Organic-inorganic lead halide perovskites have attracted great attention as light absorbers of hybrid perovskite solar cells (PSCs). Among the organic-inorganic hybrid perovskites (OIHP), CH3NH3PbI3 (MAPbI3) and CH(NH2)2PbI3 (FAPbI3) have been extensively studied as light absorbers of PSCs. PSCs are expected to be next-generation solar cells due to their high conversion efficiency and low fabrication cost. However, power conversion efficiency of PSCs declines shortly after being fabricated, because of degradation of hybrid perovskite layers. OIHP such as FAPbI3 decomposes easily into PbI2 and transforms to d-phase from a-phase, which is suitable for the application to the solar cells, under ambient condition [1]. It was reported that partial substitution of inorganic monovalent ions such as Cs and Rb ions at FA site is effective to improve the phase stability of a-phase [2]. This can be easily explained by the Goldschmidt tolerance factor [3], which is an indicator for the stability of perovskite structure. Ionic radius of FA is too large, but the partial substitution of Cs or Rb can make the average ionic radius smaller than that of FA, which stabilizes the a-phase. For example, d-phase is the stable phase in FAPbI3 at ambient condition, which transforms to a-phase above 165 ℃ [4]. However, the partially substituted FAxCsyPbI3 is stable even at room temperature. The perovskite layers are grown on electron transport layers (ETL) such as TiO2, SnO2, and ZnO by spin-coating method. Although TiO2 has been widely used for ETL of PSCs, SnO2 has been also applied to ETL because of its excellent band energy level alignment with that of the perovskite layers and its higher electron mobility [5]. In addition, devices using SnO2 as ETL decrease their photoconversion efficiency more slowly than using TiO2 under ambient condition, which suggests that decomposition processes of the perovskites on TiO2 and SnO2 is different [6]. In the fabrication process of ETL, SnO2 is annealed after spin-coating. If the annealing temperature is high above 450 degrees Celsius, rutile structured crystalline SnO2 thin film grows, while only amorphous structured SnO2 thin film appear by the low temperature annealing, such as between 150 and 200 degrees Celsius. Some studies have shown that amorphous SnO2 fabricated by low temperature process increases performances of devices compared with crystalline SnO2 [7, 8].In this study, FA1-xCsxPb(I1-yBry)3 was synthesized by spin-coating method on the SnO2 layers and annealed at various temperature to evaluate the dependence of SnO2 substrate structure on the stability of perovskite layers. Stability of all the samples here synthesized were monitored with the X-ray diffraction technique at 0, 1, 2, 3, 4, 24, 96 and 168 hours kept in air after fabrications. Fig.1 shows the intensity of diffraction peaks of d-FAPbI3 (100) as a function of time. Although the intensity of d-FAPbI3 peaks increased in FA0.85Cs0.15PbI3 and FA0.85Cs0.15Pb(I0.85Br0.15)3 formed on SnO2 annealed at 450 ℃ , it unchanged in FA0.85Cs0.15PbI3 and FA0.85Cs0.15Pb(I0.85Br0.15)3 on SnO2 annealed at 180 degree Celsius. These results sugessts that the stability of a-FAPbI3 grown on SnO2 annealed at 180 degree Celsius is better than that annealed at 450 degree Celsius. The crystal structures of SnO2 annealed at 200, 300 and 400 degree Celsius were characterized with the Glazing Incident X-ray diffraction technique. The surface morphology of SnO2 were obsereved with the scannning electron microscope, which suggest SnO2 surface became rougher as annealing temperature increased. The chemical composition of the surface of SnO2 layers were investigated by Auger electron spectroscopy, from which Cl peaks were obsererved in SnO2 annealed at 200 ℃ and 300 ℃, while Cl were not detected in SnO2 annealed at 400 degree Celsius.[1] T. Leijtens et al., J. Mater. Chem. A.5, 11483-11500 (2017)[2] M. Saliba et al., Science 354, 206-209 (2016)[3] C. C. Stoumpos et al., Acc. Chem. Res. 48, 2791-2802 (2015)[4] Z. Li et al., Chem. Mater. 28, 284-292 (2016)[5] L. B. Xiong et al., J. Mater. Chem. A.4, 8374-8383 (2016)[6] Q. Dong et al., Nano Energy 40, 336-344 (2017)[7] W. J. Ke et al., J. Mater. Chem. A.3, 24163-24168 (2015)[8] K.-H Jung et al., J. Mater. Chem. A.5, 24790-24803 (2017) Figure 1