On the combination of USEtox® and SimpleBox 4 Nano models for the derivatization of size-dependent characterization factors for engineered nanomaterials

. Even if it has been claimed that Life Cycle Assessment is an essential tool to analyze, evaluate, understand and manage the environmental and health impacts of nanotechnology, few studies incorporate characterization factors (CFs) for human toxicity and freshwater ecotoxicity accounting for the impacts of engineered nanomaterials (ENMs) beyond their manufacturing stage. The objective of the present work consisted in identifying the correspondence between the information required and outputs provided by the USEtox® consensus model (which is not nanospecific) and the SimpleBox4Nano model (which accounts for nanospecific processes, e.g. aggregation, attachment and dissolution for Fate Factor derivatization) in order to assess the possibility of integrating the two to derive size-dependent CFs for the varying sizes of ENMs throughout their life cycle. The possibility to combine and integrate the two models appeared to be limited since there is no absolute correspondence between the two of them.


Introduction
Life cycle assessment (LCA) and its corresponding ISO framework [1,2] are recognized as suitable tools to identify the potential environmental and human health impacts of nanoenabled applications (NEAs) or nano-enabled products (NEPs) along their complete life cycles covering manufacturing, use and end-of-life stages [3]. As such, a number of review articles have been published in the past decade that cover the application of LCA to nanotechnology such as the recent work by Salieri et al. [4].
The LCA methodology comprises four iterative steps: (i) goal and scope definition, (ii) inventory analysis, (iii) impact assessment, and (iv) interpretation. Life cycle impact assessment (LCIA) converts emissions into environmental damages through linked fateexposure-effect models that require robust experimental data and a mechanistic understanding for each of these components. LCIA methods describe environmental impacts in terms of characterization factors (CFs). A CF is a substance-specific quantitative representation of the (relative) impact of a substance in the environment. CFs are based on models of cause-effect chains for a specific impact category.
USEtox® [5,6] is a fate-effect model specifically made for LCA-applications as it calculates human and eco-toxicity CFs based on the properties of a substance and the environmental compartment of initial release. The model estimates CFs by multiplying three other aggregated parameters related to fate (fate factor, FF), exposure (exposure factor, XF), and toxicity (effect factor EF), respectively, of a specific chemical. USEtox® operates on four different spatial scales: indoor, urban, continental and global. The indoor and urban scales only have an air compartment, whereas the continental and global scales consist of five compartments: air, agricultural soil, natural soil, freshwater and sea water. USEtox® cannot be directly applied to the LCIAs of NEPs and NEAs since the fate modelling is not nanospecific. SimpleBox4Nano (SB4N) [7,8,9] is a fate model able to model the fate of engineered nanomaterials (ENMs) depending on their size. SB4N has three main compartments: regional, continental and global, but the inner nested compartment, regional, has not only air as a medium but also freshwater (including lake, lake sediment, freshwater and freshwater sediment), seawater (including surface sea, deep sea and marine sediment) and natural, agricultural soil and urban/industrial soil. From all these media transfers to the other compartments and media are possible. Furthermore, the global compartment is split into three sub-compartments: moderate, arctic and tropical. SB4N models the fate of (i) freely dispersed (pristine) nanoparticle, (ii) nanoparticle hetero-aggregated with natural colloid particles (<450 nm) and (iii) nanoparticle attached to larger natural particles (>450 nm) [9].
Ettrup et al. [10] adapted the USEtox® 2.0 consensus model to integrate the SB4N for the development of CF of TiO2 nanoparticles to be incorporated in future LCA studies. Also focusing on TiO2 nanoparticles, Salieri et al. [11] presented an integrative approach for USEtox® 2.0 model with SB4N and proposed CFs for the freshwater toxicity impact category. More recently, Temizel-Sekeryan and Hicks [12] have calculated freshwater ecotoxicity CFs for silver nanoparticles by combining the principles of colloidal science with the USEtox® model using data from published mesocosm conditions.
The objective of the present work was to compare the information requirements and output by the two models in order to identify possible limitations in their integration for the derivation of size-dependent CFs for the varying sizes of ENMs released throughout the life cycle of NEPs and NEAs.

Materials and Methods
SimpleBox4.0-Nano and USEtox® 2.12 versions have been used. The comparison of the two models comprised two steps: (i) definition of the USEtox® air, water and soil scenarios in which the main parameters for the Regional and Continental compartments in SB4N have been set to fit those of USEtox®' Urban and Continental compartments, respectively, and (ii) identification of rate constants that are common for the two models.

Common constants for the two models
In this section, original USEtox® values were selected to be inserted in the USEtox® Air, Water & Soil scenario(s) defined in SB4N (SB4N -Scenarios sheet. Landscape settings). As shown in Table 1, only the Area land rate constant from the Urban compartment in USEtox® (USEtox®-DEF values) needs to be fed into the rate constants of the Regional scale of SB4N. As indicated in Table 2, Area land, Area sea, Fraction fresh water, Fraction natural soil, Fraction agricultural soil and Depth fresh water constants' s values in the Continental scale of SB4N need to be adjusted to fit with USEtox® (USEtox®-DEF values).  In absence of an Urban compartment and as an alternative to k.aC.aU, SB4N uses k.aC.aR (TRANSFER rate continental air -regional air); Excel cells corresponding to AU9, AV10, AW11 and AX12 (Engine sheet, SB4N).  In absence of an Urban compartment and as an alternative to k.aU.aC, SB4N uses k.aR.aC (TRANSFER rate regional air -continental air); Excel cells corresponding to D52, E53, F54 and G55 (Engine sheet, SB4N).

Conclusions
In order to derivate FF required for the calculation of CFs, the possibility of combining two different existing models has been evaluated. USEtox® is widely proposed for the LCIA phase (though still it is mostly proposed with limitations, e.g. for the effect factor) whereas SB4N is proposed for fate calculations for ENMs, especially as SB4N allows size-dependent calculations. Our approach to integrate the two models consisted of the identification and assimilation of common mass balance rate constants and aligning common constants that define the environmental compartments. However, this possibility has revealed to be limited since there is no absolute correspondence between the two models. Different approaches to integrate USEtox® and SB4N should be developed and tested in future studies.