Normal boundary intersection applied as multivariate and multiobjective optimization in the treatment of amoxicillin synthetic solution

• Published in: Environmental monitoring and assessment Vol. 190; no. 3; pp. 140 - 12
• Main Authors: Moura, Deberton, Barcelos, Vithor, Samanamud, Gisella Rossana Lamas, França, Alexandre Boscaro, Lofrano, Renata, Loures, Carla Cristina Almeida, Naves, Luzia Lima Rezende, Amaral, Mateus Souza, Naves, Fabiano Luiz
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01-03-2018; Springer Nature B.V
• Subjects: Amoxicillin; Antibiotics; Atmospheric Protection/Air Quality Control/Air Pollution; Bacterial diseases; Bioaccumulation; Biodegradation; Bubbling; Carbon dioxide; Design of experiments; Design optimization; Drugs; Earth and Environmental Science; Ecology; Ecotoxicology; Environment; Environmental Management; Environmental monitoring; Monitoring/Environmental Analysis; Multiple objective analysis; Multivariate analysis; Organic matter; Oxidation; Ozone; Penicillin; Water pollution; Waterways; Drugs; Ozonation; Pareto frontier; Advanced oxidation process
• ISSN: 0167-6369; 1573-2959
• DOI: 10.1007/s10661-018-6523-8

Amoxicillin is a useful antibiotic to combat bacterial infections. However, this drug can cause serious problems when discarded in waterways due to its great bioaccumulation potential. This compound can be treated via advanced oxidation processes (AOPs), which are capable of converting amoxicillin into carbon dioxide and water. In this context, the use of ozone as an oxidizer has excelled in amoxicillin degradation. This paper aims at treating a synthetic solution of amoxicillin (0.1 g L
−1
) in a reactor with ozone bubbling. A Design of Experiment (DoE) with a response surface known as Central Composite Design (CCD) was used to optimize the treatment process. In addition, a Normal Boundary Intersection (NBI) method was used in the construction of a Pareto boundary chart. Results after 1-h treatment showed a reduction of 53% of the initial organic matter from a designed model using factors, such as pH, ozone generator power, and O
3
flow. A model was built from the CCD with score of 0.9929. Thus, the model was able to represent the real scenario with confidence.