Rare-earth trifluoride low-loss dielectric resonators for millimeter-wave and 6G applications
NdF3, SmF3, and TbF3 trifluoride ceramics combine moderate permittivity (~8-15) with a computed loss tangent below 1e-3 at 100 GHz, targeting high-Q resonators for 30 to 300 GHz wireless infrastructure.
The opportunity
LnF3 (Nd/Sm/Tb preferred) mmWave dielectric resonators for 30-300 GHz: moderate eps (~8-15), low loss tangent (<1e-3 @100 GHz, computed). NdF3 DFPT eps_total ~12, eps_inf ~4, tan delta 1e-4 to 1e-3. Preferred trivalent cations are paramagnetic (NOT closed-shell, corrected per candor aaa); preference re-based on computed phonon-anharmonic loss tangent; magnetic-loss characterization reserved.
Investment thesis
Rare-earth trifluorides have been largely overlooked as candidates for millimeter-wave dielectric components, while the global wireless industry is under mounting pressure to find ceramic resonator materials that maintain low dielectric loss as carrier frequencies push from sub-6 GHz into the 30-300 GHz range demanded by next-generation dense wireless infrastructure. The conventional workhorse materials — alumina, PTFE composites, and alkaline-earth titanates — perform acceptably below 30 GHz but suffer increasing phonon-anharmonic loss at higher frequencies, degrading resonator Q factors precisely where 6G system budgets are tightest. Lattice Graph's computational screen identified that lanthanide trifluorides, particularly neodymium trifluoride (NdF3) in the tysonite structure, occupy a permittivity window of roughly 8-15 — modest enough to allow compact resonator geometries — while phonon-anharmonic calculations suggest loss tangents in the range of 1×10⁻⁴ to 1×10⁻³ at 100 GHz, which would place them among the lowest-loss ceramic candidates at these frequencies. This asset is a lead filing in the rare-earth trifluoride millimeter-wave dielectric family. It claims composition and device-use rights across a defined set of trivalent lanthanide trifluorides, with NdF3, SmF3, and TbF3 as the preferred members on the basis of computed phonon-anharmonic loss rather than electronic-structure arguments about closed-shell vs. open-shell configurations. The timing aligns with a forced-substitution moment: 6G basestation and backhaul component manufacturers are making materials platform decisions now, ahead of spectrum allocation and infrastructure build-out scheduled in the late 2020s and early 2030s, and patent positions in dielectric resonator compositions are far less crowded at millimeter-wave frequencies than at microwave frequencies below 10 GHz.
Asset rating
Material identity
- Formula
- NdF3
- Class
- rare-earth trifluoride
- Space group
- P-3c1 (tysonite)
Computational validation
How this candidate was proven in silico — multiple independent physics engines, not a single model
Each candidate is validated by multiple independent machine-learning interatomic potentials. A material advances only when the engines agree on phonon (dynamic) stability — disagreement is surfaced, not hidden.
Technical deep-dive
NdF3 crystallizes in the tysonite structure (space group P-3c1), a hexagonal framework in which neodymium is nine-coordinate and fluorine occupies three distinct Wyckoff sites. This structure is mechanically robust and chemically stable in the absence of moisture, properties shared by the other rare-earth trifluorides across the lanthanide series. What distinguishes this family for millimeter-wave applications is the combination of a moderate total permittivity (DFPT-computed εtotal approximately 12 for NdF3, with εinf approximately 4) and an ionic contribution that, while significant, is governed by high-frequency optical phonon modes with relatively weak anharmonic coupling. The gap between the static and high-frequency dielectric constants — the ionic contribution — is the primary driver of microwave and millimeter-wave loss through two-phonon difference processes, and the computed phonon-anharmonic loss tangent for NdF3 falls in the range of 1×10⁻⁴ to 1×10⁻³ at 100 GHz under the simulation labeled R-NDF3-001. For context, loss tangents below 1×10⁻³ at 100 GHz are exceptional; most commercial ceramics used as resonators today exhibit loss tangents in the range of 1×10⁻³ to 5×10⁻³ at these frequencies. The computational validation was carried out using density functional perturbation theory (DFPT) for the permittivity tensor and phonon frequencies, combined with phonon-anharmonic perturbation theory to estimate intrinsic loss. Two independent machine-learning interatomic potentials — MACE and CHGNet — were run against the NdF3 structure and both agree that the material is dynamically stable: the phonon dispersion shows no imaginary modes anywhere in the Brillouin zone, confirming that the tysonite phase is not on the edge of a structural instability that would artificially suppress low-frequency modes and inflate apparent loss. This two-potential consensus is a meaningful bar; a structure that appears stable under DFT alone can still harbor pathological soft modes when a more transferable potential is applied. The DFPT calculation was performed on the DFT-relaxed structure (one DFT source), providing the full dielectric tensor and mode-resolved Born charges needed for the loss-tangent estimate. The broader rare-earth trifluoride family (La, Ce, Nd, Sm, Gd, Tb, Dy, Er) was screened for inclusion in the composition claims, with Nd, Sm, and Tb identified as preferred members based on their computed phonon-anharmonic loss performance rather than closed-shell electronic arguments. Notably, these preferred members carry paramagnetic lanthanide ions (Nd³⁺, Sm³⁺, Tb³⁺), which introduces a secondary loss mechanism — magnetic dipole coupling — that must be characterized experimentally. This is flagged openly: the computational screen addresses phonon-anharmonic (lattice) loss, not spin-lattice relaxation loss, and the two are additive at millimeter-wave frequencies. The relative magnitude of the magnetic contribution for these specific ions at room temperature is not yet quantified. The dielectric permittivity range of approximately 8-15 across the lanthanide trifluoride series is strategically valuable. Higher-permittivity resonators (ε > 20) demand precise dimensional tolerances because resonant frequency scales inversely with the square root of permittivity, and thermal expansion shifts the frequency; lower permittivity (ε < 5) requires physically larger resonator bodies that resist integration into compact module packages. The 8-15 window is well matched to the thermal budget and form-factor requirements of 5G and emerging 6G radio units and beamforming modules. The tysonite phase is also anisotropic, meaning the dielectric tensor components perpendicular and parallel to the c-axis differ; this anisotropy can be exploited by orienting the resonator to tune the effective permittivity experienced by the resonant mode, providing a materials-level degree of freedom that isotropic cubic ceramics lack.
Market & opportunity sizing
The addressable market for dielectric resonators and filters in millimeter-wave wireless infrastructure is estimated at $1-5 billion, scaling with the pace of 6G buildout. This estimate encompasses basestation dielectric resonator oscillators (DROs), filter banks in the 28 GHz, 39 GHz, 60 GHz (WiGig/backhaul), and E-band (70-80 GHz) licensed spectrum windows, plus the emerging D-band (110-170 GHz) bands under active 6G standardization discussion at ITU-R. Component-level, each macro basestation contains multiple dielectric resonator filters, and the unit economics of high-Q ceramic elements at millimeter-wave frequencies support premium pricing relative to lower-frequency analogs because the manufacturing yield and loss-performance constraints are far tighter. These are estimates based on industry projections for 6G capex; actual market realization depends on spectrum policy timelines and operator investment cycles. The customer set is concentrated: the major RF front-end module suppliers (component makers for basestation OEMs and handset integrators) are the natural buyers or licensees. These companies already purchase or manufacture dielectric resonator materials for 4G/5G sub-6 GHz applications and are actively seeking drop-in or next-generation materials for the frequency extension into millimeter-wave. A licensing or supply-agreement model is the most natural commercialization path — the intellectual property covers the composition and device-use claim, meaning a ceramic powder producer, a sintering specialist, or a module integrator can be licensed under the patent family without requiring Lattice Graph to operate a manufacturing line. Royalty rates in advanced ceramic composition patents typically range from 1-5% of component revenue, which on even a modest unit-share of the millimeter-wave resonator market produces a commercially meaningful licensing stream. The race window is real: 6G spectrum auctions and pre-commercial deployments are expected in the 2028-2032 timeframe, and component supply chains are typically locked in two to three years before volume manufacturing, meaning the design-win and materials-qualification cycle has already begun.
Market & competitive position
low-loss high-Q dielectric resonator for 30-300 GHz
The incumbent dielectric resonator materials for microwave applications are alumina (Al₂O₃, ε ≈ 9-10, tan δ ≈ 1×10⁻⁴ at 10 GHz), barium titanate-based ceramics (high ε but higher loss at millimeter-wave), and temperature-compensated PTFE composite laminates. Alumina is the most credible comparator because its permittivity overlaps with the lanthanide trifluoride window. However, alumina's loss tangent increases meaningfully above 30 GHz due to phonon scattering from grain boundaries and residual porosity in sintered parts, and its intrinsic phonon-anharmonic loss rises faster with frequency than the NdF3 computation suggests. Barium strontium titanate and lead-free titanate ceramics offer higher permittivity but at the cost of significantly elevated loss tangents at millimeter-wave frequencies, driven by the proximity of ferroelectric soft modes. No ceramic dielectric resonator material currently has both a validated loss tangent below 1×10⁻³ at 100 GHz and an established volume manufacturing base at that frequency — the field is genuinely open at 100 GHz, which is where this filing is positioned. Rare-earth oxide ceramics (Y₂O₃, La₂O₃) and rare-earth aluminate garnets (e.g., Y₃Al₅O₁₂) have been explored in academic literature for millimeter-wave substrates but not widely patented as dielectric resonator compositions for the 30-300 GHz range. The fluoride chemistry introduces advantages: the ionic fluoride bond is highly polarizable without introducing lossy d-band or f-band electronic contributions at the relevant photon energies, the crystal structures are well-ordered (minimizing defect scattering), and synthesis via fluoride precipitation routes is reasonably well-established in the rare-earth processing industry. The competitive moat here is the computational prioritization of specific lanthanide members on phonon-anharmonic grounds — a distinction that prior art based on empirical microwave measurements below 10 GHz would not have been motivated to make — combined with freedom-to-operate screening showing a clean landscape for these compositions in this application space.
| This asset | Incumbents |
|---|---|
| low-loss high-Q dielectric resonator for 30-300 GHz | alumina/ceramic resonators |
Claims & IP position
What's claimed, the protected family, and the freedom-to-operate read
The claims in this family cover two principal branches: composition claims on the rare-earth trifluoride ceramic (LnF3, where Ln is drawn from a defined set including La, Ce, Nd, Sm, Gd, Tb, Dy, and Er) and device-use claims on its application as a dielectric resonator operating in the 30-300 GHz frequency range. The preferred embodiments center on NdF3, SmF3, and TbF3 as the members with the most favorable computed phonon-anharmonic loss profiles, though the composition claims are drafted to encompass the full series of enumerated lanthanides to prevent design-around substitution. A key negative limitation has been incorporated: the preference for Nd, Sm, and Tb is explicitly grounded in their computed loss tangents rather than in any closed-shell electronic argument, which corrects a potential claim-drafting vulnerability (paramagnetic open-shell members might otherwise have been assumed unfavorable and excluded). The device-use claim structure ties composition to operating frequency range, which is deliberate. Generic rare-earth fluoride compositions appear in the literature for optical applications (fluoride crystals for lasers, scintillators) and for PFAS-treatment contexts, but no prior art is known to claim these compositions specifically as dielectric resonators at millimeter-wave frequencies. The frequency-specific device-use claim provides a second layer of protection even if the composition claim were successfully challenged on a narrow novelty ground. The family is designated as a lead filing, meaning it anchors the rare-earth trifluoride millimeter-wave dielectric patent family and is expected to carry continuation and divisional prosecution covering process (sintering conditions, grain-boundary engineering) and system-level claims (resonator-in-module assemblies) as experimental data matures.
- Claim type
- Composition+device_use
- Drafted claims
- 2 claims
- Freedom to operate
- Clear path
- Blocking patents
- None found — white space
Freedom-to-operate analysis across more than 300,000 materials patents returns a clean status for this asset — no identified blocking art covers rare-earth trifluorides as dielectric resonator compositions for millimeter-wave or 6G frequency applications. Rare-earth fluorides appear in the patent literature primarily in optical (laser gain media, scintillator hosts) and specialty coating contexts; the millimeter-wave dielectric resonator application space is unoccupied by prior composition claims on lanthanide trifluorides. The main FTO risk to monitor is any future filing by a ceramic supplier or RF component maker who independently arrives at the same composition family through empirical screening — this is a realistic scenario given that the millimeter-wave resonator materials problem is well-known in the RF industry and several large component manufacturers run internal materials programs. The current window of freedom is clear; maintaining it requires timely prosecution and, as experimental data is generated, rapid conversion to continuation filings that add process and device claims.
Validation roadmap
What's proven so far, and what a buyer would fund next
The completed computational work on this asset consists of a DFPT simulation (R-NDF3-001) that delivers the full dielectric permittivity tensor for NdF3 (εtotal approximately 12, εinf approximately 4) and a phonon-anharmonic loss-tangent estimate of 1×10⁻⁴ to 1×10⁻³ at 100 GHz. Structural dynamic stability was independently confirmed by two machine-learning interatomic potentials — MACE and CHGNet — both of which show a phonon dispersion with no imaginary modes, meaning the tysonite structure is not at risk of spontaneous distortion that would invalidate the loss calculation. One DFT calculation anchors the electronic structure and serves as the reference for both the dielectric tensor and the potential-energy surface used in anharmonic perturbation theory. The agreement between two independent potentials on the absence of soft modes is a material result: it means the stability is not an artifact of a single model's training set. What remains open is substantial and should be stated directly. Measured loss tangents at 100 GHz on sintered NdF3 ceramic do not yet exist in this program, and the computed loss tangent must be treated as an order-of-magnitude estimate pending experimental confirmation. The phonon-anharmonic perturbation approach accounts for intrinsic lattice loss but does not capture extrinsic contributions from grain boundaries, point defects, moisture absorption, and impurity scattering phases, which dominate loss in real sintered ceramics and typically elevate measured tan δ by one to two orders of magnitude above the single-crystal intrinsic floor. Additionally, the preferred members (Nd³⁺, Sm³⁺, Tb³⁺) are paramagnetic, and the spin-lattice relaxation loss contribution at room temperature and 30-300 GHz has not been simulated or measured. For diamagnetic members of the family such as LaF3 or LuF3, this concern does not apply, but those members are not identified as preferred in the current computation. The two principal validation gates — direct microwave/millimeter-wave loss measurement on sintered pellets, and magnetic-loss characterization of paramagnetic members — are the critical next experimental steps before this asset can be positioned as a fully de-risked, manufacturing-ready material.
- Independent DFT references
- 1
- Evidence receipts
- 4
Applications
Strategic fit & buyers
The natural acquirers and licensees for this asset are RF component manufacturers and advanced ceramics suppliers with active millimeter-wave product programs. On the component side, the large front-end module makers and dielectric filter specialists serving the basestation market have the manufacturing infrastructure to qualify and sinter new ceramic compositions and the sales channels to place them in 6G supply chains. On the materials side, rare-earth chemical and ceramics companies with existing lanthanide processing capabilities could license the composition claims to supply sintered resonator blanks to module integrators, without needing to own the downstream RF characterization infrastructure. A secondary buyer class is the semiconductor and packaging companies developing heterogeneous integration platforms for 6G radio units, where low-loss dielectric resonators are needed as co-packaged components; for these buyers the device-use claim is the primary value, and they would seek a license to incorporate NdF3 or SmF3 resonators into their module architecture without manufacturing the ceramic themselves. Strategic fit is strongest for companies that are currently constrained by alumina performance above 60 GHz and are actively evaluating alternative ceramic hosts. The asset also fits naturally within any portfolio acquisition strategy targeting 6G materials IP, particularly one that pairs dielectric resonator compositions with antenna substrate and waveguide materials claims — a bundled licensing position across multiple component types in the 6G front-end stack commands substantially higher valuation than any single composition claim in isolation. Within Lattice Graph's own integrated packaging, storage, and PFAS-treatment systems portfolio, this asset represents a deliberate extension into a high-value RF materials application that shares computational methodology with the rest of the portfolio but serves a distinct commercial end market, making it equally suitable for standalone licensing or portfolio sale.
Risks & roadmap
The primary technical risk is the gap between the computed intrinsic loss tangent and the likely measured loss tangent in sintered ceramic. Phonon-anharmonic calculations give the single-crystal lower bound; real ceramics are sintered polycrystalline bodies with grain boundaries, residual porosity, and trace second phases that scatter both phonons and electromagnetic fields. Industry experience with high-Q microwave ceramics suggests that achieving a measured tan δ within a factor of 2-5 of the intrinsic limit requires extremely controlled sintering, hot-pressing or spark-plasma sintering techniques, and high-purity starting powders. There is also the unresolved magnetic-loss question: at room temperature, the paramagnetic susceptibility of Nd³⁺ and Tb³⁺ is moderate, and for most RF applications the spin-lattice relaxation time at millimeter-wave frequencies is short enough that magnetic loss is small, but this has not been demonstrated experimentally for these specific compositions. The diamagnetic members (La, Ce in certain oxidation states, Lu) sidestep this concern but are not the computed-preferred members. The roadmap to de-risk the asset is straightforward in sequence if not in timeline: sinter NdF3 and SmF3 pellets from commercially available fluoride precursors under inert atmosphere, measure loss tangent by the Hakki-Coleman or TE01δ resonator method at 30-100 GHz, and characterize magnetic-loss contribution by comparing diamagnetic (LaF3) and paramagnetic (NdF3, TbF3) members at identical microstructural quality. If measured loss falls within the predicted 1×10⁻³ range at 100 GHz, the asset transitions from computationally validated to experimentally supported, substantially increasing its licensing value and enabling process-claim continuation filings grounded in demonstrated sintering protocols. The FTO landscape is currently clean but requires monitoring; accelerated prosecution to grant is advisable ahead of the 6G design-win window opening around 2027-2028.
More in Integrated systems
Related assets in the same portfolio — each a separately filed position
License or acquire Rare-earth trifluoride low-loss dielectric resonators for millimeter-wave and 6G applications
Request the full data room: complete claim set, proof packet, FTO memo, and licensing / acquisition terms.