2014 Annual Report Conference on Electrical Insulation and Dielectric Phenomena

Study and Analysis of Conduction Mechanisms and

Space Charge Accumulation Phenomena under High
Applied DC Electric Field in XLPE for HVDC Cable
Application
A. Hascoat1-2, J. Castellon1, S. Agnel1 P. Hondaa3, S. Ammi3
1 3
Institut d’Electronique du Sud RTE
Université Montpellier 2 La Défense, France
Montpellier, France
W. Frelin2, P Egrot2. D. Le Roux4
2 4
EDF R&D Boréalis AB,
Moret-sur-Loing, France Stenungsund, Sweden

Abstract— The development of High Voltage Direct Current Moreover, the local electric field due to the presence of space
(HVDC) cables requires design according to specific criteria and charges in the insulation changes the electric field distribution.
materials with appropriate properties. Cross-linked polyethylene Space charges trapped in the insulation bulk are mainly due to
(XLPE) has established itself over the past 20 years as the most defects and/or to charge injection at the interface.
used insulation material for HVAC cables, but also more recently Consequently, dielectric properties can be modified and thus
for HVDC cables. If the electrical properties of this polymer have destructive phenomena can occur, such as breakdown.
been widely studied under AC stress, the behavior of these
materials under high DC stress is less known and needs thorough This work investigates the dielectric behavior of XLPE
investigation. It is well known that, in DC conditions, the electric under different combined thermal and DC electrical stresses.
field distribution is highly dependent on operating conditions XLPE samples have been made using DC cable compound
(thermal gradient and electric field) and can be affected by provided by Borealis with a specific process (Fig. 1). The
electric charges trapped in the insulation. The resulting space insulation is first compression moulded at 120°C in order to
charge accumulation is able to increase significantly the local form the Rogowski shape. Thereafter, thin semi-conductive
electric field, thus accelerating ageing and increasing the risk of plaques are moulded together with the pre-moulded Rogowski
breakdown. Consequently, the influence of electrical and thermal shaped insulation on both sides of the sample in a final cross-
stresses on the material properties could be a key parameter on linking step at 180°C. The dimensions of active area were 50
the ageing law for HVDC insulating material. mm in diameter with 0.5 mm in thickness (for the degassed
XLPE).
The purpose of the present work is to investigate the
dielectric behavior of XLPE insulation under different combined This paper presents the results obtained with this type of
thermal and DC electrical stresses. The first step is to evaluate samples under different thermal and DC electrical stresses.
the volume electrical conduction and interface injection The dielectric behavior of XLPE samples is studied, first,
mechanisms by using current versus voltage (I-V) measurements. considering the volume electrical conduction and interface
The second step is to investigate the development of space injection mechanisms. DC electric fields from 2 kV/mm up to
charges by using a non-destructive space charge measurement 60 kV/mm have been applied at three different temperatures
technique (the Thermal Step Method).
70, 80 and 90°C. The second step investigates the
Keywords — HVDC Cable; XLPE; Conduction mechanisms; development and the accumulation of space charges in XLPE
Space charge; Electric field. by using a non-destructive space charge measurement
technique (the Thermal Step Method).
I. INTRODUCTION
Recent improvement on electronic components makes
HVDC solutions today really attractive. Polymeric insulation
(XLPE) allows a higher operating condition compared to mass
impregnated paper and are commonly used in HVDC links.
The DC electric field is strongly different from the Laplace
field (in AC) as the electrical conductivity of the insulation
depends on temperature and local electric field [1], [2]. Fig. 1. XLPE sample used in this study

978-1-4799-7525-9/14/$31.00 @2014 IEEE 530

 II. ELECTRICAL CONDUCTION AND INTERFACE INJECTION A. Study of injection phenomena
MECHANISMS Schottky injection is investigated by plotting ln(J)=f(√E)
Measurements are made on samples placed in a with linear scale (Fig. 4). This phenomenon controls the
temperature-controlled furnace at 3 different temperatures charge injection in the linear curve areas. The theoretical
70°C, 80°C and 90°C. The experimental setup is shown expression of the Schottky injection current is indicated below
schematically in Fig. 2. Surface current conduction is removed [4]:
using a guard ring. The applied voltages (up to 30 kV) were
provided by a low residual ripple 35 kV HVDC power supply. ⎛ w − βS E C ⎞
The current measurements were carried out with a Keithley J = AS T² exp ⎜ − 0 ⎟ (1)
⎜ k bT ⎟
6517A electrometer. ⎝ ⎠

The current-voltage characteristics (I-V), shown in Fig. 3, Where J is the current density (A/m²), AS the Richardson
were plotted based on the measured quasi-steady state current constant (1.2.106 A.m-2.K-2), T the temperature (K), w0 the
(in our study, quasi-steady state currents were obtained after injection level (eV), βS the Schottky constant (4.10-24 C.m for
10000 s). The principle of this technique is based on the XLPE), EC the electric field at cathode (V/m) and kb the
application of a DC voltage across a sample for a long time. Boltzmann constant (1.38.10-23 J/K).
When the transient polarization phenomenon is finished, the However, the electric field at the cathode EC is influenced
current reaches a steady state value (conduction current). After by space charges at the interfaces. Consequently, this
removing the voltage, the sample is short-circuited for a time parameter cannot be determined easily. Therefore, it has been
equivalent to the measurement duration to ensure its complete proposed to quantify the electrical contact state with a
depolarization. parameter named γ (without unit), such as EC = γE. γ<1
I(V) results show a non linear phenomenon (Fig. 3) which indicates the presence of homocharges at electrode-insulation
indicates the presence of several conduction phenomena. The interface, whereas γ>1 indicates the presence of heterocharges.
conduction current characteristics tend to decrease after a Homocharges are commonly injected at the electrodes, while
threshold electric field value, comprised between 20 and 40 heterocharges are intrinsic to the material and can possibly
kV/mm, depending on the temperature. This phenomenon originate from the peroxide decomposition products [5], [6],
corresponds to a combination of charge injection and trapping [7].
[3]. Other phenomena could be responsible of the measured The γ parameter enables to estimate the strengthening or
current as charge injections at electrodes and/or volume the weakening of the electric field for the Schottky injection
conduction mechanisms in the insulating material. Injection [8].
and conduction phenomena are investigated by using an
appropriate choice of X-Y axes of the current versus voltage The relation between γ and the current density due to the
characteristics. Schottky injection is expressed by:

⎛ w − β γE ⎞
J = AS T 2 exp ⎜ − 0 S ⎟⎟ (2)
⎜ kbT
⎝ ⎠
The logarithm of the Schottky injection current can be
written as follow:

w 0 βS γ
ln(J) = ln(AS T²) − + E (3)
kbT k bT
The linearization parameters of Equation (3) are:
w
• Y axis origin: ln(AS T²) − 0
Fig. 2. Principle of conduction current measurements kbT

14 βS γ
• Slope:
12 kbT
70°C
10 So, the γ parameter is deduced from the slope, graphically
Current (nA)

8 determined:
80°C 2
6 ⎛ Slope k b T ⎞
4 γ=⎜ ⎟ (4)
⎝ βS ⎠
2 90°C
0 Fig. 4 shows two linear domains with different slopes. For
0 10 20 30 each domain and each temperature, the slopes and
Voltage (kV) corresponding γ and w0 are reported in Table 1.
Fig. 3. I versus V characteristics at 70, 80 and 90°C

978-1-4799-7525-9/14/$31.00 @2014 IEEE 531

 -9 1.E-05

Current density J (A/m²)

Area 1 Area 2 70°C
-11 70°C
1.E-06
-13
Ln(J)

80°C 80°C
-15 Slope 1
1.E-07
-17 90°C 90°C
1.E-08
-19 1.E+15 1.E+16 1.E+17 1.E+18 1.E+19
1 3 5 7
10 3 √E V²/d 3 (V²/m3 )
Fig. 4. Linear scale representation of ln(J)=f(√E) Fig. 5. J=f(V²/d3) plot in log scale

TABLE 1. CALCULATED SCHOTTKY INJECTION PARAMETERS AND ENERGY Space charges can be trapped in the polymer. From a
LEVELS OF INJECTION chemical point of view, the peroxide decomposition products
facilitate the trap presence [14], [15] and heterocharges [16].
Area 1 Slope w0 (eV) γ Area 2 Slope w0 (eV) γ In our study, this assumption cannot be verified through space
70°C 0.00153 1.36 3.28 70°C 0.0005 1.23 0.35 charge measurements because the samples have been degassed
80°C 0.0016 1.38 3.80 80°C 0.00027 1.21 0.13 before measurement and the high poling temperatures used
90°C 0.00165 1.4 4.27 90°C 0.00017 1.22 0.06 (above 70°C) could favor the degassing of the possible
The energy levels (w0) found in our study are between 1.2 remaining species.
and 1.4 eV. These energy levels are slightly higher than those
reported in the literature with different electrode materials III. SPACE CHARGE ACCUMULATION STUDY
(1.12 eV with gold and 1.1 eV with silver). The higher After several DC electrical poling under 2, 20, 30, 40 and
injection energy levels found in this paper can be explained by 60 kV/mm at 70, 80 and 90°C during 3 hours, the space
the use of semi-conductive electrodes (carbon black in XLPE). charges have been measured using the Thermal Step Method
Furthermore, it has been shown that the Schottky injection (TSM). All space charge measurements have been performed
phenomenon is dominant at 40°C from 3 kV/mm up to 300 on two XLPE samples poled in the same conditions. Then, the
kV/mm [9]. sample is short circuited during half an hour before measuring
In area 1 (Fig. 4), below 30 kV/mm, heterocharges near the space charges in order to evacuate charges at electrodes and
electrode increase the contact electric field at the interface ensure to study charges trapped in the insulation.
(γ>1). This phenomenon facilitates the charge injection. The principle of this method is to apply a thermal step on
Beyond 30 kV/mm, in the area 2, the heterocharges presence one face of the sample. The thermal step causes a re-balance
isn’t observed. Moreover, peroxide decomposition products of charges at the electrodes and induces a capacitive current,
present in XLPE may interact with charges and charge which can be measured [7].
injection can be facilitated [10], [11].
Fig. 6 shows space charges density profiles and total
B. Study of volume conduction phenomena trapped charge amounts for each electrical and thermal poling.
At 2 kV/mm, the electric field is too important to observe The space charge densities obtained for 2 kV/mm are very
the ohmic conduction. The Poole-Frenkel conduction low (less than 0.2 C/m3). Then, a significant increase is
mechanism has been investigated too [12]. The Poole-Frenkel observed at 20 kV/mm. The space charge accumulation is
mechanism is not dominant in this study. In consequence, it strongly influenced by the applied electric field. Independently
has been decided to investigate the Space Charge Limited of the poling conditions, results show homocharges at the
Current mechanism (SCLC). Under this law, the current electrodes, due to a dominant injection phenomenon. The
density changes with V²/d3, as shown by the following maximum values of space charges density is observed at 80°C.
theoretical SCLC expression [12]: A significant decrease in negative charges at the cathode
appears at 90°C. The positive charges, trapped at the anode,
9 V² tend also to move toward the cathode. The flow of charges is
J = εμ 3 (5)
8 d favored by the higher conduction at this temperature. Because
of the increase in electron mobility under high field at high
Where d is the insulation thickness (m), ε the permittivity temperature, the charges injected at the cathode tend to cross
(F.m-1), μ the charge mobility (m².V-1.s-1) and V the applied over the sample, promoting accumulation of positive charges.
voltage (V). Results are represented in log scale by J=f(V²/d3) Total trapped charge values have been calculated from
in Fig. 5. each space charge density distribution. By analyzing these
For each temperature, the slope is equal to about 1 for most values (Fig. 6), it is shown that the sign of the total trapped
of points, which confirm that, SCLC, can be the dominant charge changes when the applied electric field increases
volume conduction mechanism at 70, 80 and 90°C. Same (electric field threshold).
conclusion has been found in [13].

978-1-4799-7525-9/14/$31.00 @2014 IEEE 532

 -20
2 kV/mm 20 kV/mm 30 kV/mm The analysis of the results has highlighted that:
ha

de
rg
C

e 40 kV/mm
0 0.1 0.260 kV/mm0.3 0.4 0.5
Thi 70°C
k ( ) • The DC conduction properties are mainly controlled by
15 2.5
Cathode Anode the Schottky injection at interfaces and Space Charge
Charge density (C/m 3)

10 2
5 Limited Current (SCLC) in the bulk;

Total charge (μC)

1.5
0 • The sign of the total trapped charge changes for lower
1
-5 applied electric field (threshold) when the temperature
0.5
-10 increases;
0
-15
-20 -0.5 • Beyond this threshold, the amount of trapped charges
0 0.1 0.2 0.3 0.4 0.5 2 20 30 40 60 becomes positive and always increases with the applied
Thickness (mm) Electric field (kV/mm) electric field;
-20
2 kV/mm 20 kV/mm 30 kV/mm
ha

de

• This threshold seems also to be visible in conduction

rg
C

40 kV/mm
0 0.1 60 kV/mm 0.3
0.2 0.4 0.5
e

15 80°C 2.5 current measurements.

Charge density (C/m 3)

10 2
5 References
Total charge (μC)

1.5 [1] C.K. Eoll, “Theory of stress distribution in insulation of high voltage DC
0 cables”, IEEE TDEI, Vol EI-10 1, pp 27-36, 1975.
1
-5 [2] B. Aladenize, R Coelho, F Guillaumond, P Mirebeau, “On the intrinsic
0.5
-10 space charge in a power cable”, Journal of Electrostatics, 39, pp 235-
-15 0 251, 1997.
-20 -0.5 [3] J. Castellon, L. Banet, I. Preda, S. Agnel, A. Toureille, E. David, A.
0 0.1 0.2 0.3 0.4 0.5 2 20 30 40 60 Sami, M. Frechette, “Dielectric properties analysis of HDPE/SiO2
Thickness (mm) Electric field (kV/mm) nanodielectric materials for future cable insulation”, JiCable 2011, pp
-20
2 kV/mm 20 kV/mm 30 kV/mm 92-96, 2011.
ha

de
rg
C

40 kV/mm
0 0.1 0.260 kV/mm0.3 0.4 0.5
e

[4] P. Llovera, “Etude des mécanismes d’injection de charge dans les

15 Thi 90°C
k ( 2.5 )
matériaux isolants au moyen de mesures électrostatiques de déclin et
Charge density (C/m 3)

10 2 retour de potentiel”, PhD thesis, 2001.

5 [5] N. Hussin, “The effects of crosslinking byproducts on the electrical
Total charge (μC)

1.5
0 properties of low density polyethylene”, PhD thesis, 2011.
1 [6] J. Yoshida, K. Matsushima, H. Miyake, Y. Tanaka, T. Takada, “Space
-5
0.5 charge accumulation and breakdown of LDPE with acetophenone under
-10
high DC stress at various temperatures”, Conference on condition
-15 0 monitoring and diagnosis, pp 515-51, 2010.
-20 -0.5 [7] R.J. Fleming, “Space charge in polymers, particularly polyethylene”,
0
0.1 0.2 0.3 0.4 0.5 2 20 30 40 60 Brazilian journal of Physics, vol 29 n°2, pp 280-294, 1999.
Thickness (mm) Electric field (kV/mm)
[8] J. Castellon et al, “Electrical properties analysis of micro and nano
Fig. 6. Space charges density distributions (on the left) and corresponding composite epoxy resin materials”, IEEE TDEI, vol 18 n°3, pp 651-658,
total trapped charges (on the right) 2011.
This threshold appears at lower applied electric fields [9] A. Toureille, “Phénomènes de conduction à champ électrique élevé dans
when the temperature increases: 30-40 kV/mm at 70°C, 20-30 les hauts polymères”, PhD thesis, 1974.
kV/mm at 80°C, and 2-20 kV/mm at 90°C. [10] N. Hussin, G. Chen, “The trapping characteristic of low density
polyethylene in the presence of crosslinking by-products”, Journal of
After this threshold, the total trapped charge always Physics : conference 18, pp 1-7, 2009.
increases significantly with the applied electric field, [11] M. Nagao, S. Mineyuki, M. Kosaki, “Effect of acetophenone coating on
regardless of the temperature. The same electric field carrier injection into polyethylene”, IEEE 8th International Symposium
on Electrets, pp 893-898, 1994.
threshold seems to appear for the I(V) characteristics (Fig. 3).
[12] P.N. Murgatroyd, Theory of space-charge-limited current enhanced by
Frenkel effect, J.Phys. D vol 3, pp 151-156, 1970.
IV. CONCLUSIONS [13] K.S. Suh, C.R Lee, J.S. Noh, J. Tanaka, D.H Damon, “Electrical
conduction in polyethylene with semiconductive electrodes”, IEEE
In this work, the electrical characteristics of XLPE used in TDEI, vol 1 n°2 pp 224-230, 1994.
HVDC cables have been studied. The following properties
[14] P. Carstensen, A.A Farkas, A. Campus, U.H Nilsson, “The role of
were investigated: peroxide by-products on the space charge accumulation in polymeric
HVDC cables, Nordic Insulation Symposium”, pp 215-219, 2005.
• Volume conduction and injection mechanisms for
[15] A. Campus, P. Carstensen, A.A Farkas, M. Meunier, “Chemical defects
different temperatures (from 70°C up to 90°C), and and electron trapping relevant to cable dielectrics”, CEIDP 2002, pp
applied electric fields (up to 60 kV/mm DC); 155-158, 2002.
• Space charge accumulation phenomena as function of [16] M. Fu, G. Chen, L.A Dissado, J.C Fothergill, “Influence of thermal
treatment and residues on space charge accumulation in XLPE for DC
temperature (70°C, 80°C and 90°C) and electric field power cable application”, IEEE TDEI, Vol. 14, pp 53-64, 2007.
up to 60 kV/mm DC.

978-1-4799-7525-9/14/$31.00 @2014 IEEE 533