Guidelines for the optimization of hafnia-based ferroelectrics through superlattice engineering
Johanna van Gent, Binayak Mukherjee, Ewout van der Veer, Ellen M. Kiens, Gertjan Koster, Bart J. Kooi, Jorge Íñiguez-González, Beatriz Noheda · 2025
Read Guidelines for the optimization of hafnia-based ferroelectrics through superlattice engineering by Johanna van Gent, Binayak Mukherjee, Ewout van der Veer, Ellen M. Kiens, Gertjan Koster, Bart J. Kooi, Jorge Íñiguez-González, Beatriz Noheda (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-05174.txt.
Introduction
Layer stacks that boost hafnia memory films are tuned by Johanna van Gent and co-authors in a 2025 arXiv paper. Hafnia films can hold a charge state, a trait that suits data storage. Their stacks swap mixed hafnia-zirconia layers with pure zirconia layers. The pure layers act as a booster for the stack's held charge.
Stacks with near nine tenths zirconia hit record held charge and stood a billion flips. Swapping rare hafnia for common zirconia also cuts cost and strain on supply. Added layer faces help block breakdown. Guides for stack build close the work.
Catalog record
| CREATOR | Johanna van Gent, Binayak Mukherjee, Ewout van der Veer, Ellen M. Kiens, Gertjan Koster, Bart J. Kooi, Jorge Íñiguez-González, Beatriz Noheda |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 33 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-05174
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Guidelines for the optimization of hafnia-based ferroelectrics through superlattice engineering Johanna van Gent, Binayak Mukherjee, Ewout van der Veer, Ellen M. Kiens, Gertjan Koster, Bart J. Kooi, Jorge Íñiguez-González, Beatriz Noheda Abstract 1 Abstract Hafnia-based ferroelectrics are revolutionizing the data storage industry and the field of ferroelectrics, with improved materials and devices being reported monthly. However, full understanding and control has not been reached yet and the ideal material still needs to be found. Here we report ferroelectric hafnia-zirconia superlattices made out of zirconium-substituted hafnia (Hf 1-x Zr x O 2 ) sublayers of varying stoichiometries alternating with pure ZrO 2 sublayers. It is observed that the ZrO 2 layers in these superlattices act as a booster for the total remnant polarization (P r ). By combining the benefits of the ZrO 2 layers and the added interfaces, which help prevent breakdown, we fabricate superlattices with a total 87.5% ZrO 2 content, exhibiting record polarizations with a 2P r value of 84 µC/cm 2 that can be cycled 10 9 times, while maintaining a 2P r > 20 \mu C/cm 2 2. Next to these attractive properties, substitution of HfO 2 by the much more abundant ZrO 2 offers a significant step towards the sustainable application of these devices. 2 Introduction Ever since the concepts of ferroelectric capacitors as non-volatile memory were birthed in the 1950s [ 1 , 2 , 3 , 4 , 5 ] , effort has been made in fulfilling the promise of such devices for faster, more efficient data storage hardware [ 6 , 7 ] . Ferroelectric devices introduce a ferroelectric thin film to established device architectures, where the addition of a switchable spontaneous polarization enables storage of binary values as up/down polarization within the ferroelectric layer. However, the prospect for ferroelectric devices was initially limited since the perovskite-based complex oxide ferroelectrics that dominated the most prominent era in ferroelectric device research of the 80s-00s, depolarize in sub-10 nm thicknesses and cannot be easily integrated into existing CMOS architectures. [ 8 , 9 , 10 ] In 2011, the first report of robust ferroelectricity in silicon-doped HfO 2 [ 11 ] was initially met with skepticism: ferroelectric HfO 2 was shown to overcome not one but both major limitations of other ferroelectric materials: i) it did not depolarize at the nanoscale and ii) its history as a high-permittivity (high-k) gate insulator for high-performance field-effect transistors (FETs) proved it was compatible with CMOS technology. Over time this behaviour was proven to be reproducible, leading to gradual acceptance and a subsequent explosion of research into ferroelectric hafnia. The origin of ferroelectric behaviour in these first polycrystalline films synthesized by atomic layer deposition (ALD) was attributed to a metastable orthorhombic phase ( Pca2_{1},OIII ), but the mechanisms that make this material so different from other ferroelectrics ( i.e. the reason why it does not depolarize at the nanoscale) remains unknown [ 12 , 10 , 13 ] . Since then, extensive research into the material has improved understanding regarding optimization of ferroelectric performance in HfO 2 -based devices. The principal hurdle in obtaining ferroelectric hafnia is that the non-polar monoclinic P2_{1}/c,\ m phase is far more stable than the metastable polar OIII phase. Consequently, any strategy for achieving the metastable ferroelectric phase hinges on avoiding the large unit cell volume of the former phase. As the window of stability of the polar phase is small, the easiest approach is to first stabilize the much lower volume tetragonal P4_{2}/nmc phase and then slightly enlarge the cell such that it may form the polar phase. There are several mechanisms capable of achieving this, which are often used simultaneously for enhanced stability and performance. Firstly (and most prominently), the use of surface tension through synthesis of small grains in polycrystalline films or thin single crystal films forces a smaller unit cell volume, thereby stabilizing the tetragonal or even cubic ( Fm\bar{3}m ) phase at annealing temperatures which, upon cooling, relax into the metastable OIII phase [ 14 , 15 ] . In the case of epitaxial thin films, the strain associated with the surface tension is further aided by the additional substrate-induced (epitaxial) strain which will also dictate the orientation of the polar axis [ 16 , 17 , 18 , 19 ] . In addition, doping HfO 2 with various elements (Si, Zr, La, Y, etc.) has also been shown to be an effective strategy in improving ferroelectric phase stability and device performance for both polycrystalline and epitaxial thin films through the associated ”chemical pressure” induced by dopants [ 20 , 21 , 22 , 23 , 24 ] . A popular choice of hafnia composition is the Hf 0.5 Zr 0.5 O 2 solid solution, which serves to both stabilize the ferroelectric phase and to increase the remnant polarization relative to pure hafnia [ 14 ] . Similarly, the strain associated with the presence of oxygen vacancies has also been shown to be beneficial in both polycrystalline and epitaxial thin films [ 25 , 26 ] . There is one important distinction between polycrystalline and epitaxial hafnia films: since epitaxial films enable control of the crystalline direction, the polar axis of the film is uniform across the layer and there is no need to ”wake-up” the films by electrical cycling after growth [ 27 ] . An additional distinguishing factor is that there are two different metastable ferroelectric phases accessible to epitaxial thin films: the OIII phase mentioned thus far [ 16 , 18 , 24 ] and two rhombohedral phases that display similar behaviour. These non-centrosymetric rhombohedral R3m [ 27 , 26 ] and R3 [ 28 , 29 , 30 ] phases have previously been found in epitaxial films under high compressive strains and exhibit large polarization values. However, they typically suffer from larger coercive fields, close to the breakdown fields, which increase the operation power and pose a challenge to the device cyclability and endurance. In the particular case of epitaxial films grown on SrTiO 3 , the choice of bottom electrode has been found to be vital for R3m phase stability through an as-of-yet unclear mechanism. It seems thus far that only La 1-x Sr x MnO 3 enables growth of high quality r-phase epitaxial hafnia films, which has been hypothesized to be a result of either its role as an oxygen reservoir, or as buffer layer providing epitaxial strain relaxation or due to charge transfer [ 31 , 32 ] . Recently, there has been increased interest in the synthesis of HfO 2 -ZrO 2 multilayers, which have been demonstrated to enhance spontaneous polarizations and cyclability relative to equivalent solid solution Hf 1-x Zr x O 2 for ALD multilayers [ 33 , 34 , 35 , 36 ] . In the context of epitaxial superlattices, only one report of Hf 0.5 Zr 0.5 -HfO 2 superlattices grown on SrTiO 3 exists, where an increase to both the polarization and dielectric response was observed at the cost of device cyclability [ 37 ] . The latter observation is surprising as increased cyclability has been readily observed in ALD multilayers, where it has been attributed to redistribution of defects along the multilayer interfaces [