Progress in Surface Modification of SnO2 Electron Transport Layers for Stable Perovskite Solar Cells

The photovoltaic (PV) performance of perovskite solar cells (PSCs) has rapidly advanced in the recent years; yet, the stability issue remains one of the last‐mile challenges on the road to commercialization. Charge transport layers and their interfaces with perovskites stand for critical tuning knobs that determine the device stability of PSCs. This review focuses on the effects of modification of SnO2 electron transport layers (ETLs) on the interfacial physicochemical properties and stability of PSC devices. In detail, the intrinsic defects, surface hydroxyls, and nonuniform morphology of SnO2 will negatively impact its interfacial physicochemical properties, thus degrading the device stability of PSCs. To tackle these existing issues, three modification approaches, such as surface morphology control, surface physicochemical modifications, and surface composite‐structure design, are categorized. Lastly, future perspectives in further promoting the stability of PSCs from SnO2 ETLs are raised based on the currently unresolved issues from both material and device levels.


Introduction
Since the first invention of perovskite solar cells (PSCs), its power conversion efficiency (PCE) has rapidly surged to 25.8%, [1] with flexible PSCs also reaching a high PCE of 23.6%. [2] However, stability of PSCs is still considerably inferior that have rendered them not able to fulfill the commercialization standards at the moment. [3] So far, the approaches adopted to mitigate the instability issues of PSCs include enhancements of the intrinsic stability of perovskite light-harvesting materials, modification of functional layers, tailored interfacial engineering, and device encapsulations. [4] Due to high ion mobility and intrinsic surface defects (e.g., dangling bonds, uncoordinated/undercoordinated sites) of perovskite materials, [5] the neighboring charge transport layers that sandwich perovskite in between serve the critical roles of not only regulating but also preserving the physical/chemical integrity of perovskites. Out of the charge transport layers, hole transport layers (HTLs) such as spiro-OMeTAD and PTAA with tailored molecular functions and interfacial properties [6] have contributed to improved stabilities of PSCs under moisture and thermal stresses. Yet, the complementary aspects such as optical, chemical, and mechanical stabilities, as mediated by electron transport layers (ETLs), remain relatively underexplored in PSCs with n-i-p architecture.
Among the many ETLs, inorganic ETLs such as SnO 2 , TiO 2 , Zn 2 SnO 4 have desirable properties such as high optical transparency, structural, and thermodynamic stabilities. Due to these advantages, ETL/ perovskite/HTL multilayer stacks can function to effectively protect perovskite materials, constituting pivotal pathways to improve stabilities at device level in response to external stress conditions. Representing an ETL that has retarded chemical reactivity or photocatalytic activity with perovskite materials, SnO 2 is a promising candidate for operationally stable and high-performance PSCs. Moreover, it possesses much greater electron mobility for inhibiting interfacial carrier accumulation and the consequent current-voltage (J-V) hysteresis in solar cells; the moisture-resistant, UV-tolerant natures of SnO 2 even make it potentially viable for long-term, steady-state power output of the incorporated PSCs. [7] So far, PSCs with rigid and flexible substrates have achieved top solar-to-electricity performance with SnO 2 being the ETLs ( Figure 1). Nonetheless, the self-doped defects at the surface of SnO 2 (such as Sn and O vacancies, surface hydroxyls, etc.) exert profound impact to the performance and stability of PSCs. [8] In detail, hydroxyl groups (-OH) and dangling bonds on SnO 2 surface decrease the formation energy of oxygen vacancies and promote oxygen diffusion toward perovskite lattice, [8a,b,d] while causing proton movements from the organic cations, thereby degrading the thermal stabilities of perovskite materials. [9] As such, it can be seen that the major degradation mechanism of SnO 2 -based PSCs of both the chemical properties and PV performance is mainly attributed to the defective buried interfaces between SnO 2 ETLs and perovskites, thus making the modifications at the SnO 2 surface/interfacial locations extremely crucial for stability enhancement in PSCs. To date, most article reviews of SnO 2 -based PSCs mainly focus on the influence of bulk doping and surface modification of ETLs in the PV performance, [10] with inner connections between the surface/interfacial modifications and long-term stabilities under moisture, elevated temperatures, and light irradiation conditions lacking in-depth analysis. The consolidation of this scientific building block, therefore, will inspire relevant ideas to further improvements of stability at both material and device levels.
This review focuses on surface morphology, defects, interfacial stress and energetics of SnO 2 , and the corresponding effects in device stability of PSCs. Based on the underlying structure-property relationships, we further generalize and categorize three surface modification approaches-morphology control, physicochemical modifications, and composite-structure design. We first introduce the optical and electronic properties of SnO 2 ETLs in Section 2, based on which we proceed to the modification approaches of SnO 2 /perovskite interfaces of PSCs in Section 3. This work will advance the understanding of the important and affecting roles of SnO 2 /perovskite interfaces in stabilizing PSCs under long-term PV operation.

Physical Properties of SnO 2 Electron Transport Layers
Since ETLs serve the roles of electron extraction and transport in PSCs, it is desirable for them to exhibit high charge extraction potential, less charge accumulation, and less charge recombination when forming interfaces with perovskite films. In addition, ETLs are desired to possess minimum parasitic absorption so to maximize the light absorbable by perovskite active layer. In this section, we will review the properties of SnO 2 from the four aspects-photoelectric properties, energy band structures, film formation, and defect properties.

Large Bandgap and Optical Transmittance
The optical bandgap of a semiconductor directly dictates the cutoff energy of photons that can be absorbed. SnO 2 has bandgaps ranging from 3.60 to 4.50 eV according to different fabrication methods (Figure 2a), [10b] which renders it transparent to most visible region when compared with other ETL materials, such as TiO 2 , ZnO, and PCBM (Figure 2b,c). In detail, it is the defects, crystallinity, and/or particle sizes from different synthetic approaches that give rise to the different optical bandgaps of SnO 2 , where the types and amounts of defects directly determine the relative coordination between Sn and O elements, thereby affecting the electron orbital overlap and thus the energy gap, [11] as similar to oxygen-deficient TiO x materials. [12] Characteristically, SnO 2 has high transmittance (T ) of solar spectrum above 80% within the spectral range of 400-800 nm (Figure 2b), and thus constitutes minimum parasitic absorption and ensures a full exploitation of solar irradiation by perovskite light-harvesting layers, where the relationship between T and absorption coefficient (α) is described as [13] T % e Àαt where t stands for film thickness, where SnO 2 film typically will not exceed 20-30 nm. Along with the very small α at wavelengths where light is transmissible, the transmittance by SnO 2 is typically closed to unity. The refractive index is also an important factor affecting the light transmittance. Since SnO 2 ETLs have particularly low refractive index in the visible region, and the overall refractive index in combination with TCOs (FTO, ITO) is also much smaller than other types of ETLs. [14,15] As such, sunlight can pass through SnO 2 ETL substrates without much loss solely due to the prolonged optical loss by refraction.  [1,5a,14] and b) on flexible substrates. [2,63] Performance of devices based on SnO 2 ETLs is shown by red curves. Inset: photographs of typical rigid and flexible PSCs. [14b,64] c-e) PSCs employing mesoporous (c), planar (d), and inverted (e) device structures.

Electrical Properties
The intrinsic electrical properties of SnO 2 ETLs govern the electron transport process herein and the eventual charge collection at the conductive electrodes, as evaluated by electrical resistivity, electron density, and mobility. Pristine SnO 2 exhibits relatively low resistivity/resistance. However, obvious rise of electrical resistivity/resistance and reduction of carrier density are resulted upon Mg metal ion doping, which reach above 100 Ω·cm resistivity and 10 15 to 10 16 cm À3 carrier concentrations in the range of 2.5%-20% Mg doping, respectively, despite considerable improvement of electron mobility that falls within 10-55 cm 2 V À1 s À1 (Figure 3a,b). [16] By engineering SnO 2 ETLs for simultaneously improved mobility and conductivity, it was previously demonstrated that SnO 2 nanocrystals (NCs) exhibited boosted charge extraction after functionalized with NbO x wrapping shells, leading to synergistically enhanced electron mobility and conductivity with 900-3000 mA cm À2 current density within 3-5 V electrical bias (Figure 3c,d). [17] The increased mobility will reduce charge accumulation at the ETL/perovskite interface for remarkably suppressed hysteretic behaviors of solar cell devices. The exceptional electron mobility and conductivity after interfacial treatments thus indicate the promising electrical performance of SnO 2 being ETLs applicable in PSCs.

Energy-Level Alignment with Perovskites
Conduction bands (CBs) of SnO 2 are notably lower in energy levels with respect to the perovskites, which renders this material a good electron extractor while effectively inhibiting hole injection from perovskite layer, thus preventing the undesirable carrier recombination. Most importantly, the CB offsets between SnO 2 and perovskites are small as compared to other ETL materials, which can therefore avoid large overpotential and V oc loss while ensuring a facilitated charge transfer, as shown by the schematic band diagrams of a typical n-i-p device (Figure 4a,b). [17] The ideal properties of SnO 2 are further exemplified by the fact that its energy band positions are highly tunable upon specific surface modification (e.g., NbO x surface wrapping) as shown in Figure 4b. This regulative approach is highly useful as perovskites with different compositions are characteristic with Fermi  [20] TiO 2 , [20c,65] ZnO, [66] C 60 , [67] PCBM, [68]   . Electrical properties of SnO 2 ETLs with different elemental doping and functionalization. a) Electrical resistivity/resistance and b) carrier density/mobility at different Mg doping contents. [16] a,b) Reproduced with permission. [16] Copyright 2016, Royal Society of Chemistry. c) Mott-Gurney plots and d) I-V plots of NbO x -enveloped SnO 2 and pristine SnO 2 NCs as ETLs. [17] c,d) Reproduced with permission. [17] Copyright 2021, Royal Society of Chemistry. . Device structure dependent energy-level alignments in PSCs with SnO 2 ETLs. a) Cross-section illustration of a regular n-i-p PSC and b) corresponding energy band diagrams of all functional layers. [17] a,b) Reproduced with permission. [17] Copyright 2021, Royal Society of Chemistry. c) Schematic depiction of inverted p-i-n PSC and d) corresponding energy band diagrams of the functional layers. [18] c,d) Reproduced with permission. [18] Copyright 2018, American Chemical Society. ETLs is not only limited to regular n-i-p device structure, where SnO 2 ETLs can be fabricated on top of perovskites in p-i-n inverted devices through spin-casting relatively nonpolar solutions containing SnO 2 NCs (Figure 4c). [18] Such device architecture, when composited with PCBM interlayer, can further achieve cascaded energy band alignments for minimized charge potential loss ( Figure 4d). As device hysteresis is mainly influenced by the two interfaces between perovskite and charge transport layers in ETL/perovskite/HTL stacks, it is further believed that the matched energy-level alignment between SnO 2 and perovskite will also contribute to the mitigation of potential hysteretic behaviors of PSC devices.

Film Formation and Low-Temperature Processing of SnO 2 Electron Transport Layers
For the preparation of SnO 2 ETLs, various synthetic methods can be adopted such as chemical bath deposition (CBD), [19] atomic layer deposition (ALD), [20c,21] solution spin-coating, dip coating, slot casting, spray coating, and so on. [22] Different synthesis approaches can result in SnO 2 films with different morphological properties that affect the stability of perovskite films and PSCs, [23] where a summary of device stability regarding PSCs with SnO 2 ETLs with different synthesis approach is explicated in Table 1.
Among the various preparation routes, SnO 2 is most well known for its low-temperature solution processing, which does not require elevated temperature conditions for annealing to attain the benign electronic properties, whereas high temperatures between 400 and 500°C are required to reach good crystallinity and structural integrity of TiO 2 as ETLs. Such intensive thermal conditions, however, have prevented the deployment of TiO 2 as ETLs in many device applications with heat-vulnerable substrates, such as flexible PSCs and silicon HJTs. It is exactly because of its low-temperature processing nature that renders SnO 2 ETLs energetically moderate and advantageous in flexible PSCs and large-area PSCs, [24] making them beneficial to the lowcost, upscaled production of PSCs with pronounced application prospects. Specifically, Qi and coworkers previously fabricated large-area SnO 2 ETLs via CBD method and the scalable perovskite solar submodules with aperture area >200 cm 2 , demonstrating a PCE > 15%, while the 5 cm Â 5 cm mini-modules using the same ETL processing exhibited a T 80 operational stability of 1020 h; [25] on the other hand, mini-modules with 22.8 cm 2 aperture areas based on room-temperature-sputtered SnO 2 ETLs were likewise fabricated, giving >12% device efficiency with T80 operational lifetime of 515 h. [26] These results clearly demonstrate the promising potential of SnO 2 ETLs in the large-area photovoltaics with scalable sizes.

Defects in SnO 2 Electron Transport Layers
Although SnO 2 is featured with benign properties such as chemical stability, optical transparency, and low-temperature processability as elaborated earlier, surface and bulk defects nonetheless exist in SnO 2 that affect the structural and electronic properties of the interfaces and PSC devices. In detail, surface defects include oxygen vacancy (V O ), [8a,27] hydroxyl groups (-OH), [8b,d] and impurity ions, while the formation and existence of such defects are related to the surface dangling bonds and uncoordinated/undercoordinated atoms due to the small crystallite sizes and low-crystallinity nature of SnO 2 from its lowtemperature processing. [9] Although numerous bulk defects, namely V O , tin vacancies (V Sn ), tin (Sn O )/oxygen (O Sn ) antisites, oxygen (O i )/tin (Sn i ) interstitials, also exist in SnO 2 ETLs, surface defects rather play more significant roles in affecting the interfacial charge extraction from perovskite absorbers and physical quality of perovskite films deposited on top. To suppress surface defects, efforts have been carried out through surface modification methods to target the large number of dangling bonds and hydroxyl terminal defects on the surface, as detailed in later context.

Approaches to Improve Stability of PSCs Based on Electron Transport Layers
Stability is an important performance parameter of PSCs, which is directly related to the commercialization of PSCs. [28] Based on the aforementioned analysis, SnO 2 has been proved to be a suitable ETL in PSCs. [29] With various surface defects in SnO 2 hindering the optimizing of device stability in PSCs, [30] we summarize three major pathways to improve the stability of PSCs from the perspective of SnO 2 ETLs, including surface morphological control, defect passivation, and composite-structure design, as shown in Figure 5.

Crystal Size Manipulation in SnO 2 Electron Transport Layers
Targeting the low crystallinity issue of SnO 2 ETLs as caused by the low-temperature processing, the in situ regrowth of SnO 2 NCs was reported via introducing optimum trace amounts of surface adsorbed water on FTO or ITO substrates by UV-ozone (UVO) treatments ( Figure 6a). [31] The surface water content exhibited a monotonic increment upon extended UVO treatment duration, but prolonged UVO treatment of TCO substrates led to excessive surface water that caused random agglomeration of SnO 2 crystallites, thus forming incomplete surface coverage of the underlying FTO substrates required for good PV performance of PSCs ( Figure 6a). [31] As a result, SnO 2 ETLs with the optimum morphology and greater crystallinity lead to PSC devices with greater PCE (>20%) with respect to the reference and control device counterparts with less optimized SnO 2 surface morphology. [31] The regulation of SnO 2 ETLs in affecting device performance and stability of PSCs is also manifested in the uniformity of SnO 2 nanoparticle sizes. Polymer heparin potassium (HP) was further utilized to realize homogeneous dispersion of SnO 2 colloidal nanoparticles in solution form, which directly resulted in ordered arrangement of SnO 2 NCs and perovskite films with vertically oriented grains after depositions ( Figure 6b). [32] The highquality growth of perovskite film on top of HP-functionalized SnO 2 ETLs ultimately gave rise to exceptional device stability under continuous one-sun illumination for 1000 h with negligible performance loss, with devices fabricated on pristine SnO 2 ETLs showing more than 20% performance degradation relative to HP-treated counterparts under identical stress conditions ( Figure 6c). [32]

Surface Wrapping of SnO 2 Electron Transport Layers
Although size and morphological control approaches have warranted robust crystallization of perovskite layers on top and stable operational stability of PSC devices, interfacial energetics and surface defects still await a synergetic manipulation for maximally enhanced device performance and stability. It was found that by wrapping SnO 2 NCs with amorphous NbO x exterior layer (Figure 7a), the NbO x -enveloped ETL exhibited optimized energy-level alignment with perovskite with smaller charge potential offset (Figure 7b), thereby facilitating carrier extracting while minimizing V oc loss. [17] Most importantly, unencapsulated PSC devices based on NbO x -enveloped SnO 2 NCs as ETLs exhibited considerably improved device stability under thermal stress conditions ( Figure 7c). [17] The beneficial effects of SnO 2 surface wrapping are further corroborated by similar works. Moreover, SnO 2 NCs as capped with polyacrylamide (PAM) also gave rise to larger perovskite grain size and matched energy-level alignment with the perovskite layer on top (Figure 7d,e). [33] As a result of the ETL modification, unencapsulated solar cell devices based on PAM-functionalized SnO 2 ETLs exhibit much enduring performance under 50% relative humidity (RH) by retaining more than 90% of the original efficiency, in contrast with the control devices based on pristine SnO 2 ETLs (Figure 7f ). [33] 3.1.3. Micro-and Nanostructure Designs in SnO 2 Electron Transport Layers Although SnO 2 ETLs possess low-temperature processability as mentioned in Section 2.3, it can nevertheless be synthesized at high temperatures through hydrothermal methods. Most importantly, in combined with acidic aqueous environment, nanostructured SnO 2 such as nanorod, nanotube, and nanofiber arrays can be grown. [34] Such nanostructured SnO 2 ETLs possessed controllable area densities or wall thicknesses with well-defined, and upright structural orientations. When applied in PSCs, devices based on these nanostructured ETLs showed much enhanced PV performance and long-term stability as compared to the compact ETL-based control PSCs. Similarly, as aggregated from constituent nanocrystals, SnO 2 with microsphere shapes with large specific surface areas, high crystallinity, and tunable sizes can be formed, [35] resulting in considerable PV performance in PSCs.

Effects of Posttreatments on the Surface Morphology of SnO 2 Electron Transport Layers
In contrast with the abovementioned approaches that all adopt in situ control over the morphology of SnO 2 ETLs, posttreatments can also effectively modify the surface morphology of SnO 2 ETLs. In specific, it was reported that the as-synthesized SnO 2 ETLs react with ammonium (NH 3 ) that triggered crystal fusion at the surface; [36] this surface microstructural modification led to optimized interfacial contact with perovskite layer on top by creating nucleation sites. Based on enhanced interfacial charge transport and suppressed structural defects, PSCs based on the NH 3 -treated SnO 2 ETLs showed improved device stability with 86% retained performance after 60 days of aging at room temperature in 40%-50% RH. [36] Likewise, posttreatment etching of SnO 2 ETLs in NH 4 F aqueous solution gave much smoother surface morphology with greater SnO 2 coverage [37] and eventually resulted in higher device PCE and fill factor than the PSCs with pristine SnO 2 ETLs. Importantly, PSCs based on SnO 2 /NH 4 F ETLs showed >80% retention of initial performance, Figure 5. Schematic illustrations of the routes for improving the stability of PSCs by modifying SnO 2 ETLs. Surface modification strategies are outlined as morphological control, physicochemical modifications, and design of composite-structure ETLs. Interfacial stress schematic is reproduced with permission. [38] Copyright 2021, Elsevier.

Surface Physicochemical Modifications
Defects of SnO 2 thin films greatly influence their electronic properties and the physical quality of perovskite crystallization atop, and the interfaces with perovskite therefore represent active sites that regulate the performance of PSCs. [38,39] Passivation of SnO 2 surface defects commonly adopted surface modification methods including UV/UVO, ionic compounds, carbon derivatives, acids, self-assembled monolayers (SAM), fullerene derivatives, macromolecular polymers, where a comprehensive summary of device performance and stability of PSCs with surface defect passivation treatments is given in Table 2.

Passivating Oxygen Vacancies at SnO 2 Surface
Among the many passivation strategies, UVO represents one that can effectively remove oxygen vacancies through decomposing hydroxyls, surface dangling bonds, and other organic contaminants from metal oxide surfaces, thereby eliminating carrier recombination centers for robust interfacial stability. The effect of UVO treatment on operation stability of perovskite lightemitting diodes was previously studied (Figure 8a), and found that the luminance of the fabricated light-emitting devices exhibited notably enhanced stability with UVO treatment on the SnO 2 ETLs (Figure 8b). [40] With a more in-depth investigation, it was confirmed that the reduced oxygen vacancies at SnO 2 surface as functionalized by UVO treatment contribute to a great crystallization quality of perovskite film on top (Figure 8c,d). [41] Furthermore, physicochemical relationship between SnO 2 's surface oxygen vacancies and thermal instability of perovskite thin films was further revealed, as indicated by plasma treatments in different atmospheric gases and a thermal stress at 85°C for 120 h (Figure 8e). [9] As a result of the oxygen vacancy elimination, PSCs based on O 2 -plasma surface treatment exhibited the greatest stability under imposed thermal stress (Figure 8f ). [9] In terms of chemical passivation of SnO 2 surface defects, passivation molecules at the SnO 2 /perovskite interfaces should have both electronegative and electropositive functional groups to passivate the undercoordinated Sn (electron-poor) and O (electronrich) atoms. After passivation, surface dangling bonds associated with these undercoordinated surface atoms as recombination centers are eliminated due to the formation of new chemical bonds. [10a] To meet such requirements of surface passivation, zwitterions are the ideal choices as passivation agents. Zwitterion-3-(1-pyridinio)-1-propanesulfonate was previously used to modify SnO 2 ETL to improve the thermal stability of solar cell device. [20b] The zwitterionic compound led to the formation of interfacial dipole ( Figure 9a) and a changed work function of SnO 2 ETL, thus preventing reversed electron transfer and inhibiting charge recombination. Furthermore, due to the positively charged ions,   (Figure 9d), thus passivating the interfacial defects in a holistic format. As a result of the interface modifications, BGCl-treated solar cell devices exhibit a remarkable 24.4% device efficiency. Without encapsulation, BGCl-treated devices only sustained less than 5% loss of efficiency after 500 h of aging under ambient conditions with 30% RH at 20°C (Figure 9e). [42] The effects of interfacial passivator in defect suppression can be further tuned by tailoring the number and position of functional groups on organic ring structures. It was previously demonstrated that chlorobenzene sulfonic potassium (Cl-BSAK) salts can fill oxygen vacancies on the SnO 2 surface via coordination bonds between Cl and Sn (Figure 10a). [43] In contrast, firstprinciples density functional theory (DFT) calculations revealed the inclined interaction between sulfonate group and perovskite, where most stable configuration was achieved by 3Cl-BSAK on perovskite surface (Figure 10b). [43] Furthermore, it was found that 3Cl-BSAK with the greatest number of Cl was conducive to the greatest performance enhancement of PSCs, in which unencapsulated devices based on 3Cl-BSAK treatment exhibited more than 95% retention of original efficiency after 2400 h of aging in dry environment and 80% retention of initial performance after 800 h of thermal stress at 80°C (Figure 10c). [43] Due to the large deformation of substrates during thermal annealing, flexible PSCs suffer from large interfacial residual stress issues that largely plague the operational stabilities of PSCs under bending or illumination conditions. From mechanistic perspective, the large interfacial residual stress usually stems from interfacial defects or discrepant thermal expansivity between ETLs and perovskites. [44] Therefore, it is imperative to develop modification strategies that can achieve defect passivation of SnO 2 ETLs to resolve the instability issues in flexible PSCs. Non-fullerene acceptor molecule Y6 (fused dithienothiophen[3,2-b]-pyrrolobenzothiadiazole derivative) was previously used as a buffer layer in SnO 2 ETLs (Figure 11a). [45] The Y6 treatment greatly resulted in relaxed interfacial stress, and the asfabricated flexible PSCs showed evidently improved mechanical Figure 8. Effects of surface defect annihilation in the device performance stability. a) UVO treatment for elimination of surface hydroxyls from SnO 2 ETLs. [40] b) Stability of light-emitting devices with or without UVO treatment. [40] a,b) Reproduced with permission. [40] Copyright 2021, The Royal Society of Chemistry. c) Oxygen vacancy healing by UVO treatment. [41] d) Cross-section SEM images of perovskite/SnO 2 without (upper) and with (bottom) UVO treatment. [41] c,d) Reproduced with permission. [41] Copyright 2020, Elsevier. e) Photographs of SnO 2 -based perovskite films with different plasma treatments before and after thermal stress. [9] f ) Device stability with SnO 2 ETLs with plasma treatments in different atmospheric conditions after different thermal stress durations. [9] e,f ) Reproduced with permission. [9] Copyright 2022, The Royal Society of Chemistry.
www.advancedsciencenews.com www.small-science-journal.com durability at different bending radius and bending cycles (Figure 11b,c). [45] Similarly, multifunctional histamine diiodide (HADI) additive was applied to cure the defects at the buried interface of SnO 2 /perovskite (Figure 11d), forming interlayer bridging structures that refine charge transfer and energy band alignments. [46] Consequently, flexible PSCs with HADI-treated SnO 2 ETLs displayed exceptional mechanical stability upon different bending curvature radii (Figure 11e), thus suggesting the transferable efficacy of multifunctional ionic additives in improving both rigid-based and flexible PSCs. [46] 3.

Inhibiting Surface Hydroxyls of SnO 2 Electron Transport Layers
As pointed out in previous context, surface hydroxyls of SnO 2 ETLs play a significant role in causing the instabilities of perovskite layers and solar cell devices, where the existence of hydroxyls not only induce intrinsic defects of SnO 2 ETLs but also negatively impact the buried interface between SnO 2 and perovskites. To mitigate the defective interface as a whole, researchers have come up with novel strategies that used SAMs to optimize the interfaces. [47] A typical SAM molecule is structurally composed of three parts: anchoring group, interterminal group, and terminal group (Figure 12a). [48] To react with surface hydroxyls, anchoring groups of SAMs (typically phosphonic, hydroxyls, and carboxylic acid groups) first bind with the hydroxyls at SnO 2 surface through condensation, whereas terminal groups of SAMs can then have the potentials to fill the halide vacancies of perovskite bottom surface, thus modifying the buried interface. A concept of "molecular glue" at the SnO 2 /perovskite interface was previously proposed by adopting 3-iodopropyl trimethoxysilane (Si(OCH 3 ) 3 CH 2 ) 3 I, I-SAM) as molecular treatment to enhance adhesion toughness between SnO 2 and perovskite, as shown in Figure 12b. [49] The mechanical strength of I-SAM-treated SnO 2 /perovskite interface could be further illustrated by the much less voids/cracks formed upon perovskite film delamination from the substrate (Figure 12d), as compared to the fractured perovskite film without SAM functionalization with rampant morphological defects (Figure 12c). [49] The increased interfacial mechanical stability resulted in profound operational stabilities of I-SAM-treated PSCs under continuous illumination, in comparison with the device counterparts treated with H-terminated SAM (H-SAM) and control devices without SAM treatments (Figure 12e). [49] The observed device stability brought by interfacial I-SAM again corroborates the importance of multifunctional treatments that not only chemically passivate SnO 2 ETLs but also modify the bottom surfaces of perovskite layers. Particularly, the application of SAMs at the ETL/perovskite interface finds the best application scenario in flexible PSCs. As demonstrated in the above context, SAM can formidably enhance interfacial toughness of SnO 2 /perovskite junction. And when applied in flexible PSCs, it was shown that their performance stability under bending stress was greatly improved (Figure 12f, left). [50] In detail, SAM-functionalized flexible PSCs with optimized upper surface LD structural engineering yielded 88% retention of original performance after 10 000 cycles at a bending radius of 5 mm, while showing T 90 operational lifetime of 1000 h under continuous one-sun illumination, which remarkably outperformed the flexible PSC counterparts with untreated SnO 2 /perovskite interfaces that exhibited T 80 lifetime Figure 11. Additive-functionalized SnO 2 ETLs for stable flexible PSCs. a) Schematic illustration of Y6 functionalization in SnO 2 for mechanically durable flexible PSCs. [45] b,c) Device stability of flexible PSCs based on pristine and Y6-treated SnO 2 ETLs at different bending curvatures (b) and at 8-mm radius for different cycles (c). [45] a-c) Reproduced with permission. [45] Copyright 2022, Wiley-VCH GmbH. d) Schematic illustration of PSC device with HADI treatment on the SnO 2 ETLs. [46] e) Stability test of flexible PSCs at different bending curvature radius. Inset: photographs of flexible PSCs under bending conditions. [46] d,e) Reproduced with permission. [46] Copyright 2022, Wiley-VCH.
www.advancedsciencenews.com www.small-science-journal.com of only 600 h, currently topping the operational performance stability among all flexible PSCs reported. [50] The corresponding cross-section morphology of untreated flexible PSCs showed omnipresent channel and interfacial cracks that were distributed after the bending stress, whereas SAM þ LD devices were free of such morphological degradation (Figure 12f, right). By comparing the abovementioned modifications in flexible PSCs, one can infer on the much greater efficacy of interfacial SAMs in curing the surface defects of SnO 2 , and thus the enhancements in mechanical and operational stabilities of flexible PSCs than just the ionic additives.

Enhancing Surface Adhesion of SnO 2 Electron Transport Layers
Although we have shown that extensive efforts have successfully passivated surface defects in SnO 2 ETLs, the stability of PSCs nevertheless also depends on the residual stress sustained by perovskite layers. [51] From fundamental perspectives, residual stress at the interface arises from discrepant thermal expansivity between ETL and perovskite film after thermal annealing. [51c] Thermally induced tensile stress in the perovskite film weakens chemical bonds, reduces the formation energy of defects, and Figure 12. SAMs in achieving surface defect passivation of SnO 2 ETL. a) Schematic diagram of synthetic process of SAMs. [48] Reproduced with permission. [48] Copyright 2020, Wiley-VCH. b) Mechanical toughness of SnO 2 /perovskite interfaces without and with SAMs functionalization. c,d) SEM images of perovskite surfaces as delaminated from substrates without SAM treatment (c) and with iodine-terminated SAM (d). e) Operational stability of PSCs without and with SAM treatments under continuous one-sun illumination and maximum power point (MPP) tracking. [49] b-e) Reproduced with permission. [49] Copyright 2021, American Association for the Advancement of Science. f ) Mechanical reliability of flexible PSCs under bending stress. Left: device stability as function of bending cycle with at a radius of 5 mm. Right: cross-section SEM images of flexible devices with and without SAM at the SnO 2 /perovskite interface after the 10 000-cycle bending. Reproduced with permission. [50] Copyright 2022, Wiley-VCH.
www.advancedsciencenews.com www.small-science-journal.com Small Sci. 2023, 3, 2200108 reduces the activation energy of ion migration, which therefore accelerates the degradation of perovskite film under illumination and electrical bias conditions. [9] In this regard, adamantane derivative molecules (AD, ADCA, and ADAA) containing carbonyl (C = O) functional groups were used to modify the SnO 2 ETL/ perovskite contact for passivating the interfacial defects, while releasing the residual stress therein. The passivation molecule coordinates with the oxygen vacancy defects in SnO 2 while simultaneously binding with the undercoordinated Pb in perovskite (Figure 13a). [52] In contrast, functionalized SnO 2 ETLs with the molecular treatments led to notable relief of residual stress in perovskite thin films (Figure 13b). [52] Among the three molecular modifiers, theoretical modeling further confirms the vital roles of steric hindrance in affecting the interfacial stress relaxation and thus the integral stability of PSC devices (Figure 13c), where the stabilizing effects exhibit an order of ADAA > ADCA > AD on unencapsulated devices under thermal stress, continuous one-sun illumination and ambient conditions (Figure 13d). [52] The results indicate that unlike chemical modification of SnO 2 /perovskite buried interface, the ordered arrangement of passivation molecules according to their mutual steric hindrance is also an important factor for consideration in regulating interfacial residual stress. Instead of regulating the interfacial residual stress through organic molecule's steric hindrance, polystyrene (PS) was introduced into PSCs as a buffer layer between SnO 2 ETL and perovskite film, as shown in Figure 14a. [53] Due to the soft polymer nature, PS formed continuous film and properly alleviated the interfacial strain between SnO 2 ETL and perovskite film by reducing the modulus mismatch, thereby improving the structural integrity of perovskite lattice. Consequently, PSCs with PS interfacial treatment showed greatly enhanced operation stability under continual one-sun illumination in ambient atmosphere, as well as illumination-dark cycle stability for more than 100 h (Figure 14b). [53] Instead of adopting the mindset of interfacial buffer layer, an idea of implanting volatile organic additiveammonium formate (NH 4 COOH) within SnO 2 ETL was previously devised, where NH 4 COOH vaporized during perovskite film annealing to realize an in situ and integral modification of SnO 2 ETL, perovskite layer and their interface (Figure 14c), thus achieving a formidably released interfacial stress (Figure 14d). [44] Most profoundly, perovskite films on flexible substrates functionalized with such preburied NH 4 COOH additive exhibited minimum morphological degradation (Figure 14f ), in contrast to the untreated films that showed obvious cracks and delamination after 4000 cycles of mechanical bending (Figure 14e), much alike to the strengthening effects of I-SAM on the interfacial toughness as discussed above. [44] However, differentiation needs to be made between the treatments on SnO 2 ETLs by halidecontaining molecules (e.g., I-SAM) and the ammonium-based volatile additives (e.g., NH 4 COOH) as owing to their different working mechanisms of interfacial modifications. In specific, previous research revealed that light-induced performance degradation of PSCs mainly arose from the generation of iodide Figure 13. Strain relaxation of SnO 2 /perovskite interfaces for stable device performance. a) Schematic illustration of SnO 2 /perovskite interfaces as passivated by adamantane derivative molecules. [52] b) Grazing-incidence angle-dependent lattice spacings of perovskite films on SnO 2 ETLs with different molecular interfacial treatments. [52] c) Theoretical models of SnO 2 and perovskite surfaces with different adamantane molecular passivators. [52] d) Stability tests of control and target devices with molecular treatments under 60°C thermal stress (top), continuous one-sun illumination (middle), and 15%-20% RH conditions (bottom). [52] a-d) Reproduced with permission. [52] Copyright 2022, Wiley-VCH.
www.advancedsciencenews.com www.small-science-journal.com interstitials (I I À and I I þ ) at the perovskite/charge transport layer interfaces; where electric-bias induced degradation also originated from the interaction between I I À and hole carriers transferred from ETL. [54] Such device degradations as dictated by halide-defect formation and evolution behaviors thus signify discrepant passivation effects by different chemical treatments of SnO 2 ETLs.

Surface Composite-Structure Designs
From the energy level and electronic property perspectives, ETLs need to have high electron mobility, fine energy-level alignment, and good surface coverage of TCO substrates. Also, an optimized interfacial energy-level structure is exceptionally important to good charge extraction and transport, a basic requirement for reduced interfacial charge accumulation for good device stability. [55,56] To meet these material characteristics, a meticulously composited ETL structure is demanded, such as double ETLs and planar/ mesoporous composited structures, where Table 3 explicates a comprehensive summary of device performance and stability based on composite-structured ETL designs from recent works.

Suppressing Interfacial Charge Accumulation by Double Electron Transport Layers
Interfacial charge accumulation is detrimental to the stabilities of PSCs by accelerating ion migrations, thus triggering lattice decomposition of perovskite materials. [57] As a typical example, it was previously shown that the energy levels of SnO 2 ETLs can be effectively tuned by composing low-amount and high-amount NH 4 Cl-doped composite SnO 2 ETLs (Figure 15a), [58] in which the charge potential difference between composite SnO 2 ETLs and perovskite layers can be properly enlarged to promote lightinduced charge extraction from the latter, thus facilitating the subsequent carrier transport to TCO substrate and avoiding charge accumulation (Figure 15b). [58] A deeper mechanistic picture of the principles was further built, which needs to be followed for designing double ETLs, where composite order of TiO 2 and SnO 2 need to be scrutinized in terms of their free energy differences (ΔG) with perovskites and the resultant electron mobility within the ETLs (Figure 15c,d). [59] Moreover, composite-structure designs featured with SnO 2 -based dual ETLs in p-i-n structured PSCs have also been reported, where C 60 /SnO 2 NCs and PCBM/SnO 2 NCs double ETLs were employed on top of perovskite layers to enhance electron extraction. These composite structures, as compared to pristine C 60 and PCBM ETLs, showed much more cascaded energy alignments with perovskites, thus resulting in mitigated hysteresis, overall performance, and long-term stability of PSCs. [18,60] Based on these designing principles, researchers have constructed double ETLs with largely enhanced operational stabilities of the corresponding PSCs under different environmental stresses, whose performance and stability profiles are summarized in Table 3. Figure 14. Residual stress release in perovskite layers via interfacial modifications. a) Schematic depiction of stress release of SnO 2 /perovskite interface by PS layer. [53] b) Operational stability under continuous one-sun illumination (left) and illumination-dark cycle stability (right) of PSC devices without and with PS buffer layer. [53] a,b) Reproduced with permission. [53] Copyright 2019, Wiley-VCH. c) Schematic illustration of preburied additive for cross-layer treatment of SnO 2 /perovskite interface. [44] d) d-spacing-sin 2 ψ method for calculating the residual stress of perovskite films with and without NH 4 COOH preburied additive. [44] e,f ) Cross-section SEM images of PSC devices after 4000 bending cycles with (e) and without (f ) preburied NH 4 COOH in SnO 2 ETLs. [44] c-f ) Reproduced with permission. [44] Copyright 2022, Wiley-VCH.

Enhancing Mechanical Stability by Interpenetrative Electron Transport Layers
In addition to the interfacial charge accumulation, device stability is also dependent on the detailed surface morphological properties of ETL composites. Different from the mesoporous ETL structures, where the porosity of ETL materials such as TiO 2 is realized through thermal sintering the solution-deposited nanoparticle precursors, [61] interpenetrative SnO 2 ETLs can also be achieved by the chemical reaction between the implanted precursor dopants therein and perovskite layers on the top. It was previously found that porous and interpenetrating SnO 2 ETL as generated from the evaporation of preburied formamidinium iodide (FAI) (Figure 16a) could led to a more cascaded energylevel structure between SnO 2 and perovskites (Figure 16b). [62] Also, the effects of the interpenetrating SnO 2 /perovskite interface can profoundly enhance physical adhesion of perovskite film on the SnO 2 ETL substrate, as manifested in the phenomenally enhanced mechanical strength that is free of interfacial delamination after bending stress (Figure 16d), in comparison with the perovskite film without the SnO 2 porosity treatment ( Figure 16c). [62] In the end, PSCs based on interpenetrating SnO 2 /perovskite interfacial design exhibited greater operational stability under continuous one-sun illumination as imparted from maximum power point (MPP) tracking ( Figure 16e) and mechanical durability of flexible PSCs with bending stress conditions (Figure 16f ). [62]

Summary and Outlook
This review analyzes the interconnection between SnO 2 ETLs and device stability as the focal point, with topics ranging from material properties and film formation of SnO 2 ETLs to the passivation and modification of interfacial defects. Summarizing the aforementioned works, degradations to the PV performance of PSCs related to the SnO 2 interface can be deduced into three categories: 1) surface defects of SnO 2 ETLs that facilitate the oxygen diffusion in perovskite layers that weaken the thermal/chemical stabilities of perovskite materials; 2) charge carrier accumulation and the related ion migration due to interfacial energy level misalignment, causing device instability under illumination and electrically biased conditions; and 3) interfacial stress induced by the lattice mismatch between perovskite and SnO 2 ETLs, causing weakened mechanical durability and performance degradation that are most significant in flexible PSCs. Accordingly, three manipulation strategies imposed on SnO 2 surface to tailor Figure 16. FAI-modified SnO 2 ETLs for interpenetrating interfaces in stable PSCs. a) Schematic illustrations of the preparation method of interpenetrative SnO 2 ETL and b) energy-level cascading between perovskite and pristine/interpenetrative SnO 2 ETLs. [62] c) Cross-section SEM images of perovskite films deposited on pristine and d) interpenetrative SnO 2 ETLs before and after mechanical bending. [62] e) Stability tests of rigid PSCs under continuous one-sun illumination and f ) flexible PSCs with mechanical bending. [62] a-f ) Reproduced with permission. [62] Copyright 2021, Springer Nature.
its physical and electronic compatibility in PSCs-surface morphological control, defect passivation, and composite-structure designs, have been utilized to improve the PSCs' PV performance and operational stability. Several substantiated conclusions are deduced as: 1) Surface morphological profiles of SnO 2 ETLs directly affect the film growth and crystallization quality of perovskite layers atop. 2) Hydroxyl groups and the existence of dangling bonds on the SnO 2 surface lead to reduced formation energy of oxygen vacancies that will accelerate deprotonation process of perovskite lattice under illumination and elevated temperature conditions, thus affecting stabilities under light and thermal stresses. 3) Composite-structure designs are conducive to reduced energy band offsets between ETLs and perovskites, thus inhibiting interfacial carrier accumulation and J-V hysteresis of PSCs. In summary, surface modifications of SnO 2 ETLs have obtained prominent effects to the stability enhancement of PSCs against different environmental stresses, based on which we propose future research directions that were not fully accounted for in prior works, as detailed below.

Multifunctional Molecular Passivation
The defective buried interfaces of SnO 2 /perovskite are complex physical and chemical systems, which not only involve surface defects such as hydroxyls and dangling bonds but also the causal lattice mismatch, interfacial delamination, and etc. Synergistic solution to all these problems demands customized passivation reagents (e.g., molecules with multiple functional groups, fieldeffect dipole molecules, polymers with special structural designs) with multiple functionality to simultaneously suppress surface defects of SnO 2 ETLs while gaining integrity of interfacial mechanics, so as to improve stability and PV performance at the same time.

SnO 2 /Perovskite Integral Modifications
Even though SnO 2 has been optimized with tailored modifications that give rise to stable performance of PSCs, it is after all inorganic in chemical composition and has large bulk modulus. It thus constitutes phenomenal lattice mismatch and large thermal expansivity discrepancy with organic-contained perovskite films on top, thereby forming unavoidable interfacial residual stress upon film annealing. Therefore, constructing tailored interlayer designs is necessary to the strain relaxation at the buried interface, which is especially meaningful to the settings of flexible PSCs due to the massive thermally induced substrate deformation. Among the many manipulative routes, preburying volatile additives in SnO 2 ETL can lead to effective regulation of electron mobility, interfacial stress, and defect density of perovskite film as a whole, thus forming gradient distribution of additive ingredients across the SnO 2 /perovskite interface.

Surface Composite-Structure Designs
In terms of double-ETL designs, while the bottom layer that interfaced with TCO substrates remain as planar structure, the top layer that contacts with perovskite can be selectively modified into porous or mesoporous morphology, which can accordingly enlarge the contact area with perovskite for enhanced interfacial adhesion while fulfilling the cascaded energy level alignments for efficient carrier extraction/transport.