-
Tribology of Si/SiO2 in Humid Air: Transition from Severe
ChemicalWear to Wearless Behavior at NanoscaleLei Chen, Hongtu He,,
Xiaodong Wang, Seong H. Kim,*, and Linmao Qian*,
Tribology Research Institute, Key Laboratory of Advanced
Technologies of Materials (Ministry of Education), Southwest
JiaotongUniversity, Chengdu 610031, ChinaDepartment of Chemical
Engineering and Materials Research Institute, The Pennsylvania
State University, University Park,Pennsylvania 16802, United
States
*S Supporting Information
ABSTRACT: Wear at sliding interfaces of silicon is a maincause
for material loss in nanomanufacturing and device failurein
microelectromechanical system (MEMS) applications.However, a
comprehensive understanding of the nanoscalewear mechanisms of
silicon in ambient conditions is stilllacking. Here, we report the
chemical wear of single crystallinesilicon, a material used for
micro/nanoscale devices, in humidair under the contact pressure
lower than the material hardness.A transmission electron microscopy
(TEM) analysis of the wear track conrmed that the wear of silicon
in humid conditionsoriginates from surface reactions without
signicant subsurface damages such as plastic deformation or
fracture. When rubbedwith a SiO2 ball, the single crystalline
silicon surface exhibited transitions from severe wear in
intermediate humidity to nearlywearless states at two opposite
extremes: (a) low humidity and high sliding speed conditions and
(b) high humidity and lowspeed conditions. These transitions
suggested that at the sliding interfaces of Si/SiO2 at least two
dierent tribochemicalreactions play important roles. One would be
the formation of a strong hydrogen bonding bridge between hydroxyl
groups oftwo sliding interfaces and the other the removal of
hydroxyl groups from the SiO2 surface. The experimental data
indicated thatthe dominance of each reaction varies with the
ambient humidity and sliding speed.
INTRODUCTIONTribological problems, such as high friction and
severe wearthat lead to energy dissipation or materials failure,
play a criticalrole in all length scales from earthquakes1,2 down
to micro/nanoelectromechanical systems (M/NEMS).3,4 Although
vari-ous contact and friction mechanisms have been proposed,
ascientic understanding of wear mechanisms at the nanoscale isstill
lacking. Normally, wear of sliding interfaces is thought tobe
material removal by mechanical separation owing tomicrofracture, by
chemical dissolution, or by melting at thecontact interfaces.58 The
wear mechanism may change withvariations in surface properties or
dynamic surface responsesduring a sliding process.9
Wear of materials is often described by the Archard lawwhich
relates the wear volume to the applied load and thehardness of
materials.57 In inert environments (such as dry orvacuum),
deformation via phase transformation,8,10 viscousow,11 and
dislocation formation12 were identied as the mainfactors for
silicon wear. With the increase of normal load, themechanical
damage on silicon could manifest as the protrusionof the surface
(forming a hillock) at low contact pressure ormechanical wear
(material removal) at high contact pressure.This transition takes
place at a normal load that surpasses thehardness of silicon (13
GPa).13 The recent nanoscaleexperiment in high vacuum proposed a
wear mechanism viaatom-by-atom removal for the materials of Si and
diamond-like
carbon (DLC), in which wear rate depends on the stress-assisted
bond dissociation of the substrate material.1417
The surface wear upon rubbing or sliding depends on notonly
intrinsic properties of materials such as hardness and bondenergy
but also many extrinsic factors of the slidinginterfaces.1119
Although the contact pressure was far lessthan the yield stress of
silicon (7 GPa), the chemical reactionsinvolving water could induce
serious damage of silicon contactinterfaces.18,19 The results
obtained in MEMS applicationsindicated that this chemical wear
would be more likely to occuron silicon surface after plasma
cleaning.20 In order to minimizeor avoid the wear of material,
depositions of hard coatings orhydrophobic organic layers on
silicon or silica substrate havebeen attempted.2123 Recent
experimental and theoreticalstudies demonstrated that water-induced
chemical wear ofsilicon substrates can be prevented in an alcohol
vaporenvironment.24 But, the success of the alcohol vapor
lubricationapproach does not advance our understanding of the
chemicalmechanism of the water-induced wear process itself.In this
study, we present the eects of sliding speed (v) and
relative humidity (RH) on the wear of single crystalline
siliconsurface rubbed with a SiO2 microsphere. Under
humidconditions, the wear of silicon originates from surface
reactions
Received: July 10, 2014Published: December 5, 2014
Article
pubs.acs.org/Langmuir
2014 American Chemical Society 149 dx.doi.org/10.1021/la504333j
| Langmuir 2015, 31, 149156
-
without signicant subsurface damages, and the degree
ofwater-induced wear of Si strongly depends on the sliding speedand
RH. It was found that the wearless state of the Si/SiO2interface
could be achieved at low RH/high v or high RH/low vconditions.
Fundamental understanding of wear mechanism inhumid air would be
useful to control or mitigate the nanoscalewear of the sliding
interfaces of silicon parts in M/NEMSapplications.
EXPERIMENTAL METHODSThe p-type Si(100) wafers with a thickness
of 0.5 mm were purchasedfrom MEMC Electronic Materials, Inc. The
root-mean-square (RMS)roughness of the silicon wafer measured by an
atomic forcemicroscope (AFM, SPI3800N, Seiko, Japan) was about 0.05
nmover a 500 nm 500 nm area. The native oxide layer on silicon
surfacewas not removed. Before nanowear test, the Si samples
wereultrasonically cleaned in methanol, ethanol, and deionized
water for10 min. The silicon wafer surface was relatively
hydrophilic and had awater contact angle of 39.All the nanowear
tests and in situ topography scanning were
performed with AFM within an environment chamber for
relativehumidity (RH) control (Figure 1). A p-doped Si(100) wafer
with a
native oxide layer was scratched with a silica ball attached to
an AFMcantilever. AFM cantilevers with spherical SiO2 tips were
purchasedfrom Novascan Technologies. The radii of the spherical
SiO2 tips weremeasured with scanning electron microscopy and
determined to be1.0 m (inset in Figure 1). The roughness of SiO2
microsphere wasmeasured as 0.4 nm over a 250 nm 250 nm area (see
Figure S1 inSupporting Information), which was smaller than the
elastic Hertziancontact deformation (1.3 nm) under our test
conditions. Using areference probe with a force constant of 2.957
N/m, the normal springconstants of the cantilevers of SiO2 tips
were calibrated as 10.513.8N/m.25 The friction forces were
calibrated by using a silicon gratingwith a wedge angle of 5444
(TGF11, MikroMasch, Germany).26 Thenanowear tests were carried out
at the sliding scan speed varying from0.008 to 50 m/s and RH from
0% to 70%. The applied normal load(Fn) was 3 N, the number of
sliding cycles (N) was 100, and thesliding distance (D) was 200 nm,
if not specially mentioned. Aftertests, the topography of wear area
was imaged with a sharp siliconnitride tip (MLCT, Veeco) which had
a nominal curvature radius of12 nm and a nominal spring constant of
0.1 N/m. All tests wererepeated multiple times to conrm the
repeatability of the data (seeFigure S2 in Supporting Information).
A selected set of wear scar onthe Si substrate was analyzed with
TEM (Tecnai G2, FEI). The cross-section sample for TEM was prepared
using a focused ion beamsystem (Nanolab Helios 400S, FEI, Holland).
During the samplepreparation, an epoxy polymer, instead of Pt, was
deposited on silicon
surface as the passivation layer to prevent the
decrystallization ofsilicon due to high-energy impact of Pt during
the deposition.
RESULTS AND DISCUSSIONTransition from Severe Chemical Wear to
Wearless
State. Figure 2 shows the topographic images of wear scars ona
silicon surface at three RH regimes (10%, 30%, and 65%).The
corresponding cross-section proles of wear scars areshown in Figure
S3 of the Supporting Information. At low RH(10%), the Si surface
wore severely when sliding speed (v) wasless than 0.1 m/s. The wear
of Si substrate was reduced as vincreased; at v > 2 m/s, there
was no visible wear on the Sisurface (Figure 2a). At intermediate
RH (30%), the wear ofsilicon resulted in grooves at all speed
conditions (Figure 2b).When wear tests were operated at relatively
high RH (65%),the silicon substrate wear was negligible at v <
0.1 m/s, andthere was sudden transition to severe wear at v >
0.1 m/s(Figure 2c). Small dierences in the length of wear
scarsformed at various RH and sliding speeds might be due to
thepile-up debris at the edge of the wear track (see Figures S4
andS5 in the Supporting Information).The entire wear rate data
measured over wide ranges of RH
and sliding speed are shown in a 2D color plot in Figure 3.
Theaverage wear rate at each condition is shown in Figures S6 andS7
of the Supporting Information. It is intriguing to note thatthere
are wearless regions in the opposite corners of the v andRH
parameter domains. In low RH conditions (from 0% to30%), the
wearless behavior was observed when v was high(lower-right blue
region in Figure 3). In contrast, the siliconsurface revealed
negligible wear at low sliding speeds when RHwas above 55%
(upper-left blue region in Figure 3). Other thanthese two regions,
the wear rate of silicon decreasedmonotonically to a constant value
as v increased or the wearrate increased to steady state as RH
increased from 0% to 70%.For the normal load of 3 N and adhesion
force (pull-o
force) of 1.35 N (see Figure S8 in the SupportingInformation) at
65% RH, the maximum contact pressure ofthe Si/SiO2 interfaces is
calculated to be 1.3 GPa based on theDerjaguinMullerToporov (DMT)
model.27 This contactpressure is much lower than the yield stress
of Si (7 GPa) andSiO2 (8.4 GPa).
28 Since SiO2 (70 GPa) has lower elasticmodulus than Si (160
GPa), the indentation deformation willbe mostly at the SiO2 sphere
at all test conditions of this study.Therefore, the mechanical
process alone cannot explain the Siwear under ambient conditions.
In fact, there was no materialremoval on Si surface when the Si
wafer was scratched with theSiO2 sphere in dry conditions at the
same mechanical loadconditions. Instead, small hillocks
(protrusion) were generatedon the silicon surface (see Figure S9 in
the SupportingInformation). This protrusion seems to be mainly due
tosubsurface deformation.11 These results indicated that
thechemical reactions induced by mechanical shear must
beresponsible for serious wear of silicon in humid air.19 This
isoften called tribochemical wear.To conrm the occurrence of
tribochemical wear without
any mechanical damage to the substrate, the microstructure ofthe
wear scar formed by sliding with a SiO2 ball in ambient airwas
analyzed by the cross-sectional TEM. Since the amorphousoxide layer
shows a featureless microstructure similar to that ofthe
passivation layer, the amorphous oxide layer on the topsurface of
silicon substrate is dicult to be identied in TEMimages. The groove
depth shown in Figure 4a was 7.5 nm,which was consistent with the
wear depth measured with AFM.
Figure 1. Nanowear test of the Si/SiO2 interfaces in humid air.
TheSiO2 microsphere with a radius of 1 m moved horizontally on
theSi(100) wafer over a distance (D) of 200 nm under applied load
Fn = 3N. The sliding speed (v) was varied from 0.008 to 50 m/s, and
thetotal number of reciprocating cycles (N) was 100. The
relativehumidity (RH) was varied within 0% and 70%.
Langmuir Article
dx.doi.org/10.1021/la504333j | Langmuir 2015, 31, 149156150
-
The good agreement between TEM and AFM wear depthmeasurements
implied that the amorphitization of Si beneaththe wear track was
negligible. Also, it was noted that the lattice-resolved image in
worn area (Figure 4b) shows the prefectcrystalline order of silicon
even very close to the sliding surface.No dislocation or defects
are observed in the subsurface ofsilicon lattice; thus, the
subsurface plastic deformation can beruled out. The absence of the
amorphous transformation orcrack formation in the silicon substrate
beneath the weartrack29,30 supports that the main wear mechanism
istribochemical, not mechanical, under the low contact pressure(1.3
GPa) and humid conditions.18,24
In previous studies, it is hypothesized that the chemical
wearproceeds via the formation of SiOSi bonds bridging the twosolid
surfaces.18,24,31 Upon the initial contact, silanol groups onthe
native oxide surface of the Si wafer can form hydrogenbonds with
the silanol groups at the SiO2 counter surface.Dehydration
reactions of these groups during the sliding could
form SisubstrateOSisphere bonds, which bridge two solidsurfaces.
The shear action may dissociate those interfacialbonds or the
subsurface SiO bonds. The latter will lead towear of silicon
substrate.18,24 During the sliding process, thetemperature rise due
to frictional heat was negligible since thesliding speed was low
(see Supporting Information for moredetails).32,33 Alternatively,
it can be conceived that themechanical stress or shear could deform
the Morse potentialof a specic chemical bond at the surface,
lowering the energybarrier for bond dissociation.34 Then, the
water-induceddissociation of the SiOSi network or SiSi network
onsilicon substrate might take place readily, which leads
totribochemical wear during mechanical shear in humid air.35,36
In any case, the data shown in Figures 2 and 3 clearly
indicatethat the degree of these tribochemical wear processes
appearsto strongly depend on the RH and sliding speed.In a previous
study, the chemical species in the wear scars on
the silicon substrate were analyzed with time-of-ight
secondary
Figure 2. AFM images of wear scars on silicon surface after
sliding by a SiO2 microsphere at various speeds v and under dierent
relative humidity.(a) RH = 10%, (b) RH = 30%, and (c) RH = 65%. Fn
= 3 N, N = 100, and D = 200 nm. Note the height full scales for two
right images in (a) andtwo left images in (c) are 5 nm, and all
others are 50 nm.
Figure 3. Nanowear map of the single crystalline silicon
surfacescratched with a SiO2 sphere (radius = 1 m) at an applied
load of 3N for 100 reciprocating cycles. There are about 150 data
points usedto show the RH and sliding speed dependence of wear
behaviors ofsilicon surface. More details can be found in the
SupportingInformation (Figures S6 and S7).
Figure 4. High-resolution TEM images of the wear scar on the
siliconsubstrate. (a) TEM image showing a 7.5 nm deep wear scar
formedon silicon surface after sliding by a SiO2 microsphere under
theconditions of Fn = 3 N, RH = 60%, v = 24 m/s, and N = 100.
Insetshows the AFM image of the wear scar. (b) Representative
lattice-resolved image in worn area marked with a box (red dotted
line) in(a). The EDX spectrum in inset reveals no oxygen is
detected in thesubsurface of silicon.
Langmuir Article
dx.doi.org/10.1021/la504333j | Langmuir 2015, 31, 149156151
-
ion mass spectrometry (ToF-SIMS).37 The ToF-SIMSintensities of
SiOH+ and SiH+ groups were found signicantlyhigher in wear debris
of silicon compared to the original siliconsurface (see Figure S10
in the Supporting Information).37
These ToF-SIMS results also support the occurrence of
thetribochemical reactions at the Si/SiO2 interface during
thesliding process in humid air.Correlation between Wear Rate and
Frictional Energy
Dissipation. Even though the main wear mechanism istribochemical
rather than mechanical, the wear rate stillcorrelated with total
frictional energy dissipated during theinterfacial slide.38,39 This
also supported that the wear reactionwas activated by the
mechanical shear at the sliding interface.Some fraction of the
dissipated energy could be channeled intothe chemical reaction
coordinate to surmount the activationbarrier of the SiOSi bond
dissociation upon reaction withH2O.
24
Since the dissipated energy is the integral of the friction
forceFt over the total sliding distance, the friction force of the
Si/SiO2 interface was recorded during the wear test. Figure 5
plots
the friction force (Ft) as a function of sliding cycles
(N)measured at = 0.04, 0.8, and 20 m/s and RH = 10, 30, and65%. The
friction force of the Si/SiO2 interface varied duringthe wear test,
and its variance also depended on RH and . Ingeneral, Ft slightly
increased initially (N < 10) and thendecreased to a steady value
as the sliding continued (N > 10).The RH and dependences of Ft
during the transient periodand at the nal steady state (see Figure
S11 in SupportingInformation) indicated that the chemical and/or
physicalconditions of the interfaces dynamically vary over
time.40,41
The increase of initial friction with increasing RH could
beattributed to the growth of a solidlike structured layer ofwater
on the clean native oxide surface.42,43 The decrease ofinitial
friction with increasing velocity could be attributed to
thedecrease in the extent of water bridge bonds.43 With theincrease
in the number of reciprocating cycles, the transition tolow
friction may happen after the removal of native oxide layersof the
silicon wafer by tribochemical wear.44
Figure 6 exhibits the correlation between wear rate of Si
andfrictionally dissipated energy of the Si/SiO2 interface over
100sliding cycles under various RH and sliding speed
conditions.With the increase of sliding speed, the total dissipated
energy ofSi/SiO2 pair over 100 sliding cycles revealed a trend
similar tothe wear rate of silicon substrate at RH = 65% (Figure
6a).Figure 6b shows the linear relationship between the averagewear
rate measured at RH = 10%, 30%, and 65% and the totaldissipated
energy through friction. A similar relationship wasreported for
wear under fretting (oscillatory and reciprocatingsliding)
conditions.4547 Density functional theory (DFT)calculations
predicted that in the absence of any mechanicaldeformation or
activation, the activation energy for the SiOSi dissociation upon
reaction with H2O molecule impingingfrom the gas phase is about
113.8 kJ/mol when the siliconsurface is terminated by hydroxyl.24
The activation energyunder mechanical stress could be lower;34 but
how much loweris not known. In any case, when the imposed energy
issuciently high enough to overcome the activation barrier,
thetribochemical reaction of Si/SiO2 pair will happen. The
resultshown in Figure 6b indicated that the dissipated energy of
Si/SiO2 pair is too low to cause the tribochemical damage ofsilicon
when the wear test was performed at high speed/lowRH (20 m/s/10%)
or at low speeds/high RH (0.04 m/s and0.08 m/s/65%)
conditions.Interfacial Chemistry at Low RH Conditions. In low
humidity (RH below 30%), the average wear rate of Si
substrate
Figure 5. Friction force vs number of reciprocating cycles
(FtN)curves at representative sliding speed (v = 0.04, 0.8, and 20
m/s) andRH = (a) 10%, (b) 30%, and (c) 65%. The error of Ft
estimated fromve dependent measurements is less than 20%. The
asterisk marks thedierent value of stable friction force (N = 100)
between theconditions of RH = 65%/v = 0.04 m/s (wearless case) and
theconditions of RH = 65%/v > 0.2 m/s (wear case).
Figure 6. (a) Correlation between the average wear rate and the
total dissipated energy during 100 cycles at RH = 65%. (b) Wear
rate of the siliconsubstrate as a function of total dissipated
energy during 100 sliding cycles at RH = 10%, 30%, and 65%.
Langmuir Article
dx.doi.org/10.1021/la504333j | Langmuir 2015, 31, 149156152
-
decreased linearly as a function of ln(v) or increased
linearlywith 1/ln(1/RH) before transforming into the wearless
state(see Figure S12 in the Supporting Information). This
mightimply that the formation of capillary water bridge with
stronghydrogen-bonded network between sliding interfaces isinvolved
in the tribochemical reaction of the Si/SiO2 interfaceduring
sliding process.18,34,37 The total free energy (E)needed for
condensation of water vapor bridging two solidsurfaces could be
expressed as48
=E k T Vln(1/RH)B
m (1)
where kB is Boltzmanns constant, T is the temperature, V is
thevolume of water bridge, and m is the volume of a watermolecular.
Assuming the condensation is an activation process,the time need to
form such a water bridge can be approximatedwith the Arrhenius
relationship4850
E k Texp[ /( )]0 B (2)Here, 0 is the critical time needed to
condense one adsorbatelayer.51,52 Combining eq 1 and eq 2, the
water bridge volume Vcan be related to the sliding speed v and RH
as49,51,52
V
c v vln( / )ln(1/RH)m 0
(3)
where c is a constant and v0 is the critical sliding
speedcorresponding to a transition to the state where
interfacialfriction becomes independent of sliding speed (see
Figure S13in the Supporting Information).51 Equation 3 indicates
that thevolume of water bridge between the substrate and the
slidingcounter surface is proportional to ln(v0/v) and
1/ln(1/RH).Based on the Kelvin equation, the water bridge volume at
the
equilibrium state can be calculated at each relative
humid-ity.53,54 In this experiment, the minimum sliding speed
was0.008 m/s. If the sliding process at this speed is assumed to
bea quasi-equilibrium state, the water bridge volume V can
beestimated for the lowest sliding speed at each RH; then V canbe
estimated at all tested sliding speeds using eq 3. Figure 7aplots
the average wear rate of the silicon substrate as afunction of
water bridge volume V at RH below 30%. The dataclearly show that
the average wear rate () is proportional tothe volume of water
bridge (V). It is also intriguing to note thatthe linear
relationship also holds at high humidity conditions(Figure 7b), but
the slope is dierent, indicating the role ordynamics of water
bridge is dierent at high humidity
conditions (see Interfacial Chemistry at High RH
Conditionssection).Although further molecular details for
tribochemical
reactions within the capillary bridge are beyond the scope
ofthis experimental work, it is possible to postulate that
thewearless behavior at high sliding speeds in the low RHcondition
(
-
and then dropped to a low value within N < 10 cycles
(seegreen line in Figure 5c and Figure S11c in the
SupportingInformation). The substrate showed 0.5 nm thick wear in
therst 20 cycles and then no more wear during the next 80
cycles(see Figure S15 in the Supporting Information). The
initialwear thickness is close to typical native oxide thickness on
thesilicon wafer.44 Once this oxide layer wears o, the chemicalbond
formation across the interfaces would be less likely sincethe Si
surface does not have OH groups. Although the siliconsurface in the
sliding interfaces can be oxidized in the presenceof water and
oxygen, the oxidation rate seems to depend on and RH. If the
surface hydroxyl group formation is facilitated byvigorous
interfacial shear, it might explain the less wear at lowsliding
speeds.57 At the same shear rate, the transition to thenearly
wearless behavior is not observed at lower RH (30%)(for example,
Figure 2b), implying that the RH dependence ofthe adsorbed water
layer structure may play a critical role.42,43,56
During the nearly wearless transition process at high
RHconditions, not only the Si substrate but also the SiO2
spheresurface was also altered. This alteration was revealed in
thefollowing control experiment. First, a new SiO2 tip was rubbedon
a fresh silicon wafer surface with a displacement of 200 nmat RH =
65% and v = 0.04 m/s. After 100 sliding cycles, thedisplacement was
increased to 400 nm, and the friction forcesFt inside and outside
the original 200 nm track were compared.As shown in Figure 8a, once
the Si/SiO2 interface was
preconditioned to the low-f riction and low-wear state,
thefriction force of the preconditioned SiO2 sphere remained
lowregardless of the substrate condition whether it is
preconditioned(center 200 nm) or fresh (outside the initial 200 nm
weartrack). A typical friction force measured on the fresh
siliconsurface by a new SiO2 tip was 3.7 N at this test
condition,but the friction force measured with the preconditioned
SiO2 tipwas only 0.61 N on the fresh substrate, which was close to
thevalue (0.45 N) measured on the preconditioned wear track(Figure
8b). Once the SiO2 sphere surface was preconditioned tothe
low-friction and wearless state, then the friction forcemeasured by
reciprocating cycles on the fresh substrate surfacewas always low
regardless of sliding speed (see Figure S16 inthe Supporting
Information). This state was stable andunaltered even after long
exposure to the ambient humid air(see Figure S17 in the Supporting
Information). The only wayto recover to the initial high friction
state of the SiO2 sphere
was to keep sliding onto the fresh native oxide surface of
thesilicon substrate (see Figure S17 in the SupportingInformation).
These results indicated that the low-frictionbehavior is the
consequence of the modication of the SiO2sphere surface.The low
friction and negligible wear state induced by rubbing
could not be due to contaminations. If contaminations
wereresponsible for this behavior, the same should have
beenobserved at all RH tested. But, the transition to the
nearlywearless state at low was observed only at high RH. The
lowfriction and negligible wear at low and high RH could
beexplained if silanol groups of the SiO2 sphere undergodehydration
reactions with adjacent silanol groups during theslow shear against
Si at high RH.58 This reaction would result inthe dehydroxylation
of the SiO2 surface.
59 If the SiO2 sphereloses silanol groups, then the
SisubstrateOSisphere bridges wouldnot be formed readily.18,31 Thus,
tribochemical wear would besuppressed. The stability of this low
friction and negligible wearstate also implies that the rehydration
rate of the dehydroxy-lated silica surface is very slow.60
CONCLUSIONIn humid air, the wear of Si rubbed with SiO2 is not
ubiquitous;it is highly dependent on the ambient humidity and the
shearrate. The Si surface is mechanically robust in dry conditions;
inhumid air, however, the water-induced chemical reactions makeit
susceptible to wear. The TEM analysis conrms that there isno
mechanically induced subsurface damage in the wear trackformed in
humid conditions, supporting the hypothesis thatwear in humid
conditions is purely tribochemical. Thetribochemical reactions
appear to involve two pathways: (1)dehydration reaction between
silanol groups at the substrateand those at the counter surface,
which leads to wear, and (2)dehydration reactions between adjacent
silanol groups on onesolid surface, which leads to a low-friction
state. Thecompetition of these reaction channels determines
whetherthe Si/SiO2 interfaces would wear severely or not.
Thedominant reaction mechanism varies depending on relativehumidity
(RH) and interfacial shear rate (v). It is especiallyintriguing to
note that wearless behavior is observed at theopposite corners of
the RH and v parameter domain: low RH/high and high RH/low v.
ASSOCIATED CONTENT*S Supporting InformationCharacterization of
SiO2 microspheric tip used in the test;conrmation of
reproducibility; cross-section proles of wearscars at selected RH
and sliding speed; eect of pile-up debrison wear scar length; wear
rate at selected relative humidity andsliding speed; humidity eect
on the adhesion force of Si/SiO2pair; wearless behavior of Si/SiO2
pair in dry air condition;estimation of frictional temperature
rise; measurement ofchemical contents on the wear scars formed in
microwear tests;comparison of the initial (N = 1) and steady (N
> 20) frictionat various RH and v; wear rate at various sliding
speeds and RH(RH < 30%); critical sliding speed v0 at various
relativehumidity (RH < 30%); negligible wear of the Si substrate
athigh RH and low v conditions; stability of the low-friction
andlow-wear state of the SiO2 sphere surface formed at high RHand
low v conditions. This material is available free of chargevia the
Internet at http://pubs.acs.org.
Figure 8. (a) Friction loop (Ft vs D) curve measured
afterpreconditioning the Si/SiO2 interfaces to the low-friction and
low-wear state by rubbing at v = 0.04 m/s and under RH = 65%. After
N= 100, the sliding displacement distance was increased from 200
to400 nm. (b) Comparison of the friction force Ft measured for
freshsurface by new SiO2 tip, for inside and outside the
preconditionedwear track by modied tip. The error of Ft estimated
from threedependent measurements is less than 20%.
Langmuir Article
dx.doi.org/10.1021/la504333j | Langmuir 2015, 31, 149156154
-
AUTHOR INFORMATIONCorresponding Authors*E-mail
[email protected] (S.H.K.).*E-mail [email protected]
(L.Q.).NotesThe authors declare no competing nancial interest.
ACKNOWLEDGMENTSThe authors are grateful for the nancial support
from theNatural Science Foundation of China (91323302, 51175441,and
51375409). S.H.K. acknowledges the support from theNational Science
Foundation (Grant DMR-1207328).
REFERENCES(1) Snyder, G. J.; Lim, J. R.; Huang, C.-K.; Fleurial,
J.-P.Thermoelectric Microdevice Fabricated by a MEMS-like
Electro-chemical Process. Nat. Mater. 2003, 2 (8), 52831.(2)
Bartsch, S. T.; Lovera, A.; Grogg, D.; Ionescu, A. M.Nanomechanical
Silicon Resonators with Intrinsic Tunable Gain andSub-nW Power
Consumption. ACS Nano 2012, 6 (1), 256264.(3) Alsem, D. H.; Stach,
E. A.; Dugger, M. T.; Enachescu, M.;Ritchie, R. O. An Electron
Microscopy Study of Wear in Poly SiliconMicroelectromechanical
Systems in Ambient Air. Thin Solid Films2007, 515, 32593266.(4)
Kim, S. H.; Asay, D. B.; Dugger, M. T. Nanotribology andMEMS. Nano
Today 2007, 2, 2229.(5) Archard, J. F. Contact and Rubbing of Flat
Surfaces. J. Appl. Phys.1953, 24, 981988.(6) Kato, K. Classication
of Wear Mechanisms/Models In Wear-Materials, Mechanisms and
Practice; Stachowiak, G. W., Ed.; Wiley:New York, 2006; pp 918.(7)
Chung, K. H.; Lee, Y. H.; Kim, D. E. Characteristics of
Fractureduring the Approach Process and Wear Mechanism of a Silicon
AFMTip. Ultramicroscopy 2005, 102, 161171.(8) Bhushan, B.;
Israelachvili, J. N.; Landman, U. Nanotribology:Friction, Wear and
Lubrication at the Atomic Scale. Nature 1995, 374,607616.(9) Kato,
K.; Adachi, K. Wear Mechanisms. In Modern TribologyHandbook;
Bhushan, B., Ed.; CRC Press: Boca Raton, FL, 2001; pp273274.(10)
Kovalchenko, A.; Gogotsi, Y.; Domnich, V.; Erdemir, A.
PhaseTransformations in Silicon under Dry and Lubricated Sliding.
Tribol.Trans. 2002, 45, 372380.(11) Zarudi, I.; Cheong, W. C. D.;
Zou, J.; Zhang, L. C. AtomisticStructure of Monocrystalline Silicon
in Surface Nano-modification.Nanotechnology 2004, 15, 104107.(12)
Ribeiro, R.; Shan, Z.; Minor, A. M.; Liang, H. In SituObservation
of Nano-abrasive Wear. Wear 2007, 263, 15561559.(13) Yu, B. J.;
Dong, H. S.; Qian, L. M.; Chen, Y. F.; Yu, J. X.; Zhou,Z. R.
Friction-induced Nanofabrication on Monocrystalline
Silicon.Nanotechnology 2009, 20, 465303.(14) Gotsmann, B.; Lantz,
M. A. Atomistic Wear in a Single AsperitySliding Contact. Phys.
Rev. Lett. 2008, 101, 1125501.(15) Lantz, M. A.; Wiesmann, D.;
Gotsmann, B. DynamicSuperlubricity and the Elimination of Wear on
the Nanoscale. Nat.Nanotechnol. 2009, 4, 586591.(16) Jacobs, T. D.
B.; Carpick, R. W. Nanoscale Wear as a Stress-Assisted Chemical
Reaction. Nat. Nanotechnol. 2013, 8, 108112.(17) Liu, J. J.;
Notbohm, J. K.; Carpick, R. W.; Turner, K. T. Methodfor
Characterizing Nanoscale Wear of Atomic Force Microscope Tips.ACS
Nano 2010, 4 (7), 37633772.(18) Yu, J. X.; Kim, S. H.; Yu, B. j.;
Qian, L. M.; Zhou, Z. R. Role ofTribochemistry in Nanowear of
Single-Crystalline Silicon. ACS Appl.Mater. Interfaces 2012, 4 (3),
15851593.(19) Varenberg, M.; Etsion, I.; Halperin, G. Nanoscale
Fretting WearStudy by Scanning Probe Microscopy. Tribol. Lett.
2005, 18 (4), 493497.
(20) Ku, I. S. Y.; Reddyhoff, T.; Holmesb, A. S.; Spikes, H. A.
Wearof Silicon Surfaces in MEMS. Wear 2011, 271, 10501058.(21)
Erdemir, A.; Donnet, C. Tribology of Diamond-like CarbonFilms:
Recent Progress and Future Prospects. J. Phys. D: Appl. Phys.2006,
39 (18), 311327.(22) Chen, L.; Yang, M. C.; Song, C. F.; Yu, B. J.;
Qian, L. M. Is 2 nmDLC Coating Enough to Resist the Nanowear of
Silicon. Wear 2013,302 (12), 909917.(23) Carpick, R. W.; Salmeron,
M. Scratching the Surface:Fundamental Investigations of Tribology
with Atomic ForceMicroscopy. Chem. Rev. 1997, 97, 11631194.(24)
Barnette, A. L.; Asay, D. B.; Kim, D.; Guyer, B. D.; Lim, H.;Janik,
M. J.; Kim, S. H. Experimental and Density Functional TheoryStudy
of the Tribochemical Wear Behavior of SiO2 in Humid andAlcohol
Vapor Environments. Langmuir 2011, 27, 1270212708.(25) Torii, A.;
Sasaki, M.; Hane, K.; Okuma, S. A Method forDetermining the Spring
Constant of Cantilevers for Atomic ForceMicroscopy. Meas. Sci.
Technol. 1996, 7, 179184.(26) Varenberg, M.; Etsion, I.; Halperin,
G. An Improved WedgeCalibration Method for Lateral Force in Atomic
Force Microscopy.Rev. Sci. Instrum. 2003, 74, 33623367.(27)
Schwarz, U. D. A Generalized Analytical Model for the
ElasticDeformation of an Adhesive Contact between a Sphere and a
FlatSurface. J. Colloid Interface Sci. 2003, 261, 99106.(28)
Petersen, K. E. Silicon as a Mechanical Material. Proc. IEEE1982,
70 (5), 420457.(29) Zarudi, I.; Cheong, W. C. D.; Zou, J.; Zhang,
L. C. AtomisticStructure of Monocrystalline Silicon in Surface
Nano-modification.Nanotechnology 2004, 15, 104107.(30) Ribeiro, R.;
Shan, Z.; Minor, A. M.; Liang, H. In SituObservation of
Nano-abrasive Wear. Wear 2007, 263, 15561559.(31) Liu, Y.;
Szlufarska, I. Chemical Origins of Frictional Aging. Phys.Rev.
Lett. 2012, 109, 186102.(32) Lim, S. C.; Ashby, M. F. Overview No.
55, Wear MechanismMaps. Acta Metall. 1987, 35 (1), 124.(33) Liu,
Y.; Erdemir, A.; Meletis, E. I. A Study of the WearMechanism of
Diamond-like Carbon Films. Surf. Coat. Technol. 1996,82, 4856.(34)
Beyer, M. K. The Mechanical Strength of a Covalent BondCalculated
by Density Functional Theory. J. Chem. Phys. 2000, 112(17),
73077312.(35) Mizuhara, K.; Hsu, S. M. Tribochemical Reaction of
Oxygen andWater on Silicon Surfaces. Tribol. Ser. 1992, 21,
323328.(36) Katsuki, F. Single Asperity Tribochemical Wear of
Silicon byAtomic Force Microscopy. J. Mater. Res. 2009, 24,
173178.(37) Yu, J. X.; Chen, L.; Qian, L. M.; Song, D. L.; Cai,
Y.Investigation of Humidity-dependent Nanotribology Behaviors
ofSi(100)/SiO2 Pair Moving from Stick to Slip. Appl. Surf. Sci.
2013,265, 192200.(38) Mohrbacher, H.; Blanpain, B.; Celis, J. P.;
Roos, J. R. LowAmplitude Oscillating Sliding Wear on Chemical
Vapour DepositedDiamond Coatings. Diamond Relat. Mater. 1993, 2,
879884.(39) Yu, J. X.; Qian, L. M.; Yu, B. J.; Zhou, Z. R.
NanofrettingBehavior of Monocrystalline Silicon (100) Against SiO2
Microspherein Vacuum. Tribol. Lett. 2009, 34, 3140.(40) Xiao, X.
D.; Qian, L. M. Investigation of Humidity-dependentCapillary Force.
Langmuir 2000, 16, 81538158.(41) Nevshupa, R. A.; Scherge, M.;
Ahmed, S. I.-U. TransitionalMicrofriction Behavior of Silicon
Induced by Spontaneous WaterAdsorption. Surf. Sci. 2002, 517,
1728.(42) Asay, D. B.; Kim, S. H. Evolution of the Adsorbed Water
LayerStructure on Silicon Oxide at Room Temperature. J. Phys. Chem.
B2005, 109, 1676016763.(43) Asay, D. B.; Kim, S. H. Effects of
Adsorbed Water LayerStructure on Adhesion Force of Silicon Oxide
Nanoasperity Contact inHumid Ambient. J. Chem. Phys. 2006, 124,
174712.(44) Chen, L.; Kim, S. H.; Wang, X. D.; Qian, L. M.
Running-inProcess of Si-SiOx/SiO2 Pair at Nanoscale Sharp Drop of
Frictionand Wear Rate during Initial Cycles. Friction 2013, 1,
8191.
Langmuir Article
dx.doi.org/10.1021/la504333j | Langmuir 2015, 31, 149156155
-
(45) Aghdam, A. B.; Khonsari, M. M. Prediction of Wear
inReciprocating Dry Sliding via Dissipated Energy and
TemperatureRise. Tribol. Lett. 2013, 50, 365378.(46) Fouvry, S.;
Paulin, C.; Liskiewicz, T. Application of an EnergyWear Approach to
Quantify Fretting Contact Durability: Introductionof a Wear Energy
Capacity Concept. Tribol. Int. 2007, 40, 14281440.(47) Bryant, M.
D.; Khonsari, M. M.; Ling, F. F. On theThermodynamics of
Degradation. Proc. R. Soc. A 2008, 464, 20012014.(48) Bocquet, L.;
Charlaix, E.; Ciliberto, S.; Crassous, J. Moisture-induced Ageing
in Granular Media and the Kinetics of CapillaryCondensation. Nature
1998, 393, 2431.(49) Riedo, E.; Levy, F.; Brune, H. Kinetics of
CapillaryCondensation in Nanoscopic Sliding Friction. Phys. Rev.
Lett. 2002,88 (18), 185505(4).(50) Tambe, N. S.; Bhushan, B.
Friction Model for the VelocityDependence of Nanoscale Friction.
Nanotechnology 2005, 16, 23092324.(51) Szoszkiewicz, R.; Riedo, E.
Nucleation Time of NanoscaleWater Bridges. Phys. Rev. Lett. 2005,
95, 135502.(52) Noel, O.; Mazeran, P.-E.; Nasrallah, H. Sliding
VelocityDependence of Adhesion in a Nanometer-Sized Contact. Phys.
Rev.Lett. 2012, 108, 015503.(53) Israelachvili, J. Intermolecular
and Surface Forces, 2nd ed.;Academic Press Inc.: San Diego,
1992.(54) Orr, F. M.; Scriven, L. E.; Rivas, A. P. J. Comments on
theCapillary Binding Force of a Liquid Bridge. Fluid Mech. 1975,
67,723742.(55) Luzar, A.; Chandler, D. Hydrogen Bond Kinetics in
LiquidWater. Nature 1996, 397, 5557.(56) Vogler, E. A. Structure
and Reactivity of Water at BiomaterialSurfaces. Adv. Colloid
Interface Sci. 1998, 74, 69117.(57) Gee, M. G.; Butterfield, D. The
Combined Effect of Speed andHumidity on the Wear and Friction of
Silicon Nitride. Wear 1993,162-164, 234245.(58) Zhuravlev, L. T.
The Surface Chemistry of Amorphous Silica.Colloids Surf., A 2000,
173, 138.(59) Feng, A.; McCoy, B. J.; Munir, Z. A.; Cagliostro, D.
E. WaterAdsorption and Desorption Kinetics on Silica Insulation. J.
ColloidInterface Sci. 1996, 180, 276284.(60) Snyder, L. R.; Ward,
J. W. The Surface Structure of PorousSilicas. J. Phys. Chem. 1966,
70, 39413952.
Langmuir Article
dx.doi.org/10.1021/la504333j | Langmuir 2015, 31, 149156156