Environ. Sci. Technol. 2007, 41, 6748-6754
Tracking Petroleum Refinery Emission Events Using Lanthanum and Lanthanides as Elemental Markers for PM2.5 P R A N A V K U L K A R N I , †,| S H A N K A R A R A M A N C H E L L A M , * ,†,‡ A N D MATTHEW P. FRASER§ Department of Civil and Environmental Engineering, University of Houston, Houston, Texas 77204-4003, Department of Chemical and Biomolecular Engineering, University of Houston, Houston, Texas 77204-4004, and Department of Civil and Environmental Engineering, Rice University, Houston, Texas 77005
Fine particulate matter levels at four air sampling stations in the Houston, TX area are apportioned to quantify the impact of emissions from a local refinery during a reported emission event. Through quantification of lanthanum and lanthanides using a recently developed analytical technique, the impacts of emissions from fluidized-bed catalytic cracking (FCC) units are quantitatively tracked across the Houston region. The results show a significant (33-106fold) increase in contributions of FCC emissions to PM2.5 compared with background levels associated with routine operation. This impact from industrial emissions to ambient air quality occurs simultaneously with a larger, regional haze episode that lead to elevated PM2.5 concentrations throughout the entire region. By focusing on detailed chemical analysis of unique maker metals (lanthanum and lanthanides), the impact of emissions from the FCC unit was tracked from the local refinery that reported the emission event to a site approximately 50 km downwind, illustrating the strength of the analytical method to isolate an important source during a regional haze episode not related to the emission event. While this source apportionment technique could separate contributions from FCC emissions, improved time-resolved sampling is proposed to more precisely quantify the impacts of transient emission events on ambient PM2.5.
Introduction Emission events refer to the periodic release of material to the atmosphere from industrial activities due to nonroutine operation. Because of their short duration and unexpected nature, emission events are difficult to capture and characterize quantitatively, even though they can influence emission inventories for geographically large areas (1-4). In Houston, TX, a significant effort has focused on the emission * Corresponding author phone: (713)743-4265; fax: (713)743-4260; e-mail:
[email protected]. † Department of Civil and Environmental Engineering, University of Houston. ‡ Department of Chemical and Biomolecular Engineering, University of Houston. § Rice University. | Present address: Trinity Consultants, Houston, Texas. 6748
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 19, 2007
events for volatile organic compounds (VOCs) and the impact of these emissions on ozone formation (5). However, episodic emissions of fine particulate matter (PM) from industrial sources can also occur and impact local air quality. Previous research on episodic PM events has predominantly focused on using organic markers to track PM emission and transport related to wood combustion (6-8) or motor vehicle emissions and road dust suspension (9-13). However, some emissions can only be effectively tracked using elemental markers such as soil resuspension using Al, Si, Ca, Mg, Fe, and Ti (14) and oil combustion using V and Ni (15) as well as catalytic refining of petroleum using lanthanum and lanthanides (2, 16)salso referred to as the rare earth elements (REEs) in this manuscript. Petroleum refining is an important industrial activity in the Houston, TX region which is home to approximately 25% of the U.S. refining capacity. Along with a positive economic impact, this aggregation of refining and petrochemical manufacturing also potentially affects the environment and public health negatively. For example, epidemiological studies have revealed an increased incidence of chronic and acute respiratory health effects, eye irritation, sleep problems, headaches, etc. in communities living in the proximity of refineries (17). Understanding how the various types of emissions, including gases and particles, from an integrated refining and petrochemical facility impact public health is not straightforward. Vital to delineating the relationship between emission and health impact is improved methods for source attribution. Research has demonstrated that emissions of zeolite catalysts are the primary cause of lanthanum and lanthanides enrichment in atmospheric PM2.5 (1, 2, 16). Exposure to REEs has been shown to cause cytotoxicity and acute lung toxicity and inhibit enzyme activity in animals (18-20). Additionally, high levels of REEs in humans have been shown to adversely impact bioelectrical activity of the central nervous system, liver, and kidney, increase the incidence of atherosclerosis at the fundus of the eye, and reduce intelligence quotients of children (21-23). For this reason, our research group has been tracking routine, continuous PM2.5 emissions from FCC units in refineries in the Houston area using REEs as elemental markers (2). In contrast to the low but consistent emissions from routine refining operations, emission events result in the release of large quantities of FCC catalysts and may periodically result in a greater elevation of local PM2.5 levels. For example, emission reports estimate that 57 kg of PM2.5 and 1535 kg VOCs were released from an oil refinery located in the Houston area on September 2-3, 2005 due to a malfunctioning wet gas compressor (24). More recently, a break in the secondary cyclone of the catalytic cracking unit caused the release of an estimated 45 tons of catalyst to the local atmosphere on August 1-2, 2006 (25). This episode was also estimated to release 39 kg of VOCs and several organic hazardous air pollutants (e.g., 1, 3-butadiene and benzene). The reported large amounts of PM released during these emission events motivate monitoring nonroutine REE emissions in addition to routine operation. The objective of this research is to quantitatively track PM2.5 releases from the FCC unit of a petroleum refinery to assess the contribution of such events to ambient PM2.5. REE concentrations from four PM2.5 monitoring sites in the Houston area following the emission event that occurred on August 1-2, 2006, were used to calculate the fraction of PM2.5 mass directly attributable to FCC emissions. During this same period, a widespread regional haze episode resulted in 10.1021/es062888i CCC: $37.00
2007 American Chemical Society Published on Web 08/28/2007
TABLE 1. a. TEOM Sampling Informationa and b. Average Wind Speed and Direction Data during the Reported Emission Event at the Four Sites Considered a. TEOM Sampling Information TEOM sampling time sampling station period (day) (latitude, longitude) Channelview (+29.802500, -95.125556) Kingwood (+30.058333, -95.189722) East Houston (+29.767778, -95.220556) Clinton Drive (+29.733611, -95.257500)
sampling duration (h)
07/30/2006-08/08/2006 11/01/2006-11/21/2006 07/16/2006-08/03/2006 08/04/2006-08/28/2006 07/15/2006-08/07/2006 08/30/2006-09/20/2006 07/17/2006-08/03/2006 08/04/2006-08/15/2006
240 528 456 600 576 528 432 288
b. Average Wind Speed and Direction Data resultant average wind wind speed direction during event during event sampling station (km/h) (deg)
FIGURE 1. A map depicting the locations of the source of the catalyst emission event (petroleum refinery) and four TEOM sampling stations in the greater Houston, TX area. elevated PM2.5 concentrations throughout the entire region. By focusing on specific metal markers (REEs) for emissions from FCC units, we demonstrate that the impact of the industrial emission event can be isolated from the more widespread regional increase in PM2.5. The research will also compare these elevated PM2.5 levels apportioned to FCC emissions to background levels during routine operation at local refineries.
Methods PM2.5 Sampling. Samples of PM2.5 collected on filters used in Tapered Element Oscillating Microbalance (TEOM) instruments were obtained from four continuous ambient monitoring (CAM) sites including Clinton Drive, Channelview, East Houston, and Kingwood (Figure 1). The Kingwood site is located near a residential area, whereas the other three sites are located near the heavily industrialized Houston Ship Channel. Table 1a summarizes geographical information on the sampling locations and duration. Samples immediately following the reported emission event (bold lettering) as well as those capturing routine operations at local refineries characterizing background REE levels (regular font) were analyzed. Because these sites are operated by Texas Commission on Environmental Quality (TCEQ) and the City of Houston primarily for monitoring PM2.5 mass concentrations as required for compliance with the National Ambient Air Quality Standards (NAAQS), filters are not archived unless an emission event is suspected. For this reason, a background sample immediately before or after the catalyst emission event (see next section) were not available at Channelview. The samples were separated into those influenced by the emission event and background samples based on the reported emission period. It should be noted that these samples were integrated over 9-24 days, encompassing the short emission event as well as a longer period of background emissions. The wind direction and velocity data for the period coinciding with the reported emission event are given in Table 1b. Metals Analysis. All samples and field blanks (blank filters for individual sites) were analyzed using our recently
(latitude, longitude) Channelview (+29.802500, -95.125556) Kingwood (+30.058333, -95.189722) East Houston (+29.767778, -95.220556) Clinton Drive (+29.733611, -95.257500)
7.84 not available
150° not available
6.58
161°
13.63
170°
a Information corresponding to samples that captured the emission event is written in bold font. Background samples are written in regular font.
developed method capable of trace to ultratrace level quantitation of REEs in PM2.5 (1, 2). Briefly, metals were first extracted from the filters using a two-stage process employing HF, HNO3, and H3BO3 in a microwave oven at 200 °C and 200 psig with a 20 min dwell time for each stage. The digestate was later analyzed for REEs and non-REEs by inductively coupled plasma-mass spectrometry (ICP-MS) with internal standardization using 115In. Non-REEs analyzed included Na, Mg, Al, Si, K, Ti, V, Mn, Fe, Co, Ni, Cu, As, Zn, Se, Rb, Sr, Cd, Cs, Pb, Th, and U. Catalyst Emission Period. Ambient PM2.5 mass concentrations in each of the local air quality monitoring sites from July 31, 2006-August 4, 2006 are depicted in Figure 2 (26). The data show that PM2.5 levels at many local sites were between 5 and 10 µg/m3 prior to the emission period and increased to as much as 40 µg/m3 coinciding with the reports of catalyst emissions from the cracking unit beginning at 5:30 a.m. on August 1, 2006. As PM2.5 levels are simultaneously elevated at air quality monitoring locations to the north, west, and south of the FCC emission site, a widespread regional haze episode occurred simultaneously with the FCC emission report. The resultant wind direction during the episode, as shown in Table 1b, was generally from the south, transporting the PM releases from the refinery source to the sampling stations to the north of the emission site. Subsequent chemical analysis (reported in this manuscript) can be used to isolate the contribution of loss of zeolite catalyst from the FCC unit from the regional contribution to elevated PM2.5 levels. The period of impact from the FCC emission event must be defined as the samples of ambient PM2.5 were collected over a period much longer than the FCC emission event. To conservatively estimate the impact at local monitoring sites, the period of elevated PM2.5 due to the regional haze episode (∼2 days from midday August 1 to midday August 3 as shown in Figure 2) will be used as the VOL. 41, NO. 19, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6749
FIGURE 2. The hourly elevated PM2.5 concentrations during the catalyst emission event at the continuous ambient PM2.5 monitoring stations. period of impact from the FCC emission event. It is unlikely that short-lived emissions from the FCC unit occurred beyond this period, as local PM2.5 levels decrease to levels similar to those measured before the haze episode and simultaneous report of release of material from the FCC unit. Source Apportionment. It has been demonstrated that the ambient REEs in PM2.5 predominantly arise from loss of catalyst in locations impacted by refinery operations (2, 16), and, therefore, these elements can be used as markers for emissions from the FCC operation. Therefore, during a catalyst emission event where REE emissions from refining operations are orders of magnitude greater than during routine operation, it is reasonable to assume that refinery sources overwhelm any other contributing sources to ambient REEs. As a result, the mass balance of REEs during the emission event is expected to follow
REEambient PM2.5 ) SFCC contributionFFCC catalyst + intercept (1) where F is the gravimetric concentration of REEs in FCC catalysts (ng/ng) and S, the slope of the best-fit line, is the estimate of the contribution of FCC emissions to ambient PM2.5. The average source profile of the six FCC catalysts based on the same analytical technique is F (1, 2). The intercept is analogous to the error term in chemical mass balance calculations (28) and should be negligible if primary refinery emissions are the major cause of enhancements in PM levels as expected during a catalyst emission event. Seven rare earth elements including La, Ce, Pr, Nd, Sm, Gd, and Dy, which together comprised more than 99% REE mass in FCC catalysts, were used to track emissions from FCC sources to the ambient environment. La alone constituted 50-80% of the total REE mass in each of these catalysts thereby dominating the unweighted least-squares regression. As a result, even though La in each of the six catalysts was accurately and precisely quantified by ICP-MS, weighted linear regressions were performed to isolate the impact of disproportional weighting on one element (La). A weighting factor inversely proportional to the precision of individual ambient REE concentrations was assigned to each elemental concentration based on the unequal standard deviation of 3-5 replicate ICP-MS measurements for each element (heteroscedastic error) (29). The simple linear regression approach used in this study has limitations: other potential sources of ambient PM were not included, and additionally only REE species were used as tracers (i.e., chemical mass balance (CMB) was not performed). Hence small uncertainties in REE ambient concentrations and source profile may result in significant changes in predicted source contributions and therefore should be cautiously interpreted. Alternatively, inclusion of 6750
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 19, 2007
FIGURE 3. Enrichment in REE concentrations (filled symbols) during the catalyst emission event. Routine or background REE concentrations are depicted as hollow symbols. Tb, Ho, Er, Tm, Yb, and Lu background concentrations are depicted as their ICP-MS detection limits because they were not detected in the corresponding samples. Also, Lu was detected only at Channelview during the emission event. REE data in CMB analysis would be essential to prevent colinearity of FCC and soil profiles because non-REEs in soil and catalyst may not be distinguishable. Unlike REEs, high and fluctuating background concentrations of nonrare earth elements disallow their isolation during an FCC emission event from the time integrated samples. Given that only recently ICP-MS methods have become widely available for REE quantification in PM source samples, linear regression of REEs was selected as the preferred source apportionment method to determine FCC emissions.
Results and Discussion REE Enrichment in PM2.5 during the Catalyst Emission Event. Since sample analysis relied on filters collected from the continuously operating TEOM instruments, sample collection intervals were significantly longer than the event duration. Hence, even samples encompassing the catalyst emission episode included several days wherein REEs were emitted at background (i.e., nonemission event) levels. To correctly attribute REE concentrations to the FCC emissions period, background levels over the extended sample period were subtracted. As observed in Figure 3, background REE concentrations in fine PM for all four sites monitored in this study were very similar and substantially lower than in samples collected during the FCC emission period. Further, background REE concentrations measured in this study agree very closely with earlier measurements at a separate air quality
FIGURE 4. Nonparametric depiction of La to light lanthanide ratios during the catalyst emission event (n ) 4) and at the background levels (n ) 4). Data for FCC catalysts (n ) 6), soil (n ) 5), and motor vehicle emissions (n ) 6) were obtained from our previous work (2, 33). The box encompasses the 25th and 75th percentiles, and the whiskers are determined by the fifth and 95th percentiles. The horizontal line inside the box is the median. and the square within the box denotes the average. Crosses (×) denote first and 99th percentiles, and the maximum and minimum values are represented by dashes (-). monitoring site in Houston in the summer of 2001 (2). Because background levels were highly consistent and low, they could be subtracted from the total loading over the extended sampling period to accurately estimate the REE concentrations during the emission event. This allowed source apportionment of the FCC contribution during the shorter emission event period using eq 1. Figure 3 shows that REE concentrations during the episode were 2-200fold higher than routine levels, demonstrating their enrichment during the episode. It should also be noted that several heavy REEs present in FCC catalysts at low levels (Ho, Er, Tm, Yb, and Lu) were not detected in the background samples but were present in elevated levels during the emission episode. This further demonstrates the common source of the REE enrichment during the emission episode. Error bars in Figure 3 during the emission event were obtained by propagating ICP-MS measurement errors in both background samples and samples influenced by the emission event by adding individual variances (30). REE and non-REE concentrations along with respective uncertainties in all TEOM samples are given in Tables S1 and S2, respectively, of the Supporting Information. Confirmation that Loss of the FCC Catalyst Is the REE Source for PM2.5 during the Episode. The relative abundance sequence of dominant REEs in ambient PM2.5 during the emission event and in FCC catalysts were identical (La > Ce > Nd > Pr > Gd > Sm > Dy) and the same as earlier measurements of REEs in PM samples from 2001 (2). Also, enrichment factors of these elements in ambient PM2.5 during the emission event with respect to FCC catalyst samples were all close to unity (0.7-2.1) when normalized by Nd. (The choice of Nd for the reference metal has been explained in ref 1.) Additionally, as shown in Figure 4, the ratios of La to Ce, Pr, Nd, and Sm for the four samples collected during the catalyst emission event were in the same range as those for
the background samples and in FCC catalysts as well as being considerably different from those in the only two other possible REE sources (local soil and vehicle emissions). Finally, Varimax rotation of the elemental data (including REEs, V, Mn, Fe, Co, Ni, Cu, As, and Pb) separated a factor with high loadings of La, Ce, Pr, Nd, Sm, and Gd. Therefore, REE source signatures were preserved during the episode confirming that the source of elevated REE concentrations in ambient PM2.5 (shown in Figure 3) is from emissions of the FCC catalyst. Hence, similar to recent analysis of routine emissions (2), REEs are the critical markers for FCC catalyst emissions during the emission event, justifying the use of eq 1 to estimate the source contribution. Contributions of FCC Operations to PM2.5 during the Catalyst Emission Event. FCC source contributions for each of the four sites were obtained through a weighted linear regression of seven dominant REEs (La, Ce, Pr, Nd, Sm, Gd, and Dy) measured in ambient PM2.5 samples during the emission episode period as shown in Figure 5 and summarized in Table 2. In all the cases, excellent correlation coefficients were obtained (0.80 < R2 < 0.99) with near zero intercept values demonstrating that eq 1 is capable of isolating the FCC source contribution. The highest contribution was estimated for Channelview, which is closest to the source (see Figure 1) and located downwind of the reported emission site based on TCEQ meteorological data. The lowest source contribution was calculated for Clinton Drive, which was west of, and not directly downwind, the reported emission event. Intermediate source contributions obtained for East Houston and Kingwood were consistent with the atmospheric transport from the reported emission site to the north and northwest as predicted by the average wind direction during the episode. The ambient PM2.5 mass concentration during the emission period was similar at all four sites investigated by detailed VOL. 41, NO. 19, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6751
FIGURE 5. Weighted linear regression of ambient REE concentrations during the catalyst emission episode against the average FCC catalyst profile on a mass basis adopted from ref 2. REE uncertainties are not seen because they are smaller than the symbol size.
TABLE 2. Summary of FCC Source Contributions to Ambient PM2.5 Mass Calculated Using the Simplified Apportionment Method (Eq 1) sampling station Channelview Kingwood East Houston Clinton Drive
PM2.5 mass apportioned to the FCC sourcea (µg/m3) 15.89 ( 0.03 0.15 ( 0.00b 4.03 ( 0.05 0.11 ( 0.00 7.79 ( 0.05 0.10 ( 0.00 3.61 ( 0.02 0.11 ( 0.00
a Note that the bold font corresponds to samples during the catalyst emission event, whereas the regular font corresponds to routine emissions for the sampling durations shown in Table 1a. The numbers after ‘(’ are the standard deviation of the slope in regression analysis. b Zero accurate to two decimal places.
chemical speciation. This indicates that the regional haze episode, leading to elevated PM2.5 concentrations at locations upwind of the emission site such as Deer Park and Galveston, has a greater contribution to measured PM2.5 levels compared to localized emissions from the FCC emission event. The lack of an increase in the measured PM2.5 mass at Channelview despite the high apportioned FCC contribution as compared to other sites cannot be readily explained without source attribution (i.e., CMB). However, as described earlier, CMB could not be performed solely for the short period of the FCC emission event using time integrated TEOM samples. An improved approach to better resolve the contribution of transient FCC emission events would include a more timeresolved collection of ambient PM for detailed elemental 6752
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 19, 2007
speciation. Nonetheless, concentration of several marker metals (Na, Mn, Fe, Ba, Zn, As, Se, V, and Cu) representing sources other than the FCC catalyst were found in lower concentrations in a time integrated Channelview sample as compared to three other sites as shown in the Supporting Information (Table S2). A CMB analysis with 13 metal markers (including 4 REEs) and 7 sources was also performed using a time integrated (240 h) sample (see Supporting Information). The FCC catalyst emission accounted for a significant amount (12%) of the total apportioned PM2.5 mass at the Channelview site. Note that this would translate into a much higher contribution during the short time duration emission event since the background FCC contribution is typically 1-2% of total PM2.5 mass. Table 2 summarizes FCC contributions to ambient PM2.5 during the FCC emission episode (in bold) as well as during background days (in regular font). As seen, FCC emissions contributed only 0.10-0.15 µg/m3 to PM2.5 levels during background days (written in regular font) comprising only 1% of the average mass. This agrees quantitatively with the analysis of samples from another air quality monitoring site in Houston from 2001 (2). In contrast, the FCC source contributed substantially to PM2.5 levels during the catalyst emission episode. In other words, the impact of emissions from refining operations on ambient PM2.5 levels during the emission event was 33-106-fold greater than PM2.5 levels in background samples. During the emission event, a catalyst from refining operations comprised a significant fraction of the PM2.5 mass indicating that primary emissions of a catalyst from the malfunctioning refinery affected air quality over a widespread area. In addition to the increases in ambient PM mass concentration, results presented herein demonstrate that
emission events from refining operations release large quantities of metals including rare earths. Hence, air quality in the neighborhood of petroleum refineries needs to be closely monitored to evaluate both chronic and acute exposures of humans to REEs to accurately isolate their adverse health effects. Also, to better characterize source contribution and human exposure, it may be necessary to obtain samples integrated over short time periods for speciating REEs in PM. Further, analytical procedures reported in this manuscript are perfectly suited to monitor the emissions from refining operations in cities with a high concentration of industrial activity and can be used to augment existing regulatory programs based solely on industry self-reporting. Interestingly, X-ray fluorescence (XRF) analysis of PM samples during the catalyst emission event at the Clinton Drive site also revealed elevated concentrations of Al, Si, and Fe (2.75, 4.89, and 2.00 µg/m3, respectively) with all REEs below detection limits (32). It should be noted that in addition to REEs, zeolite catalysts are enriched in Al, Si, and Fe (1, 2, 16). Because these elements are typically used to track sources of crustal material including road dusts, the catalyst emission event could potentially be mistakenly attributed to road dust or soil resuspension. This motivates analysis of trace-level REEs, available only through ICP-MS to augment ambient monitoring of Al, Fe, and Si commonly measured by XRF. Therefore, XRF measurements in industrialized urban environments should be supplemented with ICP-MS to better characterize and identify anthropogenic emissions in the form of industrial emission events.
Acknowledgments This project has been funded entirely with funds from the State of Texas as part of the program of the Texas Air Research Center. The contents do not necessarily reflect the views and policies of the sponsor nor does the mention of trade names or commercial products constitute endorsement or recommendation for use. We thank Dr. Wei-Yeong Wang from the Bureau of Air Quality Control, City of Houston and Ms. Marsha Hill of TCEQ for providing air quality samples.
Supporting Information Available Source apportionment of PM2.5 collected at Channelview using CMB (Figure S1) and rare (Table S1) and nonrare (Table S2) earth elements composition in ambient PM2.5 samples. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Kulkarni, P.; Chellam, S.; Mittlefehldt, D. W. Microwave assisted extraction of rare earth elements from petroleum refining catalysts and ambient fine aerosols prior to inductively coupled plasma -mass spectrometry. Anal. Chim. Acta 2007, 581 247259. (2) Kulkarni, P.; Chellam, S.; Fraser, M. P. Lanthanum and lanthanides in atmospheric fine particles and their apportionment to refinery and petrochemical operations in Houston, TX. Atmos. Environ. 2006, 40, 508-520. (3) Kleinman, L. I.; Ryan, W. F.; Daum, P. H. S.; R.;, S.; Lee, Y.-N.; Nunnermacker; J. L.; Weinstein-Lloyd, J. An ozone episode in the Philadelphia metropolitan area. J. Geophys. Res., [Atmos.] 2004, 109, D20302/1-D20302/17. (4) Murphy, C. F.; Allen, D. T. Hydrocarbon emissions from industrial release events in the Houston- Galveston area and their impact on ozone formation. Atmos. Environ. 2005, 39, 3785-3798. (5) Nam, J.; Kimura, Y.; Vizuete, W.; Murphy, C.; Allen, D. Modeling the impacts of emission events on ozone formation in Houston, Texas. Atmos. Environ. 2006, 40, 5329-5341. (6) Buzcu, B.; Yue, Z. W.; Fraser, M. P.; Nopmongcol, U.; Allen, D. T. Secondary particle formation and evidence of heterogeneous chemistry during a wood smoke episode in Texas. J. Geophys. Res., [Atmos.] 2006, 111 (D10), D10S13/1-D10S13/14.
(7) Wu, C. F.; Jimenez, J.; Claiborn, C.; Gould, T.; Simpson, C. D.; Larson, T.; Liu, L. J. Agricultural burning smoke in Eastern Washington: Part II. Exposure assessment. Atmos. Environ. 2006, 40, 5379-5392. (8) Fraser, M. P.; Lakshmanan, K. Using Levoglucosan as a Molecular Marker for the Long-Range Transport of Biomass Combustion Aerosols. Environ. Sci. Technol. 2000, 34, 4560-4564. (9) Choi, Y.-J.; Hyde, P.; Fernando, H. J. S. Modeling of episodic particulate matter events using a 3-D air quality model with fine grid: Applications to a pair of cities in the US/Mexico border. Atmos. Environ. 2006, 40, 5181-5201. (10) Gertler, A.; Kuhns, H.; Abu-Allaban, M.; Damm, C.; Gillies, J.; Etyemezian, V.; Clayton, R.; Proffitt, D. A case study of the impact of winter road sand/salt and street sweeping on road dust reentrainment. Atmos. Environ. 2006, 40, 5976-5985. (11) Mysliwiec, M. J.; Kleeman, M. J. Source Apportionment of Secondary Airborne Particulate Matter in a Polluted Atmosphere. Environ. Sci. Technol. 2002, 36, 5376-5384. (12) Schauer, J. J.; Fraser, M. P.; Cass, G. R.; Simoneit, B. R. T. Source Reconciliation of Atmospheric Gas-Phase and Particle-Phase Pollutants during a Severe Photochemical Smog Episode. Environ. Sci. Technol. 2002, 36, 3806-3814. (13) Chow, J. C.; Watson, J. G.; Zhiqiang, L.; Douglas, H.; Frazier, C. A.; Solomon, P. A.; Thuillier, R. H.; Mangliano, K. Descriptive analysis of PM2.5 and PM10 at regionally representative locations during SJVAQS/AUSPEX. Atmos. Environ. 1996, 30, 2079-2112. (14) Alastueya, A.; Querola, X.; Castilloa, S.; Escudero, M.; Avilab, A.; Cuevasc, E.; Torresc, C.; Romeroc, P.-M.; Expositod, F.; Garcia, O.; Diazd, J. P.; Van Dingenene, R.; Putaude, J. P. Characterisation of TSP and PM2.5 at Izana and Sta. Cruz de Tenerife (Canary Islands, Spain) during a Saharan Dust Episode (July 2002). Atmos. Environ. 2005, 39, 4715-4728. (15) Watson, J. G.; Zhu, T.; Chow, J. C.; Engelbrecht, J.; Fujita, E. M.; Wilson, W. E. Receptor modeling application framework for particle source apportionment. Chemosphere 2002, 49, (9), 1093-1136. (16) Kitto, M. E.; Anderson, D. L.; Gordon, G. E.; Olmez, I. Rare earth distributions in catalysts and airborne particles. Environ. Sci. Technol. 1992, 26, 1368-1375. (17) Luginaah, I. N.; Taylor, S. M.; Elliott, S. J.; Eyles, J. D. Community reappraisal of the perceived health effects of a petroleum refinery. Social Sci. Med. 2002, 55, 47-6. (18) Nakamura, Y.; Tsumura, Y.; Tonogai, Y.; Shibata, T.; Ito, Y. Differences in behavior among the chlorides of seven rare earth elements administered intravenously to rats. Fundam. Appl. Toxicol. 1997, 37, 106-116. (19) Highsmith, S. R.; Head, M. R. Tb3+ Binding to Ca2+ and Mg2+ Binding Sites on Sarcoplasmic Reticulum ATPhase. J. Biol. Chem. 1983, 258, 6858-6862. (20) Wen, H. R.; Chen, R. S. Study of biologic effects of rare earth ions; John Liberty Eurotext: Montrouge, 1998; Vol. 5, pp 199-203. (21) Zhu, W. F.; Xu, Q.; Shao, P. P.; Zhang, H.; Wu, D. S.; Yang, W. J. Bioelectrical activity of central nervous system among populations in a rare earth element area. Biol. Trace Elem. Res. 2000, 73, 1-17. (22) Zhang, H.; Feng, J.; Zhu, W.; Liu, C.; Xu, S.; Shao, P.; Wu, D.; Yang, W.; Gu, J. Chronic Toxicity of Rare Earth Elements on Human Beings. Biol. Trace Elem. Res. 1999, 73, 1-17. (23) Zhang, H.; Feng, J.; Zhu, W. F.; Wu, D. S.; Yang, W. J.; Zhang, H. Investigation on arteriosclerosis among population in a rare earth area in South China. Biol. Trace Elem. Res. 1997, 59, 9398. (24) Texas Commission on Environmental Quality. Reports of Air Emission Events; Report number: 63968; Austin, TX. http:// www.tceq.state.tx.us/compliance/field_ops/eer (accessed 12/ 01/2006). (25) Texas Commission on Environmental Quality. Reports of Air Emission Events; Report number: 79352; Austin, TX. http:// www.tceq.state.tx.us/compliance/field_ops/eer (accessed 12/ 01/2006). (26) Texas Commission on Environmental Quality. PM2.5 Data: Soot, Dust, Smoke (Fine Particulate Matter). http://www.tceq.state.tx.us/nav/data/pm25.html (accessed 12/01/2006). (27) Russell, M.; Allen, D. T.; Collins, D. R.; Fraser, M. P. Daily, seasonal, and spatial trends in PM2.5 mass and composition in Southeast Texas. Aerosol Sci. Technol. 2004, 38, (SUPPL 1), 1426. (28) Watson, J. G.; Robinson, N. F.; Fujita, E. M.; Chow, J. C.; Pace, T. G.; Lewis, C.; Coulter, T. CMB8 applications and validation protocol for PM2.5 and VOCs; 1808.2D1; Desert Research Institute: Reno, NV, 1998. VOL. 41, NO. 19, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6753
(29) Miller, J. N.; Miller, J. C. Statistics and Chemometrics for Analytical Chemistry, 4th ed.; Harlow, England, New York, 2000. (30) Massart, D. L.; Vandeginste, B. G. M.; Buydens, L. M. C.; De Jong, S.; Lewi, P. J.; Smeyers-Verbeke, J. Handbook of Chemometrics and Qualimetrics: Part A; Elsevier Science: Amsterdam, The Netherlands, 1997. (31) Fraser, M. P.; Yue, Z. W.; Buzcu, B. Source apportionment of fine particulate matter in Houston, TX, using organic molecular markers. Atmos. Environ. 2003, 37, (15), 2117-2123. (32) Sullivan, D. W. Interim Report on Research on PM2.5 Concentrations at Houston Clinton Drive CAMS 403 submitted to TCEQ;
6754
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 19, 2007
Center for Energy and Environmental Resources at The University of Texas at Austin: Austin, TX, 2006. (33) Chellam, S.; Kulkarni, P.; Fraser, M. P. Emissions of organic compounds and trace metals in fine particulate matter from motor vehicles: a tunnel study in Houston, TX. J. Air Waste Manage. Assoc. 2005, 55, (1), 60-72.
Received for review December 5, 2006. Revised manuscript received April 30, 2007. Accepted June 20, 2007. ES062888I
