Gold nanoparticle aerosols for rodent inhalation and translocation studies
- First Online:
- Cite this article as:
- Möller, W., Gibson, N., Geiser, M. et al. J Nanopart Res (2013) 15: 1574. doi:10.1007/s11051-013-1574-9
The intensive use of nano-sized particles in many different applications necessitates studies on their risk assessment as there are still open questions on their safe handling and utilization. For reliable risk assessment, the interaction of nanoparticles (NP) with biological systems after various routes of exposure needs to be investigated using well-characterized NP. We report here on the generation of gold-NP (Au-NP) aerosols for inhalation studies with the spark ignition technique, and their characterization in terms of chemical composition, physical structure, morphology, and specific surface area, and on interaction with lung tissues and lung cells after 1 h inhalation by mice. The originally generated agglomerated Au-NP were converted into compact spherical Au-NP by thermal annealing at 600 °C, providing particles of similar mass, but different size and specific surface area. Since there are currently no translocation data available on inhaled Au-NP in the 10–50 nm diameter range, the emphasis was to generate NP as small as 20 nm for inhalation in rodents. For anticipated in vivo systemic translocation and dosimetry analyses, radiolabeled Au-NP were created by proton irradiating the gold electrodes of the spark generator, thus forming gamma ray emitting 195Au with 186 days half-life, allowing long-term biokinetic studies. The dissolution rate of 195Au from the NP was below detection limits. The highly concentrated, polydisperse Au-NP aerosol (1–2 × 107 NP/cm3) proved to be constant over several hours in terms of its count median mobility diameter, its geometric standard deviation and number concentration. After collection on filters particles can be re-suspended and used for instillation or ingestion studies.
KeywordsGold nanoparticlesSpark ignitionChain aggregate/agglomerateProton irradiationAu-195 radiolabelX-ray diffraction
Nanoscale particles are widely used in industrial sectors, technology, consumer products, science, and medicine (Seaton and Donaldson 2005; Moghimi et al. 2005; Oberdörster et al. 2005; Maynard et al. 2006; Doane and Burda 2012). Among many recent NP formulations nowadays available gold nanoparticles (Au-NP) have wide applications in science and medicine. The unique surface morphology has allowed various labeling options for cell receptors and cell types including tumor cells (Arvizo et al. 2010; Papasani et al. 2012). In addition, they exhibit light and gamma ray absorption and scattering properties making them novel targets for optical and X-ray enhancement as well as contrast substances (Hainfeld et al. 2006; Rand et al. 2011). However, the safety of NP is mandatory for their use in living organisms including humans.
Many studies suggest that nanometer-sized particles show specific toxicity characteristics in cell systems and living organisms and may follow specific transport and translocation routes in the body, being different from micrometer-sized particles (Donaldson et al. 2004; Oberdörster et al. 2005; Geiser and Kreyling 2010). Materials being non-toxic in the micrometer size range can show toxicity in the nanometer size range (Ferin et al. 1992). In addition, the toxicology metrics of nanomaterials seem not to be mass based but are more related to surface area and particle number. Inhaled nanometer-sized particles may pass into the circulation and be transported to, and accumulate in, secondary organs such as the brain, possibly playing a role in neurodegenerative diseases (Oberdörster et al. 2005; Peters et al. 2006). However, these specific transport properties of NP may also offer new therapeutic opportunities, underpinning the new discipline of nano-medicine in pharmacology (Doane and Burda 2012).
distribution and transport of NP within the respiratory tract,
their potential to cross the air–blood barrier (ABB) into the circulation and accumulate in secondary organs and tissues,
toxicological responses in the lungs and other tissues to the presence and/or accumulation of NP.
In order to study and understand these mechanisms, well-characterized aerosols of nanometer-sized particles are required and in case of studies of quantitative biokinetics, a stable radiolabel is required. Recently, we have described the generation and characterization of titanium dioxide NP aerosols radiolabeled with 48V using spark ignition aerosol generation technology (Kreyling et al. 2011). The Au-NP described here fulfill all these requirements. Using the spark discharge method, NP can be produced as an aerosol for direct inhalation. Gold can be radiolabeled by neutron or proton bombardment yielding different gold isotopes, such as 198Au or 195Au, which are gamma emitters and can be quantified using gamma ray detectors. Electron microscopic analysis of NP at the individual particle level in tissues or cell cultures is also highly important as it provides information about the cells and intracellular components which are the targets of (inhaled) NP. These analyses contribute to the understanding of underlying mechanisms and the modes of action that may eventually lead to adverse health effects. Prerequisites for such studies include adequate particle application (i.e., inhalation for lung targeting) and tissue preservation, representative tissue and particle sampling for quantitative analysis (unbiased stereology), as well as unambiguous particle identification (Geiser and Kreyling 2010; Kreyling et al. 2013). Recently, we reported on an Au-NP inhalation study, in which we had used thermally untreated Au-NP agglomerates/aggregates (Takenaka et al. 2012). Based on this study, we have further developed our aerosol generation technology, as we describe here. In order to study particle biokinetics in the physiological (non-toxicological) regimen, our studies focused on inhalation studies of only few hours, where no toxicological signs in the lung or other organs could be seen. However since the aerosol is stable over hours and for several consecutive days, long-term exposure studies could be conducted. In addition, particles could be collected on filters and re-suspended for use in ingestion or instillation studies.
Materials and methods
Generation of Au-NP using spark ignition technology and online inhalation
In order to obtain compact spherical Au-NP, the agglomerate aerosol was heated up to 600 °C in a tube furnace (30 cm length) downstream of the Kr-85 neutralizer. The 600 °C heat treatment caused instantaneous melting of the agglomerates, forming single compact particles. Although the melting point of bulk gold is 1,064 °C, there is a significant melting point depression of Au for particles in the nanometer size range (Buffat and Borel 1976). Since the primary particles within the aggregates are only few nanometers in diameter they melt at temperatures of 600–700 °C. Therefore, the generator can provide two different types of aerosol for inhalation studies, being composed either of agglomerates/aggregates of small primary particles or of larger compact spherical particles.
The aerosol was fed through a rodent inhalation apparatus maintained with an absolute filter and pump unit to adjust the aerosol pressure to between 20 and 40 Pa below the laboratory pressure for radiation safety reasons (Alessandrini et al. 2008; Kreyling et al. 2011). Using this system, an aerosol that is stable over several hours is provided to the animals for inhalation within less than 5 s of the initial particle formation. After 1 h Au-NP inhalation by intubated mice, animals were killed and the lungs were processed for electron microscopy as described previously for titanium dioxide aerosol inhalation studies in rats (Geiser et al. 2005, 2008).
Radiolabeling of Au-NP
One end of each of two pure cylindrical gold electrodes (diameter 3 mm, length 4–5 mm) was proton activated in the Scanditronix MC40 cyclotron at JRC Ispra, Italy (Holzwarth et al. 2012). The proton energy was set at a value of 36 MeV. The electrodes were mounted in a closed loop water cooling system with the protons first passing through a 300 μm aluminium window and about 1–2 mm of water before impinging on the electrode tips, still with an energy high enough to efficiently create 195Hg via the (p,3n) nuclear reaction, together with various other isotopes, including 195Au via the (p,p2n) reaction. The two 195Hg isomers created have half-lives of 41.6 and 9.9 h, so the electrodes were stored after irradiation for several weeks to allow decay of the 195Hg into 195Au. This isotope of gold decays via electron capture into 195Pt with a main gamma emission at 98.9 keV and Pt X-ray emissions at 65.1, 66.8, and 75.8 keV among others (Wilson 1966; Firestone and Ekström 2004). 195Au has a half-life of 186.1 days and thus is ideal for radio-tracing studies over several weeks or months.
The irradiation was carried out over a period of 5 days for a total of 40 h irradiation. The ion beam current was 18 μA, but due to the beam diameter of several mm, only a fraction of the beam actually impinged on the gold electrodes. The depth to which the electrodes were activated was about 500 μm, and the total 195Au activity of the two electrodes after the irradiation and decay period of some weeks was about 50 MBq. A significant activity of 196Au (about 27 MBq) was also present, but this decays with a half-life of 6.2 days, therefore, decreases greatly over several weeks and was negligible during the inhalation studies. A rough estimate of the average 195Au activity concentration in the activated layer gives a value of 370 kBq/mg, but the actual activity concentration would vary somewhat with depth, with a higher activity concentration closer to the surface of the electrodes, tailing off to zero after several hundred microns.
Aerosol size characterization
Aerosol concentration and size distribution were continuously measured using a condensation particle counter (CPC 3022, TSI, Aachen, Germany) and a differential mobility particle size spectrometer (DMPS, consisting of a model 3071 differential mobility analyzer and a model 3010 CPC, TSI, Aachen, Germany), as shown in Fig. 1. The analysis of concentration and particle size was performed at the entry point to the inhalation apparatus in order to ensure correct measurement of the aerosol properties at the time of inhalation. Since the heat-treated Au-NP were solid spherical shaped NP their surface area was calculated using volume determination together with the bulk density of gold.
Au-NP sampling for morphology and chemical analysis
Agglomerate Au-NP (without heat treatment) as well as heat-treated Au-NP were directly deposited onto both silicon wafer substrates and transmission electron microscope (TEM) grids. An electrostatic precipitator (Model 3089, TSI Inc., Aachen, Germany) was connected to the spark ignition aerosol generation system to deposit Au-NP on a 2.5-cm-diameter, single crystalline silicon substrate for chemical and crystallographic analyses. In addition, Au-NP were sampled on formvar-coated copper TEM grids using a bipolar TEM sampler (FHNW bipolar TEM sampler, IAST, University of Applied Sciences of Northwestern Switzerland, Switzerland) (Fierz et al. 2007).
Au-NP morphology and crystalline structure analysis
Conventional TEM, the corresponding selected area electron diffraction patterns (SAED) and high resolution microscopic imaging (HRTEM) of the specimens were investigated using a EM10CR (Zeiss, Oberkochen, Germany) and a FEI Titan 80-300 TEM microscope equipped with a Cs corrector for the objective lens. A Fischione high angle annular dark field detector (HAADF), GATAN post-column imaging filter and a cold field emission gun operated at 300 kV as an acceleration voltage was used (Takenaka et al. 2012). Elemental microanalysis of NP was performed by the energy dispersive spectroscopy (EDX) using the same instrument described above.
Ultrastructural information of NP in three spatial dimensions was evaluated using electron tomography (3D), whereby NP projections acquired from multi-angular tilt series were reconstructed (Baumeister 2002). We performed 3D electron tomography on Au-NP collected from the aerosol line on formvar-coated TEM grids using a Tecnai F20 field emission TEM (FEI Company, Eindhoven, Netherlands) operating at 200 kV. Typical tilt series were acquired at a magnification of 34,000× covering an angular range between −70° and 70° (with an increment of 2°). The pixel size in the tomograms was 0.59 nm. Semi-manual alignment and reconstruction were performed using the FEI software package (Inspect 3D, Amira, FEI Company, Eindhoven, The Netherlands). For image processing, the ImageJ software equipped with the volume viewer plugin was used.
X-ray diffraction (XRD) of Au-NP for crystalline structure analysis was performed using a PANalytical X’Pert MPD PRO system. The instrument was equipped with Ni-filtered Cu Kα (λ = 0.154 nm) radiation, ¼° fixed divergence, primary and secondary Soller slit with 0.04 rad aperture, circular sample holder with 16 mm diameter, and X’Celerator detector (position sensitive in a range of 2.122° 2θ with 127 channels, yielding a channel width of 0.01671° 2θ), applying a continuous scan in the range of 15°–140° 2θ and an integration step width of 0.0334° 2θ. The measuring time step was 485 s, to be divided by 127 channels to yield effective measuring time of 3.81 s/0.0334° 2θ. The structural parameters were extracted using Rietveld refinement with the BRASS program. Background, scale factor, unit cell parameters, Gaussian, as well as Lorentzian peak width parameters were simultaneously refined. The structural model (ICSD-52700) for cubic Au was used for refining the X-ray patterns. The quality of the refinement was evaluated in terms of the usual R factor (Rwp) and the background corrected residual R′p2. A volume-weighted average crystallite size (dXRD) for Au-NP was determined from the line-broadening analysis. The instrumental contribution to the peak broadening was taken into account during the full profile fitting by instrumental parameters derived from a fit of standard crystalline LaB6.
In addition, samples were analyzed by XRD using a dedicated glancing angle diffractometer designed and constructed at JRC Ispra (Gibson et al. 2000). The system is optimized for analyzing thin films and thus also is ideal for very small sample quantities deposited on Si wafers (Holzwarth et al. 2012). It utilizes Cu Kα radiation, a Soller slit in the diffracted beam, and a solid state detector for background reduction and energy discrimination. The entrance slit can be set to match the chosen incident angle and sample size, and a laser alignment system is used for very accurate sample positioning. The overall angular resolution of the system depends on entrance slit width, but is in the region of 0.15°.
Au-NP analysis in tissue samples
Au-NP TEM analysis in ultrathin tissue sections was performed on chemically fixed lungs (phosphate-buffered 2.5 % glutaraldehyde) after a 1 h inhalation by intubated mice as described for a similar inhalation exposure of rats using titanium dioxide NP aerosols (Geiser et al. 2008). Tissue blocks and cell pellets were further post-fixed with osmium tetroxide and uranyl acetate. Finally, ultrathin sections were stained with heavy metals (uranyl acetate and lead citrate) for conventional TEM analysis using a CM12 TEM (Phillips, Eindhoven, The Netherlands) operating at 80 kV.
The specific surface area of the thermally treated Au-NP was calculated to be 13.5 m2/g for 23 nm diameter spheres, using the density of 19.3 g/cm3 of bulk gold. The corresponding volumetric specific surface area was 260 m²/cm³. Based on TEM measurements of the primary particles of 5–7 nm diameter, there is 3–5 times higher specific surface area of 62 m2/g for the non-treated agglomerate/aggregate Au-NP, while having the same mass.
The 195Au radioactivity concentration of the aerosol was 1.2 kBq/L at the reference date. Based on radioactivity measurement of an integral filter sample of 0.3 L/min aerosol flow during the entire exposure and the sampled aerosol volume, the aerosol mass concentration was also calculated to be ≈1.6 mg/m3. As would be expected, the 195Au radiolabels were stably integrated into the Au-NP since the radiolabel leaching rate obtained in distilled water by a filter sandwich technique (Kreyling et al. 2011) was below detection limits within the first day.
Results of the TEM ultrastructure analysis
The two different types of Au-NP obtained from different preparation techniques were investigated using TEM microscopic imaging. Without heat treatment, the Au-NP were obtained in the long chain-like agglomerated form (see Figs. 3a, 4a) as observed in our previous study (Takenaka et al. 2006, 2012). Although the particles are not sufficiently dispersed, they are crystalline with well-formed lattice planes with neck-to-neck contact. One of the single crystalline particles of the long agglomerate attached to the carbon grid is shown in Fig. 4b. The lattice spacing of this crystalline particle in the (1 1 1) plane was found to be 0.2354 nm which agrees with the highest intensity peak of the Au-NP in the XRD analysis at 38.18 2θ (2.3550 Å; see below and Fig. 8). The high resolution microscopic imaging of the long chain agglomerates and the selected particles from the agglomerate are presented in Fig. 4d, e. The lattice spacing measured at the multiple spots of the agglomerates ranged from 0.234 to 0.238 nm, supporting again phase pure Au-particles. The Au-agglomerates were analyzed using selected area diffraction patterns and the occurrence of the crystallographic rings show the high crystallinity of the particles (Fig. 4f).
After in situ heat treatment of these Au-NP at 600 °C, the primary Au-aggregates melted and formed compact spherical and/or potato-like Au-NP. The particles were well dispersed on the TEM grid (see Figs. 3, 5a) ranging from 10 to 30 nm in diameter. The magnified and Fourier transformed image of a particle observed in Fig. 5a is shown in Fig. 5b. The Au crystallite size was found to ranging from 10 to 25 nm in diameter with a well-formed lattice, suggesting, that the larger particles may be poly-crystalline. The hexagonal arrangement (in the form of dark or bright spots) was also observed within the crystal. The lattice spacing in the (1 1 1) plane and the hexagonal arrangement (two crossed directions) were found to be 2.39 and 2.38 Å, respectively; characteristic of Au-particles (see Fig. 5b, c). The investigation of the particle morphology in different spots revealed that the particles were spherical or potato-like in the size range of 10–30 nm (see Fig. 5d). These Au-particles were also found to be single or poly-crystalline with a similar lattice spacing (2.35–2.38 Å) as described above (see Fig. 5e, f).
X-ray diffraction results
Inhaled Au-NP in lung tissue
Recently, we reported on an Au-NP inhalation study in which we had used thermally untreated agglomerate/aggregate Au-NP (Takenaka et al. 2006, 2012). Based on these studies, we have further developed our aerosol generation technology in order to produce radiolabeled compact Au-NP, which we describe here.
The described spark ignition generator is capable of providing ~20 nm Au-NP aerosols at high particle number concentration of 1–3 × 107 NP/cm3 with a lognormal size distribution (median mobility diameter 20–25 nm and a geometric SD of 1.6) and the mass concentration was ~1.6 mg/m³. This number concentration is the upper limit above which diffusional coagulation is likely to modify the size distribution and number concentration within seconds. Therefore, the inhalation was set up immediately after particle generation. In our pervious study using agglomerate/aggregate Au-NP inhalation lower number concentration of only 4 × 106 NP/cm3 was chosen with 16 nm mobility diameter (Takenaka et al. 2006). The chain-like structure of the non-heat-treated aerosol was also present in this study, but because of lower number concentration there were many single primary 5–7 nm particles present in the previous study.
The aerosol production was stable over several hours and for several consecutive days. The freshly produced NP aerosols may be used within a few seconds for inhalation experiments in rodents in order to specifically study deposition of these very small particles in the respiratory tract and their interactions with lung tissues. Such small-sized NP aerosols for inhalation can usually not be achieved when dispersing NP powders since the powder cannot be completely dis-aggregated and the particles remain as sub-micrometer-sized agglomerates/aggregates (Ma-Hock et al. 2007). Other techniques for studying NP biokinetics in the lung use intratracheal instillation of a NP suspension (Semmler-Behnke et al. 2008).
Nano particles suspension instillation has the advantage that a known dose is transferred into the lung allowing dose response studies on biokinetics and toxicology. However, there are also several major disadvantages, since a single bolus of a high dose is instilled at one time into the lung, which cannot be achieved during inhalation over several hours. Therefore, the extremely high dose rate and the aqueous volume of the suspension may imply adverse reactions to the lung tissue and non-physiological conditions. In addition, since NP suspensions do usually not contain single NP, but NP agglomerates (Porter et al. 2008), the principle concept of whether NP dispersion instillation studies can mimic NP inhalation exposure is questionable.
Hence, the spark ignition particle generation methodology allows biomedical and toxicological studies with focus on the interactions of the smallest NP with respiratory tract tissues and subsequently, after passing the ABB and transport via blood circulation, with tissues of secondary organs. This is a field which is largely unknown despite the fact that the increasing use of new types of engineered NP should imply risk and safety assessments of these particles.
Nano particles were characterized very thoroughly for their aerosol characteristics and physico-chemical properties. Spark ignition generation with or without subsequent heat treatment allows the production of two different types of NP, namely agglomerate/aggregate NP and compact spherical NP with similar particle masses. Since the agglomerate/aggregate NP consist of various much smaller primary particles of 5–7 nm diameter (Takenaka et al. 2006; Liu et al. 2012), they differ significantly in their morphological structure and surface area. Because of the high density of gold (19.3 g/cm3), there is a threefold to fivefold decrease of specific surface area after heat treatment of the Au-NP. Previous toxicology studies gave evidence for the specific surface area to be the major determinant in particle toxicology and not the mass (Stoeger et al. 2006; Oberdörster et al. 2005), and for a threshold surface area above which the organism shows adverse responses. It is expected that the surface area determines the major mechanisms of interaction of the NP with the environment within biological systems (Nel et al. 2009), such as surfactant coating, NP-protein interactions and subsequent transport and translocation processes across cellular membranes and barriers, such as the ABB or the blood–brain barrier (BBB) (Kreuter 2004). Therefore, while delivering similar masses within an inhalation study significantly different surface areas can be applied allowing direct study of these influences.
Proper melting of the non-heat-treated agglomerate aerosol consisting of 5–7 nm primary particles at 600 °C, far below of the melting point of bulk gold (1,064 °C), was also shown in a previous study (Liu et al. 2012), but the authors do not report on particle morphology and crystal structure. In our study, both the primary particles of the Au-NP agglomerates and the heat-treated (melted) spherical Au-NP show the crystal structure of bulk gold with no amorphous component. This may suggest similar internal structure and surface properties of the particles (beyond effects due to decreasing particle size). Therefore, chemical stability and interactions in biological systems and possible coating by proteins in nano-medicine applications may be comparable in both particle types.
For NP radiolabeling, proton irradiation of the gold electrodes at a cyclotron caused a nuclear reaction by which 195Au isotopes were created only in the electrode tips. During spark ignition, 195Au isotopes evaporated together with stable Au atoms from the surface of the electrode and condensed to primary Au-particles. The 195Au-NP aerosol yielded a specific activity of 1.2 MBq/m3 and 0.75 MBq/mg NP mass. Since only the tip of the gold electrode is activated the specific 195Au radioactivity may decreases during erosion of the first few milligrams of Au of the gold electrodes, resulting in a decrease of specific 195Au activity of the NP aerosol. However, at the aerosol mass concentration of about 1.6 mg/m3 produced at a total aerosol flow rate of 10 L/min, about 1 h of continuous aerosol operation is required to erode each mg; therefore, major changes would only occur after more than 30 rodent exposures lasting 1 h each.
The estimated deposited dose after 1 h Au-NP inhalation of the aerosol with a mass concentration of 1.6 mg/m3, assuming a respiratory minute volume (RMV) of 0.04 L/min (Kuehl et al. 2012) and a deposition fraction of 0.35 of the inhaled aerosol is 1–2 μg. This Au-NP dose did not show any apparent toxicological signs in the lungs of the animals. Therefore, it can be assumed that the particle biokinetics happens under physiological (healthy) conditions. The deposited dose after this 1 h inhalation is by orders of magnitudes lower compared to other toxicological studies, which usually use single bolus instillation of micron- or nanometer-sized particles (Kreyling et al. 2013). Particle identification in the various lung compartments and lung cells after inhalation was previously described for Ir- or TiO2-NP (Geiser and Kreyling 2010), but the specific distribution and biokinetics of the Au-NP used in this study needs further investigation.
However, for use in medicine thorough toxicology of Au-NP has to be evaluated, i.e., the successful use of Au-NP as X-ray contrast agent in pre-clinical animal studies required much higher doses (10 mg/mL blood) (Hainfeld et al. 2006) as used in our study (2 μg), and the stability of the Au-NP in the organism and possible accumulation in secondary organs may prevent the application in humans. Since toxicology of nanomaterials seems to be primarily related to the specific surface area, including conduction band energy levels (Zhang et al. 2012), and not to the mass, the particles described here are suitable for such hypothesis testing, because particles with similar mass but different specific surface area can be produced. In addition, toxicological reactions have to be expected when the delivered dose (surface area) exceeds certain threshold values (Stoeger et al. 2006). Higher exposure doses can easily be produced by longer exposure times, including accumulating exposures over days and weeks and other than inhalation exposure is possible after particle collection on filters and re-suspension in a medium.
Spark ignition technology has been used for reproducible production of highly concentrated polydisperse Au-NP aerosols (1–3 × 107 NP/cm3; ~1.6 mg/m3, ~0.9 mg/h) which proved to be constant over several hours in terms of their CMD, geometric standard deviation and number concentration. Either chain aggregate/agglomerate Au-NP or compact spherical Au-NP were produced, respectively, without or with 600 °C heat treatment of the aerosol downstream of the spark ignition generator. The Au-NP were extensively characterized in terms of their chemical composition, physical structure, morphology, and specific surface area. Gamma emitting Au-NP were produced with long half-life of 186 days after radiolabelling of the gold electrodes using a proton beam of a cyclotron accelerator, forming 195Au isotopes. The radiolabel was stable since the radiolabel leaching rate in water was below detection limits.
This work was partially funded by HMGU: EU FP7 Projects NeuroNano, NMP4-SL-2008-214547; ENPRA, NMP4-SL-2009-228789; InLiveTox, NMP4-SL-2009-228625; UBern: Swiss National Science Foundation Project 310030-120763; UBremen: S.P. and L.M. would like to thank National Science Foundation and the Environmental Protection Agency under Cooperative Agreement Number DBI-0830117 for supporting this work. Key support was also provided by the US Public Health Service Grants U19 ES019528 (UCLA Center for NanoBiology and Predictive Toxicology), RO1 ES016746, and RC2 ES018766. S.P. and L.M. would also like to thank Prof. A. Rosenauer and M. Schowalter, Department of Physics, University of Bremen, for TEM and EDX analysis.