Growth and characterisation of ferromagnetic and antiferromagnetic Fe 2 + x V y Al Heusler alloy films

We investigated growth, annealing conditions and magnetic properties of the Heusler alloy Fe 2 + x V y Al by means of x­ray diffraction, magnetic hysteresis and exchange­bias measurements. Ferromagnetic Heusler alloy films were obtained by sputtering Fe 2 VAl and Fe 3 VAl targets and performing post­growth annealing. The characteristic (2 2 0) Heusler alloy peaks were seen in the x­ray diffraction measurements and corresponding ferromagnetic behaviours were observed. In addition, antiferromagnetic Heusler alloy films were deposited by employing Al pegs on Fe 3 VAl sputtering targets. The deposited films had elemental ratios close to the predicted Fe 2.5 V 0.5 Al phase, and a 16 Oe exchange­bias was measured in a Fe 2.3 V 0.7 Al/Co 60 Fe 40 system at 100 K.

There are over 2500 known possible Heusler alloys, hence identifying possible stable AF candidate constitutes a great challenge [7]. For the Fe 2 VAl Heusler alloy, several inconsist encies are observed in the measurements and the predictions of transport and magnetic properties. Nishino et al [12] per formed temperature dependent resistivity and photoemission spectroscopy measurements and associated the results with the response of a nonmagnetic narrow gap semiconductor. A heavyfermionic behaviour with weak magnetic ordering was observed at temperatures as low as 2 K for offstoichiometric compositions. Popiel et al [13] showed that the magnetic prop erties of Fe 2 VAl depend on the crystal structure whereupon A2 structure is FM but the B2 structure is superparamagnetic down to 4.2 K. Venkatesh et al [14] supported this by performing ab initio calculations of L2 1 ordered Fe 2 VAl where several types of disorder were considered for the unit cell. The calculations were then compared to magnetic measurements of D0 3 type Fe 2 VAl which provided further evidence for the magnetic prop erties being dependent on the atomic disorder in the lattice sites. However, the results point towards paramagnetic and weak magnetic ground states. A schematic representation of the most common type of atomic ordering in Heusler alloys is shown in figure 1 where the Heusler alloy has the general formula X 2 YZ.
Singh and Mazin [15] have predicted a metastable AF state of Fe 2 VAl when the compound is ironrich to become Fe 2.5 V 0.5 Al. Their calculations considered a supercell that was created by quadrupling the L2 1 structure along the (1 1 1) axis. The nearest neighbours of each atom were consistent with the B2 disorder allowing for Fe and V atoms to be easily inter changed. The AF ground state was found by tuning the compo sition of Fe 2+x V 1−x Al. A speculation of AF phases existing in Fe 2 VAl was discussed by Feng et al [16] as a result of a lower than expected superparamagnetic response observed. To our knowledge, there is no clear evidence that Fe 2 VAl would have a stable AF phase. It is expected however that combinations of AF and FM phases might form in polycrystalline samples as previously reported for Ni 2 MnAl [17].
A very close relative of Fe 2 VAl, Fe 2 VSi was reported to have a stable AF tetragonal ground state at low temper ature [18][19][20]. Fukatani et al used this material in a thinfilm form deposited on MgO and MgAl 2 O 4 to prove the effects of biaxial strains on the magnetic and transport properties of the Heusler alloy [21]. The results showed an increase of more than 50% of the Néel temperature with a maximum value of 193 K as determined by temperature dependent resistivity measurements.
The key parameter in the AF order is the spacing between planes where the magnetic spins are ordered ferromagn etically in the plane. For Fe 2 VAl, the lattice constant is theor etically reported to be 0.576 nm and the AF order is predicted to appear at Fe 2.5 V 0.5 Al. Hence, it is critical to control the composition and the crystallisation processes.

Experimental procedures
Fe 2 VAl films were deposited on Si/SiO 2 substrates using a PlasmaQuest high targetutilisation sputtering system (HiTUS) with a base pressure of 6 × 10 −8 mTorr. A 13.56 MHz radiofrequency (RF) antenna ionises the process gas (Ar) kept at a constant pressure of 1.86 mTorr. A bias voltage (V B ) applied to the target was varied from −300 V to −900 V in order to control the sputtering rate (0.2-0.7 Å s −1 ) and the grain size of the film. The thickness of the Heusler alloy films was kept constant at 100 nm. A 5 nmthick capping layer of Ta was used in order to prevent oxidation. The films were post annealed in a vacuum furnace at a base pressure lower than 5 × 10 −8 mTorr.
After each annealing step, magnetisation curves were meas ured using an alternating gradient force magnetometer (AGFM, Princeton Measurements Model 2900) at room temper ature (RT). The exchangebias measurements were performed using an ADE Model 10 vector VSM equipped with an open cycle cryostat. The crystalline structures were characterised using xray diffraction (XRD, Rigaku SmartLab) using a standard col limated beam and a Ge(2 2 0) 2bounce monochromator optics.
Furthermore, compositions of the sputtered Fe 2 VAl films were characterised by inductivelycoupled plasma optical emission spectroscopy (ICPOES) and energydispersive xray spectroscopy (EDX). The chemical composition analysis by ICPOES was performed by InterTek Ltd. For the ICPOES measurements, samples consisted of 100 nmthick films were deposited on glass substrates. The samples were then dis solved in aqua regia (HCl: 15 ml, HNO 3 : 5 ml) and diluted to a final volume of 30 ml with demineralised water. The solutions were then analysed by ICPOES while blank glass substrates were also analysed for unexpected contaminants. The EDX analysis was performed using a Thermo Scientific EDX unit attached to FEI Sirion SFEG/JEOL 7800F Prime scanning electron microscopes.

Results and discussion
The asdeposited Fe 2 VAl films were characterised by XRD. As shown in in figure 2, films deposited at −300 V films showed  [7]. Reprinted from [7], Copyright 2006, with permission from Elsevier. a sharp peak at 44.4° which coincided with the Heusler (2 2 0) peak position [16]. This structure disappeared when the bias voltage increased to −600 V and −900 V. While the Heusler (2 0 0) peaks at ~31° are missing for all samples, a weak (4 0 0) Heusler peak at 64° can be seen for the −300 V in a longer range scan in the inset. This suggested the formation of the sought B2 crystalline phase which is predicted to be AF. No (1 1 1) family superlattice peaks corresponding to the fully ordered L2 1 phase were visible. Throughout the XRD scans, the forbidden Si (2 0 0) reflection can be observed at ~33°. Their appearance and changing intensities were thought to be due to slight misalignments of the samples during sample loadings in the XRD setup.
In order to determine the crystallisation conditions for the alloy deposited at −900 V, a number of samples were annealed under vacuum. As shown in figure 3, the alloy crys tallises when annealed at T a = 500 °C for 2 h. The (2 2 0) peak indicates a strong A2 type structure. The absence of the (2 0 0) peak indicated a lack of the higherordered B2 phase.
By performing Scherrer analysis [22], the mean grain sizes of these two sets of films were found to be sim ilar, i.e. 14.8 ± 0.7 nm for those sputtered at −300 V and 14.3 ± 0.8 nm for Fe 2 VAl sputtered at −900 V and post annealed. This was different when compared to our previous report on the crystallisation of the FM Co 2 FeSi Heusler alloy [9]. We conclude that the grain size control during the deposi tion had little influence on the crystallisation process for the Fe 2 VAl films.
As shown in figure 4, the FM behaviour in Fe 2 VAl films sputtered at −300 V and −900 V could be a consequence of formation of A2 Heusler structure as predicted [15]. It is clear that the stoichiometric alloy has a tendency to crystal lise in a FM phase. The AF phase is predicted to form when an offstoichiometric Ferich composition is achieved.     To achieve the desired AF composition of Fe 2.5 V 0.5 Al, a more Ferich Fe 3 VAl target was used. Asdeposited films were found to have a composition of Fe 2.6 VAl 0.4 by ICPOES. From the XRD measurements, the Heusler alloy (2 2 0) peaks are observed for the samples annealed at 400 °C for 5 h and at 500 °C for 3 h (see figure 5). Temperatures between 400 °C and 500 °C were found to be the minimum temperatures for the A2 crystallisation to occur in these films.
The asdeposited Fe 3 VAl samples were measured to have the element ratio of 2.5:1:0.3 using the more Ferich Fe 3 VAl target. In order to increase the Al content to reach the pre dicted 2.5:0.5:1 AF ratio [7], highpurity Al pegs (6 mm in diameter) were placed on top of the Fe 3 VAl target during sput tering. As shown in figure 6(a), it was possible to increase the Al content of the films with increasing numbers of Al pegs. A composition ratio close to 3:1:1 was achieved. With further processes of sputtering with Al pegs, we obtained films with a composition ratio of 2.3:0.7:1, close to the AF composition of Fe 2.5 V 0.5 Al, as observed by EDX method.
Magnetic properties of a 100 nmthick Fe 2.3 V 0.7 Al films were measured as seen in figure 6(b). Compared to the Fe 2.3 V 0.7 Al/Co 60 Fe 40 stacks, the magnetic response of the Fe 2.3 V 0.7 Al layer was very weak. This suggested that the exchangeshifted FM signals were from the CoFe layer itself and its interaction with the nonmagnetic/AF Heusler alloy layer.
The exchangebias stack structure consisted of Si/ SiO 2 substrate/5 nm Cr/30 nm Ag/20 nm Fe 2.3 V 0.7 Al/3 nm Co 60 Fe 40 /2 nm Ta. The Cr is used as a smoothing layer for the substrate, the Ag is a wellknown seed layer for the Heusler structure and has a good lattice match with the alloy [23]. The sample was then fieldannealed at 500 K under −2 T applied field for 90 min, and then cooled to 100 K with the applied field, following the York Protocol [24]. Three M-H hysteresis curves in figure 6(b) show in sequence: (1) at 300 K before the field annealing, (2) at 100 K fieldcooled after the −2 T field annealing at 500 K, then (3) at 300 K after warming up from the 100 K measurement. A larger coercivity, as well as an exchangebias of around 16 Oe was observed at 100 K after the fieldset, indicating the exchangebias coupling at the Fe 2.3 V 0.7 Al/Co 60 Fe 40 interface. Due to the small exchange bias measured, we could not determine the average blocking temperature as similarly reported on Mn 2 VSi [10]. The modest loop shift values and the decreased thermal stability were attributed to the lack of anisotropy associated with the cubic structure. Most cubic Heusler alloys exhibit comparable properties with regards to the pinning strength and thermal stability of the AF configuration of the magnetic moments. In order to increase the thermal stability and pinning strength, a tetragonal distortion of the lattice is required. Further element replacement in the Fe 2 VAl alloy and strain introduction to induce magnetic anisotropy may increase the exchangebias as well as the blocking temperature for RT antiferromagnetism.

Conclusion
A systematic study of structural and magnetic properties was performed on Fe 2+x V y Al layers deposited by sputtering Fe 2 VAl and Fe 3 VAl targets with Al dopant pegs. The FM A2 structure Fe 2 VAl was confirmed by XRD and magnetic mea surements for −300 V and −900 V films after annealing at 500 °C for 2 h. By sputtering Fe 3 VAl target with Al pegs, Fe 2.3 V 0.7 Al films were deposited whose elemental ratio was close to the predicted AF Fe 2.5 V 0.5 Al. A very weak magnetic response of a 100 nmthick Fe 2.3 V 0.7 Al film as well as a 16 Oe exchangebias shift observed in a Fe 2.3 V 0.7 Al/Co 60 Fe 40 stack suggested AF nature of the Fe 2.3 V 0.7 Al films.