The method used to preserve soil samples for VOC determination can define the future of a site
Experimental evidence shows that the sampling protocol explains more variability in the recovery of Volatile Organic Compounds (VOCs) than soil texture itself, with direct implications for analytical results and for Contaminated Land Management decisions. The study compared three soil matrices under controlled, known contamination, isolating the effect of the sample collection, preservation, and transport method on the final concentration measured in the laboratory. The findings reinforce that the choice of sampling protocol can significantly alter the conclusion about the extent and severity of contamination.
The robustness of an environmental investigation depends, ultimately, on the representativeness of the analytical data relative to the real condition of the sampled matrix. This study experimentally assessed the effect of the sampling protocol on the recovery of Volatile Organic Compounds (VOCs) in three soil matrices, under controlled, known contamination.
The results demonstrate that the preservation protocol accounts, in a statistically robust way, for more variability in analytical recovery than the textural heterogeneity of the soil itself, a finding with direct implications for the interpretation of laboratory reports against CONAMA Resolution 420/2009 (Brazil's national soil and groundwater quality standard) and for decision-making in site management. None of the protocols evaluated is free of technical limitation, and the choice of the most appropriate method must be justified in light of the objective of the investigation, not assumed by default.
A single soil, spiked with a doping solution whose reference concentration is known and identical, can yield a technically “compliant” result or a result indicating the need for risk assessment and intervention, depending exclusively on the collection and preservation protocol employed, even before the sample reaches the laboratory.
The risk that begins before the laboratory
Decisions on confirmatory investigation, remediation, or closure of contaminated sites are anchored in the comparison between the concentration determined in the laboratory and a regulatory reference value, in Brazil typically the limits of CONAMA Resolution 420/2009 and the Board Decisions of CETESB (the environmental agency of the State of Sao Paulo). This process assumes that the reported concentration reflects, with acceptable fidelity, the concentration actually present in the matrix.
For VOCs, this premise is particularly fragile: they are substances with high vapor pressure and a strong tendency to partition into the gas phase, which makes every stage of sampling, transport, and preparation a potential point of loss of the analyte itself. An inadequate collection protocol does not introduce random noise: it introduces systematic bias, which underestimates the real concentration.
The reference technical standard (USEPA SW-846, Method 5021A) establishes three sampling alternatives for VOC screening by headspace: preservation in volatile-free water, preservation in methanol, and sampling in a hermetically sealed device without preservative. This study sought to quantify the profile of advantages and risks of each alternative.
Experimental design
We conducted a comparative bench-scale study, using three soil types of contrasting textures (clayey, sandy, and silty), collected in the metropolitan region of Porto Alegre/RS and free of VOC contamination. Each matrix was spiked with a reference solution of common type C gasoline and water, and subjected to the three protocols in triplicate, totaling 27 samples sent to an accredited laboratory, with determination by GC-Headspace in accordance with USEPA 5021A / 8260D. Ten compounds were quantified: trimethylbenzene isomers, styrene, isopropylbenzene, n-propylbenzene, benzene, toluene, ethylbenzene, and xylene isomers, a set characteristic of contamination by light petroleum derivatives.
Finding 1: the protocol accounts for recovery; the soil texture, in most cases, does not
The difference between protocols was substantial, consistent, and statistically robust. On average, considering the 10 compounds and the 3 soils evaluated, methanol preservation recovered 56.6% of the reference concentration; water preservation recovered 19.6%; and sampling without chemical preservation recovered only 1.7%.
A two-way ANOVA (soil and method), applied to each of the 10 compounds, confirms the pattern: the effect of the preservation protocol was statistically significant for all 10 compounds (p<0.001), while the effect of the soil matrix was significant for only 4 of them (styrene, isopropylbenzene, ethylbenzene, and m,p-xylene), including benzene itself among the 6 with no detectable effect.
In practice, this indicates that the textural heterogeneity of the site exerts a secondary influence on the reliability of the analytical result when compared with the decision, made in the field, of which preservation protocol will be adopted. The main point of analytical quality control lies neither in the laboratory nor in the selection of the sampling area, but in the collection procedure itself.
Finding 2: the most critical scenario is sampling without chemical preservation
The protocol without chemical preservation, operationally the simplest and therefore the most susceptible to adoption under cost and schedule pressure, showed the most concerning performance of the study. In two of the three soils evaluated, it recovered less than 1% of the reference concentration; in the matrix with the highest organic matter content, recovery reached 4.4%. The case of greatest regulatory relevance is that of benzene, carcinogenic to humans (Group 1, IARC) and a central parameter in any human health risk assessment involving petroleum derivatives: in the protocol without preservation, it was not detected in two of the three soils evaluated, and its average recovery did not exceed 0.12% of the reference value.
In practical terms, a laboratory report based on sampling without adequate preservation may report “benzene not detected” in an area where the compound is, in fact, present at a concentration substantially higher than measured, not because of laboratory error, but because of physical loss of the analyte between collection and analysis.
The counterpoint: methanol is not a protocol free of analytical risk
The results above might suggest, in a simplistic way, that methanol should be universally adopted as the standard. That reading would be incomplete: preservation in an organic solvent carries its own limitations, which must be weighed against the objective of each investigation.
Elevation of the quantification limit (LOQ)
The addition of solvent to the sample dilutes the soil before injection into the chromatograph, raising the quantification limits. This makes the protocol less suitable when the objective is to confirm the absence of contamination at trace concentrations, since a higher LOQ can mask compounds at very low concentration.
Risk of cross-contamination and overestimated results
Because of its high affinity for VOCs present in the atmosphere of the preparation environment, methanol is subject to absorbing external compounds, inflating the result. In Soil 1, the average recovery of isopropylbenzene in methanol reached 102%, and one replicate of n-propylbenzene reached 122%, values that should not exceed the reference of 100%.
Suitability dependent on the objective of the investigation
By minimizing losses via biodegradation and volatilization, methanol tends to be more suitable when there is already well-founded suspicion of elevated concentrations. In initial screenings, with lower expected concentrations, the elevation of the LOQ may represent a disadvantage relative to the aqueous protocols.
The picture that emerges is not one of a single “correct” method and two “incorrect” ones, but of a spectrum of protocols with distinct error profiles: the protocol without preservation tends toward the false negative through physical loss of the analyte; methanol, although superior in average recovery, presents an occasional risk of overestimated results and of an elevated LOQ for low-concentration screenings. The appropriate choice depends on the stage of the investigation, the expected concentration, and the level of confirmation required by the decision to be made.
What this means for site management
The impact of this type of analytical bias is not merely technical: it is decisional. In a contaminated site management process, the analytical result guides every subsequent stage, and a systematically underestimated result, such as that observed in the protocol without preservation, can generate three practical consequences:
A systematically overestimated result, a risk also identified in the methanol protocol, has a distinct but equally relevant consequence: it can direct resources to an unnecessary investigation or remediation, or undermine the technical credibility of a report before an environmental agency or negotiating counterparty. In both directions, the sampling protocol is not an operational field detail: it is a technical decision with weight equivalent to that of the choice of the accredited laboratory.
Practical recommendations
- Prioritize methanol preservation for confirmatory investigations, with reinforced quality control. It is technically more suitable when the objective is to characterize contamination that is already indicated, accepting the trade-off of a higher LOQ.
- Reinforce QA/QC when methanol is adopted, with the systematic inclusion of field and method blanks in every campaign, given the identified risk of overestimated results.
- Do not use sampling without chemical preservation for regulatory compliance purposes. Treat it as a preliminary screening step, never as a definitive result against regulatory limits.
- Standardize the protocol across sampling campaigns. Since the variability between protocols exceeded the variability between soils, variations in the result should reflect a real change in the environmental condition, not a change of method.
- Document and technically justify the chosen protocol in contracts and terms of reference, with the same level of detail today dedicated to the choice of the accredited laboratory.
Methodological note
This study was conducted at bench scale, with three soil matrices and controlled contamination in a laboratory environment, a design that allows the effect of the sampling protocol to be isolated more cleanly than would be possible in the field, but which also implies that the results should be read as evidence of a pattern, not as a universal correction factor applicable to any real site. The statistical tests are valid, with 27 replicates per compound; the relationship between recovery and the specific physicochemical properties of the soil, however, was the object of internal exploratory analysis, without sufficient statistical validation for generalization.
Study source: Technical team of LZ Ambiental. “Comparative Study of Efficiency between Soil Sample Preservation Methods for the Determination of Volatile Organic Compounds”. Paper presented at the III AESAS Conference, Contaminated Site Management (Sao Paulo, 2025). Study conducted by LZ Ambiental. (link pending validation) The text of this article is an original synthesis by LZ Ambiental.
LZ Ambiental designs and executes technically defensible sampling protocols for the investigation and management of contaminated sites throughout the country. To assess whether the preservation protocol for your samples is appropriate to the matrix and the contaminant of interest, talk to our technical team.
