Fe-Co-B Soft Magnetic Ribbons: Crystallization Process, Microstructure and Coercivity.

In this work, a detailed microstructural investigation of as-melt-spun and heat-treated Fe67Co20B13 ribbons was performed. The as-melt-spun ribbon was predominantly amorphous at room temperature. Subsequent heating demonstrated an amorphous to crystalline α-(Fe,Co) phase transition at 403 °C. In situ transmission electron microscopy observations, carried out at the temperature range of 25–500 °C and with the heating rate of 200 °C/min, showed that the first crystallized nuclei appeared at a temperature close to 370 °C. With a further increase of temperature, the volume of α-(Fe,Co) crystallites considerably increased. Moreover, the results showed that a heating rate of 200 °C/min provides for a fine and homogenous microstructure with the α-(Fe,Co) crystallites size three times smaller than when the ribbon is heated at 20 °C/min. The next step of this research concerned the influence of both the annealing time and temperature on the microstructure and coercivity of the ribbons. It was shown that annealing at 485 °C for a shorter time (2 s) led to materials with homogenous distribution of α-(Fe,Co) crystallites with a mean size of 30 nm dispersed in the residual amorphous matrix. This was reflected in the coercivity (20.5 A/m), which significantly depended on the volume fraction of crystallites, their size, and distribution.


Introduction
Among several groups of soft magnetic materials, Fe-based amorphous and nanocrystalline alloys are extremely interesting from both a scientific and an application point of view [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17]. They exhibit not only optimal soft magnetic properties (e.g., low coercivity (H c ) and high permeability (µ')) but also are characterized by relatively low magnetic core losses (P s ) in comparison with other materials. For this reason, they find various applications in motors, transformers, actuators, sensors, and electronic communication devices [3]. However, soft magnetic properties are sensitive to chemical composition as well as microstructural features of materials [18]. It was reported that nanocomposites consisting of α-Fe crystallites with sizes smaller than 35 nm and surrounded by an amorphous matrix showed low coercivity and high permeability [1,18]. This behavior, explained by Herzer [18] in 1989, is based on a random anisotropy model in which the averaging of magnetocrystalline anisotropy plays a key

Materials and Methods
Fe 67 Co 20 B 13 master alloy was prepared by induction melting from chemical elements Fe (99.85%), Co (99.9%), and binary compound FeB 18 in an argon atmosphere. Then, the as-cast ingot was induction melted in an argon atmosphere and ejected with 26 kPa overpressure onto a copper wheel rotating with a linear speed of 34 m/s. The obtained ribbon (termed Ribbon 0) had 10 mm in width and 16 µm in thickness. Subsequent annealing at three temperatures for various times was performed at (i) 370 • C/60 s, (ii) 410 • C/30 s, and (iii) 485 • C/2 s, in a specially designed block heating system for the ultrarapid annealing technique, hereafter referred to as Ribbon 1, Ribbon 2, and Ribbon 3, respectively. The main parts of the heating system are two bulky copper blocks, heated up to the appropriate temperature. The temperature is stabilized by the PID controller, in terms of the thermocouple's signal, situated in one of the blocks close to the sample's surface. The microstructure of ribbons was examined using a Tecnai G2 Transmission Electron Microscope (TEM) (FEI, Eindhoven, The Netherlands) equipped with an energy-dispersive X-ray microanalyzer (EDX) (EDAX, Mahwah, NJ, USA) and a High Angle Annular Dark Field (HAADF) detector (Fischione, Pennsylvania, Pittsburgh, PA, USA). The Gatan 628 heating holder was used for the in situ experiments. Thin foils for TEM observations were prepared with TenuPol-5 double jet electropolished using an electrolyte of perchloric acid (20 vol.%) and methanol (80 vol.%) at a temperature of about −20 • C. Room temperature high-energy, wide-angle X-ray diffraction measurements were carried out at DESY synchrotron in Germany, Hamburg using the beamline P07 (87.1 keV, λ = 0.0142342 nm). The samples were rotated 180 • around the ω axis in order to obtain diffraction with good grain statistics [24]. The amount of crystalline phase was calculated using Origin 2018 Academic software, from the integrated area of each peak using the following equation: area of crystalline peaks/(area of crystalline peaks+amorphous peaks). The background correction was performed using HighScore Plus software. Thermoanalysis was performed by a differential scanning calorimetry (DSC) using a thermal analyzer (Netzsch DSC 404C Pegasus, Netzsch-Gerätebau GmbH, Selb, Germany)) instrument with a heating rate of 20 • C/min. Magnetic measurements were performed in a hysteresis loop tracer, specially designed for soft magnetic materials in the ribbon shape attachment. Properties of extremely soft magnetic materials were measured with the high sensitivity in the magnetic field with a range of ±660 A/m. The measuring unit was based on the idea described by Kulik et al. [25]. Figure 1a shows the integrated high-energy synchrotron radiation diffraction patterns for the as-spun ribbon. According to the synchrotron diffraction patterns, the as-spun ribbon is almost fully amorphous with a characteristic broad diffused diffraction halo. Only a trace amount of crystallites (about 2%) in the amorphous matrix is detected. Very low peaks are observed at the 2Theta angles (~4.0, 5.6, and 7.0 degrees) that indicate the presence of a negligible amount in the primary α-(Fe,Co) phase. Han et al. investigated the influence of boron addition on the structure of ribbons and showed the two-phase structure of Fe 67 Co 20 B 13 as-spun ribbon [22]. The differential scanning calorimetry (DSC) curve ( Figure 2b) recorded on heating for Fe 67 Co 20 B 13 as-spun ribbon showed the two-stage crystallization process. The first exothermic peak (T p1 ), with the minimum at 403 • C, corresponds to the crystallization of the α-(Fe,Co) phase. The second peak (T p2 ) occurring at 522 • C is related to the crystallization processes of iron and/or cobalt borides. Based on DSC curves, the temperatures for the heat treatment process have been selected, i.e., below and above the T p1 value.

Characterization of the as Spun Fe 67 Co 20 B 13 Ribbon
Materials 2020, 13, x FOR PEER REVIEW 3 of 11 (87.1 keV, λ = 0.0142342 nm). The samples were rotated 180° around the ω axis in order to obtain diffraction with good grain statistics [24]. The amount of crystalline phase was calculated using Origin 2018 Academic software, from the integrated area of each peak using the following equation: area of crystalline peaks/(area of crystalline peaks+amorphous peaks). The background correction was performed using HighScore Plus software. Thermoanalysis was performed by a differential scanning calorimetry (DSC) using a thermal analyzer (Netzsch DSC 404C Pegasus, Netzsch-Gerätebau GmbH, Selb, Germany)) instrument with a heating rate of 20 °C/min. Magnetic measurements were performed in a hysteresis loop tracer, specially designed for soft magnetic materials in the ribbon shape attachment. Properties of extremely soft magnetic materials were measured with the high sensitivity in the magnetic field with a range of ±660 A/m. The measuring unit was based on the idea described by Kulik et al. [25]. Figure 1a shows the integrated high-energy synchrotron radiation diffraction patterns for the as-spun ribbon. According to the synchrotron diffraction patterns, the as-spun ribbon is almost fully amorphous with a characteristic broad diffused diffraction halo. Only a trace amount of crystallites (about 2%) in the amorphous matrix is detected. Very low peaks are observed at the 2Theta angles (~4.0, 5.6, and 7.0 degrees) that indicate the presence of a negligible amount in the primary α-(Fe,Co) phase. Han et al. investigated the influence of boron addition on the structure of ribbons and showed the two-phase structure of Fe67Co20B13 as-spun ribbon [22]. The differential scanning calorimetry (DSC) curve ( Figure 2b) recorded on heating for Fe67Co20B13 as-spun ribbon showed the two-stage crystallization process. The first exothermic peak (Tp1), with the minimum at 403 °C, corresponds to the crystallization of the α-(Fe,Co) phase. The second peak (Tp2) occurring at 522 °C is related to the crystallization processes of iron and/or cobalt borides. Based on DSC curves, the temperatures for the heat treatment process have been selected, i.e., below and above the Tp1 value.     (Figure 2c,d). The SADP has been indexed in accordance with the bcc α-(Fe,Co) crystal structure. The size of crystalline regions was estimated to be between 100 and 200 nm. Moreover, the high-resolution transmission electron microscopy (HRTEM) micrograph, fast Fourier transform (FFT), and inverse fast Fourier transform (IFFT) images taken from (1) crystalline and (2) amorphous regions, marked with squares, are presented in Figure 3. The FFT obtained from the HRTEM image pointed by square 1 can be well indexed based on the α-(Fe,Co) structure confirming SADPs results. Additionally, the lattice fringes of the observed crystallite correspond to the (110) planes of the bcc α-(Fe,Co) phase. The area of square 2 reveals the existence of local atomic ordered regions with a size of about 1-2 nm, called "nanocrystalline (atomic) clusters" randomly dispersed in the amorphous matrix (highlighted by yellow ovals). Furthermore, onion-like contrasts are marked by blue arrows. This kind of local microstructure was already reported and described in Fe-Si-B-P-Cu systems [19].

Characterization of the as Spun Fe67Co20B13 Ribbon
Materials 2020, 13, x FOR PEER REVIEW 4 of 11 taken from (1) crystalline and (2) amorphous regions, marked with squares, are presented in Figure  3. The FFT obtained from the HRTEM image pointed by square 1 can be well indexed based on the α-(Fe,Co) structure confirming SADPs results. Additionally, the lattice fringes of the observed crystallite correspond to the (110) planes of the bcc α-(Fe,Co) phase. The area of square 2 reveals the existence of local atomic ordered regions with a size of about 1-2 nm, called "nanocrystalline (atomic) clusters'' randomly dispersed in the amorphous matrix (highlighted by yellow ovals). Furthermore, onion-like contrasts are marked by blue arrows. This kind of local microstructure was already reported and described in Fe-Si-B-P-Cu systems [19].

In situ TEM Heating Observations
The in situ TEM experiments were performed with two heating rates of 20 and 200 °C/min to compare and recreate conditions typically employed during the experimental heat treatment. During taken from (1) crystalline and (2) amorphous regions, marked with squares, are presented in Figure  3. The FFT obtained from the HRTEM image pointed by square 1 can be well indexed based on the α-(Fe,Co) structure confirming SADPs results. Additionally, the lattice fringes of the observed crystallite correspond to the (110) planes of the bcc α-(Fe,Co) phase. The area of square 2 reveals the existence of local atomic ordered regions with a size of about 1-2 nm, called "nanocrystalline (atomic) clusters'' randomly dispersed in the amorphous matrix (highlighted by yellow ovals). Furthermore, onion-like contrasts are marked by blue arrows. This kind of local microstructure was already reported and described in Fe-Si-B-P-Cu systems [19].

In situ TEM Heating Observations
The in situ TEM experiments were performed with two heating rates of 20 and 200 °C/min to compare and recreate conditions typically employed during the experimental heat treatment. During

In Situ TEM Heating Observations
The in situ TEM experiments were performed with two heating rates of 20 and 200 • C/min to compare and recreate conditions typically employed during the experimental heat treatment. During the in situ TEM experiments, carried out with a heating rate of 200 • C/min, the evolution of microstructure was observed (Figure 4a). At 375 • C, in consistence with the DSC scan, the crystallization nuclei were revealed, followed by the dendritic growth of crystals (430 • C). The same phenomenon was reported in [23] for Fe 85-x Co x B 15 alloys for x within the range of 12 < x < 25 at.%. The presence of the crystallized nuclei at this low temperature was associated with a phase separation taking place within the amorphous matrix. This led to the formation of a dendritic α-(Fe,Co) structure, whereas the majority of the sample volume remained in its amorphous state.
Materials 2020, 13, x FOR PEER REVIEW 5 of 11 the in situ TEM experiments, carried out with a heating rate of 200 °C/min, the evolution of microstructure was observed (Figure 4a). At 375 °C, in consistence with the DSC scan, the crystallization nuclei were revealed, followed by the dendritic growth of crystals (430 °C). The same phenomenon was reported in [23] for Fe85-xCoxB15 alloys for x within the range of 12 < x < 25 at.%. The presence of the crystallized nuclei at this low temperature was associated with a phase separation taking place within the amorphous matrix. This led to the formation of a dendritic α-(Fe,Co) structure, whereas the majority of the sample volume remained in its amorphous state. More significant differences were observed in the electron diffraction patterns (SADP) with the increase of temperature ( Figure 5). The SADP taken at room temperature showed two typical diffused halo rings corresponding to the most intense reflections of the α phase, i.e., (111) and (211). At 400 and 450 °C, diffused spots appeared on the first ring (counting from the center of the diffraction pattern) and also on an additional diffraction ring that appeared between the first and the second halo ring. At 500 °C, all the diffraction rings are well-developed, and they correspond with high accuracy to the crystallographic planes of the α-(Fe,Co) phase. In addition, one can see individual reflections marked by arrows, which are characteristic of the Fe2B phase. In order to compare the microstructural changes with respect to the heating rate applied (20 and 200 °C/min), the BF image analysis was supplemented with measurements of crystallite sizes (dendrite arms cross-sections) at various temperatures.   [13]. From Figure 6, one can see that in both cases, the same crystallization mechanism occurs, including nucleation and dendritic growth. However, the crystallite size for More significant differences were observed in the electron diffraction patterns (SADP) with the increase of temperature ( Figure 5). The SADP taken at room temperature showed two typical diffused halo rings corresponding to the most intense reflections of the α phase, i.e., (111) and (211). At 400 and 450 • C, diffused spots appeared on the first ring (counting from the center of the diffraction pattern) and also on an additional diffraction ring that appeared between the first and the second halo ring. At 500 • C, all the diffraction rings are well-developed, and they correspond with high accuracy to the crystallographic planes of the α-(Fe,Co) phase. In addition, one can see individual reflections marked by arrows, which are characteristic of the Fe 2 B phase. In order to compare the microstructural changes with respect to the heating rate applied (20 and 200 • C/min), the BF image analysis was supplemented with measurements of crystallite sizes (dendrite arms cross-sections) at various temperatures.
Materials 2020, 13, x FOR PEER REVIEW 5 of 11 the in situ TEM experiments, carried out with a heating rate of 200 °C/min, the evolution of microstructure was observed (Figure 4a). At 375 °C, in consistence with the DSC scan, the crystallization nuclei were revealed, followed by the dendritic growth of crystals (430 °C). The same phenomenon was reported in [23] for Fe85-xCoxB15 alloys for x within the range of 12 < x < 25 at.%. The presence of the crystallized nuclei at this low temperature was associated with a phase separation taking place within the amorphous matrix. This led to the formation of a dendritic α-(Fe,Co) structure, whereas the majority of the sample volume remained in its amorphous state. More significant differences were observed in the electron diffraction patterns (SADP) with the increase of temperature ( Figure 5). The SADP taken at room temperature showed two typical diffused halo rings corresponding to the most intense reflections of the α phase, i.e., (111) and (211). At 400 and 450 °C, diffused spots appeared on the first ring (counting from the center of the diffraction pattern) and also on an additional diffraction ring that appeared between the first and the second halo ring. At 500 °C, all the diffraction rings are well-developed, and they correspond with high accuracy to the crystallographic planes of the α-(Fe,Co) phase. In addition, one can see individual reflections marked by arrows, which are characteristic of the Fe2B phase. In order to compare the microstructural changes with respect to the heating rate applied (20 and 200 °C/min), the BF image analysis was supplemented with measurements of crystallite sizes (dendrite arms cross-sections) at various temperatures.   [13]. From Figure 6, one can see that in both cases, the same crystallization mechanism occurs, including nucleation and dendritic growth. However, the crystallite size for  [13]. From Figure 6, one can see that in both cases, the same crystallization mechanism occurs, including nucleation and dendritic growth. However, the crystallite size for samples heated at 20 • C/min is three times larger than the crystal size of samples heated at 200 • C/min for the same measurement temperatures. In addition, by analyzing the nature of the histograms, it can be concluded that in the first case (20 • C/min), a bimodal crystallite size distribution occurs at all temperatures. In contrast, for the second case at temperatures of 410 • C and 480 • C, unimodal normal type distribution occurs. In summary, it can be stated that the crystallization process of amorphous Fe 67 Co 20 B 13 ribbons at 20 and 200 • C/min heating rates carried out by in situ TEM leads to visibly different microstructures in respect of the particular size and distribution of the crystalline phase, which in turn leads to various magnetic properties.       Figure 8 presents the synchrotron X-ray diffraction patterns of heat-treated ribbons. Samples after annealing show a two-phase structure consisting of both amorphous and α-(Fe,Co) crystalline phases. Moreover, the amount of crystalline α-(Fe,Co) phase calculated from X-ray diffraction patterns is found to be 6.5%, 38.7%, and 91.6% for ribbons annealed at 370, 410, and 485 • C, respectively. It is therefore clear that the higher the heat treatment temperature is, the greater the crystalline to amorphous phase ratio. sample heated in situ to a temperature of 500 °C with a heating rate of 200 °C/min. Reflections corresponding to interplanar distances such as 2.51 and 2.03 Å, as well as diffusion rings corresponding to the amorphous phase, are well visible on the FFT image taken from the marked area (white square). According to the angle's measurement, the presence of α-(Fe,Co) and (Fe,Co)2B phases with the [111] and [012] zone axes may be confirmed, respectively. Accordingly, the IFFT's two nano regions corresponding to the α-(Fe,Co) and Fe2B phases separated by the amorphous phase can be distinguished (Figure 7 right image).   Figure 8 presents the synchrotron X-ray diffraction patterns of heat-treated ribbons. Samples after annealing show a two-phase structure consisting of both amorphous and α-(Fe,Co) crystalline phases. Moreover, the amount of crystalline α-(Fe,Co) phase calculated from X-ray diffraction patterns is found to be 6.5%, 38.7%, and 91.6% for ribbons annealed at 370, 410, and 485 °C, respectively. It is therefore clear that the higher the heat treatment temperature is, the greater the crystalline to amorphous phase ratio.  Figure 9 shows the set of BF micrographs (a, c, e) and SADPs (b, d, f) for Fe67Co20B13 ribbons after heat treatment performed under various conditions (370 °C/60 s, 410 °C/30 s, and 485 °C/2 s). Ribbon 1 has α-(Fe,Co) crystallites with an average crystal size of 53 ± 13 nm, randomly placed in the amorphous matrix. Both the BF image and SADP indicate that the content of the crystalline phase is much smaller than the amorphous phase. Ribbon 2 contains much more α-(Fe,Co) crystal grains compared to Ribbon 1, while the crystallites' size decreases to 37 ± 8 nm. Ribbon 3 annealed at the highest temperature, within the investigated temperature range at almost fully crystalline, with the small amorphous regions surrounding α-(Fe,Co) crystallites with an average size of 30 ± 8 nm. By the increase of annealing temperature, extra spots appear. The new phase was identified to be iron and/or cobalt borides (Fe,Co)2B. The distinction between both phases cannot be made due to the same crystallites structure (I4/mcm) and similar lattice parameters (for Fe2B and for Co2B). In this case, the amount of borides was almost negligible. However, it is well known that the presence of borides in greater amounts has a negative impact on soft magnetic behavior leading, e.g., to the increase of coercive fields.  Figure 9 shows the set of BF micrographs (a, c, e) and SADPs (b, d, f) for Fe 67 Co 20 B 13 ribbons after heat treatment performed under various conditions (370 • C/60 s, 410 • C/30 s, and 485 • C/2 s). Ribbon 1 has α-(Fe,Co) crystallites with an average crystal size of 53 ± 13 nm, randomly placed in the amorphous matrix. Both the BF image and SADP indicate that the content of the crystalline phase is much smaller than the amorphous phase. Ribbon 2 contains much more α-(Fe,Co) crystal grains compared to Ribbon 1, while the crystallites' size decreases to 37 ± 8 nm. Ribbon 3 annealed at the highest temperature, within the investigated temperature range at almost fully crystalline, with the small amorphous regions surrounding α-(Fe,Co) crystallites with an average size of 30 ± 8 nm. By the increase of annealing temperature, extra spots appear. The new phase was identified to be iron and/or cobalt borides (Fe,Co) 2 B. The distinction between both phases cannot be made due to the same crystallites structure (I4/mcm) and similar lattice parameters (for Fe 2 B and for Co 2 B). In this case, the amount of borides was almost negligible. However, it is well known that the presence of borides in greater amounts has a negative impact on soft magnetic behavior leading, e.g., to the increase of coercive fields. HRTEM micrograph for Ribbon 3 is shown in Figure 10. It proves that α-(Fe,Co) crystallites with a mean size of 30 nm are spread in the amorphous matrix. The aforementioned ribbons, due to their unique microstructure, are called "nanocomposite materials", where the α-Fe crystallites are embedded in an amorphous matrix, which is beneficial from a magnetic properties point of view.  HRTEM micrograph for Ribbon 3 is shown in Figure 10. It proves that α-(Fe,Co) crystallites with a mean size of 30 nm are spread in the amorphous matrix. The aforementioned ribbons, due to their unique microstructure, are called "nanocomposite materials", where the α-Fe crystallites are embedded in an amorphous matrix, which is beneficial from a magnetic properties point of view. HRTEM micrograph for Ribbon 3 is shown in Figure 10. It proves that α-(Fe,Co) crystallites with a mean size of 30 nm are spread in the amorphous matrix. The aforementioned ribbons, due to their unique microstructure, are called "nanocomposite materials", where the α-Fe crystallites are embedded in an amorphous matrix, which is beneficial from a magnetic properties point of view.   Figure 11 presents B-H (where B is induction and H is a magnetic field) loops of heat-treated Fe 67 Co 20 B 13 ribbons in the correlation with microstructural features apparent in BF images. The coercivity (H c ) considerably diminishes with the increase of heat treatment temperature from 49.8 A/m for Ribbon 1 to 20.5 A/m for Ribbon 3. This behavior is a consequence of microstructure evolution, including both α-(Fe,Co) crystallite sizes and their volume fraction as well as the distribution in the amorphous matrix. In the case of the ribbon annealed at a temperature of 370 • C for 60 s, slightly below the onset of α-(Fe,Co) phase crystallization peak (Figure 1b), the crystallites are the largest (53 nm), unevenly embedded in the matrix. Here, the α-(Fe,Co) crystallites grow from Fe-rich regions (well observed in HRTEM- Figure 4) while the annealing time (60 s) is sufficient for crystal growth. As a consequence of the existence of larger, heterogeneously distributed α-(Fe,Co) crystallites, the coercive field value is the largest. However, Ribbon 3 annealed at 485 • C for two s is characterized by a fine microstructure, where the crystallites with an average crystallite size of 30 nm are uniformly distributed in the amorphous matrix. In this case, α-(Fe,Co) crystallites are connected with both preexisting nuclei as well as newly formed ones. Short annealing times and a large number of nuclei inhibit crystallites growth. Thus, a low coercive field (20.5 A/m) in this alloy can be explained by the fact that the exchange correlation length is larger than the crystallite's size. Additionally, it can be noticed that the saturation magnetic induction (B s ) was estimated to increase subtly with the annealing temperature.

Magnetic Properties
Materials 2020, 13, x FOR PEER REVIEW 9 of 11 Figure 11 presents B-H (where B is induction and H is a magnetic field) loops of heat-treated Fe67Co20B13 ribbons in the correlation with microstructural features apparent in BF images. The coercivity (Hc) considerably diminishes with the increase of heat treatment temperature from 49.8 A/m for Ribbon 1 to 20.5 A/m for Ribbon 3. This behavior is a consequence of microstructure evolution, including both α-(Fe,Co) crystallite sizes and their volume fraction as well as the distribution in the amorphous matrix. In the case of the ribbon annealed at a temperature of 370 °C for 60 s, slightly below the onset of α-(Fe,Co) phase crystallization peak (Figure 1b), the crystallites are the largest (53 nm), unevenly embedded in the matrix. Here, the α-(Fe,Co) crystallites grow from Fe-rich regions (well observed in HRTEM- Figure 4) while the annealing time (60 s) is sufficient for crystal growth. As a consequence of the existence of larger, heterogeneously distributed α-(Fe,Co) crystallites, the coercive field value is the largest. However, Ribbon 3 annealed at 485 °C for two s is characterized by a fine microstructure, where the crystallites with an average crystallite size of 30 nm are uniformly distributed in the amorphous matrix. In this case, α-(Fe,Co) crystallites are connected with both preexisting nuclei as well as newly formed ones. Short annealing times and a large number of nuclei inhibit crystallites growth. Thus, a low coercive field (20.5 A/m) in this alloy can be explained by the fact that the exchange correlation length is larger than the crystallite's size. Additionally, it can be noticed that the saturation magnetic induction (Bs) was estimated to increase subtly with the annealing temperature.

Conclusions
Based on the in situ TEM experiments and the nano-and microstructure observations, the following conclusions can be drawn: (i) independently of the heating rate, the crystallization process of amorphous Fe67Co20B13 melt-spun ribbons is realized by the nucleation and dendritic growth of α-(FeCo) phase, while the first crystallization effects are manifested at a temperature close to 370 °C; (ii) finer and more homogeneous microstructures are observed in the case of sample heated with the heating rate of 200 °C/min than in the one heated with 20 °C/min; (iii) formation of Fe2B phase at 500 °C during heating with the 200 °C/min heating rate is confirmed by the HREM investigations. The

Conclusions
Based on the in situ TEM experiments and the nano-and microstructure observations, the following conclusions can be drawn: (i) independently of the heating rate, the crystallization process of amorphous Fe 67 Co 20 B 13 melt-spun ribbons is realized by the nucleation and dendritic growth of α-(FeCo) phase, while the first crystallization effects are manifested at a temperature close to 370 • C; (ii) finer and more homogeneous microstructures are observed in the case of sample heated with the heating rate of 200 • C/min than in the one heated with 20 • C/min; (iii) formation of Fe 2 B phase at 500 • C during heating with the 200 • C/min heating rate is confirmed by the HREM investigations. The aforementioned results prelude the ultrarapid annealing process for Fe-based soft magnetic ribbons being interesting and prospective from the scientific and application point of view. Moreover, we have examined the microstructure of Fe 67 Co 20 B 13 ribbons after heat treatment performed under various conditions: (1) 370 • C/60 s, (2) 410 • C/30 s, and (3) 485 • C/2 s. These results were then correlated with the coercivity values. It was found that the annealing at the higher temperature (485 • C) for a very short time (2 s) provides a fine, homogenous microstructure resulting in lower coercivity H c = 20.5 A/m and magnetic induction of B > 1.5 T. Further tests, including structural and magnetic studies, are needed to optimize the ultrarapid annealing process of this material.