Capteurs de la pression atmosphérique au vide primaire Common Vacuum Gauge Technologies in the range atmospheric pressure to 10-4 mbar 2ème Rencontre du réseau "vide" du CNRS Dr Andrew Chew La Rochelle Tuesday 23rd October 2012 Confidentiality Statement This presentation has been prepared exclusively for the benefit and use of Edwards and is confidential in all respects. This presentation does not carry any right of publication or disclosure, in whole or in part, to any other party. This presentation is the property of Edwards. Neither this presentation nor any of its contents may be used for any purpose without the prior written consent of Edwards. This presentation includes certain statements, estimates, targets and projections as to anticipated future business performance. Such may reflect significant assumptions and subjective judgements by Edwards which may or may not prove to be correct. Edwards makes no representations as to the accuracy, completeness or fairness of this presentation and so far as is permitted by law, no responsibility or liability whatsoever is accepted by Edwards for the accuracy or sufficiency thereof or for any errors, omissions, or misstatements relating thereto. Disclaimer Edwards Ltd, disclaim any and all liability and any warranty whatsoever relating to the accuracy, practice, safety and results of the information, procedures or their applications described herein Contents Vacuum ranges – gauge operating ranges - Definitions: Total/partial, Direct/Indirect, absolute/non absolute, accuracy - Classification of gauges according to measurement principles • • Gauges commonly used in the vacuum range 10-4 to 1013 mbar • - U-tube manometer McLeod Rotating piston • • • • • • Mechanical Bourdon Capsule - Strain Gauge, Piezo/Piezoresistive Capacitance Manometer • Pirani Thermocouple - Convection • Spinning Rotor Gauge • • • Hot and Cold cathode ionisation gauges Vacuum Gauges • The measurement of pressure characterizes the state of a vacuum environment and/or allows process control • Accuracy/uncertainty (the level of systematic error), precision (the level of random error), resolution level, measuring range, linearity of output, reproducibility, robustness, process compatability and response time are key performance criteria • Also size, cost and serviceability Total Pressure Gauges - Classification Absolute (Primary) gauges - can be calibrated from their own physical properties in terms of fundamental units (measure relative to perfect vacuum) Non-absolute (Secondary or Transfer) gauges cannot be so – hence need to be calibrated Classification Total pressure gauges measure sum/total pressure (Dalton’s law) of all gas species and partial pressure gauges* identify and quantify individual gas species • Direct gauges- respond directly to force associated with pressure; the reading is independent of gas species - measure the force exerted by the gas on a surface of some sort • Indirect gauges – measure pressure dependent physical property of the gas (e.g. thermal conduction or viscous/molecular drag) or a physical quantity ∝ number density (e.g ions). - reading is dependent on gas species: Ptrue = Pmeasured/f where f = gauge or gas/sensitivity factor *partial pressure gauges will be discussed in a separate lecture Basics Pressure = Force per Unit Area = Pascal = N/m2 • So if we wish to measure pressure directly by measuring the mechanical force exerted on some sort of transducer of area 1cm2 the forces are: Pressure Force (N) 1013.25 mbar 10 1 mbar 10-2 10-4 mbar 10-6 10-9 mbar 10-11 Classification Elastic deformation U-tube manometer Balance of hydrostatic forces Classification Kinetic effects Charged particles Decrement/ Resonance Rotating disc Radiometric Convection U-tube manometer Many other mechanisms Flourescence Particle scattering Brownian motion Thermal radiation ( e.g. Knudsen radiometer) Radioactivity Scattering UV photons Radial drift etc………… Common Total Pressure Gauges - Hierarchy Liquid wall Direct Gauges U-tube McLeod Strain Elastic element Diaphragm Bourdon Capsule dial Piston Molecular, Viscosity Indirect Gauges Thermal conductivity (energy transfer) Spinning Rotor Others: quartz fibre, rotating disk, quartz oscillator Pirani, convection Schulz-Phelps Thermocouple Extractor Hot Cathode Ionization Capacitance Cold Cathode Others (radioactive) Bayard Alpert Penning Inverted Magnetron Gauge measurement ranges U-tube liquid McLeod Bourdon Common Total Pressure Gauges – operating ranges Gauges which are subject of this lecture Capsule Dial Strain Capacitance Diaphragm Pirani Thermocouple Convection Spinning Rotor Hot Cathode Ionisation SchulzPhelps Cold Cathode Ionisation 10-11 10-10 10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1 Range extended by special construction mbar 1 101 102 103 Torricelli 1643 Evangelista (Torr)icelli (1608-1647) Gasparo Berti….water manometer ~ 1640 Blaise Pascal 1646 refuted Aristotle’s view horror vacui that nature abhors a vacuum Liquid Level Manometers direct and absolute • U-tube manometer absolute • At equilibrium P1 – P2 = ρ∆hg • If one side is sealed and evacuated to a relatively negligible pressure the manometer is directly readable. Accuracy and operating range is limited mainly due to capillary effects at the liquid–wall interface • Measurements 103 to about 1 mbar but interferometric methods and high purity mercury enable very high accuracy readings (<±0.0001%) and pressure range to 0.01 mbar • Still most accurate gauge for pressures >1 mbar differential McLeod Gauge 1874 • McLeod compression gauges extend the measuring range utilising the Boyle-Mariotte Law: pressure of gas to be measured is confined to a fixed volume and compressed by the rising column, effectively amplifying the measurable height differential • Measures only non-condensables • Allowing for the condensation of vapours and using mercury vapour cold trap (and corrections) allows measurements at the 10-5 mbar level (<±1% accuracy) A = capillary crosssectional area p = h2 A / V Piston gauge • • The measurement range covered is below 103 mbar A low end threshold component low enough to support the calibration of typical transfer standards with ranges as low as 1.3 mbar and even 0.13 mbar Piston-type gauges counterbalance the pressure of a gas with a spring or a solid weight, in which case it is known as a dead-weight tester and may be used for calibration of other gauges Direct and absolute: accuracy ±[0.11 +(11.3 x 10-6p)] Pa = ± 0.11% @100 Pa Elastic Element Gauges • Use a diaphragm or membrane to separate an evacuated and hermetically sealed chamber (established pressure) from a volume contiguous with vacuum (effectively replacing the liquid in the liquid level manometer) • The differential pressure across membrane produces force to deflect it • The indication is independent of ambient temperature and atmospheric pressure change • These are not absolute but give a direct reading Bourdon Gauge (1849) Internal tube is bent into an arc and connected to vacuum. The external pressure (and hence differential) causes tube to deflect and move needle. Changes from 103 to 10 mbar with an accuracy ±1% (at higher pressures this can be <±0.1%) Correction for atmospheric pressure change needed Capsule dial •The capsule dial gauge Uses an hermetically sealed and evacuated capsule composed of corrugated membranes – aneroid (no fluid) •Consequently, it is barometrically compensated (independent of atmospheric pressure changes). Gears and levers Sealed diaphragm Capsule Dial Gauge • Typical full scale ranges are 0 to 25, 50, 125 and 1013 mbar measuring through to 1 mbar • Particular ranges can be selected for heightened accuracy by filling a sensitive capsule element with gas at known pressure e.g. 100-110 mbar range with ±1% accuracy. • Temperature fluctuations are important: from the ideal gas equation ∆T = 3K will give an uncertainty of 1% Generally quoted to ±1 to 2% x full scale deflection accuracy • Diaphragm Gauge Non-absolute but give a direct reading Cavity sealed Strain sensor Change in pressure causes diaphragm to move against the pressure sensitive plate. Plate voltage will change with applied pressure. Diaphragm Vacuum Strain Gauge • Deflection produces a change in electrical resistance of piezoelectric transducer • Membrane - usually silicone and the compression/extension conducting layer forms part of a resistive bridge network • Rugged and compact - balanced below 1 mbar with range 1 to 2000 mbar (~±1% of full scale deflection) A modern diaphragm gauge Grown Piezoresistive Sensor Bridge Measurement Schematic Diaphragm = Capacitance Manometer Capacitor electrodes (Inconel) diaphragm Getter Cavity sealed < 1x10-7 mbar Vacuum Capacitance Manometer - Signal Flow 2 Fixed electrodes The diaphragm will move with a change in pressure. 7 1 5 Change in pressure. 3 When the diaphragm moves, the gap from the fixed electrodes to the diaphragm will vary. 0 to10V DC Output Variable capacitance signal sent to electronics. 4 6 Electronics condition signal to supply linear output. When the gap varies, the electrical capacitance varies as well. Capacitance Diaphragm Gauge • Inductive transducers use ferromagnetic rod to penetrate the core of differential transformer coil such that deflection of the diaphragm causes movement • Capacitance diaphragm gauges are more prevalent – using FM modulated coaxial capacitor to detect very small deflections exploit fully +ive and -ive deflection of (25 µm) diaphragm 0.4 nm deflections are measurable! • Differential types for lower pressures via an established pressure (or to measure the pressure differential) Made with stainless-steel, inconel and monel, for corrosion resistance More details in a separate lecture Capacitance Diaphragm Gauge • Thermal fluctuations may cause differential expansion and they can be heated to a constant temperature • Range 10-5 mbar to 104 mbar (104 dynamic range) • Resolution usually 1 part in 104 • High repeatability Commonly quoted accuracy of ±0.15% of reading + 0.005% of full scale In practice range is ±0.3 to 10% Thermal conductivity gauges Thermal gauges utilise thermal transfer as an analogue of pressure. A filament heated in vacuum loses heat by convection, conduction . and radiation Kn =λ/d d = diameter of filament convection conduction Kn = 10 Kn = 0.01 Radiation/ conduction losses through wire mbar Non-absolute and indirect Thermal Conductivity Gauges • These exploit heat transfer between heated wire (400K) and vacuum wall (at 300K) • pressure dependent 0.01 < Kn < 10 Consider 1. In molecular flow energy current equals the conduction heat loss through gas: φ is sensitivity constant (dependent on the gas properties) and g is geometrical factor P Q =φ 1 + gP Thermal Conductivity Gauges At lowest pressures other effects become predominant: 2. Heat loss (conduction) from ends of wire (most dominant for short wires) 3. Thermal radiation losses Q = Aσε (T High wire-length/diameter ratio and low emissivity material minimises these losses to extend to lowest pressures. In practice changes in emissivity and thermal (or energy) accommodation of materials limit these gauges to accuracy of ±10 to 100%. 4 wire −T 4 wall ) Pirani Gauge (1906) The Pirani gauge operates in the pressure regime where conduction is predominant. Heated wire forms one arm of a Wheatstone bridge; as filament resistance changes bridge becomes out of balance. There are two modes of operation • the filament is maintained at a constant temperature i.e. resistance: sensitivity falls at higher pressures/temperatures • a constant voltage is applied to the filament: constant sensitivity but difficult to operate Reference compensating filament at atmospheric pressure and close temperature can minimise thermal effects The sensitivity of the gauge is both pressure dependent and gas species dependent, so calibration is essential. Pirani gauges operate between 103 mbar and 10-4 mbar. Pirani Gauge – Wheatstone Bridge Constant voltage type Pirani Gauge Calibration typically at atmosphere and lowest vacuum points Heat conductivity varies with gas species..use calibration chart The divergence at higher pressures is due to convection becoming more important These are not high accuracy gauges and contamination of the filament can cause serious shifts in sensitivity, but clean gauges accuracy ± 15% Any thermal gauge will have a relatively long time constant High pressure - heat flow is independent of pressure except small convection effect which is exploited to measure to atmospheric pressures Modern gauges can include electronics to linearise output A Solid State Pirani Sensor Construction of the sensor Cross section of the sensor Thermocouple Gauges • The filament of a thermocouple gauge is heated by a constant current and filament temperature is measured by thermoelectric indications • Same principle and limitations as Pirani gauge – different measurement technique • The lower limit of these gauges is extendable 10-3 to 10-4 mbar but dynamic range is 103 • Accuracy is >±30 to 50% Thermocouple Heated filament Thermocouple Supply Voltage heated filament and thermocouple junction thermocouple monitors the temperature of the wire filament: the hotter the wire, the lower the pressure Convection Gauge Similar to the Pirani gauge, this sensor uses a temperature compensated, gold-plated tungsten wire to measure heat transfer by both conduction and convection, and thereby give better performance at higher pressures. The wire is inside a concentric insulated cylinder In the range 10-4 to 10 mbar the gauge operates as per a Pirani ………. …...from 10 to 103 mbar convection is the heat transfer mechanism Other features and limitations of this sensor are the same as those of Pirani and most thermocouple gauges. Accuracy ±15% of reading ±3 x 10-4 mbar Repeatability ±5% of reading Higher pressure reading is susceptible to mounting orientation –needs horizontal filament? Typical materials exposed to vacuum Convection • Filament: Tungsten (Gold plated) • Sintered filter: Phosphor-bronze • Gauge tube: Aluminium or stainless steel (316L) • Other: Nickel, fluoro-elastomer, PTFE Pirani • Filament - Tungsten / Rhenium - Platinum / Iridium • Tube: Stainless Steel 316L & 304L • Filter: Stainless Steel 316L • Other Glass, Ni, NiFe, PTFE (APG100-XLC only) Thermocouple • Nickel plated steel, • alloy 52 (iron nickel) pins, • stainless steel, • gold, platinum, rhodium and palladium. Molecular Gauges • Momentum transfer or molecular gauges -exploit molecular drag phenomenon • Confusingly also referred to as friction or viscosity gauges – strictly these exploit viscous drag • Several of these span flow regimes Non-absolute and indirect The Spinning Rotor Gauge – 1924 Epstein and commercial instrument mid-1980s (Fremerey and Beams) Deceleration of a high speed (~400Hz) spinning steel ball due to molecular drag measured. The rotation of the ball – which has a small magnetic moment – is sensed by a pickup coil The sensitivity of the gauge is relatively independent of gas species and is very stable - the uncertainty is better that 3% and stability better than 2% per annum The gauge can be used as a transfer standard for calibration Range 0.1 mbar to 10-7 mbar Spinning Rotor Gauge • ω 10σ eff P = = − Fbraking ω πρrv Deceleration rate measured in practice by considering fixed number of rotations N and sequence of timing intervals τn which elapse for N rotations is measured such that • ω τ i +1 − τ i − = ω τ i .τ i +1 Spinning Rotor Gauge Other decelerating effects • • ω = − Fbraking − Fresidual − 2α T − FRN ω 1K/h drift has an equivalent N2 pressure of 1 x 10-7 mbar, ensuring thermal stability allows equivalent N2 pressures of 4 x 10-9 mbar • Reducing vibrations to < 0.03 g means that Fresidual (the pressure independent decelerating component) due to frequency dependency of eddy currents is largest ‘offset’ effect Spinning Rotor Gauge Low pressures - molecular braking effect becomes of same order as the residual effect • this determines stability and lower measurement limit (Fresidual is of order 10-7 mbar and needs to be measured repeatedly) σeff means SRG is not absolute but is used widely as a transfer (or secondary) calibration standard • In molecular flow common user operation gives < ±5% accuracy (intercomparisons by national standards laboratories suggesting accuracy at the ±1% level) Spinning Rotor Gauge For pressures >10-2 mbar a special linearization procedure is used to compensate for the diminishing pressure dependence of the rotor deceleration rate • effectively an interpolation procedure to measure to 10 mbar with an uncertainty of > ±10% Residual drag limits lowest pressure accuracy - caused e.g. by non-coincident magnetic centre and mass centre – can be compensated too some degree but varies from ball-to-ball Quartz friction gauge - electrically driven forced oscillation is damped by the gas - pressure measured by changes in resonance impedance ± 10% accuracy in range 10-3 to 103 mbar Indirect measurement Rotating disc gauge Measures molecular torque developed between coaxial discs Dushman 1914 absolute, direct 1989-1995 GD ~ pressure GI suspension asymmetries GM molecular drag GE Eddy current drag Ionisation Gauges* Molecules ionized by electron impact and resulting ion-current measured • Hot Cathode/Thermionic • Bayard Alpert Gauge (BAG) • Extractor gauge • Shultz-Phelps • Cold Cathode Discharge Gauges • Penning/Philips Gauge • Inverted Magnetron Gauge *These will be discussed in more detail in next lecture Ionisation Gauge Principle Vacuum System Electrode Electrode Current flow Ammeter Voltage Source Current flow related to pressure Bayard-Alpert Gauge (BAG): 1950 Overcame the low pressure X-ray limit of Triode gauges thermionic cathode (hot) filament K is a source of electrons I- and is parallel to positive grid anode A, ion collector C is concentric to this (greater -ive potential than other electrodes) Grid A (+ve) Filament K Collector C (-ve) Hot Cathode Ionisation Gauges I+ = n λ σ In = number density λ = electron path length σ = ionisation cross-section/ionisation probability hence there is sensitivity correction for different gases Accuracy ±10 to 100% Ionisation Gauges >10-4 mbar space charge effects become important: ions orbit around the collector rather than reach it which reduces the number of ionizing electrons and combined increased Scattering reduces the apparent sensitivity reading in BAG by 75% BAG has an upper pressure limit of about 10-3 mbar to avoid filament burn out and because of space-charge effects Schulz-Phelps (1957) version extends upper range from 10-3 mbar to 1 mbar by close-spacing the electrodes: only one ionization per electron and hence lower sensitivity Anode Collector Cathode Cold Cathode Gauges • Path length of electrons is increased c.f. hot-cathode gauge by using anode at high voltage (few kV) above cathode and strong magnetic field - Penning or Philips Gauge • In Inverted Magnetron: anode and cathode are parallel and magnetic field is parallel to Anode (electrons have tight spiral paths = high collision probability) • ‘Strike’ by heated filament, field emission, cosmic rays etc. Radioactive emitter works best at high pressure but no longer commercially available because of EHS restrictions Penning Crossed electric and magnetic fields trap electrons. The high voltage ranges from 2 to 6 kV at the anode whilst the cathode is grounded and the magnetic field 1 to2 kGauss cathode anode Ions are collected on the ring anode Anode current i+ = kps K and s (~1.1 to 1.2) are gas and gauge dependent At low pressures the discharge is unstable and the calibration can change abruptly Range ~10-3 to 10-8 mbar Penning Range of a standard Penning gauge is ~ 10-3 mbar to 10-8 mbar The accuracy of Penning gauges is poor: +20 to 100% and -50%, especially at low pressures and large changes in sensitivity are not uncommon They are susceptible to contamination leading to errors in pressure measurement. 1927 Inverted Magnetron – 1958 Hobson and Redhead With this orientation of the electric and magnetic fields, electrons are able to perform relatively long orbits around the anode increasing ionization event – allows measurement to lower pressures ~10-9 to 10-10 mbar Robust, no heat Magnet Atom Gas N Electron path N Ion Cathode Anode S Insulator S Gauge body Accuracy -50 to +100%, Wide Range Gauge Comines Pirani and Inverted Magnetron : 10-9 to 103 mbar Pirani Anode Striking filament Cathode cups Body tube Magnets Calibration* Need calibration especially for process establishment, gauge retrofit and inter-comparison and for gases other than calibration gas (N2) Establish calibration pressure by: 1. absolute gauge = primary gauge (from 103 to 10-2 mbar) 2. Static expansion ±1% 10 to 10-7 mbar; Dynamic expansion - gas at a known flow rate through known conductance 10-5 (±1%) to 10-9 mbar (±5%) 3. Molecular beams to determine a pressure with a gauge calibrated as above: 10-8 (±7%) to 10-12 mbar (±30%) *Will be discussed in separate lecture Calibration Here output voltage or pressure indication varies in proportion to applied pressure over a wide range (several decades) Used for interpolation and extrapolation between calibration points Linear: Capsule dial, Strain, HCIG, Capacitance, SRG.. Non linear: Cold-Cathode, McLeod, Pirani, Thermocouple.. Secondary gauges • Secondary (transfer) gauges used as traceable reference - not absolute but calibrated as before (often CM, SRG and HCIG) • National standards labs (NIST, BPM, NPL, PTB, LNE, INGC etc.) for primary calibration of transfer gauges then used on calibration chambers • Commercial calibration labs: typical accuracy of ±26% at 10-6 mbar to ± 2.3% at 103 mbar Linearity Here output voltage or pressure indication varies in proportion to applied pressure over a wide range (several decades) Used for interpolation and extrapolation between calibration points Linear: Capsule dial, Strain, HCIG, Capacitance, SRG.. Non linear: Cold-Cathode, McLeod, Pirani, Thermocouple.. Vacuum Gauges - Traditional Gauging System • • Gauge lead has special plug for each type of head Gauge lead is costly shielded or co-axial cable Vacuum Gauges - Active Gauge Heads • • • Surface mount technology allows amplification circuits to be mounted in gauge head rather than the readout Millivolt signal amplified to 2 to 10 v dc at gauge head Linearised output Acknowledgements Dr Karl Jousten, PTB, Germany • http://cas.web.cern.ch/cas/Spain-2006/Spain-lectures.htm Dr Ron Reid, ASTEC, UK • http://www.cockcroft.ac.uk/education/PG_courses_200910/Spring_2010/Reid%20Lecture%208.pdf • Many good texts on vacuum metrology Capteurs de la pression atmosphérique au vide primaire Common Vacuum Gauge Technologies in the range atmospheric pressure to 10-4 mbar Rencontre du réseau "vide" du CNRS Andrew Chew Dr Andrew Chew La Rochelle Tuesday 23rd October 2012
© Copyright 2026 Paperzz