Single junction silicon solar cells are approaching their practical efficiency limits, and perovskite/silicon tandems represent the most promising evolution for photovoltaics. Bifacial configurations are especially compelling as they enhance power generation density by harvesting ground reflected irradiance while shifting the optimal bandgap to a bromide lean window that suppresses phase segregation. However, commercial transition of tandem solar cells is hindered by three gaps: unknown reverse bias reliability, unclear photothermal stability, and inaccurate energy yield forecasting. 

The first challenge is reverse-bias instability under partial shading, which remains a key barrier for perovskite devices. Although perovskite/silicon tandems are viewed as a promising solution, the operative mechanism at the subcell level has not been systematically clarified. In this work, we pair controlled reverse-bias experiments with circuit analysis and a measurement-validated pseudo-tandem to link current partition to voltage distribution. We show that silicon-limited operation confines nearly all reverse bias to the robust silicon subcell, enabling high-performance retention under stress. 

The second challenge is photothermal durability, which emerges as the dominant barrier for bifacial tandems once phase segregation is suppressed in the bromide-lean bandgap window required for bifacial operation. We identify perovskite lattice compliance, the propensity of the lattice to elastically deform under stress, as the governing material parameter dictating this durability, demonstrate that compliant lattices mechanically dissipate thermally induced microstrain and suppress nonradiative losses at elevated temperature. Guided by this principle, we achieve a stabilized power generation density of 36.2 mW/cm² under 0.20 albedo, with devices retaining 90% of initial efficiency after more than 1000 hours of ISOS-L-2 testing. 

Finally, we examine temperature coefficients to improve energy yield forecasting, discovering that these coefficients are not intrinsic constants but are highly sensitive to rear irradiance. The transition between subcell limiting states near current matching causes a sign reversal in the short circuit current coefficient and a discontinuity in fill factor. By providing this comprehensive mechanistic framework and practical design rules, this dissertation enables the realization of efficient, stable, and commercially viable bifacial perovskite silicon tandem solar cells.