IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO.

4, OCTOBER 2013 2065

Comparative Study of Arc-Quenching Capabilities of

Different Ablation Materials
Erik Jonsson, Magne Runde, Gustavo Dominguez, Andreas Friberg, and Erik Johansson

Abstract—Gases released from polymers placed near a burning various aspects of arc behavior and current interruption in the
electric arc are known to influence the behavior of the arc. In presence of gassing polymers.
switching equipment, such ablation materials can be used to Many studies examine polytetrafluoroethylene (PTFE),
improve the interrupting capability. Current interruption experi-
ments have been carried out with the purpose of comparing the which is widely used as nozzle material in interrupters due to
arc-quenching capabilities of four common polymers. The test its durability and low gassing rate. Other works consider poly-
setup has static electrodes and two ablation polymer plates placed mers which take a more active part in the interruption process.
in parallel on both sides of the arc. Each ablation material is Among these, the most frequently studied are polyoxymethy-
tested according to an “up-and-down” procedure that determines lene (POM), polyamide (PA), poly(methyl methacrylate)
the current magnitude giving 50% probability for successful
interruption. Current is supplied from a capacitor unit, and a
(PMMA), and polyethene (PE). Each polymer releases a
sinusoidal waveform is created by means of a damped RLC circuit. different mixture of vapors when exposed to arcs. The arc
The electrodes and the ablation materials are replaced after each interruption performance depends, among other factors, on
interruption test. Polypropylene shows the best arc-quenching how fast these gases are released, as well as the amount and the
performance among the tested materials and interrupts about kind of atoms they consist of. The most important characteristic
2.7 times as high current as polyetetrafluoroethylene (PTFE),
of an efficient ablation material is the release of hydrogen
which is used as a reference material due to its low gassing. Poly-
carbonate and poly(methyl methacrylate) also greatly improve gas [1], [3], [11]. This is due to hydrogen’s superior thermal
the current interrupting properties, interrupting a 2.2–2.3 times conductivity and good dielectric properties compared to other
higher current compared to PTFE. As also pointed out by others, decomposition products. Fillers, including those affecting the
the arc-quenching capabilities of the polymers seem to increase color, also have been shown to have a significant effect on the
with an increasing content of hydrogen in the ablation polymer. ablation performance [9].
Index Terms—Ablation materials, arc quenching, current inter- The polymer ablation process is caused by various mech-
ruption, switchgear. anisms, primarily photo induced and thermal breakdown of
chemical bonds. According to modeling results, photo-induced
ablation is dominant after the arc reaches the steady-state condi-
I. INTRODUCTION tion [12]. The polymer surface absorbs most of the short-wave
radiation ( 350 nm) while the longer wavelengths are ab-

T HE USE of gassing polymers in the arcing chamber walls

of low-voltage (lv) switchgears is common. When the
polymer is exposed to the intense heat of the arc, the surface
sorbed inside the bulk of the ablation material or pass through.
The decomposition of the polymer occurs in the molten region
which has been estimated to be between 50 and 100 m thick
degrades and starts to evaporate. The gases that are produced [10].
change the properties of the arcing medium and may enhance This paper describes and applies an experimental
heat dissipation, increase pressure, and in other ways improve method with good statistical confidence for comparing the
the arc extinguishing capabilities of the device. When employed arc-quenching capabilities of four different ablation polymers.
in such a context, the gassing polymer is normally referred to as This is done by estimating the current magnitude, yielding
an ablation material. 50% probability for successful interruption for each material.
An experimental screening of gassing and arc extinction A simple current interruption setup is used, and the approach
properties of a large number of polymers was published as is similar to the so-called “up-and-down” procedure often used
early as 1982 [1]. More recent experimental work [2]–[9] for determining the withstand voltage of electric insulation
and theoretical modeling [10]–[15] provide details concerning materials.
Some of this work has been presented in a conference report
Manuscript received May 08, 2012; revised September 07, 2012; accepted [16].
October 29, 2012. Date of publication July 02, 2013; date of current version
September 19, 2013. This work was supported by the Norwegian Research II. POLYMER MATERIALS TESTED
Council. Paper no. TPWRD-00475-2012.
E. Jonsson is with the Department of Electrical Power Engineering, Norwe- The polymers selected for the present study are PP, poly-
gian University of Science and Technology (NTNU), Trondheim 7491, Norway
carbonate (PC), PMMA, and PTFE, all without fillers or any
(e-mail: [email protected]).
M. Runde is with the Department of Electrical Power Engineering, Norwe- coloring. The selected PP is semitransparent, PC and PMMA
gian University of Science and Technology (NTNU), Trondheim 7491, Norway, are fully transparent, and PTFE has its natural white color. The
and also with SINTEF Energy Research, Trondheim 7465, Norway.
polymers were obtained from a commercial supplier and came
G. Dominguez, A. Friberg and E. Johansson are with ABB Corporate Re-
search, Västeräs 721 78, Sweden. as 3-mm-thick plates. PTFE is known for being an almost inac-
Digital Object Identifier 10.1109/TPWRD.2012.2227834 tive polymer when exposed to electric arcs, and is included as a

0885-8977 © 2013 IEEE

2066 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO. 4, OCTOBER 2013

reference material. The other three materials are chosen because

they all consist of H, C, and O only, but have different structures
and compositions.
To have some additional information about the materials,,
thermogravimetric analyses (TGA) were carried out. Samples of
the polymers were heated up and the heat flow, mass, and tem-
perature were recorded as a function of time. The temperature
was increased at a constant rate of 10 C/min from 30 to 650 C,
in a N ambient. The energy needed to heat up and evaporate
95% of a 1-g polymer was determined and referred to as .
The reason for choosing 95% is that the remaining 5% mostly
consists of carbonized residuals that do not readily evaporate.
The melting point , density , evaporation energy , and
chemical structure, for each polymer, will be shown.
PP; C, 0.855 g/cm 664 kJ/g Fig. 1. Schematic drawing of the arc ablation assembly.

PC; C, 1.21 g/cm 889 kJ/g

PMMA; 160 C, 1.18 g/cm 1007 kJ/g Fig. 2. Arc ablation rig, with both copper electrodes but only one of the
polymer ablation plates is in place.

PTFE; 327 C, 2.20 g/cm 1485 kJ/g

Fig. 3. Capacitor discharging circuit for current interruption experiments with

an arc. Triggering circuitry, current, and voltage measurement devices are not
shown.

As can be seen, both melting point and evaporation energy

vary substantially among the examined polymers. the electrode gap. Use of an HV pulse instead of a thin metal
wire has the advantage of not contaminating the gap with metal
vapor. During the discharging, a damped sinusoidal current
III. EXPERIMENTAL SETUP
wave with a period of around 50 ms is created.
The arc ablation rig is shown in Figs. 1 and 2. The electrodes The charging voltage (i.e., the voltage of the capacitors when
are made of 2 6-mm copper bars and are separated with a the current interruption experiment starts) of the circuit is ad-
5-mm gap. The ablation material to be tested is placed on both justable and was typically between 200 and 800 V. The voltage
sides of the electrodes with 1-mm air gap to the copper bars. drop across the arc is around 70 V; thus, to some extent, lim-
The circuit used for the current interruption experiments is iting the current. The arc voltage can, however, be considered
shown in Fig. 3. The capacitors are charged with a dc voltage constant for charging voltages well above the arc voltage. For
source to the desired level. To establish an arc and initiate the charging voltages from 200 V and upwards, an almost perfect
current flow, a short HV pulse (6 s) creates a spark-over across linear relationship between voltage and current is seen. This
JONSSON et al.: COMPARATIVE STUDY OF ARC-QUENCHING CAPABILITIES OF DIFFERENT ABLATION MATERIALS 2067

makes it possible to preset the current level for each experiment terial are necessary to estimate with good statistical confi-
by simply stepping the charging voltage up and down according dence.
to the linear voltage–current relationship. It should also be noted The method was first published in 1948 [17], and the fol-
that even the transient recovery voltage (TRV) will be propor- lowing description emphasizes the most relevant parts, applied
tional to the charging voltage. to the present case. The procedure includes the following steps.
The frequency of the discharging current is approximately 20 1) Make some initial tests and set a starting test current as
Hz which makes the interruption easier than at 50 or 60 Hz. close as possible to the expected value.
Generally, it is the derivative of the current right before current 2) Choose a step as close as possible to the expected stan-
zero, together with the TRV, that determine how difficult an in- dard deviation. Normally, 2%–3% of the start current value
terruption becomes. Moreover, the maximum interruption levels is a good choice. An inappropriate step length results in
presented in this paper are also highly dependent on the geom- poor statistical confidence.
etry of the specific test setup. But for comparing and ranking 3) Start by making a test at the start current . If the interrup-
different ablation materials against each other, it is well suited. tion is successful, the next test should be , or if the
The arc voltage, charging voltage, and the current are mea- interruption fails, the next test should be .
sured and recorded by a sampling oscilloscope. 4) Run the experiment until at least 10 successful interrup-
For each ablation material, about 40 pairs of plates were pre- tions and at least 10 failures have been recorded. The
pared. Only one single current interruption test was performed number 10 is taken from the recommendation for testing
on each pair. To determine the mass loss, each pair was weighed the dielectric strength of electric insulators [18].
with a precision of 0.01 mg before and after the experiment. The lowest current value during the test series is assigned
Before the second weighing, the surfaces of the plates were in- the index 0, and the levels upwards are numbered
spected. If any copper droplets had stuck to the surface, these , with the step in between.
were removed, while carbonized material was not removed. Be- When considering a limited number of tests, only the smallest
fore the next test, the copper electrodes were replaced with a group, either the successful or the failed interruptions, should be
new set, to ensure as similar test conditions as possible. used in the analysis. This eliminates the effect of the values from
the ramping at the beginning of the sequence, which otherwise
IV. DETERMINING CURRENT INTERRUPTION CAPABILITY would give an incorrect estimate.
The approach employed in this study is to determine the cur- Define as the number of failed interruptions and as the
rent magnitude giving 50% probability for successful interrup- number of successful interruptions. Choose the smallest group
tion for each ablation material. This situation is similar to that of these and denote it as only. As stated before, in our case,
of impulse testing of electric insulation materials. In a dielectric is 10, and is the number of tests at the current level within
breakdown experiment, the critical flashover voltage (yielding a this group. can then be estimated as
50% probability for flashover) is denoted . Correspondingly,
the critical current is here referred to as .
Current interruption is a process that depends on several fac-
use negative sign if
tors. However, the two, by far, most important ones are the cur- (1)
or positive sign if
rent amplitude and the rate of rise of the recovery voltage. For
the present test circuit, the current can be used as the single where is the number of different current levels in the series.
measure of the severity of an experiment. This is because the The estimate of the standard deviation for can be found
recovery voltage is proportional to the charging voltage and, by first calculating the variance of the data obtained throughout
therefore, also to the current. Hence, the same statistical model the test procedure, here referred to as .
as when testing dielectric strength can be applied, but now with
the current as the variable.
Assume that the highest current each ablation polymer may (2)
interrupt follows a normal (Gaussian) distribution. It is not pos-
sible to measure this value directly, but only whether an inter- This is not the same as the variance for the normal distributed
ruption is successful or not for a given current. Consequently, highest current interruption value, which is the parameter of in-
the polymers have to be tested according to predefined current terest. The relationship between and has been em-
levels, which normally requires a large number of tests to give pirically determined to [17]
a statistically confident result.
In this case, each interruption experiment significantly dam- (3)
ages the electrodes and contaminates the ablation material. Both
the electrodes and the plates need to be replaced after each test. Equation (3) is an approximation which is valid if
This makes the laboratory work time consuming, so the number is larger than 0.3, which is true for all
of tests should be limited as much as possible, while still main- series in the present report.
taining sufficiently high statistical significance.
The so-called “up-and-down” method is chosen since it re- V. RESULTS
quires the lowest number of tests compared to other statistical Fig. 4 shows typical current and voltage waveforms from a
methods. Even with this procedure, 20 to 30 tests for each ma- successful interruption experiment. All successful tests yielded
2068 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO. 4, OCTOBER 2013

successful interruptions would become positioned

below , and the mean value from failed interruptions would
be positioned above.
All of the results are summarized in Table I, where the mate-
rials are ranked according to the determined values. As can
be seen from the values, the ablation process has a profound
effect on the interrupting capability. PP has the best properties
as ablation material and interrupts a 2.7 times higher current
compared to PTFE. PC and PMMA show more similar perfor-
mance, interrupting 2.2 to 2.3 times higher currents than PTFE.
The standard deviations are relatively small for all materials,
especially for PP and PTFE. Although there are clear differences
between the values, there is still a current range around 1500
A where PC sometimes interrupts the current and PMMA some-
times fails to interrupt. This shows the importance of using a sta-
tistically appropriate procedure when rating ablation materials.
Fig. 4. Current and voltage curves from the experiment with a successful cur- The total dissipated energy from the arc is calculated by
rent interruption at the first current zero crossing, approximately at time 0.03 s.
The current is interrupted, and the voltage across the two electrodes , stays integrating the product of the arc voltage and the current. This
at the same value as the source voltage. energy is then divided by the mass evaporated from the polymer
plates during the same experiment. This gives a quantity, here
referred to as “ablation energy” (fourth column in Table I),
which is a measure of the amount of arc energy needed to ablate
1 g of polymer which is a comparable unit to the TGA data. All
of the energy from the arc will not contribute to the ablation,
some will radiate in other directions and some will melt parts
of the electrodes. For all samples, the weight loss during the
experiment is small compared to the weight of the sample and,
therefore, it is difficult to determine very accurately. But it can
be seen that PC, PMMA, and PP all have values between 20–30
kJ/g, while three times more energy is required to evaporate the
same mass of PTFE.
Table I also lists the ablation volume, which is the ablation
energy multiplied by density. This is the energy needed to evap-
orate one cubic centimeter of the material. Dimensional changes
of the nozzle in a switchgear are a problem, and this is an im-
portant property when evaluating long-term performance of ab-
Fig. 5. Current and voltage curves showing a failed current interruption at the lation materials. PTFE, which is widely used as nozzle material,
first current zero crossing. This leads to reignition of the arc, and the current
flowing for one more half cycle. needs 6–8 times more energy to evaporate one unit of volume
compared to the other three tested materials. PC requires slightly
more energy than PP and PMMA.
similar traces with only minor variations in the shape of the The values obtained from the TGA are also listed in
voltage across the electrodes. In Fig. 5, curves from an inter- Table I. is measured during slow conduction heating and
ruption failure are shown. The arc reignites and the current con- not from arc radiation, and the values are much higher than the
tinues for one more half cycle. ablation energy. Even if cannot directly be compared to the
Fig. 7 shows the electrodes after one interruption experiment. ablation energy or linked to the arc ablation interaction, it is still
The arc damages the copper electrodes significantly and some obvious that correlates well with high-current interruption
metal is melting and evaporating. This is not ideal but the elec- capability for the tested polymers.
trodes are shifted to a new set between each test, ensuring equal Table I also shows the atomic percentages of H, C, and O
start conditions for each test. atoms in each material. PP, which has the highest hydrogen
The great advantage of using the “up-and-down” sequential content, also shows the highest interrupting capability, whereas
procedure is that it automatically centers most of the test cur- PMMA and PC having lower hydrogen content also yield lower
rents around . This gives narrow confidence intervals with values. This observation supports the earlier finding that hy-
a limited number of tests. In Fig. 6, the results from all current drogen is an efficient arc-quenching gas [1], [3], [11].
interruption tests are shown with a separate diagram for each The appearance of the ablation polymer surfaces after the
ablation material. In each diagram, the test procedure can be fol- interruption differs substantially among the tested materials.
lowed, and the calculated values and the associated standard The PC samples become very black and carbonized while
deviations are indicated. If the procedure had been continued PMMA remains clean and unburned. PP looks fairly similar to
for an infinite number of tests, the mean value of tests giving PMMA in this aspect. This indicates that the polymer chains
JONSSON et al.: COMPARATIVE STUDY OF ARC-QUENCHING CAPABILITIES OF DIFFERENT ABLATION MATERIALS 2069

Fig. 6. “Up-and-down” diagrams for all tested polymers. In each diagram, the calculated value and the standard deviation are indicated. The calculations of
the values here are based on the failed interruptions, since more successful than failed interruptions were recorded in every series. This eliminates the influence
of the successful interruptions in the beginning for each series. The increasing trends at the ends of the PC and PMMA series are a coincidence.

gassing from the polymers enhances the current interruption ca-

pability to a large extent. By using the “up-and-down” proce-
dure, a satisfactory statistical confidence level can be obtained
without performing an excessive number of experiments.
The TGA measurements show that the evaporation energy
follows the opposite trend as the critical currents . Since
only four materials have been tested in this study, it can be a co-
incidence, but still a clear trend appears among these materials.
Even though evaporation from thermal conduction heating dif-
fers from the arc-polymer interaction, it is likely that low evap-
oration energy is favorable for good arc extinguishing perfor-
mance.
Fig. 7. Electrodes after one current interruption experiment. The electrodes are
Pressure increase in an arc chamber has previously been
partly melted and evaporated. The darker areas on the ablation material plate studied during ablation of PE, POM, PA66, and PMMA [3].
are contaminations from the arc experiment. One of the two ablation plates is The pressure buildup in these experiments was found to in-
removed.
crease with increasing content of hydrogen in the polymers. In
general, it is well known that the current interruption capability
of a gas will benefit both from higher static gas pressure and
of PMMA and PP evaporate in a more homogeneous and com- from larger gas flow. Here, the setup is open and the static
plete manner without leaving much carbonized residuals on the pressure remains fairly constant. The release of hydrogen
surface. Moreover, as shown before, PC contains benzene rings only contributes to increase the gas flow, but still efficiently
with a large amount of carbon atoms and is known for easily enhances the interrupting performance.
becoming carbonized and black when exposed to intense heat. It is reasonable to assume that also the polymer structure in-
Hence, the blackening of the PC samples is to be expected. fluences the evaporation process. A logical hypothesis is that a
less complex polymer evaporates more homogeneously, leaving
a smother surface and with less contamination from partly de-
VI. DISCUSSION
graded polymers. This would be favorable and contribute to
Even though the setup is relatively open and the ablation ma- make the ablation material more efficient and presumably more
terials are placed as plates only on two sides of the arc, the durable.
2070 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO. 4, OCTOBER 2013

TABLE I
RESULTS FROM THE ELECTRICAL EXPERIMENT AND TGA

The electrodes are made of copper, which is not an ideal ma- [4] D. Gonzalez, H. Pursch, and F. Berger, “Experimental investigation of
terial for this purpose since it is easily eroded by the arc. The interaction of interrupting arcs and gassing polymer walls,” presented
at the IEEE Holm Conf. Elect. Contacts, Minneapolis, MN, 2011.
copper vapor from this erosion may effect the value to some [5] E. Debellut, F. Gary, D. Cajal, and A. Laurent, “Study of re-strike phe-
extent. However, it is believed that the order of ranking between nomena in a low-voltage breaking device by means of the magnetic
the tested ablation materials is not changed to any notable extent camera,” J. Phys. D: Appl. Phys., vol. 34, pp. 1665–1674, 2001.
[6] P. Rodriguez, J. Didier, G. Bernard, and S. Rowe, “Arc contact insu-
by this erosion. Replacing the electrodes before every test en- lating wall interaction in low voltage circuit breakers,” IEEE Trans.
sured that the initial geometry was the same in all experiments. Power Del., vol. 13, no. 2, pp. 480–488, Apr. 1998.
For use of ablation material in a switching device, it is impor- [7] H. Ito, T. Sakuta, and T. Kobayashi, “Measurement of arc properties
and spectrum intensity in AC arcs surrounded by plastics,” Elect. Eng.,
tant for the ablation material to not leave carbonized residuals vol. 140, pp. 38–47, 2006.
on the surface since this is likely to decrease the interruption ca- [8] Q. Yang, M. Rong, A. Murphy, and W. Yi, “The influence of medium
pability and the lifetime. Hence, a natural and obvious continu- on low-voltage circuit breaker arcs,” Plasma Sci. Technol., vol. 8, pp.
680–684, 2006.
ation of the present study would be to carry out a series of many [9] L. Yangping, Z. Manjing, Z. Qing, and G. Bo, “Research on arc resis-
tests using the same ablation plates to evaluate the long-term tance of PTFE improved by introducing inorganic fillers,” in Proc. Int.
property of the materials. Symp. Elect. Insul. Mater., 2008, pp. 259–262.
[10] V. Nossov, B. Hage, B. Jusselin, and C. Fievet, “Simulation of the
Dimensional changes of an ablation nozzle after many inter- thermal radiation effect of an arc on polymer walls in low-voltage cir-
ruptions are also an important issue to consider. As can be seen cuit breakers,” Tech. Phys., vol. 52, pp. 651–659, 2007.
in Table I, PP and PMMA evaporate about eight times more [11] P. Andre, “Comparison and thermodynamic properties of ablated
vapours of PMMA, PA66, PETP, POM and PE,” J. Phys. D: Appl.
volume compared to PTFE, for the same arc energy. Gradual Phys., vol. 29, pp. 1963–1972, 1996.
changes of nozzle geometry are unavoidable if ablation is used [12] E. Domejean, P. Chevrier, C. Fievet, and P. Petit, “Arc-wall interaction
as an arc-quenching principle and needs to be considered when modelling in a low-voltage circuit breaker,” J. Phys. D: Appl. Phys.,
vol. 30, pp. 2132–2142, 1997.
designing nozzles. [13] Q. Ma, M. Rong, A. Murphy, Y. Wu, and T. Xu, “Simulation study of
the influence of wall ablation on arc behavior in a low-voltage circuit
VII. CONCLUSION breaker,” IEEE Trans. Plasma Sci., vol. 37, no. 1, pp. 261–269, Jan.
2009.
The main conclusions from this work are as follows. [14] C. B. Ruchti and L. Niemeyer, “Ablation controlled arcs,” IEEE Trans.
• The test rig and the circuit together with the “up-and- Plasma Sci., vol. PS-14, no. 4, pp. 424–434, Aug. 1986.
down” method are an efficient and sensitive procedure for [15] L. Niemeyer, “Evaporation dominated high current arcs in narrow
channels,” IEEE Trans. Power App. Syst., vol. PAS-97, no. 3, pp.
comparing ablation materials. 950–958, May 1978.
• Even for an open arc-quenching assembly with static elec- [16] E. Jonsson, M. Runde, G. Dominguez, A. Friberg, and E. Johansson,
trodes, the presence of ablation polymer enhances the cur- “Arc quenching performance due to ablation; comparison between
four common polymers,” in Proc. Int. Conf. Elect. Contacts, 2012, pp.
rent interruption capability to an impressive degree. 41–44.
• From a large number of experiments, with pristine condi- [17] W. J. Dixon and F. J. Massey, “A method for obtaining and analyzing
tions in each test, PP interrupts almost 2.7 times higher cur- sensitivity data,” J. Amer. Statist. Ass., vol. 43, pp. 109–126, 1948.
[18] E. Kuffel, W. S. Zaengl, and J. Kuffel, High Voltage Engineering: Fun-
rent than PTFE. PC and PMMA have very similar perfor- damentals. Amsterdam, the Netherlands: Elsevier, 2000.
mance, ranging 2.2 to 2.3 times higher than PTFE.
• High content of hydrogen combined with a clean and com- Erik Jonsson, photograph and biography not available at the time of
plete ablation process are favorable properties for polymer publication.
ablation materials.
• Among the tested polymers, low “evaporation energy” (de- Magne Runde, photograph and biography not available at the time of
fined before), as determined by a TGA measurement, cor- publication.
relates well with high-current interruption performance.

REFERENCES Gustavo Dominguez, photograph and biography not available at the time of
publication.
[1] P. F. Hettwer, “Arc-interruption and gas-evolution characteristics of
common polymeric materials,” IEEE Trans. Power App. Syst., vol.
PAS-101, no. 6, pp. 1689–1696, Jun. 1982.
[2] S. Markutsya, M. Rapeaux, and V. Tsukruk, “Intensive electric arc Andreas Friberg, photograph and biography not available at the time of
interaction with polymer surfaces: Reorganization of surface mor- publication.
phology and microstructure,” Polymer, vol. 46, pp. 7028–7036, 2005.
[3] P. Andre, W. Bussiere, E. Duffour, L. Brunet, and J. M. Lombart, “Ef-
fect of dielectric material on arc pressure and ablation measurement in
high-power apparatus,” IEEE Trans. Magn., vol. 39, no. 1, pt. 1, pp. Erik Johansson, photograph and biography not available at the time of
197–201, Jan. 2003. publication.