Vacuum Gauge

What is a Vacuum Gauge

A vacuum gauge is an instrument that measures absolute pressure below atmospheric, that is, the degree of vacuum inside a sealed system. It belongs to the pressure-measurement family under test and measurement, but unlike a general pressure transmitter it operates at the low-pressure extreme where gas density becomes so small that direct mechanical force on a wall is often too tiny to read. Vacuum gauges therefore include not only mechanical force sensors but also gauges that infer pressure indirectly from gas properties such as thermal conductivity or ionization probability.

The single most important fact about vacuum measurement is that no one sensor spans the whole range. The vacuum spectrum covers roughly 9 to 12 orders of magnitude, from atmospheric pressure of 101,325 Pa (760 Torr, 1,013 mbar) down to extreme-high vacuum below 1e-12 Torr. Sensitivity, linearity, and the physical effect being measured all change across that span, so gauges are classified by principle and assigned to a pressure decade. A capacitance manometer that excels near 1 Torr is blind at 1e-9 Torr; an ionization gauge that reads 1e-10 Torr is destroyed if exposed to atmosphere. Selection is the act of mapping a process decade onto the right family.

The standard vacuum regime boundaries provide the reference frame engineers use. Rough or low vacuum runs from atmosphere down to about 1 Torr (1,013 down to about 1 mbar; 1e5 down to about 1e2 Pa), with some references placing the low-to-medium split near 25 Torr. Medium vacuum runs from about 1 Torr to 1e-3 Torr (about 1 to 1e-3 mbar; 1e2 to 1e-1 Pa). High vacuum (HV) runs from 1e-3 to 1e-7 Torr (1e-3 to 1e-7 mbar; 1e-1 to 1e-5 Pa). Ultra-high vacuum (UHV) runs from 1e-7 to 1e-12 Torr, down to about 1e-10 Pa, with some texts starting UHV at 1e-9 Torr. Below 1e-12 Torr lies extreme-high vacuum (XHV).

Units are a frequent source of error because three systems coexist. The exact conversions are: 1 Torr equals 101,325 divided by 760, which is 133.322 Pa and 1.33322 mbar; 1 mbar equals 100 Pa and 0.750062 Torr; and 1 Pa equals 0.00750062 Torr and 0.01 mbar. A Torr and a millibar are the same order of magnitude but not equal, since 1 Torr is about 1.333 mbar, a roughly 33 percent difference that quietly corrupts any spec comparison if treated as interchangeable. HVAC and general industrial work add inHg and microns, where 1 micron equals 1 mTorr equals 1e-3 Torr.

Four engineering questions drive every vacuum-gauge decision: which pressure decade must be measured, how accurate the reading must be, whether the process gas is air or something else, and how harsh the chemistry and pressure excursions are. These four collectively determine the gauge family, the diaphragm or filament material, and the maintenance burden over the instrument's service life. The remaining chapters work through each in turn.

Gauge Types by Principle

Vacuum gauges divide into four groups by the physical effect they exploit: direct or mechanical gauges that measure a real force on a wall and are therefore gas-type independent; thermal-conductivity gauges that measure heat loss through the gas and are gas-type dependent; ionization gauges that ionize the gas and measure an ion current proportional to density, also gas-type dependent; and wide-range combination gauges that stitch two principles into one head. The table below maps each principle to its usable range and headline accuracy.

Direct mechanical gauges are gas-type independent because they sense a real force, not a gas property. The Bourdon tube gauge is a simple, rugged dial type reading from atmospheric down to about 10 mbar (about 7.5 Torr); it serves rough vacuum only. The capsule (diaphragm dial) gauge uses an evacuated thin-walled capsule that deflects with pressure, covering atmosphere to about 1 mbar. Both are mechanical, inexpensive, and immune to gas composition, but neither reaches medium or high vacuum.

The capacitance diaphragm gauge (CDG), also called a capacitance manometer or Baratron-type gauge, is the most accurate working gauge of all. A thin ceramic or Inconel diaphragm forms one plate of a capacitor; gas pressure deflects it, changing capacitance. It is gas-type independent, and a single head spans about four decades of its full scale (FS). Full-scale heads range from 0.1 Torr to over 1,000 Torr across the market, and a 0.1 Torr FS head reads usefully down to about 1e-4 Torr (1.33e-5 mbar). Heated versions at 45, 100, or 160 degrees C resist condensation and deposition in process gas.

The spinning rotor gauge (SRG), or molecular drag gauge, levitates a steel ball magnetically and infers pressure from how fast gas decelerates it; its range is about 1e-2 down to 5e-7 Torr at roughly 1 percent of reading, with an accommodation coefficient stable to within 0.1 percent per year. That stability makes it the standard transfer standard for high-vacuum calibration. The McLeod gauge, a mercury compression manometer covering about 1e-1 to 1e-5 Torr, is the historic absolute or primary standard for calibrating other gauges, though it is limited to non-condensable gases and carries a mercury hazard.

Thermal-conductivity gauges infer pressure from gas heat loss and are gas-type dependent. The Pirani gauge reads a heated wire's resistance as gas conducts heat away; it covers about atmosphere (760 Torr) down to roughly 1e-4 Torr (5e-4 Torr practical), with convection-enhanced versions reaching 1e-3 to 1,000 Torr. The thermocouple gauge measures the heated-wire temperature directly with a thermocouple, covering about 100 Torr to 1e-4 Torr (1e-3 Torr practical), cheaper and less precise than a Pirani. The convection gauge is a convection-enhanced Pirani that extends the usable upper range to about 1,000 Torr (atmosphere) and down to about 1e-3 Torr; above about 1 Torr its mounting orientation matters.

Ionization gauges ionize the gas and read an ion current proportional to density. The hot-cathode or Bayard-Alpert gauge (BAG) uses a heated filament to emit electrons; its range is about 1e-3 to 1e-10 Torr, with glass tubulated tubes limited near 5e-10 Torr by the X-ray limit and nude UHV designs reaching about 2e-11 Torr or lower. The cold-cathode gauge (Penning, magnetron, inverted magnetron) sustains a Townsend discharge in crossed electric and magnetic fields with no filament, following I = K times p to the power m with m roughly 1 to 1.4; inverted-magnetron forms reach about 1e-2 to 1e-11 mbar at uncertainty around 30 to 50 percent. Wide-range combination gauges pair a Pirani or MEMS Pirani with a cold cathode, or a Piezo element with a MEMS Pirani, giving atmosphere to roughly 1e-9 mbar in a single port with automatic crossover.

Sensing Technologies in Depth

Chapter 2 grouped the families; this chapter explains how each works, what limits it, and which products embody it, because the failure modes follow directly from the physics. The four engineering principles in order of decreasing pressure are mechanical force, thermal conductivity, molecular drag, and ionization.

Capacitance diaphragm (CDG / Baratron). A thin ceramic or Inconel diaphragm separates the process side from an evacuated reference cavity and forms one plate of a capacitor. Process pressure deflects the diaphragm, the gap changes, and the capacitance change is converted to a linear output. Because it measures the mechanical force of gas on the diaphragm, the reading is independent of gas species, the defining advantage for process gas work. Accuracy is the best of any working gauge: the MKS Baratron 627H reaches plus-or-minus 0.12 percent of reading, the 690A reference reaches plus-or-minus 0.05, 0.08, or 0.12 percent of reading depending on range, and typical industrial CDGs sit at 0.2 to 0.5 percent of reading. Representative heads include the Agilent CDG-500 (1,000 to 1 Torr), the Inficon CDG025D (0.1 to 1,000 Torr), and the heated MKS Baratron 627H (0.02 to 25,000 Torr). Heated 45, 100, and 160 degrees C versions keep condensable or deposition-prone process gas from fouling the diaphragm.

Pirani and thermocouple (thermal conductivity). A heated wire sits in the gas; at higher pressure more gas molecules carry heat away, cooling the wire and changing its resistance (Pirani) or its temperature as read by a thermocouple. The effect is strong in the medium-vacuum range and saturates near atmosphere and near 1e-4 Torr, which sets the practical limits. Because heat conduction depends on the gas, these gauges are calibrated for nitrogen or air and require correction factors for other gases. Pirani gauges are fast, robust, and low cost, making them the workhorse for medium vacuum and the rough-vacuum stage of combination gauges; their specifications, calibration, and uncertainty for the 0.01 Pa to 150 kPa range are governed by ISO 19685:2017. Thermocouple gauges trade precision for lower cost.

Spinning rotor gauge (molecular drag). A small steel ball is suspended in a magnetic field and spun up to high speed, then allowed to coast; residual gas molecules strike it and slow it down, and the deceleration rate is proportional to pressure. With no filament, no discharge, and a physically stable accommodation coefficient drifting under 0.1 percent per year, the SRG is exceptionally reproducible, which is why NIST uses it as a transfer standard in key comparisons from 1e-4 to 1 Pa. Accuracy is about 1 percent of reading over its 1e-2 to 5e-7 Torr range. It is a metrology and reference instrument rather than a cheap process monitor.

Hot-cathode Bayard-Alpert ionization (BAG). A heated filament emits electrons that are accelerated toward a grid; collisions ionize gas molecules, and the resulting ion current is proportional to gas density and hence pressure. The lower bound is set by the X-ray limit: electrons hitting the grid generate soft X-rays that eject photoelectrons from the collector, mimicking ion current and reporting a residual pressure. Glass tubulated tubes bottom out near 5e-10 Torr; nude UHV designs reach about 2e-11 Torr or lower. The reading is gas dependent, with nitrogen sensitivity factors roughly 8 to 45 per Torr across designs and relative sensitivities such as helium about 0.18 and argon about 1.3 versus nitrogen at 1.0. The filament is vulnerable to burnout if exposed to high pressure or reactive gas. Hot-cathode gauges are governed by ISO 27894:2009 and remain the standard for high-vacuum and UHV research measurement.

Cold-cathode ionization (Penning, magnetron, inverted magnetron). Crossed electric and magnetic fields trap electrons in long helical paths, sustaining a self-maintained Townsend discharge without any hot filament; the discharge current follows I = K times p to the power m, with m roughly 1 to 1.4. With no filament to burn out, the cold cathode tolerates sudden venting, over-pressure, and reactive or corrosive gas, and the inverted-magnetron form reaches UHV at about 1e-11 mbar. The costs are coarse uncertainty around 30 to 50 percent, the need for periodic electrode cleaning as sputtered material accumulates, and slow or unstable starting in deep UHV where there are too few seed ions to strike the discharge.

Wide-range combination heads. To get one number across many decades from a single port, manufacturers combine principles with automatic crossover. A Pirani plus cold-cathode combo (for example KJLC, Pfeiffer MPT 200, DigiVac) reads atmosphere to about 7.6e-10 Torr (1e-9 mbar). A MEMS Pirani plus Piezo combo, such as the Inficon PPG550, uses the Piezo element for gas-independent measurement from 2 to 1,333 mbar and the MEMS Pirani to extend down to about 1e-5 mbar. The related Inficon PCG550 instead pairs a Pirani sensor with a capacitance-diaphragm element for gas-independent rough-vacuum readings. The Edwards WRG and WRG200 are a different architecture: they combine a Pirani sensor with a cold-cathode inverted-magnetron gauge, with no Piezo element, covering atmosphere to about 1e-9 mbar in one head.

Materials and Process Media

In vacuum measurement, material and media compatibility decides not only corrosion resistance but also whether the gauge can survive the process at all. The wetted parts that matter are the diaphragm of a capacitance gauge, the filament of a hot-cathode gauge, the wire of a Pirani gauge, and the discharge electrodes of a cold-cathode gauge. Each fails differently, and the failure mode dictates which family is acceptable for a given chemistry.

Capacitance diaphragm materials. CDG diaphragms are made from either alumina ceramic or Inconel. Alumina ceramic offers excellent corrosion resistance and a very stable mechanical response, well suited to clean and corrosive gases alike. Inconel diaphragms tolerate aggressive process chemistries common in semiconductor and chemical-vapor processes. For condensable vapors or deposition-prone gases, heated heads at 45, 100, or 160 degrees C keep the diaphragm above the condensation point so films do not build up and shift the zero. The choice of ceramic versus Inconel, heated versus unheated, is the central material decision for process-gas vacuum measurement.

Filament and wire materials. Hot-cathode ionization gauges use either tungsten filaments or thoria- or yttria-coated iridium filaments. Tungsten is simple and inexpensive but burns out quickly if exposed to oxygen or reactive gas; coated iridium runs cooler and survives brief air exposure and reactive species far better, which is why it is preferred for less forgiving processes. Burnout protection circuitry that cuts filament power above a pressure threshold is a standard option. Pirani sensing wires face the same exposure but degrade gradually through contamination rather than catastrophic burnout.

The table below maps common vacuum services to the gauge family and material strategy that survive them. It is for initial selection only; before implementation, confirm the specific gas, temperature, and deposition behavior against the manufacturer's compatibility documentation.

Two media-driven cautions recur in practice. First, the McLeod gauge and any mercury-based reference are limited to non-condensable gases and introduce a mercury hazard, so they have largely been displaced by spinning rotor gauges as transfer standards. Second, deep UHV and XHV work requires bakeable gauges with all-metal, CF (ConFlat) sealed construction; elastomer seals and non-bakeable heads outgas and prevent the system from ever reaching base pressure.

Key Specification Parameters

Vacuum-gauge datasheets list many figures, but only a handful drive a selection decision: measurement range and decades per head, accuracy and how it is stated, gas-type dependence, material and media compatibility, resolution and stability, output and integration, flange and fitting, and for ionization gauges the filament material and burnout protection. The comparison table consolidates the headline numbers for the main families before each is explained.

Measurement range and decades per head. State range in Torr, mbar, or Pa, and note how many usable decades a single head delivers. A CDG head reads about four decades of its full scale; thermal and ionization gauges each cover several decades but with widely varying confidence across that span. The mismatch between the process decade and the gauge's sweet spot is the most common selection error.

Accuracy and how it is stated. There are two fundamentally different accuracy statements. Gas-independent gauges quote percent of reading: CDG and SRG sit at 0.05 to 0.5 percent of reading, so the absolute error scales with the value displayed. Thermal and ionization gauges instead carry uncertainty on the order of plus-or-minus 10 to 30 percent per order of magnitude, and cold cathodes can reach 30 to 50 percent. Always note the temperature at which accuracy is specified, since heated CDG heads quote at their operating temperature.

Gas-type dependence. Direct, CDG, and SRG gauges are gas independent; thermal and ionization gauges need correction factors keyed to the gas. This is critical whenever the process gas is not air. For ionization gauges, relative sensitivities such as helium about 0.18 and argon about 1.3 versus nitrogen at 1.0 must be applied, or the reported pressure will be wrong by a factor of several.

Output, integration, and fittings. Modern gauge controllers provide analog 0 to 10 V or 4 to 20 mA, often log-linear, plus digital RS-232, RS-485, EtherCAT, or DeviceNet, and programmable setpoint relays that drive pump sequencing and interlocks. Vacuum fittings are their own discipline: KF (NW) quick flanges for rough and medium vacuum, CF (ConFlat) all-metal flanges for HV and UHV, plus VCR and ISO-K, with bakeability required for UHV. Resolution, repeatability, drift or long-term stability, and response time round out the comparison, and for ionization gauges the filament material (tungsten versus coated iridium) and the presence of burnout protection belong on the spec sheet.

Governing standards. Calibration by direct comparison with a reference gauge is defined by ISO 3567:2011, with uncertainty evaluation in ISO 27893:2011. Pirani gauge specifications, calibration, and measurement uncertainty for 0.01 Pa to 150 kPa are covered by ISO 19685:2017, and hot-cathode ionization gauges by ISO 27894:2009. JIS Z 8750 gives the Japanese calibration method, DIN 28400 defines vacuum terminology, and AVS (American Vacuum Society) recommended practices plus NIST primary standards and transfer-standard practice complete the framework.

Selection Decision Factors

To turn the preceding chapters into a specific gauge, follow the decision sequence below. Most vacuum-gauge errors come not from one wrong figure but from deciding accuracy or output before the pressure decade and gas chemistry have fixed the family. These six steps work as a fixed RFQ template.

1. Pressure decade required: this picks the family first. Rough vacuum points to Bourdon, capsule, Pirani, or CDG; medium vacuum to Pirani, convection, or CDG; high vacuum to a Bayard-Alpert hot-cathode or a cold cathode; UHV to a nude BAG or an inverted-magnetron cold cathode. If one port must span the full range, choose a combination or wide-range gauge with automatic crossover.
2. Accuracy demanded: for absolute, process-critical readings or a calibration reference, choose a CDG / Baratron or an SRG, both quoting percent of reading and gas independent. For trend monitoring or pump and interlock control, the coarser uncertainty of thermal and ionization gauges is acceptable and far cheaper.
3. Gas composition: for non-air or varying gas mixes, prefer gas-independent gauges (CDG, SRG, or Piezo), or commit to applying correction factors on thermal and ionization gauges. Skipping this step is the single most common cause of a reading that is wrong by a factor of several.
4. Process hazards: for corrosive, condensable, or deposition-prone gas, use a heated CDG with a ceramic or Inconel diaphragm, or a cold cathode that has no filament to attack. For frequent venting or over-pressure events, a cold cathode survives where a hot-filament BAG burns out.
5. Integration: specify the required output (analog or digital), setpoint relays for pump and interlock logic, the flange type (KF, CF, VCR, ISO-K), and bakeability if the system reaches UHV. The controller and connector ecosystem often locks in a manufacturer.
6. Cost and maintenance: thermal gauges are cheapest; CDG and SRG are premium; ionization gauges incur ongoing filament or electrode upkeep and recalibration. Weigh the lifecycle cost, not just the purchase price, since a mis-specified gauge that drifts or burns out repeatedly dominates total cost.

One last dimension is manufacturer ecosystem and serviceability: controller compatibility, replacement filament and sensor availability, recalibration service, and a documented correction-factor library for your process gases. The established vacuum-instrument makers each cover the full family. MKS Instruments, including Granville-Phillips, leads the market with Baratron capacitance manometers (626/627/690), Granville-Phillips and Series 274 Bayard-Alpert ionization gauges, the SRG-3 spinning rotor transfer standard, and the 9xx controllers. INFICON, including InstruTech, offers the CDG025D capacitance gauge, PSG Pirani gauges, cold cathodes, and the PPG550 and PCG MEMS Pirani-plus-Piezo wide-range gauges. Pfeiffer Vacuum supplies the CenterLine and ActiveLine families: TPR Pirani, IKR cold cathode, hot cathode, the MPT 200 combo, and CMR capacitance. Leybold (Atlas Copco) carries BOURDONVAC and capsule dial gauges, THERMOVAC Pirani, PENNINGVAC cold cathode, and IONIVAC hot cathode. Edwards Vacuum (Atlas Copco) offers APG active Pirani, AIM active inverted magnetron, the WRG and WRG200 wide-range heads, and ASG and CDG diaphragm gauges. Agilent (Varian) provides the CDG-500, ConvecTorr convection, FRG full-range, Bayard-Alpert, and inverted-magnetron gauges. Kurt J. Lesker (KJLC) supplies cold-cathode and Pirani combos, ion gauges, and controllers, and resells MKS and Inficon. Teledyne Hastings focuses on thermal and thermocouple instruments, and ULVAC, Canon ANELVA, Azbil, SRS (Stanford Research IGC100), the Fredericks Company, and DigiVac round out the field.