Crystallization Behavior and Thermal Analysis of CoFeB Thin Films

1. Introduction

In recent years, amorphous CoFeB has been found to have characteristics that can be exploited in scientific research and engineering applications, such as magnetic recording media, magnetoresistance random access memory (MRAM), and gauge sensors. CoFeB thin films with thicknesses in the range 10–50 Å are used in magnetoresistance (MR) devices, and annealed CoFeB thin films exhibit perpendicular or in-plane anisotropy and B diffusion [1–4]. CoFeB thin films demonstrate excellent magnetic and electrical properties because of their amorphous structure and high spin polarization. An as-deposited ferromagnetic CoFeB thin film is typically inserted into a spin-valve structure to form a free/pinned layer of magnetic tunneling junction (MTJ). The formation of the junction leads to increased tunneling magnetoresistance (TMR) and the development of ferromagnetism (FM)/antiferromagnetism (AFM) exchange-biasing anisotropy, which makes the structure suitable for both magnetoresistance random access memory and gauge sensor applications. The increased TMR and exchange-biasing anisotropy are caused by the mean free path of spin being shorter in amorphous materials compared with that of crystalline materials [5–11]. The concentration ratio of Co affects the stability of the amorphous state [12].

In this study, we investigated the crystalline behavior of amorphous CoFeB thin films through thermal analysis. The thermal characteristics and glass-forming ability (GFA) index of amorphous CoFeB thin films are worthy of research. The structure of CoFeB films was determined using X-ray diffraction (XRD) patterns. Nonisothermal differential scanning calorimetry (DSC; T_A Instruments DSC 2920) was used to determine thermal properties at a heating rate of 20 K/min. These properties included the glass transition temperature (T_g), onset crystallization temperature (T_x), and liquid temperature (T_l). Crucial related parameters were also measured to estimate the thermal performance, including the range of temperatures at which the film was in the supercooled region (ΔT_x = T_g − T_x) and the GFA index, which is obtained using γ = T_x/(T_g + T_l) and γ_m = (2T_x − T_g)/T_l. To evaluate the resistance to crystallization, the Kissinger formula was used to calculate the activation energy (Q) of crystallization. Thinner Co₄₀Fe₄₀B₂₀ and Co₆₀Fe₂₀B₂₀ films show a higher resistance to crystallization, a wider range of supercooled liquid temperatures, and a higher GFA index because their atomic arrangement is random. This experiment was conducted to improve the atomic crystalline behavior of transformed Co₄₀Fe₄₀B₂₀ and Co₆₀Fe₂₀B₂₀ films.

2. Experimental Details

CoFeB thin films were sputtered on a glass substrate by using DC magnetron sputtering at room temperature (RT) to obtain films with thicknesses ranging from 25 to 200 Å. The base chamber pressure exceeded 2 × 10⁻⁶ Torr, and the Ar working pressure was 5 × 10⁻³ Torr. The atomic compositions of the two CoFeB targets were 60 at % Co, 20 at % Fe, and 20 at % B and 40 at % Co, 40 at % Fe, and 20 at % B. XRD with a CuK_α1 line (Philips X’pert) was used to determine the amorphous structure. Thermal performance was investigated using nonisothermal DSC at a heating rate of 20 K/min. Heating rates in the range 10–40 K/min were applied in nonisothermal DSC analysis to determine the crystallization behavior. Several temperatures between T_g and T_x were applied in isothermal DSC analysis to examine the crystallization kinetics. DSC measurement provides qualitative and quantitative data in endothermic (heat absorbing) and exothermic (heat releasing) processes; nonisothermal and isothermal heating methods can be used to obtain information on changes in physical and/or chemical properties.

3. Results and Discussion

Figures 1(a) and 1(b) present the XRD patterns of Co₄₀Fe₄₀B₂₀ and Co₆₀Fe₂₀B₂₀ thin films in the 2θ range 20°–90°. This XRD result indicated that the XRD patterns of Co₄₀Fe₄₀B₂₀ and Co₆₀Fe₂₀B₂₀ thin films with thicknesses in the range 25–200 Å were amorphous. Thermal analysis was conducted to examine this phenomenon further (Figures 2(a) and 2(b)).

Figures 2(a) and 2(b) show that the amorphous Co₄₀Fe₄₀B₂₀ and Co₆₀Fe₂₀B₂₀ thin films undergo a change from an amorphous state to a crystalline state through an exothermic crystallization reaction and exhibit a supercooled liquid region in DSC measurements when heated at 40 K/min. Figure 2 shows that the two CoFeB thin films have a clear glass transition temperature and undergo an exothermic crystallization reaction in the supercooled liquid region.

The corresponding thermal properties of Co₄₀Fe₄₀B₂₀ thin films are shown in Table 1(a), and those of the Co₆₀Fe₂₀B₂₀ thin films are listed in Table 1(b). The evaluated parameters were used to calculate the supercooled liquid region (ΔT_x = T_g − T_x), GFA index, and glass transition temperature (T_rg = T_g/T_l, γ = T_x/(T_g + T_l)) and γ_m = (2T_x − T_g)/T_l [13–15]. The γ and γ_m values of the GFA index indicated that the Co₆₀Fe₂₀B₂₀ thin film with a thickness of 75 Å demonstrated the most effective GFA and showed a high supercooled liquid region (146 K) and high GFA index (γ = 0.350 and γ_m = 0.586). The high ΔT_x suggests that the film has high thermal stability. However, Co₄₀Fe₄₀B₂₀ yielded opposite results for the same film thickness. These results suggested that, for CoFeB alloy films, a thickness of 75 Å is critical.

The Kissinger formula was used to calculate the dependence of the Q of crystallization on the thickness of the Co₄₀Fe₄₀B₂₀ and Co₆₀Fe₂₀B₂₀ films, as shown in Figure 4. The two curves exhibit a cross feature and share a critical thickness of 75 Å. These activation energies yielded a critical thickness of 75 Å for amorphous Co₄₀Fe₄₀B₂₀ and Co₆₀Fe₂₀B₂₀ thin films. For the amorphous Co₄₀Fe₄₀B₂₀ film thickness of 75 Å, the lowest activation energy was approximately 45.78 kJ/mol. By contrast, at the same film thickness, the highest activation energy of Co₆₀Fe₂₀B₂₀ was approximately 83.44 kJ/mol. This result was consistent with the calculated DSC results. A higher activation energy corresponds with a higher resistance to crystallization. The thermal performance of the Co₆₀Fe₂₀B₂₀ thin film was higher than that of the Co₄₀Fe₄₀B₂₀ thin film with various Co and Fe concentrations [18–20].

The incubation time of the 75 Å-thick Co₆₀Fe₂₀B₂₀ thin film as a function of the isothermal annealing temperature revealed a longer time than that of the Co₄₀Fe₄₀B₂₀ thin film, as shown in Figure 6. This demonstrates that the thermal stability of Co₆₀Fe₂₀B₂₀ films is higher than that of the Co₄₀Fe₄₀B₂₀ films in isothermal analysis. According to results of DSC and Kissinger formula, the 75 Å-thick is a critical and important point in Co₆₀Fe₂₀B₂₀ and Co₄₀Fe₄₀B₂₀ thin films. The 75 Å-thick Co₆₀Fe₂₀B₂₀ thin film has the highest activation energy. In contrast, the 75 Å-thick Co₄₀Fe₄₀B₂₀ thin film has the lowest activation energy. Figures 5 and 6 investigated the crystallization fraction and incubation time to prove that the thermal stability of Co₆₀Fe₂₀B₂₀ thin film is stronger than Co₄₀Fe₄₀B₂₀ thin film.

4. Conclusions

In summary, the thermal performance, crystallization behavior, calculated anticrystallization, and structure of Co₄₀Fe₄₀B₂₀ and Co₆₀Fe₂₀B₂₀ thin films were investigated using DSC, XRD, and the Kissinger fitting method. The GFA index, defined using γ and γ_m, increased as the thickness decreased. Moreover, the Kissinger fitting indicated that the critical thickness of the CoFeB thin film was 75 Å. The performance of Co₆₀Fe₂₀B₂₀ thin films is more suitable for amorphous magnetic thin-film applications because of a high GFA index, Q, and ΔT_x. A critical result is that amorphous CoFeB thin films can be used in the magnetic recording industry and crystalline applications because of their crystalline behavior. Finally, based on the nonisothermal and isothermal analyses, the thermal stability and incubation time of Co₆₀Fe₂₀B₂₀ films were more favorable than those of Co₄₀Fe₄₀B₂₀ films.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.