Magnetic Fe-N Core-Shell Nanoparticles
Nanoscale Analytics in the Transmission Electron Microscope
- Primitive unit cells of the 4 phases used for ab initio simulations. The plots were generated using XCrysDen  but its main function is as a property analyzer program. It can run on most UNIX platforms, without any special hardware or software requirements. Special efforts were made to allow for appropriate display of 3D isosurfaces and 2D contours, which can be superimposed on crystalline structure and interactively rotated and manipulated. XCrySDen is also a graphical user interface for the CRYSTAL95/98 (Saunders, V. R., et al. CRYSTAL98—User’s Manual. University of Torino, Turin, Italy, 1999.
- Fig. 1: TEM bright-field images of (a) ε-Fe3N and (b) α’’-Fe16N2. (c) and (d) are the corresponding selected area diffraction pattern.
- Fig. 2: High-angle annular dark-field STEM images of (a) ε-Fe3N and (b) α’’-Fe16N2. (c) and (d) are the corresponding electron energy-loss spectral imaging maps (red: N-K, green: Fe-L2,3, blue: O-K).
- Fig. 3: Electron energy-loss spectra of the core of (a) ε-Fe3N and (b) α’’-Fe16N2. (c) and (d) are the corresponding shell spectra for FeO and Fe3O4, respectively. The experimental spectra are in black, the simulated spectra using the Wien2k code are in red and the FeFF are in green.
- Tab. S1: Wien2k input parameters.
- Tab. S2: FeFF input parameters.
- Tab. S3: Experimental parameters used for both types of ELNES simulations.
In this structure-property investigation we analyzed α’’-Fe16N2 and ε-Fe3N core-shell nanoparticles via analytical electron micrsocopy. Conventional electron microscopy on pristine nanoparticles was done using a vacuum transfer holder. In order to study the effect of oxidation on the nanoparticles a combined experimental and density functional theory (DFT) study of the electron near edge fine-structure (ELNES) was carried out.
Alternative magnetic materials have regained attention after the recent rare-earth crisis [1, 2]. Iron nitrides exhibit remarkable physical properties [3-6], which can be of use for several applications, e.g. magnetic recording media [6, 7], catalysis , and biomedical applications . These applications demand precise control over the particle size and corresponding magnetic properties on the nanometer scale. Iron nitride nanoparticles are inexpensive, exhibit high saturation magnetizations and large magnetic moments [5, 7]. A main difficulty in the measurement of their magnetic properties is to prepare pure phases in sufficiently large quantities. A recent review discusses the synthesis and properties of iron nitrides at reduced dimensions . Α’’-Fe16N2, ε-Fe3N and γ’-Fe4N are important ferromagnetic iron nitrides. Analytical electron microscopy in combination with ab nitio density functional simulations is of importance for establishing structure-property correlations in these materials.
Material & Methods
Sample preparation and electron microscopic acquisition conditions can be found in the literature [11, 12]. Ab-initio DFT calculations were carried out using the Wien2k  and the FeFF  codes. Simulation details are summarized in the supporting information.
Conventional transmission electron microscopy applying diffraction contrast imaging and selected area diffraction was carried out using a vacuum transfer holder to prevent oxidation and to study the pristine material. The results for ε-Fe3N and α’’-Fe16N2 are shown in figure 1 a,b and figure 1 b,c, respectively.
The selected area diffraction patterns were indexed using the structural parameters from Jacobs et al.  and Widenmeyer et al.  for ε-Fe3N and α’’-Fe16N2, respectively. It can be seen that both types of nanoparticles agglomerate hampering further atomic resolution image analysis. Nevertheless, the nanoparticle size can be determined from these images as 15.4 ± 4.1 nm for ε-Fe3N and 50 nm for α’’-Fe16N2. This is in good agreement with X-ray measurements. The different values of the particle size are due to the different synthesis method being used.
For any practical application the nanoparticles need to be exposed to ambient air. The result was a partial oxidation, which was then studied by analytical scanning transmission electron microscopy (STEM), since imaging techniques only indicate a shell structure. Analyses using energy-dispersive X-ray spectroscopy (EDX) are not a good option, because of technical and practical limitations such as low energy X-ray absorption and peak overlapping. A better alternative is to use high-angle annular dark-field images (fig. 2 a,b) in combination with electron energy-loss spectroscopy (EELS) maps (fig. 2 c,d). The shell thickness was determined to be 2.5 ± 1.0 nm for ε-Fe3N and 7.0 ± 2.0 nm for α’’-Fe16N2, respectively. The colored maps were combined by three elemental images using the electron energy-loss ranges 400-420 eV for N-K, 530-550 eV for O-K, and 705-730 eV for Fe-L2,3 as red, blue and green channel, respectively. This yields a yellow to green core and a blueish shell that can be well seperated from each other. In addition, EELS offers the possibility to simultaneously probe the local electronic structure and the phase via the ELNES. For example iron oxides can exist as FeO (wuestite), α-Fe2O3 (hematite) and Fe3O4 (magnetite). Each of them presents different magnetic properties that interacts differently with the ferromagnetic nitride core. FeO is at room-temperature and ambient pressure paramagnetic , α-Fe2O3 is antiferromagnetic , and Fe3O4 is ferromagnetic . It was not clear from the beginning of the analysis which iron oxide modification is present in the shell or if it is the same for all samples. Iron nitride phases can be treated in a similar way. The exact core-shell combination involves for example effects like spin exchange interactions, which themselves influence macroscopic magnetic measurements such as hysteresis loops .
Since only from the oxides experimental EELS reference spectra were available  and to check whether the description by bulk properties is still applicable, ab initio DFT calculations were carried out to simulate the ELNES of the N- and O-K edge of the previously mentioned phases, respectively. These simulations were then compared to our experiments. At present time there are two possibilities to do so: (i) bandstructure based calculations as implemented for example in the Wien2k code  and (ii) self-consistent real-space multiple-scattering calculations as implemented for example in the FeFF9 code . The Wien2k code is among the most accurate codes and offers a large number of features like bandstructure, density of states, X-ray emission and absorption spectra, and electron energy loss spectra. Depending on the number of atoms and the symmetry access to a supercomputer is necessary. The FeFF code on the other hand uses a completely different approach, which allows the simulation of doped structures or alloys, voids, interstitials etc. with much less computational effort than needed in Wien2k. A fast desktop computer is for most cases sufficient. Since FeFF has its origin in the calculation of X-ray spectra it offers despite ELNES also the possibility to calculate angular-momentum resolved density of states, X-ray emission spectra and X-ray absorption spectra. Core-holes are also easily introduced without the need to think in detail on supercell geometries like in Wien2k. The interested reader might find more details in [21-24].
Based on the previous considerations we were able to calculate the N- and O-K ELNES. The results of the comparison between experiment and simulation are shown in figure 3. Figure 3a shows the results for a ε-Fe3N core and figure 3b a FeO shell. Distinct peaks labelled p1-p4 can be identified in the spectra. For a better comparison the spectra were normalized. In case of the nitride core differences between experiment and simulation were found in the N-K ELNES indicating that the core is not pure ε-Fe3N anymore. Thus, the blue curve where the N atoms on the 2c position were replaced by O atoms and the 2b position is fully occupied by N atoms in figure 3a was calculated using FeFF. It is closer to the experiment than the pure ε-Fe3N curves. For the shell the case in figure 3b is more clear. When comparing the peaks p1-p4 of the FeO simulation to the experiment, one recognizes deviations of the peak form in p1 and the missing of p4. Furthermore, it might be possible that a second iron oxide phase, e.g. Fe3O4 (fig. 3d), is present in the shell. A more extended study on the oxidation behavior of ε-Fe3N would be of advantage. In the α’’-Fe16N2 case, which is shown in figure 3c,d, both Wien2k and FeFF yield a good agreement of simulation and experiment for both core and shell. All peaks (p1-p4) are reproduced by the simulation. The measurement as well as the simulations are in agreement with the O-K ELNES measured by Colliex et al. .
Summary & Conclusion
In summary, we were able to determine phase and particle sizes of iron nitride nanoparticles via conventional TEM. Furthermore, we were able to characterize partially oxidized core-shell nanoparticles via a combination of analytical TEM and DFT based ELNES simulations. For the α’’-Fe16N2 case a good agreement between experiment and simulation was found for both core and shell, whereas for the ε-Fe3N case only the shell fitted and the the experimental core spectrum deviated from the simulation indicating a structural and/or chemical change of the ε-Fe3N core.
The authors acknowledge financial support from the LOEWE project RESPONSE funded by the Ministry of Higher Education, Research and the Arts of the Hessen state.
Primitive unit cells of the 4 phases used for ab initio simulations. The plots were generated using XCrysDen  but its main function is as a property analyzer program. It can run on most UNIX platforms, without any special hardware or software requirements. Special efforts were made to allow for appropriate display of 3D isosurfaces and 2D contours, which can be superimposed on crystalline structure and interactively rotated and manipulated. XCrySDen is also a graphical user interface for the CRYSTAL95/98 (Saunders, V. R., et al. CRYSTAL98—User’s Manual. University of Torino, Turin, Italy, 1999.
1Technische Universität Darmstadt, Department of Materials and Earth Sciences, Electron Microscopy Center Darmstadt (EMC-DA), Darmstadt, Germany
Dr. Michael Dürrschnabel
Department of Materials and Earth Sciences
Technische Universität Darmstadt
 K. Bourzac, The Rare-Earth Crisis, MIT Technol Rev 114, 58 (2011)
 M. D. Kuz’min, K. P. Skokov, H. Jian, I. Radulov, O. Gutfleisch: Towards high-performance permanent magnets without rare earths, J. Phys. Condens. Matter 26, 064205 (2014) doi: 10.1088/0953-8984/26/6/064205
 T. Takahashi, J. Burghaus, D. Music, R. Dronskowski, J. M. Schneider: Elastic properties of γ′-Fe4N probed by nanoindentation and ab initio calculation, Acta Mater. 60, 2054 (2012) doi: 10.1016/j.actamat.2011.12.051
 J. M. D. Coey, P. A. I. Smith: Magnetic nitrides, J. Magn. Magn. Mater. 200, 405-424 (1999) doi: 10.1016/S0304-8853(99)00429-1
 T. K. Kim, M. Takahashi: New Magnetic Material Having Ultrahigh Magnetic Moment, Appl. Phys. Lett. 20, 492 (1972) doi: 10.1063/1.1654030
 S. Suzuki, H. Sakumoto, J. Minegishi, Y. Omote: Coercivity and unit particle size of metal pigment, IEEE Trans. Magn. 17, 3017 (1981) doi: 10.1109/TMAG.1981.1061756
 S. F. Matar, G. Demazeau, B. Siberchicot: Magnetic particles derived from iron nitride, IEEE Trans. Magn. 26, 60 (1990) doi: 10.1109/20.50490
 U. I. Kramm, J. Herranz, N. Larouche, T. M. Arruda, M. Lefèvre, F. Jaouen, P. Bogdanoff, S. Fiechter, I. Abs-Wurmbach, S. Mukerjee, J.-P. Dodelet: Structure of the catalytic sites in Fe/N/C-catalysts for O2-reduction in PEM fuel cells, Phys. Chem. Chem. Phys. 14, 11673 (2012) doi: 10.1039/c2cp41957b
 Y. Namiki, T. Fuchigami, N. Tada, R. Kawamura, S. Matsunuma, Y. Kitamoto, M. Nakagawa: Nanomedicine for cancer: lipid-based nanostructures for drug delivery and monitoring, Acc. Chem. Res. 44, 1080 (2011) 10.1021/ar200011r
 S. Bhattacharyya: Iron Nitride Family at Reduced Dimensions: A Review of Their Synthesis Protocols and Structural and Magnetic Properties, J. Phys. Chem. C 119, 1601 (2015) doi: 10.1021/jp510606z
 A.-M. Zieschang, J. D. Bocarsly, M. Dürrschnabel, L. Molina-Luna, H.-J. Kleebe, R. Seshadri, B. Albert: Nanoscale Iron Nitride, ε-Fe3N: Preparation from Liquid Ammonia and Magnetic Properties, Chem. Mater. 29, 621 (2017) doi: 10.1021/acs.chemmater.6b04088
 I. Dirba, C. A. Schwöbel, L. V. B. Diop, M. Duerrschnabel, L. Molina-Luna, K. Hofmann, P. Komissinskiy, H.-J. Kleebe, O. Gutfleisch: Synthesis, morphology, thermal stability and magnetic properties of α″-Fe16N2 nanoparticles obtained by hydrogen reduction of γ-Fe2O3 and subsequent nitrogenation, Acta Mater. 123, 214 (2017) doi: 10.1016/j.actamat.2016.10.061
 P. Blaha, K. Schwarz, G. K. H. Madsen, D. Kvasicka, J. Luitz: “WIEN2k, An Augmented Plane Wave Plus Local Orbitals Program for Calculating Crystal Properties” 2th Edition, Vienna University of Technology, Vienna, 2001
 J. J. Rehr, J. J. Kas, F. D. Vila, M. P. Prange, K. Jorissen: Parameter-free calculations of X-ray spectra with FEFF9, Phys. Chem. Chem. Phys. 12, 5503. (2010) doi: 10.1039/b926434e
 H. Jacobs, D. Rechenbach, U. Zachwieja: Structure determination of γ′-Fe4N and ϵ-Fe3N, J. Alloys Compd. 227, 10 (1995) doi: 10.1016/0925-8388(95)01610-4
 M. Widenmeyer, L. Shlyk, A. Senyshyn, R. Mönig, R. Niewa: Structural and Magnetic Characterization of Single-phase Sponge-like Bulk α′′-Fe16N2, Z. Für Anorg. Allg. Chem. 641 (2) 348 (2015) doi: 10.1002/zaac.201500013
 H. Jiang, R. I. Gomez-Abal, P. Rinke, M. Scheffler: First-principles modeling of localized d states with the GW@LDA+U approach, Phys. Rev. B 82, 045108 (2010) doi: 10.1103/PhysRevB.82.045108
 G. Rollmann, A. Rohrbach, P. Entel, J. Hafner: First-principles calculation of the structure and magnetic phases of hematite, Phys. Rev. B 69, 165107 (2004) doi: 10.1103/PhysRevB.69.165107
 J. Noh, O. I. Osman, S. G. Aziz, P. Winget, J.-L. Brédas: A density functional theory investigation of the electronic structure and spin moments of magnetite, Sci. Technol. Adv. Mater. 15, 044202 (2014) doi: 10.1088/1468-6996/15/4/044202
 C. Hébert: Practical aspects of running the WIEN2k code for electron spectroscopy, Micron 38, 12 (2007) doi: 10.1016/j.micron.2006.03.010
 M. S. Moreno, S. Lazar, H. W. Zandbergen, R. F. Egerton: Accuracy of the calculated unoccupied states in GaN phases as tested by high-resolution electron energy-loss spectroscopy, Phys. Rev. B 73, 073308 (2006) 10.1103/PhysRevB.73.073308
 M. S. Moreno, K. Jorissen, J. J. Rehr: Practical aspects of electron energy-loss spectroscopy (EELS) calculations using FEFF8, Micron 38, 1 (2007) 10.1016/j.micron.2006.03.011
 K. Jorissen, J. J. Rehr, J. Verbeeck: Multiple scattering calculations of relativistic electron energy loss spectra, Phys. Rev. B 81, 155108 (2010) doi: 10.1103/PhysRevB.81.155108
 C. Colliex, T. Manoubi, C. Ortiz: Electron-energy-loss-spectroscopy near-edge fine structures in the iron-oxygen system, Phys. Rev. B 44, 11402 (1991) 10.1103/PhysRevB.44.11402
 A. Kokalj: XCrySDen—a new program for displaying crystalline structures and electron densities, J. Mol. Graph. Model. 17, 176 (1999) 10.1016/S1093-3263(99)00028-5
 J. P. Perdew, Y. Wang: Accurate and simple analytic representation of the electron-gas correlation energy, Phys. Rev. B, 45, 13244 (1992) 10.1103/PhysRevB.45.13244
 J. P. Perdew, K. Burke, M. Ernzerhof: Generalized Gradient Approximation Made Simple, Phys. Rev. Lett. 77, 3865 (1996) 10.1103/PhysRevLett.77.3865