# Al₂O₃-on-Si thin-film benchmark handoff
**Validated public benchmark; bounded hybrid across 5–10 nm**

**Audience:** Mixed technical  
**Depth:** Explanation-first  
**Terminology:** Standard field terminology  
**Derived from:** CGRP packet `INV-2026-03-27-HT-T1`, `P16`

**Boundary + horizon:** Planar cross-plane solid multilayer thermoreflectance benchmark: **Al transducer | ALD Al₂O₃ | native-oxide-present (001) Si**. Active horizon is the benchmark-only comparison window for the **4.67 nm** and **9.39 nm** Al₂O₃ public cases.

**[STATE: Equilibrium Band]**

## Current bounded result
- **4.67 nm case:** interface-leaning
- **9.39 nm case:** near parity
- **Overall 5–10 nm read:** bounded hybrid
- **Not established here:** raw TDTR replay or per-interface TBR split

## Problem
This handoff packages the **public-benchmark path** for a bounded non-equilibrium heat-transfer question: within the present thin-film benchmark boundary, which route best explains the observed departure from classical continuum expectations? The surviving benchmark-level comparison is among interfacial thermal resistance, microscale transport nonlocality, and a bounded hybrid, with the final packet supporting a bounded-hybrid ranking across the 5–10 nm range.

## Boundary and evidence basis
Inside the final boundary are finite Al₂O₃ film resistance, the **combined** interface contribution across the ultrathin stack, the thickness change from **4.67 nm** to **9.39 nm**, and the validated benchmark comparison artifacts. Outside the boundary are internal flow, intended phase change, raw TDTR delay-time replay, justified per-interface decomposition, and any user-local specimen claim. The packet uses direct-source fields plus transparent derivations, with unknowns held as **NA**.

Core retained values:
- **4.67 nm** and **9.39 nm** XRR thicknesses
- **1.50 ± 0.09 W/m/K** intrinsic Al₂O₃ conductivity
- **6.92 ± 0.50 m²·K/GW** combined TBR

## Current bounded result
The validated comparison sheet supports a **bounded hybrid** across the 5–10 nm Al₂O₃ benchmark range overall. The thinner point is **interface-leaning**, while the thicker point shifts toward film/interface parity. This is sufficient for benchmark-level route ranking, but not for claims of raw-trace replay or separate interface resistances.

## What would reopen or invalidate it
Use the validated benchmark comparison sheet as the current decision surface. Escalate beyond this handoff only if the next inference requires raw TDTR delay-time traces, modulation-frequency replay, justified per-interface decomposition, or transfer to a different specimen.

**Constraints used:** standards-first; no new physics; no invented anchor; same Tank continuity; direct-source-only fill; combined TBR stays combined; unknowns remain NA; one-variable reassessment on thickness; benchmark-only scope explicit; workbook/CSV agreement required.

**Limits / uncertainties:** benchmark-only export; no raw TDTR delay-time traces recovered; no justified per-interface TBR decomposition; native-oxide thickness unresolved; no local-run dynamic certification.

**Kill-test / next test:** Use the validated benchmark comparison sheet as the current decision surface. The present route ranking fails if the next required inference depends on raw TDTR delay-time traces, modulation-frequency replay, or justified per-interface decomposition that the public packet does not supply.

**Derived from CGRP packet; no new investigatory claims added.**