Voltage-Controlled Magneto-Ionic Device

• Voltage-controlled magneto-ionic devices are solid-state elements where applied voltage triggers non-volatile ion migration to modulate magnetic properties.
• They leverage diverse material architectures and ionic species to dynamically control magnetization, anisotropy, and domain structures for memory and logic applications.
• Optimized device performance hinges on precise control of ion transport, readout mechanisms, and scaling strategies compatible with CMOS integration.

A voltage-controlled magneto-ionic device is a class of solid-state element in which magnetic properties—such as magnetization, anisotropy, or even electronic transport—are modulated via voltage-driven, non-volatile migration of ions within the material lattice or across interfaces. These devices operate by applying an electric field to induce motion of specific mobile ionic species (e.g., O²⁻, N³⁻, H⁺, Li⁺, OH⁻), enabling dynamic and energy-efficient control over ferromagnetism, magnetic anisotropy, and domain structure without resorting to large currents or magnetic fields. The physical mechanisms, kinetics, device performance, and application horizons depend critically on the choice of material architecture, mobile ion, and gating configuration.

1. Materials Systems and Device Architecture

Magneto-ionic devices have been demonstrated in both thin-film and nanostructured geometries. Representative device stacks and their key parameters include:

Mobile Ion	Representative Stack	Typical Operating Voltage	Key Layer Thicknesses
O²⁻	Si/SiO₂/Ta/Pt/Co(0.9 nm)/GdOx(3–30 nm)/Au(12 nm)	±4–12 V	Co: 0.9 nm, GdOx: 3–30 nm
N³⁻	Si/Ti/Cu/CoN(85 nm)	–4 to –50 V	CoN: 85 nm
H⁺	Pd/Co(1 nm)/Pd/YSZ(3–20 nm)/Pt	±1–6 V	YSZ: 3–20 nm
Li⁺	Ta/CoFeB/MgO/Ta/LiPON(70–100 nm)/Pt	±2–3 V	CoFeB: 1 nm, LiPON: 70–100 nm
OH⁻	Ta/Pd/α-Co(OH)₂ (thick, electrodeposited)	±2–8 V	α-Co(OH)₂: 500 μm
N³⁻ (nanodots)	Si/Ti/Pt/FeCoN(nanodot)/PC electrolyte/Pt	–10 to –25 V	FeCoN dots: 20–35 nm thick
C²⁻/Fe²⁺	Si/Ti/Cu/Fe–C multilayer/Ti–C cap	±50 V	Fe–C: 45–58 nm
Fe²⁺/B³⁺/O²⁻	Si/Ti/Au/Ta/FeBO(x=0–5% O)/PC electrolyte/Pt	±50 V	FeBO: 50 nm

Layers are typically fabricated by sputtering for inorganic structures, with electron-beam lithography or optical lithography for pattern definition and electrode integration. Electrolyte gating (either liquid, solid, or ion-gel) is the standard approach for applying electric fields, forming electric double layers at the magnetic interface for efficient ion transfer (Rojas et al., 2020, Tan et al., 14 Mar 2025, Das et al., 11 Nov 2025).

2. Ion Transport and Magneto-Ionic Mechanisms

The decisive attribute of magneto-ionic devices is voltage-driven migration of ions, resulting in local redox reactions, stoichiometry changes, or interfacial reconstruction. The generic magneto-ionic transport follows:

• Drift-diffusion equations: For each ionic species,

$J = -D\,\nabla c + \mu c E$

where $J$ is flux, $D$ the temperature-dependent diffusivity [$D(T) = D_0 e^{-E_a/k_B T}$], $c$ the local concentration, and $\mu$ the mobility.

• Activation barriers and threshold voltages: The onset of ion migration is governed by $E_a$ (activation energy), ionic charge $z$, and local field $E$. Threshold voltage $V_{th}$ is typically $V_{th}\sim E_c d_{EDL}$, with $E_c$ the critical field and $d_{EDL}$ the double-layer thickness.
• Redox and magnetic phase formation: For example, N³⁻ migration in CoN leaves a metallic Co sublayer with emergent ferromagnetism; O²⁻ insertion/removal in Co/oxide toggles interfacial perpendicular magnetic anisotropy (PMA); H⁺/Li⁺/OH⁻ can modulate magnetic properties via intercalation and lattice modification (Rojas et al., 2020, Bauer et al., 2014, Quintana et al., 2022).

Distinct migration fronts form: in nanocrystalline CoN and FeCoN, a planar, wavefront-like depletion is observed, allowing homogeneous control across thickness. Kinetics are generally faster for low-electronegativity ions (e.g., N³⁻, C²⁻) with lower $E_a$, yielding lower $V_{th}$ and higher rates compared to O²⁻.

3. Magnetization Control, Dynamics, and Performance Metrics

Switching phenomena in magneto-ionic devices include:

• Ferromagnetism ON/OFF: CoN transitions from paramagnetic to ferromagnetic at a critical N content ($\lesssim$50 at.% N), achieving $\Delta M_S \simeq 637\,\text{emu}/\text{cm}^3$ with coercivity $H_C \sim 12$ Oe under bias (Rojas et al., 2020).
• Anisotropy switching: PMA in CoFeB/HfO₂ or Co/GdOx can be toggled over $\Delta K_S > 0.6\,\text{erg}/\text{cm}^2$ (Bauer et al., 2014). In YSZ/Co devices, millisecond PMA–IMA toggling is possible at $V_G=+6$ V (Lee et al., 2020).
• Cyclability and endurance: CoN-based elements sustain $>10^3$ ON/OFF cycles with $<5\%$ degradation, $\sim10^4$ cycles projected under optimized $|V|<5$ V (Rojas et al., 2020).

Device	$V_{th}$(ON)	Switching Time	$\Delta M_S$	Key Endurance
CoN (85 nm)	–4 V	4 min (–4 V), 50 V faster	637 emu/cm³	$>10^3$ cycles
Co₃O₄ (130 nm)	–8 V	6.2 min (–8 V), 50 V faster	588 emu/cm³	~30 cycles
Fe–C (58 nm)	–50 V	$\sim$40 min (full)	7× initial Ms	Nonvolatile
α-Co(OH)₂	–2 to –8 V	120–2 min (–2 to –8 V)	2.5 emu/cm³	Full reversibility
YSZ/Co (1 nm)	+6 V	1 ms (6 nm YSZ)	PMA $\to$ IMA	$>10^3$ cycles

Switching rates scale strongly with the applied voltage and temperature due to the Arrhenius dependence of $D(T)$, and are further boosted in systems with highly uniform electric fields (e.g., via conducting TiN buffer) or lower ion activation barriers (Rojas et al., 2020, Quintana et al., 2022).

4. Device Physics: Models and Readout Techniques

Transport modeling uses coupled drift–diffusion (Nernst–Planck) and reaction–diffusion equations to describe the ionic migration and magnetic transformation:

$$J_i = -D_i \nabla c_i - \frac{z_i F}{RT} D_i c_i \nabla\phi;\quad \frac{\partial c_i}{\partial t} + \nabla\cdot J_i = 0$$

Magnetic property modulation is generally linked to local ion concentration via phenomenological expressions such as $M_S(x) = M_0[1 - \alpha\,c_\mathrm{ion}(x)]$ or $K_\mathrm{eff}(c_i)=K_0-\alpha c_i$ (Murray et al., 2021, Das et al., 11 Nov 2025).

Readout is performed via MOKE (magneto-optic Kerr effect) for spatially resolved domain imaging, VSM/SQUID for absolute $M_S$, and anomalous Hall effect for integrated devices. Patterned nanodot arrays and vortex-state control are accessed by AC demagnetization and statistical analysis (Shannon entropy, intra-fractional Hamming distance) (Spasojevic et al., 15 Jul 2025, Spasojevic et al., 20 Mar 2024).

Endurance and retention are determined by the non-volatility of ionic rearrangement; the switched magnetic state persists after gate removal, provided the ions remain bound and no back-diffusion or secondary reactions occur.

5. Application Spaces and Integration Guidelines

Voltage-controlled magneto-ionic devices are positioned for several advanced applications:

• Non-volatile memory and logic: Single-layer CoN or patterned FeCoN dots provide bistable or multi-state magnetization for non-volatile bits, reconfigurable memory, and logic gate design (Rojas et al., 2020, Spasojevic et al., 15 Jul 2025).
• Analog synapses and neuromorphic computing: Partial (sub-threshold or pulsed) gating yields analog, non-volatile $M_S$ tuning, supporting in situ synaptic-weight emulation, hardware reservoir computing, and multistate elements (Das et al., 11 Nov 2025, Spasojevic et al., 20 Mar 2024, Das et al., 15 Dec 2024).
• Hardware security primitives: Voltage-addressable, reconfigurable FeCoN nanodots act as physically unclonable functions, random number generators, and “p-bits” for probabilistic inference architectures (Spasojevic et al., 15 Jul 2025).
• Integration with CMOS platforms: Device stacks can be directly sputtered or patterned onto standard Si(100) substrates. On-chip gating leverages solid or microfluidic electrolytes; crossbar and vertical integration are possible.

Application	Key Advantages	Integration Aspects
Memory/logic	Non-volatility, low power	CMOS backend, single-layer CoN
Neuromorphic	Analog $M_S$, $\tau_{switch}\sim$s–min	Partial gating, fine bias control
Security	Deterministic/probabilistic states	Litho-defined nanodots, readout optics
Sensors/MEMS	Endurance, cycling, moderate speed	>100 nm films for stability

Integration is facilitated by room-temperature, electrode-compatible, and low-voltage (sub–5 V) operation in many material systems. For high-speed or sub-ms switching, ultrathin films and fast ionic conductors (solid-state LiPON, YSZ, highly conductive buffer layers) are essential.

6. Practical Considerations, Limitations, and Optimization

Endurance and reliability hinge on managing parasitic electrochemistry (e.g., O₂/H₂ evolution), optimal voltage operation (to avoid gas evolution and irreversible ion loss), and mechanical stability under repeated cycling.

Speed–endurance–power trade-offs are governed by:

• Speed $\propto \exp(-E_a/k_BT)\,V$
• Endurance degrades for $|V| > V_\mathrm{gas\,evolution}$ or excessive ion extraction rates
• Recommended $|V| \lesssim 5$ V, 1–10 min cycling for >10⁴ cycles, with activation energy tuning via heating or materials engineering (Rojas et al., 2020, Quintana et al., 2022).

Device optimization involves adjusting film thickness (thinner films for faster switching at lower $V$), enhancing field uniformity (e.g., via TiN/Pt buffer), and selecting high-mobility solid electrolytes. For sub-ms operation, proton-conducting YSZ or nanostructured interfaces are promising (Lee et al., 2020).

Scalability is enabled by standard micro-/nano-fabrication for dot arrays, CMOS-compatible layer stacks, and patternable electrolyte architectures. Voltage windows, device dimensions, and materials selection must balance speed, retention, and energy per switching event.

7. Outlook and Emerging Directions

Recent advances extend magneto-ionics to:

• Dual- and multi-ion systems: Fe–C and Fe–BO structures allow cascaded or concurrent transport of multiple ionic species, enabling new functionalities such as biocompatibility and multi-level weight storage (Tan et al., 14 Mar 2025, Ma et al., 14 Mar 2025).
• Patterned vortex-state and analog-nanodot memory: FeCoN nanodots show voltage-tunable transitions between paramagnetic, single-domain, and vortex (“vortion”) regimes, with analog tuning of magnetic properties post-fabrication (Spasojevic et al., 20 Mar 2024).
• Wireless actuation: Magneto-ionics is achievable via wireless bipolar electrochemistry, eliminating direct wiring for applications in bioelectronics and microfluidics (Ma et al., 2023).
• Sub-millisecond, high-endurance switching: YSZ-based solid-state architectures offer $\sim$1 ms switching at room temperature, $>10^3$ cycles, and CMOS technology compatibility (Lee et al., 2020).

Magneto-ionic devices thus represent a diverse and adaptable platform for dynamically tuning magnetism by voltage-induced ionic motion, with application potential spanning memory, logic, neuromorphic, security, and sensor domains (Rojas et al., 2020, Das et al., 11 Nov 2025, Spasojevic et al., 15 Jul 2025).