The Vacuum Stack

Gauges

You cannot pump what you cannot see, and no single gauge sees the whole range. Vacuum measurement runs more than seventeen decades, atmosphere down to ~10⁻¹² Pa, and no instrument covers it. You always run several, and you learn where each one lies.

The families split three ways. Thermal (Pirani, thermocouple) read pressure off how fast the gas carries heat away. Ionization (hot-cathode Bayard-Alpert and extractor, cold-cathode Penning) read it off the ion current the gas makes. Mechanical / reference (capacitance manometer, spinning rotor) read a force or a drag. [A]

Pressure in Torr (1 Torr ≈ 1.33 mbar).

The rule under everything. Every ionization gauge depends on the gas. Relative sensitivity (nitrogen = 1) runs about 4 for acetone vapour, 0.18 for helium. An ion gauge reads the wrong number unless you correct for the gas in front of it. [B]

Capacitance manometer (Baratron). Atmosphere to ~1 mbar, lower on heated models. A tensioned metal (Inconel) diaphragm, ceramic Al₂O₃ in some designs, deflects and its capacitance shifts. Independent of the gas, which makes it the precision reference at low and medium range. [B] Spinning rotor (SRG). A magnetically levitated ball whose deceleration reads pressure, a transfer reference for high-vacuum calibration. [A]

A capacitance manometer (Baratron) cutaway, the ceramic membrane and its reference cavity

Pirani. ~67 Torr to ~1e-4 Torr, but below ~5e-4 Torr it has only one good digit. A heated filament loses heat to the gas; more gas, more cooling. Gas-dependent, so it wants calibration per gas. Small, robust, cheap, the everyday rough/medium gauge. [A] Thermocouple. Same heat-loss principle, read by a thermocouple; rough and medium range. [C]

A Pirani gauge, the heated filament in its tube

A hot filament emits electrons that ionize the gas; the ion current is proportional to pressure. [A][B]

Gauge	Range (Torr)	Note
Bayard-Alpert (standard)	~1e-3 → 1e-10	the UHV reference
Bayard-Alpert (nude UHV)	~1e-4 → 1e-11	grid as an open basket on a flange
Extractor (e.g. IE514)	~1e-4 → < 1e-11	cuts soft-X-ray and ESD noise → XHV
Schulz-Phelps	near 1e-1	the high-pressure end, where the BAG saturates

A nude Bayard-Alpert gauge, fine collector wire inside the open grid basket

Shop limits. The hot filament outgasses and has a finite life. For UHV, spec gauges bakeable to 450°C, with dual / replaceable filaments and electron-bombardment degas. [A]

Penning (cold cathode, and inverted magnetron). ~1e-2 to 1e-9 Torr. A cold-cathode discharge held by a magnetic field, no filament to burn out, rugged, long-lived, made for industry. The cost: slow to strike at very low pressure, and less linear than the hot cathode. [B]

A Penning cold-cathode gauge, the discharge held in its magnetic field

Gauge	Range (Torr)	Principle	Gas-dependent	Use
Capacitance (Baratron)	1500 → ~1e-4	mechanical, membrane	no	precision reference
Pirani	~67 → ~1e-4	thermal conduction	yes	rough/medium, robust
Thermocouple	~10 → ~1e-3	thermal conduction	yes	rough, simple
Penning (cold cathode)	~1e-2 → ~1e-9	ionization, B field	yes	HV/UHV, industrial
Bayard-Alpert (hot cathode)	~1e-3 → ~1e-10 (nude 1e-11)	thermionic ionization	yes	UHV reference
Extractor	~1e-4 → < 1e-11	low-noise ionization	yes	XHV
Spinning rotor	high vacuum	levitated rotor	nearly no	calibration

The field rules that fall out of it:

Cross a (relative) ion gauge against an (absolute) capacitance manometer when the true value matters.
In UHV, measure where it counts, there can be a large pressure drop between the point of interest and the gauge.
Always correct by the sensitivity factor of the dominant gas.
A hot filament is fine but mortal. A cold cathode is rugged but coarse.

Sources · 6

Rated [A] primary, [B] manufacturer, [C] trade (cross-checked). Priority to the Cornell/USPAS course and the public SLAC spec.