Single Trap States in Single CdSe Nanoplatelets

Abstract

Trap states can strongly affect semiconductor nanocrystals, by quenching, delaying, and spectrally shifting the photoluminescence (PL). Trap states have proven elusive and difficult to study in detail at the ensemble level, let alone in the single-trap regime. CdSe nanoplatelets (NPLs) exhibit significant fractions of long-lived "delayed emission" and near-infrared "trap emission". We use these two spectroscopic handles to study trap states at the ensemble and the single-particle level. We find that reversible hole trapping leads to both delayed and trap PL, involving the same trap states. At the single-particle level, reversible trapping induces exponential delayed PL and trap PL, with lifetimes ranging from 40 to 1300 ns. In contrast with exciton PL, single-trap PL is broad and shows spectral diffusion and strictly single-photon emission. Our results highlight the large inhomogeneity of trap states, even at the single-particle level.

Keywords: CdSe nanoplatelets; single-particle spectroscopy; spectral diffusion; transient absorption; trap states.

Figure 1

(a) PL spectra as a function of delay time fitted to a combination of two Gaussians. (Black) Steady-state PL spectrum, measured using a standard spectrometer, (cyan–red) time-resolved PL spectra, showing a relative increase in trap PL intensity, and red shift of trap PL with time, measured using a custom-built setup (Supporting Information). The time-resolved spectra are broadened due to the limited resolution of our custom setup. (b) Integrated intensities of the (blue) exciton and (red) trap PL peaks, as a function of delay time; the inset shows a zoom-in of the first 100 ns. (c) The same data as in (b), but displayed on double-logarithmic axes. Solid lines in (c) are fits to power-law decay I ∝ t^–α, with I the PL intensity, t the delay time, and α the power-law slope, fitted on the interval [100, 3000] ns. (d) Three-state model explaining delayed and trap emission in CdSe nanoplatelets: after photoexcitation, one charge carrier can localize at a (surface) trap. Direct radiative recombination of the trapped charge carrier is possible, yielding trap PL, or the charge carrier may detrap. The black dashed lines indicate nonradiative transitions. (e) (Solid lines) Transient absorption spectra of an ensemble of CdSe nanoplatelets dispersed in hexane, normalized to the HH bleach feature, recorded at delay times of (black) 0.0, (blue) 11.3, (green) 119, and (red) 215 ns; (pink area) steady-state absorption spectrum of the same sample. The HH and LH absorption features are indicated. (f) HH bleach −ΔA as a function of delay time, for CdSe nanoplatelets dispersed in hexane with (black) no additives and (green) added p-BQ. (g) Time dependence of the (blue) exciton and (red) trap PL intensity, (black solid symbols) the HH bleach −ΔA, and (black open circles) the time derivative of the HH bleach, dΔA/dt, of CdSe NPLs dispersed in hexane.

Figure 2

Results of measurements on a single CdSe NPL from the untreated “A” microscopy sample: (a) PL spectra and (b) intensity as a function of time; (c) PL spectra and (d) PL decay curves, constructed from the four colored regions indicated in panels (a) and (b). (e) The same data as depicted in panel (d), but displayed on shorter time scales. Lines are fits to (c) two Gaussian peaks and (d, e) biexponential decay. The fitted (c) full-width at half-maxima and (d) long-component and (e) short-component PL lifetimes τ are indicated. The PL intensities plotted in panel (b) are the sum of the exciton and trap peak integrals, which were determined by fitting each emission spectrum to two Gaussian peaks.

Figure 3

(a–c) Three (de)trapping regimes that CdSe NPLs may experience: (a) no trapping, (b) fast trapping–detrapping cycling, and (c) irreversible trapping. (d) Calculated lifetime of the slow delay component τ_slow, normalized to the trap recombination lifetime τ_T, and (e) fractional trap PL I_T/I_tot, as a function of trapping probability P_t and detrapping probability P_dt. The dotted lines in (d) and (e) are contours of the calculated average number of trapping events N prior to recombination. (f) Configuration–coordinate diagram, illustrating exciton and trap emission. (g) Zoom-in of the diagram displayed in panel (f), indicating the energy barriers for trapping ΔE_trap^# and detrapping ΔE_detrap^#. As the single-NPL trap-PL fwhm appears to be independent of the peak energy (Figure S3; Extended Data), the difference in the equilibrium Q among trap states is negligible; that is, the trap states are spaced vertically, not horizontally, in the coordinate diagram.

Figure 4

(a) Cross-correlation function g⁽²⁾ of a CdSe NPL emitting predominantly exciton PL, for gate times of (blue, top) 0 (i.e., no time-gating), (cyan, middle) 1.00, and (green, bottom) 4.00 ns. (b) Cross-correlation function g⁽²⁾ of a CdSe NPL emitting predominantly trap PL. (c) PL spectra and (d) PL decay curves of three different NPLs. Symbols are data points, solid lines are fits to (c) two Gaussian peaks and (d) biexponential decay. The fitted (c) fwhm and (d) lifetimes are indicated. The PL decay curves were collected without spectral selectivity, so they include dominant trap PL and any weak exciton PL. (e) Two-dimensional histogram showing the correlation of fitted peak energy μ and average photon arrival time ⟨τ⟩ of a trap-emitting NPL that exhibited spectral diffusion. (f, g) Cartoons illustrating the (lack of) electron–hole pair polarizability. The cartoons depict a cross-cut of a NPL, with the x,y axes corresponding to the lateral directions, and F the external electric field. Hypothetical hole localization (f) inside the NPL or (g) at the NPL surface results in qualitatively different polarizabilities. The NPL of panel (a) belonged to the untreated “A” microscopy sample; the NPLs of panels (b)–(e) to the butylamine-treated “B” sample.