# Fileset

[10_1002admi.202400627.pdf](https://mdr.nims.go.jp/filesets/f74504dd-746d-4e06-bc12-13171045915b/download)

## Creator

Shinnosuke Yasuoka, [Takao Shimizu](https://orcid.org/0000-0001-9508-7601), Kazuki Okamoto, Nana Sun, Soshun Doko, Naoko Matsui, Toshikazu Irisawa, Koji Tsunekawa, Alexei Gruverman, Hiroshi Funakubo

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Probing of Polarization Reversal in Ferroelectric (Al,Sc)N Films Using Single‐ and Tri‐Layered Structures With Different Sc/(Al+Sc) Ratio](https://mdr.nims.go.jp/datasets/743397e4-f7bd-4036-9564-66062d359946)

## Fulltext

Probing of Polarization Reversal in Ferroelectric (Al,Sc)N Films Using Single‐ and Tri‐Layered Structures With Different Sc/(Al+Sc) RatioRESEARCH ARTICLEwww.advmatinterfaces.deProbing of Polarization Reversal in Ferroelectric (Al,Sc)NFilms Using Single- and Tri-Layered Structures WithDifferent Sc/(Al+Sc) RatioShinnosuke Yasuoka, Takao Shimizu, Kazuki Okamoto, Nana Sun, Soshun Doko,Naoko Matsui, Toshikazu Irisawa, Koji Tsunekawa, Alexei Gruverman,and Hiroshi Funakubo*Wurtzite-(Al,Sc)N films are promising candidates for ferroelectric memorydevices owing to their outstanding properties. However, there are manychallenges on the way to practical applications, including lowering an electricfield required for polarization switching. Understanding the switchingkinetics, especially the starting point of polarization reversal, is key todesigning materials with desired properties. Here, the impact of Scconcentration and segregation on the switching kinetics for (Al,Sc)Ncapacitors is investigated by evaluating time- and field-dependences of theswitching polarization for the tri-layered (Al,Sc)N films with variousSc/(Al+Sc) ratios. The remanent polarization of stacked films slightlydecreased compared to those of the single-layered films with the sameaverage Sc/(Al+Sc) ratio, while their coercive fields depended on the averageSc content in (Al,Sc)N. The ferroelectric switching behavior suggests thepossibility of nucleation originating from the Sc-rich region and the sequentialswitching mechanism for individual layers, which is unique to multilayeredfilms. This shows a possibility that nucleations of the polarization switchingstart not from the interface between the (Al,Sc)N films and the electrodes. Theunique switching kinetics in tri-layered (Al,Sc)N films have provided newinsights into the field of ferroelectric switching in wurtzite-nitrides.1. IntroductionFerroelectric polarization switching is the transition between twostable states of spontaneous polarization induced by an externalS. Yasuoka, K. Okamoto, N. Sun, H. FunakuboDepartment of Materials Science and EngineeringTokyo Institute of TechnologyYokohama 226-8502, JapanE-mail: funakubo.h.aa@m.titech.ac.jpThe ORCID identification number(s) for the author(s) of this articlecan be found under https://doi.org/10.1002/admi.202400627© 2024 The Author(s). Advanced Materials Interfaces published byWiley-VCH GmbH. This is an open access article under the terms of theCreative Commons Attribution License, which permits use, distributionand reproduction in any medium, provided the original work is properlycited.DOI: 10.1002/admi.202400627electric field. Due to the polarizationnonvolatility, materials possessing suchcapability can be used in low-energy andhigh-density integrated memory devices,such as ferroelectric random-accessmemory (FeRAM) and ferroelectrictunnel junction (FTJ).[1–3] Moreover, themultifunctionality of ferroelectrics is alsopromising for novel applications in otherareas, including negative capacitanceand neuromorphic computing.[4,5]However, the development of thenext-generation ferroelectric-baseddevices requires overcoming multiplechallenges related to the enhance-ment of performance and reliability.In contrast to conventional perovskiteferroelectrics, HfO2-based films areCMOS-compatible and exhibit robust po-larization at the nanoscale thickness.[6–8]Polarization switching in hexagonalwurtzite-structured nitride was exper-imentally demonstrated in (Al,Sc)Nfilm by Fichtner et al.[9] The AlN-basedmaterials, which are already widelyused in piezoelectric devices, providesignificant advantages from a practical perspective because theycan be fabricated at room temperature and possess good fer-roelectric properties in films on Si substrate.[10,11] Not only B-and Y-doped AlN films but even GaN-based films have beenT. ShimizuResearch Center for Functional MaterialsNational Institute for Materials ScienceTsukuba 305-0044, JapanT. ShimizuPRESTOJapan Science and Technology Agency4-1-8 Honcho, Kawaguchi-shi, Saitama 332-0012, JapanS. Doko, N. Matsui, T. Irisawa, K. TsunekawaCanon ANELVA Corporation5-1 Kurigi 2-chome, Asao-ku, Kawasaki-shi, Kanagawa 215-8550, JapanA. GruvermanDepartment of Physics and AstronomyUniversity of NebraskaLincoln, NE 68588-0299, USAAdv. Mater. Interfaces 2025, 12, 2400627 2400627 (1 of 8) © 2024 The Author(s). Advanced Materials Interfaces published by Wiley-VCH GmbHhttp://www.advmatinterfaces.demailto:funakubo.h.aa@m.titech.ac.jphttps://doi.org/10.1002/admi.202400627http://creativecommons.org/licenses/by/4.0/http://crossmark.crossref.org/dialog/?doi=10.1002%2Fadmi.202400627&domain=pdf&date_stamp=2024-10-13www.advancedsciencenews.com www.advmatinterfaces.dereported to exhibit ferroelectricity.[12–15] The ferroelectric proper-ties of wurtzite-structured nitrides are strongly influenced by thedopant elements, especially the remanent polarization (Pr) andcoercive field (Ec) of the (Al,Sc)N films, which decrease with in-creasing the Sc/(Al+Sc)N ratio in the film.[9,16] This is due to thechange in the coordination environment of the cations from four-coordinated tetrahedrons to five-coordinated bipyramids that de-scribe the metastable layered hexagonal structure of ScN, result-ing in a flattening of the energy landscape.[17,18] Although it wasshown that the Pr and Ec values of (Al,Sc)N films can be con-trolled by strain-tunable crystal anisotropy, the ferroelectric prop-erties of AlN- and GaN-based films are mainly dominated by thelevel of Sc dopants.[10,19,20]Recently, we have reported the temperature- and frequency-dependent switching kinetics in the (Al0.8Sc0.2)N capacitors,which follow the invariant polarization switching model anddomain-wall motion regime even at high temperatures andfrequencies.[21] The switching behavior of the (Al,Sc)N films canbe fitted by the Kolmogorov-Avrami-Ishibashi (KAI) model de-scribed by the following equation:ΔP = 1 − exp[−(tt0)n](1)where t0 and n are the characteristics switching time and di-mension of the domain growth, respectively.[22] Yazawa et al. re-ported a composite extended KAI model, where the nucleationrate peaks after the growth of switched nuclei.[23] Furthermore,it has been reported that the investigations of the switchingmechanisms, such as observation of polarization switching at theatomic level and prediction of the switching pathway in additionto the switching kinetics.[24] In general, ferroelectric polarizationswitching is expected to occur from inverse domain nuclei gener-ated around the electrode interface.[25] On the other hand, model-ing using density functional theory (DFT) based quantum molec-ular dynamics (QMD) predicted that inhomogeneous Sc distribu-tion reduces the activation barrier.[26] These calculations implythat the insertion of layers with small Ec at the electrode inter-faces may lead to the reduction of the intrinsically large Ec of the(Al,Sc)N systems. Up to now, there are no studies that have inves-tigated the polarization reversal in (Al,Sc)N structures composedof stacked layers with different coercive fields. In addition, theimpact of doping level on the switching kinetics in the (Al,Sc)Nfilms has not been well understood.In this study, we investigated the ferroelectric switching kinet-ics of the single- and tri-layered (Al,Sc)N structures with variousSc/(Al+Sc)N ratios observing changes in the ferroelectric prop-erties and the dimensions of the growing domains.2. Results and Discussion2.1. Crystal StructureSingle-layered films of the 150-nm-thick (Al0.9Sc0.1)N(pure Sc10%) and (Al0.7Sc0.3)N (pure Sc30%) and tri-layered films of [10-nm-thick (Al0.7Sc0.3)N]/[130-nm-thick(Al0.9Sc0.1)N]/[10-nm-thick (Al0.7Sc0.3)N] (Sc30/10/30%)[10-nm-thick (Al0.9Sc0.1)N]/[130-nm-thick (Al0.7Sc0.3)N]/[10-nm-thick (Al0.9Sc0.1)N] (Sc10/30/10%) were deposited on a(111)Pt/TiOx/SiO2/Si substrates as shown in Figure 1. Here,the average Sc/(Al+Sc) ratio of Sc30/10/30% and Sc10/30/10%films were ≈13% and 27%, respectively. The microstructure ofSc30/10/30% film is shown in Figure S1 (Supporting Informa-tion). It can be seen that the grain is connected in the tri-layerand extends continuously from one electrode to the other.We expect there might be no obvious change in the grain orcrystallite size in the tri-layer stack compared to the single-layerstructure.Figure 2a,b show the out-of-plane XRD and in-plane GIXRDpatterns of the (Al,Sc)N films, respectively. Based on the stan-dard powder diffraction patterns of wurtzite AlN (PDF#00-025-1133), the films did not show any other secondary phases. Only00l and h00/hk0 diffraction peaks were observed in the out-of-plane XRD and in-plane GIXRD patterns, respectively, exclud-ing those coming from the electrode and substrate. These resultsindicate that all samples including stacked films consisted of a(001)-out-of-plane-oriented wurtzite structure phase. It should benoted that the peaks from the single-layered films (Sc10% andSc30% films) in the in-plane XRD patterns were almost sym-metrical, whereas the coexistence of two peaks was observed forthe tri-layered films (Sc30/10/30% and Sc10/30/10% films). Thissuggests that the compositional diffusion has occurred withinthe stacked films, which is also ascertained by RBS measure-ments (See Figure S2, Supporting Information). We attemptedto separate peaks in order to investigate the influence of thestrain. Figure 3a,b shows the enlarged 110 diffraction peaks ofSc30/10/30% and Sc10/30/10% films. The asymmetric peaks de-rived from (Al,Sc)N with different Sc/(Al+Sc) ratios were sepa-rated by fitting with a Pearson VII based on the peak positionsof the single-layered films.[27] The difference in peak intensitiesmay contribute to the intensities derived from the respective lay-ers with different Sc content ratios as well as the film thick-ness, considering the decrease in peak intensity with increas-ing Sc/(Al+Sc) ratio of (Al,Sc)N films.[28] Figure 3c shows thein-plane a-axis lattice constant with error bars as a function ofthe average Sc/(Al+Sc) ratio of the entire film. This a-axis latticeconstant was derived from the deconvoluted fits of Figure 3a,b.Based on the in-plane lattice parameters of pure Sc10% and 30%films, it was found that the middle layers, which constitute themajority of the stacked films, were not significantly affected. Incontrast, the minority top and bottom layers were strained bythe stresses applied from the middle layers and/or bottom elec-trodes. Here, the a-axis lattice constant of Sc10% layers in theSc10/30/10% film increased relative to that of the pure Sc10%film, suggesting that the top layer contributes more than the bot-tom layer to the change in the lattice constant of the coveringlayer because the lower layer on Pt is expected to be under com-pressive stress.[29] Even taking into account the stress-based lat-tice changes at this interface, the inserted layers with differentSc/(Al+Sc) ratios are likely to influence the enhancement of nu-cleation at the electrode interface because the Ec of each layer isdifferent.2.2. Ferroelectric PropertiesAs the next step, the ferroelectric properties were investigated forthe four films shown in Figure 1. Figure 4a shows the electricAdv. Mater. Interfaces 2025, 12, 2400627 2400627 (2 of 8) © 2024 The Author(s). Advanced Materials Interfaces published by Wiley-VCH GmbH 21967350, 2025, 5, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/admi.202400627 by National Institute For, Wiley Online Library on [17/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advmatinterfaces.dewww.advancedsciencenews.com www.advmatinterfaces.deFigure 1. Schematic illustrations of a) the 150-nm-thick (Al0.9Sc0.1)N (pure Sc10%), b) 10-nm-thick (Al0.7Sc0.3)N/130-nm-thick (Al0.9Sc0.1)N/10-nm-thick (Al0.7Sc0.3)N (Sc30/10/30%), c) 10-nm-thick (Al0.9Sc0.1)N/130-nm-thick (Al0.7Sc0.3)N/10-nm-thick (Al0.9Sc0.1)N (Sc10/30/10%), and d)150-nm-thick (Al0.7Sc0.3)N (pure Sc30%).Figure 2. a) Out-of-plane XRD and b) in-plane GIXRD patterns for (Al,Sc)N films with various Sc/(Al+Sc) ratios.Adv. Mater. Interfaces 2025, 12, 2400627 2400627 (3 of 8) © 2024 The Author(s). Advanced Materials Interfaces published by Wiley-VCH GmbH 21967350, 2025, 5, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/admi.202400627 by National Institute For, Wiley Online Library on [17/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advmatinterfaces.dewww.advancedsciencenews.com www.advmatinterfaces.deFigure 3. Enlarged views ≈110 diffraction peaks including the results in peak fitting based on the peak position of the pure Sc content films in thea) Sc30/10/30% and b) Sc10/30/10% films. c) Average Sc/(Al+Sc) ratio dependences of strain for single- and tri-layered (Al,Sc)N films.field dependency of the switched polarization values for the re-versal from the upward to downward polarization obtained by thePUND method. The pulse and interval widths of 10 μs and 50 μswere used in the PUND measurement, as shown in the inset. Inthe PUND measurements, although a perfect saturation behavioris absent, a saturation trend can be found for all samples includ-ing the stacked films. This deviation from saturation can be par-tially attributed to the enhanced leakage current under high elec-tric fields.[30] The breakdown field change in (Al,Sc)N films withvarious average Sc/(Al+Sc) ratios was shown in Figure S3 (Sup-porting Information). It can be seen that the EBD of Sc30/10/30%film was on the line of the average composition dependency ofEBD as shown in the solid line. While Sc10/30/10% film increasedslightly against the trend. This increase is in good agreement withthe result of the multilayered architecture of (Al,Sc)N films re-ported by Zheng et al.[31]The relative dielectric constant of these films as a function ofthe average Sc/(Al+Sc) ratio is shown in Figure S4 (SupportingInformation). The relative dielectric constant of the pure 30%film is slightly larger than that of the pure 10% film, which is ingood agreement with our previous work.[29] The relative dielec-tric constant of tri-layers is between the values of pure 10% and30% films, because of the sandwiched layers with different com-positions. Figure 4b shows the 2Pr and Ec as a function of theaverage Sc/(Al+Sc) ratios in the entire (Al,Sc)N films. The 2Prof the stacked films slightly decreased in comparison to the ex-pected values based on the average composition of entire films.On the other hand, their Ec almost continuously depended onthe average Sc/(Al+Sc) ratios for entire average (Al,Sc)N films asshown in Figure 4b, indicating that the impact of the stackinglayer on Ec is not so obvious. The slight decreases in 2Pr valuesof the stacked films are presumably due to charge compensationbetween layers of different compositions or the stress from theupper and lower layer on Pt as shown in the XRD results. Withina certain voltage range, the 2Pr values of the Sc30/10/30% andSc10/30/10% films are higher than those of the pure Sc30% andSc10% films. The dependence of Ec on the average compositionin the entire film suggests that the insertion of a low-Ec layer atthe film-electrode interface, which is generally expected to pro-mote nucleation, has no large effect on decreasing the net Ecof the entire stacked structure. Furthermore, these results im-ply that the nucleation sites that control the Ec may be at the Sc-rich regions of the film, not at the interface because the genera-tion of the nuclei in the low-Ec layer does not significantly con-tribute to the acceleration of polarization switching of the capaci-tor structures. This assumption is supported by the model basedon theoretical calculations suggesting that polarization switchingoccurs at the Sc-rich region[26] and the domain nucleation energyis higher than that of domain growth.[32] Experimentally, the insitu scanning transmission electron microscopy (STEM) observa-tions have confirmed that nucleation occurs in the regions wherethe largest local electric field is applied.[24] Moreover, Ec as a func-tion of Sc content is almost on the same line for Sc-AlN and Sc-GaN.[20] The additional experiment of examining different stackorders with the same % combinations will be conducted in thenext step work.Adv. Mater. Interfaces 2025, 12, 2400627 2400627 (4 of 8) © 2024 The Author(s). Advanced Materials Interfaces published by Wiley-VCH GmbH 21967350, 2025, 5, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/admi.202400627 by National Institute For, Wiley Online Library on [17/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advmatinterfaces.dewww.advancedsciencenews.com www.advmatinterfaces.deFigure 4. a) Switching polarization values obtained by PUND in the(Al,Sc)N films with various Sc/(Al+Sc) ratios as a function of the electricfield. b) Average Sc/(Al+Sc) ration dependences of 2Pr and Ec estimatedfrom (a).2.3. Switching KineticsFigure 5a shows the electric field dependence of the current–time (I--t) curves obtained from the PUND measurements of theSc10% film. The peak representing the ferroelectric switchingcurrent exhibits a sharp shape as the applied electric field in-creases, indicating a high-speed polarization switching. The po-larization switching time (tsw) was defined as the time it takes forpolarization switching to be completed.[25] Figure 5b shows thetsw estimated from the I--t curves of each film as a function ofthe electric field normalized by Ec along with the previously re-ported data for (Hf,Ce)O2, Pb(Zr0.52Ti0.48)O3, and LiNbO3.[33–35]The normalized electric field dependencies of tsw are almost thesame regardless of film stacking, suggesting that the Sc concen-tration and the stack structure had no significant contribution tothe growth speed of the switching domain. The switching of the(Al,Sc)N capacitors for the applied electric fields normalized tothe respective Ec is faster than of the ferroelectrics with differentstructures. The relatively high linearity of (Al,Sc)N film allowsus to expect fast polarization switching at larger E because thedominant mode of switching remains unchanged, as shown bythe dashed line in Figure 5b. Merz’s law is an empirical modeldescribing the relationship between tsw and E as follows:tsw = t0 exp(EaE)(2)where t0 and Ea are the theoretical switching times for an in-finitely strong electric field and the activation field, respectively.It can be seen that Ea for (Al,Sc)N, which exhibits a similar Ectrend with respect to Sc/(Al+Sc) ratio, is significantly larger thanthose of other ferroelectric materials (Figure 5c). The activationenergy is the smallest in the tri-layer Sc10/30/10% film, due tothe well crystalline quality by inserting a pure 10% layer. Becausethe lattice constant of pure 10% film is close to the lattice constantof the underlying Pt bottom electrode, results in a small latticemismatch. Landau–Devonshire thermodynamic modeling wasintroduced for the phenomenological description of the relation-ship between the Gibbs free energy and polarization as shownin Figure 5d. This well depth corresponds to the switching bar-rier height which relates to the intrinsic ferroelectric coercivefield.[36] Based on the reported DFT calculation results, the po-larization dependence of the switching barrier was plotted.[37] Itcan be seen that the increased Sc composition reduced the en-ergy barrier for the switching, which is in good agreement withthe reduced Ea with high Sc composition. The sharp slop of theswitching barrier corresponds to the fast-switching speed of thedomain. It can be seen that (Al,Sc)N film shows a faster switchingspeed and higher material-specific deep energy landscape com-pared to those of the traditional ferroelectric materials, unlikeEc, which is generally promoted by the effects of interfaces anddefects.To reveal the effect of the Sc concentration and the stackingstructure of the switching mechanism, we investigated the fielddependencies of the switching kinetics. Figure 6a–d shows theswitched polarization fraction as a function of time for the differ-ent electric fields for pure Sc10%, Sc30/10/30%, Sc10/30/10%,and pure Sc30% films, respectively. The switching pulse and in-terval widths of 0.5–1000 μs and 2.5–5000 μs were shown inthe inset. The dashed lines correspond to the KAI fitting usingEquation (1). These results indicate that not only the polariza-tion switching occurs in a wide range of electric fields and timescales, but also that the switching behavior of the single-layeredfilms is well-fitted by the KAI model while there is a slight devia-tion from this model in the multilayered films. Figure 6e,f showthe field and the average Sc/(Al+Sc) ratio dependences of the nvalues obtained from the KAI fitting, respectively. The n values inthe pure Sc10% and Sc30% films are ≈2, increasing slightly withan increase in the field, which is also in agreement with the 3Ddomain growth reported by Calderon et al.[24] It indicates the 2Din-plane propagation of the switched domains is a rate determina-tion step for the pure films.[23] On the other hand, the n values ofthe stacked films are almost 1 regardless of the electric field, indi-cating the rate determination step of the 1D out-of-plane propaga-tion of the switched domains. It is possible that the mechanism ofpolarization switching in (Al,Sc)N multilayered films cannot beexplained by the KAI model. Figure S5a,b (Supporting Informa-tion) show the time dependence of ΔP/Pmax alongside the resultsof nucleation-limited-switching (NLS) fitting and a Lorentziandistribution function that describes the dispersion of t0 in individ-ual domains for the Sc30/10/30% films. The figures indicate thatour data can also be well-fitted to the NLS model. However, thefull width at half maximum of the distribution function showsno obvious change with respect to the applied E. The full widthat half maximum previously reported for ferroelectric materialsfitted to the NLS model exhibits a marked dependence on E.[38]Adv. Mater. Interfaces 2025, 12, 2400627 2400627 (5 of 8) © 2024 The Author(s). Advanced Materials Interfaces published by Wiley-VCH GmbH 21967350, 2025, 5, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/admi.202400627 by National Institute For, Wiley Online Library on [17/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advmatinterfaces.dewww.advancedsciencenews.com www.advmatinterfaces.deFigure 5. a) Electric field dependency of the switching current-time curves in (Al0.9Sc0.1)N film. b) tsw as a function of E normalized by Ec and c) Ea in(Al,Sc)N films, together with the reported data for (Hf,Ce)O2,[33] Pb(Zr,Ti)O3,[34] and LiNbO3.[35] d) Schematic illustrations of the switching barrier for(Al,Sc)N.Therefore, the results suggest that the KAI model is sufficient tounderstand the switching kinetics of wurtzite-type (Al0.8Sc0.2)Nfilms in this work.Based on the data in Figure 4, these results imply that the nu-cleation sites that control the Ec may be at the Sc-rich regions ofthe film, not at the interface because the generation of the nucleiin the low-Ec layer does not significantly contribute to the accel-eration of polarization switching of the capacitor structures. Thisassumption is supported by the model based on theoretical cal-culations suggesting that polarization switching occurs at the Sc-rich region[26] and the domain nucleation energy is higher thanthat of domain growth.[32] Experimentally, the in situ scanningtransmission electron microscopy (STEM) observations haveconfirmed that nucleation occurs in the regions where the largestlocal electric field is applied.[24] Moreover, Ec as a function of Sccontent isis almost on the same line for Sc-AlN and Sc-GaN.[20]In addition, the n values of ≈1 followed the 1D out-of-plane prop-agation of the switched domains. And normalized electric fielddependencies of tsw are almost the same regardless of film stack-ing, suggesting that the Sc concentration and the stack structurehad no significant contribution to the growth speed of the switch-ing domain. Following the assumption that the nuclei are notgenerated at the film-electrode interface, the switching mecha-nism in the multilayered films can be attributed to the pseudo1D growth via a two-step switching process: i) subsequent switch-ing in the Sc-poor layers with high Ec; ii) polarization switchingcomplete of the entire film, as schematically shown in Figure 6g.The net polarization switching caused by kinetics in each layersupports the invariance of Ec in the multilayered films shownin Figure 4b. Thus, investigation of the switching kinetics insingle- and tri-layered films has provided new insights into thefield of ferroelectric switching in wurtzite-structured (Al,Sc)Ncapacitors.3. ConclusionThe switching kinetics in the single- and tri-layered (Al,Sc)Nfilms with various Sc/(Al+Sc) ratios were investigated. Theirferroelectric properties and switching rates were evaluated byPUND measurements. The ferroelectric switching behaviors of(Al,Sc)N structures composed of several layers suggest a possi-bility of nucleation originating in the Sc-rich regions in a switch-ing mechanism unique to the stacked films. These results pro-vide a critical step for understanding the switching kinetics ofthe wurtzite-structured ferroelectric nitrides.4. Experimental SectionAll films shown in Figure 1 were deposited on (111)Pt/TiOx/SiO2/Sisubstrates via the dual source reactive sputtering method usingAl (99.999%) and Sc (99.99%) metal targets. The Sc/(Al+Sc) ra-tio and film thickness were controlled by RF powers and deposi-tion time, respectively, and determined by Rutherford backscatter-ing spectrometry (RBS) (HRBS-V500, KOBELCO) and wavelength-dispersive X-ray fluorescence spectrometry (XRF) (PW4400, PANalyt-ical). The results of RBS analysis indicate the fabricated stacked(Al,Sc)N structures as shown in Figure S2 (Supporting Information)Adv. Mater. Interfaces 2025, 12, 2400627 2400627 (6 of 8) © 2024 The Author(s). Advanced Materials Interfaces published by Wiley-VCH GmbH 21967350, 2025, 5, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/admi.202400627 by National Institute For, Wiley Online Library on [17/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advmatinterfaces.dewww.advancedsciencenews.com www.advmatinterfaces.deFigure 6. Electric field dependences of the fraction of the switching polarization as a function of time in the a) pure Sc10%, b) Sc30/10/30%,c) Sc10/30/10%, and d) pure Sc30% films. e) Electric field and f) average Sc/(Al+Sc) ratio dependences of n values obtained from (a–d). g) Schematicillustrations of the polarization switching process by applying an electric field.of the supplementary material. The details of the sputtering condition of(Al,Sc)N were described elsewhere.[16]The crystal structure was characterized by out-of-plane and in-planemeasurements in XRD (X’Pert-MRD, Philips) and grazing-incident XRD(GIXRD) (Smart Lab, Rigaku), respectively. 100-nm-thick Pt top-electrodesof 50 and 100 μm in diameter were fabricated via electron beam evapora-tion through a shadow mask at room temperature. The ferroelectric prop-erties and switching kinetics of Pt/(Al,Sc)N/Pt capacitors were evaluatedby measuring the switched polarization as a function of the applied elec-tric field and time. To determine the correct Pr and Ec values that eliminatethe leakage current contribution, the positive-up-negative-down (PUND)measurements were carried using a pulse and interval width of 10 μs and50 μs. In PUND measurements, the 2Pr was defined as the value 1.1 timesthe inflection point at which the switching polarization value begins to sat-urate, while Ec was defined as the electric field at which the net polarizationwas zero.Supporting InformationSupporting Information is available from the Wiley Online Library or fromthe author.AcknowledgementsThis work was partly supported by the project “Element Strategy Ini-tiative to Form a Core Research Center (JPMXP0112101001)” of MEXT,Adv. Mater. Interfaces 2025, 12, 2400627 2400627 (7 of 8) © 2024 The Author(s). Advanced Materials Interfaces published by Wiley-VCH GmbH 21967350, 2025, 5, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/admi.202400627 by National Institute For, Wiley Online Library on [17/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advmatinterfaces.dewww.advancedsciencenews.com www.advmatinterfaces.deMEXT Initiative to Establish Next-generation Novel Integrated CircuitsCenters (X-NICS)(JPJ011438), MEXT Program: Data Creation and Uti-lization Type Material Research and Development Project (Grant Num-ber: JPMXP1122683430). This work was also partly supported by theJapan Society for the Promotion of Science (JSPS) KAKENHI (Grant No.21H01617, 22K18307, and 22K20427) and by JST PRESTO (Grant Number:JPMJPR20B3), JST ASPIRE (Grant Number: JPMJAP2312), Japan. Supportof the TIT World Research Hub program is gratefully acknowledged.Conflict of InterestThe authors declare no conflict of interest.Data Availability StatementThe data that support the findings of this study are available from the cor-responding author upon reasonable request.Keywords(Al,Sc)N films, multiple layers, Sc concentration, switching kineticsReceived: July 26, 2024Revised: September 9, 2024Published online: October 13, 2024[1] J. F. Scott, C. A. Paz De Araujo, Science 1989, 246, 1400.[2] D. J. Kim, H. Lu, S. Ryu, C. W. Bark, C. B. Eom, E. Y. Tsymbal, A.Gruverman, Nano Lett. 2012, 12, 5697.[3] K. H. Kim, I. Karpov, R. H. Olsson, D. Jariwala, Nat. Nanotechnol.2023, 18, 422.[4] A. I. Khan, K. Chatterjee, B. Wang, S. Drapcho, L. You, C. Serrao, S. R.Bakaul, R. Ramesh, S. Salahuddin, Nat. Mater. 2015, 14, 182.[5] J. Li, C. Ge, J. Du, C. Wang, G. Yang, K. Jin, Adv. Mater. 2020, 32,1905764.[6] K. Uchino, E. Sadanaga, T. Hirose, J. Am. Ceram. Soc. 1989, 72, 1555.[7] P. Gao, Z. Zhang, M. Li, R. Ishikawa, B. Feng, H. Liu, Y. Huang, N.Shibata, X. Ma, S. Chen, J. Zhang, K. Liu, E. Wang, D. Yu, Nat. Com-mun. 2017, 8, 15549.[8] T. Mimura, T. Shimizu, H. Uchida, O. Sakata, H. Funakubo, Appl.Phys. Lett. 2018, 113, 102901.[9] S. Fichtner, N. Wolff, F. Lofink, L. Kienle, B. Wagner, J. Appl. Phys.2019, 125, 114103.[10] S. Yasuoka, T. Shimizu, A. Tateyama, M. Uehara, H. Yamada, M.Akiyama, H. Funakubo, Phys. Status Solidi Appl. Mater. Sci. 2021,2100302.[11] Sung-Lin Tsai, T. Hoshii, H. Wakabayashi, K. Tsutsui, T-K. Chung, E.Y. Chang, K. Kakushima, Appl. Phys. Lett. 2021, 118, 082902.[12] J. Hayden, M. D. Hossain, Y. Xiong, K. Ferri, W. Zhu, M. V. Imperatore,N. Giebink, S. Trolier-Mckinstry, I. Dabo, J. P. Maria, Phys. Rev. Mater.2021, 5, 044412.[13] D. Wang, S. Mondal, J. Liu, M. Hu, P. Wang, S. Yang, D. Wang, Y. Xiao,Y. Wu, T. Ma, Z. Mi, Appl. Phys. Lett. 2023, 123, 033504.[14] D. Wang, P. Wang, B. Wang, Z. Mi, Appl. Phys. Lett. 2021, 119, 111902.[15] M. Uehara, R. Mizutani, S. Yasuoka, T. Shiraishi, T. Shimizu, H.Yamada, M. Akiyama, H. Funakubo, Appl. Phys. Lett. 2021, 119,172901.[16] S. Yasuoka, T. Shimizu, A. Tateyama, M. Uehara, H. Yamada, M.Akiyama, Y. Hiranaga, Y. Cho, H. Funakubo, J. Appl. Phys. 128, 114103.[17] F. Tasnádi, B. Alling, C. Höglund, G. Wingqvist, J. Birch, L. Hultman,I. A. Abrikosov, Phys. Rev. Lett. 2010, 104, 137601.[18] S. Zhang, D. Holec, W. Y. Fu, C. J. Humphreys, M. A. Moram, J. Appl.Phys. 2013, 114, 133510.[19] S. Yasuoka, R. Mizutani, R. Ota, T. Shiraishi, T. Shimizu, M. Uehara,H. Yamada, M. Akiyama, H. Funakubo, ACS Appl. Electron. Mater.2022, 4, 5165.[20] M. Uehara, R. Mizutani, S. Yasuoka, T. Shimizu, H. Yamada, M.Akiyama, H. Funakubo, Appl. Phys. Express 2022, 15, 081003.[21] S. Yasuoka, R. Mizutani, R. Ota, T. Shiraishi, T. Shimizu, K. Okamoto,M. Uehara, H. Yamada, M. Akiyama, H. Funakubo, Appl. Phys. Lett.2023, 123, 202902.[22] Y. Ishibashi, Y. Takagi, J. Phys. Soc. Japan 1971, 31, 506.[23] K. Yazawa, J. Hayden, J. Maria, W. Zhu, S. Trolier-mckinstry, G. L.Brennecka, Mater. Horiz. 2023, 10, 2936.[24] S. C. V., J. Hayden, S. M. Baksa, W. Tzou, S. Trolier-mckinstry, I. Dabo,J. Maria, E. C. Dickey, Science 2023, 380, 1034.[25] W. J. Merz, Phys. Rev. 1954, 95, 690.[26] A. Krishnamoorthy, S. C. Tiwari, A. Nakano, Nanotechnology 2021, 32,49LT02.[27] H. Wang, J. Zhou, J. Appl. Cryst. 2005, 38, 830.[28] M. Akiyama, T. Kamohara, K. Kano, A. Teshigahara, Y. Takeuchi, N.Kawahara, Adv. Mater. 2009, 21, 593.[29] S. Yasuoka, R. Mizutani, R. Ota, T. Shiraishi, T. Shimizu, S. Yasui, Y.Ehara, K. Nishida, M. Uehara, H. Yamada, M. Akiyama, Y. Imai, O.Sakata, H. Funakubo, J. Ceram. Soc. Japan 2022, 130, 436.[30] D. Wang, P. Wang, S. Mondal, J. Liu, M. Hu, M. He, S. Nam, W. Peng,S. Yang, D. Wang, Y. Xiao, Y. Wu, A. Mortazawi, Z. Mia, Appl. Phys.Lett. 2023, 123, 103506.[31] J. X. Zheng, D. Wang, P. Musavigharavi, M. M. A. Fiagbenu, D.Jariwala, E. A. Stach, R. H. Olsson, J. Appl. Phys. 2021, 130, 144101.[32] Y. H. Shin, I. Grinberg, I. W. Chen, Nature 2007, 449, 881.[33] F. Cüppers, K. Hirai, H. Funakubo, Nano Converg 2022, 9, 56.[34] E. Tokumitsu, N. Tanisake, H. Ishiwara, Jpn. J. Appl. Phys. 1994, 33,5201.[35] J. Jiang, X. J. Meng, D. Q. Geng, A. Q. Jiang, J. Appl. Phys. 2015, 117,104101.[36] K. Yazawa, A. Zakutayev, G. L. Brennecka, Appl. Phys. Lett. 2022, 121,042902.[37] H. Wang, N. Adamski, S. Mu, C. G. Walle, J. Appl. Phys. 2021, 130,104101.[38] A. K. Tagantsev, I. Stolichnov, N. Setter, J. S. Cross, M. Tsukada, Phys.Rev. B. 2002, 66, 214109.Adv. Mater. Interfaces 2025, 12, 2400627 2400627 (8 of 8) © 2024 The Author(s). Advanced Materials Interfaces published by Wiley-VCH GmbH 21967350, 2025, 5, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/admi.202400627 by National Institute For, Wiley Online Library on [17/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advmatinterfaces.de Probing of Polarization Reversal in Ferroelectric (Al,Sc)N Films Using Single- and Tri-Layered Structures With Different Sc/(Al80+Sc) Ratio 1. Introduction 2. Results and Discussion 2.1. Crystal Structure 2.2. Ferroelectric Properties 2.3. Switching Kinetics 3. Conclusion 4. Experimental Section Supporting Information Acknowledgements Conflict of Interest Data Availability Statement Keywords