ORIGINAL RESEARCH ARTICLE

Concomitant Precipitation of Intermetallic b-NiAl

and Carbides in a Precipitation Hardened Steel
S. JAKOB, M. HÖRNQVIST COLLIANDER, J. KAWSER, S. RASHIDI, S.W. OOI,
and M. THUVANDER

The investigated steel is hardened by precipitation of both intermetallic b-NiAl particles and
carbides. Around peak hardness, here reached after aging at 520 °C for 6 h, the number density
of b-NiAl particles is 2.4 9 1024 and 4.4 9 1023 for carbides, respectively. The carbides and
b-NiAl often form co-precipitates with presumably the carbides nucleating on the b-NiAl
precipitates. The secondary carbides are mainly of the Cr-rich M23C6 type in the investigated
states. The steel has a good resistance against over-aging, whereby the b-NiAl coarsen
considerably slower than the carbides.

https://doi.org/10.1007/s11661-023-07291-7
Ó The Author(s) 2024

I. INTRODUCTION on the other hand. This group of steel is also called

dual-hardening steels and bridges the gap between tool
THE ever-increasing demands on steel components steels, maraging steels and lower alloyed engineering
include strength, toughness and in recent years also steels.[12–15] Other co-precipitation systems employ the
lightweight design. The latter can be achieved by formation of Cu precipitates alongside carbides.[16,17] A
increasing the yield strength of the material as long as comprehensive review can be found in Reference 18. The
the deformability can be preserved. Furthermore, the steel in this study is fabricated using ingot casting
mechanical strength as well as the fatigue resistance allowing for alloying with a high content of aluminium,
must be maintained at operating temperatures. This which in addition to formation of Al-containing inter-
requires sophisticated control of chemical composition metallic particles also provides good corrosion proper-
and thermo-mechanical processing.[1] Precipitation ties. The carbon content is elevated to facilitate the
hardening is a long-established way to improve mechan- carbide formation and the microstructure consists of
ical strength. High strength in excess of 1 GPa is martensite. Hybrid 55 displays a significant increase in
possible by having nano-scale clusters and precipitates hardness upon aging for a few hours in the temperature
in the microstructure.[2,3] Most precipitation hardening range of 500 °C to 600 °C, and the resistance to
steels obtain their strength by precipitation of inter- over-aging is better than for typical precipitation hard-
metallic particles[4–7] or nm-sized carbides.[8,9] Depend- ening steels.[11]
ing on the size and coherency of the particles, the Atom probe tomography (APT) is a unique tech-
dislocation movement is hindered either by shearing nique, allowing for the investigation of nm-sized fea-
through the particle or by formation of dislocation tures within complex engineering alloys. It has nearly
loops around the precipitates (Orowan mechanism).[10] atomic spatial resolution and yields the chemical infor-
The investigated steel, Hybrid 55Ò,[11] obtains a high mation with high sensitivity for all elements.[19]
strength by the formation of intermetallic b-NiAl The purpose of this study is to describe the basics of
precipitates (also known as B2-NiAl for its CsCl the precipitation process and to understand the reason
structure) on one hand and finely dispersed carbides for the good over-aging resistance. Therefore, APT was
conducted on peak-hardened and over-aged specimens.
Furthermore, a solution-annealed sample as well as the
heat-treated materials have been investigated using
S. JAKOB, M. HÖRNQVIST COLLIANDER, J. KAWSER, S.
synchrotron X-ray diffraction (XRD).
RASHIDI, and M. THUVANDER are with the Department of
Physics, Chalmers University of Technology, 412 96 Gothenburg,
Sweden. Contact e-mail: [email protected];
[email protected] S.W. OOI is with the Maxwell II. EXPERIMENTAL
Centre, Ovako Group R&D, JJ Thompson Avenue, Cambridge,
CB3 0HE, UK and also with the Ovako Corporate R&D, Building The composition of the investigated steel Hybrid 55
202, 813 82 Hofors, Sweden. was measured by optical emission spectroscopy and is
Manuscript submitted June 30, 2023; accepted December 19, 2023. shown in Table I. After casting, homogenization, hot
Article published online January 19, 2024

870—VOLUME 55A, MARCH 2024 METALLURGICAL AND MATERIALS TRANSACTIONS A

rolling and soft annealing treatment, the steel was this peak was ranged accordingly. The peak at 29 Da
austenitized at 950 °C for 45 minutes, followed by was ranged as Ni2+. Size, volume fraction and number
quenching. This condition is further called solution-an- density of the precipitates were calculated using iso-con-
nealed (SA) since most of the carbides are dissolved. centration surfaces as well as the parameter-free method
One aging treatment was performed at 520 °C for presented by Zhao et al.[21] This method uses the radial
6 hours to reach approximately peak hardness. Also, distribution function (RDF) of the atomic species. The
one over-aged sample was produced by heat treatment size of the precipitates was then calculated from the
at 650 °C for 4 hours. The three samples are designated number of atoms, hence, avoiding local magnification
as SA, peak-aged and over-aged, and the hardness effects.[22,23] In this study, only the RDF of Ni was used
values were 410, 600 and 360 HV, respectively. for b-NiAl precipitates to avoid the overlap of Al at 27
The APT study was carried out using a LEAP 3000X Da.
HR (Imago Scientific Instruments). The analyses were To obtain crystallographic information, XRD mea-
carried out in laser mode with a laser pulse energy of surements were made at the Swedish Material Science
0.30 nJ (frequency 200 kHz, detection rate 0.5 pct, beamline P21 at the PETRA III synchrotron (DESY,
temperature 50 K). The data were evaluated using the Germany). The energy was 60 keV, corresponding to a
software IVAS 3.6 (CAMECA). Specimen preparation wavelength k = 0.2066 Å. Two VAREX XRD4343CT
was made using standard electropolishing.[20] The col- area detectors were placed downstream at a sam-
lected datasets had mass spectra with some overlapping ple-to-detector distance of around 2.8 m (the exact
peaks. Especially the peak at 27 Da is of interest since it distance was calibrated using a LaB6 standard) to allow
has contributions from Fe2+, Cr2+ and Al+ ions. The collection of diffraction patterns in transmission. The
compositions of matrix and precipitates were measured investigated samples had a thickness of 1.0 mm, and the
by separating the reconstruction into several POS-files. beam size was 0.5 9 0.5 mm2. Data reduction was
Hereby, the peak at 27 Da could be attributed to Al for performed using PyFAI 0.20.0.[24,25]
intermetallic particles, Fe or Cr in case of carbides and
Fe in the matrix. The composition between matrix and
intermetallic particles is depicted with proximity his- III. RESULTS
tograms (proxigrams). There, the contribution of Fe and
Cr was calculated proportionally to the peaks at 26 and Figure 1 shows electron micrographs of Hybrid 55. It
28 Da, respectively, using the natural isotope abun- has a martensitic microstructure with several undis-
dance. The remaining ions at 27 Da were then attributed solved carbides visible. The carbides are Cr-rich in case
to Al. The peak at 24 Da contains C2+ and C42+. Since of M23C6 and V-rich in case of MC. The intermetallic
there is a peak at 24.5 Da for C42+ and the peak particles as well as strength-enhancing smaller carbides
decomposition revealed a majority of C42+ for 24 Da, are not visible at this length scale. Thus, advanced
investigation methods like synchrotron XRD and APT
are employed.
Table I. The Measured Chemical Composition of the Figure 2 shows diffractograms obtained before and
Investigated Steel (in Wt. Pct), Balance Fe.[11] after aging. The SA condition shows clear peaks from
retained austenite, weight fraction of around 2.6 pct
C Si Mn Cr Ni Mo V Al Cu
from Rietveld refinement, which is reduced to 0.8 pct
0.18 0.10 0.31 5.05 6.09 0.69 0.49 2.36 0.17 after 6 hours at 520 °C and almost zero (too low for

Fig. 1—Electron micrographs ((a) SEM and (b) TEM) of the Hybrid 55 microstructure. Only undissolved carbides are visible at the
magnifications used here.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 55A, MARCH 2024—871

Fig. 2—(a) XRD diffractogram of material in the solution-annealed state, and after aging at 520 °C for 6 h and 650 °C for 4 h (logarithmic
intensity scale and vertical offsets). (b) through (d) Magnified views with linear intensity scales to highlight carbides and NiAl (Color
figure online).

reliable quantification) after 4 hours at 650 °C where W is the full width at half maximum (after
(Figure 2(a)). A closer examination of the SA condition subtraction of the instrumental peak width) and h is the
(Figure 2(b)) shows the presence of both MC and Bragg angle. For the peak-aged condition, Eq. [1]
M23C6. The carbide peaks are relatively sharp, indicat- suggests an approximate diameter of around 3 nm. No
ing particle sizes above where size-related broadening significant change of the MC peaks could be observed,
occurs. This is confirmed by TEM, as seen in Figure 1, whereas the M23C6 peaks increased significantly in
where scarce MC and M23C6 carbides larger than intensity (see Figure 3), indicating an increased volume
100 nm are observed. fraction.
The b-NiAl particles are clearly visible after 6 hours In the over-aged condition there is a significant
at 520 °C (Figure 2(c)), although the peaks remain increase in the M23C6 fraction, as suggested by the
broad, which indicates a small particle size. An approx- integrated peak area (Figure 3), and a clear narrowing of
imate size, D, of the coherently diffracting domains can the b-NiAl peak (Figures 2(a) and (c)). The integrated
be obtained from the Scherrer equation intensity of the b-NiAl peak is not significantly higher
compared to the peak-aged condition, only around 5
0:9k pct, but the peak width is distinctly smaller, and Eq. [1]
D¼ ; ½1
Wcosh results in an approximate diameter of 25 nm for the
over-aged condition. There are also several additional

872—VOLUME 55A, MARCH 2024 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 3—(a) Single peak fits of MC and M23C6 peaks for the different conditions. (b) Integrated intensities of the fitted peaks, normalized by the
value for the SA condition (Color figure online).

Fig. 4—APT reconstruction of the peak-aged material (520 °C/6 h) showing co-precipitation of b-NiAl (red iso-surfaces) and Cr23C6 (blue
iso-surfaces). A large primary VC is shown in green (Color figure online).

diffraction peaks, likely a result of the precipitation of to martensite upon final cooling (after the aging). See
other carbide species. As the over-aged condition is the supplementary file for atom distribution maps.
mainly included for comparison in the present case, no In the over-aged sample (650 °C/4 hours), coarsening
further attempts were made to identify the additional of the precipitates has occurred, see the APT recon-
carbides. structions in Figure 5. The carbides were of two distinct
The APT investigation of the peak-aged (520 °C/6 types; Cr23C6 and VC. The iso-concentration surface
hours) sample reveals a very large number density of thresholds for the depicted particles are 25 at. pct Ni and
small precipitates, see Figure 4. The precipitates are of Al for the intermetallic particles, 12 at. pct Cr for Cr23C6
two types; b-NiAl and Cr23C6. The particles are dis- and 6 at. pct V for VC, respectively. The carbides always
played as iso-concentration surfaces. The thresholds are have at least one b-NiAl particle in direct contact with
25 at. pct Ni and Al for the intermetallic particles, 12 at. them, but many b-NiAl particles are isolated. Elemental
pct Cr for Cr23C6 and 30 at. pct V for a primary VC, distributions can be found in the Supplementary File.
respectively. Most of the carbides are in contact with a The composition of matrix and precipitates was
b-NiAl particle. The precipitates were partly aligned on deduced from separated POS-files and can be seen in
parallel planes, presumably consisting of arrays of Table II. The b-NiAl precipitates still contain significant
dislocations. Furthermore, a 50 nm wide section without amounts of Fe and almost equal amounts of Ni and Al.
any precipitates can be seen in Figure 4. This volume is The nm-sized Cr23C6 still incorporate large amounts of
most probably retained austenite. The interface is Fe, and are enriched in V and Mo. The primary VC
enriched in carbide forming metals as well as P, contains small amounts of Cr and Mo. In the over-aged
suggesting a former grain boundary. Throughout the material, the intermetallic particles contain about 8 at.
precipitation-free region, alloying elements, such as Al, pct Fe, as well as slightly more than 2 at. pct of Cu and
Cr, Ni and C are enriched. From the APT analysis, it Mn combined. The Cr23C6 have higher concentration of
cannot be confirmed whether this volume transformed Cr and V compared to the peak-aged condition. Also,
the measured C content is higher. The VC particles have

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 55A, MARCH 2024—873

Fig. 5—APT reconstructions of two specimens from over-aged material (650 °C/4 h). The red iso-surfaces depict b-NiAl particles, the blue
iso-surfaces are Cr23C6 and VC are depicted in green (Color figure online).

about the same amount of V and C with significantly so, the number density of carbides is much higher in the
more Cr, Mo and Fe concentration than the primary peak-aged condition. In the specimen of over-aged
VC. material, the carbides were measured only in part due
Figure 6(a) depicts the size distribution of particles to the limited field of view and the small analysis
derived from iso-surfaces for both peak-aged and volume. It is not reasonable to calculate number
over-aged samples. Herein, 25 at. pct of Ni and Al, densities from this limited sample volume, but it seems
and 12 at. pct of Cr were used for iso-surface calcula- to be in the order of 1020 m3.
tions in IVAS, for b-NiAl and carbides, respectively. A The chemical composition of b-NiAl precipitates was
mean equivalent radius of spherical particles was calcu- further investigated with proxigrams. Figure 7 depicts
lated and the results are listed in Table III. The error the chemical concentration of important elements as a
margins for these radii are calculated from variation of function of distance to the 25 at. pct Ni + Al iso-sur-
the threshold value by 2 at. pct Ni and Al and 1 at. pct face. The b-NiAl particles have a gradual change in
Cr in either direction. Figure 6(b) shows the normalised composition from the matrix. Beside the main con-
RDF for Ni–Ni and Cr–Cr pairs. From this data, stituents, Mn and Cu are incorporated within the
particle sizes can be calculated and are presented in precipitates. Fe and Cr are decreasing the further the
Table III. The errors represent the 95 pct confidence distance inwards. The chemical composition of b-NiAl
interval of the fitted radius to the measured RDF. in the over-aged condition (Figure 7(b)) has a higher
The equivalent radii of b-NiAl precipitates derived amount of Ni and Fe, whereas Al is reduced. Mn and
from iso-surface volumes are in reasonable agreement Cu are enriched at 1.5 and 1.0 at. pct, respectively,
with the RDF method. The size increases by more than whereas Cr is almost zero at the plateau value.
a factor of 5 in both cases between peak-aged and
over-aged condition. The numbers are also in very good
agreement with those obtained from XRD. Within the IV. DISCUSSION
analysed volume of over-aged material, Cr23C6 sizes
(diameters) are larger than 20 nm, which is an increase APT is a powerful technique for the detection of
by more than one order of magnitude. Volume fractions clusters and small precipitates. The chemical informa-
and number densities were determined from the iso-sur- tion is recorded, and size and number densities of
face volumes (including edge precipitates). The volume particles can be extracted. However, the analysis is
fraction of intermetallic b-NiAl was slightly higher complex and different challenges must be addressed. In
(6.9 pct) in peak-aged than in the over-aged condition the case of dual-hardening steels, like Hybrid 55 in this
(6.6 pct). This can be understood from higher solubility study, Fe, Cr and Al are integral parts of the alloy and
at the higher aging temperature. The threshold variation have isotopes that share the same mass-to-charge ratio.
shows a broad range for peak-aged volume fraction due Hence, the peak in the mass spectrum at 27 Da has to be
to the diffuse interface between matrix and precipitate. treated carefully. Since the reconstruction can be sepa-
The number density of b-NiAl precipitates was about rated into different POS-files, the different regions of
200 times higher in the peak-aged condition. Even more interest can be individually analysed, which was done in

874—VOLUME 55A, MARCH 2024 METALLURGICAL AND MATERIALS TRANSACTIONS A

 Table II. Average Composition of Matrix and Precipitates as Calculated from Separated POS-Files.

C Al Si V Cr Mn Fe Ni Cu Mo M/C
At. Pct (V + Cr + Fe + Mo)

Peak-Aged
Matrix 0.18 0.72 0.08 0.14 3.59 0.10 93.5 1.48 0.01 0.12
NiAl 1.00 30.3 0.12 0.26 2.46 1.07 36.2 27.7 0.73 0.16
Cr23C6 13.8 0.60 0.22 2.43 29.3 0.34 49.0 1.89 0.03 2.22 6.01
Primary VC 47.7 0.03 0.03 44.5 3.71 0.02 0.11 0.48 0.00 3.30 1.08
Austenite 0.64 4.02 0.13 0.32 5.06 0.23 84.2 4.94 0.10 0.33
Over-Aged
Matrix 0.04 0.09 0.20 0.39 3.91 0.20 92.1 2.50 0.06 0.43
NiAl 0.20 42.2 0.09 0.05 0.23 1.25 9.26 45.4 1.00 0.02
Cr23C6 16.9 0.01 0.01 5.34 56.5 0.54 17.5 0.69 0.00 2.42 4.84
VC 35.9 0.36 0.04 33.1 15.0 0.42 4.42 1.05 0.00 9.05 1.72

Fig. 6—(a) Size distribution of particles obtained using iso-surfaces; (b) normalised RDF of Ni–Ni and Cr–Cr pairs, respectively. Mean
precipitate sizes can be deduced, see text (Color figure online).

Table III. Mean Radii, Volume Fractions and Number Densities of the Different Precipitates in the Investigated Samples;
Iso-Surface Threshold Values were 25 ± 2 at. pct Ni & Al and 12 ± 1 at. pct Cr, respectively.

RRDF [nm] RIso-surface [nm] Volume Fraction [Pct] Number Density [1021 m3]
Peak-Aged
NiAl 1.41 ± 0.03 (1.94) 1.84 (1.76) (9.6) 6.9 (5.0) (2700) 2400 (2100)
Cr23C6 1.82 ± 0.09 (1.99) 1.99 (2.06) (2.4) 1.8 (1.4) (500) 440 (350)
Over-Aged
NiAl 10.74 ± 0.08 (9.82) 9.58 (9.25) (6.8) 6.6 (6.3) (14) 13 (12)

this study. A common method to visualise composition high evaporation fields, notably carbides. Carbon tends
at interfaces or iso-concentration surfaces from particles to diffuse on the surface of the tip. Evaporation happens
are proxigrams. In this case, the reconstruction can not often of several ions or molecules, resulting in a higher
be separated and the peak at 27 Da was decomposed in ratio of multiple hits at the detector as well as detection
proportion to the main isotopes of Fe2+ at 28 Da and of C2, C3 and C4 ions. Multiple hits in close proximity in
Cr2+ at 26 Da, for each point of the proxigrams. time and space at the detector cannot be resolved and
Another challenge is the measurement of phases with therefore the atom species is underestimated in the

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 55A, MARCH 2024—875

Fig. 7—Proxigrams of b-NiAl precipitates in (a) peak-aged and (b) over-aged material. Iso-surface values are 25 at. pct Ni and Al. Note the
different scales on the abscissas (Color figure online).

chemical composition.[26,27] Local magnification, i.e., the Fe in them. However, since they are nm-sized, a
distortion of the reconstruction due to different evapo- significant matrix contribution due to lateral resolution
ration fields of different phases in the material, as well as and local magnification is expected. Since C is often
lateral resolution, limits the depiction of a sharp underestimated in APT experiments, a physical compo-
interface of b-NiAl particles. Additionally, Fe atoms sition matching the M23C6 type is reasonable. This type
are measured within the precipitate from the surround- of carbide is also found in cold-work tool steel after
ing matrix. Lattice parameter estimation from the tempering.[29] The measured composition of Cr23C6 in
b-NiAl (100) peak in the diffractograms in Figure 2 the over-aged condition displayed a higher C content,
yields 2.89(9) Å for the peak-aged condition. This is which might result from a smaller amount of Fe from
slightly larger than the nominal value of 2.881 Å matrix contribution, since the size of the carbides are
(Crystallography Open Database COD ID 9008802), significantly larger. The V-carbide in this material has
indicating a composition close to pure NiAl, but may about the same amounts of V and C. However, there is
indeed suggest a larger presence of Fe as the lattice also about 15 at. pct Cr and 9 at. pct Mo within the
parameter will increase with Fe content. Pike et al.[28] carbides. In literature, it has been shown that M2C can
investigated a wide range of compositions of (Ni,Fe)Al be reached after long aging times, however, with
alloys and typically found lattice parameters ranging M = Cr or Mo.[30,31] The phase diagram of Fe–Cr–V–C
from 2.886 to 2.889 Å for Fe concentrations in the range suggests the existence of M3C and M7C3 beside MC and
15 to 40 at. pct, increasing with increasing Fe content. M23C6 at 700 °C.[32,33] Though, the carbide in this study
In the over-aged condition, the composition in the has about 33 at. pct V, which is more than suggested to
proxigram reaches plateau values. Therefore, it is substitute Fe and Cr in M7C3.[34] The metal to carbon
reasonable to assume a real Fe content of about 8 at. ratio is already below 2 (see Table II) without taking the
pct in the intermetallic precipitate. This is consistent underestimation of C in APT into account. Transition
with the measured lattice parameter, 2.88(3) Å, which is metal carbides are reported to have a wide range of
rather close to the nominal, indicating a lower Fe carbon deficiency without a change in crystal struc-
content. The size evaluation of particles was done as ture.[35,36] Hence, it is reasonable to assume an
well by the parameter-free method of Zhao et al.,[21] under-stochiometric MCx (x < 1) type for this carbide.
where subjective influence of the examiner is avoided. No other carbides are seen in APT, which, in combina-
Only the Ni–Ni RDFs (and Cr–Cr RDF for carbides) tion with the narrow peak width in the diffractograms,
were used since Al has an overlap at one of its main indicates that they are larger and more scarcely dis-
peaks (27 Da). The RDF method provides similar tributed than the M23C6. Hence, no significant contri-
particle sizes as when using iso-concentration surfaces. bution to strengthening is expected.
In the latter case, the iso-value is chosen by the examiner The high strength of this material is based on the high
and influences the determined size. Also, since edge number of nm-sized particles. On one hand, the inter-
particles are excluded in the size distribution, a higher metallic b-NiAl have coherent interfaces and disloca-
proportion of larger carbides is omitted. This is not the tions should be able to move through the particles. The
case for the more isolated b-NiAl particles. strength contribution can therefore be described in the
In the peak-aged condition, the primary VC has simple model from Gerold and Haberkorn[37] as
nearly stochiometric composition with little Cr and Mo
content. The Cr-carbide particles have almost 50 at. pct

876—VOLUME 55A, MARCH 2024 METALLURGICAL AND MATERIALS TRANSACTIONS A

 Ds / f1=2  r1=2 ; ½2
where Ds is the strength increase, f the volume fraction
and r the particle radius. From Eq. [2] it can be seen
that the b-NiAl precipitates contribute about the same
increase in strength between the investigated condi-
tions or even more to the overall strength in the over-
aged condition. On the other hand, Cr23C6 is assumed
to contribute by the Orowan mechanism, causing dis-
location loops. The strength increase can be described
by Eq. [3].[38]
 
Gb k Fig. 8—Elemental distribution in a precipitation-rich plane of
Ds /  ln ½3 peak-aged material. Every Cr23C6 precipitate is connected to at least
k b
one b-NiAl particle.
with G the shear modulus, b the Burgers vector and
the interparticle spacing k defined as follows: are enhanced at higher temperatures. The possible
1=3
exchange of atoms from one precipitate to the other
k¼N  2  r; ½4 facilitates coarsening of the particles within a short
where N is the number density of the particles. Using time.[43] In contrast to steels with higher alloying content
these equations for the Cr23C6 precipitates one can see of Cu, where Cu precipitates readily exchange solute
that the strength contribution drops by one order of atoms with b-NiAl particles, the co-precipitates in
magnitude from peak-aged to over-aged condition. Hybrid 55 are b-NiAl and carbides. Therefore, no
Since the particles are not isolated but co-precipitates immediate exchange of solute elements is expected. In
in a martensitic microstructure with lath and grain the over-aged state, primarily the carbides coarsen,
boundaries, the real deformation mechanism is rather while b-NiAl particles exhibit only a limited size
complex. increase, as was also observed in literature.[40] This is
The sequence in which precipitates form is of special due to the quality of the interface as well as the solubility
interest in the case of dual-hardening steels. In this of the constituent elements. The carbides form an
study, the precipitation of intermetallic particles and incoherent boundary with the matrix, whereas b-NiAl
carbides in Hybrid 55 was investigated in peak-aged and precipitates have coherent interfaces with the matrix and
over-aged condition. After quenching, the alloying a misfit depending on the alloying content. Hence, more
elements are in supersaturation. During the aging diffusion and greater driving force for coarsening is
treatment, Ni and Al form clusters, which become expected for the incoherent carbides. The same phe-
b-NiAl intermetallic particles with Fe still incorpo- nomenon was observed for Cu precipitates and carbides
rated.[17,30] b-NiAl exhibits coherent interfaces with the by Mulholland and Seidman.[31] The limited growth of
martensitic/ferritic matrix and therefore readily precip- the coherent intermetallic particles seems to be the
itates in a homogenous distribution of fine parti- reason for the improved resistance to over-aging com-
cles.[6,13,39,40] Carbides have an incoherent interface pared to other steels for similar applications.
with the matrix. Possible nucleation sites include dislo-
cations, lath and grain boundaries,[41] where C can
agglomerate and eventually form carbides. In this V. CONCLUSIONS
dual-hardening steel the intermetallic b-NiAl precipi-
tates are also nucleation sites for the carbides.[13,30] The dual-hardening steel Hybrid 55 has been inves-
During formation of the intermetallic compounds other tigated by APT and XRD in peak-aged and over-aged
elements, such as Cr, are displaced and therefore in condition. Peak-aging heat treatment at 520 °C for
excess next to the b-NiAl precipitates.[6] This partition 6 hours results in high number densities of nm-sized
of Cr might enhance the formation of carbides as particles of two types: intermetallic b-NiAl particles,
similarly observed for C next to Cu precipitates.[42] having a mean radius of 1.4 nm and a number density of
Hence, ‘‘assisted precipitation’’ of the secondary car- 2.4 9 1024 m3 as well as Cr23C6 with a mean radius of
bides connected to the b-NiAl particles is expected. This 1.8 nm and number density of 4.4 9 1023 m3. A
is visible by the fact that every carbide precipitate in the precipitation-free zone corresponding to a residual
peak-aged condition is in contact with a b-NiAl particle, austenite lath had higher alloying contents within its
as can be seen in Figure 8. The small content of Cu in volume and segregation of carbide forming elements as
Hybrid 55 is found within the b-NiAl particles without well as traces of P at the interface.
clustering. It is reported that Cu prefers the Al sublattice Over-aging at 650 °C for 4 hours results in coarsening
in the B2 ordered structure, since in this way the lattice of the precipitates. However, carbides coarsen at a faster
misfit to the ferritic matrix can be minimized.[43] rate than the intermetallic particles. b-NiAl precipitates
Similarly, Mn is found within the Al sublattice.[44] This have a mean radius of 11 nm after this over-aging
can explain the reduced Al to Ni content in over-aged whereas carbides are measured only in part due to the
precipitates. The coarsening rate of particles depend on limited field of view in APT. Furthermore, under-sto-
the diffusion rate and the matrix solubility. Both factors chiometric VCx (x < 1) secondary carbides with

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 55A, MARCH 2024—877

significant amounts of Mo and Cr were found in the REFERENCES
over-aged sample material. 1. J. Zhao and Z. Jiang: Prog. Mater. Sci., 2018, vol. 94, pp. 174–242.
Every carbide is in contact with at least one intermetal- https://doi.org/10.1016/J.PMATSCI.2018.01.006.
lic particle, indicating the heterogenous nucleation of 2. H.J. Kong and C.T. Liu: Technologies., 2018, vol. 6, p. 36. http
carbides and co-precipitation of the two types of precip- s://doi.org/10.3390/TECHNOLOGIES6010036.
3. Z. Xiong, I. Timokhina, and E. Pereloma: Prog. Mater. Sci., 2021,
itates. The different character of the interfaces seems to be vol. 118, 100764https://doi.org/10.1016/J.PMATSCI.2020.100764.
the reason for the divergent coarsening rates. 4. M. Hättestrand, J.-O. Nilsson, K. Stiller, P. Liu, and M. Ander-
sson: Acta Mater., 2004, vol. 52, pp. 1023–037. https://doi.org/10.
1016/j.actamat.2003.10.048.
5. R.F. Decker, S. Floreen, R.K. Wilson, Maraging steels: recent
ACKNOWLEDGMENTS developments and applications, in Proceedings of Symposium on
TMS Annual Meeting, 1988, pp. 1–38.
The Swedish Innovation Agency VINNOVA is 6. D.H. Ping, M. Ohnuma, Y. Hirakawa, Y. Kadoya, and K. Hono:
kindy acknowledge for funding through the Grant Mater. Sci. Eng. A, 2005, vol. 394, pp. 285–95. https://doi.org/10.
2018-04433. This research was performed in part at 1016/J.MSEA.2004.12.002.
the Chalmers Materials Analysis Laboratory, CMAL. 7. Z. Guo, W. Sha, and D. Vaumousse: Acta Mater., 2003, vol. 51,
pp. 101–16. https://doi.org/10.1016/S1359-6454(02)00353-1.
We acknowledge DESY (Hamburg, Germany), a 8. C.Y. Chen, H.W. Yen, F.H. Kao, W.C. Li, C.Y. Huang, J.R.
member of the Helmholtz Association HGF, for the Yang, and S.H. Wang: Mater. Sci. Eng. A, 2009, vol. 499, pp.
provision of experimental facilities. Parts of this 162–66. https://doi.org/10.1016/J.MSEA.2007.11.110.
research were carried out at beamline P21.2 at PETRA 9. W.B. Lee, S.G. Hong, C.G. Park, K.H. Kim, and S.H. Park: Scr.
III and we would like to thank Ulrich Lienert and Syl- Mater., 2000, vol. 43, pp. 319–24. https://doi.org/10.1016/S1359-
6462(00)00411-5.
vio Haas for assistance. Beamtime was allocated for 10. P.M. Kelly: Int. Metall. Rev., 1973, vol. 18, pp. 31–37. https://doi.
proposal I-20190135 EC. SJ gratefully acknowledges org/10.1179/IMTLR.1973.18.1.31.
Dr. Kristina Lindgren, now at RISE Research Insti- 11. J.-E. Andersson, F. Lindberg, and P. Ölund: Light. Des. Worldw.,
tutes of Sweden, for discussions regarding the calcula- 2019, vol. 12, pp. 34–39. https://doi.org/10.1007/s41777-018-0063-3.
12. W.M. Garrison and M.S. Bhat: Metall. Trans. A, 1988, vol. 19A,
tion of particle size and the RDF method. pp. 1751–60. https://doi.org/10.1007/BF02645143.
13. F. Danoix, R. Danoix, J. Akre, A. Grellier, and D. Delagnes: J.
Microsc., 2011, vol. 244, pp. 305–310. https://doi.org/10.1111/J.
1365-2818.2011.03537.X.
FUNDING 14. C.Y. Huang and H.W. Yen: Mater. Charact., 2021, vol. 178, 11-
1216https://doi.org/10.1016/j.matchar.2021.111216.
Open access funding provided by Chalmers Uni- 15. M. Hofinger, M. Staudacher, M. Ognianov, C. Turk, H. Leitner,
versity of Technology. and R. Schnitzer: Micron, 2019, vol. 120, pp. 48–56. https://doi.
org/10.1016/j.micron.2019.02.004.
CONFLICT OF INTEREST 16. Y.C. Lin, I.E. McCarroll, Y.T. Lin, W.C. Chung, J.M. Cairney,
and H.W. Yen: Acta Mater., 2020, vol. 196, pp. 516–27. https://d
The authors declare that they have no conflict of oi.org/10.1016/j.actamat.2020.06.046.
interest. 17. M. Hofinger, C. Turk, M. Ognianov, H. Leitner, and R. Schnitzer:
Mater. Charact., 2020, vol. 160, 110126https://doi.org/10.1016/j.
matchar.2020.110126.
OPEN ACCESS 18. Z.B. Jiao, J.H. Luan, M.K. Miller, Y.W. Chung, and C.T. Liu:
Mater. Today, 2017, vol. 20, pp. 142–54. https://doi.org/10.1016/J.
This article is licensed under a Creative Commons MATTOD.2016.07.002.
Attribution 4.0 International License, which permits 19. W. Lefebvre-Ulrikson, F. Vurpillot, and X. Sauvage, eds.: Atom Probe
use, sharing, adaptation, distribution and reproduction Tomography: Put Theory Into Practice, Academic Press, London, 2016.
in any medium or format, as long as you give appro- 20. M.K. Miller, R.G. Forbes, Atom-Probe Tomography. Local Elec-
trode Atom Probe, Springer, New York, 2014, pp. 1–423. https://d
priate credit to the original author(s) and the source, oi.org/10.1007/978-1-4899-7430-3/COVER.
provide a link to the Creative Commons licence, and 21. H. Zhao, B. Gault, D. Ponge, D. Raabe, and F. De Geuser: Scr.
indicate if changes were made. The images or other Mater., 2018, vol. 154, pp. 106–10. https://doi.org/10.1016/j.scrip
third party material in this article are included in the tamat.2018.05.024.
22. F. Vurpillot, A. Bostel, and D. Blavette: Appl. Phys. Lett., 2000,
article’s Creative Commons licence, unless indicated vol. 76, pp. 3127–29. https://doi.org/10.1063/1.126545.
otherwise in a credit line to the material. If material is 23. D.J. Larson, B. Gault, B.P. Geiser, F. De Geuser, and F.
not included in the article’s Creative Commons licence Vurpillot: Curr. Opin. Solid State Mater. Sci., 2013, vol. 17, pp.
and your intended use is not permitted by statutory 236–47. https://doi.org/10.1016/j.cossms.2013.09.002.
regulation or exceeds the permitted use, you will need 24. G. Ashiotis, A. Deschildre, Z. Nawaz, J.P. Wright, D. Karkoulis,
F.E. Picca, and J. Me Kieffer: J. Appl. Crystallogr., 2015, vol. 48,
to obtain permission directly from the copyright pp. 510–19. https://doi.org/10.1107/S1600576715004306.
holder. To view a copy of this licence, visit http://creat 25. J. Kieffer, V. Valls, N. Blanc, and C. Hennig: J. Synchrotron. Rad.,
ivecommons.org/licenses/by/4.0/. 2020, vol. 27, pp. 558–66. https://doi.org/10.1107/S1600577520000776.
26. M. Thuvander, J. Weidow, J. Angseryd, L.K.L. Falk, F. Liu, M.
Sonestedt, K. Stiller, and H.O. Andrén: Ultramicroscopy, 2011, vol.
111, pp. 604–08. https://doi.org/10.1016/J.ULTRAMIC.2010.12.024.
SUPPLEMENTARY INFORMATION 27. D. Shinde, M. Thuvander, A. Rehan, S. Ejnermark, and K. Stiller:
Microsc. Microanal., 2019, vol. 25, pp. 454–61. https://doi.org/10.
1017/S1431927619000424.
The online version contains supplementary material 28. L.M. Pike, I.M. Anderson, C.T. Liu, and Y.A. Chang: Acta Mater., 2002,
available at https://doi.org/10.1007/s11661-023-07291-7. vol. 50, pp. 3859–79. https://doi.org/10.1016/S1359-6454(02)00192-1.

878—VOLUME 55A, MARCH 2024 METALLURGICAL AND MATERIALS TRANSACTIONS A

29. H. Kim, J.Y. Kang, D. Son, T.H. Lee, and K.M. Cho: Mater. 37. V. Gerold and H. Haberkorn: Phys. Status Solidi, 1966, vol. 16,
Charact., 2015, vol. 107, pp. 376–85. https://doi.org/10.1016/J.M pp. 675–84. https://doi.org/10.1002/pssb.19660160234.
ATCHAR.2015.08.001. 38. J. Friedel: Dislocations, Pergamon Press, Oxford, 1964.
30. D. Delagnes, F. Pettinari-Sturmel, M.H. Mathon, R. Danoix, F. 39. R. Schnitzer, G.A. Zickler, E. Lach, H. Clemens, S. Zinner, T.
Danoix, C. Bellot, P. Lamesle, and A. Grellier: Acta Mater., 2012, Lippmann, and H. Leitner: Mater. Sci. Eng. A, 2010, vol. 527, pp.
vol. 60, pp. 5877–88. https://doi.org/10.1016/J.ACTAMAT.2012. 2065–70. https://doi.org/10.1016/J.MSEA.2009.11.046.
07.030. 40. S.D. Erlach, H. Leitner, M. Bischof, H. Clemens, F. Danoix, D.
31. M.D. Mulholland and D.N. Seidman: Acta Mater., 2011, vol. 59, Lemarchand, and I. Siller: Mater. Sci. Eng. A, 2006, vol. 429, pp.
pp. 1881–97. https://doi.org/10.1016/J.ACTAMAT.2010.11.054. 96–106. https://doi.org/10.1016/J.MSEA.2006.05.071.
32. J. Bratberg and K. Frisk: Metall. Mater. Trans. A Phys. Metall. 41. J.P. Adamson and J.W. Martin: Acta Metall., 1971, vol. 19, pp.
Mater. Sci., 2004, vol. 35A, pp. 3649–63. https://doi.org/10.1007/ 1015–18. https://doi.org/10.1016/0001-6160(71)90033-2.
s11661-004-0271-9. 42. Q. Liu and S. Zhao: MRS Commun., 2012, vol. 2, pp. 127–32.
33. V. Raghavan: J. Phase Equilib. Diffus., 2007, vol. 28, pp. 274–76. h https://doi.org/10.1557/mrc.2012.21.
ttps://doi.org/10.1007/S11669-007-9069-Y. 43. Z.B. Jiao, J.H. Luan, M.K. Miller, C.Y. Yu, and C.T. Liu: Sci.
34. M. Filipovic, E. Romhanji, and Z. Kamberovic: ISIJ Int., 2012, Rep., 2016, vol. 61(6), pp. 1–3. https://doi.org/10.1038/srep21364.
vol. 52, pp. 2200–204. https://doi.org/10.2355/isijinternational.52. 44. Z.B. Jiao, J.H. Luan, M.K. Miller, C.Y. Yu, and C.T. Liu: Acta
2200. Mater., 2015, vol. 84, pp. 283–91. https://doi.org/10.1016/J.ACT
35. W.S. Williams: Mater. Sci. Eng., 1988, vol. 105–106, p. 1. https://d AMAT.2014.10.065.
oi.org/10.1016/0025-5416(88)90474-0.
36. Y. Oba, S. Koppoju, M. Ohnuma, T. Murakami, H. Hatano, K. Publisher’s Note Springer Nature remains neutral with regard to
Sasakawa, A. Kitahara, and J.I. Suzuki: ISIJ Int., 2011, vol. 51, jurisdictional claims in published maps and institutional affiliations.
pp. 1852–58. https://doi.org/10.2355/isijinternational.51.1852.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 55A, MARCH 2024—879