Double Trillium Lattice

• Double Trillium Lattice is a 3D network of two interpenetrating trillium lattices formed from corner-sharing triangles, resulting in intrinsic chirality and geometric frustration.
• It underpins diverse magnetic phenomena including chiral helimagnetism, skyrmion-lattice signatures, and field-induced quantum spin liquid states in compounds like EuPtSi and K2Ni2(SO4)3.
• Studies reveal magnetization plateaus and reentrant symmetry recovery driven by competing exchange interactions and Dzyaloshinskii–Moriya coupling, highlighting its role in advanced magnetic phase behavior.

The double trillium lattice denotes two interpenetrating trillium networks within the same crystal. A trillium lattice is a three-dimensional network of corner-sharing equilateral triangles generated in chiral cubic structures of space group $P2_13$ from atoms on the $4a$ Wyckoff position, so geometric frustration and structural chirality are built into the connectivity. In binary B20 intermetallics the two atomic species each form a trillium network, whereas in langbeinite-family insulators two crystallographically inequivalent magnetic sites form two independent trillium sublattices, producing a bona fide double trillium lattice of spins; ternary compounds such as EuPtSi generalize the motif to three interpenetrating $4a$ networks (Khatua et al., 27 Nov 2025).

1. Definition and structural scope

In the canonical B20 setting, atoms occupy the special Wyckoff $4a$ sites with fractional coordinates $(x,x,x)$ and symmetry equivalents. The resulting site set forms a three-dimensional network of corner-sharing triangles known as the trillium lattice. In binary B20 compounds such as MnSi and FeGe, both atomic species occupy inequivalent $4a$ positions with different internal coordinates and each species forms its own trillium network. Because the two trillium nets interpenetrate, one speaks of a double trillium lattice. The structure is non-centrosymmetric and enantiomorphic, so inversion symmetry is absent and the crystal has a definite handedness (Khatua et al., 27 Nov 2025).

In insulating langbeinites and related phosphates, the phrase has a more literal magnetic meaning. In K$_2$Ni$_2$(SO$_4$)$_3$, K$4a$0Co$4a$1(SO$4a$2)$4a$3, KSrFe$4a$4(PO$4a$5)$4a$6, KBaFe$4a$7(PO$4a$8)$4a$9, and KBaCr$4a$0(PO$4a$1)$4a$2, two crystallographically inequivalent magnetic sites each form a three-dimensional network of corner-sharing triangles, and the two magnetic trillium sublattices are intertwined by additional inter-sublattice exchange paths (Zivkovic et al., 2021). In K$4a$3Co$4a$4(SO$4a$5)$4a$6, for example, Co1 and Co2 each independently form a trillium lattice and together yield a double trillium topology that survives a structural transition from high-temperature cubic $4a$7 to low-temperature monoclinic $4a$8 with a three-fold superstructure (Magar et al., 11 Aug 2025).

The same structural logic extends beyond the two-sublattice case. EuPtSi is a LaIrSi-type B20 derivative in space group $4a$9 in which Eu, Pt, and Si all occupy the $4a$0 site, so each element forms a trillium network. The magnetic Eu sublattice therefore sits on one trillium net embedded within two symmetry-equivalent nonmagnetic trillium networks, which directly generalizes the double-trillium concept to a multiple-trillium setting (Kaneko et al., 2018).

2. Geometric frustration, chirality, and exchange topology

The defining magnetic consequence of the trillium geometry is frustration on corner-sharing triangular motifs. For antiferromagnetic exchange, no spin configuration can satisfy all pairwise bonds on a triangle simultaneously. In a single trillium lattice this already generates a highly degenerate manifold; in a double trillium lattice the interpenetration of two such networks adds further competing constraints because intra- and inter-sublattice exchanges generally favor incompatible correlations (Gonzalez et al., 13 May 2026).

Chirality is equally central. In $4a$1, the absence of inversion symmetry permits antisymmetric Dzyaloshinskii–Moriya exchange on bonds. A minimal spin Hamiltonian used for chiral trillium magnets is

$4a$2

where the symmetric exchange $4a$3, DM vectors $4a$4, Zeeman term, and anisotropy collectively determine whether the ordered state is collinear, helical, or multi-$4a$5 (Kaneko et al., 2018). In the weak-chirality continuum limit, the helical pitch obeys $4a$6, so the ratio of antisymmetric to symmetric exchange sets the modulation scale.

For langbeinite double trillium systems, a widely used starting point is the isotropic Heisenberg model in a magnetic field,

$4a$7

In K$4a$8Ni$4a$9(SO$(x,x,x)$0)$(x,x,x)$1, this exchange network separates naturally into a strong-TL with $(x,x,x)$2 K, a weak-TL with $(x,x,x)$3 K, and inter-trillium couplings dominated by $(x,x,x)$4 K, with weaker $(x,x,x)$5 K and $(x,x,x)$6 K. The 2026 analysis emphasizes that this hierarchy places the material close to a spin-liquid region surrounding a “tetratrillium” limit in which each triangle of the strong-TL is connected to a single weak-TL spin, turning the triangle into an effective tetrahedron (Gonzalez et al., 13 May 2026). A closely related description in the 2025 review states that dominant inter-sublattice couplings can connect the two trillium nets into a “hypertrillium” topology of corner-sharing tetrahedra (Khatua et al., 27 Nov 2025). This suggests that coupled trillium systems interpolate between triangle-based and tetrahedron-based frustration without leaving the chiral $(x,x,x)$7 framework.

3. Chiral helimagnetism and skyrmion signatures in EuPtSi

EuPtSi is a chiral trillium-lattice antiferromagnet in which neutron diffraction establishes magnetic order at $(x,x,x)$8 K. In the ground state, magnetic peaks appear at

$(x,x,x)$9

and its cyclic permutations. Upon heating, a first-order commensurate–incommensurate lock-in transition occurs around $4a$0 K, producing

$4a$1

The coexistence of central and satellite peaks together with hysteresis identifies the transition as first order (Kaneko et al., 2018).

Half-polarized neutron diffraction resolves the handedness of the helix. The relevant polarization-resolved intensity is

$4a$2

and EuPtSi shows much stronger intensity in the $4a$3 channel than in $4a$4. The measured flipping ratios near $4a$5 are approximately $4a$6 in the ground state and $4a$7 in the incommensurate phase. This establishes a single-handed helical structure with ordered moments perpendicular to $4a$8, and the chirality is unchanged across $4a$9 (Kaneko et al., 2018).

Under a magnetic field $_2$0, EuPtSi enters an $_2$1 phase near $_2$2 T and $_2$3 K. In this phase, magnetic peaks move into the equatorial plane perpendicular to the field and form a hexagonal pattern around nuclear Bragg positions. The ordering vectors are

$_2$4

restricted to cyclic permutations with a fixed sign order, and $_2$5 is similar to $_2$6. The sixfold reciprocal-space pattern together with the field selection rule $_2$7 matches the classic skyrmion-lattice fingerprint seen in canonical chiral B20 magnets, and the EuPtSi paper states that the pattern could be the hallmark of a formation of skyrmion lattice in EuPtSi (Kaneko et al., 2018). Within the broader double-trillium context, EuPtSi shows how the chiral trillium geometry favors uniform helical chirality in zero field and multi-$_2$8 superpositions under field (Khatua et al., 27 Nov 2025).

4. Double-trillium langbeinates under magnetic field

K$_2$9Ni$_2$0(SO$_2$1)$_2$2 provides the most detailed magnetic case study of a bona fide double trillium lattice with two magnetic $_2$3 sublattices. Density functional theory identified five relevant exchanges up to about $_2$4 Å: $_2$5 K, $_2$6 K, $_2$7 K, $_2$8 K, and $_2$9 K. The strongest coupling is the interlattice antiferromagnetic $_4$0, while $_4$1 and $_4$2 frustrate the two trillium lattices independently. Experimentally the compound has $_4$3 K, a predominantly dynamic low-temperature state, a broad “liquid-like” diffuse maximum in polarized neutron scattering at $_4$4 Å$_4$5, and a field-induced quantum spin liquid for $_4$6 T, where the small static component becomes undetectable and $_4$7 retains a nearly quadratic low-$_4$8 dependence up to $_4$9 T (Zivkovic et al., 2021).

A later pulsed-field and classical Monte Carlo study extended the field scale to $_3$0 T and resolved a more structured magnetization process. Saturation occurs at $_3$1 T, corresponding in simulation units to $_3$2. Most notably, the calculations reveal a pronounced dome in the $_3$3–$_3$4 phase diagram centered near $_3$5, or $_3$6 T, whose experimental entry and exit are tracked by anomalies in $_3$7 (Gonzalez et al., 13 May 2026).

Inside that dome, the weak-TL is fully polarized, $_3$8, while the strong-TL adopts an up–up–down configuration, $_3$9, on every triangle. For equal populations of the two trillium sublattices,

$4a$00

so

$4a$01

The paper therefore identifies a $4a$02 magnetization plateau-like phase. In the classical limit no true $4a$03 plateau is expected; instead a finite-temperature dome stabilizes this configuration, producing reentrant behavior in which the system recovers the Hamiltonian symmetries as the field is increased further (Gonzalez et al., 13 May 2026).

The same work shows that once the weak-TL saturates, it acts as an effective field on the strong-TL:

$4a$04

This maps the double-trillium problem onto an effective single-trillium problem and explains why closely related plateau-like domes also occur in the single TL and in the tetratrillium limit (Gonzalez et al., 13 May 2026). A plausible implication is that magnetization plateaus and reentrant symmetry recovery are not isolated to one compound but reflect a generic field response of coupled trillium networks near a tetrahedral-coupling regime.

5. Quantum, high-spin, and oxide realizations

K$4a$05Co$4a$06(SO$4a$07)$4a$08 establishes a pseudospin-$4a$09 double trillium regime. Below $4a$10 K, magnetization and heat capacity are consistent with a $4a$11 Kramers doublet of Co$4a$12 with $4a$13 and $4a$14 K. In zero field, static order appears only below $4a$15 K and remains incomplete: $4a$16SR is best fit by a two-component exponential with an approximately $4a$17 fluctuating fraction, and the static order is fully suppressed by a small field of about $4a$18 T. Above that field the low-temperature heat capacity follows $4a$19 with $4a$20, which the paper treats as a fingerprint of proximate quantum spin liquid behavior on the double trillium lattice (Magar et al., 11 Aug 2025).

High-spin realizations display a broader phenomenology. KBaFe$4a$21(PO$4a$22)$4a$23 hosts two interpenetrating Fe$4a$24 trillium sublattices in chiral $4a$25, five antiferromagnetic exchanges, and $4a$26 K, yet no magnetic long-range order is observed down to $4a$27 mK. NMR and $4a$28SR reveal neither a static internal field nor spin freezing; instead the system supports two coexisting dynamic relaxation channels, a dominant fast sporadic channel and a slower Markovian component, together with $4a$29 below about $4a$30 K (Sebastian et al., 11 Nov 2025). By contrast, the closely related KSrFe$4a$31(PO$4a$32)$4a$33 was initially discussed as a spin-liquid candidate, but $4a$34P NMR later showed a short-range spin-freezing state below $4a$35 K: the linewidth becomes nearly field-independent, the extrapolated zero-field internal-field width is $4a$36 T, and $4a$37 remains very large and nearly temperature independent below $4a$38, so persistent dynamics coexist with quasi-static short-range order rather than a pure gapless spin liquid (Sebastian et al., 6 Nov 2025).

FeGe$4a$39O$4a$40 extends the double-trillium concept to oxide chemistry. It crystallizes in chiral cubic $4a$41, the Fe atoms form what the authors term a double-trillium lattice with the shortest Fe–Fe distance about $4a$42 Å, and magnetization at $4a$43 K reaches $4a$44/Fe$4a$45 at $4a$46 kOe without saturation. The susceptibility gives $4a$47 per Fe and $4a$48 K, while specific heat shows a broad magnetic maximum near $4a$49 K and only about $4a$50 of the expected $4a$51 entropy is recovered by $4a$52 K. Neutron scattering detects no magnetic Bragg peaks down to $4a$53 K (Boswell et al., 13 Jan 2026).

A common misconception is that a double trillium topology automatically implies a highly frustrated or spin-liquid-like antiferromagnet. KBaCr$4a$54(PO$4a$55)$4a$56 is the clearest counterexample. There, the spin lattice comprises two crystallographically nonequivalent ferromagnetic sublattices that are coupled antiferromagnetically. The dominant exchanges are $4a$57 K, $4a$58 K, $4a$59 K, $4a$60, and $4a$61 K, leading to a collinear $4a$62 two-sublattice antiferromagnet and a frustration parameter $4a$63; the paper explicitly states that frustration is eliminated in this trillium network (Kolay et al., 2024).

6. Terminological ambiguities and broader significance

The phrase “double trillium lattice” is used in at least two distinct ways. In the standard crystallographic and materials-physics sense, it refers to two interpenetrating trillium networks in the same crystal. In B20 intermetallics, two trillium sublattices are always present by symmetry, but only one is usually magnetic, so the magnetism is governed by a single trillium subnetwork even though the crystal contains a double trillium framework. In langbeinite-family insulators, by contrast, the two sublattices are both magnetic and the term denotes a true double trillium lattice of spins (Khatua et al., 27 Nov 2025).

A second, explicitly nonstandard usage appears in the study of the line graph of the trillium lattice. That paper introduces the name “trilline lattice” for the non-bipartite network obtained by placing sites on trillium bonds and connecting incident bonds. It states that, if by “Double Trillium Lattice” one means the line graph of the trillium lattice, then it is synonymous with the trilline lattice, but it also notes that the paper itself does not use the term “Double Trillium” (Fancelli et al., 2024). In that model, the classical Heisenberg Hamiltonian

$4a$64

defines a three-dimensional fragile spin liquid with exponentially decaying correlations and fractional $4a$65 orphan moments under dilution, while the Ising limit maps to triple dimer coverings and a quantum dimer model with $4a$66 winding-parity sectors (Fancelli et al., 2024). This usage is therefore related to, but distinct from, the mainstream materials usage.

Across materials classes, the double trillium lattice functions as a unifying geometry for several otherwise disparate phenomena: DM-enabled helimagnetism and skyrmion-lattice signatures in EuPtSi, field-induced quantum spin liquid behavior and plateau-like dome physics in K$4a$67Ni$4a$68(SO$4a$69)$4a$70, pseudospin-$4a$71 proximate quantum spin liquid behavior in K$4a$72Co$4a$73(SO$4a$74)$4a$75, dynamically inhomogeneous high-spin states in Fe phosphates, and frustration without long-range order in oxide FeGe$4a$76O$4a$77 (Kaneko et al., 2018). Ternary compounds such as CeIrSi show that even when only one of several interpenetrating trillium-derived sublattices is magnetic, the same corner-sharing-triangle topology can suppress long-range order and produce broad thermodynamic maxima with coherent low-temperature transport (Kneidinger et al., 2019).

The accumulated literature therefore identifies the double trillium lattice not as a single magnetic phase but as a structurally chiral, topologically frustrated platform. Its defining ingredients are interpenetrating corner-sharing-triangle networks, exchange competition across and within those networks, and symmetry-allowed antisymmetric exchange in $4a$78-derived settings. This combination is sufficient to support collinear antiferromagnets, helical single-chirality states, skyrmion-lattice signatures, field-induced dynamic regimes, reentrant plateau-like phases, and short-range frozen states, depending on the exchange hierarchy and spin degree of freedom (Khatua et al., 27 Nov 2025).