Correction Methods for Organic Carbon Artifacts When Using Quartz-Fiber Filters in Large Particulate Matter Monitoring Networks: The Regression Method and Other Options

Abstract

Sampling and handling artifacts can bias filter-based measurements of particulate organic carbon (OC). Several measurement-based methods for OC artifact reduction and/or estimation are currently used in research-grade field studies. OC frequently is not artifact-corrected in large routine sampling networks (e.g., U.S. Environmental Protection Agency (EPA)'s Chemical Speciation Network). In some cases, the OC artifact has been corrected using a regression method (RM) for artifact estimation. In this method, the gamma-intercept of the regression of the OC concentration on the fine particle (PM2.5) mass concentration is taken to be an estimate of the average OC sampling artifact (net of positive and negative artifacts). This paper discusses options for artifact correction in large routine sampling networks. Specifically, the goals are to (1) articulate the assumptions and limitations inherent to the RM, (2) describe other artifact correction approaches, and (3) suggest a cost-effective method for artifact correction in large monitoring networks. The RM assumes a linear relationship between measured OC and PM mass: a constant slope (OC mass fraction) and a constant intercept (RM artifact estimate). These assumptions are not always valid. Additionally, outliers and other individual data points can have a large influence on the RM artifact estimates. The RM yields results within the range of measurement-based methods for some datasets and not for others. Given that the adsorption of organic gases increases with atmospheric concentrations of organics, subtraction of an average artifact from all samples (e.g., across multiple sites) will underestimate OC for lower-concentration samples (e.g., clean sites) and overestimate OC for higher-concentration samples (e.g., polluted sites). For relatively accurate, simple, and cost-effective artifact OC estimation in large networks, the authors suggest backup filter sampling on at least 10% of sampling days at all sites with artifact correction on a sample-by-sample basis as described herein.

U.S. National PM2.5Chemical Speciation Monitoring Networks—CSN and IMPROVE: Description of networks

“…For the collection of nitrate, both networks use denuders and reactive nylon filters to minimize sampling artifacts caused by nitrate volatilization (Hering et al, 1988;Hering and Cass, 1999). Carbon artifacts are accounted for by using various blank correction approaches (Kim et al, 2005;Chow et al, 2008;Watson et al, 2009;Chow et al, 2010;Maimone et al, 2011), as discussed later. Therefore, these samplers are designed to minimize sampling artifacts for nitrate while allowing flexibility for estimating organic carbon artifacts.…”

“…However, while heat treating removes the initial OC artifact, QFF medium continues to sorb organic gases until reaching pseudo-equilibrium. The overall artifact is both passive and active and depends on the sampled volume and several other factors and can reach the equivalent of 4 mg/m 3 for CSN filters sampled with the MetOne sampler (McDow and Huntzicker, 1990;EPA, 1998;Solomon et al, 2000;Turpin et al, 1994;Turpin et al, 2000;Maimone et al, 2011).…”

“…Studies examined uncertainty issues in only one network or the other, whereas other studies examined it between networks. These studies include the initial CSN intercomparisons (Solomon et al, 2000;Coutant and Stetzer, 2001;Solomon et al, 2003), estimating the overall measurement uncertainty for mass or one or more chemical components for CSN and IMPROVE (Solomon et al, 2000;Chow et al, 2001;Solomon et al, 2003;Ashbaugh et al, 2004aAshbaugh et al, , 2004bChow et al, 2004;Watson et al, 2005;White et al, 2005;Yu et al, 2006;Smiley, 2013;Hyslop and White, 2008b), and estimating different uncertainty components (blanks, MDL, sampling artifacts) for OC, mass, ions, and elements by EDXRF (Iyer et al, 2000;Gutknecht et al, 2001;Kim et al, 2001;Lewtas et al, 2001;Fehsenfeld et al, 2004;Subramanian et al, 2004;Gego et al, 2005;Kim et al, 2005;Gutknecht et al, 2006;Zhao et al, 2006;Chow et al, 2008;Hyslop and White, 2008a;Watson et al, 2009;Chow et al, 2010;Gutknecht et al, 2010;Cheng et al, 2011;Maimone et al, 2011;Malm et al, 2011). Several special EPA Office of Air Quality Planning and Standards (OAQPS) studies also were conducted to examine sampling artifacts associated with QFF medium caused by a specific filter cassette (Clark, 2002b); the effect of QFF filter type (manufacturer) for collection and analysis of OC and EC (Peterson et al, 2007); the extraction efficien...…”

“…When sampling atmospheric particles, artifacts are typically due to collection of gases on a sampling substrate or volatilization of sample already collected. Determining the impact of an artifact on the measurement may be complicated (e.g., Chow, 1995;Hering and Cass, 1999;Solomon et al, 2000;Turpin et al, 2000;Kim et al, 2001;Pang et al, 2002;Ashbaugh et al, 2004aAshbaugh et al, , 2004bSubramanian et al, 2004;Chow et al, 2005aChow et al, , 2005bKim et al, 2005;White et al, 2005;Yu et al, 2006;Watson et al, 2009;Chow et al, 2010;Maimone et al, 2011). Sampling artifacts are primarily associated with semivolatile components in organic carbon and ammonium nitrate.…”

“…Examples of these include use of a single QFF backup filter behind the primary QFF, such as IMPROVE uses, or use of a QFF behind a Teflon filter; use of a denuder (e.g., activated-carbon-impregnated strips or XAD-coated glass annular denuder) in front of a reactive backup filter (e.g., carbonor XAD-impregnated glass-fiber filter) behind the primary QFF, and application of a linear regression method where OC (y-axis) is plotted versus PM 2.5 mass as obtained on a Teflon filter (Peters et al, 2000;Turpin et al, 2000;Solomon et al, 2000;Lewtas et al, 2001;Tolocka et al, 2001;Subramanian et al, 2004;Kim et al, 2005;Watson et al, 2009;Chow et al, 2010;Maimone et al, 2011). Other statistical methods also have been investigated for CSN (Kim et al, 2001).…”