WPC4 2 ZB0JXvPXPX0Í ÍX0Í ÍҫvPXP*`XaMPP)`_PP)`vPXP*`X2*2"|w   X B#aMPP#Date of Release: 2/13/96; second 15day changes Board Hearing: 9/28/95Պ  California Environmental Protection Agency r AIR RESOURCES BOARD % PROPOSED ă CALIFORNIA NON-METHANE ORGANIC GAS d TEST PROCEDURES ă D"Adopted: July 12, 1991  Amended: September 22, 1993 L Amended: _______________ Monitoring and Laboratory Division, Southern Laboratory Branch >"Mobile Source Division #9528 Telstar Avenue !El Monte, California 91731 NOTE:` ` ` Mention of any trade name or commercial product does not constitute endorsement or recommendation of this product by the Air Resources Board. ` ` ` The regulatory amendments proposed in this rulemaking are shown in underline to indicate additions and strikeout to indicate deletions from the version of the test procedures adopted on September 22, 1993. Modifications to the originally noticed text (released October 20, 1995) are designated by bold italics and  bold strikeout  to represent additions and deletions, respectively. Modifications to the modified text are designated by  underlined bold italics  and  underlined bold italic strikeout , to indicate additions and deletions, respectively.    ! TABLE OF CONTENTS ă A.General Applicability and Requirements`!(#HA1 B.Determination of NonMethane Hydrocarbon Mass Emissions by Flame Ionization Detection`!(#HB1 C.Method 1001: Determination of Alcohols in Automotive Source Samples by Gas Chromatography`!(#HC1 D.Method 1002: Determination of C2 to C5  Hydrocarbons in Automotive Source Samples by Gas Chromatography`!(#HD1 E.Method 1003: Determination of C6 to C12 Hydrocarbons in Automotive Source Samples by Gas Chromatography`"(#HE1 F.Method 1004: Determination of Aldehyde and Ketone Compounds in Automotive Source Samples by High Performance Liquid Chromatography`"(#HF1 G.Determination of NonMethane Organic Gas Mass Emissions`!(#HG1 r% APPENDICES ă Appendix 1` `  List of LightEnd and MidRange Hydrocarbons`"(#H11 Appendix 2` `  Definitions and Commonly Used Abbreviations`"(#H21 Appendix 3` `  References`"(#H31ă   @A-@ ( Part A ă GENERAL APPLICABILITY AND REQUIREMENTS ă 1.These test procedures shall apply to all 1993 and subsequent model-year transitional low-emission vehicles (TLEV), low-emission vehicles (LEV), and ultra low-emission vehicles (ULEV) certifying to non-methane organic gas (NMOG) emission standards. 2.This document sets forth the analysis and calculation procedures that shall be performed to determine NMOG mass emissions. The document consists of the following parts: A.` ` General Applicability and Requirements B.` ` ` Determination of Non-Methane Hydrocarbon Mass Emissions by Flame Ionization Detection C.` ` ` Determination of Alcohols in Automotive Source Samples by Gas Chromatography (Method No. 1001) D.` ` ` Determination of C2 to C5 Hydrocarbons in Automotive Source Samples by Gas Chromatography (Method No. 1002) E.` ` ` Determination of C6 to C12 Hydrocarbons in Automotive Source Samples by Gas Chromatography (Method No. 1003) F.` ` ` Determination of Aldehyde and Ketone Compounds in Automotive Source Samples by High Performance Liquid Chromatography (Method No. 1004). G.` ` ` Determination of NMOG Mass Emissions Appendix 1 List of LightEnd and MidRange Hydrocarbons Appendix 2 Definitions and Commonly Used Abbreviations Appendix 3 References Alternative procedures may be used if shown to yield equivalent results and if approved in advance by the Executive Officer of the Air Resources Board. 3.The analyses specified in the table below shall be performed to determine mass emission rates of NMOG in grams per mile (g/mi) or milligrams per mile (mg/mi) for vehicles operated on the listed fuel: h ddx """" ddx """"h "  " FuelNMHC by FIDNMHC by GCAlcoholsCarbonyls "  "AlcoholXXX"q q "CNGXX"q q "DieselXX"q q "GasolineXX"        "LPGXX where "LPG" means "liquified petroleum gas", "CNG" means "compressed natural gas", "NMHC" means "non-methane hydrocarbon", "FID" means "flame ionization detector", and "GC" means "gas chromatograph." The specified analyses shall be performed in accordance with the following parts of this document: NMHC by FID  --Part B.Determination of Non-Methane Hydrocarbon Mass Emissions by Flame Ionization Detection NMHC by GC  --Part D. Determination of C2 to C5 Hydrocarbons in Automotive Source Samples by Gas Chromatography (Method No. 1002); and ` ` ` Part E.Determination of C6 to C12 Hydrocarbons in Automotive Source Samples by Gas Chromatography (Method No. 1003) CARBONYLS --Part F.hhhDetermination of Aldehyde and Ketone Compounds in Automotive Source Samples by High Performance Liquid Chromatography (Method No. 1004) ALCOHOLS --Part C.Determination of Alcohols in Automotive Source Samples by Gas Chromatography (Method No. 1001) 4For those manufacturers which choose to develop reactivity adjustment factors unique to a specific engine family, exhaust NMOG emissions shall be fully speciated. NMHC emissions shall be analyzed in accordance with parts D and E (Method Nos. 1002 and 1003). In addition, aldehydes and ketones, alcohols, and ethers shall be analyzed according to parts F, C, and E (Method Nos. 1004, 1001, and 1003). Analysis for alcohols shall be required only for vehicles which are operated on fuels containing alcohols. 5.For natural gas-fueled vehicles, the methane concentration in the exhaust sample shall be measured with a methane analyzer. A GC combined with a FID is used for direct measurement of methane concentrations. SAE Recommended Practice J1151 is a reference on generally accepted GC principles and analytical techniques for this application. A density of 18.89 g/ft3 shall be used to determine the methane mass emissions. The methane mass emissions shall be multiplied by the appropriate methane reactivity adjustment factor and then added to the reactivity-adjusted NMOG emissions as specified in "California Exhaust Emission Standards and Test Procedures for 1988 and Subsequent Model Passenger Cars, Light-Duty Trucks,and Medium-Duty Vehicles." 6.The mass of NMOG emissions shall be calculated in accordance with part G, "Determination of NMOG Mass Emissions". The mass of NMOG emissions in g/mile or mg/mile shall be calculated by summing the mass of NMHC determined by the FID, the mass of aldehydes and ketones, and the mass of alcohols.   @B-@ ' PART B ă !DETERMINATION OF 0 NON-METHANE HYDROCARBON MASS EMISSIONS BY FLAME IONIZATION DETECTION ă 1.INTRODUCTION 1.1This procedure describes a method for determining non-methane hydrocarbon ( NMHC ) exhaust mass emissions from motor vehicles. Other applicable forms of instrumentation and analytical techniques which prove to yield equivalent results to those specified in this procedure may be used subject to the approval of the Executive Officer of the Air Resources Board. 1.2All definitions and abbreviations are contained in Appendix 2 of these test procedures. 2.TOTAL HYDROCARBON MEASUREMENT 2.1A flame ionization detector (FID ) is used to measure total hydrocarbon concentration in vehicle exhaust in accordance with the Code of Federal Regulations.[Ref 1] SAE Recommended Practices J254[Ref. 2] and J1094a[Ref. 3] are references on generally accepted gas analysis and constant volume sampling techniques. For Beckman 400 FIDs only, implementation of the recommendations outlined in SAE paper 770141 , "Optimization of Flame Ionization Detector for Determination of Hydrocarbons in Diluted Automobile Exhaust;" author, Glenn D. Reschke [Ref. 4] shall be required. Other FID analyzer models shall be checked and adjusted, if necessary, to minimize any non-uniformity of relative response to different hydrocarbons. 3.METHANE MEASUREMENT 3.1A gas chromatograph (GC ) combined with a FID constitute a methane analyzer and shall be used for direct measurement of methane concentrations. The SAE Recommended Practice J1151[Ref. 5] is a reference on generally accepted GC principles and analytical techniques for this specific application. 4.TOTAL HC FID RESPONSE TO METHANE 4.1The FID is calibrated to propane and therefore tends to over respond to the methane portion of the vehicle exhaust sample during hydrocarbon analysis. In order to calculate the NMHC concentration, a methane response factor must be applied to the methane concentration (as measured by the methane analyzer) before it can be deducted from the total hydrocarbon concentration. To determine the total hydrocarbon FID response to methane, known methane in air concentrations traceable to National Institute of Standards and Technology ( NIST ) shall be analyzed by the FID. Several methane concentrations shall be analyzed by the FID in the range of the exhaust sample concentration. The total hydrocarbon FID response to methane is calculated as follows: ` ` ` rCH < 4  =hh#FIDppm/SAMppm where: ` ` ` rCH $ 4  =hh#FID methane response factor. ` ` ` FIDppm = FID reading in ppmC. ` ` ` SAMppm = hh#the known methane concentration in ppmC. The FID response to methane shall be checked at each calibration interval. 5.NMHC MASS EMISSION PER TEST PHASE 5.1The following calculations shall be used to determine the NMHC mass emissions for each phase of the Federal Test Procedure. [Ref. 1]. 5.2 Non-Alcohol Fueled Vehicles 5.2.1` ` NMHCe = FID THCe - (rCH < 4   * CH4e ) ` ` NOTE: If NMHCe is calculated to be less than zero, then NMHCe = 0. 5.2.2` ` NMHCd = FID THCd - (rCH < 4   * CH4d ) ` ` NOTE: If NMHCd is calculated to be less than zero, then NMHCd = 0. 5.2.3` ` ` COe = (1 - (0.01 + 0.005 * HCR) * CO2e - 0.000323 * Ra ) * COem ` ` ` NOTE: If a CO instrument which meets the criteria specified in CFR 40, 86.111 is used and the conditioning column has been deleted, COem must be substituted directly for COe. ` ` a) For gasoline, CH1.85 , where HCR = 1.85: ` ` ` COe = (1 - 0.01925 * CO2e - 0.000323 * Ra ) * COem ` ` b) For Phase 2 gasoline, CH1.94 , where HCR = 1.94: ` `  COe = (1 - 0.01970 * CO2e - 0.000323 * Ra ) * COem ` `  b c) For LPG, CH 2.66 2.64, where HCR = 2.66 2.64: ` ` ` COe = (1 - 0.02330 0.02320 * CO2e - 0.000323 * Ra) * COem ` ` d) For CNG, CH3.78, where HCR = 3.78: ` ` ` COe = (1 - 0.02890 * CO2e - 0.000323 * Ra) * COem ` `  100 *  x  ` `  hh#x + y/2 + 3.76 * (x + y/4 - z/2) 5.2.4` ` DF  =   ` `  CO2e + (NMHCe + CH4e + COe ) * 10é4  ` ` ` (where fuel composition is CxHyOz as measured for the fuel used.) ` ` ` a) For gasoline, CH1.85 , where x = 1, y = 1.85, and z = 0: ` ` ` DF = 13.47 / [CO2e + (NMHCe + CH4e + COe ) * 10é4 ] ` ` b) For Phase 2 gasoline, CH1.94 , x = 1, y = 1.94 and z = 0.017: ` `  DF = 13.29 / [CO2e + (NMHCe + CH4e + COe ) * 10é4 ]  ` ` ` c) For LPG, CH 2.66  2.64 , where x = 1, y = 2.66 2.64, and z = 0: ` ` ` DF = 11.64 11.68 / [CO2e + (NMHCe + CH4e + COe) * 10é4 ] ` ` d) For CNG, CH3.78, where x = 1, y = 3.78, and z = 0.016: ` ` ` DF = 9.83 / [CO2e + (NMHCe + CH4e + COe) * 10é4 ] 5.3 Vehicles Operating on Fuels Containing Methanol 5.3.1` ` NMHCe = FID THCe - (rCH < 4  * CH4e) - (rCH < 3 OH * CH3OHe ) ` ` ` NOTE: If NMHCe is calculated to be less than zero, then NMHCe = 0. 5.3.2` ` NMHCd = FID THCd - (rCH < 4  * CH4d ) - (rCH < 3 OH * CH3OHd ) ` ` ` NOTE: If NMHCd is calculated to be less than zero, then NMHCd = 0. 5.3.3` ` ` COe = (1 - (0.01 + 0.005 * HCR) * CO2e - 0.000323 * Ra ) * COem ` ` ` NOTE: If a CO instrument which meets the criteria specified in CFR 40 86.111 is used and the conditioning column has been deleted, COem must be substituted directly for COe . ` ` ` a) For M100 (100% methanol), CH3OH, where HCR = 4: ` `  COe = (1 - 0.03000 * CO2e - 0.000323 * Ra ) * COem ` ` b) For M85 (85% methanol, 15% indolene), CH3.41 O0.72 , where ` `  HCR = 3.41: ` `  COe = (1 - 0.02705 * CO2e - 0.000323 * Ra ) * COem ` `  100 *  x  ` `  x + y/2 + 3.76 * (x + y/4 - z/2) 5.3.4` ` DF =   ` `  CO2e + (NMHCe + CH4e + COe + CH3 OHe + HCHOe ) * 10é4 ` ` (where fuel composition is CxHyOz as measured for the fuel used.)  X  ` ` a) For M100 (100% methanol), CH3OH, where x = 1, y = 4, and z = 1: ` ` ` DF = 11.57 / [CO2e + (NMHCe+ CH4e + COe + CH3OHe + HCHOe ) * 10é4 ] ` ` b) For M85 (85% methanol, 15% Indolene), CH3.41 O0.72 , where x = 1, ` `  y = 3.41, and z = 0.72: ` ` DF = 12.02 / [CO2e + (NMHCe + CH4e + COe + CH3OHe + HCHOe) * 10é4 ] 5.4 Vehicles Operating on Fuels Containing Ethanol 5.4.1` ` NMHCe = FID THCe - (rCH < 4   * CH4e ) - (rC < 2 H < 5 OH * C2H5OHe ) ` ` NOTE: If NMHCe is calculated to be less than zero, then NMHCe = 0. 5.4.2` ` ` NMHCd = FID THCd - (rCH < 4   * CH4d ) - (rC < 2 H < 5 OH * C2H5OHd) ` ` NOTE: If NMHCd is calculated to be less than zero, then NMHCd = 0. 5.4.3` ` ` COe = (1 - (0.01 + 0.005 * HCR) * CO2e - 0.000323 * Ra ) * COem ` ` ` NOTE: If a CO instrument which meets the criteria specified in CFR 40, 86.111 is used and the conditioning column has been deleted, COem must be substituted directly for COe . ` ` a) For E100 (100% ethanol), C2H5OH, where HCR = 3: ` `  COe = (1 - 0.02500 * CO2e - 0.000323 * Ra ) * COem ` `  100 *  x  ` `  hh#x + y/2 + 3.76 * (x + y/4 - z/2) 5.4.4` ` DF =   ` `  CO2e + (NMHCe + CH4e + COe + C2 H5 OHe + HCHOe ) * 10é4 ` ` (where fuel composition is CxHyOz as measured for the fuel used.) ` ` ` a) For E100 (100% ethanol), C2H5OH, where x = 1, y = 3, and z = 0.5: ` ` DF = 12.29 / [CO2e + (NMHCe + CH4e + COe + C2 H5 OHe + HCHOe ) * 10é4 ] 5.5 All Vehicles 5.5.1` ` NMHCconc = NMHCe - NMHCd * [1 - (1 / DF)] ` ` NOTE: If NMHCconc is calculated to be less than zero, then NMHCconc = 0. 5.5.2` ` NMHCmass < n   = NMHCconc * NMHCdens * VMIX * 10é6 6.TOTAL WEIGHTED NMHC MASS EMISSIONS 6.1All Vehicles  ` `   NMHCmass 1 + NMHCmass 2pp2 NMHCmass 3 + NMHCmass 2(#(#K6.1.1NMHCwm = 0.43 * ______________________ pp2+ 0.57 * ___________________ ` `  hh#Dphase 1 + Dphase 2pp27 Dphase 3 + Dphase 2 7.SAMPLE CALCULATIONS 7.1Given the following data for a gasoline vehicle, calculate the weighted NMHC mass emission.  XXX  Xddx Xddx 6c c 6Test PhaseFID THCe (ppmC)FID THCd (ppmC)CH4e (ppmC)CH4d (ppmC)COem (ppm)CO2e (%)VMIX (ft3)Dphase n (mile) Ra (%)6q q 6141.88.67.535.27147.21.1928463.583 386q q 6213.08.45.685.1020.80.8048563.848 386                  6315.48.96.165.2036.71.0428393.586 38  XXX For Phase 1: NMHCe = FID THCe - (rCH < 4  * CH4e ) ` `  = 41.8 ppmC - (1.04 * 7.53 ppmC) ` `  = 33.97 ppmC NMHCd = FID THCd - (rCH < 4  * CH4d ) ` `  = 8.6 ppmC - (1.04 * 5.27 ppmC) ` `  = 3.12 ppmC COe ` ` = (1 - 0.01925 * CO2e - 0.000323 * Ra ) * COem ` ` ` NOTE: If a CO instrument which meets the criteria specified in CFR 40, 86.111 is used and the conditioning column has been deleted, COem must be substituted directly for COe . ` ` = (1 - 0.01925 * 1.19% - 0.000323 * 38%) * 147.18 ppm ` ` ` = 142.0 ppm DF ` ` = 13.47 / [CO2e + (NMHCe + CH4e + COe ) * 10é4 ` ` =  13.47  ` `  1.19% + (33.97 ppmC + 7.53 ppmC + 142.0 ppmC) * 10é4 ` ` = 11.15 NMHCconc  = NMHCe - NMHCd * [1 - (1 / DF)] ` `  = 33.97 ppmC - 3.12 ppmC * [1-(1/11.15)] ` `  = 31.13 ppmC NMHCmass n  = NMHCconc * NMHCdens * VMIX * 10é6 ` `  = 31.13 ppmC * 16.33 g/ft3 * 2846 ft3 * 10é6 NMHCmass 1  = 1.45 g Similarly, for Phase 2: hh#NMHCmass 2 -= 0.33 g and for Phase 3:hh#NMHCmass 3 -= 0.27 g Therefore, ` `   NMHCmass 1 + NMHCmass 2pp2 NMHCmass 3 + NMHCmass 2 NMHCwm` ` = 0.43 * ______________________ pp2+ 0.57 * ________________ ` `  Dphase 1 + Dphase 2-pp2 Dphase 3 + Dphase 2 ` ` = 0.43 *  1.45 g + 0.33 g  + -0.57 *  0.27 g + 0.33 g  ` `   3.583 mile + 3.848 mile-pp273.586 mile + 3.848 mile NMHCwm ` ` = 0.15 g/mile 7.2Given the following data for a vehicle operating on 85% methanol and 15% gasoline (M85), calculate the weighted NMHC mass emission.  XXX #_PP# Xddx Xddx >  >Test PhaseFID THCe (ppmC)FID THCd (ppmC)CH4e (ppmC)CH4d (ppmc)CH3OHe (ppm)COem (ppm)CO2e (%)VMIX (ft3) Dphase n (mile) Ra (%) HCHOe (ppm)>P P >188.55.517.762.8272.9303.21.282832 3.570 32 0.96>P P >214.57.08.012.825.19.70.834827 3.850 32 0.10>                      >321.87.710.132.937.418.21.132825 3.586 32 0.12#vPXP#  XXX [For this example, CH3OHd was assumed to be 0.0 ppmC for all three background bag samples.] For Phase 1: NMHCe = FID THCe - (rCH < 4  * CH4e ) - (rCH < 3 OH * CH3OHe ) ` `  = 88.5 ppmC - (1.04*17.76 ppmC) - (0.66*72.9 ppmC) ` `  = 21.92 ppmC  NMHCd =FID THCd - (rCH < 4  * CH4d ) - (rCH < 3 OH * CH3OHd ) ` `  =5.5 ppmC - (1.04*2.82 ppmC) - (0.66*0.0 ppmC) ` `  = 2.57 ppmC COe` ` = (1 - 0.02705 * CO2e - 0.000323 * Ra ) * COem NOTE: If a CO instrument which meets the criteria specified in CFR 40, 86.111 is used and the conditioning column has been deleted, COem must be substituted directly for COe . ` ` = (1 - 0.02705 * 1.28% - 0.000323 * 32%) * 303.2 ppm ` ` = 289.6 ppm DF` ` = 12.02 / [CO2e + (NMHCe + CH4e + COe + CH3OHe + HCHOe ) * 10é4] =` `  12.02   1.28% + (21.92ppmC + 17.76ppmC + 289.6 ppmC + 72.9ppmC + 0.96ppm) * 10é4 ` ` = 9.10 NMHCconc = NMHCe - NMHCd * [1 - (1 / DF)] ` `  = 21.92 ppmC - 2.57 ppmC * [1 - (1 / 9.10)] ` `  = 19.63 ppmC NMHCmass n =NMHCconc * NMHCdens * VMIX * 10é6  NMHCmass 1 = 0.91 g Similarly, Phase 2:NMHCmass 2 = 0.0 g and for Phase 3:NMHCmass 3 = 0.10 g Therefore, ` `   NMHCmass 1 + NMHCmass 2pp2 NMHCmass 3 + NMHCmass 2 NMHCwm` ` = 0.43 * ______________________ pp2+ 0.57 * ________________ ` `  Dphase 1 + Dphase 2-pp2 Dphase 3 + Dphase 2 ` ` = 0.43 *  0.91 g + 0.00 g  + -0.57 *  0.10 g + 0.0 g  ` `   3.570 mile + 3.850 mile-pp273.586 mile + 3.850 mile NMHCwm ` ` = 0.06 g/mile 8. DEFINITIONS CH3OHd = ` ` ` the methanol concentration in the dilution air as determined from the dilution air methanol sample using the procedure specified in Method No. 1001, ppmC. CH3OHe =` ` ` the methanol concentration in the dilute exhaust as determined from the dilute exhaust methanol sample using the procedure specified in Method No. 1001, ppmC. CH4d =` ` ` the methane concentration in the dilution air, ppmC. CH4e =` ` ` the methane concentration in the dilute exhaust, ppmC. C2H5OHd =` ` ` the ethanol concentration in the dilution air as determined from the dilution air ethanol sample using the procedure specified in Method No. 1001, ppmC. C2H5OHe =` ` ` the ethanol concentration in the dilute exhaust as determined from the dilute exhaust ethanol sample using the procedure specified in Method No. 1001, ppmC. COe =the carbon monoxide concentration in the dilute exhaust corrected for carbon dioxide and water removal, ppm. COem =` ` ` the carbon monoxide concentration in the dilute exhaust uncorrected for carbon dioxide and water removal, ppm. CO2e =` ` ` the carbon dioxide concentration in the dilute exhaust, %. Dphase n =` ` ` the distance driven by the test vehicle on a chassis dynamometer during test phase n (where n is either 1, 2, or 3), mile. DF =` ` ` dilution factor. FID THCd =` ` ` the total hydrocarbon concentration including methane and methanol (for methanol-fueled engines) or ethanol (for ethanol-fueled engines) in the dilution air as measured by the FID, ppmC. FID THCd =` ` ` the total hydrocarbon concentration including methane and methanol (for methanol-fueled engines) or ethanol (for ethanol-fueled engines) in the dilute exhaust as measured by the FID, ppmC. HCHOe =` ` ` the formaldehyde concentration in the dilute exhaust as determined from the dilute exhaust carbonyl sample using the procedure specified in Method No. 1004, ppm. HCR =` ` ` the hydrogen-to-carbon ratio for the fuel used. NMHCconc =` ` ` the non-methane hydrocarbon concentration in the dilute exhaust corrected for background, ppmC. NMHCd =` ` ` the non-methane hydrocarbon concentration in the dilution air corrected for methane and alcohol removal, ppmC. NMHCdens =` ` ` the mass per unit volume of non-methane hydrocarbon corrected to standard conditions (16.33 g/ft3 at 293.16o K and 760 mm Hg assuming a C:H ratio of 1:1.85 and 17.28 g/ft3 for LPG at 293.16o K and 760 mm Hg), g/ft3 . NMHCe =` ` ` the non-methane hydrocarbon concentration in the dilute exhaust corrected for methane and alcohol removal, ppmC.NMHCmass n=` ` ` the mass emission of non-methane hydrocarbon per test phase n (where n is either 1, 2, or 3), g. NMHCwm =` ` ` the total weighted mass of non-methane hydrocarbon per mile for all three phases of the FTP, g/mile. Ra =` ` ` the relative humidity of the ambient air, percent. rCH < 3 OH =` ` ` the FID response factor to methanol (see CFR 40, 86.121-90(c)). rCH < 4  =the FID response factor to methane (see section 4). rC < 2 H < 5 OH =` ` ` the FID response factor to ethanol (same procedure for methanol response factor, see CFR 40, 86.121-90(c)). VMIX =` ` ` the total dilute exhaust volume measured per test phase and corrected to standard conditions (293.16o K and 760 mm Hg), ft3 . 9. REFERENCES (1)` ` ` Code of Federal Regulations, Title 40, Part 86, Subpart B (2)` ` ` SAE J254, "Instrumentation and Techniques for Exhaust Gas Emissions Measurement" (3)` ` ` SAE J1094a, "Constant Volume Sampler System for Exhaust Emissions Measurement" (4)` ` ` SAE 770141, "Optimization of a Flame Ionization Detector for Determination of Hydrocarbon in Diluted Automotive Exhausts". G.D. Reschke, Vehicle Emissions Laboratory, General Motors Proving Ground (5)` ` ` SAE J1154, "Methane Measurement Using Gas Chromatography", (revised December 1991)   @C-@   *Part C DETERMINATION OF ALCOHOLS IN AUTOMOTIVE SOURCE SAMPLES v BY GAS CHROMATOGRAPHY %METHOD NO. 1001  1. INTRODUCTION 1.1This document describes a method of sampling and analyzing automotive exhaust for alcohols in the range of 8 to 1200 micrograms g per 15 milliliters ( mL ) of solution. The "target" alcohols which shall be analyzed and reported by this method are methanol and ethanol. These alcohols, when present in concentrations above the LOD, shall be reported. 1.2This procedure is based on a method developed by the U. S. Environmental Protection Agency, (U.S. EPA) [Ref 9.1 6] which involves flowing diluted engine exhaust through deionized or purified water contained in glass impingers and analyzing this solution by gas chromatography (GC) . 1.3All definitions and abbreviations are contained in Appendix 2 of these test procedures. 2. METHOD SUMMARY 2.1The samples are received by the laboratory in impingers. Compound separation and analysis are performed using a GC. The sample is injected into the GC by means of a liquid autosampler. Separation of the sample mixture into its components constituents is performed by using a temperatureprogrammed capillary column operated with a temperature gradient . A flame ionization detector ( FID ) is used for alcohol detection and quantification. 2.2The computerized GC data system identifies the alcohol associated with each of the peak s . The alcohol concentrations are determined by integrating the peak areas and using response factors determined with from external standards. 3. INTERFERENCES AND LIMITATIONS 3.1An interference interferent is any component present in the sample with a retention time similar to that of any the target alcohol s described in this method. To reduce interference error, proof of chemical identity may require performance of periodic confirmations using an alternate method and/or instrumentation, e.g., gas chromatograph/mass spectrometer ( GC/MS ) . 3.2The concentration of the alcohols in the range of interest is stable for up to six days as long as the samples are sealed and refrigerated at a temperature below 40o F.4. INSTRUMENTATION AND APPARATUS 4.1For each mode of the CVS test, two sampling impingers, each containing a known amount of deionized or purified water (e.g. 15 mL for this procedure), are used to contain the sample. 4.1.1` ` ` A temperature-programmable GC, equipped with a DB-Wax Megabore column (30 meter ( m ) , 0.53 millimeter ( mm ) ID, 1.0 micron (  ) film thickness) and FID is used. Other columns may be used, provided the alternate(s) can be demonstrated to be equivalent or better with respect to precision, accuracy and resolution of all the target alcohols. 4.1.2` ` ` A liquid autosampler is required. 4.1.3` ` ` A PC-controlled data acquisition system for quantitation quantifying of peak areas is required. 5. REAGENTS AND MATERIALS 5.1Methanol shall have a purity of 99.9 percent, or be high performance liquid chromatography grade, EM Science or equivalent. 5.2Ethanol shall be absolute, ACS reagent grade. 5.3 American Standards for Testing Materials ASTM Type I purified or Type II deionized water shall be used. 5.4A stock solution is prepared gravimetrically or volumetrically  by diluting methanol and ethanol with deionized or purified water, e.g., for this method the stock solution contains is approximately  1 g/mL   10 mg/mL  percent by volume of each target alcohol. 5.4.1` ` ` A calibration standard within at the expected concentration range of the samples is prepared by successive dilutions of the stock solution with deionized or purified water, e.g., 50 parts per million (ppm) g/mL volume to volume (v/v) is typical. 5.4.2` ` ` A control standard containing all target alcohols is prepared by successive dilutions of a stock solution different from that of Section 5.4.1. This standard, at an approximate concentration of the samples, is used to monitor the precision of the analysis of update control charts for each target alcohol. 5.4.3` ` ` All standards should be refrigerated at less than a temperature below 40o F during storage. 5.5Gas requirements. 5.5.1` ` ` Air shall be "Zero" grade. "Ultrazero" grade may be required to achieve the LOD required by Section 8.8. 5.5.2` ` Nitrogen shall have a minimum purity of 99.998 percent. 5.5.3` ` Helium shall have a minimum purity of 99.995 percent. 5.5.4` ` Hydrogen shall have a minimum purity of 99.995 percent. 6. PROCEDURE 6.1Each of the graduated fritted sampling impingers is filled with 15 mL of deionized or purified water. 6.2The impingers are placed in an ice bath during the sample collection. 6.3After sampling, the impingers are allowed to warm to room temperature and the solution contained in each impinger is transferred to a vial and sealed. 6.3.1` ` ` Samples should shall be refrigerated ( at a temperature below 40o F or lower) if immediate analysis is not feasible, or if reanalysis at a later date may be required. 6.4One microliter aliquots of unmodified samples are injected via autosampler into a GC. Suggested standard operating conditions for the GC are:, configured as follows: Column: DB- w Wax, 30 m, 0.53 mm ID, 1.0 film thickness Carrier gas flow: Helium at 5 milliliters per minute ( mL/min ) Make-up gas flow:Nitrogen at 25 mL/min Detector: FID, Hydrogen at 30 mL/min and Air at 300 mL/min Injector: Packed column injector with Megabore adapter insert; on-column injection Column t T emperature:hhh50o C (1 min), 50o C to 70o C (5o C/min), 70o C to 110oĠC (15o C/min), 110o C (4 min) Data system: PC-based data acquisition system 6.4.1` ` ` One calibration standard, one control standard, and one deionized or purified water blank are analyzed daily at the beginning of each set of samples. 6.4.2` ` ` A replicate analysis is performed at least once per 24 hour period. 6.4.3` ` ` The control standard is analyzed performed at least once per 24 hour period. 6.4.4` ` ` For Samples containing compounds having concentrations above the documented range of instrument linearity , the sample must be diluted and reanalyzed. 6.4.5` ` ` The peak integrations are corrected as necessary in the data system. Any misplaced baseline segments are corrected in the reconstructed chromatogram. 6.4.6` ` ` The peak identifications provided by the computer are checked and corrected if necessary. 6.4.7` ` ` The target alcohol peaks at or above the maximum allowable limit of detection ( LOD ) are reported (Section 8.8). At the laboratory's discretion, peaks at or above the LOD calculated in section 8.8 may be reported. The calculated LOD must be lower than the maximum allowable LOD . 7. CALCULATIONS 7.1The concentration of each target alcohol, in g/mL, is determined by the following calculation that compares the sample peak area with that of an external standard:  Concentration (g/mL)sample = Peak Areasample x Response Factor where the response factor (RF) is calculated during the calibration by: ` `  Concentrationstandard (g\mL) ` ` RF =   ` `  Peak Areastandard 7.2This concentration is then used to calculate the total amount of alcohol in each impinger: Mass (g) = Concentration (g/mL) x Impinger volume (mL) 7.3An internal standard method may also be used. 8. QUALITY CONTROL 8.1Calibration and control standards are prepared at least every six months and analyzed daily. 8.2Blank Run. A deionized or purified water blank run is performed before running the calibration standard. All target alcohol concentrations from the blank analysis must be below the LOD before the analysis may proceed. 8.3Calibration Run. One run of the calibration standard is performed daily to generate the response calibration factor needed for quantitating quantifying sample analyses. 8.4Control Standard Run. One run of the quality control standard is performed after the calibration run. Measurements of all target alcohols in the control standard must fall within the control limits before sample analysis may proceed. To meet this requirement, it may be necessary to inspect and repair the GC, and rerun the calibration and/or control standards. 8.5Control Charts. A q Q uality control chart (s) is are maintained for each analyte in of the control standard sample . The control charts, used on a daily basis, establish es that the method is "in "statistical é control". The following describes how to construct a typical control chart: 1.` ` ` Obtain at least 20 daily control standard sample results; 2.` ` ` Calculate the average control standard sample mean concentration and standard deviation (s) for the target analyte (s) ; and 3.` ` ` Create a control chart for the target analyte (s) by placing the concentration on the Y-axis and the date on the X-axis. Establish Draw an upper warning limit and a lower warning limit at two standard deviations (2s) above and below the average concentration. Establish Draw an upper control limit and a lower control limit at three standard deviations (3s) above and below the average concentration.  The control sample must be "in-control" for The measured concentrations of all target analytes contained in the control standard must be within the control limits ("incontrol") for the sample results to be considered acceptable. A control standard sample measurement is considered to be "out-of-control" when the analyzed value exceeds the 3s limit, or two successive control standard sample measurements of the same analyte exceed the 2s limit. 8.6Duplicates. A duplicate analysis of one sample is performed at least once a day. The relative percent difference (RPD) is calculated for each duplicate run:RPD(%) = Difference between duplicate and original measurements x 100  F(#(#K` `   Average of duplicate and original measurements For each compound, the allowable RPD depends on the average concentration level for the duplicate runs, as shown in the following table: Average Measurement for Duplicate Runs-pp27Allowable RPD (%) ` ` 1 to 10hh#times LOD-pp27  <100 ` ` 10 to 20hh# " "(-pp2730 ` ` 20 to 50hh# " "(-pp2720 ` ` Greater than 50hh# " "(-pp2715 If the results of the duplicate analyses do not meet these criteria for all target alcohols, the sample must be reanalyzed. If the criteria are still not met, all sample results for the day from this instrument must be deleted and the samples reanalyzed. 8.7Linearity. A multipoint calibration to confirm check for instrument linearity is performed for all target alcohols for new instruments, after making instrument modifications which can affect linearity, and at least once every year six months . The multipoint calibration consists of at least five concentration or mass loading levels, each above the LOD, evenly distributed over the range of expected sample concentration linearity of the instrument . Each concentration level is measured at least twice. A linear regression analysis is performed using concentration and area counts to determine the regression correlation coefficient (r). The r must be greater than 0.995 to be considered sufficiently linear for one point calibrations. 8.8Limit of Detection. The LOD limit of detection for the target alcohols must be determined for new instruments, after making instrument modifications which can affect the LOD and at least once every year. every six months . To make the calculations, it is necessary to perform a multipoint calibration consisting of at least four "low" concentration levels, each above the expected LOD. A linear regression is performed on the data. The LOD must be calculated using the following equation [Ref. 7] (Ref 9.2) : ` `  hh#LOD =  b  + (t x s) ` `  hh# m   LOD = |A| + 3.3(S) where each term in the equation is expressed in concentration units, and |A| is the absolute value of the least squares X-intercept calculated from the multipoint data. S  b  is the absolute value of the yintercept, m is the slope of the linear regression, s  is the standard deviation of at least five replicate determinations of the lowest concentration standard, and t is the tfactor for 99 percent confidence for a onesided normal (Gaussian) distribution. The number of degrees of freedom is equal to the number of replicates, minus one. An abbreviated ttable is: ` ` Degrees of Freedomhh#(tvalue ` `  4hh#(3.7 ` `  5hh#(3.4` `  6hh#(3.1 ` `  7hh#(3.0  At least three replicates are required . The lowest standard must be of a concentration of at one to five times the estimated LOD detection limit . If data is not available in the concentration range near the detection limit, S may be estimated by: (S = RSD x |A|  RSD is the relative standard deviation of the lowest standard analyzed. An example of typical LODs is given in the table below: CAS Nos. COMPOUND(-LOD (ug/ml) 00067-56-1 methanolhh#(0.24 00064-17-5 ethanolhh#(-0.17 8.8.1` ` ` The maximum allowable LOD for each alcohol is 0.50 g/mL. The calculated laboratory LOD must be equal to or lower than the maximum allowable LOD. All peaks identified as target compounds that are equal to or exceed the maximum allowable LOD must be reported. If the calculated laboratory LOD is less than the maximum allowable LOD, the laboratory may choose to set its reporting limit at either the maximum allowable LOD or the calculated laboratory LOD. 8.8.2.` ` ` For the purpose of calculating the total mass of all species, the concentrations of the compounds below the LOD are considered to be zero. 9. REFERENCES 9.1U.S. Environmental Protection Agency, Characterization of Exhaust Emissions from Methanol and Gasoline Fueled Automobiles, EPA 460/3-82-004. 9.2U.S. Environmental Protection Agency, Compendium of Methods for the Determination of Toxic Organic Compounds in Ambient Air, (Method T03-15) EPA-600/4-89-017 Research Triangle Park, North Carolina, June, 1989.   @D-@ * Part D DETERMINATION OF C2 TO C5 HYDROCARBONS hIN AUTOMOTIVE SOURCE SAMPLES BY GAS CHROMATOGRAPHY %METHOD NO. 1002 ă 1. INTRODUCTION 1.1This document describes a method of analyzing, by gas chromatography, C2 to C5 hydrocarbons (light-end hydrocarbons) in the range of parts per billion carbon ( ppbC ) from automotive source samples. This method does not include sample collection procedures [Ref. 8] (Ref 9.1) . The "target" hydrocarbons which shall be analyzed and reported by this method and M m ethod 1003 are listed in Attachment Appendix 1. All compounds on this list, when present in concentrations above the LOD, shall be measured and reported ("targeted") by either Method 1002 or Method 1003. Each laboratory should divide the list into light-end (Method 1002) and mid-range (Method 1003) hydrocarbons in the manner which best suits the laboratory instrumentation. All compounds on the list not targeted by M m ethod 1002 must be targeted by M m ethod 1003. More compounds may be measured than those on the target list. 1.2All definitions and abbreviations are contained in Appendix 2 of these test procedures. 2. METHOD SUMMARY 2.1This is a rapid method intended for routine analysis. 2.2The samples are received by the laboratory in Tedlar bags, which are sub-sampled into a gas chromatograph ( GC ) for separation and analyses analysis. 2.3The gas chromatographic analysis is performed on a packed column operated isothermally at 35oC, or an Alumina (Al203) Porous Layer Open Tubular (PLOT ) column temperature programmed from OoC to 200oC. An flame ionization detector ( FID ) is used for detection and quantification. 2.4The sample is injected into the GC by means of gas sampling valves. Separation of the sample hydrocarbon mixture into its components constituents takes place in the chromatographic column. The chromatographic column and the corresponding operating parameters described in this method normally provide complete resolution of most target compounds. 2.5The computerized GC data acquisition system identifies the hydrocarbons associated with each of the peak s . The hydrocarbon concentrations are determined by integrating the peak areas and using response calibration factors determined from with NIST-traceable standards. 3. INTERFERENCES AND LIMITATIONS 3.1An interference interferent is any component present in the sample with a retention time very similar to that of any the target hydrocarbon s described in this method. To reduce interference error, proof of chemical identity may require performance of periodic confirmations using an alternate method and/or instrumentation, e.g., gas chromatograph/mass spectrometer ( GC/MS ) , photoionization detector ( PID ) , different column, etc. 3.2To maximize sample integrity, sample bags should not leak or be exposed to bright light or excessive heat. Sampling bags must be shielded from direct sunlight to avoid reactions occurring due to reactive hydrocarbons. The compound 1,3-butadiene, most of which is in CVS bag no. 1, is unstable. Therefore all CVS bag no. 1 samples must be analyzed within 8 4 hours; CVS bag no. 2, CVS bag no. 3, and background samples must be analyzed within 24 hours, although analysis within 8 hours is recommended. 4. INSTRUMENTS AND APPARATUS 4.1Tedlar bags, 2 mil in thickness, nominally 5 to 10 liters in capacity and equipped with quick-connect fittings, are used to contain the samples. 4.2For manual subésampling into a GC, a ground glass syringe is used to transfer gaseous samples from Tedlar bags to the GC sample inlet. For automated systems, a sample loop is used to transfer gaseous samples from the Tedlar bag to the sample inlet of the GC. Sample aliquot size is chosen based on considerations of instrument sensitivity and/or linearity. 4.3A temperatureéprogrammable GC equipped with a gas sampling valve system, a FID, and accessories is required. 4.4A stainless steel column [6 feet ( ft ) x 1/8 inch ( in ) ] packed with phenylisocyanate Durapak 80/100 mesh is used. An Al Alumina PLOT column ( 50 60 m x 0.32 mm) may also be used. has been shown to be equivalent A wax precolumn is recommended to prevent water damage to the PLOT column. Other columns may be used, provided the alternate(s) can be demonstrated to be equivalent or better with respect to precision, accuracy and resolution of all the target hydrocarbons. 4.5A sample trap capable of being cryogenically cooled may be used. 4. 5 6An electronic integrator for quantitation of peak areas is required. If the data acquisition system cannot record the chromatogram, an analog recorder is also required. 5. REAGENTS AND MATERIALS 5.1Helium shall have a minimum purity of 99.995 percent. Higher purity helium may be required to achieve the LOD required by Section 8.7.1.  5.2Hydrogen shall have a minimum purity of 99.995 percent. 5.3Air shall be "Zero" grade. "Ultrazero" grade may be required to achieve the LOD required by Section 8.7.1. 5.4Nitrogen shall have a minimum purity of 99.998 percent. 5.5 Calibration Standard - The quantitative calibration standard for all target hydrocarbons shall be propane at a concentration level between 0.25 and 1 ppm-mole and within the calculated linearity of the method (see section 8.6). This propane standard must be traceable to a NISTcertified SRM to a Standard Reference Material (SRM) certified for propane by the National Institute of Standards and Technology (NIST). NIST traceable means that the standard has been compared with not more than one intermediate standard to a NISTcertified SRM . A comparison between a SRM and a candidate standard will yield a secondary NIST traceable standard, while a comparison between a secondary NIST traceable standard and a candidate standard will yield a tertiary NIST traceable standard. A NIST SRM propane standard, secondary NIST traceable propane standard, or tertiary NIST traceable propane standard is required for calibration of M m ethod 1002 or 1003.  The minimum requirements for the manufacture and use of a secondary or tertiary NIST traceable propane standard shall be the following: 1) the standard must be packaged in aluminum cylinders which have been precleaned and passivated, 2) a gas chromatograph shall be used to compare the candidate cylinders with the NIST traceable standard using either direct or interpolation comparison. (In the direct comparison, the analyte concentration for the candidate standard may not differ more than 10 percent from the analyte concentration of the NIST traceable standard; while in the interpolation comparison, the analyte concentration of a candidate standard is bracketed between the analyte concentrations of the NIST SRM standards.), 3) analytes and balance gases in secondary candidate cylinders must be the same composition as the NIST SRM propane standards used in an interpolation comparison procedure, 4) hydrocarbon analytes (except propane) and balance gases may differ in secondary candidate cylinders from the NIST SRM propane standard used in a direct comparison procedure if there are no interferences with the propane measurement or stability, 5) comparison between a secondary and tertiary standard must be direct, 6) all candidate cylinders must be analyzed a minimum of 4 times and stated uncertainties must be determined for all secondary and tertiary standard analytes, and 7) gas in the cylinder cannot be used when the pressure of the cylinder falls below 300 pounds per square inch. It is recommended that either the laboratory or standard supplier(s) for the laboratory certify the secondary standards through the NIST Traceable Reference Material program or obtain Research Gas Material/Mixtures of propane standards from NIST for use as secondary standards. 5.6 Control Standard - A quality control standard, containing at least ethene, propane, n-butane, and 2-methylpropene with a concentration between 0.2 and 1 parts per million carbon ( ppmC ) based on a propane standard, is used for the following quality control purposes: ` ` 1. Daily update of control charts, and ` ` 2. Daily determination of marker retention time windows.  The control standard(s) must have concentrations verified against a NIST-traceable propane standard (See Section 5.5 for definition of NIST-traceable) when used for either LOD determinations or linearity checks. This verification can be performed at the laboratory conducting the sample analysis. 5.7A high concentration standard (higher than the calibration standard), containing the target hydrocarbons listed in Section 5.6 is used semi-annually for linearity determinations. The high concentration calibration standard must have concentrations verified against a NIST-traceable propane standard (see Section 5.5 for the definition of NIST-traceable). This verification can be performed at the laboratory performing the analysis. 5.8Liquid nitrogen may be required is used to cool the cryogenic sample trap and column oven where applicable. 6. PROCEDURE 6.1The gaseous sample is analyzed for the target hydrocarbons C2 through C5. 6.2Suggested standard operating conditions for the gas chromatograph are: 6.2.1` ` Packed Column: Helium carrier gas flow:hhh(50 milliliter/minute ( m l L/min) (packed column) Hydrogen gas flow:hh#(32 mL/min Air flow: hh#(300 mL/min Sample valve temperature: hh#(ambient Heating bath temperature: hh#(60o - 80oC Injector temperature:hh#35oC (packed column) Column temperature:hh#35oC (isothermal) (packed column) ` ` ` hhh 0oC to 200oC (temperature program) (Al PLOT Column) Detector temperature:hh#(200 to 250 oC 6.2.2` ` ` PLOT Column: Helium carrier gas velocity:hh#(30 cm/sec at 200oC Nitrogen makeup gas flow:hhhsufficient such that the total flow of helium plus nitrogen is 30 mL/min Hydrogen gas flow:hh#30 mL/min Air flow: hh#(300 mL/min Sample valve temperature:hh#(150oC (PLOT column) Column temperature:hhh0oC (hold 7 min), 10oC/min to 200oC (hold 15 min) Detector temperature: hh#250oC 6.3For automated systems, connect the samples to the GC and begin the analytical process. 6.4Introduce the sample into the carrier gas stream through the injection valve. 6.4For automated systems, connect the samples to the GC and begin the analytical process. Č6.5Each separated component exits from the column into the FID where a response is generated. 6.6Concentrations of hydrocarbons are calculated by an electronic integrator device, which has been calibrated using a NIST-traceable propane calibration standard. 6.7For compounds having concentrations above the documented range of instrument linearity, a smaller aliquot must be taken (for manual systems, a smaller syringe or smaller loop; for automated systems, a smaller loop). 6.8The peak integrations are corrected as necessary in the data system. Any misplaced baseline segments are corrected in the reconstructed chromatogram. 6.9The peak identifications provided by the computer are checked and corrected if necessary. 6.10All The peaks areas of identified as target compounds (Appendix 1) at or above the maximum allowable LOD are reported (Section 8.7). At the laboratory's discretion, peaks at or above the LOD calculated in section 8.7 may be reported. The calculated LOD must be lower than the maximum allowable LOD. 6.11Target compounds which coelute are reported as the major component, as determined by the analysis of several samples by GC/MS or other methods. An exception to this is m and pxylene, where GC/MS data and fuel profiles are used to determine the relative contribution of each component to the peak. This method was used to determine the m and pxylene MIR value given in Appendix 1. 6.11The maximum retention time in this analysis is typically about 15 30 mins. 6.12After each run, the packed column is backéflushed with helium while the oven temperature is raised and maintained at 60oC for 15 mins, or as required to flush the column. 6.13The Al Alumina PLOT column is programmed to 200o C to assure all compounds are eluted before the next run. 6.1 3 4Before the next run, sufficient time (typically 15 mins) is allowed after backéflush of the packed column to re-establish the required temperature of the column. 6.14The total run time is typically about 45-60 mins. 7. CALCULATIONS 7.1The target hydrocarbon concentrations, in ppbC, are calculated by the data system using propane as an external standard. 0 Concentrationsample (ppbC) = Peak Areasample x Response Factor where the response factor (RF) is calculated during daily calibration by:RF = Concentration of NIST-traceable propane standard, ppbC ` `  Area of propane peak 8. QUALITY CONTROL 8.1Blank Run. A blank (pure nitrogen or helium) is run once daily before running the calibration standard, control standard, and samples. All target hydrocarbon concentrations from the blank analysis must be below the LOD before the analysis may proceed. As an alternative to a daily blank run, a daily partial blank check in tandem with a weekly blank run may be used. A partial blank check is where the calibration standard, consisting of only propane and make-up gas (all organic compounds except methane and propane are below 2 percent of the propane standard concentration), is run daily and is checked for contamination except in the propane region of the chromatograph. The weekly blank run will provide a check on contamination in the propane region of the chromatograph. 8.2Calibration Run. One run of the calibration standard is performed per day to generate the response calibration factor needed for quantitating quantifying sample analyses. 8.3Control Standard Run. One run of the quality control standard is performed daily. Measurements of all compounds in the control standard must fall within the control limits before sample analysis may proceed. To meet this requirement, it may be necessary to inspect and repair the GC, and rerun the calibration and/or control standards. 8.4Control Charts. A Q quality control chart (s) are is maintained for each component of the control standard sample . The control charts, used on a daily basis, establish es that the method is "iné "statistical control." The following describes how to construct a typical control chart: ` ` 1. Obtain at least 20 daily control standard sample results; ` ` 2. Calculate the average control standard sample mean concentration and standard deviation (s) for the each target hydrocarbon; and ` ` 3. Create a control chart for the each target hydrocarbon by placing the concentration on the Y-axis and the date on the X-axis. Establish an upper warning limit and a lower warning limit at two standard deviations (2s) above and below the average concentration. Establish an upper control limit and a lower control limit at three standard deviations (3s) above and below the average concentration.  The control sample must be "in-control" for The measured concentrations of all target hydrocarbons contained listed in the control standard sample must be within the control limits ("incontrol") for the sample results to be considered acceptable. A control standard sample measurement is considered to be "out-of-control" when the analyzed value of the sample measurement exceeds the 3s limit, or two successive control standard sample measurements of the same analyte exceed the 2s limit. 8.5Duplicates. A duplicate analysis of one sample is performed at least once a day. The relative percent difference (RPD) is calculated for each duplicate run:RPD (%) =  Difference between duplicate and original measurements x 100 (#(#K(#(#K(#(#KAverage of duplicate and original measurements For each compound in the control standard, the allowable RPD depends on the average concentration level for the duplicate runs, as shown in the following table: Average Measurement for the Duplicate Runspp2Allowable RPD (%)  ` ` 1 to 10 hh#times(LOD -pp27  <100 ` ` 10 to 20hh#" ("-pp27  <30 ` ` 20 to 50hh#"("-pp27  <20 ` ` Greater than 50hh#"("-pp27  <15  If the results of the duplicate analyses do not meet these criteria for all target hydrocarbons in the control standard, the sample must be reanalyzed. If the criteria are still not met, all sample results for the day from this instrument must be deleted and the samples reanalyzed. 8.6Linearity. A multipoint calibration to confirm check for instrument linearity is performed for the target hydrocarbons in the control standard for new instruments, after making instrument modifications which can affect linearity, and at least once every year six months unless a daily check of the instrument response indicates that the linearity has not changed. To monitor the instrument response, a quality control chart is constructed, as specified in section 8.4, except using calibration standard area counts rather than control standard concentrations. When the standard area counts are outofcontrol, corrective action(s) must be taken before analysis may proceed. The multipoint calibration consists of at least five concentration or mass loading levels (using smaller or larger volume sample sizes of existing standards is acceptable), each above the LOD, evenly distributed over the range of expected sample concentration linearity of the instrument . Each concentration level is measured at least twice. A linear regression analysis is performed using concentration and average area counts to determine the regression correlation coefficient (r). The r must be greater than 0.995 to be considered sufficiently linear for one-point calibrations. 8.7Limit of Detection. The limit of detection LOD for the target hydrocarbons in the control standard must be determined must be determined at least once every year six months . unless a daily check of the instrument response indicates that the LOD has not changed. To monitor the instrument response, a quality control chart is constructed, as specified in section 8.4, except using calibration standard area counts rather than control standard concentrations. When the standard area counts are outofcontrol, corrective action(s) must be taken before analysis may proceed. To make the necessary calculations, it is necessary to perform a multipoint calibration consisting of at least four "low" concentration levels, each above the LOD. The LOD must be calculated using the following equation [Ref. 9] (Ref 9.2) : ` `  hh#LOD =  b  + (t x s) ` `  hh# m   LOD = |A| + 3.3(S) Čwhere each term in the equation is expressed in concentration units, and |A| is the absolute value of the least squares X-intercept calculated from the multipoint data. S  b  is the absolute value of the yintercept, m is the slope of the linear regression, s  is the standard deviation of at least five replicate determinations of the lowest concentration standard, and t is the tfactor for 99 percent confidence for a onesided normal (Gaussian) distribution. The number of degrees of freedom is equal to the number of replicates, minus one. An abbreviated ttable is: ` ` Degrees of Freedomhh#(tvalue ` `  4hh#(3.7 ` `  5hh#(3.4 ` `  6hh#(3.1 ` `  7hh#(3.0  At least three replicates are required . The lowest standard must be of a concentration of at one to five times the estimated LOD detection limit . If data is not available in the concentration range near the detection limit, S may be estimated by: (S = RSD x |A|  RSD is the relative standard deviation of the lowest standard analyzed. 8.7.1.` ` The maximum allowable LOD for each compound is 20 ppbC propane . The calculated laboratory LOD must be equal to or lower than the maximum allowable LOD. All peaks identified as target compounds that are equal to or exceed the maximum allowable LOD must be reported. If the calculated laboratory LOD is less than the maximum allowable LOD, the laboratory may choose to set its reporting limit at either the maximum allowable LOD or the calculated laboratory LOD. 8.7.2.` ` For the purpose s of calculating the total mass (ppbC) of all species, the concentrations of all compounds below the LOD are considered to be zero. 8.8Method No. 1002/Method No. 1003 Crossover Check For each sample, a compound shall be measured by both Method No. 1002 and Method No. 1003. The crossover compound shall be a compound that can reasonably be expected to be found and measured by both methods in the laboratory performing the analysis. The results obtained by the two methods should be compared and an acceptance criteria set for the relative percent difference. 9. REFERENCES 9.1Standard Test Method for C1 through C6 Hydrocarbons in the Atmosphere by Gas Chromatography, American Standards for Testing Materials (ASTM) Standards on Chromatography (1981). 9.2U.S. Environmental Protection Agency, Compendium of Methods for the Determination of Toxic Organic Compounds in Ambient Air (Method T03-15),EPA-600/4-84-041. Research Triangle Park, North Carolina, April 1989.   @E-@  *Part E DETERMINATION OF C6 TO C12 HYDROCARBONS hIN AUTOMOTIVE SOURCE SAMPLES BY GAS CHROMATOGRAPHY %METHOD NO. 1003 ă X` hp x (#%'0*,.8135@8: