Ma terials Sci ence & Technology

Introduction to / Overview of Plasma Diagnostics

James Whitby

Outline

What is plasma diagnostics

What do we want to know? Optical spectra: GD-OES Mass spectra: GD-MS

Free techniques Electrical parameters Physical probe techniques

Langmuir probe etc. Non perturbing

Optical techniques

Refractive index measurements e.g. interferometry, Schlieren Emission measurements e.g. cyclotron, bremsstrahlung, line widths, Zeeman splitting Laser probe measurements: Rayleigh, Thomson and Raman scattering e.g. optogalvanic spectroscopy, laser induced fluorescence, coherent anti-Stokes Raman spectroscopy

Almost non-perturbing

Other techniques

What is/are Plasma Diagnostics

Oxford English Dictionary Definitions

Plasma
a gas of positive ions and free electrons with little or no overall electric charge

Diagnostics
the practice or techniques of diagnosis Diagnosis the identification of the nature of an illness or other problem by examination of the symptoms

To deduce information about the state of the plasma from practical observations of physical processes and their effects
I R Hutchinson, Principles of Plasma Diagnostics

What do we want to know?

Time and spatially resolved knowledge of

Number densities of neutrals, ions and electrons Fluxes Excited state densities Electron energy distribution function Ion energy distribution function (velocities) Neutral gas temperature, sample temperature Electric and magnetic fields Gas flow velocities

Why do we want to know?

In order to test models of the system, and so to be able to

Confirm that the underlying physics are understood Better interpret results Predict the behaviour of new samples Optimise the analytical instrumental hardware and parameters

Free techniques

In normal operation of analytical instruments plasma parameters such as gas pressure, VDC (for RF systems), current and voltage may be routinely recorded GD-OES provides information on line emission from selected species, and perhaps a measure of background intensity. GD-MS can (depending on mass analyser) provide information on many charged species. This free information can and should be used to assess if operation is nominal (no air leaks, hardware problems, contamination from pumps or impure process gas, unexpected elements in sample) Emission spectrum or mass spectrum may sometimes provide very crude estimates of electron temperature, or of importance of electron impact vs Penning processes

Ma terials Sci ence & Technology

Electrical Parameters

Plasma is characterised electrically by its (complex) impedance: at least two sheaths plus negative glow for us
average impedance deduced from voltage and current measurements Real part gives resistivity (conductance)

Can be used to estimate electron temperature in fully ionized plasmas (Spitzer conductivity) Sheaths dominate the impedance of our plasmas

For analytical RF glow discharges we see changes in the matching and/or reflected power which relate to the plasma conditions

Not used widely for diagnostics, but c.f. impedance probes

Ma terials Sci ence & Technology

Physical Probe Techniques

An electrical conductor is introduced into the plasma to provide local (possibly time resolved) measurements of

Electron temperature (Langmuir probe) Electron number density (Langmuir probe)

It does this by measuring particle flux (current) as a function of bias potential A Langmuir probe is essentially an insulated wire with an un-insulated tip

Ma terials Sci ence & Technology

Example from EMPA RF-GD-MS

140 120 100

Ion current saturates at -5 pA below -4.6 V Floating potential is -2.8 V Electron current saturates at 116 pA above ~10 V Plasma potential

current [pA]

80 60 40 20 0 -20 -10 -5 0 voltage [V] 5 10 15

Electron temperature can be obtained from a semi-log plot of electron current against bias voltage (reciprocal of gradient). Te = 1.0 eV Rather good linear fit suggests single maxwellian electron energy distribution

enlargement of log plot to determine electron temperature from slope

4 3.5 3 2.5 2 1.5 1 -3.2

ln(electron current /pA)

y = 1.0037x + 4.5473 Results of varying bias on quadrupole rods 2 R plasma, = 0.996 30 W, 500 Pa Ar/Cu RF +ve current from probe to plasma

-2.7

-2.2 voltage /V

-1.7

-1.2

Ma terials Sci ence & Technology

Langmuir probe theory 1

A probe of area A in a thermal plasma would be hit by more electrons than ions, so emits a positive current 1 1 I = eA 1 n " # n " $ eA n" 4 i i 4 e e 4 e e

If insulated (floating) will rapidly charge to the floating potential At plasma potential, little perturbation, current hitting probe is ! predominately from electrons Higher potential

If a potential is applied:

electron current saturates

Ion current decreases, but is already small

Lower potential
some electrons reflected zero current at floating potential Lower still, ion current saturates

Not true in practice - sheath changes size

Probe potential perturbs plasma over Debye length scale

Often Debye length De << a, the probe dimension

10

Langmuir Probe Theory 2

Collisions near to probe will reduce current

By ratio of mean free path to radius of probe if <<a Can be ignored if a<< for spherical probe Need to consider probe size and shape to decide if sheath around probe is collisional

From measurements and appropriate formula, can determine Te and ne (ni) to about 10%, also plasma and floating potentials May use double or triple probes, or emissive probes to attempt to improve performance To measure ion energies need gridded energy analyser

Problems with space charge effects

11

Problems with Langmuir probes

Melting of probe Contamination of probe

Can be reduced by in situ heating or sputtering Perturbation of plasma, appropriate theory

Interpretation of results

Need physical access to plasma

12

Optical measurements

Advantage of optical techniques is that they allow

minimal perturbation of plasma

None for emission measurements and refractive index measurements Minor for LIF, OGS, CARS etc.

Can give good spatial and temporal resolution

Disadvantage need windows, and often expensive equipment (tunable lasers)

Ma terials Sci ence & Technology

13

Refractive Index Measurements

Interferometry, Schlieren/shadowgraph

Rayleigh/Thomson scattering Raman scattering

Coherent Anti-Stokes Raman Scattering (CARS)

Laser Induced Fluorescence (LIF) Emission Measurements

lineshapes

Stark effect

14

Refractive Index Measurements

Refractive index of plasma depends on electron density through the plasma frequency

Use cold plasma approximation By interferometry

Refractive index can be measured

Michelson, Mach-Zender Measured phase shift gives line integral of refractive index and electron density technically challenging

Schlieren and shadowgraph imaging

Simpler experimentally than interferometry Inherently an imaging technique Gives derivative or second derivative of refractive index

15

Thomson Scattering for electron temperature and number density

Probe laser causes electrons to oscillate, resulting Doppler shifted emission is detected

Absolute values for ne and Te deduced from magnitude and shape of scattered signal

Shape depends on Debye length of plasma and the differential scattering wave vector (observation geometry)

Lumped together as scattering parameter

Te typically derived from plot of logarithm of power (intensity of detected signal) vs square of wavelength shift

Gradient of straight line gives electron temperature Curve fitting with fewer assumptions gives better results, especially for analytical plasmas

Typical observed wavelength shifts are c. 5 nm from an analytical ICP discharge Observation geometry controls spatial resolution

Ma terials Sci ence & Technology

16

Thomson Scattering

Advantages
Spatial resolution No perturbation of plasma

(So long as laser power kept low enough)

Fairly reliable (e.g. compared to Langmuir probes for ne)

Disadvantages
Expensive (pulsed laser) Need optical access from two directions and good baffles Weak signal

Cross section comparable to classical electron E.g. Te ~ Ti, Maxwellian EEDF

Some approximations typically used in interpreting results

Careful correction for stray light necessary

Ma terials Sci ence & Technology

17

Rayleigh Scattering

Elastic scattering of incident radiation

Proportional to number density

Gives gas kinetic temperature if pressure known and e.g. ideal gas law Caution! Argon metastable atoms have 500 times the cross section of argon ground state atoms

Detect at same wavelength as the incident light

Ma terials Sci ence & Technology

18

Raman Scattering and Coherent Anti-Stokes Ramans Scattering (CARS)

Raman (inelastic) scattering can be used to give number densities from the scattered intensity (as for Rayleigh)
Species and state specific, but must have a Raman crosssection Can be used for atoms (but hardly ever is)

Resonant enhancement possible

CARS is a non-linear Raman technique that gives a stronger signal and inherent spatial resolution

Widely used in combustion research to give temperatures from vibrational spectra (e.g. of N2)

Ma terials Sci ence & Technology

19

Absorption measurements

Classical technique to establish number density

Still widely used in analytical chemistry e.g. AAS Light resonant with a spectral transition is aimed through plasma Power loss is related to number of absorbing species in light path

If appropriate constants are known e.g. oscillator strength

May not be practical for ground states if required wavelength is in deep ultra-violet Diode lasers are making this technique more attractive
Ma terials Sci ence & Technology

20

Laser Induced Fluorescence

Used to map number densities of species/states

Laser tuned to resonance with a spectral line Excites species, decay (fluorescence) observed

Either at same wavelength or another (longer)

Signal depends on population of lower state Spatial resolution possible by observing at right angles to exciting beam

Tunable diode lasers (much cheaper than dye lasers) are making this easier Signal to noise ratio often better than for absorbtion measurements
Ma terials Sci ence & Technology

21

Stark effect

The linewidth of a spectral transition contains information about the environment

At low pressures, Doppler broadening may dominate

Gives translational temperature

Pressure broadening may be dominated by effects of charged particles (Stark broadening)

Line width/shape gives ion density Theory complicated, but known for e.g. hydrogen atoms

Directed flow may result in a Doppler shift Electric fields may cause a measurable Stark shift
Ma terials Sci ence & Technology

22

OptoGalvanic Spectroscopy (OGS)

(pulsed ) tunable laser is used to perturb system (like LIF) but rather than monitoring optical emission, the electrical properties of the plasma are monitored
e.g. if laser light creates ions, then current increases transiently In general, sign and magnitude of electrical signal can be used (via a model) to interpret energy flow in plasma Electrical detection can have advantages over optical detection (cost!) c.f. presentation on Saturday by R Djulgerova

Ma terials Sci ence & Technology

23

Data Treatment

Often want 3D information, but have a technique that gives line-of-sight integrals (e.g. absorption measurements) If chordal measurements are available for a cylindrically symmetric sytem then this data can be inverted to recover the 3D distribution e.g. Abel inversion

Sensitive to noise in data

Ma terials Sci ence & Technology

24

Other techniques

Ma terials Sci ence & Technology

25