Thank you very much for giving me the chance to present some of our recent results on phase

transitions in quasi one-dimensional systems.

I want to discuss two examples we have been interested in.

So optical excitation of indium nanowires on silicon 1,1,1 that leads to insulator metal

transition and how thermal gold vibrations can lead to the formation of a spin liquid

at a silicon 5,5,3 surface.

Now let's start with the indium system.

The indium induced 4 by 1 surface reconstruction on silicon 1,1,1 has been known for a fairly

long time, actually more than 50 years.

If you look at it as an SDM image, you see indium rows and microscopically they consist

at a double row of six-sac chains, as it is shown here with the red atoms at room temperature.

If you cool the system below 120 Kelvin, it transforms into indium hexagon structures

as it is shown here, it becomes insulating.

Now can that system be triggered optically?

That question has been answered by Michael Horn von Högen in Duisburg.

He made an experiment at 20 Kelvin.

He excited the low temperature 8 by 2 structure with laser pulse and he showed that he can

transform the 8 by 2 into the 4 by 1 unit cell with the laser pulse without actually

heating the system.

So it was an optically induced phase transition in contrast to a thermal phase transition.

Now he came to us and asked what is happening here and what we did was at first to calculate

the band structure of the system.

This is a band structure of the 8 by 2 surface and then we took out electrons from the occupied

valence state and put them into formerly empty conduction states.

Then we looked at how the energy of the system changes.

We calculated potential energy surfaces as it is shown here on the right hand side.

The black line corresponds to the ground state configuration.

We have a minimum for the 8 by 2 and we have a local minimum for the 4 by 1.

But if for example you take out electrons here at the zone boundary and put them into

empty zone center states you obtain a potential energy surface like it is shown here, the

red one, which shows a steep gradient from the 8 by 2 to the 4 by 1.

Now is that indeed sufficient for a phase transition?

To answer that we performed ab initio molecular dynamics on these excited state potential

energy surfaces and we find that in dependence on how many electrons you excite and where

you excite the electrons you can indeed observe phase transitions that are fairly rapidly

on a picosecond time scale.

We find that there are at least 0.4 electrons that need to be excited before a phase transition

can occur.

We also find that exciting more than one electron does not really help to make it faster.

If you look at the potential energy surface for two electron excitation it is nearly parallel

to a one electron excitation so the phase transition happens at the same time.

And we find phase transition times between 350 and 1700 femtoseconds.

Now does it agree with experiment?

The Duisburg group performed time-resolved electron diffraction experiments.

So we first excited the system with the pump laser pulse and after some delay time we used

an electron pulse to probe the surface structure and by varying the delay time we could actually

produce a movie of how the system evolves in time.

And we confirmed very nicely our theoretical predictions.

We also find a lower and an upper threshold for the optical excitation of the system and

we find that indeed phase transition times as short as 450 femtoseconds are possible.