Identification of the cause of the difference among TOC quantitative methods according to the water sample characteristics

Abstract

Total organic carbon (TOC) analysis with accurate determination of particulate organic carbon (POC) content in suspended solids (SS) containing water is critical for evaluating the environmental impact of particulate organic pollutants in water and calculating the carbon cycle mass balance. TOC analysis is divided into the non-purgeable organic carbon (NPOC) and differential (known as TC-TIC) methods; although the selection of method is greatly affected by the sample matrix characteristics of SS, no studies have investigated this. This study quantitatively evaluates the effect of SS containing inorganic carbon (IC) and purgeable organic carbon (PuOC), as well as that of sample pretreatment, on the accuracy and precision of TOC measurement in both methods for various environmental water sample types (12 wastewater influents and effluents and 12 types of stream water). For influent and stream water with high SS, the TC-TIC method expressed 110–200 % higher TOC recovery than that for the NPOC method due to POC component losses in SS owing to its conversion into PuOC during sample pretreatment (using ultrasonic) and subsequent loss in the NPOC purging process. Correlation analysis confirmed that particulated organic matter (POM, mg/L) content in SS directly affected this difference (r > 0.74, p < 0.01, n = 24); for POC water samples (those containing >10 mg/L of POM) featuring purgeable dissolved organic matter, TC-TIC was appropriate in securing TOC measurement accuracy. In constrast, in effluent and stream water with low SS (i.e., < ∼5 mg/L) and high IC (> 70 %) contents, the TOC measurement ratios (TC-TIC/NPOC) of both methods were similar, between 0.96 and 1.08, suggesting that NPOC is appropriate for improving precision. Our results provide useful basic data to establish the most reliable TOC analysis method considering SS contents and its properties along with the matrix characteristics of the sample.