Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: dielectric constant of water



What is the conductivity of the water used? Unless it is high purity it
may be acting more like a conductor than an insulator, and the
capacitance determined by the gaps between water and plate.

On Friday, August 22, 2003, at 11:53 AM, Wolfgang Rueckner wrote:

Greetings -- I'm having difficulty understanding the results of a
seemingly simple demonstration experiment and could use your
insights. The demonstration is meant to show the effect of placing a
dielectric between capacitor plates and deducing the dielectric
constant from the change in capacitance. Let me begin by describing
the apparatus/procedure and then I'll tell you about the perplexing
results.

The apparatus consists of two parallel aluminum plates (8"x10") with
adjustable separation. The experiment I've been performing is with a
1/2" separation, well within the linear range (that is to say, the
capacitance goes down linearly as the separation is increased up to
about 2" and then starts to deviate from this ideal behavior). The
capacitance is measured with an impedance meter operating at 120 Hz
and 1 kHz -- different manufacturer models give the same results - I
haven't tried a DC measurement yet (i.e., measurement of RC time
constant in charging). When 1/2" thick slabs of some dielectric
material are inserted between the plates, the capacitance goes up and
the dielectric constant is determined by the factor of increase.
Materials such as phenolic, PVC, acrylic plastic, and polyethylene
yield results that are in excellent agreement with values found in
the literature.

Using this set-up, we try to measure the dielectric constant of
water. Our water-filled plastic cell (which gets inserted between
the two capacitor plates) has never given satisfactory quantitative
results -- the increase in capacitance is not as large as expected
and I have always waved this result away by saying we really have two
dielectrics in series inside the plates (the plexiglass cell walls
and the water). In trying to improve the experiment, I've placed a
zip-lock type plastic bag between the two plates and filled it with
water. It works really well - too well - I get a surprising number
for the dielectric constant of water = 130. That's measured at 1 kHz
and 120 Hz (no difference). The data I have found on water's
dielectric constant is

87.0 at 10^5 Hz
78.2 10^6
78 10^8
76.7 3x10^9
34 2.5x10^10

As you can see, it drops off quite a bit once you get into the
gigahertz range, but it doesn't change too much at lower frequencies.
I haven't found values in the DC to 1 kHz range, but I suspected my
value of 130 to be too high. The water is at room temperature. I
detect no difference between Cambridge tap water and distilled H2O.
The plastic bag is placed between the plates (upon which the
capacitance increases by a trivial amount) and then filled with
water. The water filled bag does protrude slightly beyond the edges
of the plates but, in moving it around a bit, I have determined that
edge effects are negligible. The numbers are:

capacitance with air between plates = 0.035 nF
capacitance with empty bag between plates = 0.035 nF
capacitance with bag full of water = 4.560 nF, which implies a
dielectric constant of 130 for water

Next I inserted an actual (trade-mark written on it) gallon-size
zip-lock bag between the two plates and found the capacitance to go
up to 7.650 nF, a factor of 219!

The plastic of the two bags looks and feels the same. The thickness
of the zip-lock is half the other (zip-lock is 0.05 mm thick plastic
whereas other bag is 0.10 mm thick), but neither of them alone affect
the overall capacitance to any significant amount. I have repeated
the experiment with a generic plastic grocery bag (I think they're
made from recycled plastic and probably have a mixture of plastics)
whose thickness is only 0.02 mm. The final capacitance went up to an
amazing 18.6 nF, implying a dielectric constant of 530!!

What's going on here? It seems that the thinner the bag material,
the further I'm off from my expected results. Granted, the data
isn't clean since I have at least two different plastic bag materials
and different thicknesses - I shall need to keep one of these two
constant in further experimentation. One thought was that the
plastic happens to be a polar molecule and the polar water molecules
will be aligned at the inside surface of the bag, giving a
polarization effect with bag alone (never mind putting the whole
shebang in a capacitor). But this polarization would be in the
opposite direction on the opposite wall of the bag and the
polarization effect at the water/plastic surfaces would cancel out.
Yes?

Finally, I just tried the thick bag (0.1 mm) with about half the
capacitor plate spacing (approx 5/16") and measured the following:

capacitance with air between plates = 0.057 nF
capacitance with empty bag between plates = 0.058 nF
capacitance with bag full of water = 5.0 nF, which implies a
dielectric constant of 87 for water!!!
This shows that one can arrange the experiment to fit the theory.

Your thoughts?? -- Wolfgang

Dr. Vern Lindberg
585-475-2546
http://www.rit.edu/~vwlsps