International Comparison CCQM-K10 Final Report Franklin Guenther 1, George Rhoderick 1, Alain Marschal 2, Anneliese Medem 3, Kenji Kato 4, Gwi Suk Heo 5, Ed de Leer 6, Annarita Baldan 6, Theo L. Hafkenscheid 6, Gerard Nieuwenkamp 6, Adriaan M.H. van der Veen 6, Leonid Konopelko 7, Chris Brookes 8, Hansa d Souza 8, Martin Milton 8 1 National Institute of Standards and Technology (NIST), Chemical Science and Technology Laboratory, 100 Bureau Drive Stop 8393, Gaithersburg, MD SA 2 BNM-LNE, Centre Métrologie et Instrumentation, 1, rue Gaston Boissier, 75724 Paris Cedex15, France 3 Federal Environmental Agency (BA), Paul Ehrlich Strasse 29, DE-63225 LANGEN, Germany 4 National Metrology Institute of Japan (NMIJ), 1-1, Hagashi, Tsukuba-shi, Ibaraki, 305, Japan 5 Korea Research Institute of Standards and Science (KRISS), Division of Chemical Metrology and Materials Evaluation, P.O. Box 102, Yusong, Taejon, Republic of Korea 6 NMi Van Swinden Laboratorium B.V. (NMi VSL), Schoemakerstraat 97, 2628 VK Delft, the Netherlands 7 D.I.Mendeleyev Institute for Metrology (VNIIM), Laboratory of State Standards in the Field of Analytical Measurements, 19, Moskovsky Prospekt, 198005 St-Petersburg, Russia 8 National Physical Laboratory (NPL), Teddington. Middlesex, TW11 OLW, K Field Amount of Substance: Gas Standards Subject Comparison of primary gas standards containing benzene, toluene, and o-xylene in a balance of nitrogen. Participants DE (BA), FR (BNM), JP (NMIJ), KP (KRISS), NL (NMi), RF (VNIIM), K (NPL), S (NIST) Organizing body CCQM Gas Standards Working Group Background This key comparison was intended to compare the capabilities for the preparation and value assignment of gas standards for benzene, toluene, and o-xylene in nitrogen (subsequently referred to as BTX), maintained at the participating national metrology institutes. The range of the nominal amount-of-substance fractions of the comparison standard is 5 nmol/mol to 10 nmol/mol, which is close to regulatory levels in most countries. Page 1 of 21
Conduct of the Comparison The National Institute of Standards and Technology (NIST) prepared two 30-liter primary standard gas mixtures (PSM) of BTX in nitrogen using normal gravimetric procedures [1]. These two PSMs were then rigorously compared to other PSMs maintained at NIST to verify the gravimetrically determined amount-of-substance fraction. A total of eight 6- liter cylinders, treated with the proprietary deactivation process called Aculife-IV, were procured from Scott Specialty Gases in Plumsteadville PA. 1 These cylinders were previously used for CCQM-K7 and were determined to have contained a stable BTX gas mixture for over a year. These cylinders were repeatedly evacuated and filled with nitrogen to purge the previous gas mixture, and the nitrogen analyzed until no BTX remained. The cylinders where then filled through a gas manifold from one of the gravimetrically prepared PSMs. The cylinders were held at NIST for a two-month period and compared to the parent PSM to confirm that the BTX samples were stable. Each participating laboratory was shipped one 6-liter cylinder. This cylinder was to be analyzed by the receiving laboratory and then returned to NIST. NIST then reanalyzed the cylinder contents to assure that the compounds remained stable throughout the comparison. The comparison was officially declared closed on November 1, 2001. All participating laboratories had submitted complete reports by this date. Gravimetric Preparation Calculations The mass determination equation of the organic liquid using the method of filling and sealing glass capillary tubes is: m = m + S + D + D Organic Diff C Liquid Vapor where m Diff is the mass difference of the full capillary and the empty capillary, S C is a correction factor for loss of mass upon sealing, D Liquid is the mass of air displaced by the liquid organic, and D Vapor is the mass of air displaced by the organic vapor. The standard uncertainty of m Diff is estimated from three repeat measurements of the mass of the capillary tube before and after filling. The value of S C was estimated by sealing 10 empty capillaries and measuring the subsequent mass loss, and is equal to 0.8 µg with a standard uncertainty of 0.5 µg. The quantity D Liquid is estimated using the mass of organic liquid (m Diff ), the organic liquid density, and the density of air. The standard uncertainty is estimated by combining the standard uncertainty of m Diff and an estimated uncertainty of 1 % for the liquid density (rectangular distribution). The quantity D Vapor is estimated using the vapor pressure of the organic and the gas equation (PV=nRT). Equilibrium conditions and ideal gas behavior are not assumed, thus only 50 % of the calculated equilibrium value is used, with an uncertainty equal to the estimated value. 1 Certain commercial equipment, instruments, or materials are identified in this paper to foster understanding. Such identification does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the material or equipment identified is and best available for the purpose. Page 2 of 21
The total moles of gas added to the cylinder is calculated by: mn m m 2 Air Z Balance = + + M M M N2 Air Organic Organic The quantity m is the mass of nitrogen, air, or organic liquid added, and the quantity M is the associated molar mass. The amount of air is quite small, equal to approximately 0.15 % of the nitrogen mass. The amount of organic added is also very small, and was ignored in this calculation. The standard uncertainty of these mass determinations is 1 g, or approximately 0.02 % relative. Finally the amount-of-substance fraction (concentration) is calculated using: Concentration = ( m Organic M P Organic Z Organic Balance ) + N Organic where M Organic is the molar mass of the organic liquid, P Organic is the purity of the organic liquid, N Organic is the moles of added BTX due to the balance gas, and Z Balance is the total moles of gas added. The molar masses are assumed to have relatively small uncertainties, which are ignored. The quantity P Organic was determined by gas chromatography with flame ionization detection, gas chromatography with mass selective detection, and by differential scanning calorimetry. The worst-case result was used as the purity estimate and the uncertainty was estimated to cover the interval over all the method results. The pure BTX components were analyzed for purity by Gas Chromatography (GC) with flame ionization detection, GC with mass selective detection(gc/msd) and Differential Scanning Calorimetry. The pure benzene and toluene were determined to be a minimum 99.99 % pure with a standard uncertainty of 0.06 %, and did not contain detectable levels of BTX impurities (< 0.01 %). The o-xylene was found to be a minimum 99.63 % pure with a standard uncertainty of 0.06%, and did not contained detectable levels of BTX impurities (< 0.01 %). The o-xylene did contain trace amounts of m-xylene and p-xylene, however these should not interfere in the analysis of these BTX samples. The nitrogen dilution gas was labeled as 99.999 % pure with a maximum of 2 µmol/mol water and 5 µmol/mol oxygen. The nitrogen and room air were analyzed and found not to contain BTX to the detection limit of the method (0.005 nmol/mol). The standard uncertainty was estimated as the detection limit value. The final combined uncertainty was expanded using a k factor of 2. All the final expanded uncertainties originating from the gravimetric preparation were less than 0.5% relative to the concentration. These uncertainties do not take into account the uncertainties of the cylinder treatment, potential problems due to transport and storage temperatures, and potential problems with the external connections made by the receiving laboratory. Therefore it is NIST s standard procedure to claim a 1 % relative expanded uncertainty for gravimetrically prepared organic standards in this Page 3 of 21
concentration range. The gravimetric amount-of-substance fraction and expanded uncertainty for the BTX are listed in Table 1. Calibration Standards of Participants KRISS: Two µmol/mol gravimetric standards were prepared in Luxfer Australia cylinders with fine-polished internal surface with stainless steel valves. A syringe micro gravimetric method was used for the preparation of the µmol/mol standards. The nmol/mol standards were prepared by static dilution of the high concentration gas mixture into Scott Specialty Gases Netherlands cylinders, which were treated with the ACLIFE process. One primary reference gas was selected and used in a one-point calibration of the instrument. A one-point calibration was used due to the very close proximity of the concentration of the gas standard to the comparison sample. LNE: Gravimetric standards were prepared in Messer cylinders. The cylinder treatment was unspecified by Messer, however a surface treatment was applied to all cylinders. The cylinders were heated overnight under vacuum by LNE and then used. A 2 µmol/mol standard was prepared by static gravimetric method using minimum 99.95 % pure liquids. Dynamic dilution of this standard was used to generate a standard gas at the level of the comparison sample. NIST: Gravimetric standards were prepared in Scott Specialty Gases SA cylinders, which were treated by the ACLIFE IV process. The cylinders were further process at NIST by repeated pressurization with pure nitrogen, and evacuation. Standards were prepared by a micro capillary technique at the µmol/mol concentration and verified against previously prepared standards. These standards were then gravimetrically diluted with nitrogen gas to the 4 nmol/mol to 10 nmol/mol concentration range. Six standards were used to calibrate the instrument and to value assign the comparison sample. NMi: Two independently prepared gas mixtures were prepared by injection of a weighed liquid BTX mixture into a stainless steel transfer line, heated to 70º C and brought into the evacuated cylinder by means of a nitrogen gas stream. Six primary standards were prepared by dilution of the parent gas standards into 5 L cylinders from Scott Specialty Gases. These cylinders were treated by the ACLIFE VI process. The standards were then verified for internal consistency and with other primary standards maintained at NMi. During this verification, three of the standards were found not to be consistent, and were eliminated from the calibration. The remaining three standards were used to value assign the comparison cylinder. An additional standard was obtained by serial dilution of a primary standard used in the CCQM-K7 comparison. A fifth standard was independently prepared from a new liquid mixture. NMIJ: All cylinders used were obtained from Takachiho Chemical Industrial Co Ltd., and were 9.5 L Australian Luxfer aluminum-alloy cylinders with Super Clean process. Before use, the cylinders were baked at high vacuum, and then filled with a BTX in N 2 standard (10 nmol/mol) at 0.5 MPa and evacuated. Five binary mixtures of benzene, toluene, o-xylene, m-xylene and ethylbenzene at a concentration of 0.1 mmol/mol, were prepared by a vapor phase transfer method from small stainless vessels to the evacuated aluminum cylinders. Then an intermediate 5-component mixture at a concentration of approximately 100 nmol/mol was prepared by a 2-stage dilution of the parent gases into an aluminum cylinder. A 10 nmol/mol calibration standard was Page 4 of 21
prepared by dilution of the intermediate gas. The 10 nmol/mol calibration standard gas was diluted into aluminum cylinders to prepare two additional calibration gases with concentrations of 2 nmol/mol and 5 nmol/mol, respectively. These three standards were used to bracket the concentration of the compounds in the comparison cylinder. BTXfree nitrogen was used for dilution throughout. Concentrations of the calibration gases are gravimetric values, which were corrected for purity of the BTX liquids and impurities in the diluting gas. ncertainties of the calibration gases included uncertainties due to instability over a 3-month storage period. In addition, uncertainty factors associated with weighing, purity measurements of the BTX liquids, and impurities in the dilution gas were included in the uncertainty budget. NPL: Gravimetric standards were prepared in cylinders obtained from Air Products and treated using their proprietary Quantum process. The standard were prepared by gravimetric techniques using a three-stage dilution of pure hydrocarbons with a purity of 99.9 %. The matrix gas used was nitrogen with a stated purity of 99.9999% supplied by Air Liquide. The total detectable hydrocarbons found in the matrix gas were less than 0.1 nmol/mol. Each of the standards were validated against 2 other mixtures at different concentration levels. Two BTX gravimetric standards were used to bracket the comparison sample. A linear interpolation method was used to assign the concentration of the components in the mixture. BA: Calibration standards were prepared by volumetric static injection followed by dynamic dilution. Known volumes of the liquid compounds are added to the complementary gas, using a µl syringe, in a vessel of well defined volume and evaporated (as described in ISO 6144). The vessel is filled with synthetic air at ambient pressure and temperature resulting in an approximately 4 mg/m 3 concentration. The vessel is raised in pressure to approximately 150 kpa by adding synthetic air. This standard is further diluted through mass flow controllers using zero air to the concentration of interest. The analyzer is calibrated with a three-point external standard, bracketing the concentration of the compounds in the comparison cylinder. VNIIM: All cylinders were 5 L aluminum obtained from Scott Specialty Gases and were treated with the ACLIFE IV process. Two Gravimetric cylinder standards were prepared by injection of a BTX liquid mixture using a sample loop valve system with addition of nitrogen by gravimetry. Both primary standards were further diluted in a twostage process to approximately 5 nmol/mol. The BTX concentration was determined by direct comparison to these two working standards. BTX purity was determined using GC/MSD and nitrogen purity was assessed for BTX, water and other components. Results The distributed protocol instructed the participants to measure the concentration of the BTX in the cylinder with respect to their primary standards of BTX a minimum of three times. The participating laboratories are listed in Table 2 along with the method used to analyze the BTX sample, and the date of the submitted report. In Tables 3 through 5, the participating laboratories reported concentrations and uncertainties are given along with the calculated relative difference from the gravimetric concentration. In Figures 1 through 3 the results are plotted in terms of the deviation in nmol/mol from the gravimetric value. The gravimetric value is represented by zero deviation. The uncertainty bars represent the expanded uncertainties at a level of confidence of 95 %, Page 5 of 21
reported by the participant. All the participants had results that were within 5 % of the relevant reference values except for one value for o-xylene, which was 10 % off the reference value. At this concentration level, these results are to be considered well with acceptable limits. Equivalence Statements Key comparison reference value: there is no single reference value for this comparison, the value x igrav is taken as the reference value for laboratory i. The degree of equivalence of each laboratory with respect to the reference value is given by a pair of terms; 2 2 2 i igrav D i = (x i x igrav ), the degree of equivalence, and = u + u, its expanded uncertainty, both expressed in nmol/mol. The degree of equivalence between two laboratories may be computed using D = D D, and its uncertainty; ij i j i ij = + 2 2 i j 2 u ( Di ) + u j ( D j ) = 2 + = i j where the factors 2 are coverage factors. In this particular case, the direct manipulation of the expanded uncertainties leads numerically to the same result as the path over the standard uncertainties. It is important to note that it is impossible to directly manipulate expanded uncertainties as if they were standard uncertainties. Limit of Claims This comparison could be used to assess calibration and measurement capability (CMC) claims for non-polar volatile organic hydrocarbons, which are liquid at room temperature, with boiling point temperatures between 80º C and 150º C, and at a concentration of 1 nmol/mol to 100 µmol/mol. A similar comparison is expected to be held within 5 years. References 1. G. C. Rhoderick and W. L. Zielinski, Jr., Preparation of Accurate Multicomponent Gas Standards of Volatile Toxic Organic Compounds in the Low-Parts-per-Billion Range, Analytical Chemistry, 7(11), 1988. Coordinator Franklin R. Guenther and George C. Rhoderick National Institute of Standards and Technology (NIST) Gaithersburg, Maryland 20899-8393 SA Page 6 of 21
Project Reference: CCQM-K10 Completion Date: November 2001 Participant Contact List Dr. Alain Marschal Bureau National de Métrologie, (LNE) 1 rue Gaston Boissier 75724 PARIS CEDEX 15 France Mrs. Anneliese Medem Federal Environmental Agency (BA) Paul Ehrlich Strasse 29 DE-63225 LANGEN Germany Dr. Kenji Kato National Metrology Institute of Japan (NMIJ) 1-1, Hagashi Tsukuba Ibaraki 305 Japan Prof. Hun-Young So Korea Research Institute of Standards and Science (KRISS) Division of Chemistry and Radiation P.O. Box 102 Yusung Taejon, 305-600 Korea Dr. Ed de Leer Nederlands Meetinstituut (NMi) Schoemakerstraat 97 Postbus 654 2600 AR DELFT The Netherlands Dr. Leonid Konopelko D.I. Mendeleyev Institute for Metrology (VNIIM) Page 7 of 21
19, Moskovsky Prospekt 198005 St. Petersburg Russia Dr. Martin J.T. Milton National Physical Laboratory (NPL) Environmental Standards Section Teddington Middlesex TW11 0LW England Dr. Franklin R. Guenther National Institute of Standards and Technology (NIST) Chemical Science and Technology Laboratory 100 Bureau Drive Gaithersburg, MD 20899-8393 SA Page 8 of 21
Table 1: Concentration of gravimetric primary standards Compound Amount Expanded % Relative fraction ncertainty ncertainty (nmol/mol) (nmol/mol) Primary Standard A Benzene 4.64 0.05 1.00% Toluene 8.12 0.08 1.00% o-xylene 6.27 0.06 1.00% Primary Standard B Benzene 4.61 0.05 1.00% Toluene 7.25 0.07 1.00% o-xylene 6.81 0.07 1.00% Table 2: Participating Laboratories methods Laboratory Analytical Sampled Date of method Volume Analysis (ml) France (LNE) GC/FID with pre-concentration on Carbotrap and Carbosieve, 25m x 0.32mm PLOT Alumina/KCl. One PSM by gravimetry followed by dynamic dilution to working level. 300 Jul-01 Germany (BA) GC/FID with cryotrap onto Carbosieve SIII, Carbotrap, Carbotrap C. 25m x 0.53mm PLOT Alumina/KCl. Calibration by volumetric static injection followed by dynamic dilution. 220/440 Jul-01 Japan (NMIJ) GC/FID with pre-concentration Tenax trap. Three PSM by gravimetry. 450 Jun-01 Korea (KRISS) Netherlands (NMi) Russia (VNIIM) nited Kingdom (NPL) SA (NIST) GC/FID with cryogenic concentrator with Tenax trap. 60m x 0.32mm DB-1 1.0 µm phase. Two PSM by gravimetry. 200 Aug-01 GC/FID with pre-contentration Tenax-TA trap, 50m x 0.32mm PLOT Alumina/KCl. Six PSM by ISO 6142 400 Aug-01 GC/FID with Tenax trap, 20m x 0.53mm SE-30 3.0 µm phase. One PSM be gravimetry 500 Jul-01 GC/FID with pre-concentration Tenax-TA trap, 50m x 0.53 megabore PLOT Alumina/KCl activated. PSM by gravimetry, bracketing method 100 Jul-01 GC/FID with cryoigenic concentrator, 60m x 0.53mm Supelcowax 10. Four PSM by gravimetry. 400 Jun-01 Page 9 of 21
Table 3: Results of CCQM K10: Comparison of Benzene Laboratory Submitted K Submitted Gravimetric Gravimetric % Relative Analytical Factor Expanded Amount Expanded Difference Result ncertainty Fraction ncertrainty (nmol/mol) (nmol/mol) (nmol/mol) (nmol/mol) France (LNE) 4.654 2 0.058 4.64 0.05 0.3% Germany (BA) 4.63 2 0.20 4.64 0.05-0.2% Japan (NMIJ) 4.624 2 0.817 4.61 0.05 0.2% Korea (KRISS) 4.63 2 0.25 4.61 0.05 0.4% Netherlands (NMi) 4.57 2 0.22 4.64 0.05-1.5% Russia (VNIIM) 4.76 2 0.14 4.61 0.05 3.2% nited Kingdom (NPL) 4.51 2 0.06 4.64 0.05-2.8% SA (NIST) 4.64 2 0.05 4.64 0.05 0.0% Figure 1. CCQM-K10 Comparison of benzene in nitrogen standards 1.0 0.8 Absolute Deviation from Reference Value (nmol/mol) 0.6 0.4 0.2 0.0-0.2-0.4-0.6-0.8 LNE BA NMIJ KRISS NMi VNIIM NPL NIST -1.0 Page 10 of 21
Table 4: Results of CCQM K10: Comparison of Toluene Laboratory Submitted k Submitted Gravimetric Gravimetric % Relative Analytical Factor Expanded Amount Expanded Difference Result ncertainty Fraction ncertrainty (nmol/mol) (nmol/mol) (nmol/mol) (nmol/mol) France (LNE) 8.154 2 0.100 8.12 0.08 0.4% Germany (BA) 7.90 2 0.25 8.12 0.08-2.8% Japan (NMIJ) 7.348 2 0.642 7.25 0.07 1.3% Korea (KRISS) 7.17 2 0.38 7.25 0.07-1.1% Netherlands (NMi) 8.2 0.48 8.12 0.08 1.2% Russia (VNIIM) 7.56 2 0.23 7.25 0.07 4.3% nited Kingdom (NPL) 7.99 2 0.10 8.12 0.08-1.7% SA (NIST) 8.08 2 0.05 8.12 0.08-0.5% Figure 2. CCQM-K10 Comparison of toluene in nitrogen standards 1.0 0.8 Absolute Deviation from Reference Value (nmol/mol) 0.6 0.4 0.2 0.0-0.2-0.4-0.6-0.8 LNE BA NMIJ KRISS NMi VNIIM NPL NIST -1.0 Page 11 of 21
Table 5: Results of CCQM K10: Comparison of o-xylene Laboratory Submitted k Submitted Gravimetric Gravimetric % Relative Analytical Factor Expanded Amount Expanded Difference Result ncertainty Fraction ncertrainty (nmol/mol) (nmol/mol) (nmol/mol) (nmol/mol) France (LNE) 6.198 2 0.076 6.27 0.06-1.2% Germany (BA) 6.30 2 0.27 6.27 0.06 0.4% Japan (NMIJ) 6.994 2 0.647 6.81 0.07 2.6% Korea (KRISS) 6.81 2 0.35 6.81 0.07-0.1% Netherlands (NMi) 6.57 2 0.52 6.27 0.06 4.7% Russia (VNIIM) 7.50 2 0.22 6.81 0.07 10.1% nited Kingdom (NPL) 5.9 0.08 6.27 0.06-5.6% SA (NIST) 6.27 2 0.11 6.27 0.06-0.1% 1.0 Figure 3. CCQM-K10 Comparison of o-xylene in nitrogen standards 0.8 Absolute Deviation from Reference Value (nmol/mol) 0.6 0.4 0.2 0.0-0.2-0.4-0.6-0.8 LNE BA NMIJ KRISS NMi VNIIM NPL NIST -1.0 Page 12 of 21
Annex 1 Entry in Appendix B Measurand: Benzene x i : result of measurement carried out by laboratory i u i x igrav : u igrav : combined standard uncertainty of x i gravimetric value of the benzene amount of substance fraction in cylinder received by laboratory i combined standard uncertainty of x igrav Lab i x i u i x igrav u igrav Date of nmol/mol nmol/mol nmol/mol nmol/mol measurement LNE 4.65 0.03 4.64 0.02 Jul-01 BA 4.63 0.10 4.64 0.02 Jul-01 NMIJ 4.62 0.41 4.61 0.02 Jun-01 KRISS 4.63 0.13 4.61 0.02 Aug-01 NMi 4.57 0.11 4.64 0.02 Aug-01 VNIIM 4.76 0.07 4.61 0.02 Jul-01 NPL 4.51 0.03 4.64 0.02 Jul-01 NIST 4.64 0.03 4.64 0.02 Jun-01 Page 13 of 21
Matrix of Equivalence Measurand: Amount-of-substance fraction of benzene in nitrogen Nominal Value: 5 nmol/mol Lab i D i i The degree of equivalence of each laboratory with respect to the reference value is given nmol/mol by a pair of terms; D i = (x i x igrav ) and i, its expanded uncertainty (k= 2 for all laboratories), LNE 0.02 0.07 BA -0.01 0.21 The degree of equivalence between any two laboratories may be calculated by; NMIJ 0.01 0.82 D ij = D i D j and its expanded uncertainty; KRISS 0.02 0.25 NMi -0.07 0.22 VNIIM 0.15 0.15 NPL -0.13 0.08 ij = + 2 2 i j 2 u ( Di ) + u ( D j ) = 2 + = i j NIST 0.00 0.07 Note: se of this equation is allowed since k=2 for all laboratories, see report for explanation. Page 14 of 21
CCQM-K10 Degrees of equivalence for benzene in nitrogen at nominal 5 nmol/mol 1.00 0.80 0.60 0.40 Di /(nmol/mol) 0.20 0.00-0.20 LNE BA NMIJ KRISS NMi VNIIM NPL NIST -0.40-0.60-0.80-1.00 Page 15 of 21
Measurand: Toluene x i : result of measurement carried out by laboratory i u i x igrav : u igrav : combined standard uncertainty of x i gravimetric value of the toluene amount of substance fraction in cylinder received by laboratory i combined standard uncertainty of x igrav Lab i x i u i x igrav u igrav Date of nmol/mol nmol/mol nmol/mol nmol/mol measurement LNE 8.15 0.05 8.12 0.04 Jul-01 BA 7.90 0.13 8.12 0.04 Jul-01 NMIJ 7.35 0.32 7.25 0.04 Jun-01 KRISS 7.17 0.19 7.25 0.04 Aug-01 NMi 8.22 0.24 8.12 0.04 Aug-01 VNIIM 7.56 0.12 7.25 0.04 Jul-01 NPL 7.99 0.05 8.12 0.04 Jul-01 NIST 8.08 0.03 8.12 0.04 Jun-01 Page 16 of 21
Matrix of Equivalence Measurand: Amount-of-substance fraction of toluene in nitrogen Nominal Value: 8 nmol/mol Lab i D i i The degree of equivalence of each laboratory with respect to the reference value is given nmol/mol by a pair of terms; D i = (x i x igrav ) and i, its expanded uncertainty (k= 2 for all laboratories), LNE 0.03 0.13 BA -0.22 0.26 The degree of equivalence between any two laboratories may be calculated by; NMIJ 0.10 0.65 D ij = D i D j and its expanded uncertainty; KRISS -0.08 0.39 NMi 0.10 0.49 VNIIM 0.31 0.24 NPL -0.13 0.13 ij = + 2 2 i j 2 u ( Di ) + u ( D j ) = 2 + = i j NIST -0.04 0.10 Note: se of this equation is allowed since k=2 for all laboratories, see report for explanation. Page 17 of 21
CCQM-K10 Degrees of equivalence for toluene in nitrogen at nominal 8 nmol/mol 1.00 0.80 0.60 0.40 VNIIM Di /(nmol/mol) 0.20 0.00-0.20 LNE BA NMIJ KRISS NMi NPL NIST -0.40-0.60-0.80-1.00 Page 18 of 21
Measurand: o-xylene x i : result of measurement carried out by laboratory i u i x igrav : u igrav : combined standard uncertainty of x i gravimetric value of the o-xylene amount of substance fraction in cylinder received by laboratory i combined standard uncertainty of x igrav Lab i x i u i x igrav u igrav Date of nmol/mol nmol/mol nmol/mol nmol/mol measurement LNE 6.20 0.04 6.27 0.03 Jul-01 BA 6.30 0.14 6.27 0.03 Jul-01 NMIJ 6.99 0.32 6.81 0.03 Jun-01 KRISS 6.81 0.18 6.81 0.03 Aug-01 NMi 6.57 0.26 6.27 0.03 Aug-01 VNIIM 7.50 0.11 6.81 0.03 Jul-01 NPL 5.92 0.04 6.27 0.03 Jul-01 NIST 6.27 0.06 6.27 0.03 Jun-01 Page 19 of 21
Matrix of Equivalence Measurand: Amount-of-substance fraction of o-xylene in nitrogen Nominal Value: 6 nmol/mol Lab i D i i The degree of equivalence of each laboratory with respect to the reference value is given nmol/mol by a pair of terms; D i = (x i x igrav ) and i, its expanded uncertainty (k= 2 for all laboratories), LNE -0.08 0.10 BA 0.03 0.28 The degree of equivalence between any two laboratories may be calculated by; NMIJ 0.18 0.65 D ij = D i D j and its expanded uncertainty; KRISS 0.00 0.36 NMi 0.30 0.52 VNIIM 0.69 0.23 NPL -0.35 0.10 ij = + 2 2 i j 2 u ( Di ) + u ( D j ) = 2 + = i j NIST 0.00 0.13 Note: se of this equation is allowed since k=2 for all laboratories, see report for explanation. Page 20 of 21
CCQM-K10 Degrees of equivalence for o-xylene in nitrogen at nominal 6 nmol/mol 1.00 0.80 0.60 VNIIM 0.40 NMi Di /(nmol/mol) 0.20 0.00-0.20 LNE BA NMIJ KRISS NIST -0.40 NPL -0.60-0.80-1.00 Page 21 of 21