FW: [SI-LIST] : Dielectric loss (Intuitive explanation)

Hemant Shah (hemants@xynetix.com)
Fri, 19 Sep 97 16:06:00 EDT

OK, here it is! (hang in there, its a little involved.)

First, let me supply the missing "intuitive explanation" for polarization
(including
the atomic mechanism) and how a dialectric behaves and how this relates
to
polarization.

Stripped of all the fancy talk, a dielectric material is any substance
which
resists the penetration of an electric field into its interior. Imagine
a plate of
a dielectric within which you are attempting to establish an electric
field by
piling negative charged along one side, and positive charges along the
other. The dielectric will resist this by magically producing opposite
charges
at its surfaces wherever you are applying your charges - the more charges
you
pile on, the more the dielectric will counter with its opposing charges.
The
effect of this is to cancel out some of the fields which would otherwise
have been
produced inside the dielectric. In fact, the dielectric constant of a
material is
DEFINED by the magnitude of this reduction - it is the ratio of the field
which
would have been produced without these opposing charge to the field which
is
actually produced. For example, if only one third of the expected field
occurs
within the dielectric, then its dielectric constant is 3. Obviously, the
dielectric constant of a vacuum is 1.

If you think about it, this explains why a capacitor works. As we pile
charges into
one end and take them out of the other, the dielectric inside the
capacitor is busy
piling up cancelling charges right next to the ones we put in, and it's
doing this
at both ends of the capacitor. These prevent a voltage from being
established
across the capacitor - it looks to the outside observer as if the
capacitor were
"gobbling up" the charges (storing them.)

OK now, how (from an atomic viewpoint) does a dielectric do this? All
substances
consist of atoms, which in turn contain a positively charged nucleus and
an equally
(but negatively) charged coterie of electrons. In a conductor, the
electrons are
free to swim about (almost like a fluid) within the crystal lattice. In
a dielectric,
however, the electrons are bound to the nuclei. However, in the presence
of an
electric, field, they may shift slightly to one side (this is a slight
simplification,
but to go any deeper requires a considerable amount of Quantum
Mechanics.) This
tiny amount of shifting (very small even, in comparison to the size of
the atom)
takes place for every atom within the dielectric, and has the effect of a
bulk
displacement of all the positive charge toward the negative pole of the
applied
field, and visa versa for the negative charge. In fact, the total amount
of charge
is any common substance is simply astounding - we are unaware of this
because the
positive and negative charges are always in near perfect balance.
Because of this,
an exceedingly slight shift can cause a significant total effect. It is
very much
like what happens on a soccer field - when the ball is reversed from the
right side
of the field to the left, and every player takes two or three steps to
the left,
the net effect is the same as if ONE player was moved from the right side
of the
field completely over to the left. The same thing happens in a
dielectric - the
net effect of the shifting is the same as if some charge were transferred
from
one side of the sample to the other (positive charge magically appears on
one
surface, and negative on the other.) This shifting is called
"polarization".

All right, how does "loss" come into all this and what does it have to do
with
resonant frequency of polarization? It all has to do with time delay
(doesn't
everything?) When a field is applied, the aforementioned shifting cannot
happen
instantaneously - there must be SOME lag (the resonant frequency of
polarization
is related to this time delay.) Thus, if you apply a rapidly reversing
field
to a dielectric, there will be a phase lag between the applied field and
the
resulting polarization. As the frequency of the applied field is
increased, the
absolute lag is constant, therefore, the phase lag increases - when it
reaches
pi/2 (90 degrees) we are at resonance.

OK, so there is a frequency dependant phase lag between applied electric
fields
and polarization - so what? Well, as we have see, the polarization is
the
mechanism by which a dielectric resists the application of a field, and
it has
the effect of making the dielectric look like it is "gobbling up" charge
(a.k.a.
drawing excess current.) This means that there will be a frequency
dependant
phase lag between applied electric fields and the resulting current
flowing in
to establish those fields. If you know something about electric power
systems
and have heard of the term "power factor", you will know exactly what
this
means - there is POWER flowing into the system (it is absorbing energy.)

Alright, so you're not a power engineer - I'll give you a molecular
explanation
for what is going on. Let's imagine a dielectric sample to which we
apply one
half cycle of a very high frequency electric field (just a single short
electric pulse, really.) As soon as the pulse arrives, all the electrons
start shifting over toward the positive side of the field, but the pulse
is so
short that by the time they have just gotten into motion, the field
disappears.
Well, now we have a whole bunch of electrons which have acquired some
kinetic
energy which has to go somewhere. In fact, as the electrons rattle
around to
return to their normal equilibrium state, this energy will get
transferred into
to crystal lattice as vibrational energy (heat.) Note that the reason
this
all happened is that the pulse was so short that the electrons couldn't
track it
fast enough (phase lag.) If the electric field were continuously
oscillating
instead of a single pulse, this effect would happen continuously - the
electrons'
shifting would be out of phase with the applied electric field. I can't
get into
the mathematics of "out of quadrature" phase (sorry, it's physics), but
it boils
down to energy being pumped into the electrons which must eventually be
dissipated into the crystal lattice, causing heating.

Robert S Holmes
Chief Technical Officer
Xynetix Design Systems, Inc.

----------
From: Hemant Shah
To: Robert Holmes
Subject: FW: [SI-LIST] : Dielectric loss
Date: Friday, September 19, 1997 9:37AM

A question for you (Dr.) Bob!

Hemant

----------
From: owner-si-list
Sent: Thursday, September 18, 1997 1:22 PM
To: si-list
Subject: [SI-LIST] : Dielectric loss

The term "lossy dielectric" implies an energy loss or joule heating
in
the dielectric material. Could someone out there explain the actual
loss mechanism? Is is a paramagnetic or molecular vibration?, or?..
I
understand the "loss tangent which is the ratio of the e" to e'
(imaginary over real) but that does not explain the physical
mechanism
of the energy conversion. My book (quote: "it is obvious that the
imaginary part is associated with power loss or dissipation within
the
dielectric"...) well, that is not so obvious to me! The book further
implies that the loss is associated with the resonant frequency of
polarization. The book, EM waves and radiating systems by E. Jordan,
eloquently discusses "electronic polarization", and "microscopic
bodies that contribute to the polarization effect".. but does not
actually give a intuitive explanation.
Thanx in advance
Hans Mellberg