Modification of SnO2 Electron Transport Layer in Perovskite Solar Cells

Rapid development of the device performance of organic-inorganic lead halide perovskite solar cells (PSCs) are emerging as a promising photovoltaic technology. Current world-record efficiency of PSCs is based on tin oxide (SnO2) electron transport layers (ETLs), which are capable of being processed at low temperatures and possess high carrier mobilities with appropriate energy- band alignment and high optical transmittance. Modification of SnO2 has been intensely investigated by various approaches to tailor its conductivity, band alignment, defects, morphology, and interface properties. This review article organizes recent developments of modifying SnO2 ETLs to PSC advancement using surface and bulk modifications, while concentrating on photovoltaic (PV) device performance and long-term stability. Future outlooks for SnO2 ETLs in PSC research and obstacles remaining for commercialization are also discussed.


Introduction
Solar energy is one of the most plentiful energy resources accessible to humankind. Among the various PV technologies, currently monocrystalline-silicon-solar cells dominate the PV market due to its high PCE of 26.1% and high working stability, but suffers from cost-intensive production cost of the highly purified monocrystalline silicon. The low production cost and rapid increase in unit-cell efficiency of organic-inorganic PSCs, which is currently 25.7% [1], enables it to compete with silicon-solar cells. Organometallic-halide perovskites are based on the chemical formula of ABX 3 , where A is organic or metal cations, such as methylammonium (CH 3 NH 3 + (MA + )), formamidinium ((NH 2 ) 2 CH + (FA + )), Rb + , or Cs + , B is metal ions, such as Pb 2+ or Sn 2+ , and X is halogen ions, such as I − , Br − , or Cl − . Organometallic-halide perovskites possess features of an ideal absorber material, including high absorption coefficients (~10 −4 cm −1 ), long carrier-diffusion lengths (>1 µm), ambipolar-charge-transport capabilities, and low exciton-binding energy (20-50 meV), [2].
A typical n-i-p PSC device structure consists of glass/transparent conducting oxide (TCO)/n-type ETL/perovskite absorber/p-type hole transport layer (HTL)/metal contact. The bottom TCO is usually fluorine-doped tin oxide (FTO) or indium tin oxide (ITO), and the top metal contact is typically gold (Au), silver (Ag), or aluminum (Al). For a typical p-i-n PSC device, the structure consists of glass/TCO/HTL/perovskite/ETL/top metal contact.

Elemental Doping
Elemental doping of SnO 2 is a straightforward method to effectively alter the conductivity, defect states, and energy level. SnO 2 can be simply doped by various elements to tune its electrical and chemical properties. SnO 2 is an n-type material. The tetravalent-Sn sites can be replaced by cations with low valence states, such as gallium (Ga 3+ ), cobalt (Co 3+ ), zinc (Zn 2+ ), magnesium (Mg 2+ ), and lithium (Li + ) for p-type doping, or can be substituted by cations with high valence states, such as antimony (Sb 5+ ), molybdenum (Mo 5+ ), tantalum (Ta 5+ ), and niobium (Nb 5+ ) for n-type doping.
Li et al. reported a significant improvement in conductivity without declining the transmittance by doping SnO 2 with Ta [22]. Ta-doped SnO 2 improved the PCE from 19.5% to 20.9% by improved fill factor (FF) and short-circuit current density (J SC ), as shown in the illuminated current-density (J-V) curves in Figure 1a. This is due to effective acceleration of electron collection and transfer, and reduction in recombination at the ETL/absorber interface, and shown in the steady state photoluminescence (PL) and time-resolved PL (TRPL) results in Figure 1b Figure 1d. Improved electron-extraction ability is suggested, based on the steady-state PL and TRPL results in Figure 1e,f. Encapsulated devices retained over 85% of its initial efficiency after stored in air at room temperature and 40-50% relative humidity for 300 h.
M. Park et al. reported a solution process to effectively dope SnO 2 with Li at a low processing temperature [24]. Li-doping enhanced the conductivity of SnO 2 and produced a reduction of the conduction-band energy, as shown in Figure 1g, facilitating the transfer of electrons and reduced the charge recombination. This resulted in improved V OC , FF, and J SC with a PCE increase of 15.3% to 18.2%, as shown in Figure 1h,i.
Niobium doping of SnO 2 was reported by Ren et al. using a solution-processable low-temperature method [25]. The improvement in PV performance originates from the increased conductivity and improved surface morphology, which lead to enhanced electron extraction and inhibited charge recombination. Unencapsulated devices maintained 90% of its initial PCE after 288 h stored in air at room temperature.
Yttrium (Y 3+ ) doping of SnO 2 reported by Yang et al. promotes more homogeneous distribution and well-aligned SnO 2 nanosheet arrays, which leads to improved electron transfer from the absorber to the ETL [26]. Enlarged bandgaps from the Y-doping and a higher conduction-band energy allows improved energy band alignment and reduced the charge recombination at the ETL/absorber interface. This improved the PCE of the PSC from 13.4% to 17.3% by increasing V OC , FF, and J SC .
Bai et al. reported that Sb:SnO 2 nanocrystals was used to replace the undoped SnO 2 ETL [27]. This shifted the Fermi-energy level upward, which improved the energy-band alignment and reduced charge recombination. Electron = recombination lifetime was longer, and V OC and FF increased with less photocurrent hysteresis. PCE values increased from 15.7% to 17.2%. Unsealed devices retained over 95% of its initial efficiency after 504 h stored in a desiccator at room temperature.
Gallium-doped SnO 2 reported by Roose et al. observed decreased trap-state density in the ETL, leading to a reduced recombination rate [28]. V OC and FF increased from 1.00 to 1.07 V and 57.0% to 70.0%, respectively, resulting in a PCE enhancement of 12.5% to 17.0%. Unencapsulated devices maintained about 70% of its initial efficiency after 1000 h under continuous 1 SUN illumination in nitrogen (N 2 ).
Xiong et al. reported Mg-doped SnO 2 as the ETL in PSCs [29]. An optimum-Mg content resulted in uniform, smooth, and dense films with reduced free-electron density, which suppressed the charge recombination, and increased electron mobility, which enabled fast extraction of electrons, contributing to improved J SC . PCE improved from 8.2% to 15.2% by doping the ETL with Mg. Unencapsulated devices maintained over 90% of its initial PCE after 720 h of storage in air with <20% relative humidity. Li et al. reported a significant improvement in conductivity without declining the transmittance by doping SnO2 with Ta [22]. Ta-doped SnO2 improved the PCE from 19.5% to 20.9% by improved fill factor (FF) and short-circuit current density (JSC), as shown in the illuminated current-density (J-V) curves in Figure 1a. This is due to effective acceleration of electron collection and transfer, and reduction in recombination at the ETL/absorber interface, and shown in the steady state photoluminescence (PL) and time-resolved PL (TRPL) results in Figures 1b,c.  Aluminum doped SnO 2 reported by Chen et al. resulted in increased J SC and FF by using a low-temperature solution-processable method [30]. Doping SnO 2 with Al enhanced the charge transport and electron extraction based on TRPL tests. PSC devices exhibited improved PCE of 9.0% to 12.1%.

Metal Oxide
Binary layers of SnO 2 with other metal oxides is another approach to modify the ETL in PSCs. This approach leads to tailoring the surface morphology and tunes the energy band alignment. Wang et al. investigated inserting indium oxide (In 2 O 3 ) between ITO and SnO 2 , which resulted in reduced trap densities in the perovskite with improved charge transfer and band alignment [31], as shown in Figure 2a. This resulted in higher V OC and FF values with a PCE enhancement of 21.4% to 23.1%, as shown in Figure 2b. Unsealed devices retained 98% of its initial efficiency after being stored in N 2 for 1920 h, as shown in Figure 2c, resulting in 91% of its initial PCE after 180 h of continuous 1 SUN illumination and about 80% of its initial PCE after 120 h of exposure to 75% relative humidity.
Europium (Eu) doped tungsten-oxide (WO x ) nanorods were inserted between the perovskite and SnO 2 layer by Chen et al. [32], which contributed to improved crystallinity of the perovskite film and enhanced conductivity and carrier mobility of both the SnO 2 ETL and spiro-OMeTAD HTL. Due to the energy level of the Eu:WO x nanorods, as shown in Figure 2d, the electron and hole extraction were remarkably boosted at the ETL/absorber and absorber/HTL interface, improving the PV performance, as shown in Figure 2e. Unencapsulated devices retained over 90% of its initial PCE after 500 h of continuous 1 SUN illumination at 16-25 • C and 20-30% relative humidity, and over 90% of its initial PCE after 2000 h of exposure to ambient air, as shown in Figure 2f.
Song et al. investigated inserting an anodized-TiO 2 between the FTO and SnO 2 ETL, which lead to a defect-free physical contact and improved electron extraction [33], as illustrated in Figure 2g. Such bi-layered ETLs resulted in a large change in free energy and moderate electron mobility, as illustrated in Figure 2h. This enhanced the V OC from 1.14 to 1.20 V, which led to an enhanced PCE of 19.0% to 21.1%, as shown in Figure 2i. Similarly, Lee et al. investigated combining a compact TiO 2 between the TCO and SnO 2 ETL, which improved charge collection due to better hole-blocking ability of the TiO 2 underlayer [34]. This resulted in increasing the PCE from 16.4% to 19.8%.
Unencapsulated devices maintained over 95% of its initial PCE after 1200 h of storage in air with 20% relative humidity.
Dagar et al. studied inserting a thin magnesium-oxide (MgO) overlayer on top of SnO 2 , which led to more uniform films and reduced interfacial-carrier recombination [35]. This resulted in better stability and enhanced device performance from a PCE of 15.2% to 19.0%. Unsealed devices maintained 67% of its initial PCE after 2568 h in 30% relative humidity.
A bilayer of lead oxide (PbO) doped-SnO 2 with undoped SnO 2 was investigated by Bi et al. [36], which improved the shunt resistance and enhanced the FF from 72.9% to 75.5% and PCE from 17.0% to 18.8%. Devices maintained over 90% of its initial efficiency after 1080 h at room temperature with 15% relative humidity.
Ultrathin MgO was also inserted between FTO and SnO 2 ETL in PSCs by Ma et al., which resulted in enhanced electron transporting and hole-blocking properties [37]. Due to MgO passivation, less FTO-surface defects were observed along with a smoother surface and suppressed carrier recombination. The PCE increased from 16.4% to 18.2%.
Hou et al. studied chemical-bath deposition of a SnO 2 /TiO 2 bilayer [38], which facilitates charge separation achieving effective extraction and transport of electrons. Higher electron mobility and suppressed recombination was observed due to the reduced energy barriers and gradual-energy levels, which lead to a PCE enhancement of 12.0% to 18.1%.
Yan et al. reported a bi-layered ETL of SnO 2 and zinc oxide (ZnO), which showed superior electron extraction and a lower charge recombination rate [39]. This resulted in a V OC enhancement of 1.06 to 1.23 V, and a PCE enhancement of 11.9% to 14.6%. Unencapsulated devices retained 80% of its initial efficiency after 300 h in N 2 at 85 • C. devices retained 98% of its initial efficiency after being stored in N2 for 1920 h, as shown in Figure 2c, resulting in 91% of its initial PCE after 180 h of continuous 1 SUN illumination and about 80% of its initial PCE after 120 h of exposure to 75% relative humidity.

Ionic Compounds 2.3.1. Surface Modification by Ionic Compounds
Adding ionic compounds into the SnO 2 solution or applying it to the surface the SnO 2 are some other cost-effective approaches to modifying the SnO 2 ETL. Compared to organic molecules and carbon materials, ionic compounds are usually more stable and lower in cost.
Zhuang et al. looked into the modification of SnO 2 by using rubidium fluoride (Rb) by using two different methods: (i) adding RbF into the SnO 2 solution and (ii) inserting RbF at the SnO 2 /perovskite interface [40]. Adding RbF to the SnO 2 bulk resulted in improved electron mobility, while adding RbF to the surface of SnO 2 resulted in inhibited ion migration and reduced carrier recombination due to the Rb + cations escaping into the bulk perovskite. This led to increased V OC with PCE of over 23%, as shown in Figure 3a with negligible hysteresis. A stronger steady-state PL intensity of the RbF-treated SnO 2 corresponds to improved perovskite-film quality, as shown in Figure 3b. Based on the TRPL results in Figure 3c, a shorter fast decay (τ 1 ) indicates enhanced electron extraction, while the longer slow decay (τ 2 ) corresponds to decreased defects/traps and improved perovskite-film quality. Unencapsulated PSCs maintained about 75% of its initial PCE after 200 h of exposure to white LED light illumination.
Chen et al. studied the combined effect of doping planar-SnO 2 (p-SnO 2 ) with different concentrations of RbF and depositing RbF onto the mesoporous SnO 2 (m-SnO 2 ) layer [41], resulting in improved PV performance, as shown in Figure 3d. RbF modification increased the conductivity of SnO 2 , as shown in Figure 3e, and passivated interfacial traps through F-Sn bonds. Rb + diffused into the perovskite which passivated the perovskite and suppressed ion migration. Unencapsulated devices maintained 90% of its initial PCE after 300 h at the maximum power point (MPP) under 1 SUN illumination, as shown in Figure 3f.
Chen et al. reported surface modification of the SnO 2 by using 4-imidazoleacetic acid hydrochloride (ImAcHCl) [42], as illustrated in Figure 3g. The chloride anion in ImAcHCl improves the crystallinity of the perovskite layer. Modifying SnO 2 with ImAcHCl shifts the conduction and valence bands up, as shown in Figure 3h, suppresses carrier recombination, and enhances carrier lifetime. This results in the enhancement of PCE from 19.5% to 21.0%, as shown in Figure 3i. Unencapsulated-PSC devices retained 95% of its initial PCE after 840 h at room temperature with 46-60% relative humidity. Unencapsulated PSCs based on PTAA maintained 90% of its initial PCE after 40 h of exposure to 85 • C in N 2 .
Potassium-hydroxide (KOH) modification of SnO 2 surfaces was investigated by Bu et al., which resulted in suppressed hysteresis and enhanced PV performance with a PCE increase from 19.3% to 20.5% [43]. Potassium cations were shown to passivate the ETL/perovskite interface, facilitate grain growth of the perovskite, and improve stability.
Cesium-carbonate (Cs 2 CO 3 ) post-treatment, studied by Li et al., improves the electrical properties of SnO 2 and passivates the ETL/perovskite interface [44]. Such Cs 2 CO 3 modification improves the surface wettability of the ETL and reduces the roughness, resulting in a perovskite film with larger grains. The Cs 2 CO 3 post-treatment lowers the work function, reducing electron-hole recombination, and enhancing the electron transfer. Devices retained 91% of its initial efficiency after stored in 35-45% relative humidity for 340 h.
Alkali-metal cations, such as lithium chloride (LiCl), sodium chloride (NaCl), potassium chloride (KCl), rubidium chloride (RbCl), and cesium chloride (CsCl), were used to modify the SnO 2 surface [45]. Such modification increases the mobility and reduces the trap density of states of SnO 2 . Efficient defect passivation suppresses the recombination at the ETL/absorber interface. Devices based on NaCl-treated SnO 2 maintained over 90% of its initial PCE after 960 h. with negligible hysteresis. A stronger steady-state PL intensity of the RbF-treated SnO2 corresponds to improved perovskite-film quality, as shown in Figure 3b. Based on the TRPL results in Figure 3c, a shorter fast decay (τ1) indicates enhanced electron extraction, while the longer slow decay (τ2) corresponds to decreased defects/traps and improved perovskite-film quality. Unencapsulated PSCs maintained about 75% of its initial PCE after 200 h of exposure to white LED light illumination.    Figure 4a; it enhanced charge extraction and suppressed interfacial recombination [46]. This enhanced the V OC up to 1.20 V for a perovskite layer with a bandgap of 1.54 eV. PSC devices with a PCE of 23.8% was achieved, as shown in Figure 4b, with enhanced stability maintaining 84% of initial efficiencies after 200 h of continuous light exposure, as shown in Figure 4c.    The contact between the SnO2 and perovskite was improved by introducing a biological polymer, heparin potassium, to the SnO2 [47]. Such bulk incorporation regulated the arrangement of SnO2 nanocrystals and induced vertically aligned crystal growth of the perovskite, as illustrated in Figure 4d. This improved the PV performance, as shown in Figure 4e. Due to the strengthened interface binding, device-operational stability was enhanced resulting in maintaining 97% of the initial efficiency after 1000 h under 1 SUN illumination at the MPP, as shown in Figure 4f.
Introducing KCl to the SnO2 ETL passivated both the grain boundaries of the perovskite and the defects at the ETL/absorber interface [48]. The Cl − and K + ions passivate the ETL/absorber contact, while the K + ions in the ETL diffuse into the perovskite layer and passivate the grain boundaries, resulting in enhanced VOC from 1.08 to 1.14 V and increased PCE from 20.2% to 22.2%, as shown in Figure 4g. The stronger steady-state PL intensity with the presence of KCl demonstrates suppressed recombination of the perovskite, as shown in Figure 4h. Based on TRPL results in Figure 4i, the decrease in τ1 suggests faster electron transfer with the incorporation of KCl, and the increase in τ2 suggests slower recombination in the perovskite film grown on KCl-incorporated SnO2. Unsealed devices maintained 88% of its initial PCE after 120 h of continuous 1 SUN illumination.
Girard's Reagent T (GRT) was introduced to the SnO2 nanoparticles by Bi et al. [49]. The carbonyl group in GRT is anticipated to prevent agglomeration of the SnO2 nanoparticles. The quaternary-ammonium-chloride salt in GRT is expected to facilitate the crystal growth of the perovskite, and the quaternary-ammonium cation and chloride anion can passivate the defects at the ETL/absorber interface. Such GRT incorporation into SnO2 resulted in better electrical properties of SnO2, promoted vertical growth of the perovskite, and reduced interfacial defects at the ETL/perovskite interface, resulting in a PCE enhancement of 19.8% to 21.6%. Unencapsulated devices retained over 99% of its initial efficiency after 720 h at 60 °C, and 59% after 672 h under 1 SUN illumination.
Liu et al. introduced ammonium chloride (NH4Cl) into SnO2, which resulted in PSC devices with negligible hysteresis and improvement in PCE from 18.7% to 21.4% [50]. The contact between the SnO 2 and perovskite was improved by introducing a biological polymer, heparin potassium, to the SnO 2 [47]. Such bulk incorporation regulated the arrangement of SnO 2 nanocrystals and induced vertically aligned crystal growth of the perovskite, as illustrated in Figure 4d. This improved the PV performance, as shown in Figure 4e. Due to the strengthened interface binding, device-operational stability was enhanced resulting in maintaining 97% of the initial efficiency after 1000 h under 1 SUN illumination at the MPP, as shown in Figure 4f.
Introducing KCl to the SnO 2 ETL passivated both the grain boundaries of the perovskite and the defects at the ETL/absorber interface [48]. The Cl − and K + ions passivate the ETL/absorber contact, while the K + ions in the ETL diffuse into the perovskite layer and passivate the grain boundaries, resulting in enhanced V OC from 1.08 to 1.14 V and increased PCE from 20.2% to 22.2%, as shown in Figure 4g. The stronger steady-state PL intensity with the presence of KCl demonstrates suppressed recombination of the perovskite, as shown in Figure 4h. Based on TRPL results in Figure 4i, the decrease in τ 1 suggests faster electron transfer with the incorporation of KCl, and the increase in τ 2 suggests slower recombination in the perovskite film grown on KCl-incorporated SnO 2 . Unsealed devices maintained 88% of its initial PCE after 120 h of continuous 1 SUN illumination.
Girard's Reagent T (GRT) was introduced to the SnO 2 nanoparticles by Bi et al. [49]. The carbonyl group in GRT is anticipated to prevent agglomeration of the SnO 2 nanoparticles. The quaternary-ammonium-chloride salt in GRT is expected to facilitate the crystal growth of the perovskite, and the quaternary-ammonium cation and chloride anion can passivate the defects at the ETL/absorber interface. Such GRT incorporation into SnO 2 resulted in better electrical properties of SnO 2 , promoted vertical growth of the perovskite, and reduced interfacial defects at the ETL/perovskite interface, resulting in a PCE enhancement of 19.8% to 21.6%. Unencapsulated devices retained over 99% of its initial efficiency after 720 h at 60 • C, and 59% after 672 h under 1 SUN illumination.
Liu et al. introduced ammonium chloride (NH 4 Cl) into SnO 2 , which resulted in PSC devices with negligible hysteresis and improvement in PCE from 18.7% to 21.4% [50]. Such improvement is due to the increased electron mobility, and improved band alignment and passivation of the ETL/perovskite interface, which also improved the device's stability. Unencapsulated devices stored in N 2 retained over 95% after 1000 h.
Sun et al. incorporated potassium sodium tartrate (PSTA) into the SnO 2 colloidal dispersion [51]. PTSA contains mobile-alkali-metal cations leading to improved uniformity and conductivity, and less defects in SnO 2 , which improves the crystallinity of the perovskite film. Sodium and potassium cations can diffuse into the perovskite and passivate the defects at the grain boundaries and surface. This results in PCE enhancement of 18.3% to 21.1% with reduced hysteresis. Device stability improves with unencapsulated devices retaining over 95% of its initial efficiency after 1440 h of exposure to ambient air with 45% relative humidity at 25 • C.
Phosphoric acid was introduced into SnO 2 by Jiang et al. to eliminate dangling bonds on the SnO 2 surface and improve the carrier-collection efficiency [52]. Electron mobility increased by 3 times and surface-trap states reduced, decreasing the electron-transport barriers of SnO 2 . Attributed to the enhanced electron-collection efficacy, the PCE increased from 19.7% to 21.0%.
Tetramethylammonium hydroxide (TMAH) was incorporated into SnO 2 by Huang et al. at low temperatures of 100-150 • C [53]. Such modification of SnO 2 attributed to higher conductivity of SnO 2 and also effectively passivated the grain boundaries of the perovskite film. Improved charge transport between the perovskite and ETL resulted in enhanced efficiencies from 18.1% to 20.5%. Encapsulated devices in 15% relative humidity maintained 97% of its initial PCE after 360 h.
Modification of SnO 2 by introducing cesium fluoride (CsF) into the ETL was investigated by Akin et al., which led to improved optoelectronic properties and rapid extraction of photogenerated electrons [54]. By combining the modification of SnO 2 and inserting zwitterion molecules at the perovskite/HTL interface, a high V OC value of 1.23 V and a PCE value of 20.5% were achieved. Device-operational stability was also improved retaining over 90% of its initial PCE after 800 h under continuous 1 SUN illumination at the MPP.
Ammonium sulfide [(NH 4 ) 2 S] is incorporated in SnO 2 reported by Ai et al., which passivated the surface defects by terminating the Sn-S dangling bonds [55]. The conductivity and electron mobility of SnO 2 are increased, enhancing the electron collection and lowering electron-hole recombination rate. The Sn-S-Pb anchors the perovskite crystals at the ETL/absorber interface, which enhances the stability and electron extraction of the PSC. PCE values increase from 18.7% to 20.0% by using this method.

Carbon Materials 2.4.1. Surface Modification by Carbon Materials
Insertion of highly conductive carbon material at the ETL/absorber interface can facilitate electron transfer. A smooth ETL surface can also be enabled to impact the growth and nucleation of the perovskite layer on top. In addition, defects at the ETL/absorber interface can be passivated by the carbon material. For example, Wang et al. investigated the application of novel fulleropyrrolidine (NMBF-X, X = H or Cl) monomers and dimers, as shown in Figure 5a, in between the perovskite and ETL [56]. The chlorinated-fullerene dimers resulted in the most efficient PCE of 22.3% with minimal hysteresis, as shown in Figure 5b,c, which stems from the passivation of the NMBF-Cl dimer with the SnO 2 and perovskite simultaneously. After 1000 h, unencapsulated PSC devices exposed to air at 25-35 • C with 45-60% relative humidity retained over 95% of its initial PCE. Nanomaterials 2022, 12, x FOR PEER REVIEW 23 of 35 Polystyrene (PS) was inserted between the ETL and absorber layer to release residual stress in the absorber during annealing, which attributes to reduced interface defects, less Polystyrene (PS) was inserted between the ETL and absorber layer to release residual stress in the absorber during annealing, which attributes to reduced interface defects, less recombination, and lower ion-migration tendencies [57]. PS was also applied on top of the perovskite film for inner-encapsulation, as shown in Figure 5d, which improves the long-term device stability maintaining over 90% or its initial efficiency after 72 h at the MPP under continuous 1 SUN illumination, as shown in Figure 5e. Devices with the PS-inner encapsulation maintained over 90% of its initial efficiency after 2800 h in air, as shown in Figure 5f.
Adding graphene quantum dots (GQDs) to the SnO 2 surface, as shown in Figure 5g, will improve the conductivity and fill the electron traps in SnO 2 , which improves the electron extraction and reduces the recombination at the ETL/absorber interface [58]. PSC devices with GQD-treated SnO 2 results in a V OC enhancement of 1.10 to 1.13 V and a PCE enhancement of 17.9% to 20.3% with little hysteresis, as shown in Figure 5h,i. Unencapsulated devices maintain over 95% of its initial efficiency in N 2 after 720 h and then an additional 1440 h in air (20-30% relative humidity).
Zheng et al. reported modifying the SnO 2 ETL surface with [6,6]-phenyl-C 61 -butyric acid methyl ester (PCBM) [59]. By simultaneously inserting PbS above the HTL and spiro-OMeTAD, a PCE of 19.6% was demonstrated with long-term storage stabilities of about 100% retained after exposure to air for 1000 h.
Ke et al. also reported fullerene, PCBM, and modification of the SnO 2 ETL surface, which resulted in an increase in PCE from 16.5% to 19.1% [60]. PCBM promotes enhanced electron transfer, suppressed carrier recombination at the ETL/absorber interface, and passivates both the ETL/absorber interface and the perovskite-grain boundaries.
Plasma-enhanced atomic layer deposition (PEALD) was reported to use a low deposition temperature of 100 • C and still attained high efficiencies by surface modification with passivation by C 60 -self-assembled monolayer (C 60 -SAM) [61]. The V OC improved from 1.07 to 1.13 V and FF increased from 75.5% to 79.1%, resulting in an efficiency enhancement from 17.2% to 19.0%. Unencapsulated devices maintained over 98% of its initial PCE after 480 h in N 2 with room-light exposure and a relative humidity below 10%.
A comparative study between passivation of SnO 2 by [6,6]-phenyl-C 61 -butyric acid (PCBA) and PCBM was performed by Wang et al. [62]. Passivation of fullerene-molecules at the SnO 2 surface efficiently decreases the defects and improves the conductivity through enhanced electron mobility. The PCBA modification showed higher passivation efficiency, resulting in a PCE increase from 15.4% to 18.6%.

Bulk Incorporation of Carbon Materials
Highly conductive carbon-based materials incorporated into the SnO 2 can enhance the conductivity of the ETL. For example, polymeric carbon nitrides (cPCNs) were introduced into the SnO 2 nanocrystals by Li et al., which led to electron mobilities three times higher than that of the undoped SnO 2 [63]. Such modification of SnO 2 led to less wettability, reduced grain boundaries from the suppressed heterogeneous nucleation of the perovskite, and improved the band alignment, as shown in Figure 6a. This resulted in negligible hysteresis with a PCE enhancement of 21.3% to 23.2%, as shown in Figure 6b. Unencapsulated devices retained 95% of its initial PCE after 2880 h in N 2 , as shown in Figure 6c.
Niu et al. studied the effect of incorporating Nb 2 C MXenes into SnO 2 , which led to increased grain growth and increased lattice-spacing facets of SnO 2 [64]. A decrease in the steady-state PL, shown in Figure 6d, demonstrates enhanced photogenerated carrier generation and reduced recombination. Superior crystallinity and effective carrier transport resulted in improved PSC PCE from 19.0% to 22.9%, as shown in Figure 6e. Unencapsulated devices retained 98% of its initial PCE after 960 h in 40-60% relative humidity at 25 • C, as shown in Figure 6f. at the SnO2 surface efficiently decreases the defects and improves the conductivity through enhanced electron mobility. The PCBA modification showed higher passivation efficiency, resulting in a PCE increase from 15.4% to 18.6%. Niu et al. studied the effect of incorporating Nb2C MXenes into SnO2, which led to increased grain growth and increased lattice-spacing facets of SnO2 [64]. A decrease in the steady-state PL, shown in Figure 6d, demonstrates enhanced photogenerated carrier generation and reduced recombination. Superior crystallinity and effective carrier transport resulted in improved PSC PCE from 19.0% to 22.9%, as shown in Figure 6e. Unencapsulated devices retained 98% of its initial PCE after 960 h in 40-60% relative humidity at 25 °C, as shown in Figure 6f.

Bulk Incorporation of Carbon Materials
Red-carbon quantum dots were used to dope solution-processed SnO2, which resulted in increasing the electron mobility by 20 times [65]. The enhanced electron mobility also showed to help passivate the traps and defects at the ETL/perovskite interface and promote growth of highly crystalline perovskite. Efficiencies improved from 19.2% to 22.8% through improvement of JSC, VOC, and FF, as shown in Figure 6g. Unencapsulated Red-carbon quantum dots were used to dope solution-processed SnO 2 , which resulted in increasing the electron mobility by 20 times [65]. The enhanced electron mobility also showed to help passivate the traps and defects at the ETL/perovskite interface and promote growth of highly crystalline perovskite. Efficiencies improved from 19.2% to 22.8% through improvement of J SC , V OC , and FF, as shown in Figure 6g. Unencapsulated devices maintained 96% of its initial PCE after 1000 h in air with 40-60% relative humidity at 25 • C, as shown in Figure 6h. Reduced steady-state PL intensity of the perovskites grown on the SnO 2 doped with red-carbon quantum dots suggest improved photogenerated-carrier generation and suppressed recombination, as shown in Figure 6i.
A water-soluble nonionic polymer, polyacrylamide (PAM), is introduced into SnO 2 be Dong et al., which improves the electron mobility, wettability, and uniformity of SnO 2 [66]. Defects in the perovskite is also reduced and grain size is increased from the PAM addition into SnO 2 . Band alignment at the ETL/perovskite interface is also improved, resulting in a PCE enhancement from 20.2% to 22.6%. Unencapsulated devices maintained 90% of its initial efficiency after 1080 h of exposure to 45%-55% humidity.
Graphitic carbon nitride (g-C 3 N 4 ) quantum dots were added to SnO 2 and applied to PSC devices by Chen et al. [67]. The oxygen-vacancy-reduced trap centers were effectively eliminated and bulk and interface-electron transport were promoted. The high conductivity and suitable energy band alignment of g-C 3 N 4 -treated SnO 2 led to a PCE improvement of 20.2% to 22.1%. Unencapsulated-PSC devices retained 90% of its initial efficiency after 1500 h in air with 60% relative humidity at 25 • C, and 80% of its initial efficiency after 75 h in air with 60% relative humidity at 85 • C.
Zhang et al. incorporated graphdiyne into SnO 2 , which led to improved electron mobility [68]. The enhanced hydrophobicity inhibits heterogeneous-perovskite nucleation, attributing to the reduced grain boundaries and less defect density. The ETL/absorber interface is also improved from passivation of the Pb-I anti-site defects. This results in a PCE increase of 19.2% to 21.1%.
Zhao et al. introduced naphthalene-diimide graphene into nanocrystal SnO 2 [69]. Such modification increases the surface hydrophobicity of SnO 2 and forms a van-der-Waals interaction at the ETL/perovskite interface. Enhanced FF values from 74.6% to 82.1% are attributed to the enhanced electron mobility and electron-extraction efficiency, and reduced carrier recombination. PCE improves from 19.0% to 20.2%.
Carbon nanodots by a hydrothermal process were incorporated into SnO 2 by Wang et al. using a simple solution process [70]. Incorporation of carbon nanodots into SnO 2 reduced the density of trap-states and increased the electron mobility of SnO 2 , resulting in a PCE of 20.0% with negligible hysteresis. As carbon materials exhibit high conductivity, incorporation of a carbon material into the ETL can enhance its charge transport and conductivity. Carbon materials can also passivate defects, improving the stability and efficiency of PSCs. Unencapsulated PSCs maintained 90% of its initial efficiency after 200 h of UV exposure in air with 20-30% relative humidity at 20 • C.

Surface Modification by Organic Molecules
Modification of SnO 2 by organic molecules can passivate defects at the ETL surface and improve the electrical properties and transport at the interface. The ETL and absorber layer can be chemically bridged by the organic molecule enhancing the interfacial-electron transfer. For example, Lou et al. investigated the introduction of π-conjugated n-type small organic molecules (BTAC4 and Y6) onto the surface of KCl-doped SnO 2 ETL [71], as illustrated in Figure 7a, which yields less trap states, suppressed carrier recombination, and improved electron transport and extraction. Band alignment improves with the modification, as shown in Figure 7b. Applying BTAC4 and Y6 to the surface of SnO 2 results in an enhanced PCE from 21.2% to 23.1% and 22.1%, respectively, as shown in Figure 7c. Unencapsulated devices maintained about 90% of its initial efficiency after 768 h in air with 35% relative humidity.   An iodine-terminated SAM, 3-iodopropyl trimethoxysilane [Si(OCH 3 ) 3 (CH 2 ) 3 I, I-SAM] was applied to the SnO 2 ETL surface in PSCs, which increased the adhesion toughness at the ETL/absorber interface-enhancing mechanical reliability [72], as shown in Figure 7d. This is attributed to the higher toughness and decreased hydroxyl groups at the interface. Without the SAMs treatment, irreversible morphological degradation, such as voids and delamination, was observed at the ETL/perovskite interface for operational stability tested-devices. Treatment with I-SAM on the SnO 2 enhanced the PCE from 20.2% to 21.4% with diminished hysteresis, as shown in Figure 7e. Long-term working stability was also improved, retaining over 90% of its initial efficiency in N 2 with continuous 1 SUN illumination at the MPP for unencapsulated devices for 1200 h, as shown in Figure 7f.
Triphenylphosphine oxide (TPPO) is an air-robust and cost-effective molecule for n-type doping of SnO 2 [73], as shown in Figure 7g. Surface modification by TPPO enhanced the conductivity and lowered the work function of SnO 2 , as shown in Figure 7h,i. The V OC improved from 1.08 to 1.11 V and PCE improved from 19.0% to 20.7% attributed to the lower recombination rate and faster electron extraction.
Thiophene-based interlayers were adopted to the SnO 2 surface to reduce the energy loss by optimizing the surface-electronic states of SnO 2 and improving the perovskite-film quality [74]. Surface modification of SnO 2 by thiophene-3-acetic acid improved the conductivity and lowered the work function of SnO 2 . Ion-defect states at the ETL/perovskite interfaces were passivated by bonding of the under-coordinated Pb 2+ of MAPbI 3 with the sulfur atoms of the thiophene rings with a lone pair of electrons. V OC improved from 1.07 to 1.12 V, and FF improved from 73.5% to 80.1%, resulting in a PCE improvement of 17.5% to 20.6%. This resulted in improved device stability retaining over 90% or its initial efficiency after 1440 h in N 2 and over 80% of its initial efficiency after 130 h in air with 70% relative humidity at 85 • C.
Aminosulfonic acid ( + H 3 N-SO 3 − , SA) is introduced to the surface of SnO 2 [75], and a chemical bridge is formed between the ETL and perovskite through the coordination bond to SnO 2 via -SO 3 − anions and electrostatic interactions with the perovskite via -NH 3 + cations. Better surface wettability of the SA-treated SnO 2 led to larger grain size of perovskite films. Attributed to the passivated-contact defects, V OC improved from 1.11 to 1.15 V, while barrier-free charge transferred led to improved FF and J SC with reduced hysteresis, resulting in a PCE improvement of 18.2% to 20.4%. Unencapsulated devices maintained over 80% of its initial PCE after 1000 h in air of 25-35% relative humidity and over 75% of its initial PCE after 500 h in N 2 at 60 • C.
The organic molecule p-amino benzenesulfonic acid (ABSA) was introduced to the surface of SnO 2 by inactivating the under-coordinated Sn ions [76]. This decreased the energy-band barrier on the surface of SnO 2 , and increased the conductivity and lessenscarrier recombination. This results in a PCE enhancement of 18.0% to 20.3%. Unsealed PSC devices retain 57% of its initial efficiency after 720 h in N 2 .
A plant-photosynthesis promoter, choline chloride, was introduced to the surface of SnO 2 by a simple molecular self-assembly method [77]. Such modification reduces the oxygen vacancies of SnO 2 , while the Cl ions form strong Pb-Cl bonds with the uncoordinated Pb ions in the MAPbI 3 . This passivates the defects at the ETL/absorber interface and reduces carrier recombination, improving the V OC from 1.07 to 1.15 V. PCE is improved from 16.8% to 18.9%.
Zuo et al. investigated various SAMs, such as 4-pyridinescarboxylic acid (PA-SAM), 4-cyanobenzoic acid (CBA-SAM), and benzoic acid (BA-SAM), and applied them to the surface of SnO 2 [78]. Proper interfacial interactions were shown to decrease trap-state density and enhance interfacial-charge transfer. Among the various SAMs, application of PA-SAM resulted in the highest efficiency enhancement of 17.2% to 18.8%, with a V OC enhancement from 1.06 to 1.10 V. This is due to improved electronic coupling and suppressed interfacial traps. Charge transfer at the ETL/absorber interface improved from the lowered work function.
Wang et al. reported interfacial-sulfur functionalization anchoring of SnO 2 by using potassium hexylxanthate to modify the surface of SnO 2 [79]. This approach effectively passivated the charge traps and suppressed carrier recombination at the interface by sulfur functionalization of the SnO 2 surface. Functionalized-sulfur atoms can also coordinate with the under-coordinated Pb 2+ ions at the interface. Such strategy resulted in improved device efficiencies of 16.6% to 18.4%. Unsealed devices maintained about 90% of its initial PCE after 1680 h at room temperature.

Bulk Incorporation of Organic Molecules
Incorporation of organic molecules possessing versatile functional groups into SnO 2 improves the dispersion of colloids, enhances the electrical properties, and passivates the defects in SnO 2 . For example, Xiong et al. introduced poly(ethylene glycol) diacrylate (PEGDA), as shown in Figure 8a, into the SnO 2 dispersion to prevent aggregations [80], which resulted in more uniform film and well-matched band-energy alignment with the perovskite. PEGDA-modified SnO 2 also attributed to passivating the defects at the ETL/absorber interface, as shown in Figure 8b,c. This showed a PCE improvement from 21.8% to 23.3%, with a V OC improvement of 1.09 to 1.14 V. Unencapsulated-PSC devices maintained over 90% of its initial PCE after 850 h in N 2 under 1 SUN illumination and 98% of its initial efficiency after 1000 h in air with 30-35% humidity.
Luan et al. incorporated 2,2,2-trifluoroethanol (TFE) into the SnO 2 ETL [81], as shown in Figure 8d, which showed enhanced electron mobility and optimized energy-band alignment. The modified SnO 2 exhibits a very smooth surface, which attributed to the less trap density at the ETL/absorber interface and inside the perovskite film, leading to lessened carrier recombination. With the addition of oxygen-plasma treatments PCE values of 21.7% were achieved with FF over 80%, as shown in Figure 8e. Unencapsulated devices maintained over 90% of its initial efficiency after 720 h in 30-40% relative humidity, as shown in Figure 8f.  [79]. This approach effectively passivated the charge traps and suppressed carrier recombination at the interface by sulfur functionalization of the SnO2 surface. Functionalized-sulfur atoms can also coordinate with the under-coordinated Pb 2+ ions at the interface. Such strategy resulted in improved device efficiencies of 16.6% to 18.4%. Unsealed devices maintained about 90% of its initial PCE after 1680 h at room temperature.

Bulk Incorporation of Organic Molecules
Incorporation of organic molecules possessing versatile functional groups into SnO2 improves the dispersion of colloids, enhances the electrical properties, and passivates the defects in SnO2. For example, Xiong et al. introduced poly(ethylene glycol) diacrylate (PEGDA), as shown in Figure 8a, into the SnO2 dispersion to prevent aggregations [80], which resulted in more uniform film and well-matched band-energy alignment with the perovskite. PEGDA-modified SnO2 also attributed to passivating the defects at the ETL/absorber interface, as shown in Figure 8b,c. This showed a PCE improvement from 21.8% to 23.3%, with a VOC improvement of 1.09 to 1.14 V. Unencapsulated-PSC devices maintained over 90% of its initial PCE after 850 h in N2 under 1 SUN illumination and 98% of its initial efficiency after 1000 h in air with 30-35% humidity.   [81], as shown in Figure 8d, which showed enhanced electron mobility and optimized energy-band alignment. The modified SnO2 exhibits a very smooth surface, which attributed to the less trap density at the ETL/absorber interface and inside the perovskite film, leading to lessened carrier recombination. With the addition of oxygen-plasma treatments PCE values of 21.7% were achieved with FF over 80%, as shown in Figure 8e. Unencapsulated devices maintained over 90% of its initial efficiency after 720 h in 30-40% relative humidity, as shown in Figure 8f.
Ethylene diamine tetraacetic acid (EDTA) was incorporated into the SnO2 ETL by Yang et al. [82]. Electron transfer is facilitated due to the optimized energy-band alignment, as shown in Figure 8g, and enhanced electron mobility of the modified SnO2. Perovskite film grown on the modified SnO2 also exhibited larger grain size and lower trap density. This led to a PCE enhancement of 18.9% to 21.6%, as shown in Figure 8h. Unencapsulated devices maintained 86% of its initial efficiency after 120 h in continuous 1 SUN illumination, as shown in Figure 8i. The large grain size of the perovskite repressed perovskite degradation at the grain boundaries.
Polyethylene glycol (PEG) introduced into the SnO2 prevented nanoparticle agglomeration, resulting in a dense and uniform film [83]. Such modification of SnO2 improved the wettability and enabled pinhole-free perovskite films, demonstrating a PCE enhancement of 19.2% to 20.8%. Unsealed devices retained over 97% of its initial efficiency after 2160 h in air with 30-80% relative humidity at 28-35 °C.
Polyethylenimine (PEIE) was added into SnO2 by a low-temperature solution process [84]. Such doping of SnO2 resulted in optimized band alignment, larger built-in potential, improved electron transport and extraction, and mitigated charge recombination. Unencapsulated devices retained 82% of its initial efficiency after1,680 h in 40% relative humidity.

Conclusions and Future Directions
In summary, recent progress in modifying the SnO2-ETL bulk and surface properties are discussed. SnO2 has been considered the most promising alternative to TiO2, as it has Ethylene diamine tetraacetic acid (EDTA) was incorporated into the SnO 2 ETL by Yang et al. [82]. Electron transfer is facilitated due to the optimized energy-band alignment, as shown in Figure 8g, and enhanced electron mobility of the modified SnO 2 . Perovskite film grown on the modified SnO 2 also exhibited larger grain size and lower trap density. This led to a PCE enhancement of 18.9% to 21.6%, as shown in Figure 8h. Unencapsulated devices maintained 86% of its initial efficiency after 120 h in continuous 1 SUN illumination, as shown in Figure 8i. The large grain size of the perovskite repressed perovskite degradation at the grain boundaries.
Polyethylene glycol (PEG) introduced into the SnO 2 prevented nanoparticle agglomeration, resulting in a dense and uniform film [83]. Such modification of SnO 2 improved the wettability and enabled pinhole-free perovskite films, demonstrating a PCE enhancement of 19.2% to 20.8%. Unsealed devices retained over 97% of its initial efficiency after 2160 h in air with 30-80% relative humidity at 28-35 • C.
Polyethylenimine (PEIE) was added into SnO 2 by a low-temperature solution process [84]. Such doping of SnO 2 resulted in optimized band alignment, larger built-in potential, improved electron transport and extraction, and mitigated charge recombination. Unencapsulated devices retained 82% of its initial efficiency after1,680 h in 40% relative humidity.

Conclusions and Future Directions
In summary, recent progress in modifying the SnO 2 -ETL bulk and surface properties are discussed. SnO 2 has been considered the most promising alternative to TiO 2 , as it has a high electron mobility and conductivity. SnO 2 also possesses a suitable band structure owning a deep conduction band allowing enhanced electron extraction at the ETL/absorber interface. The wide bandgap of SnO 2 allows most of the light to be absorbed by the perovskite-absorber layer and suppresses UV-related photochemical reaction, improving the device stability. SnO 2 can also be processed at a lower temperature than TiO 2 allowing for flexible applications to be possible. Despite the many advantages of SnO 2 , there are drawbacks, such as defects in the surface and bulk of the pristine films, deteriorating its electronic properties. Such drawbacks can be alleviated by the various surface-modification and bulk-incorporation methods discussed in this review.
Surface-modification and bulk-incorporation methods are discussed as strategies, including elemental doping, metal-oxides bilayers, incorporation of ionic compounds, carbon materials, and organic molecules. Among the five main aspects of modifying the ETL, bulk and surface modification using ionic compounds are generally lower in cost and more stable than carbon materials and organic molecules. Bilayer-metal-oxide approaches are simple and straightforward; however, an additional fabrication step and additional material cost is required. Generally, surface-modification approaches require an additional fabrication step, so bulk incorporation may be considered more fabrication-friendly. Various modification at the surface or the bulk can lead to improved morphology, conductivity, and band alignment of SnO 2 , resulting in improved electron-transport capabilities and reduced carrier recombination. Modification of the surface properties of SnO 2 can also lead to improved quality and crystallinity of the perovskite absorber and reduced interfacial defects at the ETL/absorber interface, attributing to improved PV device performance and stability.
Although such modification approaches of SnO 2 have demonstrated enhanced device performance and stability, there still requires improvement in the operational stability of the PSCs at the MPP under continuous 1 SUN illumination. There have been many reports on the long-term stability of devices exposed to elevated temperatures and high relative humidity conditions. However, there are a few reports on the MPP tracking under continuous 1 SUN illumination which are working conditions of the PSCs, and is an important evaluation of the long-term operational stability of PSCs. Such longterm working-stability evaluation methods will help create a better understanding of the mechanisms for improved working stability and will be critical for future directions for commercialization. Another important factor to consider for future directions will be the scalability of the modification approaches for the SnO 2 ETL, since developing methods compatible with large-area substrates will be essential for commercialization.
Perovskite-tandem applications, with lower bandgap absorbers, such as silicon, Cu(In,Ga)Se 2 , and tin-related absorbers [85][86][87][88][89][90][91], below wider bandgap perovskite-based solar cells, are also future steps to commercialization [92]. Thus, depending on the bottomsolar cell, there may be limitations in fabrication methods or temperature process of the ETL in the tandem configurations. Especially, tandem devices with flexible substrates will have a limitation on the process temperature of the layers in the top solar cell.