On the possibility that problem of the stickiness of the soil might be
due to the magnetic particles, the Phoenix team has done some
experiments with Mars soil simulants that contain some proportion of
magnetic particles so you would think this problem would have already
been seen in these simulations if that were the reason. Perhaps though
there was a higher proportion of magnetic particles than expected in
the actual Mars soil.
I did a web search on how you can demagnetize a permanent magnet and
found this:

Magnet.
5 Magnetization and demagnetization.
"Permanent magnets can be demagnetized in the following ways:
Heating a magnet past its Curie point will destroy the long range
ordering.
Contact through stroking one magnet with another in random fashion
will demagnetize the magnet being stroked, in some cases; some
materials have a very high coercive field and cannot be demagnetized
with other permanent magnets.
Hammering or jarring will destroy the long range ordering within the
magnet.
A magnet being placed in a solenoid which has an alternating current
being passed through it will have its long range ordering disrupted,
in much the same way that direct current can cause ordering."
http://en.wikipedia.org/wiki/Magnet#Magnetization_and_demagnetization

The TEGA instrument will heat the sample to high temperature, but of
course this can't be used to remove the magnetization if you can't get
the sample to the instrument in the first place.
The second method of demagnetizing could conceivably work, but I
doubt the Phoenix lander has a magnet that could be passed over a soil
sample. A variation on this might be to rub and mix around the sample
on itself, then the different magnetic orientations on the particles
might tend to cancel each other out.
The third possibility might also be feasible by striking hard on the
sample to demagnetize it. The scoop might be used for this purpose. My
guess though is that you would have to place the sample on a hard
surface to do this. A flat metal surface on the lander would work but
there might be a worry that this could jar the landers internal
electronics by doing this. Perhaps the sample could be placed on top
of a hard rock that the robot arm could reach.
I don't think the fourth method mentioned of passing an alternating
current over the sample is feasible since I doubt the lander has the
capability of doing this to a sample that is outside the lander.

Bob Clark