Edge Tunnel Magnetoresistance

• Edge-TMR is a magnetotransport phenomenon where tunneling conductance is controlled by edge-specific features such as spin-split states, edge terminations, and modifications to the transverse mode spectrum.
• Research across 2D altermagnetic nanoribbons, Janus MBene, and topological insulator devices shows TMR ratios ranging from modest values to ultra-giant percentages based on edge geometry and chemistry.
• Methodologies like the Landauer-Büttiker formalism and non-equilibrium Green’s function analysis are used to quantify ballistic transport and edge roughness effects, guiding the design of MRAM and spintronic sensors.

Edge tunnel magnetoresistance (edge-TMR) denotes a family of tunneling magnetoresistance phenomena in which the decisive selectivity arises from edges: from spin-split edge states, from the chemistry and symmetry of electrode terminations, or from edge-induced modifications of the transverse mode spectrum. In current arXiv literature, the term is used most explicitly for two-dimensional altermagnetic nanoribbons whose transport is carried by edge states and governed by real-space-spin coupling (Fang et al., 14 Aug 2025). Closely related edge-dependent TMR phenomena also appear in Janus MBene in-plane junctions (Sun et al., 5 Feb 2025), topological-insulator edge devices (Götte et al., 2014), crystal-facet-sensitive altermagnetic tunnel junctions (Chi et al., 2023), and nanoscale magnetic tunnel junctions whose resistance and switching-voltage statistics are controlled by edge roughness (Pandey et al., 2023).

1. Edge-TMR as a distinct class of magnetotransport

Conventional tunnel magnetoresistance in magnetic tunnel junctions is usually formulated in terms of the conductance contrast between parallel and anti-parallel magnetic configurations of electrodes separated by an insulating barrier. The edge-TMR literature preserves that core logic but relocates the decisive filtering mechanism to edge degrees of freedom. In the formulation of “Edgetronics in 2D Altermagnet via Real-Space-Spin Coupling,” the effect is explicitly defined as an unconventional TMR specific to quantum-confined edge states in two-dimensional altermagnetic nanoribbons, where transport is transferred from bulk to spin-split edge states and is governed by real-space-spin rather than momentum-spin coupling (Fang et al., 14 Aug 2025).

A second usage emphasizes edge termination rather than edge-state topology. In Janus Cr$_4$B$_3$N in-plane MTJs, the tunnel response depends strongly on whether the electrode edges are Cr-B-N, Cr-B, or mixed, because the edge terminations determine the matching or mismatch of $\mathbf{k}$-resolved spin-polarized conduction channels across the barrier (Sun et al., 5 Feb 2025). A related but not identical formulation appears in altermagnetic tunnel junctions based on RuO$_2$/TiO$_2$/CrO$_2$, where TMR is almost zero along [001] but reaches $6100\%$ along [110]; there the operative variable is crystal facet or transport direction rather than edge-state localization per se (Chi et al., 2023).

A plausible synthesis is that edge-TMR is best treated as an umbrella category for TMR effects in which edge geometry, edge chemistry, edge-state topology, or edge-derived transport anisotropy determines the tunneling contrast. This broader reading is consistent with the way edge roughness in ultrasmall MTJs perturbs both resistance states and therefore the reproducibility of TMR-related observables (Pandey et al., 2023).

2. Microscopic mechanisms

In Cr$_2$Se$_2$O, the central mechanism is one-dimensional real-space-spin coupling in a two-dimensional altermagnetic second-order topological insulator. The spin-split floating edge states are energetically isolated within the bulk band gap, each edge hosts states of opposite spin, and transport occurs exclusively through these edge states with quantized spin conductance (Fang et al., 14 Aug 2025). The switching logic is stated directly in the edge basis:

$$|r,s\rangle \rightarrow |r,s\rangle \quad \text{(P state, current ON)}$$

$_3$0

The OFF state follows from spin mismatch and orthogonality of the edge channels, while the ON state follows from matched edge and spin character.

In Janus Cr$_3$1B$_3$2N, the mechanism is different. The material is an altermagnet with zero net magnetization but anisotropic, momentum-dependent spin splitting. Three in-plane MTJs with a $_3$3 vacuum barrier were constructed, and the decisive variable is the edge assembly of the electrodes. For symmetric Cr-B vertical edge-assembled electrodes, the parallel state exhibits highly conductive transmission strips in $_3$4-space whereas the anti-parallel state is nearly blocked over the Brillouin zone, producing an ultra-giant TMR ratio of $_3$5 (Sun et al., 5 Feb 2025). For asymmetric or chemically different terminations, the separation between parallel and anti-parallel transmission channels is reduced.

In RuO$_3$6-based altermagnetic tunnel junctions, the decisive element is facet-resolved conduction-channel projection. For the (001) facet, spin-up and spin-down Fermi-surface projections are shape-equivalent and current is effectively spin-neutral, so the junction shows almost zero TMR. For the (110) facet, the projected conduction channels for opposite spins diverge substantially and the altermagnetic electrode functions as a spin polarizer, yielding giant TMR (Chi et al., 2023). This establishes that giant TMR in altermagnets does not require a net magnetic moment; momentum-resolved spin splitting can be sufficient.

Topological-insulator edge devices realize yet another microscopic route. In Bi$_3$7Se$_3$8, the helical edge states of thin strips are spin-momentum locked, so electrons with a given spin propagate only in one direction. TMR then arises because a ferromagnetic injector preferentially couples to one helical branch, strongly modulating the current collected at the output contact (Götte et al., 2014). The edge-state magnetoresistance theory for two-dimensional topological insulators further shows that magnetic-field-induced gap opening and impurity-assisted backscattering can control edge conductance, providing a complementary edge-magnetotransport mechanism adjacent to, though not identical with, tunnel MR (Braginsky et al., 2021).

3. Materials platforms and device realizations

The present literature spans altermagnets, topological insulators, and conventional MgO-based MTJ structures. The platforms differ in whether the operative edge degree of freedom is a topological edge state, an atomically defined edge termination, a facet-selected momentum filter, or a roughened sidewall that perturbs quantum confinement.

Platform	Geometry or edge condition	Reported behavior
Cr$_3$9Se$\mathbf{k}$0O monolayer	Edge-MTJ nanoribbon with type-I and type-II edge magnetic order	ON state transmission $\mathbf{k}$1, OFF state transmission $\mathbf{k}$2 at the Fermi energy; edge-TMR up to $\mathbf{k}$3 for type-I and up to $\mathbf{k}$4 for type-II
Cr$\mathbf{k}$5B$\mathbf{k}$6N	In-plane MTJs with $\mathbf{k}$7 vacuum barrier and Conf-1/2/3 edge assemblies	Conf-1: $\mathbf{k}$8; Conf-2: $\mathbf{k}$9; Conf-3: $_2$0
RuO$_2$1/TiO$_2$2/CrO$_2$3	Altermagnet/insulator/ferromagnet tunnel junction along [001] or [110]	Almost zero TMR along [001]; $_2$4 along [110]
Bi$_2$5Se$_2$6	FM/insulator/topological-insulator edge device	Room-temperature TMR of the order of $_2$7

Cr$_2$8Se$_2$9O is notable because the device is composed entirely of a single two-dimensional altermagnet in a nanoribbon geometry, with no heterojunction required; the paper states that this avoids interface issues common in conventional TMR devices (Fang et al., 14 Aug 2025). By contrast, the Bi$_2$0Se$_2$1 proposal retains a tunnel barrier and a ferromagnetic injector but avoids the use of a second ferromagnetic electrode whose magnetization needs to be pinned (Götte et al., 2014). The RuO$_2$2 and Cr$_2$3B$_2$4N works place altermagnets in the electrode role, using momentum-resolved spin splitting to achieve spin filtering despite vanishing net magnetization (Chi et al., 2023, Sun et al., 5 Feb 2025).

4. Formal descriptions and transport laws

For the Cr$_2$5Se$_2$6O edge-MTJ, ballistic transport is described with the Landauer-Büttiker expression

$_2$7

and the reported transmission plateaus at $_2$8, $_2$9, or $_2$0 indicate integer numbers of spin-resolved conduction channels (Fang et al., 14 Aug 2025). In that work, the edge-TMR ratio is defined as

$_2$1

The same study reports electrical tuning by a lateral field: one spin-polarized edge channel can be shifted away from the Fermi level, changing the quantized conductance from $_2$2 to $_2$3. The critical field is reported as $_2$4 in intrinsic Cr$_2$5Se$_2$6O and $_2$7 with electron doping (Fang et al., 14 Aug 2025).

For Bi$_2$8Se$_2$9 edge devices, the analytical expression for the two-dimensional edge-state TMR is

$6100\%$0

where $6100\%$1 is the spin polarization of the edge state, $6100\%$2 is the spin polarization of electrons injected from the ferromagnet, and $6100\%$3 is the in-plane magnetization angle (Götte et al., 2014). In the ideal limit $6100\%$4 and $6100\%$5, the analytical TMR diverges. The paper interprets this as a consequence of perfect spin selectivity in the helical one-dimensional edge channels.

The magnetoresistance theory for edge states of a two-dimensional topological insulator introduces a complementary formalism based on a field-induced gap in the edge spectrum:

$6100\%$6

with

$6100\%$7

In that treatment, the magnetic field alone opens the gap but does not by itself induce backscattering; backscattering appears through the combined action of the magnetic field and impurities (Braginsky et al., 2021). Although this is not a tunnel-junction model, it is directly relevant to edge magnetotransport because it quantifies how time-reversal-symmetry breaking modifies edge conduction.

5. Edge disorder, roughness, and variability

The most direct treatment of edge disorder in MTJs is the non-equilibrium Green’s function analysis of edge roughness in nanoscale magnetic tunnel junctions (Pandey et al., 2023). There, edge roughness is modeled as a stochastic variation of the cross-sectional radius,

$6100\%$8

with a stretched-exponential autocorrelation function

$6100\%$9

where $_2$0. The rough edge is generated by filtering white noise with the Fourier transform of the autocorrelation, and the study uses $_2$1 realizations to extract statistical variations.

The principal result is that stochastic variation in shape and size changes the transverse energy mode profile and therefore changes resistance and switching voltage. The paper reports that the variations are larger as the MTJ size is scaled down because of quantum confinement. For $_2$2, $_2$3, and $_2$4, the coefficient of variation in $_2$5 is reported as $_2$6, $_2$7, and $_2$8, respectively (Pandey et al., 2023). It also gives an area-based variance estimate,

$_2$9

which scales as $_2$0 and increases with $_2$1 up to saturation when $_2$2 is large.

The same work proposes two reduced models for efficient computation. In the circle approximation, each rough realization is mapped to a perfect circle with the same area. In the ellipse approximation, the cross section is mapped to an ellipse with the same area and a ground-state-energy constraint determined by the first transverse eigenvalue (Pandey et al., 2023). Because the impact on TMR itself is not directly tabulated, any statement about TMR must remain qualified. The paper does state that both parallel and anti-parallel resistances are perturbed; this suggests that edge roughness reduces the reproducibility of TMR in aggressively scaled devices, even when the dominant research interest is the variability of resistance and switching voltage rather than the maximization of TMR.

6. Interpretation, misconceptions, and design implications

One common misconception is that high TMR requires a ferromagnetic electrode with a large net moment. The altermagnetic literature directly contradicts that premise. Cr$_2$3B$_2$4N and RuO$_2$5 both have vanishing net magnetization in the relevant descriptions, yet their momentum-resolved spin splitting produces large or giant TMR when the transport geometry and edge or facet conditions are appropriate (Sun et al., 5 Feb 2025, Chi et al., 2023). In Cr$_2$6Se$_2$7O, the decisive object is not a ferromagnetic moment at all but a pair of spin-split floating edge states in a second-order topological setting (Fang et al., 14 Aug 2025).

A second misconception is that TMR in nanoscale structures is controlled solely by interface quality. The edge-roughness study shows that stochastic sidewall deviations modify the transverse mode spectrum, resistance, and switching voltage, with stronger effects at smaller radii (Pandey et al., 2023). The Janus MBene work adds that atomically sharp and compositionally well-defined edge terminations can change TMR from $_2$8 to $_2$9 (Sun et al., 5 Feb 2025). The RuO$$|r,s\rangle \rightarrow |r,s\rangle \quad \text{(P state, current ON)}$$0 study further shows that changing the transport direction from [001] to [110] can move the same basic junction concept from almost zero TMR to $$|r,s\rangle \rightarrow |r,s\rangle \quad \text{(P state, current ON)}$$1 (Chi et al., 2023). Taken together, these results place edge chemistry, edge morphology, and facet orientation on the same conceptual footing as barrier composition and interfacial spin filtering.

A third misconception is that edge-TMR is a single mechanism. The available work instead identifies at least three operational regimes. One regime is edge-state mediated and real-space selective, as in Cr$$|r,s\rangle \rightarrow |r,s\rangle \quad \text{(P state, current ON)}$$2Se$$|r,s\rangle \rightarrow |r,s\rangle \quad \text{(P state, current ON)}$$3O (Fang et al., 14 Aug 2025). A second is edge-termination dependent and momentum resolved, as in Cr$$|r,s\rangle \rightarrow |r,s\rangle \quad \text{(P state, current ON)}$$4B$$|r,s\rangle \rightarrow |r,s\rangle \quad \text{(P state, current ON)}$$5N (Sun et al., 5 Feb 2025). A third is helical-edge based, as in Bi$$|r,s\rangle \rightarrow |r,s\rangle \quad \text{(P state, current ON)}$$6Se$$|r,s\rangle \rightarrow |r,s\rangle \quad \text{(P state, current ON)}$$7, where a single ferromagnet and a topological edge channel suffice to generate TMR of the order of $$|r,s\rangle \rightarrow |r,s\rangle \quad \text{(P state, current ON)}$$8 at room temperature (Götte et al., 2014). A closely related non-tunnel regime is magnetic-field-controlled edge magnetoresistance in topological insulators, where impurity-assisted backscattering depends on the field-induced gap and can even be suppressed by sufficiently strong impurity interaction (Braginsky et al., 2021).

The design implications reported across these studies are consistent. Edge engineering appears as a new degree of freedom for device optimization; Néel-order control, lateral electric field, and electrostatic doping enable switching and tunability in edge-MTJs (Fang et al., 14 Aug 2025). Altermagnets are proposed as candidates for MRAM, sensors, and spin-transfer-torque devices (Sun et al., 5 Feb 2025). The roughness study targets reliable design of STT-MRAM with ultra-small MTJs (Pandey et al., 2023). This suggests that edge-TMR is not a narrow subtopic of tunnel spectroscopy but a broader design framework connecting altermagnetism, topological edge transport, and nanoscale variability in spintronic devices.