Comparison of sorption models to predict analyte loss during sample filtration and evaluation of the impact of filtration on data quality

Rahul Kumar, Sangeet Adhikari, Rolf U. Halden

Research output: Contribution to journalReview articlepeer-review

Abstract

Although filtration has been a widely applied sample pretreatment step in environmental analytical chemistry, its impact on the quality of the data produced is often underappreciated in the scientific community. The objective of this literature review and modeling exercise was to examine nine existing sorption models with input parameters including hydrophobic interactions, pH, and structural features to predict the loss of analytes during wastewater filtration due to sorption to suspended solids and to assess the impact of filtration on data quality. Models' sorption estimates were further validated with a set of comprehensive metadata collected and analyzed from 20 peer-reviewed research papers that reported physical measurements of the suspended solids sorbed fraction of analytes obtained during wastewater filtration of contaminants of emerging concern (CECs). Data on the impact of filtration were obtained from the literature for 156 organic compounds reported both for the dissolved and particulate bound analyte mass. Approximately 40% of CECs (62/156) showed significant filtration loss (>20%) as a result of the removal of suspended solids during filtration. The loss of analyte mass due to filtration ranged from <1% for atenolol to >95% for acenaphthene. Collected literature data were then used to evaluate the utility of sorption modeling to predict analyte losses during sample pretreatment. Among nine sorption models, three were found to predict filtration loss of at least 70% of the CECs evaluated within a range of ±20% of the actually measured filtration loss of analytes, assuming a suspended solid concentration of 200 mg/L and a fraction of organic carbon in suspended solids of 0.43. Thus, sorption modeling can help reduce error when calculating mass loadings based on samples filtered before analysis. It is concluded that the estimates could be further improved by considering the following factors: ionic interactions, characteristics of the water-borne sorbents, and filtration media properties.

Original languageEnglish (US)
Article number152624
JournalScience of the Total Environment
Volume817
DOIs
StatePublished - Apr 15 2022

Keywords

  • Fate
  • Hydrophobic interaction
  • PPCPs
  • Partitioning
  • QSPR
  • Sorption

ASJC Scopus subject areas

  • Environmental Engineering
  • Environmental Chemistry
  • Waste Management and Disposal
  • Pollution

Fingerprint

Dive into the research topics of 'Comparison of sorption models to predict analyte loss during sample filtration and evaluation of the impact of filtration on data quality'. Together they form a unique fingerprint.

Cite this