Power law load dependence of atomic friction
Extracted source text ยท arXiv cond-mat/0404044
\documentstyle[aps,epsfig,multicol]{revtex}
\draft
\input epsf
\begin{document}
\title{\bf Power law load dependence of atomic friction}
\author{C. Fusco\thanks{Author to whom correspondence should be
addressed. Electronic address: fusco@sci.kun.nl.} and A. Fasolino}
\address{Theoretical Physics, NSRIM, University of
Nijmegen,\\
Toernooiveld 1, 6525 ED Nijmegen, The Netherlands
}
\date{\today}
\maketitle
\begin{abstract}
We present a theoretical study of the dynamics of a tip scanning a graphite
surface as a function of the applied load.
From the analysis of the lateral forces, we extract the
friction force and the corrugation of the effective tip-surface interaction
potential.
We find both the friction force and potential amplitude to have a power law
dependence on applied load with exponent $\sim 1.6$. We interpret these
results as characteristic of sharp undeformable tips in contrast to the case
of macroscopic and elastic microscopic contacts.
\end{abstract}
\begin{multicols}{2}
It is well known that macroscopic
friction is proportional to the applied load but the load dependence of
atomistic friction is still under investigation and has been discussed in a
limited number of experimental and theoretical
works~\cite{fujisawa1,fujisawa2,ishikawa,sasaki1,zhong,sorensen,schwarz,enachescu}.
Usually the load dependence is described by means of contact mechanics
continuum theories~\cite{schwarz,enachescu} which do not give information
about the atomic interactions in the sliding contact region and assume a
spherical tip. Moreover, it has been argued that, depending on
the shape of the tip, power laws with different exponents can be
found~\cite{schwarz}.
In this Letter, we present a detailed study of the load dependence of
atomic-scale friction in the case of a sharp tip-surface contact, finding a
power law dependence with exponent larger than one. We also show how
the effective tip-surface potential energy barriers can be derived from force
measurements, thus establishing a connection between friction and the
interatomic potential.
Recent works using noncontact mode Atomic Force Microscopy (AFM) have shown
the possibility to reconstruct the tip-surface potential~\cite{holscher1},
but they lack the link between the corrugation of the potential and the
friction force. The ideal way to achieve this goal is provided by a study of
the load dependence.
We address the load dependence of friction by
simulating the dynamics of a tip scanning a rigid monolayer graphite surface,
extending to three dimensions ($3D$)
the Tomlinson-like models of AFM~\cite{tomanek}.
Experimental evidences show that the tip usually cleaves small graphite flakes
which remain attached to it~\cite{dienwiebel}.
Since the contact diameter between the tip and the surface can be very small
we consider the limiting case in which the tip is formed by a single carbon
atom connected via harmonic interactions with force constants $K_x$,
$K_y$ and $K_z$ to a support moving along the scanning direction.
The carbon atom of coordinates $(x,y,z)$ interacts
with the graphite surface via the specific
empirical potential $V_{TS}$ for graphite given by~\cite{los}:
\begin{equation}
V_{TS}=\sum_{j}\left[\theta(r_0-r_{Tj})V_{1}(r_{Tj})+\theta(r_{Tj}-r_0)V_{2}(r_{Tj})\right]
\end{equation}
where $\theta(r)$ is the Heaviside function, $r_{Tj}$ is the distance between
the tip carbon atom and the $j-$th substrate atom and
\begin{equation}
V_i(r)=\epsilon_i\left(e^{-2\beta_i(r-r_0)}-2e^{-\beta_i(r-r_0)}\right)+v_i
\qquad i=1,2
\end{equation}
with $v_1=\epsilon_1-\epsilon_2$ and $v_2=0$.
The values of the parameters are $r_0=0.371$ nm, $\epsilon_1=5.355$ meV,
$\epsilon_2=2.614$ meV, $\beta_1=14.693$ nm$^{-1}$ and
$\beta_2=21.029$ nm$^{-1}$.
The tip support is moved with constant velocity $v_{scan}$ along the scanning
line $x_{scan}=v_{scan}t$, $y_{scan}=$constant.
We solve numerically the equations of motion in the constant force mode, i.e.
we set $K_z=0$ and add a constant force $F_{load}$ in the downward $z$
direction, including also a damping term proportional to the atom velocity,
which takes into account dissipation mechanisms:
\begin{equation}
\begin{array}{ccc}
m\ddot{x}=-\frac{\partial V_{TS}}{\partial x}+K_x(x_{scan}-x)-m\eta\dot{x}\\
m\ddot{y}=-\frac{\partial V_{TS}}{\partial y}+K_y(y_{scan}-y)-m\eta\dot{y}\\
m\ddot{z}=-\frac{\partial V_{TS}}{\partial z}-F_{load}-m\eta\dot{z}
\end{array}
\end{equation}
We adopt an atomistic approach, assuming for $m$ the mass of a single carbon
atom ($m=1.92\cdot 10^{-26}$ kg) and for the damping parameter
$\eta=1$ ps$^{-1}$,
which is a value appropriate for dissipation of energy and momentum
at the atomic scale (see e.g.~\cite{ps}).
Here we show results for $K_x=K_y=1$ N/m which are typical values of AFM,
whereas our scanning velocity $v_{scan}=0.4$ m/s is much higher than in
experiments. Our choice of parameters makes the dynamics
underdamped.
\begin{figure}
\epsfig{file=Fig1.eps,width=8cm,height=6cm}
\caption{Tip-surface potential $V_{TS}$ as a
function of the tip-surface distance $z$ with $x$ and $y$ coordinates
taken as those of the hollow site (solid line).
Points give the values of $V_{TS}$ at several positions along the actual
trajectory as a function of the corresponding instantaneous value of $z$.
Results for all indicated loads are obtained for simulations where the
scanning direction is along a row of atoms ($y_{scan}=0.17$ nm).}
\label{f.potentialzV}
\end{figure}
The energetics underlying the tip motion for different loads is illustrated
in Figs.~\ref{f.potentialzV} and~\ref{f.contxy}. In Fig.~\ref{f.potentialzV}
we present the tip-surface potential $V_{TS}$ as a function of the vertical
tip-surface distance $z$, for in-plane coordinates corresponding to the
position of the hollow site, on which we superimpose the values of $V_{TS}$
at actual positions of the tip during the simulations for different values
of the load.
The tip position probes the attractive part of the potential
for low loads ($F_{load}=0.1$ nN) and moves closer and closer to the
repulsive core of the substrate atoms for increasing loads.
Clearly the whole tip-surface potential is
probed by the motion of the tip for different loads, as suggested in
Ref.~\cite{fujisawa2}.
In Fig.~\ref{f.contxy} we show the contour plots of $V_{TS}$ , calculated for
$z$ given by the average value of the tip height at two values of the load.
The scanning line $y_{scan}=0.17$ nm corresponds to scanning
along a row of atoms. The actual trajectory is also reported: for the smaller
load the motion follows a zig-zag-like pattern, while the trajectory
acquires a one-dimensional stick-slip character for the higher load
(see also~\cite{fujisawa2}).
\begin{figure}
\epsfig{file=Fig2.eps,width=8cm,height=8cm}
\caption{Contour plot of the potential $V_{TS}$ for $F_{load}=1$ nN (top) and
$F_{load}=4$ nN (bottom) with $z$ fixed at the average value $<z>$ of the
height obtained by the simulations ($<z>=0.25$ nm for $F_{load}=1$ nN and
$<z>=0.2$ nm for $F_{load}=4$ nN). The minimum $V_{min}$ and the maximum
$V_{max}$ of $V_{TS}$ are located respectively at the hollow site and on top
of one atom.
We show 10 contour lines between $V_{min}$ and $V_{max}$, separated by an
energy interval $\Delta$. $V_{min}=123.5$ meV, $V_{max}=158.5$ meV and
$\Delta=3.2$ meV for $F_{load}=1$ nN, while $V_{min}=774$ meV,
$V_{max}=1.01$ eV and $\Delta=23.7$ meV for $F_{load}=4$ nN. The horizontal
thick solid line indicates the scanning direction ($y_{scan}=0.17$ nm), while
the crosses are points along the actual atomic trajectory taken every $4$ ps.}
\label{f.contxy}
\end{figure}
Fig.~\ref{f.T0eta1Kxy1v0.0004y0.17F2-4m0.0192} presents
typical force plots as a function of $x_{scan}$ and $x$~\cite{note} for
$F_{load}=2$ nN and $F_{load}=4$ nN. The sawtooth behavior
characteristic of stick-slip motion is determined by the competition between
the elastic force ${\bf F}_{el}$ and the force ${\bf F}_{TS}$ due to the
tip-surface potential $V_{TS}$. Elastic energy, accumulated
in the spring, is counterbalanced by the substrate attraction,
until, suddenly, the tip slips to another minimum.
Therefore, while sticking, ${\bf F}_{el}$ mirrors ${\bf F}_{TS}$.
This fact can be used to derive $V_{TS}$ itself.
The solid lines in
Fig.~\ref{f.T0eta1Kxy1v0.0004y0.17F2-4m0.0192} represent
the lateral force along the scanning direction, $F_{el}^x$, as obtained
by our simulations.
Increasing the load enhances the sawtooth behavior of the stick-slip motion
and gives rise to a larger initial sticking region, often observed
experimentally (see e.g.~\cite{morita}).
As shown in Fig.~\ref{f.contxy} the actual trajectory does not necessarily
follow the scanning line so that the potential energy landscape during the
motion is not known a priori. However, we can extract the effective value of
the energetic barrier $V_0$ for a given $F_{load}$ directly from the force
plots. By assuming a sinusoidal shape of $F_{TS}$~\cite{zhong,ke} and noting
that $F_{TS}$ should average to zero for a periodic substrate,
we can reconstruct $F_{TS}^x$.
\begin{figure}
\epsfig{file=Fig3.eps,width=9cm,height=7cm}
\caption{Lateral forces along the scanning direction for $F_{load}=2$ nN
[(a),(c)] and $F_{load}=4$ nN [(b),(d)], plotted as a function of $x_{scan}$
[(a),(b)] and $x$ [(c),(d)], for $y_{scan}=0.17$ nm.
Solid lines are the elastic forces $F_{el}^x$ obtained
by simulations, while the dashed lines in (c) and (d) are static
calculations of the tip-surface force (with reverted sign) $-F_{TS}^x$
at $(y,z)$ determined by averaging $y(t)$ and $z(t)$
given by the dynamics. The dotted lines in (a), (b) give the
slope $K_{eff}$ of the
sticking part ($K_{eff}=0.78$ N/m for $F_{load}=2$ nN and $K_{eff}=0.89$ N/m
for $F_{load}=4$ nN).}
\label{f.T0eta1Kxy1v0.0004y0.17F2-4m0.0192}
\end{figure}
Then, the tip-surface potential $V_{TS}$ is simply given,
up to a constant, by integrating $F_{TS}^x$:
\begin{equation}
\label{e.potential}
V_{TS}(x)=-\int F_{TS}^xdx=\frac{V_0}{2}\sin\left(\frac{2\pi x}{a}\right)+const.
\end{equation}
with $V_0=F_0a/(2\pi)$, where $a$ and $F_0$ are the period and the amplitude
of $F_{TS}$.
In Figs.~\ref{f.T0eta1Kxy1v0.0004y0.17F2-4m0.0192}(c)-(d) we show
$-F_{TS}^x$ (dashed lines) obtained by static calculations for
$(y,z)$ determined by averaging $y(t)$ and $z(t)$ given by the dynamics.
Indeed, $-F_{TS}^x$ follows the sticking parts of $F_{el}^x$ quite well.
The first stick signal, of larger amplitude, is the most suitable
to estimate the amplitude of $F_{TS}^x$.
\begin{figure}
\epsfig{file=Fig4.eps,width=12cm,height=8cm}
\caption{Load dependence of the energy barrier (a) and of the friction
force (b) for $y_{scan}=0.17$ nm. Inset of (b): relation between friction
force and energy barrier. The solid circles are the result of the simulations
and the dashed lines are power law fits, as explained in the text.}
\label{f.loadbarr}
\end{figure}
The resulting dependence of the energy barrier $V_{0}$ on the load is shown
in Fig.~\ref{f.loadbarr}(a). An increase of $V_0$ with $F_{load}$
has also been found experimentally~\cite{fujisawa2,riedo}.
The dashed line in Fig.~\ref{f.loadbarr}(a) is a power law fit to the
numerical data with exponent $\sim 1.6$.
Fig.~\ref{f.loadbarr}(b) illustrates the lateral friction force $F_{fric}$,
obtained as the mean value of the instantaneous lateral force $F_{el}^x$,
as a function of $F_{load}$. Also these data can be fitted by a
power law with exponent $\sim 1.6$.
Thus, the linear relation between macroscopic friction and
load does not hold at the microscopic level. Moreover, the exponent is
different from the $2/3$ expected for a Hertzian
contact~\cite{schwarz,enachescu}. As pointed out in
Ref.~\cite{schwarz} even a small deviation from the spherical shape of the
tip can be responsible for a change in the power law exponent.
We conjecture that exponents larger than one are characteristic of sharp,
undeformable tip-surface contacts.
Note that, since the power law
exponents for $V_0$ and $F_{fric}$ as a function of $F_{load}$ are very close,
an approximately linear relation between $F_{fric}$ and $V_0$ is
recovered (see inset of Fig.~\ref{f.loadbarr}(b)),
indicating a direct link between atomistic friction and energy barriers for
diffusion.
In conclusion, we have presented a theoretical study of the $3D$ dynamics of
a tip scanning a graphite surface with realistic
interactions as a function of the applied load.
We predict a power law behavior with exponent $\sim 1.6$ for the friction
force as a function of applied load, at variance with macroscopic behavior.
The study of the load dependence establishes a
direct linear relation between friction and potential corrugation in the
contact region.
This work was supported by the Stichting Fundamenteel Onderzoek der Materie
(FOM) with financial support from the Nederlandse Organisatie voor
Wetenschappelijk Onderzoek (NWO). The authors wish to thank Elisa Riedo,
Jan Los, Sylvia Speller, Sergey Krylov and Jan Gerritsen for useful and
stimulating discussions.
\begin{thebibliography}{99}
\bibitem{fujisawa1}
S. Fujisawa, E. Kishi, Y. Sugawara and S. Morita, Phys. Rev. B {\bf 52},
5302 (1995).
\bibitem{fujisawa2}
S. Fujisawa, K. Yokoyama, Y. Sugawara and S. Morita, Phys. Rev. B {\bf 58},
4909 (1998).
\bibitem{ishikawa}
M. Ishikawa, S. Okita, N. Minami and K. Miura, Surf. Sci. {\bf 445}, 488
(2000).
\bibitem{sasaki1}
N. Sasaki, M. Tsukada, S. Fujisawa, Y. Sugawara, S. Morita and K. Kobayashi,
Phys. Rev. B {\bf 57}, 3785 (1998).
\bibitem{zhong}
W. Zhong and D. Tom\'anek, Phys. Rev. Lett. {\bf 64}, 3054 (1990).
\bibitem{sorensen}
M. R. S\o rensen, K. W. Jacobsen and P. Stoltze, Phys. Rev. B {\bf 53}, 2101
(1996).
\bibitem{schwarz}
U. D. Schwarz, O. Zw\"orner, P. K\"oster and R. Wiesendanger, Phys. Rev. B
{\bf 56}, 6987 (1997).
\bibitem{enachescu}
M. Enachescu, R. J. A. van den Oetelaar, R.W. Carpick,
D. F. Ogletree, C. F. J. Flipse and M. Salmeron, Phys. Rev. Lett. {\bf 81},
1877 (1998).
\bibitem{holscher1}
H. H\"olscher, W. Allers, U. D. Schwarz, A. Schwarz and R. Wiesendanger,
Phys. Rev. Lett. {\bf 83}, 4780 (1999); Phys. Rev. B {\bf 62}, 6967 (2000);
H. H\"olscher, A. Schwarz, W. Allers, U. D. Schwarz and R. Wiesendanger,
ibid. {\bf 61}, 12678 (2000); U. D. Schwarz, H. H\"olscher and
R. Wiesendanger, ibid. {\bf 62}, 13089 (2000).
\bibitem{tomanek}
D. Tom\'anek, W. Zhong and H. Thomas, Europhys. Lett. {\bf 15}, 887 (1991);
N. Sasaki, K. Kobayashi and M. Tsukada, Phys. Rev. B {\bf 54}, 2138 (1996);
H. H\"olscher, U. D. Schwarz and R. Wiesendanger, Surf. Sci. {\bf 375}, 395
(1997);
Y. Sang, M. Dub\'e and M. Grant, Phys. Rev. Lett. {\bf 87}, 174301 (2001).
\bibitem{dienwiebel}
M. Dienwiebel, Ph.D. Thesis, University of Leiden, 2003.
\bibitem{los}
J. H. Los and A. Fasolino, Comp. Phys. Comm. {\bf 147}, 178 (2002);
Phys. Rev. B {\bf 68}, 024107 (2003).
\bibitem{ps} R. Guantes, J. L. Vega, S. Miret-Art\'es and E. Pollak,
J. Chem. Phys. {\bf 119}, 2780 (2003); H. Xu and I. Harrison,
J. Phys. Chem. B {\bf 103}, 11233 (1999).
\bibitem{note}
Experimental force plots data are often given as a function of $x_{scan}$.
However, since $F_{el}^x=K_x(x_{scan}-x)$, it is straightforward
to rewrite the data $F_{el}^x(x_{scan})$ as $F_{el}^x(x)$.
\bibitem{morita}
S. Morita, S. Fujisawa and Y. Sugawara, Surf. Sci. Rep. {\bf 23}, 1 (1996).
\bibitem{ke} S. H. Ke, T. Uda, R. P\'erez, I. \u Stich and K. Terakura,
Phys. Rev. B {\bf 60}, 11631 (1999).
\bibitem{riedo}
E. Riedo, E. Gnecco, R. Bennewitz, E. Meyer and H. Brune, Phys. Rev. Lett.
{\bf 91}, 084502 (2003).
\end{thebibliography}
\end{multicols}
\end{document}