The Svalbard Global Seed Vault stores roughly 1.3 million seed samples at -18°C, deep inside a Norwegian mountain. It works. It also leaks. In 2017 meltwater from thawing permafrost reached the entrance tunnel. Svalbard at -18°C is what current best-practice cryopreservation on Earth looks like. The Lunar Ark targets something different: -196°C, the boiling point of liquid nitrogen, sustained for ten centuries, using a refrigerator the size of a planet.
Why -196°C
Below the glass-transition temperature of water (around -130°C), biochemical reactions effectively halt. DNA does not unwind. Proteins do not denature. Cellular structures, vitrified into an amorphous glass-like solid, are mechanically stable indefinitely. -196°C is conservative — well below the threshold, immersed in a fluid (LN2) that provides passive thermal buffering. It is the temperature at which seed banks, biobanks, and reproductive biology labs already store viable samples for human-civilizational timeframes today. Scaled to the Ark's vault, it preserves not just seeds but tissue samples, gametes, and embryos for the eventual reconstruction of complete species.
The trick: passive cold from a permanently shadowed crater
On Earth, holding -196°C requires either continuously replenished LN2 (Svalbard, every biobank in the world) or active cryocooler systems running on grid power (laboratories, hospitals). Both have failure modes that scale poorly to a thousand-year mission. The Lunar Ark's site selection — Shackleton Crater rim at the south pole — exists precisely to break this dependency. Permanently shadowed crater floors near Shackleton sit at 25-50 K (-223 to -248 °C) year-round, for at least the last 2 billion years per Diviner radiometer data. The Codex's L1-CRY system is designed around a passive cooling architecture (L3-CRY-COOL-PT, L3-CRY-COOL-STIR) that radiates waste heat into the PSR cold sink — no power burned on refrigeration, no mechanical pumps in the critical loop.
"Vitrified mammalian embryos stored at LN2 temperature have demonstrated viability after 30+ years with no measurable degradation. The theoretical upper bound is geological — limited by radiation-induced damage, not chemistry."
— Mazur et al., Cryobiology (2014); ongoing follow-up at Embryo Donation International.
Vitrification: skipping the ice
The enemy of long-term cryopreservation is not cold — it is the phase transition to crystalline ice. Ice crystals, expanding 9% in volume relative to liquid water, lacerate cell membranes. Vitrification protocols (L3-GEN-EMBR-VIT, L3-GEN-DNA-ENCAP) use cryoprotectant chemistry to drive the freezing process directly to a glassy amorphous solid, bypassing crystallization entirely. The 2025 Texas A&M work on raising the glass-transition temperature for whole-organ vitrification — visible in the Codex's news feed — extends this protocol from gametes and embryos to organs, which would historically crack during cooling. The Ark inherits the chemistry.
Inventory, retrieval, integrity
Storing the samples is only half the system. The L1-CRY architecture also specifies the inventory layer (L2-CRY-INV, L3-CRY-INV-TRACK): RFID-tagged samples, robotic retrieval arms cleared for cryogenic operation, a database that survives the broader collapse of any external network. Without retrievability, the vault is a tomb. The Codex treats inventory and retrieval as preservation-critical, not as logistics.
The 1000-year horizon is not a stunt. The proteins encoded in a stored embryo would still be biochemically active if rewarmed. The DNA in a stored seed would still germinate. The Ark's ambition is to make that latency — between deposit and revival — as long as the geological window the Moon offers us, which is essentially as long as we need it to be.