Dielectric-shelled titanium, zirconium, and hafnium diboride filler for electrically isolated packaging

The transition-metal diborides — titanium diboride (TiB2), zirconium diboride (ZrB2), hafnium diboride (HfB2), and the ternary boride AlMgB14 — are among the stiffest, most thermally conductive ceramic fillers available to the packaging industry. Their bulk thermal conductivities and elastic moduli rival or exceed commercial alternatives such as aluminum nitride and boron carbide, making them attractive hard-phase fillers for thermally demanding applications. The obstacle that has historically kept them out of sensitive package locations is a straightforward but consequential one: uncoated diborides are electrically conductive, and at the loadings required for effective thermal transport they breach the percolation threshold, creating leakage paths that make them incompatible with any electrically active or voltage-sensitive region of a power-electronics or RF assembly. This asset addresses that obstacle directly through a compositionally specific and dimensionally precise solution: a conformal dielectric shell, ranging from 1 to 50 nanometers in thickness, applied to the surface of diboride and related boride particles before compounding. When these coated particles are incorporated at sub-percolation loadings (5 to 25 volume percent), the resulting composite retains bulk volume resistivity above 10^8 ohm-centimeters — the threshold that enables deployment in electrically isolated package locations — while preserving the mechanical stiffness and thermal transport advantages that make diborides attractive in the first place. The commercial unlock is access to a class of package locations that refractory boride fillers have never previously occupied. This asset sits within the high-power thermal-interface materials portfolio as a dependent arm: it extends the core diboride claims to a well-defined family of coated species combined with a specific electrical-isolation design rule. Its role is both commercially oriented and defensively structured — it carves out a technologically distinct product category (coated, sub-percolation, isolated-location use) from the broader landscape of boride-filler art, and it establishes negative limitations that focus the claim scope on the computationally validated stable members of the diboride family, excluding species that do not meet the dynamic stability requirements.

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.

The diboride crystal system targeted here adopts the hexagonal P6/mmm (AlB2-type) structure, a layered arrangement in which alternating metal and boron honeycomb planes produce strongly anisotropic bonding and excellent in-plane stiffness. TiB2, ZrB2, and HfB2 are the canonical confirmed-stable members of this family. Dynamic stability — the requirement that no imaginary (negative-frequency) phonon modes exist at any point in the Brillouin zone — was evaluated using two independent machine-learning interatomic potentials representing different parameterizations and training sets. For TiB2, one potential yields a minimum phonon frequency of +0.265 THz (positive, meaning dynamically stable); the second potential initially returned a value of -2.9 THz during screening of AlMgB14, flagging it as potentially unstable. That second result was subsequently resolved through additional calculation, with AlMgB14 rehabilitated to a minimum frequency of +0.42 THz — a meaningful finding because it documents a specific stability rescue that would not appear in most screening workflows. Two independent DFT reference calculations corroborate the machine-learning results. TaB2 and NbB2 carry minimum frequencies of +0.626 and +0.814 THz respectively across the confirmed-stable class. The claim scope explicitly excludes MoB2, WB2, CrB2, MnB2, and ScB2, all of which exhibited imaginary phonon modes across all four computational engines used in the broader portfolio screening; that four-way agreement on instability is strong grounds for the exclusion, and its documentation as a negative limitation is a deliberate design choice that narrows prosecution risk. The thermal transport simulations go considerably beyond simple stability checks. Harmonic phonon calculations (referenced as workflow elements addressing diboride sound velocity and Cahill thermal conductivity estimates) establish the phonon group velocities and density-of-states features that determine the intrinsic lattice thermal conductivity floor. These were complemented by anharmonic Boltzmann transport equation (BTE) calculations for the three primary targets: TiB2 at a phonon-branch contribution baseline, ZrB2 at a higher computed value, and HfB2 at an intermediate figure — all evaluated as phonon-only contributions excluding electronic conduction, which is the regime relevant to the coated composite since the dielectric shell is specifically designed to suppress the electronic contribution that would otherwise dominate in uncoated metal-like diborides. The combination of high phonon-channel conductivity and electronic-channel suppression is the physical mechanism that makes the coated system work: the shell does not significantly impede phonon transmission at the length scales targeted (1 to 50 nm), but it does interrupt the electron percolation network at sub-percolation loadings, achieving selective transport. The dielectric shell specification — 1 to 50 nm conformal coverage — is dimensionally important. Shells thinner than 1 nm are unlikely to achieve pinhole-free coverage over irregularly faceted refractory ceramic particles, offering insufficient electrical barrier integrity. Shells thicker than 50 nm begin to add meaningful thermal interfacial resistance per particle and reduce the volumetric packing efficiency of the hard phase, degrading the modulus benefit. The 1 to 50 nm window therefore reflects a real engineering trade-off that the patent family is designed to own. The volume fraction range of 5 to 25 percent is calibrated to sit below the percolation threshold for spheroidal conductive particles in a polymer matrix (typically around 15 to 30 percent for irregular refractory particles, depending on aspect ratio and surface treatment), ensuring that even if shell coverage is imperfect, the overall composite remains in the non-percolating regime. The confirmed-stable class also includes the extended boride B4C and the layered hexagonal boron nitride (h-BN), species that share structural elements with the diboride family and whose stability across the computational screening is equally well-established. AlMgB14 is notable as a non-stoichiometric ternary boride with ultrahard character; its inclusion in the confirmed-stable class after the stability rescue calculation broadens the claim coverage to include a phase with distinct hardness and wear-resistance properties relative to the binary diborides. Comparative Example G in the experimental record documents an uncoated percolating diboride composite and serves as the negative control against which the volume resistivity improvement of greater than 10^8 ohm-centimeters is benchmarked — a concrete experimental anchor for the claim.

The addressable market for this asset sits within advanced thermal-interface and packaging materials used in power electronics and RF/radar modules. Power-electronics packaging is currently undergoing a forced transition: wide-bandgap semiconductor platforms (silicon carbide and gallium nitride) operate at higher junction temperatures, higher switching frequencies, and higher power densities than legacy silicon, and they impose correspondingly more severe demands on the package substrate, underfill, and inter-die materials. The thermal interface material market broadly is estimated at several billion dollars annually and growing, but the specific sub-segment addressable by a high-stiffness, electrically isolated filler — materials that go into non-active regions of high-power packages where both thermal load-spreading and electrical isolation are required simultaneously — is more concentrated. The estimated total addressable market for this asset is in the range of $0.5 to $1 billion, reflecting the premium that power-electronics and RF module makers pay for materials that solve both requirements without compromise. The buying logic for this type of material is typically a combination of qualifying-supplier contracts and long-term materials qualification runs with tier-one package manufacturers and module assemblers. The royalty or licensing leverage is strongest at the formulation level: a compound supplier or materials company that licenses this composition and process owns a defensible product differentiation that cannot be replicated simply by switching to an uncoated competitor's filler. Power-electronics packaging customers — automotive inverter makers, industrial drive manufacturers, RF/radar systems integrators — tend to have long qualification cycles (12 to 36 months for a new interface material) and high switching costs once qualified, which supports durable licensing economics. The RF/radar segment is a secondary but meaningful customer class, particularly for applications where filler materials must survive both high-power thermal loads and radio-frequency signal environments without introducing dielectric losses or electrical leakage. The commercial advantage relative to existing practice is structural, not incremental. Uncoated boride fillers at useful thermal loadings create percolation-mediated leakage that disqualifies them from electrically sensitive locations. The coating-plus-loading design rule solves that problem without switching to a lower-conductivity filler species, preserving the stiffness and thermal performance that makes diborides attractive in the first place. No direct analog to this product formulation is currently offered by the leading filler suppliers, which means the first mover who qualifies this approach in a commercial compound has a window to establish market position before the category is crowded.

high-stiffness hard-phase filler with electrical isolation for power-electronics-adjacent applications

dielectric shell 1-50 nm + below-electrical-percolation loading + electrically isolated package placement (vs Chomerics TiB2-in-elastomer broad blocker)

The most natural acquirers or licensees for this asset are advanced materials companies already supplying the power-electronics packaging supply chain: specialty compound makers, filler producers with ceramic surface-treatment capabilities, or vertically integrated thermal-interface material suppliers who want to move up in technical sophistication. Parker Hannifin (Chomerics), Momentive, Henkel, and Laird (part of Comfort Systems) are the tier-one thermal-interface material suppliers with the manufacturing infrastructure to commercialize a coated-boride compound and the customer relationships to qualify it with power module makers. A second category of buyer is the substrate or package manufacturer who wants proprietary materials to support a next-generation SiC or GaN module platform — companies like Kyocera, Denka, or Rogers Corporation that supply package substrates and have incentive to differentiate their platform with a higher-performance isolation-compatible filler material. In the RF/radar segment, defense prime contractors and their electronic warfare and radar subsystem suppliers represent a less price-sensitive buyer class with significant motivation to solve the dual thermal-and-electrical-isolation problem in compact high-power modules. In each case, the licensing argument is the same: the dielectric-shell-plus-sub-percolation design rule is a specific, defensible technical combination that is not currently offered by any commercial supplier, and the computational and experimental evidence base de-risks the qualification pathway relative to starting from scratch.

The principal technical risk is the open durability gate: the volume-resistivity cycling test remains unfinished, and until that data confirms that the dielectric shell survives thermomechanical cycling, the electrical isolation performance claim is unverified at the durability level. Shell integrity on hard refractory ceramic particles is a known manufacturing challenge — achieving pinhole-free, adhesion-stable conformal coatings at 1 to 50 nm thickness on faceted TiB2 or ZrB2 particles requires process control that has not yet been demonstrated in the experimental record for this family. The commercial risk is that this asset is a dependent arm, not a standalone invention: its value is linked to the broader high-power thermal-interface materials portfolio, and a licensee or acquirer would typically need to take the core family alongside this arm to obtain meaningful protection. The FTO position is narrow, meaning claim scope is constrained relative to the prior art, and a well-resourced competitor with Chomerics-adjacent IP could potentially challenge or design around specific claim elements. De-risking the technical path requires completing the cycling coupon protocol (before-and-after resistivity measurements per the Family H protocol), which is a well-defined and bounded experimental program. Process development for the dielectric shell — atomic layer deposition or sol-gel coating on diboride particles at scale — is a known industrial capability available through contract manufacturers, which limits the scale-up risk. On the IP side, the FTO risk is managed by maintaining the triple-conjunction design requirement in all claims and ensuring that prosecution history does not inadvertently broaden the footprint into the Chomerics space. A full professional FTO opinion before commercial launch is the appropriate next step for any party considering acquisition or licensing.