1 INTRODUCTION

Nanoparticles of iron carbides Fe7C3 and Fe3C, encapsulated into carbon shells, are characterized by high biocompatibility and can be controlled by an external magnetic field. Hence, they can be considered as new-type platforms for biomedical applications [1].

These nanoparticles can be prepared by the decomposition of ferrocene molecules Fe(C5H5)2 at high pressure and high temperature (HPHT) [2, 3]; they may possess different magnetic properties, depending on the synthesis conditions. Therefore, investigation of HPHT phase transformations in ferrocene is an important and urgent problem.

It is known that three radically different regions can be distinguished in the ferrocene phase diagrams. The first one, which is characterized by low pressure and temperature, corresponds to different conformations of ferrocene molecules in the monoclinic phase [4–7]. The second region corresponds to the occurrence of polymerized states of organometallic compounds [8]. The third region in the P–T phase diagram is related to the HPHT decomposition of initial ferrocene molecules. This region is of particular interest for materials scientists because the decomposition of ferrocene is accompanied by the formation of new materials with unique structures and properties. In particular, the previous investigations of Fe(C5H5)2 transformations in Toroid high-pressure cells showed that, at a pressure of 8 GPa and temperatures up to 1870 K, ferrocene is transformed into amorphous iron carbide nanoparticles with a variable composition and crystalline nanoparticles of iron carbides Fe7C3 and Fe3C. It was established that Fe7C3 and Fe3C nanoparticles can be encapsulated into carbon or complex multilayer shells [9–12].

The pressure region near 10 GPa and temperature region near 2200 K are promising conditions for the formation of nanomaterials during the decomposition of ferrocene. The reason is that a binary mixture of iron carbide Fe7C3 and diamond is formed in the two-component Fe–C system under these conditions. One would expect the formation of magnetically controlled luminescent nanoparticles upon the decomposition of ferrocene under these conditions. These materials are important for drug delivery systems, magnetic resonance tomography, and therapeutic local hyperthermia. However, phase transformations in ferrocene Fe(C5H5)2 at these pressures and temperatures have not yet been investigated.

The purpose of this study is to analyze thermal transformations of ferrocene at a pressure of 10 GPa in high-pressure diamond anvil cells (DACs) upon laser heating of the sample to 2200 K.

2 EXPERIMENT

Ferrocene was synthesized using a method similar to that proposed in [13]. This method is described in detail in the Supplementary material.

It is known that, when studying iron-containing compounds, the Mössbauer effect can be observed only for 57Fe nuclei. Natural iron contains only about 2% of this Mössbauer isotope, which makes it impossible to carry out Mössbauer analyses of microsamples in DACs. In this study, ferrocene was synthesized using the Fe2O3 reagent enriched in the 57Fe isotope to 96%.

Diamond anvils with a culet 600 μm in size were fabricated to perform high-pressure experiments. A 60-μm-thick tungsten gasket with a hole 180 μm in diameter was used. A 45-μm-thick layer of ferrocene Fe(C5H5)2 powder was placed in the gasket hole (working volume of the DAC). A layer of MgO powder with a thickness of 5–10 μm was placed between the ferrocene and the anvil, which served simultaneously as a pressure-transferring medium and a thermal insulator with a high melting temperature (about 4150 K at 8.7 GPa [14]).

The pressure on the sample was monitored by the shift of the high-frequency edge of the Raman peak from the diamond anvil [15] using a Princeton Instruments Acton SP2500 monochromator/spectrograph. The excitation source was a 671-nm solid-state laser with diode pumping. It was established that the pressure nonuniformity across the sample was no more than ±0.2 GPa.

Laser heating of the sample was performed using the large-scale research facility [16–18]. Infrared laser radiation with a wavelength of 1064 nm was generated by a single-mode (TEM00) fiber laser (IPG Photonics). The radiation was focused on the ferrocene surface to a spot 25–30 μm in diameter. The sample was heated in the cw mode for 20 s. The laser-heating setup was described in detail in the supplementary material.

X-ray diffraction (XRD) analysis of the samples obtained was carried out on the Belok-RSA beamline of the Kurchatov Synchrotron Radiation Source at the Kurchatov Complex for Synchrotron and Neutron Investigations, National Research Centre Kurchatov Institute (Moscow, Russia) [19]. The images were recorded by a Rayonix SX165 two-dimensional detector. The synchrotron radiation wavelength was 0.75 Å.

The transmission electron microscopy (TEM) analysis was performed using an Osiris TEM/STEM microscope (Thermo Fisher Scientific, United States) equipped with a high-angle annular dark-field detector (Fischione, United States). After the HPHT treatment of ferrocene, the sample was removed from the DAC. To perform TEM studies, a plate (lamella) was cut from this sample using a focused ion beam in a Scios double-beam scanning electron–ion microscope (Thermo Fisher Scientific, United States).

Mössbauer spectra on 57Fe nuclei were first obtained for the sample under high pressure in the diamond anvil cell (after laser heating) and then at ambient pressure (after unloading and opening of the DAC). A point radioactive source of gamma rays 57Co(Rh) was at room temperature. Isomer shifts were calibrated using the spectrum of the standard α-Fe absorber. Model processing of the spectra was performed using the SpectrRelax program package [20].

3 EXPERIMENTAL INVESTIGATION OF THERMAL TRANSFORMATIONS IN FERROCENE

3.1 Calculation of the Spatial Temperature Distribution

To estimate the temperature gradient on the sample under laser irradiation, the spatial temperature distribution was calculated according to the technique described in detail in [18].

Heating under laser radiation is characterized by a stepwise increase in the intensity of the optical signal recorded by a multispectral camera (see the supplementary material for details). It was established that the heating process becomes more uniform after 5–7 s; afterwards, the spatial temperature distribution can be calculated correctly. In the course of time, the sample heating zone is reduced (see Fig. 1). One can conclude that, in the beginning of laser heating, the sample thermal conductivity is higher, and the heated area around the laser spot is large. The maximum temperature in the heating zone was 2200 K ± 3% (see Fig. 1).

Fig. 1.
figure 1

(Color online) Temperature distribution on the ferrocene surface in the high-pressure cell at different instants.

Full size image

3.2 X-Ray Diffraction Analysis

The X-ray diffraction pattern of the sample after removal from the DAC is presented in Fig. 2. Bar diagrams of the compounds found in the sample by the XRD analysis are given below the X-ray diffraction patterns.

Fig. 2.
figure 2

(Color online) X-ray diffraction pattern of the sample and bar diagrams of the most likely compounds. A microsample about 40 µm in size in the holder is shown in the inset (inside the dashed circle). The bar diagrams are plotted according to the database of the International Centre for Diffraction Data (ICDD): graphite (PDF no. 00-023-0064), o-Fe7C3 (PDF no. 01-075-1499), h-Fe7C3 (PDF no. 00-017-0333), FeO (PDF no. 00-006-0615), α-Fe (PDF no. 00-006-0696), MgO (PDF no. 00-004-0829), and Mg(OH)2 (PDF no. 00-044-1482).

Full size image

Reflection peaks in the X-ray diffraction pattern are significantly broadened, which may be attributed to diffuse scattering from small particles. The peak near angles 2Θ of 13° corresponds to the reflection from the (002) planes of graphite-like carbon.

Reflections in the range of angles 2Θ of 18°–26° are characteristic of the orthorhombic phase of Fe7C3 and α-Fe. This range of angles 2Θ also contains peaks, which can be assigned to reflections from carbon layers with two- and three-dimensional ordering types. In addition, one can observe peaks from cubic phases of FeO (at 17.3° (111), 20.1° (200), and 28.5° (220)) and MgO (at 20.5° (200) and 29.2° (220)). The presence of MgO in the sample is likely due to the fact that this material was used as a heat insulating layer in the DAC.

3.3 Transmission/Scanning Electron Microscopy and Energy-Dispersive X-Ray Microanalysis

The results of scanning/transmission electron microscopy (S/TEM) studies and energy-dispersive X-ray microanalysis (EDXMA) are shown in Figs. 3 and 4. The dark-field S/TEM image (with the detection of large-angle scattered electrons) is presented in Fig. 3a. The bright contrast due to the heaviest atoms in the sample indicates that the sample consists of spherical Fe nanoparticles with a characteristic size of 10–20 nm. This suggestion is also confirmed by the EDXMA elemental mapping given in Figs. 3b–3d. The distribution of Fe in Fig. 3b repeats completely the S/TEM image of nanoparticles in Fig. 3a. Carbon is distributed uniformly (see Fig. 3c). The distribution map of O (see Fig. 3d) is disperse, which is not due to Fe nanoparticles but may be caused by MgO powder residues (the heat insulating layer in the DAC) rather than by the FeOx shell of iron-containing nanoparticles.

Fig. 3.
figure 3

(Color online) (a) Dark-field S/TEM image of sample nanoparticles obtained by the detection of large-angle scattered electrons. Elemental distribution maps for the microsample from the high-pressure cell after high-pressure high-temperature treatment of ferrocene: (b) Fe, (c) C, and (d) O.

Full size image
Fig. 4.
figure 4

(Color online) Fe nanoparticles obtained after high-pressure high-temperature treatment of ferrocene: (a) bright-field TEM image, (b) HRTEM image (arrows with numbers indicate carbon fibers (1) associated with Fe nanoparticles and (2) separated from nanoparticles), (c) HRTEM image of an individual nanoparticle (the area of analysis of the nanoparticle lattice is indicated by a square), (d) two-dimensional Fourier spectrum from the area indicated in panel (c) (the spectrum coincides with the electron diffraction pattern from bcc-Fe in the [001] projection), and (e) HRTEM image of the portion of the nanoparticle lattice after filtering.

Full size image

The bright-field TEM image of a group of Fe nanoparticles obtained during the HPHT treatment of ferrocene is presented in Fig. 4a. Note that the dispersivity of nanoparticles is rather low. Along with nanoparticles, one can also observe carbon nanofibers in the sample, which are pronounced in the high-resolution TEM (HRTEM) images (see Fig. 4b). Some carbon fibers are associated with Fe nanoparticles (for example, those indicated by arrows with number 1). The others, separated from nanoparticles, are indicated by arrows with number 2. The HRTEM image of an individual nanoparticle is presented in Fig. 4c, where the area of analysis of the nanoparticle lattice is indicated by a square. A thorough analysis of the image using the Fourier transform (see Fig. 4d) demonstrated unambiguous coincidence between the crystal lattice and the bcc Fe lattice observed in the [001] projection. The HRTEM image of the area of the nanoparticle lattice after filtering (see Fig. 4e) shows a square grid with a period of 0.2 nm. One can clearly see in the image of the nanoparticle (see Fig. 4c) that a part of its surface is conjugated with a multilayer graphene shell with a thickness of 6–8 layers and a characteristic interplanar spacing of 0.34 nm, while the other part is conjugated with amorphous carbon.

3.4 Mössbauer Spectroscopy

Figure 5 shows the Mössbauer spectra of the sample recorded in the DAC at a pressure of 10.1 GPa after (a) laser irradiation and (b) unloading and opening of the DAC at ambient pressure. The spectra were recorded at a temperature of 295 K.

Fig. 5.
figure 5

(Color online) Mössbauer spectra of the sample at a pressure of 10.1 GPa (a) after laser irradiation and (b) after unloading and opening of the high-pressure cell at ambient pressure. A micrograph of the sample (indicated by a dashed circle) in the gasket hole is given in the inset.

Full size image

The experimental spectra were analyzed using a model consisting of several partial components corresponding to the magnetically ordered and paramagnetic states of iron atoms. The hyperfine interaction parameters obtained for these components from model processing are listed in Table 1.

Table 1. Hyperfine interaction parameters for different components obtained by model processing of the experimental Mössbauer spectra: the isomer shift δ, the quadrupole splitting in doublets Δ, the quadrupole shift in sextets ε, and the hyperfine magnetic field Hhf on 57Fe nuclei
Full size table

The Mössbauer spectra of the sample at a pressure of 10.1 GPa and at ambient pressure can be described by the same set of partial components. At ambient pressure, three magnetic sextets with the isomer shift parameters δ in the range of 0.18–0.30 mm/s and hyperfine magnetic fields Hhf = 16.0–22.5 T correspond to the iron carbide (Fe7C3) phase. The low-intensity sextet parameters (δ = 0.05 mm/s, ε = ‒0.05 mm/s, and Hhf = 32.62 T) correspond to the iron (α‑Fe) phase.

An analysis of the parameters of the paramagnetic components shows that one of the doublets is related to iron ions in the ferrocene phase; its parameters are δ = 0.43 mm/s and Δ = 2.36 mm/s. The second doublet (with the hyperfine parameters δ = 0.80 mm/s and Δ = 0.76 mm/s) is characteristic of Fe2+ atoms in the wüstite (Fe1 – xO) phase [21]. It should be noted that the resonance lines of this doublet have a large width (~1.6 mm/s). Taking into account that the XRD analysis revealed the MgO and FeO phases in the sample, we may assume that the second doublet corresponds to a mixture of (MgFe)1 – xO solid solutions with different compositions. These phases appear in the sample likely because Fe2+ ions may be incorporated into the MgO structure (heat insulating layer in the DAC) during the heat treatment [22].

4 RESULTS AND DISCUSSION

The complex analysis indicate that the solid-phase products of ferrocene transformation upon laser heating to 2200 K at a pressure of 10 GPa are nanoparticles of the crystalline phases of iron carbide Fe7C3 and iron α-Fe encapsulated into carbon shells and dispersed in a carbon matrix. The carbon shells and the matrix are composed of amorphous and graphite-like carbon. They contain fragments of two- and three-dimensionally ordered carbon layers. In addition, there is a mixture of iron and magnesium oxides in small amounts.

It is noteworthy that the phase compositions of the sample determined by the XRD, electron microscopy, and Mössbauer analyses are slightly different. The reason is that the XRD and electron microscopy studies were carried out (locally) for individual microparticles of the sample extracted from the DAC. The Mössbauer spectroscopy provided data from the entire material in the gasket hole.

The found phases are products of condensation and “secondary” structural ordering of the components from the “primary” medium (ferrocene decomposition products). The sample prepared in this study at a pressure of 10 GPa and a temperature of 2200 K contains simultaneously iron and iron carbide nanoparticles. The coexistence of these phases was not observed in the studies on thermal treatment of ferrocene at 8 GPa in the temperature range up to 1870 K and during ferrocene photolysis at temperatures of 2100–3000 K [11, 23, 24].

The coexistence of iron and iron carbide phases can be explained as follows. It is known that the thermally induced fragmentation of ferrocene molecules implies breaking of bonds in the following order: C‒H, C–C, and, finally, Fe–C [24]. The occurrence of iron particles can be explained by the breaking of all bonds in a ferrocene molecule. In this case, the primary medium contains chemically unbound iron atoms, the condensation of which leads to the formation of α-Fe nanoparticles. In turn, the formation of iron carbides indicates the presence of Fe–C clusters in the primary medium. Therefore, we may assume the existence of the reaction zones of two types in the DAC, which have different compositions of the primary medium. The suggested arrangement of these zones is shown in Fig. 6.

Fig. 6.
figure 6

(Color online) Schematic of the laser heating of ferrocene Fe(C5H5)2 in the high-pressure cell. The focused laser spot is 25–30 µm in diameter.

Full size image

In Fig. 6, the upper (surface) and lower heating zones in the sample are shown in yellow and red, respectively.

The upper zone corresponds to the high-temperature region, where thermally induced and (possibly) photoinduced breaking of Fe–C bonds occurs. In this case (as described above), the primary medium contains chemically unbound iron atoms. The condensation of these atoms leads to the formation of α-Fe nanoparticles.

The lower region is characterized by lower temperatures, at which the ferrocene fragmentation is not accompanied by the total breaking of Fe–C bonds. In this case, the condensation of metal-containing components of the primary medium leads to the formation of carbide Fe7C3.

Thus, the difference in the phase composition of ferrocene decomposition products at a temperature of 2200 K and a pressure of 10 GPa (in comparison with that described in [11, 23, 24]) is related to different compositions of the primary medium in these experiments.

Note that, in contrast to the results of studying the two-component Fe–C system [25], no diamond phase was found in our sample at similar pressures and temperatures. This fact can be explained by specific features of the diamond formation in two- (Fe–C) and three- (Fe–C–H) component growth systems. Our preliminary experiments demonstrate that diamond is formed at higher pressures during the ferrocene decomposition.

5 CONCLUSIONS

A series of experiments on the decomposition of ferrocene Fe(C5H5)2 upon laser heating to 2200 K have been carried out in a high-pressure diamond anvil cell at a pressure of 10 GPa. The spatial temperature distribution on the sample upon heating has been calculated. It has been established that the main components of the products of ferrocene transformation under the above conditions are nanoparticles of the crystalline phases of iron carbide Fe7C3 and iron α-Fe. These nanoparticles are dispersed in a matrix of graphite-like and amorphous carbon. The coexistence of iron and iron carbide nanoparticles in the sample is due to the presence of two temperature zones in the high-pressure cell with different compositions of the primary medium.