Tailoring the metal-oxide interface for improving fatigue performance in HfO2-based ferroelectrics

Tailoring the metal-oxide interface for improving fatigue performance in HfO2-based ferroelectrics | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Tailoring the metal-oxide interface for improving fatigue performance in HfO 2 -based ferroelectrics Yi-Xuan Liu, Ziyi Yuan, Babak Bakhit, Han Yu, Jiahao Lu, Ji Soo Kim, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7907164/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Hafnia-based ferroelectrics have attracted significant interest for their scalability and compatibility with complementary metal-oxide-semiconductor (CMOS) technology for memory applications. However, the imprint and fatigue behaviour remain critical challenges, particularly in electrode-integrated devices. By investigating Y-doped HfO 2 films with various top electrodes (Pt, Cu, W, Mo, Ni, and Al), this study reveals that the fatigue performance is determined by the interplay between the metal work function and chemical stability through extended Frenkel pairs formation at the interface. Moreover, YHO with Ni top electrodes undergoes a unique ferroelectric failure from ferroelectric switching to filamentary conduction under electrical cycling, driven by extended Frenkel pair defects redistribution. Electrodes with high extended Frenkel pair formation energy enable YHO capacitors to maintain nearly unchanged ferroelectric polarization after 10 8 switching cycles. Our findings show that charged defect formation, rather than intrinsic work function, determines interfacial band bending and the Schottky barrier. This work highlights the importance of interface engineering for enhancing the fatigue performance and the reliability of HfO 2 -based ferroelectrics for memory applications. Physical sciences/Materials science/Materials for devices/Electronic devices Physical sciences/Materials science/Materials for devices/Information storage Ferroelectrics Hafnia Band alignment Fatigue Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Nearly fifteen years after the discovery of ferroelectricity in doped hafnium oxide (HfO 2 ), this material has emerged as a leading candidate for advanced nanoelectronic applications, owing to its compatibility with complementary metal-oxide-semiconductor (CMOS) technology. 1-3 In contrast to perovskite ferroelectrics (e.g., BaTiO 3 , PZT) that lose their polarisation below a critical thickness, ferroelectric HfO 2 defies the conventional size effect by maintaining and even enhancing polarisation at the nanometer scale, making it attractive for next-generation non-volatile memory devices. 4-7 Recent advances, including chemical doping, solid‑solution formation such as Hf 1- x Zr x O 2 (HZO), strain engineering, grain size controlling and texturing, have further expanded the application potential of HfO 2 ‑based ferroelectric thin films. 8-12 In particular, yttrium-doped HfO 2 (YHO) has attracted attention for its robust ferroelectricity with high remanent polarisation. 13 Although a substantial increase in remanent polarisation with reduced coercive field enables low-power operation, the long-term reliability of HfO 2 -based ferroelectrics, particularly their fatigue under electrical cycling, remains underexplored. Acceptor-doped HfO 2 films intrinsically host a pronounced concentration of oxygen vacancies. 14 Under electrical cycling, these defects accumulate and then either destabilise the polar phase or pin the domain walls, both effects causing ferroelectric fatigue over time. 15 Additional contributions arise from interfacial charge injection and trapping, leading to the formation of passive interfacial layers or field-induced phase changes. 16 These processes occur primarily at the electrode/oxide interface. In epitaxial films, the bottom electrode imposes strain on the oxide and participates in interfacial redox reactions, thereby affecting phase stability and the distribution of oxygen vacancies. 17-19 In contrast, the top electrode acts as a capping layer, inducing residual stress that stabilises the ferroelectric phase. 20 Through oxygen exchange and interfacial reactions with the oxide, the top electrode also determines the interfacial band alignment and the charge-injection barrier during switching. Early studies employed noble metals such as Pt or Au to suppress interfacial reactions, enabling the demonstration of the intrinsic ferroelectric response of HfO 2 -based films. 21 However, these electrodes are incompatible with CMOS integration and fail to represent the interfacial chemistry of practical device metals, which can promote defect formation and carrier injection, thereby accelerating fatigue. Moreover, the oxygen exchange at metal/HfO 2 interfaces is not solely determined by the vacancy formation energy of the metal oxide (MeO x ) but can be strongly influenced by interfacial reactions. One important process is the formation of extended Frenkel pairs (EFPs), where the removed oxygen from an oxide film is incorporated into the metal while the accompanying charge is transferred from the vacancy energy level of the induced vacancies to electrodes. 22 First-principles calculations show that at a Mo/HfO 2 interface the vacancy formation energy can decrease from ~6.3 eV in the bulk to ~1.4 eV when in contact with Mo. A similar effect occurs for other metal electrodes also. 23 Such EFP defects can create a charged dipole layer that can locally bend energy bands and modify the effective barrier at the interface, thus enhancing carrier injection and polarisation degradation during switching. 24 Therefore, the selection of metal electrodes becomes essential for understanding and improving the fatigue performance of practical HfO 2 -based ferroelectric devices. In the present study, we examine the impact of top electrode choice on the fatigue behaviour of 7% Y-doped HfO 2 ferroelectric films. To exclude microstructural variability inherent to polycrystalline films, epitaxial YHO layers were grown on La 0.7 Sr 0.3 MnO 3 -buffered SrTiO 3 substrates, yielding a well-defined distorted orthorhombic ferroelectric phase. A set of electrode metals (Cu, W, Mo, Ni, and Al) with similar work functions but different oxygen vacancy formation energy 25 , or vice versa, was selected to compare with the Pt electrode. All electrodes were deposited at room temperature to suppress extrinsic capping mechanical stress. Comprehensive electrical tests and Kelvin probe force microscopy (KPFM) were performed to evaluate polarisation fatigue and interface potential evolution for different electrodes. It was found that the selection of top metal electrodes can affect the polarisation value and coercive field before fatigue. Notably, these devices undergo a transition from ferroelectric switching to a mixed ferroelectric-resistive response, accompanied by enhanced leakage and resistive switching. Overall, our findings suggest a foundation for engineering metal-oxide interfaces to enhance the endurance of hafnium oxide ferroelectric memories. Orthorhombic polar phase with rhombohedral distortion The device structure is demonstrated in Fig. 1 a, where (111)-oriented YHO thin films were epitaxially grown on La 0.67 Sr 0.33 MnO₃ (LSMO)/SrTiO 3 (STO) substrates with various metal top electrodes (Pt, Cu, W, Mo, Ni, Al) patterned on the YHO surface. The 25 nm-thick LSMO bottom electrode was epitaxially grown on an STO (001) substrate along the (001) direction layer by layer, as confirmed by the periodic oscillations of the intensity from the reflection high-energy electron diffraction monitoring ( fig. S1 ). The X-ray diffraction (XRD) 2θ-ω pattern (Fig. 1 b) confirms that the YHO film is (111)-oriented single-phase structure. The primary YHO (111) reflection appears at the 2 θ = ~29.8°, alongside STO (001) and LSMO (001) peaks, with the well-defined Laue oscillations, reflecting smooth film interfaces, high crystallinity, and uniform thickness. The (111) peak of YHO remains unchanged with different top electrode materials, indicating that the ex-situ sputtering of metal electrodes has a negligible impact on the phase structure of YHO. The film thickness is estimated to be ~ 8 nm from the full width at half maximum (FWHM) of the (111) peak. Note that the 2 θ value of (111) peak (at 2 θ = ~29.8°) is slightly lower than that in the commonly reported polar o-phase in HfO 2 (at 2 θ = ~30.5°). 13 This suggests an expanded (111)-spacing ( d 111 ) along the out-of-plane direction, where the high structural order arises from the lattice mismatch and is consistent with the orthorhombic Pca 2 1 phase with a slight rhombohedral distortion. 26 The high crystalline quality of the YHO film is evidenced by the sharp central peak in the ω -rocking curve of the (111) reflection, confirming that most of the YHO lattice is well aligned with the substrate. To further confirm the symmetry of the YHO lattice, pole figure measurements were performed around 2 θ = 29.8°, i.e., around the YHO {111} reflection. For (111)-oriented single-domain film, three reflections, ( \(\:\stackrel{-}{1}\) 11), ( \(\:1\stackrel{-}{1}\) 1), and (1 \(\:1\stackrel{-}{1}\) ), are expected at χ ~ 71° from the out-of-plane direction, with azimuthal angles φ differing by 120°. The pole figure at 2 θ = 29.8° (Fig. 1 d) reveals 12 reflections, indicating the presence of four distinct in-plane domain variants. As shown in Fig. 1 e, these crystal domains, rotated by 90° relative to each other, align with the four-fold symmetry of the cubic STO substrate. As shown in Fig. 1 f, the 12 peaks with in-plane components of the scattering vector at χ ≈ 71° share the same 2 θ = 30.5°, suggesting an equivalent d -spacing of 2.93Å that is smaller than that of the out-of-plane (111) reflection ( d 111 ~ 2.99 Å). Therefore, these combined characterisations verify that the orthorhombic unit cells in YHO films are elongated along the (111) polar axis. The presence of ferroelectricity in YHO films can be verified by piezo-response force microscopy (PFM). After poling the film with ± 8 V, PFM phase images exhibit 180° contrast between oppositely poled domains both along out-of-plane (Fig. 1 g) and in-plane (Fig. 1 h) direction. This indicates the presence of two reversibly switchable polarisation states. Switching spectroscopy-PFM (SS-PFM) further displays a sharp phase inversion and a butterfly-shaped amplitude loop as a function of bias, determining a local coercive voltage V c of 2.7 V and a phase-loop saturation voltage of comparable magnitude. 27 This nanoscale response implies the reversible ferroelectric switching in YHO films. Schottky contact between YHO and top electrode metals In device structures, the electrode work function is often modified by interfacial reactions, charge transfer, and dipole formation, leading to an effective work function that deviates from the intrinsic value. Therefore, Kelvin probe force microscopy (KPFM) was employed to determine the contact potential difference (Fig. 2 a, fig. S2a and S2b ) between the metal electrodes and the YHO surface. Atomic force microscopy (AFM) confirmed a smooth surface morphology with uniform electrode thickness of ~ 70 nm ( fig. S2c ), ensuring that topography-induced artefacts were negligible in the KPFM measurements. Figure 2 d-h show spatial mapping of contact potential differences between the metal and YHO (Δ P Me−YHO ), with colour contrast corresponding to potential magnitude. Gaussian fitting of the potential difference distribution ( fig. S3a-e ) yields Δ P Me−YHO of 260 mV, 60 mV, 240 mV, 140 mV, and 435 mV for Pt, Cu, W, Mo, and Ni electrodes, respectively ( fig. S2f ). It is noted that the measured potential differences between the metals and YHO, Δ P Me−YHO , do not exhibit the anticipated positive correlation with metal work function (orange curve and cyan curve in Fig. 2 i). Specifically, Cu, W, and Mo, despite similar work functions, show distinctly different Δ P Me−YHO . Moreover, Ni has a considerably higher Δ P Ni−YHO ~ 435 mV than that of Δ P Pt−YHO ~ 260 mV, despite its lower work function (5.16 eV for Ni compared to 5.65 eV for Pt). The inconsistency between the measured contact potentials and the nominal metal work functions suggests the presence of interfacial effects at the metal/YHO interface. Stronger interfacial interactions generally enhance the deviation between the measured contact potential and the intrinsic work function, as interface dipoles introduce additional potential drops that shift the band alignment beyond the prediction. 29 Considering the heat of metal oxides formation per oxygen atom, certain metal such as Ni, are not expected to form stable oxides, yet they show a substantial deviation in their measured contact potential. This deviation might suggest additional mechanisms, where a key process is the formation of EFP defects. Interfacial oxygen exchange occurs when the YHO is contacted by a metal with a lower oxide formation energy than Hf, expressed as $$\:{\text{O}}_{\text{O}}+\frac{1}{n}\text{M}\text{e}={\text{V}}_{\text{O}}^{{\bullet\:}{\bullet\:}}+\frac{1}{n}\text{M}\text{e}{\text{O}}_{n}+2{e}^{}$$ where \(\:{\text{O}}_{\text{O}}\) represents the lattice oxygen in YHO that diffuses into the metal electrode. Me denotes the metal electrode and \(\:{V}_{\text{O}}^{{\bullet\:}{\bullet\:}}\) is the positively charged oxygen vacancy left in YHO. MeO n refers to the metal oxide formed at the interface. This reaction describes the migration of oxygen atoms from YHO to the metal, leading to oxidation of the electrode. Initially, these oxygen vacancies in YHO are electrically neutral (referred to as \(\:{V}_{\text{O}}\) ), containing two localized electrons at the vacancy sites with an energy level of ~ 2.5 eV above the valence band of YHO (Fig. 2 b). 28 When contacted by a high-work-function metal, of which the Fermi level lies below that of YHO, electrons from neutral \(\:{V}_{\text{O}}\) are transferred to the metal, leaving positively charged \(\:{V}_{\text{O}}^{{\bullet\:}{\bullet\:}}\) , i.e., EFP defects, between the film and metal electrode, generating an interfacial dipole layer adjacent to the metal electrode (Fig. 2 c). The accumulation of \(\:{V}_{\text{O}}^{{\bullet\:}{\bullet\:}}\) significantly modifies the Schottky barrier height, requiring an effective metal work function, \(\:{}_{\text{M}\text{e}}^{\text{e}\text{f}\text{f}}\) , to describe the interfacial band alignment. The formation energy of \(\:{\text{V}}_{\text{O}}^{{\bullet\:}{\bullet\:}}\) can be quantified as $$\:{E}_{EFP}={E}_{\text{f}\text{o}\text{r}\text{m}}+\:{E}_{\text{M}\text{O}}\left(\frac{1}{n}\text{M}\text{e}{\text{O}}_{n}\right)+2\left({E}_{{\text{V}}_{\text{O}}}-{E}_{f,\:\:\text{M}\text{e}}\right)+{E}_{electro}$$ where E form denotes the formation energy of a neutral vacancy relative to molecular O 2 , E MO represents the formation free energy of the metal oxide per oxygen atom. The term, \(\:2\left({E}_{{\text{V}}_{\text{O}}}-{E}_{f,\:\:\text{M}\text{e}}\right)\) , accounts for the energy gained from two electrons falling from \(\:{\text{V}}_{\text{O}}\) energy level ( \(\:{E}_{{\text{V}}_{\text{O}}}\) ) into the metal Fermi level ( \(\:{E}_{f,\:\:\text{M}\text{e}}\) ). E electro denotes the electrostatic contribution of EFP defects to the total energy. 25 For different metal electrodes, only the second and third terms in the formation energy expression vary, which oppose each other. Metals with lower oxide formation energies, or with higher work functions that facilitate electron transfer from vacancies, have reduced EFP formation energy ( E EFP ). According to a theoretical study 22 , Ni exhibits a particularly low formation energy of \(\:{V}_{\text{O}}^{{\bullet\:}{\bullet\:}}\) ​ ( E EFP ~1.47 eV) compared to Cu (3.35 eV), W (2.65 eV), and Mo (2.07 eV), leading to enhanced vacancy and dipole accumulation. As a result, the measured contact potentials, ΔP Me−YHO , show a strong negative correlation with E EFP (with a Pearson coefficient ρ ~-0.988) rather than solely with their intrinsic work functions or oxide formation energies. Pt, on the other hand, has high work function (5.65 eV) like Ni, yet its inert nature prevents it from EFP defect formation. Ferroelectricity and fatigue performance of YHO with different top metal electrodes The top electrode material is found to significantly influence the ferroelectric performance of YHO films. Positive-up negative-down (PUND) measurements showed a clear ferroelectric switching peak in current-voltage loops at switching fields of 3.7-5 MV/cm for Pt, Cu, W, Mo, and Ni top electrodes (Fig. 3 a and fig. S4 ). Notably, the higher E C (21.75 MV/cm) of the YHO-Al capacitor may be due to the voltage drop across the interfacial alumina layer, as Al is more reactive than Hf ( fig. S4f ). Note that alumina formation is ineluctable from being exposed to the atmosphere and/or being in contact with YHO. The functionality of the alumina layer is outside the scope of this work. The remanent polarisation (2 P r ) of YHO films with Pt, Cu, W, Mo, and Ni electrodes ranges from 14 to 18 µC/cm 2 ( fig. S4a-e ). Among these, YHO-Pt exhibits the highest 2 P r ~ 18 µC/cm 2 , whereas YHO-Ni shows a reduced 2 P r ~ 14 µC/cm 2 and a markedly higher coercive voltage 2 V C of 8.4 V ( E C ~ 4.8 MV/cm), despite having a comparable oxide formation energy to W ( E C ~ 3.7 MV/cm). These deviations cannot be rationalised by work function or E MO alone but instead correlate with EFP defects at the metal/YHO interface. As the EFP defect formation energy decreases, the accumulation of charged vacancies intensifies and thereby results in suppressed P r and an increased V C (Fig. 3 b). Furthermore, polarisation values at zero voltage differ for the positive and negative electric field ( ΔP E =0 ), indicating imprint effect due to the built-in field. This asymmetry corresponds to the difference in work function between the interfacial oxide MeO x layer at the top contact and the LSMO bottom electrode. 31 Considering the work function of metal oxides is correlated with the oxidation states of their cations, the work functions of oxides with average cation oxidation states in the literature are used for illustrative comparison (Table S1 ). 32 As shown in Fig. 3 c, ΔP E =0 varies in proportion to the work function of interfacial oxide MeO x . In particular, the YHO-W exhibits the smallest difference between φ MeO x and the LSMO bottom electrode ( Δφ , see below Fig. 3 c), and accordingly, the minimal polarisation offset. These results indicate that the interfacial MeO x layer can partially mitigate the intrinsic asymmetry between the top and bottom electrodes, minimising the built-in field within YHO capacitors and consequently reducing the imprint effect. The ferroelectric fatigue behaviour of YHO films was measured using an electric switching cycle E s ~6.25 MV/cm at a field amplitude of approximately 1.5 E c . For most YHO capacitors, the fatigue behaviour evolves in two distinct stages. Initially, the P r decreases monotonically with electrical cycling, reflecting typical ferroelectric fatigue. In some devices, an increase in ± P r appears beyond a critical cycle number, which can be attributed to the sharp rise in leakage current after ferroelectric failure. This gradual ferroelectric failure differs from dielectric breakdown that can be characterized by a linear I-V relationship. 33 The choice of metal electrode strongly influences both polarisation decay and failure onset. YHO-Pt devices exhibit exceptional fatigue resistance, showing nearly unchanged polarisation even after 10 8 cycles, consistent with the high chemical stability of Pt. YHO–Cu devices also maintain good fatigue resistance, showing only a small polarisation reduction (~ 7%). In contrast, YHO devices with W, Mo, and Ni electrodes exhibit more pronounced ferroelectric degradation, with reductions of 10%, 18%, and 26%, respectively, implying their inferior fatigue resistance. The YHO-Al device shows a distinct fatigue behaviour, with an initial gradual decay followed by a sharp polarisation drop after ~ 10 7 cycles. The selection of metal electrodes also influences the critical cycling threshold for ferroelectric failure. The ferroelectric failure occurs after approximately 10 7 -10 8 cycles for YHO with Cu electrodes, whereas W, Mo, and Ni electrodes lead to failure at lower cycles of 1×10 7 , 5×10 6 , and 2×10 6 cycles, respectively. The failure behaviour of Ni-electrode devices is particularly notable, as their failure cycles are significantly lower than those of Pt-electrode devices despite their similarly high work functions. This suggests that factors beyond metal work function, such as chemical reactivity and defect dynamics, may play crucial roles in the ferroelectric stability of YHO devices. Fatigue and ferroelectric failure mechanism The metal/YHO interface plays a critical role in determining the ferroelectric and fatigue behaviour of YHO devices, as defect chemistry within the film and charge injection during electrical cycling act primarily at the interface. In particular, the effective work function together with the \(\:{V}_{\text{O}}^{{\bullet\:}{\bullet\:}}\) formed at the interface with the electrode jointly determine the band alignment. Figure 4 a-d illustrate the band alignment for YHO contacting with different metal electrodes. The metals with lower work functions, such as Cu, W, and Mo, promote electron injection into YHO under an applied field (Fig. 4 a and 4 b). Meanwhile, their lower E EFP facilitates oxygen ion capture from YHO, leaving behind charged \(\:{V}_{\text{O}}^{{\bullet\:}{\bullet\:}}\) near the interface. During electrical cycling, additional oxygen vacancies can be generated at the metal/YHO interface, increasing the concentration of mobile \(\:{V}_{\text{O}}^{{\bullet\:}{\bullet\:}}\) within the film. 34 Compared with perovskite ferroelectric materials (typical E C ~10–100 kV/cm), the high E C of HfO 2 -based ferroelectrics, on the order of MV/cm, is sufficient to drive these \(\:{V}_{\text{O}}^{{\bullet\:}{\bullet\:}}\) across the thin films. An appropriate \(\:{V}_{\text{O}}^{{\bullet\:}{\bullet\:}}\) concentration can stabilize the ferroelectric orthorhombic phase, whereas excessive accumulation destabilizes the polar structure. 18 , 35 As a result, in YHO devices prone to EFP defect formation, the facilitated accumulation of interfacial \(\:{V}_{\text{O}}^{{\bullet\:}{\bullet\:}}\) under cycling accelerates fatigue, consistent with previous reports that electron injection and \(\:{V}_{\text{O}}^{{\bullet\:}{\bullet\:}}\) migration into HfO 2 -based films lead to domain pinning and to non-polar phases. 36 , 37 Among these electrodes, Ni induces the most severe fatigue in YHO devices, despite having a high work function comparable to Pt and an oxide formation energy akin to W, consistent with its low EFP formation energy discussed above. These vacancies at the interface pin the local Fermi level, causing stronger band bending and a lower Schottky barrier (Fig. 4 c). Consequently, both pronounced \(\:{\text{V}}_{\text{O}}^{{\bullet\:}{\bullet\:}}\) and electron injection accelerates ferroelectric degradation. By contrast, a distinct and stable alumina layer may form at the interface between the Al electrode and YHO film, as Al possesses a lower oxide formation energy than Hf. 38 This passivation interfacial layer would suppress charge transfer and oxygen exchange across the interface, but results in a high coercive field and dielectric breakdown under high electric fields. 39 Depth-resolved XPS measurements were performed to obtain the oxygen concentration inside YHO capped with Ni. Figure 4 d shows that this concentration increases by ~ 7% from the Ni/YHO interface towards the YHO/LSMO interface for the pristine device, supporting the significant oxygen scavenging effect of Ni. To elucidate the ferroelectric failure mechanism, the leakage current characteristics were examined in YHO-Ni devices with different electrode areas (314, 490, and 1960 µm 2 ), which are most prone to failure. Before fatigue cycling, the leakage current scales linearly with the tested electrode area (Fig. 4 e and fig. S5a-c ). The electrode-area-dependent scaling indicates an interfacial conduction mechanism dominated by charged carriers, e. g., \(\:{V}_{\text{O}}^{{\bullet\:}{\bullet\:}}\) (Fig. 4 g). After fatigue failure, the leakage characteristics change markedly, showing hysteresis and resistive switching that are independent of electrode area (Fig. 4 f and fig. S5d-f ). This indicates a transition to filamentary conduction, where repeated cycling drives mobile interfacial \(\:{V}_{\text{O}}^{{\bullet\:}{\bullet\:}}\) to nucleate and merge into continuous conductive paths (Fig. 4 h). This filamentary mechanism is consistent with observations in fatigued HfO 2 -based capacitors, where oxygen vacancies congregate near the electrode and eventually form conductive paths. 40 Notably, a lower E EFP (e. g., Ni) promotes the accumulation of \(\:{V}_{O}^{\bullet\:\bullet\:}\) at the interface, thereby facilitating filament formation under switching cycling (inset of Fig. 4 i). After failure, the leakage current increases by nearly three orders of magnitude, consistent with the increased concentration of charged \(\:{V}_{O}^{\bullet\:\bullet\:}\) in the YHO films, proved by the XPS-determined oxygen concentration after failure, shown in the inset in Fig. 4 d. Consequently, the conduction transition from interfacial to a filamentary mechanism emerges earlier in YHO-Ni and YHO-Mo devices, as reflected by their reduced critical failure cycles (Fig. 4 i). In contrast, ferroelectric failure hardly occurs in YHO-Pt devices, which can be attributed to the high chemical stability of Pt and the limited \(\:{V}_{\text{O}}^{{\bullet\:}{\bullet\:}}\) contributing to filament formation. Conclusions Electrode materials play a critical role in the endurance of HfO 2 -based ferroelectric devices. Low-work-function materials accelerate polarisation degradation through enhanced electron injection and oxygen vacancy generation in the ferroelectric film. However, Ni, despite its high work function, exhibits the most severe fatigue with polarisation degradation of ~ 26% before 10 7 cycling owing to an unusually low extended Frenkel Pair formation energy, which promotes charged vacancy accumulation at the interface and consequently interfacial band bending. At fatigue failure, YHO capacitors undergo a transition from ferroelectric switching to filamentary conduction, evidenced by hysteretic leakage currents and a change from area-dependent to area-independent conduction. The high extended Frenkel pair formation energy of Pt suppresses charged oxygen vacancy formation, thereby enhancing the fatigue resistance and avoiding ferroelectric failure after switching cycles. These findings demonstrate that interfacial charged defects, whose formation is jointly controlled by the metal work function and the oxide formation energy of the metal electrode, control the effective barrier height and fatigue behaviour, providing a pathway to improve endurance and long-term stability in ferroelectric memories. Methods Film growth and device fabrication of YHO with various metal electrodes Y-doped HfO 2 (YHO) films were grown by pulsed laser deposition (PLD). The YHO target (Hf 0.93 Y 0.07 O 2 ) was synthesised by weighing Y 2 O 3 and HfO 2 powders according to the stoichiometric ratio. Then, the powders were mixed in a ball miller for 12 h, followed by drying, pelletizing, and sintering at 1400°C for 8 h. A La 0.7 Sr 0.3 MnO 3 (LSMO) buffer/seed target was prepared using La 2 O 3 , SrCO 3 , and MnO 2 , with the carbonate precursors first calcined above 850°C and subsequently sintered at 1200°C. The deposition process began with the growth of 25 nm-thick LSMO films on (100)-oriented Nb-doped (0.5 wt.%) TiO 2 -terminated SrTiO 3 (Nb: STO) single-crystal substrates. The growth conditions were set at 750°C, 100 mTorr oxygen partial pressure (p O₂ ), 0.7 J/cm² laser fluence, and a 2 Hz laser frequency. The film thickness was monitored in situ using reflection high-energy electron diffraction (RHEED). YHO films were then deposited under adjusted conditions, with the substrate temperature increased to 890°C, p O₂ set to 75 mTorr, and the laser fluence maintained at 0.7 J/cm². After deposition, p O₂ was raised to 300 Torr, and the films were held at 890°C for 20 min before being cooled to room temperature at 5 K/min. The YHO film thickness was calibrated based on the growth rate determined from reference samples. Top electrodes (Pt, Cu, W, Mo, Ni, and Al) were deposited via magnetron sputtering following a standard lift-off process. AZ 4533 UV photoresist layer was spin-coated onto the sample surface at 8000 rpm for 30 s and soft-baked at ~ 100°C for 2 min. The coated samples were then exposed to UV light for 10 s through a photomask and developed using AZ 351B developer. Then the top metal electrodes were sputter-deposited onto the patterned resist. The lift-off process was performed by submerging the samples in acetone for 5 min, followed by sonication to remove the unexposed resist and lift off the excess metal. Scanning probe microscopy and X-ray diffraction (XRD) characterisation The surface topography and piezoresponse of the heterostructures were characterised using a Bruker Multimode atomic force microscope with Pt-coated NSC35 tips (MikroMasch). A DC voltage was applied to the sample via the LSMO electrode while grounding the PFM tip, and the piezoresponse was detected using an AC excitation at 15 kHz in the same configuration. The sweep rate was set at 1 Hz. Kelvin Probe Force Microscopy (KPFM) was performed using a Bruker Dimension Icon Pro and with a PtIr-coated Si tip, with a resonance frequency of ~ 70 kHz. A DC bias was applied to the probe to match the contact potential difference between the probe and the sample surface. X-ray diffraction measurements, including 2𝜃-𝜔 scans, 𝜔 scans, and 𝜑 scans, were carried out using a Panalytical Empyrean Diffractometer with parallel beam optics and CuKα₁ radiation. The system was equipped with either a single-point proportional detector or a PIXcel3D position-sensitive detector. 2𝜃–𝜔 scans were performed in scanning line detector mode to achieve high resolution and enhance d-spacing intensity information, while pole figure measurements were conducted in point detector mode, optimized for high scan speed. Electrical measurements Ferroelectric polarisation loops were measured at room temperature using an AixACCT TFAnalyzer 2000. A bias voltage was applied to the top electrode (diameter: 20–50 µm), while the LSMO bottom electrode was grounded. Polarisation hysteresis loops were obtained via positive-up negative-down (PUND) measurements, integrating the current response at 1 kHz. Leakage current was measured using a linear voltage sweep mode with an auto-ranging current detection system to ensure high-resolution measurements. Fatigue performance was assessed by applying a triangular voltage wave with an amplitude of 5 V and a frequency of 100 kHz. PUND measurements were performed at logarithmic cycle intervals to monitor the evolution of remanent polarisation. X-ray Photoelectron Spectroscopy (XPS) XPS core levels were determined using a Thermo Scientific Escalab 250Xi instrument operated with monochromatic Al Kα radiation (hν = 1486.6 eV) at ~ 1 × 10 − 9 Torr. The 10-nm-thick Ni top electrodes were kept in electrical contact with the XPS sample holder to prevent possible peak shifts arising from sample charging. Depth-resolved XPS core levels were acquired from 500 × 500 µm 2 regions located in the centre of 1 × 1 mm 2 sputter-etched areas after gentle surface sputter-etching by 400-eV Ar ions. Elemental depth-resolved quantifications were carried out using CasaXPS software. Declarations Competing interests: The authors declare no competing interests. Author contributions Y.-X. L., J. S. K., and J. L. M.-D. conceived this work, led the investigation, and developed the methodology. Y.-X. L. deposited thin film, performed and analysed electrical measurements, and contributed to all other analyses. J. S. K. fabricated devices and conducted XRD measurements with J. L. Z. Y. performed KPFM measurements. J. S. K. conducted PFM measurements and H. Y. assisted in fitting and performed SS-PFM characterisations with the support of K. W. XPS measurements were performed by B. B., who analysed the XPS data. J. P. B. S. and K. W. acquired funding. J.L.M.-D. acquired funding, provided overall project supervision measurements. Y.-X. L. drafted the first version of the manuscript. All authors contributed to reviewing and editing the manuscript, and they approved the final version. Acknowledgements We gratefully acknowledge discussions with Prof. Uwe Schroeder from NaMLab gGmbH and his valuable advice. This work was supported by the ERC grant EU-H2020-ERC-ADG 882929, EROS. The authors also acknowledge funding by National Nature Science Foundation of China (Nos. 52325204 and 52302148), the Royal Academy of Engineering Chair in Emerging Technologies grant (Grant No. CIET1819-24), and the Portuguese Foundation for Science and Technology (FCT) in the framework of the Strategic Funding UID/04650/2025. J. P. B. S. also thanks FCT for the contract under the FCT Tenure program 1st edition. Data availability The data that support the findings of this study are provided in the article and the Supplementary Information. The data sets generated and analysed during the current study are available from the corresponding author on reasonable request. References Böscke T, Müller J, Braeuhaus D, Schröder U, Böttger U (2011) Ferroelectricity in hafnium oxide thin films. Appl Phys Lett 99 Schroeder U, Park MH, Mikolajick T, Hwang CS (2022) The fundamentals and applications of ferroelectric HfO 2 . Nat Reviews Mater 7:653–669 Chen H et al (2022) HfO 2 -based ferroelectrics: From enhancing performance, material design, to applications. Appl Phys Reviews 9 De La Pérez J, Joanni E, Vilarinho P, Kholkin A (2010) Thickness effect on the dielectric, ferroelectric, and piezoelectric properties of ferroelectric lead zirconate titanate thin films. J Appl Phys 108 Kang S et al (2022) Highly enhanced ferroelectricity in HfO 2 -based ferroelectric thin film by light ion bombardment. Science 376:731–738 Cheema SS et al (2022) Ultrathin ferroic HfO 2 -ZrO 2 superlattice gate stack for advanced transistors. Nature 604:65–71 Vishnumurthy P, Alcala R, Mikolajick T, Schroeder U, Antunes LA, Kersch A (2024) Ferroelectric HfO 2 -based capacitors for FeRAM: Reliability from field cycling endurance to retention. In: 2024 IEEE International Reliability Physics Symposium (IRPS) ). IEEE Wang Y et al (2023) A stable rhombohedral phase in ferroelectric Hf(Zr) 1+x O 2 capacitor with ultralow coercive field. Science 381:558–563 Kim JS et al (2025) Coercive field control in epitaxial ferroelectric Hf 0.5 Zr 0.5 O 2 thin films by nanostructure engineering. ACS Appl Mater Interfaces 17:25442–25450 Kelley KP et al (2023) Ferroelectricity in hafnia controlled via surface electrochemical state. Nat Mater 22:1144–1151 Lee Y et al (2022) The influence of crystallographic texture on structural and electrical properties in ferroelectric Hf 0.5 Zr 0.5 O 2 . J Appl Phys 132 Liao J et al (2019) Grain size engineering of ferroelectric Zr-doped HfO 2 for the highly scaled devices applications. IEEE Electron Device Lett 40:1868–1871 Yun Y et al (2022) Intrinsic ferroelectricity in Y-doped HfO 2 thin films. Nat Mater 21:903–909 Lee J et al (2023) Role of oxygen vacancies in ferroelectric or resistive switching hafnium oxide. Nano Convergence 10:55 Rushchanskii KZ, Blügel S, Ležaić M (2021) Ordering of oxygen vacancies and related ferroelectric properties in HfO 2-δ . Phys Rev Lett 127:087602 Jia S et al (2025) Developing HZO-based superlattices to enhance fatigue‐resistance by charge injection suppression. Adv Funct Mater, 2501470 Estandia S et al (2019) Engineering ferroelectric Hf 0.5 Zr 0.5 O 2 thin films by epitaxial stress. ACS Appl Electron Mater 1:1449–1457 Shi S et al (2024) Stabilizing the ferroelectric phase of Hf 0.5 Zr 0.5 O 2 thin films by charge transfer. Phys Rev Lett 133:036202 Baumgarten L et al (2024) Smart design of fermi level pinning in HfO 2 -based ferroelectric memories. Adv Funct Mater 34:2307120 Lee Y, Goh Y, Hwang J, Das D, Jeon S (2021) The influence of top and bottom metal electrodes on ferroelectricity of hafnia. IEEE Trans Electron Devices 68:523–528 Starschich S, Menzel S, Böttger U (2017) Pulse wake-up and breakdown investigation of ferroelectric yttrium doped HfO 2 . J Appl Phys 121 Sharia O, Tse K, Robertson J, Demkov AA (2009) Extended Frenkel pairs and band alignment at metal-oxide interfaces. Phys Rev B—Condensed Matter Mater Phys 79:125305 Demkov AA (2006) Thermodynamic stability and band alignment at a metal-high-k dielectric interface. Phys Rev B 74:085310 Memories F (2023) Smart Design of Fermi Level Pinning in HfO Akasaka Y et al (2006) Modified oxygen vacancy induced fermi level pinning model extendable to P-metal pinning. Jpn J Appl Phys 45:L1289 Wei Y et al (2018) A rhombohedral ferroelectric phase in epitaxially strained Hf 0.5 Zr 0.5 O 2 thin films. Nat Mater 17:1095–1100 Hill MO et al (2024) Depth-resolved X‐ray photoelectron spectroscopy evidence of intrinsic polar states in HfO 2 ‐based ferroelectrics. Adv Mater 36:2408572 Chen G, Hou Z, Gong X, Li Q (2008) Effects of Y doping on the structural stability and defect properties of cubic HfO 2 . J Appl Phys 104 Yang Z et al (2009) Role of interface dipole in metal gate/high-k effective work function modulation by aluminum incorporation. Appl Phys Lett 94 Lee H-S, Park H-H (2015) Band structure analysis of La 0.7 Sr 0.3 MnO 3 perovskite manganite using a synchrotron. Advances in Condensed Matter Physics 746475 (2015) Fields SS, Jaszewski ST, Lenox MK, Ihlefeld JF (2023) Asymmetric electrode work function customization via top electrode replacement in ferroelectric and field-induced ferroelectric hafnium zirconium oxide thin films. Adv Mater Interfaces 10:2202232 Greiner MT, Chai L, Helander MG, Tang WM, Lu ZH (2012) Transition metal oxide work functions: the influence of cation oxidation state and oxygen vacancies. Adv Funct Mater 22:4557–4568 Liu N et al (2024) Precise control of fatigue, wake-up, charge injection, and break-down in Hf 0.5 Zr 0.5 O 2 -based ferroelectric memories. Appl Phys Lett 124 Pešić M et al (2016) Physical mechanisms behind the field-cycling behavior of HfO 2 ‐based ferroelectric capacitors. Adv Funct Mater 26:4601–4612 Cao T, Ren G, Shao D-F, Tsymbal EY, Mishra R (2023) Stabilizing polar phases in binary metal oxides by hole doping. Phys Rev Mater 7:044412 Fengler FP et al (2017) Domain pinning: Comparison of hafnia and PZT based ferroelectrics. Adv Electron Mater 3:1600505 Zeng B et al (2024) Polarization fatigue mechanism of laminated hafnium zirconium oxide ferroelectric thin films. Acta Mater 272:119920 Fillot F et al (2003) Investigations of the interface stability in HfO 2 -metal electrodes. Microelectron Eng 70:384–391 Kang H-K et al (2017) Al 2 O 3 passivation effect in HfO 2 · Al 2 O 3 laminate structures grown on InP substrates. ACS Appl Mater Interfaces 9:17526–17535 Zhang Y et al (2021) Evolution of the conductive filament system in HfO 2 -based memristors observed by direct atomic-scale imaging. Nat Commun 12:7232 Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryMaterials.docx Supporting materials for "Tailoring the metal-oxide interface for improving fatigue performance in HfO2-based ferroelectrics" Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7907164","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":533751326,"identity":"00d0a411-79e2-48af-a2aa-283b1191b64f","order_by":0,"name":"Yi-Xuan Liu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAvklEQVRIiWNgGAWjYFAC5sYHEhVwXgIxWhibDSzOkKilTaKyjRQtBjcSGyRuzrOT121gfviBsS2NCC1nDjYYztyWbLjtAJuxBGNbDhFajjc2JEtuO8C47QCDGdCRFURoOczYcPjvnAP22w6wfyNSy/HGxgbJhgOJ2w7wgGwhwmGSZw42M0gcS07edpinWCLhHBHe57uRfPyHRI2d7bbj7Rs/fChLJqxF4QCMxcxAZETKNxCjahSMglEwCkY2AACHtz1GsblYvwAAAABJRU5ErkJggg==","orcid":"","institution":"University of Cambridge","correspondingAuthor":true,"prefix":"","firstName":"Yi-Xuan","middleName":"","lastName":"Liu","suffix":""},{"id":533751327,"identity":"97c15f2d-8034-4bfd-aa57-cbc5a1f1079c","order_by":1,"name":"Ziyi Yuan","email":"","orcid":"https://orcid.org/0000-0002-4854-4926","institution":"University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Ziyi","middleName":"","lastName":"Yuan","suffix":""},{"id":533751328,"identity":"0ba64d37-5aad-46df-8d29-9aa01853c3c9","order_by":2,"name":"Babak Bakhit","email":"","orcid":"https://orcid.org/0000-0002-3083-7536","institution":"University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Babak","middleName":"","lastName":"Bakhit","suffix":""},{"id":533751329,"identity":"e76dff28-ed91-4e7d-9f84-7158799ef529","order_by":3,"name":"Han Yu","email":"","orcid":"","institution":"Tsinghua University","correspondingAuthor":false,"prefix":"","firstName":"Han","middleName":"","lastName":"Yu","suffix":""},{"id":533751330,"identity":"fc76e86a-97e2-4e70-8a8a-9f03430505da","order_by":4,"name":"Jiahao Lu","email":"","orcid":"","institution":"University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Jiahao","middleName":"","lastName":"Lu","suffix":""},{"id":533751331,"identity":"f32abbca-f240-45c2-abb0-b1e1f063028c","order_by":5,"name":"Ji Soo Kim","email":"","orcid":"https://orcid.org/0000-0003-0745-8754","institution":"University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Ji","middleName":"Soo","lastName":"Kim","suffix":""},{"id":533751332,"identity":"efe8d58a-8290-4e79-b1a4-e75df2ecc248","order_by":6,"name":"Ke Wang","email":"","orcid":"https://orcid.org/0000-0001-9840-2427","institution":"State Key Laboratory of New Ceramics and Fine Processing, School of Materials Science and Engineering, Tsinghua University, Beijing 100084, P. R. China","correspondingAuthor":false,"prefix":"","firstName":"Ke","middleName":"","lastName":"Wang","suffix":""},{"id":533751333,"identity":"45d03af9-5985-414d-b497-043db1b6e7fa","order_by":7,"name":"José Silva","email":"","orcid":"https://orcid.org/0000-0002-3485-7032","institution":"University of Minho","correspondingAuthor":false,"prefix":"","firstName":"José","middleName":"","lastName":"Silva","suffix":""},{"id":533751334,"identity":"bc09cfad-1d22-478d-b9fd-db40f3531616","order_by":8,"name":"Judith MacManus-Driscoll","email":"","orcid":"https://orcid.org/0000-0003-4987-6620","institution":"University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Judith","middleName":"","lastName":"MacManus-Driscoll","suffix":""}],"badges":[],"createdAt":"2025-10-20 15:36:07","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7907164/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7907164/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":94252034,"identity":"8ffb8b67-7f30-4539-a30b-972589fd8ca8","added_by":"auto","created_at":"2025-10-24 07:03:44","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3254057,"visible":true,"origin":"","legend":"","description":"","filename":"MainText.docx","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/5c4615653bf2235bf5b7f65e.docx"},{"id":94252033,"identity":"6d8a701c-ed5b-40b8-9758-f850f3e256f5","added_by":"auto","created_at":"2025-10-24 07:03:44","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10018,"visible":true,"origin":"","legend":"","description":"","filename":"NCOMMS2584162.json","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/f971c05dcc7dc06b5f2c1d3f.json"},{"id":94251643,"identity":"5262cc4e-abe7-4703-ab3e-354e9a00ca69","added_by":"auto","created_at":"2025-10-24 06:55:44","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1185616,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterials.docx","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/19d8394ee869a4ed77232113.docx"},{"id":94251646,"identity":"274dfbd1-3089-4e2d-ba47-6170fb3701f5","added_by":"auto","created_at":"2025-10-24 06:55:44","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":113489,"visible":true,"origin":"","legend":"","description":"","filename":"NCOMMS25841620enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/ef453c34bd21c915632d2050.xml"},{"id":94251655,"identity":"5e48964e-02d7-4277-8f05-7bdae4d550e8","added_by":"auto","created_at":"2025-10-24 06:55:44","extension":"jpeg","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1216432,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/c09c12567ba8f97284c12e47.jpeg"},{"id":94251649,"identity":"f8a5c087-793f-471b-9fbd-4661481b50fc","added_by":"auto","created_at":"2025-10-24 06:55:44","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":815143,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/bc1b7e0aaf848de4da5ab9a6.png"},{"id":94251653,"identity":"8c10dd67-c736-4e58-88d7-d54d03ac355b","added_by":"auto","created_at":"2025-10-24 06:55:44","extension":"jpeg","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":534202,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/7d6a87160a865d84b1630de0.jpeg"},{"id":94253072,"identity":"534ce91d-5e47-44f0-b113-2ea3d3dd1e05","added_by":"auto","created_at":"2025-10-24 07:11:44","extension":"jpeg","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":598516,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/6ba4cdb6f90cfc550bce62ab.jpeg"},{"id":94252037,"identity":"0c250d82-ba02-454f-8e03-d85792d68cc1","added_by":"auto","created_at":"2025-10-24 07:03:44","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":163243,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/e538296d95a7c02202cc098f.png"},{"id":94251644,"identity":"133ee0cc-cbf8-4dac-97f9-eef569560d8d","added_by":"auto","created_at":"2025-10-24 06:55:44","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":92495,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/427d23516e9c16449b95af22.png"},{"id":94251648,"identity":"a4bf2473-05b9-480a-93e9-67b5f562696b","added_by":"auto","created_at":"2025-10-24 06:55:44","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":51310,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/ef91f86dae2061c024654ff4.png"},{"id":94251650,"identity":"f6692b3e-c7bf-4157-bedb-0d710e1b89bc","added_by":"auto","created_at":"2025-10-24 06:55:44","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":56162,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/21cb29b2ea760ad4bfc2dd57.png"},{"id":94252036,"identity":"c9aee0d5-2d11-426e-adc4-2889ee903a32","added_by":"auto","created_at":"2025-10-24 07:03:44","extension":"xml","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":110598,"visible":true,"origin":"","legend":"","description":"","filename":"NCOMMS25841620structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/d60112c8c9d8eeb81b9b73c2.xml"},{"id":94251651,"identity":"5b96b6a7-0ba7-41d4-bd64-25b79ae32c69","added_by":"auto","created_at":"2025-10-24 06:55:44","extension":"html","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":123902,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/d8717c8f400051b896dff8c8.html"},{"id":94251637,"identity":"3dbcd269-15e5-4e9a-a128-cf68977eba65","added_by":"auto","created_at":"2025-10-24 06:55:44","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":936070,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eStructure characterisation and domain configuration of YHO films.\u003c/strong\u003e \u003cstrong\u003ea\u003c/strong\u003e, Schematic diagram of the device structure based on (111)-oriented YHO thin films epitaxially grown on LSMO(001)/STO(001), with various metal top electrodes (Me = Pt, Cu, W, Mo, Ni, Al). The crystal structure of YHO is described by\u003cem\u003e Pca\u003c/em\u003e2\u003csub\u003e1\u003c/sub\u003e space group with rhombohedral (\u003cem\u003eR\u003c/em\u003e) distortion. \u003cstrong\u003eb\u003c/strong\u003e,2\u003cem\u003eθ-ω\u003c/em\u003e XRD scan of YHO(111) films grown on LSMO (001)/STO(001) with various top electrodes of Pt, Cu, W, Mo, Ni, and Al. \u003cstrong\u003ec\u003c/strong\u003e, Rocking curves of YHO(111) peaks with the various top electrodes. The inset of (c) demonstrates that the rocking curve of the YHO (111) peak consists of a broad peak (the part with low crystallinity) and a sharp peak (the part with high crystallinity). \u003cstrong\u003ed\u003c/strong\u003e, Pole figure around the (111) peak of YHO film at 2\u003cem\u003eθ\u003c/em\u003e=29.8°. The radial axis corresponds to 𝜒 with a (0°-90°) range, while the azimuthal axis represents 𝜑 in 0°-360° range. \u003cstrong\u003ee\u003c/strong\u003e, Domain symmetry at 2\u003cem\u003eθ\u003c/em\u003e=29.8° in 𝜒 range between 0°-90° and 𝜑 range between 0°-360°. The sketch shows the 90°-rotated crystal domains relative to each other. \u003cstrong\u003ef\u003c/strong\u003e, 2\u003cem\u003eθ\u003c/em\u003e scans of the 12 peaks in the pole figure and the out-of-plane (111) peak. \u003cstrong\u003eg\u003c/strong\u003e, out-of-plane and \u003cstrong\u003eh\u003c/strong\u003e, in-plane PFM response of phase (above) and amplitude (below) after poling with \u003cem\u003eV\u003c/em\u003e\u003csub\u003eDC\u003c/sub\u003e of ±8 V for YHO film. \u003cstrong\u003ei\u003c/strong\u003e, Out-of-plane hysteresis loops measured by switching spectroscopy PFM in amplitude and phase collected with \u003cem\u003eV\u003c/em\u003e\u003csub\u003eAC\u003c/sub\u003e of 800 mV at 15 kHz for YHO film.\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/90d2fc2ef82c5d94540299c2.png"},{"id":94251638,"identity":"3477f263-184b-47dc-8138-f0bc667324e2","added_by":"auto","created_at":"2025-10-24 06:55:44","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":715557,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eKelvin Probe Force Microscopy (KPFM) measurement of YHO with different top metal electrodes. a\u003c/strong\u003e, Schematic and working principle of KPFM measurement. \u003cstrong\u003eb\u003c/strong\u003e, \u003cstrong\u003ec \u003c/strong\u003eBand model of the extended oxygen Frenkel pair formation (b) before and (c) after charge transfer from the energy level of \u003cem\u003eV\u003c/em\u003e\u003csub\u003eO\u003c/sub\u003e to the metal. Neutral oxygen vacancies occupied a defect state at ~2.5 eV above the valence band in YHO, which is higher than the Fermi levels of some metals with high work function.\u003csup\u003e28\u003c/sup\u003e \u003cem\u003ec\u003c/em\u003e\u003csub\u003eYHO\u003c/sub\u003e denotes electron affinity of the YHO. After band alignment, a bending barrier, \u003cem\u003ef\u003c/em\u003e\u003csub\u003eBB\u003c/sub\u003e, and effective metal work function, ϕ\u003csub\u003eMe\u003c/sub\u003e\u003csup\u003eeff\u003c/sup\u003e, \u0026nbsp;are induced by interfacial charge accumulation. Positively charged V\u003csub\u003eO\u003c/sub\u003e\u003csup\u003e••\u003c/sup\u003e, denoted by red circles, form in the pink region adjacent to the metal electrode. \u003cstrong\u003ed\u003c/strong\u003e-\u003cstrong\u003eh\u003c/strong\u003e, Potential mapping of YHO films and top metal electrodes (\u003cstrong\u003ed\u003c/strong\u003e) Pt, (\u003cstrong\u003ee\u003c/strong\u003e) Cu, (\u003cstrong\u003ef\u003c/strong\u003e) W, (\u003cstrong\u003eg\u003c/strong\u003e) Mo, and (\u003cstrong\u003eh\u003c/strong\u003e) Ni. The potential difference between the YHO films and the top metal electrodes is calculated as \u003cem\u003eDP\u003c/em\u003e marked in the bottom-right corner, respectively. \u003cstrong\u003ei\u003c/strong\u003e, Comparison of the work functions of metal electrodes (left) and the measured potential differences, \u003cem\u003eDP\u003c/em\u003e\u003csub\u003eMe-YHO\u003c/sub\u003e, between the metals and YHO (right) by KPFM with the dependency of their EFPs formation energy \u003cem\u003eE\u003c/em\u003e\u003csub\u003eEFP\u003c/sub\u003e.\u0026nbsp;\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/74231ccc92f3af5b8006d3bc.png"},{"id":94251641,"identity":"ed8a87cf-259b-4663-a323-e340c1085342","added_by":"auto","created_at":"2025-10-24 06:55:44","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":272426,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFerroelectric properties and fatigue performance of YHO films with different top electrode metals.\u003c/strong\u003e \u003cstrong\u003ea\u003c/strong\u003e, Ferroelectric \u003cem\u003eP-V\u003c/em\u003e by the PUND method on YHO films with different top electrode metals. The remanent polarisation, \u003cem\u003eP\u003c/em\u003e\u003csub\u003er\u003c/sub\u003e, and coercive voltage, \u003cem\u003eV\u003c/em\u003e\u003csub\u003eC\u003c/sub\u003e, are marked correspondingly. \u003cstrong\u003eb\u003c/strong\u003e, 2\u003cem\u003eP\u003c/em\u003e\u003csub\u003er\u003c/sub\u003e and 2\u003cem\u003eV\u003c/em\u003e\u003csub\u003ec\u003c/sub\u003e of YHO films as a function of the formation energy of EPFs for the different metal electrodes on top. \u003cstrong\u003ec\u003c/strong\u003e, Imprint polarisation, \u003cem\u003eDP\u003c/em\u003e\u003csub\u003e\u003cem\u003eE\u003c/em\u003e\u003c/sub\u003e\u003csub\u003e=0\u003c/sub\u003e, for YHO films with different top electrodes. The oxide interfaces, MeO\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e, formed between metal electrodes and the YHO film are considered, as indicated by \u003cem\u003ef\u003c/em\u003e\u003csub\u003eMeO\u003c/sub\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e on the right axis. \u003cem\u003eDf\u003c/em\u003e denotes the work function difference between \u003cem\u003ef\u003c/em\u003e\u003csub\u003eMeO\u003c/sub\u003e\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e and the LSMO bottom electrode (the metallic work function is used for Pt). A value of 4.8 eV is adopted for the work function of LSMO layer.\u003csup\u003e30\u003c/sup\u003e\u003cstrong\u003e d\u003c/strong\u003e-\u003cstrong\u003ei\u003c/strong\u003e, Fatigue measurement on YHO films with (d) Pt, (e) Cu, (f) W, (g) Mo, (h) Ni, and (i) Al top electrodes during fatigue cycling at 5 kV/mm \u0026nbsp;up to 10\u003csup\u003e8\u003c/sup\u003e cycles.\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/328d1d86ee3df7bc6b106b39.png"},{"id":94251640,"identity":"301c61ea-08de-4f12-9149-8d7579f4af4b","added_by":"auto","created_at":"2025-10-24 06:55:44","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":273272,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFerroelectric-resistive switching transition and fatigue failure mechanism. a\u003c/strong\u003e-\u003cstrong\u003ed\u003c/strong\u003e, Band alignment for YHO and (a) Pt, (b) Cu, W, Mo, and (c) Ni metal electrodes with a Fermi level pinning effect from the EFPs interface. \u003cem\u003eE\u003c/em\u003e\u003csub\u003eCB\u003c/sub\u003e and \u003cem\u003eE\u003c/em\u003e\u003csub\u003evac \u003c/sub\u003edenote the conduction band level and the vacuum level, respectively.\u003cem\u003e \u003c/em\u003e\u0026nbsp;E\u003csub\u003ef\u003c/sub\u003e\u003csup\u003eYHO\u003c/sup\u003e and E\u003csub\u003ef\u003c/sub\u003e\u003csup\u003eMe\u003c/sup\u003e refer to the Fermi level of YHO film and metal electrodes (Me=Pt, Cu, W, Mo, Ni), respectively. Φ\u003csub\u003eMe\u003c/sub\u003e and Φ\u003csub\u003eSB\u003c/sub\u003e\u003csup\u003eMe\u003c/sup\u003e indicate the work function of the metal and the effective energy barrier \u0026nbsp;in accordance with YHO after band alignment. \u003cstrong\u003ed\u003c/strong\u003e, Oxygen content across the YHO-Ni device before fatigue compared with that near the Ni/YHO interface. The inset in \u003cstrong\u003ed\u003c/strong\u003e illustrates the depth-resolved XPS measurement, where Ar ions are employed for depth sputtering. The Ni electrode of ~10 nm thick ensures non-destructive probing of the region. \u003cstrong\u003ee\u003c/strong\u003e, \u003cstrong\u003ef\u003c/strong\u003e, DC leakage current measurements on YHO with Ni top electrode (e) before and (f) after fatigue. \u003cstrong\u003eg\u003c/strong\u003e, Interfacial mechanism by oxygen vacancy accumulation and \u003cstrong\u003eh\u003c/strong\u003e, filamentary mechanism by oxygen vacancy conducting path. \u003cstrong\u003ei\u003c/strong\u003e, Fatigue performance and critical cycling number at failure for YHO films with different metal electrodes. The critical cycling numbers at failure are marked on the upper \u003cem\u003ex\u003c/em\u003e-axis, and the EFP defect formation energies of these metal electrodes are denoted on the right \u003cem\u003ey\u003c/em\u003e-axis. Note that the ferroelectric degradation of these capacitors shows a strong correlation with the effective energy barrier at the metal/YHO interfaces, as illustrated by the arrow in the lower right corner. The initial distribution of oxygen vacancies before fatigue, oxygen vacancy migration, and the formation of conductive paths are schematically illustrated throughout the fatigue process in the inset.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/55189b3de0b28ff789aa646d.png"},{"id":94253386,"identity":"9dd18055-99b3-4668-8b7e-d110b30ab8fb","added_by":"auto","created_at":"2025-10-24 07:19:46","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3090185,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/29223d6d-fe11-4648-8c6a-93953fdd5080.pdf"},{"id":94251639,"identity":"72952654-6b68-4e59-9c1a-9558da7aab20","added_by":"auto","created_at":"2025-10-24 06:55:44","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1185616,"visible":true,"origin":"","legend":"Supporting materials for \"Tailoring the metal-oxide interface for improving fatigue performance in HfO2-based ferroelectrics\"","description":"","filename":"SupplementaryMaterials.docx","url":"https://assets-eu.researchsquare.com/files/rs-7907164/v1/88a0a10ed1aa4d731fedbbaa.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Tailoring the metal-oxide interface for improving fatigue performance in HfO\u003csub\u003e2\u003c/sub\u003e-based ferroelectrics","fulltext":[{"header":"Introduction","content":"\u003cp\u003eNearly fifteen years after the discovery of ferroelectricity in doped hafnium oxide (HfO\u003csub\u003e2\u003c/sub\u003e), this material has emerged as a leading candidate for advanced nanoelectronic applications, owing to its compatibility with complementary metal-oxide-semiconductor (CMOS) technology.\u003csup\u003e1-3\u003c/sup\u003e In contrast to perovskite ferroelectrics (e.g., BaTiO\u003csub\u003e3\u003c/sub\u003e, PZT) that lose their polarisation below a critical thickness, ferroelectric HfO\u003csub\u003e2\u003c/sub\u003e defies the conventional size effect by maintaining and even enhancing polarisation at the nanometer scale, making it attractive for next-generation non-volatile memory devices.\u003csup\u003e4-7\u003c/sup\u003e Recent advances, including chemical doping, solid‑solution formation such as Hf\u003csub\u003e1-\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003eZr\u003cem\u003e\u003csub\u003ex\u003c/sub\u003e\u003c/em\u003eO\u003csub\u003e2\u003c/sub\u003e (HZO), strain engineering, grain size controlling and texturing, have further expanded the application potential of HfO\u003csub\u003e2\u003c/sub\u003e‑based ferroelectric thin films.\u003csup\u003e8-12\u003c/sup\u003e In particular, yttrium-doped HfO\u003csub\u003e2\u003c/sub\u003e (YHO) has attracted attention for its robust ferroelectricity with high remanent polarisation.\u003csup\u003e13\u003c/sup\u003e Although a substantial increase in remanent polarisation with reduced coercive field enables low-power operation, the long-term reliability of HfO\u003csub\u003e2\u003c/sub\u003e-based ferroelectrics, particularly their fatigue under electrical cycling, remains underexplored.\u003c/p\u003e\n\u003cp\u003eAcceptor-doped HfO\u003csub\u003e2\u0026nbsp;\u003c/sub\u003efilms intrinsically host a pronounced concentration of oxygen vacancies.\u003csup\u003e14\u003c/sup\u003e Under electrical cycling, these defects accumulate and then either destabilise the polar phase or pin the domain walls, both effects causing ferroelectric fatigue over time.\u003csup\u003e15\u003c/sup\u003e Additional contributions arise from interfacial charge injection and trapping, leading to the formation of passive interfacial layers or field-induced phase changes.\u003csup\u003e16\u003c/sup\u003e These processes occur primarily at the electrode/oxide interface. In epitaxial films, the bottom electrode imposes strain on the oxide and participates in interfacial redox reactions, thereby affecting phase stability and the distribution of oxygen vacancies.\u003csup\u003e17-19\u003c/sup\u003e In contrast, the top electrode acts as a capping layer, inducing residual stress that stabilises the ferroelectric phase.\u003csup\u003e20\u003c/sup\u003e Through oxygen exchange and interfacial reactions with the oxide, the top electrode also determines the interfacial band alignment and the charge-injection barrier during switching. Early studies employed noble metals such as Pt or Au to suppress interfacial reactions, enabling the demonstration of the intrinsic ferroelectric response of HfO\u003csub\u003e2\u003c/sub\u003e-based films.\u003csup\u003e21\u003c/sup\u003e However, these electrodes are incompatible with CMOS integration and fail to represent the interfacial chemistry of practical device metals, which can promote defect formation and carrier injection, thereby accelerating fatigue.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMoreover, the oxygen exchange at metal/HfO\u003csub\u003e2\u003c/sub\u003e interfaces is not solely determined by the vacancy formation energy of the metal oxide (MeO\u003cem\u003e\u003csub\u003ex\u003c/sub\u003e\u003c/em\u003e) but can be strongly influenced by interfacial reactions. One important process is the formation of extended Frenkel pairs (EFPs), where the removed oxygen from an oxide film is incorporated into the metal while the accompanying charge is transferred from the vacancy energy level of the induced vacancies to electrodes.\u003csup\u003e22\u003c/sup\u003e First-principles calculations show that at a Mo/HfO\u003csub\u003e2\u003c/sub\u003e interface the vacancy formation energy can decrease from ~6.3 eV in the bulk to ~1.4 eV when in contact with Mo. A similar effect occurs for other metal electrodes also.\u003csup\u003e23\u003c/sup\u003e Such EFP defects can create a charged dipole layer that can locally bend energy bands and modify the effective barrier at the interface, thus enhancing carrier injection and polarisation degradation during switching.\u003csup\u003e24\u003c/sup\u003e Therefore, the selection of metal electrodes becomes essential for understanding and improving the fatigue performance of practical HfO\u003csub\u003e2\u003c/sub\u003e-based ferroelectric devices.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn the present study, we examine the impact of top electrode choice on the fatigue behaviour of 7% Y-doped HfO\u003csub\u003e2\u003c/sub\u003e ferroelectric films. To exclude microstructural variability inherent to polycrystalline films, epitaxial YHO layers were grown on La\u003csub\u003e0.7\u003c/sub\u003eSr\u003csub\u003e0.3\u003c/sub\u003eMnO\u003csub\u003e3\u003c/sub\u003e-buffered SrTiO\u003csub\u003e3\u003c/sub\u003e substrates, yielding a well-defined distorted orthorhombic ferroelectric phase. A set of electrode metals (Cu, W, Mo, Ni, and Al) with similar work functions but different oxygen vacancy formation energy\u003csup\u003e25\u003c/sup\u003e, or vice versa, was selected to compare with the Pt electrode. All electrodes were deposited at room temperature to suppress extrinsic capping mechanical stress. Comprehensive electrical tests and Kelvin probe force microscopy (KPFM) were performed to evaluate polarisation fatigue and interface potential evolution for different electrodes. It was found that the selection of top metal electrodes can affect the polarisation value and coercive field before fatigue. Notably, these devices undergo a transition from ferroelectric switching to a mixed ferroelectric-resistive response, accompanied by enhanced leakage and resistive switching. Overall, our findings suggest a foundation for engineering metal-oxide interfaces to enhance the endurance of hafnium oxide ferroelectric memories.\u003c/p\u003e"},{"header":"Orthorhombic polar phase with rhombohedral distortion","content":"\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe device structure is demonstrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea, where (111)-oriented YHO thin films were epitaxially grown on La\u003csub\u003e0.67\u003c/sub\u003eSr\u003csub\u003e0.33\u003c/sub\u003eMnO₃ (LSMO)/SrTiO\u003csub\u003e3\u003c/sub\u003e (STO) substrates with various metal top electrodes (Pt, Cu, W, Mo, Ni, Al) patterned on the YHO surface. The 25 nm-thick LSMO bottom electrode was epitaxially grown on an STO (001) substrate along the (001) direction layer by layer, as confirmed by the periodic oscillations of the intensity from the reflection high-energy electron diffraction monitoring (\u003cb\u003efig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e). The X-ray diffraction (XRD) \u003cem\u003e2θ-ω\u003c/em\u003e pattern (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb) confirms that the YHO film is (111)-oriented single-phase structure. The primary YHO (111) reflection appears at the 2\u003cem\u003eθ\u003c/em\u003e = ~29.8\u0026deg;, alongside STO (001) and LSMO (001) peaks, with the well-defined Laue oscillations, reflecting smooth film interfaces, high crystallinity, and uniform thickness. The (111) peak of YHO remains unchanged with different top electrode materials, indicating that the \u003cem\u003eex-situ\u003c/em\u003e sputtering of metal electrodes has a negligible impact on the phase structure of YHO. The film thickness is estimated to be ~\u0026thinsp;8 nm from the full width at half maximum (FWHM) of the (111) peak. Note that the 2\u003cem\u003eθ\u003c/em\u003e value of (111) peak (at 2\u003cem\u003eθ\u003c/em\u003e = ~29.8\u0026deg;) is slightly lower than that in the commonly reported polar o-phase in HfO\u003csub\u003e2\u003c/sub\u003e (at 2\u003cem\u003eθ\u003c/em\u003e = ~30.5\u0026deg;).\u003csup\u003e13\u003c/sup\u003e This suggests an expanded (111)-spacing (\u003cem\u003ed\u003c/em\u003e\u003csub\u003e111\u003c/sub\u003e) along the out-of-plane direction, where the high structural order arises from the lattice mismatch and is consistent with the orthorhombic \u003cem\u003ePca\u003c/em\u003e2\u003csub\u003e1\u003c/sub\u003e phase with a slight rhombohedral distortion.\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e The high crystalline quality of the YHO film is evidenced by the sharp central peak in the \u003cem\u003eω\u003c/em\u003e-rocking curve of the (111) reflection, confirming that most of the YHO lattice is well aligned with the substrate.\u003c/p\u003e\u003cp\u003eTo further confirm the symmetry of the YHO lattice, pole figure measurements were performed around 2\u003cem\u003eθ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;29.8\u0026deg;, i.e., around the YHO\u003csub\u003e{111}\u003c/sub\u003e reflection. For (111)-oriented single-domain film, three reflections, (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\stackrel{-}{1}\\)\u003c/span\u003e\u003c/span\u003e11), (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:1\\stackrel{-}{1}\\)\u003c/span\u003e\u003c/span\u003e1), and (1\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:1\\stackrel{-}{1}\\)\u003c/span\u003e\u003c/span\u003e), are expected at \u003cem\u003eχ\u003c/em\u003e\u0026thinsp;~\u0026thinsp;71\u0026deg; from the out-of-plane direction, with azimuthal angles \u003cem\u003eφ\u003c/em\u003e differing by 120\u0026deg;. The pole figure at 2\u003cem\u003eθ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;29.8\u0026deg; (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ed) reveals 12 reflections, indicating the presence of four distinct in-plane domain variants. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ee, these crystal domains, rotated by 90\u0026deg; relative to each other, align with the four-fold symmetry of the cubic STO substrate. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ef, the 12 peaks with in-plane components of the scattering vector at \u003cem\u003eχ\u003c/em\u003e\u0026thinsp;\u0026asymp;\u0026thinsp;71\u0026deg; share the same 2\u003cem\u003eθ\u003c/em\u003e\u0026thinsp;=\u0026thinsp;30.5\u0026deg;, suggesting an equivalent \u003cem\u003ed\u003c/em\u003e-spacing of 2.93\u0026Aring; that is smaller than that of the out-of-plane (111) reflection (\u003cem\u003ed\u003c/em\u003e\u003csub\u003e111\u003c/sub\u003e\u0026thinsp;~\u0026thinsp;2.99 \u0026Aring;). Therefore, these combined characterisations verify that the orthorhombic unit cells in YHO films are elongated along the (111) polar axis.\u003c/p\u003e\u003cp\u003eThe presence of ferroelectricity in YHO films can be verified by piezo-response force microscopy (PFM). After poling the film with \u0026plusmn;\u0026thinsp;8 V, PFM phase images exhibit 180\u0026deg; contrast between oppositely poled domains both along out-of-plane (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eg) and in-plane (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eh) direction. This indicates the presence of two reversibly switchable polarisation states. Switching spectroscopy-PFM (SS-PFM) further displays a sharp phase inversion and a butterfly-shaped amplitude loop as a function of bias, determining a local coercive voltage \u003cem\u003eV\u003c/em\u003e\u003csub\u003ec\u003c/sub\u003e of 2.7 V and a phase-loop saturation voltage of comparable magnitude.\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e This nanoscale response implies the reversible ferroelectric switching in YHO films.\u003c/p\u003e"},{"header":"Schottky contact between YHO and top electrode metals","content":"\u003cp\u003eIn device structures, the electrode work function is often modified by interfacial reactions, charge transfer, and dipole formation, leading to an effective work function that deviates from the intrinsic value. Therefore, Kelvin probe force microscopy (KPFM) was employed to determine the contact potential difference (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea, \u003cb\u003efig. S2a\u003c/b\u003e and \u003cb\u003eS2b\u003c/b\u003e) between the metal electrodes and the YHO surface. Atomic force microscopy (AFM) confirmed a smooth surface morphology with uniform electrode thickness of ~\u0026thinsp;70 nm (\u003cb\u003efig. S2c\u003c/b\u003e), ensuring that topography-induced artefacts were negligible in the KPFM measurements. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed-h show spatial mapping of contact potential differences between the metal and YHO (Δ\u003cem\u003eP\u003c/em\u003e\u003csub\u003eMe\u0026minus;YHO\u003c/sub\u003e), with colour contrast corresponding to potential magnitude. Gaussian fitting of the potential difference distribution (\u003cb\u003efig. S3a-e\u003c/b\u003e) yields Δ\u003cem\u003eP\u003c/em\u003e\u003csub\u003eMe\u0026minus;YHO\u003c/sub\u003e of 260 mV, 60 mV, 240 mV, 140 mV, and 435 mV for Pt, Cu, W, Mo, and Ni electrodes, respectively (\u003cb\u003efig. S2f\u003c/b\u003e). It is noted that the measured potential differences between the metals and YHO, Δ\u003cem\u003eP\u003c/em\u003e\u003csub\u003eMe\u0026minus;YHO\u003c/sub\u003e, do not exhibit the anticipated positive correlation with metal work function (orange curve and cyan curve in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ei). Specifically, Cu, W, and Mo, despite similar work functions, show distinctly different Δ\u003cem\u003eP\u003c/em\u003e\u003csub\u003eMe\u0026minus;YHO\u003c/sub\u003e. Moreover, Ni has a considerably higher Δ\u003cem\u003eP\u003c/em\u003e\u003csub\u003eNi\u0026minus;YHO\u003c/sub\u003e ~ 435 mV than that of Δ\u003cem\u003eP\u003c/em\u003e\u003csub\u003ePt\u0026minus;YHO\u003c/sub\u003e ~ 260 mV, despite its lower work function (5.16 eV for Ni compared to 5.65 eV for Pt).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe inconsistency between the measured contact potentials and the nominal metal work functions suggests the presence of interfacial effects at the metal/YHO interface. Stronger interfacial interactions generally enhance the deviation between the measured contact potential and the intrinsic work function, as interface dipoles introduce additional potential drops that shift the band alignment beyond the prediction.\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e Considering the heat of metal oxides formation per oxygen atom, certain metal such as Ni, are not expected to form stable oxides, yet they show a substantial deviation in their measured contact potential. This deviation might suggest additional mechanisms, where a key process is the formation of EFP defects. Interfacial oxygen exchange occurs when the YHO is contacted by a metal with a lower oxide formation energy than Hf, expressed as\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:{\\text{O}}_{\\text{O}}+\\frac{1}{n}\\text{M}\\text{e}={\\text{V}}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}+\\frac{1}{n}\\text{M}\\text{e}{\\text{O}}_{n}+2{e}^{}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003ewhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{O}}_{\\text{O}}\\)\u003c/span\u003e\u003c/span\u003e represents the lattice oxygen in YHO that diffuses into the metal electrode. Me denotes the metal electrode and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}\\)\u003c/span\u003e\u003c/span\u003e is the positively charged oxygen vacancy left in YHO. MeO\u003csub\u003en\u003c/sub\u003e refers to the metal oxide formed at the interface. This reaction describes the migration of oxygen atoms from YHO to the metal, leading to oxidation of the electrode. Initially, these oxygen vacancies in YHO are electrically neutral (referred to as \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{\\text{O}}\\)\u003c/span\u003e\u003c/span\u003e), containing two localized electrons at the vacancy sites with an energy level of ~\u0026thinsp;2.5 eV above the valence band of YHO (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb).\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e When contacted by a high-work-function metal, of which the Fermi level lies below that of YHO, electrons from neutral \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{\\text{O}}\\)\u003c/span\u003e\u003c/span\u003e are transferred to the metal, leaving positively charged \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}\\)\u003c/span\u003e\u003c/span\u003e, i.e., EFP defects, between the film and metal electrode, generating an interfacial dipole layer adjacent to the metal electrode (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec). The accumulation of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}\\)\u003c/span\u003e\u003c/span\u003e significantly modifies the Schottky barrier height, requiring an effective metal work function, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{}_{\\text{M}\\text{e}}^{\\text{e}\\text{f}\\text{f}}\\)\u003c/span\u003e\u003c/span\u003e, to describe the interfacial band alignment. The formation energy of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{V}}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}\\)\u003c/span\u003e\u003c/span\u003e can be quantified as\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e\n$$\\:{E}_{EFP}={E}_{\\text{f}\\text{o}\\text{r}\\text{m}}+\\:{E}_{\\text{M}\\text{O}}\\left(\\frac{1}{n}\\text{M}\\text{e}{\\text{O}}_{n}\\right)+2\\left({E}_{{\\text{V}}_{\\text{O}}}-{E}_{f,\\:\\:\\text{M}\\text{e}}\\right)+{E}_{electro}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003ewhere \u003cem\u003eE\u003c/em\u003e\u003csub\u003eform\u003c/sub\u003e denotes the formation energy of a neutral vacancy relative to molecular O\u003csub\u003e2\u003c/sub\u003e, \u003cem\u003eE\u003c/em\u003e\u003csub\u003eMO\u003c/sub\u003e represents the formation free energy of the metal oxide per oxygen atom. The term, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:2\\left({E}_{{\\text{V}}_{\\text{O}}}-{E}_{f,\\:\\:\\text{M}\\text{e}}\\right)\\)\u003c/span\u003e\u003c/span\u003e, accounts for the energy gained from two electrons falling from \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{V}}_{\\text{O}}\\)\u003c/span\u003e\u003c/span\u003e energy level (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{E}_{{\\text{V}}_{\\text{O}}}\\)\u003c/span\u003e\u003c/span\u003e) into the metal Fermi level (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{E}_{f,\\:\\:\\text{M}\\text{e}}\\)\u003c/span\u003e\u003c/span\u003e). \u003cem\u003eE\u003c/em\u003e\u003csub\u003eelectro\u003c/sub\u003e denotes the electrostatic contribution of EFP defects to the total energy.\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e For different metal electrodes, only the second and third terms in the formation energy expression vary, which oppose each other. Metals with lower oxide formation energies, or with higher work functions that facilitate electron transfer from vacancies, have reduced EFP formation energy (\u003cem\u003eE\u003c/em\u003e\u003csub\u003eEFP\u003c/sub\u003e). According to a theoretical study\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e, Ni exhibits a particularly low formation energy of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}\\)\u003c/span\u003e\u003c/span\u003e​ (\u003cem\u003eE\u003c/em\u003e\u003csub\u003eEFP\u003c/sub\u003e ~1.47 eV) compared to Cu (3.35 eV), W (2.65 eV), and Mo (2.07 eV), leading to enhanced vacancy and dipole accumulation. As a result, the measured contact potentials, \u003cem\u003eΔP\u003c/em\u003e\u003csub\u003eMe\u0026minus;YHO\u003c/sub\u003e, show a strong negative correlation with \u003cem\u003eE\u003c/em\u003e\u003csub\u003eEFP\u003c/sub\u003e (with a Pearson coefficient \u003cem\u003eρ\u003c/em\u003e ~-0.988) rather than solely with their intrinsic work functions or oxide formation energies. Pt, on the other hand, has high work function (5.65 eV) like Ni, yet its inert nature prevents it from EFP defect formation.\u003c/p\u003e\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eFerroelectricity and fatigue performance of YHO with different top metal electrodes\u003c/h2\u003e\u003cp\u003eThe top electrode material is found to significantly influence the ferroelectric performance of YHO films. Positive-up negative-down (PUND) measurements showed a clear ferroelectric switching peak in current-voltage loops at switching fields of 3.7-5 MV/cm for Pt, Cu, W, Mo, and Ni top electrodes (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea and \u003cb\u003efig. S4\u003c/b\u003e). Notably, the higher \u003cem\u003eE\u003c/em\u003e\u003csub\u003eC\u003c/sub\u003e (21.75 MV/cm) of the YHO-Al capacitor may be due to the voltage drop across the interfacial alumina layer, as Al is more reactive than Hf (\u003cb\u003efig. S4f\u003c/b\u003e). Note that alumina formation is ineluctable from being exposed to the atmosphere and/or being in contact with YHO. The functionality of the alumina layer is outside the scope of this work. The remanent polarisation (2\u003cem\u003eP\u003c/em\u003e\u003csub\u003er\u003c/sub\u003e) of YHO films with Pt, Cu, W, Mo, and Ni electrodes ranges from 14 to 18 \u0026micro;C/cm\u003csup\u003e2\u003c/sup\u003e (\u003cb\u003efig. S4a-e\u003c/b\u003e). Among these, YHO-Pt exhibits the highest 2\u003cem\u003eP\u003c/em\u003e\u003csub\u003er\u003c/sub\u003e\u0026thinsp;~\u0026thinsp;18 \u0026micro;C/cm\u003csup\u003e2\u003c/sup\u003e, whereas YHO-Ni shows a reduced 2\u003cem\u003eP\u003c/em\u003e\u003csub\u003er\u003c/sub\u003e\u0026thinsp;~\u0026thinsp;14 \u0026micro;C/cm\u003csup\u003e2\u003c/sup\u003e and a markedly higher coercive voltage 2\u003cem\u003eV\u003c/em\u003e\u003csub\u003eC\u003c/sub\u003e of 8.4 V (\u003cem\u003eE\u003c/em\u003e\u003csub\u003eC\u003c/sub\u003e ~ 4.8 MV/cm), despite having a comparable oxide formation energy to W (\u003cem\u003eE\u003c/em\u003e\u003csub\u003eC\u003c/sub\u003e ~ 3.7 MV/cm). These deviations cannot be rationalised by work function or \u003cem\u003eE\u003c/em\u003e\u003csub\u003eMO\u003c/sub\u003e alone but instead correlate with EFP defects at the metal/YHO interface. As the EFP defect formation energy decreases, the accumulation of charged vacancies intensifies and thereby results in suppressed \u003cem\u003eP\u003c/em\u003e\u003csub\u003er\u003c/sub\u003e and an increased \u003cem\u003eV\u003c/em\u003e\u003csub\u003eC\u003c/sub\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eFurthermore, polarisation values at zero voltage differ for the positive and negative electric field (\u003cem\u003eΔP\u003c/em\u003e\u003csub\u003e\u003cem\u003eE\u003c/em\u003e=0\u003c/sub\u003e), indicating imprint effect due to the built-in field. This asymmetry corresponds to the difference in work function between the interfacial oxide MeO\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e layer at the top contact and the LSMO bottom electrode.\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e Considering the work function of metal oxides is correlated with the oxidation states of their cations, the work functions of oxides with average cation oxidation states in the literature are used for illustrative comparison (Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e).\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec, \u003cem\u003eΔP\u003c/em\u003e\u003csub\u003e\u003cem\u003eE\u003c/em\u003e=0\u003c/sub\u003e varies in proportion to the work function of interfacial oxide MeO\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e. In particular, the YHO-W exhibits the smallest difference between \u003cem\u003eφ\u003c/em\u003e\u003csub\u003eMeO\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e and the LSMO bottom electrode (\u003cem\u003eΔφ\u003c/em\u003e, see below Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec), and accordingly, the minimal polarisation offset. These results indicate that the interfacial MeO\u003csub\u003e\u003cem\u003ex\u003c/em\u003e\u003c/sub\u003e layer can partially mitigate the intrinsic asymmetry between the top and bottom electrodes, minimising the built-in field within YHO capacitors and consequently reducing the imprint effect.\u003c/p\u003e\u003cp\u003eThe ferroelectric fatigue behaviour of YHO films was measured using an electric switching cycle \u003cem\u003eE\u003c/em\u003e\u003csub\u003es\u003c/sub\u003e~6.25 MV/cm at a field amplitude of approximately 1.5\u003cem\u003eE\u003c/em\u003e\u003csub\u003ec\u003c/sub\u003e. For most YHO capacitors, the fatigue behaviour evolves in two distinct stages. Initially, the \u003cem\u003eP\u003c/em\u003e\u003csub\u003er\u003c/sub\u003e decreases monotonically with electrical cycling, reflecting typical ferroelectric fatigue. In some devices, an increase in \u0026plusmn;\u0026thinsp;\u003cem\u003eP\u003c/em\u003e\u003csub\u003er\u003c/sub\u003e appears beyond a critical cycle number, which can be attributed to the sharp rise in leakage current after ferroelectric failure. This gradual ferroelectric failure differs from dielectric breakdown that can be characterized by a linear \u003cem\u003eI-V\u003c/em\u003e relationship.\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e The choice of metal electrode strongly influences both polarisation decay and failure onset. YHO-Pt devices exhibit exceptional fatigue resistance, showing nearly unchanged polarisation even after 10\u003csup\u003e8\u003c/sup\u003e cycles, consistent with the high chemical stability of Pt. YHO\u0026ndash;Cu devices also maintain good fatigue resistance, showing only a small polarisation reduction (~\u0026thinsp;7%). In contrast, YHO devices with W, Mo, and Ni electrodes exhibit more pronounced ferroelectric degradation, with reductions of 10%, 18%, and 26%, respectively, implying their inferior fatigue resistance. The YHO-Al device shows a distinct fatigue behaviour, with an initial gradual decay followed by a sharp polarisation drop after ~\u0026thinsp;10\u003csup\u003e7\u003c/sup\u003e cycles.\u003c/p\u003e\u003cp\u003eThe selection of metal electrodes also influences the critical cycling threshold for ferroelectric failure. The ferroelectric failure occurs after approximately 10\u003csup\u003e7\u003c/sup\u003e-10\u003csup\u003e8\u003c/sup\u003e cycles for YHO with Cu electrodes, whereas W, Mo, and Ni electrodes lead to failure at lower cycles of 1\u0026times;10\u003csup\u003e7\u003c/sup\u003e, 5\u0026times;10\u003csup\u003e6\u003c/sup\u003e, and 2\u0026times;10\u003csup\u003e6\u003c/sup\u003e cycles, respectively. The failure behaviour of Ni-electrode devices is particularly notable, as their failure cycles are significantly lower than those of Pt-electrode devices despite their similarly high work functions. This suggests that factors beyond metal work function, such as chemical reactivity and defect dynamics, may play crucial roles in the ferroelectric stability of YHO devices.\u003c/p\u003e\u003c/div\u003e"},{"header":"Fatigue and ferroelectric failure mechanism","content":"\u003cp\u003eThe metal/YHO interface plays a critical role in determining the ferroelectric and fatigue behaviour of YHO devices, as defect chemistry within the film and charge injection during electrical cycling act primarily at the interface. In particular, the effective work function together with the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}\\)\u003c/span\u003e\u003c/span\u003e formed at the interface with the electrode jointly determine the band alignment. Figure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea-d illustrate the band alignment for YHO contacting with different metal electrodes. The metals with lower work functions, such as Cu, W, and Mo, promote electron injection into YHO under an applied field (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb). Meanwhile, their lower \u003cem\u003eE\u003c/em\u003e\u003csub\u003eEFP\u003c/sub\u003e facilitates oxygen ion capture from YHO, leaving behind charged \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}\\)\u003c/span\u003e\u003c/span\u003e near the interface. During electrical cycling, additional oxygen vacancies can be generated at the metal/YHO interface, increasing the concentration of mobile \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}\\)\u003c/span\u003e\u003c/span\u003e within the film.\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e Compared with perovskite ferroelectric materials (typical \u003cem\u003eE\u003c/em\u003e\u003csub\u003eC\u003c/sub\u003e~10\u0026ndash;100 kV/cm), the high \u003cem\u003eE\u003c/em\u003e\u003csub\u003eC\u003c/sub\u003e of HfO\u003csub\u003e2\u003c/sub\u003e-based ferroelectrics, on the order of MV/cm, is sufficient to drive these \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}\\)\u003c/span\u003e\u003c/span\u003e across the thin films. An appropriate \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}\\)\u003c/span\u003e\u003c/span\u003e concentration can stabilize the ferroelectric orthorhombic phase, whereas excessive accumulation destabilizes the polar structure.\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e As a result, in YHO devices prone to EFP defect formation, the facilitated accumulation of interfacial \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}\\)\u003c/span\u003e\u003c/span\u003e under cycling accelerates fatigue, consistent with previous reports that electron injection and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}\\)\u003c/span\u003e\u003c/span\u003e migration into HfO\u003csub\u003e2\u003c/sub\u003e-based films lead to domain pinning and to non-polar phases.\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eAmong these electrodes, Ni induces the most severe fatigue in YHO devices, despite having a high work function comparable to Pt and an oxide formation energy akin to W, consistent with its low EFP formation energy discussed above. These vacancies at the interface pin the local Fermi level, causing stronger band bending and a lower Schottky barrier (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec). Consequently, both pronounced \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{V}}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}\\)\u003c/span\u003e\u003c/span\u003e and electron injection accelerates ferroelectric degradation. By contrast, a distinct and stable alumina layer may form at the interface between the Al electrode and YHO film, as Al possesses a lower oxide formation energy than Hf.\u003csup\u003e38\u003c/sup\u003e This passivation interfacial layer would suppress charge transfer and oxygen exchange across the interface, but results in a high coercive field and dielectric breakdown under high electric fields.\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\u003cp\u003eDepth-resolved XPS measurements were performed to obtain the oxygen concentration inside YHO capped with Ni. Figure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ed shows that this concentration increases by ~\u0026thinsp;7% from the Ni/YHO interface towards the YHO/LSMO interface for the pristine device, supporting the significant oxygen scavenging effect of Ni. To elucidate the ferroelectric failure mechanism, the leakage current characteristics were examined in YHO-Ni devices with different electrode areas (314, 490, and 1960 \u0026micro;m\u003csup\u003e2\u003c/sup\u003e), which are most prone to failure. Before fatigue cycling, the leakage current scales linearly with the tested electrode area (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ee and \u003cb\u003efig. S5a-c\u003c/b\u003e). The electrode-area-dependent scaling indicates an interfacial conduction mechanism dominated by charged carriers, e. g., \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}\\)\u003c/span\u003e\u003c/span\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eg). After fatigue failure, the leakage characteristics change markedly, showing hysteresis and resistive switching that are independent of electrode area (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ef and \u003cb\u003efig. S5d-f\u003c/b\u003e). This indicates a transition to filamentary conduction, where repeated cycling drives mobile interfacial \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}\\)\u003c/span\u003e\u003c/span\u003e to nucleate and merge into continuous conductive paths (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eh). This filamentary mechanism is consistent with observations in fatigued HfO\u003csub\u003e2\u003c/sub\u003e-based capacitors, where oxygen vacancies congregate near the electrode and eventually form conductive paths.\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e Notably, a lower \u003cem\u003eE\u003c/em\u003e\u003csub\u003eEFP\u003c/sub\u003e (e. g., Ni) promotes the accumulation of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{O}^{\\bullet\\:\\bullet\\:}\\)\u003c/span\u003e\u003c/span\u003e at the interface, thereby facilitating filament formation under switching cycling (inset of Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ei). After failure, the leakage current increases by nearly three orders of magnitude, consistent with the increased concentration of charged \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{O}^{\\bullet\\:\\bullet\\:}\\)\u003c/span\u003e\u003c/span\u003e in the YHO films, proved by the XPS-determined oxygen concentration after failure, shown in the inset in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ed. Consequently, the conduction transition from interfacial to a filamentary mechanism emerges earlier in YHO-Ni and YHO-Mo devices, as reflected by their reduced critical failure cycles (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ei). In contrast, ferroelectric failure hardly occurs in YHO-Pt devices, which can be attributed to the high chemical stability of Pt and the limited \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{V}_{\\text{O}}^{{\\bullet\\:}{\\bullet\\:}}\\)\u003c/span\u003e\u003c/span\u003e contributing to filament formation.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eElectrode materials play a critical role in the endurance of HfO\u003csub\u003e2\u003c/sub\u003e-based ferroelectric devices. Low-work-function materials accelerate polarisation degradation through enhanced electron injection and oxygen vacancy generation in the ferroelectric film. However, Ni, despite its high work function, exhibits the most severe fatigue with polarisation degradation of ~\u0026thinsp;26% before 10\u003csup\u003e7\u003c/sup\u003e cycling owing to an unusually low extended Frenkel Pair formation energy, which promotes charged vacancy accumulation at the interface and consequently interfacial band bending. At fatigue failure, YHO capacitors undergo a transition from ferroelectric switching to filamentary conduction, evidenced by hysteretic leakage currents and a change from area-dependent to area-independent conduction. The high extended Frenkel pair formation energy of Pt suppresses charged oxygen vacancy formation, thereby enhancing the fatigue resistance and avoiding ferroelectric failure after switching cycles. These findings demonstrate that interfacial charged defects, whose formation is jointly controlled by the metal work function and the oxide formation energy of the metal electrode, control the effective barrier height and fatigue behaviour, providing a pathway to improve endurance and long-term stability in ferroelectric memories.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003eFilm growth and device fabrication of YHO with various metal electrodes\u003c/h2\u003e\u003cp\u003eY-doped HfO\u003csub\u003e2\u003c/sub\u003e (YHO) films were grown by pulsed laser deposition (PLD). The YHO target (Hf\u003csub\u003e0.93\u003c/sub\u003eY\u003csub\u003e0.07\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e) was synthesised by weighing Y\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e and HfO\u003csub\u003e2\u003c/sub\u003e powders according to the stoichiometric ratio. Then, the powders were mixed in a ball miller for 12 h, followed by drying, pelletizing, and sintering at 1400\u0026deg;C for 8 h. A La\u003csub\u003e0.7\u003c/sub\u003eSr\u003csub\u003e0.3\u003c/sub\u003eMnO\u003csub\u003e3\u003c/sub\u003e (LSMO) buffer/seed target was prepared using La\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e, SrCO\u003csub\u003e3\u003c/sub\u003e, and MnO\u003csub\u003e2\u003c/sub\u003e, with the carbonate precursors first calcined above 850\u0026deg;C and subsequently sintered at 1200\u0026deg;C. The deposition process began with the growth of 25 nm-thick LSMO films on (100)-oriented Nb-doped (0.5 wt.%) TiO\u003csub\u003e2\u003c/sub\u003e-terminated SrTiO\u003csub\u003e3\u003c/sub\u003e (Nb: STO) single-crystal substrates. The growth conditions were set at 750\u0026deg;C, 100 mTorr oxygen partial pressure (p\u003csub\u003eO₂\u003c/sub\u003e), 0.7 J/cm\u0026sup2; laser fluence, and a 2 Hz laser frequency. The film thickness was monitored in situ using reflection high-energy electron diffraction (RHEED). YHO films were then deposited under adjusted conditions, with the substrate temperature increased to 890\u0026deg;C, p\u003csub\u003eO₂\u003c/sub\u003e set to 75 mTorr, and the laser fluence maintained at 0.7 J/cm\u0026sup2;. After deposition, p\u003csub\u003eO₂\u003c/sub\u003e was raised to 300 Torr, and the films were held at 890\u0026deg;C for 20 min before being cooled to room temperature at 5 K/min. The YHO film thickness was calibrated based on the growth rate determined from reference samples.\u003c/p\u003e\u003cp\u003eTop electrodes (Pt, Cu, W, Mo, Ni, and Al) were deposited via magnetron sputtering following a standard lift-off process. AZ 4533 UV photoresist layer was spin-coated onto the sample surface at 8000 rpm for 30 s and soft-baked at ~\u0026thinsp;100\u0026deg;C for 2 min. The coated samples were then exposed to UV light for 10 s through a photomask and developed using AZ 351B developer. Then the top metal electrodes were sputter-deposited onto the patterned resist. The lift-off process was performed by submerging the samples in acetone for 5 min, followed by sonication to remove the unexposed resist and lift off the excess metal.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eScanning probe microscopy and X-ray diffraction (XRD) characterisation\u003c/h2\u003e\u003cp\u003eThe surface topography and piezoresponse of the heterostructures were characterised using a Bruker Multimode atomic force microscope with Pt-coated NSC35 tips (MikroMasch). A DC voltage was applied to the sample via the LSMO electrode while grounding the PFM tip, and the piezoresponse was detected using an AC excitation at 15 kHz in the same configuration. The sweep rate was set at 1 Hz. Kelvin Probe Force Microscopy (KPFM) was performed using a Bruker Dimension Icon Pro and with a PtIr-coated Si tip, with a resonance frequency of ~\u0026thinsp;70 kHz. A DC bias was applied to the probe to match the contact potential difference between the probe and the sample surface.\u003c/p\u003e\u003cp\u003eX-ray diffraction measurements, including 2\u0026#120579;-\u0026#120596; scans, \u0026#120596; scans, and \u0026#120593; scans, were carried out using a Panalytical Empyrean Diffractometer with parallel beam optics and CuKα₁ radiation. The system was equipped with either a single-point proportional detector or a PIXcel3D position-sensitive detector. 2\u0026#120579;\u0026ndash;\u0026#120596; scans were performed in scanning line detector mode to achieve high resolution and enhance d-spacing intensity information, while pole figure measurements were conducted in point detector mode, optimized for high scan speed.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eElectrical measurements\u003c/h3\u003e\n\u003cp\u003eFerroelectric polarisation loops were measured at room temperature using an AixACCT TFAnalyzer 2000. A bias voltage was applied to the top electrode (diameter: 20\u0026ndash;50 \u0026micro;m), while the LSMO bottom electrode was grounded. Polarisation hysteresis loops were obtained via positive-up negative-down (PUND) measurements, integrating the current response at 1 kHz. Leakage current was measured using a linear voltage sweep mode with an auto-ranging current detection system to ensure high-resolution measurements. Fatigue performance was assessed by applying a triangular voltage wave with an amplitude of 5 V and a frequency of 100 kHz. PUND measurements were performed at logarithmic cycle intervals to monitor the evolution of remanent polarisation.\u003c/p\u003e\u003cp\u003eX-ray Photoelectron Spectroscopy (XPS)\u003c/p\u003e\u003cp\u003eXPS core levels were determined using a Thermo Scientific Escalab 250Xi instrument operated with monochromatic Al Kα radiation (hν\u0026thinsp;=\u0026thinsp;1486.6 eV) at ~\u0026thinsp;1 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;9\u003c/sup\u003e Torr. The 10-nm-thick Ni top electrodes were kept in electrical contact with the XPS sample holder to prevent possible peak shifts arising from sample charging. Depth-resolved XPS core levels were acquired from 500 \u0026times; 500 \u0026micro;m\u003csup\u003e2\u003c/sup\u003e regions located in the centre of 1 \u0026times; 1 mm\u003csup\u003e2\u003c/sup\u003e sputter-etched areas after gentle surface sputter-etching by 400-eV Ar ions. Elemental depth-resolved quantifications were carried out using CasaXPS software.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003ch2\u003eCompeting interests:\u003c/h2\u003e\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\u003c/p\u003e\u003ch2\u003eAuthor contributions\u003c/h2\u003e\u003cp\u003eY.-X. L., J. S. K., and J. L. M.-D. conceived this work, led the investigation, and developed the methodology. Y.-X. L. deposited thin film, performed and analysed electrical measurements, and contributed to all other analyses. J. S. K. fabricated devices and conducted XRD measurements with J. L. Z. Y. performed KPFM measurements. J. S. K. conducted PFM measurements and H. Y. assisted in fitting and performed SS-PFM characterisations with the support of K. W. XPS measurements were performed by B. B., who analysed the XPS data. J. P. B. S. and K. W. acquired funding. J.L.M.-D. acquired funding, provided overall project supervision measurements. Y.-X. L. drafted the first version of the manuscript. All authors contributed to reviewing and editing the manuscript, and they approved the final version.\u003c/p\u003e\u003ch2\u003eAcknowledgements\u003c/h2\u003e\u003cp\u003eWe gratefully acknowledge discussions with Prof. Uwe Schroeder from NaMLab gGmbH and his valuable advice. This work was supported by the ERC grant EU-H2020-ERC-ADG 882929, EROS. The authors also acknowledge funding by National Nature Science Foundation of China (Nos. 52325204 and 52302148), the Royal Academy of Engineering Chair in Emerging Technologies grant (Grant No. CIET1819-24), and the Portuguese Foundation for Science and Technology (FCT) in the framework of the Strategic Funding UID/04650/2025. J. P. B. S. also thanks FCT for the contract under the FCT Tenure program 1st edition.\u003c/p\u003e\n\u003ch3\u003eData availability\u003c/h3\u003e\n\u003cp\u003eThe data that support the findings of this study are provided in the article and the Supplementary Information. The data sets generated and analysed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eB\u0026ouml;scke T, M\u0026uuml;ller J, Braeuhaus D, Schr\u0026ouml;der U, B\u0026ouml;ttger U (2011) Ferroelectricity in hafnium oxide thin films. Appl Phys Lett 99\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSchroeder U, Park MH, Mikolajick T, Hwang CS (2022) The fundamentals and applications of ferroelectric HfO\u003csub\u003e2\u003c/sub\u003e. Nat Reviews Mater 7:653\u0026ndash;669\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen H et al (2022) HfO\u003csub\u003e2\u003c/sub\u003e-based ferroelectrics: From enhancing performance, material design, to applications. Appl Phys Reviews 9\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDe La P\u0026eacute;rez J, Joanni E, Vilarinho P, Kholkin A (2010) Thickness effect on the dielectric, ferroelectric, and piezoelectric properties of ferroelectric lead zirconate titanate thin films. J Appl Phys 108\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKang S et al (2022) Highly enhanced ferroelectricity in HfO\u003csub\u003e2\u003c/sub\u003e-based ferroelectric thin film by light ion bombardment. Science 376:731\u0026ndash;738\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCheema SS et al (2022) Ultrathin ferroic HfO\u003csub\u003e2\u003c/sub\u003e-ZrO\u003csub\u003e2\u003c/sub\u003e superlattice gate stack for advanced transistors. Nature 604:65\u0026ndash;71\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVishnumurthy P, Alcala R, Mikolajick T, Schroeder U, Antunes LA, Kersch A (2024) Ferroelectric HfO\u003csub\u003e2\u003c/sub\u003e-based capacitors for FeRAM: Reliability from field cycling endurance to retention. In: \u003cem\u003e2024 IEEE International Reliability Physics Symposium (IRPS)\u003c/em\u003e). IEEE\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang Y et al (2023) A stable rhombohedral phase in ferroelectric Hf(Zr)\u003csub\u003e1+x\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e capacitor with ultralow coercive field. Science 381:558\u0026ndash;563\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKim JS et al (2025) Coercive field control in epitaxial ferroelectric Hf\u003csub\u003e0.5\u003c/sub\u003eZr\u003csub\u003e0.5\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e thin films by nanostructure engineering. ACS Appl Mater Interfaces 17:25442\u0026ndash;25450\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKelley KP et al (2023) Ferroelectricity in hafnia controlled via surface electrochemical state. Nat Mater 22:1144\u0026ndash;1151\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLee Y et al (2022) The influence of crystallographic texture on structural and electrical properties in ferroelectric Hf\u003csub\u003e0.5\u003c/sub\u003eZr\u003csub\u003e0.5\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e. J Appl Phys 132\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiao J et al (2019) Grain size engineering of ferroelectric Zr-doped HfO\u003csub\u003e2\u003c/sub\u003e for the highly scaled devices applications. IEEE Electron Device Lett 40:1868\u0026ndash;1871\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYun Y et al (2022) Intrinsic ferroelectricity in Y-doped HfO\u003csub\u003e2\u003c/sub\u003e thin films. Nat Mater 21:903\u0026ndash;909\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLee J et al (2023) Role of oxygen vacancies in ferroelectric or resistive switching hafnium oxide. Nano Convergence 10:55\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRushchanskii KZ, Bl\u0026uuml;gel S, Ležaić M (2021) Ordering of oxygen vacancies and related ferroelectric properties in HfO\u003csub\u003e2-δ\u003c/sub\u003e. Phys Rev Lett 127:087602\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJia S et al (2025) Developing HZO-based superlattices to enhance fatigue‐resistance by charge injection suppression. Adv Funct Mater, 2501470\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEstandia S et al (2019) Engineering ferroelectric Hf\u003csub\u003e0.5\u003c/sub\u003eZr\u003csub\u003e0.5\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e thin films by epitaxial stress. ACS Appl Electron Mater 1:1449\u0026ndash;1457\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eShi S et al (2024) Stabilizing the ferroelectric phase of Hf\u003csub\u003e0.5\u003c/sub\u003eZr\u003csub\u003e0.5\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e thin films by charge transfer. Phys Rev Lett 133:036202\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBaumgarten L et al (2024) Smart design of fermi level pinning in HfO\u003csub\u003e2\u003c/sub\u003e-based ferroelectric memories. Adv Funct Mater 34:2307120\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLee Y, Goh Y, Hwang J, Das D, Jeon S (2021) The influence of top and bottom metal electrodes on ferroelectricity of hafnia. IEEE Trans Electron Devices 68:523\u0026ndash;528\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eStarschich S, Menzel S, B\u0026ouml;ttger U (2017) Pulse wake-up and breakdown investigation of ferroelectric yttrium doped HfO\u003csub\u003e2\u003c/sub\u003e. J Appl Phys 121\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSharia O, Tse K, Robertson J, Demkov AA (2009) Extended Frenkel pairs and band alignment at metal-oxide interfaces. Phys Rev B\u0026mdash;Condensed Matter Mater Phys 79:125305\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDemkov AA (2006) Thermodynamic stability and band alignment at a metal-high-k dielectric interface. Phys Rev B 74:085310\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMemories F (2023) Smart Design of Fermi Level Pinning in HfO\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAkasaka Y et al (2006) Modified oxygen vacancy induced fermi level pinning model extendable to P-metal pinning. Jpn J Appl Phys 45:L1289\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWei Y et al (2018) A rhombohedral ferroelectric phase in epitaxially strained Hf\u003csub\u003e0.5\u003c/sub\u003eZr\u003csub\u003e0.5\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e thin films. Nat Mater 17:1095\u0026ndash;1100\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHill MO et al (2024) Depth-resolved X‐ray photoelectron spectroscopy evidence of intrinsic polar states in HfO\u003csub\u003e2\u003c/sub\u003e‐based ferroelectrics. Adv Mater 36:2408572\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen G, Hou Z, Gong X, Li Q (2008) Effects of Y doping on the structural stability and defect properties of cubic HfO\u003csub\u003e2\u003c/sub\u003e. J Appl Phys 104\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYang Z et al (2009) Role of interface dipole in metal gate/high-k effective work function modulation by aluminum incorporation. Appl Phys Lett 94\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLee H-S, Park H-H (2015) Band structure analysis of La\u003csub\u003e0.7\u003c/sub\u003eSr\u003csub\u003e0.3\u003c/sub\u003eMnO\u003csub\u003e3\u003c/sub\u003e perovskite manganite using a synchrotron. \u003cem\u003eAdvances in Condensed Matter Physics\u003c/em\u003e 746475 (2015)\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFields SS, Jaszewski ST, Lenox MK, Ihlefeld JF (2023) Asymmetric electrode work function customization via top electrode replacement in ferroelectric and field-induced ferroelectric hafnium zirconium oxide thin films. Adv Mater Interfaces 10:2202232\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGreiner MT, Chai L, Helander MG, Tang WM, Lu ZH (2012) Transition metal oxide work functions: the influence of cation oxidation state and oxygen vacancies. Adv Funct Mater 22:4557\u0026ndash;4568\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiu N et al (2024) Precise control of fatigue, wake-up, charge injection, and break-down in Hf\u003csub\u003e0.5\u003c/sub\u003eZr\u003csub\u003e0.5\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e-based ferroelectric memories. Appl Phys Lett 124\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePešić M et al (2016) Physical mechanisms behind the field-cycling behavior of HfO\u003csub\u003e2\u003c/sub\u003e‐based ferroelectric capacitors. Adv Funct Mater 26:4601\u0026ndash;4612\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCao T, Ren G, Shao D-F, Tsymbal EY, Mishra R (2023) Stabilizing polar phases in binary metal oxides by hole doping. Phys Rev Mater 7:044412\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFengler FP et al (2017) Domain pinning: Comparison of hafnia and PZT based ferroelectrics. Adv Electron Mater 3:1600505\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZeng B et al (2024) Polarization fatigue mechanism of laminated hafnium zirconium oxide ferroelectric thin films. Acta Mater 272:119920\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFillot F et al (2003) Investigations of the interface stability in HfO\u003csub\u003e2\u003c/sub\u003e-metal electrodes. Microelectron Eng 70:384\u0026ndash;391\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKang H-K et al (2017) Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e passivation effect in HfO\u003csub\u003e2\u003c/sub\u003e\u0026middot; Al\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e3\u003c/sub\u003e laminate structures grown on InP substrates. ACS Appl Mater Interfaces 9:17526\u0026ndash;17535\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhang Y et al (2021) Evolution of the conductive filament system in HfO\u003csub\u003e2\u003c/sub\u003e-based memristors observed by direct atomic-scale imaging. Nat Commun 12:7232\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Ferroelectrics, Hafnia, Band alignment, Fatigue","lastPublishedDoi":"10.21203/rs.3.rs-7907164/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7907164/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eHafnia-based ferroelectrics have attracted significant interest for their scalability and compatibility with complementary metal-oxide-semiconductor (CMOS) technology for memory applications. However, the imprint and fatigue behaviour remain critical challenges, particularly in electrode-integrated devices. By investigating Y-doped HfO\u003csub\u003e2\u003c/sub\u003e films with various top electrodes (Pt, Cu, W, Mo, Ni, and Al), this study reveals that the fatigue performance is determined by the interplay between the metal work function and chemical stability through extended Frenkel pairs formation at the interface. Moreover, YHO with Ni top electrodes undergoes a unique ferroelectric failure from ferroelectric switching to filamentary conduction under electrical cycling, driven by extended Frenkel pair defects redistribution. Electrodes with high extended Frenkel pair formation energy enable YHO capacitors to maintain nearly unchanged ferroelectric polarization after 10\u003csup\u003e8\u003c/sup\u003e switching cycles. Our findings show that charged defect formation, rather than intrinsic work function, determines interfacial band bending and the Schottky barrier. This work highlights the importance of interface engineering for enhancing the fatigue performance and the reliability of HfO\u003csub\u003e2\u003c/sub\u003e-based ferroelectrics for memory applications.\u003c/p\u003e","manuscriptTitle":"Tailoring the metal-oxide interface for improving fatigue performance in HfO2-based ferroelectrics","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-24 06:55:39","doi":"10.21203/rs.3.rs-7907164/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"982f58eb-88c2-4528-ba93-2d48fd572e6e","owner":[],"postedDate":"October 24th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":56746292,"name":"Physical sciences/Materials science/Materials for devices/Electronic devices"},{"id":56746293,"name":"Physical sciences/Materials science/Materials for devices/Information storage"}],"tags":[],"updatedAt":"2026-04-20T12:30:52+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-24 06:55:39","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7907164","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7907164","identity":"rs-7907164","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}