R1: the annealing baseline

Sweep executed 18 August 2026. The finding: the annealing mechanism is real and defect-resolved, and ultra-thin defect-engineered silicon cells recover fully at 65–100 °C under light — inside the realistic sunlit LEO operating band. But the effect is strongly thickness-gated (full recovery at 20 µm, near-zero at 80 µm) and doping-gated (p-type recovers, n-type does not). At the program's 100 µm n-type design point, the annealing claim is currently undemonstrated on both axes.

Why this baseline exists

R1 bets that thin silicon can be radiation-hard through self-annealing rather than shielding. The directions page stated the mechanism exists in the literature but is not yet creditable at the 100 µm design point; R1 wins when end-of-life degradation becomes predictable and boring — an actuarial input, not a research question. This page records what the literature actually supports: the defect physics, the thickness scaling, the operating-temperature arithmetic, and the named gaps between the published evidence and a creditable model.

Defect inventory and annealing stages

established Radiation damage in silicon is displacement damage: the primary event is the Frenkel pair (displacement threshold ≈ 11–13 eV), and the surviving stable defects are impurity complexes. Carrier-lifetime damage follows1/L² = 1/L₀² + K_L·Φ with K_L ≈ 1.9×10⁻¹⁰ per 1 MeV electron for n/p 1 Ω·cm cells (Rosenzweig 1963). Each defect has its own annealing temperature and activation energy:

Device level: conventional Si cells show no significant annealing below ~200 °C; significant recovery needs 200–400 °C (JPL Radiation Handbook). Low-oxygen/low-carbon cells recover 96% of pre-irradiation I_sc at 200 °C and fully at 275–300 °C (second-order, Ea 1.5±0.2 eV — Weinberg & Swartz 1981). Ambient annealing exists but is slow and partial: ~10% of I_sc over days-to-a-month for electrons, ~20% after 22 months for protons; Telstar-era cells recovered 15–20% of diffusion length in two weeks at room temperature. Reverse annealing — continued degradation during post-irradiation heating — is documented above 400 °C at fluences >10¹⁴ e/cm². Injection-enhanced annealing (recombination energy driving defect reactions) is established mechanism-level physics in Si: Lang & Kimerling 1974, Arimura 1969, Makarenko 2018.

Thickness scaling: the short-base argument, quantified

established When the base is thinner than the radiation-degraded diffusion length, lifetime damage stops limiting collection. This is not a heuristic — it is quantified and flight-proven:

The modern annealing result — and its gates

established-with-gap The strongest recent result in the program's direction: ultra-thin silicon heterojunction cells (Solestial/ASU, independently validated by CEA) recoverfully from 1 MeV electron and 300 keV proton damage at 65–100 °C under illumination. But the same dataset shows the two gates:

The design-point problem, stated plainly: the published recovery data brackets the program's design point from both sides — full recovery at 20 µm, almost none at 80 µm — and the doping axis points away from n-type: Cz p(B) recovers ~90% at 100 °C, Cz n(P) does not recover at all. The "100 µm n-type" assumption inherited by the directions page is challenged by this evidence on both axes.

Operating-temperature realism

evidence-backed Measured LEO/GEO array temperatures span −150 to +110 °C (Vayner 2016); ISS-class silicon arrays run ~+55 °C typically, ~28 °C at equinox insolation; the qualification envelope is ~30,000 cycles between −100 and +100 °C. The sunlit operating point — where annealing under illumination could actually run — is roughly +50 to +80 °C. Against that: conventional Si needs ≥200 °C (never seen in orbit), while the thin-Si result works at 65–100 °C — the upper end of the realistic sunlit band, and reachable by thermal design. The arithmetic closes, but only for the thin, p-type, defect-engineered cell, and only at the warm end of the orbit.

The gap, decomposed

• Thickness-resolved annealing at the design point — Published recovery data: 20 µm full, 80 µm near-zero. The 80–100 µm band — where the program lives — has no published annealing rates. This is one irradiation-and-anneal campaign on existing cells.
• Mechanism attribution — Injection-enhanced bulk annealing vs surface proximity / gettering / out-diffusion is unresolved; the strongest practitioner (Solestial) treats the defect engineering as proprietary. The mechanism decides whether the effect transfers to other cell architectures.
• In-orbit flight telemetry — No peer-reviewed flight dataset shows Si cell recovery in orbit (APEX recorded degradation; the documented in-orbit annealing case is GaAs, Hughes HS 393). Company on-orbit claims exist; published data does not.
• Combined-protocol annealing — No campaign anneals under radiation + illumination + thermal cycling at a realistic LEO profile. Single-stressor annealing data may not predict recovery under the coupled environment.

What this changes

Three program consequences. First, the R1 design point should be revisited: the evidence favors Ga-doped p-type SHJ at the thinnest manufacturable thickness, not 100 µm n-type — the directions page should carry that as an open design revision, not a settled assumption. Second, the single decisive experiment is cheap and ground-based: irradiate-and-anneal at 80–100 µm to locate the annealing cliff between 20 and 80 µm — if the cliff is sharp and below 100 µm, the R1 annealing thesis as scoped is wrong and the design moves down-thickness. Third, watch items: Solestial/York on-orbit data (flown Transporter-10, "nominal", unpublished), and any peer-reviewed flight telemetry showing in-orbit Si recovery — none exists today.

Perimeter and provenance

Three parallel literature sweeps, 25 sources, admitted to the Surya epistemic graph as event ev-000000000008 with assessmentsurya-assessment-r1-annealing-2026-08, judgmentestablished-with-gap. Sweep-level gaps, flagged honestly: several AIP/Wiley/HAL full texts were bot-walled, so some numbers rest on abstracts and citing works rather than primary text; the Medjoubi 2021 damage-coefficient table and Guillot 2024 thickness-resolved numbers could not be verified firsthand; the classic defect annealing temperatures rest on 1960s EPR/IR work read through later kinetics papers; the Solestial mechanism is partly proprietary; and no peer-reviewed in-orbit Si annealing telemetry was located at all. None of these change the shape of the baseline — they are recorded as perimeter, not smoothed over.

Sources

• Tada et al., Solar Cell Radiation Handbook, 3rd ed., JPL (1982) — The canonical damage-law and annealing-temperature compilation
• Rosenzweig, Gummel, Smits, BSTJ 42 (1963) — Original diffusion-length damage law; K_L ≈ 1.9×10⁻¹⁰ per 1 MeV electron
• Watkins & Corbett, Phys. Rev. 138 (1965) — EPR of the divacancy; migration Ea 1.3 eV
• Mooney et al., Phys. Rev. B 15 (1977) — Defect census for B-doped Si under 1 MeV electrons
• Mikelsen et al., Phys. Rev. B 72 (2005) — Divacancy annealing kinetics: first-order, Ea 1.30±0.02 eV
• Londos et al., J. Appl. Phys. 84 (1998) — A-center annealing behavior
• Weinberg & Swartz, NASA TM-82597 (1981) — Reduced annealing temperatures: 96% Isc recovery at 200 °C in low-O/C cells
• Weinberg & Swartz, Appl. Phys. Lett. 36 (1980) — Origin of reverse annealing in radiation-damaged Si cells
• Srour, Marshall & Marshall, IEEE TNS 50 (2003) — Displacement-damage (NIEL/DDD) framework review
• Lang & Kimerling, Phys. Rev. Lett. 33 (1974) — Recombination-enhanced defect reactions — the injection-annealing mechanism
• Arimura et al., J. Appl. Phys. 40 (1969) — Injection-enhanced divacancy annealing in n-type Si
• Makarenko et al., J. Appl. Phys. (2018) — Electron injection alters defect reaction rates in irradiated Si
• DiNetta, Rand, Barnett et al., NASA CP (1991) — <25 µm cells tolerate <250 ns lifetimes; InP-class radiation resistance projected
• Landis, NASA CP (1991) — Optimum EOL thickness may be as low as 2 µm
• Cariou et al., IEEE J. Photovoltaics 14 (2023) — 60/110/160 µm Ga-doped SHJ vs PERC under 1 MeV electrons
• Matsuda et al., NASDA, NASA CP (1994) — 50 µm cells flown on ERS-1/ETS-VI/ADEOS; ~60,000 100 µm CICs qualified
• Herasimenka et al., 40th Space Power Workshop (2023) — Full recovery at 80 °C under light; thickness- and doping-gated
• Gurimskaya et al., ESPC (2023) — Peer-reviewed conference version: effective annealing in ultra-thin Si
• Aponte et al., 41st Space Power Workshop (2024) — Annealing at 65 °C under maximum-power-point operation
• Solestial / CEA (2023) — Independent CEA validation: recovery under 1 sun at 90 °C after a 10-year-LEO-equivalent dose
• Wang et al., Science Advances (2025) — Perovskite/p-Si tandem on 100 µm p-Si HJT bottom cell; radiation-resilient
• Khan, Yamaguchi et al., Sol. Energy Mater. Sol. Cells 75 (2003) — Ga-doped Si more radiation tolerant than B-doped
• Vayner, NASA Glenn (2016) — Measured array temperatures: −150 to +110 °C
• Gruenbaum, APEX/PASP-Plus, NASA/JPL (1995) — Si degradation recorded; in-orbit annealing documented for GaAs, not Si
• Brown & Gabbe et al., BSTJ 42 (1963) — Telstar: 15–20% diffusion-length recovery after two weeks at room temperature