Development of Diagnostic Methods for Polyethylene

Insulations in Transmission Cables

1001393

0
0
Development of Diagnostic Methods for Polyethylene
Insulations in Transmission Cables

1001393

Technology Review, February 2001

EPRI Project Manager

W. Zenger

EPRI • 3412 Hillview Avenue, Palo Alto, California 94304 • PO Box 10412, Palo Alto, California 94303 • USA
800.313.3774 • 650.855.2121 • [email protected] • www.epri.com

0
DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES
THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN ACCOUNT OF
WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH INSTITUTE, INC. (EPRI).
NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE ORGANIZATION(S) BELOW, NOR ANY
PERSON ACTING ON BEHALF OF ANY OF THEM:

(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I) WITH
RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM
DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR INTERFERE WITH PRIVATELY OWNED
RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL PROPERTY, OR (III) THAT THIS DOCUMENT IS
SUITABLE TO ANY PARTICULAR USER'S CIRCUMSTANCE; OR

(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER (INCLUDING
ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE HAS BEEN ADVISED
OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR SELECTION OR USE OF THIS
DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN
THIS DOCUMENT.

ORGANIZATION(S) THAT PREPARED THIS DOCUMENT

EDF

This is an EPRI Level 2 report. A Level 2 report is intended as an informal report of continuing research, a
meeting, or a topical study. It is not a final EPRI technical report.

ORDERING INFORMATION
Requests for copies of this report should be directed to the EPRI Distribution Center, 1355 Willow Way, Suite 2478,
Concord, CA 94520, (800) 313-3774.

Electric Power Research Institute and EPRI are registered service marks of the Electric Power Research Institute, Inc.
EPRI. ELECTRIFY THE WORLD is a service mark of the Electric Power Research Institute, Inc.

Copyright © 2001 Electric Power Research Institute, Inc. All rights reserved.

0
CITATIONS

This report was prepared by

Electricite de France
Research and Development Division
Electrical Equipment Department
Les Renardieres – CIMA 8
77.818 Moret-sur-Loing, France

Principal Investigator
R. Clavreul
A. Courtot-Descharles

This report describes research sponsored by EPRI.

The report is a corporate document that should be cited in the literature in the following manner:

Development of Diagnostic Methods for Polyethylene Insulations in Transmission Cables, EPRI,

Palo Alto, CA: 2001. 1001393.

iii
0
0
CONTENTS

1 PHYSICAL AND CHEMICAL CHARACTERISTICS............................................................ 1-1

Introduction ........................................................................................................................ 1-1
Measurement of birefringence of insulating material........................................................... 1-2
Principle of the method .................................................................................................. 1-2
Sampling procedure....................................................................................................... 1-2
Results .......................................................................................................................... 1-3
Discussion ..................................................................................................................... 1-4
Conclusion..................................................................................................................... 1-5
Infra-red spectroscopy polymer structure identification ....................................................... 1-5
Principle of the method .................................................................................................. 1-5
Sampling procedure....................................................................................................... 1-5
Result ............................................................................................................................ 1-5
Discussion ..................................................................................................................... 1-6
Conclusion..................................................................................................................... 1-6
Detection of mineral impurities by measurement of Proton Injection X-ray Emissions
(PIXE) ................................................................................................................................ 1-6
Principle of the method .................................................................................................. 1-6
Sampling procedure....................................................................................................... 1-7
Result ............................................................................................................................ 1-7
Conclusion..................................................................................................................... 1-7
Search for electrochemical tree structures by optical microscopy....................................... 1-8
Principle of the method .................................................................................................. 1-8
Sampling procedure....................................................................................................... 1-8
Results .......................................................................................................................... 1-9
Conclusion..................................................................................................................... 1-9
Determination of trace moisture by Karl Fisher electrochemicaI method ............................ 1-9
Principle of the method .................................................................................................. 1-9
Application to cable isolation.......................................................................................... 1-9

v
0
 Sampling procedure......................................................................................................1-10
Results .........................................................................................................................1-10
Conclusion....................................................................................................................1-10
Measurement of density under ASTM D756 standard .......................................................1-11
Principle of the method .................................................................................................1-11
Sampling procedure......................................................................................................1-11
Results .........................................................................................................................1-11
Conclusion....................................................................................................................1-12
Measurement of the crystalline melting temperature and of crystallinity content by
Differential Scanning Calorimetry (DSC) method...............................................................1-12
Principle of the method .................................................................................................1-12
Sampling procedure......................................................................................................1-13
Results .........................................................................................................................1-13
Conclusion....................................................................................................................1-14
Measurement of mechanical relaxation .............................................................................1-14
Principle of the measurement .......................................................................................1-14
Results .........................................................................................................................1-15
Conclusion....................................................................................................................1-15
Determination of micro-cavities by small angle X-ray spectroscopy (SAXS) ......................1-16
Principle of the measurement .......................................................................................1-16
Sampling procedure......................................................................................................1-17
Results .........................................................................................................................1-17
Conclusion....................................................................................................................1-17
Discussion - Conclusion ....................................................................................................1-17
Physical and chemical characterisation methods ..........................................................1-17
Additional material assessments after a failure .............................................................1-18
Methods used to investigate the ageing of extruded cables ..........................................1-19
Recommendations........................................................................................................1-20

2 DIELECTRIC TESTS ........................................................................................................... 2-1

Introduction ........................................................................................................................ 2-1
Measurement over time of dielectric relaxation by spectroscopy ........................................ 2-1
Principle of the method .................................................................................................. 2-1
Sampling procedure....................................................................................................... 2-2
Results .......................................................................................................................... 2-3

vi
0
 Conclusion..................................................................................................................... 2-3
Measurement of breakdown field with voltage ramp tests................................................... 2-4
Principle of the method .................................................................................................. 2-4
Sampling procedure....................................................................................................... 2-6
Results .......................................................................................................................... 2-6
Conclusion..................................................................................................................... 2-6

3 CHARACTERISATION OF POLYETHYLENE SLABS AND CABLES BY SPACE

CHARGE MEASUREMENTS ................................................................................................. 3-1
Space charge measurement in cables................................................................................ 3-1
Definition of methods: thermal and pressure wave measurements of space charge ...... 3-1
Pressure wave method.............................................................................................. 3-1
Principle of the method......................................................................................... 3-1
Example : polymer studies.................................................................................... 3-3
Details of method [5] :........................................................................................... 3-4
Thermal wave method............................................................................................... 3-5
Principle of the method......................................................................................... 3-5
Presentation and analysis of experimental results on aged materials (effect of
pressure, temperature and applied electrical field)......................................................... 3-7
Pressure wave method.............................................................................................. 3-7
Thermal wave method............................................................................................... 3-9
Pressure effect: .................................................................................................... 3-9
Proposed theoretical explanation of evolution of space charges ...................................3-12
Pressure wave method.............................................................................................3-12
Thermal wave method..............................................................................................3-13
Conclusion....................................................................................................................3-13

4 CORRELATION OF ELECTROLUMINESCENCE WITH EXTRUDED CABLE

AGEING.................................................................................................................................. 4-1
Electroluminescence Measurement.................................................................................... 4-1
Definition of methods and its applications on material films ........................................... 4-1
Parameters influencing EL mechanisms ........................................................................ 4-5
Excitation mechanisms.............................................................................................. 4-5
Electroluminescence relaxation mechanisms ............................................................ 4-7
Comparison between cables aged under accelerated processes (electrical stress)
and non-aged cables ..................................................................................................... 4-9

0
 Cables studied .......................................................................................................... 4-9
Results on new cables isolation materials ................................................................4-10
Results on aged cables ............................................................................................4-11
Conclusion ...............................................................................................................4-11
Current Questions .............................................................................................................4-12

5 REFERENES ....................................................................................................................... 5-1

viii
0
LIST OF FIGURES

Figure 1-1 Sampling for measurement of birefringence results ............................................... 1-3

Figure 1-2 Observation of low and high birefringence in XLPE new cable............................... 1-4
Figure 1-3 Electrochemical tree structure in bush shape......................................................... 1-8
Figure 1-4 Electrochemical tree structure in « bow-tie » shape ............................................... 1-8
Figure 1-5 Sampling for moisture measurement by Karl Fisher method .................................1-10
Figure 1-6 Taking sample for density measurement...............................................................1-11
Figure 1-7 Density measurement of cables B1 & D1 (new XLPE cables from different
manufactures), B2 (cable B1 aged during 2700 hours at 24 kV/mm at room
temperature), D3 (cable D1 aged during 1300 hours at 27 kV/mm with 8-hours
thermal cycles) ...............................................................................................................1-12
Figure 1-8a DSC of XLPE non-aged cable.............................................................................1-13
Figure 1-9 Samples for measurement of mechanical relaxation .............................................1-15
Figure 1-10 Mechanical relaxation of an XLPE cable before and after ageing.......................1-16
Figure 2-1 Circuit diagram....................................................................................................... 2-2
Figure 2-2 Sampling procedure............................................................................................... 2-3
Figure 2-3 Gold electrode with guard ring on one face of the sample...................................... 2-3
Figure 3-1 Principle of pressure wave method: Pressure pulse, (p(z,t)), abscissa of wave
leading edge (zf), induced surface charges at electrodes (σ0, σd).................................... 3-2
Figure 3-2 Displacement of fine layer charge on z axis, and effect on space charge
distribution, under conditions of short circuit .................................................................... 3-2
Figure 3-3 Charge signal of virgin sample of 1 mm thick XLPE ............................................... 3-4
Figure 3-4 Charge signal of sample subjected to voltage : space charge in a 1 mm thick
XLPE............................................................................................................................... 3-4
Figure 3-5 Standardised charge signals at the end of each polarisation (a) at 20°C, (b) at
70°C [1]. .......................................................................................................................... 3-8
Figure 3-6 (a) distribution of electrical field E=f(x) for different pressures (graph A : 300
bar, graph B : 150 bar, graph C : 50 bar, 30kV and 70°C) ............................................... 3-9
Figure 3-7 (a) distribution of space charge ρ=f(x) for different pressures (graph A : 300
bar, graph B : 150 bar, graph C : 50 bar, 30kV and 70°C) [5] .........................................3-10
Figure 3-8 (a) distribution of space charge ρ=f(x) in XLPE for a voltage of 30 kV,
temperature of 30°C and pressure 50 bar.......................................................................3-11
Figure 3-9 (b) distribution of space charge ρ=f(x) in XLPE for voltage 30 kV, temperature
of 70°C and pressure 150 bar.........................................................................................3-11

ix
0
Figure 3-10 (b) distribution of space charge ρ=f(x) in HDPE for voltage 30 kV,
temperature of 70°C and pressure 150 bar.....................................................................3-11
Figure 3-11 Space charge dynamics at the outset of sample de polarisation at 20°C. The
arrows symbolise the charge injection [1]. ......................................................................3-12
Figure 4-1 Jablonsky’s diagram molecular concept of luminescence ...................................... 4-2
Figure 4-2 Band diagram of three categories of material : (a) isolation, (b) semi-
conductor, c) conductor ................................................................................................... 4-2
Figure 4-3 Electron energy distribution curve in a material with a large band forbidden
band ................................................................................................................................ 4-3
Figure 4-4 Configuration of measurement enclosure............................................................... 4-4
Figure 4-5 Divergent field configuration................................................................................... 4-5
Figure 4-6 Uniform field configuration ..................................................................................... 4-5
Figure 4-7 Electroluminescence as a function of electrical field............................................... 4-6
Figure 4-8 Electroluminescence as a function of frequency..................................................... 4-7
Figure 4-9 Electroluminescence phase angle in relation to applied voltage............................. 4-7
Figure 4-10 Luminescence spectrum due to charge recombination......................................... 4-8
Figure 4-11 PE photoluminescence spectrum ......................................................................... 4-9
Figure 4-12 Results of comparative tests on new and aged cables ........................................4-10
Figure 4-13 Determination of emission thresholds of A1, D3 and C2 cable samples from
left to right ......................................................................................................................4-11

x
0
1
PHYSICAL AND CHEMICAL CHARACTERISTICS

Introduction

Crosslinked polyethylene insulated energy cables have been characterised in optimal manner, in
terms of the analysis of the physical and chemical characteristics on which ageing studies are
based. The following measurements are detailed in this document:
• Birefringence ;
• Infra-red spectroscopy ;
• Proton Injection X-ray Emission (PIXE) ;
• Optical microscopy ;
• Karl Fisher method electrochemical analysis ;
• Density ;
• Differential Scanning Calorimetry (DSC) analysis ;
• Mechanical relaxation ;
• Small Angle X-ray Spectroscopy (SAXS).

Very High Voltage crosslinked polyethylene insulated cables (crosslinking agent : dicumyl
peroxide) were manufactured, then aged in the laboratory under the following conditions :
• Cable A1 : new XLPE cable ;
• Cable B1 : new cable with the same XLPE insulating from another manufacturer ;
• Cable A2 : cable A1 aged during 2,700 hours at 24 kV/mm and room temperature ;
• Cable B2 : cable B1 aged during 1,300 hours at 27 kV/mm with 8-hour thermal cycles,
raising temperature from ambient to 90 °C, followed by 16 hours without heating.

Results obtained on these cables are presented.

1-1
0
Physical and Chemical Characteristics

Measurement of birefringence of insulating material

Principle of the method

The measurement of birefringence involves the detection of the material’s capacity to polarise
light directed at it in a controlled manner, so quantifying intensity. The principle of
measurement is as follows :

O-------------|---------------|------------------||-------------------|-----------------|--------------------|

Source Lens Polariser Sample Analyser Lens Observation

When a light beam is transmitted through a non-birefringent material, the light leaves with the
same polarisation. But for a birefringent material, polarised light from the sample breaks down
into two orthogonal axes.

These two components are propagated in the material at different speeds, resulting in fazing of
light output, to a degree proportional to the thickness of the sample and differences in material-
related indices. The difference in the optical path of the two components is known as the travel
difference. Both components are aggregated at the analyser level. The intensity of the light
transmitted by the analyser is the sum of the contribution of all the wavelengths of the spectrum
after interference from the two perpendicular components.

To obtain the birefringence value, the travel distances are calculated using the Newton’s colour
scale divided by the thickness of the material.

The analysis of birefringence in cable isolation materials provides information on residual

stresses:
• When macromolecular chains are highly deformed, a high degree of birefringence is
observable ;
• In the event of low deformation of macromolecular chains, birefringence is low.

Sampling procedure

The sampling procedure is as follows :

• The outer sheath and metal shielding are first removed from the cable;
• A 20 mm long sample of cable is then taken;
• A 400 µm thickness sample including the isolation material and the inner and outer semi-
conductors is obtained, using a microtome (figure 1-1).

1-2
0
 Physical and Chemical Characteristics

Figure 1-1
Sampling for measurement of birefringence results

Results

After placing the sample between two cross-polarisers under white light, the following
observations were made (figure 1-2) :
• Very high level of birefringence against the inner semi-conductor: blue colour ;
• Low birefringence in rest of sample: yellow color.

1-3
0
Physical and Chemical Characteristics

Figure 1-2
Observation of low and high birefringence in XLPE new cable

No difference was observed between the non-aged crosslinked polyethylene cable, and the cable
subjected to electrical ageing of around 2,700 hours at 24 kV/mm & room temperature.

Discussion

The origin of the residual stresses observed in the cable is not correlated with ageing, but is
related to the manufacturing process. A high degree of birefringence against the inner semi-
conductor probably indicates a greater level of polyethylene crosslinking at this point.

1-4
0
 Physical and Chemical Characteristics

Conclusion

1. The birefringence measurement does not detect any electrical or thermal ageing ;

2. The birefringence method points at manufacturing defects in the isolation material. It is

recommended to first make this measurement on non-aged cable in order to detect the low
birefringence area. Then, sampling for physical and chemical methods like infra-red
spectroscopy, PIXE, density, DSC, mechanical relaxation and SAXS measurements, shall be
done in the same birefringence area (low birefringence area) in order to prevent any
misunderstanding in the interpretation of the ageing phenomena.

Infra-red spectroscopy polymer structure identification

Principle of the method

The infra-red spectroscopy method is frequently used to characterise macro-molecular and

organic materials. It enables the observation of the rotation and vibration movements of specific
groups of the product under investigation, when it is subject to electromagnetic radiation in the
infra-red and near ultra-violet frequencies (600 to 4,000 cm-1). Characteristic absorption bands
of crosslinked polyethylene organic groups are observable, particularly in the following regions:
• 1,650 cm-1-1,740 cm-1 : carbonyl groups;
• 1,300 cm-1-1,380 cm-1 : methyl groups.

Infra-red spectroscopy measurements enable the quantitative analysis of organic products, as

determined by the Beer Lambert law, as follows:

Absorbance A = Log (Io/I) = K*L*C with

Io = Intensity of incident radiation;
I = Intensity of transmitted radiation;
K = Absorption constancy;
L = Sample thickness
C = Concentration of product under study

Sampling procedure

The sampling procedure was the same as that used for the birefringence measurement.

Result

Measurements were performed using an infra-red spectrometer operating under Fourier

transformation, and supplied with micro-beam system for transmission analysis, on a surface of
50 µm diameter. The following results were obtained :

1-5
0
Physical and Chemical Characteristics

• Identification of four absorption bands in the region of 1,650 cm-1-1,740 cm-1 :

– 1,735 cm-1 : vibration C=O of ester link; this is probably a vinyl acetate group in the
semi-conductors (Ethylene Vinyl Acetate) ;
– 1,720 cm-1 : vibration C=O in ketone groups in the polyethylene ; this band is
characteristic of material oxidation ;
– 1,710 cm-1 : COOH vibration of the carboxylic acid groups contained in the polyethylene;
this band is also characteristic of oxidation in the product under study ;
– 1,695 cm-1 : ketone associated with aromatic cycle; this is an acetophenone, a secondary
crosslinking product derived from dicumyl peroxide. As this product is highly volatile,
quantitative analysis is not recommended.
• Band of characteristic methyl group at 1,377 cm-1 was also demonstrated.

Discussion

Crosslinked polyethylene oxidation bands were observed both in non-aged cable and in cable A
electrically aged for 2,700 hours at 24 kV/mm & room temperature, without substantial
difference in terms of intensity. When differences were observed, these were more related to
birefringence (High/low birefringence) than to ageing. For example, the quantity of methyl
groups was greater in the low birefringence zone.

Conclusion

The infra-red microscopy method does not detect any electrical or thermal ageing.

It detects oxidation that could occur during ageing in non-waterproof cables by oxygen
penetration from outside. This phenomena is very slow and it is independent of the electrical
stress. Besides, the thermal stress applied on non-waterproof cables has only an acceleration
effect on oxygen penetration in polyethylene and oxidation.

This infra-red method also detects polyethylene oxidation occurring during manufacturing and
can been applied on new cables to check the quality of their isolation.

Detection of mineral impurities by measurement of Proton Injection X-ray

Emissions (PIXE)

Principle of the method

The energy of protons contained in the measurement beam is enough to disturb the electrons in
the deep-lying atomic layers of the sample (layers K, L, M). The de-excitation of these
electrons, from a high level to lower level is accompanied by X-ray emissions at energies
characteristic of those of the levels in question, hence of atomic excitation. Knowledge about the
electron transmission energies for each element make sample analysis possible.

1-6
0
 Physical and Chemical Characteristics

Sampling procedure

The sampling procedure adopted was the same as that used in the birefringence measurement.

Result

To identify the diffusion of mineral impurities (if any) from the semi-conductors into the
isolation material, scanning was performed throughout the thickness of the insulating material, at
50 µm steps:
• The following products were detected:
– Silicon;
– Sulphur ;
– Chlorine ;
– Calcium ;
– Aluminium ;
– Magnesium ;
– Potassium.
• The dispersion of mineral impurities was substantial in the insulating material;
• No profiles for inorganic impurities were observed in cable isolation;
• No significant difference was observed between the non-aged cable insulated with
crosslinked polyethylene and the cable subjected to electrical ageing for 2,700 hours at 24
kV/mm and room temperature, or the cable aged electrically and thermally for 1,300 hours at
27 kV/mm with thermal cycles of 90 °C (cycles of 8 hours of raising temperature from
ambient to 90°C, followed by 16 hours of cooling by natural convection).

Conclusion

1. The PIXE method does not detect any electrical or thermal ageing ;

2. It gives a photography of the quality of the interfaces « isolation/semiconductor » obtained

during manufacturing ;

3. In non-waterproof cables, mineral impurities can diffuse with water from semiconductors to
the isolation. So, the PIXE measurement is mainly recommended in this last case.

1-7
0
Physical and Chemical Characteristics

Search for electrochemical tree structures by optical microscopy

Principle of the method

The method involves identifying tree structure electrochemical faults possibly present within the
insulating material by optical microscopy in transmission mode. Two types of water tree
structure are possible (Figures 1-3, 1-4) :
• Water tree structures in «bush» shape, which are formed from the inner or outer semi-
conductor, usually originate in a protrusion of the relevant semi-conductor. The source of the
water tree is characterised by a local electrical field that can be abnormally higher than the
applied stress. Such a phenomena leads firstly to the formation of a canal, then to the
development of branches.
• Electrochemical tree structures in «bow-tie» shape are formed from mineral or organic
impurities in the insulating material.

Figure 1-3 Figure 1-4

Electrochemical tree structure in bush Electrochemical tree structure in « bow-
shape tie » shape

These water tree structures are made up of a multitude of micro-channels formed in the
insulating material. Impregnation by a coloured solution leads to the faults in the material,
highlighted by penetration of the colour, which can then be easily identified by optical
microscopy in transmission mode.

Sampling procedure

The sampling procedure used was the same as that for birefringence measurements. The samples
obtained are impregnated in a solution of methylene blue (0.5 % by weight) in distilled water for
around 8 hours at 80 °C, and then rinsed in distilled water and wiped on filter paper.

Samples are then observed by optical microscopy in transmission mode, at magnifications

ranging from 10 to 1,000.

1-8
0
 Physical and Chemical Characteristics

Results

No electro-chemical tree structures were detected in the non-aged cables, or in cables subject to
ageing for investigation purposes.

Conclusion

1. The microscopy method is not adapted to detect any electrical or thermal ageing in
waterproof cables ;

2. This method is recommended only for non-waterproof cables when the presence of both
water and electrical field in the isolation usually leads to the formation of water trees after
several years of ageing in normal conditions.

Determination of trace moisture by Karl Fisher electrochemicaI method

Principle of the method

The Karl Fisher electrochemical method is the most sensitive measurement technique for the
detection of trace moisture in the crosslinked polyethylene isolation material used for energy
cables. The principle is as follows:

1. Iodine reduction : I2 + SO2 + H2O Æ 2 HI + H2SO4

2. Coulometrical reaction or iodine oxidation : 2I-Æ I2 + 2 e-

According to Faraday’s law, when one mole of water reacts with a mole of iodine, two electrons
are detected by electrical measurement.

Application to cable isolation

For solid isolation material, such as crosslinked polyethylene, measurement is indirect, and
performed according to the following process:

1. Introduction of sample into the oven, preheated to 170 °C ;

2. Water present in the sample is extracted and then vaporised in the oven. ;

3. The vaporised water is then transferred into the electrochemical cell by an inert and dry
vector gas (pure nitrogen) ;

4. As soon as it enters the electrochemical cell, the water reacts immediately with the Karl
Fisher reagents.

1-9
0
Physical and Chemical Characteristics

Sampling procedure

Sampling procedure was specifically developed for Karl Fisher measurement (Figure 1-5) :

1. The outside sheath and metal shielding are first removed from the cable;

2. A 20 mm length of cable is then cut out;

3. A sample of size (20 mm*10 mm*2 mm) is then taken using a microtome*.

Figure 1-5
Sampling for moisture measurement by Karl Fisher method

Results

The water content in crosslinked polyethylene cable isolation material is around (25 ± 5) ppm.
This level is independent of the zone sampling (close to the semi-conductors or in the middle of
the insulating material). It is also independent of the cable type (non-aged cable, cable subjected
to ageing for 2,700 hours at 24 kV/mm, cable subjected to ageing for 1,300 hours at 27 kV/mm
with thermal cycles at 90 °C).

Conclusion

1. The Karl Fisher method has not to been applied to waterproof cables where the isolation has
been crosslinked in nitrogen atmosphere ;

2. For non-waterproof cables, it is recommended to compare non-aged and aged cables by Karl-
Fisher method where an important quantity of water coming from outside could lead to the
formation of water trees during long term ageing.

1-10
0
 Physical and Chemical Characteristics

Measurement of density under ASTM D756 standard

Principle of the method

The measurement of density under standard ASTM D756 is performed using a density gradient
column. As this column is graduated, samples whose density is known with precision are first
measured, in order to check the correspondence of the graduations to density. For crosslinking
polyethylene, measurements are performed in a water/isopropylic alcohol mixture to give a
dynamic measurement magnitude for the polyethylene density field (0.9).

Sampling procedure

The final sample was a crosslinked polyethylene ribbon obtained from a numerically controlled
lathe, the following stages being necessary (Figure 1-6) :

1. The outer sheath, metal shielding and metal conductors are removed from the cable;

2. A sample cable length of 50 mm is then taken;

3. A ribbon of 150 µm thick is then unrolled from the outer semi-conductor to the inner semi-
conductor, using a numerically controlled lathe, the blade angle 45 °C.

Figure 1-6
Taking sample for density measurement

Results

Density values range between 0.918 et 0.9235 in non-aged and electrically aged samples, with or
without thermal cycle. The relatively low observed divergence of 0.5 % should rather be
attributed to the change in birefringence (strong and low birefringence), rather than to ageing
(Figure 1-7).

1-11
0
Physical and Chemical Characteristics

Figure 1-7
Density measurement of cables B1 & D1 (new XLPE cables from different manufactures),
B2 (cable B1 aged during 2700 hours at 24 kV/mm at room temperature), D3 (cable D1 aged
during 1300 hours at 27 kV/mm with 8-hours thermal cycles)

Conclusion

1. This method does not detect any electrical or thermal ageing ;

2. The density measurement is only recommended in new cables in order to check the
homogeneity of the isolation. In case of heterogeneity, we can consider that problems may
have occurred during manufacturing and would merit to be discussed with the manufacturer.

Measurement of the crystalline melting temperature and of crystallinity

content by Differential Scanning Calorimetry (DSC) method

Principle of the method

The differential scanning calorimetry method involves detecting any physical modification in a
material (melting point, crystallisation, vaporisation) leading to the temperature related
absorption or transmission of energy. This method makes it possible to determine the melting
point and the crystalline area of the polyethylene, and of the crystallinity rate, for the surface area
of the melting peak is proportional to the latter.

1-12
0
 Physical and Chemical Characteristics

Sampling procedure

Sampling procedure involves cutting out 500 µm diameter samples, and using a numerically
controlled lathe to shave off a 150 µm thin ribbon. The method is the same as described for
density measurement.

Results

Measurement performed in the range of 0 °C to 130 °C with temperature rise of 10 °C/minute

showed :

In non-aged cable and in cable electrically aged at 24 kV/mm for 2,700 hours (Figure 1-8a):
• A first pre-melting peak at 75 °C corresponded to the heat treatment temperature of the cable
at the time of manufacture; heat treatment is performed to eliminate the volatile by-products
of peroxide degradation at the time of the crosslinking reaction ;
• Crystalline melting peak at around (97.0 ± 1.0) °C

Figure 1-8a
DSC of XLPE non-aged cable

In the cable aged at 27 kV/mm for 1,300 hour with 90 °C thermal cycles (Figure 1-8b):
• The pre-melting 75°C peak observed in the non-aged cable disappeared ;
• A new peak around 90°C was observed, this being the highest temperature reached during
cable ageing.

1-13
0
Physical and Chemical Characteristics

Figure 1-8b
DSC of XLPE aged cable

The crystallinity rate was determined by reference to the 69 cal/g reference value for the totally
crystalline polyethylene. This rate is around 45 %, and is independent of ageing.

Conclusion

1. No effect of ageing has been detected in the studied cables. We don’t recommend this
method for the study of any electrical ageing.

2. The melting point depends of the crosslinking degree. The DSC method is allowed in new
cables in order to check the melting point and detect any manufacturing problem : when
crosslinking is incomplete, this phenomena usually continue during thermal ageing and leads
to a change of the melting point.

3. We also recommend this method to detect any important increase of the temperature over
65-70°C during ageing.Measurement of mechanical relaxation

Principle of the measurement

The method consists of applying to the material a static stress over which a weaker sinusoidal
stress is superimposed, and to record the amplitude of the displacement and the phase, as
compared to mechanical excitation. The tangent of the angle of mechanical loss is determined:

Tan (delta) = Imaginary part/Real part (in phase)

1-14
0
 Physical and Chemical Characteristics

Sampling procedure

Cylindrical samples of 9.5 mm radius and 10 mm height were obtained in the isolation using a
sample cutter. These samples were then machined by lathe to obtain uniform surfaces and
eliminate all surface defects which may have been caused by semi-conductors (Figure 1-9).

Figure 1-9
Samples for measurement of mechanical relaxation

Applied stresses were 600 g static and 200 g in dynamic. The measurements were made at
temperature steps of 5 °C from ambient temperature to 120°C, and at a number of frequencies,
from 10-5 Hz to 10 Hz.

Results

Measurements showed no significant difference between non-aged cable (Cable B1) and cable
aged electrically at 24 kV/mm for 2,700 hours (Cable D3). On the other hand, an evolution was
observable in cable electrically aged for 1,300 hours at 27 kV/mm, with thermal cycles at 90 °C
(cable B2). Ageing resulted in an increase in the tangent of the angle of loss at 10-2 Hz, and a
displacement of the alpha relaxation peak in the lower frequencies (Figure 1-10).

Conclusion

The tangent of the angle of mechanical loss of crosslinked polyethylene isolation used in very
high voltage energy cables tends to increase on the simultaneous application of electrical and
thermal stresses, as a result of applying temperature cycles from ambient temperature to 90 °C,
whereas electrical stress alone causes no notable evolution. The ageing time in our study was
insufficient in order to be able to correlate Tan(delta) with ageing.

1-15
0
Physical and Chemical Characteristics

Cable D3

Cable B1 (new) and B2

Figure 1-10
Mechanical relaxation of an XLPE cable before and after ageing

Determination of micro-cavities by small angle X-ray spectroscopy (SAXS)

Principle of the measurement

The method consists of producing a beam of X-rays in a tube by the bombardment of an anode
by a beam of electrons. When the beam passes through the sample, it detects particles or micro-
cavities from 10 to 1,000 Angstroms, resulting in the dispersion of the incidence beam by
diffusion. This dispersion is related to the difference in index or in density, and to the size of
particles. The larger the particles, the less the deviation of the beam. The recorded signal is the
number of blows by unit of time, as a function of the angular position of the detector, which is a
scintillation type meter. Particle size is analysed by the Frankuchen method, based on the
Guinier equation, valid for small angles where:

I(s) = Ie*M*n2*exp(-S2.Ro2/3)
where
Ie = intensity of diffraction of an electron
M = number of particles

n = number of electrons per particle

s = vector of diffraction
Ro = radius of gyration of particle

The mass fraction of populations of particles makes it possible to quantify the probability of their
presence by the equation:

1-16
0
 Physical and Chemical Characteristics

Kn=k*W*(Ron)*Ro3
when
k = constant
W(Ron) = fraction of mass of population of particles n

Sampling procedure

Samples of size (20 mm*15 mm) were taken from the ribbon made for density measurement
purposes by the numerically controlled lathe for density measurements.

Results

Particles of diameter 70 and 120 Angstroms were detected in the isolation materials of non-aged
cables, and cables electrically aged for 2,700 hours at 24 kV/mm :
• The magnitude of the small particles of 70 Angstroms corresponded to crystallites ;
• Particles of 120 Angstroms are of long period size (crystallites and size of amorphous phase).
The difference between the two sizes is the size of the amorphous phase, which is in the
order of 50 Angstroms.

The size of the amorphous phase is around 5 % higher in the electrically aged cable. As
precision is around 3 %, this finding requires prudent interpretation.

Conclusion

SAXS measurements enable the determination of the size of crystallites and of the amorphous
phase of the polymer. No micro-void was detected by this method. On the other hand, the
number of micro-voids may have been too small by comparison to other elements measured to
enable their identification.

Discussion - Conclusion

Physical and chemical characterisation methods

Physical and chemical characterisation methods in crosslinked polyethylene insulating material

relate to :

1. Morphological analysis:
• Measurement of birefringence ;
• Measurement of density;
• Determination of crystalline melting point and of crystallinity rate using the
differential scanning calorimetry method;

1-17
0
Physical and Chemical Characteristics

2. Analysis of electrical characteristics:

• The methods relating to the analysis of electrical characteristics are presented in tasks
2 (usual methods) and tasks 3 (new methods in the course of the development).

3. Analysis of mechanical characteristics:

• Measurement of mechanical relaxation

4. Chemical characterisation of materials studied:

• Identification of polymer structure by infra-red spectroscopy

5. Search of physical faults, of micro-cavity type or of tree structure type:

• Detection of micro-cavities by Small Angle X-ray Spectroscopy;
• Search for electrochemical tree structures by optical microscopy

6. Search for impurities:

• Detection of mineral impurities by Proton Induced X-ray Emissions (PIXE) method ;
• Detection of moisture traces by the Karl Fisher electrochemical method

Additional material assessments after a failure

The methods below are commonly used in the event of an assessment after a failure :

1. Case of cable subjected to abnormally high heating:

A cable subjected to abnormally high heating caused by high loadflows may present visually
observable anomalies. Among these are the deformation of the outer sheath or the
displacement of the central conductor, leading to a change in dimensions.

The recommended method of characterisation is differential scanning calorimetry method,

making it possible to check:
• Presence of the pre-melting peak around 75 °C ; if this peak is no longer in evidence, the
cable has been subjected to heating beyond 75°C;
• Possible presence of a further peak around the temperature of crystalline melting point :
this shows the actual temperature reached by the cable in operation ;
• Difference at various radial locations.

2. Unsealed cable subjected to accidental penetration of moisture:

If an unsealed cable installed on the network is in a humid region for at least 5 years,
accidental penetration of water may be the cause of the defect. This can be checked by
measuring the moisture rate. The moisture rate is around 30 ppm in a crosslinked
polyethylene isolation surrounding a normal cable. If the moisture rate exceeds 50 ppm, the
water present in the cable must in part have originated from the outside environment. In this

1-18
0
 Physical and Chemical Characteristics

case, it is recommended to check that the water is the origin of the incident by searching for
electro-chemical tree structures using optical microscopy.

3. Defective cable which has not been affected by an operating defect arising from over-heating
or accidental water penetration :
Cases such as these occur relatively frequently. It is recommended to search for faults
originating in cable manufacture. The presence of faults generally results in the reduction of
the dielectric strength of the material. This can be checked by the measurement of the
breakdown field on a voltage ramp test.

At a second stage, there should be a search for faults causing a deterioration of dielectric
withstand characteristics. Faults are of two types:
• Physical faults, such as cavities or tree-structures : the optical microscopy is perfectly
suitable for the identification of cavities and electrical or electro-chemical tree-structures.
• Faults of chemical composition or impurity : the most frequently encountered impurity is
moisture. There may also be mineral impurities, which may be detected by the PIXE
method, or organic impurities identifiable by infrared microscopy.

Methods used to investigate the ageing of extruded cables

When undertaking the investigation of the ageing of extruded cable isolation, it is first
recommended to search for a deterioration in the dielectric withstand characteristics of the
cable, as compared to non-aged cable. If and only if properties are adversely affected, is it
recommended to identify the factors causing the adverse trend, by means of physical and
chemical analysis. The following methods can be used:

1. Inspection of isolation morphology:

• Measurement of birefringence, other methods of characterisation can be undertaken in a
single zone of birefringence (strong or weak birefringence), in order to limit the dispersion of
results;
• Measurement of density;
• Measurement of crystalline melting point and crystallinity rate using differential thermal
analysis.

2. Characterisation of the thermal evolution of the material:

• Chemical characteristics of the material will be analysed by optical microscopy, a technique
identifying polymer oxidation, or possible alternative structural changes.

3. Search for micro-faults using Small Angle X-ray Spectroscopy

• This method detects faults as small as a few tens of Angstroms, possibly formed by
breakdown of chemical bonds during ageing.

1-19
0
Physical and Chemical Characteristics

4. Analysis or evolution of mechanical characteristics

• The measurement of mechanical relaxation enables the detection of possible hardening of the
material, itself possibly linked to a reticulation mechanism. If not, breakdown of
macromolecular chains might be the cause of ageing.

Recommendations

Physical and chemical analytical methods complement electrical measurements, and allow an
understanding of the parameters or mechanisms at the root of the malfunction or ageing of the
XLPE cables. Minimisation of the error assessment margin due to misinterpretation of potential
chemical and physical evolution can be achieved by comparing aged to non-aged cable, and
ensuring that all samples are taken from the same manufacturing batch.

1-20
0
2
DIELECTRIC TESTS

Introduction

This chapter describes the dielectric tests commonly used in the study of crosslinked
polyethylene as an isolation material for extruded cables. The following two methods are
covered :
• Measurement over time of dielectric relaxation by spectroscopy ;
• Measurement of the breakdown field with a voltage ramp test.

Results obtained with theses methods on very high voltage cable isolation, both new and aged
under accelerated procedures, are presented.

Measurement over time of dielectric relaxation by spectroscopy

Principle of the method

The principle of this method is to apply to the sample high voltage. A voltage of opposite
polarity is simultaneously applied to a variable reference capacitance, previously positioned at a
value close to that of the sample. The transient made up of the charge difference between the
reference capacitance and that of the sample is detected by a charge amplifier. The output
voltage from this charge amplifier is connected to the sample capacitance Cs, which is
measured as a function of time T. The information acquired over time is then transformed into
frequency information in the 10-4 to the 10+4 Hz range, by application of a Laplace
transformation (Figure 2-1).

2-1
0
Dielectric Tests

Figure 2-1
Circuit diagram

Laplace transformation [Cs(t)] = C*(ω) = C’(ω) -iC ‘’(ω) [2.1]

where
C*(ω) = Complex capacity as function of frequency ω
C’(ω) = Real part of complex capacity
C’’(ω) = Imaginary part of complex capacity

Conductance S of the sample takes the form :

S=C’’ [2.2]

The tangent of the angle of loss Tan (δ) is determined by:

δ)=C’’/C’
Tan(δ [2.3]

Sampling procedure

Initially, the sample used is a crosslinked polyethylene ribbon made by using a numerically
controlled lathe, in the following steps (Figure 2-2):
1. The outer sheath, the shielding and the metallic conductor are removed from the cable;
2. A 50 mm sample length of cable is then removed;
3. A ribbon 150 µm thick is made by numerically controlled lathe and is run from the outer
semi-conductor to the inner semi-conductor, the lathe blade being angled at 45°.

At a second stage, square samples of 50mm width are cut. Electrodes are placed on each side of
the samples, these being disks of pure (99.99%) gold, deposited by evaporation under secondary
vacuum, with a guard ring on one of the faces (Figure 2-3).

2-2
0
 Dielectric Tests

Figure 2-2 Figure 2-3

Sampling procedure Gold electrode with guard ring on one
face of the sample

Results

Measurements were performed on very high voltage crosslinked polyethylene insulated non-aged
cable, and the same cable aged for 2700 hours at 24 kV/mm, or for 1300 hours with thermal
cycles up to 90°C, under the following conditions:
• Frequency field of 10-3 Hz to 10+3 Hz ;
• Temperature variation range from 30 °C to 90 °C in steps of 10 °C.

No location-dependent difference was observed for a given cable : near the inner semi-
conductor, in the middle of the isolation, or close to the outside semi-conductor. Similarly, no
specific difference was observed for measurements at ambient temperature between non-aged
and electrically aged cables, with or without thermal cycles. However, it was observed that
losses were a little greater for measurements performed at 70°C and above in the cable isolation
electrically aged without thermal cycle. These measurements were in the magnitude of 5x10-4 in
the frequency range of 10-1 - 10 Hz, and increased to 10-3 when measurements were performed
between 70 °C and 90 °C.

Conclusion

The measurement of dielectric relaxation demonstrated an observable fact of electrical ageing :

the tangent of the angle of loss tends to increase slightly with electrical ageing. So, we
recommend this method to detect any electrical ageing. A correlation of dielectric relaxation with
ageing could be possible, but the too short ageing time of cables didn’t allow us to do it.

2-3
0
Dielectric Tests

Measurement of breakdown field with voltage ramp tests

Principle of the method

This is a test of dielectric breakdown which allows the characterisation and evaluation of the
electrical properties and degree of ageing of XLPE insulating material. This is a short term
characterisation in which samples to be studied are subject to an alternating voltage ramp at
frequency 50 Hz, that is at a voltage which increases at a constant speed. The measurement
system includes:
• Application of electrical stress: an adjustable auto-transformer connects to the primary of a
high voltage transformer from 500V to 100kV. On the transformer secondary, a voltage
ramp is obtained.
• Detection circuit: a detection resistance is placed in series with the tested sample in order to
detect dielectric breakdown by passing a short circuit current through the terminals of the
resistor, so protecting the measurement device, and measuring the breakdown voltage by a
voltmeter placed on the secondary of the auto-transformer, upstream of the relay contact for
low voltage breakdown relay contact. Hence, when the high voltage power supply is cut off
after the breakdown of the dielectric, only the auto-transformer secondary remains live,
which enables the measurement of the breakdown voltage.

The measurements are performed on crosslinked polyethylene films, between pairs of Rogowski
electrodes, one face being connected to the applied voltage, the other being connected to the
ground. The voltage ramp is applied up to the dielectric breakdown, taking the physical form of
a channel crossing the thickness of the insulating material.

At the moment of dielectric breakdown, the power supply circuit is immediately cut off. This
enables determination of the breakdown voltage using a voltmeter located on the auto-
transformer secondary. The magnitudes observed depend on :
• Characteristics of the insulating material: electric breakdown is related to the characteristics
of the insulating material, and maybe initiated by an electrical discharge even of thermal
origin.
• Experimental parameters: This involves the shape of the voltage applied, the nature and
profile of the electrodes, the presence of immersion oil in most of the tests, and devices
enabling the measured parameters to be sampled.

For samples from a same material obtained under the same experimental conditions, we do not
obtain a single breakdown field, but a statistical distribution of values which must be processed
using appropriate statistical models. The Weibull statistical model, originally used in the
mechanical field, is used. The probability of breakdown of a chain composed of N links, is
considered to be equal to that of the weakest link in the chain. Three parameters are basically
involved in the breakdown mechanism for a solid isolation:
• Time t after which breakdown occurs;
• Electrical field E applied to the sample until breakdown occurs;

2-4
0
 Dielectric Tests

• Volume V the sample subject to testing.

The probability of breakdown of a sample of volume Vi at time Ti subject to electrical field, Ei

takes the form of :

 
(t ) V .dt.dV 
ti

P(ti , Ei ) = 1 − exp − k1 ∫ ∫ t E
a −1 b
[2.4]
 Vi 0 

In which k1, a and b are isolation constants

In the case of a solid isolation whose volume is fixed, k2= k1.V is posited, whence the
relationship:

 
(t ) dt 
ti

P(ti , Ei ) = 1 − exp − k 2 ∫ t E
a −1 b
[2.5]
 0 

For a field ramp whose speed of ramping up is kept constant, the relationship is α=a+b :

[
P( Ei ) = 1 − exp − k . Ei
α
] [2.6]

In which α is a parameter of shape

Generally speaking, the Weibull relationship can be written as :

  E α 
P( E ) = 1 − exp − k    [2.7]
  Eo  

In which Eo is the scale parameter

And also:

 Eα 
P( E ) = 1 − exp − α  [2.8]
 E'o 

with E’o=k1/αEo

If the expression is considered in its linear form such that:

1
LogLn = αLogE − αLogE ' o [2.9]
1− P( E )

For E=E’o we obtain a, whence P=63,2 %

2-5
0
Dielectric Tests

1
LogLn =0 [2.10]
1− P

Sampling procedure

A crosslinked polyethylene ribbon of 150 µm thickness was made, using a numerically

controlled lathe, for the measurement of dielectric relaxation. Samples of 50mm side dimensions
are then cut out for the purposes of measurement of dielectric strength, the population used for
statistical processing for each material being 80.

The measurements of dielectric strength are made under the following conditions:
• Stainless steel electrodes 20 mm diameter, Rogowski profile;
• Silicone impregnation oil 500 cst;
• Voltage around 5 kV/s until dielectric breakdown.

Results

Experimental results are obtained in graphic form with the abscissa Xi=Log Ei and the ordinate
Yi=Log Ln 1/(1-Pi), for determination of the nominal breakdown voltage field Eo for Pi=63.2 %,
in the order of 135 kV/mm, showed no change with ageing and for 2700 hours limited ageing
time) at 24kV/mm or 1300 hours at 27kV/mm with thermal cycles at 90°C.

Conclusion

The measurement of dielectric breakdown strength enables the physical faults (if any) occurring
during the course of ageing, to be identified, for example micro-cavities or tree structures,
possibly inducing a nominal breakdown field with less withstand than the new sample. As these
faults are irreversible, we recommend their detection from the outset, for example at the level of
formation of space charges.

2-6
0
3
CHARACTERISATION OF POLYETHYLENE SLABS
AND CABLES BY SPACE CHARGE MEASUREMENTS

Space charge measurement in cables

Plane geometry single-pole conduction models do not fully describe the dielectric behaviour of
cable isolation. Even so, the study of the correlation, between the space charge inside the
dielectric material itself (spatial distribution of the electrical field) and the outer current, is
important. The reason is that isolation breakdown in the samples generally requires a very strong
current at the electrode terminals [1].

Definition of methods: thermal and pressure wave measurements of space

charge

Generally speaking, these measurement methods use the propagation of a pressure pulse in the
sample to make initial approximation of space charge distribution. The disturbance pulse may be
created either outside the insulating material by a pulse laser, or by a piézo-electrical transducer.
Alternatively it may be set up in the sample by applying a voltage pulse to the sample terminals
[1-4].

Thermal methods provide good resolution over a shallow depth, and are preferred for the study
of thin films. But it is also possible to use thermal methods to study far thicker isolation
materials, although resolution will be adversely affected close to the interfaces.

These two methods are therefore suited to the study of the materials used to insulate high voltage
cables [1].

Pressure wave method

Principle of the method

The principle is one of the propagation of a rapid transient pressure wave (2 * 103 m/s, steep
leading-edge), produced in a shock wave tube. The change of the wave in the material will
depend on the material’s thickness [5].

For a plane sample of an insulating material containing charges ρ(z) and permanent dipoles, two
conductor electrodes are bonded to or placed on either side of the dielectric (figure 3.1).

3-1
0
Characterisation of Polyethylene Slabs and Cables by Space Charge Measurements

Figure 3-1
Principle of pressure wave method: Pressure pulse, (p(z,t)), abscissa of wave leading edge
(zf), induced surface charges at electrodes (σ0, σd)

A pressure pulse is generated on the surface of one of the two electrodes and is propagated at the
speed of sound through the sample.

This mechanical disturbance modifies the atomic and molecular structure of the material,
displacing charges and causing the concentrations at the permanent dipoles to vary. The result is
a change in the distribution of the electrical field, and a variation in the induced charges at the
electrodes (figure 3.2).

Charges

Field

Figure 3-2
Displacement of fine layer charge on z axis, and effect on space charge distribution, under
conditions of short circuit

Depending on the impedance of the outer circuit (open or short circuit), the modification of the
interface charges generates a potential difference or current. Information on the spatial

3-2
0
 Characterisation of Polyethylene Slabs and Cables by Space Charge Measurements

distribution of the space charges can be extracted from the study of electrical behaviour over
time.

In open circuit conditions, the relation ship is as follows:

Vco (t) = χG (εr) ∫0 zf(t) E(z, 0) p(z, t) dz (3.1)

Vco (t) : potential difference

p(z, t) : propagation of rapid transient pressure wave
E(z, 0) : electric field spatial distribution
zf : abscissa of wave leading edge
χ : material compressibility
G (εr) : relative permitivity versus pressure

In closed circuit conditions, the current circulating through the exterior circuit is linked to
potential by:

Icc(t) = Co dVco(t) / dt (3.2)

Co : capacity of the capacitor at rest

The current is then:

Icc(t) = Co χG (εr) ∫0 zf(t) E(z, 0) (δ p(z, t) / δt) dz (3.3)

Using simple calculations (relating the mobility of the charge to network elasticity theory under
mechanical compression), it is easy to express the critical mobility above which the charge is no
longer detected (versus the electrical charge induced by the pressure pulse and of the speed of
network displacement). Electrical disturbance is related to thickness :

∆E = ρ νs τ δ l / ε d = ρ (νs τ)2/ ε d * ∆l/l (3.4)

ρ : charge density
νs τ : thickness zone
δ l : network deformation
d : sample thickness

Example : polymer studies

In PE, for example, a pressure pulse of around 100 bars is obtained, corresponding to a relative
displacement ∆l/l of about 2.5.10-3 , and an electrical disturbance ∆E in short circuit conditions
(operating mode) of less than 10 V/mm.

Signals recovered, may for example, be of the following type:

• virgin sample (figure 3.3):

3-3
0
Characterisation of Polyethylene Slabs and Cables by Space Charge Measurements

propagation of
pressure in insulator
pressure output

pressure input
in insulator

Figure 3-3
Charge signal of virgin sample of 1 mm thick XLPE

• sample with space charge (figure 3.4):

propagation of
pressure in insulator pressure output

pressure input
in insulator

Figure 3-4
Charge signal of sample subjected to voltage : space charge in a 1 mm thick XLPE

Details of method [5] :

• Improvement of the method for the distribution of charges in insulating materials (LDPE,
Teflon, XLPE): sample shock by laser pulse.
• This is a non-destructive method requiring electronic equipment of low time constant (rapid
propagation) and good resolution (material thickness: thin film, up to a few millimetres
thickness).

3-4
0
 Characterisation of Polyethylene Slabs and Cables by Space Charge Measurements

Thermal wave method

Principle of the method

This non-destructive method on solid isolation relies on the diffusion of a heat front applied to
one of the faces of the sample, and on the resulting non-uniform dilation [5].

Dilation induces a small electrical disturbance in the insulating material which is detected on the
terminals of the sample and digitally processed to give the electrical field profile [1]. The
resolution of the heat equation gives the relationship between the measured current and the
charge densities. This method can be applied to a cylindrical structure such as a cable.
Furthermore, the wave crosses interface boundaries [5].

The theoretical base of this method is the single dimension heat equation:

∂ 2T ( x, t ) ∂T ( x , t )
=W (3.5)
∂x 2
∂t

T(x,t) : temperature on abscissa x

t : time
where W = inverse of diffusivity taking into account the specific mass, the specific heat and the
thermal conductivity.

Qi2 Qi1

3-5
0
Characterisation of Polyethylene Slabs and Cables by Space Charge Measurements

It is considered that the insulating plate of a thickness «d» is placed between two parallel
electrodes connected by short circuit, in direction x, and the charge Qi in the material induces on
the electrodes (at distance*and d-x) image charges Qi1 and Qi2. The propagation of a heat-wave
and the accompanying non-homogenous dilation modify the relative values of x and d-x and
hence the balance between the image charges Qi1 and Qi2. As the sum of Qi1 and Qi2 must
remain constant, the exchanges of charges between the two electrodes translate into a current in
the outer circuit, which can be measured and recorded.

If the value of the current is known at all times, along with the temperature distribution in the
sample, the value of charge Qi and its position x can be found.

A measurement cell at ambient temperature is used. One of the electrodes is in contact with a
cold source (refrigerated liquid) and the other, due to its mass and thermal inertia, is the hot
source.

The electrical conditions fixed on the system can be expressed as follows:

Qi1 + Qi2 + Qi = 0 (3.6)

The temperature is taken into account by means of a linear thermal dilation coefficient for the
material in direction x, and a coefficient of permitivity variation versus temperature.

The current in the outer circuit is then expressed as follows:

I(t) = -S δQi2 / δt (3.7)

where S is the surface area of the sample

If the space charge density ρ(x) is distributed in the thickness of the material, Qi may be replaced
by:

∫
x 0+ d
ρ (x) dx (3.8)
x0

Finally, the current can be expressed as follows:

∫
x 0+ d
I(t) = - α C E(x) (δT(x,t) / δt) dx (3.9)
x0

C : sample capacity
E(x) : local electrical field in sample

Measurements were made at the Paul Sabatier University in Toulouse [5], in two stages. First the
samples undergo electrical, thermal and pressure treatment cycles to simulate the conditions
under which insulating material operates. This is followed by the measurement of charge
distribution under atmospheric pressure.

3-6
0
 Characterisation of Polyethylene Slabs and Cables by Space Charge Measurements

Example : case of cables used in the submarine environment. Samples of LDPE and XLPE (2
mm diameter) undergo the following treatment cycles: polarisation at 30 kV, at 70 °C and 50
bars for 12 hours, followed by several minutes of immersion, and short circuit discharge at 70 °C
for 12 hours.

To treat samples under pressure, interference charges are first eliminated (depolarisation for
around 12 hours at 70 °C). Then the samples are placed in a high pressure enclosure, and
subjected to continuous voltage for 24 hours at fixed pressure and temperature, in order to form
space charges. The samples are then maintained under short circuit in an aluminium sheet. The
charges are fixed by subsequently working at 0°C and atmospheric pressure.

Presentation and analysis of experimental results on aged materials (effect of

pressure, temperature and applied electrical field)

Pressure wave method

The temperature effect while measurements on material properties is substantial. Among such
effects are degradation by oxidisation, which promotes PE radical reactions at very high
temperature, which may reduce the breakdown field in the cables; increase in material density
during successive annealing, due to the reorganisation of molecular chains, which may increases
the breakdown behaviour of dielectric [6]; prolonged high temperature which increases diffusion
of impurities or of oxygen from the electrodes into the insulating material, causing the presence
of ionic species at the interfaces, which may have an adverse effect on the breakdown field under
direct current conditions [1].

For these reasons, the heat cycle procedure needs to be controlled, and also the measurements on
LDPE samples, (with a view to extrapolation to the LDPE, HDPE and XLPE used in high
voltage cables).

The maximum temperature chosen for annealing periods is the highest temperature withstood by
LDPE cables without deformation. The electrodes are short-circuited during heating (which
tends to reduce behaviour variances between samples), and then the samples are cooled at
ambient temperature.

Polarisation sequences are limited to 8 hours, the field applied is 30 kV/mm and the polarisation
temperature is from 20 °C to 70 °C. The heat cycle sequences when followed by polarisation
tend to stabilise sample behaviour.

On the other hand, it is important to know whether the polarisation phase has a more important
role to play than the thermal treatment phase on charge migration.

Results on LDPE samples have shown that:

• when the first heat cycle is undertaken before electrode bonding, the development of the
space charge is rapid, and the charge densities are substantial.

3-7
0
Characterisation of Polyethylene Slabs and Cables by Space Charge Measurements

• there is a tendency to a reduction in charge density with the number of heat cycle-
polarisation sequences.
• if the initial annealing duration increases, the density of the charges trapped in the isolation at
the time of first polarisation is reduced [1].
• in an annealing-polarisation sequence, the temperature effect can be shown to condition the
sample ageing: the longer the heat cycle, the better the dielectric withstand.
• The physical and chemical phenomena involved are essentially the result of the diffusion of
impurities from the isolation electrode interface, or of oxidisation, or of the increase in the
crystal content in the isolation material itself [7].

Interpretation of charge distribution from signals obtained by this method:

General behaviour when voltage is applied is governed by the presence (or absence) of
homocharges (charges of the same polarity as the electrode) close to the two electrodes. Also, a
large zone of positive space charges is concentrated in the centre of the isolation material.

A temporary change is clearly perceptible.

It has been clearly established that a high temperature directly affects the kinetics of the
construction or relaxation of the space charge. Similarly, the electrical field positively impacts
signal change.

Effect of electrical field: the following figures (Figure 3-5) illustrate the fact that the density of
charges trapped on this type of material rises with the polarisation field. If the temperature is
kept unchanged, the maximum quantity of trapped charges as a function of the electrical field
can be obtained.
standardised amplitude
standardised amplitude

standardised thickness standardised thickness

Figure 3-5
Standardised charge signals at the end of each polarisation (a) at 20°C, (b) at 70°C [1].

In conclusion, heterocharges dominate at ambient temperature, whereas at 70 °C, charges of

the same polarity as the nearest electrode become perceptible at the time of polarisation.

3-8
0
 Characterisation of Polyethylene Slabs and Cables by Space Charge Measurements

The charge dynamics are far less perceptible at 20°C than under hot conditions, when balance is
achieved in a few hours.

As soon as the voltage is cut off, the presence of zero electric field planes prevent the charges
crossing the isolation material.

The duration of polarisation and the field value applied both affect the amplitude of the space
charges observed .

Thermal wave method

Experimental parameters such as the effect of pressure, or temperature, or the applied electrical
field value are so many parameters which can influence the distribution of the space charge. For
example, it appears that a pressure increase may cause a reduction in the density and distribution
of the charge.

Pressure effect:

Experimental results obtained on HDPE and XLPE samples at 30°C and 70°C for pressures from
50 bar to 300 bar (in the case of cable studies in the submarine environment) and for 10 to 30 kV
have shown that the overall quantity of charge reduces as pressure is increased (Figures 3-6 and
3-7).

The effects of other experimental parameters have been evaluated by comparing a material with
crosslinked matrix (XLPE) to a non-crosslinked matrix (HDPE).

Figure 3-6
(a) distribution of electrical field E=f(x) for different pressures (graph A : 300 bar, graph B :
150 bar, graph C : 50 bar, 30kV and 70°C)

3-9
0
Characterisation of Polyethylene Slabs and Cables by Space Charge Measurements

Figure 3-7
(a) distribution of space charge ρ=f(x) for different pressures (graph A : 300 bar, graph B :
150 bar, graph C : 50 bar, 30kV and 70°C) [5]

In XLPE, the pressure variation for given values of temperature and voltage will cause a
quantitative variation in the charge. On the other hand, temperature variation (other parameters
being fixed) appears not to quantitatively modify the charges.

Comparison with HDPE shows that:

• variation in pressure also causes a quantitative modification in charge which is, however,
substantially less than in the case of the XLPE.
• a higher temperature (70 °C) causes more charges to accumulate than at 30 °C. This is the
contrary of XLPE results, for a given voltage of 30 kV. On the other hand, for a 10 kV
voltage, the temperature effect is no longer significant.

To summarise, in HDPE, the charge quantity is substantially and less than the one in XLPE, and
appears to be greater at 70 °C than at 30 °C. It would appear that the phenomena in question
are more complex, due to the heterogeneity of the materials (this may also be related to the
manufacturing process). The charge distribution on XLPE samples is more homogenous,
presenting a zone of positive charges in the centre and negative charges at the ends. These
differences are thought to be linked to the type of carries brought into play (mainly electrons in
XLPE and ions in HDPE).

In conclusion, phenomena such as electron and hole injections, and polarisation have been
shown to occur, and to be influenced by pressure in the anode, volume and cathode zones. The
basic factors for space charge formation are therefore the free volume, depth of the traps, their
energy levels, and the mobility of the carriers.

Measurements were undertaken on a 20 cm length cable, with 11 mm isolation thickness at a

temperature of 70 °C, a polarisation voltage of 45 kV DC, formation time 89 hours and pressure
300 bars. Measurements of the electrical field and of space charges were carried out after 48
hours of short circuit conditions. The maximum residual field (1.8 kV/mm) is less than the
average applied field (4.1 kV/mm) [5] (Figures 3.8 to 3.10).

3-10
0
 Characterisation of Polyethylene Slabs and Cables by Space Charge Measurements

Figure 3-8
(a) distribution of space charge ρ=f(x) in XLPE for a voltage of 30 kV, temperature of 30°C
and pressure 50 bar

Figure 3-9
(b) distribution of space charge ρ=f(x) in XLPE for voltage 30 kV, temperature of 70°C and
pressure 150 bar

Figure 3-10
(b) distribution of space charge ρ=f(x) in HDPE for voltage 30 kV, temperature of 70°C and
pressure 150 bar

3-11
0
Characterisation of Polyethylene Slabs and Cables by Space Charge Measurements

The corresponding charge density is a maximum of 18 mC/m3 , relatively lower than under the
same conditions when atmospheric pressure applies (50 mC/m3). At 300 bars, however, a
reduction of 65% in charge density is observable.

The interest of high pressure measurements is to extrapolate to normal conditions of seabed use
(depth up to 3,000 m, at 70 °C).

Proposed theoretical explanation of evolution of space charges

Pressure wave method

A qualitative analysis of space charge dynamics (Figure 3.11) shows that the charges pressed
against each of the electrodes can only be released by a transfer from the insulating material to
the semi-conductor, as they are of the same polarity as the electrodes. This tends to distance them
from the interface, due to the local field. It is therefore likely that the discharge process is
mainly due to the re-combination of the charge carriers in the volume of the insulating material.

Figure 3-11
Space charge dynamics at the outset of sample de polarisation at 20°C. The arrows
symbolise the charge injection [1].

Furthermore, a quantitative digital treatment [1] of current measurements and field distributions
may enable a calculation of the displacement currents, and hence the conduction currents at all
points in the sample, during charging and discharging. This would make it possible to translate
transport phenomena and charge transfers at all points and at all times.

P.Morin has shown the correlation of outer current measurements during polarisation. The Many
and Rakavy model [8] makes it possible to determine effective mobility in the PE, and the
corresponding activation energy. Hence, on the basis of these values and of the conduction
current, the effective carrier concentrations can be deduced.

The pressure wave method provides insight into the nature and density of charge carriers in
polyethylene materials, as their mobility is sufficiently low to be detectable.

3-12
0
 Characterisation of Polyethylene Slabs and Cables by Space Charge Measurements

It may also be possible to couple the use of this method with that of thermally stimulated currents
to determine the trap densities in the gap. This knowledge is essential if conduction phenomena
in insulating materials are to be characterised.

Thermal wave method

Given that [5] basic factors in space charge formation are the free volume, the depth of traps and
carrier mobility, simple theoretical models based on the existence of a number of zones where
positive or negative carriers predominate have been proposed.

The results of such calculations highlight the observable difference between non-crosslinked
materials and crosslinked materials:
• for the various voltage values applied at fixed pressures and temperatures, the carriers do not
all have the same mobility along the cross-section of XLPE and LDPE samples.

In XLPE, average mobility at electrode level is of the order of 10-16 to 10-17 m2/V.s. In the
centre however, the average mobility is 10-15 m2/V.s. Mobility values in HDPE are
substantially the same as those in XLPE. Generally speaking, for materials of this type, the
carriers are more mobile in the centre of the sample.
• pressure and temperature have very little effect on mobility; on the other hand, mobility
reduces by one factorial of ten for applied voltage in the range of 10 kV to 30 kV.
• this model should make it possible to estimate energy levels in the trap, and the influence of
the electrical field and pressure on this energy. Trapping and untrapping phenomena may
occur. Preferably, these should be taken into account to improve the model, and hence the
parameters influencing space charge.

A more complex model could thus be set up using other hypotheses for the calculation of
mobility and trap energy levels.

Conclusion

Space charge measurement is a very interesting method for the study of electrical ageing ; it can
detect a very important field that could lead to a breakdown. Experiments are actually done up to
year 2002 in universities in Europe in order to correlate space charge results with XLPE ageing
and develop an ageing model.

3-13
0
0
4
CORRELATION OF ELECTROLUMINESCENCE WITH
EXTRUDED CABLE AGEING

Electroluminescence Measurement

Definition of methods and its applications on material films

Electroluminescence (EL) is a term covering all the mechanisms causing light emission when a
potential difference is applied to a given material (contrary to photoluminescence or
electromagnetic excitation). Electrical energy is hence totally or partially transformed into light
energy hν.

This method was first used for inorganic materials (SiC) and inorganic semi-conductors (GaAs),
and was then validated on organic semi-conductors. Light emitting structures
(electroluminescent diodes) were hence developed.

Luminescent emission is also possible in polymers, although the phenomenon is adversely

affected by polymer degradation under electrical stress. The understanding of these phenomena
is therefore important, in order to control and assess the effects of this method on the
sustainability of the electrical properties or ageing of polymers in cables [9].

The principle of the method used, and sample preparation are as follows:

EL may generate two different mechanisms [10] :

• Ionisation impact mechanism:
In the presence of an electrical field, an electron may acquire sufficient kinetic energy to
excite another electron by impact. In this way, the number of free electrons in a material will
be multiplied by successive collisions. If a combination of holes and electrons are both
involved in the process, an avalanche ensue. The pairs created in this process will contribute
to light emission.
• charge recombination mechanism:
This occurs when the material is subjected to alternating stresses. The electrons and holes
injected into the material during the negative and positive half phases in alternating current
are captured by the different centres deep in the insulating material acting as traps. Beyond a
given threshold voltage, light emission is observable during the positive alternating half
phase, and is the result of a re-combination of the injected holes with the electrons trapped

4-1
0
Correlation of Electroluminescence with Extruded Cable Ageing

during the previous half cycle. Similarly, during the following half cycle, light emission will
occur when the ejected electrons re-combine with the holes captured in the deep-lying traps.

The concepts used to describe polymer luminescence are [9] : the molecular concept (relaxation
of excited molecule: Figure 4-1), and the band theory concept (Figure 4-2).

Figure 4-1
Jablonsky’s diagram molecular concept of luminescence

Figure 4-2
Band diagram of three categories of material : (a) isolation, (b) semi-conductor, c)
conductor

4-2
0
 Correlation of Electroluminescence with Extruded Cable Ageing

The electroluminescence of polymers with small gap may lead to the application of
electroluminescent diodes (DEL).

For polymers with higher gap, under the excitation of the direct current field, the ionisation of a
centre with electron transmission, due to tunnel effect may occur (Valence bond Æ Conduction
band ; Zener effect); alternatively there may occur inelastic collisions between hot electrons and
the localised centre; or the generation of charges of the two polarities with re-combination
(Electroluminescent diodes).

The existence of hot electrons in polymer materials has been recognised, but their involvement in
ageing and dielectric breakdown is a matter of controversy.

Knowledge of space charges, both trapped and mobile, fixed by the electrical field, is related to
electro-luminescent excitation.

Electro-luminescence makes it possible to define the processes (Figure 4-3) involved in the
polymer reactions (excitation, ionisation, as a function of electron kinetic energy).

The device includes an excitation device, an optical detection enclosure for spectral analysis, and
electrical power supply (application of high voltage at 104V) (Figure 4.4). Luminescent
excitation is generated by high intensity UV monochromatic radiation from a 150 W xenon
lamp. This radiation provides a continuous spectrum in a wavelength ranging from 200 nm to
1000 nm.
electronic density

impact impact
excitation ionisation
mobile space charge
fixed space charge

internal
ageing rupture

kinetic energy (eV)

bottom of BC

Figure 4-3
Electron energy distribution curve in a material with a large band forbidden band

4-3
0
Correlation of Electroluminescence with Extruded Cable Ageing

monochromatic
radiation from photomultiplier
motor
Xenon lamp
window for UV - visible

camera
CCD

mobile mirror

focal lens
heating
high voltage
vacuum
gas
controled Temperature

fluid

Figure 4-4
Configuration of measurement enclosure

The samples studied are polyethylene films of 40 µm to 120 µm thickness.To study the
electroluminescent excitation mechanisms in polymer isolation material, there are two types of
test:
• Divergent field tests, in a point-to-point or point-plane system to locate the discharge. For
thermoplastics, point moulding is undertaken for heat hardening materials. These materials
are directly polymerised around the point (Figure 4-5). Optical measurements are made
normal to the electrode axis. This configuration provides a check on the electrical stability of
the sample, and establishes a link between electroluminescence and electrical ageing.
• uniform field tests: a thin film is placed between two plane electrodes, of a thickness ranging
between 10 µm and 100 µm. The field is therefore homogenous over all the material (Figure
4-6). This configuration makes it possible to study the electroluminescence of all the
isolation materials in film form.

The preparation of samples requires a constant thickness of polymer films (gold cathode 30 nm
atomisation of fine particle deposits of semi-transparent metal, or 15 nm silver evaporation over
a 5 cm diameter).

The main results so obtained highlight the influential parameters.

4-4
0
 Correlation of Electroluminescence with Extruded Cable Ageing

distance between
electrode electrode-surface (mm)

polymer bloc

lens optical detection

mass electrode

distance between
electrode-mass (mm)

Figure 4-5
Divergent field configuration

photomulitiplier

electrode

polymer film metallisation

mass electrode

silicone layer

Figure 4-6
Uniform field configuration

Parameters influencing EL mechanisms

Excitation mechanisms

To understand electroluminescent excitation mechanisms in polyethylene, measurements took

place under alternating voltage. The following parameters were studied:
• Sample thickness effect: a study of luminescence as a function of the applied electrical field
was performed on polyethylene films of thickness ranging between 25 and 100 µm. This
study demonstrated the presence of an apparent threshold of light emission around 15-20
4-5
0
Correlation of Electroluminescence with Extruded Cable Ageing

kV/mm, the degree of precision being of the order of the cps (cycle per second). Beyond this
threshold field, light emission varied with the thickness of the film.
• Electrical field effect: for a LDPE containing no additive, it was noteworthy that for an
operating gradient between 10 kV/mm and 20 kV/mm, the creation of a space charge resulted
locally in large electrical field values. Electroluminescence was therefore due to the injection
of electrical charges from the electrodes, and to the induced effects by these charges in the
polymer material (Figure 4-7).
• Effect of stress frequency: luminescence was directly proportional to the stress frequency in
the range of 10-1 and 102 Hz (Figure 4-8) : the luminescence effect occurred with each
alternation without significant polarity effect, although light emission was 45° ahead of the
applied voltage.
• Study of the effect of the phase angle variation confirmed that the trapping of same polarity
charges took place, reducing the field at the electrodes (Figure 4-9).
• Effect of electrode metals: several metal compounds were studied, gold, silver and calcium.
Light emission is strongly dependent on the metallic nature of the electrode metal, but no
correlation was established with the injection current. It would appear that oxidation
phenomena play an important role.

Figure 4-7
Electroluminescence as a function of electrical field

4-6
0
 Correlation of Electroluminescence with Extruded Cable Ageing

Figure 4-8
Electroluminescence as a function of frequency.

Excitation

Time

Figure 4-9
Electroluminescence phase angle in relation to applied voltage.

Electroluminescence relaxation mechanisms

The identified relaxation mechanisms were as follows:

• Photoluminescence related to the presence of unsaturated groups in the polyethylene:
Fluorescence observed with maximum emission of 334 nm was associated with a presence of
carbonyl groups of enone type -CH=CH-CO-R. Phosphorescence emitted at 450 nm is linked
to the presence of carbonyl groups of di –enone type -CH=CH-CO-CH=CH-R. The
electronic transitions involved in the poly-enones are of type (π Æ π*) and (n Æ π*).

4-7
0
Correlation of Electroluminescence with Extruded Cable Ageing

• Induced luminescence by charge recombination: The photoluminescence spectrum was

compared to that of the emission induced by charge re-combination following the charging of
polyethylene by cold plasma film surface contact (Figures 4-10 and 4-11) : there is no
contribution from fluorescence in the re-combination spectrum, whereas spectrum
components at 465 nm, 500 nm and 530 nm were common to the re-combination and
phosphorescence spectra. This shows that the excitation states involved in the emission due
to charge recombination are low energy.
• The result is a re-combination process of a re-combining centre, when that centre is
sufficiently deep to durably stabilise an electrical charge.
Polyene less polar groups than poly-enone -(CH=CH)- provide localised traps around 0.1
eV ; poly-enone polar groups provide deep trapping levels at 0,5-0,6 eV.
• Electroluminescence related to plasma radiative relaxation: Electroluminescence
measurements on polyethylene films of different thickness exhibited two spectrum
components. The first component was located at 500-600 nm. The second component at 750
nm is thought to be due to plasma relaxation or to the «collective electron oscillations»
excited by the injection current at the electrode surface.

Temperature
of liquid nitrogen
Standardised
intesity

Wave length

Figure 4-10
Luminescence spectrum due to charge recombination

4-8
0
 Correlation of Electroluminescence with Extruded Cable Ageing

Temperature
of liquid nitrogen

under excitation
decreasing

Standardised
intensity

Wave length

Figure 4-11
PE photoluminescence spectrum

One of the main conclusions of this study is that the luminescent centres involved in the LDPE
luminescence are carbonyl groups, including di-enone unsaturated ones, which are thought to
play a preponderant role:

enone : -CH=CH-CO-R

di-enone : -CH=CH-CO-CH=CH-R or -CH=CH-CH =CH-CO-R

These are particular examples of the poly-enone series : -(CH=CH)nCO-R

Comparison between cables aged under accelerated processes (electrical stress)

and non-aged cables

A study relating to electroluminescence mechanisms on new and aged High Voltage cable
samples was performed at the Laboratoire de Génie Electrique (Electrical and Engineering
Laboratory) of Toulouse, and at Pirelli Cable Company.

The purpose was to understand electroluminescence mechanisms in PE cable isolation materials,

to study the effects of thermal and electrical ageing on electroluminescence, and to validate these
phenomena observed on cable samples.

Cables studied

Comparative measurements on new and aged cable isolation materials at the EDF test bench at
Les Renardières, involved on the following equipment:
• Cable A1 : new cable with crosslinked polyethylene isolation;

4-9
0
Correlation of Electroluminescence with Extruded Cable Ageing

• Cable D1 : new cable with the same crosslinked polyethylene isolation (from another
manufacturer);
• Cable D3 : after ageing of A1 for 2835 hours 30 kV/mm ;
• Cable C2 after ageing of C1 for 2775 hours at 35 kV/mm, C1 insulator being identical to the
one of A1 and D1.

The measurements were performed on polyethylene ribbons of 100 µm-120 µm thick and 5 cm
wide. Zones subjected to electroluminescence analysis were the polyethylene insulator material
close to the inner semi-conductor.

Results on new cables isolation materials

The electroluminescence measurements as a function of the applied electrical field gave similar
results for new cables A1 and D1 with a transmission threshold at 15 kV/mm (Figure 4-12). This
threshold was identical to the one obtained on polyethylene films directly prepared by raw
material suppliers Atochem and Boréalis for study purposes.

Figure 4-12
Results of comparative tests on new and aged cables

4-10
0
 Correlation of Electroluminescence with Extruded Cable Ageing

Results on aged cables

Electroluminescence obtained for a given potential difference were the lower as the cable was
subject to a strong electrical field for ageing purposes. Furthermore, the transmission threshold
(ES) increases as the electrical ageing field increases (Figure 4-13) :

A1 D3

C2

Figure 4-13
Determination of emission thresholds of A1, D3 and C2 cable samples from left to right

[Es(C2)=35 kV:mm/2775h] > [Es(D3)=30 kV:mm/2835h] > [Es(A1)=Es(D1)=0

kV:mm/t=0, new cable]

Conclusion
Results obtained on new and aged cables confirm the existence of a correlation between the
electrical field applied to the polyethylene and its electroluminescent response. However,
supplementary measurements would be necessary to understand the effect of ageing time on the
evolution of the electroluminescent effect.

1. The charge injection current excites plasmons (collectives electron oscillations ) at the
surface of the electrodes. Their relaxation may cause radiation. The optical signature of
such phenomena is at its maximum around 750 nm. This contribution to the
electroluminescence spectrum is not related to the material, but reveals the existence of an
injection current leading to the setting-up of a space charge.

2. Injected charges which are subsequently trapped, recombine under radiation on polymer
chromophores . The optical signature of such recombination phenomena centres around 500
nm, where the recombining centres appear to be the poly-enones groups in the material. This

4-11
0
Correlation of Electroluminescence with Extruded Cable Ageing

component of electroluminescence will therefore depend upon the chemical composition of

the polymer, and on its possible evolution in the course of ageing.

The poly-enones groups of the material have not been brought to light by physical and chemical
methods. However their existence is assumed, it would explain the results obtained from the
study of electroluminescence phenomena. This assumption deserves validation. This would
involve taking measurements on samples in which poly-enone groups have been shown to exist
by infrared microscopy, particularly so when working with samples taken from energy cables.

We recommend the electroluminescence method for XLPE cables ageing study because :
• non-aged and aged cables give different results ;
• electroluminescence seems to be the first step of ageing before the space charge development
in XLPE cables ;
• more work on electroluminesence measurement with cables time ageing would merit to be
done in order to estimate the lifespan of XLPE cables.

Current Questions

The measurements of electroluminescence performed on populations of new and aged

cables enable the classification of the samples by their electrical ageing parameters.
A dominant phenomenon of electroluminescence is the recombination component in the thick
samples investigated (100-120 µm). What is the significance of the observed thresholds of
electroluminescence on cable samples? Are electroluminescent thresholds correlated to a
threshold of irreversible ageing of the polyethylene isolation of energy cables?

To eliminate doubts about electroluminescent thresholds, the first solution would be to improve
the precision of the measurement. This will not be easy in the case of polyethylene, as this
material emits very little light as compared to materials made up of molecules including a large
number of conjugated bonds. Indeed, for polyethylene, luminescence can only be explained by
the presence of defects in the molecule itself, such as conjugated bonds combined with ketone
groupings. Do these thresholds depend on the field applied to the polyethylene during the ageing
process? Measurements of the cable samples aged under different electrical fields, and sampled
over time, should make it possible to check whether these thresholds are actually correlated to
electrical ageing.

The following is necessary if electroluminscence is to be developed fully as a diagnostic method

for HV and VHV energy cables:
• Use of electroluminescence correlated with the study of space charges, when continuous
stresses are applied
• Electroluminescence measurements to be performed with semi-conductor electrodes similar
to those of cable semi-conductors;

4-12
0
 Correlation of Electroluminescence with Extruded Cable Ageing

• Search for the physical significance of electroluminescence detection thresholds. Are

electroluminescence thresholds correlated to an ageing threshold for the polyethylene used in
energy cables, or are they correlated to the detection limit of the luminescence signal?
• Measurement of electroluminescence can be applied to a number of new cables, and in cables
aged under different electrical fields, and sampled by ageing time. The effect of thermal
cycles applied to cables will also be studied. Can measurements of electroluminescence be
used as a diagnostic method of the ageing of polyethylene insulating materials in a HV/VHV
cables.

4-13
0
0
5
REFERENCES

1. PhD : Thèse de Doctorat de l’Université de Paris VI - Physic, Electronic, P. MORIN,

October 1995

2. C. ALQUIE, G. DREYFUS, J. LEWINER, Phys.Rev.Lett. 47, P.1483-1487, 1981

3. W. EISENMENGER, M. HAARDT, Solid State Comm., 41, p.917-920, 1982

4. T. TAKADA, T. SAKAI, IEEE Trans.Elec.Ins.18, p.619-928, 1983

5. PhD : Thèse de Doctorat de l’Université Paul Sabatier de Toulouse - Génie Electrique, B.

GARROS, December 1992

6. E.W. FISHER, Pure and Applied Chem., 31, p.113-131, 1972

7. Y. SUZUOKI, T. FURUTA, H. YAMADA, S.O. HAN, T. MITZUTANI, M. IEDA, N.

YOSHIFUJI, IEEE Trans.Elec.Ins., 26, p.1073, 1991

8. MANY, G. RAKAVY, Phys.Rev.126, p.1980, 1962

9. PhD : Thèse de Doctorat de l’Université Paul Sabatier de Toulouse – Electronic, C. LADJI,

November 1998

10. EDF Technical Report, HM-65/98/016, February 1998, R. CLAVREUL

5-1
0
0
0
 About EPRI
EPRI creates science and technology
solutions for the global energy and energy
services industry. U.S. electric utilities
established the Electric Power Research
Institute in 1973 as a nonprofit research
consortium for the benefit of utility members,
their customers, and society. Now known
simply as EPRI, the company provides a wide
range of innovative products and services to
more than 1000 energy-related organizations
in 40 countries. EPRI’s multidisciplinary team
of scientists and engineers draws on a
worldwide network of technical and business
expertise to help solve today’s toughest
energy and environmental problems.

EPRI. Electrify the World

© 2001 Electric Power Research Institute (EPRI), Inc. All

rights reserved. Electric Power Research Institute and EPRI
are registered service marks of the Electric Power Research
Institute, Inc. EPRI. ELECTRIFY THE WORLD is a service
mark of the Electric Power Research Institute, Inc.

1001393

Printed on recycled paper in the United States

of America

EPRI • 3412 Hillview Avenue, Palo Alto, California 94304 • PO Box 10412, Palo Alto, California 94303 • USA
800.313.3774 • 650.855.2121 • [email protected] • www.epri.com