Unexpectedly large tritium amounts were trapped in the Plasma Facing Components of JET and TFTR during the respective tritium campaigns. Newly created co-deposited layers of carbon and hydrogen were identified as the main sinks. The first wall of ITER in contrast to JET and TFTR will be covered with beryllium, whereas the divertor tiles will be built of tungsten with the exception of a relatively small area of carbon fibre composites. Due to these three materials the composition of the newly created layers will change as a function of plasma operation. Their possible hydrogen content is not known yet and as a consequence the estimates of potentially trapped tritium differ strongly. To respect safety limits measurements of the mobilisable tritium inventories inside the vacuum vessel are required. The present strategy is to rely on the accountancy of the accessible tritium inside the fuel cycle and to derive the quantity of tritium trapped inside the vessel by difference. The tritium injected into the machine is only measured by mass flow meters and no effort is made to determine the tritium exhausted.

Enhancements to determine the tritium and deuterium amounts injected into the torus and first proposals for enabling accountancy of the tritium and deuterium released from the torus cryo-pumps on a shot-by-shot basis are given. Only few additional buffer volumes and a micro gas chromatograph are required as the solutions are simple and inexpensive. These tools could be used already in the H-phase of ITER to obtain an integral value of the hydrogen trapped in the co-deposited layers by simple addition of small concentrations of deuterium to the protium and measuring the injected and released deuterium amounts.