Low-temperature solution processed ternary chalcogenide halide solar cellsContentsTitle                        PageAbstract           31. Introduction           41.1  Solar cells          41.2  Perovskite solar cells         51.2.1        Crystal structure        71.2.2        Evolution of perovskite solar cells      81.2.3        Challenges and Issues        101.2.4        Alternate materials        112        Research objective         133        Experimental          163.1  Chemicals & materials         163.2  Synthesis          163.2.1        Synthesis of chalcogenide complexes      163.2.2        Synthesis of ternary chalcogenides      17Thermal decomposition      17Microwave synthesis      18Solar cell fabrication.         18Characterization         20     Material characterization       20     Optoelectronic characterization      214        Results and discussions         214.1 Material characterization        214.2 Optoelectronic characterization       325        Scope and future plan         415.1 Milestone plan         436    Acknowledgements         447    References           45 Chapter 1IntroductionAbstractPerovskite solar cells (PSCs) are a promising renewable alternative technology for generating electricity at low manufacturing cost.  Devices based on alkylammonium lead halides have achieved an impressive efficiency of over 20%,[1] but there is now a focus on replacing lead with non-toxic substitutes as this is particularly relevant for commercialisation. Considering non-toxic hetero-valent cationic substitution for lead could result in the formation of a double perovskite with an expected formula of A2MCX6.[2] Oxides of these type of multi cationic compounds have been already explored due to their multiferroic and ferroelectric properties.[3] Ternary metal chalcogenides materials with their broader absorption and suitable bandgap tunability can be considered as promising solution-processable photovoltaic materials. Appropriate selection of materials for ternary chalcogenides could result in the high flexibility of bandgap tuning without the toxicity issues. AgBiS2 with its favourable photovoltaic and optoelectronic properties,[4],[5],[6],[7] achieved encouraging photovoltaic efficiency. [8]  But, the need for post-treatment for the pure material and recombination loss issues with the device architecture remains as an issue.Here, we introduce two simple methods for the synthesis of pure colloidal AgBiS2 nanoparticles under ambient conditions at low temperature, using the bismuth- aryl dithiolate as a starting material which has not been previously investigated. Additionally, this I–III–VI2 family of compounds, synthesised by this novel methods, shows the potential to act as a photo-absorbing and charge-transporting layer simultaneously in mesoporous solar cells. This material will subsequently be used for the development of materials with a double perovskite structure.Introduction1.1    Solar CellsThe ever-burgeoning demand for energy and resources is leading to scientific investigations that aim to mitigate and eventually solve the looming energy crisis – where to find an additional 13.5 TW of energy required by 2050 to secure our collective future?[9]. Considering the impact of climate change caused by rising CO2 levels, other environmental effects and resource depletion, renewable energy sources represent a superior option. Among the renewable energy sources, sunlight has the highest potential as it delivers 174000 TW of solar energy to the earth’s surface, more than 8000 times current energy consumption. I  Conventional silicon solar cells are very efficient in energy conversion but the need for high purity materials demanded higher fabrication cost, which resulted in a longer payback time.[10] Even though module prices were brought down by a factor of five in last six years10,slow progress directed researchers to search for alternative materials a few decades ago. These alternatives comprise of a broad variety of photoactive thin film materials, ranging from gallium arsenide (GaAs), cadmium telluride (CdTe), indium phosphide (InP), copper indium gallium selenide (CIGS) to organic, inorganic or hybrid composites and even thin film crystalline and amorphous silicon.[11] However, the toxicity and the scarcity of raw materials limited further development of many of these devices.Third generation photovoltaic technologies, principally developed following O’Regan and Grätzel’s 1991 paper reporting dye-sensitised colloidal TiO2 film based solar cells with efficiencies exceeding 7 %, then came into the picture.[12],[13]In more recent years, the application of organic dyes and porphyrin dyes in DSSCs has resulted in champion devices performances above 13%.[14] Although the utilisation of less complex manufacturing processes reduces the expenses and energy reimbursement time, the conversion efficiencies (PCEs) of these devices still fall notably behind those of conventional silicon solar cells and there are longevity issues related to the use fluid electrolytes.[15] In a parallel development, the organic dyes were substituted by quantum confined semiconductor nanoparticles (e.g., CdTe, CdSe, PbS, etc.) with the assembled devices termed quantum dot sensitised solar cells (QDSSC). But, toxicity issues and failure to implement the multi-exciton generation effect somewhat disappointed the enthusiasts.The advancements in the field with the knowledge gained from three generations of photovoltaic devices gave birth to entirely new kind of organic-inorganic perovskite solar cells (PSCs).Figure 1: Photovoltaic device efficiency chart (replotted with reference to Ref [16]).1.2 Perovskite Solar CellsThe perovskite solar cells evolved from mimicking DSSCs but later transformed as an entirely novel technology (shown in Figure 2), bringing greater attention as its PCE exceeded 22% only in few years (shown in Figure 1). An explosion of interest in PSCs occurred which  led researchers to investigate combinations of different electron transport materials (TiO2, SnO2, Al2O3, ZrO2, ZnO, PCBM, etc.), perovskite materials (CH3NH3PbX3, ((NH2)2CH)PbX3, CH3NH3SnX3 with X = Cl, Br, I)[17], hole transport materials (Spiro-OMeTAD (2,2?,7,7?-tetrakis-(N,N-di-p-methoxy phenylamine)9,9?-spirobifluorene), P3HT, PTAA, CuI[18] Cu-phthalocyanine[19] CuSCN,[20] various metal oxides and etc.)[21], and counter electrodes (Ag, Au, graphite, etc.)[22] based on band alignment (Shown in Figure 3) to increase the efficiency of these promising devices.      Figure 2: Evolution of perovskite solar cell architecture from DSSCs (reproduced from ref [23]).Figure 3: A summary of the energy levels of various hybrid perovskites, charge transport materials and the work function of counter electrode materials (reproduced from ref [24]).1.2.1 Crystal StructurePerovskite is a general nomenclature for a class of compounds having a three-dimensional crystal structure with formula ABX3 (Figure 4), where the cation A occupies at the eight corners of the cube and cation B is situated at the body centre and is surrounded by 6 X-anions (resides at the face centres) in an octahedral [BX6]4- cluster.[25]Figure 4: Crystal structures of standard perovskite materials and perovskite solar cell absorber materials. (Reproduced from 23)In the case of organic?inorganic hybrid perovskites, at least one of the “A”, “B”, and the “X” ions are inorganic; the “A” cations that are typically organic (most commonly CH3NH3+ or HC(NH2)2+), the “B” metal cations are normally divalent metal ions, such as Pb2+, Sn2+, Eu2+, Cu2+ etc.)[26], while the anions “X” are halides such as Cl–, Br– ,I–. Combinations of these components form an expansive group of organic–inorganic perovskites.17 Classical inorganic perovskite materials comprise strontium titanate (SrTiO3)and calcium titanate (CaTiO3). Due to the interaction of the charge, spin electronic properties and structural properties, this group of materials display some exciting phenomena such as superconductivity, ferroelectricity, magnetoelectricity, magnetoresistance, anti-ferroelectricity, anti-ferromagnetism, and so forth.[27]Treating all ions as rigid spheres and considering close packing yields the Goldschmidt’s Tolerance Factor concept, [28],[29],[30] given by equation 1.t=RA+RX?2(RB+RX)        (1)where RA, RB, and RX are the ionic radii for the corresponding ions. The tolerance factor should be close to 1 to form the cubic perovskite structure. Experimentally, most of the perovskites have tolerance factors between 0.8 and 0.9, although their structure is slightly distorted.29,30 The evolution of the perovskite solar cell architectures, such as mesoporous, planar, inverted, tandem and so on, are shown in Figure 1 and the most common of these in Figure 5. In mesoporous PSCs (Figure 5 a), the perovskite material is infiltrated into a layer of mesoporous TiO2. The n-type material is in contact with the active layer for the extraction of electrons (hole blocking layer) and a p-type material is used for hole extraction (electron blocking layer). When the perovskite material absorbs light, the photo-generated electrons are injected into the mesoporous TiO2. Figure 5 b demonstrates a schematic diagram of a planar device architecture where a perovskite flat layer is sandwiched between two chosen contacts. In planar, after light absorption both charge generation as well as charge extraction occurs in the perovskite layer. These two architectures are then altered in different approaches to accomplish diverse device configurations. For example, starting the assembly from the counter electrode to anode results in an inverted structure whereas stacking one device above another (perovskite or some other solar cells) results in tandem architectures.Figure 5: Examples of device architectures. a) Mesoporous and b) Planar perovskite solar cells.Perovskite solar cells are considered as promising solar cells because of their high power conversion efficiency (PCE) alongside low material expenses. The ease of fabrication, abundant material source and high PCE rapidly caught the attention of scientists. Although the first promising solid-state perovskite cells were only reported just in 2012 mid, swift advance was made within a few years (Figure 1).17,[31]Methyl ammonium lead trihalides were first applied as sensitizers in dye-sensitised solar cells in combination with an iodide/triiodide liquid electrolyte and yielded an efficiency of 3.8%[32]. However, the perovskite material was found to be unstable in the liquid electrolyte[33]. Substituting the liquid electrolyte with a Spiro-OMeTAD solid-state HTM resulted in efficiencies of 7.6 % for a CH3NH3PbI3-xClx based sensitiser[34] and 9.7 % for a CH3NH3PbI3 based sensitiser[35] deposited on a TiO2 nanoparticle layer.In 2012, a solid-state dye sensitised solar cell with a PCE of 10.2%[36] was obtained by using N719 as a sensitizer, tin-based perovskite CsSnI3 as a co-sensitizer and hole conductor. Then, in late 2012, Miyasaka, Snaith and co-workers followed the extremely thin absorber approach (ETA) using mixed perovskite CH3NH3PbI2Cl layer on Al2O3 with spiro-OMeTAD as the p-type hole conductor and obtained a PCE of 10.9%, a major milestone in the progress towards the highly efficient perovskite photovoltaics developed in subsequent years[37].Figure 6: Electron and hole flow in a typical perovskite solar cells with respect to the energy bands of the materials (Reproduced from Ref 17).Solar cell efficiencies of over 10 %, in the absence of a conductive nanoparticle layer, indicated that the perovskite material does not function solely as a photoactive layer, but could also operate as an electron transporting material.36 This device architecture renders the nanoparticle scaffold as an optional layer, as illustrated in Figure 6. This realisation led Snaith and co-workers to fabricate a CH3NH3PbI3-xClx-based planar device using a dual-source thermal evaporation to deposit the perovskite layer with an efficiency of 15.4%.[38]Subsequently, in 2013, Liu et al reported an efficiency of 15.7% through the combination of both planar and mesoporous architectures[39] whereas Jeon et al reported efficiencies near and above 18% in 2014. In this latter case, methylammonium lead bromide (MAPbBr3) was combined with formamidinium lead iodide (FAPbI3) as the light harvesters in a bilayer (one above other) solar cell architecture[40]and subsequently increased this to 19.3% by interface engineering on a planar device[41],. By November 2014, a PSC efficiency of 20.2 % was reported,[42] with an NREL-certified nonstabilized efficiency of 20.1%.16 Later, the same group of researchers hit the top of the efficiency chart in 2016, achieving a power conversion efficiency of 22.1%. Very recently, the scientists of EPFL, led by Michael Saliba, have developed the first perovskite compound with triple cations (cesium/ methylammonium/ formamidinium) reaching an efficiency of 21.1%, and 18% in operational conditions for periods longer than 250 h.1). These new composites were shown to be more temperature stable and less affected by environmental moisture. However, these results were achieved under ambient conditions and, with the toxicity of lead, remains a bottleneck that limits their commercialization.Before discussing the operational principles of a photovoltaic device, it is necessary to clarify a few fundamental parameters that are used to describe device performance, as well as how a device is measured. A photovoltaic device is measured by placing a cell in the light and measuring the current output at a range of applied voltages. An example of a so called current-voltage plot is shown in Figure XX. The solid line is from a device measured in the light and the dotted line is from a device measured in the dark.Perhaps the most important definition is the overall efficiency of the cell.  This is given by the ratio of the maximum power output of the cell (Pmax) and the incident light power (IL)??PCE =  (Pmax/IL)× 100%     Eqn 1–Power conversion efficiency (PCE) Other than the Pmax, there are two other key parameters that are important in defining the performance of a solar cell. These points are the current when the voltage is zero, referred to as the short-circuit current (JSC), and the voltage when the current in zero, called the open-circuit voltage (VOC).  Another term often quoted is the fill factor (FF), which is defined as the ratio of the maximum power of the cell and the maximum theoretical power. This is sometimes referred to as the ‘squareness’ of the solar cell. For a perfect diode the FF would be 1 and the solar cell would have its PMax at JSC and VOC.  ??FF = (Pmax / Jsc.Voc) ??? ? ??         Eqn 2 – Fill factor (FF)In reality, solar cells never display ideal diode behaviour because power is dissipated through parasitic series and shunt resistances. An equivalent circuit for a solar cell is shown in Figure 2.3 that allows for a better understanding of the non-idealities shown in the curve in Figure 2.2. The series resistance (Rs) arises due to the resistance of the cell materials and contacts. This value should ideally be zero. The parallel shunt resistance (Rsh) should be infinite in a perfect solar cell. In a non-ideal device, it is lower, due to leakage losses within the solar cell, arising from alternative electrical pathways through the device. The shape of the current-voltage curve gives an indication of how Rs and Rsh limit performance.  Both cause the fill factor to decrease; a non-zero RS decreases the slope of the curve near VOC, as it means the current is not increasing with voltage at the ideal rate, and a lower than infinite Rsh leads to a slope in the current-voltage curve near JSC due to leakage losses.1.2.3 Challenges and LimitationsDespite the rapid advances made, much work is still required to reduce the toxicity of the materials used and increase the stability of perovskite solar cells. To reduce toxicity, the replacement of Pb2+ with Sn2+ has already been successfully demonstrated,[43]but unfortunately, chemical stability has been hard to achieve.[44] Degradation of the perovskite material with thermal stress, light, oxygen or moisture causes device instability and leads to anxieties over long-term performance.[45] Smith et al. reported a moisture stable layered hybrid perovskite in ambient humidity[46], also It has been found bromide-doping could improve the stability of the hybrid perovskites. [47] However, environmental issues due to leaching of harmful Pb are not solved.45 And one more relevant issue to be noted is hysteresis of these devices. There are plenty of reports proving significant variation of device efficiency depending upon the scan direction (from forward to reverse bias or vice-versa) and the scan rate.40 This phenomenon, termed as JV hysteresis, hinders the precise determination of a steady-state device efficiency.  Moreover, the Monash group has recently shown that planar PSC device undergoes more extreme degradation when the testing protocol follows the diurnal solar radiation cycle.[48] Considering all this, there is a critical need to investigate new non-toxic and stable materials that are capable of substituting the hybrid lead-based perovskite materials in PSCs.1.2.4  Alternate MaterialsThe oxide based Inorganic perovskites are the stable alternative to the hybrid perovskites. They, however, failed to meet the efficiency expectations.49LiNbO3, BaTiO3, Pb(Zr, Ti)O3 and BiFeO3, which are being actively investigated for photovoltaic applications, managed to reach power conversion efficiencies of up to 8% by bandgap engineering. [49],[50],[51] Still, the complex processing and limited possibility for improvement make them unlikely to match the performance of hybrid metal-halide perovskites.Whereas the ardent search for non-toxic perovskite halides by lead replacement, without compromising efficiency, has led to so many alternative cations with the most obvious being another group 14 elements, such as Sn and Ge to other divalent cations. Tin perovskite materials have achieved efficiencies up to 6.4%,[52],43 However, ready oxidation of Sn2+ and Ge2+ to Sn4+ and Ge4+ in air at room temperature results in poor intrinsic stability.44 This issue was difficult to solve by using alternative alkyl ammonium halides.  The hybrid perovskites with other homovalent elements resulted in higher band gap based on computation studies.[53] Some recent reports are there on alternative materials like (CH3NH3)2CuCl4-xBrx with better air stability. However, the efficiency achieved so far is only 0.02% [54]. Considering these limitations, identifying a stable, non-toxic cation is a key challenge in the area of perovskite solar cells.  Moving beyond divalent cations resulted in some alternatives like bismuth with its potential semiconducting character, rich structural diversity and interesting electronic and optical properties.[55],[56],[57],[58] The efficiency of these devices was poor presumably due to nonradiative recombination of defect states.[59]The triple cationic perovskites (caesium/ methylammonium/ formamidinium) reported by Saliba et al. were impressive showing temperature and moisture stability, but the need for ambient conditions and the toxicity of lead remained a bottleneck that limits their commercialization.  On the other hand, double and triple metal cation oxide perovskites are well known for their multiferroic and ferroelectric properties.[60],3 There has been an advancement of very poor performance to the recent success of 8% with some of these oxide ferroelectric perovskites by tuning the bandgap, controlling the cation proportion and distribution in the double perovskite Bi2FeCrO6.[61],[62] A constrained visible range spectral response, due to large band gaps or the complex process to tune the bandgap, remains a limiting factor for oxide perovskites. Chalcogenides are a very promising stable alternative due to their wide absorption of solar radiation. Chalcogenides, like CdSe, CdTe, PbS are limited by their toxicity but they have been used in thin film or sensitised solar cells. By comparison, other less-toxic chalcogenides or chalcogenide perovskites based on sulphur or selenium have not attracted much attention even though they were synthesised a long time.[63]The need for either high temperature or vacuum processing for synthesis remains as a hindrance in exploiting other non-toxic and abundant chalcogenides in solar cells.[64],[65] Non-toxic chalcogenides based on bismuth, silver, antimony, gallium shows impressive potential, but, have not reached high efficiencies in quantum dot and solid state devices. The reason for the lack of efficiency is understood as due to high interlayer recombination and low charge transfer rates through these materials. This could be solved by, using hetero bimetallic chalcogenides, which could simultaneously act as photo-absorbing and charge transporting medium.The I–V–VI and III-V-VI ternary chalcogenide semiconductors, in particular, are attractive due to their optoelectronic and photovoltaic potential.[66],[67] Among these, materials like AgBiS2 have shown potential in solid-state and sensitised devices.4,5,6,7The device architectures that achieved a 6.3% efficiency is shown in Figure 7.8 It has been reported that replacing MoOx layer with Spiro-OMeTAD in thin film solar cells could improve open circuit voltage in solid state devices.[68] It is well known that there exist two phases of AgBiS2, namely, the phase ?-AgBiS2 with a hexagonal structure and  ?-AgBiS2 with a cubic structure. The phase transition temperature is 468±5 K.[69],[70] AgBiS2 has been synthesised through vacuum fusing[71], flux technique[72] solvothermal[73] and ligand assisted method.8However, simple solution processable methods or size control ling methods are yet to be explored.Figure 7: Reported device architecture of AgBiS2 solid state devices(reproduced from Ref 8).Considering, non-toxic chalcogenide halides (MCX) in hybrid perovskites , is a topic studied little before, despite some properties, such as photoconductivity and ferroelectricity have been explored of BiSI or BiSeI.[74],[75],[76] Horak et al. analysed the photoelectric properties of single crystals of BiSI, witnessing n-type conductivity and anisotropic PV effects. [77]Later there were reports on synthesising polycrystalline BiSI by both spray pyrolysis and hydrothermal reaction from solution phase precursors, still, the studies were confined to physiochemical characterization in each case.[78],[79] Hahn et al. studied BiSeI and BiSI for solar cell applications in 2012 and realised they are n-type materials, with high absorption coefficients and indirect band gaps of 1.57 eV and 1.29 eV, respectively. However, an extremely poor PCE (0.012%) was obtained.62 To date there have not been any studies investigating the capability of this material in perovskite solar cells.Research ObjectivesOne of the practical ways to solve major challenges of current solar cells will be choosing earth abundant, non-toxic, non-volatile material as a replacement for the current cations in perovskites and to synthesise them by solution processable method at low temperature. A literature survey indicated that an attractive option is to consider heterovalent cationic substitution, which could lead to the formation of a double perovskite with an expected formula of A2MCX6. There are some reports on speculated crystal structures of cesium-based bimetallic halide double perovskites calculated via computation analysis (crystal structure shown in Figure 8).2 Devices based on double and triple metal cation oxide perovskites with multiferroic and ferroelectric properties,2 have shown an efficiency advancement up to 8%, but lack room for improvement due to higher bandgap and complex route to broadening their absorption spectrum. The novel idea in this project will be filling this gap by developing less explored double perovskites halides based on non-toxic chalcogenides. Very few halide double perovskites have been reported, but these are based on alkali or rare-earth metals and used for radiation detectors.[80]Figure 8: The polyhedral model of the conventional (left) and reduced (right) unit cell of the hypothetical halide double perovskites(reproduced from Ref 2).Focusing on non-toxic metal chalcogenide cations, bismuth-based binary sulphides attracted some attention considering their low bandgap, 1.3eV, and wide absorption until 950 nm, proving photovoltaic potential in thin film and sensitised solar cells.[81],[82]Antimony, silver, indium and gallium also look promising based on elementary considerations and known electrical conductivity. Antimony is less attractive owing to its toxicity. Silver sulphide showed narrow bandgap of 1.1eV with extended absorption edge up to 1100nm with chemical stability. In addition, in an octahedral environment, the ionic radii of Ag+ (1.29 Å) are similar to those of Pb2+ (1.19 Å) and Bi3+ (1.03 Å).[83] Following this simple reasoning, we will investigate new halide double perovskites with the pairs M1/M2 where M1 = Bi and M2 = Ag, Cu, etc.The idea of using multiple cations leads to ternary chalcogenides, which offer a high flexibility of bandgap tuning without the toxicity issues. I–V–VI family materials like AgBiS2 has shown promising potential in photovoltaics.8 But, synthesis methods are either complex or needs a vacuum and dark conditions. Our focus will initially be on a simplified synthesis of AgBiS2 from binary dithiolate complexes, an approach not used so far. Most of the reports provide little evidence for the exact composition and crystallinity of the material. Therefore, convenient environmental friendly methods of synthesising AgBiS2 nanoparticles by controlling the particle size, morphology, composition and crystal structure of silver bismuth sulphide will be an important focus. Low-temperature calcination and microwave irradiation are attractive due to their simple rapid convenient and environment-friendly features. Hence, we introduce two methods for the synthesis of these ternary chalcogenides; one is two-step thermolysis and the other is microwave-assisted synthesis method. Metal aryl dithioate complexes [e.g. Bi(S2C(C6H4)-4-CH3)3], will be synthesised using aryl dithioate ligands in collaboration with the Andrews group (Figure 9). [84] and converted to AgBiS2 by thermolysis and microwave synthesis methods.Figure 9: Synthesis of bismuth sulphide via thermal decomposition of a bismuth aryl dithioate complex synthesised.84The initial focus will be on synthesising silver bismuth sulphide (AgBiS2) and indium bismuth sulphide (InBiS3).After the initial screening by analysing the photovoltaic potential of these ternary compounds, the combination of lanthanum and gallium with bismuth thiolates will be studied and subsequently reacted with organic halides to form double perovskites. Since the device architectures studied todate8 have lower than expected Voc and recombination losses are likely to be significant, different device architecture will be examined with a view to optimising the performance of devices made with the newly synthesised materials.  References