Information on the development of CSMs is readily available in a number of guidance documents including the following:

• ITRC Triad Implementation Guide, ITRC SCM-3, 2007 ITRC 2007
• Guidance for Conducting Remedial Investigations and Feasibility Studies Under CERCLA USEPA 1992
• Data Quality Objectives Process for Hazardous Waste Site Investigations: Final Guidance USEPA 2000b
• Standard Guide for Developing Conceptual Site Models for Contaminated Sites ASTM 2014a
• Environmental Cleanup Best Management Practices: Effective Use of the Project Life Cycle Conceptual Site Model USEPA 2011

TPH is made up of many hundreds to thousands of individual compounds that can change composition both spatially and temporally; therefore, it is necessary to use an investigative strategy that is considerate of TPH properties. TPH can be identified at a site in a variety of media (soil, water, air), can be measured in LNAPL, and it is essential for a practitioner to understand how the TPH analytical methods vary, and when to apply them. Furthermore, a successful risk assessment is dependent on an iterative and frequently updated CSM.

TPH has specific physical, chemical, and biological behaviors that need to be considered while developing a quantitative CSM, and therefore, those unique aspects of TPH will also need to be considered in the investigative strategy. Understanding petroleum hydrocarbon chemistry, and how that chemistry affects the fate and transport in the environment, is essential. If any data are missing from the CSM, the investigative strategy should focus on filling those data gaps using the appropriate sampling methods, handling procedures, data quality objectives, laboratory and field analyses, and evaluation and reporting of results. In general, myriad decisions need to be made on the quantity of data and the spatial distribution of data collection, what samples to collect from various media using appropriate field tools, and which analytical methods to use. Therefore, any sampling and analysis related to petroleum contamination should be discussed with a hydrogeologist and risk assessor prior to preparing a TPH data collection work plan to ensure that the appropriate data for conducting the TPH risk assessment is available. Sampling locations, sample density, and other basic assessment considerations should also consider guidance from ITRC, EPA, state, and other resources.

Although pipelines and aboveground storage tanks can carry/hold most forms of petroleum, underground storage tanks typically contain only refined petroleum products. These initial indications can also be supplemented through a review of historic documents (inventories, facility engineering/design reports, and release histories), interviews, or early assessment data collection. Knowing the type of petroleum is crucial to guide the investigative strategy that can be employed to refine a qualitative CSM to become quantitative (including understanding the compounds to sample for in a more detailed risk assessment).

Potential maximum release volumes or rates vary with storage volumes or the nominal volume flows associated with facility operations. A finite release volume of petroleum will spread over time to impact a finite area and volume of porous soils and sediments (much as any spilled liquid on porous media expands to a finite wet area) before receding due to attenuation. A continuing release of NAPL will expand to impact a finite area and volume of soil or sediment. At the maximum extent of impact, the release rate and the attenuation rate (due to dissolution, volatilization, and degradation) are equal. Mobile NAPL might directly intercept receptors as NAPL.

TPH data collection, analyses, and interpretation are just parts of a CSM. Careful field screening, using light-induced fluorescence and/or membrane interface probe tools and an understanding of chromatograms will enhance the information gained from conducting “bulk” TPH analyses. Detections on the light-induced fluorescence monitor or detections above the calibrated baseline for membrane interface probe response can be used as a rough yes/no determination as to the presence of TPH contamination. Furthermore, if the nature of the release, other than the location, is unknown, assuming a 200-foot radius from the source can be a starting point for delineating the TPH impacts. This is based on studies of soluble plumes of hydrocarbon chemicals, predominantly BTEX chemicals in unconsolidated sediments where limited extents have been documented to be less than 200 feet in length API 1998; Connor et al. 2015.

When released to land, NAPL will generally seep downward into soil toward the water table due to gravity and pressure head. A residual trail of NAPL will remain in soil as the NAPL front descends. Lenses of low-permeability or wetter soils may induce some intermediate lateral NAPL migration. The fraction of mobile NAPL depletes as a portion is left behind as immobile residual chemical in the soil column in which the NAPL is descending. NAPL migration may be limited by this depletion, or by physical barriers such as low permeability layers. Thus, smaller or more viscous hydrocarbon releases may be completely trapped as residual in the unsaturated soil zone. Larger volume and less viscous releases may migrate to the water table and become submerged, trapped at the water table and capillary fringe interface. As the water table level rises and falls, the NAPL will begin to spread laterally and “smear” into the soil matrix due to a fluctuating water table.

Furthermore, BTEX chemicals are fractionally soluble in water, have higher pure chemical solubility than other hydrocarbon chemicals, and, when present, are the dominant constituents in soluble groundwater contaminant plumes originating from petroleum hydrocarbons. Other petroleum constituents have lower pure chemical aqueous solubility and dissolved-phase mobility, and show less migration in groundwater than the BTEX chemicals. However, these other petroleum constituents may still be in the groundwater. Their suspected presence should be confirmed through sampling and analysis.

Most soluble hydrocarbon plumes in unconsolidated sediments are less than 200 feet in length measured from the downgradient edge of the NAPL plume, and stable or shrinking in extent. In terms of the time duration, studies of soluble petroleum plumes have been shown API 2012; McHugh et al. 2014 to be of limited extent with a median half-life for benzene of approximately 3.9 years due to source zone depletion. Groundwater contaminants in fractured bedrock or karst or in aquifers with high seepage velocities (> 10 cm day^-1 to 1 m day^-1) may migrate farther distances than the nominal 200-foot plume length for BTEX API 1998; Connor et al. 2015. This should be considered when developing your investigative strategy.

Degradation and transformation of petroleum hydrocarbon constituents in the environment is widely reported Zobell 1946; Atlas 1981; Leahy and Colwell 1990 and recognized under aerobic conditions, in the presence of other external electron acceptors (nitrate, manganese, ferric iron, sulfate, carbon dioxide), and in fermentative/methanogenic conditions. With O₂ present, the degradation of all gasoline range hydrocarbon constituents has been reported Prince, Parkerton, and Lee 2007. These may readily degrade to CO₂ and H₂O in a water-soluble phase. Similar results are reported Prince et al. 2013 for crude oil, with the exception of hopanes as a class. However, hopanes are not unique to crude oil but are present as components in bacterial cell membranes, and, as such, are known to be ubiquitous in the environment. Table 5-2 summarizes the relevant biodegradation processes for TPH in different media.

Table 5‑2. Transformation processes (aerobic, anaerobic, fermentative/methanogenic) for typical compositions of TPH found in impacted media

In terms of pathway-specific evaluations, surface soils, subsurface soils, and saturated soil and sediment can be directly impacted with NAPL (residual) or indirectly impacted with TPH dissolved in soil moisture. Soil exposure pathways mainly include direct contact by receptors, including humans, plants, plant roots, and burrowing animals. Migration of TPH contamination in groundwater, and the leaching or dissolution of TPH into groundwater, are potential groundwater risks, and ultimately, groundwater exposure can occur via ingestion. Similarly, NAPL advection/dispersion on surface water, and the dissolution of TPH into the surface water, are potential risks to human and ecological receptors (aquatic vegetation, fish, benthic organisms, amphibians/reptiles) through direct contact and ingestion pathways.

For potential inhalation exposures, burrowing animals can be at risk to volatile TPH in soil gas. Likewise, the transport of soil gas or direct volatilization of vapors from NAPL or dissolved TPH into the breathing zone or indoor structures can be other inhalation exposure pathways to consider in a risk assessment. Comprehensive guidance is available USEPA 2015c; ITRC 2014 on risk evaluation of hydrocarbon vapors from subsurface petroleum sources to indoor air. The guidance highlights the significant and extensive attenuation of hydrocarbon vapors over short distances in aerobic soils due to biodegradation.

Risk-based concentration criteria for different land uses, such as residential or commercial/industrial, and for each relevant environmental media (soil, groundwater, air) may be developed and often include significant conservative factors of safety. These criteria are applicable only when the exposure route is present (direct contact, ingestion, inhalation) now and in the future and complete from the contaminant source to the point of exposure (i.e., there is a complete source-pathway-receptor linkage). Although there may be no risk, there may still be regulatory requirements to fulfill (see Tiered TPH risk assessment framework).

Exclusion distances may be applied specifically for the petroleum vapor intrusion pathway USEPA 2015c; ITRC 2014; Lahvis 2017. The distances defined in these guidance documents incorporate the protective aerobic biodegradation zone between the source (NAPL or dissolved petroleum plume) and the indoor structure to mitigate any risks of petroleum vapors. Similarly, exclusion criteria can be applied to determine if an ecological assessment is warranted. As outlined in ASTM E2205 2014b, ecological assessments can be excluded based on a series of conditional questions summarized in Table 5-3. However, exclusion distances have not been adopted by all state agencies. In such cases, screening should follow local regulatory requirements.

Table 5‑3. Representative exclusion criteria for ecological risk assessment
(Source: ASTM E2205,2014.)

Understanding petroleum hydrocarbon chemistry and how that chemistry affects the fate and transport in the environment helps to define what data inputs are sufficient. Determining what and how much data are needed from the CSM to make decisions about risk specific to TPH can be difficult. Therefore, any sampling and analysis related to petroleum contamination should be discussed with a risk assessor prior to preparing a TPH data collection work plan to ensure that the appropriate data for conducting the TPH risk assessment are available.

The presence of organic materials (e.g., grass particles, algae, etc.) should be noted and minimized in the sample because these may cause interferences with TPH analyses. Likewise, physical characteristics of the sample (e.g., petroleum odor, colorations, and/or viscosities) should be noted. In particular, other odors such as solvent odors, rotten-egg odors, etc., help indicate different petroleum types in different regions, weathering and biological changes in the petroleum, or potential interferences (e.g., sulfur). Whenever these characteristics are encountered, the depth in the vadose or unsaturated zone and other descriptive characteristics (odors, staining) should also be noted in field logs. Contact between the soil or sediment samples and plastics or gloves should be minimized during the collection procedure, as these often contain organic materials that could interfere with TPH analyses.

Samples being submitted to the laboratory for volatiles analyses (i.e., BTEX or TPH fractions of ≤C12) should be immediately preserved in methanol or collected in airtight containers (e.g., EnCore samplers) to avoid volatilization of the constituents of concern. If compositing or mixing incremental samples in the field, ensure the procedures are in accordance with the methods described in ITRC ISM Section 5.4.2, where multiple incremental samples are preserved together in a large container of methanol, when possible, or as otherwise directed by the overseeing regulatory agency. Results for soils not collected in airtight containers or immediately field-preserved in methanol may be biased low due to volatilization and biodegradation of contaminants prior to laboratory extraction/preservation. This bias can be increased if methanol is not used by the laboratory due to the superior extraction efficiency of solvent extraction versus vapor partitioning Hewitt 1998. Soil and sediment intended to be used for field vapor screening should not be used for laboratory analysis of TPH due to volatile losses as well.

In terms of the sample volumes needed, ensure that minimum mass requirements are met if ISM samples are to be collected ITRC 2012. Otherwise, consult with the analytical laboratory to specify and provide sample containers (typically glass), preservatives, and shipping instructions needed for the expected TPH and indicator compound analyses (see TPH Analytical Methods).

When methanol is being considered as the preservative, detection limits should be sufficiently low to be below action/risk levels and used to determine if an additional sample needs to be collected for dry weight correction. Consult the DQOs set during the TPH data collection planning. Furthermore, the laboratory should be asked to provide methanol containing the appropriate surrogates for analysis rather than adding surrogates to the samples once they arrive at the laboratory (e.g., see Alaska GRO method AK101). This will allow for the evaluation of potential losses of analyte during the collection process and allows the surrogates to interact with the soil to estimate extraction efficiency. Note that the process of prespiking surrogates in methanol-preserved samples is a modification of some of the more commonly used methods (e.g., GRO, MassDEP VPH method).

For water samples, every effort should be taken in the field to minimize excessive turbidity. Turbid samples will include petroleum hydrocarbons that are adhering to soil particles rather than the dissolved-phase TPH in the water. Excessive turbidity, as well as droplets of NAPL being introduced into the sample, is usually induced by turbulence during fast drawdowns in the groundwater and rapid recharge or in the surface water column. While it has been known for years that bailers produce turbid samples, methods are available for reducing turbidity in samples including “no-purge” sampling tools (e.g., the SNAP sampler or Hydrasleeve), low-flow purge and sampling, redeveloping monitoring wells, and using prepack screens and “well development” for open boring/direct push sampling.

Filtering of water samples is another option but should be discussed with the risk assessor and/or regulator beforehand. Filtering may be needed to remove droplets of LNAPL, sheens, colloids, or sediment particles from the water samples because relatively large differences in TPH analytical results from sample duplicates can be attributed to these heterogeneities. However, filtering may limit the representativeness of the water sample in the risk assessment context because receptors (e.g., direct human ingestion or ecological contact) would not generally encounter filtered water.

Although it may not be possible to entirely avoid turbidity, the observation should be documented along with the presence of any NAPL, discolorations, odors, natural organic materials (e.g., algae), etc. Each of these causes interferences in TPH analyses or provides TPH data that may be misrepresentative. In addition, the sampling location, depth, lithology, and additional field screening data such as dissolved oxygen, redox, pH, temperature, etc., should be properly documented to provide a comprehensive data package for the risk assessment.

Additional groundwater or surface water samples may also need to be taken to measure geochemistry, natural attenuation parameters, and/or the toxicity of TPH degradation products. If silica gel cleanup is to be included, sufficient sample volumes will need to be collected, and the appropriate sample containers, compatible sampling supplies, preservatives, and storage and shipping details planned in advance.

For sampling soil gas vapors for TPH, one of the main considerations is determining the appropriate type of sampling instrument to use depending on the carbon range of TPH to be targeted. Suitable containers for TPH soil gas samples include Tedlar bags, gas-tight vials (glass or stainless steel), sorbent tubes, and passivated stainless-steel canisters (Summa). Tedlar bags are generally not considered to be reliable for more than 48 hours, but some agencies may have different requirements ITRC 2014. Furthermore, different types of sample containers may be needed to cover the complete TPH carbon range desired. Specifically, Summa canisters should be used for carbon ranges <C12, while sorbent samples can be used to collect soil gases containing a wider range of carbon ranges (e.g., bulk gasoline and diesel range).

Another consideration for sampling soil gas vapors is the compounds used as tracer gases during leak check procedures. Certain tracer gas compounds (isobutane, isopropyl alcohol) have a direct impact on the quantitation of TPH if present at elevated concentrations greater than 0.01% by causing (1) false positives and (2) elevated reporting limits due to significant dilutions performed by the laboratory. Other tracer gas compounds such as 1,1-difluoroethane or sulfur hexafluoride may not cause false positives, but may lead to elevated reporting limits if also present at elevated concentrations. Helium is generally considered the least problematic tracer gas because it can be measured in the field and does not lead to either type of interference with TPH analyses. However, helium itself may have to be quantified in the sample using additional analysis at the laboratory. For additional information, see ITRC 2014.

Considerations in selecting one or more analytical methods may include:

• Project objectives
• Regulatory requirements
• Application (detection, delineation, monitoring, risk assessment, etc.)
• Petroleum type, if known

Method selection includes specification of:

• Extraction solvent(s)
• Applied calibration standards
• Carbon number reporting ranges

Potential issues may include:

• Over- or under-representation of certain ranges/fractions or compounds
• Applicable method boiling point/carbon number range relative to the sample range
• Elevated reporting limits
• Method interferences
• Accuracy and precision

Multiple methods may be specified for different purposes at the same site. Consultation between the risk assessor, project manager, and analyst may be required.

SGC methods, including EPA Methods 418.1/1664A (SGT-HEM) and 3630C, and ISO Methods (used in Europe) 11046B (for soil) and 9377-2 (for water), generally involve the introduction of SG to the extraction sample to partition the polar compounds from the hydrocarbons and/or fractionate the sample into aliphatics and aromatics. However, the effectiveness of these methods can vary depending on the amount of SG used, the rinse solvent, and the separation step (shaking/stirring in the extraction vial or running the extract through an SG column) called for in the method Zemo, Synowiec, et al. 2013.

The nonhydrocarbons found in soil and groundwater at biodegrading petroleum release sites are most frequently oxygen-containing molecules that are metabolites from biodegradation of the petroleum, such as organic acids/esters, alcohols, ketones, aldehydes, and phenols Zemo and Foote 2003; Zemo, O’Reilly, et al. 2013; Lundegard and Sweeney 2004; Mohler et al. 2013. However, oxygen-containing nonhydrocarbons in a crude oil could have molecular structures similar to biodegradation products from oil recently spilled. In a somewhat special case of nonhydrocarbons, sulfur or sulfur-containing compounds have been found in groundwater extracts at sites where nonpetroleum organic matter from any origin is undergoing sulfate reduction Lundegard and Sweeney 2004.

Currently, it is still a challenge separating polar from nonpolar compounds using SGC (e.g., any moderately polar compounds will be retained in the silica matrix of a silica gel column, including any that increase in polarity as a result of biotransformation). This may result in incomplete separation of metabolites from nonpolar fractions. Although some well-resolved components could be eliminated by subtraction, incomplete separation does not address any unresolved complex mixture present.

Furthermore, interferences are most commonly attributed to low concentrations of contamination in the SG or lab glassware, or could be due to insufficient activation of the SG.

Sometimes with lower RLs, the laboratory may have to perform dilutions that cause the RLs to be elevated. To ensure that the dilution performed was reasonable, a check should be made to determine if there are elevated concentrations of other target analytes in the sample. If there are other target analytes present at very high levels, then the dilution is likely justified and the presence of elevated RLs may not be an issue. However, if a dilution was performed without an obvious reason (e.g., low concentrations or nondetect results for target analytes), then the laboratory should provide an answer as to why the dilution was performed. This could happen due to elevated concentrations of nontarget compounds, typically for indicator compounds in VOC and SVOC GC/MS analyses.

For some methods that extract organics from soil and water, typically a 1 mL extract is generated. However, if the sample is excessively viscous, the laboratory may not be able to reduce the volume to the target amount. In this case, the excess volume of the extract results in a dilution factor that should be explained in the report from the laboratory. For example, if the target extract volume is 1 mL but the sample could be reduced only to 5 mL, then a 5-fold dilution should be reported and explained. For soil vapor or indoor air samples in canisters, if the vacuum is too high, the laboratory may overpressurize the canisters because the autosampler may not be able to extract the sample from the canister under such high vacuum. The overpressurization similarly results in a dilution factor.

TPH chromatograms have been used for decades to identify the petroleum constituents and “interferences” present in soil and groundwater samples Zemo, Bruya, and Graf 1995; Kaplan and Galperin 1996. The chromatograms from routine regulatory TPH analyses such as Method 8015 can be used if the resolution and scale are adequate (refer to Figure 2.1). However, the fact that the sample is separated into purgeable and extractable portions (except when using Method TX1005) must be considered in the interpretation. The highest quality chromatograms are those generated using high resolution GC/FID or GC/MS (total ion) for “forensic” analyses.

5.12.1 Potential Effects of Holding-Time Exceedances on TPH Results

When holding times are missed, there is the potential for the TPH data to be biased low. The magnitude of the bias is dependent on the extent of the holding-time exceedance, the matrix (including the presence and abundance of hydrocarbon-degrading microbes), and preservation/collection methods, as well as the type of petroleum product causing the contamination. The lighter, more volatile petroleum products (e.g., gasoline, Stoddard solvent, kerosene) will be affected more significantly by a holding-time exceedance than the heavier petroleum products (e.g., fuel oil #6, motor oil).

Generally, the rule of thumb is that if the extraction and/or analysis holding time is grossly exceeded USEPA 2017d, nondetect TPH results may be considered unusable for project objectives. However, professional judgment based on the nature of the petroleum product, etc., may be used to determine the ultimate effect on the data.

5.12.2 Potential Effects of Blank Detections on TPH Data Interpretation

If there is TPH or hydrocarbon range contamination in a laboratory method or field blank, this may be caused from extraneous peaks, column bleed, equipment contamination, or possibly something that is within the laboratory’s control that probably should have been fixed prior to analyzing the samples. Many times, it is not due to a petroleum hydrocarbon pattern so when comparing blank results to sample results, it is best to review the chromatograms of both the blanks and the samples to make sure the contamination is the same pattern in both before applying the 2x rule described below. If the chromatographic profile does not match the profile found in the blank, applying the 2x rule would result in a false negative.

The general rule of thumb to follow during this evaluation is that if the concentration in the sample is less than 2x the blank concentration and the petroleum hydrocarbon contamination pattern in the sample matches the petroleum hydrocarbon contamination pattern in the blank, the result in the sample is potentially a false positive. Note that the use of 2x the blank concentration is provided as a general guideline USEPA 2017d; local regulatory data evaluation guidelines should be referenced.

Furthermore, blanks should be scrutinized to help determine whether results near the method detection limit are indeed accurate or potentially false positives.

5.12.3 Potential Effects of Laboratory Control Sample Results on TPH Data Interpretation

For different TPH analyses, the analytes spiked into a laboratory control sample (LCS) vary depending on the method to be performed. For example, fuel oil #2 may be used for DRO analyses, individual hydrocarbon components may be included for VPH/EPH analyses, etc. In the case of EPA 8015 analyses, a study USDOD 2018 generated the performance-based limits summarized in Tables 5-6 and 5-7 (adapted from DOD QSM 5.1 Appendix C).

Table 5‑6. LCS control limits for SW-846 8015(mod) solid matrix in percent
(Source: DoD,2017.)

Table 5‑7. LCS control limits for SW-846 8015(mod) water matrix in percent
(Source: DoD,2017.)

The results of the LCS affect the entire batch of samples prepared or analyzed with the LCS, depending on the method. If LCS recoveries are outside the acceptance limits, the following guidelines are used in the assessment of the data:

• If low recoveries, then there is a potential low bias for that analyte in all associated samples in the batch: affects detected and nondetect results.
• If high recoveries, then there is a potential high bias for that analyte in all associated samples in the batch: affects only detected results.
• If recoveries are significantly low (<10%), nondetect results for that analyte in all associated results in the batch may not be usable for project objectives.
• If LCS/lab control sample duplicate (LCSD) relative percent differences (RPDs) are higher than the acceptance criteria, both detected and nondetect results are uncertain but the bias is considered indeterminate.

5.12.4 Potential Effects of Surrogate Recoveries on TPH Data Interpretation

Table 5-8 summarizes some of the analytical/extraction surrogates typically used in TPH and related analyses. Surrogate recoveries can be impacted by different pH conditions. If not noted, QA/QC results could indicate that the sample result is invalid because of low surrogate recovery, although the sample results are in fact due to something inherent in the sample that can have important implications on site conditions.

Table 5‑8. Surrogate compounds typically used in TPH analyses

If surrogate recoveries are outside of the acceptance limits, the guidelines shown in Table 5-9 are used in the assessment of the data:

Table 5‑9. Out of bounds surrogate recovery guidelines

Note that there would be no effect on the usability of the data if the surrogate recovery is outside acceptance limits and a significant dilution was performed on the sample (greater than 5-fold dilution). The guidelines listed above may not apply if the sample chromatogram indicates significant interference in the area where the surrogate elutes. In these instances, matrix interference is causing problematic surrogate recoveries and the evaluation of sample biases due to surrogate recoveries may not be possible. Therefore, it is important that the chromatogram is provided by the laboratory and included in the evaluation of sample data due to surrogate nonconformances.

5.12.5 Potential Effects of Matrix Spike (MS)/MS Duplicates (MSD) on TPH Data Interpretation

If MS or MSD recoveries are outside the acceptance limits, the following guidelines are used in the assessment of the data:

• If low recoveries, then there is a potential low bias for that analyte in the sample that was spiked: affects detected and nondetect results.
• If high recoveries, then there is a potential high bias for that analyte in the sample that was spiked: affects only detected results.
• If recoveries are significantly low (<10%), nondetect results for that analyte in the sample that was spiked may not be usable for project objectives.

The above guidelines do not apply to samples that already contain TPH at concentrations significantly (greater than 4x) exceeding the spike amount.

5.12.6 Evaluating and Interpreting Breakthrough

The amount of solvent (i.e., hexane) used to elute the aliphatic component of the hydrocarbon mixture is critical. An excessive volume of solvent may cause the lighter aromatics to break through and be captured in the aliphatic fraction while an insufficient volume of solvent may allow some of the heavier aliphatic hydrocarbons to be retained on the silica gel cartridge/column, resulting in a lower recovery for these aliphatic fractions. Depending on the analytical conditions, this could result in an underestimation of the aromatic carbon range concentration for the excessive solvent condition or an overestimation of the aromatic carbon range concentration for the deficient solvent condition. If aromatic breakthrough is suspected, the aliphatic fraction may be analyzed to determine if naphthalene or any of the other more “mobile” aromatics are present.

Each sample (field and QC sample) must be evaluated for potential breakthrough on a sample-specific basis by evaluating the percent recovery of the fractionation surrogates and on a batch basis by quantifying naphthalene and 2-methylnaphthalene in both the aliphatic and aromatic fractions of the LCS and LCSD. If the concentration of either naphthalene or 2-methylnaphthalene in the aliphatic fraction exceeds 5% of the total concentration for naphthalene or 2-methylnaphthalene in the LCS or LCSD, then there are potential biases to the data.

• Potential low bias exists for the aromatic fraction (≥C11), naphthalene, and 2-methylnaphthalene results in all associated samples.
• Potential high bias exists for the lower carbon range aliphatics (≥C9) in all associated samples.

It should be noted that breakthrough could also occur on an individual sample basis, regardless of whether the LCS exhibited acceptable fractionation results. If the sample contains a high content of PAHs/hydrocarbons, it is possible that PAHs may break through into the aliphatic fraction. This would not be as obvious, because analyses of sample aliphatic fractions for naphthalene and 2-methylnaphthalene are not required. Review of fractionation surrogate recoveries and sample chromatograms can assist in the determination of whether any significant breakthrough has occurred in a sample; low recoveries of fractionation surrogate in the aromatic fraction may indicate a potential breakthrough issue in a sample. In such cases, special considerations may be required (e.g., dilution required prior to re-fractionation). It should be noted that acceptable recovery of the fractionation surrogates may not always provide absolute confirmation that effective separation of the aliphatic fraction from the aromatic fraction of the sample extract has been accomplished.

There is also the possibility of breakthrough of aliphatics, like n-alkanes, in crude oils and diesel fuels to the aromatic fraction. This could be a source of high bias in the aromatic fraction. This is relatively easy to “see” upon inspection of the chromatograms but not easy to correct.

It should be noted that SGC is not 100% selective. There may be some aliphatics in the aromatic fractions and vice versa. For un-degraded crude oils, for example, that have high n-alkane content, even 90%+ selectivity/efficiency of the silica gel would result in some of the n-alkanes being visible in the aromatic fraction. This may not be breakthrough per se, just not 100% selectivity. Also, as the degree of substitution increases around an aromatic core, the aromatic character is reduced and not necessarily selective to the silica gel; some of the alkylated substituted aromatics with aliphatic side chains will split into the aliphatic and aromatic fractions and at some point, the highly alkylated aromatics will end up with the aliphatics. This may be acceptable from a risk assessment perspective because the aromaticity of these highly alkylated aromatic compounds is not dominant and these compounds may likely behave as aliphatics in the environment.

5.12.7 Potential Effects of Co-Eluting Contaminants on TPH Results

Chlorinated solvents can give rise to false positives for TPH. Table 5-10 summarizes some common nonpetroleum contaminants that can cause false positive results in the volatile TPH (e.g., VPH and GRO) analyses.

Table 5‑10. Potential nonpetroleum interferences MADEP 2009
(Source: MaDEP, 2009.)

5.12.8 Avoiding Double Counting of Indicator Compounds in Fractionated TPH Data

Many of the fractionated TPH methods (e.g., VPH and EPH) report aliphatic and aromatic hydrocarbon ranges as well as target indicator compounds (e.g., BTEX or PAHs). During the analysis of samples by these methods, the concentrations of target compounds that fall chromatographically within specific hydrocarbon ranges are excluded from the hydrocarbon range data because the target compounds and hydrocarbon ranges are evaluated for risk separately. Double counting refers to the instance where this correction is not made to the hydrocarbon range data, thereby resulting in high-biased hydrocarbon range results and an overestimation of TPH risk. It should be noted that correction to the hydrocarbon range data for target compounds should be done using the target compound concentrations derived from the method of analysis used for reporting the hydrocarbon range data. For example, correcting VPH by GC/PID/FID data by subtracting out BTEX concentrations determined by Method 8260 is not appropriate. Instead, the BTEX values determined by the VPH by GC/PID/FID method should be used for correcting the hydrocarbon range data in this case. Risk must be evaluated using benchmarks based on whether the data are or are not corrected for target compound concentrations.

5.12.9 Potential Issues Associated with TPH Chromatograms

Due to the nonspecific nature of the TPH analysis, TPH chromatograms should be routinely reviewed as part of the QA/QC process before TPH analytical results are used for decision making. TPH chromatograms should be reviewed to resolve issues associated with overlapping reporting ranges even though a single product type is known to exist at a site. As shown in Figure 5-4, the only product present in this example is diesel fuel #2; however, concentrations of GRO would also be reported even though gasoline was not present. It should be noted that some analytical methods do not have overlapping carbon ranges, thus avoiding this problem.

Figure 5‑4. Example of overlapping reporting ranges for TPH Zemo 2016.
(Source: Zemo, 2016.)

TPH chromatograms can also exhibit “drifting baselines.” This phenomenon can occur due to carryover from previous analyses of dirty samples or an instrument gradually cleaning itself if it was calibrated following dirty samples. Likewise, the FID signal can gradually increase as the GC column heats, producing a baseline that ramps up on a very subtle level. The most common problem that this presents is that laboratories may select the baseline as the bottom of the resolvable peaks that do not include all of the mass in the unresolved complex mixture (UCM). This improper baseline integration can result in reported concentrations that are significantly lower than what is actually present. Figure 5-5 illustrates this concept.

Figure 5‑5. Baseline integration example: Drifting baseline.

5.12.10 Evaluating Potential Uncertainty in TPH Data

Estimating the error associated with individual results will ultimately impact the uncertainty associated with the assessment of risk. The heterogeneity of the sampled matrix, particularly soil or sediment, may be of concern, as it may cause the overall error to be unacceptably large, may limit the usability of the data (or may even render data unusable), and could ultimately result in the need for resampling Gy 1998; Pitard 1993; HIDOH 2016; ITRC 2012. The uncertainties in results are exacerbated by releases of multiple types of petroleum products, releases over time, and/or if the evaluation of degradation products are included in the investigation.

To identify uncertainties and errors in data, DQOs set during the development of the TPH data collection plan should be consulted and thoroughly discussed while the plan itself is being developed. If unacceptable errors are found, adjustments to the data collection plan may need to be implemented for any additional sampling, including additional replicates, QA/QC samples, and a deeper evaluation of the source(s) of the errors. In addition to using the DQOs included in a data quality review, the evaluation of the usability of TPH data should always include a review of sample chromatograms (see the 2x blank rule, for example). Even if data should be rejected based on severe QC issues, this does not automatically indicate that resampling/reanalysis is required. Rather, the entire data set and other lines of evidence should be evaluated to determine whether the data gap is critical, thereby requiring corrective action.