The discovery of ferroelectric hafnium oxide enabled a variety of non-volatile memory devices, like ferroelectric tunnel junctions or field-effect transistors. Reliable application of hafnium oxide based electronics in space or other high-dose environments requires an understanding of how these devices respond to highly ionizing radiation. Here, the effect of 1.6 GeV Au ion irradiation on these devices is explored, revealing a reversible phase transition, as well as a grain fragmentation process. The collected data demonstrate that non-volatile memory devices based on ferroelectric hafnia layers are ideal for applications where excellent radiation hardness is mandatory.

Hafnium oxide (hafnia)-based functional layers are one of the most promising material systems for emerging memory applications, usable as high-k dielectrics,1,2 as resistive switching layers for resistive random access memory (RRAM),3–5 and as ferroelectrics.6–9 Since the discovery and first publication of ferroelectricity in hafnia,10,11 the interest in possible technological applications and corresponding research activities have immensely increased.12,13 Memory concepts like the usage of hafnia in ferroelectric field effect transistor (FeFET), in ferroelectric random-access memory (FeRAM), or in ferroelectric tunneling junctions (FTJs) offer benefits like a high scalability and good retention for highly scaled devices and can be integrated in the conventional complementary metal–oxide–semiconductor (CMOS) front-end-of-line (FEoL) and back-end-of-line (BEoL) processes.14–17 Even more applications are possible, e.g., pyroelectric and piezoelectric sensors as well as energy storage systems.18–20 The most promising and intensively studied hafnia-based ferroelectric material systems developed within the last decade are silicon doped hafnium oxide (HSO) and zirconium hafnium oxide (HZO).6,10 In FeRAM, two different memory states can be achieved in distinct electrical polarization states of a polar structure in the crystalline ferroelectric material. Compared to the conventional charge-based Flash technology, this information storage mechanism is, therefore, not directly based on charge flow, but strongly on the crystal structure and defects of the active material.9,21,22

The phase responsible for ferroelectricity in polycrystalline hafnia films has been identified as the orthorhombic Pca21 phase.23 Since this phase is metastable, polymorphism is observed in hafnia films, often containing an additional contribution of the monoclinic phase,24 which forms the thermodynamical ground state. Moreover, hafnia films often suffer from a severe wake-up effect, which describes the transition from an initially antiferroelectric-like behavior toward ferroelectric behavior upon electric field cycling.10 Multiple explanations have been suggested, such as defect redistribution and phase transitions.10,25 Recent results point toward 90° domain wall movement and stress relaxation being responsible for the observed behavior.9,26,27 Moreover, hafnia films can be affected by imprint, which has often been related to domain wall pinning and internal bias fields caused by oxygen vacancies.28 

In the past, non-hafnia-based ferroelectric electronics have already been used for application in radiation-harsh environments.29–31 Also, hafnia-based electronics for resistive memory applications have recently been demonstrated to show an improved resilience toward ionizing radiation compared to CMOS based technologies.32–39 Similarly, studies investigating electronics consisting of ferroelectric (and anti-ferroelectric) hafnia films exposed to γ, proton or ion radiation were reported.40–46 Only a minor influence of the irradiation on the electrical properties was reported, which is mainly related to single event effects due to the creation of oxygen vacancies in the oxide layers when exposed to radiation of up to a dose of about 13 Mrad. These promising radiation hardness results are making these material systems interesting candidates for radiation-hard memories. As the functionality of oxide-based ferroelectric devices is strongly dependent on the amount and distribution of defects, changes induced by irradiation might influence the electrical behavior of devices (single event effects). Additionally, high ion energies combined with high fluences can lead to a phase transition in hafnium oxide (multi event effects). Such phenomena were reported to appear in crystalline hafnium oxide bulk47 and films48 when a specific energy loss threshold (about 18 keV/nm in monoclinic HfO2) is exceeded and a high fluence in the range of 1012ions/cm2 (leading to a defective track overlap) is reached. The phase transition seems to be due to a double-hit process, where a first ion impact creates oxygen defects, while the hit of a second ion on the impact zone predamaged by the first ion triggers a crystalline-to-crystalline phase transition. Similar to the recently reported oxygen defect-induced phase transition in crystalline oxygen engineered films49 and in non-doped hafnia films exposed to heavy ion radiation,48,50 where the important role of the oxygen stoichiometry and oxygen defects in the phase transition and phase stabilization process was revealed, an ion irradiation-induced phase transition was also reported for ferroelectric (doped) hafnia.46 Kämpfe et al. described structural and electrical polarization (P–V) changes depending on the ion fluence of films irradiated with 1.63 GeV Au-ions. At low fluences, no structural changes were detected and only a minor influence on the electrical properties was found, while at fluences above 1×1012ions/cm2, similar to findings in non-doped hafnia layers, crystalline-to-crystalline structural changes and reduced remanent polarization values were found. In addition, recent reports reported improved ferroelectric properties after light ion irradiation,51 thus providing additional evidence for phase transitions under ion irradiation. This raises important questions regarding a possible temporary or permanent loss of stored information after irradiation and the connected mechanisms. The objective of this study was to investigate the response of the structural and ferroelectric properties of Si:HfO2 and Hf0.5Zr0.5O2-based metal–ferroelectric–metal (MFM) capacitor stacks exposed to 1.635 GeV Au ions at various fluences, including additional post-irradiation P–V cycling and systematic characterization of the remanent and spontaneous polarization, endurance, and the influence of pre-poling. The results offer a better understanding of the functionality and of possible failure mechanisms of ferroelectric emerging memory cells, which are required for the successful incorporation of new emerging memory materials and advanced microelectronic systems into radiation applications.

Metal–ferroelectric–metal (MFM) capacitors were fabricated on highly boron-doped silicon wafers. The TiN bottom electrode was deposited utilizing atomic layer deposition (ALD) at a temperature of 450°C employing TiCl4 and NH3 as precursors. 20 nm thick amorphous hafnium-silicon-oxide (HSO) and hafnium-zirconium-oxide (HZO) films as well as 10 nm thin HZO films were also deposited via ALD, utilizing HfCl4 and SiCl4 as precursors together with H2O as the oxidizing reactant and Ar as a purging gas. A constant silicon doping concentration of around 3 at. % was achieved by using a precursor cycling ratio of 16:1. Similarly, HZO films were produced using HfCl4 and ZrCl4 as the precursors with a cycling ratio of 1:1 to achieve Hf0.5Zr0.5O2 film stoichiometry. The film thickness was adjusted by the number of total deposition cycles. The TiN top electrode was fabricated by sputtering at low temperature to avoid in situ crystallization of the hafnia layer. To achieve a polar ferroelectric structure, crystallization of the doped hafnium oxide layer was triggered in N2 atmosphere by rapid thermal processing treatment for 20 s at a temperature of 650°C for HSO films and by furnace annealing at 400°C for 1 h for HZO films.

The heavy ion irradiation experiments were performed at the X0-beamline at the UNILAC accelerator of the GSI Helmholtzzentrum fuer Schwerionenforschung in Darmstadt with Au ions of 1.635 GeV kinetic energy and fluences between 5×109 and 7×1012ions/cm2. The flux was limited to 5×108ions/cm2s to avoid macroscopic heating of the samples. A fluence study was performed within a range from 5×109 to 7×1012ions/cm2. A schematic of the investigated stacks is shown in Fig. 1.

FIG. 1.

Schematic illustration of a hafnia-based metal–ferroelectric–metal (MFM) capacitor irradiated with 1.635 GeV Au ions. At this energy, the ions pass completely through the TiN and HfO2 layers and stop in the Si substrate.

FIG. 1.

Schematic illustration of a hafnia-based metal–ferroelectric–metal (MFM) capacitor irradiated with 1.635 GeV Au ions. At this energy, the ions pass completely through the TiN and HfO2 layers and stop in the Si substrate.

Close modal

X-ray diffraction investigations before and after heavy ion irradiation experiments were carried out using a Rigaku SmartLab diffractometer with a 9 kW rotating Cu anode (Kα-radiation) to obtain the characteristics of the crystalline structures on full sheet-layer samples.

To test the electrical properties of the MFM capacitors, Ti/Pt top electrode dots were deposited by electron beam evaporation utilizing a shadow mask. Afterward, the TiN top electrode material between the dot-contacts was removed by a wet etching step to form separated MFM capacitors of 0.150–0.006mm2. Electrical programming and measurements were carried out by utilizing a cascade microtech bench with an Agilent B1500 parameter analyzer. Voltage-dependent electrical polarization (P–V) hysteresis measurements were performed with an Aixacct TF 3000 FE analyzer using a triangular waveform at a frequency of 1 kHz. Electric field cycling is performed at the same frequency with an amplitude of 3 MV/cm. The highly doped silicon wafers thereby acted as a bottom ground during the electrical measurements. For poling prior irradiation, a triangular pulse of positive or negative amplitude (3 MV/cm) was performed to set the respective state.

After irradiation, the structural and electrical behavior is connected and compared to the properties of the non-irradiated reference samples.

The XRD patterns of the three different sample series containing 20 nm HSO, 20 nm HZO, and 10 nm HZO before and after ion exposure at different fluences are shown in Fig. 2. As-grown stacks consist of a mixture of the monoclinic (space group: P21/c; ICDD: 00 034 0104) and the non-centrosymmetric ferroelectric orthorhombic (space group Pca21; ICDD: 04-005-5597) phases (in 20 nm thick films). A monoclinic phase is not observed in the pattern of as grown stacks containing 10 nm HZO. For irradiations up to a fluence of 5×1011ions/cm2, the initial crystallinity is preserved. Larger ion fluences starting at 3×1012ions/cm2 have a clear impact on the crystalline structure of the films. In 20 nm thick films [Figs. 2(a) and 2(b)], a decrease/vanishing of the monoclinic phase (at around 2θ=28.5°) is observed at the fluence of 3×1012ions/cm2. The characteristics orthorhombic (111) reflection (2θ30.5°30.75°) is starting to shift toward larger diffraction angles (2θ30.75°30.9°) in all ferroelectric layers. At first glance, this is indicating a phase transition, as the behavior is directly comparable to reported irradiation-induced effects in non-doped hafnium oxide films.48,50 Based on the peak position, the resulting phase is most likely the cubic Fm3¯m phase that has an expected (111) line at slightly higher angles than the coinciding orthorhombic (111) and tetragonal (101) lines. This is in agreement with hafnium oxide phases reported in the literature.49,50 At the same time, the intensity of the reflections corresponding initially to the (111) orthorhombic phase is decreasing and in width broadened [full width at half maximum (FWHM): from about 0.32°0.46° in 20 nm HSO, 0.34°0.45° in 20 nm HZO, 0.43°0.49° in 10 nm HZO]. Reflex broadening of a reflection is visible in irradiated hafnium oxide-based films in literature (compare Refs. 48 and 50) and is a direct indication of irradiation-induced grain fragmentation at ion fluences above 1×1012ions/cm2, occurring in a range, where a defective track overlap is occurring. Here, initially large columnar grains were found, for irradiations above 1×1012ions/cm2, small and more circular grains were obtained. Both induced grain fragmentation and phase transition are interconnected phenomena, which are most likely a result of an increased number of induced (oxygen) defects in the oxide layer. This gets especially relevant also for electrical results of devices containing such oxide layers. An electronic energy loss of the ions in the ferroelectric layers of about 53 keV/nm can be estimated by TRIM calculations55 using a ferroelectric layer density of 8.79g/cm2. The damage process is ascribed to the electronic energy loss of the Au ions, which by far exceeds the reported threshold of about 18 keV/nm in hafnium oxide for a crystalline-to-crystalline phase transition47 and the corresponding nuclear loss is only about 68 eV/nm with the number of displacements per atom (dpa) in Hf0.5Zr0.5O2 being as low as 1E-4, according to Full Cascade TRIM calculations. The orthorhombic (020) reflection seems to be unchanged in all films, which is a result of the similar 2θ-position of the ferroelectric orthorhombic, and non-polar tetragonal, cubic, and other orthorhombic phases at this angle. As such phases are not (easily) distinguishable by XRD for strained, oriented films, electrical polarization measurement results (Fig. 3) are combined with the structural results to obtain a clearer picture.

FIG. 2.

XRD patterns of ferroelectric doped hafnium oxide films before and after irradiation, containing (a) 20 nm HSO, (b) 20 nm HZO, and (c) 10 nm HZO. With increasing fluence, a shift of the initially orthorhombic (111) reflection toward larger angles occurs. (d) Reference patterns of the monoclinic, ferroelectric orthorhombic, tetragonal, and cubic phase of HfO2, based on previously reported structures.52,53 Data for (a) and (b) have been reproduced from Ref. 46. Reflections obtained around 2θ=32.6°34° originate from the forbidden Si (200) Umweganregung.54 

FIG. 2.

XRD patterns of ferroelectric doped hafnium oxide films before and after irradiation, containing (a) 20 nm HSO, (b) 20 nm HZO, and (c) 10 nm HZO. With increasing fluence, a shift of the initially orthorhombic (111) reflection toward larger angles occurs. (d) Reference patterns of the monoclinic, ferroelectric orthorhombic, tetragonal, and cubic phase of HfO2, based on previously reported structures.52,53 Data for (a) and (b) have been reproduced from Ref. 46. Reflections obtained around 2θ=32.6°34° originate from the forbidden Si (200) Umweganregung.54 

Close modal
FIG. 3.

The difference in the initial polarization response after ion irradiation can be observed in (a). After 10 000 cycles, the polarization hystereses have re-opened, showing ferroelectric behavior. Comparing the remanent polarization (b) after irradation, a clear trend to lower values with increased fluences can be observed.

FIG. 3.

The difference in the initial polarization response after ion irradiation can be observed in (a). After 10 000 cycles, the polarization hystereses have re-opened, showing ferroelectric behavior. Comparing the remanent polarization (b) after irradation, a clear trend to lower values with increased fluences can be observed.

Close modal

The electrical polarization of devices based on 20 nm HSO-, 20 nm HZO-, and 10 nm HZO before (Ref.) and after irradiation at various fluences (1×1010, 5×1011, and 2.4×1012ions/cm2) are shown in Fig. 3(a). Additionally, the results obtained after 10 000 cycles after heavy ion irradiation are presented. Boxplots (statistics) of the corresponding remanent polarization values after irradiation (not cycled) are shown in Fig. 3(b). At a fluence of 1×1010ions/cm2, no significant impact of the ions on the PV-characteristics [Fig. 3(a)] is visible, while at higher fluences a pinching of the loops accompanied by a gradual reduction of the remenant (PR) and saturation polarization (PS) is observed for all stacks. The PR is found to be about 3.7μC/cm2 before irradiation for HSO-containing stacks, about 7.6μC/cm2 for 20 nm HZO-, and about 10.7μC/cm2 for 10 nm HZO-containing devices. After being exposed to heavy ions, these PR values decrease [20 nm HSO: 3μC/cm2 (1×1010ion/cm2), 1.7μC/cm2 (5×1011ions/cm2), 0.6μC/cm2 (2.4×1012ions/cm2); 20 nm HZO: 7.1μC/cm2 (1×1010ion/cm2), 5.9μC/cm2 (5×1011ions/cm2), 3.4μC/cm2 (2.4×1012ions/cm2); 10 nm HZO: 10.4μC/cm2 (1×1010ions/cm2), 9.8μC/cm2 (5×1011ions/cm2), 6.2μC/cm2 (2.4×1012ions/cm2)]. Overall, HZO films show larger overall values when compared to the HSO-containing stacks. Specifically, 10 nm thin films reveal a high polarization before and after being exposed to heavy ions.

In addition, the 20 nm HSO samples reveal some interesting details in their P–V characteristics. While initially all samples behave similarly antiferroelectric-like, their slope differs, indicating different relative permittivity. Here, a decrease with increasing fluences is observed. Moreover, for the highest fluence, this reduction in permittivity even persists postcycling. This trend can also be observed, but to a much smaller degree, in the 20 nm HZO films as well. Lower relative permittivity indicates a reduction in symmetry, fitting well to an increased number of grain boundaries, equivalent to a larger disordered volume in the ferroelectric layer. This is often referred to as an amorphization in ferroelectric hafnium oxide, still after the exposure to high energy heavy ions a grain fragmentation and presence of nanocrystalline grains is expected.50,56 Additionally, PS is reduced initially, even though the same ferroelectric behavior is reached postcycling. This indicates a reduction of switchable orthorhombic phase in the initial case. Possible explanations are phase transitions, or domain wall pinning due to the generation of oxygen vacancies. An increase in mechanical stress, on the one hand, would alter the ferroelastic switching behavior,9 on the other hand, does not match the observed behavior, as the position of the displacement current peaks and, therefore, the inflection points do not shift in position.

Combining the XRD and polarization hysteresis results, it can be concluded that a phase transition is likely to occur at large fluences (above 1×1012ions/cm2), where major changes of the polarization are visible [Fig. 3(a)], which can be directly related to the shifts of the reflections in the XRD patterns (Fig. 2). At a fluence of 2.4×1012ions/cm2, the initial ferroelectric orthorhombic phase is partially transformed to a non-polar phase, i.e., either the tetragonal or the cubic phase (as suggested recently49,50), since similar results are observed in the XRD patterns of all three different stack configurations. Still, in all cases the ferroelectric orthorhombic phase seems to be partially preserved even after irradiation at huge fluences above 1×1012ions/cm2. As in this region significant overlapping of defective track is expected,47,48 a preservation of a polar crystalline structure with still good ferroelectric properties is remarkable. At fluences below 1×1012ions/cm2, where phase transitions in hafnium oxide-based films are less probable to occur, the visible minor polarization changes are most likely a result of introduced defects in the HSO and HZO layers.

Another remarkable result is the finding that the irradiation-induced pinching of the P–E loops can be re-opened by applying wake-up field cycles [Figs. 3(a) and 3(b)]. Hereby, the PR and PS are increasing again, even for devices being exposed to fluences as large as 5×1011ions/cm2 or 2.4×1012ions/cm2. A re-opening of the loops is a clear indication of a field-induced phase transformation back to the ferroelectric orthorhombic phase. In Fig. 3(b), statistical values of the remanent polarization obtained after cycling are given. The general trend toward a reduced polarization with increasing fluence is visible, still, for all fluences, the obtained PR exceeds the values of non-cycled irradiated devices. Interestingly, a larger re-opening [e.g., 20 nm HSO: 8.2μC/cm2 (1×1010ions/cm2), 7.3μC/cm2 (5×1011ions/cm2), 4μC/cm2 (2.4×1012ions/cm2] is especially seen in stacks with 20 nm thick HSO and HZO films when compared to the non-irradiated reference curves. This phenomenon can be explained by the missing monoclinic phase fraction of the final irradiated plus field-cycled stacks. The reference stacks were initially containing a small part of the non-polar monoclinic phase, which was reduced during irradiation and is not reformed after cycling either.

A comparison of the PS values over the electric field cycling is also validating the aforementioned trends (see Fig. 4). For all samples, independent of the material stack, an initial reduction of PS with increasing fluence is observed. However, upon cycling, a strong increase is directly observed and all further cycling does not change the value significantly. This is in Stark contrast to common antiferroelectic-like wake-up effects, where a constant PS is expected during electric field cycling.9 Instead, the observed behavior either fits to the expected behavior of sudden wake-up9 or a superposition of both classical and sudden wake-up. However, sudden wake-up has only been observed in initially amorphous hafnia films, crystallizing under the application of an electric field. This effect shows Arrhenius-like behavior and an effective activation energy barrier of 0.45 eV has been reported.9 Furthermore, it has been shown that the applied voltage allows to precisely control this crystallization process.8 While an increased amount of grain boundaries, which are amorphous in nature, due to grain fragmentation have been observed in the here presented samples, as discussed above, the applied electric field does not reach the prior reported necessary amplitudes required for the field-induced crystallization process.8 Since the reduction in PS persists even after cycling for high fluence samples, which has been related to the increased fraction of grain boundaries above, it is therefore unlikely that an electric-field induced crystallization is responsible for the observed behavior. Instead, a phase transition from a non-polar phase to the orthorhombic phase, similar to the irreversible phase-transition from the tetragonal phase, as reported by other groups,57,58 appears to be consistent with the here reported results.

FIG. 4.

Evolution of PS over electric field cycling (up to 10 000 cycles) at an amplitude of 3 MV/cm. A significant increase after initial cycling is observed, saturating after the first 1000 cycles. A reduction in PS even after cycling is observed for high fluences.

FIG. 4.

Evolution of PS over electric field cycling (up to 10 000 cycles) at an amplitude of 3 MV/cm. A significant increase after initial cycling is observed, saturating after the first 1000 cycles. A reduction in PS even after cycling is observed for high fluences.

Close modal

Finally, the impact of ion irradiation on pre-poled material is investigated to get insight into the influence of the polarization state on the irradiation induced defects. Figure 5 shows the polarization hystereses of samples poled positive and negative prior to the irradiation. While the observed trends are in superposition with the previously discussed effects, a clear imprint (shift of the coercive fields) can be observed in addition. This is especially pronounced in the HZO samples, whereas the HSO samples show only minor asymmetries. One remarkable observation is that the curves of the irradiated samples show a steeper switching transition around EC and a smaller coercive field window (2EC). Interestingly, only the coercive field of switching to the previously poled state shifts to lower values with higher fluence. While imprint has been related to the presence of oxygen vacancies before,28 the here observed trend is not expected. Instead, the defects and phase transitions generated by the irradiation in addition to the oxygen vacancies influence the domain dynamics, resulting in the observed behavior. More in-depth analysis of the oxygen vacancy dynamics in hafnia will be required in future studies to understand the here observed behavior fully.

FIG. 5.

Polarization hysteresis of capacitors poled prior to Au irradiation. A shift of the coercive fields to positive or negative fields is observed for negatively or positively poled samples, respectively.

FIG. 5.

Polarization hysteresis of capacitors poled prior to Au irradiation. A shift of the coercive fields to positive or negative fields is observed for negatively or positively poled samples, respectively.

Close modal

The here presented results provide a clear picture on the phase transitions and defects landscape of Si- and Zr-doped hafnia under ion irradiation. Based on the XRD and P–V data, the model presented in Fig. 6 can nicely explain the observed behavior.

FIG. 6.

Schematic illustration of the changes in the hafnia layer upon heavy ion irradiation and subsequent electric field cycling. The phase transition of monoclinic and orthorhombic grains to cubic phase can be transformed to the orthorhombic phase upon cycling. The grain fragmentation, on the other hand, appears to be irreversible.

FIG. 6.

Schematic illustration of the changes in the hafnia layer upon heavy ion irradiation and subsequent electric field cycling. The phase transition of monoclinic and orthorhombic grains to cubic phase can be transformed to the orthorhombic phase upon cycling. The grain fragmentation, on the other hand, appears to be irreversible.

Close modal

First, heavy ion irradiation induces structural changes: Monoclinic and partially orthorhombic phases undergo a phase transition to higher symmetry non-polar phases. Based on the XRD data, the resulting phase is the cubic Fm3¯m phase since a right shift of the line at about 30° is observed. This is also confirmed by the absence of additional lines from the orthorhombic, tetragonal, monoclinic phase close to 35°, which do not coincide with the cubic phase. This phase transition also explains the reduced polarization switching in the P–V hystereses. Moreover, grain fragmentation and an increased fraction of grain boundaries explain the reduction in remanent polarization, and the non-recoverable polarization loss for high fluences, as well as the line broadening in the XRD data.

Upon cycling, the cubic phase is transformed back to the orthorhombic phase, explaining the rapid transition upon electric field cycling as well as the slightly higher polarization values compared to the reference sample. The latter is a result of the originally monoclinic phase fraction now transformed to the orthorhombic phase. This highlights that the heavy ion irradiation induced orthorhombic-cubic phase transition is reversible.

Finally, the induced oxygen vacancies affect the switching behavior, resulting in pinched hysteresis curves for high fluences and influencing the imprint behavior of the devices.

The presented results highlight the high resilience of ferroelectric hafnium oxide-based films and electronics against high energy heavy ion irradiation in the electronic energy loss regime.

In conclusion, the presented data clearly indicate a heavy ion irradiation induced phase transition from the orthorhombic and monoclinic to the cubic phase. At the same time, grain fragmentation is likely to take place. The overall ferroelectric properties do not degrade for low fluences. Additional insights into the wake-up effect in ferroelectric hafnia revealed a reversible transition back to the ferroelectric orthorhombic phase and improved ferroelectric properties. Overall, ferroelectric non-volatile memory devices show excellent radiation hardness, thus being suitable for applications in, e.g., aero-space, where radiation hardness plays an important role.

The work leading to this publication has been undertaken in the framework of the projects WAKeMeUP and StorAIge, which received funding from the Electronic Components and Systems for European Leadership Joint Undertaking in collaboration with the European Union’s H2020 Framework Programme (No. H2020/2014-2020) and National Authorities, under the Grant Agreement Nos. 783176 and 101007321, respectively. Funding by the Federal Ministry of Education and Research (BMBF) under Contract Nos. 16ESE0298 and 16MEE0154 is gratefully acknowledged. The results presented are based on a UMAT experiment, which was performed at the X0-beamline of the UNILAC at the GSI Helmholtzzentrum für Schwerionenforschung, Darmstadt (Germany) in the frame of FAIR Phase-0.

The authors have no conflicts to disclose.

M.L. and T.V. contributed equally to this work.

Maximilian Lederer: Data curation (equal); Formal analysis (equal); Investigation (equal); Writing – original draft (equal); Writing – review and editing (equal). Tobias Vogel: Data curation (equal); Investigation (equal); Visualization (equal); Writing – original draft (equal); Writing – review and editing (equal). Thomas Kämpfe: Conceptualization (equal); Project administration (equal); Supervision (equal); Writing – review and editing (equal). Nico Kaiser: Investigation (equal); Writing – review and editing (equal). Eszter Piros: Methodology (equal); Writing – review and editing (equal). Ricardo Olivo: Data curation (equal); Formal analysis (equal); Writing – review and editing (equal). Tarek Ali: Data curation (equal); Writing – review and editing (equal). Stefan Petzold: Investigation (equal); Writing – review and editing (equal). David Lehninger: Data curation (equal); Formal analysis (equal); Investigation (equal); Project administration (equal); Writing – review and editing (equal). Christina Trautmann: Methodology (equal); Supervision (equal); Writing – review and editing (equal). Lambert Alff: Conceptualization (equal); Project administration (equal); Supervision (equal); Writing – review and editing (equal). Konrad Seidel: Project administration (equal); Supervision (equal); Writing – review and editing (equal).

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
J.-P.
Locquet
,
C.
Marchiori
,
M.
Sousa
,
J.
Fompeyrine
, and
J. W.
Seo
,
J. Appl. Phys.
100
,
051610
(
2006
).
2.
J. H.
Choi
,
Y.
Mao
, and
J. P.
Chang
,
Mater. Sci. Eng.: R: Rep.
72
,
97
(
2011
).
3.
S.
Petzold
,
A.
Zintler
,
R.
Eilhardt
,
E.
Piros
,
N.
Kaiser
,
S. U.
Sharath
,
T.
Vogel
,
M.
Major
,
K. P.
McKenna
,
L.
Molina-Luna
, and
L.
Alff
,
Adv. Electron. Mater.
5
,
1900484
(
2019
).
4.
S. U.
Sharath
,
S.
Vogel
,
L.
Molina-Luna
,
E.
Hildebrandt
,
C.
Wenger
,
J.
Kurian
,
M.
Duerrschnabel
,
T.
Niermann
,
G.
Niu
,
P.
Calka
,
M.
Lehmann
,
H.-J.
Kleebe
,
T.
Schroeder
, and
L.
Alff
,
Adv. Funct. Mater.
27
,
1700432
(
2017
).
5.
S.
Dirkmann
,
J.
Kaiser
,
C.
Wenger
, and
T.
Mussenbrock
,
ACS Appl. Mater. Interfaces
10
,
14857
(
2018
).
6.
J.
Müller
,
T. S.
Böscke
,
U.
Schröder
,
S.
Müller
,
D.
Bräuhaus
,
U.
Böttger
,
L.
Frey
, and
T.
Mikolajick
,
Nano Lett.
12
,
4318
(
2012
).
7.
T.
Ali
,
P.
Polakowski
,
S.
Riedel
,
T.
Büttner
,
T.
Kämpfe
,
M.
Rudolph
,
B.
Pätzold
,
K.
Seidel
,
D.
Löhr
,
R.
Hoffmann
,
M.
Czernohorsky
,
K.
Kühnel
,
X.
Thrun
,
N.
Hanisch
,
P.
Steinke
,
J.
Calvo
, and
J.
Müller
,
Appl. Phys. Lett.
112
,
222903
(
2018
).
8.
M.
Lederer
,
S.
Abdulazhanov
,
R.
Olivo
,
D.
Lehninger
,
T.
Kämpfe
,
K.
Seidel
, and
L. M.
Eng
,
Sci. Rep.
11
,
22266
(
2021
).
9.
M.
Lederer
,
R.
Olivo
,
D.
Lehninger
,
S.
Abdulazhanov
,
T.
Kämpfe
,
S.
Kirbach
,
C.
Mart
,
K.
Seidel
, and
L. M.
Eng
,
Phys. Status Solidi RRL
15
,
2100086
(
2021
).
10.
T. S.
Böscke
,
J.
Müller
,
D.
Bräuhaus
,
U.
Schröder
, and
U.
Böttger
,
Appl. Phys. Lett.
99
,
102903
(
2011
).
11.
J.
Müller
,
U.
Schröder
,
T. S.
Böscke
,
I.
Müller
,
U.
Böttger
,
L.
Wilde
,
J.
Sundqvist
,
M.
Lemberger
,
P.
Kücher
,
T.
Mikolajick
, and
L.
Frey
,
J. Appl. Phys.
110
,
114113
(
2011
).
12.
T.
Schenk
,
M.
Pešić
,
S.
Slesazeck
,
U.
Schröder
, and
T.
Mikolajick
,
Rep. Prog. Phys.
83
,
086501
(
2020
).
13.
T.
Mikolajick
,
U.
Schröder
, and
S.
Slesazeck
, in 5th IEEE Electron Devices Technology and Manufacturing Conference (EDTM) (IEEE, 2021), pp. 1–3.
14.
M.
Trentzsch
,
S.
Flachowsky
,
R.
Richter
,
J.
Paul
,
B.
Reimer
,
D.
Utess
,
S.
Jansen
,
H.
Mulaosmanovic
,
S.
Muller
,
S.
Slesazeck
,
J.
Ocker
,
M.
Noack
,
J.
Muller
,
P.
Polakowski
,
J.
Schreiter
,
S.
Beyer
,
T.
Mikolajick
, and
B.
Rice
, in 2016 IEEE International Electron Devices Meeting (IEDM) (IEEE, 2016), pp. 11.5.1–11.5.4.
15.
S.
Dünkel
,
M.
Trentzsch
,
R.
Richter
,
P.
Moll
,
C.
Fuchs
,
O.
Gehring
,
M.
Majer
,
S.
Wittek
,
B.
Muller
,
T.
Melde
,
H.
Mulaosmanovic
,
S.
Slesazeck
,
S.
Müller
,
J.
Ocker
,
M.
Noack
,
D.-A.
Lohr
,
P.
Polakowski
,
J.
Müller
,
T.
Mikolajick
,
J.
Hontschel
,
B.
Rice
,
J.
Pellerin
, and
S.
Beyer
, in IEEE International Electron Devices Meeting, IEDM (IEEE, Piscataway, NJ, 2017), pp. 19.7.1–19.7.4.
16.
D.
Lehninger
,
R.
Olivo
,
T.
Ali
,
M.
Lederer
,
T.
Kämpfe
,
C.
Mart
,
K.
Biedermann
,
K.
Kühnel
,
L.
Roy
,
M.
Kalkani
, and
K.
Seidel
,
Phys. Status Solidi A
217
,
1900840
(
2020
).
17.
X.
Wang
,
T.
Mikolajick
, and
M.
Grube
,
Phys. Status Solidi RRL
2022
, 2100572.
18.
C.
Mart
,
T.
Kämpfe
,
S.
Zybell
, and
W.
Weinreich
,
Appl. Phys. Lett.
112
,
052905
(
2018
).
19.
C.
Mart
,
T.
Kämpfe
,
R.
Hoffmann
,
S.
Eßlinger
,
S.
Kirbach
,
K.
Kühnel
,
M.
Czernohorsky
,
L. M.
Eng
, and
W.
Weinreich
,
Adv. Electron. Mater.
49
,
1901015
(
2020
).
20.
K.
Kühnel
,
M.
Czernohorsky
,
C.
Mart
, and
W.
Weinreich
,
J. Vac. Sci. Technol. B
37
,
021401
(
2019
).
21.
U.
Schröder
,
M.
Materano
,
T.
Mittmann
,
P. D.
Lomenzo
,
T.
Mikolajick
, and
A.
Toriumi
,
Jpn. J. Appl. Phys.
58
,
SL0801
(
2019
).
22.
S.
Starschich
,
S.
Menzel
, and
U.
Böttger
,
Appl. Phys. Lett.
108
,
032903
(
2016
).
23.
X.
Sang
,
E. D.
Grimley
,
T.
Schenk
,
U.
Schröder
, and
J. M.
LeBeau
,
Appl. Phys. Lett.
106
,
162905
(
2015
).
24.
T. D.
Huan
,
V.
Sharma
,
G. A.
Rossetti
, and
R.
Ramprasad
,
Phys. Rev. B
90
,
38
(
2014
).
25.
M.
Pešić
,
F. P. G.
Fengler
,
L.
Larcher
,
A.
Padovani
,
T.
Schenk
,
E. D.
Grimley
,
X.
Sang
,
J. M.
LeBeau
,
S.
Slesazeck
,
U.
Schröder
, and
T.
Mikolajick
,
Adv. Funct. Mater.
26
,
4601
(
2016
).
26.
S.
Kirbach
,
M.
Lederer
,
S.
Eßlinger
,
C.
Mart
,
M.
Czernohorsky
,
W.
Weinreich
, and
T.
Wallmersperger
,
Appl. Phys. Lett.
118
,
012904
(
2021
).
27.
M.
Lederer
,
K.
Mertens
,
R.
Olivo
,
K.
Kühnel
,
D.
Lehninger
,
T.
Ali
,
T.
Kämpfe
,
K.
Seidel
, and
L. M.
Eng
,
J. Mater. Res.
36
,
4370
(
2021
).
28.
Y.
Higashi
,
K.
Florent
,
A.
Subirats
,
B.
Kaczer
,
L.
Di Piazza
,
S.
Clima
,
N.
Ronchi
,
S. R. C.
McMitchell
,
K.
Banerjee
,
U.
Celano
,
M.
Suzuki
,
D.
Linten
, and
J.
van Houdt
, in IEEE International Reliability Physics Symposium, IRPS (IEEE, 2019), pp. 1–7.
29.
S.
Gerardin
and
A.
Paccagnella
,
IEEE Trans. Nucl. Sci.
57
,
3016
(
2010
).
30.
R.
Sayyah
,
T. C.
Macleod
, and
F. D.
Ho
,
Ferroelectrics
413
,
170
(
2011
).
31.
S. C.
Philpy
,
D. A.
Kamp
,
A. D.
DeVilbiss
,
A. F.
Isacson
, and
G. F.
Derbenwick
, in IEEE Aerospace Conference (IEEE, Piscataway, 2000), pp. 377–383.
32.
M. J.
Marinella
,
IEEE Trans. Nucl. Sci.
68
,
546
(
2021
).
33.
S.
Petzold
,
S. U.
Sharath
,
J.
Lemke
,
E.
Hildebrandt
,
C.
Trautmann
, and
L.
Alff
,
IEEE Trans. Nucl. Sci.
66
,
1715
(
2019
).
34.
N.
Arun
,
L. D. V.
Sangani
,
K.
Vinod Kumar
,
A.
Mangababu
,
M. G.
Krishna
,
A. P.
Pathak
, and
S. V. S.
Nageswara Rao
,
J. Mater. Sci.: Mater. Electron.
32
,
2973
(
2021
).
35.
J. S.
Bi
,
Z. S.
Han
,
E. X.
Zhang
,
M. W.
McCurdy
,
R. A.
Reed
,
R. D.
Schrimpf
,
D. M.
Fleetwood
,
M. L.
Alles
,
R. A.
Weller
,
D.
Linten
,
M.
Jurczak
, and
A.
Fantini
,
IEEE Trans. Nucl. Sci.
60
,
4540
(
2013
).
36.
X.
He
and
R. E.
Geer
, in 2013 IEEE Aerospace Conference (IEEE, 2013), pp. 1–7.
37.
S. L.
Weeden-Wright
,
W. G.
Bennett
,
N. C.
Hooten
,
E. X.
Zhang
,
M. W.
McCurdy
,
M. P.
King
,
R. A.
Weller
,
M. H.
Mendenhall
,
M. L.
Alles
,
D.
Linten
,
M.
Jurczak
,
R.
Degraeve
,
A.
Fantini
,
R. A.
Reed
,
D. M.
Fleetwood
, and
R. D.
Schrimpf
,
IEEE Trans. Nucl. Sci.
61
,
2972
(
2014
).
38.
X.
He
,
W.
Wang
,
B.
Butcher
,
S.
Tanachutiwat
, and
R. E.
Geer
,
IEEE Trans. Nucl. Sci.
59
,
2550
(
2012
).
39.
R.
Fang
,
Y.
Gonzalez Velo
,
W.
Chen
,
K. E.
Holbert
,
M. N.
Kozicki
,
H.
Barnaby
, and
S.
Yu
,
Appl. Phys. Lett.
104
,
183507
(
2014
).
40.
F.
Huang
,
Y.
Wang
,
X.
Liang
,
J.
Qin
,
Y.
Zhang
,
X.
Yuan
,
Z.
Wang
,
B.
Peng
,
L.
Deng
,
Q.
Liu
,
L.
Bi
, and
M.
Liu
,
IEEE Electron Device Lett.
38
,
330
(
2017
).
41.
W.
Zhang
,
G.
Li
,
X.
Long
,
L.
Cui
,
M.
Tang
,
Y.
Xiao
,
S.
Yan
,
Y.
Li
, and
W.
Zhao
,
Phys. Status Solidi B
257
,
1900736
(
2020
).
42.
C.
Liu
,
W.
Xiao
,
Y.
Peng
,
B.
Zeng
,
S.
Zheng
,
L.
Yin
,
Q.
Peng
,
X.
Zhong
,
M.
Liao
, and
Y.
Zhou
,
IEEE Trans. Electron Devices
68
,
4368
(
2021
).
43.
X.
Yang
,
J.
Bi
,
Y.
Xu
,
K.
Xi
, and
L.
Ji
,
Appl. Phys. Express
14
,
061001
(
2021
).
44.
Y.
Wang
,
F.
Huang
,
Y.
Hu
,
R.
Cao
,
T.
Shi
,
Q.
Liu
,
L.
Bi
, and
M.
Liu
,
IEEE Electron Device Lett.
39
,
823
(
2018
).
45.
W. L.
Zhang
,
Y. H.
Mao
,
L.
Cui
,
M. H.
Tang
,
P. Y.
Su
,
X. J.
Long
,
Y. G.
Xiao
, and
S. A.
Yan
,
Phys. Chem. Chem. Phys.
22
,
21893
(
2020
).
46.
T.
Kämpfe
,
T.
Vogel
,
R.
Olivo
,
M.
Lederer
,
N.
Kaiser
,
S.
Petzold
,
T.
Ali
,
D.
Lehninger
,
C.
Trautmann
,
L.
Alff
, and
K.
Seidel
, in Joint Conference of the IEEE International Frequency Control Symposium and International Symposium on Applications of Ferroelectrics (IFCS-ISAF) (IEEE, 2020), pp. 1–3.
47.
A.
Benyagoub
,
Phys. Rev. B
72
,
1277
(
2005
).
48.
T.
Vogel
,
N.
Kaiser
,
S.
Petzold
,
E.
Piros
,
N.
Guillaume
,
G.
Lefevre
,
C.
Charpin-Nicolle
,
S.
David
,
C.
Vallee
,
E.
Nowak
,
C.
Trautmann
, and
L.
Alff
,
IEEE Trans. Nucl. Sci.
68
,
1542
(
2021
).
49.
N.
Kaiser
,
T.
Vogel
,
A.
Zintler
,
S.
Petzold
,
A.
Arzumanov
,
E.
Piros
,
R.
Eilhardt
,
L.
Molina-Luna
, and
L.
Alff
,
ACS Appl. Mater. Interfaces
14
,
1290
(
2022
).
50.
T.
Vogel
,
A.
Zintler
,
N.
Kaiser
,
N.
Guillaume
,
G.
Levèvre
,
M.
Lederer
,
A. L.
Serra
,
E.
Piros
,
T.
Kim
,
P.
Schreyer
,
R.
Winkler
,
D.
Nasiou
,
R.
Olivo
,
T.
Ali
,
D.
Lehninger
,
C.
Charpin-Nicolle
,
G.
Bourgeois
,
L.
Grenouillet
,
M.-C.
Cyrille
,
G.
Navarro
,
K.
Seidel
,
T.
Kämpfe
,
S.
Petzold
,
C.
Trautmann
,
L.
Molina-Luna
, and
L.
Alff
, “Structural and electrical response of emerging memories exposed to heavy ion radiation,” ACS Nano (to be published).
51.
S.
Kang
,
W.-S.
Jang
,
A. N.
Morozovska
,
O.
Kwon
,
Y.
Jin
,
Y.-H.
Kim
,
H.
Bae
,
C.
Wang
,
S.-H.
Yang
,
A.
Belianinov
,
S.
Randolph
,
E. A.
Eliseev
,
L.
Collins
,
Y.
Park
,
S.
Jo
,
M.-H.
Jung
,
K.-J.
Go
,
H. W.
Cho
,
S.-Y.
Choi
,
J. H.
Jang
,
S.
Kim
,
H. Y.
Jeong
,
J.
Lee
,
O. S.
Ovchinnikova
,
J.
Heo
,
S. V.
Kalinin
,
Y.-M.
Kim
, and
Y.
Kim
,
Science
376
,
731
(
2022
).
52.
R.
Materlik
,
C.
Künneth
, and
A.
Kersch
,
J. Appl. Phys.
117
,
134109
(
2015
).
53.
M.
Lederer
,
T.
Kämpfe
,
R.
Olivo
,
D.
Lehninger
,
C.
Mart
,
S.
Kirbach
,
T.
Ali
,
P.
Polakowski
,
L.
Roy
, and
K.
Seidel
,
Appl. Phys. Lett.
115
,
222902
(
2019
).
54.
P.
Zaumseil
,
J. Appl. Crystallogr.
48
,
528
(
2015
).
55.
J. F.
Ziegler
,
M. D.
Ziegler
, and
J. P.
Biersack
,
Nucl. Instrum. Methods Phys. Res., Sect. B
268
,
1818
(
2010
).
56.
M.
Dhanunjaya
,
D. K.
Avasthi
,
A. P.
Pathak
,
S. A.
Khan
, and
S. V. S.
Nageswara Rao
,
Appl. Phys. A
124
,
587
(
2018
).
57.
Y.
Tashiro
,
T.
Shimizu
,
T.
Mimura
, and
H.
Funakubo
,
ACS Appl. Electron. Mater.
3
,
3123
(
2021
).
58.
T.
Shimizu
,
Y.
Tashiro
,
T.
Mimura
,
T.
Kiguchi
,
T.
Shiraishi
,
T. J.
Konnno
,
O.
Sakata
, and
H.
Funakubo
,
Phys. Status Solidi RRL
15
,
2000589
(
2021
).