Statistical modelling approach to derive quantitative nanowastes classification index; estimation of nanomaterials exposure

Abstract

The widespread use of engineered nanomaterials (ENMs) in consumer products and industrial applications has raised concerns because of their potential impacts in the environmental systems. These concerns are due to dramatically increasing reports on the toxicity of ENMs to biological life forms in the ecosystems, and secondly, owing to rapid generation of new waste streams previously unknown and generically regarded as nanowastes. Currently, nanowastes classification remains poorly studied despite the large diversity of nanoproducts (e.g. cosmetics, textiles, etc.) being introduced in the global markets. Systematic classification of nanowastes is one of the useful tools to support efficient and effective nanowastes management by stakeholders, for example, governments, regulatory agencies, industries producing nanoproducts and ENMs, waste management industry, among others. In this paper, we present preliminary results on quantitative nanowastes classification through estimation of exposure parameter. Risk is a function of hazard and exposure; here results on the potential exposure of ENMs from nanowastes are presented. The estimation was done using statistical modelling to model the dynamics of aggregation as a surrogate parameter for exposure. In this work, statistical inference approach specifically the non-parametric bootstrapping and linear model were applied. Data used to develop the model were sourced from the literature. 104 data points with information on aggregation, natural organic matter (NOM), pH (including pHpze), ionic strength (IS), and size were analysed. Results from linear regression analysis of aggregation as a function of four variables (NOM, IS, pH, size) indicated IS as the single most significant variable (at 5% level). Notably, IS accounted for aggregation with a linear coefficient and standard error 4.466 and 1.543, respectively. In addition, 5000 boot-strap estimates were performed to obtain the sampling distribution for the linear regression coefficient of aggregation on IS. The boot-strapping results were consistent as they yielded a linear coefficient of 4.466 and standard error of 1.754. Thus, our results support the hypothesis that IS is an adequate abiotic parameter to account for the aggregation of ENMs in the aquatic systems.