Changes to TNI Standard Module 6 Quality Systems

Changes to TNI Standard Module 6 Quality Systems for Radiochemical Testing Presented at the 2015 ASP Workshop Charleston, South Carolina 17 September 2015

Radiochemistry Expert Committee Members David Fauth C. Martin Johnson, Jr. Sreenivas Komanduri Nile Ludtke Keith Mc. Croan Larry Penfold Bob Shannon Tom Semkow Richard Sheibley Carolyn Wong Associate Members Ariana Mankerian Joe Pardue Bill Ray Ron Houck Bryan Miller Terry Romanko Carl Kircher Tom Patten Jim Chambers Virgene Mulligan Reed Jeffery Yoon Cha TNI Staff Ilona Taunton

Background ¨ ¨ Environmental testing for inorganic and organic, and for radiochemical parameters, have evolved separately Protocols and concepts used for “stable” chemistry may not always be applicable to radiochemistry, although they often have been used as basis for quality requirements

Regulatory Background ¨ Safe Drinking Water Act (to a lesser extent also Clean Water Act) Ø Narrow, specific set of tests Promulgated methods largely based on 1950 s to 1970 s EPA procedures ª Expanded to include DOE, USGS, ASTM and SM ª Instruments (and even sometimes reagents) not always available ª Quality requirements sparse (to put it politely…) ª Little or no reliable performance/validation data ª

What are Key Differences? ¨ Lab-developed methods are the rule Ø Measure radioisotopic composition Screening tests: gross alpha/beta, ª Gamma emitters, 3 H, etc. ª Ø ¨ Chemical separations use yield tracers to minimize bias Measurements relative to background Ø Ø Results are not censored (e. g. , against MDL, RL) rather positive, negative or zero results are reported “as measured” Measurement uncertainty estimated and reported with each result

MARLAP Ø Cross-agency document developed and approved by eight federal agencies: ª EPA, DOD, DOE, DHS, NRC, FDA, USGS, NIST u u Part I for project planners / managers of radioanalytical projects Parts II and III address technical topics

TNI Radiochemistry Expert Committee ¨ Formed in 2012 Ø Consists of ten radiochemists Lab managers, QA officers, and regulators ª Experience in state, federal, and commercial environmental programs ª Ø Reviewed and updated TNI Standard Module 6: Working Draft Standard posted for comment ~1 year ago ª “Modified WDS” posted for comment in December 2014. ª Voting Draft Standard (VDS) balloted in April 2015 ª Interim Standard (IS) posted in August ª IS will become TNI Standard late in September 2015 ª

Terms and Definitions Added definitions specific to Module 6 to add clarity For example: critical value: value to which a measurement result is compared to make a detection decision (also known as critical level or decision level) Note: The critical value is designed to give a specified low probability, α, of false detection in an analyte-free sample, which implies that a result that exceeds the critical value gives high confidence (1 − α) that the radionuclide is actually present in the material analyzed. For radiometric methods, α is often set at 0. 05. detection limit (DL) for Safe Drinking Water Act (SDWA) compliance: Laboratories that analyze drinking-water compliance samples for SDWA must use methods that provide sufficient detection capability to meet the detection limit (DL) requirements established in 40 CFR 141. The SDWA DL for radioactivity is defined in 40 CFR Part 141. 25(c) as the radionuclide concentration which can be counted with a precision of plus or minus 100% at the 95% confidence level (1. 96σ where σ is the standard deviation of the net counting rate of the sample).

Uncertainty is Clearly Defined Measurement Uncertainty: parameter associated with the result of a measurement that characterizes the dispersion of the values that could reasonably be attributed to the measurand. Standard Uncertainty: an estimate of the measurement uncertainty expressed as a standard deviation (c. f. , Expanded Uncertainty). Expanded Uncertainty: the product of the standard uncertainty and a coverage factor, k, which is chosen to produce an interval about the result that has a high probability of containing the value of the measurand. (c. f. Standard Uncertainty) Counting Uncertainty: the component of measurement uncertainty attributable to the random nature of radioactive decay and radiation counting (often estimated as the square root of observed counts) (after MARLAP). Older references sometimes refer to this parameter as Counting Error or Count Error. (c. f. , Total Uncertainty). Total Uncertainty: an estimate of the measurement uncertainty that accounts for contributions from all significant sources of uncertainty associated with the analytical preparation and measurement of a sample. Such estimates are also commonly referred to as Combined Standard Uncertainty or Total Propagated Uncertainty, and in some older references as the Total Propagated Error, among other similar terms. (c. f. , Counting Uncertainty).

New Concept !!! Radiation Measurements Batch (RMB): added to address QC needs for tests that are non-destructive in nature. An RMB is composed of one (1) to twenty (20) environmental samples that are counted directly without preliminary physical or chemical processing that affects the outcome of the test (e. g. , non-destructive gamma spectrometry, alpha/beta counting of air filters, or swipes on gas proportional detectors). The samples in an RMB share similar physical and chemical parameters, and analytical configurations (e. g. , analytes, geometry, calibration, and background corrections) and the maximum time between the start of processing of the first and last samples in an RMB is fourteen (14) calendar days. Preparation batch: Concept unchanged - applicable to tests that require physical or chemical preparation that affects the outcome of the test. Analytical Batch: Concept unchanged (and rarely used)

Exclusions and Exceptions ¨ ¨ Module 6 is applicable to measurements used to monitor radioactivity or determine compliance with regulations pertaining to radioactivity Non-radiometric measurements of radionuclide parameters were previously excluded from Module 6 (e. g. , KPA) and thus effectively not covered anywhere Ø In the update, ª ª ª May refer to Chemistry Module 4 for technique-specific requirements or QA/QC not addressed in Module 6. The source of requirements used must be clearly documented in the lab’s Quality System / procedures For example, calibrations, calibration verifications, and detection statistic determinations, and method-specific quality control for (radio)isotopic determinations by ICP-MS

“Detection Capability” ¨ Detectable Activity renamed Detection Capability Ø Definitions provided in Section 1. 3 MDA, Critical Value, SDWA DL, etc. Requirements similar to current requirements. ª Must empirically verify detection capability prior to analyzing samples ª Precludes use of MDA/MDC for detection decisions ª Ø

Validation of Methods ¨ ¨ ¨ Method performance for all methods must be characterized. Performance data must be available that address key performance indicators, such as detection capability, precision, bias, and selectivity (consistent with published guidelines such as MARLAP, FEM, EUROCHEM) May be taken from published validation data, historical QC results, or method validation

Validation of Precision and Bias ¨ ¨ The range must include zero activity unless not applicable to use of results… Acceptance criteria for performance must be sufficient to meet intended use of data, Ø ¨ i. e. , based on DQOs / MQOs Must be published in the lab’s Quality System / SOPs

Verification of Measurement Uncertainty ¨ ¨ Compare results of precision evaluation to uncertainty to verify uncertainty estimate validity Uncertainty reported in association with all measurements Ø Ø Ø To comply with SDWA, other regulations, or contracts, labs may report the counting uncertainty; All other measurements report total/combined uncertainty consistent with the Guide to the Expression of Uncertainty in Measurement (GUM), MARLAP, or equivalent approaches; Report must specify the type of uncertainty reported (counting or total), and its significance, (e. g. , 95%, 1σ, or k=1)

Verification of Selectivity ¨ The laboratory must qualitatively evaluate selectivity by addressing sample and matrix characteristics: Ø Ø Ø ¨ ¨ Effect of matrix on the ability of method to detect analyte; Ability of the method to chemically separate the analyte from the interfering analytes; Spectral and instrumental interferences. May be accomplished by testing matrix blanks, spiked matrix blanks, worst-case samples, or reference materials. If applicable, a qualitative selectivity statement should be included in the SOP.

Demonstration of Capability (DOC) ¨ Analyze four samples and four blanks Ø Blanks added Results are not censored ª “Absolute bias” (i. e. , bias at zero activity) can compromise low-activity results ª Ø ¨ Ongoing QC measurements can be used! Otherwise, essentially unchanged

Technical Requirements Reorganized and clarified to address the “calibration life-cycle” ¥ 1) Set-up of instrumentation (1. 7. 1. 1) 2) Establish and perform instrument QC (1. 7. 1. 4) (eliminates misleading term “CCV”) 3) Initial calibration for method (1. 7. 1. 2) 4) Calibration verification - true verification of methodspecific calibrations (1. 7. 1. 3) ¥See ASTM D 7282 -Standard Practice for Set-up, Calibration, and Quality Control of Instruments Used for Radioactivity Measurements

Calibration by Mathematical Modeling ¨ The section on calibration Ø Ø Reiterates the need for physical calibration of instruments against traceable reference materials Opens the door for applying mathematical or statistical corrections based on techniques such as Monte Carlo simulations, etc.

Instrument Performance Checks ¨ Gamma-ray detectors Ø Semiconductor detectors (e. g. , HPGe) ª ª Ø Scintillation detectors (e. g. , Na. I) ª ¨ ¨ Semiconductor detectors added Liquid Scintillation Spectrometers Ø ¨ Day of use. Alpha/beta counters (previously gas-proportional counters) Ø ¨ Twice weekly for continuously operating detectors (c. f. , ANSI N 42. 15) Day of use for non-continuously operating detectors System normalization as recommended by manufacturer Alpha-particle Spectrometry Systems – unchanged Performance check frequency may be extended for long sample counts, runs on automated sample changers Ø Ø Run duration limited to 7 days for automated sample changers Requires bracketing counts on manual counters

Background Measurements ¨ Previous TNI Standard requirements confused contamination control and control of backgrounds Ø Ø Ø Provided little guidance to labs and auditors Effective oversight difficult Implementation at labs very inconsistent ª ª Elevated levels of low-activity bias and Type I and II decision errors on detections

Background Measurement, Checks, and Contamination Monitoring ¨ The update differentiates between: Ø Ø Ø ¨ ¨ Subtraction backgrounds; Short-term background checks Contamination controls Minimum frequency and functional requirements Recognizes different approaches in use for determining background to minimize prescription Ø Paired measurements, historical compositing of backgrounds, blank populations, etc.

Quality Control for Radiochemistry ¨ ¨ General statements applicable to all QCs are incorporated into Section 1. 7. 2. 1. Documented quality control program must incorporate requirements imposed by regulations, methods, and the TNI standard. Regulations and method requirements take precedence ª Minimum frequency and acceptance criteria for QCs must be documented in quality manual. ª

Radiation Measurements Batch (RMB) ¨ Current batch QC requirements: Ø Ø ¨ LCS, duplicate, MS, and blank For gamma spec, an MS was required but no blank! RMB applies to non-destructive measurements Ø “Prep” has no effect on the measurement ª Ø ¨ ¨ ¨ Direct counting (alpha/beta or gamma spectrometry of air filters) Must share common physical/chemical parameters and analytical configurations Batch size ≤ 20 samples May add samples to batch for up to 14 days from initiation One LCS (can reuse), blank, and duplicate per batch

Negative Control - Method Blank ¨ Laboratory must have procedures to determine when method blank is different from zero Ø ¨ ¨ e. g. , compare result to CSU Blanks must be evaluated for long term trends / bias No subtraction of batch method blank Ø Ø May correct for average historical activity of method blanks to address demonstrated bias, But must account for additional uncertainty resulting from the correction

Positive Control Laboratory Control Sample ¨ One per batch Ø Ø Minimum spike concentration based on the relative uncertainty of acceptance criteria For RMBs ª LCSs need not be prepared for each batch u u By definition there is no prep for RMBs May use calibration standard (although a second independent source is still required for verifying calibrations)

Positive Control Laboratory Control Sample ¨ Must include all radionuclides being determined except: Ø Ø Gross activity measurements may use an appropriate surrogate (e. g. 230 Th for gross α) Alpha spectrometry measurements ª Ø For multiple radionuclides with similar characteristics determined simultaneously, only one analyte/isotope needs to be included Gamma-ray spectrometry using energy/efficiency calibration curve ª May use a radionuclide similar energy – minimum of low-energy and high-energy gamma

Sample Specific QC Measures ¨ Matrix Spikes Ø One per preparation batch ª ª ª ¨ Duplicate Ø Ø ¨ Components consistent with those of the LCS Not required for non-destructive methods (e. g. , gamma spec) Not required for methods with tracers or carriers One per batch For Radiation Measurement Batches when multiple detectors used, a 2 nd count on a different detector is required Tracers/Carriers Ø Essentially unchanged

No Special Handling of QC Samples ¨ No systematic preference of detectors, equipment, or glassware for QC samples Ø ¨ e. g. , always counting blank in same detector, or dedicating equipment or glassware for QA samples Does not preclude dedicating glassware or instrumentation for different activity samples if QC is handled the same as samples

Reagent Quality, Water Quality, and Checks ¨ Reference standards may be obtained from Ø Ø Ø ¨ ¨ ¨ a national metrology institute (NMI), e. g. NIST. an ISO/IEC Guide 34 accredited reference material provider, or an ANSI N 42. 227 reference material manufacturer. Provides guidance for when there is no known provider of a traceable standard, Information as required by Section 8, Certificates, of ANSI N 42. 22 -1995, and The laboratory must verify standards before use

Data Evaluation and Reporting ¨ ¨ Evaluation of tracers and carriers discussed (missing from 2009 Standard) Discussion of QC acceptance criteria in Section 1. 7. 2. 1 – 1. 7. 2. 3 Ø ¨ May report qualified results if activity measured in samples is greater than 5 times activity found in blank Requires documentation of QC calculations

Reporting ¨ Results must be reported: Ø As measured; ª Uncensored at detection limit or reporting limit u Ø Include positive, negative and zero values With laboratory’s estimate of the measurement uncertainty; e. g. , 10. 2± 1. 4 p. Ci/L or 0. 03± 0. 26 p. Ci/L Ø ¨ And with activity reference date, Project- or client-specified requirements may supersede requirements of the standard.

Sample Handling ¨ The laboratory must: Ø Ø ¨ Verify that samples have been preserved as required (regulation, method, contract, or quality system). Document timing, methods used, acceptance range, or other conditions indicating acceptable preservation. If verification does not meet criteria, the lab must: Ø Ø Perform corrective actions (e. g. , notify client, preserve sample), and Qualify impacted test results in the report

Our Next Steps? ¨ ¨ ¨ Assessment checklist Training for assessors Training for laboratories

Thank you!!! Questions? For more information on TNI: http: //nelac-institute. org