Recommended practice for calibrating vacuum …

1 Recommended practice for calibrating vacuum gaugesof the thermal conductivity typeR. E. Ellefsona)Leybold Inficon, Incorporated, East Syracuse, New York 13057A. P. MiillerNational Institute of Standards and Technology, Gaithersburg, Maryland 20899~Received 12 August 1999; accepted 24 March 2000!This article describes and recommends various methods for calibration of thermal conductivityvacuum gauges in the pressure range of the order of 1021Pa~1023 Torr!to an atmosphere and isone of a series published by the American vacuum Society. It contains data from many sources andrepresents the opinions of a number of experts in the field.

2 The text was developed by asub-committee of the Recommended practice Committee of the American vacuum Society. TheThermal Conductivity Gauging Committee is made up of users and manufacturers of vacuumgauges who have a variety of experience in the practical use of thermal conductivity gauges. 2000 American vacuum COMMITTEE MEMBERSR ecommended practice Committee for thermal Conduc-tivity Gauging included the following people: D. H. Baker,Teledyne Electronic Technologies, H. M. Brady, ElectronTechnology, E.

3 Drubetsky, Televac, R. E. Ellefson, Chair,Leybold Inficon, Inc, A. P. Miiller, National Institute ofStandards and Technology~NIST!, R. N. Peacock, HPSD ivision/MKS Instruments, and N. T. Peacock, HPSD ivision/MKS contributors include: R. Clark, MKS Instruments,H. Gray, Edwards High vacuum International, R. W. Hy-land, NIST, and S. C. Hisel, Leybold Inficon, DISCLAIMERThis Recommended practice is based on sources and in-formation believed to be reliable, but the American VacuumSociety and the authors disclaim any warranty or liabilitybased on or relating to the contents of this American vacuum Society does not endorse prod-ucts, processes, manufacturers, and suppliers.

4 Nothing in thisarticle should be interpreted as implying such SCOPEP rocedures and apparatus are described for calibratingvacuum gauges of the thermal conductivity type by directcomparison with a calibrated reference instrument such as acapacitance diaphragm gauge~CDG!, quartz Bourdon gauge~QBG!, or spinning rotor gauge~SRG!. The pressure rangeconsidered is of the order of 1021Pa~1023 Torr! 105Pa~760 Torr!. This Recommended calibration procedure can beused for lower pressures. However, uncertainties in readingsincrease significantly for thermal conductivity gauges~TCGs!

5 Operated at very low nitrogen gas is commonly used as the test gas forTCG calibration although any dry inert gas may be used. Inthis article we do not address condensable gases, that is,gases whose saturation vapor pressure may be reached underthe temperature and pressure conditions of the explosive gas mixtures are not considered and arediscouraged from use since TCGs could be an text is directed primarily to the calibration of TCGswith their readouts~panel meter display and analog or digitaloutput!, including thermocouple gauges, Pirani gauges, andthermistor gauges, with or without convection-mode opera-tion.

6 The procedure may be adapted readily to the calibrationof other vacuum gauges whose operation is based on thermalconductivity or on a hybrid of pressure measurement meth-ods. For example, there are TCGs that do not include a pres-sure display but rather provide a dc analog voltage calibration of these gauges involves determining theanalog output voltage as a function of PRINCIPLE OF OPERATION OF THERMALCONDUCTIVITY GAUGEST hermal conductivity gauges1,2are a class of pressuremeasuring instruments in which the measured response isassociated with energy loss from a heated element~usually aheated wire!

7 The energy loss is due to thermal conductionthrough the surrounding gas and to the wire supports, and byradiation and by convection. The low-pressure limit of aTCG is reached when the pressure dependent energy lossbecomes significantly less than the pressure independentlosses due to radiation and conduction to the wire supports,as shown in Fig. 1. Energy loss by gas conduction has alinear pressure dependence at low pressures where the en-ergy transfer by gas molecules is proportional to the numberdensity of molecules.

8 As the gas density goes up and themean free path becomes shorter than the wire-to-wall dis-tance, gas molecule collisions increasingly become importanta!Electronic mail: Vac. Sci. Technol. A 18 5 , Sep Oct 20000734-2101 2000 18 5 2568 10 $ 2000 American vacuum Societyin energy transport to the wall and eventually at high pres-sures~viscous flow!the energy transfer rate becomes pres-sure independent~excluding convection effects!.1,3 Deviationfrom linear response in a wire filament Pirani4,5and aminiature6 Pirani is observed to occur at pressures of 10 100Pa~Fig.

9 1!where the mean free path is a few to 10 times thewire diameter. The upper pressure limit of a TCG may beextended to approximately 105Pa by taking advantage of the~weak!pressure dependence of convection losses. PracticalTCGs measure pressures between 1021and 103Pa~1023 10 Torr!. thermal conductivity gauges that utilize convec-tion of the gas within the gauge envelope to extend theirmeasurement range above 103Pa to 105Pa~the convectionmode of operation!are also commercially available. Occa-sionally TCGs have calibrations that extend below responses of a TCG to different gases are shown inFigs.

10 2 and are two principal thermal conductivity gauge types:the Pirani and the thermocouple. The typical heated-wiretemperature for both gauge types is 120 150 C. Figures4~a!and 4~b!show basic electrical measurement circuits forthe two types of gauges. In a Pirani gauge, the wire tempera-ture~ , its resistance!is usually kept constant and the re-quired voltage across the bridge provides a pressure-dependent signal. In a thermocouple gauge, the wire isusually heated with constant power and the pressure-dependent wire temperature is measured directly with a ther-mocouple.