[geeks] jackassery from the BBC

[Phil Stracchino]

der Mouse wrote:
>> I heard somewhere that SEM doesn't work on the perpendicular write
>> drives anymore anyway because the data density is a higher resolution
>> than a SEM can resolve.  Does anyone know if there is any truth to
>> this?
> 
> I for one don't believe it until I see something firmer than "I heard
> somewhere".  SEMs, if I'm thinking of the right device (maybe I'm
> thinking of atomic force microscopes?), can resolve individual atoms,
> and if the magnetic domains are smaller than that then (a) we the disk
> customers being grossly cheated on storage density and (b) I want to
> know what's storing the data.

That would be atomic force microscopes, yes.  A conventional scanning
electron microscope can resolve surface details as small as 1 to 5
nanometers.  Scanning tunneling microscopes, the direct ancestor of the
AFM, can give resolution as fine as 0.1nm, with 0.01nm depth resolution,
but I don't know if an STM or an AFM can be used to map magnetic
domains.  I suspect it is slightly more likely to be doable with an STM
than an AFM simply by virtue of the mechanisms involved.

Toshiba released 1.8" disk drives in 2004 with an areal density of 177
to 200 Gb per square inch; Hitachi demonstrated perpendicular recording
at areal densities of 230GB/in^2; and Seagate demonstrated 421 Gb/in^2
in 2006, using then-current production equipment.  It's currently
estimated that perpendicular recording can be scaled to around 1Tb/in^2.

Seagate's demonstration was over two years ago.  If we make an
assumption that 400Gb/in^2 is possible right now in a production
commercial device, that's around 6.2x10^14 bits per square meter, or
roughly 6x10^-4 bits per square nanometer, or alternately, around
1610nm^2 per bit.  From there it's trivial to calculate that if we
assume a square magnetic domain, we require approximately a 40nm square
per bit.  40nm physical features are well within the resolution
capability of a SEM.  Even the 1Tb/in^2 estimated limit of perpendicular
recording only reduces that domain size to around 25nm, while Toshiba's
2004-vintage 177Gb/in^2 drives should feature around 60nm domain sizes.

Now, that said, I don't know precisely how a SEM resolves magnetic
domains, or whether it can resolve magnetic domains as finely as it can
resolve physical features.  Searching for articles that would clarify
this question, I was able to find this paper, copyrighted 2008,
describing work apparently done principally at Brookhaven National
laboratory:

http://cat.inist.fr/?aModele=afficheN&cpsidt=1634956

I quote the following partial paragraph of interest:

"The optimum scanning electron microscope imaging conditions are
attained with an incident-electron energy set at 5 to 6kV, which
produced images with resolution on the order of 1N<m. The simplicity of
the technique and the ready adaptability of the SEM to such
modifications as in situ current and magnetic field application suggest
the extension of these to investigations of other materials of
technological interest, such as perpendicular media disks."

This seems to imply that the current limit on resolution of magnetic
domains using a SEM is on the order of one micrometre, which is not
sufficient to resolve the 40nm magnetic domains of Seagate's 400-plus
Gb/in^2 disks.  It also seems to imply that the authors believe their
techniques may be used to scan perpendicular-domain disks, though given
the resolutions they cite being able to achieve, it seems unlikely that
a SEM using these techniques would suffice for any perpendicular-domain
device using Toshiba's 2004 technology or newer unless the resolution
for magnetic domains could be improved by at least 1.5 orders of
magnitude (and a full two orders would be better).



-- 
  Phil Stracchino, CDK#2     DoD#299792458     ICBM: 43.5607, -71.355
  alaric at caerllewys.net   alaric at metrocast.net   phil at co.ordinate.org
         Renaissance Man, Unix ronin, Perl hacker, Free Stater
                 It's not the years, it's the mileage.