TY - JOUR
T1 - Electrochemical fingerprints of brominated trihaloacetic acids (HAA3) mixtures in water
AU - Cetó, Xavier
AU - Saint, Christopher
AU - Chow, Christopher W K
AU - Voelcker, Nicolas H
AU - Prieto-Simón, Beatriz
PY - 2017
Y1 - 2017
N2 - In this work, we explore the capabilities of combining electrochemical sensors and chemometrics towards the analysis of haloacetic acids (HAAs) in water samples. Our approach is based on electronic tongue principles. It combines voltammetric measurements on a gold electrode with chemometric data processing, to extract characteristic fingerprints for HAAs. Cyclic and square wave voltammograms were pre-processed by means of fast Fourier transform (FFT) to provide the coefficients used as subsequent inputs for an artificial neural network (ANN) model. We were able to quantitatively detect and discriminate each HAA under study. Quantitation of HAA3 mixtures (i.e. bromodichloroacetic acid, dibromochloroacetic acid and tribromoacetic acid) was achieved at the μg/L level, with a normalized root mean square error (NRMSE) of 0.054 for the validation subset. Finally, successful analysis of spiked water samples was achieved demonstrating the operation of the sensor in the absence of matrix effects.
AB - In this work, we explore the capabilities of combining electrochemical sensors and chemometrics towards the analysis of haloacetic acids (HAAs) in water samples. Our approach is based on electronic tongue principles. It combines voltammetric measurements on a gold electrode with chemometric data processing, to extract characteristic fingerprints for HAAs. Cyclic and square wave voltammograms were pre-processed by means of fast Fourier transform (FFT) to provide the coefficients used as subsequent inputs for an artificial neural network (ANN) model. We were able to quantitatively detect and discriminate each HAA under study. Quantitation of HAA3 mixtures (i.e. bromodichloroacetic acid, dibromochloroacetic acid and tribromoacetic acid) was achieved at the μg/L level, with a normalized root mean square error (NRMSE) of 0.054 for the validation subset. Finally, successful analysis of spiked water samples was achieved demonstrating the operation of the sensor in the absence of matrix effects.
KW - Artificial neural networks
KW - Disinfection by-products
KW - Electronic tongue
KW - Haloacetic acids
KW - Principal component analysis
UR - http://www.scopus.com/inward/record.url?scp=85014786360&partnerID=8YFLogxK
U2 - 10.1016/j.snb.2017.02.179
DO - 10.1016/j.snb.2017.02.179
M3 - Article
AN - SCOPUS:85014786360
VL - 247
SP - 70
EP - 77
JO - Sensors and Actuators B: Chemical
JF - Sensors and Actuators B: Chemical
SN - 0925-4005
ER -