SEARCH

SEARCH BY CITATION

Keywords:

  • Manufactured nanomaterials;
  • Cerium oxide;
  • Microscopy;
  • Crystal structure;
  • Chemical composition;
  • Surface properties

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. References

Part 1 (see companion paper) of the present study discussed the application of a multimethod approach in characterizing the size of cerium oxide nanoparticles (NPs). However, other properties less routinely investigated, such as shape and morphology, structure, chemical composition, and surface properties, are likely to play an important role in determining the behavior, reactivity, and potential toxicity of these NPs. The present study describes the measurement of the aforementioned physicochemical properties of NPs (applied also to nanomaterials [NMs]) compared with micrometer particles (MPs). The authors use a wide range of techniques, including high resolution-transmission electron microscopy, scanning electron microscopy, atomic force microscopy, X-ray diffraction, X-ray energy dispersive spectroscopy, electron energy loss spectroscopy, X-ray photoelectron spectroscopy, and electrophoresis, and compare these techniques, their advantages, and their limitations, along with recommendations about how best to approach NM characterization, using an application to commercial cerium oxide NPs and MPs. Results show that both cerium oxide NPs and MPs are formed of single polyhedron or truncated polyhedron crystals. Cerium oxide NPs contain a mixture of Ce3+ and Ce4+ cations, whereas the MPs contain mainly Ce4+, which is potentially important in understanding the toxicity of cerium oxide NPs. The isoelectric point of cerium oxide NPs was approximately pH 8, which explains their propensity to aggregate in aqueous media (see companion paper). Environ. Toxicol. Chem. 2012; 31: 994–1003. © 2012 SETAC


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. References

Quantitative studies of the effect of the physicochemical properties of nanomaterials (NMs) on their toxicity under realistic exposure scenarios are still scarce 1. As discussed in Part 1 of the present study (see companion paper 2), it is clear that, in some cases, particle size and specific surface area can be key determinants of NM toxicity, but other properties such as shape, morphology, crystal structure, composition, purity, surface chemistry, and particle reactivity may also play a significant role.

The shape of NMs may play an important role in determining their toxicity. Certain types of carbon nanotubes have been shown to have asbestos-like behavior, dependent on length, aspect ratio, and biopersistence 3. Nanoparticles (NPs; e.g., TiO2, Al2O3, ZrO2, Si3N4, and SiC) containing more edges have been shown to have higher toxicity to exposed murine fibroblasts and macrophages 4. Particle shape may also influence the biocidal property of silver NPs, with truncated triangular silver particles showing the strongest effect, compared with the less toxic spherical and rod-shaped NPs 5.

Crystal structure may have a determining role in NP toxicity and may be the underlying reason for the toxic effects of particle shape. For example, the crystal structure of silicon dioxide and titanium dioxide NPs controls their ability to induce pulmonary inflammation and fibrosis in mice 6. In addition, the crystal-phase composition of NPs can control their photocatalytic and toxic effects. For instance, amorphous titania has been shown to be capable of producing more reactive oxygen species than the crystalline (anatase or rutile) phases, and anatase NPs have been found to be more cytotoxic than rutile 7.

Particle composition and surface chemistry may be important in determining biological effects. Comparing the toxicity of silver and cerium oxide (CeO2) NPs and MPs across a variety of biological systems and models suggests that silver particles are more toxic than CeO2 particles and that NPs are more toxic than MPs 8. In addition, interaction of NPs with contaminants such as trace metals (Cd) may enhance toxicity 8. The toxicity and protective role of CeO2 NPs have both been related to the oxidation state of the surface atoms (Ce3+/Ce4+). Interaction of NPs with organic molecules such as humic substances and proteins or other organic molecules in biological media may alter their surface chemistry 9, 10 and biological impact 11. Surface charge is another key property, which often, but not always, controls the colloidal stability of NP dispersion and hence will affect measurements of exposure concentration, nature of toxicant (dispersed or agglomerated, for charge stabilized dispersions), and transport in environmental and biological systems. For instance, the surface charge of NPs may change their ability to be transported through the blood–brain barrier and alter their permeability to mammalian cells 12.

Because of this relationship between NM properties and toxicity, several reviews have recommended the quantification of the aforementioned properties as a minimum requirement for the physicochemical characterization of NMs prior to any toxicological studies 1, 13. Nevertheless, most toxicological and ecotoxicological studies still report few physicochemical characteristics, often incompletely measured, and rarely relate these properties to toxic effects. Clearly, both a detailed characterization of NM physicochemical properties, including nonsize properties, and biological effects investigations are essential to improving our understanding of NM toxicity 1, 13. Only in this way will it be possible for direct data comparisons to be made allowing the construction of a reliable hazard databases for NMs and providing the underpinning data for the development of quantitative structure–activity relationship models for predicting NM toxicity.

The present article reports the characterization of commercially available CeO2 NPs and micrometer particles (MPs), including their nonsize-based characteristics; in particular, their morphology, shape, structure, chemical composition, and surface charge and chemistry were quantified. Particle size characterization, along with related parameters such as specific surface area, are discussed in detail in Part 1 2. Here we discuss characterization of nonsize properties as assessed with various techniques (single-particle and ensemble techniques), along with their basic principles, advantages and limitations, sample requirements, and data treatment, as well as providing recommendations on how to overcome some of the characterization problems. Cerium oxide NPs are currently used as fuel additives to enhance fuel efficiency and reduce harmful emissions. In addition, ceria NPs are used in a wide range of applications, including catalysts, surface coating, and polishing materials. This wide use has resulted in their release to the environment, in which their fate, exposure levels, and potential impact are largely unknown.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. References

The NPs used, atomic force microscopy (AFM), transmission electron microscopy (TEM), and X-ray diffraction (XRD) techniques are described in detail in Part 1 of the present study 2: a short summary is given below. In addition, we describe methods that were not included in Part 1, including scanning electron microscopy (SEM), X-ray energy dispersive spectroscopy (X-EDS), electron energy loss spectroscopy (EELS), X-ray photoelectron spectroscopy (XPS), and laser Doppler electrophoresis. The present study focuses only on the characterization of the chemical composition, oxidation state, and surface charge of the CeO2 particles. However, other chemical and surface properties that can also be obtained using other analytical tools have not been applied in the present study such as Fourier transform infrared spectroscopy, nuclear magnetic resonance spectroscopy, time-of-flight secondary-ion mass spectroscopy, low-energy ion scattering, Raman spectroscopy, and so on, all of which provide complementary information on particle chemistry and surface properties. A review of all of the different chemical and surface analytical techniques is beyond the scope of the present study, and more information can be found elsewhere 13.

Cerium oxide NPs (nominal size <25 nm, determined via Brunauer, Emmett, and Teller (BET) measurement, manufacturer's data), and MPs (nominal size <5 µm; manufacturer's data) were purchased from Sigma Aldrich, and several ecotoxicology exposures have been reported 8, 14. Stock suspensions of CeO2 NPs and MPs were prepared by adding approximately 10 mg to ultrahigh-purity water (R > 18 MΩ and total organic carbon < 2 µg C L−1) to obtain 100 mg L−1 of particles in concentration, which was sonnicated for 30 min to enhance particle dispersion. The stock solution was further diluted to 0.1 and 1.0 mg L−1 in 10 mM NaCl for sample analysis. All the suspensions were left to equilibrate for 24 h.

The TEM used was an FEI Tecnai F20 field emission gun coupled with an X-EDS from Oxford Instruments and EELS from Gatan, Inc. Transmission electron microscopy samples were prepared by ultracentrifugation at 150,000 g using a Beckman ultracentrifuge (L7-65 ultracentrifuge) with a swing-out rotor SW40Ti. In all cases, at least 100 NPs were analyzed by TEM to construct a representative particle size distribution.

An XE-100 AFM (Park Systems) was used in the present study. Samples were prepared by applying an adsorption in which freshly cleaved muscovite micas sheets were immersed vertically in suspensions of NPs and MPs for 30 min. Subsequently, the mica sheets were withdrawn from the solution and gently rinsed by immersion in deionized water to remove nonadsorbed particles and excess NaCl. The specimen was allowed to dry in a covered Petri dish overnight. The measurements were carried out under ambient conditions in true noncontact mode using a Silicon cantilever with a spring constant of 42 (10–130) N m−1 (Nanosensors; 910 (U)-NCHR).

Scanning electron microscopy

The SEM used here was a Philips XL-30 field emission gun-SEM. Samples were imaged with a standard secondary electron detector using a 10-kV electron beam. Samples were prepared by depositing the powder of NPs on double-sided sticky tape; the other side was placed on the sample holder, to which it strongly adhered. The samples were then coated with a thin layer of gold by sputtering under vacuum.

X-ray energy dispersive spectroscopy

The X-EDS analysis was performed using a Link ISIS EDS detector from Oxford Instruments, which uses a thin window Li(Si) detector capable of detecting all elements with an atomic number ≥5, that is, above Be, in the periodic table. To obtain an optimal takeoff angle of the photon beam, the TEM grid was tilted toward the detector by approximately 20°.

Electron energy loss spectroscopy

The EELSs were recorded on a TEM (described above). The EELS spectra were collected with a Gatan PEELS 666 parallel electron energy spectrometer and EL/P data acquisition software. The energy spread of the incident electron beam (as measured by the full width at half-maximum of the zero-loss peak) is approximately 1.0 eV. All the spectra were collected in diffraction mode and noncollimated electrons were effectively blocked by using a 10-µm selected area aperture. The O K-edge and the Ce M-edge EELS spectra were collected with a convergence angle of 2α = 8.2 mrad, a collection angle of 2β = 8.9, a 1.0-mm EELS aperture, and 0.5 eV/channel energy dispersion. Cerium oxide particles are very beam sensitive, with a reduction of Ce4+ to Ce3+ with time through irradiation-induced reduction 9. A time series of EELS spectra was collected every 3 s and for 3 min; no irradiation damage was observed within the first 9 s, so the collection time for these spectra was confined to 3 s to prevent irradiation damage from affecting the data. The Ce M4–5 core-loss ionization edges and the corresponding low-loss EELS spectrum, including the zero-loss peak, were acquired consecutively from the same specimen region. Here we limit our investigation to qualitative analysis. Quantitative analysis of the oxidation state of cerium ions in CeO2 NPs and MPs in different media is presented elsewhere 9, 10.

X-ray photoelectron spectroscopy

X-ray photoelectron spectroscopy measurements of the dry NP samples (powder) were performed using an ESCLab instrument (Thermo VG scientific) equipped with Al Kα X-ray source (1,486.68 eV), which was operated at 15 kV. The NP samples were placed on a standard sample stud using double-sided adhesive tape, and the takeoff angle was fixed at 90°. Low-resolution survey spectra were obtained over a binding energy range of 0.0 to 1,050 eV using 1-eV increments. High-resolution spectra were obtained over a binding energy range of 860 to 920 eV using 0.1-eV increments.

Laser Doppler electrophoresis

Electrophoretic mobility and zeta potential can be measured in many ways, and laser Doppler velocimetry is one that has wide availability and is used in our laboratories. Laser Doppler electrophoresis measures the movement of charged particles in an electric field (known as electrophoretic mobility), using a well-known Doppler effect (the change in the frequency of light scattered from moving particles relative to the incident frequency of the light). The electrophoretic mobility of particles was evaluated using a Zetasizer (model ZEN3600; Malvern Instruments) operating with an He-Ne laser at a wavelength of 633 nm using back scattered light. Calculation of zeta potential, which is related but not identical to surface charge, from the measured electrophoretic mobility is discussed below. Several other techniques can be used to measure particle electrophoretic mobility, with different principles of operation, including electroacoustic effects and colloid vibration current 15, 16. To determine the isoelectric point (IEP; i.e., point of zero electrophoretic mobility/zeta potential 17), the sample pH was altered by adding small aliquots of 0.1 or 0.01 M HCl or NaOH.

RESULTS AND DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. References

Shape and morphology

Figures 1a and 1b show SEM micrographs of CeO2 MPs and NPs, respectively. Shape and morphology of MPs are clearly observed, whereas those of NPs are not fully resolved. Scanning electron microscopy has a more limited resolution (generally, a few nanometers) compared with TEM or AFM, and this resolution is approximately at the size of the NPs (see companion paper 2). Figure 1a shows a pseudo-3D image of MPs having a range of shapes, including triangular pyramid, truncated octahedral, truncated hexagonal pyramid with equal edge lengths, and truncated hexagonal pyramid with unequal edge lengths. However, Figure 1b simply shows an indistinct mass of particles resulting from resolution constraints.

thumbnail image

Figure 1. (a) Typical scanning electron microscopy (SEM) micrographs of cerium oxide microparticles showing a range of shapes, including triangular pyramid (1), truncated octahedral (2), truncated hexagonal pyramid with equal edge lengths (3), and truncated hexagonal pyramid with unequal edge lengths (4). (b) Cerium oxide nanoparticles (NPs).

Download figure to PowerPoint

Figure 2 shows images taken by AFM (amplitude images), again showing different shapes of MPs similar to those observed with SEM. Atomic force microscopy offers 3D visualization with high resolution and accuracy (0.1 nm; Table 1) in the vertical, Z axis and relatively poor accuracy at the nanoscale in the horizontal, X-Y axis, limited by tip–particle interactions at the nanoscale. In the present study, a tip with a radius of approximately 10 nm has been used, so the shape and morphology of NPs could not fully be resolved.

thumbnail image

Figure 2. Typical atomic force microscopy (AFM) amplitude images of cerium microparticles (MPs; a, b) and cerium oxide nanoparticles (NPs; c, d).

Download figure to PowerPoint

Table 1. Comparison between the different techniques presented in the present study
MethodAFMHR-TEMSEMX-EDS in TEMX-EDS in SEMEELSXPSXRDLDE
  • a

    Depending on the electron source and operational conditions.

    AFM = atomic force microscopy; HR-TEM = high-resolution transmission electron microscopy; SEM = scanning electron microscopy; X-EDS = X-ray energy dispersive spectroscopy; EELS = electron energy loss spectroscopy; XPS = X-ray photoelectron spectroscopy; LDE = laser Doppler electrophoresis; PSD = particle size distribution; UHV = ultrahigh vacuum.

Measured parameterPSD, shape, tip–specimen interaction forcesPSD, morphology, crystallography structure, defectsPSD, morphology,topographyElemental compositionElemental composition, chemical bonds, coordination number, nearest-neighbor distribution electronic structure, oxidation stateElemental composition, bond type and length, electronic structure, oxidation stateStructural analysis of crystals, amounts of different crystalline phases, crystal sizeElectrophoretic mobility
PrincipleScanning a probe on a surfaceInteraction of electrons with matterInteraction of electrons with matterEmission of photon resulting from relaxation of excited atomsInelastic interaction of electrons with matterEmission of electrons due to relaxation of excited atomsElastic interaction of X-rays with matterDoppler laser effect
Spatial resolution0.1 nm height0.1–0.2 nm5–20 nm2–20 nm≥300 nm∼1 nm in diffraction mode3 µm10 µm
Depth resolutionSample thickness0.5–3 µmSample thickness1–10 nm
Energy resolutionTypically 110–150 eV0.3–2.5 eVa0.25–1 eVa
Detectability0.1–1.0 weight %0.1–1.0 weight %0.1–1.0 weight %0.1–1.0 weight %Ten times better than EDS Single-atom detection has been demonstrated0.1–1.0 atom % 1,000–10,000 ppm
Sample environmentAmbient air LiquidUHVUHVUHVUHVUHVAmbient conditionsLiquid
ApplicabilityRaw material and in situRaw material and in situRaw material and in situRaw material and in situRaw material and in situRaw material and in situRaw/extracted materialIn situ
Optimum sample thickness15–20 nm15–20 nm<20 nm
Elemental rangeZ ≥ 5Z ≥ 3Z ≥ 3
Quantification requirementWith calibrationWith calibrationPeak fitting
CalibrationStandard grid of specific heightCalibration of image scale with defined lattice parameter   Using standards of known oxidation stateUsing standards of known oxidation state  
Measurement timeSlow (hours)Slow (hours)Slow (hours)Fast (minutes)Slow (hours)Slow (hours)Moderate-slow (>20 min-h)Fast (minutes)
Data analysisSlow (hours)Slow (hours)Slow (hours)Fast (minutes)Slow (hours)Slow (hours)Moderate-slow (> 20 min-h)Fast (minutes)
AdvantagesHigh-resolution analysis can be performed under ambient pressure and in aqueous mediaHigh resolution, visual observation of the particlesDirect observation of the particles, gives semi-3D informationStraight forward for qualitative analysis, spectrum can be fully acquired in one passSuited to light elements that are difficult to analyze by X-EDS, gives detailed chemical informationSurface sensitive, relatively nondestructivenondestructive, fast, and averaging of propertiesEasy, fast
DifficultyTime-consuming, requires large number of particles for representative PSD, TEM requires special sample preparation, requires UHVLow energy resolution, results in peaks overlap between certain elementsTime-consuming, requires to identify many parameters for the TEM and the EELS systemsExpensive, slow, poor spatial resolution, require UHVPossibility of sample damage, low spatial resolutionRepresentation of the calculated ζ potential

Figure 3 shows images from the same sample taken by TEM, revealing a projection of the particles in two dimensions rather than pseudo-3D information as obtained from SEM and AFM. Transmission electron microscopy images (Fig. 3a) of MPs show different shapes, including rhombus, hexagonal with equal edge lengths, and hexagonal with unequal edge lengths, in good agreement with the shapes observed via SEM and AFM. In contrast to SEM, the high resolution of TEM (∼0.1–0.2 nm; Table 1) allows resolution of shape, morphology, and lattice structure (Fig. 4; discussed below) of individual NPs. Figures 3b and 3c show different shapes of CeO2 NPs, including rhombus (b) with equal and unequal edge lengths and hexagonal with equal edge lengths (c). Transmission electron microscopy is a preferred method for assessing the morphometry of these NPs, and AFM and SEM provide important supplementary data, and, as in all cases, a multimethod approach is favored. Clearly, in electron microscopy (EM), as in other analyses, issues of sample preparation, as discussed in the present article and in Part 1 2, must be considered fully. Quantifying particle shape is not a trivial process, except for the most symmetric systems (such as spheres or rods). Even here, a great deal of time-consuming work is required to obtain representative information by microscopy such as measurement of a large number of randomly selected particles and conservation of the particle shape from any perturbations caused by the ultrahigh vacuum, drying effects, or aggregation processes. The combination of techniques such as flow field-flow fractionation and AFM 18 has allowed the quantification of shape for very small NPs (<25 nm) in natural waters without the ultrahigh vacuum. A combination of flow field-flow fractionation with multiangle laser light scattering allows the quantification of the shape of larger NPs (25–450 nm) 19. Other methods, such as small-angle X-ray scattering (not employed here), can also be used to determine NP size and shape as well as aggregate structure (fractal dimension) but have not yet been validated for environmental samples 20, 21.

thumbnail image

Figure 3. Typical transmission electron microscopy (TEM) micrographs of cerium microparticles (MPs; a), cerium oxide nanoparticles (NPs; b, c), aspect ratio distribution of cerium oxide NPs (d), and circularity parameter distribution of cerium oxide NPs (e).

Download figure to PowerPoint

thumbnail image

Figure 4. Typical high-resolution transmission electron microscopy (HR-TEM) micrograph of cerium oxide nanoparticles (NPs). [Color figure can be seen in the online version of this article, available at wileyonlinelibrary.com].

Download figure to PowerPoint

Figures 2 and 3 show the wide differences in shape of the NPs used in the present study. Along with the polydispersity and extensive aggregation and agglomeration, which are widely seen in commercial samples, this heterogeneity in material characteristics limits understanding of fundamental toxicological mechanisms of action or environmental chemistry and transport, which require extensive quantification of NP properties. For instance, nonuniform shape and deviations from circularity affect the meaningfulness of derived particle size distribution. In many samples, as with these NPs, particles may have a range of different shapes (i.e., a shape distribution). Several methods of calculating shape have been used to quantify the shape and geometry of particles, such as sphericity or circularity, aspect ratio, fractal dimension.

Figure 3d shows the aspect ratio (i.e., length/width) of CeO2 NPs calculated from TEM images, with a range of aspect ratio from 1.0 to 2.6. Approximately 44% of particles have an aspect ratio of 1.0 to 1.1, and this corresponds to the rhombus or hexagonal shape with equal edge lengths; higher values may correspond to elongated rhombus or hexagonal shapes. Nanoparticles produced by different methods will have different shapes—even NPs from the same preparation method might have different shapes—the importance of this shape distribution is unknown in environmental processes, and only a small amount of information is available for shape effects on toxicity 5, with one significant exception, the asbestos-like behavior of high-aspect-ratio NMs such as carbon nanotubes 3 and most likely other high-aspect-ratio NMs made of other materials.

Particle shape can also be described by a circularity factor, which can be defined as in Equation 1

  • equation image(1)

where A and P are the area and perimeter of the particle, respectively. The circularity parameter equals 1 for a circle, and smaller values indicate deviation from the circular shape, with a square being equal to 0.80, a triangle of equal sides 0.61, with lower, smaller values (0.17–0.54) for more elongated objects 22. Figure 3e shows the circularity factor of CeO2 NPs calculated from TEM images. Approximately 44% of the particles have a shape factor of 0.886, which corresponds to a shape between a sphere and a square and is likely to be the rhombus shape or hexagonal shape with equal edge lengths, in good agreement with the aspect ratio data.

Crystallography: XRD and high-resolution TEM

Figure 3a (presented in Part 1 of the study) shows the XRD patterns of CeO2 NPs and MPs. All samples exhibited eight typical peaks corresponding to (111), (200), (220), (311), (222), (400), (331), and (420) planes, which are typical of face-centered cubic fluorite structure of CeO2 (standard data JCPDS 34-0394), although peaks for NP samples were significantly broader compared with the reference spectrum because of the small size of the NPs. The lattice parameter (i.e., size of the unit cell in a crystal lattice) for CeO2 fluorite structure is 5.4113 Å (standard data JCPDS 34-0394). The lattice parameters calculated from XRD spectra (averages over the first three peaks) for the nanometer (5.40614 ± 0.01068) and micrometer (5.410548 ± 0.02156) CeO2 particles were found to match well with the standard value. No change in NP lattice relaxation (i.e., increase of lattice parameter) was observed in the case of CeO2 NPs compared with MPs. Our findings are consistent with other studies on nanocrystalline CeO2 prepared by different methods 23.

Figure 4 shows a high-resolution TEM micrograph of CeO2 NPs. It shows the lattice images of CeO2 NPs, illustrating that CeO2 NPs have a well-defined crystalline structure. The interplanar spacing was approximately 0.32 ± XX nm, a typical value of the interplanar spacing of the (111) atomic planes. This value is in good agreement with that measured by XRD and that reported in the literature from HR-TEM for CeO2 NPs 24. It was not possible to observe the lattice image (crystal structure) of MPs because of increased thickness, resulting in opaque images (Fig. 3a). More detailed investigation of the crystallography, surface limiting atomic planes, and 3D reconstruction of particle shape is presented elsewhere 9.

As with other parameters, the use of several techniques is recommended, and XRD and TEM are particularly complementary. X-ray diffraction is a bulk analysis method of sample crystallography, whereas HR-TEM is a single-particle analysis method with attendant advantages and limitations. High-resolution TEM has excellent spatial resolution and requires small sample masses but does not necessarily provide an accurate representation of the sample; XRD requires larger amounts for analysis to ensure the quality of the data but gives a good overview of the whole sample. Transmission electron microscopy operates in vacuum, whereas XRD operates in air; TEM, especially, requires particular efforts to avoid sample perturbation. High-resolution TEM allows the observation of crystal structures of individual NPs (optimal sample thickness <15–20 nm), but not that of the MPs, whereas XRD has better resolution for larger particles, experiencing peak broadening in the nanometer range.

Artifacts may occur with both methods and should be treated carefully. In XRD, amorphization (i.e., transformation from crystalline to amorphous structure) of NPs and overlap of the position of peaks for different minerals can hamper the identification. This is of particular concern for NPs, in that amorphization and peak width increase with a decrease in particle size. Amorphization can also occur in HR-TEM because of the high beam energy and should be accounted for during sample observation by performing time-series analysis and immediate collection of data.

Chemical composition and surface chemistry: X-EDS, EELS, and XPS

Figures 5a and 5b compare an X-EDS spectrum with an EELS spectrum from the same sample (CeO2 NPs). It can be seen clearly that the two methods suggest, as does the XPS (data not presented here), that the particles are composed of cerium and oxygen and confirm the absence of any impurities or contaminants at high concentrations (>1%; see Table 1). The X-EDS data (Fig. 5a) show characteristic peaks corresponding to both cerium and oxygen, and the EELS spectrum (Fig. 5b) shows edges or steps corresponding to cerium and oxygen. In addition, the X-EDS spectrum shows a peak corresponding to Cu from the Cu TEM grid in the analysis. One of the principal assumptions in X-EDS analysis is that the collected X-rays originate from the area of interest in the samples; however, this is not always the case. X-rays can be generated from atoms or particles near the area of interest, and this should be carefully considered when analyzing environmental samples containing NMs. Such considerations may limit the applicability of X-EDS for a fully quantitative analysis of NMs.

thumbnail image

Figure 5. (a) Representative X-ray energy dispersive spectroscopy (X-EDS) spectra of cerium oxide nanoparticles (NPs) taken during transmission electron microscopy (TEM) analysis. The Cu peak is a contamination from the grid material. (b) Electron energy loss spectroscopy (EELS) spectra of cerium oxide NPs. (c) Ce 3d EELS spectra of NPs versus microparticles (MPs). (d) Ce 3d XPS spectra of NPs versus MPs.

Download figure to PowerPoint

Figure 5c shows the Ce-3d EELS spectra for CeO2 NPs and MPs. Figure 5c shows two peaks at approximately 880 to 883 and approximately 897 to 901, indicating the cerium M5 and M4 edges. It is clear that both NPs and MPs have different peak shapes, heights, and positions on the energy scale (chemical shifts). The NPs M5 and M4 edges consist of single peaks at approximately 880 and 987 eV, with higher M5 peak intensity. The MPs M5 and M4 edges consists of two main symmetrical maxima at approximately 883 and 901 eV of equal heights, followed by two peaks of lower intensity (shoulders) at approximately 889 and 906.4 eV. The M5,4 peaks for MPs are centered at higher energy loss than those of NPs, suggesting that MPs contains mainly CeIV, whereas NPs contain a mixture of CeIII and CeIV. Further quantitative interpretation of these samples is given elsewhere 9, 10.

Figure 5d shows an XPS of Ce 3d of CeO2 NPs and MPs. The main peaks occur at kinetic energies of 884 and 900.5 for the NPs and 886 and 902 for the MPs. The XPS spectra of MPs are shifted toward higher binding energy compared with NPs, indicating that the MPs have a higher oxidation state than the NPs, in good agreement with the results obtained by EELS. This ability to quantify and discriminate between oxidation states and chemical environments is one of the major strengths of the EELS and XPS techniques. Quantitative analysis using these methods is particularly useful for CeO2, for which a proposed mode of toxicity is based on the biological reduction of CeIV to CeIII with the production of reactive oxygen species 25. Electron energy loss spectroscopy and XPS (along with synchrotron-based X-ray spectroscopy) are useful tools for quantitatively determining the oxidation state of CeO2 NPs 9, 26, 27, helping to test the hypothetical mode of action. Similarly, EELS and XPS can be applied to other types of NPs; for instance, to distinguish metallic from oxidized NPs (i.e., Ag vs. Ag2O).

Depending on the sensitivity of the material to the type of radiation used, the total dose of electrons to which the sample is exposed, the temperature of the surface, and the level of vacuum, sample degradation may take place. A high vacuum level may remove oxygen and water that are initially trapped within or on the surface of the sample, resulting in an alteration of the sample chemistry. This has been observed with both EELS 9 and XPS 28, and great care should be taken to obtain representative spectra with minimum perturbation.

Surface charge: Laser Doppler electrophoresis

The relationship between pH and surface charge is important in understanding processes such as NP transport in environmental or biological systems, because NP suspensions are generally stable well below and above the IEP. As a rule of thumb, particles are stable for zeta potential (ζ) greater than ± 30 mV (unless sterically stabilized). Figure 6a shows the dependence of the surface charge (represented as electrophoretic mobility) of CeO2 NPs on suspension pH. It shows that CeO2 NPs are positively charged at low pH, are negatively charged at high pH, and have an IEP at approximately pH 8. Our results also indicate that the presence of biological exposure media, buffers, and natural organic macromolecules all influence charge greatly and result in a shift of the IEP 10, as would be expected as a result of sorption and formation of surface coatings, or “coronas.”

thumbnail image

Figure 6. (a) Electrophoretic mobility (EPM; in micrometers.centimeters/volts.seconds) of cerium oxide nanoparticles (NPs) in 10 mM NaNO3 as a function of sample pH. (b) Henry function f(κa) as a function of particle size at a given ionic strength, 1, 10, and 100 mM, calculated from Equation 6. [Color figure can be seen in the online version of this article, available at wileyonlinelibrary.com].

Download figure to PowerPoint

The exact value of the point of zero charge (PZC) may depend on other parameters such as the method applied to determine the PZC (electrophoresis IEP, potentiometric titration “point of net proton charge,” and nonspecific ion adsorption “point of zero net charge”) and the order of titration process (i.e., from acid to base or vice versa) 29. The effect of these parameters on the measured PZC has been investigated elsewhere on other types of particles (goethite and kaolinite) and the parameters were shown to result in a small variability of the order of 1 pH unit 29. More detail on the different methods of determining the PZC and the parameters determined can be found elsewhere 17, 29. In addition, specific sorption of anions (e.g., nitrate, phosphate, and sulfate) and organic molecules (e.g., surfactants, natural organic matter, biopolymers) tend to form inner-sphere surface complexes with CeO2 surfaces, alter the surface charge, and shift the PZC 30. Therefore, nitric acid, phosphoric acid, and sulfuric acid should not be used to determine the PZC.

Nanoparticle surface charge can be determined indirectly by determining the zeta potential of the particles from the measured electrophoretic mobility (µ), defined as the velocity of particles (v) per electric field unit (E), µ = v/E. The ζ can be defined as the electrical potential at the so-called shear plane, a plane that divides the surrounding fluid into two parts, (1) an inner part that moves with the particles because of the strong coupling of the electric double layer to the colloidal surface and (2) an outer part that typically moves in the opposite direction because of the current of the counterions 31. Therefore, the ζ differs from the potential at the particle surface, which can be quantified by titrations 32. The most widely used models of electrophoresis that can be used to calculate ζ were developed by Smoluchowski in 1903 (Eqn. 2) and Hückle in 1924 (Eqn. 3).

  • equation image(2)
  • equation image(3)

where ε0 and εr are the dielectric constant of the vacuum and relative dielectric constant of the medium, respectively, µ is the electrophoretic mobility, and η is the viscosity of the media. The Smoluchowski model assumes that the Debye length (κ−1, the distance over which mobile charge carriers in the vicinity of a solid surface [e.g., NPs] screen the electric field on the surface and can be calculated according to Eqn. 4) is very small compared with the particle radius, a (i.e., κa >> 1), whereas the Hückel model assumes that Debye length is larger than particle radius (i.e., κa << 1) 33.

  • equation image(4)

where e is the elementary charge, zi and ni are the charge number and number concentration of ion i, N the number of ionic species contained in the solution, KB is the Boltzman constant, and T is the absolute temperature.

For an arbitrary value of κa, the electrophoretic mobility becomes a function of κa and can be given by Henry's equation

  • equation image(5)

where f(κa) is a complex function called Henry's function that can be calculated by using a simplified analytical expression (Eqn. 6) derived by Ohshima 34.

  • equation image(6)

Henry's function is related to particle size and double layer thickness. As the double layer thickness has a constant value for a given ionic strength (Eqn. 3), Henry's function becomes a function of particle size only (Fig. 6b) at constant ionic strength, suggesting that knowledge of particle size is essential to calculate zeta potential from electrophoretic mobility. Figure 6b shows that, for a particle radius greater than approximately 200 nm for environmentally relevant ionic strengths (1–100 mM), the Smoluchowski assumption (f(κa) = 1.5) applies, whereas Hückel's assumption (f(κa) = 1.0) applies for particle radius approximately 0.5 to 5 nm for ionic strengths similar to those given above, which is at the lower end of the range of NPs. Therefore, for a particle radius in the range 0.5 to 50 nm, that is, for NPs, it becomes necessary to calculate the value of Henry's function to calculate the zeta potential from the electrophoretic mobility.

It is worth mentioning here that the use of zeta potential for NPs coated in permeable materials such as proteins and humic substances is questionable 35, and electrophoretic mobility or surface charge (by titration) should be reported. This implies the need to apply soft colloid theories in such cases, as described elsewhere 35. Biogenic and geogenic macromolecular coatings are likely to be ubiquitous under realistic environmental and biological conditions 36, so the use of zeta potentials is questionable under in situ conditions. In addition, for nonspherical particles, such as carbon nanotubes or fractal or porous aggregates, zeta potential values are unlikely to be meaningful 37.

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. References

The present study, Parts 1 and 2, provides a full set of properties of CeO2 NPs that has been used in a range of ecotoxicological studies and is currently in use for other studies that will provide a tool to relate the toxic effect of these materials to their properties as evidence accumulates. The present study shows clearly that NM characterization is not a trivial task and is particularly complicated for commercial NMs, for which properties are heterogeneous and poorly controlled 1, 13, 38. In particular, we highlight that particle shape should be reported as a distribution similar to particle size distribution and calculation of zeta potential from the measured electrophoretic mobility should be considered carefully for NMs, because neither the Smoluchowski nor the Hückle assumption is applicable to NMs. Therefore, these findings stress the need for a more careful approach toward the interpretation of NM characterizations in current and past ecotoxicological research, in particular, in light of the misuse of some analytical tools, the lack of standardized and validated sample preparation protocols, and the lack of critical evaluation of the published data.

Characterization of NMs also becomes more complicated in environmental and biological media as a result of various processes that might take place such as agglomeration and sedimentation. Quite clearly, there is no single best technique for particle characterization, and a multimethod approach is essential to enhance data quality, reduce bias, and improve our understanding of the properties of NMs that may control their reactivity and toxicity. Understanding of the physical principles of the different techniques, measured parameters, sample requirement, sample preparation, and data analysis and treatment are key to obtaining representative, coherent, and accurate results.

In addition, it is recommended here, as elsewhere, that NM characterization for ecotoxicological studies should be performed before, during, and after exposure because of the dynamic nature of the system. Different techniques are suitable at different stages of this characterization paradigm. The applicability of the characterization techniques as well as the analysis environment are summarized in Table 1.

Today, two fundamental pressing questions in the area of toxicology and ecotoxicology of manufactured NMs are (1) what properties of NMs should be characterized to make a reliable assessment of NM toxicity? and (2) how are NM properties and effect or behavior linked 39? The nature and extent of characterization required depends on the nature of the study being performed and is hypothesis driven, although it is possible to define a minimum set of characterizations for all studies 1, 13. We recommend that comprehensive characterization of the NMs is performed to ensure data accuracy and completeness and that potentially valuable biological effects information is not lost for future assessment and for data comparability. As knowledge of the most important NP characteristics increases, NP characterization will become more selective.

The properties required for all NMs are synthesis method, size, specific surface area, shape, core composition, surface coating, aggregation state, purity, and stability. Nanomaterial properties are dissolution, structure, and surface speciation. For instance, it is important to determine the dissolution of silver particles, which are highly soluble 11, whereas it is irrelevant to determine dissolution of CeO2 NPs, which are highly insoluble 10. Another example is the determination of titanium dioxide NP structure. Titanium dioxide particles can have different phases such as rutile, anatase, and brookite, which control particle photocatalytic properties, production of reactive oxygen species, and toxicological effects 7, 40.

The second question seems to be far too complicated to tackle presently, given the current state of knowledge. However, we think that this can be achieved through two progressive steps. The first is the synthesis of very well-defined NMs with tailored properties that can be linked to their effects via quantitative structure–activity relationship-type models. The second is an extensive characterization of a set of manufactured NMs that can be used to verify the developed quantitative structure–activity relationship models, which can then be used as predictive models for other manufactured NMs.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. References

We acknowledge funding from the NERC (NE/E009204/1, NE/D004942/1, NE/G010641/1, and the Facility for Environmental Nanoscience Analysis and Characterisation). We also thank Jenny Readman and Louise Male for their help with the X-ray diffraction analysis. The Siemens D5000 Kristalloflex used in the present study was obtained through Birmingham Science City: Creating and Characterising Next Generation Advanced Materials (West Midlands Centre for Advanced Materials Project 1), with support from Advantage West Midlands and part funded by the European Regional Development Fund.

References

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. References
  • 1
    Powers KW, Brown SC, Krishna VB, Wasdo SC, Moudgil BM, Roberts SM. 2006. Research strategies for safety evaluation of nanomaterials. Part VI. Characterization of nanoscale particles for toxicological evaluation. Toxicol Sci 90: 296303.
  • 2
    Baalousha M, Ju-Nam Y, Cole PA, Tyler CR, Stone V, Fernandes T, Jepson MA, Hriljac JA, Jones IP, Lead JR. 2012. Characterization of cerium oxide nanoparticles. Part 1: Size measurements. Environ Toxicol Chem (this issue).
  • 3
    Poland CA, Duffin R, Kinloch I, Maynard A, Wallace WAH, Seaton A, Stone V, Brown S, MacNee W, Donaldson K. 2008. Carbon nanotubes introduced into the abdominal cavity of mice show asbestos-like pathogenicity in a pilot study. Nat Nano 3: 423428.
  • 4
    Yamamoto A, Honma R, Sumita M, Hanawa T. 2003. Cytotoxicity evaluation of ceramic particles of different sizes and shapes. J Biomed Mater Res 68A: 244256.
  • 5
    Pal S, Tak YK, Song JM. 2007. Does the antibacterial activity of silver nanoparticles depend on the shape of the nanoparticle? A study of the gram-negative bacterium Escherichia coli. Appl Environ Microbiol 73: 17121720.
  • 6
    Wiessner JH, Henderson JD, Sohnle PG, Mandel NS, Mandel GS. 1988. The effect of crystal structure on mouse lung inflammation and fibrosis. Am Rev Respir Dis 138: 445450.
  • 7
    Sato T, Taya M. 2006. Enhancement of phage inactivation using photocatalytic titanium dioxide particles with different crystalline structures. Biochem Eng J 28: 303308.
  • 8
    Gaiser BK, Fernandes TF, Jepson MA, Lead JR, Tyler CR, Baalousha M, Biswas A, Britton GJ, Cole PA, Johnston BD, Ju-Nam Y, Rosenkranz P, Scown TM, Stone V. 2012. Interspecies comparisons on the uptake and toxicity of silver and cerium dioxide nanoparticles. Environ Toxicol Chem 31: 144154.
  • 9
    Baalousha M, Le Coustumer P, Jones I, Lead JR. 2010. Characterization of structural and surface speciation of representative commercially available cerium oxide nanoparticles. Environ Chem 7: 377385.
  • 10
    Rogers NJ, Franklin NM, Apte SC, Batley GE, Lead JR, Baalousha M. 2010. Physico-chemical behaviour and algal toxicity of nanoparticulate CeO2 in freshwater. Environ Chem 7: 5060.
  • 11
    Fabrega J, Fawcett SR, Renshaw JC, Lead JR. 2009. Silver nanoparticle impact on bacterial growth: Effect of pH, concentration, and organic matter. Environ Sci Technol 43: 72857290.
  • 12
    Lockman PR, Koziara JM, Mumper RJ, Allen DD. 2004. Nanoparticle surface charges alter blood–brain barrier integrity and permeability. J Drug Target 12: 635641.
  • 13
    Baer DR, Amonette JE, Engelhard MH, Gaspar DH, Karakoti AS, Kuchibhatla S, Nachimuthu P, Nurmi JT, Qiang Y, Sarathy V, Seal S, Sharma A, Tratnyek PG, Wang C-M. 2007. Characterization challenges for nanomaterials. Surf Interf Anal 40: 529537.
  • 14
    Gaiser B, Biswas A, Rosenkranz P, Jepson M, Lead JR, Stone V, Tyler CR, Fernandes T. 2011. The effect of silver and cerium dioxide micro- and nano-sized particles on Daphnia magna. J Environ Monit 13: 12271235.
  • 15
    Kosmulski M, Hartikainen J, Maczka E, Janusz W, Rosenholm JB. 2002. Multiinstrument study of the electrophoretic mobility of fumed silica. Anal Chem 74: 253256.
  • 16
    O'Brien RW, Cannon DW, Rowlands WN. 1995. Electroacoustic determination of particle size and zeta potential. J Colloid Interf Sci 173: 406418.
  • 17
    Sposito G. 1998. On points of zero charge. Environ Sci Technol 32: 28152819.
  • 18
    Baalousha M, Lead JR. 2007. Characterization of natural aquatic colloids (<5nm) by flow-field flow fractionation and atomic force microscopy. Environ Sci Technol 41: 11111117.
  • 19
    Baalousha M, Kammer F, Motelica-Heino M, Coustumer P. 2005. 3D characterization of natural colloids by FlFFF-MALLS-TEM. Anal Bioanal Chem 308: 549560.
  • 20
    Renaud G, Lazzari R, Leroy F. 2009. Probing surface and interface morphology with grazing incidence small angle X-ray scattering. Surf Sci Rep 64: 255380.
  • 21
    Hyeon-Lee J, Beaucage G, Pratsinis SE, Vemury S. 1998. Fractal analysis of flame-synthesized nanostructured silica and titania powders using small-angle X-ray scattering. Langmuir 14: 57515756.
  • 22
    Mavrocordatos D, Perret D, Leppard GG. 2007. Strategies and advances in the characterisation of environmental colloids by electron microscopy. In Wilkinson KJ, Lead JR, eds, Environmental Colloids and Particles: Behaviour, Structure and Characterisation. John Wiley & Sons, Chichester, United Kingdom.
  • 23
    Morris VN, Farrell RA, Sexton AM, Morris MA. 2006. Lattice constant dependence on particle size for ceria prepared from a citrate sol-gel. J Phys: Conference Series 26: 119122.
  • 24
    Chen HI, Chang HY. 2004. Homogeneous precipitation of cerium dioxide nanoparticles in alcohol/water mixed solvents. Colloid Surf A: Physicochem Eng Aspects 242: 6169.
  • 25
    Auffan M, Rose J, Wiesner MR, Bottero JY. 2009. Chemical stability of metallic nanoparticles: A parameter controlling their potential cellular toxicity in vitro. Environ Pollut 157: 11271133.
  • 26
    Thill A, Zeyons O, Spalla O, Chauvat F, Rose J, Auffan M, Flank AM. 2006. Cytotoxicity of CeO2 nanoparticles for Escherichia coli. Physico-chemical insight of the cytotoxicity mechanism. Environ Sci Technol 40: 61516156.
  • 27
    Pfau A, Schierbaum KD. 1994. The electronic structure of stoichiometric and reduced CeO2 surfaces: An XPS, UPS and HREELS study. Surf Sci 321: 7180.
  • 28
    Dengis PB, Gerin PA, Rouxhet PG. 1995. X-ray photoelectron spectroscopy analysis of biosurfaces: Examination of performances with yeast cells and related model compounds. Colloid Surf B: Biointerf 4: 199211.
  • 29
    Appel C, Ma LQ, Dean Rhue R, Kennelley E. 2003. Point of zero charge determination in soils and minerals via traditional methods and detection of electroacoustic mobility. Geoderma 113: 7793.
  • 30
    Baalousha M. 2009. Aggregation and disaggregation of iron oxide nanoparticles: Influence of particle concentration, pH and natural organic matter. Sci Tot Environ 407: 20932101.
  • 31
    Handy RD, Von der Kammer F, Lead JR, Hassellöv M, Owen R, Crane M. 2008. The ecotoxicology and chemistry of manufactured nanoparticles. Ecotoxicology 17: 287314.
  • 32
    Lead JR, Hamilton-Taylor J, Nesketh N, Jones MN, Wilkinson AE, Tipping E. 1994. A comparative study of proton and alkaline earth metal binding by humic substances. Anal Chem Acta 294: 319327.
  • 33
    Elimelech M, O'Melia CR. 1990. Kinetics of deposition of colloidal particles in porous media. Environ Sci Technol 24: 15281536.
  • 34
    Ohshima H. 1994. A simple expression for Henry's function for the retardation effect in electrophoresis of spherical colloidal particles. J Colloid Interf Sci 168: 269271.
  • 35
    Duval JFL, Wilkinson KJ, vanLeeuwen HP, Buffle J. 2005. Humic substances are soft and permeable: Evidence from their electrophoretic mobilities. Environ Sci Technol 39: 64356445.
  • 36
    Lynch I, Dawson KA. 2002. Protein–nanoparticle interactions. Nano Today 3: 4047.
  • 37
    Smith B, Wepasnick K, Schrote KE, Bertele AR, Ball WP, O'Melia C, Fairbrother DH. 2008. Colloidal properties of aqueous suspensions of acid-treated, multi-walled carbon nanotubes. Environ Sci Technol 43: 819825.
  • 38
    Domingos RF, Baalousha M, Ju-Nam Y, Reid M, Tufenkji N, Lead JR, Leppard GG, Wilkinson KJ. 2009. Characterizing manufactured nanoparticles in the environment—Multimethod determination of particle sizes. Environ Sci Technol 43: 72777284.
  • 39
    Alvarez PJJ, Colvin V, Lead J, Stone V. 2009. Research priorities to advance eco-responsible nanotechnology. ACS Nano 3: 16161619.
  • 40
    Hartmann NB, Von der Kammer F, Baalousha M, Ottofuelling S, Baun A. 2010. Algal testing of titanium dioxide nanoparticles—Testing considerations, inhibitory effects and modification of cadmium bioavailability. Toxicology 269: 190197.