Heat Treatment Effects on Pristine and Cold-Worked Thin-Walled Inconel 625

Thin-walled Inconel 625 sheet metal was sectioned into tensile specimens, plastically strained, and then heat treated. Specimens were pulled to a targeted strain, unloaded, and then subjected to one of two heat treatments with the goal of restoring the full ductility and total plastic strain capability of the material. Post-heat treatment tensile testing was performed at room temperature to evaluate the heat treatment efficacy and then followed by hardness and microstructural analysis. The results showed the amount of material recovery was affected by the initial amount of plastic strain imparted to the tensile specimen before heat treatment. Although recrystallization was not observed, grains did elongate in the load direction, and the Kernel average misorientation (KAM) increased with heat treatment. Furthermore, specimens prestrained to 40% and heat treated at 980 °C successfully recovered 88% of pre-heat treatment strain capability prior to fracturing.


Introduction
Inconel 625 is a solid solution strengthened nickel-based superalloy commonly used to manufacture components requiring high temperature strength, excellent corrosion resistance, good weldability, and processability in harsh environments [1][2][3]. Inconel 625LCF is a special grade of Inconel 625 used in applications requiring improved low cycle fatigue (LCF) properties. The improved LCF properties are the result of controlled microstructures with fine grain sizes (American society for testing materials (ASTM) No. 5 or finer), which are achieved by appropriate heat treatments and tight control of impurities [4]. Because of Inconel 625's outstanding properties, it is a material of choice for numerous applications across several industries, including aeronautical, aerospace, chemical, petrochemical, marine, and nuclear [1,3,5]. During manufacturing, the material might undergo several machining, forming, and shaping processes to obtain its final shape. These processes introduce significant deformation in the part, leading to hardening, with an associated increase in strength and reduction in ductility, thus presenting properties that may significantly differ from the baseline material properties.
Metallic materials subjected to prestrain cold-work are well known to display an increase in strength and decrease in ductility due to work hardening [6][7][8][9][10], while heat treatments generally reduce strength and increase ductility [2,11,12]. Heat-treatment effects on Inconel 625 were studied under different thermo-mechanical conditions (i.e., as solution treated, cold rolled, and cold rolled and annealed) [13]. It was determined that cold rolling followed by annealing at 800 • C for 30 min resulted in a strength increase from 291 to 676 MPa and a ductility reduction from 75 to 50%. The strength increase was attributed to grain boundary and twin boundary strengthening [13]. Additionally, a study of the influence of prestrain and heat treatment on aluminum sheet mechanical properties found that deformed and heat-treated specimens exhibited higher strength when compared to undeformed and heat-treated specimens or deformed and non-heat-treated specimens [6]. Furthermore, Gao et al. showed that the yield strength and elongation of as cast Inconel 625 specimens increased from 247 MPa and 41% to 366 MPa and 60%, respectively, after being hot extruded and heat treated at 1150 • C for 1h, and to 361 MPa and 53%, respectively, after cold rolling and heat treatment to 1120 • C for 20 min [1]. Since material property variation during operation may lead to undesirable effects, Shankar et al. suggested aging Inconel 625 above the temperature it is exposed to during service [14].
The effects of heat treatment on the hardness and microstructure of Inconel 625 have also been previously studied [13][14][15][16][17][18]. The microhardness of additively manufactured Inconel 625 did not appear to change significantly with heat treatment at 980 • C with varying hold times. However, the ultimate tensile strength (UTS) showed a gradual increase with increasing hold time [2,14,15]. Furthermore, with an increase in temperature to 1050 • C for 2 h, the hardness of Inconel 625 presented softening after being annealed, which was attributed to the reduction of dislocation density coupled with the hardening process and formation of smaller grains [16]. Additionally, annealing can improve the mechanical properties by inducing static recrystallization, that is, recrystallization occurring during annealing, which is normally observed in cold-worked material subjected to an annealing treatment [19].
Three types of heat treatment for Inconel 625 were identified: (1) stress relief, which aims to relieve the stress of the material without altering the microstructure; (2) recrystallization annealing, which promotes recrystallization but no significant grain growth; and (3) solution annealing, which promotes recrystallization and grain growth [20]. Although the temperature for dynamic recrystallization, which is recrystallization during hot deformation, occurs at higher temperatures (1200 • C [21]) than heat-treatment temperatures, a reduction in work hardening rate, and thus an increase in dislocation density, with increasing temperature has been shown to occur for hot deformation [21,22]. Additionally, an increase in dislocation density has been shown to reduce the temperature required for recrystallization [21,23,24]. Therefore, for fatigue-critical applications, heat treatments have to be chosen carefully to avoid excessive grain growth, since grain boundaries provide barriers to crack propagation [25,26], and thus having fewer grain boundaries is detrimental to fatigue properties, especially in thin walls.
Heat treatment of Inconel 625LCF can potentially relieve material stresses that originate during forming operations and restore some degree of ductility to the material, allowing for additional forming processes or operational loads with a lower risk of crack initiation and propagation. This study investigated the employment of intermediate heat treatment processes after forming operations to restore the strain capability of the Inconel 625LCF material as well as the microhardness and microstructural response to deformation and heat treatment of the same material.

Specimen Cut Plan
Inconel 625LCF sheets, 0.33 mm thick, were received as formed tubes (length of 216 mm and inner diameter of 63.5 mm) that were seam welded axially to form a cylinder. The weld was subsequently planished to the thickness of the sheet metal. Specimens were electrical discharge machined (EDM) from regions unaffected by the welding and planishing processes. Note that specimen orientations were defined relative to the axial weld and did not necessarily correspond to the sheet rolling direction, as the rolling direction was unknown. Figure 1 shows the welded tubes with axial hourglass and transverse tensile specimens' locations overlaid in Figure 1a,b, respectively, and their specimen drawings in Figure 1c,d. Note that the hourglass specimen was designed with a "reference notch", shown on the upper edge of the gage section in Figure 1c, to correlate the local strain to the microhardness. Drawings of tubes with cut plans for (a) axial specimens, (b) transverse specimens, (c) hourglass specimens (dimensions in mm), and (d) tensile specimen (dimensions in mm). Axial specimens had some minor curvature and were tested as machined, and transverse test specimens required flattening prior to test.
The tensile samples were used for standard room-temperature tensile testing to investigate the effects of heat treatment and prestrain on the material properties. The hourglass specimens were used for the correlation between strain gradients and microhardness testing because of the continuous gauge radius.

Heat Treatment
Two heat treatment temperatures were used in this investigation. The temperatures were selected to balance effects beneficial for reducing the propensity for cracking with possible effects that could be deleterious to fatigue performance. The American Society for Metals (ASM) 4/4E suggested annealing treatment for Inconel 625LCF includes heating to at least 980 • C, holding for a time commensurate with section thickness, and cooling at a rate equivalent to air cool or faster [27]. Additionally, the ASM Handbooks [27,28] provided several recommendations for heat treatment of Inconel 625, including stress relief, which is intended to relieve residual stresses without significant recrystallization, and annealing, which is performed to restore partially worked material to a state in which it can undergo further deformation operations. Lastly, the Special Metals information sheet for 625LCF [4] recommended an annealing temperature of 980 • C. Two heat treatment cycles were selected to use in this investigation and are presented on Table 1. Table 1. Heat-treatment processes used for this study, informed by [27,29]. The stress relief treatment was expected to relax residual stresses while avoiding complete recrystallization, and the annealing treatment was expected to significantly restore the material ductility, though it can result in some recrystallization and undesirable grain growth. Although heat treatments at these temperatures for short periods of time (5-20 min) were not expected to strongly influence the material response [30], the amount of deformation imposed prior to heat treatment is known to affect the temperature and time required for recrystallization [27].

Tensile Test
Tensile tests were conducted at room temperature using an electromechanical Instron 2580 Series Static Load Cell and a reaction frame with a 9000 N load cell. The displacement rate was 12.7 mm/min, which corresponded to strain rates on the order of 10 −3 s −1 . Tests were run on specimens until the desired stop condition, which depended on the goal of the test. Tensile specimens were prestrained to a target strain value and then tested to failure following heat treatment. Hourglass specimens were prestrained to a target local strain value then tested either to failure or to maximum load post-heat treatment.
Tensile specimens were prestrained using standard tensile testing techniques. Testing was halted when an extensometer with a 25.4 mm gage length reached the target strain value, and the specimens were unloaded. Following prestrain, specimens were subjected to one of the two heat treatments under investigation. After heat treatment, specimens were tested either to failure or to maximum load per the ASTM E8/E8M Standard Test Methods for Tension Testing of Metallic Materials [30] and/or sectioned for hardness measurements to evaluate the specimen response to the heat treatment. Note that the extensometer was reset to a 25.4 mm gage length for the post-heat treatment tests; the strain gradient in the tensile specimens was expected to be consistent, so this gage length reset was expected to have a minimal effect on the resulting strain measurement.
Hourglass specimens were tested using similar techniques. All hourglass specimens were prestrained to a target local maximum strain of 40% as measured using digital image correlation (DIC). Because of the continuous radius of the test specimen, this maximum local strain occurred at the gage midpoint, with gradually decreasing strain as the gage section widened. Post-heat treatment, hourglass specimens were tested either to failure or to maximum load. Table 2 contains a comprehensive list of the specimens used during this investigation as well as their orientations relative to the formed tube, the sectioned shape, targeted prestrain, heat treatment temperature, and final test performed on the specimen. A total of eighteen specimens were used in this study; they were subdivided into: (1) two tensile specimens tested for baseline data; (2) two pristine tensile specimens, each subjected to one of two heat treatment effects; (3) four tensile specimens with 10% prestrain, each subjected to one of two heat treatments; (4) four tensile specimens with 40% prestrain, each subjected to one of two heat treatments; (5) one hourglass specimen tested for base line strain-microhardness correlation; (6) four hourglass specimens with 40% local prestrain, each subjected to one of two heat treatments for strain-microhardness correlation.

Sample Preparation
Metallographic specimens were mechanically sectioned, mounted in phenolic resin, and manually ground using 1200 grit wet silicon carbide (SiC) paper. To obtain a mirror-like finish, samples were polished using an automatic polisher operating at 150 RPM and 10 N of force. A 1 µm diamond lubricant suspension was used to polish the samples for 5 min, followed by a final polish using 0.05 µm colloidal silica suspension for 10 min.
The samples were etched to reveal the underlying microstructure using an electrolytic etch of waterless Kalling's etchant [31] (100 mL ethyl alcohol; 80 mL HCl; 4 g CuCl). Samples were fully submerged in the etchant, and a voltage of 3 V was applied to them for 7 s, followed by water rinse and ethyl alcohol rinse to remove water impurities. Additionally, Adler's etchant [32] (75 mL D.I. water; 150 mL HCl; 9 g copper ammonium chloride; 45 g ferric chloride) was used for pristine samples by light cotton swab followed by water rinse and ethyl alcohol rinse to remove water impurities.

Microscopy
Etched sample cross-sections were documented using bright field imaging on a Leica M205A optical stereomicroscope at 200×. Bright field lighting was used to obtain sufficient grain boundary contrast for grain size measurement. Grain measurements were performed using the grain size measurement function in the Leica Application Suite software using the linear intercept method.
As-polished cross-sections were documented using electron backscatter diffraction (EBSD). EBSD was performed on a FEI Quanta 600FEG environmental scanning electron microscope (ESEM) coupled with a NordlysMax 3 EBSD detector operating at 20 kV, a spot size of 6.0, and a scanning step size of 1.02 µm at a magnification of 500×. Grain size analysis was performed through the built-in functionality of the Oxford AZtec software.

Microhardness
Microhardness testing was performed in the sample cross-sections along the load direction. A Struers DuraScan microhardness tester was employed and equipped with a 136 • pyramid indenter. All hardness indentations used a 50 g-force load with a 10 s dwell time. The hardness point locations were centered in the middle of the cross-sectional thickness with a 0.25 mm test point spacing to prevent plasticity interference between points. Before cross-sectional mounting and polishing, the specimen grip sections were removed at the reference notches to maintain a spatial reference between the hardness data and the DIC strain data. Longitudinal cross-sections were obtained by sectioning the specimens at less than half-width, then polishing to approximately half-width to obtain a surface appropriate for microhardness measurements. Measurements of the specimens preand postpolishing provided an indication of the relative location of the cross-section.
Hourglass shaped tensile specimens included a constant radius gage section, which led to a longitudinal strain gradient through the gage section when pulled in tension. These specimens were prestrained to a target maximum localized strain matching the standard tensile specimens as measured by DIC, then subjected to the heat treatments and finally tensile tested. Localized strain distributions obtained from DIC techniques were compared to posttest microhardness measurements obtained from sections of the hourglass specimens.

DIC
For the hourglass specimens, strain was measured using the DIC technique with ARAMIS software. Imaging was performed using a GOM ARAMIS adjustable dual camera equipped with two VQXT-120M.K06 cameras with resolutions of 4096 × 3000 px. A frame rate of 1 frame per second was used throughout testing. The samples were illuminated using a GOM GmbH light projector. For DIC analysis, the facet size of 10 × 10 pixels and a step size of 7 pixels were used, resulting in a 30% overlap between facets.
Three datasets were obtained from the DIC data for analysis: (1) "prestrain" data, consisting of node locations and strains for the end of the prestrain test; (2) "baseline" data, consisting of node locations for the start of the post-heat treatment test; and (3) "failure" or "max load" data, consisting of node locations and strains for the end of the post-heat treat test. All node location data for each data set were zeroed on the reference notch on the upper left corner of the gage section using spatial offset values determined from measurements of the DIC images. These locations were used to establish scales for image measurements and to provide the initial zero point.
For each of the DIC datasets, a surface fit was generated from the node location and strain data through linear interpolation. The surface fits for the "prestrain" and "failure" or "max load" datasets were added to calculate a "total" strain field. The linearly interpolated surface fit of post-heat treatment strain and total strain values were used in this analysis. A drawback to this method is that the DIC speckle pattern had to be removed for heat treatment and reapplied for the post-heat treatment tests. Thus, the node locations for the post-heat treatment tests did not correspond to the same locations in the pre-heat treatment tests, although the interpolated strain field data were spatially aligned.

Mechanical Test
The tensile specimens were used to investigate the effects of both heat treatments by subjecting them to varying degrees of prestrain before undergoing one of the heat treatments. The specimens were subsequently tested post-heat treatment to investigate strength and ductility recovery. In the following plots and tables, the load data obtained from the load cell is displayed, rather than the stress data. This is to remove bias that results from using stress values calculated from specimen dimensions that were remeasured following the heat treatment. All specimens were machined to the same nominal specimen dimensions. Eleven test specimens were evaluated for heat treatment response as shown in Table 3. Baseline material (parent) specimens (1021-3 and 1021-10), with no heat treatment (no HT) and no prestrain, were compared against the 40% prestrained specimen 1021-5 to verify the test methods. As Figure 2a shows, prestraining affected neither the total accumulated strain nor the load carrying capability of the as-received pristine material. The strain data results were similar to those for the as-received parent material, indicating minimal effect of the extensometer gage length reset on the resulting observed strain data. Additionally, Figure 2b shows the tensile properties of the heat-treated specimens 1021-7 and 1021-8 as compared to those of the parent metal. Neither of the investigated heat treatments had a significant effect on the strength or ductility of the parent material in the absence of pre-heat treatment cold work.  Figure 3a,b shows the prestrain curves prior to heat treatment and the test to failure results after heat treatment, respectively. Significant scatter in the maximum load capacity was present for the post-heat treatment tests. Post-heat treatment, fracture elongation values for specimens with some degree of prestrain were reduced by 11% to 80% when compared to the parent material fracture elongation. All heat-treated specimens exhibited some degree of ductility recovery relative to the prestrained and non-heat-treated control specimen (1021-5). In general, specimens that were heat treated to 900 • C exhibited higher load capacity than specimens that were heat treated to 980 • C. Note that specimen 1021-4, which was prestrained to 40% and heat treated to 900 • C, showed an increase in ductility when compared to specimen 1021-5, which was prestrained to 40% and not heat treated. However, it showed a significantly lower ductility and higher strength than other specimens prestrained to 40% and heat treated to 980 • C. Table 4 shows the load at yield (load corresponding to the 0.2% offset yield strength), load at target prestrain, and prestrain for samples pre-heat treatment and their respective loads at yield, maximum loads, and strains at fracture after heat treatment. The last column shows the reduction of strain at fracture for the post-heat-treatment strain compared to the parent strain at fracture for the parent specimens (51%). Table 4. Pre-and post-heat treatment specimens' loads at yield, maximum loads, and fracture elongation percentages.

Pre-Heat Treatment
Post  2  995  1536  10  900  1307  2100  41  20  1022-4  964  1481  10  980  1198  2071  41  20  1024-6  996  1502  10  900  1274  2056  39  24  1024-8  1007  2099  40  980  740  1562  44  14  1021-4  1078  2164  40  900  1606  2020  20  61  1021-5  1069  2135  40  -2161  2159  10  81  1021-6  1083  2161  40  980  815  1623  45  12  1021-3 1089 Comparing the post-heat treatment and prestrained specimens to the parent material, specimens prestrained to 40% and heat treated to 980 • C almost fully recovered their ductility. Specimen 1021-4, heat treated to 900 • C, recovered less than 40% of the original material's ductility, and specimen 1021-5, with no HT, fractured at around 10% strain, at the same total strain as the parent material. Furthermore, the load at yield is generally lower than the load seen at the targeted prestrain for all specimens after being heat treated, indicating some softening due to heat treatment for the prestrained specimens. Furthermore, specimens heat treated to 980 • C displayed lower load at yield than specimens heat treated to 900 • C regardless of the prestrain level. This behavior was consistent with that described in literature of service-exposed and then heat-treated Inconel 625 [2,13] and aluminum [33], which implies that although an increase in strength was observed at the beginning of annealing, higher temperatures and/or longer hold times led to a decrease in strength and increase in ductility.
Additionally, the maximum load for post-heat-treatment specimens and the parent material were comparable, at approximately 2150 N, for specimens prestrained to 10% and heat treated to 900 • C as well as for the specimen that was not heat treated. However, the maximum load was lower for specimens heat treated to 980 • C for both 10% and 40% prestrain and higher for specimen 1021-4, which was prestrained to 40% and heat treated to 900 • C. Similar results were observed on additively manufactured Inconel 625; stress relieving did not significantly alter its tensile properties, while annealing resulted in higher ductility and a lower tensile strength [34]. Although contradicting results were observed on additively manufactured Inconel 718 [35], which presented an increase in strength with each subsequent heat treatment (stress relief, hot isostatic pressing, and solution treatment and aging), these results were predominantly associated with the formation of an equiaxed grain structure providing barriers for dislocations and with the presence of annealing twins with further heat treatments. Figure 4 shows an integration of the pre-and post-heat-treatment tests for prestrained specimens. The post-heat-treatment strain values are offset by the final prestrain value. Thus, the total strain values are the sum of the accumulated strain from the pre-and postheat treatment tests. Figure 4a displays the accumulated results for the 10% prestrained specimens. The total strain to failure after heat treatment was similar to the as-received parent material with no HT. This indicates that heat treatment did little to change the maximum load or ductility in the 10% prestrain condition. Figure 4b displays the accumulated results for the 40% prestrained specimens. Heat treatment at 980 • C (1024-8 and 1021-6) almost fully recovered the accumulated 40% plastic strain, indicating that appropriate heat treatment can restore the ductility of the original undeformed condition. However, specimen 1021-6 exhibited a 31% decrease in maximum load post-heat treatment; this decrease in strength and increase in elongation was in accordance with [36]. Specimen 1021-4, heat treated at 900 • C, exhibited a higher strength than the specimens heat treated to 980 • C, but it recovered only 45% of the strain when compared to samples 1024-8 and 1021-6, indicating that heat treatment at 900 • C was not as effective at recovering ductility. In general, this behavior was shown by Sukumaran et al., demonstrating that ductility increases and strength decreases with increasing heattreatment temperature [12]. Specimens prestrained to 10% exhibited minimal change in total strain capability relative to the as-received parent material, regardless of the heat treatment, which indicates that the degree of prestrain affected ductility recovery. For the specimens prestrained to 40%, the 980 • C heat treatment resulted in a greater ductility recovery than the 900 • C heat treatment. Given some amount of prestrain, the higher heat-treatment temperature resulted in lower load at yield (more yield strength recovery) than the lower-heat treatment temperature. Specimens prestrained to 40% and heat treated to 980 • C exhibited postheat-treatment tensile properties in line with those of the original baseline material. These specimens exhibited a total ductility nearly twice that of the baseline material with similar or reduced strength.
Therefore, the heat treatments appeared to have minimal impact on the tensile properties without some degree of prestrain. Specimens with 10% prestrain did not exhibit improved total ductility, while specimens with 40% prestrain exhibited improved ductility; of the latter, specimens heat treated to 980 • C exhibited the most additional ductility. Thus, it can be concluded that ductility recovery was most successful with specimens prestrained to 40% and heat treated to 980 • C. Similar results were demonstrated in cold-rolled and annealed Inconel 625 by Wang et al. [13] and in hot/cold deformation annealed Inconel 625 by Gao et al. [1], the latter of which also demonstrated a progressive strength decrease and ductility increase on specimens cold rolled and annealed at temperatures higher than 800 • C, which was attributed to grain refinement and the formation of annealing twins to impede dislocations.

Microscopy
Microscopy was used to analyze the grain structure of deformed and undeformed specimens. Optical microscope was used to investigate etched specimens, and scanning electron microscopy (SEM) was used for higher magnification and crystallographic analysis.

Optical
Two areas were investigated on the specimens based on the amount of strain each region was expected to experience during tensile testing. Figure 5 shows an image of an hourglass specimen with the regions of interest highlighted. The first region, 1, is located on the grip section, which was expected to experience little to no strain, and the second region, 2, was located on the gage section of the sample, where most deformation was expected. By comparing the deformation at both locations, it was possible to observe the effects of heat treatment on deformed and undeformed microstructure. Grain length, along the load direction, was measured by sectioning, mounting, polishing, and etching each specimen. The length was then measured by drawing a straight line along numerous grains in the load direction to increase statistical significance. Because of the highly deformed nature of the grains in region 2, along with the prevalence of annealing twins, determining grain size through traditional methods was impractical. Therefore, grain length measurements provided a convenient method to compare grain elongation within the bounds of this study. Figure 5 shows the etched microstructure on the undeformed region (Figure 5a-c) and deformed region (Figure 5d-f). Note that the grains in the deformed region were severely deformed and elongated along the load direction. Also, note that the grain was difficult to define at lower magnification because of the deformation and twinning.
The undeformed region, 1, of all specimens exhibited similarly sized and equiaxed grains, while the deformed region, 2, presented elongated grains along the load direction due to the mechanical work imparted by the tensile test. Note that neither heat treatment generated a fully restored, equiaxed grain structure in the deformed region. However, annealing twins were present throughout all regions of the specimen. Figure 6 shows the undeformed region, 1, and deformed region, 2, of specimen 1041-3, which experienced no prestrain and was tested to failure. Specimen 1041-3/1 was examined before heat treatment, while specimen 1041-3/2 was examined after heat treatment at 980 • C. There was little difference between the undeformed microstructures in each specimen. In the deformed region, the grain lengths for specimen 1041-3/1 were difficult to distinguish because of the amount of twinning, while specimen 1041-3/1 presented a higher grain definition and generally larger grains, indicating some amount of grain growth during heat treatment. Note that in sample 1041-3/1, the grain length determination in the deformed region was convoluted because of the excessive deformation imposed by the mechanical strain during failure. Thus, the values reported are subject to a greater uncertainty.  Figure 7 shows the statistics for the grain elongation measurements for (a) tensile specimens prestrained to 10% and 40% and (b) hourglass specimens prestrained to 40% and failure. Note that all specimens showed longer grains in the deformed region when compared to the undeformed region independent of heat treatment, but no evidence of recrystallization was observed. The only exception was specimen 1041-3/1, which, as mentioned, had significant deformation coupled with twinning that made grain boundary identification difficult and thus decreased the confidence of the results. Specimens subjected to heat treatment exhibited longer grains in the gage area of the sample, where the most deformation occurred, and grip area, where deformation severity was reduced, indicating some degree of grain growth during the investigated heat-treatment cycles. Table 5 lists the grain length for undeformed and deformed specimens measured through optical microscopy. Severe grain deformation was observed on the gage section of prestrained specimens. Specimen 1041-3/1 showed a decrease in grain length, which was attributed to the difficulty of performing measurements due to its high deformation. Post-heat treatment, grains located in the deformed region were longer than grains in the undeformed region in specimen 1041-3/2.

SEM/EBSD
Further grain analysis was performed through EBSD, in which the sample crosssection was scanned and mapped for grain size and texture changes due to heat treatment. Since the specimens were thin walled, two locations were chosen for each specimen to examine a higher number of grains and gain higher statistical significance. One specimen for each heat treatment condition was selected and had the undeformed region scanned parallel to the load direction. Table 6 shows the statistics for each specimen. Grain area average and standard deviation increased as the heat treatment temperature increased, indicating that microstructure changes (grain growth) occurred during each treatment. However, the observed increase was minor and did not represent a significant change in tensile strength of elongation. Figure 8a-c shows the texturing for the specimens with no HT, 900 • C, and 980 • C, respectively, as well as their respective inverse pole figure (IPF) maps. Figure 8d-f shows the corrected grain structure maps for the specimens with no HT, 900 • C, and 980 • C, respectively, with twins and small angles transitions removed from consideration for a more accurate grain size measurement. Grain analysis was conducted with a 6 • misorientation angle and was set to disregard special boundaries on axis (111) with an angle of 60 • and a deviation of 5 • on Aztec software; this was done to bypass twins and more accurately obtain the overall grain size. The crystallographic texture was quantified through the IPF multiples of uniform distribution (MUD), measuring the texture orientation density, with a random texture being characterized by having an MUD of 1 [37]. Although the texture appeared to change from [101] to [111] with increasing heat treatment temperature, the overall texture was relatively weak, with a 1.92 MUD for the no HT specimen that decreased to to 1.64 and 1.56 MUDs for the 900 • C and 980 • C specimens, respectively.
Through the information obtained from the EBSD scans, it was possible to measure the Kernel average misorientation (KAM), shown in Figure 9, which measures the average misorientation angle compared to neighbors [38]. KAM maps with a 2.5 • threshold showed an increase in the misorientation between the grains as the heat treatment temperature increased. Since these analyses were performed on the undeformed region, this confirmed that heat treatment caused changes to the microstructure. There was a visible increase in KAM as the heat treatment temperature increased, which is a strong indicator of grain growth being promoted by temperature. Additionally, the higher KAM values were observed mainly on grain boundaries, agreeing with previous studies performed on dynamic recrystallization [21,39] that stated that dynamic recrystallization starts predominantly at grain boundaries.
Although grains seemed to increase in size with increasing in heat treatment temperature, they did not vary significantly, and the standard deviation showed an increase with temperature. This is consistent with literature, which has reported not having significant change with either of the proposed heat treatments on undeformed locations [5,40], observing little to no change in grain structure at temperatures below 1050 • C.

Strain and Hardness
The hourglass specimen geometry facilitated the development of a strain gradient, representing a range of local strain up to the maximum, which allowed for investigations of the heat treatments' effects on the material with a varying strain history. Hardness measurements obtained along the longitudinal cross-section of the specimen were then correlated to the localized strain experienced by the material in the region of the hardness indentation.

Strain Gradient
Strain traces were calculated from the DIC data to facilitate comparisons between the strain gradients and midthickness hardness measurements. Five hourglass test specimens, listed in Table 7, were evaluated to determine how the investigated heat treatments affected localized strain and material hardness. The procedure for prestraining the hourglass specimens was established by performing a preliminary test using a 25.4 mm extensometer on the gage section and straining it to approximately 40% local strain. Following prestraining, specimens were heat treated and retested using the DIC system, in which one specimen from each heat treatment was tested to failure or stopped at maximum load. Figure 10a,b shows the load and displacement data from the prestrain tests and post-heat treatment results of the tensile test for hourglass specimens, respectively. Both stress and strain values were inconsistent throughout the test specimen gage because of the hourglass shape; as a result, load and displacement provided the most convenient method of comparison between tests. All specimens were machined to the same nominal dimensions. Figure 10 shows the calculated strain data for a prestrained and a heat-treated specimen, highlighting the strain gradient developed along the specimen gage section as measured by DIC. Similarly to tensile specimens, hourglass specimens heat treated at 900 • C exhibited limited ductility recovery and comparable strength to the parent material, while hourglass specimens heat treated at 980 • C exhibited lower strength and lower ductility relative to the parent material.

Strain-to-Hardness
Following tensile testing, hourglass specimens were sectioned through the longitudinal centerline to obtain microhardness measurements along the exposed cross-section. Microhardness traces, shown in black in Figure 11, indicate the locations where hardness measurements were obtained. Using the reference notches and the indentation spacing, microhardness values were correlated to the interpolated local strain field values obtained from DIC data. The plots depicted in Figure 11 show the microhardness, prestrain, postheat-treatment failure strain, and total strain values along the data trace (the total strain values are the sum of the prestrain and failure strain). From the plots, it is evident that microhardness values increased as the local strain increased. The plots depicted in Figure 12 show comparisons between localized strain values obtained from DIC and microhardness measurements. All specimens exhibited similar maximum hardness values, regardless of test procedure or heat treatment.  Figures (a,b) show specimens that that underwent heat treatment at 900 • C, and Figures (c,d) show specimens that underwent heat treatment at 980 • C.
Some increased hardness would have been expected in specimens heat treated to 900 • C because of the higher strength displayed. However, the maximum hardness values were associated with higher localized total strain values in the specimens heat treated to 980 • C. Figure 13 shows the relationship between total localized strain (including prestrain) and microhardness. For specimens heat treated at 900 • C, the total strain correlation curves were consistent with those of the parent material, which is consistent with the tensile data, indicating that the 900 • C treatment had limited effect on the material hardness. Specimens heat treated at 980 • C exhibited a similar hardness at low local strain values as the parent material. However, above 15% local strain, these specimens exhibited reduced hardness at a given local strain value when compared to the parent material. Results from the 980 • C specimens indicate that it is possible that heat treatments increased the amount of total strain required to obtain the maximum hardness observed in the materials. Thus, the strain hardening imparted into Inconel 625 by forming operations may be recovered. Softening behavior was shown for Inconel 625 after being annealed at 1050 • C for 2 h [16], while at 980 • C, additively manufactured Inconel 625 did not present significant change in hardness [15]. The hourglass specimens heat treated to 980 • C were more ductile than those heat treated to 900 • C and were able to sustain higher levels of localized strain. However, when accounting for the prestrain test, specimens with any heat treatment exhibited greater total ductility than the as-received parent material. Figure 13. Total strain and micro hardness correlation for prestrain summed with post-heat-treatment strain. Figures (a,b) show the effect of 900 • C heat treatment on micro hardness and (c,d) show the effect of 980 • C heat treatment on micro hardness.
While the specimen gage section geometry developed the strain gradient during the prestrain tests, the heat treatment appeared to influence the gradients observed in the post-heat treatment tests. Post-heat treatment strain gradients differed between the two heat treatment temperatures, indicating that the material heat treatment response was affected by the degree of initial prestrain.
All specimens exhibited similar maximum hardness values, accounting for some increased hardness near the fracture surface for the specimens that were tested to failure. Hardness values for specimens heat treated to 900 • C were consistent with hardness measurements obtained from the parent material, and the correlations between hardness and total strain were also in line with those of the as-received parent material, indicating minimal effect of the 900 • C treatment on final properties. Specimens heat treated to 980 • C displayed a higher hardness at low local strains and lower hardness at high local strains relative to the parent material, indicating that heat treatment to 980 • C may lead to an increased capacity to sustain higher total strains prior to reaching critical hardness values.

Conclusions
An investigation into the effect of heat treatment on the room-temperature tensile, microstructural, and microhardness properties of Inconel 625LCF was conducted to study influences on ductility recovery and strength reversibility. This study was performed on as-received, parent specimens, prestrained specimens, heat-treated specimens, and a combination of these to understand the influence of heat treatment on the specimens' response. The results indicated that heat treatment at 980 • C successfully recovered much of the material's room temperature strain capacity and strength properties. Furthermore, the material response to the heat treatment was found to vary depending on the amount of prestrain imparted to the specimen prior to heat treatment.

•
The results from the room-temperature tensile tests showed that appropriate heat treatments cold restore the ductility of prestrained material. Ductility recovery was most successful with specimens prestrained to 40% and heat treated to 980 • C.

•
The material response to the heat treatments was affected by the degree of mechanical work imparted by the prestrain procedure. Specimens with 10% prestrain did not exhibit improved ductility, regardless of the heat treatment temperature. However, specimens with 40% prestrain exhibited improved ductility. Of these, specimens heat treated to 980 • C exhibited the most ductility recovery. • Metallographic evaluations indicated that heat treatments did not lead to full recrystallization and that the combination of the prestrain and heat treatment may lead to increased grain size. Additionally, post-heat-treatment mechanical strain appeared to affect the resulting grain size to a greater extent than the heat treatment. • EBSD results showed a slight increase in grain size and increase in KAM with increasing heat treatment temperature. The KAM increase confirmed that misorientation increased with heat treatment. Additionally, texturing changed from a weak [101] to a weak [111] with increasing heat treatment.

•
Specimens heat treated to 900 • C exhibited microhardness values similar to the parent material. The correlations between hardness and total strain were similar to those for the as-received parent material, indicating minimal, if any, effect of the 900 • C treatment on final properties. • Specimens heat treated to 980 • C exhibited microhardness values that were higher at low local strains and lower at high local strains compared to the as-received parent material, indicating that the 980 • C treatment resulted in an increased capacity to sustain higher total strains prior to reaching critical hardness values.
Overall, the study showed that intermediate heat treatments could recover the ductility in the material and thus reduce the propensity for crack initiation. However, the degree of applied strain prior to heat treatment had a significant impact on the resulting properties. More refinement in heat treatment parameters is warranted, and additional metallographic work is needed to fully understand the interaction between the prestrain procedure and the heat treatments. Data Availability Statement: Data will not be publically available. But can be shared upon request.