Rationale for Soil‐Gas Sampling to Improve Vapor Intrusion Risk Assessment in China

Abstract

Contaminated site investigation and remediation are becoming increasing common in China since the implementation of the Action Plan for Prevention and Control of Soil Pollution (2016), the Soil Pollution Control Act (2019), and an overall desire by the national government to return >100,000 industrial sites closed since 2001 and >2 million hectares of brownfield land to productive use. Many of these sites are contaminated with volatile organic compounds (VOCs) that represent a potential risk for vapor intrusion (VI) (Ma 2019). To date, environmental guidelines and regulations for VOC and soil-gas sampling on a national level are lacking in China even though VI risks are predicted from soil concentration measurements assuming equilibrium partitioning relations (MEE 2014). Future VOC guidelines for VI risk assessment in China should consider building on experience and learnings from the United States on soil-gas sampling. A rationale and recommended methods for soil-gas sampling for VI risk assessment are provided in this editorial. VOC concentrations in soil gas cannot be adequately predicted based on soil concentration measurements (McAlary et al. 2011; Ma et al. 2018; Ma 2019). The authors are not aware of any field studies that clearly show meaningful correlations between paired VOC concentrations in soil and soil gas (Ma 2019). In particular, benzene has been observed at very high concentrations in soil gas (>2000 μg/m3) where corresponding soil concentrations are below detection limits (Zhang et al. 2019). VOC vapor concentrations can also be orders of magnitude less than those predicted based on soil concentrations and equilibrium vapor partitioning relations (Golder-Associates 2008). VOC concentrations in soil gas can also be overestimated in soil samples where nonaqueous phase liquid (NAPL) is present and unaccounted for in three-phase (soil, pore-water, vapor) partitioning relations (U.S. Environmental Protection Agency [U.S. EPA] 2015). The distribution of NAPL in soils can also be highly heterogeneous, contributing to the variability and poor correlation (Ma 2019). Volatilization (loss) of VOCs during soil sampling is difficult to avoid (Ma 2019). Based on these learnings, soil-gas sampling is preferred over soil or groundwater sampling for VI risk assessment. In addition, soil-gas sampling is preferred because the methodology provides a direct measure of the media (vapor) involved in the exposure pathway (Ma 2019). Depth-discrete soil-gas sampling is recommended to validate the variability of soil-gas data between vapor sources and building foundations, which is of particular importance at petroleum VI sites where biodegradation of VOCs is often significant. Soil-gas sampling from immediately below building foundations (subslab sampling) will also be needed in China to support the development of regionally specific risk-based screening and cleanup levels. U.S. regulatory values derived from empirical studies of vapor attenuation across building foundations will likely have limited applicability in China given the contrast in typical residential building types (single-family in the United States vs. high-rise in China) and transferability of empirically derived slab attenuation factors. Subslab sampling in China will also be affected by regulatory acceptance of institutional and engineering controls. Although soil-gas sampling is recommended over soil and groundwater sampling for VI investigations, it is important to understand that soil-gas sampling has certain limitations (Ma 2019). In particular, VOC soil-gas concentration data can exhibit substantial spatiotemporal variability, that will generally increase with increasing distance from a VOC source and proximity to building foundations. Soil-gas concentration data may also be a poor predictor of VI risks at sites with future building construction because of potential effects of buildings and building foundations on VOC distributions in the subsurface. Such scenarios are likely to be a common occurrence in China given the large number of redevelopment sites anticipated in the coming years. These sites will require predictions of VI risks based on soil and groundwater data. A variety of methods have been developed for soil-gas sampling, including Tedlar bags, canisters, and sorbent tubes. Tedlar bags are generally used for field screening but not suitable for laboratory analysis due to low recoveries (U.S. EPA 2015; Ma 2019). Canisters are commonly used in North America while sorbent tubes are more popular in Europe (Woolfenden 2010). Each method has its own advantages and disadvantages. Sorbent tubes are better suited for sampling less volatile VOCs and semivolatile VOCs, although some more volatile VOCs, such as vinyl chloride, can also be adequately trapped using proper sorbents (Ma 2019). Canisters are ideal for most VOCs of concern for vapor intrusion, including ultra-volatile reactive compounds such as H2S, but the recoveries of certain longer-chain (heavier) hydrocarbons (>C10 hydrocarbon) can be poor (Ma 2019). Based on this experience, sorbent tubes are generally the preferred methodology for soil-gas sampling in China because the method: (1) is less prone to false positive VOC concentrations caused by residual contamination from previous sampling events (McHugh et al. 2018); (2) provides a means to detect lower and wider ranges of VOC concentrations (Woolfenden 2010); and (3) is easier to handle, transport, and ship to analytical laboratories, which are still relatively scarce within the country. Guidance and training will be needed in China to promote best practices for sorbent tube sampling that include addressing inherent complexities of measuring flow rates before, during and after sampling (using high-precision flow meters), selecting appropriate sorbents, and screening soil-gas concentrations to prevent VOC breakthrough. Sorbent tubes may be less relevant at petroleum VI sites because associated analytical methods do not include the quantification of fixed gases (oxygen, carbon dioxide, methane, and nitrogen), which are key indicators of hydrocarbon biodegradation. These gases, with the exception of nitrogen, can, however, be analyzed in situ using landfill gas meters and generally at sufficient quality supporting vapor intrusion risk assessment (Ma 2019). In summary, we urge China to strongly consider the development and promotion of national and provincial guidelines for soil-gas sampling and analysis in an effort to improve VI risk at VOC-contaminated sites. The soil-gas sampling guidelines can largely be adopted from the United States and other countries where established practices and standard operating procedures already exist. It is however, important for China to adapt the guidelines to address a VI exposure model that is unique with respect to future redevelopment and building construction. This work was supported by Natural Science Foundation of China (21407180, 21878332), Beijing NOVA program (Z181100006218088), PetroChina Innovation Foundation (2018D-5007-0607), Science Foundation of China University of Petroleum-Beijing (2462018BJC003), and the Independent Project Program of State Key Laboratory of Petroleum Pollution Control (Grant No. PPCIP2017004), CNPC Research Institute of Safety & Environment Technology. J. Ma, PhD, corresponding author, is at State Key Laboratory of Heavy Oil Processing, Beijing Key Lab of Oil & Gas Pollution Control, China University of Petroleum-Beijing, Beijing 102249, China; ORCID ID: 0000-0003-3719-3814, email: rubpmj@sina.com M. Lahvis, PhD, is at Shell Global Solutions (US) Inc., Shell Technology Center, TX 77082.