1. INTRODUCTION

As members of the colloidal semiconductor nanocrystals (NCs) family, cesium lead halide perovskites in the form of CsPbX3 (X represents Cl, Br, I) offer outstanding optical properties such as high photoluminescence (PL) quantum yield and defect tolerance ^{[13" target="_self" style="display: inline;">–3]}. Their wavelength tunability can cover the full visible spectrum through either composition or size manipulation. CsPbX3 NCs are potential candidates for a variety of optoelectronic and photonic applications such as chiroptics, light-emitting diodes, photodetectors, and nonlinear optical devices ^{[4–9" target="_self" style="display: inline;">–9]}.

Doping is a well-known strategy that may lead to novel functionalities and significant improvements in device efficiency. To date, various ions that have been used as dopants for CsPbX3 NCs have been reported ^{[1012" target="_self" style="display: inline;">–12]}. Among those, the doping of CsPbX3 NCs with Mn2+ has recently attracted much attention because the presence of Mn2+ can introduce novel optical and magnetic features to the NCs ^[13,14]. For instance, the energy transfer from excitons to Mn2+ results in a strong emission from the field transition of Mn2+ at approximately 600 nm with a lifetime of ∼1ms ^[13,14]. The long lifetime of the PL emission can efficiently enhance the temporal discrimination of signals from the autofluorescence background ^[15,16]. Additionally, Mn-doped CsPbX3 NCs often exhibit a superior quantum yield of their exciton emission than their undoped counterparts, thus making them attractive for many applications. Additionally, Mn-doped perovskite NCs have been found to exhibit strong two-photon absorption (2PA) cross sections on the order of ∼105GM ^[17]. For weakly confined semiconductor NCs, whose sizes are much smaller than the exciton Bohr radius, the multiphoton absorption (MPA) cross sections scale linearly with their volume and show more efficient MPA in NCs with larger sizes ^[18]. When the size of the perovskite NCs is much smaller than the exciton Bohr radius, the increased quantum confinement effect may prevail over the decreased density of state, which may further enhance their MPA properties ^[19]. However, to date, studies of the quantum confinement-induced enhancements of MPA properties in Mn-doped perovskite NCs have not been reported.

In this study, the influences of the geometry of Mn-doped CsPbCl3 NCs on their MPA properties were investigated. By performing femtosecond transient absorption (fs-TA) spectra measurements, we found that more efficient energy transfer from the excitons to Mn2+ can be induced in the Mn-doped CsPbCl3 two-dimensional nanoplatelets (2D NPs) compared with those in their cubic counterparts. More importantly, the 2D geometry of the NCs can greatly amplify their MPA under excitation from the visible to near-infrared wavelengths. Our experimental results indicate that Mn-doped perovskite 2D NPs are better alternative materials for probes of multiphoton fluorescence lifetime imaging (FLIM) systems in light of their high MPA, long PL lifetime, and superior anti-Stokes shift.

2. EXPERIMENT

2.1 A. Sample Synthesis

The experimental conditions used for the synthesis of the Mn-doped CsPbCl3 2D NPs and cubic NCs in this study were similar to those reported in the previous literature ^[20]. During the fabrication of Mn-doped CsPbCl3 cubic NCs, a mixture of oleylamine and HCl was first introduced in a reaction flask along with PbCl2, MnCl2, and other reagents. When the temperature reached 180°C, a Cs precursor was quickly injected. The heating mantle was removed, and the reaction mixture was cooled to room temperature. For the fabrication of the Mn-doped CsPbCl3 2D NPs, the experimental conditions were similar to those mentioned above. The main difference was that the Cs-to-Pb precursor ratio was 1:5 for the cubic NCs and 1:2 for the 2D NPs. The 2D NPs and cubic NCs were precipitated with acetone and redispersed in hexane. The obtained solutions were then diluted with hexane to the desired optical density for use in spectroscopic experiments. The doping concentrations of Mn2+ in both the 2D NPs and cubic NCs were equal and approximately 0.2%.

2.2 B. Sample Characterization

The PL quantum yields of the perovskite 2D NPs and the cubic NCs were determined using quinine sulfate monohydrate (η350nm=0.58) as a standard. The size and shape of the NCs were recorded using a high-resolution transmission electron microscope (HR-TEM) (JEOL, JEM-2100F) at 200 kV. The linear absorption spectra of the samples were measured using a ultraviolet–visible (UV–vis) spectrometer (Lambda 950, PerkinElmer, Inc.), while their PL spectra were measured using a fluorescence spectrophotometer (SENS-9000, Zolix).

2.3 C. Femtosecond TA Measurements

The samples were pumped at 350 nm and probed with a white-light continuum. The probe pulses (350–750 nm) were generated by focusing a small portion (∼5μJ) of fundamental 800 nm laser pulses into a thin CaF2 plate. The linear polarization of the pump pulse was adjusted to be perpendicular to that of the probe pulse using a polarizer and a half-wave plate. Cross-polarization helps to eliminate any early contribution from coherent artifacts. The pump-induced changes in the transmission (ΔT/T) of the probe beam were monitored using a monochromator/photomultiplier configuration with lock-in detection. The pump beam was chopped at 100 Hz, which was used as the reference frequency for the lock-in amplifier.

2.4 D. Z-Scan Measurements

The MPA cross sections of the 2D NPs and the cubic NCs were measured via Z-scan measurements ^[21], where a beam splitter was used to divide the incident laser beam into two parts. The first part served as the reference, while the second part functioned as the signal beam. The latter was focused using a circular lens with a focal length of 30 cm into a 1 mm thick quartz cuvette filled with the sample solution. The beam transmitted through the samples that could be influenced by the MPA was detected by a signal power detector. From the open-aperture Z-scan curves, we can obtain 2PA and three-photon absorption (3PA) coefficients (α2 and α3) from the following equations: ΔT(2PA)=11+α2(1−e−α0Lα0){I0/[1+(z/z0)2]},(1)ΔT(3PA)=1{1+2α3(1−e−2α0L2α0){I0/[1+(z/z0)2]}2}0.5,(2)where ΔT(2PA) and ΔT(3PA) are the normalized transmittance of the Z-scan measurements, α0 is the linear absorption of the samples, L is the thickness of the samples, I0 is the peak intensity, z is the sample position, z0=πω02/λ is the Rayleigh range, ω0 is the beam waist at the focal point, and λ is the laser wavelength ^[22]. The relevant 2PA cross section (σ2, in units of cm4·s) can be calculated from the equation σ2=α2hνNAd0×10−3,(3)where hν is the photon energy of the input light beam, NA is Avogadro’s number, d0 is the molar concentration of the samples (in units of mol/L) ^[23]. Accordingly, the 3PA cross section (σ3, in units of cm6·s2) can be calculated from the equation ^[24] σ3=α3(hν)2NAd0×10−3.(4)

2.5 E. Multiphoton-Excited PL Measurements

For the multiphoton excited PL emission measurements, fs pulses with wavelengths ranging from 620 to 700 nm and from 1300 to 1500 nm were used as the excitation sources. The signals were dispersed using a 750 mm monochromator combined with suitable filters and detected using a photomultiplier via the standard lock-in amplifier technique. Two continuously variable neutral density filters were employed to control the incident energy of the laser pulses. A short-pass filter cut at 680 nm was placed in front of the spectrometer to remove the scattered light at the excitation laser frequencies.

3. RESULTS AND DISCUSSION

The HR-TEM images of 2D NPs and cubic NCs are shown in Figs. 1(a) and 1(b). As shown in Figs. 1(c)–1(e), the 2D NPs had non-identical lateral sizes of ∼10.1 and ∼9.6nm, while the majority of 2D NPs had a thickness of ∼3.0nm, which corresponds to five unit cells ^[25]. The as-synthesized cubic NCs had a cubic geometry with an average edge length of ∼7.8nm, as shown in Fig. 1(f). The calculated volumes were 291nm3 for the 2D NPs and 475nm3 for the cubic NCs. Figure 2(a) shows the UV-vis and PL spectra for both the 2D NPs and the cubic NCs excited at 350 nm. We found that the 2D NPs showed more obvious features of quantum confinement effects, including stronger excitonic absorption peaks and smaller Stokes shifts compared with the cubic NCs. More specifically, the first exciton absorption peaks of the 2D NPs and cubic NCs were located at 381 nm and 385 nm, respectively, indicating band-gap shifts from 3.25 to 3.22 eV, whereas their emission peaks redshifted to 395 nm and 402 nm. The observed blueshift in the UV-vis and PL spectra of the 2D NPs can be explained by their stronger quantum confinement effects compared with the cubic NCs ^[26]. Along with their excitonic PL, both the Mn-doped 2D NPs and the cubic NCs showed broad PL characteristics with peaks at ∼592nm because of the Mn2+ d-d transition. After absorbing the light-forming excitons, the host transfers its energy to the T14 state of the Mn2+ d-electrons, followed by an emission from the T14 to A16 state. Additionally, the absolute quantum yields of the PL emission at ∼592nm for the 2D NPs and the cubic NCs were determined to be 2% and 16%, respectively. Compared with cubic NCs, 2D NPs exhibit much lower PL quantum yield. The 2D NPs are strongly susceptible to surface defects due to a large surface-to-volume ratio, rendering their PL quantum yield typically quite low ^[27]. As shown in Fig. 2(b), their excited-state decay curves suggest that the average PL lifetime values were ∼1.8 and ∼1.6ms, respectively. The long lifetime values of the Mn2+ d-d emissions are due to the nature of the spin-forbidden T14 to A16 transitions. Compared with their cubic counterpart, the 2D NPs exhibited even longer lifetimes due to their stronger quantum confinement effects.

Fig. 1. HR-TEM micrographs depicting the atomic resolution of both (a) flat-lying and stacked Mn-doped CsPbCl3 2D NPs and (b) cubic NCs. Subfigures (c), (d), and (e) show the distribution of the lengths, widths, and thicknesses of the 2D NPs, respectively, while (f) shows the edge size distribution of the cubic NCs.

下载图片 查看所有图片

Fig. 2. (a) UV-vis absorption and PL spectra of the 2D NPs and the cubic NCs. (b) Time-dependent PL intensity of Mn2+ at 592 nm measured from the 2D NPs and the cubic NCs. The insets show their colloidal solutions under UV illumination (λ=365nm).

下载图片 查看所有图片

The fs-TA spectra of the 2D NPs and the cubic NCs pumped at 350 nm are shown in Fig. 3(a). Both the 2D NPs and the cubic NCs exhibited three distinctive spectral features, including ground-state bleaching (GSB) bands centered at 383 nm and 387 nm, respectively. The energies of these GSB bands are in good agreement with the lowest energy excitonic bands, and they can thus be attributed to state-filling-induced bleaching ^[28]. Furthermore, there are two bands of photo-induced absorption (PIA) for both the 2D NPs and cubic NCs. The short-lived PIA band at the energies just below the band gap is mostly likely derived from band-gap renormalization. The PIA band above the band gap likely corresponds to increased scattering from the photorefractive effect of the suspension.

Fig. 3. (a) fs-TA data for the Mn-doped 2D NPs and cubic NCs pumped at 350 nm. (b) Recovery curves of the bleaching process at 383 nm for the 2D NPs and that at 387 nm for the cubic NCs. (c), (d) The excitation-intensity-dependent GSB signal amplitude at a time delay of 1 ns is shown for the 2D NPs and the cubic NCs. The curves are the best fit lines calculated based on Eq. (5).

下载图片 查看所有图片

To further explore the photophysical properties of the 2D NPs and cubic NCs, we measured their exciton–Mn2+ energy transfer time (τET), which can be used to estimate the relative strength of their exciton–dopant coupling ^[29]. Because doping with Mn2+ in the host introduces an additional pathway for the depletion of excitons, the differences in exciton relaxation dynamics can reveal the exciton–dopant coupling information for the 2D NPs and cubic NCs. The τET values were then measured by monitoring the dynamics of the exciton relaxation. Figure 3(b) shows these relevant dynamic processes near the exciton absorption peaks in the 2D NPs and cubic NCs. The recovery curves of the bleaching process for the 2D NPs and cubic NCs show three distinct components with very different time scales. The time constants (relative amplitudes) obtained via the multiexponential fitting of the fs-TA data were τ1=5.5ps (A1=0.58), τ2=100ps (A2=0.21), and 1.93 ns (A3=0.21) for the 2D NPs and τ1=6.2ps (A1=0.64), τ2=140ps (A2=0.18), and 6.7 ns (A2=0.18) for the cubic NCs. The fast recovery component τ1 is attributed to the rapid electron trapping at the defects of the Cl vacancies. The second component (τ2) represents the exciton–Mn2+ energy transfer, and its relative amplitude can be interpreted as the energy-transfer-related population. The nanosecond-scale recovery component (τ3) is related to the exciton relaxation in the subpopulation of the CsPbCl3 NCs, while its relative amplitude is associated with the subpopulation that does not undergo energy transfer. Comparing the exciton–Mn2+ energy transfer times between the 2D NPs (∼100ps) and the cubic NCs (∼140ps) suggests that the exciton–Mn2+ exchange coupling in the former is 1.4 times stronger than that in the latter due to its enhanced quantum confinement properties.

Then, we estimated the linear absorption cross sections (σlin) of the 2D NPs and the cubic NCs from the excitation-intensity-dependent one-photon-induced GSB signals via fs-TA spectroscopy. After fast Auger recombination within the initial hundreds of picoseconds, the samples contained only a single exciton in the subsequent time period, which is demonstrated by the parallel decay lines for all the excitation intensities after a long delay (>0.5ns) ^[17]. The amplitude of the GSB signal under different excitation intensities varied based on the following equation: −A(I/I0)=−Amax[1−e−(I/I0)·σlin·I0],(5)where A(I/I0) denotes the GSB signal amplitude of samples after a long time delay as a function of excitation intensity and I0 is the minimum excitation intensity used in the fs-TA experiment ^[18]. As shown in Figs. 3(c) and 3(d), the excitation-intensity-dependent GSB signal amplitude with a delay time of 1 ns could be well fitted with Eq. (5), from which the values of σlin were extracted and determined to be ∼1.84×10−14cm2 for the 2D NPs and ∼0.7×10−14cm2 for the cubic NCs. Correspondingly, their molar distinction coefficients were determined to be ∼4.78×106 and ∼1.82×106L·cm−1·mol−1 at 350 nm. Additionally, compared with cubic NCs, the linear absorption cross section of the 2D NPs is considerably larger. This is the result of a higher optical field penetration in the x and y directions where the local field factor in these two directions approaches unity. For cubic NCs, the local field factor is always much smaller than 1, resulting in their much smaller linear absorption cross section ^[30,31].

Based on the UV-vis absorption spectra of the 2D NPs and the cubic NCs, we expected that under excitation with fs pulses in the range of 450–800 nm, they might exhibit 2PA behavior. Figure 4(a) shows the PL spectra of the 2D NPs and the cubic NCs under excitation with fs pulses at 680 nm. We found that the emission spectra of Mn2+ were indistinguishable from those obtained under excitation at 350 nm, indicating that the same emissive states were involved under excitation of both 350 nm and 680 nm. Although both the Mn-doped 2D NPs and NCs exhibited strong exciton emission under one-photon excitation, their emission was completely quenched under two-photon excitation, and only the emission from Mn2+ was observed. This implies that there are large differences between the cross sections of the CsPbCl3 host and the Mn2+ under one- and two-photon excitation ^[32]. To explore whether the observed PL of the 2D NPs and the cubic NCs originated from 2PA processes, their PL intensities were measured under different excitation intensities. As shown in Fig. 4(b), the clear square relationship between the PL and excitation intensities suggests that 2PA mechanisms are truly responsible for the PL of the 2D NPs and cubic NCs ^[33]. Additionally, we easily observed two-photon excited PL with the naked eye in solutions of the 2D NPs and cubic NCs under excitation with a power intensity of several gigawatts per centimeter squared (GW/cm2), as shown in the inset of Fig. 4(a). Z-scan measurements were then carried out to quantitatively determine the relevant 2PA spectra of the 2D NPs and the cubic NCs, as presented in Fig. 4(c). Excitation wavelengths in the range of 620–700 nm were chosen to gain insight into the contributions of the electronic confinement to the 2PA cross sections from wavelengths that were resonant to one-photon excitonic transitions up to the 2D continuum. For comparison, the corresponding linear absorption spectra are plotted over their one-photon energies. For the cubic NCs, a weaker contribution of the first excitonic resonance to the 2PA spectrum due to weak selection rules in the continuum was found in comparison to the energies of the quasi-continuum, which have been reported for many other semiconductor NCs ^[19]. In contrast, the 2PA signals for the 2D NPs showed a dominant peak at 381 nm. The similarity between the peak positions and full width at half-maximum of the linear absorption and the 2PA spectrum clearly confirms the excitonic enhancement of the 2PA process, which should also be induced by strong quantum confinement effects ^[34]. Analyzing the 2PA spectra of the 2D NPs and the cubic NCs shown in Fig. 4(c) shows that there is a shape dependence for the 2D NPs and the cubic NCs. The 2D NPs showed a maximum 2PA cross section of 3.6×105GM at 620 nm, while that for the cubic NCs was 0.87×105GM at 620 nm. Compared with cubic NCs, the 2D NPs displayed a much larger 2PA over the entire measured wavelength range, which is attributed to the confinement-related increase of the transition dipole moments and band mixing ^{[3537" target="_self" style="display: inline;">–37]}. Since the thickness (3 nm) of the 2D NPs is much smaller than the Bohr excitonic diameter of CsPbCl3 (5 nm), strong quantum confinement can be induced in the 2D NPs, which can result in the quasi-2D colloidal quantum-well-type structure experienced by the charge carriers. In turn, this significantly enhances the 2PA of the 2D NPs.

Fig. 4. (a) PL spectra for the 2D NPs and cubic NCs excited by fs pulses at 680 nm. The inset shows the PL emissions of the 2D NPs and cubic NCs. (b) PL intensity versus excitation intensity with log–log plot slopes of 1.95 and 2.02, respectively. (c) 2PA and linear absorption spectra of the 2D NPs and cubic NCs plotted over the two- and one-photon energies. (d) Photo-induced energy transfer process from the host (CsPbCl3) to Mn2+ in the Mn-doped 2D NPs and cubic NCs under two-photon excitation.

下载图片 查看所有图片

Having demonstrated the strong 2PA from the 2D NPs and cubic NCs in the visible light band, we move forward to investigate their MPA in the second biological window (1000–1700 nm), which could offer even greater advantages ^[38,39]. The PL properties of the 2D NPs and the cubic NCs were measured using fs pulses at 1300 nm. As expected, the obtained PL spectra of Mn2+ were similar to those obtained under the one-photon excitation, as shown in Fig. 5(a). Moreover, we also plotted the relationships between up-converted PL intensity and excitation intensity in Fig. 5(b). Unexpectedly, the PL intensities of both the 2D NPs and the cubic NCs had a cubic dependence on the excitation intensity, confirming that the observed PL was induced by 3PA ^[33]. Although the host CsPbCl3 may exhibit four-photon absorption at 1300 nm, the observed PL at 592 nm was not induced by four-photon excited energy transfer from the host to the Mn2+ and subsequent T14−A16 Mn d-electron emission; otherwise, a fourth-power dependence would have been observed. Additionally, we found that the four-photon excited PL emission of the 2D NPs and cubic NCs made of undoped CsPbCl3 was difficult to measure under the same excitation intensity, as shown in the inset of Fig. 5(c). Therefore, the underlying PL mechanism for the Mn2+ emission at 592 nm was attributed to direct 3PA transitions from the A16 state to the T14 state of Mn2+, as shown in Fig. 5(d). The 3PA spectra of the 2D NPs and the cubic NCs were then measured using the Z-scan measurements. While tuning the excitation wavelength from 1500 to 1300 nm, the 3PA cross sections of the 2D NPs and the cubic NCs continuously increased, as shown in Fig. 5(c). The maximum 3PA cross sections of the Mn-doped NPs and NCs were 6.54×10−76 and 1.49×10−76cm6·s2·photon−2 at 1300 nm, respectively. Compared with the cubic NCs, it can be seen that the maximum 3PA cross section of the 2D NPs was again larger due to their stronger quantum confinement effects. To further confirm the possibility of such a transition, we measured the linear excitation spectra of the 2D NPs and cubic NCs detected at 592 nm. As shown in Fig. 5(d), a very weak PL emission at 592 nm was obtained under excitation wavelengths ranging from 420 to 500 nm, which were likely induced by the direct one-photon transitions of the A16 state to the T14 state but not by the excitonic energy transferred from the host to the Mn2+. The 3PA spectra of the 2D NPs and the cubic NCs resembled their linear excitation spectra, providing supporting evidence for the direct 3PA transition from the A16 state to the T14 state.

Fig. 5. (a) PL spectra for the 2D NPs and cubic NCs excited by fs pulses at 1300 nm. The inset shows the PL emissions of the Mn-doped 2D NPs and cubic NCs. (b) PL intensity of the 2D NPs and cubic NCs versus the optical intensity with log–log plot slopes of 2.98 and 3.02, respectively. (c) 3PA spectra of the 2D NPs and cubic NCs. The inset shows solutions of the 2D NPs and cubic NCs in a pure host (CsPbCl3) under the same excitation intensity used for the inset of (a). (d) One-photon excitation spectra of the 2D NPs and cubic NCs detected at 592 nm. The inset shows the direct 3PA process from the A16 state to the T14 state in the Mn2+ of the 2D NPs and cubic NCs.

下载图片 查看所有图片

In addition to the absolute MPA cross section of the NCs, we also calculated the maximum volume-normalized (VN) MPA cross sections, which yield more intrinsic values, to study the influences of geometry on the MPA ^[19,37]. As shown in Fig. 6, the maximum VN 2PA cross sections of the 2D NPs and cubic NCs are determined to be 1237GM/nm3 and 183GM/nm3, respectively, while the relevant values of the 3PA cross sections are 2.24×10−78cm6·s2·photon−2/nm3 and 3.1×10−79cm6·s2·photon−2/nm3, respectively. Additionally, the relevant values for the Mn-doped ZnS spherical NCs were also added for comparison purposes. Compared with their cubic counterparts, the 2D geometry of the Mn-doped CsPbCl3 NPs led to VN 2PA and 3PA values that were 6.8 and 7.2 times larger, respectively, due to the stronger confinement-related increases of the transition dipole moments and band mixing. Additionally, the VN 2PA and 3PA cross sections of the Mn-doped CsPbCl3 2D NPs and cubic NCs were 1–3 orders of magnitude larger than those of the Mn-doped ZnS spherical NCs ^[40,41], which can be ascribed to the stronger quantum confinement effects of the CsPbCl3 2D NPs.

Fig. 6. Comparison of the maximum VN 2PA and 3PA cross sections of Mn-doped CsPbCl3 2D NPs and cubic NCs, as well as Mn-doped ZnS spherical NCs.

下载图片 查看所有图片

4. CONCLUSION

In summary, we synthesized Mn-doped CsPbCl3 2D NPs and cubic NCs for comparative purposes. Their linear absorption cross sections and the strengths of their exciton–dopant couplings were investigated via fs-TA spectroscopy. Most importantly, we determined that the strong quantum confinement effects of the 2D NPs can be used to enhance their MPA. Our broadband Z-scan measurements show that the 2D NPs exhibit much better MPA from the visible light band to the second biological window compared to their cubic counterparts. Additionally, we confirmed that the MPA mechanisms of the 2D NPs and cubic NCs for the excitation wavelengths in the 620–700 nm and 1300–1500 nm ranges can be attributed to two-photon excited excitonic energy transfer from the host to the Mn2+ and direct 3PA transition from the A16 state to the T14 state, respectively.