Measurement of fine particulate matter nonvolatile and semi-volatile organic material with the Sunset Laboratory Carbon Aerosol Monitor.

The combination of technology used in the High-Volume Brigham Young University Organic Sampling System (BIG BOSS (18)) and the Harvard particle concentrator (19) has resulted in the PC-BOSS. (4,18-20) The configuration and operation of the PC-BOSS as used in the studies reported here has been previously described. (11) The PC-BOSS was used for sample collection to determine fine particulate mass, sulfate, carbonaceous material (elemental and organic), nitrate, SVOM, and semi-volatile nitrate. Samples for the chemical characterization of [PM.sub.2.5] in the minor flow following a particle concentrator and a BOSS diffusion denuder were collected in a filter pack containing a pre-fired (800 [degrees]C) 47-mm quartz filter (Pallflex) followed by 47-mm CIG fiber filter to determine fine particulate carbonaceous material and nitrate, including semi-volatile species lost from the particles during sampling. A second parallel filter pack containing a 47-mm Teflon (Whatman) filter followed by a 47-mm Nylon (Gelman, Nylasorb) filter was used to determine [PM.sub.2.5] filter retained (nonvolatile) mass, sulfate and nitrate, including any nitrate lost from the particles during sample collection. A side flow filter pack, before the particle concentrator, containing a 47-mm polycarbonate (Corning, Nuclepore, 0.4-[micro]m pore size) filter followed by a 47-mm CIG filter, was used to collect particles (excluding semi-volatile species lost during sampling) and gas-phase organic material after the 2.5-[micro]m outlet cut. These data were compared with data from the minor flow filters to determine the particle concentrator and denuder efficiencies. The quartz filter nonvolatile OC (NVOC) and EC, and CIG filter SVOC, determined with the PC-BOSS, were used for comparison with the results obtained with the two Sunset monitors.

Temperature Programmed Volatilization (4,21) was used in the analysis of PC-BOSS collected samples for total carbonaceous material. In this method, the various sample filters are heated from ambient temperature at a predetermined ramp rate to a predetermined termination temperature. The ramp rate and termination temperatures are dependent on the type of filter being analyzed. Quartz filters are heated to 800 [degrees]C in an [N.sub.2]/[O.sub.2] atmosphere. Charcoal-impregnated filters are heated to 400 [degrees]C in an [N.sub.2] atmosphere. Carbon in compounds desorbed from the filters during the heating process is catalytically converted to [CO.sub.2] and detected by NDIR absorption.

The PC-BOSS results provided a comparison for the data obtained with the two Sunset monitors because total fine particulate carbonaceous material, including the NVOC, SVOC, and EC, are also measured with this instrument. A disadvantage of the PC-BOSS is that the limits of detection of this sampler were higher than for the Sunset monitors. Therefore, 3-hr integrated data were obtained with the PC-BOSS.

Sample Collection

Initial studies were conducted in February 2003 in Lindon, UT. The Lindon sampling site has been previously described. (11) In these experiments, results obtained with the modified Sunset monitor were compared with fine particulate SVOC determined in 6-hr integrated samples with the PC-BOSS. More extensive studies were conducted during July 2003 at the South Coast Air Quality Management District (SCAQMD) sampling site in Rubidoux, CA. In these experiments, both the conventional and modified Sunset monitors were simultaneously used. Results were compared with 3-hr integrated PC-BOSS sample results.

RESULTS AND DISCUSSION

Efficiency of the Sunset Monitor Denuder for Removal of Gas-Phase Organic Material

The analysis of material collected by the CIG filter in the modified Sunset instrument is illustrated in Figure 1B. On the basis of data obtained in several studies with a charcoal-impregnated filter using a parallel plate denuder (4,5,8,9,14,18,20) the initial peak in the data is due to the small fraction of the gas-phase organic material entering the denuder that is not removed by the denuder, and this material is well separated from SVOC lost from particles. For the sample given in Figure 1B, this initial peak is about half the size of the larger peakin the 220-450 [degrees]C temperature step. The gas-phase organic material collected on the side-flow filter of the PC-BOSS is a measure of the total gas-phase organic material that enters the diffusion denuder of the Sunset instrument. The initial peak in Figure 1B is the amount of this total that is not removed by the denuder. Comparison of these two values gives the denuder efficiency. Initially the efficiency of the denuder was better than 99.5%. However, as observed in the data obtained over a 3-week period in Lindon, the size of the first peak, relative to the SVOC peak, increases with time as the efficiency of the denuder drops to approximately 98% over this time period. The data illustrated in Figure 1B are after approximately 2 weeks of sampling with the denuder. To help maintain high denuder efficiency, new charcoal-impregnated cellulose fiber filter strips (Schleicher and Schuell) were placed in the denuder about every 10 days during the Lindon study. These strips are cut to size for the denuder box and used without further treatment as received from the manufacturer. Denuder strips were not changed during the Rubidoux study because the study duration of 4 weeks was not expected to result in a significant degradation of denuder efficiency on the basis of previous studies. (14)

Determination of Fine Particulate Carbonaceous Material with the Sunset Monitors

The two Sunset monitors used at Rubidoux were intended to allow for the semicontinuous measurement of total fine particulate carbonaceous material, nonvolatile carbonaceous material (including both nonvolatile OC and EC) being determined on the instrument with the quartz filter, and fine particulate SVOC lost from the particles collected on the quartz filter being determined on the instrument with the CIG filter. The data obtained with the modified Sunset instrument using a CIG filter are illustrated in Figure 1B and have been discussed in the proceeding section. Typical data obtained with the conventional Sunset monitor are illustrated in Figure 1A. As shown, the organic carbonaceous material is initially evolved in four temperature steps in a He atmosphere. Any pyrolyzed carbon and EC are then removed in an [O.sub.2]/He atmosphere, with the amount of EC identified from the point at which the measured absorbance returns to the original value for the sample. Although the data obtained with the Sunset instrument will allow determination of the amount of carbon evolved in each of the temperature steps and the identification of the concentrations of EC, only total carbonaceous material is used in the comparisons made in this paper.

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

The concentrations of nonvolatile and SVOC determined with the two instruments are illustrated with concentrations determined over a 7-day period in the Rubidoux study in Figure 2. As illustrated, the results indicate that SVOC was a significant fraction of the total fine particulate material during the study. Concentration of SVOC and nonvolatile quartz filter carbon were both high and showed strong (but different) diurnal patterns during the first 4 days of this period. Concentrations were much lower during the last 2 days.

The accuracy in the measurement of SVOC, NVOC, and total [PM.sub.2.5] C (calculated as the sum of NVOC, SVOC, and EC) as illustrated by the data in Figure 2, can be evaluated by comparison of the measurement obtained with the modified Sunset instrument and the SVOC, NVOC, and total [PM.sub.2.5] C (also calculated as the sum of NVOC, SVOC, and EC) measured using the PC-BOSS as described above. These two measurements for each species are compared for the 3-hr integrated time periods over which the PC-BOSS measurements were made in Figure 3 and the linear regression statistics are given in Table 1. The larger number of data points in Figure 3 for SVOC as compared with NVOC results from two time periods when the Sunset quartz data (NVOC) were not valid. As illustrated, measurements for total [PM.sub.2.5] C made by the two instruments agreed. Linear regression analysis of the data given in Figure 3 (N = 21, [R.sup.2] = 0.92) gives a zero intercept slope of 0.99 [+ or -] 0.02. The slope obtained with a fit intercept is similar, 0.90 [+ or -] 0.06. The precision of the comparison is [sigma] = [+ or -]1.5 [micro]g C/[m.sup.3] (8%), and the average measured concentration of [PM.sub.2.5] C is 18.8 [micro]g C/[m.sup.3], with no bias between the two measurements. The precision of the comparison is similar to the expected precision of the PC-BOSS 3-hr integrated measurement of [+ or -]1-2 [micro]g [PM.sub.2.5] C/[m.sup.3], on the basis of the blank filter reproducibility and the uncertainty in the flow data, and also on the basis of results of past studies. (14,18,20) Thus, this comparison indicates that the [PM.sub.2.5] C measurements by the two independent techniques are equivalent. During the time period when SVOC was measured, the SVOM (assuming SVOM was 61% C) (22) averaged 21% of the total [PM.sub.2.5], as previously reported for measurements by an R & P filter dynamics measurement system tapered element oscillating microbalance. (23)