• 沒有找到結果。

Magnetic properties of Ni/Co/Pt(111) and

N/A
N/A
Protected

Academic year: 2021

Share "Magnetic properties of Ni/Co/Pt(111) and"

Copied!
28
0
0

加載中.... (立即查看全文)

全文

(1)

Magnetic properties of Ni/Co/Pt(111) and

Co/Ni/Pt(111)

6.1 Magnetic properties of the initial growth

6.1.1 dN i Ni/1 ML Co/Pt(111)

The Kerr signals measured at room temperature of dN i Ni/1 ML Co/Pt(111) are shown in Fig. 6.1(a). The easy axis of the magnetization for 1 ML Co/Pt(111) is out-of-plane, same as that of our previous study [12] Therefore only the polar Kerr hysteresis loop was observed for dN i= 0 ML.. For the pure Ni films deposited on Pt(111), the direction of the easy axis seems to have a cant angle Θ of 41.3with respect to the surface normal. We were surprised that only the polar Kerr signals could be observed for dN i Ni/1 ML Co/Pt(111) when 0 ≤ dN i ≤ 24 ML. This indicates that the easy axis of the magnetization for dN i Ni/Pt(111) changes the direction from the cant to the out-of-plane when 1 ML Co buffer layer is inserted. The polar Kerr signals also enhance when Co buffer layer is inserted. The evolutions of polar Kerr intensity and coercivity versus Ni coverage are shown in Fig. 6.1(b). The Kerr intensity is 0.033 of 1 ML Co/Pt(111) (Kerr intensity magnetization). When 1 ML Ni deposited on 1 ML Co/Pt(111), Kerr intensity increases to 0.046. The Kerr intensity increases very slow when 1 ML≤ dN i ≤ 8 ML. It increases gradually between 8 ML≤ dN i≤ 12 ML, and then it increases slowly to 0.131 when the thickness of Ni is 24 ML. Note that Kerr intensity is about 0.02 in 24 ML Ni/Pt(111)

75

(2)

Figure 6.1: (a) The polar and longitudinal Kerr hysteresis loops of different coverages of Ni deposited on 1 ML Co/Pt(111). (b) The evolutions of polar Kerr intensity and coercivity versus Ni coverage.

(3)

system. This indicates that Kerr intensity has about seven folds after 1 ML Co buffer layer is inserted. It is interesting to understand why the magnetic properties have dramatic changes after inserting Co buffer layer, especially the spin reorientation transition occurs.

The magnetic anisotropy energy can be used to describe the tendency of the alignment of the magnetization [12]. The magnetic anisotropy energy density can be written as E = Kef fsin2Θ if we neglect the high-order terms. The easy axis of the magnetization prefers the out-of-plane direction (Θ = 0) when Kef f > 0. On the other hand, the easy axis of the magnetization prefers the in-plane direction (Θ = 90) when Kef f < 0. The effective magnetic anisotropy Kef f of the Ni/Pt(111) can be written as [12, 100]

Kef f = KvN i+KsN i−P t

dN i +KsN i−vacuum

dN i (6.1)

where Kv is the volume anisotropy, and Ks is the interface anisotropy. In a system of ultrathin films, Kef f is dominated by Ks. Krishnan et al. reported KsN i−P t = 0.17 mJ/m2 and KvN i < 0 for Ni/Pt multilayers at low temperature [14]. The competition between interface anisotropy and volume anisotropy causes the easy axis of the magnetization in the cant direction when 4 ML≤ dN i ≤ 24 ML. The effective anisotropy Kef f of Ni/1 ML Co/Pt(111) is

Kef f = KvN i+KsN i−Co

dN i +KsN i−vacuum

dN i + KsCo−P t

dCo +KsN i−Co

dCo (6.2)

Comparing equation (6.1) and equation (6.2), KsN i−P t changes to KsN i−Co and KsCo−P t when Co buffer layer is inserted between Ni and Pt(111). From the publication data, KsN i−Co = 0.42 mJ/m2 [57], KsCo−P t = 0.57 mJ/m2 [101]. Note that dCo = 1 ML only, Kef f is dominated by the last two terms in equation (6.2). The positive value and larger increase of Kef f after inserting 1 ML Co buffer layer cause the spin reorientation transition. The easy axis of the magnetization changes from the cant direction to the out- of-plane direction as shown in Fig. 6.1. In addition, our previous study has shown that the easy axis of the magnetization is out-of-plane when the thickness of Co on Pt(111) is less than 3.5 ML [12]. Therefore the polar Kerr intensity increases from 0.033 of 1 ML Co/Pt(111) to 0.046 of 1 ML Ni/1 ML Co/Pt(111) is reasonable.

It is worth noticing that the coercivity decreases rapidly when 1 ML≤ dN i ≤ 8 ML, after which it reaches a stable value of about 250 Oe. That means the Ni films of this coverage region decrease the coercivity rather than increasing the Kerr intensity. Alterna- tively, the Ni films increase the Kerr intensity but leaving the coercivity remain the same

(4)

when dN i≥ 8 ML. Our group has been studied the magnetic properties of nonmagnetic Ag overlayer deposited on Co/Pt(111) [12], the coercivity increases with thickness of Ag films. It seems that the influences in coercivity of magnetic and nonmagnetic overlayers are very different, the magnetic one helps the spin reversal and however, the nonmagnetic one prevents the domain wall motion. The influence of Ni content on the coercivity of Ni-Co-Pt alloy has been studied by Zou et al. [102]. They reported that the coercivity of sputtered Co1−xNixPt3 decreases with the increase of Ni content when x≥ 0.25. To our knowledge, the depression of coercivity by Ni overlayer has not been observed by other group.

6.1.2 dCo Co/ 1 ML Ni/Pt(111)

We used MOKE to measure the hysteresis loops of ultrathin Co films deposited on 1 ML Ni/Pt(111) at room temperature. The hysteresis loops measured at room temperature are shown in Fig. 6.2(a). Neither polar Kerr signal nor longitudinal Kerr signal is observed when the thickness of Co film is below 3 ML. Even when the sample is cooled down to 150 K to avoid the thermal fluctuation, the Kerr signals still does not appear. The longitudinal Kerr signal is detected when the thickness of Co is larger than 3 ML. This phenomenon is different from its mirror system dN i Ni/1 ML Co/Pt(111) as discussed in sec. 6.1.1. At room temperature, 1-24 ML Ni/1 ML Co/Pt(111) system exhibits strong perpendicular magnetic anisotropy. M. Seddat et al. [103] reported that there exists a charge transfer from Pt to 3d bands of Ni at the interface, which leads to the loss of ferromagnetism in Ni. The phenomenon of charge transfer also has been observed in Ni/Cu(001) system [104]. Robach et al. [105] proposed that the hybridization of the 5d orbital of Pt with 3d states of Co through the strong spin-orbit coupling in the Pt atoms causes an enhancement of the number of Pt-Co bonds along the surface normal, and thus enhancing the perpendicular anisotropy. In our system, the Ni buffer layer may form a non-magnetic layer between Co and Pt layers as shown in the schematic plot of Fig. 6.2(a). This is one of the possible reasons why the Kerr rotation was too small to be observed between 150 K and 300 K when the thickness of Co film was below 3 ML. The other possible reasons are that the Curie temperature is too low or the coercivity is too large. The minimum temperature and maximum magnetic field in our system are around 150 K and 900 Oe, respectively. According to the report in the system of CoxNi1−xPt [16], the addition of Co to Ni leads to an increase in both the surface anisotropy and Curie

(5)

Figure 6.2: The Kerr hysteresis loops of ultrathin Co films deposited on 1 ML Ni/Pt(111).

All the samples were prepared at room temperature and the MOKE measurements were performed at (a) 300 K and (b) 450 K. The schematic plots indicate the surface compo- sition of the samples.

temperature. The inserted Ni layer in our system may dilute the Co magnetic moment and decrease the Curie temperature to lower than 150 K. The coercivity increases when intercalating thin Ni layers on Co/Pt multilayers have been reported by R. Krishnan et al [106]. The coercivity may increase to a strength higher than 900 Oe in our system.

A sequence of samples of dCo Co /1 ML Ni/Pt(111) (dCo = 1, 2, 3, 4, and 5) were prepared at room temperature, and then annealed at 450 K for 20 minutes. Fig. 6.2(b)

(6)

shows the polar and longitudinal hysteresis loops of these annealed samples that were measured at 450 K. Only polar Kerr hysteresis loop is observed when dCo = 1 ML. The longitudinal hysteresis loop can be observed after dCo > 3 ML and it becomes more square as the Co coverage increases. However, both Kerr signals disappeared when 2 ML≤ dCo 3 ML. The most possible factor for causing the polar Kerr intensity of dCo = 1 ML at 450 K is the diffusion of Co. In 1 ML Co/ 1 ML Ni/Pt(111), the starting temperature of Co and Ni atoms in the intermixing process is 420 K. When the Co atoms intermix with the Ni atoms, some Co atoms touch and couple with the Pt atoms. The Pt-Co bonds enhance perpendicular magnetic anisotropy. But when 2 ML ≤ dCo ≤ 3 ML, this intermixing process does not occur unless the annealing temperature is higher than 450 K. The initial growth mode of Co ultrathin films deposited on 1 ML Ni/Pt(111) is layer-by-layer when dCo ≤ 2 ML, and it turns into a 3-D island growth as shown in Fig 5.1(b). The surface morphology of the island structure causes the samples to favor in-plane magnetic anisotropy when dCo ≥ 3 ML [105]. In addition, the volume anisotropy KV of Co is negative; it favors the in-plane magnetization when the Co films become thicker. Therefore, the easy axis of the magnetization changes from the out-of-plane to the in-plane direction when the thickness of Co is 4 ML and 5 ML as shown in Fig. 6.2(b).

Figure 6.3: The evolutions of polar Kerr intensities versus sample temperature during the first heating process.

(7)

6.2 Comparative study in magnetic properties between 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111)

Fig. 6.3 shows the evolutions of polar Kerr intensities for 1 ML Ni/1 ML Co/Pt(111) (curve (a)) and 1 ML Co/1 ML Ni/Pt(111) (curve (b)) during the first heating processes.

The polar Kerr intensity of curve (a) decreases with the increase of sample temperature.

Since the thermal fluctuation will decrease the saturated magnetization, the depression of polar Kerr intensity as shown in curve (a) is reasonable. The polar Kerr intensity is too small for us to measure when 300 K < Ts< 420 K for the 1 ML Co/1 Ml Ni/Pt(111) as shown in curve (b). It increases when Ts > 420 K and enhances dramatically when Ts > 580 K. By comparing with the AES studies as shown in Fig. 5.2, this anomalous behavior of curve (b) can be understood. The intermixing process of Ni and Co atoms

Figure 6.4: The evolutions of polar Kerr intensity versus sample temperature after an- nealing at each Tafor 20 minutes are plotted in (a) 1 ML Ni/1 ML Co/Pt(111) and (b) 1 ML Co/1 ML Ni/Pt(111). The Curie temperature as a function of Ta is plotted in frame (c).

(8)

causes the increase of polar Kerr intensity when Ts > 420 K, while the process of the Ni-Co intermixing layer diffusing into the Pt substrate enhances the polar Kerr intensity when Ts> 580 K. That means the most important mechanism of the observation of polar Kerr intensity is whether or not the Co atoms contact with the Pt atoms. The polar Kerr intensities of both curve (a) and (b) decrease rapidly when the sample temperature is higher than 660 K. Although the perpendicular magnetic anisotropy is strong when the Co atoms coupled with Pt atoms, the thermal energy can overcome the anisotropy energy at a certain high temperature. This causes the polar Kerr intensity decreases rapidly near the Curie temperature. The Curie temperatures of the first heating processes are 710 K and 730 K for 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111).

Curie temperature is the temperature that separates the magnetic ordered state (Ts <

TC), where the internal field dominates the thermal effect, from the magnetic disordered state (Ts > TC), where the thermal disorder regions [107]. For a useful magnetic ultra- high density recording media, the Curie temperature must be approximately at the range around 200C [2, 108]. Therefore, Curie temperature is an important topic of application in ultra-high density magnetic recording materials. We measured the Kerr intensity at different sample temperature after the sample was annealing at high temperature for 20 minutes. The results of 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111) are shown in Fig. 6.4(a) and (b), respectively. It is easy to see that the Curie temperatures of both Fig. 6.4(a) and (b) decrease with the increase of annealing temperature. The evolution of Curie temperature versus annealing temperature is plotted in Fig. 6.4(c).

The Curie temperature of both systems has linear relationship with annealing temperature except a kink at 780 K. When the annealing temperature is between 750 K and 780 K, the decreasing rates of Curie temperature are 1.9 and 3.0 for 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111), respectively. The decreasing rate for 1 ML Ni/1 ML Co/

Pt(111) is smaller than that of 1 ML Co/1 ML Ni/Pt(111) at this temperature range.

When the annealing temperature is higher than 780 K, the decreasing rates are 4.0 and 4.1 for 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111), respectively. The difference between the decreasing rates is very small when Ta> 780K.

As Ts approaches to TC, the saturated magnetization goes to zero following a power law as shown in eq. (6.3), where β is called critical point exponent.

M s(Ts) = M s(0)(1Ts

Tc)β (6.3)

(9)

Figure 6.5: The evolutions of critical exponent, β value, versus annealing temperature for 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111).

Here, M s(Ts) and M s(0) are the saturated magnetization at sample temperature Ts and 0 K. β is called critical point exponent and is dependent on the magnetic structure, especially depends on the dimension of the magnetic structure. The β value can be obtained from linear fitting the curve of n(M s(Ts)) versus n(1TTsc), β is the slope of the fitting results. Fig. 6.5 shows the evolutions of β value versus annealing temperature of 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111). The β value of 1 ML Ni/1 ML Co/Pt(111) is around 0.198 when the annealing temperature is between 730 K and 780 K, it has a sudden crossover from 0.198 to 0.345 when the annealing temperature is between 780 K and 800 K. The tendency of the β value of 1 ML Co/1 ML Ni/Pt(111) is same as that of 1 ML Ni/1 ML Co/Pt(111). The β value of 1 ML Co/1 ML Ni/Pt(111) is around 0.226 when the annealing temperature is between 750 K and 780 K, it increases significantly from 0.226 to 0.363 when the annealing temperature increases from 780 K to 800 K. The β values of 2D Ising model [45], 2D XY model [47] and 3D Heisenberg spin-lattice model [48] are, respectively, 0.125, 0.23 and 0.365. The values of β measured in 2D regime are often intermediate between the 2D Ising value of 0.125 and the 3D value of 13 [109]. The β values of 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111) change approximately from that of 2D XY to that of 3D Heisenberg, this implies that the surface magnetic structure changes from a 2D-like phase to a 3D-like phase. This

(10)

result is reasonable because the surface magnetic film becomes thicker when the Co and Ni atoms diffuse deeper after higher temperature annealing. Therefore, the thickness d of the ferromagnetic films is not very less than the sizeof the largest lateral fluctuations.

Since d , all films fall into 2-D universality class [51], it is reasonable the film with large thickness becomes more 3D-like and falls into 3-D universality class. Huang et al. [24] have been studied the relation between β and the thickness of ferromagnetic films deposited on Cu(100) and Cu(111) in 1994. In their report, the β value presents a thickness independence in the ultrathin region, but a sudden crossover form 2D to 3D behavior occurs at a certain thickness of Ni or Co films. The actual crossover thickness value is affected by the microstructure of the films. Since the magnetic properties are strongly related to the surface properties, the increase of β value can be an evidence of the change of surface structure. Therefore, the change of surface magnetic structure from a 2D-like to a 3D-like may be resulted from the change of surface alloy from NixCo1−xPt to NixCo1−xPt3.

The Curie temperature of Co1−xPtx ordered alloy has been studied by Kashyap et al.

[84]. They indicated that the Curie temperatures of Co3Pt, CoPt, and CoPt3 ordered alloys are 1000 K, 710 K, and 290 K, respectively. J. Crangle and W.R. Scott [90] have been studied the Curie temperatures of PdFe, PtFe, PtCo, and PdNi alloys in 1965. They

Figure 6.6: Curie temperature as a function of Pt concentration for 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111).

(11)

indicated that the Curie points rise rapidly with solute content (Fe%, Co%, and Ni%) of the dilute ferromagnetic alloys. In other words, the Curie points decrease rapidly with the increase of solvent content (Pd% and Pt%). According to the report of J. Crangle and W.R. Scott, the Curie temperatures of Co5.17Pt94.83, Co10.2Pt89.8, and Co15.2Pt84.8 are 104 K, 218 K, and 315 K, respectively. The Curie temperature decreases with the increase of Pt content, it decreases more rapidly in the CoPt3 region. From the study of subsurface concentration in sec. 5.4, we know that the Pt concentration is dependent on the annealing temperature in the subsurface region. We reach a Pt-rich surface structure after high temperature annealing, thus, Pt is the solvent for this moment. We plot the Curie temperatures of 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111) as a function of shifted Pt concentration in Fig. 6.6. The Curie temperatures of both systems decrease with the increase of Pt%, this result is very similar to those of the reports we introduced above [84, 90]. This trend of Curie temperature becomes more obvious when Pt% is higher than 66%. The decreasing rate of the Pt% > 66% region is higher than that of the Pt% < 66%. Comparing the result we obtained in sec. 5.4, we can conclude that the Curie temperature decreases rapidly after the surface alloy changes form NixCo1−xPt to NixCo1−xPt3.

The Curie temperatures of both 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111) systems are lower than those of Co1−xPtxwith x being kept at the same value.

The addition Ni atoms seems to low the Curie temperature of 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111) systems. The depression of Curie temperature has been reported by other groups [24, 110]. In theoretically, S. Basu and K. Ghatak [111] have been predicted the depression of Curie temperature of (A1−xBx)1−yCy alloy in which A and B are magnetic and C is nonmagnetic. For a given set of positive JAA and JBB (Jij is the exchange integral between atoms i and j ), they inferred that the negative exchange interaction JAB may cause a depression of Curie temperature between A-C and B-C alloys. The Curie temperatures of bulk Co and Ni are respectively around 1388 K and 631 K [83], the exchange integral JCo−Co is reasonable higher than the exchange integral JN i−N i. From the AES studies in sec. 5.2, we know that the attractive force of Ni-Co is higher than those of Ni-Ni and Co-Co. Hence, we can predict that the Co and Ni atoms prefer bonding together during the diffusion process. Thus, increasing the Ni content will decrease the number of Co-Co exchange coupling of the system. From this point of view, the Curie temperature should be depressed. It is interesting that the Curie temperature

(12)

Figure 6.7: Ni% versus Co% plots of 1 ML Ni/1 ML Co/Pt(111) (”◦ ”) and 1 ML Co/1 ML Ni/Pt(111) (”• ”).

of 1 ML Ni/1 ML Co/Pt(111) is always higher than that of 1 ML Co/1 ML Ni/Pt(111) in Fig. 6.6 under the condition of fixing Pt concentration. Since the LEED patterns of 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111) can not be distinguished, we believe that this phenomenon in resulted from the difference of surface concentrations of Ni and Co between 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111).

Fig. 6.7 shows the relations between Ni% and Co% of these two systems. The Ni % has a linear relationship with the Co% for both 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111) systems. The ratios of Ni% to Co% are 0.43± 0.01 and 0.34 ± 0.01, that means the Ni% of 1 ML Ni/1 ML Co/Pt(111) is higher than that of 1 ML Co/1 ML Ni/Pt(111) in the subsurface region. Zhou et al. have been studied the magnetic prop- erties of Co1−xNixPt3 [87]. They reported that the Curie temperature decreases from 550 K to 313 K with the Ni content x increases from 0 to 0.66. We can estimate the concentrations of Ni and Co with a 75% concentration of Pt from Fig. 6.7. The estimated composition are Co0.69Ni0.31Pt3 for 1 ML Ni/1 ML Co/Pt(111) and Co0.75Ni0.25Pt3 for 1 ML Co/1 ML Ni/Pt(111). Furthermore, we can obtain the approximate Curie temper- atures from Fig. 6.6. From our estimation, the Curie temperature are 360 K and 270 K for Co0.69Ni0.31Pt3 and Co0.75Ni0.25Pt3. Through this estimation, the Curie temperature is higher when the Ni content is higher in the subsurface region. This is a very different result to that of the report of Zhou et al. [87]. The differences are as the followings:

1. According to the report of Zhou et al., the Curie temperatures of Co0.67Ni0.33Pt3

and Co0.75Ni0.25Pt3 are 503 K and 515 K, respectively. Our estimated results is lower than those of Zhou et al. for around 200 K

(13)

2. In this mirror systems, after high temperature annealing, TC is higher when the concentration of Ni is higher in the subsurface region.

The first difference in TC can be understood by the phenomenon of the depression of TC with increasing the ordering of CoPt3 alloy. Menzinger et al. have been reported that the Curie temperature of CoPt3 alloy decreases with the increase of ordering, Tc is 320 K for a CoPt3 alloy which its measured order parameter is higher than 0.9 [88]. From our LEED study in sec. 5.3, we can observe sharp LEED patterns at a Pt concentration of 75

%. This can be an evidence of the formation of ordered NixCo1−xPt3 alloy. Hence, the Curie temperature of our system is lower than that of the report of Zhou et al. is because of the ordering of the NixCo1−xPt3 alloy. In other words, the NixCo1−xPt3 alloy of our systems is more ordered than that of Zhou’s system.

Several groups have been reported that TC decreases with the increasing the Ni content in Ni-Co-Pt alloy [87, 24, 110]. Thus, it is very interesting that the TC is higher when the concentration of Ni is higher in the subsurface region. From Fig. 6.7, we know that the Ni% of 1 ML Ni/1 ML Co/Pt(111) is higher than that of 1 ML Co/1 ML Ni/Pt(111) in the

”subsurface region”. Alternatively, this causes the Ni% of 1 ML Ni/1 ML Co/Pt(111) is smaller than that of 1 ML Co/1 ML Ni/Pt(111) in the deeper region below the ”subsurface region”, for convenience, we call the region as ”bulk region” of the alloy. Owing to the mean free paths of Pt 237 eV, Ni 848 eV, and Co 656 eV which were used to calculate the subsurface composition are around 2-4 ML of the Ni-Co-Pt alloy, the thickness of the

”subsurface region” may be approximately 2-4 ML. The schematic plot of the ”subsurface region” and ”bulk region” is shown in Fig. 6.8. Since the penetration depth of the laser light is large enough for the light to penetrate the entire film, the Kerr signal comes from the deeper films must be taking into consideration. Therefore, we conjecture that the magnetic contribution comes from two parts: the first comes from the ”subsurface region” of Ni-Co-Pt alloy and the second comes from the ”bulk region” of Ni-Co-Pt

           





            

           





            

           





            

Figure 6.8: The schematic plot of the ”subsurface region” and ”bulk region”.

(14)

alloy. Comparing our results with those of Zhou et al. [87], we conclude that the Curie temperature is dominated by the competition between the ”subsurface region” and the

”bulk region” of Ni-Co-Pt alloy. The Curie temperature is higher when the ratio of Ni%

to Co% of the ”bulk region” is higher than that of the ”subsurface region”. This result is very important to the magnetic phenomena which we are going to discuss in the next section, especially in sec. 6.4 in which the spin reorientation transition is induced by the annealing process.

Figure 6.9: Frame (a) is the Kerr response hysteresis loops of 3 ML Co/1 ML Ni/Pt(111), which was measured at room temperature after annealed at each Ta for 20 minutes. Frame (b) is the evolution of polar and longitudinal remanence Kerr intensities versus Ta.

(15)

6.3 Comparative study in magnetic properties between dCo

Co/1 ML Ni/Pt(111) and dN i Ni/1 ML Co/Pt(111)

We then investigated the evolutions of magnetic properties of 3 ML Co/1 ML Ni/Pt(111) after the sample was annealed at different annealing temperatures Ta. The polar and longitudinal hysteresis loops are shown in Fig. 6.9(a). All the loops were measured at room temperature after annealing at each Ta for 20 minutes. No Kerr signal is observed until the sample is annealed at temperatures higher than 625 K. When the annealing temperature is greater than 625 K, the polar hysteresis loop appears gradually. The polar hysteresis loops become square when 750 K ≤ Ta ≤ 800 K. No longitudinal Kerr signal can be observed in the annealing process. These results indicate that the easy axis of the magnetization of 3 ML Co/1ML Ni/Pt(111) is in the out-of-plane direction.

However when the annealing temperature is at 850 K, the polar hysteresis loop is no longer saturated. The remanence Kerr intensity versus annealing temperature is shown in Fig. 6.9(b). The polar Kerr intensity increases slightly when the annealing temperature is between 625 and 700 K. It increases significantly when the annealing temperature is between 700 K and 800 K. For further annealing at temperatures higher than 800 K, the polar Kerr intensity decreases. By comparing Fig. 6.9 with Fig. 5.3(b), the enhancement of perpendicular magnetic anisotropy can be associated with the diffusion process of Co atoms. Co atoms start to intermix with Ni atoms at 610 K as shown in Fig. 5.3(b), but the Kerr signal starts to increase at 625 K as shown in Fig. 6.9. This is likely because only a few Co atoms intermix with Ni in the beginning of the intermixing process; hence the Kerr signal is too weak to be detected. More and more Co atoms intermix with Ni atoms when Ta increases from 625 K to 700 K, the number of Co-Pt bonds is high enough for detection. Therefore, the out-of-plane magnetization starts to increase at 625 K. However, the Kerr signal is still small unless the annealing temperature is higher than 700 K, which is the temperature of the formation of Co-Ni-Pt surface alloy as shown in Fig. 5.3(b). Further comparing the polar hysteresis loops at different Ta in Fig.6.9(a), the coercivity increases with the annealing temperature; the coercivity measured at room temperature after annealing at 750 K is 400 Oe, and it is greater than 900 Oe after annealing at 850 K. The same phenomenon has been found in 7 ML Co/Pt(111) [112]

and Pt(15 ML)/Co(9 ML)/Pt(4 nm) trilayers [113]. Inase et al. [114] have studied the influence of Pt concentration in the coercivity of CoNiPt thin films. They indicated that

(16)

the coercivity increases with increasing Pt concentration when fixing the concentration ratio of Co to Ni. In our system, the ratio is nearly the same after annealing at 800 K and 850 K (ratio∼ 5), but the concentration of Pt increased from 75% to 88%. The more concentration of Pt is, the higher coercivity of the Co-Ni-Pt alloy will be. Our result is consistent with the report of Inase [114]. Compare those systems with ours, the Pt atoms may act as domain wall pinning sites. It is valuable to note that the coercivity of the Ni-Co-Pt system can be adjusted by changing the annealing temperature, due to the variety in concentrations of the alloy formation.

It is interesting that the remanence Kerr intensity measured at room temperature

Figure 6.10: (a) The polar Kerr hysteresis loops of 3 ML Co/1 ML Ni/Pt(111) versus sample temperature. The sample had been pre-annealed at 850 K for 20 minutes. (b) The polar Kerr intensity θK and coercivity HC versus sample temperature of Ta = 850 K.

(17)

decreases rapidly when the annealing temperature is greater than 800 K in Fig. 6.9(b).

Two possible reasons for this reduction in the Kerr intensity: the concentrations of Ni and Co are too dilute, or the coercivity is too large to be measured in our laboratory at room temperature. To determine the true physical origin, we measured the polar hysteresis loops at different sample temperature (Ts) for 3 ML Co/1 ML Ni/Pt(111) after it was annealed at 850 K for 20 minutes. The result is shown in Fig. 6.10(a). The polar Kerr signal can be detected when Ts ≤ 700 K. The hysteresis loop becomes square when the sample temperature is between 440 K and 480 K. The polar hysteresis loops are not saturated when Ts ≤ 440 K, and no Kerr rotation can be observed after Ts < 300 K under a maximum magnetic field of 900 Oe. The curve of the remanence Kerr rotation versus Ts is shown in Fig. 6.9(b). It increases when Ts decreases from 700 K to 440 K. It is interesting that when the concentration of Co dilutes to about 10 % in the subsurface region, the out-of-plane magnetization is still strong. J. Kim et al. [112] reported that the polar Kerr signal continues to increase as Co atoms form a diluted alloy with Pt in a concentration as low as probably a few percent. The dependence of coercivity as a function of Ts for 3 ML Co/1 ML Ni/Pt(111) is shown in Fig.6.10(b) after annealing at 850 K. Since the hysteresis curves are not saturated when the temperature is lower than 440 K, we only plot the curve for Ts≥ 440 K. Hcis lower at higher temperatures, because the thermal energy helps to reverse the magnetic domains. These results indicate that the decrease of the Kerr signal after the high temperature annealing shown in Fig. 6.9(b) is due to the high coercivity. The coercivity increases with the decrease of the sample temperature as shown in Fig. 6.10(b). It is greater than the maximum magnetic field that we can apply (900 Oe) when Ts ≤ 300 K for 3 ML Co/1 ML Ni/Pt(111) after annealing at 850 K. The magnetic behavior of 2 ML Co/1 ML Ni/Pt(111) is almost same as that of 3 ML Co/1 ML Ni/Pt(111). But for the systems, 4 ML Co/1 ML Ni/Pt(111) and 5 ML Co/1 ML Ni/Pt(111), only longitudinal Kerr intensity can be observed before and after annealing (not shown here). Since the volume anisotropy of Co favors an in-plane easy axis, the contribution of the volume anisotropy may overcome that of the surface anisotropy when the thickness of Co film is large enough.

Fig. 6.11(a) shows the temperature dependence of polar saturated Kerr intensities of dCoCo/1 ML Ni/Pt(111) (dCo= 1, 2, and 3 ML) after annealing at 750 K for 20 minutes.

There is no longitudinal Kerr intensity has been observed before and after annealing in this system when dCo≤ 3ML. The saturated polar Kerr intensities of the three curves in

(18)

Fig. 6.11(a) decrease with the increase of sample temperature. The Curie temperatures of these samples are very close after annealing at 750 K, they are 645 K, 640 K and 650 K for dCo = 1, 2, and 3 ML, respectively. At room temperature, the polar Kerr intensities are 0.05, 0.075, and 0.11 for dCo= 1, 2, and 3 ML, respectively. Since the depth of Ni-Co

Figure 6.11: The temperature dependence of polar Kerr intensities of dCo Co/1 ML Ni/Pt(111) (dCo= 1, 2, and 3 ML) after annealing at (a) 750 K and (b) for 20 minutes.

Frame (c) is the coverage dependence of Curie temperature.

(19)

intermixing layer diffusing into the Pt substrate for low coverage is greater than that of high coverage, it is reasonable that the enhancement of PMA after annealing at 750 K is more large for 1 ML Co/1 ML Ni/Pt(111). From Fig. 6.10, we know that, due to a large coercivity after 850 K annealing, the magnetization of 3 ML Co/1 ML Ni/Pt(111) is hard to be saturate. Therefore, in Fig. 6.11(b), we only show the temperature dependence of remanence polar Kerr intensities after annealing at 850 K for 20 minutes. The deeper the ferromagnetic atoms diffusing into the Pt substrate, the more obviously the difference in Curie temperature become. The Curie temperatures of these samples after annealing at 850 K are 260 K, 500 K and 700 K for dCo= 1, 2, and 3 ML, respectively. In Fig. 6.11(b), the remanence Kerr intensities of both dCo = 2 ML and 3 ML decrease at certain low temperatures. This is because the coercivity is too high to exceed the maximum field that we can applied. Fig. 6.11(c) shows the Curie temperature versus Co coverage after annealing at 750 K and 850 K. There exists a significant decrease of Curie temperature with Co coverage after annealing at 850 K for 20 minutes. Since the temperature of the structure changing form Co1−xNixPt to Co1−xNixPt3 is higher for thicker film, we believe that the rapid decrease of Curie temperature with Co coverage after 850 K annealing is mainly resulted from the change of structure. There is another possibility should be considered. From the comparative study between 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111), we know that the Curie temperature is also related to the ratio of N% to Co% in the bulk region of ferromagnetic alloy. The Ni atoms are more sufficient for separating the Co atoms from the Pt atoms and diluting the Co magnetic dipole momentum for lower coverage after high temperature annealing; this causes the Curie temperature decreases rapidly with Co coverage.

Fig. 6.12 shows the temperature dependence of polar Kerr intensities for dN iNi/1 ML Co/Pt(111) with dN i= 1, 2, 12, and 24 ML. In order to investigate the contributions of Ni and Co in the surface and bulk regions, we also plot the temperature dependence of polar Kerr intensity for 2 ML Co/1 ML Ni/Pt(111) in Fig. 6.12. We have been discussed the magnetic properties of the initial growths of dN i Ni/1 ML Co/Pt(111) in sec. 6.1.1.

The easy axis of the magnetization is in the out-of-plane direction. Even the thickness of Ni is up to 24 ML, the direction still does not change. The data shows in Fig. 6.12(a) were measured during the first heating processes of the as-deposited systems. The intensities of these systems decrease with the increase of sample temperature, except for the 2 ML Co/1 ML Ni/Pt(111) system which presents a non-magnetic-like behavior at room temperature

(20)

before annealing.

Fig. 6.12(b) shows the measured data after annealing at 750 K for 20 minutes. Owing to the temperatures of the magnetization decreasing to zero in the first heating process of 12 ML Ni/1 ML Co/Pt(111) and 24 ML Ni/1 ML Co/Pt(111) are higher than 750 K,

Figure 6.12: The temperature dependence of polar Kerr intensities of 2 ML Co/1 ML Ni/Pt(111) and dN i Ni/1 ML Co/Pt(111) with dN i = 1, 2, 12, and 24 ML. (a) First heating process; after annealing at (b) 750 K and (c) 850 K for 20 minutes.

(21)

we only plot the data of 1 ML Ni/1 ML Co/Pt(111), 2 ML Ni/1 ML Co/Pt(111), and 2 ML Co/1 ML Ni/Pt(111) in Fig. 6.12(b). The Curie temperatures are 680 K, 720 K, and 640 K for 1 ML Ni/1 ML Co/Pt(111), 2 ML Ni/1 ML Co/Pt(111), and 2 ML Co/1 ML Ni/Pt(111), respectively. The thickness of the ferromagnetic films of 2 ML Ni/1 ML Co/Pt(111) is thicker than that of 1 ML Ni/1 ML Co/Pt(111), it is reasonable that the temperature of the phase transition (NixCo1−xPt → NixCo1−xPt3) of 2 ML Ni/1 ML Ni/Pt(111) is higher than that of 1 ML Ni/1 ML Co/Pt(111). Thus, this cause the Curie temperature of 2 ML Ni/1 ML Co/Pt(111) is higher than 1 ML Ni/1 ML Co/Pt(111).

But this reason can not explain the Curie temperature of 2 ML Co/1 ML Ni/Pt(111).

Although the thickness of the ferromagnetic films of 2 ML Co/1 ML Ni/Pt(111) is thicker than that of 1 ML Ni/1 ML Co/Pt(111), the Curie temperature of 2 ML Co/1 ML Ni/Pt(111) is lower than that of 1 ML Ni/1 ML Co/Pt(111). From the AES studies in sec. 5.2, we know that the Ni-Co intermixing layer diffusing into the Pt substrate is nearly at the same temperature for 2 ML Ni/1 ML Co/Pt(111) and 2 ML Co/1 ML Ni/Pt(111) (different in temperature∼ 5 K). That means the difference in temperatures of the phase transition from NixCo1−xPt to NixCo1−xPt3 between 2 ML Ni/2 ML Co/Pt(111) and 2 ML Co/2 ML Ni/Pt(111) is very small. Therefore, the thickness of the ferromagnetic film and the temperature of phase transition are not the principle reasons of this phenomenon.

Subsequently, the fives samples were heated up to 850 K and annealed at 850 K for 20 minutes. The results are shown in Fig. 6.12(c). Neither polar nor the longitudinal Kerr signals can be observed for 12 ML Ni/1 ML Co/Pt(111) and 24 ML Ni/1 ML Co/Pt(111).

This can be occurred if the Co and Ni atoms continue to intermix with each other at 850 K, this means that the Ni atoms continue to dilute the magnetic moments of Co atoms and the Curie temperature is below the lowest temperature that we can apply. It is interesting that the trend of the Curie temperatures of 1 ML Ni/1 ML Co/Pt(111), 2 ML Ni/1 ML Co/Pt(111), and 2 ML Co/1 ML Ni/Pt(111) is inverse to that of after 750 K annealing. The Curie temperatures are 360 K, 260 K, and 480 K for for 1 ML Ni/1 ML Co/Pt(111), 2 ML Ni/1 ML Co/Pt(111), and 2 ML Co/1 ML Ni/Pt(111), respectively.

Note that the Ni content of 2 ML Ni/1 ML Co/Pt(111) is always higher than that of 2 ML Co/1 ML Ni/Pt(111) before and after annealing. From this point of view, we predict that the rapid depression of Curie temperature for 2 ML Ni/1 ML Co/Pt(111) after 850 K annealing is because of the rapid increase of Ni% to Co% ratio in the Ni-Co-Pt alloy.

That means the Ni atoms which do not form Ni-Co-Pt alloy does not affect the value of

(22)

the Curie temperature too much. The Curie temperature of 2 ML Ni/1 ML Co/Pt(111) depresses rapidly when more and more Ni atoms diffuse into the Pt substrate to form Ni-Co-Pt alloy. Comparing Fig. 6.12(b) and (c), we can reach the same result developed in sec. 6.2. When the Ni and Co form alloy with Pt, the Curie temperature is dominated not only by the surface structure and film thickness, but also by the competition of the ratio of Ni% to Co% between ”subsurface region” and ”bulk region” of the Ni-Co-Pt alloy.

When the Ni and Co form alloy with Pt, the Curie temperature is higher when the ratio of Ni% to Co% of the ”bulk region” is higher than that of the ”subsurface region”.

6.4 Annealing induced spin reorientation transition (SRT)

From the above discussions, we understand that the phase transition, film thickness, and the ratio of Ni% to Co% in the subsurface and bulk regions of Ni-Co-Pt alloy affect the depression of Curie temperature. In the calculation of Pt concentration for 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111) The question arising now is what difference does it make if the surface composition becomes less homogeneous? In order to investigate this problem, we prepared two mirror systems of 2 ML Co/2 ML Ni/Pt(111) and 2 ML Ni/2 ML Co/Pt(111). The ferromagnetic films of theses samples are thicker than those of 1 ML Co/1 ML Ni/Pt(111) and 1 ML Ni/1 ML Co/Pt(111), thus the difference in the ratio of Ni% to Co% between the subsurface region and bulk region becomes more obvious than those of 1 ML Co/1 ML Ni/Pt(111) and 1 ML Ni/1 ML Co/Pt(111). This helps us to investigate the influence of the ratio of Ni% to Co% of the subsurface region.

Fig. 6.13(a) shows the polar and longitudinal hysteresis loops of 2 ML Ni/2 ML Co/Pt(111) after annealing at each Ta for 20 minutes. All the loops were measured at room temperature. Only polar Kerr signal can be observed unless annealing the sam- ple at a temperature higher than 675 K. When the annealing temperature is 675 K, both the polar and longitudinal Kerr signals can be observed, after which only the longitudinal can be detected. The easy axis of the magnetization is initially in the out-of-plane direc- tion, it changes from the out-of plane to the in-plane after annealing at a temperature higher than 675 K. That means a spin reorientation transition occurred by the annealing.

The annealing temperature dependence of Kerr intensities is shown in Fig. 6.13(b). The polar Kerr intensity increases a little when the annealing temperature increases from 300 K to 525 K, after which it starts to decrease and approaches a value of zero when the

(23)

Figure 6.13: (a) Polar and longitudinal hysteresis loops of 2 ML Ni/2 ML Co/Pt(111).

Each curve was measured at room temperature after annealing at Ta for 20 minutes. (b) The annealing temperature dependences of polar and longitudinal Kerr intensities. (c) The evolutions of Ni 848 eV, Co 656 eV, and Pt 237 eV AES signals versus annealing temperature.

annealing temperature is higher than 675 K. The longitudinal Kerr intensity stays zero when 300 K ≤ Ta ≤ 600 K, after which it starts to increase except for a decrease at a temperature higher than 850 K. The Auger intensities of Pt 237 eV, Ni 848 eV and Co 656 eV were measured to monitor the diffusion process of 2 ML Ni/2 ML Co/Pt(111) and the result is shown in Fig. 6.13(c). The Ni 848 eV AES intensity starts to decrease at 525 K while the Co 656 eV AES intensity starts to increase at the same annealing temperature.

No significant increase of Pt 237 eV AES intensity is observed unless Ta≥ 800 K in which the Co 656 eV AES intensity starts to depress. From the AES studies in sec. 5.2, we know

(24)

that the intensity of Pt 237 eV will not change when the intermixing process of Ni and Co atoms occurred. Therefore, the Ni and Co atoms of 2 ML Ni/2 ML Co/Pt(111) inter- mix to each other when 300 K≤ Ta ≤ 525 K. Comparing Fig. 6.13(b) with (c), we know that the Ni-Co intermixing process decreases the perpendicular magnetic anisotropy when 525 K≤ Ta≤ 750 K. The Ni-Co intermixing layer starts to diffuse into the Pt substrate when Ta ≥ 800 K. The interaction between neighboring Co and Pt decreases and the interactions between neighboring Ni and Pt or Ni and Co increase when the intermixing process occurred. Hence, the perpendicular magnetic anisotropy deceases.

Figure 6.14: (a) Polar and longitudinal hysteresis loops of 2 ML Co/2 ML Ni/Pt(111).

Each curve was measured at room temperature after annealing at Ta for 20 minutes. (b) The annealing temperature dependences of polar and longitudinal Kerr intensities. (c) The evolutions of Ni 848 eV, Co 656 eV, and Pt 237 eV AES signals versus annealing temperature.

(25)

Fig. 6.14(a) shows the polar and longitudinal hysteresis loops of 2 ML Co/2 ML Ni/Pt(111) after annealing at each Ta for 20 minutes. All the loops were measured at room temperature. Only longitudinal Kerr signal can be observed unless annealing the sample at a temperature higher than 850 K. When the annealing temperature is higher than 850 K, only polar Kerr signal can be observed. The easy axis of the magnetization is initially in the in-plane direction, it changes from the in-plane to the out-of plane after annealing at a temperature higher than 850 K. That means, as 2 ML Ni/2 ML Co/Pt(111), a spin reorientation transition occurred by the annealing. The annealing temperature dependence of Kerr intensities is shown in Fig. 6.14(b). When the annealing temperature is between 300 K and 600 K, the longitudinal Kerr intensity does not change and so does the polar Kerr intensity. The longitudinal Kerr intensity increase with Ta when 600 K≤ Ta ≤ 750 K, after which it starts to decrease and reach a value of zero at 850 K. It is interesting that the polar Kerr intensity appears abruptly when the annealing temperature is 850 K. The annealing temperature dependence of Pt 237 eV, Ni 848 eV and Co 656 eV AES signals are shown in Fig. 6.14(c). The Co 656 eV intensity starts to decrease at 600 K while the Ni 848 eV intensity starts to increase at the same annealing temperature. No significant increase of Pt 237 eV AES intensity is observed unless Ta 800 K. The Ni and Co atoms intermix to each other when 600 K ≤ Ta ≤ 800 K. The Ni-Co intermixing layer diffuses into the Pt substrate increase the perpendicular magnetic anisotropy when 800 K≤ Ta ≤ 900 K. Comparing the annealing induced SRT between 2 ML Ni/2 ML Co/Pt(111) and 2 ML Co/2 ML Ni/Pt(111), the change of the surface composition by annealing may change the effective anisotropy of the ferromagnetic film thus causes the spin reorientation transition.

The effective anisotropies the as-deposited films of 2 ML Ni/2 ML Coi/Pt(111) and 2 ML Co/2 ML Ni/Pt(111) can be expresses as eq. (6.4) and eq. (6.5), respectively. The two groups marked as (1) and (2) have the same value when the buffer layer changes from Co films to Ni films. The values of KsCo−P t, KsN i−P t, KsCo−vacuum, and KsN i−vacuum are 0.57 mJ/m2 [101], 0.17 mJ/m2 [16], −0.28 mJ/m2 [59], and −0.48 mJ/m2 [60], respec- tively. The value of KsCo−P t is about 3.5 times the value of KsN i−P t, but the value of KsN i−vacuum is only 1.5 times that of KsCo−vacuum. Therefore, it is reasonable that the direction of the magnetization easy axis changes from the out-of-plane direction to the in-plane direction when the buffer layer changes from Co films to Ni films. (out-of plane:

(26)

Kef f > 0; in-plane: Kef f < 0)

Kef f2N i2Co= KvN i+ KvCo

  

(1)

+KsCo−P t

dCo +KsN i−Co

dCo +KsN i−Co dN i

  

(2)

+KsN i−vacuum

dN i (6.4)

Kef f2Co2N i= KvN i+ KvCo

  

(1)

+KsN i−P t

dN i +KsN i−Co

dN i +KsN i−Co dCo

  

(2)

+KsCo−vacuum

dCo (6.5)

After high temperature annealing, the Co and Ni atoms diffuse into the Pt substrate.

It is straightforward that, after higher temperature annealing, the ratio of Ni% to Co%

of 2 ML Ni/2 ML Co/Pt(111) is higher than that of 2 ML Co/2 ML Ni/Pt(111) in the subsurface region. Alternatively, in the bulk region, the ratio of Ni% to Co% of 2 ML Ni/2 ML Co/Pt(111) must be smaller than that of 2 ML Co/2 ML Ni/Pt(111). The schematic diagrams of the surface compositions are shown in Fig. 6.15. The ratio of the black region is lower than that of the white region. Comparing the spin reorientation transitions of 2 ML Ni/2 ML Co/Pt(111) and 2 ML Co/2 ML Ni/Pt(111), it seems that the sample favors an in-plane easy axis when the ratio is larger in the top surface region and smaller in the bulk region. But it is interesting that no SRT was observed for the 1 ML Ni/1 ML Co/Pt(111) and 1 ML Co/1 ML Ni/Pt(111) systems. Unlike the one-ML mirror systems, the ratios of Ni% to Co% for 2 ML Ni/2 ML Co/Pt(111) and 2 ML Co/2 ML Ni/Pt(111) systems do not keep constant during high temperature (not shown here). This implies that, after high temperature annealing, the surface compositions of the two-ML mirror systems are less homogeneous than one-ML mirror systems. This may be the reason that causes the phenomenon of SRT after high temperature annealing.

           





            

           





            

           





            

           





            

           





            

           





            

Figure 6.15: Surface structures after high temperature annealing for 2 ML Ni/2 ML Co/Pt(111) and 2 ML Co/2 ML Ni/Pt(111)

數據

Figure 6.1: (a) The polar and longitudinal Kerr hysteresis loops of different coverages of Ni deposited on 1 ML Co/Pt(111)
Figure 6.2: The Kerr hysteresis loops of ultrathin Co films deposited on 1 ML Ni/Pt(111).
Figure 6.3: The evolutions of polar Kerr intensities versus sample temperature during the first heating process.
Fig. 6.3 shows the evolutions of polar Kerr intensities for 1 ML Ni/1 ML Co/Pt(111) (curve (a)) and 1 ML Co/1 ML Ni/Pt(111) (curve (b)) during the first heating processes.
+7

參考文獻

相關文件

Particles near (x, y, z) in the fluid tend to rotate about the axis that points in the direction of curl F(x, y, z), and the length of this curl vector is a measure of how quickly

remember from Equation 1 that the partial derivative with respect to x is just the ordinary derivative of the function g of a single variable that we get by keeping y fixed.. Thus

The following theorem states that (1 + x) k is equal to the sum of its Maclaurin

The third step is to express the proper rational function R(x)/Q(x) (from Equation 1) as a sum of partial fractions of the

The best way to picture a vector field is to draw the arrow representing the vector F(x, y) starting at the point (x, y).. Of course, it’s impossible to do this for all points (x,

Master of Arts in Interdisciplinary Cultural Studies FT, PT Master of Mediation and Conflict Resolution FT, PT Master of Social Sciences in Psychology PT The University of

 Teachers have to understand the salient features of the three pedagogical approaches of Direct Instruction, Enquiry Learning and Co-construction outlined below and

Then, we tested the influence of θ for the rate of convergence of Algorithm 4.1, by using this algorithm with α = 15 and four different θ to solve a test ex- ample generated as