8.4 Recycling and Retention of Fuels
153
flux, or fluence of hydrogen into PFM in present tokamaks is much less than that
expected in a fusion reactor, the estimation of the fuel inventory in the reactor should
be done by simulation with a laboratory work using H or D ion beams and/or linear
plasma machines or by extrapolation of the retention data in the present tokamaks.
To make reliable estimation both the data in laboratory works and in tokamaks
should be carefully compared. Until now, however, both do not give reliable results,
mainly because of uncertainty in fluence dependence which changes with the flux
and temperature of PFM.
For determination of the hydrogen retention in present tokamaks, there are two
ways; one is to integrate flux differences of throughput and exhaust with discharge
duration. This is very difficult to get reliable results, because the difference is very
small, and the determination of the exhaust flux includes large error. The other is
to make post-mortem analysis of retained hydrogen in PFM. Although the latter is
basically more reliable than the former, it is not easy to analyze all PFM with larger
area. In the following, details of the fuel retention are discussed.
8.4.3.1 Fuel Retention in Carbon and Metals
The reemission from carbon materials (C) includes hydrocarbons resulting from
chemical sputtering and causes significant surface recession (erosion). Initially, fuel
retention mostly occurs in surface and near-surface regions. Afterward, deposited
layers gradually grow, and the fuel retention is dominated in the deposited C layers.
In particular, the deposited C layers on plasma shadowed area continuously grow so
as the fuel retention.
In case of metals, fuel easily diffuses in and out. Therefore, after stopping the
injection, large amount of fuel retained during the injection is released out. Accordingly, the difference between the dynamic retention under plasma exposure and the
static retention after stopping the exposure is quite large. In contrast, fuel injected in
C is mostly trapped near the surface within the projected range making D-C and T-C
bonds until C is saturated with D and T. Once trapped making such bonds hydrogen
(D and T) is hardly released. In addition, very small hydrogen diffusivity in C inhibits
its penetration into the bulk. Therefore, the difference between the dynamic retention and the static retention in C is smaller than that of the metals as appeared in
Fig. 8.13. However, the porous nature of C allows hydrogen molecules to penetrate
deep through open pores resulting in the depth profiles given in Fig. 9.12. This means
that residual fuel gases in a reactor vessel can penetrate through the whole thickness
of a carbon tile, resulting in all carbon crystallites (C grains) surfaces being saturated with the fuel. Actually, T was observed at the backside of carbon tiles used in
DT discharge campaign in JET Mark-IIA divertor [38]. However, fuel particles can
hardly penetrate inside of the C grains.
Unfortunately, little systematic data for D retention with metallic plasma-facing
wall have been available except recent JET experiments with ITER-like wall (ILW),
i.e. W divertor and Be first wall. Nevertheless, above mentioned differences between
the C-wall and ILW clearly appeared as shown in Fig. 8.13 [39, 40], which shows
153
flux, or fluence of hydrogen into PFM in present tokamaks is much less than that
expected in a fusion reactor, the estimation of the fuel inventory in the reactor should
be done by simulation with a laboratory work using H or D ion beams and/or linear
plasma machines or by extrapolation of the retention data in the present tokamaks.
To make reliable estimation both the data in laboratory works and in tokamaks
should be carefully compared. Until now, however, both do not give reliable results,
mainly because of uncertainty in fluence dependence which changes with the flux
and temperature of PFM.
For determination of the hydrogen retention in present tokamaks, there are two
ways; one is to integrate flux differences of throughput and exhaust with discharge
duration. This is very difficult to get reliable results, because the difference is very
small, and the determination of the exhaust flux includes large error. The other is
to make post-mortem analysis of retained hydrogen in PFM. Although the latter is
basically more reliable than the former, it is not easy to analyze all PFM with larger
area. In the following, details of the fuel retention are discussed.
8.4.3.1 Fuel Retention in Carbon and Metals
The reemission from carbon materials (C) includes hydrocarbons resulting from
chemical sputtering and causes significant surface recession (erosion). Initially, fuel
retention mostly occurs in surface and near-surface regions. Afterward, deposited
layers gradually grow, and the fuel retention is dominated in the deposited C layers.
In particular, the deposited C layers on plasma shadowed area continuously grow so
as the fuel retention.
In case of metals, fuel easily diffuses in and out. Therefore, after stopping the
injection, large amount of fuel retained during the injection is released out. Accordingly, the difference between the dynamic retention under plasma exposure and the
static retention after stopping the exposure is quite large. In contrast, fuel injected in
C is mostly trapped near the surface within the projected range making D-C and T-C
bonds until C is saturated with D and T. Once trapped making such bonds hydrogen
(D and T) is hardly released. In addition, very small hydrogen diffusivity in C inhibits
its penetration into the bulk. Therefore, the difference between the dynamic retention and the static retention in C is smaller than that of the metals as appeared in
Fig. 8.13. However, the porous nature of C allows hydrogen molecules to penetrate
deep through open pores resulting in the depth profiles given in Fig. 9.12. This means
that residual fuel gases in a reactor vessel can penetrate through the whole thickness
of a carbon tile, resulting in all carbon crystallites (C grains) surfaces being saturated with the fuel. Actually, T was observed at the backside of carbon tiles used in
DT discharge campaign in JET Mark-IIA divertor [38]. However, fuel particles can
hardly penetrate inside of the C grains.
Unfortunately, little systematic data for D retention with metallic plasma-facing
wall have been available except recent JET experiments with ITER-like wall (ILW),
i.e. W divertor and Be first wall. Nevertheless, above mentioned differences between
the C-wall and ILW clearly appeared as shown in Fig. 8.13 [39, 40], which shows