Basic Mechanisms in Plasma Etching
Vincent M. Donnelly
University of Houston
1
2
Outline
• Brief history of plasma etching, anisotropic vs. isotropic etching, and mechanisms for 
anisotropy
• Etching product volatility
• Adsorption and reactant sticking coefficients
• Chlorination of Si during chlorine plasma etching
• Chemical etching of Si by F atoms 
• Unwanted photo‐assisted etching of Si
History of plasmas and plasma etching
• In the early days of integrated‐circuit processing, wet‐etching was used for 
pattern transfer. 
• In late 1960s – early 1970s, plasma‐based pattern transfer replaced wet etching 
for two reasons:
1) anisotropic etching for finer feature pattern transfer
2) eliminate wet chemical waste disposal.
• Early sputtering done with Ar plasmas.
• O2 plasmas used to strip photoresist.
• Fluorine and chlorine‐containing plasmas used to etch Si, SiO2 and Al.
• Use of additives (e.g. H2) or alternative gases to CF4 (C2F6, C3F8, CHF3) to 
increase selectivity of etching of SiO2 with respect to Si (Heinecke).
3
EARLY 1960’s BARREL PLASMA ETCHING
• Early plasma reactors were barrel type.
• Wafers were placed in a quartz chamber with external electrodes or coil.
• Adequate for resist stripping. Lacked wafer temperature control. Suffered from 
etch nonuniformity. 
• Low energy ions and high pressure led to isotropic etching. 
4
RADIAL FLOW REACTOR (“REINBERG REACTOR”)
• Inward (radial) flow was thought to compensate for gas consumption.
• Wafers on the grounded electrode; RF applied to opposite electrode.
• Some anisotropy, but still not adequate.
5
MIGRATION OF WAFERS TO THE SMALLER ELECTRODE
low-energy ion high-energy ion high-energy ion
bombardment bombardment bombardment
• Motivated by the (incorrect) theory that voltage is inversely proportional to the 
electrode area ratio to the 4th power. 
• Actually it scales more like the inverse area ratio to the 1st power. Nonetheless, 
smaller electrode receives high‐energy ion bombardment. 6
THE NEED FOR ANISOTROPIC PLASMA ETCHING
7
Some Basic Considerations in Ion-Assisted Etching
Ion‐Neutral Synergy and Ion‐Assisted Etching
• Etching “reactions”:
Feed gas + plasma  neutral radicals, (+) ions
radicals + (+) ions + chamber surface  other products
neutrals (inc. radicals) + (+) ions + wafer  volatile products
• Characteristics of anisotropic plasma etching:
- Perpendicular bombardment of the surface by + ions that are
accelerated by a sheath potential
- Anisotropic etching relies on a synergy between ion and neutral
reactions on the surface (“the whole is greater than the sum of its
parts”). First explained in an experiment by Coburn and Winters at
IBM Research Laboratories in San Jose, CA. They used ion and
molecular beams to simulate a plasma:
8
9
Some Basic Considerations in Ion-Assisted Etching
Ion‐Neutral Synergy and Ion‐Assisted Etching (cont.)
• Coburn and Winters’ famous ion and molecular beams experiment:
10
Etching Mechanism 1:
substrate
or film
substrate
or film
substrate
or film
+
sputtered
product
Physical sputtering (slow) Chemical etching (slow)
+
Chemical sputtering (fast)
Chemical Sputtering Mechanism for Anisotropic Etching
+ = positive ion = neutral chemisorbed layer
• Good example of chemical sputtering: Si etching in a Cl2 plasma.
10
11
Etching Mechanism 2: Sidewall Inhibitor
• Isotropic chemical etching is sometimes very fast. This process can be slowed or
stopped by adding an “inhibitor” to the plasma gas.
• Inhibitor coats vertical sidewall but is sputtered away on horizontal surfaces.
• Good examples of sidewall inhibitor: 1) etching of photoresist‐masked Al in a Cl2/BCl3
plasma; 2) Bosch process for MEMS etching in SF6 and C4F8 pulsed gas plasmas.
• Plasma etching is the gasification of a solid by reactive species that
are formed in the plasma.
• For example, fluorine and chlorine atoms are known to convert solid
silicon into gaseous SiF4 and SiCl4. The overall reactions are:
4F(g) + Si(s)  SiF4(g)
4Cl(g) + Si(s)  SiCl4(g)
• Ion bombardment also accelerates these reactions.
THE IMPORTANCE OF VOLATILE PRODUCTS
12
Solid elements to be etched                        gaseous etchant elements
13
Volatile the etching products in halogen, carbon, hydrogen, and oxygencontaining
plasmas are SiF4, SiF2, SiCl4, SiCl2, SiBr4, SiBrxHy, SiClxBryHz,
SiOF2, CO, CO2, O2, COF2, metal halides and metal oxy-halides.
• For a compound to be sufficiently volatile, its evaporation rate should
be much higher than the desired etching rate.
• The maximum evaporation rate is computed by the principle of detailed
balance: at equilibrium, the forward and reverse rates of every elementary
process are equal.
• Consequently, for a gaseous species at a number density, n, in
equilibrium with its liquid or solid state in a closed system, its evaporation
rate equals its impingement rate on the solid or liquid.
• The impingement rate or flux (molecules-cm2s-1) of species onto a
surface is
4
nv
fi 
Some Basic Considerations in Etching
Vaporization of Products
14
• The thermal speed, v (in cm/s), is given by
where k is the Boltzman constant (8.314 x 107 erg K-1mole-1), T is the
temperature in Kelvin, and m is the mass in grams/mole.
• For an ideal gas at pressure p (in dyne-cm-2 [1 dyne-cm-2 = 0.1 Pa = 7.502 x
10-5 Torr]), the impingement rate is
• The equilibrium vapor pressure, pV, is given by the Clausias-Claperon
equation:
Some Basic Considerations in Etching
Vaporization of Products (cont.)
2/1
8







m
kT
v

mkT
p
fi
2






 

RT
H
ppV exp0
15
where H is the heat of vaporization and p0 is a constant of integration.
Consequently, at equilibrium, the evaporation rate (equals the impingement
rate) is given by
• NOTE: The evaporation rate is given by the right side of this equation,
regardless if the system is closed or open and far from equilibrium, as it
always in in plasma etching.
Some Basic Considerations in Etching
Vaporization of Products (cont.)
mkT
p
ff V
ie
2

16
2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2
1
10
100
CBr4
CCl4
150 100 50 -2525 0
SiBr4
SiCl4
T(
o
C)
VAPORPRESSURE(TORR)
1000/T(K)
Some Basic Considerations in Etching
Vaporization of Products (cont.)
• Equilibrium vapor pressures
of Si and C halides.
• (SiF4 and CF4 are gases).
• Note the much lower vapor
pressures of the Br
compounds relative to Cl (and
F) products.
• Does this explain slower
etching rates for Br? No.
17
2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2
1
10
100
CBr4
CCl4
150 100 50 -2525 0
SiBr4
SiCl4
T(
o
C)
VAPORPRESSURE(TORR)
1000/T(K)
Some Basic Considerations in Etching
Vaporization of Products (cont.)
Etching rates (Å/min) if
limited by product
vaporization rate.
•  Si and C etching in Cl
or Br-containing plasmas
not limited by vaporization
of tetrahalide compounds.
~109
~108
~107
18
Some Basic Considerations in Etching
Product Desorption Rates
• For a species to be present on the surface during etching, it must be
strongly adsorbed. The rate of thermal desorption (s-1) is given by
• 0 is the pre-exponential or so-called attempt frequency and Ea is the
activation energy for desorption (i.e. binding energy) of the adsorbate.
• 0 is often set equal to a typical vibrational frequency of ~1013s-1, but in
fact 0 = kT/h(qts/qr), where qts and qr are partition functions of the
transition state for desorption and the reactant state.
• 0 can vary from typically 108 to 1015 s-1.
• Ea can also span a wide range of values, reflecting the complex nature
of the surface layer and the multitude of bonding configurations.
)/exp(0 RTEk ad  
19
Some Basic Considerations in Etching
Product Desorption Rates (cont.)
• Using  0 = 1013s-1 and above eq. for kd, it follows that for a species to
have a lifetime (1/kd) ~0.1s on the surface, comparable to the time
required to etch 1 monolayer, must have a binding energy, Ea = 16
kcal/mole (0.69 eV) at room temperature.
• Ea = 16 kcal/mole exceeds physisorption energies for most adsorbatesubstrate
combinations. Products like SiBr4 should be much more weakly
bound and therefore desorb nearly instantaneously after being formed.
• Ea = 16 kcal/mole is less than most chemical bonds, consequently any
chemisorbed species will likely live indefinitely on the surface.
• Therefore, surfaces are covered with chemisorbed species during and
after etching.
20
Some Basic Considerations in Etching
Adsorption and Etching by Neutrals
Unoccupied site Occupied site
SURFACE SITES IN L-H MODEL
21
• A sticking coefficient can be defined
by the Langmuir-Hinshelwood (L-H)
adsorption model.
•  = relative density of adsorption sites
(0 <  < 1). The probability for
adsorption is
where S is the sticking coefficient (or
sticking probability) at an unoccupied
site.
• When all sites are occupied, the
probability for adsorption is zero. Most etching processes operate near this
limit.
• On atomically rough surfaces that are present during etching, the adsorption
and diffusion processes are more complicated, with a range of differing
adsorption sites and rates.
)1(  Skads
22
Cl Uptake on Si (100) in a Cl2 plasma
0 10 20 30 40 50 60 70 80
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
P=2.1 mTorr Cl2
ClCOVERAGE(X10
15
)
P=0.6 mTorr Cl2
SiClLD-LIFINTENSITY(ARB.UNITS)
LASER REPETITION RATE (Hz)
excimer laser pulse
Laser‐induced heating
SiCl desorption
SiCl excitation (by laser,
close to the substrate)
Plasma
SiClx
layerSi
1)
2)
3)
• A simple Langmuir‐Hinshelwood (L‐H) model was fit to the measurements.
• The model reproduces that trends at both pressures with a single sticking 
coefficient times flux, where the chlorinating species could be Cl, Cl2, Cl+ and 
Cl2
+.  The chlorination rate constant is 8 x 104s‐1Torr‐1 for a plasma with ~10% 
dissociation of Cl2 and an ion density of ~5 x 1010cm‐3.
23
97 98 99 100 101 102 103 104 105
0
10
20
30
SiCl4
(none)
SiCl3
SiCl2
SiCl
Si
Si·
10.0 sccm Cl2
 = 30°
Intensity(counts)
Binding Energy (eV)
Cl Uptake on Si (100) in a Cl2 plasma
0 5 10 15 20 25 30 35 40 45 50
0.00
0.05
0.10
0.15
0.20
0.25
Si
Si·
SiCl3
SiCl2
SiCl
RelativeConcentration
Depth (Å)• SiClx ~ 16 Å
• SiCl>SiCl2>SiCl3>SiCl3O depth and concentration
• Si∙deep into layer
24
0 2 4 6 8 10 12 14 16 18
1.0
1.5
2.0
2.5
3.0
3.5
Si(100)
HR plasma, XPS
HR plasma, LD-LIF
ICP plasma, LD-LIF
ClCOVERAGE(10
15
cm
-2
)
E
1/2
(eV
1/2
)
Cl Uptake on Si (100) in a Cl2 plasma: Dependence on Ion Energy
25
Cl Uptake on Si (100) in a Cl2 plasma: Dependence on Ion Energy
Some Basic Considerations in Etching
Adsorption and Etching by Neutrals (cont.)
• Some materials (e.g. Si) will be etched by neutrals (e.g. F-atoms) in the
absence of ion bombardment. This reaction is mainly responsible for
deep Si etching in SF6 plasmas.
• Define a reaction coefficient for etching, X(S), as the probability that an
atomic or molecular material, S, will be converted into a volatile products per
impinging neutral, X (in the absence of ion bombardment or other sources of
energetic particles),
RX(S) = etching rate, NA = Avogadro’s number, rS and MS are density and
mass of the substrate. (e.g. if 1000 impinging F atoms per area per time
leads to 1 Si atom being etched, X(S) = 0.001.
• If X(S) << S (S = sticking coefficient defined 2 slides back), then   1.
)4/(S
)S(S
)S(
XX
XA
X
vnM
RN 
 
26
27
One Way to Measure X(S) : Downstream Reactor
D.L. Flamm, V.M. Donnelly, and J.A. Mucha, The reaction of fluorine
atoms with silicon. J. Appl. Phys, 1981. 52: p. 3633.
• Eliminate ion bombardment
• Measure absolute F number density by chemical titration.
28
Summary of Published Reaction 
Probabilities for F +Si
• Reported values (when 
corrected) range from 
0.00042 to 0.1!
• How to explain this 240‐fold 
discrepancy?
29
10
11
10
12
10
13
10
14
10
15
10
16
10
17
10
18
10
19
10
20
10
-3
10
-2
10
-1
ReactionProbability(F(Si)
)
F Flux (cm
-2
s
-1
)
F + Si Reaction Probability Exhibits Strong Flux Dependence
30
10
11
10
12
10
13
10
14
10
15
10
16
10
17
10
18
10
19
10
20
10
-3
10
-2
10
-1
ReactionProbability(F(Si)
)
F Flux (cm
-2
s
-1
)
In SF6
plasma
F + Si Reaction Probability Much Higher in SF6 Plasmas
Motivation : Precise Control of IED for ALE
(I) Low energy ions : No reaction occurs
except for pure chemical etching, if any.
(II) High energy ions: chemical
sputtering, sub-surface ion penetration,
physical sputtering  not well-controlled
etching, substrate damage.
(III) Medium energy ions: Only chemical
sputtering. More confined to surface.
Etching slows or stops when adsorbed
etchant reacts away.  minimum damage,
layer-by-layer etching.
Physical 
Sputtering Eth
Chemical 
Sputtering Eth
• Layer‐by‐layer etching, high selectivity, and no substrate damage are/will be 
critical requirements. To achieve these, one needs to control the IED.
• Accurate control of peak ion energy.
• Narrow width of IED.
• Work with medium energy ions (region III).
(I) (II)(III)
31
Ion Energy Dependence of Si Etching in
Chlorine-Containing Plasmas
32
“Kinetic study of low energy argon ion‐enhanced plasma etching 
of polysilicon with atomic/molecular chlorine”,
J. P. Chang, J. C. Arnold, G. C. H. Zau, H‐S. Shin, and H. H. Sawin, J. 
Vac. Sci. Technol. A, 15, 1853 (1997).
Yield (Si atoms etched per ion)
Sqrt. ion energy (eV1/2)
Characteristics of polycrystalline Si etching with Cl
and Ar+ beams in high vacuum:
• Etching exhibits a threshold ion energy of 16 eV.
• Above this threshold, etching rate increases with the 
square root of ion energy.
Poly‐Si
Eth  16 eV
0 1 2 3 4 5 6 7 8
ER(Å/min)@50mTorr
E
1/2
(eV
1/2
)
0
100
200
300
400
500
Eth~16 eV
What We Expect in a Pulsed Plasma for Si 
Etching
33
0 2 4 6 8 101214161820222426283032
0.00
0.01
0.02
0.03
0.04
0.05
Eth~16 eV
Time-AveragedIonEnergyDistribution
Energy (eV)
50 mTorr
etching
no etching
0 10 20 30 40
0.00
0.01
0.02
0.03
0.04
0.05 Eth~16 eV
Time-AveragedIonEnergyDistribution
Energy (eV)
50 mTorr
• Ar/Cl2 (few %) pulsed plasma.
• 20 s ON 80 s OFF.
• Synchronous DC bias in the
afterglow at 70‐97 s.
• 1015 /cm3 p‐type Si
• Above threshold etching looks like what we 
expect.
• Big question (and big problem): What causes 
sub‐threshold etching?
34
What Causes Sub‐threshold Etching?
• Spontaneous chemical etching by Cl atoms?
‐ It is widely reported that this does not happen for p‐type or i‐Si
(only heavily doped n‐type).
SiO2 mask
Si
‐ This is confirmed in our cross section SEMs: no undercutting of the mask.
• Therefore spontaneous chemical etching by Cl atoms is not the reason.
35
Grids / Substrate Bias Experiments
1
-30 -20 10 20 30
-4
Current (mA)
-2
Voltage (V)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
RelativeEtchingRate
energetic positive ions
Cl, Cl2, photons
positive ions
Cl, Cl2, photons
V
Si
Cl2/Ar Plasma
‐5V
0.1 L/0.1S 0V‐20V
Conclusion: Sub‐threshold etching is 
due to photo‐assisted etching induced 
by plasma light.
36
Summary
• After nearly 50 Years, plasma etching is well established yet 
challenges persist.
• As processes evolve to atomic layer etching, ion‐induced lattice 
damage and mixing of etchant and substrate atoms present major 
problems, requiring controlled fluxes of low energy ions.
• Photo‐assisted reactions also complicate etching with low energy 
ions.
37
More details can be found in:
• “Plasma Etching: Yesterday, Today, and Tomorrow”, V. M. Donnelly and A. Kornblit, 
J. Vac. Sci. Technol. A, 31, 050825‐1 (2103).
• “Critical review: Plasma‐surface reactions and the spinning wall method”, V. M. 
Donnelly, J. Guha, and L. Stafford, J. Vac. Sci. Technol. A, 29, 010801‐1 (2011).
• “The surprising Importance of Photo‐assisted Etching in Chlorine‐Containing 
Plasmas”, H. Shin, W. Zhu, C. M. Donnelly and D. J. Economou, J. Vac. Sci. Technol. 
A, 30, 201306 (2012).
• “Review: Reactions of fluorine atoms with silicon, revisited, again”, V. M. Donnelly, 
J. Vac. Sci. Technol. A, 35, 05C202‐1 (2017).