Holmium disilicate (Ho2Si2O7) phonon-stable member of the rare-earth-silicate dielectric platform

Holmium disilicate (Ho2Si2O7) represents a deliberately constructed genus-broadening member of the rare-earth-silicate dielectric platform within the critical-mineral recovery and recycling separations portfolio. Its strategic purpose is not to be the highest-performing composition in the series — that distinction belongs to the yttrium and lutetium disilicate leads, which carry computed static dielectric tensors — but rather to demonstrate that phonon and thermal stability within the rare-earth disilicate structural family extends meaningfully beyond those leads, thereby strengthening the coverage and defensibility of the underlying composition-and-device-use claim family. This is a classic genus-fortification move: by establishing that a structurally distinct rare-earth member (here, holmium, sitting between dysprosium and erbium in the lanthanide series) is dynamically stable and survives thermal excitation at realistic operating temperatures, the claim family resists narrowing challenges that might otherwise argue the genus reduces only to the specifically exemplified members. The timing logic for this filing is straightforward. Advanced semiconductor packaging dielectrics — particularly redistribution-layer (RDL) dielectrics and glass-core panel dielectrics — are under intense material innovation pressure as 2.5D and 3D integration push interconnect pitch below five microns, demanding dielectrics that combine low loss, controlled permittivity, and chemical compatibility with glass substrates. The rare-earth disilicate structural class is attractive here because these compounds are chemically stable, crystallographically well-characterized, and amenable to film deposition. If the lead members of this platform demonstrate the performance targets the portfolio requires, the Ho2Si2O7 member preserves freedom to practice across the holmium compositional space and forecloses a competitor from independently claiming it as a workaround. That is its commercial function: genus insurance rather than genus flagship. Candidly, the asset is early. Phonon stability has been established computationally, but the dielectric permittivity, dielectric loss tangent, glass-transition temperature, and coefficient of thermal expansion for this specific composition have not yet been measured or computed. No performance claim is asserted for Ho2Si2O7 at this stage — only structural attestation. A buyer evaluating this asset should treat it as a defensive broadening position with measurable, well-defined validation gates remaining before it can support an independent performance narrative.

Candidates are screened for imaginary phonon modes with the machine-learning interatomic potentials named above. This is a computed harmonic check, not a measurement; where engines disagree, that is shown.

Ho2Si2O7 adopts the rare-earth disilicate crystal structure class, a well-characterized family of mixed silicon-oxygen tetrahedral frameworks in which the rare-earth ion occupies a high-coordination interstitial site. The material is indexed in the Materials Project database under mp-18662. Holmium's ionic radius (approximately 0.901 Å in eight-coordination) sits between that of dysprosium and erbium, placing it squarely in the mid-heavy lanthanide range where the lanthanide contraction begins to modestly tighten the unit cell relative to lighter members such as lanthanum or neodymium disilicate. This structural tightening has downstream implications for phonon branch stiffening and potentially for dielectric response, though those implications remain to be quantified for Ho2Si2O7 specifically. The computational validation performed for this composition centers on two independent simulation protocols using the MACE-MP-0 machine-learning interatomic potential. First, a phonon band structure calculation was executed on a 2x2x2 supercell with a 6x6x6 q-point mesh — a standard convergence target for disilicate-class unit cells. The result is unambiguously stable: the minimum phonon frequency across the entire Brillouin zone is +0.134 THz, with zero imaginary modes detected. A positive minimum frequency across all branches and all q-points is the operationally meaningful definition of dynamic stability, confirming that the crystal structure represents a true local minimum on the energy landscape rather than a saddle point or transition state. A material with imaginary modes at the zone boundary, for instance, would spontaneously reconstruct at finite temperature and could not serve as a reliable thin-film dielectric. Ho2Si2O7 passes this gate cleanly. The second protocol addresses finite-temperature behavior, which phonon calculations at 0 K cannot fully capture. A 2,000-step molecular dynamics trajectory was run with the MACE-MP-0 machine-learned potential (not ab initio MD) at 350 K, which is below the temperatures of glass-substrate packaging process steps and does not cover them. The trajectory shows no bond breaking, no amorphization, and no chemical dissociation events over its full duration. This machine-learned-potential MD result complements the zero-temperature phonon result: it suggests that thermal fluctuations at the simulated temperature do not kick the structure out of its free-energy basin. Together, the phonon calculation and the AIMD trajectory establish a consistent picture of a structurally and thermally robust compound. What remains open is equally important to state clearly. Only one machine-learning potential (MACE-MP-0) was applied to Ho2Si2O7. The lead members of this disilicate family — Y2Si2O7 and Lu2Si2O7 — have benefited from higher-level density functional perturbation theory (DFPT) calculations that yield static dielectric tensors and electronic bandgaps, which are the properties that directly underpin any performance claim for a packaging dielectric. No DFPT calculation has been performed for the holmium member, and no bandgap value is available. Similarly, no measured dielectric permittivity, loss tangent, glass-transition temperature, or coefficient of thermal expansion has been reported for a densified Ho2Si2O7 film. These are the open validation gates. Until at least the DFPT static dielectric tensor is computed and cross-checked against a second machine-learning potential or DFT benchmark, the holmium member cannot be promoted from a structural-attestation position to a performance-attested position in the claim family.

The primary commercial context for this asset is advanced semiconductor packaging, specifically the dielectric layer applications in panel-level redistribution layers and glass-core substrates. These are high-growth segments driven by the industry's transition to heterogeneous integration architectures — chiplets, 2.5D interposers, and 3D stacked packages — that require backend dielectrics with more precise electrical and thermomechanical properties than conventional polymer-based or standard oxide dielectrics. The market for advanced packaging materials broadly is estimated to be in the multi-billion-dollar range annually, with glass-core substrates and RDL dielectric films representing a rapidly expanding segment as major logic and memory manufacturers build out panel-level packaging capacity. The buyer logic for a genus-member asset like Ho2Si2O7 is not direct royalty from holmium-disilicate film sales — it is the contribution this member makes to the enforceability and breadth of the broader rare-earth-disilicate dielectric claim family. A claim family that can point to multiple independently confirmed stable members across the rare-earth series is substantially harder to design around than one relying on a single exemplified composition. From a licensing standpoint, the value of this member is inseparable from the value of the platform: it is appropriately bundled with the Y2Si2O7 and Lu2Si2O7 leads rather than licensed independently. The royalty model for the platform as a whole would most naturally attach to material supply agreements, process licenses, or device-level claims in packages using a rare-earth disilicate dielectric layer, not to individual composition claims in isolation. No addressable market figure should be attributed specifically to the holmium member at this stage. The composition contributes to claim coverage but has no demonstrated property advantage over the lead members and no independently validated dielectric performance. Any TAM discussion belongs at the platform level, where the Y/Lu leads anchor the performance narrative. The appropriate framing for this asset in a transaction is as a defensive-broadening member that reduces the risk of a competitor entering the holmium compositional space and narrows the attack surface for invalidation by narrowing the genus.

genus completeness / priority preservation around the EF5 disilicate leads; no asserted property advantage until DFPT/film gate closes

RDL/glass-core packaging-dielectric use distinguishes thermal-barrier / chamber-coating prior art (per EF5 lead)

Preliminary screen only. This is not a freedom-to-operate opinion.

The natural acquirers for this asset are the same parties who would be interested in the rare-earth-disilicate dielectric platform as a whole, because the value of the holmium member is inseparable from the platform it supports. Semiconductor packaging material suppliers — companies developing advanced dielectric films for panel-level RDL processes or glass-core substrate stacks — are the primary audience, particularly those who are already evaluating rare-earth oxide or silicate systems as candidates to replace conventional polymer dielectrics in next-generation packages. Substrate and interposer manufacturers building out glass-core processing capacity would also have strategic interest, since controlling the intellectual property around rare-earth disilicate dielectrics in that application space reduces future licensing exposure. A secondary acquirer category is defensive: a materials company or packaging IDM that wants to ensure freedom to operate in the holmium compositional space without risking a blocking position from this claim family. In that scenario, the acquisition is primarily about FTO insurance rather than active use of the technology. Given that the asset is a genus member rather than a lead composition, licensing it as part of a platform package — bundled with the Y2Si2O7 and Lu2Si2O7 leads — is the most commercially logical structure. Standalone licensing of Ho2Si2O7 before the dielectric property gate is closed would be difficult to price with confidence, as the performance case remains to be made.

The primary risk is the open validation gap: without a computed or measured dielectric tensor, bandgap, and thin-film physical properties for Ho2Si2O7 specifically, this composition cannot support an independent performance-based narrative. A buyer must either (a) accept the structural-attestation-only claim at the current stage and plan to close the DFPT and measurement gates post-acquisition, or (b) wait for those gates to close before assigning full value to this member. The computational work required — a DFPT static dielectric tensor calculation using DFT, ideally followed by cross-validation with a second machine-learning potential — is well-defined and executable within a reasonable timeframe using the Lattice Graph platform's existing workflows. The measurement work (depositing and characterizing a densified Ho2Si2O7 film) is more time- and resource-intensive, but it is the same measurement workflow already targeted for the lead compositions. A secondary risk is the single-potential computational basis: the phonon and AIMD results rely entirely on MACE-MP-0 without corroboration from CHGNet, MatterSim, or ORB, which are standard in the portfolio's consensus protocol. This does not mean the stability conclusion is wrong — MACE-MP-0 is a high-quality, broadly validated potential — but a skeptical reviewer could note the absence of multi-potential consensus. Running the three additional potentials on this composition is a low-cost, near-term de-risking step. A third consideration is holmium's status as a critical mineral with a constrained global supply chain, which could create raw material access risks at scale if a Ho2Si2O7-based dielectric were ever commercialized; however, at this stage of development and given the asset's role as a genus member rather than a preferred embodiment, this supply-chain consideration is a second-order concern relative to the validation gaps.