ArticlePDF Available

Abstract and Figures

A new technique of magnetic imaging on a spin-stand [Mayergoyz et al., J. Appl. Phys. 87, 6824 (2000)] is further developed and extensively tested. The results of successful imaging of digital patterns overwritten with misregistration ranging from 0.3 to 0.07 μm are reported. The results are compared with magnetic force microscopy (MFM) images and the conclusion is reached that the spin-stand imaging technique can provide (at least) the same level of resolution and accuracy as the MFM imaging technique. © 2001 American Institute of Physics.
Content may be subject to copyright.
Spin-stand imaging of overwritten data and its comparison with magnetic
force microscopy
I. D. Mayergoyza) and C. Tse
Electrical and Computer Engineering Department, University of Maryland, College Park, Maryland 20742
C. Krafft
Laboratory for Physical Sciences, College Park, Maryland 20740
R. D. Gomez
Electrical and Computer Engineering Department, University of Maryland, College Park, Maryland 20742
A new technique of magnetic imaging on a spin-stand Mayergoyz et al., J. Appl. Phys. 87, 6824
2000兲兴 is further developed and extensively tested. The results of successful imaging of digital
patterns overwritten with misregistration ranging from 0.3 to 0.07
m are reported. The results are
compared with magnetic force microscopy MFMimages and the conclusion is reached that the
spin-stand imaging technique can provide at leastthe same level of resolution and accuracy as the
MFM imaging technique. © 2001 American Institute of Physics. DOI: 10.1063/1.1359233
It has long been recognized that the imaging of magne-
tization patterns recorded on hard drive disks is a source of
valuable information that may enhance our understanding of
recording processes and assist in the design of new recording
systems. The magnetization imaging is routinely performed
by using magnetic force microscopy MFM.1It has been
realized that MFM has the following intrinsic limitations: 1
low rate of image acquisition, 2special requirements for
the preparation of the sample to be imaged, and 3virtual
impracticality of fast accumulation of numerous images of
the same target area in order to increase the signal-to-noise
Recently, a new technique of magnetic imaging on a
spin-stand has been developed.2In this technique, raw image
acquisition is performed by scanning a target area of a hard
drive disk by a magnetoresistive MRhead in the along- and
cross-track directions. Scanning in the along-track direction
is realized due to the rotation of the disk, while scanning in
the cross-track direction is achieved by using very small and
accurately controlled radial displacements of the head. As a
result of this scanning mechanism, the spin-stand imaging
technique has the following advantages over conventional
MFM imaging: a high rate of image acquisition, increase in
the signal-to-noise ratio due to multiple imaging of the same
target area, and performance of imaging under similar con-
ditions as in conventional hard disk drives. However, due to
the nonlocalized nature of the magnetoresistive head in the
cross-track direction, the collected images can be quite dis-
torted. In addition, the collected images are scalar in nature,
while magnetization distributions are vector fields. For these
reasons, the collected images must be treated as raw images,
and image reconstruction is needed in order to retrieve the
actual magnetization distributions from the raw images. The
image reconstruction technique is based on the response
function characterization of the MR reading element and can
be described as follows.
The position of the scanning MR element can be identi-
fied by the xand ycoordinates of its center. The recorded
magnetization distribution can be characterized by the
equivalent distribution of virtual magnetic charges
where his the thickness of the recording media and it is
tacitly assumed that the recorded magnetization is uniform
over the media thickness and, for this reason, divMhas the
meaning of ‘‘surface’’ divergence.
The distribution of virtual magnetic charges is equiva-
lent in the sense that they create the same magnetic field as
the actual magnetization distribution. This magnetic field
causes the signal collected by the MR element. This signal
can be viewed as the superposition of the signals due to the
elementary magnetic charges distributed over the disk sur-
face. The last assertion can be mathematically expressed as
Here S(x,y) is the signal of the MR element, while
R(xx,yy) can be interpreted as the response function
of the MR element. This function has the physical meaning
of the signal induced in the MR element at position (x,y)by
the point unit magnetic charge located at position (x,y).
In order to experimentally determine the response func-
tion, an isolated sharp transition is first written. This transi-
tion is then trimmed by using dc erasure on both sides of the
same track. As a result, a ‘‘tiny’’ isolated spot of magnetic
charges is written that can be viewed as an approximation to
a point charge. The MR reading element can now be used to
measure the signal as a function of relative position with
respect to the recorded ‘‘point’’ magnetic charge. This signal
can be interpreted as a scaled version of R(xx,yy).
An example of the spin-stand measurement of the response
function of the MR element is shown in the left plot of Fig.
1. To get the information about geometric dimensions of the
aElectronic mail:
67720021-8979/2001/89(11)/6772/3/$18.00 © 2001 American Institute of Physics
recorded tiny spots of magnetic charges, the magnetic force
microscope was used. An example of the MFM image of the
‘‘point’’ magnetic charge is shown in the right plot of Fig. 1.
It is worthwhile to mention that we were able to record and
measure the response function for such tiny spots of mag-
netic charges that their counterpart MFM images were found
to be elusive and could not be clearly observed.
Having determined the response function, formula 2
can be viewed as a convolution integral equation that relates
the raw image S(x,y) to the distribution of virtual magnetic
m(x,y) which in turn is related to the divM
through Eq. 1. There is no way to reconstruct the actual
vectorial field of recorded magnetization by using only for-
mulas 1and 2. To circumvent this difficulty, we have
used the known fact3–5 that only the curl-free component of
magnetization can be retrieved from MR measurements. This
is because the curl-free component of magnetization distri-
bution is the field producing part of the total magnetization
distribution. For this reason, only this component is sensed
by the MR element. The last statement can be best illustrated
by an example of a dc erased track. In this example, there
exists nonzero magnetization within the track, however, this
FIG. 2. ColorReconstructed image of F6 overwritten by F9: 0.15
FIG. 3. ColorReconstructed image of F6 overwritten by F9: 0.09
FIG. 4. ColorReconstructed image of F6 overwritten by F9: 0.07
FIG. 1. ColorSpin-stand image of a head response function of a MR
element left plotand a MFM image of a tiny spot of magnetic charges
right plot.
6773J. Appl. Phys., Vol. 89, No. 11, 1 June 2001 Mayergoyz
et al.
track does not produce any magnetic field. This is because
the magnetization within the track is divergence-free but not
Thus Mxand Myin formula 1are meant to satisfy the
additional equation:
y0. 3
Now, by using Eqs. 13, the measured response
function Rof the MR reading element, and the Fourier trans-
form technique, the curl-free component distribution of mag-
netization can be fully retrieved from the scalar raw image
S(x,y). The mathematical details of the reconstruction tech-
nique can be found in our previous work.2
The imaging technique described above has been imple-
mented and extensively tested using a Guzik model 1701 MP
spin-stand. The main emphasis has been on imaging of edge
areas of tracks overwritten with small misregistrations rang-
ing from 0.3 to 0.07
m. In our experiments, giant magne-
toresistive GMRheads produced by ALPSwith write
widths of 1.1
m and read widths of 0.7
m have been used.
First, F6 patterns hexadecimal F611110110 in binary no-
tationwere recorded and then they were overwritten by F9
patterns hexadecimal F911111001 in binary notation
with controlled misregistrations ranging from 0.3 to 0.07
m. The overwritten tracks were scanned and the collected
raw images were reconstructed. Figures 2, 3, and 4 show the
reconstructed images of F6 patterns overwritten by F9 pat-
terns with misregistration of 0.15, 0.09, and 0.07
m, respec-
tively. These figures show only the Mxalong-trackcompo-
nent of magnetizations. In order to emphasize the binary
nature of the patterns, the color contrast of the images of Mx
has been deliberately saturated. The artifact of this saturation
is that noise has also been enhanced, as can be seen from
Figs. 2, 3, and 4.
To assess the accuracy and the resolution of the de-
scribed spin-stand imaging technique, extensive comparison
of this technique with MFM imaging has been carried out.
The sample results of this comparison are shown in Figs. 5
and 6 for misregistrations of 0.3 and 0.15
m, respectively.
Any such comparison, however, should be carried out in the
context that MFM images represent magnetic charges of the
patterns while the reconstructed spin-stand images are the
magnetization distributions of the patterns. Still, it remains
apparent from Fig. 6 that remnants of the overwritten F6
patterns are barely visible on the MFM image, while on the
spin-stand image see Fig. 2these remnants are well-
pronounced with many interesting details. This comparison
suggests that the developed spin-stand imaging technique has
at least the same level of resolution and accuracy as the
MFM imaging technique and it is clearly superior to the
latter as far as the rate and conditions of image acquisition
are concerned.
1D. Rugar, H. J. Mamin, P. Guethner, S. E. Lambert, J. E. Stern, I. Fadyen,
and T. Yogi, J. Appl. Phys. 68,11691990.
2I. D. Mayergoyz, C. Serpico, C. Krafft, and C. Tse, J. Appl. Phys. 87,
6824 2000.
3R. Madabhushi, R. D. Gomez, E. R. Burke, and I. D. Mayergoyz, IEEE
Trans. Magn. 32, 4147 1996.
4I. A. Beardsley, IEEE Trans. Magn. 25, 671 1989.
5I. D. Mayergoyz, A. A. Adly, R. D. Gomez, and E. R. Burke, J. Appl.
Phys. 73, 5799 1993.
FIG. 5. ColorMFM image of F6 overwritten by F9: 0.3
m misregistra-
tion. FIG. 6. ColorMFM image of F6 overwritten by F9: 0.15
m misregistra-
6774 J. Appl. Phys., Vol. 89, No. 11, 1 June 2001 Mayergoyz
et al.
... 10,11 Spin-stand magnetoresistive microscopy is another magnetic imaging technique, which utilizes a conventional magnetic read/write head (RWH) of a hard disk drive (HDD) as a sensor and a spinning disk as the sample. 12,13 However, the requirement for laminar air flow during sample measurement, a) Author to whom correspondence should be addressed. Electronic mail: and the restriction to samples that can be rotated at a high angular velocities under the head, adds utilization complexity that limits this technique to samples with extremely flat surfaces. ...
... Please note that the obtained values are representing the full system resolution, including errors from write pole fringing fields and grain size distribution of the media. 12,13 Ultimately, the spatial magnetic imaging resolution of the SMRM is limited by the physical dimensions of the TMR sensor. ...
An advanced scanning magnetoresistive microscopy (SMRM) — a robust magnetic imaging and probing technique — will be presented, which utilizes state-of-the-art recording heads of a hard disk drive as sensors. The spatial resolution of modern tunneling magnetoresistive sensors is nowadays comparable to the more commonly used magnetic force microscopes. Important advantages of SMRM are the ability to detect pure magnetic signals directly proportional to the out-of-plane magnetic stray field, negligible sensor stray fields, and the ability to apply local bipolar magnetic field pulses up to 10 kOe with bandwidths from DC up to 1 GHz. Moreover, the SMRM can be further equipped with a heating stage and external magnetic field units. The performance of this method and corresponding best practices are demonstrated by presenting various examples, including a temperature dependent recording study on hard magnetic L10 FeCuPt thin films,imaging of magnetic vortex states in an in-plane magnetic field, and their controlled manipulation by applying local field pulses.
This chapter provides an introduction to hard disk data storage and retrieval. It begins by describing the technological evolution of hard disk drives. The issues related to data loss are subsequently discussed. This is followed by a survey of selected magnetic-microscopy techniques available for the imaging of hard disk data. A novel spin-stand-based magnetic-microscopy technique, which is free from the magnetic-force-microscopy (MFM) limitations, is introduced along with advanced spin-stand-based data-forensic methods. This technique is capable of producing high-resolution images of magnetization patterns in a rapid and convenient way. It offers several advantages—a high rate of data acquisition; a large image area; a good signal-to-noise ratio because of rapid accumulation of numerous scanned images; the absence of scanning-induced hysteresis; and imaging under conditions similar to those in actual hard disk drives.
Scaling of head response function in spin-stand imaging is proposed. This scaling is performed in order to improve the accuracy of the measured head response function. This response function is measured by imaging a small spot of magnetic charges formed as a result of dc-trimming of isolated transitions. The theoretical justification for the scaling is the ``nearly'' self-similar nature of the measured response function with respect to the cross-track dimension of the charged spot. This scaling technique has been tested experimentally. It is demonstrated that the scaling of head response function allows one to reconstruct magnetization images of overwritten data that cannot be discerned otherwise.
The spatial and vectorial characterization of thermal relaxation of recorded magnetization patterns by using the spin-stand imaging technique [I. D. Mayergoyz et al., J. Appl. Phys. 87, 6824 (2000); 89, 6772 (2001)] is reported. In order to obtain such characterization, a recorded track is scanned at successive instants of time over periods of 70 hours. As a result, the spatial distributions of the read-back voltages (that constitute the “raw” images of the track) are consecutively collected. The images of the vectorial magnetization are then reconstructed and local magnetization relaxation rates are subsequently evaluated. It is demonstrated that the spatially inhomogeneous and vectorial nature of thermal relaxation of recorded patterns may result in temporal track broadening. It is found that this temporal track broadening is more pronounced for disks with higher coercivities. © 2002 American Institute of Physics.
High-speed massive imaging of hard disk data by using the spin-stand imaging technique [I. D. Mayergoyz, C. Serpico, C. Krafft, and C. Tse, J. Appl. Phys. 87, 6824 (2000) and I. D. Mayergoyz, C. Tse, C. Krafft, and R. D. Gomez, J. Appl. Phys. 89, 6772 (2001)] is reported. In order to obtain these large-scale images of hard disk data, disks from commercial hard drives were scanned by a giant magnetoresistive head in the along- and cross-track directions. A special method of triggering has been devised to capture the data nondestructively. Challenges related to the eccentricity of the disk and the instability of the trigger have also been addressed. By using this massive imaging technique, we were able to image disk data with track densities as high as 60 000 tracks per inch. With a specially designed automated algorithm, the developed technique can be programmed to image the drive data of the whole disk surface with high resolution and speed. © 2003 American Institute of Physics.
In this paper, an advanced extraction technique of the response function of giant-magnetoresistive (GMR) head for use in spin-stand imaging is presented. The essence of this technique is the spatial averaging of readback signals of many identically written "point-like" charges resulted from dc-trimmed transitions. As a result of the averaging, the noise generated by random fluctuations of medium properties is filtered out. This extraction technique also restores the symmetry of the head response function by employing a nonstandard dc-trimming procedure and a modified bias condition for the read process. By using the extracted head response function, we were able to reconstruct magnetization images of overwritten data with a misregistration of 50 nm.
Defects in magnetic hard disk media reduce the performance of hard disk drives. A consequence of the increasing storage density is that submicrometer defects in the magnetic storage media become more relevant. However, finding and visual inspection of these defects under an optical microscope is problematic. Therefore, a spin test stand technique was used to reliably map the performance of relevant storage media defect positions, and a computer controlled technique was developed which positions those defects relative to a local analysis tool within 45 s. The accuracy and repeatability was measured to be ±0.05° in angular (approximately ±30 μm of a circular arc) and ±5 μm in radial direction, respectively. Here, surface analysis of defects has been performed using scanning force microscopy. Defect and disk features with heights < 1 nm have been measured routinely while defect diameters were in the range of 1 μm to 250 nm. In particular, nontopographic defects are visualized by means of magnetic force microscopy.
Conference Paper
In our paper we generalize and extend model of storing data on magnetic drives. This model allows to erase data from magnetic drives even in the presence of a very strong adversary that can read the old data overwritten arbitrary number of times. However, from the physical point of view the adversary can be even more powerful, e.g. in addition be can be able to determine the order in which bits were stored. Such an assumption in the case of very well equiped adversary trying to retrive important data can be realistic. For that reason we introduce an extended model of the adversary, that we called time-aware adversary model. We show that the solution for secure data deletion from is not suitable for the time-aware model, i.e. the adversary can reconstruct whole data stored on a magnetic drive with high probability. We investigate time-aware model and its properties and propose some solutions which (to some extent) allow provable deletion in newly introduced model.
Full-text available
User data is often unprotected on disk and tape drives or not erased when no longer needed, creating data security vulnerabilities that many computer users are unaware of. Federal and state laws require data sanitization, which comprises a variety of data eradication methods. Secure sanitization refers to methods meeting those federal and state laws. Companies that fail to meet these laws can be subject to fines of $5 million, and individuals can be imprisoned for up to 10 years. Physical destruction of storage devices offers the highest security. But executing the disk drive internal secure-erase command also offers a higher security level than external-block-overwrite software, according to federal guideline NIST 800-88. Recent disk drives with internal full disk encryption now implement an enhanced secure-erase command that takes only milliseconds to complete.
A new technique of magnetic imaging on a spin stand has been developed. In this technique, raw image acquisition is performed by scanning a target area of a disk by a conventional magneto-resistive (MR) read head in two orthogonal (along- and cross-track) directions. Due to the nonlocalized nature of the MR reading head in the cross-track direction, image reconstruction is needed in order to retrieve the actual distribution of magnetization. It is demonstrated that the image reconstruction can be performed by using the response function characterization of the MR head and the specially designed deconvolution technique, which yields the curl-free (field producing) component of magnetization. The technique developed is illustrated by the sample examples of imaging of overwritten tracks with small misregistrations. © 2000 American Institute of Physics.
The problem of magnetization image reconstruction from magnetic force scanning tunneling microscopy images is discussed. The reconstruction problem is reduced to some convolution‐type integral equations and analytical solutions to these equations are obtained by using the standard Fourier transform technique. The theoretical discussion is illustrated with some numerical examples.
This paper discusses the principles of magnetic force microscopy (MFM) and its application to magnetic recording studies. We use the ac detection method which senses the force gradient acting on a small magnetic tip due to fields emanating from the domain structure in the sample. Tip fabrication procedures are described for two types of magnetic tips: etched tungsten wires with a sputter‐deposited magnetic coating and etched nickel wires. The etched nickel wires are shown to have an apex radius on the order of 30 nm and a taper half‐angle of approximately 3°. Lorentz‐mode transmission electron microscopy of the nickel tips reveals that the final 20 μm is essentially single domain with magnetization approximately parallel with the tip axis. Images of written bit transitions are presented for several types of magnetic media, including CoPtCr, CoSm, and CoCr thin films, as well as γ‐Fe 2 O 3 particulate media. In general, the written magnetization patterns are seen with high contrast and with resolution better than 100 nm. A number of magnetic recording applications are discussed, including the investigation of overwrite behavior and the writing characteristics in CoSm media at high data density. Computer calculations were performed to simulate the MFM response to written magnetic transitions. By including the extended geometry of the tip, the nonparallel orientation of the cantilever, and the finite width of the magnetic transitions, good agreement with experiment was obtained. The model calculations correctly predict the experimentally observed change in image contrast that occurs as a function of tip orientation. Computer calculations showing the dependence of resolution on tip geometry are also presented.
A novel experimental technique of magnetic biasing for the Magnetic Force Microscopy (MFM) imaging is introduced. This technique enforces linearity of the imaging system and, in this way, it facilitates the interpretation of images of magnetization patterns. The technique of numerical reconstruction of the curl-free vector magnetization from a raw image obtained by the linearized MFM system is presented. An example is given which demonstrates the significant difference between the reconstructed magnetization and the raw image. This stresses the usefulness of magnetic biasing and image reconstruction for MFM image interpretation
A combination of two measurements is shown to be sufficient to determine the magnetization in a thin film except for a constant out-of-plane component, under the condition that the magnetization is uniform through the thickness. The measurements consist of DPC (differential phase contrast) Lorentz microscopy, which measures the path integral of the in-plane B field, and some measurements of the out-of-plane component of H above and below the film, for example by Hall probe. Field measurements on one side only are sufficient if the magnetization is in-plane. The in-plane magnetization is decomposed into a divergence-free part and a part with zero component of curl perpendicular to the plane of the film. The DPC measurement directly yields the divergence-free component, and the curl-free as well as the out-of-plane parts are found from a deconvolution of the perpendicular field. Several examples are given, and the interpretation of the two magnetization components in terms of B and H provides a qualitative description of the physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at
  • Mayergoyz
J. Appl. Phys., Vol. 89, No. 11, 1 June 2001 Mayergoyz et al. [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded to ] IP: