Crystallisation of seeded Silicon, impact of Light Impurities and Defects – CrySaLID

Task 1 is devoted to the grain structure formation and control in solidification processes from seed using: a unique solidification experiments with in situ X-ray imaging (IM2NP-MCA), experiments on bi-crystals growth (NTNU), experiments to produce mono-like ingots (SIMAP). All these contributions will generate results needed to understand the grain structure formation and evolution when starting growth from seeds.
Task 2 addresses the key point of light impurities in particular the carbon and oxygen impurities and of structural defects known to deteriorate, alone or in combination with impurities, the PV efficiency. Simulations of the thermodynamic and chemical environment in the EMIX furnace in order to develop concrete solutions for the reduction of contamination (SIMAP, EMIX), mono-like crystallisation using silicon feedstock with impurities in the X-ray imaging device (IM2NP-MCA) and in the medium scale furnace (SIMAP) are performed. Additionally, characterisation of structural defects and of the implied mechanisms is also conducted by X-ray diffraction imaging on mono-like samples (IM2NP-MCA), study of dislocations in between seeds in bi-crystals (NTNU), and of dislocations linked to the grain structure (SINTEF). All these results are put in perspective with the characterisation of the PV properties (IM2NP-OptoPV, KAU).
Task 3 is devoted to the development and validation of multi-scale modelling and simulation of grain formation and evolution: phase field model of grain boundary grooves in silicon (IM2NP-TMS), ingot scale grain structure simulation in 3D (CEMEF).
Task 4 is devoted to the synthesis of results, industrial scale assessment and feedback to the EMIX industrial process. Following EMIX closure, the project was reorganised without modifying the structure and the main project objectives.

The project is in essence international with four French partners, two Norwegian partners and one Swedish partner.
A unique result obtained in the project is the observation by IM2NP of the development of dislocations and their interaction with grain and twin boundaries during growth. Indeed, dislocations are major defects for PV silicon and the knowledge of their formation and development mechanisms is strongly requested to improve processes. These in situ experiments and their analysis allowed following dislocation propagation from the seed into the crystal along the growth direction. This work created the conditions for an extended collaboration with SINTEF and NTNU within the INSIDES (In situ characterisation and simulation of defect evolution in silicon) project (national Norwegian funding) submitted by SINTEF in Norway and awarded in 2015.
Moreover, first simulations of facetted silicon growth where obtained using phase field method at IM2NP. At another scale and using the silicon parameters, the CEMEF team could model the formation of twin grains at the level of {111} facets. These simulations at several scales are absolutely needed to yield predictive and quantitative simulation tools to be used at the laboratory and industrial scale.
In clonclusion, oOriginal results were obtained by the CrySaLID team. One of the highlights is the characterization by IM2NP of growth by in situ X-ray imaging. CEMEF simulated in 3D Si growth and twin formation with a model intended to be installed in commercial software. The 3D global modelling and growth of Kyropoulos Si ingots was obtained (SIMAP). Two collaboration projects emerged during the project: one within the consortium (INSIDES) and another one between IM2NP and NTU, Taiwan. The originality of the results leads to the continuation of these researches beyond the end of the project.

The project will contribute to the increase of the knowledge concerning the main features of silicon crystal growth: defects, grain competition, twins. The mastering of this knowledge interests fundamental research but is also essential for the development of predictive simulation codes requested by research and industry; and to improve industrial processes that all without exception presently face limitations due to the control of structural defects like dislocations and twinning.

The project has been at the origin of 12 publications in international peer-reviewed journals and 28 presentations in national and international conferences since its beginning in January 2015. Additional publications are in preparation between IM2NP and CEMEF and between IM2NP, SINTEF and NTNU. Moreover, Nathalie Mangelinck-Noël from IM2NP participated as a trainer to the French National Training Action organised by the Cristech network with a lecture on : “Methods and in situ techniques during crystallization” in October 2017.
Additionally, efforts have been done to present this work to the public and to school children and students during the whole project with rewarding and interesting feedback for both researcher and non-specialist public. An article was recently published in the French magazine «La Recherche«: N. Mangelinck-Noël, Etienne Pihan, « Le renouveau du silicium dope les panneaux solaires », Article La Recherche N° 539, Septembre 2018.

The development of the photovoltaic (PV) sector requires significant progress in performance and reductions in cost. For crystalline Si solar cells in particular, it is well-known that the grain structure, impurities and defects left after the crystallisation step, have a major impact on the final PV properties. However, the involved fundamental mechanisms are still not fully understood which hinder an efficient and reproducible control of the crystallisation processes.
In the last few years, research efforts concerning mono-like silicon crystal growth which is an interesting compromise between single crystalline and multicrystalline silicon both used for PV applications were resumed. Indeed, single crystals yield high PV efficiency but their fabrication requires costly crystal growth processes. Oppositely, multicrystalline silicon results in lower PV efficiency but can be fabricated using cheaper solidification processes close to metallurgical casting processes. Knowing this, the objective is to produce mono-like crystals i.e. with as few grains as possible using casting or directional solidification processes being initiated on seeds or by selecting the crystalline orientation by dendrite growth. For successful mono-like growth, the nucleation of parasitic grains and twins during crystallisation must be minimized and the subsequent growth controlled to favour the desired crystalline orientation. In addition, impurities and defects especially dislocations have major detrimental effects and are closely linked to the grain growth from seeds.
The scientific objectives of the CrySaLID project are to deepen the understanding of the mechanisms of crystallisation of silicon grown from seed as parasitic grains nucleation, grain competition and twins with pure silicon or silicon containing light impurities. Moreover, the characterisation of defects and light impurities linked to the crystallisation and grain structure will be addressed to understand the generation of structural defects and the impact of impurities. The correlation with PV property measurement will also be conducted. The first technological objective is to define processing conditions for improved grain structure, impurity segregation and defect control for a growth from a selected crystalline orientation. The second aims at developing a 3D and predictive Si growth simulation tool at industrial scale. Concerning the company EMIX which is part of the consortium, one major objective within the project is to fabricate silicon with lower C levels suitable for subsequent mono-like crystal growth. The objectives of the CrySaLID project will be achieved by: i) complementary investigations of Si crystallisation mechanisms (grain growth and competition) by in situ X-ray imaging of its growth, Kyropoulos and mono-like crystal growth, bi-crystal growth investigation, ii) structural defects, impurity and PV characterisation, iii) multi-scale modelling (thermodynamic environment, impurity segregation, phase field modelling of grain boundary grooves and 3D grain structure industrial scale modelling), iv) Feedback to the industrial process.