Exchange-biasing phenomenon can induce an evident unidirectional hysteresis loop shift by spin coupling effect in the ferromagnetic (FM)/antiferromagnetic (AFM) interface which can be applied in magnetoresistance random access memory (MRAM) and recording-head applications. However, magnetic properties are the most important to AFM texturing. In this work, top-configuration exchange-biasing NiFe/IrMn(x Å) systems have been investigated with three different conditions. From the high-resolution cross-sectional transmission electron microscopy (HR X-TEM) and X-ray diffraction results, we conclude that the IrMn (111) texture plays an important role in exchange-biasing field (Hex) and interfacial exchange energy (Jk).HexandJktend to saturate when the IrMn thickness increases. Moreover, the coercivity (Hc) dependence on IrMn thickness is explained based on the coupling or decoupling effect between the spins of the NiFe and IrMn layers near the NiFe/IrMn interface. In this work, the optimal values forHexandJkare 115 Oe and 0.062 erg/cm2, respectively.
Keywords:Exchange biasing; Texture; Coupling or decoupling effect
The exchange-biasing phenomenon using the IrMn basing layer can be applied in magnetoresistance random access memory (MRAM) and recording-head applications extensively because Ir20Mn80 exhibits great characteristics: high interfacial exchange energy (Jk) (or exchange-biasing field (Hex)), low coercivity (Hc), high blocking temperature (TB), and good thermal stability in device performance [1-5]. Moreover, the NiFe/IrMn also can be applied in the high-frequency ferromagnetic resonance (FMR) . In a ferromagnetic (FM)/antiferromagnetic (AFM) system, the texturing in the AFM layer can have an important impact on the magnetic properties of the system. In the past, a NiO/NiFe system with varied AFM NiO thicknesses was studied . In this paper, we will show how the magnetic properties, such as HexHc, and Jk, of the IrMn/NiFe top-configuration system may vary as a function of the IrMn layer thickness (x). It is found that these magnetic properties are closely related to the degree of the (111) texture in the IrMn layer [8-10]. Hex and Jk tend to saturate as x increases beyond 90 Å. Hc is inversely proportional to x, which is caused by the spin coupling or decoupling effect near the NiFe/IrMn interface.
The top-configuration NiFe/IrMn system was made by DC magnetron sputtering onto a glass substrate. The deposition sequences were: glass/Ta(30 Å)/NiFe(50 Å)/IrMn(x Å)/Ta(100 Å), where x = 15, 30, 60, 90, 110, and 150 Å. For this system, we have applied three different conditions during and/or after deposition: (a) the substrate temperature (Ts) was kept at room temperature (RT) only; (b) Ts was at RT with an in-plane external field (h) = 500 Oe during deposition; and (c) Ts = RT, with h during deposition and post-deposition annealing in the field at TA = 250 °C for 1 h, and then field-cooling to RT. The seed Ta layer was used in order to induce a stronger (111) texture in the NiFe or IrMn layer . The cap Ta layer was used to protect the IrMn layer from oxidation. The target compositions of the IrMn and NiFe alloy are 20 at.% Ir, 80 at.% Mn and 80 at.% Ni, 20 at.% Fe, respectively. The typical base chamber pressure was better than 1 × 10−7 Torr, and the Ar working chamber pressure was 5 × 10−3 Torr.
The degree of the (111) texture of the Ir20Mn80layer was characterized by the X-ray diffraction method using a CuKα1line. In order to observe the growth texture and the interfacial morphology directly, we performed high-resolution cross-sectional transmission electron microscopy (HR X-TEM). The exchange-biased magnetic hysteresis loop was measured by a LakeShore Model 7300 vibrating sample magnetometer (VSM).
Results and Discussion
Figure 1shows a typical unidirectional shifted hysteresis loop for the top-configuration NiFe(50 Å)/IrMn(90 Å) sample grown under condition (c). From this figureHexandHcare defined: i.e.,Hex ≡ (H1 + H2)/2 andHc ≡ (H1 − H2)/2. We find thatHex = 112 Oe andHc = 42 Oe in this sample.
Figure 1. The hysteresis loop of a glass/Ta(30 Å)/NiFe(50 Å)/IrMn(90 Å)/Ta(100 Å) sample. This sample was post-annealed atTA = 250 °C andh = 500 Oe for 1 h and then field-cooled to RT. The switching fieldsH1andH2, exchange-biasing field (Hex), and coercivity (Hc) are indicated in the figure
Figure 2shows the X-TEM images of the three NiFe(50 Å)/IrMn(90 Å) samples made under three different conditions, from (a) to (c), respectively. In Fig. 2a, the IrMn (111) crystal plane has grown randomly on the underneath NiFe layer. This indicates that condition (a) is not sufficient to induce the stronger IrMn texture. Under condition (b), as shown in Fig. 2b, the (111) texture arrangement seems better than that in Fig. 2a, but it is still not the best. In contrast, condition (c) can induce an almost perfect IrMn (111) texturing, which follows the underlying NiFe (111) growth texture closely. This clearly indicates that the (111) texturing can cross the NiFe/IrMn interface whenTAis raised to 250 °C. In short, the post-annealing at elevatedTAand the deposition field h are necessary conditions to produce the strongest IrMn (111) texture in the NiFe/IrMn system.
Figure 2. The X-TEM images of glass/Ta(30 Å)/NiFe(50 Å)/IrMn(90 Å)/Ta(100 Å) samples under three different deposition conditions:adeposited at RT only,bdeposited at RT with an external fieldh = 500 Oe, andcthe same film-growth procedure as in (b), post-annealing atTA = 250 °C withhon for 1 h, and then field-cooling to RT
Figure 3shows different degrees of the IrMn (111) texture in the NiFe(50 Å)/IrMn(x) system with X-ray diffraction.Iois the intensity of the IrMn (111) line andIbis the background intensity. According to this figure, there is a higher IrMn (111) texture in conditions (b) or (c). As to condition (a) in Fig. 3, the (111) texture is clearly not well developed yet. These phenomena are consistent with the results from X-TEM images.
Figure 3. The degree of the IrMn (111) texture, as determined from the X-ray diffraction studies, is shown as a function ofxfor glass/Ta(30 Å)/NiFe(50 Å)/IrMn(x Å)/Ta(100 Å).Iois the intensity of the IrMn (111) line andIbis the background intensity
Figure 4 shows Hex plotted as a function of the IrMn thickness (x) for the NiFe(50 Å)/IrMn(x) system under various conditions. As x ≤ 15 Å, there is almost no exchange-bias interaction, since Jk > KAFx, where KAF is the anisotropy energy of IrMn . When x increases from 15 Å to 60 Å, the IrMn pinning action becomes more effective, or Jk = KAFx, which indicates that Hex should increase with increasing x. Moreover, we find that as x ≥ 90 Å under conditions (a)–(c), Hex tends to saturate. The last phenomenon is consistent with X-ray and X-TEM results indicating that the continuation of the (111) perpendicular texture across the NiFe/IrMn interface should stop Hex from decreasing.
Figure 4. IrMn thickness (x) dependence of the exchange-biasing field (Hex) for the glass/Ta(30 Å)/NiFe(50 Å)/IrMn(x Å)/Ta(100 Å) samples under conditions (a) to (c)
According to the well-known theory based on the interfacial exchange-biasing phenomenon,
where Ms is the saturation magnetization of the NiFe layer. Since the ferromagnetic thickness tFM = 50 Å is fixed for these NiFe/IrMn systems, Ms is constant. Therefore, from Eq. 1Jk is proportional to Hex. The x dependence of Jk in Fig. 5 should look similar to that of Hex in Fig. 4. Note that the largest Jk value, about 0.062 erg/cm2, has been realized in this study, as shown in Fig. 5. The value is about half of that found in reference .
Figure 5. IrMn thickness (x) dependence of the interfacial energy (Jk) is shown for the glass/Ta(30 Å)/NiFe(50 Å)/IrMn(x)/Ta(100 Å) systems
The Hc is plotted as a function of x in Fig. 6. In general, Hc increases in the x range from 15 Å to 30 Å (or 60 Å) and decreases in the x range thereafter. According to reference , the Hc behaviors are caused by the spin coupling and decoupling effects at the NiFe/IrMn interface as x increases. As discussed before, when x increases from 30 Å to 60 Å, Hex increases gradually, which implies the NiFe/IrMn coupling drag interaction. In turn, the coupling force between the NiFe and the nearest IrMn spins at the interface is larger than that between neighboring IrMn spins. The external field (H) needs to rotate not only the NiFe spins but also the IrMn spins on top together. As a result, the resistance to domain wall motion is higher, and Hc should increase as x increases from 15 Å to 60 Å. However, as x continues to increase, Hc eventually decreases, due to the decoupling effect between the interfacial NiFe spin and the IrMn spin on top. The reason for the decoupling is that as x continues to increase, Hex is fully developed, and even the lowest-level IrMn spin (at the interface) is strongly pinned by the IrMn spins above. Therefore, when the external field is large enough to switch the NiFe spin at H = Hc, the neighboring IrMn spin does not rotate together anymore. Hence, Hc decreases as x ≥ 60 Å (Fig. 6).
Figure 6. Coercivity (Hc) versus the IrMn thickness (x) for the glass/Ta(30 Å)/NiFe(50 Å)/IrMn (x Å)/Ta(100 Å) systems
In conclusion, under the various conditions (a)–(c) for the top-configuration NiFe/IrMn systems, the magnetic properties, such as HexJk, and Hc, have been investigated. These magnetic properties are closely related to the growth IrMn (111) texturing. From HR X-TEM and X-ray diffraction results, we conclude that the strongest IrMn (111) texture appears in condition (c). Therefore, condition (c) should induce the highest Hex and Jk. Furthermore, the Hc value first increases and then decreases as x increases from 15 Å to 150 Å. This is due to the spin coupling and decoupling drag effects at the NiFe/IrMn interfaces. The optimal Hex and Jk values obtained from this study are 115 Oe and 0.062 erg/cm2, respectively. This Hex value of NiFe/IrMn is larger or equal to the optimal Hex in the NiO/NiFe systems [13,14].
This work was supported by the National Science Council and I-Shou University, under Grant Nos. (NSC97-2112-M214-001-MY3), (ISU97-S-03), and (ISU97-02-20).
J. Magn. Magn. Mater.. 2002, 240:130.
Bibcode number [2002JMMM..240..130Y]Publisher Full Text
J. Magn. Magn. Mater.. 2004, 270:403.
COI number [1:CAS:528:DC%2BD2cXht1altLk%3D]; Bibcode number [2004JMMM..270..403L]Publisher Full Text
Thin Solid Films. 2003, 441:111. Publisher Full Text
J. Appl. Phys.. 2000, 88:975.
Bibcode number [2000JAP....88..975V]Publisher Full Text
J. Appl. Phys.. 2000, 87:4927.
COI number [1:CAS:528:DC%2BD3cXjtFamurc%3D]; Bibcode number [2000JAP....87.4927C]Publisher Full Text
J. Appl. Phys.. 2003, 93:6668.
COI number [1:CAS:528:DC%2BD3sXjslSksb0%3D]; Bibcode number [2003JAP....93.6668A]Publisher Full Text
J. Appl. Phys.. 1996, 79:6389.
COI number [1:CAS:528:DyaK28Xis1Knu7Y%3D]; Bibcode number [1996JAP....79.6389L]Publisher Full Text
J. Appl. Phys.. 2006, 100:13906.
Bibcode number [2006JAP...100A3906W]Publisher Full Text
J. Magn. Magn. Mater.. 1999, 200:552.
COI number [1:CAS:528:DyaK1MXmtVGksrk%3D]Publisher Full Text
J. Appl. Phys.. 2005, 98:113903.
Bibcode number [2005JAP....98k3903M]Publisher Full Text
J. Alloy. Compd.. 2008, 448:59.
COI number [1:CAS:528:DC%2BD2sXht1ylsrjF]Publisher Full Text
Phys. Rev. B. 2003, 68:214420.
Bibcode number [2003PhRvB..68u4420A]Publisher Full Text
J. Magn. Magn. Mater.. 2001, 224:61.
COI number [1:CAS:528:DC%2BD3MXht1Okt7k%3D]; Bibcode number [2001JMMM..224...61Y]Publisher Full Text
J. Magn. Magn. Mater.. 1998, 186:265. Publisher Full Text