Excimer laser annealing of sputtered three-layer stack (1)

Hello everyone! I plan to write several posts in the next couple of weeks about the laser annealing of our sputtered three-layer stack. This laser annealing experiment is done at Center for Nanophase Materials Sciences (CNMS) at Oak Ridge National Laboratory (ORNL). It is a collaborational  proposal we submitted at the end of last year and approved early this year. The CNMS offers the free use of their facilities which were written in the proposal. To begin with, I would like to welcome more comments on my posts since it’s not easy for me to attend our the GREATEST NRG meeting after I moved to Vanderbilt. I’ll start with some basic information about samples, laser, setup and so on.

Sample info: NSN stack with 50nm a-Si layer and 30nm Si3N4 protective layers and the Si substrate was removed

Laser info: pulsed KrF excimer laser, 248nm wavelength, 25ns FWHM, spot size: ~2.2 mm diameter

Experiment setup: Ar was flowing through the chamber during laser annealing.

We have tried four different energy densities to do the laser annealing experiments. They are ~160mJ/cm2, ~110mJ/cm2, 80mJ/cm2 and 60mJ/cm2. The first set of images are samples annealed at around 160mJ/cm2. More than 50% of free-standing membranes broke during the laser annealing process at this energy density.

The top four are SEM images on the free-standing membrane area.  First of all, it seems that the free-standing area shows a different morphology than the film on the substrate. This may due to the effect from the Si substrate as a sink to suck the heat from the top films through heat transfer where the free-standing stack absorbs all the energy from the laser. Second of all, the white dots feather looks like particles but it is not very clear at this point. Keep in mind that I didn’t strip nitride at this point and I believe they still sandwich the Si film. Another interesting feather would be the tiny dark dots feather, which looks like nanopores with 10 to 2onm in diameter. This also has not confirmed yet. A staff at CNMS helped me did a quick TEM on my samples and the followings are the TEM images of this sample.

From the TEM images, it first proves that those white area in the SEM image (black area in the TEM image) are not particles. The interesting part is that there are some feathers like void inside those black area in the TEM image. Another useful information from the diffraction image shows that Si film is partly crystallized. The majority part is still amorphous.

Next sample is annealed with energy density of 110mJ/cm2 and it shows very different results.

The top four SEM images are at different magnitude from low to high. This sample shows very different but interesting morphology that we have never seen in our RTP annealed samples. The SEM shows that Si film becomes to porous net structure after laser annealing! This porous net structure seems quite irregularly in terms of pores with different shapes and sizes. It also looks like unfinished Si agglomeration that Si film agglomerated to form tortured Si rods but still connecting to each other. Si grains and grain boundary can be clearly seen at a higher magnification. The size of Si grans ranges from several tens nanometers up to several hundreds nanometers. If you look at the images at lower magnitude carefully, you could see some quite bright spots. I believe these are bumped Si films (nodes ?) which have a higher height. Due to these bumped Si, the top nitride layer is pop out after annealing, which can be seen from some broken membranes at shown in the following.

These are two interesting SEM images of the porous Si net structure with broken curved top nitride layer. The top nitride film shows some bright circular features which I think are some bumped features corresponding to the underneath bumped Si. In the second SEM image, the bended broken Si film shows a little cross-sectional view of the surface of the porous net structure Si, which seems pretty rough due to the bumped feature. Now let’s think this unique structure in a big picture, here is what happens in my thought. Si film tends to agglomerate during the laser annealing process at this energy density. However, the fully agglomeration is suppressed due to the existence of the cap/top nitride layer. Si film agglomerates to tortured porous net  structure with different height in the thickness direction. These higher parts in the silicon film become to some bumpers with different sizes and the cap/top nitride layer deforms due to the compressive stress from these bumpers, which maybe the reason for the formation of the bulged structures. However, the compressive stress is quite big which can not offset by the plastic deformation of the silicon nitride and it results in the popping out the nitride film finally. This is a very preliminary speculation which needs to be further verified. The follows are some TEM images of this sample.

From the TEM images we can directly see the crystallized Si in tortured rod structure as well as the surrounded pores. The crystalline fraction seems pretty high in this sample.

Samples which annealed at the rest two energy density seems not crystallize yet, which can be proved from the diffraction pattern of a sample annealed with 80mJ/cm2 in the following TEM images.

The bright view TEM image shows featureless amorphous phase and the diffraction pattern proves the amorphous nature of the silicon film. This means that the energy density of 80mJ/cm2 is not enough to crystallize amorphous silicon.

In sum, Si film shows very different morphologies with excimer laser annealing with different energy densities. Since this is our first attempt to crystallize Si film using excimer laser, further in-depth characterization of these interesting structures is needed to fully understand this different crystallization process. Any questions and comments are very welcome to bring in and would be very help for the further research.