Real-World Application of Open-Path UV-DOAS, TDL, and FT-IR Spectroscopy for Air Quality Monitoring at Industrial Facilities

Open-path spectroscopy is known for its ability to provide real-time measurements of dozens of compounds over sampling paths of up to 1000 meters in length. Advances in open-path monitoring technology and data processing techniques, coupled with new regulatory requirements, have greatly increased the acceptance and widespread application of spectroscopy-based open-path measurements. Large industrial facilities adjacent to residential communities are a particular application of interest, because traditional fixed-point analyzers lack the spatial coverage of the open-path instruments. This work discusses technical and practical considerations for the installation and operation of more than 120 open-path analyzers that are currently providing continuous data at several oil refineries in California. Open-path analyzers include ultraviolet differential optical absorbance spectroscopy (UV-DOAS), Fourier transform infrared (FT-IR), and tunable diode laser (TDL) technologies. We will discuss lessons learned from these projects, including fundamental approaches to compound identification, target species detectability, interferences, and data management.

The term fenceline monitoring refers to air quality measurements collected at or around the perimeter of a facility, and is typically performed at industrial facilities with processes that may generate airborne emissions. A fenceline monitoring network often contains a range of instrumentation depending on the facility and monitoring objectives. Figure 1 shows a generic example of a large monitoring network, including fenceline open-path, point, meteorological monitors, and community point monitors.

FIGURE 1: A large-scale monitoring system at a generic petroleum refinery with open-path, point, and meteorological monitors (collocated with fixed- point and community sites).

Motivations for fenceline monitoring often include compliance with regulations, addressing community concerns about fugitive emissions, or consent agreements. Most facilities conducting fenceline monitoring do so to comply with state or local regulations. For example, in 2017, California Assembly Bill 1647 (AB-1647) (1) mandated fenceline monitoring programs at petroleum refineries, to be regulated by local Air Districts. Examples of the local regulations include Regulation 12 Rule 15 (2) from the Bay Area Air Quality Air Management District (BAAQMD), Rule 1180 (3) from the South Coast Air Quality Management District (SCAQMD), and Rule 4460 (4) from the San Joaquin Valley Air Pollution Control District (SJVAPCD). Two smaller air districts—San Luis Obispo and Santa Barbara—also instituted regulations, but there are no active refineries in those areas that currently meet the required production levels. Each of these regulations require facilities to establish air monitoring systems to provide air quality data along refinery boundaries. Similar regulations, such as Colorado House Bill 21-1189 (5), have been passed within the past year, and more are expected in other states with large industrial manufacturing presences. These state and local rules enhance the breadth and depth of other recent federal fenceline regulatory requirements, such as the Refinery Technology Review via the Clean Air Act Section 112(d) that resulted in the widespread application of passive fenceline monitoring for benzene (6).

These industrial monitoring programs typically require real-time monitoring of a range of compounds over large distances, and at detection limits relevant for comparison to health thresholds. The measurements must meet strict uptime requirements, and be displayed on a publicly accessible website shortly after collection, in some cases requiring public alerts during an emission event. This unique combination of requirements requires a robust technical approach, for which open-path spectroscopy is particularly well-suited.

Technical Background

Open-path analyzers measure absorbance of airborne compounds between a light source and a detector. Different compounds have unique absorbance characteristics at discrete wavelengths which correspond to absorption spectra when scanning over a range of wavelengths. The agreement between measured spectra and reference spectra (the spectral match) is used to identify compounds present. Quantification is achieved by scaling of the reference spectra. The appropriate open-path analyzer is selected based on the absorbance wavelengths of the compounds of interest.

A monostatic configuration uses a combined light source and detector at one end of the open sampling path and a reflector located several hundred meters away, thus effectively doubling the path length. Figure 2 shows an example of a monostatic configuration. A bistatic system has the light source and detector on opposite sides of the path, across a single path. Importantly, while an open-path system can measure compounds over a large distance (500+ meters), it cannot distinguish between a concentrated or dispersed plume within that path.

FIGURE 2: Operating principle of a monostatic open-path analyzer.

Open-Path Spectroscopy Techniques

Differences between open-path spectroscopy techniques are primarily due to the light source used, which subsequently determines the available wavelengths for measuring compounds.

UV-DOAS analyzers contain an ultraviolet (UV) light source—commonly a xenon or deuterium lamp, operating in the range of 225 nm to 400 nm. Target compounds for UV-DOAS analyzers include air toxics such as benzene, toluene, ethylbenzene, xylenes, and sulfur dioxide. This instrument simultaneously measures multiple compounds in real time at a time resolution as fast as every 30 sec. Detection limits are down to low ppb and, in some cases, sub-ppb levels. Typically, one-way path lengths range from 300–700 meters for monostatic configurations, and up to 1000 meters for bistatic configurations.

FT-IR analyzers contain an infrared laser operating in the wavelength range of 670 cm^-1 to 4400 cm^-1 at a resolution of 0.5 cm^-1. Target compounds for FT-IR analyzers primarily include short-chain hydrocarbons, methane, aldehydes, ammonia, hydrogen cyanide, and other volatile organic compounds (VOCs). FT-IR analyzers can be used to measure multiple compounds simultaneously in real time at a time resolution of approximately once every two to three minutes. Similar to UV-DOAS, typical one-way path lengths range from 300–700 meters for monostatic configurations. Bistatic configurations are not typically used for FT-IR.

Tunable diode laser (TDL) technology uses a diode laser that emits a narrow wavelength of light tuned to a specified range of interest, which is typically in the mid-IR spectrum. This technique is typically used to measure a single compound of interest, with one or two additional compounds (such as water or methane) measured to verify operations under normal ambient conditions. Example compounds measured by this instrument include hydrogen sulfide, hydrogen cyanide, and hydrogen fluoride.

Minimum Detection Limits (MDLs)

With current regulations (see above) requiring detection of compounds at or below health thresholds, there has been increased focus on how to achieve the lowest possible detection limits for each compound of interest. In addition, the standardization of methodologies to calculate the minimum detection limits (MDLs) has become a growing topic of discussion. Broadly speaking, laboratory-derived MDLs do not represent actual performance in real-world applications, and real-world MDLs are dependent on operating conditions and interference from other similarly absorbing compounds.

Currently, there is no broadly accepted standard for MDL determination with open-path spectroscopy. Two common approaches include the U.S. Environmental Protection Agency (EPA) Compendium of Methods for the Determination of Toxic Organic Compounds in Ambient Air, known as TO-16 (7), and the American Society for Testing and Materials (ASTM) International Standard Test Method for Determination of Gaseous Compounds by Extractive Direct Interface Fourier Transform Infrared (FT-IR) Spectroscopy, known as ASTM D6348-03 (8). Both techniques generally provide similar results, but differ in their technical approach, and therefore implementation.

Operational Considerations

Selection of the instrument type to deploy depends on the compounds of interest. Many fenceline monitoring networks require a combination of open-path technologies to provide comprehensive coverage of many target compounds. TDL open-path analyzers generally provide single-gas measurements, where UV-DOAS and FT-IR can simultaneously monitor dozens of compounds in real time.

The definition of “real-time monitoring” depends on the regulatory driver, or the program goals, but most open-path analyzers can provide real-time measurements every five minutes and MDLs below health thresholds. Typical instrumental time resolutions of UV-DOAS, FT-IR, and TDL open-path analyzers are on the order of 30 sec, 3 min, and 5 min, respectively. If possible, it is preferable to select open-path analyzers that retain raw instrument data independent of calibration, which may be used at any time to verify real-time measurements.

Detectability of target compounds in the presence of atmospheric interferents is an important consideration in designing open-path fenceline monitoring networks. Many compounds can be measured by multiple types of open-path analyzers, but the appropriate open-path technology should be selected to meet the needs of the monitoring program. For example, benzene—a common target compound—can be measured by UV-DOAS and FT-IR open-path spectroscopy, but the two techniques have significantly different detection capabilities. For benzene, the Agency for Toxic Substances and Disease Registry (ATSDR) cites an inhalation reference concentration (RfC), which is an estimate of the concentration that a population can be exposed to without an “appreciable risk” of noncancer health effects, of 9.4 ppb (9). The Refinery Risk and Technology Review regulation cites benzene concentrations down to 2.8 ppb (10), and the California Office of Environmental Health Hazard Assessment (OEHHA) acute benzene REL of 8.6 ppb (11), so applications that require comparison to the applicable health thresholds must be capable of measuring well below them. With an appropriate application of these technologies (in this case UV-DOAS, and not FT-IR), these health-based fenceline concentration limits can be achieved (12).

Figure 3 shows absorbance spectra for benzene in the FT-IR and UV-DOAS spectral analysis regions, which illustrate the differences in detectability for these techniques.

FIGURE 3: Representative benzene absorbance spectra in the (a) FT-IR and (b) UV-DOAS spectral analysis regions.

As an example, for FT-IR analysis of benzene, the maximum usable absorbance units (AU) amplitude is 0.006 AU at 1037 cm^-1 The absorbance feature at 674 cm^-1 is extremely sensitive but is obscured by interfering CO₂ absorbance in ambient long path open-path applications. While FT-IR can be used to monitor benzene, the practical minimum detection limit (MDL) is on the order of greater than 100 ppb for a 300 m path, which is significantly higher than the applicable health thresholds. For UV-DOAS analysis of benzene, the maximum usable absorbance amplitude is 0.26 AU at 253 nm, and the practical MDL is on the order of less than 1 ppb for a 300 m path. Therefore, for applications requiring fenceline monitoring of benzene with comparison to health thresholds, UV-DOAS open-path spectroscopy is the ideal technique.

Some target compounds with few unique spectral features or spectral overlap with atmospheric interferents can be challenging to measure with open-path spectroscopy. An example of this is hydrogen sulfide—associated with a variety of industrial processes—which has an OEHHA (13) Acute Reference Exposure Level (REL) of 30 ppb (also the California 1-hour air quality standard [14]). Hydrogen sulfide can be measured by UV-DOAS open-path spectroscopy, but absorbance spectrum overlap with benzene, and other aromatics may result in false positive detections. FT-IR open-path spectroscopy can be used, but the MDL is commonly too high. The preferred technique for open-path monitoring of hydrogen sulfide is TDL, with high selectivity and lower detection limits compared to other available open-path technologies. Moreover, recent advances in signal processing and wavelength modulation approaches have allowed for enhanced signal-to- noise (S/N) ratios, more robust data post-processing, and lower MDLs (on the order of 25 ppb for a 300 m path). These technological advances have precipitated regulations from local air districts in California to encourage use of TDL for open-path monitoring of hydrogen sulfide.

Table I shows considerations for selecting an open-path technology for hydrogen sulfide monitoring based on selectivity, detectability, time resolution, single-gas or multiple-gas capability, and cost.

Practical Aspects of Implementation

Considerations for installation of open-path analyzers can vary depending on the facility. For example, topography or fenceline perimeter distance typically determines the number instruments needed. Another consideration is whether a facility needs total fenceline coverage, or if partial coverage may be sufficient because of the proximity to local communities and typical wind conditions. Short paths with temporary installation may be achieved with basic tripods for the analyzer components, whereas permanent installations with longer pathlengths require more substantial infrastructure, such as concrete foundations or support pillars. If a monostatic configuration is selected, power and shelter requirements can be reduced to one side of the path, as those requirements for a reflector on the other side are minimal, or in some cases unnecessary. A bistatic configuration requires power and shelters at both ends of the path. In either case, placement of sampling paths can be optimized to minimize the total number of power and shelter installations to reduce costs while meeting the program objectives. As open-path analyzer operations depend on the amount of light that returns to the detector, instrument alignment is a critical component of continuous operations. Monostatic configurations double the effective pathlength, which can be helpful for achieving lower detection limits, but necessitates more frequent alignment. Conversely, bistatic configurations may achieve shorter paths that the light travels and require less-frequent alignment. Regardless, robust and stable infrastructure is critical.

Fenceline monitoring programs that incorporate open-path spectroscopy typically require an up-front capital investment from the facility. Although the per-analyzer cost may be greater for open-path technologies as compared to traditional point monitors, the spatial coverage provided by open-path analyzers is usually comparable to that provided by several point monitors. Moreover, depending on the target compounds and the corresponding point monitor technology, ongoing operations and maintenance activities may provide cost savings over time for open-path installations as compared to point monitors. The cost-benefit analysis of open-path vs. point monitors is a consideration when designing the fenceline monitoring program and will ultimately depend on the goals of the program and the regulatory drivers.

Data Management

All fenceline monitoring programs require a data management system for the ingest, storage, quality control, and dissemination of data. Programs that require data display to the public in real time necessitate additional infrastructure, such as reliable internet connection at each sampling site and an automated quality control approach that can filter unreliable data points in real-time.

Depending on the size of the fenceline monitoring system, data flow can be on the order of 250 data points every five minutes, or 6.5 million data points per calendar quarter. Therefore, a robust data management solution is needed for this scale. Data quality control is typically performed in several ways. On a real-time basis, the data management system performs quality control based on instrument diagnostics to ensure proper operations, and flag data according to their quality indicators. Qualified personnel subsequently review the data on a daily, weekly, monthly, and/or quarterly basis, and may generate reports depending on the requirements of the program. For projects that require real-time display on a publicly accessible website, data must be displayed shortly (10–20 min) after they are collected. This requires rigorous automated quality control to minimize false positive alerts and ensure alerts are issued during potential emission events. Additional analysis tools may be used in conjunction with the data management system, such as source-receptor analysis metrics to assess origination of detections at the fenceline, as well as where potential emissions may travel based on local meteorological conditions.

Conclusion

Over the past five years, over 120 open-path analyzers (along with dozens of meteorological and point monitors for other compounds) have been installed and operated at the fenceline of several large oil refineries in California. Similar programs exist in other states or are under consideration. The open-path monitor types employed include UV-DOAS, FT-IR, and TDL, depending on the target species and site characteristics. These analyzers measure many key pollutants defined by regulatory or other site requirements at detection limits appropriate for the specified health-based standards. The open-path instruments cover large fenceline distances and provide better coverage than traditional point instruments, though point monitors are used for target compounds not suited to be measured by open-path methods. Besides the instrument selection process, robust infrastructure and data processing systems are critical elements due to the high data throughput and demanding public reporting requirements. These operating installations continue to successfully provide continuous data for public and regulatory use.

References

(1) California State Assembly, Assembly Bill No. 1647, Chapter 589, AB-1647 Petroleum refineries: air monitoring systems. Available at https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=201720180AB1647 (Accessed Oct. 4, 2022).

(2) Bay Area Air Quality Management District, Air District Regulation 12, Rule 15: Petroleum Refining Emissions Tracking. Available at https://www.baaqmd.gov/~/media/files/planning-and-research/rules-and-regs/workshops/2015/1215-1216-workshop/1215_amg_-01282015.pdf. (Ac- cessed Oct. 4, 2022).

(3) South Coast Air Quality Management District, Rule 1180: Refinery Fenceline and Community Air Monitoring. Available at http://www.aqmd.gov/docs/default-source/rule-book/reg-xi/r1180.pdf?sfvrsn=9. (Accessed Oct. 4, 2022).

(4) San Joaquin Valley Air Pollution Control District, Rule 4460: Petroleum Refinery Fenceline Air Monitoring. Available at https://www.valleyair.org/Workshops/postings/2022/08-15-22/Proposed-Rule-4460.pdf. (Accessed Oct. 4, 2022). Final rule adopted December 19, 2019.

(5) Colorado State Assembly, House Bill 21–1189. Available at https://www.leg.colorado.gov/sites/default/files/2021a_1189_signed.pdf. (Accessed Oct. 4, 2022).

(6) U.S. Environmental Protection Agency, Petroleum Refinery Sector Rule (Risk and Technology Review and New Source Performance Standards). Available at https://www.epa.gov/stationary-sources-air-pollution/petroleum-refinery-sector-rule-risk-and-technology-review-and-new (Accessed Oct. 4, 2022).

(7) U.S. Environmental Protection Agency Office of Research and Development, “Compendium Method TO-16; Long-Path Open-Path Fourier Transform Infrared Monitoring of Atmospheric Gases,” EPA/625/R-96/0101b, (January 1999).

(8) ASTM International, Standard Test Method for Determination of Gaseous Compounds by Extractive Direct Interface Fourier Transform Infrared (FT-IR) Spectroscopy, ASTM D6348-03. Available at https://www.astm.org/d6348-03.html (Accessed Oct. 4, 2022).

(9) Agency for Toxic Substances and Disease Registry, Toxicological Profile for Benzene. Available at https://wwwn.cdc.gov/TSP/ToxProfiles/ToxProfiles.aspx?id=40&tid=14. (Accessed Oct. 4, 2022).

(10) U.S. Environmental Protection Agency, Petroleum Refinery Sector Rule (Risk and Technology Review and New Source Performance Standards). Available at https://www.epa.gov/stationary-sources-air-pollution/petroleum-refinery-sector-rule-risk-and-technology-review-and-new. (Accessed Oct. 4, 2022).

(11) California Office of Environmental Health Hazard Assessment, Benzene. Available at https://oehha.ca.gov/air/chemicals/benzene. (Accessed Oct. 4, 2022).

(12) C. MacDonald and J.E. Marrero, “Ensuring High Quality Data from Industrial Fenceline Air Quality Monitors for Real-Time Public Reporting,” presented at the A&WMA West Coast Conference. (November 2020).

(13) California Office of Environmental Health Hazard Assessment, Hydrogen Sulfide. Available at https://oehha.ca.gov/chemicals/hydrogen-sulfide. (Accessed Oct. 4, 2022).

(14) California Air Resources Board, Air Quality Standards. Available at https://ww2.arb.ca.gov/resources/background-air-quality-standards. (Accessed Oct. 4, 2022).

Steven R. Schill, R. Scott McEwan, Ryan C. Moffet, Josette E. Marrero, Clinton P. MacDonald, and Eric D. Winegar are with Sonoma Technology, Inc., in Petaluma, California. Direct correspondence to: ewinegar@sonomatech.com ●