Turning Self‐Trapped Exciton Emission to Near‐Infrared Region in Thermochromism Zero‐Dimensional Hybrid Metal Halides

Low dimensional lead‐free metal halides have become the spotlight of the research on developing multifunctional optoelectronic materials as their properties show a wide range of tunability. However, most reported low dimensional metal halides only function in the ultra‐violet to visible range due to their large bandgap. Moreover, the organic cation based low dimensional metal halides show limited thermal stability; on the other hand, their inorganic cation based counterparts suffer from limited solution processability. A hybrid cation approach is proposed, where a zero dimensional (0D) metal halide ((DFPD)2CsBiI6) is developed by using mixed organic–inorganic cations: 4, 4‐difluoropiperidine (DFPD) and cesium (Cs+). This ensures both thermal stability and solution processability. Furthermore, [BiI6]3− octahedra are serving as active light absorption units, which ensures the bandgap to be located at the visible region. Its photoluminescence (PL) is further shifted to the near infrared (NIR) region by doping (DFPD)2CsBiI6 with antimony (Sb3+). The developed materials show multifunctional properties: thermochromic behavior, light detection, and NIR light emitting. This study expands the scope of developing multifunctional 0D metal halides.


Introduction
Metal-halide perovskites and their derivatives with unique optoelectronic properties, for example, high absorption coefficient, glass. [3]he optoelectronic properties of metal halides can also be tuned by their chemical compositions as one of the internal factors.The change of chemical composition leads to the change of the dimensionality of the metal halides, [4] specifically, low dimensional structures are formed, typically with the introduction of large molecular size cations.In this context, the term "dimensionality" pertains to the way metal-halide units are connected within the crystal structure at the molecular level, rather than the physical appearance of materials.In low dimensional metal-halides, their ionic nature and soft lattice structure cause their optoelectronic properties to be sensitive to temperature.This ensures their application as thermochromic materials with high sensitivity.Furthermore, in low dimensional metal halides, the strong lattice distortion by pseudo-Jahn−Teller effect can enhance the electron-phonon interactions and promote the formation of self-trapped excitons (STEs). [5]The luminescent STEs in metal halides have attracted tremendous attentions, due to its unique broadband emission with high luminescence efficiency (close to 100%) and large Stoke shift (≈100 nm) that enables to achieve highly efficient LEDs.
However, most reported low dimensional metal halides with STE luminescence only show visible region light emission, which is limited by their bandgap (mostly located in the ultra-violet region).It is attractive to expand their emission color to near infrared (NIR) region, as NIR emission shows a broad scope of applications, such as fiber-optic communication, quantum computing, and optical diagnosis. [6]There are only few reports about low dimensional metal halides with NIR STE luminescence.Among these reports, there are three strategies have been used: I) introducing bivalent Sn 2+ as metal site for perovskite structure; [7] II) incorporating rare earth elements; [8] and III) doping other metal ions (e.g., Cr 3+ ions). [9]However limited progress has been made due to Sn 2+ is unstable and easily to be oxidized. [10]5a] The Cr 3+ ions are easily to be oxidized to highly toxic Cr 6+ . [11]ere we propose a new approach to develop multifunctional low dimensional metal halides which show both thermochromic property and STE emission at NIR region.In this approach, as shown in Figure 1, we developed a mono organic cation (4,4-difluoropiperidine, DFPD) based zero dimensional (0D) (DFPD) 3 BiI 6 and a mixed organic-inorganic (DFPD-Cs) cation based 0D (DFPD) 2 CsBiI 6 metal halide crystals.In the two materials, [BiI 6 ] 3− octahedra are serving as active light absorption units, which ensure the bandgap of the formed 0D metal halides to be located at the visible region.They both show thermochromic properties that is, their colors change with temperature, which can be directly observed by naked eyes.By using the mixed organic and inorganic cations, we could not only improve the thermal stability of (DFPD) 2 CsBiI 6 , but also introduce luminescent STEs with emission at the deep-red region.Furthermore, we prepared Sb doped (DFPD) 2 CsBiI 6 crystals with STEs emission at the NIR region.The mechanism of the thermochromic property of the prepared crystals has been studied by combining experimental and theoretical methods, we found the lattice expansion is response for the color change at different temperatures.The STEs emission at NIR region has been studied by time-resolved spectroscopies and density functional theory (DFT) calculations.We found the doped Sb changes the in-direct band structure to a semidirect band structure, which makes the STEs emission to be possible.As a proof-of-concept, we also demonstrated the use of these inorganic-organic mixed cation-based crystals as active material for photodetection and light emitting device thereby emphasizing their utility and taking one step toward developing multifunctional low dimensional metal halides.

Results and Discussion
The (DFPD) 2 CsBiI 6 and (DFPD) 3 BiI 6 single crystals were synthesized by using a hydrothermal method, which produced needlelike single crystals which present as 0.1-cm-wide and 0.6-cmlong rods (see details in the Supporting Information).Their crystal structures were determined via single-crystal X-ray diffraction (SCXRD) analysis.The details of SCXRD data are summarized in Tables S1 and S2, Supporting Information.As shown in Figure 1, both (DFPD) 2 CsBiI 6 and (DFPD) 3 BiI 6 exhibit a 0D structure at the molecular structure level with spatially isolated octahedrons [BiI 6 ] 3− separated by large cations.One has to note that in (DFPD) 2 CsBiI 6 the [BiI 6 ] 3− octahedra are well separated by DFPD + and Cs + .Hence, it is still a 0D structure.In (DFPD) 2 CsBiI 6 , one can notice the strong interaction between the F atom (from the organic cation DFPD) and the Cs atom (inorganic cation): they form a Cs-F chemical bond with bond length of 3.37 Å.This is different from most reported mix organicinorganic cation-based metal halides, in which Cs is mostly encapsulated by the organic cations (e.g., amines). [12]To the best of our knowledge, such metal halides with chemical bond formed between organic and inorganic cations are rare. [13]Here, we are reporting one of the few examples.This approach opens up a new way to develop organic-inorganic cation mixed metal halides and tune their properties.The crystal structures are further confirmed by carrying out powder X-ray diffraction (PXRD).The measured PXRD patterns of both crystals are consistent with the calculated ones.This indicates the purity of the samples.The energy dispersive spectroscopy (EDS) elemental mapping of (DFPD) 2 CsBiI 6 single crystal is shown in Figure 1e together with the scanning electron microscope (SEM) image, wherein the homogeneous distribution of F, Cs, Bi, and I suggests the character of single-crystalline compounds and further confirms the purity of the prepared single crystals.
The optical properties of the prepared crystals were characterized by steady-state UV-vis absorption and photoluminescence (PL) spectroscopies.As shown in Figure 2a, the prepared (DFPD) 2 BiI 6 and (DFPD) 2 CsBiI 6 show broad absorption.Their bandgaps are estimated to be 2.12 and 2.01 eV for (DFPD) 3 BiI 6 and (DFPD) 2 CsBiI 6 , respectively based on Tauc plot (insert of Figure 2a).It is difficult to determine the band structure type (direct or indirect bandgap) only based on the absorption spectra.We will use DFT calculations to answer this in the coming section.The PL of (DFPD) 3 BiI 6 and (DFPD) 2 CsBiI 6 crystals were measured at both room temperature and cryogenic temperature (80 K, Figure 2b).(DFPD) 3 BiI 6 crystals do not show any observable PL at room temperature (Figure S1, Supporting Information).At 80 K, only a weak PL with a lifetime of about 4 ns in the range of 540-900 nm is observed for (DFPD) 3 BiI 6 crystals.This weak emission at low temperatures can be ascribed to the nonradiative recombination processes (e.g., phonon modes) are significantly suppressed.As the population of both low and high frequency phonon modes can be decreased at low temperatures.3a,14] (DFPD) 2 CsBiI 6 crystals only show weak PL in the range of 600-800 nm at room temperature (Figure 2b).The non or weak PL of the two crystals might be ascribed to their electronic band structure (indirect bandgap, see the section below), which is widely observed in Bi-based perovskite materials. [15]At 80 K, the PL of (DFPD) 2 CsBiI 6 crystals is enhanced without clear peak shift compared to that at room temperature.This indicates at the two different temperatures the PL processes are the same in terms of energy level.However, we observed the PL excitation (PLE) spectrum at 80 K is blue shifted (≈60 nm) compared to that at room temperature.The blue shift indicates the bandgap of the material might be changing with temperature.To confirm this, we measured the absorption spectrum at different temperatures (Figure 2c): with increasing temperature from 273 to 473 K, the absorption spectrum becomes broader, and the bandgap becomes smaller (Figure 2c insert).
To understand the mechanism for the temperature induced absorption, we measured the XRD patterns of (DFPD) 2 CsBiI 6 at different temperatures (Figure 2d).As temperature increases from 80 to 473 K, the XRD peaks shift to the lower angles, which indicate that the temperature causes thermal expansion.The thermal expansion is confirmed by extracting the unit cell volume and Bi-I bond length based on the XRD at different temperatures (Figure S2, Supporting Information).With increasing temperature, the unit cell volume becomes large and Bi-I bond becomes longer.As a result, the bandgap of the material might become lower at higher temperature, [16] which ensures the prepared (DFPD) 2 CsBiI 6 can be used as thermochromic materials.To demonstrate their thermochromic property, we took optical images of (DFPD) 2 CsBiI 6 at different temperatures (Figure 2e).The color of the (DFPD) 2 CsBiI 6 powder gradually changed from yellow to dark red with temperature from 88 to 473 K.A similar temperature dependent absorption/color change is observed for (DFPD) 3 BiI 6 (Figure S3, Supporting Information).The color change is reversible for (DFPD) 2 CsBiI 6 sample, but it is irreversible for (DFPD) 3 BiI 6 .The XRD results (Figure S4, Supporting Information) indicate an irreversible phase transition has happened for (DFPD) 3 BiI 6 at high temperature.The phase transition temperature can be observed from the TG-DSC curve (Figure S5, Supporting Information).Based on the above-mentioned results, we can conclude the mixed organic-inorganic cationbased metal halide ((DFPD) 2 CsBiI 6 ) shows improved thermal stability compared to mono organic cation-based metal halide ((DFPD) 3 BiI 6 ).Furthermore, the (DFPD) 2 CsBiI 6 crystals show high air and water stability, as shown in Figure S6, Supporting Information, there is no clear change in the XRD data comparing the as-prepared (DFPD) 2 CsBiI 6 single crystal and after 30 days of storage in air with 66% relative humidity.
The above-mentioned experimental results demonstrate the prepared (DFPD) 3 BiI 6 and (DFPD) 2 CsBiI 6 show thermochromic property (bandgap change) but negligible or weak PL.However, we are still lacking a clear understanding about the mechanism at microscopic molecular level.Here we use DFT method to calculate their electronic structure to explain these observations.As shown in Figure 3a,c, the calculated bandgap of (DFPD) 3 BiI 6 and (DFPD) 2 CsBiI 6 is 2.11 and 1.96 eV, which matches well with the experimental results (Figure 2a).This indicates our calculation method is reliable for studying the electronic structure of the systems.It is clear the two materials have an indirect bandgap structure, which explains their non/weak PL.Based on the projected density of state (PDOS, Figure 3b,d), the VBM and CBM are contributed mainly by Bi and I, which means the electronic transition at energy around bandgap is mainly involving the [BiI 6 ] 3− octahedra.Hence, the thermochromic property of the two materials is mainly caused by the structural change of the [BiI 6 ] 3− octahedra.To elucidate this, we used ab initio molecular dynamics (AIMD) to calculate the electronic structure and PDOS of (DFPD) 2 CsBiI 6 at different temperatures (300, 373, and 423 K).The stable structures at different temperatures were confirmed by calculating their free energy up to 15 ps by molecular dynamic simulations (Figure S7a, Supporting Information).Based on the PDOS (Figure S7b, Supporting Information), the CBM and VBM are contributed mainly from Bi and I at different temperatures.Furthermore, the bandgap becomes smaller as temperature increases.The AIMD theoretical calculation results match well with the XRD analysis at different temperatures: the thermal expansion (structural expansion) changes the crystal field strength around Bi 3+ , and as a result the change of the bandgap of the material is observed.This offers the foundation for using (DFPD) 3 BiI 6 and (DFPD) 2 CsBiI 6 as thermochromic materials for visual thermometers applications.
Except for the temperature dependent color change, we can also use PL to carry out remote temperature sensing.It has been widely observed that the PL spectrum and lifetime are temperature dependent. [17]Moreover, it is well-known that PL-based sensing is more sensitive than absorption (or color change)-based methods, because of the different ways of measuring them. [18]articularly, materials with NIR PL are attractive for sensing, as NIR light has longer penetration depth and less scattering compared to visible light.Here the prepared (DFPD) 2 CsBiI 6 show weak PL at the deep red region (Figure 2b).It is promising to develop NIR luminescent materials based on (DFPD) 2 CsBiI 6 .Herein, we use the ion doping approach to tune its photoluminescent to NIR.As ion doping had been proven to be an efficient way to tailor the PL properties of 0D metal halides. [19]Various of ions, such as Sb 3+ , Mn 2+ , Cu + have been successfully incorporated into the lattice of 0D metal halides. [20]12b 23] Furthermore, Sb 3+ has efficient radiative decay channels for PL.
We prepared Sb 3+ doped (DFPD) 3 BiI 6 and (DFPD) 2 CsBil 6 single crystals by mixing different amounts of Sb 2 O 3 , in the precur-sor solution for single crystal growth (details in Supporting Information).The molar percentage of Sb 3+ has been varied from 1% to 10% in the precursor solution.The formed single crystals were confirmed by XRD measurement (Figure 4a, Figure S9a, Supporting Information).With the increase of doping percentage of Sb 3+ , the XRD peaks shift to the larger angels, this indicated the compression of the crystal structure, which is due to the smaller size of Sb 3+ (1.28 Å) compared to Bi 3+ (1.63 Å).There is no obvious color change with the Sb doping, which can be seen from the absorption spectra (Figure 4b).In the PL spectra (Figure S9b, Supporting Information), we did not observe any clear PL for Sb 3+ doped (DFPD) 3 Bil 6 .Interestingly, we can observe a broadband STEs emission with emission peak at ≈850 nm for Sb 3+ doped (DFPD) 2 CsBil 6 (Figure 4c).It is clear the broadband STEs emission is located at the NIR region.The PL spectra were recorded only until 950 nm, due to the limited detection range of the detector we used.Hereafter, we only focus on the luminescent Sb 3+ doped (DFPD) 2 CsBil 6 samples.By comparing the PL intensity for different amounts of Sb doping, with 5% Sb in the precursor solution, the prepared crystals showed the highest PL.The measured PL quantum yield (PLQY) is about 1% (Table S4, Supporting Information).The reason for the sample, prepared with 5% Sb in the precursor solution, shows the highest PLQY is due to the [SbI 6 ] 3− octahedron can act as luminescent center as well as change the band structure of the material (see the section below).However, the further increase of Sb content will introduce concentration quenching effect. [24]As a result, we observed the decreased PLQY for 10% Sb sample.This phenomenon had been reported by several different groups. ,[14b, 24,25]  This explanation is supported by the time-resolved PL for the samples with different amounts of Sb doping (Figure S10, Supporting Information).As we can see 5% doping sample gives the longest PL decay lifetime (Table S4, Supporting Information).For this sample, the actual doping amount of Sb 3+ was identified by inductively coupled plasma mass (ICP-MS) spectroscopy.The obtained actual doping amount of Sb 3+ is 2%.Hereafter, we focus on the Sb 3+ doped (DFPD) 2 CsBiI 6 sample with actual Sb doping amount of 2% and name it as (DFPD) 2 CsBiI 6 :Sb in the following discussion.The EDS elemental mapping of (DFPD) 2 CsBiI 6 :Sb single crystal is shown in Figure S8, Supporting Information, wherein the homogeneous distribution of F, Cs, Bi, I, and Sb suggests the character of single-crystalline compounds and further confirms the purity of the prepared single crystals.The STEs emission of Sb doped samples is confirmed by low temperature PL measurement (Figure 4d).The PL spectrum at 80 K is much narrower than that at room temperature (Figure 4d).The temperature dependent full width at half maximum (FWHM) can be ascribed to the phase transition or strong electron-phonon coupling.We can exclude the former one, as we did not observe any new peaks from XRD measurement (Figure 2d).We would ascribe the strong electron-photon coupling to such FWHM change with temperature.The strong coupling also accounts for the low PLQY of the sample. [26]Meanwhile, we observe the PLE spectrum show a blue shift (≈50 nm) at 80 K (Figure 4d).This indicates the thermochromic property of (DFPD) 2 CsBil 6 is kept even with Sb doping.Moreover, the NIR PL is temperature dependent: as the temperature decreases from 300 to 80 K, the PL intensity increases (Figure S11, Supporting Information).This indicates the prepared (DFPD) 2 CsBiI 6 :Sb can act as active material for remote temperature sensing in its PL mode as well.
The NIR STEs emissions of the undoped and doped samples show relatively lower PLQY compared to other reported Sb doped 0D metal halides. [27]To understand the reason for this, we first carried out time-resolved PL measurement for both samples at room temperature and 80 K (Figure 4e,f).At both temperatures, the Sb doped sample ((DFPD) 2 CsBiI 6 :Sb) shows longer PL lifetime (Table S4, Supporting Information), this matches with its relatively higher PLQY.This might be ascribed to the triplet 3 P 1 states ( 3 P 1 → 1 S 0 transitions) of Sb 3+ , [5b,c,23a,28] which are introduced by Sb doping.At low temperature the PL becomes stronger and PL lifetime becomes longer than that at room temperature, this can be ascribed to several reasons, such as restricted phonon modes, structural change of the crystals, and decreased thermal energy for nonradiative recombination processes.To gain a clear picture about the PL enhancement with Sb doping, we use DFT method to calculate the electronic band structure, PDOS and charge distribution of VBM/CBM of both doped and undoped samples.Comparing the calculated electronic band structure (Figure 3a; Figure S12, Supporting Information) of undoped and doped samples, it is clear that the Sb dopant has changed the band structure from an indirect band gap ((DFPD) 2 CsBiI 6 ) to a semi direct band gap ((DFPD) 2 CsBiI 6 :Sb), in which direct and indirect gap values show small difference).The semi direct band gap of (DFPD) 2 CsBiI 6 :Sb is supported by the Tauc plot (Figure S14, Supporting Information).Because of the modification of the band structure, the CB and VB edges of (DFPD) 2 CsBiI 6 :Sb are less dispersive than that of (DFPD) 2 CsBiI 6 .The semi direct band gap structure offers higher possibility for the excitons to recombine radiatively, which explains the enhanced PL in Sb doped sample.Furthermore, the calculated band gap is narrowed to 1.78 eV after Sb doping, which matches with the experimental observation (1.81 eV, insert of Figure 4b).By analyzing PDOS and the charge distribution of VBM/CBM, we can see the Sb dopant contributes substantially to the VB and CB around the band gap transition.Especially, after Sb doping, the main contribution for VBM is changed from Bi(p) and I(p) orbital to Sb(p) and I(p) orbital, which means the VBM is localized on the [SbI 6 ] 3− octahedron.5b] As the energy at NIR region is rather low, the thermally activated vibrations could provide effective channels for the excitons to recombine non-radiatively and within a rather short timescale.This also explains their low PLQY.The multi-exponential decay of the PL (Table S4, Supporting Information) at room temperature supports such a hypothesis.
The time-dependent PL decay curve gives information about the STEs dynamic process at nanosecond timescale.26a,29] To gain insight about this information, we used ultrafast femtosecond transient absorption (fs-TA) spectroscopy to study the dynamics of STEs.The fs-TA spectra of (DFPD) 2 CsBiI 6 :Sb is shown in Figure 5a,b with distinct features: three negative peaks and multiple positive peaks in the broad spectral window.The three negative peaks located at about 430, 500, and 650 nm can be assigned as ground state bleach (GSB) signals of the excitonic states, as the peak positions match with the steady-state absorption spectrum of the sample (Figure 4b).There are three excited state absorption (ESA) signals with distinct peaks located at around 480, 570, and 700 nm.We can observe a broadband ESA at the spectral range of 700-1050 nm, which has been assigned to the character of STEs in metal-halide perovskites. [30]As shown in Figure 5b, apart from the decay of the signals' intensity, we can also see the spectral shift of the two ESA signals (at 570 and 700 nm) and GSB signals (at 500 and 650 nm).Such spectral shift has been widely observed in semiconductor materials including metal-halide perovskites and has been assigned to optical stark effect caused by exciton-exciton interaction. [31]All the GSB and ESA signals reach their absolute maximum within 200 fs (Figure 5c), we did not observe any clear state filling process during our experiment.This means immediately on excitation, the excitons relax to the band edge state within 200 fs after above bandgap photon (400 nm, 3.1 eV) excitation.Then the signals decay with multiple time-constants.The absolute amplitude of GSB signal at 430 and 650 nm show a slightly increase over time (Figure 5b).This might be due to the overlap between the GSB and ESA signal, and the positive ESA signal show spectral shift and decrease in amplitude with time.Overall, some amount of GSB and ESA signals remain at our maximum detection window (3 ns), which indicates there are excitons with long lifetime.This matches well with the observation from the time-resolved PL results, in which the PL lifetime of (DFPD) 2 CsBiI 6 :Sb is more than 20 ns.
This qualitative data analysis directly reveals information about the excited-state dynamics in (DFPD) 2 CsBiI 6 :Sb, but does not provide a quantitative relaxation model.For this purpose, our strategy is to carry out global fitting analysis by a sum-ofexponential-decays model fit to the data (details in Supporting Information).We get these spectra from the minimally complex representation of multi-exponential decays: the so-called decay associated spectra (DAS), as shown in Figure 5d.The first component with a lifetime of 1 ps can be assigned to the band edge state, which shows GSB peaks in agreement with the steady-state absorption spectrum of the sample.Then it decays to two distinct states with a lifetime of 12 ps and long-lived state with lifetime longer than the detection time-window (3 ns).The component with 12.5 ps lifetime shows a broadband ESA signal, which indicates it is the STEs state, which might relax to the emissive long-live state and/or back to the ground state non-radiatively.This agrees with the low PLQY (≈1%) of the sample, as the STEs state can have efficient non-radiative channel.Here the 12.5 ps might be involving spin flipping process from singlet STEs to triplet STEs, as the timescale matches with most reported intersystem crossing timescale in Sb doped metal-halide perovskites and molecular materials. [32]Based on the global kinetic analysis, we can construct a quantitative model to support the abovementioned qualitative analysis.Figure 5e shows the extracted relaxation scheme with decay parameters by global kinetic fitting of the experimental data.
The organic-inorganic mixed cation-based metal-halide perovskites (compared to their full inorganic counter parts, such as Cs 3 Bi 2 I 9 and Cs 3 BiI 6 ) have improved solution processability for film deposition, specifically, they can be more easily dissolved in organic solvents (e.g., dimethylformamide, dimethyl sulfoxide, methanol) and beneficial for optoelectronic devices' fabrication, such as solar cells and photodetectors.Herein, we tested its processability for film deposition (details in Supporting Information).Briefly, we dissolved the prepared crystals in organic solvent (N,N-dimethylacetamide), then spin coated the precursor solution on glass substrates.We measured the XRD patterns, absorption, PL, and PLE spectra of the film samples and compared them with the crystals/powder samples (Figure S15, Supporting Information).As we can see, films and crystals have the same crystal structure and optical property.The thermochromic property is retained in the prepared film samples (Figure S15b, Supporting Information insert) as well.
Considering the above-mentioned optical properties of the prepared (DFPD) 2 CsBiI 6 :Sb material, here we use it to fabricate photodetectors based on the (DFPD) 2 CsBiI 6 :Sb film.We deposited the film on a planar glass substrate with two Ti/Au electrodes (with 5 μm channel width and 1 mm length, see Figure 6a).The response of the detector was tested at illumination under a 30 mW cm −2 xenon lamp, we can see the photodetector show a good response to light based on the measured currentvoltage curve (Figure S16).As we discussed above, the films have the thermochromic property.We heated up the photodetector to different temperatures, and the responsivity (photocurrent) of the detector increases with increasing of the temperature (Figure 6c,d).This can be ascribed to the synergy effects: with increasing temperature the band gap decreases and the mobility of the charge carrier increases. [33]Then we tested the excitation wavelength-dependent photo responsivity, we found the photodetectors show the highest responsivity at 365 nm among three test wavelengths (365, 440, and 520 nm, Figure 6e).This indicates the charge carries generated by high energy photons tend to contribute more current than that generated by low energy photons in the circuit.Such interesting observation is opposite of the PL process, as shown in Figure 4d, the charge carries generated by low energy photons give more PL than that generated by high energy photons.This might be ascribed to the selection rule for the photoexcitation and STEs formation.It is well known that STEs have rather low charge carrier mobility, which is not favorable for photocurrent generation. [34]We further measured the response speed of the photodetectors at 365 nm light with intensity of 111.2 mW cm −2 .The response speed of the detectors is rather slow (at both 3 and 5 V) with a rise time (photocurrent changing from 0% to 70% of the maximum value) of 110 ms and decay time of 162 ms (Figure 6f and Figure S17, Supporting Information).The speed is probably limited by the low charge carrier mobility of the 0D perovskites.It is possible to increase the speed by reducing the area of the deposited (DFPD) 2 CsBiI 6 :Sb film and optimizing the circuit, as the response time is related to the RC time constant of the circuit. [35]However, it is out of the scope of our current work.
To show the multifunction of the developed material, we used the (DFPD) 2 CsBiI 6 :Sb film for NIR imaging.We fabricated a NIR LED by coating a thin layer of (DFPD) 2 CsBiI 6 :Sb film on the surface of a commercially available LED lamp (with emission at 600-605 nm).The fabricated NIR LED shows luminescence at the NIR region, which matches well with the PL spectra of (DFPD) 2 CsBiI 6 :Sb film and powder (Figure 4c; Figures S15c and  S19a, Supporting Information).We took photographs of a bonsai under Xe lamp and our NIR LED with a CCD camera (spectra sensitive in 700-1100 nm).When the fabricated NIR LED device is switched off, the NIR camera fails to capture any image (Figure S19b, Supporting Information).However, upon turning on the NIR LED, the NIR camera can effectively capture the outline of the bonsai under the NIR LED light (Figure S19b, Supporting Information).These findings demonstrate the practicality of (DFPD) 2 CsBiI 6 :Sb for NIR imaging and night vision purposes.

Conclusion
In conclusion, we developed a new type of 0D metal-halide materials which show thermochromic property for temperature sensing and NIR PL.The mixed organic-inorganic cation approach ensures the good solution processibility of the prepared (DFPD) 2 CsBiI 6 and its Sb doped materials.By combining experimental and theoretical approach we find the mechanism of the thermochromic behavior of (DFPD) 3 BiI 6 and (DFPD) 2 CsBiI 6 is a result of the thermal caused structural expansion and changes of their bandgaps.Furthermore, we turned on the NIR PL of (DFPD) 2 CsBiI 6 by Sb doping.The PL turn-on is due to changing the bandgap structure from indirect bandgap to semi-indirect bandgap.We used ultrafast fs-TA to find the reason for the rather low PLQY of (DFPD) 2 CsBiI 6 :Sb, in which a non-emissive state is account for this.We further demonstrated the multifunctional application of (DFPD) 2 CsBiI 6 :Sb as active material for photodetection and NIR LEDs.Our findings open new avenues for exploring organic-inorganic cations of low-dimensional metal halides to endow new structures, and hence enable the development of new compositions with multifunction for optoelectronics.

Figure 5 .
Figure 5. fs-TA spectra of (DFPD) 2 CsBiI 6 :Sb.a) Wavelength versus probe-delay map of the fs-TA experiment after excitation at 400 nm.The total timerange is 3 ns.Note the change from linear to logarithmic scale at 10 ps.b) Transient spectra extracted from panel a) over time-ranges of 0 ps to 3 ns.c) Kinetic traces extracted at representative detection wavelengths (black circle: 475 nm; orange circle: 500 nm, purple circle: 1000 nm, signal multiplied by a factor of 4) from the fs-TA experiment.Data overlaid by the corresponding kinetic trace extracted from the global fit (black, orange, and purple solid lines).Note the change from linear to logarithmic scale at 2 ps.d) Decay associated spectra (DAS) extracted from global kinetic analysis of fs-TA spectra.e) Schematic illustration of the photoinduced dynamics in (DFPD) 2 CsBiI 6 :Sb.HES: high energy state; BES: band edge state; NES: non-emissive state; ES: emissive state; GS: ground state.

Figure 6 .
Figure 6.(DFPD) 2 CsBiI 6 :Sb thin film and its photodetector device.a) Schematic of (DFPD) 2 CsBiI 6 :Sb based photodetector.b) Current-voltage curves of the device measure in the dark and the illumination of 365 nm LED light at different light fluences.c) Temporal photocurrent response of the device measured in the illumination under a 111.2 mW cm −2 , 365 nm LED light at different temperatures.d) Current-voltage curves of the device measured in the illumination under a 111.2 mW cm −2 , 365 nm LED light at different temperatures.e) Temporal photocurrent response of the device at different wavelengths.f) Temporal photocurrent response of the device for a wavelength of 365 nm (at 3 V bias).