Perspective. Method validation in the Ontario Geoscience Laboratories

Abstract
In 1992, the Ontario Geoscience Laboratories (hereafter referred to as the Geoscience Laboratories) moved from Toronto to Sudbury into a new 25 000 ft2 facility. The acquisition of an extensive bank of new equipment and many staffing changes meant all methods were revised and then validated. Thus 7 single element tests, 2 precious metal packages, 2 major element packages by XRF, 7 elements determined by AA, 11 elements by ICP-OES, 8 elements by XRF and 24 elements by ICP-MS were validated. These methods represent the program needs of much of the Ontario Geological Survey. The task had to be completed ‘as soon as possible’ to minimize the impact on the mapping programs of the Survey which were not interrupted by the move. The tasks of building implementation, method development and validation were undertaken by approximately 12 staff members and were completed in 12 to 18 months. Method validation ensures that a particular analytical methodology using a specific combination of techniques will yield an analytical result that is proven to be correct. It is critical to perform such exercises when procedures are instituted or changed, new equipment procured or staff hired to ensure continuity of data quality. A generic validation program was documented, streamlined and used by all. A detailed outline was drawn up that explained the exercise and listed the steps required. For each method and element, precision, trueness, detection limit, quantification limit, working range, blank level and lack of inherent bias were quantified. These parameters were based on data from a minimum of seven replicates of five to seven SRMs that spanned a wide geological range. SRMs used in calibration or method development could not be included. The data quality objectives were approximately ± 5% precision (2σ) and trueness to meet the stringent data quality objectives of the Geoscience Laboratories. Here, over a 5-year period, precision and trueness are not expected to exceed ± 10%(2σ) when the concentration of an element is more than three times the limit of quantification or 10 times the limit of detection (LOD = 3so). Overall, the exercise was a resounding success. Data from the validation of Ba is used throughout the text to illustrate the principles discussed. The problems that were encountered typically included the availability of certain rock compositions, the level of characterization of some reference materials, the uncertainty on ‘known’ values and thus method trueness, the definition of detection limit and the difficulty in getting realistic ‘blank’ determinations for some methods. For the Geoscience Laboratories, the method validation exercise provides a standardized, non-subjective means to ensure the data quality objectives of the facility are met. It also convinces management, staff and clients that an appropriate level of expertise is achieved.