Recovery and reactivity of polycyclic aromatic hydrocarbons collected on selected sorbent tubes and analyzed by thermal desorption-gas chromatography/mass spectrometry

Abstract

This article describes the optimization of methodology for extending the measurement of volatile organic compounds (VOCs) to increasingly heavier polycyclic aromatic hydrocarbons (PAHs) with a detailed focus on new sorbent tube technology. Although PAHs have lower volatility than compounds such as benzene, toluene, ethylbenzene and xylenes, these semi-volatile compounds can be detected in air and breath samples. For this work, PAHs were captured on sorbent tubes and subsequently analyzed using automated thermal desorption gas chromatography – mass spectrometry (ATD-GC/MS). While many different sorbent tubes are commercially available, optimization for airborne PAH sampling using sorbent tubes has not been previously considered. Herein, several commercially available sorbent tubes, including Carbograph 2TD/1TD, Tenax TA, XRO-440, and inert-coated PAH tubes are compared to determine the relative recovery for eight PAHs commonly found in the environment. Certain types of sorbent materials were found to be better suited for PAH recovery during thermal desorption and some sorbent tube materials may react with PAHs during sampling or thermal desorption. As such, selection of sorbent tube media should be carefully considered prior to embarking on a PAH study.

1. Introduction

Polycyclic aromatic hydrocarbons (PAHs) are organic compounds composed of two or more fused hydrocarbon rings. PAHs are prevalent in environmental media, including air, water, dust, soil, and food [[1], [2], [3]]. They are found or created as products of incomplete combustion of tobacco smoke, coal, crude oil, waste incineration, domestic heating, wood and gasoline as well as from forest and building fires [[2], [3], [4]]. Human exposure to PAHs occurs through multiple routes, including inhalation, ingestion, and dermal contact, and PAHs are prevalent in both indoor and outdoor air [4,5]. Exposure to PAHs has been linked to increased risk for cancer, predominantly lung cancer, which accounts for the majority of cancer-related deaths worldwide [5,6]. Other adverse outcomes are related to DNA and human serum albumin adducts, particularly with respect to adverse birth outcomes [[7], [8], [9], [10]]. Multiple PAHs have been classified as potentially carcinogenic to humans, including benzo[a]pyrene [2,3,5], and the mechanisms of PAH toxicity to humans have been well documented [6].

Individuals with certain occupations may have increased exposure to PAHs and risk for adverse health effects. For example, coke oven workers [11] and firefighters [[12], [13], [14]] are exposed to byproducts of combustion that produce PAHs, and roof and road pavers working with hot asphalt [15,16], welders [17], and military personnel exposed to jet fuel [18] have also been shown to have PAH exposure. Coke oven workers have been shown to have increased diabetes risk [11] and declines in lung function [19], and firefighters have increased incidence of certain types of cancer [20,21] and acute cardiovascular events [22,23]. While total PAH concentrations in North America have decreased in the last 30 years, PAHs continue to present a threat to human health and the environment [1].

To assess human exposure to PAHs and other environmental contaminants, air samples or biological samples are often obtained for routine analytical analyses. Urinary testing for PAH metabolites, such as 1-hydroxypyrene [[15], [16], [17],24], 1-hydroxynaphthalene [19], and 4-hydroxyphenanthrene [11] has been routinely performed. PAHs have also been assessed in breast milk [25] and blood [26,27] but obtaining these biological fluids may be more invasive for the subject and require specialized equipment and personnel. Thus, few individuals may be willing to provide these types of samples. In contrast to blood and urine collection, breath sampling is non-invasive and can be performed with relative ease [28]. PAHs in breath are generally present in their native form before metabolism. As such, breath samples can be taken immediately after a suspected inhalation exposure. Breath sampling has been employed to assess PAHs in firefighters [12,13], and naphthalene has been studied in breath as a biomarker of jet fuel exposure (JP-8) [18].

Despite the apparent advantages, breath sampling has not been used extensively for PAH assessment. PAHs span a wide range of volatility and are considered “phase-distributed” contaminants because they occur in ambient air as a gas and are also condensed onto airborne particles, depending on the temperature [29,30]. As such, sampling for PAHs is difficult because both phases must be captured and subsequently recovered for analysis [31]. Originally, PAH measurement methodologies involved some form of combined filter and vapor trap media to avoid losses during sampling. These methods were first documented in EPA Method TO-13 in 1989, and then revised for TO-13A in 1999 [32]. PAHs are collected onto filters and sorbent cartridges containing PUF or XAD-2 with air sample volumes of approximately 300 m³ (3 × 10⁸ mL). Compounds are extracted from the cartridges by Soxhlet extraction, and PAHs are ultimately analyzed by gas chromatography/mass spectrometry (GC/MS) [4,32]. As sampling and analysis instrumentation evolved, these volume constraints were alleviated, and 200–500 mL air samples can be used to analyze PAHs using GC/MS systems. Current methods utilize standard adsorbent cartridges (3 mm diameter ×75 mm length or 5 mm diameter ×90 mm length) to capture both gas and particle phase PAHs. Stainless steel sampling tubes can be packed with filters and glass wool plugs for PAH sampling. Therefore, thermal desorption (TD)-GC/MS methods for measuring single-ring aromatic compounds such as benzene, toluene, and xylenes, are now being extended to two- to six-ring PAHs [33,34].

Sorbent selection is an important factor to consider for PAH sampling and analysis. Many types of sorbent tubes with different packing materials are available depending on the size, volatility and properties of the compounds to be sampled [35]. Three different tube materials are commonly used: stainless steel, inert-coated stainless steel, and glass. Stainless steel tubes are used to capture VOCs for many gas-phase monitoring applications and can be utilized during both pumped and diffusive sampling. Inert-coated stainless steel tubes contain a very thin inert layer that is bonded to the interior surface of the tube [36]. These tubes can be used to sample compounds that are reactant and thermally labile. Glass tubes can be used to sample reactive compounds, but contain about two-thirds of the sorbent volume compared to stainless steel tubes [37]. Common hydrophobic sorbent materials for capturing higher boiling compounds include Tenax TA, XRO-440, and graphitized carbon black, such as Carbograph 1TD and 2 TD [34,37]. Tenax TA is used to analyze compounds from n-C6/7 to n-C30, while Carbograph 1TD and 2 TD can capture compounds from n-C6 to n-C12/14 and n-C8 to n-C20, respectively. Dual bed sorbent tubes can be utilized to cover a wider volatility range [37,38]. Sorbent tubes packed with Tenax TA and polydimethylsiloxane (PDMS) foam have been used to collect SVOCs, including PAHs and polybrominated diphenyl ethers (PBDEs), and TD-GC/MS methods have been optimized for analysis of these compounds [[39], [40], [41]]. Multi-sorbent tubes containing glass beads, Carbopack C, and Carbopack B have also been used for airborne PAH analysis using a short path thermal desorption [42]. The PerkinElmer XRO-440 hydrophobic multi-bed sorbent tube was designed to capture compounds from n-C4 to n-C44, including all regulated PAHs up to benzo[ghi]perylene [34]. The Markes PAH (chemical weapons) sorbent tubes were also developed to capture all 16 regulated PAHs (both gas- and particulate-phase compounds) [43].

Methods that are selective for PAHs and provide high percent recoveries for these target compounds are needed for air and breath sampling. For some applications, identification and quantification of both VOCs and PAHs is necessary, and thereby a sorbent tube that can be utilized to capture both types of compounds is required. Sorbent tube selection can be difficult, as each formulation has advantages and disadvantages toward sampling particular compounds based on factors such as the strength of interactions between the sorbent and target compounds, hydrophobicity, inertness, and formation of artifacts [36,38]. Additionally, previous studies have shown that larger sample volumes and higher sampling rates may be required for optimal PAH measurements, which may compromise the retention of VOCs [41]. As such, methods must be carefully optimized prior to sampling.

During a series of laboratory experiments for assessing exposures to toxic and carcinogenic contaminants, exhaled breath samples were collected with a focus on single-ring aromatic compounds, including benzene, toluene, ethyl benzene and xylenes (BTEX). Exposures to the related class of known human carcinogens, PAHs, were further explored using the same methodology. In previous work, a method for dual VOC and PAH analysis was developed using Carbograph 2 TD/1TD sorbent tubes [44]. While this method was found to be useful for analyzing VOCs, the recovery of the PAHs from the stainless steel Carbograph 2 TD/1TD tubes was not optimal. Here the experimental parameters selected for PAH standard loading in the previously published article are further investigated, and several different types of sorbent tubes, including Tenax TA, XRO-440, Carbograph 2 TD, and PAH tubes are compared for performance. The effect of method parameters such as standard loading time, temperature, desorption flow rate, sorbent selection, and type of tube selected (e.g., stainless steel, glass, or inert-coated) on PAH recovery are discussed.

2. Materials and Methods

2.1. Sorbent Tube Selection

The sorbent tubes utilized for experimental analysis of PAHs are listed in Table 1, and were supplied by Markes International (Gold River, CA, USA) or PerkinElmer Life & Analytical Sciences (Shelton, CT, USA). All sorbent tubes were conditioned prior to sampling using a TC-20 tube conditioner (Markes International) except for the XRO-440 sorbent tubes, which were conditioned using a desorption cycle on the PerkinElmer automated thermal desorber (ATD). XRO-440 sorbent tubes were conditioned for 2–3 cycles on the ATD, and the chromatograms were assessed to ensure there were no PAH backgrounds prior to sampling. All sorbent tubes were blanked prior to first use. Desorption temperatures and times varied and were based on supplier recommendations for the different types of sorbents.

Table 1:

Sorbent tubes selected for analysis of PAHs and conditioning parameters. All sorbent tubes were analyzed using the TC-20 tube conditioner except for the stainless steel and glass XRO-440 sorbent tubes, which were analyzed using 2–3 conditioning cycles on the ATD.

Sorbent Tube	Type	Condition T (°C)	Condition Time (min)	Catalog Number	Supplier
Carbograph 2TD/1TD	Stainless steel	380	120	C2-AAXX-5126	Markes International
Quartz wool-Carbograph 2TD	Glass	380	120	C2-CAXX-5141	Markes International
Carbograph 2TD	Inert-coated stainless steel	380	120	C2-CAXX-5141	Markes International
Tenax TA	Glass	335	90	C1-BAXX-5039	Markes International
Tenax TA	Stainless steel	335	90	C1-AAXX-5003	Markes International
PAH (Chemical Weapons)	Inert-coated stainless steel	330	60	C2-CAXX-5138	Markes International
XRO-440	Stainless steel	375	20	N9307121	PerkinElmer
XRO-440	Glass	375	20	Custom packed with same sorbents included in N9307121	PerkinElmer

2.2. PAH Standard Preparation

A 2000 ng/μL Restek Corporation (Bellefonte, PA, USA) PAH standard (catalog no. 31011, lot no. A099403) with 16 compounds was diluted in HPLC grade methanol (Fisher Scientific, Hampton, NH, USA) to prepare 1.0 and 2.0 ng/μL PAH standards. PAH standards were loaded onto sorbent tubes in 1 or 2 μL volumes. Standards were either loaded by spiking the PAHs liquid directly onto the quartz wool (glass tubes) or by injecting the liquid into the tube using a loading port (stainless steel and inert-coated tubes). The loading port was fashioned out of a single tube conditioner (model 10) from Dynatherm Analytical Instruments, Inc. (CDS Analytical LLC, Oxford, PA, USA). The gas flow and temperature settings were controlled by the loading port and confirmed using a flow meter and thermometer. Sorbent tubes were inserted into the loading port in the direction of helium gas flow, and Hamilton glass syringes were used to inject PAH standards onto the sorbent tube through the loading port. All sorbent tubes were placed onto the loading port after or during PAH injection in a steady stream of helium at a flow rate of 50 mL/min for either 30 s or 4 min. Some sorbent tubes were loaded at room temperature while others were loaded using the loading port after it was heated to approximately 127 °C. Additional details regarding PAH standard loading procedures can be found elsewhere [44].

2.3. ATD-GC/MS Analysis

Chromatographic separation was achieved using a Rxi-5Sil MS capillary GC column with a 5 m Integra Guard column, 0.25 mm ID, 30 m length, and 0.25 μm film thickness (part no. 13623–124) purchased from Restek Corporation. Research grade helium gas (99.9999%) and ultra zero air were obtained from Airgas (Morrisville, NC, USA).

A PerkinElmer (PE) 650 TurboMatrix ATD system and an Agilent 6890 N GC coupled to an Agilent 5975 inert XL MS instrument were utilized for analysis. Thermal desorption was achieved using one of the methods described in Table 2. Each method had a purge time of 5 min, a column flow rate of 2 mL/min, an outlet split of 6 mL/min, and the inlet split option was not selected. The valve temperature was set at 270 °C, the transfer line 290 °C, and the trap temperature was set from 10 to 385 °C with a 10 min hold.

Table 2.

Thermal desorption method parameters used throughout the study to analyze sorbent tubes with different packing materials. Method parameters were selected based on supplier recommendations.

No.	Method	T (°C)	Flow rate (mL/min)	Time (min)
1	Carbograph 1	375	40	20
2	Carbograph 2	375	20	15
3	Carbograph 3	375	60	15
4	Tenax	280	40	15
5	Inert-Carbograph 2TD	325	40	20

The GC oven temperature was held at 35 °C for 2 min and increased to 190 °C at 6 °C/min, then increased to 310 °C at 28 °C/min and held for 8 min for a 40.12 min run time. The quadrupole, ion source, and transfer line temperatures were 176, 290, and 290 °C, respectively. Ions were monitored from 35–300 m/z. Scan spectra were collected at a rate of 2^2 and the SIM/scan method option was utilized. Of the 16 components in the PAH standard, naphthalene, acenaphthylene, acenaphthene, fluorene, phenanthrene, anthracene, fluoranthene, and pyrene were targeted in the SIM method.

The stainless steel Carbograph 2 TD/1TD and glass quartz wool-Carbograph 2 TD sorbent tubes were desorbed using methods 1–3 in Table 2; the glass and stainless steel Tenax TA and the inert-coated stainless steel tubes with PAH sorbent were desorbed using methods 4 and 5, and the inert-coated stainless steel and quartz wool-Carbograph 2 TD sorbent tubes were desorbed using method 6. Glass and stainless steel XRO-440 sorbent tubes were desorbed using methods 1–3. Most sorbent tubes were analyzed within 1–2 days of PAH standard loading. Some tubes were stored for a maximum of 2 weeks – 1 month before analysis based on instrument availability, therefore meeting the maximum recommended sample storage length of 30 days [36].

2.4. Experimental Designs

A series of experiments were designed to determine the reasons for poor recovery of PAHs from Carbograph 2 TD/1TD sorbent tubes and to investigate PAH recovery from additional types of sorbent tubes. The experimental parameters are included in Table 3. While ideally the number of replicates within an experiment would be the same for all of the sorbent tubes tested, this was not possible for the comparison of additional types of sorbent tubes due to limited availability of some of the tubes tested and instrument time. For each sorbent type, all replicates analyzed were included to limit favoritism of data.

Table 3.

Description of experimental conditions used for PAH analyses. Parameters were altered in the experiments to determine the effect of each condition on PAH percent recovery. When heat was used during PAH standard loading, the temperature of the heated loading port was approximately 127 °C.

Experiment	Sorbent Tube	Type	Desorption Method	PAHs (ng)	Loading Time (min)	Heat	Number
Incomplete Desorption	Carbograph 2TD/1TD	Stainless Steel	Carbograph 1	1.0, 2.0	4	Yes	2
	XRO-440	Glass	Carbograph 1	2.0	4	Yes	3
Method Optimization	Carbograph 2TD/1TD	Stainless Steel	Carbograph 1	1.0	4	Yes	15
			Carbograph 2	1.0	4	Yes	15
			Carbograph 3	1.0	4	Yes	15
Breakthrough	Carbograph 2TD/1TD	Stainless Steel	Carbograph 1	2.0	0.5	No	3
				2.0	4	No	3
	XRO-440	Glass	Carbograph 1	2.0	0.5	No	3
				2.0	4	No	3
Heat	Carbograph 2TD/1TD	Stainless Steel	Carbograph 1	2.0	0.5	Yes	3
				2.0	4	Yes	3
	XRO-440	Glass	Carbograph 1	2.0	0.5	Yes	3
				2.0	4	Yes	3
Additional Sorbent Tubes and Reaction Products	Carbograph 2TD/1TD	Stainless Steel	Carbograph 1	2.0	0.5	Yes	8
	QW-Carbograph 2TD	Glass	Carbograph 1	2.0	0.5	Yes	9
	Inert-coated Carbograph 2TD	Inert-coated	Inert-Carbograph 2TD	2.0	0.5	Yes	6
	QW added to Carbograph 2TD/1TD	Stainless steel	Carbograph 1	2.0	4	Yes	3
	XRO-440	Glass	Carbograph 1	2.0	0.5	Yes	6
	XRO-440	Stainless Steel	Carbograph 1	2.0	0.5	Yes	6
	Tenax TA	Stainless Steel	Tenax	2.0	0.5	Yes	12
	Tenax TA	Glass	Tenax	2.0	0.5	Yes	12
	PAH	Inert-coated Stainless steel	Tenax	2.0	0.5	Yes	6

2.5. Data Analysis

ChemStation software version D.02.00 was used to integrate chromatographic peaks. The QA report included with the Restek Certificate of Analysis for the PAH SV calibration mix no. 5 (Lot A099403) shows that for the eight PAHs tested (naphthalene – pyrene), the average area counts obtained from three GC/MS analyses of the standard were within 2.5–5.9% of the average area counts obtained for naphthalene. Thus, the expected percent recoveries of these PAHs should be close to 100% in comparison to the recovery of naphthalene. Therefore, in this experiment percent recoveries were calculated using eqn. 1 with the assumption that naphthalene was fully recovered in each experiment.

$$\text{ Percent Recovery }=(\text{ area counts of PAH/area counts of naphthalene })\times 100$$. [1]

Comparisons of the raw area counts obtained for naphthalene using different desorption tubes revealed differences ranging from 5 to 10% for this VOC, indicating that naphthalene was consistently desorbed from all sorbent tubes. While not all PAHs may expect to give the same response factor through thermal desorption as they do in direct liquid injection, sorbent tube performance was evaluated based on highest obtainable percent recovery for each PAH (even if less than 100%).

GraphPad Prism version 7 and Microsoft Excel were used for graphing and statistical analyses. Unpaired two-tailed t-tests (α = 0.05) were used to determine statistical significance of differences between experimental conditions. In the analysis of incomplete desorption, the percentage of PAHs remaining on the sorbent tubes after the first desorption cycle was determined using Eq. (2):

$$\text{\%PAHs}=(\text{PAH area in second desorption/area of PAHs spiked })\times 100$$ [2]

3. Results and Discussion

3.1. Recovery of PAHs from Carbograph 2TD/1TD Sorbent Tubes

In a method previously developed to analyze both VOCs and PAHs during breath sampling, stainless steel Carbograph 2 TD/1TD sorbent tubes were selected to capture the wide range of compounds required for GC/MS analysis [44]. The method was created to analyze BTEX compounds and eight priority PAHs of interest that could all be retained using Carbograph 2 T/1TD sorbent tubes with a compound range of n-C6 to n-C20 (naphthalene through pyrene). However, the recoveries of some PAHs were found to be much lower than expected in comparison to other PAHs during method development. Fig. 1 shows a representative TIC of a 2.0 ng PAH sample desorbed from a stainless steel Carbograph 2 TD/1TD sorbent tube. Peaks for acenaphthene, fluorene, and anthracene were lower than expected.

Figure 1:

SIM chromatogram of 2.0 ng PAH desorbed from a stainless steel Carbograph 2TD/1TD sorbent tube.

For comparison, the PAH standard was also spiked onto an XRO-440 sorbent tube to check if the poor recovery of acenaphthene, fluorene, and anthracene was due to degradation of the PAH standard or was specific to the use of the stainless steel Carbograph 2 TD/1TD sorbent tube. As shown in Fig. 2, the responses for acenaphthene, fluorene, and anthracene from XRO-440 tubes are consistent with those of the other PAHs, indicating that the poor recovery of PAHs observed in Fig. 1 may be due to the use of the stainless steel Carbograph 2 TD/1TD sorbent tubes. Due to the observed low recoveries of PAHs, a series of experiments were performed to improve the recovery of PAHs from stainless steel Carbograph 2 TD/1TD sorbent tubes and determine what factors may have contributed to differences in recovery between the sorbent tubes. The flow diagram in Fig. 3 was used to guide this investigation.

Figure 2:

SIM chromatogram of 2.0 ng PAH desorbed from a glass XRO-440 sorbent tube.

Figure 3:

Flow diagram depicting potential reasons for unsuccessful desorption of PAHs. The formation of reaction products may occur during PAH standard loading, sorbent tube storage, or during desorption/injection into the TD.

As seen in Fig. 3, unsuccessful desorption may be due to factors such as incomplete desorption, compound breakthrough, formation of reaction products, or irreversible binding of the compound to the sorbent. Incomplete desorption of compounds can occur if the TD method has not been optimized to properly desorb compounds; in this case, PAHs would remain on the sorbent tubes after the desorption cycle. Optimization of compound loading time is important because a loading time that is too long could lead to compound breakthrough (compound loss of approximately 50% in the helium gas effluent), while conversely a loading time that is too short could cause PAHs to not fully enter the sorbent beds [[45], [46], [47]]. Breakthrough volume is typically assessed by placing a second clean sorbent tube after the first tube to determine if any target compounds are collected onto the second tube [46]; in this study, breakthrough of PAHs was assessed at both short (30 s) and long (4 min) loading times to determine if PAH recovery decreased after longer loading times.

Desorption may also appear to be unsuccessful if reaction products/artifacts are formed during sample loading onto the sorbent tube, sample storage, or desorption/injection of the sample [48,49]. Chemical reactions involving sorbents have been previously reported in the literature [30,46,50]. Finally, if compounds collected on multi-bed sorbents move from weaker to stronger sorbent beds, they can become irreversibly bound to the sorbent and will not be recovered during analysis. To reduce this risk, multi-sorbent tubes should be stored under refrigeration and analyzed within 30 days [36,49]. The risk for compound migration and irreversible sorbent binding was mitigated in this study by following recommendations to store tubes under refrigeration, and all samples were analyzed within 30 days.

A previous study has also shown that PAHs with low volatility can become trapped at the 4-port valve within the TD system, leading to up to 10% loss during two-stage desorption [51]. While losses of PAHs within the desorption system were not considered in the current study, this factor should be equal for all sorbent tubes after desorption has taken place and is not likely to cause significant differences in percent recoveries of PAHs from different sorbent tubes.

3.2. Investigation of Incomplete Desorption of PAHs

To determine if PAHs were not fully desorbing from stainless steel Carbograph 2 TD/1TD sorbent tubes in a single analysis, the recovery of PAHs from these stainless steel sorbent tubes was compared to recovery from glass XRO-440 tubes (see Table 3 for experimental conditions). After the initial desorption cycle and GC/MS analysis, the sorbent tubes were desorbed a second time using Carbograph method 1 without tube conditioning, and Eq. 2 was used to determine the relative amounts of PAHs remaining on the tubes.

The second desorption of three XRO-440 glass tubes showed that small amounts of naphthalene, acenaphthene, fluorene, fluoranthene and pyrene remained on the tubes. However, the concentrations of naphthalene, acenaphthene, and fluorene remaining on the tubes were less than 2% of the original area counts, and the area counts of fluoranthene and pyrene ranged from 0.7 to 7.2% of the original area counts. No PAHs were observed after the second desorption of Carbograph 2 TD/1TD tubes. Therefore, poor PAH recovery from Carbograph 2 TD/1TD sorbent tubes does not appear to result from incomplete desorption of PAHs.

3.3. Thermal Desorption Method Optimization to Improve Recovery

To further investigate whether PAHs were optimally desorbed from the stainless steel Carbograph 2 TD/1TD sorbent tubes, three different thermal desorption methods were developed and tested for percent recovery. Forty-five sorbent tubes were loaded with 1.0 ng PAHs using a heated loading method [44], and fifteen sorbent tubes were desorbed per Carbograph desorption method (no. 1–3; see Table 1). The spread of the percent recoveries for each method can be seen in Fig. 4.

Figure 4:

Percent recoveries of PAHs by desorption method. Increasing the desorption flow rate did not have a significant effect on the average percent recovery of PAHs from stainless steel Carbograph 2TD/1TD sorbent tubes. Desorption methods: Carbograph 2 (20 mL/min); Carbograph 1 (40 mL/min); Carbograph 3 (60 mL/min).

While slight differences in recovery were observed by changing the thermal desorption method parameters, these data show that acenaphthene and fluorene displayed such low recovery from Carbograph 2 TD/1TD sorbent tubes that they were barely detectable despite the selected desorption method. Anthracene and pyrene were also recovered at less than 50% using all three methods. Due to the similarities in the results obtained using the three different desorption methods, Carbograph desorption method 1 with a 40 mL/min flow rate was utilized for future experiments.

3.4. Analysis of PAH Breakthrough

According to Fig. 1, one explanation for the poor recovery of PAHs could be compound breakthrough if analyte concentrations are too high or standard loading times are too long. Thus, breakthrough analysis was important during method development. The breakthrough of PAHs from stainless steel Carbograph 2 TD/1TD sorbent tubes was compared to breakthrough from glass XRO-440 tubes. Breakthrough of VOCs and PAHs has been previously assessed using the XRO-440 sorbent tubes, which did not show breakthrough when up to 205 mg of analyte was loaded onto the sorbent tube [34]. Therefore, the glass XRO-440 sorbent tubes provided an excellent control for the experiment.

Three stainless steel Carbograph 2 TD/1TD and three glass XRO-440 sorbent tubes were loaded with 2.0 ng PAHs for 30 s using the loading station and were analyzed by ATD-GC/MS. Three additional stainless steel Carbograph 2 TD/1TD and three glass XRO-440 sorbent tubes were loaded with 2.0 PAHs for 4 min and then analyzed by ATD-GC/MS. No significant differences in percent recovery were observed using either sorbent tube when the loading time was increased from 30 s to 4 min (two-tailed t-tests, α = 0.05). Therefore, since no evidence of PAH breakthrough was observed, the selected sample loading times do not appear to lead to poor recovery of PAHs. Since the loading time was found to not significantly affect PAH recovery from either the Carbograph 2 TD/1TD or XRO-440 sorbent tubes, the shorter loading time of 30 s was used to in future experiments to increase the efficiency of standard loading.

3.5. The Effect of Heat on PAH Percent Recovery

Differences in percent recovery between heated and room temperature injections were also investigated at the two different loading times. The injection port was heated to approximately 127 °C during PAH standard loading to help volatilize PAHs in the stream of helium gas. Tests were first performed to determine if any differences in percent recovery were observed between 30 s and 4 min loading times when a heated injection port was utilized. Three stainless steel Carbograph 2 TD/1TD and three glass XRO-440 sorbent tubes were loaded for either 30 s or 4 min with a heated injection port at the loading station. The percent recoveries obtained for PAHs with the heated injection were compared to those obtained without heat (room temperature), as investigated during the breakthrough study. As previously observed, PAHs did not show differences in percent recovery after 30 s or 4 min loading times with room temperature injection, but interestingly, three PAHs showed statistically significant differences in percent recovery from stainless steel Carbograph 2 TD/1TD sorbent tubes between 30 s and 4 min loading times in the presence of heat: fluorene (p = 0.0485), anthracene (p = 0.0002), and pyrene (p = 0.0003) (two-tailed t-tests, α = 0.05). All three of these PAHs had higher percent recoveries when loaded for 30 s compared to 4 min, indicating that some PAH loss may have occurred during the 4 min loading time when the injection port was heated. PAHs did not show any significant differences in recovery from the glass XRO-440 sorbent tubes at the two different loading times when the injection port was heated.

Next, the percent recoveries obtained for the PAHs were directly compared at each loading time to determine if the heated injection port significantly improved PAH percent recovery compared to room temperature injection. A few PAHs showed significant differences in percent recovery between room temperature and heated injections using stainless steel Carbograph 2 TD/1TD sorbent tubes, as shown in Table 4. Phenanthrene and anthracene showed significant increases in percent recovery using a heated injection port when loaded for only 30 s compared to room temperature injection (Table 4). In contrast, fluorene showed a significant decrease in percent recovery when loaded using heated injection instead of RT injection at both loading times (Table 4). Therefore, the use of heat during PAH loading in the current method may have contributed to the decreased percent recovery observed for fluorene. None of the PAHs had statistically significant differences in percent recovery from the glass XRO-440 tubes.

Table 4:

Differences between room temperature and heated loading injections. Significant differences between percent recoveries of PAHs desorbed from stainless steel Carbograph 2TD/1TD sorbent tubes using heated versus room temperature loading were found at two loading times in three replicate experiments. Two-tailed t-tests (α=0.05) were performed to determine significance between the percent recovery for each method.

Compound	Loading Time (min)	Percent recovery without heat (Avg. %)	CV without heat (%)	Percent recovery with heat (Avg. %)	CV with heat (%)	p-value
Fluorene	4	19.73	29.26	4.607	2.985	0.0105
Fluorene	0.5	27.02	7.824	13.29	40.10	0.0144
Phenanthrene	0.5	101.6	3.018	112.0	1.433	0.0066
Anthracene	0.5	58.02	4.706	70.70	6.385	0.0141

3.6. Observation of Reaction Products on Carbograph 2TD/1TD Sorbent Tubes

As shown in the flow chart in Fig. 3, the formation of reaction products/artifacts due to chemical reactions with the sorbent can also lead to incomplete/unsuccessful desorption of PAHs. Reaction products/artifacts may be formed by chemical reactions during standard loading/sampling, thermal desorption, sorbent tube storage, or the process of desorption/injection. Since the analytical method for PAHs analysis employed SIM/scan mode, the scan TICs of the stainless steel Carbograph 2 TD/1TD tubes were inspected for potential reaction products. SIM/scan mode has been previously shown to be advantageous for identifying non-targeted compounds such as artifacts [44]. Using this technique, two peaks were observed in the scan chromatograms from the stainless steel Carbograph 2 TD/1TD sorbent tubes that did not match the targeted PAHs. The spectra of these unknown compounds were searched using the NIST library, and the tentative identifications for these compounds are provided in Table 5. The identities of these reaction products have not been confirmed by standards.

Table 5.

PAH reaction products tentatively identified in chromatograms. Reaction products were tentatively identified using a NIST 17 library search.

Compound Name	Primary Ion (m/z)	Secondary Ion (m/z)	Formula	Retention Time (min)	Structure
9-Fluorenone	152	180	C13H8O	26.5
9,10-Anthracenedione	180	208	C14H8O2	29.5

Interestingly, these non-targeted compounds appear to be reaction products of the targeted PAHs that were included in the Restek PAH standard (e.g., fluorene to 9-fluorenone and anthracene to 9,10-anthracendione) that exhibited low recovery from Carbograph 2 TD/1TD tubes in the previous experiments (shown in Fig. 1). The use of a heated loading port appears to have facilitated the production of 9-fluorenone by increasing the reactivity of fluorene, which was shown to have a significantly decreased percent recovery when loaded onto Carbograph 2 TD/1TD sorbent tubes using a heated injection port compared to room temperature injection, as shown in Table 4. Likewise, the percent recoveries of the reaction products 9-fluorenone and 9,10-anthracenedione were higher after the heated loading injection compared to the room temperature loading injections for the stainless steel Carbograph 2 TD/1TD sorbent tubes, as shown in Table 6. The glass XRO-440 sorbent tubes showed lower recoveries of the reaction products than the stainless steel Carbograph 2 TD/1TD sorbent tubes, and only slight differences were observed between the room temperature and heated loading injections for the glass XRO-440 sorbent tubes. 9-Fluorenone and 9,10-anthracendione have been previously reported as oxidation reaction products of fluorene and anthracene. Interestingly, in the previous study, the reaction was conducted in a cylindrical stainless steel reactor at 300 °C [52]. Some charcoal sorbents also contain metals that may catalyze PAH degradation during thermal desorption, especially at higher temperatures [49]. Therefore, the use of stainless steel tubes and charcoal sorbents, such as graphitized carbon black may have contributed to the observed oxidation of PAHs.

Table 6.

Differences in percent recovery of reaction products between room temperature and heated loading injections for stainless steel Carbograph 2 TD/1TD and glass XRO-440 sorbent tubes.

Compound	Sorbent Tube	Loading Time (min)	Percent recovery without heat (Avg. %)	Percent recovery with heat (Avg. %)
9-Fluorenone	Stainless Steel Carbograph 2 TD/1TD	0.5	9.203	29.08
9-Fluorenone	Glass XRO-440	0.5	2.972	2.546
9,10-Anthracenedione	Stainless Steel Carbograph 2 TD/1TD	0.5	1.457	9.019
9,10-Anthracenedione	Glass XRO-440	0.5	0.4289	0.6961

3.7. Investigation of additional types of sorbent tubes

After observing the low recovery and reactivity of PAHs from Carbograph 2 TD/1TD sorbent tubes, additional types of sorbent tubes with similar compound ranges capable of retaining these eight PAHs were analyzed to determine if the reactivity was specific to stainless steel Carbograph 2 TD/1TD tubes and to discover other sorbent tubes that may be amenable to gas-phase PAH analysis. The following sorbent tubes were evaluated: glass quartz wool (QW)-Carbograph 2 TD, inert-coated stainless steel Carbograph 2 TD, glass XRO-440, stainless steel XRO-440, glass Tenax TA, stainless steel Tenax TA, and inert-coated stainless steel PAH tubes (Table 1). The sorbent tubes were loaded according to the experimental parameters in Table 3. The percent recoveries of PAHs observed for each of these sorbent tubes are shown in box-and-whisker plots in Fig. 5. The reaction products listed in Table 5 were also investigated in the scan chromatograms of the additional sorbent tubes (Fig. 6).

Figure 5:

Percent recoveries for PAHs desorbed from different types of sorbent tubes. Blue: stainless steel; red: glass; green: inert-coated. Box-and-whisker plots are shown from min to max. Many of the PAHs had inconsistent percent recoveries from stainless steel Carbograph 2TD/1TD sorbent tubes. PAHs appeared to desorb better from glass or inert-coated tubes. Note that the scales of the y-axes are larger for some PAHs than others depending on variability in overall percent recovery. C,S (n=8); C,G (n=9); C,SQ (n=3); C,I (n=6); X,S (n=6); X,G (n=6); T,S; (n=12); T,G (n=12); P,I (n=6).

Figure 6:

Percent recoveries of reaction products detected in scan chromatograms of different types of sorbent tubes. Blue: stainless steel; red: glass; green: inert-coated. Box-and-whisker plots are shown from min to max. C,S (n=8); C,G (n=9); C,SQ (n=3); C,I (n=6); X,S (n=6); X,G (n=6); T,S; (n=12); T,G (n=12); P,I (n=6). See figure legend of Figure 5 for complete tube descriptions.

3.7.1. Comparison of PAH recoveries from Carbograph sorbent tubes

Differences between the percent recoveries of PAHs obtained for stainless steel Carbograph 2 TD/1TD sorbent tubes were compared to glass quartz wool-Carbograph 2 TD, inert-coated Carbograph 2 TD, and stainless steel Carbograph 2 TD/1TD sorbent tubes with a wisp of quartz wool manually added to the sampling end of the tube. For all PAHs, the lowest percent recoveries were obtained using the stainless steel Carbograph 2 TD/1TD sorbent tubes. The stainless steel Carbograph 2 TD/1TD sorbent tubes performed much better when the PAHs were injected onto a wisp of quartz wool that was manually added to the sampling end of the tube. The increases in percent recoveries upon the addition of quartz wool were statistically significant for acenaphthene, fluorene, and anthracene, which were the three PAHs with the lowest recoveries from Carbograph 2 TD/1TD tubes without quartz wool. Quartz wool and glass wool have been reported to be advantageous for trapping semi-volatile compounds, such as PAHs and phthalates, and improving compound recovery [53].

The glass quartz wool-Carbograph 2 TD and inert-coated Carbograph 2 TD sorbent tubes only contain a single sorbent bed, the weaker sorbent Carbograph 2 TD. These sorbent tubes are amenable to collecting semi-volatile PAHs but would not be amenable to dual VOC and PAH analysis, as lighter VOCs would not be retained on these sorbent tubes. PAHs showed similar percent recoveries from these sorbent tubes, which were higher and more consistent than the stainless steel Carbograph 2 TD/1TD sorbent tubes for most PAHs.

PAH reaction products were also observed on the Carbograph sorbent tubes. Stainless steel Carbograph 2 TD/1TD sorbent tubes showed the highest percent recoveries of 9-fluorenone and 9,10-anthracendione, while the tubes with Carbograph 2 TD sorbent showed lower levels of reaction products. Therefore, switching from stainless steel to glass QW-Carbograph 2 TD or inert-coated stainless steel PAH tubes appeared to significantly decrease the incidence of PAH reaction products.

3.7.2. Comparison of PAH recoveries from XRO-440 sorbent tubes

As seen in Fig. 5, the stainless steel and glass XRO-440 tubes exhibited similar percent recoveries for all PAHs except for fluorene. Fluorene showed a statistically significant improvement in percent recovery from glass compared to stainless steel XRO-440 sorbent tubes (two-tailed t-test, α = 0.05). The stainless steel XRO-440 sorbent tubes also showed a higher average incidence of 9-fluorenone percent recovery compared to glass tubes but 9,10-anthracenedione was observed at less than 1%.

3.7.3. Comparison of PAH recoveries from Tenax TA sorbent tubes

PAHs showed average percent recoveries of 60% or greater from both stainless steel and glass Tenax tubes. No significant differences were observed between PAH percent recovery from stainless steel and glass sorbent tubes. Very low background percentages of 9-fluorenone were observed on both stainless steel and glass Tenax TA (1% or less). Thus, the tube material and Tenax TA sorbent do not appear to cause PAH reactivity.

3.7.4. Comparison of PAH recoveries from Inert-coated stainless steel PAH sorbent tubes

All PAHs showed high percent recoveries between 62–127% from inert-coated stainless steel PAH sorbent tubes with coefficients of variation (CVs) below 0.10. Less than 2% recovery of 9-fluorenone was seen on some of these sorbent tubes. The inert-coated stainless steel PAH tubes were amenable to analysis of the eight tested PAHs.

3.7.5. Trends observed in PAH recovery and reactivity from comparison of all sorbent tubes

Trends in the percent recovery and reactivity of PAHs observed from experiments using all nine combinations of sorbent tubes were assessed to draw some conclusions about factors leading to low PAH recovery. The combination of a stainless steel tube and graphitized carbon black sorbent (e.g., Carbograph 1TD/2 TD tubes) showed the lowest percent recoveries of acenaphthene, fluorene, and anthracene and thus the highest percent recovery of reaction products 9-fluorenone and 9,10-anthracenedione. In contrast, the Carbograph 2 TD sorbent tubes with the inert-coated stainless steel showed increased percent recoveries for all of these PAHs and very low levels of reaction products, indicating that the uncoated stainless steel may contribute to PAH reactivity with graphitized carbon black sorbents. Reactivity was not an issue for stainless steel tubes with different sorbent materials, such as Tenax TA, which showed similar PAH recoveries for both stainless steel and glass tube materials (see Fig. 5). While XRO-440 sorbent tubes with stainless steel did not have PAH recoveries that were as low as the stainless steel Carbograph 2 TD/1TD recoveries, XRO-440 tubes also contain some graphitized carbon black sorbent. The graphitized carbon black sorbent may have caused some reactivity with the stainless steel tube material, as evidenced by the presence of 9-fluorenone on the stainless steel XRO-440 sorbent tubes in conjunction with the decreased percent recovery of fluorene observed on these tubes compared to glass XRO-440 sorbent tubes.

3.8. Extension of Methodology to 4- and 5-ring PAHs

The utilized ATD-GC/MS method was developed to analyze VOCs and PAHs from breath using a single sorbent tube and GC column [44]. PAHs that eluted after pyrene were not well recovered from stainless steel Carbograph 2 TD/1TD sorbent tubes in the original method because some of these compounds are close to or outside of the compound size range that Carbograph 2 TD can retain (C20-C22). PAH analyses using Tenax and inert-coated stainless steel PAH tubes, which can better retain these heavier compounds, revealed improved peak shapes and recovery of benzo(a)anthracene, chrysene, benzo(b)fluoranthene, benzo(k)fluoranthene, and benzo(a)pyrene. Benzo(a)pyrene is of interest due to its carcinogenicity and use as a biomarker [2,4]. These results indicate that the method can be used to focus on analysis of 4-, 5-, and 6-ring PAHs by selecting a sorbent tube that can recover these heavier compounds. TD-GC/MS methods that focus on the analysis of 4–6 ring PAHs have been developed using sorbent tubes with quartz wool and Carbograph 2 TD [54] or by directly desorbing quartz fiber filters and other materials used to collect PAHs [[55], [56], [57]]. While the present study was designed to analyze both VOCs and light PAHs, improvements in PAH desorption and quantitation may be achieved by using a different sorbent tube and focusing the method on the analysis of heavier PAHs in the future.

4. Conclusion

The investigation of PAH recoveries from a suite of commercially available sorbent tubes revealed some interesting trends that appear to be linked to the sorbent and tube material selections. Recovery of fluorene, acenaphthene, and anthracene was lowest from stainless steel Carbograph 2 TD/1TD sorbent tubes. The poor recovery of fluorene and anthracene from stainless steel Carbograph 2 TD/1TD sorbent tubes appears to be linked to the formation of two reaction products, 9-fluorenone and 9,10-anthracenedione. These reaction products had the highest observable recoveries from sorbent tubes containing graphitized carbon black sorbents (Carbograph 2 TD/1TD, Carbograph 2 TD, and XRO-440) and stainless steel tube materials. The combination of the stainless steel sorbent tube material and the graphitized carbon black sorbent, which showed lower levels of target PAHs and higher levels of the reaction products than the glass and inert-coated stainless steel tubes with Carbograph 2 TD sorbents, may lead to PAH reactivity due to the presence of metals within the graphitized carbon black sorbent. The use of heat during standard loading may have also contributed to the formation of 9-fluorenone on stainless steel Carbograph 2 TD/1TD sorbent tubes, as fluorene showed significantly decreased percent recovery from these sorbent tubes during heated sample loading compared to room temperature. Acenaphthene reaction products may have been formed but were not detected in this analysis, and therefore the reason for the poor recovery of acenaphthene is unclear. However, the results suggest that the stainless steel tube material combined with the graphitized carbon black sorbent also affected either the capture or recovery of acenaphthene from these sorbent tubes.

As a result of these experiments and observations, some of the experimental parameters in the previously published method for dual VOC and PAH analyses can be altered [44]. In the published method, a 20 mL/min desorption flow rate was utilized. In this investigation, flow rates of 40 and 60 mL/min were also tested (Fig. 4). Since there were no apparent differences in recovery between these three flow rates, the flow rate of the original method can be increased if desired. In the published method, PAHs were loaded for 4 min with heat. In this investigation, however, some PAHs (fluorene, anthracene, and pyrene) showed optimal recovery when loaded for 30 s with heat. Therefore, the authors recommend loading PAH standards for 30 s with heat as long as graphitized carbon black sorbents, such as Carbograph 1TD or 2 TD, are not being used for sample collection, as heat may contribute to PAH reactivity on these sorbent tubes. In future, a range of temperatures for standard loading may be investigated to determine an ideal temperature that yields high recovery of PAHs while minimizing the formation of reaction products.

The selection of a sorbent tube that gives consistent and high recoveries of PAHs is important for method development, and this study provides useful comparisons of several popular sorbent tubes that may be considered for thermal desorption analysis of PAHs. This study also shows how the use of heat during standard loading and the loading time can affect the recovery of PAHs and potentially lead to the development of reaction products using certain sorbent tubes, especially uncoated stainless steel tubes with graphitized carbon black sorbents. Depending on study goals, a sorbent tube that can capture both VOCs and PAHs simultaneously may be desired. Tenax TA and XRO-440 sorbent tubes can both capture a wide range of compound sizes and are therefore both amenable to combined VOC/PAH analyses with high recovery of PAHs and minimal formation of reaction products [34]. This information can be valuable for study design and thermal desorption method development for the analysis of PAHs in breath or environmental samples of interest.