EPA/600/R-10/129 I September 2010 lwww.epa.gov/athens
vvEPA
United States
Environmental Protection
Agency
State-of-the-Science Report on
Predictive Models and Modeling
Approaches for Characterizing and
Evaluating Exposure to
Nanomaterials
Ecosystems Research Division
National Exposure Research Laboratory
Office of Research and Development
-------
EPA/600/R-10/129
September 2010
www.epa.gov/athens
State-of-the-Science Report on
Predictive Models and Modeling
Approaches for Characterizing and
Evaluating Exposure to Nanomaterials
by
John M. Johnston
Ecosystems Research Division
National Exposure Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
960 College Station Rd.
Athens, GA 30605
Michael Lowry, Stephen Beaulieu and Evan Bowles
RTI, International
3040 Cornwallis Road
Research Triangle Park, NC 27709
Contract No. EP-W-09-004
Work Assignment 1-17
EMRAD Risk Assessment Support to OSW
Project Officer:
Shannon Sturgeon
Office of Resource Conservation and Recovery
Washington, DC 20024
Work Assignment Manager:
Gerry Laniak
Office of Research and Development
Athens, GA 30605
National Exposure Research Laboratory
Office of Research and Development
Washington, DC 20460
-------
Notice and Disclaimer
The U.S. Environmental Protection Agency through its Office of Research and
Development (ORD), funded and collaborated in the research described here under
Contract No. EP-W-09-004, Work Assignment 1-17 to Research Triangle Institute
International. It has been subjected to the Agency's peer and administrative review and
has been approved for publication as an EPA document. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
This is a contribution to the EPA Office of Research and Development's Nanomaterials
Research Program.
The appropriate citation for this report is:
Johnston, J.M., Lowry, M., Beaulieu, S., and Bowles, E. 2010. State-of-the-Science
Report on Predictive Models and Modeling Approaches for Characterizing and
Evaluating Exposure to Nanomaterials. U.S. Environmental Protection Agency, Office of
Research and Development, Athens, GA. EPA/600/R-10/129, September 2010.
-------
Contents
Contents
Acronyms vii
Acknowledgements x
Executive Summary 1
Chapter 1.0 Introduction 4
1.1 Background 5
1.2 Purpose and Scope of this Report 8
1.3 Overview of Re view Methodology 10
1.3.1 Information Search and Review 10
1.3.2 Characterization of Models and Approaches 11
1.4 Roadmap to Report 12
Chapter 2.0 Nanomaterials and their Environmental Transport and Fate 15
2.1 Engineered Nanomaterials and their Classification 15
2.1.1 Carbonaceous ENMs 16
2.1.2 Metal ENMs 16
2.1.3 Semiconductor Materials, Including Quantum Dots 17
2.1.4 Nanopolymers/Dendrimers 17
2.2 Properties of ENMs that Influence Environmental Behavior 17
2.2.1 Size Characteristics 18
2.2.2 Surface Area and Charge Characteristics 18
2.2.3 Chemical Composition and Structure Characteristics 19
2.2.4 Reactivity Characteristics 19
2.2.5 Partitioning Characteristics 19
2.3 Key Processes Influencing Environmental Behavior 20
2.3.1 Aggregation and Deposition 20
2.3.2 Disaggregation and Detachment 21
2.3.3 Settling and Sedimentation 21
2.3.4 Filtering and Enhanced Transport in Porous Media 21
2.3.5 Particle Diffusion 22
2.3.6 Redox Reactions 22
2.3.7 Biodegradation 22
2.3.8 Hydrolysis 22
2.3.9 Photolysis 23
2.3.10 Phase Partitioning 23
2.4 Considerations for the Fate and Transport of ENMs in Environmental Media 24
2.4.1 Fate and Transport in Aquatic Systems 25
IV
-------
Contents
2.4.2 Fate and Transport in Terrestrial Systems 27
2.4.3 Uptake and Accumulation of Nanomaterials in Biological Systems 29
2.5 Challenges to Modeling Nanomaterials 29
2.5.1 Complexity of Transport Characteristics and Behaviors and Associated
Data Gaps 30
2.5.2 Variability in Nanomaterial Types and Properties 31
2.5.3 Limitations of Traditional Risk Assessment Models 31
2.5.4 Need for Near-Term Risk Management Decisions 32
Chapter 3.0 Review Methodology for Relevant Models and Methods 33
3.1 Developing the Search Strategy 33
3.2 Identifying Key Information Sources 35
3.2.1 Journals 35
3.2.2 Reports 38
3.2.3 Research Centers 40
3.2.4 Informational Web Sites 41
3.3 Summary of Recent Reports and Compendia 44
3.3.1 Summary Reports of State of Current Knowledge 44
3.3.2 Governance Frameworks 46
3.3.3 Other Relevant Nanomaterial Reports and Compendia 47
Chapter 4.0 Model Reviews 49
4.1 Model/Method Evaluation Criteria 49
4.2 Review of Environmental Fate and Transport Models 50
4.2.1 Surface Water Models 51
4.2.2 Subsurface Models 55
4.2.3 Bioaccumulation Models 58
4.2.4 Multimedia Models 61
4.3 Alternative Approaches 64
4.3.1 Adaptive Management and Evaluation Frameworks 66
4.3.2 Multi-Criteria Decision Analysis (MCDA) 68
4.3.3 Bayesian Approaches 70
Chapter 5.0 Conclusions 74
Chapter 6.0 Bibliography 78
Appendix A Titles Pertaining to the Use of Exposure Models for Nanomaterials 90
Appendix B Exposure Model/Method Summaries for NMs in the Environment 99
-------
Contents
List of Figures
Figure 1-1. A predictive modeling strategy to evaluate ENM risks 5
Figure 1-2. Source-to-outcome framework for ecological exposure research (US EPA,
2009a) 6
Figure 1-3. Framework for protecting human health and the environment (US EPA, 2009a) 9
Figure 1-4. Model review process for exposure models and alternative approaches 11
Figure 1-5. Report roadmap 12
Figure 2-1. Illustration of chemical processes relevant to transport in aquatic (surface
water) and terrestrial (groundwater, soils, and sediments) environmental systems 25
Figure 3-1. Summary of review methodology 33
Figure 3-2. Key search terms organized by theme 34
Figure 4-1. Conceptual exposure pathway utilized by the SMARTEN model 68
Figure 4-2. Conceptual network of ENM flows over value chain and into environmental
compartments 73
List of Tables
Table 2-1. Chemical Properties of Nanomaterials Relevant to Environmental Fate and
Transport 17
Table 2-2. Comparing the Properties Needed to Model ENMs Versus Other Contaminants 31
Table 3-1. Journals Producing the Highest Number of Relevant Studies 35
Table 3-2. Major Reports on ENM Research Relevant to Environmental Exposure
Modeling 38
Table 3-3. Key Research Centers for Nanotechnology Research 40
Table 3-4. Publically Available Websites Providing Valuable Information ENM Exposure
Modeling 41
Table 3-5. Summary Reports on ENM Research Relevant to Environmental Exposure
Modeling 45
Table 3-6. Governance Frameworks on the Production of Nanomaterials 46
Table 3-7. Other Relevant Reports and Compendia on the Research of Nanomaterials 48
Table 4-1. Summary Evaluation of Alternative Approaches to ENM Risk Evaluation 65
VI
-------
Acronyms
Acronyms
Acronym Definition
3MRA Multimedia, Multipathway, Multireceptor Risk Assessment
AAN average agglomeration number
ADE advection-dispersion equation
ADME absorption, distribution, metabolism, and elimination
AGU American Geophysical Union
AHP Analytical Hierarchy Process
ALARA as low as reasonably achievable
ANSI American National Standards Institute
ARAMS Army Risk Assessment Modeling System
ASTM American Society for Testing and Materials
BCF bioconcentration factor
BIOPLUME 2-D transport model of dissolved hydrocarbons
BSAF biota sediment accumulation factor
CalTOX California Total Exposure Model for Hazardous Waste Sites
CBEN Center for Biological and Environmental Nanotechnology
CEINT Center for Environmental Implications of NanoTechnology
CFC critical flocculation concentration
CFT classic filtration theory
CNT carbon nanotube
CSC critical salt concentration
DEFRA Department for Environment, Food and Rural Affairs
DIAS Dynamic Information Architecture System
DLVO Derjaguin and Landau, Verwey and Overbeek
EHP Environmental Health Perspectives
ENM Engineered nanomaterials
ENRHES Engineered Nanoparticles: Review of Health and Environmental Safely
EOR enhanced oil recovery
EPA U.S. Environmental Protection Agency
EPI Suite Estimation Programs Interface Suite
ERED Environmental Residue Effects Database
ES&T Environmental Science & Technology
ETH Eidgenossische Technische Hochschule (Swiss Federal Institute of
Technology)
FRAMES Framework for Risk Analysis in Multimedia Environmental Systems
GIS geographic information systems
HEAST Health Effects Assessment Summary Table
HSPF Hydrological Simulation Program-FORTRAN
HWIR Hazardous Waste Identification Rule
HYDRUS water flow and solute transport model in variably saturated porous media
ICON International Council on Nanotechnology
IOM Institute of Occupational Medicine
IRIS Integrated Risk Information System
vn
-------
Acronyms
ISO International Organization for Standardization
Krf the equilibrium ratio of the concentration in water to the concentration in the
solid
K;, the equilibrium ratio of the concentration in air to the concentration in water
Koc organic carbon partition coefficient
MAUT Multi-Attribute Utility Theory
MAVT Multi-Attribute Value Theory
MCDA multi-criteria decision analysis
MEMS MicroElectroMechanical Systems
MFA material flow analysis
MIMS Multimedia Integrated Modeling System
MO A modes of action
MOC Method of Characteristics (2-D transport model for groundwater)
MODFLOW finite difference flow model
MRC military relevant compounds
MWCNT multiwalled carbon nanotube
NCEA National Center for Environmental Assessment
NEMS NanoElectroMechanical Systems
NIOSH National Institute for Occupational Safety and Health
NM nanomaterial
NMT Nanotechnology, Molecular Nanotechnology
NNI National Nanotechnology Initiative
NNI Nanotechnology Now
NP nanoparticles
NPDES National Pollutant Discharge Elimination System
NRC National Research Council
NSF National Science Foundation
PAH polycyclic aromatic hydrocarbon
PEC predicted environmental concentration
PM precautionary matrix
PMFA probabilistic material flow analysis
PNEC predicted no effect environmental concentration
PRZM Pesticide Root Zone Model
PZC point of zero charge
QSAR quantitative structure-activity relationship
QUAL2K River and Stream Water Quality Model
RESRAD Residual Radioactivity Models
SAB Science Advisory Board
SADA Spatial Analysis Decision Assistance
SMAA-TRI stochastic multicriteria acceptability analysis
SMARTEN Strategic Management and Assessment of Risks and Toxicity of Engineered
Nanomaterial s
STELLA modeling software package
TMDL total maximum daily load
TOPSIS Techniques for Order Preference by Similarity to Ideal Solution
Vlll
-------
Acronyms
TOUGH2 general purpose numerical simulation program for porous and fractured media
TRACT Tools for the Reduction and Assessment of Chemical and Other
Environmental Impacts
TRIM Total Risk Integrated Methodology
UC CEIN University of California Center for Environmental Implications of
Nanotechnol ogy
UCLA University of California, Los Angeles
UK United Kingdom
UV ultraviolet
Vol value of information
WASP Water Quality Analysis Simulation Program
WRR Water Resources Research
WWTP wastewater treatment plant
-------
Acknowledgements
The authors would like to thank Michele Conlon (EPA ORD/NERL) for her vision and support
of this effort. The external peer reviewers for this report—Ms. Christine Hendren, Dr. Igor
Linkov, and Dr. Sam Luoma—provided a number of comments on the draft that significantly
improved the quality and transparency of the report. The authors would like to thank Dr. Mark
Wiesner for reviewing the report and providing helpful guidance, also for sharing information on
his research at the Center for the Environmental Implications of Nanotechnology at Duke
University, sponsored by the EPA and the National Science Foundation. Our thanks to Dr.
MacArthur Long (EPA ORD/NERL) for managing the external peer review.
-------
(This page intentionally left blank.)
XI
-------
Executive Summary
Executive Summary
This state-of-the-science review was undertaken to identify fate and transport models and
alternative modeling approaches that could be used to predict exposure to engineered
nanomaterials (ENMs) released into the environment, specifically, for aquatic systems. The
development of modeling frameworks that represent the unique complexities of ENM behavior
in the environment is in its infancy, and a critical mass of researchers actively engaged in model
development efforts has yet to be achieved. Further, it is widely recognized that there are many
obstacles to model development and, in general, to conducting environmental risk assessments of
ENMs that provide meaningful information for risk managers. Nevertheless, the U.S.
Environmental Protection Agency (EPA) will be required to manage potential risks across the
life cycle of ENMs, from production through the disposition of wastewaters and residuals
containing ENMs. Therefore, this state-of-the-science review included traditional modeling
frameworks as well as approaches that are considered relatively new to environmental modeling
science and risk management (e.g., adaptive management, multi-criteria decision analysis). In
essence, this review sought to answer five basic questions:
1. What models and approaches have been used successfully to simulate nanomaterial
behavior in environmental systems?
2. What models and approaches cannot be used to predict exposures to ENMs in
ecosystems?
3. What models and approaches can be used in the near term, and what types of predictions
can be supported by available models?
4. What techniques can be used to address uncertainties and support risk management
decisions in the near term given obvious gaps in information?
5. What does the state-of-the-science suggest with respect to long-term research goals that
can be undertaken to improve fate and transport modeling tools for ENMs?
To describe the state-of-the science landscape of fate, transport, and exposure models, we
conducted a focused review of the literature (published and grey literature), research centers,
conference proceedings, and related organizations (e.g., trade associations). We investigated a
wide range of information sources to identify models and methods that could be used in
evaluating exposures associated with the environmental release of ENMs. Initially, the
information search was limited to current models and approaches used to simulate fate and
transport of engineered ENMs in aquatic systems. However, because the literature on
environmental exposure modeling of ENMs was extremely limited, the search criteria were
expanded to include other types of particles (e.g., aerosols, polymers, and colloids) that exhibit
transport behaviors similar to ENMs. In addition, we expanded the scope of our review to
include modeling approaches that could be useful to risk managers in the near term (e.g.,
decision analysis methods). Therefore, the literature review identified (1) specific types of
modeling approaches (e.g., colloid models) considered highly relevant to exposure modeling of
ENMs, (2) traditional environmental exposure models applied to conventional chemicals, and (3)
other approaches that could potentially offer modeling solutions for the exposure assessment of
ENMs. In essence, our expanded search recognized that ENMs behave as both chemicals and
-------
Executive Summary
particles and, because few models have been developed to simulate materials that exhibit
properties of both chemicals and particles, we defined the state-of-the-science for exposure
modeling in broad terms.
Our review strategy involved the development of a thematic key word search based, in part, on
our preliminary review of major reports and literature reviews on the environmental modeling of
ENMs. We conducted a systematic search of literature databases using different key word
combinations and tailored those searches to focus on the most prolific researchers and specific
modeling topics (e.g., excluded health and safety literature). Although we recognized that ENM
research is a dynamic field, it quickly became apparent that the majority of cutting-edge research
on environmental exposures and modeling of ENMs is attributable to a relatively small subset of
scientists and modelers. Specifically, our search strategy included
• Elsevier's on-line technical documents service (ScienceDirect)
• Google Scholar search engine, particularly to identify significant reports
• ISFs on-line technical documents in the Web of Science
• Specific sources on nanotechnology, including nano-specific journals, research centers,
and nanotechnology trade association web sites
• Online libraries at the University of North Carolina at Chapel Hill
(http://www.lib.unc.edu/) and North Carolina State University (http://www.lib.ncsu.edu/)
• Personal communications with experts in nanotechnology research.
The results of this review are presented in the bibliography in Section 6, and a subset of these
results in Appendix A was organized by topical area (e.g., models currently used in fate and
transport simulations of ENMs) to support at-a-glance usage of the information provided in this
report.
The review focused on models and approaches that could be useful in assessing the multimedia,
multipathway impacts associated with nanomaterial releases into the environment; therefore, the
evaluation included single media models (e.g., porous media colloid models) that could be
integrated into a larger modeling framework, as well as multimedia modeling frameworks and
systems. In characterizing the state-of-the-science, we concentrated on fate and transport models
for ENMs, i.e., those models designed to predict the migration and transformation of chemicals
in the environment in support of exposure and risk assessment. Although we recognized that
bioaccumulation may be an important determinant of exposure for certain types of ENMs (e.g.,
nanoscale metals), the focus of this review was clearly on fate and transport modeling as the
means to predict exposure.
The information on fate and transport modeling approaches for ENMs is presented at several
levels of detail. For example, we developed summaries of models for specific environmental
media (i.e., surface water, subsurface, and biological media) as well as multimedia models. In
addition, we present (1) models developed and used specifically to evaluate ENMs, (2)
established regulatory models used for risk assessment purposes, (3) models that have potential
applicability to ENMs, and (4) alternative approaches to traditional environmental fate and
transport models. We identified a short list of applicable models and approaches, and developed
detailed reviews that could be useful to ENM researchers as a foundation in building a predictive
-------
Executive Summary
modeling capacity for estimating environmental exposures to ENMs. The detailed reviews also
include alternative approaches because risk management decisions are likely to be required
before the data and modeling science are sufficiently mature to produce reliable quantitative risk
estimates for exposures to ENMs.
Lastly, this report suggests several conclusions regarding the state-of-the-science for
environmental fate and transport models and alternative approaches that could be useful in
supporting an assessment of the potential environmental exposures to ENMs. These conclusions
are intended to inform the development of a long-term research strategy and offer insight into
future directions that may be productive. In summary, the conclusions presented in this report are
• Research priorities should continue to emphasize the development of empirical studies to
characterize fate and transport behavior under laboratory and field conditions
• Field testing of currently available fate and transport models could provide significant
insight into the limitations of these models when applied to ENMs
• Development of new models to replace or modify the partitioning approach used in most
multimedia fate and transport models should be a priority given the importance of these
concepts for conventional organics
• A parallel research track to adopt alternate approaches describe in this review (e.g.,
decision analysis) should be pursued to meet immediate needs and provide improved
decision-making support as data and models specific to ENMs continue to evolve
• The primary focus of this state-of-the-science review was on ENMs with an organic base
and, therefore, a similar review specific to metals should be conducted
• The development of a standard ENM data model that introduces consistency in
nomenclature and testing requirements and is driven by fate and transport modeling
needs would support an integrated approach to data/model development for ENMs.
-------
1.0 Introduction
Chapter 1.0
Introduction
Among the many emerging contaminants confronting the U.S. Environmental Protection Agency
(EPA) are the various nanoscale materials used in manufacturing commercial nanotechnology
products. Engineered nanomaterials (ENMs)1 are being incorporated into new commercial
products at an increasing rate (Wiesner et al., 2006). There are many unknowns regarding the
ecological or human health risk associated with exposure to nanoscale materials used in the
manufacturing process or residues released during normal use or after the useful lifetime of the
product. Environmental processes such as weathering of ENMs may even create a wide spectrum
of additional, ill-defined transformation products. Perhaps the biggest unknown is whether these
residues enter (or persist in) the ambient environment in forms or concentrations that pose health
or environmental concerns.
To determine the nature and extent of possible exposure to ENMs in the environment, methods
will be needed to predict their fate and transport in environmental media, understand the
biologically relevant forms of ENMs that persist in the environment and, ultimately, confirm
their occurrence in media (e.g., drinking water and foods) to which humans and animals may be
exposed. Because of the sheer numbers and variety of ENMs that have been and will be created,
and the profound influence of type, purity, purpose, and characteristics (e.g., coated versus
uncoated) on the environmental behavior of ENMs, it will be a virtual impossibility to conduct a
full battery of tests on each nanomaterial that adequately describes interactions among
environmental compartments and biological systems. At the nano scale, chemicals can exhibit
behaviors that are unique when compared to behaviors of materials in a larger scale,
conventional form2. For example, nano-scale materials may exhibit unique electromagnetic and
optical properties. In addition, ENMs tend to be more reactive than larger-sized materials due to
a much higher surface area to mass ratio, potentially resulting in faster kinetics (e.g., oxidation-
reduction reactions, dissolution) than might otherwise be expected.
As suggested in Figure 1-1, predictive modeling will be required to represent the relationships
among: (1) the manner in which ENMs are released into the environment, (2) the behavior, fate,
and transport of ENMs in various environmental compartments, (3) the exposure of human and
ecological receptors to ENMs, and (4) the adverse effects to ENMs as exposures occur over time
and space. However, these models can only be developed if the foundation of basic information
needs for ENMs is met, including chemical-physical properties, environmental behavior, and
relevant health and ecological endpoints. Based on the information provided by predictive
models, decisions can be supported to invest in additional data collection efforts, consider risk
management options, and so forth. As EPA carries out its mission to safeguard public health and
1 It is important to distinguish between (1) intentionally produced, engineered nanomaterials (ENMs), (2) naturally
occurring nanomaterials (e.g., soil colloids), and (3) incidental nanomaterials produced unintentionally through
some anthropogenic process (e.g., combustion by-products). Given the potential need for regulations covering their
production, use, and disposal, the focus of this report is on ENMs, though many of the fate and transport concepts
and approaches may apply to natural and incidental nanomaterials.
2 The literature often describes materials sized greater than the nanoscale as bulk materials. Consistent with EPA's
National Center for Environmental Assessment (NCEA), we have avoided this use of the term in order to prevent
confusion with large-volume, bulk production of materials, which may include the bulk production of ENMs.
-------
1.0 Introduction
the environment from chemical stressors, the information that predictive models provide will be
an essential component of the decision-making process regarding safe management, use, and
disposal of ENMs.
ENM
Release
Predictive Modeling
ENM
Fate &
Transport
ENM
Exposure
ENM
Effects
"Safe" for intended uses?
Life cycle issues?
Risk management options?
Additional research?
More simulations?
More data collection?
Information Needs on ENMs
Chemi cal-Physica 1
Properties
Environmental
Behavior
Health and
Ecological Endpoints
Figure 1-1. A predictive modeling strategy to evaluate ENM risks.
1.1 Background
In considering the information presented in this report, it is first necessary to understand the
context for this effort, namely, EPA's conceptual framework for exposure science (US EPA,
2009a). Exposure assessment is the process of measuring and modeling the magnitude,
frequency, and duration of contact between the potentially harmful agent and a target population,
including the size and characteristics of that population (Zartarian et al., 2005). As shown in
Figure 1-2, EPA has adapted a source-to-outcome framework to operationalize this definition
for the purposes of exposure assessment. Interestingly, this framework includes many of the
same features as the predictive modeling strategy illustrated in Figure 1-1. For example, EPA's
conceptual framework also begins with the release of a stressor (e.g., ENMs) into the
environment and ends with a dose-response characterization to determine the nature and
significance of the toxicological endpoints. Following release, ENMs may be transformed and
move through environmental media; thus, there is an implicit recognition of the importance of
multimedia (versus single medium) behavior. The intensity of the exposure is defined in terms of
the concentration in the contact medium, as well as the length of time that a receptor remains in
contact with the contaminated medium; the exposure becomes a dose only after the stressor has
crossed the body barrier. However, this figure also describes the interactions of environmental
factors that contribute to exposure and, importantly, illustrates the types of feedbacks that are
possible for ecological receptors and the environment. The impacts of exposure to a chemical
stressor in a defined ecosystem or habitat can include a cascade of effects that represent both
direct (e.g., significant reduction in a valued species due to direct toxic effects) and indirect (e.g.,
a shift in the vegetative community) effects. Note that this figure was developed, primarily, to
illustrate the importance of these interactions in spatially defined ecosystems in the sense that
-------
1.0 Introduction
ecologists often think about the spatial boundaries of an ecosystem (e.g., an old growth forest).
The energy and matter fluxes in an ecosystem are tightly coupled to plant and animal
communities and, as a result, these feedback loops are common. Although similar types of
feedback loops are possible in human exposure scenarios, human-ecosystem feedback is less
common for human receptors, because humans tend to be less tightly coupled with their
ecosystems.
Cascade of Effects
Environmental
Concentration/
Intensity
Atmopshere
Vegetation
Habitat Conditions
Hydrosphere
Lithosphere
Flow Dynamics
Dispersion
Kinetics
Thermodynamics
Spatial Variability
Distribution
Variability
Meteorology
Wate
Soil
Food
V
X.
^•s
Exposure
Pathway
Duration
Intensity
^^^^ Frequency
• Individual
• Community
• Population
~~» • Ecosystem
Effects
Figure 1-2. Source-to-outcome framework for ecological exposure research (US EPA, 2009a).
Based on this figure, it is clear that exposure science—as evidenced by EPA research
programs—seeks to represent the critical processes and flows of materials within an ecosystem
as the means of predicting potential effects associated with the introduction of a physical,
chemical, or biological stressor. This provides a critical context for this state-of-the-science
report, because it suggests that the focus of this review should, appropriately, be on modeling
approaches that are capable of representing the complex interactions among ENMs, abiotic, and
biotic compartments. Of critical importance is the recognition that nanomaterials are particles
and chemicals. This means that traditional partition coefficients (e.g., solid-water partition
coefficients) that drive multimedia modeling for most conventional chemicals cannot provide an
appropriate theoretical basis with which to predict the environmental exposures to ENMs. In
essence, nanomaterials may exist in aqueous solution both in truly dissolved form and as
suspended particles; traditional aqueous-solid phase partition coefficients only consider
dissolved mass versus mass associated with a stationary solid phase. Furthermore, many
processes are of critical importance to ENMs that may not be relevant to the environmental
behavior of conventional chemicals (e.g., processes determining the stability of aqueous
suspensions of nanoparticles).
Second, it is important to recognize how EPA's research strategy for nanotechnology (US EPA,
2009b) is related to the conceptual framework for exposure science. The purpose of EPA's
-------
1.0 Introduction
nanotechnology research program is to conduct focused research to inform risk management
decisions under various environmental statutes for which EPA is responsible and, more than
likely, under new environmental statutes that will be developed in the future as the
nanotechnology industry continues to mature. The EPA research strategy is structured around
several research themes and associated science questions. The first research theme is "Sources,
Fate, Transport and Exposure" with the associated science questions:
Key Science Question 1. What technologies exist, can be modified, or must be developed to
detect and quantify manufactured nanomaterials in environmental media and biological
samples?
Key Science Question 2. What are the major processes and/or properties that govern the
environmental fate, transport, and transformation of manufactured nanomaterials, and how
are these related to the physical and chemical properties of those materials?
Key Science Question 3. What are the exposures that will result from releases of
manufactured nanomaterials?
The state-of-the-science report on sampling and analysis (US EPA, 2008) focused on the first
key science question, providing a comprehensive review of the literature and research on what is
currently available so that the Agency could identify gaps in the technology and methods to
detect and quantify ENMs in environmental media. Similarly, this state-of-the-science report
focuses on the second and third key science questions through a comprehensive review of the
literature and research on currently available models and modeling approaches, so that the
Agency can determine what types of research are needed and what the current body of work will
and will not support with respect to the prediction of ecological (and human health) exposures to
ENMs released into the environment.
Lastly, the National Research Council's (NRC's) publication Science and Decisions: Advancing
Risk Assessment (NRC, 2009, commonly referred to as the "silver book") provides specific
guidance on transport, fate, and exposure assessment that reiterates major themes in both EPA's
conceptual framework for exposure science as well as in the Agency's research plan for
nanotechnology. Although the silver book primarily focused on human health risk assessment
(and, by extension, exposure assessment for human receptors), the insights and recommendations
expressed by the Committee clearly resonate with ecological exposure science. For example,
among the recommendations that the NRC Committee provided, the following pertain
specifically to exposure science:
• Exposure assessment should characterize sources, routes, pathways and the attendant
uncertainties linking source to dose
• Recognition of the multiple possible exposure pathways highlights the importance of a
multimedia, multipathway exposure framework
• A critical insight that should be recognized by EPA and other practitioners is that there is
no ideal transport, fate, or exposure model that can be used under all circumstances
-------
1.0 Introduction
• A lower resolution model (e.g., screening) that produces more timely outputs (at greater
uncertainty) may be required to support decisions in the near term when available data
and models are not suitable to support a more refined analysis
• Guidelines to help the risk analyst or risk manager understand how model uncertainty and
data limitations affect overall uncertainty in exposure assessment are needed
• The communication of uncertainty and variability should be part of key computational
steps of risk assessment—e.g., exposure assessment and dose-response assessment.
Several themes emerge from the nexus of these three reports that heavily influenced the
development of this state-of-the science report, from the review strategy through the
development of criteria with which to evaluate various models and exposure assessment
approaches.
• First, multimedia, multipathway exposure frameworks are preferable because they
support the development of exposure assessments that reflect the movement of ENMs
across and between abiotic and biotic compartments. As suggested in Figure 1-1, the
complexities and feedback loops inherent in a functioning ecosystem should be
represented to the greatest extent possible.
• Second, currently available models are often not ideal (or in some cases, even useful) for
ENMs; the chemical properties required by traditional fate and transport models to
simulate major environmental processes (e.g., equilibrium partition coefficients) are not
likely to be the same properties that drive those processes for ENMs.
• Third, the purpose of an exposure assessment for ENMs is, ultimately, to support the
characterization of potential risks to health and the environment. Consequently, modeling
approaches should be considered that will support decisions in the immediate future (i.e.,
models that could be used right now) as well as approaches that would require additional
data and model development. In either case, it will be critical to communicate the
uncertainty and variability associated with exposure assessments.
1.2 Purpose and Scope of this Report
The development of modeling frameworks that represent the unique complexities of
nanomaterial behavior in the environment is in its infancy, and a critical mass of researchers
actively engaged in model development efforts has yet to be achieved. Further, it is widely
recognized that there are many obstacles to model development and, in general, to conducting
environmental risk assessments of ENMs that provide meaningful information for risk managers
(e.g., Greiger et al., 2009, 2010; Wiesner et al., 2009). Nevertheless, EPA will be required to
manage potential risks associated with nanomaterials, from the production stage through ultimate
discharge and disposal of wastewaters and other residuals containing ENMs. Therefore, this
state-of-the-science review included traditional modeling frameworks as well as approaches that
are considered relatively new to environmental modeling science and risk management (e.g.,
adaptive management, multi-criteria decision analysis). Because of the general lack of ENM-
specific models, as well as numerous data deficiencies that have been reported in many of the
references included in Section 6, these alternative approaches may provide a bridge between the
immediate management needs for ENMs and the longer term exposure research interests and
-------
1.0 Introduction
goals of EPA. We recognize that these alternative approaches cut across risk management and
decision analysis sciences; however, given the significant data deficiencies and lack of
standardized analytical methods, we believe that it was well within the scope of this report to
explore near term solutions to characterizing potential exposures to ENMs. Put succinctly, the
purpose of this report on nanomaterial exposure models was to
• Provide a targeted review of the literature (published and grey literature), research
centers, researchers, conference proceedings, and entities (e.g., trade associations) that
report on current technologies
• Develop a synthesis of promising models and approaches that could be useful in building
capabilities in modeling environmental exposures to ENMs, especially for surface water,
groundwater, soil, and sediments.
It should be pointed out that this review primarily focused on the fate and transport of ENMs
with a base substance composed of an organic chemical (e.g., fullerenes, carbon nanotubes).
Given the breadth of this report, we did not specifically investigate the ENMs with a base
substance composed of metal (e.g., nano-scale silver). Although many of the concepts and
models discussed in this report are relevant to all types of ENMs, the fate and transport modeling
of nanoscale metals deserves separate treatment due to the inherent complexities in the
environmental behavior of metals (e.g., geochemical speciation, unique sorption-desorption
dynamics, nonlinear behavior in the subsurface).
Within the broader context of EPA's mission to protect human health and the environment, the
approach taken in this report implies that the domain of models/approaches that were of greatest
interest were those that fell into the space identified in Figure 1-3 by the dashed box.
Is mitigation necessary?
I
,^~^^^^
Compliance
Monitoring
Risk Management
(Stressor Exposure
Reduction)
— — — "~^ A
1
L
How best to mitigate?
r— ^
Was mitigation successful?
Figure 1-3. Framework for protecting human health and the environment (US EPA, 2009a).
However, some of the alternative approaches that are described in this report are appropriate for
the entire framework. Thus, the value of information that could be provided by using, for
-------
1.0 Introduction
example, a multi-criteria decision analytic (MCDA) framework or a Bayesian network approach,
extends beyond exposure assessment and crosses over into risk management of outcomes.
As in the silver book, we recognized that the universe of potential models was so extensive that
we needed to create a set of science questions to guide this review, and focus the state-of-the-
science review on exposure models that can be used or adapted for ENMs. Therefore, we
developed the following list of questions to guide the review and, ultimately, to organize the
conclusions of this review.
1. What models and approaches have been used successfully to simulate nanomaterial
behavior in environmental systems?
2. What models and approaches cannot be used to predict exposures to ENMs in
ecosystems?
3. What models and approaches can be used in the near term, and what types of
predictions can be supported by available models?
4. What techniques can be used to address uncertainties and support risk management
decisions in the near term given obvious gaps in information?
5. What does the state-of-the-science suggest with respect to long-term research goals
that can be undertaken to improve fate and transport modeling tools for ENMs?3
This list of questions represents a distillation of the essential goals and purpose of this report and
provided a compass that was enormously useful in considering which models to include in this
state-of-the-science review.
1.3 Overview of Review Methodology
1.3.1 Information Search and Review
A wide range of information sources was evaluated to identify models and approaches that could
be used in evaluating ecosystem exposures associated with the environmental release of
engineered ENMs. A literature review was performed to identify specific types of modeling
approaches (e.g., colloid models) considered highly relevant to exposure modeling of ENMs,
along with other potentially useful modeling frameworks and methods. Sources addressing
exposure and environmental fate and transport modeling of ENMs were the primary focus of the
search. The search engines ScienceDirect, the ISI Web of Science, and Google Scholar were
used extensively to perform the state-of-the-science literature review using different
combinations of search criteria described in Section 3.1. Titles pertaining to modeling
environmental transport of ENMs in soils and aquatic systems were identified and further
evaluated, and key sources of information such as key journals, reports, research centers, and
informational websites were identified and catalogued. A complete and categorized listing of
titles relevant to modeling the fate, transport, and exposure to ENMs released into the
environment is presented in Appendix A. Each of these references was thoroughly reviewed,
3 It should be noted that, although this state-of-the-science report is intended to inform the development of long-term
research goals, the purpose of this report was not to develop long-term research goals.
10
-------
1.0 Introduction
and a subset of these titles (highlighted in blue) was selected for the detailed model evaluations
provided in Appendix B.
1.3.2 Characterization of Models and Approaches
The model evaluation framework implemented for this assessment provides a systematic and
consistent approach for reviewing and summarizing information about models. The review
categories were developed to be consistent with the NRC paper, Models in Environmental
Regulatory Decision Making (NRC, 2007). In this report, the NRC assesses how models support
the EPA's environmental regulatory process. The development and application of regulatory
models is described along with recommended considerations for selecting and using models to
support EPA programs. The NRC document describes criteria for evaluating whether a model
and its results provide a sound basis for regulatory decision making. In reviewing the NRC
document, we compiled a series of key considerations for model evaluation and organized them
into general categories with specific questions. For example, under the category of Purpose and
Scope, the types of questions that are relevant include: What is the model purpose?; What
transport media are considered?; and What spatial and temporal scales does the model consider?
Because risk management decisions are likely to be required before the data and modeling
science are sufficiently matured to reliably produce quantitative risk estimates for ecological
exposures to ENMs, we also examined non-traditional modeling frameworks and methods that
are more closely related to decision analysis and risk management. These alternative methods do
not necessarily fit the traditional mold of environmental fate and transport models but may still
be useful in the risk assessment of nanotechnology in the near term.
Figure 1-4 presents a flow chart for characterization of the models identified in the search
strategy.
Categorize by transport media:
air, subsurface, surface water,
biological uptake, or multimedia
Model
used for
NMs?
NM fate and
transport model
\
^/
•\
Yes .-''''
Potential NM
transport model
"X. No
Non-applicable
transport model
i
\
Alternative
approach for NM
Figure 1-4. Model review process for exposure models and alternative approaches.
11
-------
1.0 Introduction
1.0 Introduction
2.0 Nanomaterials and
their environmental fate
and transport
3,0 Review methodology
for relevant models and
methods
4.0 Model Reviews
1.4 Roadmap to Report
Figure 1-5 presents the organization of this state-of-the-science
report. In Section 2, we begin by providing a brief overview of how
ENMs are defined and classified, describe properties of ENMs that
are key determinants of environmental behavior, and identify major
processes (e.g., aggregation of particles) that strongly influence the
mobility and fate of ENMs in the environment. The purpose of the
introductory material in Section 2 is to provide the reader with
general information on ENMs, a refresher of sorts needed to
understand the review of exposure models and approaches. Therefore,
the presentation of this material was intentionally brief and the reader
is encouraged to review any of a number of references included in the
bibliography that offer a more thorough treatment of these and other
issues, notably, EPA's nanotechnology white paper (US EPA, 2007),
EPA's state-of-the-science report on sampling and analysis (US EPA,
2008), Nanotechnology and the Environment (Wiesner and Bottero,
2007), Considerations for environmental fate and ecotoxicity testing
to support environmental risk assessment from engineered
nanoparticles (Tiede, et al., 2009), and Nanomaterials in the
environment: behavior, fate, bioavailability, and effects (Klaine et al.,
2008) as excellent sources of information. The remainder of Section
2 discusses the salient features of exposure modeling for ENMs
released into the environment and identifies key challenges associated
with predicting exposures to ENMs (e.g., limitations of current risk
assessment modeling frameworks).
Section 3 presents the search strategies and key information sources
that were included in the search including, for example, major
reports, journals, and research centers. Section 3 then discusses the
model/method evaluation criteria that we selected for this state-of-
the-science review. Note that the dynamic nature of ENM-related
5.0 Conclusions
§.0 Bibliography
Appendix A - Titles
pertaining to the use of
exposure models for NMs
Appendix B - Exposure
model/method summaries
for NMs in the environment
Figure 1-5. Report
roadmap.
research has resulted in a proliferation of publications over the past several years, with the
landscape of relevant literature changing on almost a monthly basis. Therefore, we limited our
search to materials that were either published or extracted from other sources (e.g., personal
communications) before April 30, 2010. Section 3 concludes with a summary of major
compendia and reviews of models/methods for ENM risk assessment that have been conducted
during the past three years. These summary reports were critical in shaping our search strategy
and, in addition, we relied on the combination of these reports to ensure that we were not
duplicating previous efforts by other researchers. Rather than attempt to identify every possible
report, article, or paper related to the fate and transport modeling of ENMs, we developed the
search/review strategy to describe the state-of-the-science landscape of models and approaches
that have been, or could be, useful in developing a research strategy for ENMs. Thus, seminal
journal articles, modeling reviews, and major reports were reviewed in detail to ensure that this
12
-------
1.0 Introduction
report captured the state-of-the-science with respect to modeling the fate and transport of ENMs
in soil and aquatic systems, and estimating exposures to ecological receptors.
Section 4 summarizes the review methodology and presents results of our review of models and
alternative methods used for, or of potential use in, the exposure modeling of ENMs. The
descriptions presented in this section distill the information on each model and methodology into
practical summaries intended to provide the reader with a thumbnail understanding of the
model/method, and its potential relevance to the exposure assessment of ENMs. From this
universe of models and methods, we applied the evaluation criteria presented in Section 4.1 to
select the most promising models and approaches and developed the detailed model/method
reviews found in Appendix B.
Section 5 summarizes our conclusions regarding the state-of-the-science review of models and
methods relevant to the exposure assessment of ENMs released into the environment focusing on
exposure pathways related to contaminant movement among the soil, subsurface, sediment,
surface water, and biological compartments (i.e., this review did not evaluate air models). As
discussed above in Section 1.2, our purpose was not to solve the problem with this review;
rather, the intent was to provide a comprehensive understanding of the types of modeling
approaches available and identify those models/methods that could potentially be applied (with
or without modification) to predict exposures to ENMs in the environment. Therefore, the
conclusions section offers insight into future directions that may be productive but does not
provide a definitive research agenda for ecological exposure assessment of ENMs.
Section 6 presents a full bibliography of relevant reports, publications, web sites, and
communications. This bibliography represents approximately two-thirds of the larger set of
references that we identified using our search strategy and initial review. To avoid diluting the
references with related but non-essential information, we attempted to focus the bibliography
somewhat narrowly on models and methods relevant to ecological exposure modeling. As
suggested above in the Section 3 summary, this bibliography is not comprehensive in the sense
that it includes every journal article with any relationship to fate and transport modeling of
ENMs. However, this bibliography does represent the state-of-the-science (as published) with
regard to fate and transport modeling of ENMs and will provide the reader with a thorough
understanding of the models, approaches, gaps, and current research in this field.
Appendix A reorganizes and refines the bibliography to focus on key references, in particular,
references for models/methods that are described in more detail in Appendix B. The references
are organized around five themes
• Exposure science and model evaluation
• Recent reports and compendia
• Models that simulate particle, aerosol, polymer, and colloid behavior
• Multimedia models currently used in ENM fate and transport simulations
• Alternative approaches and models.
13
-------
1.0 Introduction
Lastly, Appendix B presents detailed reviews for exposure models and other relevant alternative
approaches (e.g., multi-criteria decision analysis) in a standardized format which includes
• Model Summary
• Key References
• Contact/Availability Information
• Purpose and Scope
• Evaluation (e.g., mathematical representation; complexity; consideration of uncertainty;
applicability to ENMs).
14
-------
2.0 Nanomaterials and their Environmental Transport and Fate
Chapter 2.0
Nanomaterials and their Environmental Transport and Fate
This section of the state-of-the-science report provides background information on ENMs and
their behavior once released to environmental systems. This background material is needed to
understand the review of exposure models and approaches. The presentation of this material is
intentionally brief, and the reader is encouraged to review any of a number of references
included in the bibliography that offer a more thorough treatment of these issues (US EPA, 2007;
Wiesner and Bottero, 2007; Tiede et al., 2009; Klaine et al., 2008).
Section 2.1 defines nanomaterials and describes several nanomaterial classes. Next, Sections 2.2
and 2.3 discuss the chemical properties and processes that influence the behavior of ENMs in the
environment. Section 2.4 provides a summary of key transport behaviors of ENMs in aquatic
and terrestrial environmental systems. This information provides important context for the
evaluation of various environmental fate models discussed in Section 4 (e.g., What properties
and processes need to be considered by models of ENM fate and transport?). Section 2.5
summarizes several major challenges associated with modeling the environmental behavior of
ENMs, including transport complexity, variability in types of ENMs, limitations of traditional
modeling approaches, and the need for near-term decision making.
2.1 Engineered Nanomaterials and their Classification
ENMs are generally defined as having at least one dimension less than 100 nm and exhibiting
properties that are in some way unique relative to the same materials in larger, conventional
forms. Some researchers (e.g., Auffan et al., 2009) argue that only materials exhibiting novel,
scaled behaviors (e.g., surface area-normalized toxicity or adsorption) should be considered from
a regulatory perspective differently from conventionally sized materials.
Given the emerging and dynamic nature of nanoscience, consistency in the terminology used to
describe NMs in industry and in the academic literature has not yet been achieved, representing a
potentially significant source of ambiguity when considering the environmental behavior of
ENMs. Several national and international organizations (e.g., ANSI, ASTM) are developing
standards for consistent terminology that will, among other things, provide standard definitions
of different classes of ENMs. The reader is referred io ASTM StandardE2456-06, Standard
Terminology Relating to Nanotechnology, for additional information.
Although there are a number of excellent articles and reports describing types ENMs (e.g., US
EPA, 2007; Hansen et al., 2007; Ju-Nam and Lead, 2008; Wiesner and Bottero, 2007), we
adopted the chemical classification scheme following the work of Klaine et al. (2008), which is
based primarily on the chemical composition of the base substance. Importantly, given the broad
range of nanomaterial types and properties, this classification system is not internally consistent
in all cases. For example, many dendrimers are carbonaceous, and carbon nanotubes and metal
oxides can be semiconductors. In addition, important differences exist between ENMs within a
given category (e.g., fullerenes versus carbon nanotubes, both of which are carbonaceous
ENMs). Additional complications arise from modifications often made to ENMs such as surface
coatings that give them desirable properties, altering their basic chemistry and often significantly
15
-------
2.0 Nanomaterials and their Environmental Transport and Fate
changing their behaviors with respect to environmental fate and toxicity. Despite these
complexities, the classification scheme described below provides a useful overview of the
various types of nanomaterials and current applications. Note that the evaluation of fate and
transport models in this report focuses mainly on the behavior of organic-based, carbonaceous
nanomaterials. Although many of the concepts are relevant to other classes of nanomaterials,
important modeling approaches required to evaluate other types of nanomaterials (e.g.,
geochemical speciation models) were not reviewed in this report.
2.1.1 Carbonaceous ENMs
This class of ENMs is defined by the presence of carbon atoms in the nanomaterial structure.
Although carbonaceous ENMs share this fundamental similarity in their chemical composition,
there is significant diversity with respect to properties and environmental behavior. For example,
fullerenes are 60-carbon-atom hollow spheres also known as buckyballs; carbon nanotubes
(CNTs) include multi and single-walled carbon nanotubes. CNTs exhibit strong thermal and
electrical conductivity properties. CNTs often have very high aspect ratios, which are similar to
the high aspect ratios of asbestos, and thus warrant further toxicity evaluation. Carbon ENMs are
often hydrophobic in aqueous systems. Thus, unmodified carbon ENMs typically aggregate
together and/or attach to other surfaces in aqueous systems. Significant efforts have been
undertaken to reduce their hydrophobicity and to increase the stability of aqueous suspensions
(e.g., through functionalization of CNTs with polyethylene glycol or phospholipids). Common
applications of carbon ENMs include plastics, catalysts, battery and fuel cell electrodes, super-
capacitors, water purification, orthopedic implants, conductive coatings, adhesives and
composites, sensors, and as components in electronics, aircraft, aerospace and automotive
industries.
2.1.2 Metal ENMs
Metal containing ENMs include metal oxides as well as zero-valent metals. Titanium dioxide has
been used as a photocatalyst in solar cells, paints, and coatings. Both titanium and zinc oxide
have been used in sunscreens, cosmetics, and bottle coatings due to their UV blocking properties.
Cerium dioxide has been applied as a combustion catalyst in diesel fuels, improving emission
quality. Cerium dioxide has also been used in solar cells, gas sensors, oxygen pumps, and
glass/ceramics.
Nanoparticulate, zero-valent iron has been used extensively for remediation of waters, sediments,
and groundwater through chemical oxidation (e.g., of nitrates, chlorinated solvents) (Kanel et al.,
2008). It is noteworthy that a voluntary moratorium on the use of zero valent iron nanoparticles
for remediation has been in effect in the UK due to unknown potential environmental transport
behaviors and impacts.
The use of nanoparticulate silver for its antimicrobial properties represents by far the most
prevalent use of ENMs in consumer products (Klaine et al., 2008). Specific applications include
medical wound dressings, textiles (e.g., undergarments), air filters, toothpaste, baby products,
clothes washing machines and vacuum cleaners. Colloidal gold has also been used in medicine
(e.g., tumor therapy) and electronics.
16
-------
2.0 Nanomaterials and their Environmental Transport and Fate
2.1.3 Semiconductor Materials, Including Quantum Dots
Quantum dots are semiconductor nanocrystals. These ENMs have a reactive semiconductor core
surrounded by a shell made up of another material (e.g., silica). Quantum dots often include
surfactant coatings. The uses have primarily been medical (e.g., imaging) as well as in solar cells
and photovoltaics.
2.1.4 Nanopolymers/Dendrimers
This class of ENMs includes polymers with controllable size, topology, and molecular weight.
They have many applications in biology, material science, and catalysis. Specific applications
include chemical sensors, electrodes, transfecting agents, prion disease therapy, and drug
delivery.
2.2 Properties of ENMs that Influence Environmental Behavior
This section provides a summary of key properties that can influence the environmental behavior
of ENMs, particularly in aquatic and terrestrial systems. Given their strong surface reactivity and
unique behaviors at the nanoscale, different chemical properties can be relevant for ENMs versus
conventionally sized chemicals. In addition, nanomaterials exhibit properties of chemicals as
well as particles, thus complicating the understanding of their behavior in the environment and
requiring a more extensive set of properties when compared to many traditional contaminants.
For example, nanomaterials exist in aqueous solution both as truly dissolved molecules as well
as suspended particles, so that many traditional partitioning relationships (e.g., Kd [the
equilibrium ratio of the concentration in water to the concentration in the solid] relating solid-
aqueous phase concentrations) are insufficient to characterize nanomaterial behavior.
Table 2-1 provides a summary listing of the properties discussed in this section and groups the
properties into several categories of relevance to ENMs. A primary references for much of the
discussion of ENM properties in this section is Chappell (2008).
Table 2-1. Chemical Properties of Nanomaterials Relevant to Environmental Fate and Transport
Size
Characteristics
Surface Area and
Charge
Characteristics
Chemical
Composition and
Structure
Characteristics
Reactivity
Characteristics
Partitioning
Characteristics
Average particle size
Particle size range
Effective particle size
Aggregate size
Average
agglomeration
number
Surface area
Specific surface area
Surface charge
Hydrophobicity
Point of zero charge
Zeta potential
Elemental
composition
Cry stall nity
Surface coatings
Surface
functionalization
Aspect ratio
Bulk density
Speciation
Mean speciation
Degradability
Hydrolysis rate
Biodegradation rate,
Photolysis rate
Redox reaction rates
Solubility
Effective solubility
Volatility
Partition
coefficients
17
-------
2.0 Nanomaterials and their Environmental Transport and Fate
2.2.1 Size Characteristics
A sample of ENMs generally contains a distribution of variably sized nanoparticles. Thus,
parameters characterizing the size distribution of nanoparticles can be critical. For example, the
average particle size and particle size range are descriptors often provided by manufacturers.
Nanoparticles in aqueous suspensions may be surrounded by more or less permanent layers of
molecules due to inter-particle attractive forces. This behavior may result in a far larger effective
particle size relative to the size of the original nanoparticle. Particle aggregates may also be
characterized by the aggregate size and average agglomeration number (AAN). The AAN is the
average number of primary particles contained in an aggregate of particles. The AAN is typically
strongly dependent on environmental conditions (e.g., pH, ionic strength) as well as nanoparticle
surface coatings. Note that some research communities differentiate between aggregates and
agglomerates based on the degree of reversibility of the bonds holding particles together,
whereby aggregates are strongly bonded particles and agglomerates are more loosely bound
(British Standards Institute, 2007; Luoma, 2008).
2.2.2 Surface Area and Charge Characteristics
The surface area describes the total exposed surface for a particle. The strong surface activity of
ENMs accounts for much of the unique behavior associated with particles in the nano versus
larger size range. As the particle size decreases approximately below 20 nm, a very large fraction
of the total number of atoms exists at the particle surface, which can give rise to quantum
mechanical effects and associated unique properties and behaviors. The related specific surface
area is the ratio of the surface area to the mass for a particle. Some researchers suggest that, due
to their surface activity, surface area may be a better measure of potential health effects than
concentration, which is mass based (Bell, 2007). For example, some effects may be correlated
more closely with the surface area than the concentration (Oberdorster, 1996; Oberdorster et al.,
2007; Stoeger et al., 2006, 2007).
The surface charge is a measure of the density of charged entities on a particle surface and may
affect its propensity to interact with other surfaces and ions. The surface charge is typically pH
dependent (e.g., with oxide minerals). The surface charge is related to the hydrophobicity and
solubility, often determining the stability of nanoparticle suspensions, interactions with other
materials (e.g., attachment to natural colloids or immobile soil solids). Hydrophobicity describes
a particles interaction with water. More hydrophobic materials interact to a lesser degree with
water molecules, resulting in a reduced affinity for aqueous solutions and greater difficulty in
creating stable nanoparticle suspensions. Nanoparticles are often treated in order to decrease
their hydrophobicity (e.g., functional groups added to CNTs). The point of zero charge (PZC) is
the pH at which the number of positively charged sites on a surface that interact with protons is
equal to the negatively charged sites. This parameter is critical for determining the stability of
nanoparticle suspensions and thus their aqueous mobility. Below the PZC, water donates more
protons than hydroxide groups, and so the adsorbent surface is positively charged (and attracts
anions). Conversely, the surface is negatively charged above PZC (attracting cations/repelling
anions). The PZC determines how readily particles will adsorb to surfaces. At the PZC, a
colloidal system exhibits zeta potential of zero (i.e., the particles remain stationary when an
electric field is applied). The zeta potential refers to the electrical potential at a short distance
18
-------
2.0 Nanomaterials and their Environmental Transport and Fate
from a particle surface. This potential arises from the surface charge and is often used to
approximate the surface potential of the particle.
2.2.3 Chemical Composition and Structure Characteristics
Elemental composition describes what elements make up ENMs. Importantly, the composition
of ENMs is often modified with surface coatings or other treatments. A related issue is that the
purity of manufactured ENMs varies widely, so that other chemicals are often present within a
given batch of ENMs. The crystalinity refers to the stable three-dimensional arrangement of
atoms. The crystalinity can determine or strongly influence other parameters such as surface
area, charge, and aspect ratios. In order to maintain stable aqueous suspensions of nanoparticles
(without particle aggregation and settling), many ENMs are modified with surface coatings such
as polymers, polyelectrolytes, and surfactants. In addition to the desired changes in the resulting
ENM surface chemistry, such modifications may also alter a chemical's environmental transport
behavior and toxicity. When surface modifications involve changes in surface functional groups
it is referred to as surf ace functionalization. The aspect ratio describes the ratio between the
longest and shortest lengths of a particle. This parameter can be important for mobility and
uptake in organisms.
Speciation refers to chemical form. Different chemical species of the same element often have
very different properties influencing environmental fate and toxicity (e.g., solubility, volatility).
For some ENMs (e.g., CNTs), surface functionalization can lead to different speciation
properties for different parts of the surface, which may be represented by a "mean speciation" of
the material's surface (Chappell, 2008). Speciation can strongly influence other properties such
as particle size, solubility, and particle charge and, as with all chemicals, speciation can
significantly affect the mobility and toxicity of the nanomaterial.
2.2.4 Reactivity Characteristics
The degradability of ENMs refers to their persistence under various environmental conditions.
Generally, only organic ENMs biodegrade, and the rates of biodegradation vary widely. Mineral
ENMs typically do not biodegrade; however, they may be differentially susceptible to other
degradation/transformation processes such as oxidation.
2.2.5 Partitioning Characteristics
The solubility refers to whether the material dissolves in water or other substances (e.g., acids,
bases, solvents, biological media). In many studies, the dissolved fraction is defined
operationally using filters (e.g., dissolved materials are those that pass through a 200 or 450 nm
filter). However, the behavior of ENMs below this size fraction can be very different (e.g.,
aggregation, attachment) than truly dissolved molecules. Therefore, more complete descriptions
of nanoparticle suspensions (e.g., particle size distribution, surface activity) and a better
understanding of solubility are necessary to fully describe aqueous nanomaterial systems. The
effective solubility reflects the presence of suspended ENM aggregates in addition to truly
dissolved molecules. Thus, effective solubilities can vary significantly with environmental
19
-------
2.0 Nanomaterials and their Environmental Transport and Fate
conditions (e.g., pH, ionic strength) and can greatly exceed the solubility of conventionally sized
materials (Fortner et al., 2005).
2.3 Key Processes Influencing Environmental Behavior
This section describes key environmental processes that influence the mobility and fate of ENMs
in aqueous and terrestrial environments. In considering whether a model may be useful for
predicting exposures to ENMs, it is critical to understand what processes should be represented
within the modeling construct.
2.3.1 Aggregation and Deposition
Both aggregation and deposition of
nanoparticles can be considered as a two-
step process by which particles are first
transported to the proximity of a surface
(including for example another particle)
followed by an attachment step. In
aggregation, particles interact with other
moving surfaces while in deposition
particles interact with stationary surfaces
(including previously deposited particles).
As particles are transported and collide in
water or air, they may attach to each other
through forces of attraction. Particle-particle
interactions that control attachment may
result from relatively weak van der Waals
forces, stronger polar and electrostatic
interactions, or covalent bonding (Brar,
2009). The aggregation/disaggregation and
attachment/detachment behavior of
particulates in aqueous systems may be
predicted mechanistically using classic
DLVO theory and its extensions (see Box 2-
1). The term agglomeration is sometimes
used interchangeably with aggregation;
however, some sources use agglomeration to
describe particle groupings that are held
together by weaker forces than aggregation.
For example, irreversible groupings of
primary particles may be referred to as
"hard" aggregates in contrast to "soft"
(reversible) agglomerates (Brar, 2009).
Because aggregates tend to settle out of solution more readily than primary particles, the
aggregation process can have a fundamental influence on the mobility of ENMs. For example,
Box 2-1. DLVO Theory and Its Extensions
Classic DLVO (Derjaguin and Landau, Verwey and
Overbeek) theory describes the attractive and repulsive
forces between charged surfaces in a liquid medium.
The sum of these forces determines whether attraction
or repulsion forces will control particle aggregation and
attachment behavior. Fundamental DLVO theory
accounts for van der Waals attractive forces and
electrostatic repulsion (Grasso et al., 2002). Classic
DLVO theory has been found inadequate to fully
describe particulate behavior in some situations.
Additional forces that may be important in environmental
systems but are not considered by DLVO theory include
hydrogen bonding and the hydrophobic effect, hydration
pressure, non-charge transfer Lewis acid base
interactions, and steric interactions (Grasso et al.,
2002). An additional force active at the nanoscale with
potential importance for nanomaterials is Born repulsion
(Brant et al., 2007). Other explanations for non DLVO
behavior include heterogeneous surface charge
conditions (Bradford and Toride, 2007).
Although models of particulate behavior in aqueous
systems based on DLVO theory and its extensions may
be useful, Brant et al. (2007) emphasize several
important issues to consider for potential application of
the theory for ENMs. Nanoparticles at the lower size
range (approximately smaller than 20 nm) can
increasingly resemble molecular solutes, and
intermolecular forces can become more important.
Particle diffusion (Section 2.3.5) can occur at very fast
rates for nanoparticles, thereby increasing their potential
contact time with other particles and possible
attachment surfaces. This contact efficiency may
increase the rate of attachment (and decrease mobility)
beyond DLVO predictions. Development of extensions
to DLVO theory to account for ENM behaviors is an
active area of research (e.g., Loux and Savage, 2008).
20
-------
2.0 Nanomaterials and their Environmental Transport and Fate
aggregates that settle out will tend to preferentially reside in sediments at the bottom of surface
water bodies rather than being entrained within flowing water (if their settling rate is large
compared with the hydraulic residence time of the system). In addition, aggregates in porous
media may become trapped (filtered) and thus immobile. Aggregates also have a lower surface
area to mass ratio than primary particles, so they may be relatively less reactive (Johnson et al.,
1996).
The degree of aggregation and deposition is dependent on particle characteristics as well as
properties of the environmental system. Important particle characteristics include type, size, and
surface properties. Important environmental characteristics include pH, ionic strength, dissolved
carbon content, and the presence of dissolved organic matter and multivalent cations. In general,
greater aggregation (and settling) occurs under higher ionic strength and pH conditions. A study
by Former et al. (2005) demonstrated a dependence of C60 suspensions aggregate size
distribution on mixing rate, pH, and ionic strength, whereby lower pH and ionic strength led to
smaller particles (less aggregation). In many cases, the ionic strength of natural waters is
sufficiently large for particles to aggregate and settle to the bottom sediments.
2.3.2 Disaggregation and Detachment
Disaggregation and detachment are essentially the inverse processes relative to aggregation and
deposition. Specifically, disaggregation occurs when an aggregate suspended in solution
separates in to its component particles. Detachment occurs when a particle detaches from a
stationary surface it had previously attached to. In general, processes that favor particle stability
(low attachment probabilities) also tend to favor disaggregation and detachment.
2.3.3 Settling and Sedimentation
Settling is the process whereby particulates in aqueous solution sink due to gravity. Settling may
lead to sedimentation, the deposition of settled particulates onto sediments present at the bottom
of a water body. In very simple systems, the process of settling may be simulated using the
Stokes equation, which shows that the settlement velocity is exponentially dependent on the
particle diameter. Accordingly, larger particles are much more likely to settle out of suspension.
Particles with a settling velocity greater than a critical velocity equal to the mean depth divided
by the hydraulic residency time will preferentially be removed from suspension. Particles with a
lower settling velocity than the critical velocity will be removed proportionally to the ratio of the
settling velocity and the critical settling velocity (Boxall, 2007a). However, in some cases,
Stokes law may underpredict the settling velocity of an aggregate by an order of magnitude or
more (Wiesner, 1999).
2.3.4 Filtering and Enhanced Transport in Porous Media
Filtering refers to the process whereby aqueous phase particulates are deposited in porous media.
Mechanisms for particle filtering include attachment to the porous medium as well as straining,
which occurs when particulates are too large to pass through pore spaces. The formation of
particle aggregates can increase the likelihood of straining and the associated particulate filtering
in porous media.
21
-------
2.0 Nanomaterials and their Environmental Transport and Fate
An alternative potential mechanism in porous media is due to enhanced particle transport from
size exclusion. Because of their size and charge characteristics, particulates may be excluded
from some regions of the pore space, flowing primarily within the larger pore spaces. When this
occurs, the average velocity of parti culates can exceed the average water velocity, thus allowing
particulates to travel faster than inert solutes (van de Weerd et al., 1998).
2.3.5 Particle Diffusion
Diffusion refers to the process whereby particles spread from areas of higher concentration to
areas of lower concentration. Particle diffusion constants are predicted to be inversely
proportional to the particle size, so that the rate of diffusion increases as particle size decreases
towards the nanoscale (Brant et al., 2007).
2.3.6 Redox Reactions
Some ENMs are specifically designed to stimulate redox reactions. The redox chemistry often
provides the desirable mechanisms responsible for useful applications of ENMs. For example,
zero-valent iron has been used extensively in the remediation of waters contaminated with
chemicals such as chlorinated solvents. In its nanoscale form, zero-valent iron can undergo the
chemical reduction reactions that degrade contaminants faster than larger sized zero-valent iron.
Oxidation-reduction reactions can be critical for many ENMs.
2.3.7 Biodegradation
Biodegradation refers to the biologically mediated transformation of chemicals (parent
compounds) into other forms (daughter products). Ultimately, the products of biodegradation
may be carbon dioxide and water; however, other chemicals may also be formed (in some cases
more toxic than the parent compounds). Many ENMs are composed of inherently non-
biodegradable inorganic chemicals (e.g., metals) and not expected to biodegrade. However, some
carbon ENMs have been shown to be metabolized biologically (Filley et al., 2005). In addition,
some polymer based ENMs (and surface coatings) are known to be biodegradable. For some
ENMs (e.g., polymers evaluated for use in drug transport), biodegradability is integral to the
materials design and function (US EPA, 2007).
2.3.8 Hydrolysis
Hydrolysis describes the reaction of a chemical with water, whereby a chemical bond is broken
between a carbon atom and some functional group and a new carbon-oxygen bond is formed
with oxygen derived from the water molecule. Hydrolysis generally breaks chemicals down into
simpler molecules, which are usually (but not always) less toxic. Given their much greater
surface area to mass ratios, ENMs are generally anticipated to undergo transformations such as
hydrolysis more readily and at a faster rate than larger-sized materials. However, few ENMs
have been thoroughly characterized with respect to hydrolysis, particularly under variable and
complex environmental conditions.
22
-------
2.0 Nanomaterials and their Environmental Transport and Fate
2.3.9 Photolysis
Photolysis refers to a reaction whereby a chemical compound is broken down by photons.
Photolysis generally breaks chemicals down into simpler molecules, which may be (but are not
always) less toxic. Some nanomaterials, such as TiO2 and C60 are photocatalytic in the sense
that when in water and exposed to UV light, they convert energy in photons to chemical
compounds that may directly or indirectly react with other compounds in water.
2.3.10 Phase Partitioning
Phase partitioning refers the transfer of mass between aqueous, solid, and air phases, including
the processes of dissolution, volatilization, and solid/aqueous partitioning (sorption).
Traditionally, equilibrium mass partitioning between solid and aqueous media has been
described using partition coefficients (Kj, or the equilibrium ratio of the concentration in water to
the concentration in the solid). Equilibrium mass partitioning between water and air has
traditionally been characterized using Henry's constants (Kh, or the equilibrium ratio of the
concentration in air to the concentration in water). Other partitioning coefficients may
characterize the mass uptake into other phases (e.g., plants, fatty tissue). Solubility describes the
equilibrium partitioning between a pure solid or liquid phase and water. For non-equilibrium
partitioning, mass transfer may be described kinetically using parameters that characterize the
rate of mass transfer from one phase to another (e.g., first-order mass transfer rate coefficients).
The field of environmental science has studied the partitioning of various chemicals under a wide
range of conditions, yielding extensive summaries of partitioning data used to predict where
chemicals will tend to accumulate in the environment (e.g., air, water, soil, fatty tissues). In
addition, various approaches for estimating parameters for chemicals based on their chemical
structure (e.g., quantitative structure-activity relationships or QSARs) have been developed. This
concept of predicting chemical fate based on partitioning behavior is central to most
fate/transport and exposure models, and the theoretical underpinnings of phase partitioning (for
conventionally sized chemicals) have been widely accepted in risk assessment modeling.
However, the utility of existing partition coefficient values and estimation approaches for ENMs
is limited given the unique properties and behaviors of chemicals at the nano-scale as discussed
below.
ENMs will dissolve to differing degrees into aqueous solution, yielding truly dissolved
molecules in addition to nanoscale particulates and larger-sized aggregates suspended in
solution. Therefore, nanomaterials in solution exhibit properties and behaviors of chemicals as
well as particles. Understanding the dissolution of ENMs is significantly complicated by
commonly used operational definitions of solubility (e.g., materials passing through 200 nm
filters being considered dissolved; see Section 2.2.5). The presence of nanoparticles and
associated stable aggregates can greatly increase total concentrations in solution relative to the
molecular solubility of a compound. For example, Former et al. (2005) show that C60 fullerenes
may form negatively charged, water-stable colloidal aggregates, increasing the effective aqueous
concentration by approximately 11 orders of magnitude greater than the estimated molecular
solubility of C60. Given the potentially strong influence of suspended nanoparticulates on the
effective aqueous concentrations of ENMs, traditional solubility and partitioning relationships
23
-------
2.0 Nanomaterials and their Environmental Transport and Fate
will have limited applicability for predicting the dissolution and the relative mass of ENMs in
solid and aqueous phases.
As with aqueous/solid phase mass transfer, the process of chemical volatilization (i.e.
aqueous/air phase mass transfer) has traditionally been described using partition coefficients,
specifically Henry's constant, which is the ratio of the vapor concentration of a chemical to the
aqueous concentration of that chemical. The volatilization of ENMs is also complicated by the
potentially significant enhanced solubility effects of nanoparticle suspensions as described
above. Therefore, the utility of traditional Henry's constant values may be limited for predicting
the partitioning of ENMs to air.
Given their greater surface area per mass ratio and reactivity, it would stand to reason that ENMs
could partition at higher rates and thus reach equilibrium more quickly than conventional
materials. This behavior may be particularly relevant for mass transfer processes that are
kinetically limited (that reach equilibrium relatively slowly). However, this difference in ENM
versus conventional chemical behavior has not been extensively evaluated, particularly under
variable and complex environmental conditions.
Phase partitioning is also related to biological media and the bioaccumulation of chemicals.
Traditional environmental fate evaluations have relied on bioconcentration factors (BCF), which
is generally defined as the equilibrium ratio of chemical concentrations in an organism relative to
the chemical concentrations in the environmental medium of interest. BCFs characterize the
uptake of chemicals to organisms in aqueous environments as well as the uptake of chemicals by
plants from soil. Due to the potentially strong influence of ENMs on effective solubility
described above, the utility of existing BCFs for the evaluation of the potential to accumulate
ENMs in organisms is limited. Furthermore, many researchers have proposed mechanisms of
biological uptake that may be possible only at the nanoscale (e.g., direct penetration of
nanoparticles into organisms).
In summary, because that ENMs exhibit properties of chemicals and particles, partitioning data
heavily utilized in traditional environmental fate modeling (e.g., values measured or estimated
for Kd and Kh) will not be valid for ENMs. Nevertheless, the ability to predict in which medium
substances prefer to exist will still be a critical concept for environmental modeling of ENMs.
However, methods of measuring and predicting partitioning for ENMs will need to be modified
relative to approaches currently used for chemicals in conventional forms. Specifically, there will
need to be (1) new model constructs that accurately capture the complexities of ENM
partitioning, and (2) new experimental data will be required to parameterize these models
accurately under different environmental conditions.
2.4 Considerations for the Fate and Transport of ENMs in Environmental Media
This section summarizes properties and processes that need to be considered by models of ENM
fate and transport. This information provides important context for the evaluation of various
environmental fate models discussed in Section 4. The discussion first considers behavior in
aquatic systems (Section 2.4.1) and subsequently considers behavior in terrestrial systems
(Section 2.4.2). For the purposes of this discussion, aquatic systems are surface waters, whereas
24
-------
2.0 Nanomaterials and their Environmental Transport and Fate
terrestrial systems include soils, groundwater, and sediments. Figure 2-1 provides a summary of
fate and transport processes and their relevance to behaviors of ENMs in aqueous versus
terrestrial systems. This figure ties together the discussion of processes in Section 2.3 with the
consideration of ENM behaviors in environmental systems, focusing on processes specific to
particles like many ENMs versus traditional processes that are often considered in models
currently applied for environmental risk assessment.
Aquatic
Sedimentation
Phase partitioning:
Dissolution
Volatilization
Adsorption
Biological uptake
Chemical Tranformations:
Photolysis
Redox reactions
Hydrolysis
Biodegradation
Terrestrial '*
Straining
I
I
I
gregation and
Attachment
"Traditional"
Disaggregation
and
Detachment
Figure 2-1. Illustration of chemical processes relevant to transport in aquatic (surface water) and
terrestrial (groundwater, soils, and sediments) environmental systems.
2.4.1 Fate and Transport in Aquatic Systems
Traditional evaluation of chemical fate in aquatic environmental systems has focused on the
following processes and parameters (US EPA, 1996):
• Dissolution characterized by the solubility
• Volatilization characterized by the Henry's constant
• Adsorption to sediments and/or suspended particulates (i.e., solid-aqueous phase
partitioning)
• Biological uptake characterized by bioconcentration/bioaccumulation factors
• Photolysis characterized by photolysis rates
• Hydrolysis characterized by hydrolysis rates
• Biodegradation characterized by biodegradation rates.
25
-------
2.0 Nanomaterials and their Environmental Transport and Fate
Many traditional environmental fate assessments consider some or all of the processes listed
above. Associated evaluation techniques (e.g., decision frameworks and models) have been
developed to predict potential environmental concentrations and to support regulations for a
range of environmental contaminants such as pesticides, industrial solvents, and agricultural
chemicals. Somewhat less common are environmental fate models that consider processes of
aggregation, attachment, settling, and sedimentation. These processes, which involve the
behavior of aqueous-phase particulates, are critical in understanding and predicting the
environmental fate of ENMs in aqueous systems.
A key consideration influencing the mobility of ENMs in surface waters is whether a distribution
of suspended nanoparticles (a nanoparticle suspension) will be stable. In other words, will the
nanoparticles tend to aggregate, forming larger assemblies that may settle out, will the
nanoparticles tend to hetero-aggregate with other solids in aqueous solution (e.g., soil particles or
naturally occurring colloids), or will the nanoparticles remain in a stable suspension? The
aggregation/disaggregation process is generally dependent on the solution pH and ionic strength,
the presence of other solutes, the properties of naturally occurring particles, and on the properties
of the specific ENM (e.g., surface charge, size distribution). Typically, higher pH and ionic-
strength solutions give rise to larger particles and subsequent settling. The ionic strength of
natural waters (particularly seawater) is often sufficient for particles to aggregate and settle to
bottom sediments.
Researchers into ENM behavior in aqueous systems emphasize that surface water chemistry can
have a profound influence on the extent of particle aggregation, thus controlling whether
nanoparticle suspensions are stable and flow with surface water or are unstable, tending to
aggregate and settle (Brar, 2009; Boxall, 2007a). The more prevalent compounds in aquatic
ecosystems—proteins, humic acids, organic matter, and natural colloids—may have a profound
and complex influence on nanoparticle suspensions and behavior in surface water. Additional
research is required to characterize behaviors of ENMs under highly variable natural conditions.
In addition to nanoparticle interactions such as aggregation, ENMs have the capacity to sorb to
sediments, other suspended particulates, and/or other solid surfaces (Oberdorster, 2005a). ENMs
sorbed to other particulates that are stably suspended in solution (e.g., colloids) would have
enhanced mobility relative to nanoparticles that aggregate and settle out of solution.
An additional complexity associated with ENMs in the environment is the impact of
transformation processes on ENM properties and behavior. So-called weathering processes
occurring in natural systems may include aggregation, hydrolysis, loss or acquisition of surface
coatings, photolysis, etc. Each of these processes may have a marked influence on the ENM
surface chemistry, reactivity, bioavailability, toxicity, etc. under natural conditions (Brar, 2009).
Given their unique properties and behaviors at the nano scale, ENM may be expected to exhibit
different transformation behaviors (e.g., rates, dependencies) than conventionally sized materials.
The study of ENM transformations is an active field of research (e.g., Hotze et al., 2010).
26
-------
2.0 Nanomaterials and their Environmental Transport and Fate
2.4.2 Fate and Transport in Terrestrial Systems
Traditional evaluation of chemical fate in terrestrial (soil and groundwater) systems has focused
on the following processes and parameters (US EPA, 1996)
• Dissolution characterized by the solubility
• Volatilization characterized by the Henry's constant
• Adsorption to organic matter (e.g., organic carbon partition coefficients, Koc)
• Adsorption to inorganic matter
• Biological uptake by plants characterized by BCFs
• Photolysis (at the soil surface) characterized by photolysis rate
• Redox reactions and associated rates
• Hydrolysis characterized by hydrolysis rate
• Biodegradation characterized by biodegradation rates.
Many traditional environmental fate assessments consider some or all of the processes listed
above. Associated evaluation techniques have been developed to predict potential environmental
concentrations and to support regulations for a range of environmental contaminants. Somewhat
less common are environmental fate models that consider processes of aggregation, attachment,
straining, and enhanced porous media transport. These processes, which involve the behavior of
particulates, are critical in understanding and predicting the environmental fate of ENMs in
terrestrial systems.
Until the last two decades, subsurface transport was thought to be mediated only by mobile
liquid and gaseous phases. However, we now recognize that a separate, solid phase consisting of
particles may be present and mobile under some conditions and may facilitate or retard
contaminant transport in porous media (Sen and Khilar, 2006). In some cases, contaminants have
migrated much farther than would be predicted based on their solubility and sorption
characteristics, a behavior that can be explained by colloid-facilitated transport. Such facilitated
transport may occur when colloids (and nanoparticles) are excluded from some regions of the
pore space due to their size and charge characteristics (e.g., occlusion from zones with small
pores and electrostatic repulsion from solid surfaces). Under these conditions, the average
velocity of the parti culates may exceed the average water velocity (i.e., colloids can travel faster
than inert solutes) (van de Weerd et al., 1998). In contrast to the enhanced mobility associated
with colloid facilitated transport, colloid transport has also been associated under some
conditions with reduced contaminant transport through porous media (Sen, Mahajan, et al., 2002;
Sen, Nalwaya, et al., 2002; Bekhit and Hassan, 2005). Retardation of transport may occur when
colloidal particles are filtered (trapped) within a porous medium due to their size and/or due to
attachment to immobile solid surfaces. Filtering of parti culates will be enhanced under
conditions favoring aggregation and attachment. A related transport retardation mechanism may
occur when particulates plug the pore space and thereby reduce the hydraulic conductivity and
groundwater flow velocities.
Colloidal particles may be contaminants (e.g., radioactive metal particulates, nanoparticles with
potentially toxic properties) or the particles may mediate the transport of other contaminants in
27
-------
2.0 Nanomaterials and their Environmental Transport and Fate
the subsurface (e.g., chemicals or nanoparticles sorbed to and thus potentially migrating with the
colloidal particles). With their high surface area to volume ratios, the surface activity of colloids
may be very high, potentially leading to faster kinetics and irreversible association (e.g., fast,
irreversible sorption).
Colloids generally may undergo the following relevant behaviors in the subsurface:
• Attachment (also called colloid deposition) generally refers to the binding of colloidal
particles onto the stationary soil solid surface
• Detachment (also called colloid release) generally refers to the release of colloids from
the soil surface into the flowing liquid phase
• Aggregation refers to the process whereby colloids join together forming larger sized
aggregates; in many cases these particles may be too large to fit through aquifer pore
spaces, and are filtered by the porous medium and no longer migrate with the flowing
groundwater (they may also clog a porous medium, reducing its hydraulic conductivity)
• Disaggregation refers to the breaking apart of colloid aggregates (the opposite of
aggregation).
There is typically an abundance of natural colloidal particles attached to soil grain surfaces in the
environment (e.g., clay colloids, humic and fulvic acids). Under certain conditions, these colloids
can be released from the soil matrix and transported with the mobile liquid phase. Mechanisms
leading to colloid mobility could include an increase in the groundwater flow rate (e.g., around a
pumping well), which may result in sufficient advective shear forces for colloid detachment. An
additional important detachment mechanism is tied to the aqueous geochemistry. For example, a
decrease in the salt concentration below the critical salt concentration (CSC) may release
colloids from the solid by decreasing the attractive forces and increasing the repulsive forces
between the colloids and the soil (Blume et al., 2005). Note that this behavior is analogous to the
criticalflocculation concentration (CFC) in surface waters, which refers to the concentration
above which flocculation is favored (e.g., flocculation in wastewater treatment settlement
reactors). These properties are also related to the zeta potential and the point of zero charge as
described in Section 2.2.2. Groundwater generally has higher ionic strength than rainfall but
lower ionic strength than marine and many freshwater systems. Typically, higher ionic strengths
increase the likelihood of particle aggregation and attachment due to the reduced inter-particle
electrostatic repulsion.
The primary relevant processes determining colloid transport behavior include advection,
dispersion, particle-particle physico-chemical interactions, and particle-soil physico-chemical
interactions. Other potentially relevant phenomena include acid-base relationships, steric
repulsions (e.g., with long-chain polymers), and magnetic interactions (e.g., with iron) (Tosco
and Sethi, 2009).
As described in Section 2.3, nanoparticle surface modifications and coatings can greatly
influence their transport behavior. One application of such surface coatings is for zero-valent
iron nanoparticles utilized in the remediation of groundwater contamination (e.g., chlorinated
solvents). If the iron nanoparticles are made more hydrophilic through surface modifications,
nanoparticle aggregation is greatly reduced along with potential sorption and filtration by the
28
-------
2.0 Nanomaterials and their Environmental Transport and Fate
aquifer material (US EPA, 2008). This modification to iron nanoparticles can greatly increase the
mobility of the nanoparticles in groundwater, thus increasing the effectiveness of iron
nanoparticles for remediation. Of course, similar enhanced transport would not be desirable for
nanoparticles that may have environmental risk implications.
Surfactants are commonly utilized in the engineering of nanoparticles to help stabilize
nanoparticle suspensions. These surfactants can have a clear impact on the partitioning between
aqueous, solid, and particulate phases and thus on the fate of nanoparticles in subsurface
systems. In addition to engineered surfactants, many systems have natural surfactants (e.g.,
natural organic carbon), and some researchers have found evidence that such natural surfactants
can help stabilize aqueous nanomaterial suspensions, thus enhancing subsurface transport.
Several studies have evaluated the mobility of nanosized materials in a porous medium under
different conditions (Zhang, 2003; Lecoanet and Wiesner, 2004; Lecoanet et al., 2004).
2.4.3 Uptake and Accumulation of Nanomaterials in Biological Systems
Published, quantitative research on the bioavailability, uptake, and bioaccumulation of ENMs in
plants and animals is scarce (Klaine et al., 2008). However, the body of research on effects of
ENMs is far more developed and provides some clues as to the potential for bioaccumulation of
ENMs released into the environment. In fact, the vast majority of research involving ENMs and
biological systems is focused on interactions (e.g., positive effects on plant growth) and potential
toxicity, especially on standard animal models such as Daphnia magna as well as on test
organisms that will likely be exposed to ENMs through close contact (e.g., soil bacteria). Based
on first principles, it is evident that organisms exposed to environmentally relevant
concentrations of ENMs will be capable of accumulating these materials. Nano-sized particles—
due to their size—can diffuse through cell membranes, can by engulfed by cells, or can adhere to
cells. Because some ENMs are designed specifically for drug delivery purposes, it is reasonable
to assume that they will interact with proteins and other cellular components and, especially, be
taken up by the gut (Klaine et al., 2008). Despite the fact that little published research exists to
determine how efficiently different types of organisms may accumulate and, possibly, translocate
ENMs within the body, it is clear that potential interactions at the cellular level may allow for the
relatively efficient accumulation of ENMs and, possibly, some level of magnification in the food
chain.
2.5 Challenges to Modeling Nanomaterials
Predicting the behavior of ENMs in the environment requires an understanding of: (1) the
potential sources of ENMs; (2) the distribution of ENMs once released into the environment; and
(3) the transformations and persistence of ENMs in the environment (Lowry and Gasman, 2009).
Significant modeling challenges make current estimations of ENM fate highly uncertain. This
section of the state-of-the-science report discusses the challenges listed below:
• Complexity of ENM transport characteristics and associated data gaps
• Variability in nanomaterial types and properties
• Limitations of current modeling approaches
29
-------
2.0 Nanomaterials and their Environmental Transport and Fate
• Need for near term risk management decisions.
2.5.1 Complexity of Transport Characteristics and Behaviors and Associated Data Gaps
The locations, concentrations, and properties of ENMs released to the environment will affect
their distribution, concentration, and ultimately their effects on the health of receptors, including
humans and ecosystems (Lowry and Gasman, 2009). Two aspects of this complexity bear noting.
First, as described in detail above, properties and processes that define nanomaterial transport
(e.g., stability of dispersions; simultaneous chemical and particle characteristics) can be quite
different from properties and processes considered in most traditional risk assessment (e.g.,
Henry's constants, K^/ for aqueous-solid phase partitioning). Second, at each stage of modeling
ENM transport, there are large uncertainties that cannot be quantified given currently available
data and models. These uncertainties do not purely surround the values of traditional parameters,
but they also include potentially unique causal mechanisms of transport that, as of this report,
have not been fully addressed (SCENIHR, 2007). Thus, identifying the chemical and physical
properties required to predict the transport of ENMs in a natural system is crucial to developing
predictive exposure models. ENMs do not conform to the behavior of conventional chemicals
(Lowry and Gasman, 2009) and, therefore, studies must be conducted to determine the driving
properties^or specific ENMs (e.g., key chemical properties) in specific environmental settings
(e.g., key water quality characteristics). Before adequate models for ENMs can be developed,
the critical properties and mechanisms must be understood and characterized.
It should be noted, however, that there is more than just a lack of knowledge surrounding
properties specific to ENM transport. Grieger et al. (2009) address the issue of uncertainty in
ENMs by evaluating 31 reports and papers pertaining to ENMs and environmental health and
safety. The authors present a table describing each uncertainty identified. Each uncertainty is
categorized into locations and sub-locations that define the lack of knowledge area within the
different environmental, human health and safety aspects of ENM exposure. A few categories of
uncertainties that are of importance to environmental fate and transport taken from this chart
include:
• Initial concentration levels
• Release points during the ENM lifecycle (production, use, and waste)
• The form of the released ENM (agglomerates, composites, mixtures)
• Environmental processes relevant to ENMs
• Degradation
• Bioaccumulation
• Cellular uptake.
Thus, to predict or assess the risks that ENMs may pose in the environment, we must understand
the sources, characteristics, transformations, and the effect on the surface properties of those
materials.
30
-------
2.0 Nanomaterials and their Environmental Transport and Fate
2.5.2 Variability in Nanomaterial Types and Properties
Nanomaterials exist as a widely divergent array of chemicals with very different behaviors and
properties. In fact, ENMs should not be considered as a single class of substances due to their
extreme diversity (e.g., Hansen, 2009; Wiesner and Bottero, 2007). Due to the wide variety in
ENM properties, size, morphology, chemical composition, and the even greater variability in the
types of surface coatings, an understanding of ENM processes from first principles is needed
before we can generalize these findings across classes or types of ENMs (Lowry and Gasman,
2009). The sheer range of types and properties poses a particular challenge in terms of
developing and understanding chemical behaviors of ENMs relevant to environmental transport.
2.5.3 Limitations of Traditional Risk Assessment Models
In general, conventional models used for chemical environmental fate and exposure assessment
are not directly applicable in their current form for manufactured ENMs (US EPA, 2007). For
example, the Estimation Programs Interface Suite (EPI Suite) has little applicability to ENMs,
because it is based on equilibrium partition coefficients and does not consider the behavior of
particulates (US EPA, 2009b). Even though some models (e.g., EPA's MINTEQA2) and
approaches (e.g., DLVO theory) may be useful in modeling ENMs, these models/methods will
need to be modified and validated to ensure that they adequately represent the chemical
properties and/or transformation processes relevant to ENMs (US EPA, 2009b). Therefore, while
most established models may still provide insight, their potential application is relatively narrow
for ENM risk assessment purposes (Wiesner et al. 2009).
The chemical properties required by traditional risk assessment models are unlikely to be the
same properties required to model the environmental transport of ENMs. Chemical and physical
properties of ENMs (Section 2.2) are strongly related to the processes that control movement
(Section 2.3). Therefore, typical chemical properties for predicting chemical fate and transport
such as water solubility, octanol-water partition coefficient, and vapor pressure are not as
important for ENMs as particle size, surface charge and surface potential (US EPA, 2009b).
Metcalfe et al. (2009) produced Table 2-2 comparing some the characteristics needed for
environmental fate and transport modeling of ENMs versus conventional compounds. This
evaluation further supports the consensus that established risk assessment models (and, indeed,
current thinking) must be adapted before they can be applied reliably to predict the
environmental behavior of ENMs.
Table 2-2. Comparing the Properties Needed to Model ENMs Versus Other Contaminants
Characteristic Nanoparticles Other contaminants
Distribution in water Dispersivity Solubility
Distribution in porous media Filtration Adsorption/desorption
Biologically availability Sorption? Lipophilicity
Cellular uptake Vesicular transport? Passive or facilitated diffusion
31
-------
2.0 Nanomaterials and their Environmental Transport and Fate
Table 2-2. Continued
Characteristic Nanoparticles Other contaminants
Toxic mechanisms Steric hindrance, photo-chemical Interactions with cellular
effects, oxidative damage, macromolecules and receptors,
inflammation narcosis
Target trophic systems Bottom of the food chain? Top of the food chain
2.5.4 Need for Near- Term Risk Management Decisions
A critical issue with the development of environmental fate and transport models for ENMs and
their subsequent validation is the relatively long time required to gain knowledge upon which to
make decisions versus the very rapid pace of nanotechnology development (Owen et al., 2009).
Therefore, the development of regulations based on quantitative risk assessments (using tools
including fate and transport models) will be an inherently slow governance process. Given that
reliable, mechanistic ENM risk assessment may not be available for years or even decades
(Grieger et al. 2009), risk managers are in need of tools over the short term to aid in the decision
making process. Thus, in the absence of reliable quantitative and qualitative data or evaluation
methods, novel approaches must be attempted to gain insight for near term decision making.
Given this critical issue and need, Section 4 discusses several alternative approaches that derive
more from risk management and decision analysis sciences than from traditional fate and
transport analysis.
32
-------
3.0 Review Methodology for Relevant Models and Methods
Chapter 3.0
Review Methodology for Relevant Models and Methods
This section describes the methodology used to search for relevant information on the fate and
transport modeling of ENMs released in the environment. Figure 3-1 depicts the review
methodology used to identify information sources that describe the state-of-the-science for
exposure modeling of ENMs. Section 3.1 presents the search strategy developed to gather
information on the type of fate and transport models needed to estimate exposure concentrations
for ENMs in aquatic systems. This section includes the criteria, tools, and methodology used to
frame the retrieval of appropriate references and other information. Section 3.2 describes the
sources of information that we identified as particularly useful in describing the state-of-the-
science for ENM exposure modeling, including key journals, reports, research centers, and
informational websites. Section 3.3 summarizes the recent reports and compendia that are
relevant to the area of nanotechnology research but only peripherally related to this report.
3.0 Review
Methodology
3.1 Developing the
search strategy
i
3.2.1
Journals
3.2 Identifying key
information sources
J
3.2.2
Key reports
3.3 Summary of
recent reports and
Compendia
|
3.2.3
Research Ctrs
3.2.4
Websites
Figure 3-1. Summary of review methodology.
3.1 Developing the Search Strategy
Identifying references, published studies, and research describing the state-of-the-science in
modeling the environmental fate and transport of ENMs was the primary objective of the search.
We investigated a wide range of information sources, and developed an iterative search strategy
to identify specific types of modeling approaches (e.g., colloid models) considered highly
relevant to exposure modeling of ENMs, along with other potentially useful modeling
frameworks and methods.
The search engines ScienceDirect and Google Scholar were used extensively to perform a
comprehensive literature review using different combinations of search criteria as described
below. Based on peer review comments, we also performed a search of the literature using the
33
-------
3.0 Review Methodology for Relevant Models and Methods
ISI Web of Knowledge database (http://apps.isiknowledge.com/), which includes the Web of
Science. The purpose of this additional search was to ensure that the use of ScienceDirect,
Google Scholar, and local university searches provided adequate coverage in describing research
related to the state-of-the-science.
Information pertaining to modeling ecological exposures to ENMs, particularly in aquatic
systems, was further evaluated. The websites of National Science Foundation (NSF)-funded
research centers were explored, as were publicly available conference proceedings and
nanotechnology-related informational websites. Initially, the information search was limited to
current models and approaches used in simulating fate and transport of engineered ENMs in
aquatic systems. However, because technical reports that describe exposure modeling of ENMs
were extremely limited, the search criteria were expanded to include other ultrafine particle types
such as aerosols, polymers, and colloids. In addition, the search included information that could
be relevant to fate and transport modeling (e.g., physical and chemical properties).
Elsevier's on-line technical documents service (ScienceDirect) was used as the primary source
for our literature searches. ScienceDirect is one of the largest online collections of published
scientific research available, containing over 9.6 million articles from over 2500 journals, and
over 6,000 e-books, reference works, book series and handbooks. Searches were also conducted
using the Google Scholar search engine, as well as the online libraries of North Carolina State
University (http://www.lib.ncsu.edu/) and the University of North Carolina at Chapel Hill
(http://www.lib.unc.edu/). As suggested above, the ISI Web of Science was used to verify that
we had characterized the landscape of published literature relevant to the fate and transport, and
exposure modeling of ENMs. This resource includes over 12,000 journals and more than 46
million records. The ISI Web of
Science is particular helpful in
identifying articles with the
greatest impact to the field; for
instance, the influence of a paper
on the field can be determined
based on the number of papers that
have cited it.
Nanotechnology
Nano scale
Nanoparticles
Etc.
Transport
Mobilization
We adopted a broad search
strategy by incorporating
combinations of key words, and
this set of search criteria was used
to identify models and approaches
that could potentially be used to
simulate the fate and transport of
ENMs. Figure 3-2 outlines the
most productive search terms—by
technical theme—that we used in
various combinations to identify
additional references.
Figure 3-2. Key search terms organized by theme.
34
-------
3.0 Review Methodology for Relevant Models and Methods
The largest number of documents retrieved from the expanded searches pertained primarily to
exposure modeling of ENMs, ENM toxicity, or safety assessments for workers involved in the
manufacturing of nanoparticles (NPs). In general, we found relatively few references (compared
with the results of our search) that could be considered specific to modeling the fate and
transport of, or characterizing the exposure to, ENMs released into the environment.
Relevant information sources were identified from several areas of study (e.g., mining, metals
transport, sedimentation processes) that had potential applicability. Understanding the key
processes and research challenges specific to ENMs was critical in the assessment of models
developed for other purposes. For instance, many researchers have pointed out that the
environmental behavior of some ENMs is similar to colloids; thus, part of our revised search
strategy focused on identifying colloidal fate and transport models for aqueous systems. In
addition, research on ENMs in other fields was considered (e.g., use of iron oxide ENMs in
groundwater remediation). For completeness, modeling tools were also searched across EPA
research offices/programs, other federal agencies, states and EPA regions, and international trade
and research organizations.
A complete listing of titles relevant to modeling the fate, transport, and exposure to ENMs
released into the environment is presented in Appendix A. Each of these references was
thoroughly reviewed, and a subset of these titles (highlighted in blue) was selected for the
detailed model evaluations provided in Appendix B.
3.2 Identifying Key Information Sources
The search strategy produced a tremendous number of journal articles, major reports, research
centers, and informational websites; however, few useful conference proceedings were identified
that provided materials relevant to fate and transport modeling of ENMs. The following
subsections highlight the most important information sources that we identified.
3.2.1 Journals
Table 3-1 lists the journals that produced the majority of relevant research articles that we
identified. Although a much wider variety of journals was included in the search, this list
presents those journals that published the highest numbers of relevant articles on modeling the
fate and transport of engineered ENMs (and similarly sized particles) in aquatic systems and
soils. These journals include well-established titles dealing with hydrology and environmental
modeling, as well as more recently published titles focusing specifically on nanotechnology.
Table 3-1. Journals Producing the Highest Number of Relevant Studies
Journal title First published
Water Resources Research 1965
Environmental Science and Technology 1967
Environmental Pollution 1970
Environmental Health Perspectives 1972
35
-------
3.0 Review Methodology for Relevant Models and Methods
Table 3-1. Continued
Journal title First published
Journal of Environmental Quality 1972
Advances in Water Resources 1977
Journal of Contaminant Hydrology 1986
Waste Management 1989
Ecotoxicology and Environmental Safety 1995
Environmental Toxicology and Chemistry 1996
Journal of Nanoparticle Research 1999
Nanomedicine: Nanotechnology, Biology, and Medicine 2005
Nanotoxicology 2007
Brief descriptions of these journals are provided below, adapted from each journal's website.
• Water Resources Research (WRR), 1965: a peer-reviewed scientific journal published
by the American Geophysical Union (AGU). AGU states that WRR is an
"interdisciplinary journal integrating research in the social and natural sciences of water."
• Environmental Science & Technology (ES&T), 1967: Published by the American
Chemical Society, the journal combines magazine and research sections. The news and
features section of ES&T presents objective reports and analyses of the major advances,
trends, and challenges in environmental science, technology, and policy for a diverse
professional audience. The research section seeks to publish papers that are particularly
significant and original. The types of papers published in the research section of ES&T
are research article, policy analysis, critical review, correspondence (comment/rebuttal),
and correction/addition (errata).
• Environmental Pollution, 1970: an international journal that focuses on papers that
report results from original research on the distribution and ecological effects of
pollutants in air, water and soil environments and new techniques for their study and
measurement. Findings from re-examination and interpretation of existing data are also
included. The editors are focusing on papers that provide new insights into environmental
processes and or the effects of pollutants.
• Environmental Health Perspectives (EHP), 1972: a monthly journal of peer-reviewed
research and news published by the U.S. National Institute of Environmental Health
Sciences, National Institutes of Health, Department of Health and Human Services.
EHP's mission is to serve as a forum for the discussion of the interrelationships between
the environment and human health by publishing in a balanced and objective manner the
best peer-reviewed research and most current and credible news of the field.
• Journal of Environmental Quality, 1972: covers various aspects of anthropogenic
impacts on the environment, including terrestrial, atmospheric, and aquatic systems.
Emphasis is given to the understanding of underlying processes rather than to monitoring.
Contributions reporting original research or brief reviews and analyses dealing with some
aspect of environmental quality in natural and agricultural ecosystems are accepted from
all disciplines for consideration by the editorial board.
36
-------
3.0 Review Methodology for Relevant Models and Methods
Advances in Water Resources, 1977: provides a forum for the presentation of
fundamental scientific advances in the understanding of water resources systems. The
scope of Advances in Water Resources includes any combination of theoretical,
computational, or experimental approaches used to advance fundamental understanding
of surface or subsurface water resources systems or the interaction between these
systems.
Journal of Contaminant Hydrology, 1986: an international journal publishing scientific
articles pertaining to the contamination of groundwater. Emphasis is placed on
investigations of the physical, chemical, and biological processes influencing the
behavior of organic and inorganic contaminants in both the unsaturated and saturated
zones. Articles on contamination of surface water are not included unless they
specifically deal with the link between surface water and groundwater.
Waste Management, 1989: an international journal devoted to the presentation and
discussion of information on the generation, prevention, characterization, monitoring,
treatment, handling, reuse and ultimate residual disposition of solid wastes, both in
industrialized and in economically developing countries.
Ecotoxicology and Environmental Safety, 1995: focuses on integrated mechanistic
research related to short- and long-term pathways and interactions of substances and
chemical mixtures in environmental systems and subsystems on their bioavailability,
circulation, and assimilation in target organisms, as well as biological responses of these
organisms, and damage mechanisms (endocrine disruption, genotoxicity); and on their
subsequent fate in the food chain, including humans.
Environmental Toxicology and Chemistry, 1996: seeks to publish papers describing
original experimental or theoretical work that significantly advances understanding in the
area of environmental toxicology, environmental chemistry and hazard/risk assessment.
Emphasis is given to papers that enhance capabilities for the prediction, measurement,
and assessment of the fate and effects of chemicals in the environment, rather than simply
providing additional data.
Journal of Nanoparticle Research, 1999: is an academic journal published by Springer.
It focuses mainly on physical, chemical and biological phenomena and processes in
structures of sizes comparable to a few nanometers. It covers the synthesis, assembly,
transport, reactivity, and stability of nanostructures and devices obtained via precursor
NPs, in various fields such as physics, chemistry, biology and health care.
Nanomedicine: Nanotechnology, Biology, and Medicine, 2005: presents theoretical
and experimental research results related to nanoscience and nanotechnology in life
sciences, including Basic, Translational, and Clinical research, and commercialization of
results. Article formats include Communications, Original Articles, Reviews,
Perspectives, Technical and Commercialization Notes, and Letters to the Editor. In
addition, regular features on our website will address commercialization, funding
opportunities, and societal, Public Health, and ethical issues of nanomedicine.
Nanotoxicology, 2007: is the first peer-reviewed academic journal devoted entirely to the
publication of research that addresses the potentially lexicological interactions between
nano-structured materials and living matter. The journal publishes the results of studies
that enhance safety during the production, use and disposal of ENMs.
37
-------
3.0 Review Methodology for Relevant Models and Methods
3.2.2 Reports
The reports shown in Table 3-2 were produced by governmental agencies and significant
research institutions.
Table 3-2. Major Reports on ENM Research Relevant to Environmental Exposure Modeling
Report title
Date
Institute
Author
Engineered Nanoparticles: Review of Health 2009
and Environmental Safety (ENRHES)
EMERGNANO: A review of completed and 2009
near completed environment, health and safety
research on ENMs and nanotechnology
Nanomaterial Research Strategy 2009
Approaches to Safe Nanotechnology: 2009
Managing the Health and Safety Concerns
Associated with Engineered ENMs
Edinburgh Napier University,
EU Joint Research Center,
Institute for Occupational
Medicine, Technical University
of Denmark
UK Department for
Environment, Food and Rural
Affairs (DEFRA)
US EPA
National Institute for
Occupational Safety and
Health (NIOSH)
Stone, V. (project
coordinator) of
Edinburgh Napier
University
Institute for
Occupational
Medicine
Morris, J., Wentsel, R.
(US EPA)
NIOSH
Sampling and Analysis of ENMs in the
Environment: A State-of-the-Science Review
Environmental Fate and Ecotoxicity of
Engineered Nanoparticles
Nanotechnology White Paper
Nanotechnology: A Research Strategy for
Addressing Risk
2008 US EPA
2007
2007
2006
Norwegian Pollution Control
Authority
US EPA
Varner, K.
Bioforsk
Nanotechnology Work
Group (US EPA)
Woodrow Wilson International Maynard, A.
Center for Scholars
Brief descriptions of the reports are provided below, adapted from the executive summary.
• Engineered Nanoparticles: Review of Health and Environmental Safety (ENRHES),
2009: presents a comprehensive and critical scientific review of the health and
environmental safety of four classes of ENMs: fullerenes, CNTs, metals and metal
oxides. The review considers sources, pathways of exposure to the health and
environmental outcomes of concern, followed by a risk assessment based upon this
information. The report includes an illustration of the state-of-the-art as well as on-going
work, while identifying knowledge gaps in the field. Prioritized recommendations have
been developed and set in the context of informing policy makers in the development of
methods to address exposure as it relates to the potential hazards posed by engineered
NPs, and in the development of appropriate regulation.
• lOM's EMERGNANO project for UK Department for Environment, Food and
Rural Affairs (DEFRA), 2009: this project involved a detailed review and analysis of
research carried out worldwide on Environment, Health, and Safety aspects of engineered
NPs, including issues relating to hazard, exposure and risk assessment and regulation,
38
-------
3.0 Review Methodology for Relevant Models and Methods
and made an assessment of how far 18 of 19 Research Objectives from this report's
predecessor have been met and which gaps still remain to be filled.
US EPA's Nanotechnology Research Strategy, 2009: The purpose of the Nanomaterial
Research Strategy is to guide the EPA's Office of Research and Development's program
in conducting focused research to inform ENM safety decisions that may be made under
the various environmental statutes for which EPA is responsible. This report focuses on
four areas that take advantage of EPA's scientific expertise as well as fill gaps not
addressed by other organizations. The four research themes are: (1) Identifying sources,
fate, transport, and exposure, (2) Understanding human health and ecological effects to
inform risk assessments and test methods, (3) Developing risk assessment approaches,
and (4) Preventing and mitigating risks.
NIOSH, Approaches to Safe Nanotechnology: Managing the Health and Safety
Concerns Associated with Engineered ENMs, 2009, this report aims to provide an
overview of what is known about the potential health hazards of engineered NPs and
measures that can be taken to minimize workplace exposures. It provides a detailed list of
potential health concerns, reviews numerous related studies, and offers a number of
recommendations to improve safety regulations and monitoring.
US EPA's State-of-the-Science Review, 2008: this state-of-the-science review was
undertaken to identify and assess currently available sampling and analysis methods to
identify and quantify the occurrence of ENMs in the environment. The environmental
and human health risks associated with ENMs are largely unknown, and methods needed
to monitor the environmental occurrence of ENMs are very limited or nonexistent.
Because this research is current and ongoing, much of the applicable information is found
in gray literature (e.g., conference proceedings, communications with research scientists
and other experts).
Norwegian Pollution Control Authority, Environmental Fate and Ecotoxicity of
Engineered Nanoparticles, 2008: This report is an overview of the scientific knowledge
on potential negative effects that engineered NPs may have on the environment. So far,
scientific evidence show that some NPs have toxic effects under laboratory conditions,
but practically nothing is known about their mobility and uptake in organisms under
environmental conditions. This report was written by researchers at Bioforsk Soil and
Environment.
US EPA's Nanotechnology White Paper, 2007: the purpose of this paper is to inform
EPA management of the science needs associated with nanotechnology, to support
related EPA program office needs, and to communicate these nanotechnology science
issues to stakeholders and the public. The Nanotechnology Research Framework outlines
how EPA will strategically focus its own research program to provide key information on
potential environmental impacts from human or ecological exposure to ENMs in a
manner that complements other federal, academic, and private-sector research activities.
Woodrow Wilson International Center for Scholars, Nanotechnology: A Research
Strategy for Addressing Risk, 2006: this report addresses the current state of
nanotechnology risk research and what needs to be done to help ensure the technology's
safe development and commercialization. A strategic research framework is developed
that identifies and prioritizes what the author believes are the critical short-term issues.
39
-------
3.0 Review Methodology for Relevant Models and Methods
Recommendations are made on how a viable strategic research plan might be
implemented.
3.2.3 Research Centers
We identified several leading U.S. research centers that are at the forefront of studies on the
behavior of ENMs in the environment. Each of these research centers are the recipients of grants
from the National Science Foundation and/or the EPA, and are performing cutting edge research
on a variety of technical issues central to understanding the fate and transport characteristics of
ENMs in the environment. We recognize that there are also other research centers in the U.S. as
well as international research centers (especially in the EU) that are involved in ENM research;
however, our search strategy did not identify other centers (domestic or international) engaged in
significant research that was specific to environmental fate and transport modeling of ENMs in
aquatic systems. In addition, the intent was to identify centers that focused on the environmental
behavior of ENMs rather than on characterizing ENM properties or developing engineering
applications for nanomaterials. Table 3-3 lists the university and the name of the research center,
and provides a link to the center's website.
Table 3-3. Key Research Centers for Nanotechnology Research
Academic Institution Research Center Web Address
Duke University Center for Environmental Implications of http://www.ceint.duke.edu/
NanoTechnology (CEINT)
Rice University Center for Biological and Environmental http://cben.rice.edu/
Nanotechnology (CBEN)
University of California, Los University of California Center for http://cein.cnsi.ucla.edu/paqes/
Angeles Environmental Implications of
Nanotechnology (UC CEIN)
Brief descriptions of each research center are provided below, adapted from the center's website.
• Duke University's CEINT was created in 2008 with funding from the National Science
Foundation and the US EPA. As described on its website, CEINT is elucidating the
relationship between a vast array of ENMs and their potential environmental exposure,
biological effects, and ecological consequences. Headquartered at Duke University,
CEINT is a collaboration between Duke and a number of leading universities and
researchers (comprehensive list found here: http://www.ceint.duke.edu/participating-
institutions ). CEINT performs fundamental research on the behavior of nano-scale
materials in laboratory and complex ecosystems. Research includes all aspects of ENM
transport, fate and exposure, as well as ecotoxicological and ecosystem impacts. CEINT
is developing risk assessment models to provide guidance in assessing existing and future
concerns surrounding the environmental implications of ENMs. This address provides a
detailed list of CEINT's partners and participating institutions:
• CBEN's mission is to discover and develop ENMs that enable new medical and
environmental technologies. The mission is accomplished by the following:
40
-------
3.0 Review Methodology for Relevant Models and Methods
o Fundamental examination of the wet/dry interface between ENMs, complex aqueous
systems, and ultimately our environment (Theme 1).
o Engineering research that focuses on multifunctional NPs that solve problems in
environmental and biological engineering (Themes 2, 3).
o Educational programs that develop teachers, students, and citizens who are well
informed and enthusiastic about nanotechnology.
o Innovative knowledge transfer that recognize the importance of communicating
nanotechnology research to the media, policymakers, and the general public.
The Center's research focuses on investigating and developing nanoscience at the wet/dry
interface. Water, the most abundant solvent present on Earth, is of unique importance as
the medium of life. The Center's research activities explore this interface between ENMs
and aqueous systems at multiple length scales, including interactions with solvents,
biomolecules, cells, whole-organisms, and the environment. These explorations form the
basis for understanding the natural interactions that ENMs will experience outside the
laboratory, and also serves as foundational knowledge for designing biomolecular/ENM
interactions, solving bioengineering problems with nanoscale materials, and constructing
nanoscale materials useful in solving environmental engineering problems.
• UC CEIN is the sister center to Duke's CEINT. Headquartered at the University of
California, Los Angeles (UCLA), it includes the University of California at Santa
Barbara and other UC partners. According to UC CEIN's website, UC CEIN will explore
the impact of libraries of engineered ENMs on a range of cellular life forms, organisms
and plants in terrestrial, fresh water and sea water environments. By being able to predict
which ENM physicochemical properties are potentially hazardous, the UC CEIN will be
able to provide advice on the safe design of engineered ENMs from an environmental
perspective. While UCLA serves as the lead campus for the UC CEIN, researchers from a
range of other institutions and organizations are involved in the UCLA-based UC CEIN
research, and a comprehensive list of domestic and international partners may be found
here: http://cein.cnsi.ucla.edu/pages/institutions.
3.2.4 Informational Web Sites
The publicly-available websites shown in Table 3-4 were identified because they provided
valuable information related to fate and transport of engineered ENMs in aquatic ecosystems.
The table provides a brief summary of their mission along with the current web address.
Table 3-4. Publically Available Websites Providing Valuable Information ENM Exposure Modeling
Website Mission Web Address
National Nanotechnology coordinate Federal nanotechnology research http://www.nano.qov/
Initiative (NNI) and development
International Council on improve communication between http://icon.rice.edu/
Nanotechnology (ICON) stakeholders involved with risk assessment
and research
NanoRISK newsletter addressing nanotechnology risk http://www.nanorisk.org/
assessment
41
-------
3.0 Review Methodology for Relevant Models and Methods
Table 3-4. Continued
Website
Mission
Web Address
NanoScienceWorks
Nanotechnology Now
(NN)
Nanowerk
Project on Emerging
Nanotechnologies
not-for-profit portal for the nanoscience
research community
created to serve the information needs of
business, government, academic, and public
communities
serve as portal for nanotechnology and
nanosciences; provide links and editorial
content
ensure that risks are minimized, consumer
engagement remains strong, potential
benefits realized
http://www.nanoscienceworks.orq/
http://www.nanotech-now.com/
http://www.nanowerk.com/
http://www.nanotechproiect.org/
Brief descriptions of these websites are provided below, adapted from each website.
• http://www.nanoscienceworks.org/
NanoScienceWorks.org is a not-for-profit community portal for the nanoscience research
community. The website provides a comprehensive variety of nanotechnology and
nanoscience-related information, in addition to an encyclopedia and a free monthly
newsletter.
"NanoScienceWorks.org serves the nano community as a gateway to the news, journals,
books, and articles that support and drive nano research and development. We invite you to
explore these resources, view our slidecasts, and join our networking database of nano-
involved people and institutions from around the world."
• http://icon.rice.edu/
The creation of a sustainable nanotechnology industry requires meaningful and organized
relationships among diverse stakeholders. The International Council on Nanotechnology
(ICON) aims at providing such interactions for a broad set of members. Managed by Rice
University's Center for Biological and Environmental Nanotechnology,
ICON activities promote effective nanotechnology stewardship through risk assessment,
research and communication. By pooling the resources of the nanotechnology industry,
government and non-government organizations and academia, ICON can cost-effectively
provide a wide range of synergistic projects that serve the interests of all stakeholders. CBEN
provides a financial and administrative structure for ICON, such that ICON members are also
members of CBEN.
• http://www.nano.gov/
The National Nanotechnology Initiative (NNI) is the program established in fiscal year 2001
to coordinate Federal nanotechnology research and development. The NNI provides a vision
of the long-term opportunities and benefits of nanotechnology. By serving as a central locus
for communication, cooperation, and collaboration for all Federal agencies that wish to
42
-------
3.0 Review Methodology for Relevant Models and Methods
participate, the NNI brings together the expertise needed to guide and support the
advancement of this broad and complex field.
• http://www.nanowerk.com/
Nanowerk is committed to educate, inform and inspire about nanosciences and
nanotechnologies. As the leading nanotechnology and nanosciences portal, nanowerk.com
delivers useful, entertaining and cutting-edge information from all things nano. Nanowerk
has become the premier nanotechnology portal due to the depth, rich scope and relevance of
our unique editorial content and the comprehensive resources that we put at users' fingertips.
Editorial content Scientists appreciate the publicity they receive through our articles and in
turn help spread the word about Nanowerk among their colleagues and scientific
communities. On average Nanowerk run between 70-100 news articles every week. T news
section is separated into Business News and Research & General News. Nanowork also
produces a newsletter deals explicitly with the risks involved in nanotechnology. "Much of
nanotechnology today is about producing nanoscale particles that, due to their size, have
significantly more catalytic active surfaces. [This newsletter tries] to support a debate on the
very real issues that we are facing today: the fact that engineered ENMs such as carbon
nanotubes or titanium dioxide particles are finding their way from scientists' laboratories into
commercial products and we don't understand the risks they pose to health and
environment." Nanorisk is a bi-monthly newsletter published by Nanowerk LLC.
• http://www.nanotechproject.org/
The Project on Emerging Nanotechnologies was established in April 2005 as a partnership
between the Woodrow Wilson International Center for Scholars and the Pew Charitable
Trusts. The Project is dedicated to helping ensure that as nanotechnologies advance, possible
risks are minimized, public and consumer engagement remains strong, and the potential
benefits of these new technologies are realized. The Project on Emerging Nanotechnologies
collaborates with researchers, government, industry, NGOs, policymakers, and others to look
long term, to identify gaps in knowledge and regulatory processes, and to develop strategies
for closing them.
• http://www.nanotech-now.com/
Nanotechnology Now (NN) covers future sciences such as Nanotechnology, Molecular
Nanotechnology (MNT), MicroElectroMechanical Systems (MEMS),
NanoElectroMechanical Systems (NEMS), Nanomedicine, Nanobiotechnology,
Nanoelectronics, Nanofabrication, Computational Nanotechnology, Quantum Computers,
and Artificial Intelligence. NN was created to serve the information needs of business,
government, academic, and public communities. And with the intention of becoming the
most informative and current free collection of nano reference material. We will cover:
related future sciences, issues, news, events, and general information, and make this a place
to come for information, stimulating debate, and research info.
43
-------
3.0 Review Methodology for Relevant Models and Methods
• http ://www.nanorisk. org/
This newsletter deals explicitly with the risks involved in nanotechnology. "Much of
nanotechnology today is about producing nanoscale particles that, due to their size, have
significantly more catalytic active surfaces. [This newsletter tries] to support a debate on the
very real issues that we are facing today: the fact that engineered ENMs such as carbon
nanotubes or titanium dioxide particles are finding their way from scientists' laboratories into
commercial products and we don't understand the risks they pose to health and
environment." Nanorisk is a bi-monthly newsletter published by Nanowerk, LLC.
3.3 Summary of Recent Reports and Compendia
As compared to chemicals in their conventional form, the unique properties of ENMs have led to
concerns about the potential health and ecological risks that might be associated with exposure to
ENMs following environmental release. Given the rapid growth in the manufacture and use of
ENMs, there have been a number of recent research initiatives and published reports that focus
on developing basic information about ENMs. In general, these reports fall into one of the
following four categories:
1. Reports that characterize the basic physical and chemical properties of ENMs
2. Reports that address aspects of modeling (e.g., particle transport) relevant to the
predictive risk assessment of environmental releases of ENMs
3. Reports that summarize the state of current knowledge on the health and
environmental risk assessment of ENMs
4. Reports that propose governance frameworks for the handling and safe management
of ENMs.
The first category covers a wide range of research, much of which is geared towards the
manufacturing aspects of ENMs. Although this information on ENMs is critical to fate and
transport modeling, it was not the focus of this effort and, to some degree, was included in EPA's
state-of-the-science review on sampling and analysis of ENMs in the environment (US EPA,
2008). As discussed in Section 1, the second category was the primary focus of this report, and
the results of the model reviews are presented in Sections 4.2 and 4.3. Sections 3.3.1 and 3.3.2
summarize recent reports and compendia that fall into the last two categories. Section 3.3.3
addresses recent reports and compendia that pertain to the risk assessment of ENMs but were
beyond the scope of this report.
3.3.1 Summary Reports on the Current State of Knowledge
This section describes several recent articles on the health and environmental risk assessment of
ENMs that provide significant insight into the current state of risk assessment science as it
pertains to ENMs. Table 3-5 lists selected articles that we believed provide the most salient
discussion of issues, uncertainties, and available data for the risk assessment of ENMs. It should
be emphasized, again, that ENMs cover a wide range of compounds and that our use of this term
should not be interpreted as a science-based simplification of nanoscale materials. Rather, it is a
convention adopted for the purposes and readability of this report.
44
-------
3.0 Review Methodology for Relevant Models and Methods
Table 3-5. Summary Reports on ENM Research Relevant to Environmental Exposure Modeling
Report title Date Author
The known unknowns of nanomaterials: Describing and characterizing onnQ C' t I
uncertainty within environmental, health, and safety risks neger e a .
Redefining risk research priorities for nanomaterials 2010 Grieger et al.
Nanomaterial risk assessment and risk management: Review of regulatory 200g ... . „ .. . F
frameworks ' '' '
Decreasing Uncertainties in Assessing Environmental Exposure, Risk, and onno w t i
Ecological Implications of Nanomaterials 2UUa wiesner et al-
Grieger et al. (2009) address the shortfalls surrounding the modeling of ENMs by
identifying the uncertainties associated with these models. By discovering these known
unknowns, the paper intends to generate a pathway to shrink the knowledge gaps that
currently persist in the modeling of ENMs. The article identifies 31 reports and articles
published by leading scientists and authorities for ENMs, classifying the uncertainties
addressed in each report and presenting them in a table that categorizes each uncertainty
with respect to environmental health and safety. From the assessment, the article
concludes the current level of knowledge to be in an early state in need of additional
studies to decrease knowledge gaps. The article thus estimates that current quantitative
risk assessments may produce premature results. Lastly, the article recommends that the
focus of the research be given to the assessment and development of test procedures and
equipment to fully characterize ENMs so that uncertainties will be most effectively
reduced in the near term.
Grieger et al. (2010) discuss the recent advancement in the risk assessment approach for
ENMs. The authors argue that due to the timeframe for the current approach to
responsible development of ENMs, alternative approaches must be explored. The article
evaluates possible alternative approaches and adaptive evaluation frameworks
(precautionary matrices, multi-criteria decision analysis, etc.) and governance
frameworks (International Risk Governance Council, Environmental Defense and
Dupont, etc.) currently in place. Intended for decision makers, this report suggests the
need for surveillance of ENMs in light of the high levels of uncertainty with ENMs.
Linkov and Satterstrom (2009) review the current ENM risk management frameworks
using a regulatory pyramid. Thirteen frameworks and related documents were reviewed
in all, including the USEPA white paper on nanotechnology (US EPA, 2007), DEFRA
(2005), and SCENfflR (2005). The authors reviewed the frameworks by the following
categories: (1) science and research aspects; (2) legal and regulatory aspects; (3) social
engagement and partnerships; and (4) leadership and governance. The regulatory pyramid
consists of 4 levels relating to the time frames of immediate, short term, medium term,
and long term ordered from bottom to top respectively. The authors find that appropriate
tools in the bottom (immediate) are largely lacking and recommend an adaptive
framework be utilized to manage nanotechnology risk.
Wiesner et al. (2009) address the uncertainties surrounding environmental exposure, risk
and ecological implications of ENMs. The authors characterize these uncertainties by
asking 4 questions concerning: (1) ENM properties and environmental conditions that
45
-------
3.0 Review Methodology for Relevant Models and Methods
control the spatial and temporal distribution of ENMs in the environment; (2)
fundamental differences between natural, incidental and manufactured ENMs; (3) nano
effects on bioavailability, toxicity, and other environmental end points; and (4) effects
ENMs may have on ecosystems; The authors then use the findings from these 4 questions
to suggest methods for conducting risk assessments of emerging ENMs.
3.3.2 Governance Frameworks
This section summarizes recent proposals to govern the production and risk management of
ENMs in the face of significant uncertainty due to a lack of research data on the environmental
behavior and effects of ENMs. Renn (2008) define risk governance as a process that includes the
totality of actors, rules, conventions, processes and mechanisms concerned with how relevant
risk information is collected, analyzed and communicated, and how management decisions are
taken. Table 3-6 lists the articles and reports that represent the most recent governance
proposals.
Table 3-6. Governance Frameworks on the Production of Nanomaterials
Report title Date Author
Essential features for proactive risk management 2009 Murashov and Howard
Strategic approaches for the management of environmental risk 2009 Owen et al.
uncertainties posed by nanomaterials
Moving toward exposure and risk evaluation of nanomaterials: challenges 2009 Thomas et al.
and future directions
The Framing Nano Governance Platform: A New Integrated Approach to the 2009 Widmer et al.
Responsible Development of Nanotechnologies
Murashov and Howard (2009) propose a proactive approach to the management of
occupational health risks based on six guidelines. The authors suggest first to utilize
qualitative risk assessments based on expert judgments and extrapolations from existing
data for similar materials. Secondly, develop strategies that quickly adapt to
accumulating risk assessment information and to refine risk management requirements.
The authors promote the ISO (International Organization for Standardization) Concept
Database, which aims to update risk databases in real time to aid in decision making. The
third guideline is to embody an appropriate level of precaution due to the lack of
qualitative risk assessment because of the amount of uncertainty present. The fourth
guideline is to generate global governance so that risk management is equivalent across
the spectrum of emerging technology firms. The fifth guideline suggests the ability to
elicit strong voluntary cooperation among firms. Lastly, the authors suggest that there be
a high level of stakeholder involvement based on the belief that involving all stakeholders
will broaden the knowledge of potential benefits and risks as they pertain to ENMs and
their production.
Owen et al. (2009) identifies one area of uncertainty surrounding the modeling of
ENM—the complexity of their behavior in natural systems. The article then outlines two
methods to address this issue: (1) a hazard-driven approach; and (2) an exposure-driven
approach. The hazard-driven approach suggests a large data gathering task utilizing
46
-------
3.0 Review Methodology for Relevant Models and Methods
extensive toxicity assessments over many types of organisms using endpoints that would
cover all potential exposure routes. The exposure-driven approach would focus on the
development and subsequent validation of a conceptual model of exposure for the ENM
of concern by utilizing a life cycle assessment that considers sources and pathways of
exposure during production, use and end-of-life. The model would need to incorporate
properties specific to the transport of ENMs rather than properties used for traditional
chemical assessments. The article concedes that both approaches would depend on a
considerable time lag between data gathering and resultant decision making based on the
information obtained. Therefore, the article suggests that environmental surveillance
approaches could be used in the near term to act as a safety net, but also advocates that it
is not yet clear how fit for purpose this monitoring may be.
• Thomas et al. (2009) implicate the need for exposure assessment due to the explosive
growth of nanotechnology. This call for action stresses this need as a critical factor in the
risk assessment of ENMs, which is important in fostering their sustainable development.
The article argues that an assessment that identifies and characterizes the contact and
uptake of compounds into organisms which may result in health effects is essential in
eliminating harmful chemical exposures of humans and the environment. The authors
explore the avenues of exposure from a life cycle perspective, exposure metrics, and
exposure assessment activities relevant to ENMs to explain the shortcomings of current
efforts, as well as justify the need for more action.
• Widmer et al. (2009) describe the FramingNano governance platform, which was
undertaken to create proposals for a workable governance platform bases on: (a) the
analysis of regulatory processes of nanotechnology; (b) consultation with stakeholders to
define key issues; and (c) dissemination of information on governance of nanotechnology
to allow input to its development. The overall objective of the FramingNano Governance
Platform is to promote responsible development of nanotechnology without hindering
innovation and commercial growth. As a result, the platform proposes guidance on 4
different levels on nanotechnology development: (1) Technical and organizational -
prioritizing research needs; (2) Communication and dialogue - effective information
passing to aid in policy implementation; (3) Institutional - management of policy for
responsible development of nanotechnologies; and (4) International harmonization - to
aid in global governance. The research conducted by the FramingNano project concluded
that governance and regulation of nanotechnologies is a dynamic process and must be
continuously adapted as emerging scientifically relevant information becomes available.
3.3.3 Other Relevant Nanomaterial Reports and Compendia
This section (Table 3-7) presents other articles that may be of importance to the progression of
nanotechnology risk assessment but fell outside the scope of this review (i.e., toxicology). We
did not report the summaries of these reports and compendia, but recommend these articles for
readers that wish to gain insight on related aspects of risk assessment.
47
-------
3.0 Review Methodology for Relevant Models and Methods
Table 3-7. Other Relevant Reports and Compendia on the Research of Nanomaterials.
Report title Date Author
Factors Influencing the Partitioning and Toxicity of Nanotubes in the Aquatic 2008 Kennedy et al.
Environment
Ecotoxicity of engineered nanoparticles to aquatic invertebrates: a brief 2008 Baun et al.
review and recommendations for future toxicity testing
Nanoparticles: Their potential toxicity, waste and environmental 2009 Bystrzejewska-Piotrowska
management (literature review) et al.
Safety Assessment for Nanotechnology and Nanomedicine: Concepts of 2010 Oberdorster et al.
Nanotoxicology
48
-------
4.0 Model Reviews
Chapter 4.0
Model Reviews
In this section of the state-of-the-science report we evaluate several modeling approaches and
specific models according to their relevance and applicability for predicting ENM transport. The
focus of the review is on aquatic and terrestrial systems, including transport in surface water,
sediments, groundwater, and soils. The evaluation also considers biological uptake and
multimedia models (air models are not considered). In addition, the focus of this review is on
organic-based nanomaterials. Many of the approaches reviewed are relevant to other
nanomaterials (notably metals); however, evaluation of these materials can require additional
modeling tools (such as geochemical speciation modeling) that were outside the scope of this
review.
Section 4.1 presents the model evaluation framework, which provides a systematic approach for
reviewing information about models and is based on National Research Council guidelines
(NRC, 2007). The remainder of Section 4 presents results of the model reviews. Summary
descriptions of specific models and approaches are provided in this section. Appendix B
provides more detailed reviews for the models and approaches that show more promise for ENM
transport modeling. The model reviews are divided into two main categories: fate and transport
models (Section 4.2) and alternative approaches (Section 4.3).
4.1 Model/Method Evaluation Criteria
The model evaluation framework implemented for this assessment provides a systematic and
consistent approach for reviewing and summarizing information about models. The review
categories were developed to be consistent with the National Research Council paper, Models in
Environmental Regulatory Decision Making (NRC, 2007). In this report, the NRC assesses how
models support the EPA's environmental regulatory process. The development and application
of regulatory models is described along with recommended considerations for selecting and
using models to support EPA programs. The NRC document describes criteria for evaluating
whether a model and its results provide a sound basis for regulatory decision making. Using the
NRC document as a guide, we compiled a series of key considerations for model evaluation and
organized them into the general categories used in the model reviews. Appendix B provides
model reviews that are structured according to these categories. We have provided reviews for
models that have been developed for and/or applied to ENMs as well as other models that have
potential applicability (in their present form) for the evaluation of ENM transport in the
environment.
Purpose and Scope. What is the model purpose? What transport media are considered (i.e.,
groundwater, surface water, multimedia)? What processes are simulated? What is the conceptual
basis and what are the primary assumptions? What spatial and temporal scales does the model
consider? What kind of results does the model produce (e.g., media concentrations, risk,
probability)?
Background and History. How extensively has the model been used and applied (e.g., a single
academic study versus an established model used for extensive regulatory decision making)?
49
-------
4.0 Model Reviews
What kind of peer review has the model been subjected to? Has the model been verified through
comparison with other models or laboratory experimental results? Has the model been validated
through comparisons with measured environmental data?
Complexity. What physical and chemical processes are considered? What is the mathematical
representation (e.g., analytical or numerical solution)? How extensive are the input data
requirements?
Consideration of Uncertainty. Does the model account for uncertainty and, if so, how? Is the
model process based or statistical? Is the model deterministic or probabilistic? Does the model
implementation include tools for sensitivity and/or uncertainty analysis?
Availability and Usability. Is a user-friendly interface available? Does the model rely on
proprietary algorithms and/or user interfaces? Is the model documentation complete and
transparent? Is the source code available for potential enhancements/modifications?
Applicability to ENM Behavior. Does the model consider key processes and chemical
properties relevant for the specific environmental medium or media considered in determining
ENM behavior (e.g., aggregation/disaggregation, attachment/detachment in aqueous systems)?
How does the model account for the gaps in the data that are necessary for traditional models?
How well does this model respond to updated scientifically relevant information? What kind of
information can be gathered from these types of models? What interpretations can be made from
the findings of these types of models?
4.2 Review of Environmental Fate and Transport Models
In this section we describe several environmental fate and transport models and their potential
applicability for evaluating ENMs. The model categories considered are based primarily on
NRC's review of regulatory modeling practices at EPA (NRC, 2007). There is overlap in model
constructs between some of these categories. For example porous media transport mechanisms
may be important both for sediments in a surface water model as well as a groundwater transport
model. Nevertheless, the model categorization used in this document is consistent with the
typical structure and scope of many existing environmental fate models.
Models within the following general categories are considered:
• Surface water models. Most surface water quality models account for interactions
between surface water and underlying sediments (e.g., sedimentation and resuspension).
Thus, this category may include models that evaluate fate and transport within the water
column as well as within underlying sediments.
• Subsurface models. This category includes models that simulate environmental fate
within soils, the unsaturated zone (below the soil zone and above the groundwater table),
and saturated groundwater.
• Biological uptake models. This category focuses on models predicting the uptake of
chemicals into biological organisms and associated potential bioaccumulation.
50
-------
4.0 Model Reviews
• Multimedia models. These models account for processes and mass transfer across
multiple environmental media. There is redundancy between multimedia models and the
other model categories. In fact, some multimedia modeling systems explicitly include
media-specific submodels (e.g., the EXAMS surface water model within the Multimedia,
Multipathway, Multireceptor Exposure and Risk Assessment [3MRA] modeling system).
Nevertheless, multimedia models deserve focused consideration given their extensive use
in many traditional environmental risk assessments.
Within each of the model categories listed above, we have evaluated models of three types:
• Models specific to nanomaterials. These models were developed for and/or used
specifically to evaluate ENM transport. The models considered were identified during the
literature review documented in Section 3.
• Established regulatory models. Given the focus of this assessment on supporting EPA's
evaluation of models for regulatory support, we have provided a discussion of several
existing, established models currently used by EPA to evaluate chemical fate in the
environment. In most cases, these models are not appropriate for modeling ENMs in their
present form, because they do not consider critical properties and processes for ENMs.
The discussion will highlight limitations of such models and modeling approaches. The
regulatory models evaluated correspond to those listed in NRC's review of EPA
regulatory modeling practice (NRC, 2007). Although these models represent only a
subset of the available models used by EPA, they are among the most widely applied, and
they are representative of typical risk assessment modeling practice by the agency.
• Other models. In some cases, existing models developed for other contaminants may
provide potentially useful approaches for simulating ENM environmental transport. For
example, given their particulate nature, modeling approaches for colloid transport in
porous media may be relevant to nanoparticles.
4.2.1 Surface Water Models
As discussed in Section 2.4.1, many traditional environmental fate models of chemicals in
aquatic systems consider some or all of the following important processes: dissolution,
volatilization, adsorption, biological uptake, photolysis, hydrolysis, and biodegradation (see
Section 2.4.1). Less common are environmental fate models that consider processes of
aggregation, attachment, and sedimentation, all processes critical to understanding and predicting
the environmental fate of ENMs (Sections 2.3 and 2.4). Models predicting ENM behavior in
aquatic environments should account for these critical processes related to particulates in natural
systems. The evaluations of surface water fate and transport models in this section primarily
consider whether models account for these key processes.
4.2.1.1 Surface Water Models of Nanomaterials
This section describes surface water models identified in the literature that have been developed
and/or applied specifically for the evaluation of ENM fate in the environment. Appendix B
provides more detailed reviews of each of these models.
51
-------
4.0 Model Reviews
Mackay et al. (2006) developed a stochastic probability model predicting the environmental
stability of nanoparticle suspensions in aqueous solutions and the associated uncertainty. The
model simulates settlement utilizing critical buoyancy properties and the Boltzmann equation.
Rates of aggregation are estimated based on molecular collision and adhesion coefficients.
Appendix B provides a more detailed review of the Mackay et al. (2006) model.
Boncagni et al. (2009) implemented an experimental study of the exchange of titanium dioxide
nanoparticles between streams and streambed sediments. They evaluated the degree of
aggregation and sedimentation under a range of conditions (pH and water flow velocity). They
utilized the process based model of colloids in surface water systems developed by Packman et
al. (2000) to interpret the results. The model was formulated based on advective pumping theory,
colloid filtration, and settling. The Packman model is discussed further in Section 4.2.1.3 and is
reviewed in Appendix B.
Koelmans et al. (2009) performed a compartmental modeling analysis of mass transfer between
surface water and sediments, considering particulate transport processes. The model estimated
steady state concentrations of carbon-based nanoparticles by accounting for processes of
sedimentation, aggregation, degradation, and burial in deeper sediment layers. The model
assumes: (1) a distinct, mixed biologically active layer; (2) transport of manufactured carbon
nanoparticles to sediment is through sedimentation; and (3) the removal of manufactured carbon
nanoparticles can be modeled as a first order decay process. Their analysis suggested that
concentrations of manufactured carbon-based nanoparticles in aquatic sediments will likely be
negligible relative to levels of black carbon nanoparticles (incidental ENMs generated as
combustion byproducts). Appendix B provides a more detailed review of the Koelmans et al.
(2009) model.
4.2.1.2 Regulatory Surface Water Models
This section discusses several established surface water models that have been used in risk
assessments to support EPA regulatory programs. The models evaluated correspond to those
listed in NRC's review of EPA regulatory modeling practice (NRC, 2007). Although these
models represent only a subset of the available surface water models used by EPA, they are
among the most widely applied, and they are representative of typical risk assessment modeling
practice by the agency.
The three reviewed models are moderate complexity, conceptual models (see Box 4-1). The
models do consider some of the key processes characterizing particulate transport in aqueous
systems, including sedimentation, resuspension, and particulate advection. However, their
current applicability for predicting EMM behavior is significantly limited by lack of knowledge
and lack of available, empirical data characterizing ENMs. Furthermore, it is as yet unknown
whether the associated lumped-parameter formulations will be adequate for simulating the
environment behavior of ENMs or whether alternative modeling constructs will be required. For
example, if sufficient empirical knowledge becomes available to support model
parameterization, can mass transfer between the water column and underlying sediments be
adequately modeled using a lumped mass-transfer-rate formulation? Given the limited
applicability of regulatory surface water models for evaluating ENMs at this time, we provide
52
-------
4.0 Model Reviews
only a brief review of each model
and have not included more detailed
reviews of these models in
Appendix B
HSPF (The Hydrological
Simulation Program-FORTRAN)
is a modeling package for
simulating watershed hydrology and
water quality. HSPF adopts a basin-
scale approach, incorporating
pollutant source models and fate
and transport in one dimensional
stream channels. The model
accounts for watershed hydrology,
including sediment runoff processes
along with in-stream hydraulic and
sediment-chemical interactions.
Simulation results include time
series of the runoff flow rate,
sediment load, and contaminant
concentrations, as well as water
quantity and quality. HSPF
considers up to three sediment types
(sand, silt, and clay) in addition to a
single dissolved organic chemical
and transformation products of that
chemical. The model considers the
following reaction processes:
hydrolysis, oxidation, photolysis,
biodegradation, volatilization, and
sorption. The model accounts for
particulate settling and potential
resuspension. Resuspension is
modeled based on the shear stress at
the sediment water interface and the
capacity to transport particulates at
a particular flow. Mass transfer with
sediments is modeled as
sorption/desorption and
deposition/scour processes. Fate and
transport mechanisms within the
sediments (e.g., pore water flow,
bioturbation) are not modeled.
HSPF has been used in hydrologic
Box 4-1. Complexity and Empiricism in Environmental Risk
Assessment Modeling
Fate and transport models used in risk assessment often do not
include extremely detailed, mechanistic formulations of fate
processes. Rather, these models often encapsulate detailed
processes using simpler formulations. One example would be a
surface water model that describes the mass transfer between a
water column and underlying sediments using a single mass
transfer rate even though the underlying processes may be quite
varied and complex (sedimentation, scouring, adsorption,
bioturbation, diffusion, pumping exchange). In this case, the mass
transfer rate could be considered a lumped parameter designed to
capture the cumulative effect of a range of relevant processes.
Models that adopt such an approach have been referred to as
conceptual models in that the model provides a conceptual
framework for the underlying processes (Wainright and Mulligan,
2004).
One can consider a continuum of model types from empirical
models that rely only on data with no underlying conceptualization
of the system (e.g., a fitted regression equation) to physical models
that are based on a detailed, mechanistic and spatially explicit
understanding of the underlying processes. In general, physical
models are fully distributed spatially; empirical models are fully
lumped with no explicit spatial representation. Conceptual models
(i.e., most risk assessment fate and transport models) are typically
semi-distributed, falling in the middle of the continuum between
physical and empirical models.
Conceptual models may appear to compromise scientific rigor by
ignoring known complexities. However, such an approach is often
necessary in order to predict environmental behaviors given the
typically extreme variability, complexity, and uncertainty associated
with natural systems. In other words, additional scientific detail and
complexity do not necessarily increase model reliability, particularly
when dealing with highly variable and uncertain systems. However,
it is generally not possible to make predictions using conceptual
models a priori (i.e., based on theoretical considerations alone).
Rather, conceptual models must be grounded in empirical evidence
in order to produce realistic predictions of environmental behavior.
For example, model input parameters such as mass transfer rates
and reaction terms may be adjusted until model predictions are
reasonably close to measured data (i.e., model calibration). In
addition, conceptual models may parameterized using regressions
previously developed from empirical data (e.g., correlations relating
surface water volatilization rates to wind speed). Large-scale (e.g.,
national) risk assessments may not have sufficient available data to
support model calibration; nevertheless, the model input
distributions are typically based on empirical knowledge gained
through laboratory experimentation and field data collection.
In summary, models often used in regulatory risk assessment are
moderately complex, conceptual models. Parameterization of such
models must be based on empirical evidence in order to make
reliable predictions. Very limited knowledge (experimental and field
data) of ENM transport is available to support the use of such
modeling frameworks at this time.
53
-------
4.0 Model Reviews
and water quality evaluations, including analysis of pesticide runoff and agricultural best
management practices.
WASP provides a dynamic compartment-modeling approach for aquatic systems, including both
the water column as well as the underlying sediments. The model can evaluate 1, 2, and 3
dimensional systems and a variety of pollutant types, including particulates. The model considers
processes of advection, dispersion, point and diffuse mass loadings and boundary mass transfers.
Sediment transport processes include advection, dispersion, settling (and sedimentation), as well
as erosion to the water column from the sediment layer. Example uses of the WASP model
include evaluations of eutrophication, phosphorus loading, bacterial contamination, as well as
PCB, VOC, and heavy metal pollution.
QUAL2K provides a relatively simple model for simulating flow and water quality in rivers and
streams. The model has been used to evaluate the environmental impact of pollution discharges
along rivers from point and non-point sources. The model has been used extensively to support
National Pollutant Discharge Elimination System (NPDES) wastewater discharge permit
applications, total maximum daily load (TMDL) studies, and environmental impact statements
for proposed development. A wide range of chemical and biological pollutants within a river can
be modeled, including carbonaceous biochemical oxygen demand, nitrogen and phosphorus,
suspended solids, algae, pathogens, phytoplankton and detritus. Physical-chemical processes
simulated by the model include water quality kinetics, chemical equilibrium, advection,
dispersion, settling, and interactions with the atmosphere (re-aeration) and riverbed (sediment
oxygen demand). Water quality parameters predicted throughout the modeled river domain
include dissolved oxygen concentration, pH, salinity and temperature, in addition to the various
pollutant quantities.
4.2.1.3 Other Surface Water Models
This section describes surface water models that appear to have potential utility for modeling
ENMs but that were not developed specifically for that purpose.
Packman et al. (2000) developed a process-based model to simulate the transport of colloids in
surface water systems, including the mass exchange between the water column and underlying
sediments. The model accounts for particle settling and sedimentation processes as well as
pumping exchange of parti culates due to water flow through sediment bedforms induced by
stream flow. Their formulation also considered particulate filtration in the porous bed sediments.
Using their model, solute and colloid exchanges may be predicted without fitting coefficients and
only requiring measurable hydraulic and particle parameters as inputs. One limitation of the
model is that it does not account for changes in the particulate suspension (e.g., due to time
varying pH or ionic strength). The Packman et al. (2000) model was used by Boncagni et al.
(2009) to simulate the behavior of titanium dioxide nanoparticles in surface water systems (see
Section 4.2.1.1). Appendix B provides a more detailed review of the model.
54
-------
4.0 Model Reviews
Box 4-2. Subsurface Colloid Transport
Modeling
Colloid transport is often modeled using an
advection-dispersion equation modified to account
for colloid attachment and detachment. The most
common approach relies on CFT (also called
clean bed filtration theory). This formulation
accounts for first-order kinetic attachment and
assumes that detachment is negligible (Tosco et
al., 2009). However, experiments have shown that
CFT theory is not always valid. In addition to
detachment, CFT does not account for so-called
blocking effects (also known as ripening), whereby
a maximum, threshold concentration of colloids is
able to attach to the solid. Such blocking behavior
may be described using a Langmuirian isotherm
approach. Other mechanisms not considered in
CFT include straining and enhanced transport (see
Section 2.3.4). In addition, CFT approaches do
not typically consider the potentially strong effects
of solution chemistry on colloid
attachment/detachment behavior (as described by
DLVO theory and its extensions). Some models
have extended CFT in order to account for some
of these additional processes.
4.2.2 Subsurface Models
As discussed in Section 2.4.2, many traditional
environmental fate models of chemicals in
terrestrial systems consider some or all of the
following important processes: dissolution,
volatilization, adsorption to organic and inorganic
matter, biological uptake, photolysis, hydrolysis,
and biodegradation. Less common are
environmental fate models that consider
processes of aggregation, attachment, and
sedimentation, all processes critical to
understanding and predicting the environmental
fate of ENMs. Models predicting ENM behavior
in the terrestrial environments should account for
these critical processes related to particulates in
natural systems. The evaluations of subsurface
fate and transport models in this section primarily
consider whether models account for these key
processes. Some established modeling
approaches show promise for modeling
subsurface ENMs, particularly colloid transport
models (US EPA, 2007). Box 4-2 provides some
additional background information about
subsurface colloid transport modeling.
4.2.2.1 Subsurface Models of Nanomaterials
This section describes subsurface models identified in the literature that have been developed
and/or applied specifically for the evaluation of ENM fate in the environment. Appendix B
provides more detailed reviews of each of these models.
Tosco and Sethi (2009) developed a one-dimensional model called MNM1D (micro and
nanoparticle transport model in porous media in ID geometry). The model considers constant or
transient hydrochemical parameters (ionic strength) and describes attachment and detachment
phenomena. The model accounts for multiple attachment sites, one based on linear and another
based on Langmuirian isotherms (thus accounting for blocking effects as described in Box 4-2).
The governing partial differential equations were solved using a finite difference solution, and
the model was validated through comparison with other models (HYDRUS ID and Stanmod).
The model was developed in Matlab and may be downloaded from the website
www.polito.it/groundwater/software. Appendix B provides a more detailed review of the Tosco
and Sethi (2009) model.
Ju and Fan (2009) developed a nanoparticle transport model for use in enhanced oil recovery
(EOR) applications. The proposed EOR approach involves the injection of poly silicon
nanoparticles to change the solid matrix from an oil wet to a water wet system. The model
55
-------
4.0 Model Reviews
includes both oil and water phases and accounts for blocking as well as permeability reduction.
Thus, the model accounts for complex processes (e.g., multiphase flow, permeability reduction
from particle straining). However, the model is one dimensional. Given that the purpose of this
model is to support enhanced oil recovery involving multiphase flow processes, we have not
performed a more detailed review of this model.
Li et al. (2008) developed a model to evaluate the transport of fullerene (C60) nanoparticles.
This model accounts for nonequilibrium attachment kinetics and maximum retention capacity
(site blocking). The authors developed a correlation for the maximum retention capacity
allowing prediction based on flow velocity, nanoparticle size, and mean grain size of the porous
medium. The authors determined that patch-wise surface charge heterogeneity on the sand grains
is probably the reason that observations deviated from classical DLVO theory. They concluded
that modifications to clean-bed filtration theory and accounting for surface heterogeneity are
necessary to predict nC60 transport behavior in saturated porous media. Appendix B provides a
more detailed review of the Li et al. (2008) model.
Liu et al. (2009) modeled experimental transport results for engineered multiwalled carbon
nanotubes (MWCNTs) using a one-dimensional model. This model includes a new theoretical
collector efficiency relationship to describe colloid attachment behavior. The model is based on
traditional colloid filtration theory (CFT) modified with a site-blocking term. The model
provided good agreement with experimental results. Appendix B provides a more detailed
review of the Liu et al. (2009) model.
Cullen et al. (2010) simulated the transport of nano-fullerenes (C60) and MWCNTs using a two-
dimensional finite element model. The model considered heterogeneity in permeability. The
model is based on classical CFT modified with a maximum retention capacity term. Their results
indicated that carbon nanotubes are more mobile than C60. This study utilized the commercially
available model, COMSOL Multiphysics version 3.4a. Their results show that nanoparticle
transport and maximum concentrations are very sensitive to collision efficiency factors and
blocking factors (parameters controlling colloidal attachment). As these authors emphasized,
accurate methods to predict these parameters from soil and nanoparticle characteristics have not
been developed, especially for natural environmental conditions. Appendix B provides a more
detailed review of the Cullen et al. (2010) model.
4.2.2.2 Regulatory Subsurface Models
This section discusses several established subsurface models that have been used in risk
assessments to support EPA regulatory programs. The models evaluated correspond to those
listed in NRC's review of EPA regulatory modeling practice (NRC, 2007). Although these
models represent only a subset of the available subsurface models used by EPA, they are among
the most widely applied, and they are representative of typical risk assessment modeling practice
by the agency.
None of the three evaluated regulatory models accounts for key processes of ENM subsurface
transport, including aggregation, attachment, and porous media filtering. The models therefore
56
-------
4.0 Model Reviews
are not suitable in their present form for evaluating ENM transport. We have thus provided only
brief reviews of these models below.
It is worth mentioning that some researchers have simulated colloid facilitated transport (see
Section 2.3.4) using conventional porous media transport models (such as the three regulatory
models reviewed in this section). For example, Contardi et al. (2001) and Vilks et al. (1998)
utilized a transport model accounting only for advection, dispersion, sorption, and decay. They
recognized that the model did not account for colloid behavior explicitly; however they utilized
an approximate approach to decrease the degree of sorption in order to account for colloid
facilitated transport. This modeling approach essentially utilizes a lumped-parameter approach to
simulate complex processes (see Box 4-1). If such an approximate approach is effective for
simulating ENMs, the regulatory models described in this section may have potential use.
However, given the unique and complex behaviors exhibited by ENMs in porous media, it seems
unlikely that this approach would provide reliable predictions for a broad range of conditions.
The PRZM modeling package couples a model of pesticide and nitrogen fate in the crop root
zone with a variably saturated flow and transport model of the deeper unsaturated zone. The one-
dimensional root zone model is solved using a finite difference approach (formulated from
multiple homogeneous compartments in series). The deeper unsaturated zone model is based on
a finite-element solution of Richard's equation for flow and an advection-dispersion equation for
transport. PRZM accounts for processes of advection, dispersion, sorption, biodegradation
(including up to two degradation products). The model also simulates surface runoff and
sediment erosion, including the transport of contaminants sorbed to sediments. The model
includes a Monte Carlo pre-and post-processor that supports probabilistic simulations. PRZM
has been used extensively to evaluate the fate of pesticides in agricultural settings. Other than
surface runoff the model does not include processes specific to particulate transport such as
aggregation, attachment, and porous media filtering. Therefore, PRZM is not suitable in its
current form to simulate the behavior of ENMs.
MODFLOW is a modular three-dimensional finite-difference groundwater flow model
developed by the U.S. Geological Survey and first published in 1984. MODFLOW is one of the
most widely used groundwater flow and transport models. Although the original version of the
model only considered groundwater flow, MODFLOW's modular structure has allowed
integration of many additional capabilities. The MODFLOW modeling system now includes
capabilities to simulate coupled groundwater/surface-water systems, solute transport, variable-
density and unsaturated-zone flow, aquifer-system compaction and land subsidence, parameter
estimation, and groundwater management. The model is based on a finite difference numerical
solution to the groundwater flow and transport equations. The model does not include processes
specific to particulate transport such as aggregation, attachment, and porous media filtering.
Therefore, MODFLOW is not suitable in its current form to simulate the behavior of ENMs in
the environment.
BIOPLUME is a two-dimensional finite difference model utilized to simulate processes of
natural attenuation of organic contaminants in ground water. Attenuation processes considered
include advection, dispersion, sorption, and biodegradation. BIOPLUME was developed from
the U.S. Geological Survey solute transport model MOC. The model considers the fate and
57
-------
4.0 Model Reviews
transport of the contaminant as well as several aerobic and anaerobic electron acceptors,
including oxygen, nitrate, sulfate, iron (III), and carbon dioxide. Three kinetic formulations are
available to simulate biodegradation reactions, including first-order decay, instantaneous
reaction, and Monod kinetics. BIOPLUME does not include processes specific to particulate
transport such as aggregation, attachment, and porous media filtering. Therefore, BIOPLUME is
not suitable in its current form to simulate the behavior of ENMs in the environment.
4.2.2.3 Other Subsurface Models
This section describes subsurface models that appear to have potential use for modeling ENMs
but that were not developed specifically for that purpose. Several researchers have developed
modeling approaches for simulating colloid transport in porous media. We have provided a brief
summary of several of these studies below. More detailed reviews are provided in Appendix B
for the established models TOUGH2 and HYDRUS.
Corapcioglu and Choi (1996) developed a one-dimensional model describing colloid transport in
unsaturated porous media with four phases (aqueous, air, solid matrix, and colloid). They
concluded that the air-water interface could strongly limit colloid transport due to colloid
attachment to the air-water interface. Johnson et al. (2007) incorporated geochemical
heterogeneity and random sequential deposition dynamics. Sun et al. (2001) developed a two-
dimensional colloid transport model for heterogeneous porous media. Ryan et al. (1999)
considered the importance of the geochemical environment on colloid attachment/detachment
behavior by developing a two-dimensional model accounting for physical and geochemical
heterogeneity. Bradford and Toride (2007) attempted to account for non-CFT behavior using a
conventional advection-dispersion equation (ADE) model with first order kinetic deposition and
release; they allowed some parameters to vary stochastically in successfully simulating
experimental results. Bekhit and Hassan (2005) developed a 2D colloid transport model
accounting for potentially facilitated and retarded colloid transport.
Moridis et al. (2003) utilized the TOUGH2 (Pruess, 1991) model to develop three-dimensional
simulations of a proposed nuclear waste disposal facility and associated colloid transport. This
effort accounted for colloid transport using the EOS9nT module (Moridis et al., 1999).
Appendix B provides a more detailed review of the EOS9nT module of TOUGH2.
HYDRUS is a software package for simulating water, heat, and solute movement in two- and
three-dimensional variably saturated media. The model includes several optional mechanisms of
colloid transport as documented in Simunek et al. (2006). Appendix B provides a more detailed
review of the HYDRUS model.
4.2.3 Bioaccumulation Models
Although the focus of this state-of-the-science review was on fate and transport models, we
recognize that the bioaccumulation of ENMs may, for some materials, represent a significant
exposure pathway. Therefore, we include this short summary of bioaccumulation models for
completeness, focusing primarily on organic chemicals. It should be noted that there is a
considerable body of research and attendant models available to estimate tissue concentrations of
58
-------
4.0 Model Reviews
organic chemicals, and the past few years have seen considerable advances in our ability to
predict uptake and accumulation of metals.
There are several types of mathematical modeling approaches that have been developed and used
in predicting exposure concentrations in biota, especially in aquatic systems. These approaches
can be classified as (1) quantitative structure-activity models, (2) mass balance models, and (3)
food web bioaccumulation models. The distinction among these models are certainly blurred
because QSAR elements are found—either explicit or imbedded—in all bioaccumulation models
and, similarly, kinetics (i.e., absorption, distribution, metabolism, and elimination [ADME]) tend
to be represented in most models, often using the log of the octanol-water partition coefficient as
a surrogate; nevertheless, it is useful to organize the types of predictive bioaccumulation models
into these categories.
Based on this review, no models or modeling approaches were identified that (1) have been
applied to ENMs or, (2) because of their theoretical underpinnings, could readily be used to
predict the bioaccumulation of ENMs in aquatic biota. Therefore, this section summarizes each
approach and provides a brief discussion of the potential relevance and applicability to ENM
bioaccumulation. Given the plethora of published research on the development and validation of
models to predict the uptake and accumulation of conventional chemicals, we identified a
handful of articles and reports that provide an excellent overview of methods as well as the
uncertainties associated with predictive bioaccumulation models. The primary sources of
information identified for this review were
• Bioaccumulation Assessment Using Predictive Approaches (Nichols et al., 2009)
• Uncertainties in ecological, chemical, and physiological parameters of a bioaccumulation
model: Implications for internal concentrations and tissue-based risk quotients
(DeLaender et al., 2010)
• Evaluation of Chemical Bioaccumulation Models for Aquatic Ecosystems - Final Report
(Aqua Terra, 2004).
QSAR Models. The earliest approaches to predict chemical concentrations in aquatic organism
is based on the relationship between an organism's BCF4 and the log of the w-octanol-water
partition coefficient (log Kow). As pointed out by Nichols et al. (2009), a curvilinear relationship
is obtained when plotting the BCF versus the log Kow up to a log Kow value of roughly 6. The use
of log Kow as a predictor of bioconcentration potential is based on the behavior of hydrophobic
organic chemicals, namely, the partitioning of hydrophobic organics into the lipid tissue of
animals. Because w-octanol tends to be a useful surrogate for lipid, this approach has proven to
be very useful within the range of chemicals for which uptake across the gills (versus uptake
through the food web) tends to be the driving exposure route. More recent development of
QSAR algorithms adjusts the baseline (i.e., the BCF based strictly on the log Kow) by chemical-
specific attributes such as ionizability. This same type of approach has proven to be extremely
useful in predicting toxicity for chemical classes with similar modes of action (MOA) such as
organic chemicals that cause adverse effects via narcosis (see, for example, Netzeva et al., 2007).
4 For the purposes of this discussion, the BCF is defined simply as the ratio of the chemical concentration in fish per
unit mass over the chemical concentration in water per unit volume.
59
-------
4.0 Model Reviews
However, because these QSAR methods are often derived from empirical studies, and because
they rely heavily on the assumption that w-octanol is an appropriate surrogate for lipid (implying
that the chemical partitions to lipid preferentially), these methods are unlikely to support
predictions for ENMs without considerable research demonstrating how ENMs partition from
the gut into other tissues following exposure. It is reasonable to assume that the mechanism for
partition for many ENMs (certainly quantum dots) is very different than the mechanism for
conventional organic chemicals and, therefore, QSAR approaches based on the log KOW are
unlikely to produce reliable predictions without extensive study into the actual mechanisms that
drive partition of ENMs.
Mass Balance Models. These types of models predict bioaccumulation in various body parts, in
essence, by representing processes associated with chemical uptake (e.g., the amount of water
that passes across the gill) and elimination (e.g., the dilution of chemical mass associated with
growth of the animal). The model conceptualizes the animal as one or more compartments (or
boxes) and the concentration of chemical in each box is a function of the processes that affect the
throughput of the chemical mass (Aqua Terra, 2004). In this engineering type approach, some
mass of chemical enters the box (e.g., parent compound that is not biotransformed), some mass
of chemical remains in the box (i.e., accumulation), and some mass of chemical leaves the box
(i.e., elimination). Most of these approaches are developed to solve the equation for steady-state
conditions and, therefore, supporting studies must demonstrate that steady state has been
achieved to provide reliable data for model validation. Naturally, the development of suitable
study data must address ADME and, as Nichols et al. (2009) point out, there is already a need to
improve the representation of ADME processes for conventional chemicals. The authors point
out that metabolism has long been a significant source of uncertainty for hydrophobic chemicals
(consider the importance of metabolism in predicting the tissue concentrations of poly cyclic
aromatic hydrocarbons [PAHs]). Given the hydrophobicity of certain classes of ENMs (e.g., low
solubility of fullerenes), significant research may be required before a mass balance approach
may be applied reliably to ENMs. However, mass balance approaches may be developed that
simplify the ADME paradigm by eliminating processes that are not relevant to ENMs (e.g.,
certain elimination mechanisms may not be relevant for ENMs that bind strongly to cellular
proteins).
Food Web Bioaccumulation Models. Whereas mass balance models are designed to predict
uptake and accumulation from water only exposures, food web models account for the exposure
to chemicals via water passing across the gills as well as through the diet (i.e., consuming prey
species that have accumulated some level of the chemical). These types of models can be limited
to aquatic food webs or they can be extended to terrestrial organisms that consume aquatic
animals. The primary difference between the mass balance class of models and food web models
is that food web models consider (and solve for, mathematically) multiple trophic levels
simultaneously. For chemicals that are efficiently metabolized by lower trophic level organisms
the predicted tissue concentrations may approximate the predictions generated using a mass
balance approach because water is the dominant exposure pathway (i.e., the chemical may not
biomagnify up the food chain). In contrast, if dietary exposure to a chemical is dominant, food
web models will produce a much higher estimate of tissue concentrations than a mass balance
model that only considers water only exposures. As suggested above, the food web class of
60
-------
4.0 Model Reviews
models is also based on kinetics principles (i.e., ADME), and require suitable studies to measure
parameters related to the chemistry (e.g., log Kow), ecology (e.g., nonlipid organic carbon in
detritus), and physiology (e.g., diffusion resistance for uptake) to produce useful predictions (De
Laender et al., 2010). This strongly suggests that, for food web models to be considered for
ENMs, substantial research would be needed across all trophic levels.
It should be noted that, although these classes of bioaccumulation models may not be
immediately useful in predicting tissue concentrations in aquatic ecosystems for the purposes of
exposure assessment, the theoretical bases of these approaches have been developed and
validated over many years and, in general, model performance has been considered appropriate
to support risk management decisions. A variety of models have been created and used for both
organic chemicals and metals (e.g., EPA's Bioaccumulation and Aquatic System Simulator, or
BASS, Barber, 2008) and, therefore, there is significant potential for further development of
these concepts to represent processes that are specific to ENM behavior in aquatic ecosystems.
4.2.4 Multimedia Models
Multimedia models treat various environmental media (e.g., surface water, groundwater and
atmosphere) as an integrated system, synthesizing information about chemical partitioning,
reaction, and intermedia transport. Multimedia models have been used to estimate regional and
global contaminant migration based on mass balance relationships (Fenner et al., 2005).
Multimedia models have also been used to assess transport at more local scales, including risk
assessments of point contamination sources (e.g., industrial sources of hazardous waste). This
section reviews several multimedia models that have been developed specifically to evaluate
ENMs (Section 4.2.4.1). In addition, several multimedia modeling frameworks established
within the risk assessment community are discussed (Section 4.2.4.2).
Many multimedia models are compartmental models based on a mass balance formulation. Such
models estimate the transport of material through (often homogeneous) compartments during the
life cycle of a chemical and may include the following steps: (1) characterize the source and
production volumes of material (compounds or chemicals); (2) estimate the emissions of material
to environmental compartments (air, sediment, soil, surface water, etc.); (3) specify the fate in
the environment; and (4) derive distributions of predicted environmental concentrations (PECs)
and predicted no effect environmental concentrations (PNECs) for the studied material.
One approach to calculate PECs and PNECs is material flow analysis (MFA) (also known
substance flow analysis, SFA), which is a method of analyzing the flows of material or substance
in a well-defined system. Generally, the goal of a MFA is to obtain an understanding of the
material flows, calculate indicators, and develop strategies and measures for improving the
material flow system. MFA can be used to determine flows to and amounts of materials within
the studied environmental compartments. It is also possible to extend MFA into a probabilistic
material flow analysis (PMFA), in which the goal is to derive probability distributions of PECs.
The PMFA is designed to calculate concentrations of possible contaminants in environmental
compartments and life stages associated with these contaminants.
61
-------
4.0 Model Reviews
The goal of many multimedia modeling analyses is to estimate PECs of potential hazards as well
as PNECs such that the risk quotient (PECs/PNECs) can be calculated. From this risk quotient,
risk managers can determine which chemicals are at greater risk (typically a risk quotient >1).
Sensitivity and uncertainty analysis (when available) can also be beneficial in the development
of intervention strategies for the chemicals associated with higher risk.
4.2.4.1 Multimedia Models ofNanomaterials
This section describes multimedia models identified in the literature that have been developed
and/or applied specifically for the evaluation of ENM fate in the environment. Each of the
models will be further reviewed in Appendix B.
Boxall et al. (2007a) developed a deterministic model by deriving dilution equations predicting
the environmental concentrations of ENMs in surface water, sludge, and soil. This model
determines the PECs of specific ENMs after a life cycle that includes production, use, emission,
and disposal. The model uses estimations and data values for parameters such as concentration of
the ENM within the product, daily usage of the product, fraction of the ENM removed during
sewage treatment, and sludge application rates. Uncertainty is introduced into the model when
data is not available for some of the necessary parameters. Uncertainty is also evaluated by
allowing certain parameters (e.g. concentration of ENM within the product) to vary to calculate a
range of PECs.
Blaser et al. (2008) modeled the emissions of silver (Ag) from biocidal products that held nano-
silver. The model was designed to estimate the emissions of silver and analyze the mass flow as
a result of emission, assess the fate and estimate the PECs of silver in a river system, estimate the
PNECs through critical evaluation of available toxicity data for environmentally relevant forms
of silver, and characterize the risk. Many simplifying assumptions such as neglecting emissions
from production or solid waste, as well as the removal of marine environments from the system
provide a simple model that may not encapsulate the characteristics of the broader life cycle of
ENMs.
Mueller and Nowack (2008) present a model intended to address the quantities of engineered
ENMs released into the environment from a life-cycle perspective. Using material flow analysis,
three types of nanoparticles were studied: nano-silver, nano-titanium oxide, and carbon
nanotubes. The model incorporated estimated worldwide production, particle release from
products, and flow coefficients within the compartments selected for the model. The different
life cycles of the three products generated varied results for the PECs. These generated PECs
were then compared to the PNECs specific to each material in order to estimate potential risk.
Gottschalk et al. (2010a) developed a probabilistic material flow analysis (PMFA) to calculate
distributions PECs and PNECs in a system comprised of 11 compartments. The paper used
PMFA specifically to address the lack of data concerning environmental fate, exposure,
emission, and transmission characteristics of ENMs. This stochastic approach allows the model
to represent uncertainties based on estimated input parameters. The authors propose that the use
of Monte Carlo simulations and Markov Chain Monte Carlo modeling is appropriate to estimate
PECs when faced with limited of data.
62
-------
4.0 Model Reviews
4.2.4.2 Regulatory Multimedia Models
The six multimedia risk assessment models described in this section represent current, accepted
approaches for multimedia modeling within the regulatory community. None of these models
provides a comprehensive solution for estimating the fate of ENMs in the environment. Many of
the associated sub-models do not account for key processes of particulate transport in the
environment such as aggregation, attachment, settling, and porous media filtering. In addition,
these traditional multimedia models are strongly reliant on chemical property estimation tools
(e.g., QSARs) that were developed for chemicals other than ENMs (see Section 2.3.10). Some of
the multimedia modeling frameworks are highly abstracted and only describe mass transfer
between environmental compartments using simple mass transfer functions rather than
mechanistic formulations that account explicitly for underlying processes. Such highly abstracted
multimedia models may be useful for screening type evaluations of ENM transport; indeed,
multimedia models specific to ENMs discussed in the previous section fall within this category.
However, parameterization of such models generally will require knowledge (e.g., empirical
data) that is currently unavailable for ENMs. Given these significant limitations, we have only
provided brief summary descriptions of the traditional multimedia models described below.
FRAMES (Framework for Risk Analysis in Multimedia Environmental Systems and subsequent
versions) and MIMS (the Multimedia Integrated Modeling System) are two multi-agency
software frameworks regarded as the best available in United States for multimedia risk
assessment. These two frameworks borrow concepts and codes from other frameworks, such as
STELLA and DIAS. The most notable use of FRAMES is the integration of 17 scientific models
in the 3MRA model to accomplish multimedia, multi-pathway, and multi-receptor risk
assessment. MIMS is an object-oriented framework, especially suitable for models with different
spatial and temporal scales. Its conceptual design supports interchanging models and data sets
(and modeling of physical, chemical, biological, and human systems), cross-platform portability
to support off-the-shelf models and distributed computing. MIMS differs from FRAMES in that
it provides mechanisms to allow feedbacks between models (i.e., dynamically coupled system).
The 3MRA Modeling System is a suite of 17 environmental risk assessment modules originally
designed to support the Hazardous Waste Identification Rule (HWTR). It uses FRAMES to allow
integration of these varied modules and data. This model has been peer-reviewed by EPA's
Science Advisory Board (SAB) and is currently supported by ongoing activities at ORD to
develop automated systems to populate the extensive databases required to run simulations.
TRIM (Total Risk Integrated Methodology) is an elaborate collection of multiple models (e.g.,
fugacity-based models, simple air quality models, human exposure models) developed to
perform deterministic multimedia health and ecological risk assessments for hazardous air
pollutants. It uses the MIMS framework for integration of the risk assessment process. TRIM
provides risk metrics tables that can be used to further analyze, interpret, and visualize the
results, (http://www.epa.gov/ttn/fera/trim_gen.html)
The RESRAD (Residual Radioactivity Models) family of codes is a comprehensive set of
components that allow probabilistic multimedia risk assessment that are fully interoperable. The
codes support multimedia modeling and provide capabilities for sensitivity and uncertainty
63
-------
4.0 Model Reviews
analysis. RESRAD uses an OpenLink software framework for integration of environmental risk
analysis and management. In addition, the technical support for RESRAD is extremely good.
RESRAD uses as low as reasonably achievable (ALARA) analysis or a cost-benefit analysis that
can help in the cleanup decision-making process. The code is supported by Argonne National
Labs and frequently updated to enhance functionality; for example, recent updates include non-
radioactive chemicals, (http ://web.ead.anl.gov/resrad/home2/)
CalTOX (California Total Exposure Model for Hazardous Waste Sites) is an Excel-based
fugacity model for multimedia risk assessment. Due to the simplicity of the model, its results
have been incorporated into life-cycle assessment models (e.g., TRACT). The CalTOX model has
been used primarily for the assessment of contaminated soils as the primary source of
contamination; however, it has been adapted for multiple purposes, including the support of risk
ranking schemes and life cycle assessments. The California Exposure Modeling Research Center
at Berkeley has an active program that involves the continuing development of this model, in
part, to run nested spatial scale calculations, (http://eetd.Ibl.gov/ie/ERA/)
ARAMS (Army Risk Assessment Modeling System) is a multimedia risk assessment tool that
specially addresses human health and the ecological risks associated with military relevant
compounds (MRCs); however, it is applicable to any setting with contaminated sources or
media. ARAMS uses FRAMES to integrate environmental models and databases. ARAMS
considers temporal and spatial distribution of contaminants and lends itself to sensitivity and
uncertainty analyses. ARAMS has functional links to multiple existing databases, such as the
Integrated Risk Information System (IRIS), Health Effects Assessment Summary Table
(HEAST), Environmental Residue Effects Database (ERED), and BSAF
(http://el.erdc.usace.army.mil/arams/).
SADA (Spatial Analysis Decision Assistance) is unique among all the other risk assessment
models presented because it is a decision analysis and support tool for risk assessment. SADA
combines risk assessment with geographic information systems (GIS) and statistical analysis
methods and sampling design to determine remedial design and cost-benefit analysis. Use of GIS
provides the capability to explore data that is spatially distributed
(http://www.tiem.utk.edu/~sada/index.shtml).
4.3 Alternative Approaches
It is apparent that risk assessment of ENMs will depend critically and sensitively on the issues
and uncertainties surrounding their fate and transport in the environment (Wiesner et al., 2009).
Because of this, we must address the question: How can environmental behavior and risk be
characterized for an emerging technology? The logical first step might be to modify a
conventional risk assessment to incorporate the environmental interactions and properties
relevant to ENMs. However, traditional risk assessment modeling will introduce substantial and
unquantifiable uncertainties due to paucity of data surrounding the persistence of
environmentally relevant forms of ENMs. Thus, while these models may provide some insight
into the complex systems surrounding ENM transport in the environment, they offer limited
guidance as to the actual potential for adverse health and environmental effects. This creates a
need to develop models and approaches that can explicitly address the uncertainties surrounding
64
-------
4.0 Model Reviews
these complexities and behaviors of ENMs in the environment, and provide meaningful
information to risk managers. Therefore, we must rethink the existing assessment paradigms with
respect to the nature of ENMs and their transformations, biological interactions, and
environmental transport so that effective risk management can be developed for ENMs.
Other authors have supported this position (Linkov et al., 2009a). For example, Grieger et al.
(2009) argue that although conventional risk assessments are needed for responsible
development, the process may take decades, leaving decision makers with little support in the
near term. Given the immediate demands placed on decision makers, we must design more
adaptive risk governance frameworks and alternative methods to support the characterization of
potential risks associated with ENMs released into the environment (Hansen, 2009). Lowry and
Gasman (2009) also stress the need for developing new frameworks to describe the potential
risks of ENMs in the environment. They suggest integration of laboratory results into risk
analytic frameworks such that preliminary risk analyses can prioritize and identify the most
relevant data gaps needed to aid in traditional risk assessment. Thus, with the current limitations
of traditional risk assessment and the future impact on every aspect of our lives and society that
nanotechnology is expected to have, this state-of-the-science review has incorporated alternate
approaches to more traditional fate and transport models including5:
• Adaptive management and evaluation frameworks
• MCDA
• Bayesian approaches.
The above approaches can be used for relatively near-term decision making for exposure to
ENMs and will be discussed in the rest of this section. Appendix B provides more detailed
reviews of each of these approaches. Table 4-1 summarizes advantages and limitations of each
of these alternative approaches.
Table 4-1. Summary Evaluation of Alternative Approaches to ENM Risk Evaluation
Advantages Limitations
Adaptive • flexible and adaptable because of the • limited quantitative data
Management and acceptance of available data . |ack£ thorough testjng and va|idatjon
Evaluation . explores atypical pathways of exposure
Frameworks
5 Two additional alternative approaches were identified in the search process for this report: Precautionary matrices
(PM) and Value of Information (Vol). PM is a simple scoring tool designed for use in early assessments of the
potential exposure risk of a substance to human health and the environment (Hock et al., 2008). The associated
safety matrix incorporates information about potential harmful effects, product life cycle, chemical properties, and
potential exposure routes in order to gain a general understanding of the risks that may arise from these substances.
The designers stress that while this tool cannot replace traditional risk assessments, these matrices can be used to
prioritize research needs for emerging technologies. Given the simplicity of the approach and the relatively narrow
applicability of the technique, we have not provided additional review information for PM. Vol is intended to
quantify the improvement in expected value from obtaining new information before making a decision and can
reveal methods to reduce risk or increase potential value. While Vol may be a valuable tool for risk management, its
potential application for evaluating ENM risk has yet to be documented. Therefore, the approach has not been
reviewed in this report.
65
-------
4.0 Model Reviews
Multicriteria
Decision Analysis
Bayesian
Networks
balances societal benefits against
unintended side effects and risks
combines multiple lines of evidence to
estimate the toxicity, risk, or exposure to
ENMs given limited information on physical
and chemical properties
scientifically sound decision analytical
framework
ranks or groups all the alternatives through
a structured process rather than suggesting
a single replacement
generally robust to imperfect knowledge
easily updated/modified as new scientifically
relevant information becomes available
provides optimal decisions based on the
parameters assigned to the model
• does not predict environmental fate
• interpretation of MCDA is subject to parameter
definitions defined by the user
• outcomes may depend on the decision maker
which can be influenced by personal goals
and preferences
some networks can be too large and complex
for current Bayesian algorithms
4.3.1 Adaptive Management and Evaluation Frameworks
An adaptive evaluation framework, which is a form of adaptive management, is an alternative
method that can be used to resolve challenges in modeling ENMs. Adaptive management is an
atypical environmental management method in which the process involves: (1) Setting goals and
management objectives; (2) development of a model of the system being managed;
(3) development of a range of management choices; (4) monitoring and evaluating outcomes of
management decisions; and (5) development of a mechanism in which new information can be
incorporated into the system for future decisions (learning attribute). Along with this process,
adaptive management allows for, and encourages, revisiting and revising goals and objectives of
the project as well as a collaborative structure for stakeholder participation and learning (Linkov
et. al, 2006). Adaptive management can be divided into two approaches: passive and active. Both
follow the first three steps of the adaptive management process however passive management
studies only one alternative experiment at a time, while active management implements multiple
alternative strategies and examines the outcomes.
Following the above structure, adaptive evaluation frameworks can be used to circumvent the
lack of data needed for the traditional risk assessment of ENMs by identifying many exposure
potentials based on criteria typically not included in the risk assessment paradigm. Possible
added avenues of evaluation include the location of the ENM within a product (Hansen et al.,
2007), the product life cycle and potential release points, and anticipated volumes of production
(Metcalfe et al., 2009). Adaptive evaluation frameworks typically unite these parameters with
more common evaluation parameters such as basic physical and chemical information available
for the ENM.
It is important to note that because each model approach relies on differing parameters, there are
multiple techniques in which these types of specific frameworks can be developed. However,
most frameworks use a conceptual guideline in the early development of the framework to map
potential pathways of exposure. These pathways can be as simple or complex as the developer
chooses. For instance, the categorization framework presented by Hansen et al. (2008) considers
only consumer exposure to products containing ENMs, thus neglecting possible exposure from
many environmental factors. However, a more extensive example of a conceptual framework for
66
-------
4.0 Model Reviews
potential exposure can be shown in Figure 4-1, which is the exposure pathway conceptualized
by the SMARTEN (Metcalfe et al., 2009) technique for adaptive frameworks. This framework
utilizes information from the ENM manufacture as well as generalized environmental processes
and exposure pathways to make predictions about the environmental fate and effects of ENMs.
Once pathways are identified, adaptive management requires alternative solutions be explored in
order to produce lower risk exposure potentials. The decision maker takes a decision which is
then interpreted as a hypothesis that needs to be tested and validated. Validation could involve
monitoring exposure levels such as environmental surveillance (Metcalfe et al., 2009) or
evaluating potential exposures due to the location and concentration levels of ENMs within the
product (Hansen et al., 2008). The findings are evaluated to determine if the hypothesis is to be
confirmed or rejected. If rejected, a new hypothesis is generated and the process starts again.
Therefore, adaptive evaluation frameworks view the management of a risk as a process
consisting of many small decisions rather than a single decision (Hansen, 2009).
Because technologies are evolving that constantly generate new safety and health information,
adaptive evaluation frameworks must be able to accommodate new data so that the most accurate
risk assessment can be performed on emerging ENMs. This methodology allows risk assessors to
be more proactive in evaluating all aspects of the life cycle of an ENM, thus aiding in decisions
to produce lower risk ENM products.
Some of the emerging adaptive evaluation frameworks are: (a) Categorization frameworks
(Hansen et al., 2007); and (b) SMARTEN (Metcalfe et al., 2009), which reports a governance
framework for adaptive evaluation frameworks. Hansen et al. (2007) will be evaluated in detail
in Appendix B using the guidelines from Section 4.1.
67
-------
4.0 Model Reviews
Sources
2. Media and
Transport Processes
3. Exposure Pathways
and Receptors
Engineered
Nanostructures
Engineered
Nanodevices
Reaction
Intermediates
Production
Waste
Product
Degradation
Pioduct i
rtt€n(!»ire*l
'•lac
/
/
t'
/
\
\
as, Splia
application
-eat
Aggregation/ ' .
UV Degradalion .
Settling / suspension
^ _ Soii._
Degradation •
Sorption / desorption
1 Water
Aggregation/ •:
Degradation
Setting / suspension
* Sfrdjjnent
Degradation •
""^*,
^•^.
rff^'i' /"
^/
/
Mf A
^?^" ,^X;
/ /Ji
/ /V'
/^^ijonlact
"^1
^
^
/.
/ UTCltSCt A
: j- - > „ p|
"" " 1
nta!2iS2- — — " —
>-j Terrestrial
7 InvertebralBS
,-•** Contact,
/
/
s -,
*^;"
Aquatic
rwertebrates
Fish I—
Benthic
Invertebrates
\
_,s— -•H*
^
X
rnesBo
Terrsslrial ,^
Vertebrates
^
Inpssuon
\
iHesiic«ix^
•*-}
//
/
Fi
InqasllBn
3h
Figure 4-1. Conceptual exposure pathway utilized by the SMARTEN model.
4.3.2 Multi-Criteria Decision Analysis (MCDA)
This section presents a basic overview of MCDA techniques, including characteristics shared by
different approaches. Within MCDA, almost all methodologies share similar steps of
organization and decision matrix construction, but each methodology synthesizes information
differently (Yoe, 2002; Figueira et al., 2005; Belton and Stewart, 2002). Different methods
require diverse types of value information and follow various optimization algorithms. Some
techniques rank options, some identify a single optimal alternative, some provide an incomplete
ranking, and others differentiate between acceptable and unacceptable alternatives.
The MCDA methods are classified into two general categories of elementary methods and multi-
objective methods that are considered more sophisticated and briefly discussed in the following.
4.3.2.1 Elementary Methods
Elementary MCDA methods (e.g., maximum method, conjunctive method, lexicographic
method, and TOPSIS method) can be used to reduce complex problems to a singular basis for
selection of a preferred alternative. However, these methods do not necessarily weight the
relative importance of criteria and combine the criteria to produce an aggregate score for each
alternative. While elementary approaches are simple and can, in most cases, be executed without
the help of computer software, these methods are best suited for single-decision maker problems
68
-------
4.0 Model Reviews
with few alternatives and criteria, a condition that is rarely characteristic of environmental
projects.
In the maximum method, each alternative is scored based on the performance of its weakest
attribute. The analogous maximax method scores each alternative based on the performance of
its strongest attribute. Comparison of the alternatives requires that all attributes be scored on
comparable scales.
The conjunctive method is designed to screen alternatives based on whether they exceed
minimum performance thresholds for all criteria. One useful application of the conjunctive
approach is to decrease a large number of alternatives to allow more detailed evaluation of a
subset. The conjunctive method does not require attributes to be scored on a common scale,
thereby limiting the effort needed for the analysis. In the analogous disjunctive method,
alternatives pass the screening test if they exceed the minimum performance threshold for at least
one attribute (as opposed to all attributes in the conjunctive method).
In the lexicographic method, the criteria are ordered in terms of importance. The alternative with
the best performance is the alternative with the strongest performance for the most important
criterion. If multiple alternatives are tied with respect to the most important criterion, these
alternatives are compared for the next criterion, and so on, until the highest performing
alternative is selected.
In the TOPSIS method (technique for order preference by similarity to ideal solution), the
selected alternative should be as close to the ideal as possible and as far from the negative ideal
as possible. The ideal is defined as a hypothetical alternative with the highest individual criteria
scores. The negative ideal is the combination of minimum scores.
4.3.2.2 Multi-objective Methods
Some of the main multi-objective decision analysis methods include Multi-Attribute Utility
Theory (MAUT), Multi-Attribute Value Theory (MAVT), Analytical Hierarchy Process (AHP),
and outranking. Table 4-1 summarizes important elements, strengths and weaknesses of these
methods (Linkov et al., 2007). The first three methods are more complex methods that use
optimization algorithms, whereas outranking uses a dominance approach. The optimization
approaches employ numerical scores to communicate the merit of each option on a single scale.
Scores are developed from the performance of alternatives with respect to individual criteria and
then aggregated into an overall score. Individual scores may be simply summed or averaged, or a
weighting mechanism can be used to favor some criteria more heavily than others. The goal of
MAUT is to find a simple expression for the net benefits of a decision. Through the use of utility
or value functions, MAUT transforms diverse criteria into one common scale of utility or value.
MAUT relies on the assumptions that the decision maker is rational (preferring more utility to
less utility, for example), that the decision maker has perfect knowledge, and that the decision
maker is consistent in his judgments. The goal of decision makers in this process is to maximize
utility or value. Because poor scores on criteria can be compensated for by high scores on other
criteria, MAUT is part of a group of MCDA techniques known as compensatory methods.
69
-------
4.0 Model Reviews
MAVT refers to decision analysis without formal uncertainty analysis, while MAUT refers to
methodologies that formally account for uncertainty. In the literature, MAVT is typically treated
as a subset of MAUT, and the more general term (MAUT) is more commonly used.
Similar to MAUT, AHP (Saaty, 1994) aggregates various facets of the decision problem using a
single optimization function known as the objective function. The goal of AHP is to select the
alternative that results in the greatest value of the objective function. Like MAUT, AHP is a
compensatory optimization approach. However, AHP uses a quantitative comparison method
that is based on pair-wise comparisons of decision criteria rather than utility and weighting
functions. All individual criteria must be paired against all others and the results compiled in
matrix form. For example, in examining the choices in the selection of an ENM, AHP would
require the decision maker to answer questions such as, "With respect to the selection of an
ENM, which is more important, its economic impacts or its environmental impacts?" The user
uses a numerical scale to compare the choices and AHP moves systematically through all pair-
wise comparisons of criteria and alternatives. AHP thus relies on the supposition that humans are
more capable of making relative judgments than absolute judgments. Consequently, the
rationality assumption in AHP is more relaxed than in MAUT.
Unlike MAUT and AHP, outranking is based on the principle that one alternative may have a
degree of dominance over another (ODPM, 2004). Dominance occurs when one option performs
better than another on at least one criterion and no worse than the other on all criteria (ODPM,
2004). However, outranking techniques do not presuppose that a single best alternative can be
identified. Outranking models compare the performance of two (or more) alternatives at a time,
initially in terms of each criterion, to identify the extent to which a preference for one over the
other can be asserted. Outranking techniques then aggregate the preference information across all
relevant criteria and seek to establish the strength of evidence favoring selection of one
alternative over another. For example, an outranking technique may entail favoring the
alternative that performs the best on the greatest number of criteria. Thus, outranking techniques
allow inferior performance on some criteria to be compensated for by superior performance on
others. They do not necessarily, however, take into account the magnitude of relative
underperformance in a criterion versus the magnitude of over-performance in another criterion.
Therefore, outranking models are known as partially compensatory. Outranking techniques are
most appropriate when criteria metrics are not easily aggregated, measurement scales vary over
wide ranges, and units are incommensurate or incomparable.
4.3.2.3 Recent Reports and Models
There are many different forms of MCDA available for use, such as the stochastic multicriteria
acceptability analysis (SMAA-TRI), AHP, and MAUT. Linkov et al. (2007) and Linkov et al.
(2009b) explore some of these techniques and how they can be utilized in ENM decision making.
These reports will be evaluated in detail in Appendix B using the guidelines from Section 4.1.
4.3.3 Bayesian Approaches
Bayesian Networks, or BayesNets, provide a framework for adaptable risk assessment that can
account for various types of uncertainty and may be easily updated/modified as new
70
-------
4.0 Model Reviews
scientifically relevant information becomes available. Much of this discussion is based upon
work at CEINT directed by Dr. Mark Wiesner. Bayesian approaches are a major focus of their
efforts to develop approaches for evaluating ENM behavior in the environment.
BayesNets are probabilistic networks. A network, or graph, provides a mathematical structure
composed of nodes (vertices) and edges. Edges join pairs of vertices and represent a pairwise
relationship between two nodes. Two nodes are said to be connected if a path of edges exists that
can be followed from one vertex to the other. Graphs can be undirected or directed. In a directed
graph, relationships move in one direction, whereby one vertex influences another but not vice-
versa. In an undirected graph, influence can occur in either direction. A network is probabilistic
if probabilities (also known as weights) are assigned to the edges. The weights represent the
likelihood of a relationship occurring between nodes.
Bayesian networks are based on directed, acyclic graphs representing a set of random variables
(nodes) and their conditional dependences (edges). For the modeling of ENM exposure to the
environment, a Bayesian network may represent the probabilistic relationship between
environmental media and the amount of ENM present in the system. Given an amount of ENM
produced, the network can be used to estimate the amount of ENM in specific environmental
media based on the likelihoods of material flow through the network.
The development of a Bayesian network offers two significant advantages: (1) because a
Bayesian network only connects nodes that are probabilistically related, an enormous
computational saving can result; and (2) Bayesian networks are extremely adaptable.
Traditionally, probabilistic models could lead to excessive numbers of potential states to be
solved, which could require impractical computational efforts. Bayesian networks offer a
solution to the computational challenges by limiting the possible combinations of states based on
probabilistic relationships. The adaptability of Bayesian networks lies in the fact that networks
can be expanded or modified as scientifically relevant information emerges.
Some have refrained from using BayesNets due to the belief that they will only work well if the
probabilities upon which they are based are exact. In actuality, approximate probabilities, even
those based on professional judgment, can provide very useful results. In other words, BayesNets
are generally robust to imperfect knowledge. Thus, the combination of several strands of
imperfect knowledge can still allow surprisingly strong conclusions.
Figure 4-2 displays a network developed by CEINT designed to predict environmental exposure
to ENMs. Network nodes represent ENM mass residing within the system, and edges represent
material flows through the system. The nodes may be either: (1) a source of ENM; or (2) an
environmental compartment in the system where ENMs may reside. Sources of ENMs may
include: initial production sources of raw ENMs (S); intermediate products containing ENMs (I);
and final products containing ENMs (P). For the CEINT model, the environmental compartments
included are atmosphere, wastewater treatment plant (WWTP), storage, landfill, effluent, sludge,
natural waters, and agricultural land. The directed edges indicate the flow of ENM from one
compartment (node) to another. In the case of edges from products to environmental
compartment, this represents the leakage (i.e., release of ENMs into the environment). Leakage
can be aggregated over environmental compartments or over specific stages of the value chain
71
-------
4.0 Model Reviews
(NM production, use, and transport as products are produced). For example, the directed edge
from S into the atmosphere compartment represents the potential loss of ENM from the raw
ENMs used for production into the immediate atmosphere. A probabilistic relationship for each
flow path (edge) must be defined for this to be a Bayesian network. The probabilistic relationship
assigned to each edge designates the fraction of ENM that moves from one node to the next. A
framework for describing ENM production and incorporation into products as well as leakage to
the environment can now be explored.
From this generalization (Figure 4-2), flows in this system can be characterized. Conceptually,
the description of all flows within this network represents a very high demand for information on
trends in commercialization, product use, product degradability, and ENM transformation and
transport. However, the framework provides the ability to aggregate across the value chain or
across receiving compartments such as wastewater, thereby reducing the number of unknowns at
the cost of loss of detail. For example, the amount of ENM entering the wastewater compartment
can be expressed as the product of the ENM source term and the sum of products of coefficients
representing all pertinent intermediate flows. This aggregation yields a single coefficient that
captures ENM production and use profiles relevant to the wastewater compartment. Though the
value of this coefficient may not be known initially, it may be estimated from measurements of
the quantities of ENMs in wastewater or from commercial projections and assumptions of use of
these products. Moreover, assumptions regarding the amount of ENMs entering wastewater are
made explicit through the specification of the coefficient and can be examined in what-if
scenarios. In this fashion the concentration of NPs that make their way into in wastewater sludge
and can be estimated by the above network using equations that will be available upon
completion of the work. Differences in production/usage profiles and the physical-chemical
characteristics of the ENMs determine their environmental fate with respect to wastewater
treatment residuals (sludge and treated water). Similar conceptual equations can be developed for
ENMs entering surface waters, landfills, and the atmosphere. Associated production/usage
profiles and transfer functions will be generated through the CEINT research.
The broad variety of materials made into NPs (e.g. metals, oxides, or carbon-based), the
technical difficulties associated with measuring NPs at low concentrations, and the added
complexity of detecting particles in the complex media that constitute natural waters, soils, and
air, present significant challenges to estimating potential exposures to ENMs. The Bayesian
network approach for ENM exposure assessment described above requires quantitative
relationships between the amounts of ENMs entering disposal and treatment systems and
environmental compartments. Thus, transfer functions must be developed for each environmental
compartment considering properties necessary for transport in the specified media.
72
-------
4.0 Model Reviews
Nano-Enafaled ProJ'.i' I
L I
Agricultural
Land
S]u,ls;v F.tlluent j
Natural
Waters
Figure 4-2. Conceptual network of ENM flows over value chain and into environmental
compartments.
73
-------
5.0 Conclusions
Chapter 5.0
Conclusions
In this section, conclusions are drawn from the state-of-the-science review of models and
methods relevant to the exposure assessment of ENMs released into the environment. These
conclusions are presented with respect to the five basic questions that this review was designed
to answer, for convenience, repeated below.
1. What models and approaches have been used successfully to simulate nanomaterial
behavior in environmental systems?
2. What models and approaches cannot be used to predict exposures to ENMs in
ecosystems?
3. What models and approaches can be used in the near term, and what types of
predictions can be supported by available models?
4. What techniques can be used to address uncertainties and support risk management
decisions in the near term given obvious gaps in information?
5. What does the state-of-the-science suggest with respect to long-term research goals
that can be undertaken to improve fate and transport modeling tools for ENMs?
The remainder of this section discusses each of these science questions with respect to results of
the state-of-the-science for exposure modeling of ENMs in the environment.
What models and approaches have been used successfully to simulate nanomaterial
behavior in environmental systems?
Fate and transport models that predict the behavior of ENMs in the environment must consider
particulate transport behaviors, including processes of aggregation and disaggregation,
attachment and detachment, settling and sedimentation, filtering and enhanced transport in
porous media, as well as particulate diffusion (see Section 2.3 for further discussion of these
processes related to ENMs). We identified several models (see Section 4) that account for many
of these particulate processes, including models developed specifically for ENMs, some
established regulatory models, and models originally developed for other purposes (e.g., colloid
transport models). However, even if models incorporate descriptions of key particulate-transport
behaviors, there remain significant knowledge gaps for ENMs. Reliable parameterization of such
models will not be possible until sufficient data (laboratory and field) are available. Therefore,
the utility of these models to predict ENM concentrations in the environment in the near term is
limited.
Several alternative approaches to traditional risk assessment that are rooted in risk management
and decision analysis offer the ability to evaluate ENMs risks in the short term without
necessarily requiring extensive data collection. Furthermore, these alternative approaches offer
distinct advantages in terms of integrating different types of information (e.g., expert knowledge
and scientific judgment) and account for uncertainty and incomplete knowledge. Section 4.3
provided an introduction to several promising alternative approaches, including adaptive
evaluation frameworks, MCDA, and Bayesian approaches. Although these alternative
approaches may not be sufficiently robust to support prescriptive regulatory requirements in the
74
-------
5.0 Conclusions
near term (e.g., setting allowable concentration limits), they may provide information on relative
risks, which can be useful in risk management. Risk ranking results could be used to prioritize
research studies (i.e., characterize those ENMs with the greatest potential risk) as well as to
prioritize regulatory initiatives such as voluntary agreements with industry to avoid potentially
high risk practices.
What models and approaches cannot be used to predict exposures to ENMs in ecosystems?
Models that do not consider critical particulate-transport behaviors and associated properties
show little promise for the evaluation of ENM transport. Many established regulatory models do
not account for key ENM behaviors (see Section 4). In addition, many established fate and
transport models are based on traditional equilibrium partitioning relationships. Existing
estimates of partition coefficients (e.g., those based on QSARs) are generally invalid for
materials at the nanoscale due primarily to the fact that ENMs exhibit properties of particles as
well as chemicals (see Section 2.3.10). Furthermore, enhanced partitioning models will likely be
required to predict ENM behaviors reliably (i.e., models that account for the distribution of mass
between solid, aqueous, as well as particulate phases and potentially nonequilibrium, kinetic
mass transfer).
It is important to recognize that environmental risk assessment often relies on low or moderate
complexity models that describe detailed transport mechanisms using relatively simple model
constructs (sometimes referred to as lumped parameter, conceptual models; see Box 4-1). Such
approaches are necessary (and appropriate) to simulate the fate and transport of conventional
chemicals, even though they are simplifications of natural systems that are variable and complex.
However, the use of such models involves uncertainty, as emphasized in the silver book, and it is
critical to characterize and, where possible, quantify the uncertainty in the risk estimates. It is as
yet unknown whether such lumped-parameter formulations will be appropriate for simulating the
behavior of ENMs in environmental systems or whether alternative modeling constructs will be
required. Regardless, lumped parameter models must be grounded in empirical evidence in order
to make reliable predictions (given their empirical or semi-empirical basis, they cannot make
reliable predictions a priori). Therefore, even if the underlying model constructs ultimately are
shown to be appropriate (i.e, model structural analysis), the current applicability of these models
for predicting ENM behavior is significantly limited by a lack of knowledge and lack of
available, empirical data to ensure reliable predictions.
What models and approaches can be used in the near term, and what types of predictions
are currently supported?
The alternative approaches described in Section 4.3 can be applied in the near term to evaluate
risks associated with nanomaterials released to the environment. Some of these approaches result
in a qualitative (or semi-quantitative) relative ranking of potential risks from specific ENMs. To
the extent that these approaches produce quantitative results (e.g., predicted environmental
concentrations), current gaps in our knowledge will create significant and possibly
unquantifiable uncertainties. Nevertheless, the relative risk results should provide important
insights for regulatory decision makers in the near term.
75
-------
5.0 Conclusions
What techniques can be used to address the uncertainties and support risk management
decisions in the near term given obvious gaps in information?
The alternative approaches described in Section 4.3 provide promising methods for near term
evaluation of ENM exposures. Several of these approaches address uncertainties explicitly. For
example, the BayesNet approach is a probability-based approach that produces probability
distributions for estimated exposures and effects. Some of the reviewed fate and transport models
(Section 4.2) show promise for ENMs, and many of these models may be implemented using
methods developed to represent variability and uncertainty (e.g., Monte Carlo analysis).
However, the extensive data gaps associated with ENMs limit the utility of fate and transport
models in the near term.
What does the state-of-the-science suggest with respect to long-term research goals that can
be undertaken to improve fate and transport modeling tools for ENMs?
In concluding this state-of-the-science report, we offer the following to inform the development
of an integrated research strategy for exposure modeling of ENMs:
• As emphasized in this report, there are significant data gaps in the understanding of
nanomaterial behaviors in the environment. Naturally, research should continue to
support the development of a basic understanding of the fundamental mechanisms
controlling fate and transport. This research should include: (1) empirical studies
(laboratory and field) to characterize ENM transport under a variety of natural conditions
and to develop parameters in support of ENM modeling; (2) the field testing of existing
models to develop insight into the magnitude of their current limitations; and (3) given
the importance of the partitioning approach in multimedia modeling, new modeling
approaches should be developed to replace or modify the partitioning approach used for
conventional organic chemicals.
• The prevalent data gaps in characterizing ENMs will severely limit the ability to predict
ENM transport using existing fate and transport models, even if these models account for
key particulate-transport processes associated with ENMs. It may require years to
develop sufficient knowledge and modeling expertise that support reliable predictions of
the environmental behavior of ENMs. Given these challenges, alternative approaches
(described in Section 4.3) can be used to support science-based risk management,
explicitly acknowledging uncertainties in the estimation of exposures. Therefore, we
recommend a parallel research track (along with fundamental fate and transport research)
that promotes decision analytic and/or adaptive management approaches that, ultimately,
can be linked to mechanistic fate and transport models/data under development.
• We recommend a state-of-the-science evaluation similar to the current report that is
focused on non-organic ENMs, most notably metals. The focus of the current report was
on organic-based ENMs, although many of the concepts discussed are more broadly
applicable.
• Multiple references emphasized the lack of consistency in reporting (and measuring)
ENM properties as well as ambiguity in nomenclature for this emerging field. In addition,
the large number of ENMs and their highly variable properties and behaviors suggest that
76
-------
5.0 Conclusions
different modeling and parameterization approaches will be required for different types
of ENMs. Therefore, we suggest a model-based classification system for ENMs that
captures differences and similarities in environmental behaviors and dependencies. Such
a classification system would (at a minimum) need to consider the chemical composition
of the base material (e.g., organic versus metal) as well as the composition of any surface
modification to the ENM. Development of a standard ENM data model that links fate and
transport modeling needs to basic research standards on ENM properties would provide a
more integrated approach to environmental modeling of ENMs. Such a data model would
support key input parameter requirements for fate and transport models—a core data set
required for each class of ENM. This data model may in turn lead to characterization and
reporting recommendations for ENM manufacturers, thereby providing much needed data
for environmental fate and transport modeling and risk assessment.
77
-------
6.0 Bibliography
Chapter 6.0
Bibliography
Albert, J., 2007. Bayesian Computation with R. Springer, New York.
Allen, B., Kichambare, P, Gou, P., Vlasova, I, Kapralov, A., Konduru, N., Kagan, V. and Star,
A. 2008. Biodegradation of Single-Walled Carbon Nanotubes through Enzymatic Catalysis.
Nemo Letters. 8 (11), pp 3899-3903.
AquaTerra. 2004. Evaluation of Chemical Bioaccumulation Models of Aquatic Ecosystems.
Prepared for the US EPA National Exposure Research Laboratory. February, 2004.
Auffan, M.; Rose, J.; Bottero, J.Y.; Lowry, G.V.; Jolivet, J.P.; Wiesner, M.R. 2009. Towards a
definition of inorganic nanoparticles from an environmental, health, and safety perspective.
Nature Nanotechnology., 4(10):634-641.
Balbus, J.M., Maynard, A.D., Colvin, V.L., Castranova, V., Daston, G.P., et al. 2007. Meeting
Report: Hazard Assessment for Nanoparticles—Report from an Interdisciplinary Workshop.
Environmental Health Perspectives 775(11): doi: 10.1289/ehp. 10327
Barber, M.C.. 2008. Bioaccumulation and Aquatic System Simulator (BASS) User's Manual
Version 2.2. EPA Report No. 600/R-01/035. March, 2008.
http://www.epa.gov/ceampubl/fchain/bass/BASS%20Manual.pdf
Bauer, C., Buchgeister, J., Hischier, R., Poganietz, W.R., Schebek, L., Warsen, J. 2008. Towards
a framework for life cycle thinking in the assessment of nanotechnology. Journal of Cleaner
Production. 16 (8-9), pp. 910-926.
Baun, A., Hartmann, N.B., Grieger, K., Kusk, K.O., 2008. Ecotoxicity of engineered
nanoparticles to aquatic invertebrates: a brief review and recommendations for future toxicity
testing. Ecotoxicology 17, 387-396.
Bekhit, H., and Hassan, A., 2005. Two-Dimensional Modeling of Contaminant Transport in
Porous Media in the Presence of Colloids. Advances in Water Resources, v.12. s.28. pp. 1320-
1335.
Bell, T. 2007. Understanding risk assessment of nanotechnology.
http://www.nano. gov/Understanding_Risk_Assessment.pdf
Belton, V., and Stewart, T. 2002. Multiple Criteria Decision Analysis: An Integrated Approach.
Boston: Kluwer Academic Publishers.
Biswas, P., and Wu, C.Y., 2005. Critical Review: Nanoparticles and the environment. Journal of
the Air and Waste Management. 55, 708-746.
Blaser, S.A., Scheringer, M., MacLeod, M., Hungerbuhler, K. 2008. Estimation of cumulative
aquatic exposure and risk due to silver: contribution of nanofunctionalized plastics and
textiles. Science of the Total Environment 390, 396-409.
Blume, T., Weisbrod, N., and Selker, J. 2005. On the critical salt concentrations for particle
detachment in homogeneous sand and heterogeneous Hanford sediments. Geoderma. 124: 1-
2 121-132, ISSN 0016-7061, DOI: 10.1016/j.geoderma.2004.04.007.
78
-------
6.0 Bibliography
Boncagni, N., Otaegui, J., Warner, E., Curran, T., Ren, J., and de Cortalezzi, M. 2009. Exchange
of TiO2 Nanoparticles between Streams and Streambeds. Environmental Science &
Technology. 43: 20 7699-7705.
Boxall, A., Chaudhry, Q., Sinclair, C., Jones, A., Aitken, R., Jefferson, B., and Watts, C. 2007a.
Current and Predicted Environmental Exposure to Engineered Nanoparticles. Central Science
Laboratory, York.
Boxall, A., Tiede, K., and Chaudhry, Q. 2007b. Engineered nanomaterials in soils and water:
how do they behave and could they pose a risk to human health? Nanomedicine (Future
Medicine} 2, 919-927.
Bradford, S. A., and Toride, N. 2007. A Stochastic Model for Colloid Transport and Deposition.
Journal Environmental Quality 36: 1346-1356.
Brant, J.; Lecoanet, H.; and Wiesner, M.R. 2005. Aggregation and deposition characteristics of
fullerene nanoparticles in aqueous systems. Journal of Nanoparticle Research. 7:545-553.
Brant, J., Labille, J., Bottero, J.-Y., and Wiesner, M.R. 2007. Nanoparticle Transport,
Aggregation, and Deposition, in Environmental Nanotechnology'. Wiesner and Bottero, eds.,
McGraw Hill, New York.
Brar, S., Verma, M., Tyagi, R.D., and Surampalli, R.Y. 2009 Engineered nanoparticles in
wastewater and wastewater sludge - Evidence and impacts. Waste Management. 30: 3 504-
520.
Breggin, L.K., and Pendergrass, J., 2007. Where Does the Nano Go? End-of-Life Regulation of
Nanotechnologies. Woodrow Wilson International Center for Scholars, Washington.
British Standards Institute. 2007. Terminology for nanomaterials. PAS 136. London: British
Standards Institute, ISBN 978 0 580 61321 0. 16 pp..
Bystrzejewska-Piotrowska, G., Golimowski, J., and Urban, P. 2009. Nanoparticles: Their
potential toxicity, waste and environmental management (literature review). Waste
Management, 29 (9), pp. 2587-2595.
Chappell, M.A. 2008. Solid-phase characteristics of engineered nanoparticles: A multi-
dimensional approach, p. 109-122, In I. Linkov and J. A. Steevens, eds. Nanotechnology:
Risks and Benefits. Springer, The Netherlands.
Choy, C., Wazne, M, and Meng, X. 2008. Application of an empirical transport model to
simulate retention of nanocrystalline titanium dioxide in sand columns. Chemosphere. 71:9
1794-1801.
Contardi, J.S., Turner, D.R., and Ahn, T.M. 2001. Modeling colloid transport for performance
assessment. Journal of Contaminant Hydrology 47 323-333
Corapcioglu M., and Choi H. 1996. Modeling colloid transport in unsaturated porous media and
validation with laboratory column data. Water Resources Research 32(12):3437-49.
Cullen, E., O'Carroll, D., Yanful, E.K., and Sleep, B. 2010. Simulation of the Subsurface
Mobility of Carbon Nanoparticles at the Field Scale, Advances in Water Resources. 33: 361-
371.
79
-------
6.0 Bibliography
DEFRA (Department for Environment, Food and Rural Affairs). 2007. Characterising the
Potential Risks posed by Engineered Nanoparticles A Second UK Government Research
Report. Defra Publications, London.
De Laender, F., Van Oevelen, D., Middelburg, J.J., and Soetaert, K. 2010. Uncertainties in
ecological, chemical, and physiological parameters of a bioaccumulation model: Implications
for internal concentrations and tissue-based risk quotients. Ecotox & Envr Safety. 73: 240-
246.
Environmental Defense (ED)—DupontNano Partnership. 2007. Nano risk framework.
http://nanoriskframework.com/page. cfm?tagID=l 081.
Espinasse, B., Hotze, E., and Wiesner, M. 2007. Transport and retention of colloidal aggregates
of C60 in porous media: Effects of organic macromolecules, ionic composition, and
preparation method. Environmental Science & Technology 41:7396-7402.
Fenner,K. M. Scheringer, M. MacLeod, M. Matthies, T. McKone, M. Stroebe A. Beyer, M.
Bonnell, A. C. Le Gall, J. Klasmeier, D. Mackay, D. van de Meent, D. Pennington, B.
Scharenberg, N. Suzuki, and F. Wania. 2005. Comparing Estimates of Persistence and Long-
Range Transport Potential among Multimedia Models. Environ. Sci. Technol., 39 (7), pp
1932-1942
Figueira J., Greco, S., and Ehrgott, M. 2005. Multiple Criteria Decision Analysis: State of the
Art Surveys. New York: Springer.
Filley, T.R., Ahn, M., Held, B.W., and Blanchette, R.A. 2005. Investigations of Fungal Mediated
(C60-C70) Fullerene Decomposition. Preprints of Extended Abstracts Presented at the ACS
National Meeting, American Chemical Society, Division of Environmental Chemistry 45(1),
446-450.
Former, J.D.; Lyon, D.Y.; Sayes, C.M.; Boyd, A.M.; Falkner, J.C.; Hotze, E.M.; et al. 2005. C60
in Water: Nanocrystal Formation and Microbial Response. Environmental Science &
Technology. 39 (11), 4307-4316.
Franklin, N., Rogers, N., Apte, S., Batley, G., Gadd, G., and Casey, P. 2007. Comparative
toxicity of nanoparticulate ZnO, bulk ZnO and ZnC12 to a freshwater microalga
(Pseudokirchneriella subcapitata): The importance of particle solubility. Environmental
Science & Technology 41:8484-8490.
Fu, K., and Sum, Y. 2003. Dispersion and solubilization of carbon nanotubes. Journal of
Nanoscience Technology. 3:351-364.
Goldman L., and Coussens C., (Eds.). 2005. Implications of Nanotechnology for Environmental
Health Research, Roundtable on Environmental Health Sciences, Research, and Medicine.
Board on Health Sciences Policy, The National Academies Press.
Gottschalk, F., Scholz, R.W., and Nowack, B. 2010a. Probabilistic material flow modeling for
assessing the environmental exposure to compound: Methodology and an application to
engineered nano-TiO2 particles. Environmental Modelling & Software 25 320-332.
80
-------
6.0 Bibliography
Gottschalk, F., Sonderer, T., Scholz, R., andNowack, B. 2010b. Possibilities and limitations of
modeling environmental exposure to engineered nanomaterials by probabilistic material flow
analysis. Environmental Toxicology and Chemistry. 29, 5, 1036-1048.
Grasso, D., Subramaniam, K., Butkus, M., Strevett, K., and Bergendah, J. 2002. A review of
non-DLVO interactions in environmental colloidal systems. Reviews in Environmental
Science & Biotechnology 1: 17-38, 2002.
Greaves-Holmes, W.A. 2009. Guide for the Safe Handling of Engineered and Fabricated
Nanomaterials. The Journal of Technology Studies. 1, 33-39
Grieger, K., Hansen, S., and Baun, A. 2009. The known unknowns of nanomaterials: Describing
and characterizing uncertainty within environmental, health and safety risks. Nanotoxicology.
3:3.
Grieger, K., Baun, A., and Owen, R. 2010. Redefining risk research priorities for nanomaterials.
Journal of Nanoparticle Research. 12: 383 -392.
Hansen, S. 2009. PhD thesis, Regulation and Risk Assessment of Nanomaterials - Too Little,
Too Late? Department of Environmental Engineering, Technical University of Denmark.
Hansen, S. F., Larsen, B. H., Olsen, S. I, Baun, A. 2007. Categorization framework to aid
Hazard Identification of Nanomaterials. Nanotoxicology 1:243-250
Hansen, S.F., Michelson, E.S., Kamper, A., Borling, P., Stuer Lauridsen, F., and Baun, A. 2008.
Categorization framework to aid exposure assessment of nanomaterials in consumer
products. Ecotoxicology 17:438-447.
Hardman, R. 2006. A toxicologic review of quantum dots: Toxicity depends on physicochemical
and environmental factors. Environmental Health Perspectives. 114:165-172.
Hock, J., Hofmann, H., Krug, H., Lorenz, C., Limbach, L., Nowack, B., et al. 2008. Guidelines
on the Precautionary Matrix for Synthetic Nanomaterials. Federal Office for Public Health
and Federal Office for the Environment, Berne.
Holsapple, M.P., Farland, W.H., Landry, T.D., Monteiro-Riviere, N.A., Carter, J.M., Walker
N.J., and Thomas, K.V. 2005. Research Strategies for Safety Evaluation of Nanomaterials,
Part II: Toxicological and Safety Evaluation of Nanomaterials, Current Challenges and Data
Needs. Toxicological Sciences. 88(1): 12-17.
Hotze, E.M., Bottero, J., and Wiesner, M. 2010. Theoretical Framework for Nanoparticle
Reactivity as a Function of Aggregation State. Langmuir 26(13), 11170-11175.
IOM. (R.J. Aitken, S.M. Hankin, B. Ross, C.L. Tran, V. Stone, T.F. Fernandes, K. Donaldson, R.
Duffm, Q. Chaudhry, T.A. Wilkins, S.A. Wilkins, L.S. Levy, S.A. Rocks, A. Maynard).
2009. EMERGNANO: A review of completed and near completed environment, health and
safety research on nanomaterials and nanotechnology. U.K. Defra project CB0409; report
TM/09/01. http://www.safenano.org/Uploads/EMERGNANO CB0409 Full.pdf
Johnson, C.P., Li, X., and Logan, B.E., 1996. Settling velocity of fractal aggregates.
Environmental Science & Technology. 30, 1911-1918.
81
-------
6.0 Bibliography
Johnson, W., Li, X., and Assemi, S. 2007. Deposition and re-entrainment dynamics of microbes
and non-biological colloids during non-perturbed transport in porous media in the presence
of an energy barrier to deposition. Advances in Water Resources 30 1432-1454.
Ju, B., and Fan, T. 2009. Experimental study and mathematical model of nanoparticle transport
in porous media. Powder Technology 192, 195-202.
Ju-Nam, Y., and Lead, J. 2008. Manufactured nanoparticles: An overview of their chemistry,
interactions and potential environmental implications. Science of the Total Environment 400
396-414.
Kanel, S., Goswami, R., Clement, T., Barnett, M., and Zhao, D. 2008. Two Dimensional
Transport Characteristics of Surface Stabilized Zero-valent Iron Nanoparticles in Porous
Media. Environmental Science & Technology. 42: 3 896 - 900.
Kennedy, A, Hull, M., Steevens, J., Donstova, K., Chappell, M., Gunter, J., Weiss, C. 2008.
Factors Influencing the Partitioning and Toxicity of Nanotubes in the Aquatic Environment.
Environmental Toxicology and Chemistry. 27, 9, 1932-1941
Klaine, S., et al. 2008. Critical Review: Nanomaterials in the Environment: Behavior, Fate,
Bioavailability, and Effects. Environmental Toxicology and Chemistry. 27, 9, 1825-1851.
Koelmans AA, Nowack B, and Wiesner MR. 2009. Comparison of manufactured and black
carbon nanoparticle concentrations in aquatic sediments. Environmental Pollution 157:1110-
1116.
Krayer von Krauss M., van Asselt, M., Henze M., Ravetz J., and Beck M. 2005. Uncertainty and
precaution in environmental management. Water Science Technology 52(6): 19.
Lam, C.W., James, J., McCluskey, R., Sivaram, A., & Hunter, R. (2006). A review of carbon
nanotube toxicity and assessment of potential occupational and environmental health risks.
Critical Reviews in Toxicology., 36(3): 189-217.
Lecoanet, H. and M.R. Wiesner. 2004. Velocity Effects on the deposition of fullerene and oxide
nanoparticles in porous media. Environmental Science and Technology, 38: 16 4377 - 4382.
Lecoanet, H., J.Y. Bottero, and M.R. Wiesner. 2004. Laboratory assessment of the mobility of
several commercial nanomaterials in porous media. Environmental Science & Technology.,
38: 195164-5169.
Li, Y., Wang, S., Luan, Z., Ding, J., Xu, C., Wu, D., 2003. Adsorption of Cadmium (II) from
aqueous solution by surface oxidized carbon nanotubes. Carbon 41, 1057-1062.
Li, Y. S., Y. G. Wang, K. D. Pennell, and L. M. Abriola. 2008. Investigation of the Transport
and Deposition of Fullerene (C60) Nanoparticles in Quartz Sands under Varying Flow
Conditions, Environmental Science & Technology, 42(19), 7174-7180.
Linkov, I, K. Satterstrom, S. Zemba & J. Steevens, 2006. Use of multi-criteria decision analysis
tools to facilitate weight-of-evidence evaluation in nanotechnology risk assessment. NSTI
Nanotechnology Conference and Trade Show, Boston, MA.
http://www.nsti.org/Nanotech2006/showabstract.html?absno=1611
82
-------
6.0 Bibliography
Linkov, I, Satterstrom, F.K., Steevens, J., Ferguson, E., Pleus, R. 2007. Multi-criteria decision
analysis and environmental risk assessment for nanomaterials. Journal of Nanoparticle
Research 9: 543-554.
Linkov, I. and Satterstrom, F. K.. 2009. Nanomaterial Risk Assessment and Risk Management
Review of Regulatory Frameworks. Real-Time and Deliberative Decision Making. NATO
Science for Peace and Security Series,2, 129-157.
Linkov, I, Satterstrom, F.K., Monica, J.C., Jr., Foss Hansen, S. and Davis, T.A. 2009a. Nano
Risk Governance: Current Developments and Future Perspectives. Nanotechnology: Law and
Business 6:203.
Linkov I, Steevens J, Chappell M, Tervonen T, Figuera JR, Merad M. 2009b. Classifying
Nanomaterial Risks Using Multi-Criteria Decision Analysis. In: Linkov I, Steevens JA (eds)
Nanomaterials: risks and benefits. Springer, Dordrecht, pp 179—191.
Liu, X. Y., D. M. O'Carroll, E. J. Petersen, Q. G. Huang, and C. L. Anderson. 2009. Mobility of
Multiwalled Carbon Nanotubes in Porous Media. Environmental Science & Technology, 25
43(21), 8153-8158.
Long, T.C., Saleh, N., Tilton, R.D., Lowry, G.V., Veronesi, B., 2006. Titanium dioxide (P25)
produces reactive oxygen species in immortalized brain microglia (BV2) implications for
nanoparticle neurotoxicity. Environmental Science and Technology. 40, 4346-4352.
Long, T.C., Tajuba, J., Sama, P., Saleh, N., Swartz, C., Parker, J., Hester, S., Lowry, G.V.,
Veronesi, B., 2007. Nanosize titanium dioxide stimulates reactive oxygen species in brain
microglia and damages neurons in vitro. Environmental Health Perspectives 115, 1631-
1637.
Loux, N.T. and Savage, N. 2008. An assessment of the fate of metal oxide nanomaterials in
porous media. Water, Air and Soil Pollution, 194:227-241
Lowry, G., and Cassman, E. 2009. Nanomaterial Transport, Transformation, and Fate in the
Environment: A Risk-Based Perspective on Research Needs. In Nanomaterials: Risks and
Benefits (NATO Science for Peace and Security Series C: Environmental Security) edited by
Igor Linkov and Jeffery Steevens.
Luoma, S.N., 2008, Silver Nanotechnologies and the Environment: Old Problems or New
Challenges: Project on Emerging Nanotechnologies Publication 15. Woodrow Wilson
International Center for Scholars and PEW Charitable Trusts, Washington D. C. 67p.
Mackay, C., Johns, M., Salatas, J., Bessinger, B., Perri, M. 2006. Stochastic probability
modelling to predict the environmental stability of nanoparticles in aqueous suspension.
Integrated Environmental Assessment and Management 2(3):293-298.
Maynard, Andrew D. 2006. Nanotechnology: A Research Strategy for Addressing Risk.
Woodrow Wilson International Center for Scholars. PEN 3.
Meerschaert, M.M., 2007. Mathematical Modeling. Elsevier Academic Press, London.
Metcalfe C, Bennett E, Chappell MA, Steevens JA, Depledge M, Goss G, Goudey S, Kaczmar S,
O'Brien N, Picado A. 2009. SMARTEN: strategic management and assessment of risks and
83
-------
6.0 Bibliography
toxicity of engineered nanomaterials. In: Linkov I, Steevens JA (eds) Nanomaterials: risks
and benefits. Springer, Dordrecht, pp 95-109.
Money, E, Robichaud C., Reckhow K. 2009. Nano-risk and macro-uncertainty: Using
probability networks to model the environmental implications of nanotechnology (abstract).
Moore, M.N. 2006. Do Nanoparticles Present Ecotoxicological Risks for the Health of the
Aquatic Environment? Environment International. 32(8): 967-976.
Morgan, K. 2005. Development of a preliminary framework for informing the risk analysis and
risk management of nanoparticles. Risk Analysis. 25(6): 1621-1635.
Moridis, G., Wu, Y., Pruess, K., 1999. EOS9nT: a TOUGH2 Module for the Simulation of Flow
and Solute/Colloid Transport. Report LBNL-42351. Lawrence Berkeley National
Laboratory, Berkeley, CA.
Moridis, G.J., Q. Hu, Y.-S. Wu, and G.S. Bodvarsson. 2003. Preliminary 3-D site-scale studies
of radioactive colloid transport in the unsaturated zone at Yucca Mountain, Nevada. Journal
of Contaminant Hydrology 60 (2003) 251- 286.
Mueller, N. and Nowack,. B.2008. Exposure Modeling of Engineered Nanoparticles in the
Environment. Environ. Sci. Technol. 42, 4447—4453.
Murashov, V., Howard, J. 2009. Essential features for proactive risk management. Nature
Nanotechnology. 4: 467-470. doi:10.1038/nnano.2009.205
Netzeva, Tatiana, Pavan, M, Worth A. 2007. Review of Data Sources, QSARs, and Integrated
Testing Strategies for Aquatic Toxicity. European Commission, JRC Scientific and Technical
Reports. EUR 22943 EN - 2007.
Nichols, JW, Bonnell M, Dimitrov, SD, Escher, B, Han, X, Kramer, NI. 2009. Bioaccumulation
Assessment Using Predictive Approaches. Int.Envr. Assessment & Mngmnt. Vol 5, No 4. pp
577-597.
NIOSH (National Institute for Occupational Safety and Health). 2009. Approaches to Safe
Nanotechnology: Managing the Health and Safety Concerns Associated with Engineered
Nanomaterials. http://www.cdc.gov/niosh/topics/nanotech/safenano/.
Norwegian Pollution Control Authority (NPCA). 2008. Environmental fate and ecotoxicity of
engineered nanoparticles. Report no. TA 2304/2007. Eds: EJ. Joner, T. Hartnik and C.E.
Amundsen. Bioforsk, As. 64.
Nowack, B., Bucheli, T.D. 2007. Occurrence, Behavior and Effects of Nanoparticles in the
Environment, Environmental Pollution. 150, 5-22.
Nowack, B., Mueller, N. 2008. Is anything out there? What life cycle perspectives of nano-
products can tell us about nanoparticles in the environment. Environmental Science &
Technology. 42 4447-4453.
NRC (National Research Council) Committee on Models in the Regulatory Decision Process.
2007. Models in Environmental Regulatory Decision Making. National Academies Press. 500
Fifth Street, NW Washington, DC 20001. www.nap.edu
84
-------
6.0 Bibliography
NRC (National Research Council) Committee on Improving Risk Analysis Approaches Used by
the U.S. EPA. 2009. Science and Decisions: Advancing Risk Assessment. National
Academies Press. 500 Fifth Street, NW Washington, DC 20001. www.nap.edu
Oberdorster, G. 1996. Significance of particle parameters in the evaluation of exposuredose-
response relationships of inhaled particles. Particu Sci Technol 14: 135-151.
Oberdorster G. 2010. Safety Assessment for Nanotechnology and Nanomedicine: Concepts of
Nanotoxicology. Journal of Internal Medicine; 267: 89-105.
Oberdorster, G., Maynard, A., Donaldson, K., Castranova, V., Fitzpatrick, J., Ausman, K.,
Carter, J. Karn, B., Kreyling, W., Lai, D., Olin, S., Monteiro-Riviere, N., Warheit, D., Yang,
H, and a report from the ILSI Research Foundation/Risk Science Institute Nanomaterial
Toxicity Screening Working Group. 2005a. Principles for characterizing the potential human
health effects from exposure to nanomaterials: elements of a screening strategy. Particle and
Fibre Toxicology 2.
Oberdorster, G., Oberdorster, E. & Oberdorster, J. 2005b. Nanotoxicology: An emerging
discipline evolving from studies of ultrafme particles. Environmental Health Perspectives,
113(7): 823-839.
ODPM (Office of the Deputy Prime Minister), 2004. DLTR Multi-Criteria Decision Analysis
Manual. Office of the Deputy Prime Minister.
http://www.communi ties.gov.uk/index.asp?id=l 142251.
Oberdorster G, Oberdorster E, Oberdorster J. 2007. Concepts of Nanoparticle Dose Metric and
Response Metric. Environmental Health Perspectives. 115(6): A 290.
Owen, R., M. Crane, K. Grieger, R. Handy, I. Linkov, and M. Depledge. 2009. Strategic
Approaches for the Management of Environmental Risk Uncertainties Posed by
Nanomaterials. Nanomaterials: Risks and Benefits. Springer, The Netherlands.
Packman, A., Brooks, N., Morgan, J. 2000. A physiochemical model for colloid exchange
between a stream and a sand streambed with bed forms. Water Resources Research. 36: 8
2351—2361.
Petersen E., Huang Q, Weber W. 2008. Ecological uptake and depuration of carbon nanotubes
by Lumbriculus variegates. Environmental Health Perspectives 113:1-32
Pruess, K. 1991. TOUGH2 - A General Purpose Numerical Simulator for Multiphase Fluid and
Heat Flow, Lawrence Berkeley Laboratory Report LBL-29400, Berkeley, CA.
Renn, O. 2008. Risk Governance in a Complex World. Earthscan, London.
Roberts, A., Mount, AS, Seda, B, Souther J., Qiao, R., Lin S, Ke, PC, Rao, AM, Klaine,
SJ.,2007. In vivo Biomodification of Lipid-Coated Carbon Nanotubes by Daphnia magna.
Environmental Science & Technology. 41 (8): 3025-3029.
Robichaud, C. O., A. E. Uyar, M. R. Darby, L. G. Zucker, and M. R. Wiesner. 2009. Estimates
of Upper Bounds and Trends in Nano-TiO2 Production As a Basis for Exposure Assessment.
Environ. Sci. Technol., 43 (12), pp 4227-4233.
85
-------
6.0 Bibliography
Roco, M. 2007. International Risk Governance Council Policy Brief: Nanotechnology Risk
Governance: Recommendationsybr a Global, Coordinated Approach to the Governance of
Potential Risks (IRGC, 2007); available at http://www.irgc.org/.
Ross, S.M., 2003. Introduction to Probability Models, Eighth Edition. Elsevier Academic Press,
San Diego.
Rozhkov SP, Goryunov AS, Sukhanova GA, Borisova AG, RozhkovaNN, Andrievsky GV.
2003. Protein interaction with hydrated C(60) fullerene in aqueous solutions. Biochemical
and Biophysical Research Communications. 303:5 62-5 66.
Rushton E., Jiang J., Leonard S., Eberly S., Castranova V., Biswas P., Elder A., Han X., Gelein
R., Finkelstein J., Oberdorster G. 2010. Concept of assessing nanoparticle hazards
considering nanoparticle dosemetric and chemical/biological response metrics. Journal of
Toxicology and Environmental Health. Part A. 73(5):445-61.
Ryan, Joseph N., Rebecca A. Ard, Robin D. Magelky, Menachem Elimelech, Ning Sun, and Ne-
Zheng Sun. 1999. Colloid Mobilization and Transport in Contaminant Plumes: Field
Experiments, Laboratory Experiments, and Modeling. EPA Environmental Research Brief.
EPA/600/S-99/001 February 1999.
Saaty, T.L., 1994. Fundamentals of Decision Making and Priority Theory with the Analytic
Hierarchy Process, 1st edn. Analytic Hierarchy Process Series; vol. 6. RWS Publications,
Pittsburgh, PA.
SCENIHR European Commission Scientific Committee on Emerging and Newly Identified
Health Risks (European Commission Scientific Committee on Emerging and Newly
Identified Health Risks SCENIHR). 2005. The Appropriateness of Existing Methodologies to
Assess the Potential Risks Associated with Engineered and Adventitious Products of
Nanotechnologies. Document number SCENIHR/002/05.
Sen TK, Mahajan SP, Khilar KC. 2002a. Colloid-associated contaminant transport in porous
media: 1. Experimental studies. AICHE J 2002;48(10):2366-74.
Sen TK, Nalwaya N, Khilar KC. 2002b. Colloid-associated contaminant transport in porous
media: 2. Mathematical modeling. AICHE J 2002;48(10):2375-85.
Sen, T.K., and Khilar, K.C. 2006. Review on subsurface colloids and colloid-associated
contaminant transport in saturated porous media. Advances in Colloid and Interface Science.
vol. 119, issue 2-3, 71 -96.
Shvedova A, Castranova V, Kisin ER, Schwegler-Berry D, Murray A., Gandelsman V., et al.
2003. Exposure to carbon nanotube material: assessment of nanotube cytotoxicity using
human keratinocyte cells. Journal of Toxicology and Environmental Health A 66:1909-1926.
Simunek, J., Changming He, J. L. Pang, and S. A. Bradford. 2006. Colloid-facilitated transport
in variably-saturated porous media: Numerical model and experimental verification, Vadose
Zone Journal, 5, 103 5 -1047.
Smith, C., Shaw, B., Handy, R,. 2007. Toxicity of single-walled carbon nanotubes to rainbow
trout (Oncorhynchus mykiss): Respiratory toxicity, organ pathologies, and other
physiological effects. Aquatic Toxicology. 82:94-109.
86
-------
6.0 Bibliography
Stoeger, T., Reinhard, C., Takenaka, S., Schroeppel, A., Karg. E., Ritter, B. Heyder, J., Schultz,
H. 2006. Instillation of six different ultrafme carbon particles indicates surface area threshold
dose for acute lung inflammation in mice. Environmental Health Perspectives. 114(3): 328-
333.
Stoeger, T., Schmid, O., Takenaka, S., Schulz, H. 2007. Inflammatory Response to TiO2 and
Carbonaceous Particles Scales Best with BET Surface Area. A 290-291. Environmental
Health Perspectives. 115(6): A290-A291.
Stone, V. (Project Coordinator). 2009. Engineered Nanoparticles: Review of Health and
Environmental Safety (ENRHES). Project Final Report. Edinburgh Napier University.
Sun, N., Elimelech, M., Sun, NZ, Ryan, J. 2001. A novel two-dimensional model for colloid
transport in physically and geochemically heterogeneous porous media. Journal of
Contaminant Hydrology 49.173-199.
Templeton, R., Ferguson, P., Washburn, K., Scrivens, W., Chandler, G. 2006. Life-cycle effects
of single-walled carbon nanotubes (SWNTs) on an estuarine meiobenthic copepod.
Environmental Science & Technology 40:7387-7393.
Ter Laak, T.L., Gebbink, W.A., Tolls, J. 2006. Estimation of soil sorption coefficients of
veterinary pharmaceuticals from soil properties. Environmental Toxicology and Chemistry.
25(4): 933-941.
Tervonen, T., Linkov, I. Figueira, J., Steevens, J., Chappell, M., Merad, M. (2009). Risk-based
Classification System of Nanomaterials. J. of Nanoparticle Research 11:757-766.
Thomas, K., Sayre, P. 2005. Research strategies for safety evaluation of nanomaterials, Part I:
Evaluating the human health implications of exposure to nanoscale materials. Toxicological
Sciences. 87(2), 316-321
Thomas, T., Thomas, K., Sadrieh, N., Savage, N., Adair, P., Bronaugh, R. 2006. Research
strategies for safety evaluation of nanomaterials, Part VII: Evaluating consumer exposures to
nanoscale materials. Toxicological Sciences. 91(1), 14-19
Thomas, T., Bahadori, T., Savage, N., Thomas, K., 2009. Moving toward exposure and risk
evaluation of nanomaterials: challenges and future directions. Nanomedicine and
Nanobiotechnology. 4 426-433
Tiede, K. Hassellov, M. Breitbarth, E. Chaudry Q. Boxall, A.B.A. 2009. Considerations for
environmental fate and ecotoxicity testing to support environmental risk assessment from
engineered nanoparticles, Journal Chromatography. A 1216, 503-509.
Tosco T., Sethi R. (2009). MNM1D: A Numerical Code for Colloid Transport in Porous Media.
Implementation and Validation. American Journal of Environmental Science 4: 516-524.
Tran, C.L., Donaldson, K., Stone, V., Fernandez, T., Ford, A., Christofi, N., Ayres, J.G., Steiner,
M., Hurley, J.F., Aitken, RJ. and Seaton, A. 2005, December. A scoping study to identify
hazard data needs for addressing the risks presented by nanoparticles and nanotubes. A
Research Report of the Institute of Occupational Medicine, United Kingdom.
Treder, M., Phoenix, C. 2003. Precautionary Principle Analyzed. Center for Responsible
Nanotechnology. http://www. crnano. org/precautionary.htm.
87
-------
6.0 Bibliography
Tsuji J., Maynard A., Howard P, James J, Lam C, Warheit D, Santamaria A.. 2006. Research
strategies for safety evaluation of nanomaterials, Part IV: risk assessment of nanoparticles.
Toxicological Sciences, 89:42-50.
US EPA (Environmental Protection Agency). 1996. Cleaner Technologies Substitutes
Assessment (CTSA): A Methodology and Resource Guide [Web Only] (EPA/744-R-95-002)
December 1996, 509 pages, http://www.epa.gov/dfe/pubs/tools/ctsa/
US EPA (Environmental Protection Agency). 2007. Nanotechnology White Paper. US EPA
Office of the Science Advisor. EPA 100/B-07/001 February
US EPA (Environmental Protection Agency) [Varner, K.]. 2008. Sampling and Analysis of
Nanomaterials in the Environment: A State-of-the-Science Review. Final Report. Prepared
for the US EPA Office of Research and Development by the Eastern Research Group.
EPA/600/R-08/098.
US EPA (Environmental Protection Agency). 2009a. A Conceptual Framework for US EPA's
National Exposure Research Laboratory. US EPA Office of Research and Development.
EPA/600/R-09/003.
US EPA (Environmental Protection Agency). 2009b. Nanomaterial Research Strategy. US EPA
Office of Research and Development. EPA 620/K-09/011.
van de Weerd, H., Leijnse, A., van Riemsdijk, W.H., 1998. Transport of reactive colloids and
contaminants in groundwater: effect of nonlinear kinetic interactions. Journal of
Contaminant Hydrology. 32, 313-331.
Vilks, P. F. Caron and M. K. Haas. 1998. Potential for the formation and migration of colloidal
material from a near-surface waste disposal site. Applied Geochemistry, Vol. 13, pp. 31-42.
Wainwright, John and Mark Mulligan. 2004. Environmental modelling: finding simplicity in
complexity. John Wiley and Sons, West Sussex, England.
Wang, Y., Li, Y., Former, J., Hughes, J., Abriola, L., Pennell, L. 2008. Transport and Retention
of Nanoscale C60 Aggregates in Water-Saturated Porous Media. Environmental Science &
Technology. 42: 10 3588 - 3594.
Warheit, D. B., Webb, T. R., Colvin, V. L., Reed, K. L., Sayes, C. R. 2007a. Pulmonary bioassay
studies with nanoscale and fine-quartz particles in rats: Toxicity is not dependent upon
particle size but on surface characteristics. Toxicological Sciences. 95 (1): 270-280.
Warheit, D. B., Webb, T. R., Reed, K. L., Frerichs, S., and Sayes, C. M. 2007b. Pulmonary
toxicity study in rats with three forms of ultrafme-TiO2 particles: Differential responses
related to surface properties. Toxicology 230: 90-104.
Warheit, D.B. 2008. How meaningful are the results of nanotoxicity studies in the absence of
adequate material characterization? Toxicological Sciences 101(2): 183-185.
Westerhoff, P. 2006. RFA: Exploratory Research to Anticipate Future Environmental Issues:
Impacts of Manufactured Nanomaterials on Human Health and the Environment Fate,
transformation and toxicity of manufactured nanomaterials in drinking water. 2003.
http://cfpub.epa.gov/ncer abstracts/index.cfm/fuseaction/display.publications/abstract/7387
-------
6.0 Bibliography
Widmer, M., Meili, C., Mantovani, E., Porcari, A. 2009. Mapping Study on Regulation and
Governance of Nanotechnologies, published under the FramingNano Project: A
multistakeholder dialogue platform framing the responsible development of Nanosciences &
Nanotechnol ogi es.
http://www.nanowerk.com/nanotechnology/reports/reportpdf/reportl26.pdf
Wiesner, M.R. 1999. Morphology of Particle Deposits, Journal of Environmental Engineering,
125(12), 1124-1132, 1999.
Wiesner and Bottero. 2007. Nanotechnology and the Environment. Chapter in Environmental
Nanotechnology: Applications and impacts ofnanomaterials. McGraw Hill, NY.
Wiesner MR, Lowry GV, Alvarez P, Dionysiou D, Biswas P. 2006. Assessing the risks of
manufactured nanomaterials. Environ Sci Technol. 40(14):4336-45.
Wiesner, M., Lowry, G., Jones, K., Hochella, M., Di Giulio. R., Gasman, E., Bernhardt, E. 2009.
Decreasing Uncertainties in Assessing Environmental Exposure, Risk, and Ecological
Implications of Nanomaterials. Environmental Science & Technology. 43 (17), 6458-6462.
Williams, AGB, Gregory, KB, Parkin, GF, Scherer, MM. 2005. Hexahydro-l,3,5-trinitro-l,3,5-
triazine Transformation by Biologically Reduced Ferrihydrite: Evolution of Fe Mineralogy,
Surface Area, and Reaction Rates. Environmental Science & Technology. 39 (14): 5183-
5189.
Wittmaack, K. 2007a. In search of the most relevant parameter for quantifying lung
inflammatory response to nanoparticle exposure: Particle number, surface area, or what?
Environmental Health Perspectives. 115:187-194.
Yoe, C., 2002. Trade-Off Analysis Planning and Procedures Guidebook. Prepared for Institute
for Water Resources. U.S. Army Corps of Engineers.
Zartarian, V., T. Bahadori, and T. McKone. 2005. Adoption of an official ISEA glossary. J.
Expo. Anal. Environ. Epidemiol. 15(l):l-5.
Zhang WX (2003) Nano scale iron particles for environmental remediation: an overview.
Journal of Nanoparticle Research. 5(3-4):323-332
Ziolli RL, Jardim WF. 2002. Photocatalytic decomposition of seawater-soluble crude-oil
fractions using high surface area colloid nanoparticles of TiO2. Journal of Photochemistry
andPhotobiology A: Chemistry 147: 205-212.
89
-------
Appendix A
Appendix A
Titles Pertaining to the Use of Exposure Models for Nanomaterials
(references highlighted in blue are reviewed in Appendix B)
Exposure Science and Model Evaluation
Title
Meeting Report: Hazard Assessment for Nanoparticles — Report
from an Interdisciplinary Workshop
Towards a framework for life cycle thinking in the assessment of
nanotechnology
Understanding risk assessment of nanotechnology
Nanoparticles and the environment
Where Does the Nano Go? End-of-Life Regulation of
Nanotechnologies
Nanoparticles: Their potential toxicity, waste and environmental
management
Nano risk framework
The Appropriateness of Existing Methodologies to Assess the
Potential Risks Associated with Engineered and Adventitious
Products of Nanotechnologies
Redefining risk research priorities for nanomaterials
Research Strategies for Safety Evaluation of Nanomaterials, Part
II: Toxicological and Safety Evaluation of Nanomaterials, Current
Challenges and Data Needs
Uncertainty and precaution in environmental management
First Author
Balbus, J.
Bauer, C.
Bell, T.
Biswas, P.
Breggin, L.
Bystrzejewska-
Piotrowska, G.
Environmental
Defense Fund
European
Commission
SCENIHR
Grieger
Holsapple, M.
Krayer von Krauss,
M.
Date
2007
2008
2007
2005
2007
2009
2007
2006
2010
2005
2005
Source
Meeting Report
Article
Article
Article
Report
Article
Report
Report
Article
Article
Article
Content
Risk assessment
Research
Risk assessment
Review
Review
Review
Risk assessment
Review, risk
assessment
Review, risk
assessment
Framework, review
Model development
Number of
References
15
60
44
387
233
104
70
approx. 220
71
36
40
-------
Appendix A
A review of carbon nanotube toxicity and assessment of
potential occupational and environmental health risks
Nano-risk and macro-uncertainty: Using probability networks to
model the environmental implications of nanotechnology
Do Nanoparticles Present Ecotoxicological Risks for the Health of
the Aquatic Environment?
Exposure Modeling of Engineered Nanoparticles in the
Environment
Models in Environmental Regulatory Decision Making
Science and Decisions: Advancing Risk Assessment
Occurrence, Behavior and Effects of Nanoparticles in the
Environment
Ecological uptake and depuration of carbon nanotubes by
Lumbriculus variegates
Safety Assessment for Nanotechnology and Nanomedicine:
Concepts of Nanotoxicology
Principles for characterizing the potential human health effects
from exposure to nanomaterials: elements of a screening
strategy
Nanotoxicology: An emerging discipline evolving from studies of
ultrafine particles
Risk Governance in a Complex World
In vivo Biomodification of Lipid-Coated Carbon Nanotubes by
Daphnia magna
Lam, C.
Money, E.
Moore, M.
Mueller, N.
NRC, Committee
on Models in the
Regulatory
Decision Process
NRC, Committee
on Improving Risk
Analysis
Approaches used
by the U.S. EPA
Nowack, B.
Petersen E.
Oberdorster G.
Oberdorster, G.
Oberdorster G.
Renn, O.
Roberts, A.
2006
2009
2006
2008
2007
2009
2007
2008
2010
2005a
2005b
2008
2007
Article
Abstract
Article
Article
Book
Book
Article
Article
Article
Article
Article
Book
Article
Review
Model development
Review, risk
assessment
Review, model
development
Model development,
review
Risk assessment
Review
Research
Review
Risk assessment
Review
Research, risk
assessment
Research
105
3
101
34
approx. 150
approx. 150
255
34
71
183
93
n/a
16
-------
Appendix A
International Risk Governance Council Policy Brief:
Nanotechnology Risk Governance: Recommendations for a
Global, Coordinated Approach to the Governance of Potential
Risks
Exposure to carbon nanotube material: assessment of nanotube
cytotoxicity using human keratinocyte cells
Research strategies for safety evaluation of nanomaterials, Part 1:
Evaluating the human health implications of exposure to
nanoscale materials
Research strategies for safety evaluation of nanomaterials, Part
VII
Considerations for environmental fate and ecotoxicity testing to
support environmental risk assessment from engineered
nanoparticles
Research strategies for safety evaluation of nanomaterials, Part
IV: risk assessment of nanoparticles
Nanotechnology White Paper
A Conceptual Framework for U.S. EPA's National Exposure
Research Laboratory
Hexahydro-1,3,5-trinitro-1,3,5-triazine Transformation by
Biologically Reduced Ferrihydrite: Evolution of Fe Mineralogy,
Surface Area, and Reaction Rate
Roco, M.
Shvedova, A.
Thomas, K.
Thomas, T.
Tiede, K.
Tsuji, J.
US EPA
US EPA
Williams, A.
2007
2003
2005
2006
2009
2006
2007
2009a
2005
Presentation
Article
Article
Article
Article
Article
Report
Report
Article
Risk governance
Research
Review
Review
Review
Review, risk
assessment
Framework
Framework
Research
0
53
16
12
85
61
182
13
50
Recent Reports and Compendia
Title
Ecotoxicity of engineered nanoparticles to aquatic invertebrates:
a brief review and recommendations for future toxicity testing
A Guide for the Safe Handling of Engineered and Fabricated
Nanomaterials
The known unknowns of nanomaterials: Describing and
characterizing uncertainty within environmental, health and safety
risks
First Author
Baun, A.
Greaves-Holmes,
W.
Grieger, K.
Date
2008
2009
2009
Source
Article
Article
Article
Content
Review
Review
Review
Number of
References
39
31
64
-------
Appendix A
Factors Influencing the Partitioning and Toxicity of Nanotubes in
the Aquatic Environment
EMERGNANO: A review of completed and near completed
environment, health and safety research on nanomaterials and
nanotechnology
Nanotechnology: A Research Strategy for Addressing Risk
Approaches to Safe Nanotechnology: Managing the Health and
Safety Concerns Associated with Engineered Nanomaterials
Environmental fate and ecotoxicity of engineered nanoparticles
Engineered Nanoparticles: Review of Health and Environmental
Safety (ENRHES)
Moving toward exposure and risk evaluation of nanomaterials:
challenges and future directions
A scoping study to identify hazard data needs for addressing the
risks presented by nanoparticles and nanotubes
Characterising the potential risks posed by engineered
nanoparticles
Sampling and Analysis of Nanomaterials in the Environment: A
State-of-the-Science Review. Final Report
Nanomaterial Research Strategy
Decreasing Uncertainties in Assessing Environmental Exposure,
Risk, and Ecological Implications of Nanomaterials
Kennedy, A,
IQM.forU.K.
DEFRA
Maynard, A.
NIOSH
Norwegian
Pollution Control
Authority
Stone, V. (project
coordinator) of
Edinburgh
Napier University
Thomas, T.
Iran, C.
DEFRA
US EPA [Varner,
K.]
US EPA
Wiesner, M.
2008
2009
2006
2009
2008
2009
2009
2005
2005
2008
2009b
2009
Article
Report
Report
Report
Report
Report
Article
Report
Report
Report
Report
Article
Research
Review
Review, risk
assessment
Review, risk
assessment
Review
Review, risk
assessment
Review, risk
assessment
Review, risk
assessment
Review, risk
assessment
Review, risk
assessment
Framework,
review, risk
assessment
Review
39
71
44
approx. 180
125
approx. 90
25
264
50
39
43
62
-------
Appendix A
Models that Simulate Particle, Aerosol, Polymer, and Colloid Behavior
Title
Two-Dimensional Modeling of Contaminant Transport in Porous
Media in the Presence of Colloids
Exchange of TiO2 Nanoparticles
between Streams and Streambeds
A Stochastic Model for Colloid Transport and Deposition
Aggregation and deposition characteristics of fullerene
nanoparticles in aqueous systems
Application of an empirical transport model to simulate retention
of nanocrystalline titanium dioxide in sand columns
Modeling colloid transport for performance assessment
Modeling colloid transport in unsaturated porous media and
validation with laboratory column data
Simulation of the Subsurface Mobility of Carbon Nanoparticles at
the Field Scale
Transport and retention of colloidal aggregates of C60 in porous
media: Effects of organic macromolecules, ionic composition, and
preparation method
Comparative toxicity of nanoparticulate ZnO, bulk ZnO and ZnCI2
to a freshwater microalga (Pseudokirchneriella subcapitata): The
importance of particle solubility
Dispersion and solubilization of carbon nanotubes
A review of non-DLVO interactions in environmental colloidal
systems
Deposition and re-entrainment dynamics of microbes and non-
biological colloids during non-perturbed transport in porous media
in the presence of an energy barrier to deposition
First Author
Bekhit, H.
Boncagni, N.
Bradford, S.
Brant, J.
Choy, C.
Contardi, J.
Corapcioglu, M.
Cullen, E.
Espinasse, B.
Franklin, N.
Fu, K.
Grasso, D.
Johnson, W.
Date
2005
2009
2007
2005
2008
2001
1996
2010
2007
2007
2003
2002
2007
Source
Article
Article
Article
Article
Article
Article
Article
Article
Article
Article
Article
Article
Article
Content
Model development
Research
Model development
Research
Research
Model development
Model development,
research
Model development,
research
Research
Research
Review
Model development,
review
Research, review
Number of
References
49
32
65
24
21
18
18
51
31
38
88
approx. 130
154
-------
Appendix A
Experimental study and mathematical model of nanoparticle
transport in porous media
Two Dimensional Transport Characteristics of Surface Stabilized
Zero-valent Iron Nanoparticles in Porous Media
Critical Review: Nanomaterials in the Environment: Behavior,
Fate, Bioavailability, and Effects. Environmental Toxicology and
Chemistry
Adsorption of Cadmium (II) from aqueous solution by surface
oxidized carbon nanotubes
Investigation of the Transport and Deposition of Fullerene (C60)
Nanoparticles in Quartz Sands under Varying Flow Conditions
Mobility of Multiwalled Carbon Nanotubes in Porous Media
Stochastic probability modelling to predict the environmental
stability of nanoparticles in aqueous suspension
Preliminary 3-D site-scale studies of radioactive colloid transport
in the unsaturated zone at Yucca Mountain, Nevada
EOSQnT: a TOUGH2 Module for the Simulation of Flow and
Solute/Colloid Transport
A physiochemical model for colloid exchange between a stream
and a sand streambed with bed forms
Protein interaction with hydrated C(60) fullerene in aqueous
solutions
Colloid Mobilization and Transport in Contaminant Plumes: Field
Experiments, Laboratory Experiments, and Modeling
Colloid-associated contaminant transport in porous media: 1.
Experimental studies
Colloid-associated contaminant transport in porous media: 2.
Mathematical modeling
Review on subsurface colloids and colloid-associated
contaminant transport in saturated porous media
Ju, B.
Kanel, S.
Klaine, S.
Li, Y.
Li, Y.
Liu, X.
Mackay, C.
Moridis, G.
Moridis, G.
Packman, A.
Rozhkov, S.
Ryan, J.
Sen, T.
Sen, T.
Sen, T.
2009
2008
2008
2003
2008
2009
2006
2003
1999
2000
2003
1999
2002a
2002b
2006
Article
Article
Article
Article
Article
Article
Article
Article
Report
Article
Article
Report
Article
Article
Article
Model development,
research
Research
Review
Model development,
research
Model development,
research
Model development,
research
Model development
Research
Model development,
research
Model development
Research
Model development,
research
Research
Model development
Review
20
30
249
23
34
36
11
52
46
40
16
111
28
21
215
-------
Appendix A
Toxicity of single-walled carbon nanotubesto rainbow trout
(Oncorhynchus mykiss): Respiratory toxicity, organ pathologies,
and other physiological effects
A novel two-dimensional model for colloid transport in physically
and geochemically heterogeneous porous media
Life-cycle effects of single-walled carbon nanotubes (SWNTs) on
an estuarine meiobenthic copepod
MNM1 D: A Numerical Code for Colloid Transport in Porous
Media. Implementation and Validation
Transport of reactive colloids and contaminants in groundwater:
effect of nonlinear kinetic interactions
Potential for the formation and migration of colloidal material from
a near-surface waste disposal site
Transport and Retention of Nanoscale C 60 Aggregates in
Water-Saturated Porous Media
Photocatalytic decomposition of seawater-soluble crude-oil
fractions using high surface area colloid nanoparticles of TiO2
Smith, C.
Sun, N.
Templeton, R.
Tosco, T.
van de Weerd, H.
Vilks, P.
Wang, Y.
Ziolli, R.
2007
2001
2006
2009
1998
1998
2008
2002
Article
Article
Article
Article
Article
Article
Article
Article
Research
Model development
Research
Model development,
research
Research
Research
Research
Research
44
52
40
31
27
30
35
19
-------
Appendix A
Multimedia Models Currently Used in Fate and Transport Simulation
Title
Estimation of cumulative aquatic exposure and risk due to silver:
contribution of nanofunctionalized plastics and textiles.
Current and Predicted Environmental Exposure to Engineered
Nanoparticles
Engineered nanomaterials in soils and water: how do they
behave and could they pose a risk to human health?
Probablilistic material flow modeling for assessing the
environmental exposure to compound: Methodology and an
application to engineered nano-TiO2 particles
Possibilities and limitations of modeling environmental exposure
to engineered nanomaterials by probabilistic material flow
analysis
Comparison of manufactured and black carbon nanoparticle
concentrations in aquatic sediments
First Author
Blaser, S.
Boxall, A.
Boxall, A.
Gottschalk, F.
Gottschalk, F.
Koelmans, A.
Date
2008
2007a
2007b
201 Oa
2010b
2009
Source
Article
Report
Article
Article
Article
Article
Content
Model development,
research
Research
Review, risk
assessment
Model development
Model development
Research
Number of
References
64
112
50
55
71
57
Alternative Approaches and Models
Title
Categorization framework to aid exposure assessment of
nanomaterials in consumer products
A toxicologic review of quantum dots: Toxicity depends on
physicochemical and environmental factors
Guidelines on the Precautionary Matrix for Synthetic
Nanomaterials
Classifying Nanomaterial Risks Using Multi-Criteria Decision
Analysis
Multi-criteria decision analysis and environmental risk
assessment for nanomaterials
First Author
Hansen, S.
Hardman, R.
Hock, J.
Linkov, 1.
Linkov, 1.
Date
2008
2006
2008
2009b
2007
Source
Article
Article
Paper
Book
Article
Content
Framework, research
Review
Model development
Model development,
risk assessment
Model development
Number of
References
12
48
n/a
37
25
-------
Appendix A
Use of multi-criteria decision analysis tools to facilitate weight-of-
evidence evaluation in nanotechnology risk assessment
From Comparative Risk Assessment to Multi-Criteria
Decision Analysis and Adaptive Management: Recent
Developments and Applications.
SMARTEN: strategic management and assessment of risks and
toxicity of engineered nanomaterials
Development of a preliminary framework for informing the risk
analysis and risk management of nanoparticles
Is anything out there? What life cycle perspectives of nano-
products can tell us about nanoparticles in the environment
DLTR Multi-Criteria Decision Analysis Manual
Concept of assessing nanoparticle hazards considering
nanoparticle dosemetric and chemical/biological response
metrics
Colloid-Facilitated Solute Transport in Variably Saturated Porous
Media: Numerical Model and Experimental Verification
[HYDRUS]
Precautionary Principle Analyzed
Trade-Off Analysis Planning and Procedures Guidebook
Linkov, 1.
Linkov, 1.
Metcalfe, C.
Morgan, K.
Nowack, B.
Office of the
Deputy Prime
Minister (ODPM),
UK
Rushton, E.
Simunek, J.
Treder, M.
Yoe, C.
2006
2006
2009
2005
2008
2004
2010
2006
2003
2002
Conference
Article
Book
Article
Article
Report
Article
Article
Article
Report
Model development,
review
Model development,
review
Model development,
risk assessment
Review, risk
assessment
Review
Model development,
review
Research
Model development
Review
Model development,
review
n/a
81
44
21
1
approx. 50
41
69
9
>150
-------
Appendix B
Appendix B
Exposure Model/Method Summaries for NMs in the Environment
List of Reviewed Models and Methods
Surface Water Models 101
Mackay et al. (2006) 101
Koelmans et al. (2009) 104
Packman etal. (2000) 107
Subsurface Models 110
MNM1D 110
Li et al. (2008) 113
Liu et al. (2009) 115
Cullen et al. (2009) 117
TOUGH2 120
HYDRUS 123
Multimedia Models 126
Boxall et al. (2007) 126
Blaser et al. (2008) 129
Mueller and Nowack (2008) 133
Gottschalk et al. (2009) 136
Alternative Approaches 141
Hock et al. (2008) 141
Hansenetal. (2008) 145
Linkov et al. (2007) 148
Linkov et al. (2009) 151
99
-------
Appendix B
Appendix B provides model reviews that are structured according to the categories discussed in
Section 4.1. We have provided reviews for models that have been developed for and/or applied
to ENMs as well as other models that have potential applicability (in their present form) for the
evaluation of ENM transport in the environment. Some review categories were not applicable to
all models. In those cases, we note that the categories/questions are not applicable to the model.
Table B-l provides a summary categorical overview of the models reviewed within this
Appendix.
Table B-l. Summary of Models Reviewed
Model Reference
Overview
Natural
Engineered
Model
Type
Deterministic (D) or
Probabilistic (P)
Process-based (P) or
Statistical (S^
Peer
Review
Journal Article
External Peer Review
Availability
and Usability
User-interface
Use of proprietary
algorithms
Source code
available
Page Reference
Surface Water Models
Mackay et al. (2006)
Koelmans et al. (2009)
Packman et al. (2000)
X
X
X
X
P
D
D
s
P
P
X
X
B-4
B-7
B-10
Subsurface Models
MNM1D
Li et al. (2008)
Liu et al. (2009)
Cullen et al. (2009)
TOUGH2
HYDRUS
X
X
X
X
X
X
D
D
P
D
D
D
P
P
P
P
P
P
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
B-13
B-16
B-19
B-21
B-24
B-27
Multimedia Models
Boxall et al. (2007)
Blaser et al. (2008)
Mueller and Nowack (2008)
Gottschalk et al. (2009)
X
X
X
X
D
D
D
P
P
P
s
s
X
X
X
B-30
B-33
B-37
B-40
Alternative Approaches
Hock et al. (2008)
Hansen et al. (2008)
Linkov et al. (2007)
Linkov et al. (2009)
X
X
X
X
X
D
D
D
P
P
P
P
s
X
X
X
X
X
X
B-45
B-49
B-52
B-55
100
-------
Appendix B
Surface Water Models
Mackay et al. (2006)
Summary: This model provides an approach for predicting whether nanoparticle suspensions in
aqueous systems will be stable - i.e., whether the particles will aggregate and settle out or
remain stable in solution. The model is inherently probabilistic and results in a probability
distribution for the predicted behavior. The model was developed to support the documented
analysis. No other use, verification, or validation is known.
KEY REFERENCES:
Mackay, C., Johns, M., Salatas, J., Bessinger, B., Perri, M. 2006. Stochastic probability
modelling to predict the environmental stability of nanoparticles in aqueous suspension.
Integrated Environmental Assessment and Management 2(3):293-298.
CONTACT/AVAILABILITY INFORMATION:
Christopher E Mackay (mackayc@exponent.com)
Exponent, 15375 Southeast 30th Place, Suite 250, Bellevue, Washington 98007, USA
PURPOSE AND SCOPE:
What is the model purpose? This model is designed to predict whether nanoparticle suspensions
will be stable in aqueous systems - i.e., whether the particles will aggregate and settle out or
remain in a stable solution.
What processes are simulated? The model considers buoyancy, aggregation, settling of
nanoparticles in solution.
What are the primary assumptions? Insufficient information is available to document the model
assumptions.
What transport media are considered? Aqueous solutions are considered in this model.
What spatial and temporal scales does the model consider? The scales are unspecified, however
it could be any scale at which a nanoparticle aqueous suspension may be present.
What are the forms of results produced? The primary model result is the probability that a
nanoparticle suspension will be stable in aqueous solution.
EVALUATION:
Background and History
How extensively has the model been used and applied? To our knowledge, the model has only
been used for the referenced study.
101
-------
Appendix B
What verification and validation has been conducted? There is no known verification or
validation of the model with experimental results or environmental data.
Complexity
What physical and chemical properties are considered? The model inputs include: fluid column
height, nanomaterial density, aggregation rate constant, dissociation constant, number of
particles, probability of aggregation per collision, particle radius, temperature, particle volume,
and fluid viscosity.
What is the mathematical representation? The model evaluates the critical buoyancy properties
using the Botzmann equation and potential aggregate settling using the Stokes-Einstein equation.
Kinetic aggregation is treated as a chain of binary reactions characterized by an aggregation rate
constant. The model implements a stochastic Monte Carlo solution and predicts a probability
distribution of apparent solubilities for nanoparticle suspensions.
What are the data requirements? The model input parameters are not extensive, however many
of the parameters (e.g., aggregation rate constant) are likely unavailable currently for most
nanoparticle solutions, particularly in environmental conditions (e.g., non-ideal aqueous
mixtures).
Consideration of Uncertainty
How does the model account for uncertainty? The model is implemented within a stochastic
Monte Carlo framework and thus inherently considers uncertainty. However, given that input
data are unavailable for many nanomaterials and natural conditions, realistic uncertainty
predictions may not be possible in most cases.
Availability and Usability - not applicable
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? This model was developed specifically to evaluate
nanoparticle specific behavior in aqueous systems. However, the model assumes ideal conditions
and input parameter knowledge that currently significantly limit its use for evaluating many
environmental nanomaterial transport problems.
How does the model address input data gaps associated with many traditional models? The
model accounts for key processes of nanoparticle stability and allows user-specification of
associated inputs (e.g., aggregation rate constant). However, the model does not provide any
guidance on parameterization for different nanomaterials. Also, it is unclear if the model can
easily be updated for emerging, scientifically relevant information for nanomaterial behavior.
What kind of interpretations/predictions can be made from this model? This model framework
may be useful to incorporate into other fate and transport models in order to extend their utility
to nanomaterials. For example, this modeling approach could predict the stability of nanoparticle
102
-------
Appendix B
suspensions under different environmental conditions (i.e, the effective aqueous concentration
accounting for the presence of stable particles). An environmental flow model then could predict
the migration of nanoparticle contaminants at the predicted effective concentration (assuming
that environmental conditions do not change).
103
-------
Appendix B
Koelmans et al. (2009)
Summary: This mass balance model was used to compare the concentrations of manufactured
carbon nanoparticles (MCNPs) to naturally occurring black carbon nanoparticles (BCNPs) in
aquatic sediments. The model is a relatively simple compartmental model accounting for
sedimentation, burial, and degradation. The analysis concluded that MCNP concentrations in
sediments are likely to be negligible relative to concentrations of BCNPs. This conclusion is
possible even when considering the significant uncertainties in the estimate due to the very
large magnitude difference in the estimated concentrations.
KEY REFERENCE:
Koelmans AA, Nowack B, Wiesner MR. 2009. Comparison of manufactured and black carbon
nanoparticle concentrations in aquatic sediments. Environ Pollut 157:1110-1116.
CONTACT/AVAILIBILITY INFORMATION:
A. A. Koelmans (bart.koelmans@wur.nl Phone: 31 0 317 483201)
Aquatic Ecology and Water Quality Management Group. Wageningen University, PO Box 47,
6700 AA Wageningen, The Netherlands.
PURPOSE AND SCOPE:
What is the model purpose? This mass balance model is used to compare the concentrations of
manufactured carbon nanoparticles (MCNPs) to naturally occurring black carbon nanoparticles
(BCNPs) in aquatic sediments. This model calculates the steady state concentrations of carbon
nanoparticles using the parameters of concentration in the sediment, nanoparticle sedimentation
flux, sediment thickness, accumulation, and a first order removal rate of manufactured carbon
nanoparticles.
What processes are simulated? This model simulates the processes of sedimentation, burial, and
removal due to aggregation and degradation to calculate the concentrations of carbon nanotubes
within aquatic sediments.
What are the primary assumptions? The model is characterized by three main assumptions: (1)
There is a distinct, mixed biologically active layer; (2) MCNPs enter the sediment through
sedimentation; (3) MCNP removal can be modeled as a first order decay process.
What transport media are considered? This model considers transport between a water column
and underlying sediments.
What spatial and temporal scales does the model consider? The specific spatial and temporal
scales are not specified. However, the sediment layer is assumed to be 10 cm thick.
What are the forms of results produced? The model outputs are MCNP concentrations and
MCNP to BCNP weight ratios.
104
-------
Appendix B
EVALUATION:
Background and History
How extensively has the model been used and applied? To our knowledge, the model has only
been used for the referenced study.
What verification and validation has been conducted? There is no known verification or
validation of the model with experimental results or environmental data.
Complexity
How many physical and chemical properties are considered? Input parameters include: sediment
layer thickness, sediment accumulation rate, first-order removal rate constant, sedimentation flux
of settling solids in the water column, total concentration of MCNPs in the water column,
concentration of settling solids.
What is the mathematical representation? The model is a relatively simple mass balance
compartmental model characterized by mass inflows and internal transformations.
What are the data requirements? To calculate the concentration of CNPs, this model requires
data inputs of: (1) sediment thickness; (2) sediment accumulation rate; (3) removal rate constant;
(4) sedimentation flux; (5) total CNP concentration in the water column; (6) concentration of
settling solids; and (7) Conditional distribution ratio between MCNP concentration in settling
particles and MCNP concentration in water. These values are presented in table 1 of the article.
Consideration of Uncertainty
How does the model account for uncertainty? The model is deterministic and does not explicitly
consider uncertainty.
Availability and Usability - not applicable
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? The model does not explicitly represent fundamental
mechanisms of particulate transport (e.g., the model does not rely on DLVO theory). Rather, the
model aggregates several potential mechanisms into single mass transfer and transformation
rates (e.g., first order decay, settling rate). This approach allows straightforward predictions to be
made. The model does not allow any theoretical estimation of behavior and instead relies
completely on the reliability and accuracy of the user-specified inputs.
How does the model address input data gaps associated with many traditional models? The
authors have been explicit in the basis for input parameters and have utilized several documented
estimation approaches - e.g., estimates of water concentrations based on Mueller and Nowack
(2008). However, the authors readily state that the results are highly uncertain due to lack of
knowledge regarding nanomaterials in the environment. The model is relatively simple and
should be easily updated to account for emerging, scientifically relevant information describing
nanomaterial behavior.
105
-------
Appendix B
What kind of interpretations/predictions can be made from this model? A major conclusion of
the analysis was that MCNPs are unlikely to be present in aquatic sediments at concentrations
significant relative to concentrations of BCNPs.
106
-------
Appendix B
Packman et al. (2000)
Summary: The model simulates mass exchange of colloids between a stream and a streambed.
The model considers key behaviors associated with particulates, including settling and porous
media filtering. The model has been applied successfully by Boncagni et al. (2009) to evaluate
an experimental study of titanium dioxide nanoparticle transport.
KEY REFERENCES:
Packman, A., Brooks, N., Morgan, J. 2000. A physiochemical model for colloid exchange
between a stream and a sand streambed with bed forms. Water Resources Research. 36: 8 2351 -
2361.
Packman, A. I, N.H. Brooks, and J. J. Morgan, Kaolinite exchange between a stream and s
treambed: Laboratory experiments and validation of a colloid transport model, Water Resources
Research 36: 8.
CONTACT/AVAILABILITY INFORMATION:
Aaron I. Packman (a-packman@northwestern.edu. Phone: 847-491-9902)
Department of Civil and Environmental Engineering, Northwestern University, Evanston, IL.
PURPOSE AND SCOPE:
What is the model purpose? The model simulates the mass exchange of colloids between a
stream and a streambed.
What processes are simulated? The model considers bed form driven advective pore water flow
driven by stream flow over the bed forms. The advective pore water flow along with
sedimentation and filtering in the porous medium resulted in pumping exchange of suspended
sediment between the water column and the sediments.
What are the primary assumptions? Assumptions of the model documented in the reference
included: the only effect of the bedform on the flow field in the bed is to produce a sinusoidal
pressure distribution at the bed surface (no erosion or resuspension of particulates);
homogeneous porous medium; suspension and bedform properties do not change with time;
What transport media are considered? The model simulates mass exchange between a stream
water column and underlying sediments.
What spatial and temporal scales does the model consider? The model formulation is in terms of
dimensionless parameters, which theoretically should allow the approach to be used at a wide
range of spatial scales.
What are the forms of results produced? The primary results of the model are predicted
concentrations in the water column and in the bed sediments as well as associated mass fluxes.
107
-------
Appendix B
EVALUATION:
Background and History
How extensively has the model been used and applied? The model used in the primary reference
as well as by Boncagni et al. (2009) to evaluate the transport of titanium dioxide nanoparticles.
Section 4.2.2.1 of the report describes the Boncagni (2009) modeling effort. We are not aware of
uses of the model other than these two academic studies.
What verification and validation has been conducted? The initial paper documenting the model
(Packman et al., 2000) was published along with a companion paper documenting a model
verification effort.
Complexity
What physical and chemical properties are considered? The model input parameters include:
particle diameter, particle settling velocity, dune wavelength, bedform height, filtration
coefficient, porosity of bed sediment, stream width, average stream velocity, dispersion
coefficient, depth of sand bed, stream depth, concentration in stream.
What is the mathematical representation? The model is developed using a series of equations
characterizing advective puming theory, colloid filtration, and settling. The formulation results in
an analytical expression.
What are the data requirements? There is an extensive list at the end of the document which
labels all the input parameters needed for the model.
Consideration of Uncertainty
How does the model account for uncertainty? The model does not explicitly consider
uncertainty.
Availability and Usability - not applicable
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? The model considers key behaviors of parti culate
transport including settling and porous media filtering. In addition, the model has been
successfully applied in a study of titanium dioxide nanoparticles (Boncagni et al., 2009).
How does the model address input data gaps associated with many traditional models? The
Boncagni et al. (2009) study involved calibration of the model to experimental data. The model
does not provide the ability to estimate input parameters for nanomaterials without calibration.
The model is also set up to be easily updated to account for emerging, scientifically relevant
information regarding nanomaterial behavior.
108
-------
Appendix B
What kind of interpretations/predictions can be made from this model? The primary model
results include predicted concentrations within surface water and underlying sediments as well as
the mass flux rate between these media.
109
-------
Appendix B
Subsurface Models
MNM1D
Summary: The model simulates one-dimensional transport of nanoparticles in porous media.
The model accounts for key nanoparticle behaviors, including attachment, detachment, and
blocking, as well as transient ionic strength effects. The authors have developed a unique
empirical relationship for attachment, detachment, and blocking coefficients as a function of
ionic strength.
KEY REFERENCES:
Tosco T., Sethi R. (2009). MNM1D: A Numerical Code for Colloid Transport in Porous Media.
Implementation and Validation. American Journal of Environmental Science 4: 516-524.
CONTACT/AVAILABILITY INFORMATION:
Tiziana Tosco (tiziana.tosco@polito.it Phone: +39 011-564-7670)
DITAG-Department of Land, Environment and Geotechnologies, Politecnico di Torino, C.so
Duca Degli Abruzzi 24, 10129, Torino, Italy
PURPOSE AND SCOPE:
What is the model purpose? This model was developed to simulate the one-dimensional transport
of nanoparticles in porous media.
What processes are simulated? The model considers potentially transient ionic strength
conditions, which can impact the stability of nanoparticle suspensions. The model accounts for
particle attachment and detachment using one or two linear and/or langmuirian sorption sites and
first-order kinetic attachment coefficients. The model also considers potential blocking
phenomena, whereby all sorption sites become occupied.
What are the primary assumptions? Information regarding model assumptions is not provided in
the reference.
What transport media are considered? Transport in saturated porous media is considered for this
model.
What spatial and temporal scales does the model consider? The potential spatial scale for the
model is not explicitly provided, however the example problems used for verification involve
small scale problems (0.1 m, similar to a laboratory column) over relatively short time periods
(minutes).
What are the forms of results produced? The model predicts nanoparticle concentrations as a
function of time and one-dimensional distance from the source.
110
-------
Appendix B
EVALUATION:
Background and History
How extensively has the model been used and applied? To our knowledge, the model has only
been used for the referenced study.
What verification and validation has been conducted? In this paper, the authors verified the
model algorithms through comparisons with existing groundwater transport models (HYDRUS-
1D and STANMOD).
Complexity
What physical and chemical properties are considered? Input parameters include: inlet colloid
concentration, solid bulk density, dispersion coefficient, Darcy velocity, porosity, attachment and
detachment coefficients, maximum attached particle concentration (for blocking), inlet salt
concentration (for ionic strength effects).
What is the mathematical representation? The model is formulated using an advection-
dispersion equation along with the relevant source/sink and transformation terms describing
particulate transport processes along with the coupled transport of a conservative tracer (salt, for
transient ionic strength effects). The model solution is implemented in Matlab using a finite
difference solution to the partial differential equations.
What are the data requirements? The data requirements for each equation presented in the model
are listed after the equations.
Consideration of Uncertainty
How does the model account for uncertainty? It does not consider uncertainty.
Availability and Usability - not applicable
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? This model was developed specifically for the
simulation of nanparticle transport in porous media. The model does account for some of the key
processes for nanomaterial behavior (attachment/detachment, blocking, and ionic strength
effects).
How does the model address input data gaps associated with many traditional models? The
model relies on an empirical relationship developed by the authors based on laboratory
experiments and theoretical considerations, which estimates attachment, detachment, and
blocking coefficients as a function of ionic strength. Also, the model is implemented in Matlab
and likely can readily be updated as emerging, scientifically relevant information becomes
available.
What kind of interpretations/predictions can be made from this model? Several of the input
parameters are difficult and potentially impossible to predict a priori (e.g., attachment and
111
-------
Appendix B
detachment rates for multiple types of sorption sites). This limitation does not necessarily
prevent the use of this model to evaluate environmental transport problems. However, it does
indicate that the model must be calibrated to laboratory and/or environmental data in order to
develop reasonable input parameter ranges.
112
-------
Appendix B
Li et al. (2008)
Summary: The one-dimensional model of nanoparticle transport in porous media accounted for
attachment and site blocking. The model was used to interpret experimental fullerene (c60)
transport. The experimental results and the calibrated model results were utilized to develop
correlations relating the maximum retention capacity to the flow velocity, nanoparticle size, and
mean grain size of the porous medium. The authors also estimated collision efficiency factors
based on their experimental and modeling results and compared them with theoretical
predictions using DLVO theory. The fitted values were more than one order of magnitude
greater than the theoretically predicted efficiency factors. The authors attribute this to surface
heterogeneities and suggest that clean bed filtration theory may need to be modified (to consider
surface heterogeneity) in order to accurately simulate nC60 transport in porous media.
KEY REFERENCES:
Li, Y. S., Y. G. Wang, K. D. Pennell, and L. M. Abriola. 2008. Investigation of the Transport
and Deposition of Fullerene (C60) Nanoparticles in Quartz Sands under Varying Flow
Conditions, Environmental Science & Technology, 42(19), 7174-7180.
CONTACT/AVAILABILITY INFORMATION:
Linda M. Abriola (linda.abriola@tufts.edu. Phone: 617-627-3237)
Department of Civil and Environmental Engineering, Tufts, University, 200 College Avenue,
Medford, Massachusetts
PURPOSE AND SCOPE:
What is the model purpose? The model simulates the one-dimensional transport of nanoparticles
in porous media, considering particle attachment and maximum retention (blocking). The model
was utilized in the referenced study to interpret laboratory results from a study of nC60 transport
in porous media.
What processes are simulated? The model considers particle advection, dispersion, attachment,
and maximum retention (site blocking).
What are the primary assumptions? Model assumptions are not documented.
What transport media are considered? The model evaluates transport in saturated porous media.
What spatial and temporal scales does the model consider? The model has been utilized to
evaluate laboratory data at a scale of approximately 16 cm (a laboratory column). The time scale
was not specified.
What are the forms of results produced? The primary model predictions are concentrations as a
function of space and time.
113
-------
Appendix B
EVALUATION:
Background and History
How extensively has the model been used and applied? To our knowledge, the model has only
been used for the referenced study.
What verification and validation has been conducted? Model results were compared successfully
with results measured in laboratory experiments.
Complexity
What physical and chemical properties are considered? Input parameters include: inlet
concentration, soil bulk density, porosity, dispersion coefficient, pore-water velocity, particle
attachment rate, and the particle retention capacity.
What is the mathematical representation? The model partial differential equations are solved
using a finite difference numerical approach.
What are the data requirements? Table 1 of the document outlines the data requirements for this
model.
Consideration of Uncertainty
How does the model account for uncertainty? The model does not explicitly consider
uncertainty.
Availability and Usability - not applicable
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? The model considers several key processes for
particulates in porous media, including attachment and blocking.
How does the model address input data gaps associated with many traditional models? The
experimental results and the calibrated model results were utilized to develop correlations
relating the maximum retention capacity to the flow velocity, nanoparticle size, and mean grain
size of the porous medium. It is unknown if the model can be easily updated to account for
emerging, scientifically relevant information regarding nanomaterial behavior.
What kind of interpretations/predictions can be made from this model? The authors estimated
collision efficiency factors based on their experimental and modeling results and compared them
with theoretical predictions using DLVO theory. The fitted values were more than one order of
magnitude greater than the theoretically predicted efficiency factors. The authors attribute this to
surface heterogeneities and suggest that clean bed filtration theory may need to be modified (to
consider surface heterogeneity) in order to accurately simulate nC60 transport in porous media.
114
-------
Appendix B
Liu et al. (2009)
Summary: The model was used to interpret experimental multiwalled carbon nanotube
(MWCNT) transport data. The model simulates one-dimensional transport of MWCNTs in
porous media. The model is based on colloid filtration theory (attachment) with an added site-
blocking term. The model successfully reproduced the experimental results. Results showed
that MWCNTs were relatively mobile under the higher flow rates evaluated in this study.
Because these flow rate conditions are similar to conditions in a drinking water treatment
system sand filter, the results suggest that augmented treatment technologies may be necessary
to remove MWCNTs from drinking water.
KEY REFERENCES:
Liu, X. Y., D. M. O'Carroll, E. J. Petersen, Q. G. Huang, and C. L. Anderson. 2009. Mobility of
Multiwalled Carbon Nanotubes in Porous Media. Environmental Science & Technology., 25
43(21), 8153-8158.
CONTACT/AVAILABILITY INFORMATION:
Denis M. O'Carroll (docarroll@eng.uwo.ca. Phone: 519-661-2193)
Department of Civil and Environmental Engineering, University of Western Ontario, London,
ON, Canada N6A5B9
PURPOSE AND SCOPE:
What is the model purpose? The model was used to interpret experimental multiwalled carbon
nanotube (MWCNT) transport data.
What processes are simulated? The model simulates one-dimensional transport of MWCNTs in
porous media. The model is based on colloid filtration theory (attachment) with an added site-
blocking term.
What are the primary assumptions? A complete set of model assumptions is not provided in the
reference.
What transport media are considered? Saturated porous media
What spatial and temporal scales does the model consider? The model has been used to evaluate
laboratory scale column experiments.
What are the forms of results produced? The primary model results are concentrations as a
function of time and space.
EVALUATION:
Background and History
How extensively has the model been used and applied? To our knowledge, the model has not
been used beyond the referenced study.
115
-------
Appendix B
What verification and validation has been conducted? The model was successfully used to
simulate measured data from laboratory experiments.
Complexity
What physical and chemical properties are considered? Input parameters include: inlet
concentration, soil bulk density, porosity, dispersion coefficient, pore-water velocity, particle
attachment rate, and the particle retention capacity.
What is the mathematical representation? The model is based on a finite-element solution to the
underlying partial differential equations.
What are the data requirements? The data requirements are the same as the physical and
chemical properties used in this model.
Consideration of Uncertainty
How does the model account for uncertainty? The model does not explicitly account for
uncertainty.
Availability and Usability - not applicable
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? The model considers several key processes for
particulates in porous media, including attachment and blocking.
How does the model address input data gaps associated with many traditional models? The
model was used successfully to simulate experimental data. Beyond the conditions of the
experiments, however, the model does not provide an approach or recommendations for
parameterization for other nanomaterials. It is unknown if the model can be easily updated to
account for emerging, scientifically relevant information regarding nanomaterial behavior.
What kind of interpretations/predictions can be made from this model? The MWCNTs were
quite mobile at the relatively high flow rates associated with the study (similar conditions that
exist in sand filtration drinking water treatment systems). This results suggests that traditional
filtration systems that do not incorporate additional treatment steps such as coagulation may not
adequately remove MWCNTs. Under natural subsurface conditions, where pore water velocities
would be in the lower range of those used in this study, the MWCNTs are substantially less
mobile. The MWCNTs employed here were specifically engineered to be stable in aqueous
solutions.
116
-------
Appendix B
Cullenetal. (2009)
Summary: This model of nanoparticle transport in porous media is based on colloid filtration
theory and associated porous media filtration mechanisms. In addition, the model allows
specification of a maximum paniculate retention (blocking). The model was developed to
support an academic study and has not been widely used or tested. The model was developed
using the COMSOL Multiphysics modeling system. Key conclusions of the analysis included
that carbon nanotubes were more mobile than fullerenes and that the particulate transport
behavior was strongly dependent on collision efficiency and blocking factors.
KEY REFERENCES:
Cullen, E., O'Carroll, D., Yanful, E.K., Sleep, B., 2010. Simulation of the Subsurface Mobility of
Carbon Nanoparticles at the Field Scale, Advances in Water Resources. 33: 361-371.
CONTACT/AVAILABILITY INFORMATION:
Denis M. O'Carroll (docarroll@eng.uwo.ca. Phone: 519-661-2193)
Department of Civil and Environmental Engineering, University of Western Ontario, London,
ON, Canada N6A 5B9
PURPOSE AND SCOPE:
What is the model purpose? The model was developed to support an academic study of nano-
fullerenes (nC60) and multi-walled carbon nanotubes (MWCNTs) and their transport in porous
media.
What processes are simulated? The model simulated nanoparticle advection, dispersion,
straining, attachment, and blocking in heterogeneous porous media. The model does not address
straining due to increased particle aggregation (i.e., temporal changes in the particulate
suspension).
What are the primary assumptions? Removal of the nanoparticles from the aqueous phase was
assumed to adhere to colloid filtration theory and associated mechanisms (e.g., deposition,
interception and sedimentation). The aquifer was assumed to be completely saturated and under
steady-state flow conditions. The vertical to horizontal permeability ratio was assumed to be 0.5
to account for anisotropy. The top and bottom domain boundaries were subject to Type II
(Neumann) no flow boundary conditions and the right and left side boundaries were subject to
Type I (Dirichlet) constant head boundary conditions. No nanoparticle flux occurred across the
top and bottom boundaries.
What transport media are considered?The model simulates transport in saturated porous media.
What spatial and temporal scales does the model consider? The model considered a two-
dimensional domain of approximately 10 square meters and time scales of several days.
What are the forms of results produced? The primary results of the model include particulate
concentrations with respect to time and space.
117
-------
Appendix B
EVALUATION:
Background and History
How extensively has the model been used and applied? To our knowledge, the model has only
been used for the referenced study.
What verification and validation has been conducted? The reference does not document any
verification or validation analysis associated with this model. However, the model was
implemented using the COMSOL multiphysics system, which has been subject to extensive
verification and validation.
Complexity
What physical and chemical properties are considered? Input requirements for the model
include: source concentration, hydraulic gradient, longitudinal and transverse dispersion,
porosity, collision efficiency factor, collector removal efficiency, and the particle retention
capacity.
What is the mathematical representation? The model is implemented within the COMSOL
multiphysics system, which provides a finite element solution to the underlying partial
differential equations.
What are the data requirements? Table 1 of the document outlines the input parameters needed
for this model.
Consideration of Uncertainty
How does the model account for uncertainty? The model does not consider uncertainty.
Availability and Usability
Is a user-friendly interface available? The model is implemented within the COMSOL
Multiphysics system, which does provide a user interface.
Does the model rely on proprietary algorithms and/or use interfaces? Yes, COMSOL
Multiphysics is aproprietary modeling system.
Is the documentation complete and transparent? Yes, the model is well documented within the
provided reference, and the COMSOL Multiphysics system is well documented.
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? Yes, the model incorporates key processes associated
with nanoparticle transport in porous media, including porous media filtering and blocking.
How does the model address input data gaps associated with many traditional models? The
model accounts for particle transport processes and allows user specification of the associated
118
-------
Appendix B
inputs. However, no guidance is provided for estimating these inputs. It is unknown if the model
can be easily updated to account for emerging, scientifically relevant information regarding
nanomaterial behavior.
What kind of interpretations/predictions can be made from this model? Nanoparticle transport
and maximum concentrations are very sensitive to collision efficiency factors and blocking
factors. At present, accurate methods to predict these factors a priori from soil and nanoparticle
characteristics have not been developed. For the conditions evaluated the carbon nanotubes are
much more mobile than nC60 due to the smaller collector efficiency associated with carbon
nanotubes. However, the mobility of nC60 increased significantly when a maximum retention
capacity term was included in the model. Model results also demonstrate that, for the systems
examined, nanoparticles were predicted to be less mobile in heterogeneous systems compared to
the homogeneous systems with the same average hydraulic properties.
119
-------
Appendix B
TOUGH2
Summary: The TOUGH2 simulators are recognized for their powerful simulation capabilities
involving complex fluid flow and heat transfer in porous and fractured media. The TOUGH2
codes have been applied to problems ranging from Yucca mountain groundwater flow to multi-
component environmental remediation. A module has been developed to support modeling of
the transport of colloids in porous media. The model simulates the potential filtering of colloids
through linear kinetic rate constants characterizing attachment and detachment.
KEY REFERENCES:
Pruess, K., C. Oldenburg and G. Moridis. TOUGH2 User's Guide, Version 2.0, Lawrence
Berkeley National Laboratory Report LBNL-43134, Berkeley, CA, November 1999. (2.2
Megabytes)
Pruess, K. TOUGH2 - A General Purpose Numerical Simulator for Multiphase Fluid and Heat
Flow, Lawrence Berkeley Laboratory Report LBL-29400, Lawrence Berkeley Laboratory,
Berkeley, C A, May 1991.
Moridis, G. J., Y. S. Wu, and K. Pruess, EOSSnT: A TOUGH2 Module for the Simulation of
Nonisothermal Fluid Flow and Solute/Colloid Transport in the Subsurface; LBNL Report No.
44260, August 1999
CONTACT/AVAILABILITY INFORMATION:
Karsten Pruess (K Pruess@lbl.gov. Phone: 510-486-6732)
Senior Scientist, Earth Sciences Division, Lawrence Berkely National Laboratory
PURPOSE AND SCOPE:
What is the model purpose? TOUGH2 is a general-purpose numerical simulation program for
multi-dimensional fluid and heat flows of multiphase, multicomponent fluid mixtures in porous
and fractured media. Chief application areas are in geothermal reservoir engineering, nuclear
waste isolation studies, environmental assessment and remediation, and flow and transport in
variably saturated media and aquifers. Although primarily designed for geothermal reservoir
studies and high-level nuclear waste isolation, TOUGH2 can be applied to a wider range of
problems in heat and moisture transfer, and in the drying of porous materials. The TOUGH2
simulator was developed for problems involving strongly heat-driven flow. To describe these
phenomena a multi-phase approach to fluid and heat flow is used, which fully accounts for the
movement of gaseous and liquid phases, their transport of latent and sensible heat, and phase
transitions between liquid and vapor. TOUGH2 takes account of fluid flow in both liquid and
gaseous phases occurring under pressure, viscous, and gravity forces according to Darcy's law.
Interference between the phases is represented by means of relative permeability functions. The
code includes Klinkenberg effects and binary diffusion in the gas phase, and capillary and phase
adsorption effects for the liquid phase. Heat transport occurs by means of conduction (with
thermal conductivity dependent on water saturation), convection, and binary diffusion, which
includes both sensible and latent heat.
120
-------
Appendix B
What processes are simulated? Processes simulated include advection, diffusion, dispersion,
equilibrium, kinetic, or combined sorption following linear, Langmuir, and/or Freundlich
isotherms, radioactive decay including daughter products, linear chemical reactions, colloid
filtration, and colloid-assisted solute transport.
What are the primary assumptions? Some of the assumptions include: water flow is isothermal;
concentration are at a trace level without an effect on flow properties; the gas phase pressure
does not deviate from the reference pressure of the system; there is no phase change.
What transport media are considered? Porous and fractured media
What spatial and temporal scales does the model consider? The spatial and temporal scales
considered are flexible. The model can evaluate small or large-scale problems at variable time
scales.
What are the forms of results produced? The primary results are predicted material flows and
concentrations as well as and fluxes associated with the simulated system.
EVALUATION:
Background and History
How extensively has the model been used and applied? Since its introduction in 1991, TOUGH2
has been extensively utilized and applied in the academic, regulatory, and industry realms. The
model has been utilized by approximately 300 organizations in over 30 countries.
What verification and validation has been conducted? The model has been associated with many
verification and validation analyses.
Complexity
What physical and chemical properties are considered? The model requires many input
parameters. The reader is referred to the user's guide for additional information.
What is the mathematical representation? TOUGH2 uses an integral finite difference method for
space discretization, and first-order fully implicit time differencing. A choice of either a sparse
direct solver or a various preconditioned conjugate gradient algorithms is available for linear
equation solution. Thermophysical properties of water are represented, within experimental
accuracy, by steam table equations provided by the International Formulation Committee. The
program provides options for specifying injection or withdrawal of heat and fluids. Double-
porosity, dual-permeability, and multiple interacting continua (MINC) methods are available for
modeling flow in fractured porous media
What are the data requirements? Section 5 of the Moridis (199) User Guide document outlines
the input parameters needed for the model.
121
-------
Appendix B
Consideration of Uncertainty
How does the model account for uncertainty? Although the model is deterministic and does not
explicitly account for uncertainly, several studies have ealuated uncertainty using the model.
Availability and Usability
Is a user-friendly interface available? No GUI is available inthe public domain, however several
utility programs supporting use of the model can be found here:
http://esd.lbl.gov/TOUGH2/PROGRAMS/FREEPROGRAMS .html. In addtion, a proprietary
user interface named PetraSim is available from Rockware, Inc. Golden, Colorado.
Does the model rely on proprietary algorithms and/or use interfaces? No, the model is in the
public domain. However, a proprietary user interface named PetraSim isavailable.
Is the documentation complete and transparent? Yes. Also the source code for TOUGH2,
written in standard FORTRAN??, is available from the Energy Science and Technology
Software Center (ESTSC) of the U.S. Department of Energy. The LBNL group, headed by
Karsten Pruess, serves as custodians of the code, and provides limited technical support.
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? Yes, the model includes a module (EOS9nT) that
implements aspects of colloid filtration theory including attachment/detachment behavior.
Although the model accounts for some critical processes for particulate transport in porous
media, it cannot provide guidance for parameterizing the model for nanomaterials. Also, the
model has been revised numerous times since its inception, demonstrating that enhancements can
be made to account for emerging, scientifically relevant information.
What kind of interpretations/predictions can be made from this model? The primary model
predictions include concentrations and flow rates over time as well as boundary fluid and
contaminant mass fluxes.
122
-------
Appendix B
HYDRUS
Summary: HYDRUS simulates the movement of water, heat, multiple solutes, and particulates
in variably saturated porous media (unsaturated and saturated zones). In addition to key
processes relevant to transport in porous media, the model utilizes colloid filtration theory to
describe the attachment/detachment behavior of particulates in porous media systems. The
model has been extensively used, verified, and peer reviewed. A graphical user interface is
available. The ID version of the model is public domain, however the 2D and 3D versions are
proprietary.
KEY REFERENCES:
Simunek, J., Changming He, J. L. Pang, and S. A. Bradford, Colloid-facilitated transport in
variably-saturated porous media: Numerical model and experimental verification, Vadose Zone
Journal, 5, 1035-1047, 2006.
Simunek, J., and M. Th. van Genuchten, Using the Hydrus-lD and Hydrus-2D codes for
estimating unsaturated soil hydraulic and solute transport parameters, in van Genuchten, M. Th.,
F. J. Leij, and L. Wu (eds.) Characterization and Measurement of the Hydraulic Properties of
Unsaturated Porous Media, University of California, Riverside, CA, 1523-1536, 1999.
CONTACT/AVAILABILITY INFORMATION:
Jirka SimunekJirka Simunek (Jiri.Simunek@ucr.edu. Phone: 951-827-7854)
Professor and Hydrologist, Department of Environmental Sciences, University of California
Riverside, Riverside, CA 92521, USA
PURPOSE AND SCOPE:
What is the model purpose? HYDRUS simulates the movement of water, heat, multiple solutes,
and particulates in variably saturated porous media (unsaturated and saturated zones).
What processes are simulated? HYDRUS considers advection, diffusion and dispersion,
sorption, and degradation of up to 15 solutes. The model also simulates diffusive transport in the
gas phase. In addition, HYDRUS can evaluate non-equilibrium mass transfer through two-
region, dual porosity sorption formulation, which considers mobile and immobile regions of the
pore space. Filtration theory is used to describe attachment/detachment behavior of particulates
(viruses, colloids, or bacteria).
What are the primary assumptions? The technical reference and user guide do not provide a
listing of assumptions.
What transport media are considered? Partially saturated and/or fully saturated porous media
may be simulated.
What spatial and temporal scales does the model consider? The spatial and temporal scales that
may be considered are flexible. The finite element mesh may be set up to conform to irregular
boundaries that encompass very small or very large systems (e.g., a few centimeters or miles).
123
-------
Appendix B
What are the forms of results produced? The primary model predictions include concentrations
and flow rates over time as well as boundary water and contaminant mass fluxes.
EVALUATION:
Background and History
How extensively has the model been used and applied? From its introduction in the mid 1990's,
HYDRUS has been used very extensively, primarily (though not exclusively in the
environmental field to evaluate potential contaminant migration under various scenarios. A web
page listing HYDRUS-related references includes well over 100 citations (http://www.pc-
progress. com/en/Default. aspx?h3 d-references).
What verification and validation has been conducted? The HYDRUS technical manual (2006)
provides a series of comparisons with other models and with laboratory and field data. Several of
the references on the HYDRUS website (http://www.pc-progress.com/en/Default.aspx7h3d-
references) include model verification and validation analyses.
Complexity
What physical and chemical properties are considered? The model includes a relatively
complex, spatially explicit representation of porous media. The model input includes a large
number of potential parameters for a series of optional modules simulating a range of different
processes (e.g., constructed wetlands design, root water uptake). The reader is referred to the
user manual (2007) for additional information about the model input parameters.
What is the mathematical representation? The flow model is based on a finite element solution
to Richard's equation for variably saturated flow. The transport model is based on a finite
element solution to the advection-dispersion equation.
What are the data requirements? The model input includes a large number of potential
parameters. The reader is referred to the user manual (2007) for additional information about the
model input parameters.
Is the model linked with any corresponding input databases? We are not aware of explicit
database linkages with HYDRUS, however we have not performed a comprehensive search of
the extensive HYDRUS documentation and references. Given its extensive use, it would not be
surprising if the model had been incorporated into a data-driven modeling system.
Consideration of Uncertainty
How does the model account for uncertainty? The model is deterministic and the graphical user
interface does not appear to support uncertainty analysis. As with any deterministic model,
HYDRUS could be implemented for a probability-based analysis using Monte Carlo techniques.
Availability and Usability
Is a user-friendly interface available? Yes, a user interface is available at the product website
(http://www.pc-progress.com/).
124
-------
Appendix B
Does the model rely on proprietary algorithms and/or use interfaces? The ID version of the
model and an associated user interface are in the public domain. The 2D and 3D versions of the
model are proprietary.
Is the documentation complete and transparent? Yes. Also, the source code is available for the
ID code. The code for the 2D and 3D versions may not be available, because the model is
proprietary.
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? Yes, the model relies on colloid filtration theory and
accounts for processes of particulate attachment and detachment. The model does not appear to
account for potential blocking or enhanced transport processes. The model also does not consider
potentially changing aqueous chemical conditions (e.g., pH, ionic strength) which can have a
profound impact on particulate dispersions.
How does the model address input data gaps associated with many traditional models? Although
the model accounts for some critical processes for particulate transport in porous media, it cannot
provide guidance for parameterizing the model for nanomaterials. Also, the model has been
revised numerous times since its inception, demonstrating that enhancements can be made for
emerging, scientifically relevant information.
What kind of interpretations/predictions can be made from this model? The primary model
predictions include concentrations and flow rates over time as well as boundary water and
contaminant mass fluxes.
125
-------
Appendix B
Multimedia Models
Boxall et al. (2007)
Summary: This model develops a series of algorithms to calculate predicted environmental
concentrations (PECs) of engineered nanoparticles in air, soil, and water that arise from the use
of a single product containing nanoparticles. The algorithms combine to form equations to
calculate PECs based on the concentration of nanoparticles within the product combined with
hypothetical daily usage of the product. This simplistic approach is applied to a limited range of
products, environmental compartments, and life cycle stages, thus limiting the information that
can be used. The parameter values are also entered as point estimates, addressing uncertainty at
a minimum thus requiring high confidence in the data used to populate the model.
KEY REFERENCES:
Boxall, A., Chaudhry, Q., Sinclair, C., Jones, A., Aitken, R., Jefferson, B., Watts, C. 2007.
Current and Predicted Environmental Exposure to Engineered Nanoparticles. Central Science
Laboratory, York.
CONTACT INFORMATION:
Alistair B A Boxall (abab500@york.ac.uk. Phone: 01904 434791)
Senior lecturer, Environment Department, University of York, Heslington, York, YO10 5DD
PURPOSE AND SCOPE:
What is the model purpose? This model is designed to calculate the PECs of engineered
nanoparticles in air, soil, and water arising from a range of applications of products containing
nanoparticles. In order to calculate the PECs, the authors develop a series of algorithms to
determine master equations for PECs of air, water, and soil.
What processes are simulated? The modeled processes are broken down into the specific media
they are simulated for. The model simulates dispersion and emissions for transport in air, dilution
and application processes for water, and application processes for soil.
What are the primary assumptions? This model assumes that the system is in a steady state.
What transport media are considered? The model considers the media of air, water, sludge, and
soil.
What spatial and temporal scales does the model consider? Spatial and temporal scales are
neglected in this model.
What are the forms of results produced? This model produces point estimates for the PECs of
nanomaterials in the specified environmental compartment as a result of a specific consumer
product.
126
-------
Appendix B
EVALUATION:
Background and History
How extensively has the model been used and applied? To our knowledge, the model has only
been used for the referenced study.
What verification and validation has been conducted? There is no known verification or
validation of the model with experimental results or environmental data.
Complexity
What physical and chemical properties are considered? There are no considerations of specific
physical and chemical properties of the nanomaterials. Most parameter values considering
nanomaterials are estimated, such as concentration levels within the product and application
rates.
What is the mathematical representation? PECs within the air compartment are calculated by
computational fluid dynamic models to capture the dispersion after emissions from personal
hygiene products, skin care products, traffic, and industrial stack sources. The model also
includes wind speeds and dilution factors in the calculations.
To calculate the PECs for water, the model uses dilution equations for direct application of
nanoparticles to the surface water, application via runoff and spray drift, and application via the
sewage system dependent upon the amount of wastewater produced per capita per day. These
three equations combine to determine the PECs for surface water. The dilution factor is specific
to the receiving water.
Soil PECs are calculated based on the processes for direct application (through remediation
technologies, plant protection products, excretion of nanomedicines used in veterinary products,
and from aerial deposition) and the application of sewage sludge. To calculate the PECs as a
result of sludge application, the authors first calculate the PECs of the sewage sludge based on
parameters for concentration of the nanoparticles within the product, daily usage, percent
removed via sewage treatment, market penetration, and sludge production.
What are the data requirements? The data requirements for calculations of PECs are presented in
the table below. Along with the data characteristic are a description of the data needed and the
type of form the data is inputted as (e.g. probability distribution, point estimate, ranking, etc.).
The model input parameters are not extensive, however many of the parameters such as
application rates, concentrations, and removal fractions are likely unavailable for most
nanoparticle containing products.
127
-------
Appendix B
Data Characteristic Type of Input Description
Application rate(s)
Dimensions
Runoff
Spray drift
Nanoparticle
concentration
Daily usage
Market penetration
Removal percentage
Wastewater amount
Dilution factor
Sludge production
Point estimate This is the rate at which nanoparticles are applied to the
environmental source. Application can be directly, via runoff, via
spray drift, or via sludge.
Point estimate Length, width, depth, and density of the environmental media.
Point estimate Fraction of the nanoparticle applied via runoff.
Point estimate Fraction of the nanoparticle release in spray drift.
Point estimate Percentage of nanoparticle contained in the product being evaluated.
Point estimate This represents the amount of nanoparticle emitted due to usage of
the product per day.
Point estimate The amount of the population using the product.
Point estimate Fraction of nanoparticle removed by sewage treatment
Point estimate The amount of wastewater produced and applied to the
environmental compartment of concern per capita per day
Point estimate The dilution factor in the receiving water
Point estimate The amount of sludge produced and applied to the environmental
compartment of concern per capita per day
Table. Data Requirements.
Consideration of Uncertainty
How does the model account for uncertainty? The model does not allow for parameters to be
inputted in distributional form, but rather as point estimates. The authors consider uncertainty by
calculating PECs for the different concentrations of nanoparticles within the product generating a
three-scenario analysis.
Availability and Usability - not applicable
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? This model was developed specifically for predicting
environmental exposure of nanomaterials, however does not model the chemical and physical
properties specific to the behaviors of nanomaterial s.
How does the model address input data gaps associated with many traditional models? The data
gaps associated with traditional models are not addressed in this model. The model does not
include parameters for the chemical and physical properties pertinent to nanomaterial behavior
so it cannot be updated with this emerging information.
What kind of interpretations/predictions can be made from this model? Because this model is a
simplistic approach based on a limited range of products, life-cycle processes, and environmental
compartments and processes, it is difficult to accurately calculate the PECs of air, soil, and water
for engineered nanoparticles without sufficient data.
128
-------
Appendix B
Blaser et al. (2008)
Summary: This multimedia model incorporates the use of the Rhine river model in estimating
the predicted environmental concentrations (PECs) of nanomaterials in environmental
compartments of an aquatic setting (sewage treatment plants, freshwaters, and freshwater
sediments). However, the model has many limiting assumptions pertaining to emissions of
nanomaterials, most likely generating underestimates of the actual concentrations in the
environmental compartments.
KEY REFERENCES:
Blaser, S.A., Scheringer, M., MacLeod, M., Hungerbuhler, K., 2008. Estimation of cumulative
aquatic exposure and risk due to silver: contribution of nanofunctionalized plastics and textiles.
Science of the Total Environment 390, 396-409.
CONTACT/AVAILABILITY INFORMATION:
Martin Scheringer (scheringer@chem.ethz.ch. Phone: +41 (44) 632 30 62)
Senior Scientist, Safety and Environmental Technology Group, Institute for Chemical and
Bioengineering, ETH Zurich, Wolfgang-Pauli-Str. 10, Room HCI G 127, CH-8093 Zurich,
Switzerland
PURPOSE AND SCOPE:
What is the model purpose? The model is designed to ultimately characterize the risk of aquatic
exposure of silver through 4 stages: (1) Conduct mass flow analysis of silver and emissions
estimates; (2) Estimate the PEC of silver in a river system; (3) Estimate PNEC of silver; and (4)
Quantify the risk.
The development of a 13 compartment mass-balance model is used to determine the flow of
silver within a freshwater system, utilizing the Rhine river model to assess the flow within the
river system. The model then aims to quantify the PECs of silver within the environmental
compartments of STPs, freshwater, and freshwater sediments. Once the PECs are calculated, the
model calculates risk quotients (PEC/PNEC) by developing PNEC estimates from the literature.
What processes are simulated? The model simulates mass flow of silver into the environment
based on emission scenarios and mass balance models for correspond transport media. The
processes include sedimentation, exchange, diffusion, water flow, burial, and bed load shift.
What are the primary assumptions? This model considers the emissions of engineered
nanomaterials for nano-silver as the only source of silver into the river stream. This assumption
neglects sources such as particulate emissions, production of silver-containing plastics and
textiles, aerial deposition, and leachates. The model bases its assessment on estimated silver use
in the year 2010.
What transport media are considered? This model considers transport media of natural
freshwaters and freshwater sediments.
129
-------
Appendix B
What spatial and temporal scales does the model consider? The spatial scale is selected to be the
year of 2010 and the spatial domain is the 25 member countries of the European Union.
What are the forms of results produced? The results are in the form of PEC s and ultimately risk
quotients. These results are point estimates for each of the scenario (low, intermediate, and high)
EVALUATION:
Background and History
How extensively has the model been used and applied? To our knowledge, the model has only
been used for the referenced study.
What verification and validation has been conducted? The model compares its outputs to those
of empirical data from 16 different sources relating to the environmental concentrations in STP,
dry sewage sludge, river water, and river sediments. This comparison is presented in figure 5 of
the article.
Complexity
What physical and chemical properties are considered? It is not clear what physical and
chemical properties are considered in the model.
What is the mathematical representation? The emission of silver into the aquatic environment is
determined by first calculating the amount of silver released into wastewaters from biocidal
products (equation 1 in the article). The authors then use this output in equation 2 to calculate the
amount of silver input into the natural water system. Values are calculated for all 3 emission
scenarios.
PECs are estimated utilizing the Rhine river compartment model which includes moving water,
stagnant water, and the top layer of the equatic sediment and displayed in figure 3 of the article.
Permanent sediment is included in the representation, but is not modeled. This model
incorporates the processes of water flow, bed load shift, sedimentation, burial, exchange, and
diffusion. The parameter values
What are the data requirements? The data requirements for calculations of PECs are presented in
the table below. Along with the data characteristic are descriptions of the data needed and the
type of form the data is inputted as (e.g. range, point estimate, ranking, etc.). Model input
parameters for the Rhine river model were more extensive, utilizing a range of values with a
confidence factor.
130
-------
Appendix B
Data Characteristic Type of Input Description
Silver emissions for
into the environment
Emission scenario
parameters
Mass balance
parameters
Predicted No-Effect
Concentrations
Point estimates There were 17 different input parameters for silver emissions into
the environment. These input parameters were characterized by the
media in which silver was emitted (wastewater, STP, natural water,
sewage sludge, solid waste, landfills, slag, fly ashes, and the
atmosphere) as well as the form from which the silver came
(biocidal products, other sources). These parameters were given 3
different values pertaining to the emission scenarios of minimum,
intermediate, and maximum scenarios.
Point estimates There were 14 input parameters pertaining to the emission scenarios
which quantified the population, silver release rate, amount of silver,
wastewater produced, and fraction removed during different stages
of the silver life cycle.
Range There were 7 input parameters to model the mass balance model
once the silver had entered the aquatic environment. These
parameters included: (1) water flow velocity; (2) Concentration of
suspended paniculate matter (SPM); (3) Sediment density; (4)
Porosity of sediment; (5) Settling velocity of SPM; (6) Resuspension
rate; and (7) SPM-water partition coefficient.
Point estimates These values were taken from the literature and used to calculate the
risk quotient in the environmental compartments of STP, freshwater,
and freshwater sediments
Table. Data Requirements
Consideration of Uncertainty
How does the model account for uncertainty? The model accounts for uncertainty by conducting
a three scenario analysis labeled minimum-, intermediate-, and maximum emission scenarios.
The intermediate scenario considers the most probable assumptions, while the minimum and
maximum scenarios consider the assumptions that lead to lower or elevated PECs. The model
performs a first-order error propagation to assess the uncertainty in the model outputs.
Confidence factors in some of the model inputs are used to address the uncertainty of model
inputs such as those used in the mass balance models within the aquatic environment.
Availability and Usability - not applicable
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? The chemical and physical properties for this model
only pertain to the behavior of the nanomaterials within the aquatic system. Dispersivity
properties (as suggested in section 2.2 of this report when considering nanomaterial behavior in
aquatic systems) do not appear to be considered in the Rhine river model, so all relevant
nanospecific properties are not considered in this model.
How does the model address input data gaps associated with many traditional models?
Parameter values can be altered in the Rhine river model as information is made available.
However, many of the values pertinent to nanomaterial behavior may not be included.
131
-------
Appendix B
What kind of interpretations/predictions can be made from this model? The model is intended to
calculate PECs and ultimately the risk characterization of silver within different environmental
compartments (STPs, freshwater, and freshwater sediment). However, the model only considers
nano-silver in biocidal products as the exclusive source of silver emission, neglecting other
sources such as particulate emissions. Therefore, the PECs are most likely underestimates of the
actual concentrations.
132
-------
Appendix B
Mueller and Nowack (2008)
Summary: This model was the first attempt at simulating the flow of nanomaterials through the
environment based on a life cycle assessment. Though much of the data required to model this
flow is uncertain and estimated on best guesses or worst case scenarios, the model sets up a
framework to calculate predicted environmental conditions (PECs) for environmental
compartments such as air, soil, sewage treatment plants (STP), and surface waters. This
framework can be built upon as scientifically relevant information on nanomaterials becomes
available.
KEY REFERENCES:
Mueller, N. Nowack,. B.2008. Exposure Modeling of Engineered Nanoparticles in the
Environment. Environ. Sci. Technol. 42, 4447—4453.
CONTACT/AVAILIBILITY INFORMATION:
Bernd Nowack (nowack@empa.ch. Phone: +4171 2747 692)
Empa-Swiss Federal Laboratories for Materials Testing and Research, Technology & Society
Laboratory, Environmental Risk Assessment and Management Group, Lerchenfeldstrasse 5, CH-
9014 St. Gallen
PURPOSE AND SCOPE:
What is the model purpose? This model was designed to calculate the quantities of engineered
nanoparticles in different compartments of the environment based on a life cycle assessment. Per
this assessment, the model is set up to utilize material flow analysis based on the parameters of
estimated worldwide production volume, allocation of the production of nanomaterial into
product categories, particle release from products, and flow coefficients between environmental
compartments. Ultimately the model is intended to calculate PECs such that risk quotients can be
determined based on PNECs located in the literature.
What processes are simulated? The processes simulated are substance flows based on life cycle
assessments and emissions into the environment based on production and product use.
What are the primary assumptions? The primary assumptions of the model were: (1) Primary
compartments of air, soil, and water were considered to be homogeneous; (2) For simplicity the
system was considered to be in a steady-state; (3) In the absence of real data, substance flow
rates between environmental compartments were the same for all nanoparticles; and (4)
Secondary compartments (i.e. sediments and groundwater) were not considered.
What transport media are considered? The transport media of air, soil, and water compartments
are considered in this model.
What spatial and temporal scales does the model consider? The model is spatially representative
of Switzerland and bases its parameters for flow coefficients on tons/year. However, the system
is assumed to be in a steady state such that transfer coefficients can be used to calculate
concentrations.
133
-------
Appendix B
What are the forms of results produced? The model produces PEC for the three types of
nanoparticles studied. These calculations are then compared to predicted no effect concentrations
(PNEC) derived from the literature to estimate a possible risk quotient (PEC/PNEC). Typically,
risk quotients are categorized by a threshold of 1.
EVALUATION:
Background and History
How extensively has the model been used and applied? To our knowledge, the model has only
been used for the referenced study.
What verification and validation has been conducted? There is no known verification or
validation of the model with experimental results or environmental data.
Complexity
What physical and chemical properties are considered? The model is based on material flow
analysis and used flow coefficients between environmental compartments to determine PECs
within each compartment. It is not clear if physical and chemical properties are considered
within these compartments.
What is the mathematical representation? The calculation of PEC for specific environmental
compartments is determined by substance flow through waste incineration plants (WIPs)
landfills, and/or sewage treatment plants (STPs). This produced point estimates for PECs
dependent upon the input values into the system.
What are the data requirements? The data requirements for calculations of PECs are presented in
the table below. Along with the data characteristic are descriptions of the data needed and the
type of form the data is inputted as (e.g. range, point estimate, ranking, etc.). The model input
parameters are not extensive, however many of the parameters such as application rates,
concentrations, and removal fractions are likely unavailable for most nanoparticle containing
products. Thus, the input requirements were based on worst-case scenarios in the absence of
sufficient data.
134
-------
Appendix B
Data Characteristic Type of Input Description
Worldwide production Point estimates
volumes
Allocation of the Point estimates
production volumes to for weighting
product categories factors
Particle release from
products
Flow coefficients
Point estimates
Point estimates
Production volumes were taken at a best guess based on the
production volumes of 10 companies in Switzerland.
Different product categories were assigned to a specific nanoparticle
if present in that product category. Weight factors allocated how
much of the total production volume was contained in that product
category.
Each nanospecific product category had 2 to 6 different release
points during the life-cycle. A percentage of nanomaterial release
was given to each of these release points. The compartment to which
the nanomaterial was released was also specified.
Flow coefficients between different compartments in the model were
quantified by tons/year.
Table. Data Requirements.
Consideration of Uncertainty
How does the model account for uncertainty? The model accounts for uncertainty by developing
a realistic and high exposure scenario (RE- and HE scenario). The HE scenario relied on worst-
case scenario estimations, which lead to higher concentrations in the environment. This two-
scenario analysis was the extent of uncertainty analysis conducted.
Availability and Usability - not applicable
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? This model was developed to estimate PECs of
nanomaterials with the lack of data concerning relevant chemical and physical properties specific
to them.
How does the model address input data gaps associated with many traditional models? The
model addresses data gaps associated with many traditional models by assigning flow
coefficients between environmental compartments, rather than calculating the flow within each
compartment, and modeling a flow-in, flow-out process. However, calculation of the flow
coefficients may not include all the necessary parameters. The model can be updated as
information regarding percentage of particle release from products is quantified, as well as more
data that can increase the certainty of flow coefficients.
What kind of interpretations/predictions can be made from this model? This model is a
simplified approach to quantifying PECs of nanomaterial based on a life cycle assessment. The
framework seems applicable, but requires more certainty in the parameter values before this
assessment can be used in traditional risk assessments.
135
-------
Appendix B
Gottschalk et al. (2009)
Summary: This multimedia environmental model was designed to calculate the predicted
environmental concentrations (PECs) to be used towards a quantitative assessment of the risks
of nanomaterials in the environment. Because of the current data gaps surrounding
nanomaterials, this model was developed to model basically any substance with a distinct lack
of data concerning environmental fate, exposure, emission and transmission characteristics.
Thus, by combining methods of sensitivity and uncertainty analysis, Monte Carlo simulation,
and Markov Chain Monte Carlo modeling, the proposed model can realistically calculate PECs
when facing significant data gaps.
KEY REFERENCES:
Gottschalk, F., Scholz, R.W., Nowack, B., 2009. Probabilistic material flow modeling for
assessing the environmental exposure to compound: Methodology and an application to
engineered nano-TiO2 particles. Environmental Modelling & Software 25 (2010) 320-332.
Gottschalk, F., Sonderer, T., Scholz, R., & Nowack, B.. 2010. Possibilities and limitations of
modeling environmental exposure to engineered nanomaterials by probabilistic material flow
. Environmental Toxicology and Chemistry. 29, 5, 1036-1048.
Gottschalk, F., Sonderer, T., Scholz, R., & Nowack, B.. 2009. Modeled Environmental
Concentrations of Engineered Nanomaterials (TiC>2, ZnO, Ag, CNT, Fullerenes) for Different
Regions. Environmental Science and Technology. 43, 24, 9216-9222.
CONTACT/A VAILIBILITY INFORMATION:
Fadri Gottschalk (fadri . gottschalk@env.ethz . ch. Phone: +41 44 632 63 25)
ETH Zurich, Institute for Environmental Decisions (IED), CFtN J 73.2, Universitatstrasse 22,
8092 Zurich
PURPOSE AND SCOPE:
What is the model purpose? The proposed exposure assessment develops its methods from a
material flow analysis (MFA) modeling approach and is accomplished via the steps:
(1) characterize source and production volumes of material (compounds or chemicals);
(2) estimate the emissions of material to environmental compartments (air, sediment, soil,
surface water, etc.); (3) specify the fate in the environment; and (4) derive distributions of
predicted environmental concentrations (PECs) for the studied material.
The proposed PEC modeling approach combines sensitivity and uncertainty analysis, Monte
Carlo (MC) simulation, Bayesian and Markov chain modeling, and is intended for cases
characterized by a distinct lack of data concerning environmental fate, exposure, emission, and
transmission characteristics. The simulated PECs for materials in the desired environmental
compartments provide the basis for the quantitative exposure assessment and are derived from
the results of the probabilistic material flow analysis (PMFA).
136
-------
Appendix B
What processes are simulated? The model simulates the material flow of substance through 11
compartments consisting of life-cycle (production, manufacturing, and consumption), recycling,
waste incineration plants, sewage treatment plants, untreated sewage, landfills, atmosphere, soil,
sediments, groundwater, and surface water.
What are the primary assumptions? The model assumptions are presented in section 2.2 of the
article title system analysis. The section presents the system flow chart to explain all of the inputs
and outputs between each environmental compartment.
What transport media are considered? This model considers the transport media of air, surface
water, subsurfaces, sewage treatment plants, and waste incineration plants.
What spatial and temporal scales does the model consider? The model considered the flow of
engineered nanoparticles in Switzerland as a spatial scale.
What are the forms of results produced? The results of the model give distributions of PECs for
different environmental compartments.
EVALUATION:
Background and History
How extensively has the model been used and applied? This method and approach has been
utilized in several recent publications (Gottschalk et al. 2010, Gottschalk et al. 2009). The former
is a recent follow-up in which the author addresses both the possibilities and limitations of using
this particular method for modeling environmental fate and transport. Within this article, the
author discusses the other models currently used to model PECs with regards to nanomaterials.
Comparatively, the author justifies why this model is more useful than the models currently
developed. As of yet, there have been no new authors that have used this model in their own
nanomaterial risk assessment.
What verification and validation has been conducted? Due to the novelty of its design and
limitations of other research in this area, this model cannot be verified using similar models.
Complexity
What physical and chemical properties are considered? The model is based on material flow
analysis and used flow coefficients between environmental compartments to determine PECs
within each compartment. It is not clear if physical and chemical properties are considered
within these compartments.
What is the mathematical representation? A flow chart (Figure 1) of the mathematical
methodology displays the steps necessary in deriving the most informative probability
distributions of PECs for any material using the PMFA modeling framework. The remainder of
this summary will discuss this process and the methods presented by Gottschalk which follow
the steps: (1) develop a system model with a corresponding general system of linear equations to
describe the flow model; (2) use a probabilistic Monte Carlo method to determine a probability
distribution of the PECs in selected environmental compartments; (3) utilize Markov chain
137
-------
Appendix B
Monte Carlo algorithms based on the previous findings to more accurately determine the
probability distribution of the PECs; and (4) perform sensitivity on the PMFA to give the risk
manager more insight on the model.
Steps 1 through 3 of the methodology conceptualization are followed to develop a system design
of the desired mass balance compartment flow model which specifies the
compartments/processes in which the material will flow through and the directional flows
associated with them. Often, a flow chart is prepared to aid in the visual understanding of the
model as well as the mathematical formulation. The compartments or processes represent the
different stages of material flow and directional flows represent the transfer coefficients of
material. Transfer coefficients imply the transport of material between and within the
compartments.
If all transfer coefficients of the material are known (or estimated by point values), the flow of
the material in the system can be determined mathematically through a stationary input-output
model of a set of n linear equations containing n unknowns. Matrix algebra can be used in this
case and solutions for PECs can be found deterministically by computing inverse matrices.
However, if transfer coefficients are not known, a probabilistic approach must be used to capture
the uncertainty of these values. Density functions that represent the uncertainty of transfer
coefficient values will be used in the probabilistic methods used to determine PECs for the
desired environmental compartments.
Before the probabilistic methods are employed, a system of linear equations must be defined
which calculates flows to and deposition within the examined compartments of the system. This
system follows the balance principle that the mass of all inputs into a process equals the mass of
all outputs of the process and includes accumulation or depletion of mass within the
compartment. The balance principle is used to define the transfer coefficients between each
compartment. Matrix algebra, input values of transfer coefficients, and values of periodic input
of material to specified compartments are then combined to define a matrix equation that can
solve the steady-state of the desired system of linear equations. As mentioned, point estimates of
the input parameters can be used to solve the system straightforwardly and may be done for
model validation (MFA standard).
Once the system of linear equations has been determined, the study employs a probabilistic
determination of the unknown output variables (storage of material within the processes) via
Monte Carlo (MC) methods. The MC methods in this study follow two steps: (1) model the
probability distributions of all model input parameters (transfer coefficients); and (2) a repeated
computation of a proposed linear equation system to determine the output variables. The first
step determines a sequence of random variables for each model input parameter following a
desired probability distribution (picked to represent the probability that the input parameter falls
within a particular interval). The second step solves the new system of linear equations that result
from these input parameters. At the conclusion of the MC method, this algorithm has solved
thousands of systems of linear equations, generating a probability distribution of the unknown
output variables. This is a crucial step in the modeling procedure, because it allows the modeler
to gain insight on the PECs of the studied material without a full understanding of all input
138
-------
Appendix B
parameters. By simulating this method enough (Meerschaert, 2007), we can develop a smooth
probability density curve of the output variables.
To more accurately describe the probability distribution of the unknown output variables,
Markov chain Monte Carlo (MCMC) methods were employed. Markov chains are discrete
random processes with the Markov property (Ross, 2003). A discrete random process is defined
as a system that can be in various states and changes randomly in discrete steps. The Markov
property distinguishes that the probability distribution of the next step is determined only by the
current state of the system, and not by the previous states of the system. The state of the system
changes according to the transition probabilities of the chain. Over time, regardless of the
starting state of the system, this discrete process converges to an equilibrium distribution.
MCMC methods are a class of algorithms that sample from probability distributions based on
constructing a Markov chain that has a desired distribution as its equilibrium distribution
(stationary).
Gottschalk uses Bayesian inferences to provide posterior distributions to define what is known
about unobservable model input parameters given measured or simulated data. The simulated
data refers to the probability distributions of the output variables found by the MC methods
above. This posterior distribution is used in the MCMC algorithms to improve upon the
probability distributions of the PECs. Bayesian techniques require a proposal distribution and
posterior distribution. In particular, Gottschalk utilized Metropolis algorithms (Albert, 2007)
with a symmetric proposal distribution and the above posterior distribution as the stationary
distribution of the Markov chain to approach the optimal input parameter values. The MCMC
method is repeated until the Markov chain is considered well mixed, or close to its steady state
solution.
System Model
and System of
Linear Equations
Probabilistic
modeling of the
input parameters
Definition of
goods, products,
processes and
functions
Sensitivity and
uncertainty analysis
Probability
distributions of
PECs
Improved PEC
probability
distributions
Markov Chain
Monte Carlo
(MCMC) modeling
Interpretation,
assessment of
PECs
Figure 1. Basic flow chart of the PMFA methodology.
139
-------
Appendix B
What are the data requirements? Table 3 of the Gottschalk et al. (2009) reference 30 modeling
parameters needed for the model. Each parameter is inputted as a probability density function.
Is the model linked with any corresponding input databases? Currently, the model is not linked
with any input databases, but potentially could be developed to utilize an input database.
Consideration of Uncertainty
How does the model account for uncertainty? A probabilistic approach is employed capture the
uncertainty of transfer coefficient values. Uncertainty in the model is addressed by assigning
probability density functions to determine transfer coefficient values. The (drawn) values are
used in the probabilistic methods detailed above to calculate PECs for the desired environmental
compartments.
Sensitivity analysis is conducted on the model to determine which input parameters have the
greatest influence on output variables. This analysis allows the risk manager more information of
the model at hand. Identifying the most influential parameter inputs will help the risk manager
prioritize which inputs should be studied in order to reduce the uncertainty of the model and
improve the overall exposure modeling process.
Availability and Usability - not applicable
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? This model was designed to specifically model
nanomaterials. Though the model does not explicitly model the relevant physical and chemical
properties needed to model nanomaterial behavior, it does employ probabilistic and Bayesain
approaches to capture these behaviors.
How does the model address input data gaps associated with many traditional models? The
model addresses an overall lack of data by utilizing a probabilistic and Bayesian approach to
develop a probabilistic material flow analysis to quantify the output variables. The model
parameters can be updated to account for relevant information regarding nanomaterial behavior,
thus more accurately describing the steady state of the system.
What kind of interpretations/predictions can be made from this model? This model allows for a
calculation of PECs in different environmental compartments in a steady-state system. From this
assessment, risk managers would be able to calculate risk quotients of nanomaterials by
obtaining predicted no effect concentrations PNECs.
140
-------
Appendix B
Alternative Approaches
Hock et al. (2008)
Summary: The precautionary matrix serves as a guide to determine the potential health concern
of nanorelavent materials in the absence of extensive data. Though this safety matrix in no way
replaces the traditional risk assessment methods, it does allow a pre-characterization of
potential risks and non-risk nanomaterials.
KEY REFERENCES:
Hock J., Hofmann H., Horner K., Krug H., Lorenz C., Limbach L., Nowack B., Riediker M.,
Schirmer K., Som C., Stark W., Studer C., von Gotz N., Wengert S., Wick P.: Guidelines on the
Precautionary Matrix for Synthetic Nanomaterials. Federal Office for Public Health and Federal
Office for the Environment, Berne 2008.
The PDF version of the guidelines can be downloaded from
http://www.bag.admin.ch/themen/chemikalien/00228/00510/05626/index.html?lang=en.
CONTACT/AVAILABILITY INFORMATION:
Jiirgen Hock (1 uergen.hoeck@temas.ch. Phone: +4171 446 50 30)
Federal Office of Public Health FOPH, CH-3003 Bern
PURPOSE AND SCOPE:
What is the model purpose? The precautionary matrix is a safety matrix developed for the Swiss
action plan for synthetic materials. The safety matrix is intended to estimate "nanospecific
precautionary need" of synthetic nanomaterials and their applications for employees, consumers,
and the environment. The model is intended to facilitate communication between all interest
groups, increasing responsibility in the development of nanotechnology. However, it should be
noted that the safety matrix does not substitute for an actual risk assessment. Rather than
evaluating the dangers and risks associated with specific nanoparticles, this model should only be
used to identify key areas for action (which may include an extensive risk assessment).
What processes are simulated? This model calculates values for nanorelevance, specific
framework conditions, potential effects of the nanomaterials, and potential human and
environmental exposure to nanomaterials.
What are the primary assumptions? The precautionary matrix operates under two assumptions:
(1) Treat all nanomaterials as if no investigations have been carried out for specific cases to
allow consistently objective evaluations; and (2) In the case of data gaps, use the worst case
scenario.
What transport media are considered? The precautionary matrix does not model the transport of
substance through media, rather estimates the potential risk of input into the environment.
141
-------
Appendix B
What spatial and temporal scales does the model consider? The precautionary matrix includes
the life cycle of the nanomaterial based on user inputs. The spatial and temporal scales are
specific to each nanomaterial, as well as the input of the user.
What are the forms of results produced? The precautionary matrix produces a risk value based
on a function of the nanorelevance, specific framework conditions, potential effect of the
nanomaterial, and potential human and environmental exposure of the nanomaterial. This risk
value is then classified as a low rating of need for nanospecific action or a need for nanospecific
action.
EVALUATION:
Background and History
How extensively has the model been used and applied? To our knowledge, the model has only
been used for the referenced study.
What verification and validation has been conducted? This novel approach has not been verified
by experimental or empirical data. Also, no other models have validated this construction.
Complexity
What physical and chemical properties are considered? The model considers redox activity,
catalytic activity and stability of the nanomaterial. However, these values are quantified as low,
medium or high and are not actually modeled.
What is the mathematical representation? The risk value of a given (nano)material is based on a
function of the four modeled processes given above. This estimation of precautionary need is
presented in section 5.2 of the guideline document where section 5.1 explains the sub-
calculations needed for this overall estimation. Section 5.3 then classifies the estimation of
precautionary need into two categories: (1) A score of 0-20 rates the nanospecific need for action
as low; and (2) A score >20 suggests a need for nanospecific action (existing measures should be
reviewed, further clarification undertaken, and measures to reduce the risk associated with
manufacturing, use and disposal implemented in the interests of precaution).
What are the data requirements? The data requirements for calculations of nano-specific risk are
presented in the table below. Along with the data characteristic are a description of the data
needed and the type of form the data is inputted as (e.g. probability distribution, point estimate,
ranking, etc.). The model input parameters are not extensive and, in the case of uncertainty, are
based on a worst case scenario.
142
-------
Appendix B
Data Characteristic Type of Input Description
Nanorelevance
Specific framework
characteristics
Potential effect
Yes/No given
values of 1 or 0
Yes/Partly/No
given values of
0,3, or 5
Low/Medium/High
given values of 1,
5, and 9
Potential Human and Various rankings;
Environmental {0 or 1}, or {1, 5,
Exposure or 9}
There are 5 parameters that determine the nanorelevance of a
material. These parameters are based on particle and agglomerate
sizes (e.g. does the material form agglomerate > 500nm?)
There are 4 parameters that quantify the risk surrounding specific
framework characteristics. These parameters quantify the data gaps
in the current knowledge about the material (e.g. Is the origin of the
starting materials known?)
There are 3 parameter values for the potential effect of a material
and they are based on the redox activity, catalytic activity, and the
stability of the material. For the given nanomaterial a score of
low(l), medium(5), or high(9) is given for each characteristic
above.
There are 9 parameter values to determine the potential human and
environmental exposure and depend on: (a) the physical
surroundings in production or application; (b) the contact per day;
(c) potential input into the environment. An example parameter is
frequency with which a consumer uses the nanomaterial product. It
is given corresponding values to monthly(l), weekly(5), or
daily(lO).
Table. Data Requirements.
Consideration of Uncertainty
How does the model account for uncertainty? The model accounts for uncertainty addressing the
data gaps in a separate variable in the overall risk calculation: specific framework conditions. If
information is not known about the nanomaterial, the risk value is increased. If it is determined
that the material is nanorelevant, and there is no knowledge about the material, then the
mathematical representation will generate a value that is classified as a "need for nanospecific
action" based solely on the specific framework conditions category.
Availability and Usability
Is a user-friendly interface available? The precautionary matrix excel form is downloadable
from the website. The cells are programmed to calculate the individual category rankings as well
as the overall risk value.
Is the documentation complete and transparent? Yes. Also, the embedded functions in the excel
sheet are accessible such that they can be evaluated to determine if the source code is consistent
with the documentation.
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? The precautionary matrix considers redox activity,
catalytic activity, and stability of nanomaterials. However, it does not model these properties, nor
simulate transport in the environment.
143
-------
Appendix B
How does the model address input data gaps associated with many traditional models? This
model is used to quantify a precautionary need when developing nanomaterials. Thus, in the
absence of extensive data, the precautionary matrix can serve as a tool to determine which
nanomaterials may need precautionary action in development. The ability to store information
and values in a precautionary matrix allows for an easy update of the matrix in the presence of
emerging, scientifically relevant information.
What kind of interpretations/predictions can be made from this model? The precautionary matrix
cannot be used to predict the risks of nanomaterials, however it can be used to evaluate whether
or not more information should be gathered on a particular nanomaterial and the process in
which it is used.
144
-------
Appendix B
Hansen et al. (2008)
Summary: This adaptive framework model permits classification of products containing
nanomaterials into categories of expected, possible, and no expected exposure, based on
information not typically used for risk assessment but more accessible. Using (a) information
from the Woodrow Wilson International Center pertaining to the location and concentration of
nanoparticles within consumer products and (b) best estimates available or worst-case
assumptions, the model estimates consumer exposure to selected nanomaterials. Though this
model is not designed specifically for the environmental modeling of nanomaterials, a similar
framework could be employed to generate preliminary estimates of environmental exposure.
KEY REFERENCES:
Hansen SF, Michelson ES, Kamper A, Borling P, Stuer Lauridsen F, Baun A (2008)
Categorization framework to aid exposure assessment of nanomaterials in consumer products.
Ecotoxicology 17:438-447.
CONTACT/AVAILABILITY INFORMATION:
Steffen Foss Hansen (sfha@env.dtu.dk. Phone +45 45251593)
Postdoc, Department of Environmental Engineering, Technical University of Denmark, DTU
Bygningstorvet, Building 113, room 072, 2800 Kgs. Lyngby, Denmark
PURPOSE AND SCOPE:
What is the model purpose? This model proposes a framework to aid in exposure assessment of
consumer products containing nanomaterials in the absence of traditional risk assessments. The
model is designed to group nanomaterial-containing consumer products into three separate
categories of: (1) expected to cause exposure; (2) may cause exposure; and (3) no expected
exposure to the consumer. The model is also designed to categorize nanomaterial-containing
products into groups pertaining to the: (a) chemical composition of the nanomaterial within the
product; (b) location of the nanomaterial within the product; and (c) the type or class of
nanomaterial within the product.
What processes are simulated? The only process simulated in this model is an exposure
assessment which is based on default values and equations taken from the Technical Guidance
Document (TGD) on risk assessment for existing substance (European Commission JRC 2003).
What are the primary assumptions? The primary assumption of this model is that exposure
assessment of products based on nanotechnology must take into account the location of the
nanomaterial in the product.
What transport media are considered? No transport media are considered in this model because
it is designed to predict exposure from the use of consumer products.
What spatial and temporal scales does the model consider? Spatial and temporal scales are not
included in this model. However, different life cycle stages of the consumer products are
145
-------
Appendix B
considered in the model to conduct exposure assessment (e.g. paint containing nanomaterials is
evaluated for when it is in liquid form and in dried form).
What are the forms of results produced? This model gives results of exposure as point estimates
in terms of mg kg"1 bw d"1.
EVALUATION:
Background and History
How extensively has the model been used and applied? To our knowledge, the model has only
been used for the referenced study.
What verification and validation has been conducted? There is no known verification or
validation of the model with experimental results or environmental data.
Complexity
What physical and chemical properties are considered? The only physical properties considered
are concentration of the nanomaterial within the consumer product. However, the model uses the
TGD on risk assessment for existing substances from the European Commission JRC (2003), so
it is possible that more properties are considered.
What is the mathematical representation? The model calculates exposure assessments by
estimating the quantity of active nanomaterial either: (a) on the skin per application of product
(lotions); or (b) inhaled per application of the product. This value is given as a point estimate
based on parameters such as chemical make-up of the nanomaterial, concentration of
nanomaterial in the product, location of the nanomaterial in the product, and type of
nanomaterial class. This value is then used to classify the product as: (1) expected to cause
exposure; (2) may cause exposure; and (3) no expected exposure to the consumer. The authors
stress that the estimated values should not be used as the basis for a risk assessment because
many of the information needed for the exposure assessment is unavailable.
What are the data requirements? The data requirements for the categorization framework are
presented in the table below. Along with the data characteristic are a description of the data
needed and the type of form the data is inputted as (e.g. categorical, point estimate, ranking,
etc.). The model input parameters are not extensive, however many of the parameters such as
concentrations of active substance within the product are likely unavailable for most nanoparticle
containing products so are estimated to generate different scenarios.
146
-------
Appendix B
Data Characteristic Type of Input Description
Chemical identity
Product Categories
Categorical
Categorical
Nanomaterial location Categorical
Nanomaterial type Categorical
Exposure assessment Point estimates
parameters
The nanomaetials are classified by the chemical composition of the
nanomaterial (e.g. silver, silica, etc.)
The products are categorized into 8 different types: appliances, food
and beverages, health and fitness, home and garden, automotive,
cross-cutting, electronics and computers, and goods for children
The 3 categories to characterize the distribution of nanomaterials
within the products are: bulk, surface, and particles.
The type of nanomaterials included one- or multiphase materials,
patterned- or unpatterned films, and particles that are surface bound,
suspended in liquids, suspended in solids, or free airborne particles.
These values include amount of product used per application,
respiratory rates, body weight, body area, and other parameters
applicable to the use of the product in calculation of exposure.
Table. Data Requirements.
Consideration of Uncertainty
How does the model account for uncertainty? The main approach for uncertainty is to use worst-
case assumptions or best estimates in the case of data gaps.
Availability and Usability - not applicable
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? The model does not consider many nanomaterial
properties. The only physical properties considered are the chemical composition of the actual
nanomaterial and the type of nanomaterial, which is presented in the data requirements section.
How does the model address input data gaps associated with many traditional models? The
model attempts to bypass traditional exposure assessment by using non-traditional parameters
such as location of the nanomaterial in the product and concentration of nanomaterial within the
product. However, some of this quantification was uncertain as well. The model design is
intended to be updated as more information concerning nanomaterial location and nanomaterial
concentration becomes available. However, this particular model is not intended for
environmental exposure, so does not account for transport properties of nanomaterials.
What kind of interpretations/predictions can be made from this model? Currently, the model is
intended for exposure rates to humans. However, it is feasible that the model could be modified
to calculate the expected exposure to the environment based on the parameters presented above.
Once the estimated levels are calculated, the results could be used with models for transport
within the environment. This might require an extensive amount of data though.
147
-------
Appendix B
Linkovetal. (2007)
Summary: The multicriteria decision analysis (MCDA) presented below develops a method in
which to prioritize several nanomaterials according to specified management objectives by
evaluating nanomaterials based on the criteria of potential health and ecological impacts,
societal importance, and stakeholder preference. By assigning relative importance weights to
the above criteria and ranking different nanomaterials according to these criteria, decision
makers can determine which nanomaterials to pursue (whether in production or research) based
on multiple avenues of evaluation.
KEY REFERENCES:
Linkov, I, Satterstrom, F.K., Steevens, J., Ferguson, E., Pleus, R. 2007. Multi-criteria decision
analysis and environmental risk assessment for nanomaterials. Journal of Nanoparticle Research
9: 543-554.
CONTACT/AVAILABILITY INFORMATION:
Igor Linkov (Igor.Linkov@usace.army.mil, Phone: 617-225-0812)
Research Scientist, US Army Engineer Research and Development Center, Concord, MA
PURPOSE AND SCOPE:
What is the model purpose? This model illustrated an example of MCDA application to the
problem of prioritizing several nanomaterials according to specified management objectives.
Three hypothetical alternative nanomaterials were considered, each with different social and
economic values as well as environmental properties and associated risks and benefits.
What processes are simulated? The processes simulated for this model are health and ecological
effects based on public health effects, occupational exposure-related effects, and environmental
effects, societal importance based on potential uses of the nanomaterials in manufacturing,
potential uses in consumer products, and availability of alternatives, and stakeholder preference
based on political, public, and scientific preferences.
What are the primary assumptions? The main assumptions of the model inherently deal with the
associated risks of corresponding model criteria.
What transport media are considered? Transport media were not considered in this model.
What spatial and temporal scales does the model consider? The model does not incorporate
spatial and temporal scales.
What are the forms of results produced? Based on their example, Nanomaterial #1 had the most
benefits for industry, while Nanomaterial #2 and Nanomaterial #3 were more environmentally
friendly, although the knowledge on the potential environmental risks and benefits of these
materials was very uncertain. In their example, Nanomaterial #2 scored the highest and was thus
the preferred alternative. Although Nanomaterial #1 might have been better economically and
148
-------
Appendix B
preferred socially, the decision process showed that Nanomaterial #2 was better overall because
of its favorable environmental effects.
EVALUATION:
Background and History
How extensively has the model been used and applied? To our knowledge, the model has only
been used for the referenced study. However, MCDA has been used for multiple applications.
What verification and validation has been conducted? There is no known verification or
validation of the model with experimental results or environmental data.
Complexity
What physical and chemical properties are considered? No properties are considered for this
alternative approach; however there are 9 sub-criteria evaluated for the calculation of the overall
score.
What is the mathematical representation? This model utilized the Analytic Hierarchy Process
(AHP), one of the most widely used tools to help decision makers assign weights. In AHP, the
category weightings were derived from a series of relative judgments in the form of a weightings
ratio. The original AHP algorithms require assignments of the value on the scale from 1 to 9,
while recent AHP adaptation allows incorporation of different scales including experimental and
measurement values (Figueira et al., 2005). In their example, they assigned the weighting
themselves.
Based on their relative weightings, AHP derived normalized weightings for the criteria. Sub-
criteria and measures were compared and weighted in a pairwise manner similar to that for the
main criteria. Once relative weightings were given for each of the sub-criteria, normalized
weightings were calculated for use in scoring different alternatives (see breakdown in Table 2).
The goal of the weighting process was to set absolute weights that reflect as closely as possible
the relative ratings input by the user. In AHP procedures, weightings are calculated by finding
the eigenvector corresponding to the highest eigenvalue of the weightings matrix. Other MCDA
methods may use different procedures to elicit/calculate weights.
What are the data requirements? The data requirements for quantification of the above simulated
processes values for the 9 sub-criteria needed to calculate the scores for each main criteria,
followed by an overall risk score. The model input parameters are not extensive and require
assignment from the decision maker. Input parameters were also assigned relative weighting to
aid in the quantification of the main criteria.
Consideration of Uncertainty
How does the model account for uncertainty? The model does not consider uncertainty because
the decision makers assigned all the weights and evaluation criteria values as point estimates.
149
-------
Appendix B
Availability and Usability - not applicable
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? The properties of nanomaterials were not included in
the model.
How does the model address input data gaps associated with many traditional models? The data
gaps posed by many traditional models were addressed by developing a ranking system for
nanomaterials based on the criteria of health effects, social importance, and stakeholder
preference. However, the model cannot be updated to account for emerging, scientifically
relevant information unless it changes the ranking values for each nanomaterial as they pertain to
health impacts, social importance, and stakeholder preference. For instance, if information
suggesting that nanomaterial #1 was extremely harmful to the environment, the ranking assigned
to that nanomaterial would be lowered.
What kind of interpretations/predictions can be made from this model? The model is designed to
rank the different nanomaterials across all evaluation criteria. From this ranking, risk managers
can make decisions on which nanomaterials to produce based on the relative health impacts,
social acceptance, and stakeholder preference. The risk manager can also rerun the model with
different weights to generate different scenarios, such that all angles of the development of
nanotechnology can be examined.
150
-------
Appendix B
Linkovetal. (2009)
Summary: This multicriteria decision analysis (MCDA) application addresses the problem of
prioritizing several nanomaterials according to the potential risk to the environment. The
MCDA attempts to incorporate several biological processes (bioaccumulation, bioavailability,
toxic effects) as well as physical and chemical properties of nanomaterials (agglomeration, size,
etc.) to categorize nanomaterials into groups of extreme-, high-, medium-, low-, and very low
risk. By employing Monte Carlo methods, the MCDA explores all feasible values for criteria
measurements and weights so that the robustness of nanomaterial categorization can be
assessed.
KEY REFERENCES:
Linkov I, Steevens J, Chappell M, Tervonen T, Figuera JR, Merad M. (2009) Classifying
Nanomaterial Risks Using Multi-Criteria Decision Analysis. In: Linkov I, Steevens JA (eds)
Nanomaterials: risks and benefits. Springer, Dordrecht, pp 179—191.
CONTACT/AVAILABILITY INFORMATION:
Igor Linkov (Igor.Linkov@usace.army.mil, Phone: 617-225-0812)
Research Scientist, US Army Engineer Research and Development Center, Concord, MA
PURPOSE AND SCOPE:
What is the model purpose? This model is designed to guide developers in nanomaterial research
and application as well as promote the safe use and handling of nanomaterials by employing the
use of a MCDA support system. The model uses performance metrics that define: (a) toxicity
and physico-chemical characteristics of nanomaterials; and (b) expected environmental impacts
through the nanomaterial-containing product life cycle. This model was explored because
traditional risk assessment models for chemical and biological materials may not be applicable to
nanomaterials for some time.
This model uses a stochastic multicriteria acceptability analysis (SMAA-TRI) to explore all
reasonable values for all model criteria parameters so as to group each product into specific risk
categories. The SMAA-TRI was chosen because of its ability to address uncertainty by utilizing
Monte Carlo simulations such that all possible parameter values and criteria weights could be
assessed for the decision making tool. The goal of this MCDA was to rank the alternatives rather
than select a single best alternative. In this case, the alternatives are the different nanomaterials
and are ranked to prioritize the materials that need further study based on risk potentials.
What processes are simulated? The model simulates biological processes based on
bioavailability potential, bioaccumulation potential, and toxic potential are given subjective
probabilities based on the nanomaterial characteristics/properties.
What are the primary assumptions? The main assumptions of the model inherently deal with the
associated risks of corresponding model criteria. For instance, a smaller-sized nanoparticle
represents higher risk.
151
-------
Appendix B
What transport media are considered? Actual transport is not considered in this model but
bioaccumulation potential considers the accumulation of particles absorbed from all sources of
soil, water, air, and food.
What spatial and temporal scales does the model consider? The model does not incorporate
spatial and temporal scales.
What are the forms of results produced? The results of this MCDA group nanomaterials into a
risk categorization of: extreme-, high-, medium-, low-, and very low risk. The analysis also
quantifies the percentage of scenarios that a particular nanomaterial fell into the relevant risk
category (figure 2 of the article).
EVALUATION:
Background and History
How extensively has the model been used and applied? To our knowledge, the model has only
been used for the referenced study. However, MCDA has been used for multiple applications.
What verification and validation has been conducted? There is no known verification or
validation of the model with experimental results or environmental data.
Complexity
What physical and chemical properties are considered? The model includes five properties of
nanomaterials relevant to fate and transport: (1) agglomeration; (2) reactivity/charge; (3) Critical
functional groups; (4) contaminant dissociation; and (5) particle size.
What is the mathematical representation? The intention of the model is to classify the alternative
nanomaterials into the categories of: (1) extreme risk; (2) high risk; (3) medium risk; (4) low
risk; or (5) very low risk. The results are shown in section 5 of the article.
What are the data requirements? The data requirements for the MCDA are presented in the table
below. Along with the data characteristic are a description of the data needed and the type of
form the data is inputted as (e.g. categorical, point estimate, ranking, etc.). The model input
parameters are not extensive, however many of the parameters such as bioavailability potential
are likely unavailable for most nanoparticle containing products so are estimated to generate
different scenarios
152
-------
Appendix B
Data Characteristic Type of Input Description
Extrinsic nanomaterial
characteristics
Biological processes
Criteria weights
Rates, The model includes data estimations for agglomeration rates,
categorizations, reactivity/charge, critical function groups, contaminant dissociation
and estimates in rates, and size. These values are in distributional form to capture the
distributional uncertainty surrounding the knowledge gaps concerning these
form. characteristics
Subjective Bioavailabilty potential, bioaccumulation potential, and toxicity
probabilities potential are based on the nanomaterial characteristics given above.
Probability The above inputs are given weights to develop multiple scenarios for
distributions a given nanomaterial. These weights are in distributional form to
capture the uncertainty surrounding the importance of the above
characteristics.
Table. Data Requirements.
Consideration of Uncertainty
How does the model account for uncertainty? The model accounts for uncertainty by conducting
numerical simulations by comparing the effect of changing parameter values and criteria
evaluations (weighting) on the modeling outcomes. Monte Carlo simulations quantify parameter
imprecision by drawing parameter values from specified probability distributions. This method
explores all feasible values for criteria measurements so that the robustness of nanomaterial
categorization can be assessed.
Availability and Usability - not applicable
Application to nanomaterial behavior
Does the model consider relevant chemical and physical properties to capture nanomaterial
behavior (specific to the media modeled)? The evaluation criteria are aimed to quantify the
relevant chemical and physical properties relevant to environmental transport of nanomaterials.
Although the transport of material is not modeled, the potential of transport is estimated based on
this quantification.
How does the model address input data gaps associated with many traditional models? This
model bypasses the data gaps associated with many traditional models not quantifying the actual
values of necessary data. Rather, the model allows the decision maker to rank the importance of
these data gaps. As emerging, scientifically relevant information becomes available, the decision
maker will be able to alter the uncertainty levels of the model, as well as assign appropriate
weights to the evaluation criteria. Thus, the decision makers will be able to make more accurate
decisions about which nanomaterials need more research pertaining to risk assessment.
What kind of interpretations/predictions can be made from this model? This model is designed to
rank the nanomaterials into different categories of potential risk. By categorizing them, risk
managers can prioritize the research needs for current nanotechnology.
153
-------
vvEPA
United States
Environmental Protection
Agency
Office of Research
and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
EPA/600/R-10/129
September 2010
www.epa.gov
Please make all necessary changes on the below label,
detach or copy, and return to the address in the upper
left-hand corner.
If you do not wish to receive these reports CHECK HERE
D; detach, or copy this cover, and return to the address in
the upper left-hand corner.
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Recycled/Recyclable
Printed with vegetable-based ink on
paper that contains a minimum of
50% post-consumer fiber content
processed chlorine free
------- |