46 Experimental Methods
pressure will result in useful deposition rates. The mean free path, , can be
evaluated from the following equation, assuming the ideal gas law:
=
kT
p
2P
(3.1)
where k is the Boltzmann constant, T the Temperature, the cross sectional
area and P the pressure. For a reasonable rate of deposition the mean free path
should be in the same order of magnitude as the target to substrate distance.
For = 10 cm the pressure should be around 10��3 mbar, which is in the range
of pressures described above.
In practice the growth parameters can lead to stress in the lms, which may
detach from the substrate during testing or in case of very thin substrates, a
highly strained lm can cause the substrate to bend or even break. Obtaining
strained lms or quantication of stress/strain in the lms produced have not
been pursued here but it was in some instances observed as rapid lm detachment.
In those cases sputtering parameters dierent from the standard had
been used and was not used again.
An important aspect that has so far been briey touched upon is the source
of the Ar ions, the plasma. A plasma is created by applying a large potential
dierence between a cathode and an anode in the chamber. With a sucient
energy input atoms become ionized and a large number of ions and electrons
are created. Due to the high density of charged particles, the plasma will react
to magnetic elds, which is exploited in magnetron sputtering systems. In such
systems, which was used throughout this project, a strong magnetic eld traps
the electrons and concentrates the plasma close to the target material, which has
a negative bias. The magnetron systems make it possible to lower the sputtering
pressure signicantly compared to traditional glow discharge systems, due to the
conned plasma, while achieving high deposition rates.
3.1.1.1 Reactive sputter deposition
In reactive sputter deposition the parameter space is further enlarged by adding
the partial pressure of a reactive gas. The reactive gas in this project is always
oxygen due to the focus on preparing metal oxides. The presence of oxygen in
the chamber complicates the deposition process in several ways. Oxygen will
interact with the target causing surface oxidation and changes in the sputter
yield. Most of the transition metals oxidise with a very negative heat of formation
and the sputter yields go down. Furthermore, the deposited atoms oxidize
and the density is not always easily obtained for the oxide prepared. The exact
oxide phase will also be unknown until further characterization has been carried
out. To explore the eects of adding oxygen a rate vs. oxygen partial pressure
curve is typically made.
There is a hysteresis behavior when increasing and decreasing the oxygen partial
pressure. While increasing the partial pressure more of the sputtered atoms will