En poursuivant votre navigation sur ce site, vous acceptez le dépôt de cookies dans votre navigateur. (En savoir plus)

Learning-Based Characterization Models for Quality Assurance of Emerging Memory Technologies (H/F)

This offer is available in the following languages:
Français - Anglais

Date Limite Candidature : lundi 31 mai 2021

Assurez-vous que votre profil candidat soit correctement renseigné avant de postuler. Les informations de votre profil complètent celles associées à chaque candidature. Afin d’augmenter votre visibilité sur notre Portail Emploi et ainsi permettre aux recruteurs de consulter votre profil candidat, vous avez la possibilité de déposer votre CV dans notre CVThèque en un clic !

Faites connaître cette offre !

General information

Reference : UMR5506-PATGIR-002
Date of publication : Monday, April 19, 2021
Scientific Responsible name : Patrick GIRARD
Type of Contract : PhD Student contract / Thesis offer
Contract Period : 36 months
Start date of the thesis : 1 September 2021
Proportion of work : Full time
Remuneration : 2 135,00 € gross monthly

Description of the thesis topic

Context and motivation:
Semiconductor memories are inherent parts of modern System-on-Chip (SoC) designs. Due to the increase in computing capacity, a proportional increase in the amount of data to be processed is observed in SoCs, thus creating larger memory capacity requirements. The latest technologies make these memories denser and thus defects due to the manufacturing process are more prone to occur not only in the memory array but also in the periphery of the memory.
Magnetic Random Access Memory (MRAM) is one of the most mature emerging technology considered in several industrial demonstrators by major semiconductor foundries. In this technology, data is stored in terms of resistive states by using the spin of electrons for storage instead of their charge. MRAM has several characteristics that make it useful for many applications (automotive, health, security, etc.), such as non-volatility, fast access time, zero leakage power of the memory array, endurance, reliability and CMOS-process compatibility.
Memory test is currently based on the use of March algorithms targeting Functional Fault Models (FFMs). However, with shrinking technologies, these solutions will shortly become insufficient to achieve high coverage of new defect types that may occur during the manufacturing process. One solution to this problem is to adapt Cell-Aware (CA) test concepts successfully developed and fully deployed for logic circuits to non-volatile MRAM memories. CA test assumes that many escapes during testing are due to defects within standard cells. CA test uses a cell-internal-fault dictionary (called CA model) describing the detection conditions of each potential defect affecting a standard cell.
By adapting innovations from the digital world to emerging memory testing, and by replacing functional testing of MRAM memories by structural testing, this project will satisfy two objectives: anticipate the widening gap between correct functional modeling and realistic behavior of defects in memories, and consider new failure mechanisms in MRAM emerging technologies that cannot always be modeled by FFMs due to the stochastic nature of the switching and the inherent thermal instability of MRAMs.

The goal of the PhD thesis is hence to develop test characterization models (i.e., CA models) for the gate-cells (elementary cells, simple and complex logic gates) extracted from the description of an MRAM memory. These models will be enriched with layout information to allow complete coverage of realistic defects. Considering the significant number and the diversity of gate-cells for all considered memory technologies, Machine Learning techniques will be used for the automated generation process of these models in a time-efficient manner.

Keywords: Magnetic Memories, Defect, Characterization models, Test, Diagnosis, Quality, Machine Learning

Required skills:
The applicant must have a Master and/or an Engineering degree, with skills in the fields of artificial intelligence, digital circuit design, and test of circuits and systems. Good knowledge of programming languages (Python, Matlab, etc.) and CAD tools (Synopsys, Cadence) is also mandatory.

Work Context

Funding and partners:
The PhD thesis will be funded through a doctoral contract allocated by the CNRS in the
framework of the Call « 80PRIME » 2021. The work will be done in collaboration
between LIRMM (UMR 5506 Université de Montpellier / CNRS), SPINTEC (UMR 8191
CEA / CNRS / Université Grenoble-Alpes / Grenoble INP) and STMicroelectronics
LIRMM, 161 rue Ada, 34095 Montpellier, France http://www.lirmm.fr/
SPINTEC, 17 rue des Martyrs, 38054 Grenoble, France http://www.spintec.fr/
STMicroelectronics, 850 Rue Jean Monnet, 38920 Crolles, France https://www.st.com/

Starting date / duration / location:
September 2021 / 3 years / Montpellier & Grenoble

We talk about it on Twitter!