The common theme
of the samples were that they are all
magnetic recordings made throughout history, which included a
meteorite shard, a floppy disk, a zip drive disk, and a hard disk. The Campo
del Cielo meteorite shard originated
from a region in Argentina
and was expected to have made impact approximately four to five
ago, it was acquired from geological sample vendor and was expected to
authentic. It turns
out, that meteorites
are an incredibly useful tool for the paleomagnetism community as it
to look back to a fixed moment in time.
the meteorite makes impact, the shards undergo tremendous heating and
allowing for the magnetic domains to interact with the
earth’s current magnetic
field before cooling back down and becoming fixed.
The floppy disk and zip drive disk are both
information storages devices that encoded information into magnetic
they used similar technologies that allowed them to be relatively
information storage devices that could be transported between locations. The devices main component
is a thin circular
Mylar section that has a magnetic coating where the information is
main difference between the two types of disks are their section sizes
domain size where the smaller domains allow for high information
the technology that allowed
them to be so inexpensive came at the cost of capacity, and largely
fell out of
fashion by the mid-2000s. A
information storage device to the floppy and zip disk is the hard disk
and operates under similar principals to the floppy and zip disk. The technology has been
further refined to
where a magnetic coating is placed on circular aluminum platters, and
become the prevailing secondary storage devices for general purpose
a read-write head of hard disk drive
was also examined.
often viewed under a light microscope initially
to ensure proper sample preparations before placing into the AFM or SEM
wasting valuable instrument time.
shard was naturally rough and required
extensive polishing before viewing.
Sequential polishes on a polishing wheel with silicon
carbide grinding sheets
were performed starting at grit sizes (average grit size in microns)
800 (12), and 1200 (2.5).
turned out that meteorite was initially rather dull and the main
were just the polishing lines. To
the material grain sizes, the shard was placed in a Nital solution for
minute, a mixture of nitric acid and alcohol.
Extreme care was used due to the volatility of the
– Conductive Coating
floppy disks were partly made of Mylar, which is insulating, and to
potential charging effects when viewing the samples in the SEM, the
were coated with a thin layer of gold using a Denton Sputtering system. Proper groundings to the
samples were also
added with the use of conductive carbon paint.
was performed before attempting MFM to
determine sample surface morphology.
would have greatly helped if the tips had not been broken in that it
allowed me to see if the tip was seeing the surface correctly with the
pass. There was the
potential for the
first pass to become severely distorted due to both the surface and
offered a way to look
at the surface roughness and verticality, which the SEM is often unable
Electron Microscopy – SE2, Inlens, BSD
The SEM with
appropriate viewing parameters offers superior visualization
of the surface morphology compared to AFM, and potentially useful
was taken to ensure
the samples were firmly secure to protect the instrument from any
interactions before being placed into the Zeiss Auriga SEM.
Dispersive X-ray Spectroscopy – X-ray collection, and Mappings
EDX allows for
the determination of sample elemental
composition, and when mapping, spatial variations in elemental
Results & Discussions
Due to changes
in the project direction midway, I ended up collecting
a tremendous of data and images. Some were good images and some were
awful images. For
brevity’s sake, I am
only displaying a collection of images from each sample that I believe
represent the sample while still allowing for comparisons between the
sample surfaces. Most
of the samples
have an SEM images taken at a lower magnification (1kX) to see large
morphology and an SEM image taken at a higher magnification (10kX) to
coatings/grain surface morphologies.
measurements are usually presented for 30umx30um and 2umx2um scale with
and 3D perspectives. Lastly,
I tried to incorporate
one or two additional bits of information that I believe were
SEM: In the pre-etched SEM images shown below, the main surface
streaks were related to the polishing. The large clumps are most
likely dust and polishing debris not properly removed before imaging.
You can clearly make out polishing streaks going in two different
direction, which is due to lack of prior knowledge polishing
on behalf. During polishing, I rotated the shard to get a better
grip on it with respect to the wheel changing the direction of the
AFM: The AFM is very clearly able to resolve the two different
polishing streaks. In the images there are vertical streaks
present, which are an artifact of the tip striking the surface a bit to
hard. Roughness calculations were performed in regions lacking
the large polishing and dust contaiminations.
|Root Mean Square
EDX Spectrum: The only elements visible in the characteristic
x-ray spectrum is iron and nickel, which is a strong suggestion that
the elemental composition is similar to verified meteorites.
Light Microscope Image: Shown
below is one of the light microscope images taken of the pre-etched
sample. The polishing lines are very clear.
SEM: After the Nital etching process, the meteorite shard's
polishing streaks are sufficiently suppressed and the grains become
easily visible. The surface is now remarkably patterned in a
rough pattern reminiscent of sea waves.
AFM: The AFM very easily is able to resolve the etched shards
roughness and is able to pick up the rolling up and down features of
the grains. The sample roughness for both scans has gone up a
reasonable amount compared to the pre-etched shard.
|Root Mean Square
EDX Mapping: It was my hope that by etching the shard, the grain
boundaries would display some variations of elemental composition.
Unfortunately, nital does not etch based off of elemental
composition but instead by the grains phase orientation. Thus the
EDX mapping of the grains proved not useful. Afterward, I tried
taking an EDX mapping of a crack in the sample. The first EDX map
is of iron and the second is of nickel. From the mappings, it is
clear that the crack is deficient in iron, and the nickel is pretty
uniformly spread out.
SEM: The SEM images of he floppy disk disply the spindly
structures made by the magnetic coating on the mylar sheet. There
are regions of bright and dark spots in both SE2 and Inlens detector
images, at both high and low accelerating voltages. EDAX mappings
also provided little in the way of determining the variations in
images. As best as I can reason, the surface is not as flat as it
appears from the AFM images. Instead of flat the surface is
forming mountaints and valleys and some of the escaping electrons are
being blocked from reaching the detectors leading to regions of dark
and bright. It should also be noted that the bright edge effect
is clearly visible in the second image, indicating an exposed surface.
These features are stationary indicating that it is not a
AFM: The AFM picture further confirm the mountain and valley
theory, which are clearly visible in the 2umx2um image. It again
appears that the tip was striking the surface a bit to hard leading to
the streaking image shown in the 2um image. Lastly below the
roughness measurements, are two 30umx30um images of the similar areas.
The left image was scanned top to bottom, while the right image
was scanned left to right. The bright spot orientation appear to
change their shape depending on the scan direction, indicating that
there was most likely something wrong with the tip, perhaps it picked
something up on one side leading to distorted scans.
|Root Mean Square
SEM: The zip disk sem images again show a spindly pattern on the
surface, and they also display the dark and bright regions. This
is most likely due to how the surface coatings are deposited since the
processes are similar for both zip and floppy disks. The edge
effect is particular visible in the 10kX magnification, and this again
is not due to charging.. I also included images of the samples
after taking a full EDX mapping of the bright and dark spots. The
zip disk post EDX image is on the left and the floppy disk post EDX
image is on the right. It is very easy to see the rip in surface
exposing the insulating mylar films underneath. What was so
remarkable about these photos, is that the damage was not occuring
where the beam was striking the surface, but instead further away.
The floppy disk image in particular even shows the carbon deposit
from the beam showing very clearly where the mapping took place, yet
the damage occured some distance away.
|Root Mean Square
EDX Spectrum: The EDX spectrum for the zip disk is particulary
puzzling. It appears to have another magnetic iron oxide
coating, but it also features paramagnetic titanium. I was
unable to find any work suggesting the presence of titanium, but I was
also not able to verify the coating used on zip disks. What I do
know is that the zip disk's magnetic coating was able to increase the
storage capacity by about a factor of 10.
SEM: The hard disk was particulary difficult to collect SEM
images of, due to the incredibly smooth surface, there was just very
little topologically going on. So secondary electron detectors
were unable to pick up much of anything. There is also a
backscattered detector image featuring a bit of dirt for contrast shown
below, but it was also fairly dull of an image.
AFM: The AFM was able to clearly pick up streaking patterns
across the surface. While the scratches are incredibly small ~0.5
nm, shown by the surface roughness calculation and sectional analysis,
it is likely damage caused by the head which lead to the crashing of
the hard disk. Even this small variation likely lead to the disk
failing, which is how I managed to obtain a piece of one.
|Root Mean Square
EDX Spectrum: The EDX spectrum for the hard disk clearly
shows peaks for cobalt, nickel, and phosphorus. The magnetic
coating for hard disk are usually an outermost 5um layer of cobalt,
with a 20 um layer of nickel phosphorous. The platter is mainly
aluminum with a bit of magnesium, but the core of the platter is
covered by too much magnetic coating to see any characteristic xrays.
The phosphorus was being used to make the nickel nonmagnetic, and
perhaps the titanium and chlorine was being used in the zip disk in a
Hard Disk Head
SEM: Lastly, images of the read-write head of the solid disk were
imaged in the SEM. The image on the left shows the entire cross
section of one such head, where the contact is the visibly bright
area on the left of the image. The middle section of the image is
another magnetic coating using to read the magnetic field of the disk.
The right images shows the magnetic coating.
EDX Mapping & SEM Position: EDX mapping were taken of the
contact shown in the above image. The first two images show the
contact position and the SEM image of the spot where the mapping was
taken. The mapping shown below correspond to oxygen, lead,
aluminum, magnesium, iron, and silicon going from left to right.
The black contact clearly contains oxygen, magnesium, and iron,
while it is deficient in lead, aluminum, and silicon.
All Atomic Force Microscopy Roughness Comparison
The atomic force microscope software program was able to calculate the
surface roughness of the scanned regions. Where dust or
contamination was present, regions avoid the dust were chosen for
roughness calculations. Changing directions for scanning had very
little effect to surface roughness, even when the tip appeared to be
slightly defected. From the table, it is very clear that etching
increased the surface roughness by a factor of about ~2-3. It is
also clear that as the magnetic domain size decreased from floppy, to
zip, to hard disk, the surface roughness of the corresponding magnetic
coatings also decreased.
|Root Mean Square
I want to express my deepest thanks to Brian McIntyre, the course
instructor, for all of his help with the class work and graduate work
related questions. Without his help and guidance, it would have
been very unlikely that I got any data or results at all. I would
also like to thank my teaching assisntant Hanyuan Zhang for all his
help during the labs, and my labmate Congcong Wang who helped keep me
on time to class and lab this semester. Lastly, I want to thank
the Dean's office for making this course available to graduate
students. The knowledge and experience gained in this course,
will prove invaluable in the future.