Popcorn-Shaped Fe x O (Wüstite) Nanoparticles from a Single-Source Precursor: Colloidal Synthesis and Magnetic Properties.

Colloidal nanoparticles (NPs) with myriads of compositions and morphologies have been synthesized and characterized in recent years. For wüstite Fe x O, however, obtaining phase-pure NPs with homogeneous morphologies have remained challenging. Herein, we report the colloidal synthesis of phase-pure Fe x O (x ≈ 0.94) popcorn-shaped NPs by decomposition of a single-source precursor, [Fe3(μ3-O)(CF3COO)(μ-CF3COO)6(H2O)2]·CF3COOH. The popcorn shape and multigrain structure had been reconstructed using high-angle annular dark-field scanning transmission electron micrograph (HAADF-STEM) tomography. This morphology offers a large surface area and internal channels and prevents further agglomeration and thermal tumbling of the subparticles. [Fe3(μ3-O)(CF3COO)(μ-CF3COO)6(H2O)2]·CF3COOH behaves as an antiferromagnetic triangle whose magnetic frustration is mitigated by the low symmetry of the complex. The popcorn-shaped Fe x O NPs show the typical wüstite antiferromagnetic transition at approximately 200 K, but behave very differently to their bulk counterpart below 200 K. The magnetization curves show a clear, unsymmetrical hysteresis, which arises from a combined effect of the superparamagnetic behavior and exchange bias.

Introduction

Broadening the toolbox for the rational synthesis of nanomaterials is currently one of the most important objectives in inorganic chemistry. Nanosized materials such as NPs, have shown properties (optical, magnetic, and electronic) and possess attributes (colloidal state, tunable surface chemistry, etc.) that can benefit fields including bioimaging,[1] radiation detection,[2] light emission,[3] electrochemical energy storage,[4] photovoltaics,[5] and many more. State-of-the-art synthetic methods set high standards in the uniformity of the size, shape, and composition of NPs in a colloidal ensemble.

Wüstite FeO (in general x = 0.84–0.95), an antiferromagnetic compound, crystallizes in a NaCl-type defective crystal structure. The stoichiometric composition (FeO) does not exist at ambient pressure and room temperature (RT). All obtained substoichiometric wüstite phases have a deficiency of iron, and hence the mixed valency (Fe2+/3+) is necessary to maintain charge balance. Fe3+ cations move to unoccupied tetrahedral sites in FeO, leaving behind vacancies on octahedral sites.[6] According to the phase diagram, bulk FeO disproportionates into Fe3O4 and Fe below 570 °C. However, this disproportionation reaction becomes immeasurably slow below 300 °C.[7] Thus, FeO exists as a metastable phase at room temperature.[6,8] Yin and ÒBrien[9] showed that colloidal wüstite NPs can be synthesized at 320 °C due to the already strongly reduced rate of redox disproportionation close to 300 °C. Thus, FeO is kinetically stable at RT in bulk and as NPs.[10] FeO NPs are expected to be faster oxidized in air as compared to the bulk due to their larger surface area. However, fast oxidation is generally not observed, and the oxidation rates strongly depend on the capping ligands and the particle sizes/morphology. Thus, oxidation can be slow and occurs mostly on the surface, leading to a ferromagnetic shell covering the antiferromagnetic wüstite core.[11−13] Previously, FeO NPs with shapes such as spheres, cubes, octapods, and truncated octahedra[9,14−18] were synthesized. Typically, single-source precursors such as iron(II) acetate,[9,14,18] or Fe(acac)2–3,[14,15] or long-chain carboxylates[16,17] are decomposed in high boiling solvents or by selective oxidation of Fe(CO)5 with pyridine N-oxide.[14] Another approach is to decompose Fe-oleates, leading to a wide variety of NP shapes[16,17] and allowing the synthesis of heterostructures such as wüstite/spinel[19] or wüstite/ferrite[20,21] core–shell NPs. However, a clean wüstite phase has scarcely been obtained at NP sizes below ∼20 nm because of the aforementioned surface oxidation of FeO and the limitations of these precursors in producing only Fe2+ species.

In this report, we present a synthesis of novel, popcorn-shaped FeO (x ≈ 0.94) NPs (Figure ), with high structural and size uniformity. The synthesis decomposes an iron(III) oxo-trifluoroacetate single-source precursor (denoted as “Fe3OTFA”) in squalane as an inert solvent together with trioctylphosphine (TOP) as a mild reductant and capping agent and oleic acid (OA) and hexadecylamine (HDA) as capping ligands. The composition of the FeO NPs with x ≈ 0.94 was elucidated by a Rietveld refinement of a powder XRD pattern and by elemental mapping on the atomic scale using analytical electron microscopy. We also detail the crystal structures of the main precursor and three additional iron trifluoroacetate complexes discovered in this work. Additionally, the magnetic properties of the synthesized FeO NPs and their precursor are presented and discussed.

Figure 1

Structural characterization of popcorn-shaped FeO NPs (x ≈ 0.94): (a, b) overview TEM images, (c) slice through the reconstructed HAADF-STEM tomogram, (d) reaction scheme for the NP synthesis, and (e) powder XRD (Cu Kα1, 1.540598 Å); inset: SAED spectrum of FeO NPs. (f) 3D tomography image reconstructed by tilt-series of HAADF-STEM images.

Experimental Section

Synthesis of “Fe3OTFA” Precursor

Ten grams (62 mmol) of FeCl3 (Alfa Aeser, 98%) were mixed with trifluoroacetic acid (100 mL, 1.3 mol, TFA, Fischer) in a 250 mL two-neck round-bottom flask under N2 flow and degassed three times by applying vacuum. Subsequently, the reaction mixture was heated to 86 °C and left stirring for 2 h until a homogeneous yellow suspension was obtained. Then 1 mL of deionized water was injected, quickly turning the color of suspension into a dark red. The solution was left stirring for additional 24 h at 86 °C. After cooling, TFA excess was distillated off under vacuum. The remaining solid product, “Fe3OTFA”, was dried at 60 °C under vacuum for at least 5 h, yielding a fine red powder, which is directly useable without further workup. Despite “Fe3OTFA” appearing stable under ambient atmosphere, it was stored under vacuum until further use.

Synthesis of Popcorn-Shaped FeO (x ≈ 0.94) NPs

In a typical synthesis, “Fe3OTFA” (337 mg, 0.3 mmol) was mixed with HDA (1.304 g, 5.4 mmol, Aldrich, 90%), squalene (5 mL, Merck, ≥ 99%), OA (3.6 mmol, 1.143 mL, Aldrich, 90%), TOP (5 mL, Strem, 15–6655, > 97%) in a 25 mL three-neck round-bottom flask, forming a dark solution. This mixture was degassed by alternating between vacuum and N2 flow for 3 times at room temperature, followed by heating under vacuum to 110 °C and left drying for another 1.5 h. Afterward, the mixture was heated (19 °C/min) to 260 °C under N2. At ca. 210 °C, the solution changed color from black to a clear, light yellow. Subsequent NC nucleation and growth processes were visible as a color change from light yellow via brown to black at 260 °C. After 20 min at 260 °C, the solution was rapidly cooled to 110 °C and 20 mL of toluene were added, followed by complete cooling. For purifying the particles, the reaction mixture was mixed with 30 mL of ethanol as a nonsolvent and centrifuged (8000 rpm, 3 min), the supernatant was discarded and the precipitate was dispersed in 10 mL of toluene.

Materials Characterization

Low and high resolution transmission electron microscopy images (LRTEM/HRTEM) as well as selected area electron diffraction (SAED) patterns were obtained with a JEOL JEM-2200FS microscope operating at 200 kV. NCs were deposited onto a carbon-coated Cu grids (Ted-Pella). A tomographic tilt-series from −76 to 72° (2° steps) were recorded on an STEM-aberration-corrected FEI Titan Themis operated at 300 kV with a probe semiconvergence angle of 18 mrad (beam current 70 pA). The limits of the angular tilt-range were given by shadowing of the Cu grid. The tomogram was reconstructed based on high-angle annular dark-field scanning transmission electron micrographs using FEI’s Inspect3D, and visualized using Amira software. For element mapping using high-angle annular dark-field scanning transmission (HAADF-STEM) electron microscopy, combined with energy dispersive X-ray spectroscopy (EDX), a beam current of 6 nA (no difference with 1 or 3 nA) and detection with a SuperEDX system (4 detectors) were used. Powder X-ray diffraction (XRD) patterns were obtained on a Stoe STADI P powder X-ray diffractometer (Cu Kα1 radiation, λ = 1.540598 Å, germanium monochromator). The powders were measured on a transmission sample holder or in a capillary (0.5 mm, Hilgenberg). For single-crystal measurements, a suitable crystal was selected and tip-mounted either on a Bruker SMART platform with an Apex I-detector diffractometer for “Fe3OTFA” at 250 K or on a Bruker D8 SMART platform equipped with an Apex II detector for the other structures at 100 or 250 K using graphite-monochromated Mo Kα radiation (λ = 0.71073 Å). Using Olex2,[22] the structure was solved with the ShelXD[23] structure solution program using Dual Space or with the Superflip[24−26] structure solution program using charge flipping and refined with the ShelXL[27] refinement package using least-squares minimization. Thermogravimetric analysis (TGA) coupled to mass spectrometry (MS) was performed using a Netzsch simultaneous thermal analyzer (STA 449 F5 Jupiter) coupled (tubes at 200 °C) with a quadrupole mass spectrometer (QMS 403 D Aëolos). “Fe3OTFA” (13.1 mg) was placed in an aluminum crucible and heated under Ar gas flow (40 mL min–1) to 500 °C (5 °C min–1). Fourier transform infrared spectroscopy (FTIR) was measured on a Thermo Scientific Nicolet iS5 with an iD5 ATR accessory (diamond). The device is located in an argon glovebox (O2 < 0.1 ppm, H2O < 0.1 ppm).

Magnetic Measurements

The magnetic properties were measured on a Quantum Design PPMS-9. The magnetic susceptibility data of “Fe3OTFA” were measured in an applied magnetic field of 0.1 T between 2 and 300 K after cooling the sample in zero field. The temperature-dependence of the magnetization of FeO NPs was measured between 2 and 300 K in a field of 0.1 T after cooling without an applied field (zero-field cooled, ZFC) and after cooling in a magnetic field of 0.1 T (field cooled, FC). The field-dependence of the magnetization of FeO NPs was measured at 2 K after cooling in zero-field, and at 2, 10, 50, 100, 200, and 300 K after cooling to 2 K in a field of 1 T.

Results and Discussion

A direct and reproducible synthesis of a single-source precursor is the backbone of a successful nanoparticle synthesis method. For the presented synthesis of wüstite NPs, the precursor quality has a direct influence on the phase purity, size and shape of the FeO NPs. A compound obtained by reacting FeCl3 with TFA under reflux yields has been used as a catalyst in several organic syntheses.[28−35] Our investigation of this reaction identified a mixture of several products (Figures S1–S4, Tables S1–S8). Instead, [Fe3(μ3-O)(CF3COO)(μ-CF3COO)6(H2O)2]·CF3COOH as a single product was obtained in this work when the synthesis was modified by adding a defined amount of water at a later stage of the reaction. The phase-pure precursor was synthesized by reacting commercial FeCl3 and TFA at 86 °C for 26 h with the injection of water after 2 h. After evaporation of the residual TFA, the precursor was obtained as a red, air-stable powder and used for the synthesis of wüstite NPs without further purification. We hypothesize that water ensures sufficient level of hydrolysis of the preformed, suspended Fe(TFA)3. Instead of adding water, FeCl3·6H2O as a starting reagent also yields an identical precursor. However, the wüstite NPs obtained from a precursor produced by water-assisted synthesis were substantially more uniform and reproducible. The crystal structure of our precursor, denoted as “Fe3OTFA”, was determined using single-crystal XRD (see Figure a and the Supporting Information for all the crystallographic details) and has an overall composition of [Fe3(μ3-O)(CF3COO)(μ-CF3COO)6(H2O)2]·CF3COOH (Figure a and Figure S1 and Tables S1 and S2). The powder XRD pattern indicates phase purity (Figure b). “Fe3OTFA” crystallizes in the monoclinic space group P21/n. Three Fe3+ ions are arranged in a triangle around a central oxygen atom. Each side of the triangle comprises two bridging trifluoroacetates. Trans to the central oxygen atom, the sixth coordination site of the octahedral coordinated Fe3+ ions is occupied either by H2O (2x) or by a trifluoroacetate. This trifluoroacetate group forms a hydrogen bond to a TFA molecule. This “Fe3OTFA” and another earlier published structure[36] are the only known compounds with exclusively Fe3+ ions in the triangular core unit. Typically, this unit has only been observed with mixed (Fe2+/3+) valency.[37−39] “Fe3OTFA” is isostructural to [Cr3(μ3-O)(CF3COO)6(H2O)2(CF3COO)]·CF3COOH.[38] To elucidate the decomposition of “Fe3OTFA”, a TGA-MS analysis was conducted (Figure S5). In addition to the structure of “Fe3OTFA”, three additional crystal structures were obtained by varying the growth conditions of the single-crystal or by reducing the synthesis temperature. All three structures and the crystallographic details are presented in the Figures S2–S4 and Tables S3–S8. In this report, we focus exclusively on “Fe3OTFA” as a precursor for the synthesis of wüstite NPs, because this compound can be obtained in high yield without further purification.

Figure 2

Structural analysis of the precursor: (a) the crystal structure of [Fe3(μ3-O)(CF3COO)(μ-CF3COO)6(H2O)2]·CF3COOH as obtained from the single-crystal XRD and (b) simulated and experimentally measured powder XRD patterns of this compound indicating the excellent phase purity of the precursor.

Popcorn-shaped FeO (x ≈ 0.94) NPs (Figure a–c,f) were synthesized in a colloidal heating-up synthesis method at 260 °C (heating rate: ∼19 °C/min) by decomposing “Fe3OTFA” in TOP/squalane (1:1 by mass), HDA and OA under N2. The starting solution had a dark-red to black appearance. It turned into a clear, slightly yellowish solution at approximately 210 °C. A subsequent color change to brown and then to metallic black indicated the nucleation and growth of NPs, respectively. The growth was finished when the solution turned shiny black. A reaction temperature of 260 °C for 20 min and a 12:18:1 OA:HDA:“Fe3OTFA” molar ratio yielded the best size homogeneity. The size of the popcorn FeO NP clusters is slightly tunable by varying the precursor amount between 0.3 and 0.4 mmol (Figure S6). Higher precursor amounts led to inhomogeneous size distributions. For the synthesis of popcorn-shaped FeO NPs, obtaining the correct HDA/OA molar ratio and molar quantity is crucial. The molar ratio has to be between 1.3 and 2 (HDA/OA), and at least a 6-fold molar excess of OA to “Fe3OTFA” is needed. When the HDA amount is increased to a 22-fold excess, the resulting FeO NPs are separated, and no popcorn clusters are built (Figure S7). The growth mechanism of such polycrystalline popcorn-shaped NPs might first involve the formation of small (sub-10 nm) nuclei, which then aggregate into polycrystalline NPs. Such aggregation might be caused by the reduced solubility of NPs as they grow. The temperature at which the process occurs is presumably too low to cause efficient fusion of these crystalline domains.

The surface of the isolated FeO NPs appears to be covered exclusively by oleates, according to FTIR (Figure S8). TOP is known as an efficient phosphor source[40] and hence might contaminate the oxide. This possibility had been ruled-out by high-resolution elemental mapping of FeO NP using EDX measurements in HAADF-STEM mode (Figure ). A very small phosphide signal indicates no distinct P surface shell, but at most a minute residual contamination. The growth of the popcorn shape may be influenced by three ligands: OA, HDA and an amide. The formation of the latter might be catalyzed in situ by Fe3+ ions.[41,42] When adding “Fe3OTFA” only after the drying step (110 °C, 1.5 h, Figure S9), no popcorn-shaped FeO NPs are obtained, which supports the assertion that Fe3+-catalyzed amide formation is crucial to obtaining the popcorn-shape. The role of such a NP-yielding side reaction, whether by the amide as a ligand or a fluorine radical trap or by inducing a structural change in “Fe3OTFA”, remains uncertain. To obtain the wüstite phase, all four reactants (HDA, OA, TOP, and squalene) are needed, which was verified by excluding them one by one. The decomposition of “Fe3OTFA” in HDA/OA or HDA/TOP/squalene produces Fe3O4 NPs, indicating that the combination of OA with TOP/squalane reduces Fe3+ to Fe2+. Interestingly, in addition to the popcorn-shaped FeO, a second phase of highly monodisperse wüstite NPs appears when using slightly increased amounts of OA/HDA and synthesis time (Figure S10). Herein, we focused on the synthesis of popcorn-shaped NPs as this novel structure offers a unique morphology by combining approximately 14 nm large subparticles into clusters measuring up to 60 nm. The individual crystallite size was determined by Rietveld refinement (Figure S11) using a Si standard reference (NBS, no. 640, PDF 027–1402) to deconvolute the instrumental broadening. Such clusters, unlike completely fused nanocrystals, still provide a large surface area, but further agglomeration seems inhibited. Additionally, they have intrinsic channels that can be visualized by TEM tomography (Figure c and Video S1).

Figure 3

Elemental mapping using EDX measurements of FeO nanoparticle in the HAADF-STEM mode. (a) STEM image of a FeO nanoparticle. The yellow frame shows the area #1, from where the integrated STEM EDX spectra (see f and g) were extracted. The green line (integrated) shows the taken line profiles of the background corrected peak intensities. (b) Line profiles representing intensity distribution of Fe (red), O (green), P (blue), and HAADF (black). The elemental maps of the FeO particle for (c) Fe, (d) O, and (e) P. (f) Integrated STEM EDX spectrum taken from the yellow area #1 indicated in a. (g) Zoom into the STEM EDX spectrum showing a minor peak of P, which is 200 times smaller than the Fe–Kα peak. Cu arises from the sample holder (Cu grids) and the small Si peak from the detector.

For magnetic iron oxide-based NPs the combined effect of the granularity and the coexistence of the antiferromagnetic wüstite and uncompensated surface spins or surface oxides (γ-Fe2O3 and/or Fe3O4) can overcome the superparamagnetic magnetic limit due to the exchange bias arising at the interface of the oxides,[43] and by reducing the thermal tumbling of small subparticles because the cluster holds them in place. The phase composition of these FeO NPs remained constant for weeks. To obtain more insights into the properties of the wüstite NPs and the precursor (“Fe3OTFA”), we conducted magnetic measurements.

Figure a shows the magnetic susceptibility of the precursor “Fe3OTFA”, plotted as the χT product as a function of the temperature. χT at 300 K is 5.3 emu K mol–1, and it decreases to 0.45 emu K mol–1 at 2 K. The room temperature value is lower than expected for three uncoupled spins s = 5/2 of high- spin Fe3+ ions, χT = 13.125 emu K mol–1 (for g = 2). This lower-than-expected value at RT and the steady decrease in χT with decreasing temperature indicate strong antiferromagnetic interactions between the high-spin Fe3+ ions. The magnetic properties of a polynuclear spin-only complex can be modeled based on the spin Hamiltonian, as shown in eq. :[44]where J are the exchange coupling constants, are the spin operators, g are the Landé factors, μ is the Bohr magneton and is the magnetic field vector. Using PHI,[45] the data can be fit based on eq with i,j = 1,2,3 enumerating the Fe3+ ions. The best fit, shown as a solid black line (Figure a), was obtained with J12 = J13 = −28.7 cm–1, J23 = −43.2 cm–1, and g1 = g2 = g3 = 2.32, comparable to a recently studied oxygen-centered Fe3+ triangle.[46] This fit represents an isosceles triangle. While the symmetry of the “Fe3OTFA” core is low enough that three different exchange coupling constants and three different g-values could be used, such a model does not improve the fit significantly but instead overparameterizes the model. An argument can be made for an isosceles triangle based on the coordination of the Fe3+ ions: one has a different coordination than the other two (trifluoroacetates vs water trans to the central oxygen atom). Therefore, the complex contains two interactions between dissimilar Fe3+ ions and one between similar ones. Antiferromagnetic interactions in a triangle cannot all be satisfied at the same time; they are in competition. In an equilateral triangle, i.e., where all three exchange coupling parameters are the same, the molecule is spin frustrated with a doubly degenerate S = 1/2 ground state, which is unstable toward distortion or higher order exchange interactions (e.g., Dzyaloshinskii-Moriya interactions). In “Fe3OTFA”, the degeneracy is lifted by the unequal interactions in the triangle, leading to a gap between the two lowest S = 1/2 states (inset of Figure a).[46]

Figure 4

(a) Magnetic susceptibility (as the χT product) of “Fe3OTFA” vs temperature. Solid line: best fit to the Hamiltonian (eq ) with parameters g = 2.32, J12 = J13 = −28.7 cm–1, J23 = −43.2 cm–1. Inset: ground state of an equilateral (frustrated) and an isosceles triangle. (b) Magnetization vs temperature of popcorn-shaped FeO (x ≈ 0.94) NPs at 0.1 T, zero-field-cooled (ZFC, black squares) and field-cooled (FC, red circles). (c) Magnetization vs magnetic field of FeO NPs at 2 K after zero-field cooling (ZFC, dotted black line) and after cooling in a field of 1 T (FC, solid blue line); magnetization at 300 K (red dashed line). Inset: expanded region around the origin.

The magnetic properties of FeO NPs are presented in Figure b. Both the zero-field-cooled (ZFC) and field-cooled (FC) magnetization, at a field of 0.1 T between 300 and 2 K, show a broad maximum at approximately 220 K and a decrease below 200 K, which is steeper in the ZFC curve. The FC magnetization is almost flat between 150 and 25 K, after which a relatively steep increase occurs with decreasing temperatures. The ZFC magnetization is only flat down to 100 K, followed by a decrease trending to zero. Below 25 K, however, an increase similar to that in the FC magnetization is observed. The broad maxima at 220 K and the sharp drop below 200 K are a clear indication of the antiferromagnetic order known for wüstite, which has a Néel temperature of ∼200 K. In bulk FeO, TN increases from 192.4 to 199.2 K between x = 0.95 and 0.929.[47,48] Therefore, TN ≈ 200 K for our FeO NPs is slightly higher than expected, possibly indicating some nanosize effects. The divergence of the ZFC and FC magnetization below the Néel temperature, however, is unexpected for an antiferromagnet but indicative of superparamagnetism, ferromagnetism or spin-glass freezing. The increase at low temperatures can probably be ascribed to a paramagnetic impurity with a Curie or Curie–Weiss temperature-dependence of the magnetization.

The field-dependence of the magnetization at 2 K shows a clear opening around the origin, i.e., magnetic hysteresis. While the hysteresis after ZFC is symmetrical around zero, it is shifted to negative fields after cooling in a +1 T field. The coercive field of the 2 K ZFC magnetization is ±0.2 T. The negative FC coercive field is −0.32 T, and the positive side almost goes through the origin (0.02 T). Between 2 and 200 K, the FC coercivity narrows with increasing temperature (Figures S12 and S13) and disappears by 300 K. The opening of a hysteresis at a low temperature supports superparamagnetism as an explanation for the ZFC-FC divergence. The shift of the hysteresis curve after field-cooling is reminiscent of exchange bias.[43,49] Put together, these observations indicate that the FeO NPs can be viewed as antiferromagnetic cores with uncompensated surface spins. The surface spins lead to a net magnetic moment in the NPs. Due to the coupling of the surface spins to the antiferromagnetic core of the NPs, these surface spins have a preferred orientation and can be frozen in that orientation below the Néel temperature. That the hysteresis curves never saturate, not even at 9 T and 2 K, may also be attributed to the antiferromagnetic core of the FeO NPs. Similar effects have been observed in NiO and CuO NPs[50,51] but also in FeO/Fe3O4 core–shell NPs.[52−55] We cannot entirely rule out the partial oxidation of the FeO surface leading to a similar core–shell structure, as in earlier publications.[52−55] In our case, however, any hypothetical oxidized phase is below the detection limit of powder XRD and is invisible in TEM. The presence of an exchange bias alone might not be a sufficient evidence for the oxidation of the surface to form FeO/Fe3O4 core–shell NPs, as the examples of NiO and CuO NPs indicate.[50,51]

Conclusions

A gram-scale synthesis of a phase-pure iron oxo-trifluoroacetic precursor [Fe3(μ3-O)(CF3COO)(μ-CF3COO)6(H2O)2]·CF3COOH and its crystal structure are presented. Besides this “Fe3OTFA” compound, three additional compounds were obtained by slight variations of the reaction conditions or crystallization procedure. “Fe3OTFA” had been then used as a single-source precursor to obtain colloidal wüstite FeO NPs by decomposition in a high-boiling solvent. These wüstite NPs show an unprecedented popcorn morphology that offers large surface area and internal channels. A 3-D tomography model created from HAADF-STEM images reveals the striking details of these popcorn-shaped FeO NPs (see Video S1). Magnetic measurements of the FeO NPs show signs of the antiferromagnetic order expected for wüstite but also shows a divergence of the FC and ZFC magnetizations below TN, hysteretic magnetization curves and exchange bias. The magnetism of the precursor, “Fe3OTFA”, can be modeled by an isosceles antiferromagnetic triangle.