conversion efficiencies (PCEs) of perovskite solar cells (PeSCs) have already surpassed 25%, [4] which is comparable to current industrial-grade mono-crystalline silicon solar cells. This fantastic efficiency combined with potentially low-cost fabrication makes them more attractive than the most efficient photovoltaic technologies. [5] Similarly, perovskite light-emitting diodes (PeLEDs) are shown to have an extraordinary performance with external quantum efficiencies (EQEs) exceeding 20% in devices made from the solution phase at low temperatures. [6][7][8] We associated these notable improvements on device performance with lessons learned from organic optoelectronics and other semiconductor devices. Typically, many empirical approaches learned from others such as structural optimization, [9] interface engineering, [10] defects passivation, [11] etc., have been considered optimizing the carrier injection/extraction and mitigating undesirable non-radiative recombination losses. [12][13][14][15][16] Nonetheless, there is plenty of room optimizing nonradiative recombination losses in the bulk and at the interfaces to reach the efficiency limits. [17,18] Historically, band misalignments and interfacial defects have been considered the predominant factors responsible for the undesirable recombination at the various interfaces in devices. [19][20][21] In light of band misalignment, the resulting nonradiative recombination losses at the perovskite interfaces cause a substantial reduction in open-circuit voltages (V oc ) of the PeSCs, [14] mainly because of charge extraction suppression. For example, an energy barrier height of >0.1 eV is sufficient to block the extraction of majority carriers and to block the minority carriers from entering the undesirable transport layer, resulting in efficiency losses. [22] The presence of a charge injection barrier is also known to impact the turn-on voltage and device efficiency of PeLEDs. [23] Moreover, the deep-level defect-assisted recombination provides an additional path for intrinsic loss of charge carriers. [24][25][26] However, the origin and distribution of deep-level defects, a kind of defect that are far away from the band edge and can trap charge carriers, are highly debated topics. To date, there are several experimental observations confirming that a considerable density of deep-level defects retains at the interfaces of both the single-crystal and polycrystalline devices, and some surface passivation strategies can considerably eliminate the defects on the surfaces. [27][28][29] Based on photoemission Surfaces and heterojunction interfaces, where defects and energy levels dictate charge-carrier dynamics in optoelectronic devices, are critical for unlocking the full potential of perovskite semiconductors. In this progress report, chemical structures of perovskite surfaces are discussed and basic physical rules for the band alignment are summarized at various perovskite interfaces. Common perovskite surfaces are typically decorated by various compositional and structur...