Fig. 1. （a） Schematic of a hexahedral element formed by infinite material division； （b-c） Schematic XRD configurations for the （b） out-of-plane and （c） in-plane measurement^［11］； （d） Residual strain distribution in the depth of 50， 200， 500 nm for the tensile-strained film （measured （points） and Gauss fitted （line） diffraction strain data as a function of sin²φ）. The error bar indicates standard deviation of the 2θ^［13］； （e） Schematics for in-situ tensile testing of perovskite materials； （f） Tensile stress-strain curves for MAPbI₃ and MAPb（I_0.87Br_0.13）₃^［14］； （g） Chart of （G_C/E）^1/2 value for photovoltaic materials （semiconductors， metals， and elastomers）^［15］

Fig. 2. Illustrations of perovskite solar cells （a1） without and （a2） with a protective layer consisting of a parylene film^［27］； （b） The angular view SEM image of nanocone arrays of PDMS with aspect ratio of 0.5； （c） The angular view SEM image of i-cone arrays of epoxy with aspect ratio of 0.5； （d） Finite element mechanical modeling for perovskite solar cell based on flat and 0.5-aspect ratio-i-cone-epoxy substrates^［28］

Fig. 3. （a） Dark-field optical microscopy image of the orthogonal AgNW arrays； The inset shows fast fourier transform image of orthogonal AgNW arrays. （b） An ultrathin and lightweight PSCs device mounted onto the surface of a leaf； （c） Variation in normalized device efficiency of PSCs using ITO and orthogonal AgNW electrodes under repeated （1 000 times） bending cycles^［31］； （d） Device architecture of the all-carbon-electrode-based flexible PSCs； （e） PCE evolution of the flexible devices upon increasing bending curvature radius after 200 bending cycles； （f） Bending durability of the ITO/PEN-based and all-carbon-electrode-based flexible solar cells as a function of bending cycles （bending radius is fixed at 4 mm）^［30］； （g） Resistance changes of PET/ITO and PET/IZVO TCEs during the dynamic fatigue twisting test with a fixed twisting angle of 20°； （h） Surface FESEM images of PET/ITO and PET/IZVO TCEs after the dynamic fatigue twisting； （i） XRD patterns of the IZO thin film and IZVO thin films deposited with varied concentration of V atoms on glass substrates； （j） Normalized PCE， （k） R_s and R_sh （measured on a flat surface） of PSCs with ITO and IZVO electrodes as a function of bending cycle at a radius of 5 mm^［33］

Fig. 4. （a-d） The SEM images of CH₃NH₃PbI_3-xCl_x films without and with different amount of H₂O； （e） PCE of FPSCs as a function of bending cycles with different radii of 10 mm and 5 mm. R_b is the radius of curvature^［34］； （f） Schematic illustration of interaction between PCL and perovskite； （g） Efficiency change of PSCs optimized by PCL additive after continuous operation for 400 h； （h） The evolution of normalized PCE as a function of different curvature radius of the FPSC^［35］； （i） J-V curves of FPSCs with different concentrations of DS additive； （j） Schematic of the chelated intermediate and formation mechanism of perovskite using the DS additive^［34］

Fig. 5. （a） Schematic illustration of grain boundary passivation of the perovskite film with s-ELA and its stitching effect for high mechanical endurance during stretch and release cycles； QNM-AFM characterization of MHP film （b） W/O s-ELA and （c） W/ s-ELA， with the corresponding （i） topography， （ii） Young’s modulus and （iii） adhesion images^［37］； （d） The schematic diagram of PSCs structure of PEN/ITO/SnO₂/Perovskite/Spiro-OMeTAD/MoO₃/Ag and the enlarged illustration of the structure of borax as well as the interaction between borax and perovskite at grain boundaries； The indoor PCE decay curves of the pristine and target flexible devices （e） under different bending radii and （f） under different bending numbers with the bending radius of 5 mm^［38］

Fig. 6. （a） Schematic diagram of the formation mechanism and working principle of the PEDOT：GO adhesive interface layer； （b） Viscosity tests at room temperature of different adhesive interface layer^［40］； （c） Cross-sectional SEM image of FPSCs and schematic illustration of the device structure； （d） Normalized PCE for the FPSCs with different bending radii and the inset shows the corresponding photo of the cell under different bending radius； （e） PCE changes of FPSCs with or without Al（acac）₃ boundary layer as a function of time （up to 1000 hours） under a working environment with relative humidity above 50%^［41］

Fig. 7. Optical images of perovskite thin films containing different mass ratios of TUEG3 additives （a-b） before and （c-d） after annealing at 100 ℃ for 1 h； （e） Bending steps at a 1.5 mm radius， 50 times， and healing steps at 85 ℃ for 30 min. Solid lines represent median data， while dashed lines represent mean data^［43］； （f） Supramolecular repair mechanism of the AD-23 adhesive； （g） J-V test of FPSCs without and with AD-23 adhesive； （h） Performance change of PSCs with AD-23 after bending 10，000 times at 2.5 mm radius； （i） SEM images of flexible control device， FPSC with AD-23 addition after bending test and FPSC with AD-23 addition after bending test and then thermal healing at 70℃^［44］； （j） Preparation and dynamic nature of the self-healing PU； （k） Normalized average PCE as a function of stretching cycles with 20% stretching for PSCs without s-PU， with s-PU， with s-PU and annelaing^［42］