Plasma Assisted Physical Vapour Deposition
Plasma Assisted Physical Vapour Deposition
Plasma Assisted Physical Vapour
Bangalore Plasmatek Pvt.Ltd.
(Off): 302, RV Apartments, Sri Rama Mandira Road, Basavanagudi,
Bangalore 560 004
Modification of the surface properties of a solid by appropriate coatings
has been practised by man since prehistoric days. Introduction of new
technologies like electroplating and physical vapour deposition in vacuum
over the past hundred years has resulted in unprecedented growth in the in
this field. During the past two decades non equilibrium plasmas have
revolutionised the surface coating technologies opening up a variety of
applications which are otherwise impossible. Principles and practise of some
of these new coating techniques would be briefly presented in this talk.
Bangalore Plasmatek has developed a prototype coating system based on vacuum
arcs and magnetron sputtering. Our experience in developing special coatings
acceptable to end-users would also be discussed.
Keyword: Plasma, Coatings, Brazing, Titanium, Graphite
A large number of objects are used in our daily life. Some
are natural and some are man made. Each one is expected to
meet a variety of specifications which could be conflicting
with each other. An iron tool should be strong but not
corrode. Gold ornaments should be glittering but not too
soft. An ornament should look like gold but may not be as
expensive. These are only few examples. Since prehistoric
times man has been developing different techniques to
achieve the required material properties. Alloying, heat
treatment, painting, electroplating etc are some of the
techniques developed over the centuries. Demand and search
for novel materials and techniques continue. Developments in
vacuum technology and plasma science have given new
directions to surface engineering – thin film deposition in
particular. Vacuum environment protects the processes
against attack from atmospheric gases. Large mean free paths
and low boiling points in vacuum are helpful in depositing
Low pressure plasmas can provide more energy to the
depositing films at lower temperatures. Plasmas can be
manipulated with electric and magnetic fields. This has been
exploited in exploiting plasma interactions with materials
and developing novel techniques of thin film deposition.
Further, non-equilibrium properties of low pressure plasmas
promote certain reactions which are otherwise difficult and
would require high temperatures. Some of these thin film
depositions methods would be discussed.
Physical vapour deposition (pvd)
A conventional vacuum coating system is shown schematically
in Fig.1. Material to be deposited is heated and vaporised
in vacuum. The vapours condense on the substrates which are
kept at specified distance from the vapour source. Often the
substrates are cleaned in situ using a glow discharge
plasma. This improves the adhesion of the film. Increasing
the substrate temperature improves the adhesion further and
also helps in getting denser films with fewer voids.
Resistive heating, high energy electron beam heating and
induction heating are some of the options available for
vaporising the source material.
Using several sources it is possible to deposit multi layer
films. Optical interference filters are routinely made by
this technique. Other applications include decorative
coatings, metallization for packaging and electronic
industries etc. In spite of its wide spread applications
there are several draw backs in the conventional physical
vapor deposition (PVD) technology. Some of them are poor
adhesion, low density with voids, difficulties in depositing
alloys and compounds etc. Many of these difficulties can be
over come by plasma based techniques.
Figure 1. Vacuum Coating Unit
The basic principles
The basic principles of plasma assisted physical vapour
deposition (PAPVD) are illustrated in Fig 2. The
substrate is bombarded with energetic ions while the
film is being deposited . The ions – could be argon –
may have energies up to 1keV. These ions share their
energy and momentum with the depositing atoms. As a
result the mobility of the depositing atoms increase
leading to denser films. It is like shaking a bottle
while filling it with grains. This is illustrated in
Fig.3 which is adapted from published results of
computer simulation . As the ratio of ion flux to the
flux of the depositing atoms increase the number of
voids in the film decrease making it denser. Increasing
the energy of the ions or the temperature of the
substrate has a effect. However, the bombarding atoms
can also sputter away some of the atoms being deposited
and reduce the rate of film growth. Ion bombardment
increases the substrate temperature and influences the
film growth further. In addition, if the same ions
bombard the substrate before deposition, the impurities
sitting on the surface and possibly some atoms of the
substrate as well get etched away. This enhances the
bonding between the substrate and the depositing atoms.
There are several ways in which ion bombardment can be
realised. In the system shown in Fig.1, an ion source is
added to a conventional PVD system. Argon ions up to
1keV are normally used. Ion flux and energy can be
independently controlled. Such systems are difficult to
scale up for industrial use.
Energetic ions bombarding a surface can dislodge the
atoms from the solid when the kinetic energy imparted by
the ion is sufficient to overcome the binding energy of
the atoms. This is called sputtering. The sputtered
atoms have energies in the range of few eV. This is
quite large compared to a fraction of an eV of the
thermally evaporated atoms. The films obtained when
these atoms condense are superior to thermally
Figure 2. Principles of PAPVD
Figure 3. Films become denser with increasing ion flux
A simple coating unit based on sputtering uses a glow
discharge plasma. The material to be deposited is taken
as the cathode and the chamber walls act as anode. The
substrates are fixed in front of the cathode. A glow
discharge is created by applying high voltage (~ few
hundred volts) under low pressure of argon. A bright
glow is seen around the cathode as the discharge
strikes. The extent and nature of the glow depends on
the pressure, voltage and the cathode surface. Most of
the applied voltage falls over a short distance in front
of the cathode. The rest of the space between the
cathode and the anode contains a low density plasma and
is almost field free. The ions from the plasma gain full
energy in the cathode fall area - called the sheath -
and bombard the cathode with energy nearly equal to the
(a) Magnetron sputtering
(b) top view of the discharge
The secondary electrons released from the cathode
surface gain energy from the sheath and enter the plasma
and ionise the gas there to maintain the discharge. The
energetic ions impinging the cathode cause sputtering.
The neutral atoms emitted from the cathode reach the
substrates and deposit as a film. The substrate is also
exposed to the plasma. Negative bias to the substrates
results in ion bombardment of the growing film. But the
growth rate is limited by the ion flux to the cathode
which is nominal in the usual glow discharge.
Plasma density close to the cathode can be increased by
using a confining magnetic field as shown in fig. 4. One
pole is in the center of the cathode. The other pole
forms an annular ring around it. This results in a
highly non uniform magnetic field in front of the
cathode. Such a device is called a magnetron. Electrons
gyrate around the field lines and find more time to
ionise the gas near the cathode. Thus plasma density and
sputtering rate increases. The crossed electric and
magnetic fields give rise to ExB drift of the plasma as
shown in the figure. This makes the plasma move in
circle as shown in fig 4. As ions are not magnetised
this drift is only for the electrons. This results in a
toroidal electron current. The vertical components of
the magnetic field exert jxB forces which stabilise the
plasma in the zero vertical field region. Plasma is
therefore constrained to move in a narrow circular path
in the toroidal region. Sputtering is maximum from this
part of the cathode. By adjusting the parameters like
discharge voltage/current, pressure, magnetic field and
coating geometry one can get high deposition rates.
It is possible to have large sputtering cathodes and
they can be fixed in any desired position. This a
decided advantage over the conventional PVD. Of course,
the sputtered atoms are more energetic than evaporated
atoms. This enhances the density of the films. However,
because of magnetic confinement, plasma density close to
the substrates is rather low. Hence, ion to atom flux is
poor at the substrates. But the advantage is that
substrate heating is now reduced and it is easier to
coat thermally sensitive materials.
Ion bombardment of the film can be realised if it is
surrounded by a plasma. In the magnetron this avoided by
confining the plasma close to the cathode. Window and
Savvides  opened up some field lines and made them go
all the way to the substrate. This was achieved by
introducing some imbalance between the inner and outer
magnets of the magnetron. One useful way of doing this
is shown in fig. 5. Some electrons from the cathode
region escape along the field lines towards the
substrate. This would be followed by the ions. In
addition, the electrons can ionise the gas in front of
the substrate creating a dense plasma. A negative bias
to the substrate results in the desired ion bombardment.
Ion energy can be adjusted by varying the bias. It is
possible to achieve high ionic to atomic flux ratios
with unbalanced magnetrons.
Reactive deposition is possible with magnetron
discharges. In the presence of dense plasma reactivity
of gases like nitrogen, methane etc with metals like Ti,
Figure 5 Unbalanced magnetron
etc is enhanced. Nitrides, carbides and carbo-nitrides
are routinely deposited for industrial applications.
However, operating parameters need to be controlled
carefully. For example, if the partial pressure of
reactive gasses is too low sub stochiometric films are
formed. On the other hand at higher pressures compound
are formed on the cathode and sputtering rate reduces.
This is called target poisoning.
Deposition of insulating films like oxides is not
straight forward. The first problem is that the cathode
becomes prone to arcing as an oxide film is formed on it
and electric charge is built up. This takes the
discharge from glow to arc region which is detrimental
to the system. The second problem is that as the anode
gets coated it becomes insulating. The path for the flow
of discharge current is blocked and the discharge is
quenched. This is called the case of disappearing anode.
Both these problems can be overcome by using the so
called bipolar or dual magnetron . This is a system
of two identical magnetrons. Instead of supplying dc
power, alternating high voltage at high frequency of few
tens of kHz is supplied to the two as shown in Fig.6.
Figure 6. Dual Magnetron
There is no separate anode. For one half of the cycle
one magnetron acts as the cathode and the other as the
anode. In the second half their roles are reversed.
There is no oxide coating or charge build up on either
of the magnetrons.
Magnetrons are widely used in industries. Cathodes from
25 mm to 3500 mm are in use. Power up to 200 kW is
supplied. Most commonly used coatings are metals and
their nitrides, carbides oxides and their combinations.
Applications range from cutting tools, decorative
coatings of a few centimeters to several meters long
architectural glass and steel sheets.
An electric arc is generated when the flow of
current is suddenly interrupted. It is seen as
bright flash of light. This is normally occurs when
mechanical switches are operated, at loose contacts
etc. During an arc the electrode material evaporates
and reacts with atmospheric gases. This gets
deposited as a coloured layer on nearby surfaces. An
electric arc can also be generated in vacuum. This
illustrated in Fig.7. A voltage is applied between
the cathode and the anode which could be the wall of
the vacuum chamber. The cathode is brought in
contact with the anode using the igniter. A large
current flows. When the igniter is withdrawn the arc
is generated on the cathode surface. One can see
bright spots – called the cathode spots - moving
randomly all over the cathode surface. A current is
maintained in the circuit. It could be few tens of
amperes to few kilo amperes.
Figure 7 Cathode spot phenomenon
It is limited only by the power source. The voltage
across the electrodes falls to about 25V. This is
called a random arc.
Most of the processes of the arc occur very close to
the cathode spot which has a short life of 1–5 s
and occupies a tiny space of 10 – 30 m. This
transient nature makes it difficult to study the
arc. These spots move randomly in the absence of any
magnetic field. As one spot dies another is born
some where along the periphery of the dying spot.
The physical phenomena of cathode spots are roughly
understood as follows . The current density is
106 – 108A/m2 This locally heats up the cathode to
very high temperatures and an area corresponding to
the cathode spot melts. This micro pool of molten
cathode emits electrons by thermionic and field
emission processes. These electrons ionise the
vapours of the cathode material escaping the molten
pool. Extremely high-density plasma (1024 – 1026
m-3) is formed very close to the cathode. This
plasma is of the cathode material. It is not
necessary to have any gas to maintain the arc
discharge. Cathodic arc is also known as vacuum arc.
Some of the ions from the plasma are attracted to
the cathode. The escaping vapours also exert a
pressure on the molten pool. These two factors cause
splattering of the molten cathode material leading
to emission of droplets of various sizes –
sub-microns to a few microns-at large angles to the
normal of the cathode surface.
It is observed that the ions are multiply ionised.
Ions with charge states up to 6+ have been observed.
Ions from low melting point metals like lead are
singly ionised while for refractory metals average
ion charges up to 3+ are observed. The degree of
ionisation is seen to decrease with increasing arc
current. It is possible that multiple ionisation is
caused by multiple collisions with electrons in the
high density plasma close to the cathode. But the
details need to be understood.
The ions from the cathodic arc can have energies up
to 150 ev. This is more than the cathode voltage of
about 25 v. Two alternate explanations have been put
forward to explain this anomaly. As per the
potential hump theory , a region of higher
potential is formed near the cathode due to excess
ionisation. Ions in this region are accelerated
large energies. The alternate explanation is called
the gas dynamic theory . According to this,
multiple collisions with electrons in the region of
high density plasma leads to ion acceleration. If
this is true, the ion energy should be independent
of its charge state while as per potential hump
theory higher charge state ions should have higher
energy. There is no unambiguous evidence in favour
of either of the explanations.
Coatings obtained from cathodic arc sources are very
dense and adhesion to the substrates is good. This
is because the bombarding ions have high energy
which can be further increased by biasing the
substrates. Since the plasma is almost fully ionised,
the ion to atomic flux ratio is very high. The films
are free of voids. There is no gas entrapment as no
working gas like Ar is used. Reactive deposition is
more efficient as the ions are energetic and ionised
unlike in magnetron plasmas. Operating pressures
with reactive gases like nitrogen are not critical
as the optimum value has a broad window. One can
have very large cathodes similar to magnetrons.
The only problem with the cathodic arc is due to
micro droplets that are emitted from the molten pool
of the cathode spot. These make the coatings rough
and textured. This is not acceptable for some
applications like optical coatings. Several
techniques have been devised to get smooth films.
Some are briefly discussed below.
The droplets are emitted preferentially at large
angles with respect to the normal of the cathode
surface. Using collimators one can prevent most of
the macro particles from reaching the substrates.
This imposes geometrical restrictions for industrial
applications and the films are not totally
Here the attempt is to minimise the splattering and
production of droplets . This is done by
imparting a large velocity to the arc using a
magnetic field as shown in Fig.8.
(a) Steered arc cathode and
(b) Arc as seen from above
The radial component Br and the arc current j which
is perpendicular to the cathode surface conspire to
give rise to j x B force. This causes the spots to
trace a circle but in a direction opposite to the
force! This is the so called retrograde motion of
the arc. Various explanations are offered to
understand this . Trans. On Plasma Science
18(1990)883]. Here we would not go into these. But
it is important to note following experimental
observations. (i) Arc velocity increases with
magnetic field and the arc current. (ii) Velocity
depends on the material of the cathode. For example,
it is visibly slow for aluminium and it is very
difficult to make the arc move on a graphite
cathode! (iv) At higher magnetic fields the arc
tends to become unstable. This effect is more
pronounced in the presence of a reactive gas like
nitrogen. These factors are important in an
industrial coating unit and need to be understood
better. The ExB force gives rise to a current loop
in the toroidal channel. As in the case of the
magnetron, this current channel is stabilised by the
vertical component of the magnetic field in a region
of the minimum vertical field. This is called acute
angle effect because the plasma is pushed inwards
away from the acute angle the field line makes with
the cathode surface. But it is not clear as to what
makes the arc unstable.
The velocity steered arc is in the range of 2 m/s to
45 m/s [10. The residence time of the cathode spot
at each position is reduced. The size of the molten
metal pool and hence the amount of splattering is
minimised. The number of macro particles in the film
is reduced considerably. The angular distribution of
the macro-particles also change. Magnetic field can
not be increased indefinitely as at higher fields
(~500 gauss) the arc becomes unstable. The reduction
is more pronounced in the case reactive deposition.
This could be due to the formation of a thin layer
of high melting point film – say titanium nitride –
on the cathode. On such films cathode spots move
faster. Complete elimination of macro particles is
not possible with the steered arc. In addition to
reducing macro particles, steered arc is more
favourable for reactive deposition. This is due to
the fact that magnetised electrons near the cathode
excite, disassociate and ionise reactive gases like
nitrogen. This is seen as a deep purple hue
emanating at the arc and spreading out along field
lines. Large steered arc cathodes are being used in
industry for cutting tools, decorative coatings and
Figure 9 Enhanced arc
A schematic drawing of an enhanced arc is shown in
Fig. 9 
It essentially consists of non uniform magnetic
field generated by a combination of coils and pole
pieces. The plasma flows through this field in a
narrow duct. This appears to be a modification of
the steered arc. At the cathode the field lines form
an acute angle. This confines the plasma to a small
region in the center. The emission of droplets is
reduced as in the steered arc. The expanding plasma
encounters the non uniform and strong magnetic field
near the pole gaps. Here as it gets compressed and
density increases. The electrons and ions bombard
the droplets escaping from the cathode. Smaller
drops are completely vaporised and larger ones
become smaller. The net effect is an increase in
plasma density and a decrease in the number of macro
particles. It is said that the coating could be made
almost free of macro-particles. But it is difficult
to have large cathodes for industrial requirement.
As an alternate one can increase the number of
cathodes in the coating system.
Here the aim is to separate the macro particles from
ions and electrons using a curved magnetic field.
Such a filter was first used by Askenov and his
co-workers . Since then plasma transport through
such filters has been studied by various workers [8,
14, 15, 16, 17]. An example of this is shown in
Figure 10. Filtered arc
Here a 900 bend or a quarter of a toroid is used.
Cathode is at end and the substrates are at a
distance from the other end. The magnetic field is
such that the diameter of the duct is much smaller
than the ion gyro radius and much larger than
electron gyro radius. In other words, the electrons
are magnetised and the ions are not. Electrons
spiral around the field lines and go out. Ions
follow the electrons to maintain plasma neutrality.
The macro particles, even if they are charged, are
not affected by the magnetic field as they are too
heavy. They travel in straight lines and hit the
walls of the duct. The duct is designed such that it
is optically blind so that macro particles can not
reach the substrates unless they are multiply
reflected from the walls. Thus one can get almost
macro particle free films. All the plasma generated
at the cathode would not reach the exit of the duct.
Some plasma is lost to the walls due to cross field
diffusion and various drifts caused by curved
magnetic fields.. Transmission efficiency can be
increased by biasing the duct as shown in the figure
10. Transmission efficiency can also be improved by
increasing the magnetic field. However, beyond a
certain field efficiency starts reducing due to the
onset of instabilities in the plasma .
Filters come in various shapes and sizes [17, 18,
19]. They include S filters  and off plane
double bend filters . Design objects are to
eliminate macro particle flux and increase plasma
transport efficiency. It depends on various factors
like the diameter and other geometrical factors of
the duct, magnetic field , bias on the duct, cathode
material etc. The efficiency is specified by the
system coefficient which is as the ratio of the
total ion current at the exit of the filter to the
arc current. For unfiltered arcs this is usually in
the range of 10% to 20%. For 900 filters a system
coefficient of 2.5% has been achieved. About 0.6% is
reported for an S filter. Filtered arc have limited
industrial use due to low deposition rates and
restricted geometry make.
Electron beam evaporation
This resembles conventional PVD system. The material to
be evaporated is placed in a water cooled copper hearth.
It is heated by an intense beam of electrons of about
100 ev energy in the presence of an inert gas like argon
or a reactive gas like nitrogen or both. The electron
impact ionisation cross section peaks for most elements
around these energies. Therefore, the electron beam, in
addition to acting as source of heat, forms a plasma of
the ambient gas as well as that of evaporating material.
By suitably biasing the substrates one can have desired
ion bombardment of the growing films. The electron beam
is derived from a hot cathode filament or a plasma
source. A tungsten filament at the top emits the
electron beam. The copper hearth containing the
evaporant is at the bottom of the chamber. The entire
system is immersed in a magnetic field. This helps in
guiding and focusing the electron beam to the hearth. On
their way the electron ionise the back ground gas as
well as the metal vapours coming from the hearth. Ion to
atom flux ratio at the substrates could be high and good
quality films are obtained. Deposition rates can be
adjusted by the beam power. However, too intense beams
can lead to splattering and droplets.
Unlike in magnetron and cathodic arc systems, here it is
not possible to have large evaporating sources and
coating geometry is restricted. On industrial scale,
electron beam evaporating systems are being used for
Coating at Bangalore plasmatek
Bangalore plasmatek has developed a flexible coating system.
It can accommodate two large (500 mm) steered cathode arc
sources and several smaller magnetrons and steered arc
sources. A photograph of the system is shown in Fig. 11.
Figure 11. Coating Chamber at BPT
The cubical vacuum chamber has internal dimensions of 1000 x
1000 x 600 mm. About two hundred 10 mm drill bits can be
coated in one batch. The machine is generally used for hard
coatings like TiN, TiCN and decorative coatings like TiN/Au.
Some of the coated items are shown in Figs. 12 and 13.
Figure 12. Coated Cutting tools
Figure 13. TiN /Au coated watch straps
A typical coating cycle for TiN coating involves following
steps. Ultrasonic cleaning either in a solvent or aqueous
based cleaning solution, drying, loading, evacuating, glow
discharge cleaning, high energy ion etching, reactive
deposition, cooling and unloading. For decorative
applications, gold is deposited by magnetron sputtering
after depositing TiN without venting the system.
The system has also been used for aluminising telescopic
mirrors. X–ray mirrors of an x-ray diffraction machine have
been coated with nickel.
Coating gold on glass is difficult. Special procedure has
been developed to achieve good adhesion of gold films on
Brazing of ceramics to metals has many applications.
Conventionally it is achieved by a series of complicated
steps of moly-manganese metalisation and vacuum brazing. At
Bangalore Plasmatek we have reduced the number of steps of
metalisation. The ceramic parts are coated with titanium
film of about 12-15 m. Brazing is then done in vacuum with
usual filler material. This has been successfully tried out
for brazing alumina ceramics to stainless steel, copper and
titanium metals. Some of the brazed samples are shown in
(a) Ceramic brazed to titanium (b) Graphite brazed to copper
(c) Ceramic brazed to stainless steel
For certain applications it may not be possible to use
brazing alloys in the form of foils, wires , powders or
pastes due to geometrical and other considerations. A
special composite cathode of copper and silver has been
developed for depositing CuSil coatings for such
applications. Using the CuSil coatings in conjunction with
Ti metallisation good ceramic to metal brazing has been
Joining of graphite to metals has many important
applications. For good thermal and electrical contact at the
junction brazing is unavoidable. Bangalore Plasmatek has
been successful in brazing graphite to copper by titanium
metallisation of graphite.
Physical vapour deposition has come a long way from the days
of simple vacuum coating – thanks to plasma based
technologies. Today large coating machines have become
integral part industries like cutting tool manufacturing,
architectural glass production, steel mills, watch
components manufacturing units etc. In India several such
coating units are operating. But developmental work in this
field is in its infant stage compared to many other
developed and developing countries. Bangalore Plasmatek has
taken a small step in this direction. An integrated approach
involving vacuum technology, power electronics, automation
and plasma technology is required for the development modern
coating machines. Without that we can at best be assemblers.
[ 1] D.M. Mattox, J. Appl.Phys. 34(1963)2493
 K.H. Muler, J.Vac. Sci. Technol. A4,184(1986); A5,
[3 ] B. Window and N. Savvides, J.Vac. Sci. Technol.
A4(1986)196, 453, 504]
[4 ] T. Rettich and P. Wiedemuth, J. Non-Crsyt. Solids,
 A.E.Guile and B.Jutner, IEE. Traans. on Plasma Science
 Lloyd, 1(1971)184], a Lloyd, 10th Intl Conf. Phen.
Ionised Gases, Oxford 1(1971)184]
 Lyubimov, Sov. Phys. Tech. Phys. 22(1977); 18(11973)565]
 D.M. Sanders, D.B.Boercker and S. Falabella, IEEE.
Trans. On Plasma Science 18(1990)883].
 S. Ramalingam, Int. Patent No. WO 85/03954, 1985
 Chang Rung feng PhD Thesis, University of Minnesota
 B.F.Coll, and D.M.Sanders, Surf. Coat. Technol
 I.I.Askenov, V.A.Blous, V.G. Padalka and V.M. Khoroshik,
Sov.J.Plasma Phys. 4(1978)425; Instrum. Exp.
 Marcela M. M. Bilek and Andre' Anders, Plasma sources
Sci. Technol. 8(1999)488].
 Andre' Anders, Simone Anders and Ian G Brown, Plasma
Sources Sci. Technol. 4(1995)1
 I.I.Askenov, A.N. Belokhvostnik, V.G. Padalka ,N.S.
Repalov and V.M. Khoroshik Plasma Phys.. Control Fusion
 V.S. Veeraswamy, G.A.J. Amaratunga and W.I. Milne IEEE.
TRAANS. Plasma Sci. 21(1993)322
 S.Anders et.al. IEEE. Traans. Plasma Sci. 26(1997)670]
 Andre' Anders, Sur. Coat. Technol. 120-121(1999)319]
 D.A.Karpov, Sur. Coat. Technol 96(1997)22
 A. Harkovita, J. Salo, A.Anttila, R.Lappalainen,
Diamond Relat. Mater. 4(1995)1335
 X. Shi et.al. Thin Solid Films 345(1999)1
Plasma Assisted Physical Vapour Deposition