Ferroelectric [HfO2/ZrO2] Superlattices with Enhanced Polarization, Tailored Coercive Field, and Improved High Temperature Reliability

Abstract

Modern microelectronic systems and applications demand an every increasing amount of non‐volatile memories that are fast, reliable, and consume little power. Memory concepts based on ferroelectric HfO₂ like the ferroelectric field effect transistor (FeFET) and the ferroelectric random access memory (FeRAM) are promising to satisfy these requirements. As a consequence, continuing high attention is given to improve the ferroelectric properties and the reliability characteristics of the ferroelectric HfO₂ films – for instance by using different dopant elements, dopant concentrations, and film thicknesses. Superlattices (i.e., a periodic structure of two materials stacked upon each other) are a promising alternative approach. Herein, [HfO₂/ZrO₂] superlattices of various sublayer thicknesses and a constant total thickness of 10 nm are embedded into metal‐ferroelectric‐metal (MFM) capacitors and then electrically as well as structurally characterized with special focus on remanent polarization, coercive field, endurance, and high temperature reliability. Compared to a 10 nm (Hf,Zr)O₂ solid solution reference film, the use of superlattice stacks significantly improves the above mentioned parameters. In addition, most of these parameters depend on the sublayer thickness, which allows, for instance, tailoring the coercive field of the whole device.

Introduction

The demand for data storage increases persistently in modern microelectronic systems. Many applications like for instance sensors for internet of things devices or microcontrollers for automotive solutions, require non-volatile memories that are fast, reliable, and consume little power.^[1,2] Memories based on ferroelectric materials like the ferroelectric field effect transistor (FeFET), the ferroelectric random access memory (FeRAM), and the ferroelectric tunnel junction (FTJ) are auspicious to satisfy these requirements.^[3]

At zero external electric field, ferroelectric materials have two distinguishable remanent polarization states (P_r⁺ and P_r^-) that can be used to store information. Switching between both states is triggered by applying an external electric field that is larger than the coercive field (|E_c|) of the ferroelectric.

First memory concepts based on perovskite-type ferroelectrics were suggested already in the late 1950s.^[4] However, these conventional ferroelectrics require high layer thicknesses (due to the small coercive and breakdown fields) and their integration into complementary metal–oxide–semiconductor (CMOS) technologies is rather challenging. Consequently, their use is limited to niche applications.^[5]

In 2011, ferroelectricity was detected in HfO₂,^[6] a material successfully integrated into standard CMOS high-k metal gate technologies just a few years before.^[7] The ferroelectric properties of the HfO₂-based films are attributed to the polar, non-centrosymmetric, orthorhombic (o) phase of space group $$Pca2_{1}$$,^[8] which can be stabilized, for instance, by introducing impurity atoms.^[9,10] These findings have led to a renaissance of the above mentioned ferroelectric memory concepts^[11–15] and a continuing high demand to improve the ferroelectric and reliability characteristics of the HfO₂ films,^[16] for instance by testing different dopant elements,^[10,17] dopant concentrations,^[18] and film thicknesses.^[19–21]

To further optimize important characteristics of the ferroelectric films, superlattices are a promising alternative approach. Superlattices are periodic structures of two (or more) materials with layer thicknesses in the range of nanometers. This concept is not new. In 2014, Park et al. reported on ferroelectric characteristics of HfO₂/ZrO₂ stacks^[22] followed by other groups.^[23–28] A review about the development of that topic is given in reference.^[29]

The observation of ferroelectricity in [HfO₂/ZrO₂]_x superlattices is (at least) unexpected, since the increased surface/interface energy effect (without doping) does not necessarily explain the stabilization of the ferroelectric phase. A first insight into the mechanism of the formation of the o-phase in superlattices was given by Park et al.^[30] Just recently, superlattices were also embedded into the gate stack of FeFETs.^[31–33]

Herein, we report on the ferroelectric characteristics of [HfO₂/ZrO₂] superlattices with various sublayer thicknesses at a constant total thickness of 10 nm. The stacks are characterized with respect to remanent polarization, coercive field, endurance, and high temperature reliability. These parameters clearly depend on the sublayer thickness, which allows tailoring of for instance the coercive field within certain limits.

Metal-ferroelectric-metal (MFM) capacitors are used as simple test vehicle to characterize the ferroelectric stack. They are also an important building block for embedded ferroelectric memories. Placed in the back-end-of-line (BEoL) of microchips, they can be connected either to the gate or drain contact of a standard logic device to realize a 1T1C FeFET^{[12,13,34,35]} and a 1T1C FeRAM,^[36] respectively.

Experimental Section

MFM capacitors were fabricated on highly boron doped 300 mm silicon wafers. The 20 nm thick TiN bottom and top electrodes were deposited via reactive magnetron sputtering. The ferroelectric films were deposited by atomic layer deposition (ALD) at 300°C. As precursors, hafnium tetrachloride (HfCl₄), and zirconium tetrachloride (ZrCl₄) were used together with water (H₂O) and argon (Ar) as oxidizing reactant and purging gas, respectively.

Five superlattice stacks with different sublayer thicknesses and a constant total thickness of 10 nm (confirmed by spectroscopic ellipsometry) were embedded as ferroelectric into the MFM structures: [0.25 nm HfO₂/0.25 nm ZrO₂]₂₀, [0.5 nm HfO₂/0.5 nm ZrO₂]₁₀, [1 nm HfO₂/1 nm ZrO₂]₅, [2.5 nm HfO₂/2.5 nm ZrO₂]₂, and [5 nm HfO₂/5 nm ZrO₂]. As reference, a 10 nm (Hf,Zr)O₂ solid solution film with constant stoichiometry and approximately equal proportions was added.

In case of the superlattices, the number of deposition cycles necessary to meet the targeted HfO₂ and ZrO₂ sublayer thicknesses were calculated from the growth per cycle (GPC) of the individual material. The (Hf,Zr)O₂ reference film was achieved by periodically depositing one cycle of HfO₂ and one cycle of ZrO₂. Therefore, it could be considered as the “superlattice” with the smallest sublayer thickness in the range of the GPC (≈ 0.05 nm). The crystallization anneal was done under N₂ atmosphere within a rapid thermal annealing (RTA) tool. The hold temperature, hold time, and ramp up rate were 400°C, 60 s, and 50 K s^-1, respectively. Cool down was done passively.

Discrete Ti/Pt dots were deposited with electron beam evaporation by using a shadow mask. Subsequently, the TiN between the dot contacts was removed by wet etching (SC1) to form individual MFM capacitors. Figure 1 depicts a schematic drawing of all samples. In addition, Figure 2 depicts the MFM stack with the 1 nm sublayer thickness, which was visually analyzed on cross-section specimen using a FEI Tecnai F20 transmission electron microscope (TEM).

[IMAGE OMITTED. SEE PDF]

[IMAGE OMITTED. SEE PDF]

The structural characteristics of the crystalline phases were concluded from X-ray diffraction experiments in a grazing-incident diffraction geometry (GIXRD) with a thin-film diffractometer (Bruker D8 Discover). The incident angle was set to 0.5°. To avoid the substrate reflection (311), all samples were mounted with a 45° rotation between the <100 > directions of the (100)-oriented silicon wafer and the diffraction plane.

The polarization versus electric field (P–E) measurements were carried out with an Aixacct TF 3000 FE analyzer using a triangular waveform at a frequency of 1 kHz. The endurance tests were performed with an electrical cycling frequency of 10 kHz.

Results and Discussion

The back-end of line (BEoL) compatible thermal budget used here (60 s at 400°C) is sufficient to crystallize the solid solution (Hf,Zr)O₂ reference film and all tested [HfO₂/ZrO₂] superlattices with a total thicknesses of 10 nm (cf., Figure 3). The XRD analysis can clearly distinguish the non-polar monoclinic (m)-phase from the cubic (c)-, the tetragonal (t)-, and the ferroelectric (non-centrosymmetric) o-phase. It is worth to mention that the lamination has no considerable impact on the formation of the m-phase, since the fraction of this phase is negligible for all films studied in here. For superlattices starting with the HfO₂ sublayer, a similar result was reported.^[30] On the contrary, if the first layer directly on the bottom TiN electrode is ZrO₂, a significant fraction of the m-phase may be found, as partly reported in the literature, for instance in [30].

[IMAGE OMITTED. SEE PDF]

It is apparent from the GIXRD pattern that the area of the diffraction line at 2θ ≈ 35° (corresponding to {200} planes of the o-phase) changes with the sublayer thickness. To investigate this, the area of this diffraction line is normalized to that of the main diffraction line at 2θ ≈ 30.5°, which corresponds to {111} planes of the same phase (cf., Figure 4). Superlattices with sublayer thicknesses between 0.25 nm and 1 nm have the largest fraction of the {200} diffraction line. If the sublayer is thicker and / or in case of the (Hf,Zr)O₂ (HZO) reference, this fraction is significantly smaller. In the literature, an increase of the {200}intensity/area is related to a crystallographic texture that enhances the polarization, as studied by transmission Kikuchi diffraction.^[37,38] Since the o-phase diffraction line at 2θ ≈ 18°, which belongs to the (100) plane, is not apparent in Figure 4, the same is expected to apply for the superlattices studied here.

[IMAGE OMITTED. SEE PDF]

The Scherrer equation can be used to calculate the grain size within the superlattices and the reference device. Due to the GIXRD geometry, the grain size is estimated at ≈15° off-axis along the vertical direction. By using the main diffraction line at 2θ ≈ 30.5° and a shape factor of 0.9, the estimated size is almost equal to the film thickness (cf. Figure 4). Thus, it can be concluded that the crystal growth is not interrupted at the sublayer interfaces and that the grains are most likely of columnar shape.

Since XRD cannot be used to clearly distinguish between the c-, t-, and o-phase (all mentioned phases have diffraction lines at similar 2θ angles), polarization versus electric field (P–E) measurements were conducted to verify the presence of the ferroelectric o-phase. As seen in Figure 5, all MFM capacitors are ferroelectric, which proofs the presence of the o-phase. It is worth to mention that devices with the solid solution (Hf,Zr)O₂ reference film depict almost no pinching of the hysteresis loop in the pristine state. In contrast, whenever the superlattice films are integrated into the MFM capacitors, pronounced antiferroelectric-like characteristics are observed before wake-up cycling. We conclude that the laminates either crystallize preferentially in the antiferroelectric-like t-phase^[39] and / or stabilize a higher amount of in-plane oriented domains, which can act antiferroelectric-like within the o-phase.^[38] However, less than 1000 field cycles are needed to fully wake-up these integrated superlattice devices (i.e., eliminate the antiferroelectric-like characteristics). During wake-up either a field-induced phase transformation from a t-rich to an o-rich state occurs with subsequent stabilization of the o-phase due to defect migration,^[39] and / or destabilization of antiferroelectric-like in-plane domains due to stress relaxation caused by, for example, defect (dislocation) generation might be present.^[38]

[IMAGE OMITTED. SEE PDF]

Assuming that crystallization occurs initially at the interface between the first deposited HfO₂ sublayer and the bottom TiN interface, as suggested for instance in [Ref. [30]], we may derive the following model-based-description to explain ferroelectricity of the [HfO₂/ZrO₂] superlattice stacks: The crystallization is triggered from the bottom TiN electrode either during deposition (depending on the temperature of the ALD process and the sublayer thickness) or during a post metallization anneal, as applied in this study. During this process, the first HfO₂ sublayer gets orientation and phase information, which are determined (among other factors) by the stress conditions introduced from the bottom TiN electrode. Thus, a significant fraction of the o-phase is stabilized. This information is then propagated across the entire laminated stack. This model also explains the significantly higher amount of m-phase in the stack, should the superlattice start with ZrO₂ as a first sublayer.^[30] For that, the crystallization (triggered again from the bottom TiN electrode) may start from the ZrO₂ sublayer and other crystallization information is propagated. In the case of thick ZrO₂ sublayers, the propagation of the crystallization information across the stack may be impeded, which explains that the fraction of the t-phase and/or of in-plane domains within the o-phase increases and a pinching of the hysteresis loops is observed that becomes more pronounced for an increasing ZrO₂ sublayer thickness.

One important reliability parameter that can be directly estimated qualitatively from the shape of the P–E curve is the leakage-density, since leakage leads to a “cigar-like” broadening of the hysteresis loop. For the measurement conditions used in Figure 5, no significant leakage is observed for the reference and the superlattice MFM capacitors.

The remanent polarization (2P_r) is another important parameter that characterizes ferroelectric films. For instance in the case of FeFETs, too low 2P_r-values could result in a reduced memory window (MW). Figure 6 depicts the remanent polarization of the solid solution reference device and the superlattice devices with various sublayer-thickness in the pristine state (a) and after field cycling (b–d). The superlattice itself and the thickness of the sublayer have strong impact on the measured remanent polarization. In the pristine state, the superlattices show smaller remanent polarization than the reference device. Furthermore, the remanent polarization of the superlattices decreases with increasing sublayer-thickness. This can be explained by the antiferroelectric-like characteristics of the superlattices, which become more prominent for thicker sublayers. After wake-up, the dependency between the sublayer thickness and the remanent polarization changes as follows: Starting from the solid solution reference device, the remanent polarization increases with increasing sublayer thickness. However, whenever surpassing the 1 nm sublayer thickness, this trend reverts and the remanent polarization starts to decrease with further increasing sublayer thickness. This result is in good agreement with the GIXRD data, where the largest fraction of {200} planes were observed for sublayer thicknesses between 0.25 and 1 nm. That a sublayer thickness of 1 nm results in the highest remanent polarization was reported before for similar stacks, for instance in reference.^[24,30] In addition, it should be mentioned that it is not limited to [HfO₂/ZrO₂] superlattices that the sublayer thickness influences the remanent polarization. Similar characteristics were observed for [Si-doped HfO₂/SiO₂] stacks with various SiO₂ thicknesses.^[40]

[IMAGE OMITTED. SEE PDF]

The coercive field (|E_c|) is another important parameter. Integrated into the gate stack of a logic device to build a FeFET, the MW is defined by the coercive field of the ferroelectric material and its film thickness (t_F): MW ≈ 2t_FE_c.^[41] Furthermore, the coercive field directly impacts the program and erase voltage and thus can be used to tune the power consumption during device operation. Also the endurance limitations related to high coercive field of HfO₂ based ferroelectrics is one important challenge for future memory applications. Using superlattices is a rather easy approach to tailor the coercive field, as the following results demonstrate.

Figure 7 shows the coercive field of the solid solution reference device and the superlattices with different sublayer-thicknesses for the pristine state (a) and after field cycling (b–d). In the case of the pristine state, the superlattices have significantly smaller coercive fields than the reference. In addition, the coercive field further decreases with increasing sublayer thickness. Similar to the trend observed in Figure 5a before, this can be explained by the increasing antiferroelectric-like characteristics of the superlattice devices with increasing sublayer thickness. However, in contrast to Figure 5b–d, the same trend is observed in Figure 7b–d after cycling, although the antiferroelectric-like characteristics of the hysteresis loops are vanished. While for the sublayers of 1 nm thickness and below, this might be explained by the out-of-plane texture of the <200 > axes, the origin for the reduced coercive field in the case of thicker sublayers is not expected. From the P–E loops, a slightly decreased slope is observed around the coercive field, indicating that small antiferroelectric-like fractions in the domain configuration remain and thus result in a lower effective coercive field.

[IMAGE OMITTED. SEE PDF]

The superlattice approach opens the possibility to tailor the coercive field (2|E_c|) between 3.1 MV cm^-1 (reference film after 10³ field cycles) and 2.1 MV cm^-1 (superlattice with 2.5 nm sublayers after 10³ field cycles), which corresponds to a reduction of about 30%. Alternatively, the coercive field can be tuned for instance by using hybrid ferroelectric/anti-ferroelectric stacks,^[42] Hf_xZr_1-xO₂ films of various compositions,^[18] or an interface modification of the bottom electrode through ultraviolet-ozone irradiation.^[43] Although the last mentioned approaches allow a modification of the coercive field within similar ranges, they suffer from disadvantages like for instance polarization reduction^[18] and/or hysteresis loop shape distortion.^[42]

Figure 8 displays the endurance characteristics. Cycling tests can basically be divided into three stages: wake-up, fatigue, and dielectric breakdown.^[44,45] During wake-up, the remanent polarization growths with increasing number of field cycles due to the de-pinching of the prior pinched hysteresis loop. This is observed in particular for the superlattice devices (cf., Figure 8b–d), which are more pinched in the pristine state. The impact of the sublayer thickness on the number of wake-up cycles is negligible for sublayer thicknesses up to 2.5 nm. However, stacks with 5 nm sublayers need a reduced number of field cycles for wake-up. This stage is attributed to a gradual decrease in the built-in bias field^[44] and/or a partial phase transition of the dielectric material from the t- to the o-phase due to a redistribution of defects (oxygen vacancies).^[46] Recently also stress relaxation due to defect generation and thus the reduction of in-plane domains was suggested as the underlying mechanism.^[38] After the wake-up process is completed, the reference sample and the superlattice devices get straight into the fatigue (or aging) phase. During this stage, the remanent polarization decreases with increasing field cycles due to the generation and field drift of defects like oxygen vacancies that facilitate charge trapping and / or domain pinning. Finally, the main failure mode is a dielectric breakdown, which occurs abrupt for all tested devices. This stage is attributed to the formation of permanent conduction paths that substantially promote enhanced leakage currents. Usually the field drift and accumulation of oxygen vacancies are attributed to this failure mode, similar to filament formation in HfO₂-based resistive random access memory (RRAM) devices.

[IMAGE OMITTED. SEE PDF]

The maximum endurance (defined as the number of field cycles before dielectric breakdown) is shown in Figure 9 for the solid solution reference device and the superlattice of different sublayer-thicknesses after cycling at different electrical fields. As expected, the endurance decreases with increasing electrical fields applied to the devices and with increasing sublayer-thickness. It was reported elsewhere that devices with HfO₂/ZrO₂/HfO₂ trilayers have improved endurance characteristics compared to superlattices and solid solution (Hf,Zr)O₂ reference films,^[30] which can, based on the results observed here, be due to the increased single layer thickness of the trilayered devices.

[IMAGE OMITTED. SEE PDF]

Certain applications, for instance in the automotive sector, require appropriate reliability characteristics at elevated operation temperatures. Figure 10 depicts the polarization versus electric field (top row) and current density versus electric field (bottom row) characteristics of the superlattice with 1 nm sublayer thickness measured at chuck temperatures between 75 and 150°C. For a constant bias stress like for example 3.5 MV cm^-1, the leakage increases with rising chuck temperature. This expected trend can be seen qualitatively from the “cigar-like” shape of the polarization loops and quantitatively from the current versus electric field plots. It is worth to point out that even at the most challenging bias and temperature stress conditions (3.5 MV cm^-1 and 150°C) the superlattice with 1 nm sublayers survives 10³ field cycles, which is in strong contrast to the other superlattices and the reference device that break under the same stress condition after a significantly lower number of cycles.

[IMAGE OMITTED. SEE PDF]

Figure 11 summarizes the endurance of the solid solution reference device and the superlattices of different sublayer-thicknesses for chuck temperatures up to 150°C. The high temperature reliability strongly depends on the structure of the ferroelectric material: Starting from the reference film, the temperature reliability significantly improves with increasing sublayer thickness up to 1 nm, which is the most promising film. In the case of thicker sublayers, the endurance characteristics degrade again and reach at 5 nm almost the same results as observed earlier for the reference device.

[IMAGE OMITTED. SEE PDF]

This behavior is rather analogue to another class of superlattices that use Al₂O₃ interlayers between sublayers of Zr-doped HfO₂.^[47] It was reported that Al₂O₃ interlayers might facilitate the formation of smaller grains. The reliability improvement can thus be explained by the lengthened leakage paths. However, we observe grain sizes in the range of the total superlattice thickness. Another explanation for the improved reliability characteristics of the superlattice devices can be the comb like band-structure originating from the different band-gap values of the HfO₂ sublayer (5.3–5.9 eV^[48]) and the ZrO₂ sublayer (3.8–4.5 eV^[49]). As a consequence, the interfaces between the sublayers act as a barrier that suppresses electron and hole transport. Contrary, in the case of a perfect solid solution, an intermediary band-gab is formed.

Conclusion

Herein, [HfO₂/ZrO₂] superlattices with sublayer thicknesses ranging from 0.25 nm to 5 nm and an overall thickness of 10 nm were embedded into MFM capacitors. The electrical and structural characterization revealed that, compared to a pure 10 nm (Hf,Zr)O₂ reference film, superlattice-based devices have higher remanent polarization, allow tailoring of the coercive field, and provide better reliability characteristics, in particular at high operation temperatures. Regarding remanent polarization and high temperature reliability, a sublayer thickness of 1 nm seems to be the most promising. The corresponding superlattice is, thus, very interesting for high temperature applications as found for instance in the broad automotive sector. Another important advantage of the superlattices is the possibility to tailor the coercive field by the sublayer thickness. A sublayer thickness of 2.5 nm provides the smallest coercive field without significantly sacrificing the remanent polarization and the reliability characteristics (both are superior compared to the reference device). This, in turn, makes the corresponding superlattice promising for low power applications.

Acknowledgements

This project has received funding from the ECSEL Joint Undertaking (JU) under grant agreement no. 101007321. The JU received support from the European Union's Horizon 2020 research and innovation programme and France, Belgium, Czech Republic, Germany, Italy, Sweden, Switzerland, and Turkey.

Open access funding enabled and organized by Projekt DEAL.

Correction added on 23 August 2023, after first online publication: Projekt Deal funding statement has been added.

Conflict of Interest

The authors declare no conflict of interest.

Data Availability Statement

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

References

G. Croce, in 2022 IEEE International Memory Workshop (IMW), IEEE, ISBN 978‐1‐6654‐9947‐7, 2022, pp. 1–4.

A. Keshavarzi, W. van den Hoek, IEEE Des. Test 2019, 36, 41.

A. Keshavarzi, K. Ni, W. van den Hoek, S. Datta, A. Raychowdhury, IEEE Micro 2020, 40, 33.

D. S. Campbell, J. Br. Inst. Radio Eng. 1957, 17, 385.

H. Takasu, J. Electroceram. 2000, 4, 327.

T. S. Böscke, J. Müller, D. Bräuhaus, U. Schröder, U. Böttger, Appl. Phys. Lett. 2011, 99, [eLocator: 102903].

K. Mistry, R. Chau, C.‐H. Choi, G. Ding, K. Fischer, T. Ghani, R. Grover, W. Han, D. Hanken, M. Hattendorf, J. He, C. Allen, J. Hicks, R. Huessner, D. Ingerly, P. Jain, R. James, L. Jong, S. Joshi, C. Kenyon, K. Kuhn, K. Lee, C. Auth, H. Liu, J. Maiz, B. McIntyre, P. Moon, J. Neirynck, S. Pae, C. Parker, et al., in 2007 IEEE International Electron Devices Meeting. IEEE, ISBN 978‐1‐4244‐1507‐6, 122007, pp. 247–250.

J. Müller, U. Schröder, T. S. Böscke, I. Müller, U. Böttger, L. Wilde, J. Sundqvist, M. Lemberger, P. Kücher, T. Mikolajick, L. Frey, J. Appl. Phys. 2011, 110, [eLocator: 114113].

T. S. Böscke, St. Teichert, D. Bräuhaus, J. Müller, U. Schröder, U. Böttger, T. Mikolajick, Appl. Phys. Lett. 2011, 99, [eLocator: 112904].

F. Ali, T. Ali, D. Lehninger, A. Sünbül, A. Viegas, R. Sachdeva, A. Abbas, M. Czernohorsky, K. Seidel, Adv. Funct. Mater. 2022, 32, [eLocator: 2201737].

H. Mulaosmanovic, E. T. Breyer, S. Dünkel, S. Beyer, T. Mikolajick, S. Slesazeck, Nanotechnology 2021, 32, 50.

D. Lehninger, H. Mahne, T. Ali, R. Hoffmann, R. Olivo, M. Lederer, K. Mertens, T. Kampfe, K. Biedermann, M. Landwehr, A. Heinig, D. Wang, Y. Shen, K. Bernert, S. Thiem, K. Seidel, in 2022 IEEE International Memory Workshop (IMW), IEEE, ISBN 978‐1‐6654‐9947‐7, 52022, pp. 1–4.

K. Seidel, D. Lehninger, R. Hoffmann, T. Ali, M. Lederer, R. Revello, K. Mertens, K. Biedermann, Y. Shen, D. Wang, M. Landwehr, A. Heinig, T. Kampfe, H. Mahne, K. Bernert, S. Thiem, in 2022 IEEE Symposium on VLSI Technology and Circuits (VLSI Technology and Circuits), IEEE, ISBN 978‐1‐6654‐9772‐5, 6122022, pp. 355–356.

S. Beyer, S. Dunkel, M. Trentzsch, J. Muller, A. Hellmich, D. Utess, J. Paul, D. Kleimaier, J. Pellerin, S. Muller, J. Ocker, A. Benoist, H. Zhou, M. Mennenga, M. Schuster, F. Tassan, M. Noack, A. Pourkeramati, F. Muller, M. Lederer, T. Ali, R. Hoffmann, T. Kampfe, K. Seidel, H. Mulaosmanovic, E. T. Breyer, T. Mikolajick, S. Slesazeck, in IEEE IMW 2020, pp. 1–4.

J. Okuno, T. Kunihiro, K. Konishi, M. Materano, T. Ali, K. Kuehnel, K. Seidel, T. Mikolajick, U. Schroeder, M. Tsukamoto, T. Umebayashi, IEEE J. Electron Devices Soc. 2022, 10, 29.

H. Mulaosmanovic, P. D. Lomenzo, U. Schroeder, S. Slesazeck, T. Mikolajick, B. Max, in 2021 IEEE International Reliability Physics Symposium (IRPS), IEEE, ISBN 978‐1‐7281‐6893‐7, 32021, pp. 1–6.

M. H. Park, Y. H. Lee, H. J. Kim, Y. J. Kim, T. Moon, K. D. Kim, J. Müller, A. Kersch, U. Schroeder, T. Mikolajick, C. S. Hwang, Adv. Mater. 2015, 27, 1811.

D. Lehninger, M. Ellinger, T. Ali, S. Li, K. Mertens, M. Lederer, R. Olivio, T. Kämpfe, N. Hanisch, K. Biedermann, M. Rudolph, V. Brackmann, S. Sanctis, M. P. M. Jank, K. Seidel, Adv. Electron. Mater. 2021, 7, [eLocator: 2100082].

D. Lehninger, R. Olivo, T. Ali, M. Lederer, T. Kämpfe, C. Mart, K. Biedermann, K. Kühnel, L. Roy, M. Kalkani, K. Seidel, Phys. Status Solidi A 2020, 217, [eLocator: 1900840].

D. Lehninger, T. Ali, R. Olivo, M. Lederer, T. Kampfe, K. Mertens, K. Seidel, in 2020 Joint Conference of the IEEE International Frequency Control Symposium and International Symposium on Applications of Ferroelectrics (IFCS‐ISAF), IEEE, ISBN 978‐1‐7281‐6430‐4, 2020, pp. 1–3.

E. Yurchuk, J. Müller, S. Knebel, J. Sundqvist, A. P. Graham, T. Melde, U. Schröder, T. Mikolajick, Thin Solid Films 2013, 533, 88.

M. H. Park, Y. J. Kim, H. J. Kim, T. Moon, K. D. Kim, Lee, in ISAF‐ISIF‐PFM‐2015, 2015.

Y. W. Lu, J. Shieh, F. Y. Tsai, Acta Mater. 2016, 115, 68.

S. L. Weeks, A. Pal, V. K. Narasimhan, K. A. Littau, T. Chiang, ACS Appl. Mater. Interfaces 2017, 9, [eLocator: 13440].

Y. Chen, L. Wang, L. Liu, L. Tang, X. Yuan, H. Chen, K. Zhou, D. Zhang, J. Mater. Sci. 2021, 56, 6064.

Y.‐K. Liang, W.‐L. Li, Y.‐J. Wang, L.‐C. Peng, C.‐C. Lu, H.‐Y. Huang, S. H. Yeong, Y.‐M. Lin, Y.‐H. Chu, E.‐Y. Chang, C.‐H. Lin, IEEE Electron Device Lett. 2022, 43, 1451.

N. Bai, K.‐H. Xue, J. Huang, J.‐H. Yuan, W. Wang, G.‐Q. Mao, L. Zou, S. Yang, H. Lu, H. Sun, X. Miao, Adv. Electron. Mater. 2022, 9, [eLocator: 2200737].

Y. Peng, W. Xiao, Y. Liu, C. Jin, X. Deng, Y. Zhang, F. Liu, Y. Zheng, Y. Cheng, B. Chen, X. Yu, Y. Hao, G. Han, IEEE Electron Device Lett. 2022, 43, 216.

J. Y. Park, D. H. Lee, K. Yang, S. H. Kim, G. T. Yu, G. H. Park, E. B. Lee, K. H. Kim, M. H. Park, ACS Appl. Electron. Mater. 2022, 4, 1369.

M. H. Park, H. J. Kim, G. Lee, J. Park, Y. H. Lee, Y. J. Kim, T. Moon, K. D. Kim, S. D. Hyun, H. W. Park, H. J. Chang, J.‐H. Choi, C. S. Hwang, Appl. Phys. Rev. 2019, 6, [eLocator: 041403].

K. Ni, J. Smith, H. Ye, B. Grisafe, G. B. Rayner, A. Kummel, S. Datta, in 2019 IEEE International Electron Devices Meeting (IEDM), IEEE, ISBN 978‐1‐7281‐4032‐2, 07.12.2019‐11.12.2019, 28.8.1–28.8.4.

S. Cheema, N. Shanker, L.‐C. Wang, C.‐H. Hsu, S.‐L. Hsu, Y.‐H. Liao, M. S. Jose, J. Gomez, W. Li, J.‐H. Bae, S. Volkman, D. Kwon, Y. Rho, G. Pinelli, R. Rastogi, D. Pipitone, C. Stull, M. Cook, B. Tyrrell, V. Stoica, Z. Zhang, J. Freeland, C. Tassone, A. Mehta, G. Soheli, D. Thompson, D. I. Suh, W.‐T. Koo, K.‐J. Nam, D. J. Jung, et al., 2021, [DOI: https://dx.doi.org/10.21203/rs.3.rs-413053/v1].

S. S. Cheema, N. Shanker, L.‐C. Wang, C.‐H. Hsu, S.‐L. Hsu, Y.‐H. Liao, M. San Jose, J. Gomez, W. Chakraborty, W. Li, J.‐H. Bae, S. K. Volkman, D. Kwon, Y. Rho, G. Pinelli, R. Rastogi, D. Pipitone, C. Stull, M. Cook, B. Tyrrell, V. A. Stoica, Z. Zhang, J. W. Freeland, C. J. Tassone, A. Mehta, G. Saheli, D. Thompson, D. I. Suh, W.‐T. Koo, K.‐J. Nam, et al., Nature 2022, 604, 65.

D. Lehninger, M. Lederer, T. Ali, T. Kampfe, K. Mertens, K. Seidel, in IEEE, editor, 2021 IEEE International Interconnect Technology Conference (IITC), IEEE, ISBN 978‐1‐7281‐7632‐1, 2021, pp. 1–4.

D. Lehninger, R. Hoffmann, A. Sunbul, H. Mahne, T. Kampfe, K. Bernert, S. Thiem, K. Seidel, IEEE Electron Device Lett. 2022, 43, 1866.

J. Okuno, T. Kunihiro, K. Konishi, H. Maemura, Y. Shuto, F. Sugaya, M. Materano, T. Ali, M. Lederer, K. Kuehnel, K. Seidel, U. Schroeder, T. Mikolajick, M. Tsukamoto, T. Umebayashi, in 2021 IEEE International Memory Workshop (IMW), IEEE, ISBN 978‐1‐7281‐8517‐0, 2021, pp. 1–3.

M. Lederer, P. Bagul, D. Lehninger, K. Mertens, A. Reck, R. Olivo, T. Kämpfe, K. Seidel, L. M. Eng, ACS Appl. Electron. Mater. 2021, 3, 4115.

M. Lederer, R. Olivo, D. Lehninger, S. Abdulazhanov, T. Kämpfe, S. Kirbach, C. Mart, K. Seidel, L. M. Eng, Phys. Status Solidi RRL 2021.

S. E. Reyes‐Lillo, K. F. Garrity, K. M. Rabe, Phys. Rev. B 2014, 90, 14.

M. Neuber, M. W. Lederer, K. Mertens, T. Kämpfe, M. Czernohorsky, K. Seidel, Crystals 2022, 12, 1115.

S. L. Miller, P. J. McWhorter, J. Appl. Phys. 1992, 72, 5999.

T. Ali, D. Lehninger, M. Lederer, S. Li, K. Kühnel, C. Mart, K. Mertens, R. Hoffmann, R. Olivo, J. Emara, K. Biedermann, J. Metzger, R. Binder, M. Czernohorsky, T. Kämpfe, J. Müller, K. Seidel, L. M. Eng, Adv. Electron. Mater. 2022, 8, [eLocator: 2100837].

Y. Zhang, D. Wang, C. Luo, J. Cheng, S. Huo, B. Zhang, R. Tao, D. Chen, Z. Fan, J.‐Y. Dai, X. Lu, J.‐M. Liu, IEEE Trans. Electron Devices 2022, 69, 3094.

M. Pešić, F. P. G. Fengler, L. Larcher, A. Padovani, T. Schenk, E. D. Grimley, X. Sang, J. M. LeBeau, S. Slesazeck, U. Schroeder, T. Mikolajick, Adv. Funct. Mater. 2016, 26, 4601.

S. J. Kim, J. Mohan, H. S. Kim, J. Lee, C. D. Young, L. Colombo, S. R. Summerfelt, T. San, J. Kim, Appl. Phys. Lett. 2018, 113, [eLocator: 182903].

P. Jiang, Q. Luo, X. Xu, T. Gong, P. Yuan, Y. Wang, Z. Gao, W. Wei, L. Tai, H. Lv, Adv. Electron. Mater. 7, 2000728.

M. Lederer, K. Seidel, R. Olivo, T. Kämpfe, L. M. Eng, Front. Nanotechnol. 2022, 4, [eLocator: 900379].

J. H. Choi, Y. Mao, J. P. Chang, Mater. Sci. Eng., R 2011, 72, 97.

F. Zandiehnadem, R. A. Murray, W. Y. Ching, Physica B+C 1988, 150, 19.