Since the initial reports of perovskite solar cells (PSCs) in 2009 the power conversion efciencies (PCEs) of such devices have risen from 3.8%1 to 25.2% . Perovskites have many properties which make them attractive materials for solar cell applications, including efcient light absorption, tuneable band gap, long charge-carrier lifetimes and high defect tolerance3–8 . However it is the relative ease by which perovskite flms can be deposited from solu-tion that has generated the greatest interest, as this potentially allows high volume manufacture of photovoltaic modules at low cost and low temperature. Tis could allow a dramatic reduction in the energy payback time of a commercial module to less than half a year9 . In order for this to become a reality, it is necessary to demonstrate that perovskite solar cells can be fabricated using industrially compatible coating techniques. Currently most perovskite device optimisation is performed using spin coating; a simple and reliable technique capable of producing highly uniform thin flms. However spin coating is usually only suitable for coating small substrates on the order of square centimetres, rather than square meters10. Spin coating is also
a wasteful coating process, as the majority of the ink to be coated is thrown from the substrate during deposi-tion. Consequently, increased research efort is being focused towards scalable deposition techniques such as blade-coating, slot-die coating, inkjet printing, and spray-coating.
based on an ‘inverted’ (p-i-n) architecture where the PEDOT:PSS, CH3NH3PbI3−xClx perovskite and PCBM were deposited by spray-coating, with devices reported having an average PCE of 7.1%28. By switching to an n-i-p architecture and sequentially spray-coating compact TiO2, mesoporous TiO2, CH3NH3PbI3−xClx perovskite and spiro-OMeTAD, we have also been able to create PV devices yielding an average PCE of 9.2%29. Recently perovskite solar cells having a PCE of 20% have been demonstrated in which all solution processed layers (namely SnO2, perovskite, and spiro-OMeTAD) were deposited via air-blading30. Tis arguably represents the state-of-the-art for scalable perovskite deposition methods. In this article we perform a similar study, build-ing upon our previous work to spray-coat all solution processable layers within a device, and incorporate a low vacuum treatment step to crystallise a “triple-cation” perovskite layer. We then scale up our fabrication process to larger area substrates to fabricate efcient PSCs. We believe that this represents an important proof-of-concept that could be transferable to an industrial manufacturing environment.
spraying this layer. For a more in depth analysis of spray-coated np-SnO2 flms we direct the reader to other recent work in which we develop this process33. On spray-coating the spiro-OMeTAD (Device C) we fnd that there is only minimal loss in device perfor-mance, with devices having an average PCE of 17.0±2.9%. Here, we have carefully controlled the thickness of the spray-cast spiro-OMeTAD (≈200 nm) by adjusting the concentration of spray-solution so that it matches that of the spin-cast layer. Indeed, we have found that if the spiro-OMeTAD layer thickness is greater than this optimum value, it results in reduced performance (see Fig. S4) and increased hysteresis. As a result of this optimi-sation, when all three layers were spray-coated (Device D) we were able to produce PSCs with an average PCE of 16.6±2.4% and a peak efciency of 19.4%.