Science of the Total Environment 496 (2014) 122–131
Contents lists available at ScienceDirect
Science of the Total Environment
j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv
Combining life cycle assessment and qualitative risk assessment: Thecase study of alumina nanofluid production
Grazia Barberio a,⁎, Simona Scalbi a, Patrizia Buttol a, Paolo Masoni a, Serena Righi b
a ENEA—Italian National Agency for New Technologies, Energy and Sustainable Economic Development, Lungotevere Thaon di Revel, 76-00196 Rome, Italyb University of Bologna, C.I.R.S.A., Ravenna, Italy
H I G H L I G H T S
• RA and LCA are applied for assessing the sustainability of new technologies• A framework for combining RA and LCA and overcome their limits is proposed• A case study of alumina nanofluid production is presented• Two different pilot lines are analyzed: single-stage and two-stage• Results show that RA and LCA have a complementary role in the impact assessment
In this paper the authors propose a framework for combining life cycle assessment (LCA) and Risk Assessment(RA) to support the sustainability assessment of emerging technologies. This proposal includes four steps ofanalysis: technological system definition; data collection; risk evaluation and impacts quantification; resultsinterpretation. This scheme has been applied to a case study of nanofluid alumina production in two differentpilot lines, “single-stage” and “two-stage”. The study has been developed in the NanoHex project (enhancednano-fluid heat exchange). Goals of the study were analyzing the hotspots and highlighting possible trade-offbetween the results of LCA, which identifies the processes having the best environmental performance, andthe results of RA, which identifies the scenarios having the highest risk for workers. Indeed, due to lack of dataabout exposure limits, exposure–dose relationships and toxicity of alumina nanopowders (NPs) and nanofluids(NF), theworkplace exposure has been evaluated bymeans of qualitative risk assessment, using StoffenmanagerNano. Though having different aims, LCA and RA have a complementary role in the description of impacts ofproducts/substances/technologies. Their combined use can overcome limits of each of them and allows awider vision of the problems to better support the decision making process.
The environmental effects of a technology depend on the use doneby society, the interaction with technological systems, the physical con-text and the quantity of use (Mulder et al., 2011). The environmental as-sessment of emerging technologies is particularly challenging. Firstly,these technologies propose new products with different functions anda wide range of (unforeseen) applications. Secondly, the environmentalassessment is often conducted when the emerging technologies aredeveloped only at laboratory scale with high uncertainty on scaling-up effects. Finally, the emerging technologies could produce reboundeffects in the market, society and environment (Zamagni et al., 2012).
In order to assess their sustainability the European Commission encour-ages life cycle thinking (LCT) and related life cycle-based methods. LCTis a holistic approach to avoid shifting burdens to other life cycle stages,regions of the world and environmental impacts (UNEP, 2011). Lifecycle assessment (LCA), regulated by the international standards ISO14040 series (ISO, 2006a,b), is the main tool of LCT for environmentalevaluations.
To date, nanotechnology is an emerging sector; about 1.720 newnanoproducts have been introduced into the market since 2005.1
Databases have been developed in the framework of projects and na-tional initiatives to provide consumers, citizens, policymakers, andothers stakeholderswith information about the nanotechnologymarket(some examples are reported in the Nanotechnology Consumer
1 http://www.nanotechproject.org/cpi/, accessed February 3rd 2014.
123G. Barberio et al. / Science of the Total Environment 496 (2014) 122–131
Products Inventory (NCPI),2 inWissensplattformDaNa,3 Nanotechnolo-gy Products database4 and the RIVM initiative “Nanomaterials inconsumer products” (2010)). These new products have a broad rangeof applications, often with innovative functions that cover new produc-tion sectors. An LCT approach is suitable to promote a sustainabledevelopment of nanotechnologies. LCA is especially important in theearly development stages, because it helps to consider the environmen-tal impacts during the product-design process, to suggest improvementmeasures for scaling-up and to compare scenarios based on innovativeand conventional processes. Nevertheless, the data gap on LCA in thearea of nanotechnology environmental assessment is still broad, severalissues are not clearly defined and more information is necessary(Gavankar et al., 2012; Hischier and Walser, 2012; Kim and Fthenakis,2013).
Moreover, health and environment effects of nanotechnologiesand nanomaterials (NMs) are still much uncertain. The internationalscientific community is working hard on this issue. Recommendationsof the European Commission for a responsible strategy that aims toenable the safe development and use of NMs and nanotechnology(COM 243, 2005; COM 338, 2004) include the proposal of integratingrisk assessment (RA) of chemicals at all stages of the life cycle ofthe nanotechnology-base product. Human Health Risk Assessment(HHRA) and Ecological Risk Assessment (ERA) are the methods toevaluate the possible risks due to the exposure to dangerous substancesand probability of adverse health effects in humans and ecosystems,now or in the future (U.S. EPA, 2000). However, toxicity of NMscannot be extrapolated by the toxicity of their related bulk form(COM 572, 2012; SCENIHR, 2007). OECD (Organisation for EconomicCo-operation and Development) Working Party on ManufacturedNanomaterials proposes a document including current practices,challenges and strategies for assessing risk although the available dataare limited. It underlines the necessity for more research on specificrisk assessment issues and defines a brief strategy and research needsfor RA of NMs (OECD, 2012).
LCA and RAhave different aims but they seem tohave a complemen-tary role in the description of environmental impacts. Complementaryuse of LCA and RA has been suggested by many authors in the lastdecades for overcoming limits and optimizing benefits. They haveanalyzed differences at methodological level (Bare, 2006; De Haes Udoet al., 2006; Hellweg et al., 2009) and have proposed applications onchemicals and bulk forms related to NMs (Nishioka et al., 2005; Riberaet al., 2014; Scheringer et al., 2001; Socolof and Geibig, 2006; Walseret al., 2014;Wang et al., 2013;Wright et al., 2008). Different approachesfor combining these methods could be found in the literature. Theydescribe the degree of integration from the complete separation to theperfect complementarity (Flemström et al., 2004), the relevance oftaking into consideration spatial and temporal differences and thelevel only above threshold (Potting et al., 1999) and the possibility ofusing risk or hazard phrases (Askham et al., 2013). Currently, the mostcommon approach of combining LCA and RA is to use the ecotoxicolog-ical and toxicological parameters in the development of the life cycleimpact assessment (LCIA) methods. There is a general consensus inthe scientific community that toxicity indicators in LCIA cannot providethe same detailed and specific information as RA does, but they can helpidentifying where RA is necessary (Pant et al., 2004). As a consequence,the two tools should be used in combination to better support businessand policy decisions.
The aim of this paper is to propose an approach for combining RA andLCA and to present its application to theNMs. Some literature studies andalso some specific initiatives at national level propose recommendationsand suggestions concerning the use ofmethodological frameworks, addi-tional knowledge instruments such as experts' elicitation, decision tree,
multicriteria analysis (Davis, 2007; Grieger et al., 2012; Seager andLinkov, 2008; Shatkin, 2008; Som et al., 2010; Sweet and Strohm, 2006;Wardak et al., 2008). However, it is very difficult to find quantitativecase studies and robust results due both to data gap onNMS and tometh-odological barriers. Starting from the analysis of the main characteristicsand differences of RA and LCA applied to NMS, a framework for theircombined use and its application to the production of alumina NF withtwo different processes, single-stage and two-stage, are here presented.5
The qualitative RA has shown the potential effects of these productionson workers who are directly exposed to relatively high direct concentra-tions and LCA has allowed analyzing their environmental profile.
2. Method
RA (TGD, 2003) is focused on chemicals and their effects on theenvironment and human health. For a specific substance release, itallows the identification at a local scale of situations above thresholdand possible contaminations. Notably, RA is focused on the toxicity ofmaterials and doesn't take in account other environmental impacts.LCA (ISO, 14040 2006a,b), instead, allows the environmental assess-ment of products/services throughout their life cycle by consideringseveral environmental issues. Existing impact assessment methodsinclude the evaluation of the impacts at a global/regional scale, thoughthe development of spatial differentiated Characterisation factors andthe collection of site-specific inventory data could allow overcomingthis limit of the method (Zamagni et al., 2008a,b). A significant differ-ence between the twomethods concerns the reference flow considered.RA identifies the risk of each substance and the cumulative effects, sothe emissions are expressed as total emissions into an environmentalmedium (soil, water or air) with volume known, in order to obtain aconcentration value. In LCA, the Functional Unit (FU), a quantitativemeasure of the functions that the goods (or service) provide, is thebasis for comparing products: all data collected throughout the lifecycle and the potential impacts are referred to it.
As regards the toxicity assessment, RA aims to risk minimization(“only above threshold”), in agreement with the assumption that toxiceffects are caused by concentrations above a certain threshold; whileLCA has a prevention approach (“less is better”), with the assumptionthat the relationship between emissions and environmental/humanhealth damage is linear (Sleeswskij, 2011). Therefore also the goals ofthe two methods are different. RA guarantees the safety of the popula-tion and/or the environment by modeling the impact caused by theabsolute quantities of toxic substances emitted with focus on receptors.LCA assesses the overall pressure on the environment of a product fromcradle to grave (life-cycle perspective), focusing on the product's totalreleases and resources consumption and offering the best frameworkto avoid shift of burdens.
The differences above mentioned, which are weaknesses in case ofseparate use of each method, could be overtaken by their combineduse. In this study, the development of a framework for a combined useof RA and LCA started with an analysis highlighting synergisms andinteraction levels between the two methods.
In Fig. 1, the potential similarities between the RA and LCAframeworks are critically analyzed:
- The left side of Fig. 1 shows themain steps of RA, reworked fromTGD(2003). In agreement with this document, the first phase is thehazard definition and concerns the identification of the substances,the targets and the whole scenario/context. This phase is calledProblem Formulation in Fig. 1. Other phases are Exposure Assessment,Effect Assessment and Risk Characterization. Fig. 1 highlights the stepof data collection, which is included in the exposure assessment
5 This case study is part of the research project NanoHex - enhanced nanofluid heatexchange- aimed at translating promising laboratory nanotechnology results into pilotlines for the production of nanofluid coolant.
Fig. 1.Methodological framework of risk assessment/risk management (RA/RM) and life cycle assessment (LCA).
124 G. Barberio et al. / Science of the Total Environment 496 (2014) 122–131
of the TGD, and an additional step of RiskManagement, distinct fromRA. In the risk management process, the results of theRA are integrated with other considerations, such as economic orlegal concerns, to reach decisions regarding the need for and thepracticability of implementing various risk reduction activities(U.S. EPA, 2000).
- In the right side of Fig. 1, LCA is presented with its four main steps,as defined by ISO (2006a,b): Goal and scope definition, Life CycleInventory (LCI), Life Cycle Impact Assessment (LCIA) and Resultsinterpretation and the iterative procedure among these steps. Fig. 1highlights the characterization models, which are used for quantify-ing the environmental impacts and are included in the impactassessment step.
The similarities among the steps are highlighted by means of boxeswith the same filling and graphical symbol, so that Fig. 1 shows aparallel between the methodological frameworks of LCA and RA. Fromthis parallel, a new framework for a combined use of LCA and RAcould be defined throughout four steps (Fig. 2) briefly described below:
- Step 1: technological system definition, which includes the cleardescription of the technologies involved in the analysis and conse-quently the Problem Formulation in RA and the Goal and ScopeDefinition in LCA. The Problem Formulation defines the goals, thedangerous substances to be investigated and the focus of the assess-ment, e.g. workplace and/or ecological risk. The Goal and ScopeDefinition includes: the objective of the study (goals), the FU, thesystem description and the boundaries, i.e. the phases of the systemthat are included in the study. For bothmethods this step consists inhighlighting the breadth of the problem. In a combined use,elements of integration can be identified.
- Step 2: data collection, which includes the Emission Identificationof the RA and the Inventory Analysis of the LCA. Any informative re-port regarding Safety and Health of substances (i.e. environmentalreport, registration dossier for REACH, etc.), workers' exposure
Fig. 2. Proposal of methodological framewo
(frequency and time) during production to the substances of inter-est, information on working area, is a source of data for the RA. Forthe LCA production data refer tomaterials and energy consumption,transport and packaging, direct emissions into the environment.Moreover information on the environmental exposure to referencesubstances and cumulative effects (needed for RA) and on the useand end of life of the product (needed for LCA) has to be collected.Data can be collected by companies along the supply chain, ormodeled and estimated on the basis of literature and experts'judgment. The use of a common form to collect data and informationneeded for the RA and the LCA is an integration aspect that canoptimise this step.
- Step 3: risk evaluation (risk characterization) and impactsquantification (life cycle impact assessment—LCIA), which areobtained by combining data (from step 2) and mathematicalmodels, such as the Exposure Assessment and Effect Assessment inRA and in Characterization Models in LCA. The models used in RAand in LCA are different. The exposure assessment models for theRA can achieve high levels of complexity: they analyze a specificflux release infixed time and space; they can be analytical or numer-ical; they can analyze the contaminants dispersion in two or three-dimension, in static or dynamic conditions. The commonest LCIAmodels are spatially and temporally aggregated. Regarding theeffects assessment, RA considers the impacts on human healthand the ecosystems, while LCA analyses many impact categories(e.g. acidification, global warming, depletion of resources, humanhealth, …). Due to these differences, they could be combined intwo different ways: a) applying separately the models to the sametechnological system (defined in the step 1) and analyzing theresults or b) performing an RA study and then using those resultsin the characterization phase of LCIAmodels (human and eco toxic-ity categories).
- Step 4: results interpretation. The results interpretation for the RAcan be identified in the Risk Management. Starting from the RA
125G. Barberio et al. / Science of the Total Environment 496 (2014) 122–131
results and other considerations – involving political, social,economic and science/engineering factors – Risk Managementdevelops, analyses and compares options and selects the optimalresponse for safety from that hazard (U.S. EPA, 2000). The iterativerelationship between RA and RM has been highlighted in Leeuwen(2007). In LCA this mandatory phase is named Interpretationphase. The life cycle interpretation phase considers severalelements, as defined in the ISO standard, a) identification of theenvironmental significant issues; b) evaluation of study complete-ness, sensitivity and consistency; and c) definition of conclusions,limitations, and recommendations for reducing the environmentalimpact of the product system. This introduces an iterative processof revision of all the previous steps of the LCA study.
3. Case study
The framework proposed has been applied to the aluminananofluid6 (NF) production, as herein described:
1. an LCA-based approach has been adopted for the technologicalsystem definition in order to consider the upstream and down-stream stages and their relevance for the inventory and the impactassessment. The technological system definition is shared by boththe qualitative RA (problem formulation) and the LCA (goal andscope definition);
2. the data collection combines information to develop the qualitativeRA and the LCA study;
3. for the risk evaluation and impacts quantification, methodsand assessment models have been chosen separately and appliedindependently;
4. results interpretation has been carried out in an integrated way asit gives important information on different aspects, i.e. the bestproduction process performance and the workplace exposure.
3.1. Step 1: technological system definition
In line with the framework, the problem formulation for qualitativeRA and the goal and scope definition for LCA are shared by means of adetailed technological system definition. In the NanoHex project apilot line has been developed to produce NF coolants, which can beused as an enhanced heat transfer fluid in the heat radiator, engine,refrigeration or air conditioning systems.
The technological system is the pilot line producing Al2O3 (alumina)NF coolants with:
• a single-stage process: wet chemical synthesis allows the productionand dispersion of tailored nanoparticles within a carrier fluid;
• a two-stage process: pre-produced nanoparticles are added to acarrier fluid.
Fig. 3 shows thephases for theproduction of aluminaNFwith single-stage (on the left) and two-stage (on the right) process. The single-stageprocess incorporates in the same production line the production of theNF, from particle formation to stable dispersion. However, it is alsopossible to purchase commercially available NPs of alumina for the NFproduction. In this case the NF production is divided into two complete-ly separate steps (external particle production and proprietary fluidformulation), i.e. a two-stage process. For the two-stage process a vesselfor homogenization, amill or an ultrasonifier for dispersion and a rotaryvacuum evaporator to remove ethanol are the necessary equipment. Areactor and an autoclave are also needed for particle synthesis in thesingle stage route. All in all, from a technical viewpoint, single stageproduction is far more versatile, while two-stage production is fasterand more cost effective.
6 A nanofluid is afluid, generallywater, oil orwater and glycol, which could contain sev-eral nanoparticles as metals, metals oxides, carbon nano tube, depending on the final use.
3.1.1. Goal and scope of the LCA studyThe goal of the study is to assess and compare the environmental
impacts of two different processes of alumina NF production: single-stage and two-stage. The function of the systems is the production ofNF, the functional unit is the production of 1000 kg of NF and the systemboundaries are from cradle to gate. Being a comparative study, onlydifferences among the two production systems are modeled.
Processes included in the aluminaNF single-stage production are thefollowing:
o materials for poly-aluminium chlorohydrate production;o production and waste treatment of aluminium oxide, ethanol and
water;o European electricity low voltage productionmix (Ecoinvent DB v.2,
2010), which considers 29.2% of nuclear energy, 14.4% of hydroenergy, 1.1% of hydropower energy (pumping systems), 50.7% offossil fuel, 3.3% of renewable energy, 1.2% of energy from waste;
o transport of materials;o packaging of materials and final product; ando waste disposal, which includes the incineration of cardboards and
hazardous waste (ethanol solution).
Processes included in the alumina NF two-stage production are thefollowing:
o production of NPs of Al2O3, including energy consumption andprimary aluminium production;
o production and waste treatment of ethanol and water;o European electricity low voltage production mix from Ecoinvent DB
v.2 (2010);o transport of materials;o packaging of materials and final product; ando waste disposal, which includes the incineration of cardboard and
hazardous waste (ethanol solution).
The LCA studyhas been carried out according to ISO 14040 standards(ISO, 2006a,b) and considering the technical handbook published by theEuropean Platform on LCA as part of the International Reference LifeCycle Data System (JRC-IES, 2010).
3.1.2. Problem formulation of RAThe technological system investigated includes processes occurring
in closed loops. As regards the nanoparticles (NPs) production in inputinto the two-stage process, data from literature justify the assumptionthat no airborne emissions occur (see Section 3.2). For these reasons,outdoor potential release of NMs or NPs can be considered negligibleand the RA problem formulation, which includes the definition ofhazard and targets of interest, has led to focus the study on HHRA ofthe alumina NF as directly relevant for workers, i.e. on the occupationalexposure. For the single-stage process the RA has been performed onthe production of NF starting from bulk precursors; for the two-stageprocess the RA has been performed on the production of NF and of theNPs of alumina, which is the nano-precursor used in the process. Lackof data about exposure limits, exposure–dose relationships and toxicityof alumina NPs and NF, did not allow quantitative risk assessment. Thisproblem, which is often encountered in studies on NMs (OECD, 2012),can be overtaken by using qualitative risk assessment. In the scientificcommunity several methods are available: weight of evidence (Zuinet al., 2011), categorization (Hansen et al., 2007), prioritization.Moreover the interest in applications is increasing and some tools arealso available: Advanced RECH Tool (ART) (van Tongeren et al., 2011),Control Banding (CB) tools (NIOSH, 2009) and others.
In this study a qualitative risk assessment of nano-objects has beenperformed using Stoffenmanager Nano, a ‘work-in-process' online toolthat reflects the current knowledge on risks related to working withNMs (van Duuren-Stuurman et al., 2011).
Nanopowder n-Al2O3
Suspension 20wt% Al2O3 in ethanol
MillingBreaking agglomerates,
Surface modification
Octyltriethoxysilan
Solvent exchange by distillation in a
rotary vacuum evaporator
and dilution to 9w% Al2O3
Two-stage nanoaluminaproduction
Al(OH)nCl6-n
50% aqueous solutionAl2O3
Seed crystals
Oven Evaporisation at 120 C Calcination at 1000 C
Suspension20wt% Al2O3 in ethanol
MillingBreaking agglomerates,
Surface modification
Octyltriethoxysilan
Solvent Exchange by distillation in a
rotary vacuum evaporator
Deionised water
Nanofluid
Ethanol
Single-stage nanoaluminaproduction
Ethanol
Nanofluid
Transport and packaging
T&P
T&P
T&P
Included processes Excluded process
T&P
T&P
Ethanol incineration
and packagingtreatment
Deionised water
Ethanol incineration
and packagingtreatment
T&P T&P
T&P
Legend:
Fig. 3. Technological system definition of single-stage and two-stage processes.
126 G. Barberio et al. / Science of the Total Environment 496 (2014) 122–131
3.2. Step 2: data collection
Concerning the second step, data needed to develop the qualitativeRA and the LCA study have been effectively collected in agreementwith the requirements of both methods and the definition of theboundaries.
Primary data collection has been performed by providing question-naires and spread-sheets to NanoHex partners. The questionnairesaimed to collect general information about:
• availability of internal reports regarding Safety and Health of nano-particles, production and use (i.e. environmental report, registrationdossier for REACH,…);
• NPs production process, with a detailed description of the technology(i.e. flow-chart …) and the source domain of potential release of NPs(handling of NPs, release of primary particles during synthesis, …);description of working area; duration and frequency of productiontask; steps in which the operator may enter in contact with NPs andrelated safety measures (personal protective equipment and localcontrol measure of room); and
• production of NF as: a) transport ofmaterials with detail on transport,frequency and distance between suppliers and producers; b) estima-tion of energy consumption due to NPs and/or NF production;c) description of waste (amount and types) generated during NPsand/or NF production and their final treatment or disposal; andd) emissions into the environment during the production.
Tables 1 and 2 show inventory data for the single and two-stage NFproduction, respectively.
Primary data of the pilot lines producing nanoalumina with single-stage and two-stage process have been collected at ItN7 factory. Primarydata on materials, as aluminium chlorohydrate and NPs of alumina
7 www.itn-nanovation.de.
could be obtained neither from suppliers nor from commercial data-bases, so a literature analysis has been performed and data on laborato-ry processes could be found. Being the amount of octyltriethoxysilanethe same for both alumina NF system productions, its production andtransport have been kept out of the system boundaries. All the otherbackground data are from the commercial database Ecoinvent 2.0.
A sensitivity analysis was performed to check the robustness of theresults (see Section 3.3.2). As regards the poly-aluminium chloride(PAC, [Al(OH)nCl6 − n]m) production, only qualitative data of inventorycould be found, so the production has been modeled by using thestoichiometry of reaction of aluminium hydroxide and aluminiumchloride, with n varying between 1 and 5:
nAlCl3 þ 6−nð ÞAl OHð Þ3→ 3Al2 OHð Þ6−nCln ð1Þ
The reaction having n = 5 has been chosen for the “BASE CASENanofluid of alumina single-stage, Al2OHCl5”. Only material productionhas been considered, while energy consumption and production wastehave not been included. Primary data for AlCl3 production are from afactory8 and the data on aluminium hydroxide production are fromEcoinvent 2.0 (RER: aluminium hydroxide, at plant). As regards theNPs production different methods could be found in the literature,which are used in industrial applications such as electronics,metallurgy,optoelectronics and fine ceramic composites (Hassanzadeh-Tabrizi andTaheri-Nassaj, 2009; Prete, 2010): microwave synthesis (Ebadzadehand Asadian, 2009), precipitation (Hassanzadeh-Tabrizi and Taheri-Nassaj, 2009), sol–gel (Shojaie-Bahaabad and Taheri-Nassaj, 2008)and production by pulsedwire discharge (PWD) in air flow atmosphere(Cho et al., 2003; Ishihara et al., 2012). PWD has been selected in thisstudy because it can produce different nanosized powders, it has beenwidely investigated and some inventory data are available. The amountof nanoalumina has been calculated based on the stoichiometry ofreaction. Estimation of energy consumption and weight loss of PWDwas assumed by Cho et al. (2003). These authors have observed a
Table 1Inventory for the production of 1000 kg of Alumina NF with single-stage process.
Type Source of data Amount(kg/a)
Unit Type ofpackaging
Type oftransport
Average distance to distributioncenter (km)
Al(OH)n Cln-6 Data from literature 400 kg Drum Truck 200Al2O3 Ecoinvent v.2.0 [RER: aluminium oxide, at plant/benefication] 2 kg cardboard bag Truck 200Deionized water Ecoinvent v.2.0 [CH: water, deionized, at plant] 900 kg IBC Truck 15Ethanol Ecoinvent v.2.0 [RER: ethanol from ethylene, at plant] 400 kg Drum Truck 200Electricity Ecoinvent v.2.0 [RER: electricity, low voltage, production RER, at grid] 107999 MJIncineration of cardboard CH: disposal, packaging cardboard, 19.6% water, to municipal
incineration [municipal incineration]0.04 kg Transport to incinerator included
400 kg Transport to incinerator includedin the DB process
127G. Barberio et al. / Science of the Total Environment 496 (2014) 122–131
weight loss due only to adhesion of particles to the chamber, electrodesand other components inside the chamber, and not to airborneemissions.
3.3. Step 3: risk evaluation and impacts quantification
In this third step, RA and LCA models have been applied separately.Here models and results are presented.
3.3.1. RA models of risk evaluation and resultsStoffenmanager Nano tool allows the qualitative assessment of
occupational health risks from inhalation exposure to ManufacturedNano-objects (MNO) using the Safety Data Sheets (SDS) and/or productinformation sheets. Stoffenmanager Nano is a risk-banding tool devel-oped for employers and employees to prioritize health risks in exposuresituations where quantitative risk assessment is not possible yet and toassist implementation of control measures to reduce exposure levels(van Duuren-Stuurman et al., 2011, 2012). The risk is characterized inlow, average and high priority by combining the available hazard infor-mation of a substance (classified from “very hazardous” (class E) to “nothazardous” (class A)) with an estimation of inhalation exposure(classified from 1 (low exposure) to 4 (very high exposure) by usinginformation on substance, production process and working area.
Five phases are necessary to develop a qualitative risk assessment inthe Stoffenmanager tool:
• phase 1: a “general step”, which considers information on the mainprocess and the source domain of potential release of NPs (handlingof NPs, release of primary particles during actual synthesis, sprayingor dispersion of a ready-to-use nanoproducts, fracturing and abrasionof nanoproducts-embedded end products).
• phase 2: “Product characteristics”, which includes information asproduct information sheets and material safety data sheets (if avail-able); dustiness; moisture content; concentration of nanocomponentin the product; inhalation hazard.
• phase 3: “Handling process”, which takes into account the main ele-ments for characterizing tasks such as handling of products with low
Table 2Inventory for the production of 1000 kg of alumina NF with two-stage process.
Type Source of data
Nanopowder of Al2O3 Literature dataDeionized water Ecoinvent v.2.0 [CH: water, deionized, at plant]Ethanol Ecoinvent v.2.0 [RER: ethanol from ethylene, at plant]Energy consumption RER: electricity, low voltage, production RER, at grid [productionIncineration of cardboard CH: disposal, packaging cardboard, 19.6% water, to municipal
speed or little force or in medium quantities (several kilograms) orhandling of products with a relatively high speed/force which leadsto dispersion of dust and so on; duration task; frequency task;information on employees and their distance from breathing zone.
• phase 4: “working area” with information on frequency of cleaning;volume and ventilation of working room.
• phase 5: “Local control measures and personal protective equipment”.
In this study, the analysis has been carried out for three “nano-objects”: Al2O3 NPs, Al2O3 NF/two-stage production, Al2O3 NF/single-stage production. For a comprehensive evaluation of the Al2O3 NFproduced by the two-stage process we have to consider both the resultsof Al2O3 NPs and Al2O3 NF production/two-stage. Themain results fromStoffenmanager analysis are given in Table 3.
If there are no sufficient toxicological data, as in this case study, thepossible classification of substance is in the range from class C to classE and the assignment is given using information from parent materials.In this case the hazard class is D, which means very hazardoussubstance, and has been estimated considering parent materials.
The exposure score is calculated by multiplying the respectiveweighting factors by the parameters, requested in the five phasesduring compilation of Stoffenmanager data collection and summingthe results (on a logarithmic scale).
The exposure and risk can be evaluated for the single event, namely“task”, and for the annual production, namely “time weighted”.Concerning the exposure, as shown in Table 3, low task and timeweight-ed exposure classes have been assigned to Al2O3 NF, in both single- andtwo-stage process. Average time weighted and high task exposureclasses have been assigned to Al2O3 NPs production. Finally, concerningrisk evaluation, all the processes considered (Al2O3 NF, single- andtwo-stage, Al2O3 NPs) have average risk score in the class timeweighted.The risk scores of the task are average for the production of Al2O3 NFsingle- and two-stage process and high for Al2O3 NPs production.
The reason for higher exposure, and consequently higher risk, of NPsproduction if compared to NF production can be found in the “datacollection” of these processes. Lack of data for parameters requested inStoffenmanager Nano is broader for NPs than for NF production. In
Amount(kg/a)
Unit Type ofpackaging
Type oftransport
Average distance to distributioncenter (km)
90 kg Cardboard bag Truck 200901 kg IBC Truck 15360 kg Drum Truck 200
mix] 35,999.71 MJ1.8 kg Transport to incinerator included
in the DB process360 kg Transport to incinerator included
in the DB process
Table 3Final results of qualitative risk assessment for the three systems analyzed, using Stoffenmanager tool.
Al2O3 Hazard class Exposure class (time weighted) Exposure class of task Risk score of task Risk score (time weighted)
Two-stage, nanopowder D 2 3 3 2Two-stage, nanofluid production D 1 1 2 2Single-stage, nanofluid production D 1 1 2 2
Hazard range is from A (low) to E (extreme). Exposure is classified in 4 levels from 1 (low) to 4 (very high). Risk priority is classified as low (1), middle (2), high (3).
128 G. Barberio et al. / Science of the Total Environment 496 (2014) 122–131
these cases the respective weighting factors to compile the exposurescore algorithmic calculation cannot be assigned, so conservativedefault factors are used in line with the precautionary principle. ForNPs production, the parameters with scarcity of data can be found: inthe source domain that generate the release (phase 1 of general step);in the handling process (phase 3); some data of working area (phase 4)such as for ventilation; local control measures (phase 5). Moreover,concentration of nanoparticles, which is higher in NPs production thanin NF, strongly influences the results of exposure score.
3.3.2. LCA models of impact assessment and resultsThe characterizationmethod adopted to evaluate the environmental
impacts is IMPACT 2002+ (Jolliet et al., 2003). According to Jolliet et al.(2003), for the midpoint category Global Warming Potential (GWP),temporal scale level considers overall long-term effects through theuse of infinite time horizons (i.e. 500 years horizon in GWP), but asrecommended by the ILCD Handbook (JRC-IES, 2011) also the IPCC100 years is considered for the GWP impacts. To date, characterizationfactors of NMs for toxicity categories (aquatic ecotoxicity, carcinogens,non carcinogens and terrestrial ecotoxicity) are lacking, so all the poten-tial environmental impacts refer to the bulk chemicals. The scientificcommunity is developing the standard tests to measure the toxicity ofNMs, but nowadays information is not sufficient to develop characteri-zation factors for LCA impact categories.
Table 4 shows the characterization results for the NF pro-duction single-stage (“BASE CASE Nanofluid of alumina single-stage(Al2OHCl5)”) and two-stage (“BASE CASE Nanofluid of alumina two-stage”) in the gray columns. The former has higher potential impactsdue to higher energy and materials consumption in the single-stageproduction.
LCA results show that “BASE CASE Nanofluid of alumina two-stage”is the preferable production process with the data presently available.Therefore, a deeper analysis of NF production processes has beenperformed to better understandwhether environmental improvementsare possible.
Following the priorities set by the “A resource-efficient Europe—Flagship initiative of the Europe 2020 Strategy”9 and the “EuropeanClimate Change Program”,10we analyzed themineral extraction catego-ry, which refers to the depletion ofmetals ores (e.g. copper and bauxite)and the global warming potential, which considers the rise in theaverage temperature of Earth's atmosphere and oceans since the late19th century.
For the mineral extraction (Fig. 4) the production of electricityand poly-aluminium chlorohydrate contributes to the total results ofthe NF single-stage process by 69% and 27% respectively, followed byethanol production (4%). In NF two-stage, the production of electricityand alumina NPs contributes by 57% and 34%, followed by ethanolproduction (9%).
In single-stage production the potential impacts are due to the use ofmetals in the production of electricity and poly-aluminiumchlorohydrate. In particular the main impacts are due to copper andnickel consumption for electricity production (about 38% and 29%respectively) and to aluminium for the production of poly-aluminium
chlorohydrate (18% of the total mineral extraction impacts). In two-stage production, relevant flows are copper and nickel for the electricityproduction and aluminium for the production of NPs of alumina,responsible for 35%, 36% and 27% respectively.
Concerning global warming potential 500 years (Fig. 5), energyconsumption gives the highest contribution (94%) in the single-stageproduction process. In the two-stage process, energy consumptioncontributes the 80%, followed by the production of NPs and ethanol(12% and 6% respectively).
The most relevant elementary flows for the GWP 500 years are theemissions into air of CO2 and VOC, respectively responsible for 93%and 1% in the single-stage process, and 91% and 3% in the two-stageone. The high impact of CO2 emissions is due to the large contributionof fossil fuels to the production of electricity.
For the sensitivity check of the PAC production the process“Nanofluid of alumina single-stage (Al2(OH)5Cl)” (where n has beenput equal to 1 in the reaction (1)) has been assessed. The results showthat the single-stage productionwith Al2(OH)5Cl has potential environ-mental impacts comparable (in a range of 1%–5%) with Al2(OH)Cl5(Table 4), except for the ozone layer depletion, where the difference(about 16%) is due to the production of hydrochloric acid. Hence theuse of Al2OHCl5 as a proxy of the entire family of PAC does not introducesignificant differences in the environmental assessment.
As regards the two-stage process, the “BASE CASE Nanofluid ofalumina two-stage” has also been compared with processes whereatomized Al powder (Scalbi and Masoni, 2012), NPs of zirconia(data supplied by ItN) and NPs of SiC (Reau et al., 2012) substitute foralumina NPs (Table 4). The use of different raw materials affects theresults of LCA impact assessment up to 36%. Though the differencesare large, the results show that in any case the two-stage process hasbetter environmental performance if compared to the single-stageprocess.
3.4. Step 4: Interpretation
In the case study the workplace exposure has been evaluatedby means of a qualitative RA and the following impact categorieshave been analyzed in the LCA study: aquatic acidification, aquaticecotoxicity, aquatic eutrophication, carcinogens, global warming(500 years), Global warming (100 years), ionizing radiation, landoccupation, mineral extraction, non carcinogens, non-renewableenergy, ozone layer depletion, photochemical oxidation, respiratoryeffects, terrestrial acidification/nitrification, and terrestrial ecotoxicity.The results of occupational risk and environmental impacts areconflicting:
• On the basis of qualitative RA results, the risk of “Nanofluid of aluminasingle-stage” is medium and comparable to that for “Nanofluid ofalumina two-stage”. However in the latter we have to consider alsothe contribution to the risk of “Nanopowder of alumina for two-stage”, which is high.
• On the basis of LCA results and of the data presently available,“Nanofluid of alumina two-stage” has better environmental profilethan “Nanofluid of alumina single-stage”, due to lower energy andmaterials consumption respect to the single-stage production.A sensitivity analysis on the data affected by major uncertainty,
Table 4Characterization results for NF production with different NPs.
Impact categories Unit BASE CASE nanofluidof alumina single-stage (Al2OHCl5)
Nanofluid ofaluminasingle-stage(Al2(OH)5Cl)
BASE CASEnanofluidof aluminatwo-stage
Nanofluid of aluminatwo-stage withatomizer Al powder
Nanofluid ofaluminatwo-stagewith SiC
Nanofluid ofaluminatwo-stagewith zirconia
Aquatic acidification kg SO2-eq 5.7E+01 5.6E+01 2.2E+01 2.3E+01 1.9E+01 1.9E+01Aquatic ecotoxicity kg TEG-EQ 7.3E+07 7.7E+07 2.9E+07 3.2E+07 2.0E+07 2.3E+07Aquatic eutrophication kg PO4 eq 8.9E−01 9.2E−01 5.8E−01 6.2E−01 5.0E−01 5.1E−01Carcinogens kg C2H3Cl eq to air 4.1E+01 4.2E+01 2.1E+01 2.5E+01 1.4E+01 1.7E+01Global warming 500 years kg CO2 eq 1.7E+04 1.7E+04 6.8E+03 7.2E+03 6.0E+03 6.6E+03Global warming 100 years kg CO2 eq 1.8E+04 1.8E+04 7.1E+03 7.4E+03 6.2E+03 6.8E+03Ionizing radiation Bq-C14 1.5E+06 1.5E+06 5.3E+05 5.4E+05 4.9E+05 5.1E+05Land occupation m2yr eq 7.4E−03 7.3E−03 2.8E−03 3.0E−03 2.5E−03 2.7E−03Mineral extraction MJ surplus 1.3E+03 1.3E+03 5.3E+02 6.2E+02 3.5E+02 3.6E+02Non Carcinogens kg C2H3Cl eq to air 3.6E+02 3.7E+02 1.5E+02 1.6E+02 1.1E+02 1.3E+02Non-renewable energy MJ surplus 3.5E+05 3.5E+05 1.4E+05 1.4E+05 1.3E+05 1.4E+05Ozone layer depletion kg CFC-11 eq 1.1E−03 9.2E−04 3.3E−04 3.9E−04 2.9E−04 3.1E−04Photochemical oxidation kg C2H4 eq 2.7E+00 2.7E+00 1.5E+00 1.6E+00 1.4E+00 1.5E+00Respiratory effects PM2.5 eq 1.4E+01 1.4E+01 5.4E+00 5.8E+00 4.5E+00 4.7E+00Terrestrial acidification/nitrification kg SO2-eq 2.3E+02 2.3E+02 8.8E+01 9.3E+01 7.8E+01 8.2E+01Terrestrial ecotoxicity kg TEG-EQ soil 2.5E+05 2.4E+05 9.1E+04 9.6E+04 8.1E+04 8.3E+04
129G. Barberio et al. / Science of the Total Environment 496 (2014) 122–131
confirmed that the two-stage process has always better environmen-tal performance if compared to the single-stage process.
A trade-off could be achieved by reducing the uncertainty of NPsproduction data. Moreover, thanks to a deeper knowledge of the NPsproduction processes, some practices of Risk Management could beapplied to reduce the risk. On the other side a trade-off could be obtain-ed by reducing the energy consumption in the single-stage productionto improve the environmental assessment.
4. Discussion and conclusions
This paper proposes a framework for the combined use of LCA andRA to provide scientifically sound information for the early assessmentof potential impacts on health, safety and the environment of emergingtechnologies, notably nanotechnologies. The case study was focused onthe production of NF.
The first challenge of the framework application was to combinetwo different scales of spatial resolution: in fact, RA is site-specific and
Fig. 4. Characterization results of mineral
therefore high spatial resolution is required, while LCA is typically atregional/global scale and has low spatial resolution. The combined useof the two methods has required the definition of a common scope,which includes both site-generic and site-specific data (Step 1). TheRA was performed on the work environment by collecting site-specificdata, such as the pollutant concentration and the worker exposureconditions, while site-generic characterization factors have been usedin the LCA study.
Both methods have presented difficulties with data availability buttheir integration has provided an opportunity for saving time andreducing costs. In fact, if the technological system is well defined, i.e. ifprocesses are known andmodels and tools for the analysis are selected,a complete data collection (Step 2) at enterprises can be carried out forboth studies simultaneously. This has been tested in this case study,where detailed and integrated data sheets have been distributed tothe enterprises for optimizing the collection of data needed for RA andLCA.
The case study has demonstrated that the integration of the twomethods increases the significance of the assessment and provides a
Fig. 5. Characterization results of global warming and contribution analysis.
130 G. Barberio et al. / Science of the Total Environment 496 (2014) 122–131
more comprehensive evaluation (Step 3). The lack of characterizationfactors in the LCIAmethods impedes the evaluation of the toxic impactsdue to the release of the nanoproducts in the environment, but this limitcan be overcome by integrating the RA results. Actually, the combina-tion of the twomethodswould be particularly effective if values obtain-ed in RA were used for the development of the characterization factorsin the toxicity impact categories of LCA (Olsen et al., 2001), to fill the gapof knowledge on the environmental fate and behavior of nanoproductsand on the effects due to the exposure to them. As far as we know, onlytwo published studies report effect- and/or characterization factor - ofnanoproducts. Eckelman et al. (2012) estimated the effect factor ofCNTs for freshwater ecotoxicity and Salieri et al. (2013) calculated thecharacterization factor of TiO2 also for freshwater ecotoxicity. No char-acterization factors are available for human toxicity of nanoproducts.Even if in this study the lack of toxicity data on alumina NF preventedfrom producing characterization factors, the combined use of themethods has allowed the assessment of the occupational risk, which isnot taken into account in the life cycle impact assessment methodscurrently available. Further research should be addressed to create asub-category of human toxicity that deals with the occupational riskto obtain robust results for evaluations of products, especially theemerging ones (Hellweg et al., 2005; Hidalgo et al.;, 2013; Scanlonet al., 2013).
Thanks to the integration of the two methods, a more tailored RAmight be performed by addressing the hotspots highlighted by LCAthrough the entire life-cycle. In this case study the application of theframework has been somehow limited, because the system boundariesinclude the production phase of the NF, but not the use and end of life.Indeed, a complete LCA study, including all life cycle phases, couldlead to identify the most critical environmental phases and to focusthe analysis on specific substances and targets of these phases.
The combined use of RA and LCA in the interpretation step (Step 4)allowed an in-depth analysis of the system and a wider vision of theproblems to better support the decisionmaking process. Indeed, the in-terpretation requires analyzing the hotspots and suggesting possibletrade-off between the (qualitative) RA results, which identify thescenarios having the highest risk for workers, and the LCA results,which identify the processes having the best environmental perfor-mance. This trade-off can be obtained by working on the reduction of
risk through Risk Management or by improving the environmentalassessment through eco-design criteria. Further investigation isnecessary to have more data for the NPs evaluation, in particular toimprove the quality of the RA study and to expand the boundaries ofthe technological system.
The study has highlighted the hotspots in the use of combinedmethods, showing conflicting results between RA and LCA. Weightingconflicting results is the same problem encountered in the multicriteriaanalysis. Linkov et al. (2011) propose a model, based on multicriteriadecision analysis (MCDA) and a value of information (VoI) approach,for prioritizing research strategies in nanotechnologies in termsof EHS. In any case, a trade-off in the system could be obtained in afuture industrialization phase, where an appropriate design of theproduction processes in all phases could reduce the risk, by usingstrategies of Risk Management or by improving the efficiency in theuse of resources.
As outlined in Introduction, there are three main challenges inconducting an assessment of emerging technologies, i.e. wide range ofpossible applications, scale up and rebound effects. No one of theseaspects could be covered in this case study. As it has considered onlythe production phase, without including application, use and end oflife, the innovative functions of NF were not analyzed neither possiblerebound effects could be clearly identified. Moreover, this study refersonly to a pilot-line. Nevertheless, though the study here presented can-not be considered a full assessment of theNF, it has been an opportunityfor testing and demonstrating the effectiveness of a combined use of RAand LCA. Their combined use should be encouraged, especially foremerging technologies, with the aim to take into consideration thesafe-by-design concept and to stimulate a responsible and sustainabledevelopment based on economic growth, social values and reductionof environmental burdens.
Acknowledgments
The contribution of the NanoHex project (www.nanohex.eu),funded by the Seventh Framework Programme for collaborativeprojects, of Heiko Poth of ITN and Alessandro Stocco of Carpi s.r.l., whoprovided useful data for the study, is acknowledged.
131G. Barberio et al. / Science of the Total Environment 496 (2014) 122–131
References
Askham C, Gade AL, Hanssen OJ. Linking chemical risk information with life cycleassessment in product development. J Clean Prod 2013;51:196–204.
Bare JC. Risk assessment and life-cycle impact assessment (LCIA) for human healthcancerous and noncancerous emissions: integrated and complementary withconsistency within the USEPA. Human Ecol Risk Assess 2006;12:493–509.
Cho C, Yoshiaki K, Hisayuki S, Weihua J, Kiyoshi Y. Enhancement of nitridation in synthe-sis of aluminum nitride nanosize powders by pulsed wire discharge. Jpn J Appl Phys2003;42:1763–5. [Part 1, No. 4A, April 2003].
COM 243. Nanosciences and nanotechnologies: an action plan for Europe 2005–2009.European Commission; 2005.
COM 338 (communication). Towards a European strategy for nanotechnology. COM 338(communication); 2004.
COM 572. Second regulatory review on nanomaterials. European Commission; 2012.Davis JM. How to assess the risks of nanotechnology: learning from past experience.
J Nanosci Nanotechnol 2007;7:402–9.De Haes Udo, Sleeswijk AW, Heijungs R. Similarities, differences and synergisms between
HERAandLCA—an analysis at three levels. HumanEcol RiskAssess 2006;12(3):431–49.Ebadzadeh T, Asadian K. Microwave-assisted synthesis of nanosized alfa-Al2O3. Powder
Technol 2009;192:242–4.Eckelman MJ, Mauter MS, Isaacs JA, Elimelech M. New perspectives on nanomaterial
aquatic ecotoxicity: production impacts exceed direct exposure impacts for carbonnanotoubes. Environ Sci Technol 2012;46(5):2902–10.
Ecoinvent database version 2. report available at http://www.ecoinvent.org/database/ecoinvent-version-2/, 2010.
Flemström K, Carlson R, Erixon M. Relationships between life cycle assessment and riskassessment—potentials and obstacles. Report 5379, Industrial Environmental Infor-matics (IMI). Chalmers University of Technology; 2004.
Gavankar S, Sangwon Suh, Keller Arturo. Life cycle assessment at nanoscale: review andraccomandations. Int J Life Cycle Assess 2012;17:295–303.
Grieger KD, Laurent A, Miseljic M, Christensen F, Baun A, Olsen SI. Analysis of current re-search addressing complementary use of life-cycle assessment and risk assessmentfor engineered nanomaterials: have lessons been learned from previous experiencewith chemicals? J Nanopart Res 2012;14:958.
Hansen SF, Larsen BH, Olsen SI, Baun A. Categorization framework to aid hazard identifi-cation of nanomaterials. Nanotoxicology 2007;1(3):243–50.
Hassanzadeh-Tabrizi SA, Taheri-Nassaj E. Economical synthesis of Al2O nanopowderusing a precipitation method. Mater Lett 2009;63:2274–6.
Hellweg S, Demou E, Scheringer M, McKone TE, Hungerbühler K. Confronting workplaceexposure to chemicals with LCA: examples of trichloroethylene and perchloroethy-lene in metal degreasing and dry cleaning. Environ Sci Technol 2005;39(19):7741–8. [2005 Oct 1].
Hellweg S, Demou E, Bruzzi R, Mejer A, Rosenbaum RK, Huijbregts MAJ, et al. Integratinghuman indoor air pollutant exposure within life cycle impact assessment. Environ SciTechnol 2009;43:6.
Hidalgo C, González-Gálvez D, Janer G, Escamilla M, Vázquez-Campos S. Life cycle assess-ment (LCA) of nanomaterials: a comprehensive approach. The 6th InternationalConference Life Cycle Management in Gothenburg 2013; 2013.
Hischier R, Walser T. Life cycle assessment of engineered nanomaterials: state of the artand strategies to overcome existing gaps. Sci Total Environ 2012;425:271–82.
Ishihara S, Hisayuki S, Tadachika N, Tsuneo S, Koichi N. Synthesis of nanosized aluminapowders by pulsedwire discharge in airflowatmosphere. Ceram Int2012;38:4477–84.
ISO (International Organization for Standardization) (UNI EN) 14040. Environmentalmanagement–life cycle assessment (LCA)—Principles and Framework. Switzerland:Geneve; 2006a.
ISO (International Organization for Standardization) (UNI EN) 14044. EnvironmentalManagement–LCA—requirements and guidelines. Switzerland: Geneve; 2006b.
Jolliet O, Margni M, Charles R, Humbert S, Payet J, Rebitzer G, et al. IMPACT 2002+: a newlife cycle impact assessment methodology. Int J LCA 2003;8(6):324–30.
Kim HC, Fthenakis V. Life cycle energy and climate change implications of nanotechnol-ogies. Journal of Industrial Ecology 2013;17:528–41.
Leeuwen CJ. General introduction. In: Leeuwen CJ, Vermeire TG, editors. Risk assessmentof chemicals. Dordrecht, the Netherlands: Springer edition; 2007.
Linkov I, Bates ME, Canis LJ, Seager TP, Keisler JM. A decision-directed approach for prior-itizing research into the impact of nanomaterials on the environment and humanhealth. Nat Nanotechnol 2011;6:784–7. http://dx.doi.org/10.1038/nnano.2011.163.
Mulder K, Ferrer D, van Lente H. What is sustainable technology? Perceptions, paradoxesand possibilities. Sheffield, UK: Greeleaf Publishing; 2011.
NIOSH (National Institute for Occupational Safety and Health). Qualitative risk character-ization and management of occupational hazards: control banding (CB): a literaturereview and critical analysis; 2009 [CDC Aug. 2009].
Nishioka Y, Levy JI, Norris GA, Bennett DH, Spengler JD. A risk-based approach to healthimpact assessment for input-output analysis. Part 1: methodology. Int J LCA 2005;10(3):193–9.
OECD (Organisation for Economic Co-operation and Development). Important issues onrisk assessment of manufactured nanomaterials. Series on the Safety ofManufacturedNanomaterials; 2012 [March 2012].
Olsen S, Christensen F, Hauschild M, Pedersen F, Larsen H, Tørsløv J. Life cycle impact as-sessment and risk assessment of chemicals—a methodological comparison. EnvironImpact Assess Rev 2001;21(4):385–404.
Pant R, Van Hoof G, Schowanek D, Feijtel TCJ, De Koning A, HauschildM, et al. Comparisonbetween three different LCIA methods for aquatic ecotoxicity and a product environ-mental RA: insights from a detergent case study within OMNIITOX. InternationalJournal of LCA 2004;9(5):295–306.
Potting J, Haushild M, Wenzel H. Less is better and only above threshold: two incompat-ible paradigms for human toxicity in life cycle assessment? Int J LCA 1999;4(1):6–24.
Prete F. Sintesi e caratterizzazione di nanopolveri composite allumina-zirconia”, PhD inMaterial Engineer Alma Mater Studiorum Bologna University; 2010.
Reau A, Guizard B, Canel J, Galy J, Tenegal F. Silicon carbide nanopowders: the parametricstudy of synthesis by laser pyrolysis. J Am Ceram Soc 2012;95(1):153–8. [January 2012].
Ribera G, Clarens F, Martínez-Lladó X, Jubany I, Martí V, Rovira M. Document Life cycleand human health risk assessments as tools for decision making in the design andimplementation of nanofiltration in drinking water treatment plants. Sci Total Envi-ron 2014;466-467:377–86. [1 January 2014].
Salieri B, Olsen SI, Righi S. How to calculate the characterisation factor for nanoparticle? Acase study on n-TiO2. In: Reale F, Scalbi S, editors. Life Cycle Assessment eottimizzazione ambientale: esempi applicativi e sviluppi metodologici. VII Convegnodella Rete Italiana LCA. Milan. 27–28 June 2013 978-88-8286-292-3; 2013. p. 330–5.
Scalbi S, Masoni P. Analisi LCA di materiali ultraleggeri per il trasporto: aluminum foamsandwich. RICERCA DI SISTEMA ELETTRICO, Report RdS/2012/094; 2012.
Scanlon KA, Gray GM, Royce AF, Lloyd SM, La Puma P. The work environment disability-adjusted life year for use with life cycle assessment: a methodological approach.Environ Health 2013:12:21.
SCENIHR (Scientific Committee on Emerging, Newly Identified Health Risks). Opinion onthe appropriateness of the risk assessmentmethodology in accordance with the tech-nical guidance documents for new and existing substances for assessing the risks ofnanomaterials. SCENIHR approved this opinion for the public consultation at the17th plenary on 29 March 2007. Washington D.C.: Science Policy Council; 2007
Scheringer M, Vögl T, von Grote J, Capaul B, Schubert R, Hungerbühler K. Scenario-basedrisk assessment of multi-use chemicals: application to solvents. Risk Anal 2001;21:3.
Seager TP, Linkov I. Coupling multicriteria decision analysis and life cycle assessment fornanomaterials. J Ind Ecol 2008;12(3):282–5.
Shatkin JA. Nanotechnology: health and environmental risks. Boca Raton, FL, USA: CRCPress/Taylor & Francis Group; 2008. p. 167.
Shojaie-Bahaabad M, Taheri-Nassaj E. Economical synthesis of nano alumina powderusing an aqueous sol–gel method. Mater Lett 2008;62:3364–6.
Sleeswskij AW. Regional LCA in a global perspective. A basis for spatially differentiatedenvironmental life cycle assessment. Int J Life Cycle Assess 2011;16:106–12.
Socolof ML, Geibig JR. Evaluating human and ecological impacts of a product life cycle: thecomplementary roles of life-cycle assessment and risk assessment. Human Ecol RiskAssess 2006;12:510–27.
SomC,BergesM, ChaudryQ,DusinskaM, Fernandes TT,Olsen SI, et al. The importance of lifecycle concepts for the development of safe nanoproducts. Toxicology 2010;269:160–9.
Sweet L, Strohm B. Nanotechnology—life-cycle risk management. Human Ecol Risk Assess2006;12:528–51.
TGD (Technical Guidance Document for Risk Assessment)Institute for Health andConsumer Protection European Chemicals Bureau. EUR 20418 EN/2; 2003.
UNEP (United Nations Environment Programme). Global guidance principles for life cycleassessment databases. A basis for greener processes and products; 2011 [Reportavailable at //lcinitiative.unep.fr].
van Duuren-Stuurman B, Vink SR, Brouwer DH, Kroese DED, Heussen HGA, Verbist KJM,Tielemans E, Van Niftrik MFJ, Fransman W. Stoffenmanager Nano: description ofthe conceptual control banding model. TNO report 2011(V9216).
van Duuren-Stuurman B, Vink SR, Verbist KJM, Heussen HGA, Brouwer DH, Kroese DED,et al. Stoffenmanager Nano Version 1.0: a Web-based tool for risk prioritization ofairborne manufactured nano objects. Ann Occup Hyg 2012:1–17.
van Tongeren M, Fransman W, Spankie S, Tischer M, Brouwer D, Schinkel J, et al.Advanced REACH Tool: development and application of the substance emissionpotential modifying factor. Ann Occup Hyg 2011;55(9):980–8.
Walser Tobias T, Juraske R, Demou E, Hellweg S. Indoor exposure to toluene from printedmatter matters: complementary views from life cycle assessment and risk assess-ment. Environ Sci Technol 2014;48:689–97.
Wang J, Gerlach JD, SavageN, CobbGP. Necessity and approach to integratednanomateriallegislation and governance. Sci Total Environ 2013;442:56–62. [1 January].
Wardak A, Gorman ME, Swami N, Deshpande S. Identification of risks in the life cycle ofnanotechnology-based products. J Ind Ecol 2008:435–48.
Wright HE, Zhang Q,Mihelcic JR. Integrating economic input–output life cycle assessmentwith risk assessment for a screening-level analysis. Int J Life Cycle Assess 2008;13:412–20.
Zamagni A, Buonamici R, Buttol P, Porta PL, Masoni P. Main R&D lines to improve reliabil-ity, significance and usability of standardised LCA. Technical report; 2008a [availableat www.calcasproject.net].
Zamagni A, Buttol P, Porta PL, Buonamici R, Masoni P, Guinée J, et al. Critical review of thecurrent research needs and limitations related to ISO-LCA practice. ENEA Report;2008b [available at old.enea.it/produzione_scientifica/edizioni_tipo/volumi.html].
Zamagni A, Barberio G, Rigamonti L. What is sustainable technology? The role of life cycle-basedmethods in addressing the challenges of sustainability assessment of technologies.Position paper of proceeding of DIRE meeting, 29/09/2012 978-88-8286-270-1; 2012.
Zuin S, Micheletti C, Critto A, Pojana G, Johnston H, Stone V, et al. Weight of Evidenceapproach for the relative hazard ranking of nanomaterials. Nanotoxicology 2011;5(3):445–58. [September 2011].
