1. PERSISTENT SURVEILLANCE FOR PIPELINE PROTECTION AND THREAT INTERDICTION Carbon as a flow-through, consumable PFC material Peter Stangeby University of Toronto Presented at the ReNew Theme III workshop; Taming the Plasma Material Interface UCLA, March 4-6, 2009 UNIVERSITY OF TORONTO Institute for Aerospace Studies

2. Tungsten is an attractive PFC material for reactors, however… Tungsten PFCs may not be compatible with AT operation since AT may involve high plasma temperature at the edge, resulting in high sputtering rates from walls and targets. Concentrations of high-Z impurities in the confined plasma must be kept to very low levels. Much higher concentrations are permissible for the low-Z refractory carbon. Graphite, however, may not be compatible with the high neutron fluence of devices like FDF, DEMO and reactors due to swelling damage.

3. Tungsten is an attractive PFC material for reactors, however… Tungsten experiences a surface roughening effect due to He bombardment at energies below the threshold for physical sputtering (simultaneous D + T + 5-10% He, as will occur in burning plasmas), creating a surface “foam” (nanoscopic morphology) [Nishijima, Baldwin, Doerner]. “ Such structures are potentially large sources of high- Z dust, could harbour significant levels of retained hydrogen isotopes in the plethora of helium nano-bubbles/voids within nano-structures and may have a dramatic influence on surface-thermal properties of W plasma facing components.” [Baldwin and Doerner]. The effect becomes significant over ~1000K and could be the most serious issue for W in DEMO [Konishi and Ueda].

4. A potential solution to the W ‘foam’ problem: low-Z coatings - A solution to the problem has been proposed [Baldwin, Doerner, Nishijima, Tokunaga, Ueda ]: coating the W with a low-Z material such as Be, B or C. - Coating W with a low-Z material could also solve the problem of unacceptably high concentrations of W in the confined plasma. - The low-Z coatings would be consumed and replenished continuously and so neutron damage would not be an issue. - Continuous refurbishment of the plasma-facing surfaces with coatings would also solve the basic problem of loss of structure in high duty cycle devices.

5. Very little tritium retention in low-Z coatings at high temperature Reactors will operate with hot walls, ~500C-1000C, for reasons of thermal efficiency. At high temperatures there is little tritium retention in any materials, including C, where H/C 1000C [J Davis]. Low-Z coatings in hot wall devices will result in little tritium retention by co-deposition, which is a serious issue for cold (200-300C) wall devices like ITER. In hot wall devices the tritium control problem is less likely to be retention than permeation into the cooling system. Low-Z coatings may also act as a permeation barrier, although if they reduce the surface recombination rate, that could increase T retention and permeation; R&D is required, including into finding the best substrate material.

6. Unique advantage of C as low-Z coating: unwanted deposits easily removed by O 2 -baking The continual replenishing of the low-Z coatings will inevitably result in thick deposits accumulating at some locations, even if they contain little tritium, will cause other problems, including dust formation and UFOs causing disruptions. It will therefore be as important to be able to remove, in situ, the low-Z material as to be able to create, in situ, the coating in the first place. Carbon appears to be unique in this regard since it is possible both to introduce it into the vessel as a gas e.g. vapor deposition and to remove it as a gas, as CO and CO 2 using oxygen baking. Since little tritium would be contained in the carbon deposits, there would be little creation of tritiated water, which is problematic to handle and process.

7. Estimates of carbon gross, net erosion rates and tritium codeposition rates For a reactor having P heat = 400 MW, and assuming P rad = 75% P heat, thus 0.25P heat = γkT s φ s, where γ = sheath heat transmission coefficient = 7; T s = plasma average temperature in contact with surfaces = 10 eV assumed here; φ s = total D/T-ion flux to all surfaces [ions/s], targets and walls. C physical sputtering due to D/T-ions and self- sputtering, Y eff = 0.005 (Eckstein 2002 yields for maxwellian ions plus a 3kT-sheath). Chemical sputtering and RES assumed to be not significant at assumed surface temperature of 1000C. 80% duty cycle. Then gross erosion rate = 23 tons/yr.

8. No chemical sputtering or radiation enhanced sputtering (RES) of carbon at 1000C Chemical sputtering → 0 for T > ~ 700C. [Toronto, JNM 255 (1998) 153] RES ~ 0 until T > 1100C [Tore Supra, PPCF 40 (1998) 1335]. vs RES ~ 0 until T > 1900C [TEXTOR, JNM 220-222 (1995) 467].

9. Estimates of carbon gross, net erosion rates and tritium codeposition rates, cont’d Net erosion is far more difficult to calculate reliably. Here, for concreteness, we assume net/gross = 0.1. Thus net erosion rate = carbon throughput rate ~2 tons C/yr. O 2 -baking at >700C is extremely rapid, such that even a graphite substrate, if present, would be removed quickly. At 1000C, T/C < ~ 0.0015 in carbon codeposits, thus T-throughput via codeposition ~ 1.5 kg T/yr.

10. Reactivity of carbon with O 2 is very high for T > 700C. 2 gmC/m 2 /s removed in 1 atm air 2 tonsC/yr net deposits Assume C deposited on 100 m 2. Thus time to remove: once/yr: 3 hrs once/mo: 15 min once/wk: 3 min once/day: 30 sec W substrate, if exposed, releases volatile oxides for T > 1250C. SiC substrate, if exposed, degrades for T >1400C. pure graphite codeposits pure graphite

11. Conclusion Like Li, C may be usable as a flow-through PFC material Experiments are required to assess this possibility: - studies of very high temperature O 2 -baking of codeposits - assessment of collateral damage to very high temperature baking - identification of best substrate material