The results of experimental studies on He embrittle­ment ofAuSS are broadly consistent with the concepts described here. However, the literature for neutron irradiations is much more limited than in the case of microstructural evolution and matrix swelling, especially for the most pertinent data from reliable in­reactor creep rupture tests. Indeed, there is little quan­titative characterization of grain boundary cavity and other microstructures for neutron-irradiated alloys. The most consistent trend for neutron irradiations is that high-temperature postirradiation tensile tests show significant to severe reductions in tensile ductility and creep rupture times and IG rupture along GBs.11

As noted above, there is a much more significant body of work for well-characterized high-energy He ion implantation studies. Helium can be preimplanted at various temperatures and further subjected to vari­ous postimplantation annealing treatments, prior to tensile or creep testing, or simultaneously with creep testing. The different modes of He implantation result in very different creep rupture behavior.90 Helium implantation during high temperature in-beam creep is perhaps the most relevant, controlled, and systematic approach to studying HTHE. A series of implantation studies carried out at the Research Center Julich in Germany, coupled with the models described above, are the most comprehensive and insightful examples of this research.90’91’99’100’192’199-201 Figure 22(a) shows the mean trend lines for tr versus applied stress for SA 316SS at 1023 K for in-beam creep’ at an implantation rate of 100 appm He/h’ compared with unimplanted controls.90 Clearly’ HTHE leads to a very large reduction in the tr especially at lower stress. The stress power is rк4 for the in-beam creep condition’ compared with к 9 for the unimplanted control. Figure 22(b) shows a corresponding plot for a Ti-modified AuSS (DIN 1.4970) in-beam creep tested at 1073 K.90 HTHE is observed’ but the magni­tude of the reduction in tr is less in this case. The stress power in-beam creep condition is rк 2.85 compared with 5.7 for the control. As expected’ HTHE also reduces er; in the case of Ti-modified steel’ er decreases from к 10% to 1%90 and for the annealed 316SS from more than 30% to 1% or less.99 Similar

comparisons for a test of these two alloys at 873 K also show severe reductions in tr and er in 316SS, whereas there is a much smaller effect in the Ti-modified AuSS.192 This work also showed that the Ti-modified alloy is much stronger in CW condi­tion and suffers only moderate HTHE.

Helium implantation of unstressed specimens at 1023 K to levels between 10 and 1000 appm was used to evaluate the critical bubble size resulting in rapid rupture in subsequent creep tests at the same tem — perature.9 A rapid drop-off in tr and er occurred between 300 and 1000 appm at a stress of 90 MPa. This He concentration correlated with an average grain boundary bubble size of ~13-17nm, which provides a reasonable estimate of the corresponding critical bubble size for these conditions.

However, it must be emphasized that in-beam creep tests do not ‘simulate’ neutron-irradiation con­ditions since the He implantation rates are highly accelerated and yield very high He/dpa ratios. Further, the duration of these tests is limited by schedules of the ion accelerators. The in-beam creep data show that the tr and er decrease with decreasing implantation rates scaling as « G^3 (Schroeder et at91 and see previous discussion). This may be due to the effect of GHe on the number density of matrix bubbles that help shield GBs from He accumulation, as suggested by in eqn [16]. Indeed, matrix bubble densities, Nb, in 316SS increase with the concentration of He preimplanted at 1023 K, scaling as « XH{2 at greater than 100 appm He.

The corresponding number of grain boundary bubbles, Ngb, is insensitive to XHe. 9 Of course, the scaling applies only to high implantation rates that result in significant He concentrations at all GHe. A more relevant scaling law would be based on the He required for creep rupture (XHer), which also scales with tr, as

XHer / tr / GHe [17]

The in-beam creep tests suggest that q« 1/3. Thus, for example, if tr = 10 h at XHe = 1000 appm (see above), for in-beam creep tests, this scaling predicts tr« 2700 h at XHe « 60 appm for He generated at fusion reactor rates of 200 appm year- . Note that this is not a reliable absolute estimate of the creep rupture time since the in-beam irradiation experi­ments involved very thin specimens (0.1 mm).

Figure 23 shows tr versus stress for in-reactor creep tests at 993 K under neutron irradiation. The corresponding ‘control’ curve at 993 K is based on a logarithmic interpolation between curves for

specimens preimplanted with up to 80 appm He at ambient temperature and tested at 973 and 1073 K. Note that data at 1073 K suggest that at such low — temperature preimplantation has little effect on tr. Severe HTHE is again observed in this case at stresses below about 200 MPa for in-reactor creep conditions.

The key results of these studies can be summar­ized as follows:

• High implantation rates of the order 100 appm h-1 result in the formation of a high number density of bubbles in the matrix and on GBs up to and in excess of 1015 m-2.

• Matrix and grain boundary bubbles form both as isolated cavities as well as in association with other features such as second-phase particles, disloca­tions and grain edges, and triple points.

• Essentially all the implanted He precipitates in bubbles.

• Matrix bubbles influence the amount of He flow­ing to GBs.

• The grain boundary bubbles grow stably with the addition of He, until some reach the critical size where they convert to stress-driven growing creep cavities.

• Postimplantation tests at 90 MPa and 750 °C for a wide range of He contents suggest a critical size of 13-17 nm.

• Creep cavity growth and coalescence kinetics are rapid, and tr is dominated by the time needed to establish a population of bubbles and grow (gas driven) them to the critical size.

• The creep rupture time is generally lower for implantation coincident with creep, compared to preimplantation followed by creep testing.

• The stress power r relating rupture time and the applied stress (tr / a^r) is decreased for in-beam and in-reactor creep tests, with r« 2-4 when com­pared with postimplantation and unimplanted values of r« 6-9.

• The tr and XHer decrease with decreasing He gen­eration rates with a scaling law « GH^.

• However, significant He is required to cause HTHE in all cases.

• Alloys with fine-scale matrix and grain boundary precipitates (e. g., TiC) that trap He in a larger number of smaller bubbles mitigate HTHE.

• The CBM also rationalizes the degradation of high-temperature fatigue properties at high He levels.91,99,204

The implantation studies show that HTHE is most severe in conventional AuSS, like 316SS, that contain only coarse carbides. Fine-scale TiC (and phosphide) phases that trap He in a high density of fine- scale matrix bubbles n90,192,203-205 provide greatly enhanced HTHE resistance. Key issues are optimiz­ing the ability of the precipitates to trap He in small bubbles and ensuring their thermal and irradiation stability needed for long-term service. Note that a fine-scale matrix also increases the strength of AuSS alloys, thus enhancing creep constraint reductions of cavity growth rates; and fine grain boundary phases can also impart further resistance to grain boundary cavitation. Indeed, Maziasz and coworkers extended these concepts to developing a new series of AuSS alloys with remarkable unirradiated creep strength (in terms ofboth creep rates and rupture times) based on precise control of microalloyed fine-scale matrix and grain boundary phases.206

A number ofstudies have also shown that the FMS are very resistant to HTHE as well as void swelling.12,13,192,201,207 The most obvious explanation is that the sink densities and numerous trapping sites for He keep bubbles finely distributed and protect prior austenite GBs from He accumulation.176 Lath boundaries may be especially effective if, due to their special nature, they are effective in trapping He in small bubbles, but at the same time, resistant to cavity growth by vacancy accumulation. However, it has been suggested that the bubble microstructures are not that dissimilar in AuSS and FMS and that at least part of the difference in the HTHE sensitivity is that the FMS are inherently weaker at the same tempera­tures and that the corresponding lower grain bound-

12,199

ary stresses increase m.