Quality Assurance / Quality Control (QA/QC) Systems – Guiding Principles
As part of our ongoing commitment to producing quality results, Lambton Scientific follows a systematic approach to batch analytics. Analytical runs follow a pattern as can be seen the figure below.
- Initial Calibration
- Calibration Standards can consist of the following:
- Prepared in-house:
- Documented in "Reagent Preparation Log Book"
- "Critical Components" must have a supplier Certificate of Analysis
- All other ingredients must be ACS grade or better
- Any gravimetric weighing to be at least 2 significant digits (by difference), 2 or 4 preferred
- Prepared using DI waster (18 Mohms-cm)
- Use Trace Metals grade acids (if required)
- Use virgin and/or pre-rinsed storage containers
- Purchase Standards:
- Must be traceable to a Standard Reference Material (NIST preferred)
- Documented Certificates of Analysis
- Prepared in-house:
- Calibration Standards are:
- Typically analyzed from lowest to highest concentration
- Have the lowest concentration standard to be 5 - 10 times the MDL
- Have the highest concentration less than the maximum linearity value
- Have concentrations with no more than one order of magnitude between concentrations (i.e. 1, 10, 100, 1000, etc)
- For nonlinear curves (i.e. quadratics) - minimum of five standards (plus blank)
- For linear curves:
- Minimum of three standards (plus blank)
- Mininum acceptable correlation coefficient value is 0.995
- Treatment of intercepts:
- For positive intercepts:
- Leave as a positive intercept, however, the calculated concentration must not exceed 2 x MDL
- For negative intercepts:
- Force the regression thru origin
- Force the regression thru origin
- For positive intercepts:
- Calibration Standards can consist of the following:
- Batch 1 of Analytical Run
- Length of batch is dictated by the Reference Method (typically between 10 and 20 samples)
- QA/QC of Batch 1:
- Instrument Blank (IB)
- External Reference (QCS - Quality Control Start of Run)
- Calibration Standards and QA/QC samples must be completely independent of each other
- Preferably, use standards/reagents from different suppliers
- Alternately, use standards/reagents from a single supplier, but different lot numbers
- Calibration Standards and QA/QC samples must be completely independent of each other
- Method Blank (MB)
- Spiked Blank (SB)
- Sample 1 (S1)
- Sample 1 Replicate (S1 REP)
- must be completely independent from Sample 1 and follow all analytical steps
- Sample 1 Matrix Spike (S1 SPK)
- A known concentration of the target analyte added to sample
- Sample 1 Matrix Spike Duplicate (S1 SPK DUP)
- As above and must be completely independent from Sample 1 Matrix Spike
- Remaining of Batch 1
- Samples 2 - 10 (S2..S10)
- Continuing Calibration Verification (CCV)
-
A periodic confirmation, by analysis of a certified external standard (near midpoint of range) that the instrument performance has not change significantly (typically within 10%) from the initial calibration
- Batch 2 of Run
- QA/QC of Batch 2
- As above
- Remaining of Batch 2
- As above
- External Reference (QCE - Quality Control End of Run)
Lambton Scientific's Typical Rules of Quality Control:
- Maximum intercept (b) is 2 x MDL
- Minimum correlation coefficient (r²) is 0.995
- External Reference Target Limit is 10% (90-110% of Target) - see QA/QC Charting
- External Reference Relative Percent Difference (RPD) is 10%
- Sample Replicates must be within 10% of average
- Matrix Spike must be within 20% of original sample value
- Calculated concentrations must not be more than 10% extrapolated beyond the top of the calibration curve