156
8 PMI in Large Tokamaks
retention was observed in Tore-Supra, which had water cooled divertor/limiter blades
at the bottom of its tokamak vessel [42]. In the next chapter (Chap. 9), the detailed
estimation of the fuel retention in a reactor is made.
8.5 T-Related Issues on the In-Vessel T Inventory
Evaluations of the in-vessel fuel inventory have been mostly done by integration
of local retentions in redeposited layers and/or near surfaces assuming toroidal
symmetry in tokamaks. Although the bulk retention should have certain contribution, it has been scarcely assessed. The bulk retention in the plasma-facing materials
must be very small in its concentration. Nevertheless, its integrated amount over full
volume of the PFM and vacuum vessel could not be small. Here, the maximum bulk
retention for metallic PFM is estimated according to crude assumptions as follows.
It is well known that outgassing rates from metals are ranging from 10
−6 to 10
−3
Pa·m
3 ·m
−2 ·s
−1 [43]. Taking 10
−5 Pa·m
3 ·m
−2 ·s
−1 as the outgassing rate from plasmafacing surfaces with their surface area of 1000 m
2 , the total outgassing rate would be
10
−2 Pa·m
3 ·s
−1 . This seems negligible compared to the maximum fuel through-put
in ITER of 200 Pa·m
3 ·s
−1 (~2.5 g·s
−1 ) [44].
Since the source of the H outgassing is very likely H retained in the materials as
impurity and H uptake can occur at the backside surface with oxidation by water.
Therefore, it might be reasonable to assume tritium uptake in the bulk proceeds
just opposite to the outgassing. Then about 10
−2 Pa·m
3 ·s
−1 (0.1 mg·s
−1 ) orders of
tritium would be continuously retained in the bulk, i.e. every ITER discharge (400 s)
would remain 0.4 g of T in the bulk and 25000 shots pile up 1 kg. This might be
overestimation. However, another estimation given below would probe this T pile-up
scenario is realistic considering that any materials retain H as an impurity with more
than 10 ppm. Suppose T can be replaced by the impurity H, though the reaction rate
is totally depending on diffusion time. Then, 10 ppm T could be dissolved in the
bulk after many discharges or at steady state. This suggests that usage of 1000 tons
of structure materials for a reactor could result in 10 kg of T at the maximum. This
type of tritium inventory is well known in the detection of T by an ionizing chamber
as a memory effect, i.e. initially introduced T to the ionization chamber and piping
disappeared and certain time is required to start the detection, and the detection
continues for while after stopping the T introduction as explained in Chap. 9 in ref.
[33]
From the aspect of tritium safety, this type of inventory is often categorized to
be immobilized T inventory and might not be a serious concern, but kg orders of T
would disappear in the bulk of structure materials of all tritium handling systems at
the starting phase of every reactor. This will be discussed later again.
Difficulty of quantitative analysis of T in the in-vessel components adds additional
problems. It is ironical that the accuracy in detecting low levels of T (below 10
9 Bq)
by its β decay is better than that in the very high levels which are determined by mass
and/or pressure measurements and calorimetry with the accuracy of only 10
−2 –10
−4 .
8 PMI in Large Tokamaks
retention was observed in Tore-Supra, which had water cooled divertor/limiter blades
at the bottom of its tokamak vessel [42]. In the next chapter (Chap. 9), the detailed
estimation of the fuel retention in a reactor is made.
8.5 T-Related Issues on the In-Vessel T Inventory
Evaluations of the in-vessel fuel inventory have been mostly done by integration
of local retentions in redeposited layers and/or near surfaces assuming toroidal
symmetry in tokamaks. Although the bulk retention should have certain contribution, it has been scarcely assessed. The bulk retention in the plasma-facing materials
must be very small in its concentration. Nevertheless, its integrated amount over full
volume of the PFM and vacuum vessel could not be small. Here, the maximum bulk
retention for metallic PFM is estimated according to crude assumptions as follows.
It is well known that outgassing rates from metals are ranging from 10
−6 to 10
−3
Pa·m
3 ·m
−2 ·s
−1 [43]. Taking 10
−5 Pa·m
3 ·m
−2 ·s
−1 as the outgassing rate from plasmafacing surfaces with their surface area of 1000 m
2 , the total outgassing rate would be
10
−2 Pa·m
3 ·s
−1 . This seems negligible compared to the maximum fuel through-put
in ITER of 200 Pa·m
3 ·s
−1 (~2.5 g·s
−1 ) [44].
Since the source of the H outgassing is very likely H retained in the materials as
impurity and H uptake can occur at the backside surface with oxidation by water.
Therefore, it might be reasonable to assume tritium uptake in the bulk proceeds
just opposite to the outgassing. Then about 10
−2 Pa·m
3 ·s
−1 (0.1 mg·s
−1 ) orders of
tritium would be continuously retained in the bulk, i.e. every ITER discharge (400 s)
would remain 0.4 g of T in the bulk and 25000 shots pile up 1 kg. This might be
overestimation. However, another estimation given below would probe this T pile-up
scenario is realistic considering that any materials retain H as an impurity with more
than 10 ppm. Suppose T can be replaced by the impurity H, though the reaction rate
is totally depending on diffusion time. Then, 10 ppm T could be dissolved in the
bulk after many discharges or at steady state. This suggests that usage of 1000 tons
of structure materials for a reactor could result in 10 kg of T at the maximum. This
type of tritium inventory is well known in the detection of T by an ionizing chamber
as a memory effect, i.e. initially introduced T to the ionization chamber and piping
disappeared and certain time is required to start the detection, and the detection
continues for while after stopping the T introduction as explained in Chap. 9 in ref.
[33]
From the aspect of tritium safety, this type of inventory is often categorized to
be immobilized T inventory and might not be a serious concern, but kg orders of T
would disappear in the bulk of structure materials of all tritium handling systems at
the starting phase of every reactor. This will be discussed later again.
Difficulty of quantitative analysis of T in the in-vessel components adds additional
problems. It is ironical that the accuracy in detecting low levels of T (below 10
9 Bq)
by its β decay is better than that in the very high levels which are determined by mass
and/or pressure measurements and calorimetry with the accuracy of only 10
−2 –10
−4 .