When the Blank Tells a Story
Your firm is validating soil data for a risk assessment at a former dry cleaner. Twelve samples were analyzed for volatile organic compounds. Trichloroethene (TCE) was detected at 3.2, 4.8, and 2.9 µg/kg. The method blank contained TCE at 2.1 µg/kg. The laboratory control sample recovered 68% for TCE; surrogates were within limits. The report states samples were collected in EnCore samplers and analyzed 12 days after collection. Two samples were not preserved in methanol. The laboratory reporting limit is 1.0 µg/kg. The risk assessor wants to use the three detects as valid because they exceed the blank. The project decision is whether to excavate a 15-by-20 m hot spot. The client wants the report by Friday.
1. Apply data validation principles to the TCE results. How should the method blank, LCS recovery, and holding time affect the usability of the three detects?
Read the worked answer
The three detects are not usable as reported without qualification. The method blank contains TCE at 2.1 µg/kg, which is 72% of the 2.9 µg/kg result and 44% of the 4.8 µg/kg result. A common validation rule is that sample results below ten times the blank concentration are suspect; all three detects are below 21 µg/kg, so they should be qualified as estimated or potentially affected by blank contamination. The LCS recovery of 68% is below the typical 70–130% acceptance window, indicating a low bias; that means the true concentrations could be higher, but the blank problem still prevents using the numbers as definitive. Holding time is the larger issue: EnCore samplers without methanol preservation have a short holding time, typically 48 hours, and these samples were analyzed 12 days after collection. The two unpreserved samples are outside holding time and should be rejected for TCE. For any methanol-preserved samples, the 14-day holding time may be met, but the report does not clearly identify which samples were preserved. The data are therefore usable only as screening-level information, not as the basis for a hot-spot excavation decision.
2. What resampling and laboratory controls would you require before the risk assessment decision, and how would you document the data quality?
Read the worked answer
Before making the excavation decision, I would require resampling of the affected locations using proper VOC soil procedures. Samples should be collected in EnCore or methanol-preserved containers according to the method, chilled, and analyzed within the holding time; the sampling plan should specify which preservation option applies to each sample. I would add a trip blank, a field duplicate, and a matrix spike/matrix spike duplicate for TCE, and I would require the laboratory to demonstrate a method blank below the reporting limit and an LCS within acceptance limits before sample analysis. If the hot spot is defined by the 4.8 µg/kg result, I would resample that boring and adjacent locations. The data validation report should list each sample, its preservation, holding time, blank qualification, LCS recovery, and final usability code. The risk assessment should state that the original three detects are estimated and were not used quantitatively; the resampled data should be the basis for the excavation boundary. The Friday deadline should not override the need for defensible data, and the client should be told that the schedule may need to move.
3. The risk assessor argues that because all three detects exceed the blank, they are valid. How would you respond, and what does the project decision require?
Read the worked answer
I would explain that exceeding the blank is not sufficient. A result must also be sufficiently above the blank to be distinguishable from laboratory contamination, and the common ten-times rule is a conservative screen, not a pass/fail test. Here the 2.9 µg/kg result is only 0.8 µg/kg above the blank, which is within the laboratory's own variability at a 1.0 µg/kg reporting limit. The 68% LCS recovery adds a low-bias uncertainty, and the unpreserved samples have a holding-time failure. The risk assessor is treating 'greater than the blank' as proof of a real detection when the data are at best estimated. For a 15-by-20 m excavation decision, the consequences of a false positive include unnecessary soil removal, disposal cost, and disturbance; a false negative could leave contamination in place. The project needs resampled, properly preserved data with acceptable QC. I would document the disagreement and the basis for qualifying the original results, and I would not sign a risk assessment that uses the three detects as definitive.