Self charging behaviour of electrolyte capacitors

During my physics lab this afternoon, I found that discharged electrolyte capacitors re-charges to a percentage of its original voltage when left alone as open circuit. I was informed this effect is called dielectric relaxation by one of the tutors, but no explanation was forth coming.

So I did some research and here is what I have found:
"A capacitor exhibiting dielectric absorption acts as if during its long precharge time the dielectric material has soaked up some charge that remains in the dielectric during the brief discharge period. This charge then bleeds back out of the dielectric during the relaxation period and causes a voltage to appear at the capacitor terminals. Fig 2 depicts a simple model of this capacitor: When 10V is applied for 1 min, the 0.006-µF capacitor gets almost completely charged, but during a 6-sec discharge period it only partially discharges. Then, over the next minute, the charge flows back out of the 0.006-µF and charges the 1-µF capacitor to a couple of dozen millivolts." [http://www.national.com/rap/Application/
"An electrolytic capacitors is, amongst other things, an electrochemical
cell, and as such can store energy as polarization of the electrolyte/electrodes, and as chemical change. Discharging the surface charge briefly does not release this stored energy, which will subsequently give rise to a terminal voltage as the cell settles back to equilibrium. " [http://www.du.edu/~jcalvert/phys/caps.htm#Diel]
The following are posts to sci.physics newsgroup thread regarding the same topic:
"The way I remember it was that the capacitors had TWO "neutral" states: electrically neutral and mechanically neutral.
Upon discharge the capacitor plates were electrically neutral, but as they mechanically relaxed the diaelectric became polarized and thus produced a charge difference on the plates."

"The situation in electrolytics is complicated by the active nature of the electrolyte, i suspect. If this is correct, an initial zero volt indication merely means that the charge closest to the electrofes has been bled off. If investigating, leaving a milliameter (other wise known as a "calibrated short sircuit... 8)>>) and reading it at intervals (or equivalent datalogger, for the instrumentally endowed...) would be instructive."
The same thread also offered alternatives to the dielectric relaxation explanation
"What you have experienced is common- you have to think about what you have there, an ELECTROLYTIC capacitor. Electrolytics contain "electrolyte", a fluid dielectric betweenn the plates. This fluid is not pure, and it is the impurities that
cause the "battery" action you have noticed. Short the pins for a couple of days and THEN retry. Chemically the impurities in the oil will neutralize and the capacitor will no longer "charge itself up". The action is similar to a very poor battery (try
sticking dissimilar metal nails into an orange and measure the voltage across the nails- you should get about half a volt or more)."

"If one can buy electrolytics with pure electrolyte, and plates made of pure aluminum, then such reactions would not be seen. However, traces of copper and other metals in the plates as small as they may be, are responsible for this "battery" action. As long as industrial grade metals and electrolyte oils are used, we will constantly see this effect...............sq"
It would seem there are two schools of thought on this subject:
  1. Dielectric relaxation is responsible for the self-charging behaviour of electrolyte capacitors.
  2. Chemical reaction similar to that of a battry causes the self-charging behaviour of electrolyte capacitors.
Personally the results of the "memory" effects of capacitors observed supports the theory of dielectric relaxation being responsible. If the effect is due to a reversible chemical reaction, then by conservation of energy the longer the capacitor is held at a certain potential, stronger the self-charged voltage.

The following is my own uninformed and most likely wrong explanation of why this happens, as I interpret it:
When a capacitor initally charges its dielectric material's structure is "stressed" or polarised due to the presence of the electric field. This causes a small amount of energy to be stored in the inter-molecular bonds with in the dielectric material. When the capacitor discharges rapidly, the dieletric material's structure is still under stress even though the potential across the terminals is 0 due to its slower response. As it reverts back to its normal, minimal-energy state in the absence of an electric field, a small increase in electric potential develops across the capacitor's terminals after some time. If the dielectric material is held by an electric field for a long period of time the internal structure will re-arrange graduly itself as to achieve minimal-energy under influence. Thus when discharged after a long time with in an electric field the self-charging effects are smaller as a lower amount of stress is present in the molecular structure.
Should any one more knowledgeable stumble on this topic, please correct any mistakes I have made.