[FoRK] Annihilation from within...

Dave Long < dave.long at bluewin.ch > on > Mon Dec 11 03:33:20 PST 2006

> BTW, the point (which I failed to make in the last message) is this
> --- it seems to me that if you can make the SQUIDs small enough and
> cheap enough

SQUIDS *are* small (if not cheap).  At least the last ones I worked 
with (back in the 80's) were tiny.  It was the huge dewar and fully 
mumetal-shielded room[0] that made things bulky.[1]

> and have them work at room temperature,

If superconduction is the only reason for the cryostat, that may be 
relatively easy to arrange, but my understanding of squids is that they 
count flux quanta, so the other reason to keep things cold is to keep 
transitions rare enough that thermal noise doesn't out-slew the 
counter.  (but see [2] infra for an indication that it's become much 
easier to keep a lock over the last decade or two)

There are plenty of hits for "room temperature squid", which seems to 
be an advance in the sense that the sample no longer needs to be dunked 
in the dewar -- it can remain at room temperature, but the squid itself 
still needs to be cooled.  (this technique seems to be behind the 
process for the low-field NMR/MRI I'd wondered about upthread, but 
unfortunately having the insulation between the sample and the sensor 
tends to add to the read distance)

>> Shielding and other environmental and power considerations are
>> admittedly thornier...
> As far as I can see, an array of NEMS atomic magnetometers would
> very nicely fit into a skullcap, requiring no particular shielding
> or power requirements below what's available for mobile applications.

Without shielding, you're trying to pick up 100 fT variations while 
sitting in a 50 uT field.  1e-8 is not the strongest of signals.  My 
guess as to why the pictures in the PDF only show the device fitting 
over the top of the head is that the rooms in which they are installed 
have been wrapped in mu-metal[2] (not the most mobile of solutions).

Given all the above objections to squids, I'll agree that atomic 
magnetometers have more promise, subject to the following questions: 
(a) do they have the resolution for MEG? (b) are they any easier to 


:: :: ::

>> OK, so high-rez says to me "steep spatial gradients", and as Eugen 
>> points out, real-time says "steep temporal gradients".  Given that 
>> low-power is all about "shallow gradients", how is materials science 
>> supposed to beat the physics?
> There's no conflict here --- we're not talking physics problems, we're 
> talking engineering problems, specifically miniaturization problems.

So how do you propose to miniaturize a waveform (increase its 
resolution by increasing its curvature) without increasing the power 
requirements?  Isn't power pretty much inverse wavelength?

:: :: ::

[0] kind of like for an MRI, one didn't take metal objects into the 
squid closet -- but failure to observe that precaution would only ruin 
your data, not your day

[1] in general I work with instruments, that one I worked in.

[2] it appears that DSP can help here:
> With DFC, the residual magnetic field at the SQUID can be kept close 
> to zero even if the device is moved in the Earth's field. Therefore, 
> the noise level of a high-T[c] magnetometer measured inside a 
> magnetically shielded room (60 fT Hz[-][1][/][2] with a 1/f corner at 
> 2 Hz) remained unchanged after moving the device in the magnetic field 
> outside the room (60 μT dc plus 0.8 μT peak-to-peak power line 
> interference).
as can fabrication techniques?
> Using SQUIDs fabricated with precise slots and holes, researchers have 
> reduced the problem of intrinsic magnetic noise, previously a major 
> impediment to the application of high-Tc SQUIDs in the earth's 
> magnetic field, to an artifact of design.
(where high-Tc is still about 77 K, or -320 F)

More information about the FoRK mailing list