Magnetron Sputtering

Samer Elatrache V-17810600 estudiante de CRF

Electron Motion in Parallel Electric and Magnetic Fields. Let
us now examine what happens when a magnetic field of strength B is
superimposed on the electric field 8 between the target and substrate. Such a
situation arises in magnetron sputtering as well as in certain plasma etching
configurations. Electrons within the dual field environment experience the
well-known Lorentz force in addition to electric field force, i.e.,
m dv
dt
F = - - - -q(g+ v X B),
where q, m and u are the electron charge, mass, and velocity, respectively
First consider the case where B and 8 are parallel as shown in Fig. 3-20a.
When electrons are emitted exactly normal to the target surface and parallel to
both fields, then v x B vanishes; electrons are only influenced by the 8 field,
which accelerates them toward the anode. Next consider the case where the 8
field is neglected but B is still applied as shown in Fig. 3-20b. If an electron is
launched from the cathode with velocity u at angle 8 with respect to B, it
experiences a force quB sin 8 in a direction perpendicular to B. The electron
now orbits in a circular motion with a radius r that is determined by a balance
of the centrifugal (m(v sin 8 ) * / r ) and Lorentz forces involved, i.e., r =
mu sin 8 /qB. The electron motion is helical; in corkscrew fashion it spirals
down the axis of the discharge with constant velocity u cos 8. If the magnetic
field were not present, such off-axis electrons would tend to migrate out of the
discharge and be lost at the walls.
The case where electrons are launched at an angle to parallel, uniform
and B fields is somewhat more complex. Corkscrew motion with constant
radius occurs, but because of electron acceleration in the 8 field, the pitch of
the helix lengthens with time (Fig. 3-2Oc). Time varying fields complicate
matters further and electron spirals of variable radius can occur. Clearly
magnetic fields prolong the electron residence time in the plasma and thus
enhance the probability of ion collisions. This leads to larger discharge
currents and increased sputter deposition rates. Comparable discharges in a
simple diode-sputtering configuration operate at higher currents and pressures,
Therefore, applied magnetic fields have the desirable effect of reducing
electron bombardment of substrates and extending the operating vacuum range
Perpendicular Electric and M8gnetiC Fields. In magnetrons,
electrons ideally do not even reach the anode but are trapped near the target,
enhancing the ionizing efficiency there. This is accomplished by employing a
magnetic field oriented parallel to the target and perpendicular to the electric
field, as shown schematically in Fig. 3-21. Practically, this is achieved by
placing bar or horseshoe magnets behind the target. Therefore, the magnetic
field lines first emanate normal to the target, then bend with a component
parallel to the target surface (this is the magnetron component) and finally
return, completing the magnetic circuit. Electrons emitted from the cathode are
initially accelerated toward the anode, executing a helical motion in the
process; but when they encounter the region of the parallel magnetic field, they
are bent in an orbit back to the target in very much the same way that electrons
are deflected toward the hearth in an e-gun evaporator. By solving the coupled
differential equations resulting from the three components of Eq. 3-39, we
readily see that the parameric equations of motion are where y and x are the distances above and along the target, and w, = qE/m.
These equations describe a cycloidal motion that the electrons execute within
the cathode dark space where both fields are present. If, however, electrons
stray into the negative glow region where the 8 field is small, the electrons
describe a circular motion before collisions may drive them back into the dark
space or forward toward the anode. By suitable orientation of target magnets, a
"race track" can be defined where the electrons hop around at high speed.
Target erosion by sputtering occurs within this track because ionization of the
working gas is most intense above it.
Magnetron sputtering is presently the most widely commercially practiced
sputtering method. The chief reason for its success is the high deposition rates
achieved (e.g., up to 1 pm/min for Al). These are typically an order of
magnitude higher than rates attained by conventional sputtering techniques.
Popular sputtering configurations utilize planar, toroidal (rectangular cross
section), and toroidal-conical (trapezoidal cross section) targets (Le., the
S-gun). In commercial planar magnetron sputtering systems, the substrate
plane translates past the parallel facing target through interlocked vacuum
chambers to allow for semicontinuous coating operations. The circular
(toridal-conical) target, on the other hand, is positioned centrally within the
chamber, creating a deposition geometry approximating that of the analogous
planar (ring) evaporation source. In this manner wafers on a planetary substrate
holder can be coated as uniformly as with e-gun sources.

Reactive Sputtering
In reactive sputtering, thin films of compounds are deposited on substrates by
sputtering from metallic targets in the presence of a reactive gas, usually mixed
with the inert working gas (invariably Ar). The most common compounds
reactively sputtered (and the reactive gases employed) are briefly listed:
1. Oxides (oxygen)-Al,O,, In,O,, SnO,, SO,, Ta,O,
2. Nitrides (nitrogen, ammonia)-TaN, TiN, AlN, Si,N,
3. Carbides (methane, acetylene, propane)-Tic, WC, Sic
4. Sulfides (H,S)-CdS, CuS, ZnS
5. Oxycarbides and oxynitrides of Ti, Ta, Al, and Si
Irrespective of which of these materials is considered, during reactive
sputtering the resulting film is either a solid solution alloy of the target metal
doped with the reactive element (e.g., TaN,,,,), a compound (e.g., TiN), or
some mixture of the two. Westwood (Ref. 25) has provided a useful way to
visualize the conditions required to yield alloys or compounds. These two
regimes are distinguished in Fig. 3-22a, illustrating the generic hysteresis
curve for the total system pressure (P) as a function of the flow rate of
reactive gas (Q,) into the system. First, however, consider the dotted line
representing the variation of P with flow rate of an inert sputtering gas (Q,).
Clearly, as Qi increases, P increases because of the constant pumping speed
(see Eq. 2-16). An example of this characteristic occurs during Ar gas
sputtering of Ta. Now consider what happens when reactive N, gas is
introduced into the system. As Q, increases from Q,(O), the system pressure
essentially remains at the initial value Po because N, reacts with Ta and is
removed from the gas phase. But beyond a critical flow rate QF, the system
pressure jumps to the new value P,. If no reactive sputtering took place, P
would be somewhat higher (i.e., P3). Once the equilibrium value of P is
established, subsequent changes in Q, cause P to increase or decrease linearly
as shown. As Q, decreases sufficiently, P again reaches the initial pressure.
The hysteresis behavior represents two stable states of the system with a
rapid transition between them. In state A there is little change in pressure,
while for state B the pressure varies linearly with Q,. Clearly, all of the
reactive gas is incorporated into the deposited film in state A-the doped metal
and the atomic ratio of reactive gas dopant to sputtered metal increases with
Q,. The transition from state A to state B is triggered by compound formation
on the metal target. Since ion-induced secondary electron emission is usually
much higher for compounds than for metals, Ohm's law suggests that the
plasma impedance is effectively lower in state B than in state A. This effect is
reflected in the hysteresis of the target voltage with reactive gas flow rate, as
schematically depicted in Fig. 3-22b.
The choice of whether to employ compound targets and sputter directly or
sputter reactively is not always clear. If reactive sputtering is selected, then
there is the option of using simple dc diode, RF, or magnetron configurations.
Many considerations go into making these choices. and we will address some
of them in turn.
Target Purity. It is easier to manufacture high-purity metal targets
than to make high-purity compound targets. Since hot pressed and sintered
compound powders cannot be consolidated to theoretical bulk densities, incorporation
of gases, porosity, and impurities is unavoidable. Film purity using
elemental targets is high, particularly since high-purity reactive gases are
commercially available.
Deposition Rates. Sputter rates of metals drop dramatically when
compounds form on the targets. Decreases in deposition rate well in excess of
50% occur because of the lower sputter yield of compounds relative to metals.
The effect is very much dependent on reactive gas pressure. In dc discharges,
sputtering is effectively halted at very high gas pressures, but the limits are
also influenced by the applied power. Conditioning of the target in pure Ar is
required to restore the pure metal surface and desired deposition rates. Where
high deposition rates are a necessity, the reactive sputtering mode of choice is
either dc or RF magnetron.

Stoichiometry and Properties. Considerable variation in the
composition and properties of reactively sputtered films is possible, depending
on operating conditions. The case of tantalum nitride is worth considering in
this regard. One of the first electronic applications of reactive sputtering
involved deposition of TaN resistors employing dc diode sputtering at voltages
of 3-5 kV, and pressures of about 30 x torr. The dependence of the
resistivity of "tantalum nitride" films is shown in Fig. 3-23, where either Ta,
Ta,N, TaN, or combinations of these form as a function of N, partial
pressure. Color changes accompany the varied film stoichiometries. For
example, in the case of titanium nitride films, the metallic color of Ti gives
way to a light gold, then a rose, and finally a brown color with increasing
nitrogen partial pressure.