We have synthesized and applied GO-COOH as an effective dopant for Spiro-OMeTAD in C-PSCs processed at low temperature, enhancing charge transfer across the HTL interface and facilitating the construction of an ideal HTL-carbon electrode contact. C-PSCs based on a bilayer carbon electrode realized an outstanding PCE of 23.6% and demonstrated a VOC-FF product that exceeds 80% of the S-Q limit. Both theoretical calculations and experimental studies confirmed the transfer of valence electrons from Spiro-OMeTAD to GO-COOH, leading to notably enhanced conductivity and optimized band alignment. The conjugated network surface promotes the formation of robust π-π interactions with the carbon electrode and facilitates efficient charge transfer kinetics in C-PSCs. Moreover, the incorporation of GO-COOH into Spiro-OMeTAD allows for the fixation of mobile Li+ ions through intercalation and the formation of Li-C bonds, which enhances the stability of devices. This work provides insights into realizing the oxygen-free, rapid oxidation of Spiro-OMeTAD and the construction of an ideal HTL interface to boost the efficiency and stability of low-temperature processed C-PSCs.
Unless stated otherwise, all chemicals were obtained from commercial sources and used as received, including FTO-coated glass substrates (surface resistivity ≈ 11 Ω sq, Suzhou Shangyang Solar Technology), SnO colloidal dispersion (15 wt% in HO, Alfa Aesar), lead iodide (PbI, >99.999%, Sigma-Aldrich), formamidinium iodide (FAI, 99.8%, Xi'an Polymer Light Technology), methylammonium chloride (MACl, 99.9%, Xi'an Polymer Light Technology), methylammonium iodide (MAI, 99.9%, Xi'an Polymer Light Technology), Spiro-OMeTAD (99.8%, Xi'an Polymer Light Technology), LiTFSI (99.95%, Sigma-Aldrich), 4-tert-butylpyridine (t-BP, 96%, Sigma-Aldrich), N,N-dimethylformamide (DMF, 99.8%, Sigma-Aldrich), dimethylsulfoxide (DMSO, 99.9%, Sigma-Aldrich), acetonitrile (99.8%, Sigma Aldrich), CB (99.8%, Sigma Aldrich), deionized water (Sinopharm Chemical Reagent), ethanol (Sinopharm Chemical Reagent), isopropanol (99.8%, Sinopharm Chemical Reagent), multilayer graphene (Nanjing XFNANO Materials Tech) and commercial highly conductive carbon material (Jiangxi DASEN Technology).
GO-COOH was prepared by a chemical method. First, we mixed HCl and HNO in a volume ratio of 3:1 to obtain a mixed acid solution. At room temperature, we added 200 mg multilayer graphene powder to 20 ml of the mixed acid solution under magnetic stirring, continued stirring for a further 30 s, poured the dispersion into 2,000 ml deionized water and quickly filtered the mixture under reduced pressure. Shortening the stirring time or increasing the volume ratio of the mixed acid solution could lead to less oxidized GO (with reduced COOH content) and other oxygen-containing functional groups, including -O-, -C=O and -OH. We rinsed the residue with deionized water three times under vacuum filtration conditions and collected the black precipitate on the filter membrane. GO-COOH powder was obtained after drying under vacuum at 60 °C for 12 h. Unless noted otherwise, the GO-COOH reported in the text was prepared by this method, and was shown to have an oxygen content of 17% by XPS.
C-PSC devices were fabricated by separately preparing two semi-cells (A and B) and stacking them face to face to mechanically connect them through carbon. For semi-cell A, the FTO substrate was sequentially cleaned with an aqueous solution of detergent, deionized water, anhydrous ethanol and isopropanol in an ultrasonic cleaning machine for 15 min, followed by UV-ozonation treatment for 15 min. A colloidal dispersion of SnO and deionized water were mixed in a volume ratio of 1:4 and shaken for 2 min to disperse the colloids evenly throughout the solution. After filtration, the dispersion was spin-coated onto the treated FTO substrate at 3,000 r.p.m. for 30 s and then heated at 150 °C for 30 min. The perovskite layer was deposited on the prepared SnO film using a two-step process. First, a 1.5 M PbI solution was prepared by mixing PbI in a solution of DMF and DMSO in a volumetric ratio of 9:1. The mass of PbI used to prepare the 1.5 M solution was 691.5 mg dissolved in a mixed 1 mL solvent of DMF and DMSO. The PbI solution was then spin-coated onto the SnO film at 2,000 r.p.m. for 30 s, followed by heating at 70 °C for 20 s. After cooling the annealed film to room temperature, we spin-coated the organic salt solution. MACl (27 mg), FAI (270 mg) and MAI (19.2 mg) were added to 3 ml isopropanol and the mixture was stirred and filtered to obtain an organic salt solution, which was then spin-coated onto the PbI film at 2,000 r.p.m. for 15 s, followed by annealing at 150 °C for 10 min to obtain the perovskite film. Next, we prepared the HTL. The control solution was obtained by dissolving Spiro-OMeTAD (72.3 mg), LiTFSI in acetonitrile (17.5 μl, 520 mg ml) and t-BP (28.8 μl) in 1 ml CB. Doping solutions were obtained by adding GO-COOH or G powder, at concentrations of 0.5, 1.0, 1.5 mg mL, to the control solution. The control and doped solutions were both spin-coated onto the perovskite film at 3,000 r.p.m. for 30 s at room temperature (20-25 °C). To deposit the bilayer electrode, 20 mg of multilayer graphene powder was added to 5 ml isopropanol and ball-milled for 12 h to obtain a graphene dispersion. The graphene dispersion was then spray-coated onto the prepared HTL under 85 °C annealing to remove isopropanol until reaching a target thickness of ~3 μm, finally producing semi-cell A. For semi-cell B, the commercially highly conductive carbon film was spread out and secured on bare glass with adhesive tape to act as a charge collector, and then the graphene slurry was spray-coated onto it under 85 °C annealing to remove isopropanol, obtaining semi-cell B. For each semi-cell, the thickness of the sprayed graphene layer was around 3 μm. The C-PSCs were then assembled by stacking semi-cell B and semi-cell A face to face under a low pressure without any resins or glues. The thickness of the whole carbon electrode was calculated to be around 8.5 μm (Supplementary Fig. 24), which may correspond to the actual thickness of the carbon under working conditions. Finally, an anti-reflection layer (evaporated MgF with a thickness of 150 nm) was deposited onto the glass side of semi-cell A. All complete C-PSCs devices mentioned in this Article and its Supplementary Information possess this anti-reflection layer of MgF.
We used the Molclus program to analyse the spatial distribution of the four types of oxygen-functionalized groups surrounding the GO or G sheets in the vicinity of Spiro-OMeTAD. The system was treated as a cluster molecular model. Initially, we generated 40 different spatial configurations of Spiro-OMeTAD and the GO or G sheets using the genmer subroutine. Subsequently, we used the Molclus program in calling the Gaussian16 software and applied the PM6 semi-empirical method to optimize and order the initial conformations according to their energies. For each system, we selected the three most energetically stable conformations to conduct optimizations and infrared frequency calculations using the B3LYP functional with the 6-31 g(d) basis set, correcting the imaginary frequencies to obtain accurate structures. Finally, we identified the most stable configurations, as depicted in Supplementary Fig. 6.
Next, we used the M062X functional with the DEF2TZVP basis set to calculate the single-point energies of the systems, which allowed us to obtain information on the frontier molecular orbitals, specifically, the HOMO and the LUMO. We also calculated the single-point energies of the systems using the B3LYP functional with the 6-31g(d) basis set and obtained the expansion coefficients of all orbitals to determine the TDOS of the systems. Finally, we used the Multiwfn program to plot the frontier molecular orbital and TDOS diagrams and calculate the ADCH charge of the GO or G fragments in the systems.
The J-V characteristics and steady-state current and PCE outputs of the PSCs were measured with a source measure unit (2450, Keithley) under AM1.5G 1-sun illumination (100 mW cm) generated with a solar simulator (Sol3A Class AAA, Oriel). J-V curves were obtained using the sweep mode of the reverse scan (from 1.20 to -0.1 V) and forward scan (from -0.1 to 1.20 V) with a voltage step of 0.02 V and a delay time of 10 ms. A metal mask was used to control the effective area of the C-PSCs to 0.075 cm during tests. For the large-area C-PSCs, a metal mask with an area of 1 cm was used. The surface chemical states of samples were analysed by XPS using a Kα X-ray photoelectron spectrometer (ESCALAB250Xi, Thermo Scientific). UPS was performed using a Kratos Analytical spectrometer (ESCALAB250Xi, Thermo Scientific) with monochromatic He Iα radiation at 21.2 eV. UV-visible absorption spectra were obtained using a UV-visible spectrophotometer (UV-3600 Plus, Shimadzu). CV measurements were conducted in a three-electrode cell on an electrochemical workstation (CHI700E, Shanghai Chenhua) over a scan range of -0.8 to 0.8 V. Ag/AgCl with a saturated KCl solution was used as the reference electrode and platinum rods were used as the counter electrode and work electrode with 0.1 M tetrabutylammonium hexafluorophosphate in acetonitrile as the electrolyte. Open-circuit potentials and Tafel polarization curves were also measured in a three-electrode cell. The test sample served as the working electrode and was placed parallel, face to face, with platinum foil, which served as the counter electrode. Ag/AgCl with a saturated KCl solution was used as the reference electrode. The crystalline properties of samples were analysed by XRD (SmartLab 9 kW, Cu Kα radiation, λ = 0.15418 nm). FTIR spectra were acquired in the range of 500-4,000 cm using an IR spectrometer. (6700, Thermo Fisher). The structural morphologies of samples were analysed using a laser confocal Raman microscope (DXR Microscope, Thermo Fisher) at a wavelength of 532 nm. ToF-SIMS was performed with an ultra-high-resolution mass spectrometer (Model ToF-SIMS 5, ION-ToF). For the ToF-SIMS measurements, all samples were heated at 85 °C for 24 h to accelerate the diffusion of ions. AFM and C-AFM images of the HTL films were obtained with a Bruker AFM instrument. PL and TRPL spectra were measured on a steady-state fluorescence and phosphorescence lifetime spectrometer (FLS1000, Edinburgh Instruments) operating at 450 nm and a pulse frequency of 0.2 MHz. The average power density was 0.15 mW at 20 MHz and the laser fluence was estimated to be ~5 nJ cm. CPD measurements were conducted using a scanning Kelvin probe (M470, Bio-logic). Electrochemical impedance spectroscopy was conducted on an electrochemical workstation (Zennium Zahner) at a bias voltage of 0.1 V with an alternative signal amplitude of 20 mV in the frequency range of 0.05 Hz to 1 MHz. The microstructures of samples were observed using a field-emission SEM (JSM-7610F Plus, Hitachi). EQE spectra were obtained using an EnliTech EQE measurement device. Resistance was measured using a four-probe tester (HPS2523, Helpass), which was calibrated against a standard resistance block (HPS4001, Changzhou Helpass Electronic Technology) before use.
For the light-soaking tests (ISOS-L-1 and ISOS-L-2 protocols), unencapsulated devices were maintained at 30 or 85 °C under a nitrogen atmosphere and subjected to continuous white-LED illumination (DR-L-BWL22/22, Derui Keyi). The light intensity (100 mW cm) was calibrated using a calibrated Si reference (National Institute of Metrology, China). Their photovoltaic performance parameters were tracked at the MPP under open-circuit voltage conditions. For the thermal stability test, unencapsulated devices were annealed at 30 or 85 °C under a nitrogen atmosphere for 500 h, with J-V curves measured every 100 h. For the thermal cycling test (ISOS-T-1), unencapsulated devices were subjected to thermal cycling in air between 25 °C (12 h) and 85 °C (12 h) at a ramp rate of 30 °C h in the dark. After every 5 cycles, up to 50 cycles, the devices were removed from the chamber for J-V measurement. For the dark storage test (ISOS-D-2 protocol), unencapsulated devices were maintained at a constant 85 °C in an ambient environment (20% relative humidity in the dark) throughout testing. J-V measurements were performed at 100 h intervals until 1,000 h.
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.