obtaining more information from existing filter samples in

A closer look at fine particulate matter data samples taken from long-term

speciation networks.

Obtaining More Information from Existing Filter Samples in PM

Speciation Networks by Judith C. Chow, Xiaoliang Wang, Mark C. Green, and John G. Watson

The Lava Beds National Monument in northern California, where high levels ofcarbon (including black and brown) were recorded after a 2018 regional fire.

em • The Magazine for Environmental Managers • A&WMA • May 2019

Analyzing National PM Speciation Network Data by Judith Chow et al.


The United States has operated the long-term non-urban

Interagency Monitoring of PROtected Visual Environ-

ments (IMPROVE) network since 1987 and the urban

Chemical Speciation Network (CSN) since 2000. The 2017

Annual A&WMA Critical Review and Discussion papers1-3

describe these networks and provide references for greater

detail. Solomon et al.4 summarizes similarities and differences

for these networks and changes in methodology that oc-

curred during their long-term operation. A limited number of

analyses are routinely performed on these samples, although

modern laboratory technology now permits more specific

speciation to determine source contribution and evaluate

adverse environmental effects.

Fine particulate matter (PM2.5; particles with diameters 2.5

micrometers and smaller), is collected on Teflon-membrane,

quartz-fiber, and nylon-membrane filters to quantify mass

concentrations by gravimetry, elements by x-ray fluorescence

(XRF), carbon fractions (including organic and elemental

carbon [OC and EC]) by thermal/optical reflectance (TOR)

analysis, and water-soluble ions by ion chromatography (IC),

as illustrated in the sidebar on the following page, Collecting,

Analyzing, and Storing PM Samples. The 2017 Critical

Review Discussion Paper noted the need for long-term

consistency of these methods to detect trends that indicate

the effectiveness of emission reduction strategies, as illus-

trated in Figure 1. In addition to these trends, the data from

these networks have been used to: 1) identify and quantify

source contributions using receptor models; 2) improve

emission inventories; 3) validate urban- and regional-scale

source models; and 4) elucidate atmospheric processes.

PM2.5 speciation data can be downloaded in comma-

delimited formats5,6 for any or all of these uses.

The Critical Review also indicated complementary analyses

that can be applied to the archived filters or water extracts to

better address these goals:

• IC with pulsed amperometric and conductimetric detection

can examine the water extract residual for bioaerosol,


Figure 1. Trends in annual CSN PM2.5 compositions at Fresno, CA. Chemical components indicate where emissions reductions are needed. Combustion sources, such as biomass burning, engine exhaust, and formation of secondary organic aerosols, are causes of high organic matter (OM = 1.4 x OC) levels. NH3 and NOx emissions are causes ofelevated ammonium and nitrate levels. Sulfate was previously reduced with strict SO2 emission controls in California’s central valley. The downward trend in these values from 2004 to 2011 indicates the effectiveness of these controls. Recent years show greater variability, possibly due to wet vs. dry climates and wildfires.



Collecting, Analyzing, and Storing PM Samples

Gravimetric analysis is non-destructive and weighs the Teflon-membrane filter before and after sampling in a controlled environmentto determine PM2.5 mass concentration.

X-ray fluorescence is non-destructive and bombards the Teflon-mem-brane filter with high energy x-ray electromagnetic radiation that re-moves an inner shell electron, thereby emitting a lower energy photon when outer shell electrons replace the missing inner-shell electron. The wavelength of this energy indicates the element and thenumber of photons is proportional to the elemental concentration.

Ion chromatography acts on the destruction of the nylon-membranefilter by extracting it in ultrapure distilled-deionized water. This extractthen passes through an ion exchange column that separates the anions or cations in time for detection of their electrical conductivity.The retention time identifies the ion and the level of conductivity determines the concentration.

Thermal/optical reflectance carbon analysis submits a small punchfrom the quartz-fiber filter to step-wise heating in inert (He), then oxidizing (98% He, 2% O2) environments to determine OC, EC, andthermal carbon fractions that are indicative of sources. Laser light reflectance and transmittance are monitored to determine and correctfor OC charring. Since 2016, these optical measurements have beenmade at seven wavelengths ranging from 405 nm to 980 nm, permitting estimation of brown carbon (BrC), commonly found in thesmoldering phase of biomass combustion. A different method was usedin the CSN prior to 2009.

Sample Storage is in sealed Petri slides in a refrigerated environmentto preserve sample integrity. Although stored at the analysis facilities,these Teflon-membrane and quartz-fiber filter remnants are the property of the state agencies participating in the CSN and can be retrieved for more complex analyses when a request is made to EPA.



spectrum of absorption properties for the aerosol deposit, as

illustrated in Figure 2.

Biomass burning produces light absorbing OC, termed

“brown carbon” (BrC) in their aggregate, that are indicative

of their contributions. Per unit of black carbon, which is

present to some extent in all combustion samples, the

smoldering peat and burning pine needles in Figure 2

absorb light more strongly at shorter wavelengths than do

the particle emissions from diesel-engine exhaust. This

difference permits identification of the biomass burning

contribution.7 While smoke from residential wood combus-

tion is a long-established wintertime PM2.5 contributor in

many areas, and is being effectively controlled through stove

design, fuel specifications, and burning bans, non-winter

contributions from ever increasing wildfires are more difficult

to identify with currently available data. These contributions

are important when defining natural visibility conditions

related to the Clean Air Visibility Rule,8 formerly the Regional

Haze Rule, and in exempting exceptional events from

compliance determination.

Figure 3 shows an example of black carbon (BC) and BrC

contributions during a Northern California wildfire that

adversely affected air quality across the region. Lava Beds

National Monument (NM) is located in far northeastern

California, close to the Oregon border. The Carr fire started

on July 23, 2018 approximately

150 km southwest of Lava Beds

NM and burned over 90,000

hectares before containment oc-

curred on August 30, 2018. Ele-

vated OC and EC concentrations

were sustained from July 24 to

August 24 with a total carbon

(TC = OC+EC) concentration

reaching 50 µg/m3. For much of

this time period the Carr Fire

was active, but little or no growth

of the Carr fire area was occur-

ring toward the end of August.

Averaged over the summer, BC

and BrC absorption contributed

about 12 Mm-1 and 10 Mm-1 to

total absorption at Lava Beds

NM, respectively. On five days

(July 4, 25, and 31 and August

24 and 30), all light absorption

was attributed to BrC, indicative

of the biomass contribution.

biomass burning, and secondary organic aerosol markers.

The procedures are the same as those for the inorganic

ions with different columns, detectors, and calibrations.

• Thermal desorption-gas chromatography/mass spectrometry

extracts organic species through heating of a small quartz-

fiber filter section and directs ions with a mass spectrometer.

These spectra yield concentrations of polycyclic aromatic

hydrocarbons (PAH) and other markers for fossil fuel and

biomass combustion sources.

• Inductively coupled plasma-mass spectrometry (ICP-MS)

complements, rather than replaces, XRF by detecting

rare-earth elements and lead isotopes that may differentiate

among fugitive dust source contributions. This method is

applied to a strong acid extract of the Teflon-membrane

filter that leave the sample unavailable for other analyses.

• Ultraviolet/Visible spectrometer measures light transmission

through the Teflon-membrane filter that illustrate different

wavelength-specific absorption characteristics among

fossil fuel combustion, biomass burning, and fugitive

dust sources.

Related to the last technique is the multiwavelength thermal/

optical analysis that has been applied in IMPROVE and CSN

since January 2016.7 This is a minor modification to the

method applied previously that does not affect the long-term

OC and EC quantification. However, with the same proce-

dures and analysis times used before, the method provides a

Figure 2. Absorption spectra of different source emissions sampled on Teflon-membrane filters and analyzed with an Ultraviolet/Visible spectrometer. Light absorption through the filter is normalized to absorption at 1,000 nm, which is dominated by the EC component in each of the source emissions.

While these contributions are

currently expressed as contribu-

tions to light absorption, work is

underway to develop source-

specific mass absorption

efficiencies for fires that will

translate these into PM2.5

source contributions. There is

growing evidence that second-

ary organic aerosol contains a

substantial BrC component that

could also be determined with

these data.

As the obvious PM2.5

emission sources become

more completely controlled, it

will be necessary to develop

more specific markers for the

remaining sources that cause

exceedances of National

Ambient Air Quality Standards.

Filter samples from the

national PM speciation

networks have a high potential

to assist in determining these

contributions. em



Figure 3. Organic and elemental carbon (OC and EC) concentrations, as well as browncarbon (BrC) contributions at Lava Beds National Monument during the 2018 northernCalifornia Carr Fire that destroyed large parts of the city of Redding, CA. Besides theCarr Fire, the Klondike Fire, about 200 km west northwest of Lava Beds NM in Southwestern Oregon from mid-July to November 2018, also impacted the area.

Drs. Judith C. Chow, Xiaoliang Wang, Mark C. Green, and John G. Watson are Research Professors in the Division of AtmosphericSciences at the Desert Research Institute, Reno, NV. E-mail: [email protected].

Acknowledgment:This work was jointly supported by National Science Foundation grants AGS-1464501 and CHE-1214163, and the National Park ServiceIMPROVE Carbon Analysis Contract (P16PC0029). Thanks to Megan Johnson, now at North Carolina State University, for acquiring thedata illustrated in Figure 2.

References1. Hidy, G.M.; Mueller, P.K.; Altshuler, S.L.; Chow, J.C.; Watson, J.G. 2017 Critical Review: Air quality measurements—From rubber bands to tapping the

rainbow; J. Air Waste Manage. Assoc. 2017, 67 (6), 637-668; https://doi.org/10.1080/10962247.2017.1308890.2. Hidy, G.M.; Chow, J.C.; Watson, J.G. A Summary of the 47th Annual A&WMA Critical Review: Air quality measurements—From rubber bands to tapping

the rainbow; EM June 2017; http://pubs.awma.org/flip/EM-June-2017/crsummary.pdf.3. Kleinman, M.T.; Head, S.J.; Morris, R.E.; Stevenson, E.D.; Altshuler, S.L.; Chow, J.C.; Watson, J.G.; Hidy, G.M.; Mueller, P.K. Critical Review Discussion:

Air quality measurements—From rubber bands to tapping the rainbow; J. Air Waste Manage. Assoc. 2017, 67 (11); 1159-1168;https://doi.org/10.1080/10962247.2017.1370309.

4. Solomon, P.A.; Crumpler, D.; Flanagan, J.B.; Jayanty, R.K.M.; Rickman, E.E.; McDade, C.E. U.S. National PM2.5 Chemical Speciation Monitoring Networks-CSN and IMPROVE: Description of networks; J. Air Waste Manage. Assoc. 2014, 64 (13), 1410-1438; https://doi.org/10.1080/10962247.2014.956904.

5. CIRA 2019. Federal Land Manager Environmental Database (FED). Colorado State University, Fort Collins, CO. See http://views.cira.colostate.edu/fed/ (accessed February 15, 2019).

6. EPA 2019. Pre-generated data files. U.S. Environmental Protection Agency, Research Triangle Park, NC. See https://aqs.epa.gov/aqsweb/airdata/download_files.html (accessed February 15, 2019).

7. Chow, J.C.; Watson, J.G.; Green, M.C.; Wang, X.L.; Chen, L.-W.A.; Trimble, D.L.; Cropper, P.M.; Kohl, S.D.; Gronstal, S.B. Separation of brown carbon from blackcarbon for IMPROVE and CSN PM2.5 samples; J. Air Waste Manage. Assoc. 2018, 68 (5), 494-510; https://doi.org/10.1080/10962247.2018.1426653.

8. EPA 2005. Final Clean Air Visibility Rule Will Restore Visibility in America’s National Parks and Wilderness Areas. U.S. Environmental Protection Agency, Research Triangle Park, NC. See https://www.epa.gov/visibility/final-clean-air-visibility-rule-will-restore-visibility-americas-national-parks-and (accessed February 15, 2019).

obtaining more information from existing filter samples in

Documents