|
|
How AI can Accelerate R&D for Solder Paste Formulations
Materials Tech |
|
Authored By:Mélanie MATHON, Mehdi GUELMOUSSI, Lucillianna DA COSTA, Valérie LUCAS, Christophe DEHON Inventec Performance Chemicals France SummarySolder pastes formulations constantly encounter new challenges such as the incorporation of new alloys, the refinement of powder granulometry, reduction of voids, and mitigation of residue corrosion. Each solder paste must adhere to certain minimum specifications, including solderballing, low slump, and wettability, regardless of its intended application. When developing new formulations, approximately fifteen ingredients are adjusted or created to meet these specifications. Due to the complexity of the formulations, the studies are time-consuming. However, the use of new methods based on Artificial Intelligence (AI) enables optimizing the composition of solder paste with specific properties and cutting development times by drastically reducing the number of tests to be carried out. As a field of research, the full understanding of the factors influencing formulated products remains incomplete and the scientific literature about this topic is scarce. To address this problem, AI is proving capable of going beyond human detection of correlations between the features of a phenomenon [1]. Our focus lies in leveraging AI to predict the key solder paste properties to comply with new market specifications and exploring the potential for generating new formulas. This resulted in the testing of hypotheses and the development of a solution, harnessing the efficiencies offered by AI in reducing both time and resources during the formulation development phases of solder paste. In conclusion of this article, a comparison is made between the experimental results and the AI anticipated results. Analyzing these results in alignment with the methods used in data science not only advocates for the significance of experimentation in AI but also enhances the agility of AI project, particularly in the field of solder paste applications and broader chemicals. ConclusionsIn this study, AI was used to predict the results of the key characteristics to better select the choice of formulas, to understand the relationship between molecules and to spend time on other aspects of the developments (specific requirements, processes, etc.). To be able to exploit AI data in a relevant way, we highlight that the input database itself is crucial to generating reliable output data. A great deal of work needs to be done upstream on data relevance before AI capabilities can be exploited. As part of the development of a new solder paste formulation, AI made possible to:
AI is also helping to increase our knowledge of the interactions between raw materials and their impact on the various characteristic tests. The use of artificial intelligence is a great opportunity to improve the efficiency and relevance of the projects we monitor. It is important to note that AI is a tool, not a replacement for a high-performance R&D team with strong know-how. In the near future, AI could be used to predict the behavior of soldering pastes by varying the alloys used. This perspective is also in line with the objective of saving time and improving efficiency. Initially Published in the SMTA Proceedings |
|
Comments
|
|
|
|
|