Phosphonic Acid Modification of the Electron Selective Contact

Mar 29, 2019 - Massachusetts Institute of Technology , 77 Massachusetts Avenue, ... School of Materials Science and Engineering, Georgia Institute of ...
0 downloads 0 Views 775KB Size
Subscriber access provided by UNIV AUTONOMA DE COAHUILA UADEC

Letter

Phosphonic Acid Modification of the Electron Selective Contact: Interfacial Effects in Perovskite Solar Cells Rebecca B. M. Hill, Silver Hamill Turren Cruz, Federico Pulvirenti, Wolfgang Tress, Sarah Wieghold, Shijing Sun, Lea Nienhaus, Moungi G. Bawendi, Tonio Buonassisi, Stephen Barlow, Anders Hagfeldt, Seth R. Marder, and Juan-Pablo Correa-Baena ACS Appl. Energy Mater., Just Accepted Manuscript • DOI: 10.1021/acsaem.9b00141 • Publication Date (Web): 29 Mar 2019 Downloaded from http://pubs.acs.org on March 29, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Energy Materials

Phosphonic Acid Modification of the Electron Selective Contact: Interfacial Effects in Perovskite Solar Cells Rebecca B. M. Hill,1 Silver-Hamill Turren-Cruz,2 Federico Pulvirenti,1 Wolfgang R. Tress,2 Sarah Wieghold,3 Shijing Sun,3 Lea Nienhaus,3 Moungi Bawendi,3 Tonio Buonassisi,3 Stephen Barlow,1 Anders Hagfeldt2*, Seth R. Marder1*, Juan-Pablo Correa-Baena2,3,4*

Affiliations: 1Center for Organic Photonics and Electronics (COPE) and School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, Georgia 30332-0400, United States. 2Laboratory of Photomolecular Science (LSPM) and Institute of Chemical Sciences and Engineering, Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Vaud CH-1015, Switzerland. 3Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139, United States. 4School of Materials Science and Engineering, Georgia Institute of Technology, Atlanta, GA, 30332, USA

*Correspondence to: JPCB [email protected], [email protected].

AH

[email protected],

SM

Keywords: perovskite solar cells, interfaces, open-circuit voltage, hysteresis, solar cell stability.

ACS Paragon Plus Environment

1

ACS Applied Energy Materials 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 20

ABSTRACT

The role electron-transport layers (ETLs) play in perovskite solar cells (PSCs) is still widely debated. Conduction band alignment at the perovskite/ETL interface has been suggested to be important role for the performance of the solar cells. However, little is known about the effects of work-function shifts on the solar-cell performance, and specifically, the open-circuit voltage (VOC). Here the effects of surface modification of SnO2 ETLs using polar phosphonic acids are investigated, including the effects on work-function, surface energy, device performance, and device stability in inert atmosphere. The phosphonic acid modifications did not have a large effect on VOC; however, a sharp decrease in overall device performance was found, mostly due to reduced fill factors. When exposed to conditions of low oxygen concentration, the phosphonic acidsurface-modified devices yielded current-voltage (J-V) curves with considerably lower hysteresis than those based on unmodified SnO2. This suggests that this modification method may be valuable for achieving stabilized PCE without hysteresis.

ACS Paragon Plus Environment

2

Page 3 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Energy Materials

Organic-inorganic perovskite solar cells (PSCs) are a promising emerging technology, with the power conversion efficiencies (PCEs) currently reaching values above 23%.1 Electron-transport layers (ETLs) and ETL/perovskite interfaces are an important area of research, 2,3 as they can have a large effect upon charge collection,4,5 and, in the case of n-i-p structures, perovskite crystal growth and overall device stability.6 The performance of planar perovskite devices has been shown to be especially affected by the ETL/perovskite interface.2 SnO2 has potential advantages over TiO2 as an ETL for PSCs due to its improved UV stability.7,8 The SnO2 layer has been used as an ETL in PSCs,9 and spin-coated SnO2 ETLs have also been developed.10 SnO2 ETLs can be processed at low temperatures (~120 °C) suitable for flexible substrates. The metal oxide-perovskite band alignment for SnO2 may be responsible for the improved performance seen relative to TiO2.9 SnO2 ETLs also exhibit enhanced stability in dry air, but exhibit degradation of performance after exposure to inert atmosphere. This sensitivity to inert atmosphere may indicate loss of oxygen from the SnO2 surface as molecular oxygen, thus modifying the work-function of this layer and/or introducing trap states at the interface. Phosphonic acid (PA) surface modifiers have been used in organic solar cells for the modification of indium tin oxide (ITO) and ZnO to impart increased stability,11 improved open-circuit voltage,12 and suppressed surface recombination.13 For PA-modified ZnO in organic solar cells, the increase in stability seen was attributed to passivation of the ZnO surface to molecular oxygen adsorption.14 Phosphonic acids have been previously used in PSCs by Guerra et al. on a TiO2 ETL.15 A slight increase in PCE was seen from a decrease in VOC and an increase in JSC, when using the modifiers. However, there is still no consensus of the role that work-function plays in determining VOC.

ACS Paragon Plus Environment

3

ACS Applied Energy Materials 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 20

Here, we report the effects of phosphonic acid modification on SnO2 ETLs, by atomic layer deposition (ALD), in PSCs. In particular, we investigate its effects on the oxide work-function, surface energy, perovskite morphology, and solar-cell performance and stability. Gold Spiro-OMeTAD

Perovskite

BPA

pCN-BPA

DEA-PCNVPA

SnO2 FTO

Figure 1. Schematic of the perovskite solar cell stack used in this work. The phosphonic acid modifiers are adsorbed by the SnO2 and the perovskite photoabsorber is deposited atop the modified ETL. The SnO2 (15 nm by ALD) on fluorine doped-SnO2 was treated with three phosphonic acids, BPA, pCN-BPA, and DEA-P-CNVPA. To modify the work-function of the SnO2 ETL, three phosphonic acids (PAs) were used: benzylphosphonic acid (BPA), (4-cyanobenzyl)phosphonic acid (pCN-BPA), and (E)-(1-cyano-2(4-(diethylamino)phenyl)vinyl)phosphonic acid (DEA-P-CNVPA). They were expected to shift work-function in varying ways based on their molecular dipoles; specifically the BPA, pCN-BPA, and DEA-P-CNVPA were expected to slightly increase, substantially increase, and decrease the work-function, respectively. For all the following studies, SnO2 was UV-ozone cleaned and modified by dipping in a 0.05 mM phosphonic acid solution in ethanol for 3 hours and rinsed with EtOH. While other dipping times where explored, it was found that less than 1 min did not have

ACS Paragon Plus Environment

4

Page 5 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Energy Materials

any affect in on the electronic or performance metrics, which may indicate that this short time is not enough for surface modification. Very similar metrics were shown for samples that were dipped for more than 1 min, and therefore we used 3 hr of dipping time throughout this study, as it will be discussed below. Solar cells were then prepared (Figure 1) by depositing a multicomponent perovskite, containing Cs, methylammonium (MA), formamidinium (FA), Pb, iodine, and bromine. A hole-transport layer (Spiro-OMeTAD) and a gold top electrode were subsequently deposited on the perovskite absorber.

Figure 2. Work-function changes upon modification of SnO2. a) Secondary electron edge onset, and b) work-function of the control and phosphonic acid-modified SnO2 substrates. c) The band energetics of perovskites adapted from Correa-Baena et al.9 and the work-function of the SnO2 ETLs studied. Prior to measurement, the control was treated with UV-Ozone for 15 min, while the SnO2 was modified with 0.05 mM of phosphonic acid in ethanol for 3 h.

ACS Paragon Plus Environment

5

ACS Applied Energy Materials 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 20

To understand the effects of PA modification on the work-function of the SnO2 ETLs, we used ultraviolet photoelectron spectroscopy. The work-function values were acquired from the secondary electron edge shown in Figure 2a and are summarized for all measured samples in Figure 2b (see Table S1 for details). The bare SnO2 samples with UV ozone treatment exhibit a work-function of 4.8 ± 0.1 eV. As expected, surface modification with pCN-BPA, the electronwithdrawing cyano group is oriented such that the negative end of associated dipole points away from the modified surface, exhibited the largest work-function of the modified substrates at 4.5 ± 0.1 eV. Surface modification with BPA produced an intermediate work-function change (4.3 ± 0.1 eV), while modification by DEA-P-CNVPA, in which a donor-acceptor push-pull π system with an amino donor and cyanovinyl acceptor leads to a dipole oriented with its negative end toward the surface, attained the lowest work-function at 4.1 ± 0.1 eV. Figure 2c shows the work-function values for each sample, and compares these values to the perovskite band energies with respect to the vacuum level9 to demonstrate on the potential effects of these work-function shifts. Based on these changes of up to 0.20 eV in the work-function of pCN-BPA with respect to BPA, one might expect noticeable changes in the device VOC if this modified potential is retained upon deposition of the perovskite. It is well known that the work-function of some metal oxides, such as ITO, can be decreased upon being washed by a solvent,16 and it is unclear how the solution deposition of a perovskite will modify the work-function of 4.8 eV for bare SnO2. However, if we assume that its work-function is retained upon perovskite deposition, the VOC effect should be even more pronounced if BPA is compared with the control SnO2. Decreasing the work-function of the ETL toward the conduction band of the perovskite, as is the case for BPA, should increase the VOC, as the increased built-in potential of the solar cell could lead to reduced interface recombination.17 Based on this hypothesis, the BPA-modified SnO2 is expected to have the highest VOC, as its work-

ACS Paragon Plus Environment

6

Page 7 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Energy Materials

function is positioned closest to the conduction band of the perovskite. In contrast, pCN-BPA, with a work-function 0.25 eV below the conduction band of the perovskite, is expected to have a lower built-in potential, and a lower VOC. It is important to note that potentials tend to be screened in the perovskite due to mobile ions.

Figure 3. Perovskite solar-cell characteristics of the materials studied. Distribution of photovoltaic results a) open-circuit voltage, b) short-circuit density, c) fill factor, and d) power conversion efficiency for SnO2 modified with 0.05 mM phosphonic acids. Prior to device fabrication, the control SnO2 substrate was treated with UV-Ozone for 15 min, while the modified substrate was treated with 0.05 mM in ethanol for 3 h. To understand whether different work-functions can lead to changes in VOC and overall performance, PSCs were prepared. A 500 nm-thick perovskite layer was deposited by spin coating the perovskite solution based on cesium, methylammonium, and formamidinium cations and iodide/bromide formulation, and an antisolvent procedure.13,18–20 The perovskite on top of the modified SnO2 was examined via atomic force microscopy (AFM)21,22 and scanning electron microscopy (SEM)19,23,24 for changes in morphology and roughness (Figure S1). The perovskite layers are composed of large crystals with dimensions around 300 nm. No major change in the

ACS Paragon Plus Environment

7

ACS Applied Energy Materials 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 20

perovskite morphology can be observed with respect to the unmodified SnO2 samples, with the exception of a small decrease in grain size for the samples modified with DEA-P-CNVPA. Additionally, to understand the effect of phosphonic acid on the perovskite structures, we performed X-ray diffraction measurements on full devices. As shown in Figure S2, the control device shows the same perovskite structure as the reference perovskite film (excess PbI2 was added for passivation), confirming that the cubic structure of the mixed ion perovskite is stable in the device environment. Additional peaks detected in the control device are from the Au. All the impurity peaks are labelled. Adding PA does not affect the perovskite structure, as device with pCN-PA and DEA-P-CNVPA show the same XRD patterns as the control device. Solar-cell results (Figure 3) indicate that all phosphonic acid surface modifications of dipping in a 0.05 mM PA solution in ethanol for 3 h resulted in decreased overall PCE relative to the control based on UV-ozone-treated SnO2. Samples dipped in pCN-BPA showed the PCE successively decreased with increasing the dipping time from 1 min to 5 min to 1 h (Figure S3). Further increase of the dipping time to 3 h led to the same PCE as a 1 h dip. The VOC values for the devices based on PA-modified SnO2 are all lower than that of the control (Figure 3a). According to the energy diagram in Figure 2c, the changes in ETL work-function should yield a high built-in potential for the BPA-modified SnO2, and therefore a high VOC. However, we found discrepancies when looking at the work-functions of different PA-modified samples, and comparing them to changes in VOC. For example, the pCN-BPA-modified SnO2 shows a work-function that is 0.20 eV larger than that of BPA-modified SnO2, and so built-in potential arguments would suggest a VOC that is 0.20 V lower than in the BPA case. Instead we saw a slight decrease in VOC of about 0.025 V (for the averages). The changes are also smaller than the effects on VOC obtained in similar studies of the modification of ZnO for organic solar cells.25 This indicates that for this system the work-function

ACS Paragon Plus Environment

8

Page 9 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Energy Materials

of the ETL surface, before deposition of the perovskite active layer, does not have a significant effect on the VOC obtained. Other work has shown similar effects on the HTL, where no VOC changes are seen for different work-functions or ionization energies.26 Differences in average JSC (Figure 3b) were also very slight and the major contributor to differences in PCE between devices based on the three different phosphonic acid-modified surfaces is from variations in the fill factor (FF, Figure 3c and d). Overall, BPA-modified devices exhibit the highest FF and overall PCE among the treated samples.

Figure 4. Understanding the effects of phosphonic acids on SnO2. a) Surface energies of SnO2 modified by three phosphonic acids. b) Schematic of the photoluminescence measurements. Samples were excited from the glass side under 405 nm excitation to ensure most charges are produced close to the SnO2/perovskite interface. c) Time-resolved photoluminescence at room temperature for all samples. To understand the discrepancy in the work-function changes without affecting VOC, we studied the effects of the different PAs on the surface energy (Figure 4a), which dictates the interaction at the perovskite/SnO2 interface. The contact angles of the modified substrates with water and CH2I2 were measured and the surface energy estimated following Fowkes (Table S2).27 The largest difference seen between the substrates was in the contact angle of water and the corresponding polar component of the surface energy. The SnO2 control sample exhibits a value of 5.4° ± 2.9°,

ACS Paragon Plus Environment

9

ACS Applied Energy Materials 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 20

corresponding to a rather polar surface, whereas the PA-modified substrates are less polar, showing values of 27.7° ± 5.3°, 37.8° ± 3.4°, and 47.9° ± 2.1°, for the BPA, pCN-BPA, and DEA-PCNVPA, respectively. We investigated the quenching of the photoluminescence (PL) signal of a perovskite film in contact with the SnO2 and modified substrates, by probing the response of excitation with a 405 nm laser from the substrate side (Figure 4b). As expected, the SnO2 substrates quench the PL signal considerably (Figure 4c). However, the signal quenching is significantly reduced upon modification of the SnO2 substrate by any of the PAs. The PA interlayer may reduce the oxideperovskite electronic coupling and, thus, the rate of interfacial electron transfer. This effect may then cause a decrease in FFs, as is the case for the devices with modified ETLs. The decrease in VOC with respect to bare-SnO2-based cells can also be explained by the reduced electronic coupling; the slower rate of injection into the ETL allows recombination to become more competitive.28

ACS Paragon Plus Environment

10

Page 11 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Energy Materials

Figure 5. Solar-cell performance in a nitrogen environment over time. Forward and reverse current-voltage curves of a) the control and b) the DEA-P-CNVPA modified devices before (blue traces) and after (red traces) resting in inert atmosphere for 19 h. The photocurrent was normalized to emphasize the difference between backward and forward scans. c) PCE for control and DEAP-CNVPA-modified device as a function of time resting in nitrogen atmosphere. Prior to device fabrication, the control SnO2 substrate was treated with UV-Ozone for 15 min, while the modified substrate was treated with 0.05 mM in ethanol for 3 h. Solid lines indicate PCE value obtained from the reverse scan, where the dashed lines indicate the forward scans. As noted previously, degradation of photovoltaic performance has been seen for PSCs with an ALD SnO2 ETL kept under inert atmosphere, presumably associated with oxygen loss. Given that phosphonic acids can help passivate oxide surfaces,14 we studied the stabilization of solar cell performance by phosphonic acid modification for ALD SnO2 ETL devices kept under inert atmosphere (Figure 5). The control sample showed a marked increase in hysteresis (Figure 5a) and its average PCE decreased to 65% of the original value after 19 h (Figure 5c and S4). By contrast, a DEA-P-CNVPA-modified sample maintained similar hysteresis and the PCE decreased to 86%

ACS Paragon Plus Environment

11

ACS Applied Energy Materials 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 20

of the original value after 19 h (Figure 5c and S4), ending at an average PCE higher than that of the control (8.6% vs. 7.5% PCE). It is possible that the unchanged hysteresis for the modified samples relates to work function stabilization by the PAs. Phosphonic acid modification of the work-function of the ETL does not have a significant effect on the open-circuit voltage of the solar cell, whereas it has a positive effect in the long-term performance. The work-function has an insignificant influence on the device performance when comparing samples treated with two different phosphonic acids. Unmodified SnO2 samples showed the highest surface energy and the best device performance, whereas all modified samples exhibited lower surface energy and performance. A phosphonic acid-modified sample maintained similar hysteresis and exhibited a lower decrease in PCE after 19 h than a control, which showed increased hysteresis. These results show the importance and promise of phosphonic acid surface modification of SnO2 ETLs for PSC stability under low-oxygen environments.

Materials Tetrakis(dimethylamino)tin(IV) (TDMASn, 99.99%-Sn, Strem Chemicals INC), oxygen gas (99.9995% pure, Carbagas), nitrogen (99.9999% pure, Carbagas), zinc powder, titanium isopropoxide (TTIP), lead (II) iodide (PbI2), lead(II) bromide (PbBr2), 4-tert-butylpyridine (tBP), bis(trifluoromethane)sulfonimide lithium salt (Li-TFSI, Sigma Aldrich), hydrochloric acid (HCl), ethanol (EtOH), acetone, isopropanol (IPA), chlorobenzene (CB), dimethyl sulfoxide (DMSO), dimethylformamide (DMF), 1.2-dichlorobenzene (DCB), were purchased from Sigma-Aldrich; 2,2´,7,7´-tetrakis(N,N´-di-p-methoxyphenylamine)-9,9´-spirobifluorene (Spiro-OMeTAD) was purchased from Merck; and fluorine-doped SnO2 (FTO) glass was purchased from Nippon Sheet Glass. Experimental details Preparation of the phosphonic acids

ACS Paragon Plus Environment

12

Page 13 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Energy Materials

Benzylphosphonic acid (97%) was purchased from Alfa Aesar. (1-Cyano-2-(4(diethylamino)phenyl)vinyl)phosphonic acid (DEA-P-CNVPA) was synthesized according to the literature.17 The preparation of 4-cyanobenzylphosphonic acid (pCN-BPA) is described below. 4-Cyanobenzylphosphonic acid (pCN-BPA): A degassed solution of 4(bromomethyl)benzonitrile (9.0 g, 45.9 mmol), triethyl phosphite (23.6 mL, 138 mmol) was refluxed overnight at 150 °C under nitrogen. The excess triethyl phosphite was removed by vacuum distillation to afford the product as yellow oil (9.43 g, 81%). The 1H NMR spectrum was consistent with the literature for diethyl 4-cyanophenylphosphonate.29 A solution of this material (5.0 g, 19.7 mmol) was made in dry dichloromethane (63.3 mL) and bromotrimethylsilane (7.82 mL, 59.2 mmol) was added via syringe. The mixture was capped with a greased stopper and allowed to stir at room temperature overnight. The volatiles were removed under reduced pressure to produce a light-yellow oil. The oil was dissolved in methanol and water. The vessel was sealed with a rubber septum and a needle inserted for the release of any pressure. The mixture was allowed to stir at room temperature overnight. The resulting white solid (3.39 g, 87%) was recrystallized from acetonitrile. The 1H NMR spectrum was consistent with the literature. Modification of SnO2 by phosphonic acids Glass preparation: F:SnO2 (FTO) substrates were chemically etched with zinc powder and 4 M HCl solution and then cleaned through immersing in piranha solution (H2SO4/H2O2 = 3:1) for 10 min. All substrates were further cleaned by UV-ozone for 15 min before deposition of SnO2. Preparation of SnO2 by ALD: SnO2 control planar devices were deposited through atomic layer deposition of tetrakis(dimethylamino)tin(IV) (heated at 55 °C) and ozone in a Savannah ALD 100 instrument (Cambridge Nanotech Inc.), at 118 °C. Oxygen gas was used for production of ozone (13% in O2) by a generator (AC-2025, IN USA Incorporated). The carrier gas was Nitrogen with a flow rate of 10 sccm. The growth rate (0.065 nm/cycle) was measured by ellipsometry. Phosphonic acid modification: SnO2 substrates were dipped in ethanol solutions of the 3 different phosphonic acids at a concentration of 0.05 mM. Several dipping times were investigated, including 1 min, 5 min, 1 h, and 3 h. Except where specified, measurements were done with a dipping time of 3 h for all phosphonic acids. Lead halide perovskite precursor solution The mixed perovskite precursor solutions are composed of a mixed-ion recipe reported earlier.18 The precursor solution contains FAI (1 M), PbI2 (1.1 M), MABr (0.2 M) and PbBr2 (0.22 M) in anhydrous DMF:DMSO 4:1 (v:v). A 1.5 M stock solution of CsI (abcr GmbH) in DMSO was added to above solution (MA/FA perovskite) in a 5:95 molar ratio. Solar cell Preparation The perovskite solution was spin coated in a two-step program (10 s at 1000 rpm and 20 s at 6000 rpm). During the second step, 200 μL of chlorobenzene was poured on the spinning substrate 20 s prior to the end of the program. The substrates were then annealed (at 100 °C unless stated otherwise) for 1 h in a nitrogen-filled glove box. After the perovskite annealing, the substrates were cooled down for few minutes and a spiro-OMeTAD solution (70 mM in chlorobenzene) was spin coated at 4000 rpm for 20 s. Spiro-OMeTAD was doped with Li-TFSI, FK209, and tBP with

ACS Paragon Plus Environment

13

ACS Applied Energy Materials 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 20

molar ratios of: 0.5, 0.03 and 3.3 respectively. Finally, 70-80 nm of a gold top electrode was thermally evaporated under high vacuum. Samples were placed in a dry box after solar cell deposition to activate the Li-TFSI doping. Solar cell characterization The solar cells were measured using a 450 W xenon light source (Oriel). A Schott K113 Tempax filter (Präzisions Glas & Optik GmbH) was used to reduce the spectral mismatches between AM 1.5G and the light source. A Si photodiode equipped with an IR-cutoff filter (KG3, Schott) was used as a reference. Current-voltage data were obtained by applying an external voltage bias and measuring the current response with a digital source meter (Keithley 2400). The voltage scan rate was set to 10 mV s-1. The photovoltaic data were collected without any device preconditioning, such as light soaking. In order to fix the active area of the devices and avoid artifacts produced by scattered light a black metal mask was used during the measurements. The IV data for champion devices were gathered with antireflective coating on a 0.16 cm2 mask aperture, whereas the data for the statistical analysis devices collected without antireflective coating. General Characterization Field emission scanning electron microscopy was performed using a Zeiss SEM (Supra55VP), with an InLens detector and an acceleration voltage of 0.5 kV. Atomic force microscopy was employed using a Cypher ES AFM (Asylum Research). Images were collected with an uncoated silicon tip (300 kHz, 26 N/m, OPUS) in tapping mode. All images are shown with line-wise flattening to remove tilting effects of the substrate plane. Time-resolved photoluminescence measurements were obtained by time-correlated single photon counting (TCSPC). The samples were excited through the glass side by a pulsed 405 nm wavelength laser (PicoQuant LDH-P-C405) at a repetition rate of 500 kHz. The laser power was adjusted by a neutral density filter wheel to obtain a