Rapid Screening of Benzene and Aromatic Hydrocarbons in Soil Samples Using Portable Gas Chromatography with Photoionization Detection

A procedure for rapid screening of aromatic hydrocarbons was developed to analyze soil samples using a portable gas chromatograph with a photoionization detector. Three types of soil samples were selected and spiked with different amounts of aromatic hydrocarbons, e.g., benzene, toluene, and xylenes, to determine the detection limits of the instrument and study the influence of the sample matrix on the detection of the studied compounds. Good chromatographic separation between the compounds was achieved at specific analytical conditions. The calibration curves for all target compounds showed correlation, R2 above 0.995 with RSD below 10%, demonstrating very good linearity and precision. The recoveries for all three types of soil samples were within 20%. The limits of quantitation for all compounds were determined between 2 and 7 µg/L in aqueous media, equivalent to 10 to 35 µg/kg in soil, which corresponds well with the requirements of EPA Method 8021B.

KEY WORDS: aromatic hydrocarbons, benzene, gas chromatography, headspace sampling, photoionization detection

Benzene, toluene, ethylbenzene, and xylenes, referred to as BTEX, are aromatic hydrocarbons that naturally occur in crude oil. These compounds are found in gasoline and used in a wide range of industrial processes. Benzene is determined as a carcinogen to humans by both the International Agency for Cancer Research (IACR) and the Environmental Protection Agency (EPA). The EPA’s Mobile Source Air Toxics (MSAT) gasoline fuel program aims at reducing hazardous air pollutants, also known as air toxics, emitted by cars and trucks, including benzene (EPA, n.d.a). According to the program, average gasoline benzene content should be decreased to 0.62 volume percent (vol%) for both reformulated and conventional gasoline, with a maximum average benzene standard of 1.3 vol%. Nationwide, there are approximately 550,000 underground storage tanks (USTs) for petroleum or hazardous substances (EPA, n.d.b). This poses a high risk for contamination of soil and groundwater, which is a source of drinking water for nearly half of the population.

According to the Toxic Release Inventory (TRI) from 2006, an estimated release of 24,033 pounds (~11 metric tons) of benzene to soils from 968 domestic manufacturing and processing facilities in 2004 accounted for about 0.3% of the estimated total environmental releases from facilities required to report to the TRI (ATSDR, 2007). An additional +435,000 pounds (~197 metric tons), constituting about 6% of the total environmental emissions, were released via underground injection. Benzene is released to soils through industrial discharges, land disposal of benzene-containing wastes, and gasoline leaks from underground storage tanks. There is a potential for release of benzene to soil from hazardous waste sites. Benzene has been detected in soil samples collected at 436 of the 1,684 sites and in sediment samples collected at 145 of the 1,684 sites where benzene has been reported in some medium.

Despite advances in technologies and safety practices in place, hazardous spills of chemical products containing aromatic hydrocarbons are not uncommon. The Center for Biological Diversity (n.d.) reported that between 1986 and 2013, there were nearly 8,000 significant incidents with hazardous liquid spills from pipelines, resulting in an average quantity per spill of 76,000 barrels per year or more than 3 million gallons.

On December 7, 2022, a leak in the Keystone Pipeline released 14,000 barrels of oil into a creek in Washington County, Kansas. The leak was the largest in the United States since the 2013 North Dakota pipeline spill. It was the largest oil leak in the history of the Keystone Pipeline. As a result, the pipeline was shut down, and isolation valves were commanded closed (United States Department of Transportation [USDOT], 2023). The EPA recommended building an earthen dam to contain the spill. Clean-up efforts and reopening of the pipeline were time-sensitive. According to TC Energy, the spill cost about $480 million in clean-up efforts (TC Energy, 2023).

Monitoring of benzene and aromatic hydrocarbons in soil is of critical importance during emergencies and oil spills to avoid huge financial losses. The EPA has developed various methods for analysis of volatile organic compounds (VOC) in solid and liquid matrices, including soil. Methods 5021, 8021, and 8260 involve headspace sampling and gas chromatography analysis with an applicable concentration range between 0.1 and 200 μg/L (EPA, 2017; Zhao and Kira, 2017). Despite the robustness and low limits of detection, offsite analysis of samples has certain disadvantages. The normal procedure requires the collection of samples onsite and analysis of the samples offsite in a laboratory. This includes sample transportation, storage, and handling at specific conditions, such as cooling and longer sample preparation before analysis. In addition, recovery of target analytes may decrease over time due to volatility and degradation in the sample matrix.

In addition to offsite laboratory analysis, alternative methods for screening of VOCs in soil are also available. Hand-held photoionization detectors (PID) are often employed to measure VOC concentrations in water or soil (RAE Systems, 2005). Hewitt and Lukash (1999) reported a linear correlation when studying the headspace PID response of benzene, toluene, xylenes, dichloroethylenes, trichloroethylene, and perchloroethylene concentrations in soil samples with concentrations between 0.2 and 10 mg VOC/kg. Hand-held PIDs nowadays can detect VOCs at parts per billion (ppb) levels while also being capable of direct benzene measurements with pre-filter tubes (Draeger, n.d; Honeywell, 2004). Conventional benzene detector tubes could also be used to report benzene levels. However, both methods present challenges and do not provide as accurate results as field-portable and benchtop gas chromatographs.

During the last decades, a new segment has emerged in the field of analytical chemistry with the availability of portable gas chromatographs capable of analyzing samples onsite, straight after collection with little or no sample preparation (EPA, 1998; Soo et al., 2018). Portable gas chromatographs are equipped with PIDs, electron capture detectors (ECD), or even mass selective detectors (MSD). PIDs and ECDs are beneficial for portable applications due to their compact size, high sensitivity, and selectivity. Moreover, such detectors do not require auxiliary gases, e.g., hydrogen, nitrogen, or air to operate.

Photovac Voyager (Waltham, MI) is a field-portable, computer-controlled gas chromatograph that incorporates three columns and dual detectors, a PID, and an ECD to achieve broadened analytical capabilities (EPA, 1998). This instrument was developed with consideration of ergonomic and analytical performance demands in field environments. Photovac Voyager has a unique internal analytical engine that includes a specially designed miniature stainless steel valve array to provide fast sample delivery and minimize sample carryover and contamination caused by high VOC concentrations. The instrument also incorporates a unique triple-column arrangement, with pre-column and backflush, a port with a pump for direct air sampling, and a syringe injection port for headspace sampling of aqueous and soil extract media. Ultra-pure air or nitrogen (99.999%) can be used as a carrier gas. Columns A, B, and C are intended for analysis of heavy (C7 to C12), midrange (C4 to C7), and light (C1 to C3) hydrocarbon compounds, respectively. The internal sampling train, sample loop, GC columns, valves, and injection port are heated isothermally at temperatures from 55 to 80C. The Voyager is also unique being classified as intrinsically safe (Class 1, Division I, Groups A, B, C, and D), rendering it useful in hazardous locations. The Voyager can be effectively used to monitor many of the VOCs listed in EPA Method 8260D (EPA, 2017), including chlorinated and aromatic hydrocarbons. Sample matrices of applicability include soil, soil gas, water, and ambient air. The MDLs for VOCs range from parts per trillion (ppt) in water (ng/L) to about 500 parts per million (ppm) in ambient air, depending upon the type of compound and detector used. The use of robust columns and a photoionization detector give the instrument an additional advantage for large-volume gaseous samples, even with the presence of water vapors from headspace or ambient air.

An analysis of VOCs in aqueous and soil samples has been extensively studied, with results meeting or exceeding the requirements outlined in the abovementioned EPA methods. Some of the studies (EPA, 1998; Zhang, 2019; Soo et al., 2018) point out the challenges in separating two of the xylene isomers, specifically m-, and p-xylene. The use of polar columns, e.g., polyethylene glycol (PEG) stationary phase column, also known as WAX column, provides good separation between the components in the BTEX mixture despite the increase of the peak width for high boiling compounds.

This study focuses on developing a rapid procedure for the analysis of BTEX compounds in soil samples using a portable gas chromatograph with a photoionization detector. The soil samples are dissolved in purified water, and after an equilibrium is reached in the vial, a headspace sample is introduced into the system for analysis. The obtained results can be viewed on the instrument screen and transferred to a computer for further data processing.

2.1 Chemicals and Standards

A certified mixture of BTEX, including benzene, toluene, ethylbenzene, p-xylene, m-xylene, and o-xylene with concentrations of 2,000 µg/mL analytes in methanol was obtained from Restek (Bellefonte, PA) to prepare calibration standards in step dilutions. Fluorobenzene, 2,000 µg/mL in methanol, used as an internal standard, was also obtained from Restek. Analytical grade methanol (>99.9%) was purchased from Sigma-Aldrich (St. Louis, MO), and ultrapure water (MilliporeSigma, Rockville, Maryland) was used for sample preparation.

Calibration standards were prepared from stock standards with dilution in methanol, resulting in final concentrations in sample vials of 10, 20, 50, 100, and 200 µg/L for the BTEX mixture and 100 µg/L for fluorobenzene at all calibration levels.

Three types of soil matrices were used in the study, including sand, farmland soil, and clay. The soils were air-dried and passed through no. 18 soil sieve (diameter <0.991 mm) but retained on a no. 120 soil sieve (diameter >0.125 mm). Glass beads with similar size were added to blank and calibration samples to compensate for volume change in the sample vials. All matrices were analyzed for interfering compounds prior to experiments.

2.2 Sample Preparation

The sample procedure was as follows: certified pre-cleaned 40-mL vials were filled with 10 mL of deionized water. Two grams of solid samples, e.g., glass beads or soil, were added to the vials followed by spiking with 100 µL stock standard of BTEX solution and 50 µL of internal standard. The vials were sealed immediately and shaken for two minutes. After that, the vials were equilibrated at room temperature (25°C) for 30 minutes. Samples (200 µL) were withdrawn manually with a gas-tight syringe penetrating through the septum of the vial cap and inserted straight into the injection port of the gas chromatography system.

2.3 Sample Analysis

A portable gas chromatograph (GC) with a PID and Photovac Voyager (Waltham, MI) was selected as an instrument to analyze the samples containing aromatic hydrocarbons. The design allows the instrument to be carried on site and analyze samples straight after collection or requires little sample preparation. The portable GC can be deployed and reach equilibrium within one hour (EPA, 1998). The selection of PID has many benefits for this application, including compact size, no need for auxiliary or makeup gases, wide linear range, and selectivity based on the ionization potential of the analytes. Although the instrument can operate with pure air or nitrogen (99.999%) as a carrier gas, nitrogen was selected due to improved peak shape and resolution. Voyager SiteChart LX Version 1.23 software was used to transfer and integrate the results of the analysis, and Microsoft Excel was used to further process the obtained data.

2.4 Method Optimization

During the method development, different parameters were optimized to improve the resolution of target compounds and reduce the time of analysis, including carrier gas pressure and column temperature. Initially, samples were analyzed on both A and B columns to determine the presence of co-eluting peaks of the target analytes due to different stationary phases and elution order of compounds. It was found that the compounds of interest have well-separated peaks without interferences on Column B (20 m x 0.32 mm x 1.0 µm, Supelcowax10). The carrier gas pressure was set to 10 psi with a column temperature of 60°C and a total run time of 10 minutes without sacrificing peak resolution. Detailed information on the analytical parameters is presented in Table 1. The injection volume was also optimized, and it was found that the best reproducibility was achieved at 200 µL, compared to injection volumes between 100 and 500 µL. The presence of traces of methanol in the samples from the spiking with calibration solutions did not have an observable effect on the early eluting compounds, e.g., benzene. This can be explained by the high value of the partition coefficient (K) of methanol in aqueous solution. Furthermore, the UV lamp used in the PID with an intensity of 10.6 eV has a low affinity to methanol with an ionization potential of 10.85 eV (Rae Systems, 2005).

A typical chromatogram of the analysis of BTEX mixture with fluorobenzene as an internal standard demonstrating the elution of compounds and separation of peaks is depicted in Fig. 1. The results of analysis for linearity, precision (%RSD), method limits of detection (LODs), and limits of quantitation (LOQs) are presented in Table 2. The linearity was determined from a five-point calibration between 10 μg/L and 200 μg/L with an internal standard concentration of 100 μg/L. The analysis of the target compounds demonstrated good linearity, with coefficients of the linear regression R2 exceeding 0.995. The precision was measured by analysis of six replicate samples at 50 μg/L and 100 μg/L. Based on the results, the relative standard deviations for all six compounds were in the range between 8% and 10% at 50 μg/L and between 0.4% and 1.7% at 100 μg/L, thus demonstrating excellent quantitation precision. The limits of detection and limits of quantitation were calculated from seven replicate samples at a concentration of 10 μg/L. It was determined that the detection limits were between 0.5 and 2.0 μg/L, which translated into quantitation limits between 1.8 and 7.1 μg/L. Furthermore, samples with concentrations of the target compounds at 2 μg/L, corresponding to 10 μg/kg in soil, were analyzed to verify the limits of quantitation. The results confirmed that all six analytes were detected and quantitated at the expected quantitation limits.

The recoveries of the analytes were determined by analyzing six replicate samples at 50 μg/L and 100 μg/L in three different solid matrices: sand, farmland soil, and clay. The results of the analysis are presented in Table 3. The recovery varied from 74% to 99% for sand, 87% to 112% for soil, and 91% to 123% for clay.

In this study, various analytical parameters were optimized toward peak resolution, peak shape, and run time of the instrument. Co-elution of compounds in the BTEX mixture, more specifically p-, and m-xylene, were reported in multiple studies (EPA, 1998; Soo et al., 2018). The selected instrument has three analytical columns, including a polar and a PEG capillary column reported to separate the isomers of xylene. To achieve a good separation between the peaks of p-, and m-xylene, nitrogen was selected instead of oxygen as a carrier gas. Column temperature and pressure were optimized to achieve good separation between the peaks. It was found that good analytical conditions are met at column temperature of 60°C and pressure at 10 psi, maintained throughout the analysis. All analytes in the mixture elute within 600 seconds (10 minutes) at the selected conditions. No co-eluting peaks were registered from the selected solid matrices. In addition, maintaining the temperature of the detector at 60°C improves the peak shapes. Although no late eluting compounds were detected in the samples, in the case of less volatile components present in solid samples, the runtime of the instrument can be extended to 3200 seconds. The addition of fluorobenzene to the BTEX mixture improved the calibration and precision of the results.

The headspace sampling procedure was also optimized in relation to amount of water added to the vials and injection volume of samples in the instrument. Vials with volume of 40 mL were selected to facilitate the handling of solid samples, considering the procedure is applied onsite. Different volumes of ultrapure water, between 10 and 20 mL, were added to the vials to study detector response and repeatability. It was found that adding 10 mL of water provides better response of the detector compared to samples with 20 mL of water. In addition, increasing the headspace allows for larger volume injections in the system for analysis. The volume of injected samples was also optimized. The experiments demonstrated that the injection volume of 200 µL provided excellent results for calibration and repeatability.

The presence of traces of methanol in the samples from the spiking solutions did not interfere with elution of peaks, especially benzene, the first eluting compounds in the mixture. This can be explained by the high partition coefficient of methanol. Moreover, the photoionization detector has low affinity to methanol due to higher ionization potential.

The results of the analysis demonstrated that the selected instrument is capable of separating, detecting, and quantifying all six components in the BTEX mixture. The achieved linearity was above 0.995, precision below 10%, recovery within 25%, and detection limits in the single range parts per trillion (ppt). The detection limits can be further improved by reaching headspace equilibrium at elevated temperatures using heating blocks for the sample vials, although this would require the use of addition equipment. The injection volume can also be increased to 500 µL or even larger volumes. The recoveries for the selected solid matrices did not represent a significant drop in the response at the selected spike concentrations. This can be attributed to the short time between spiking and analysis.

This short study has certain limitations. The study involved specific portable gas chromatograph with analysis conducted at constant column temperature and pressure. Portable gas chromatographs vary in parameters, including temperatures of column and detector, column stationary phase, detector sensitivity, and sample introduction in the system. Although direct comparison between instruments from different manufacturers is not always possible, optimized method parameters from this study can be tested and transferred to other systems. Sample preparation and analysis were conducted in an offsite laboratory with maintained temperature and relative humidity. Nevertheless, the preparation of samples is pretty straightforward and requires only basic analytical equipment, readily available in mobile and field laboratories. The study was also limited to three types of solid matrices. Further study of different soils can be beneficial to evaluate the recoveries of analytes depending on the matrix.

This procedure for rapid screening can provide valuable results for the presence of specific VOCs in solid samples straight onsite with limited time for instrument equilibration and sample preparation. It has to be noted that the instrument is intrinsically safe and can operate without power and carrier gas supply for 6 to 8 hours using a battery pack and built-in cylinder. The procedure can be applied successfully in emergency response clean-up operations and environmental remediation. The advantages of this approach include a simplified sample preparation procedure (compared to laborious analytical methods), analysis of samples onsite without the need to transport samples, short duration between sampling and reporting results, and low cost. These types of instruments can also be used to aid onsite exposure monitoring of the personnel involved in sampling and excavation of soils.

Despite the limitations of this study, the results provide relevant information on the preparation and analysis of solid sample matrices containing aromatic hydrocarbons using portable chromatography with photoionization detection. The results of the analysis are similar to the expected lower limits of quantitation, as outlined in EPA Method 8021—approximately 1 μg/kg for soil/sediment samples and 1 μg/L for groundwater. In addition, samples prepared and analyzed onsite can yield higher recoveries due to the short time between sample collection and analysis, thus overcoming losses in offsite analysis due to sample transportation, storage, matrix effects, and handling.

Although this study covers the use of a single instrument, sample preparation in an offsite laboratory, and selection of three types of solid samples, the results demonstrate that portable gas chromatographs can be utilized and successfully used in the field for emergency response and remediation cleanup operations in the screening of contaminated soil for VOCs. Such instruments are available in different configurations, including mass selective detection to provide relevant information on the type of contaminants and levels of contamination in low parts per billion concentrations. The procedure can also be applied for routine screening of underground storage tanks or pipelines to detect early leakage and contamination of soil and groundwater.

Agency for Toxic Substances and Disease Registry, Toxicological Profile for Benzene. https://www.atsdr.cdc.gov/toxprofiles/tp3.pdf, 2007.

Center for Biological Diversity, America’s Dangerous Pipelines. https://www.biologicaldiversity.org/campaigns/americas_dangerous_pipelines/

Draeger. (n.d.). PID Sensors. https://www.draeger.com/en-us_us/Products/PID-sensors

Environmental Protection Agency. (n.d.a.) Gasoline Mobile Source Air Toxics. https://www.epa.gov/gasoline-standards/gasoline-mobile-source-air-toxics

Environmental Protection Agency. (n.d.b.). Underground Storage Tanks (USTs). https://www.epa.gov/ust

Environmental Protection Agency. (1998). Field-Portable Gas Chromatograph Perkin-Elmer Photovac, Voyager. https://archive.epa.gov/nrmrl/archive-etv/web/pdf/01_vr_pep.pdf

Environmental Protection Agency. (2014). Aromatic and Halogenated Volatiles by Gas Chromatography Using Photoionization and/or Electrolytic Conductivity Detectors. https://www.epa.gov/sites/default/files/2015-12/documents/8021b.pdf

Environmental Protection Agency. (2018). Volatile Organic Compounds by Gas Chromatography/Mass Spectrometry. https://www.epa.gov/sites/default/files/2018-06/documents/method_8260d_update_vi_final_06-11-2018.pdf

Hewitt, A.D. & Lukash, J.E. (1999). PID-based estimation of the concentration of aromatic and chlorinated VOCs in soil. Field Analyt. Chem. Technol. 3(3):193-200.

Soo, J.C., Lee, E.G., LeBouf, R.F., Kashon, M.L., Chisholm, W, & Harper, M. (2018). Evaluation of a portable gas chromatograph with photoionization detector under variations of VOC concentration, temperature, and relative humidity. Journal of Occupational and Environmental Hygiene, 15:4, 351–360, DOI: 10.1080/15459624.2018.1426860.

Rae Systems. (2004). RAE Systems PID Training Outline. https://prod-edam.honeywell.com/content/dam/honeywell-edam/sps/his/en-us/products/gas-and-flame-detection/documents/Application-Note-000_RAE-Systems-PID-Training-Outline_04-05.pdf

Rae Systems. (2005). Using Rae Systems PIDs for Measuring Soil Headspace and in Soil Vapor Extractions. https://sps.honeywell.com/content/dam/his-sandbox/products/gas-and-flame-detection/documents/Technical-Note-118_Using-RAE-Systems-PIDs-for-Measuring-Soil-Headspace-and-in-Soil-Vapor.pdf

TC Energy. (2023). Milepost 14 Incident. https://www.tcenergy.com/incident/milepost-14-incident/

United States Department of Transportation. (2023). Amended Corrective Action Order. https://primis.phmsa.dot.gov/enforcement-documents/32022074CAO/32022074CAO_Corrective%20Action%20Order%20(Amended)_03072023_(22-261792).pdf

Zhang, J. (2019). Determination of Volatile Organic Compounds in Soil and Sediments. https://www.agilent.com/cs/library/applications/application-voc-in-soils-and-sediments-8890-GC-5994-0662en-agilent.pdf

Zhao fa, H., & Kira, Y. (2017). Determination of Volatile Organic Compounds in
Soils By HS-GC/MS. https://www.perkinelmer.com/lab-solutions/resources/docs/APP_Determination_Volatile_Organic_Compounds_in_Soils_013764_01.pdf