Validation of a method for the analysis of volatile organic compounds in water

Dissertação de mestrado, Qualidade em Análises, Faculdade de Ciências e Tecnologia, Universidade do Algarve, 2015

Many of the volatile organic compounds (VOCs) which can be harmful to humans have their origin in petroleum products. The VOCs have been found in water sources and leakage from storage tank and accidental spills have been regarded as the main causes of contamination from VOCs. The main objective of this study was to validate detection method of some 15 VOCs by solid-phase microextraction – gas chromatography – mass spectrometry. SPME-GC-MS has been a widely accepted method for analysis of VOCs. The compounds analyzed in this study are; MTBE, 3-ethyltoluene, 4-ethyltoluene, 2-ethyltoluene, 1,2,4-trimethylbenzene, 4-isopropyltoluene, 1,3-diethylbenzene, Indane, 1,4-diethylbenzne, 1,3-dimethyl-5-ethylbenzene, 1,2-diethylbenzene, 1,4-dimethyl-2-ethylbenzenene, 1,3-dimethyl-4-ethylbenzene, 1,2-dimethyl-4-ethylbenzene, 1,2-dimethyl-3-ethylbenzene and hexachlorobutadiene. After separation by the gas chromatograph the compounds were detected in full scan mode and later further studies were carried out in selected ion monitoring (SIM) mode of mass spectrometer. Method validation parameters for the detection of these compounds included selectivity, linear working range, limit of detection (LOD), limit of quantification (LOQ) precision, accuracy and measurement uncertainty. Various statistical tools like regression analysis, residual analysis, Mandel’s test of linearity, RIKILT, and normalized area test were applied to derive and ascertain the results and arrive at a conclusion. The retention time and representative mass fragments were identified for each compound. A linear curve (regression analysis) in the working range was also identified for each of these compounds after suitable dilution of the pure compounds. Working range was between less than 0.1 μg/L and 0.5 μg/L (the minimum and maximum calibration standards) for all the compounds except for MTBE and indane. Linearity was confirmed by residual analysis and Mandel’s test for linearity. Two of the compounds 1,4-diethylbenzene and 1,3-dimethyl-5-ethylbenzene coelute and appear as a single peak in the chromatogram and therefore, their quantity is expressed as the combined quantity of the two. LODs are well above the baseline and LOQs are either equal to or lower than the lowest calibration standards. LOD and LOQ were also quantified from precision data. Precision was studied by determining repeatability and intermediate precision and was expressed as relative standard deviation (RSD %). Council Directive 98/83/EC has prescribed a limit value of 25% for precision. None of the values of repeatability and intermediate precision exceeded the limit of 25%. Accuracy was determined by recovery study of three types of spiked water matrices; tap water, river water and groundwater. Recovery was expressed by comparing the spiked results with the theoretical value (a value provided by the commercial supplier) of a compound in terms of percentage of recovery. Also 10 replicate analysis of the spiked sample gave its precision. Most of the recovery results have been found between 90 and 115%. All the recovery values meet the criterion of 25% recovery set by the Council Directive 98/83/EC. ISO 17025 requires that the laboratories express the results accompanied by the estimated uncertainty. Expanded uncertainty of the method was determined for each compound by combining the component uncertainty of precision, calibration standards and regression interpolation and then multiplying the combined uncertainty by a coverage factor of two for a 95% confidence level. Uncertainty values ranged from 8.2% to 23%. It has been found that for the same compound the uncertainty values for the three different matrices are similar. VOCs targeted in this study can be used as possible indicators of petroleum product contamination of water sources. Each compound has its own retention time and mass spectra which can be used for its detection. Linearity of the working range has been confirmed by various statistical tests. LOD, LOQ and recovery results meet European regulation requirement and this indicates validity of the method and can be applied to detect the compounds in water. HS-SPME sampling is solvent free and less time consuming and therefore is preferable. There has been only limited research in method validation of many target VOCs. So this study contributes to methods of analysis used to detect the target VOCs in different water matrices.