Effect of gamma-ray irradiation on ferroelectric, hafnia-based, non-volatile memory for use in extreme environments

Effect of gamma-ray irradiation on ferroelectric, hafnia-based, non-volatile memory for use in extreme environments

Domain, Specialties: Material physics, surface science

Keywords: HfO₂, ferroelectricity, radiation hardness, point defects, HAXPES, extreme environments

Location: Laboratory for the Study of NanoStructures and Surface Imaging (LENSIS) at CEA Saclay, France

Type of contract: PhD offer, expected start date October 2026

Description

The emergence of hafnia-based ferroelectric (FE) memories [Böscke2011] has opened a new paradigm for ultra-low-power edge computing. Hafnia is fully compatible with CMOS technology [Trentzsch2017], can be aggressively scaled [Martin2024], offers low access latency (~ns), and requires ~fJ/bit to switch—three orders of magnitude less than other emerging non-volatile memories. These properties make FE memories ideally suited for smart edge systems, overcoming the memory wall while enabling secure local data processing. Advanced 1T-1C FeRAM architectures, where the FeCAP is connected to the drain of a CMOS transistor, also provide promising paths toward NVDRAM [Micron2024] and NVSRAM [Okuno2025]. Moreover, hafnia is a by-product of the French nuclear industry giving France and Europe a major strategic advantage in global supply compared to China. These advantages align with strategic applications in space, defense, medical, nuclear safety, and heavy-duty transport, where electronics face harsh radiation environments. For example, space missions endure continuous radiation exposure for 5–15 years (hundreds of krad/year for ESA’s JUICE mission [Furano2012]), in addition to sporadic solar flare spikes. On-board systems must thus tolerate both steady-state and transient radiation doses with reliability far beyond that of consumer-grade devices. In this context, ultra-low-power FE logic and memory—especially in FD-SOI SoC implementations—could deliver up to 3× lower payload and 10× lower power consumption, while maintaining the non-volatile and radiation-hard performance required for space, defense, and nuclear applications.

Nevertheless, stable imprint behavior still remains a significant challenge [Mikolajick2021]. Imprint induces a shift of the Polarization-Voltage (P-V) curve along the voltage axis [Alhada2025] as shown in Fig. 2 and is attributed to charge trapping/detrapping, FE domain pinning and concentration profile of charged defects. All may be accentuated under irradiation. The seminal paper of Lucovsky et al. demonstrated that the key point defect is the V_O, which provides both shallow and deep charge traps, enhanced by g- and X-ray radiation [Lucovsky2006]. First detailed DFT calculations confirmed the nature of the traps at V_O [Xiong2005]. Incident radiation creates e-h pairs as suggested in Fig. 1, which can then be swept to either electrode depending on the internal field or applied bias.

Figure 1. (left) Schematic of the formation of shallow and deep trap states at oxygen vacancies and possible charge states (0 and 2+) (right) Co60 sources at LABRA CEA Saclay

Ding et al reported trapped charged following generation of e-h pairs under g-ray radiation [Ding2021]. Kaya et al studied the effect of g-radiation up to 5 MRad demonstrating a high threshold voltage shift in HfO₂ MOSFETs.

Figure 2. (Left) Phase field modelling of P-V imprint(blue) compared with experimental results (red) [Alhada2025] (right) Schematic of HAXPES geometry allowing operando VO quantification as a function of cycling conditions and bias.

Radiation enhanced diffusion of defects in solids is a long-standing historical problem, originating in the need to characterize nuclear materials [Dienes1958]. Charged V_O may drift under the internal polarization field in addition to the classic diffusion problem represented by Fick’s law leading to diffusion enhancement. Furthermore, V_O diffusion to electrode/ferroelectric interfaces may be self-limiting, resulting in accumulation of V_O in the bulk film. Trapped charge at the electrode/ferroelectric interface may further limit drift, resulting in the accumulation of excess charge in the bulk. Modest γ-radiation can degrade FeCAP polarization [Zhang2020], however, the reported resilience metrics do not include FeRAM at the memory array-level with large statistical data collection.

The correlation of radiation induced changes in V_O concentration with electrical performance is an open question which will be addressed in this thesis work.

The project will use advanced photoelectron spectroscopy techniques including synchrotron radiation induced Hard X-ray photoelectron spectroscopy and complementary structural analysis including high-resolution electron microscopy, X-ray diffraction and near field microscopy. Using appropriate, low energy, ion beam etching and angular dependence of the XPS, the electrode/ferroelectric interfaces will be studied following the methods we have already developed [Hamouda2020, Hamouda2022]. Hard X-ray photoemission (HAXPES) using synchrotron radiation will be used to probe the interface chemistry and band alignment in FeCAP in a non-destructive manner (Fig. 2 (right)).

Figure 3. Inelastic mean free path of electrons as a function of kinetic energy. The darker rectangle represents the domain covered by laboratory XPS systems. The larger, light blue rectangle represents the HAXPES regime up to 8 keV photon energy.

We will design samples with suitable geometry (Fig. 2) for operando measurements of the interface band line-ups as a function of the applied bias Operando experiments

There are two specific objectives to the thesis work:

1.  Gamma radiation effects on Hf_xZr_1-xO₂ (HZO) material properties. Radiation induced orthorhombic to tetragonal/monoclinic/antipolar phase transitions. Generation and transport of mobile charge electron-hole pairs, charge trapping/de-trapping, enhanced V_O mobility. Radiation induced point defects including V_O and V_Hf. Radiation enhanced diffusion.

2.  Quantification of FE device properties in HZO-based FeCAPs as a function of radiation dose. Outcome of the characterization will be exploited to develop and calibrate FeCAP compact models to be used at the circuit level.

The results will be compared to circuit level characterization of irradiation effect on wafer level and packaged 1T-1C 16kb and 512kb HZO-based FeRAMs, respectively. Circuit/system level simulations of irradiation. Quantification of radiation sensitivity of peripheral circuits and underlying CMOS

FeCAP fields and FeRAM circuit integrated onto 130nm CMOS test vehicles will be fabricated at CEA/Leti, Grenoble. The first task will focus on understanding the radiation induced changes in ferroelectric properties of FeCAPs. The second task will focus on the characterization of FeRAM circuit integrated on 200mm test vehicles with high statistics (at least 16kb).

The irradiations will be done at LABRA (CEA DRF) using the PAGURE irradiator allowing controlled irradiations up to 472 Gy/hr. Three campaigns are foreseen in the project, allowing to study in parallel the modifications of material properties, the FeCAP behaviour and the FeRAM and peripheral circuit response to gamma irradiation. The irradiation campaign spacing will also allow material or processing optimization between runs.

The physical and chemical analysis will be done in laboratory and synchrotron environments. X-ray and ultra violet photoelectron spectroscopy will be carried out at CEA Saclay. Hard X-ray photoemission using synchrotron radiation (Soleil, Petra-III, NSLS-2, Spring-8) will allow probing the V_O concentration profile in FeCAPs [Iung2025, Yadav2025] after irradiation. XRD will probe the phase composition as a function of dose   and cycling.

At each level, the experimental material and electrical characterization will be accompanied by theoretical calculations to simulate the material response to irradiation and compact modelling of the resulting device performance (collaboration with the Institut des Nanotechnologies de Lyon).

The PhD project will be carried out in the framework of a close collaboration between the CEA/Leti in Grenoble providing the samples, integrated devices and wafer scale characterization and the CEA/Iramis in Saclay for the fundamental analysis of the material properties, irradiation experiments and device scale characterizations.

References

[Böscke2011] T. S. Böscke et al., Ferroelectricity in hafnium oxide thin films Appl. Phys. Lett. 99 102903, (2011) doi: 10.1063/1.3634052

[Mikolajick2021] T. Mikolajick et al., Next generation ferroelectric materials for semiconductor process integration and their applications J. Appl Phys. 129, 100901, (2021) doi: 10.1063/5.0037617

[Alhada-Lahbabi2025] K. Alhada-Lahbabi et al., Investigating Experimental Short-Term Imprint Dynamics in Ferroelectric Hafnium Oxide Through Phase-Field Modeling Adv. Funct. Mater. (2025), e14094

[Dienes1958] G.J. Dienes and A.C. Damask Radiation Enhanced Diffusion in Solids, J. Appl. Phys. 29, 1713 (1958)

[Ding2021] M. Ding and X. Liu Damage effect of hafnium oxide gate dielectric-based metal–oxide–semiconductor structure under gamma-ray irradiation AIP Advances 11, 065304 (2021); doi: 10.1063/5.0048080

[Furano2012] G. Furano et al., Review of radiation hard electronics activities at European Space Agency Workshop on Electronics for Particle Physics 2013 IOP

[Hamouda2020] W. Hamouda et al., J. Appl. Phys. 127, 064105 (2020) https://doi.org/10.1063/1.5128502

[Hamouda2022] W. Hamouda et al., Appl. Phys. Lett. 120, 202902 (2022) https://doi.org/10.1063/5.0093125

[Iung2025] T. Iung et al., Oxygen vacancy distribution and phase composition in scaled, Hf0.5Zr0.5O2-based ferroelectric capacitors Appl. Phys. Lett. 126, 062903 (2025) doi: 10.1063/5.0245595

[Kaya2018] S. Kaya et al., Co-60 gamma irradiation effects on electrical characteristics of HfO2 MOSFETs and specification of basic radiation- induced degradation mechanism Radiation Physics and Chemistry 149 (2018) 7

[Lucovsky2006] G. Lucovsky et al., Differences Between Charge Trapping States in Irradiated Nano-Crystalline HfO2 and Non-Crystalline Hf Silicates IEEE Trans Nucl. Sci. 53, 3644 (2006)

[Martin2024] S. Martin et al., Hf0.5Zr0.5O2 FeRAM scalability demonstration at 22nm FD-SOI node for embedded applications, 2024 IEEE International Electron Devices Meeting (IEDM), doi: 10.1109/IEDM50854.2024.10873378

[Trentzsch2017] M. Trentzsch et al. A 28 nm HKMG super low power embedded NVM technology based on ferroelectric FETs,” in IEDM Tech. Dig., Dec. 2016, pp. 11.5.1–11.5.4, doi: 10.1109/IEDM.2016.7838397

[Yadav2025] G. Yadav et al., Polarization-dependent, oxygen vacancy distribution in ferroelectric Hf0.5Zr0.5O2 capacitors IEEE 2025

[Zhang2020] W. Zhang et al., IEEE Access, 8,108121(2020), doi: 10.1109/ACCESS.2020.3000865

Funding acquired, selection based on CV and eventual interview

Les équipes de l’IRAMIS mènent des recherches en physique et chimie, au carrefour des mondes académiques et des missions du CEA, ainsi que des enjeux sociétaux et de l’innovation. Ses recherches portent sur les nouvelles technologies pour l’énergie, les technologies quantiques, les nouveaux matériaux, les systèmes complexes et l’interaction rayonnement-matière.

La physique de la matière condensée est étudiée au SPEC depuis ses aspects les plus fondamentaux jusqu’aux applications dans les cas qui s’y prêtent. Les approches sont extrêmement variées et permettent l’exploration de mondes qui vont de l’échelle nano aux objets macroscopiques. Fort de ses publications dans des revues scientifiques spécialisées, le SPEC est aussi un lieu de transfert technologique et de brevets. Le SPEC est impliqué dans de nombreuses collaborations, aussi bien avec le CEA qu’au niveau national et international.

La recherche au LENSIS se concentre sur l’étude de la structure électronique et chimique des surfaces, interfaces et films d’oxydes fonctionnels. Pour ce faire, nous utilisons un large éventail de techniques d’analyse de surface basées sur la photoémission, telles que XPS, HAXPES, ARPES et PEEM, ainsi que des sondes électroniques telles que LEEM. Les matériaux et les dispositifs que nous étudions sont réalisés en collaboration étroite avec nos partenaires de recherche technologique, notamment, le CEA LATI (Grenoble) et NaMLab (Dresde), ainsi que des partenaires indutriels tels que ST Microelectronics et la Ferroelectric Memory Company (FMC) dan le cadre deu projet européen Ferro4EdgeAI (www.ferro4edgeai.eu) coordonné par le laboratoire.

Le candidat devrait avoir de très bonnes bases en physique de l'état condensé. Des bonnes notes et classement en Master sont nécessaires.Previous experience/skills on electron spectroscopy and/or image processing is not mandatory but useful.

Solid grounding in condensed matter physics. Good academic marks and high ranking in Masters courses required