# Reason behind heating the substrate in Pulsed Laser Deposition

Why do we need to supply a constant heat to the substrate while depositing thin films in Pulsed Laser Deposition technique?

I have a lot of experience with CVD and sputtering, but limited experience in PLD; however, several of my colleagues did this all the time in our shared a laser lab. When they were attempting to reproduce a specific result all of the parameters had to be systematically varied, from the laser fluence to the substrate conditioning and temperature, and more.

The substrate temperature has an impact on the rate at which the surface growth occurs, and the form of the growth (growth mode). If the temperature is too low, you may get a polycrystalline (random) surface; if it is too high your adhesion may fail, with very slow rates of growth.

Somebody else may be able to provide some theory.

I have no experience with either CVD or PLD, but it was interesting to think about this question. In a humble attempt to build on Peter Diehr's answer, here is some theory (at a very heuristic and simplified level).

The deposition of each new layer can be thought of as being governed at large scales by some mixture of Laplacian and Eden growth in two dimensions (or some anisotropic variation thereof). Laplacian growth tends to form intricate branch-like structures, while Eden growth (associated with uniform deposition on the surface) would allow adjacent branches to collide. Unwanted dislocations and grain boundaries can arise both from collisions between previously disconnected nucleation regions as well as collisions between adjacent branches of growth structures. The regularity of growth (i.e. the typical length of branches and number of offshoots) is controlled by the rate of diffusion along the boundary of the new layer. When the diffusion rate is small, growth structures tend to look very rough. When the boundary diffusion rate is large, the structures appear much smoother because particles on the boundary have time to reach rough spots before being frozen in place by incoming particles. Since Einstein's relation tells us that the diffusion constant is proportional temperature, i.e. $D=\mu k_BT$, increasing the temperature allows for smoother growth (and presumably lowers the chance of forming dislocations, or of collisions between mismatched crystals, in cases with ambiguity in the natural crystal structure).

Of course, the temperature also controls the rate at which particles evaporate from surface aggregates (and from the surface itself). Hence, there is a tradeoff between the rate of growth and number of uncontrolled defects in the final structure, with the limit on performance controlled by the phase diagram (and to some extent, the rate at which the laser pulses provide new material).

TLDR's point on the temperature-dependence of the diffusion coefficient is indeed an interesting point of departure: substrate temperature alters the crystal surface onto which you deposit, for instance it alters the diffusion coefficient D(r, T) 'experienced' by the arriving species.

This means that, in straightforward terms, you can control the growth mechanism at the surface by controlling substrate temperature.

Note that there are more factors with which you can control the growth mechanism:

• for instance there is still a choice on what crystal face you want to use, for instance one can instead of a Miller Index (0 0 1) face use a (1 1 1) face of a material, which results in a different lattice parameter and thus a different lattice mismatch;

• one an alter background pressure or choose to deposit in a reactive atmosphere (for instance if you deposit a highly reactive element such as Zr in an oxygen-rich atmosphere there is a tendency for oxides (ZrO2 or even higher charge state oxides, this depends on the amount of energy you deposit onto the Zr, which consequently populates the plasma cloud with higher charge species) to form during the plasma expansion, so then part of your species arriving at the substrate will be a Zr oxide instead of elemental Zr. This is called reactive PLD).

There's quite some tricks that one can apply to improve film quality in case of PLD growth, but the advantage of temperature control is that changing the temperature is relatively straightforward: all you have to do is mount your substrate on a heater stage. Many PLD setups provide either resistive of laser based heating stages (ideally i.c.m. a pyrometer setup for a full temperature control)).

To give an example of how substrate temperature influences morphology, we can take a look at binary carbides deposited by PLD. In deposition of binary carbides of refractory metals (ZrC, TiC, TaC, etc.), there is a clear tendency that at low substrate temperature, the resulting samples are amorphous, at intermediate to high temperatures the samples lean towards polycrystalline, and at high temperature these samples become highly single-crystalline. One can check this by performing x-ray diffraction (XRD) on the resulting samples.

Long story short: by altering substrate temperature one alters the surface chemistry of the substrate, which can result in morphological control.

Sources:

Stafe M., Marcu A., Puscas N. Pulsed Laser Ablation of Solids: Basics, Theory, and Application. Springer Series in Surface Sciences 53, 2014.