Exploration of MAB phase formation in the Fe-Y-Al-B system using thin film materials libraries

Transition metal ceramics are characterized by their high melting points, high thermal stability and oxidation resistance as well as excellent mechanical properties, such as high hardness. A special subclass is nanolaminated ceramics with MAX phases at the forefront [1] followed by a more recent addition of MAB phases [2], [3]. MAX and MAB phases are both ternary systems composed of a transition metal (M), an A-group element (A) as well as C and/or N (X) or B. MAB phases in particular are interesting since the favored M−elements are shifted towards higher group numbers compared to MAX phases and thus include magnetic elements such as Mn, Fe, Co, and Ni in thermodynamically stable ternary MAB phase configurations. The resultant properties make MAB phases, e.g., Fe₂AlB₂[4], promising candidates for magnetic and magnetocaloric applications.

Owing to their structural and chemical versatility, MAX and MAB phases provide an extensive platform for materials design. Adjustments of composition is a viable approach for tuning materials properties and adding new functionalities. For MAX phases in particular, M−site solid-solutions have been extensively explored, where mixing two transition metals results in either a disordered distribution [5], [6] or a chemically ordered arrangement [7], [8] on the M−sublattice. The latter is particularly intriguing and has led to realization of two-dimensional (2D) derivatives (MXenes) with remarkably improved capacitance and catalytic properties [9]. The prospect of ordered MAB phases was opened by theoretical predictions and subsequent synthesis of in-plane ordered (Mo_2/3Y_1/3)₂AlB₂ and (Mo_2/3Sc_1/3)₂AlB₂[10] as well as out-of-plane ordered Ti₄MoSiB₂[11], coined i-MAB and o-MAB phases, respectively, realized by Dahlqvist et al., immediately expanded to theoretical high-throughput materials stability studies [12], [13]. First-principles density functional theory (DFT) calculations were used to assess thermodynamic stability of a MAB phase by comparing its energy to the energy of a set of the most competing phases determined by linear optimization procedure. Chemical disorder on the M−sublattice was modelled using special quasi-random structure (SQS) method. Contribution from configurational entropy was considered and formation energies were calculated at typical MAB phase bulk synthesis temperature of 2000 K. Additional temperature effects, such as lattice vibrations, were disregarded, since such contributions tend to cancel out in the calculated stability. The predictive power of the studies is corroborated by all to date experimentally realized MAB phases identified as thermodynamically stable. Most importantly, a large subgroup of 39 in-plane ordered and 52 chemically disordered quaternary MAB phases were predicted as thermodynamically stable [12]. The established theoretical protocol confirmed by experimental synthesis motivates further exploration of new element combinations in quest of novel MAB phases.

The present work was motivated by the theoretical study of Dahlqvist and Rosen [12] and aims for the experimental verification of the MAB phase (Fe_2/3Y_1/3)₂AlB₂. In Ref. [12] this phase is predicted as the most stable quaternary i-MAB phase with M₂AB₂ stoichiometry, but it has not yet been synthesized. The ternary phase Fe₂AlB₂ has been thoroughly investigated due to its promising magnetic and magnetocaloric properties [4], [14], [15]. Furthermore, Y is a common element in i-MAX phase synthesis [16] and Y-containing i-MAB phases have been reported [10]. We investigate the Fe-Y-Al-B system by using combinatorial thin film synthesis followed by high-throughput X-ray diffraction (XRD) and energy dispersive X-ray spectroscopy (EDX) characterizations as well as further transmission electron microscopy (TEM) and atom probe tomography (APT) analysis for samples selected from the materials libraries (MLs).

Two MLs of the system Fe-Y-Al-B were synthesized as thin films with continuous composition spreads on 100 mm diameter single crystal Al₂O₃(0001) wafers by confocal co-sputtering from 100 mm diameter elemental Fe (purity 99.99 %), Y (99.9 %), Al (99.9995 %), and two B (99.5 %) targets in an ultra-high vacuum (UHV) magnetron sputtering system (DCA Instruments, Finland). The targets were positioned at 19 cm target-to-substrate distance with an inclination angle of 45° with respect to the substrate surface. Al and the two B targets were operated in radio frequency (RF) mode at 84 W and 300 W, respectively, while Fe and Y were sputtered in direct current (DC) mode, both at 17 W for ML1 and 17 W and 8 W, respectively, for ML2. Two B targets were operated at maximum power in order to compensate for the low sputtering rate inherent to B. In order to prevent thermal shock, the B targets were bonded on a Cu plate and the applied power was increased at a rate of 50 W/min. Prior to the depositions, the substrates were cleaned in acetone and isopropanol ultrasonic baths for 5 min each, heated at 50 °C/min to the deposition temperature of 700 °C and held at it for 10 min to ensure equal temperature distribution and removal of surface adsorbents. The base pressure was ∼ 2 x 10^-6 Pa at room temperature (RT) and ∼ 2 x 10^-5 Pa at 700 °C. 6 N grade Ar (99.9999 at.%) was used as the sputter gas at 20 sccm flow and 0.26 Pa partial pressure. The depositions were carried out for 30 min and 1 h for ML1 and ML2, respectively. The experimental conditions were based on previously optimized synthesis of Cr₂AlB₂ MAB phase [17].

Images of as-deposited MLs were taken using an in-house built photo setup [18]. A high-throughput approach was implemented for investigations of the thin film composition spreads [19]. On each ML, a grid consisting of 342 measurement areas (MAs) with the size of 4.5 mm x 4.5 mm was defined. MLs were characterized by automated methods including XRD for phase analysis and EDX measurements in a scanning electron microscope (SEM) for composition analysis. The composition of selected MAs closest to each target were further investigated with X-ray photoelectron spectroscopy (XPS) in order to determine the composition range of ML2. Subsequently, regions and specific MAs of interest were identified and further investigated by high-resolution (scanning) TEM (HR(S)TEM) and APT.

Needle-shaped APT samples were prepared using a dual beam FIB/SEM system (FEI Helios G4 CX) following a standard approach [20]. To avoid any loss of the Fe-Y-Al-B thin film during the FIB-based milling process, a 160 nm thick Ni protective layer was deposited on the thin film. APT measurements were performed on a LEAP 5000 XR™ (CAMECA Instruments) using laser pulsing at 65 K with 30 pJ laser energy at a pulse repetition rate of 200 kHz and a detection rate of 0.006 atoms per pulse. The APT data was reconstructed and analyzed by IVAS 3.8.12 software.

Fig. 1b shows a photograph of the as-deposited ML2. The sputter target positions relative to the substrate are indicated. Two MAs, corresponding to two different Fe:Y ratios of 4:1 and 2:1 and selected for further cross-sectional TEM analysis, are marked.

The measured composition of the elongated grains corresponds to Fe_1.5AlB_1.1, which deviates from the ideal Fe₂AlB₂ MAB phase stoichiometry. However, in the previously conducted thin film investigations of structurally related Cr₂AlB₂[21] as well as MoAlB [22], broad phase stability ranges in terms of composition have been determined. Cr₂AlB₂ in particular has formed within the composition range of Cr: Al: B = 31–39: 18–29: 34–49 at.%, where Cr_1.3AlB_1.2 was obtained for synthesis in Al excess. This is very close to here measured stoichiometry of Fe_1.5AlB_1.1 formed in MA1. Based on (S)TEM-EDX and APT results, MA1 represents Al-rich synthesis conditions as well.

In contrast to theoretical predictions [12], we have not observed the formation of the quaternary phase (Fe_2/3Y_1/3)₂AlB₂ or incorporation of Y into Fe₂AlB₂ to a lesser degree. Several factors may be responsible for this disagreement. First of all, the exact starting stoichiometry may play a decisive role. Bulk synthesis study on MAX phase (Mo_1-xSc_x)₂AlC with x = 0.33, 0.5, and 0.66 revealed, that the Mo:Sc ratio must be close to the 2:1 for formation of high crystal quality in-plane ordered (Mo_2/3Sc_1/3)₂AlC [24]. Deviation to Mo:Sc = 1:1 resulted in severely defected MAX phase structure and secondary phases. Furthermore, the correct ratio of the two transition metals must be matched to stoichiometric amounts of Al and B. In the quaternary ML with composition gradients for all four elements, these two requirements do not necessary coincide at the same region. As local composition determined by APT reveals, the Fe:Y in MA1 is close to the required 2:1, but Al and B are present in excess. Previous thin film studies of Cr₂AlB₂ and MoAlB indicate, that MAB phases may be stabilized in relatively wide composition ranges with respect to their ideal stoichiometry. In theoretical phase stability calculations, however, crystal structures with exclusively full site occupancy were considered. Consequently, it is not known how deviations in composition and presence of defects affect the phase stability of quaternary MAB phases in particular. Finally, under off-stoichiometric experimental conditions, the set of competing phases is different from the one considered theoretically and may include other secondary phases with even lower formation energies, further suppressing nucleation of the targeted MAB phase.