Solidification of Unseeded high-PERformance MULTIcrystalline silicon

The solar cell industry has for two decades been dominated by two silicon-based technologies; (i) multicrystalline solar cells with low production cost, and (ii) monocrystalline solar cells with higher solar efficiency, but also higher production cost. Reduced production costs for monocrystalline processes in recent years have led to multicrystalline solar cells having to increase their solar efficiency in order to provide the same performance / cost ratio and maintain their competitiveness. Multicrystallinesilicon (mc-Si) solar cells consist of several crystals, and the areas between the single crystals, called grain boundaries, are the source of the loss of generated current. The properties of the grain boundaries are a function of the orientation both to the grains on each side, and orientation to the grain boundary itself. Some types of grain boundaries generate harmful structural crystal defects that propagate inward into the crystal, while other types of grain boundaries can absorb and eliminate the same crystal defects. Through these indirect mechanisms, it is the grain boundaries of mc-Si solar cells that largely determine the concentration of structural crystal defects and the loss of generated current compared to monocrystalline solar cells. The main objective of the project at start-up was to develop new solidification processes with a lower concentration of structural crystal defects by changing the ratio between harmful and favorable grain boundary types. The grain boundary types and the distribution between them are largely defined during the first part of the solidification process, where the first thin layer of solid silicon formed along the bottom of the quartz crucible serves as a template for the further crystal growth. A key tool for this part of the project was a pilot-scale furnace for the production of mc-Si ingots that allows "destructive" testing with different bottom substrates and with temperature conditions beyond what can be run in a production scale furnace. Necessary physical modifications of the pilot furnace, development of a new reference process corresponding to industry standards, as well as tests of two different bottom substrates were carried out in 2018. The project has used Laue-scan to identify grain boundary types and other crystal defects in mc-Si wafers. Unlike traditional methods, this method can map grain boundary types and defects over larger areas that provide a more robust data set. In addition to mapping grain boundary types and defects in test wafers, the same has been done over height and width in two commercial reference ingots. During the first part of the project period, the market share of mc-Si solar cells fell dramatically. Market overcapacity clearly showed that state-of-the-art mc-Si solar cells could not compete with monocrystalline solar cells. Initial results from the project and external publications showed that the potential for improvements in the solidification process (0.2-0.3% abs) would not be sufficient to ensure the competitiveness of mc-Si solar cells. Both internal data and recent studies from other groups indicated a greater potential for improvement in reducing the electrical activity of existing structural defects. The focus of the project was therefore in 2019 changed from reducing the amount of crystal defects formed during the solidification process, to understanding how the electrical activity of existing crystal defects is affected by contaminants and furnace atmosphere during the solidification process, and how the activity can be changed with heat treatment at wafer level. For this work, 8 full-scale ingots with different levels of iron and aluminum were produced; 4 mc-Si ingots, and 4 monocrystalline ingots as references without grain boundaries and widespread crystal defects. Based on this test kit, the project has developed a process for cleaning of silicon wafers using heat treatment without the use of added chemicals. The method is compatible with processing large batches with low cost. The cleaning method improves all areas of mc-Si wafers, but is not able to eliminate the worst performing areas. The process has the highest industrial relevance for emerging low-temperature cell processes, which in contrast to high-temperature processes do not include process steps with an inherent cleaning effect. Three master's theses at NTNU have been completed as part of the project.

Status ved avslutning av prosjektet er at den multikrystallinske delen av solbransjen er marginalisert og ser ut til å forsvinne helt i løpet av 2021. REC Solar produsere fremdeles små volumer med multikrystallinsk silisium til spesialmarkeder, men har i 2020 dreid sin virksomhet til å levere råstoff med markedets laveste klimaavtrykk til produksjon av monokrystallinske solceller. Varmebehandlingsprosessen utviklet i prosjektet har høy relevans også for selskapets monokrystallinske virksomhet, og kan legge til rette for økt bruk av vårt råstoff i en stadig voksende del av markedet som betaler høyere pris for produkter med lavt klimaavtrykk. Denne aktiviteten blir videreført i andre interne og eksterne prosjekter.

In the project, Solidification of Unseeded high-PERformance MULTIcrystalline Silicon (SUPERMULTI), we aim to secure the future competitiveness of REC Solar's tall multicrystalline ingots by influencing the recombination activity of structural defects through controlling how they interact with impurity elements and through in-situ hydrogen passivation. REC Solar captures a large synergy effect by using their own feedstock source in their own crystallization process for multicrystalline silicon, which enables the growth of very tall ingots, thus significantly increasing the productivity of the process. The product satisfies current material specifications, but due to increasing quality requirements for tomorrow's higher efficiency cell processes, this is not expected to last. Due to an increase in the density of crystal defects towards the top of the ingot, tall ingots may be particularly challenging for high efficiency cells. However, what renders the structural defects recombination active are the impurity elements decorating them. In the SUPERMULTI project we will therfore work to understand and change the recombination activity of structural defects, like grain boundaries and dislocation clusters, in order to make them less harmful. This can be achieved by influencing the distribution of impurity elements (dissolved/precipitated states), and by increasing the hydrogen passivation of the defects. It is important to understand how the impurity elements are distributed on different types of grain boundaries. Mapping of grain boundary types with Laue diffractometer will be conducted at Sintef and followed by lifetime characterisation of the samples, combined with modeling of recombination activity. Dedicated ingots will be grown with added impurity elements by REC Solar, as well as ingots with in-situ hydrogen during growth. Both p-type and n-type ingots will be made. The material will also be processed to solar cells (PERC type) to screen effects down to cell level. Systematic investigations of heat treatment wil be conducted on selected samples to influence the distrubution of the impurity elements and hydrogen's ability to passivate defects. In the last phase we will look into cross effects between impurity additions and hydrogen in-situ during growth and how it influences the recombination activity of the specific impurity elements. In addition to Elkem Solar and SINTEF staff, bachelor and/or master students at NTNU will be involved in characterization as well as R&D activities in Elkem Solar's labs and production facilities.