Processes in the environment are often complex and the description and prediction of the behaviour of environmental pollutants requires sophisticated methodology including multidimensional technology. Multivariate statistical methods including principal component analysis (PCA) and partial least squares projection to latent structures (PLS) were used to characterise several nitroaromatic compounds according to a large number of chemical and physical properties. These descriptors were derived from databases and semi-empirical, empirical and quantum-chemical calculations. A model was calculated and used for prediction of experimentally determined relevant environmental parameters, including volatility through gas chromatographic retention times and the distribution between air and water phase using solid-phase microextraction measurements. Analytical microextraction techniques are very well suited for both monitoring and measuring relevant environmental properties. Two such methods, solid-phase microextraction (SPME) and microextraction by packed syringe (MEPS), which are both faster and less resource demanding than traditional methodology, were used to improve the analysis of nitroaromatic compound and haloanisoles. Both techniques showed few interfering compounds and good reproducibility in combination with GC-MS detection. The reproducibility for SPME and MEPS was further enhanced by the usage of isotopically labelled internal standards (nitroaromatic compounds) or internal standards with similar chemical properties as the analytes (haloanisoles).

The SPME method for nitroaromatic compounds was tested using industrial wastewater from a wastewater treatment facility receiving input from several nitrogen based industries. Using gas chromatography-negative chemical ionisation-mass spectrometry(GC-NCI-MS) as the final detection, limits of detection (LODs) in the range 0.02–38 ng L−1 were achieved. Results for the 2,4,6-TCA (TCA) and 2,4,6-TBA (TBA) using high resolution GC-MS were even more impressive obtaining LODs in the range 0.02-0.03 ng L-1, which are currently the lowest reported in the international literature. Although the newly developed MEPS technology is generally faster than SPME, similar LODs as for SPME could not yet be achieved. Both methods were found to be so sensitive that TCA and TBA could be detected in wine before it could be sensorically noticed as cork tainted.