Very well written, to the point.

Mahendra 

-----Original Message-----
From: EmbeddedNet [mailto:[log in to unmask]] On Behalf Of Richard
Ulrich
Sent: Tuesday, March 07, 2006 6:10 AM
To: [log in to unmask]
Subject: [EM] Circuitree Embedded Passive Column (DRAFT)

Here's a draft of my next column on embedded passives.  Any comments
would be appreciated.

- Rick

for "Embedded Passive Update" column, CircuiTree, May 2006


High-Capacitance Thick Film Dielectrics?

Rick Ulrich

In my last column, two months ago, I bemoaned the sad state of affairs
with screen-printable thick film dielectrics.  Even when packed with
95%+ by weight with randomly-dispersed high-k barium titanate and
printed to only 10 microns thick, the overall k is a mere 30 and
delivers only about 3 nF/cm2.  One way of increasing the capacitance of
this approach would be to somehow stack the high-k phase so that it is
vertically continuous over as wide a lateral area as possible.  Some
kind of self-assembly or self-orientation would be necessary, and I know
of work going on in that direction.  But there's another trick that's
been receiving new attention lately that may be able to significantly
increase the capacitance density of screen-printable thick films.  The
concept is pretty simple: load the polymer up with conductive metal
particles, as opposed to the  
insulating high-k particles that I described in the previous column.   
Rao Tummala, C. P. Wong, and their team at Georgia Tech have been
exploring this lately and have managed to get effective dielectric
constants of over 100 by dispersing Ag and Al particles in epoxy.

To understand how this works, think about drilling a vertical hole one
molecule wide all the way through the dielectric.  The materials and
their configurations encountered along the way are what the field sees.
In the case of randomly-dispersed 10 micron k = 1000 barium titanate
particles scattered in k = 3 epoxy, this hole would pass through
alternating regions of high and low k material, which the field would
see as capacitors in series.  The result is an overall k in between the
two, but much closer to the low end at about 30 since adding caps in
series lowers the overall value of the set.  But, if the dispersed
particles are instead conductive, then all of the voltage drop is over
the epoxy phase alone.  The hole might pass through 10 microns of total
applied film thickness, but only a small fraction of that will be actual
dielectric, depending the concentration of conductive particles that
were mixed in.  In essence, this is a clever way to make a very thin
polymer film that is easily screen printed.  As you approach the
percolation point, the minimum concentration of metal particles that
causes them to start touching each other, the effective film thickness
is very small and high specific capacitance can be obtained.

Using 80 wt% of 3.0 micron Al in epoxy, they've managed to get an
effective k of  109 and a dissipation factor of only 2%.  I say
"effective k" because the material that is actually acting as the
dielectric has the same k it had in bulk form, about 3, but is spread
very thinly.  The value of 109 is based on the entire film thickness so
this is the k it acts like.  Printed to a total thickness of 10 microns
gives a specific capacitance of almost 10 nF/cm2, a factor of three
better than high-k dispersed in epoxy and better than 30 times what
you'd get from unfilled epoxy.  This latter number also represents the
ratio of the printed film thickness to the effective thickness.  The
downside to this method is that you have to be very close to the
percolation point to get these high values and, in this vicinity, the
effective k and the dissipation factor change rapidly with particle
loading, making tolerance a major challenge.  Once the particles reach
percolation and actually begin touching, the dissipation factor
skyrockets since the film becomes leaky.  Much past this, it would not
function as a dielectric at all, so you need to be just a fraction of a
percent under percolation for best results.  But Georgia Tech is working
on getting around this issue by using metal particles that have an oxide
coating so that even when they touch, there is no electrical continuity.

The primary application for this would be decoupling, since tolerance is
not a major issue as long as a certain minimum amount of capacitance is
achieved.  Also, dissipation factor is not nearly so much a detriment in
decoupling as it is in filters and A/C converters.  In fact, some folks
claim that a high dissipation factor is beneficial to decoupling since
it helps remove high-frequency noise on the power/ground planes.

Their results will soon be published in IEEE Transactions of Advanced
Packaging along with a considerable amount of information on the
processability of this sort of material.  Much work remains to be done,
but it's an interesting and promising approach that could eventually
give tens of nF/cm2 with a screen-printable film.



******************************************************
Dr. Richard Ulrich
Professor
Dept. of Chemical Engineering
3202 Bell Center
Univ. of Arkansas
Fayetteville, AR   72701

[log in to unmask]
(479) 575-5645
******************************************************