Tensile Performance Sensitivity to Variations of Standard 17-4 PH Heat Treatments on LPBF-Produced Material

Abstract

Standard heat treatments for metals of a particular composition are typically designed with the assumption of a conventional starting microstructure, such as that produced by casting or wrought processing. When applied to metals fabricated by Laser Powder Bed Fusion (LPBF) metal additive manufacturing (AM), these heat treatments can produce inconsistent performance due to the unique as-built microstructures. This study investigates how modifications to standard heat treatments for 17-4 PH steel influence the microstructure and mechanical properties of LPBF-fabricated material. Specimens were produced and subjected to varying solutionizing and homogenizing treatments followed by standard aging treatments. Microstructures were characterized using optical microscopy, Electron Backscatter Diffraction (EBSD), and X-ray diffraction (XRD), and mechanical properties were evaluated through uniaxial tensile testing. Based on these results, recommendations are provided for achieving improved wrought-like performance in LPBF 17-4 PH steel.

1. Introduction

When processed by Laser Powder Bed Fusion (LPBF), materials have a unique as-built microstructure characteristic of the process resulting from the rapid solidification and high cooling rates on the order of ${10}^6$ °C ${\mathrm{s}}^{-1}$ [1]. These conditions can lead to small grain size [2], high dislocation density [3], high amounts of residual stress [4], and retention of metastable phases [5]. As such, when heat treatments designed around the conventional starting microstructures of a wrought (e.g., hot rolling and drawing) or cast material are applied to LPBF-produced material, the results can lead to varying microstructures and a mechanical performance that does not achieve the desired results. An understanding of the augmentation of typical heat treatment schedules for LPBF materials for predictable performance is necessary, especially for those that rely on thermal processing for their properties such as 17-4 PH.

17-4 PH steel is a precipitation hardening (PH) martensitic steel used in a variety of applications in the aerospace, oil and gas, and medical industries due to its high strength and corrosion resistance. This versatility in application is in part due to the range of standard and well-characterized heat treatment states available, enabling the right balance of strength and toughness to be achieved for the end application. These standardized treatments, as dictated through industry standards such as ASTM A564 [6], follow the same sequence: a solution annealing treatment followed by a low temperature aging treatment. Achieving the solution-annealed Condition A per ASTM A564 is a treatment at approximately 1040 °C for one hour. The aging treatments precipitate a Cu-rich phase for additional strengthening with standard ranges in time and temperature to achieve the desired performance. The aging treatments follow a naming designation describing the temperature used in degrees Fahrenheit, such as the H900 treatment, which is a 900 °F (480 °C) treatment for one hour. These treatments are predicated on the constancy of the Condition A microstructure, deviation from which leads to variation in predictable performance.

It is well documented that the as-built microstructure of LPBF 17-4 PH differs from what is seen in typical wrought 17-4 PH. High cooling rates leading to increased amounts of retained $\delta$-ferrite [7] are exacerbated by variation in composition [8], resulting in an as-built structure with a mix of $\delta$-ferrite, austenite, and martensite in varying proportions [9,10,11,12,13]. With further increase in austenite stabilizing elements such as nitrogen, a primarily austenitic microstructure has been observed with drastically varying performance from a martensitic 17-4 PH [14,15]. The elevated nitrogen is mainly a result of nitrogen atomizing gas used in the powder manufacturing that is retained though the LPBF fabrication process [16,17]. Observed in situ absorption from the LPBF process has also been demonstrated, but at insufficient amounts to result in austenite stabilization [18].

With the possibility of these varying starting conditions, several studies have investigated the heat treatment behavior of LPBF 17-4 PH. Cheruvathur et al. [19] showed that a 1150 °C homogenization treatment breaks up the dendritic solidification and homogenizes microsegregation, resulting in a primary martensitic structure from an initial mixed phase structure. Li et al. [20] also used 1150 °C treatments and demonstrated an optimal time of one hour for the best performance after evaluating a range of durations. Alnajjar et al. [21] demonstrated that a martensitic structure was achievable and all $\delta$-ferrite could be reverted by reaustenizing material at 1100 °C. Considering copper segregation, An et al. [22] determined that for proper precipitation hardening, a homogenization of at least 1100 °C is needed. Sun et al. [23] observed, however, that when a conventional 1040 °C solution annealing treatment is used on a mixed $\delta$-ferrite and martensite structure, the Cu-rich precipitates are still observed after a subsequent H900 aging treatment. Some studies have also considered the effects of forgoing the typical solutionizing treatment and directly aging the as-built material resulting in reduced strength but improved ductility due to the elevated amount of retained austenite and $\delta$-ferrite [24,25].

In this work, the effects of varying the initial treatment needed to achieve Condition A before undergoing a range of standard aging treatments is explored. The current state of the literature indicates that a higher temperature homogenization treatment above 1100 °C, rather than the traditional solution annealing treatment at 1040 °C, is necessary for LPBF-produced 17-4 PH and will be explored as part of this study, while previous studies have evaluated the effects of the progression from solutionizing to homogenizing treatments, subsequent aging has been focused on the peak age condition of H900. For this current work in addition to the H900 (480 °C for 1 h) condition, H1025 (550 °C for 4 h) and H1150 (620 °C for 4 h) over-age conditions are evaluated. For this study, when aging treatments are named the aforementioned ASTM A564 naming convention will be used. Microstructure by optical microscopy and EBSD, phase composition by XRD, and mechanical performance as determined by uniaxial tensile testing will all be presented for a better understanding of how best to design heat treatment schedules specifically for the unique microstructures seen in LPBF 17-4 PH.

2. Materials and Methods

Samples for this study were produced with a Renishaw (Wotton-under-Edge, UK) AM250 LPBF system under high purity argon atmosphere. The AM250 uses a 200 W Continuous Wave (CW) Yb:Fiber laser ($\lambda$ = 1070 nm) which operates in a power-modulated mode where laser power, point distance, exposure time, and hatch spacing are controlled as primary laser control parameters. A summary of the build parameters used for fabricating material in this study is found in

Table 1. LPBF build parameters used on the AM250 for the fabrication of samples.

Laser Power (W)	Hatch Spacing (μm)	Point Distance (μm)	Exposure Time (μs)	Layer Thickness (μm)	Layer Rotation (Deg.)
200	90	55	85	50	67

. Border scans were not applied as the surfaces of all samples were intended to be machined. Argon-atomized 17-4 PH powder of 15–45 $\mathsf{\mu }$m particle size distribution with its composition summarized in

Table 2. Composition of powders used in experiments compared to the ASTM industry standard for 17-4 PH. ${\mathrm{Cr}}_{\mathrm{Eq}}$ and ${\mathrm{Ni}}_{\mathrm{Eq}}$ calculated per WRC-1992 equations. Equations (1) and (2), respectively.

Element	ASTM A564 (wt.%)	Argon Atomized Powder (wt.%)
Cr	15.00–17.00	16.25
Ni	3.00–5.00	4.336
Cu	3.00–5.00	4.21
Mn	1.0 Max.	0.1968
Si	1.0 Max.	0.39
Nb	0.15–0.45	0.3
C	0.07 Max.	0.0171
P	0.04 Max.	0.0117
O	-	0.0422
N	-	0.0219
Co	-	0.0024
Mo	-	0.0068
V	-	0.05
W	-	0.001
Al	-	0.002
Fe	Bal.	Bal.
${\mathrm{Cr}}_{\mathrm{Eq}}$	-	16.47
${\mathrm{Ni}}_{\mathrm{Eq}}$	-	6.43
${\mathrm{Cr}}_{\mathrm{Eq}}$/${\mathrm{Ni}}_{\mathrm{Eq}}$	-	2.56

was used for sample builds. Inductively coupled plasma optical emission spectroscopy (ICP-OES), inert gas fusion, and combustion analysis were used for all composition measurements. ICP-OES was used for all elements except for oxygen, nitrogen, and hydrogen which were analyzed by inert gas fusion, and combustion analysis was used for measurement of carbon and sulfur. The typical 17-4 PH composition as defined by ASTM A564 is also listed in Table 2 for reference. Cr and Ni equivalency was calculated per the WRC-1992 equations where element wt. % is used and is presented as Equations (1) and (2), respectively.

$${\mathrm{Cr}}_{\mathrm{Eq}}=\mathrm{Cr}+\mathrm{Mo}+(0.7\times \mathrm{Nb})$$

(1)

$${\mathrm{Ni}}_{\mathrm{Eq}}=\mathrm{Ni}+(35\times \mathrm{C})+(20\times \mathrm{N})+(0.25\times \mathrm{Cu})$$

(2)

For the uniaxial tensile samples, vertical cylinders of 10 mm diameter were oriented in an array with consistent 20 mm spacing. Samples were removed from the build plate by wire EDM. Heat treatments were performed per ASTM A564 on the cylindrical samples in vacuum furnaces with control thermocouples. Furnaces were heated to the scheduled temperature and held to +/−5.6 °C. Samples were held to the scheduled time +10/−0 min, followed by a gas cool to room temperature. The schedules for the various treatments used are summarized in

Table 3. Aging treatment schedules used for samples per ASTM A564 [6] with solutionizing and homogenizing treatment details.

Treatment Type	Treatment Name	Time (h)	Temperature (°C)
Solutionize	930 °C	1	930
Solutionize	1040 °C	1	1040
Homogenize	1150 °C	1	1150
Age	H900	1	480
Age	H1025	4	550
Age	H1150	4	620

. Samples were machined to final geometry after heat treatment per ASTM E8 [26] for round subsize specimen 4 as detailed in Figure 1. Machined tensile samples were tested at a constant crosshead speed of 0.5 mm/min until 2.0% strain, where the speed was then increased to 5 mm/min through failure with continuous loading.

Characterization of as-built material was performed on samples that were taken from the fabricated cylinders after their removal from the build plate and before heat treatment. For material in a heat-treated state, samples were collected after the completion of the various heat treatment cycles. Optical microscopy samples were mounted, ground, polished, and etched with Fry’s Reagent per ASTM E407 [27]. Electron Backscatter Diffraction (EBSD) was collected from a sample prepared in the same manner as those for optical microscopy, with the etching step replaced by an additional vibratory polishing operation. An EDAX velocity detector on a Zeiss (Oberkochen, Germany) Gemini SEM was used for data collection at 1 $\mathsf{\mu }$m step size and was processed using Oxford Instruments (Abingdon-on-Thames, UK) AztecCrystal software. X-ray diffraction (XRD) measurements were taken using a Bruker (Billerica, USA) D8 Advance diffractometer, equipped with a sealed tube and Cu X-ray source ($\lambda$ = 1.54060 Å), incident beam parabolic mirror and 1 mm pinhole collimator optics, an Eulerian cradle, and a Bruker Eiger2 500K detector. Then, 2D coupled $\theta$-2$\theta$ scans were collected on each sample and subsequently integrated into 1D data for use in phase analysis. Using the Eulerian cradle, the samples were rocked 0–40 degrees in $\chi$ and rotated completely in $\varphi$ during data collection to minimize the effect of texture on observed relative intensities. Measurements were collected in 0.05 degree increments with 1 s exposures at each step. Pole figure data were collected on the same diffractometer using 2D side-inclination with 5 degree steps in $\chi$ and optimized $\varphi$ coverage.

3. Results and Discussion

3.1. Microstructure by Optical Microscopy and EBSD

Optical microscopy and EBSD were performed on material in the as-built condition, and samples were treated at two solutionizing and one homogenizing temperatures to assess the microstructural effects of these treatments. Across all samples small amounts of porosity can be observed. Samples were cross-sectioned and analyzed in the plane perpendicular to the build direction for optical microscopy and parallel to the build direction for EBSD.

Optical microscopy for the as-built microstructure is found in Figure 2A and Figure 3A. Here, the typical LPBF scan path features can be easily observed with a mixture of large and small grains. A mixture of the $\delta$-ferrite and martensite structures characteristic of this material can also be observed. As presented by An et al. [28] with the calculated ${\mathrm{Cr}}_{\mathrm{Eq}}$/${\mathrm{Ni}}_{\mathrm{Eq}}$ ratio seen in this work of 2.56, a near equal mixture of $\delta$-ferrite and martensite is to be predicted. From etched micrographs using Fry’s Reagent, the lighter, large grains indicate $\delta$-ferrite, where the dark regions with small grains indicate the presence of martensite. Figure 4A presents the EBSD IPF map for the same as-built condition. Here, the imaging plane is perpendicular to that of the optical microscopy, but the same typical as-built scan path features can still be observed. As with the optical microscopy, a mix of large grains at the core of the scan tracks with smaller grains concentrated on the edges of the tracks are seen. The microstructure for the as-built condition samples used in this study are consistent with what has previously been presented in the literature for similar LPBF 17-4 PH heat treatment studies [21,29,30].

The typical treatment for wrought 17-4 PH to achieve Condition A is at 1040 °C. The first solution treatment evaluated is a reduced temperature treatment at 930 °C. Previous work has shown experimentally for the LPBF as-built microstructure containing a significant fraction of $\delta$-ferrite, reversion to austenite begins to occur at approximately 800 °C [21,31,32] and as such some amount of solutionizing could be expected at this reduced temperature treatment. Figure 2B and Figure 3B show the optical microscopy results of the 930 °C solutioning treatment. At the lower magnification presented in Figure 2B, some remnants of the as-built scan path structure can still be seen. At the higher magnification image in Figure 3B, a lathy structure is present that has not been previously observed in LPBF 17-4 PH material. Analysis of the EBSD IPF maps in Figure 4B reveals a structure similar to that of the as-built material with the presence of additional fine grains located on the existing as-built structure grain boundaries. The retention of the as-built structure for this heat treat condition indicates that at 930 °C full austenite reversion does not occur.

Figure 2C and Figure 3C show the effects of the typical 1040 °C treatment. With this treatment there are no longer obvious remnants of the as-built structure, and a common martensitic structure can be identified. This microstructure is similar to what has been observed in previous studies for a 1040 °C treatment of LPBF material and is analogous to a typical wrought Condition A. EBSD IPF maps in Figure 4C reveal a fine lath structure characteristic of 17-4 PH and also characteristic of what has been previously presented in the literature. The microstructural effects of the elevated temperature homogenizing treatment at 1150 °C are presented in Figure 2D and Figure 3D. Similar to the 1040 °C treatment, the results here show no remaining as-built structure and a typical martensitic structure. EBSD IPF maps in Figure 4D reveal a coarser lath structure compared to what is seen at the 1040 °C treatment but overall a finer grain size. As with the results of the 1040 °C, the 1150 °C samples reflect a typical Condition A microstructure. The microstructure of both the 1040 °C and 1150 °C treatments is similar and indicates that a full austenite reversion occurred, allowing for a complete martensite transformation upon cooling to room temperature.

3.2. XRD Analysis

For further quantification of the effects of the two solutionizing and one homogenizing treatments, XRD was performed on the samples. Measurements were taken perpendicular to the build direction and data were analyzed by the Rietveld method. The diffractograms, pole figures, and data summary are found in Figure 5 and Figure 6, and

Table 4. Phase composition and lattice parameters for the various solution- and homogenization-treated conditions as determined by XRD. As-built, 930 °C for 1 h, 1040 °C for 1 h, and 1150 °C for 1 h.

	BCC		FCC
	wt. %	a (Å)	wt. %	a (Å)
As-Built	99.5	2.87750	0.5	3.602
930 °C	98.9	2.87673	1.1	3.6180
1040 °C	100	2.87769	0	-
1150 °C	100	2.87778	0	-

, respectively. The distinction between the two BCC $\delta$-ferrite and martensite structures cannot be made with XRD and is denoted as the singular BCC in the presented results. All samples exhibited a primary BCC crystal structure with small amounts of FCC austenite phase observed in the as-built and 930 °C treatment samples at 0.5% and 1.1%, respectively. FCC austenite was not measured in the 1040 °C and 1150 °C treatment samples. This further indicates as we discussed in Section 3.1 that at the 1040 °C and 1150 °C treatments complete austenite reversion occurs allowing for a complete martensitic transformation to occur upon cooling to room temperature. At the 930 °C treatment, a structure that is primarily BCC with 1.1% retained austenite is more similar to the as-built state with its measured 0.5% retained austenite.

Further analysis of the diffraction data indicates the presence of additional crystallographic features. Peak broadening was observed in all samples indicating small crystallite size, microstrain, and/or inhomogeneous chemistry. Refined microstructure values suggest small crystallite sizes on the order of 100 nm for the as-built condition and microstrain values of approximately 0.001 for samples in the two solutionized and one homogenized conditions resulting in the observed peak broadening. Systematic deviations from expected relative intensities indicate the presence of crystallographic texture. As seen in the collected pole figures found in Figure 6A–D, a (2 0 0) fiber-like texture parallel to the build direction was also observed in all samples. This texture observed across all heat treatments was also observed by Clausen et al. [14]. All samples exhibited compressive stress in the plane perpendicular to the build direction. The average stress from two perpendicular measurements was calculated to be −264 MPa for the as-built condition, −351 MPa for the 930 °C treatment, −386 MPa for the 1040 °C treatment, and −376 MPa for the 1150 °C treatment. Negligible shear stresses were calculated.

3.3. Mechanical Properties and Fracture Surfaces

Uniaxial tensile performance was assessed on samples treated with the two solutionizing and one homogenizing temperatures as well as the three aging treatments. SEM fractography was performed on the fracture surface as well. Stress–strain curves for samples with the 930 °C solution treatment are found in Figure 7, 1040 °C samples are found in Figure 8, and 1150 °C samples are found in Figure 9. Corresponding fracture surfaces for the 930 °C solution treatment are found in Figure 10, 1040 °C samples in Figure 11, and 1150 °C samples in Figure 12. A summary table of all mechanical results is found in

Table 5. Summary of mechanical performance with 1 standard deviation of the sample population noted. Each data entry represents three tested samples.

Solution/ Homogenizing Treatment	Aging Treatment	Elastic Modulus (GPa)	0.2% YS (MPa)	UTS (MPa)	Strain at Failure (%)
930 °C	No Age	182 ± 8	696 ± 22	896 ± 12	17 ± 0.6
930 °C	H900	197 ± 11	1069 ± 7	1165 ± 8	11 ± 3.5
930 °C	H1025	187 ± 6	951 ± 12	979 ± 14	13 ± 5.0
930 °C	H1150	179 ± 6	676 ± 7	827 ± 10	18 ± 4.5
1040 °C	No Age	189 ± 7	738 ± 14	986 ± 7	9.3 ± 3.2
1040 °C	H900	195 ± 3	1255 ± 4	1392 ± 7	9.0 ± 4.4
1040 °C	H1025	197 ± 7	1117 ± 14	1172 ± 14	9.3 ± 3.2
1040 °C	H1150	192 ± 7	841 ± 4	938 ± 7	15 ± 3.8
1150 °C	No Age	189 ± 8	786 ± 14	1027 ± 7	9.3 ± 3.8
1150 °C	H900	207 ± 12	1330 ± 12	1496 ± 12	5.0 ± 2.0
1150 °C	H1025	209 ± 13	1172 ± 24	1254 ± 12	8.0 ± 4.6
1150 °C	H1150	190 ± 6	876 ± 14	972 ± 10	13 ± 4.1

. A summary table of mechanical minimums per ASTM A564 is found in

Table 6. Summary of mechanical performance requirements in aged condition per ASTM A564 [6]. Condition A is equivalent to the 1040 °C treatment.

Solution Anneal Treatment	Aging Treatment	Elastic Modulus (GPa)	0.2% YS (MPa)	UTS (MPa)	Strain at Failure (%)
Condition A	H900	-	1170	1310	10
Condition A	H1025	-	1000	1070	12
Condition A	H1150	-	725	930	16

.

Across all three solutionizing and one homogenizing treatments, mechanical performance follows the trend typical of wrought material where strength increases from the no-age condition with the H900 treatment being the highest strength, and reducing strength in the over-age H1025 and H1150 treatments. In all cases the H1150 strength is below what is seen in the no-age condition. All samples also exhibited spread in the strain at failure attributable to the porosity observed in the microscopy and EBSD analysis presented in Figure 2, Figure 3 and Figure 4. Samples also exhibited some spread in modulus as is typical in LPBF material. Although not utilized in this study, the loading cycle procedure allowed for in ASTM E111 [33] should be applied for more accurate results. All fracture surfaces exhibit ductile failure with the surface containing dimples across all treatment conditions. At the H900 peak-age treatment, cracks consistent with intergranular cleavage are observed for all starting conditions.

In evaluating the performance of the mechanical samples, a trend of comparable and lower strain at failure for the 1040 °C and 1150 °C treatments in comparison to those at 930 °C is also observed and can be correlated with the presence of retained austenite as indicated by XRD. These results are in line with those of studies that have evaluated the direct aging of LPBF 17-4 PH as opposed to the sequence of a solutionizing/homogenizing treatment followed by an aging treatment like those of LeBrun et al. [24] and Eskandari et al. [25]. Here, there is a lower overall strength and higher ductility was presented for all treatments.

A summary of mechanical performance in heat-treated LPBF 17-4 PH is presented by Huang et al. [34]. In comparison to the results presented in this study, the application of a 1040 °C solutionizing treatment followed by aging provided material with performance similar to what is seen and presented in the literature as well as in typical wrought processing as seen in Table 6. An elevated 1150 °C homogenization treatment results in higher strengths with some reduction in ductility. The reduced temperature 930 °C solutionizing treatment and subsequent aging yields material that performs between the direct age and the 1040 °C treatments, with a balance of strength and ductility.

4. Conclusions

In this study, the effects of varying the solutionizing and homogenizing temperatures on microstructure and in combination with standard aging treatments on the mechanical performance of Laser Powder Bed Fusion (LPBF) 17-4 PH were explored and the following conclusions were reached:

• A Condition A-like (solution-annealed) microstructure is achievable by using either a 1040 °C or 1150 °C treatment on the as-built LPBF microstructure.
• For all evaluated starting conditions resulting from a 930 °C, 1040 °C, and 1150 °C treatment, the typical progression seen in wrought material for aging per ASTM A564 was demonstrated. Peak aging occurred at the H900 treatment with the highest strength, and the over-age treatments of H1025 and H1150 demonstrated reduced strength and increased ductility.
• The use of a 1150 °C homogenization treatment results in improved strengths over a 1040 °C solution-treated material at the cost of reduced ductility.
• Material treated at 930 °C exhibits a unique primary BCC microstructure with small amounts of retained austenite on the order of what was measured in the as-built material without typical martensitic features. This treatment results in a lowered overall strength, but demonstrates improved ductility.