ANALYSIS OF PROTOCOL GASES An On-Going Quality Assurance Audit (Results Through March 1995) Avis P. Hines Quality Assurance Support Branch Quality Assurance and Technical Support Division Atmospheric Research and Exposure Assessment Laboratory U.S. Environmental Protection Agency Research Triangle Park, NC 27711 1.0 INTRODUCTION The Atmospheric Research and Exposure Assessment Laboratory (AREAL) of the U.S. Environmental Protection Agency (EPA) operates a nationwide audit on the vendors of Protocol Gas Standards. The intent of this program is as follows: 1. Increase the acceptance and use of Protocol Gases as secondary standards by the air monitoring community. 2. Provide a quality assurance check for the vendors of these gases. 3. Assist users of Protocol Gases to identify vendors who can consistently provide accurately certified Protocol Gases. For the first audits of nitric oxide, carbon monoxide, and sulfur dioxide, gas cylinders obtained through third parties were analyzed in triplicate by EPA and by a non-EPA laboratory. Each laboratory used its own SRMs and followed the Protocol Gas certification procedure. Because a statistical analysis showed that the results from EPA and the independent laboratory were indistinguishable, EPA is now the primary auditing laboratory. The other laboratory serves as the referee laboratory to resolve differences between EPA and the vendors. 2.0 PROCEDURE Either directly or through third parties, EPA procures Protocol Gases from commercial sources, checks the accuracy of the vendors' certification of concentration, and examines the accompanying documentation for completeness and accuracy. The vendors are not aware that EPA is obtaining the gases for a check on the completeness of the documentation and accuracy of the certification of concentration. Protocol Gases have a maximum allowable deviation of 2% from the certified value. Accuracy of the certification is checked using Standard Reference Materials (SRMs). If the difference between the EPA-determined and the vendor-determined concentration is more than 2%, or if the documentation is incomplete, EPA notifies the vendor immediately to resolve and correct the problem. Results of EPA certification checks are placed on two bulletin boards, EMTIC (Emission Measurement Technology Information Center) and AMTIC (Ambient Monitoring Technology Information Center), of the Technology Transfer Network of the EPA Office of Air Quality Planning and Standards. Bulletin board entries are organized in tables by gas mixture type and by vendor. Numerical data are supplemented by narrative footnotes explaining the results of any corrective action taken by the vendor. Thus the entries provide a continuous record of all audit activities. Users who believe that their Protocol Gas has been certified incorrectly are encouraged to contact Ms. Avis Hines of AREAL (919-541-4001) to request an EPA certification check. If EPA accepts the gas cylinder for testing, the results of these tests will also be posted on the bulletin boards. 3.0 REQUIRED DOCUMENTATION The Protocol Gas procedure specifies two types of documentation that must accompany the gas cylinder: a Certificate of Analysis, which may be mailed separately or attached to the cylinder; and a cylinder tag which must be attached to the valve under the valve cap. Documentation is incomplete until the vendor provides every item shown in Tables A and B for the certificate and the tag, respectively. These tables reflect the requirements specified in the revised Protocol Gas guidance issued by this Laboratory in September 1993 (U.S. EPA Traceability Protocol for Assay and Certification of Gaseous Calibration Standards (Revised September 1993), EPA 600/R03/224). Table A. REQUIRED DOCUMENTATION FOR A CERTIFICATE OF ANALYSIS Cylinder ID number Reference standard data Certified concentration Protocol statement of analyte Balance gas Analytical method used in assay Cylinder pressure All analyzer readings Certificate date Calculations to three significant figures Expiration date Name and signature of analyst and laboratory name Certification period (months) TABLE B. REQUIRED DOCUMENTATION FOR A CYLINDER TAG Cylinder ID number Certification date Expiration date Protocol statement Laboratory ID Balance gas Cylinder pressure Analyte Concentration 4.0 RESULTS This section of the audit report, organized by gas mixtyre type and by vendor, is updated whenever EPA conducts a new audit or receives corrective action reports from a vendor. It allows users of Protocol Gases to easily review the comparative performances of the vendors. The standard of comparison used in reporting the certification of concentration check results is the relative percent difference between the vendor and the EPA values. Prior to this latest update only single component gases had been checked. With the inclusion of multi-component gases the reporting format used previously became unwieldy. Therefore, the tables were simplified to ensure they still provided information in a user-friendly form. This was accomplished by: (1) removing the vendor and the EPA concentration values from the tables and reporting the nominal value to the nearest 10 ppm; and (2) removing the cost information. (Since the prices of Protocol Gases are subject to change and differ markedly between vendors, users of these gases should obtain them through a competitive procurement procedure whenever possible.) The tables summarize audit results for each gas mixture type and include footnotes which describe corrective actions taken by the vendors. In each table each vendor has been assigned an acronym and a footnote letter. If a vendor has more than one plant, each plant is assigned its own acronym and footnote. Notes may not be necessary for every vendor on every audit. The following acronyms have been assigned to each vendor: ACRONYM VENDOR AGA, OH AGA Gas, Inc. Maumee, OH APC-NC Air Prod. and Chem Durham, NC APC-IL Air Prod. and Chem Chicago, IL AIG-MI Airco Industrial Gases Royal Oak, MI AIG-NC Airco Industrial Gases Research Triangle Park, NC AIG-NJ Airco Industrial Gases Riverton, NJ ASG-PA Alphagaz Spec. Gas Div. Morrisville, PA ASG-CA Alphagaz Spec. Gas Div. Long Beach, CA MGP-OH Matheson Gas Products Twinsburg, OH LCC Liquid Carbonic Cylinder Gas Products Bethlehem, PA MGP-NJ Matheson Gas Products East Rutherford, NJ MGI-PA MG Industries Gas Prod. Valley Forge, PA NSG-NC National Specialty Gases Durham, NC SMG-CA Scott Marrin Gases Riverside, CA SSG-NC Scott Specialty Gases Durham, NC SSG-NJ Scott Specialty Gases Plainille, NJ SSG-PA Scott Specialty Gases Plumsteadville, PA When using the data in the following tables for procuring Protocol Gases readers should bear in mind the following points. First, the information in the footnotes may be important and should not be ignored. Second, if the difference between EPA's value and a manufacturer's value differs by 2% or less, then (because of the uncertainties in the total measurement system ) statistically there is no difference between the two values. Thus, a difference of 2.0% is the same as one of 0.57%. Third, EPA has not assigned a rating to the vendors concerning who is the best, who is approved, who is not approved, etc. The information is presented without making such judgments. TABLE I. NO Protocol Gas Results Nominal ppm Date Date % Difference Complete Vendor NO Certified Checked (Vendor-EPA) Doc. Notes APC-NC 40 4-92 7-92 2.0 Yes AIG-MI 40 5-92 7-92 -0.5 No a ASG-PA 40 4-92 7-92 -0.7 No b MGP-NJ 40 4-92 7-92 0.7 No c MGI-PA 40 4-92 7-92 0.2 No d NSG-NC 40 4-92 7-92 -1.7 Yes SMG-CA 40 4-92 7-92 0.0 Yes SSG-NC 40 4-92 7-92 -0.7 No e a. Documentation problems corrected. b. Documentation problems corrected. c. Documentation problems corrected. d. Documentation problems corrected. e. Documentation problems corrected. TABLE II. SO2 Protocol Gas Results Nominal ppm Date Date % Difference Complete Vendor SO2 Certified Checked (Vendor-EPA) Doc. Notes APC-NC 50 9-92 12-92 -3.8 No a AIG-NJ 50 9-92 12-92 -0.7 No b AIG-NC 50 6-93 7-93 1.3 Yes AGA-OH 20 7-92 7-93 0.8 Yes ASG-PA 50 8-92 12-92 1.1 No c MGP-OH 50 9-92 12-92 -3.9 No d MGP-NJ 50 5-93 5-93 3.1 Yes i MGI-PA 50 12-92 2-93 -5.5 No e MGI-PA 99 8-94 2-95 -0.8 Yes NSG-NC 50 8-92 12-92 -3.6 Yes f SMG-CA 50 8-92 12-92 -0.2 No g SMG-CA 100 10-92 9-93 1.1 Yes SSG-NC 50 7-92 12-92 0.6 Yes SSG-NJ 50 7-93 5-94 0.4 No h a. Vendor replaced calibration standard and reported (3/93) new value of 50.2 ppm (-1% difference). Documentation problems corrected. b. Prior to receiving EPA results, vendor notified all recent purchasers that Airco's calibration standard was defective and requested gas be returned for re-analysis at no cost. c. Documentation problems corrected. d. Vendor replaced calibration standard and reported (2/93) new value of 56.8 ppm (0.7% difference). Documentation problems corrected. e. Vendor found error on part of analyst (3/93). Analyst given additional training. Documentation problems corrected. f. Vendor replaced calibration standard. No re-analysis done Documentation problems corrected. g. Documentation problems corrected. h. Documentation problems corrected. i. Vendor replaced pulsed fluorescence detector with a NDIR detector and reported a new value of 52.6 ppm (1.3% difference). TABLE III. CO Protocol Gas Results Nominal ppm Date Date % Difference Complete Vendor CO Certified Checked (Vendor-EPA) Doc. Notes APC-NC 40 1-93 4-93 0.3 No a AIG-NJ 40 1-93 4-93 -0.3 No APC-IL 9 12-94 2-95 -0.8 Yes ASG-PA 40 1-93 4-93 -0.7 Yes MGP-OH 40 1-93 4-93 0.3 No b MGP-NJ 40 5-93 6-93 -0.6 Yes MGI-PA 40 1-93 4-93 -1.2 No c NSG-NC 40 12-92 4-93 1.0 No d SMG-CA 40 1-93 4-93 -0.5 Yes SSG-NC 40 1-93 4-93 0.0 Yes a. Certificate missing; sent when EPA requested but did not agree with cylinder tag; second certificate matched cylinder tag but calibration standard identified as having gone out of calibration in January 1992; sent third certificate with corrected calibration date. b. Certificate sent only after EPA requested it. c. Documentation problems corrected. d. Certified only for six months rather than the 24 months allowed. TABLE IV. Mixture NO/SO2 Protocol Gas Results Note: The first group of multi-blend Protocol Gases was checked November 1993 through May 1994. The results are listed in Table IV. Initially, the results for five of these mixtures were disappointing, but in four of these five cases incorrectly certified SRM's caused the inaccurate certifications. Once the vendors had obtained the correct certified value for their SRM's, their revised SO2 concentrations differed by less than 1% from the EPA-determined value. All vendors supplied complete documentation. % Difference Nominal ppm Date Date (Vendor-EPA) Complete Vendor NO SO2 Certified Checked NO SO2 Doc. Notes AGA-OH 400 300 11-93 12-93 1.5 -1.7 Yes 900 1500 11-93 12-93 2.0 0.7 Yes AIG-NJ 400 300 9-93 11-93 1.6 0.3 Yes 900 1500 9-93 11-93 0.5 2.4 Yes a ASG-CA 400 300 9-93 11-93 3.2 -2.7 Yes 900 1500 9-93 11-93 -0.6 1.4 Yes b LCC-PA 400 300 8-93 11-93 0.5 2.5 Yes 900 1500 8-93 11-93 -0.1 -0.5 Yes c MGP-OH 400 300 10-93 12-93 -1.7 1.7 Yes 900 1500 10-93 12-93 0.4 -1.4 Yes MGI-PA 400 300 9-93 11-93 0.5 -0.6 Yes 900 1500 10-93 11-93 0.7 -0.8 Yes NSG-NC 400 300 9-93 11-93 0.0 -4.3 Yes 900 1500 9-93 11-93 0.3 -1.9 Yes d SMG-CA 400 300 11-93 12-93 1.5 -1.4 Yes 900 1500 11-93 12-93 0.6 -0.7 Yes SSG-NC 400 300 9-93 11-93 -1.2 -3.2 Yes 900 1500 9-93 11-93 -0.1 -0.7 Yes e a. Manufacturer used corrected value for NIST SRM, and reported (1-94) a new value of 1468 ppm SO2 (0.1% difference). b. Manufacturer reported a new NO value (4-94) of 401 ppm NO (-2.4% difference), and no change for the SO2 value. c. Manufacturer used corrected value for NIST SRM, reanalyzed the gas and reported (2-94) a new value of 311.4 ppm SO2 (-1.3% difference). d. Manufacturer purchased new analyzer and reanalyzed cylinder, reported a new value (5-94) 302 ppm SO2 (0.7% difference). e. Manufacturer used corrected value for NIST SRM, reanalyzed the gas and reported (1-94) a new value of 302.9 ppm SO2 (-1.7% difference). TABLE V. Mixture NO/CO2 Protocol Gas Results % Difference Nominal ppm Date Date (Vendor-EPA) Complete Vendor NO CO2 Certified Checked NO CO2 Doc. Notes AIG-NJ 910 18% 7-93 5-94 -2.9 --- Yes a AIG-NJ 540 18% 8-93 5-94 -2.2 --- Yes b MGI-PA 880 18% ? ? 5-94 1.4 --- No c MGI-PA 900 18% ? ? 5-94 1.8 --- No d a. Client owned cylinder, has not been returned to vendor for recertification. CO2 concentration well within 2%; exact value not determined. b. Client owned cylinder, has not been returned to vendor for recertification. CO2 concentration well within 2%; exact value not determined. c. CO2 concentration well within 2%; exact value not determined. d. CO2 concentration well within 2%; exact value not determined. TABLE VI. Mixture NO/SO2/CO2 Protocol Gas Results % Difference Nominal ppm Date Date (Vendor-EPA) Comp. Vendor NO SO2 CO2 Cert. Checked NO SO2 CO2 Doc. Notes APC-IL 550 500 18% 4-94 5-94 -0.9 -0.2 --- Yes a SSG-PA 570 530 18% 4-94 5-94 0.4 -0.2 --- Yes b a. CO2 concentration well within 2%; exact value not determined. b. CO2 concentration well within 2%; exact value not determined.