Fully Compensated Ferrimagnetic Metal

• Fully compensated ferrimagnetic metals are materials with chemically or crystallographically inequivalent sublattices that align antiparallel to yield a nearly zero net magnetization while retaining metallic and spin-polarized electronic states.
• These systems are engineered via techniques such as valence-electron tuning in Heusler compounds and cluster compensation, resulting in high ordering temperatures and robust spin functionality.
• Their unique combination of antiferromagnetic compensation and ferromagnetic spin properties enables applications in spin-transfer-torque devices, exchange bias engineering, and advanced spintronic systems.

A fully-compensated ferrimagnetic metal is a magnetically ordered material in which inequivalent magnetic sublattices, or in some cases inequivalent atomic clusters, carry substantial antiparallel moments whose vector sum is zero or nearly zero, while the electronic structure remains metallic and often strongly spin-polarized. In contrast to a conventional antiferromagnet, the compensating moments are not related by simple symmetry equivalence, so spin-resolved bands and transport channels need not be identical. This class therefore combines antiferromagnet-like vanishing macroscopic magnetization with ferromagnet-like spin functionality, and it appears in several experimentally studied Heusler and non-Heusler compounds as well as in a growing set of two-dimensional and symmetry-engineered models (Midhunlal et al., 2018, Stinshoff et al., 2016, Giri et al., 2021, Liu et al., 21 Feb 2025).

1. Definition and distinguishing characteristics

In a collinear ferrimagnet, the total magnetization is the sum of oppositely oriented sublattice contributions,

$M(T)=M_A(T)+M_B(T),$

with $|M_A|\neq |M_B|$ in general. A fully compensated ferrimagnet is the limiting case in which these inequivalent sublattice moments cancel,

$M_{\mathrm{tot}}\approx 0,$

even though each sublattice remains strongly magnetized. The defining distinction from a conventional antiferromagnet is that the compensating units are chemically or crystallographically inequivalent, so zero net moment is not enforced by a symmetry that would also force identical spin-up and spin-down spectra (Midhunlal et al., 2018, Stinshoff et al., 2016).

Because the metallic state can remain highly spin-polarized, the literature also uses the labels “half-metallic fully compensated ferrimagnet”, “fully compensated ferrimagnet”, and, historically, “half-metallic antiferromagnet”. The more precise usage in the cited works is that true half-metallicity is compatible with compensated ferrimagnetism rather than with a symmetry-enforced antiferromagnet, since a conventional antiferromagnetic symmetry would make the two spin channels equivalent (Stinshoff et al., 2016, Semboshi et al., 2021).

A useful modern contrast is with altermagnetism. In an altermagnet, symmetry implies $g_{\uparrow}(E)=g_{\downarrow}(E)$ for the spin-resolved density of states, which rules out an altermagnetic half-metal. In a fully compensated ferrimagnet, by contrast, the opposite-spin magnetic atoms are not symmetry-related, so $g_{\uparrow}(E)\neq g_{\downarrow}(E)$ is allowed while the integrated moment still vanishes. This is why recent work treats fully compensated ferrimagnetic metals as a distinct zero-net-moment class rather than as a subset of antiferromagnets (Guo et al., 5 Dec 2025).

2. Structural and electronic design principles

A dominant design route runs through Heusler compounds. Full Heuslers of the form $X_2YZ$ often crystallize either in the cubic $L2_1$ structure ($Fm\bar{3}m$) or in the inverse or $X_a$ structure ($F\bar{4}3m$), and compensation is commonly targeted by tuning the valence-electron count to the Slater–Pauling condition

$|M_A|\neq |M_B|$0

for full Heuslers. At $|M_A|\neq |M_B|$1, a zero total moment is expected (Midhunlal et al., 2018, Stinshoff et al., 2016, Midhunlal et al., 2020).

This strategy is explicit in $|M_A|\neq |M_B|$2, where the valence-electron count is 24 and the calculated total moment is essentially zero, while local moments remain sizable: Mn on the $|M_A|\neq |M_B|$3 site carries about $|M_A|\neq |M_B|$4, Mn on $|M_A|\neq |M_B|$5 about $|M_A|\neq |M_B|$6, Fe on $|M_A|\neq |M_B|$7 about $|M_A|\neq |M_B|$8, and V on $|M_A|\neq |M_B|$9 about $M_{\mathrm{tot}}\approx 0,$0 (Stinshoff et al., 2016). The same logic underlies $M_{\mathrm{tot}}\approx 0,$1 ($M_{\mathrm{tot}}\approx 0,$2), where 50–50 V/Co mixing tunes the sublattice balance while preserving high ordering temperatures (Midhunlal et al., 2018).

Compensation need not be site-based in the usual Heusler sense. In Pt- and Ni-doped cubic $M_{\mathrm{tot}}\approx 0,$3, the compensating objects are two 26-atom clusters within the 52-atom $M_{\mathrm{tot}}\approx 0,$4-brass-type cell. Density functional theory gives $M_{\mathrm{tot}}\approx 0,$5 and $M_{\mathrm{tot}}\approx 0,$6, producing an effective moment of about $M_{\mathrm{tot}}\approx 0,$7. This “intra-unitcell cluster-cluster magnetic compensation” is structurally different from Heusler sublattice compensation but serves the same fully compensated ferrimagnetic function (Giri et al., 2021).

Recent two-dimensional work has added a filling-enforced perspective. In that formulation, one spin channel is gapped, the electron filling fixes $M_{\mathrm{tot}}\approx 0,$8 and $M_{\mathrm{tot}}\approx 0,$9 as integers, and the compensation condition

$g_{\uparrow}(E)=g_{\downarrow}(E)$0

can coexist with spin-split, ferromagnetic-like bands because no symmetry links the opposite-spin sublattices. This framework extends fully compensated ferrimagnetism beyond the classical Heusler setting (Liu et al., 21 Feb 2025, Wang et al., 27 Feb 2026).

3. Representative materials and experimental realizations

Several bulk materials now serve as reference systems for the subject, and they span high-$g_{\uparrow}(E)=g_{\downarrow}(E)$1 Heuslers, cluster-compensated cubic alloys, spin-gapless chalcogenides, and disorder-stabilized Heuslers.

System	Reported compensated-state signature	Key scale
$g_{\uparrow}(E)=g_{\downarrow}(E)$2 (Midhunlal et al., 2018)	Room-temperature moment $g_{\uparrow}(E)=g_{\downarrow}(E)$3 in ribbon and $g_{\uparrow}(E)=g_{\downarrow}(E)$4 in bulk	$g_{\uparrow}(E)=g_{\downarrow}(E)$5–$g_{\uparrow}(E)=g_{\downarrow}(E)$6; ribbon $g_{\uparrow}(E)=g_{\downarrow}(E)$7
$g_{\uparrow}(E)=g_{\downarrow}(E)$8 (Midhunlal et al., 2018)	Room-temperature moment $g_{\uparrow}(E)=g_{\downarrow}(E)$9 in ribbon and $g_{\uparrow}(E)\neq g_{\downarrow}(E)$0 in bulk	$g_{\uparrow}(E)\neq g_{\downarrow}(E)$1–$g_{\uparrow}(E)\neq g_{\downarrow}(E)$2; ribbon $g_{\uparrow}(E)\neq g_{\downarrow}(E)$3
$g_{\uparrow}(E)\neq g_{\downarrow}(E)$4 (Stinshoff et al., 2016)	Calculated total moment $g_{\uparrow}(E)\neq g_{\downarrow}(E)$5 per primitive cell; experimentally compensated at low $g_{\uparrow}(E)\neq g_{\downarrow}(E)$6	$g_{\uparrow}(E)\neq g_{\downarrow}(E)$7; $g_{\uparrow}(E)\neq g_{\downarrow}(E)$8 in slightly overcompensated samples
Pt/Ni-doped $g_{\uparrow}(E)\neq g_{\downarrow}(E)$9 (Giri et al., 2021)	Lowest-energy FCF state with $X_2YZ$0	Exchange bias $X_2YZ$1–$X_2YZ$2; $X_2YZ$3
NiAs-type $X_2YZ$4 (Semboshi et al., 2021)	Half-metallic fully compensated ferrimagnet with zero magnetization and spin-polarized Fermi surfaces	$X_2YZ$5 around $X_2YZ$6; coercivity $X_2YZ$7 at $X_2YZ$8
A2-disordered $X_2YZ$9 (Philip et al., 11 Dec 2025)	Ordered moment $L2_1$0; calculated $L2_1$1	$L2_1$2

Among Mn-based Heuslers, $L2_1$3 is notable because the parent alloys $L2_1$4 and $L2_1$5 carry moments near $L2_1$6, whereas the mixed-Y compounds reduce the net moment by more than an order of magnitude while maintaining $L2_1$7 (Midhunlal et al., 2018). In $L2_1$8, neutron diffraction and first-principles analysis identify $L2_1$9 as the high-$Fm\bar{3}m$0 fully compensated ferrimagnetic composition, with compensation arising in a disordered $Fm\bar{3}m$1 structure through antiparallel coupling between ferromagnetically aligned $Fm\bar{3}m$2 and $Fm\bar{3}m$3 atom pairs (Midhunlal et al., 2020).

Outside the Mn-based systems, $Fm\bar{3}m$4 was predicted to adopt an inverse Heusler structure with a total moment of about $Fm\bar{3}m$5, and experiment found a low-moment ferrimagnetic phase with antiferromagnetically oriented Cr and Co moments, although synchrotron diffraction also revealed CoAl and Cr secondary phases (Jamer et al., 2015). This suggests that approaching full compensation can be easier than isolating the single-phase equilibrium structure.

Non-Heusler systems broaden the concept substantially. Pt/Ni-doped cubic $Fm\bar{3}m$6 realizes a compensated ferrimagnetic background together with uncompensated ferrimagnetic clusters, enabling very large exchange bias (Giri et al., 2021). Hexagonal NiAs-type $Fm\bar{3}m$7 was reported as an experimentally realized HM-FCFM material with a compensation temperature around $Fm\bar{3}m$8, linear $Fm\bar{3}m$9 below compensation, and high magnetic coercivity (Semboshi et al., 2021). A later pyrrhotite-type $X_a$0 study found fully compensated ferrimagnetic behavior with $X_a$1, N-type ferrimagnetism, a pseudo-gap at the Fermi level in the up-spin band, and coercivity of $X_a$2 at $X_a$3 (Yin et al., 16 Dec 2025).

4. Compensation temperatures, Néel classifications, and magnetic thermodynamics

In experiment, fully compensated ferrimagnetism is rarely a perfectly temperature-independent $X_a$4 state. More commonly, the sublattice moments have different thermal evolutions and cross at a compensation temperature $X_a$5,

$X_a$6

This produces the characteristic N-type ferrimagnetic response described by Néel: the net magnetization decreases with increasing temperature, crosses zero, changes sign, and then vanishes at the ordering temperature. By contrast, P-type ferrimagnetism does not cross zero within the ordered phase (Midhunlal et al., 2018).

This contrast is unusually clear in $X_a$7. The melt-spun ribbons are N-type ferrimagnets with $X_a$8 for $X_a$9 and $F\bar{4}3m$0 for $F\bar{4}3m$1, while the arc-melted bulk samples are P-type ferrimagnets whose $F\bar{4}3m$2 increases toward $F\bar{4}3m$3 without crossing zero in the measured range (Midhunlal et al., 2018). In $F\bar{4}3m$4, slight deviations from the ideal composition move the compensation point from near $F\bar{4}3m$5 to about $F\bar{4}3m$6, and the crossing is accompanied by magnetic reversal and a sign change of the anomalous Hall effect (Stinshoff et al., 2016).

A recent mean-field generalization of the Néel diagram identifies a critical regime in which full compensation is maintained below the Curie temperature and extends past the nominal compensation point. For a two-sublattice ferrimagnet the net moment may be written as

$F\bar{4}3m$7

and the critical condition requires both moment balance and exchange balance. In GdCo$F\bar{4}3m$8-type ferrimagnets this regime is predicted to combine nearly zero net moment across a broad temperature range with enhanced coercive fields and altermagnetic-like band features (Ali et al., 30 Jan 2026). This suggests that full compensation need not be confined to a single crossing temperature.

Compensation can also underwrite other collective phenomena. In Pt/Ni-doped $F\bar{4}3m$9, the fully compensated ferrimagnetic background coexists with nearby ferrimagnetic states only a few $|M_A|\neq |M_B|$00s of meV/f.u. higher in energy, and the exchange interaction between that background and uncompensated ferrimagnetic clusters produces exchange bias values in the range $|M_A|\neq |M_B|$01–$|M_A|\neq |M_B|$02 (Giri et al., 2021). This shows that fully compensated ferrimagnetism is not merely a small-moment limit of ferrimagnetism but can control the entire free-energy landscape.

5. Disorder, sample preparation, and phase stability

The experimental literature consistently shows that compensation is highly sensitive to stoichiometry, site occupancy, and synthesis route. In $|M_A|\neq |M_B|$03, arc-melting and annealing produce bulk materials with one magnetic-temperature profile, whereas melt-spinning and rapid quenching produce ribbons with different compensation temperatures, different $|M_A|\neq |M_B|$04, and even a different Néel type. The reported compositions deviate slightly from the nominal formulae, and the authors explicitly conclude that even a slight variation in stoichiometry and sample preparation method can influence the physical properties of a compensated system (Midhunlal et al., 2018).

Neutron diffraction has shown that disorder can be constructive rather than merely detrimental. In $|M_A|\neq |M_B|$05, the $|M_A|\neq |M_B|$06 compensated composition is not a perfectly ordered inverse Heusler. Instead, the key state is a disordered $|M_A|\neq |M_B|$07 structure with $|M_A|\neq |M_B|$08-Co disorder, and the compensation mechanism differs from previously reported $|M_A|\neq |M_B|$09. The identified ferrimagnetic motif is antiparallel coupling between the ferromagnetically aligned magnetic moments of $|M_A|\neq |M_B|$10 and $|M_A|\neq |M_B|$11 atom pairs (Midhunlal et al., 2020).

Cr$|M_A|\neq |M_B|$12Al pushes this point further. Comprehensive single-crystal XRD, synchrotron powder XRD, neutron powder diffraction, magnetization, and XMCD show complete Cr/Al site mixing in an A2-disordered Heusler structure, yet the alloy still exhibits a robust compensated ferrimagnetic state with a vanishingly small ordered moment of $|M_A|\neq |M_B|$13, $|M_A|\neq |M_B|$14, and spin-gapless-semiconducting transport. First-principles calculations on an A2-disordered SQS structure reproduce a negligibly small magnetization of $|M_A|\neq |M_B|$15 (Philip et al., 11 Dec 2025). This suggests that disorder can itself mediate the compensated state rather than merely perturb an ordered one.

At the same time, thermodynamic stability remains a major constraint. A formation-enthalpy study of chromium-based inverse Heuslers concluded that all investigated $|M_A|\neq |M_B|$16 compounds were unstable: $|M_A|\neq |M_B|$17 and $|M_A|\neq |M_B|$18 were stable with respect to elemental constituents but decomposed into binary phases, whereas $|M_A|\neq |M_B|$19, $|M_A|\neq |M_B|$20, $|M_A|\neq |M_B|$21, and $|M_A|\neq |M_B|$22 were unstable with respect to the elements themselves (Meinert et al., 2013). Experiment on $|M_A|\neq |M_B|$23 is consistent with this picture: the desired low-moment ferrimagnetic phase coexists with CoAl and Cr (Jamer et al., 2015). A plausible implication is that metastability and controlled disorder are not accidental complications but central parts of the materials design problem.

6. Two-dimensional, symmetry-engineered, and spin-ordering-induced realizations

Recent theory has extended fully compensated ferrimagnetism into low-dimensional and symmetry-controlled settings. One line of work treats two-dimensional fully compensated ferrimagnetism as a filling-enforced phase in which spin-split, ferromagnetic-like bands coexist with $|M_A|\neq |M_B|$24 because one spin channel is gapped and the electron filling enforces $|M_A|\neq |M_B|$25. Three realization schemes were proposed for 2D van der Waals materials: Janus structures, staggered potentials from electric fields or substrates, and element substitution or alloying. Candidate systems include bilayer NiICl, electric-field-driven bilayer CrI$|M_A|\neq |M_B|$26, electric-field-driven YI$|M_A|\neq |M_B|$27, and CrMoC$|M_A|\neq |M_B|$28S$|M_A|\neq |M_B|$29 (Liu et al., 21 Feb 2025).

A second line starts from altermagnetic metals. In monolayer $|M_A|\neq |M_B|$30, either electric field or uniaxial strain can break the $|M_A|\neq |M_B|$31 symmetry and drive a transition from an altermagnetic metal to an FC-FIM metal. In that framework, charge-carrier doping cannot generate a net magnetic moment in an altermagnet because $|M_A|\neq |M_B|$32, but it can do so in a fully compensated ferrimagnet because the spin-resolved densities of states are inequivalent (Guo et al., 5 Dec 2025).

A third route changes the spin order without changing the lattice. In a bilayer built from monolayer $|M_A|\neq |M_B|$33, first-principles calculations show that flipping the Néel vector of one layer can switch the same bilayer lattice between a $|M_A|\neq |M_B|$34-symmetric antiferromagnetic state and a fully compensated ferrimagnet with global non-relativistic spin splitting. The proposal was presented as “spin ordering-induced fully-compensated ferrimagnetism” (Guo et al., 14 Jul 2025).

Spontaneous interaction-driven fully compensated ferrimagnetism has also been proposed in Hubbard-model studies. Hartree–Fock calculations identify a broad stability regime of spontaneous filling-enforced fFIM, and the same work argues that defect engineering can realize spontaneous fFIM in nominally nonmagnetic graphene (Wang et al., 27 Feb 2026). In parallel, Janus Mn$|M_A|\neq |M_B|$35BrI monolayers have been proposed as fully compensated ferrimagnetic platforms with a Néel temperature above room temperature, spontaneous spin splitting from built-in layer-dependent electrostatic potential, SOC-induced valley polarization for an out-of-plane Néel vector, and AHE-compatible symmetry under appropriate hole doping (Lv et al., 24 Jun 2026).

These developments substantially broaden the meaning of “fully compensated ferrimagnetic metal.” The older Heusler-centered notion emphasized 24-electron bulk half-metals; the newer symmetry literature shows that fully compensated ferrimagnetism can also arise from broken rotational or mirror symmetries, from spin-order engineering, or from disorder-mediated filling constraints (Liu et al., 21 Feb 2025, Guo et al., 5 Dec 2025, Guo et al., 14 Jul 2025).

7. Functional consequences and applications

The main technological appeal of fully compensated ferrimagnetic metals is the coexistence of high spin functionality with negligible stray field. The Heusler literature repeatedly identifies such materials as candidates for spin-transfer-torque-based magnetic tunnel junctions, next-generation MRAM, and spin-polarized STM tips, because near-zero macroscopic moment suppresses magnetic cross-talk while high spin polarization at the Fermi level preserves efficient spin transport (Midhunlal et al., 2018, Stinshoff et al., 2016).

The application space is broader than spin torque alone. In Pt/Ni-doped $|M_A|\neq |M_B|$36, the compensated ferrimagnetic background supports gigantic exchange bias and an extra Hall effect associated with noncoplanar interfacial moments, making the material relevant for exchange-bias engineering and Hall-based readout (Giri et al., 2021). In two-dimensional filling-enforced fFIMs, theory predicts fully spin-polarized currents in half-metallic states, a non-zero anomalous Hall conductivity, and Kerr and Faraday responses despite zero net magnetization (Liu et al., 21 Feb 2025). In Janus Mn$|M_A|\neq |M_B|$37BrI and related ferrivalley proposals, valley polarization and anomalous valley Hall responses become symmetry-allowed specifically because the fully compensated ferrimagnet breaks $|M_A|\neq |M_B|$38 while retaining zero net moment (Lv et al., 24 Jun 2026, Xie et al., 17 Apr 2026).

Compensated ferrimagnetism also appears in magnonic and thermospintronic settings. A theoretical triple-$|M_A|\neq |M_B|$39 collinear state on the frustrated kagome lattice exhibits a compensated ferrimagnetic pattern at the triangle level and gives $|M_A|\neq |M_B|$40-wave-type spin splittings in both magnon and electron bands. The same work predicts an antiferromagnetic spin Seebeck effect at zero field in insulating systems and filling-controlled polarized states in metallic systems (Aoyama et al., 21 Apr 2026). This suggests that the fully compensated ferrimagnetic metal is not only a transport material but also a platform for zero-field spin-current generation.

Across these examples, a common principle emerges. Fully compensated ferrimagnetic metals are most valuable when the small net moment is not the result of weak magnetism but of strong, well-ordered, and oppositely directed sublattice moments. That condition supports high ordering temperatures, high spin polarization, exchange-bias phenomena, magneto-optical activity, and valley-selective transport, while preserving the central practical advantage of vanishing or near-vanishing macroscopic magnetization (Midhunlal et al., 2018, Giri et al., 2021, Liu et al., 21 Feb 2025).