Analysis of carbonaceous aerosols using the Thermal Optical Transmittance and Thermal Optical Reflectance Methods 99-439 Gary A. Norris*, M. Eileen Birch**, Michael P. Tolocka*, Charles W. Lewis*, Paul A. Solomon*, James B. Homolya*** * U.S. EPA, ORD/NERL, Research Triangle Park, NC 2771 1 ** National Institute for Occupational Safety And Health, Cincinnati, Ohio 45226-1998 *** U.S. EPA OAQPS, Research Triangle Park, NC 27711 ABSTRACT Carbonaceous particulate Typically represents a large fraction ofPM;>.5(20 - 40%). Two primary techniques presently used for the analysis of particulate carbon arc Thermal Optical Transmission (TOT - NIOSII Method 5040) and Thermal Optical Reflectance (TOR). These two methods both quantify carbon by heating filters and volatilizing the carbon that is oxidized in a granular bed of Mn02, reduced to CH4 in a Ni methanator, and quantified as CH.i with a flame ionization detector. However, the methods use different techniques to correct for the formation of pyrolysis products and the temperature programs for defining organic and elemental carbon. The TOT and TOR measurement techniques are being compared using samples from the Chemical Speciation Monitor Evaluation Field Study. All of the samples will be measured with TOR and a subset of samples representing a range of mass concentrations will be measured with TOT. This comparison will provide insight into the effect of the measurement technique parameters on organic and elemental carbon concentrations. ------- INTRODUCTION Carbonaceous particulate represents a significant fraction of airborne particulate matter less than or equal to 2.5 |im in aerodynamic diameter (PM2.5). Carbonaceous particulate is classified into three main categories: organic carbon (OC), elemental carbon (EC), and carbonate carbon (CC)1. These categories are operationally defined by the method or analysis technique. OC sources include combustion, industrial, and photochemical process. EC sources are mainly combustion related. CC is found in soil and generally constitutes less than 5 percent of the total carbon2. OC and EC typically represent 20 to 40 percent of PM2.5. As a result, quantifying OC and EC is required for reconstruction of the gravimetric mass. In addition, OC and EC can be used in receptor models and as exposure variables for health effects studies. Two primary thermal-optical methods are used to quantify particulate carbon: thermal optical reflectance (TOR) and thermal optical transmission ( TOT). These two methods both quantify carbon by heating a quartz filter punch (0.5 cm2 for TOR and 1.5 cm2 for TO T) and volatilizing the carbon which is oxidized in a granular bed of MnN02, reduced to CI-I4 in a Ni methanator, and quantified as methane (CH4) with a flame ionization detector. However, the methods differ in the technique used to correct for the formation of pyrolysis products and they use different temperature programs. The TOR particulate carbon analysis method has been previously described by Chow2. The temperature profile and the range of times required for the response at each step to ------- become constant is shown in Table 1. The rumple oven is stepped to 500 °C in 4 steps (120 °C (OC1), 250 °C (OC2), 450 °C (OC3), 500 °C (OC4)) to vaporize the organic carbon in a helium atmosphere. In the second part of the analysis a 2 percent (V98 percent He mix is introduced, the temperature is then stepped to 800 °C in 3 steps (550 °C (EC1), 700 °C (EC2), and 800 °C (EC3)). The amount of time for each temperature step is operationally defined based on the time required for the FID response to return to baseline. The result of this temperature step routine is the OC and EC are quantified in terms of the temperature steps (OC1, OC2, OC3, OC4, EC1. EC2, EC3). Methane (CH4) calibration gas is introduced at the end of each sample cycle. Pyrolysis correction is made by monitoring the filter reflectance of a lle-Ne laser at 632.8 nm with a photodetector. Pryolizcd OC is quantified as the carbon evolved from the time the carrier gas is changed from He to 2 percent O2 in He to the time the laser measured filter reflectance reaches its initial value. Carbonate carbon Ls determined by acidifying the sample with HC1 and measuring evolved carbon at ambient temperature in a 2 percent oxygen/98 percent helium atmosphere. The TOT method used in this analysis in specified in NIOSH Method 5040 3*4. In this method, the sample oven is purged with helium and the temperature is stepped (to 205 °C. 500 °C, 650 °C, and 850 °C) to volatilize the OC. and CC. In the second part of the analysis the temperature is lowered to 650 °C, and a 2 percent O2/98 percent He mix is introduced, the temperature is then stepped to 940 °C. Table 1 specifies the temperature program and time periods for the steps. At the end of the analysis, a calibration gas standard (CH4) is introduced. Correction for pyroloysis of the OC is accomplished by ------- monitoring the 'ransmittance of a He-Ne laser though the filter at 670 nm The point at which the filter transmittance reaches its initial value is defined as the split between OC and EC. Carbonate carbon is determined using HC1 pretreatment or the thermogram will show a quant ifiable peak around 820 °C. The National Air Monitoring Station (NAMS) chemical spcciation program specifics the NIOSH Method 5040 as the method for OC, EC, and total carbon (TC = EC + OC). The TOR and TOT methods were evaluated in a methods comparison in 1986 5. Comparisons were based on ambient PMi0, automotive, wood smoke, pvrolized ambient PMio, organic aerosol from a smog chamber, and a blank. Results were presented for the laboratories used for this study with Desert Research Institute (DR1) and Sunset Labs (SL) conducting the TOR and TOT analysis, respectively. The coefficient of variation for the TC. OC, and EC for the ambient samples for TOR and TOT were 3.0, 2.5, 3.0 percent; and 2.2, 2.5, and 6.2 percent, respectively. Results for each laboratory were reported as TC normalized to the mean,of all of the participants (n - 10) and the ratio of EC to TC (see Table 2). The TC results for both of the methods were similar, however, the EC/TC ratio for TOT was lower than TOR for the ambient, unleaded auto, and wood smoke samples. Wood smoke had the largest difference with TOR and TOT having EC/TC ratios of 0.36 and 0.09, respectively. An interlaboratory comparison between TOR (DR1) and TO T (SL) was conducted as part of the DRI study on Phoenix and Tuscon Urban Haze and PMjo. This study also found good agreement on three TC measurements of potassium hydrogen phthalate, with an average absolute percent difference less than 3 percent2. ------- The TOR and TOT methods were also evaluated in an interlaboratory comparison6. Samples from an urban location, a loading dock with diesel vehicles, and a firehouse bay were evaluated. The comparison also included filters spiked with sucrose, and disodium salt of ethylenediaminetetraacetic acid (EDTA) that have no EC to evaluate the TC and the pryrolysis correction. Table 3 summarizes the results from the methods comparison. TOR results are from one laboratory while the TOT results represent an average of 5 laboratories. Both the TOR and TOT techniques quantified less than 3 percent of the EDTA and sucrose standards as EC. The absolute percent difference between the TOR and TOT methods for the EC measurements were 33, 76. 58, 27, 23 for the wood 1 , wood 2, urban, dicsei truck, and fire station samples, respectively. The absolute percent difference between the TOR and TOT for the OC measurements were 1, 6, 3 for the diesel truck, and lire station samples, respectively. EC had largest percentage difference for the wood smoke and urban samples. This difference in the EC measurements may be due to either the temperature program, length of analysis at each temperature, ( Table 1) or the method used to correct for pryrolysis of the OC. EXPERIMENTAL DESIGN This comparison of the TOR and TOT particulate carbon analysis techniques will use ambient PM2.5 samples from 3 cities with different sources of carbon; and quality assurance standards of sucrose, wood smoke, diesel, dicsei spiked with carbonate, and ------- blank filters (Table 4). Sunset Laboratory will conduct the TOT analysis using NIOSH Method 5040 and DRI will conduct the TOR analysis following the temperature program in Table 1. Quality assurance samples were sent to each laboratory. Sunset Laboratory and DRI will each be sent a sample set: sucrose solution and a 47-mm quartz filter, carbonate solution, quartz filter with diesel particulate, and a quartz filter with wood smoke particulate. To determine the potential interference of carbonate when present in a carbonaceous particulate sample (in this case diesel), the sample portion would be analyzed with and without a spike. The labs should obtain the same EC results in both cases if carbonate does not interfere. All of the standards will be provided in triplicate to allow for calculation of the measurement precision of the calibration and source samples. The ambient samples will be from the Philadelphia and Rubidoux PM2 5 Chemical Speciation Sampler Evaluation sites, and the Spokane Particulate Matter and Health Study 1. Philadelphia particulate carbon is a mix of automobile and industrial source while Rubidoux is a mix of automobile and photochemical organic aerosols. The Philadelphia and Rubidoux samples will be evaluated with and without XAD denuders upstream of the quartz filter to remove organic gases. Spokane filters will be collected in November and December when the particulate carbon is predominately from wood smoke7. Two additional wood smoke source samples from EPA will also be evaluated. Samples from the 3 cities and the source samples will represent a range of OC concentrations and FC/OC ratios. Differences between the methods will be highlighted ------- by using the combination of samples from cities with different sources of carbon, and source samples (see Table 4). CONCLUSIONS Particulate carbon typically represents a large fraction of PM2.5. OC and EC measurements arc routinely used for reconstruction of the gravimetric mass, and receptor modeling. The two primary carbon analysis methods arc TOR and TOT. OC and EC arc determined operationally by each of the methods. These techniques differ in their temperature programs, step time, and pryrolysis correction. Differences between the methods need to be fully understood to help interpret past and future particulate carbon results. This comparison of the TOR and TOT particulate carbon analysis methods will add to the analyses conducted by Countess5 and Birch6. Analysis of PM2 5 samples from three cities with different sources of particulate carbon will help provide information on any systematic differences in the OC and EC measurements. In addition, source samples of diesel and wood smoke will be analyzed to evaluate the measurement precision. Particulate carbon measurements of samples with and without an XAD denuder will provide data on the differences between the TOR and TOT methods for samples which do not have the potential OC artifact caused by gas phase organic compounds. Organic gas denuders have not been used in previous intcrcomparisons. ------- DISCLAIMER This paper has been reviewed in accordance with the U.S. Environmental Protection Agency's peer and administrative review policies and approved for presentation and publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. ------- REFERENCES 1. Fung, K. Aerosol Sci. TechnoL, 1990, 12, 122-127 2. Chow, J.C.; Watson, J.G.; Pritchett, L.C.; Pierson, W.R.; Frazier, C.A.; Purcell, R.C. Atmos. Environ, 1993, 8, 1185 - 1201. 3. Birch, MA. Analyst 1998,123, 851-857 4. Eller, P.M.; Cassinelli, M.E. Elemental Carbon (Diesel Particulate): Method 5040. N10SH Manual of Analytical Methods, 4th ed. (1st Supplement) National Insitute for Occupational Safety and Health, DliliS (N10S1I), Cincinnati, Oil; Publicaiton No. 96- 135, 1996. 5. Countess, R.J. Aerosol Sci. TechnoL, 1990, 12, 114-121 6. Birch, M.H.; Cary, R.A. Aerosol Sci. TechnoL, 1996. 25, 221-241 7. Haller, L.; Claiborn C.; Larson, T.; Kocnig J.; Norris G.; Edgar R. J. Air Waste Manage. Assoc. ------- Table 1. Optical correction and thermal programs for the TOT and TOR particulate carbon measurement techniques. Method TOT Optical correction Transmission Program OC EC Helium 250 °C, 1 min 2 % O2 in Helium 650 °C, 30 sec 500 °C, 1 min 750 °C, 30 sec 650 °C, 1 min 850 °C, 1 min 850 °C, 1.5 min 940 °C, 2 min Reduce to 650 °C TOR Optical correction Reflectance Program OC EC Helium OC1: 120 °C, 3 - 10 min 2 % O2 in Helium EC1: 550 °C 3-10 min OC2: 250 °C, 3-10 min OC3: 450 °C, 3 - 10 min FC2: 700 °(\ 3-10 min EC3: 800 °C, 3-10 min OC4: 550 °C, 3 10 min OC - OC1 t OC2 t OC3 + OC4 pryrolized carbon EC - EC1 + EC2 -* EC3 pryrolized carbon Table 2. Comparison of the TOR and TOT analysis methods (Countess) TC normalized to the mean of the 10 participants in the comparison Method Ambient PMm Unleaded Auto Diesel Wood smoke sample 1 TOR 1.00 0.98 1.00 1.02 TOT 0.99 0.88 0.95 0.^4 Mean loading ((ig/cm2) 30.8 34.0 106.8 96.7 IC/TC ratio Method Ambient PM10 Unleaded Auto Diesel Wood smoke sample 1 TOR 0.30 0.83 0.87 0.36 TOT 0.15 0.70 0.88 0.09 Mean ratio 0.22 0.72 0.81 0.16 ------- Tabic 3. Comparison of the TOR and TOT analysis methods (ng/cm2) (Birch) TOR TOT Sample OC EC OC EC Sucrose 0.57 0.26 EDTA 0.20 0.02 Wood 1 6.20 4.42 Wood 2 — 2.17 — 0.30 Urban 9.70 3.00 10.42 1.65 Diesel Truck 17.33 8.20 18.48 6.25 Fire Station 136 20.27 139.80 16.10 ------- Table 4. Samples to be analyzed using TOT and TOR Speciation Site Number of Samples W/O XAD Denuder Philadelphia 10 Carbon from motor vehicles and industrial processes Rubidoux 10 Carbon from motor vehicles and secondary aerosols Spokane 4 Carbon from motor vehicles and wood smoke Field blanks 4 2 field blanks from Philadelphia, and Rubidoux Wood smoke source samples 2 2 wood smoke source samples from the EPA, Research Triangle Park. NC Source sample blank 1 1 source sample blank from the EPA, Research Triangle Park, NC With XAD Dcnudcr Philadelphia 5 Rubidoux 5 Quality Assurance Sucrose 3 Triplicate evaluation of the calibration Wood Smoke 3 Triplicate evaluation of a wood smoke sample Diesel 3 Triplicate evaluation of a diesel sample : Diesel spike with carbonate 3 Field blanks 2 ------- NERL-RTF-O-647 TECHNICAL REPORT DATA I. REPORT NO. EPA 600/R-99/059 2. 3.RECIPIENT'S ACCESSION NO. 4. TITLL AND SUBTITLE Analysis of carbonaceous aerosols using the Thermal Optical Transmittance and Thermal Optical Reflectance Methods 5.REPORT DATE 6.PERFORMING ORGANIZATION CODE 7. AUTHOR(S) Gary A. Norris*, M. Eileen Birch**, Michael P. Tolocka*. Charles W. Lewis*. Paul A. Solomon*, James B. Homolya*** • U.S. EPA, ORD/NERL, Research Triangle Park, NC 27711 " • National Institute for Occupational Safety And Health, Cincinnati, Ohio 45226- 1998 * * • U.S. EPA OAQPS, Research Triangle Park, NC 27711 8.PERFORMING ORGANIZATION REPORT NO. 9. PERFORMING ORGANIZATION NAME AND ADDRESS USFPA/NFRL/SACB 79 TW Alexander Drive MD-47 RTP, NC, 27711 10.PROGRAM ELEMENT NO. 11. CONTRACT/GRANT NO. RTI Contract 68D50040, WA 111 30 12. SPONSORING AGENCY NAME AND ADDRESS USHPA/NERL/SACB 79 TW Alexander Drive MD-47 RTP, NC, 27711 13.TYPE OF REPORT AND PERIOD COVERED 14. SPONSORING AGENCY CODE 15. SUPPLEMENTARY NOTES 16. ABSTRACI Carbonaceous paniculate typically represents a large fraction of PM?S (20 -40%). Two primary techniques presently used for the analysis of particulate carbon are Thermal Optical Transmission (TOT - NIOSH Method 5040) and Thermal Optical Reflectance (TOR). These two methods both quantify Carbon by heating filters and volatilizing the carbon that is oxidized in a granular bed of MnO,, reduced to CH4 in a Ni methanator, and quantified as CH4 with a flame ionization detector. However, the methods use different techniques to correct for the formation ofpyrolysis products and the temperature programs for defining organic and elemental carbon. The TOT and TOR measurement techniques are being compared using samples from the Chemical Speciation Monitor Evaluation Field Study. All of the samples will be measured with TOR and a subset of samples representing a range of mass concentrations will he measured with TOT. This comparison will provide insight into the effect of the measurement technique parameters on organic and elemental carbon concentrations. 17. KEY WORDS AND DOCUMENT ANALYSIS a. DESCRIPTORS h.IDENTIFIERS/ OPEN ENDED TERMS c.COSATI 18. DISTRIBUTION STATEMENT RELEASE TO PUBLIC 19. SECURITY CLASS (This Report) UNCLASSIFIED 21.NO. OF PAGES 12 20. SECURITY CLASS (This Page) UNCLASSIFIED 22. PRICE ------- |