Environmental Impact Assessment of Nanoparticles and Nanoenabled Products Using LCA Frameworks

Transcription

1 Environmental Impact Assessment of Nanoparticles and Nanoenabled Products Using LCA Frameworks A Dissertation Presented By Leila Pourzahedi to The Department of Civil and Environmental Engineering in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the field of Environmental Engineering Northeastern University Boston, Massachusetts May 2016

2 i Abstract Forecasts show rapid expansion of the nanotechnology industry, with a global annual growth rate of more than 20% through this decade. Despite the performance benefits of using engineered nanomaterials (ENMs) for various applications, the environmental and health implications are not yet fully understood. Uncertainties still exist regarding potential nanomaterial emissions, their fate and transport in the environment, the human or environmental exposure and effects, and their relationship to particle specific characteristics. To address the environmental concerns, this research developed and applied life cycle assessment (LCA) models to evaluate ENMs and nanoenabled products. LCA is a multi-criteria, systems analysis tool for quantifying their ecological and human health impacts of different products and processes during their life cycle, from material extraction to end-of-life. Environmental LCA entails coupled emission inventory, fate and transport, exposure, and effect modeling of pollutants, linking nanomaterials and nano-enabled products to various impact categories, including global warming potential, human health effects, and ecotoxicity. Four core nano-lca studies were carried out, including: evaluation of a nanosilver-enabled commercial bandage from its material extraction to disposal, comparison of various nanosilver synthesis routes, investigation of contribution from nanosilver to production environmental impacts of various nanosilver containing products, and quantification of life cycle benefits of using nanotechnology as a substitute for conventional systems for a case of carbon nanotube-enabled electromagnetic interference (EMI) shields for satellites. In contrast to the focus on nanomaterial releases in the literature, nanosilver synthesis impacts were shown to be of significance, driven by the upstream indirect impacts from silver extraction, regardless of the nanoparticle synthesis method. Ecotoxicity impacts of ENM manufacturing were found to be several times those of direct particle release during product use and disposal. The contribution of nanosilver to the overall burdens of nano-enabled product manufacturing was shown to be a function of nanosilver loading, product composition, and particle surface area. Finally, net energy benefits were established for the case of carbon nanotube EMI shielding in terms of savings in weight, fuel use and primary energy demand.

3 ii This work makes meaningful contributions to the field of nano-lca through generation of consistent and comparative life cycle inventory data sets on manufacturing nanoparticles and nano-enabled products, as well as model development through function-based expression of results. Findings of these studies could potentially influence material and process selection, help prioritize of research and development measures including green chemistry efforts, and guide evolving policy discussions on nano labeling and regulation.

4 iii Acknowledgements I would like to thank my advisor, Professor Matthew Eckelman, for his continued guidance and patience. I am very grateful to have had the opportunity to work with you. I would also like to thank Professor Jacqueline Isaacs for her many invaluable contributions in these past years. Special thanks to my committee members, Professor Christopher Bosso and Professor Philip Larese-Casanova, for the insightful discussions and recommendations during my studies. I would like to thank all the members of the SNM group. It was a pleasure to work with you all. My sincerest thanks to all my officemates for all the support they have given me. It was a blessing to work amongst you. Most importantly, my deepest gratitude to my parents, Hossein and Forouzan Pourzahedi, and my brother Ali, for their endless love, support and encouragements in these past years. Words cannot describe how grateful I am for the opportunities you have set before me. This work was supported by National Science Foundation grant SNM through the Nanoscale Science and Engineering Center for High-rate Nanomanufacturing at Northeastern University.

5 iv Table of Contents Chapter Introduction to Nanotechnology Nanotechnology Market, Application and Benefits Environmental Implications of Nanomaterials Introduction to Life Cycle Assessment Life Cycle Impact Assessment Practices US EPA TRACI Impact Assessment Method USEtox Model for Human- and Eco- Toxicity Nano-LCA; Identifying Current Gaps and Challenges Motivation and Research Objectives Dissertation Structure Contribution to Other Research End-of-life Management of CNT-enabled Lithium-Ion Batteries Evaluating Research of ENM MFA for Compliance with Current Nano-Policy Chapter Introduction Silver Nanoparticle Releases from Textiles Life Cycle Modeling of Nanomaterials and Nano-enabled Products Methods and Modeling AgNP Synthesis Bandage Manufacturing and Packaging End of Life Life Cycle Assessment Modeling Uncertainty Analysis Results AgNP Synthesis Bandage Production End of Life Life Cycle Stage Comparisons Discussion and Implications... 43

6 v Chapter Introduction Methods and Modeling Scope and System Boundary AgNP Synthesis Routes Chemical Reduction of Silver Nitrate with Trisodium Citrate (CR-TSC) Chemical Reduction of Silver Nitrate with Sodium Borohydride (CR-SB) Chemical Reduction of Silver Nitrate with Ethylene Glycol (CR-EG) Green Synthesis using Soluble Starch (CR-Starch) Flame Spray Pyrolysis (FSP), Utilizing a Methane-Oxygen Flame Arc Plasma Reactor (AP) Reactive Magnetron Sputtering of Silver with Ar-N 2 Gas Mixture (RMS-AR-N) AgNP Functional Unit Life Cycle Assessment Modeling Comparative Results and Discussion Synthesis Comparison by Mass Syntheses Comparison by Particle Size/Function Analysis of Process Contribution Comparison to Previous Results Impact Assessment in Context of Products Chapter Introduction Methods and Modeling Scope and System Boundary Product Description Purchased Products Products Based on Literature Experimental Procedures Life Cycle Assessment Modeling Results and Discussions Analysis of Process Contributions Analysis of AgNP Release... 79

7 vi Chapter Introduction LCA Concepts for CNTs CNTs for Electromagnetic Interference Shielding Methodology Analysis Results EMI Mass Calculation Fuel Consumption Analysis Total CED for EMI Shielding Options Uncertainty Analysis Discussion Chapter Summary of Chapters Contributions of Study Future Work and Research Direction References Appendix A 115 Appendix B 129 Appendix C 146 Appendix D 157

8 vii List of Figures Figure 1.1. Global flow of ENMs in 2010, Keller et al., (2013) Figure 1.2. Relationship between LCA phases as described by ISO Figure 1.3. Midpoint and endpoint impacts of GHG emissions. Bare, Figure 1.4. ZOI test results for three different AgNP-enabled bandages Figure 2.1. System boundary and life cycle stages Figure 2.2. Nano scale SEM image of the Acticoat 7 bandage Figure 2.3. Process contribution for all TRACI impact categories of RMS-Ar-N synthesis route (silver source is solid silver, gaseous elements are argon and nitrogen). Labels demonstrate the total absolute value of each impact category. Abbreviations: ozone depletion (OD), global warming (GW), photochemical smog (PS), acidification (ACF), eutrophication (EU), human health: carcinogens (HHC), human health: non-carcinogens (HHNC), human health: criteria air pollutants (HHCR), ecotoxicity (EC), fossil fuel depletion (FF) Figure 2.4. Bandage manufacturing process contributions. RMS-AR-N AgNPs were assumed to be incorporated. The labels on bars correlate with the CV of that impact category. Abbreviations: ozone depletion (OD), global warming (GW), photochemical smog (PS), acidification (AC), eutrophication (EU), human health: carcinogens (HHC), human health: non-carcinogens (HHNC), human health: criteria air pollutants (HHCR), ecotoxicity (EC), fossil fuel depletion (FF) Figure 2.5. (a) Process contribution to bandage incineration, case of Acticoat 7, for ecotoxicity and GWP, (b) life cycle stage comparison of Acticoat 7 bandage in terms of ecotoxicity and GWP Figure 3.1. Relative environmental impacts of multiple AgNP synthesis routes (a) TRACI 2.1 life cycle impact assessment method, (b) re-scaled impacts with respect to size-dependent bioactivity. Abbreviations: chemical reduction with trisodium citrate (CR-TSC), chemical reduction with sodium borohydride (CR-SB), chemical reduction with ethylene glycol (CR-EG), chemical reduction with soluble starch (CR-Starch), flame spray pyrolysis (FSP), arc plasma (AP); potential impact categories are ozone depletion (OD), global warming (GW), photochemical smog (PS), acidification (AC), eutrophication (EU), human health: carcinogens (HHC), human health: noncarcinogens (HHNC), human health: criteria air pollutants (HHCR), ecotoxicity (EC), fossil fuel depletion (FF)... 59

9 viii Figure 3.2. (a) Process contribution for all TRACI impact categories of FSP synthesis route (silver source is silver octanoate, reagents are 2-ethylhexanoic acid and xylene, gaseous elements are oxygen and methane). (b) Process contribution for all TRACI impact ca Figure 3.3. Absolute environmental impacts of bandage production shown for GWP and ecotoxicity. Error bars demonstrate the high and low bounds for impacts depending on the production route. Abbreviations: potential impact categories are global warming (GW) and ecotoxicity (EC) Figure 4.1. Process contribution to environmental impacts of different products. Components were categorized generally as paper packaging, nanosilver, extrusion and molding processes for plastics, cellulosic and cotton fibers, and polymers. Concentration of silver is shown in parentheses in percentage to total product. Products are listed in ascending silver concentration from the top left corner with the lowest percentage of AgNPs, to the bottom right with the highest amount of AgNPs. Abbreviations of TRACI s potential impact categories: ozone depletion (OD), global warming (GW), photochemical smog (PS), acidification (AC), eutrophication (EU), human health: carcinogens (HHC), human health: non-carcinogens (HHNC), human health: criteria air pollutants (HHCR), ecotoxicity (EC), fossil fuel depletion (FF) Figure 4.2. Result of different release studies on the products considered in this study. Due to the differences in the experimental methods, they are shown in varying colors. Axes are in log-log format Figure 5.1. (a) Tubular struts, and (b) engine cover components of the Juno spacecraft, both constructed using the CNT M55J/CE laminate. Image courtesy NASA/JPL-Caltech Figure 5.2. (a) Mass distribution of LEO satellites, (b) distribution of fuel consumption for launching the EMI shielding under various shielding scenarios, (c) total CED for Iridium NEXT and ORBCOMM satellites under various shielding scenarios with error bars representing 95% confidence interval from Monte-Carlo simulation on input parameters (CV values can be found in Table 5.2)

10 ix List of Tables Table 1.1. Review of existing Nano-LCA Table 3.1. Synthesis route summary and particle size distribution Table 4.1. Previous literature on silver release from AgNP-enabled commercial consumer products Table 5.1. Current CNT polymer composites technologies and their SE Table 5.2. Various EMI options, mass of each component, the fuel required and cumulative energy demand (CED) for each scenario

11 1 Chapter 1 Introduction to Nanotechnology and Nano-LCA 1.1. Introduction to Nanotechnology Nanotechnology Market, Application and Benefits The properties of nanomaterials are different to that of the bulk material, due to higher specific surface areas, surface reactivity and quantum-related effects. 1 To exploit these properties, manufacturers have begun to incorporate them into products for various applications. Silver nanoparticles (AgNPs) have been used for their remarkable antimicrobial, antifungal and antiviral properties, 1 in consumer products used in homes, 2 targeted for children, 3 or in textile and athleticwear. 4 6 Carbon nanofibers (CNFs) can be used for polymer reinforcement with added benefits of electrical conductivity in applications such as polymer composites to catalytic support materials. 7 Reportedly, carbon nanotubes (CNTs) possess exceptional electrical conductivity and mechanical properties, and are finding increasing application in electronics, such as rechargeable lithium ion batteries, 8,9 and semiconductors. 10 Titanium dioxide nanoparticles, one of the most extensively used nanomaterial, 11,12 are used in coatings, paints, and pigments, as a result of their photocatalytic properties. 12 Forecasts show constant growth in nanotechnology industry, reaching $75.8 billion by the year The Project on Emerging Nanotechnologies by the Woodrow Wilson Center has listed nearly 2000 consumer products containing nanoparticles to date, all readily available in the market. 14 Keller et al. (2013) looked at the global flow of 10 different ENMs for market data of 2010, modeling the amount of each nanomaterial in industry section and environmental compartment. 11 Their collective results showed high levels of nanomaterial consumption for coating applications, followed by the electronics and cosmetic industries. Figure 1.1 depicts this flow of ENMs in metric tons/year, from manufacturing to various end uses and disposal/release scenarios.

12 2 Figure 1.1. Global flow of ENMs in 2010, Keller et al., (2013). 11 The majority of scientific research on nanotechnology has focused on the application of ENMs. 15 Another category of research on this cutting edge technology is the evaluation of the inherent risk of human or environmental exposure to nanomaterials. 16 The current stance of nanotechnology with regards to environmental implications is predominantly hindered by the complications during manufacture. The use of harmful solvents, energy intensive and inefficient synthesis methods requiring further purification steps and generating substantial amounts of waste, are examples of concerns facing nanomanufacturing processes today. 15 Additionally, uncertainties still exist regarding ENM emissions, their fate and transport in the environment, and the human or environmental levels of effect and exposure. 17 Therefore, as acknowledged by Eckelman et al. (2008), 15 and Gilbertson et al. (2015), 17 there is the need to bridge the gap between the studies on potential environmental health and safety implications of nanomaterials, and redesigning nanotechnology for reduced overall impacts Environmental Implications of Nanomaterials To address the environmental concerns of nanomaterials, scientific endeavor has focused on identifying environmental transportation pathways, and evaluating the behavior, fate and

13 3 concentrations within different mediums. Mass flow models have been developed to predict the concentration of nanomaterials from manufacturing and after use to various compartments of the environment. 11,18 21 While some researched particle release at larger scales, others studied product-specific nano-emissions under hypothetical use phase scenarios. 11,18,19,22 Fate, dissolution, residence time, and species change of nanoparticles have been investigated for different environmental mediums: soil, water, and air. 32 With use and disposal of nanoenable products, particles can ultimately reach waste streams. Simulations have been performed to assess the physical and chemical transformations of nanomaterials during waste management practices such as wastewater treatment, incineration, and landfilling. 39 The same features of ENMs that makes them a desirable choice for unique applications, could potentially be a cause for concern during exposure to these particles. 40 Nanotoxicology is another emerging field in this category, with the intention to determine the potential hazards of ENMS and the particle-specific characteristics that drive harmful impacts towards organisms. 16,41 46 Hischier et al. (2012) mention that development in nanotechnology, although beneficial, presents caveats from a sustainability perspective. 47 Despite the increasing body of knowledge on toxicological impacts and behavioral aspects of human and environmental exposure to nanomaterials, implications of nanomanufacturing were found to be generally less well reported. 48 Therefore, in recent years, researchers have recognized the importance of considering potential costs to the environment and human health during every nano-related process, which could potentially offset advantages gained by using ENMs or nano-enabled products. In a study on AgNPs, researchers at the US Environmental Protection Agency (EPA), have acknowledged the need for processes to assess the risks of using ENMs, in addition to considering other aspects such as environmental costs of transporting nanomaterials. 1 They suggested the use of comprehensive assessment tools such as Life Cycle Assessment (LCA) to ensure benefits of using nanomaterials are analyzed in conjunction with the hazards imposed to the consumer and the environment, to avoid any unintended consequences. 1 Reviewing research on nanoscale products, Gavankar et al. (2012) identified the demand for considering potential environmental effects of ENMs throughout their life cycle as many literature have only focused on their development and application. 40 Product-related assessment tools such as LCA have been suggested for a system-level evaluation of nanotechnology, as this framework focuses on consumption and production of goods and

14 4 services. 40,47 LCA is a comprehensive tool that can be applied to different products and processes during their life cycle, to quantify their ecological and human health impacts Introduction to Life Cycle Assessment Environmental impacts are caused by emissions to the environment or the consumption of resources during material extraction, manufacturing of product, its use phase or at end-of-life. 49 Environmental LCA concepts have been introduced over three decades ago, to estimate and quantify the impacts associated with a product or a process, by evaluating the inputs and outputs, from and to the environment, during their life cycle. 50 Early forms of LCA developed in early 1970s tackled the issues of energy consumption, but were then further improved to include analysis of resource efficiency, pollution levels and solid waste generation as well. 50 In the 1990s, efforts were made to standardize the framework of LCA. 50 This can be seen in the works of International Organization for Standardization (ISO) in 1994: ISO and ISO series on Environmental Management Life Cycle Assessment. 51,52 ISO recognized LCA as a tool to identify potential opportunities for reducing environmental impact of products at different steps of their life cycle. 51 It also acknowledges the inability of LCA models to include socio-economic aspects of products. ISO defines LCA in four phases, shown in Figure 1.2, all described in detail below. 1. Goal and Scope Definition is the initial step of performing an LCA. In this step the purpose and underlying reasons of the study should be defined. This step includes technical details of the product under study, the boundary of the system, chosen impact categories, any made assumptions, and most importantly, the functional unit. 51 Functional unit is the foundation of an LCA study, and is defined as the quantification of the service provided by the product under study. 49 Functional unit links inputs and outputs of the model and enables easy comparisons of LCA results by providing a common basis. 51 In this step the boundary and scope of the study is also defined as cradle-to-grave (from material extraction to end-of-life), or cradle-to-gate (partial assessment from material extraction to factory gate, omitting use phase and end-of-life).

15 5 2. Life Cycle Inventory (LCI) Analysis is often the most time consuming step of the LCA. 49 It involves collection and compilation of input and output data, to and from the product system. 51 Material and energy inputs, co-products, wastes, and emissions to air, water and soil, and the relationship between the data, are defined in this step for unit processes (smallest elements of the life cycle inventory with quantifiable inputs and outputs) within the system boundary. 51 Extra attention should be given to processes yielding multiple products to ensure correct allocation of the resources and emissions. 3. Life Cycle Impact Assessment (LCIA) is a process that links inventory data to specific environmental impacts. 51 It requires choosing a set of impact categories fitted to the premises of the study, classification of the inventory data for the chosen impact categories, and quantifying the impact categories using characterization factors. This step is explained in more detail in Section Life Cycle Interpretation is the phase where the results of steps 2 and 3 are analyzed together. This step explains the conclusions, limitations and recommendations based on the results of the LCA modeling. 52 Despite the beneficial information gained regarding the environmental performance of a product systems by using this framework, limitations and unresolved issues still exist at every step of the process. The most important fact to consider while performing an LCA is the inability of this model to predict absolute or precise measures of environmental impacts. 51 This is due to the relativity of potential results to a reference unit, and intrinsic uncertainties of impact models. 51 Assigning correct functional units to the product system under study can become challenging when the product is multi-functional, or the function is difficult to quantify. 53 Incorrect selection of the system boundary can result in truncation errors from cutoffs. 53 In curating the life cycle inventory, allocation issues could be a cause of concern, especially in multi-functional processes or multiproduct systems. ISO generally recommends avoiding allocation of impacts when possible by unit processes division or system expansion procedures, and using physical or other relationships between products when allocation cannot be avoided. 51

16 6 Figure 1.2. Relationship between LCA phases as described by ISO Life Cycle Impact Assessment Practices As mentioned earlier in Section 1.2 LCIA involves the quantification of relative environmental impacts of a product based on its function. It consists of selecting the impact categories, their indicators, assigning them to the inventory data, and calculating the final impact category indicator results with respect to the functional unit. 54 It should be noted that results of LCIA are limited to those defined by the goal and scope. 51 An important decision affecting the final outcomes of an LCIA is the selection of midpoint or damage oriented (endpoint) impact categories. 55 Midpoint impacts refer to the effect of a stressor midpoint within a cause-effect chain. 56 As an example CO 2 emissions can have a midpoint effect of global warming potential and could ultimately damage human health, water resources, agriculture and other species, as shown in Figure Typically uncertainty associated with the endpoint impacts are much greater than midpoint impacts as they are quantifiable and well defined. 55 When selecting impact categories, it is suggested they include the three broad categories of resource consumption, human health impacts and ecological impacts. 54 For a generic situation (not time and location specific), indicators for an impact category can be defined using the equation below 54 Category Indicator = S Characterisation Factor (s) Emission Inventory (s) (1)

17 7 where s stands for the chemical. Emissions data are linked by the characterization factors (CFs) to the indicator of the impact category. Based on equation (1), characterization factors are the contribution of a unit mass of an emission to the impact category, if the emission inventory is expressed on a mass basis per functional unit. 54 For non-generic situations, characterization factors can be defined as shown below in equation (2) for the case of human or ecological toxicity 54 CF = Effect(s,j,t) = ( Fate(s,j,t) j Emissions (s,i) j ) Emission (s,i). ( Exposure(s,j,t) Fate (s,j,t) ). ( Effect(s,j,t) Exposure (s,j,t) ) (2) where s,i, j, and t are defined as the chemical, location of emission, location of exposure, and the time period, respectively. 54 Depending on the impact method, other different factors can be introduced to the equation. For environmental decision making, evaluation of a large number of potential impacts within the mentioned categories is required. 57 This process requires significant amount of time, data, and resources. 57 In 1995, the US Environmental Protection Agency (EPA) began working on developing a simple and accessible impact assessment tool based on the available knowledge and technology in the US, as the common practice was to use European based stressors for chemical potencies developed by the Dutch. 56 The result of this work was the Tool for Reduction and Assessment of Chemical and other Environmental Impacts or TRACI. Section explains this impact assessment method and its categories as it is the cornerstone of US based LCIA. Figure 1.3. Midpoint and endpoint impacts of GHG emissions. Bare,

18 US EPA TRACI Impact Assessment Method TRACI is a software and a tool for quantifying environmental impacts by characterizing environmental stressors that have potential negative effects. EPA developed this tool after realizing the need for sustainability metrics and methodologies applicable to the US conditions. 56 The first step to developing this method was the selection of a comprehensive list of impact categories. 56 Taxonomy studies of potential impacts were conducted by the EPA and refined to include a short list of categories complying with the existing policies followed by the EPA. 57 This list includes both depletion categories and pollution categories. 56 TRACI is a midpoint impact assessment method. 56 This decision was made due to the reduced amount of forecasting and effect modeling required by these methods, resulting in a less complex model that is easily interpreted. 56 A short description of each impact category included in TRACI is given below. 1. Stratospheric Ozone Depletion: Ozone is our main protectant against radiation. In addition to humans, ozone has effects on crops, and other species as well. 57 World Meteorological Organization has calculated the ozone depleting potentials of substances capable of breaking down ozone such as chlorofluorocarbons (CFCs), hydrochlorofluorocarbons (HFCs) and halons. These ozone depleting potentials of chemicals have been reported relative to that of CFC Global Warming Potential: Global warming is the average warming in the temperature of the atmosphere, as a result of increase in greenhouse gas (GHG) emissions such as carbon dioxide, methane and nitrous oxide, from natural or human sources. 56,57 TRACI uses global warming potential indices proposed by the International Panel on Climate Change (IPCC), based on the potency of GHGs relative to that of CO 2, for a 100 tear time frame Acidification: This impact is caused by the increase in the concentration of H + in aquatic environments. Acidification could be a result of addition of acids, chemical or biological reactions, or deposition of air emissions onto aquatic media. 57 Sulfur dioxide (SO 2) and nitrogen oxide (NO x) emissions are the main contributors to acidification of aquatic environments. TRACI quantifies the potential increase in H + relative to the impact of SO Eutrophication: Increase in the concentration of nutrients such as nitrates and phosphates are the main contributors to biological productivity and subsequently

19 9 eutrophication. 57 This could be an unintended consequence of fertilizer used for agricultural lands, depending on their topography and soil type. 56 Eutrophic impacts of chemicals are quantified relative to that of direct nitrogen (N) release to an aquatic system. 5. Human Health Criteria Air Pollutants: This category quantifies the negative impacts of particulate matter emissions to human health. 57 These emissions could be classified into direct emissions such as those from combustion or dust (inhalable coarse particles with micrometer diameters, or fine particles with diameters less than 2.5 micrometers), or they could be secondary particulates, byproduct of chemical reactions of precursors such as SO 2 or NO x. 57 Exposure to these particulate matters were characterized with reference to PM Human Health Cancer, Noncancer, and Ecotoxicity: In 2011, after the introduction of TRACI 2.0, the basis for this impact category changed. Previously, TRACI used a multimedia fate and exposure pathway model named CalTOX to determine its human toxicity potentials (HTP). Similar to HTP, ecological toxicity potential (ETP) was developed to quantify the ecological harm caused by a unit release of chemical to the environment, by using a modified version of CalTOX in conjunction with chemical and landscape datasets. 56 Current versions of TRACI comply with the Life Cycle Initiative of United Nations Environment Program and Society of Environmental Toxicology and Chemistry (UNEP-SETAC) by using the developed global consensus model USEtox. This model is described in Section below. The impacts are expressed in terms of Comparative Toxic Units (CTU) or affected cases per kg emissions. 7. Photochemical Smog Formation: This impact category relates to the formation of ozone molecules in the troposphere. 56 This could be caused by increase in the concentration of NO x, volatile organic compounds (VOCs), carbon monoxide, methane, temperature, and radiation. 56 TRACI uses the Maximum Incremental Reactivity (MIR), a model developed to quantify the human and environmental effect of smog specifically for the US. 57 MIR database consists of 1200 polluting substances Fossil Fuel Depletion: This metric was quantified in TRACI by taking account the surplus energy required to produce fossil fuel in the future (assuming fixed technology) when the economically extractable portion is depleted. 56 This estimation is not restricted to a specific site.

20 10 On-going research for TRACI is being conducted to further develop the resource depletion category, to include land use change and water consumption. 56,57 Complication of these two impact categories rise from the fact that they are extremely location specific. 56 These set of categories were acknowledged to be a minimal list, subject to further expansion in the future. 57 TRACI was developed with the intention to quantify impacts and aid decision making in sustainability, LCIA, emission control and process design USEtox Model for Human- and Eco- Toxicity This model calculates midpoint characterization factors for human toxicity and freshwater ecotoxicity for use in LCIA. 58 USEtox database includes 3000 organic and inorganic substances. 57 It uses multimedia mass balance concepts to determine the concentration of chemicals released to the environment s various compartments. 58 The model uses a matrix based algebra to link environmental exposure, fate and effects. 59 Characterization factors (in cases/kg emission) were calculated using the equation (3) below 59 CF = EF. XF. FF (3) where EF is the effect factor for human toxicity (in cases/kg intake) or ecotoxicity (in Potentially Affected Fraction (PAF).m 3 /kg), XF corresponds to exposure factor for human toxicity only (in day - 1 ), and FF denotes the fate factor (in day). Fate factor is equal to the residence time of a chemical in a specific compartment of the environment, which is dependent on the type of chemical and the properties of the compartment. 60 USEtox also accounts for inter-media transport and removal of the chemicals. 59 Exposure factor for environmental ecotoxicity is determined by calculating the dissolved fraction of chemical in freshwater, and for humans is defined as the rate at which the chemical is transferred to human body from the compartments through multiple exposure pathways (e.g. air, water, food, etc.). 60 For ecotoxicological impacts, the effect factor is determined using the linear slope of dose-response curves of the chemical up to the point where 50% of the species are affected. 60 Therefore, the effect factor for ecotoxicity can be calculated as EF = 0.5 HC50, where HC50 is the hazardous concentration at which 50% of the species are affected above their EC50 (concentration at which 50% of the species population are affected). 60 For human toxicity, with the assumption of linear dose-response relationship, effect factor is calculated as EF =

21 ED50, where ED50 is the lifetime daily dose resulting in 50% probability of effect, and is determined for both carcinogenic and non-carcinogenic impacts through inhalation and ingestion. 59 Defining the previous factors and using equation (3), characterization factors for both ecotoxicity potential and human toxicity potential can be determined. These characterization factors are expressed in Comparative toxic Units (CTU). For human toxicity, the characterization factor unit is CTU h which provides estimates for increase in human population morbidity for a unit mass of chemical emissions (cases/kg). 59 For ecotoxicity, characterization factor unit is CTU e, and it estimates the PAF of species over time per unit mass of chemical emissions (PAF.m 3.day/kg). 59 These characterization factors are used as is in TRACI impact categories of human health carcinogenic and non-carcinogenic, and ecotoxicity. It should be noted that for metals, ionizing compounds and amphiphilics, the USEtox characterization factors are defined as interim. 58 It has been recommended to use these factors with extreme caution due to their high level of uncertainty and insufficient information on their fate and exposure, but avoiding them would mean no impacts from the emission of the material. 57, Nano-LCA; Identifying Current Gaps and Challenges It has been established that, despite their benefits, environmental performance of ENMs should be evaluated, as they could impose negative ecological and human health impacts. These impacts can occur during all the life cycle stages of ENMs, from emissions during their synthesis, to their release during use and end-of-life. For these reasons, it is imperative to assess the environmental impact associated with the commercialization of nanotechnologies, to avoid any unintended consequence. 61 Being in the developmental phase, early application of LCA can affect policy and decision making aspects of adopting this technology. 8,61 Applying LCA to emerging technologies is a challenging task, imposing high levels of uncertainties and complexities from lack of representative data. 61 Lack of data availability creates difficulties in developing LCIs at every step of the life cycle. Data on nano-synthesis processes could be based on lab-scale experiments, as information on commercial-scale manufacturing is mostly confidential. These methods are highly inefficient in terms of material use and particle yield, and do not monitor all inputs and emissions (such as electricity consumption, use of catalyst, or CO 2

22 12 release). As a result, LCA modeling based on lab-scale data could potentially be an inaccurate portrayal of the industrial-scale technique, over- or under-estimating environmental impacts due to the uncertainties. Behavior of nano-enabled products during use could be different to the conventional product. 61 This would mean the assumption of similar use phase scenarios has to be verified. For example, in a life cycle comparison of a conventional and AgNP-enabled T-shirt, 62 Walser et al. (2011) accounted for less washing of the AgNP-enabled T-shirt. For other applications where the benefits of using nano-enabled products are not so easily quantifiable, adjustment of use phase material and energy inputs could prove to be more challenging. This could in part be accommodated through correctly defining the functional unit of the study. This aspect of LCA could also be affected by the inherent lack of data. Upadhyayula et al. (2012) explain the difficulty of functional unit selection in such situations, especially when the boundaries are cradle-to-grave, with the example of comparing automobiles: vehicle miles traveled could be used for vehicle comparison, however, accounting for CNT-enabled batteries is more difficult due to the unavailability of performance data. 61 As an important procedural step, definition of functional unit in existing LCA of ENMs was analyzed through a thorough literature review, for nanoparticles and nano-enabled products. Table 1.1 describes these studies, the impact categories they considered, and the scope of their analysis, in addition to their decision on functional unit. Analyzing the components of these studies showed that only less than half of the LCAs on nano-enabled products considered a functional unit based on the product performance or service, while all the LCAs on the nanoparticles themselves were based solely on mass. Depending on their application, nanomaterials feature specific enhanced functions not attainable using conventional materials, therefore, performing nano-lca solely on the basis of weight for these cases would not be sufficient. 47 Twenty out of 29 LCAs of ENMs mentioned in Table 1.1, focus on cradle-to-gate, therefore omitting possible impacts from end-of-life. Hischier and Walser (2012) mention six studies out of 17 they reviewed on nano-lca which consider incineration in their scope, but opt to use incineration models for traditional materials, failing to account for the behavior and fate specific to ENMs. 47 Environmental models have predicted, even though small, a percentage of nanoparticulate release from the filtration system incineration plans, that is not accounted for by

23 13 modeling the traditional or bulk material. 18,63 This release to the environment, in addition to nanoemissions during use bring up the difficulties regarding LCIA of ENMs. Another issue facing nano- LCA is the lack of nano-specific impact characterization factors. Behavior of material at nanoscale greatly differs to that of the bulk material. 40 These factors are highly variable and impose significant levels of uncertainty to LCIA results, as they are dependent on intrinsic factors such as shape, size, capping agent, and functionality, and extrinsic factors such as the composition of medium. 40 As explained in Section 1.2.3, human and ecotoxic characterization factors can be derived by quantifying effect, fate and exposure factors for materials. Ecotoxicity characterization factors for nanomaterials have so far been attempted for CNTs, 64 and nano-tio For calculating effect factor, knowing PAF of aquatic organisms and the effective concentration factors are required. Due to the lack of consensus among toxicological studies regarding their decision on type and characterization of nanomaterials and response metrics, a wide range of results are to be expected, as eluded to by Eckelman et al. (2012) and Salieri et al. (2015), affecting correct and confident estimation of this factor. Fate factor includes the fate and transport kinetics of materials in the environment. For nanomaterials, partitioning, aggregation and transport parameters should be well defined to achieve representable characterization factors. For the case of CNTs difficulties arose due to their colloidal nature, controlling their behavior through aggregation, filtration, deposition, straining and settling kinetics, as opposed to equilibrium partitioning behavior of simply small molecules. 64 Such components are not well defined for all nanomaterials. Fate factor also requires site-specific information such as ph, organic matter levels, or ionic strength, which could potentially affect the kinetics of nanomaterial behavior. 64 Other nanomaterial aspects that affect the fate factor, as also seen in both case studies of CNT and nano- TiO 2, are the variation in size, shape, functionalization or coating. 64,65 There is no set standard defining the relationship between these variables and the behavior of nanomaterials in aquatic environments. Finally, calculating exposure factors requires information on dissolved fraction of the material. USEtox exposure factor is derived for organic chemicals and its partitioning coefficients were shown to be invalid for ENMs with low solubility and high surface reactivity. 65 For an insoluble nanomaterial such as nano-tio 2, estimating such factor was shown to be challenging, resorting to eliminating this factor altogether from the characterization factor calculation, by setting it equal to unity. 65

24 14 Table 1.1. Review of existing Nano-LCA. Citation Greijer et al., 2001 Lloyd and Lave, 2003 Lloyd et al., 2005 Environmental Impact Assessment Method Eco-Indicator 95, ET long, EPS, ECO Sweden, ECO Netherlands and EDIP (weighing methods to determine the most significant environmental aspect). Air related emissions. Hybrid LCA CO2 reduction analysis + cost, materials, fuel inputs, emissions, GHGs, toxic releases, RCRA hazardous waste, etc.) Hybrid LCA CO2 eq, resource use, hazardous waste, toxic release, energy use Product Nano-enabled Products LCA ENM Type and Functional Unit Comparative? Life Cycle Stages Characteristics Solar cell Nanocrystalline dye 1 kwh electricity output from the solar cell system Polymer (polypropylene) composites Automotive catalysts Nano-clay Nanoscale platinum-group metal (PGM) particles Calculation for body panels production of one year s fleet of lightduty vehicle in the U.S. (16.9 million vehicles), 210 million vehicles on the road, miles and 10 year life span PGM required for U.S. vehicle fleet Yes, (1) between solar cells and natural gas power plant [g CO2/kWh], (2) between cells with different efficiencies [g CO2/kWh and g CO2/m 2 ] Yes, substituting nanocomposite s or aluminum instead of steal in body panels [16.9 million vehicles] Yes, current technology vs. nanotechnology [%changes in life cycle effects for reaching new emission standards] Cradle-to-grave Cradle-to-grave Cradle-to-grave

25 15 Table 1.1. Continued Citation Roes et al Bauer et al., 2008 Environmental Impact Assessment Method NREU, climate change, abiotic depletion, ozone layer depletion, photochem oxidant form, acidification, eutrophication, cost CED, aquatic ecotoxicity, ionizing radiation, photochem oxidant formation, ozone depletion, human toxicity, climate change, eutrophication, acidification, resource depletions Product Polymer Nanocomposites: (1) thin film for packaging, (2) thick film for agricultural use, (3) injectionmolded panels used in cars Field emission display Nano-enabled Products LCA ENM Type and Characteristics Nano-clay (silicate) Functional Unit Comparative? Life Cycle Stages (1) Amount of packaging film for 1000 bags for 200 g candies, (2) amount of fill to cover a standard greenhouse of 650 m3, (3) body panels of a low weight car that runs 150,000 km in its lifetime Yes, conventional ((1) pure PP, (2) PE, (3) PP-glass fiber composite) material vs. polymer nanocomposite s for said functional units CNT 15 screen Yes, FED, CRT and LCD comparison Cradle-to-grave Cradle-to-grave

26 16 Table 1.1. Continued Citation Joshi, 2008 Krishnan et al., 2008 Environmental Impact Assessment Method Non-renewable energy use, GHG emission Hybrid LCA GHG emissions, energy use Product Nanoclay biopolymer: organically modified montmorillonite (OMMT) composites Nano-scale CMOS microprocessor Dahlben et al., 2009 Eco-Indicator 1999 (1) CNT switch, (2) CNT polymer mesh Nano-enabled Products LCA ENM Type and Functional Unit Comparative? Life Cycle Stages Characteristics Nanoclay Per kg composite Yes, various polymer composites Cradle-to-gate - One wafer No Manufacturing SWNT (1) one device, (2) unit area of mesh No Cradle-to-gate (manufacture to factory gate) Khanna and Bakshi, 2009 Life cycle energetic impact nonrenewable energy consumption (1) Polymer composites (carbon nanofiber and hybrid carbon nanofiber and glass fiber reinforced) (2) Automotive body panels Carbon nanofiber (1) Equal stiffness (compared to 48x96x in 3 steel plate) (2) lifetime of a midsize 3300 lbs car with 150,000 vehicle miles traveled Yes, PNC compared with steel for said functional units Cradle-to-gate

27 17 Table 1.1. Continued Citation Environmental Impact Assessment Method Hassan, 2010 Hybrid LCA BEES + EIO Merugula et al., 2010 Meyer et al., 2010 Sengul and Theis, 2010 Walser et al., 2011 CED Hybrid LCA TRACI 2 CED, GWP, aquatic acidification, heavy metal emission Climate footprint, sea & freshwater toxicity, NREU Product Nano-enabled Products LCA ENM Type and Functional Unit Comparative? Life Cycle Stages Characteristics Photocatalyst coatings for concrete pavement Glass fibre/epoxy Carbon nanofiber matrix for reinforcing large wind turbine blades Socks Nanosilver: nm Nanophotovoltaic: Quantum dot photovoltaic module Nano TiO2 1 lane km No Cradle-to-grave Cadmium selenide QDPV T-shirt Nanosilver (FSP: 1-2 nm, nano silvertricalciumphosphate (nanoag-tcp): nm, PlasPu: 2-10 nm elliptical per kwh electricity generated One pair of socks Per square meter of PV + Per kwh energy output Being dressed with the polyester T-shirt for outdoor activities once a week during one year in Switzerland Yes, blades with various CNF content Yes, conventional socks vs. nanosilver socks with three production routes Yes, between different photovoltaics and other renewable and non-renewable energy sources Yes, conventional vs. nano-enabled t- shirt + pure nanosilver and nanoag-tcp Cradle-to-gate Cradle-to-gate Cradle-to-gate Cradle-to-grave

28 18 Table 1.1. Continued Citation Wender et al., 2011 Babaizadeh and Hassan, 2012 Manda et al., 2012 Environmental Impact Assessment Method Life cycle energy tradeoffs BEES + cost CED (REU & NREU), GHG emissions (IPCC 2007 GEP 100a), ReCiPe Product Lithium-ion battery TiO2 coated glass for residential windows Paper with nano TiO2 coating Nano-enabled Products LCA ENM Type and Functional Unit Comparative? Life Cycle Stages Characteristics SWCNT Nano TiO2 Nano TiO2 Energy storage capacity (kwh) One square meter of glass, 40 years lifetime 1 ton of printing and writing paper No Yes, Conventional glass vs. coated Yes, Conventional paper vs. various NP paper configuration Cradle-to-gate Cradle-to-grave Cradle-to-grave Merugula et al., 2012 EROI, ozone depletion, GWP, marine aquatic toxicity, human toxicity, fresh water ecotoxicity, terrestrial ecotoxicity Glass fibre/epoxy matrix for reinforcing large wind turbine blades Carbon nanofiber Per kwh Yes, various CNF loadings and blade mass reductions Cradle-to-gate Weil et al., 2012 LCA (CED) Automotive MWCNT, SWCNT Per kg ENM Yes Cradle-to-gate supercapacitor Dahlben et al., 2013 TRACI 2 CMOS transistor SWNT Conductivity No Cradle-to-gate (extended to include use phase and endof-life) Gilbertson et al., 2014 TRACI 2, ReCiPe H2S chemical sensor SWNT 1 chip No Cradle-to-gate

29 19 Table 1.1. Continued Citation Environmental Impact Assessment Method Product Nanomaterial LCA ENM Type and Characteristics Functional Unit Comparative? Life Cycle Stages Isaacs et al., 2006 EPS SWNT 1 g SWNT Yes, between arc ablation, chemical vapor deposition and HiPco processes Osterwalder et al., 2006 Khanna et al., 2007 Healy et al., 2008 CED - Oxide nanoparticles (TiO2, ZrO2) Ti: sulfate and chloride + octanoate, pentanoate and isopropoxide ; Zr: chloride CML & Eco-Indicator 99 energy use, GWP, marine ecotoxicity, human health toxicity, damage to ecosystem Climate change, acidification, land use, mineral depletion, ecotoxicity, carcinogens, respiratory inorganics Per kg product Yes, between wet chemistry and dry processes - Carbon nanofibres Per kg product Yes, between nanofibres, aluminum, steel and polypropylene - SWNT Per gram SWNT Yes, between arc ablation, chemical vapor deposition and HiPco processes Manufacturing Cradle-to-gate Cradle-to-gate Manufacturing

30 20 Table 1.1. Continued Citation Kushnir and Sanden, 2008 Grubb and Bakshi, 2010 LeCorre et al., 2012 Environmental Impact Assessment Method Product Nanomaterial LCA ENM Type and Characteristics CED - Carbon nanoparticle (CNP) Functional Unit Comparative? Life Cycle Stages Per kg Yes, between various CNP production methods and CNP vs. aluminum production Eco-Indicator 99 - Nano TiO2 Per kg product Yes, between TiO2, steel, aluminum and polypropylene, polysilicon and CNF TRACI 2 and Eco- Indicator 99 - Starch nanocrystals, organically modified nanoclay Pati et al., 2013 CED, GWP - Gold nanoparticles: size range nm Per kg product Per kg NPs Yes, between starch nanocrystals ad organically modified nanoclay Yes, between various green synthesis routes Cradle-to-gate Cradle-to-gate Manufacturing Cradle-to-gate

31 Motivation and Research Objectives With extensive use of ENMs, it is expected to see an increase in the production rate of these particles and their presence in the environment. This can lead to significantly higher environmental impact from this sector compared to the past. However, there is a lack of casespecific data on the impacts associated with the production and release of various ENMs. Moreover, as eluded to in Section 1.3, there is no consensus yet, with regards to quantifying the characterization factors of ENMs, which is a key step in linking the potential ENM emissions to specific environmental impact categories. A study performed by Eckelman et al. (2012), on potential impacts of CNTs, attempted to account for these stated gaps. 64 They developed characterization factors for ecotoxicity impacts of CNTs using the USEtox model, and compared the indirect non-nano impacts from upstream manufacturing using multiple routes, to direct ecotoxic nano impacts. 64 Their results suggested that the ecotoxicity impacts from nanomaterial production were roughly equivalent to the direct impacts of their hypothetical worst case release scenario, and almost three folds greater than their realistic case. 64 Another motivational LCA study was performed by Dahlben et al. (2013) on cradle-to-grave environmental impact of a CNTenabled semiconductor device, to assess the effect of addition of CNTs on the environmental impacts profile of the product. 10 Their results showed that despite the findings of the previous study, where high environmental impact were associated with the production of CNTs, their impacts in the context of the product were insignificant due to their low concentration. 10 These studies sparked the questions that motivated this research: 1. What are the contributing factors to the life cycle environmental impacts of other nanomaterial manufacturing processes? Do said factors comply with the findings of Eckelman et al. (2012) for the LCA case study on CNTs? What are the factors affecting the life cycle environmental impacts of producing nanoenabled products? Besides nanomaterial loading percentage, what other processes can potentially influence or drive the overall environmental burdens of the product? Do results comply with findings of Dahlben et al. (2013)? How can the environmental impacts of nanomaterials and nano-enabled products be defined in terms of their performances? What are some suggestions of functional units? 4. How can life cycle benefits of substituting conventional methods with nanotechnology be quantified?

32 22 The studies in the dissertation at hand were motivated by the aforementioned articles, to expand them and fit them to other less researched ENMs, such as AgNPs. Production of AgNPs is predicted to grow four folds by 2022, finding more applications in healthcare and textiles. 66 It is important to evaluate the environmental burdens imposed by producing and using AgNPs while still in the developing stages. To address the questions above, research objectives below were defined: 1. Comparing environmental impacts of producing AgNPs through multiple routes using correct LCA notation in terms of functional unit. 2. Evaluating the cradle-to-grave life cycle environmental impacts of an AgNP-enabled product. 3. Comparing indirect impacts of nano-enabled products with the impacts of direct particle release for the case of AgNPs. 4. Completing previous LCA on CNTs by determining life cycle benefits of products on the basis of their function Dissertation Structure The work shown in this dissertation falls into the scopes of the National Science Foundation grant on Designing and Integrating LCA Methods for Nanomanufacturing Scale-up. This Scalable Nanomanufacturing (SNM) research uses LCA methodology to address the ethical, legal and societal issues of commercializing nanotechnology. This research is important while nanotechnology is in the developmental phase, as it could affect current policies and incorporate sustainability metrics in decision-making. Chapter 2 describes a study on cradle-to-grave environmental impacts of AgNPs in the context of a commercial nano-enabled product. By utilizing LCA methodology, the magnitude of primary sources of environmental impacts of an AgNP-enabled wound dressing can be quantified. This study determined the contributing processes to AgNP synthesis (specific to the wound dressing), evaluated the contribution of AgNPs to the overall impacts of the medical bandage, and compared the environmental impacts of various life cycle stages of the product. With the prevailing use of AgNPs in the medical field, it is imperative to understand their unintended consequences beforehand. This is the first LCA study performed for nanomaterials in healthcare settings.

33 23 Findings of this study emphasized the importance of including indirect non-nanosilver emissions when considering the environmental performance of nano-enabled products. Modeling the life cycle impacts of this AgNP synthesis demonstrated the domination of production impacts by the resource intensive and highly polluting silver mining processes upstream. The significance of these impacts affected the contribution of AgNPs, making them accountable for more than 50% of the overall production impacts of the product despite the low concentration of AgNPs (almost 10%). Another interesting finding of this study was that the indirect impacts of upstream processes greatly outweigh those from its disposal, including nanosilver releases to the environment. This study was published in 2014, 67 and was the foundation of other works on nano-lca, all featured in this dissertation. Chapter 3 describes a study in continuation of the work shown in Chapter 2. Recognizing the significant environmental damage caused by AgNP production formed the idea of analyzing the potential range of impacts from varying AgNP synthesis routes. Multiple routes were assessed for their environmental performance utilizing cradle-to-gate LCA concepts. The routes were chosen to represent the current trends in AgNP production, and included physical, chemical and biobased methods. Findings of this study complied with the work presented in Chapter 1; across all synthesis routes, impacts from upstream processing of silver dominated almost all impact categories, contributing to over 90% in some cases. Electricity intensive methods such as flame spray pyrolysis showed to have higher impact levels compared to chemical reduction routes, on a mass basis (production of 1 kg AgNPs). For bio-based synthesis route, important tradeoffs were seen in ozone depletion potential and ecotoxicity, caused by fertilizer and pesticide runoff from crop cultivation processes. In addition to comparison by mass, the results were also compared on the basis of function, as an important element to LCA modeling. In this sense, this study is unique, as no studies prior to this have defined functional units for nanomaterials themselves. The function of the AgNP being antimicrobial activity, changed the rank order of impact levels from synthesis routes, as particles with smaller diameter are considered to be more potent. These multiple synthesis scenarios were applied to the case of AgNP-enabled wound dressing to evaluate the life cycle environmental impacts of production methods in the context of a product, showing that overall environmental burdens are highly sensitive to the synthesis route by which the AgNPs are produced. This study was published in

34 24 Chapter 4 describes cradle-to-gate LCAs of various products. It was seen from the LCA of AgNPenabled bandage that impacts of nanoparticles dominated the overall environmental burdens of the product. Dahlben et al. (2013) had found the opposite for their CNT-enabled semiconductor, an insignificant contribution form CNT manufacturing, as a result of the minimal concentration of nanoparticles on the device. 69 Therefore, the contribution of ENMs to the overall impacts of nanoenabled products are not only dependent on the production route, but to the ENM loading on the product as well. Having done the study on AgNP-enabled bandage and access to a database of AgNP synthesis requirements, created an opportunity to model the life cycle environmental impacts of an array of AgNP-enabled consumer products, to assess the effect of AgNP concentration on its contribution to the total life cycle burdens of production. Cradle-to-gate impacts of production were compared to direct impacts from release during use of the product. The contribution of silver was found to rely not only on its concentration, but to the composition of the product and the amount of other components as well. This resulted in a wide range of 1% to 99% AgNP contribution to impact category, depending on the product. Analyzing the release studies showed higher impacts from upstream production processes compared to direct impacts of AgNP release, in terms of ecotoxic potentials. Compiling release experiments showed solid polymeric samples with surface coating lost more particles during use, compared to textile samples with particles embedded in the fiber, emphasizing the importance of particle incorporation method. To compensate for the lack of product-specific AgNP synthesis information, results of the previous study on AgNP manufacturing were used to show the potential range of impacts. The list of products assessed contained several products in the same category, such as wound dressings, food containers, clothing, and products used for children. To compare the environmental performance of products in the same category, such as medical bandages, it is important to define a basis of function. Previously we had proposed the use of size-dependent toxicity of AgNPs to aquatic species as a basis of comparison among particles synthesized through various methods. For an application such as wound dressing where AgNPs are enabled in the product to serve as an antimicrobial agent, their efficacy towards bacteria can be used to quantify their performance. The Kirby-Bauer antibiotic testing or disc diffusion antibiotic sensitivity testing, was used to determine the level of efficacy of different nano-enabled bandages towards gram-

35 25 negative bacteria. Three sample sets of 6 mm diameter discs from three acquired AgNP-enabled bandages were placed on agar plates loaded with Escherichia-coli bacteria and incubated at 37 C for 3 days to simulate use conditions. If the samples contain antibiotic agents, there will be an area around the samples with diffused antibiotics where no bacteria will grow, called the zone of inhibition (ZOI). The level of antimicrobial efficacy of the products were quantified by measuring the diameter of the ZOI. Results of this tests can be seen in Figure 1.4. Outcomes of the product LCAs can be rescaled on the basis of ZOI for these bandages, to determine the relative levels of environmental impacts of a functional unit basis. This proposition for functional unit is the first to quantify impacts of antimicrobial products for the purpose of LCA. This is a detrimental component of comparative LCA where comparing product-based impacts is not acceptable. These experiments were carried out in the laboratory as part of research on nanomaterial characteristics and function. The results are not yet published. Figure 1.4. ZOI test results for three different AgNP-enabled bandages. Chapter 5 describes a study on functional unit based comparative LCA. Functional unit is not only a means to compare environmental impacts, it can also be a basis for quantifying the benefits of nano-enabled products during use, compared to conventional methods. Incorporating AgNPs in products in many cases would just provide additional antimicrobial function to a product. For a scenario where including nanoparticles would improve upon the current functional state of the product, an extreme application for CNTs was considered to lay an outline for future assessment of ENMs and applications. This study quantifies the life cycle energy benefits of using CNT composites for electromagnetic interference (EMI) shielding purposes. The only similar study for CNTs is by Zhai et al. (2015), which quantifies benefit for using CNTs in reinforced cement, flash memory switches, and lithium ion batteries. They concluded that life cycle net energy benefits

36 26 from CNT-enabled products are highly dependent on the application. 70 This illustrates the importance of evaluating life cycle ecological impacts and energy use of a technology prior to its adoption. While CNT composites are costly and energy-intensive to produce relative to existing EMI shielding materials, they may offer both economic and energy benefits when considered on a life cycle basis. When comparing nano-enabled products with the current technology, it is imperative to account for both the differences in production and performance. This study compares CNT composites as a replacement for aluminum as an EMI shielding component of satellites on the basis of their function, shielding effectiveness (in db). The findings of this study showed that, CNT composites with the performance levels similar to aluminum could be developed to provide more than 50% reduction of shielding mass, which is an important factor when launching objects into orbit. This reduction in EMI mass leads to lower levels of fuel consumption during launch, which subsequently affects the life cycle primary energy demand. Results showed nearly 30% reduction of cumulative energy demand for CNT composites compared to aluminum, which consequently affects the ultimate environmental impact levels. As the final chapter, Chapter 6 presents a critical summary of all the research completed in this dissertation and explains potential future work to follow these studies Contribution to Other Research This section explains the contributions of this author to two studies that are not included in the dissertation at hand, yet in line with the overall scheme of research on nano-lca End-of-life Management of CNT-enabled Lithium-Ion Batteries To this point, production and use phase stages of the LCA of nanoparticles and nano-enabled products have been covered. For the case of new technologies such as nanotechnology, their emergence could start to outpace the rate at which their potential environmental impacts are assessed at end-of-life. 9 For the case of AgNPs, extensive literature is available for their fate in the environment at end-of-life. The case of AgNP-enabled bandage also provides a scenario that includes environmental impacts of particle release from incineration plants. In an effort to predict the concentration of nanomaterials in the environment by adopting these new technologies, this author added a waste management analysis to a study on material flow analysis (MFA) of CNT-

37 27 enabled lithium ion batteries for portable computers, by Espinoza et al. (2014). 9 Amount of CNTs generated as waste from product obsolescence was modeled to assist in determining the fate of these nanomaterials at end-of-life. This work was shown as a section of the mentioned study. For the case of US, state level information on electronic waste collection, recycling, incineration, and landfilling were gathered. To date, only 25 states have regulation on battery recycling. A baseline scenario with current rate and an optimistic scenario of a nationwide recycling rate of 85%, a target set by the Waste Electrical and Electronic Equipment (WEEE) directive in the EU, were both modeled. The results suggested almost 4000 tons of discarded CNTs from this application over the 25 years time period of this study. Of this amount, for the current state of recycling in the US, 8% were projected to be incinerated, and almost 80% landfilled, the rest assumed to be collected. By reaching an 85% recycling rate nationwide, the landfilled amount over the 25 years could be reduced by 60%. These projections can help in decision making for future planning on recycling facilities or development of extended producer responsibility programs, to ensure proper control and handling of CNT emissions. Further information regarding this analysis can be found in the publication by Espinoza et al. (2014) in a special issue of ACS Sustainable Chemistry & Engineering Evaluating Research of ENM MFA for Compliance with Current Nano-Policy LCA is inherently a tool developed to aid in decision making. In 2011, the National Nanotechnology Initiative (NNI), called for the integration of risk assessment (RA) and LCA concepts into decision making on environmental health and safety (EHS) issues of nanotechnology. 71 A study by Walker et al. (2015), reviewed the LCA studies on nanotechnology from , to evaluate their compliance with the NNI priorities. This study selected 31 LCAs on nanotechnology or nanoenabled products and assessed whether the articles include any RA methods or mention/follow any mandate on nanotechnology. As an addition to this list of LCA studies a second list on MFA, fate and transport studies on nanomaterials was prepared by this author for a more comprehensive outlook on this matter. Mass flow studies at spatial levels are often carried out on a life cycle basis and predict the environmental concentration of nanomaterial in each media. 71 Therefore, their inclusion in this analysis are of importance. This work can be found as an added section to the aforementioned study. The list contained 20 articles, and showed increasing sophistication with the progression of these types of studies. Despite the lack of RA measures in the studies, its importance was acknowledged, with some considering these fate ad transport

38 28 models fit for regulatory purposes. Although sustainable development of nanotechnologies require both LCA and RA models, they were found to be separate entities to this point in time. 71 Details of this review can be found in the work published by Walker et al. (2015) in the Journal of Nanoparticle Research. 71 LCA offers insights on how to reduce the environmental damage of nanotechnology by identifying the hotspots at each life cycle stage for possible improvements, deeming it a great candidate to assist in decision making on nanotechnology as it advances to commercialized levels.

39 29 Chapter 2 Environmental Life Cycle Assessment of Nanosilver-Enabled Bandages This study has been published Pourzahedi, L., & Eckelman, M. J. (2014). Environmental life cycle assessment of nanosilver-enabled bandages. Environmental science & technology, 49(1), Over 400 tons of silver nanoparticles (AgNPs) are produced annually, 30% of which are used in medical applications due to their antibacterial properties. The widespread use of AgNPs has implications over the entire life cycle of medical products, from production to disposal, including but not limited to environmental releases of nanomaterials themselves. Here a cradle-to-grave life cycle assessment (LCA) from nanoparticle synthesis to end-of-life incineration was performed for a commercially available nanosilver-enabled medical bandage. Emissions were linked to multiple categories of environmental impacts, making primary use of the TRACI 2.1 impact assessment method, with specific consideration of nanosilver releases relative to all other (nonnanosilver) emissions. Modeling results suggest that (1) environmental impacts of AgNP synthesis are dominated by upstream electricity production, with the exception of life cycle ecotoxicity where the largest contributor is mining wastes, (2) AgNPs are the largest contributor to impacts of the bandage for all impact categories considered despite low AgNP loading, and (3) impacts of bandage production are several times those bandage incineration, including nanosilver releases to the environment. These results can be used to prioritize research and policy measures in order to improve the overall ecotoxicity burdens of nano-enabled products under a life cycle framework.

40 Introduction For hundreds of years, silver has been widely used in a range of technologies for its physical and chemical properties, such as high electrical conductivity, photosensitivity, and bioactivity. Production and applications of nanosilver have also existed for more than a century, typically in colloidal form, first for ingestion as a medicinal tonic, and later for direct application as a biocidal agent. 72 More recently, nanosilver has been incorporated within a broad range of consumer products. Over 400 metric tons of silver nanoparticles (AgNPs) are produced globally each year, with applications in cosmetics, textiles and electronics. 11 Almost 30% of this net global production of AgNPs are incorporated into medical supplies and devices due to their bactericidal properties, 11 yet there has been limited research on the fate of AgNPs used in health care settings, or on the significance of the life cycle impacts of these nanoparticles in the context of medical devices A commercial application of AgNPs in the medical field is nanosilver-enabled wound dressing that is applied directly to severe burns and open wounds and is the focus of this study. These bandages are produced by embedding silver particles within the fabric of the bandage using a variety of physical techniques including impregnation, deposition, and coating. 1 The antimicrobial properties of AgNPs allows the bandages to be effective against common gram-negative and - positive bacteria, thus promoting healing of the wounds and preventing infection. 77 AgNPs work against bacteria by releasing silver ions that adhere to the cell membrane, penetrate it, and generate reactive oxygen species that cause oxidative stress and potential DNA damage. 77 Numerous experiments have been carried out to test the effectiveness of these particles against common bacteria such as E. coli, 78,79 with little evidence of bacterial resistance. These bandages have shown promising results in terms of reducing antibiotics use and decreasing the length of stay in hospitals and increasing healing rate of the wounds The potential advantages of using nanosilver-enabled bandages notwithstanding, there is also concern regarding the health and environmental implications associated with the life cycle of AgNPs Silver Nanoparticle Releases from Textiles The growing number of nanosilver applications in consumer products worldwide has led to increasing concerns regarding the potential for direct human exposure and releases to the environment. Both experimental and mathematical models have been developed in past efforts

41 31 to quantify potential AgNP releases from specific products. For textiles, releases of dissolved silver and silver nanoparticles have been reported at laboratory scales for various AgNP-enabled textiles, particularly during mechanical agitation. 6,83,84 Modeling of AgNP releases from textiles has been used to determine the projected level of nanosilver in effluents, suggesting AgNPincorporated textiles will have higher release probabilities than all other applications, due to relatively high concentrations of particles and continuous use and wash cycles. 85 AgNPs released during washing cycles are conveyed with municipal and industrial effluents to wastewater treatment plants (WWTPs), 33 35,84,86 contributing to measurable AgNP concentrations in wastewater. 6,83 AgNPs in WWTPs may be removed by treatment processes but could be reintroduced to the environment through the incineration or use of biosolids as soil amendments or, ultimately contributing to terrestrial and aquatic ecotoxicity. 84 AgNPs may also be released from textiles during waste management and disposal, either as particulate emissions from incineration or leached from solid waste. 39,87,88 These end-of-life releases are particularly relevant for single-use textiles that are not subject to washing, such as the bandages under consideration here. AgNPs in textiles and medical applications have also been considered in larger-scale release and fate models that have tracked multiple engineered nanomaterials (ENMs) across a range of enduses. Early models based on partition coefficients estimated predicted environmental concentrations in Switzerland; 88 these were later extended to include additional environmental compartments and geographic areas, using a probabilistic approach. 89 At a global level, production, use, treatment, and disposal pathways have been reported for ten categories of ENMs, 11 and later geographically disaggregated to allow for location-specific modeling of predicted concentrations. 20 At a more fundamental level, Bayesian modeling has been applied to physicochemical parameters of ENMs to model fate and transport behavior and to estimate some categories of environmental impacts. 90 This latter approach is particularly promising as fate and effects can vary significantly even within each class of ENMs depending on parameters such as size, surface chemistry, and environmental conditions variations that are not generally captured in the larger-scale release models. 91

42 Life Cycle Modeling of Nanomaterials and Nano-enabled Products AgNPs and other ENMs can be released from products not just during cleaning and disposal, but also from product manufacturing and use stages, or at any point during a product s life cycle, both accidentally and deliberately. 87,92 While ENM risk assessment has focused on environmental releases, fate and transport, and subsequent exposure and effects on humans and aquatic invertebrates, impacts from non-nano emissions that occur throughout the ENM life cycle have received less attention. Compared to naturally occurring nanoparticles, ENMs are precisely manufactured to possess certain physical, chemical, or biological properties and many are highly energy intensive to produce, with low material efficiencies compared to other high-tech materials and pharmaceuticals. 15 Sources of manufacturing emissions and environmental impacts differ depending on the production route, but in general, high energy demands, low material efficiencies, and intensive purification requirements can all contribute to the overall impacts of the nanomaterials and nano-enabled products. Nanomanufacturing impacts derived from conventional emissions may exceed impacts from any potential releases of nanomaterials. For example, an assessment of life cycle aquatic ecotoxicity of CNTs demonstrated that non-nano emissions from CNT synthesis and upstream production of materials and energy resulted in ecotoxicity impacts that were several orders of magnitude greater than those from CNT releases. 93 Many of these impacts can occur upstream from the actual synthesis, during production of chemicals or energy, 94 and several studies have highlighted the dominant contribution of emissions of upstream electricity generation to overall emissions. 95,96 Here, the distribution of life cycle environmental impacts for AgNPs were studied and checked for coherence with prior studies. Nanomaterials are rarely used by consumers in their pure form, thus, modeling of nano-enabled products must consider the benefits and potential impacts of nanomaterials in the context of the entire product and its function in the economy. As Life Cycle Assessment (LCA) models utilize functional comparisons, a product-based rather than a material-based assessment framework for engineered nanomaterials has been recommended. 47 Different product categories utilize different concentrations of nanomaterials, and this obviously affects their impacts relative to the larger product. For a nanomaterial-enabled product, the presence of ENMs is not necessarily a driver of total life cycle impacts. For example, in a study of CNTs in electronic memory devices,

43 33 the environmental impacts of CNTs were insignificant compared to those of the metals (particularly gold) and process inputs used during fabrication, indicating that nano-enabled devices may have negligible nano-related impacts. 69 Several studies have examined life cycle impacts of nanosilver applications in textiles. A comparative cradle-to-grave LCA has been performed between a regular T-shirt and one containing AgNPs with two different routes of AgNP synthesis, stating higher global warming potential (GWP) for production stages in comparison to other life cycle phases, due to emissions during silver mining. 97 A similar screening-level study for nanosilver containing socks has been performed also considering for different AgNP production methods, concluding that the overall environmental burden of the product is highly dependent on how AgNPs are synthesized and incorporated into the textiles. 98 The present study extends this previous work by evaluating life cycle environmental impacts of a single-use, nanosilver-coated medical bandage, Acticoat 7. The study uses LCA to contribute to the investigation of the magnitude and primary sources of environmental impacts from nanoenabled products through three interrelated analyses: (i) process contributions to environmental impacts of AgNP synthesis (based on Acticoat 7 specifications); (ii) relative contribution of AgNPs to the life cycle impacts of the overall medical bandage; and (iii) comparison of production vs disposal stages for various categories of environmental impact.

44 Methods and Modeling Scope and System Boundary CRADLE-TO-GATE LIFE CYCLE IMPACTS END-OF-LIFE IMPACTS textiles, plastics, adhesives, energy Material Extraction & Production Ag, chemicals, energy AgNP Synthesis stabilized AgNPs Bandage Manufacture Application and Use Medical Waste Incineration mining, industry, transport emissions AgNP, direct and indirect process emissions AgNP, direct and indirect process emissions (potential AgNP releases) particulates and ash Figure 2.1. System boundary and life cycle stages. Figure 2.1 depicts the system boundary for this analysis, showing the life cycle of Acticoat 7 medical bandage. As specified by the manufacturer, AgNPs used in Acticoat 7 are produced via reactive magnetron sputtering. 99 The synthesis route and associated material inputs and energy consumption data are described in Section Production of basic materials, chemicals, and fuels for subsequent manufacturing steps was traced back to raw material extraction and processing. It should be noted that deposition of particles onto the bandage polymer substrate is also included in the synthesis step as it takes place simultaneously. Textile fabrication and packaging were included in the bandage manufacturing stage described in Section Transportation of the final product was also included in this stage. Application of the bandage was assumed to take place at a hospital setting. Resources and supplies required for bandage application during use phase, such as disposable gloves, were excluded from the analysis since they are not specific to the use of nanosilver. While on the wound, bandages are passive and require no additional resource inputs for use. Leaching into the wound and subsequent bodily uptake was estimated from medical literature to be modest, at approximately 0.1% of silver contained in the bandages. 100 Another article estimated 10% loss of silver from bandages to the skin surface, which may later be cleaned with sterile pads that are then discarded as medical waste. 85 After removal from the wound, used bandages and cleaning pads are treated as biohazardous waste and are either incinerated (assumed here) or sterilized in an autoclave prior to disposal. Acticoat 7 s modeled end-of-life incineration is described in Section The scope of the assessment includes all natural resource inputs and emissions that occur at each point along the life cycle, including emissions of AgNPs and other forms of silver during production and

45 35 incineration, as well all relevant non-nano, non-silver emissions, such as NO x from electric power plants AgNP Synthesis Acticoat 7 antimicrobial barrier dressings use SILCRYST TM silver nanocrystal coating technology. 101 These AgNPs are developed using reactive magnetron sputtering (RMS), 99 a form of physical vapor deposition (PVD). PVD is a family of processes in which an elemental, alloy, or compound target is vaporized from a solid phase to atoms or molecules in a low pressure or vacuum chamber. 102 Particles can be deposited on a substrate via sputtering, a stage where the solid surface is bombarded with the ionized gas mixture in the chamber. 102 The collision results in the ejection of particles and their deposition onto the substrate surface. Sputtering commonly involves plasma and movement of the positively charged ions towards the target surface due to their electric potential difference. 102 Magnetrons are often used to provide strong magnetic fields that hold plasma particles close to the target surface. 103 Magnetrons can use both direct current (DC) and radio frequency (RF) for sputtering. RF sputtering helps avoid positive charge buildup on the target and can also be used for insulating targets. 102 RMS is a PVD technique in which the deposited film is formed via chemical reaction between the gas mixture and the metal target. In RMS, a mixture of argon and reactive gases is used to deposit a film of oxides or nitrides of the target onto the substrate surface. 104 Here, reactive sputtering of silver using RF (13.56 MHz) in argon-nitrogen (RMS-Ar-N) was modeled based on an experimental study and scaled up to the reference flow of 1 kg AgNPs. 104 In the absence of reported data, 100% conversion of the silver target to AgNPs was assumed. Resource inputs to produce 1 kg of AgNPs using the RMS-AR-N method are 10.4 g nitrogen gas mixed with 124 g argon and 28 kwh of electricity. 90% recycling was considered for all carrier gases, excluding capture and repressurization. The RMS-Ar-N method produces particles with a size distribution between 50 to 60 nm.

46 Bandage Manufacturing and Packaging In this study the medical bandage manufacturing process is modeled based on the components of a commercially available wound dressing, Acticoat This bandage is composed of five layers infused with AgNPs for faster healing of the wounds. An SEM image of this bandage can be seen in Figure 2.2 showing the existence of spherical nano-sized silver particles, with composition confirmed by EDAX. The samples were mounted on an aluminum pin using a carbon coated adhesive and sputter coated with conductive platinum prior to imaging. Hence the presence of aluminum, carbon, and platinum in the EDAX plot, respectively. The inner layer and two outer layers are a high density polyethylene (HDPE) mesh, with AgNPs incorporated using the RMS method. 105 The two layers in between the HDPE mesh are made of non-woven rayon and polyester fabric. These layers are ultrasonically welded together. Silver concentration for a 10 cm 12.5 cm piece of Acticoat 7 bandage was empirically derived using acid digestion and shown to be 104 mg. 106 The amounts of HDPE and polyester fabric used for a 125 cm 2 piece of bandage were also calculated by weight. The ultrasonic welding process was not modeled in the inventory due to unavailability of data. Individual packaging slips were assumed to be from bleached supercalendared paper. A summary of the bandage inventory can be found in Table A2 in Appendix A. Figure 2.2. Nano scale SEM image of the Acticoat 7 bandage

47 End of Life Used bandages are collected in hospitals as hazardous waste and are transported to medical waste incineration plants. 107 It was assumed that the packaging and textile burn completely during the process, which is in excess of 800 C. Behavior of ENMs in incineration plants is highly dependent on the configuration and operating conditions, as well as the physicochemical characteristics of the nanoparticles under study. For example, CeO 2 nanoparticles were found in a full-scale test as likely to remain with residues and be deposited in landfills, with current filtration technology capable of removing any nanoparticles in the flue gas. 37 Other metallic nanoparticles may volatilize during incineration and condense into nanoparticles after passing through the filters. 38 Incinerated AgNPs either remain in the slag or become airborne. 11 Even though a majority of these particles can be captured by filtration, it was estimated that between % of the total AgNPs can be released into the atmosphere. 11 All other incinerator emissions are modeled using existing unit processes, as described in Section Subsequent fate of AgNPs after their release from incinerators was modeled as follows. AgNPs in sealed landfills are assumed to remain deposited there or recirculated with leachate. 89 Settling of airborne particles was allocated based on area covered by water and land. For the US, 93.2% of the airborne particles were assumed to settle onto soils and the rest on surface waters. 89 Longterm studies on AgNP fate in soils have demonstrated that almost 70 wt% of AgNPs remain in soil or sediment they were added to, with partial transformation to Ag 2S in soils and a combination of Ag 2S and Ag-sulfhydryl compounds in sediments 23, 29 This study focuses on the remaining 6.8% of airborne AgNPs that deposit onto surface waters. Nanoparticles in aquatic environments participate in specific processes that should be included in any multi-media fate and transport models, such as aggregation, sedimentation, adsorption, and dissolution. 108 Multi-media models have been developed for that aid in determining concentrations of engineered nanoparticles across air, rain, surface water, soil and sediments. 21 AgNPs under oxidative environments in water will not undergo complete dissolution, while partial dissolution depends on their size and shape, 109 residence time, and ambient temperature. 91 Water chemistry is another important factor in dissolution since chlorides, sulfides, dissolved organic carbons (DOCs) and natural organic matter (NOM) directly affect the ion release rate. 30 Reaction modeling with inclusion of NOMs can be used to predict Ag ion release, 110 but first order kinetics for dissolution have been shown to be adequate. 21,91,108,111 Here, we based Ag dissolution rates on recent experimental studies looking

48 38 at ion release from a range of particle sizes and capping agents into natural waters. 28 Particles less than 10 nm in diameter showed almost 80% loss, and larger particles (50 nm) lost at most 50% of their mass over a 4 month period. 28 RMS-Ar-N synthesized AgNPs have diameters between nm, and so 50% dissociation was assumed here. The characterization factor for ecotoxicity impacts associated with free silver ion fraction of AgNPs in surface waters was used in subsequently LCA modeling (Section 2.2.5). Table A3 contains the life cycle inventory data for this stage of the life cycle Life Cycle Assessment Modeling This analysis was performed using SimaPro 8.1 as the life cycle modeling platform, which allows for convenient life cycle inventory data management and assessment of environmental impacts stemming from each of the ~10,000 discrete substances emitted over the life cycle of the bandages. Each entry life cycle inventory developed for this study was matched with an appropriate unit process from the US-EI database (Earthshift, Huntington, VT), the ecoinvent database adjusted for U.S. energy inputs, with details provided in the supplement Tables A1-A3, found in Appendix A. Environmental impacts stemming from all direct and indirect emissions were modeled using The U.S. EPA s Tool for the Reduction and Assessment of Chemical and Other Environmental Impacts (TRACI) 2.1 impact assessment model. TRACI is a collection of linked fatebehavior-effect models that considers the following impact categories and equivalent units: ozone depletion (kg CFC-11 eq), global warming potential (kg CO 2 eq), and smog formation (kg O 3 eq); respiratory effects (kg PM 10 eq); water and soil quality impacts including acidification (mol H + eq) and eutrophication (kg N eq); human health impacts from toxic carcinogenic and noncarcinogenic substances (health comparative toxic units, or CTUh); and ecotoxicity (CTUe). TRACI was chosen because it is a widely used midpoint model that expresses impacts in terms of discrete environmental problems and was developed for the US context. TRACI 2.1 uses the USEtox consensus multi-media fate and exposure model to model effects on human health and ecotoxicity. The USEtox characterization factor for silver ions was combined with estimates of silver dissociation from AgNPs deposited in surface waters to estimate aquatic ecotoxicity due from nanosilver releases. USEtox ecotoxicity characterization factors for metals are interim, and their parameter uncertainty was considered in subsequent analysis, described in the following section.

49 Uncertainty Analysis Monte Carlo simulation was used for uncertainty analysis associated with all inputs and model parameters. Default, log-normally distributed probability distributions were used from all background US-EI database unit processes. For direct inputs and foreground processes, probability distributions were calculated using the pedigree-matrix approach, with details provided in Table A5. Uniform distributions were assumed for AgNP release from incineration and ecotoxic effects, using the range suggested in the USEtox model. Simulation was performed over 1,000 iterations using SimaPro 8.1 and summary statistics extracted. Simulation results are described for each set of results presented in Section Results The results of this study have been organized to address the three research questions posed in the Introduction section, namely process contribution to environmental impacts of the AgNP synthesis route (Section 2.3.1), the relative contribution of AgNPs to overall bandage production (Section 2.3.2) and end-of-life impacts (Section 2.3.3), and the relative contribution of different life cycle stages (Section 2.3.4) AgNP Synthesis Analysis of process contributions to AgNP synthesis demonstrated in Figure 2.3, was performed based on the requirements for producing 1 kg of AgNP. Analysis of process contributions reveals that the top contributing processes to the global warming potential (GWP) are all combustion related, specifically of hard coal, natural gas, lignite, and diesel used for power generation. The largest contributor to ecotoxicity is related not to nanosilver releases but to environmental releases from bulk silver processing, namely the disposal of sulfidic tailings from mining and refining processes of copper and lead ores that contain silver, emissions that occur far upstream from AgNP synthesis. The aforementioned processes also contribute significantly to carcinogenic, non-carcinogenic and eutrophication impact categories. Acidification is mainly due to refining of copper sulfide ores through the generation of SO x emissions, blasting, and power generation processes. Figures A1-A10 demonstrate the contributing processes to silver production for each of the TRACI 2.1 impact categories, while Monte Carlo simulation results for all impact categories

50 40 can be found in Table A4 in Appendix A. Uncertainty for carcinogenicity and ecotoxicity were found to be the highest among all impact categories, with coefficients of variation (CV) of % and stemming in large part from the uncertainties in incinerator emissions and characterization factors, in particular the interim ecotoxicity characterization factor for free ionic silver. These large uncertainties hinder decision-making on the basis of life-cycle toxicity results, and future efforts should focus on improving modeling capabilities and data quality in order to reduce uncertainty. The environmental impact categories of global warming, fossil fuel depletion, acidification, ozone depletion, and smog formation all had CVs in the 35-50% range. 100% 80% 9.5E-06 kg CFC E+02 kg CO2 eq 2.3E+01 kg O3 eq 2.4E+00 kg SO2 eq 6.7E+00 kg N eq 5.0E-05 CTUh 1.6E-04 CTUh 2.4E-01 kg PM E+03 CTUe 1.4E+02 MJ surplus Synthesis emissions 60% Carrier gases 40% Electricity 20% 0% OD GW PS AC EU HHC HHNC HHCR EC FF Silver Figure 2.3. Process contribution for all TRACI impact categories of RMS-Ar-N synthesis route (silver source is solid silver, gaseous elements are argon and nitrogen). Labels demonstrate the total absolute value of each impact category. Abbreviations: ozone depletion (OD), global warming (GW), photochemical smog (PS), acidification (ACF), eutrophication (EU), human health: carcinogens (HHC), human health: non-carcinogens (HHNC), human health: criteria air pollutants (HHCR), ecotoxicity (EC), fossil fuel depletion (FF).

51 Bandage Production Life cycle environmental impacts for the production of Acticoat 7 were modeled to determine the contribution of AgNPs to the overall impacts of bandage manufacture. Figure 2.4 shows the results for the scenario where RMS-Ar-N would be used to produce and deposit AgNPs. This figure illustrates that the largest contributor to impacts in all categories is nanosilver, even though AgNPs make up just 6% of the bandage mass. CVs from the uncertainty analysis can be seen as the labels in Figure 2.4. The CV of the ecotoxicity results is among the highest, again due to large uncertainties in the fate and effect modeling for AgNPs emitted from incineration. Figure 2.4. Bandage manufacturing process contributions. RMS-AR-N AgNPs were assumed to be incorporated. The labels on bars correlate with the CV of that impact category. Abbreviations: ozone depletion (OD), global warming (GW), photochemical smog (PS), acidification (AC), eutrophication (EU), human health: carcinogens (HHC), human health: non-carcinogens (HHNC), human health: criteria air pollutants (HHCR), ecotoxicity (EC), fossil fuel depletion (FF).

52 End of Life Used bandages are treated as medical waste and may be incinerated after disposal. Figure 2.5a below represents a scenario where a 125 cm 2 piece of Acticoat 7 and its packaging are burnt and between 0.05 and 1% of the nanosilver in the bandage is released into the atmosphere as particulate emissions. Incineration products of paper and plastic components of the bandage are the largest contributors for most categories of environmental impact, with the exception of ecotoxicity, where the dissociated silver has the highest contribution, as shown in Figure 2.5a (see Figure A6 in Appendix A for additional details). For the 6.8% deposition rate onto surface water and 50% dissociation, it was calculated that between mg of silver will dissociate per bandage. Based on the USEtox characterization factor for Ag + in freshwater ( CTUe/kg), the ecotoxicity impact of the dissociated silver ions were calculated to be between and CTUe per bandage. For comparison, the normalization factor (average annual impacts per capita) for ecotoxicity from metal pollution in the US has been estimated at CTUe, 112 so these AgNPs make a relatively small contribution to overall ecotoxicity Life Cycle Stage Comparisons Figure 2.5b illustrates a comparison among the life cycle stages of Acticoat 7 for GWP and ecotoxicity impact categories. Bandage manufacturing is responsible for higher impacts compared to conventional and AgNP emissions during disposal. Impacts from production are predominantly due to non-nano emissions from conventional industrial facilities, especially from silver mining and processing facilities and electric power plants. Thus efforts to reduce environmental impacts, including ecotoxicity specifically, can be advanced by not just reducing the risks of direct nanoparticle exposures but also by improving the energy and material efficiency of AgNP synthesis and reducing industrial process emissions upstream of AgNP production.

53 43 (b) Figure 2.5. (a) Process contribution to bandage incineration, case of Acticoat 7, for ecotoxicity and GWP, (b) life cycle stage comparison of Acticoat 7 bandage in terms of ecotoxicity and GWP Discussion and Implications AgNP-enabled medical devices are an important category of nano-enabled products. 113 These results characterize and quantify the life cycle impacts associated with the Acticoat 7 bandage at three levels: AgNP synthesis, AgNP impacts relative to the entire bandage, and across bandage life cycle stages. Impacts associated with AgNP synthesis dominate the cradle-to-grave impacts of the bandage, and emissions from AgNP and bandage production are several times more impactful than emissions from bandage incineration, including direct releases of AgNPs. For comparison, Walser et al. reported that AgNP synthesis dominated the cradle-to-gate impacts of a nano-enabled t-shirt when a plasma polymerization with Ag co-sputtering process was assumed, but were relatively insignificant when flame-spray pyrolysis AgNP synthesis was considered. 97

54 44 These diverse results underscore the need to use product-specific synthesis data as opposed to generic production inventories when assessing nano-enabled products. AgNPs are finding increasing direct contact applications, yet there has been limited research on the fate of AgNPs used in health care settings, or on the significance of the life cycle impacts of these nanoparticles in the context of medical devices. Wound dressings were chosen here as the product under study as a representative of this category with high nanoparticle loading. But nanosilver is present in several types of medical supplies and a wide range of consumer products. Manufacturers choose synthesis route and subsequent purification and surface functionalization of AgNPs depending on performance specifications for the intended application, given rise to a wide range of physicochemical properties. In the context of life cycle assessment, these differences should be considered when determining an appropriate functional unit. It is important to note that the AgNP physicochemical properties and morphology affects not just product function but also environmental fate and transport of silver particles in the environment and their stability. Hence, particles developed with different techniques will both perform very differently and have different environmental impacts. 1 It is therefore critical that life cycle design and assessment including consideration of ENM size, shape, and surface chemistry The fate of AgNPs in products and in the environment is under active investigation and further work will serve to reduce the uncertainty of the results presented here. For example, it has been previously shown that only 10% of AgNPs are released from bandages during use, 117 but additional empirical data of AgNPs leaching from bandages and subsequent fate will support the modeling of additional routes of AgNP disposal and potential release. Second, the fate of AgNPs in incineration plants or other solid waste management facilities has not yet been fully investigated It is hypothesized that AgNP will partially volatilize in the high furnace temperatures and then condense and have surface reactions with other flue gas constituents, 85 but physicochemical characterization and toxicity testing of AgNPs following incineration has yet to be performed. Sulfidation in particular has been shown to reduce the toxicity of AgNPs in the environment, due to reduction of Ag + availability in the environment. 34,110,119 A general difficulty of applying LCA to nanomaterials and nano-enabled products is that the use confers superior properties or performance that is not attainable by the current technologies to

55 45 which they are to be compared. In the case of bandages, AgNPs might replace the use of a topical antibiotic such as triclosan, and one could use standard system expansion techniques in LCA to show the benefits of reducing triclosan manufacturing. But other benefits of using AgNPs would be more difficult to capture using current LCA methods, such as reduced bacterial resistance to antibiotics. 78,79 One area of potential future research is to develop quantitative techniques for assessing the potential health and societal benefits of nanomaterial applications that can been expressed in the same units as current LCIA methods, such as disability-adjusted life years (DALYs), as was reported recently for a LCA of a CNT-enabled gas sensor. 120 Finally, as with many emerging technologies, the data used in this and nearly every other LCA study for engineered nanomaterials were gathered from bench-scale investigations. Commercial production facilities are more efficient in terms of material and energy use, hence our results should be seen as an upper bound for impacts from the production phase. Here 100% conversion of the silver target to AgNPs was assumed, and in general yield values are not reported in the literature and must also be bounded in nano LCA studies. Life cycle impacts will likely decrease as nanomaterial production volumes grow and nanotechnologies become more mature, but life cycle modeling can continue to identify opportunities for process improvements and provide context for understanding relative risks and benefits of nanotechnologies.

56 46 Chapter 3 Comparative Life Cycle Assessment of Silver Nanoparticle Synthesis Routes This study has been published Pourzahedi, L., & Eckelman, M. J. (2015). Comparative life cycle assessment of silver nanoparticle synthesis routes. Environmental Science: Nano, 2(4), Silver nanoparticles (AgNPs) can be produced through a variety of synthesis routes with differing mechanisms, inputs, yields, reaction conditions, and resulting size distributions. Recent work has focused on applying green chemistry and sustainable manufacturing principles to nanomaterial synthesis, with the goal of reducing life cycle energy use and environmental impacts. Life cycle assessment (LCA) is used here to analyze and compare the environmental impacts of AgNPs produced through seven different synthesis routes (cradle-to-gate). LCA reveals both direct and indirect or upstream impacts associated with AgNPs. Synthesis routes were chosen to represent the current trends in nanoparticle synthesis and include physical, chemical and bio-based methods of production. Results show that, across synthesis routes, impacts associated with the upstream production of silver itself were dominant for nearly every category of environmental impact, contributing to over 90% of life cycle burdens in some cases. Flame spray methods were shown to have the highest impacts while chemical reduction methods were generally preferred when AgNPs were compared on a mass basis. The bio-based chemical reduction route was found to have important tradeoffs in ozone depletion potential and ecotoxicity. Rescaling results by the size-dependent antimicrobial efficacy that reflects the actual function of AgNPs in most products provided a performance-based comparison and changed the rank order of preference in every impact category. Comparative results were also presented in the context of a nanosilver-doped wound dressing, showing that the overall environmental burdens of the product are highly sensitive to the synthesis route by which the AgNPs are produced.

57 Introduction Silver nanoparticles (AgNPs) are a versatile class of engineered nanomaterials (ENMs), produced in the hundreds of tons annually and used in a wide range of consumer and high-tech products. Major end-uses categories include medical products (30%), paints and coatings (25%), textiles (15%), and cosmetics (15%). 11 Silver has the highest electrical and thermal conductivity and reflectivity of all the elements, but the function of AgNPs in the majority of applications currently is as an antimicrobial agent. Antimicrobial activity of AgNPs is a function of particle morphology and surface chemistry, as well as environmental conditions including ionic strength and the presence of natural organic matter (NOM). 21 Increasing commercial production volumes of ENMs generally has led to extensive research into potential environmental and health implications of their use, particularly given the projections for their widespread incorporation into consumer products. This research includes mechanisms and projected magnitudes of ENM releases from products, 3,84 treatment and control technology efficiencies, 34,35 fate and transport in the environment, 21,87,90 and potential exposures and effects. 22,88,108 Based on a mechanistic understanding of ENM fate and effects, research is also increasingly being directed at the design of ENMs that mitigate environmental exposures and potential toxic effects, for example through addition of surface functional groups. 121 In the case of AgNPs, considerations of impact have centered on ecotoxicity to aquatic organisms. Thousands of tons of silver compounds are estimated to be released worldwide to the environment annually, only a fraction of which is nanosilver, but their impacts vary widely across the different chemical and physical forms of the emissions. 122 In addition to releases of ENMs themselves, the resource intensity of nanomaterial synthesis has also been identified as of environmental concern. Production of nanomaterials can be orders of magnitude more energy- and materials-intensive than for fine chemicals or pharmaceuticals, 15 often with low yields and involving toxic or hazardous reagents and solvents. 94 Life cycle assessment (LCA) studies of ENMs that consider the environmental implications across production, use, and release have found that energy and chemicals use during ENM synthesis contributes a significant, and often dominant share of total life cycle environmental impacts. 64,97 In earlier work, we reported LCA results for a medical bandage treated with AgNPs synthesized via magnetron sputtering, examining the relative contribution of AgNPs to the life cycle impacts

58 48 of bandage production and end-of-life incineration. 123 The production of AgNPs was the largest contributor to product impacts across all environmental categories, despite low concentrations in the bandage. These impacts were driven by extraction and processing of the silver, as well as synthesis-related resource inputs, particularly electricity. Metallic nanoparticles such as gold and silver can be synthesized using a wide variety of techniques, including vapor deposition, ablation and sputtering, flame spray pyrolysis, electrochemical methods, microwaves, and chemical reduction methods. 94,124 Green chemistry and green nanoscience researchers have been successful at developing novel synthesis routes that reduce resource requirements and use benign chemicals, while maintaining control of ENM size and morphology. 125 Given the importance of synthesis processes to life cycle impacts of ENMs, using synthesis-specific life cycle data is critical to achieving accurate LCA results for existing nano-enabled products. Unfortunately, representative and validated life cycle inventory data are lacking for many nanoparticle synthesis methods that have been reported in the literature, and there have been calls to accelerate efforts to make these data sets available. 47 In addition, most nano LCA studies have considered only a single synthesis method, yet comparisons across synthesis routes can also identify drivers of overall impacts and the relative environmental advantages of a certain route, complementing other considerations of purity, yield, and cost. 8,97,98 For example, a recent article by Pati et al. compared chemical reduction routes for the synthesis of gold nanoparticles using 3 conventional and 13 bio-based reducing agents. 126 Gold itself was the most important driver of life cycle energy use, such that syntheses using bio-based reducing agents but with low yields and long reaction times were not favored. Such information can also be used to prioritize research for improving novel synthesis routes prior to commercial scale-up. Here we present life cycle inventory data and detailed assessment results for seven AgNP synthesis routes (including one bio-based route) in order to compare the magnitude and patterns of impacts across multiple environmental categories. AgNPs produced through different routes have varying size distributions and morphologies, and so results are presented both per unit mass and also normalized by their estimated antimicrobial efficacy. These data and results are intended

59 49 for use in further research into ENM synthesis as well as subsequent environmental assessments of AgNP-enabled products Methods and Modeling Scope and System Boundary This cradle-to-gate analysis considers the direct emissions from AgNP production and indirect impacts associated with upstream resource extraction and energy generation. Production methods of seven different nanosilver synthesis routes and their corresponding material and energy inputs are described in Section Environmental impacts were analyzed and compared initially using a mass-based functional unit of 1 kg of AgNPs, and later normalized by their estimated antimicrobial efficacies (Section 3.2.3). Additionally, potential impacts of AgNPs were studied in context of a commercial product, Acticoat 7, a commercially available medical bandage containing nanosilver. Life cycle inventory data for this product can be found in our earlier work. 67 Here we model the relative contribution of AgNPs to the overall product, considering each of the modeled AgNP synthesis routes AgNP Synthesis Routes The AgNP methods of production considered here can be classified into two general categories: wet chemistry and physical techniques. Wet chemistry through chemical reduction utilizes solutions of silver reacted with an appropriate reducing agent to produce metallic nanoparticles, while capping agents are typically added to stabilize the colloidal solution. Here, reduction of silver nitrate with trisodium citrate (CR-TSC), 127 sodium borohydride (CR-SB), 128 and ethylene glycol (CR-EG) were considered. 129 Chemical reduction of silver nitrate with soluble starch from potatoes (CR-Starch) was also modeled in this study as a representative, novel bio-based chemical reduction method, 130 though citric acid for the CR-TSC method is also industrially produced through biotechnological means. Material input was based on data provided by the literature. Energy required for heating solutions was calculated based on the heat capacity of water. Physical routes considered in this were flame spray pyrolysis (FSP), 97 arc plasma (AP), 131 and reactive magnetron sputtering with an argon and nitrogen gas mixture (RMS-AR-N). 104 Further details and life cycle inventory (LCI) data for all methods can be found in the subsections below and Tables

60 50 B1-B7 of the SI. A summary of all methods and their resulting particle size ranges is provided in Table Chemical Reduction of Silver Nitrate with Trisodium Citrate (CR-TSC) Synthesis of citrate-stabilized nanosilver was first reported in The method considered here reduces a 1 mm silver nitrate solution using 1% trisodium citrate. The process reaction is as follows, Ag + + C 6H 5O 7Na H 2O 4 Ag 0 + C 6H 5O 7H 3 + 3Na + + H + + O 2 To simulate this reaction, chemical elements must not be in their ionized form, hence it is assumed that the previous reaction is a result of the reaction below, 12 AgNO C 6H 5O 7Na H 2O 12 Ag + 4 C 6H 5O 7H NaNO 3 +4 H 2 +5 O 2 Silver nitrate is heated to boiling prior to the addition of trisodium citrate. The mixture is further heated until the formation of AgNPs. Stoichiometric relations suggest that, production of 1 kg AgNPs from this reduction requires 1.57 kg silver nitrate, 0.8 kg trisodium citrate, and 0.14 liter of water. Byproducts of this reaction are 0.6 kg citric acid, 0.8 kg sodium nitrate, kg hydrogen gas and 0.12 kg oxygen. The minimum heating energy required to boil the solution was calculated based on the heat capacity of the solution. Production of trisodium citrate was studied as part of this synthesis route. One kg of this compound is obtained by reacting 0.74 kg citric acid with 0.46 kg sodium hydroxide, as per the chemical reaction assumed below (0.2 liter water is also produced alongside trisodium citrate), C 6H 8O NaOH Na 3C 6H 5O H 2O Citric acid is industrially manufactured using the fungus Aspergillus niger and a source of sugar. This is done by a submerged fermentation technique. Here, sugarcane was used as the source of sugar. The reaction below shows the process performed by the fungus, in which 0.93 kg glucose is needed to produce 1 kg citric acid, 132 C 6H 12O 6 + 3/2 O 2 C 6H 8O H 2O After this step, citric acid is treated with calcium hydroxide and sulfuric acid for further purification, reactions shown below, 133

61 51 3 Ca(OH) C 6H 8O 7 Ca 3(C 6H 5O 7) 2. 4 H 2O + 2 H 2O 3 H 2SO 4 + Ca 3(C 6H 5O 7) 2. 4 H 2O 2 C 6H 8O CaSO 4. 2 H 2O + 2 H 2O Treatment of 1 kg citric acid requires 0.76 kg sulfuric acid and 0.57 kg calcium hydroxide (hydrated lime). Hydrogen and calcium sulfate (gypsum) are both by-products of these reactions Chemical Reduction of Silver Nitrate with Sodium Borohydride (CR-SB) Reducing silver salt solutions with sodium borohydride is the most common method used among chemical reduction syntheses. 134 The reduction of a 1 mm silver nitrate solution with a 2 mm sodium borohydride solution was considered based on the reaction below, 128 AgNO 3 + NaBH 4 Ag + ½ H 2 + ½ B 2H 6 + NaNO 3 Due to the exothermic nature of this process, sodium borohydride is chilled in an ice bath prior to the addition of silver nitrate, in order to decrease the rate of reaction. As sodium borohydride undergoes side reactions during silver nitrate reduction, it is generally used in excess. 128 The molar ratio of sodium borohydride to silver nitrate directly affects the rate of formation of AgNPs, with a range of ratios reported. 128,135,136 Any unreacted sodium borohydride that could be recovered and recycled would be subtracted from the life cycle inventory. Therefore, as a baseline the above reaction was used here to calculate the stoichiometric amount of NaBH 4 needed to reduce silver nitrate kg silver nitrate and 0.35 kg sodium borohydride are needed to obtain 1 kg silver particles alongside 0.12 kg diborane, 0.78 kg sodium nitrate and kg hydrogen as emissions and co-products. Production of Sodium borohydride was further studied. Sodium borohydride is produced by reacting 2.27 kg trimethyl borate and 2.5 kg sodium hydride. The aforementioned amounts result in 1 kg sodium borohydride and 4.28 kg sodium methoxide as a co-product. Below is the chemical reaction for this process, 137 B(OCH 3) NaH NaBH NaOCH 3 Here, sodium hydride was modeled using inventory data for sodium hydroxide, as both are produced by electrolysis of sodium chloride.

62 Chemical Reduction of Silver Nitrate with Ethylene Glycol (CR-EG) In this technique, ethylene glycol is used as the reducing agent to obtain silver particles from silver nitrate (with 3.81 g/l silver concentration) and poly n-vinylpyrrolidone (PVP) is used as the stabilizing agent. 138 The reaction takes place at room temperature For the production of 1 kg silver particles, 1.57 kg silver nitrate, 291 kg ethylene glycol and 47 kg PVP are required. A recycling rate of 90% was assumed for ethylene glycol, consistent with most LCI datasets for chemicals, thus only 29.1 kg of this product is required as input per kg AgNPs. To produce PVP, first 1,4-butanediol is dehydrogenated over copper at 200 o C to form gammabutyrolactone. Gamma-butyrolactone is then reacted with ammonia to obtain pyrrolidone, which is subsequently treated with acetylene yielding VP monomers. Reactions below describe these set of processes, 141 C 4H 10O 2 C 4H 6O H 2 C 4H 6O 2 + NH 3 C 4H 7NO + H 2O C 4H 7NO + C 2H 2 C 6H 0NO Stoichiometric relations require that for producing 1 kg PVP, 0.8 kg 1,4-butanediol, 0.15 kg ammonia and 0.23 kg acetylene is required. Emissions from these reactions were calculated to be 0.04 kg hydrogen and 0.16 liter water Green Synthesis using Soluble Starch (CR-Starch) To reduce life cycle impacts due to the chemical reducing agents, plant extracts can be used as an bio-based alternative. 142 Here we consider synthesis of AgNPs using a starch solution as both the reducing and stabilizing agent for the silver nitrate solution. 130 Here, 0.9 g of a chemically modified starch (hydroxypropyl starch, with molar substitution ratio of 0.84, resulting in its complete solubility in water) is added to 100 ml deionized water and heated up to 70 C. The dissolved starch solution is used to reduce the silver nitrate solution. An AgNP solution with a concentration of 100 ppm was formed after 60 minutes of heating. The heat required to bring the solution up to temperature was subsequently calculated. With ~100% yield, 143 to produce 1 kg of silver nano particles, 90 kg of starch is required alongside 1000 liter of water to reduce 1.57 kg of silver nitrate.

63 Flame Spray Pyrolysis (FSP), Utilizing a Methane-Oxygen Flame This technique produces nano-particles of metal or metal oxides by combusting metal precursor solutions. The particles are formed just above the flame and collected from a filter. 97 Here, a precursor solution of silver octanoate was sprayed into a methane-oxygen flame. 97 Producing 1 kg of pure silver nano-particles using this route requires 33.4 kg oxygen, 1.53 kg methane, 62.8 liter water, 2.35 kg silver octanoate, 6.29 kg 2-ethylhexanoic acid, 6.29 kg xylene, and 25.1 kwh of electricity. In addition, silver octanoate and 2-ethylhexanoic acid lacked existing LCI data and so were modeled separately. To produce 1 kg of 2-ethylhexanoic acid, 1.02 kg n-butyraldehyde is needed, while 0.05 kg CO 2 emissions and 7 g of steam are generated as byproducts. N-butyraldehyde is an organic compound, commercially manufactured using the oxo process, where propylene, carbon monoxide and hydrogen are combined with in the presence of a rhodium or cobalt catalyst, 97 2 CH 3CH=CH CO + 2 H 2 CH 3(CH 2) 2CHO + (CH 3) 2(CH) 2O From propylene, n- and iso-butyraldehyde are produced. This reaction gives a 4:1 ratio of n- to iso-butyraldehyde. 97 Following Walser et al., we use an economic allocation factor of 90% to assign process burdens to the n-butyraldehyde. By this reaction, manufacturing 1 kg n- butyraldehyde requires 0.58 kg propylene, 0.02 kg oxygen, and 0.39 kg carbon monoxide. Silver octanoate is derived from the reaction between octanoic acid (a fatty acid component of coconut oil), silver nitrate, sodium hydroxide, and deionized water. 97 To obtain 1 kg of silver octanoate, 2.7 kg crude coconut oil, 2.57 kg silver nitrate, 0.38 kg sodium hydroxide, and 0.9 liter deionized water are required. Silver nitrate is commercially produced by dissolving silver metal in cold and diluted nitric acid in a pure oxygen environment. This process produces nitrogen monoxide and water as a result of the oxidation, chemical reaction shown below, Ag + 4 HNO 3 3 AgNO H 2O + NO Based on this stoichiometry, to produce 1 kg of silver nitrate, 0.49 kg nitric acid and 0.64 kg silver are needed. In this process 0.07 liter water and 0.06 kg nitrogen monoxide are also produced as byproducts.

64 Arc Plasma Reactor (AP) Arc Plasma is a physical vapor deposition method modeled in this study. 131 Particles are created by vaporizing pure silver in a reactor using a high electricity discharge. They are then condensed in an argon chamber, simultaneously depositing onto the substrate with a rate of 1.2 g/min. Based on previous studies, an average argon flow rate of 5 l/min was assumed here. 145,146 Assuming 100% yield, to produce 1 kg of silver nano-particles, 1 kg of pure silver, 7.4 kg of argon gas and is required using a DC current of 120 A with a voltage of 25 V Reactive Magnetron Sputtering of Silver with Ar-N 2 Gas Mixture (RMS-AR-N) Reactive magnetron sputtering (RMS) is another type of physical vapor deposition. This process uses argon as an inert gas for bombardment of the sputtering target, silver. The collision results in the ejection of silver ions and their movement toward the substrate. In reactive sputtering, a mixture of argon and reactive gases is used to deposit films of oxide or nitride forms of the target onto the substrate surface. Here, material and energy consumption for producing 1 kg of AgNPs were estimated based on experimental data for an argon-nitrogen gas mixtures. 104 To produce 1 kg of AgNPs from 1 kg pure silver (assuming 100% yield), 10.4 g nitrogen gas and 124 g argon is required and 27.8 kwh of electricity is used for this process. Table 3.1. Synthesis route summary and particle size distribution. Synthesis Category Wet Chemistry Mechanism Source of Silver Abbreviation Size (nm) Source Reduction with trisodium citrate CR-TSC 100 Sileikaite et al. 127 Reduction with sodium borohydride AgNO 3 CR-SB Mulfinger et al. 128 Reduction with ethylene glycol CR-EG 10.4 Slistan-Grijalva et al. 138 Reduction with soluble starch CR-Starch El-Rafie et al. 130 Physical Routes Flame spray pyrolysis AgC 8H 15O 2 FSP Arc plasma Ag (s) AP 100 Walser et al. 97 Zhou et al. 131 Reactive magnetron sputtering in Ar-N 2 Ag (s) RMS-AR-N Pierson et al. 104

65 AgNP Functional Unit Functional unit is a central consideration in life cycle assessment and other comparative analytical techniques. The antimicrobial efficacy of AgNPs in products depends on characteristics such as particle size, shape and capping agent, as well as how the AgNPs are incorporated. 73,77 As shown in Table 3.1, the various synthesis routes considered here can produce particles over a wide size range. Nanoparticle size and shape can be tuned by controlling reaction conditions, depending on synthesis route. In the FSP method, the concentration of the metal precursor and the rate at which it is fed to the spray gun affect the particle size. 147 In the arc plasma technique, higher arc current results in larger particle sizes. 148 As the evaporation rate of the precursor metal increases the size of silver particles and the standard deviation of their size distribution also increases. 145 For chemical reduction methods, the size and shape of the particles are affected by the type of stabilizing agent, reducing agent and its redox potential, molar ratios, and other conditions of the experiment. 1 Ordinarily, the particles formed via chemical reduction have higher impurity levels (depending on the reducing efficiency) than other methods but are capable of producing homogenous nanoparticles with regular shapes. 1 Even though green synthesis routes use the same chemical reduction techniques, there is typically less control over the size of particles when using bio-based reducing agents. 1 To account for properties and characteristics of nanoparticles, as an alternate functional unit to 1 kg AgNP production, the bioactivity of AgNPs was explored in the form of LC 50. This is the lethal concentration where 50% loss of the sample population is observed. Antimicrobial efficacy of AgNPs produced by each method were estimated using size-dependent LC 50 values (µg/ml). 114 A linear regression was obtained using experimental toxicity data for zebrafish embryos subjected to a range of nanosilver particle sizes: LC D + 18; R 2 = 0.41, where D is the diameter of the nanoparticles (in nm). In general, larger particles require higher concentrations to kill 50% of the population, illustrating the potency of AgNPs smaller in size. Only relative values are required, and it is assumed here that nanosilver will display the same sizedependent toxicity trend for bacteria.

66 Life Cycle Assessment Modeling Analysis of life cycle impacts was performed using SimaPro 8.1 as the modeling platform. Each record of energy and chemical use derived in Section was matched with an appropriate unit process from the US-EI database, the Ecoinvent 3.1 database adjusted for U.S. energy inputs. Environmental impacts were modeled using the U.S. EPA s Tool for the Reduction and Assessment of Chemical and Other Environmental Impacts (TRACI) 2.1 assessment method. 57 This method considers the impact categories of global warming potential (kg CO 2 equivalents [eq]), ozone depletion (kg CFC-11 eq), smog formation (kg O 3 eq), respiratory effects (kg PM 10 eq), acidification (mol H + eq), eutrophication (kg N eq), human health impacts from toxic carcinogenic and noncarcinogenic substances (health comparative toxic units, or CTUh), and ecotoxicity (CTUe). Detailed life cycle inventory, and additional process contribution analysis for synthesis routes can be found in Appendix B. Tables B1 to B7 include the full life cycle inventory from synthesis route modeling Comparative Results and Discussion Results and discussion for this analysis have been organized into five sections. Section discusses the comparative environmental impact assessment on a mass basis between the synthesis routes. Section considers the rescaling of the previous comparative results for the functional unit. Processes contributing to the overall impacts of production are explained in Section Comparison of the findings of this study to LCAs previously performed on AgNPs is discussed in Section Finally, Section demonstrates the effect of synthesis choice in the context of a nano-enabled product. Figures B1 to B7 of Appendix B illustrate the LCA results for each synthesis route and the relative contribution of their components. A comparative massbased LCA of the reducing agents used in this study is shown in Figure B8. Figures B9 and B10 clarify the main contributors of global warming and ecotoxicity impacts of silver itself Synthesis Comparison by Mass Comparison among synthesis routes was first performed using a mass-based functional unit of 1 kg of AgNPs, with relative results shown in Figure 3.1a (absolute values can be found in Table B8). Analysis indicates that manufacturing silver particles using the flame spray pyrolysis method with

67 57 silver octanoate as the precursor sprayed in a methane-oxygen flame has the highest of impacts across all categories except fossil fuel use. These relatively high impacts are largely due to the low AgNP production yield of the FSP synthesis route (roughly 50%) relative to other routes, as well as significant direct electricity use (see Table B5 and Figure B5). Yield has been shown in several previous reports to drive environmental impacts, 126 because life cycle assessments account for all of the energy and materials used to produce reagents, including those that are used in excess. Chemical reduction with ethylene glycol had the highest level of fossil fuel use overall, due to the substantial use of fossil-based ethylene in upstream chemical production. Just among chemical reduction methods, CR-EG had the highest impacts for smog, carcinogenics, and fossil fuel depletion impact categories. This is due both to the silver precursor and to the reagents used, especially EG and the PVP used as a capping agent. On a unit mass basis, the life cycle environmental impacts of ethylene glycol are modest when compared to other reducing agents, with the highest relative impacts in the categories of acidification and respiratory disease, where volatilized ethylene glycol acts as a precursor to secondary particulate matter (see Figure B8). But as ethylene glycol is being used in large quantities for this chemical reduction route (29.1 kg/kg AgNPs), the upstream production of chemicals drives impacts in most categories for this synthesis route. Reduction with trisodium citrate dominates for global warming potential as the reaction is carried out at boiling temperature, requiring comparatively more heating energy compared to other chemical reduction methods, which in turn results in higher GHG emissions. Again among chemical reduction methods, CR-Starch had the highest impacts for ozone depletion, acidification, eutrophication, non-carcinogenics and ecotoxicity, despite using biobased reagents. Ozone depletion potential for this bio-based route is due to the use of trichloromethane in pesticides for potatoes, while runoff of fertilizers and pesticides causes higher eutrophication and ecotoxicity impacts compared to conventional chemical reduction methods. These results are also highly dependent on the temperature at which the reaction occurs. For comparison, the method described in Section was compared to another CR- Starch report which used an autoclave to increase the reaction temperature to 121 C at 15 psi (103.4 kpa). 149 Discussion of these results (Figure B9) can be found in Section 1 of Appendix B. This result emphasizes that the use of bio-based solutions will not necessarily result in reduction of

68 58 impacts on a life cycle basis and thorough analyses are required prior to adopting alternative synthesis methods Finally, reduction with sodium borohydride was shown to have the lowest impact levels despite the high impacts per unit mass from producing sodium borohydride compared to other reducing agents. Modeling sodium borohydride inputs in excess (3 stoichiometric requirements) does not noticeably change these results Syntheses Comparison by Particle Size/Function Using mass as a basis for comparison ignores performance differences among AgNPs produced through these various synthesis routes. Figure 3.1b shows the comparative results rescaled with respect to their size-dependent toxicity (LC 50) levels. Even though CR-TSC AgNPs had relatively low levels of impact per kg of synthesized particles, in Figure 3.1b they are revealed to have the highest impacts in all categories. As reported, CR-TSC AgNPs had larger diameters (100 nm) and thus smaller surface-to-volume ratios than AgNPs synthesized through other means, such that more of these particles are required in order to perform the same antimicrobial functionality as particles with smaller diameter (killing 50% of the bacterial population). High mass-based impact values for FSP and CR-Starch resulted in these routes being the second most impactful after rescaling by function. These rescaled results are sensitive to the average particle sizes assumed for each synthesis route. In reality, AgNP size can be controlled by varying reaction conditions or inputs, depending on the synthesis route. It may be possible for a single route to produce AgNPs of different sizes with only marginally different mass-based results but with appreciably different function-based results. Here we rely on a single size value for each route to demonstrate that function-based results (as all LCA results should be) shown in Figure 3.1b depend both on the environmental burdens of production and the bioactivity of the nanoparticles. These results are specific to antimicrobial products; results for other product categories would have to be rescaled separately according to their primary function.

69 59 100% 80% 60% 40% 20% 0% 100% 80% 60% 40% 20% 0% (a) (b) CR-EG CR-SB CR-TSC CR-Starch FSP RMS-AR-N AP OD GW PS AC EU HHC HHNC HHCR EC FF OD GW PS AC EU HHC HHNC HHCR EC FF Figure 3.1. Relative environmental impacts of multiple AgNP synthesis routes (a) TRACI 2.1 life cycle impact assessment method, (b) re-scaled impacts with respect to size-dependent bioactivity. Abbreviations: chemical reduction with trisodium citrate (CR-TSC), chemical reduction with sodium borohydride (CR-SB), chemical reduction with ethylene glycol (CR-EG), chemical reduction with soluble starch (CR-Starch), flame spray pyrolysis (FSP), arc plasma (AP); potential impact categories are ozone depletion (OD), global warming (GW), photochemical smog (PS), acidification (AC), eutrophication (EU), human health: carcinogens (HHC), human health: noncarcinogens (HHNC), human health: criteria air pollutants (HHCR), ecotoxicity (EC), fossil fuel depletion (FF) Analysis of Process Contribution In order to show which inputs and processes are contributing most to impacts in each category, detailed results are shown for two synthesis routes in Figure 3.2: FSP and CR-Starch, representing physical and chemical methods, respectively. CR-Starch is highlighted as well to explore the significance of using a bio-based reducing agent. Corresponding figures for all other routes are provided in Appendix B. For all routes, the contribution of the source of silver (in most cases silver nitrate) is much greater than the other chemical inputs. This is largely due to industrial activity involved in the mining and refining of silver and its transport (commonly by air). In terms of ecotoxicity, the impacts stem from result from leaching of mining tailings (Figures B10 and B11). These results mirror that found for gold nanoparticles by Pati et al. and suggests that life cycle

70 60 impacts from precious metal nanoparticle synthesis can be most effectively reduced by improving synthesis yields and recovering silver from spent solutions. Additionally, synthesis routes that use bio-based reducing agents but sacrifice nanoparticle yields may end up increasing environmental impact overall. 126 For the case of FSP, as seen in Figure 3.2a, an analysis of process contributions demonstrates that nearly 80% of the GHG emissions result from upstream industrial energy use, including combustion of coal for electricity and diesel for transportation; only 10% of global warming impact is due to the flame used in the synthesis process itself. For soluble starch reduction, Figure 3.2b illustrates that the contribution of the starch solution ranges from as low as 10% to carcinogenics up to 60% to ecotoxicity. It is important to note that, while potato starch is benign compared to more toxic chemical reducing agents, indirect toxicity impacts occur during potato cultivation from the runoff of pesticides and herbicides and their subsequent ecotoxic and eutrophic effects. Perhaps surprisingly, production of deionized water also results in noticeable impacts, particularly for ozone depletion from trichloromethane used in the production of anionic resins. For global warming potential and fossil fuel depletion, energy use contributes to more than 50% of the impacts. 100% Synthesis emissions Gaseous elements Electricity 100% Silver source Reagents DI water Heat 80% 80% 60% 60% 40% 40% 20% 20% 0% OD GW PS ACF EU HHC HHNC HHCR EC FF (a) 0% OD GW PS ACF EU HHC HHNC HHCR EC FF (b) Figure 3.2. (a) Process contribution for all TRACI impact categories of FSP synthesis route (silver source is silver octanoate, reagents are 2-ethylhexanoic acid and xylene, gaseous elements are oxygen and methane). (b) Process contribution for all TRACI impact ca

71 Comparison to Previous Results A previous screening-level LCA performed on AgNP-enabled socks also compared a flame spray pyrolysis and arc plasma method, as well as an undisclosed CR route, using an Economic Input- Output (EIO) LCA model. 98 It should be noted here that process-based (used here) and EIO LCA models are different in structure and scope. The former is a bottom-up engineering model while the latter is a top-down economic model, with production of goods aggregated into broad economic sectors. Unlike process-lca where chemical and energy inputs can be specified individually and in physical units, in EIO-LCA there is a single sector input for silver, gold, and other precious and non-ferrous metals. In spite of these model differences, comparison of our results with those of Meyer et al. show the same relative pattern across all impact categories: FSP having higher impact levels than AP and chemical reduction having the lowest impacts among synthesis methods Impact Assessment in Context of Products A final analysis was performed as an extension to previous work on nano-enabled bandages. 123 Previous findings for the specific AgNP synthesis route used for the Acticoat 7 product, RMS-AR- N, showed the upstream electricity production dominated the global warming impacts of synthesis and that mining emissions dominated ecotoxicity impacts. Here, we present the range of product-level results for these two impact categories when the exact synthesis route is not known. Figure 3.3 represents the absolute potential global warming and ecotoxicity impact results for a cm 2 piece of Acticoat 7 bandage (material components of this bandage can be found in Table B9). The contribution of AgNPs to product-level impacts is shown as the average for all synthesis routes, while the error bars represent the AgNPs produced through the most- and least- impactful synthesis routes. Physical synthesis routes directly deposit the AgNPs onto the substrate with no extra energy requirements. For the chemical synthesis methods, dip coating of the substrate by submerging it into the final AgNP solution was assumed. Pad-dry-cure techniques have been reported for coating textiles with nanoparticles. 150,151 Figure 3.3 illustrates dominant contribution of AgNPs to product life cycle impacts, regardless of how they are produced and despite being incorporated into the bandage at low loading levels. These results also show the sensitivity of the overall impacts of the bandage to the method of nanoparticle synthesis, again

72 EC GW 62 underlining the importance for product manufacturers and researchers of using synthesis-specific LCI data when conducting environmental assessments of nano-enabled products. HDPE fabric Polyester fabric Packaging AgNP Kg CO2 e CTU e Figure 3.3. Absolute environmental impacts of bandage production shown for GWP and ecotoxicity. Error bars demonstrate the high and low bounds for impacts depending on the production route. Abbreviations: potential impact categories are global warming (GW) and ecotoxicity (EC). AgNPs are projected to be an increasingly important material for a range of modern technologies. Anticipatory analysis of potential environmental and health impacts is prudent, while life cycle based results can uncover opportunities to improve efficiencies and lower impacts.

73 63 Chapter 4 The Life Cycle of Nanosilver-enabled Consumer Products: Investigating Hotspots Increasing use of silver nanoparticles (AgNPs) in consumer products as antimicrobial agents has prompted scientific research towards evaluation of their toxicity to higher organisms and their potential release to the environment. It has been shown that manufacturing of engineered nanomaterials can also pose significant burdens to the environment. Here a cradle-to-gate life cycle assessment for production of 15 different AgNP-enabled consumer products was performed. Environmental burdens were assessed over multiple impact categories defined by US- EPA s TRACI method. Results showed considerable environmental impacts associated with product manufacturing. Depending on the product composition, and silver loading, the contribution of AgNP synthesis to the overall impacts was seen to vary over a wide range, from 1% to 99%. Analyzing release studies showed the importance of AgNP incorporation methods. It was found that solid polymeric samples lost more silver during wash compared to fibrous materials. Direct ecotoxicity impacts of AgNP exposure from all the products were calculated as a worst case scenario, showing lower impact levels compared to total ecotoxic upstream production processes, a difference of 99%. Results of this study can be used to determine a threshold AgNP concentration at which direct impacts outweigh production, using a life cycle framework.

74 Introduction There has been an increase in the use of engineered nanomaterials (ENMs) in numerous sectors of the economy over the past decade, including energy, medical, electronics and an array of everyday household consumer products. 11 Larger surface area of nano sized particles results in increased surface reactivity, enhancing their physical and chemical properties compared to their bulk form, 1 deeming them suitable alternatives for improving product performances. In addition, incorporation of nanomaterials could be promising in terms of minimizing material use and subsequently waste at end of life of the product. 152 According to the Project of Emerging Nanotechnologies in Consumer Products Inventory, 14 as of April 2016, nearly 2000 products were listed as containing nanomaterials, 25% of which claimed to have silver nanoparticles (AgNPs). AgNPs have gained popularity among nanomaterials due to their unique physicochemical characteristics. 153 These nanoparticles particularly exhibit superior biological properties, 74 and are known for their broad spectrum antimicrobial function against Gram-negative and Gram-positive bacteria. 1,73,74 Contrary to popular belief, AgNPs have been in use for over 100 years in colloidal form for biocidal purposes, 72 and are currently being incorporated in health care products and textiles with an increasing rate. 1,11 Their antibacterial mode of action is believed to be through the generation of reactive oxygen species (ROS) during the penetration of silver ions through bacterial cell membrane. 73,74 A global material flow analysis for AgNPs was carried out based on the world economy in 2010, showing a production rate of nearly 500 tons/year. 11 The main applications of these nanoparticles were found to be in medical devices, coatings, textiles, and cosmetics. 11 To date, 442 AgNP containing consumer products were reported by the Project on Emerging Nanotechnologies to be readily available in the market. The list includes products such as linen and fabrics, tools and appliances, food packaging, healthcare devices, and personal care products such as toothpaste and cleansing bars. The inventory was recently categorized by various nano-specific aspects. Classifying nano-enabled products by the location of nanomaterial on the product showed for products containing metals or metal oxides, the particles were either suspended in a solution (colloidal form) or bound to the surface of the product. 154 It was also shown that antimicrobial protection was the most desired function of the products utilizing nanomaterials, especially those containing silver. 154

75 65 AgNPs have been receiving significant attention from toxicologists as a result of concerns with the environmental and human health risks caused by their potential unintended exposure. 32,75 In addition to the established bactericidal properties, 73,86,153,155 AgNPs have displayed cytotoxic effects towards higher organisms. 73 Researchers have already gathered evidence of this toxicity against higher cell lines such as zebrafish, rats, and humans, 73,116,156,157 in aquatic environments or as aerosolized particles in the atmosphere. 32 The toxicity of AgNPs were shown to depend greatly upon particle characteristics such as size, shape, and capping agent and environmental conditions. 73,115,158 Particles of smaller size exhibit higher levels of toxicity towards organisms. 114,115 It has also been shown that AgNPs interaction with gram-negative organisms is extremely shape-dependent, 159 and that surface charge can greatly affect the toxicity of these particles towards gram-positive bacteria. 160 It is yet unclear what causes this toxicity towards microorganisms, but it is speculated that the contributing factor to this process is the dissolution of AgNPs and generation of silver ions. 73,74,153,156 When tested under strictly anaerobic conditions, AgNPs showed no particle-specific toxicity and ion release was shown to be the toxic mode of action. 161 To make use of the unique characteristics of AgNPs, manufacturers have started incorporating these particles within the matrix of their products. The method of incorporation depends on the application of the product and the exposure level required for maximum product performance. Subsequently the rate of silver ion release is a function of the deposition method and the concentration of silver particles on the substrate. 162 For glass substrates, spin coating at different speeds, 163 and dip coating by simply immersing the substrate in solutions have been reported. 164 Various methods have been developed for embedding AgNPs in polymer matrices. Examples are, polymerization of polyethylene samples in the presence of AgNPs, 165 vapor deposition of particles on nylon 11 matrix followed by heat treatment, 166 simultaneous silver sputtering and plasma polymerization of an organosilicon matrix, 167 layer-by-layer assembly of AgNP containing films on polyethylene terephthalate subtrates, 168 melt blending of silver and polyethylene composites and laminating by extrusion, 162 homogenous casting of an AgNP-polyethylene suspension, 162 and deposition of an AgNP layer on polyethylene using a metallic sprayer. 162

76 66 To apply AgNPs to textiles, a variety of processes exist that either impregnate the particles within the fibers or distributes them as a coating. The application method is an important determinant of the final concentration of silver on fabrics after washing processes. Prior literature have assessed the relation between the incorporation method and amount of particle release by synthesizing AgNP-enabled textiles and analyzing the wash water. Perelshtein et al. used ultrasound irradiation to deposit AgNPs on nylon, polyester and cotton fabrics, and reported no loss of silver after 20 wash cycles as a result of the high temperature and speed of AgNP dispersion. 169 Another method reported by Radetic et al. used corona treatment, a pulsed high voltage electric discharge, to aid with silver adhesion onto fabric surface, and saw higher antimicrobial performance of the fabric before washing the samples. 170 Maneerung et al. developed a method to impregnate AgNPs in bacterial cellulose fibers using chemical reduction, and saw accumulation of particles at fabric surface, but signs of deeper particle penetration and slower silver ion release with higher concentration of the reducing agent. 171 Other reported methods are electrospinning AgNP containing gelatin fibers, 172 thermal reduction during melt processing of polyamide 6 fibers, 173 and pad-dry-cure techniques. 174,175 A recent study by Reed et al. looked into the after laundering antibacterial activity of four fabric samples with different silver integration methods, namely, covalently tethered, electrostatically attached, silver salt coated, and metallic silver coated fibers. 5 They concluded that the silver incorporation method and initial loading contribute mainly to the silver release levels during use. 5 Current literature on AgNP-enabled products have predominantly focused on their antimicrobial efficacy towards bacteria or higher organisms, or the release of particles under various conditions. Table 4.1 shows a list of studies on particle release from commercially available products. These experiments were performed so as to simulate particle leaching during use conditions. Quadros et al. studied the release of silver from products for children to various liquid media, and found higher silver dissolution levels in media with high salt concentrations. 3 Benn et al. also experimented on some AgNP containing products for home, and reported a variety of release rates. They observed no correlation between the amount of silver on the product and the released amount, for fabric samples. 2 In a study on AgNP-enabled socks, Benn and Westerhoff detected colloidal and ionic silver release from the samples, and suggested that the manufacturing process could be a cause of the different release rates from the socks. 4

77 67 Table 4.1. Previous literature on silver release from AgNP-enabled commercial consumer products. Reference Consumer Products Release Media Parsons et al., different wound dressings Deionized water Benn and Westerhoff, 2008 Benn et al., different socks Ultra-pure water Athletic shirt, unfinished cloth fabric, medical mask, medical cloth, toothpaste, shampoo, detergent, towel, teddy bear, humidifier Quadros et al., 2011 Anti-odor spray, surface disinfectant, throat spray Air Water, air, soil Huang et al., 2011 Food bag Ultrapure water, acetic acid, ethanol, hexane Walser et al., 2011 T-shirt Water Goetz et al., 2012 Food bags, food containers Water, ethanol, acetic acid, olive oil Quadros et al., 2013 Plush toy, baby blanket, sippy cup, breast milk storage bag Echegoyen et al., 2013 Food bags, food containers Ethanol, acetic acid Tap water, saliva, sweat, urine, orange juice, milk formula, HCl, saline solution, air, dermal wipes On a larger scale, release of AgNP to various compartments of the environment have been mathematically modeled to assess the risk and exposure levels. Using mass flow analysis, it was predicted that a significant contribution to silver environmental emissions will be made by using antimicrobial textiles and plastics. 22 Concentration of AgNPs in various environmental media have also been predicted using probability density functions showing promising results when faced with lack of nano-specific data. 63,176 To account for non-nano emissions upstream of manufacturing and during use of AgNP-enabled products, life cycle assessment (LCA) models have been developed to quantify the overall environmental burdens of the products from material extraction to end-of-life. Several studies have looked into quantifying the life cycle environmental impacts of AgNP-enabled textiles. A screening level LCA on an AgNP-enabled sock was performed, suggesting use phase as the main drive of environmental impacts. This study has also considered various AgNP synthesis techniques, showing the dependence of the impacts to the manufacturing method. 177 This was confirmed by another cradle to gate LCA of different physical, chemical, and bio-based AgNP synthesis methods, providing results based on function as well as mass, to reflect the performance level of AgNPs in a comparative analysis. 68 To assess the change in the impact profile of products due to the addition of AgNPs, a comparative cradle-to-grave LCA of a conventional T-shirt and an AgNP-enabled one was carried out, demonstrating higher impacts at the production level, mainly driven by silver mining. 62 Another cradle-to-grave study looked at the impacts of a commercial

78 68 AgNP-enabled wound dressing. 67 This study assessed the environmental impacts of nanomanufacturing, and found the impacts of production were caused by silver extraction processes. During production, significant use of electricity and subsequent emissions form power generation facilities were major contributors to global warming potential, while disposal of sulfidic tailings from silver mining contributed to the toxic impact to aquatic organisms. This study also found AgNPs to be the main contributor to the environmental impacts of the product manufacture, despite their low concentration (only 6% by weight). These impacts were due to the AgNP production process. Manufacturing of the bandage was also found to result in higher environmental impact levels compared to the incineration processes at its end-of-life. Therefore, for this AgNP-enabled product, indirect non-nano emissions upstream were found to be the predominant drive for impacts, compared to emissions and direct AgNP releases during disposal. This study aims to expand on the previous LCA studies on AgNP-enabled products, to determine the effects of AgNP incorporation on the impact profile of products. This is done by comparing indirect non-nano impacts of production to direct impacts of AgNP release. For this purpose, cradle-to-gate LCA of 15 different products utilizing AgNPs for their antimicrobial properties were modeled. The amount of bioavailable silver released from the products were quantified by simulating usage, and factors affecting particle separation were discussed. These AgNP release values were based on both experimental data and previous literature. Environmental impacts of silver release were then compared to product manufacture for each product. This comparison will ultimately allow for the determination of a threshold silver concentration; a relative weight percentage of AgNPs in a product that could cause significant environmental damage Methods and Modeling Scope and System Boundary This LCA considers the environmental impacts of AgNPs in the context of consumer products. A cradle-to-gate LCA was performed for 15 commercially available products to quantify the impacts associated with their production and determine the relative contribution of AgNPs to the overall impacts of the product. Products were chosen based on their availability in the market, some purchased for the purposes of this study, and some were based on prior release literature. The material composition and manufacturing steps of these products are explained in detail in section 2.2. Out of 15 products, six were purchased for the purpose of this study. Acid digestion and

79 69 release experiments were performed on product samples to determine the silver concentration and leaching respectively, described in section 2.3. For these products, if the concentration of AgNPs derived by acid digestion was lower than reported amounts, literature values were used for LCA modeling to project a worst case scenario. The remaining products were based on available data in literature Product Description Purchased Products Products obtained were readily available in the market. Samples from products were imaged using Scanning Electron Microscopy (SEM) to confirm the presence of AgNPs. Energy Dispersive X-ray Spectroscopy (EDX) analysis was also performed to configure all of the sample elements. For EDX, samples were mounted on an aluminum pin using a carbon coated adhesive, and sputtered with a coating of platinum prior to imaging. These images can be seen in Figure C1 in Appendix C. Product compositions were gathered from their labeling. Acticoat 7 (by Smith & Nephew) is a nanosilver-enabled wound dressing. This bandage contains 5 alternating layers, with the 2 outer layers and the inner most layer being a high density polyethylene (HDPE) mesh containing AgNPs, and the 2 remaining layers sandwiched between the HDPE layers are made of a nonwoven rayon and polyester fabric. 101 Presence of nano sized particles of silver were confirmed with SEM, Figure C1a. A cm 2 piece of this bandage was purchased for this study, and the amount of material use (HDPE and polyester) was calculated for this piece. The energy requirements for polyester spinning and knitting and fabric production were based on a previous LCA study. 62 These values were adapted to HDPE fabric production. Details can be found in Table C1 of Appendix C. All 5 layers of the bandage are reported to be ultrasonically welded together. This process was not modeled as part of our analysis due to the unavailability of data. Individual packaging slips were assumed for each bandage to be made of supercalendared paper. The amount of silver derived using acid digestion conformed to the amount reported in previous literature. 106 A summary of the inventory can be found in Table C1. Silvercel (by Systagenix) is a silver containing antimicrobial alginate dressing used on wounds with higher risk of infection. This dressing uses silver coated fibers (X-STATIC ), and upon further

80 70 inspection with SEM, Figure C1b, nano sized particles of silver were found on the strands. It has been reported that by weight, this dressing uses 36% high G calcium alginate fibers, 6% carboxymethyl cellulose (CMC), 28% silver coated fibers, and 30% ethylene methyl acrylate film for tensile support and combatting adhesion. 178 Acid digestion of the samples resulted in silver concentrations close to the reported amounts by literature, 178 and manufacturer (8% elemental silver with a sustained release formulation). The weight of the bandage was based on literature, 106 therefore, by using the given weight percentages, the weight of the bandage components were calculated. Individual packaging as mentioned earlier was assumed for each bandage. A summary of the inventory for a 5 5 cm 2 piece of bandage can be found in Table C2. Calcium alginate fibers are algae sourced biopolymers. 179 They are synthesized through the reaction of calcium chloride (CaCl 2) with sodium alginate, itself extracted from brown algae using a series of processes. Sodium alginate (NaC 6H 7O 6) is the sodium salt of alginic acid or alginate(c 6H 8O 6), a polysaccharide found in algae cells. 183 The extraction process was modeled in SimaPro using values from the literature, which can be found in detail in Table C2. Dried algae was soaked in 2% formaldehyde to remove its pigments, then added to a 0.2 M hydrochloric acid (HCl) solution before extraction with 2% sodium carbonate (Na 2CO 3). 183 It has been reported that generally, sodium carbonate is used in excess during the extraction processes, 184 but due to the unavailability of data, as a baseline, a stoichiometric value was used based on the reaction assumed below C 6H 8O 6 + Na 2CO 3 NaC 6H 7O 6 + NaHCO 3 Calcium alginate was then modeled based on the reaction below 2NaC 6H 7O 6 + CaCl 2 Ca(C 6H 7O 6) 2 + 2NaCl Reports on the synthesis of calcium alginate fibers have suggested use of microfluidic devices, where the reaction between CaCl 2 and sodium alginate takes place within the device. 180,182 Energy consumption of this process was assumed to be negligible. Algae growth and dewatering was based on an analysis done on the life cycle of algal fuels for the GREET model (Argonne National Laboratory, IL, USA). 185 Nitrogen and phosphorous nutrients were assumed to be consumed stoichiometrically. Carbon dioxide was incorporated by CO 2 capture of other plants. Energy use for delivery of this CO 2 was included in the model. Water use was estimated by calculating the replenished water after loss due to evaporation, as 95% of the water is recovered through the

81 71 growth process. Electricity consumption of pumps, centrifuges and dissolved air flotation processes were also included as part of this analysis. Detailed inventory can be found in Table C2. Aquacel Ag (by ConvaTech) is another silver impregnated wound dressing using Hydrofiber technology. Product specification reports 12 mg of silver per gram of dressing. 186 Acid digestion results showed silver concentrations close to this value as well as the amount reported in previous literature. 106 The silver was reported to be in ionic form, which was confirmed by the SEM imaging, Figure C1c, showing silver chloride (AgCl) crystals on the fibers. Silver chloride and silver nitrate were modeled using the reactions below AgNO 3 + NaCl AgCl + NaNO 3 3Ag + 4HNO 3 3AgNO 3 + 2H 2O + NO Hydrofiber technology uses carboxymethyl cellulose fibers held together using needlebonding. 186 Individual packaging as mentioned earlier was assumed for each bandage. The weight of the bandage was based on the literature, 106 and was adjusted for the cm 2 piece of bandage modeled for this study, with inventory shown in Table C3. Polymem (by Ferris Mfg. Corp.) is a polyurethane foam membrane impregnated with small particles of silver used for wound management. SEM imaging showed nano sized particles of silver embedded in the foam matrix, Figure C1d. Silver concentration derived from acid digestion conformed to findings of prior literature. 106 Individual packaging as mentioned earlier was assumed for each bandage. Detailed inventory for a cm 2 piece of dressing can be found in Table C4. GoGreen Food Container (by Kinetic) is a nanosilver-enabled food storage with dimensions of cm 3 (0.76 QT or 860 ml) made from polypropylene. SEM imaging showed no nanoparticulate of silver on the surface of the container and acid digestion showed only trace amount of silver for the samples. Therefore, the silver concentration for LCA study was based on another study on the same container, showing 0.32% silver by product weight. 187 The product was weighed to determine the amount of polymer used. Injection molding of the polymer was assumed for forming the container. Details of the inventory can be found in Table C5.

82 72 Silver containing socks #1 (by Fox River Mills Inc.) use the X-STATIC technology in their fiber blend to provide antibacterial and anti-odor features. According to the product specification, the socks contain 60% polypropylene, 20% nylon, 19% X-STATIC nylon, and 1% spandex. Similar to Silvercel dressing which used the same silver fiber technology, nano sized particles of silver were found on the sock fibers using SEM, Figure C1e. Energy and material requirements for nylon and polypropylene spinning and fabric production were adapted from data in previous literature, 62 and information in ELCD 2.0 database (JRC, EPLCA). Details can be found in Table C6. Silver concentration found using acid digestion was significantly greater than the amount reported in the literature. Here, the values based on published data were used to model this product Products Based on Literature From a study done by Quadros et al. on leaching from AgNP containing products, 3 products were chosen. 3 Silver concentration reported for all products was derived using thermally assisted nitric acid digestion. The weight of all the products were based on measurements. Chosen products are as follows: Baby blanket (by Baby Pink or Blue Ltd.) with AgNPs embedded in the products. It was assumed that the blanket is composed of polyethylene fleece. Detailed inventory can be found in Table C7. Children s cup (by Baby Dream Co. Ltd.) using AgNPs for maintaining freshness and deodorizing purposes. Injection molded polypropylene was assumed to be used for this product. Details can be found in Table C8. Plush toy (by Pure Plushy Inc.) infused with AgNPs to protect against bacteria. The interior foam and exterior fur were inspected for presence of AgNP, showing considerable amount available within the filling of the toy. This part was modeled assuming a polystyrene filling. Details can be found in Table C9. Another study by Benn et al. looked into AgNP release from consumer products used at home. 2 From the list of products in this study, 3 medical related products were chosen. Silver concentrations were found using nitric acid digestion of samples. Product weights were reported in the study. Product descriptions are as follows: Medical mask (by Nanbabies) containing silver to work against bacteria and fungi. It was assumed the mask is composed of polypropylene fibers. Details can be found in Table C10. Medical cloth (by Nanbabies) with silver was assumed to be made from cotton. Details can be found in Table C11. Towel (by Good4U) with AgNPs was assumed to be composed of a fiber blend of 80% polyester and 20% nylon. Details can be found

83 73 in Table C12. Athletic T-shirt (by Puckskin) containing nanosilver made from polyester. Detailed inventory can be found in Table C13. From a study on release from silver containing food containers by Echegoyen et al., one product was chosen. 187 Concentration of silver was determined through calcination in a muffled oven followed by nitric acid digestion of the remaining ashes. Product description is as follows: FreshLonger plastic bag (by Sharper Image ) is an AgNP infused re-sealable bag for food storage to prevent food from spoiling fast. It was assumed that the bag is produced from LDPE using plastic film extrusion. Detailed inventory for a cm 2 can be found in Table C14. Thickness of each sheet of the bag was assumed to be 2 mils (1 mil is a thousandth of an inch, equal to mm). AgNP release from sock fabrics was studied by Benn and Westerhoff, for several commercially available socks. For comparison with the purchased sock, one was chosen from this study with different fiber composition and silver concentration. Silver containing socks #2 (by Arctic Shield) uses X-Systems AgNP fiber technology for odor protection. These socks are composed of 50% cotton, 39% polyester, 6% nylon, and 5% spandex. Silver concentration was determined through a heated nitric acid digestion process. Product weight was determined by the study. Detailed inventory for the product can be seen in Table C Experimental Procedures To determine the total silver content of products, digested triplicate samples from each product were ashed by adapting methods used by Echegoyen and Nerín, Tulve et al.. 187,188 Pieces weighing 3 10 mg from each product were cut and placed into 2-ml glass vials. The vials were then placed into a tube furnace (Thermo Scientific Lindberg Blue M Mini-Mite) at 600 C until samples were completely combusted (3 10 min). Then ash-containing vials were filled with nitric acid (HNO3, 70% ACS certified, Spectrum) and heated to ~100 C for 4 hours, with additional nitric acid added over time to keep vials full. After this period, samples were removed from heat and let to cool. Amount of 1 ml hydrogen peroxide (30% ACS certified, Fisher) was added and samples were diluted to a volume of 250 ml with ultrapure water. Samples were analyzed for silver concentration by inductively-coupled plasma mass spectroscopy (ICP-MS, X-Series, Thermo Electron), which had a calibration curve for Ag ranging from ppb.

84 74 To mimic a situation in which silver is released from products during contact to skin or a wound during 3 days, a method used by Wu et al. was adapted. 189 Circles with a diameter of 10 mm were cut from each product and immersed into 9 ml of phosphate buffered saline (PBS 1X, Ricca). Samples were placed in a water bath at 37 C for 3 days, then filtered (450-nm, Teflon), acidified with the addition of 1 ml HNO3 (70%), and analyzed by ICP-MS Life Cycle Assessment Modeling Life cycle impacts were modeled using the SimaPro 8.1 platform (PRé Consultants, Amersfoort, the Netherlands). Material and energy inputs were associated with their respective unit process in the US-EI database (Earthshift, Huntington, VT), the Ecoinvent database adjusted for U.S. energy grid system. Environmental impacts were assessed using the U.S. EPA s Tool for the Reduction and Assessment of Chemical and Other Environmental Impacts (TRACI) 2.1 method. 57 This method considers ozone depletion (OD) in kg CFC-11 equivalents (eq.), global warming potential (GW) in kg CO2 eq., photochemical smog formation (PS) kg O3 eq., acidification (AC) in kg SO 2 eq., eutrophication (EU) in kg N eq., human health impacts from carcinogenic and noncarcinogenic substances (HHC and HHNC) in health comparative toxic units or CTUh, human health impacts from criteria air pollutants (HHCR) in kg PM2.5 eq., and ecotoxicity (EC) in CTUe, and fossil fuel depletion (FF) in MJ surplus. Due to the uncertainty in the method of AgNP production for products, a single synthesis route was assumed for all products for consistent comparison. Chemical reduction of silver nitrate (AgNO 3) with trisodium citrate was the method chosen for this study. Effects of using different synthesis routes is illustrated as a sensitivity analysis in Figure C2 and described in result section. Life cycle inventory data for all production methods can be found in our earlier work Results and Discussions Analysis of Process Contributions Life cycle impact assessment was performed on a product basis with relative results illustrated in Figure 4.1. Contribution of silver to total impacts covered a wide range, in general, from almost 1% up to 99%. In general the environmental impacts of AgNPs were found to be due to the electricity consumption and emissions of the silver mining processes, consistent with previous literature. For all products, eutrophication and human health non-carcinogenic impact categories

85 75 show considerable contribution by the AgNPs. Further investigation showed this contribution is a result of the sulfidic tailings and emissions from silver mining processes. For products such as the baby blanket, T-shirt, and towel, due to the low concentration of silver particles, AgNP impacts were overshadowed by the polymeric components of the products. These impacts were driven by the production of polyester or polyethylene terephthalate (PET) fabric. Electricity consumption during the spinning and knitting processes and the emissions related to the power source were the contributing factors to the global warming potential, ozone depletion, smog, and acidification impact categories. This significant electricity use was also the source of disposal of spoils from coal mining to landfills downstream, which in turn contributes to the eutrophication and carcinogenic impacts. Natural gas consumption for electricity generation and the emissions of combustion were found to be driving the non-carcinogenic and respiratory effects. Use of xylene as a precursor to terephthalic acid was found to contribute to the ecotoxic impacts in addition to disposal of incineration residues of PET. Fossil fuel depletion was due to natural gas extraction for electricity generation and use of xylene for producing terephthalic acid and ethylene for producing ethylene glycol, both used in manufacturing PET granulates. For sock #2, silver contribution was also negligible compared to the other components. For this product, cotton based fibers dominated the overall burdens, due to upstream impacts from cotton cultivation and electricity consumed during cotton yarn and fabric manufacturing. In ozone depletion, global warming, smog and acidification categories, electricity consumption during cotton weaving processes and yarn production, and emissions associated with electricity generation were found to be the source of impacts. Another contributing factor to ozone depletion was the use of trichloromethane as pesticide for cotton. Eutrophication and ecotoxicity impacts were driven by pesticide and fertilizer runoff. Human health impacts, carcinogenic, noncarcinogenic and respiratory, were caused by disposal of coil mining spoils and hard coal ashes during electricity generation, and emissions from the process of applying plant protecting products. Fossil fuel depletion impacts were due to natural gas extraction and coal mining for electricity generation. Compared to sock #2, sock #1 had higher silver concentration therefore, contribution of silver was higher. This ranged from 9% increase in the global warming category, reaching up to 70% increase

86 76 in the total impacts in a category such as non-carcinogenic human health effects. Sock #2 weighed more than #1, and was predominantly composed of cotton fabric, which compared to polymer based textiles is significantly more resource intensive on a per kg basis, except for fossil fuel depletion category, as it is the main source for polymeric products. In the case of plush toy, the process of styrene production and its ultimate disposal in municipal incineration plants contributed significantly to the overall impacts. Eutrophication and noncarcinogenic health impacts of this product are also affected by the emissions of sulfidic tailing from mining silver. In the ozone depletion category, the contribution of AgNPs are a result of heating processes during manufacture. Low concentration of AgNPs in the children s drinking cup resulted in negligible impacts, compared to the electricity used during the injection molding process, except for the ecotoxicity impact category which was dominated by the disposal and incineration of the polypropylene, and the fossil fuel depletion impact category as a result of crude oil consumption. Process of injecting molding plastics require more energy than transforming them into films. Therefore, higher impacts can be seen due to these processes in the food container than the plastic bag. Fossil fuel depletion categories of these products (food container and plastic bag) are dominated by the polypropylene production, as explained earlier. Medical cloth and mask were the products with the highest concentration of silver, as a result all the impact categories are dominated by the impacts of AgNPs. Source of impacts for ozone depletion, global warming, and fossil fuel depletion was found to be the heating component during particle synthesis. Contributing factor to smog and acidification impacts were emissions from blasting processes during silver extraction and refining from combined metal mining. Sulfidic tailing disposal was found to be key factor affecting the eutrophication, carcinogenic and ecotoxic impacts. Emissions from refining copper ores and desilverizing lead were the main causes of noncarcinogenic human health impacts. The same processes contributed to respiratory effects in addition to emissions from natural gas production. Between the wound dressings, contribution of AgNP inclusion was highest in Acticoat 7. Silvercel dressing has the same concentration of AgNPs as Acticoat 7, but due to the use of calcium alginate fibers, the contribution of silver was from 10% (in eutrophication or non-carcinogenic impact categories) to 60% (in ozone depletion) less. Use of precursors such as chlorine and formaldehyde

87 77 drive the contribution of these fibers. For Aquacel Ag and Polymem with lower AgNP to product weight ratios, impacts were driven by the production of CMC fibers and polyurethane foam respectively. Polyurethane impacts are driven by upstream impacts of consuming toluene diisocyanate and polyols. Figure 4.1. Process contribution to environmental impacts of different products. Components were categorized generally as paper packaging, nanosilver, extrusion and molding processes for plastics, cellulosic and cotton fibers, and polymers. Concentration of silver is shown in parentheses in percentage to total product. Products are listed in ascending silver concentration from the top left corner with the lowest percentage of AgNPs, to the bottom right with the highest amount of AgNPs. Abbreviations of TRACI s potential impact categories: ozone depletion (OD), global warming (GW), photochemical smog (PS), acidification (AC), eutrophication (EU), human health: carcinogens (HHC), human health: non-carcinogens (HHNC), human health: criteria air pollutants (HHCR), ecotoxicity (EC), fossil fuel depletion (FF). In general, it can be concluded from the manufacturing results that, products with silver concentrations less than 0.004% show minimal to negligible impacts associated with AgNPs themselves. In products with concentrations between %, the contribution of AgNP was highly dependent on the composition of the products and the amount of other components being

88 78 used. For AgNP loadings %, it was seen that AgNP impacts dominated all the categories. On a product basis, an AgNP concentration of 10% or above has resulted in a significant contribution by these particles to the overall impacts of the products. Therefore, additional attention is required when evaluating the environmental performance of these products. Figure C2 examines the relationship between the concentration of silver and the percentage contribution of AgNPs to the overall impacts of the product. It can be seen from this figure that high levels of correlation exists between the silver concentration and the contribution to fossil fuel depletion and global warming potential impact categories. This is due to the fact that reduction of AgNPs with trisodium citrate required external heating. Additionally, AgNPs contributed to the smog impact category from the blasting processes during silver extraction, thus, the higher the silver concentration, the higher the contribution to this impact category. Hence, the relatively good correlation between these two components as shown in Figure C3. The remaining aquatic impact categories showed lower correlation levels, especially eutrophication and noncarcinogenic impacts. The contributions to these impact categories from silver were present regardless of the silver concentration, as a result of the sulfidic tailings, but the ratio was highly dependent to the composition of the product. These correlations are highly dependent on the method of particle synthesis and can change by considering other energy inputs or precursors. The contribution of AgNPs to the overall impacts of the product can greatly vary by considering a different synthesis method. Figure C3 demonstrates a sensitivity analysis performed to quantify the range of potential environmental impacts of AgNPs, using methods described in our earlier work. 68 Results in Figure C3 are presented on the basis of producing 1 kg AgNPs. For categories such as ozone depletion, global warming and fossil fuel depletion, the contribution of AgNPs could drop as much as 80% or increase by 20% by using other synthesis routes. In eutrophication, carcinogenic, non-carcinogenic, and ecotoxicity impact categories, the chemical reduction with trisodium citrate represents the lower bounds of potential impacts, but using other methods could result in a 300% increase of AgNP specific impacts. Impacts of acidification and respiratory effects can decrease by 40% or potentially increase by almost 100%. Smog impacts can range from 20% lower or 230% higher levels as a result of change in the production method. These results depend on the precursors used during AgNP synthesis, the heating requirements and electricity consumption. The diverse results indicate the need for using product specific synthesis data to assess the environmental impacts of nano-enabled products.

89 Analysis of AgNP Release Our experimental release studies were performed using a phosphate buffered saline solution to maintain a constant ph and simulate contact with skin. Therefore our release results differed to the ones already existing on the purchased materials. Figure 4.2 demonstrates the results of our study in addition to the release experiments performed on the remaining products, with respect to the AgNP loading on the product. The concentrations are depicted as those derived by the acid digestion. The axes are shown in log-log format. Release from the Aquacel Ag samples were the greatest due to the characteristics of the Hydrofiber technology, which dissolves in contact with solutions, enabling high releases of AgNPs. Among the products, non-textile polymeric samples showed higher amounts of AgNP release, suggesting more surface bound particles compared to fibrous material. This emphasizes the importance of the method of AgNP incorporation in the amount released from products. Silver released to aquatic environments can be a source of ecotoxic impacts due to release of free silver ions. The highest silver release amount is from the medical cloth, releasing 0.16 mg silver during wash. As a scenario for direct impacts from AgNP release, ecotoxicity impacts of dissociation of 0.16 mg AgNPs in aquatic environments were calculated. For a dissociation of 50% and USEtox Ag + characterization factor of CTUe/kg, 67 an ecotoxicity impact of CTUe was calculated, which is a significantly lower value than the ecotoxicity of the upstream processes of the same product (2.11 CTUe, see Table C16). For a worst case scenario of release from all products, the total amount of leached AgNPs add up to 0.95 mg, and results in a total ecotoxicity impact of CTUe. This is significantly less than the ecotoxicity of the manufacturing processes combined, and only slightly greater than the three lowest impactful products in this category: Aquacel Ag, Silvercel, and the plastic bag (0.009, 0.074, and CTUe respectively, see Table C16).This suggests that the inclusion of upstream environmental impacts from production is detrimental when considering the life cycle impacts of AgNP-enabled products. Additionally, looking at Figures 4.1 and 4.2, products with silver concentrations less than 0.03% AgNP, showed insignificant nano-specific impacts during production, despite high AgNP release rates (compared to other products under study) ranging from %. This illustrates the complementary nature of LCA models and release studies and the importance of including both in the assessment of environmental impacts of products.

90 80 Figure 4.2. Result of different release studies on the products considered in this study. Due to the differences in the experimental methods, they are shown in varying colors. Axes are in log-log format Larger scale mathematical models have been previously developed to determine the probability of AgNP release from products, or their concentration in different compartments of the environment. Using particle flow analysis concepts, Arvidsson et al. developed estimations for AgNP emissions during use of wound dressings, textiles and electronics, resulting in 10% release from wound dressings and up to 100% release from textiles and electronics. 85 Keller et al. utilized global market information in conjunction with material flow analysis tools to determine the likelihood of nanomaterial emissions to the environment and landfills. 11 They compiled estimates for release of particles during use from available studies of different nanomaterials. 4,18,22,24,63,190 For AgNPs, they used a high estimate of 25% release from medical products, and almost 100% release from textiles to waste water treatment plants. Our analysis shows up to 5% AgNP release from AgNP-enabled wound dressings, complying with results of Arvidsson et al., but significantly lower release rates from AgNP-enabled textiles. This is due to the assumption made by prior studies of 100% release of particles from textiles during use. Our results show this percentage is highly dependent on the fabrication method of the textiles, and the silver incorporation techniques. This is supported by a recent work by Reed et al., showing a 3-80% AgNP release range, after four wash cycles, for four different textiles. 5

91 81 AgNP-enabled products comprise the majority of nano-enabled product market. The results of this study quantify the life cycle environmental impacts associated with the production of these products and their potential silver emissions during use. As shown in this study, in addition to our previous work, 67 impacts associated with indirect and non-nano emissions upstream dominate the overall environmental burdens of products, and outweigh direct nano-specific impacts. Release experiments can help determine the concentration of AgNPs left on the product at endof-life, thus enabling more accurate modeling of silver concentrations reaching landfills or waste incineration plants. Additionally, impacts of products should be assessed in the context of their function. By understanding the burdens imposed by every aspect of the life cycle of these products, opportunities for improvement on technology can be more easily identified. Information on contribution of silver to environmental impacts and release from the product matrix, can assist in identifying the minimum AgNP concentration for optimal product performance, as well as the desired level of impregnation (surface coating or subsurface embedding) for a more controlled release, therefore minimizing impacts over their life cycle and on a function basis.

92 82 Chapter 5 Life Cycle Energy Benefits of Carbon Nanotubes for EMI Shielding Application This study has been submitted to Journal of Cleaner Production Carbon nanotube (CNT) composites have been developed for use as electromagnetic interference (EMI) shielding in satellites as well as other land-based applications. While CNT composites are costly and energy-intensive to produce relative to existing EMI shielding materials, they may offer both economic and energy benefits when considered on a life cycle basis. CNTs unique characteristics and enhanced physical and chemical properties have made them preferable as alternatives to conventional material. In comparing nano-enabled products with existing technologies, it is imperative to account for differences not just in production but also in performance and delivery. In this study, we considered CNT polymer composites as a replacement for aluminum (the current dominant EMI shielding material) on the basis of shielding effectiveness (in db). Results demonstrate that CNT-enabled composites offer opportunities for more than 50% reduction in the mass of EMI shielding, while maintaining performance levels within ±10 db of the shielding effectiveness of aluminum. Reduction in the mass of the EMI component of the satellite subsequently affects the total fuel required to launch it into orbit, reducing it by 6% when using CNT composites. Results show nearly 30% reduction of life cycle cumulative energy demand for the CNT polymer composite shielding compared to aluminum, with concomitant environmental benefits.

93 Introduction Concerns regarding the safety and potential environmental impacts of nanoparticles have been increasing with the rapid growth in the nanotechnology market. 191,192 Such concerns can be addressed through prospective assessments of ecological and human health risks, as well as consideration of occupational hazards and environmental sustainability. 193 A holistic perspective on benefits, risks, and potential impacts of engineered nanomaterials (ENMs) to health and environment can also guide technology development, and many have suggested the use of life cycle assessment (LCA) for these purposes. 40,47,191,192 The LCA framework is more expansive than traditional risk analysis as it considers both direct emissions of nanomaterials as well as (nonnano) emissions of substances that result as a consequence of ENM production, use, and eventual disposal. 71 LCA quantifies both positive and negative effects, and so considers environmental benefits of ENM incorporation in products against resource- or emissions-intensive processes in the ENM life cycle that could potentially offset or negate those benefits. Therefore, LCA is an ideal method when the assessment goal is to evaluate unintended consequences and potential tradeoffs between benefits and impacts to the environment and public health LCA Concepts for CNTs The market for carbon nanotubes (CNTs) is one of the fastest growing among ENMs, 61 estimated to reach $6 billion by the end of Due to their high electrical and thermal conductivity and superior mechanical properties, 195 CNTs have found various applications in electronics, composites, coatings, sensors, automotive and aerospace industries. 11,196 Tests on CNT composites have shown an increase in the intrinsic material properties, such as modulus, strength and conductivity (both thermal and electrical), compared to the original polymer matrix, 197 due to the high aspect ratio of the filler CNTs. 197 To assess the tradeoffs of these particles, to date, a handful of studies have performed LCAs on CNT manufacturing and CNT-enabled products. 48,69,120 These studies have mainly set cradle-to-gate boundaries for their analysis and have shown that CNTs have significant mass-based environmental burdens due to energy intensive manufacturing processes. 48,69,199,200 Despite the extensive use of energy during nanomanufacture, accurate estimations of the contribution of nanocomponents, to environmental impacts, relies on defining the functional unit of the study. 201 Therefore, fair comparison of products is only made by quantifying the benefits of nano-enabled products during

94 84 use, 202 while accounting for implications of energy and resource requirements of manufacturing and particle release at end-of-life. 48,61,191 The assessment of product benefits is highly dependent on the product application and the concentration of particles in the product. 61 Most impact categories in LCA determine the environmental burdens associated with a process or product. A series of life cycle studies for carbon nanofibers (CNFs) was previously conducted, which not only determined their energy consumption and environmental impacts with a holistic outlook, but also accounted for savings of using this nanotechnology compared to the materials for which they substituted. 7,203,204 LCAs on CNTs and CNT-enabled products have predominantly attempted to quantify the environmental impacts, 69,198 or the energy use during manufacture. 48 Despite acknowledging the importance of evaluating environmental benefits of nanotechnology, 48,61,191 only a few studies have looked into energy tradeoffs, 48 and many have overlooked accounting for the technological advantages from superior properties that could not otherwise be achieved by using the conventional materials. Nanoparticle benefits have commonly been defined in terms of differences in the environmental impacts or energy use. For instance, an environmental assessment on nanoclay polymer composites performed by Roes et al. (2007) compared the difference in their energy use and environmental impacts to conventional materials. 205 Another study on CNF polymer composites by Khanna and Bakshi (2009) looked at their application in body panels of automobiles, comparing life cycle energy savings of this technology with steel. 7 Few LCAs have accounted for the benefits gained due to the function of the nanomaterial, and the primary motivation for its inclusion in the product. Examples of such benefits include: fuel savings due to reduced weights of nanoclay polymer composites, 206 improved functionalities of a CNT-enabled switch, 198 enhanced performance of chemical gas sensors and their subsequent human health benefits, 120 and weight savings from use of CNFs in turbine blades, 204 each compared against current technologies being employed CNTs for Electromagnetic Interference Shielding One of the primary applications of ENMs has been in electronics and optics. 20 Electromagnetic interference (EMI) of radio frequency radiation with electronic appliances is a major concern that has grown with increasing dependence on electronics and telecommunications. 196, EMI

95 85 shielding refers to the protection of electronic compartments against electromagnetic radiation. 207,210 Shielding against EMI becomes even more critical in highly sensitive applications such as aerospace control and communication systems. 210,211 For instance, significant EMI shielding is required for satellites at launch and during orbit to avoid any malfunction in appliance performances. 210 Satellites at low earth orbits (LEO) are far more likely to encounter EMI due to solar winds and cosmic rays in space environments, and accumulation of charged particles from Van Allen belts on the probe s surface. 212,213 Metallic plates are most commonly used as EMI shields. 196,207,214 Their shielding mechanism is mainly by reflection of the electromagnetic radiation rather than absorption. 207 Metals typically used for shielding include aluminum, nickel, stainless steel, brass, and a highly permeable mumetal (a nickel-iron alloy). 215 For spacecraft applications metallic boxes provide the shielding function by acting as a Faraday cage. 211 Additionally, reinforcing honeycomb structures of the satellite comprised of an aluminum foil surface can also act as an EMI shield. 197 Metallic shields have the disadvantage of being heavy, prone to corrosion, with low mechanical flexibility. 209 To compete with the high density of sheet metals, at times metal polymer composites with steel fibers or nickel coated carbon fibers have been used for shielding. 211,216 For satellite applications, any savings in mass directly influences fuel consumption, thus if polymer composites with lower densities than pure metallic sheet could yield technical performance with reduced mass, then significant fuel savings could result. 211 In a recent effort, NASA developed a series of roadmaps and priorities for their future improvements in space technologies. 217 A high priority was given to research on structural and material weight reductions, as it could significantly improve the performance of the launch vehicles. 217 Additionally, the implications of nanoscale technologies were explored, recognizing their influence on material performance, and prioritizing the development of lightweight structures by ENM incorporation. 217 Processing advantages, flexibility, corrosion resistance, and reduced mass are the main reasons for using conductive polymer composites. 196, ,218 CNTs are favorable as a filler material for composites as they possess high electrical conductivity, high aspect ratios, and exceptional mechanical properties. 196,208,209,218 Many studies have looked into the shielding effectiveness (SE) of CNT polymer composites, showing promising results, some comparable to the SE of metals. SE is a function of the rate of the incoming wave to the transmitted wave by the shield, measured in

96 86 decibel (db) units. 216 Table 5.1 summarizes the SEs for CNT composites, where for the same type of nanocomposite, higher CNT loadings results in higher levels of SE. Only the composites developed by Rawal et al. (2013) perform at the same SE levels of conventional shielding materials such as aluminum with SE of ~70-80 db, 197 and pass the minimum SE requirements of NASA, which is 40 db, as well. 219 Table 5.1. Current CNT polymer composites technologies and their SE. Reference CNT Type Polymer Type Thickness (mm) Al-Saleh et al., 2009 CNT loading% SE Average Range (db) MWCNT Polypropylene vol.% X-band vol.% vol.% >24 Frequency Range* Yang et al., 2005 MWCNT Polystyrene wt.% X-band Huang et al., 2007 Long SWCNT Epoxy wt.% 2-25 X-band Short SWCNT wt.% Annealed SWCNT wt.% 1-21 Wang et al., 2012 PES-MWCNT Polyether ether wt.% X-band ketone Rawal et al., 2013 CNT sandwich panel M55J/Cyanate NA 82 X-band CNT co-cured sheet Ester Zhang et al., 2007 MWCNT Polyurethane wt.% 4-16 K-band shape memory wt.% 5-35 polymer Thomassin et al., MWCNT Polycaprolactone vol.% Microwave 2008 foamed Polycaprolactone solid vol.% *X-band ( GHz), K-band ( GHz), Microwave (25-40 GHz) Considering the benefits of using CNT-enabled polymer composites for satellite applications, namely reduction of shielding mass and the subsequent fuel efficiency, we seek to compare these composites to the conventional metallic sheets on a mass and net primary energy demand basis. Assuming communication satellites launched to LEO, savings and benefits are to be quantified on the basis of EMI shielding function for this specific application, i.e., effectively shielding the satellite from interferences in X-band frequencies ( GHz) Methodology Here we quantify the reduction in EMI shielding mass by using CNT polymer composites compared to conventional EMI shields in satellites, the fuel used for each scenario during launch, and the savings in total primary energy use upstream. This analysis focuses on the net energy savings of incorporating CNTs in satellites in the context of their function as EMI shields. Except for energy use, other ecological impacts associated with CNT manufacturing are not quantified in this

97 87 analysis but they can be found in previous LCA studies Satellite applications are interesting in that there are no direct CNT emissions during the use phase that would affect the terrestrial environment. Additionally, product disposal is not relevant for satellites as during their deorbiting, they either burn at re-entry, or are sent to a graveyard orbit at a higher altitude. 220 This simplifies the scope of our analysis to comparing the primary energy used for the EMI shields and their respective fuel requirements, assuming no change in other components of the satellite. The analysis breaks down into three parts: 1) calculating the mass of conventional and CNT composite EMI shield for satellites, 2) calculating the fuel that is required for launch, and 3) determining the primary energy used for each shielding scenario. In order to calculate the mass of EMI shield on a satellite, a breakdown of satellite components by mass is required. Tsiolkovsky s rocket equation can be used to determine the required fuel for launch, and finally, having the mass of the shields and the fuel, the total primary energy demand of these components can be calculated and compared for various shielding scenarios. These steps are explained in detail in Sections through For satellite EMI shielding components, the cumulative energy demand (CED, or primary energy) over the life cycle of the satellite was modeled using the CED 1.08 model as implemented in SimaPro 8.1 software (PRé Consultants, Amersfoort, the Netherlands) and the primary embodied energy of materials modeled in CES EduPack 2015 (Granta Design Limited, Cambridge, United Kingdom). The energy use of components was linked to the respective unit processes in the US-EI life cycle inventory database (Earthshift, Huntington, VT), the ecoinvent database adjusted for U.S. energy grid inputs. Uncertainty analysis for all of the inputs to the model was performed using Monte-Carlo simulation over 10,000 iterations in SimaPro 8.1, with a default log-normal distribution for the US-EI database unit processes. Monte-Carlo simulation was also used to assess the uncertainty associated with the effect of change in satellite mass to the final results, utilizing software (Palisdale Corporation, Ithaca, NY). Section shows the results of the uncertainty analysis Analysis Results EMI Mass Calculation For the purpose of this study, communication satellites launched to lower earth altitudes were considered. It has been reported that typically 20% of the mass of the communication and electrical systems of satellites can be made up of EMI shielding. 211 Based on available data for satellite component breakdown by mass, for communication satellites an average of 23% of the

98 88 satellite s mass is dedicated to the communication systems, and nearly 27% consists of the electrical and power systems. 221,222 To assess the wide span in the range of mass for satellites, two satellite sizes were assumed. First, an extremely light sample from the ORBCOMM network of satellites was chosen, with a launch mass of 45 kg, from an existing database. 223 On the high end of the range, the Iridium NEXT satellite launched by Orbital ATK Inc. was considered, weighing 860 kg at launch. 224 The launch mass of a satellite is the dry mass of the satellite, which includes the propulsion systems, a total of 93% of the launch mass, and the propellant required for orbit transfer and insertion. 221 By calculation, 4.9 kg and 93 kg of EMI mass out of the total launch mass of the satellite is assumed for low and high satellite scenarios, respectively. The conventional method of shielding requires an aluminum sheet with a thickness of 2 mm. 211 With the total mass of the shield, its thickness, and the density of aluminum (2.643 g/cm 3 ), the area covered by the EMI shielding was calculated to be 17.6 m 2 and 0.92 m 2 for the high and low cases, respectively. When replacing the sheet of aluminum with a lighter composite, the technological performance must be maintained at the same level as aluminum for shielding. The performance of EMI shields can be quantified in terms of SE (with units of db). For an aluminum sheet thicker than its skin depth (the depth at which the amplitude of the wave has diminished to approximately e -1 of its initial value), 216 the primary mechanism of shielding is by reflection. The skin depth is defined as: δ = ( πfμσ) 1 (1) where δ is the skin depth (m), f is the frequency of the wave (Hz), μ is defined as the magnetic permeability (H/m), and σ is the electrical conductivity (S/n) of the shield. These parameters are intrinsic to the material. For an aluminum sheet at the X-band frequency (8-12 GHz), the average skin depth is calculated to be 0.8 µm (μ= (H/m), σ= (S/m)). At the current 2 mm thickness of the aluminum sheet, it is safe to assume that the only mode of shielding is by reflection. A simplified formulation for calculating the SE by reflection is as, 216 SE R = log σ 2πfμ (2) For aluminum in X-band frequencies, the average SE was calculated to be 66 db. Minimum requirements of SE in space is set to 40 db. 219 Thus the composite chosen to replace aluminum should at least meet the minimum requirements. The CNT composite that was selected for substitution was developed by Lockheed Martin. 197 This composite consisted of a 25 µm thick multi-walled nanotube (MWCNT) sheet (areal density of 15 g/m 2 ) co-cured as an outer ply on a mm thick M55J/CE (cyanate ester, composite density ~1.6 g/cm3) polymer in an autoclave, with an average SE of 75 db. For comparison, an aluminum composite by the same company

99 89 (Lockheed Martin) was also modeled. 197 This composite was composed of a thin aluminum foil cocured on a M55J/CE substrate of the same thickness. An average value of 0.8 mm was assumed based on the available literature on similar technologies. 210,219 Using the high and low values for area covered by the EMI shields, masses of composite components (CNT, aluminum foil, and polymer) can be calculated, as shown in Table 5.2. Compared to solid aluminum, the CNT composites have lower mass in total and also meet the required mechanical properties as laminates with 0.05 cm thickness were successfully used in the launch of Juno spacecraft. 197 Figure 5.1 shows the CNT composites used in various components of the Juno spacecraft. Figure 5.1. (a) Tubular struts, and (b) engine cover components of the Juno spacecraft, both constructed using the CNT M55J/CE laminate. Image courtesy NASA/JPL-Caltech Fuel Consumption Analysis The next step is to calculate how much fuel is actually being used to launch these components into space. We can use Tsiolkovsky s rocket equation to calculate that value: v = v e ln m 0 m 1 (3) where v is the maximum change in velocity (km/s), v e is the effective exhaust velocity (km/s) and can also be calculated as v e = g. I sp (I sp is the specific impulse of the fuel used and g the acceleration by gravity), m 0 is the initial total mass at launch and m 1 is the mass without propellant (m 0 M fuel ). Communication satellites are typically launched to LEO altitudes, and

100 90 occasionally to geosynchronous earth orbits (GEO) or geostationary transfer orbits (GTO). For the satellites modeled here, launch to LEO levels were assumed. To launch satellites to LEO altitudes, an approximate v of 9.1 km s is required.225 The most common and most reliable fuel type (for its handling ease for satellite launch to LEO) is liquid fuel, a specific liquid oxygen (LOX)-kerosene mixture. 226,227 The LOX-kerosene liquid bipropellant has an I sp of 289 seconds, 225 resulting in a v e of 2.83 km/s. Accounting only for the mass of the EMI shields, the amount of fuel that is required for their launch was calculated using the rocket equation, with values shown in Table 5.2. The mass of the liquid fuel components can be calculated by having the oxidizer to fuel ratio. This value was found to be 2.58 for LOX-kerosene propellants, 228 therefore the mass of LOX and kerosene for each scenario were subsequently determined. Compared to their conventional aluminum based counterparts, the CNT polymer composites require less propellant for launch as a result of lower mass. Table 5.2. Various EMI options, mass of each component, the fuel required and cumulative energy demand (CED) for each scenario. EMI options CNT M55J/CE Laminate Components Mass (kg) CED (GJ) Total CED (GJ) CV% Iridium ORBCOMM Iridium ORBCOMM Iridium ORBCOMM Iridium ORBCOMM CNT sheet Polymer Carbon fiber Cyanate ester Propellant 1, LOX Kerosene Autoclave curing Aluminum M55J/CE Laminate Aluminum foil Polymer Carbon fiber Cyanate ester Propellant 1, LOX 1, Kerosene Autoclave curing Sheet rolling Aluminum Sheet Aluminum sheet Propellant 2, LOX 1, Kerosene Sheet rolling

101 Total CED for EMI Shielding Options Analysis of primary energy consumption for the scenarios mentioned was also carried out as part of this study. A representation of this rate is the CED of the processes, in units of MJ. CED accounts for both the direct and indirect energy use - fossil based, nuclear, hydro, and renewables from raw material extraction to the end-of-life of a product or process The primary energy consumption for the various shielding options can be determined by summing the CED of EMI shielding components and the embodied energy of their respective fuel consumption. The CEDs were modeled using the SimaPro software. For the liquid fuel, the embodied energy for kerosene and LOX were modeled based on the unit processes available in the US-EI LCI database, Kerosene, at regional storage and Oxygen, liquid, at plant, and were 54 and 9.72 MJ/kg respectively. The composites consist of the shielding element and the polymeric component. The M55J/CE is a carbon-cyanate ester composite with typical resin content of approximately 30%. 232 Cyanate ester was modeled using batch scale production data reported in literature. 233 This resin is synthesized by the reaction of cyanogen bromide and bisphenol A in the presence of trimethylamine as the catalyst. 233 Bisphenol A unit process already exists in the US-EI database as Bisphenol A, powder, at plant, but cyanogen bromide and triethylamine required further modeling and addition to the database. Triethylamine was modeled using the stoichiometric reaction of ammonia and ethanol (NH C 2H 5OH N(C 2H 5) H 2O). 234 Cyanogen bromide was also modeled using stoichiometric proportions, relying on the reaction of sodium cyanide and bromine (NaCN + Br 2 BrCN + NaBr). 235 Bromine was modeled as part of this process using lab scale production based on stoichiometry (2 NaBr + 3 H 2SO 4 Br 2 + SO NaHSO H 2O). 236 Average heating, electricity and infrastructure use for cyanate ester synthesis were based on ecoinvent reports on life cycle inventory of chemicals. 237 Additionally, a recycling rate of 90%, consistent with most LCI datasets for chemicals, was assumed for triethylamine. These modeling assumptions resulted in a CED of 212 MJ/kg of cyanate ester resin. To determine the primary energy required for 1 kg of carbon fiber, its manufacture was broken down into two processes: 1) the production of polyacrylonitrile (PAN) precursor fibers, and 2) PAN conversion to carbon fiber. 238 Based on the available literature, per 1 kg of PAN fibers 233 MJ in the form of heating by natural gas and 2.78 MJ of electricity is required. 238 Converting PAN to carbon fibers is done through a series of oxidation and etching processes. The total energy for this step per kg carbon fiber in the form of heating by natural gas was reported to be 97.8 MJ, in

102 92 addition to 72.4 MJ electricity used. 238 Modeling these components in SimaPro resulted in a total embodied energy of 696 MJ per kg carbon fiber. Aluminum that is used for EMI shielding is typically a 6061 alloy, 239 with an embodied energy ranging from 190 to 210 MJ/kg according to the CES EduPack. An average value of 200 MJ/kg was used to represent the CED of aluminum. A study on energy requirements for CNT manufacturing at industrial scale reported the thermal and electrical energy consumptions per kg MWCNTs as 295 and 187 MJ respectively. 240 These values resulted in a total CED of 1040 MJ per kg MWCNT. The raw materials require further processing prior to their use as EMI shields. For raw aluminum sheet and foil, sheet rolling to the required thickness was modeled using the respective unit process in SimaPro, resulting in an energy demand of 12.3 MJ per kg aluminum. Co-cured composites, both with CNT sheet and aluminum foil, require processing in autoclave to further cure the polymer. Based on the available data in the literature, this process has an energy intensity of 21.9 MJ per kg of composite. 241 With information on all of the embodied energy per kg of material and the mass of all the components, the CED for each shielding scenario was calculated as shown in Table 5.2, illustrating the saving in primary energy consumption with the use of nanocomposites. The aluminum composite resulted in higher CEDs, despite weighing less than pure aluminum sheet (13.6 and 0.7 kg less for high and low scenarios respectively), due to the combined effect of considerable aluminum use and high embodied energy of the polymer Uncertainty Analysis Uncertainty analysis for the cumulative energy results was carried out using the Monte-Carlo simulation in SimaPro. For all inputs and background processes, the default log-normal distribution set in the US-EI database was used. For any foreground processes, the probability distribution was determined using the pedigree matrix method, found in Table D1 of Appendix D. Uncertainty was highest among the CNT composites, with highest coefficients of variation (CVs), followed by the aluminum composites, as shown in the last two columns of Table 5.2. This largely stems from the high uncertainties of the MWCNT, carbon fiber and cyanate ester production processes. The CV for the aluminum sheet was the lowest among the various EMI shields as all the modeling components in this shielding option were previously established processes and required no further modeling of new technology.

103 93 As mentioned earlier, a wide range of launch mass for LEO communication satellites exists, consisting of very small nanosatellites weighing as low as 10 kg, to ones weighing almost a ton. This variability in mass can cause dramatic changes in the amount of fuel required to launch the equipment into orbit. Monte-Carlo simulation was used to explore the effect of this uncertain parameter (satellite launch mass), on fuel consumption for the three shielding scenarios. Based on a database published by the Union of Concerned Scientists (gathered from a collective of publicly accessible academic articles, and governmental and non-governmental sources), the distribution of mass for satellites launched to LEO by the U.S. was plotted, shown in Figure 5.2a. 223 Figure 5.2b illustrates the probability density of fuel use under the three shielding materials (aluminum sheet, CNT M55J/CE laminate and aluminum M55J/CE laminate), emphasizing the possibility of lower consumption rates by adopting the lighter CNT-enabled polymer composite EMI shields. Figure 5.2c is the graphical representation of total calculated CEDs, with error bars illustrating the 95% confidence intervals derived from Monte-Carlo simulation on input parameters Discussion A comparison of current technologies in CNT-based EMI shielding with conventional aluminum sheets shows that use of CNT-based shields would allow for more savings in satellite mass and fuel use, while performing at the same level of SE. In addition, the use of CNT sheets as the outer ply of the laminate reduces the cost, time, and labor intensity of surface preparation, as these materials do not require any surface abrasion steps. 197 The composites also satisfy the mechanical requirements of spacecraft design based on the results of flexural tests comparing them with the solid polymeric sheet. 197 These properties favor the use of nanocomposites for future developments in space applications.

104 94 Figure 5.2. (a) Mass distribution of LEO satellites, (b) distribution of fuel consumption for launching the EMI shielding under various shielding scenarios, (c) total CED for Iridium NEXT and ORBCOMM satellites under various shielding scenarios with error bars representing 95% confidence interval from Monte-Carlo simulation on input parameters (CV values can be found in Table 5.2). It is noteworthy to mention that the energy require for cutting and shaping the shields were not included as part of this study and were assumed to be similar between the shielding options. Another issue that introduces a level of uncertainty to the analysis is the method of resin curing. Resins can be cured through a variety of methods including thermal curing with autoclaves at different durations and temperatures (room temperature to 200 C), microwave curing, 246,247 electron beam, ultraviolet- and X-rays. 248,249 The curing method used is highly dependent on the manufacturer and application of the composite. Energy consumption of these methods differ greatly; as an example, electron beam curing has been reported to use only 10% of the energy used by thermal curing. 248 Ecoinvent s report on life cycle inventory of chemicals also

105 95 acknowledges the variety in resin curing agents in addition to the different routes, where it is assumed that the data on the resin in many cases will provide a sensible approximation. 237 Based on the rocket equation (3), the fuel requirements for launch can increase exponentially with respect to the velocity ratios. Thus the analysis results can change by using a fuel with higher specific impulses such as LOX-hydrazine or solid rocket propellants. With regard to the CED of aluminum based technologies, the final outcomes can be affected by the thickness of the shielding device. The values used in this study were based on available literature, but with technological advancements, the thickness could be reduced further, resulting in lower CED levels. We have shown results for two satellite cases with launch to LEO, but as mentioned, there are various mass classes for satellites, for example, satellites launched to GTO can weigh as much as 5400 kg. Thus depending on the satellite s primary application and the final orbital altitude, the mass and ultimately the amount of propellant used could be influenced. Prior LCA studies on CNT manufacture have mainly quantified their environmental implications. In this study we have attempted to capture the increased performance of the current technology by quantifying the benefits gained with respect to the substituted material. For many other cases, where there are societal and health gains in using nanotechnology, these benefits are not so easily quantifiable, such as determining quality of life or ethical concerns. Hence, it is important to develop quantitative methods to assess life cycle trade-offs of products. A recent study uses disability-adjusted life years (DALYs) as a measure for potential human health benefits of a CNTenabled chemical gas sensor, defining impact-benefit ratios to assess the feasibility of the product. 120 Another study investigated the net energy benefit of using CNTs for four CNT applications (CNT-incorporated reinforced cement, flash memory switches and lithium-ion batteries). 251 Their results demonstrated that life cycle net energy benefits from CNT-enabled products are dependent on the application. CNTs in cell phone flash memory devices, cement reinforcement and MWCNT for cathode of batteries showed positive net energy benefits, while for batteries with SWCNT anodes, net energy benefits were negative. These results illustrate the importance of evaluating life cycle ecological impacts and energy use of a technology prior to its adoption. Net energy benefits of nano-enabled products will likely increase with increased efficiencies from technological developments, but LCA concepts can still help to identify potential opportunities for improvement.

106 96 Chapter 6 Conclusion and Future Work 6.1. Summary of Chapters This dissertation focused on quantifying the life cycle environmental performance of nanomaterials and nano-enabled products. LCAs with different scopes were performed for two different nanomaterials in compliance with the scope of the NSF Sustainable Nanomanufacturing Award. The LCAs were developed to serve as foundations for future development in nano-lca. Following the research objectives mentioned in the Introduction Section 4, studies were carried out on different aspects of nano-lca. These studies emphasized on proper assessment of the environmental performance of nanomaterials and nano-enabled products by complying with guidelines of ISO In Chapter 2, a comprehensive cradle-to-grave LCA for an AgNP-enabled wound dressing was performed. This study demonstrated the dominancy of AgNPs, for all TRACI environmental impact categories, in the overall impacts of the product. This was found to be caused by upstream emissions from silver extraction. It was also found that environmental impacts of production were several times those of end-of-life disposal scenarios (incineration). Chapter 3 illustrated the effects of varying AgNP synthesis methods, on a mass basis and the basis of function. Comparative cradle-to-gate LCA of several AgNP syntheses were performed, suggesting silver source to be the main contributor to environmental impacts for nearly all impact categories for every route. Energy intensive physical synthesis methods showed the highest level of impacts on a mass basis. This study also showed that seemingly green synthesis concepts should be assessed with LCA frameworks as results can prove otherwise. Results were also presented on a functional unit basis and in the context of a product. Chapter 4 discussed the importance of AgNP loading on products to the overall environmental burdens by performing cradle-to-gate LCA analysis for an array of products. Results showed the contribution of AgNPs to production impacts is both a function of AgNP concentration, and the composition of the product. This LCA analysis was coupled with release studies first to show the importance of particle incorporation method on the amount leached from the product, and second, to demonstrate the greater environmental impacts due to

107 97 upstream processes compared to impacts from direct exposure during use phase release. To quantify life cycle benefits gained by substituting conventional methods with nanotechnology, the case of CNT-enabled EMI shielding was assessed in Chapter 5. This analysis compared the advantages of adopting CNT composites as opposed to aluminum on the basis of their performance for satellite applications. Using CNT composites suggested reduction in the EMI shielding mass, which translates to less fuel required for launch, lower primary energy requirements, and subsequently, lower environmental impacts Contributions of Study The works depicted in this dissertation have made novel contributions to the nano-lca body of knowledge. With the growth in the market of medical applications of AgNP, it is important to assess its environmental impacts beforehand. The LCA of AgNP-enabled wound dressing was the first nano-lca performed for healthcare settings. Comprehensive and transparent data were published for nano-synthesis, and product manufacturing. These datasets fill the current knowledge gaps in creating an LCI for AgNPs, as current LCAs on AgNPs are limited with LCI often being under non-disclosure agreements. Methodological advancements were made in appropriate usage of functional unit in nano-lca to better understand the relative impacts of nanomaterials. All the comparative LCAs were performed on the basis of function as opposed to conventional mass-based or product based assessments. This is an important procedural step of LCA and was defined for nanomaterials for the first time. For AgNPs, their size-dependent bioactivity towards organisms were set as their function. It is encouraged for future LCAs to adopt a function based approach when comparing multiple scenarios. Nanomaterials, even AgNPs, provide many functions, all acceptable as a basis depending on their ultimate end-use. Additionally for nano-enabled product comparison, for the case of AgNPs, quantifying results based on antimicrobial efficacy was proposed for the first time for product comparisons. Using zone of inhibition results for LCA purposes had never been attempted prior to this date. For the case of CNTs, the unique application of satellite EMI shielding was chosen with comparison of scenarios on the basis of shielding effectiveness. This analysis could serve as a basis for potential future LCAs, emphasizing the need for incorporating the basic science related to the field of study within the life cycle analysis.

108 98 Cradle-to-gate LCAs were incorporated with research on other nano-related disciplines for the first time. This is a step towards the development of comprehensive environmental assessment for nanomaterials and nano-enabled products. The contributions of this dissertation could potentially aid in decision making and prioritizing policies regarding nanomanufacturing and nano-enabled products. Results of these studies could be used in the following ways: More studies on nanotechnology is required to fill the data gaps for creating inventories in nano-lca. LCIs created here for nano-synthesis and products could be used or built up on, by LCA practitioners, for future environmental assessments of nanoparticles or nanoenabled products. Environmental impacts of nanomaterials were shown to be dominated by synthesis. With nanotechnology moving towards commercialization, it is important to control the manufacturing aspect of life cycle, to affect the overall environmental impacts of ENMs. The comparative synthesis results can be used to guide manufacturers in opting for a life cycle perspective, to avoid the unintended consequences of nano synthesis. The study on life cycle net energy benefit could serve as an example for future comparative LCAs between advanced and current technology, as it included performance based comparison among products. This would allow for a fair comparison of results and determination of their true environmental tradeoffs. Using the outcomes of cradle-to-gate assessment combined with use phase release could potentially affect regulations regarding threshold concentrations of nanomaterials for specific applications for consumer and environmental safety. This threshold level can be used for the purpose of labeling nano-enabled products, or for changing waste management procedures through safe disposal of products or extended producer responsibility acts. Despite the focus of this dissertation on indirect environmental burdens of nanomaterials, the concerns regarding their direct effects should not be neglected. RA methodologies have the ability to determine the potential consequences of exposure to nanomaterials for different human and environmental aspects. LCAs provide information on impactful upstream processes that would otherwise not be included in a RA. In that sense, LCA works are complementary to RA allowing for a thorough assessment of potential causes of human and environmental impact from nano- and non-nanoemissions. Therefore LCAs and RAs can be combined to ensure the responsible

109 99 development of nanotechnology in a cost effective manner, by identifying hotspots for improvement throughout the life cycle of ENMs or nano-enabled products Future Work and Research Direction The works in this study can be extended in various ways. Questions still remain regarding the environmental performance of other nanomaterials, their exact fate and behavior in the environment, the factors affecting their toxicity toward human health and the environment, and what the collective conclusion of all these studies mean for the future of nanotechnology. Additional research is required to answer these concerns with the development of nanotechnology. Some specific areas of future development on studies are as follows: Expansion of nano-lca to nanomaterials less studied such as zinc oxides or Fe oxides. Considering LCA of other areas AgNP application, such as electronics, comparing products to find any similarities in their environmental performance. Product based LCAs could be further studied for their impacts during end-of-life to determine the fate of nanomaterials at that stage. Extending the work on performance based LCA to include other applications and nanomaterials. Extensive research on definition of characterization factors of AgNPs and other nanomaterials for proper assessment of their environmental impacts. This requires further insight into nanomaterial dissolution, fate and transport kinetics. Also, additional studies on the toxicity of nanomaterials with set standards are required for this purpose. More LCAs on different nano-enabled products and other nanomaterials can expand the environmental outlook of nanotechnology. These LCAs can provide insight to specific areas of a product life cycle that would require potential improvements as this technology ascends to commercialization. To correctly assess the environmental impact of direct exposure to nanomaterials, nano-specific characterization factors are required. This field should include collaborations with particle fate modelers, and nanotoxicologists. Affecting these characterization factors are particle characterization, size and shape, surface coating, agglomeration levels, species change, environmental conditions and many more. Further research in this area is required to ensure the accuracy of particle-specific characterization factors. Finally, scale-up of nanomanufacturing could potentially result in a rebound effect. There is greater efficiency in a

110 100 commercialized manufacturing plant, resulting in reduced costs and increased use. This increased use of nanotechnology would potentially offset the benefits gained through manufacturing efficiency. It is imperative to use LCA or RA tools as precautionary measures, to ensure safe and sustainable development of nanotechnology as they focus on complementary issues regarding the concerns of nanotechnology, focusing on both indirect and direct implications of their use, respectively.

111 101 References (1) Varner, K.; El-Badaway, A.; Feldhake, D.; Venkatapathy, R. State of the science literature review: everything nanosilver and more. US Environ. Prot. Agency Wash. DC 2010, 363. (2) Benn, T.; Cavanagh, B.; Hristovski, K.; Posner, J. D.; Westerhoff, P. The release of nanosilver from consumer products used in the home. J. Environ. Qual. 2010, 39 (6), (3) Quadros, M. E.; Pierson IV, R.; Tulve, N. S.; Willis, R.; Rogers, K.; Thomas, T. A.; Marr, L. C. Release of silver from nanotechnology-based consumer products for children. Environ. Sci. Technol. 2013, 47 (15), (4) Benn, T. M.; Westerhoff, P. Nanoparticle silver released into water from commercially available sock fabrics. Environ. Sci. Technol. 2008, 42 (11), (5) Reed, R. B.; Zaikova, T.; Barber, A.; Simonich, M.; Lankone, R.; Marco, M.; Hristovski, K.; Herckes, P.; Passantino, L.; Fairbrother, D. H.; et al. Potential Environmental Impacts and Antimicrobial Efficacy of Silver-and Nanosilver-Containing Textiles. Environ. Sci. Technol (6) Mitrano, D. M.; Rimmele, E.; Wichser, A.; Erni, R.; Height, M.; Nowack, B. Presence of Nanoparticles in Wash Water from Conventional Silver and Nano-Silver Textiles. ACS Nano (7) Khanna, V.; Bakshi, B. R. Carbon Nanofiber Polymer Composites: Evaluation of Life Cycle Energy Use. Environ. Sci. Technol. 2009, 43 (6), (8) Wender, B. A.; Seager, T. P. Anticipatory life cycle assessment of SWCNT-enabled lithium ion batteries; Nanotechnology for Sustainable Manufacturing, (9) Espinoza, V. S.; Erbis, S.; Pourzahedi, L.; Eckelman, M. J.; Isaacs, J. A. Material flow analysis of carbon nanotube lithium-ion batteries used in portable computers. ACS Sustain. Chem. Eng. 2014, 2 (7), (10) Dahlben, L. J.; Eckelman, M. J.; Hakimian, A.; Somu, S.; Isaacs, J. A. Environmental life cycle assessment of a carbon nanotube-enabled semiconductor device. Environ. Sci. Technol. 2013, 47 (15), (11) Keller, A. A.; McFerran, S.; Lazareva, A.; Suh, S. Global life cycle releases of engineered nanomaterials. J. Nanoparticle Res. 2013, 15 (6), (12) Babaizadeh, H.; Hassan, M. Life cycle assessment of nano-sized titanium dioxide coating on residential windows. Constr. Build. Mater. 2013, 40, (13) Nanotechnology Market Outlook market research report (accessed Apr 5, 2016). (14) Project on Emerging Nanotechnologies, Consumer Products Inventory; (15) Eckelman, M. J.; Zimmerman, J. B.; Anastas, P. T. Toward green nano. J. Ind. Ecol. 2008, 12 (3), (16) Nel, A.; Xia, T.; Meng, H.; Wang, X.; Lin, S.; Ji, Z.; Zhang, H. Nanomaterial toxicity testing in the 21st century: use of a predictive toxicological approach and high-throughput screening. Acc. Chem. Res. 2012, 46 (3), (17) Gilbertson, L. M.; Wender, B. A.; Zimmerman, J. B.; Eckelman, M. J. Coordinating modeling and experimental research of engineered nanomaterials to improve life cycle assessment studies. Environ. Sci. Nano 2015, 2 (6),

112 (18) Mueller, N. C.; Nowack, B. Exposure Modeling of Engineered Nanoparticles in the Environment. Environ. Sci. Technol. 2008, 42 (12), (19) Gottschalk, F.; Nowack, B. The release of engineered nanomaterials to the environment. J. Environ. Monit. 2011, 13 (5), (20) Keller, A. A.; Lazareva, A. Predicted releases of engineered nanomaterials: From global to regional to local. Environ. Sci. Technol. Lett (21) Meesters, J.; Koelmans, A. A.; Quik, J. T.; Hendriks, A. J.; Van de Meent, D. Multimedia Modeling of Engineered Nanoparticles with SimpleBox4nano: Model Definition and Evaluation. Environ. Sci. Technol (22) Blaser, S. A.; Scheringer, M.; MacLeod, M.; Hungerbühler, K. Estimation of cumulative aquatic exposure and risk due to silver: Contribution of nano-functionalized plastics and textiles. Sci. Total Environ. 2008, 390 (2), (23) VandeVoort, A. R.; Tappero, R.; Arai, Y. Residence time effects on phase transformation of nanosilver in reduced soils. Environ. Sci. Pollut. Res. 2014, (24) Boxall, A. B.; Tiede, K.; Chaudhry, Q. Engineered nanomaterials in soils and water: how do they behave and could they pose a risk to human health? Nanomed. 2007, 2 (6), (25) Jaisi, D. P.; Elimelech, M. Single-Walled Carbon Nanotubes Exhibit Limited Transport in Soil Columns. Environ. Sci. Technol. 2009, 43 (24), (26) Chen, C. Y.; Jafvert, C. T. Sorption of buckminsterfullerene (C60) to saturated soils. Environ. Sci. Technol. 2009, 43 (19), (27) Klaine, S. J.; Alvarez, P. J. J.; Batley, G. E.; Fernandes, T. F.; Handy, R. D.; Lyon, D. Y.; Mahendra, S.; McLaughlin, M. J.; Lead, J. R. Nanomaterials in the environment: Behavior, fate, bioavailability, and effects. Environ. Toxicol. Chem. 2008, 27 (9), (28) Dobias, J.; Bernier-Latmani, R. Silver release from silver nanoparticles in natural waters. Environ. Sci. Technol. 2013, 47 (9), (29) Lowry, G. V.; Espinasse, B. P.; Badireddy, A. R.; Richardson, C. J.; Reinsch, B. C.; Bryant, L. D.; Bone, A. J.; Deonarine, A.; Chae, S.; Therezien, M.; et al. Long-term transformation and fate of manufactured Ag nanoparticles in a simulated large scale freshwater emergent wetland. Environ. Sci. Technol. 2012, 46 (13), (30) Mitrano, D. M.; Ranville, J. F.; Bednar, A.; Kazor, K.; Hering, A. S.; Higgins, C. P. Tracking dissolution of silver nanoparticles at environmentally relevant concentrations in laboratory, natural, and processed waters using single particle ICP-MS (spicp-ms). Environ. Sci. Nano 2014, 1 (3), (31) Taghavy, A.; Mittelman, A.; Wang, Y.; Pennell, K. D.; Abriola, L. M. Mathematical Modeling of the Transport and Dissolution of Citrate-Stabilized Silver Nanoparticles in Porous Media. Environ. Sci. Technol. 2013, 47 (15), (32) Quadros, M. E.; Marr, L. C. Environmental and human health risks of aerosolized silver nanoparticles. J. Air Waste Manag. Assoc. 2010, 60 (7), (33) Brar, S. K.; Verma, M.; Tyagi, R. D.; Surampalli, R. Y. Engineered nanoparticles in wastewater and wastewater sludge - Evidence and impacts. Waste Manag. 2009, 30 (3), (34) Kaegi, R.; Voegelin, A.; Sinnet, B.; Zuleeg, S.; Hagendorfer, H.; Burkhardt, M.; Siegrist, H. Behavior of metallic silver nanoparticles in a pilot wastewater treatment plant. Environ. Sci. Technol. 2011, 45 (9), (35) Hou, L.; Li, K.; Ding, Y.; Li, Y.; Chen, J.; Wu, X.; Li, X. Removal of silver nanoparticles in simulated wastewater treatment processes and its impact on COD and NH< sub> 4</sub> reduction. Chemosphere 2012, 87 (3),

113 (36) Holder, A. L.; Vejerano, E. P.; Zhou, X.; Marr, L. C. Nanomaterial disposal by incineration. Environ. Sci. Process. Impacts 2013, 15 (9), (37) Walser, T.; Limbach, L. K.; Brogioli, R.; Erismann, E.; Flamigni, L.; Hattendorf, B.; Juchli, M.; Krumeich, F.; Ludwig, C.; Prikopsky, K.; et al. Persistence of engineered nanoparticles in a municipal solid-waste incineration plant. Nat. Nanotechnol. 2012, 7 (8), (38) Wiesner, M. R.; Plata, D. L. Environmental, health and safety issues: Incinerator filters nanoparticles. Nat. Nanotechnol. 2012, 7 (8), (39) Bolyard, S. C.; Reinhart, D. R.; Santra, S. Behavior of Engineered Nanoparticles in Landfill Leachate. Environ. Sci. Technol. 2013, 47 (15), (40) Gavankar, S.; Suh, S.; Keller, A. F. Life cycle assessment at nanoscale: review and recommendations. Int. J. Life Cycle Assess. 2012, 1 9. (41) Oberdorster, G.; Oberdorster, E.; Oberdorster, J. Nanotoxicology: An emerging discipline evolving from studies of ultrafine particles. Environ. Health Perspect. 2005, 113 (7), (42) Adams, L. K.; Lyon, D. Y.; McIntosh, A.; Alvarez, P. J. Comparative toxicity of nano-scale TiO2, SiO2 and ZnO water suspensions. Water Sci. Technol. 2006, 54 (11-12), (43) Aschberger, K.; Johnston, H. J.; Stone, V.; Aitken, R. J.; Hankin, S. M.; Peters, S. A.; Tran, C. L.; Christensen, F. M. Review of carbon nanotubes toxicity and exposure-appraisal of human health risk assessment based on open literature. Crit. Rev. Toxicol. 2010, 40 (9), (44) Baun, A.; Hartmann, N.; Grieger, K.; Kusk, K. Ecotoxicity of engineered nanoparticles to aquatic invertebrates: a brief review and recommendations for future toxicity testing. Ecotoxicology 2008, 17 (5), (45) Bystrzejewska-Piotrowska, G.; Golimowski, J.; Urban, P. L. Nanoparticles: their potential toxicity, waste and environmental management. Waste Manag. 2009, 29 (9), (46) Blaise, C.; Gagné, F.; Férard, J. F.; Eullaffroy, P. Ecotoxicity of selected nano-materials to aquatic organisms. Environ. Toxicol. 2008, 23 (5), (47) Hischier, R.; Walser, T. Life cycle assessment of engineered nanomaterials: State of the art and strategies to overcome existing gaps. Sci. Total Environ. 2012, 425, (48) Wender, B. A.; Seager, T. P. Towards prospective life cycle assessment: Single wall carbon nanotubes for lithium-ion batteries. In Sustainable Systems and Technology (ISSST), 2011 IEEE International Symposium on; 2011; pp 1 4. (49) Rebitzer, G.; Ekvall, T.; Frischknecht, R.; Hunkeler, D.; Norris, G.; Rydberg, T.; Schmidt, W.- P.; Suh, S.; Weidema, B. P.; Pennington, D. W. Life cycle assessment: Part 1: Framework, goal and scope definition, inventory analysis, and applications. Environ. Int. 2004, 30 (5), (50) Guinee, J. B.; Heijungs, R.; Huppes, G.; Zamagni, A.; Masoni, P.; Buonamici, R.; Ekvall, T.; Rydberg, T. Life cycle assessment: past, present, and future. Environ. Sci. Technol. 2010, 45 (1), (51) ISO 14040: Environmental management -- Life cycle assessment -- Principles and framework (accessed Apr 2, 2016). (52) ISO 14044: Environmental management -- Life cycle assessment -- Requirements and guidelines (accessed Apr 3, 2016). (53) Reap, J.; Roman, F.; Duncan, S.; Bras, B. A survey of unresolved problems in life cycle assessment-part I goals and scope and inventory analysis. Int. J. Life Cycle Assess. 2008, 13 (4),

114 (54) Pennington, D. W.; Potting, J.; Finnveden, G.; Lindeijer, E.; Jolliet, O.; Rydberg, T.; Rebitzer, G. Life cycle assessment Part 2: Current impact assessment practice. Environ. Int. 2004, 30 (5), (55) Reap, J.; Roman, F.; Duncan, S.; Bras, B. A survey of unresolved problems in life cycle assessment. Int. J. Life Cycle Assess. 2008, 13 (5), (56) Bare, J. C. TRACI. J. Ind. Ecol. 2002, 6 (3-4), (57) Bare, J. TRACI 2.0: the tool for the reduction and assessment of chemical and other environmental impacts 2.0. Clean Technol. Environ. Policy 2011, 13 (5), (58) Henderson, A.; Hauschild, M.; van de Meent, D.; Huijbregts, M.; Larsen, H.; Margni, M.; McKone, T.; Payet, J.; Rosenbaum, R.; Jolliet, O. USEtox fate and ecotoxicity factors for comparative assessment of toxic emissions in life cycle analysis: sensitivity to key chemical properties. Int. J. Life Cycle Assess. 2011, 16 (8), (59) Rosenbaum, R. K.; Bachmann, T. M.; Gold, L. S.; Huijbregts, M. A. J.; Jolliet, O.; Juraske, R.; Koehler, A.; Larsen, H. F.; MacLeod, M.; Margni, M. USEtox the UNEP-SETAC toxicity model: recommended characterisation factors for human toxicity and freshwater ecotoxicity in life cycle impact assessment. Int. J. Life Cycle Assess. 2008, 13 (7), (60) Huijbregts, M.; Hauschild, M.; Jolliet, O.; Margni, M.; McKone, T.; Rosenbaum, R.; Van De Meent, D. USEtox User Manual. USEtox Team February (61) Upadhyayula, V. K. K.; Meyer, D. E.; Curran, M. A.; Gonzalez, M. A. Life cycle assessment as a tool to enhance the environmental performance of carbon nanotube products: A review. J. Clean. Prod. 2012, 26, (62) Walser, T.; Demou, E.; Lang, D. J.; Hellweg, S. Prospective environmental life cycle assessment of nanosilver T-shirts. Environ. Sci. Technol. 2011, 45 (10), (63) Gottschalk, F.; Sonderer, T.; Scholz, R. W.; Nowack, B. Possibilities and limitations of modeling environmental exposure to engineered nanomaterials by probabilistic material flow analysis. Environ. Toxicol. Chem. 2010, 29 (5), (64) Eckelman, M. J.; Mauter, M. S.; Isaacs, J. A.; Elimelech, M. New Perspectives on Nanomaterial Aquatic Ecotoxicity: Production Impacts Exceed Direct Exposure Impacts for Carbon Nanotoubes. Environ. Sci. Technol. 2012, 46 (5), (65) Salieri, B.; Righi, S.; Pasteris, A.; Olsen, S. I. Freshwater ecotoxicity characterisation factor for metal oxide nanoparticles: A case study on titanium dioxide nanoparticle. Sci. Total Environ. 2015, 505, (66) Silver Nanoparticles Market By Application And Segment Forecasts To (accessed Apr 5, 2016). (67) Pourzahedi, L.; Eckelman, M. J. Environmental Life Cycle Assessment of Nanosilver- Enabled Bandages. Env. Sci Technol. (68) Pourzahedi, L.; Eckelman, M. J. Comparative life cycle assessment of silver nanoparticle synthesis routes. Environ. Sci. Nano 2015, 2 (4), (69) Dahlben, L. J.; Eckelman, M. J.; Hakimian, A.; Somu, S.; Isaacs, J. Environmental Life Cycle Assessment of a Carbon Nanotube-Enabled Semiconductor Device. Environ. Sci. Technol (70) Pei Zhai; Matthew J. Eckelman. Net energy benefits of carbon nanotube applications. Press. (71) Walker, W. C.; Bosso, C. J.; Eckelman, M.; Isaacs, J. A.; Pourzahedi, L. Integrating life cycle assessment into managing potential EHS risks of engineered nanomaterials: reviewing progress to date. J. Nanoparticle Res. 2015, 17 (8),

115 (72) Nowack, B.; Krug, H. F.; Height, M. 120 years of nanosilver history: implications for policy makers. Environ. Sci. Technol. 2011, 45 (4), (73) Marambio-Jones, C.; Hoek, E. M. A review of the antibacterial effects of silver nanomaterials and potential implications for human health and the environment. J. Nanoparticle Res. 2010, 12 (5), (74) Wijnhoven, S. W.; Peijnenburg, W. J.; Herberts, C. A.; Hagens, W. I.; Oomen, A. G.; Heugens, E. H.; Roszek, B.; Bisschops, J.; Gosens, I.; Van De Meent, D. Nano-silver-a review of available data and knowledge gaps in human and environmental risk assessment. Nanotoxicology 2009, 3 (2), (75) Wardak, A.; Gorman, M. E.; Swami, N.; Deshpande, S. Identification of Risks in the Life Cycle of Nanotechnology-Based Products. J. Ind. Ecol. 2008, 12 (3), (76) Fabrega, J.; Luoma, S. N.; Tyler, C. R.; Galloway, T. S.; Lead, J. R. Silver nanoparticles: behaviour and effects in the aquatic environment. Environ. Int. 2011, 37 (2), (77) Rai, M.; Yadav, A.; Gade, A. Silver nanoparticles as a new generation of antimicrobials. Biotechnol. Adv. 2009, 27 (1), (78) Saengkiettiyut, K.; Rattanawaleedirojn, P.; Sangsuk, S. A study on antimicrobial efficacy of nano silver containing textile. CMU J Nat Sci 2008, 7, (79) Castellano, J. J.; Shafii, S. M.; Ko, F.; Donate, G.; Wright, T. E.; Mannari, R. J.; Payne, W. G.; Smith, D. J.; Robson, M. C. Comparative evaluation of silver-containing antimicrobial dressings and drugs. Int. Wound J. 2007, 4 (2), (80) Tonkin, C.; Wood, F. Nanocrystalline silver reduces the need for antibiotic therapy in burn wounds. Prim. Intent. 2005, 13 (4), 163. (81) Fong, J.; Wood, F. Nanocrystalline silver dressings in wound management: a review. Int. J. Nanomedicine 2006, 1 (4), 441. (82) Huang, Y.; Li, X.; Liao, Z.; Zhang, G.; Liu, Q.; Tang, J.; Peng, Y.; Liu, X.; Luo, Q. A randomized comparative trial between Acticoat and SD-Ag in the treatment of residual burn wounds, including safety analysis. Burns 2007, 33 (2), (83) Geranio, L.; Heuberger, M.; Nowack, B. The behavior of silver nanotextiles during washing. Environ. Sci. Technol. 2009, 43 (21), (84) Benn, T. M.; Westerhoff, P. Nanoparticle silver released into water from commercially available sock fabrics. Environ. Sci. Technol. 2008, 42 (11), (85) Arvidsson, R.; Molander, S.; Sandén, B. A. Impacts of a Silver-Coated Future. J. Ind. Ecol. 2011, 15 (6), (86) Choi, O.; Deng, K. K.; Kim, N.-J.; Ross Jr, L.; Surampalli, R. Y.; Hu, Z. The inhibitory effects of silver nanoparticles, silver ions, and silver chloride colloids on microbial growth. Water Res. 2008, 42 (12), (87) Gottschalk, F.; Nowack, B. The release of engineered nanomaterials to the environment. J. Environ. Monit. 2011, 13 (5), (88) Mueller, N. C.; Nowack, B. Exposure modeling of engineered nanoparticles in the environment. Environ. Sci. Technol. 2008, 42 (12), (89) Gottschalk, F.; Sonderer, T.; Scholz, R. W.; Nowack, B. Modeled environmental concentrations of engineered nanomaterials (TiO2, ZnO, Ag, CNT, fullerenes) for different regions. Environ. Sci. Technol. 2009, 43 (24), (90) Money, E. S.; Reckhow, K. H.; Wiesner, M. R. The use of Bayesian networks for nanoparticle risk forecasting: model formulation and baseline evaluation. Sci. Total Environ. 2012, 426,

116 (91) Kittler, S.; Greulich, C.; Diendorf, J.; Koller, M.; Epple, M. Toxicity of silver nanoparticles increases during storage because of slow dissolution under release of silver ions. Chem. Mater. 2010, 22 (16), (92) Nowack, B.; Bucheli, T. D. Occurrence, behavior and effects of nanoparticles in the environment. Environ. Pollut. 2007, 150 (1), (93) Eckelman, M. J.; Mauter, M. S.; Isaacs, J. A.; Elimelech, M. New perspectives on nanomaterial aquatic ecotoxicity: Production impacts exceed direct exposure impacts for carbon nanotoubes. Environ. Sci. Technol. 2012, 46 (5), (94) Sengul, H.; Theis, T. L.; Ghosh, S. Toward sustainable nanoproducts: An overview of nanomanufacturing methods. J. Ind. Ecol. 2008, 12 (3), (95) Healy, M. L.; Dahlben, L. J.; Isaacs, J. A. Environmental Assessment of Single-Walled Carbon Nanotube Processes. J. Ind. Ecol. 2008, 12 (3), (96) Pati, P.; McGinnis, S.; Vikesland, P. J. Life Cycle Assessment of Green Nanoparticle Synthesis Methods. Environ. Eng. Sci. 2014, 31 (7), (97) Walser, T.; Demou, E.; Lang, D. J.; Hellweg, S. Prospective environmental life cycle assessment of nanosilver T-shirts. Environ. Sci. Technol. 2011, 45 (10), (98) Meyer, D. E.; Curran, M. A.; Gonzalez, M. A. An examination of silver nanoparticles in socks using screening-level life cycle assessment. J. Nanoparticle Res. 2011, 13 (1), (99) Silcryst nanosilver technology (100) Moiemen, N. S.; Shale, E.; Drysdale, K. J.; Smith, G.; Wilson, Y. T.; Papini, R. Acticoat dressings and major burns: systemic silver absorption. Burns 2011, 37 (1), (101) Acticoat 7, description and features nephew.com/professional/products/advanced-wound-management/acticoat/acticoat- 7/. (102) Mattox, D. M. Handbook of physical vapor deposition (PVD) processing; William Andrew, (103) Musil, J.; Baroch, P.; Vlček, J.; Nam, K. H.; Han, J. G. Reactive magnetron sputtering of thin films: present status and trends. Thin Solid Films 2005, 475 (1), (104) Pierson, J. F.; Wiederkehr, D.; Billard, A. Reactive magnetron sputtering of copper, silver, and gold. Thin Solid Films 2005, 478 (1), (105) Dunn, K.; Edwards-Jones, V. The role of Acticoat TM with nanocrystalline silver in the management of burns. Burns 2004, 30, S1 S9. (106) Parsons, D.; Bowler, P. G.; Myles, V.; Jones, S. Silver antimicrobial dressings in wound management: a comparison of antibacterial, physical, and chemical characteristics. Wounds 2005, 17 (8), 222. (107) Office of Air Quality Planning and Standards; Office of Air and Radiation. Emission Factor Documentation for AP-42 Section 2.6 Medical Waste Incineration. U.S. Environmental Protection Agency July (108) Quik, J. T.; Vonk, J. A.; Hansen, S. F.; Baun, A.; Van De Meent, D. How to assess exposure of aquatic organisms to manufactured nanoparticles? Environ. Int. 2011, 37 (6), (109) Zhang, W.; Yao, Y.; Sullivan, N.; Chen, Y. Modeling the primary size effects of citratecoated silver nanoparticles on their ion release kinetics. Environ. Sci. Technol. 2011, 45 (10), (110) Levard, C.; Hotze, E. M.; Lowry, G. V.; Brown Jr, G. E. Environmental transformations of silver nanoparticles: impact on stability and toxicity. Environ. Sci. Technol. 2012, 46 (13),

117 (111) Liu, J.; Hurt, R. H. Ion release kinetics and particle persistence in aqueous nano-silver colloids. Environ. Sci. Technol. 2010, 44 (6), (112) Ryberg, M.; Vieira, M. D. M.; Zgola, M.; Bare, J.; Rosenbaum, R. K. Updated US and Canadian normalization factors for TRACI 2.1. Clean Technol. Environ. Policy 2014, 16 (2), (113) Project on Emerging Nanotechnologies. Consumer Products Inventory (accessed Dec 2, 2014). (114) Tuttle, G. R. Size and surface area dependent toxicity of silver nanoparticles in zebrafish embryos (Danio rerio) (115) Park, M. V.; Neigh, A. M.; Vermeulen, J. P.; de la Fonteyne, L. J.; Verharen, H. W.; Briedé, J. J.; van Loveren, H.; de Jong, W. H. The effect of particle size on the cytotoxicity, inflammation, developmental toxicity and genotoxicity of silver nanoparticles. Biomaterials 2011, 32 (36), (116) Powers, C. M.; Slotkin, T. A.; Seidler, F. J.; Badireddy, A. R.; Padilla, S. Silver nanoparticles alter zebrafish development and larval behavior: distinct roles for particle size, coating and composition. Neurotoxicol. Teratol. 2011, 33 (6), (117) Gibbins, B.; Warner, L. The role of antimicrobial silver nanotechnology. Med. Device Diagn. Ind. Mag. 2005, 1, 1 2. (118) Holder, A. L.; Vejerano, E. P.; Zhou, X.; Marr, L. C. Nanomaterial disposal by incineration. Environ. Sci. Process. Impacts 2013, 15 (9), (119) Choi, O.; Clevenger, T. E.; Deng, B.; Surampalli, R. Y.; Ross Jr, L.; Hu, Z. Role of sulfide and ligand strength in controlling nanosilver toxicity. Water Res. 2009, 43 (7), (120) Gilbertson, L. M.; Busnaina, A. A.; Isaacs, J. A.; Zimmerman, J. B.; Eckelman, M. J. Life Cycle Impacts and Benefits of a Carbon Nanotube-Enabled Chemical Gas Sensor. Environ. Sci. Technol. 2014, 48 (19), (121) Pasquini, L. M.; Hashmi, S. M.; Sommer, T. J.; Elimelech, M.; Zimmerman, J. B. Impact of surface functionalization on bacterial cytotoxicity of single-walled carbon nanotubes. Environ. Sci. Technol. 2012, 46 (11), (122) Eckelman, M. J.; Graedel, T. E. Silver emissions and their environmental impacts: a multilevel assessment. Environ. Sci. Technol. 2007, 41 (17), (123) Pourzahedi, L.; Eckelman, M. J. Environmental Life Cycle Assessment of Nanosilver- Enabled Bandages. Environ. Sci. Technol. 2014, 49 (1), (124) Cushing, B. L.; Kolesnichenko, V. L.; O Connor, C. J. Recent advances in the liquid-phase syntheses of inorganic nanoparticles. Chem. Rev. 2004, 104 (9), (125) Dahl, J. A.; Maddux, B. L. S.; Hutchison, J. E. Toward Greener Nanosynthesis. Chem. Rev. 2007, 107 (6), (126) Pati, P.; McGinnis, S.; Vikesland, P. J. Life Cycle Assessment of Green Nanoparticle Synthesis Methods. Environ. Eng. Sci. 2014, 31 (7), (127) Šileikaite, A.; Prosycevas, I.; Puiso, J.; Juraitis, A.; Guobiene, A. Analysis of Silver Nanoparticles Produced by Chemical Reduction of Silver Salt Solution. Mater. Sci. MEDŽIAGOTYRA 2006, 12 (4). (128) Mulfinger, L.; Solomon, S. D.; Bahadory, M.; Jeyarajasingam, A. V.; Rutkowsky, S. A.; Boritz, C. Synthesis and study of silver nanoparticles. J. Chem. Educ. 2007, 84 (2), 322. (129) A. Slistan-Grijalva, R. H.-U. Assessment of growth of silver nanoparticles synthesized from an ethylene glycol silver nitrate polyvinylpyrrolidone solution. Phys. E-Low-Dimens. Syst. Amp Nanostructures - Phys. E 2005, 25 (4),

118 (130) El-Rafie, M. H.; El-Naggar, M. E.; Ramadan, M. A.; Fouda, M. M.; Al-Deyab, S. S.; Hebeish, A. Environmental synthesis of silver nanoparticles using hydroxypropyl starch and their characterization. Carbohydr. Polym. 2011, 86 (2), (131) Zhou, M.; Wei, Z.; Qiao, H.; Zhu, L.; Yang, H.; Xia, T. Particle size and pore structure characterization of silver nanoparticles prepared by confined arc plasma. J. Nanomater. 2009, 2009, 3. (132) Verhoff, F. H.; Spradlin, J. E. Mass and energy balance analysis of metabolic pathways applied to citric acid production by Aspergillus niger. Biotechnol. Bioeng. 1976, 18 (3), (133) Tariq, V.-N.; Stefani, L. A. J.; Butcher, A. C. Citric acid production: a practical introduction to biotechnology for first-year undergraduate students. Biochem. Educ. 1995, 23 (3), (134) Tolaymat, T. M.; El Badawy, A. M.; Genaidy, A.; Scheckel, K. G.; Luxton, T. P.; Suidan, M. An evidence-based environmental perspective of manufactured silver nanoparticle in syntheses and applications: A systematic review and critical appraisal of peer-reviewed scientific papers. Sci. Total Environ. 2010, 408 (5), (135) Song, K. C.; Lee, S. M.; Park, T. S.; Lee, B. S. Preparation of colloidal silver nanoparticles by chemical reduction method. Korean J. Chem. Eng. 2009, 26 (1), (136) Drogat, N.; Granet, R.; Sol, V.; Krausz, P. One-Pot Silver Nanoring Synthesis. Nanoscale Res. Lett. 2010, 5 (3), (137) Health, U. D. of; Services, H.; others. Hazardous substances data bank (HSDB, online database). Natl. Toxicol. Inf. Program Natl. Libr. Med. Bethesda MD (138) Slistan-Grijalva, A.; Herrera-Urbina, R.; Rivas-Silva, J. F.; Avalos-Borja, M.; Castillon- Barraza, F. F.; Posada-Amarillas, A. Assessment of growth of silver nanoparticles synthesized from an ethylene glycol silver nitrate polyvinylpyrrolidone solution. Phys. E Low-Dimens. Syst. Nanostructures 2005, 25 (4), (139) Carotenuto, G.; Pepe, G. P.; Nicolais, L. Preparation and characterization of nano-sized Ag/PVP composites for optical applications. Eur. Phys. J. B-Condens. Matter Complex Syst. 2000, 16 (1), (140) Li, J.; Zhu, J.; Liu, X. Ultrafine silver nanoparticles obtained from ethylene glycol at room temperature: catalyzed by tungstate ions. Dalton Trans. 2014, 43 (1), (141) Budavari, S.; O neil, M. J.; Smith, A.; Heckelman, P. E. The merck index: an encyclopedia of chemicals, drug, and biologicals (142) Hebbalalu, D.; Lalley, J.; Nadagouda, M. N.; Varma, R. S. Greener techniques for the synthesis of silver nanoparticles using plant extracts, enzymes, bacteria, biodegradable polymers, and microwaves. ACS Sustain. Chem. Eng. 2013, 1 (7), (143) Hebeish, A.; El-Rafie, M. H.; El-Sheikh, M. A.; El-Naggar, M. E. Nanostructural features of silver nanoparticles powder synthesized through concurrent formation of the nanosized particles of both starch and silver. J. Nanotechnol. 2013, (144) James, D. H.; Castor, W. M. Kirk-Othmer Encyclopedia of Chemical Technology; John Wiley & Sons: New York, (145) Chen, J.; Lu, G.; Zhu, L.; Flagan, R. C. A simple and versatile mini-arc plasma source for nanocrystal synthesis. J. Nanoparticle Res. 2007, 9 (2), (146) Chazelas, C.; Coudert, J.-F.; Jarrige, J.; Fauchais, P. Synthesis of ultra fine particles by plasma transferred arc: Influence of anode material on particle properties. J. Eur. Ceram. Soc. 2006, 26 (16), (147) Mäkelä, J. M.; Keskinen, H.; Forsblom, T.; Keskinen, J. Generation of metal and metal oxide nanoparticles by liquid flame spray process. J. Mater. Sci. 2004, 39 (8),

119 (148) Wei, Z.; Xia, T.; Bai, L.; Wang, J.; Wu, Z.; Yan, P. Efficient preparation for Ni nanopowders by anodic arc plasma. Mater. Lett. 2006, 60 (6), (149) Vigneshwaran, N.; Nachane, R. P.; Balasubramanya, R. H.; Varadarajan, P. V. A novel onepot green synthesis of stable silver nanoparticles using soluble starch. Carbohydr. Res. 2006, 341 (12), (150) Hebeish, A.; El-Naggar, M. E.; Fouda, M. M.; Ramadan, M. A.; Al-Deyab, S. S.; El-Rafie, M. H. Highly effective antibacterial textiles containing green synthesized silver nanoparticles. Carbohydr. Polym. 2011, 86 (2), (151) Buşilă, M.; Muşat, V.; Textor, T.; Mahltig, B. Synthesis and characterization of antimicrobial textile finishing based on Ag: ZnO nanoparticles/chitosan biocomposites. RSC Adv. 2015, 5 (28), (152) EPA. Nanotechnology White Paper; EPA, United States, 2005; p 134. (153) Sotiriou, G. A.; Pratsinis, S. E. Antibacterial activity of nanosilver ions and particles. Environ. Sci. Technol. 2010, 44 (14), (154) Vance, M. E.; Kuiken, T.; Vejerano, E. P.; McGinnis, S. P.; Hochella, M. F.; Rejeski, D.; Hull, M. S. Nanotechnology in the real world: Redeveloping the nanomaterial consumer products inventory. Beilstein J. Nanotechnol. 2015, 6, (155) Chaloupka, K.; Malam, Y.; Seifalian, A. M. Nanosilver as a new generation of nanoproduct in biomedical applications. Trends Biotechnol. 2010, 28 (11), (156) Griffitt, R. J.; Luo, J.; Gao, J.; Bonzongo, J.-C.; Barber, D. S. Effects of particle composition and species on toxicity of metallic nanomaterials in aquatic organisms. Environ. Toxicol. Chem. 2008, 27 (9), (157) Asharani, P. V.; Wu, Y. L.; Gong, Z.; Valiyaveettil, S. Toxicity of silver nanoparticles in zebrafish models. Nanotechnology 2008, 19 (25), (158) Choi, O.; Hu, Z. Size dependent and reactive oxygen species related nanosilver toxicity to nitrifying bacteria. Environ. Sci. Technol. 2008, 42 (12), (159) Pal, S.; Tak, Y. K.; Song, J. M. Does the antibacterial activity of silver nanoparticles depend on the shape of the nanoparticle? A study of the gram-negative bacterium Escherichia coli. Appl. Environ. Microbiol. 2007, 73 (6), (160) El Badawy, A. M.; Silva, R. G.; Morris, B.; Scheckel, K. G.; Suidan, M. T.; Tolaymat, T. M. Surface charge-dependent toxicity of silver nanoparticles. Environ. Sci. Technol. 2010, 45 (1), (161) Xiu, Z.; Zhang, Q.; Puppala, H. L.; Colvin, V. L.; Alvarez, P. J. Negligible particle-specific antibacterial activity of silver nanoparticles. Nano Lett. 2012, 12 (8), (162) Sánchez-Valdes, S.; Ortega-Ortiz, H.; Ramos-de Valle, L. F.; Medellín-Rodríguez, F. J.; Guedea-Miranda, R. Mechanical and antimicrobial properties of multilayer films with a polyethylene/silver nanocomposite layer. J. Appl. Polym. Sci. 2009, 111 (2), (163) Chou, K.-S.; Huang, K.-C.; Lee, H.-H. Fabrication and sintering effect on the morphologies and conductivity of nano-ag particle films by the spin coating method. Nanotechnology 2005, 16 (6), 779. (164) Aslan, K.; Leonenko, Z.; Lakowicz, J. R.; Geddes, C. D. Fast and slow deposition of silver nanorods on planar surfaces: application to metal-enhanced fluorescence. J. Phys. Chem. B 2005, 109 (8), (165) Zapata, P. A.; Tamayo, L.; Páez, M.; Cerda, E.; Azócar, I.; Rabagliati, F. M. Nanocomposites based on polyethylene and nanosilver particles produced by metallocenic in situ polymerization: synthesis, characterization, and antimicrobial behavior. Eur. Polym. J. 2011, 47 (8),

120 (166) Akamatsu, K.; Takei, S.; Mizuhata, M.; Kajinami, A.; Deki, S.; Takeoka, S.; Fujii, M.; Hayashi, S.; Yamamoto, K. Preparation and characterization of polymer thin films containing silver and silver sulfide nanoparticles. Thin Solid Films 2000, 359 (1), (167) Saulou, C.; Despax, B.; Raynaud, P.; Zanna, S.; Marcus, P.; Mercier-Bonin, M. Plasma deposition of organosilicon polymer thin films with embedded nanosilver for prevention of microbial adhesion. Appl. Surf. Sci. 2009, 256 (3), S35 S39. (168) Fu, J.; Ji, J.; Fan, D.; Shen, J. Construction of antibacterial multilayer films containing nanosilver via layer-by-layer assembly of heparin and chitosan-silver ions complex. J. Biomed. Mater. Res. A 2006, 79 (3), (169) Perelshtein, I.; Applerot, G.; Perkas, N.; Guibert, G.; Mikhailov, S.; Gedanken, A. Sonochemical coating of silver nanoparticles on textile fabrics (nylon, polyester and cotton) and their antibacterial activity. Nanotechnology 2008, 19 (24), (170) Radetić, M.; Ilić, V.; Vodnik, V.; Dimitrijević, S.; Jovan\vcić, P.; \vsaponjić, Z.; Nedeljković, J. M. Antibacterial effect of silver nanoparticles deposited on corona-treated polyester and polyamide fabrics. Polym. Adv. Technol. 2008, 19 (12), (171) Maneerung, T.; Tokura, S.; Rujiravanit, R. Impregnation of silver nanoparticles into bacterial cellulose for antimicrobial wound dressing. Carbohydr. Polym. 2008, 72 (1), (172) Rujitanaroj, P.; Pimpha, N.; Supaphol, P. Wound-dressing materials with antibacterial activity from electrospun gelatin fiber mats containing silver nanoparticles. Polymer 2008, 49 (21), (173) Damm, C.; Münstedt, H.; Rösch, A. The antimicrobial efficacy of polyamide 6/silver-nanoand microcomposites. Mater. Chem. Phys. 2008, 108 (1), (174) Wasif, A. I.; Laga, S. K. Use of nano silver as an antimicrobial agent for cotton. AUTEX Res J 2009, 9 (1), (175) Dastjerdi, R.; Montazer, M.; Shahsavan, S. A new method to stabilize nanoparticles on textile surfaces. Colloids Surf. Physicochem. Eng. Asp. 2009, 345 (1), (176) Gottschalk, F.; Sonderer, T.; Scholz, R. W.; Nowack, B. Modeled Environmental Concentrations of Engineered Nanomaterials (TiO2, ZnO, Ag, CNT, Fullerenes) for Different Regions. Environ. Sci. Technol. 2009, 43 (24), (177) Meyer, D. E.; Curran, M. A.; Gonzalez, M. A. An examination of silver nanoparticles in socks using screening-level life cycle assessment. J. Nanoparticle Res. 2011, 13 (1), (178) SILVERCEL Non-Adherent Made Easy - Wounds International (accessed Mar 15, 2016). (179) Bhattarai, N.; Zhang, M. Controlled synthesis and structural stability of alginate-based nanofibers. Nanotechnology 2007, 18 (45), (180) Yamada, M.; Sugaya, S.; Seki, M. Microfluidic synthesis of complex alginate fibers for the direction control of cell growth. Proc MicroTAS , 866. (181) Papageorgiou, S. K.; Katsaros, F. K.; Kouvelos, E. P.; Nolan, J. W.; Le Deit, H.; Kanellopoulos, N. K. Heavy metal sorption by calcium alginate beads from Laminaria digitata. J. Hazard. Mater. 2006, 137 (3), (182) Shin, S.-J.; Park, J.-Y.; Lee, J.-Y.; Park, H.; Park, Y.-D.; Lee, K.-B.; Whang, C.-M.; Lee, S.-H. On the fly continuous generation of alginate fibers using a microfluidic device. Langmuir 2007, 23 (17),

121 111 (183) Fenoradosoa, T. A.; Ali, G.; Delattre, C.; Laroche, C.; Petit, E.; Wadouachi, A.; Michaud, P. Extraction and characterization of an alginate from the brown seaweed Sargassum turbinarioides Grunow. J. Appl. Phycol. 2010, 22 (2), (184) Basha, N. S.; Rekha, R.; Letensie, A.; Mensura, S. Preliminary investigation on sodium alginate extracted from Sargassum subrepandum of Red Sea of Eritrea as Tablet Binder. J. Sci. Res. 2011, 3 (3), (185) Frank, E. D.; Han, J.; Palou-Rivera, I.; Elgowainy, A.; Wang, M. Q. Life-cycle analysis of algal lipid fuels with the GREET model. Cent. Transp. Res. Energy Syst. Div. Argonne Natl. Lab. Oak Ridge 2011, (186) Queen, D.; Walker, M.; Parsons, D.; Rondas, A. Aquacel Ag Dressings Made Easy. Wounds Int. 2011, 2 (2). (187) Echegoyen, Y.; Nerín, C. Nanoparticle release from nano-silver antimicrobial food containers. Food Chem. Toxicol. 2013, 62, (188) Tulve, N. S.; Stefaniak, A. B.; Vance, M. E.; Rogers, K.; Mwilu, S.; LeBouf, R. F.; Schwegler- Berry, D.; Willis, R.; Thomas, T. A.; Marr, L. C. Characterization of silver nanoparticles in selected consumer products and its relevance for predicting children s potential exposures. Int. J. Hyg. Environ. Health 2015, 218 (3), (189) Wu, J.; Zheng, Y.; Song, W.; Luan, J.; Wen, X.; Wu, Z.; Chen, X.; Wang, Q.; Guo, S. In situ synthesis of silver-nanoparticles/bacterial cellulose composites for slow-released antimicrobial wound dressing. Carbohydr. Polym. 2014, 102, (190) Kaegi, R.; Voegelin, A.; Sinnet, B.; Zuleeg, S.; Hagendorfer, H.; Burkhardt, M.; Siegrist, H. Behavior of metallic silver nanoparticles in a pilot wastewater treatment plant. Environ. Sci. Technol. 2011, 45 (9), (191) Som, C.; Berges, M.; Chaudhry, Q.; Dusinska, M.; Fernandes, T. F.; Olsen, S. I.; Nowack, B. The importance of life cycle concepts for the development of safe nanoproducts. Toxicology 2010, 269 (2), (192) Miseljic, M.; Olsen, S. I. Life-cycle assessment of engineered nanomaterials: a literature review of assessment status. J. Nanoparticle Res. 2014, 16 (6), (193) Bauer, C.; Buchgeister, J.; Hischier, R.; Poganietz, W. R.; Schebek, L.; Warsen, J. Towards a framework for life cycle thinking in the assessment of nanotechnology. J. Clean. Prod. 2008, 16 (8-9), (194) MarketsandMarkets. Carbon Nanotubes Market by Type & Application MarketsandMarkets (accessed Dec 10, 2015). (195) Zhirnov, V.; Herr, D.; Meyyappan, M. Electronic applications of carbon nanotubes become closer to reality. J. Nanoparticle Res. 1999, 1 (1), (196) Li, N.; Huang, Y.; Du, F.; He, X.; Lin, X.; Gao, H.; Ma, Y.; Li, F.; Chen, Y.; Eklund, P. C. Electromagnetic interference (EMI) shielding of single-walled carbon nanotube epoxy composites. Nano Lett. 2006, 6 (6), (197) Rawal, S.; Brantley, J.; Karabudak, N. Development of carbon nanotube-based composite for spacecraft components. In Recent Advances in Space Technologies (RAST), th International Conference on; IEEE, 2013; pp (198) Dahlben, L.; Isaacs, J. A. Environmental Assessment of Manufacturing with Carbon Nanotubes. In IEEE International Symposium on Sustainable Systems and Technology; Institute of Electrical and Electronics Engineers Inc., (199) Healy, M.; Dahlben, L.; Isaacs, J. Environmental Assessment of Single Walled Carbon Nanotube Processes. J. Ind. Ecol. 2008, 12 (3),

122 (200) Isaacs, J. A.; Tanwani, A.; Healy, M. L. Environmental assessment of SWNT production. In Electronics and the Environment, Proceedings of the 2006 IEEE International Symposium on; IEEE, 2006; pp (201) Meyer, D. E.; Curran, M. A.; Gonzalez, M. A. An Examination of Existing Data for the Industrial Manufacture and Use of Nanocomponents and Their Role in the Life Cycle Impact of Nanoproducts. Environ. Sci. Technol. 2009, 43 (5), (202) Köhler, A. R.; Som, C.; Helland, A.; Gottschalk, F. Studying the potential release of carbon nanotubes throughout the application life cycle. J. Clean. Prod. 2008, 16 (8-9), (203) Khanna, V.; Bakshi, B. R.; Lee, L. J. Carbon nanofiber production. J. Ind. Ecol. 2008, 12 (3), (204) Merugula, L.; Khanna, V.; Bakshi, B. R.; others. Comparative Life Cycle Assessment: Reinforcing wind turbine blades with carbon nanofibers. In Sustainable Systems and Technology (ISSST), 2010 IEEE International Symposium on; IEEE, 2010; pp 1 6. (205) Roes, A. L.; Marsili, E.; Nieuwlaar, E.; Patel, M. K. Environmental and cost assessment of a polypropylene nanocomposite. J. Polym. Environ. 2007, 15 (3), (206) Lloyd, S. M.; Lave, L. B. Life cycle economic and environmental implications of using nanocomposites in automobiles. Environ. Sci. Technol. 2003, 37 (15), (207) Chung, D. D. L. Materials for electromagnetic interference shielding. J. Mater. Eng. Perform. 2000, 9 (3), (208) Huang, Y.; Li, N.; Ma, Y.; Du, F.; Li, F.; He, X.; Lin, X.; Gao, H.; Chen, Y. The influence of single-walled carbon nanotube structure on the electromagnetic interference shielding efficiency of its epoxy composites. Carbon 2007, 45 (8), (209) Thomassin, J.-M.; Pagnoulle, C.; Bednarz, L.; Huynen, I.; Jerome, R.; Detrembleur, C. Foams of polycaprolactone/mwnt nanocomposites for efficient EMI reduction. J. Mater. Chem. 2008, 18 (7), (210) Turer, I.; Aydin, K. Electromagnetic shielding properties of satellites. In Recent Advances in Space Technologies (RAST), th International Conference on; IEEE, 2015; pp (211) Gaier, J. R.; Terry, J. Emi shields made from intercalated graphite composites; National Aeronautics and Space Administration, 1995; Vol (212) Uhlig, T.; Sellmaier, F.; Schmidhuber, M. Spacecraft Operations; Springer, (213) Macdonald, M.; Badescu, V. The International Handbook of Space Technology; Springer, (214) Yang, S.; Lozano, K.; Lomeli, A.; Foltz, H. D.; Jones, R. Electromagnetic interference shielding effectiveness of carbon nanofiber/lcp composites. Compos. Part Appl. Sci. Manuf. 2005, 36 (5), (215) Geetha, S.; Satheesh Kumar, K. K.; Rao, C. R.; Vijayan, M.; Trivedi, D. C. EMI shielding: Methods and materials A review. J. Appl. Polym. Sci. 2009, 112 (4), (216) Al-Saleh, M. H.; Sundararaj, U. Electromagnetic interference shielding mechanisms of CNT/polymer composites. Carbon 2009, 47 (7), (217) Steering Committee for NASA Technology Roadmaps. NASA Space Technology Roadmaps and Priorities: Restoring NASA s Technological Edge and Paving the Way for a New Era in Space; National Academies Press, (218) Yang, Y.; Gupta, M. C.; Dudley, K. L.; Lawrence, R. W. Novel carbon nanotube-polystyrene foam composites for electromagnetic interference shielding. Nano Lett. 2005, 5 (11), (219) Ryschkewitsch, M. G. Mitigating in Space Charging Effects A Guideline. Natl. Aeronaut. Space Adm. Tech. Handb. NASA-HDBK-4002A

123 (220) Cornara, S.; Beech, T.; Belló-Mora, M.; Martinez de Aragon, A. Satellite constellation launch, deployment, replacement and end-of-life strategies (221) Lutz, E.; Werner, M.; Jahn, A. Satellite Systems for Personal and Broadband Communications; Springer Science & Business Media, (222) Pritchard, W. L. Estimating the mass and power of communications satellites. Int. J. Satell. Commun. 1984, 2 (2), (223) UCS Satellite Database (accessed Dec 16, 2015). (224) Orbital ATK. Iridium NEXT The Next-generation Satellite Constellation of Iridium Communications Inc. (225) Ward, T. A. Aerospace Propulsion Systems; John Wiley & Sons, (226) Pelton, J. N.; Oslund, R. J.; Marshall, P. Communications Satellites: Global Change Agents; Routledge, (227) Chartrand, M. R. Satellite Communications for the Nonspecialist; SPIE Press, (228) Burkhardt, H.; Sippel, M.; Herbertz, A.; Klevanski, J. Comparative study of kerosene and methane propellant engines for reusable liquid booster stages. In 4th Conference (International) on Launcher Technology Space Launcher Liquid Propulsion. Liege; (229) Patel, M. Cumulative energy demand (CED) and cumulative CO 2 emissions for products of the organic chemical industry. Energy 2003, 28 (7), (230) Huijbregts, M. A.; Rombouts, L. J.; Hellweg, S.; Frischknecht, R.; Hendriks, A. J.; van de Meent, D.; Ragas, A. M.; Reijnders, L.; Struijs, J. Is cumulative fossil energy demand a useful indicator for the environmental performance of products? Environ. Sci. Technol. 2006, 40 (3), (231) Huijbregts, M. A.; Hellweg, S.; Frischknecht, R.; Hendriks, H. W.; Hungerbu hler, K.; Hendriks, A. J. Cumulative energy demand as predictor for the environmental burden of commodity production. Environ. Sci. Technol. 2010, 44 (6), (232) DoD. MIL-HDBK-17-2F: Composite Materials Handbook. Dep. Def. Handb. U. S. 2002, 17. (233) Lakshmi, M. S.; Reddy, B. S. R. Synthesis and characterization of new epoxy and cyanate ester resins. Eur. Polym. J. 2002, 38 (4), (234) Eller, K.; Henkes, E.; Rossbacher, R.; Höke, H. Amines, Aliphatic. In Ullmann s Encyclopedia of Industrial Chemistry; Wiley-VCH Verlag GmbH & Co. KGaA, (235) Morris, J.; Kovács, L.; Ohe, K. Cyanogen Bromide. In Encyclopedia of Reagents for Organic Synthesis; John Wiley & Sons, Ltd, (236) Gannon, P. Revise AS Chemistry for AQA; Heinemann, (237) Althaus, H. J.; Chudacoff, M.; Hischier, R.; Jungbluth, N.; Osses, M.; Primas, A. Life cycle inventories of chemicals. Ecoinvent report No. 8, v2.0; 8; Swiss Centre for Life Cycle Inventories: EMPA Dubendorf, (238) Das, S. Life cycle assessment of carbon fiber-reinforced polymer composites. Int. J. Life Cycle Assess. 2011, 16 (3), (239) Cain, B. L.; Croisant, W. J. Electrical contact resistance of titanium nitride coatings for electromagnetic shielding applications. Surf. Coat. Technol. 1996, 82 (1), (240) Kushnir, D.; Sandén, B. A. Energy Requirements of Carbon Nanoparticle Production. J. Ind. Ecol. 2008, 12 (3), (241) Song, Y. S.; Youn, J. R.; Gutowski, T. G. Life cycle energy analysis of fiber-reinforced composites. Compos. Part Appl. Sci. Manuf. 2009, 40 (8), (242) Chatterjee, A.; Gillespie, J. W. Room temperature-curable VARTM epoxy resins: Promising alternative to vinyl ester resins. J. Appl. Polym. Sci. 2010, 115 (2),

124 (243) Chaowasakoo, T.; Sombatsompop, N. Mechanical and morphological properties of fly ash/epoxy composites using conventional thermal and microwave curing methods. Compos. Sci. Technol. 2007, 67 (11), (244) Pillai, V. K.; Beris, A. N.; Dhurjati, P. S. Implementation of model-based optimal temperature profiles for autoclave curing of composites using a knowledge-based system. Ind. Eng. Chem. Res. 1994, 33 (10), (245) Strong, A. B. Fundamentals of Composites Manufacturing, Second Edition: Materials, Methods and Applications; Society of Manufacturing Engineers, (246) Tanrattanakul, V.; Jaroendee, D. Comparison between microwave and thermal curing of glass fiber epoxy composites: Effect of microwave-heating cycle on mechanical properties. J. Appl. Polym. Sci. 2006, 102 (2), (247) Lee, W. I.; Springer, G. S. Microwave Curing of Composites. J. Compos. Mater. 1984, 18 (4), (248) Drobny, J. G. Ionizing Radiation and Polymers: Principles, Technology, and Applications; William Andrew, (249) Janke, C. J.; Norris, R. E.; Yarborough, K.; Havens, S. J.; Lopata, V. J. Critical parameters for electron beam curing of cationic epoxies and property comparison of electron beam cured cationic epoxies versus thermal cured resins and composites; Oak Ridge National Lab., TN (United States), (250) Berejka, A. J.; Eberle, C. Electron beam curing of composites in North America. Radiat. Phys. Chem. 2002, 63 (3), (251) ISSST 2015 Abstracts (accessed Feb 9, 2016). (252) Sandhu, K. S.; Singh, N. Some properties of corn starches II: Physicochemical, gelatinization, retrogradation, pasting and gel textural properties. Food Chem. 2007, 101 (4), (253) Ratnayake, W. S.; Otani, C.; Jackson, D. S. DSC enthalpic transitions during starch gelatinisation in excess water, dilute sodium chloride and dilute sucrose solutions. J. Sci. Food Agric. 2009, 89 (12), (254) Ratnayake, W. S.; Jackson, D. S. Gelatinization and solubility of corn starch during heating in excess water: new insights. J. Agric. Food Chem. 2006, 54 (10), (255) Biliaderis, C. G.; Maurice, T. J.; Vose, J. R. Starch gelatinization phenomena studied by differential scanning calorimetry. J. Food Sci. 1980, 45 (6), (256) Coral, D. F.; Pineda-Gómez, P.; Rosales-Rivera, A.; Rodriguez-Garcia, M. E. Determination of the gelatinization temperature of starch presented in maize flours. In Journal of Physics: Conference Series; IOP Publishing, 2009; Vol. 167, p (257) Moseley, J. D.; Kappe, C. O. A critical assessment of the greenness and energy efficiency of microwave-assisted organic synthesis. Green Chem. 2011, 13 (4),

125 115 Appendix A Environmental Life Cycle Assessment of Nanosilver Enabled Bandages These data have been published as Supporting Information to the study Pourzahedi, L., & Eckelman, M. J. (2014). Environmental life cycle assessment of nanosilverenabled bandages. Environmental science & technology, 49(1),

126 116 Table A1. Summary of the created life cycle inventory, for production of 1 kg AgNP. Method Source Section Material Amount Unit Corresponding LCI unit process Reactive Magnetron Sputtering with Ar-N2 Pierson et al., Input Silver 1 kg Silver, at regional storage/rer with US electricity U Argon g Argon, liquid, at plant/rer with US electricity U Nitrogen 1.04 g Nitrogen, liquid, at plant/rer with US electricity U Electricity 27.7 kwh Electricity, medium voltage, at grid/us with US electricity U Table A2. Acticoat 7 wound dressing material composition, 10 cm x 12.5 cm. Material Source Amount Unit Corresponding LCI unit process Acticoat7 3 layers of HDPE mesh Smith & Nephew g Adapted for HDPE from Walser et al., layers of non-woven Smith & Nephew g Walser et al., rayon/polyester Silver content Parsons et al., 0.13 g Modeled Packaging, 2 layers of supercalendered paper Estimated 3.13 g Paper, wood-containing, supercalendared (SC), at regional storage/rer with US electricity U Transportation Estimated tkm Transport, lorry >16t, fleet average/us- US-EI U Table A3. Modeled emissions and processes from the incineration plant. Material Amount Unit Corresponding LCI unit process Silver 130 mg Emissions to air> Silver Plastic incineration 1.39 g Disposal, plastics, mixture, 15.3% water, to municipal incineration/ch with US electricity U Paper incineration 3.13 g Disposal, packaging paper, 13.7% water, to municipal incineration/ch with US electricity U Table A4. Mean and coefficient of variation of uncertainty analysis for all synthesis routes. Impact category Unit RMS-AR-N Mean CV% Acidification kg SO2 eq % Carcinogenics CTUh 4.84E % Ecotoxicity CTUe % Eutrophication kg N eq % Fossil fuel depletion MJ surplus % Global warming kg CO2 eq % Non carcinogenics CTUh % Ozone depletion kg CFC-11 eq 9.52E % Respiratory effects kg PM2.5 eq % Smog kg O3 eq %

127 117 Table A5. Pedigree matrix defined for LCA processes. All distributions are assumed to be lognormal. Reactive Magnetron Sputtering with Ar-N2 Processes U1 U2 U3 U4 U5 U6 Ub SD^2 Silver NA Argon NA Nitrogen NA Electricity NA Acticoat7 Silver NA HDPE mesh NA non-woven rayon/polyester NA Packaging NA Transportation 5 5 NA NA NA NA HDPE Fabric HDPE granulate NA HDPE spinning NA Heat, natural gas NA Disposal, polyethylene NA HDPE Spinning Water, deionised NA Electricity, medium voltage NA Treatment, sewage NA Polyester fabric PET granulate NA PET spinning NA Transport, lorry >32t NA Heat, natural gas NA Transport, freight, rail NA Disposal, PET NA PET Spinning Water, deionised NA Electricity, medium voltage NA Treatment, sewage NA

128 kg CFC-11 eq Crude oil, at production onshore Crude oil, at production onshore Crude oil, at production onshore Crude oil, at production Transport, natural gas, pipeline, long distance Uranium, enriched 3.8%, at USEC enrichment plant Remaining processes Figure A1. Process contribution to ozone depletion for silver at regional storage.

129 kg CO2 eq Lignite, burned in power plant Diesel, burned in building machine Operation, aircraft, freight Hard coal, burned in industrial furnace 1-10MW Natural gas, burned in industrial furnace >100kW Operation, lorry >16t, fleet average Diesel, burned in diesel-electric generating set Hard coal, burned in power plant Hard coal, burned in power plant Quicklime, in pieces, loose, at plant Heavy fuel oil, burned in industrial furnace 1MW, non-modulating Natural gas, burned in power plant (23% shale) Parkes process crust, from desilverising of lead/glo US-EI U Hydrogen cyanide, at plant Natural gas, at consumer (23% shale blend) Clinker, at plant Refinery gas, burned in furnace Nitric acid, 50% in H2O, at plant Conventional natural gas well US at plant Hard coal, burned in power plant Remaining processes Figure A2. Process contribution to GWP for silver at regional storage.

130 kg O3 eq Blasting Diesel, burned in building machine Diesel, burned in diesel-electric generating set Operation, lorry >16t, fleet average Lignite, burned in power plant Operation, aircraft, freight Hard coal, burned in industrial furnace 1-10MW Figure A3. Process contribution to smog for silver at regional storage.

131 kg SO2 eq Copper, primary, at refinery Blasting Lignite, burned in power plant Parkes process crust, from desilverising of lead Natural gas, at production Diesel, burned in building machine Diesel, burned in diesel-electric generating set Hard coal, burned in industrial furnace 1-10MW Operation, lorry >16t, fleet average Operation, aircraft, freight Figure A4. Process contribution to acidification for silver at regional storage.

132 kg N eq Disposal, sulfidic tailings, off-site Disposal, spoil from lignite mining, in surface landfill Blasting Remaining processes Figure A5. Process contribution to eutrophication for silver at regional storage.

133 CTUh Copper, primary, at refinery Ferrochromium, high-carbon, 68% Cr, at plant Disposal, lead smelter slag, 0% water, to residual material landfill Disposal, slag, unalloyed electr. steel, 0% water, to residual material landfill Disposal, lignite ash, 0% water, to opencast refill Parkes process crust, from desilverising of lead Disposal, basic oxygen furnace wastes, 0% water, to residual material landfill Disposal, sulfidic tailings, off-site Disposal, redmud from bauxite digestion, 0% water, to residual material landfill Transmission network, electricity, medium voltage Disposal, sludge from steel rolling, 20% water, to residual material landfill Disposal, dust, alloyed EAF steel, 15.4% water, to residual material landfill Remaining processes Figure A6. Process contribution to carcinogenics for silver at regional storage.

134 CTUh Copper, primary, at refinery Parkes process crust, from desilverising of lead Lead concentrate, at beneficiation Silver, from combined metal production, at refinery Silver, from combined gold-silver production, at refinery Remaining processes Figure A7. Process contribution to non-carcinogenics for silver at regional storage.

135 kg PM2.5 eq Copper concentrate, at beneficiation Lignite, burned in power plant Copper, primary, at refinery Parkes process crust, from desilverising of lead Diesel, burned in building machine Blasting Diesel, burned in diesel-electric generating set Natural gas, at production Hard coal, burned in industrial furnace 1-10MW Lead concentrate, at beneficiation Ferrochromium, high-carbon, 68% Cr, at plant Operation, lorry >16t, fleet average Heavy fuel oil, burned in industrial furnace 1MW, non-modulating Remaining processes Figure A8. Process contribution to respiratory effects for silver at regional storage.

136 CTUe Copper, primary, at refinery Disposal, lead smelter slag, 0% water, to residual material landfill Natural gas, unprocessed, at extraction Lead concentrate, at beneficiation Ferrochromium, high-carbon, 68% Cr, at plant Heavy fuel oil, burned in industrial furnace 1MW, non-modulating Disposal, lignite ash, 0% water, to opencast refill Disposal, sulfidic tailings, off-site Figure A9. Process contribution to ecotoxicity for silver at regional storage.

137 MJ surplus Natural gas, unprocessed, at extraction Crude oil, at production onshore Crude oil, at production offshore Crude oil, at production onshore Crude oil, at production offshore Crude oil, at production onshore Crude oil, at production Hydrogen cyanide, at plant Figure A10. Process contribution to fossil fuel depletion for silver at regional storage.

138 % Silver Plastic incineration Packaging incineration 80% 60% 40% 20% 0% Figure A11. Process contribution for all TRACI impact categories of bandage incineration.

139 129 Appendix B Comparative Life Cycle Assessment of Silver Nanoparticle Synthesis Routes These data have been published as Supporting Information to the study Pourzahedi, L., & Eckelman, M. J. (2015). Comparative life cycle assessment of silver nanoparticle synthesis routes. Environmental Science: Nano, 2(4),

140 130 Table B1. Life cycle inventory of CR-TSC synthesis route. Method Section Material Amount Unit Corresponding LCI unit process Comment 1 kg AgNP CR with trisodium citrate Input Silver 1.57 kg Created, see below Sileikaite et al., 2006 nitrate Trisodium 0.80 kg Created, see below Sileikaite et al., 2006 Citrate Water 9277 kg Water, deionized, at plant/us* Sileikaite et al., 2006 Reaction water + dissolution water Heat kj Heat, unspecific, in chemical plant/us- US-EI U Solution heated to 100 C Values checked with Aspen Plus Output Citric acid 0.59 kg Emissions to water> Citric acid Sileikaite et al., 2006 Sodium 0.79 kg Emissions to water> Sodium Sileikaite et al., 2006 Nitrate nitrite Hydrogen kg Emissions to air> Hydrogen Sileikaite et al., 2006 Oxygen 0.12 kg Emissions to air> Oxygen Sileikaite et al., kg Silver nitrate 1 kg Trisodium citrate 1 kg Citric Acid Input Nitric acid 0.49 kg Nitric acid, 50% in H2O, at plant/us 3Ag + 4HNO3 (cold and diluted) --> 3AgNO3 + 2H2O + NO Silver 0.64 kg Silver, at regional storage/us 3Ag + 4HNO3 (cold and diluted) --> 3AgNO3 + 2H2O + NO Output Water 0.07 kg Emissions to air> Water 3Ag + 4HNO3 (cold and diluted) --> 3AgNO3 + 2H2O + NO Nitrogen monoxide 0.06 kg Emissions to air> Nitrogen monoxide 3Ag + 4HNO3 (cold and diluted) --> 3AgNO3 + 2H2O + NO Input Citric acid 0.74 kg Created, see below C6H8O7 + 3NaOH --> Na3C6H5O7 + 3H2O Sodium hydroxide 0.46 kg Sodium hydroxide, production mix, at plant/kg NREL/RNA C6H8O7 + 3NaOH --> Na3C6H5O7 + 3H2O Output Water 0.26 kg Emissions to air> Water C6H8O7 + 3NaOH --> Na3C6H5O7 + 3H2O Input Sugar 0.94 kg Sugar, from sugarcane, at sugar refinery/us US-EI U C6H12O6 + 3/2 O2 --> C6H8O7 + 2 H2O, (Verhoff et al.) Oxygen 0.25 kg Oxygen, liquid, at plant/us* C6H12O6 + 3/2 O2 --> C6H8O7 + 2 H2O, (Verhoff et al.) Lime 0.57 kg Lime, hydrated, loose, at plant/us* Sulphuric acid 0.76 kg Sulphuric acid, liquid, at plant/us 3 Ca(OH)2 + 2 C6H8O7 --> Ca3(C6H5O7)2.4 H2O + 2 H2O, (Tariq et al.) 3 H2SO4 + Ca3(C6H5O7)2.4 H2O --> 2 C6H8O7 + 3 CaSO4.2 H2O + 2 H2O (Tariq et al.) Output Water 0.37 kg Emissions to air> Water Sum of water output water from above reactions. The initial acid production by fungi is an exothermic reaction. Gypsum 1.06 kg Disposal, gypsum, 19.4% water, to sanitary landfill/us* 3 H2SO4 + Ca3(C6H5O7)2.4 H2O --> 2 C6H8O7 + 3 CaSO4.2 H2O + 2 H2O (Tariq et al.)

141 131 Table B2. Life cycle inventory of CR-SB synthesis route. Method Section Material Amount Unit Corresponding LCI unit Comment process 1 kg AgNP Input Silver nitrate 1.57 kg Created, see Table S1 Solomon et al., 2007 CR with Sodium 0.35 kg Created, see below Solomon et al., 2007 sodium borohydride borohydride Water kg Water, deionized, at plant/us* Water for cooling m 3 Water, cooling, unspecified natural origin/m3 Solomon et al., 2007 Dissolution water Ecoinvent average cooling water Output Hydrogen kg Emissions to air> Solomon et al., 2007 Hydrogen Diborane 0.13 kg - Solomon et al., 2007 Sodium Nitrate 0.79 kg Emissions to water> Sodium nitrite Solomon et al., kg Sodium borohydride Input Trimethyl borate Sodium hydride 2.74 kg Trimethyl borate, at plant/glo 2.54 kg Sodium hydride, production mix, at plant/kg NREL/RNA B(OCH3)3 + 4NaH --> NaBH4 + 3NaOCH3 B(OCH3)3 + 4NaH --> NaBH4 + 3NaOCH3

142 132 Table B3. Life cycle inventory of CR-EG synthesis route. Method Section Material Amount Unit Corresponding LCI unit Comment process 1 kg AgNP Input Silver nitrate 1.57 kg Created, see Table S1 Slistan-Grijalva et al., 2005 CR with ethylene glycol Ethylene Glycol 29.1 kg Ethylene glycol, at plant/us Slistan-Grijalva et al., % recycling conidered PVP 47.2 kg Created, see below Slistan-Grijalva et al., 2005 Water 261 kg Water, deionized, at plant/us* Slistan-Grijalva et al., 2005 Dissolution water 1 kg PVP Input 1,4- butanediol 0.8 kg Butane-1,4-diol, at plant/us C4H10O2 --> C4H6O2 + 2H2, ( dehydrogenation of 1,4- butanediol over copper at 200 C forming gamma-butyrolactone) Ammonia 0.15 kg Ammonia, liquid, at regional storehouse/us* Acetylene 0.23 kg Acetylene, at regional storehouse/us* C4H6O2 +NH3 --> C4H7NO + H2O, ( gamma-butyrolactone reacting with ammonia yields pyrrolidone) C4H7NO + C2H2 --> C6H9NO, (subsequent treatment of pyrrolidone with acetylene gives VP monomer) Hydrogen 0.04 kg Emissions to air> Hydrogen C4H10O2 --> C4H6O2 + 2H2 Water 0.16 kg Emissions to water> Water C4H6O2 +NH3 --> C4H7NO + H2O

143 133 Table B4. Life cycle inventory of CR-Starch synthesis route. Method Section Material Amount Unit Corresponding LCI unit process 1 kg AgNP CR with soluble starch Comment Input Potato starch 90 kg Potato starch, at plant/us** El-Rafie et al., 2011 Water kg Water, deionised, at plant/us* El-Rafie et al., 2011 Silver nitrate 1.57 kg Created, see Table S1 El-Rafie et al., 2011 Heat kj Heat, unspecific, in chemical plant/us- US-EI U El-Rafie et al., 2011 Heat for bringing to 70 C Checked with Aspen Plus

144 134 Table B5. Life cycle inventory of FSP synthesis route. Method Section Material Amount Unit Corresponding LCI unit Comment process 1 kg AgNP Input Oxygen 33.4 kg Oxygen, liquid, at Walser et al, 2011 FSP plant/us Methane 1.53 kg Methane, 96 vol-%, from Walser et al, 2011 biogas, at purification/us* Water 62.8 kg Water, deionised, at Walser et al, 2011 plant/us* Silver 2.35 kg Created, see below Walser et al, 2011 octanoate kg Created, see below Walser et al, 2011 ethylhexanoic acid Xylene 6.29 kg Xylene, at plant/us Walser et al, 2011 Electricity kwh Electricity, medium Walser et al, 2011 voltage, at grid/us Output Nitric oxide 0.4 kg Emissions to air> Nitric Walser et al, 2011 oxide Carbon 43.9 kg Emissions to air> Carbon Walser et al, 2011 dioxide dioxide, fossil Water 16.8 kg Emissions to water> Walser et al, 2011 Water Wastewater treatment 0.06 m3 Emissions to water> Waste water/m3 Walser et al, kg Silver octanoate 1 kg 2- ethylhexaxnoic acid 1 kg N- butyraldehyde Input Coconut oil 2.7 kg Crude coconut oil, at Walser et al, 2011 plant/ph Silver nitrate 2.57 kg Created, see Table S1 Walser et al, 2011 Avoided product Sodium hydroxide 0.38 kg Sodium hydroxide, 50% in H2O, production mix, at plant/us Water 0.9 kg Water, deionised, at plant/us* Steam kg Steam, for chemical processes, at plant/us Input n- butyraldehyde Output Walser et al, 2011 Walser et al, 2011 Walser et al, kg Created, see below Walser et al, 2011 Transport 10.8 tkm Transport, lorry 16-32t, EURRO 5/US Carbon 0.05 kg Emissions to air> Carbon dioxide dioxide, fossil Walser et al, 2011 Walser et al, 2011 Input Propylene 0.58 kg Propylene, at plant/us 2CH3CH = CH2 + 2CO + 2H2 --> CH3(CH2)2CHO + (CH3)2(CH)2O ( N to iso ration is 4:1, hence, iso is neglected), (Walser et al.) Carbon monoxide 0.38 kg Carbon monoxide, CO, at plant/us Hydrogen 0.03 kg Hydrogen, liquid, at plant/us 2CH3CH = CH2 + 2CO + 2H2 --> CH3(CH2)2CHO + (CH3)2(CH)2O ( N to iso ration is 4:1, hence, iso is neglected), (Walser et al.) 2CH3CH = CH2 + 2CO + 2H2 --> CH3(CH2)2CHO + (CH3)2(CH)2O ( N to iso ration is 4:1, hence, iso is neglected), (Walser et al.)

145 135 Table B6. Life cycle inventory of AP synthesis route. Method Section Material Amount Unit Corresponding LCI unit process Comment 1 kg AgNP AP Input Silver 1.00 kg Silver, at regional storage/us Zhou et al., 2009 Argon 7.4 kg Argon, liquid, at plant/us Zhou et al., 2009 Electricity kwh Electricity, medium voltage, at grid/us Zhou et al., 2009

146 136 Table B7. Life cycle inventory of RMS-AR-N synthesis route. Method Section Material Amount Unit Corresponding LCI unit process Comment 1 kg AgNP RMS-Ar-N Input Silver 1.00 kg Silver, at regional storage/us Pierson et al., 2005 Argon g Argon, liquid, at plant/us Pierson et al., 2005 Nitrogen 10.4 g Nitrogen, liquid, at plant/us Pierson et al., 2005 Electricity 27.8 kwh Electricity, medium voltage, at grid/us Pierson et al., 2005

147 137 Table B8. Acticoat 7 wound dressing material composition. Product Section Material Amount Unit Corresponding LCI unit process Comment Acticoat7 Input 3 layers of HDPE 0.69 g Created, see below Smith & Nephew mesh 2 layers of nonwoven 0.7 g Created, see below Smith & Nephew rayon/polyester Silver 0.13 g Modeled using one of the eight Parsons et al., 2005 synthesis routes Packaging, g Paper, wood-containing, Estimation layers of supercalendered paper supercalendred (SC), at regional storage/us* Transport tkm Transport, lorry >16t, fleet average/us Estimation 1 kg HDPE Input Polyethylene 1.05 kg Polyethylene, HDPE, granulate, at plant/us Heat, natural gas 9.7 MJ Heat, natural gas, at boiler modulating <100kW/US Assumed 5% loss of plastic, (adapted from Walser et al., 2011) Assumed HDPE fibers require the same heat input as PET fibers, (adapted from Walser et al., 2011) Water kg Water, deionised, at plant/us* Assumed HDPE fibers require the same water input as PET fibers, (adapted from Walser et al., 2011) Electricity 5.9 kwh Electricity, medium voltage, at grid/us Output Polyethylene 0.05 kg Disposal, polyethylene, 0.4% water, to municipal incineration/us* Water m 3 Treatment, sewage, to wastewater treatment, class 3/US* Assumed HDPE fibers require the same electricity input as PET fibers, (adapted from Walser et al., 2011) Assumed 5% loss of plastic, (adapted from Walser et al., 2011) 1 kg Polyester Input Output Polyethylene terephthalate 1.05 kg Polyethylene terephthalate, granulate, bottle grade, at plant/us 9.7 MJ Heat, natural gas, at boiler modulating <100kW/US Walser et al., 2011 Heat, natural Walser et al., 2011 gas Water kg Water, deionised, at plant/us* Walser et al., 2011 Electricity 5.9 kwh Electricity, medium voltage, at Walser et al., 2011 grid/us Transport lorry 9.7 tkm Heat, natural gas, at boiler Walser et al., 2011 modulating <100kW/US Transport freight 0.2 tkm Transport, freight, rail/us Walser et al., 2011 Polyethylene 0.05 kg Disposal, polyethylene Walser et al., 2011 terephtalate terephtalate, 0.2% water, to municipal incineration/us* Water m 3 Treatment, sewage, to Walser et al., 2011 wastewater treatment, class 3/US*

148 138 Table B9. Comparative TRACI environmental impacts of all methods. IMPACT CATEGORY UNIT CR-EG CR-SB CR-TSC CR- STARCH FSP RMS-AR-N AP OD kg CFC-11 eq 2.27E E E E E E E-05 GW kg CO2 eq 4.32E E E E E E E+02 PS kg O3 eq 3.40E E E E E E E+01 AC kg SO2 eq 4.55E E E E E E E+00 EU kg N eq 7.19E E E E E E E+00 HHC CTUh 6.00E E E E E E E-05 HHNC CTUh 1.74E E E E E E E-04 HHCR kg PM2.5 eq 3.76E E E E E E E-01 EC CTUe 2.34E E E E E E E+03 FF MJ surplus 7.69E E E E E E E+02

149 % % 139 Silver source Reagents DI water Heat OD GW PS AC EU HHC HHNC HHCR EC FF Figure B1. Process contribution for all TRACI impact categories of CR-TSC synthesis route (silver source is silver nitrate, reagent is trisodium citrate). 100 Silver source Reagents DI Water OD GW PS AC EU HHC HHNC HHCR EC FF Figure B2. Process contribution for all TRACI impact categories of CR-SB synthesis route (silver source is silver nitrate, reagent is sodium borohydride).

150 % % 140 Silver source Reagents DI water OD GW PS AC EU HHC HHNC HHCR EC FF Figure B3. Process contribution for all TRACI impact categories of CR-EG synthesis route (silver source is silver nitrate, reagents are ethylene glycol and PVP). 100 Silver source Reagents DI water Heat OD GW PS AC EU HHC HHNC HHCR EC FF Figure B4. Process contribution for all TRACI impact categories of CR-Starch synthesis route (silver source is silver nitrate, reagent is soluble starch solution).

151 % % 141 Silver source Reagents DI water Gaseous elements Electricity Synthesis emissions OD GW PS ACF EU HHC HHNC HHCR EC FF Figure B5. Process contribution for all TRACI impact categories of FSP synthesis route (silver source is silver octanoate, reagents are 2-ethylhexanoic acid and xylene, gaseous elements are oxygen and methane). 100 Silver source Gaseous elements Electricity OD GW PS AC EU HHC HHNC HHCR EC FF Figure B6. Process contribution for all TRACI impact categories of AP synthesis route (silver source is solid silver, gaseous elements are argon and nitrogen).

152 % 142 Silver source Gaseous elements Electricity OD GW PS AC EU HHC HHNC HHCR EC FF Figure B7. Process contribution for all TRACI impact categories of RMS-AR-N synthesis route (silver source is solid silver, gaseous elements are argon and nitrogen). 100% Trisodium citrate Sodium Borohydride Potato starch Ethylene glycol 80% 60% 40% 20% 0% OD GW PS ACF EU HHC HHNC HHCR EC FF Figure B8. Comparison of reducing agents for all impact categories.

153 Comparison of Green Synthesis using Soluble Starch (CR-Starch) The method described in section is compared here with another reported CR- Starch technique, the main difference being the temperature at which the reaction occurs. 149 AgNPs in this method are produced using soluble starch as both the reducing and stabilizing agent for the silver nitrate solution. 149 Here, soluble starch is added to deionized water and heated in microwave until complete dissolution. The dissolved starch is used to reduce a 100 mm silver nitrate solution. The particles were formed in an autoclave at 15 psi and 121 C after 5 minutes. Energy consumed by the microwave in order to dissolve the starch in water was calculated using the average gelatinization temperature of starch (70 C), and an average efficiency for microwave equipment (50%). 257 Autoclave heat was also calculated for the solution mixture and checked for consistency using the chemical process optimization software Aspen Plus. Assuming 100% yield, to produce almost 1 kg of silver nano particles, 100 kg of potato starch is required alongside 1000 kg of water, to reduce 1.69 kg of silver nitrate. Life cycle inventory for this route can be found in Table S10. From figure S9, it can be seen that the significant heat and electricity consumption of this method outweighs the contribution of the other constituents. Figure S10 shows relative results between the two mentioned CR- Starch routes, and a scenario in which instead of autoclaving, the above reduction takes place at atmospheric conditions with a reaction temperature of 100C. The results emphasize the high dependency of impacts to the reaction conditions. For this synthesis route, changing the parameters such as temperature, heating duration, or silver nitrate concentration affects the diameter and the size distribution of the silver particles. 130 Lower temperatures and durations result in smaller particles. Modifications of the starch also directly affects its solubility in water, requiring less energy for dissolution. Hence, determining the application of these particles are crucial to estimating the overall cradleto-gate impacts of their synthesis.

154 % 144 Table B10. Life cycle inventory of autoclaved CR-Starch synthesis route. Method Section Material Amount Unit Corresponding LCI unit process 1 kg AgNP CR with soluble starch Input Potato starch 100 kg Potato starch, at plant/us** Water kg Water, deionised, at plant/us* Comment Vigneshwaren et al., 2006 Vigneshwaren et al., 2006 Silver nitrate 1.57 kg Created, see Table S1 Vigneshwaren et al., 2006 Heat kj Heat, unspecific, in chemical plant/us- US-EI U Electricity kj Electricity, medium voltage, at grid/us Vigneshwaren et al., 2006 Heat for bringing to psi Checked with Aspen Plus Vigneshwaren et al., 2006 Microwave electricity concumptionn 100% Silver source Reagents DI water Heat Electricity 80% 60% 40% 20% 0% OD GW PS ACF EU HHC HHNC HHCR EC FF Figure B9. Process contribution for all TRACI impact categories of autoclaved CR-Starch synthesis route (silver source is silver nitrate, reagent is soluble starch solution) CR-Starch, 70 C, 14.7 psi CR-Starch, 121 C, 15 psi CR-Starch, 100 C, 14.7 psi OD GW PS AC EU HHC HHNC HHCR EC FF Figure B10. Comparison of CR-Starch routes for all impact categories.

155 145 25% 20% 15% 10% 5% 0% Lignite, burned in power plant 1 Diesel, burned in building machine Operation, aircraft, freight Hard coal, burned in industrial furnace Natural gas, burned in industrial furnace Operation, lorry >16t Diesel, burned in diesel-electric generating set Hard coal, burned in power plant Hard coal, burned in power plant Quicklime, in pieces, loose, at plant Heavy fuel oil, burned in industrial furnace Natural gas, burned in power plant Parkes process crust, from desilverising of lead Remaining processes Figure B11. Process contribution to GWP for silver at regional storage. 100% 80% 60% 40% 20% 0% Disposal, sulfidic tailings, off-site Copper, primary, at refinery Disposal, lignite ash, to opencast refill Remaining processes 1 Disposal, lead smelter slag, to landfill Disposal, spoil from lignite mining, to landfill Disposal, slag, unalloyed electr.steel, to landfill Figure B12. Process contribution to ecotoxicity for silver at regional storage.

156 146 Appendix C The Life Cycle of Nanosilver-enabled Consumer Products: Investigating Hotspots

157 Figure C1. SEM imaging (left) and EDX analysis (right) of (a) Acticoat 7, (b) Silvercel, (c) Aquacel Ag, (d) Polymem, and (e) Sock #1 silver containing fabric. 147

158 148 Table C1. Life cycle inventory for a 10cm x 12.5cm x 0.1 mm piece of Acticoat 7 bandage Product Section Material Amount Unit LCI unit Process Comment Acticoat Input Nanosilver 0.13 g Created, see Pourzahedi et al. Parsons et al., (1 p) 3 layers of HDPE g Created, see below Smith & Nephew fabric 2 layers of 0.7 g Created, see below Smith & Nephew Polyester fabric 2 layers of paper 3.13 g Paper, wood-containing, Estimation packaging supercalendred (SC), at regional storage/us* Transport tkm Transport, lorry >16t, fleet Estimation average/us HDPE Input HDPE granulate 1.05 kg Polyethylene, HDPE, granulate, at Walser et al. fabric (1 kg) Spinning, knitting 1.05 kg plant/us- US-EI U HDPE spinning, knitting, making up Walser et al. and making up - Adapted Heat 9.7 MJ Heat, natural gas, at boiler modulating <100kW/US- US-EI U Output Disposal, HDPE 0.05 kg Disposal, polyethylene, 0.4% water, to municipal incineration/us* US-EI U Assumed HDPE fibres require the same heat input as PET fibres Assumed 5% loss of plastic (Walser et al.) HDPE Spinning, Knitting and making up (1 kg) PET fabric (1 kg) Input Water 24.8 kg Water, deionised, at plant/us* US- EI U Electricity 5.62 kwh Electricity, medium voltage, at grid/us US-EI U Input PET granulate 1.05 kg Polyethylene terephthalate, granulate, bottle grade, at plant/us- US-EI U PET Spinning, Knitting, and making up 1.05 kg PET spinning, knitting, making up - Walser Assumed HDPE fibres require the same water input as PET fibres, (Walser et al.) Assumed HDPE fibres require the same electricity input as PET fibres, (Walser et al.) Walser et al. Walser et al. Transport, lorry 0.1 tkm Transport, lorry >32t, EURO5/US- Walser et al. US-EI U Heat 9.7 MJ Heat, natural gas, at boiler Walser et al. modulating <100kW/US- US-EI U Transport, rail 0.2 tkm Transport, freight, rail/us- US-EI U Walser et al. Output Disposal, PET 0.05 kg Disposal, polyethylene terephtalate, 0.2% water, to municipal incineration/us* US-EI U Walser et al.

159 149 Table C2. Life cycle inventory for a 5cm x 5cm x 0.1 mm piece of Silvercel bandage Product Section Material Amount Unit LCI unit Process Comment Silvercel (1 Input Nanosilver mg Created, see Pourzahedi et al. Clark et al. & Parsons et p) al. Calcium Alginate 0.47 g Created see below Clark et al. fibers CMC fibers g Carboxymethyl cellulose, powder, at plant/us- S Clark et al. EMA polymer 0.25 g Created see below Clark et al. Paper packaging 1.25 g Paper, wood-containing, supercalendred (SC), at regional storage/us* US-EI U Assumed 2 layers of paper, 5.5cm x 5.5 cm x 0.1 mm, used average density of 1.16 g/cm3 Transport tkm Transport, lorry >16t, fleet average/us- US-EI U Estimated Calcium Alginate (1 kg) Sodium Alginate (1 kg) Algae growth and dewatering (1 kg, 2.46 kg algae gives 1 kg dried algae) Input Sodium alginate kg Created see below 2 NaC6H7O6 + CaCl2 -> Ca (C6H7O6)2 + 2NaCl Calcium chloride 0.28 kg Calcium chloride, CaCl2, at regional storage/us* US-EI U 2 NaC6H7O6 + CaCl2 -> Ca (C6H7O6)2 + 2NaCl Output Sodium chloride kg Emissions to water 2 NaC6H7O6 + CaCl2 -> Ca (C6H7O6)2 + 2NaCl Input Dry algae 10 kg Created see below Fenoradosoa et al. Formaldehyde 5.22 kg Formaldehyde, production mix, at Fenoradosoa et al. plant/us- US-EI U HCl 2.3 kg Hydrochloric acid, from the reaction of Fenoradosoa et al. hydrogen with chlorine, at plant/us- US-EI U Water kg Water, deionised, at plant/us* US-EI U Fenoradosoa et al. Na 2CO kg Sodium carbonate from ammonium C6H8O6 + Na2CO3 chloride production, at plant/glo US-EI NaC6H7O6 + NaHCO3 U Input Water kg Water, decarbonised, at plant/us- US-EI U Ammonia kg Ammonia, liquid, at regional storehouse/us- US-EI U Diammonium kg Diammonium phosphate, as P2O5, at Phosphate regional storehouse/us- US-EI U Diammonium Phosphate kg Diammonium phosphate, as N, at regional storehouse/us- US-EI U Electricity kwh Electricity, medium voltage, at grid/us US-EI U GREET GREET GREET GREET GREET EMA Input Water, cooling m3 Water, cooling, unspecified natural origin/m3 Adapted from ethylene vinyl acetate unit process Water, resource m3 Water, unspecified natural origin/m3 Adapted from EVA Ethylene 0.75 kg Ethylene, average, at plant/us- US-EI U Methyl acrylate 0.25 kg Methyl acrylate, at plant/glo US-EI U Adapted from EVA Heat 2 MJ Heat, natural gas, at industrial furnace Adapted from EVA >100kW/US- US-EI U Electricity kwh Electricity, medium voltage, production Adapted from EVA UCTE*, at grid/ucte US-EI U Transport, lorry tkm Transport, lorry >16t, fleet average/us- US-EI U Adapted from EVA Transport, rail tkm Transport, freight, rail/us- US-EI U Adapted from EVA Chemical plant 4E-10 p Chemical plant, organics/rer/i US-EI U Adapted from EVA

160 150 Table C3. Life cycle inventory for a 4.5cm x 4.5cm x 0.1 mm piece of Aquacel ag bandage Product Section Material Amount Unit LCI unit Process Comment Aquacel Ag (1 p) Input Nanosilver 2.55 mg Created, see Pourzahedi et al. (Queen et al. & Parsons et al.), derived from 12 mg Ag/g dressing CMC fibers g Carboxymethyl cellulose, powder, (Queen et al.) at plant/us- S Paper packaging 1.25 g Paper, wood-containing, supercalendred (SC), at regional storage/us* US-EI U Transport tkm Transport, lorry >16t, fleet average/us- US-EI U Assumed 2 layers of paper, 5.5cm x 5.5 cm x 0.1 mm, used average density of 1.16 g/cm3 Estimation Table C4. Life cycle inventory for a 10.8cm x 10.8cm x 3mm piece of Acticoat 7 bandage Product Section Material Amount Unit LCI unit Process Comment Polymem Input Nanosilver mg Created, see Pourzahedi et al. Parsons et al. (1 p) Polyurethane foam 7.56 g Polyurethane, flexible foam, at Parsons et al. plant/us- US-EI U Paper packaging 2.9 g Paper, wood-containing, supercalendred (SC), at regional storage/us* US-EI U Assumed 2 layers of paper, 11 cm x 11` cm x 0.1 mm, used average density of 1.16 g/cm3 Table C5. Life cycle inventory for GoGreen food container Product Section Material Amount Unit LCI unit Process Comment Gogreen Input Nanosilver 0.44 g Created, see Pourzahedi et al. Echegoyen et al. food Polypropylene 137 g Polypropylene, granulate, at Echegoyen et al. container (1 p) granulate plant/us- US-EI U Injection molding 137 g Injection moulding/us- US-EI U Echegoyen et al.

161 151 Table C6. Life cycle inventory for sock #1 Product Section Material Amount Unit LCI unit Process Comment Sock #1 Input Nanosilver 31.2 mg Created, see Pourzahedi et al. Benn et al. (1 p) PP fabric 13.8 g Created see below Product description Nylon 66 fabric 9.2 g Created see below Product description PP fabric (1 kg) Input PP granulate 1.05 kg Polypropylene, granulate, at plant/us- US-EI U PP spinning, knitting, and making up Considered plastic loss of 5% (ELCD database 2.0) 1.05 kg Created see below Considered plastic loss of 5% (ELCD database 2.0) Heat 9.7 MJ Heat, natural gas, at boiler modulating <100kW/US- US-EI U Output Disposal PP 0.05 kg Disposal, polypropylene, 0% water, to municipal incineration/us* US-EI U Assumed PP fibres require the same heat input as PET (Walser et al.) Considered plastic loss of 5% (ELCD database 2.0) PP spinning, knitting, and making up (1 kg) Nylon 66 fabric (1 kg) Input Electricity 1.8 MJ Electricity, medium voltage, at grid/us US-EI U Input Nylon 66 granulate Nylon 66 spinning, knitting, and making up Average energy requirements per kg polyamide/nylon fibres (ELCD database 2.0) kg Nylon 66, at plant/rer U Considered plastic loss of 1.5% (ELCD database 2.0) kg Created see below Considered plastic loss of 1.5% (ELCD database 2.0) Heat 9.7 MJ Heat, natural gas, at boiler modulating <100kW/US- US-EI U Output Disposal kg Disposal, plastics, mixture, 0% water, to municipal incineration/us* US-EI U Assumed Nylon fibres require the same heat input as PET fibres, (Walser et al.) Considered plastic loss of 1.5% (ELCD database 2.0) Nylon 66 spinning, knitting, and making up (1 kg) Input Electricity 7.5 MJ Electricity, medium voltage, at grid/us US-EI U Average energy requirements per kg polyamide/nylon fibres (ELCD database 2.0)

162 152 Table C7. Life cycle inventory for baby blanket Product Section Material Amount Unit LCI unit Process Comment Baby Input Nanosilver 7.69 mg Created, see Pourzahedi et al. Quadros et al. blanket (1 p) Fleece 70 g Fleece, polyethylene, at plant/us- US-EI U Quadros et al. Table C8. Life cycle inventory for children s cup Product Section Material Amount Unit LCI unit Process Comment Children s Input Nanosilver 0.14 mg Created, see Pourzahedi et al. Quadros et al. cup (1 p) Polypropylene granulate 90 g Polypropylene, granulate, at plant/us- US-EI U Quadros et al. Injection molding 90 g Injection moulding/us- US-EI U Quadros et al. Table C9. Life cycle inventory for plush toy interior foam Product Section Material Amount Unit LCI unit Process Comment Plush toy Input Nanosilver 26 mg Created, see Pourzahedi et al. Benn et al. (1 p) Polystyrene foam 377 g Polystyrene, general purpose, GPPS, at plant/us- US-EI U Benn et al. Table C10. Life cycle inventory for medical mask Product Section Material Amount Unit LCI unit Process Comment Medical Input Nanosilver 590 mg Created, see Pourzahedi et al. Benn et al. mask (1 p) PP fabric 2.2 g Created see Table S6 Benn et al. Table C11. Life cycle inventory for medical cloth Product Section Material Amount Unit LCI unit Process Comment Medical Input Nanosilver 810 mg Created, see Pourzahedi et al. Benn et al. cloth (1 p) Cotton fabric 3.5 g Textile, woven cotton, at plant/glo US-EI U Benn et al.

163 153 Table C12. Life cycle inventory for towel Product Section Material Amount Unit LCI unit Process Comment Towel (1 Input Nanosilver 2.3 mg Created, see Pourzahedi et al. Benn et al. p) Nylon 66 fabric 1.72 g Created see Table S6 Estimated PET fabric 6.88 g Created see Table S1 Estimated Table C13. Life cycle inventory for T-shirt Product Section Material Amount Unit LCI unit Process Comment T-shirt (1 Input Nanosilver 5.3 mg Created, see Pourzahedi et al. Benn et al. p) PET fabric kg Created see Table S1 Estimated Table C14. Life cycle inventory for plastic food bag Product Section Material Amount Unit LCI unit Process Comment Plastic Input Nanosilver g Created, see Pourzahedi et al. Echegoyen et al. food bag (1 p) LDPE granulate 2.8 g Polyethylene, LDPE, granulate, at plant/us- US-EI U Estimated Plastic extrusion 2.8 g Extrusion, plastic film/us- US-EI U Table C15. Life cycle inventory for sock # 2 Product Section Material Amount Unit LCI unit Process Comment Sock #2 Input Nanosilver mg Created, see Pourzahedi et al. Benn et al. (1 p) Cotton fabric 29.3 g Textile, woven cotton, at Product description plant/glo US-EI U PET fabric g Created see Table S1 Product description Nylon 66 fabric g Created see Table S6 Product description

164 154 Table C16. Absolute values for impacts of products OD kg CFC-11 eq GW kg CO2 eq PS kg O3 eq AC kg SO2 eq EU kg N eq HHC CTUh HHNC CTUh HHCR kg PM2.5 eq EC CTUe FF MJ surplus Acticoat 7 4.5E E E E E E E E E E-01 Aquacel Ag 4.5E E E E E E E E E E-03 Socks # 2 2.8E E E E E E E E E E-01 Baby blanket Children's cup 1.8E E E E E E E E E E E E E E E E E E E E+00 Socks #1 2.5E E E E E E E E E E-01 Plastic bag Food container Medical cloth Medical mask 3.5E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E-01 T-shirt 6.2E E E E E E E E E E+00 Polymem 9.6E E E E E E E E E E-01 Silvercel 3.1E E E E E E E E E E-02 Plush toy 1.5E E E E E E E E E E+00 Towel 2.6E E E E E E E E E E-01

165 Figure C2. Evaluation of relationship between silver loading and contribution of silver to impact category. The X axis shows the loading percentage, the Y axis shows the percentage of impact by silver. 155

166 Figure C3. High and low boundaries for environmental impacts of AgNP, with varying synthesis method, columns showing impact levels of chemical reduction with trisodium citrate for producing 1 kg AgNPs 156