Multifold Enhanced Photon Upconversion in a Composite Annihilator System Sensitized by Perovskite Nanocrystals

Photon upconversion via triplet–triplet annihilation (TTA-UC) provides a pathway to overcoming the thermodynamic efficiency limits in single-junction solar cells by allowing the harvesting of sub-bandgap photons. Here, we use mixed halide perovskite nanocrystals (CsPbX3, X = Br/I) as triplet sensitizers, with excitation transfer to 9,10-diphenylanthracene (DPA) and/or 9,10-bis[(triisopropylsilyl)ethynyl]anthracene (TIPS-An) which act as the triplet annihilators. We observe that the upconversion efficiency is five times higher with the combination of both annihilators in a composite system compared to the sum of the individual single-acceptor systems. Our work illustrates the importance of using a composite system of annihilators to enhance TTA upconversion, demonstrated in a perovskite-sensitized system, with promise for a range of potential applications in light-harvesting, biomedical imaging, biosensing, therapeutics, and photocatalysis.

I n organic molecules, upon light excitation, electrons from the ground state are excited to the singlet excited state and leave behind positively charged holes, producing singlet excitons that can show efficient radiative recombination. 1 In contrast, triplet excitons cannot be directly photoexcited and are typically generated from singlet excitons via intersystem crossing.Once formed, returning to the ground state requires a spin-flip, and therefore triplet excitons are typically nonemissive and long-lived.A way to overcome this consists of taking advantage of triplet−triplet annihilation (TTA). 2 This is a process whereby two triplet excitons from different molecules encounter each other, resulting in one emitter in the ground state and another in the singlet excited state that can undergo fluorescence.The mechanism of TTA was discovered in the 1960s, 3 and TTA was utilized for photon upconversion in the 2000s, 4,5 with applications including solar energy, 6,7 opto-genetics, 8,9 catalysis, 10,11 lighting, 12,13 and biomedical imaging. 14,15In solar photovoltaics, harvesting sub-bandgap photons by means of upconversion is a path toward overcoming the thermodynamic efficiency limits in singlejunction solar cells. 16In biomedical imaging, upconverted luminescence can be isolated from the excitation light by using short-pass filters, increasing the signal-to-noise ratio. 17−28 However, each approach has its limitations.For instance, metal−organic complexes have significant exchange energy losses of up to 300 meV.Phosphorescent molecules are in general costly.Lead chalcogenide nanocrystals have poor exciton diffusion lengths, resulting in inefficient exciton transport, which hinders their triplet sensitization ability in films. 29,30−47 These include tunable bandgaps via facile compositional changes, efficient light absorption, long carrier lifetimes and diffusion lengths, low nonradiative recombination, shallow defects, and solution-processability.Additionally, strong spin−orbit coupling due to the abundance of heavy atoms can rapidly spinmix electrons and holes, providing the possibility of triplet sensitization via separate injection of free electrons and holes into the upconverting organic semiconductor.
Perovskite nanocrystals have also lately received immense attention for optical applications. 48,49Kimizuka and co-workers demonstrated the potential of perovskite nanocrystals for photon upconversion. 29They used halide-exchanged CsPbBr 3 nanocrystals as triplet sensitizers, which undergo energy transfer to surface-bound triplet acceptors possessing an amino group, which in turn readily relay triplets to free 9,10diphenylanthracene (DPA) molecules in solution that undergo TTA.They obtained a low threshold intensity of 25 mW/cm 2 , where the excitation intensity dependence of the upconversion emission intensity changes from quadratic to linear.The maximum TTA efficiency (normalized to 100%) was 1.3%.
Since then, Wu et al. demonstrated triplet energy transfer from perovskite nanocrystals to organic ligands in a series of works using time-resolved photoluminescence and transient absorption studies.They showed that triplet energy transfer can also be mediated by electron transfer. 31In addition, they demonstrated that quantum-confined nanocrystals improve the triplet energy transfer efficiency significantly compared to larger, bulk-like nanocrystals.This is because quantumconfinement enables a strong donor−acceptor wave function overlap required for Dexter-type triplet energy transfer.For instance, the energy transfer efficiency rises from 0.006% at 11.2 nm nanocrystals to 99.4% at 3.5 nm quantum-confined nanocrystals. 32This approach was adopted to achieve above 10% efficiency for visible-to-ultraviolet upconversion, with 1naphthalenecarboxylic acid as the organic ligand and 2,5diphenyloxazole as the emitter. 33The efficiency achieved is 2.5 times more than a similar system but using bulk nanocrystals. 34ost recently, Koharagi et al. provided a demonstration of TTA-UC from green-to-UV light sensitized by CsPbBr 3 perovskite nanocrystals. 35The triplet energy transmitter was 4-(2-phenyloxazol-5-yl)benzenesulfonate with a sulfonate group, and the emitter was TIPS-naphthalene.The TTA-UC efficiency obtained was 0.014% at an excitation intensity of 16 W/cm 2 .−40 However, the efficiencies of perovskite-sensitized TTA-UC still often lag behind those of systems sensitized by organic molecules.For example, unlike an efficiency of 2.55% by Askes et al. 50or ∼4% by Cao et al. 51 for platinum(II) Scheme 1. Energetics of the System Design for Green-to-Blue TTA-UC Sensitized by CsPbX 3 (X = Br/I) Nanocrystals a a Photoexcitation (green arrow) of the nanocrystals is followed by triplet energy transfer (TET) to the surface-bound 1-PCA.This is then followed by TET to free emitter molecules in solution, which undergo triplet−triplet annihilation (TTA), and finally higher-energy upconverted emission (lighter blue arrows).The free emitter molecules could be (a) DPA, (b) TIPS-An, or (c) both DPA and TIPS-An.octaethylporphyrin (PtOEP) as the triplet sensitizer and DPA as the annihilator, the early demonstration of perovskites as triplet sensitizers using DPA as the annihilator as well obtained an efficiency of only 1.3%.
Here, we report triplet−triplet annihilation in a composite system of annihilators sensitized by perovskite nanocrystals, as a means to obtain significantly enhanced upconversion efficiency.The system design is summarized in Scheme 1, where photoexcitation of the nanocrystals is followed by triplet energy transfer to the organic ligand, then triplet energy transfer to the free emitter molecules in solution, which undergo TTA.Two blue emitters are used for our work: DPA and TIPS-An.The upconversion yield of the composite system (Scheme 1c) is more than five times that of the sum of the single-acceptor systems (Scheme 1a,b).Our work presents a simple and promising approach for highly efficient perovskitesensitized triplet−triplet annihilation upconversion systems that could be utilized in a range of light-harvesting, biomedical, and photocatalysis applications.

RESULTS AND DISCUSSION
We use mixed halide perovskite nanocrystals as the triplet sensitizers.CsPbBr 3 perovskite nanocrystals with oleic acid and oleylamine as ligands were first synthesized using a hot injection method.PbI 2 was then added to get the final mixed halide composition of approximately 30% iodide out of total halide by halide substitution, obtaining chemically stable nanocrystals without noticeable halide segregation (see the Methods section for details).Figure 1a−c shows the steadystate absorption and photoluminescence spectra of the pristine  organic molecules in our system (1-pyrenecarboxylic acid (1-PCA), DPA, and TIPS-An).Figure 1d−g shows these spectra for the pristine perovskite nanocrystals and the surfacemodified perovskite nanocrystals with various emitters in toluene.The pristine perovskite nanocrystals have an absorption onset at ∼550 nm (∼2.26 eV) with a corresponding emission peak at ∼568 nm (Figure 1d).To avoid the long alkyl chains of oleic acid and oleylamine on the pristine nanocrystal surface from hindering triplet energy transfer from the nanocrystals to free emitter molecules in solution by interrupting the short-range Dexter electron transfer, 29 we use 1-PCA to modify the surface state of perovskite nanocrystals (Figure 1a).1-PCA was chosen as the ligand given its enhanced quenching efficiency of the perovskite nanocrystals as compared to other ligands considered (Figure S1).TEM measurements show that the size distribution (10.6 ± 2.3 nm) and morphology of the nanocrystals are retained after surface modification with the 1-PCA ligand (Figure S2).XRD measurements also confirm that the cubic perovskite structure has been retained (Figure S3).The triplet energy is transferred from the photoexcited nanocrystals (bandgap, 2.26 eV) to 1-PCA ligands (triplet energy level, 2.00 eV), 52 and the 1-PCA ligands will pass on the energy from their triplet states to that of free molecules in solution which then emit.The blue emitters used for our work are DPA (triplet energy level of 1.77 eV) 53 (Figure 1b) and TIPS-An (with a lower triplet energy level between 1.40 and 1.55 eV) 54 (Figure 1c).These are complementary emitters at both extremes of the blue range of the visible spectrum (the emission of TIPS-An only begins at 425 nm while that of DPA begins at 400 nm), giving us room to differentiate their upconverted signals in the composite system (Figure 1e−g).
To understand how upconversion is promoted by using 1-PCA to mediate triplet energy transfer from the nanocrystals to the organic emitters, we employ transient absorption (TA) spectroscopy to access the early dynamics of the charge carriers upon photoexcitation.Figure 2a shows the TA spectra (up to 1500 ps) of the composite system under a 540 nm pump with a fluence of 3.65 μJ/cm −2 (see Figure S4 for the TA spectra of the control and single-acceptor systems).Under this excitation, only the nanocrystals are photoexcited by the pump.The positive signal peaking at ∼560 nm results from the groundstate bleach (GSB) of perovskite nanocrystals, which also reflects the carrier density of the photoexcited nanocrystals.Figure 2b shows the kinetics of the GSB of different systems under the same excitation conditions.The pure nanocrystals (Figure 2b, red) have the longest lifetime corresponding to intrinsic carrier recombination processes.Upon adding 1-PCA, a faster initial decay on the time scale of ∼100 ps appears as compared to the pristine nanocrystals.This demonstrates the quenching of charge carriers from nanocrystals to the 1-PCA ligands (Figure 2b, yellow). 55The transfer rate, k et , is calculated to be 1.4 × 10 8 s −1 by using k et = 1/τ 1-PCA − 1/ τ NC , where τ NC (Figure 2b, red curve) and τ 1-PCA (Figure 2b, yellow curve) are the initial decay times of the pristine NCs and the NCs with 1-PCA measured using transient absorption.The energy transfer efficiency can be estimated as 1 − τ 1-PCA / τ NC = 0.12 (see the SI for discussion on low efficiency).Furthermore, this energy transfer is maintained when the blue emitter molecules are added to the ensemble (Figure 2b, brown, blue, and green).
Figure 3a−c shows the normalized upconversion emission of surface-modified mixed halide nanocrystals with a single acceptor of DPA, single acceptor of TIPS-An, and dual acceptors of DPA and TIPS-An respectively, with excitation densities from 0.4 to 3.4 W/cm 2 .After preliminary optimization, we have chosen the concentrations for TTA to be 15 mM for DPA and 1 mM for TIPS-An (Figure S5).The upconversion emission for DPA peaks at 435 nm, while the upconversion emission for TIPS-An peaks at 475 nm.The upconversion emission from the nested system of DPA and TIPS-An closely resembles the emission spectrum of TIPS-An, indicating that the upconversion emission originates from TIPS-An only, instead of DPA or a mixture of both.As DPA is much more concentrated than TIPS-An, we propose that DPA acts as a mediator for TIPS-An.
Figure 3d shows the relative intensity of upconversion emissions of the three systems under study, for the same excitation intensity of 430 mW/cm 2 (excitation wavelength at 532 nm).As mentioned, the intensity of upconversion emission for the nested system (Figure 3d, brown) is much higher than the linear sum of the emissions from the singleacceptor systems (Figure 3d, blue and green).Control experiments with identical concentrations were performed to confirm that the emissions result from the upconversion of the composite system.No upconversion emission is observed in the absence of organic emitters (Figure 3d, yellow).Upconversion emission is also not detected for a mixture of the mixed halide nanocrystals and DPA/TIPS-An without surface modification by 1-PCA (Figure 3d, orange).This result confirms that the surface modification by 1-PCA is critical for triplet energy transfer to occur from the nanocrystals to 1-PCA, and onward to the free emitter molecules in solution.Finally, no upconversion emission is observed from a solution of 1-PCA, DPA, and TIPS-An without perovskite nanocrystals (Figure 3d, red).This shows that the triplet excited state of organic molecules is populated by direct excitation of the perovskite nanocrystals.
To discard changes in absorption affecting the relative upconversion signals, we measured the quantum yield of the upconversion emissions, as shown in Figure 3e (see Methods for details of the TTA efficiency calculations).Since the maximum yield of a bimolecular TTA-UC process is 50%, we normalize the upconversion quantum yield against 100% and denote it by Φ' UC .With the increase of excitation intensity, Φ' UC gradually increases and becomes saturated.The highest upconversion quantum yield of the nested system is 0.172% (Figure 3e, brown), much higher than the linear sum of both single-acceptor systems (0.019%, blue and 0.015%, green, respectively).We also performed direct excitation of singlet states using a 400 nm pump for the nested system.We observed an approximate linear sum of the emissions of the single-acceptor systems (Figure S6), confirming the distinct nature of the nonlinear enhancement of TTA efficiency in the nested system via the mediator.
Figure 4 shows the excitation density dependence of the integrated upconversion emission of all of the TTA-UC systems.The TTA-UC emission intensity shows quadratic dependence at low excitation intensity and linear dependence at high excitation intensity.The quadratic regime is dominated by radiative and nonradiative decay of emitter triplets, whereas TTA becomes the main deactivation channel for the acceptor triplet states at the linear regime. 56The threshold where the quadratic-to-linear transition takes place is 390 mW/cm 2 for DPA, 158 mW/cm 2 for TIPS-An, and 260 mW/cm 2 for DPA/ TIPS-An.
Having demonstrated the nonlinear enhancement of TTA efficiency in the composite system, we turn our attention to the possible mechanistic pathways.The 1-PCA ligands do not have optical signals in a wavelength region distinct from the annihilator molecules�either ground-state bleach (GSB, bleaching of S 0 to S n ) or excited-state absorption (absorption from S 1 to S n )�so it would be hard to directly probe carrier density change in the 1-PCA ligands.We therefore rely on the time-dependent carrier densities in the perovskite nanocrystals from TA measurements for insights.A comparison of the GSB kinetics between different samples in Figure 2b shows that all samples with additives (1-PCA, DPA, TIPS-An) quench the pristine nanocrystals to the same degree, as evidenced by the similar initial decay of the GSB.This indicates that the addition of DPA or TIPS-An would not dramatically change the carrier extraction from the nanocrystals.This is consistent with Figure 4 where the threshold of the DPA + TIPS-An system (brown) lies between that of the individual DPA (green) and the TIPS-An systems (blue).In addition, we note that the observed enhancement cannot be due to absorption differences alone, due to the same concentration of perovskite nanocrystals in all systems.We now examine possible pathways for this enhancement (Figure S7).In mechanism (I), triplet excited states from two DPA molecules interact to produce a singlet excited state of DPA, which immediately transfers its energy to the singlet of TIPS-An, i.e., T 1 (DPA) + T 1 (DPA) → S 1 (DPA) → S 1 (TIPS-An).In mechanism (II), the lower T 1 (TIPS-An) as compared to T 1 (DPA) facilitates an efficient energy cascade of triplets via T 1 (DPA) to T 1 (TIPS-An), leading to enhanced concentration of triplet excited states in TIPS-An.Two triplet excited states from two TIPS-An molecules then interact to produce a singlet excited state of TIPS-An.
The relatively lower threshold of the composite system compared with the DPA-only system in Figure 4 shows that it is unlikely that the TTA in the composite system is only related to DPA.This indicates that mechanism (I) probably plays a less significant role in accounting for the enhancement.Additionally, the degree of quenching is very similar for surface-modified nanocrystals with different emitters in Figure 2b (green for DPA; blue for TIPS-An; brown for DPA and TIPS-An), indicating that the addition of an emitter does not significantly promote the process of energy transfer from nanocrystals to organic molecules.It is therefore unlikely that the mere addition of DPA, behaving as a triplet intermediary between 1-PCA and TIPS-An, would, on its own, lead to the nonlinear enhancement of TTA between TIPS-An molecules.Furthermore, the photoluminescence quantum yield of TIPS-An is lower than that of DPA.Hence, this indicates that mechanism (II) also probably plays a diminished role in accounting for the enhancement.
There is a third mechanism that could be consistent with the above results: a composite TTA process between a DPA molecule and a TIPS-An molecule.This is illustrated in Scheme 1c, where a triplet excited state from DPA and a triplet excited state from TIPS-An undergo TTA to produce a bright singlet state.The singlet of DPA S 1 (DPA, 3.16 eV) and the singlet of TIPS-An S 1 (TIPS-An, 2.87 eV) can be reached from T 1 (DPA, 1.77 eV) + T 1 (TIPS-An, 1.40−1.55eV).As S 1 (DPA) is resonant with T 1 (DPA) + T 1 (TIPS-An), upconversion to S 1 (DPA) is more likely to happen.The excitation then undergoes ultrafast energy transfer from S 1 (DPA) to S 1 (TIPS-An), leading to the observation of the lower-energy PL from TIPS-An dominating the composite system's emission spectrum.It may also be possible for T 1 (DPA) and T 1 (TIPS-An) to undergo TTA to populate S 1 (TIPS-An) directly, although we believe this is less likely due to the resonance of S 1 (DPA) with T 1 (DPA) + T 1 (TIPS-An) instead.Scheme 1b represents the underlying TTA-UC arising from direct triplet transfer from 1-PCA to T 1 (TIPS-An) in the nested system.We provide a mathematical model to provide insights for why we believe Scheme 1c may lead to an overall higher TTA-UC emission intensity, taking into consideration that both the emitters DPA and TIPS-An individually have TTA-UC efficiencies close to their theoretical limits when used in other systems (see Section 2 in the SI).It is however challenging to obtain direct time-resolved evidence due to the low overall upconversion efficiencies (see Section 3 in the SI for further discussion), and further work utilizing more efficient TTA systems will be valuable and the subject of our future studies.
Finally, we comment on similarities and differences between our work and others (Figure S8).Cao et al. 51 demonstrated a synergistic effect in phosphor-sensitized TTA-UC systems arising from TTA upconversion between triplet acceptors of different types, which they term "hetero-TTA", and the upconversion signals from both acceptors are observed (Figure S8a,b).We also note a recent report by Yurash et al. 57 which utilized a three-component system employing thermally activated delayed fluorescence sensitizers, achieving a doubling of upconversion efficiency, with upconversion observed only from one emitter (Figure S8c,d).Our work in fact combines features of both reports, where we employ a composite system of annihilators, and we observe upconversion signals purely from TIPS-An in this composite system.Our system design has generated a 5-fold enhancement compared to the sum of individual systems.In addition, our work represents the successful report of significant enhanced TTA-UC in a perovskite-sensitized system involving a mix of annihilators.Our work confirms the generality of multiacceptor systems in enhancing TTA-UC efficiencies, and will be highly applicable to other upconversion systems employing different sensitizers.The higher efficiencies to be gained, for example, using quantum-confined nanocrystals (see Figure S9 and SI for further discussion), will further exemplify the detailed photophysical processes and limits of such enhancements, in turn bringing further exciting advances to these highly enhanced photon upconversion systems.

CONCLUSIONS
We report the observation of significantly enhanced TTA upconversion in a composite system comprising mixed halide perovskite nanocrystals as the triplet sensitizers, 1-PCA ligands to transfer triplet energy to the triplet states of free molecules in solution, and both DPA and TIPS-An as the annihilators.The highest upconversion quantum yield of the composite system is 0.172%, much higher than the linear sum of both single-acceptor systems (0.019 and 0.015%).The upconversion emission of the composite system originates from TIPS-An only.We obtain the threshold values of the quadratic-tolinear transition and the transient absorption kinetics, and discuss three possible mechanistic pathways of TTA-UC in the composite system.Our results suggest that the enhanced efficiency may arise from an energy resonance between the sum of the triplet excited states of DPA and TIPS-An with the singlet excited state of DPA.This resonance is followed by ultrafast energy transfer to the emissive singlet state of TIPS-An, leading to upconversion emission dominated by TIPS-An.Our work demonstrates the distinct opportunities and benefits that a composite system of annihilators provides to significantly increase the performance of TTA upconversion systems, utilized with mixed halide perovskite nanocrystals as triplet sensitizers, and highlights the need for further in-depth photophysical investigations in more efficient systems.The simple approach here will find practical use in a range of applications such as light-harvesting, biosensing, therapeutics, biomedical imaging, and photocatalysis.
Nanocrystal Synthesis.The synthesis of CsPbBr 2.1 I 0.9 nanocrystals was performed by modifying the procedure by Kovalenko et al. 58 Cesium carbonate (Cs 2 CO 3 , 0.814 g, 2.5 mmol), oleic acid (OA, 2.5 mL, 7.5 mmol), and octadecene (ODE, 30 mL) were added into a three-neck flask, degassed under vacuum for 1 h at 110 °C, and then heated under N 2 to 150 °C until all of the Cs 2 CO 3 dissolved and reacted with OA.Oleylamine (OLA, 1.5 mL, 4.5 mmol), OA (1.5 mL, 4.5 mmol), ODE (15 mL), and PbBr 2 (0.207 g) were added into another three-neck flask and degassed for 1 h at 110 °C.After complete solubilization of PbBr 2 salt, a clear solution was formed.Then, the temperature was raised to the reaction temperature (170 °C) under N 2 and the Cs-oleate precursor (1.2 mL) was quickly injected into the PbBr 2 solution.After 5 s, the reaction mixture was cooled by an ice-cold water bath to room temperature (25 °C), followed by centrifuging at 12,000 rpm (16728 RCF) for 5 min.The supernatant was discarded, and the nanocrystals were redispersed in toluene.For mixed halide nanocrystals, the PbI 2 solution prepared by adding 0.2 mmol of PbI 2 into 10 mL of toluene with 0.2 mL of OA and OLA as ligands was added into the CsPbBr 3 nanocrystals.The mixture was stirred at 40 °C until the anion exchange was fully complete.The obtained mixed halide nanocrystals were purified by centrifuge at 3000 rpm (1045 RCF), without antisolvent.
Sample Preparation.For surface modification, equal volumes of mixed halide nanocrystals solution and 1-pyrenecarboxylic acid (1-PCA) in toluene were added, to obtain final solution concentrations of 5.0 mg/mL mixed halide nanocrystals and 0.95 mM 1-PCA respectively.The anhydrous toluene solvent was not degassed, and used as received from Sigma-Aldrich.The mixture was stirred at room temperature for 6 h at 100 rpm in capped vials further sealed with parafilm.The obtained mixture was filtered through a 0.45 μm PTFE syringe filter.15 mM DPA, 1 mM TIPS-An, or both 15 mM DPA and 1 mM TIPS-An were then added and mixed thoroughly before the solution was transferred into a 1 mm path length quartz cuvette which was sealed with a PTFE cap and parafilm.The concentration of 15 mM DPA or 1 mM TIPS-An for single-acceptor systems was chosen after preliminary optimization (Figure S5).The entire procedure, except stirring, was conducted in a N 2 -filled glovebox.
Steady-State Absorption and Photoluminescence Spectroscopy.Absorption spectra were measured using a Shimadzu UV3600-Plus UV−vis−NIR spectrophotometer.Photoluminescence spectra were measured on an Edinburgh Instruments FLS980 fluorimeter with a 450 W continuous xenon arc lamp.The samples were filled in a 1 mm path length quartz cuvette.
Measurement of Upconversion Spectra with Integrating Sphere.Upconverison spectra were measured in an integrating sphere with a detector constructed with a spectroscopy camera (Andor iDus DU420A Si detector, ANDOR) coupled to a spectrograph (Shamrock SR303i, ANDOR) and samples filled in a 1 mm path length quartz cuvette excited by a 532 nm CW laser (Thorlabs) passing through an FES0500 short-pass filter (Thorlabs).
Quantum Efficiency Measurement of Mixed Halide Nanocrystals Solution with Integrating Sphere.PLQE measurements were made following the procedure of de Mello et al. 59 Temperatureand current-controlled laser diodes (Thorlabs) were used to generate stable laser beams with wavelength 405 or 520 nm.After attenuation to the desired intensity, these were aligned through a small hole, onto samples suspended, in a Spectralon-coated integrating sphere (Newport 819C-SL-5.3)modified with a custom baffle extension.Light from the experiment was collected using an optical fiber connected to an Andor Kymera 328i spectrograph, and spectra were recorded using a Si-CCD (Andor iDus 420).
Quantum Efficiency Measurement of TTA-UC Samples with Relative Method.The normalized upconversion quantum yield Φ' UC was calculated indirectly with the following equation: 60 = i k j j j j j y { z z z z z i k j j j j j y { z z z z z i k j j j j j y { z z z z z i k j j j j j y where Φ R is the quantum yield of the reference, A is the absorption at the excitation wavelength, E is the integrated photoluminescence spectral profile, I is the light intensity at the excitation wavelength, and hν is the energy of excitation photons.The subscripts UC and R denote the upconversion and reference samples respectively.The mixed halide nanocrystals in deaerated toluene solution with an excitation of 405 nm (Φ R = 0.290) and 520 nm (Φ R = 0.353) were used as the reference, and the upconversion efficiency was calculated from the average of both independent values.The factor of 2 arises because the maximum yield of a bimolecular TTA-UC process is 50% since two photons are required to generate one upconverted photon.Therefore, we normalize the upconversion quantum yield to 100% and denote it by Φ' UC .This allows comparison with other recent works in perovskite-sensitized TTA-UC systems. 29,33,34,39,61We acknowledge the push in the field to report unnormalized TTA-UC efficiencies. 62Without normalization, the upconversion quantum yields would be half that reported above in the main text.The TTA efficiency could not be obtained directly because a 500 nm short-pass filter was necessary to prevent the saturation of the detector.Transmission Electron Microscopy.Transmission electron microscopy samples were prepared by putting a small drop of mixed halide nanocrystals solution onto a carbon-coated copper grid in ambient air.Transmission electron microscopy images were recorded using an FEI Tecnai F20 (200 kV).
Time-Correlated Single Photon Counting (TCSPC).TCSPC plots were measured on an Edinburgh Instruments FLS1000 fluorimeter equipped with a 450 W continuous xenon arc lamp.A picosecond pulsed diode laser at 505.8 nm (EPL-510) was used (repetition rate of 5 MHz at a fluence of 16.2 nJ/cm 2 /pulse), which was connected to the spectrometer using a coupling flange.
Transient Absorption Spectroscopy.The output of a Ti:sapphire amplifier system (Spectra Physics Solstice Ace) operating at 1 kHz and generating ∼100 fs pulses was split into the pump and probe beam paths.The 540 nm pump pulses were created by sending the 800 nm fundamental beam of the Solstice Ace into a home-built noncollinear optical parametric amplifier (NOPA).The pump goes through a 540 nm bandpass filter and then was blocked by a chopper wheel rotating at 500 Hz.The ultraviolet−visible broadband beam (330−700 nm) was generated by focusing the 800 nm fundamental beam onto a CaF 2 crystal (Eksma Optics, 5 mm) connected to a digital motion controller (Mercury C-863 DC Motor Controller) after passing through a mechanical delay stage (Thorlabs DDS300-E/M).The transmitted pulses were collected with a monochrome line scan camera (JAI SW-4000M-PMCL, spectrograph: Andor Shamrock SR-163) with collected data fed straight into the computer.

Figure 2 .
Figure 2. (a) TA spectra of the composite system under a 540 nm pump (up to 1500 ps).(b) TA kinetics of the TTA-UC systems (normalized at peak) for TIPS-An only (blue), DPA only (green), and both DPA and TIPS-An (brown).The control samples are the pure mixed halide nanocrystals (red) and the surface-modified mixed halide nanocrystals (yellow).

Figure 3 .
Figure 3. Normalized upconversion emission spectra of surface-modified mixed halide nanocrystals (5.0 mg/mL with 0.95 mM 1-PCA in toluene) with (a) single acceptor of DPA (15 mM), (b) single acceptor of TIPS-An (1 mM), and (c) dual acceptors of DPA (15 mM) and TIPS-An (1 mM).The spectra in (c) closely resemble that of (b).A 532 nm CW laser was used with excitation densities from 0.4 to 3.4 W/ cm 2 .A 500 nm short-pass filter was used to remove pump and nanocrystal emissions.(d) Upconversion emission spectra of surface-modified mixed halide nanocrystals (yellow) with TIPS-An only (blue), DPA only (green), and both DPA and TIPS-An (brown), at 430 mW/cm 2 excitation density, with concentrations identical to (a−c).Control samples are in the absence of emitters (yellow), 1-PCA (orange), or nanocrystals (red) where no upconversion can be observed.(e) Excitation density dependence of Φ' UC for TTA-UC samples (in toluene).The samples were excited with 532 nm CW laser at sufficiently high intensities until the Φ' UC values are saturated.

Figure 4 .
Figure 4. Excitation density dependence of integrated upconversion emission intensity for TTA-UC samples (in toluene).Solid lines show the linear fitting with slopes indicated.The intersection gives the threshold excitation density, denoted by vertical dashed lines.