Unveiling Dopant‐Induced Ultrafast Exciton Dynamics in Mn/Yb Codoped Perovskite Nanocrystals

Perovskite nanocrystals (NCs) with intentionally introduced Mn2+/Yb3+ activators enable tunable emissions covering UV‐orange‐NIR spectral range. However, the exact microscopic energy transfer mechanisms in this system remain unknown. Herein, Mn/Yb codoped CsPbCl3 perovskite NCs with triple emissions originated from exciton recombination of host, 3d–3d transition of Mn2+ and 4f–4f transition of Yb3+ are prepared. Femtosecond resolution transient absorption spectra performed on the pristine CsPbCl3, Mn‐doped, Yb‐doped and Mn/Yb codoped samples clarify efficient and simultaneous energy transfer (ET) from excitons to Mn2+ and Yb3+ dopants. It is testified the sensitizations of dopants mainly result from the trapped hole, taking 285 ps for Mn2+ and 17 ps for Yb3+ respectively, which make less trapped hole recombine with de‐localized carriers. Importantly, energy transfer processes from host to Mn2+ and Yb3+ activators emerge as competition, and the ET probability of exciton‐to‐Mn2+ is higher than that of exciton‐to‐Yb3+. Finally, control experiments further prove that tunable Mn2+ orange emission and Yb3+ NIR emission are achievable via elaborate adjustment of the dopant concentrations.

To deeply understand energy transfer from excitons to dopants, transient dynamics of excited states in the Mn-doped and Yb-doped APbX 3 QDs have been studied by femtosecond (fs) pump and probe spectroscopy. The fitting models for chargetransfer kinetics from colloidal CsPbX 3 NCs to dopants have been proposed and gradually modified by T. Q. Lian, D. H. Son, R. Gamelin et al. [10,31,44,45] It is evidenced that shallow hole traps mediated ET from excitons to Mn 2+ with a spin exchange in approximately hundreds of picoseconds [31,32] results in orange Mn 2+ dd emission. [46] Likewise, the Yb 3+ -induced shallow defects [10,47] near the valence band play a critical role in facilitating nonradiative ET process within tens of picoseconds that de-excites excitons and simultaneously excites two Yb 3+ dopants. [10,45] In fact, codoping Mn/Yb into perovskite NCs has been reported to produce a unique triple-wavelength emitting feature covering UV/blue, visible, and NIR regions with PLQY up to 125.3% and minimal reabsorption loss, enabling them as efficient emitters in luminescent solar concentrators. [39] Time-resolved PL spectra reveal that these two different kinds of dopants can be simultaneously sensitized by excitons in one perovskite host. However, the exact microscopic ET mechanisms in the Mn/Yb codoped system remain unknown. Herein, a series of Mn 2+ : CsPbCl 3 , Yb 3+ : CsPbCl 3 and Mn 2+ /Yb 3+ : CsPbCl 3 NCs are successfully prepared, and transient absorption (TA) spectra are scrutinized to explore the ultrafast exciton-to-dopant energy transfer in the codoped perovskite for the first time.

Results and Discussion
Mn/Yb codoped CsPbCl 3 NCs were synthesized by a modified hot-injection procedure reported in our previous work. [36] Herein, the reaction temperature was elevated to 240°C to enable easy doping of Mn 2+ and Yb 3+ ions into CsPbCl 3 lattice. [10,27,30] XRD patterns (Figure 1a, Figures S1 and S2, Supporting Information) evidence that all the products are coincident with cubic CsPbCl 3 phase (JCPDS No. 75-0411). [30,36] Introducing Mn 2+ and Yb 3+ dopants will not result in the formation of impurity phases. The slight shift of Bragg diffraction peak toward high-angle direction proves that the Pb 2+ (1.19 Å, CN = 6) ions [48,49] are substituted by smaller Mn 2+ (0.97 Å, CN = 6) [20,30,35] and Yb 3+ (1.02 Å, CN = 6) [21] ones. Under UV light irradiation, the un-doped, Mndoped, Yb-doped, Mn/Yb-doped samples produce violet, orange, violet and orange luminescence taken by visible camera, respectively, while only the Yb-doped and Mn/Yb-doped samples yield bright NIR luminescence taken by NIR camera (Figure 1b). The orange and NIR emissions come from Mn 2+ : d-d and Yb 3+ : f-f transitions, respectively. X-ray photoelectron spectroscopy (XPS) measurements on the CsPbCl 3 , Mn: CsPbCl 3 , Yb: CsPbCl 3 and Mn/Yb: CsPbCl 3 samples were performed ( Figures S3-S6, Supporting Information). Typical Cs 3d, Pb 4f, and Cl 2p signal peaks were detected in the high-resolution XPS spectra, and extra Mn 2p and Yb 4d peaks were observed for the Mn/Yb: CsPbCl 3 sample. Additionally, as shown in Figure 1c, the shifting of the Pb 4f peak toward higher energy is observed after Mn/Yb doping. Generally, the substitution of Pb 2+ ions by Mn 2+ and Yb 3+ dopants in CsPbCl 3 leads to the alteration of Pb 2+ ligand environment and thus a slight shift to a higher binding energy of the Pb 2+ 4f XPS peak.
The carrier transition and ET mechanisms in the pristine CsPbCl 3 NCs, Mn 2+ single-doped and Yb 3+ single-doped and Mn 2+ /Yb 3+ codoped CsPbCl 3 NCs were examined by transient absorption (TA) experiments (details in Note S1, Supporting Information). The 360 nm pump pulse was used to pump the samples with a low power density of 0.9 μJ cm −2 to avoid nonlinear auger process. The induced absorption change (ΔA) as a function of probe wavelength and pump-probe delay (Δt) within the time window of 0-8 ns was recorded. [31,44] The fs-resolved TA spectra for the corresponding samples are shown in Figure 3 and Figures S16 and S17 (Supporting Information). A negative signal dominates in the TA spectra for all the samples, which is assigned to exciton bleach (XB) in line with previously reported. [44,49,54,55] The XB peak of the undoped sample is located at 403 nm ( Figure 3a) and tends to blue-shift after Mn/Yb doping (Figure 3b-d). The result is consistent with absorption phenomenon ( Figure S18, Supporting Information), confirming that the XB is originated from the inter-band transition. [31,56] Compared to the un-doped sample, the weaker intensities of XB and shorter de-bleach times for the doped samples indicate the faster depletion of excitons on the present of Mn 2+ and Yb 3+ activators. [54] We further extracted and fitted the kinetics of XB to reveal the carrier recombination pathways in the Mn/Yb doped samples by the method proposed by Gamelin and Lian et al. (Note S1, Figure S19, Supporting Information and Figure 4a). [10,44] The TA kinetic time constants and amplitudes obtained by multiexponential function fitting are tabulated in Table 1, Tables S2 and  S3 (Supporting Information).
For the undoped sample, the kinetic curve can be well fitted by a five-exponential decay function, the related kinetic processes are schematically illustrated in Figure 4a and Scheme S3 (Supporting Information). [10,44] (Figure 2d), is categorized into band edge exciton recombination. [10,31,44] With the introduce of Mn and/or Yb dopants, the shortened 5 value verifies the occurrence of efficient ET from host to Mn/Yb. [10,30,31] Therefore, extra ET channel should be added into the fitting function for the Mn/Yb doped samples. [10,57] As tabulated in Table 1 and presented in Figure 4b, the energy transfer lifetime components ( 6 , 7 ) for Mn-doped and Yb-doped samples are determined to be 285 and 17 ps, respectively. These two values are well consistent with those previously reported. [10,56,57] Energy transfer from host to dopants usually requires the following physical processes at least: 1) photoinduced carriers rapidly relax to band edges in sub-picosecond timescale, [49,58,59] 2) exciton generations, [58,60] and 3) impurity ions sensitized due to energy transfer from excitons. [10,37,56,57] Therefore, the time of sensitiza-tions of the doped activators longer than the one of XB setup is reasonable.
Upon photoexcitation, both holes and electrons rapidly localize to available trap states on the time scale of sub-ps to 10's of ps, and these trapped carriers can recombine with delocalized band edge carriers within 10's to 100's of ps. [10] The probabilities of electron trapping and their recombination show negligible changes after doping (Figure 4b), clarifying that the trapped electron decays are insensitive to Mn or Yb dopants in the host. On the contrary, the probabilities of hole trapping and their recombination are dramatically altered upon the introduce of Mn or Yb dopants (Figure 4b). The probability of hole trapping significantly increases while that of hole recombination with delocalized carriers decreases, indicating the hole-mediated ET from host to dopants. [61] All the results indicate that the trapped holes are responsible to excite Mn 2+ and Yb 3+ activators, decreasing the probability of trapped hole recombination with carriers (Schemes S4 and S5, Supporting Information). This leads to the consuming of excitons and the shortening of lifetime for exciton recombination. [10,57] Furthermore, the TA kinetic curve of the Mn/Yb codoped sample was fitted by fixing the carriers trapping lifetimes and energy transfer lifetimes to understand the changes of the related processes. Similar to the cases of the single-doped samples, the trapped holes play the key role to assist ET from host to both activators (Table 1, Figure 4c). Mn or Yb sensitized processes involved two steps: a fast hole trapping process followed by energy transfer to dopants involving band-edge electrons and trapped holes. [32] The only difference is that the energy is divided into two parts, one is transferred to Mn 2+ ions and the other is transferred to Yb 3+ ions. As tabulated in Table 1, the ET probabilities for Mn and Yb in the codoped sample is 0.16 and 0.11, respectively. Herein, energy transfer efficiency ( ) is evaluated by the formula Mn = a 6 /(a 5 +a 6 +a 7 ) or Yb = a 7 /(a 5 +a 6 +a 7 ), [32,56] where a 5 , a 6 and a 7 represent the amplitudes (probabilities) for exciton recombination, ET-to-Mn 2+ and ET-to-Yb 3+ , respectively. The corresponding values are determined to be 39% and 27% for exciton-to-Mn 2+ and exciton-to-Yb 3+ energy transfer processes, respectively. Energy transfer to Mn 2+ occurs on the time scale of hundreds of picoseconds, but energy transfer to Yb 3+ appears to occur on the time scale of tens of picoseconds. Notably, ET probability is highly limited by the matching degree of energy states. Therefore, although the time of ET-to-Yb 3+ is much faster than that of ET-to-Mn 2+ , ET probability of the former is lower than that of the latter since the energy mismatch degree (≈1.85 eV) between bandgap of CsPbCl 3 and Yb 3+ : 2 F 5/2 → 2 F 7/2 is far larger than that (≈1.03 eV) between bandgap of CsPbCl 3 and Mn 2+ : 6 A 1 → 4 T 1 .
Our finding confirms that the dopant concentrations of Mn 2+ and Yb 3+ have significant impact on the spectral profile. As shown in Figure 5, controlling experiments were conducted by comparing PL spectra and decay behaviors of the codoped samples with various Mn 2+ /Yb 3+ concentrations. With fixed Mn 2+ content and increase of Yb 3+ content, NIR PL intensity of Yb 3+ gradually enhances while orange PL intensity of Mn 2+ weakens monotonously (Figure 5a). The opposite results are found for the codoped samples with fixed Yb 3+ content and increase of Mn 2+ content (Figure 5d). The decay curves by monitoring exciton recombination show fast deexcitation of excitons for both two types of samples, and the exciton PL decay monotonously decreases with increase of Yb 3+ or Mn 2+ doping concentrations, confirming efficient energy transfer from excitons to dopants (Figure 5b,e).
Importantly, it is evidenced that increasing Yb 3+ doping concentration will prolong Mn 2+ decay lifetime (Figure 5c) for the present Mn/Yb codoped samples. Similar result is detected for the Yb 3+ decay with increase of Mn 2+ doping content (Figure 5f). These results verify that the introduced Yb 3+ codopants will result in the decreased electron population for the Mn 2+ dopants and vice versa. For the sample with fixed Mn 2+ doping concentration and increased Yb 3+ concentration, ET probability from excitons to Yb 3+ ions will enhance, which leads to the reduced electron population of Mn 2+ and the increased decay lifetime of Mn 2+ ; similarly, for the sample with fixed Yb 3+ doping concentration and increased Mn 2+ concentration, ET probability from excitons to Mn 2+ ions will enhance, which leads to the reduced electron population of Yb 3+ and the increased decay lifetime of Yb 3+ . Generally, decreased electron population for Mn 2+ or Yb 3+ dopants will reduce concentration quenching effect and prolong Mn 2+ or Yb 3+ decay lifetime. These results exclude the possibility of energy transfer from Mn 2+ to Yb 3+ or Yb 3+ to Mn 2+ in the present samples, and energy transfer processes from host to Mn 2+ and Yb 3+ activators emerge as competition, which   [10,35,45,56,57] our results not only demonstrate the trapped holes mediated ET from excitons to Mn 2+ in approximately hundreds of picoseconds, but also evidence that the trapped hole also play a crucial role in energy transfer from excitons to Yb 3+ ions for the first time. In addition, since there is no inter-dopant energy interaction between Mn 2+ and Yb 3+ ions for their large mismatching energy states in the present samples, we can analyze the kinetic behaviors for the Mn/Yb codoped system by a similar fitting process and conclude that energy transfer processes from CsPbCl 3 host to Mn 2+ and Yb 3+ activators emerge as competition and the former has higher probability than the latter.

Conclusion
In summary, transient absorption spectra are scrutinized to explore ultrafast exciton-to-dopant energy transfer in the Mn/Yb codoped CsPbCl 3 perovskite NCs. It is evidenced that the trapped holes are responsible to excite Mn 2+ or Yb 3+ activators and the probability of energy transfer to Mn 2+ is higher than that to Yb 3+ owing to their different matching degrees of energy states. Therefore, it is important to elaborately control Mn 2+ /Yb 3+ doping concentrations to tune the dual-emissions for the competitive excitons-to-dopants sensitization processes. It is believed that the understanding of ultrafast energy transfer mechanisms in the doped perovskites will provide a new avenue to design optically tunable light-harvest device and optical modulation device.

Experimental Section
Materials:  2 , 200 μL of Cs-ethanol, and 0-0.8 mmol of Mn(Ac) 2 , 0-1 mmol of Yb(OA) 3 were dissolved in 10 mL of octadecene, 2 mL of oleic acid, and 1 mL of oleylamine. After degassing, the temperature was raised to 240°C and 1 mL of previously synthesized OAM-Cl was swiftly injected to produce perovskite NCs. Detailed preparations of mentioned all NCs and the synthesis of OAM-Cl, ytterbium-oleate and lead-oleate are shown in Schemes S1 and S2 (Supporting Information).
Characterizations: X-ray diffraction (XRD) measurements were carried out on a Bruker D8 Advance X-ray powder diffractometer with Cu K radiation ( = 0.154 nm) operating at 40 kV. The actual chemical compositions were measured by inductively coupled plasma-mass spectrometry (ICP-MS) using a Perkins-Elmer Optima 3300DV spectrometer. Transmission electron microscopy (TEM) image of the sample was obtained on a JEOL JEM-2010 at an acceleration voltage of 200 kV. Scanning TEM (STEM) images were taken on an FEI aberration-corrected Titan Cubed S-Twin TEM operating in a high-angle annular dark-field mode (HAADF). XPS was performed with a VG Scientific ESCA Lab Mark II spectrometer equipped with two ultrahigh vacuum 6 (UHV) chambers. Photoluminescence (PL) spectra, PL excitation (PLE) spectra and decay curves were recorded on an Edinburgh Instruments FLS1000 spectrofluorometer equipped with a continuous xenon lamp (450 W), a pulsed flash lamp, and a 375 nm picosecond pulsed laser. The absolute photoluminescence quantum yields (PLQYs) of the samples were obtained by employing a standard barium sulfate-coated integrating sphere (150 mm in diameter, Edinburgh) as the sample chamber that was mounted on the FLS1000 spectrometer with the entry and output port of the sphere located at a 90°angle from each other in the plane of the spectrometer.
Transient Absorption (TA) Spectroscopy: TA data were collected with a HARPIA spectroscopy system. The output of a commercial Yb:KGW amplifier (1030 nm, ≈200 fs, 8.9 kHz, 10 W, Pharos) was split into two beams. One beam with 9 W power was used to pump an optical parametric amplifier (Orpheus, Light Conversion) and a second harmonic generation (Lyra, Light Conversion) to generate the tunable pump beam (315-2700 nm). The central wavelength of pump beam employed in this work was 360 nm with a pulse energy density of 0.9 μJ cm −2 . The residual 1 W output from the amplifier was first pumped a BBO crystal to generate 515 nm beam, and then focused on a Ti: Sapphire crystal to generate the probe beam (350-750 nm) in a Harpia spectrometer. Pump beam and probe beam were overlapped on the sample. The transmitted probe beam was dispersed by a monochromator (Andor, Andor Kymera 193i) and then detected by CCD. Time delays up to ≈8 ns were achieved via an optical delay line. Colloidal NCs dispersed in hexane were placed in a sealed 2 mm path length quartz cuvette (Hellma Analytics) equipped with a Teflon stir bar. To ensure comparison, we have carefully controlled the density of CsPbCl 3 NCs, the volume of CsPbCl 3 solution injected into the cuvette, light absorbance and other experimental conditions to be the same during the measurement. Moreover, all the TA measurements were performed under sufficiently low power density of fs pump pulse to ensure the average number of excitations per nanocrystal 〈N〉≪ 1. In this case, the TA signals were dominated by the single exciton state and the possibility of biexciton formation is excluded.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.