1 000 000 VOLTS

Abatement of High-Field Conduction in Liquid Dielectrics by Electrode Conditioning with Non-lonic Cage-Forming Polymers:A Novel Avenue to High-Power Engineering


(1)Technical University of Munich (Germany)

(2) C.N.R.S.-Electrostatics.and Dielectrics Lab. Grenoble (France)

  Except hydrocarbons, no organic liquide in a conventional state of purity can be said a high-voltage insulator, for its conductivity is far too large.

  Polar liquids with high permittivity are particularly liable to excessive conductivity. Nitrobenzene, for instance (e/e0 = 36) rarely oversteps 108ohm.cm after distillation. Applying a field of 100 kV/cm would bring the liquid to boiling point in a matter of seconds.

  High-permittivity liquids, however, could be very useful for important applications like capacitors, electrostatic converters, electrorheology, if they were amenable to resistivities comparable to that of hydrocarbons (> 1011 ohm.cm ).

  'l'he pioneer work  of CARVALLO [1], in the 1910s, definetely showed that conduction is not intrinsic, and can be lowered by orders of magnitude by extremely slow and careful distillation. As an example, CARVALLO obtained a sample of acetone with 4 x 109 ohm.cm , while the usually quoted value was 106 ohm.cm .

  About 1955, GARTON [2] succeeded in purifying fluorinated liquids with e/e0 ~ 10 , but failed to use them in a technical appliance, since the trace electrolytes everywhere present would spoil the resistivity in a matter of minutes.

  It was obvious that permanent elimination of' trace clectrolytes was absolutely necessary. This was achieved by FELICI [3] by treating acetone or nitrobenzene with a mixed-bed ion exchanger. Very qulckly, resistivities better than CARVALLO's (> 1010 ohm.cm ) were reached and maintained over time. Unfortunately, as soon as a field in excess of 10 kV/cm was applied, conduction surged anew to unacceptable levels. By an electrochernical process, charge carriers are created at the electrodes in large numbers, a phenomenon henceforward called "injection".

  Soon thereafter, BRIERE [4] proposed to replace lon-exchange by electrodialysis between semipermeable membranes, a much more efficient process (Figure 1). A very important benefit of BRIERE's arrangement is the complete suppression of injection, for the injected ions are blocked by the membranes.

Semipermeables membranes

Figure 1 : A, C hemipermeable membranes E, liquid under de-ionisation 1,3 lateral collection chambers

  BRIERE's apparatus was radically simplified by FILIPPINI [5] who showed that the cumbersome collection chambers could be dispensed with. One had inerely to stretch the membranes on a metallic foundation to obtain electrodes that automatically deionized the liquid as soon as the operating voltage was applied. Resistivities far in excess of CARVALLO's (1013 ohm.cm ) could be reached in short time and maintained under fields of 500 kV/cm.

  Later on, FILIPPINI and al. [6] developed ion-exchange varnishies on a polyvinylic alcohol base that adhered to a metallic surface. This was an important improvement, as the strong electrostatic forces could not be easily withstood by deformable membranes.

  From the standpoint of electrical engineering, however, these remarkable achievements were practically of'no avail, since field reversals are unescapable. In this case, ion-exchange membranes or varnishes would cause a very large injection, only 1imited by space-charge effects. One has a "perfect injector". This, however, makes the operation of' an electric machine impossible.

  In a very long series of experiments, DITTRICH, studying the propylene carbonate liquid (e/e0 = 70) discovered that the injection processes thoroughly documented by TOBAZEON and THEOLEYRE [7] would wane over time, while a layer visible to the naked eye would form on the cathode. This layer obviously allows ions from trace electrolytes to pass through and neutralize, while also preventing injection. By reversing the voltage a number of times, DITTRICH obtained full-grown layers that perform as as BRIERE's membranes,but quite differently. They allow reversal of the field or AC voltage, and perfectly adhere to the electrodes [8].

  This unexpected phenomenon could be interpreted at last, thank to the help of YAZAMI.

    Injection into propylene carbonate occurs at the cathode. Because of the electronic affinity of oxygen, the bonds between 0 and CH2 or CH are broken. Two electrons are accepted by the carbonate residue, giving rise to the CO3- - ion which is injected, while on the propylene side the opened bonds close again, giving a CH2 = CH- CH3 propylene molecule (Figure 2). The opened bonds, however, can also close between neighbouring molecules, thus starting polymerization.

Catodic reaction of propylene carbonate

Figure 2 : Catodic reaction of propylene carbonate.

  What exactly happens is not yet known, for research is still in progress. However, it is possible to make a useful inference from the behaviour of ethylene carbonate in lithium cells, as described by YAZAMI [9]. When electrons are donated, the ethylene carbonate not only gives off CO3- - and ethylene, but also poly(ethylene carbonate) and poly (ethylene oxide). This compound polymer allows the positive ions to pass through in large numbers, but these are forcibly deprived of their solvation shells. It is thus very likely that in the case of DITTRICH's experiments the molecules of propylene carbonate cannot penetrate into the polymer layer. Now, for injection to occur, the reactive species must first adsorb on the metal until electron transfer takes place. Thus, blocking access of propylene carbonate molecules to the cathode will prevent injection, a mechanism totally different from membranes.

   Formerly, free electron emission was invoked as the cause of injection. This phenomenon does actually occur, as shown by DENAT [10], but at fields well in excess of'1,000 kV/cm.

  As a matter of fact, the energy density achievable with DITTRICH's arrangement is quite comparable to that present in the gap of an iron-core electromagnet. With e/e0 =80 (a mixture of propylene and ethylene carbonates) and an easily maintained field of 350 kV/cm, one has the energy density of 1.05 T, or 0.4 J/cm3.

  To illustrate this point, DITTRICH built a 55 kV ; 1,000 W alternator (Figure 3) substantially smaller and lighter than its electromagnetic counterpart, for only lightweight electrodes, are needed, instead of magnetic circuit and copper windings. Another application has been a high-voltage, DC-pulse generator with a pulse energy. 0.5 J ; peak power 100 kW and voltage 100 kV. In both cases, the insulating liquid undergoes purification by the mere application of the operating voltage [11], [12].

   Last but not least, the polymer layers spontaneously formed in DITTRICH's experiments are by no means unique [13]. There is variety of compounds having similar capabilities. A broad spectrum of applications may be envisaged.

Alternator's cross section

Figure 3 : Cross-section of 1,000 W alternator (outer diameter 89 mm).

1. Insulating frame of stator.

2. Inductors (under DC voltage) .

3. Rotor electrodes.

4. Rotor insulating frame

5. Statoric connections

6. Insulating liquid.


[1] J. Carvallo, Ann. Phys., Paris, Vol. 2, pp.142, 19 14.

[2] Personal communication from Mr. C. G. Garton, who was then with Electrical Research Association, U.K.

[3] N. Felici, "Obtention de liquides a grand pouvoir dielectrique et de grande resistivite" C.R.A.S. Paris, T. 249, pp. 654-655, 1959.

[4] G. Briere and N. Felici, "Desionisation des solvants polaires par electrodialyse", C.R.A.S. Paris, T. 259, pp. 3237-3240, 1964.

[5] G. Briere, N. Felici and J.C. Filippini, "Desionisation du nitrobenzene par electrodialyse", C.R.A.S. Paris, T. 261, pp- 50975099,1965.

[6] J.C. Filippini, M. Sauviat and R. Tobazeon, "Interet de nouveaux revetements d'electrodes pour l'etude et la mise en oeuvre des dielectriques liquides", C.R.A.S. Paris, T. 271, pp. 600-603, 1970.

[7] S. Theoleyre and R. Tobazeon, "Conduction electrique aux champs tres intenses dans le carbonate de propylene", C.R.A.S. Paris, T. 296, Serie 11, pp. 241 - 244, 1983.

[8] W. Dittrich, Hohe elektrische Felddichten bei zugleich hohen spezifischen elektrischen Widerstanden- ein Markstein in der Entwicklung der Elektrofluidmechanik, etz Archiv, Bd. 33 (1981 ), Heft 1 1.

[9] R. Yazami, "Le lithiurn-carbone, une electrode en plein essor", Bulletin de la Societe Française de Physique, pp. 14-16,Mai 1994.

[10] A. Denat, "Etude de la conduction dans les solvants non-polaires", These de Doctorat d'Etat, Grenoble, 1982.

[11] W. Dittrich, Neue Generation elektrischer Maschinen mit HD-Fluiden, etz Bd. 108 (1987), Heft 8, pp. 328-333.

[12]W. Dittrich, H. Hill, El. Kinzel "Method for preparing high-purity propylene carbonate and for simultaneously making passivated electrodes", United States Patent Number 5, 437, 775, Aug. 1.1995.

[13] C.P. Klages, Fraunhofer Institute for Surface Engineering and Thin Films, Brunswick, Germany, personal communication.



© 2016 Lyonel Baum , France (EC) lyonelb@aol.com.