Self-ordering of the Lubricant Film


Introduction goto top

As was shown above, the use of solid lubricants may be a very promising way, especially in micro-devices, where conventional lubricants may often be ineffective or may even work like a glue. In an ideal case of contact of two rigid crystalline incommensurate surfaces the static frictional force is zero (moreover, the same is true even for commensurate surfaces if they are not perfectly aligned). Analogously, if the lubricant film has a crystalline structure and is confined between two flat substrates with the ideal structure, the friction coefficient in such perfect-sliding system may be as low as  μ ~ 10–3 to 10–2  or even lower. The critical velocity of the transition from stick-slip to smooth sliding is also quite small,  vc~10–2c. Unfortunately, such an ideal system can hardly be reproduced experimentally: even specially prepared surfaces are not perfectly smooth on a mesoscopic scale, and a lubricant has typically numerous structural defects. As a result, the static frictional force  fs  is large enough, and the solid lubricant will melt at the onset of sliding. Then, at stick, the film solidifies back, but, again, either in an amorphous state or in a state with many defects, because the cooling of the confined film is very rapid due to very good thermal contact with the substrates. According to simulation, in such a system one finds μ > 0.1 and vc~ 0.1c, i.e., the tribological characteristics are even more worse than those of liquid lubricants. In what follows we will explore whether the system itself may approach the desired perfect-sliding regime for a suitable choice of the solid lubricant. Indeed, the energy pumped into the system due to external driving has to be converted into heat. Therefore, the effective lubricant temperature T* must increase during sliding. It is this increase of the temperature that leads to melting of the lubricant in the melting-freezing mechanism of stick-slip. However, if T* will remain lower than the melting temperature Tm, the lubricant film could remain solid and, at the same time, its structure may become more ordered due to annealing of structural defects, especially if T* will be close to Tm.  Because Tm is determined by the interactions within the lubricant, one may play with the parameters of interaction looking for a situation when Tm  is close to (but higher than) T*. Below we show that, indeed, a suitable choice of the parameters may lead to the desired self-ordering of the lubricant and consequently to the low values of the frictional forces.

 

Some examples goto top

Hard lubricant, Nl=1 goto top

Sliding-induced ordering of the  Nl=1  system for  vspring= 0.1

a03.gif  

 

Configuration before

self-ordering

last_t02-1.gif

 

 

Configuration after

self-ordering

last_t03-1.gif
 

 

 

Hard lubricant, Nl=3 goto top

Sliding-induced ordering of the  Nl=3  system for  vspring= 0.3  &  movie (avi 1.5 Mb)

a57.gif

 

Curved substrate goto top

Sliding-induced ordering of the  Nl=3  system with  Vll= 0.5 (so that  Vll/Vsl= 1.5) and the curved top substrate for  vspring= 0.1

fig03.gif

Configurations:
before self-ordering (commensurate) after self-ordering (incommensurate)
before403.jpg after995.jpg
before403_bottom.jpg after995_bottom.jpg
movie (mpeg 7.2 Mb) movie (mpeg 9 Mb)
full movie (mpeg 15 Mb)

 

Dependence on Vll (interaction in the lubricant) goto top

fig04.gif

Figure: Static  fs  and kinetic  fk  frictional forces for three values of the driving velocity (vs= 0.1, 0.3 and 1 as shown in legend) as functions of the interaction amplitude  Vll   in semi-logarithmic scale. The error bars show deviation of the simulation results in different runs.

 

From this figure we can observe two clear-cut features: First, as one can judge from the dependence of  fk on  vs, the mechanism of the stick-slip motion changes from the melting-freezing to the inertia mechanism at  Vll0.5, i.e., for  Vll/Vsl1.5. Most importantly, we observe that for  Vll≈ 0.8 the kinetic frictional force  fk  achieves a minimum as low as  fk ≈ 10–4 to 10–3.  The friction coefficient in this case takes values of order  μ < 10–2  which are more than one order of magnitude lower  than those attainable with conventional liquid lubricants. Namely, for the optimum value of the interaction, Vll= 0.75, at  v = 0.03  we get smooth sliding with  fkinetic≈ 10–4, so that  µ ≈10–3  (see movie mpeg 3.3 Mb).

 

Conclusion goto top
It exists an optimal choice of the strength of interatomic interaction Vll  within the lubricant, that leads to the minimization of the kinetic friction as well as to the low critical velocity of the stick-slip to smooth-sliding transition. The optimal value of Vll  should be high enough (relatively to the amplitude Vsl  of the interaction of lubricant atoms with the substrates) so that the lubricant remains in the solid state during sliding. At the same time, the value of Vll  should be not too high, in order to allow annealing of the structural defects in the lubricant. For the parameters used in the simulations, the optimum was achieved at Vll ≈ 2.5Vsl.

 

Next: Role of lubricant molecule shape goto top

goto main Back to main page or tribology page


Last updated on April 1, 2004 by O.Braun.          Translated from LATEX by TTH