Simultaneous analysis of nitro compounds by Voltammetric method combined with the partial least squares (PLS) and the principal component regression (PCR)
Abstract
Nitro aromatic compounds exhibit a strong electrochemical behavior on hanging mercury drop electrodes (HMDE). The ability to simultaneously determine 6 nitro compounds including nitrobenzene (NB), 2-nitrophenol (2-NP), 3- nitrophenol (3-NP), 4-nitrophenol (4-NP), 2,4- dinitrophenol (DNP) and trinitrophenol (TNT) in the same samples was investigated by differential pulse Voltammetry. The method is based on the electrochemical process at HMDE in acetate buffer of pH 4.6. It was found that peak potentials of voltammetric reduction waves of NB, 2-NP, 3-NP and 4-NP appeared at - 0.332, - 0.281, - 0.311 and - 0.387 V respectively. Under the same conditions there are three peak potentials at -0.131, - 0.227, - 0.332 V for TNT and two peaks at tại - 0.221, - 0.332V for DNT. The peaks of these compounds are very close together. Due to the serious overlapping of Voltammetric peaks of these compounds in the mixtures, so that by a conventional Voltammetry it not possible to individually determine each compound. In this study, linear multivariate regression methods such as partial least squares (PLS), principal component regression (PCR) was used to resolve the overlapped Voltammograms. The obtained relative standard errors of both methods are within ± 10%. The PLS (10.8% RSEt) results in data better than PCR (11.8% RSEt).
Keywords
Download Options
Introduction
The simultaneous determination methods of the nitro - compound mixtures in the environmental samples have received considerable attention, because they are widely used to in industries to control the pollution of individual pollutants present in waste water. The high level toxicity of nitro- compounds and their propagation through environment are capable of polluting land, water, air and affecting on human, animal health, fish, aquatic organism and other life forms. There is a variety of analytical methods applied to determine these compounds including performance liquid chromatography (HPLC), UV-vis spectrophotometry and electrochemistry. These methods are often complex and time consuming and requiring sample pretreatments involving separation, extraction before analyzing [3-7]. For the Voltammetric methods, due to the serious overlapping of their reductive peak potentials caused by the general structural formula, the nitro-compounds cannot be quantitatively determined individually. Therefore these methods are usually limited to analyzing a single chemical composition or determining the sum of the nitro compounds in the mixtures [1,2,5-7].
In this study, linear multivariable regression methods such as partial least squares (PLS), principal component regression (PCR), (applying mathematical, statistical, graphics methods, etc) were applied for experimental planning, optimization of obtained experimental data used to resolve the overlapped Voltammograms of nitro-compounds [4]. Based on the PLS and PCR models, each nitro compound such as NB, 2-NP, 3-NP, 4-NP, DNT and TNT was simultaneously determined from Voltammograms of their mixtures.