Journal of Alloys and Compounds 870 (2021) 159484

Contents lists available at ScienceDirect

Journal of Alloys and Compounds

journal homepage: www.elsevier.com/locate/jalcom

Physical and thermodynamic simulations of gamma-prime precipitation

in Haynes® 282® using arc heat treatment ]]
]]]]]]
]]

⁎
Kevin Vattappara a, Vahid A Hosseini b, , Ceena Joseph c, Fabian Hanning b, Joel Andersson b
a
Department of Materials Science and Engineering, School of Industrial Engineering and Management, KTH Royal Institute of Technology, SE-100 44 Stockholm,
Sweden
b
Department of Engineering Science, University West, SE-461 86 Trollhättan, Sweden
c
GKN Aerospace AB, SE-461 38 Trollhättan, Sweden

a r t i cl e i nfo a bstr ac t

Article history: Haynes 282 is a Ni-based gamma prime (γ′) strengthening alloy with a balanced combination of high
Received 5 October 2020 temperature properties and fabricability. This paper aims to study the evolution of the microstructure and
Received in revised form 8 March 2021 hardness using a novel physical simulation method called arc heat treatment, followed by thermodynamic
Accepted 9 March 2021
modeling of γ′ precipitation. For the arc heat treatment, a steady state temperature gradient was generated
Available online 15 March 2021
using a stationary tungsten inert gas arc on a sample mounted onto a water-cooled chamber. The steady
state condition ranged from room temperature to the liquidus and was achieved within the first few sec­
Keywords:
High-temperature alloys onds. Aged and solutionized samples were arc heat treated for 1.5 min, 30 min, and 4 h. The experiments
Precipitation were complemented with temperature modeling, equilibrium calculations, and γ′ precipitation simulations.
Grain boundaries A unique graded microstructure formed in the arc heat-treated samples. It consisted of a fusion zone with a
Kinetics dendritic microstructure; a region with the dissolution of all secondary phases (MC carbides, grain
Thermodynamic modeling boundary carbides, and γ′); a region with MC and grain boundary carbides; a γ′ precipitation zone; and the
base metal. The temperature range of the dissolution area extended to lower temperatures with increasing
arc heat treatment dwell time. The γ′ precipitation zone showed a distinct etching response coupled with
high hardness. The hardness and the temperature range of the γ′ precipitation zone increased with in­
creasing arc heat treatment time. The γ′ radii increased with increasing time and temperature. The γ′
precipitation model, simulated with TC Prisma, showed very good agreement with the experimental results.
Finally, the results were used to develop time-temperature precipitation and hardness diagrams.
© 2021 The Authors. Published by Elsevier B.V.
CC_BY_4.0

1. Introduction the grain boundaries are essential for improving the creep re­
sistance, strength, and grain size stability [6,9,10].
Nickel-based superalloys are mainly used in the aerospace and Haynes 282 is a strengthening alloy that was developed in
turbine industries, where high resistance to creep, fatigue, and cor­ 2005 [3]. Its major alloying elements are Cr, Co, Mo, Al, and Ti [3,11].
rosion are required [1,2]. Their major precipitation strengthening The alloy was designed for fabricability and creep resistance, and its
phase, the phase, has an L12 structure with a general composition fabricability properties are better than its predecessors, such as
of Ni3(Al, Ti) [3,4]. The strength of the alloys is primarily based on Waspaloy and Rene 41 [3]. The standard heat treatment (SHT) for
the size, fraction, morphology, and distribution of the phase [5–7]. Haynes 282 begins with a solutionizing process at 1121–1149 °C to
In addition to , various carbides contribute to the properties of dissolve and grain boundary carbides. This is followed by a two-
nickel-based superalloys. Primary carbides (MC) form by the reac­ step aging process: 1010 °C/2 h/air-cooled for the precipitation of
tion of C with Ti, Ta, and Hf during solidification. In addition to MC grain boundary carbides and 788 °C/8 h/air-cooled for precipita­
carbides, M23C6 (Cr rich) and M6C (Mo rich) with various morphol­ tion. After the SHT, the precipitates have a spherical morphology
ogies (discrete, blocky, and film-like) typically precipitate on the with a unimodal distribution [3]. Joseph et al. [6,8] showed that with
grain boundaries [1,3,8]. Carbides with a discrete morphology along alternative heat treatments, a cuboidal morphology and bimodal
distribution can be achieved. The effects of the various heat treat­
ments on the precipitation of in Haynes 282 are still of great in­
⁎
Corresponding author.
terest to the research and industrial communities.
E-mail address: [email protected] (V. A Hosseini).

https://doi.org/10.1016/j.jallcom.2021.159484
0925-8388/© 2021 The Authors. Published by Elsevier B.V.
CC_BY_4.0
K. Vattappara, V. A Hosseini, C. Joseph et al. Journal of Alloys and Compounds 870 (2021) 159484

The progress of isothermal transformations is visualized by plot­

ting the fraction of phase transformation with time at various tem­
peratures, which is known as a time-temperature-transformation
(TTT) diagram. This information provides an important input for
designing proper heat treatments for different alloys, particularly
Ni-based superalloys. Various techniques, such as dilatometry [12],
electrical resistivity [13], Gleeble thermomechanical simulation
[14], differential thermal analysis (DTA) [13], and furnace treatments
[8,15], are utilized to create TTT diagrams by measuring the change in
microstructure and material properties resulting from phase trans­
formations. Conventional methods, such as furnace heat treatments,
involve extensive experiments and time-consuming material char­
acterizations [16,17]. In addition, the heating rate is not sufficiently
high to provide the correct information for shorter heat treatment
times. The methods using thermomechanical simulation tools, such as
Gleeble, are applicable for the development of TTT diagrams with high
flexibility; however, their accessibility is limited, and experimenting
with melted metal can be challenging. The technique is expensive, and
a large number of samples is needed to produce TTT diagrams. In 2017,
Hosseini et al. [17] introduced a new heat treatment technique that
produces a graded microstructure, which encompasses temperatures
from room temperature to the liquidus in a single test piece. The
method is called arc heat treatment, in which a stationary tungsten
inert gas (TIG) arc is applied to a disc mounted onto a water-cooled
chamber for a selected time and produces constant temperatures at
different locations within the sample. For each heat treatment dwell Fig. 1. (a) Schematic showing the experimental setup for the arc heat treatment [17].
time, only one single sample is needed to cover all temperatures, (b) Arc heat treatment in progress with various areas marked as shown in the sche­
matic. (c) Actual experimental setup with thermocouple attachment. The thermo­
which not only considerably reduces the heat treatment and
couples are arranged in a circular pattern to capture the temperatures around the
characterization time but also increases the understanding of the fusion zone.
entire range of microstructure evolution [18–21]. This method was
primarily developed for duplex stainless steels.
Apart from the aforementioned experimental techniques, com­ 100 mm in diameter × 6 mm in thickness were used for the experi­
putational thermodynamics is also used to evaluate the equilibrium ment. Two initial conditions were investigated in this study: aged
phase fractions, precipitation and coarsening kinetics, diffusion of (fully age-hardened by a standard heat-treatment; solutionized at
phases, and phase transformations. Different software packages have 1120–1149 °C followed by quenching; age-hardened in two steps:
been developed, which are predominately based on CALPHAD 1010 °C for 2 h followed by air cooling and 788 °C for 8 h followed by
(CALculation of PHAse Diagrams). The Thermo-Calc software air cooling) and solutionized (1120 °C for 90 min).
package [22] uses material input data calculated using the CALPHAD
databases to provide information about the phase fractions [22,23]. 2.2. Arc heat treatment
Lindwall et al. [23] studied the precipitation behavior of an addi­
tively manufactured IN625 alloy using TC Prisma and constructed a A schematic and the actual setup of the arc heat treatment
TTT diagram for the different phases, which was in good agreement equipment are shown in Fig. 1. A stationary TIG arc was applied on a
with the experimental TTT diagram. disc that was water cooled from the backside to produce the steady-
Since the development of the Haynes 282 alloy in 2005, its micro­ state fusion zone and temperature gradient. The sample was sealed
structure, mechanical properties, weldability, and thermomechanical with a holder (formed by a rubber gasket, as shown in Fig. 1). Water
response have been investigated in different studies [6,24–28]. However, was circulated in this chamber for symmetric heat removal during
information about the precipitation and stability of the phase, the the heat treatment (the water flow pattern is marked by arrows in
high-temperature stability of carbides, and the validity of thermo­ Fig. 1). The water flow rate was maintained at 1.4 L/min. The inlet
dynamic calculations is still lacking [28,29]. Therefore, this study aims to water was maintained at 8.5 ± 1 °C throughout the experiment to
investigate the evolution of the microstructure and hardness for short maintain a steady state condition in the sample.
heat treatment dwell times using the novel arc heat treatment approach A TIG torch was used to produce the arc with a programmable
combined with thermodynamic and phase precipitation simulations. TIG COMMANDER AC/DC 400 power source. The electrode had a
diameter of 2.4 mm with a tip angle of 60°. The arc length was set to
2. Methodology 3 mm. Fig. 1b shows the heat treatment in progress. Thermal cycles
at various locations were recorded during the 30 min arc heat
2.1. Material treatment of the solutionized sample using several K-type thermo­
couples (TC). The results were used to calibrate the temperature
Haynes 282 discs (cut from a wrought bar) with the chemical distribution model and correlate the thermal cycles at various lo­
composition (in wt%) listed in Table 1 and with dimensions of cations with the resulting microhardness and microstructure. The

Table 1
Chemical composition in wt% of the wrought Haynes 282.

Ni Cr Co Mo Ti Al Si B C Fe Mn P

Bal. 19.55 10.46 8.70 2.02 1.45 0.07 0.004 0.063 1.17 0.06 0.003

2
K. Vattappara, V. A Hosseini, C. Joseph et al. Journal of Alloys and Compounds 870 (2021) 159484

arc heat treatment was conducted for 1.5 min, 30 min, and 4 h for
each initial condition.

2.3. Characterization

2.3.1. Sample preparation

The arc heat-treated samples were waterjet cut, rough ground
from 80 to 500 grit, and polished with 9, 3, and 0.5 µm alumina.
Finally, the samples were electrolytically etched using 2% oxalic acid
at 3 V for 10 s

2.3.2. Optical microscopy

Stereo microscopy was performed using an Olympus SZX stereo Fig. 3. Comparison of the calibration and experimental radii [8,35].
microscope to characterize the fusion zone dimensions and the
secondary phase dissolution area. A ZEISS Axio Imager 2 optical
microscope was used to characterize the microstructure, and images was converted from a 3D to a 2D geometry to reduce the compu­
were obtained at 50 × and 100 × magnifications, which were then tational time. Because the aim of creating this model was to predict
stitched together using the ZEISS software ZEN lite. the constant temperature distribution in the sample, the fusion zone
boundary was used as the boundary condition [17,30]. Material
property data, including the density (g/cm3), thermal conductivity
2.3.3. Hardness testing
(W/(m*K)), and specific heat (J/(g*K)), were calculated using the
Hardness mapping with more than 400 indents was performed
JMatPro software (version 6.2.1) because the material data sheet
on the cross section of each sample using a Qness 10 microhardness
only provided the physical properties for temperatures below
tester with a load of 200 g.
1000 °C [11,31].

2.3.4. Scanning electron microscopy

2.4.2. Equilibrium and precipitation simulations
Scanning electron microscopy (SEM) analysis was performed
The equilibrium phase fraction was calculated using the JMatPro
using a LEO 1550 high-resolution field emission gun scanning elec­
software (version 6.2.1) with the actual chemical composition of the
tron microscope with a working distance of 9 mm and an accel­
samples.
erating voltage of 10 kV in the secondary electron (SE) mode using a
Precipitation calculations were performed using TC Prisma
high-resolution in-lens SE detector. Images were obtained at various
(2019a) with the initial state of solutionized conditions. The ther­
locations in the samples to characterize the microstructure from the
modynamic and mobility databases used for these calculations are
fusion zone to the base metal. The γ′ precipitate size was analyzed
TCNI9 and MOBNI4.1/5, respectively. The input for the simulations
using SEM images and ImageJ software.
included the material composition and the desired temperatures,
times, and phases for defining the matrix as well as phase pre­
2.4. Simulation cipitation. Of the several parameters that were used for the calcu­
lations, two parameters (interfacial energy and nucleation sites)
2.4.1. Temperature modeling were calibrated for the analysis in this work. A test matrix for the
The COMSOL software (version 5.4) was used to model the two intended parameters was created, and a set of values for the
temperature distribution of the solutionized samples during the arc parameters was selected based on values obtained from the litera­
heat treatment. The model is based on the steady-state energy ture [8,32,33] and also obtained from the TC Prisma user guide [34].
conservation equation with temperature-dependent variables: These parameters were optimized by calibrating them with the radii
( CT ) of the precipitates obtained in Ref. [8] for temperatures of 750,
·( T) = S + Q 850, and 950 °C with a mill-annealed initial condition. The optimized
t (1)
value for interfacial energy was 10 mJ/m2 , which is in close agree­
where is the material density, C is the specific heat capacity, T is ment with the value determined by Fahrmann et al. [29]. The
the temperature difference, is the thermal conductivity, S is the number of nucleation sites was determined to be 8.68 × 1031 m−3. A
source term for the phase change, and Q is the heat generation term. comparison of the radii with the experimental values is shown in
The source term and the heat generation terms are not considered Fig. 3. The calibrated values were further used as inputs for the
for this calculation, and the model has been adapted from an earlier thermodynamic simulations.
study performed by Kumara [30].
The various regions shown in Fig. 2 are considered to be the 3. Results
boundary conditions for the model, where the values for the top and
bottom boundary conditions of the modeled sample were chosen 3.1. Temperature modeling
using real-time TC readings taken on the sample and the water
temperature. As the heat transfer was axisymmetric, the geometry The simulated temperature distribution map is shown in Fig. 4a,
in which the temperatures were calculated based on the afore­
mentioned boundary conditions and the temperature readings from
the TC measurements.
An area below the fusion zone (along the z-axis or the symmetry
line, marked by a-a) was used to analyze the microstructural
changes (moving from the fusion zone to the base of the sample).
The temperature profile along the line a-a is shown in Fig. 4b. The
fusion zone geometry was different for the 30 min and 4 h samples,
Fig. 2. Domain for temperature modeling. The terms marked on the image depict the which influenced the predicted temperatures. Therefore, two
various regions of interest in the sample. different time profiles were calculated for the microstructural

3
K. Vattappara, V. A Hosseini, C. Joseph et al. Journal of Alloys and Compounds 870 (2021) 159484

locations in the cross section were exposed to a temperature similar to

that shown in Fig. 4a. Interestingly, different locations have different
etching responses, as shown in Fig. 7. This indicates the formation of
different microstructures in the sample during the arc heat treatment.
As explained above, MC carbides were present in the micro­
structure under both initial conditions. After arc heat treatment, it
areas beneath the fusion zone did not contain MC carbides, as in­
dicated for the aged samples in Fig. 6. The MC carbide dissolution
area increased with increasing arc heat treatment dwell time.
The optical micrographs in Fig. 7 show the etching response for
two different initial conditions and three arc heat treatment dwell
times. Far from the fusion zone in the base metal, the aged sample
showed a darker etching response compared to that of the solutio­
nized sample due to the precipitates. In contrast, the etching re­
sponse near the fusion zone (bright area) was similar in both
Fig. 4. (a) Temperature profile for 30 min sample simulated using the COMSOL
cases. Outside the bright area, at temperatures below 1000 °C, a
software. (b) Distance vs. temperature for the different heat treatment dwell times dark-etched zone formed in all samples. This is where begins to be
across the a-a section. The different gradients for the 30 min and 4 h samples were stable in this alloy. The width of the zone increased with in­
due to the different fusion zone geometries. creasing heat-treatment dwell time for both initial conditions.
Fig. 8a shows the microstructural changes that occurred in the
characterization. The change in the fusion zone geometry should be solutionized samples. The micrograph shows where MC carbides
investigated in future studies. were not observed, grain boundary carbides formed, and pre­
cipitated. In Fig. 8b, no secondary phases are present in the top of the
3.2. Initial conditions for the samples image, but fine MC carbides are still visible in the bottom of the
image. This boundary moved toward lower temperatures with in­
The samples have two different initial conditions, aged and so­ creasing heat-treatment dwell time. Below the area of dissolution,
lutionized, with hardness values of 336 and 185 HV, respectively. The an area of the grain boundary carbide precipitation was observed, as
initial microstructures for both conditions are shown in Fig. 5. In the shown in Fig. 8c, which contained grain boundary carbides and MC
solutionized samples, small and large MC carbides were 2–4 µm and carbides. The next zone, shown in Fig. 8d, was dark-etched owing to
7–12 µm, respectively, which fit well with the size reported by Liu the precipitation of .
et al. [36]. (only large MC carbides were observed, while the grain
boundaries were clean). In the aged samples, carbides were also 3.3. Scanning electron microscopy
observed along the grain boundaries.
Cross sections of the samples after arc heat treatment are shown in An SEM micrograph of the fusion zone boundary in the solutio­
Fig. 6 and Fig. 7, which includes the fusion zone in the center. This zone nized sample is shown in Fig. 9. In the top of the micrograph, the
was in the liquid state during the arc heat treatment. The other dendritic microstructure and inter-dendritic phases are visible. In

Fig. 5. (a) Optical micrograph of the solutionized sample with MC carbides, (b) SEM micrograph of solutionized sample showing the MC carbides with higher resolution but clean
grain boundaries, (c) Optical micrograph of the aged sample showing grain boundary carbides, (d) SEM micrograph of the aged sample showing the γ′ precipitates in the aged
sample.

4
K. Vattappara, V. A Hosseini, C. Joseph et al. Journal of Alloys and Compounds 870 (2021) 159484

Fig. 6. Stereo microscope images showing the cross sections, including the fusion zone (FZ) and the dissolution zone (DZ) of MC carbides (original images on the left-hand side
and corresponding images with annotation on the right-hand side), of the aged + arc heat-treated samples. Dissolution of the secondary phases below the fusion zone began in the
1.5 min sample. The 30 min and 4 h samples show larger areas subjected to dissolution.

the bottom of the micrograph, the dissolution area is visible where temperature values. The precipitate radii decreased with decreasing
all phases were dissolved, as discussed in the previous section. temperature, as detailed in Table 2. The morphology of the pre­
The evolution of the radii in the solutionized samples is shown cipitates is spherical. It should be noted that at lower temperatures
in Fig. 10. The temperatures indicated in the figure are based on the (below 860 °C) and for shorter heat treatment dwell times, the
distance measured from the fusion zone coupled with the modeled particles were too small to be measured.

Fig. 7. Evolution of γ′ precipitation zone with increasing dwell time for both initial conditions. The aged samples etched in the γ′ zone and the base metal, while the solutionized
samples only etched in the γ′ zone.

5
K. Vattappara, V. A Hosseini, C. Joseph et al. Journal of Alloys and Compounds 870 (2021) 159484

Fig. 8. (a) Variation in the microstructure in the different regions of the sample arc heat-treated for 4 h, including (b) MC carbide dissolution, (c) grain boundary carbide
formation, and (d) γ′ formation.

As detailed in Table 2, at approximately 1000 °C, the radii region (336 HV) is similar to that of the zone in the 1.5 min and
reached ~24 nm after 30 min and ~45 nm after 4 h. However, at 30 min arc heat-treated solutionized samples, which confirms the
870 °C, they were only ~10 nm after 30 min and ~23 nm after 4 h. presence of in the aged metal. In the solutionized samples, high
Fig. 11 illustrates the grain boundary carbides and distribution of hardness values were observed in the zone, which were indicated
precipitates along the grain boundaries in the solutionized sam­ by the distinct etching. As can be seen, the dark-etching and high-
ples. The grain boundary carbides change from a discrete mor­ hardness zones are in good agreement with each other. The maps for
phology to a continuous film morphology with decreasing the arc heat-treated samples show increasing hardness within the
temperature. zone for longer heat treatment times, which is more evident in the
At higher temperatures, a lower density of particles was seen solutionized samples. In addition to the zone, the hardness in the
next to the grain boundary carbides. The width of this region de­ aged samples dropped considerably above the precipitation zone,
creased with decreasing temperature, where a more homogenous as shown in Fig. 12. This is where the dissolution of phases occurred.
distribution was found at lower temperatures.
3.5. Thermodynamic calculations
3.4. Microhardness
3.5.1. Equilibrium calculations
Fig. 12 shows a microhardness map for the various arc heat- Fig. 13 shows an equilibrium phase fraction diagram for Haynes
treated samples. In the aged samples, the hardness of the base metal 282 calculated using the JMatPro software (version 6.2.1). The

Fig. 9. SEM micrograph showing the fusion zone with a dendritic structure and MC carbides that formed owing to segregation in the fusion zone.

6
K. Vattappara, V. A Hosseini, C. Joseph et al. Journal of Alloys and Compounds 870 (2021) 159484

Fig. 10. Evolution of the γ′ precipitates within the γ′ zone for the solutionized samples.

Table 2 metal (6). A previous study of arc heat treatment on duplex stainless
γ′ particle radii at various times and temperatures for the solutionized samples. steel also showed the formation of a graded microstructure [38].
1.5 min 30 min (Average Std 4 h (Average Std
Dev: 1.97) Dev: 7.14) 4.1.1. High-temperature stability of phases (regions 2 and 3)
̴ 990–1000 °C Unresolvable 24 nm 45 nm In the initial condition of the solutionized samples, all phases
̴ 970–980 °C Unresolvable 21 nm 43 nm were removed except for MC carbides. After arc heat treatment, the
̴ 950–960 °C Unresolvable 20 nm 35 nm MC carbides were dissolved in region 2, while the equilibrium phase
̴ 930–940 °C Unresolvable 19 nm 28 nm
fraction (Fig. 15) indicates that they should be stable. This area had
̴ 900–920 °C Unresolvable 14 nm 26 nm
̴ 885–895 °C Unresolvable 13 nm Unresolvable the lowest hardness compared to those of the other regions. In re­
̴ 860–870 °C Unresolvable 10 nm 23 nm gion 2 of the aged samples, all phases (MC carbides, grain boundary
̴ 840–850 °C Unresolvable Unresolvable 21 nm carbides, and ) were dissolved, and the hardness dropped mark­
edly compared to the initial aged condition, as expected. The in­
creasing size of region 2 with time is the result of chemical
temperature ranges of interest are labeled 1–5, showing the dif­ homogenization [39–41]; the dissolution of MC carbides occurs at
ferent regions observed in the microstructure, and are discussed in higher rates at higher temperatures because of faster diffusion. The
Section 4.1. Based on the phase diagram, MC carbides form during dissolution of MC carbides indicates that they are not stable, at least
solidification and are stable at high temperatures. M6C carbides are during the arc heat treatment. Further study is needed to verify this
stable below 1180 °C and forms below 1000 °C. observation by heat treatment in a furnace. If confirmed, an im­
provement in the thermodynamic databases is needed.
3.5.2. TC Prisma calculations
The precipitation calculations were performed for the solu­ 4.1.2. Grain boundary carbides (regions 4 and 5)
tionized state. Particle sizes were calculated for 760, 820, 950, and The grain boundary carbide morphology found in the high-
980 °C, as shown in Fig. 14. The temperatures 760 and 820 °C were temperature region has also been reported by Polkowska et al. [42].
chosen to compare the simulated values with the values reported by The precipitate-free zone next to grain boundary carbides at high
Haas et al. [37], and the other temperatures (950 and 980 °C) were temperatures was also observed by Brittan et al. [43]. This phe­
chosen to compare their simulated values with the experimentally nomenon can lead to lower mechanical properties during service.
obtained values in this work. As can be seen, higher temperatures Alexandratou et al. [44] reported the formation of γ′ close to grain
result in larger precipitates. boundary carbides at lower temperatures (after a standard heat
treatment; solutionized at 1120–1149 °C followed by quenching;
4. Discussion age-hardened in two steps: 1010 °C for 2 h followed by air cooling
and 788 °C for 8 h followed by air cooling), similar to Fig. 11f, which
4.1. Microstructural map for the arc heat-treated samples can result in improved mechanical properties [45].

Correlating the arc heat-treated cross-section and modeled 4.1.3. precipitation (region 5)
temperature profiles shows that the temperature range and the For the arc heat treatment dwell times of 30 min and 4 h, the γ′
etching response are in close agreement with the thermodynamic radii are compared with the predicted values from TC Prisma in
calculations (Fig. 15). The following regions can be defined: the fu­ Fig. 16. The TC Prisma model was calibrated based on previous ex­
sion zone (1), dissolution of phases (2), primary carbides (3), perimental size measurements [8]. At 980 °C, the experimental
grain boundary carbide precipitation (4), zone (5), and finally base and calculated values were 21 and 23 nm for the 30 min samples and

7
K. Vattappara, V. A Hosseini, C. Joseph et al. Journal of Alloys and Compounds 870 (2021) 159484

Fig. 11. Micrographs of the grain boundary carbides and along the grain boundary in the solutionized sample arc heat-treated for 4 h. (a) Grain boundary at a higher
temperature showing only grain boundary carbides with a discrete morphology. (b) Grain boundary area within the beginning of the zone showing a low fraction of
precipitates within the grain and a depleted region near the grain boundary. (c) & (d) Grain boundary at a lower temperature showing a higher fraction of precipitates in the grain
but still a reduced fraction near the grain boundary. (e) & (f) A similar distribution of precipitates is shown in the grain boundaries area and the grain interior.

Fig. 12. Microhardness map showing the evolution of the hardness. The aged samples show higher initial hardness values than those of the solutionized samples. Arc heat
treatment increased the hardness within the regions showing a darker etching response.

8
K. Vattappara, V. A Hosseini, C. Joseph et al. Journal of Alloys and Compounds 870 (2021) 159484

Fig. 16. Comparison of the experimentally obtained and simulated γ′ radii.

Fig. 13. Equilibrium diagram for Haynes 282 calculated using the JMatPro software
(version 6.2.1). The temperature ranges of interest are labeled 1–5, which are corre­
lated with the microstructure in Section 4.1.

Fig. 17. Time-temperature hardness diagram of solutionized and arc heat treated
sample for a hardness of 350 HV.

760 °C for 30 min yielded ~3 nm radii, which is lower than that in

Haas et al. [37].
Fig. 14. Simulated precipitate sizes at 760, 820, 950, and 980 °C. All simulations
Therefore, the observations showed that precipitation simula­
were performed for the solutionized initial condition. The dashed lines indicate the tions are in close agreement with the results of this study for the
heat treatment times of 30 min and 4 h. solutionized initial condition, but they underestimate the particle
radii compared to results from other studies. The possible reasons
for these observations are as follows:

• The initial state of the material: the thermomechanical history

before the experiment can substantially influence the results. For
instance, the mill-annealed condition is expected to contain
some before aging, which could result in larger radii after
aging.
• Heating rate to the specific soak temperature: In other studies
[37,46], generally slower heating rates to the soak temperatures
(2 min to the desired temperature in furnaces based on our ob­
servations) were achieved compared to the rapid heating (10 s) in
this study. Therefore, it is expected that some will form before
reaching the desired temperature in the furnace, resulting in
larger sizes in those studies. Arc heat treatment, in contrast,
Fig. 15. Various regions and their corresponding temperature range in the 4 h arc
heat-treated sample. provided rapid heating and cooling and represented the results
closer to the selected temperatures.

43 and 46 nm for the 4 h samples, respectively. At 950 °C, the ex­ 4.2. TTT diagrams
perimental and calculated values were 20 and 19 nm for the 30 min
samples and 35 and 38 nm for the 4 h samples, respectively. Based on the results presented in the previous sections, time-
Therefore, the radii observed in SEM and predicted with TC Prisma temperature-hardness (TTH) and time-temperature-precipitation
were in close agreement with each other. The experimental values (TTP) diagrams were generated for the precipitates.
were reported for a mill-annealed sheet material at 950 °C by Joseph
et al. [6] and the sheet material at 950 and 850 °C by Hanning et al. 4.2.1. Time-temperature-hardness diagram
[46]. The particle sizes obtained by Hanning et al. [46] and Joseph The TTH diagram shown in Fig. 17 was drawn for the solutionized
et al. [6] were larger than those predicted by TC Prisma in this study. samples by using 350 HV as the reference hardness, which was
Haas et al. [37] investigated a Haynes 282 sheet solutionized at identified on the samples from the microhardness maps. The dis­
1120 °C for 30 min, followed by aging at 820 and 760 °C. The radii tance from the fusion zone on the hardness maps was compared
were ~12 nm for 820 °C and ~6 nm for 760 °C. In the present study, with the temperature calculations.
the radii for 865–870 °C were ~10 nm, which is lower than those In the 1.5 min sample, only one temperature point achieved 350
found by Haas et al. at 820 °C [37]. The TC Prisma simulations at HV. For the 4 h sample, the temperature map includes a slight

9
K. Vattappara, V. A Hosseini, C. Joseph et al. Journal of Alloys and Compounds 870 (2021) 159484

• Arc heat treatment successfully produced the representative

microstructure of the material at various temperatures with only
one sample for one heat treatment time.
• The arc heat-treated samples showed a graded microstructure
with various regions from the liquidus to room temperature as
follows: fusion zone, dissolution of secondary phases region,
primary carbide region, grain boundary carbide region, pre­
cipitation region, and base metal.
• The MC carbides were dissolved at high temperatures because a
longer heat treatment time decreased the upper stability tem­
perature.
Fig. 18. TTP plot showing where γ′ was observed in the SEM analysis. The blue
markers represent a heat-treatment time of 30 min, and the dark red markers • precipitation was observed below 1000 °C, where the hardness
represent a time of 4 h. The cross represents the time of 1.5 min where the γ′ of the precipitation zone increased with increasing arc heat
precipitates were present but could not be resolved using SEM. treatment time in both the aged and solutionized samples.
• The TC Prisma simulated values of the γ′ radii for 980 and 950 °C
were in good agreement with the experimentally obtained
uncertainty due to the influence of the shape of the fusion zone on values.
the temperature prediction (shown in Fig. 4). Based on the TTH • The hardness and etching response reflected the change in the
diagram, the nose of the diagram is located between 700 and 900 °C. microstructure better than SEM analyses because SEM was
This is in good agreement with Polkowska et al. [42]; however, they unable to resolve very fine particles.
used 30 samples to produce a hardness map, which is far more time
consuming compared to what has been achieved using the arc heat CRediT authorship contribution statement
treatment technique. As shown in Fig. 17, the hardness value of 350
HV was achieved after 1.5 min. A further work, however, is needed to Kevin Vattappara: Original Draft, Writing, Methodology,
map the hardness values for the time periods between the 1.5 min Investigation, Vahid Hosseini: Supervision Conceptualization,
and 30 min to improve the accuracy of the TTH diagram. It should be Validation, Writing - Review & Editing. Ceena Joseph: Supervision
noted that Polkowska et al. [42] investigated the thermal exposure of Conceptualization, Validation, Writing - Review & Editing. Fabian
8 h, however, the lower and upper boundaries of the hardness dia­ Hanning: Investigation Review & Editing. Joel Andresson:
gram were similar to those reported by Haas et al. [36] and Hanning Supervision, Writing - Review & Editing, Methodology.
et al. [45] for shorter exposure times (up to 30 min). A slightly faster
hardness increase in the arc heat treated samples compared to the Declaration of Competing Interest
results reported by Haas et al. [36] and Hanning et al. [45] was ob­
served, which could be related to the difference in the starting mi­ The authors declare that they have no known competing fi­
crostructure, heating rates, and thermal stresses induced during the nancial interests or personal relationships that could have appeared
arc heat treatment. to influence the work reported in this paper.
The TTH diagram was not drawn for the aged condition because
the high initial hardness indicates that some age hardening had al­
Acknowledgments
ready occurred before the arc heat treatment was performed. The arc
heat treatment further increased the hardness values and changed
The authors would like to thank Dr. Sten Wessman for his as­
the etching responses. Polkowska et al. [42] also reported that it is
sistance with automatic hardness testing and Mr. Kjell Hurtig for
possible to achieve higher hardness values than those produced by
performing the heat treatments.
the standard heat treatment procedure; however, it is known that
the hardness at room temperature might not provide a complete
References
picture of the properties. Therefore, further studies are needed to
investigate the possibility of achieving better properties by mod­ [1] R.C. Reed, C.M.F. Rae, 22 - Physical Metallurgy of the Nickel-Based Superalloys,
ifying the heat treatment. in: D.E. Laughlin, K. Hono (Eds.), Physical Metallurgy, Fifth edition, Elsevier,
Oxford, 2014, pp. 2215–2290.
[2] L.O. Osoba, A.M. Oladoye, V.E. Ogbonna, Corrosion evaluation of superalloys
4.2.2. Time-temperature-precipitation diagram Haynes 282 and Inconel 718 in Hydrochloric acid, J. Alloy. Compd. 804 (2019)
376–384.
The TTP diagram in Fig. 18 shows the temperatures at which the [3] L. Pike, Development of a fabricable gamma-prime (γ′) strengthened superalloy,
precipitates were observed. The temperatures were correlated in: Proceedings of The 11th International Symposium on Superalloys, 2008, pp.
with of each characterized location using the temperature model. 191–200.
[4] W. Xia, X. Zhao, L. Yue, Z. Zhang, Microstructural evolution and creep mechan­
The temperature ranges are smaller compared to the changes in the isms in Ni-based single crystal superalloys: a review, J. Alloy. Compd. 819 (2020)
hardness maps because small (<10 nm) precipitates were un­ 152954.
resolvable with the SEM but still increased the hardness values [5] R.C. Reed, The Superalloys: Fundamentals and Applications, Cambridge uni­
versity press, 2008.
considerably. Therefore, hardness mapping represents the transfor­
[6] C. Joseph, C. Persson, M.H. Colliander, Influence of heat treatment on the mi­
mation better than SEM analysis in the present study. crostructure and tensile properties of Ni-base superalloy Haynes 282, Mater. Sci.
Eng. A 679 (2017) 520–530.
[7] E. Richards, R. Cook, Factors influencing the stability of nickel-base high-tem­
5. Conclusions perature alloys, IND HEATING, 36, 1969, 1717–1718.
[8] C. Joseph, Microstructure Evolution and Mechanical Properties of Haynes 282,
Chalmers University of Technology, 2018.
The evolution of the microstructure and hardness of Haynes 282 [9] C. Sullivan, M. Donachie, Some effects of microstructure on mechanical prop­
was investigated using a novel arc heat treatment technique. The TIG erties of nickel-base superalloys, Metals Eng. Q. 7 (1967) 36.
arc was applied for 1.5 min, 30 min, and 4 h on samples with aged [10] H. Collins, R. Quigg, Carbide and intermetallic instability in advanced nickel-base
superalloys(Relative stability of gamma prime and minor phases in five nickel
and solutionized initial conditions. The experiments were com­ base superalloys, discussing effect of time, aging temperature and stress), ASM
plemented with temperature modeling, thermodynamic calcula­ Trans. Q. 61 (1968) 139–148.
tions, and precipitation simulations. [11] H. International, HAYNES® 282® alloy, in: H. International (Ed.), 2018.

10
K. Vattappara, V. A Hosseini, C. Joseph et al. Journal of Alloys and Compounds 870 (2021) 159484

[12] M. Hunkel, H. Surm, M. Steinbacher, Dilatometry, Handbook of Thermal Analysis [29] M. Fahrmann, D. Metzler, Simulation of γ′ precipitation kinetics in a commercial
and Calorimetry, Elsevier, 2018, pp. 103–129. Ni-base superalloy, JOM 68 (2016) 2786–2792.
[13] P. Archambault, D. Godard, High temperature precipitation kinetics and ttt curve [30] C. Kumara, Modelling of the temperature field in TIG arc heat treated super
of a 7xxx alloy by in-situ electrical resistivity measurements and differential duplex stainless steel samples, in, 2016, pp. 36.
calorimetry, Scr. Mater. 42 (2000) 675–680. [31] K. Kruger, HAYNES 282 alloy, Materials for Ultra-supercritical and Advanced
[14] N. Chobaut, D. Carron, J.M. Drezet, Monitoring precipitation kinetics in heat Ultra-supercritical Power Plants, Elsevier, 2017, pp. 511–545.
treatable aluminium alloys using in situ resistivity in gleeble thermomechanical [32] A. Plati, Modelling of γ′ Precipitation in Superalloys, Materials Science and
simulator, Materials Science Forum, Trans Tech Publ, 2014, pp. 921–925. Metallurgy Department, University of Cambridge, 2003, pp. 1–4.
[15] M. Fahrmann, L. Pike, Experimental TTT Diagram of HAYNES 282 Alloy, in: [33] K. Wu, Y. Wen, J.A. Hawk, Phase Field Simulations on the Precipitation Kinetics of
Proceedings of the 9th International Symposium on Superalloy 718 & Derivatives: γ′ in Ni‐Base Superalloy Haynes 282, TMS2013 Supplemental Proceedings, 2013,
Energy, Aerospace, and Industrial Applications, Springer, 2018, pp. 565–578. 711–718.
[16] V.A. Hosseini, L. Karlsson, D. Engelberg, S. Wessman, Time-temperature-pre­ [34] Q. Chen, H. Jou, G. Sterner, TC-PRISMA User’s Guide and Examples, Thermo-Calc
cipitation and property diagrams for super duplex stainless steel weld metals, Software AB, Stockholm, Sweden, 2011.
Weld. World 62 (2018) 517–533. [35] K. Vattappara, Understanding the effect of temperature and time on Gamma
[17] V.A. Hosseini, L. Karlsson, K. Hurtig, I. Choquet, D. Engelberg, M.J. Roy, C. Kumara, prime coarsening for Nickel-base superalloy Haynes 282, in, 2019.
A novel arc heat treatment technique for producing graded microstructures [36] L. Liu, F. Sommer, H. Fu, Effect of solidification conditions on MC carbides in a
through controlled temperature gradients, Mater. Des. 121 (2017) 11–23. nickel-base superalloy IN 738 LC, Scr. Metall. Mater. 30 (1994) 587–591.
[18] V.A. Hosseini, M. Thuvander, S. Wessman, L. Karlsson, Spinodal decomposition in [37] S. Haas, J. Andersson, M. Fisk, J.-S. Park, U. Lienert, Correlation of precipitate
functionally graded super duplex stainless steel and weld metal, Metall. Mater. evolution with Vickers hardness in Haynes® 282® superalloy: in-situ high-en­
Trans. A 49 (2018) 2803–2816. ergy SAXS/WAXS investigation, Mater. Sci. Eng. A 711 (2018) 250–258.
[19] V.A. Hosseini, L. Karlsson, D. Engelberg, S. Wessman, Time-temperature-pre­ [38] V. Hosseini, Super duplex stainless steels: Microstructure and propertiesof
cipitation and property diagrams for super duplex stainless steel weld metals, physically simulated base and weld metal, in, University West, 2018.
Weld. World 62 (2018) 517–533. [39] M.K. Banerjee, 2.1 Fundamentals of Heat Treating Metals and Alloys, in:
[20] V.A. Hosseini, L. Karlsson, Physical and kinetic simulation of nitrogen loss in high M.S.J. Hashmi (Ed.), Comprehensive Materials Finishing, Elsevier, Oxford, 2017,
temperature heat affected zone of duplex stainless steels, Materialia 6 (2019) pp. 1–49.
100325. [40] P. Berthod, L. Aranda, C. Vébert, S. Michon, Experimental and thermodynamic
[21] V.A. Hosseini, L. Karlsson, C. Örnek, P. Reccagni, S. Wessman, D. Engelberg, study of the high temperature microstructure of tantalum containing nickel-
Microstructure and functionality of a uniquely graded super duplex stainless based alloys, Calphad 28 (2004) 159–166.
steel designed by a novel arc heat treatment method, Mater. Charact. 139 (2018) [41] A.K. Bhargava, M.K. Banerjee, 2.14 Heat-Treating Copper and Nickel Alloys, in:
390–400. M.S.J. Hashmi (Ed.), Comprehensive Materials Finishing, Elsevier, Oxford, 2017,
[22] J.-O. Andersson, T. Helander, L. Höglund, P. Shi, B. Sundman, Thermo-Calc & pp. 398–420.
DICTRA, computational tools for materials science, Calphad 26 (2002) 273–312. [42] A. Polkowska, W. Polkowski, M. Warmuzek, N. Cieśla, G. Włoch, D. Zasada,
[23] G. Lindwall, C. Campbell, E. Lass, F. Zhang, M. Stoudt, A. Allen, L. Levine, P.M. Purgert, Microstructure and hardness evolution in haynes 282 nickel-based
Simulation of ttt curves for additively manufactured inconel 625, Metall. Mater. superalloy during multi-variant aging heat treatment, J. Mater. Eng. Perform.
Trans. A 50 (2019) 457–467. (2019).
[24] J. Andersson, G. Sjöberg, M. Chaturvedi, Hot Ductility Study of HAYNES® 282® [43] A. Brittan, J. Mahaffey, M. Anderson, The performance of Haynes 282 and its
Superalloy, in: Proceedings of the 7th International Symposium on Superalloy, 2010. weld in supercritical CO2, Mater. Sci. Eng. A 759 (2019) 770–777.
[25] L. Pike, Long Term Thermal Exposure of Haynes 282 Alloy, in: Proceedings of the [44] A. Alexandratou, S. Deligiannis, D. Liolios, P. Tsakiridis, TEM study of precipita­
7th International Symposium on Superalloy, 2010, pp. 645–660. tion sequences in haynes 282 Ni superalloy, Microsc. Microanal. 24 (2018)
[26] B.T. Alexandrov, J.M. Rodelas, M.R. Kittila, J.C. Lippold, D.A. Metzle, Continuous 2200–2201.
Heating and Cooling Transformation Diagram in Ni-base Superalloy 282, in: [45] R. Maldonado, E. Nembach, The formation of precipitate free zones and the
Sustainability of Superalloys Symposium, MS&T′11, ASM, October 2011. growth of grain boundary carbides in the nickel-base superalloy NIMONIC PE16,
[27] L. Pike, HAYNES® 282™ Alloy: A New Wrought Superalloy Designed for Improved Acta Mater. 45 (1997) 213–224.
Creep Strength and Fabricability, ASME Turbo Expo 2006: Power for Land, Sea, and [46] F. Hanning, A.K. Khan, J. Steffenburg-Nordenström, O. Ojo, J. Andersson,
Air, American Society of Mechanical Engineers, 2006, pp. 1031–1039. Investigation of the effect of short exposure in the temperature range of
[28] L. Osoba, R. Ding, O. Ojo, Improved resistance to laser weld heat-affected zone 750–950 °C on the ductility of Haynes® 282® by advanced microstructural
microfissuring in a newly developed superalloy HAYNES 282, Metall. Mater. characterization, Metals 9 (2019) 1357.
Trans. A 43 (2012) 4281–4295.

11