Bakeout Physics · Part 3 of 3

Module 04: Vent Physics

The vent: what opening a baked UHV system puts back on the wall

The system was assembled, pumped, and baked 48 h at 150 °C. It sits at UHV. Now someone opens it to air for a few minutes. This page shows, molecule by molecule, what that vent gives back to the wall, what it can never touch, and why the choice of vent gas matters more than the stopwatch.

Vent gas

Exposure time log scale, 1 second to 12 days

Wall temperature during vent venting at 60–80 °C limits multilayer condensation, see below

THE PHYSICS CARRIED OVER FROM PART 2 · same two laws, now applied to our 150 °C fleet

Frenkel · desorption

τ = τ₀·e^E/kT

At 150 °C, bound water (0.9–1.05 eV, this card uses 1.05) sits ~0.3 s per site: that is why the 48 h bake empties the water reservoir. The hydrocarbon (1.6 eV) still sits ~13 days: out of reach.

Fick · H diffusion in the bulk

D = D₀·e^−E/kT, L = √(Dt)

D(150 °C) ≈ 3.4×10⁻⁹ cm²/s, so 48 h drains a ~240 µm hydrogen skin. Fully draining the bulk would need 400 °C and up, which our fleet cannot reach: see the grey zone on the Part 2 Arrhenius map.

New for the vent · adsorption

t_ML ≈ 3×10⁻⁶/P, dose = P·t

The same wall physics run in reverse: impingement fills the sites the bake emptied. Everything below follows from these two expressions.

STEP 1The wall during the vent, live

Same stratified wall as before, but now the story runs backwards: the gas rains onto the surface instead of leaving it. At 150 °C, the hydroxyl groups (OH chemically grafted onto the Cr₂O₃ oxide) never left. Removing them takes more than ~350 °C. So everything your bake actually removed sits above them, and everything the vent restores lands right back in the same place. Drag the exposure slider or hit Replay.

CROSS-SECTION DURING THE VENT · exposure time set by the slider

OH hydroxyls, grafted on the oxide: never left at 150 °C strongly bound water (~1.05 eV): removed by the bake, comes back with dose physisorbed multilayers: only if P/P_sat is high (humid air) capillary / subsurface inventory: the slow, deep reservoir H in the bulk, with its bake-depleted skin

The wall is a sponge you just wrung out. Open it under a running tap (humid air) and it is soaked again before you can blink, then keeps drinking for hours into its deep pores. Open it in a dry room (N₂) and almost nothing happens: a damp film at most. The water can never reach the hydrogen inside the rubber of the sponge.

STEP 2Two clocks: monolayer time and water inventory

Your engineer is racing the wrong clock. The first clock, monolayer formation, is lost before the reflex arc of his hand finishes opening the valve. The second clock, water inventory, is the one a human can act on, and it scales with the dose, P_H2O × t.

CLOCK 1 · THE MONOLAYER · unwinnable

t_ML ≈ 3 × 10⁻⁶ / P Torr·s, sticking ≈ 1

Humid air, P_H2O = 8.8 Torr

0.4 µs

The first monolayer is complete about a million times faster than the valve opens.

Dry N₂, P_H2O = 2×10⁻³ Torr

sub-monolayer forever

Impingement would allow a monolayer in ~2 ms, but at P/P_sat ≈ 10⁻⁴ the equilibrium coverage stays a fraction of one layer. It never builds.

CLOCK 2 · THE INVENTORY · winnable, by dose

dose = P_H2O × t this is the number that sets the next bake

Your vent right now

Same duration, other gas

Halving the open time buys a factor 2. Switching air to dry N₂ buys a factor ~4000 on P_H2O.

STEP 3The dose chart: water taken back, versus exposure

Re-adsorbed water inventory against exposure time, for both vent gases at the wall temperature you chose. The dashed line is roughly what your 48 h / 150 °C bake removed: when a curve reaches it, the water part of your bake is fully undone. Hover to read, and click to set the slider.

Hover the chart to compare both gases at any exposure time.

Illustrative model, in monolayer-equivalents: fast strongly-bound refill, BET-type multilayers gated by P/P_sat, and a slow logarithmic capillary/subsurface term. Absolute values are indicative, and the scaling with dose and humidity follows from the model.

STEP 4What survives, what is lost, what it costs

The bake removes two inventories. The hydrogen depleted from the metal skin comes through the vent intact. The water removed from the surface comes back with the vent, in proportion to the dose.

THE H₂ WORK · SURVIVES THE VENT

48 h at 150 °C drains hydrogen from a skin of depth L = √(Dt) ≈ 240 µm (D ≈ 3.4×10⁻⁹ cm²/s). At room temperature, hydrogen from the vent gas does not re-dissolve into stainless in minutes or hours. The depleted skin, and the lower H₂ outgassing it buys you, comes through the vent intact.

THE WATER WORK · SPENT BY THE VENT

Water inventory restored by your scenario

With a 150 °C ceiling, a hotter bake is not available, so bake duration is the only recovery variable, and the vent dose has to be kept low.

THREE VENT SCENARIOS · computed from the same model

Vent scenario	Dose [Torr·s]	Water back [ML eq.]	Field consequence

Field protocol to minimise the water dose. Vent while the system is still warm (60–80 °C: the wall’s local P/P_sat collapses, multilayers cannot condense). Vent with dry N₂ at a slight overpressure so ambient air never flows in. Keep the opening as short as the job allows. Pump down immediately. Do all four and the rebake often disappears entirely, and the pump-down just runs a little longer.

Sources · 3

[A] P.A. Redhead, Recommended practices for measuring and reporting outgassing data, JVST A 20, 1667 (2002).
[A] P. Chiggiato, Outgassing properties of vacuum materials for particle accelerators, CERN Accelerator School : Vacuum for Particle Accelerators, Glumslov 2017, arXiv:2006.07124.
[A] M.R. Louthan & R.G. Derrick, Hydrogen transport in austenitic stainless steel, Corrosion Science 15, 565 (1975).

Why We Bake Learning Hub Pumps & Conductance