ALD process for depositing bound-chlorine aluminum oxychloride passivation films

The glass-core advanced-packaging substrates portfolio addresses a quiet but consequential materials problem in semiconductor packaging: how to protect the fine redistribution-layer (RDL) dielectrics and copper routing structures in next-generation glass-core substrates from moisture ingress, ion migration, and chemical attack. Conventional atomic layer deposition of alumina (Al2O3) has been the industry's default passivation approach for decades, but it leaves no residual halogen in the film. This process invention carves out a structurally distinct class of amorphous aluminum oxychloride (AlOxCly) films deposited by a cyclic ALD sequence in which a chlorine-bearing reactant is deliberately retained in the film at 2–25 atomic percent after a defined thermal anneal. That retained bound-chlorine window — verified by X-ray photoelectron spectroscopy or time-of-flight secondary-ion mass spectrometry — is what separates this process from every prior-art alumina ALD route. The commercial logic is straightforward. As the packaging industry shifts from organic laminate to glass-core substrates to achieve tighter pitch, lower loss, and better dimensional stability, the interface passivation requirements tighten correspondingly. Glass surfaces are chemically reactive, and thin RDL dielectrics are intolerant of the pinholes and moisture paths that alumina ALD alone can leave at certain anneal conditions. An AlOxCly film with tunable bound-chlorine concentration offers an additional lever — chlorine-terminated surface sites that can modify the dielectric interface chemistry, suppress copper hillock growth, or serve as an etch-stop differential — giving fab engineers a process knob that simply does not exist in conventional alumina practice. The invention is a process claim, which means that any fab that deposits an amorphous Al-O-Cl film with retained bound chlorine by a cyclic ALD route lands inside the claim space regardless of the specific precursor pairing chosen, provided the composition window and measurement conditions are met. The timing aligns with a meaningful forced-substitution cycle. Glass-core substrate capacity is being qualified at multiple OSAT and IDM fabs through the late 2020s, and passivation process choices made during qualification become locked into the process of record for years. A process patent filed and granted before widespread glass-core adoption gives the holder a licensing position against every fab that qualifies an AlOxCly ALD step — a structural advantage that compounds as glass-core volumes scale.

The deposited material is amorphous aluminum oxychloride, compositional shorthand AlOxCly, where the chlorine is not a surface contaminant or a transient etch residue but a deliberately retained, chemically bound constituent of the bulk film, present at 2–25 atomic percent after a defined post-deposition thermal anneal. The amorphous nature of the film is important: without long-range crystalline order, there are no grain boundaries to act as moisture-diffusion highways, which is one of the known failure mechanisms for polycrystalline Al2O3 passivation layers on glass and dielectric surfaces. The chlorine incorporation modifies the local bonding environment around aluminum — XPS spectra of correctly processed films show Al-Cl bonding contributions distinct from the Al-O contributions that dominate conventional alumina — and that local bonding change is what differentiates the film's surface chemistry from standard ALD alumina. The preferred precursor pairing identified by computational screening is trimethylaluminum (TMA) combined with hydrogen chloride (HCl) as the chlorine reactant, an optionally supplemented oxidant step included in the cyclic sequence. TMA is already the dominant aluminum precursor in industrial ALD, which matters for integration: the process does not require introduction of an exotic organometallic that lacks safety data or supply chains. The computational precursor ranking was performed using the extended tight-binding method (xTB), a semi-empirical quantum-chemistry approach well-suited to rapid screening of reaction enthalpies and leaving-group stability for ALD surface reactions. In that screen, TMA+HCl emerged as the preferred combination (ranked first among the candidates evaluated), with methylaluminum dichloride (MADC), dimethylaluminum chloride (DMACl), and aluminum trichloride (AlCl3) identified as alternative precursor routes that also fall within the claim space. The xTB screen evaluated chlorine-vacancy formation energetics — essentially asking how strongly the chlorine is bound in the deposited film and under what thermal conditions it desorbs — which directly informs the anneal-condition window needed to retain 2–25 at% Cl rather than driving chlorine out entirely. Because AlOxCly is an amorphous process-defined material rather than a bulk crystalline phase, the conventional computational-stability validation workflow (multi-potential phonon consensus, DFT geometry optimization of a crystal structure) does not apply directly: there is no periodic unit cell to relax. The computational work therefore focused on the chemistry of the deposition and retention mechanism rather than bulk structural stability. The xTB precursor ranking assessed the energetics of chlorine retention versus desorption at the film surface during and after ALD cycling. A separate chlorine-vacancy energetics calculation examined the binding energy of Cl in representative AlOxCly local environments, providing a theoretical basis for understanding why chlorine remains in the film at the concentrations claimed rather than volatilizing during anneal as it does in many conventional ALD sequences using chloride precursors. These are the technically load-bearing computational inputs: they rationalize the process window and support the XPS/ToF-SIMS measurement claims with a physical mechanism. The key measurable process outcome — retained bound-chlorine concentration in the 2–25 at% window after a defined anneal, confirmed by XPS or ToF-SIMS — is the claim-defining metric. ToF-SIMS is particularly well-suited here because it provides depth-resolved chlorine profiles, enabling a distinction between surface-adsorbed Cl (which would be dismissed as contamination) and bulk-retained Cl distributed through the film thickness. XPS provides bonding-state information, distinguishing Al-Cl from Al-O contributions and confirming that the chlorine is chemically integrated rather than physisorbed. Both characterization paths are explicitly recited in the claim, giving the process a well-defined analytical verification protocol that is routinely available in semiconductor analytical labs.

The addressable market for ALD passivation process licensing sits within the broader advanced packaging substrate and RDL fab segment. Glass-core substrate manufacturing is an emerging high-growth vertical; analyst estimates place the glass-core substrate market at several hundred million dollars in the near term, scaling toward a multi-billion dollar segment by the early 2030s as hyperscaler and AI accelerator packaging demands drive adoption. The total addressable market for ALD process licensing into this segment is estimated at $1–3 billion, which is consistent with the scale of ALD equipment and process spending across the RDL passivation supply chain, though these are estimates and actual licensing yield depends on adoption rates and royalty terms negotiated. The buyer profile for a license or acquisition of this process patent is any organization operating ALD equipment in an advanced packaging or RDL fab context — OSATs qualifying glass-core lines, IDM packaging divisions, or substrate manufacturers building out ALD capability for passivation steps. The process nature of the claim means the licensing hook is at the fab level: any fab running a cyclic ALD sequence that produces an amorphous Al-O-Cl film with 2–25 at% retained bound chlorine after anneal, verified by XPS or ToF-SIMS, is practicing the process regardless of whether they use TMA+HCl or one of the alternative precursor routes. That broad precursor coverage maximizes the claim's reach across the real-world diversity of fab process choices. Royalty logic for a process patent in semiconductor manufacturing typically attaches to wafer starts or substrate area processed under the claimed method. Given the high per-unit value of glass-core substrates and the relatively thin royalty rates conventional in the ALD licensing space (fractions of a percent of substrate value), even modest penetration of the glass-core qualification pipeline represents meaningful royalty income. Alternatively, an outright acquisition by an ALD equipment company seeking to bundle process IP with tool sales, or by a substrate manufacturer seeking to defensively control passivation process freedom, would price this on the basis of blocking value and the cost of designing around — which, for a process claim covering the full retained-Cl composition window, is high.

process protection of the bound-Cl film

deliberate bound-Cl retention distinguishes conventional alumina ALD

The natural acquirers and licensees for this process patent fall into two categories. The first is ALD equipment and process companies — Applied Materials, Lam Research, ASM International, and their peers — that are actively developing and qualifying ALD process recipes for advanced packaging customers. These companies have strong incentives to hold process IP that differentiates their recipe portfolios, and a process patent covering AlOxCly ALD with retained bound-Cl fits directly into their advanced packaging product lines. An ALD company that acquires this IP can bundle the process recipe with tool sales or license it to fabs as part of an applications development agreement, creating a recurring revenue stream on top of equipment sales. The second category is glass-core substrate manufacturers and advanced packaging OSATs — Ibiden, Shinko, NGK, and major OSAT players qualifying glass-core lines — who have a defensive interest in securing process freedom for the passivation steps in their glass-core qualification flows. If an AlOxCly ALD process proves to be the preferred passivation solution for glass RDL interfaces (a plausible outcome given the surface chemistry advantages), then owning or licensing this process patent before glass-core volumes scale is strategically valuable to avoid royalty exposure at scale. Chemical precursor suppliers (Entegris, Merck KGaA, SK Materials) that supply HCl or aluminum-chloride precursors into ALD process flows represent a third category with licensing interest, as process IP that specifies a preferred precursor pairing can underpin supply agreements with higher switching costs.

The primary technical risk is that the XPS coupon measurement has not yet been performed. The process is computationally motivated and chemically plausible — TMA+HCl ALD is a known chemistry, and residual chlorine in ALD films deposited from chloride precursors is an observed phenomenon in the literature — but the specific claim that bound-Cl can be retained at 2–25 at% after the claimed anneal conditions using the preferred TMA+HCl route needs experimental confirmation. If the anneal drives chlorine out below 2 at% under conditions relevant to actual packaging process flows, the composition window would need to be revised and the anneal conditions tightened, potentially narrowing the claim. The path to de-risking this is a straightforward coupon experiment: deposit AlOxCly by TMA+HCl ALD, anneal at defined conditions, and characterize by XPS — a weeks-long experiment at any equipped fab or university surface-science lab. The commercial risk is adoption timing. Glass-core substrate qualification is progressing but remains in early stages at most fabs, and the process-of-record choices for passivation are not yet locked. If fabs converge on silicon nitride or silicon oxide CVD rather than ALD for glass-core passivation before this process establishes a licensing position, the addressable royalty base shrinks. That said, the process claim is not limited to glass-core substrates specifically; any RDL passivation or dielectric interface application that uses cyclic ALD to deposit an Al-O-Cl film with retained bound Cl is within the claim, which broadens the licensing surface beyond the glass-core vertical alone. The defensive value of the clean freedom-to-operate position means that even as a portfolio asset rather than a standalone flagship, this process claim provides meaningful protection for a Lattice Graph partner or acquirer operating in the RDL passivation space.