Batch QC Failure: Duplicate Discrepancy, Elevated Blank, and Expired CRM Calibration
A measurement professional deploys twelve charcoal liquid scintillation detectors for a 72-hour test in a 12-unit apartment building, with one duplicate pair in unit 4B and one field blank. Results: the 4B pair reads 3.1 and 9.4 pCi/L; the field blank reads 0.8 pCi/L; the laboratory lower limit of detection is 0.6 pCi/L. The project QC plan requires duplicates to agree within 20 percent relative percent difference and blanks to be below the lower limit of detection. A concurrent CRM check in the basement reads 2.1 pCi/L while the charcoal detector in the same room reads 5.6 pCi/L. The professional's CRM calibration certificate is 13 months old, and the manufacturer specifies annual calibration. The building owner is under contract to sell and wants all results released as-is.
1. Evaluate the QC data: which results can be released?
Read the worked answer
The duplicate pair fails. The relative percent difference is the absolute difference divided by the mean, times 100: 6.3 divided by 6.25, times 100, which is about 101 percent. That is far above the 20 percent criterion and invalidates the 4B result; a single value cannot be chosen as the unit result.
The field blank at 0.8 pCi/L exceeds the lower limit of detection of 0.6 pCi/L, which fails the blank criterion and signals possible contamination during handling, transport, or storage, or elevated laboratory background. Until the blank issue is resolved, the batch is suspect. The remaining ten unit results are not automatically invalid, but their reliability depends on the blank investigation; if contamination is confirmed, all results from that batch may be biased high and may need retesting.
The 5.6 pCi/L charcoal result cannot be confirmed or refuted by the CRM because the CRM calibration is expired.
2. What is the most likely explanation for the duplicate and blank results, and what should you do next?
Read the worked answer
A 101 percent duplicate disagreement is too large for detector precision alone. Likely causes include one detector being exposed to a different micro-environment, such as near a door, window, or radon entry point; a handling error, such as a detector left open or stored near a source; or a laboratory analysis error. The elevated blank points to handling or laboratory background.
Next steps: check deployment logs and photos for the 4B pair; confirm detector serial numbers and return dates; contact the laboratory to review the blank and request re-analysis or a replacement batch; review storage and shipping records. If the cause is field placement, retest 4B with two new detectors placed side by side and a new blank. Do not release 4B until a valid duplicate is obtained.
Also review whether the blank was opened at the site and resealed properly, and whether detectors were stored in a basement or near a radon source before deployment. Correct the procedure that allowed the blank to be exposed.
3. How do you handle the CRM discrepancy and the owner's request to release all results?
Read the worked answer
The CRM cannot adjudicate the 5.6 pCi/L result because its calibration is 13 months old against a 12-month requirement. Its readings are for information only and should not be reported as a valid measurement. Recalibrate or replace the CRM before further use, and document the calibration status.
The charcoal result stands unless the blank investigation implicates the batch. Tell the owner in writing: 4B is invalid pending retest; the blank is under investigation; the remaining units are being reviewed. Provide a revised report rather than releasing invalid data. If the closing schedule is tight, offer expedited retesting of 4B and any affected units.
Releasing known-invalid results would violate the QC plan and could mislead the buyer. The professional's obligation is to report only data that meet the project QC criteria, even when the client pressures for speed.