5.4 Properties of Electrodeposited CMAs
169
One feasible strategy is to use a substrate that can be dissolved later. This works for
Cu substrate and Ni–Co/Cu deposits since Cu can be dissolved from a Ni-rich alloy
with sufficient selectivity. The substrate removal is also possible by simply peeling
off the deposit if it is thick enough and self-supporting. In this case, the substrate
is a flexible Ti sheet on which the adherence of a metallic deposit is week because
of the native TiO 2 layer. The peeling-off method works only if a thick deposit can
be grown with a small roughness, which provides that the sample is not torn apart
during the removal from the substrate. Another common approach is the application
of a thin evaporated metal film on a non-conducting substrate whose conductivity
does not disturb significantly the magnetoresistance measurement. As the last option,
semiconductor wafers are also commonly used, and their resistance is high enough
so that the sample is not shunted during the resistivity test. The resistivity test on
multilayer films is always carried out in a current-in-plane configuration. Depending
on the direction of the external magnetic field, the longitudinal (H I) and the
transverse (H ⊥ I) component of the magnetoresistance can be determined.
For many systems, there is a full analogy between metallurgically processed and
electrodeposited samples. However, this is absolutely not true for multilayers with
GMR. For evaporated, sputtered or MBE-deposited multilayers, the smooth growth
of the subsequent layer could be achieved with a monoatomic surfactant metal layer
(mostly Pb or Ag) that provided a smooth growth of subsequent layers that would
grow otherwise as a granular film. Although there were similar attempts in multilayer electrodepositon experiments, no analogous multilayers could be produced
[176, 177]. In the absence of a surfactant layer, the grain growth takes place as it
is shown in Figs. 5.10 and 5.11b. This means that the layers are not parallel to the
substrate, especially at large layer thicknesses, which also means that the application
of an external magnetic field parallel to the substrate will not lead to a fully in-plane
magnetization. This gives rise to a magnetostatic coupling between the neighbouring
FM layers. This magnetostatic coupling (which is often named as Néel-type “orange
peel” coupling) always leads to a parallel magnetization alignment between the
FM layers. The orange peel coupling easily overwrites the coupling effect dictated
by the spacing of the FM layers by the NM one, hence destroying the antiparallel
alignment of the neighbouring FM layer magnetizations [101, 178]. The geometric
representation of the magnetostatic coupling effect is shown in Fig. 5.14.
Indeed, the oscillatory parallel/antiparallel coupling between the neighbouring
FM layers has not been evidenced for any electrodeposited FM/NM multilayer
sample [156]. Instead of an oscillatory GMR as a function of the NM layer thickness,
the following general trend could be seen for electrodeposited multilayers.
At small NM layer thickness, the NM layer is discontinuous, which allows direct
contact (i.e., direct ferromagnetic coupling) between the neighbouring FM layers.
In fact, we cannot speak about true “layers”; instead, the correct term would be “the
material deposited during the subsequent high-current pulses”, but this complicated
phrasing is never applied. When the dominating effect is the coupling of the neighbouring FM layers via pinholes in the NM layer, the magnetoresistance character
is the anisotropic magnetoresistance (AMR) that is a feature of the bulk ferromagnets. The pinholes in the FM layers are difficult to observe directly, especially with
169
One feasible strategy is to use a substrate that can be dissolved later. This works for
Cu substrate and Ni–Co/Cu deposits since Cu can be dissolved from a Ni-rich alloy
with sufficient selectivity. The substrate removal is also possible by simply peeling
off the deposit if it is thick enough and self-supporting. In this case, the substrate
is a flexible Ti sheet on which the adherence of a metallic deposit is week because
of the native TiO 2 layer. The peeling-off method works only if a thick deposit can
be grown with a small roughness, which provides that the sample is not torn apart
during the removal from the substrate. Another common approach is the application
of a thin evaporated metal film on a non-conducting substrate whose conductivity
does not disturb significantly the magnetoresistance measurement. As the last option,
semiconductor wafers are also commonly used, and their resistance is high enough
so that the sample is not shunted during the resistivity test. The resistivity test on
multilayer films is always carried out in a current-in-plane configuration. Depending
on the direction of the external magnetic field, the longitudinal (H I) and the
transverse (H ⊥ I) component of the magnetoresistance can be determined.
For many systems, there is a full analogy between metallurgically processed and
electrodeposited samples. However, this is absolutely not true for multilayers with
GMR. For evaporated, sputtered or MBE-deposited multilayers, the smooth growth
of the subsequent layer could be achieved with a monoatomic surfactant metal layer
(mostly Pb or Ag) that provided a smooth growth of subsequent layers that would
grow otherwise as a granular film. Although there were similar attempts in multilayer electrodepositon experiments, no analogous multilayers could be produced
[176, 177]. In the absence of a surfactant layer, the grain growth takes place as it
is shown in Figs. 5.10 and 5.11b. This means that the layers are not parallel to the
substrate, especially at large layer thicknesses, which also means that the application
of an external magnetic field parallel to the substrate will not lead to a fully in-plane
magnetization. This gives rise to a magnetostatic coupling between the neighbouring
FM layers. This magnetostatic coupling (which is often named as Néel-type “orange
peel” coupling) always leads to a parallel magnetization alignment between the
FM layers. The orange peel coupling easily overwrites the coupling effect dictated
by the spacing of the FM layers by the NM one, hence destroying the antiparallel
alignment of the neighbouring FM layer magnetizations [101, 178]. The geometric
representation of the magnetostatic coupling effect is shown in Fig. 5.14.
Indeed, the oscillatory parallel/antiparallel coupling between the neighbouring
FM layers has not been evidenced for any electrodeposited FM/NM multilayer
sample [156]. Instead of an oscillatory GMR as a function of the NM layer thickness,
the following general trend could be seen for electrodeposited multilayers.
At small NM layer thickness, the NM layer is discontinuous, which allows direct
contact (i.e., direct ferromagnetic coupling) between the neighbouring FM layers.
In fact, we cannot speak about true “layers”; instead, the correct term would be “the
material deposited during the subsequent high-current pulses”, but this complicated
phrasing is never applied. When the dominating effect is the coupling of the neighbouring FM layers via pinholes in the NM layer, the magnetoresistance character
is the anisotropic magnetoresistance (AMR) that is a feature of the bulk ferromagnets. The pinholes in the FM layers are difficult to observe directly, especially with