Version of Record: https://www.sciencedirect.

com/science/article/pii/S002231151731471X
Manuscript_5ed90db852444ce4564853a4cae55ccf

THE INFLUENCE OF HELIUM ON CAVITY EVOLUTION IN ION-IRRADIATED T91

Anthony M. Monterrosa1*, Zhijie Jiao1, Gary S. Was1

1
University of Michigan, Ann Arbor, MI 48109 USA

Abstract

The effect of helium pre-implanted in concentrations spanning 0 to 1000 appm on the swelling
evolution in ferritic-martensitic alloy T91 was explored. Irradiation with 5 and 4.4 MeV Fe2+
ions at 460°C up to damage levels of 350 dpa were performed with pre-implanted helium levels
of 0, 1, 10 and 1000 appm. Low concentrations of helium were found to enhance nucleation at
low damage levels. However, at high damage levels, differences in nucleation became
insignificant, and nearly identical cavity distributions were created. At low helium levels, the
cavity evolution was facilitated by similar cavity sink strengths. High helium concentrations
reduced the critical cavity radius and resulted in the formation of a high-density bubble
population. The bubbles were < 4 nm and remained stable even up to high damage levels, while
the growth of larger cavities was suppressed. The suppression of swelling was found to be a
result of helium trapping or alteration of cavity and dislocation bias. Dislocation and precipitate
evolution was also studied and exhibited relative insensitivity to helium concentration, and did
not significantly influence the cavity evolution.

Keywords: helium, ion irradiation, ferritic-martensitic, swelling, void, cavity

*Corresponding Author: Anthony M. Monterrosa, 2355 Bonisteel Blvd, Ann Arbor, MI, 48109,
[email protected]

1. Introduction

Of the Generation IV reactor concepts, the sodium fast reactor is the most mature and
closest to deployment. Components in fast reactors will experience damage levels in excess of
200 displacements per atom (dpa). To reach these high damage levels, ion irradiation has been
used as an effective surrogate to neutron irradiation [1–3]. This method is ideal for assessing the
irradiated microstructure of promising candidate alloys such as ferritic-martensitic steels, which
exhibit high strength at elevated temperatures, swelling resistance, and low activation. However,
damage induced with ion beams does not produce transmutation gasses such as helium, produced
in reactor. Therefore, helium is often introduced via pre-implantation or co-injection.
The ability of helium to enhance cavity nucleation across multiple alloy systems is well
documented [4–9]. However, studies on the interaction of helium within the ferritic-martensitic
microstructure are very limited, and studies at high damage levels are even fewer. Smidt et al.
[10] studied the swelling behavior of ferritic-martensitic steels EM-12 and HT9 using heavy
ions. Prior to irradiation, a single helium level of 1 appm (at the peak) was implanted into the
surface of the sample. A 30 MeV helium beam was degraded using an aluminum sheet to obtain
a helium profile in the shape of a half-Gaussian with a 11.8 µm full width half. Irradiations with
Fe2+ ions were performed up to 250 dpa at temperatures between 450 and 650°C in an effort to
determine the temperature dependence of swelling. EM-12 and HT9 exhibited swelling rates of
~0.011 %/dpa and ~0.017%/dpa at their peak swelling temperatures. The authors did not

© 2018 published by Elsevier. This manuscript is made available under the Elsevier user license
https://www.elsevier.com/open-access/userlicense/1.0/
consider the effect of helium concentration on swelling, or the effect of varying temperature on
the helium behavior.
Kai and Kulsinki [11] irradiated HT9 specimens up to 200 dpa with either 0 or 100 appm
pre-implanted He. Cavities were observed only in the 100 appm He samples, suggesting that He
was required to nucleate a significant cavity population. Since only one damage level was
considered, no conclusion could be made about the effect of He on growth.
Toloczko et al. [12] conducted ion irradiation on ferritic ODS MA957, HT9, and ferritic
alloy EP-450 up to 500 dpa. Without the aid of helium, swelling rates of 0.2%/dpa were found
for both HT9 and EP-450, with total swelling exceeding 30% by 500 dpa for HT9, suggesting
that helium was not required to facilitate significant swelling. The swelling rate in this study was
a factor of ~10x higher than in similar studies.
A study on ferritic-martensitic steels CNSI (nominally 9Cr) and CNSII (nominally 12Cr)
by Wang et al. [13], explored the effect of 10 and 100 appm He implantations up to 188 dpa at
460°C. Within CNSI, higher helium levels resulted in higher swelling, however no change in the
swelling rate of ~0.02%/dpa was observed regardless of helium content. It was concluded that
helium served to decrease the incubation period, but did not have a significant effect on the
swelling rate.
The most complete work on the effect of helium to date performed by Getto et al. [14]
studied the effect of pre-implanted helium levels in HT9 of 0, 1, 10 and 100 appm He. A
difference in the swelling rate was observed between the 1 and 10 appm He conditions at 460°C,
and between the 0, 10 and 100 conditions at 440°C, with higher amounts of helium generally
resulting in a higher swelling rate. The effect of helium at different temperatures was not
considered.
While these studies demonstrate that the addition of helium enhances nucleation and
increases cavity densities, the effect of helium on the cavity growth rate and evolution at a high
level of damage remains unclear. Across these studies, helium levels ranging from 0 to 100 appm
have been studied, all with variable behavior in swelling. Additionally, variations in alloy,
irradiation temperature, damage level, and helium implantation method further complicate the
interpretation of the effect of helium. This work seeks to provide a clear understanding on the
effect of pre-implanted helium by performing a comprehensive study on a single alloy, at a
single temperature over the span of four orders of magnitude in helium implantation.

2. Experiment

2.1. Sample Preparation

The alloy used for this work was T91, heat C2269. The composition in weight percent is
shown in Table 1. This alloy was given a heat treatment consisting of a 46 min anneal at
1066°C, followed by an air cool, and tempering at 790°C for 42 min, followed by an air cool.
Prior to irradiation, samples of T91 were cut in the form of 1.5 mm x 1.5 mm x 20 mm
bars using electrical discharge machining. The bar samples were then hand-polished using a
variable speed grinding wheel with silicon carbide grinding paper. Grits of 240, 320, 400, 600,
800 and 1200 were used to polish the sample. The surface was further prepared for irradiation by
polishing with a cloth pad and diamond slurries of 1 µm and 0.25 µm to provide a mirror-like
finish. To remove any damage layer induced by mechanical polishing, the samples were then

2
electropolished in a solution consisting of 10% perchloric acid and 90% methanol at -40 to -
50°C. The applied voltage was -40V for approximately 20 s.
A 100 nm thick coating of alumina was applied by atomic layer deposition (ALD) to the
surface of the samples to prevent carbon uptake during ion irradiation [15]. ALD is a thin film
deposition technique that utilizes gaseous precursors to deposit the film (in this case Al2O3). The
T91 samples were heated to 150°C in an inert argon environment. Trimethylaluminum (TMA,
Al(CH3)3) was bubbled into the chamber and reacted with hydroxyl (OH) groups on the surface
of the substrate. The TMA was then pumped out of the chamber and water was bubbled into the
chamber to react with the methyl (CH3) groups to produce alumina and methane. The water was
then pumped out of the chamber. This process was repeated for 980-1000 cycles to achieve a 100
nm thick film.

2.2 Helium Implantation

Prior to irradiation with Fe2+ ions, samples were implanted with helium to levels of 0, 1,
10, 100, and 1000 appm. One sample of T91 was used for two helium conditions. Half of the
sample was implanted with one helium level and the other half was implanted with a second
helium level. A National Electrostatics Company 400kV implanter at the Michigan Ion Beam
Laboratory was used to implant the helium. The helium implantations were performed with a
raster-scanned beam of He+, with the samples at room temperature. Energies of 80, 140, 220, 310
and 420 kV were used to obtain an approximately flat damage profile (±10%) over a depth range
of 300-1000 nm from the sample surface. The implantation profile was estimated using SRIM
(Stopping Range of Ions in Matter) in quick Kinchin-Pease mode [16,17] with a displacement
energy of 40 eV. The resulting profile predicted by SRIM after He implantation at various
energies is shown in Figure 1.
For the 1 appm helium pre-implantation, a damage rate of approximately 8 x 10-7 dpa/s
was used. For the higher helium levels, the ion beam current was increased by approximately an
order of magnitude for each level. At the highest helium level (1000 appm), the damage caused
by pre-implantation was estimated to be ~0.02 dpa. He implantation at this damage level and at
room temperature were not expected to induce significant changes to the pre-existing
microstructure, aside from the addition of helium.

2.3. Ion Irradiations

The pre-implanted T91 samples were irradiated to a high dpa using a 3 MV NEC
Pelletron accelerator at the Michigan Ion Beam Laboratory. Irradiations in this work were
performed using a focused and raster-scanned beam of either 5 MeV or 4.4 MeV Fe2+ ions.
High-speed raster-scanning provided full coverage of the irradiated area every 3.92ms.
All irradiations in this work were performed at a temperature of 460°C. J-type
thermocouples were used to calibrate a 2-D FLIR® Thermal Imaging System so that the
temperature could be monitored on at least three areas of interest on each sample throughout the
entire irradiation. Temperature control was maintained within a 2-sigma value of ±6°C by
utilizing an electronic cartridge heater and air cooling of the back of the irradiation stage. The
pressure of the vacuum in the irradiation chamber was maintained at <10-7 torr (< 1.3 x 10-5 Pa).
The damage level for each experiment was determined at a depth of 600 nm from the
surface of the alumina coating using SRIM in quick Kinchin-Pease mode. Figure 1 shows the

3
damage rate and implanted Fe profiles with depth using 5 MeV Fe2+ ions in T91 with a 100 nm
alumina coating. Experiments at nominal damage levels of 50, 150, and 300 dpa (fluence of 1.2 x
1017, 3.6 x1017, and 7.2 x 1017 ions/cm2 respectively), were performed for all helium pre-
implantation conditions (0, 1, 10, 100, 1000 appm). The experiments performed to 50 and 150
dpa utilized a 4.4 MeV Fe2+ beam, while the experiment performed up to 300 dpa utilized a 5.5
MeV beam. Regardless of beam energy, a damage rate of 1.4 x 10-3 dpa/s at 600 nm depth was
maintained for all experiments by controlling the ion beam current.

2.4 Post-Irradiation Characterization

Post-irradiation characterization was performed at the Michigan Center for Materials

Characterization (MC)2. After irradiation, TEM foils were prepared using a focused-ion beam
(FIB) in either a FEI Helios NanoLab™ Dual Beam™ or a Nova NanoLab™ DualBeam™. Due
to the shallow penetration depth of the Fe2+ ions, the FIB lift-out method was used to prepare
TEM samples. TEM specimens were thinned to 100 nm or less using the focused ion beam.
Imaging for all of the TEM specimens in this study was performed using a JEOL 2100F
Cs Corrected Analytic Electron Microscope (AEM) at the Michigan Center for Materials
Characterization (MC)2. The JEOL 2100F is a 200keV microscope that operates mainly in
scanning TEM (STEM) mode. Imaging of each TEM specimen was performed using STEM
mode, capturing a high angle annular dark field (HAADF) image and a bright field (BF) image
simultaneously. The contrast in STEM HAADF mode arises mainly from thickness and “Z-
contrast,” which is dependent on the atomic number of the material, making it very suitable for
imaging cavities. Additionally, HAADF images ensure more accurate size measurements than
traditional BF through-focused images, which can cause size variations depending on degree of
focus. For adequate depth profiling, images were taken at approximately 500kx magnification
that allowed for a field of view of about 2 µm x 2 µm. This permitted characterization of
microstructural features in defined regions within 400-800 nm from the surface of the sample in
100 nm bins. In many cases, higher magnification images (up to 2Mx) were taken to acquire
more detailed views of smaller cavities and microstructural features.
While the nominal damage levels of 50, 150, and 300 dpa for a given irradiation were
taken at a depth of 600 nm, other depth regions were also characterized. From each experiment,
three distinct depths from each experiment were characterized, namely 400-600 nm, 500-700
nm, and 600-800 nm from the surface of the sample. These ranges avoid any possible influence
of the surface region and of the injected interstitial [3], and remain in the desired helium content
range. Three depth regions from three irradiation experiments resulted in viable data from nine
damage levels: 45, 50, 58, 134, 150, 175, 276, 300, and 356 dpa. It is important to note that the
damage rate for the three depth regions varies slightly. However, the variation between damage
rate in each region was minimal with 1.2 x 10-3 dpa/s, 1.4 x 10-3 dpa/s, and 1.6 x 10-3 dpa/s for
the 400-600, 500-700, and 600-800 nm depth regions, respectively.
The thickness of each TEM specimen was measured using EELS (electron energy loss
spectroscopy). This method was used to estimate the thickness of the sample by measuring the
amount of electron energy loss as the beam passed through the sample. The error in a logarithmic
fit of the inelastic mean free path of the electrons through the sample was calculated.
Measurements were taken on at least six different areas across the sample. The average of these
thickness measurements taken from the EELS map was used as the nominal thickness of the
region where the measurements were made.

4
 Cavities were characterized through the depth of the samples. The open source ImageJ
software was used to divide the image into depth bins of 100 nm width. The cavities in each bin
were counted and measured. During counting, the HAADF images were cross-referenced with
the bright-field images to ensure the contrast was consistent with cavities. The volume of each
individual cavity was calculated, assuming spherical cavities. The sum of the volume for each
cavity was tallied for each depth bin. This value was effectively the change in volume, ∆V, of
the bin. Swelling was calculated as the change in volume divided by the original volume, so the
swelling of any particular bin in %, could be expressed as:
π
∑
N
d3
∆V 6 i =1 i
Swelling (%) = ∗100 = ∗100 , (1)
Vbin − ∆V π N 3
l ∗ w ∗ δ − ∑ i =1 di
6
where l is the length of the TEM specimen, w is the width of the bin, δ is the thickness of the
TEM specimen, N is the number cavities in the bin, and di is the diameter of the ith cavity. For
the nominal swelling value within the 500-700 nm depth region, the bin size was increased to
200 nm and the volume change calculation included cavities from both the 500-600 nm and the
600-700 nm bins. Average cavity diameter and number densities were also determined as a
function of depth, along with cavity size distributions.
The error in cavity measurements was determined by considering the error in thickness
and the error in cavity diameter. The thickness fit using the EELS zero-loss method has an error
of 10% [18]. Errors in cavity diameter were determined by the pixel resolution of the TEM
image (1 nm), resulting in a 1 nm error across all sizes. The errors in density and size were
applied to swelling using standard error propagation. A similar error methodology was applied to
the characterization of precipitates and dislocations.
While the focus of this work is on the cavity evolution, the dislocation and precipitate
microstructure were also characterized to determine whether their behavior would influence the
cavity results. Two types of loops exist in ferritic-martensitic alloys: sessile b<100> loops
aligned on the [100] planes and ½ b<111> loops. On-zone bright field STEM imaging was used
for simultaneous imaging of all possible diffraction vectors, allowing for viewing of b<100>, ½
b<111> loops, and dislocation lines. In practice, the loops were imaged in STEM mode just
slightly off of the [100] zone axis so that perpendicular loops would not appear as lines, and a
double line contrast could be seen.
Precipitates were also characterized for a number of conditions in this study. Beyond pre-
existing M23C6 carbides, no additional carbides were observed. Bright field STEM combined
with EDS mapping were used to identify the presence of any other precipitates, specifically Ni-
Si rich clusters. Their size and density were measured in the 500-700 nm depth region only. Size,
density, and volume fraction were determined in a manner similar to that for cavities.
Sink strengths for cavities, the dislocation network, dislocation loops, and the Ni-Si rich
precipitates were also calculated. The cavity sink strengths were calculated as 4π r ρ , assuming
spherical cavities where r is the average cavity radius and ρ is the cavity density. The Ni-Si rich
precipitates were also assumed incoherent and spherical, and their sink strength was calculated in
a similar manner. The dislocation network density was calculated using a method established by
Smith et al. [19] utilizing an equidistant circular grid, and counting the number of intersections
of dislocation lines with the grid. The dislocation loops were converted into line length by
calculating their effective surface area, and then multiplied by their density to achieve a sink

5
strength, calculated as 2π d ρ d , where d is the loop diameter and ρ d is the loop density. The
addition of loop and network sink strengths determined the total dislocation sink strength.

3. Results

Cavities were observed at all combinations of dpa and helium levels within the analysis
regions. Figure 2 shows representative HAADF STEM images of the T91 microstructure at 300
dpa and at all levels of helium pre-implantation. A generous population of cavities is seen within
the helium-implanted and ion-damaged region. Figure 3 shows how the cavity size, density, and
swelling in T91 evolve at various damage and helium pre-implantation levels. Increasing the
amount of helium from 0 to 1000 appm at any damage level resulted in a decrease in diameter
and an increase in cavity density. The 0 appm He condition always exhibited the largest cavity
size, which decreased monotonically with He up to 1000 appm. The lower helium samples (0, 1,
and 10 appm) exhibited the highest levels of swelling, while the swelling in the 100 and 1000
appm conditions was greatly suppressed. Table 2 summarizes the results of the cavity
measurements as a function of depth, including the average size, density, swelling, and sink
strength at all damage levels and helium levels of 0, 1, 10, 100, and 1000 appm. The following
sections present results in the various dpa / He regimes.

3.1 Low damage, low helium regime

At low damage (45-58 dpa) and low helium levels (0, 1, 10 appm He), similar cavity
sizes were observed. The 0 appm He condition retained the highest size at about 10.3 nm, closely
followed by the 1 and 10 appm conditions (8.9 and 8.8nm, respectively). The main difference
between these helium conditions at low damage levels was in the cavity density. Density
increased with helium content, with the biggest difference occurring between the un-implanted
case, and the 1 appm He case. The 10 appm He condition exhibited a density of ~25 x 1020 m-3,
the 1 appm a slightly lower density of ~20 x 1020 m-3, and then the 0 appm He case showed a
drop-off to about half that value to ~11.5 x 1020 m-3.

3.2 High damage, low helium regime

As the damage level was increased to the high damage regime (134-356 dpa), the cavity
size distributions of the 0, 1, 10 appm He conditions began to converge. Figure 4 shows how the
cavity distributions evolved with damage. The size distributions of these three helium conditions
all followed a single, roughly Gaussian profile. Differences between the helium levels are
noticeable at low damage (50 dpa), but by 300 dpa the distributions appear very similar. Despite
having different helium levels, very similar final cavity densities and sizes were observed
between the 0, 1, and 10 appm conditions at 300 dpa (20.1, 21.6, and 18.9 x 1020 m-3, and 29.3,
28.2, and 26.4 nm, respectively). Figure 5 shows representative images of the cavity evolution
of the 0, 1 and 10 appm He conditions. Differences in the cavity densities are apparent in the 50
dpa condition, however by 300 dpa, the microstructures and distributions were almost
indistinguishable (Figure 4).
Throughout the entire damage range, the 0 appm He condition maintained a slightly
larger cavity size (Figure 3a). However, by 150 dpa the cavity density in the 0 appm He case

6
was indistinguishable from that of the 1 and 10 appm conditions (Figure 3b), confirming that it
had been in the nucleation phase at 50 dpa. Beyond 150 dpa, no significant increases in cavity
density were observed for the 0, 1, and 10 appm helium conditions. In fact, over the entire
damage range studied, only the 0 appm helium condition showed any sign of increasing cavity
density. For any amount of He, there was no significant change in density with damage (Figure
3b), indicating that in any pre-implantation condition, the final cavity density was determined
very early in the irradiation (< 50 dpa).

3.3 High helium regime

Inspection at high magnifications revealed the presence of an extremely high density of

very small cavities in the 100 and 1000 appm He conditions at all damage levels, Figure 6.
Cavities in this size range (<~4 nm) were not observed in any helium condition lower than 100
appm. Because of local variation, characterization of these small cavities in the 100 and 1000
appm He conditions was difficult to perform as a function of depth. Therefore, measurements of
the maximum observed cavity densities in the 500-700 nm region were used to estimate the
maximum swelling contribution of the small cavities. The resulting densities and swelling values
are summarized in Table 2 for the 100 and 1000 appm conditions.
The small cavities present in the 100 and 1000 appm He conditions were likely helium
bubbles. The evolution of the bubble and cavity distributions with damage is shown in Figure 7
(note logarithmic scale on Y-axis). Both the 100 and 1000 appm He conditions exhibit high
cavity densities of sizes < ~4 nm, with the 1000 appm He condition having a noticeably higher
density than that of the 100 appm He. These small cavities appeared to be insensitive to damage,
as their densities remained relatively unchanged across the entire damage range (Figure 3b). As
the damage increased to 300 dpa, the cavity distributions evolved from unimodal to bimodal. At
the larger sizes, it is evident that a small amount of growth occurred, as the tails of the
distributions in Figure 7 extended to higher sizes at higher damage levels. The limited growth
was slightly higher in the 100 appm He condition increasing from 3.2 nm to 5.7 nm, while the
1000 appm He only increased from 1.2 nm to 1.8 nm over the addition of ~300 dpa. These very
low amounts of growth resulted in minimal swelling, even at high damage levels (Figure 3c).

3.4 Dislocations and precipitates

The loop diameter and densities, along with representative images, at 50, 150, and 300
dpa are shown in Figure 8 for the 10 appm He condition. For both b<100> and ½ b<111> loops,
a higher density was observed at 50 dpa, which dropped by 150 dpa and stabilized by 300 dpa.
Both types of loops also experienced a small amount of growth from 50 to 300 dpa. Overall,
b<100> loops were observed in a higher density by 3-4 times, and about twice the size of ½
b<111> loops at any given damage level.
Type <100> loops were also measured as a function of helium content. Figure 9 shows
the variation in loop size and diameter with helium content at 150 dpa. For all of the pre-
implanted cases, the loop diameter ranged between 21 and 29 nm at 150 dpa, with no obvious
trend with helium content. A similar behavior was observed for the loop density, which ranged
between 19 and 25 x 1020 m-3. The 0 appm He condition exhibited a slightly larger loop size of
36.3 nm.

7
 The dislocation network showed an even higher insensitivity to helium. Across all helium
contents, the dislocation network density ranged from 2.2 to 4.2 x 1014 m-2, well within any
measurement error.
Figure 10a shows the <100> loop line length, dislocation network, and total line length
as a function of helium content at 150 dpa. The total line length remained insensitive to He
content, which varied between 6-7 x 1014m-2 for the pre-implanted cases, and was slightly higher
for the 0 appm He condition. A comparison of the total line length to the cavity sink strengths at
150 dpa for all helium levels is shown in Figure 10b. The ratio of sink strengths is above 1 for
all but the 1000 appm He condition, indicating a dislocation dominant microstructure. For 1000
appm, the ratio falls slightly below 1, suggesting a cavity dominant microstructure. All of the
results for the dislocation loops and network are summarized in Table 3.
The 100 nm alumina coating on the surface of the T91 samples prevented any significant
uptake of carbon, and thus no additional carbides formed. The presence of Ni-Si rich clusters
was identified by EDS. Figure 11 shows an EDS map of the Ni-Si rich precipitates. The
stoichiometry was not confirmed and diffraction patterns could not be obtained, but these
clusters are consistent with G-phase precipitates or precursors. The appearance of G-phase in
irradiated ferritic-martensitic steels has been commonly reported in previous studies [20–22].
The evolution of Ni-Si rich clusters (G-phase or G-phase precursors) with damage and 10 appm
He implanted is shown in Figure 12 and Table 3. As the damage level increased, the precipitates
exhibited slight coarsening as the size increased and the density decreased, resulting in an
increase in the total volume fraction up to about 0.04% by 300 dpa. The sink strength for these
clusters remained low, ranging from 0.36 to 0.58 x 1014m-2.

4. Discussion

The cavity evolution depended highly on both the amount of implanted helium and the
damage level. The following discussion addresses the influence of helium in causing sink
strength-moderated behavior at low helium levels (0-10 appm), and on the formation of a
bimodal distribution and suppression of growth at high helium levels (100 and 1000 appm).
Lastly, it is shown that other microstructural features such as dislocations loops and network, and
G-phase precipitates had a negligible influence on the cavity evolution.

4.1 Sink strength-moderated cavity evolution

At low helium levels (0, 1, and 10 appm), the cavity behavior was controlled by the
cavity sink strengths. The effect of helium in promoting cavity nucleation is clearly seen in
Figure 3b. At the lowest damage level, the swelling and cavity sink strengths were dominated by
the cavity density. Therefore, an increase in cavity volume occurred with increasing He from 0,
to 1, to 10 appm He, (0.091%, 0.107%, and 0.124%, respectively at 50 dpa). At 50 dpa, the sink
strength of the 10 appm He condition was almost double that of the 0 appm He condition. Figure
13 shows how the cavity sink strengths varied with He content at different damage levels. The
slightly higher size of cavities in the 0 appm He case was not large enough to overcome the ~2x
density for the 1 and 10 appm He conditions.
The presence of pre-implanted helium served to create nucleation sites such that a high
density of cavities was established early on in irradiation. However, the 0 appm He condition did
not have the aid of helium to nucleate, and thus the formation of cavities (or voids in this case),

8
occurred purely by vacancy agglomeration. The 0 appm He condition experienced a much slower
nucleation rate such that at 50 dpa it was still in the nucleation-dominant regime.
By 150 dpa, the 0 appm condition achieved the highest swelling of any of the helium
levels studied. Cavities in the 0 appm He case grew at a faster rate, as seen by the larger cavity
sizes (Figure 3a), while simultaneously nucleating new cavities over a longer period (Figure
3b). The same number of vacancies produced by irradiation were distributed over a higher
number of cavities in the pre-implanted cases, resulting in lower growth rates compared to the 0
appm He case. This difference in growth rates resulted in a cross-over point in swelling between
50 and 150 dpa (Figure 3c), where the 0 appm He case achieved a similar cavity sink strength as
in the 1 and 10 appm conditions (Figure 13). At 150 dpa, cavity sink strengths for the 0, 1, and
10 appm He conditions were essentially the same.
After 150 dpa, no further increases in cavity density occurred, suggesting that vacancies
were being absorbed into cavities at nearly the same rate across the three helium levels.
Nucleation of new cavities was suppressed, and the sink strength of the established cavity
microstructure was strong enough to assure a net absorption of vacancies causing growth in
cavity size. Ultimately, similar sink strengths at 150 dpa with no further increases in density
resulted in similar growth behavior and a convergence of the 0, 1, and 10 appm cavity
distributions by 300 dpa. Figure 4 shows how very different size distributions between the 0, 1,
and 10 appm He conditions at 50 dpa converged to nearly identical distributions by 300 dpa.
Small additions of helium did not change the final distribution or total amount of vacancies in
cavities –it only provided more sites for the cavities to nucleate.
Only a small distinction could be made between additions of 1 and 10 appm helium, as
both had achieved their highest densities by the lowest damage level. However, it is suspected
that the 10 appm He condition nucleated more quickly than the 1 appm condition, by providing a
higher density of helium-assisted nucleation sites. This difference in nucleation was not
significant enough to establish different sink strengths, and thus the cavity evolution of the 1 and
10 appm conditions remained largely the same. A swelling rate of ~0.02%/dpa was maintained
for these low helium conditions. This swelling rate compares well with ion irradiations in
ferritic-martensitic steels with 10 appm He performed by Getto et al [14,20] and Wang et al [13],
and compares well to a study of HT9 by Toloczko [12], which contained no helium. Toloczko
did observe a higher swelling rate (~0.2%/dpa), but this only occurred after 400 dpa, while the
swelling rate up to 300 dpa is much lower and exhibits similar total swelling to the work
presented here.
Overall, the role of small additions of helium was simply to accelerate cavity nucleation,
but not to alter the final cavity microstructure. The difference in cavity microstructure was
largest at low damage levels, where the swelling behavior was dominated by nucleation.
However, similar cavity sink strengths between the 0, 1, and 10 appm He conditions ultimately
led to nearly identical size distributions by 300 dpa.

4.2 Formation of bimodal distributions

The cavity behavior in the high helium conditions (100 and 1000 appm) was
fundamentally different than that of the low helium conditions. A significant bubble density at
sizes < 4 nm (Figure 7) and a clear suppression of growth resulting in low swelling (Figure 3c)
were observed at all damage levels. Hishinuma and Mansur [21] postulated that bubble
formation and the eventual evolution of a bimodal distribution is due to a reduction in the critical

9
radius in the presence of helium. Bubbles below the critical radius are able to remain relatively
stable and insensitive to damage as long as they retain a particular number of helium atoms.
The analysis from [21] can be applied to the T91 system to create a map in which a
bubble population would be expected to be stable and eventually result in a bimodal distribution.
The critical radius of a cavity as a function of its helium gas pressure can be expressed by:
 kT  Z ic Z vd  Cv Z ic Z vd  
rc = 2γ /  pg +
crit
ln 1 −  +  , (2)
 Ω  Z vc Z id  Cvo Z vc Z id  

where γ is the surface energy, p g is the helium gas pressure of the cavity, k is Boltzmann’s
constant, T is irradiation temperature, Ω is the atomic volume, Z ic is the cavity capture
efficiency for interstitials, and Z id is the dislocation capture efficiency for interstitials, and Z vc
and Z vd are the capture efficiencies for vacancies and lastly, Cv / Cvo is the vacancy super
saturation. The gas pressure in the cavity, p g , can be expressed by the number of helium atoms
in the cavity:
4 
pg = ng kT /  π rc3 − ng B  , (3)
3 
where ng is the number of helium atoms, rc is the cavity radius, and B is the Van der Waals
coefficient for helium. Using these equations, a cavity stability map (Figure 14) was created that
shows the dependence of the critical cavity radius on the actual cavity radius and the number of
gas atoms in the cavity. The key parameters used for this calculation are summarized in Table 4.
In Figure 14, the estimated critical radius when no gas is present ( ng = 0 ) is shown by
the horizontal dashed line at approximately 2 nm. When helium gas atoms are added to a cavity,
the critical radius falls below this 2 nm line, indicating that the presence of helium results in a
smaller critical radius than with no gas. As higher amounts of helium are added to the cavity, the
effect of lowering the critical radius extends to larger cavity sizes.
The diagonal dashed line in Figure 14 denotes the equality of the critical radius and the
actual cavity radius. The intersection point of any of the curves with the diagonal line indicates a
point of stability. At any point below the diagonal line, a cavity with a given number of helium
atoms will grow along its respective curve until it reaches the diagonal. At that point, the cavity
cannot grow any further because at all points above the diagonal line, the critical radius is greater
than the actual cavity radius. This results in a relatively stable bubble at a size lower than the
critical radius without gas (~2 nm). For the T91 system, curves describing cavities with less than
~400 He atoms contain intersection points with the diagonal line, indicating that < 400 He
atoms/cavity are required to form a bimodal distribution. In practice, cavities with radius smaller
than 0.5 nm were not observed or could not be resolved reliably under STEM. Using this 0.5 nm
lower limit, a more realistic threshold for the creation of a bimodal distribution would be a range
of ~50-400 He atoms/cavity. The greyed area in Figure 14 highlights this range.
To determine whether a bimodal distribution is reasonable to expect in the 100 and 1000
appm He conditions, an estimation of the total amount of He in bubbles was made. Assuming
that the smaller size groups (< 4 nm) were entirely equilibrium pressurized bubbles, Eq. 3 was
set equal to the cavity surface energy of an equilibrium bubble, 2γ/r, and integrated over the size
distribution to determine the possible total number of gas atoms. For the 1000 appm He

10
condition, the small size group contains approximately 380 He atoms per cavity, which
accommodates at most ~400 appm of He. The 100 appm He distribution with ~1000 He atoms
per cavity accommodates ~84 appm of He. Thus, both the 100 and 1000 appm He conditions are
capable of satisfying the required He threshold (~50-400 He atoms/cavity) to form a stable
bimodal distribution. High amounts of helium allow for sufficient reduction of the critical cavity
radius to establish a bubble-dominated microstructure.

4.3 Helium-suppressed cavity evolution

In the 100 appm and 1000 appm conditions, a suppression of total swelling was observed.
Suppression of swelling with high helium levels has often been attributed to the development of
a very high cavity sink strength [7,22,23]. At a sufficiently high cavity sink strength, it is
expected that both vacancies and interstitials would annihilate in equal numbers, thus resulting in
a suppression of growth. However, based on Figure 13, the 100 appm condition exhibits a very
similar cavity sink strength as the 0, 1, and 10 appm conditions, which did not experience
significant suppression of growth, suggesting that sink strength alone cannot account for the
suppression of swelling. In fact, such suppression would only occur if cavities were the dominant
sink in the system. At 150 dpa in the 100 appm He case, dislocations are still the dominant sink
as shown in Figure 10b, where the dislocation sink strength is ~3x higher than the cavity sink
strength. Similar cases were reported in stainless steel [9,24] where cavity sink strength alone
was not sufficient to account for swelling suppression. Therefore, other mechanisms must
contribute to swelling suppression in the presence of helium.
Hishinuma [21,24] suggested that the presence of helium could lower the bias of the
system and result in reduced growth. Small cavities or helium clusters could serve as obstacles
for dislocation climb, thus reducing their effectiveness in capturing interstitials and reducing
vacancy supersaturation. Since growth is bias-driven, this effect would preferentially influence
the larger size group.
Alternatively, elasticity theory and kinetic Monte Carlo, and molecular statics models
have demonstrated that cavities of very small size do not behave as neutral sinks, and instead are
biased toward interstitials [25–28]. For bubbles up to a size of 4 nm, this bias toward interstitials
can range from 10-20%. It is typically assumed that dislocations are heavily biased toward
interstitials, with bias factors reported anywhere from 1.01 to 1.25, while cavities, precipitates,
and grain boundaries are neutral sinks [29–31]. Considering the cavity bias effect, a significant
population of small bubbles may compete with dislocations for interstitials. If this is the case,
then the role of helium is to create a distribution of small cavities whose growth is restricted
(stabilized) by this cavity bias and remain relatively insensitive to damage. Because of the bias
toward interstitials, the net vacancy flow to cavities is reduced, resulting in a suppression of
swelling. The residual vacancies likely still flow to the bubbles due to the high bubble sink
strength or are eliminated via recombination due to helium trapping of vacancies. Only minimal
bubble growth is observed because these vacancies are diluted over a very high density of
bubbles. The vacancies which do make it to bubbles are less likely to be emitted due to strong
binding with helium. The bias toward interstitials prevents bubble growth while helium prevents
shrinkage, so the bubble population remains relatively stable. Molecular statics calculations by
Kohnert et. al [28], showed that the bias is still present in helium-filled bubbles as long as the
bubbles are not over-pressurized.

11
 A helium trapping mechanism has also been proposed [7,32–37] as a possible method of
swelling suppression. Helium within the matrix binds with vacancies and reduces the vacancy
mobility. This prevents the vacancies from diffusing to other sinks (such as cavities) and
enhances recombination, which will tend to suppress swelling. This mechanism may also serve
to eliminate residual vacancies formed by the bubble bias. As mentioned before, the small size
group of the 1000 appm He distribution could accommodate at most ~400 appm of He. This
means that ~600 appm He is elsewhere in the lattice. Even in the assumption that all cavities
observed in the 1000 appm He condition at 50 dpa were equilibrium bubbles, only ~800 appm
He could be accommodated in the entire cavity distribution. A similar estimation of helium
content in bubbles was performed by Chen et al. [38] in a high entropy alloy. The authors
observed that the helium content in bubbles could only account for up to 51.4% of the implanted
helium. It is likely that not all of the implanted helium accumulates in observed bubbles, and
instead contributes to swelling suppression via a trapping mechanism.
In the current study, there was no evidence that helium escape was occurring. Helium is
known to be a very fast diffuser in iron with a migration energy of only ~0.06 eV [39]. However,
under irradiation conditions and in ferritic-martensitic alloys, helium’s mobility can be severely
limited by the density of traps in the material. Molecular dynamics calculations by Heinisch et
al. [40] estimate that the binding energy of helium atoms at dislocations to be approximately 2.5
eV, which is strong enough to prevent helium emission up to temperatures of ~530°C. High
concentrations of vacancies, and the presence of Cr up to 10% have also been known to limit the
mobility of helium in complex ferritic alloys [41]. Given that helium bubbles at grain boundaries
were not observed in any condition in this work, it is likely that helium mobility was not high
enough to escape trapping at dislocations or other features within the matrix. Therefore, small,
invisible helium clusters which remain in the matrix could serve to immobilize the vacancies,
enhancing recombination and preventing cavity growth.
The high cavity sink strength of the 100 and 1000 appm conditions was not sufficient to
explain the observed suppression in swelling. Therefore, an alternative mechanism, such as
helium trapping, or helium-induced alteration of cavity and dislocation bias are the likely causes
of the reduction in swelling.

4.4 Influence of dislocations and precipitates

The behavior of dislocations with damage (Figure 8) was not unusual and similar
behavior has been reported in previous high damage irradiations on HT9 [20], which showed
loop diameter was relatively insensitive to damage beyond 200 dpa. A total sink strength of ~3-6
x 1014 m-2 was reported in HT9, which is slightly lower than observed in this study’s T91 (5-7
x1014 m-2, Figure 10a) due to a higher density of <100> loops observed in T91. At 300 dpa with
10 appm He implanted, the <100> loops exhibited a higher sink strength when compared to the
<111> loops (3.3 x 1014m-2 versus 0.5 x 1014m-2). Therefore, the <100> loops would be expected
to be a much stronger sink over the <111> loops.
The dislocation behavior was very similar across helium levels, suggesting that it did not
influence the cavity evolution differently across the various helium levels. The biggest difference
occurred in the 0 appm He loop size, which was slightly larger at 36 nm versus about 21-28 nm
for the pre-implanted conditions at 150 dpa. It is possible that the lack of helium in this condition
allowed the loops to grow to larger sizes. Small pre-implanted helium clusters may have served
as pinning sites for dislocation loops, restricting their growth. However, this trend did not

12
continue with addition of helium beyond 1 appm. The 0 appm He condition exhibited similar
loop densities as the pre-implanted cases, suggesting that the presence of helium had little to no
effect on the nucleation of dislocation loops. Irradiation with helium has been shown to increase
the dislocation density [5], however, the implantation was performed at room temperature and
ferritic-martensitic steels naturally contain a high dislocation density. Furthermore, any effect on
the dislocations by irradiation with helium was insignificant compared to the high damage
induced by the 5 MeV Fe2+ ions. The dislocation density of the 1000 appm He implanted
condition measured 3.71 x 1014 m-2, which was well within the measurement error of the
unirradiated condition of 3.0 x 1014 m-3.
To determine any possible effect the dislocations may have had on cavity behavior, their
sink strengths need to be considered. Figure 10a plots the sink strengths of the <100> loops, the
dislocation network, along with the total line length at 150 dpa. The sink strengths ranged
between 6–7 x 1014 m-2, with the exception of the 0 appm He case, which exhibited a sink
strength of about 1 x 1015m-2 due to the aforementioned larger loop size, and a slightly higher
network dislocation sink strength. These sink strengths are reasonable, considering typical values
for ferritic-martensitic steels reported in the literature range from ~1 x 1014 – 1 x 1015 m-2
[33,42,43]. Previous work on HT9 [44] showed that changes in the dislocation sink strength
within this range would not dramatically alter the swelling rate. Figure 10b compares the
dislocation sink strengths at 150 dpa with the cavity sink strengths across all helium levels. At
every helium level except 1000 appm, the ratio between sink strengths was above 1, signifying
that the sink balance and relative importance of dislocations did not change across helium levels.
Figure 10a shows that the total line length did not change significantly with helium content,
therefore the dislocations did not influence the sink structure differently across helium levels.
Considering the consistency of the dislocation sink strength across four magnitudes of pre-
implanted helium content, changes in the cavity microstructure cannot be attributed to changes in
the dislocation microstructure.
Ni-Si clusters (G-phase) experienced a coarsening behavior as damage level was
increased (Figure 12). This growth behavior was also observed in high damage irradiation on
HT9 [20]. It was determined that G-phase in this alloy would have a negligible effect on the
swelling behavior up to 650 dpa [44]. G-phase in the HT9 study was observed to have a sink
strength of ~1 x 1014m-2. By 300 dpa, the Ni-Si clusters in T91 in this study achieved a sink
strength of only 0.3 x 1014 m-2, or a factor of 20 less than that due to dislocations. Therefore, even
less of an effect on cavity evolution would be expected.
In summary, <100> loops and network dislocations appear to be insensitive to implanted
He content, and G-phase and <111> loops exhibit very low sink strengths which are not expected
to influence cavity evolution. Therefore, changes in the cavity microstructure cannot be
attributed to changes in other microstructural features. The determining factor in the differences
in cavity evolution was the concentration of implanted helium.

5. Conclusions

As the helium content in T91 was increased from 0 appm to 1000 appm, dislocation and
precipitate sizes and densities were largely unaffected in contrast to a complex evolution of
cavities. In the low helium (0, 1, and 10 appm) and low damage (< 50 dpa) regime, the cavity
evolution was largely controlled by the sink strengths. Helium levels of 1 and 10 appm were
sufficient to establish a high density of cavities early on in the irradiation at the cost of slightly

13
suppressed growth. The 0 appm He condition took longer to reach a similar density, but achieved
a similar sink strength by 150 dpa. The similar sink strengths achieved by the 0, 1 and 10 appm
He conditions guided the evolution of the cavities along a similar path and resulted in nearly
identical cavity distributions by 300 dpa.
The high helium conditions (100 and 1000 appm) exhibited a fundamentally different
cavity behavior. High helium levels lowered the critical cavity radius and facilitated the
formation of a high-density bubble population. These distributions resulted in a high cavity sink
strength accompanied by a suppression in overall swelling. The bubbles remained stable even at
high damage levels, while a population of larger cavities experienced very limited growth.
Neither the high cavity sink strength nor changes in the dislocation and precipitate
microstructure could account for the observed swelling suppression. Therefore, either trapping
by small helium clusters, alteration of dislocation bias via helium, or the population of small
biased cavities were responsible for swelling suppression.

Acknowledgements

The authors gratefully acknowledge O. Toader, F. Naab, T. Kubley, and E. Uberseder at

the Michigan Ion Beam Laboratory for their assistance with helium implantations and ion
irradiations and the Michigan Center for Materials Characterization for use of the instruments
and staff assistance and NSF grant #DMR-0320740 for support of the JEOL 2100F TEM. This
work was supported by DOE under award number DE-AC07-05ID14517.

14
 Table 1. Composition of alloy T91 heat C2269 in weight percent.
Cr Mo Ni Mn Si V C Cu Fe
8.37 0.9 0.21 0.45 0.28 0.22 0.1 0.17 Bal

Table 2. Summary of cavities in T91 after ion irradiation as a function of dpa and implanted He
content.
Pre-Implanted Average Cavity Cavity Sink
Damage level Depth of Swelling
He Diameter Density Strength
(dpa) Analysis (%)
(appm) (nm) (1020 m-3) (1014 m-2)
45 400-600nm 0 9.7 11.5 0.079 0.70
1 8.6 19.5 0.098 1.05
10 8.5 25.5 0.124 1.36
100 3.2 85.7 0.048 1.72
1000 1.6 904.0 0.061 9.09
50 500-700nm 0 10.1 11.7 0.091 0.74
1 8.72 20.1 0.107 1.10
10 8.5 24.2 0.124 1.29
100 3.3 87.8 0.052 1.82
1000 1.56 907.0 0.062 8.89
58 600-800nm 0 10.3 11.5 0.094 0.74
1 8.9 19.9 0.114 1.11
10 8.8 24.8 0.147 1.37
100 3.3 87.2 0.051 1.81
1000 1.6 908.0 0.065 9.13
134 400-600nm 0 17.0 17.8 0.67 1.90
1 13.9 17.6 0.37 1.54
10 13.3 23.6 0.44 1.97
100 4.9 70.4 0.17 2.17
1000 2.09 657.7 0.10 8.64
150 500-700nm 0 19.2 17.6 0.92 2.12
1 15.3 20.7 0.58 1.99
10 14.3 24.0 0.56 2.16
100 5.07 72.4 0.20 2.31
1000 2.1 658.6 0.11 8.69
175 600-800nm 0 21.2 17.7 1.1 2.36
1 16.6 21.9 0.76 2.28
10 15.5 23.0 0.65 2.24
100 5.1 72.7 0.21 2.33
1000 2.08 655.7 0.11 8.57
276 400-600nm 0 26.0 18.5 2.6 3.02
1 25.5 21.3 2.9 3.41
10 24.1 16.4 1.8 2.48
100 4.8 61.6 0.44 1.86
1000 1.79 757.7 0.12 8.52
300 500-700nm 0 29.3 20.1 3.9 3.70
1 28.2 21.6 3.9 3.83
10 26.4 18.9 2.7 3.14
100 5.3 62.3 0.61 2.07
1000 1.79 757.0 0.13 8.51

15
 356 600-800nm 0 31.8 19.8 4.8 3.96
1 30.3 19.7 4.2 3.75
10 28.8 19.3 3.5 3.49
100 5.7 63.3 0.78 2.27
1000 1.80 755.8 0.15 8.55

Table 3. Summary of size, density and volume/length fraction of dislocation loops, dislocation
network, and G-phase in ion irradiated T91.

Pre- <100> <100> <111> <111> G-phase G-phase

Damage Network Total Line G-phase G-phase
Implanted Loop Loop Loop Loop Volume Sink
level Density Length Diameter Density
He Diameter Density Diameter Density Fraction Strength
(dpa) (1014 m-2) (1014 m-2) (nm) (1020 m-3)
(appm) (nm) (1020 m-3) (nm) (1020 m-3) (%) (1014 m-2 )

50 10 16.9 59.3 7.9 24.8 3.04 10.6 4.5 14.9 0.012 0.42
150 0 36.3 22.4 *N.M. N.M. 4.27 9.88 N.M. N.M. N.M. 0.45
1 24.0 19.2 N.M. N.M. 3.05 5.96 N.M. N.M. N.M. 0.55
10 28.7 17.5 16.4 4.4 3.07 6.88 7.0 13.3 0.033 0.58
100 27.1 17.2 N.M. N.M. 3.59 6.53 N.M. N.M. N.M. 0.42
1000 21.6 24.4 N.M. N.M. 2.2 5.54 N.M. N.M. N.M. 0.35
300 10 24.0 16.7 18.4 6.9 2.45 5.67 7.6 7.57 0.041 0.36

*N.M = Not Measured

Table 4. Key parameters for calculating the T91 cavity stability map.

Symbol Definition Value

Irradiation
T 460°C
Temperature
γ Surface Energy 1.7 J/m2
Ω Atomic Volume 2.38 x 10-29 m-3

Zic Z vd Capture Efficiency 0.95

Zvc Zid
*Ko Damage Rate 1 x 10-3 dpa/s
Vacancy Migration
* Evm 0.63 eV
Energy
Vacancy
* Evf 1.6 eV
Formation Energy
Interstitial
* Eim 0.22 eV
Migration Energy
*used to determine vacancy supersaturation, Cv / Cvo

16
 3.5 7
Helium
Implanted Fe
3 6

Implanted Fe Ions (% per 100 dpa)

SRIM Damage Rate (10 dpa/s) or
Pre-Implanted Helium (a.u.)

2.5 5
-3

2 4

1.5 Damage 3

1 2

0.5 1

0 0
0 500 1000 1500 2000
Depth (nm)
Figure 1. The helium profile (blue) overlaid with the damage rate (red dashed) and the implanted
ion distribution for 5 MeV Fe2+ ions (green dotted) taken from SRIM simulations in T91.

17
Figure 2. HAADF STEM images of T91 irradiated to 300 dpa with pre-implanted helium levels
of 0, 1, 10, 100, and 1000 appm. The damage (dashed) and helium (solid) profiles are overlaid on
the 1 appm He image.

18
Figure 3. a) Average cavity diameter b) number density, and c) swelling as a function of
damage in T91 for pre-implanted He levels of 0, 1, 10, 100, and 1000 appm.

19
Figure 4. Cavity size distributions for the 0, 1, and 10 appm He conditions at 50, 150,
and 300 dpa (500-700 nm depth) in ion irradiated T91.

20
Figure 5. HAADF STEM images showing the evolution of the T91 cavity microstructure with
damage for the 0, 1 and 10 appm He conditions.

21
Figure 6. High magnification (1Mx) HAADF STEM images of T91 irradiated to 150 dpa (500-
700 nm depth) with pre-implanted helium levels of 0, 1, 10, 100, and 1000 appm.

Figure 7. Cavity size distributions for the 100 and 1000 appm He conditions at 50, 150, and 300
dpa (500-700 nm depth) in ion irradiated T91.

22
Figure 8. Dislocation loop densities and diameters for b<100> and ½ b<111> loops are plotted
as a function of damage for the 10 appm He condition. Some loops are indicated in the images
with white circles.

23
Figure 9. Densities and diameters for b<100> dislocation loops are plotted as a function of
helium content at 150 dpa. Some loops are indicated in the images with white circles.

24
 Figure 10. a) Network and <100> loop sink strength are plotted along with the total line
length and b) ratio of total line length to cavity sink strength as a function of helium content at
150 dpa.

Figure 11. EDS mapping shows the presence of Ni-Si rich clusters after irradiation in
T91 irradiated to 300 dpa (500-700 nm depth) with 10 appm pre-implanted He.

25
 20
Density Volume Fraction
Diameter (nm) 0.05
Number Density (10 m )
-3

15

Volume Fraction (%)

20

0.04
Diameter (nm)

10
0.03

5
0.02

0 0.01
0 50 100 150 200 250 300 350
Damage (dpa)

Figure 12. Number density, diameter, and volume fraction of G-phase plotted as a
function of damage with 10 appm He pre-implanted in ion irradiated T91.

26
Figure 13. Cavity sink strength is plotted as a function of helium content at damage
levels of 50, 150, and 300 dpa in ion irradiated T91.

27
Figure 14. Cavity stability map for T91 showing dependence of the critical radius on the actual
cavity radius for various numbers of contained gas atoms. The highlighted grey area shows the
possible size range for the presence of stabilized bubbles.

28
References

[1] G.S. Was, Z. Jiao, E. Getto, K. Sun, A.M. Monterrosa, S.A. Maloy, O. Anderoglu, B.H.
Sencer, M. Hackett, Emulation of reactor irradiation damage using ion beams, Scr. Mater.
88 (2014) 33–36. doi:10.1016/j.scriptamat.2014.06.003.
[2] G.S. Was, J.T. Busby, T. Allen, E.A. Kenik, A. Jenssen, S.M. Bruemmer, J. Gan, A.D.
Edwards, P.M. Scott, P.L. Andresen, Emulation of neutron irradiation effects with
protons : validation of principle, J. Nucl. Mater. 300 (2002) 198–216.
[3] S.J. Zinkle, L.L. Snead, Opportunities and limitations for ion beams in radiation effects
studies : Bridging critical gaps between charged particle and neutron irradiations, Scr.
Mater. 143 (2018) 154–160. doi:10.1016/j.scriptamat.2017.06.041.
[4] K. Farrell, Experimental Effects of Helium on Cavity Formation During Irradiation - A
Review, Radiat. Eff. 53 (1980) 175–194.
[5] K. Farrell, P.J. Maziasz, E.H. Lee, L.K. Mansur, Modification of Radiation Damage
Microstructure by Helium, Radiat. Eff. 78 (1983) 277–295.
[6] G.R. Odette, P.J. Maziasz, J.A. Spitznagel, Fission-Fusion Correlations for Swelling and
Microstructure in Stainless Steels: Effect of the Helium to Displacement Per Atom Ratio,
103 & 104 (1981) 1289–1304.
[7] R.E. Stoller, The influence of helium on microstructural evolution: Implications for DT
fusion reactors, J. Nucl. Mater. 174 (1990) 289–310.
[8] J.L. Brimhall, E.P. Simonen, Effect of Helium on Void Formation in Nickel, J. Nucl.
Mater. 68 (1977) 235–243.
[9] K. Farrell, M.B. Lewis, N.H. Packan, Simultaneous Bombardment with Helium,
Hydrogen, and Heavy Ions to Simulate Microstructural Damage from Fission or Fusion
Neutrons, Scr. Metall. 12 (1978) 1121–1124.
[10] F.A. Smidt, P.R. Malmberg, J.A. Sprague, J.E. Westmoreland, Swelling Behavior of
Commercial Ferritic Alloys, EM-12 and HT-9, as Assessed by Heavy Ion Bombardment,
Irradiat. Eff. Microstruct. Prop. Met. (1976) 227–241.
[11] J.J. Kai, G.L. Kulcinski, 14 MeV nickel-ion irradiated HT-9 ferritic steel with and without
helium pre-implantation, J. Nucl. Mater. 175 (1990) 237–243. doi:10.1016/0022-
3115(90)90212-6.
[12] M.B. Toloczko, F.A. Garner, V.N. Voyevodin, V. V. Bryk, O. V. Borodin, V. V.
Mel’Nychenko, A.S. Kalchenko, Ion-induced swelling of ODS ferritic alloy MA957
tubing to 500 dpa, J. Nucl. Mater. 453 (2014) 323–333.
doi:10.1016/j.jnucmat.2014.06.011.
[13] X. Wang, A.M. Monterrosa, F. Zhang, H. Huang, Q. Yan, Z. Jiao, G.S. Was, L. Wang,
Void swelling in high dose ion-irradiated reduced activation ferritic–martensitic steels, J.
Nucl. Mater. 462 (2015) 119–125. doi:10.1016/j.jnucmat.2015.03.050.
[14] E. Getto, Z. Jiao, A.M. Monterrosa, K. Sun, G.S. Was, Effect of pre-implanted helium on
void swelling evolution in self-ion irradiated HT9, J. Nucl. Mater. 462 (2015) 458–469.
doi:10.1016/j.jnucmat.2015.01.045.
[15] G.S. Was, S. Taller, Z. Jiao, A.M. Monterrosa, D. Woodley, D. Jennings, T. Kubley, F.
Naab, Resolution of the carbon contamination problem in ion irradiation experiments,
Nucl. Inst. Methods Phys. Res. B. 412 (2017) 58–65. doi:10.1016/j.nimb.2017.08.039.
[16] J.F. Ziegler, M.D. Ziegler, J.P. Biersack, Nuclear Instruments and Methods in Physics
Research B SRIM – The stopping and range of ions in matter ( 2010 ), Nucl. Inst.

29
 Methods Phys. Res. B. 268 (2010) 1818–1823. doi:10.1016/j.nimb.2010.02.091.
[17] R.E. Stoller, M.B. Toloczko, G.S. Was, A.G. Certain, S. Dwaraknath, F.A. Garner, On the
use of SRIM for computing radiation damage exposure, Nucl. Instruments Methods Phys.
Res. Sect. B Beam Interact. with Mater. Atoms. 310 (2013) 75–80.
doi:10.1016/j.nimb.2013.05.008.
[18] T. Malis, S.C. Cheng, R.F. Egerton, EELS Log-Ratio Technique for Specimen-Thickness
Measurement in the TEM, J. Electron Microsc. Tech. 8 (1988) 193–200.
doi:10.1002/jemt.1060080206.
[19] S. Smith, L. Guttman, Measurement of Internal Boundaries in Three-Dimensional
Structures By Random Sectioning, J. Met. Trans. (1953) 81–87.
[20] E. Getto, K. Sun, A.M. Monterrosa, Z. Jiao, M.J. Hackett, G.S. Was, Void swelling and
microstructure evolution at very high damage level in self-ion irradiated ferritic-
martensitic steels, J. Nucl. Mater. 480 (2016) 159–176.
doi:10.1016/j.jnucmat.2016.08.015.
[21] A. Hishinuma, L.K. Mansur, Critical Radius for Bias-Driven Swelling - A Further
Analysis and its Application to Bimodal Cavity Size Distributions, 118 (1983) 91–99.
[22] B.H. Sencer, J.R. Kennedy, J.I. Cole, S.A. Maloy, F.A. Garner, Microstructural stability
of an HT-9 fuel assembly duct irradiated in FFTF, J. Nucl. Mater. 414 (2011) 237–242.
doi:10.1016/j.jnucmat.2011.03.050.
[23] N.H. Packan, K. Farrell, Simulation of First Wall Damage: Effects of the Method of Gas
Implantation, J. Nucl. Mater. 86 (1979) 677–681.
[24] A. Hishinuma, J. Vitek, J. Horak, E. Bloom, Effect of Preinjected Helium on Swelling and
Microstructure of Neutron Irradiated Pressurized Tubes of Type 316 Stainless Steel, Eff.
Radiat. Mater. 11th Int. Symp. ASTM STP 782. (1982) 92–107.
[25] D. Carpentier, T. Jourdan, Y. Le Bouar, M.. Marinica, Effect of saddle point anisotropy of
point defects on their absorption by dislocations and cavities, Acta Mater. 136 (2017)
323–334. doi:10.1016/j.actamat.2017.07.013.
[26] V.A. Borodin, A.I. Ryazanov, C. Abromeit, Void bias factors due to the anisotropy of the
point defect diffusion, J. Nucl. Mater. 207 (1993) 242–254.
[27] M.P. Surh, W.G. Wolfer, Accurate Mean Field Void Bias Factors for Radiation Swelling
Calculations, J. Comput. Mater. Des. 14 (2007) 419–424. doi:10.1007/s10820-007-9052-
2.
[28] A.A. Kohnert, M. Alice, B.D. Wirth, Molecular statics calculations of the biases and point
defect capture volumes of small cavities, J. Nucl. Mater. 499 (2018) 480–489.
doi:10.1016/j.jnucmat.2017.12.005.
[29] Z. Chang, D. Terentyev, N. Sandberg, K. Samuelsson, P. Olsson, Anomalous bias factors
of dislocations in bcc iron, J. Nucl. Mater. 461 (2015) 221–229.
doi:10.1016/j.jnucmat.2015.03.025.
[30] T. Jourdan, Influence of dislocation and dislocation loop biases on microstructures
simulated by rate equation cluster dynamics, J. Nucl. Mater. 467 (2015) 286–301.
doi:10.1016/j.jnucmat.2015.09.046.
[31] W.G. Wolfer, The Dislocation Bias, J. Comput. Mater. Des. 14 (2007) 403–417.
doi:10.1007/s10820-007-9051-3.
[32] Q. Li, C.M. Parish, K.A. Powers, M.K. Miller, Helium solubility and bubble formation in
a nanostructured ferritic alloy q, 445 (2014) 165–174.
[33] I.-S. Kim, J.D. Hunn, N. Hashimoto, D.L. Larson, P.J. Maziasz, K. Miyahara, E.H. Lee,

30
 Defect and void evolution in oxide dispersion strengthened ferritic steels under 3.2 MeV
Fe+ ion irradiation with simultaneous helium injection, J. Nucl. Mater. 280 (2000) 264–
274. doi:10.1016/S0022-3115(00)00066-0.
[34] R. Sugano, K. Morishita, H. Iwakiri, N. Yoshida, Effects of dislocation on thermal helium
desorption from iron and ferritic steel, J. Nucl. Mater. 307–311 (2002) 941–945.
doi:10.1016/S0022-3115(02)01098-X.
[35] P.D. Edmondson, C.M. Parish, Y. Zhang, A. Hallén, M.K. Miller, Helium bubble
distributions in a nanostructured ferritic alloy, 434 (2013) 210–216.
[36] G.R. Odette, On mechanisms controlling swelling in ferritic and martensitic alloys, J.
Nucl. Mater. 157 (1988) 921–927.
[37] L.K. Mansur, Effects of Point Defect Trapping and Solute Segregation on Irradiation-
Induced Swelling and Creep, J. Nucl. 83 (1979) 109–127.
[38] D. Chen, Y. Tong, H. Li, J. Wang, Y.L. Zhao, A. Hu, J.J. Kai, Helium accumulation and
bubble formation in FeCoNiCr alloy under high fluence He+ implantation, 501 (2018)
208–216. doi:10.1016/j.jnucmat.2018.01.041.
[39] M.J. Caturla, C.J. Ortiz, Effect of self-interstitial cluster migration on helium diffusion in
iron, 362 (2007) 141–145. doi:10.1016/j.jnucmat.2007.01.017.
[40] H.L. Heinisch, F. Gao, R.J. Kurtz, E. a. Le, Interaction of helium atoms with edge
dislocations in α-Fe, J. Nucl. Mater. 351 (2006) 141–148.
doi:10.1016/j.jnucmat.2006.02.027.
[41] M.K. Miller, P.D. Edmondson, C.M. Parish, Y. Zhang, A. Halle, Helium entrapment in a
nanostructured ferritic alloy, 65 (2011) 731–734. doi:10.1016/j.scriptamat.2011.07.024.
[42] D.R. Harries, R.L. Klueh, High-Chromium Ferritic and Martensitic Steels for Nuclear
Applications, 2001.
[43] P. Dubuisson, D. Gilbon, J. Seran, Microstructural evolution of ferritic-martensitic steels
irradiated in the fast breeder reactor Phenix, J. Nucl. Mater. 205 (1993) 178–189.
[44] E. Getto, G. Vancoevering, G.S. Was, The co-evolution of microstructure features in self-
ion irradiated HT9 at very high damage levels, J. Nucl. Mater. 484 (2017) 193–208.
doi:10.1016/j.jnucmat.2016.12.006.

31
Highlights

• Cavity density increased monotonically with helium content

• Swelling generally decreased monotonically with helium content
• Low helium levels resulted in increases in cavity nucleation at low damage levels, but did not
influence cavity evolution at high damage levels
• High helium levels lowered the critical cavity radius and produced bimodal distributions
• Cavity evolution was suppressed at high helium levels due to helium trapping, or alteration of
cavity and dislocation bias
• Dislocations and precipitates did not significantly influence cavity evolution