Applications of Open-Path Fourier Transform Infrared for Identification of Volatile Organic Compound Pollution Sources and Characterization of Source Emission Behaviors

By Lin, Chitsan Liou, Naiwei; Sun, Endy

ABSTRACT An open-path Fourier transform infrared spectroscopy (OP- FTIR) system was set up for 3-day continuous line-averaged volatile organic compound (VOC) monitoring in a paint manufacturing plant. Seven VOCs (toluene, m-xylene, p-xylene, styrene, methanol, acetone, and 2-butanone) were identified in the ambient environment. Daytime- only batch operation mode was well explained by the time-series concentration plots. Major sources of methanol, m-xylene, acetone, and 2-butanone were identified in the southeast direction where paint solvent manufacturing processes are located. However, an attempt to uncover sources of styrene was not successful because the method detection limit (MDL) of the OP-FTIR system was not sensitive enough to produce conclusive data. In the second scenario, the OP- FTIR system was set up in an industrial complex to distinguish the origins of several VOCs. Eight major VOCs were identified in the ambient environment. The pollutant detected wind-rose percentage plots that clearly showed that ethylene, propylene, 2-butanone, and toluene mainly originated from the tank storage area, whereas the source of n-butane was mainly from the butadiene manufacturing processes of the refinery plant, and ammonia was identified as an accompanying reduction product in the gasoline desulfuration process. Advantages of OP-FTIR include its ability to simultaneously and continuously analyze many compounds, and its long path length monitoring has also shown advantages in obtaining more comprehensive data than the traditional multiple, single-point monitoring methods.

INTRODUCTION

Since the 1970s, open-path Fourier transform infrared spectroscopy (OP-FTIR) systems have been commercially developed and used in the area of air pollutant monitoring. 1 Because of its good sensitivity for certain species, simultaneous and continuous monitoring of many species, 2 and its long monitoring path length, many successful applications of OP-FTIR have been presented in the literature,3 including power stations, petrochemical plants, refineries, hazardous waste treatment facilities, biomass burning, agriculture emissions, roadway emissions, and metropolitan air quality studies.

In a concentrated swine production facility, Childers et al.4 successfully demonstrated OP-FTIR for the measurement of ammonia, methane, carbon dioxide, and nitrous oxide. With the measurement made along different monitoring paths, the authors were able to identify the confinement barns as a significant source of ammonia emission, whereas the waste treatment lagoon was a major source of methane. At an Air Force base in United States, Hall5 applied OP- FTIR at a wastewater treatment facility for identification of fugitive volatile organic compound (VOC) emission sources and to estimate emission rates. At a municipal landfill in Sweden, Galle et al.6 took advantage of the real-time analysis capability of FTIR absorption spectroscopy to estimate inhomogeneous methane emissions from a landfill, and successfully demonstrated advantages of FTIR over traditional point-measurement methods by providing detection over large sampling areas. Another similar research was conducted to monitor methane and carbon dioxide emitted from a landfill in northern Taiwan.7 Thorn et al.8 utilized OP-FTIR to measure phosphine concentrations in the air surrounding the large fumigated structures of a tobacco warehouse and reported advantages of obtaining more comprehensive data than could be obtained with multiple single-point detection devices. In Southeast Asia, Walter et al.9 and Kagann et al.10 successfully demonstrated the application of OP-FTIR for the measurements of air quality criteria pollutants (ozone [O^sub 3^], carbon monoxide [CO], sulfur dioxide [SO^sub 2^], and nitrogen dioxide [NO^sub 2^]) in ambient air.

Advantages of OP-FTIR include its ability to simultaneously and continuously analyze many compounds, and its long path length monitoring has also shown advantages over traditional single-point monitoring methods. Therefore, it is a very useful real-time monitoring tool. The literature indicates that OP-FTIR has been widely used for methane, CO, carbon dioxide (CO^sub 2^), SO^sub 2^, nitrous oxide (NO), and ammonia monitoring, mainly because these species have unique absorptions in the infrared (IR) wavelength range, and thus have good detection sensitivity and less interference from water moisture. However, applications of OP-FTIR in the area of VOC monitoring were shown less often.5,8,10-15 Recently OP-FTIR was introduced into Taiwan because of the need to monitor VOCs.15-17 In this paper, we present some of our experiences utilizing OP-FTIR to monitor VOC emissions in chemical manufacturing processes. Two scenarios are discussed: one in a paint manufacturing plant, and the other in a lubricant manufacturing process.

EXPERIMENTAL METHODS

The OP-FTIR system utilized in this research was the RAM 2000 Remote Air Monitoring System, manufactured by AIL System, Inc., consisting of a FTIR unit, a retroreflector, and an operational computer with RMMSoft software. AIL Systems’ FTIR is composed of an IR source, a standard mid-band mercury cadmium telluride (MCT) detector, and a telescope. A liquid nitrogen automatic filling system was equipped to maintain the MCT detector at the temperature of 77 K. The retroreflector consists of a panel of gold-coated precision corner cube mounted in a protective housing and placed at the end of the measurement path facing the telescope.

The FTIR data were analyzed by using a multilevel classical least- square (CLS) method in combination with reference IR spectra generated from the MIDAC database. The modulated signal was an interferogram, which was Fourier transformed to produce a spectrum from 700 to 4000 cm-1 at a resolution of 0.5 cm-1. A total of 128 interferograms were co-added for each spectrum, and on average a spectral data was produced in every 2 min. Wave numbers of detection are listed in Tables 1 and 2 for the species identified. For each monitoring scenario, a 3-day continuous monitoring was performed.

In addition, a moveable meteorological tower was set up at a suitable place in the studying area, and data including ambient air temperature, relative humidity, barometric pressure, wind speed, and wind direction were recorded to assist in interpretation of the origins of VOCs.

In the first scenario, the OP-FTIR was set up 3 m above ground on the road close to the process zone to assess VOC emissions of the paint and varnish manufacturing plant, as indicated in Figure 1. The single path length in between the IR source and the retroreflector was 109 m. The decision of path length location depended on the local prevailing wind direction, topography, and the distribution of potential emission facilities. As shown in Figure 1, on the right side of the optical path length were the active manufacturing units, where M1 is an acrylic resin manufacturing process, M2 is a colorful steel coating process, M3 is a paint solvent factory, M4 is an anti- rust paint manufacturing process, and M5 is an anti-fouling paint manufacturing process. On the left side of the optical path length were mainly storage warehouses (SA), M6 is a water paint manufacturing unit, and M7 is a minor paint solvent blending operation. The plants use pigments, resins, and organic solvents as raw material that pass through high-speed dispersion, grinding, mixing and packaging processes to produce various types of can- packaging paints. Fugitive emissions were expected from the preassembly, dispersing, finishing, and packaging units. Major air pollution control devices included bag filters and activated carbon adsorption units.

In the second scenario, the OP-FTIR was set up 3 m above the ground in the canyon of the lubricant manufacturing facility that is located in the middle (Figure 2) of multiple petrochemical manufacturing plants to assess the origins of VOC emissions. As shown in Figure 2, the lubricant manufacturing facility is surrounded by a butadiene manufacturing facility, a gasoline desulfuration facility, a petroleum coke manufacturing facility, and the storage tank area. According to Kaohsiung’s Department of Environmental Protection,22 the lubricant manufacturing company had the third most fugitive VOC emissions (415 t/yr) among the listed petrochemical manufacturing plants, and its nearby residential area recorded the most foul odor complaints.

RESULTS

Performance of OP-FTIR at the Paint Manufacturing Plant

After signal processing work, 10 compounds (Table 1) emerged as the identifiable main species. On the basis of an emissions inventory, methane, O3, and ammonia were not expected to be emitted from the paint manufacturing processes; therefore, these data will not be discussed for this scenario. Emission of toluene, m-xylene, p- xylene, styrene, methanol, acetone, and 2-butanone are all expected from the paint manufacturing processes. Table 1 summarizes the performance statistics of the OP-FTIR monitoring data for the paint manufacturing study. It was partly cloudy with scattered rain during the 3-day continuous monitoring period. Therefore, 295 spectral data were not produced, or 13.6% of the 2160 expected spectral data for the 2-min data resolution. As will be discussed later, percentages of detectable points for some species were not satisfied. Nevertheless, the continuous monitoring dataset provides ample valuable information. Using toluene as an example, there were 1865 data points, consisting of 1858 (99.6%) nondetectable (below field detection limit) points and 7 (0.38%) detectable points. The number of detectable points was related to the process background conditions and the field detection limits. The low percentage of detectable points indicated that OP-FTIR was relatively less sensitive to toluene, as were other aromatic species (m-xylene, p- xylene, and styrene), corresponding to their high minimum reported concentrations. As for the cases of methane, methanol, acetone, 2- butanone, and ammonia, the percentage of detectable points was much higher, therefore more information was provided and better representation was expected. Note that the field detection limits of the OP-FTIR monitoring system were corresponding to the field conditions such as moisture content and the particulate light scattering effect. Therefore, the field detection limits of the measurement could not be accurately determined and could only be approximated to the minimum reported concentrations listed in Table 1. Because of the variety of field detection limits, it was also decided not to take the nondetectable points into account when calculating the averaged concentration presented in Tables 1 and 2.

Also shown in Table 1 are the maximum concentrations reported during the 3-day monitoring period. Except for methanol and acetone, the maximum concentrations of other species were higher than their low odor threshold values, indicating that strong solvent odors are expected in plants during the worst condition and related pollution prevention practices should be enhanced. For the cases of toluene, m- xylene, p-xylene, styrene, and 2-butanone, the averaged concentrations (which did not take the nondetectable points into account) were still much higher than their odor threshold values. Therefore, related odor problems should be expected during the routine production processes and should be treated as priority pollutants for odor combat in the paint manufacturing plant. Nevertheless, the averaged and maximum concentrations did not exceed the time-averaged threshold limit values (TLVs); therefore the occupational exposures were in compliance for this particular example. Moreover, it is worth noting that the recorded maximum concentration of toluene was 3166 ppb, which exceeds the 2-ppm maximum allowable boundary concentration regulated by the Taiwan Environmental Protection Agency (EPA) stationary air pollutant emission standards. The above results indicate that OP-FTIR can be of great assistance in both industrial hygiene and environmental air pollutant regulatory enforcement.

Temporal Characterization of the Paint Manufacturing Activities

Temporal characterization of the paint manufacturing activities is better depicted using a 10-min averaged spectra data. As shown in Figure 3, methanol, 2-butanone, and m-xylene showed similar time- series patterns. It is clear that high concentrations emerge during daytime and diminish at night. The results correspond to the daytimeonly operation of the plant.

The multiple-mode concentration behavior agrees with the batch operation behavior of the plant. Using data from day 2 as an example, concentrations of all three species increased quickly after the beginning of daytime office hours, and the first peak concentrations arose between 8:00 a.m. and 9:00 a.m. 2-Butanone and m-xylene demonstrated an even closer relationship; their second peak concentrations appeared around 11:00 a.m., whereas the third peak concentration was shown at around 4:00 p.m. It is also clear that 2- butanone and m-xylene concentrations were not elevated during the lunch hour (12:00 to 1:00 p.m.) of the first and second day. The above results correspond well with the three daily batch operation activities of the plant. Methanol is mainly used as a solvent for cleaning reactor tanks, whereas 2-butanone and m-xylene are base solvents to produce paints. It is clearly shown in Figure 3 that 2- butanone and m-xylene exhibited very similar time-series patterns, indicating that they were likely emitted from similar production lines where 2-butanone is used as a co-solvent to dissolve flaxseed oil into m-xylene.18 However, some discrepancies were observed on the first and third day between 4:00 p.m. and 12:00 a.m., in which m- xylene concentrations were not shown. The most likely explanation was that m-xylene concentrations were below the field detection limit (approximated with the minimum reported concentration of 92 ppb in Table 1) during the period, thus not reported. This finding reveals one of the system’s major challenges, if the field detection limits are not improved, the system can only be used in plant areas where concentrations are higher, but not in urban or residential ambient air quality studies with lower concentrations.

Spatial Characterization of the Paint Manufacturing Facilities

Taking advantage of the continuous monitoring capability of OP- FTIR and the meteorological data recorded, one can distinguish sources of emissions. The pollutant detected wind-rose percentage plots in Figure 1 show that methanol, 2-butanone, acetone, m- xylene, and toluene are mainly from southeast (SE) windward sources that correspond to manufacturing processes M3, M4, and M5, which utilize them as cleaning or base solvents. It is also clear that M7 (solvents preparation/mixing process) releases significant amount of methanol, 2-butanone, and acetone from the west-southwest (WSW) direction. For the case of acetone, in addition to the SE and WSW sources, a new source of emission was identified from the east- northeast (ENE) direction. This was not an unusual event, because acetone is used more extensively in M2 for blending-tank cleaning purposes.

It is of merit to mention that styrene was used as one of the major solvents along with methyl methacrylate (MMA) to produce acrylic resin in M1. Styrene is a suspected carcinogenic compound,19 and its low odor threshold value (5 ppb)20 often irritates the nearby neighborhood. In the study planning stage, we were hoping to find evidence of styrene coming from the north-northeast (NNE) direction (M1). However, only four data points were reported above field detection limit during the 3-day monitoring period. Therefore, no conclusive result could be drawn. As discussed earlier, for some aromatic compounds such as styrene, the detection limit of OP-FTIR may not have sufficient sensitivity. Alternatively, a point sampling method such as the U.S. Environmental Protection Agency’s TO- 1521,23 can be used in conjunction with OP-FTIR for monitoring these compounds.

Data Performance of OP-FTIR in the Petrochemical Complex

In the second scenario, the OP-FTIR was set up in a lubricant manufacturing plant to assess VOC emissions in the petrochemical complex (Figure 2). The single path length in between the IR source and the retroreflector was 83 m. The plant uses heavy oil as raw material and goes through vacuum distillation, air stripping, and a drying process to produce lubricant and base oil. Few VOC emissions are expected in these processes, and if produced, they should be collected and treated in a flare.

After signal processing work, eight major VOCs (Table 2) emerged as the identifiable species, including ethylene, propylene, toluene, 2-butanone, propane, n-butane, methanol, and ammonia. Table 2 summarizes the performance statistics of the OP-FTIR monitoring data. It was sunny and less humid during the 3 days and 8-hr continuous monitoring period. Therefore, all 2388 spectral data were produced for the 2-min data resolution. This was an example in which weather conditions favored the operation of OP-FTIR, whereas at the first scenario was not.

Using ethylene as an example, the total effective data were 2388 points, consisting of 1738 (73%) nondetectable (below detection limit) points and 650 (27%) detectable points. The number of detectable points was much higher than in the first scenario. Moreover, the minimum concentrations detected in Table 2 were lower than that of in Table 1, taking toluene, 2-butanone, methanol, and ammonia as examples. Therefore, better sensitivity and more explainable results could be provided by the OP-FTIR in a favorable weather condition that has less interference.

Also shown in Table 2 are the concentration statistics reported during the 3-day monitoring period. Except for toluene, the averaged concentrations of all other species were lower than their low odor threshold values, indicating that strong solvent odors are not expected in the plant during regular operations. Furthermore, the maximum concentrations recorded did not exceed the TLVs.

Source Identification of VOCs in the Petrochemical Complex

Among the eight VOCs monitored in Table 2, ethylene, propylene, and methanol were the major pollutants that showed unique time- series patterns (Figure 4). It is clear that ethylene and propylene have similar behavior, indicating that they originated from the same source, whereas methanol was emitted from different sources in the opposite wind direction. However, on the first and second day between 4:00 p.m. and 8:00 a.m. a different pattern was observed in comparison with ethylene and propylene. A possible explanation is that the ambient concentration of propylene was approximately 5 to 10 times lower than that of ethylene. Therefore, during this period propylene concentrations may have been too low to be recorded. It is also possible that the minimum reported concentration of ethylene (13 ppb) was lower than that of propylene (27 ppb); therefore, more ethylene data points were detected (27.2%) than for propylene (5.3%), as shown on Table 2. Alternatively, one can also reason that a minor amount of ethylene was emitted together with methanol from a northern wind direction. Figure 2 shows the pollutant-detected wind- rose percentage plots for the corresponding VOCs. It is clear that ethylene, propylene, 2-butanone, and toluene mainly originated from the southward direction where the tank storage area of the petroleum refinery is located. On the other hand, n-butane mainly originated from the northward direction where the butadiene manufacturing process of the refinery is located. Ammonia also originated from the northward direction, and its origin is assumed to be the gasoline desulfuration plant. This was expected because ammonia is an accompanying reduction product of hydrogen sulfide in the desulfuration process. Finally, propane originated from both the south and north, and methanol originated from various sources from the SSE to the north. Therefore, no definite conclusions could be made for the identification of propane and methanol sources.

CONCLUSIONS

An OP-FTIR system was set up for continuous VOC monitoring with the purpose of characterizing source emission behavior and to identify emission sources. In the paint manufacturing plant study, OP-FTIR successfully identified toluene, m-xylene, acetone, 2- butanone, methanol, styrene, and p-xylene as the process-related emission species. Continuous on-scene monitoring produces highquality time-series concentration plots that correspond well with the daytime-only batch production activities of the plant. With the help of wind rose plots we were able to distinguish the specific manufacturing process responsible for certain VOC emissions. In the second scenario, OP-FTIR was set up to assist in distinguishing sources of certain VOCs in a multi-industrial complex, and the origins of many VOCs were reasonably uncovered. Weather played an important role in producing high-quality OPFTIR monitoring data, favoring clear and less humid conditions. The OP-FTIR real-time continuous VOC line monitoring technique allowed monitoring and identification of emission sources of selected VOCs.

IMPLICATIONS

With the help of meteorological data, the OP-FTIR system was shown to be an effective tool to depict spatial variations in identifying sources of VOC emissions. The continuous monitoring capability of OP-FTIR also provides useful time-series concentration data to depict temporal emission behavior. Such information can be very useful for engineers to identify and control VOC emissions.

REFERENCES

1. Gosz, J.R.; Dahm, C.N.; Risser, P.G. Long-Path FTIR Measurement of Atmospheric Trace Gas Concentrations; Ecology 1988, 69, 1326-1330.

2. Herget, W.F.; Brasher, J.D. Remote Fourier Transform Infrared Air Pollution Studies; Opt. Eng. 1980, 19, 508-514.

3. Bacsik, Z.; Mink, J.; Keresztury, G. FTIR Spectroscopy of the Atmosphere Part 2. Applications; Appl. Spectroscopy Rev. 2005, 40, 327-390.

4. Childers, J.W.; Thompson, E.L.; Harris, D.B.; Kirchgessner, D.A.; Clayton, M.; Natschke, D.F.; Phillips, W.J. Multi-Pollutant Concentration Measurements around a Concentrated Swine Production Facility Using Open-Path FTIR Spectrometry; Atmos. Environ. 2001, 35, 1923-1936.

5. Hall, F.E., Jr. Case Study: Environmental Monitoring Using Remote Optical Sensing (OP-FTIR) Technology at the Oklahoma City Air Logistics Center Industrial Wastewater Treatment Facility; Fed. Facilities Environ. J. 2004, 15, 21-37.

6. Galle, B.; Samuelsson, J.; Svensson, B.H.; Borjesson, G. Measurements of Methane Emissions from Landfills Using a Time Correlation Tracer Method Based on FTIR Absorption Spectroscopy; Environ. Sci. Technol. 2001, 35, 21-25.

7. Hegde, U.; Chang, T.C.; Yang, S.S. Methane and Carbon Dioxide Emissions from Shan-chu-ku Landfill Site in Northern Taiwan; Chemosphere 2003, 52, 1275-1285.

8. Thorn, T.G.; Marshall, T.L.; Chaffin, C.T. Open-Path FTIR Air Monitoring of Phosphine around Large Fumigated Structures; Field Anal. Chem. Technol. 2001, 5, 116-120.

9. Walter, W.T.; Perry, S.H.; Han, J.S.; Park, C.J. Open-Path FTIR Ozone Measurements in Korea; Proc. SPIE 1999, 3534, 133-139.

10. Kagann, R.H.; Wang, C.D.; Chang, K.L.; Lu, C.H. Open-Path FTIR Measurement of Criteria Pollutants and Other Ambient Species in an Industrial City; Proc. SPIE 1999, 3534, 140-149.

11. Carter, R.E., Jr.; Thomas, M.J.; Marotz, G.A.; Lane, D.D.; Hudson, J.L. Compound Detection and Concentration Estimation by Open- Path Fourier Transform Infrared Spectrometry and Canisters under Controlled Field Conditions; Environ. Sci. Technol. 1992, 26, 2175- 2181.

12. Ross, K.R.; Todd, L.A. Field Evaluation of a Transportable Open-Path FTIR Spectrometer for Real-Time Air Monitoring; Appl. Occup. Environ. Hyg. 2002, 17, 131-143.

13. Russwurm, G.M.; Kagann, R.H.; Simpson, O.A.; McClenny, W.A.; Herget, W.F. Long-Path FTIR Measurements of Volatile Organic Compounds in an Industrial Setting; J. Air & Waste Manage. Assoc. 1991, 41, 1062-1066.

14. Tsai, S.Y.; Chen, J.D.; Chao, W.Y.; Wang, J.D. Neurobehavioral Effects of Occupational Exposure to Low-Level Organic Solvents among Taiwanese Workers in Paint Factories; Environ. Res. 1997, 73, 146-155.

15. Lin, C.; Liou, N.; Chang, P.E.; Yang, J.C.; Sun, E. Fugitive Coke Oven Gas Emission Profile by Continuous Line Averaged Open- Path FTIR Monitoring; J. Air & Waste Manage. Assoc. 2007, 57, 472- 479.

16. Hong, Y.J.; Jeng, H.A.; Gau, Y.Y.; Lin, C.; Lee, I.L. Distribution of Volatile Organic Compounds in Ambient Air of Kaohsiung, Taiwan; Environ. Monit. Assess. 2006, 119, 43-56.

17. Jeng, H.A.; Lee, I.L.; Gau, Y.Y.; Yang, C.T.; Lin, C.; Hong, Y.J. Changes in Immunological and Hematological Parameters of Female Residents Exposed to Volatile Organic Compounds in the City of Kaohsiung, Taiwan; J. Environ. Health 2006, 69, 20-25.

18. Yang, S.L. Industrial Chemical Processes; Five-State Publisher: Taipei, Taiwan, 1981.

19. Group 2B: the Agent Is Possibly Carcinogenic to Humans, Vol. 60, 233, Styrene; CAS No. 100-42-5; International Agency for Research on Cancer (IARC): Lyon Cedex, France, 1994.

20. Haz-Map: Occupational Exposure to Hazardous Agents; U.S. National Institutes of Health; available at http:// hazmap.nlm.nih.gov (accessed August 1, 2007).

21. Compendium Method TO-15: Determination of Volatile Organic Compounds (VOCs) in Air Collected in Specially Prepared Canister and Analyzed by Gas Chromatography/Mass Spectrometry (GC/MS); U.S. Environmental Protection Agency; Center for Environmental Research Information; Office of Research and Development: Cincinnati, OH, 1997.

22. Annual (2004-2005) Volatile Organic Compounds Management Plan of Kaohsiung City; Prepared for Kaohsiung City Department of Environmental Protection, Kaohsiung, Taiwan, DEPA-A-94-01-03 by Environmental Science Corporation: Kaohsiung, Taiwan, 2005.

23. Lin, C.; Shern, C.C.; Huang, C.L.; Kuo, F.L. Spatial Characterization of VOC Air Quality in an Industrial and Residential Complex; J. Environ. Eng. Manage. 2006, 16, 413-419.

Chitsan Lin and Naiwei Liou

Department of Marine Environmental Engineering, National Kaohsiung Marine University, Kaohsiung, Taiwan, Republic of China

Endy Sun

Environmental Science Corporation, Taipei, Taiwan, Republic of China

About the Authors

Chitsan Lin is an associate professor at the National Kaohsiung Marine University (NKMU) in Kaohsiung, Taiwan. Naiwei Liou was a graduate student at NKMU during this study, and is now a doctoral student in the Environmental Institute of the National Sun Yet-Shan University in Kaohsiung, Taiwan. Endy Sun is CEO of Environmental Science Corporation in Taipei, Taiwan. Please address correspondence to: Chitsan Lin, Ph.D., Department of Marine Environmental Engineering, National Kaohsiung Marine University, 142 Haijhuan Road, Nanzih District, Kaohsiung 81143, Taiwan, Republic of China; phone: +886-7-3651472; fax: +886-7-3651472; e-mail: ctlin@mail.nkmu.edu.tw.

Copyright Air and Waste Management Association Jun 2008

(c) 2008 Journal of the Air & Waste Management Association. Provided by ProQuest Information and Learning. All rights Reserved.