4.2 Growth studies on the Co-layers
4.2.1 Varying the pressure and power during the Co layer deposition 48
Figure 4.6 shows plots of the saturation fields and RKKY-coupling strength values versus the thickness of the spacer layer for samples for which the Co layer is grown at 10 W (top), 20 W (middle) and 40W (bottom). The pressure used during the Co deposition is varied as well between 7 sccm (black data) and 15 sccm (red data). Table 4.2 shows an overview of important parameters obtained from the experimental data.
Increasing the pressure from 7 sccm to 15 sccm during the Co deposition increases the strength of the RKKY-coupling significantly. As the pressure increases, the amount of particles present in the Argon plasma increases. When a Co particle is knocked off from the target and moves towards the substrate, it will collide with more particles
7)sccm) 15)sccm) Co 3 nm Ru wedge 0-4 nm Co 3 nm
10W)
7)sccm) 15)sccm) Co 3 nm Ru wedge 0-4 nm Co 3 nm
20W)
7)sccm) 15)sccm) Co 3 nm Ru wedge 0-4 nm Co 3 nm
40W)
Figure 4.6: plot of saturation fields and RKKY coupling strengths versus thickness for Ta(2)/Co(3)/Ru(wedge 0-4 nm)/Co(3)/Ta(5) samples. During deposition, the Co-layers are grown with a pressure of 7 sccm (black data) and 15 sccm (red data).
The power during the Co deposition is set at 10W (top), 20 W (middle) and 40 W (bottom).
on the way to the substrate. The energy with which the particle arrives on the substrate is therefore lower, and the impact on the substrate and growth mode can change due to this difference in energy. Increasing the pressure to 15 sccm therefore leads to a better growth of the Co layers and less interdiffusion of the Co layer with the Ta and Ru layers. A better Ta/Co and Ru/Co interface increases the RKKY-coupling strength, as is observed from the experimental data. This agrees with the
10 W Pressure 7 sccm 15 sccm
20 W Pressure 7 sccm 15 sccm
tAF −peak (10−9 m) 0.80 1.06
40 W Pressure 7 sccm 15 sccm
tAF −peak (10−9 m) 1.07 0.91
Table 4.2: Overview of the location and width of the AF regions and the maximum saturation field and RKKY-coupling strength values for Ta(2)/Co(3)/Ru(wedged 0-4 nm)/Co(3)/Ta(5) samples.
presence of a second AF domain which is only found when the Co layers are grown at 15 sccm.
The power during the Co layer depositions affect the RKKY-coupling strength as well. The power used during the deposition mainly affects the growth speed of the Co layers, as increasing the power increases the amount of particles that are knocked off from the target. This however should not change the energy of the Co particles
thickness at which the first peak is found does not change significantly except for the sample grown at 7 sccm and 20 W. The start of the Ru-wedge is determined by hand, and a small variation in the position of the peaks is expected, however not as large as is observed. The position at which the second AF peak is found is the same for all samples when the Co is grown at 15 sccm. The width of this domain is significantly higher for the sample using a Co layer deposition power of 20 W.
Samples were grown with a pressure of 30 sccm during the Co deposition, however these samples do not show any magnetic response when analysing them with the MOKE or VSM-SQUID setup. We conclude that the optimal growth conditions for the Co layers are a pressure of 15 sccm and a power of 20 W. During this growth study, only the growth conditions during the Co deposition are varied. The RKKY-coupling could be improved by optimizing the growth conditions of the spacer layer as well. However, using a pressure of 15 sccm during the Co deposition already resulted in a second AF domain, which is one of the goals of the growth study. We decided to not do a second growth study on the spacer layer and start to investigate the thermal stability of the antiferromagnetically coupled domains.
4.3 Investigating the thermal stability of the SAF
In order to investigate the thermal stability of the SAFs, synthetic antiferromagnets with a fixed Ru and Ir spacer layer thickness are annealed and magnetic loops are measured. Figure 4.7 shows magnetic loops obtained during these measurements. In both samples, the Co layers are coupled antiferromagnetically at room temperature.
The samples are annealed and measured at different temperatures. The Co/Ru/Co sample showed no significant change for the magnetic loops up to 230° C. At 250°
C the magnetic loop starts to change, as the saturation field is decreasing. At this temperature, the magnetization switches at low field, and for increasing field a ’rounding off’ of the magnetic loop is observed. This magnetic behavior can be the result of biquadratic coupling, which as explained in section 2.4.6.3 can be the result of loose spins in the spacer layer. Loose spins result in frustrations between the spins and can result in a preferential angle of 90° between the magnetizations at low fields. Loose spins can diffuse into the spacer layer during the annealing process.
At 290° C the AF coupling is completely gone. The same change is observed for the Co/Ir/Co samples, however the temperatures at which the AF coupling starts to decrease is much lower, as at 190° C the coupling starts to change significantly.
When the temperature is cooled down to room temperature, the measured loops do not change and the loss of AF coupling is irreversible. This forms a problem when the synthetic antiferromagnets are combined with an AF layer to create EB.
As explained in section 2.2, in order to set the EB, the magnetic multilayer need to be annealed up to TN < T < TC. The AF layer used during this thesis is an IrMn
CHAPTER 4. RESULTS 52
Figure 4.7: Magnetic loops of an SAF containing a fixed Ru (left) and Ir (right) spacer layer. In these samples, tRu = 1.2 nm and tIr = 0.8 nm, resulting in an AF coupling. All layers are grown at standard conditions and are annealed using the VSM-SQUID oven module. The samples are annealed and measured at the indicated temperatures.
layer, which requires an annealing of 30 mins at a temperature of 250° C.
A Ta(2)/Co(3)/Ru(wedged 0-4 nm)/Co(3)/Ta(5)-sample is annealed with the Argon oven and magnetically characterized. The sample is grown at standard growth conditions, except for the Co layers which are grown at a pressure of 15 sccm and a power of 20 W. The samples are measured with the MOKE setup to track the change of the AF domains when the annealing temperature is progressively increased. Figure 4.8 shows the saturation field and RKKY coupling versus the spacer layer thickness for 4 different measurements. The sample is heated up to the annealing temperature Tannealing, annealed at this temperature for 1 minute and then cooled down back to room temperature. The annealing temperature is progressively increased, starting from 70° C and increased with 50° C per measurement to a maximum of 320° C . Magnetic loops are measured at room temperature using the MOKE setup and saturation field values and RKKY coupling strengths are extracted. When the sample is annealed at 220° C, significant changes in the first AF coupling domain occur.
The domain starts to shrink and the maximum saturation field and RKKY-coupling strength are decreased by 30%. When the annealing temperature is increased to 270°
C an even more dramatic decrease of the RKKY-coupling is observed. Although barely visible due to a large amount of noise in the bottom left figure of 4.7, analysis
Tannealing=270°#C# Tannealing=320°#C#
JRKKY (mJ/m2) JRKKY (mJ/m2) JRKKY (mJ/m2) JRKKY (mJ/m2)
Figure 4.8: Plots of the saturation field and RKKY coupling strength versus the Ru spacer layer thickness for a Ta(2)/Co(3)/Ru(wedged 0-4 nm)/Co(3)/Ta(5)-sample.
The data is obtained by analysing the stack by MOKE directly after the deposition process (top left), after annealing the stack at 220° C (top right), 270° C (bottom left) and 320° C (bottom right). As explained in the previous section, when the Co layers are antiferromagnetically coupled, a peak in both saturation field and RKKY-coupling strength is found. Two AF domains are found for samples measured after the deposition and after annealing the samples at 220° C and 270° C. After annealing the sample at 320° C, the first AF domain has disappeared.
of the hysteresis loops showed the presence of a weakly coupled AF domain. The maximum RKKY-coupling is reduced to 3% of its value as measured right after the deposition process. When increasing the annealing temperature to 320° C, the first AF domain is completely gone. The second AF domain is much more stable.
The width of the domain only decreases when an annealing temperature of 320° C is used. The maximum saturation fields and RKKY-coupling strengths do not decrease as dramatically as in the first AF domain. A complete overview of the saturation fields and coupling strengths of the AF domains is given in table 4.3.
Figure 4.9 illustrates the position of the AF coupling domains as found in the samples. The first AF domain disappears in an asymmetric way, as the first bound-ary position changes more rapidly than the position of the second boundbound-ary. The reproducability is checked by repeating the annealing measurements on an identical sample. This second sample showed the same behavior.
Tannealing(°C) Hsat,max (10−3 T) JRKKY,max (10−3J/m2)
Table 4.3: Overview of important parameters of the AF domains. In the upper part the maximum saturation field Hsat,max and corresponding RKKY-coupling strength JRKKY found in the first region are given. The lower part of the table presents the same parameters for the second AF domain.
0)nm)
Figure 4.9: Overview of the AF regions of an Ta(2)/Co(3)/Ru(wedged 0-4 nm)/Co(3)/Ta(5)-sample as grown (black) and after annealing it at at 220° C (blue), 270° C (red) and 320° C (orange). As the annealing temperature is increased, the first AF domain disappears asymmetrically. The second AF domain is more stable against temperature. Figure not drawn on scale.
An5ferromagne5c)coupling) Ferromagne5c)coupling)
))
Increasing)spacer)layer)thickness)
tRu=)0.34)nm) tRu=)0.78)nm) tRu=)0.81)nm) tRu=)0.86)nm)
ΔM)=)2.00) ΔM)=)0.59) ΔM)=)0.31) ΔM)=)0)
Hsat)=)0)mT)) Hsat)=)504)mT)) Hsat)=)549)mT)) Hsat)=)575)mT))
Figure 4.10: The evolution of hysteresis loops for a Ta(2)/Co(3)/Ru(wedged 0-4 nm)/Co(3)/Ta(5)-sample in the first transition region. A gradual decrease of the magnetic switch is observed. The sample is grown at standard conditions except for the Co layer, which is grown at a pressure of 15 sccm and a power of 20 W.