The Vacuum Stack

Block 1 — Regimes & Pumps

Vacuum is not one thing. It runs across more than fifteen orders of magnitude, and the physics changes as you go down. This block sets the map, then walks the pumps that get you there.

Source rating, used throughout:

  • [A] primary: standard (ISO, AVS), peer-reviewed paper, lab or CERN/USPAS.
  • [B] manufacturer: technically sound, commercial interest. Cross-check.
  • [C] secondary: trade page or recopy. Flagged. Never the only source.

Pressure in Torr (1 Torr is about 1.33 mbar).

1. The master abacus — vacuum regimes

Regime Range (Torr) Mean free path Flow Pumps Gauges Seals
Rough / primary 760 to 1e-3 < ~0.1 mm viscous to transition rotary vane, scroll, diaphragm, Roots Pirani, capacitance elastomer (Viton), KF/ISO
Medium 1e-3 to 1e-5 mm to cm transition to molecular Roots+backing, turbo (low stage) Pirani, Penning (low) elastomer or metal
High (HV) 1e-5 to 1e-8 cm to m molecular turbo, diffusion, cryo Penning, Bayard-Alpert metal (CF copper) preferred
Ultra-high (UHV) 1e-8 to 1e-11 m to km molecular ion, Ti sublimation, NEG, metal turbo Bayard-Alpert, extractor metal (CF), bakeable
Extreme (XHV) < 1e-11 > km molecular ion + NEG + sublimation, cryo extractor, special CF, hard bakeout, selected materials

How to read it: below about 1e-3 Torr you enter the molecular regime. The mean free path grows past the size of the chamber. Molecules stop colliding with each other and fly straight from wall to wall. Everything changes there. Pumping becomes a matter of surfaces, capture, and how much gas the walls release.

2. Two kinds of pump

Transfer pumps move gas out of the chamber. Rotary vane, Roots, turbo, diffusion. Capture pumps trap gas inside. Ion, titanium sublimation, NEG, cryo. A capture pump has no exhaust, so it has a finite capacity. It saturates or it regenerates.

3. Primary pumps (rough)

Rotary vane (oil-sealed). One stage reaches ~1e-2 Torr, two stages ~1e-3 Torr. An eccentric rotor with vanes, oil for sealing, exhausts to air. The catch is oil back-streaming, which can contaminate a sensitive process. [B] Kurt J. Lesker, vacuum pump technical notes, https://www.lesker.com/newweb/vacuum_pumps/vacuumpumps_technicalnotes_1.cfm

Scroll, diaphragm, screw (dry). Oil-free. Scroll reaches ~1e-2 to 1e-3 Torr. Clean primary, or backing for a dry turbo. [B] Leybold, choosing vacuum pumps, https://www.leybold.com/en-us/knowledge/blog/factors-to-consider-when-choosing-vacuum-pumps

Roots (lobe booster). Two interlocking lobes, no contact, high throughput between roughly 10 and 0.01 Torr. Needs a backing pump under it. [B] Lesker, same note.

4. Transfer pumps for high vacuum

Turbomolecular. Reaches 1e-9 to 1e-10 Torr on a metal-sealed inlet, down to ~4e-11 Torr in tandem systems. It works in the molecular regime, below ~1e-3 Torr. A high-speed bladed rotor kicks molecules toward the exit. It needs a backing pump on the foreline. Clean, oil-free, fast to start. It is replacing diffusion in most labs. [B] Lesker (ultimate); [C] Ideal Vacuum (tandem 4e-11), https://www.idealvac.com/files/literature/Sec_07_Ideal_Vacuum_Turbo_Molecular.pdf

Diffusion. Reaches below ~7.5e-11 Torr. Huge throughput, 10 to 50,000 L/s. A heated oil jet drags gas down. No moving parts, reliable, cheap. The cost is possible oil contamination. Old technology, mostly replaced by the dry turbo, but unbeaten at very large throughput. [C] Anderson Process, https://www.andersonprocess.com/how-do-vacuum-pumps-work/ ; [B] Leybold (turbo throughput ~5% of diffusion).

5. Capture pumps for UHV / XHV

Cryopump. Effective ~1e-6 to 1e-9 Torr, ultimate ~7.5e-10 Torr, speed ~1200 to 4200 L/s. Surfaces at 10 to 20 K freeze and trap the gas. Finite storage, so it must regenerate, warm up and re-pump. Strong on atmospheric gases and water, weak on helium and hydrogen. [C] VacAero (figures relayed), https://vacaero.com/information-resources/vacuum-pump-technology-education-and-training/1039-an-introduction-to-vacuum-pumps.html (note: the round figures here recur across trade pages, treat as typical orders of magnitude, not a single calibrated model.)

Sputter ion pump. The heart of static UHV.

Titanium sublimation (TSP). A titanium filament sublimes a fresh getter film on the walls. Chemisorption.

  • Range 1e-5 to 1e-12 mbar [C]. In practice used below ~1e-4 Torr. Continuous filament at that pressure (fast saturation), intermittent sublimation lower down. [B] Duniway, TSP lifetime notes.
  • Bakeable to 400-450°C. Always paired with an ion pump. [B] Duniway.
  • Pumps chemically active gases only. Sublimed Ti reacts with O2 and N2, dissociates and diffuses H2. Does not pump noble gases or methane. [A] Benvenuti, CERN getter pumping lecture, https://cds.cern.ch/record/1047072
  • Sticking probability: H2 1e-2 to 5e-2, CO 0.4 to 0.6 at room temperature, higher at liquid nitrogen. [A] Benvenuti, CERN.

Non-evaporable getter (NEG). A getter alloy (St101 Zr-Al, St707 Zr-V-Fe) or thin film (TiZrV) activated by bakeout, pumping at the surface. Widely used in accelerators as a chamber coating.

  • Ultimate below 1e-12 mbar. Low 1e-14 Torr reached at CERN with NEG, TSP and ion pump combined. [A] Benvenuti et al., Vacuum (ScienceDirect), https://www.sciencedirect.com/science/article/abs/pii/0042207X9390084N
  • Activation: St101 (Zr-Al) at ~700-740°C; St707 (Zr-V-Fe) below 450°C during bakeout; TiZrV thin-film coating activated at 180°C for 24 h, compatible with aluminium chambers. [A] Benvenuti, CERN; [B] SAES Getters, NEG activation data.
  • Chemisorbs CO, CO2, H2O, N2, O2. O2 chemisorption is irreversible. H2 is reversible: it dissociates and diffuses into the bulk of the getter, and can be released by heating. Does not sorb Ar, He, Kr, Xe. [A] USPAS, NEG session; [B] SAES.
  • Always paired with an ion pump (neither TSP nor NEG pumps noble gases or methane).

Method note

Every figure is rated A/B/C. The ion pump rests largely on primary sources (CERN, arXiv, the CERN Accelerator School via Agilent). Cryopump round figures come from trade pages and are flagged as typical. No internal lab data is used.

← The Vacuum Stack