Enhanced Photovoltaic Performance of Perovskite Solar Cells by Co-Doped Spinel Nickel Cobaltite Hole Transporting Layer

Solution combustion synthesized hole transport layer of co-doped spinel nickel cobaltite were fabricated using doctor-blading technique for planar inverted perovskite solar cells. Perovskite solar cells incorporating Co-doped spinel nickel cobaltite shown an increase in device performance parameters compared to un-doped spinel nickel cobaltite, leading to PCE of 16.5%. X-ray photoelectron spectroscopy measurements revealed the tendency of Cu cations to replace preferably the surface Ni atoms changing the surface stoichiometry of spinel nickel cobaltite inducing a cathodic polarization. Ultraviolet photoelectron spectroscopy measurements unveiled the increase of the ionization potential by 0.1 eV for co-doped spinel nickel cobaltite film compare to undoped spinel nickel cobaltite-based hole transporting layer. We attribute the enhanced PCE of inverted perovskite solar cells presented due to improved hole extraction properties of co-doped spinel nickel cobaltite hole transporting layer.

Regarding the investigation of functional hole transporting layers (HTLs), a wide variety of organic and inorganic materials have been implemented to improve hole extraction, with some of the latter's advantage to be the wide optical band gap (thus high transparency in the visible range) and superior hole mobility, while they can be solution processed.Some promising inorganic HTLs are NiOx, 34 Cu:NiOx [35][36][37] , CuOx [38][39][40] , CuI 41 , CuSCN, 42 CuGaO2, 43 CuCrO2 44 .Recently, we have reported combustion synthesis of monodispersed spinel NiCo2O4 nanoparticles of ~4 nm diameter forming a compact layer with electrical conductivity of ~4 S/cm.The developed films were applied as an efficient and reliable HTL for inverted structure perovskite solar cells (PVSCs) using 230 nm thick perovskite layer. 45In order to increase the PCE of the devices a thicker perovskite layer is needed.The enhancement in light absorption leads to an increase in photogenerated carriers which accumulate at the perovskite/HTL interface (accumulation zone) and subsequently collected by the contact 46 .Thus, HTL with enhanced hole collection capability are required to increase the PCE.A common method to enhance HTL charge collection efficiency is by incorporation of intentional defects through extrinsic doping.8][49][50][51] For example, recently a co-doping strategy of NiOx with Cu/Li or Li/Mg elements has been successfully applied to enhance the PCE of PVSC. 52,53 this paper, we report the use of solution combustion synthesized NiCo2O4 codoped with 3 mol% Cu and 2 mol% Li (3% Cu -2% Li) as efficient HTL to increase the performance of inverted PVSCs.Initially, NiCo2O4 film doped with 5 mol% Cu was incorporated as HTL in PVSC exhibiting an increased Voc.However, the Jsc of the corresponding PVSC has declined significantly compare to unmodified NiCo2O4based PVSC due to lower electrical conductivity.We show that an increase in the electrical conductivity can be achieved, by 3% Cu and 2% Li co-doping of the NiCo2O4-HTL resulting in PVSCs with enhancement on both Voc and Jsc compare to NiCo2O4-HTL based PVSCs.X-ray photoelectron spectroscopy (XPS) investigation on co-doped NiCo2O4-HTL showed a decrease on the Ni/Co atomic ratio compared to unmodified NiCo2O4-HTL, indicating the preferable surface substitution of nickel by copper cations which has been previously reported that it induces a cathodic polarization. 54As a result, an increase of the ionization potential by 0.1 eV was observed for 3% Cu -2% Li NiCo2O4-HTLs compare to stoichiometric NiCo2O4-HTL using ultraviolet photoelectron spectroscopy (UPS).The increased performance of the reported PVSCs could be attributed to the cathodic polarization potential and thus better hole collection efficient of the 3% Cu -2% Li NiCo2O4 layer.
The PVSCs under investigation were prepared on top of glass/Indium Tin Oxide (ITO)/NiCo2O4 for the different doping types processed as described in detail at the supplementary section.The perovskite solution was prepared 30 min prior spin coating by mixing Pb(CH3CO2)2.3H2O:methylamoniumiodide (1:3) at 40 wt% in dimethylformamide (DMF) with the addition of 1.5% mole of MABr (methylammonium bromide).Briefly, ~350 nm perovskite active layer was spin-coated on top of each substrate followed by 50 nm spin-coated PC70BM (serving as the electron selective contact) and 100 nm thermally deposited Al.More details for the materials and processing conditions can be found within the supplementary section.In order to investigate the reduced photocurrent of 5% Cu-doped NiCo2O4, we first excluded any possible optical losses induced by the doping.Figure 1(a) demonstrates the transmittance of ~15 nm-thick NiCo2O4 layer on glass/ITO.It is obvious that the difference on transmittance is negligible for all films under study, where the extracted Tauc-plot (Fig. S1) for direct transitions [ ( α.E ) 2 = A.( E -Eg) ] show similar optical band gaps (Egs).Further, the similar morphology in all types of NiCo2O4 films was confirmed excluding, also, differences in electrical losses related to films quality (e.g.shunting current).Fig. S2 and Fig. S3 illustrates the AFM topography images of (a) 5% Cu and (b) 3% Cu -2% Li NiCo2O4 films fabricated on quartz and glass/ITO substrates, while Fig. S3(c) illustrates the topography of the ITO underlayer.In both cases the films exhibit similar roughness between them ( 0.7 -0.8 nm for quartz and 2.9 -3.0 nm for glass/ITO substrate) comparable to the ones measured for the pristine NiCo2O4 films, affirming the similar quality of different types of NiCo2O4 films. 45us, electrical characterization of PVSC were performed using Electro impedance spectroscopy (EIS) measurements under illumination and zero bias on the previous described PVSCs configurations.As it is observed at Figure 1(c) all the spectra show the characteristic two frequency response, where the first arc (higher frequencies) is ascribed to charge transfer resistance (Rct) while the second larger arc (lower frequencies) at the charge carrier recombination resistance (Rrec).recombination model, the slope between logarithmic light intensity and Voc must be equal to 2kT/q for trap-assisted and kT/q for trap-free recombination [60][61][62][63][64] .As shown in Fig. 2(a), the Voc -light intensity curves scale equal to kT/q, implying that a trap-free recombination mechanism is dominant for all the PVSCs within this paper.Thus, steady state photoluminescence (PL) measurements (Fig. 2(b)) are adequate to evaluate the degree of charge recombination at each configuration.The PL intensity for undoped NiCo2O4-HTL is much higher compared to 3% Cu -2% Li NiCo2O4-HTL, implying that a much higher number of electron-hole pairs recombine for the case of the undoped HTL justifying the lower PCE of the corresponding undoped NiCo2O4-HTL based PVSCs.The experimental results presented indicate that 3% Cu -2% Li NiCo2O4 HTL transfers and collects hole charges more efficient than the undoped NiCo2O4 HTL.
A deeper material properties and device physics investigation was performed to better understand the origin of the enhanced hole collection properties for 3% Cu -2% Li NiCo2O4.Structural characterization with X-ray diffraction (XRD) on the corresponding NiCo2O4 samples (Fig. S6) matched the cubic face-centered lattice structure of NiCo2O4 (PDF#20-0781), implying single-crystalline structure.X-ray photoelectron spectroscopy (XPS) measurements were also performed on doped and undoped NiCo2O4 HTLs.The Co 2p spectrum (Fig. S7) was best fitted by using two spin-orbit doublets for the tetrahedral Co +2 and octahedral Co +3 oxidation states and with two shake-up satellites located at the higher binding energy (BE) side of the main peaks.The peak located around 779.7 eV can be attributed to the octahedral Co +3 observed in Co3O4, 65 while the higher binging energy peak around 780.9 eV can be assigned to the tetrahedral Co +2 similar to CoO. 66 The spectrum of the Ni 2p3/2 region was fitted using three components (Fig. S8).The peak at 854.3 eV corresponds to Ni +2 ions, while that at 856.0 eV is attributed to Ni +3 . 65,67The shake-up satellite at around 861.8 eV was fitted considering one broad line.For Cu doped films the Cu 2p spectra were recoded and are displayed in Fig. S9.The Cu 2p doublet is well resolved.The Cu 2p3/2 peak at 934.6 eV and the satellite at higher binding energies indicate that Cu is oxidized and can be identified as Cu +2 ions in octahedral coordination. 54,67,68The intensity of Cu 2p3/2 peak for the 3% Cu -2% Li NiCo2O4 is low and the satellite structure is not resolved.Nevertheless, the peak is located at BEs around 934.Additional ultraviolet photoelectron spectroscopy (UPS) measurements were also performed on doped and undoped NiCo2O4 films to determine the energy levels.

Enhanced Photovoltaic Performance of Perovskite Solar Cells by Co-Doped Spinel Nickel Cobaltite Hole Transporting Layer
Cu -2 mol% Li respectively.Then, 150 uL HNO3 (69 wt% HNO3) were added slowly into the mixture, and the solution stirred up to almost complete homogeneity.The whole solution was allowed under stirring for 30 min at 60 °C.The ratio of the total metal nitrates and tartaric acid was 1.Thereafter, the violet colored solution was used for the combustion synthesis of the NiCo2O4 NPs on the various substrates.Doctor blade technique was applied for the fabrication of the precursor films on the various substrates.The resulting light violet colored films were dried at 100 °C for 30 min and used as a precursor for the combustion synthesis of NiCo2O4 NPs.Subsequently, the obtained films were heated at 250 °C in ambient atmosphere for 1 h in a preheated oven to complete the combustion process and then left to cool down at room temperature.

Device Fabrication:
The

Figure 1 (
Figure 1(b) demonstrates the current density -voltage (J -V) measurements Schottky (Fig.1(d)) measurements were carried out on devices sweeping from higher to lower voltage under dark conditions.The crossing of the curves at 1/C 2 = 0 is attributed to the flat band potential of the device.58,595% Cu and 3% Cu -2% Li NiCo2O4 -HTL based PVSCs show a higher built-in potential compare to unmodified NiCo2O4-HTL based PVSCs which is consistent with the increased Voc value achieved for the 5% Cu and 3% Cu -2% Li NiCo2O4 -HTL based PVSCs.. Further investigation of the charge carrier recombination dynamics was conducted to elucidate the enhanced device performance of 3% Cu -2% Li doped compare to undoped NiCo2O4-HTL based PVSCs.We first exclude any difference on the perovskite film morphology.AFM topography images (Fig.S4) of perovskite surface revealed similar surface roughness (12.5 ± 0.4 nm) and grain sizes (ca.110-123 nm) as shown within the paper supplementary information (Fig.S5), indicating that PVSCs under study comprise similar morphology within the active layer.Moreover, Voc -light intensity measurements were performed to investigate the recombination mechanism within PVSCs under study.According to simplified Shockley Reed Hall

Fig. 2 (
Fig. 2 (a) Voclight intensity measurements of PVSC using 15 nm-sized undoped and /PC[70]BM/Al.ITO substrates were sonicated in acetone and subsequently in isopropanol for 10 min and heated at 100 °C on a hot plate 10 min before use.The perovskite solution was prepared 30 min prior spin coating by mixing Pb(CH3CO2)2.3H2O:methylamoniumiodide (1:3) at 40 wt% in dimethylformamide (DMF) with the addition of 1.5% mole of MABr.The precursor was filtered with 0.1 µm polytetrafluoroethylene (PTFE) filters.The perovskite precursor solution was deposited on the HTLs by static spin coating at 4000 rpm for 60 s and annealed for 5 min at 85 °C, resulting in a film with a thickness of ~350 nm.The PC[70]BM solution, 20 mg mL −1 in chlorobenzene, was dynamically spin coated on the perovskite layer at 1000 rpm for 30 s. Finally, 100 nm Al layers were thermally evaporated through a shadow mask to finalize the devices giving an active area of 0.9 mm 2 .Encapsulation was applied directly after evaporation in the glove box using a glass coverslip and an Ossila E131 encapsulation epoxy resin activated by 365 nm UV irradiation.Characterization:For UV-vis absorption, XRD, four-point probe and AFM measurements the films were fabricated on quartz substrates.Transmittance, AFM, XPS and UPS measurements were conducted on films fabricated on glass/ITO substrates.XRD patterns were collected on a PANanalytical X'pert Pro MPD powder diffractometer (40 kV, 45 mA) using Cu Kα radiation (λ = 1.5418Å).Transmittance and absorption measurements were performed with a Schimadzu UV-2700 UV-vis spectrophotometer.The thickness of the films was measured with a Veeco Dektak 150 profilometer.The current densityvoltage (J/V) and Voc-intensity were obtained using a Botest LIV Functionality Test System measured with 10 mV voltage steps and 40 ms of delay time.For illumination, a calibrated Newport Solar simulator equipped with a Xe lamp was used, providing an AM1.5G spectrum at 100 mW cm −2 as measured by a certified oriel 91150 V calibration cell.A shadow mask was attached to each device prior to measurements to accurately define the corresponding device area.Steady-state PL experiments were performed on a Fluorolog-3 Horiba Jobin Yvon spectrometer based on an iHR320 monochromator equipped with a visible photomultiplier tube (Horiba TBX-04 module).The PL was non-resonantly excited at 550 nm with the line of a 5 mW Oxxius laser diode.AFM images were obtained using a Nanosurf easy scan 2 controller under the tapping mode.Electrical conductivity measurements were performed using a four-point microposition probe, Jandel MODEL RM3000.EIS and MS measurements were performed using a Metrohm Autolab PGSTAT 302N, where for the EIS a red light-emitting diode (at 625 nm) was used as the light source calibrated to 100 mW cm −2 .For EIS a small AC perturbation of 20 mV was applied to the devices, and the different current output was measured throughout a frequency range of 1 MHz to 1 Hz.The steady state DC bias was kept at 0 V throughout the EIS experiments.X-ray photoelectron spectra (XPS) and Ultraviolet Photoelectron Spectra (UPS) were recorded by using a Leybold EA-11 electron analyzer operating in constant energy mode at a pass energy of 100 eV and at a constant retard ratio of 4 eV for XPS and UPS, respectively.The spectrometer energy scale was calibrated by the Au 4f7/2 core level binding energy, BE, (84.0 ± 0.1 eV) and the energy scale of the UPS measurements was referenced to the Fermi level position of Au at a binding energy of 0 eV.All binding energies were referred to the C 1s peak at 284.8 eV of surface adventitious carbon.The X-ray source for all measurements was a non-monochromatized Al Kα line at 1486.6 eV (12 keV with 20 mA anode current).For UPS measurements, the He I (21.22 eV) excitation line was used.A negative bias of 12.22 V was applied to the samples during UPS measurements in order to separate secondary electrons originating from the sample and the spectrometer.The sample work function was determined by subtracting the high binding energy cut-off from the He I excitation energy(21.22eV).The position of the high-energy cut-off was determined by the intersection of a linear fit of the high binding portion of the spectrum with the background.Similarly, the valence band maximum is determined with respect to the Fermi level, from the linear extrapolation of the valence band edge to the background.

Figure S5 .
Figure S5.Distribution of perovskite grain size and the extracted parameters of mean

Table I
54V, thus even for lower concentration of Cu there are Cu +2 ions.The table III summarize the Ni:Co atomic ratio values obtained from the processing of the reported XPS spectra.The surface sensitivity of XPS and material precursor stoichiometry reveals that a small excess of Ni ions is identified at the surface of the undoped NiCo2O4 as the XPS calculated Ni:Co ratio is 0.55.For 5% Cu doped and 3% Cu -2% Li NiCo2O4 a decrease at Ni:Co ratio confirms the deficiency of Ni ion at the surface, resulting in 0.43 and 0.45 ratios, respectively, which have been preferentially replaced by the Cu ions.These findings agree with previous reported results of A.C. Tavaresa et al.54where the introduction of Cu replace surface Ni ions at the NiCo2O4 electrodes, which indeed induces a similar effect to cathodic polarization (downshift of the energy bands).