Physical Vapor Deposition

EE-527: MicroFabrication
Physical Vapor Deposition
R. B. Darling / EE-527 / Winter 2013

Physical Vapor Deposition (PVD)
Gas Phase Gas Phase
Transport
Evaporation Condensation
Condensed Phase Condensed Phase

(solid or liquid) (usually solid)

Equilibrium Vapor Pressure
• P* is the partial pressure of a gas in equilibrium with its
condensed phase at a given temperature T.
– There is no net transfer of material from one state to the other.
• For a given material, P* is only a function of T.
– But the dependence of P* on T is rather complicated.

Evaporation Rates - 1
• P* is the equilibrium vapor pressure of the evaporant at T.
• P is the ambient hydrostatic pressure acting upon the
evaporant in the condensed phase.
• Heinrich Hertz found experimentally that the evaporation
rate was proportional to (P* - P).
– This is consistent with kinetic theory in which the impingement
rates are proportional to pressure.
– Hertz also found that the evaporation rate could not be increased
by supplying more heat unless the equilibrium vapor pressure was
also increased by this action.
– Thus, there is a maximum evaporation rate set by P*, and this is
only achieved in a vacuum, where P = 0.

• This can be viewed as two opposing fluxes:
The net evaporation flux is the difference

between the impingement rates for the
evaporant vapor two fluxes:
 
rate goes as P
dN e
 2mk BT 
1 / 2
P*  P
Ae dt
rate goes as P*
condensed evaporant

• Hertz only measured rates of about 1/10 of the above using
Hg vapor.
• Knudsen postulated that the evaporant vapor molecules
impinging upon the condensed phase surface may be
reflected back.
– v = sticking coefficient for vapor molecules onto the surface.
– Then a (1 - v) fraction of the vapor molecules contribute to the
evaporant pressure, but not to the evaporant flux.
– Therefore, the vapor pressure must be higher by a factor of 1/v to
obtain the same evaporation rate.
• This gives the general Hertz-Knudsen equation:
  v 2mk BT  P *  P 
dN e 1 / 2
Ae dt
Mass Evaporation Rates
•  = mass evaporation rate in g/cm2-sec:

1/ 2
 m 
m
dN e
  v   P *
P 
Ae dt  2k BT 
• For most elements,  ~ 10-4 g/cm2-sec at P* = 10-2 torr.
• The mass of the evaporated material is
t Ae
M e     dAe dt
0 0

Free Evaporation Versus Effusion
• Evaporation from a free surface is termed Langmuir
evaporation.
• Because v is often much less than unity, the general
Hertz-Knudsen expression must be used.
• Effusion refers to evaporation through an orifice by which
the area of the orifice appears as an evaporation source of
the same area.
• Free evaporation is isotropic.
• Effusion is somewhat directional.
– Ideally, it is a Lambertian angular distribution.

Knudsen Cell
• In 1909 Knudsen invented a technique to force v to 1.

• This is called a Knudsen cell, effusion cell, or “K-” cell.
• The orifice acts like an evaporating surface of area Ae at P*,
but it cannot reflect incident vapor molecules, so v = 1.
Isothermal enclosure with a small orifice of area Ae.
Surface area of the evaporant inside the enclosure

is large compared to the area of the orifice.
An equilibrium vapor pressure P* is maintained

inside the enclosure.
Effusion from a Knudsen cell is therefore:
dN e
dt
 Ae 2mk BT 
1 / 2
P*  P  
Directional Dependence of a Knudsen Cell - 1
• First find the evaporant flux into a solid angle of d from a

source area of dAe:
– In a time dt, molecules that are within vdt of the orifice will exit.
– If molecules are uniformly distributed within the volume V, then
the fraction of molecules which are within striking distance in dt of
the orifice with an exit angle of  is vdt cos  dAe/V.
– The fraction of molecules entering the solid angle d is d/4.
The number of molecules emitted from an area dAe
 d
of the source in a time dt into a solid angle of d
about an exit angle of  with velocities in the range
vdt of v to v + dv is:
dAe v dt cos  dAe d
d 4 Ne  N  (v 2 ) dv
volume V V 4
• Integrate over all velocities:
– A Maxwellian distribution:
3/ 2
 m   mv 2

 (v )  4 
2
 v exp 
2

 2k BT   2 k BT 
 1/ 2
 8k BT 
1/ 2
4  2 k BT  4
0 v  v dv     vm   
2
( )
 m   m
• Obtain:
1/ 2
N  8k B T  d
d N e  
3
 cos  dt dAe
V m 4

• Since PV = NkBT, the impingement rate out through the
orifice into a solid angle of d about an exit angle of  is:
1/ 2
d Ne
3
N  8k BT  d d
 2mk BT  P cos 
1 / 2
   cos 
dAe dt V   m  4 
• The total impingement rate out through the orifice into all
forward solid angles is:
d 2 Ne
 2mk BT  P
1 / 2
dAe dt
2  / 2
Note that:  cos  d    cos  sin  d d  

0 0
• The mass evaporation rate through the orifice is
1/ 2
d 2 Ne  m 
m    P
dAe dt  2 k BT 
• The mass flux into a solid angle of d about an exit angle

of  is thus d
3
d M e  m d N e   cos 
3
dAe dt

t Ae
• The total mass of evaporated material is M e     dAe dt

0 0
• Cosine law of emission: (Effusion; Lambertian in optics)
d
dM e  M e cos 

• Normalized angular distribution function:
1
f ( ,  ) d d  cos  sin  d d

Directional Dependence of Condensation
• The area upon which the evaporant is condensed is
r 2 d
dAc 
cos
• The deposited evaporant mass per unit area of
condensation surface is thus:
dM c M e
dAc  cos  cos
 r
dAc  r 2
K-cell

Application of the Cosine Law - 1
• The cosine law was verified by Knudsen by depositing a
perfectly uniform coating inside a spherical glass jar:
r
cos   cos 
2ro

r
ro dM c Me
   uniform coating!
dAc 4 ro2
K-cell This geometry is commercially used for coating

the inside surfaces of spherical vessels, e.g. light
bulbs, as well as for planetary wafer tooling in
vacuum coating equipment.

Application of the Cosine Law - 2
• Uniformity of an evaporated film across a wafer:
substrate wafer dM c M e
 cos  cos
rw dAc  r 2

reduction factor for edge thickness:
2
ro re d edge  ro2 
ro2 1
 2 2 cos  cos   2 2  
d center ro  rw  ro  rw    r 2 
2
 1   w  
  ro  
Example:
For a 3-inch diameter wafer suspended 18 inches above a Knudsen cell:
K-cell dedge/dcenter = 0.986, or a non-uniformity of 1.4 %

Application of the Cosine Law – 3
• Computing the required source mass to create a given
thickness on the substrate:
– df = thickness of the film [cm]
– ρf = density of the film [g/cm3]
dM c M e
df f   cos  cos
dAc  r 2
• For normal incidence, φ = ψ = 0,
M e   r 2d f  f
• Example: To deposit df = 100 nm of Au (ρf = 19.32 g/cm3)

at normal incidence r = 40 cm away from the source, a
source charge of Me = 0.971 grams is needed.
Directional Dependence of a Free Point Source
• The evaporated mass from a point source is isotropic:
d d
d 3 M e   dAe dt dM e  M e
4 4
• For a condensation surface of dAc = r2d/cos, obtain:
dM c Me
 cos
dAc 4 r 2
• To put a uniform coating inside a spherical surface, a point

source must be located at the center of the sphere.

Thermodynamic Potentials - 1
• U = U(S,V,N1,N2,…,Nr) is the internal energy of a system
of particles.
– S = entropy
– V = volume
– N1, N2, … ,Nr are the number of particles in each of r components.
• For a monoatomic gas,
– Equation of state: PV = NRT
– Internal energy is U = 3/2 NRT
• U is a measure of the potential for work by the system.

U
 T  temperature
S V , N1 ... N r
U
  P  pressure
V S , N1 ... N r
U
  j  electrochemical potential
N j
S ,V , N i  N j
dU  TdS  PdV  1dN1     2 dN 2

• Helmholtz potential:
– F = F(T,V,N1,N2,…,Nr) = U - TS
– potential for work by system at constant temperature
• Enthalpy:
– H = H(S,P,N1,N2,…,Nr) = U + PV
– potential for work by system at constant pressure
– examples: heat engines, turbines, diesel motors
• Gibbs free energy:
– G = G(T,P,N1,N2,…,Nr) = U - TS + PV
– potential for work by system at constant temperature and pressure
– examples: chemical reactions, electrochemistry

Maxwell Relations
V T

S P P S dU  TdS  PdV    k dN k
k
S V
 dF   SdT  PdV    k dN k
P T T P k
P S dG   SdT  VdP    k dN k

T V V k
dH  TdS  VdP    k dN k
T
T P
 k
V S S V

Mnemonic Diagram
• Originally due to Max Born:
V F T
U G
S H P

van der Waals Equation of State
NRT aN 2
P  2
V  bN V
– {a,b} are van der Waals parameters.
– For large volumes and small particle numbers, van der Waals
equation of state is approximated by the ideal gas law: P = NRT/V.
P
isotherms
T4
T3
T2
T1
V
Phase Transitions - 1
– A system will only remain homogeneous if certain thermodynamic
stability criteria are met. If not, it will separate into two or more
portions.
P 100 % condensed phase

mixture of condensed and gas phases
100 % gas phase
c g
T = constant isotherm
V

• Similarly:
T
P = constant isobar
gas
liquid
solid
V

• Internal energy = U = U(S,V,N)
– dU = TdS - PdV + dN, in general.
– No material is created or destroyed during a phase transition: dN = 0.
• U = Ug - Uc = T(Sg - Sc) - P(Vg - Vc)
– U = Lgc - P(Vg - Vc),
– where Lgc = T(Sg - Sc) is the latent heat of the phase transition.
• For water:
– At 0°C the latent heat of fusion is 80 cal/g = 1440 cal/mole
– At 100°C the latent heat of vaporization is 540 cal/g = 9720 cal/mole

Equilibrium Vapor Pressure - 1
• Pressure of a gas in equilibrium with its condensed phase.
• Evaporation of the condensed phase converts thermal
energy into mechanical energy (expansion of the vapor).
• Process is less than 100 % efficient, because some thermal
energy must be used to increase the entropy of the system.
– (Second Law of Thermodynamics)

P 100 % condensed phase

mixture of condensed and gas phases
100 % gas phase
I
P* c g
II
T = constant isotherm
V
Vc Vg
P* is the pressure which makes areas I and II equal.

• Need to integrate along the isotherm to find S = Sg - Sc.

• In general, S S
dS  dV  dT
V T T V
• Temperature is constant during the phase change, so:
S P
dS  dV  dV
V T T V
– (Using the Maxwell relation)
• (P/T)V is independent of volume during the phase
transition, so P dP*

T dT
• P* is a function of only T.
• Integrating along the isotherm then produces
dP * S S g  S c
 
dT V Vg  Vc
• The latent heat of the phase transition is Lgc = T(Sg - Sc).
• Substituting gives the Clapeyron Equation:
dP * Lgc

dT T (Vg  Vc )
• For low vapor pressures, Vg - Vc  Vg  NRT/P, which
gives the Clausius-Clapeyron Equation:
dP* Lgc dT

P *
NRT 2
• If Lgc is constant with temperature, integrating the
Clausius-Clapeyron equation produces an Arrhenius
relationship for the equilibrium vapor pressure:
Lgc
ln P  
*
 C1
NRT
 Lgc 
P  C2 exp 
*

 NRT 
• Lgc/N is tabulated for many substances, but it does have a
temperature dependence which can alter the above
integration.

• The heat capacity of a material at constant pressure is:
H
cp   a  bT  cT 2  dT  2
T
– The {a,b,c,d} parameters are given in tables of thermochemical
data, e.g. the CRC Handbook of Chemistry and Physics.
• Integrating over temperature gives the temperature
dependence of the latent heat of the phase transition:
T
d
Lgc  Lgco   (c pg  c pc ) dT  Lgco  aT  bT  cT 
1
2
2 1
3
3
T0
T

• Integrating Lgc(T) in the Clausius-Clapeyron equation gives:
Lgco a bT cT 2 d
ln P  
*
 ln T     C3
NRT NR 2 NR 6 NR 2 NRT 2
– Both Lgco and C3 are unknown integration constants.

• Standard thermochemical data is specified at a “standard”
state of 1 atmosphere of pressure and a temperature of 25°C
or 298.15 K, the triple point of water.
– Need to rephrase the Clausius-Clapeyron equation to be able to use
this data.
– The standard state of temperature and pressure (STP) is usually
denoted by a “0” subscript or superscript.

• Standard state chemical potentials for the gas and
condensed phases are go and co.
• Equilibrium at a given temperature occurs by letting the
chemical potentials change until they equal each other at
the equilibrium vapor pressure of P*:
P*
  g    c 
*
P
 go (T )     dP   co (T )     dP
1 atm 
P T 1 atm 
P T
• The molar volume is given as shown, and for an ideal gas,
  g  Vg RT
   
 P T N P
• Since Vc is small, the integral for the condensed phase can
be ignored.
• Integrating gives the standard free energy of evaporation:
Geo (T )   go (T )   co (T )   RT ln P *
– P* is expressed in atmospheres.
– This now gives an expression for P* which eliminates the
unknown integration constants, but Geo is only listed for 298 K.
• Use the facts that:
 Geo   S eo  cp
G  H  TS ,    S eo (T ),    .
 T  P  T  P T
• Integrate twice to obtain:
T T
c p
Geo (T )  H eo (298 K )  TS eo (298 K )   
298 K 298 K
T
dTdT

• A working formulation of the equilibrium vapor pressure
from standard thermochemical data is therefore:
H ( 298 K ) S ( 298 K ) 1
T T
c p
ln P   eo
 eo
   dTdT
*
RT R RT 298 K 298 K T
– P* is in units of atmospheres.
– Heo(298K), Seo(298K), and cp(T) may be found in standard
tables of thermochemical data.
– Note: 1 kcal = 1 Cal = 1000 cal = 4186.8 J = 3.97 BTU
– See Section D of the CRC Handbook of Chemistry and Physics.
– Example: Aluminum:
• Heo(298K) = 70 kcal/mole, Seo(298K) = 30 kcal/mole-K, at the
boiling point of P = 1 atm and T = 2327°C = 2600 K.
• Hmo(298K) = 2.57 kcal/mole, at the melting point of P = 1 atm and T
= 658.5°C = 931.7 K.
1.00E+03
1.00E+02
1.00E+01
1.00E+00
Aluminum
1.00E-01 0 1000 2000 3000 4000 5000
Vapor Pressure, Torr
Gallium
1.00E-02
Indium
1.00E-03
Silicon
1.00E-04
Gold
1.00E-05
Platinum
1.00E-06
Chromium
1.00E-07 Titanium
1.00E-08
1.00E-09
1.00E-10
1.00E-11
Temperature, K

1.00E+03
1.00E+02
1.00E+01
1.00E+00
1.00E-01 0 5 10 15 20 Aluminum
Vapor Pressure, Torr
Gallium
1.00E-02
Indium
1.00E-03
Silicon
1.00E-04
Gold
1.00E-05
Platinum
1.00E-06
Chromium
1.00E-07 Titanium
1.00E-08
1.00E-09
1.00E-10
1.00E-11
Reciprocal Temperature, 10000/T(K)

Evaporation System Requirements
• Vacuum:
– Need 10-6 to 10-8 torr for medium quality films.
– Ultra-high quality films can be prepared in UHV at 10-9 to 10-11 torr.
• Cooling water:
– Hearth
– Thickness monitor
– Bell jar
• Mechanical shutter:
– Evaporation rate is set by temperature of source, but this cannot be
turned on and off rapidly. A mechanical shutter allows evaporant flux
to be rapidly modulated.
• Electrical power:
– Either high current or high voltage, typically 1-10 kW.
Evaporation Support Materials
• Refractory metals:
– Tungsten (W); MP = 3380°C, P* = 10-2 torr at 3230°C
– Tantalum (Ta); MP = 3000°C, P* = 10-2 torr at 3060°C
– Molybdenum (Mo); MP = 2620°C, P* = 10-2 torr at 2530°C
• Refractory ceramics:
– Graphitic Carbon (C); MP = 3700°C, P* = 10-2 torr at 2600°C
– Alumina (Al2O3); MP = 2030°C, P* = 10-2 torr at 1900°C
– Boron nitride (BN); MP = 2500°C, P* = 10-2 torr at 1600°C
• Engineering considerations:
– Thermal conductivity
– Thermal expansion
– Electrical conductivity
– Wettability and reactivity
Resistance Heated Evaporation
• Simple, robust, and in widespread use.
• Can only achieve temperatures of about 1800°C.
• Use W, Ta, or Mo filaments to heat evaporation source.
• Typical filament currents are 200-300 Amperes.
• Exposes substrates to visible and IR radiation.
• Typical deposition rates are 1-20 Angstroms/second.
• Common evaporant materials:
– Au, Ag, Al, Sn, Cr, Cu, Sb, Ge, In, Mg, Ga
– CdS, PbS, CdSe, NaCl, KCl, AgCl, MgF2, CaF2, PbCl2

Resistance Heated Evaporation Sources
foil dimple boat
wire hairpin
alumina coated foil dimple boat
wire helix
foil trough
chromium coated tungsten rod

wire basket
alumina crucible with wire basket alumina crucible in tantalum box

Characteristics of Specific Resistance Evaporator Sources
• Principal considerations:
– Required temperature to achieve P* ~ 10-2 torr, TP*-2.
• About 1800C is normally maximum for most resistance heated
systems.
– Wetting of support material
– Alloying with support material
– Sublimation versus evaporation from a liquid state
• Melting point temperature, Tmp versus TP*-2.
– Decomposition of non-elemental sources
– Expansion or contraction upon freezing
– Tendency to splatter
– Reactivity with oxygen
– Toxicity
– Effects of chamber wall deposits
Aluminum (Al)
• Most commonly used for silicon IC interconnects.
• Tmp = 660C; TP*-2 = 1220C.
• Wets most support materials and crawls out of its containers.
– Much of the initial charge can go only towards wetting the container.
– Wire filaments can become shorted out by Al crawling along their
length.
• BN crucibles with a W wire heater on the outside work well.
• e-Beam evaporation is very effective, but can cause
splattering at high rates.

Gold (Au)
• Most commonly used for compound semiconductor
interconnects.
• Tmp = 1063C; TP*-2 = 1400C.
• Wets W and Mo supports.
• Alloys with Ta boats, usually resulting in their breakage.
• Usually a well-behaved evaporant.
• Capable of creating very thin films with high uniformity.
• Substrate heating often used to create single crystal gold
films on mica substrates for AFM use.

Silver (Ag)
• Mostly used for chemical sensor electrodes; too reactive
for most interconnect purposes.
• Tmp = 961C; TP*-2 = 1030C.
• Wets Ta and Mo supports, but not W.
• Generally a well-behaved evaporant.

Tin (Sn)
• Assorted uses in chemical sensors; less commonly used as
an interconnect metal because of its low melting point.
• Tmp = 232C; TP*-2 = 1250C.
• Alloys with and destroys Mo supports.
• Works well with Ta and W supports.
• Works well in Al2O3 lined boats and crucibles.
• Generally a well-behaved evaporant.

Chromium (Cr)
• Most commonly used as an adhesion film; too resistive for
most interconnects.
• Tmp = 1900C; TP*-2 = 1400C; sublimes.
• Evaporation from Cr-coated W rods is usually the easiest,
but the amount that can be deposited from one rod is
limited.
• Can be used in Ta or W boats, but thermal contact to boat
from Cr shot or pellets may be a problem because the
charge will not melt.

Copper (Cu)
• Creates high conductivity films; useful for interconnects
where it will not interfere with the semiconductor.
• Tmp = 1084C; TP*-2 = 1260C.
• Works well with Ta, W, and Mo support materials.
• Works well in graphite, Mo, and Al2O3 crucibles.
• Generally, a well-behaved evaporant.

Electron Beam Heated Evaporation - 1
• More complex, but extremely versatile.
• Can achieve temperatures in excess of 3000°C.
• Uses evaporation cones or crucibles in a copper hearth.
• Typical acceleration voltage is 8-10 kV.
• Exposes substrates to secondary electron radiation.
– X-rays can also be generated by high voltage electron beam.
• Typical deposition rates are 10-100 Angstroms/second.
• Common evaporant materials:
– Everything a resistance heated evaporator will accommodate, plus:
– Ni, Pt, Ir, Rh, Ti, V, Zr, W, Ta, Mo (all noble and refractory metals)
– Al2O3, SiO, SiO2, SnO2, TiO2, ZrO2 (most metal oxides)

Electron Beam Heated Evaporation Source
270 degree bent

electron beam
magnetic
field
evaporation cones
pyrolytic graphite of material
hearth liner
4-pocket rotary
copper hearth
(0 V)
beam forming
cathode aperture
filament
recirculating (-10,000 V)
cooling water

Electron Beam Heated Evaporation - 2
• 270° bent beam electron gun is most preferred:
– Filament is out of direct exposure from evaporant flux.
– Magnetic field can be used for beam focusing.
– Magnetic field can be used for beam positioning.
– Additional lateral magnetic field can be used produce X-Y sweep.
• Sweeping or rastering of the evaporant source is useful for:
– Allows a larger evaporant surface area for higher deposition rates.
– Allows initial charge to be “soaked” or preheated.
– Allows evaporant source to be more fully utilized.
• Multiple pocket rotary hearth is also preferred:
– Allows sequential deposition of layers with a single pump-down.
– Allows larger evaporation sources to be used.

Characteristics of Specific e-Beam Evaporator Sources
• Principal considerations:
– Thermal conductivity
– Sublimation versus evaporation from a liquid state
– Decomposition of non-elemental sources
– Expansion or contraction upon freezing
– Soft x-ray emission threshold
– Tendency to splatter
– Reactivity with oxygen
– Reactivity with copper, i.e. the hearth
– Toxicity
– Effects of chamber wall deposits
– Need for a hearth liner, e.g. C or BN
– Magnetic elements (Co, Ni, Mn) can create some problems.
Titanium (Ti)
• Useful as an interconnect metal and as an adhesion film.
• Tmp = 1700C; TP*-2 = 1750C; usually sublimes.
• Evaporation cones that directly drop into the hearth
pockets are most common.
• Ti coated surfaces are extremely reactive!
– Surfaces become a getter pump for the chamber – chamber
pressure drops during evaporation due to this!
– Surfaces are extremely reactive with oxygen – chamber must be
backfilled with dry N2 to avoid combustion until fresh surfaces are
cool.

Tungsten (W)
• Useful as an interconnect metal; has the useful property of
precisely matched thermal expansion to that of silicon.
• Tmp = 3380C; TP*-2 = 3230C; usually sublimes.
• Usually requires a lot of heat, but easily accomplished in
an e-beam evaporator.

Silicon (Si)
• Amorphous silicon has several applications for both
devices and sensors. Polycrystalline silicon can be
produced with additional substrate heating and annealing,
dependent upon the substrate.
• Tmp = 1420C; TP*-2 = 1350C; usually sublimes, but
puddling is also common.
• Most commonly done using a graphite crucible.
• Si expands upon freezing and can crack crucibles.
• Si sources tend to spatter unless evaporation rate is kept
low.

Silicon Dioxide (SiO2)
• One simple method to deposit glass insulating layers, but
not to any great thickness.
• Tmp = 1730C; TP*-2 = 1250C.
• Reacts with and destroys Ta, W, and Mo supports.
• Usually used in a graphite hearth liner for e-beam
evaporation.
• Usually decomposes into SiO and O2 unless evaporated at
a significant rate.

High Throughput Evaporation Techniques
• Box coaters are used for evaporating large substrate
materials, sometimes up to several meters in size.
• Large amounts of source material are required, but cannot
be all heated at once because of realistic power limitations.
• Two popular techniques:
– Powder trickler source
– Wire feed source
• Both can be adapted for either resistance heated or electron
beam heated evaporation systems.

Problematic Source Materials
• Arsenic (As) – highly toxic, spoils vacuum chamber.
• Cadmium (Cd) – highly toxic, spoils vacuum chamber.
• Selenium (Se) – highly toxic, spoils vacuum chamber.
• Tellurium (Te) – highly toxic, spoils vacuum chamber.
• Zinc (Zn) – wall deposits spoil vacuum chamber.
• Bismuth (Bi) – toxic.
• Lead (Pb) – toxic.
• Magnesium (Mg) – can be done, but extremely reactive
and prone to combustion upon exposure to air.

Adsorption
• Adsorption is the sticking of a particle to a surface.
– Absorption (note the spelling difference!) is the collection of a
external substance within the interior of a media.
• Physisorption:
– The impinging molecule loses kinetic (thermal) energy within
some residence time, and the lower energy of the molecule does
not allow it to overcome the threshold that is needed to escape.
• Chemisorption:
– The impinging molecule loses its kinetic energy to a chemical
reaction which forms a chemical bond between it and other
substrate atoms.

Condensation of Evaporant - 1
• Condensation of a vapor to a solid or liquid occurs when
the partial pressure of the vapor exceeds the equilibrium
vapor pressure of the condensed phase at this temperature.
– Pvapor > P*(Tsubstrate)
• The vapor is “supersaturated” under these conditions.
• This is only true if condensation takes place onto a
material which is of the same composition as the vapor.
• When a material is first deposited onto a substrate of a
different composition, a third adsorbed phase must be
included to describe the process.

• Molecules impinging upon a surface may:
– Adsorb and permanently stick where they land (rare!).
– Adsorb and permanently stick after diffusing around on the surface
to find an appropriate site.
• This can lead to physisorption or chemisorption.
– Adsorb and then desorb after some residence time a.
– Immediately reflect off of the surface.
• Incident vapor molecules normally have a kinetic energy
much higher than kBT of the substrate surface.
• Whether an atom or molecule will stick depends upon how
well it can equilibrate with the substrate surface,
decreasing its energy to the point where it will not
subsequently desorb.
• Thermal accommodation coefficient:
Ev  Er Tv  Tr
T  
Ev  Es Tv  Ts
– Ev, Tv = energy, temperature of impinging vapor molecules.
– Er, Tr = energy, temperature of resident vapor molecules;
• (Those which have adsorbed, but have not permanently found a site.)
– Es, Ts = energy, temperature of substrate surface.
• If T < 1, (Er > Es), then some fraction of the impinging
molecules will desorb from the surface.

• Mean residence time for an adsorbed molecule:
1  Gdes 
 a  exp 
0  k BT 
– 0 = kBT/h = vibrational frequency of adsorbed molecule (~1014 Hz)
• This is the frequency at which the molecule “attempts” to desorb.
– Gdes = free activation energy for desorption.
• Under a constant impinging vapor flux of R, the surface
density of the deposit is then:
R  Gdes 
ns  R a  exp 
0  k BT 
– R = deposition rate in molecules/cm2-sec.
– ns = surface density of deposited molecules in cm-2.
• If the impingement rate stops, then the adsorbed molecules
will all eventually desorb.
• Condensation of a permanent deposit will not occur, even
for low substrate temperatures, unless the molecules
interact.
• Within the mean residence time, surface migration occurs
and clusters form.
• Clusters have smaller surface-to-volume ratios, and
therefore desorb at a reduced rate.
• Nucleation of a permanent deposit is therefore dependent
upon clustering of the adsorbed molecules.

Observed Growth Features of a Deposited Film
• Adsorbed monomers
• Subcritical embryos of various sizes
• Formation of critically sized nuclei
• Growth of nuclei to supercritical size and depletion of
monomers within their capture zones
• Nucleation of critical clusters within non-depleted areas
• Clusters touch and coalesce into new islands, exposing
fresh substrate areas
• Adsorption of monomers onto fresh areas
• Larger islands grow together leaving holes and channels
• Channels and holes fill to form a continuous film
Stages of Thin Film Growth
• Island Stage
• Coalescence Stage
• Channel Stage
• Continuous Film Stage

Modes of Thin Film Growth
(1) Volmer-Weber: (island growth):
M. Volmer and A. Weber, Z. Phys. Chem. 119, p. 277 (1926).
(2) Frank-Van der Merwe: (layer growth; ideal epitaxy):
F. C. Frank and J. H. Van der Merwe, Proc. R. Soc. London, Ser. A 198, p. 205 (1949).
(3) Stranski-Krastanov: (layers + islands):
J. N. Stranski and L. Krastanov, Ber. Akad. Wiss. Wien 146, p. 797 (1938).

Condensation Control
• Condensation control of the evaporant is achieved through
control of the substrate temperature Ts.
• Higher substrate temperatures:
– Increase thermal energy of adsorbed molecules.
• (Shortens the residence time.)
– Increase surface diffusivity of adsorbed molecules.
– Performs annealing of deposited film.
• Substrate heaters:
– Quartz IR lamps from frontside or backside
– Ta, W, or Mo foil heaters from backside
– Graphite impregnated cloth heaters from backside
• Too much heat will desorb the deposited film, evaporating
it away! (But this can be used for cleaning…)
Deposition End Point Control
• Evaporate the source material to completion
– Compute the required mass of source material to provide the
desired coating thickness using Lambertian source characteristics.
– Only works for sources that do not wet their supporting materials.
• Quartz crystal thickness monitors (QCMs)
– Based on a quartz crystal oscillator where the quartz crystal is
exposed to the evaporant flux.
– An increase in mass on the surface of the crystal will shift the
oscillation frequency in a predictable manner, from which the
deposited thickness can be determined from the change in the
oscillation frequency.
• Optical interference, absorption, or reflectivity
– A more sensitive technique better suited for optical coatings.

Quartz Crystal Microbalances (QCMs) - 1
• Quartz is piezoelectric; application of a voltage will create
an elastic deformation of the crystal; a sinusoidal voltage
will create a transverse acoustic wave.
• The acoustic wave will set up a resonance when the
thickness of the crystal is half of the acoustic wavelength.
• This will create an electrical series resonant equivalent
circuit which can be used to pull an oscillator circuit onto
this frequency.
• Adding mass to the surface of the quartz crystal is
equivalent to increasing the effective thickness of the
crystal, which will change the resonant frequency.

• Resonance condition: Gq
2 f 0 t q  va 
q
– fo = resonant frequency, usually 5 or 6 MHz.
– tq = thickness of quartz crystal.
– va = acoustic velocity for a shear wave = 3336 m/s.
– Gq = shear modulus of quartz = 2.947 x 1011 g/cm-sec2.
– q = density of quartz = 2.648 g/cm2.
– For fo = 6 MHz, tq = 275 m.
• Sauerbrey equation: 2 f 02 m
f  
Zq A
– Zq = (qGq)1/2 = acoustic impedance of quartz = 8.834 x 105 g/cm2-sec.
– m = deposited mass.
– A = piezoelectric active area of quartz (electrode area).
– Valid in the limit of small added mass (thin deposits).
• The deposited film thickness can be determined from the
added mass: m
tf 
Af  f
– Af = area of deposited film.
• For thicker films, acoustic reflections between the quartz
and the deposited film must be taken into account:
m Z quartz   fU  f L  
 tan  Z tan 
1

A 2Zf L   fU  
– “Z-match method”; reduces to the Sauerbrey eqn. for Z = 1.
– fU = unloaded (upper) frequency; fL = loaded (lower) frequency.
– “Z-ratio” = acoustic impedance ratio of quartz to that of the film:
Z quartz  q Gq
Z 
Z film  f Gf R. B. Darling / EE-527 / Winter 2013
• Temperature effects:
– Quartz crystal oscillators are also sensitive to temperature.
– Temperature dependence is minimized by using AT-cut quartz.
– Water cooling of the QCM oscillator head is also used to stabilize T.
• Chamber geometry effects:
– Deposited thickness on QCM is usually not the same as that deposited
onto the substrates.
– A multiplicative tooling factor (TF) is used to correct for this.
– The TF depends purely upon the relative location of the QCM to that of
the wafers; once calibrated, it does not change.
• Material data:
– The density (f) and Z-ratio (Z) are both needed to accurately calibrate
the QCM for a given deposited film.
– Usually tabulated for most common films.
Material Density (g/cm3) Z-Ratio
Aluminum (Al) 2.70 1.080
Chromium (Cr) 7.20 0.305
Copper (Cu) 8.93 0.437
Germanium (Ge) 5.35 0.516
Gold (Au) 19.30 0.381
Nickel (Ni) 8.91 0.331
Platinum (Pt) 21.40 0.245
Silicon (Si) 2.32 0.712
Silver (Ag) 10.50 0.529
Tin (Sn) 7.30 0.724
Titanium (Ti) 4.50 0.628

Physical Vapor Deposition

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Physical Vapor Deposition

Uploaded by

Copyright:

Available Formats

EE-527: MicroFabrication

Physical Vapor Deposition

R. B. Darling / EE-527 / Winter 2013

Gas Phase Gas Phase

Condensed Phase Condensed Phase

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

The net evaporation flux is the difference

R. B. Darling / EE-527 / Winter 2013

•  = mass evaporation rate in g/cm2-sec:

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

• In 1909 Knudsen invented a technique to force v to 1.

Surface area of the evaporant inside the enclosure

An equilibrium vapor pressure P* is maintained

Effusion from a Knudsen cell is therefore:

• First find the evaporant flux into a solid angle of d from a

R. B. Darling / EE-527 / Winter 2013

Note that:  cos  d    cos  sin  d d  

• The mass flux into a solid angle of d about an exit angle

• The total mass of evaporated material is M e     dAe dt

R. B. Darling / EE-527 / Winter 2013

K-cell This geometry is commercially used for coating

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

• For normal incidence, φ = ψ = 0,

• Example: To deposit df = 100 nm of Au (ρf = 19.32 g/cm3)

• To put a uniform coating inside a spherical surface, a point

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

dU  TdS  PdV  1dN1     2 dN 2

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

P 100 % condensed phase

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

P 100 % condensed phase

P* is the pressure which makes areas I and II equal.

R. B. Darling / EE-527 / Winter 2013

• Need to integrate along the isotherm to find S = Sg - Sc.

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

– Both Lgco and C3 are unknown integration constants.

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

foil dimple boat

alumina coated foil dimple boat

chromium coated tungsten rod

alumina crucible with wire basket alumina crucible in tantalum box

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013

R. B. Darling / EE-527 / Winter 2013