Metallization by Sputtering

1
Metallization by Sputtering As mentioned
evaporation (filament and e-beam) suffers from
several problems poor step coverage, damage due
to radiation and limited number of materials that
can be deposited by this method. Sputtering is an
alternate deposition method that overcomes all
the above problems. It is now the preferred
method of metallization for the semiconductor
industry. Sputtering is a physical process
whereby energetic ions of inert gas such as argon
in a plasma (glow discharge) are made to impinge
on the surface of a target of the material to
be deposited. Atoms/molecules of the target are
physically detached from the target i.e.
sputtered, and allow to deposit on the wafers to
be coated. The formation of the plasma and
acceleration of ions are achieved in a low
pressure sputtering chamber between two
conducting plates biased by a dc potential (dc
sputtering). A schematic diagram of the process
is shown below
2
The target forms the negative plate (the cathode)
of the parallel plate system and the wafers are
placed on the anode (grounded) plate. The two
plates arrangement is referred to as diode
sputtering. The sputtering of the target atoms
involve complex processes in the acceleration of
argon ions in the plasma and energy exchange
between the target and argon ions when they
arrive at the target.
3

• Note that the plasma is made up mainly of excited
  argon atoms and approximately equal number of the
  positive (argon ions) and free electrons. The
  plasma has characteristic glow regions and dark
  zones.
• For sputtering of insulating material such as
  SiO2, it is necessary to introduce radiofrequency
  power to the target to cause the plasma.
• The spacing between the cathode and anode are
  such that at the pressure of operation (
  millitorr range) the sputtered atoms/molecules
  arrive at the anode without suffering many
  collisions with the gas atoms.
• The deposition rate is related to the generation
  of ions, their arrival rates at the target as
  well as the effectiveness of these ions in
  sputtering atoms/molecules from the target.

4
For example it is known that increasing the
cathode bias, hence energy of the argon ions
arriving at the target, initially increases the
efficiency of sputtering of the target atoms at
the surface but excessive energy cause the ions
to be go deeper into the target where energy is
absorbed (implanted) without producing sputtered
atoms. The deposition rate is a function of the
gas pressure and bias which these are used as
control parameters to achieve the optimal
deposition rate. Various techniques are used to
increase the sputtering rates - use of magnetic
field near the target to increase the generation
of ions magnetron sputtering. Introduction of
electrons by the use of an electron gun as a
third electrode was also tried. triode
sputtering.
5
The spacing between the cathode and anode are
such that at the pressure of operation (
millitorr range) the sputtered atoms/molecules
arrive at the anode without suffering many
collisions with the gas atoms. Schematic diagram
of a sputtering system is shown below.
Cryo pump is a type of high vacuum pump Shutter
is used to shield the wafers (refer to as the
substrate in this diagram) during the initial
phase of the sputtering process so that oxide etc
present on the surface of the target does not get
deposited on the wafers In some systems the roles
of the target and wafer plate are reversed
initially to sputter clean the wafers before
deposition of the target material (of course the
shutter is used to shield the target at this
stage).
6
In general sputtering have the following
advantages over evaporation Better step coverage
because sputtering occurs over the whole surface
of the target extended source versus what is
approaching a point source in evaporation Better
uniformity over a wafer and across wafers (same
reason as above) Possible to sputter wide range
of materials metals, insulators and even alloys,
whereas evaporation is more or less limited to
metals.
7
Chemical Vapour Deposition of Metals A number of
metal films can be deposited by chemical vapour
deposition. Tungsten (W) film is one that is used
most extensively because of its excellent step
coverage property, particularly in submicron
contact holes or via holes (inter-layer metal
connection) as well as for its low resistivity
and ability of forming silicide. The most common
tungsten source is WF6 (tungsten hexafluoride, a
gas) through hydrogen reduction or decomposition
in a cold-wall LPCVD reactor at 350 to 450
C WF6? W 3F2 WF6 3H2 ? W 6HF Figure on
the right shows the ability of deposited tungsten
to fill a very high height-to-width aspect ratio
contact hole. CVD tungsten is used to produce
what is referred to as tungsten plug to fill
contact holes or inter-layer metal vias to
planarize the surface as well as avoiding poor
step coverage if aluminium is used. See diagram
below (Damascene process)
8
The titanium nitride (TiN) film is needed as a
barrier layer to prevent the reaction of WF6 with
the underlying silicon during tungsten CVD and to
increase the adhesion of tungsten plug to the
silicon dioxide. This layer is deposited by what
is referred to as reactive sputtering of
titanium, a process in which nitrogen is
introduced into the plasma to react with the
sputtered tungsten and deposit on the wafer as
TiN. TiN is itself a reasonably good conductor
similar to doped polysilicon.
9
The above is an example of a two-level aluminium
(actually Al-Cu alloy) structure with what is
effectively another level known as local
interconnect provided by a TiN layer. In this
case, the TiN is formed from the same Ti layer
used to form the TiSi2 silicide the Ti layer
over the oxide is allowed to react with nitrogen
during silicide formation. Over the source and
drain where the silicide is formed, the top of
the deposited Ti film also forms TiN in this
process. Note in this way contact to to the n
source or drain is n-TiSi2-TiN (similarly for
the p S/D).