Dear Orsino,

In a gas, also at low temperature, some free electrons are present, due to the cosmic ray radiation. When the electric field is applied, new electrons are produced by the applied field. If we neglect the losses, the electron density will increase exponentially. however, the frequency of growing depends on the the number of electrons. If you consider also electron losses due to recombination, diffusion and so on, and the loss of electron energy due to the elastic or excitation collision, you have loss characteristic times, some depending on the electron density (as recombination), and some only on heavy particle density (and temperature also). Therefore if the initial density is not high enough, for the given applied field, the losses are higher than the gains, and therefore electron density does not grow.

Another effect can be a decreasing of the field as the current grows, and if you are only a little higher than the threshold limit, you cha easily goes below such a value. 

Dear Mr. Filho, I am afraid that PD are always present, even if they are not significant and much lower than the accepted by IEC specs. Besides, as Mr. Colonna reffered, some free electrons are present due to the cosmic radiation always. In vacuum switch gears (in MV or HV applications) theoretically there are no charged particles, in fact there are not particles at all, and yet PD can take place, although not so easy as in normal pressure gases. However, to deal in practice with your question, PD are not take place (they are insignificant), when E is above a critical value and at the same time there are not charged particles (or there only a few of them)

Dear Orsino.

Adding to what has been said by colleagues.

I find your question interesting. The partial discharge in gases is a type of an avalanche breakdown where primary electrons are sufficiently accelerated by the applied electric field and get kinetic  energy equal or greater than the ionization energy of the gas atoms or molecules. This electric field is the critical electric field defining the dielectric strength of the material. So, to occur primary electrons must be available. 

It is so that the presence of proper radiation can ionize some atoms in the ensemble and or the metallic electrodes of the discharge tubes may emit electrons to the outside of the metal surface by thermal energy or by incident radiation or by field assisted tunneling at high electric fields. So, one can say that the condition of the existence of the primary electrons are piratically realized.

From the precision point of view, you are right.

MY BEST Wishes. 

Dear Gianpiero Colonna, A.A.P. Paraskevopoulos and Abdelhalim Zekry,

Many thanks for your answers and comments. I enjoyed them very much and they added significant knowledge to the subject.

Complementary to the question, I would like to pose a situation where a cavity is located deep inside a solid insulating body far from electrodes in a high voltage equipment. Such situation I had in mind when I thought about this question. It is a condition where a relative small volume of gas is confined with very low or none availability of electric charges, but with great possibility of being stressed by high local electric field. Should we worry about PD in such case?

Dear Dr Borges Filho,

The answer to your question "Is there any real condition ... charges? " results in a clear no. This question can be answered in the affirmative only under the purely unphysical condition represented by a lack of initiatory electrons (ambient ionization). Even neglecting cosmic radiation, (very-high-electric) field effect and some thermal mechanisms taken somewhere into account in the previous valuable responses, radon exhalation becomes, by elimination, the cause of the required ambient ionization. By the way, consider that even moisture is a common (often underestimated) vehicle for radon exhalation.        

The other mechanism of breakdown is field emission. Field emission does not need primary electrons. It is so that the electric field becomes high enough to torn electrons out of the materials by breaking the bonds of the valence electrons to the parent atoms.

But in this case the critical electric field may be much higher than that of the avalanche effect. These breakdown phenomena are very demonstrative in the breakdown of the depleted layers of the pn  diode and in electron emission in vacuum tubes. The idea of the cold cathode is based on electron emission from the surface of the material by applying very high electric fields. This effect is known quantum mechanically as tunneling. I want to point out in case of field emission the critical electric field will be much than that of the avalanche.

Dear Mr. Orsino Borges Filho,

In smal cavities inside solid or liquid dielectrics it could the problem for free electron appearance.  Cosmic rays as source of electrons is doubtful.  In my opinion absorbed negative ions on the cavity surface will be cause for electrons appearance. Due to lower possibility of electron appearance in comparison with large volumes and gaps, contacted with metal surfaces the electric strength of cavity will be more than defined by Pashen's law.

·

In gaseous dielectric medium or in vacuum(??) free electron will be present(may be insignificantly low but enough to occur PD) under high electric field due to phenomenon like cosmic radiation field emission  etc. If we consider the solid insulating material and hypothetical case of trapping of void(??) within it, can we consider the solid bulk absolutely free of defect or the void a real void? For insulating material with perfect molecular structure the potential well is deep and the ground state energy level of valance electron is very low particularly at extremely low temperature and even extremely high field may not be able to bring out an electron in the conduction band but in real  solid insulating material, be it a polar or non polar dielectric, some defects exists which form deep and shallow charge traps within the continuum and different form of charge conduction takes place under different magnitude of field the extreme case being conduction break down of insulation.

           We may conclude that so long we consider real life insulating material at normal temperature some charges will always be present to make partial discharge happen if the local field becomes high enough.

As already pointed out by others for a partial discharge to occur there should be presence of some gas. If there is no gas in the void no charged particles could be created. Hence both the presence of gas molecules/atoms and strong non uniform electric field are required for partial discharge to occur. Under this condition the gas breakdown creates free electric charges. Only when there are enough of such free charges does the partial discharge can be observed. Hence there is an inception voltage and extinction voltage defined for partial discharges.   

If there is a stress due to high local electric field (in HV equipment) then the gas molecules will break down and partial discharge condition may occur and its effect might have to be taken into consideration.  

Yes if the size of the cavity is low and the Paschen criterion is not verified. This criterion depends on the pressure in the cavity

A few details as possible clarification of some of the above comments (I am sure these points were considered in developing he comments, but I thought that it would be useful to articulate them):

1.  Partial breakdown (partial discharge) means that the entire insulation system has not broken down, or a puncture has not occurred in solid insulation.  However, this does not mean that a complete breakdown (sparkover) does not occur within the voids themselves in solid insulation.

2. Complete breakdown of a void can occur even when the total average applied field across the entire insulation is relatively low - consider the different in dielectric constants of the insulation and the gas or vacuum within the void.  The electric field within the void in enhanced compared with the electric field in the solid insulation.

3. The enhanced field within the void may result in the mechanisms described in previous comments: avalanche - which requires free electrons to initiate the avalanche, field emission - which does not rely on the availability of free electrons.

The possibility of void (complete) breakdown in practice is discussed in previous comments.

Regards, GG