ARTICLE

Ultrafast transient infrared spectroscopy forprobing trapping states in hybrid perovskite filmsAhmed M. El-Zohry1,2✉, Bekir Turedi3, Abdullah Alsalloum3, Partha Maity1, Osman M. Bakr3, Boon S. Ooi4 &

Omar F. Mohammed1✉

Studying the charge dynamics of perovskite materials is a crucial step to understand the

outstanding performance of these materials in various fields. Herein, we utilize transient

absorption in the mid-infrared region, where solely electron signatures in the conduction

bands are monitored without external contributions from other dynamical species. Within the

measured range of 4000 nm to 6000 nm (2500–1666 cm−1), the recombination and the

trapping processes of the excited carriers could be easily monitored. Moreover, we reveal

that within this spectral region the trapping process could be distinguished from recombi-

nation process, in which the iodide-based films show more tendencies to trap the excited

electrons in comparison to the bromide-based derivatives. The trapping process was

assigned due to the emission released in the mid-infrared region, while the traditional band-

gap recombination process did not show such process. Various parameters have been tested

such as film composition, excitation dependence and the probing wavelength. This study

opens new frontiers for the transient mid-infrared absorption to assign the trapping process

in perovskite films both qualitatively and quantitatively, along with the potential applications

of perovskite films in the mid-IR region.

https://doi.org/10.1038/s42004-022-00683-7 OPEN

1 Division of Physical Sciences and Engineering, King Abdullah University of Science and Technology (KAUST), Thuwal 23955-6900, Saudi Arabia.2 Department of Physics, AlbaNova Center, Stockholm University, 10691 Stockholm, Sweden. 3 KAUST Catalysis Center, King Abdullah University of Scienceand Technology (KAUST), Thuwal 23955-6900, Saudi Arabia. 4 Photonics Laboratory, King Abdullah University of Science and Technology (KAUST),Thuwal 23955-6900, Saudi Arabia. ✉email: ahmed.elzohry@fysik.su.se; omar.abdelsaboor@kaust.edu.sa

COMMUNICATIONS CHEMISTRY | (2022) 5:67 | https://doi.org/10.1038/s42004-022-00683-7 | www.nature.com/commschem 1

1234

5678

90():,;

Hybrid perovskite materials have recently attracted lots ofattention due to their unique photo-physical propertiesand their high performances in various applications such

as solar cells and light emitting diodes1–17. However, still con-trolling the amount of traps present in these materials especiallyupon making thin films is a challenging procedure18,19. Thepresence of trap states either through structural defects or othertypes can quench the charge carriers motilities inside the mate-rials and thus reducing both the device’s performance and itsstability8,18,20–23. Various direct and indirect methods have beenapplied to track and quantify trap states such as electrical oroptical measurements, however, most of these still have somedrawbacks8,18–22,24. For instance, conductivity measurementshave low time resolutions and can’t distinguish between variouscarriers such as electrons and holes especially upon having closemotilities25. Also, the most commonly used optical measurementssuch as time-resolved photoluminescence or transient absorptionin the visible range couldn’t afford direct spectral signatures fortrapping states except providing variations of multi-exponentialkinetic rates between different samples according to the estimatedtraps present in the investigated samples1,2,7,13,15,26,27. Thus, stillthere is a need for a direct transient optical method to track andquantify the trapping process in perovskite materials, and cor-relate their presence directly to charge dynamics. For that sake,we utilized mid-infrared (mid-IR) probe to monitor the chargedynamics of four different perovskite films with different com-positions. Following the charge dynamics using femtosecondtransient absorption (fs-TA) in the mid-IR has been used pre-viously for several systems including metal complexes28–31,organic dyes32–37, metals38, and semiconductors21,25,39–43.

Basically for classical semiconductors, the electron’s absorptionin the conduction band with high density of states has a broadspectral signature extending from 3333 nm (3000 cm−1) to11,111 nm (900 cm−1), in which other contributions from catio-nic or anionic molecular species present can be easilyquantified25,28,32,33. The positive signature in the mid-IR uponphoton excitation is ascribed to the presence of intra-transition offree electrons in/into the conduction band of the semiconductorused25,43,44. The transient mid-IR was used to follow trappedelectrons at the mid-gap shallow states in the platinized TiO2

system, in which IR emission is evolved in the IR region as aresult of electron trapping process25,43. Recent mid-IR studieshave been done on perovskite materials, however in those studiesthe authors focused on following the NH vibrational modes,present in the organic cationic part of the perovskites20,21,26,45. Incontrast, in our selected mid-IR region, we don’t have any con-tribution of the vibrational modes of the organic part, onlytransient signal of electrons in the conduction band, see Fig. S1.

Herein, we propose using fs-TA in the mid-IR region as a sen-sitive tool to follow the presence of traps in hybrid perovskite films.In the current study, various hybrid perovskite films have beensynthesized and utilized to study the charge dynamics in the mid-IRregion extending from ca. 4000–6000 nm (2500–1666 cm−1). Wefound that perovskite films with methyl-ammonium cation (MA)and halides of Iodide (I) derivatives tend to emit mid-IR than otherfilms with formamidinium (FA) and bromide (Br) derivatives,depending on the working conditions and the quality of the pre-pared films (see steady state measurements in Figs. S2–4). Thisstudy presents deeper understandings of charge dynamics in per-ovskite films, and potential applications of perovskite films in themid-IR region.

Results and discussionFigure 1A shows the false 2-D plot of fs-TA in the mid-IR regionwith a central detection window of 5000 nm (2000 cm−1) for

MAPbBr3 thin film using an excitation wavelength of 530 nm. Attime zero, an intense positive signal appears due to the populationof electrons in the conduction band as described for previoussemiconductors such as TiO2

28,32,43,46. The contribution of holesin the valence band is expected to be minimum due the lowenergy of the probed light in the mid-IR range, however, furtherstudies are needed to certify that. This positive signal decaysexponentially toward zero within few nanoseconds. However, theextracted spectra show negative features at longer time scale>1.0 ns; see Fig. 1B. The extracted kinetic trace at 4900 nm showsa multi-exponential decay for the positive signal with an averagelifetime of 120 ps, followed by a small negative feature beyond2 ns; see Fig. 1C.

The same film (MAPbBr3) was measured by fs-TA in thevisible range, and an extracted kinetic trace at 550 nm corre-sponding to the ground state bleach (GSA) is compared with theextracted kinetic trace at 4900 nm from the mid-IR region; seeFig. 1D and Fig. S5. The comparison shows that both normalizedkinetic traces from different spectral regions are very similar,except the presence of a new feature at the extracted kinetic tracefrom the mid-IR range; see Fig. 1C–D. The similarity between thetwo kinetic traces for MAPbBr3 film confirms the validity of mid-IR signal to trace charge dynamics in perovskite films. However,the charge dynamics in the mid-IR region are not similar. Forexample, negative features at the red-part of the false 2D plot (at4900 nm) in Fig. 1A, appears differently than in the blue-part (at5120 nm). Extracting a kinetic trace at 5120 nm (1953 cm−1)shows earlier conversion of positive to negative signals at ca.200 ps; see Fig. 1C. Also, comparing this kinetic trace with the oneextracted from the GSB in the visible region, shows differentbehavior than kinetic trace at 4900 nm; see Fig. 1D. This high-lights the dependence of such negative feature on the probedspectral window.

Upon measuring the iodide-derivative, MAPbI3 film, adetectable fs-TA mid-IR signal was also found, but with differentbehavior, see Fig. 2A, B. Interestingly, the transient mid-IR signalfor MAPbI3 film was changing over minutes time scale (min-utes); see Fig. 2C. Thus, various kinetic traces were extracted atdifferent times and compared together at the same probedwavelength. For instance, the fresh-irradiated film (~ 0 min.),appositive signal was measured until ca. 100 ps, and then a smallnegative signal started to emerge. The disappearance of thepositive signal became faster with the longer the exposure pro-cess associated with an increase of the negative signal at earlytimes; see Fig. 2C. For example, after 26 min of irradiation, thepositive signal converted into a negative signal within 10 ps; seeFig. 2C. Interestingly, this process is reversible, in whichswitching off the irradiation for almost 8 min, and re-measurethe dynamics again at the same irradiated spot (@ ~ 34 min inthe Fig. 2C), the dynamics slowly started to be similar to the10 min irradiation measurements; see Fig. 2C. In all cases, afterthe appearance of the negative signal, it decays later on to zerodue to the expected recombination process, see Fig. 2C.

Upon measuring the same film in the visible range, a strongGSB signal at ca. 760 nm was observed overlapping with an ESAspectra extending from 650 to 850 nm, see Fig. S7. However, nounique change in dynamics has been observed in the visible rangesimilar to the shown data in the mid-IR range. And upon com-paring the extracted normalized visible kinetic trace at 760 nmwith the ones from the mid-IR, it is evident that the visible kinetictrace is similar to the one from mid-IR at 26 min only at earlytimes; see Fig. 2D. However, still at later times, the mid-IR signalswitches its sign, but not the visible kinetic trace at 760 nm. Thechange of charge dynamics upon irradiation in the mid-IR regionhas been assigned in iodide-rich perovskite materials to halide-defects assisted by the low energy needed for defects

ARTICLE COMMUNICATIONS CHEMISTRY | https://doi.org/10.1038/s42004-022-00683-7

2 COMMUNICATIONS CHEMISTRY | (2022) 5:67 | https://doi.org/10.1038/s42004-022-00683-7 | www.nature.com/commschem

formation18,47–49. This comparison highlights also that thisappearance of negative features in the mid-IR region is a dyna-mical process and also can be reversible (MAPbI3 case).

The extracted kinetic trace at 4900 nm from the MAPbI3 showsstronger and earlier negative signal than the one shown in MAPbBr3case. Moreover, for further confirmation of the working conditions,a reference silicon wafer substrate was measured under the sameprocedure, and no negative features have been detected, see Fig. S8.Thus, we assign these strong negative signatures in the mid-IR forMAPbI3 film to the emissive trapping process near the VB of per-ovskites, confirming previous studies about tendencies of iodide-based perovskites to form more trap states than Br ones1,8,15,18.

To scrutinize our interpretation about the sensitivity of mid-IRtoward trapping process, we performed the same measurementson other perovskite films including FAPbBr3, FAPbI3 and mix-ture of their halides. For the FAPbBr3 film, the mid-IR signalshows primarily a strong negative signal close to time zero con-verting into a positive signal with a lifetime of ca. 60 fs, which hasbeen assigned to the exciton thermalization/dissociation process;7

see Fig. 3A, B. Previously, in the MAPbBr3 film, exciton bindingenergy seems to be smaller, thus, no detection of exciton dis-sociation process could be seen; see Fig. 1. The charge recombi-nation in the FAPbBr3 film (decay of the TA signal) has beenfitted with multi exponential behavior, giving an average lifetimeis about 10 ps, see Fig. 3B.

For the FAPbI3 film, similar observation was estimated for thelifetime needed of exciton dissociation in FAPbI3 film; see Fig. 3C, D.

However, instead of charge recombination, the measured posi-tive signal converted again to a negative signal due to a trappingprocess of time component ca. 8.5 ps; see Fig. 3D. Then the trappedelectrons recombine slowly with a lifetime higher than 1 ns. It isclear now that this negative signature in the mid-IR range is asso-ciated with the iodide derivative of hybrid perovskite films.

To verify the role of iodide anion for the formation of emissivetrapping centers, we also synthesized other set of various per-ovskite films of different ratios between the iodide and Br halides(MAPbInBr3-n) to investigate the effect of doping with iodide ionson the presence of such negative signal. Figure 4A shows thekinetic traces of mid-IR signals for four films of MAPbInBr3-n, inwhich n varies from 0 to 3. The extracted kinetics at 4900 nmshow the appearance of negative signals at different times rangingfrom 100 ps to 1 ns depending on the amount of iodide halidepresent, in which higher content of iodide shows faster appear-ance of negative mid-IR signal, showing that the iodide content inperovskite films controls the appearance of the negative signal(emission of mid-IR).

To study the dependence of appearance of these negative sig-nals on the energy of the mid-IR probe, the MAPbI3 film isexcited at 520 nm (300 µW), and probed at various mid-IR energyranging from 4000 nm to 6000 nm, as shown in Fig. 4B. Asexpected, the appearance time of the negative signal depends onthe utilized mid-IR energy, in which the switching points frompositive to negative signal happen at ca. 500 ps when using4000 nm, and at ca. 20 ps upon using 6000 nm probes.

Fig. 1 Transient mid-IR data for MAPbBr3 film. (A) 2D-false color plot of fs-transient absorption in the mid-infrared regions for MAPbBr3 film using530 nm as an excitation source. B Extracted spectra for transient spectra in mid-infrared region, the spectra are corrected for the wavelength scale.C Extracted kinetic trace at 4900 and 5120 nm with corresponding fitting, green shaded area highlights the appearance of negative signal. D Comparisonbetween normalized kinetic traces in the infrared and visible ranges for MAPbBr3 film.

COMMUNICATIONS CHEMISTRY | https://doi.org/10.1038/s42004-022-00683-7 ARTICLE

COMMUNICATIONS CHEMISTRY | (2022) 5:67 | https://doi.org/10.1038/s42004-022-00683-7 | www.nature.com/commschem 3

Furthermore, by changing the excitation wavelength from410 nm to 520 nm for another MAPbI3 film, the extracted kinetictraces at 4500 nm show a dependence of the kinetic decay withthe wavelength used; see Fig. 4C. The appearance of negativesignal is faster upon using 410 nm then became slower with440 nm, and 520 nm respectively. Upon using the excitation lightat 520 nm, the kinetic trace decay to zero with no signature ofnegative signal, despite that higher power was used, almost 20times higher (300 µW) than at 410 nm; see Fig. 4C. This excita-tion energy dependence shows that the higher the electron can bepromoted in the excited state, the higher chances to be trapped inemissive centers. Apparently, these emissive centers are formedduring the excitation process of perovskite films.

From the above mid-IR transient measurements, the followingmechanism can be drawn, see Fig. 5. Upon exciting the perovskitefilm, electrons in the CB should absorb the following mid-IRprobe to populate various vibrational levels in the excited state,giving a positive transient absorption signal. Due to the presenceof trap states within the bandgap of the perovskite film, electronsin the CB decay non-radiatively to the ground state (channel 1 inFig. 5), in which the transient signal decays to zero as shown inFAPbBr3, Fig. 3. Interestingly, upon synthesizing perovskite filmsusing chemical species such as MA+ or I− or both of them as inMAPbI3, negative transient signal in the mid-IR region starts toappear. And since perovskite films have no characteristic featuresin this mid-IR region, this negative transient signal can be onlydue to emission of mid-IR signal or mid-IR gain (channel 2 in

Fig. 5). It has been already established in the literature that MA/iodide perovskite films show more potential to form trap statesthan other derivatives (FA/Br)1,15. In addition, these mid-IR gaincan be controlled by the incident excitation energy, see Fig. 4C.Thus, we postulate that in MAPbI3 films, different kind ofemissive trap states are additionally formed related to ionmigration and formation of transient phonon modes that are notpresent for instance in the FAPbBr3 films1,15,50.

Interestingly, to detect these mid-IR emissive states, suitableprobe energy should be utilized as shown in Fig. 4B. The probeenergy data presents the influence of the energy carried by theprobing photons to free/decouple the trapped electrons if suffi-cient energy is present. For instance, the probed pulse at 4000 nm(2500 cm−1) can liberate the trapped electron more efficient thanat 6000 nm (1666 cm−1), and the appearance of mid-IR negativesignal upon using 4000 nm (2500 cm−1) will not appear early,until 500 ps. In the same way, the appearance of negative signal at6000 nm (1666 cm−1) is much faster, due to the low energycarried by the probe pulse to liberate the trapped electron in theemissive states, instead allowing for deactivation through the IRstimulated emission. These observations are consistent with apreviously proposed mechanism that trapped electrons can beexcited thermally if the energy difference between the trap stateand the CB is small, <50 meV25. This also illustrates the incap-ability of transient absorption in the visible region to detect such atrapping process due to the higher energy carried by the visibleprobed light, that have the potential to liberate the trapped

Fig. 2 Transient mid-IR data for MAPbI3 film. (A) 2D-false color plot for fs-transient absorption in the mid-infrared regions for MAPbI3 film using 530 nmas an excitation source after 26minutes of irradiation. B Extracted spectra for transient spectra in mid-infrared region (A), the spectra are corrected for thewavelength scale. C Extracted kinetic trace at 4900 nm at different irradiation time shown in minutes for MAPbI3 film, using 520 nm with excitation powerof 500 µW. D Comparison between normalized kinetic traces in the infrared and visible ranges for MAPbI3 film.

ARTICLE COMMUNICATIONS CHEMISTRY | https://doi.org/10.1038/s42004-022-00683-7

4 COMMUNICATIONS CHEMISTRY | (2022) 5:67 | https://doi.org/10.1038/s42004-022-00683-7 | www.nature.com/commschem

carriers into higher excited state, producing undistinguishablesignal for the trapping process in the visible region, in which theexcited electrons can be only deactivated via non-emissive trap-ping centers. According to the current range of probed energyutilized, it is expected that the high energy probe >1 eV will leadto deactivation through non-radiative centers, while <0.3 eV willstimulate mid-IR emission, see Fig. 5.

Moreover, upon using higher band-gap excitations such as410 nm, the probability of electron trapping in these emissivestates is increased despite the excitation intensity used, matchingwith the expected distribution for the states of trap-density pre-sent. We also show that continuous irradiation at high excitationenergy for the MAPbI3 film (Fig. 2C) increases the rate of trap-ping (channel 2 in Fig. 5), as well as the intensity of the transient

Fig. 3 Transient mid-IR data for MAPb(Br3)/I3 films. (A) 2D-false color plot for fs-transient absorption in the mid-infrared regions for FAPbBr3 film using530 nm as an excitation source. B Extracted kinetic trace at 4880 nm for FAPbBr3 film. C 2D-false color plot for fs-transient absorption in the mid-infraredregions for FAPbI3 film using 530 nm as an excitation source. D Extracted kinetic trace at 4960 nm for FAPbI3 film.

Fig. 4 Dependence of various factors on the observed transient mid-IR signal. Dependence on chemical composition: (A) Normalized kinetic traces at4900 nm extracted form MAPbInBr3-n films using 520 nm as excitation showing the appearance of negative signals. Dependence on mid-IR detectionwindow: (B) Normalized Kinetic traces for MAPbI3 film under excitation of 520 nm and at different probing wavelengths showing the appearance ofnegative signals. Dependence on excitation wavelengths: (C) Normalized kinetic traces for MAPbI3 film under various excitation wavelengths showing theappearance of negative signals.

COMMUNICATIONS CHEMISTRY | https://doi.org/10.1038/s42004-022-00683-7 ARTICLE

COMMUNICATIONS CHEMISTRY | (2022) 5:67 | https://doi.org/10.1038/s42004-022-00683-7 | www.nature.com/commschem 5

signal. This indicates toward the exciting correlation between thelight irradiation and ion migration process in perovskite films51.This means that changes in the perovskite lattice by the incidentlight (depending on energy) can lead to the formation of theseemissive trapping states at different energy levels above the VB.

ConclusionWe show herein for the first time that transient absorption in themid-IR region is a suitable spectroscopic tool to explore trappingprocess in addition to follow the behavior of free electrons in theconduction band without other contributions of other species(reduced/oxidized species). Significantly, the selected region of themid-IR spectrum can be utilized to follow the actual trapping pro-cess of electrons by detecting the negative appearance of the mid-IRtransient signal, which is likely due to the formation of emissive trapstates. We also figure out that similar measurements in the visibleregion could not be monitored due the energy of the probed lightthat can liberate the trapped carriers into higher excitation levels,providing additional complexity to distinct between tarped carriersand other species such as excitons and free carriers absorption.Interestingly, these emissive trap states in the mid-IR can be con-trolled by film quality, film chemical composition, and utilized band-gap excitation energy. This work will open frontiers toward under-standing and controlling the nature of trapping centers in perovskitefilms, along with the potential of iodide-based perovskite films togenerate emission in the mid-IR range.

MethodsFilm preparations. Preparing the perovskite films: the CaF2 substrates werecleaned with DI water, acetone, and IPA, followed by a 10 min UV ozone treat-ment. FAPbBr3 and FAPbI3 thin films were prepared using a modified antisolventdripping technique52. 1.1 M of FAX and PbX2 were dissolved in DMF:DMSO (9:1ratio), and 100 µl of the solution was spin-coated for 15 s at 4000 rpms. 300 µl oftoluene was drop-casted during the sixth second, and the films were annealedimmediately after the spin-coating process for 10 min at 170° for FAPbI3 and 100°for FAPbBr3.

Femtosecond transient absorption setup. Briefly, an excitation wavelength of410–520 nm was utilized, while the probe light was also changing from 4000 nm to6000 nm. The power used for the excitation wavelength depends on the utilizedwavelength but was changing from ca. 100–300 μW28,37,53–55. The mid-IR light wasdetected on a N2-cooled CCD that is sensitive to mid-IR photons.

Data availabilityAll relevant data are available from the corresponding authors upon request (A.M.El-Z.& O.F.M.).

Received: 31 January 2022; Accepted: 9 May 2022;

References1. Alarousu, E. et al. Ultralong Radiative States in Hybrid Perovskite Crystals:

Compositions for Submillimeter Diffusion Lengths. JPCL 8, 4386–4390 (2017).2. Guo, Z. et al. Long-Range Hot-Carrier Transport in Hybrid Perovskites

Visualized by Ultrafast Microscopy. Science 356, 59–62 (2017).3. Zuo, Z. Y. et al. Enhanced Optoelectronic Performance on the (110) Lattice

Plane of an MAPbBr3 Single Crystal. J. Phys. Chem. Lett. 8, 684–689 (2017).4. Bi, D. et al. Efficient Luminescent Solar Cells Based on Tailored Mixed-Cation

Perovskites. Sci. Adv. 2, 1–7 (2016).5. Kadro, J. M. et al. Proof-of-Concept for Facile Perovskite Solar Cell Recycling.

Energ. Environ. Sci. 9, 3172–3179 (2016).6. Colella, S., Mazzeo, M., Rizzo, A., Gigli, G. & Listorti, A. The Bright Side of

Perovskites. JPCL 7, 4322–4334 (2016).7. Zhai, Y. X., Sheng, C. X., Zhang, C. & Vardeny, Z. V. Ultrafast Spectroscopy of

Photoexcitations in Organometal Trihalide Perovskites. Adv. Funct. Mater. 26,1617–1627 (2016).

8. Shi, D. et al. Low Trap-State Density and Long Carrier Diffusion inOrganolead Trihalide Perovskite Single Crystals. Science 347, 519–522 (2015).

9. Huang, J., Shao, Y. & Dong, Q. Organometal Trihalide Perovskite SingleCrystals: A Next Wave of Materials for 25% Efficiency Photovoltaics andApplications Beyond? JPCL 6, 3218–3227 (2015).

10. Fang, H. H. et al. Photophysics of Organic-Inorganic Hybrid Lead IodidePerovskite Single Crystals. Adv. Funct. Mater. 25, 2378–2385 (2015).

11. Green, M. A. & Bein, T. Photovoltaics Perovskite Cells Charge Forward. Nat.Mater. 14, 559–561 (2015).

12. Gao, P., Gratzel, M. & Nazeeruddin, M. K. Organohalide Lead Perovskites forPhotovoltaic Applications. Energ. Environ. Sci. 7, 2448–2463 (2014).

13. Stranks, S. D. et al. Electron-Hole Diffusion Lengths Exceeding 1 Micrometer inan Organometal Trihalide Perovskite Absorber. Science 342, 341–344 (2013).

14. Jena, A. K., Kulkarni, A. & Miyasaka, T. Halide Perovskite Photovoltaics:Background, Status, and Future Prospects. Chem. Rev. 119, 3036–3103 (2019).

15. Sarmah, S. P. et al. Double Charged Surface Layers in Lead Halide PerovskiteCrystals. Nano Lett. 17, 2021–2027 (2017).

16. Pan, J. et al. Halogen Vacancies Enable Ligand‐Assisted Self‐Assembly ofPerovskite Quantum Dots into Nanowires. Angew. Chem. 131, 16223–16227(2019).

17. Zhumekenov, A. A. et al. Reduced Ion Migration and EnhancedPhotoresponse in Cuboid Crystals of Methylammonium Lead IodidePerovskite. J. Phys. D: Appl. Phys. 52, 054001 (2018).

18. Meggiolaro, D. et al. Iodine Chemistry Determines the Defect Tolerance ofLead-Halide Perovskites. Energ. Environ. Sci. 11, 702–713 (2018).

19. Fang, H.-H. et al. Ultrahigh Sensitivity of Methylammonium Lead TribromidePerovskite Single Crystals to Environmental Gases. Sci. Adv. 2, 1–9 (2016).

20. Munson, K. T., Kennehan, E. R., Doucette, G. S. & Asbury, J. B. DynamicDisorder Dominates Delocalization, Transport, and Recombination in HalidePerovskites. Chem 4, 2826–2843 (2018).

21. Munson, K. T., Grieco, C., Kennehan, E. R., Stewart, R. J. & Asbury, J. B.Time-Resolved Infrared Spectroscopy Directly Probes Free and TrappedCarriers in Organo-Halide Perovskites. ACS Energy Lett. 2, 651–658 (2017).

22. Leblebici, S. Y. et al. Facet-Dependent Photovoltaic Efficiency Variations inSingle Grains of Hybrid Halide Perovskite. Nat. Energ. 1, 1–7 (2016).

23. Krückemeier, L., Krogmeier, B., Liu, Z., Rau, U., Kirchartz, T. UnderstandingTransient Photoluminescence in Halide Perovskite Layer Stacks and SolarCells. Adv. Energy Mater. 11, 2003489 (2021).

24. Simpson, M. J., Doughty, B., Yang, B., Xiao, K. & Ma, Y.-Z. Imaging ElectronicTrap States in Perovskite Thin Films with Combined Fluorescence andFemtosecond Transient Absorption Microscopy. JPCL 7, 1725–1731 (2016).

25. Yamakata, A., Ishibashi, T. & Onishi, H. Time-Resolved Infrared AbsorptionSpectroscopy of Photogenerated Electrons in Platinized TiO2 Particles. Chem.Phys. Lett. 333, 271–277 (2001).

26. Guo, P. J. et al. Infrared-Pump Electronic-Probe of Methylammonium LeadIodide Reveals Electronically Decoupled Organic and Inorganic Sublattices.Nat. Commun. 10, 1–8 (2019).

27. Yin, J. et al. Point Defects and Green Emission in Zero-DimensionalPerovskites. J. Phys. Chem. Lett. 9, 5490–5495 (2018).

28. Abdellah, M. et al. Time-Resolved Ir Spectroscopy Reveals a Mechanism withTiO2 as a Reversible Electron Acceptor in a TiO2–Re Catalyst System for CO2

Photoreduction. J. Am. Chem. Soc. 139, 1226–1232 (2017).

Fig. 5 The overall scheme. Schematic representation for the mechanism ofcharge recombination and charge trapping in various perovskite filmsespecially (MAPbI3). See text for more information.

ARTICLE COMMUNICATIONS CHEMISTRY | https://doi.org/10.1038/s42004-022-00683-7

6 COMMUNICATIONS CHEMISTRY | (2022) 5:67 | https://doi.org/10.1038/s42004-022-00683-7 | www.nature.com/commschem

29. Asbury, J. B. et al. Femtosecond Ir Study of Excited-State Relaxation andElectron-Injection Dynamics of Ru(Dcbpy)(2)(Ncs)(2) in Solution and onNanocrystalline TiO2 and Al2O3 Thin Films. J. Phys. Chem. B 103, 3110–3119(1999).

30. Furube, A. et al. Near-Ir Transient Absorption Study on Ultrafast Electron-Injection Dynamics from a Ru-Complex Dye into Nanocrystalline In2o3 ThinFilms: Comparison with SnO2, ZnO, and TiO2 Films. J. Photoch. Photobio. A182, 273–279 (2006).

31. Ellingson, R. J. et al. Dynamics of Electron Injection in NanocrystallineTitanium Dioxide Films Sensitized with [Ru(4,4 ‘-Dicarboxy-2,2 ‘-Bipyridine)(2)(Ncs)(2)] by Infrared Transient Absorption. J. Phys. Chem. B 102,6455–6458 (1998).

32. El-Zohry, A. M. The Origin of Slow Electron Injection Rates for Indoline DyesUsed in Dye-Sensitized Solar Cells. Dyes Pigm. 160, 671–674 (2019).

33. El-Zohry, A. M. & Karlsson, M. Gigantic Relevance of Twisted IntramolecularCharge Transfer for Organic Dyes Used in Solar Cells. J. Phys. Chem. C. 122,23998–24003 (2018).

34. Juozapavicius, M. et al. Evidence for “Slow” Electron Injection inCommercially Relevant Dye-Sensitized Solar Cells by Vis-Nir and Ir Pump-Probe Spectroscopy. J. Phys. Chem. C. 117, 25317–25324 (2013).

35. Juozapavicius, M., Kaucikas, M., van Thor, J. J. & O’Regan, B. C. Observationof Multiexponential Pico- to Subnanosecond Electron Injection in OptimizedDye-Sensitized Solar Cells with Visible-Pump Mid-Infrared-Probe TransientAbsorption Spectroscopy. J. Phys. Chem. C. 117, 116–123 (2012).

36. Ghosh, H. N., Asbury, J. B. & Lian, T. Q. Direct Observation of UltrafastElectron Injection from Coumarin 343 to TiO2 Nanoparticles by FemtosecondInfrared Spectroscopy. J. Phys. Chem. B 102, 6482–6486 (1998).

37. El-Zohry, A. M. & Zietz, B. Electron Dynamics in Dye-Sensitized Solar CellsInfluenced by Dye-Electrolyte Complexation. J. Phys. Chem. C. 124,16300–16307 (2020).

38. Furube, A., Du, L., Hara, K., Katoh, R. & Tachiya, M. Ultrafast Plasmon-Induced Electron Transfer from Gold Nanodots into TiO2 Nanoparticles. J.Am. Chem. Soc. 129, 14852–14855 (2007).

39. Pavliuk, M. V. et al. Magnetic Manipulation of Spontaneous Emission fromInorganic Cspbbr3 Perovskites Nanocrystals. Adv. Optical Mater. 4,2004–2008 (2016).

40. Cieslak, A. M. et al. Ultra Long-Lived Electron-Hole Separation within Water-Soluble Colloidal ZnO Nanocrystals: Prospective Applications for SolarEnergy Production. Nano Energy 30, 187–192 (2016).

41. Harrick, N. J. Lifetime Measurements of Excess Carriers in Semiconductors. J.Appl. Phys. 27, 1439–1442 (1956).

42. Tang, J. et al. Colloidal-Quantum-Dot Photovoltaics Using Atomic-LigandPassivation. Nat. Mater. 10, 765–771 (2011).

43. Tamaki, Y. et al. Trapping Dynamics of Electrons and Holes in a NanocrystallineTiO2 Film Revealed by Femtosecond Visible/near-Infrared Transient AbsorptionSpectroscopy. Comptes Rendus Chim. 9, 268–274 (2006).

44. Heimer, T. A. & Heilweil, E. J. Direct Time-Resolved Infrared Measurement ofElectron Injection in Dye-Sensitized Titanium Dioxide Films. J. Phys. Chem. B101, 10990–10993 (1997).

45. Guo, P. J. et al. Slow Thermal Equilibration in Methylammonium Lead IodideRevealed by Transient Mid-Infrared Spectroscopy. Nat. Commun. 9, 1–9(2018).

46. El-Zohry, A. M., Cong, J., Karlsson, M., Kloo, L. & Zietz, B. Ferrocene as aRapid Charge Regenerator in Dye-Sensitized Solar Cells. Dyes Pigm. 132,360–368 (2016).

47. Barker, A. J. et al. Defect-Assisted Photoinduced Halide Segregation in Mixed-Halide Perovskite Thin Films. ACS Energy Lett. 2, 1416–1424 (2017).

48. Samu, G. F., Janáky, C. & Kamat, P. V. A Victim of Halide Ion Segregation.How Light Soaking Affects Solar Cell Performance of Mixed Halide LeadPerovskites. ACS Energy Lett. 2, 1860–1861 (2017).

49. Chen, Y. H. et al. Impacts of Alkaline on the Defects Property and CrystallizationKinetics in Perovskite Solar Cells. Nat. Commun. 10, 1–9 (2019).

50. Ghosh, S., Pal, S. K., Karki, K. J. & Pullerits, T. Ion Migration Heals TrappingCenters in CH3NH3PbBr3 Perovskite. ACS Energy Lett. 2, 2133–2139 (2017).

51. Kamat, P. V. & Kuno, M. Halide Ion Migration in Perovskite Nanocrystalsand Nanostructures. Acc. Chem. Res. 54, 520–531 (2021).

52. Jeon, N. J. et al. Solvent Engineering for High-PerformanceInorganic–Organic Hybrid Perovskite Solar Cells. Nat. Mater. 13, 897 (2014).

53. Hussain, M. et al. Spin–Orbit Charge-Transfer Intersystem Crossing ofCompact Naphthalenediimide-Carbazole Electron-Donor–Acceptor Triads. J.Phys. Chem. B 125, 10813–10831 (2021).

54. Imran, M. et al. Intersystem Crossing Via Charge Recombination in aPerylene-Naphthalimide Compact Electron Donor/Acceptor Dyad. J. Mater.Chem. C. 8, 8305–8319 (2020).

55. El-Zohry, A. M. Excited-State Dynamics of Organic Dyes in Solar Cells. InSolar Cells - Theory, Materials and Recent Advances, https://doi.org/10.5772/intechopen.94132 (IntechOpen, London, 2020).

AcknowledgementsThe authors thank for the financial support provided by KASUT to carry out this work.

Author contributionsA.M.El-Z. designed the concept of the paper, measured the mid-IR data wrote the paper,and put the overall scheme. B.T. and A.A. prepared the perovskite films and revised thepaper. P.M. performed the transient data in the visible range along with revising thepaper. O.M.B. supervised the project and revised the paper. B.S.O. supervised the projectand revised the paper. O.F.M. supervised the project and revised the paper.

FundingOpen access funding provided by Stockholm University.

Competing interestsThe authors declare no competing interests.

Additional informationSupplementary information The online version contains supplementary materialavailable at https://doi.org/10.1038/s42004-022-00683-7.

Correspondence and requests for materials should be addressed to Ahmed M. El-Zohryor Omar F. Mohammed.

Peer review information Communications Chemistry thanks the anonymous reviewersfor their contribution to the peer review of this work. Peer reviewer reports are available.

Reprints and permission information is available at http://www.nature.com/reprints

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.

Open Access This article is licensed under a Creative CommonsAttribution 4.0 International License, which permits use, sharing,

adaptation, distribution and reproduction in any medium or format, as long as you giveappropriate credit to the original author(s) and the source, provide a link to the CreativeCommons license, and indicate if changes were made. The images or other third partymaterial in this article are included in the article’s Creative Commons license, unlessindicated otherwise in a credit line to the material. If material is not included in thearticle’s Creative Commons license and your intended use is not permitted by statutoryregulation or exceeds the permitted use, you will need to obtain permission directly fromthe copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

© The Author(s) 2022

COMMUNICATIONS CHEMISTRY | https://doi.org/10.1038/s42004-022-00683-7 ARTICLE

COMMUNICATIONS CHEMISTRY | (2022) 5:67 | https://doi.org/10.1038/s42004-022-00683-7 | www.nature.com/commschem 7