T.S. Elleman

Hydrogen diffusion and solubility in pyrolytic carbon

Abstract Tritium diffusion coefficients and deuterium solubilities have been measured for laminar pyrolytic carbon in the temperature range 900–1500°C. The tritium diffusion coefficients were much lower than those for metals at equivalent temperatures, but the activation energy for the diffusion was much higher (∼-100 kcal/mole). Tritium diffusion coefficients measured for silicon-doped pyrolytic carbon were over an order of magnitude higher than the values for the undoped laminar pyrolytic carbon. The solubility of deuterium in laminar pyrolytic carbon was found to decrease with increasing temperature and exhibited a pressure dependence of p. 1 2

Tritium diffusion in 304- and 316-stainless steels in the temperature range 25 to 222 °C☆

Tritium diffusion measurements in 304- and 316-stainless steels were carried out over the temperature range 25 to 222 °C by direct measurement of tritium diffusion gradients. The 6Li (n, α)3H reaction was used to inject tritium into the specimens and to produce initial tritium concentrations in the range 0.0005 ppm to 0.007 ppm 3H by weight. Three components were identified from the concentration profiles: a surface region approximately 5 μm thick where tritium trapping occurred, a normal diffusion profile which appeared characteristic of bulk diffusion, and a rapid diffusion “tail” which was tentatively attributed to grain boundary diffusion. Surface release measurements of tritium verified the existence of a surface trapping layer. The bulk diffusion component was consistent with classical diffusion solutions and gave: D = 0.018 ( + 0.011 − 0.007) × exp − (0.61 ± 0.01 eV/kT) (cm2/sec). The surface trapping was tentatively attributed to the presence of helium stabilized voids in the specimen surface layers.

Surface effects on tritium diffusion in niobium, zirconium and stainless steel

Abstract Tritium diffusion in niobium, Zircaloy-2 and stainless steel has been studied by measurement of both tritium concentration profiles and surface tritium release rates. Concentration profiles show buildup of tritium in the surface layers of a specimen with classical diffusion behavior at depths greater than about 5 μm from the surface. Application of a two-region diffusion model to the experimental data gives tritium diffusion coefficients in the surface films which are lower than the bulk diffusion coefficients by two orders of magnitude in stainless steel and eight to ten orders of magnitude in niobium and Zircaloy over a temperature range of interest for fission and fusion reactor systems. The surface effect appears to be a consequence of oxide film formation and is not due to the helium injected into specimens along with the tritium.

Tritium diffusion in zircaloy-2 in the temperature range −78 to 204° C

Tritium diffusion measurements in Zircaloy-2 were carried out over the temperature range −78 to 204 °C by direct measurement of tritium diffusion gradients. The 6Li (n, α)3H reaction was used to inject tritium into the specimens and to produce initial tritium concentration in the range 0.0065 ppm to 0.013 ppm 3H by weight. Two diffusion components were identified from the concentration profiles: a surface trapping region approximately 5 μm thick and a normal diffusion profile characteristics of bulk diffusion. Surface release measurements of tritium verified the existence of a surface trapping layer. The bulk diffusion component was consistent with classical diffusion solutions and was given by: D = 0.00021−0.00018+0.005 exp−(8500 ± 200 cal/RT) cm2 · sec−1. The surface trapping was attributed to oxide films formed on the Zircaloy-2 at room temperature. The apparent diffusion coefficients for the surface region were consistent with: D = 4.0−3.3+19.7 × 10−14 exp−(7200 ± 1500 cal/RT) cm2 · sec−1 over the temperature range 25 to 411°C.

Surface effects on the diffusion of tritium in 304-stainless steel and zircaloy-2☆

Abstract Tritium diffusion measurements following tritium recoil injection into austenitic stainless steels and Zircaloy-2 exhibit slower hydrogen transport in the surface regions than in the bulk. Kinetics of tritium release from these materials indicate that the release is controlled by a diffusion coefficient that is two to three orders of magnitude lower than the bulk diffusion coefficient for tritium in stainless steel and seven to eight orders of magnitude lower in Zircaloy-2. A two-region classical diffusion model with different diffusion coefficients in each region has been developed which appears to adequately represent the surface data for short heating times. Release rates at long heating times are apparently influenced by trapping of hydrogen in surface films. The surface effects are shown not to be due to the helium which is injected into the specimens along with the tritium.

Influence of defects on rare-gas diffusion in solids☆

The diffusion of 133Xe in CsI was studied as an aid to understanding rare gas diffusion in more complex systems, such as nuclear reactor fuels. Diffusion at low gas concentrations and negligible radiation damage levels was measured by growing single crystals of CsI containing radioactive 133I which decayed to 133Xe (T12 = 20.8 h). Both the time rate of release and temperature dependence of the rare gas diffusion coefficient were consistent with classical diffusion solutions, giving: D = D0 exp (−Q/kT), D0 = (0.57+2.30−0.43) cm2/sec, Q = (1.01 ± 0.04) eV. Crystals containing high concentrations of defects exhibited trapping of the rare gas and anomalous diffusion kinetics. Trapped gas atoms tended to stabilize the defects and prevent their annealing during heating. Diffusion of fission gas recoiled into CsI specimen surface layers from an external fissionable source obeyed classical diffusion solutions at low fission recoil concentrations, while at high concentrations, radiation damage created traps which decreased gas diffusion rates. These traps differed significantly from natural defects in trap concentration, gas atom binding energies, and annealing characteristics. Increasing the gas concentration independent of radiation damage also lowered rare gas diffusion coefficients, showing that formation of small gas atom clusters also produced trapping. The results showed that classical rare gas diffusion could be obtained under ideal conditions and that distinctive trapping behavior with different characteristics could be associated with the presence of natural defects, radiation damage, and high gas concentration.

Hydrogen permeation through non-metallic solids

Abstract Hydrogen permeation rates through high-density sintered aluminum oxide and KT-silicon carbide tubes have been measured as a part of an ongoing program to obtain hydrogen transport parameters in non-metallic solids of interest to fusion reactors. The temperature and pressure ranges were 1200°C – 1450°C and 2–50 kPa respectively. The permeability of Al2O3 appears to be consistent with independent measurements of solubility and diffusivity, indicating no accelerated diffusion due to microstructural defects. The KT-SiC tube gave hydrogen permeability values somewhat higher than those of Al2O3, although both of these materials were found to have about two to three orders of magnitude lower hydrogen permeability than the least permeable of the refractory metals. If suitable fabrication techniques can be developed, both Al2O3 and SiC can be useful as coatings on metals to limit hydrogen permeation.

Diffusion of Xenon in Ceramic Oxides

Postirradiation measurements were made at 700° to 1500°C of the release of 133Xe recoiled into single‐crystal α‐Al2O3, BeO, MgO, and ZrO2 crystals during neutron irradiation in contact with UO2 powder. The release pattern from most of the specimens was identical: a high rate of loss of 133Xe for several minutes followed by a slower release for which the cumulative fractional release was linear with the square root of the heating time. Diffusion coefficients for 133Xe have been calculated from the slope of the release curve following the initial release. The magnitudes of the diffusion coefficients above 1100°C for α‐Al2O3, BeO, and MgO were nearly the same whereas the diffusion coefficients in ZrO2 were ten to a hundred times larger. The activation energy for 133Xe release from α‐Al2O3 above 1100°C was 64 kcal/g‐mole. The xenon release behavior has been compared to anion diffusion in these oxides.

Tritium diffusion through oxide surface films on niobium

Abstract Diffusion of tritium in niobium has been investigated using recoil implantation with particular emphasis on the role of surface oxide films. In conformity with previously reported data, bulk diffusion in niobium is found to be extremely rapid and surface oxide films significantly retard the release rate of tritium. The observed tritium release data indicate diffusion through cracks or defects in the oxide film below 500°C and transport through an intact film at temperatures higher than 500°C. The release rates are described in terms of a mathematical model with the necessary model parameters determined from experimental data. The oxide films could be removed with a reducing hydrogen atmosphere, thereby resulting in classical bulk diffusion controlled release above 700°C. Below 700°C the oxide layers were stable in hydrogen, presumably because of a moderately high oxygen potential resulting from minute amounts of moisture in the sample chamber.

Diffusion and trapping of tritium in grainboundaries of 304L stainless steel

Abstract Permeation rates of tritium through 304L stainless steel are calculated using a grainboundary diffusion model and measured values of bulk and grainboundary diffusion coefficients. Computed results show that the tritium release rate through grainboundaries should be significantly greater than through the lattice. These predictions, however, are not borne out by permeation experiments. In order to explain this discrepancy, autoradiographic examinations were carried out. The results show that although grainboundaries are indeed fast diffusion paths, their contribution to the overall permeation rate may be minimized by trapping in surface oxide films and possibly by low solubility of tritium within the grainboundaries.