l-HA United States
Environmental Protection
^*^1 Agency
Screening Methods for
Metal-Containing
Nanoparticles in Water
APM 32
RESEARCH AND DEVELOPMENT
-------
-------
EPA/600/R-11/096
September 2011
www.epa.gov
Screening Methods for
Metal-Containing
Nanoparticles in Water
APM 32
Prepared by
Edward M. Heithmar
U.S. Environmental Protection Agency
National Exposure Research Laboratory
Environmental Sciences Division
Environmental Chemistry Branch
Las Vegas, NV 89119
Although this work was reviewed by EPA and approved for publication, it may not necessarily reflect
official Agency policy. Mention of trade names and commercial products does not constitute
endorsement or recommendation for use.
U.S. Environmental Protection Agency
Office of Research and Development
Washington, DC 20460
-------
-------
Abstract
Screening-level analysis of water for metal-containing nanoparticles is achieved with
single particle-inductively coupled plasma mass spectrometry (SP-ICPMS). This method
measures both the concentration of nanoparticles containing an analyte metal and the mass of the
metal in each particle. SP-ICPMS is capable of sample throughputs of over twenty samples per
hour. In this report, the screening capability of SP-ICPMS is demonstrated in a study of
transformations of silver nanoparticles in surface water. Test water samples were collected from
two fresh water sites and two estuary sites. The effects of salinity, particle concentration,
particle size, and particle surface chemistry on relative rates of transformations were studied. At
1 i
high silver particle concentration (2.5x10 mL" ) shifts in the particle silver mass distribution
measured by SP-ICPMS indicated increased aggregation rate at high salinity, as reported by
others. However, at low silver particle concentration (2.5x10s mL"1), which is closer to expected
environmental concentrations, aggregation was minimal even in highly saline estuary water. At
the low concentration, a much more pronounced increase in either dissolved silver or silver-
containing nanoparticles too small to be distinguished from dissolved silver was observed.
These data were operationally defined in this report as "dissolved" silver. Comparison of
transformations of 50-nm and 100-nm silver indicated that rates of both aggregation and
apparent dissolution are higher for the smaller particles at the same particle concentration.
Transformations for citrate-capped and polyvinylpyrrolodone-capped silver nanoparticles were
similar.
li
-------
Notice
The United States Environmental Protection Agency's Office of Research and
Development partially performed and funded the research described here. Mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
111
-------
Contents
Abstract ii
Notice iii
List of Figures v
List of Tables vi
Abbreviations and symbols vii
Acknowlegements viii
Background 1
Rationale for research on metrology methods for metal-containing ENMs 2
Current methods for detecting, quantifying, and characterizing ENMs 4
Single particle - inductively coupled plasma mass spectrometry 6
Theory and Calibration Principles of SP-ICPMS 7
SP-ICPMS as a screening method for metal-containing nanoparticles in water 9
Experimental 10
Standards and reagents 10
Water sampling sites 10
Water sampling, sample handling, and storage 12
Transformation study procedures 13
SP-ICPMS analysis 13
Calculations 13
Results and Discussion 15
Sample water chemistry 15
Transformation of 50-nm citrate-capped silver ENM at high particle concentration 16
Comparison of transformations of 50-nm and 100-nm citrate-capped silver ENM 19
Comparison of transformations of PVP-capped and citrate-capped 50-nm silver ENM 21
Transformation of 50-nm citrate-capped silver ENM at low particle concentration 22
Conclusions and Future Work 25
References 26
iv
-------
List of Figures
Figure 1. Fresh water sampling site locations, with latitude and longitude 11
Figure 2. Water samples as received from the sampling sites before filtration 11
Figure 3. Estuary water sampling site locations, with latitude and longitude 12
Figure 4. Mass-based mean particle mass of citrate-capped 50-nm Ag over time in deionized
1 i
water and low- and high-salinity estuary waters (2.5x10 mL" ) 16
Figure 5. Polydispersity index of citrate-capped 50-nm Ag over time in deionized water and low-
1 i
and high-salinity estuary waters (2.5x10 mL" ) 17
Figure 6. Change in measured "dissolved" silver over time in suspensions of citrate-capped 50-
1 i
nm Ag in deionized water and low- and high-salinity estuary waters (2.5x10 mL" ) 18
Figure 7. Change in total measured silver over time in suspensions of citrate-capped 50-nm Ag
1 i
in deionized water and low- and high-salinity estuary waters (2.5x10 mL" ) 18
Figure 8. Increase in PDI over time for suspensions of 50-nm and 100-nm citrate-capped Ag in
1 i
high-salinity estuary water (2.5x10 mL" ) 20
Figure 9. Increase in "dissolved" silver over time for suspensions of 50-nm and 100-nm citrate-
1 i
capped Ag in high-salinity estuary water (2.5x10 mL" ) 20
Figure 10. Increase in PDI over time for suspensions of PVP-capped and citrate-capped 50-nm
1 i
Ag in high-salinity estuary water (2.5x10 mL" ) 21
Figure 11. Increase in "dissolved" silver over time for suspensions of PVP-capped and citrate-
1 i
capped 50-nm Ag in high-salinity estuary water (2.5x10 mL" ) 22
Figure 12. Changes in suspension metrics of 50-nm citrate-capped silver after 1400 minutes
(2.5\105 mL"1) 23
Figure 13. Effect of concentration on changes in suspension metrics of 50-nm citrate-capped
silver after 1400 minutes (2.5x10s mL"1) 24
v
-------
List of Tables
Table 1. Ranking (in descending order) of 14 key research priorities for eco-responsible ENM
design 2
Table 2 Quantification and characterization metrics produced by SP-ICPMS and by hyphenated
analytical methods 7
Table 3. Size, mass, and particle-concentration metrics for silver nanosphere standards 10
Table 4. Measured water chemistry parameters for water samples used in transformation studies.
15
vi
-------
Abbreviations and symbols
AFM Atomic force microscopy
APHA American Public Health Association
ca Analyte concentration in aqueous sample (g/mL)
cp Particle concentration in aqueous sample (mL"1)
DI Deionized
DLS Dynamic light scattering
EDS Energy-dispersive X-ray spectrometry
sn Nebulization transport efficiency (dimensionless)
ENM Engineered nanomaterials
EPA Environmental Protection Agency
FFF Field flow fractionation
Flow-FFF Flow field flow fractionation
HDC Hydrodynamic chromatography
ICON International Council on Nanotechnology
ICPMS Inductively coupled plasma mass spectrometry
IRZ Initial radiation zone
LDPE Low-density polyethylene
ma p Analyte element mass in the particle (g)
n. Number of analyte ions detected (demensionless)
NOM Natural organic matter
NTA Nanoparticle tracking analysis
PDI Polydispersity index (dimensionless)
PSU Practical salinity units (dimensionless)
PVP Polyvinylpyrrolodone
qLa Ionized analyte flux (s"1)
qp Particle flux (s"1)
qs Sample uptake rate (mL s"1)
rg Radius of gyration
rh Hydrodynamic radius
RSD Relative standard deviation
Sed-FFF Sedimentation field flow fractionation
SEM Scanning electron microscopy
SLS Static light scattering
SP-ICPMS Single particle - inductively coupled plasma mass spectrometry
TEM Transmission electron microscopy
WDS Wavelength-dispersive X-ray spectrometry
C, Zeta (potential; usually mv)
vii
-------
Ackiiowlegemeiits
I thank Tara Schraga of the U.S. Geological Survey and David Carpenter of the Southwest
Florida Water Management District for collecting the water samples, and Jay Kuhn of Analytical
Resources, Inc. for access to the PerkinElmer NexION ICPMS. Chady Stephan of PerkinElmer
conducted some of the experiments with me and participated in valuable discussions regarding
the data. Emily Siska, a student contractor for the Environmental Protection Agency, provided
much needed laboratory support.
viii
-------
Background
Engineered nanomaterials (ENMs) are increasingly being incorporated in industrial,
consumer, medical, and agricultural products. This is because ENMs exhibit unique optical,
electrical, and chemical properties that can impart beneficial characteristics to the product into
which they are incorporated. However, these unique properties also affect the environmental
behavior of ENMs. Their transport, fate, exposure potential, and effects are not predicted by
those of either the corresponding bulk materials or dissolved chemicals.
Most ENMs currently in production can be categorized as either metal-containing ENMs
(i.e., metals, metal oxides, or semiconducting quantum dots) or carbon-based (i.e., fullerenes and
their derivatives, and carbon nanotubes). ENMs containing metals have a greater potential to
enter the environment than carbon-based ENMs. This is a result of the fact that the major uses of
metal-containing ENMs are in dispersive applications, while carbon-based ENMs are generally
incorporated into solid composites. This increased exposure potential for metal-containing
ENMs has motivated intense research into their environmental processes, such as transformation,
transport and fate, exposure pathways, and potential adverse effects on humans and sensitive
organisms.
Detection, quantification, and characterization of ENMs, including measurement of
nanoparticle concentration and characterization of particle size distribution, are critical to all
aspects of this exposure research. Nanoparticle concentration and size distribution largely
control how they behave in the environment. Highly selective detection, quantification, and
characterization methods are important for many types of nanomaterials environmental research;
less selective methods that can rapidly screen samples for metal-containing nanoparticles are also
needed.
This report focuses on the application of a new method, single-particle-inductively coupled
plasma mass spectrometry (SP-ICPMS), for rapid screening-level measurement of nanoparticle
dispersions. The method is being developed at the Environmental Sciences Division of EPA's
National Exposure Research Laboratory. SP-ICPMS has demonstrated promise as a practical
analytical method for characterization of metal-containing ENMs in environmental waters. This
report will: (1) briefly review of the role of characterization of metal-containing ENMs in the
exposure research of these novel materials; (2) discuss methods currently available for the
measurement of various ENM exposure metrics; and (3) describe the limitations of existing
methods for studying ENMs in real environmental systems. A brief review of the theory of SP-
ICPMS will also be provided. A study that applies SP-ICPMS in a stand-alone screening-level
mode will be described in detail. Specifically, SP-ICPMS will be used to study transformations
of silver nanoparticles in surface water.
1
-------
Rationale for research on metrology methods for metal-containing
ENMs
A workshop on the eco-responsible design and development of ENMs was conducted by
the International Council on Nanotechnology (ICON) in 2009 (Alvarez, Colvin et al. 2009).
Fourteen key research priorities that would inform designing, using and disposing of ENMs to
enhance responsible development of these materials were identified by the approximately 50
researchers invited to the workshop. The fourteen priorities were then ranked against two
criteria. First, the relative importance of the research in understanding how to design and
develop ENMs in an eco-responsible manner was assessed. Second, the current gap in the state-
of-knowledge for each research priority was ranked. The workshop also assessed the amount of
effort required to sufficiently satisfy the science gap posed by each research priority. The results
of these three assessments are summarized here (Table 1).
Table t, Ranking (in descending order) :ey research priorities for eco-
responsible ENM design (adapted from Alvarez, Colvin et al. 2009).
Importance for eco-responsible design
Gap in state-of-knowledge
Effort needed
Metrology, analytical methods
Metrology, analytical methods
High
Predictive models of release
Structure-activity relations
High
Structure-activity relations
Boavailability and bioaccumulation
Medium-high
Dose-response (sub-lethal)
Sources/environmental fluxes
Medium
Boavailability and bioaccumulation
Trophic transfer
Medium
Identify relevant sentinel organisms
Uptake mechanisms
Low
Trophic transfer
Intra-organism distribution
Low
Industrial ecology/green chemistry
Industrial ecology/green chemistry
Medium
Sources/environmental fluxes
Impact on environmental infrastructure
Medium
Impact on environmental infrastructure
Predictive models of release
High
Uptake mechanisms
Assessing regulatory framework
Medium
Assessing regulatory framework
Dose-response (sub-lethal)
Medium-high
Intra-organism distribution
Waste minimization/recycling
Medium
Waste minimization/recycling
Identify relevant sentinel organisms
Low
Metrology to detect, quantify, and characterize ENMs ranked highest in importance for
enabling eco-responsible design of ENMs. This area also ranked as least developed in terms of
state-of-knowledge. Consequently, the workshop recommended a great deal of effort be applied
in metrology to detect, quantify, and characterize ENMs.
The workshop's emphasis on the need for research on detection, quantification, and
characterization of ENMs stems from a number of factors. First, these metrology tools are vital
for the success of virtually every area of research on the environmental behavior, exposure
potential, and possible adverse effects of ENMs. Monitoring environmental occurrence and
distribution of ENMs, as well as determining temporal trends in these data, requires metrology
methods. Process research into ENM transformation, transport, and fate requires methods to
measure and characterize ENMs in each compartment of a laboratory system. Toxicity research
requires quantifying and characterizing the ENMs in the original dosing material, and possible
changes in concentration and size distribution induced by the test system must also be measured.
2
-------
Often, toxicity testing requires measuring internal ENM dose in the test organisms. Finally,
emerging exposure models for ENMs (Mueller and Nowack 2008) require quantification of
ENM releases and ENM measurements in the environment for model validation.
The low level of the state-of knowledge for detection, quantification, and characterization
of ENMs is based in part on the complex set of metrics that need to be measured. Compound
concentration is often the only relevant metric for conventional pollutants in evaluating exposure
potential, because these are released and transported in dissolved form. Conversely, for ENMs,
in addition to mass concentration, measurement of particle concentration and particle size
distribution is usually required, because the environmental stability and mobility of ENMs are
influenced by their particle size. There are also several ways of measuring particle size that may
be important depending on the application, including particle mass, volume, and hydrodynamic
diameter [the theoretical diameter of an equivalent spherical particle and its electric double layer
(Hassellov, Readman et al. 2008). In addition to mass concentration, particle concentration, and
size distribution, other metrics sometimes affect environmental behavior of ENMs. Surface area
affects reaction rates of ENMs; therefore, their catalytic activity, and surface area can also
mediate ENM toxicity (Schulte, Geraci et al. 2008). Surface charge or zeta potential (Q affects
the inter-particle repulsive forces of ENM suspensions, which influences their tendency to
aggregate (Kim, Lee et al. 2008). ENM aspect ratio, the ratio of the longest to the shortest
dimension of the nanoparticles, sometimes influences toxicity of long and narrow ENMs such as
carbon nanotubes (Takagi, Hirose et al. 2008). Therefore, in nearly all research efforts related to
the environmental behavior, exposure potential, and possible adverse effects of ENMs, methods
are required for quantifying mass and particle concentration, and at least one metric related to
size distribution. Other metrics such as surface area, surface charge, and particle shape may also
be required, depending on the environmental behavior or effect being studied.
There is a paucity of practical methods for detecting, quantifying, and characterizing
nanomaterials in complex media. For ENMs in pure suspensions, there are several methods
available to quantify mass and particle concentrations, as well as characterize size distributions.
For example, characterizing ENM starting materials used in laboratory studies is often relatively
straightforward. However, the test systems in which the starting materials are studied often alter
these metrics, and there are few methods for measuring them in these more complex systems
(Alvarez, Colvin et al. 2009). Natural environmental samples are usually even more analytically
challenging than laboratory test systems, and no practical methods have yet been published for
detecting, quantifying, and characterizing ENMs in environmental media (Handy, von der
Kammer et al. 2008). Several common properties of natural environmental media contribute to
the dearth of applicable ENM metrology tools for these sample types. Natural systems generally
have very low concentrations of ENM particles (Kiser, Westerhoff et al. 2009), so the sensitivity
of many analytical techniques is insufficient. In addition, natural colloids containing the analyte
element used to detect the ENM are often present (Klaine, Alvarez et al. 2008). In the case of
metal-containing ENMs, most analytical techniques measure total element concentration, and
many natural colloids include minerals that contain a wide range of metals, so background
interference is an issue. Finally, particle size determination can be confounded by dissolved
metals adsorbed to relatively large natural organic matter (NOM) particles.
3
-------
Methods for a wide range of characterization metrics might be developed for simple
systems such as laboratory media; however, it is unlikely that methods for many ENM
characteristics, such as surface charge and aspect ratio, can be developed for natural systems.
However, methods must be developed that can at least selectively detect, and quantify ENM
particle concentration, as well as characterize the size distributions of metal-containing ENMs, in
natural environmental media. In addition, screening-level methods must be developed that will
allow rapid analysis of samples for potential ENM content. These screening techniques will be
required because selective techniques will likely involve time-consuming separations, and
therefore will not be applicable to large numbers of samples. In addition, rapid screening-level
techniques could be used to study environmental processes, such as aggregation and dissolution,
that occur on time scales too rapid for separation-based techniques.
Current methods for detecting, quantifying, and characterizing ENMs
Imaging techniques are currently the most common methods for characterizing the size
and shape of ENMs (Lin and Yang 2005; Pyrz and Buttrey 2008). Either scanning electron
microscopy (SEM) or transmission electron microscopy (TEM) can be employed. In some
studies, atomic force microscopy (AFM) can also be used (Ebenstein, Nahum et al. 2002). In
some cases, AFM can provide additional information such as adsorption forces. SEM or TEM
coupled with X-ray spectrometry, in either the energy-dispersive (EDS) or wavelength-
dispersive (WDS) modes, can definitively identify metal-containing nanoparticles. However,
particle concentrations must be high (generally >109 mL"1) to reliably find nanoparticles using
any of the imaging methods. Also, the lack of representative sampling techniques for imaging
methods precludes quantification of particle concentration. Current sampling methods can also
produce changes in the size distribution of the ENM (Tang, Wu et al. 2009). Finally, imaging
techniques are generally not practical for environmental characterization in natural media
because of background colloids. In environmental media, metal-containing ENMs cannot be
distinguished from colloid particles containing the same metal (Tiede, Hassellov et al. 2009).
Representative nanoparticle size distributions can be obtained by a variety of light
scattering methods, because they rely on measuring a signal that is produced by the collection of
all ENM particles in a large volume of sample, in contrast to the minute volume sampled by the
imaging methods. ENM size distribution by light scattering methods is also possible at lower
concentrations (ca. 106-107 mL"1) than allowed by imaging techniques. Dynamic light scattering
(DLS) is the most prevalent light scattering characterization tool. It uses the autocorrelation
function of scattered laser light to calculate the hydrodynamic radius (rh) distribution (Filella,
Zhang et al. 1997). DLS can also measure C, potential when performed on a sample in an
oscillating electric field. A second light scattering method, static light scattering (SLS), also
known as multi-angle light scattering, measures the radius of gyration (rg) (Kammer, Baborowski
et al. 2005). It has been demonstrated that the ratio rg/rh can provide information on ENM aspect
ratio (Schurtenberger, Newmen et al. 1993). DLS and SLS are most applicable for ENM
suspensions with narrow size distributions. Light scattering intensity is related to particle
diameter in a non-linear fashion, and in more polydisperse samples, scattering by large particles
distorts the calculation of the size distribution. In contrast to DLS and SLS, nanoparticle
4
-------
tracking analysis (NTA) measures the scatter from individual particles. Therefore, NTA can
accurately determine size distributions in polydisperse suspensions. None of the light scattering
techniques are element selective. They measure the size distribution of all nanoparticles
regardless of composition. Therefore, light scattering methods are not applicable in complex
natural media.
To detect, quantify, and characterize metal-containing ENMs in natural media, an
element-selective detection method is required to reduce interference from background natural
colloids. Inductively coupled plasma mass spectrometry (ICPMS) is by far the most common
such detection approach, because of its high sensitivity. However, most element-selective
detection techniques, including ICPMS, cannot characterize particle size distribution. This has
led to the increasing use of hyphenated analytical methods, where a size separation, usually by
some form of chromatography, is coupled on-line to ICPMS. Recently, the most common size
separation technique has been a form of field flow fractionation (FFF). FFF comprises several
variants. In each, a laminar flow of an eluent (usually an aqueous surfactant solution) carries a
plug of sample down a narrow channel. A force field is applied orthogonal to the laminar flow,
forcing nanoparticles toward the channel wall. Size separation is effected by the balance
between this force and the different diffusivities determined by particle sizes. Smaller particles
diffuse higher, into a faster portion of the laminar-flow profile. Therefore, smaller particles elute
before larger particles. The intensity vs. retention time profile produced by FFF is called a
fractogram. A number of force fields can be used. The FFF technique most commonly used for
size separation of ENMs is flow-FFF (Lesher, Ranville et al. 2009), where the orthogonal force
is produced by a flow through a channel wall that is porous to eluent but not to nanoparticles. A
gravity field has been used less often in a technique known as sedimentation-FFF (or sed-FFF).
The two forms of FFF are complementary in that separations are produced by different particle
properties - rh in flow-FFF, and buoyant mass in sed-FFF. While sed-FFF affords higher size
resolution, flow-FFF is easier to implement and is applicable to a wider range of ENM particle
sizes: hence, the greater popularity of flow-FFF. Flow-FFF-ICPMS does have an experimental
complication, in that nanoparticle interactions with the porous membrane wall can lead to peak
tailing and even irreversible adsorption.
Another hyphenated analytical approach for ENM size characterization uses a separation
technique known as hydrodynamic chromatography (HDC) (Tiede, Boxall et al. 2010). In HDC-
ICPMS, the ENM particles pass through a column packed with a material such as silica in a
eluent under laminar-flow conditions. Small nanoparticles can approach the particle surfaces
closer than large particles, therefore experiencing a lower average flow velocity. The fractogram
of HDC-ICPMS is inverted compared to that of FFF-ICPMS: small particles elute after larger
particles in HDC-ICPMS. FFF-ICPMS offers higher resolution, while HDC-ICPMS often
effects a more rapid separation.
Both flow-FFF-ICPMS and HDC-ICPMS provide elemental information in addition to
size distribution. Either can be applied to moderately complex systems like laboratory test
media. However, the application of either for natural environmental media is limited.
Fractograms measure the total analyte metal concentration flowing to the ICPMS. Therefore,
neither FFF-ICPMS nor HDC-ICPMS can determine the chemical state of the metal. For
example, a 10 ng/mL concentration of silver detected at a retention time related to a given n,
5
-------
could be produced by 10 ng/mL silver ENM, but it could also result from approximately 130
ng/mL of AgCl nanoparticles having the same rh. Furthermore, a high concentration of NOM
with the same rh could adsorb dissolved silver and produce a similar signal at the same retention
time. Therefore, these hyphenated methods provide an upper-limit for the potential
concentration of ENMs in natural media, but not a definitive determination of ENMs.
Single particle - inductively coupled plasma mass spectrometry
The limitations of hyphenated approaches to metal-containing ENM size characterization
when applied to natural media have encouraged development of single particle - inductively
coupled plasma (SP-ICPMS). This technique was introduced by Degueldre's group to
characterize various colloids (Degueldre and Favarger 2004; Degueldre, Favarger et al. 2004),
and later gold nanoparticles (Degueldre, Favarger et al. 2006). Recently, Ranville's group
described a preliminary application of SP-ICPMS to characterize silver nanoparticles in
municipal waste water (Monserud, Lesher et al. 2009), and Hassellov coupled SP-ICPMS with
flow-FFF (Hassellov 2009).
The principle underlying SP-ICPMS is simple. Metal-containing nanoparticles entering
an ICPMS plasma produce discrete ion plumes of the analyte metal isotopes over short time
periods (e.g., < 1 ms). If the ICPMS signal is monitored with high temporal resolution (e.g., <10
ms per data point) background in each data point from dissolved analyte metal or plasma matrix
ions (Lam and Horlick 1990) diminishes to a very low average ion count per data point. The ion
plume pulses from the nanoparticles produce several ions in a single data point, making them
easily distinguishable. The metal-containing nanoparticle concentration in the sample is
obtained by measuring the frequency of the ion plume pulses. The analyte metal mass contained
in each individual nanoparticle is independently calculated by the ion intensity of its
corresponding ion plume.
The quantification metrics produced by SP-ICPMS are complementary to those produced
by hyphenated analytical methods like FFF-ICPMS and HDC-ICPMS (Table 2). Either SP-
ICPMS or a hyphenated method, used as stand-alone techniques, can only be a screening
technique giving an upper bound for the concentration of metal-containing ENMs. Although SP-
ICPMS measures the metal of interest in each particle, it provides no direct measurement of the
particle diameter. Using the example of silver ENM again, detection of SP-ICPMS pulses of a
given intensity could be caused by silver ENM nanoparticles of a certain diameter, or they could
be produced by larger nanoparticles of insoluble silver salts, such as AgCl or AgS. Because of
the complementary nature of the metrics produced by SP-ICPMS and hyphenated analytical
methods, FFF-SP-ICPMS or HDC-SP-ICPMS could provide very selective detection,
quantification, and size characterization in natural environmental media.
6
-------
Table 2 Quantification and characterization metrics produced by SP-ICPMS and by
hyphenated analytical methods.
SP-ICPMS
Measures the particle concentration of metal-
based nanoparticles, as well as the mass of
metal in each particle.
Does not provide direct information on particle
diameters.
Hyphenated Analytical Methods
Measures total metal concentration as a
function of nanoparticle size fraction.
Does not provide information on number or
characteristics of metal-based particles.
Theory and Calibration Principles of SP-ICPMS
A complete understanding of the theory of SP-ICPMS requires consideration of some
general processes that affect signal in all forms of ICPMS, including conventional ICPMS of
analyte in solution. The processes are the same in conventional and single particle
implementations of ICPMS. Partly because of these common processes, dissolved analyte
measurements can be used in the calibration of particle element mass metric determined by SP-
ICPMS.
In conventional ICPMS of an aqueous sample, an aerosol of droplets is produced by
some form of nebulizer (Taylor 2001). Large aerosol droplets (> 5-10 [j,m in diameter) do not
contribute to analyte signal and cause noise in the plasma; therefore, a spray chamber is usually
employed to remove these large droplets by collisions with the chamber walls. This results in a
fine aerosol entering the plasma, the droplet flux of which is typically greater than 106 s"1. At
ng/mL analyte concentrations, each droplet contains less than a few thousand analyte atoms. The
efficiency of the total sampling process that results in analyte-containing aerosol in the plasma is
termed the nebulization transport efficiency, sn. This quantity is usually between 2% and 30%,
depending on specific sample introduction system and operating conditions. Aerosol is
evaporated in the region within the ICP load coil known as the preheating zone of the plasma.
The salts containing the analyte element vaporize in this region. Atomization occurs in the
region from zero to several millimeters downstream from the load coil. This region is the initial
radiation zone (IRZ) (Koirtyohann, Jones et al. 1980). Downstream from the IRZ is the normal
analytical zone; ionization occurs in this region (Thomas 2004). In a typical ICP, vaporization,
atomization, and ionization processes once aerosol is in the plasma are usually >80% efficient
for most elements (O'Connor and Evans 1999). This is true for both conventional ICPMS and
SP-ICPMS
ICPMS of nanoparticles differs from conventional ICPMS of dissolved analyte because
of the way analyte is distributed within the aerosol. In conventional ICPMS, a each droplet
contains a small number of analyte atoms. The large number of droplets results in a relatively
7
-------
constant flux of ionized analyte; therefore, a fairly constant signal is produced. Typically, the
sampling time per data point (called the dwell time) is set at >100 ms, to minimize signal
variation due to counting statistics or noise in the sampling process. In the case of ICPMS of
nanoparticles, the great majority of aerosol droplets contain no nanoparticles. If there is any
dissolved analyte in the sample, it is distributed as in conventional ICPMS, and it contributes to a
fairly constant background. Only a small fraction of droplets contain a nanoparticle at low ENM
concentration, and generally only one nanoparticle is contained in a droplet. However, each
nanoparticle can produce a plume of millions of ions that enter the mass spectrometer interface
over a period of about 500 |is (Gray, Olesik et al. 2009; Heithmar 2009). The resulting signal is
composed of the low constant background with periodic pulses of large numbers of analyte ions
detected. In contrast to conventional ICPMS, SP-ICPMS is implemented with very short dwell
times to improve the contrast between signal produced by dissolved analyte and the large, fast
pulses from analyte in nanoparticles. For example, if the dwell time is set at 1 ms, a 1 pg/mL
dissolved silver background would typically result in an average background signal of 0.1-0.5
ions in a single dwell period. By contrast, a 50 nm silver nanoparticle containing about 6.2x10"16
g silver (3.4xl06 atoms) would typically produce a signal of 20-50 ions in the same 1 ms dwell
period (transmission efficiencies of quadropole mass filters are typically ca. 10"5).
This theory of SP-ICPMS results in an easily calibrated signal if two assumptions are
met. First, every nanoparticle that reaches the plasma must be detected as an ion plume. This
requires a sufficiently long residence time, so the ion plume expands enough to substantially fill
the cross section of the central channel of the plasma. If so, equation 1 is valid.
(1) qp/cp - qs sn,
where qp = flux of particles detected in plasma (s"1), cp = concentration of nanoparticles
11
containing the detected metal in the sample (mL" ), qs = sample uptake rate (mL s" ), and sn =
nebulization efficiency (dimensionless). Note that qs and sn are properties of the ICPMS
instrument conditions and independent of the element. Therefore, equation 1 can be used to
calculate their product using any type of nanoparticle suspension of known cp.
The second assumption of quantitative SP-ICPMS calibration theory is that ICPMS
sensitivity is constant for an analyte, irrespective of whether it is dissolved or contained in
nanoparticles. Again, this requires that the residence time in the plasma to be sufficiently long.
If so, equation 2 is valid.
(2) nia,P — [qs £n Ca / qi,a] Hi,P — k n j p,
where map = mass of analyte element in a single nanoparticle (g), nLp = number of ions of
analyte element detected in the corresponding plume (number of ions detected in a single SP-
ICPMS pulse), ca = the analyte concentration in a dissolved standard of the analyte (g mL"1), qLa
= ion flux measured for the dissolved standard (s"1). For each analyte element, calibration of the
element mass in individual particles (calculation of the response factor k) requires only the qs sn
product from Equation 1 and analysis of a known concentration of dissolved analyte element
using a conventional ICPMS standard).
8
-------
Equation 2 provides calibration of nanoparticle element mass. The calibration of
nanoparticle concentration is provided by rearrangement of equation 1 for any unknown
nanoparticle suspension, once qs sn has been determined:
(3) Cp - qp / qs 8n-
SP-ICPMS as a screening method for metal-containing nanoparticies
in water
As a screening tool for metal-containing nanoparticies in natural media, stand-alone SP-
ICPMS is potentially more useful than hyphenated methods of analysis. Analysis by SP-ICPMS
can be applied to large numbers of samples, because it can achieve throughputs of over twenty
samples per hour. By comparison, typical hyphenated analyses can take 20-40 minutes each.
Also, SP-ICPMS has less potential for artifacts due to ENM adsorption. Finally, detection limits
of ENMs by SP-ICPMS are potentially much lower than those that have been obtained by flow-
FFF-ICPMS.
In addition to the ability to analyze large sample sets, the speed of screening-level SP-
ICPMS also allows rapid transformations of metal-containing ENMs to be studied. The particle
size distribution must not change significantly during the time required for size distribution
analysis, which limits the hyphenated sizing methods to relatively slow transformations. In
contrast, batch analysis by SP-ICPMS can be easily applied to studying size-distribution
transformation processes with characteristic times of 2-3 minutes, and flow-injection
experiments can be designed to study even faster processes.
This report will describe the application of SP-ICPMS as a stand-alone screening method
to study time-resolved size-distribution transformations of silver ENM suspensions in natural
surface waters. Size distribution is a critical parameter in predicting the mobility of
nanoparticies, because larger particles fall out of the water column more quickly than small
particles. The size distributions of ENMs are known to be affected by water chemistry,
particularly by ionic strength (as determined in most surface waters by salinity) and NOM
(French et al. 2009; Liu et al. 2010). In the study described in this report, the effects of water
chemistry (particularly salinity), nanoparticle concentration, nanoparticle size, and nanoparticle
surface chemistry on the size-distribution will be measured as a function of exposure time. Four
surface waters of divergent water chemistry will be studied. Size distributions at exposure times
of two to 3000 minutes will be measured. Transformation processes will be studied at both a
1 i
high particle concentration typical of previously published studies (2.5x10 mL" ) and a more
environmentally relevant 2.5x10s mL"1. Transformations of silver nanoparticle suspensions with
mean diameters of 50 nm and 100 nm will be investigated. Finally, transformation of citrate and
polyvinylpyrrolodone (PVP) capped silver nanoparticies will be compared.
9
-------
Experimental
Standards and reagents
The purchased silver nanosphere standard suspensions were NanoXact™
(Nanocomposix, San Diego, CA). The suspensions were quite monodisperse (RSD 8-10%) and
the total silver concentration of each was 20 mg/L. The mean silver nanoparticle mass in
femtograms and the suspension particle concentration were calculated from the mean
nanoparticle diameter determined by the manufacturer by TEM (Table 3). Dilutions were made
with laboratory deionized water.
Table 3, Size, mass, and particle-concentration metrics for silver nanosphere
standards,
Nanoparticle
(capping agent)
Mean Diameter
(nm)
Diameter RSD
(%)
Ag nanoparticle
mass (fg)
Particle concentration
(mL1)
50 nm (citrate)
49.1
9.1
0.65
3.1xl010
50 nm PVP
53.4
9.4
0.84
2.3xl0lu
100 nm (citrate)
99.1
8.3
5.30
3.7xlOy
Water sampling sites
Sites for the silver transformation studies were selected with the intention of getting a
range of salinities and NOM. One lake, one river, and two estuary sites were selected as water
sources for the transformation study. The lake site was at Lake Hancock and the river site was
on the Alafia River, both in west-central Florida (Figure 1). Lake Hancock is heavily impacted
by agricultural runoff and its unfiltered sample was highly colored with algae. The Alafia River
is typical of central Florida rivers, having high tannin content and a brown unfiltered sample
color (Fgure 2).
The two estuary sampling sites were both located on the San Francisco Bay and are
regular sampling locations (Stations 7 and 31) of the fresh water sampling site locations, with
latitude and longitude.of the U.S. Geological Survey (USGS 2011). Station 7 is isolated from
the mouth of the Bay by narrow straights, while Station 31 is in the wide lower arm of the Bay
with direct access to the mouth of the Bay (Figure 3). Therefore, salinity at Station 7 is generally
lower than at Station 31.
10
-------
Figure 1. Fresh water sampling site locations, with latitude and longitude.
27 9908 -81 8493 '*>
' a Lake Hancock
Tampa, FL
: ::
Alalia River
'Pata"SK^NO'AA.»U. S jjNavy, NGA, GEBCO
".G
-------
Figure 3. Estuary water sampling site locations, with latitude and longitude.
Water sampling, sample handling, and storage
The fresh water samples were collected on April 10, 2010. They were collected as a
single grab sample by opening an empty 1-L low-density polyethylene (LDPE) bottle
approximately 0.3 m below the surface of the water and capping the completely filled bottle
before bringing it to the surface. The estuary samples were collected at a depth of 1 m via the
sampling pump on the USGS research ship and collected in 1-L LDPE bottles. Salinities were
measured on-board with a salinity meter in practical salinity units (PSU).
Water samples were maintained at ca. 4 °C until analysis. Samples were filtered through
a 20 (im polypropylene membrane (Polycap 36 HD, Whatman, Inc., Florham Park, NJ)
immediately upon arrival at the laboratory. Aliquots of the filtered samples samples were
analyzed for water chemistry parameters at TestAmerica, Phoenix, AZ. The water chemistry
tests were run within two weeks of the transformation studies. Total organic carbon was
determined by Method SM5310 B [Standard Methods For The Examination of Water and
Wastewater; American Public Health Association (APHA)]; metals were determined by Method
12
-------
6010 [U.S. Environmental Protection Agency (EPA)]; anions were determined by EPA Method
300.0; suspended solids were determined by APHA Method SM2540 D.
Transformation study procedures
Water samples were allowed to reach room temperature (23-25 °C). 20 mL of each
sample was transferred to a 50-mL polypropylene centrifuge tube. An appropriate volume of
silver nanosphere suspension was pipetted into the sample to achieve a final particle
7 1 5 1
concentration of either 2.5x10 mL" or 2.5x10 mL" . Each spiked sample was inverted several
times, and an appropriate volume was immediately taken and spiked into reagent water to a final
concentration of 2.5xl04 mL"1 for analysis by SP-ICPMS (i.e., the 1 minute sample point).
Identical volumes of each spiked sample (i.e., to achieve a 2.5xl04 mL"1 analysis concentration)
were taken at 12.5, 25, 50, 1400, and for some samples 2900 minutes, diluted to 2.5xl04 mL"1,
and analyzed by SP-ICPMS. Initial test sample preparations were staggered so all water samples
at a given testing condition (nanomaterials and concentration) were analyzed together. Because
the primary focus of the studies was to examine transformations at the more environmentally
relevant low particle concentration, and transformations at high particle concentration were
expected to be predominantly ionic strength related, freshwater samples were not studied at high
particle concentration.
SP-ICPMS analysis
SP-ICPMS analyses were performed on a NexION 300Q (PerkinElmer, Waltham, MA).
The ICPMS was tuned with multi-element tuning solution for maximum overall sensitivity and
oxide and doubly charged levels conforming to manufacturer's specifications. The dwell time
for all SP-ICPMS analyses was 1 ms. Each analysis consisted of a time-resolved analysis of
25,000 dwell periods, the maximum allowed by the instrument. The settling time between dwell
periods was changed in the instrument registry from 100 j_is to 50 j_is. Shorter settling times
improve the accuracy of particle element mass distributions by minimizing the loss of part of
some ion plumes during the settling time (Heithmar 2009).
Calculations
The particle element mass distribution was characterized by the number-based mean
particle element mass (equation 1), the mass-based mean particle element mass (equation 2), and
the polydispersity index (PDI) of particle element mass (equation 3).
Mmean,n = S M; / n (equation 1)
13
-------
Mmea„,m = 2 Mi2 / 2 Mi
(equation 2)
PDIm Mmeanm / Mmeann (equation 3)
Here, M, is the mass of the ith particle, and the summations are over all n particles
detected.
The polydispersity index is a measure of the width of the particle element mass
distribution. PDIm = 1 for a perfectly monodisperse dispersion with respect to particle element
mass. The polydispersity increases as the width of the particle element mass distribution
broadens.
In addition to the determination of nanoparticle silver mass, an operational definition of
"dissolved" silver was applied to the sum of the ion counts in dwell periods containing four or
fewer ions, and the total silver mass in the sample was determined from the sum of the ion
counts in all dwell periods of the analysis. The "dissolved" silver value was interpreted as the
silver mass consistent with free silver ion. It should be noted that "dissolved" silver is an
operational definition. It can be produced by free ionic silver. However, silver-containing
particles that generate only a few ions and therefore cannot be distinguished from the dissolved
background are also included in the "dissolved" silver mass.
14
-------
Results and Discussion
Sample water chemistry
The salinity value of the open-bay Station 31 was much higher than that of the isolated
bay-arm Station 7, as expected (Table 4). Also consistent with significant sea-water content, the
major anion and cation concentrations in Station 31 water were much higher than any of the
other three water samples. Total organic carbon and suspended solids were comparable for the
two estuary samples, and much lower for the fresh water samples. Most of the organic material
that colored the fresh water samples was apparently particulate matter that was retained by the
20-[j,m filter. In fact, there was little color in any of the four filtered samples.
Table 4, Measured water chemistry parameters for water samples used in
transformation studies, ^
Chemistry parameter
S.F. Bay Station 7
S.F. Bay Station 31
Lake Hancock
Alafia River
Salinity (PSU)
0.07
15.8
NA
NA
Total Organic Carbon
(mg/L)
26
24
2.6
2.4
Calcium (mg/L)
11
190
17
14
Magnesium (mg/L)
6.0
490
4.5
8.2
Chloride (mg/L)
8.6
9100
27
14
Sulfate (mg/L)
9.5
1200
6.8
67
Suspended Solids
(mg/L)
64
48
Not detected
Not detected
15
-------
Transformation of 50-nm citrate-capped silver ENM at high particle
concentration
Citrate-capped 50-nm silver ENM suspensions were spiked in the two estuary waters and
7 1
deionized water at a particle concentration of 2.5x10 mL" . Transformations were monitored by
SP-ICPMS up to 2900 minutes exposure. The mass-based mean particle Ag mass did not change
over that time for either deionized water or the low salinity Station 7 water (Figure 4). In the
high salinity Station 31 water, the mean particle Ag mass increased by more than four-fold
compared with the other waters after 2900 minutes of exposure. This is consistent with both
theory and previous laboratory studies in saline water. Both simple Deijaguin, Landau, Verwey
and Overbeek (DLVO) theory and the more detailed Sogami-ISe theory predict that increased
ionic strength will increase nanoparticle aggregation, by shrinking the electrical double layer;
therefore, decreasing interparticle repulsion (Saleh, Kim et al. 2008; French, Jacobson et al.
2009). Initially, average particle Ag mass increases very rapidly in the high-salinity water (i.e.,
>1.5-fold increase in less than one minute).
Figure 4. Mass-based mean particle mass of citrate-capped 50-nm Ag over time in deionized
water and low- and high-salinity estuary waters (2.5xl07 mL"1).
4
3.5
155 3
ro
E
(0
(0
E
2.5
2
1.5
1
~ A
~ D.I. water
¦ Station 7
Station 31
0.5
—i—
10
100
time (minutes)
1000
10000
The PDI of the 50-nm Ag particle Ag mass is not affected by either Station 7 or
deionized water (Figure 5). Again, there is substantial increase in the PDI of the particle Ag
mass in the high salinity Station 31 water compared with the other waters, much of this increase
occurring in less than 1 minute. This indicates that, in addition to shifting to higher average
particle Ag mass, the particle Ag-mass distribution is broadened in Station 31 water.
16
-------
Figure 5. Polydispersity index of citrate-capped 50-nm Ag over time in deionized water and
low- and high-salinity estuary waters (2.5xl07 mL"1).
2.5
~
2
~
I I "
~ ¦
~ D.I. water
¦ Station 7
>
o
Q.
Station 31
0.5
0
1
10
100
1000
10000
time (minutes)
There is significant increase over time in the measured "dissolved" Ag in all three waters
(Figure 6). This relative increase is 24% for deionized water, 45% for Station 7 water, and 60%
for Station 31 water. As discussed previously, SP-ICPMS cannot distinguish between dissolved
silver and very small silver particles. Partial dissolution of silver nanoparticles might be
explained by the dilution of the excess capping agent when the standards are spiked into water.
The total mass of silver measured decreased over time for all three waters (Figure 7).
The reason for this loss of silver is not fully explained at present. Some loss in total silver could
be expected for the Station 31 water from precipitation of larger aggregates, but there was no
observable aggregation in the other waters. One possible cause is a slow precipitation, or
adsorption on the container walls, of primary particles.
17
-------
Figure 6. Change in measured "dissolved" silver over time in suspensions of citrate-capped
50-nm Ag in deionized water and low- and high-salinity estuary waters (2.5xl07 mL"1).
120-
100-
80 ¦
60 ¦
40 ¦
20 ¦
~ D.I. water -
¦ Station 7 -
Station 31 ¦
10
100
time (minutes)
1000
10000 -
Figure 7. Change in total measured silver over time in suspensions of citrate-capped 50-nm
7 1
Ag in deionized water and low- and high-salinity estuary waters (2.5x10 mL" ).
900
TS 800
.bjQ
-------
Comparison of transformations of 50-nm and 100-nm citrate-capped
silver ENM
The transformation rates of 50-nm and 100-nm citrate-capped silver ENM were
1 i
compared at particle concentrations of 2.5x10 mL" over 1400 minutes. Because transformation
rates for 50-nm silver were fastest in the high-salinity Station 31 water, that sample was used for
the comparison. The PDI of the particle Ag mass was used as a measure of particle aggregation.
The PDI of the 50-nm silver suspension increased 16% over 1400 minutes (Figure 8). The PDI
of the 100-nm suspension increased by 36% over the same time. This supports a conclusion that
the rate of aggregation of silver nanoparticles increases with particle size. He et al. studied the
effect of particle size on hematite nanoparticle aggregation and concluded that "at the same ionic
strength, aggregation rates are higher for smaller particles (He, Wan et al. 2008). However, that
study was done at constant hematite mass concentration, so it is not necessarily contradictory to
the conclusion of this study. The silver mass concentration was 8 times higher in the 100-nm
1 i
silver suspension than in the 50-nm suspension at the same 2.5x10 mL" particle concentration.
The increase in the "dissolved" silver concentration in the 100-nm silver suspension was
3.4 times the increase in the 50-nm suspension (Figure 9). Because the dissolution rate should be
proportional to surface area, this is in fair agreement with the 4-fold greater surface total silver
area in the former suspension.
19
-------
Figure 8. Increase in PDI over time for suspensions of 50-nm and 100-nm citrate-capped Ag
in high-salinity estuary water (2.5xl07 mL"1).
2.5
X
O
CL
1.5-
0.5
~ 50 nm Ag
¦ 100 nm Ag
—i—
10
100
time (minutes)
1000
10000 -
Figure 9. Increase in "dissolved" silver over time for suspensions of 50-nm and 100-nm
citrate-capped Ag in high-salinity estuary water (2.5xl07 mL"1).
txo
<
T3
O
U)
U)
b
180
160
140
120
100
80
60 -
40
20
0
~ 50 nm Ag
¦ 100 nm Ag
—i—
10
100
time (minutes)
1000
10000 -
20
-------
Comparison of transformations of PVP-capped and citrate-capped 50-
nm silver ENM
The transformation rates of PVP-capped and citrate-capped 50-nm silver ENM were
7 1
compared at particle concentrations of 2.5x10 mL" over 1400 minutes. As with the previous
comparison of different size particles, the high-salinity Station 31 water was used for the
comparison, and the PDI of the particle Ag mass was used as a measure of particle aggregation.
The PDI increase over 1400 minutes was 45% for the PVP-capped silver compared with 16% for
the citrate capped silver (Figure 10).
Figure 10. Increase in PDI over time for suspensions of PVP-capped and citrate-capped 50-
nm Ag in high-salinity estuary water (2.5xl07 mL"1).
X
O
CL
2.5 n
2
1.5
1-
0.5
¦ *
~ Citrate
¦ PVP-
10
100
time (seconds)
1000
10000
The increase in the "dissolved" silver concentration in the PVP-capped suspension over
1400 minutes is 2.25 times greater than that in the citrate-capped suspension (Figure 11),
although the PVP surface area is only 18% larger.
21
-------
Figure 11. Increase in "dissolved" silver over time for suspensions of PVP-capped and
citrate-capped 50-nm Ag in high-salinity estuary water (2.5xl07 mL"1).
140
120
100
80
T, 60
40
20
~ Citrate
¦ PVP
10
100
time( minutes)
1000
10000
Transformation of 50-nm citrate-capped silver ENM at low particle
concentration
The transformations of 50-nm citrate capped silver ENM suspensions in all four surface
waters were monitored over 1400 minutes, this time at a concentration of 2.5x10s mL"1, 100-fold
more dilute than the previous experiments. The percent changes in mass-based mean particle Ag
mass, PDI of the particle Ag, measured "dissolved" Ag mass, and measured total Ag mass were
calculated for 1400 minutes exposure (Figure 12). At this concentration, there is very little
evidence of aggregation, in terms of either average particle Ag mass or PDI of the particle Ag
mass. The single exception is a moderate increase in average particle Ag mass in the Alafia
River water. These results are consistent with aggregation following pseudo second-order
kinetics. There is evidence of some total silver mass loss. By far the most dominant process at
this low concentration is an increase in "dissolved" silver.
22
-------
Figure 12. Changes in suspension metrics of 50-nm citrate-capped silver after 1400 minutes
(2.5x10s mL"1)
200
Station 31
¦ Station 7
Lake Hancock
¦ Alafia River
-50
mass-average polydispersity "dissolved" Ag total mass Ag
particle mass index
The overall % changes over 1400 minutes for 2.5x10s mL"1 (Figure 13 - blue bars) and
f 1
2.5x10 mL" (Figure 13 - red bars) were compared. Not only was apparent dissolution the
dominant process at the lower concentration, the degree of apparent dissolution was more than 5-
fold greater over 1400 minutes than at high concentration as a percentage of the original value.
The silver mass concentration of the suspension at a particle concentration of 2.5x10s mL"1 was
0.16 [j,g/L (as calculated from 20 mg/L concentration of the purchased standard). This
concentration is much more realistic than the 16 [j,g/L concentration at a particle concentration of
7 1*
2.5x10 mL" with respect to the concentrations of engineered nanoparticles likely to be currently
encountered in the environment.
23
-------
Figure 13. Effect of concentration on changes in suspension metrics of 50-nm citrate-capped
silver after 1400 minutes (2.5xl05 mL1)
200
2.5xl05mL":
150
100
2.5xl07mL":
a>
m
\o
50
0 P
-50
-100
mass-average polydispersity "dissolved" Ag total mass Ag
mass index
24
-------
Conclusions and Future Work
The study reported here demonstrates the effectiveness of SP-ICPMS for rapidly
screening surface water for metal-containing nanoparticles. This capability would allow large
numbers of water samples to be screened. The samples showing possible presence of target
nanoparticles could then be analyzed by more selective techniques that are much more time-
consuming and costly. The study reported here shows another advantage of the rapid screening
capability of SP-ICPMS. Rates of transformations, such as aggregation and dissolution, can be
studied with high temporal resolution. In this case, the data related to aggregation is reliable
even in complex water matrices. Dissolution information is screening-level, because SP-ICPMS
cannot differentiate dissolved metal from very small nanoparticles. The sensitivity of current
ICPMS instruments precludes SP-ICPMS from distinguishing particles smaller than about 15-20
nm from dissolved silver. This would prohibit its use in studying primary particles of consumer
products containing small nanoparticles. On the other hand, SP-ICPMS could also be used to
study disaggregation if the primary particles are larger than this lower limit.
This study confirms the results of other studies of aggregation of nanoparticles at high
1 i
ionic strength and high (>10 mL" ) nanoparticle concentrations in both synthetic laboratory
media (Saleh, Kim et al. 2008; French, Jacobson et al. 2009; Trinh, Kj0niksen et al. 2009) and
natural water (Chinnapongse, MacCuspie et al. 2011). Those studies all found that aggregation
of a variety of nanoparticles increased with increasing ionic strength. However, this study also
shows that at low silver nanoparticle concentration (2.5x10s mL"1), aggregation is minimal over
24 hours, even in highly saline estuary water. Instead, the dominant process is one that is
consistent with either dissolution of the nanoparticles or formation of smaller nanoparticles that
cannot be experimentally distinguished from dissolved. This has not been observed in previous
studies because the sensitivity of the analytical techniques used could not study transformation at
low nanoparticle concentration. Studies in this concentration regime are critical, because it is the
range of expected environmental concentrations of engineered nanoparticles (Kiser, Westerhoff
et al. 2009).
These transformation study results are preliminary and more detailed research is planned.
These additional studies will investigate the role of natural organic matter on transformations at
low nanoparticle concentration, to compare with previous studies at high concentration that
found suppression of aggregation by NOM (Liu, Wazne et al. 2010; Thio, Zhou et al. 2011). The
effects of pH, temperature, and oxygen saturation should also be investigated. Finally,
ultrafiltration and other approaches will be explored to elucidate the nature of the "dissolved"
silver. Finally, hyphenated methods will be developed.
25
-------
References
Alvarez, P. J. J., V. Colvin, et al. (2009). "Research Priorities to Advance Eco-Responsible
Nanotechnology." ACS Nano 3(7): 1616-1619.
Chinnapongse, S. L., R. I. MacCuspie, et al. (2011). "Persistence of singly dispersed silver
nanoparticles in natural freshwaters, synthetic seawater, and simulated estuarine waters."
Science of The Total Environment 409(12): 2443-2450.
Degueldre, C. and P. Y. Favarger (2004). "Thorium colloid analysis by single particle
inductively coupled plasma-mass spectrometry." Talanta 62(5): 1051-1054.
Degueldre, C., P. Y. Favarger, et al. (2004). "Zirconia colloid analysis by single particle
inductively coupled plasma-mass spectrometry." Analvtica Chimica Acta 518(1-2): 137-
142.
Degueldre, C., P. Y. Favarger, et al. (2006). "Gold colloid analysis by inductively coupled
plasma-mass spectrometry in a single particle mode." Analvtica Chimica Acta 555(2):
263-268.
Ebenstein, Y., E. Nahum, et al. (2002). "Tapping Mode Atomic Force Microscopy for
Nanoparticle Sizing: Tipa"'Sample Interaction Effects." Nano Letters 2(9): 945-950.
Filella, M., J. Zhang, et al. (1997). "Analytical applications of photon correlation spectroscopy
for size distribution measurements of natural colloidal suspensions: capabilities and
limitations." Colloids and Surfaces A: Physicochemical and Engineering Aspects 120(1-
3): 27-46.
French, R. A., A. R. Jacobson, et al. (2009). "Influence of Ionic Strength, pH, and Cation
Valence on Aggregation Kinetics of Titanium Dioxide Nanoparticles." Environmental
Science & Technology 43(5): 1354-1359.
Gray, P. J., J. W. Olesik, et al. (2009). Particle Size Effects on Vaporization in Laser Ablation
Inductively Coupled Plasma Mass Spectrometry. Pittsburgh Conference on Analytical
Chemistry and Applied Spectroscopy, Chicago, IL.
Handy, R., F. von der Kammer, et al. (2008). "The ecotoxicology and chemistry of manufactured
nanoparticles." Ecotoxicology 17(4): 287-314.
Hassellov, M. (2009). Detection and Characterization of engineered nanoparticle in the
enviroment using Field-Flow Fractionation coupled to ICP-MS in single nanoparticle
detection mode. Goteborg Setae Europe.
Hassellov, M., J. Readman, et al. (2008). "Nanoparticle analysis and characterization
methodologies in environmental risk assessment of engineered nanoparticles."
Ecotoxicology 17(5): 344-361.
He, Y., J. Wan, et al. (2008). "Kinetic stability of hematite nanoparticles: the effect of particle
sizes." Journal of Nanoparticle Research 10(2): 321-332.
Heithmar, E. (2009). Characterizing Metal-based Nanoparticles in Surface Water by Single-
Particle ICPMS. 237th ACS National Meeting, Salt Lake City, Utah.
Kammer, F. v. d., M. Baborowski, et al. (2005). "Field-flow fractionation coupled to multi-angle
laser light scattering detectors: Applicability and analytical benefits for the analysis of
environmental colloids." Analvtica Chimica Acta 552(1-2): 166-174.
Kim, T., C.-H. Lee, et al. (2008). "Kinetics of gold nanoparticle aggregation: Experiments and
modeling." Journal of Colloid and Interface Science 318(2): 238-243.
26
-------
Kiser, M. A., P. Westerhoff, et al. (2009). "Titanium Nanomaterial Removal and Release from
Wastewater Treatment Plants." Environmental Science & Technology 43(17): 6757-
6763.
Klaine, S. J., P. J. J. Alvarez, et al. (2008). "NANOMATERIALS IN THE ENVIRONMENT:
BEHAVIOR, FATE, BIOAVAILABILITY, AND EFFECTS." Environmental
Toxicology & Chemistry 27: 1825-1851.
Koirtyohann, S. R., J. S. Jones, et al. (1980). "Nomenclature system for the low-power argon
inductively coupled plasma." Analytical Chemistry 52(12): 1965-1966.
Lam, J. W. H. and G. Horlick (1990). "A Comparison of Argon and Mixed Gas Plasmas for
Inductively Coupled Plasma-mass Spectrometry." Spectrochimica Acta Part B: Atomic
Spectroscopy 45.
Lesher, E. K., J. F. Ranville, et al. (2009). "Analysis of pH Dependent Uranium(VI) Sorption to
Nanoparticulate Hematite by Flow Field-Flow Fractionation - Inductively Coupled
Plasma Mass Spectrometry." Environmental Science & Technology 43(14): 5403-5409.
Lin, W.-C. and M.-C. Yang (2005). "Novel Silver/Poly(vinyl alcohol) Nanocomposites for
Surface-Enhanced Raman Scattering-Active Substrates." Macromolecular Rapid
Communications 26(24): 1942-1947.
Liu, X., M. Wazne, et al. (2010). "Effects of natural organic matter on aggregation kinetics of
boron nanoparticles in monovalent and divalent electrolytes." Journal of Colloid and
Interface Science 348(1): 101-107.
Monserud, J. H., E. K. Lesher, et al. (2009). Real Time Single Particle-inductively Coupled
Plamsa-mass Spectrometry for Detection and Characterization of Nanoparticles. 237th
ACS National Meeting, Salt Lake City, Utah.
Mueller, N. C. and B. Nowack (2008). "Exposure Modeling of Engineered Nanoparticles in the
Environment." Environmental Science & Technology 42: 4447-4453.
O'Connor, G. and E. H. Evans (1999). Fundamental aspects of ICP-MS. Inductively Coupled
Plasma Spectrometry and its Applications. S. J. Hill. Sheffield, Sheffield Academic
Press: 119-144.
Pyrz, W. D. and D. J. Buttrey (2008). "Particle Size Determination Using TEM: A Discussion of
Image Acquisition and Analysis for the Novice Microscopist." Langmuir 24(20): 11350-
11360.
Saleh, N., H.-J. Kim, et al. (2008). "Ionic Strength and Composition Affect the Mobility of
Surface-Modified FeO Nanoparticles in Water-Saturated Sand Columns." Environmental
Science & Technology 42(9): 3349-3355.
Schulte, P., C. Geraci, et al. (2008). "Occupational Risk Management of Engineered
Nanoparticles." Journal of Occupational & Environmental Hygiene 5: 239-249.
Schurtenberger, P., M. E. Newmen, et al. (1993). "Environmental Particles." J. Buffle. H.P. van
Leeuwen (Eds) 2: 37.
Takagi, A., A. Hirose, et al. (2008). "Induction of mesathelioma in p53+/- mouse by
intraperitioneal application of multi-wall carbon nanotube." Journal of Toxicol Science
33: 105-116.
Tang, Z., L. Wu, et al. (2009). "Size fractionation and characterization of nanocolloidal particles
in soils." Environmental Geochemistry and Health 31(1): 1-10.
Taylor, H. E. (2001). Inductively Coupled Plasma-Mass Spectrometry: Practices and
Techniques. San Diego, Academic Press.
27
-------
Thio, B. J. R., D. Zhou, et al. (2011). "Influence of natural organic matter on the aggregation and
deposition of titanium dioxide nanoparticles." Journal of Hazardous Materials 189(1-2):
556-563.
Thomas, R. (2004). Practical Guide to ICP-MS. New York, Marcel Dekker, Inc.
Tiede, K., A. B. A. Boxall, et al. (2010). "Application of hydrodynamic chromatography-ICP-
MS to investigate the fate of silver nanoparticles in activated sludge." Journal of
Analytical Atomic Spectrometry 25(7): 1149-1154.
Tiede, K., M. Hassellov, et al. (2009). "Considerations for environmental fate and ecotoxicity
testing to support environmental risk assessments for engineered nanoparticles." Journal
of Chromatography A 1216(3): 503-509.
Trinh, L., A.-L. Kj0niksen, et al. (2009). "Slow salt-induced aggregation of citrate-covered silver
particles in aqueous solutions of cellulose derivatives." Colloid & Polymer Science
287(12): 1391-1404.
28
-------
-------
4>EFW
United States
E n v i ro n m o n t a I Prote ct i o n
Agency
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
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
Office of Research
and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
EPA/600/R-11/096
September 2011
www.epa.gov
Recycled/Recyclable
Printed with vegetable-based ink on
paper that contains a minimum of
50% post-consumer fiber content
processed chlorine free
------- |