Predictive Reliability for High Power RF MEMS

Informations projet

PRIME

N° de convention de subvention: 101032925

DOI

10.3030/101032925

Projet clôturé

Date de signature de la CE 9 Mars 2021

Date de début 1 Juillet 2021

Date de fin 30 Juin 2023

Financé au titre de

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Coût total

€ 153 085,44

Contribution de l’UE

€ 153 085,44

Coordonné par

IDRYMA TECHNOLOGIAS KAI EREVNAS
Greece

Periodic Reporting for period 1 - PRIME (Predictive Reliability for High Power RF MEMS)

Période du rapport: 2021-07-01 au 2023-06-30

Résumé du contexte et des objectifs généraux du projet

High-power radio-frequency communication systems (e.g. Radars and Satellite Communications) are supporting our everyday activities by enabling safe transportation of people and goods around the planet as well as long-distance communications. These systems are relying on electronic components capable to deal with the corresponding signals including signal routing switches. Micro-Electro-Mechanical-Systems for Radio Frequency applications (RF MEMS) are considered as devices with great potential to take over critical functionalities, such as signal routing. Thus, their implementation in next generation communication systems will further support the realization of novel and/or additional applications in the corresponding domains with valuable benefits on our lifestyle and quality and security of our lives. However, for RF MEMS, despite the widely accepted exceptional features they own, their reliability remains an open issue and a general concern. This is due to their strongly interdisciplinary nature involving the mechanical, electrical, micro/nano fabrication and microwave domains that renders reliability studies a very challenging task to accomplish.
The PRIME vision was to introduce a fresh perspective on the assessment of (high-power) RF MEMS reliability and physics. The way to achieve this was by combining conventional reliability testing with machine learning techniques towards enabling failure related predictive diagnostics. Nevertheless, the project was a holistic approach and apart from the machine learning reinforced studies, also included RF-design tasks, micro/nano-fabrication activities, and implementation of diverse characterization techniques having always in mind to support the development of the fellow’s profile, to maximize the transfer of knowledge between the fellow and the host and to disseminate the results to the broader possible audience.
Overall PRIME successfully achieved to develop a family of RF MEMS switches capable to support diverse requirements in terms of RF response and power handling. Further to these, aspects related to high-power reliability physics have been identified and studied. Finally, machine learning based methodologies supported by either simulated or experimentally obtained data have been introduced. These were achieved via a workplan that strengthened the fellow skills and presence in this field.

Travail effectué depuis le début du projet jusqu’à la fin de la période considérée dans le rapport et principaux résultats atteints jusqu’à présent

PRIME workplan included the design of the devices to be studied, their fabrication, extensive characterization, monitoring of the RF response and power handling capabilities, study of aspects related to reliability physics and development of machine learning based methodologies enabling predictions concerning their performance and reliability. All these tasks were performed in the facilities of the host (IESL-FORTH) but also through two secondment placements related to RF design (RF-Microtech, Italy) and to machine learning tasks (Imperial College London, UK).
Following this workplan (and despite the Covid-pandemic) PRIME achieved to deliver a family of RF MEMS switches supporting different RF responses and power handling capabilities. In addition, reliability physics issues anticipated to appear during high power operation of RF MEMS were identified and studied. Finally, machine learning based methodologies utilising datasets generated by either simulation studies or by experimental results have been introduced.
PRIME research outcomes and ideas have been disseminated or are about/planned to be disseminated. More specifically so far:
• A high-power related reliability issue identified, studied, and presented as oral presentation at ESREF 2022.
• An extended version of this work has been published in Microelectronics Reliability (Microelectr. Reliab. 138, 114678, 2022).
• A machine learning based RF MEMS design methodology presented at EuMC/EuMW 2022. This work also published in the corresponding conference proceedings (2022 52nd European Microwave Conference (EuMC), Milan, Italy, 2022, pp. 369-372).
• The ideas of PRIME were also disseminated through two seminar talks delivered by the fellow, at the University of Edinburgh, UK and at Imperial College London, UK.
• The ideas of PRIME were also disseminated through FORTH RETREAT 2022, an internal event of the host institution.
• An idea about an RF MEMS novel configuration generated and a patent application submitted (under evaluation) to the Hellenic Industrial Property Organization.
• The fellow participated in the Researchers’ Night 2022 and communicated the ideas of PRIME to the broader audience.
• The fellow participated in the Researchers at School 2023 event.
• PRIME ideas and news were also disseminated through the web and through the social media (LinkedIn).

Progrès au-delà de l’état des connaissances et impact potentiel prévu (y compris l’impact socio-économique et les conséquences sociétales plus larges du projet jusqu’à présent)

PRIME vision was to introduce a fresh perspective on the study of RF MEMS by combining conventional methods with the machine learning techniques. This approach allows to deal with interdisciplinary problems such as those that typically exist in RF MEMS and more importantly during high-power operation. As results PRIME delivered new machine learning enabled methodologies for studying the design and reliability aspects of RF MEMS as well as new insights from the point of view of physics during high-power operating conditions.
Further to these, PRIME approach to combine conventional studies with machine learning in conjunction to the interdisciplinary nature of the research tasks included is expected to induce impact to diverse scientific domains. These are the i) ICT community through the machine learning tasks, ii) material community and iii) manufacturing community, related to the processes and the materials utilized to fabricate high-power RF MEMS, iv) engineering community, through the reliability outcomes and v) physics through all the gained knowledge on the phenomena governing the high-power operation.
Moreover, PRIME is an individual MSCA, thus was also focusing on the fellow. Therefore, the workplan was designed in such a way as to offer the complementary skills that were missing from the fellow’s profile. Beyond the purely research skills the fellow was also heavily involved in managerial tasks and through the secondment placements as well as through the participation on scientific events, a broad international network of collaborators was established. These will act as an important background for the fellow’s future career steps. Actions also considered (e.g. a student co-supervision) towards improving the fellow’s academic profile.
Going beyond the scientific perspective, it is worth to mention that PRIME objectives are timely dealing with the critical field of the high-power communication systems. Such systems are essential for securing effective, “greener” and safe transportation of people and goods around the world, affecting directly or indirectly several domains ranging from ecology (e.g. towards more efficient thus “greener” transportation) to tourism. Therefore, the outcomes of the project are anticipated to induce a broad and important impact to the society and economy as they directly or indirectly affect numerous of our everyday needs.
Finally, PRIME dissemination plan also included several public engagement events, thus it is expected to impact the wider public by the communicated concepts of micro/nano technologies, MEMS and machine learning to people beyond the scientific community or to pupils aiming to spark their interest about science and engineering.

prime.jpg