Structural Studies of Carbon Dust Samples Exposed to NSTX Plasma

Structural Studies of Carbon Dust Samples Exposed to NSTX Plasma

Yevgeny Raitses and Charles H. SkinnerPrinceton Plasma Physics Laboratory

Fuming Jiang and Thomas S. DuffyDepartment of Geosciences, Princeton University

18th International Conference on Plasma Surface Interactions May 26-30, 2008, Toledo, Spain

PrincetonUniversityGeosciences

Abstract

Raman analysis has shown that the production of carbon dust particles in NSTX involves strong modifications of the physical and chemical structure of the original graphite. Raman spectra were measured for different dust samples exposed to the NSTX plasma, unexposed dust samples, carbon deposits produced in an atmospheric pressure helium arc discharge, and heat-treated carbon samples. The arc electrodes and heat-treated samples were made from the same graphite material as the NSTX tiles. For the unexposed particles, the high energy G-mode peak (Raman shift ~1580 cm-1) is much stronger than the defect-induced D-mode peak (Raman shift ~ 1350 cm-1), a pattern that is consistent with Raman spectrum for commercial graphite materials. For dust particles exposed to the plasma, the ratio of G-mode to D-mode peaks is lower and becomes even less than 1. This behavior indicates on a strong increase of structural disordering in plasma exposed samples. We also demonstrate experimentally that heating to 1500-2500 C alone can not explain the observed structural modifications indicating that they must be due to higher temperatures needed for graphite vaporization, which is followed either by condensation or some plasma-induced processes leading to the formation of more disordered forms of carbon material than the original graphite. Moreover, because we found similar plasma-induced structural modifications in dust particles exposed under the different background gases (hydrogen and helium) and the different plasma conditions of the arc and NSTX experiments, our results suggest that the observed structural modifications can be unrelated to hydrocarbon compositions on the dust surface or implantation or thermal diffusion.

This work was supported by the US DOE under contract No. DE-AC02-76CH03073

Motivation

Dust from graphite tiles - tokamak safety issue.

Possible mechanisms of dust production:

• Evaporation of wall material during disruptions etc. • Spallation and flaking of thin films or redeposited material. • Sputtering during wall cleaning or conditioning.• Plasma-induced synthesis and growth.

Different formation mechanisms of carbonaceous dust should leave their signatures in their macro and micro-structures.

Results from previous dust studies

Raman spectroscopy studies suggest that tokamak-produced dust consists of graphite-like particles with different degree of disordering.1,2

Ion implantation,2,3 plasma-induced growth,4,5 heating3,6 etc may be responsible for the observed structural modifications of dust.

Note: Similarities in structural modifications of carbon dust produced in JT-60 and by the low pressure (deuterium) arc discharge.7

1. P. Roubin et al., J. Nucl. Mater. 337-339, 990 (2005) 2. T. Tanabe, “Application of optical techniques…”9th ITPA, Garching (2007)3. K. Niwase et al., J. Nucl. Mater. 179-181, 214 (1991)4. J. Winter, Plasma Phys. Control. Fusion 40, 1201 (1998)5. Ph. Chappuis et al., J. Nucl. Mater. 290-293, 245 (2001)6. C. Arnas et al., J. Nucl. Mater. 337-339, 69 (2006)7. H. Yoshida et al., J. Nucl. Mater. 337-339, 604 (2005)

Objectives and approach of this study

Explore plasma and thermal effects on structural modifications of graphite-type dust particles.

Compare Raman spectra from:

- NSTX (deuterium discharge)- Atmospheric pressure helium arc discharge- Vacuum heat treatment (no plasma)- Unused graphite tile

Note that we study carbon dust particles from very different plasma regimes, including pressure, gas and plasma properties.

NSTX

• Exploring the physics of high beta and high confinement in a low aspect ratio device

• Demonstrate non-inductive current generation and sustainment.

• plasma major radius: 0.85 m,

• minor radius: 0.68 m,

• toroidal field of up to 0.55 T,

• plasma current up to 1.5 MA

• pulse duration up to 1.5 s.

• 7 MW neutral beam injection

• 6 MW of high harmonic fast wave RF at 30 MHz. Plasma facing components that are in

contact with the plasma are protected by a combination of graphite and CFC tiles.

The National Spherical Torus Experiment (NSTX) research program is aimed at:

NSTX dust (collaboration with Phil Sharpe INL.)

Dust sampled by filtered vacuum collection technique (2001).

SEM image of dust obtained from plasma-exposed surfaces in the upper divertor region of NSTX.

Count based dust size distribution.

Count median diameter 3.3 µmComposition: carbon, some particles also contain boron.

J. P. Sharpe et al., J. Nucl. Mater, 337-339 (2005) 1000.

PPPL arc setup

Power Source -+

Ballast Resistor

Anode positioner

Anode Cathode

Vacuum chamber

Power Source -+

Ballast Resistor

Anode positioner

Anode Cathode

Vacuum chamber

- Movable graphite anode rod ( 0.5 cm) made from an unused NSTX tile and brass cathode disk ( 2.5 cm).

- Arc chamber is evacuated by a vacuum pump and then filled to ~ 1 atm with Helium gas.

Positioner

Anode

Graphiteanode

Brasscathode

Arc operation and heating experiments

Typical arc parameters:

Current : 80 A, Voltage: < 20 V Pressure: 1 atm (Helium)Anode temperature >3000 C

Plasma: mostly carbon:

Te ~ 1-5 eVTi ~ 0.5 eVNe ~ 1015/cm3

Heating experiments:

• Vacuum furnace (up to 1700 C)

• Ohmic heating in arc setup (1400-2000 C)

• High heat flux from the plasma leadsto vaporization of a graphite anode.

• Vapor products are deposited on the cathode (anodic arc)

Graphite anode

Brasscathode

Arc

1-2 mm

Soot deposit

Examples of test samples used in this study

0.1 cm 2 cm

2 cm 2 cm

Tip

NSTX flake and dust Post-run-arc cathode soot

Post-run-arc anodeGraphite sample used in heating experiments

0.1 cm

Raman spectroscopy offers a sensitive measure of the microstructure of carbon dust

0.2 W Ar laser

Dust SampleBeamSplitter

Spectrometer

vibrationalstates

Frequency shift of scattered light

Micro-Raman system* consists of

• An air-cooled 200 mW argon ion laser was operated at < 3 mW.

• holographic optics

• 0.5-m spectrometer

• Liquid nitrogen cooled CCD detector (1100 x 330 pixels).

*S.-H. Shim and T. S. Duffy, American Mineralogist 87, 318 (2002).

Energy leveldiagram

Raman spectra contain information on vibrational energies, chemical bonds, molecular symmetry and the physical-chemical environment.

Depth of absorption (=514.5 nm) ingraphite:~ 50 nm

Raman spectrum (RS) of carbon materials

1200 1400 1600 1800

Raman shift (1/cm)

1200 1400 1600 1800Raman shift (1/cm)

1332 2 cm-1 1580 cm-1

1555 cm-1

RS of graphite samples from the tile has 2 characteristic peaks:

- Graphitelike G-mode (~1580 cm-1) - Disordered D-mode (~1350 cm-1)

1200 1400 1600 1800Raman shift (1/cm)

Natural diamond Carbon

nanotube(not foundin NSTX)

Sample scraped from unused NSTX tile 1350 cm-1

D-mode

1580 cm-1

G-mode

• The peak intensities : ID/IG 1 indicate on strong disordering.1,2

• Similar Raman spectra for dust in Tore Supra3, JT-604.

also, for high energy irradiated graphite samples5 and arc4.

Note:2 carbon black and glassy carbon have RS with ID/IG > 1

Post-run dust samples

1200 1400 1600 1800Raman shift (1/cm)

Intensity (arb. units)

1200 1400 1600 1800Raman shift (1/cm)


Dust from NSTX: STRONG modifications

Dust sample from unused graphite tile

ID/IG 0.2

1200 1400 1600 1800Raman shift (1/cm)

D-mode

G-mode

1. Tuinstra and Koenig, J. Chem. Phys. (1970)2. R. Nemanich, S. Solin, Phys. Rev. B (1979)3. P. Roubin et. al, J. Nucl. Mater. (2005)4. H. Yoshida et al., J. Nucl. Mater. 2005) 5. K. Niwase et. al, J. Nucl. Mater. (1991)

ID/IG 1.1

ID/IG 1.5

Carbon ions, atoms and clusters from vaporized graphite anode reaching the brass cathode form a soot.

Post-run soot samples

Arc-produced soot: STRONG modifications

Brasscathode

Soot deposit

1200 1400 1600 1800Raman shift (1/cm)


ID/IG 1.15

ID/IG 0.8

Raman spectra of cathode soot and post-run NSTX dust samples are similar.


ID/IG 0.2

1200 1400 1600 1800Raman shift (1/cm)

1350 cm-1

D-mode

1580 cm-1

G-mode

1200 1400 1600 1800Raman shift (1/cm)

• In the arc operation, graphite anode is ablating: T > 3000C

• Anode samples are essentially heat treated (T ~ 2500 C measured a few mm below the anode tip).

• Post run anode has a thin film deposit on it.

Post-run samples from the anode tip

Graphite Anode

Anode tip and NSTX dust samples have very different Raman spectra.


ID/IG 0.2

1200 1400 1600 1800Raman shift (1/cm)

1350 cm-1

D-mode

1580 cm-1

G-mode

1200 1400 1600 1800

Raman shift (1/cm)


Sample with thin deposit

withoutdeposit

(removed)

ID/IG 0.2

Post-run anode tip: GRAPHITIC order

Heating experiments: WEAK disordering

Vacuum furnace: Samples heated to 1700C

at < 10-5 torr, for 1 hour


ID/IG 0.2

1200 1400 1600 1800Raman shift (1/cm)

1350 cm-1

D-mode

1580 cm-1

G-mode

1200 1400 1600 1800Raman shift (1/cm)


1200 1400 1600 1800Raman shift (1/cm)


Anode sideID/IG 0.1

Cathode sideID/IG 0.4

Ohmic heating (2000C)

Anode side

Cathodeside

Hot part

Thermal effects up to 2000 C alone do not lead to a strong disordering at temperatures insufficient for vaporization of the graphite

• Samples were observed on a dust detector (P1-42) after retrieval from NSTX.

• Similar to previous samples, the ratio ID/IG increases indicating on strong disordering.

• Different Raman spectra in the range of 2800-3100 cm-1 .• Note that peaks in this range

are usually assigned to heavier hydrocarbons involving C=C and C≡C bonds

Preliminary Results

Dust from Recent NSTX Experiments


ID/IG 0.2

1200 1400 1600 1800Raman shift (1/cm)

D-mode

G-mode

1. Ushizawa et al., Phys. Rev. B (1999).2. D. Lin-Vien et al., Handbook of Infrared and Raman Characteristic Frequencies of Organic Molecules (Academic Press, Inc., 1991).

ID/IG 0.7

1200 1400 1600 1800

Raman shift, 1/cm

2500 2700 2900 3100 3300Raman shift, 1/cm


2900 3100 3300Raman shift, 1/cm


Concluding remarks

Raman measurements indicate that the production of carbon dust particles in NSTX and arc discharge involves ablation processes (e.g. graphite vaporization), that are followed by the deposition and the formation of similarly disordered forms of carbon material.*

Note: more disordered forms than the original graphite

The double peaked shape of the Raman spectrum of the NSTX dust and post-arc cathode soot is similar to the results from Tore Supra and JT-60.

Note: this spectral shape is also found in highly disordered carbon black, and less disordered microcrystalline graphite and glassy carbon.

Higher resolution Raman measurements are required to investigate these structural similarities and to determine a more precise structure of the dust particles.

* Y. Raitses, C. H. Skinner, F. Jiang, T. S. Duffy, “Raman spectroscopy of carbon dust samples from NSTX”, J. Nucl. Mater. 375 (2008), 365-369

Documents

Structural Studies of Carbon Dust Samples Exposed to NSTX Plasma