3.1 Sample Preparation 45
Figure 3.2: Sputter yield as function of the atomic number, from 0 to 90. Lines
between points indicate 3d, 4d and 5d metal series. Figure taken from 187.
yields have not been carried out in this thesis but in practice the eect of
dierent sputter yields is easily observed when comparing the power required
for various targets to obtain desired deposition rates. For every new target a
calibration of rates vs. power is carried out in order to know the thickness of the
lms prepared. The rate calibration is done with an in-chamber quartz crystal
microbalance. In short a quartz crystal is used as deposition substrate while the
change in resonant frequency is measured over time. The change in frequency
is linearly related to the amount of material deposited. When the density is
known, the change in frequency can be directly converted into thickness growth
(Angstrom) per second. Details on how a quartz crystal microbalance works is
given in section 3.4.1.
The power input and chamber pressure also aect how the sputtered particles
grow on the substrate 188, 189. The energy of incoming Ar ions aect the
sputtered atoms and the energy with which they leave the target surface. In
general for low energies, a few hundred electron volts, the sputtered species will
be single neutral atoms while for very high energies, several thousands electron
volts, clusters are formed. In our case we use from 100 to 750 eV, hence the
lm growth is expected to be primarily from single atoms. The energy of the
sputtered atoms also depends on the chamber pressure. A higher pressure leads
to more collisions and by the time the sputtered atoms hit the substrate their
average energy will be lower. The mean free path of the sputtered atoms can be
a useful way to estimate whether the target to substrate distance and chamber