From Chalcogen Bonding to S–π Interactions in Hybrid Perovskite Photovoltaics

Abstract The stability of hybrid organic–inorganic halide perovskite semiconductors remains a significant obstacle to their application in photovoltaics. To this end, the use of low‐dimensional (LD) perovskites, which incorporate hydrophobic organic moieties, provides an effective strategy to improve their stability, yet often at the expense of their performance. To address this limitation, supramolecular engineering of noncovalent interactions between organic and inorganic components has shown potential by relying on hydrogen bonding and conventional van der Waals interactions. Here, the capacity to access novel LD perovskite structures that uniquely assemble through unorthodox S‐mediated interactions is explored by incorporating benzothiadiazole‐based moieties. The formation of S‐mediated LD structures is demonstrated, including one‐dimensional (1D) and layered two‐dimensional (2D) perovskite phases assembled via chalcogen bonding and S–π interactions. This involved a combination of techniques, such as single crystal and thin film X‐ray diffraction, as well as solid‐state NMR spectroscopy, complemented by molecular dynamics simulations, density functional theory calculations, and optoelectronic characterization, revealing superior conductivities of S‐mediated LD perovskites. The resulting materials are applied in n‐i‐p and p‐i‐n perovskite solar cells, demonstrating enhancements in performance and operational stability that reveal a versatile supramolecular strategy in photovoltaics.


Benzo[c][1,2,5]thiadiazol-4-ylmethanaminium bromide ((BTDZ)Br) was synthesized by protonation of benzo[c]
Perovskite powders were prepared using mechanosynthesis based on (BTDZ)2PbX4 (n =1; X = I, Br) nominal compositions.Stoichiometric amounts of PbX2 and (BTDZ)X were ground with steel beads in a Retsch Ball Mill MM-200 using a 10 mL grinding jar and a ⌀10 mm ball.The grinding process was carried out for 30 min at 25 Hz, followed by subsequent annealing at 150 °C for 15 min to obtain the final powder.
Perovskite thin films of (BTDZ)2PbX4 (n = 1) or (BTDZ)2FAPb2X7 (n = 2) nominal compositions were prepared using solution-processing of stoichiometric amounts of (BTDZ)X, PbX2, and FAX in concentrations of 0.4 M (X = I, Br).The solutions were mixed to obtain the perovskite precursor solution, which was applied onto the prepared glass substrate using a two-step spin-coating technique.In the initial phase, the coating process was executed at a spin speed of 1000 rpm, with an acceleration rate of 200 rpm/s, sustained for 10 s.Subsequently, the subsequent step was performed at 6000 rpm, with an acceleration rate of 2000 rpm/s, for a period of 30 s.The substrate was subsequently annealed at 150 °C for 15 min.

Direct (n-i-p) solar cell fabrication
involved FTO (fluorine-doped tin oxide) glass substrates that were sequentially cleaned with 2% Hellmanex solution, isopropanol, and ethanol in an ultrasonic bath for 15 min and then dried with N2.The cleaned FTO glass substrates were treated with UV-Ozone for 5 min before the deposition of TiO2.The TiO2 layer was deposited by aerosol spray pyrolysis using oxygen as a carrier gas.
For approximately 30 samples of 1.5 x 2.5 cm (112.5 cm 2 ), 0.2 mL of acetylacetone and 0.3 ml of titanium diisopropoxide bis(acetylacetonate) stock solution (75 wt.% in isopropanol) is diluted in 4.5 mL of ethanol for a total of 5 ml of solution (10% conc.).Substrates are heated to 450 °C for 15 min and 30 min after the spray of the precursor solution.A layer of mesoporous TiO2 (m-TiO2) was deposited onto the compact TiO2 (c-TiO2) substrate through a spin-coating process lasting 10 se, conducted at 4000 rpm with an acceleration rate of 1000 rpm/s.This deposition was achieved utilising the 30NRD TiO2 paste, with a quantity of 0.3 g per 2 mL of ethanol.Following this, the substrates dried at 100 °C, after which they underwent sintering at 450 °C for a duration of 30 min to form FTO/c-TiO2/m-TiO2. Subsequently, this substrate was coated with a perovskite solution that was prepared according to the method described below.For the hole-transport material, 100mg Spiro-OMeTAD was dissolved in 1.08 mL CB, doped by 23 μL Li-TFSI (1.8 mol/L in acetonitrile) and 39 μL tBP.The mixed Spiro-OMeTAD solution was spin-coated on the surface of the perovskite at 4000 rpm for 20 s with an acceleration rate of 2000 rpm/s.Finally, an 80 nm Au electrode was deposited by thermal evaporation through a shadow mask to form a device with an active area of 0.1 cm 2 .LD/3D perovskite layer was prepared by dissolving a mixture of PbI2 (543.0 mg), PbBr2 (21.0 mg), FAI (189.3 mg), MABr (6.4 mg), and CsI (15.8 mg) Cs0.05FA0.9MA0.05Pb(I0.95Br0.05)3with 5% excess of PbI2 in 1 mL of DMF and DMSO mixture (4:1 v/v) at 70 °C.The perovskite active layer was deposited using an antisolvent method with chlorobenzene.The perovskite precursor solution was deposited on the freshly prepared FTO/c-TiO2/m-TiO2 substrate, and a two-step spin-coating method was applied.The first step was carried out at 2000 rpm with an acceleration rate of 200 rpm/s for 10 s.The second step followed at 6000 rpm with an acceleration rate of 2000 rpm/s for 30 s. CB (200 μL) was slowly dripped at 15 s before the second step.After this, the substrate was annealed at 150 °C for 10 min.To treat the device, we applied a 1 mg/mL solution of (BTDZ)X (X = I, Br) in ethanol, which was then spin-coated on the surface of the perovskite at 4000 rpm for 20 s with an acceleration rate of 2000 rpm/s followed by annealing at 130 °C for 5 min.p-i-n) planar perovskite solar cells were fabricated in a nitrogen-filled glovebox.First, the NiO-sputtered ITO glass substrate was cleaned by ultrasonic cleaning in isopropanol and ethanol and subsequently dried in oven.The ITO/NiO substrate was then heated at 300 °C for 10 min.The hole transport layer was deposited on the ITO/NiO substrate by spin-coating the MeO-2PACz ([2-(3,6-dimethoxy-9Hcarbazol-9-yl)ethyl]phosphonic acid) solution at 3000 rpm for 30 s, followed by annealing at 100 °C for S4/30 10 min.The perovskite precursor solution was then spin-coated on the substrate in a one-step method at 3000 rpm for 10 s.The substrate was immediately put into a sample chamber connected to a vacuum pump, and the perovskite film was immediately exposed to low pressure for 15 s.The substrates were then transferred to a hot plate and annealed at 100 °C for 20 min.For the ETL, C60 was deposited using a thermal evaporator at a thickness of about 15 nm.The BCP solution was then spin-coated onto the C60 films at 4000 rpm and annealed at 70 °C for 5 min.Finally, a 100 nm Ag electrode was deposited by a thermal evaporator.

The inverted (
Ab initio calculations were performed within the Density Function Theory (DFT) framework as implemented in the CP2K software. [1]Both static optimisation and molecular dynamics simulations adopted the GGA-PBE functional with D3 dispersion corrections. [2]GTH pseudopotentials [3,4] together with the DZVP basis set [5] were utilised with an energy cutoff at 370 Ry.All three simulated n=1 RP systems (parallel in-plane, T-shaped interlayer and T-shaped intralayer) comprised a 4×4×2 slab of (BTDZ)2PbX4 with dimensions 27.67×27.67×40.5Åcontaining 32-unit cells for a total of 1376 atoms.Each system was then progressively heated up from 100 K to 300 K with a temperature step of 100 K by using the Bussi thermostat [6] with a 0.5 fs integration step.At each temperature increase, the systems were equilibrated for 8 ps.At 300 K, the systems were sampled for a total of 14 ps, the last 4 ps of which were averaged to compute the energy.The final equilibrated configurations were then used for the visualisation of CBs with the help of NCIPLOT, [7] and subsequently, the entire MD trajectory was analysed with in-house codes for the extraction of the conformational statistics (Fig. S7).
Classical MD simulations were performed with the LAMMPS software package. [8]The equations of motion were integrated using the Velocity-Verlet [9] algorithm with a 1 fs time step.Isothermal and isobaric conditions were maintained with the help of the Nosé-Hoover [10][11] thermostat and barostat with time constants 0.1ps and 0.35ps, respectively.Periodic boundary conditions were applied in all three directions.All classical systems, comprising a 9×9×2 slab of (BTDZ)2PbX4 with dimensions 62.26×62.26×40.5Åand a total of 6966 atoms, were equilibrated for 1ns at 300 K in the canonical NVT ensemble, where the total number of atoms N, the system's volume V and the temperature T were held constant.Next, production simulations were performed for 10 ns in the isothermal-isobaric NPT ensemble, where the sampling process took place every 500 fs for the computation of ensemble averages.Interatomic interactions for the perovskite part were described by an in-house developed and thoroughly validated classical force field. [12]The BTDZ cations were parametrised in the context of the OPLS all-atom force field by using LigParGen. [13]The van der Waals and electrostatic non-bonded interactions were calculated using a real-space cutoff radius of 10 and 8 Å, respectively, whereas the particle-particle-particle-mesh (PPPM) scheme [14] with an accuracy of 0.0001 (kcal/mol)/Å was used for the consideration of long-range electrostatic interactions.

The construction of the low-dimensional 1D δ-phase was realised starting from the δ-FAPbI3 and δ-CsPbI3
structures and expanding affinely the lattice vectors by 100% so that the BTDZ molecules could fit and replace all FA and Cs, respectively.The built systems comprised a 4×3×1 supercell.Next, the systems were equilibrated with AIMD simulations at 300 K and 1 atm in the NPT ensemble.To consider relaxation modes with longer characteristic times, the equilibrated structures were replicated in all three dimensions 3×4×7 times for the initialisation of classical MD simulations for 20 ns in the NPT statistical ensemble.The XRD diffraction patterns of the equilibrated structures were generated using CrystalDiffract (CrystalMaker Software Ltd, Oxford, England; www.crystalmaker.com).Interestingly, the simulated XRD pattern of the 1D δ-phase originating from CsPbI3 reproduced well the peaks of the experimental XRD diffraction patterns originally assigned to different layered polymorphs (Fig. S5 and Fig. S11).The organic component of this LD phase formed a continuous cross-linked network of BTDZ molecules interacting through CBs (Fig. S5).

The relative formation energies (Ef)
were computed for all considered systems with the help of AIMD simulations complemented by DFT calculations.Initially, each system was relaxed with AIMD simulations at 300 K and 1 atm.Following that, each equilibrated system was quenched down to 0 K.For the formation energies, we computed via DFT the energetic contribution of each ion of the system -Pb 2+ , I -, BTDZ + , Br - -as if it was isolated and we applied the following formula: where Es is the energy of the system, i is the ion type, ni indicates the number of ions of type i in the system, Ei is the DFT energy of ion i and m is the number of stoichiometric units of the model system.This analysis identified the 2D (BTDZ)2PbBr4 system as the energetically most favourable one, whereas the LD systems were the least stable ones.The resulting energy relations are shown in Table S1.Similarly, the calculated projected density of states (PDOS) of all systems (Fig. S8) suggests that only the 2D (BTDZ)2PbI4 and 1D δ-phase systems contributed favourably towards supporting charge transfer, with the molecular orbitals of BTDZ of these systems overlapping with the bottom of the perovskite conduction band, thereby reducing the band gap.The remaining LD systems, i.e., 1D (BTDZ)PbI3 and 2D (BTDZ)2PbBr4 (Fig. S8b,c) exhibited trapped states within the band gap, inhibiting potential charge transfer (Fig. S8a,d).

X-ray diffraction measurements were conducted using a PANalytical Empyrean Series 2 instrument in
Bragg-Brentano configuration.The instrument utilised Cu Kα radiation with a voltage of 40 kV and a current of 40 mA.For the specific purpose of grazing incidence measurements, the X-ray incidence angle was set at 2°. Diffracted X-rays were detected during the experiments employing a PIXcel3D detector.The chemical composition of films was measured using XPS (AXIS SUPRA) with an Al Ka radiation source, and all binding energies were calibrated by C 1s (248.8 eV) as a reference.

Grazing incidence wide-angle X-ray scattering (GIWAXS)
was employed to analyse the thin films deposited on microscope and ITO glass substrates.GIWAXS and XRR measurements were performed at DESY beamline P08. [15]The beam energy was 18 keV, the incidence angle of 0-0.3°, and the beam size was 300x100 µm.A Perkin Elmer XRD 1621 detector with a spatial resolution of 200 µm was used to record two-dimensional diffraction patterns with a sample-to-detector-distance of 700 mm.For the recording of XRR data, a Dectris Pilatus 100 K detector was used.Additional GIWAXS measurements were performed at the ESRF beamline ID10.An Eiger 4M detector was used for the diffraction measurement with a sample-to-detector distance of 625 mm.The beam energy was 22.5 keV with a beam size of 17 x 28.5 µm and an angle of incidence of 0-0.3°.All measurements were done in an inert nitrogen atmosphere.

Single crystal X-Ray diffraction data for (BTDZ)I, (BTDZ)PbI3 and (BTDZ)2PbBr4 were collected on a
Rigaku Oxford Diffraction XtaLAB Synergy-R DW diffractometer equipped with a HyPix ARC 150° Hybrid Photon Counting (HPC) detector at 100 K using Cu Kα (λ = 1.54184Å).Data were processed using the CrysAlisPro software. [16]The structures were solved by intrinsic phasing with SHELXT [17] and refined by full-matrix least-squares methods based F 2 using SHELXL. [18]The single crystals were obtained by using the following procedures.The vial was opened and placed on a hotplate set to 65°C.After evaporation of ca.90% of the solvent, small crystals formed on the walls of the vial.The hot plate was turned off, the solution was allowed to cool slowly to room temperature, and the colourless block crystals were collected.The structure of (BTDZ)2PbBr4 is available on the Cambridge Structural Database under CCDC number 2342702.
UV-vis absorption measurements were conducted using a Shimadzu UV-2600 spectrophotometer.
Transient absorption spectra of perovskite films were recorded using a previously described setup [19] using modules supplied by Light Conversion, with a 1030 nm seed laser (PHAROS, Light Conversion, Yb:KGW lasing medium, 400 µJ pulse energy, 200 fs duration, 50 kHz repetition rate).The 515 nm or 343 nm pump     Table S1.Relative formation energies per stoichiometric unit of all LD systems considered in this study.
Energies are given at 0 K in reference to the most stable, i.e., (BTDZ)2PbBr4, LD system.(d-f) A comparison of angular profiles extracted from GIWAXS reciprocal space maps (in b, left and g) and powder diffraction from the simulated structures.(g) GIWAXS reciprocal space map measured at 0.3° incidence angle, probing the whole thickness of the films with marked structures, namely n = 2 2D structure (20 Å, white ovals), (BTDZ)PbI3 (cyan circles), and 3D perovskite (α-FAPbI3 phase, white diamonds and ovals > 1 Å -1 ) in n = 2 (BTDZ)2FAPb2I7 nominal composition on glass.Bragg peak positions and corresponding structures are listed in Table S2-S3 and the details are included in the Supplementary Discussion section.

LD System
Table S2.List of Bragg peaks from LD perovskite structures based on GIWAXS profiles of n = 1 nominal composition on glass (Fig. S11d).#1-2 indicate other LD phases identified in GIWAXS profiles.

Bragg Peak Position Structure
Table S3.List of Bragg peaks from LD and 3D perovskite structures based on GIWAXS profiles of n = 2 nominal composition on glass (Fig. S11e-f).#1-2 indicate other LD phases identified in GIWAXS profiles.

Bragg Peak Position Structure
were recorded at 11.7 T (125 MHz for 13 C) and 20 T (213.79MHz for 13 C, 86.15 MHz for 15 N).The 11.7 T instrument (EPFL) was equipped with a Bruker Avance III console and used 3.2 mm three-channel low-temperature magic angle spinning (MAS) probe.The samples were packed into 3.2 mm zirconia rotors under ambient conditions.The 20 T instrument (University of Warwick) was equipped with a Bruker Avance NEO console, and it used a 4 mm threechannel MAS probe.The samples were packed into zirconia rotors under ambient conditions.
For (BTDZ)I: The crystallisation process involved a slow crystallisation from a saturated solution of BTDZI in tetrahydrofuran.The structure is available on the Cambridge Structural Database under CCDC number 2342700.(BTDZ)PbI3: In a 10 mL vial 5.9 mg (12.8 μmol) of PbI2 was mixed with 7.5 mg (25.6 μmol) of BTDZI in 5 mL of tetrahydrofuran and sonicated until a clear yelloworange solution was obtained.The vial was open and placed on a hotplate set to 85 °C.After evaporation of ca.80% of the solvent volume, small crystals formed on the walls of the vial.The hot plate was turned off, the solution was allowed to cool down slowly to room temperature, and the colourless block crystals were collected.The structure of (BTDZ)PbI3 is available on the Cambridge Structural Database under CCDC number 2342701.(BTDZ)2PbBr4: In a 20 mL vial 6.0 mg (16.3 μmol) of PbBr2 was mixed with 8.0 mg (32.5 μmol) of BTDZBr in 10 mL of acetone and sonicated until a clear yellow-orange solution was obtained.

Fig. S3 .
Fig. S3.Examples of interlayer CB formation from an equilibrated AIMD trajectory of the rotated interlayer model.In all cases, CBs are denoted with black broken lines and visualized via NCIPLOT green isosurfaces.

Fig. S4 .
Fig. S4.Examples of intralayer CB formation within the same organic layer from an equilibrated AIMD trajectory of the rotated interlayer model.In all cases, CBs are denoted with black broken lines and visualized through NCIPLOT with green isosurfaces indicating weak noncovalent interactions.

Fig. S5 .
Fig. S5.Radial distribution function g(r) for an n = 1 RP model system between a) Pb atoms belonging to different inorganic layers, b) nitrogen (N1 and N2, see Fig. S1) and sulfur atoms.The first peak in part (b) corresponds to the strong covalent bond formed between N and S atoms in a single BTDZ molecule.The second peak, at r = 3.8 Å, is indicative of the formation of CBs between adjacent interacting BTDZ molecules, whereas the third peak, at r = 5.7 Å, describes weaker interactions between the S atom of one molecule and the more distant N atom of the thiadiazole ring of the second molecule.(c) Left: Simulated XRD diffraction pattern of the (BTDZ)PbI3-based 1D δ-phase after replacing Cs with BTDZ.The three lines indicate the position of the experimental peaks (Fig. S11 and TableS2).Right: Molecular representation of the equilibrated 1D (BTDZ)PbI3 δ-phase showing the continuous cross-linked network of the organic phase.The face-sharing PbI6 octahedra are normal to the shown plane and are hidden for clarity.

Fig. S6 .
Fig. S6.Schematic for the definition of the four conformational angles: a) the molecule planarity angle φ, b) the substituent rotation angle ξ1, c) the off-plane angle ω and d) the ring rotation angle θ.In all parts, red arrows denote normal vectors to corresponding planes.In part (c), one red vector connects the two centers of mass of the thiadiazole rings, and the remaining vector indicates the normal to one thiadiazole ring.

Fig. S7 .Fig. S8 .
Fig. S7.Histograms of the four conformational angles considered here averaged over the AIMD trajectory for the most stable, rotated interlayer, starting configuration.Part (a) shows that the ring system of the BTDZ molecule exhibits increased stiffness with a small bending angle of only 6 o , whereas (b) shows that the methylammonium substituent exhibits a dominant rotation angle of about 90 o off the ring plane.Parts (c) and (d) summarize the prevalent molecular orientation during CB formation, dictating that the two BTDZ molecules are predominantly out of plane by about 45 o and the ring planes are rotated by about 20 o with respect to each other.Interestingly, a secondary smaller peak at 90 o is observed in part (d) which corresponds to the perfect T-shaped π-stacking of the two BTDZ molecules.The third peak appearing at 160 o is the supplementary angle of the first peak at 20 o .

Fig. S13 .
Fig. S13.(a-b) GIWAXS reciprocal space maps measured at 0.3° incidence angle for (BTDZ)2PbBr4 (n = 1) and (BTDZ)2FAPb2Br7 (n = 2) nominal compositions.Red markers: (BTDZ)2PbBr4 layered structure (based on the crystal structure stabilizsd by S-π interactions) in an orientation with the layers parallel to the substrate.Blue markers: a similar structure with an extended c-axis (i.e., stacking distance), corresponding to the layered structure stabilised by CB.Orange circles: unidentified peaks, likely corresponding to 1D or 2D n = 2 phases.(c-d) Comparison of radial profiles extracted from GIWAXS reciprocal space maps (IN black, blue, purple, and red) and simulated diffraction patterns (in green) of layered structures stabilised by S-π and CB interactions.A close-up of the 2D phase signal is shown in d), showing a double peak, indicating two different structures corresponding to CB and S-π interactions.Control is a reference 3D perovskite thin film, while Br-LD/3D indicates perovskite thin films treated with (BDTZ)Br.

Fig. S14. 15 N
Fig. S14. 15N CP NMR spectra of the (BTDZ)I and corresponding mechanosynthesised n = 1 2D perovskite recorded at 20 T (12 kHz MAS) at room temperature.Note that the material after mechanosynthesis contains a mixture of n = 1 and unreacted (BTDZ)I.Acquisition parameters were as follows.(BTDZ)I: recycle delay 1 s, number of scans 10000; n = 1: recycle delay 1 s, number of scans 57548.Lorentzian broadening of 50 Hz was applied to both spectra.A contact time of 1 ms was used for the CP.

Fig. S16 .
Fig. S16.(a) UV-vis SEC record of the (BTDZ)I spacer precursor in solution during the cathodic CV scan shown in the inset; coloured circles signify potentials at which identically coloured UV-vis spectra were recorded.(b) EPR/vis-SEC record of the organic spacer precursor solution.Spectra were recorded duringthe cathodic CV scan shown in the inset, with the coloured circles corresponding to the potentials identically coloured in UV-vis absorption spectra (with Pt mesh working electrode 3 mVs -1 scan rate).After the forward cathodic scan, the potential was maintained at -2.1 V for 5 min to accumulate the generated paramagnetic species before re-oxidation anodic sweep.Single line with giso of 2.0048 ± 0.0002 and ΔBpp of 0.18 mT.[]

Fig. S19 .
Fig. S19.XRD pattern of (a) 3D perovskite films without (control) and with (treated) BTDZ overlayer and (b) 3D perovskite films without (control) and with (treated) (BTDZ)I, highlighting the low-angle reflection peak at 8 degrees (a, inset) that is more apparent in higher concentrations of (BTDZ)I (5 mg/mL in c).(d) GIWAXS reciprocal space map of 3D perovskite films with (BTDZ)I overlayer measured at (left) 0.1° and (right) 0.3° incidence angle, probing the surface and the whole thickness of the film, respectively.Bragg peaks at 0.45 Å -1 (surface) and 0.42 Å -1 (whole sample) correspond to LD phases on top of 3D perovskite.

Fig. S21 .
Fig. S21.Cross-sectional STEM image of perovskite solar cell with BTDZ and (b) TEM images of the perovskite film with BTDZ, highlighting the co-existence of 3D and low-dimensional (LD) perovskite phases at the interface with the hole-transport material (HTM).

Fig. S24 .
Fig. S24.(a) UV-vis absorption and (b) photoluminescence (PL, dashed lines) emission spectra of 3D (black), and mixed LD/3D (red) perovskite thin films on microscope glass.(c-d) Transient absorption spectra of 3D (f) and (g) mixed LD/3D perovskite thin films on microscope glass at varying time delaysupon excitation at 515 nm (6.9 mJ/cm 2 fluence).(e) Kinetics from the ground state bleach at 777 nm, highlighting faster recovery of the 3D perovskite (black) as compared to the mixed LD/3D system (red) that can be associated with a passivation effect of the LD overlayer on the 3D perovskite thin films.

Fig. S25 .
Fig. S25.(a) Schematic of p-i-n perovskite solar cell architecture and the corresponding (b) box charts illustrating the statistical distribution of photovoltaic parameters: open-circuit voltage (VOC), short-circuit current density (JSC), fill factor (FF), and power conversion efficiency (PCE).(c-d) IPCE spectra of the (c) n-i-p and (d) p-i-n perovskite solar cells without (control, black) and with (treated, red) (BTDZ)I.

Fig. S27 .
Fig. S27.(a) Box charts illustrating the statistical distribution of photovoltaic parameters upon (BTDZ)Br (red) and (BTDZ)I (blue) treatment as compared to the control (black) in p-i-n device architectures.(b) Current-voltage characteristics of the champion devices for upon (BTDZ)Br (red) and (BTDZ)I (blue)treatment as compared to the control (black) in p-i-n device architectures.The results suggest improvements in short-circuit currents and fill factor for (BTDZ)Br systems, which could be associated with the higher contribution of the 2D perovskite phase for bromide-based perovskite as compared to iodide compositions.

Fig. S29 .
Fig. S29.J-V curve of BTDZ-based perovskites of n = 4 nominal compositions in n-i-p devices under AM 1.5G illumination.While further optimisation of the film fabrication, phase purity, and device architecture could contribute to enhancing the device performances, the photovoltaic performances of n = 4

Fig. S30 .
Fig. S30.(a) Contact angle of the water droplet on top of the 3D perovskite film without (control, left) and with (treated, right) (BTDZ)I overlayer.(b) Scatter plot of the corresponding contact angles (in a) as a function of the loading time of the water droplet on top of the 3D perovskite film without (control, black) and with (treated, red) (BTDZ)I overlayer in LD/3D perovskite films.