Exchange bias in nanostructures

This study was concerned with two issues facing the technological application of exchange bias. The first is the ferromagnet (F)/antiferromagnet (AF) interface which was investigated via the effect of the IrMn (111) in-plane texture on the magnetisation reversal of IrMn/CoFe samples. The second is t...

Full description

Bibliographic Details
Main Author: Carpenter, Robert
Other Authors: O'Grady, K. ; Vallejo-Fernandez, G.
Published: University of York 2015
Subjects:
530
Online Access:http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.647745
Description
Summary:This study was concerned with two issues facing the technological application of exchange bias. The first is the ferromagnet (F)/antiferromagnet (AF) interface which was investigated via the effect of the IrMn (111) in-plane texture on the magnetisation reversal of IrMn/CoFe samples. The second is the change in the magnitude of the exchange bias (Hex) and the median blocking temperature (Tbmed)with element size in sub-500nm nanostructures. The evolution of the magnetisation reversal over the annealing/field cooling process in CoFe and CoFe/IrMn thin films was measured for samples deposited on Cu and NiCr seed layers. The samples deposited on the Cu seed layer were found to be polycrystalline but randomly oriented whilst those on a NiCr seed layer had a strong (111) in-plane texture. The training effect is the change in the first point to reversal (Hc1) between the first and second hysteresis loops measured after field cooling. This was found to vary drastically with texture where deltaHc1 was found to be (10+/-2.5)Oe and (60+/-2.5)Oe for the samples deposited on the NiCr and Cu seed layers respectively. This was hypothesised to be due to a distribution of orientation of the easy axis of the interfacial spin clusters. In the case of the sample deposited on the Cu seed layer there is a 3-D random distribution of easy axis orientations similar to that of Stoner-Wohlfarth entities whereas for the sample grown on the NiCr seed layer there is a 2-D random distribution of easy axis. It is the tensioning of the cluster-cluster interactions with the easy axis orientation that is thought to give rise to the training effect and other phenomena in exchange bias. The change in Hex and Tbmed with element size was measured in 4x4mm arrays of nanodots patterned through e-beam lithography and fabricated using the lift-off method. Following an initial decrease of 66% due to the patterning process, Hex decreased from (105+/-5)Oe to (40+/-5)Oe with element size from 425 to 80nm. However whilst an initial decrease in Tbmed from (448+/-5)K to ~405K occured due to the patterning process no significant change is seen with element size. This implies that the change in Hex is due to a modification of the interface i.e. non-planar deposition as a result of depositing the film through a mask.