Volatile Organic Compounds (VOC) at trace level are common indicators for climate change and ambient air quality. The calibration of detection systems at trace level needs VOC mixtures at nmol/mol level with 5% uncertainty, the dynamic preparation can fulfil it. A source of uncertainty for dynamic generation is the stability of the VOC molar fraction averaged on the sampling time. A general method for the calculation of the relevance of the stability contribution to the total uncertainty is here proposed. VOC molar fraction is here considered as a composed quantity. Threshold values for stability contribution were set to consider it either as negligible and relevant. The stability of the operative quantities of a device, for the dynamic preparation of accurate VOC reference mixtures, was measured at the working conditions by the Allan deviation algorithm. Stabilities of operative quantities were composed to calculate the VOC molar fraction stability and their contribution to total uncertainty of VOC molar fraction. The stability of the air inlet pressure gave the main stability contribution. The stability contribution to the total uncertainty resulted to be negligible for observation times below 60 min for VOC molar fraction uncertainty of 2% at 30 nmol/mol level

The growing need for global volatile organic compounds (VOC) data at trace level for industrial applications, domestic and working health, air quality and climate change is driving research towards the preparation of traceable and accurate standards for the VOC amount of substance fraction in gas phase.The World Meteorological Organization (WMO) has implemented a worldwide VOC monitoring programme, named Global Atmosphere Watch (GAW), which data quality objectives refer to an expanded uncertainty lower than 10% for VOC measurement at amount of substance fractions lower than 100 nmol∙mol−1.Currently, gaseous standard atmospheres are prepared in high pressure gas cylinders, but they do not fulfil GAW requirements of accuracy and stability at ppb level, especially for oxygenated VOCs. Dynamic methods are proposed as an alternative to provide a stable amount of substance fraction of the mixture over time. Diffusion rate, dilution and carrier air flow rates, purity of air, VOC losses due to wall-interactions and leakages were considered as influence quantities of the VOC amount of substance fraction dynamically prepared in gas phase by the diffusion method. The contribution of the VOC losses to standard mixtures uncertainty and their sampling procedure are under study and a solution has not yet been proposed.In this paper, a conservative quantification of VOC losses is proposed, discussed and applied to a case study: a single stage dynamic dilutor with an acetone generation. An uncertainty budget was prepared to calculate target uncertainty for VOC losses due to leakages and VOC–wall interactions on Teflon®, Pyrex® and stainless steel tubing. Teflon® showed negligible interaction effects when the pipe was clean and long term effects were not considered. Uncertainty was not negligible for glass and stainless steel for which more accurate analyses are necessary. A leakage test was developed to assure negligible leakage contribution to uncertainty

Volatile organic compounds (VOCs) in gas mixtures at trace level (nmol/mol) are routinely measured by chemical and biochemical laboratories as climate indicators, indoor air quality pollutants from building materials emissions, contaminants in food and beverages, and biomarkers in body fluids (blood, urine, breath) of occupational exposure or human diseases. Current analytical instruments used for measurements are gas chromatographs equipped with various injector and detector configurations. The assurance of measurement quality is done by using a huge amount of certified liquid VOC standard solutions (or gaseous VOC standard cylinders) with multiple dilutions to reach the required trace level. This causes high standard uncertainty in instrument calibrations, high cost, and high consumption of analysis and laboratory personal time. In this paper, we present the implementation of portable generators producing VOC gas standards at trace level for automatic and direct calibration of VOC detectors employed in various contexts, removing the need for preparation of matrix calibration standards in cylinders. Two compact devices in-house developed by two national metrology institutes-the Istituto Nazionale di Ricerca Metrologica (INRIM) and the Federal Institute of Metrology (METAS)-are here used to dynamically generate reference gas mixtures in an SI traceable way. The two devices are based on different technologies: diffusion and permeation, for INRIM and METAS, respectively. A metrological characterization is given and the practical implementation at chemical and biochemical laboratories is discussed. Graphical abstract Onsite calibration with transportable generation system with similar performances to primary laboratory devices.

The main features of the Calibration Curve Computing (CCC) Software, a tool for regression analysis developed at the INRIM, the Italian National Metrology Institute, are described in this paper. The first release of the software tool, version 1.3, has been available on the INRIM website for free download since 2015. It was developed under a Matlab environment and distributed as an executable file. The software tool featured a friendly graphical user interface, to encourage even non-Matlab users to use it. The software is able to perform regression fitting according to ordinary,weighted and weighted total least square models and it can be applied to datasets that can be characterized by uncertainties and covariances in both the dependent and the independent variable. The curves that can be fitted to the data are fractional polynomials with positive and negative exponents up to the fifth order. Recently, version 2.0 of the CCC Software has been developed and tested and will be uploaded early in 2020. The main improvements of this new version, concerning the graphical aspects and the calculation functionalities, are here described in detail, with a major focus on the enhancements obtained in the estimated accuracy of the weighted total least square regression. Validation results and the application to a real case study are also shown. The present paper is a significant expansion of a relevant abstract presented at the 'Mathematical and Statistical Methods for Metrology'