2D-Self-Assembled Organic Materials in Undoped Hole Transport Bilayers for Efficient Inverted Perovskite Solar Cells

Interfaces between photoactive perovskite layer and selective contacts play a key role in the performance of perovskite solar cells (PSCs). The properties of the interface can be modified by the introduction of molecular interlayers between the halide perovskite and the transporting layers. Herein, two novel structurally related molecules, 1,3,5-tris(α-carbolin-6-yl)benzene (TACB) and the hexamethylated derivative of truxenotris(7-azaindole) (TTAI), are reported. Both molecules have the ability to self-assemble through reciprocal hydrogen bond interactions, but they have different degrees of conformational freedom. The benefits of combining these tripodal 2D-self-assembled small molecular materials with well-known hole transporting layers (HTLs), such as PEDOT:PSS and PTAA, in PSCs with inverted configuration are described. The use of these molecules, particularly the more rigid TTAI, enhanced the charge extraction efficiency and reduced the charge recombination. Consequently, an improved photovoltaic performance was achieved in comparison to the devices fabricated with the standard HTLs.


Charge transport measurements
Hole only devices with the architecture: ITO/MoO3 (10 nm)/TACB (80 nm) or TTAI (100 nm) /MoO3 (10 nm)/Ag (100 nm) were fabricated by thermal evaporation under high vacuum conditions. Current density versus voltage curves were measured using a Keithley 2636A SMU. These curves were fitted to the Murgatroyd equation where J is the current density, 0 the vacuum permittivity (8.85 × 10 -14 F cm -1 ), r the relative permittivity of the organic layer (≈3), 0 the zero-field mobility, L the thickness of the organic layer, and  the field-activation factor. The hole mobility was determined using the Poole-Frenkel relationship, =0exp(E 1/2 ), at an electric field of 10 5 V cm -1 .

Solar cells fabrication and characterization.
Pre-patterned ITO glass substrates were cleaned by sequential sonication in an Extran aqueous solution, deionized water, ethanol, acetone, and isopropanol for 15 minutes each step. Right before using, the substrates were subject to an UV-ozone treatment for 30 minutes. Then, the hole transporting materials were deposited. PEDOT:PSS (Heraeus, Clevios PVP AI 4083) filtered with a 0.45 m PVDF filter was spin coated at 5000 rpm for 40 seconds under ambient conditions. The substrates were then annealed at 130 °C for 20 minutes. A 2 mg/mL solution of PTAA in toluene was spin coated at 5000 rpm for 30 seconds in a nitrogen-filled glovebox and the substrates were annealed at 100 °C for 10 minutes. Those devices incorporating the small molecular materials were transferred to a vacuum chamber to deposit a thin layer of about 10 nm by thermal evaporation under high vacuum (~10 −7 mbar). The perovskite active layer was prepared in a nitrogen-filled glovebox with the composition Cs0.05(MA0.15FA0.85)0.95Pb(Br0.15I0.85)3. The perovskite precursor solution was prepared by mixing in a 85:15 volume ratio a formamidium iodide solution (1.24 M), dissolved in a solution of PbI2 (1.5 M) in DMF:DMSO (4:1 volume ratio), and a methylamonium bromide solution (1.24 M), dissolved in a solution of PbBr2 (1.5 M) in DMF:DMSO (4:1 volume ratio). The resulting solution was mixed with CsI (1.5 M) dissolved in DMSO in a 0.95:0.05 volume ratio. The freshly prepared precursor solution was deposited on the HTL-coated substrates at 4000 rpm for 30 seconds. After 25 seconds, 250 L of chlorobenzene was dropped onto the spinning substrate. Subsequently, an annealing step was performed at 100 °C for 30 minutes. The devices were finished by the thermal evaporation under high vacuum conditions (~10 −7 mbar) of a 23 nm thick layer of C60, a 8 nm thick layer of bathocuproine, and a 100 nm thick Ag top electrode. The current-voltage curves were measured using a Keithley 2612 source meter under AM 1.5 G (1000 Wm -2 ) in a Solar Simulator Abet, Xenon short-arc lamp Ushio 150 watts, in air at room temperature. A scan rate of 10 mVs -1 was used. External Quantum Efficiency measurements were carried out using a QEPVSI-b Oriel system also in air.

Photoluminescence measurements of complete devices at different potentials
The photoluminescence of the complete devices at different applied potentials were recorded on a homemade setup that consists of a 532 nm solid state laser diode (5 mW, 0.0314 cm -2 beam spot) as an optical excitation source, a CCD detector (Andor-iDUS DV420A-OE) coupled to a spectrograph (Kymera-193I-B2) setup for the optical detection, and a potentiostat (GAMRY Reference 3000) for electrical biasing.

Electrochemical Impedance Spectroscopy (EIS)
Measurements were carried out on complete devices performed by applying 20 mV of a voltage perturbation at different frequencies from 1MHz to 0.1Hz in a PGSTAT-30 from Autolab and under 1 sun illumination at different applied forward biased, from 0 to 1.1 volts (up to VOC value). The equivalent circuit used to fitting the measurements it is shown below, [1] where transport resistance has been neglected.

Morphology
Field Emission Scanning Electron Microscopy images of the samples mounted on aluminum stubs were examined using a FE-SEM (ApreoS Lovac IML Thermofisher) with a selected voltage of 5 kV and 0.10 nA under high vacuum conditions. T3 detector and immersion mode were used for sample imaging.

4.-Computational details
Geometry optimizations were carried out by DFT calculations using the ORCA program, being run in redundant internal coordinates with tight convergence criteria. [2] All calculations were performed using the B3LYP functional, [3] the def2-TZVP basis set, [4] the RIJCOSX algorithm, [5] and dispersion corrections (DFT-D4). [6] All geometry optimizations took into account the effect of the solvent (dimethylformamide) by means of the CPCM model. [7] Figure S15. Electroluminescence spectra registered at an applied constant forward voltage of 1.3 V.