Plasma Diagnostics

1
Plasma Diagnostics

• Monday Afternoon Tutorial for UC-DISCOVERY Major
  Program Award on
• Feature Level Compensation and Control
• Eray S. Aydil
• Chemical Engineering Department
• University of California Santa Barbara
• 12/01/2003

2
Central Problem in Plasma Etching

• To understand how externally controlled variables
  affect the process outcome through the internal
  plasma parameters.

Internal plasma parameters
ion flux, J radical fluxes, Gi ion energy, E

• Plasma diagnostics are experimental methods based
  on various electrical and spectroscopic
  techniques that allow the measurement of
  internal plasma parameters.

3
Ion and etchant fluxes impinging on the wafer
surface determines the etch rates and profile
evolution in plasma etching processes.

• ER f (J, E, Gi, T)
• Example SF6/O2 etching of Si ER f (J, E, GF,
  GO, T)
• Would like to measure or estimate J, E, GF, GO

Passivating oxide layer
4
Langmuir Probes
www.hidenanalytical.com
www.staldertechnologies.com
5
On Wafer Ion Flux Probe Measurements

• Probe mounted on 8 heavily doped Si wafer.
• Probe biased at -70 V with respect to the Si
  wafer
• Ion current determined by measuring the voltage
  drop across a known resistance.
• Both reference and measurement probe are isolated
  from ground using a floating power supply.
• Plasma sees the same surfaces during etching of a
  wafer.
• Probe and reference are etched but measurement is
  not affected.

Heavily Doped Conducting Si wafer
Measurement Probe
R
rf filter
Bias Voltage
V
Si
Kapton
6
Ion flux measurements in SF6/O2 plasmas

• Measurements were done in plasmas containing SF6,
  O2, HBr, Cl2, NF3 and probe worked well for
  extended periods of time.

? I-V probe ? ion flux probe
7
Measuring radical concentrations in a plasma
Line of Sight Appearance Ionization Mass
Spectrometry
Time, Cost, Footprint
Laser Induced Fluorescence
UV Absorption
IR Absorption
Optical Emission Spectroscopy with Actinometry
Accuracy
8
Optical Emission Spectroscopy
25 K
Imaging Spectrographs with CCDs
10 K
Monochromator PMT
Cost
Integrated spectrographs and data acquisition
3 K
Photodiode and narrow pass filter
0.5 K
10 1
0.1
Resolution, nm
9
Optical Emission Spectra
10
Optical Emission
u
hn
?
e-
ground state

• Emission intensity depends on nx, ne and Te
• Emission intensity is not a measure of X
  concentration

11
Optical Emission Actinometry
J. Coburn and M. Chen, J. Appl. Phys. 51,
3134(1980).
12
Actinometry Requirements

• Excitation to the emitting states of X and
  actinometer (e.g., Ar) must have similar
  magnitude cross sections and thresholds.
• sex, X ? sex, Ar then a is a weak f(Te).
• sex, X ? sex, Ar then a is f(Te) which must be
  determined.
• Emitting state must only be populated by electron
  impact excitation of the ground state.

13
Example Use of OES and ion flux measurements in
SF6/O2 etching of Si
10 mT 25 mT 40 mT 75 mT

• Etch rate has a maximum at some intermediate
  pressure (25 mTorr).

800W TCP/-20 V rf-bias/40 SF6/40 O2/150 sec
14
Example Absolute measurements of Cl and Cl2
concentrations in Cl2 plasma
Donnelly, J. Vac. Sci. Technol. A 14, 1076
(1996)Malyshev, Donnelly, Kornblit, and Ciampa,
J. Appl. Phys. 84, 137 (1998) Ullal, Singh,
Daugherty, Vahedi and Aydil J. Vac. Sci. Technol.
A 20, 1195 (2002).

• A number of emission lines for Cl2, Cl and Ar
  studied for suitability (Donnelly, et al.)
• 305 nm Cl2 emission (Eth 8.4 or 9.2 eV)
• 822 nm Cl emission (Eth 10.5 eV)
• 750.4 nm Ar emission (Eth 13.5 eV)
• The Ar emitting state has unusually low s for
  excitation from Arm but threshold does not match
  the Cl2 or Cl thresholds
• a must be corrected for Te dependence

15
From emission intensities to absolute
concentrations

• In the limit power? 0
• dissociation ? 0 nCl2 ? ng Pg/kBTg
• Repeat zero-power extrapolation at different
  pressures to determine aCl2 (Te)
• Determine Cl concentration by mass balance or
• Use nCl at single point to determine aCl,Ar

P 10 mTorr, no wafer, Q 100 sccm Cl2
16
Studying the Effect of Walls on the Cl2
Dissociation Using OES and Actinometry

• SiO2 covered walls low Cl sticking probability
  g0.03

• Alumina reactor walls high Cl sticking
  probability g1

17
Mass Spectrometry
http//www.mcb.mcgill.ca/hallett/GEP/PLecture1/Ma
ssSpe_files/image011.gif
18
Line of Sight Threshold Ionization Mass
Spectrometry

• Threshold ionization can be used for detecting
  ALL radicals in a plasma
• Density of radicals is obtained at the substrate
  plane

Principle of TIMS
O e ? O 2e 13.6 eV (E1) O2 e ? O
O 2e 19.0 eV (E2)

• Since E1 gt E2, an electron energy scan can
  differentiate the two products
• E2-E1 is typically equal to the bond energy of
  the bond that is broken during dissociative
  ionization

19
Dissociation on the ionizer filament also
produces radicals which must be distinguished
from the radicals in the beam extracted from the
plasma
Molecules are thermally dissociated on the
filament and ionized resulting in a spurious
background signal.
20
25-200 mTorr
10-5 Torr
10-9 Torr
10-7 Torr
21
Beam-to-Background Ratio

• Pure O2 Beam-to-background ratio 3.2 at 25 mTorr
  and 2.0 at 200 mTorr.
• For radicals, the beam-to-background ratio will
  depend on the sticking probability of the radical.

22
O atom detection in O2 plasma
O e ? O 2e 13.6 eV O2 e ? O O 2e
19.0 eV
m/e 16

23
O in the beam Signal w/chopper open Signal
w/chopper closed
O e ? O 2e 13.6 eV O2 e ? O O 2e
19.0 eV
24
Quantifying the Mass Spectrometer signal
where, S QMS signal in c/s a product of
m/e-ratio dependent factors Ie electron
current of the ionizer s cross-section of the
ionization process nionizer number density of
neutrals in the ionizer (nbeam nbackground)
25
Calibration

• CH4 (m/e 16) is used for calibration.
• QMS signal for CH4 is measured for a known
  pressure of the gas in the plasma chamber under
  plasma-off condition
• CH4 calibration must be done right after the O
  concentration measurements to avoid the effect of
  drifts in the SEM sensitivity

Singh, Coburn, and Graves, JVST A 17, 2447
(1999). Singh, Coburn, and Graves, JVST A 18, 299
(2000). Agarwal, Quax, van de Sanden, Maroudas
and Aydil, JVST A 22, in press (2004). Agarwal,
Hoex, van de Sanden, Maroudas and Aydil, Appl.
Phys. Lett, in press (2003).
26
Absolute O atom concentrations in O2/Ar discharge
27
Example N2 (metastable A3Su state) and N
concentrations in N2 plasma
N2 e ? N2 2e 9.4 eV (??) N2 e ? N2
2e 15.6 eV

• In plasma assisted MBE of GaN, N2 may be
  preferred over N as the nitrogen precursor.
• Can N2 be detected and absolute concentrations
  of N and N2 measured?

28
Probable Franck-Condon Transitions
Transition a 11 eV Transition b 12
eV Transition c 14 eV
29
Absolute N2 and N Concentrations
30
Summary

• Simultaneous with the emergence of plasma
  processing as an enabling technology, a variety
  of plasma diagnostic methods have been developed
  over the last two decades to measure internal
  plasma properties.
• Ion current probe
• OES and actinometry
• Line of sight threshold ionization mass
  spectrometry
• Ease of implementation range from methods that
  take days-week to Ph.D. lifetime.
• To save time and money the first ask What do we
  want to measure and how accurately do we want to
  measure it?
• Measurement of radical concentrations over the
  wafer and ion flux impinging on its surface help
  in process development and improve fundamental
  understanding of etching and deposition processes.

31
Acknowledgements

• Sumit Agarwal (now _at_ University of Massachusetts)
• Jun Belen (UCSB)
• Dr. Sergi Gomez (UCSB)
• Bram Hoex (now _at_ Eindhoven Univ. of Technology)
• Guido Quax (now _at_ Eindhoven Univ. of Technology)
• Saurabh Ullal (now Lam Research Corporation)