Rather than relying on templates, processing complex metrology data can be automated to reduce both the time required and the potential for transcription and processing errors. The Environmental Metrology Measurement Assistant (EMMA) accepts data from a variety of software platforms in order to allow single pass-through processing of several sample batches in a guided manner, increasing efficiency and ensuring standardized processing methods.
Quantification of environmental metrology data can be complex at the level on which NIST operates. Projects often include dozens of samples from multiple matrices and can target well over 300 analytes of vastly different chemical properties, each of which must be individually assessed. Quality assurance guidelines necessitate rigorous calibration and value checking protocols to ensure that data produced are of the highest quality in the world.
Most of the processing of data generated at NIST by gas- or liquid-chromatographic techniques happens by exporting instrument software reports into an external spreadsheet program. The Environmental Metrology Measurement Assistant (EMMA) eases this process considerably compared with more traditional templates by taking an instrumentation-software-agnostic approach. It pulls instrument-generated reports and parses them into an internally consistent format, allowing simultaneous processing of all data for a given project irrelevant of the technique used to generate it. Analysts are then guided through the processing in a stepwise manner from import to the calculation of final values. Multiple check points are built into the framework to trigger potential issues from instrumentation or sample-specific oddities to data entry errors. Since all data sources are combined into a single project workbook, this also serves to collect and store NIST data in a structured and consistent manner. Reviewing the way data were processed for any given project is much simpler, with targeted information available in a familiar presentation regardless of project.
Cutting and pasting combined with overly complex spreadsheet layouts give rise to operator-induced errors are the single largest source of human error in any data processing task. Broken or misaligned formulae can cause hours of quality assurance checking to correct a single value. By eliminating much of the rearrangement and manual entry requirements, EMMA's import routines alone save hours of data processing time. As EMMA constrains the data processing workflow in a NIST approved manner, normally unavoidable human errors are minimized, the final data product presentation is standardized, report generation is made simpler, and the confidence in the data product is greatly enhanced without time consuming quality checking of each detail. Automated quality control metrics and constraints are used at each step to track progress and make certain the data are within acceptable bounds and the quantification process adheres to NIST practices.
EMMA operates as a standalone workbook built with Visual Basic for Applications and runs in the Microsoft Excel 2010 (and later) environment. Other environments are being tested and will be incorporated where appropriate and useful.
EMMA currently accepts reports from the ChemStation, MassHunter, and Analyst software platforms. Other platforms will be added to the import routines as the use case expands.
EMMA currently focuses on the analysis paradigm of environmental organic metrology at NIST (i.e. quantification by area- and mass-ratio of target analytes with mass-labelled internal standards added prior to extraction). Other techniques will be considered for future inclusion including standard addition, external calibration, and others as needed.
Calibration in EMMA allows for including or excluding any given calibration point, provides a graphical overlay of where samples fall on the curve, and provides real time feedback regarding calibration parameters under a variety of calibration types. Users may select linear or quadratic regression, to force through the origin or not, which internal standard to use for quantification, and does not save changes to calibration parameters until directed by the analyst. This ensures a directed, eyes-on approach to calibration parameter review in addition to providing basic quality metrics for the entire project at a glance.
Calculation of the limit of detection can be performed in three manners: the lowest calibration point, mass present in blanks, or by minimum detectable peak area (doi: 10.1021/ac501615n). The largest value of any combination of these three techniques may be used.