for fit results). (c) Dependence of the intensity changes at delay time of 1000 ps on squared scattering vector. Inset, dependence of root-mean-square atomic displacement on fluence, derived from the (010), (002), (011), (102), and (110) peaks. (d) Comparison of the interplanar spacing changes of (013) and (010) to the inversed averaged intensity change (thermalization, see text). All traces are normalized. Inset, temporal evolution of distortion ratio. (e) Schematic illustration of the interplanar spacing changes contributed from zinc atom motions. Filled balls, atomic sites at equilibrium.

Planes(010)(011)(110)(013)(112)
A1A_1 (a. u.)0.00330.00580.01310.01920.0217
τ1\tau_1 (ps)13.915.616.922.117.8
A2A_2 (a. u.)0.00890.01080.02420.03050.0399
τ2\tau_2 (ps)95.085.282.3119.7129.4

Table S6. Fit results of the intensity drop kinetics of the five selected diffraction peaks in Figure 3b in the main text.

The intensity changes of diffraction peak show two-stage decays with characteristic times of ~17 ps and ~100 ps, as depicted in Figure 3b. The rapid process is comparable to those observed in other quantum dots,[18] [31, 45] reflecting the energy transfer from hot excitons to lattice via phonon emission and Auger recombination,31,46 i.e., carrier–phonon coupling. The slow process involves the thermal equilibration within phonon system and the heat conducting induced by the carbon film substrate.47 We quantified the atomic displacements at equilibrium under the Debye-Waller description, −ln⁡(I/I0)=Δ⟨u2⟩ΔT×Q2-\ln(I/I_0) = \Delta\langle u^2\rangle_{\Delta T} \times Q^2,[16] as depicted in Figure 3c. The high-order peaks of (013) and (112) deviate noticeably from the linear dependence on squared scattering vector, Q2Q^2, suggesting local structural distortions besides the global thermal fluctuations.[13, 18] [40] Such deviations are indicative of the expected behavior of small polarons driven by the hole trapping through surface defects, see Supplementary Discussion 4 in Supporting Information.

Besides the intensity changes, the interplanar spacing changes also encode the information of local distortion. We employed the averaged intensity change of (013) and (112) peaks, which show identical evolutions (see Figure S5 in Supporting Information), as the assessment of global lattice thermalization.48 As the thermalization process shows a biexponential rise with characteristic times of 16.7 ps and 92 ps, the interplanar spacing of (010) and (013) show distinct evolutions, in which the (013) plane exhibits a rapid expansion in 1.3 ps while the (010) plane exhibits an initial compression in the same duration, see Figure 3d. Both planes then undergo a much slower expansion with characteristic times of 95 and 106 ps, respectively, aligning with the slow thermalization rate. We calculated the ratio of the interplanar spacing change of (013) over the global thermalization, denoted as R(d/int)R(d/\mathrm{int}), to quantify the distortion rate, as depicted in the inset of Figure 3d. Fitting the trace of R(d/int)R(d/\mathrm{int}) yields a rise time of 1.8 ps and a decay time of 8.5 ps, corresponding to the establishment of local distortion and the recovery of distortion towards the global thermalization, respectively (see Supplementary Discussion 5 and 6 in Supporting Information for more discussions). The 1.8 ps rise time closely matches the duration of hole diffusion and trapping associated to the singly charged oxygen vacancies in ZnO-NPs.40 In contrast, the identified hot hole capturing process with τ0PIA1∼200\tau_{0}^{\mathrm{PIA1}}\sim200 fs in our TA measurement, appears apparently faster due to the sensitivity of TA spectroscopy to the carrier cooling and exciton formation, both of which occur on a sub-picosecond timescale. The subsequent stabilization of excitons by local lattice distortion is more prominent when it is observed from the perspective of structural dynamics.40 With further inspecting the time spans of anomalous interplanar spacing change (see Supplementary Discussion 7 in Supporting Information), we are able to conclude that the preceding time of distortion reflects the formation and stabilization of excited-state hole-polarons.[12] [40, 49]

Normalized intensity changes of (013) and (112) peaks

Figure S5. Normalized intensity changes of (013) and (112) peaks, showing identical temporal evolutions.

As schematically illustrated in Figure 3e, the reflection of closely spaced (013) planes is particularly sensitive to the displacement of zinc atoms along the cc-axis, while the (110) plane spacing primarily captures the basal-plane distortions. When holes are captured, the separations between the oxygen vacancy and its nearest four zinc atoms increase resulting in the observed rapid expansion of (013) plane. Due to the large projections of normal vector lying in the aa-bb plane, the correlated (010) and (110) planes experience compression when the four zinc atoms repel each other during hole capturing,[10] as manifested by the spacing decrease of (010) shown in Figure 3d. The reported decrease from 0.87 to 0.79 for the ratio of Young’s moduli in the direction of aa-axis over that in the cc-axis after the hole capturing by oxygen vacancies,50 probably facilitates a broader compressive response in the aa-bb plane, hence making the compressive deformation along the cc-axis less possible.

Further details of atomic motions involved in the formation and relaxation of hole polarons are inspected by the analysis of pair-distribution function (PDF). The calculated differential PDF, ΔG(r,t)\Delta G(r,t), accesses four distinct atomic pairs in a distance range of 10 Å, as depicted in Figure 4a.The peak at 1.9 Å (G1.9A˚G_{1.9\text{Å}}) corresponds to the first (nearest-neighboring) Zn-O pairs, representing the interactions between the oxygen atom and its surrounding zinc atoms. The 3.3 Å peak (G3.3A˚G_{3.3\text{Å}}) corresponds to the first Zn-Zn pairs, representing the separation of zinc atoms impacted directly by the oxygen vacancy. These two peaks reflect the local atomic environment near oxygen vacancies. In contrast, the peaks at 6.0 Å (G6.0A˚G_{6.0\text{Å}}) and 8.7 Å (G8.7A˚G_{8.7\text{Å}}) correspond to the fifth and the seventh neighboring Zn-Zn pairs with large distance, respectively, reflecting the lattice-scale atomic motions, e.g., the global lattice thermalization. The profiles of ΔG(r,t)\Delta G(r,t) at selected delay times (see Figure 4a) and the intensity changes of four peaks (see Figure 4b) reveal that, the peaks with short distance undergo more pronounced intensity reduction than those with large distance, due probably to the local distortions perturbing the uniform lattice heating. The local distortions modify the interatomic separations of low-order atomic pairs, leading to the deviations from Debye-Waller effect during thermalization,[18] consistent with the behavior of high-indexed diffraction peaks shown in Figure 3c.

Analysis of pair distribution functions

Figure 4. Analysis of pair distribution functions. (a) Calculated differential PDFs at selected delay times upon excitation of 2.21 mJ/cm2^{2}. Inset, schematic of atomic shell structure of wurtzite-phase ZnO marked with the resolved atom pairs. (b) Kinetics of ΔG\Delta G of four atom pairs in (a). Solid lines, biexponential fits. (c) Comparison of the position change to the inversed intensity change (thermalization, see text) for the first zinc-zinc pairs, G3.3A˚G_{3.3\mathrm{\AA}}. Solid lines, biexponential fits. Inset, temporal evolution of the ratio of P3.3A˚P_{3.3\mathrm{\AA}} over lattice thermalization. (d) Initial compression of the first zinc-oxygen pair followed by a thermal expansion. Inset, schematic illustration of the outward displacement of zinc atoms due to the charge repulsion induced by the captured hole at oxygen vacancy site.

The temporal change of PDF peak position corroborates the observed evolution of interplanar spacing.

Similar to the kinetics of (013) interplanar spacing (see Figure 3d), the relative distance change of the nearest Zn-Zn pairs (G3.3A˚G_{3.3\mathrm{\AA}} peak position, P3.3A˚P_{3.3\mathrm{\AA}}) shows a rapid expansion in 1.6 ps, preceding the decrease of peak intensity that also reflects the lattice thermalization, as depicted in Figure 4c. The expansion rate significantly surpasses the rate of lattice thermalization, suggesting the formation of oxygen-vacancy polarons, in which the adjacent zinc atoms move outward when an oxygen vacancy captures a hole, resulting in the increase of the nearest Zn-Zn pair separation.[13] We again calculated the ratio of P3.3A˚P_{3.3\text{\AA}} over thermalization, which shows similar evolution to that of (013)(013) plane, as depicted in the inset of Figure 4c, again corroborating the scenario of polaron formation and stabilization.

We also examined the evolution of first Zn-O pairs (G1.9A˚G_{1.9\text{\AA}}), finding an initial reduction of distance similar to the kinetics of (010)(010) interplanar spacing, as depicted in Figure 4d. This reduction can be interpreted as the zinc atoms squeeze towards the adjacent oxygen atoms during they displace away from the hole-captured vacancy. Such a coupling effect enables the spin-forbidden singlet-triplet transitions, as discussed in Supplementary Discussion 2 in Supporting Information, facilitating the radiative transitions of polaronic excitons and resulting in the green luminescence that characterizes ZnO-NPs.51

In summary, we resolved the pathway and time spans of the formation of exciton-polarons in ZnO-NPs, through the observations in both degrees of freedom of charge and lattice. Our observations highlight the occurrence of local distortion around oxygen vacancy and its neighboring zinc tetrahedron in ∼\sim1 ps, followed by the lattice stabilization through the cooperative Zn-O bond rearrangement within several picoseconds. Thus, the structure relaxes into a lower energy state, forming the excited-state polarons, which stabilize into small hole polarons. The stabilized polarons52 can recombine with the electrons in CB via Auger capturing or defect-assisted processes, resulting in the confined exciton states. Furthermore, these long-lived polarons continuously perturb the lattice and serve as a key-mechanism underlying the persistent green luminescence.

Strongly polar metal oxide materials like ZnO, are particularly susceptible to the intrinsic defect formation that generates a substantial population of small polarons,53 laying significant impacts on the material properties in many ways, e.g., enhancing photogenerated carrier mobility,54 modifying exciton binding energies, and inducing polaronic absorption bands and band bending.55 On the other hand, the strong electron–phonon interactions associated with small polarons can facilitate carrier localization and mediate the non-radiative recombination, by providing additional relaxation pathways, thereby suppressing radiative emission56 and quenching visible emission,57 which are not favored by light emitting applications. Our findings bring better understanding of small polaron formation and relaxation dynamics in ZnO-NPs, and possibly in nanomaterials with alike structure, providing essential information for designing strategies to effectively optimize the material’s performance.

Associated Content

Supporting Information

Experimental details of sample preparation, characterization, transient absorption spectroscopy, and ultrafast electron diffraction measurements; Discussions on the origins of visible TA signals, band-edge bleaching dynamics, defect-related absorption, and lattice distortions linked to Auger-induced thermalization and small-polaron formation; Additional figures S1–S7, tables S1–S7. (PDF)

Author Information

Corresponding Author

Author Contributions

W.L. conceived of and supervised the project. F.H. performed the measurements with the supports from Z.Z., Z.W., C. G. and C. H. Y.L synthesized the ZnO nanoparticles under the supervision from Y.Z. F.H. and W.L. analyzed the data with discussions with all authors. F.H. and W.L. wrote the paper with the contributions from all authors.

Acknowledgement

We thank the Analytical and Testing Center in Huazhong University of Science and Technology for the supports. Y.L. and Y.Z. thank the support from the National Natural Science Foundation of China (52452307).

Supporting Information for

*Corresponding Author. Email: [email protected]

Experimental section

Sample preparation

The ZnO nanoparticles were synthesized following a methanolic sol-gel/precipitation method adapted from Spanhel and Anderson.[1] In a typical procedure, zinc acetate dihydrate (Zn(CH3COO)2⋅2H2O\mathrm{Zn(CH_3COO)_2 \cdot2H_2O}) was dissolved in methanol to form solution A (0.03–0.1 M). Separately, potassium hydroxide (KOH) was dissolved in methanol to form a basic solution B (KOH:Zn = 2–3 mol ratio). Solution A was added dropwise into solution B within 10 min under vigorous stirring at room temperature (water-bath at 25–30 °C). The mixture was stirred for 2–3 hours until the reaction completed, and then allowed to settle.

The resulting white precipitate was collected by centrifugation, then washed three times with methanol, and finally redispersed in a 7:3 (v/v) mixture of n-butanol and chloroform. The obtained ZnO nanoparticles typically exhibit diameters below 10 nm, with the size distribution dependent on the Zn:KOH ratio, adding rate, and reaction temperature.

The obtained solution of ZnO nanoparticle was subsequently diluted tenfold with methanol to prepare samples for the experiments of carrier and structure dynamics observation. In transient absorption measurements, the diluted solution was transferred into a 1 mm-thick optical cuvette for the spectroscopic examination. In ultrafast electron diffraction measurements, a dip-coating method was employed, in which a carbon-coated copper grid was vertically withdrawn from the diluted solution at a speed of 0.05 mm/s. This operation allowed the formation of a monolayer of uniformly dispersed ZnO nanoparticles on the surface of carbon surface.

Sample characterization

Figure S1 and S1b show the morphology of the as-prepared ZnO nanoparticles deposited on carbon-coated copper grid, which is imaged by the high-resolution transmission electron microscopy (HRTEM, JEM-ARM 200F, JEOL). An average radius of 6–7 nm is observed. The nanoparticles are uniformly distributed across the substrate surface, showing no significant aggregation or overlapping. The fast Fourier transform (FFT) of the HRTEM image confirms the hexagonal wurtzite crystal structure, indicating a high degree of crystallinity.

Figure S1 shows the steady-state absorption spectrum (Lambda 35, PerkinElmer). The Tauc plot yields an optical bandgap of 3.36 eV, with the Urbach tail clearly visible in the sub-bandgap range.

Figure S1 shows the photoluminescence spectrum (QuantaMaster 8000, HORIBA) excited at 340 nm. The Gaussian fit yields two peaks centered at 2.3 eV and 3.1 eV, denoted as PLVIS\mathrm{PL}_{\mathrm{VIS}} and PLUV\mathrm{PL}_{\mathrm{UV}}, respectively.

Transient absorption spectroscopy

The transient absorption (TA) measurements in transmission geometry were performed under ambient conditions using a Helios spectrometer (Ultrafast Systems), driven by a Ti:sapphire regenerative amplifier (Legend, Coherent Inc.) operating at a repetition rate of 5 kHz and delivering ~35 fs pulses centered at 800 nm. The fundamental output was directed into a TOPAS optical parametric amplifier (Light Conversion Inc.) to generate 320 nm pump pulses. The pump beam was mechanically chopped to 2.5 kHz before exciting the sample, while a white-light continuum spanning the UV–visible range was generated by focusing a small portion of the 800 nm beam into a CaF2_2 crystal, serving as the probe. The diameters of the pump and probe spots on sample were set to ∼200 μm\sim200\,\mu\mathrm{m} and ∼60 μm\sim60\,\mu\mathrm{m}, respectively. The instrument response function of the setup was measured to be ∼100\sim100 fs.

Ultrafast electron diffraction

The ultrafast electron diffraction (UED) measurements were performed under ultrahigh vacuum conditions using a home-built ultrafast electron diffractometer, which temporal resolution was estimated to be less than 1 ps. The samples were excited by 343 nm pulses, which were the third harmonic of the 1028 nm output from femtosecond laser (CARBIDE, Light Conversion), with a repetition rate of 14.3 kHz. The probe electron pulses were accelerated to 30 keV, impinging in a transmission geometry. Diffraction patterns were intensified by a microchannel plate detector (VID240, Photek) then recorded by an sCMOS camera (ORCA flash 4.0, Hamamatsu), with 1.43×1051.43 \times10^{5} pulses accumulated for each frame. The diameters of the pump and probe spots on sample were set to ∼600 μm\sim600\,\mu\mathrm{m} and ∼100 μm\sim100\,\mu\mathrm{m}, respectively. Time-resolved diffraction signals were radially integrated to yield the scattering-vector-dependent intensity profiles for quantitative analyses.

Supplementary Discussions

Origins of different signals in visible TA spectra

In ZnO with low defect density, such as single-crystal films[2] and nanowires,[3] the TA responses in visible range are dominated by the photo-bleaching due to the hole saturation in defect-states. In contrast, the high defect density in single crystals,[4] nanobulks,[5] and nanoparticles,[6] exhibits pronounced photoinduced absorption features, which are originated from the carrier localization and self-trapping within a continuum of defect states. The carrier trapping opens up additional ultrafast relaxation channels, for example, continuum defect band diffusion,[4] trap-assisted hopping recombination,[5] and the formation of singlet/triplet mixed self-trapped excitons.[6]

Origins of the slow decays of band-edge PB

The slow relaxation of free exciton with characteristic time of τ23.4 eV\tau_{2}^{3.4\ \mathrm{eV}}, is assigned to the radiative recombination, consistent with the 20–110 ps of radiative recombination observed in other ZnO nanomaterials.[7, 8] In contrast to the first decay, the amplitude of this component decreases with increasing fluence, as tabulated in Table S4, indicating the competition of carrier relaxation channels through non-radiative Auger and radiative recombination.[9] The measured PLUVPL_{\mathrm{UV}} accounting for only 5.1% of the total emission intensity (see Figure S1) also demonstrates that the radiative recombination is not dominant in our samples under high excitation fluence. Unlike the direct band-gap excitation, the extra carriers that are excited into the conduction band during three-body processes, such as hole Auger capturing[3] and defect-assisted Auger recombination,[10] eventually relax into the deep defect states, at where they recombine with the previously captured holes.[11] This relaxation gives rise to the measured τ33.4 eV\tau_{3}^{3.4\ \mathrm{eV}}, which also serves as the transition time of band-edge PB to PIA2.

Fluence (mJ/cm2^{2})0.120.330.611.231.351.471.842.443.67
ΔA03.4 eV\Delta A_{0}^{3.4\,\mathrm{eV}} (mOD)−1.27−2.65−3.78−6.20−7.19−8.27−8.44−9.94−9.30
τ03.4 eV\tau_{0}^{3.4\,\mathrm{eV}} (ps)0.230.170.140.160.150.150.170.160.11
ΔA13.4 eV\Delta A_{1}^{3.4\,\mathrm{eV}} (%)26.133.841.64952.353.15459.775.8
τ13.4 eV\tau_{1}^{3.4\,\mathrm{eV}} (ps)13.612.210.89.48.47.36.86.96.3
ΔA23.4 eV\Delta A_{2}^{3.4\,\mathrm{eV}} (%)34.225.425.521.125.526.325.925.612.5
τ23.4 eV\tau_{2}^{3.4\,\mathrm{eV}} (ps)58.55755.55452.649.346.936.946.6
ΔA33.4 eV\Delta A_{3}^{3.4\,\mathrm{eV}} (%)49.740.832.929.922.220.62014.611.7
τ33.4 eV\tau_{3}^{3.4\,\mathrm{eV}} (ps)492484475468454445432421434
ΔA1PIA2\Delta A_{1}^{\mathrm{PIA2}} (mOD)000−0.59−0.52−0.51−0.6−1.42−1.04
τ1PIA2\tau_{1}^{\mathrm{PIA2}} (ps)infinfinf700586218477897498477206

Table S4. Fit results of the fluence-dependent kinetic traces of band-edge PB in Figure 2b in the main text.

PIA2 originated from the defect-bound excitons

Photoinduced absorption signals usually correspond to interband transitions,[12, 13] but PIA2 stems likely from the excited-state absorption of defect-bound excitons.[14, 15] The large exciton binding energy of ZnO-NPs stabilizes the localized excitons against the thermal dissociation,[16] leading to a prerequisite of energy for probe photons to dissociate and excite these excitons to the band-edge states. Consequently, only the photons with energy exceeding the sum of the transition from defect levels and the exciton binding energy can stimulate the PIA2 signal.[17, 18] The decay of PIA2 hence reflects the recombination of defect-bound excitons, which possess a long-lived lifetime more than 7 ns.[3, 19] Similar lifetimes have been observed in the excitons trapped in deep-level defects.[2, 20] In our case, the spin-orbit coupling induced by the lattice distortion around oxygen vacancies facilitates the otherwise forbidden triplet–singlet transitions, resulting in the nanosecond luminescence process.[6, 20, 21]

Deviation from Debye-Waller effect indicating the local distortions

When localized charge–lattice interactions induce asymmetric lattice distortions, those high-order diffraction peaks are sensitive to such distortions due to their small interplanar spacing.[22] The observed deviation is valid across various excitation fluences (see Figure S4), indicating that the residual local distortions may persist at thermal equilibrium. When only the first five peaks with low-order are considered, the derived root-mean-square atomic displacements follow the linear fluence-dependence of Debye-Waller effect, as depicted in the inset of Figure 3c in the main text, testifying the global lattice thermalization through coupling to hot carriers.

Dependence of relative intensity changes on squared scattering vector

Figure S4. Dependence of the relative intensity changes at delay time of 1000 ps on squared scattering vector, upon excitation of various fluences. The high-QQ diffraction peaks, i.e., the (013) and (112) peaks, apparently deviate from the linear relationship described by the Debye-Waller effect, indicating the contributions of non-thermal structural change to the measured diffraction intensity.

Correlation of local distortion with the Auger-induced thermalization

The rise and decay times of R(d/int)R(d/\mathrm{int}) gives a sum time of 10.3 ps, representing the total time of the local distortion motions preceding the global lattice thermalization. This characteristic time is close to the rapid thermalization given by the averaged intensity drops of (013) and (112) peaks (see the main text), suggesting that the local distortion is a concomitant of the rapid thermalization. The fluence-dependent measurements reveal that the rapid thermalization process is expedited with increasing fluence while the fit weight is increased, see Figure S6 and Table S7 for details. Such trends are consistent with the Auger heating mechanism,[23] in which the increased density of hot carriers in confined nanoparticles enhance the three-particle interactions, leading to exciton annihilations and the accelerated lattice heating.[24]

Intensity drops of (112) peak upon excitation of various fluences

Figure S6. Intensity drops of (112) peak upon excitation of various fluences. (a) Kinetic traces. Solid lines, biexponential fits (See Table S7 for fit results). The biexponential fit yields two processes, a fast

Fluence (mJ/cm2^2)1.01.61.92.12.2
A1A_1 (%)34.938.239.243.044.2
τ1\tau_1 (ps)31.224.521.119.717.8
A2A_2 (%)65.161.760.857.055.8
τ2\tau_2 (ps)193.5183.4146.7134.5129.4

Table S7. Fit results of the intensity drop kinetics of (112) peak in Figure S6.

Anisotropic lattice responses as the precursor to the formation of small polarons

The distinct lattice motions of local distortion over the global thermalization also exhibit in the interplanar spacing changes, which are calculated using the centroid method, as depicted in Figure S7. During the early 10 ps, the (010) and (110) planes show anomalous contractions while the (002) and (013) planes show rapid expansions. Moreover, the expansion rates of (102) and (112) planes are apparently slower than those of (002) and (013) planes. Such transient structural responses are often associated with the formation of localized excited states, which induce the lattice deformations[25] that can be viewed as an initial step in stabilizing the excited-state small polarons.

Figure S7. Change rates of lattice plane separation calculated by using the centroid method. In the early 10 ps after excitation, rapid expansions are observed in the (002) and (013) planes, coinciding with the extracted differential intensity changes shown in Figure S3. In contrast, the (010) and (110) planes show remarkable contractions in the same period, followed by thermal expansions. Other planes show gradual interplanar expansions over time.

Formation time of the oxygen-vacancy polarons

Time-resolved X-Ray absorption spectrum (TRXAS) measurement and ab-Initio calculations demonstrated that the formation of oxygen-vacancy polarons induced a local lattice distortion characterized by the damped oscillation with period of 300 fs,[11] which was not resolved in our UED results due probably to the limited temporal resolution. Instead, our UED results resolved the subsequent picosecond-scale expansion and relaxation dynamics, which manifested the stabilization of oxygen-vacancy polarons through local lattice distortions.

Supplementary Figures

at fluence of 2.44 mJ/cm2^{2} due to the governing Auger processes. Lower panel: Dependence of the maximum amplitude of PIA1 at delay time of 40 ps on fluence, extracted at 3.1 eV whereby the contributions of both PB band and PIA1 overlap. The saturation of this mixed response occurs at 1.23 mJ/cm2^{2}, the same fluence for the observed amplitude saturation of PIA2 at 3.4 eV (see the inset in Figure 2b in the main text), suggesting a shared origin related to the defect-assisted carrier trapping.

component of a couple tens of picoseconds, which is associated with the Auger thermalization, and a slow component in the range of 130–190 ps, which is associated with the heat transfer towards substrate.

(b) Dependences of the characteristic time and the fit weight of the fast component on excitation fluence, showing a trend consistent with Auger processes.

Supplementary Tables

References

  1. [1]Spanhel, L.; Anderson, M. A. Semiconductor Clusters in the Sol-Gel Process: Quantized Aggregation, Gelation, and Crystal Growth in Concentrated Zinc Oxide Colloids. J. Am. Chem. Soc. 1991, 113, 2826–2833. DOI: 10.1021/ja00008a004.DOI
  2. [2]Foglia, L.; Vempati, S.; Tanda Bonkano, B.; Gierster, L.; Wolf, M.; Sadofev, S.; Stähler, J. Revealing the Competing Contributions of Charge Carriers, Excitons, and Defects to the Non-Equilibrium Optical Properties of ZnO. Struct. Dyn. 2019, 6, 34501. DOI: 10.1063/1.5088767.DOI
  3. [3]Li, M.; Xing, G.; Xing, G.; Wu, B.; Wu, T.; Zhang, X.; Sum, T. C. Origin of Green Emission and Charge Trapping Dynamics in ZnO Nanowires. Phys. Rev. B 2013, 87, 115309. DOI: 10.1103/PhysRevB.87.115309.DOI
  4. [4]Gu, C.; Zhang, H.; Liu, Y.; Yu, J.; Pan, J.; Luo, G.; Shen, Q.; Tang, J.; Hu, J. Time-Domain Observation of Spectral Diffusion in Defective ZnO. ACS Omega 2021, 6, 15442–15447. DOI: 10.1021/acsomega.1c01890.DOI
  5. [5]Cavaleri, J. J.; Skinner, D. E.; Colombo, D. P.; Bowman, R. M. Femtosecond Study of the Size-dependent Charge Carrier Dynamics in ZnO Nanocluster Solutions. J. Chem. Phys. 1995, 103, 5378–5386. DOI: 10.1063/1.470573.DOI
  6. [6]Zhou, R.; Sui, L.; Liu, X.; Liu, K.; Guo, D.; Zhao, W.; Song, S.; Lv, C.; Chen, S.; Jiang, T.; et al. Multiphoton Excited Singlet/Triplet Mixed Self-Trapped Exciton Emission. Nat. Commun. 2023, 14, 1310. DOI: 10.1038/s41467-023-36958-3.DOI
  7. [7]Ou, P.-C.; Liu, W.-R.; Ton, H.-J.; Lin, J.-H.; Hsieh, W.-F. Ultrafast Relaxation and Absorption Saturation at near Exciton Resonance in a Thin ZnO Epilayer. J. Appl. Phys. 2011, 109, 13102. DOI: 10.1063/1.3525993.DOI
  8. [8]Appavoo, K.; Liu, M.; Sfeir, M. Y. Role of Size and Defects in Ultrafast Broadband Emission Dynamics of ZnO Nanostructures. Appl. Phys. Lett. 2014, 104, 133101. DOI: 10.1063/1.4868534.DOI
  9. [9]van Dijken, A.; Meulenkamp, E. A.; Vanmaekelbergh, D.; Meijerink, A. The Kinetics of the Radiative and Nonradiative Processes in Nanocrystalline ZnO Particles upon Photoexcitation. J. Phys. Chem. B 2000, 104, 1715–1723. DOI: 10.1021/jp993327z.DOI
  10. [10]Cohn, A. W.; Schimpf, A. M.; Gunthardt, C. E.; Gamelin, D. R. Size-Dependent Trap-Assisted Auger Recombination in Semiconductor Nanocrystals. Nano Lett. 2013, 13, 1810–1815. DOI: 10.1021/nl400503s.DOI
  11. [11]Milne, C. J.; Nagornova, N.; Pope, T.; Chen, H.-Y.; Rossi, T.; Szlachetko, J.; Gawelda, W.; Britz, A.; Van Driel, T. B.; Sala, L.; et al. Disentangling the Evolution of Electrons and Holes in Photoexcited ZnO Nanoparticles. Struct. Dyn. 2023, 10, 64501. DOI: 10.1063/4.0000204.DOI
  12. [12]Bauer, C.; Boschloo, G.; Mukhtar, E.; Hagfeldt, A. Ultrafast Relaxation Dynamics of Charge Carriers Relaxation in ZnO Nanocrystalline Thin Films. Chem. Phys. Lett. 2004, 387, 176–181. DOI: 10.1016/j.cplett.2004.01.106.DOI
  13. [13]Liu, Q.; Chen, H.; Wang, Z.; Weng, Y. Observation of the Polaron Excited State in a Single-Crystal ZnO. J. Phys. Chem. C 2021, 125, 10274–10283. DOI: 10.1021/acs.jpcc.1c01787.DOI
  14. [14]Wagner, M. R.; Callsen, G.; Reparaz, J. S.; Schulze, J.-H.; Kirste, R.; Cobet, M.; Ostapenko, I. A.; Rodt, S.; Nenstiel, C.; Kaiser, M.; et al. Bound Excitons in ZnO: Structural Defect Complexes versus Shallow Impurity Centers. Phys. Rev. B 2011, 84, 35313. DOI: 10.1103/PhysRevB.84.035313.DOI
  15. [15]Meyer, B. K.; Sann, J.; Eisermann, S.; Lautenschlaeger, S.; Wagner, M. R.; Kaiser, M.; Callsen, G.; Reparaz, J. S.; Hoffmann, A. Excited State Properties of Donor Bound Excitons in ZnO. Phys. Rev. B 2010, 82, 115207. DOI: 10.1103/PhysRevB.82.115207.DOI
  16. [16]Zeng, H.; Duan, G.; Li, Y.; Yang, S.; Xu, X.; Cai, W. Blue Luminescence of ZnO Nanoparticles Based on Non-Equilibrium Processes: Defect Origins and Emission Controls. Adv. Funct. Mater. 2010, 20, 561–572. DOI: 10.1002/adfm.200901884.DOI
  17. [17]Jollivet, A.; Quach, P.; Tchernycheva, M.; Ferreira, R.; Di Russo, E.; Rigutti, L.; Vinter, B.; le Biavan, N.; Lefebvre, D.; Hugues, M.; et al. Exciton Ionization Induced by Intersubband Absorption in Nonpolar ZnO-ZnMgO Quantum Wells at Room Temperature. Phys. Rev. B 2022, 105, 195143. DOI: 10.1103/PhysRevB.105.195143.DOI
  18. [18]de Boer, W. D. A. M.; Timmerman, D.; Gregorkiewicz, T.; Zhang, H.; Buma, W. J.; Poddubny, A. N.; Prokofiev, A. A.; Yassievich, I. N. Self-Trapped Exciton State in Si Nanocrystals Revealed by Induced Absorption. Phys. Rev. B 2012, 85, 161409. DOI: 10.1103/PhysRevB.85.161409.DOI
  19. [19]Meyer, B. K.; Alves, H.; Hofmann, D. M.; Kriegseis, W.; Forster, D.; Bertram, F.; Christen, J.; Hoffmann, A.; Straßburg, M.; Dworzak, M.; et al. Bound Exciton and Donor–Acceptor Pair Recombinations in ZnO. Phys. Status Solidi B 2004, 241, 231–260. DOI: 10.1002/pssb.200301962.
  20. [20]Janßen, N.; Whitaker, K. M.; Gamelin, D. R.; Bratschitsch, R. Ultrafast Spin Dynamics in Colloidal ZnO Quantum Dots. Nano Lett. 2008, 8, 1991–1994. DOI: 10.1021/nl801057q.DOI
  21. [21]Angelis, F. D.; Armelao, L. Optical Properties of ZnO Nanostructures: A Hybrid DFT/TDDFT Investigation. PCCP 2010, 13, 467–475. DOI: 10.1039/C0CP01234C.DOI
  22. [22]Guzelturk, B.; Cotts, B. L.; Jasrasaria, D.; Philbin, J. P.; Hanifi, D. A.; Koscher, B. A.; Balan, A. D.; Curling, E.; Zajac, M.; Park, S.; et al. Dynamic Lattice Distortions Driven by Surface Trapping in Semiconductor Nanocrystals. Nat. Commun. 2021, 12, 1860. DOI: 10.1038/s41467-021-222116-0.DOI
  23. [23]Achermann, M.; Bartko, A. P.; Hollingsworth, J. A.; Klimov, V. I. The Effect of Auger Heating on Intraband Carrier Relaxation in Semiconductor Quantumrods. Nature Phys 2006, 2, 557–561. DOI: 10.1038/nphys363.DOI
  24. [24]Yue, L.; Li, J.; Qi, Y.; Chen, J.; Wang, X.; Cao, J. Auger Recombination and Carrier–Lattice Thermalization in Semiconductor Quantum Dots under Intense Excitation. Nano Lett. 2023, 23, 2578–2585. DOI: 10.1021/acs.nanolett.2c04804.DOI
  25. [25]Seiler, H.; Zahn, D.; Taylor, V. C. A.; Bodnarchuk, M. I.; Windsor, Y. W.; Kovalenko, M. V.; Ernstorfer, R. Direct Observation of Ultrafast Lattice Distortions during Exciton–Polaron Formation in Lead Halide Perovskite Nanocrystals. ACS Nano 2023, 17, 1979–1988. DOI: 10.1021/acsnano.2c06727.DOI
  26. [26]Zhang, H.; Li, W.; Essman, J.; Quarti, C.; Metcalf, I.; Chiang, W.-Y.; Sidhik, S.; Hou, J.; Fehr, A.; Attar, A.; et al. Ultrafast Relaxation of Lattice Distortion in Two-Dimensional Perovskites. Nat. Phys. 2023, 19, 545–550. DOI: 10.1038/s41567-022-01903-6.DOI

Paper details

vlnc:W_e5a942kv
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents