Force Microscopy Experiments: From

Nanotribology to Molecular Electronics

Ernst Meyer
Institute of Physics
University of Basel, Switzerland

•Atomic-scale stick slip

- velocity dependence
- loading dependence

•Control of friction
Ultrahigh vacuum force microscopy
 Ultra-sensitive non-
contact force microscope
combined with STM
Friction in Every-day Life

FN

FL
 Friction on the Nanometer-scale: Atomic-
Stick Slip

Atomic stick-slip Friction loop

FN = 0.44 nN

Ediss = 1.4 eV
KBr(001)-crystal
(per slip)
 Velocity dependence of atomic friction

0,55

0,50
F L (nN)

0,45

0,40

0,35
1 2 3 4 5 6 7
ln (v / 1 nm /s)

• Friction increases with the logarithm of velocity

• The slope of the curve increases with the applied load

E. Gnecco et al., Phys. Rev. Lett., 84, 1172 (2000)

 Interpretation of velocity
dependence
Tomlinson model: thermal activation:

k BT v
FL (v) = FL + ln
λ v1
 Transition to Ultralow Friction on NaCl(001)
UHV FFM with sharp tip vs. Prandtl-Tomlinson model
kx = 29 N/m, kz = 0.05 N/m, vx = 3 nm/s, const. z
Scans along [100] showing maximum variation

Stick-slip Continuous sliding in contact!

η>1 η<1 mean load FN = Fz + 0.7 nN
A. Socoliuc et al., Phys. Rev. Lett. 92, 134301 (2004)
 1d-Prandtl-Tomlinson-Model
E0 ⎛ xt ⎞ 1
Potential energy: U = − cos⎜ 2π ⎟ + k ( xt − xs ) 2
2 ⎝ a⎠ 2
∂U 2 2π 2 ⎛ xt ⎞
Stability criterion: = E cos⎜ 2π ⎟+k > 0
∂xt
0
⎝ a⎠
2 2
a

2π 2 E0 E0
η= = π 2
η>1 η<1
ka 2 ka 2
2 η=1
η=3

η < 1: unique sliding solution

η > 1: instabilities
 Instability Criterium

1d: Effective spring constant equals 2nd derivative of

adiabatic potential between tip and sample
Lateral force

x
Dependence of Tomlinson Parameters
max
aFL
FL
max
E0 =
π

− kexp 2πFL
max
η= −1
kexp a
η +1
k= kexp
η

E0: linear increase

with normal forces

k: rather independent
(contact area const.?)
 Lateral contact stiffness

1 1 1 1
= + +
k klever ktip kcontact

R. Carpick et al, Appl. Phys. Lett. 70, 1548-1550 (1997)

Here: klever=29N/m » k ⇒ kcontact ≈ 1-2N/m

Continuum model :

kcontact= 8 a G ⇒ a < 1Å ?

⇒ Atomistic model needed

Dependence of Tomlinson Parameters
max
aFL
FL
max
E0 =
π

− kexp 2πFL
max
η= −1
kexp a
η +1
k= kexp
η

E0: linear increase

with normal forces

k: rather independent
(contact area const.?)
 Lateral contact stiffness

1 1 1 1
= + +
k klever ktip kcontact

R. Carpick et al, Appl. Phys. Lett. 70, 1548-1550 (1997)

Here: klever=29N/m » k ⇒ kcontact ≈ 1-2N/m

Continuum model :

kcontact= 8 a G ⇒ a < 1Å ?

⇒ Atomistic model needed

 MD-simulations on Copper:
Shear along (111)- face

Typical lateral forces <Fx>=0.5-2nN

⇒ 5-30 atoms in contact

M.R. Soerensen et al. Phys. Rev. B 53, 2101 (1996)

 Simulations: KBr cluster tips against KBr(001)
10x10x6 slab (fixed boundaries), SciFi code (L.N. Kantorovich et al.)
Buckhingham short-range + shell model Coulomb pair potentials
U. Wyder, A. Baratoff, E. Gnecco, T. Trevethan and L. N. Kantorovich

cube or [111]
stub
single ion [100] edge

[001] [011]

s
stab pyramid
t flat bottom
a
b

Tip top layer(s) frozen; <100> scans at constant z (corrugation unaffected

by van der Waals attraction which, together with soft cantilever kz causes
jump to/from contact in experiment
 Atomistic Simulation of the
Tip-Sample Interaction

Quasi-static atomistic simulation using pair potentials

L. Kantorovich, T. Trevethan, King’s College London

U. Wyder, A. Baratoff, University Basel
 [111] K-terminated Tip

0.2 forward
backward
Lateral force [nN]

0.1

0.0

-0.1

-0.2

-1 0 1 2 3 4 5
y - displacement (units of a)
perpendicular to tip edge

0.1
Lateral force [nN]

0.0

-0.1

-0.2

forward
-0.3
backward

-1 0 1 2 3 4 5 6
x - displacement (units of a)
parallel to tip edge
⇒ Reasonable contact stiffness of
about 1-2N/m
 Commensurate [001] Tip: [100] scans

Pyramid with 2 x 2 apex in contact (Fz ≈ 0)

1.0
forward •a model for
creation and pickup
backward
0.5
Lateral force [nN]

of wear debris
0.0 • Apparent stick slip
behaviour while cluster
-0.5
passes under tip
-1.0
0.0 0.5 1.0 1.5 2.0
x - displacement (units of a)

1.0
MD (Berendsen)

0.5
filtered
Low-pass filtered traces from Molecular
Lateral force [nN]

Dynamics simulation at 300 K under the

0.0 same conditions reproduce the results
of energy minimizations at closely spaced
-0.5
x intervals + fluctuations between succesive
-1.0
stick-slip events.
0.0 0.5 1.0 1.5 2.0
x - displacement (units of a)
Complex motions: 110-Tip

„Nano-Walker“
 Heteroepitaxial growth of KBr on NaCl

Along <100> direction

200nm x 200nm 10nmx10nm

f0=153.6kHz df=-5Hz, A=11nm,Q=38’000 f0=152.6kHz,df=-28.4 Hz, Q=38’000 A=11nm

•Homogenously distributed islands, two and three

monolayer high, but none with one monolayer
->No carpet growth

•Superstructure with a periodicity of 3.95nm along the

<100>-direction is found, which fits well to the 6:7 ratio
of the lattice constants .
100nm x 100nm
f0=155.2kHz, df=-20.4Hz, Q=25’067 A=7nm •Corrugations between 0.4 and 1.3 Å are observed
 Contact mode imaging of hetero-structures:
KBr/NaCl(001) super-structure

Contact mode imaging Super-Periodicty of 6x6 KBr-units observed

of a super-structure Corrugation of about 0.1nm
(Moirée-pattern)
 Monte Carlo Simulations of KBr/NaCl(001)
J. Baker and P.A. Lindgard, Phys. Rev. B 54, R 11137 (1996)

17% misfit leads to superstrucutre

6 KBr-units fit on 7 NaCl-units

Rumpling of NaCl-interface is observed on the KBr-surface (0.01nm for 2ML)

Agreement with Helium scattering data; Duan et al. Surf. Sci. 272, 220 (1992)
 Atomic friction of KBr/NaCl(001)

Super-structure is observed in atomic friction

Atomic-scale defects are observed!
 Loading dependence of atomic friction:
KBr/NaCl(001) superstructure

friction force (nN)

0.1

0.0

-0.1

0 2 4 6 8 10
FN=0.01 nN
distance (nm )

10x10nm2: friction force (nN)

0.1

Friction force map at 2 loads

0.0

-0.1

0 2 4 6 8 10
distance (nm) FN=-0.31 nN
Superstructure is less pronounced at lower loads Close to jump-off
Variations of the slope of the sticking phase within the unit
cell: Effect of η(FN)/E0(FN) or contact stiffness k?
0.2
0.49 N/m
0.47 N/m
0.34 N/m
friction force (nN)

0.37 N/m
0.41 N/m 200 μ=0.36 N/m
0.1 σ=0.1 N/m

number of points
0.33 N/m 150

100
0.0
50

0
-0.1
0.0 0.4 0.8 1.2 1.6
stiffness (N/m)
0 1 2 3 4 5
distance (nm)

η 2π 2 E0 E0
kexp = k η= = π 2

η +1 ka 2 ka 2
2

S. Maier et al., unpublished results

 Can we switch friction on and off?

AC voltages were applied across thin KBr and NaCl crystals:

• Capacitive interaction between lever and sample holder ∝ UB2

• Coulomb interaction ∝ UB
 Controlling Friction: Actuation of Nanometer-
Sized Contacts

A. Socoliuc, E. Gnecco, S. Maier, O. Pfeiffer, A. Baratoff, R. Bennewitz, E. Meyer, Science 313, 207 (2006)
 Frequency dependence of friction

Friction:

Thermal
noise:

• Friction is “switched off” only if fexc = fnorm or (1/2) fnorm

• No effect when fexc = ftors !
Frequency dependence of friction

Reduction of friction at and resonance

Frequencies of the the bending modes f and f/2
 Voltage dependence of friction

• Friction goes down to zero increasing the excitation amplitude !

 Interpretation of Dynamic Superlubricity
• In the Tomlinson model: We replace E0 with E0 (1+α cos ωt)
• The parameter α increases with the applied voltage

η=7
E0

η=5

η=3

η=1

Friction decreases
when α → 1
 New parameter ηmin
• The parameter ηmin = η (1-α) determines superlubricity

“Static” case
 „Phase“ diagram of friction
• A "phase diagram" in the η-α plane can be drawn:

1
α cr = 1 −
η

"Static" SL

A. Socoliuc et al., Science 2006

 Modulation of the Energy barrier
by actuation of the nano-contact

Standard parameters for the Tomlinson model with excitation:

0.5 times critical damped
Parameters: eta=4, alpha=0.9, f=567Hz, v=10e-9m/s, gamma=1e-6kg/s,
m=8e-13kg, a=0.5e-9m, c=1N/m
 Ultralow friction on the macroscopic scale?

• Normal force per asperity is limited to 1nN

• Macroscopic weight of 1g ≈ 10mN has to be distributed
to 107 mini-tips (≈ array of 3’000 x 3’000 tips)

Tips with a
spacing of 3μm

10mm
IBM Zurich
1000 AFM’s
 MEMS-Devices

Courtesy of Sandia National Laboratories, SUMMiTTM Technologies,

www.mems.sandia.gov"
 Gecko uses nanometer-sized contacts
to climb walls

Gecko is able to control the

contact area on all length scales

From B. Persson and S. Gorb

JCP, 119, 11437 (2003)
 Other forms of "superlubricity "
• Thermolubricity (Krylov et al., PRE 2005):

(taking "backjumps" into account → friction vanishes at low speed)

 Other forms of "superlubricity"
• Structural lubricity (Dienwiebel et al., PRL 2004):

(two mismatched graphite flakes sliding past each other)

 Nanomotor with Multiwalled Carbon Nanotubes:
Low friction and negligible wear

A. Zettl, University of California Berkeley

Sliding of rolled graphene!

A. Fennimoore et al., Nature 424, 408 (2003).
 Transitions to negligible friction in differnt dry contacts
Graphite against rotated flake picked up by tip NaCl(001) cleaved in UHV

0.9
stick-slip
FN
0.6

0.3

FL (nN)
0.0

-0.3

-0.6
"superlubricity"
-0.9
0.0 0.5 1.0 1.5 2.0 2.5 3.0
x (nm)

M. Dienwiebel et al. PRL 92, 126101 (2004) A. Socoliuc et al. PRL 92, 134301 (2004)

Effective corrugation on graphite reduced at low velocities at fixed room T

Thermal activation:
small velocity ~ high temperature

Like creep of dislocations

or vortices in type II superconductors

S.Yu Krylov et al. PRE 71, 065101(R) (2005)

 Conclusions

• Observation of ultralow friction without atomic-stick slip

• Continuum elasticity theory is not applicable

Modelling needing

•Superlubricity is observed for low loads, low velocity or

incommensurate structures

•Wear can be avoided at low loads

•Control of friction: Switch off/on friction