Iron nanoparticles by inert gas condensation
Xing, Lijuan
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2018
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Xing, L. (2018). Iron nanoparticles by inert gas condensation: Structure and magnetic characterization. University of Groningen.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Chapter 3
Synthesis and Morphology of Iron‐Iron Oxide Core‐
Shell Nanoparticles
C
ore‐shell structured Fe nanoparticles (NPs) produced by high
pressure magnetron sputtering gas condensation were studied
using transmission electron microscopy (TEM) techniques, electron
diffraction, electron energy‐loss spectroscopy (EELS), tomographic
reconstruction, and Wulff shape construction analysis. The core‐
shell structure, which is composed of an iron (Fe) core surrounded
by a maghemite (γ‐Fe
2O
3) and/or magnetite (Fe
3O
4) shell, was
confirmed by Fast Fourier transform (FFT) analysis combined with
EELS. It was found that the particle size and shape strongly depend
on the gas environment. Moreover, extensive analysis showed that
NPs with a size between 15 ‐ 24 nm possess a truncated cubic
morphology, which is confined by the 6 {100} planes that are
truncated by the 12 {110} planes at different degrees. For NPs larger
than 24 nm, the rhombic dodecahedron defined by the 12 {110}
planes is the predominant crystal shape, while truncated rhombic
dodecahedrons, as well as non‐truncated and truncated cubic NPs
were also observed. The NPs without truncation showed a
characteristic
inward
relaxation
indicating
that
besides
thermodynamics also kinetics play crucial role during particle
3.1 Introduction
Magnetic core-shell nanoparticles (NPs), which are typically constituted by a ferromagnetic core surrounded by a weakly magnetic or ferrimagnetic oxide shell, are considered promising candidates for applications in magnetic recording 1–3,
microwave absorption 4–6 , ground water remediation 7 as well as biomedical
treatment and analysis such as drug delivery 8–11, magnetic resonance imaging (MRI)
contrast enhancement 12,13 and hyperthermia 14–16. Strong impetus has been given to
the study of iron nanoparticles in biomedical applications due to their high magnetization coupled with the biocompatibility. While zero-valence iron (Fe0)
nanoparticles exposed to air will be quickly oxidized forming a thin layer of oxide (typically with an equilibrium thickness of 2-3 nm, showing no systematic dependence on the particle size 1,17) on the surface leading to a core-shell structure 18. The oxide shell not only can protect the metallic core from further oxidation (for
NP sizes larger than 8 nm), but also it can modify the magnetic properties of nanoparticles 19. The resulting iron-iron oxide core-shell NPs are potentially ideal
materials for tumor diagnostic and therapeutics since they can be used respectively in MRI 12,20–22 and hyperthermia 21,23–26. In MRI, the magnetic (core-shell form) NPs
serve as contrast enhancement agents which is expected to be more effective than the contemporarily used iron oxide NPs due to the higher saturation magnetization
21, while in hyperthermia iron NPs are able to generate localized heat and kill
malignant tumor cells selectively. This is achieved by applying an alternating field where the iron NPs due to the greater magnetization may result in higher heating efficiency at a safe biological frequency ranges (50 kHz to a few hundred kHz) 21,25
compared to iron oxide NPs. For NPs with a size smaller than 8 nm, the exposure to air could result in full oxidation throughout the whole particle 27.
Iron NPs can be produced by various methods, including sputter gas condensation 28,29, gas evaporation 30, reduction of iron salts/oxides and thermal
decomposition 31. The high pressure condensation system used in our work is a
combination of magnetron-sputtering and the inert gas aggregation technique giving solvent-free (without extra processing to remove chemical contaminants) and monodispersed iron NPs with higher diversity of crystal motifs than chemical techniques, which can be also prepared with tunable size. Regardless of the synthesis method of Fe NPs, the subsequent exposure to air will result in the initial oxidation of particles leading to the desired core-shell nanoparticles. The properties of nanoparticles could be tuned by proper control of their size and shape 32–34. It is
well known that tremendous changes in properties can be observed when the physical size of matter is reduced down to nanometer length scales 35,36. The
properties of nanoparticles, normally distinct to their bulk counterparts 37, show
strong sensitivity to the size. This modification of properties can be attributed to the increased surface-volume ratio38, the different geometric, electronic, and
magnetic structure at the atomic level of the metal particle surface and the metal– oxide interface 39. As a result, besides the size of the particle, the particle shape,
which is closely related to the crystallographic surfaces that enclose the particle, also contributes largely to the chemical and physical properties of nanoparticles. Different surfaces of metal nanocrystals have different atom densities, electronic structures, bonding and consequently different surface relaxations, energies as well as possibly chemical reactivities 40,41, leading therefore to distinct properties of the
enclosing particles.
In fact Saito et al. 42 and Hayashi et al. 43 have made systematic investigations
on the morphological structural features of ultrafine metal particles including Fe NPs. In their study, Fe metal was heated in vacuum until melting, and the subsequent heating by electric current produced evaporation at high temperature (HT) of 1850-2000 . The vapor was collected on liquid N2 cooled grids where Fe
NPs were formed with diameters in the range of 10 to 100 nm. Their results showed that NPs mainly crystallized as truncated or non-truncated rhombic dodecahedron depending on the deposition temperature. The rhombic dodecahedrons were enclosed by the 12{110} planes and partly truncated by the 6{100} with different degrees of truncation. Wang et al. 28 synthesized Fe NPs of variable mean size of
diameters from 2 to 100 nm by a nanoparticle source, which is high pressure gas magnetron-sputtering with water cooled aggregation volume, and it is termed as room temperature (RT) NP deposition. Convergent beam electron diffraction (CBED) was carried out on large particles to confirm that these particles possessed a bcc structured NP core, i.e. α-Fe. Comparison has been made between the morphology of Fe particles deposited both at HT and RT conditions. Indeed, Fe particles produced by RT sputter-gas-aggregation process showed significant different morphology from those observed at HT. It was found that the particles were predominantly faceted on the {100} planes, forming a cubic morphology, and showed different degrees of truncation by the {110} planes. Truncated rhombic dodecahedrons, faceted with {110} and truncated by {100} planes, were also observed but regular rhombic dodecahedron were not found at RT. In addition, in all studies so far, no particular attention was paid to the precise structure and constitution of the NPs inner core, which is of primary importance for magnetic applications.
Therefore in our study the focus is on the core-shell structure of Fe NPs produced by RT sputter-gas aggregation, where also strong attention has been paid
to the constitution and structure of NPs with sizes of ~20 nm, which was not feasible for aforementioned CBED studies 28. Indeed, the crystal habiting planes of
Fe NPs produced by NP deposition at RT appeared to have a predominant morphology that varied with particle size. Regular rhombic dodecahedrons, which were absent in the previous RT deposition report, were found to be the dominant shapes in our experiment for NP sizes larger than 24 nm. In contrast to previous studies we demonstrate that without mass filtering it is still possible to produce the Fe core – oxide shell particles with relatively narrow size distribution.
3.2 Experimental Methods
Fe NPs were deposited directly onto holy carbon-coated transmission electron microscopy (TEM) grids with a home modified high pressure magnetron sputtering source (www.mantisdeposition.com). The sample chamber was evacuated to a base pressure of ~10-8 mbar (aggregation chamber: ~10-6 mbar). The applied sputtering
current was 0.250 A and the aggregation length was 5 cm. Supersaturated Fe vapor was produced by magnetron sputtering of an Fe target (diameter 50.8 mm, thickness 0.5 mm, purity 99.9% from K•tech Supplies LTD) in an inert gas atmosphere (Ar: 15-50 sccm, He: 0-20 sccm. Besides Ar and He, also minute amounts of CH4 was
used at to achieve system pressures ~10-5 mbar (yielding a reducing gas
environment) prior to inert gas insertion and subsequent deposition 44. ) The
supersaturated Fe vapor is then mainly cooled by collision with Ar ions forming NP nuclei and subsequently the desired size NPs. The formed NPs in the aggregation chamber were then carried by mainly Ar (main sputtering gas) and, possibly He, to the sample chamber leading to soft landing (and therefore no NP deformation) onto various substrates. Fe nanoparticles exposed to ambient (e.g., during transfer to TEM) and/or to oxygen containing atmosphere oxidize rapidly, forming a thin oxide layer surrounding the Fe core, which thus leads to a core-shell structure. The particle size can be adjusted in the range of 5-100 nm by controlling the aggregation length, the pressure, the type of sputtering gas, and the magnetron power 44,45.
Furthermore, high-resolution transmission electron microscopy (HRTEM) and electron diffraction analysis were performed locally on a JEOL 2010(F) operating at 200 kV. High angle annular dark field (HAADF) imaging in combination with EELS were carried out simultaneously in a scanning transmission electron microscope (STEM) at the Institute for Electron Microscopy and Nanoanalysis at the Graz University of Technology using a probe-corrected FEI Titan3 G2 60-300 microscope equipped with a Schottky field-emission gun
(X-FEG) operating at 300 kV. A convergence semi-angle of 20 mrad was used. The size of EEL spectrum images was 200 x 200 x 2048 pixel, with a pixel size of 0.17 nm in x and y direction and 0.25 eV in z direction ranging from 400 to 912 eV with an exposure time 1 ms per spectrum. Using the dual EELS method, low loss spectra including the Zero-Loss, which consists of elastically scattered electrons, were acquired simultaneously to correct the data in terms of possible energy shifts. The elemental maps are formed by inelastic scattered electrons that have lost energies corresponding to inner-shell ionization edges, which are characteristic for particular elements, allowing thus spatially resolved chemical analysis of materials down to atomic level 46,47. For analysis of the 3D shape of a particle, tomographic
reconstruction 48,49 was applied to a single HAADF STEM projection employing
prior knowledge about particle symmetry (see Figure 3.8 in Appendix).
3.3 Results and Discussion
3.3.1 NP Size Distributions
Typical TEM images of NPs prepared without and with helium (He) present in the sputtering gas (besides Ar as the main sputtering gas) are shown in Figure 3.1(a) and (c), respectively. NPs with different sizes and morphologies can be observed, showing an inner contrast variation which suggests the presence of a core-shell structure. The corresponding size distributions are shown in Figure 3.1(b) and (d). The presence of He leads to a considerable decrease of the NP size, and the width of the size distribution indicating the strong sensitivity of the NP formation on the sputtering environment. The reduction of NP size as well as the size distribution may be attributed to the higher thermal conductivity for He compared to Ar, which could allow faster spread of thermal energy away from the sampling position and lead to shorter condensational growth of particles resulting, therefore, in smaller size and a more dense size distribution 50. Moreover, the introduction of He in the
aggregation volume can further increase the pressure difference between the main and aggregation volume and shortens thereby the dwell time of the nanoparticles in the aggregation volume, which also contributes to the decrease of the mean particle size and distribution.
Figure 3.1 Low magnification overview images of Fe nanoparticles fabricated without He (a) and with He (c) show NPs with inner contrast variation, which indicates a core‐shell structure. The corresponding size distributions are shown in schematics (b) and (d) with D representing the mean NP diameter, and δ the standard deviation. It can be observed that the presence of He leads to a considerable decrease of the NP size and the width of the size distribution.
Moreover, since Fe is prone to oxidize, combined with the high reactivity in general of NPs, the Fe NPs are highly sensitive to oxidation. Hence, in order to minimize oxygen contamination and any remnant oxidizing impurities into the NP core, reducing gas CH4 was introduced, at low concentrations, into the NP
deposition system. The CH4 would react in situ with remnant oxygen and water
vapor during the sputtering process and convert these impurities into volatile species that could be pumped away. On the other hand, the byproduct of the involved gas reactions, i.e., carbon, can aid the nucleation of nanoparticles, which circumvents any weakening to the nucleation process resulting from the absence of nucleation seeds caused by system purification 44. Subsequently the increased
200 nm
200 nm
(a)200 nm
200 nm
(c)
20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 Co un t Size [nm](b)
D=34.8 nm =3.79 14 15 16 17 18 19 20 21 22 23 24 0 20 40 60 80 100 C ount Size [nm](d)
D=17.5 nm =1.28nuclei concentration can help to narrow the size distribution of particles in order to obtain monodispersed Fe NPs (with a typical variation smaller than 10%) 28.
3.3.2 NP Structure
HRTEM images of the core-shell nanoparticles deposited on a holy carbon-coated TEM grid are shown in Figure 3.2(a) and (b), the corresponding fast Fourier transforms (FFT) indicate body centered cubic (bcc) iron (α-Fe) and wustite (FeO), respectively. Since we cannot exclude the possibility that some remnant oxygen and water contamination within the deposition system may lead to partial oxidation of Fe, then FeO can also be an alternative phase of the NP core. The FFT of the shell part of the HRTEM image shown in the inset of Figure 3.2(c) indicates a magnetite (Fe3O4) or maghemite (γ-Fe2O3) shell, and it remains a cumbersome
problem to distinguish the two phases though many efforts have been made by the existing surface analysis techniques including X-Ray Photoelectron Spectroscopy (XPS) 51 and EELS 52. A characteristic selected area electron diffraction (SAED)
pattern from an area containing several nanoparticles is also given in Figure 3.2(d). Assuming Fe3O4 or γ-Fe2O3 as a reference, the corresponding lattice spacings can
be obtained from the diffraction spots in the diagram as indicated in Figure 3.2(d). However, spots showing a lattice space of 0.206 and 0.149 nm can overlap with those arising from the diffraction of either α-Fe or FeO. Therefore, it is not possible at this stage to unravel, with absolute certainty, the specific constitutions of the NP core and shell based solely on HRTEM analysis. This point requires further analysis since the knowledge of the NP core structure is crucial for widespread applications of Fe NPs.
Therefore in order to investigate further the NP core constituents we show in Figure 3.3(a) an HAADF-STEM image of a single particle, representing the Z-contrast of the material, with a lateral dimension of ~30 nm. The corresponding elemental maps of O-K and Fe-L2,3 are given in Figure 3.3(b) and (c), respectively.
Both measurements show a clear core-shell structure, consistent with the HAADF image. Although oxygen was detected in the central region, this part of oxygen signal arises from the shell instead of the NP core as otherwise a higher ratio of oxygen would be expected if the NP core would be oxide.
Figure 3.2 HRTEM images of Fe‐based core‐shell nanoparticles with FFTs as inset. The insets in (a) and (b) with lattice spaces shown in Å, indicate the α‐Fe and FeO, respectively. The shell is proved to be Fe3O4 or γ‐Fe2O3 according to the diffraction pattern shown in the
inset of (c) and being consistent with the lattice spacing measured in the HRTEM image (the inset in (c) shows for clarity of the NP shape with a core‐shell structure). The SAED pattern in (d) agrees with the above results. The labels in (d) indicate the dhkl interplanar
Figure 3.3 HAADF image of a single particle with clear core‐shell contrast (a), O‐K map (b), and Fe‐L2,3 map (c). Although oxygen was detected in the central region, it originates most
likely from the oxide shell around the NP core.
Figure 3.4 shows EEL spectra collected on the center and shell part of the investigated NP with the background subtracted, which allow the relative quantification of the Fe to O ratio. Comparing previous reports 53,54 on the O-K and
Fe-L2,3 signal in the EEL spectra (ELNES) with our results, we did not get any
conclusive results most likely because of short acquisition times (required because of stability reasons of the Fe-O nanoparticle) and therefore a reduced signal to noise ratio. In addition, a possible adsorbed layer containing O2 and/or H2O would further
change the fine structure of the O-K peak, which hinders a clear interpretation. Instead we used another approach, where we quantified the O/Fe ratio considering some geometrical issues of the nanoparticle and its surrounding (see Appendix). Assuming the core and shell to be of a cube shape, and the carbon substrate and the layer on it to be homogenous (see inset Figure 3.4, Figure 3.10 and Figure 3.11 in Appendix), calculations (see Equation (3.1) to (3.5) in Appendix) gave ~98-99% Fe and ~1-2% O for the core. The calculated results, though of qualitative nature, support further the fact that the NP core is pure Fe instead of FeO (the Fe to O ratio is much closer to 100/0 than to 50/50).
Figure 3.4 Comparison of the EEL spectra collected on the center and shell area respectively with the background subtracted from the original. The inset shows the investigated NP, and 5 different areas corresponding to the central region (1) and shell area (2 – 5). While the spectrum displayed in red originates from region 1, the blue spectrum is summed up over the shell areas 2 ‐ 5, giving 2 distinct spectra indicating different ratios of Fe to O.
3.3.3 NP Crystal Morphology
Since we already proved, using diffraction patterns in combination with the elemental mapping, that the as-prepared Fe NPs have a bcc-Fe single nanocrystalline core, now the NP crystal shape will be further investigated. For this purpose Figure 3.5 shows an overview with a low magnification TEM image to illustrate the different sizes and morphologies of the core-shell Fe NPs. The overview image shows that the isolated NPs have sizes ranging from 5~100 nm. NPs larger than 15 nm showed a well-defined morphology, and are faceted along certain crystallographic lattice planes. Indeed, when the particles grow large, they
tend to adopt a morphology for which the external surfaces will be dominated by a group of specific lattice planes. Normally, this group of faceting planes corresponds to the planes with a lowest surface energy. For bcc structured α-Fe the {110} plane is the closest packed, and thus the lowest surface energy plane. Besides energy minimization, also kinetics can significantly affect the NP shape.
Figure 3.5 Low magnification overview image of Fe NPs shows distinct size‐dependent particle shapes including rhombic dodecahedron, truncated cube and truncated rhombic dodecahedron.
Furthermore, for Fe with a bcc structure distinct crystal morphologies can be formed depending on the ratio R of the growth speed along the <100> and <110> directions. Figure 3.6 gives the possible shapes of Fe nanocrystals extracted by the Wulff shape construction 41 as a function of R, ranging from cube bounded by the
6{100} planes (R=0.71) to rhombic dodecahedron bounded by {110} planes (R=1.41). Truncated cubic and rhombic dodecahedral NPs can also be obtained with different ratios R in the range 0.7<R<1.41. The NPs having the shape of a truncated rhombic dodecahedron, e.g. for R=1.06, are enclosed by 12 {110} planes truncated by 6 {100} planes. On the other hand cubic particles without any
0.2 µm
0.2 µm
rhombic dodecahedron truncated cube truncated rhombic dodecahedrontruncation are confined only by 6 {100} planes, while truncation leads to additional {110} facets. Figure 3.6 Geometrical shapes of Fe NPs as a function of the ratio R of the growth rate along the <100> and the <110> directions. Experimental TEM images are also given for each type of morphology, and the corresponding projection and truncation degree are also indicated. (C: Cube, T: truncated, RD: Rhombic Dodecahedron
Based on a large number of area surveys performed here, Fe core-shell NPs produced by the high pressure magnetron sputtering and gas condensation at RT show distinct shapes depending on the particle size. It was found that the truncated cube, faceted by the 6{100} planes with different degrees of truncation by the 12{110} planes, is predominant for particles with a size between 15-24 nm. For particles larger than 24 nm the most stable structure is the rhombic dodecahedron with 12 pseudoclose-packed {110} faces 29 (as it is shown in Figure 3.7), which,
however, was not observed in previous studies 28. Truncated rhombic dodecahedral
as well as cubic particles with and without truncations could also be observed in large NPs.
Figure 3.7 (a) HAADF image of a core‐shell particle corresponding to the [110] projection of rhombic dodecahedron. (b‐d) Reconstructed 3D structure from the single projection shown in (a) illustrating the concave structure of the {110} planes. The core is displayed in red, the shell in blue. Reconstruction is displayed (b) in a <110> direction, (c) in a <100> direction (d) in a <111> direction.
The HAADF image of a particle in Figure 3.7(a) agrees well with the [110] projection of rhombic dodecahedron. This is also proved by the 3D core-shell structure shown in Figure 3.7(b) to (d) reconstructed from a single HAADF STEM projection (see Part. A in Appendix). This reconstruction is possible under the assumption that the same projection can be observed in all <110> directions and that a sharp interface exist between core and shell. A clear inward relaxation could be noticed on the {110} planes (see arrows in Figure 3.7(a)) enclosing the rhombic
dodecahedron. Moreover, a concave surface structure could also be observed on the {100} planes in cubic NPs as well.
In general, both thermodynamics and kinetics contribute to the control of the morphology of a crystal. Thermodynamics can generate nanocrystals enclosed by a convex surface and low energy facets, including the aforementioned truncated and untruncated cubes and rhombic dodecahedrons, following the requirement to minimize the total surface energy of a system. In contrast, kinetic factors shapes the crystal morphology by manipulating the growth rate at which atoms are generated and added to the surface of a growing seed, and it is not limited by thermodynamics. Crystals enclosed by high-energy facets and/or with a concave structure on the surface can be obtained by kinetic approaches 55. It can be
concluded from the discussion by Jin et al. 56 that the diffusion rate of Fe clusters
from the atmosphere (within the aggregation volume during sputtering) to the surface of Fe seeds is much smaller than the growth rate at which clusters are generated and added to a growing seed when concave cubic Fe NPs are formed (Figure 3.7). In this condition, there will be a concentration gradient of Fe clusters, which shows a decrease from the atmosphere to the surface of the Fe seed. As a result Fe clusters will be preferentially added to the site with the highest reactivity due to an insufficient supply of clusters. In the case of cubic Fe precursor, the reactivity of different sites is supposed to decrease in the order of corner, edge and side face. Therefore, the preferential overgrowth at corners and edges of a cubic cluster is enhanced by kinetics, and results in the aforementioned concave cubic faceted Fe NPs.
3.4 Conclusion
Core-shell structured Fe NPs were produced by magnetron sputtering and gas condensation, obtaining particles constituted by a single crystal Fe core, as our extensive analysis indicated, and a polycrystalline shell (Fe3O4 and/or γ-Fe2O3).
The NPs revealed different morphologies depending on their size. For NPs with a size smaller than 24 nm, cubic particles are predominant, which are formed and confined by the 6 {100} planes but can be truncated by the 12{110} facets. The evolution of the NP shape is somewhat more complicated for particles larger than 24 nm. The most stable structure is the rhombic dodecahedron enclosed with 12 {110} faces, while truncated rhombic dodecahedral and cubic particles with no truncation could also be observed. The facets of NPs without truncation showed a characteristic inward relaxation indicating that besides thermodynamics also kinetics play a crucial role during particle growth. In general, the current results
indicate that specific particles with different size and shape can be produced by altering the synthesis parameters during gas deposition, which provides a feasible method to tailor particles with distinct properties, especially for magnetic applications. Magnetization studies are in progress in this direction via a superconducting quantum interference device (SQUID) and magnetic force microscopy (MFM) to elucidate the structure-property relation of the produced assemblies of Fe/Fe-oxide shell NPs.
Appendix
A. Tomographic reconstruction from a single HAADF STEM
projection
Reconstruction of the 3D morphology of a rhomic docecahedral particle is done from a single HAADF STEM projection taken along a <110> direction using prior knowledge about the particle symmetry. The projection is centered using the center-of-mass of the image and rotated to orient the symmetry axes (Sx, Sy) along the x- and y-direction (see Figure 3.8(a)). Then the projection is mirrored around its symmetry axes and a mean of the four mirror images is calculated (see Figure 3.8(b)). This averaged projection is now used as input for a tomographic reconstruction. The same projection is used 6 times as projection along all 110 directions of the particle (see Figure 3.8(b)). Tomographic reconstruction is done using a total-variation (TV) minimization algorithm49. This type of algorithm has
proven efficient for reconstruction from a few projections, if sharp interfaces between different materials can be expected, as in the present case. Segmentation of the reconstructed volume is done using absolute threshold values to define the interfaces between core and shell, and the exterior surface of the particle as well.
Figure 3.8 Principle of tomographic reconstruction from a single projection. (a) A projection along the <110> direction is centered and oriented along the symmetry axes Sx and Sy. (b) Mirror images around the symmetry axes are averaged and used as input for a tomographic reconstruction. The arrows indicate the additional <110> directions along which the same projection can be observed.
B. EEL spectra of Fe NP
Figure 3.9 Comparison of the EELS spectra collected on the shell (a) and center (b) area, showing the Fe, O and the background models.
The quantification of Fe and O was realized using a model based fitting approach57.
The background was fitted by a power-law function, while the ionization edges for O and Fe were described using Hartree-Slater cross sections. In addition, plural scattering which affects the shape of the ionization edges, was also taken into account by a convolution of a single-scattering core-loss model with the experimental low-loss spectrum, including the elastically scattered electrons
(Zero-Loss peak). Finally the low-loss spectrum was used to model the electron loss near edge structure (ELNES) by an iterative refinement of ELNES intensities.
Table 3.1 Elemental content obtained from the EELS spectra.
Element Comp. (at.%) Areal Dens.
(at./nm2)
Vol. Dens. (at./nm3)
O (shell) 74.5 2131.8 ± 106.7 42.9 ± 2.1 Fe (shell) 25.5 728.0 ± 72.8 14.6 ± 1.5 O (center part) 33.0 1297.8 ± 65.1 27.4 ± 1.4 Fe (center part) 67.0 2630.1 ± 263.0 55.5 ± 5.5
C. Geometric calculation of core Fe atom percentage
Figure 3.10 High Angle Annular Dark Field (HAADF) image of particle for EELS investigation with marked areas corresponding to the EELS spectra. Each area is a projection composed by the shell, the core, the carbon substrate on the TEM grid, and the possible layer containing water and/or oxygen adsorbed on the surface of the particle and carbon substrate.
Figure 3.11 Geometric model of the core‐shell structured nanoparticle investigated by EELS (assuming a homogenous thickness of the substrate and a homogenous oxygen content for the whole substrate). Four distinct sections contribute to the EELS spectrum: core, shell, carbon substrate on the TEM grid, and the possible adsorbed layer containing O2 and/or H2O on the surface of particle and carbon substrate.
For both the shell and the center part of the nanoparticle, a projection with an area S is selected to perform the calculation.
NO,layer - the areal density of O atoms in the projection contributed by the adsorbed
layer and the carbon substrate ( atoms/nm2, assuming a homogenous thickness of
the substrat.)
Nshell - the volume density of atoms (both O and Fe) in the shell (atoms/nm3);
NFe,core - the volume density of Fe atoms in the core (atoms/nm3);
NO,core - the volume density of O atoms in the core (atoms/nm3);
ωO,shell - the atom percentage of O in the shell;
ωFe,shell - the atom percentage of Fe in the shell (ωFe,shell=1- ωO,shell);
ωO,core - the atom percentage of O in the core;
ωFe,core - the atom percentage of Fe in the core;
tshell - the thickness of the shell, tshell=4.1±0.2 nm;
tcore - the lateral length of the core, tcore=22.6±0.8 nm;
tcarbon - the thickness of the carbon substrate, tcarbon=18.3±0.8 nm;
The shell is supposed to be γ-Fe2O3 or Fe3O4 or a combination of both, so ωO,shell
should lie somewhere between 57.1% and 60%. Here, we perform quantification assuming both cases of γ-Fe2O3 and Fe3O4, respectively.
carbon layer
30.8±0.4 nm
18.3±0.8 nm
4.1±0.2 nm
center part shell
carbon layer
center part shell Absorbed layer on
the surface of carbon and particle. O atoms in the absor-bed layer.
O atoms in the shell. Fe atoms in the shell. Fe atoms in the core.
The estimated thickness of the projection (tprojection=49.1±1.2 nm), including
the NP and the carbon layer, was obtained by EEL spectra analysis. tshell and tcore
are measured by Digital Micrograph. tcarbon= tprojection -2 tshell - tcore.
The adsorbed layer on the surface of the NP and the carbon substrate is assumed to be a multilayer58,59. The thickness of the adsorbed layer can be ignored
in this calculation, and the oxygen content for the adsorbed layer is assumed to be uniform throughout the layer.
The volume density of each element can be described as follows, containing contributions from distinct sections:
i i i N t S t S
(3.1)(Note: the contribution from the adsorbed layer is NO,layer S. )
Then the volume density of Fe and O in the shell can be calculated by the following equations:
Fe:
,
2
14.6 1.5
2
Fe shell shell shell core shell core carbon
N
t
t
S
t
t
t
S
(3.2) O:
,2
,
42.9 2.1
2
O shell shell shell core shell core carbon
N
t
t
S NO layer S
t
t
t
S
(3.3)The volume density of Fe and O in the center area can be calculated by the following equations:
Fe:
, ,
2
55.5 5.5
2
Fe core core Fe shell shell shell shell core carbon
N
t
S
N
t
S
t
t
t
S
(3.4)O:
,
2
, ,27.4 1.4
2
O shell shell shell O core core O layer shell core carbon
N
t
S N
t
S N
S
t
t
t
S
(3.5)Assuming the shell to be γ-Fe2O3 ( ωO,shell=60%), the results obtained from
Equation 3.2 to 3.5 are as follows:
Nshell=59±6 atoms/nm3, NO,layer=1016±103 atoms/nm2,
NFe,core=112±12 atoms/nm3, No,core=2±0.2 atoms/nm3 .
Therefore, we obtain the atom percentages of O and Fe in the core ωO,core=2±0.2%
and ωFe,core=98±0.2%, respectively.
Assuming the shell to be Fe3O4 (ωO,shell=57.1%), the results obtained from
Equation 3.2 to 3.5 are as follows:
Nshell=55±6 atoms/nm3, NO,layer=1071±101 atoms/nm2,
NFe,core=112±12 atoms/nm3, No,core=1±0.1 atoms/nm3.
Therefore, we obtain the atom percentages of O and Fe in the core ωO,core=1±0.2%
and ωFe,core=99±0.2%, respectively.
The above results support the conclusion that the core is pure Fe instead of FeO.
References
1. Signorini, L. et al. Size-dependent oxidation in iron/iron oxide core-shell nanoparticles. Phys. Rev. B 68, 195423 (2003).
2. Park, J.-I. & Cheon, J. Synthesis of “solid solution” and “core-shell” type cobalt-platinum magnetic nanoparticles via transmetalation reactions. J. Am. Chem. Soc. 123, 5743-5746 (2001).
3. Nogués, J., Skumryev, V., Sort, J., Stoyanov, S. & Givord, D. Shell-driven magnetic stability in core-shell nanoparticles. Phys. Rev. Lett. 97, 157203 (2006).
4. Liu, X. G., Geng, D. Y., Meng, H., Shang, P. J. & Zhang, Z. D. Microwave-absorption properties of ZnO-coated iron nanocapsules. Appl. Phys. Lett. 92, 173117 (2008).
5. Xiao, H.-M., Liu, X.-M. & Fu, S.-Y. Synthesis, magnetic and microwave absorbing properties of core-shell structured MnFe2O4/TiO2 nanocomposites.
Compos. Sci. Technol. 66, 2003-2008 (2006).
6. Lu, B. et al. Microwave absorption properties of the core/shell-type iron and nickel nanoparticles. J. Magn. Magn. Mater. 320, 1106-1111 (2008).
7. Martin, J. E. et al. Determination of the oxide layer thickness in core-shell zerovalent iron nanoparticles. Langmuir 24, 4329-4334 (2008).
8. Kim, J. et al. Multifunctional uniform nanoparticles composed of a magnetite nanocrystal core and a mesoporous silica shell for magnetic resonance and fluorescence imaging and for drug delivery. Angew. Chem. Int. Ed. 47, 8438-8441 (2008).
9. Xu, Z., Hou, Y. & Sun, S. Magnetic core/shell Fe3O4/Au and Fe3O4/Au/Ag
nanoparticles with tunable plasmonic properties. J. Am. Chem. Soc. 129, 8698-8699 (2007).
10. Lyon, J. L., Fleming, D. A., Stone, M. B., Schiffer, P. & Williams, M. E. Synthesis of Fe oxide core/Au shell nanoparticles by iterative hydroxylamine seeding. Nano Lett. 4, 719-723 (2004).
11. Dobson, J. Magnetic nanoparticles for drug delivery. Drug Dev. Res. 67, 55-60 (2006).
12. Cheong, S. et al. Simple synthesis and functionalization of iron nanoparticles for magnetic resonance imaging. Angew. Chem. 123, 4292-4295 (2011). 13. WangWang et al. Monodispersed core-shell Fe3O4@Au nanoparticles. J. Phys.
Chem. B 109, 21593-21601 (2005).
14. Binns, C. et al. Preparation of hydrosol suspensions of elemental and core-shell nanoparticles by co-deposition with water vapour from the gas-phase in ultra-high vacuum conditions. J. Nanoparticle Res. 14, 1-16 (2012).
15. Goetze, T. et al. Biocompatible magnetic core/shell nanoparticles. J. Magn. Magn. Mater. 252, 399-402 (2002).
16. Zhang, J. & Misra, R. D. K. Magnetic drug-targeting carrier encapsulated with thermosensitive smart polymer: Core-shell nanoparticle carrier and drug release response. Acta Biomater. 3, 838-850 (2007).
17. Herman, D. A. J. et al. Hot-injection synthesis of iron/iron oxide core/shell nanoparticles for T2 contrast enhancement in magnetic resonance imaging. Chem. Commun. 47, 9221 (2011).
18. Pratt, A. et al. Enhanced oxidation of nanoparticles through strain-mediated ionic transport. Nat. Mater. 13, 26-30 (2014).
19. Kuhn, L. T., Bojesen, A., Timmermann, L., Nielsen, M. M. & Mørup, S. Structural and magnetic properties of core-shell iron-iron oxide nanoparticles. J. Phys. Condens. Matter 14, 13551 (2002).
20. Huber, D. L. Synthesis, properties, and applications of iron nanoparticles. Small 1, 482-501 (2005).
21. Hadjipanayis, C. G. et al. Metallic iron nanoparticles for MRI contrast enhancement and local hyperthermia. Small 4, 1925-1929 (2008).
22. Bomatí-Miguel, O. et al. Fe-based nanoparticulate metallic alloys as contrast agents for magnetic resonance imaging. Biomaterials 26, 5695-5703 (2005). 23. Baker, I., Zeng, Q., Li, W. & Sullivan, C. R. Heat deposition in iron oxide and
iron nanoparticles for localized hyperthermia. J. Appl. Phys. 99, 08H106 (2006).
24. Taylor, A. et al. Cisplatin-loaded carbon-encapsulated iron nanoparticles and their in vitro effects in magnetic fluid hyperthermia. Carbon 48, 2327-2334 (2010).
25. Zhang, G., Liao, Y. & Baker, I. Surface engineering of core/shell iron/iron oxide nanoparticles from microemulsions for hyperthermia. Mater. Sci. Eng. C 30, 92-97 (2010).
26. Balivada, S. et al. A/C magnetic hyperthermia of melanoma mediated by iron(0)/iron oxide core/shell magnetic nanoparticles: a mouse study. BMC Cancer 10, 119 (2010).
27. Wang, C. M. et al. Void formation during early stages of passivation: Initial oxidation of iron nanoparticles at room temperature. J. Appl. Phys. 98, 094308 (2005).
28. Wang, C. M. et al. Morphology and oxide shell structure of iron nanoparticles grown by sputter-gas-aggregation. Nanotechnology 18, 255603 (2007).
29. Vystavel, T., Palasantzas, G., Koch, S. A. & Hosson, J. T. M. D. Nanosized iron clusters investigated with in situ transmission electron microscopy. Appl. Phys. Lett. 82, 197-199 (2003).
30. Gong, W., Li, H., Zhao, Z. & Chen, J. Ultrafine particles of Fe, Co, and Ni ferromagnetic metals. J. Appl. Phys. 69, 5119-5121 (1991).
31. Peng, S., Wang, C., Xie, J. & Sun, S. Synthesis and stabilization of monodisperse Fe nanoparticles. J. Am. Chem. Soc. 128, 10676-10677 (2006). 32. de Montferrand, C. et al. Iron oxide nanoparticles with sizes, shapes and
compositions resulting in different magnetization signatures as potential labels for multiparametric detection. Acta Biomater. 9, 6150-6157 (2013).
33. Steigerwald, M. L. & Brus, L. E. Semiconductor crystallites: a class of large molecules. Acc. Chem. Res. 23, 183-188 (1990).
34. Chithrani, B. D., Ghazani, A. A. & Chan, W. C. W. Determining the size and shape dependence of gold nanoparticle uptake into mammalian cells. Nano Lett.
6, 662-668 (2006).
35. Zhang, J. Z. Ultrafast studies of electron dynamics in semiconductor and metal colloidal nanoparticles: effects of size and surface. Acc. Chem. Res. 30, 423-429(1997).
36. Baletto, F. & Ferrando, R. Structural properties of nanoclusters: Energetic, thermodynamic, and kinetic effects. Rev. Mod. Phys. 77, 371-423 (2005). 37. Zhang, L. & Webster, T. J. Nanotechnology and nanomaterials: Promises for
improved tissue regeneration. Nano Today 4, 66-80 (2009).
38. Davis, M. E., Chen, Z. (Georgia) & Shin, D. M. Nanoparticle therapeutics: an emerging treatment modality for cancer. Nat. Rev. Drug Discov. 7, 771-782 (2008).
39. Freund, H.-J. Clusters and islands on oxides: from catalysis via electronics and magnetism to optics. Surf. Sci. 500, 271-299 (2002).
40. Spencer, M. J. S., Hung, A., Snook, I. K. & Yarovsky, I. Density functional theory study of the relaxation and energy of iron surfaces. Surf. Sci. 513, 389-398 (2002).
41. Wang, Z. L. Transmission electron microscopy of shape-controlled nanocrystals and their assemblies. J. Phys. Chem. B 104, 1153-1175 (2000). 42. Saito, Y., Mihama, K. & Uyeda, R. Formation of ultrafine metal particles by
gas-evaporation VI. bcc metals, Fe, V, Nb, Ta, Cr, Mo and W. Jpn. J. Appl. Phys. 19, 1603 (1980).
43. Hayashi, T., Ohno, T., Yatsuya, S. & Uyeda, R. Formation of ultrafine metal particles by gas-evaporation technique. IV. crystal habits of iron and fcc metals, Al, Co, Ni, Cu, Pd, Ag, In, Au and Pb. Jpn. J. Appl. Phys. 16, 705 (1977).
44. Brink, G. H. ten, Krishnan, G., Kooi, B. J. & Palasantzas, G. Copper nanoparticle formation in a reducing gas environment. J. Appl. Phys. 116, 104302 (2014).
45. Gracia-Pinilla, M. Á., Ferrer, D., Mejía-Rosales, S. & Pérez-Tijerina, E. Size-selected Ag nanoparticles with five-fold symmetry. Nanoscale Res. Lett. 4, 896-902 (2009).
46. Hofer, F., Grogger, W., Kothleitner, G. & Warbichler, P. Quantitative analysis of EFTEM elemental distribution images. Ultramicroscopy 67, 83-103 (1997). 47. Krivanek, O. L., Kundmann, M. K. & Kimoto, K. Spatial resolution in EFTEM
elemental maps. J. Microsc. 180, 277-287 (1995).
48. Midgley, P. A. Electron tomography and holography in materials science. Nature Mater. 8, 271-280 (2009).
49. Haberfehlner, G., Orthacker, A., Albu, M., Li, J. & Kothleitner, G. Nanoscale voxel spectroscopy by simultaneous EELS and EDS tomography. Nanoscale
6, 14563-14569 (2014).
50. Horn, I. & Günther, D. The influence of ablation carrier gasses Ar, He and Ne on the particle size distribution and transport efficiencies of laser ablation-induced aerosols: implications for LA–ICP–MS. Appl. Surf. Sci. 207, 144-157 (2003).
51. Shavel, A., Rodríguez-González, B., Spasova, M., Farle, M. & Liz-Marzán, L. M. Synthesis and characterization of iron/iron oxide core/shell nanocubes. Adv. Funct. Mater. 17, 3870-3876 (2007).
52. Wang, C. et al. Morphology and electronic structure of the oxide shell on the surface of iron nanoparticles. J. Am. Chem. Soc. 131, 8824-8832 (2009). 53. Krivanek, O. L. & Paterson, J. H. ELNES of 3d transition-metal oxides: I.
Variations across the periodic table. Ultramicroscopy 32, 313-318 (1990). 54. Colliex, C., Manoubi, T. & Ortiz, C. Electron-energy-loss-spectroscopy
near-edge fine structures in the iron-oxygen system. Phys. Rev. B 44, 11402-11411 (1991).
55. Zhang, H. et al. Controlling the morphology of rhodium nanocrystals by manipulating the growth kinetics with a syringe pump. Nano Lett. 11, 898-903 (2011).
56. Jin, M., Zhang, H., Xie, Z. & Xia, Y. Palladium concave nanocubes with high-index facets and their enhanced catalytic properties. Angew. Chem. Int. Ed. 50, 7850-7854 (2011).
57. Thomas, P. J. & Twesten, R. d. A simple, model based approach for robust quantification of EELS spectra and spectrum-images. Microsc. Microanal. 18, 968-969 (2012).
58. Palasantzas, G., Svetovoy, V. B. & van Zwol, P. J. Influence of ultrathin water layer on the van der Waals/Casimir force between gold surfaces. Phys. Rev. B
79, 235434 (2009).
59. van Zwol, P. J., Palasantzas, G. & De Hosson, J. T. M. Influence of roughness on capillary forces between hydrophilic surfaces. Phys. Rev. E 78, 031606 (2008).
