www.acsami.org Research Article Magneto-Ionics in Single-Layer Transition Metal Nitrides Julius de Rojas, Joaquín Salguero, Fatima Ibrahim, Mairbek Chshiev, Alberto Quintana, Aitor Lopeandia, Maciej O. Liedke, Maik Butterling, Eric Hirschmann, Andreas Wagner, Llibertat Abad, José L. Costa-Krämer, Enric Menéndez,* and Jordi Sort* Cite This: ACS Appl. Mater. Interfaces 2021, 13, 30826−30834 Read Online ACCESS Metrics & More Article Recommendations * sı Supporting Information ABSTRACT: Magneto-ionics allows for tunable control of magnetism by voltage-driven transport of ions, traditionally oxygen or lithium and, more recently, hydrogen, fluorine, or nitrogen. Here, magneto-ionic effects in single-layer iron nitride films are demonstrated, and their performance is evaluated at room temperature and compared with previously studied cobalt nitrides. Iron nitrides require increased activation energy and, under high bias, exhibit more modest rates of magneto-ionic motion than cobalt nitrides. Ab initio calculations reveal that, based on the atomic bonding strength, the critical field required to induce nitrogen-ion motion is higher in iron nitrides (≈6.6 V nm−1) than in cobalt nitrides (≈5.3 V nm−1). Nonetheless, under large bias (i.e., well above the magneto-ionic onset and, thus, when magneto-ionics is fully activated), iron nitride films exhibit enhanced coercivity and larger generated saturation magnetization, surpassing many of the features of cobalt nitrides. The microstructural effects responsible for these enhanced magneto-ionic effects are discussed. These results open up the potential integration of magneto-ionics in existing nitride semiconductor materials in view of advanced memory system architectures. KEYWORDS: voltage control of magnetism, magneto-ionics, electrolyte gating, transition metals, metal nitrides, nitrogen ions, open volume defects, positron annihilation spectroscopy ■ INTRODUCTION Modern electronic devices store data using electric current to irreversible structural changes. Recently, ionic motion on the order of 103 Hz has been successfully demonstrated with good manipulate the magnetization orientation of magnetic endurance via a proton-based (H+) mechanism, albeit with domains. With device miniaturization pushing nominal device some limitations in the operational film thicknesses and dimensions toward 10 nm, greater amounts of energy are hydrogen retention.11 expended due to resistive heating and device cooling. This Recent studies of the magneto-ionic properties of single- challenge has spurred research toward discovering novel layer thin films with structural oxygen (Co3O4) or nitrogen materials and developing new devices for next-generation (CoN) ions, present in the as-prepared thin-film state, have technologies with improved energy efficiency, robust thermal demonstrated that fully reversible and cyclable magnetic stability, and precise control of magnetic properties. Voltage- transitions between a nonferromagnetic (OFF) and a controlled magnetism (VCM) tackles this challenge by ferromagnetic state (ON) are indeed possible.5,12,24 Interest- replacing electric current with an applied voltage, potentially ingly, and in contrast to the diffusion channels observed in leading to significant energy savings.1,2 Magneto-ionics,3−18 a Co3O4, CoN films transport nitrogen via a planar ion migration branch of VCM in which ions such as O2−, H+, Li+, F−, or front and possess both superior cyclability and lower operating N2−/3− are injected into and withdrawn from a target material voltages than Co3O4,24 hinting that metal nitrides may under an applied bias, has been shown to be capable of compare favorably with their metal oxide counterparts. generating large, nonvolatile, and reproducible modulations of Previous ab initio calculations of the enthalpy of formation bulk magnetic observables. A typical magneto-ionic structure is composed of a ferromagnetic layer in contact with an oxide Received: April 2, 2021 reservoir layer from which oxygen ions are transported, Accepted: June 9, 2021 modifying the target material’s structure and stoichiometry, Published: June 22, 2021 with corresponding changes in coercive field, exchange bias field, magnetic easy axis, or magnetic anisotropy.8−10,12,17,19−23 Cyclability remains an issue as ionic transport may result in © 2021 American Chemical Society https://doi.org/10.1021/acsami.1c06138 30826 ACS Appl. Mater. Interfaces 2021, 13, 30826−30834 ACS Applied Materials & Interfaces www.acsami.org Research Article of CoN have predicted values of ΔHf ≈ −50 kJ mol−1,25 Table 1. Iron and Cobalt Nitride Films, Argon/Nitrogen significantly higher (i.e., less negative) than experimental Partial Pressure during Deposition, Crystallite Size, and estimates conducted on CoO and Co3O4 of ΔHf ≈ −237.9 and Resistivity at Room Temperaturea −910.02 kJ mol−1, respectively.26−28 This is consistent with the PN2 N concentration ⟨D⟩ ρ (T = 300 K) difference in electronegativity between nitrogen and oxygen: Co−N (%) (%) (nm) (μΩ·cm) the lower electronegativity of nitrogen results in weaker bonds FeN−S1 25 49 10 304 with Co cations, suggesting increased magneto-ionic mobility. FeN−S2 50 51 14 509 Properly tuned FeN may be a tantalizing alternative target CoN−S1 25 33 9 156 material for magneto-ionics, as ab initio calculations show that CoN−S2 50 50 13 411 the enthalpy of formation of FeN is comparable to that of CoN a and significantly higher (i.e., less negative) than that of Crystallite sizes obtained from Rietveld refinement of the X-ray diffraction patterns.51 For comparison, the resistivities of pure Fe and FeO.25,29 In addition, magnetic nitrides30,31 such as Fe−N32−37 Co films are 21 and 11 μΩ·cm, respectively (measured at room have recently drawn significant research interest due to their temperature). array of desirable properties, including high hardness, melting point, incompressibility, cost efficiency, and greater magnet- ization than iron oxides,38−43 another class of magneto-ionic target materials.18,21,23 Iron nitrides also span a wide range of mechanical and magnetic properties, which can be tuned by varying the nitrogen concentration in FexNy,44−47 and can be easily integrated with semiconductor electronics. These factors, along with the relative abundance of Fe over Co,48 suggest iron nitride could be a prime candidate for magneto-ionics. In this work, voltage-driven nitrogen transport in iron nitride films is demonstrated, and the magneto-ionic performance is evaluated and compared with cobalt nitride films. The iron nitride films are found to have, under large bias (i.e., well above the magneto-ionic onset voltage and, thus, when magneto- ionics is fully activated), greater total magnetizations, larger coercive fields, lower magneto-ionic rates, and lower (i.e., more negative) onset voltages than the examined cobalt nitrides (−8 vs −4 V). The microstructural effects responsible for the enhanced magneto-ionically induced coercivity and magnet- ization in iron nitride films are discussed, while ab initio calculations reveal that the formation energy of FeN requires a greater critical field (≈6.6 V nm−1) to induce magneto-ionic motion than CoN (≈5.3 V nm−1), consistent with the experimentally observed voltages needed to initiate magneto- Figure 1. Structural, transport, and defect characterization of iron nitride films. (a) θ/2θ XRD diffraction patterns of the as-prepared ionics. ■ iron nitride films. (b) Resistivity ρ measured as a function of temperature from 20 to 300 K for all as-prepared iron nitride films. (c RESULTS AND DISCUSSION and d) Positron lifetime components τi=1−2 and their relative Iron nitride (FeN) and cobalt nitride (CoN) films (85 nm) intensities Ii=1−2 as a function of positron implantation energy EP were grown atop Cu (60 nm)/Ti (20 nm)/[1 0 0]-oriented Si for all as-prepared iron nitride films. substrates. Deposition of iron and cobalt was performed using nitrogen partial pressures of 25% and 50%, respectively, the CoN−S2 film exhibits a single peak which is consistent resulting in differing nanocrystallinities and electrical transport with the (1 1 1) diffraction peak of the expanded Fm3m cubic behavior. To distinguish between the films by nitrogen partial CoN phase (see Figure S1), while CoN−S1 shows a peak pressure and relative electric resistivity, they are labeled FeN− consistent with hexagonal (1 0 1) Co3N1+x phase.50 Crystallite S1, FeN−S2, Co−S1, and Co−S2 (see Table 1). size and nitrogen concentration values are listed for Structural characterization of the films was carried out using comparison (Table 1). θ/2θ X-ray diffraction (XRD). Figure 1a shows the XRD Resistivity measurements as a function of temperature (20− patterns of the as-prepared iron nitride films. The nearly 300 K) were performed on both as-prepared FeN films as well stoichiometric FeN films (i.e., FeN−S1, FeN−S2) exhibit a as purely metallic iron for reference (Figure 1b). The resistivity single peak which is consistent with the (1 1 1) textured ρ at room temperature ranges from ≈21 μΩ·cm in pure Fe to hexagonal close-packed (hcp) P63/mmc FeN diffraction peak ≈500 μΩ·cm in FeN−S2, increasing gradually with N content, (Materials Project ID 12120) or possibly the (0 0 1) face- as expected. The pure iron film shows a monotonic increase of centered cubic (fcc) F4̅3m FeN diffraction peak (Materials resistivity throughout the temperature range (dρ/dT > 0, Project ID 6988).49 The patterns also show the (1 1 1) Cu where ρ and T are resistivity and temperature, respectively), peak arising from the buffer layer. Rietveld refinement of the consistent with metallic behavior. In the case of FeN−S2 FeN XRD patterns reveals that the hcp phases of both films are (Figure 1b), resistivity is observed to monotonically decrease distorted, and the films are highly nanocrystalline, with the (dρ/dT < 0) throughout the temperature range, consistent smallest crystallite sizes in the range of 10−14 nm.. The with semiconducting behavior. In contrast with FeN−S2, nitrogen content is estimated by electron energy loss FeN−S1 has an overall lower resistivity than FeN−S2 with the spectroscopy (EELS) (see Table 1). As reported earlier,24 sign of dρ/dT negative up to 220 K, then basically zero up to 30827 https://doi.org/10.1021/acsami.1c06138 ACS Appl. Mater. Interfaces 2021, 13, 30826−30834 ACS Applied Materials & Interfaces www.acsami.org Research Article 270 K, and finally negative to 300 K. This exhibits a semiconducting contribution and thus a more complex transport behavior than a typical insulating film. These relative differences in behavior mark the FeN−S1 film as relatively conductive and the FeN−S2 film as relatively resistive. CoN− S2 exhibits a similar transition in electrical transport behavior (see Table 1, Figure S2), whereas CoN−S1 is rather conductive, although with a higher electric resistivity than metallic Co. In order to characterize the depth-dependent defect structure of the as-prepared iron nitride films, variable energy positron annihilation lifetime spectroscopy (VEPALS) experi- ments were conducted.52−57 Contributions from positron annihilation lifetimes12 τ1, corresponding to localized vacan- cies, and τ2, corresponding to a mixture of signals from surface states and grain boundaries, are seen in both films (Figure 1c and 1d). The annihilation lifetimes, τi, decrease as the positrons penetrate deeper into the film, ascribed to an increase in vacancy size close to the electrolyte-side surface of the FeN films and a decrease in vacancy size deeper in the Figure 2. Magneto-ionic characterization of iron nitride films by in- films, rendering a graded defect structure. The positron plane vibrating sample magnetometry under electrolyte gating. (a) lifetimes, τ1 and τ2, increase with resistivity, indicating that Schematic of the capacitor-like structure used during the measure- resistivity is tightly related to vacancy and grain boundary ment of each film. Anhydrous liquid electrolyte (LE) forms an electric formation (i.e., the nanostructuring of the films). Examining double layer (EDL) at the film surface. Thin layer of metallic iron is the as-prepared FeN−S1 film, only contributions from τ1 and drawn to represent the appearance of a ferromagnetic layer after τ2 are observed, with no contributions from lifetime τ3 (“void- voltage actuation. (b) First hysteresis loops under −50 V bias for iron like” structures) present.12 τ1, representing small vacancy nitride films. (c) Saturation magnetization (MS) measured as a function of time for iron nitride and cobalt nitride films, taken from clusters, reaches a maximum of ≈0.24 ns near the surface, M−H loops measured between −20 and 20 kOe. (d) θ/2θ XRD before dropping to ≈0.23 ns in the film region (Figure 1c, diffraction patterns of FeN−S1 before and after −50 V biasing open squares). This suggests a higher density larger defect treatment for 75 min. complex size consisting of 3−4 vacancies within a cluster near the top of the film and 2−3 vacancies within a cluster near the working electrode (see Figure S3 for details). τ2, representing a of <0.70% (0.42%) by volume of metallic iron (cobalt), a result convolution of surface states (subsurface region) and grain of off-stoichiometric regions in the film and/or ferromagnetic boundaries, remains above 0.4 ns throughout the film, impurities in the substrate (Figures S5a and S5b). Magneto- indicating the presence of larger vacancy clusters and small ionic rates were assessed by measuring magnetic hysteresis voids near the surface. FeN−S1 shows a residual contribution loops as a voltage of −50 V was applied across the film. from larger vacancy complexes, whereas FeN−S2 exhibits a The first M−H hysteresis loop measured for each iron larger density of open volumes closer to the interface with the nitride film is shown in Figure 2b. The films clearly buffer layer. The relative intensity I1’s more rapid decrease (I2 demonstrate the appearance of a ferromagnetic phase with increases) is consistent with the appearance of the increasing an increase in MS and the coercive field HC. During the first influence of the Cu and Ti layers as the positron implantation loop the magnetization of both FeN−S1 and FeN−S2 clearly energy approaches Ep = 4 keV (Figure 1d). This overall increases, although more in FeN−S1 than in FeN−S2 (Figure behavior of both of the as-prepared iron nitride films in the 2b). VEPALS measurements shows that FeN−S1 and FeN−S2 are The first two loops measured during −50 V biasing are comparable to the as-prepared state of CoN−S2 (Figure S4) plotted in Figure S5c−f, so that the changes in magnetic with slightly lower lifetimes. behavior can be more clearly observed for all films. In Figure All magneto-ionic measurements were preformed using a 2c, the overall change in MS is plotted as a function of time for liquid electrolyte (i.e., propylene carbonate with Na+) in a FeN−S1, FeN−S2, CoN−S1, and CoN−S2. Cobalt nitride capacitor-like configuration (Figure 2a), where the Cu/Ti films are included for ease of comparison. For CoN−S2 and buffer layer acts as a working electrode and the Pt wire acts as a FeN−S2 films, MS quickly increases during the initial stages of counter electrode. The use of a nonaqueous, aprotic polar voltage application (first 12 min, Figure 2c) and subsequently liquid electrolyte will prevent electrochemical oxidation during levels off. Interestingly, FeN−S1 exhibits a distinct multistep voltage biasing.5,10,12,24 This configuration generates an electric process: in the first stage, MS increases at a rate on the order of double layer (EDL) at the film surface which applies a uniform, FeN−S2, which can be ascribed to the formation of a out-of-plane electric field.5,10,58−62 In-plane measurements of ferromagnetic Fe3N phase (Figure 2d), before continuing to M−H hysteresis loops, spanning the range between −20 and increase, albeit at a slower rate, up to a maximum MS (898 emu 20 kOe, were measured using a vibrating sample magneto- cm−3). The MS values reached by each film after long-term meter. The upper and lower branches were measured in 12.5 biasing (6 h) are presented in Table 2. The initial slope of the min for a total of 25 min for each major loop. As-prepared magnetization was fitted using a linear regression with rates FeN−S1, Fe−S2, CoN−S1, and CoN−S2 films show residual found to be 502 emu cm−3 h−1 for FeN−S1 and 267 emu cm−3 ferromagnetic behavior and a saturation magnetization MS h−1 for FeN−S2, evidencing magneto-ionics in FeN, although between 3 and 12 emu cm−3 corresponding to contamination with lower rates than CoN−S1 and CoN−S2 (Table 2). 30828 https://doi.org/10.1021/acsami.1c06138 ACS Appl. Mater. Interfaces 2021, 13, 30826−30834 ACS Applied Materials & Interfaces www.acsami.org Research Article Table 2. Magneto-Ionic Rates, Saturation Magnetization, Squareness (Ratio of Remanence to Saturation Magnetization, MR/ MS), Normalized Slope at the Coercive Field, and Coercive Field of Treated Iron and Cobalt Nitride Filmsa Co−N dM/dt (emu cm−3 h−1) MS (emu cm−3) MR/MS (%) 1/MS dM/dH@HC (kOe−1) HC (Oe) FeN−S1 502 898 87 17 86 FeN−S2 267 137 63 5 140 CoN−S1 1012 90 84 21 65 CoN−S2 2602 637 96 97 17 a Values listed reached after 6 h of gating under −50 V bias. CoN−S2 shows a tremendous increase in MS while sweeping parameters reached by the iron and cobalt nitride films after through the first quadrant of the hysteresis loop under bias, 6 h of biasing at −50 V are reported in Table 2. Broadly, greater than FeN−S1 and FeN−S2. Both cobalt nitride films greater squareness and lower coercive fields indicate a steeper show clear increases in MS as a function of time with rates of slope of the M−H curve and a narrower, more uniform 1012 emu cm−3 h−1 for CoN−S1 and 2602 emu cm−3 h−1 for coercive field distribution.65 FeN−S1 demonstrates lower CoN−S2 and CoN−S1 and CoN−S2 reaching final values of squareness and a higher coercive field than CoN−S2 (Figure MS near 90 and 637 emu cm−3, respectively. S6), correlating with a less uniform coercive field distribution The lower magneto-ionic rates of iron nitrides can be than CoN−S2. The coercive fields of each film also reflect attributed to a convolution of several factors: a slightly lower intrinsic differences in the generation of magnetic states. The enthalpy of formation of FeN, requiring greater energy to coercive fields of FeN−S2 and CoN−S1 monotonically reach break the bonds between iron and nitrogen,25,63 reduced the highest values of the iron and cobalt films, respectively positron lifetimes τ1 and τ2, which can be correlated with (140 and 65 Oe). This can perhaps be attributed to a smaller vacancy clusters and grain boundaries and thus a paramagnetic to ferromagnetic transition or possibly the reduced ionic mobility, and variation in electrical transport formation of a high density of small, isolated clusters properties. An increase in the required energy to separate Fe developing from the superparamagnetic state of small, weakly and N in turn requires a greater applied electric field (voltage) coupled grains in a nonmagnetic nitride matrix. The behavior to the film, reducing the ionic drift velocity at a given voltage of FeN−S1 and CoN−S2 (Figure S6) resembles the well- when compared with cobalt nitride. Smaller vacancies and known variation of coercivity with increasing size of a single grain boundaries (cross sections) reduce the conductivity of particle, first increasing from a superparamagnetic to a single the ionic pathways, requiring greater voltage to transfer ions domain state and then decreasing HC65 as the multidomain through the film, again reducing the drift velocity at a given state is reached. FeN−S1 and CoN−S2 differ, however, in the voltage when compared with cobalt nitride. A higher resistivity maximum coercive field and time scales: FeN−S1 reaches a allows the applied electric field to penetrate deeper into the larger coercive field HC more slowly (148 Oe in 2 h vs 50 Oe film, which most likely boosts magneto-ionic rates. Higher in 12 min in CoN−S2). Both films asymptotically approach a resistivity can also be correlated with increased bonding lower coercivity with increasing biasing time. It should be between Fe and N, thereby increasing the energy required to noted that FeN−S1 reaches a higher coercive peak than the begin ionic motion and thus reducing magneto-ionic rates. A coercive peaks of CoN−S1 and CoN−S2, suggesting that balance between these factors must be achieved to optimize FeN−S1 may be more useful for magnetic memory magneto-ionic performance. Since FeN−S1 shows a greater applications, which often require magnetic stability. In general, magneto-ionic rate than FeN−S2, XRD diffraction was carried both FeN films reached higher HC values than either CoN film on the FeN−S1 film post-treatment (Figure 2d). After biasing throughout the biasing process. This can be indicative of at −50 V for 75 min, FeN−S1 shows traces of a new peak comparatively smaller activated magnetic regions, comprising compatible with an iron nitride with lower nitrogen single- or few-domain structures in mostly denitrided regions, concentration: (1 0 0) Fe3N (PDF 00-001-1236) or (1 0 0) perhaps the result of a less complete, planar ionic diffusion Fe2N (PDF 00-002-1206). Indeed, the increase of magnet- front in the iron nitride films. ization MS of FeN−S1 as nitrogen is removed from the film To further understand the structural and compositional can be correlated with the appearance of a magnetic Fe3N changes the iron nitride films undergo during gating, cross- phase, whose magnetic moment has been calculated ab initio sectional lamellae of the as-prepared and treated FeN−S1 to be on the order of 1.44 μB.64 After long-term voltage films, capped with a protective platinum layer, were application, the MS of FeN−S1 is found to be larger than the characterized by high-angle annular dark-field scanning trans- MS of the Co nitrides. Considering that the ratios of Fe:N and mission electron microscopy (HAADF-STEM) and EELS. Co:N are nearly identical, this can be attributed to the larger Images of FeN−S1 are shown in Figure 3; analogous images of magnetic moment per atom of metallic Fe (2.22 μB) compared CoN−S2 can be found in a previous work.24 The as-prepared to metallic Co (1.72 μB),48 which becomes increasingly present FeN−S1 film is highly nanostructured and isotropic (Figure as more and more nitrogen is removed from the system. 3a). Fe (red) and N (green) are homogeneously distributed in The evolution of the shape of the M−H loops under bias can the films (Figure 3b), similar to the CoN−S2 as-deposited also provide key insights into the differences between the films. After treatment with −50 V for 75 min, FeN−S1 generated ferromagnetic phases. The squareness, defined as the undergoes a moderate change, roughly corresponding to a ratio of the remanence, MR, to MS (MR/MS), the slope of the more nanoporous structure closer to the electrolyte side and a hysteresis loop at the coercive field, HC, normalized to MS highly nanostructured structure near the working electrode (MS−1dM/dH [H = HC]), and HC have been assessed for each side, suggesting a moderate denitriding process (Figure 3c). A branch of the measured M−H loops as a function of time concentration front appears, reflected in the homogeneous (Figure S6a−f). The asymptotic values of the magnetic presence of Fe and N (red and green, respectively) near the 30829 https://doi.org/10.1021/acsami.1c06138 ACS Appl. Mater. Interfaces 2021, 13, 30826−30834 ACS Applied Materials & Interfaces www.acsami.org Research Article Figure 3. Structural and compositional characterization via high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM, left images) and electron energy loss spectroscopy (EELS, right images). HAADF-STEM and elemental EELS mappings corresponding to the areas marked in orange in as-prepared (a and b) FeN−S1and (c and d) FeN−S1 films negatively biased at −50 V for 75 min, respectively. Copper and platinum layers serve as the working electrode and a protective capping layer for lamellae preparation, respectively. Colors corresponding to each element are noted at the bottom of the figure. electrode and the reduced presence of N (orange/red), consistent with the formation of Fe3N near the electrolyte side (Figure 3d). Conversely, CoN−S2 has been shown to Figure 4. Magneto-ionic behavior of iron and cobalt nitrides during undergo a dramatic change with almost complete denitriding onset and recovery experiments. Structural and compositional under −50 V for 75 min but again with a clearly defined characterization in CoN−S1 film using high-angle annular dark-field concentration front appearing parallel to the surface.24 In scanning transmission electron microscopy (HAADF-STEM) and FeN−S1, the migration front is less defined but relatively electron energy loss spectroscopy (EELS). (a and b), Saturation planar without the existence of cross-sectional channels as magnetization (MS) as a function of time under applied onset and happens in other magneto-ionic systems such as Co3O4.12 recovery voltages for FeN−S1, FeN−S2, and CoN−S1, CoN−S2. Understanding the minimum voltage required to initiate HAADF-STEM and elemental EELS mappings corresponding to the magneto-ionic motion is key to evaluating the magneto-ionic areas marked in orange in (c and d) CoN−S2 negatively biased at −4 V and (e and f) CoN−S2 subsequently positively biased at +4 V for performance, so the onset voltages were determined by 75 min, respectively. Colors corresponding to each element are noted monotonically increasing the applied bias in steps of −2 V at the bottom of the figure. (for 25 min each) and noting when the films started to show an increase in ferromagnetic signal (Figure 4a and 4b). Both the magnitude and the duration of the applied voltage affect were completed (Figure 4c−f) to establish how the film the magneto-ionic response, so the measured threshold changes under threshold voltage cycling and, to lesser extent, voltages should be taken as approximate values. Once a what extent nitrogen is still present in the film, since previous threshold voltage was reached (i.e., permanent magnetism is HDAAF-STEM observations made after applying −50 V for 75 induced), the applied voltage was held for 75 min before min showed that all nitrogen had been released to the inverting the polarity (+8 and +4 V, respectively). The onset electrolyte.24 After negative biasing of −4 V for 75 min, CoN− voltage for FeN−S1 and FeN−S2 (−8 V) is higher than that S2 undergoes a very mild change, corresponding to a slightly Co−S1 and CoN−S2 (−4 V), and all films clearly recover the nanoporous structure near the electrolyte side as well as a as-prepared MS under inverted bias (i.e., MS modulation is fully decrease in the nitrogen content. Remarkably, under +4 V reversible). The measured onset voltages reflect a lower biasing, the film is nearly fully recovered with small nitrogen- activation energy for N ion motion in cobalt nitrides than in deficient regions still remaining postrecovery. Additional iron nitrides, presumably due to the lower electronegativity VEPALS measurements were carried out on FeN−S1 and between nitrogen and iron compared to cobalt (1.64 vs. CoN−S2 (Figure S4) for the as-prepared films, films treated at 1.7)66,67 and an increased cohesive energy.25 FeN−S1 shows the onset voltage (−8 and −4 V for CoN and CoN, an excellent ability to completely recover not only the respectively), and films treated at −50 V. It is observed, magnetization of as-grown film but also the squareness and postonset treatment, that τ1 remains very similar to the as- slope at the coercive field (Figure S7), which could make it prepared film (besides the subsurface region, where it useful for practical applications. increases) while the intensity I1 increases slightly across all CoN−S2 films were also examined via HDAAF-STEM and films. This suggests a very mild increase in the number of local EELS after onset and recovery (−4 and +4 V) experiments vacancy clusters under onset biasing, consistent with the 30830 https://doi.org/10.1021/acsami.1c06138 ACS Appl. Mater. Interfaces 2021, 13, 30826−30834 ACS Applied Materials & Interfaces www.acsami.org Research Article limited removal of N from the film. Post −50 V treatment, a voltage (≈6.6 V nm−1) to induce magneto-ionic motion than local peak in the intensity of lifetime τ2 (grain boundaries) is CoN (≈5.3 V nm−1). The larger total magnetization and observed in both systems, characteristic of a migration front coercivity achieved upon biasing show iron-based nitrides are a moving through the sample as nitrogen is removed. viable material for magneto-ionic applications and appealing as To understand the differences in the energetics between iron potential materials in existing nitride semiconductor devices and cobalt nitrides, ab initio calculation of the energy required and memory system architectures. ■ to induce ionic motion was performed. Using the nudged elastic band method (NEB) (Methods), minimum energy METHODS pathways were calculated for the insertion of a nitrogen atom Sample Preparation. Eighty-five nanometer thick FeN was into an iron slab with hcp FeN structure. The obtained total grown by reactive sputtering on boron-doped, highly conducting energy per atom (normalized to the global minimum) is [100], 500 μm thick silicon wafers, previously coated with 20 nm of plotted in Figure 5 as a function of the displacement z of the titanium and 60 nm of copper. The copper was masked during deposition to serve as a working electrode. The expanded FeN films were grown in a homemade triode sputtering system with a base pressure in the range of 10−8 Torr. Ultrahigh vacuum was ensured to minimize oxygen contamination. The target to substrate distances were around 10 cm, and the sputtering rate was around 1 Å s−1. CoN films were grown in a range of nitrogen partial pressure (100% Ar/0% N2, 75% Ar/25% N2, 50% Ar/50% N2) environments at a total pressure of 8 × 10−3 Torr. Magnetic Characterization. Magneto-electric measurements were performed by vibrating sample magnetometry while electrolyte gating the film in a capacitor-like configuration at room temperature. The films are mounted in a homemade electrolytic cell containing anhydrous propylene carbonate with sodium cation-solvated species (5 ppm). The Na+-solvated species in the electrolyte are present to react with any trace amounts of water in the propylene carbonate.69 The magnetic properties of the films were measured in plane while applying different voltages. This was done using a Micro Sense (LOT, Quantum Design) magnetometer with a maximum field of 20 kOe. Figure 5. Ab initio calculations of the threshold energy FeN system. Voltages were applied using an Agilent B2902A power supply Calculated total energy per atom, normalized to the minimum energy between the working electrode and the counter electrode, as value, as a function of the displacement between the outermost Fe demonstrated in previous works.12,61,69 The magnetic signal was surface atom and the inserted N atom for both hexagonal close- normalized to the volume exposed to the electrolyte during the gating packed (hcp) structured Fe slabs. Five-monolayer-thick Fe slab process. All measured hysteresis loops were background corrected, schematic is shown in the right panel, where the dashed line indicates carried out at high fields (always above the saturation field), to the reference z position, which is the outermost Fe surface monolayer. eliminate linear contributions (paramagnetic or diamagnetic signals). Structural and Compositional Measurements. The θ/2θ X- ray diffraction patterns were recorded on a Materials Research nitrogen atom from the iron reference monolayer (top layer). Diffractometer (MRD) from the Malvern PANalytical Co., equipped Setting the outermost iron layer as z = 0 Å, the global energy with a PIXcel1D detector, using Cu Kα radiation. The XRD patterns minimum for hcp is found at z = 1.15 Å with another local were analyzed using Rietveld refinement to obtain lattice cell minimum located around z = −0.91 Å. The critical electric parameters and crystallite size (average size of coherently diffracting field, EC, can be estimated using the electric potential per atom domains).51 between minima (1.38 eV atom−1) required to move a High-resolution transmission electron microscopy (HRTEM), high-angle annular dark-field scanning transmission electron micros- nitrogen atom between minima, leading to EC ≈ 6.6 V nm−1, copy (HDAAF-STEM), and electron energy loss spectroscopy similar to the values observed for onset in FeN−S1 and FeN− (EELS) were performed on a TECNAI F20 HRTEM/STEM S2 (−8 V, see Figure 4a) and larger (in absolute values) than microscope operated at 200 kV. Cross-sectional lamellae were the critical electrical fields needed to induce nitrogen ion prepared by focused ion beam, capped with a sacrificial platinum motion in CoN. These results complement recent calculations layer, and placed onto a Cu transmission electron microscopy grid. which show CoN has a lower calculated energy barrier24 as Transport Measurements. Both cobalt nitride and iron nitride well as previous simulations where FeN was found to have a films were deposited onto high-resistivity Si substrates. Resistivity (slightly) higher cohesive energy than CoN.25,68 values were acquired from 30 to 300 K, all using the van der Pauw ■ configuration. CONCLUSION Variable Energy Positron Annihilation Lifetime Spectros- copy. Variable energy positron annihilation lifetime spectroscopy Magneto-ionics has been demonstrated in single-layer, 85 nm (VEPALS) measurements were performed at the monoenergetic thick near-stoichiometric FeN films through electrolyte gating, positron source (MePS) at the radiation source ELBE (Electron Linac capable of controllably generating and removing a ferromag- for beams with high Brilliance and low Emittance) at Helmholtz- netic state (ON−OFF ferromagnetism). The magneto-ionic Zentrum Dresden-Rossendorf (Germany).52 A CeBr3 detector properties of FeN are compared with CoN. FeN is found to coupled to a digital lifetime spectrometer with homemade software have a greater total magnetization, a higher coercivity, and a employing a SPDevices ADQ14DC-2X with 14-bit vertical resolution and 2 GS s−1 (gigasamples per second) horizontal resolution was lower rate of magnetization generation under larger bias as well used. The time resolution function was estimated to be about 0.205 as a higher onset voltage than the examined cobalt nitrides. ns. The resolution function required for spectrum analysis uses two The available open volume supporting ion transport is slightly Gaussian functions with distinct intensities depending on the positron lower for FeN as indicated by PALS. Ab initio calculations implantation energy, Ep, and appropriate energy shifts. All spectra show that the formation energy of FeN requires a greater onset measured contain at least 107 counts. 30831 https://doi.org/10.1021/acsami.1c06138 ACS Appl. Mater. Interfaces 2021, 13, 30826−30834 ACS Applied Materials & Interfaces www.acsami.org Research Article Ab Initio Calculations. The first-principles calculations were Llibertat Abad − Institut de Microelectrònica de Barcelona, based on the projector-augmented wave (PAW)70 method as IMB-CNM (CSIC), Barcelona E-08193, Spain implemented in the VASP package71−73 using the generalized José L. Costa-Krämer − IMN-Instituto de Micro y gradient approximation.74 The virtual crystal approximation75 was Nanotecnología (CNM-CSIC), Madrid 28760, Spain used to model the variation of nitrogen per unit cell. To calculate the Fe−N formation energy, the nudged elastic band method (NEB)76,77 Complete contact information is available at: on the nitrogen pathway into a five-monolayer thick (0001) hexagonal https://pubs.acs.org/10.1021/acsami.1c06138 close-packed Fe slab was used. At each step, the atomic coordinates were relaxed until the forces became smaller than 1 meV Å−1. A Notes kinetic energy cutoff of 500 eV was used for the plane-wave basis set 25 × 25 × 1. k-point meshes were used to construct the Brillouin zone The authors declare no competing financial interest. in the Fe slab in the NEB calculations. The data used in this article are available from the corresponding authors upon request. ■ ACKNOWLEDGMENTS ■ Financial support by the European Research Council (SPIN- ASSOCIATED CONTENT PORICS 2014-Consolidator Grant, Agreement No. 648454, and the MAGIC-SWITCH 2019-Proof of Concept Grant, * sı Supporting Information Agreement No. 875018), the Spanish Government The Supporting Information is available free of charge at (MAT2017-86357-C3-1-R), the Generalitat de Catalunya https://pubs.acs.org/doi/10.1021/acsami.1c06138. (2017-SGR-292 and 2018-LLAV-00032), the European X-ray diffraction patterns of CoN films, resistivity Regional Development Fund (MAT2017-86357-C3-1-R and measurements on CoN films, positron lifetime experi- 2018-LLAV-00032), and the French ANR (ANR-18-CE24- ments for FeN and CoN films, hysteresis loops and 0017 “FEOrgSpin” and ELECSPIN ANR-16-CE24-0018 magnetoelectric data for CoN films, and magnetoelectric “ELECSPIN”) is acknowledged. This work was partially recovery data for FeN films (PDF) supported by the Impulse- und Net-working fund of the ■ AUTHOR INFORMATION Corresponding Authors Helmholtz Association (FKZ VH-VI-442 Memriox) and the Helmholtz Energy Materials Characterization Platform (03ET7015). The PALS measurements were carried out at Enric Menéndez − Departament de Física, Universitat ELBE at the Helmholtz-Zentrum Dresden-Rossendorf e. V., a Autònoma de Barcelona, Cerdanyola del Vallès E-08193, member of the Helmholtz Association. We thank the facility Spain; orcid.org/0000-0003-3809-2863; staff for assistance. L.A. thanks MINECO for a Ramón y Cajal Email: enric.menendez@uab.cat Contract (RYC-2013-12640). Jordi Sort − Departament de Física, Universitat Autònoma de Barcelona, Cerdanyola del Vallès E-08193, Spain; Institució Catalana de Recerca i Estudis Avançats (ICREA), Barcelona ■ REFERENCES (1) Hu, J. M.; Nan, C. W. Opportunities and Challenges for E-08010, Spain; orcid.org/0000-0003-1213-3639; Magnetoelectric Devices. APL Mater. 2019, 7 (8), 080905. Email: jordi.sort@uab.cat (2) Song, C.; Cui, B.; Li, F.; Zhou, X.; Pan, F. Recent Progress in Voltage Control of Magnetism: Materials, Mechanisms, and Perform- Authors ance. Prog. Mater. Sci. 2017, 87, 33−82. Julius de Rojas − Departament de Física, Universitat (3) Vasala, S.; Jakob, A.; Wissel, K.; Waidha, A. I.; Alff, L.; Clemens, Autònoma de Barcelona, Cerdanyola del Vallès E-08193, O. Reversible Tuning of Magnetization in a Ferromagnetic Spain; orcid.org/0000-0002-1206-4744 Ruddlesden-Popper-Type Manganite by Electrochemical Fluoride- Joaquín Salguero − IMN-Instituto de Micro y Nanotecnología Ion Intercalation. Adv. Electron. Mater. 2020, 6 (2), 1900974. (CNM-CSIC), Madrid 28760, Spain (4) Tan, A. J.; Huang, M.; Sheffels, S.; Büttner, F.; Kim, S.; Hunt, A. Fatima Ibrahim − Univwesity of Grenoble Alpes, CEA, CNRS, H.; Waluyo, I.; Tuller, H. L.; Beach, G. S. D. Hydration of Grenoble 38000, France Gadolinium Oxide (Gd Ox) and Its Effect on Voltage-Induced Co Mairbek Chshiev − Univwesity of Grenoble Alpes, CEA, Oxidation in a Pt/Co/Gd Ox/Au Heterostructure. Phys. Rev. Mater. CNRS, Grenoble 38000, France; Institut Universitaire de 2019, 3 (6), 064408. (5) de Rojas, J.; Quintana, A.; Lopeandía, A.; Salguero, J.; Costa- France, Paris 75231, France; orcid.org/0000-0001-9232- Krämer, J. L.; Abad, L.; Liedke, M. O.; Butterling, M.; Wagner, A.; 7622 Henderick, L.; Dendooven, J.; Detavernier, C.; Sort, J.; Menéndez, E. Alberto Quintana − Department of Physics, Georgetown Boosting Room-Temperature Magneto-Ionics in a Non-Magnetic University, Washington, District of Columbia 20057, United Oxide Semiconductor. Adv. Funct. Mater. 2020, 30 (36), 2003704. States; orcid.org/0000-0002-9813-735X (6) Dasgupta, S.; Das, B.; Li, Q.; Wang, D.; Baby, T. T.; Indris, S.; Aitor Lopeandia − Departament de Física, Universitat Knapp, M.; Ehrenberg, H.; Fink, K.; Kruk, R.; Hahn, H. Toward On- Autònoma de Barcelona, Cerdanyola del Vallès E-08193, and-Off Magnetism: Reversible Electrochemistry to Control Magnetic Spain; Catalan Institute of Nanoscience and Nanotechnology Phase Transitions in Spinel Ferrites. Adv. Funct. Mater. 2016, 26 (41), (ICN2), Barcelona E-08193, Spain 7507−7515. Maciej O. Liedke − Institute of Radiation Physics, Helmholtz- (7) Dasgupta, S.; Das, B.; Knapp, M.; Brand, R. A.; Ehrenberg, H.; Zentrum Dresden−Rossendorf, Dresden 01328, Germany Kruk, R.; Hahn, H. Intercalation-Driven Reversible Control of Magnetism in Bulk Ferromagnets. Adv. Mater. 2014, 26 (27), Maik Butterling − Institute of Radiation Physics, Helmholtz- 4639−4644. Zentrum Dresden−Rossendorf, Dresden 01328, Germany (8) Bauer, U.; Yao, L.; Tan, A. J.; Agrawal, P.; Emori, S.; Tuller, H. Eric Hirschmann − Institute of Radiation Physics, Helmholtz- L.; Van Dijken, S.; Beach, G. S. D. Magneto-Ionic Control of Zentrum Dresden−Rossendorf, Dresden 01328, Germany Interfacial Magnetism. Nat. Mater. 2015, 14 (2), 174−181. Andreas Wagner − Institute of Radiation Physics, Helmholtz- (9) Baldrati, L.; Tan, A. J.; Mann, M.; Bertacco, R.; Beach, G. S. D. Zentrum Dresden−Rossendorf, Dresden 01328, Germany Magneto-Ionic Effect in CoFeB Thin Films with in-Plane and 30832 https://doi.org/10.1021/acsami.1c06138 ACS Appl. Mater. Interfaces 2021, 13, 30826−30834 ACS Applied Materials & Interfaces www.acsami.org Research Article Perpendicular-to-Plane Magnetic Anisotropy. Appl. Phys. Lett. 2017, Pascals) Pressure and at Higher Temperatures. U.S. Geol. Surv. Bull. 110 (1), 012404. 1995, 1−461. (10) Navarro-Senent, C.; Fornell, J.; Isarain-Chávez, E.; Quintana, (27) Glasser, L.; Sheppard, D. A. Cohesive Energies and Enthalpies: A.; Menéndez, E.; Foerster, M.; Aballe, L.; Weschke, E.; Nogués, J.; Complexities, Confusions, and Corrections. Inorg. Chem. 2016, 55 Pellicer, E.; Sort, J. Large Magnetoelectric Effects in Electrodeposited (14), 7103−7110. Nanoporous Microdisks Driven by Effective Surface Charging and (28) Chase, M. NIST-JANAF Thermochemical Tables, 4th ed.; Magneto-Ionics. ACS Appl. Mater. Interfaces 2018, 10 (51), 44897− American Institute of Physics, 1998. 44905. (29) Tessier, F.; Navrotsky, A.; Niewa, R.; Leineweber, A.; Jacobs, (11) Tan, A. J.; Huang, M.; Avci, C. O.; Büttner, F.; Mann, M.; Hu, H.; Kikkawa, S.; Takahashi, M.; Kanamaru, F.; DiSalvo, F. J. W.; Mazzoli, C.; Wilkins, S.; Tuller, H. L.; Beach, G. S. D. Magneto- Energetics of Binary Iron Nitrides. Solid State Sci. 2000, 2 (4), Ionic Control of Magnetism Using a Solid-State Proton Pump. Nat. 457−462. Mater. 2019, 18 (1), 35−41. (30) Coey, J. M. D.; Smith, P. A. I. Magnetic Nitrides. J. Magn. (12) Quintana, A.; Menéndez, E.; Liedke, M. O.; Butterling, M.; Magn. Mater. 1999, 200 (1−3), 405−424. Wagner, A.; Sireus, V.; Torruella, P.; Estradé, S.; Peiró, F.; (31) Jhi, S. H.; Louie, S. G.; Cohen, M. L.; Morris, J. W. Mechanical Dendooven, J.; Detavernier, C.; Murray, P. D.; Gilbert, D. A.; Liu, Instability and Ideal Shear Strength of Transition Metal Carbides and K.; Pellicer, E.; Nogues, J.; Sort, J. Voltage-Controlled ON-OFF Nitrides. Phys. Rev. Lett. 2001, 87 (7), 075503. Ferromagnetism at Room Temperature in a Single Metal Oxide Film. (32) Wriedt, H. A.; Gokcen, N. A.; Nafziger, R. H. The Fe-N (Iron- ACS Nano 2018, 12 (10), 10291−10300. Nitrogen) System. Bull. Alloy Phase Diagrams 1987, 8 (4), 355−377. (13) Molinari, A.; Hahn, H.; Kruk, R. Voltage-Control of Magnetism (33) Jack, K. H. Binary and Ternary Interstitial Alloys I. The Iron- in All-Solid-State and Solid/Liquid Magnetoelectric Composites. Adv. Nitrogen System: The Structures of Fe 4 N and Fe 2 N. Proc. R. Soc. Mater. 2019, 31 (26), 1806662. London, Ser. A. Math. Phys. Sci. 1948, 195 (1040), 34−40. (14) Ibrahim, F.; Hallal, A.; Dieny, B.; Chshiev, M. Establishing (34) Jack, K. H. The Occurrence and the Crystal Structure of α”- Characteristic Behavior of Voltage Control of Magnetic Anisotropy by Iron Nitride; a New Type of Interstitial Alloy Formed during the Ionic Migration. Phys. Rev. B: Condens. Matter Mater. Phys. 2018, 98 Tempering of Nitrogen-Martensite. Proc. R. Soc. London, Ser. A. Math. (21), 214441. Phys. Sci. 1951, 208 (1093), 216−224. (15) Gößler, M.; Albu, M.; Klinser, G.; Steyskal, E. M.; Krenn, H.; (35) Jack, K. H. The Iron-Nitrogen System: The Crystal Structures Würschum, R. Magneto-Ionic Switching of Superparamagnetism. of α-Phase Iron Nitrides. Acta Crystallogr. 1952, 5 (4), 404−411. Small 2019, 15 (46), 1904523. (36) Jack, K. H. Results of Further X-Ray Structural Investigations of (16) Gilbert, D. A.; Olamit, J.; Dumas, R. K.; Kirby, B. J.; Grutter, A. the Iron-Carbon and Iron-Nitrogen Systems and of Related Interstitial J.; Maranville, B. B.; Arenholz, E.; Borchers, J. A.; Liu, K. Controllable Alloys. Acta Crystallogr. 1950, 3 (5), 392−394. Positive Exchange Bias via Redox-Driven Oxygen Migration. Nat. (37) Kardonina, N. I.; Yurovskikh, A. S.; Kolpakov, A. S. Commun. 2016, 7, 11050. Transformations in the Fe - N System. Met. Sci. Heat Treat. 2011, (17) Gilbert, D. A.; Grutter, A. J.; Arenholz, E.; Liu, K.; Kirby, B. J.; 52 (9−10), 457−467. Borchers, J. A.; Maranville, B. B. Structural and Magnetic Depth (38) Zhang, Y.; Wang, Q. Magnetic-Plasmonic Dual Modulated Profiles of Magneto-Ionic Heterostructures beyond the Interface FePt-Au Ternary Heterostructured Nanorods as a Promising Nano- Limit. Nat. Commun. 2016, 7, 12264. Bioprobe. Adv. Mater. 2012, 24 (18), 2485−2490. (18) Duschek, K.; Uhlemann, M.; Schlörb, H.; Nielsch, K.; Leistner, (39) Zhang, X.; Yang, M.; Jiang, Y.; Allard, L. F.; Wang, J. P. K. Electrochemical and in Situ Magnetic Study of Iron/Iron Oxide Thermal Stability of Partially Ordered Fe16N2 Film on Non- Films Oxidized and Reduced in KOH Solution for Magneto-Ionic Switching. Electrochem. Commun. 2016, 72, 153−156. Magnetic Ag under Layer. J. Appl. Phys. 2014, 115 (17), 17A767. (19) Wolf, S. A.; Awschalom, D. D.; Buhrman, R. A.; Daughton, J. (40) Mahmoudi, M.; Shokrgozar, M. A. Multifunctional Stable M.; von Molnár, S.; Roukes, M. L.; Chtchelkanova, A. Y.; Treger, D. Fluorescent Magnetic Nanoparticles. Chem. Commun. 2012, 48, M. Spintronics: A Spin-Based Electronics Vision for the Future. 3957−3959. Science 2001, 294 (5546), 1488−1495. (41) Liu, J.; Zhang, Q.; Yu, H.; Sun, F. Magnetic Properties of the (20) Eerenstein, W.; Mathur, N. D.; Scott, J. F. Multiferroic and Ising Ferromagnetic Iron Nitride System. J. Magn. Magn. Mater. 2005, Magnetoelectric Materials. Nature 2006, 442, 759−765. 288, 48−53. (21) Duschek, K.; Petr, A.; Zehner, J.; Nielsch, K.; Leistner, K. All- (42) Hang, X.; Matsuda, M.; Held, J. T.; Mkhoyan, K. A.; Wang, J. P. Electrochemical Voltage-Control of Magnetization in Metal Oxide/ Magnetic Structure of Fe16 N2 Determined by Polarized Neutron Metal Nanoislands. J. Mater. Chem. C 2018, 6 (31), 8411−8417. Diffraction on Thin-Film Samples. Phys. Rev. B: Condens. Matter (22) Bi, C.; Liu, Y.; Newhouse-Illige, T.; Xu, M.; Rosales, M.; Mater. Phys. 2020, 102 (10), 104402. Freeland, J. W.; Mryasov, O.; Zhang, S.; Te Velthuis, S. G. E.; Wang, (43) Pavanati, H. C.; Straffelini, G.; Maliska, A. M.; Klein, A. N. Dry W. G. Reversible Control of Co Magnetism by Voltage-Induced Sliding of Plasma-Sintered Iron-The Influence of Nitriding on Wear Oxidation. Phys. Rev. Lett. 2014, 113 (26), 267202. Resistance. Wear 2008, 265 (3−4), 301−310. (23) Cialone, M.; Nicolenco, A.; Robbennolt, S.; Menéndez, E.; (44) Mohn, P.; Matar, S. F. The γ-Fe4N System Revisited: An Ab Rius, G.; Sort, J. Voltage-Induced ON Switching of Magnetism in Initio Calculation Study of the Magnetic Interactions. J. Magn. Magn. Ordered Arrays of Non-Ferrimagnetic Nanoporous Iron Oxide Mater. 1999, 191 (1−2), 234−240. Microdisks. Adv. Mater. Interfaces 2021, 8 (1), 2001143. (45) Gupta, M.; Tayal, A.; Gupta, A.; Gupta, R.; Stahn, J.; (24) de Rojas, J.; Quintana, A.; Lopeandía, A.; Salguero, J.; Muñiz, Horisberger, M.; Wildes, A. Iron and Nitrogen Self-Diffusion in Non- B.; Ibrahim, F.; Chshiev, M.; Nicolenco, A.; Liedke, M. O.; Butterling, Magnetic Iron Nitrides. J. Appl. Phys. 2011, 110, 123518. M.; Wagner, A.; Sireus, V.; Abad, L.; Jensen, C. J.; Liu, K.; Nogués, J.; (46) Chen, H. T.; Yan, M. F.; You, Y. Effect of N Distribution on Costa-Krämer, J. L.; Menéndez, E.; Sort, J. Voltage-Driven Motion of Elastic and Electronic Properties of Hexagonal ε-Fe6Nx by First- Nitrogen Ions: A New Paradigm for Magneto-Ionics. Nat. Commun. Principles Calculations. J. Magn. Magn. Mater. 2014, 354, 200−204. 2020, 11 (1), 5871. (47) Matar, S. F.; Demazeau, G.; Siberchicot, B. Magnetic Particles (25) Häglund, J.; Fernández Guillermet, A.; Grimvall, G.; Körling, Derived from Iron Nitride. IEEE Trans. Magn. 1990, 26 (1), 60−62. M. Theory of Bonding in Transition-Metal Carbides and Nitrides. (48) In CRC Handbook of Chemistry and Physics, 97th ed.; Rumble, J. Phys. Rev. B: Condens. Matter Mater. Phys. 1993, 48 (16), 11685− R., Ed.; CRC Press, 2016. 11691. (49) Jain, A.; Ong, S. P.; Hautier, G.; Chen, W.; Richards, W. D.; (26) Robie, R.; Hemingway, B. S. Thermodynamic Properties of Dacek, S.; Cholia, S.; Gunter, D.; Skinner, D.; Ceder, G.; Persson, K. Minerals and Related Substances at 298.15 K and 1 bar (10−5 A. Commentary: The Materials Project: A Materials Genome 30833 https://doi.org/10.1021/acsami.1c06138 ACS Appl. Mater. Interfaces 2021, 13, 30826−30834 ACS Applied Materials & Interfaces www.acsami.org Research Article Approach to Accelerating Materials Innovation. APL Mater. 2013, 1 A Boost toward Energy-Efficient Magnetic Actuation. Adv. Funct. (1), 011002. Mater. 2017, 27 (32), 1701904. (50) Mader, K. H.; Thieme, F.; Knappwost, A. Magnetische (70) Blöchl, P. E. Projector Augmented-Wave Method. Phys. Rev. B: Eigenschaften Der Kobalt-Stickstoff-Phase Co3N. Z. Anorg. Allg. Condens. Matter Mater. Phys. 1994, 50 (24), 17953−17979. Chem. 1969, 366 (5−6), 274−279. (71) Kresse, G.; Furthmüller, J. Efficient Iterative Schemes for Ab (51) Lutterotti, L.; Scardi, P. Simultaneous Structure and Size-Strain Initio Total-Energy Calculations Using a Plane-Wave Basis Set. Phys. Refinement by the Rietveld Method. J. Appl. Crystallogr. 1990, 23 (4), Rev. B: Condens. Matter Mater. Phys. 1996, 54 (16), 11169−11186. 246−252. (72) Kresse, G.; Furthmüller, J. Efficiency of Ab-Initio Total Energy (52) Wagner, A.; Butterling, M.; Liedke, M. O.; Potzger, K.; Krause- Calculations for Metals and Semiconductors Using a Plane-Wave Rehberg, R. Positron Annihilation Lifetime and Doppler Broadening Basis Set. Comput. Mater. Sci. 1996, 6 (1), 15−50. Spectroscopy at the ELBE Facility. AIP Conf. Proc. 2017, 1970, (73) Kresse, G.; Hafner, J. Ab Initio Molecular Dynamics for Liquid 040003. Metals. Phys. Rev. B: Condens. Matter Mater. Phys. 1993, 48 (17), (53) Wada, K.; Hyodo, T. A Simple Shape-Free Model for Pore-Size 13115−13118. Estimation with Positron Annihilation Lifetime Spectroscopy. J. Phys. (74) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Conf. Ser. 2013, 443, 012003. Approximation Made Simple. Phys. Rev. Lett. 1996, 77 (18), 3865− (54) Tuomisto, F.; Makkonen, I. Defect Identification in Semi- 3868. conductors with Positron Annihilation: Experiment and Theory. Rev. (75) Bellaiche, L.; Vanderbilt, D. Virtual Crystal Approximation Mod. Phys. 2013, 85 (4), 1583. Revisited: Application to Dielectric and Piezoelectric Properties of (55) Olsen, J. V.; Kirkegaard, P.; Pedersen, N. J.; Eldrup, M. Perovskites. Phys. Rev. B: Condens. Matter Mater. Phys. 2000, 61 (12), PALSfit: A New Program for the Evaluation of Positron Lifetime 7877−7882. Spectra. Phys. Status Solidi C 2007, 4 (10), 4004−4006. (76) Henkelman, G.; Jónsson, H. Improved Tangent Estimate in the (56) Liedke, M. O.; Anwand, W.; Bali, R.; Cornelius, S.; Butterling, Nudged Elastic Band Method for Finding Minimum Energy Paths M.; Trinh, T. T.; Wagner, A.; Salamon, S.; Walecki, D.; Smekhova, A.; and Saddle Points. J. Chem. Phys. 2000, 113 (22), 9978−9985. Wende, H.; Potzger, K. Open Volume Defects and Magnetic Phase (77) Henkelman, G.; Uberuaga, B. P.; Jónsson, H. Climbing Image Transition in Fe60Al40 Transition Metal Aluminide. J. Appl. Phys. Nudged Elastic Band Method for Finding Saddle Points and 2015, 117, 163908. Minimum Energy Paths. J. Chem. Phys. 2000, 113 (22), 9901−9904. (57) Krause-Rehberg, R.; Leipner, H. S. Positron Annihilation in Semiconductors; Springer Series in Solid-State Sciences; Springer: Berlin Heidelberg, 1999; Vol. 127, p 383. (58) Weisheit, M.; Fähler, S.; Marty, A.; Souche, Y.; Poinsignon, C.; Givord, D. Electric Field-Induced Modification of Magnetism in Thin-Film Ferromagnets. Science 2007, 315 (5810), 349−351. (59) Robbennolt, S.; Nicolenco, A.; Mercier Fernandez, P.; Auffret, S.; Baltz, V.; Pellicer, E.; Menéndez, E.; Sort, J. Electric Field Control of Magnetism in Iron Oxide Nanoporous Thin Films. ACS Appl. Mater. Interfaces 2019, 11 (40), 37338−37346. (60) Robbennolt, S.; Menéndez, E.; Quintana, A.; Gómez, A.; Auffret, S.; Baltz, V.; Pellicer, E.; Sort, J. Reversible, Electric-Field Induced Magneto-Ionic Control of Magnetism in Mesoporous Cobalt Ferrite Thin Films. Sci. Rep. 2019, 9, 10804. (61) Navarro-Senent, C.; Quintana, A.; Menéndez, E.; Pellicer, E.; Sort, J. Electrolyte-Gated Magnetoelectric Actuation: Phenomenol- ogy, Materials, Mechanisms, and Prospective Applications. APL Mater. 2019, 7 (3), 030701. (62) Leighton, C. Electrolyte-Based Ionic Control of Functional Oxides. Nat. Mater. 2019, 18, 13−18. (63) Soni, H. R.; Mankad, V.; Gupta, S. K.; Jha, P. K. A First Principles Calculations of Structural, Electronic, Magnetic and Dynamical Properties of Mononitrides FeN and CoN. J. Alloys Compd. 2012, 522, 106−113. (64) Matar, S.; Mohn, P. Electronic and Magnetic Properties of Fe2N and FeN: Trends of The Magnetism of the Fe-N System. Act. Passive Electron. Compon. 1993, 15 (2), 89−101. (65) Leslie-Pelecky, D. L.; Rieke, R. D. Magnetic Properties of Nanostructured Materials. Chem. Mater. 1996, 8, 1770−1783. (66) Allred, A. L. Electronegativity Values from Thermochemical Data. J. Inorg. Nucl. Chem. 1961, 17 (3−4), 215−221. (67) Mann, J. B.; Meek, T. L.; Knight, E. T.; Capitani, J. F.; Allen, L. C. Configuration Energies of the D-Block Elements. J. Am. Chem. Soc. 2000, 122 (21), 5132−5137. (68) Liu, Z. T. Y.; Zhou, X.; Khare, S. V.; Gall, D. Structural, Mechanical and Electronic Properties of 3d Transition Metal Nitrides in Cubic Zincblende, Rocksalt and Cesium Chloride Structures: A First-Principles Investigation. J. Phys.: Condens. Matter 2014, 26, 025404. (69) Quintana, A.; Zhang, J.; Isarain-Chávez, E.; Menéndez, E.; Cuadrado, R.; Robles, R.; Baró, M. D.; Guerrero, M.; Pané, S.; Nelson, B. J.; Müller, C. M.; Ordejón, P.; Nogués, J.; Pellicer, E.; Sort, J. Voltage-Induced Coercivity Reduction in Nanoporous Alloy Films: 30834 https://doi.org/10.1021/acsami.1c06138 ACS Appl. Mater. Interfaces 2021, 13, 30826−30834