Interface Engineering in Perovskite Solar Cells by low concentration of PEAI solution in the antisolvent step

In spite of the outstanding properties of metal halide perovskites, its polycrystalline nature induces a wide range of structural defects that results in charge losses that affect the final device performance and stability. In this work, a surface treatment is used to passivate interfacial vacancies and improve moisture tolerance. A functional organic molecule, phenylethyl ammonium iodide (PEAI) salt, is dissolved with the antisolvent step. The additive used at low concentration does not induce formation of low-dimensional perovskites species. Instead, the organic halide species passivate the surface of the perovskite and grain boundaries, which results in an effective passivation. For sake of generality, this facile solution-processed synthesis was studied for halide perovskite with different compositions, the standard perovskite MAPbI3, and double cation perovskites, MA0.9Cs0.1PbI3 and

In spite of the outstanding properties of metal halide perovskites, its polycrystalline nature induces a wide range of structural defects that results in charge losses that affect the final device performance and stability.In this work, a surface treatment is used to passivate interfacial vacancies and improve moisture tolerance.A functional organic molecule, phenylethyl ammonium iodide (PEAI) salt, is dissolved with the antisolvent step.The additive used at low concentration does not induce formation of low-dimensional perovskites species.
Instead, the organic halide species passivate the surface of the perovskite and grain boundaries, which results in an effective passivation.For sake of generality, this facile solution-processed synthesis was studied for halide perovskite with different compositions, the standard perovskite MAPbI3, and double cation perovskites, MA0.9Cs0.1PbI3and

Introduction
Metal halide perovskite solar cells (PSCs) have attracted considerable attention due to their good light absorption, tunable band gap, long charge diffusion lengths, and low manufacturing costs. [1]To date, the highest power conversion efficiency (PCE) reported and certified has already exceeded 25.5 %, [2] being the most promising next-generation photovoltaic technologies in renewable energy.Interesting, this outstanding performance has been reached with polycrystalline thin films less demanding, from the industrial point view, [1b] than their crystalline counterparts.However, despite the benign defect physics of halide perovskites, the polycrystalline character of these materials leads to disorderly distribution of defects in the perovskite or at the grain boundaries, surfaces, and interfaces, that turns in losses of on the power conversion efficiency (PCE) of perovskite solar cells (PSCs). [3]Therefore, advances in passivating-strategies are highly demanded and would ensure a viable commercial future for PSCs.Recently, significant attention has been paid to improve the stability of PSCs via passivation of grain-boundary and interface engineering. [4] This iodide salt is composed by two components, a bulky organic chain with a benzene ring and an ammonium group (R-NH3 + ) which represents PEA + .Added to perovskite precursors, the large PEA + cation causes anionic layers in the 3D architecture to be isolated and transform into a 2D perovskite of general formula (RNH3)2An−1BnX3n+1 (n=1) [6] or into a quasi-3D perovskite compositions. [7] In addition, PEAI increases black phase stability, [10] as in all-inorganic perovskite CsPbI3 stabilizing the orthorhombic black phase (γ-CsPbI3) under ambient condition and to avoid the yellow -CsPbI3 formation, achieving high open circuit voltage over 1.3 V. [11] Moreover, pure 2D (PEA)2PbI4 perovskite films prepared by blade-coating has an highly crystalline nature [12] and pure 2D (PEA)2SnI4 prepared by sequential vapor process of PEAI/SnI2 shows less tendency to Sn oxidation. [13]sides, the PEA + cation has hydrophobic nature, which improves the moisture resistance of interfaces perovskite/transporting layers.In this sense, PEA + cation is considered an excellent additive if it is added properly in the 3D halide perovskite thin film.In fact, the PEAI salt was deposited as buffer layer in the PSCs [14] to control surface recombination.In particular, either on the top of the perovskite surface in order to retard the charge-carrier recombination process; [5b, 14b] or on the bottom of the perovskite layer to passivate defects of NiOx and enhance the interface contact properties [5a] were successfully employed in p-i-n configuration.
In all cases, PEAI-modified PSCs show better moisture resistance and superior thermal stability.14a] The above-mentioned study demonstrated a high PCE of 23.34 %.PEAI salt was also studied in carbon electrode-based PSCs without hole transport materials to improve the poor perovskite/carbon contact.Specially, PEAI was added in a post-treatment carbon electrode once the PSC was prepared and PEAI film was deposited between the perovskite/carbon as an ultrathin PEA2PbI4 layer. [15]Due to the hydrophobic nature of carbon and 2D perovskite layers a large cell stability over 1000 h of exposure to ambient conditions was achieved.
However, this use of PEAI requires an additional fabrication step.Here, we propose to remove this further step by incorporating PEAI during the antisolvent step.The antisolvent additive engineering is another strategy to improve the perovskite structural properties and several additives have been used with success. [16]However, PEAI has been less explored in these conditions. [17]Bai et al. demonstrated a novel solution process to growth in situ a 2D layer together with 3D perovskite in order to suppress ion migration in the device and enhance the cell ambient stability.However, when 2D perovskite is synthetized, a balance between stability and efficiency is necessary due to the lower carrier mobility of 2D structure compared to 3D perovskites. [18]rein, we delve into the study of the PEAI effects caused in several metal halide compositions when the organic iodide salt is dissolved in the antisolvent at low concentration (0.0002-0.012M).5a, 6- 15] Instead, the PEAI is introduced in low concentrations.As far as we know, there is not references using this additive in such low amount, that it is an advantage in terms of material waste optimization.In order to study the generality of the approach, three different halide perovskite compositions, MAPbI3, MA0.9Cs0.1PbI3,and MA0.5FA0.5PbI3,and two different solar cell structures, regular n-i-p and inverted p-i-n configurations, were deeply investigated and characterized.These materials have been selected as the two formers are broadly studied and present a tetragonal structure and the later because it presents a cubic structure, again for sake of the generality of the analysis.It was found that the introduction of the PEAI salt in the perovskite during the antisolvent step has an impact on reducing the crystallite domain size for all structures, as confirmed by XRD, and stabilizes their optical absorption up to 1300 h in ambient conditions.The role of the additive in the solar cells induces an increase of the PCE in all cases, as results of the reduction of the charge recombination processes confirmed by impedance spectroscopy.

Perovskite film
We exploit the antisolvent additive engineering approach [16a] to introduce the PEAI salt at very low concentrations, 0.0002-0.012M (see Experimental procedure for more details), in three representative metal halide perovskites compositions, mainly standard MAPbI3, and double cation MA0.9Cs0.1PbI3,and MA0.5FA0.5PbI3.Henceforth, these perovskites are called MA, MA0.9Cs0.1, and MA0.5FA0.5, respectively.Figure 1a-d shows the room temperature x-ray powder diffraction (XRD) profiles for the samples with different A-site compositions in the ABX3 perovskite structure, with and without PEAI addition in the synthesis.The XRD data for all samples indicate that perovskite polycrystalline films were obtained for all compositions.The XRD profiles were indexed to an I4cm tetragonal cell for MA and MA0.9Cs0.1 since the small MA cation gives rise to tetragonal structures with elongation in the c axis.But changing the small MA cation for the larger FA cation (case MA0.5FA0.5)increases the Goldschmidt tolerance factor (t) beyond the tetragonal limit (tlim = 0.972), [19] forming a cubic cell.In this case, MA0.5FA0.5 perovskite has been indexed to a Pm-3m cubic cell.The halide perovskite lattice parameters were refined by a profile matching approach (see Figure S1).For a better comparison of the obtained values, tetragonal parameters could be converted to pseudo-cubic parameters according to these equations:   =   √2 ⁄ and   =   2 ⁄ .The lattice parameters as a function of the A-site composition could be seen in Figure 1e.In this figure, it can be seen how the presence of PEAI slightly increases the lattice parameters for the tetragonal compositions (MA and MA0.9Cs0.1),while it does not affect the cubic composition (MA0.5FA0.5).The perovskite peak phase at 14.1º, which corresponds to the (110) reflection for MA and MA0.9Cs0.1 and to the (100) reflection for MA0.5FA0.5 have been followed over 1500 h by XRD of the perovskite layers under ambient conditions (see Figure S3).Interestingly, perovskite phase hardly changes with time and remains almost constant with the presence of PEAI.
The top-view scanning electron microscopy (SEM) images were carried out (see Figure S4).
We found that the perovskite morphology was significantly altered by adding PEAI in the MAPbI3 and MA0.9Cs0.1PbI3perovskites during the antisolvent, decreasing the grain size in both samples.At this point, it has to be noted that reducing the grain size is expected to be detrimental to efficiency due to recombination sites, [20] as grain boundaries are the major recombination sites in iodide-based perovskites. [21]However other effects should be considered as the grain boundary passivation.The result of a trade-off between these two effects can cause that lower size grains produce high performance devices as we observed by the addition of PEAI into the antisolvent step.The phase purity of the synthesized perovskite samples of bare MAPbI3 perovskite and MAPbI3 with iodide salt PEAI were verified by proton nuclear magnetic resonance ( 1 H NMR) spectroscopy.In order to have a net correspondence between the characterizations, we dissolved the prepared films in the proper deuterated solvent (DMSO-d6).Due to this preparation method the samples were quite diluted (Figure S5a) and the solvent peaks are predominant (DMSO and residual water).In Figures S5b and S5c it is clear the presence of the singlet corresponding to the ammonium group at 7.46 ppm and of the singlet from the aliphatic hydrogen of the MAI at 2.36 ppm respectively, [22] but there is no detectable trace of PEAI neither at high or low frequencies of the protonic spectra. [23]Each characteristic peak mentioned above is independent of the PEAI addition.Therefore, both samples maintain similar chemical environments, thus indicating that the PEAI is not incorporated into the perovskite structure.Considering on the other hand the effect in the morphology these studies point out to the grain boundaries for PEAI location.
In spite of the significant morphology changes induced by PEAI, in-plane DC dark resistivity 3.6•10 6 and 0.9•10 6 •cm, respectively, in accordance with previous works for MAPbI3 polycrystalline samples. [24]e optical absorption spectra of bare perovskite and PEAI-doped perovskite samples are displayed in Figure S7.The local minima of the second derivatives of the optical density (O.D.) spectra is useful to estimate the optical transition energies, [25] which in this case, the first optical transition are located at 1.65  0.04 eV for MA and MA0.9Cs0.1, and 1.59  0.04 eV for MA0.5FA0.5, i.e. the optical band gap Eg, corresponding to the described direct semiconductor type transitions at the R point in the pseudo-cubic Brillouin zone for perovskites. [26]The full width at half-maximum of the second derivative is assumed as the error.PEAI addition does not modify the optical band gap Eg, indicating that there are no changes in the stoichiometry of the synthesized thin films, in line with the NMR characterization.However, we found that the addition of PEAI has an important effect on the evolution of the absorption spectra under ambient conditions.In particular, the absorption coefficient remains constant up to 2.3 eV, which in concordance with the perovskite phase time evolution measured by XRD (Figure S3 shows (110) peak evolution).Thus, it is confirmed that the PEAI salt stabilize the perovskite absorption in thin films over time, up to 1200 h, independently of the perovskite analyzed (Figure S8), especially in the MA0.5FA0.5 thin films (Figure 2), possibly helped by the more packed morphology.

Perovskite device
To study the influence of PEAI salt on the photovoltaic performance, two types of solar cells were fabricated (see Figure 3a), in regular and inverted architectures to further generalize the  For the optimal PSCs, the J-V characteristics under a simulated air mass (AM) of 1 sun illumination (100 mW/cm 2 ) are shown in Figure 3b, and the corresponding photovoltaic parameters are summarized in Table 1.As we expected from the optical band gap Eg, the MA0.5FA0.5-basedsolar cells displayed lower open-circuit voltage, Voc, close to 0.78 V, and an enhancement of Voc for MA-and MA0.9Cs0.1-basedsolar cells over 1 V, however these values are also influence by the architecture and recombination processes.Interestingly, a Voc increment is observed for all PEAI-based PSCs, which could be associated with a lower recombination rate, [28]  Interestingly, for MA and MA0.9Cs0.1, the most efficient ones, average PCE is more significantly enhanced by 18 and 32 %, respectively.
In order to further investigate the effect of PEAI treatment, impedance spectroscopy has been measured at open circuit conditions with different light intensities for MA-and MA0.9Cs0.1basedsolar cells. [29]The Nyquist plots at Voc of impedance spectra at 1 sun are represented in Error!No s'ha trobat l'origen de la referència.a.As it is conventionally observed in high efficient PSCs, we observed mainly two different semicircles, at low frequency (LF) and high frequency (HF) ranges.In both devices measured it is observed a decrease in width of the arc when PEAI is used which indicates a lower resistance or better transport due to the presence of the cation at grain boundary as mentioned before which will result in higher PCE.In order to analyze more careful these data, an equivalent circuit model previously reported has been used and it is represented in Figure 4b. [30]In particular, the recombination resistance, Rrec, see  is in parallel to a HF capacitance, Cg, the geometrical capacitance, an in parallel with a RC branch (Cdr and Rdr), that splits the pattern in the HF and LF arcs, [32] and c) recombination resistance at open circuit conditions for the PSCs w/o and w/PEAI in regular architecture.

Conclusions
In summary, we have demonstrated that adding PEAI salt at low concentration during the antisolvent step is one alternative to improve perovskite solar cells performance, without increasing the number of fabrication steps.The proposed method is tested in three representative compositions, the standard perovskite MAPbI3, and double cation perovskites, MA0.9Cs0.1PbI3and MA0.5FA0.5PbI3,and both, regular and inverted solar cells.PEAI reduces the crystallite domain size, enhancing the morphology.XRD and optical absorption confirms a stabilization effect of the perovskite phase in thin films for all compositions.For solar devices, the addition of PEAI leads to a PCEs increase, respect reference cells without treatment, by a very significant 18 and 32 % for MA and MA0.9Cs0.1 respectively and by a 4 % for MA0.5FA0.5, and to a reduction of the recombination rate in all the cases.Besides, from NMR analysis, PEAI is not incorporated into the halide perovskite structure, reinforcing their passivating role at the grain boundaries, reducing traps states which results in the observed higher Voc.Those findings found for a broad range of halide perovskite materials, with both tetragonal and cubic phase, and for both regular and inverted architectures, highlighting the generality of the simple approach.Decreasing the number of fabrication steps would have an important effect in the environmental impacts and cost of large-scale perovskite fabrication.

Device fabrication n-i-p
Structure ITO/SnO2/MAPbI3 or MA0.9Cs0.1PbI3/spiro-OMeTAD/Au.Glass substrates coated with ITO were etched by using zinc powder and pour over it drops of HCl 2 M.Then, glass was cleaned with deionized water, acetone and ethanol in an ultrasonic cleaner for 15 min for each solvent one after the other.After being dried by air flow, the substrates were treated in Finally, Au electrode was deposited by thermal evaporation with a thickness of 80 nm.

Device fabrication p-i-n
Structure ITO/PEDOT:PSS/MA0.5FA0.5PbI3/PCBM/BCP/Al.Glass substrates patterned with ITO were treated in the same way than in the proceeding described previously.Once the substrates were completely cleaned, the next steps were carried out in the glove box filled with nitrogen until end of the procedure.A diluted PEDOT:PSS solution (PEDOT:PSS:2propanol, 5:1, v:v) were deposited by spin coater 4000 rpm for 45 s.Then, the HTL were heated at 130 ºC for 30 min.The perovskite solution (150 L) was dropped onto the PEDOT:PSS film and then spin coated at 1000 rpm for 10 s and then, at 6000 rpm for 30 s.
As has been mentioned, the anti-solvent method used includes at 15 s of the second step add 450 L either EA or PEAI filtered solution (0.05 mg/mL solved with EA).The perovskite films were kept at room temperature for 30 min and then heated at 100 ºC for 40 min.The ETL was formed by spin coating a solution of PCBM in CB (40 mg/mL) of 2000 rpm for 40 s and then, the film was treated at 60 ºC for 10 min.On the top of the PCBM, a thin layer of BCP (5 mg/mL in 2-propanol) was added by spin coater as 5000 rpm for 40 s.Finally, a thermal evaporation was carried out to evaporate Al.

Film characterization
X-Ray Diffraction (XRD): The XRD diffractograms of the perovskite thin films were measured by Bruker D8 Advance diffractometer using Cu Kα radiation over a 2θ range between 5° and 65° with a step size of 0.02°.
Optical characterization: UV-Vis absorption of the thin films was characterized using a UV-VIS-NIR spectrophotometer (Varian, Cary 5000) in the wavelength range of 300-900 nm.

Steady state photoluminescence emission (PL): confocal PL emission was measured with an
inverted confocal microscope Leica TCS SP8 using an excitation wavelength of 561 nm with DPSS 561 laser employing a spectral resolution of 5 nm and 63 times of zoom image.

Electrical characterization:
In-plane DC dark resistivity-temperature dependence was obtained from current-voltage characteristics with two-wire configuration due to the high resistance of the samples and a homemade Faradaic box.The dependences with increasing temperature have been measured in the 298-373 K range, with 10 K step size.The samples (thin films on glass substrate) were located on a hot plate with a K-type thermocouple right beside the sample to a process controller Electemp-TFT (Selecta).Current-voltage curves were measured at each temperature after 3 min stabilization, using a probe station and a Keithley 2450 Sourcemeterb between two evaporated Aluminium contacts.
Scanning electron microscopy (SEM): SEM measurements were used to analyse the perovskite surface by employing a JEOL 7001F microscope with an electron gun of 0.1 -30 kV power which allow us to obtain an image at 20.000 times magnification.
Hydrogen Nuclear magnetic resonance (H-NMR): NMR measurements were taken using a spectrometer Bruker Avance III HD 400 MHz employing solutions dissolved in DMSO-d6 to analyze chemical shift between 0-10 ppm.

Device characterization
Current-voltage (J-V) measurement: J-V measurements were performed under AM 1.5G (100 mW/cm 2 ) conditions with a Wavelabs Sinus-70 AAA LED solar simulator with a Keithley 2450 sourcemeter.Each J-V curve was carried out from 0 V to 1.15 V (forward direction) in a scan rate of 90 mV/s without preconditioning.The spectrum of the solar simulator is monitored with reference intensity sensor in test plane in combination with fast feedback loop for automatic intensity correction and temperature control for the LEDs.The aging test was evaluated keeping the PSCs under dark and N2 atmosphere.
Incident photon to current efficiency (IPCE): IPCE measurements were performed using a Xenon lamp with a monochromator Oriel Cornestone 130 which was used to measure along the wavelength of the spectrum.Prior measurement, calibration was done using a reference photodiode of silicon and each measurement was obtained using TRACQ BASIC software.
Finally, EQE scans were taken from 300 nm to 810 nm every 10 nm.
Impedance spectroscopy: The IS was measured using a Potentiostat Autolab-PGSTAT204 in open circuit conditions.The light intensity was controlled by the Wavelabs Sinus-70 AAA LED solar simulator.For each light intensity, an AC 20 mV voltage perturbation was carried out and the frequency range was from 1 MHz to 100 mHz.Z-View software was used to fit the impedance spectra.
Supporting Information.Details of experiments and additional supplementary figures.This material is available free of charge via the Internet.

Figure 1 .
Figure 1.(a)-(c) XRD profiles (Cu Kα1/α2) of the MA, MA0.9Cs0.1 and MA0.5FA0.5 (with and 5 F A 0 . 5-P E A I The average crystal domain size has been obtained using the Williamson-Hall (WH) methodology (Figure S2).The domain size as a function of the A-site composition in the films and the PEAI addition is shown in Figure 1f.PEAI reduces the crystallite size, likely indicative of more packed layers.When PEAI is not used, the domain size of MA and MA0.5FA0.5 (ρ) remains almost unaltered in the three compositions.The ρ-dependence with temperature in solar cells operation range temperature was measured for MA, MA0.9Cs0.1, and MA0.5FA0.5 thin films and is shown in Figure S6.The resistivity for MA and MA0.9Cs0.1 compositions are slightly reduced and MA0.5FA0.5 perovskite remains unaltered with PEAI addition in the cubic regime.This reduction in ρ could be associated with an increase in the packing of the layer and an improvement in the grain interfaces that favours charge transport.ρ values at room temperature are 1.2•10 7 , 5.1•10 6 and 0.4•10 6 •cm for pristine MA, MA0.9Cs0.1, and MA0.5FA0.5 perovskites, respectively, and, for perovskite thin films PEAI-doped are 2.4•10 6 ,

Figure 3 .
Figure 3. a) Device architectures n-i-p and p-i-n, b) photovoltage performance for champion Figure S11), which reveals that the Jsc with a value of 20.23, 20.59, 20.19, and 20.38 mA/cm 2

Figure 4c ,
Figure 4c, has been obtained by the fitting of impedance measurements, as the sum of the resistances of high and low frequency arcs, considering transport resistance negligible.[30]It

Figure 4 .
Figure 4. a) Nyquist plots at open circuit voltage under 1 sun illumination intensity, b)

(
Pilkington TEC15, ∼ 15 Ω sq -1 ).Preparation of perovskite solutions: Three compositions have been synthesized, such as MAPbI3, MA0.9Cs0.1PbI3and MA0.5FA0.5PbI3.The solution preparation of each composition is as follow.Perovskite MAPbI3 was prepared with a mixture of perovskite precursors of PbI2 and MAI with a concentration of 1.4 M for each precursor and solved with a mixture of solvents, DMF:DMSO (4:1, v:v).Perovskite MA0.9Cs0.1PbI3follows the same trend than the perovskite standard MAPbI3 and the precursor CsI is added with a concentration of 1.4 M and the precursors were solved with pure DMSO.The perovskite composition MA0.5FA0.5PbI3with a concentration of 0.65 M was prepared with a mixture of PbI2, MAI and FAI and dissolved with DMF:DMSO (4:1, v:v).
an ultraviolet-ozone (UV-O3) for 15 min to remove organic residues.Once the substrates were cleaned, the electron transport layer ETL SnO2 was deposited in ambient at 25 °C and 30 % RH by preparing a solution of SnO2 3 % in water from the Alfa Aesar solution of 15 %.The ETL was spin-coated onto the ITO substrates with a speed of 3000 rpm for 30 s, and then heated at 150 °C for 30 min.Once the SnO2 film was prepared, the substrates were submitted to 20 min of UV-O3 previously to perovskite deposition inside glovebox.A quantity of 50 L of perovskite was deposited over SnO2 film by one-step spin coating at 4000 rpm for 20 s.At 8 s after starting the second step, an aliquot of 400 L was added, either toluene or PEAI filtered solution with a concentration of 0.05 mg/mL.As soon as the spin coating was finished, the sample was moved to a hotplate at 130 ºC for 10 min.After perovskite preparation, a 50 L of hole transporting layer HTL (spiro-OMeTAD) solved in chlorobenzene (85.5 mg/mL) doped with 28.8 μL of TBP and 17.8 μL of a stock solution of 520 mg/mL of Li-TFSI in acetonitrile was spin-coated at 4000 rpm for 20 s onto the top annealed perovskite layers.

Figure S4 .Figure
Figure S4.Top-view SEM images of MA and MA0.9Cs0.1 perovskites w/o and w/PEAI.The substrate was glass/ITO.The bar size is 1 m.

Figure S7 .Figure S8 .
Figure S7.a) Second derivative of the optical density of the thin films and b) optical absorption.c) Normalized PL emission of the thin films.Solid and dashed lines correspond to samples without and with PEAI, respectively.

Table 1 .
Photovoltaic parameters of the p-i-n and n-i-p PSCs based on PEAI salt.It is included the champion solar cell and the average values with standard errors from 10 devices fabricated in parallel.