Book Description

Solar photovoltaic modules provide clean electricity from sunlight but will not be able tocompete on an open market until the cost of the electricity they produce is comparable to thatproduced by traditional methods. At present, modules based on crystalline silicon wafer solarcells account for nearly 90% of photovoltaic production capacity. However, it is anticipatedthat the ultimate cost reduction achievable for crystalline silicon solar cell production will besomewhat limited and that thin film solar cells may offer a cheaper alternative in the longterm. The highest energy conversion efficiencies reported for thin film solar cells have beenfor devices based around chalcopyrite Cu(In, Ga)(Se, S)2 photovoltaic absorbers. The most efficient Cu(In, Ga)(Se, S)2 solar cells contain absorber layers deposited by vacuumco-evaporation of the elements. However, the cost of ownership of large area vacuumevaporation technology is high and may be a limiting factor in the cost reductions achievablefor Cu(In, Ga)(Se, S)2 based solar cells. Therefore, many alternative deposition methods areunder investigation. Despite almost thirty companies being in the process of commercialisingthese technologies there is no consensus as to which deposition method will lead to the mostcost effective product. Non-vacuum deposition techniques involving powders and chemical solutions potentiallyoffer significant reductions in the cost of Cu(In, Ga)(Se, S)2 absorber layer deposition ascompared to their vacuum counterparts. A wide range of such approaches has beeninvestigated for thirty years and the gap between the world record Cu(In, Ga)(Se, S)2 solarcell and the best devices containing non-vacuum deposited absorber layers has closedsignificantly in recent years. Nevertheless, no one technique has demonstrated its superiorityand the best results are still achieved with some of the most complex approaches. The work presented here involved the development and investigation of a new process forperforming one of the stages of non-vacuum deposition of Cu(In, Ga)(Se, S)2 absorber layers. The new process incorporates copper into an initial Group III-VI precursor layer, e.g. indiumgallium selenide, through an ion exchange reaction performed in solution. The ion exchangereaction requires only very simple, low-cost equipment and proceeds at temperatures over1000?C lower than required for the evaporation of Cu under vacuum. In the new process, indium (gallium) selenide initial precursor layers are immersed insolutions containing Cu ions. During immersion an exchange reaction occurs and Cu ionsfrom the solution exchange places with Group III ions in the layer. This leads to theformation of an intimately bonded, laterally homogeneous copper selenide? indium (gallium)selenide modified precursor layer with the same morphology as the initial precursor. These modified precursor layers were converted to single phase chalcopyrite CuInSe2 andCu(In, Ga)Se2 by annealing with Se in a tube furnace system. Investigation of the annealingtreatment revealed that a series of phase transformations, beginning at low temperature, leadto chalcopyrite formation. Control of the timing of the Se supply was demonstrated toprevent reactions that were deemed detrimental to the morphology of the resultingchalcopyrite layers. When vacuum evaporated indium (gallium) selenide layers were used asinitial precursors, solar cells produced from the absorber layers exhibited energy conversionefficiencies of up to 4%. While these results are considered promising, the devices werecharacterised by very low open circuit voltages and parallel resistances. Rapid thermal processing was applied to the modified precursor layers in an attempt tofurther improve their conversion into chalcopyrite material. Despite only a small number ofsolar cells being fabricated using rapid thermal processing, improvements in open circuitvoltage of close to 150mV were achieved. However, due to increases in series resistance andreductions in current collection only small increases in solar cell efficiency were recorded. Rapid thermal processing was also used to demonstrate synthesis of single phase CuInS2from modified precursor layers based on non-vacuum deposited indium sulphide. Non-vacuum deposition methods provide many opportunities for the incorporation ofundesirable impurities into the deposited layers. Analysis of the precursor layers developedduring this work revealed that alkali atoms from the complexant used in the ion exchangebaths are incorporated into the precursor layers alongside the Cu. Alkali atoms exhibitpronounced electronic and structural effects on Cu(In, Ga)Se2 layers and are beneficial in lowconcentrations. However, excess alkali atoms are detrimental to Cu(In, Ga)Se2 solar cellperformance and the problems encountered with cells produced here are consistent with theeffects reported in the literature for excess alkali incorporation. It is therefore expected thatfurther improvements in solar cell efficiency might be achieved following reformulation ofthe ion exchange bath chemistry.