The wurtzite-structure ternary alloys of (Al,Ga,Sc)N films with a high Sc/(Al+Ga+Sc) ratio are successfully deposited by a reactive magnetron sputtering method. The breakthrough in ScN solid-solution limit in AlN with wurtzite structure [up to Sc/(Al + Ga + Sc) = 0.47] is achieved, and its wurtzite-phase diagram region is expanded by adding a relatively small amount of Ga [Ga/(Al+Ga+Sc) < 0.1] into the (Al,Sc)N film. The c- and a-axes lattice constants depend strongly on Ga/(Al+Ga+Sc) and Sc/(Al+Ga+Sc) ratios, respectively, and the Sc/(Al+Ga+Sc) ratio primarily determines the internal structure showing a crystal anisotropy parameter, u parameter. The dielectric constant increases with the increase of Sc/(Al+Ga+Sc) ratio. In addition, the coercive field significantly decreases from 5.8 to 1.8 MV/cm with an increasing Sc/(Al+Ga+Sc) ratio from 0.33 to 0.47, while polarization slightly reduces with the increased u parameter. This result provides a new way to expand the content of the ScN solid solution in the wurtzite film to further decrease the coercive field for the memory device applications.

Recently, there has been a growing interest in non-volatile ferroelectric memories by means of characteristics of ferroelectrics, including maintaining a spontaneous electric polarization without the need for constant external energy and switchable polarization.1 Ferroelectricity of a (Al,Sc)N wurtzite film was experimentally demonstrated in 2019, which exhibited a very large remanent polarization value (Pr, >100 μC cm−2) compared to those of conventional ferroelectrics, such as lead zirconate titanate [Pb(Zr,Ti)O3, PZT] and hafnium-oxide-based ferroelectric thin films.2,3 This provides significant advantages for increasing the capacity of non-volatile memory, which makes it one of the promising candidates for memory applications. Higher stability under the H2-included gas heat treatment compared to PZT and hafnium-oxide-based ferroelectric thin films is also an advantage for highly integrated memory fabrications.4 In addition, only a small degradation in polarization is observed with decreasing film thickness down to 10 nm.5 However, the coercive field (Ec) of the (Al,Sc)N film required for polarization switching is relatively high (∼6 MV/cm), which results in high switching voltage.4 From the viewpoint of the memory applications, reducing the high drive voltage associated with this large Ec is required for the low-power consumption of the memory devices.

Wurtzite-type structure (space group P63mc) ferroelectrics possess a spontaneous polarization along their c-axis, which originates in the separation of the metal and nitrogen atoms in individual planes along the c-axis.6 Alloying is considered as the simplest but most effective way to modify Ec. Various dopants, such as Sc,2,4,7–9 B,3 and Y,10 have been demonstrated to be incorporated into the AlN and GaN films for alloying, which achieves polarization switching or reduces the Ec value. Among them, the Sc dopant is most actively investigated because incorporating ScN into AlN was found to lead to softening, increased a/c-lattice parameter ratio, or improved piezoelectric response.11–13 In addition, the ferroelectric properties of AlN- and GaN-based films are mainly dominated by the content of Sc dopant.2,7–9 Sc incorporation reduces the crystal anisotropy, and the resultant lowering the energy barrier between the two polarization states makes the Ec decrease with the increasing Sc content.2 Therefore, a further increase in the amount of ScN solid solution is considered to be the key to reduce Ec. However, it is difficult to fabricate ferroelectric films in regions with high ScN solid solubility since the pure ScN has a non-polar NaCl structure. The solubility limit of ScN in AlN was experimentally demonstrated to be stable in the wurtzite structure up to 0.43.2 

In order to reduce Ec, we tried to search for a composition system that increases the amount of the ScN solid solution. The (Ga,Sc)N films with wurtzite structure were theoretically and experimentally demonstrated to show higher ScN solid solubility (0.5–0.6) than that of the (Al,Sc)N film.14,15 Therefore, we proposed the ternary alloys (Al,Ga,Sc)N to expand the solid-solution content of ScN. The multiple elements would form solid solutions with random and homogeneous distribution, which might offer additional degrees of freedom for tuning the ferroelectric property of wurtzite films. Simultaneously, the entropy increases with increasing the number of alloys, and the free energy decreases, which is popular as “high-entropy alloys” (HEAs) in metallurgy and is also introduced in the perovskite oxides films to facilitate large electrocaloric effects by promoting polar entropy.16,17 However, the research of ternary (Al,Ga,Sc)N is limited in the composition of Sc/(Al+Ga+Sc)∼0.2 by molecular beam epitaxy.15,18 There was no systematic research on the impact of the film composition on the crystal structure and ferroelectricity, especially by the magnetron sputtering method.

In this work, we successfully prepared the ternary alloy (Al,Ga,Sc)N films by a reactive magnetron sputtering method and systematically investigated the composition dependence of crystal structure and ferroelectricity. Then, the wurtzite- and NaCl-phase diagram for the AlN–GaN–ScN system is established. The solubility of ScN in wurtzite-structure AlN is enhanced, and accordingly, the Ec significantly decreases.

150 nm-thick (Al,Ga,Sc)N films with various compositions were deposited on (111)Pt/TiOx/SiO2/(001)Si substrates by the reactive radio frequency (RF) sputtering method. Al, Sc, and GaN were used as targets, and the substrate temperature was set as 400 °C. The working pressure was kept at 5 mTorr by introducing pure N2 gas. The composition of the film was controlled by adjusting the RF power of the targets. Total RF powers of Sc and Al were fixed at 200 W, and the RF power of GaN was set from 5 to 50 W. Here, the Ga content and Sc content are referred to as x = Ga/(Al+Ga+Sc), and y = Sc/(Al+Ga+Sc) in (Al1−xyGaxScy)N, respectively. In order to prepare the film with uniform compositions, the substrate was rotated.

The out-of-plane and in-plane crystal structures were detected by the x-ray diffraction (XRD) 2θθ scan (X’Pert-MRD, Philips) and the grazing-incident XRD (GIXRD) (Smart Lab, Rigaku), respectively. X-ray fluorescence spectrometry (XRF) (M4 Tornade, Bruker) and Rutherford backscattering spectrometry (RBS) (National Electrostatics Corporation and Charles Evans & Associates) were used to evaluate the composition and thickness by calibration with the standard samples. Scanning transmission electron microscopy (STEM) (JEM-2100F, JEOL) was used to observe the microstructure of the films.

For electrical property characterization, 100 nm-thick Pt top electrode dots with a diameter of 50 μm were fabricated via electron beam evaporation using a shadow mask.4 The positive-up-negative-down (PUND) measurement with pulse width and interval of 10 and 50 μs, respectively, was used to evaluate the switched polarization with the elimination of leakage current by a ferroelectric tester (TOYO, FCE-1A) at room temperature.4 

The summary of the constituent phase against the film composition in the AlN–GaN–ScN system plotted based on the previous work2,14,15,18,19 and obtained XRD data in this work are shown in Fig. 1, which displayed two regions corresponding to the wurtzite- and NaCl-type structure phases. Figure 1(a) shows the estimated one based on the previous studies of binary alloys (Al,Sc)N2,14 and (Ga,Sc)N19 films as well as reported data of the (Al,Ga,Sc)N film.15,18 It was reported that the solubility limits of ScN in AlN with wurtzite structure were about x = 0.34 and 0.43 by Yasuoka et al.14 and Fichtner et al.,2 respectively. While the solubility limits of ScN in GaN with wurtzite structure was about 0.41.19 Based on these results, the boundary line between the wurtzite and NaCl structures is estimated by plotting the dashed line assuming a linear change of the solubility against the composition as shown in Fig. 1(a). However, for ternary alloys (Al,Ga,Sc)N in this work, the wurtzite-structure region is expanded and is much larger than that of the estimated region based on the previous work of binary alloys, as shown in Fig. 1(b). This result indicates that the limit of the solid solution of ScN in a traditional binary system of the previous work is broken through by using the ternary alloys of the (Al,Ga,Sc)N film. This breakthrough might be due to the increased entropy in ternary alloys of (Al,Ga,Sc)N film, which would decrease the free energy and form the solid solutions with random and homogeneous distribution.16,17 

FIG. 1.

Wurtzite- and NaCl-structure phase diagrams for the AlN–GaN–ScN system: (a) estimated one based on the previous studies2,14,15,18,19 and (b) present data plotted based on the XRD results of obtained films in this work.

FIG. 1.

Wurtzite- and NaCl-structure phase diagrams for the AlN–GaN–ScN system: (a) estimated one based on the previous studies2,14,15,18,19 and (b) present data plotted based on the XRD results of obtained films in this work.

Close modal

Taking the four samples (i–iv) marked in the phase diagram of the AlN–GaN–ScN system [Fig. 2(a)] as examples, their out-of-plane and in-plane XRD patterns, as well as the compositions, are shown in Figs. 2(b)2(d). XRD patterns reveal that the films doped with only Sc element [film (i)] take a wurtzite structure with the (001) orientation, and only 002 and hk0 diffraction peaks were detected in the out-of-plane and in-plane XRD patterns, respectively. After co-doping with Ga element in (Al,Sc)N [film (ii)] with the Sc content of 0.33, which is similar to that of film (i), the film still showed wurtzite structure, and the 2θ value of diffraction peaks in both out-of-plane and in-plane XRD patterns was slightly shifted. The solubility of ScN in AlN of films [film (iii)] increased to 0.48 after adding a slight amount of Ga content (x = 0.07), together with the changes of peak intensity and 2θ position of 002 diffraction peak and hk0 diffraction peaks. Importantly, the films still keep the wurtzite structure, and the limit of the Sc content in the previous studies was broken through. The cross-sectional TEM image and diffraction patterns of the film (iii) are shown in Fig. S1 of the supplementary material. A dense film with clear interfaces between the film and Pt electrodes was observed. Selected area electron diffraction acquired at the film marked in the yellow circle area confirms the (002) c-axis orientation, which is in good agreement with the XRD result shown in Fig. 2. The visualization of the atomic arrangements, particularly for compositions near the wurtzite–NaCl phase boundary, will be analyzed in future work. The uniform N, Sc, Al, and Ga concentrations within the film thickness by Rutherford Backscattering Spectrometry (RBS) shown in Fig. S2 of the supplementary material indicates the homogeneity of the ternary alloys across the film thickness. The compositional mapping would be analyzed for further verifying the homogeneity of the ternary alloys across the film thickness in the next step. However, further increasing the Ga content under a high Sc content resulted in the appearance of NaCl-structure [film (iv)], the decreased peak intensity in in-plane data, and the shifted 2θ value of peak diffractions. This result suggests that the solid solution of ScN in AlN can be enhanced to expand the region of the wurtzite structure in the phase diagram by adding a small amount of Ga element.

FIG. 2.

(a) The wurtzite- and NaCl-phase diagrams of the AlN–GaN–ScN system and four marked films [films (i)–(iv)] shown in the inset. (b) Out-of-plane and (c) in-plane XRD patterns of the marked (Al,Ga,Sc)N films in (a) are selectively presented. (d) The details of the compositions of the marked (Al,Ga,Sc)N films.

FIG. 2.

(a) The wurtzite- and NaCl-phase diagrams of the AlN–GaN–ScN system and four marked films [films (i)–(iv)] shown in the inset. (b) Out-of-plane and (c) in-plane XRD patterns of the marked (Al,Ga,Sc)N films in (a) are selectively presented. (d) The details of the compositions of the marked (Al,Ga,Sc)N films.

Close modal

In order to better analyze the composition dependence of the lattice constants in the ternary alloys of (Al,Ga,Sc)N films, the c- and a-axes lattice constants obtained respectively from the out-of-plane and in-plane XRD results of wurtzite structure as a function of various Ga and Sc contents are plotted in Fig. 3. The left and right phase diagrams show the trends of composition changes of the Ga content [Figs. 3(a) and 3(b)] and the Sc content [Figs. 3(c) and 3(d)], respectively, for guiding to the eyes. The data in the previous studies14,15,19–21 and theoretical values of bulk ScN,22 GaN,23 and AlN24 are also shown alongside. The lattice constants of the films with the same Ga content are shown by using the same symbol and color, and the Ga content is presented in the inset for better comparison. Note that the c-axis lattice constant gradually increased as the x value (the Ga content) increased up to 0.6 when the Sc content was similar, which is mainly consistent with the estimated line based on theoretical values of GaN and AlN, following Vegard’s law. This is because the Ga3+ has a larger ionic radius (0.47 Å)25 than that of Al3+ (0.39 Å),26 assuming a larger c-axis lattice parameter in GaN than that of AlN. Besides this change, an increased trend of c-axis lattice constants at various Sc contents is also found at a fixed Ga content because of a larger c-axis lattice parameter in ScN [see theoretical values of the dashed line of bulk ScN in Fig. 3(a)] with a larger ionic radius (0.75 Å) compared to that of GaN.27 This result is in good agreement with the reported data of (Al1−yScy)N14 and (GaxScy)N19 on Pt and Hf bottom electrodes, respectively, and (Al1−yGaxScy)N on GaN substrate.15 However, the a-axis lattice constant as a function of x did not show a clear trend since the theoretical values of GaN and AlN are relatively small, as shown in Fig. 3(b). For the c- and a-axes lattice constants vs Sc content (y) shown in Figs. 3(c) and 3(d), only a value as a function of the Sc content shows a clear growing trend, except the film with the Ga content of 0.5 because of the coexistence of wurtzite and NaCl structures of these films. This trend follows the changes of theoretical values of (Al1−yScy)N20 and (GaxScy)N21 and is also consistent with the reported data of the binary alloys (Al1−yScy)N film14 and (GaxScy)N film19 and ternary alloys (Al1−yGaxScy)N film.15 These results indicate that the c-axis lattice constant is primarily controlled by the Ga content, while the a-axis value is primarily controlled by the Sc content.

FIG. 3.

The left and right phase diagrams show the composition changes with the Ga and Sc contents, respectively. (a) and (c) c-axis and (b) and (d) a-axis as a function of the (a), (b) x = Ga/(Al+Ga+Sc), and (c), (d) y = Sc/(Al+Ga+Sc) in (Al1−xyGaxScy)N, respectively, and the data in the previous work14,15,19–21 are also shown.

FIG. 3.

The left and right phase diagrams show the composition changes with the Ga and Sc contents, respectively. (a) and (c) c-axis and (b) and (d) a-axis as a function of the (a), (b) x = Ga/(Al+Ga+Sc), and (c), (d) y = Sc/(Al+Ga+Sc) in (Al1−xyGaxScy)N, respectively, and the data in the previous work14,15,19–21 are also shown.

Close modal

Internal structure parameter u = (a2/3c2) + 0.25 was reported to be related to the crystal structure anisotropy of the wurtzite structure, where a and c are the lattice constants of the a- and c-axes, respectively, as shown in Fig. 4(a). Figures 4(b) and 4(c) show the u parameter calculated based on the data shown in Fig. 3 as a function of the x (Ga content) and y (Sc content), together with the previous data.14,15,19–21 It can be seen that u strongly depends on the Sc content but not on the Ga content, and it increases with the increase in the Sc content, which is in good agreement with the results of the (Al1−yScy)N film14 and (GaxScy)N film,19 and their calculated values,20,21 as well as ternary alloys (Al1−yGaxScy)N film.15 This can be explained by the relatively large change of c- and a-axes lattice constants against the Sc content as shown in Figs. 3(a) and 3(d), respectively. The film with the Ga content of 0.5 showed a special change, which might be due to the coexistence of wurtzite and NaCl-structure in these films.

FIG. 4.

(a) Crystal structure of wurtzite and equation of u parameter, and u parameter as a function of (b) x = Ga/(Al+Ga+Sc) and (c) y = Sc/(Al+Ga+Sc) in (Al1−xyGaxScy)N is presented, together with the reported data of (Al1−yScy)N film,14 (GaxScy)N film,19 and their calculated values,20,21 as well as ternary alloys (Al1−yGaxScy)N film.15 

FIG. 4.

(a) Crystal structure of wurtzite and equation of u parameter, and u parameter as a function of (b) x = Ga/(Al+Ga+Sc) and (c) y = Sc/(Al+Ga+Sc) in (Al1−xyGaxScy)N is presented, together with the reported data of (Al1−yScy)N film,14 (GaxScy)N film,19 and their calculated values,20,21 as well as ternary alloys (Al1−yGaxScy)N film.15 

Close modal

The dielectric constants as a function of the Ga and Sc contents are investigated as shown in Fig. 5. Reported data for (Al,Ga)N film,28 (Sc,Al)N films,14,29,30 and (Sc,Ga)N film19 are also added in Fig. 5. Note that the dielectric constant is related to the Sc content [see Fig. 5(b)] rather than the Ga content [see Fig. 5(a)], which is due to the weak dependency of Ga/(Al + Ga) in the relative dielectric constant, εAl1-xGaxN = 0.03(1 − x) + 10.28.30 These results of Sc content dependency are in good agreement with the wealth of reports of those of the (Ga,Sc)N film19 and (Al,Sc)N film.14,29,30 

FIG. 5.

Dielectric constant changes with (a) x = Ga/(Al+Ga+Sc) and (b) y = Sc/(Al+Ga+Sc) in (Al1−xyGaxScy)N. Reported data for (Al,Ga)N film,28 (Ga,Sc)N film,19 and (Al,Sc)N films14,29,30 are also added for the comparison.

FIG. 5.

Dielectric constant changes with (a) x = Ga/(Al+Ga+Sc) and (b) y = Sc/(Al+Ga+Sc) in (Al1−xyGaxScy)N. Reported data for (Al,Ga)N film,28 (Ga,Sc)N film,19 and (Al,Sc)N films14,29,30 are also added for the comparison.

Close modal

The PUND measurement is performed to reduce the contribution of the leakage. Current density as a function of time in PUND measurement for the (Al1−xyGaxScy)N [films (i–iv)] marked in the phase diagram of the AlN–GaN–ScN system is shown in Fig. S3 of the supplementary material. In the PUND measurement, the current responses to the N pulse include the contributions from polarization switching, dielectric response, and leakage current, while the current responses to the D pulse do not include switching current. Here, the current density of the absolute value of negative pulse minus down one (N–D) corresponds to the switching current after the elimination of the leakage current. It can be seen that a clear switching current was observed for all films, confirming the ferroelectric switching in these films. The current density decreases with increasing Sc content, corresponding to the decrease of Pr in the high Sc content film that would be explained in Fig. 6. Figure 6 shows the Pr as a function of the E of the marked films (i)–(iv) shown in the phase diagram, which is the same films shown in Fig. 2. It can be seen that for the film doped without Ga [film (i)], the well-saturated Pr against E is observed, but its Ec is relatively high. Importantly, the Ec values were significantly reduced for films (i–iii) with increasing Sc content from 0.33 to 0.48 after slightly co-doping with Ga, while the Pr value decreased with the increase in the Sc content. In addition, the Ga content contribution for the decrease of Ec is also found for films (iii) and (iv), which is possibly due to the appearance of in-plane (001) oriented wurtzite crystals, as shown in Fig. 2(c). It must be noted that the saturation of Pr became worse, especially for the film with the Ga content of 0.28 and the Sc content of 0.47 [film (iv)], and this is considered to be partially attributed to the enhanced leakage current under high electric fields.

FIG. 6.

(a) The selective films marked in the phase diagram of the AlN–GaN–ScN system. (b) PUND data of the marked films (a) as a function of the electric field.

FIG. 6.

(a) The selective films marked in the phase diagram of the AlN–GaN–ScN system. (b) PUND data of the marked films (a) as a function of the electric field.

Close modal

Figure 7(a) displays the film composition region showing Pr value as pink and light pink hatched areas. Note that the ferroelectricity region of the films is also expanded and shows a larger area than the estimated boundary line (light pink hatched area) based on the reported values, as shown in the pink hatched area. Ferroelectricity was not observed for the films with the composition near GaN, as shown in the blue hatched area. This is considered to be mainly due to the relatively large leakage contribution in the PUND measurement, even for the fact that the films consist of a wurtzite structure, as shown in XRD results (Fig. 2). The fundamental reason is mainly attributed to the relatively narrower bandgap of GaN than that of AlN.31 However, the good saturation property was reported by Uehara et al.9 recently for Sc-doped GaN so that this is expected to be improved by the optimization of the deposition condition.

FIG. 7.

(a) Film composition showing Pr value as pink and light pink hatched areas, while non-ferroelectricity as the blue hatched area in the AlN–GaN–ScN system. (b) Pr vs u parameter, and (c) Ec as a function of y = Sc/(Al+Ga+Sc) in (Al1−xyGaxScy)N of the films marked in the phase diagram. Reported data of (Al,Sc)N film14 and (Ga,Sc)N film32 are also shown alongside, and the dashed circle value is referred to as not saturated Pr.

FIG. 7.

(a) Film composition showing Pr value as pink and light pink hatched areas, while non-ferroelectricity as the blue hatched area in the AlN–GaN–ScN system. (b) Pr vs u parameter, and (c) Ec as a function of y = Sc/(Al+Ga+Sc) in (Al1−xyGaxScy)N of the films marked in the phase diagram. Reported data of (Al,Sc)N film14 and (Ga,Sc)N film32 are also shown alongside, and the dashed circle value is referred to as not saturated Pr.

Close modal

Pr and Ec obtained from PUND data of the samples marked in the phase diagram as pink and light pink hatched areas [Fig. 7(a)] are summarized as a function of the u parameter and the Sc content, as shown in Figs. 7(b) and 7(c), respectively. The Pr value mainly decreased as the u parameter increased for the films almost irrespective of the Ga content except for x = 0.3, which is due to the insufficient saturation of Pr as shown in Fig. 6(b). This Pr change is in good matching with the previously reported ones.14,32 On the other hand, Ec decreased significantly from 5.8 to 3.8 MV/cm with the increase of the Sc content almost independent of the Ga content, which is consistent with that of the previous studies of Sc-doped AlN- and GaN-based films.14,32 It is due to the decreased energy barrier for polarization switching because of the increased Sc content, as mentioned in Sec. I. Furthermore, the Ec continues to decrease to 1.8 MV/cm, as the Ga content increases for films (iii) and (iv), suggesting the possible contribution of in-plane (001) oriented wurtzite crystals with 90° domain. This phenomenon is consistent with that of the BaTiO3 film, which has shown that Ec values based on 90° domain wall motion are quite lower than that of 180° domain wall by first-principles density functional theory (DFT) calculations.33 This Ec value is much lower than most of the reported values in the previous work for AlN- and GaN-based wurtzite films with various dopants,4,5,7,14,32,34 which is very promising for the development of memory devices. The reliability testing, such as fatigue and retention, would be conducted to fully assess the potential for memory applications in the next step.

These results show that the Sc content can increase while keeping a wurtzite structure in (Al,Ga,Sc)N films even if the Ga content is relatively small (less than 0.1). As a result, wider composition region having a wurtzite structure was found in ternary (Al,Ga,Sc)N films compared to that of traditional binary films (Fig. 1). A noticeable result in ferroelectricity is that Ec was found to decrease with the increase in the Sc content almost independent of the Ga content. These findings provide a new way to expand the dimension of the ScN solid solution in the wurtzite film to further decrease the coercive field for the memory device applications.

The ternary alloys (Al,Ga,Sc)N wurtzite films with a wide range of compositions were fabricated by the sputtering method. The wurtzite- and NaCl-phase phase diagram in the AlN–GaN–ScN system was established based on the out-of-plane and in-plane XRD data. The limit of the solid solution of ScN in a traditional binary system was broken through by making the ternary alloys (Al,Ga,Sc)N films. c- and a-axes lattice constants are primarily dependent on the Ga and Sc contents, respectively, and the u parameter is mainly determined by the Sc content. The dielectric constant increased with the increase in the Sc content. The Pr decreased as the u parameter increased. The Ec decreased significantly from 5.8 to 1.8 MV/cm with the increase in the Sc content as expected. This Ec value is much lower than most of the reported values in the previous work of various dopants of AlN- and GaN-based wurtzite films, which is very promising for the development of memory devices.

See the supplementary material for the analysis of microstructure and composition, as well as raw data of the PUND measurement.

This work was partly supported by the project “Element Strategy Initiative to Form a Core Research Center (JPMXP0112101001)” of MEXT, MEXT Initiative to Establish Next-generation Novel Integrated Circuits Centers (X-NICS) (JPJ011438), MEXT Program: Data Creation and Utilization Type Material Research and Development Project (Grant No. JPMXP1122683430) and the Japan Science and Technology Agency (JST) as part of Adopting Sustainable Partnerships for Innovative Research Ecosystem (ASPIRE) (JPMJAP2312). This work was also partly supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI (Grant Nos. 21H01617, 22K18307, and 22K20427.

The authors have no conflicts to disclose.

Reika Ota: Data curation (lead); Formal analysis (lead); Investigation (lead); Methodology (lead); Writing – review & editing (equal). Nana Sun: Investigation (equal); Methodology (equal); Writing – original draft (lead); Writing – review & editing (lead). Kazuki Okamoto: Formal analysis (equal); Funding acquisition (equal); Investigation (equal); Methodology (equal); Writing – review & editing (equal). Shinnosuke Yasuoka: Formal analysis (equal); Investigation (equal); Methodology (equal). Yoshihiro Ueoka: Formal analysis (equal); Funding acquisition (equal); Project administration (equal); Resources (equal). Daiki Shono: Formal analysis (equal); Funding acquisition (equal); Project administration (equal); Resources (equal). Masami Mesuda: Formal analysis (equal); Funding acquisition (equal); Project administration (equal); Resources (equal). Hiroshi Funakubo: Funding acquisition (lead); Investigation (equal); Resources (equal); Supervision (lead); Writing – review & editing (lead).

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

2.
S.
Fichtner
,
N.
Wolff
,
F.
Lofink
,
L.
Kienle
, and
B.
Wagner
,
J. Appl. Phys.
125
,
114103
(
2019
).
3.
S.
Calderon V
,
J.
Hayden
,
S. M.
Baksa
,
W.
Tzou
,
S. T.
McKinstry
,
I.
Dabo
,
J. P.
Maria
, and
E. C.
Dickey
,
Science
380
,
1034
(
2023
).
4.
N.
Sun
,
K.
Okamoto
,
S.
Yasuoka
,
S.
Doko
,
N.
Matsui
,
T.
Irisawa
,
K.
Tsunekawa
,
T.
Katase
,
T.
Koganezawa
,
T.
Nakatani
,
R.
Kumara
,
O.
Sakata
, and
H.
Funakubo
,
Appl. Phys. Lett.
125
,
032904
(
2024
).
5.
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
, and
H.
Funakubo
,
J. Ceram. Soc. Jpn.
130
,
436
(
2022
).
6.
F.
Bernardini
,
V.
Fiorentini
, and
D.
Vanderbilt
,
Phys. Rev. B
56
,
R10024(R)
(
1997
).
7.
S.
Yasuoka
,
R.
Mizutani
,
R.
Ota
,
T.
Shiraishi
,
T.
Shimizu
,
M.
Uehara
,
H.
Yamada
,
M.
Akiyama
, and
H.
Funakubo
,
ACS Appl. Electron. Mater.
4
,
5165
(
2022
).
8.
M.
Uehara
,
R.
Mizutani
,
S.
Yasuoka
,
T.
Shiraishi
,
T.
Shimizu
,
H.
Yamada
,
M.
Akiyama
, and
H.
Funakubo
,
Appl. Phys. Lett.
119
,
172901
(
2021
).
9.
M.
Uehara
,
K.
Hirata
,
Y.
Nakamura
,
S. A.
Anggraini
,
K.
Okamoto
,
H.
Yamada
,
H.
Funakubo
, and
M.
Akiyama
,
APL Mater.
12
,
121102
(
2024
).
10.
D.
Wang
,
S.
Mondal
,
J.
Liu
,
M.
Hu
,
P.
Wang
,
S.
Yang
,
D.
Wang
,
Y.
Xiao
,
Y.
Wu
,
T.
Ma
, and
Z.
Mi
,
Appl. Phys. Lett.
123
,
033504
(
2023
).
11.
D.
Jena
,
R.
Page
,
J.
Casamento
,
P.
Dang
,
J.
Singhal
,
Z.
Zhang
,
J.
Wright
,
G.
Khalsa
,
Y.
Cho
, and
H. G.
Xing
,
Jpn. J. Appl. Phys.
58
,
SC0801
(
2019
).
12.
M.
Akiyama
,
T.
Kamohara
,
K.
Kano
,
A.
Teshigahara
,
Y.
Takeuchi
, and
N.
Kawahara
,
Adv. Mater.
21
,
593
(
2009
).
13.
K. R.
Talley
,
S. L.
Millican
,
J.
Mangum
,
S.
Siol
,
C. B.
Musgrave
,
B.
Gorman
,
A. M.
Holder
,
A.
Zakutayev
, and
G. L.
Brennecka
,
Phys. Rev. Mater.
2
,
063802
(
2018
).
14.
S.
Yasuoka
,
T.
Shimizu
,
A.
Tateyama
,
M.
Uehara
,
H.
Yamada
,
M.
Akiyama
,
Y.
Hiranaga
,
Y.
Cho
, and
H.
Funakubo
,
J. Appl. Phys.
128
,
114103
(
2020
).
15.
P.
Wang
,
D.
Wang
,
Y.
Bi
,
B.
Wang
,
J.
Schwartz
,
R.
Hovden
, and
Z.
Mi
,
Appl. Phys. Lett.
120
,
012104
(
2022
).
16.
H.
Wang
,
Q.
He
,
X.
Gao
,
Y.
Shang
,
W.
Zhu
,
W.
Zhao
,
Z.
Chen
,
H.
Gong
, and
Y.
Yang
,
Adv. Mater.
36
,
2305453
(
2024
).
17.
Y.
Son
,
S.
Udovenko
,
S. V. G.
Ayyagari
,
J. P.
Barber
,
K.
Nakamura
,
C. M.
Rost
,
N.
Alem
, and
S.
Trolier-McKinstry
,
Acta Mater.
283
,
120576
(
2025
).
18.
S.
Yang
,
D.
Wang
,
M. M. H.
Tanim
,
D.
Wang
, and
Z.
Mi
,
Appl. Phys. Lett.
124
,
193501
(
2024
).
19.
M.
Uehara
,
T.
Mizuno
,
Y.
Aida
,
H.
Yamada
,
K.
Umeda
, and
M.
Akiyama
,
Appl. Phys. Lett.
114
,
012902
(
2019
).
20.
S.
Zhang
,
W. Y.
Fu
,
D.
Holec
,
C. J.
Humphreys
, and
M. A.
Moram
,
J. Appl. Phys.
114
,
133510
(
2013
).
21.
K.
Hirata
,
Y.
Ikemoto
,
M.
Uehara
,
H.
Yamada
,
S. A.
Anggraini
, and
M.
Akiyama
,
J. Appl. Phys.
135
,
164101
(
2024
).
22.
S.
Limpijumnong
and
W. R. L.
Lambrecht
,
Phys. Rev. B
63
,
104103
(
2001
).
23.
Y.-N.
Xu
and
W. Y.
Ching
,
Phys. Rev. B
48
,
4335
(
1993
).
24.
F.
Litimein
,
B.
Bouhafs
,
Z.
Dridi
, and
P.
Ruterana
,
New J. Phys.
4
(
1
),
64
(
2002
).
25.
V.
Nikolova
,
S.
Angelova
,
N.
Markova
, and
T.
Dudev
,
J. Phys. Chem. B
120
,
2241
(
2016
).
26.
N.
Jahana
,
J. I.
Khandakerb
,
S. I.
Libac
,
S. M.
Hoquec
, and
M. N. I.
Khan
,
J. Alloys Compd.
869
,
159226
(
2021
).
27.
V.
Migliorati
and
P. D.
Angelo
,
Inorg. Chem.
55
,
6703
(
2016
).
28.
O.
Ambacher
,
J.
Majewski
,
C.
Miskys
,
A.
Link
,
M.
Hermann
,
M.
Eickhoff
,
M.
Stutzmann
,
F.
Bernardini
,
V.
Fiorentini
,
V.
Tilak
,
B.
Schaff
, and
L. F.
Eastman
,
J. Phys.: Condens. Matter.
14
,
3399
(
2002
).
29.
T.
Yanagitani
and
M.
Suzuki
,
Appl. Phys. Lett.
105
,
122907
(
2014
).
30.
G.
Wingqvist
,
F.
Tasnádi
,
A.
Zukauskaite
,
J.
Birch
,
H.
Arwin
, and
L.
Hultman
,
Appl. Phys. Lett.
97
,
112902
(
2010
).
31.
A.
Rubio
,
J. L.
Corkill
,
M. L.
Cohen
,
E. L.
Shirley
, and
S. G.
Louie
,
Phys. Rev. B
48
,
11810
(
1993
).
32.
M.
Uehara
,
R.
Mizutani
,
S.
Yasuoka
,
T.
Shimizu
,
H.
Yamada
,
M.
Akiyama
, and
H.
Funakubo
,
Appl. Phys. Express
15
,
081003
(
2022
).
33.
S.
Liu
,
I.
Grinberg
, and
A. M.
Rappe
,
Nature
534
,
363
(
2016
).
34.
K.-H.
Kim
,
I.
Karpov
,
R. H.
Olsson
, and
D.
Jariwala
,
Nat. Nanotechnol.
18
,
441
(
2023
).