The relation between magnetic hysteresis and the micromagnetic state explored by quantitative magnetic force microscopy

Since the discovery of magnetic lodestones in 600 BC described by ancient
Greek philosophers, the mystery of magnetism has continued to intrigue. For
thousands of years, these lodestones were more of a curiousity until medieval
explorers discovered how to use them to produce a magnetic compass. The
compass was studied and explained by William Gilbert in his 1600 treatise on
magnetism called "de Magnete\1{3. In his work, Gilbert proposed the Earth
itself to be magnetic, helping to open up the modern ¯eld of magnetism. Only
at the start of the nineteenth century was magnetism discovered to be due
to elementary magnets analogous to how matter is made up of atoms and
molecules. This idea was then developed further by Pierre-Ernest Weiss in
19074, who discovered that these elementary magnets grouped together in
ferromagnets to form magnetic domains. In 1933, these domains were exper-
imentally con¯rmed by Barkhausen5 after observing that magnetisation was
a discontinuous process. He ampli¯ed the sound produced by these discontin-
uous jumps and attributed it to domain switching. In a contemporary study,
Bitter developed a technique using a suspension of magnetic colloids to visu-
alise the domains6. Domain switching was subsequently found to be lacking
as an explanation of magnetisation by Langmuir in 19317, who proposed that
the propagation of the walls surrounding the domains was responsible for the
magnetisation reversal. This domain wall motion also explained the appear-
ance of hysteresis which characterises the tendency of a magnetic system to
behave di®erently depending on its ¯eld history. The study of hysteresis re-
ally took o® with the advent of magnetometry. These experimental techniques
are capable of measuring the average magnetisation of a system as a function
of applied ¯eld. One of these, vibrating sample magnetometry (VSM), was
invented in 19598 and has been used ever since as the principle means of ac-
cessing a systems magnetic parameters. Since VSM is an volume averaging
technique, the microscopic nature of domains remained elusive until the ad-
vent of magnetic force microscopy (MFM) in 19879. This powerful technique
can image the magnetic ¯eld emanating from the surface which gives much
needed information about the magnetisation for 100'snm below the surface.
This is therefore especially useful when studying magnetic thin ¯lms.
The widespread use of magnetic thin ¯lms in contemporary data storage
technology, sensors of magnetic ¯eld, strain and acceleration, has been made
possible by the optimisation of the materials involved, aimed at attaining
speci¯c magnetisation structures and hysteresis curve characteristics. However
the understanding at the microscopic scale of the reversal process remains a
challenge.
This thesis presents a study of magnetism in thin ¯lms, more speci¯cally
thin ¯lms with perpendicular magnetic anisotropy (PMA). These mag-
netic thin ¯lms with an easy axis of magnetisation perpendicular to the surface
have attracted particular attention due to their potential advantage in high-
density magnetic recording media10{13. Magnetic thin ¯lms with PMA also
provide a model system of domain behaviour in a wide variety of materials.
More particulary, thin ¯lms of nickel have been shown to be of great interest
because of their intrinsic magnetic properties favouring PMA14{17.
The origin, advantages and removal of PMA in thin ¯lms are investigated
in this thesis. While both Chapters 3 and 4 present studies of the origin of
PMA in nickel ¯lms, Chapter 3 explains how the conventional interpretation of
the magnetisation loops can give an incorrect picture of both the magnetisa-
tion processes and magnetic domain structures in the ¯lms. The observations
defy explanation based on volume averaged magnetometry measurements but
come to light when magnetometry is combined with quantitative magnetic
force microscopy (qMFM). qMFM revealed a speci¯c magnetisation pattern,
namely perpendicular stripe domains with closure caps. Furthermore, the do-
main wall is shown to be as important as the domain itself in the average mag-
netisation. These new ¯ndings come from using the well established method
of magnetometry complimented by the latest qMFM techniques. Chapter 4
concentrates on how ion irradiation can be used to remove PMA. Furthermore,
it shows how temperature variation can be used to control the perpendicular
component of the magnetisation.
A common theme through this thesis is the coercivity of the ferromagnetic
¯lm. Hysteresis lies at the very foundation of the magnetic recording indus-
try10,12. Hysteretic systems are employed as recording media because they
retain their magnetic state for a long period after a writing operation, namely
they exhibit magnetic memory. This memory has been extensively studied
and exploited. However, despite decades of intense study and signi¯cant re-
cent advances18,19, a fully satisfactory microscopic understanding of magnetic
hysteresis is still lacking20,21. The origin of hysteresis and its relation to the
micromagnetic state has been studied in Chapters 3, 4 and 5. Our novel results
show how domain wall motion can be hysteresis free (Chapter 3 and 4).
Domain nucleation and wall motion were studied in more detail in Chap-
ter 4, in which we show that increasing the defect density does not necessarily
increase the hysteresis. Indeed, it can even decrease it by acting on the domain
nucleation and therefore on the domain density. The low pinning action of the
domain wall is shown to be due to the small size of the domain, implying
short distance to travel for the domain wall to allow the magnetisation rever-
sal. Thus, the hysteresis is shown to depend strongly on the micromagnetic
domain structure and particularly on the domain width.
The reproducibility of the domain nucleation and wall motion is a key
factor in microscopic memory20 and is studied in Chapter 6. The e®ect of
defects on this microscopic reproducibility is shown to be more complicated
than the common understanding. We show that defects act as nucleation
centres for domains during the magnetisation reversal. This implies a good
reproducibility of the way domains nucleate after saturation. On the other
hand, defects also act to break up the domains while a demagnetised sample is
brought towards saturation. This decreases the reproducibility of the domain
evolution on a minor loop excursion.
Chapter 7 presents a study of a perpendicular magnetised array of nanois-
lands. This has been previously proposed with a view to high density recording
media10. The hysteresis of such systems has to be characterised and controlled.
We show that decreasing the island size under 50nm increases the coercivity
distribution, to detriment of their archival potential.
In summary, this thesis presents novel information about the mechanism
leading to hysteresis and the related micromagnetic state. Yet, it also im-
proves the understanding of magnetic measurements two folds. It highlights
the danger of quick interpretation of magnetometry. Moreover, it presents a
novel way of using state-of-the-art qMFM.