technologies

Article
An Improved Calculation Model for the Prediction of
the Wear of Coated Electrical Contacts
Haomiao Yuan and Jian Song *
Precision Engineering Laboratory, Ostwestfalen-Lippe University of Applied Sciences and Arts,
32657 Lemgo, Germany; haomiao.yuan@th-owl.de
* Correspondence: jian.song@th-owl.de; Tel.: +49-5261-702-5028




Received: 28 August 2019; Accepted: 27 October 2019; Published: 31 October 2019


Abstract: To connect terminals in a cyber–physical system, large quantities of electrical contacts are
used. In order to guarantee a high reliability of the system, the lifetime of the electrical contacts should
be very long. Thus, it is of great importance to understand the failure mechanism and then to predict
the lifetime of the electrical contacts. For the applications under high thermal and/or mechanical
loads, noble plating is a good choice, considering its inertness to oxidation. For noble plating, one of
the most critical failure mechanisms is the fretting wear. Wear debris generated in the contact area,
acting as the third bodies, will greatly influence the further wear behavior and electrical performance.
In this study, the state of the art regarding third bodies is firstly reviewed, and then the influence of
the third bodies on the wear and electrical performance is investigated, from the aspects of lifetime
and the element distributions in contact area. Finally, an example of prediction of the wear of noble
plating is shown with the consideration of the third bodies. Based on this study, by involving the
third bodies, the wear of noble plating can be predicted with a higher accuracy.

Keywords: electrical contacts; noble plating; third bodies; wear prediction

1. Introduction
In recent decades, the development of cyber–physical systems has gained a great boost, which
stimulates the demand of high-quality electrical contacts. The electrical contacts are the critical
components for the transmission of the current and signals between the devices in the cyber–physical
systems. Even if one electrical contact fails, it may lead to the shutdown of a function or even the whole
system. Therefore, it is of great importance to prolong the lifetime of the electrical contacts. Generally,
in order to improve the performance of the electrical contacts, plating is used on the base materials.
The plating can be briefly divided into two groups, namely non-noble and noble plating. For the
applications under high thermal and/or mechanical loads, the noble plating is a good choice, because
the noble materials are inert to oxygen and they have a very high corrosion resistance. However, the
electrical contacts can experience the micro motions, which result from the vibration or the different
thermal expansion coefficients of the holding materials attached to the contacts. When the temperature
increases, the holding materials will expand differently, making the electrical contacts to move different
distance from their original positions. When the temperature varies, cyclic micro motions will be
generated. Due to these micro motions the noble plating will suffer fretting wear. The wear debris can
be aggregated in the contact area and change the surface condition greatly. As a consequence, the wear
behavior and the electrical performance will be changed. Therefore, understanding the formation and
the distribution of the wear debris and its influence on the wear properties can help to predict the wear
process and the lifetime of the noble plating on electrical contacts.
The designed lifetime of the systems is usually very long, for example over 15 years or for
autonomous vehicles over 300,000 km mileages. The electrical contacts are expected to serve for this

Technologies 2019, 7, 77; doi:10.3390/technologies7040077 www.mdpi.com/journal/technologies

Technologies 2019, 7, x FOR PEER REVIEW 2 of 13

Technologies 2019, 7, 77 2 of 13
The designed lifetime of the systems is usually very long, for example over 15 years or for
autonomous vehicles over 300,000 km mileages. The electrical contacts are expected to serve for this
duration.
duration.However,
However,ititisisimpossible
impossibletotoconduct
conductsuch suchaalong-term
long-termtest
testconsidering
consideringits itsunaffordable
unaffordable
time-consumption
time-consumptionasaswellwellasasthe high
the highcost. ForFor
cost. thisthis
reason, a methodology
reason, a methodology to predict the wear
to predict on the
the wear on
noble plating should be established. Then, with this methodology, the wear of the
the noble plating should be established. Then, with this methodology, the wear of the contacts after contacts after
long-term
long-termoperation
operationcan
canbe bepredicted
predictedwith
withthe thedata
datafrom
fromthetheshort-term
short-termtesttestunder
underthethe operating
operating
condition.
condition.Since
Sinceafter
afteraalong-term
long-termapplication,
application,the thesurface
surfaceisisworn
wornand
andthethewear
weardebris
debrisisisgenerated
generated
and
andaggregated
aggregatedininthe
thecontact
contactarea,
area,the
theeffect
effectof ofthe
thethird
thirdbodies
bodiesononthe
thewear
wearbehavior
behaviorof ofthe
theplating
plating
should be analyzed and considered, in order to accomplish a precise
should be analyzed and considered, in order to accomplish a precise prediction. prediction.
Many
Manymechanisms
mechanismscan canproduce
producewear
weardebris,
debris,as asshown
shownininFigure
Figure11[1].
[1].The
Thewear
weardebris
debriscan
canbe
be
generated within different physical mechanisms, depending on materials and the sliding
generated within different physical mechanisms, depending on materials and the sliding conditions. conditions.
In
In general, the wear
general, the weardebris
debrisisisdetached
detached from
from thethe surface
surface by micro
by micro stresses
stresses due todue
the to theand
force force and
relative
relative motion. The possible mechanisms may work individually or jointly to
motion. The possible mechanisms may work individually or jointly to produce the wear debris [1]: produce the wear
debris [1]:
• Micro deformation: the surface is plastically deformed in the form of grooves and scratches.
 Micro deformation: the surface is plastically deformed in the form of grooves and scratches.
• Micro ploughing: the surface is plowed by hard particles due to such as the oxides of metal wear
 Micro ploughing: the surface is plowed by hard particles due to such as the oxides of metal
debris or the contaminants from the environment.
wear debris or the contaminants from the environment.
• Micro cutting: the adhesive junctions are broken down between the surfaces thereby resulting in
 Micro cutting: the adhesive junctions are broken down between the surfaces thereby resulting
mechanical failure of the contacting asperities.
in mechanical failure of the contacting asperities.
• Micro cracking: the nucleation and propagation of the surface cracks and voids generate the
 Micro cracking: the nucleation and propagation of the surface cracks and voids generate the
particles between the surfaces.
particles between the surfaces.

(a) (b) (c) (d)

Figure
Figure 1. Wear
Wearmicro-mechanisms
micro-mechanisms to produce
to produce wearwear
debris:debris: (a) deformation;
(a) micro micro deformation;
(b) micro (b) micro
ploughing;
ploughing; (c) micro
(c) micro cutting; (d)cutting; (d) micro
micro cracking cracking [1].
[1].

Althoughaalot
Although lotofofwork
workhas hasbeen
been conducted
conducted to to investigate
investigate thethe
wearwear debris,
debris, it isitstill
is still difficult
difficult to
to clearly identify and understand the third-body behavior, since it requires
clearly identify and understand the third-body behavior, since it requires direct observations of the direct observations of
the state
state of theof the particles
particles andand the the contact
contact areaarea during
during thethe wear
wear process.However,
process. However,most mostofofthethewear
wear
observations
observations are are conducted post factum, which means that the wear process should
post factum, which means that the wear process should be disrupted, be disrupted, and
thenthen
and the samples
the samples should
should be disassembled
be disassembled for analysis.
for analysis. During this process,
During the distribution
this process, the distribution of wear
of
particles
wear and the
particles andstate of theofcontact
the state area can
the contact areabecanchanged. Thus,Thus,
be changed. the way thetoway
characterize and reflect
to characterize and
the presence
reflect of wear of
the presence debris
wearshould
debrisbeshould
carefully beselected.
carefullyIt is found that
selected. It isbesides
foundthe traditional
that besides waythe
of measuring
traditional waythe weight lossthe
of measuring or weight
wear depth after
loss or wearthedepth
tests, after
the coefficient of friction
the tests, the canof
coefficient befriction
a good
indicator
can be a goodto interpret
indicatorthe state of contact
to interpret the statearea as well area
of contact as theaswear
well debris.
as the wearThe debris.
condition Theofcondition
the wear
debris is reflected by the coefficient of friction [2]. Stability in the coefficient of friction
of the wear debris is reflected by the coefficient of friction [2]. Stability in the coefficient of friction reveals that the
amountthat
reveals of the
the debris
amount inofthethecontact
debrisarea is constant
in the contact area[1]. is constant [1].
The morphology
The morphology of of the wear
wear debris can be different,for
debris can be different, forexample
example flakes, chips,
flakes, thin
chips, plates
thin etc.etc.
plates For
the the
For contacts experiencing
contacts experiencing oscillating motions,
oscillating the wear
motions, thedebris
wear of flake and
debris chip and
of flake forms canforms
chip be aggregated
can be
and rolled into
aggregated andcylindrical,
rolled intospherical and needle-like
cylindrical, spherical and particles, due to particles,
needle-like the oppositeduetangential forces at
to the opposite
the top and
tangential bottom
forces at surfaces
the top and of the particles.
bottom Theseofkinds
surfaces of forms can
the particles. Theseact as “solid
kinds of lubricants”
forms can act so that
as
the coefficient
“solid lubricants”of friction
so that theis reduced [1,3–5]
coefficient (Figureis2)reduced
of friction resulting[1,3–5]
in the(Figure
wear reduction
2) resulting[1]. in the wear
reduction [1].
Technologies 2019, 7, 77 3 of 13
Technologies 2019, 7, x FOR PEER REVIEW 3 of 13

Figure
Figure 2. 2. Development of coefficient of of friction
friction in in aa wear
wear process,
process, fad
fad:: the coefficient of
of adhesive
adhesive
friction;
friction; ffad,max
ad,max: :maximum
maximumvalue
valueofoffadfad
; f;steady
fsteady : the
: the steady
steady coefficient
coefficient of friction.
of friction.

Hintikka
Hintikka et et al. have also
al. have also found
found thatthat the
the loose
loose wear
wear particles
particles in in the
the contact
contact area
area cancan protect
protect
against the wear of quenched and tempered steel specimens. If
against the wear of quenched and tempered steel specimens. If the wear particles are periodically the wear particles are periodically
removed
removed from from the the interface,
interface, the the fretting
fretting wear
wear will will significantly
significantly increase
increase [6]. [6]. For
For the
the application
application of of
wheel–rail
wheel–rail contact,
contact,the thepresence
presenceofofthe the“natural”
“natural”third thirdbodies
bodiescan canhelp
helptoto limit thethe
limit wear
wearand fatigue
and fatigueof
the surface
of the surface[7].[7].
In
In investigation
investigation conducted
conducted by by the
the research
research group group of of Fouvry,
Fouvry, the the energy
energy dissipation
dissipation is is used
used toto
quantify the wear in oscillating sliding contacts. The frictional force
quantify the wear in oscillating sliding contacts. The frictional force is integrated to calculate the is integrated to calculate the
accumulated
accumulated friction friction work
work dissipated
dissipated through
through the the interface
interface and and compared
compared with with the
the wear
wear volume
volume for for
the
the metallic
metallic surface
surface [8]. When the
[8]. When the electrical
electrical contacts
contacts endure
endure the reciprocating sliding,
the reciprocating sliding, the
the wear
wear debris
debris
can
can remain
remain in in the
the interface
interface to to form
form aa thick
thick third-body
third-body layer.layer. TheThe coefficient
coefficient of of friction
friction can
can be be reduced
reduced
and
and the friction energy dissipation causing the wear of the surface also decreases. Thus, wear
the friction energy dissipation causing the wear of the surface also decreases. Thus, the rate
the wear
decreases
rate decreases [9]. When appliedapplied
[9]. When in the numerical simulation,
in the numerical with the consideration
simulation, of the formation
with the consideration of theof
the third bodies, the simulated results agree better with the measured
formation of the third bodies, the simulated results agree better with the measured results [10]. results [10].
Apart
Apart from from thethe investigation
investigation from from Fouvry’s
Fouvry’s group,group, the the energy
energy based
based approach
approach is is also
also used
used by by
other
other researchers. In [11], the intensity of the wear can be related to the friction work according to
researchers. In [11], the intensity of the wear can be related to the friction work according to
Fleischer
Fleischer [12], [12],and andby bymerging
merging to Archard’s
to Archard’s equation [13], the
equation wear
[13], coefficient
the is positively
wear coefficient is correlated
positively
to the coefficient
correlated to theofcoefficient
friction. Foko Foko et al.
of friction. Foko alsoFoko
use the energy
et al. also based
use the calculation
energy basedto simulate the wear
calculation to
of a seal ring [14]. Hence, we have used the coefficient of friction as an
simulate the wear of a seal ring [14]. Hence, we have used the coefficient of friction as an indicator indicator for the change of the
wear
for the coefficient
change ofinthe thiswearstudy.
coefficient in this study.
In
In this
thisstudy,
study, twotwo arrangements
arrangements of theofelectrical contactscontacts
the electrical are usedare to generate
used to different
generatethird-body
different
distribution.
third-body distribution. The lifetime of the arrangements is determined. Based on this, thesample
The lifetime of the arrangements is determined. Based on this, the critical critical
arrangement
sample arrangement is found.isThe element
found. The content
elementincontentthe contact
in thearea of thearea
contact critical arrangement
of the is measured.
critical arrangement is
Then a methodology to calculate the thickness reduction of the
measured. Then a methodology to calculate the thickness reduction of the critical arrangement is critical arrangement is established
with the consideration
established of the influence
with the consideration of theofinfluence
the thirdofbodies.
the third Finally,
bodies. the modelthe
Finally, is verified
model iswith the
verified
experimental
with the experimentalresults. results.

2. Materials and
2. Materials and Methods
Methods
The
The base material of
base material of the
the samples
samples waswas bronze
bronze (CuSn
(CuSn44). It was
). It was firstly
firstly cut
cut to
to the
the flat
flat parts
parts and
and then
then
half
half of
of them
them are
are deformed
deformed toto the
the spherical
spherical parts.
parts. The
The radius
radius at at the
the contact
contact point
point was
was 4.5
4.5 mm.
mm. The
The
base
base materials
materials were
were degreased
degreased in in aa degreasing
degreasing bath. Silver plating
bath. Silver plating was
was used
used asas an
an example
example of
of the
the
noble plating. The contacts were electroplated with silver to the nominal thickness of 6
noble plating. The contacts were electroplated with silver to the nominal thickness of 6 μm, which µm, which was
confirmed by anby
X-ray ® X-Ray system XDAL, from Helmut
was confirmed an fluorescence device device
X-ray fluorescence (XRF, Fischerscope
(XRF, Fischerscope ® X-Ray system XDAL, from
Fischer GmbH + Co.KG, Sindelfingen, Germany).
Helmut Fischer GmbH + Co.KG, Sindelfingen, Germany).
The wear property of the silver plated electrical contacts was analyzed by the wear and fretting
corrosion test. The test device was developed in our laboratory and shown in Figure 3. The test
 Technologies 2019, 7, 77 4 of 13
Technologies 2019, 7, x FOR PEER REVIEW 4 of 13

The wear
parameters property
are listed of the 1.
in Table silver plated
During theelectrical
test, thecontacts
normal was force, analyzed by force,
frictional the wear and fretting
coefficient of
friction and contact resistance are recorded. The contact resistance was measuredparameters
corrosion test. The test device was developed in our laboratory and shown in Figure 3. The test by the
are listed in Table 1. During
“four-point-method”. Two samplethe test, the normal force,
arrangements werefrictional force, “sphere
used, namely coefficientonofplane”
frictionand and“plane
contact
resistance are recorded. The contact resistance was measured by the
on sphere”. The “sphere on plane” arrangement means that the sphere part was clamped to the “four-point-method”. Two sample
arrangements
upper were used,
sample holder and the namely
plane“sphere
part was on clamped
plane” and to “plane
the lower on sphere”. The “sphere
sample holder; the “planeon plane”
on
sphere” arrangement consisted of the plane part to the upper sample holder and the sphere part part
arrangement means that the sphere part was clamped to the upper sample holder and the plane to
was
the clamped
lower sampleto the lowerTwo
holder. sample
test holder; the were
strategies “plane on sphere”One
employed. arrangement consisted
of the strategies was oftotheterminate
plane part
to wear
the the upper sample corrosion
and fretting holder andtest theatsphere part to the
a predefined lower
cycle, andsample
then the holder. Two test
thickness strategies
reduction were
at this
employed. One of the strategies was to terminate the wear and fretting
cycle can be indicated by the difference between the initial coating thickness and the remaining corrosion test at a predefined
cycle, and
coating then the
thickness thickness
after the wear reduction at thiscorrosion
and fretting cycle can be indicated
test. The coatingby thethickness
difference wasbetween
measured the initial
by
coating
XRF. The thickness and the remaining
XRF measurements coating thickness
were conducted on the after
sphere theandwear theand fretting
plane corrosion
separately fortest.
everyThe
coating thickness was measured by XRF. The XRF measurements were
sample. The data from the distance sensor only acted as reference, and they were not used for theconducted on the sphere and the
plane separately for every sample. The data from the distance sensor
quantitative characterization, since the remaining wear debris in the contact area distorted theonly acted as reference, and they
were not
results usedmeasurement.
of the for the quantitative characterization,
The other was to runsince the the
testremaining
until thewear debris
contact in the contact
resistance reachedarea
distorted the results of the measurement. The other was to run the test until
300 mΩ. The corresponding fretting cycle was recorded as the lifetime of the sample in the wear and the contact resistance reached
300 mΩ.
fretting The corresponding
corrosion test. Based on fretting cycle the
these tests, wascritical
recorded as the lifetime
arrangement of the
is then sample in
identified. the the
After wear and
test,
fretting corrosion test. Based on these tests, the critical arrangement
the element content of the samples with the critical arrangement was measured by the is then identified. After the test, the
element content of the samples with the critical arrangement was
energy-dispersive X-ray spectroscopy (EDS, Bruker XFlash Detector 4010, Ettlingen, Germany)
® measured by the energy-dispersive
X-ray spectroscopy (EDS, Bruker XFlash ®
with a scanning electron microscope (SEM, Detector
Zeiss EVO 4010, Ettlingen,
50 VP, Germany). Germany) with a scanning
The distribution of theelectron
wear
microscope (SEM, Zeiss EVO 50 VP, Germany). The distribution
debris was observed by the digital microscope (digital microscope VHX-2000 by Keyence of the wear debris was observed by the
digital microscope
Corporation, Osaka,(digital
Japan).microscope VHX-2000 by Keyence Corporation, Osaka, Japan).

Samples:

Figure 3. 3.
Figure Wear and
Wear fretting
and corrosion
fretting test
corrosion device
test and
device samples.
and samples.

Table
Table 1. 1. Test
Test parameters.
parameters.
Parameter
Parameter Value Value
Amplitude
Amplitude peak-peak
peak-peak 200 μm200 µm
Normal
Normal force
force 3N 3N
Temperature ambient temperature (20 ◦ C~23 ◦ C)
Temperature ambient temperature (20 °C~23 °C)
Fretting frequency 1 Hz
Fretting frequency
Relative humidity 1 Hz 35~45%
Relative humidity 35~45%

The
The wear
wear prediction
prediction is is conductedfor
conducted forthe
thesamples
sampleswith
withthe
thecritical
criticalarrangement.
arrangement.The
The fundament
fundament
for the wear prediction is based on Archard’s wear equation as Equation
for the wear prediction is based on Archard’s wear equation as Equation (1) [15]: (1) [15]:

Vwear== k·F·s,
∙ F ∙ s, (1)(1)

where
whereVwear is the
Vwear volumetric
is the wear,
volumetric k isk the
wear, specific
is the wear
specific coefficient,
wear F is
coefficient, the
F is normal
the force
normal and
force s is
and the
s is the
sliding distance. The sliding distance is calculated by Equation
sliding distance. The sliding distance is calculated by Equation (2):(2):
=2∙ ∙ , (2)
where N is the fretting cycles, and s0 is the fretting amplitude from peak to peak.
 Technologies 2019, 7, 77 5 of 13

s = 2·N·s0 , (2)
Technologies 2019, 7, x FOR PEER REVIEW 5 of 13
where N is the fretting cycles, and s0 is the fretting amplitude from peak to peak.
InInorder
ordertotopredict
predictthe
thewear
wearprocess
processofofthe
thesilver
silverplating
platingmore
moreprecisely,
precisely,many
manyaspects
aspectsshould
shouldbebe
considered.
considered. It is found that the wear curve consists of different phases, Figure 4. Phases I andIIIIwere
It is found that the wear curve consists of different phases, Figure 4. Phases I and were
investigated in this study, since the wear through often happens during these phases. In
investigated in this study, since the wear through often happens during these phases. In each phase, each phase,
the
thedominant
dominantfactorfactorfor
forthe
thethickness
thicknessreduction
reductionisisdifferent.
different.The
Thespherical
sphericalparts
partswere
werethe
thefocus,
focus,since
since
they
theyhad
hadlarger
largerthickness
thicknessreduction
reductioncompared
comparedtotothe theplane
planeparts
parts[16].
[16].

(a) (b)
Figure
Figure4.4.Wear
Wearprocess
processand
andthe
thedominant
dominantfactor
factorinineach
eachphase,
phase,schematically:
schematically:(a)(a)Different
Differentphases
phasesinin
a awear
wearprocess;
process;(b)
(b)dominant
dominantfactor
factorcausing
causingthickness
thicknessreduction
reductioninineach
eachphase:
phase:deformation
deformationofofthe
the
roughness
roughnesspeaks
peaksininSub-phase
Sub-phaseIa;Ia;wear
wearoffoffofofthe
theroughness
roughnesspeakspeaksininsub-phase
sub-phaseIb;
Ib;contact
contactsurface
surfaceisis
smoothed
smoothedwhenwhenthethewear
wearprocess
processmoves
movestotophase phaseII,II,and
andthe
thewear
wearoccurs
occurson
onthe
thespherical
sphericalsegment
segment
approximately
approximatelyforforthe
thespherical
sphericalpart;
part;after
afterlong-term
long-termoperation,
operation,the
thethird
thirdbodies
bodiesare
areaggregated
aggregatedininthe
the
contactarea
contact areaand
andinfluence
influencethe
thewear
weargreatly.
greatly.

Forthe
For theprediction
predictionofofthe
thewear
wearcurve,
curve,aamethodology
methodologywith
withfollowing
followingsteps
stepsisisestablished:
established:

• Calculation
Calculationofofthe thedeformation
deformationafter afterthe thecontact
contactofofthe thetwotwocontact
contactparts
partsconsidering
consideringthe thesurface
surface
roughness
roughnesswith withassistance
assistanceofoffinitefiniteelement
elementanalysis
analysis(FEA), (FEA),involving
involvingthe theelastic-plastic
elastic-plasticmaterial
material
behavior
behaviorand andthe thesurface
surfaceroughness.
roughness.The Thegeometry
geometryisisdrawn drawncorresponding
corresponding toto
thethereal
realsamples.
samples.
The
Theroughness
roughnessisisdefineddefinedon onthetheflat
flatpart
partaccording
accordingtotothe theroughness
roughnessmeasured
measuredwith witha aconfocal
confocal
microscope
microscopeμsurf µsurfexplorer
explorerfrom fromNanofocus
NanofocusAG, AG,Germany.
Germany.According
Accordingtoto[16],[16],thetheroughness
roughness
structure
structure for
for thethetested
testedsamples
samples cancan be simplified
be simplified as triangular
as triangular prisms. prisms.
The boundaryThe boundary
conditions
conditions
are set basedare onset the
based on the condition
condition in the test.in the test.
 • Calculation
Calculationofofspecific
specificwear wearcoefficient
coefficientusing usingdatadataininphasephaseI. I.For
Forthis
thispurpose,
purpose,more morewear
wearandand
fretting
frettingcorrosion
corrosion tests are areconducted
conductedwith withthethe critical
critical arrangement
arrangement for different
for different predefined
predefined fretting
fretting
cycles incycles in the running-in
the running-in phase. The phase. The thickness
thickness reductionreduction
is measuredis measured after According
after the tests. the tests.
According
to [16], thetospecific
[16], the specific
wear wear coefficient
coefficient kcal , using thekcal, data
usinginthe dataI and
Phase in Phase I and considering
considering the
the topography
topography
of the surface of and
the surface
its change anddue its to
change due to
the plastic the plastic can
deformation, deformation,
be calculatedcan as befollowing:
calculated as
following: !
Rsm Xm D0,i 3 2
kcal = ·h1 + 3D0,i ·h1 + 3D0,i ·h0,i ·h1 , (3)
6Rk ·s1 ·F i=1 , h0,i
= ( ∙ ℎ + 3 , ∙ ℎ + 3 , ∙ ℎ , ∙ ℎ ), (3)
6 ∙ ∙ ℎ ,
where h1 is the wear depth in Phase I based on an assumption that every triangular prism
whereexperiences
h1 is the wear the samedepth wearin Phase
depth in I based
the wear on process,
an assumptionm is thethat
numbereveryof triangular
triangular prism
prisms
experiences the same wear depth in the wear process, m is the
involved in contact area, s1 is the sliding distance in Phase I, Rsm is the measured number of triangular prisms involved
average groove
in contact
width, area, s1 the
Rk is is the sliding core
measured distance in Phase
roughness depth,I, Rhsm isisthethe measured
reduction average
of coating groove
thicknesswidth, Rk isto
parallel
0,i
the measured
the normal core roughness
force simulated depth,
by FEA,h0,i isDtheisreduction
the lengthofofcoating thickness
the deformed parallelprism
triangular to thesimulated
normal
0,i
force simulated
by FEA. by FEA, D 0,i is the length of the deformed triangular prism simulated by FEA.

 Modification of the calculated specific wear coefficient according to the topography:

simplification of the roughness on the flat part and the gaps on the roughness ribs
following Equation (4):

, = ∙ ∙ , (4)
where, according to [16], αsimplification and αgap are 0.5 and 0.9, respectively.
Technologies 2019, 7, 77 6 of 13

• Modification of the calculated specific wear coefficient according to the topography: simplification
of the roughness on the flat part αsimpli f ication and the gaps on the roughness ribs α gap following
Equation (4):
kmodi, Phase I = αsimpli f ication ·α gap ·kcal , (4)

where, according to [16], αsimplification and αgap are 0.5 and 0.9, respectively.
• Involving the influence of the third bodies after a long-term sliding:

A steady coefficient of friction indicates a stable amount of wear debris in the contact area [1].
Thus, the influence of wear debris is characterized by the change of the friction during the sliding as
shown in Figure 2 following Equation (5):

fsteady,l
αthird−body = , (5)
fad,max

where the fsteady,l is the friction of coefficient at the lth steady period, since we notice that more than one
steady coefficient of friction can exist and fad,max is the maximum value of the coefficient of adhesive
friction. The αthird-body considers the reduced friction energy dissipation which in turn decreases the
wear rate.
This coefficient is further multiplied to the kcal to modify the specific wear coefficient in phase II:

kmodi, Phase II,l = αsimpli f action ·α gap ·αthird−body ·kcal (6)

• Prediction of the wear curve:

The correlation between the sliding distance and the wear depth in phase I can be determined
according to [15]:
1  
s1 = · 1674·h1 3 + 3.3·h1 2 + 0.0026·h1 , (7)
kmodi,Phase I
where s1 is the sliding distance in phase I, h1 is the wear in phase I.
The correlation between the sliding distance and the wear depth in phase II during the first steady
state of coefficient of friction is according to [15]:

π 1
   
s2,1 = − h2,1 3 + (R − Rk )h2,1 2 + 2RRk − Rk 2 h2,1 , (8)
kmodi,Phase II,1 ·F 3

where s2,1 and h2,1 are the sliding distance and the wear in phase II during the first steady state of
coefficient of friction, respectively, and R is the radius of the spherical part.
For the lth steady period of coefficient of friction (l ≥ 2), the correlation between the sliding
distance and the wear depth is:

π 1
     
s2,l = − h2,l 3 + R − R2,l−1 h2,l 2 + 2RR2,l−1 − R2,l−1 2 h2,l , (9)
kmodi,Phase II,l ·F 3

where s2,l and h2,l are the sliding distance and the wear in phase II during the lth steady state of
coefficient of friction, respectively, R2,l−1 is the total thickness reduction at the end of the (l − 1)-th
steady coefficient of friction in phase II, which can be calculated by:
Xn=l−1
R2,l−1 = h0 + h1,end + h2,n , (10)
1

where h2,n is the wear at the end of the (l − 1)-th steady state of coefficient of friction in phase II, h0 is
the thickness reduction due to deformation after the contact of two contact parts, h1,end is the wear at
the end of phase I.
Technologies 2019, 7, 77 7 of 13

Total sliding distance in phases I and II is:

Xl
stot = s1,end + s , (11)
1 2,l

where stot is the total sliding distance in phases I and II, s1,end is the sliding distance at the end of phase I.
Total thickness reduction is:
Xl
htot = h0 + h1,end + h (12)
1 2,l

where htot is
Technologies the7,total
2019, x FORthickness reduction in phases I and II.
PEER REVIEW 7 of 13

3.3.Results
Results
3.1.Lifetime
3.1. LifetimeofofElectrical
ElectricalContacts
ContactsTested
Testedwith
withDifferent
DifferentSample
SampleArrangements
Arrangements
Thedevelopment
The developmentof ofthe
thecontact
contactresistance
resistanceisisschematically
schematicallyillustrated
illustratedininFigure
Figure5.5.The
Thelifetime
lifetime
ofthe
of thesamples
samplestested
testedwith
withdifferent
differentsample
samplearrangements
arrangementsisislisted
listed in
inTable
Table2.2. The
The “sphere
“sphere on
on plane”
plane”
arrangement was the critical one. Since both of the arrangements were used in the most
arrangement was the critical one. Since both of the arrangements were used in the most applications, applications,
thecritical
the criticalarrangement
arrangement“sphere
“sphereon onplane”
plane”waswasused
usedforforthe
thefurther
furtherinvestigation.
investigation.

Figure5.5.Development
Figure Developmentof
ofcontact
contactresistance
resistanceduring
duringtests.
tests.
Table 2. Lifetime of the electrical contacts with different arrangements.
Table 2. Lifetime of the electrical contacts with different arrangements.
Arrangement
Arrangement
Sphere on Plane Plane on Sphere
Sphere on Plane Plane on Sphere
Lifetime(cycles)
(cycles) 4 5
Lifetime 10×4 10
5.4 ×5.4 1.31.3× ×1010
5
4 4
10×4 104
5.1 ×5.1 5.95.9× ×1010
4
5.0 × 10 1.0 × 105
5.0 × 104 4 1.0 × 105 5
5.5 × 10 1.0 × 10
5.5 ×5.2
10×4 104 1.01.2× ×1010
5 5

5.2 ×6.2
10× 10
4 4 1.29.8× ×1010
5 4

Average lifetime (cycles) 4 4 5

10×4 10
6.2 ×5.4 9.81.0× ×1010
Standard deviation (cycles) 4 × 10 3 2 × 104
Average lifetime
5.4 × 104 1.0 × 105
(cycles)
Standard deviation
4 × 103 2 × 104
(cycles)

3.2. Element Content in Contact Area of Samples with Critical Sample Arrangement
The element distribution in the contact area of the samples after 3000 and 5000 fretting cycles
with the “sphere on plane” arrangement is tested with EDS in SEM, shown in Figures 6 and 7. The
Technologies 2019, 7, 77 8 of 13

3.2. Element Content in Contact Area of Samples with Critical Sample Arrangement
The element distribution in the contact area of the samples after 3000 and 5000 fretting cycles
with the “sphere on plane” arrangement is tested with EDS in SEM, shown in Figures 6 and 7. The
sphere part endures more severe wear than the plane part, indicated by the larger decrease in the silver
content in contact area [17]. After 3000 fretting cycles, only very little local wear-through took place on
the sphere part, due to the surface roughness and unevenness of the silver plating. After 5000 fretting
cycles, the silver plating is worn, but the silver element is still the dominant element in the contact area
with the content
Technologies 2019, 7,mostly over
x FOR PEER 50 at.%.
REVIEW 8 of 13
Technologies 2019, 7, x FOR PEER REVIEW 8 of 13

(a) (b)
(b)
Figure6.6.Scanning
Figure electronmicroscope
Scanningelectron
electron microscope (SEM)
microscope (SEM) images
(SEM) images of
imagesof the
ofthe wear
thewear scars
wearscars and
and
scars energy-dispersive
energy-dispersive
and energy-dispersive X-ray
X-ray
X-ray
spectroscopy(EDS)
spectroscopy X-rayLinescans
(EDS)ininX-ray
X-ray Linescansmode
Linescans modeacross
mode acrossthe
across thecontact
the contactarea
contact areaafter
area after3000
after 3000fretting
3000 frettingcycles:
fretting cycles:(a)
cycles: (a)on
(a) onthe
on
the sphere;
sphere; (b) on(b) onplane.
the plane.
the plane.

(a)
(a) (b)
(b)
Figure7.7.
Figure
Figure SEM
7.SEM images
imagesofof
SEMimages of the
the
the wear
wear
wear scars
scars and
scars and
and EDS
EDSEDSinin X-ray
X-ray
in Linescans
Linescans
X-ray mode
mode
Linescans across
across
mode the
the
across contact
contact
the area
area
contact after
area
after
5000 5000
afterfrettingfretting cycles:
cycles:cycles:
5000 fretting (a)
(a) on(a) on
theon the sphere;
sphere; (b)
(b) on(b)
the sphere; on
the the
onplane. plane.
the plane.

3.3. Distribution
3.3. Distribution of
of Wear
Wear Debris
Debris after
after the
the Wear
Wear and
and Fretting
Fretting Corrosion
CorrosionTest
Test
The distribution
The distribution of
of the
the wear
wear debris
debris after
after the
the electrical
electrical failure
failurein
inthe
thewear
wearand
andfretting
frettingcorrosion
corrosion
test is shown in Figure 8. With regard to the “plane on sphere” sample arrangement, the thirdbodies
test is shown in Figure 8. With regard to the “plane on sphere” sample arrangement, the third bodies
are more
are more easily
easily removed
removed from
from the
the contact
contact area,
area, compared
compared to to the
the “sphere
“sphere on
on plane”
plane” sample
sample
Technologies 2019, 7, 77 9 of 13

3.3. Distribution of Wear Debris after the Wear and Fretting Corrosion Test
The distribution of the wear debris after the electrical failure in the wear and fretting corrosion test
is shown in Figure 8. With regard to the “plane on sphere” sample arrangement, the third bodies are
more easily removed from the contact area, compared to the “sphere on plane” sample arrangement.
Thus, wear debris distributes in a large range outside the contact area. Third bodies influence the
wear coefficient in different ways. From the point of view of the friction energy, the wear coefficient
Technologies 2019, 7, x FOR PEER REVIEW 9 of 13
decreases with the decreasing coefficient of friction. On the other hand, since the partial wear through
took place
from aftermaterial
the base several thousand
can occurfretting
in thecycles,
contactthearea.
oxides of the
Then thewear
thirddebris
bodies from the base
consist material
of different
can occur in the contact area. Then the third bodies consist of different
composition, the silver, the base material and the oxides. The oxides were harder than the composition, the silver, thesilver
base
material and
plating, and thus
the oxides.
they canThe oxides were
increase harder
the wear. than the after
Moreover, silverthe
plating, and thus they
wear-through of thecan increase
silver the
plating,
wear. Moreover, after the wear-through of the silver plating, third bodies consisted
third bodies consisted of more oxides of the wear debris from base material, and this intensified the of more oxides of
the wear
wear debris
on the basefrom base material,
material, producingandmore
this intensified
wear debris thethat
wearcanon be
thefurther
base material,
oxidized producing
in turn. Ifmore
the
wear debris that can be further oxidized in turn. If the oxides cannot be removed
oxides cannot be removed from the contact area, they will be aggregated, blocking the current path, from the contact area,
theythus
and will causing
be aggregated, blocking
the electrical the current
failure. Since the path, and thus
removal of thecausing the electrical
third bodies was more failure. Since
difficult in the
the
removal of the third bodies was more difficult in the sample arrangement
sample arrangement “sphere on plane”, it took less time to aggregate the oxides causing the “sphere on plane”, it took
less time to
electrical aggregate
failure thearrangement.
for this oxides causing the electricalthe
Consequently, failure for this
lifetime arrangement.
of the “sphere onConsequently, the
plane” is shorter,
lifetime of the “sphere on plane” is shorter, and this arrangement is the critical
and this arrangement is the critical one. Therefore, the investigation is focused on the critical sample one. Therefore, the
investigation “sphere
arrangement is focused onon the critical sample arrangement “sphere on plane”.
plane”.

(a) (b)
Figure 8. Distribution of the wear debris of different sample arrangement: (a) on the plane part after
× 1044 fretting
6.2 × fretting cycles,
cycles, arrangement
arrangement “sphere on plane”; (b) on the sphere
sphere part
part after 1.3 ×× 105 fretting
after 1.3
cycles, arrangement “plane on sphere”.

3.4. Thickness
3.4. Thickness Reduction
Reduction of
of the
the Samples
Samples with
with Critical
Critical Arrangement
Arrangement
Since the
Since the degradation
degradationof ofsilver
silverplating
platingon
onthe
thesphere
sphereisismore
moresevere according
severe according toto
thethe
EDSEDSresults in
results
Section 3.2, the investigation is focused on the wear on the sphere parts. The thickness reduction
in Section 3.2, the investigation is focused on the wear on the sphere parts. The thickness reduction on the
spherical
on parts along
the spherical partswith thewith
along sliding
the distance is shownisin
sliding distance Figurein9.Figure
shown The phase I isphase
9. The identified with the
I is identified
rapid increase in the thickness reduction, and then in phase II, the thickness reduction
with the rapid increase in the thickness reduction, and then in phase II, the thickness reductionincreases slowly.
increases slowly.
in Section 3.2, the investigation is focused on the wear on the sphere parts. The thickness reduction
on the spherical parts along with the sliding distance is shown in Figure 9. The phase I is identified
with the rapid increase in the thickness reduction, and then in phase II, the thickness reduction
increasesTechnologies
slowly.2019, 7, 77 10 of 13

Figure 9. Thickness reduction of the spherical parts with “sphere on plane” arrangement.
Technologies 2019, 7, x FOR PEER REVIEW 10 of 13
Figure 9. Thickness reduction of the spherical parts with “sphere on plane” arrangement.
3.5. Coefficient of Friction of Samples with Critical Arrangement
3.5. Coefficient of Friction of Samples with Critical Arrangement
A typical development
A typical developmentof the
of coefficient of friction
the coefficient of frictionof the “sphere
of the on plane”
“sphere arrangement
on plane” arrangement is shown
is
in Figure
shown 10.inThe
Figuremaximum coefficient
10. The maximum coefficient of ffriction
of friction ad ,max , fand
ad,max,the
andsteady coefficients
the steady of friction
coefficients fsteady,1
of friction
and fsteady,2 are fpointed
fsteady,1 and steady,2 are out in Figure
pointed 10.Figure
out in Therefore, the influence
10. Therefore, of the third
the influence bodies
of the thirdwas calculated
bodies was in
calculated
Table 3. in Table 3.

Figure Typical
10.10.
Figure Typicaldevelopment of the
development of thecoefficient
coefficientofof friction.
friction.

Table 3. Coefficient
Table ofoffriction
3. Coefficient frictionand
and the influenceofofthe
the influence the third
third bodies.
bodies.

fad,max fsteady,1 αthird-body,1 Average αthird-body,1

Average fsteady,2 αthird-body,2 Average
Averageαthird-body,2
fad,max fsteady,1 αthird-body,1 fsteady,2 αthird-body,2
1.5 1.25 0.83 αthird-body,1 0.95 0.63 αthird-body,2
1.7 1.5
1.35 1.25
0.79 0.83 0.81 0.95
1.1 0.63
0.65 0.64
1.6 1.7
1.30 1.35
0.81 0.79 0.81 1.1
1.0 0.65
0.62 0.64
1.6 1.30 0.81 1.0 0.62

4. Discussion
4.1. Calculation and Modification of Specific Wear Coefficient
According to Equation (3), the specific wear coefficient is calculated as shown in Figure 11. The
points with the sliding distance from 200 mm to 800 mm are used for calculation. Below this range,
Technologies 2019, 7, 77 11 of 13

4. Discussion

4.1. Calculation and Modification of Specific Wear Coefficient

According to Equation (3), the specific wear coefficient is calculated as shown in Figure 11. The
points with the sliding distance from 200 mm to 800 mm are used for calculation. Below this range,
the thickness reduction is very close to the deformation simulated in FEA, so it is dominated by the
deformation of roughness peaks. Above this range, the wear can develop to phase II, and Equation (3)
is invalid. The kcal calculated at each sliding distance is in a range from 3.3 × 10−8 mm3 /(N·mm) to
5.1 × 10−8 mm3 /(N·mm). The mean value of kcal at all sliding distances is 3.9 × 10−8 mm3 /(N·mm). The
standard −9 mm3 /(N·mm).
× 10REVIEW
Technologiesdeviation is 8PEER
2019, 7, x FOR 11 of 13

Figure 11.
Figure 11. Calculated specific
specific wear
wear coefficient
coefficient and
and the
the standard
standarddeviation.
deviation.

According
According to
to Equations
Equations (4)–(6),
(4)–(6),the
thespecific
specificwear
wearcoefficient
coefficientcan
canbe
bemodified
modifiedas
aslisted
listedin
inTable
Table4.4.

Table4.4. Modified
Table Modified specific
specificwear
wearcoefficient.
coefficient.

kmodi,Phase I [mm3 /(N·mm)]

, kmodi,Phase II,1 , [mm3 /(N·mm)]
, kmodi,Phase
, II,2 [mm
,
3 /(N·mm)]

2.3 ×[mm
10−8 /(N∙mm)] [mm
1.9 ×/(N∙mm)] [mm3/(N∙mm)]
3 3
Maximum 10−8 1.5 × 10−8
Minimum −8
1.5 × 102.3 −8 −9
Maximum × 10−8 1.9××1010−8
1.2 1.5 ×9.410×−810
Average 1.8 × 10−8 1.4 × 10−8 1.1 × 10−8
Minimum 1.5 × 10−8 1.2 × 10−8 9.4 × 10−9
Average 1.8 × 10−8 1.4 × 10−8 1.1 × 10−8
4.2. Wear Curves
According
4.2. Wear Curvesto Equations (7)–(12), the wear curve can be calculated and illustrated as Figure 12. The
deviation between the average predicted value and the measured value is −0.9% after the long-term
According
operation, whichtoisEquations (7)–(12),compared
greatly improved the wear curve
to the can be calculated
deviation and illustrated
of 16% from as Figure
the prediction 12.
without
The deviation between the average predicted value and the measured value is −0.9%
consideration of third bodies and −3.6% from the prediction with the consideration of third bodiesafter the
long-term
only by the operation, which is greatly
last steady coefficient improved
of friction in [15]. compared to the deviation of 16% from the
prediction without consideration of third bodies and −3.6% from the prediction with the
consideration of third bodies only by the last steady coefficient of friction in [15].
4.2. Wear Curves
According to Equations (7)–(12), the wear curve can be calculated and illustrated as Figure 12.
The deviation between the average predicted value and the measured value is −0.9% after the
long-term operation, which is greatly improved compared to the deviation of 16% from the
prediction2019,
Technologies without
7, 77 consideration of third bodies and −3.6% from the prediction with 12 ofthe
13
consideration of third bodies only by the last steady coefficient of friction in [15].

Figure 12. Prediction

Prediction of wear curve with the consideration of the third bodies in different
different stages.
stages.

The wear of the contacts after a long-term operation can be predicted based on the wear data
collected from a short period test under the operating condition. For instance, for the automobile, the
contacts in the vehicles experiencing different mileages during the test driving are disassembled to
measure the wear and then the wear at a high mileage can be extrapolated according to the methodology
established in the study, with the assumption of the average thermal and/or mechanical loads on the
electrical contacts in the operation in each mileage being the same.

5. Conclusions
In this paper, the influence of third bodies on friction and wear was firstly reviewed. Two
sample arrangements, namely “sphere on plane” and “plane on sphere”, are used to generate different
distribution of third bodies, and then the “sphere on plane” is found to be the critical arrangement.
Third bodies influence the wear coefficient in different ways. From the point of view of the friction
energy, the wear coefficient decreases with the decreasing coefficient of friction. On the other hand,
with the occurrence of the partial wear through, which will take place after several thousand cycles
of micro motion, the base material begins to oxidize. The oxides are harder than the silver plating
and they in turn can increase the wear coefficient. In the case of “sphere on plane”, more oxide
debris will remain in the contact area, which makes the “sphere on plane” the critical arrangement.
A methodology is developed to predict the wear process with regard to this arrangement. The influence
of the third bodies is taken into consideration, and different steady coefficient of friction is referred to.
The predicted wear curve is validated by the experimental results. A good agreement between the
predicted value and the measured value after the long-term operation is obtain. The deviation is −0.9%.
The prediction precision is greatly improved, compared to the prediction without consideration of
the third bodies or only with the consideration of third bodies by the last steady coefficient of friction.
Based on this study, it is possible to predict the wear of the electrical contacts with noble plating after
long-term operation with the data from the short-term test under the operating condition.

Author Contributions: H.Y. conducted the simulation, experiments, analysis, established the methodology,
validated the calculation model and wrote the original draft; J.S. initialized the investigation and the methodology,
conducted the supervision and reviewed the paper.
Funding: This study is partly financed by the German Federal Ministry for Economic Affairs and Energy (BMWi,
IGF, 20139 N).
Conflicts of Interest: The authors declare no conflict of interest.
Technologies 2019, 7, 77 13 of 13

References
1. Zmitrowicz, A. Wear debris: a review of properties and constitutive models. J. Theor. Appl. Mech. 2005, 43,
3–35.
2. Suh, N.P.; Sin, H.-C. On the genesis of friction and its effect on wear. In Solid Contact and Lubrication;
Cheng, H.S., Keer, L.M., Eds.; American Society of Mechanical Engineers: New York, NY, USA, 1980;
Volume 39, pp. 167–183.
3. Song, J.; Schinow, V. Correlation between friction and wear properties and electrical performance of silver
coated electrical connectors. Wear 2015, 330–331, 400–405. [CrossRef]
4. Zum Gahr, K.H. Abrasiver Verschleiß metallischer Werkstoffe; VDI-Verlag: Düsseldorf, Germany, 1981.
5. Zanoria, E.S.; Danyluk, S.; Mcnallan, M.J. Formation of cylindrical sliding-wear debris on silicon in humid
conditions and elevated temperatures. Tribol. Trans. 1995, 38, 721–727. [CrossRef]
6. Hintikka, J.; Lehtovaara, A.; Mäntylä, A. Third particle ejection effects on wear with quenched and tempered
steel fretting contact. Tribol. Trans. 2016, 60, 70–78. [CrossRef]
7. Niccolini, E.; Berthier, Y. Wheel–rail adhesion: laboratory study of “natural” third body role on locomotives
wheels and rails. Wear 2005, 258, 1172–1178. [CrossRef]
8. Fouvry, S.; Liskiewicz, T.; Kapsa, Ph.; Hannel, S.; Sauger, E. An energy description of wear mechanisms and
its applications to oscillating sliding contacts. Wear 2003, 255, 287–298. [CrossRef]
9. Fouvry, S.; Laporte, J.; Perrinet, O.; Jedrzejczyk, P.; Graton, O.; Alquier, O.; Sautel, J. Fretting Wear of Low
Current Electrical Contacts: Quantification of Electrical Endurance. In Proceedings of the 63rd IEEE Holm
Conference on Electrical Contacts, Denver, CO, USA, 10–13 September 2017; pp. 1–11. [CrossRef]
10. Arnaud, P.; Fouvry, S.; Garcin, S. A numerical simulation of fretting wear profile taking account of the
evolution of third body layer. Wear 2017, 376–377, 1475–1488. [CrossRef]
11. Terwey, J.T.; Poll, G. Energy based modelling of adhesive wear in the mixed lubrication regime. In Proceedings
of the 60th German Tribology Conference 2019 (60. Tribologie-Fachtagung 2019), Göttingen, Germany,
23–25 September 2018. Presentation No. 61.
12. Fleischer, G.; Wolmirstedt, H.G.; Thum, H. Verschleiß und Zuverlässigkeit; VEB Verlag Technik: Berlin, Germany,
1980.
13. Archard, J.F. Contact and Rubbing of Flat Surfaces. J. Appl. Phys. 1953, 24, 981–988. [CrossRef]
14. Foko Foko, F.; Heimes, J.; Magyar, B.; Sauer, B. Reibenergiebasierte Verschleißsimulation für Radialwellendichtringe.
In Proceedings of the 60th German Tribology Conference 2019 (60. Tribologie-Fachtagung 2019), Göttingen,
Germany, 23–25 September 2018. Presentation No. 47.
15. Yuan, H.; Song, J.; Schinow, V. A Modification of the Calculation Model for the Prediction of the Wear of
Silver-Coated Electrical Contacts with Consideration of Third Bodies. In Proceedings of the 64th IEEE Holm
Conference on Electrical Contacts, Albuquerque, NM, USA, 14–18 October 2018; pp. 310–316. [CrossRef]
16. Yuan, H.; Song, J.; Schinow, V. Simulation methodology for prediction of the wear on silver-coated electrical
contacts with a sphere/flat configuration. IEEE Trans. Compon. Packag. Manuf. Technol. 2018, 8, 364–374.
[CrossRef]
17. Song, J.; Schinow, V.; Yuan, H. Third bodies in electrical contacts — Wear and electrical performance.
In Proceedings of the 63rd IEEE Holm Conference on Electrical Contacts, Denver, CO, USA, 10–13 September
2017; pp. 117–124. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).