Effects of crack orientation and heat treatment on fatigue-crack-growth behavior of AM 17-4 PH stainless steel

Journal Pre-proofs Effects of Crack Orientation and Heat Treatment on Fatigue-Crack-Growth Behavior of AM 17-4 PH Stainless Steel Aref Yadollahi, Mohamad Mahmoudi, Alaa Elwany, Haley Doude, Linkan Bian, James C. Newman Jr PII: DOI: Reference:

S0013-7944(19)31233-0 https://doi.org/10.1016/j.engfracmech.2020.106874 EFM 106874

To appear in:

Engineering Fracture Mechanics

Received Date: Revised Date: Accepted Date:

4 October 2019 22 November 2019 4 January 2020

Please cite this article as: Yadollahi, A., Mahmoudi, M., Elwany, A., Doude, H., Bian, L., Newman, J.C. Jr, Effects of Crack Orientation and Heat Treatment on Fatigue-Crack-Growth Behavior of AM 17-4 PH Stainless Steel, Engineering Fracture Mechanics (2020), doi: https://doi.org/10.1016/j.engfracmech.2020.106874

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2020 Published by Elsevier Ltd.

Effects of Crack Orientation and Heat Treatment on Fatigue-CrackGrowth Behavior of AM 17-4 PH Stainless Steel

Aref Yadollahi1*, Mohamad Mahmoudi2, Alaa Elwany2, Haley Doude1, Linkan Bian3, and James C. Newman Jr.4

1

Center for Advanced Vehicular Systems (CAVS), Mississippi State University, Starkville, MS

39759, USA 2Department

of Industrial and Systems Engineering, Texas A&M University, College Station, TX

77843, USA 3

Department of Industrial and Systems Engineering, Mississippi State University, Mississippi State,

MS 39762, USA 4

Department of Aerospace Engineering, Mississippi State University, Mississippi State, MS

39762, USA

*Corresponding author: Email: [email protected] Phone: (662) 617-9534

Submitted to: Engineering Fracture Mechanics (EFM)

Abstract The effects of heat treatment and crack orientation on fatigue-crack-growth (FCG) behavior of 17-4 precipitation hardening (PH) stainless steel (SS) fabricated via laser powder bed fusion (LPBF) additive manufacturing (AM) were investigated.

Accurate representation of FCG

thresholds, the region defining crack growth as either very slow or nonexistent, is extremely important in adoption of AM for structural applications. In this study, new methods called compression pre-cracking constant-amplitude (CPCA) and load-reduction (CPLR) were used to generate FCG rate data in the near-threshold (low-rate) regime using non-standard or modified compact specimens. For comparison, the current ASTM load-reduction (LR) method was also used. Results indicate that FCG behavior of AM 17-4 PH SS in the Paris regime was comparable to that of wrought 17-4 PH SS in H900 condition. Although post-manufacturing heat treatment (including solution annealing plus peak-aging) was found to be necessary in order to improve tensile strength of this alloy, the results of this study indicate that this particular heat treatment process had no influence on FCG behavior of LPBF 17-4 PH SS. The effect of crack orientation with respect to the build direction however was evident. FCG rates were slightly lower for specimens with crack parallel to the build direction (i.e. longitudinal crack) in the Paris regime as compared to that of specimens with perpendicular cracks (i.e. transverse crack). The nearthreshold FCG behavior for specimens with longitudinal crack did not show the usual stress ratio (R)-shift from low to high R. The low-R data was vastly different than the data generated using the specimens with a transverse crack. Our results also imply that both the current ASTM standard LR method and the new CPLR method induce remote closure, which prematurely slows down crack growth and produces an abnormally high threshold.

Keywords: Additive manufacturing; Laser powder bed fusion; Fatigue-crack-growth; Fracture; Fatigue-crack thresholds

Glossary AM BFS CA CPCA CPLR C(T) MC(T) FCG EBSD EDM LAB LOF LPBF LR SEM TAB THT

Additive manufacturing Back-face strain Constant amplitude Compression pre-cracking constant-amplitude Compression pre-cracking load-reduction Compact specimen Modified compact specimen Fatigue-crack-growth Electron backscatter diffraction Electrical discharge machining Specimens with a longitudinal crack in as-built condition Lack-of-fusion Laser powder bed fusion Load-reduction Scanning electron microscopy Specimens with a transverse crack in as-built condition Specimens with a transverse crack in heat-treated condition

Nomenclature c cn C da/dN E F h K Kc Kmax Ncp Pmax Pmin R Ra w α ΔK ΔKeff (ΔKeff)th Kth σut σys

Crack length, mm Initial EDM notch length, mm Load reduction rate, mm−1 Crack-growth rate, m/cycle Modulus of elasticity, GPa Stress-intensity boundary-correction factor Initial EDM notch height, mm Stress intensity factor, MPa√m Fracture toughness, MPa√m Maximum stress intensity factor, MPa√m Number of pre-cracking cycles Maximum applied remote load, kN Minimum applied remote load, kN Stress ratio Surface roughness, µm Width of specimen, mm Constraint factor Stress intensity factor range, MPa√m Effective stress intensity factor range, MPa√m Effective stress intensity factor range threshold, MPa√m Threshold stress intensity factor range, MPa√m Ultimate tensile strength, MPa Yield stress (0.2% offset), MPa

1. Introduction During the last decades layer-wise production techniques have evolved from a rapid prototyping concept to an additive manufacturing (AM) technique for production of functional parts within many industrial sectors (such as aerospace, tooling, and biomedical applications) [1,2]. The Additive Manufacturing (AM) industry has proved to be a thriving industry. In 2017, AM products and services worldwide grew 21% to about $7 billion [3]. Metal AM processes have opened a completely new era in design freedom and innovation by having the ability to produce parts with complex geometries that are difficult or even impractical to produce with conventional production methodologies [4]. The other key capabilities of metal AM, particularly laser powder bed fusion (LPBF) processes, are part consolidation [5], design optimization [6], and material waste reduction [7]. These capabilities have attracted considerable research effort, which in turn has resulted in new industry trends such as 4D printing, alloy design, functional grading, and linking AM to cyber-physical systems in order to contribute to the fourth industrial revolution [8,9]. However, there are a number of challenges to be solved before AM processes can be fully utilized in industries for fabricating fracture-critical components [4,10]. LPBF processes involve complex multi-physics phenomena – such as fluid flow, heat transfer, phase change, and solidification – which makes the fabricated parts highly prone to residual stresses, microstructure heterogeneity, and manufacturing defects, including pores and lack-of-fusion (LOF) defects [4,11– 13]. The lack of understanding of AM process-structure-property relationships, particularly related to the mechanical performance and structural durability of AM parts, is a major concern that must be addressed to fulfill the operational requirements and certification constraints for applications with high quality requirements (e.g., healthcare and aerospace) [4,11,14]. Performance of many AM materials under static loading – such as tensile strength, hardness,

and impact toughness – has been well characterized and was found to be comparable with those of conventionally manufactured ones [4,11,15]. However, performance of AM material under cyclic loading, including fatigue life and particularly the crack-growth resistance, are associated with more substantial complexities as compared to that of wrought counterparts and generally lower [11,15–17]. The crack-growth resistance of long cracks is influenced by material microstructure, which particularly manifests itself in the threshold region [18]. Several studies have reported that AM materials exhibit significantly lower threshold stress intensity range values as compared to conventionally produced materials [15,17,19,20]. In addition, for the given value of the stress intensity factor, a considerable scatter of the experimentally determined crack growth rate data has been reported for AM materials [15,21,22], which can be attributed to the texture and residual stresses. A number of investigations have also indicated that the crack growth rate – especially in threshold regions – of AM materials are dependent on the orientation of crack with respect to the build direction [17,21], which is a consequence of microstructure directionality (i.e. texture) and residual stresses. More research is needed to better understand and to more accurately evaluate the fatigue-crack-growth (FCG) behavior of AM materials in the threshold region, particularly with respect to build orientation and post-manufacturing heat treatment. Fatigue strength and resistance to propagation of cracks are important material parameters for design of components. The increasing demand for damage-tolerant design of AM materials necessitates a detailed knowledge on FCG behavior [23,24]. More importantly, accurate representation of fatigue-crack thresholds, the region defining crack growth as either very slow or nonexistent, is vital for design of many structural applications [25]. In North America, the threshold crack-growth regime is experimentally defined using the ASTM Standard E647 [26], which has been shown to exhibit anomalies due to the load-reduction (LR) test procedure for many

materials [25].

Recently, compression pre-cracking constant-amplitude (CPCA) and load-

reduction (CPLR) test procedures have been proposed as alternative methods, to generate fatiguecrack-growth (FCG) rates in the near-threshold regime [27–29]. The CPLR method is similar to the current LR method, but the test generally starts at an order-of-magnitude lower crack-growth rate than allowed in the current standard. To date, no FCG test data has been published on AM materials using CPCA or CPLR methods. While there are fair amounts of data available on the microstructural characteristics and static performance of AM materials, limited work can be found discussing the FCG behavior. For instance, extensive work conducted pertaining to the characterization of microstructure and static mechanical properties, and even a few works published on fatigue and fracture behaviors of 17-4 precipitation hardening (PH) stainless steel (SS) [30–36]. However, there are no readily-available reports on the FCG data in the “threshold” regime without load-history effects, of this particular alloy in the open literature to the authors’ knowledge. Thus, the goal of this research is to evaluate the effects of post-manufacturing heat treatment and crack orientation on the FCG behavior of AM 17-4 PH SS using non-standard or modified compact, MC(T), specimens. In addition, a CPLR method was used to generate crack-growth rates in the near-threshold and threshold regimes. This paper experimentally determines the FCG rates from threshold to fracture using four methods: (i) Kmax testing, (ii) compression pre-cracking constant-amplitude (CPCA) loading, (iii) compression pre-cracking load-reduction (CPLR) loading, and (iv) ASTM load-reduction (LR) testing. FCG tests were conducted over a wide range in stress ratios (R = 0.1 to 0.7) using MC(T) specimens in different conditions (i.e. as-built and heat-treated) and crack orientations (i.e. parallel and perpendicular to the build direction).

2. Material and Experimental Procedures 2.1. Specimen fabrication and preparation Commercially-available, gas-atomized (under argon) 17-4 PH stainless steel (SS) powder (3D Systems) – with 80% of its particle size distribution below 22 µm – was utilized in this study to fabricate samples using an LPBF system (3D Systems ProX 100). Optimized process parameters, reported in [33], were chosen for fabricating 17-4 PH SS rectangular blocks under inert argon atmosphere, as shown in Fig. 1(a). All the rectangular blocks were manufactured in the same direction with a height of 48.8 mm, width of 58.8 mm, and thickness of 6.35 mm. Modified compact, MC(T), specimens were extracted from these blocks (see for example ASTM E647-15 [26]), using electrical discharge machining (EDM). In order to consider the effect of material anisotropy on the FCG behavior, two sets of MC(T) specimens were made by changing the initial notch orientation: (i) crack propagation direction perpendicular to the build direction (i.e. transverse crack) and (ii) crack propagation direction parallel to the build direction (i.e. longitudinal crack), as shown in Figs. 1(b) and 1(c), respectively. For MC(T) specimens with the transverse crack, duplicate sets of samples were manufactured and subjected to solution annealing for 30 min at approximately 1040 ºC, air cooling (AC) to room temperature (Condition A) and age hardening heat treatment for 1 hour at 482 ºC, followed by AC (Condition H900 or peak-aging). Thus, a total of three sets of AM 17-4 PH SS were taken into account: (i) MC(T) specimens with transverse crack in as-built condition (TAB), (ii) MC(T) specimens with longitudinal crack in as-built condition (LAB), and (iii) heat-treated MC(T) specimens with transverse crack (THT). Tensile properties of AM 17-4 PH SS in different orientations and conditions are listed in Table 1. Tension tests were conducted on vertical (i.e. parallel to the build direction) and horizontal

(i.e. perpendicular to the build direction) tensile specimens using a servo-hydraulic testing machine at a 0.001/s nominal strain rate [33]. As can be seen from Table 1, as-built AM specimens (in both orientations) show significantly lower yield and ultimate tensile strengths as compared to wrought materials in H900 condition (i.e. peak-aging) [37]. However, after heat treatment, the yield and ultimate tensile strengths of AM 17-4 PH SS became comparable to that of wrought material in H900 condition. This is due to the formation of fine, coherent precipitates in the martensite matrix during the peak-aging (i.e. at 450-510ºC), causing the material to become significantly harder with higher tensile strengths [38]. MC(T) specimens had a width (w) of 40.6 mm, an initial EDM notch height (h) of ~0.4 mm and notch length (cn, measured from the pin-hole centerline) of ~12.2 mm, as illustrated in Fig. 1(d). The specimens were the same as a standard C(T), except the pin-holes which were were located further from the notch than the standard specimen. The stress-intensity factor (K) and backface strain (BFS) relations for the modified and standard C(T) specimens were compared with the ASTM E647 equations [26]. The calculations indicated that the standard K solution and BFS equations [26] can be used for the MC(T) specimens, since the K values were within ±2% and BFS values were within 1.4% for c/w ≥ 0.3. One of the side faces for the MC(T) specimens was mechanically polished – in water and under very low pressure– to a mirror finish (surface roughness of ~0.7 µm) in order to better monitor the crack propagation. The other side face was slightly polished only to remove the roughness arising from partially-melted powder particles. The edges of the pin-holes in the specimens were beveled to avoid or minimize undesired out-of-plane bending moments (pins forced to contact near midthickness of the specimen). 2.2. Test procedures

The FCG tests were conducted over a wide range of stress ratios (R = 0.1 to 0.7) and produced rates from near-threshold to fracture on MC(T) specimens. All the FCG tests were performed under laboratory air conditions at room temperature on 5 kN servo-hydraulic test machines. The loads were applied in sinusoidal waveform at 18 Hz in the low-rate regime and about 3–5 Hz in the highrate regime. Crack lengths were monitored by using a BFS gage compliance procedures, as outlined in ASTM E-647 [26]. Test control was provided by a data acquisition/test control system for threshold FCG rate testing. The load sequences applied to the MC(T) specimens are shown in Fig. 2. To generate data in the near-threshold regime, all specimens were pre-cracked under constant-amplitude compression–compression loading (R = 22 or 40) for about 38,000 cycles on 5 or 25 kN servohydraulic test machines. Using these procedures, the pre-notched specimens were cycled under compression-compression loading to initiate a crack at the EDM crack-starter U-notch. Typical crack sizes at the notch tip were 0.1 to 0.35 mm (0.004 to 0.014 in), resulting from compression cycling under two loading conditions. FCG rate tests were then conducted using either (i) constant Kmax testing, (ii) compression pre-cracking constant-amplitude (CPCA) loading, or (iii) compression pre-cracking load-reduction (CPLR) at constant R after a small amount of crack extension under CPCA loading. The shed rate for the Kmax and load-reduction testing was -0.08 mm-1. For comparison, the ASTM load-reduction (LR) test method was also used in this study to generate crack-growth data in the near-threshold regime. The constant Kmax test is a K-controlled test, where the Kmax value is held constant and the minimum K value is slowly increased with crack length [39], as schematically shown in Fig. 2(a). As the ΔK value decreases, the corresponding crack-growth rate will also decrease and approach a threshold condition. In general, larger values of Kmax generates ‘lower’ thresholds, which has

been attributed to crack-closure behavior [40]; that is, higher Kmax values show less crack closure. Yamada and Newman [41,42] showed what is called “high-R” closure using the load-reduction methods to generate low crack-growth rates. Under constant-amplitude loading, the high-R tests will exhibit no crack closure, and the cracks will be fully open. The ‘compression-compression pre-cracking constant-amplitude’ (CPCA) loading, as shown in Fig. 2 (b), was designed to generate FCG rates in the threshold regime under constant-amplitude loading conditions with minimal or no load-history effects, after the crack has grown several compressive plastic-zone sizes. Using this procedure, pre-notched specimens were cycled under compression-compression pre-cracking to produce an initial crack (Pmin = −5 to −7 kN; R = 22 or 40; f = 8 Hz; Ncp ≈ 38,000 cycles). The resulting crack tip is enveloped by a small tensile residual stress field instead of the typical compressive plastic zone normally resulting from tension-tension loading and in general, free of any crack closure caused by crack surface roughness or the compressive plastic zone. Then the specimen was subjected to a constant-amplitude (CA) fatigue loading to generate FCG-rate data in the near-threshold regime at the desired stress ratio. ASTM load reduction (LR) test method [26] was designed to fully reproduce the range of fatigue-crack thresholds (e.g. low and high stress ratios), based on stress-intensity factors changing at an exponential rate. Initial starting load levels were carefully selected to ensure that growth rates immediately from the starting notch were less than 10−8 m/cycle (4×10−7 in/cycle), as required in the standard [26]. A load reduction rate of C = −0.08 mm−1 (−2 in−1) was maintained in all LR tests. Upon developing rates at or near the target 10−10 m/cycle, test control was changed to CA loading, ΔK increasing, and the crack was grown to failure. The loading for the compression pre-cracking load reduction (CPLR) testing method is depicted in Fig. 2(c). This method is basically CPCA followed by load reduction using current

ASTM E-647 LR procedures [26], except that the initial stress-intensity factor range and crackgrowth rate at the start of the load-reduction test was much lower than the maximum allowed rate in the standard [29]. Threshold testing for both procedures at a given stress ratio was done in pairs (e.g. first CPLR, followed immediately by a test under the CPCA procedures) to minimize any laboratory environment influences that might occur over time. Testing procedures for the MC(T) specimens are summarized in Table 2. The microstructures, crystallographic texture and orientation and phase fraction of cylindrical sections of samples were examined using an optical microscopy and a field emission gun scanning electron microscopy (SEM) equipped with an electron backscatter diffraction (EBSD) detector. Fracture surface of MC(T) specimens were examined using SEM to determine fracture characteristics.

3. Microstructural Characterization Microstructural examination of AM 17-4 PH SS samples revealed the presence of large lackof-fusion (LOF) defects (or un-melted regions), which can be attributed to low laser penetration depth. These types of defects are slit-shaped voids (flaws), which can cover a broad region with small volume, as shown in Fig.3. Thus, they may not be detected by density measurement techniques that depend on pore volume. This indicates that process parameter optimization based on density alone may not be a sufficiently accurate criterion [33]. Electron backscatter diffraction (EBSD) was used to quantify the crystallographic texture, microstructure (e.g. grain size and orientation), and phase volume fraction of 17-4 PH SS samples before and after heat treatment. Figure 4 shows EBSD grain-structure and phase distribution maps for the as-built 17-4 PH SS samples in three perpendicular planes, parallel and perpendicular to the build direction. For as-built 17-4 PH SS samples, the results of microstructural study displayed the presence of larger elongated grains around the track boundaries and finer grains concentrated near the center of the melt pool. Columnar grains oriented toward the top center of the melt pool can be noticed for as-built 17-4 PH SS samples from Figs. 3 and 4. This directionality is attributed to a higher temperature gradient between the melt pool and previous layer, causing a more-unidirectional heat flux. The EBSD phase fraction map confirmed the presence of both martensitic and austenitic phases in the AM 17-4 PH SS samples. Figure 4 reveals the distribution of two phases, consisting of a dominant martensite phase and less-prevalent finely dispersed austenite phase. Although fullymartensitic material can be expected at equilibrium for this alloy [31], small amounts of retained austenite “islands” (approximately 5-10%) were homogeneously formed within the boundaries of

martensite grains, as can be seen in Fig. 4. Formation of retained austenite during AM process can be attributed to several reasons. Existing literature on AM 17-4 PH SS has reported strong dependence of material microstructure and fraction of retained austenite on process parameters [31]. In addition, powder manufacturing environment (argon or nitrogen gas), formation of smaller grain sizes and interdendritic spacing, presence of strain at high-angle grain boundaries, and relatively higher dislocation density in AM process can affect the amount of retained austenite in AM 17-4 PH SS [30–32]. A higher amount of retained austenite was observed in the plane perpendicular to the build direction, Fig. 4(c), as compared to the planes parallel to the build direction, Figs. 4(a) and (b). This is attributed to the formation of columnar grains, resulting from directionality of the heat flux vector which contains a dominant component parallel to the build direction. The EBSD grain-structure and phase distribution maps for the heat-treated AM17-4 PH SS in the plane parallel to the build direction is shown in Fig. 5. As can be seen, post-fabrication heat treatment (i.e. solution annealing and peak-aging processes) resulted in recrystallization and homogenization of microstructure and removed interfacial regions (i.e. laser ‘footprint’ or track boundaries) between deposited layers of as-built samples. However, the utilized heat treatment did not change the average grain size significantly. EBSD mapping data for phase fraction revealed approximately 5% reduction in volume fraction of retained austenite after solution annealing and tempering during peak-aging.

4. Fatigue-Crack-Growth Behavior Fatigue-crack-growth (FCG) tests were conducted on AM 17-4 PH SS MC(T) specimens over a wide range of load-ratio conditions (0.1 ≤ R ≤ 0.7) and constant Kmax tests. All specimens were compression pre-cracked (CP) before FCG testing, except the ASTM load-reduction test. The test parameters for all investigated conditions are summarized in Table 2. The experimentally determined FCG rate data for long cracks under three different conditions generally ranged from near-threshold to fracture. The crack-growth rates are plotted against the stress intensity factor range (K). In general, the experimental data in different conditions (as-built and heat-treated) and crack orientations (transverse and longitudinal) showed a little more scatter than wrought alloys. In the case of the as-built specimens, the scatter might be attributed to the residual stresses and LOF present in the specimens. Occurrence of non-uniform residual stresses approaching the yield stress has been reported for the as-built AM materials [17,20]. However, the same amount of scatter can be noticed for the heat-treated specimens as the as-built ones, indicating that the scatter can also be related to the non-continuous crack growth due to the presence of large defects (i.e. LOF). Figure 6(a) shows the test data generated on the MC(T) specimens with transverse crack in the as-built condition (TAB), which were tested at high load ratio (R = 0.7) and Kmax equal constant tests. The vertical dashed line is an estimated effective threshold stress intensity factor range [43]. The value of the effective stress-intensity factor, (ΔKeff)th, is given by 1.2×10−5 E (MPa√m), where E is the modulus of elasticity. For one of the MC(T) specimens (TAB-6), the first part of the test was a CPLR test at R = 0.7 that went down to below 2×10−10 m/cycle. Then, after terminating the LR test, a CA test at R = 0.7 was conducted to growth the crack to failure. The ΔK-rate data agreed well with one of the other R = 0.7 test specimens that underwent the same testing procedures (TAB-

4). However, one of the other specimens (TAB-5), in which the first part of the test was a CPLR test at R = 0.1, showed more scatter and higher values of crack growth rates for ΔK values between 11 and 17 MPa√m. This specimen (TAB-5) fractured at nearly the same ΔK value as the other two tests. The FCG rate data approaching fracture was used to estimate the fracture toughness (Kc) value of 70 MPa√m. The Kmax-equal-constant tests, which is a load-reduction test that generates crack-growth data at very high R, were conducted to determine the effective stress-intensity factor (ΔKeff) against rate curve. Since the cracks are nearly crack-closure free, that is, ΔK = ΔKeff. In the near-threshold regime, the constant Kmax test produced lower K values than the R = 0.7 test, which can be related to high-R closure effects in R = 0.7 LR test [40,41]. Test data below 10−10m/cycle is very hard to generate, and load-reduction tests are expected to generate load-history effects. Thus, new methods are needed to generate the extremely low crack-growth-rate data. For the AM 17-4PH SS specimens with transverse crack in the as-built condition (TAB), the estimate (ΔKeff)th was 2.25 MPa√m, which is shown by the vertical dashed line. Figure 6(b) shows the tests conducted at low load ratios (R = 0.1 and R = 0.4) on MC(T) specimens with transverse cracks in as-built condition (TAB). A CPCA test at R = 0.1 was conducted that started at about 2×10−10m/cycle and was grown under CA loading to failure (TAB3). Two MC(T) specimens underwent the CPLR tests at R = 0.1 (TAB-5 and TAB-7). As can be seen from Fig. 6(b), the two CPLR tests were in excellent agreement. After reaching nearly 10−10 m/cycle, CA tests at R = 0.4 and R = 0.7 were conducted on these specimens. It should be noted that to avoid load-history affects after conducting a LR test at R = 0.1, the constant-amplitude test was conducted at higher R, such as 0.4 or 0.7. These new testing conditions are being included in the new ASTM E-647 testing standard [26].

4.1. Effect of Heat Treatment Figures 7(a) and (b) show the test data generated on the heat-treated MC(T) specimens with transverse crack (THT), which were tested at high load ratio (R= 0.7) and Kmax test as well as low load ratios (R = 0.1 and R = 0.4), respectively. Similar to the testing procedures for TAB specimens, one MC(T) specimen (THT-2) was first tested at R = 0.7 using CPLR method that went down to below 2×10−10 m/cycle. Then, after terminating the LR test, a constant-amplitude test at R = 0.7 was conducted to growth the crack to failure. The ΔK-rate data agreed well with one of the other R = 0.7 test specimens (THT-4), in which the first part of the test was a CPLR test at R = 0.1. This specimen fractured at nearly the same ΔK as the other test (THT-2). A constant Kmax test was conducted at shed rate C = −0.08 mm−1 and later at a faster shed rate of C = −0.16 mm−1 was used until a very slow crack growth rate was achieved. Again, as can be seen from Fig. 7(a) rates for the constant Kmax test were higher than the R = 0.7 rates at the same ΔK values in the near-threshold regime. This can be related to the closure effect in R = 0.7 load-reduction tests. The estimated (ΔKeff)th for the heat-treated AM 17-4PH SS specimens with transverse crack (THT) was 2.25 MPa√m, the same as the non-heat-treated specimens (TAB), shown by the vertical dashed line in Fig. 7(a). Low load ratios (R = 0.1 and 0.4) test data for heat-treated MC(T) specimens with transverse crack (THT) are shown in Fig. 7(b). A CPCA test at R = 0.1 was conducted that started at ~5×10−10 m/cycle and was grown under CA loading to failure (THT-3). One MC(T) specimen underwent the CPLR tests at R = 0.1 (THT-4). After reaching nearly 2×10−11 m/cycle, a CA test at R = 0.7 was conducted on this specimen. A CA test was conducted at R = 0.4 on another specimen after conducting an ASTM load-reduction (LR) test at R = 0.1. The results for the standard ASTM LR test are shown in Fig. 8. The crack was grown to the

maximum allowed rate (10−8 m/cycle) under tensile pre-cracking loads and then the standard LR scheme was used to reach nearly 10−10 m/cycle. As seen from Fig. 8, the CPCA and CPLR tests generated a slightly lower threshold and faster rates than using the standard ASTM load-reduction method in the low-rate regime (R = 0.1). ASTM load-reduction test generated a threshold at about 5.5 MPa√m, which is approximately 10% higher than the threshold obtained using CPLR test method (5.0 MPa√m). Comparing the FCG results of MC(T) specimens with transverse crack before and after heat treatment (Figs. 6 and 7) indicates that heat treatment did not change the longcrack behavior of AM 17-4 PH SS noticeably. As seen, FCG test results are very similar for asbuilt and heat-treated specimens from near-threshold to fracture. Furthermore, it can be seen from Figs 6 and 7 that the FCG behavior in the Paris regime for AM 17-4 PH SS specimens with transverse cracks in both as-built and heat-treated conditions (i.e. TAB and THT) are comparable to that of wrought 17-4 PH SS in H900 condition [44]. This implies that as opposed to fatigue life, which is considerably lower for AM 17-4 PH SS as compared to its wrought counterparts due to presence of process-induced defects [33], FCG behavior of AM 17-4 PH SS is not significantly influenced by such defects. 4.2. Effect of Crack Orientation Figure 9 shows the test data generated on the MC(T) specimens with longitudinal crack in asbuilt condition (LAB), tested at high load ratio (R= 0.7) and Kmax test as well as low load ratio (R = 0.1). One of the MC(T) specimens (LAB-1) was compression pre-cracked (CP) and then subjected to CA loading at R = 0.6 to grow the crack away from the crack-starter notch. Then a constant Kmax test was conducted at shed rate C = −0.08 mm−1 and later at a faster shed rate of C = −0.16 mm−1 was used until a very slow crack growth rate was achieved. During this phase, the crack was growing at about a 15-degree angle to the normal crack plane. Then a CA test at R = 0.7

was conducted, but the crack bifurcated (see Section 5 for details). Non-straight crack growth, with respect to the notch orientation, has also been reported in the literature for as-built AM specimens as-built conditions as a result of high residual stresses [17]. For one of the other specimens (LAB2), the first part of the test was a CPLR test at R = 0.7 with a shed rate of C = −0.16 mm−1. The test data (open circles) agreed very well with the previous Kmax test in the low-rate regime. During this phase, the crack grew fairly straight. Then, after terminating the load-reduction test, a CA test at R = 0.7 was conducted to growth the crack to failure (open squares). Again, the crack grew nonstraight (curved crack), therefore, the test data is invalid during this phase. After removing the test specimen from the lower clevis, the specimen seemed to be tight, indicating some unsymmetrical loading may have been applied to the specimen. Another MC(T) specimen (LAB-3) was tested using CPLR method at R = 0.1. Load-history effects are expected, and the test data should go to a threshold at a higher ΔK value, however, test data at R = 0.1 surprisingly follow the Kmax and R = 0.7 test data, as can be seen from Fig. 9. Although the reason for this observation is not clear, it may be attributed to having large amounts of residual stress in the specimen. After achieving the very low rates, this specimen was tested at R = 0.1 under CA loading to generate test data at high rates approaching fracture. The selected load level was high enough, so that the past load history would not affect the test data at the high rates. The crack grew straight for a while, but the rapid rise in ΔK against rate was when the crack began to fork into two cracks; then, one crack became dominant and the crack straightened out (16 to 22 MPa√m) and later bifurcated into two vertical cracks (see Section 5 for details). Anisotropic FCG behavior has been reported for AM materials in many studies [15,18,21,45– 47]. The reason was sought in the residual stresses and directionality of microstructure (i.e. texture). However, some studies were reported an equivalent FCG behavior for as-built AM

specimens with different crack orientations (i.e. transverse and longitudinal) [19,47–49]. Their results indicated that the differences in crack growth rates as a function of crack or loading orientation, in both threshold and Paris regimes, were not discernible [19,48–50]. In this study, AM specimens with different crack orientations displayed distinct FCG behavior with respect to each other. Comparing the FCG results of MC(T) specimens with longitudinal crack (LAB) and transverse crack (TAB) indicates that the effect of crack orientation in regard to the build direction is evident on the long crack behavior of AM 17-4 PH SS. As seen, the estimate (ΔKeff)th for the LAB specimens is very close to that of TAB specimens, i.e. 2.25 MPa√m, shown by the vertical dashed line in Fig. 9. However, the near-threshold FCG behavior for specimens with longitudinal crack did not show the usual R-shift from low to high R. The low-R data was vastly different than the data generated using the specimens with transverse crack. The FCG rate is also slightly lower for LAB specimens in the Paris regime as compared to that of TAB counterparts, indicating the crack growth resistance is weaker when crack is perpendicular to the build direction (i.e. along the layers). This can be attributed to greater tendency of crack to propagate along lower strength melt pool boundary layers in TAB specimens [51].

5. Failure Mechanism A typical fracture appearance of FCG test demonstrates the pre-crack, crack growth, and final failure regions. However, the pre-cracking zone herein – with a size of 100 µm – was not clearly distinguishable from crack growth zone for AM 17-4 PH SS. Analysis of the fracture surfaces of AM 17-4 PH SS specimens revealed no significant difference in the fracture surface characteristics between the as-built and heat-treated (i.e. TAB versus THT) specimens. In fact, very similar fracture surface morphology was observed for different conditions. Moreover, comparing the fracture surface of the fatigue life specimens from our previous study [33] with that of the fatigue crack propagation specimens herein revealed very comparable fracture features for the specimens of the two fatigue tests. For both near-threshold region, where the crack rate is low (i.e. Region I), and the crack growth region (i.e. Region II or the Paris Law region) crack propagated in a transgranular way with cleavage fracture mode. The fine fatigue striations, which are perpendicular to the propagating direction was identified in several locations over the crack growth zone. As the propagating crack approaches to the point of rapid crack growth (i.e. the final failure region, Region III), the fracture surface translates from the brittle transgranular fracture (i.e. cleavage) to a quasi-cleavage – which is a mixed mechanism involving both microvoid coalescence and cleavage – and finally, to a ductile transgranular fracture mode (i.e. dimpled rupture). A magnified view of the transition area from cleavage to microvoid coalescence is presented in Fig. 10(a). The overload fracture surface (i.e. final failure zone) demonstrated very rough surface containing a large number of pores, LOF defects, and un-melted powder particles, as depicted in Fig. 10(b). Similar morphology and roughness were observed for fracture surface of crack growth region at the different stress ratios. Moreover, the typical fracture surface appearance did not change with

increasing crack propagation rate along the crack growth region, e.g. from the low crack growth rate regime (10−9 m/cycle) up to the crack rate of the order 10−7 m/cycle. This can be seen by comparing Figs. 11(a) and 11(b), showing the fracture surface morphologies for a crack length of c ≈ 12 mm (ΔK = 6 MPa√m) and c ≈ 28 mm (ΔK = 21 MPa√m), respectively. These images were taken from a MC(T) specimen tested at R = 0.1 under CPCA method. A distinct fracture surface morphology, however, was identified for the crack growth region under the CA test method as compared to the CPLR method – for similar loading ratios. This difference can be noticed from Fig. 12, showing the interface of these two areas, where the CPLR test was terminated (ΔK = 5×10−10 m/cycle) and the CA test was started (3×10−8 m/cycle), both tests were conducted at R = 0.7. Different fracture surface morphologies can be attributed to the applied load levels since the Pmax was 1.69 kN (380 lbf) when the LR test was terminated and 4.45 kN (1,090 lbf) when the CA test was started (the crack length can be considered the same for both tests at this location). Analysis of the fracture surfaces in the crack growth zone revealed that the crack-growth velocity below the surface is slightly greater than that at the surface, as can be seen from Fig.13(a). However, the crack-growth velocity in the midsection of specimens was found to be lower as compared to that at near the side walls, where a relatively smooth region (slant fracture region) can be observed. The appearance of a crack front in the crack growth zone, shown in Fig.13(b), indicates retardation of the fatigue crack propagation in the midsection. This can be attributed to the presence of large LOF defects in the midsection of specimens, which can blunt the crack tip and retard the fatigue crack propagation rates. Although the reduction in load-bearing area due to these defects can increase the fatigue crack propagation speed, the crack blunting effects seem to be more significant. The magnified view of the crack retardation event caused by defect can be clearly identified in Fig.13(c).

For some of the MC(T) specimens with transverse crack in both as-built and heat-treated conditions (i.e. TAB and THT), we noticed that the crack tends to propagate more on the mirrorpolished edge, as can be seen from Fig.14. Examination of the side view of the crack path revealed a very flat crack growth region for the mirror-polished side and a tortuous crack propagation for the other side. This phenomenon may not be attributed to the change in microstructure caused by polishing, since the polishing procedures were conducted under very low pressure in water. Although the reason for this observation is not clear – since it is not a case for all the specimens – it could be related to the crack retardation induced by the surface roughness. Large delamination (i.e. splitting in the crack growth plane) were identified in the crack growth region of a few specimens with transverse crack in both as-built and heat-treated conditions (i.e. TAB and THT), as shown in Fig. 15. Since crack path is aligned with the deposited layer in these specimens, it would be possible that crack propagates along the interface between the layers (interfacial delamination). Thus, this delamination can be attributed to interfacial debonding of the deposited layers, resulting from the poor metallurgical bonding between layers – due to low laser penetration. The size of these features was found to be several millimeters in some cases. In general, small pores hardly have any influence on the large crack growth behavior [52]. However, the large debounding observed in this study can be expected to cause crack growth retardation by changing the crack growth direction and/or the creation of secondary cracks, which redistributes the stresses around the crack tip, and consequently, reduces the crack driving force. Specimens with longitudinal crack (LAB) were observed to have an apparent smoother fracture surface in the final failure zone than MC(T) specimens with transverse crack in both as-built and heat-treated conditions (TAB and THT). However, similar morphology and roughness were observed for fracture surface of crack growth region for both crack orientations. The rougher final

fracture surface of specimens with transverse crack can be attributed to the orientation of the LOF defects with respects to the crack path. When the crack propagates along the layers, and consequently, a larger cross-sectional area of LOF defects is subjected to the loading direction along the crack, causing a rougher fracture morphology. No delamination was also identified in the crack growth region of the LAB specimens since crack propagates across the layers, thus, no interfacial debonding of the deposited layers can be expected. For one of the LAB specimens, crack was observed to grow along a path approximately 15° inclined with respect to the notch orientation, as illustrated in Fig. 16. For one of the other specimens, the crack grew straight for a while, but the crack began to fork into two cracks, and then, one crack became dominant and the crack straightened. After several millimeters of crack extension in both cases, the crack growth path was suddenly bifurcated into two vertical cracks (i.e. along the deposited layers), as shown in Fig.16. The inclined crack propagation path can be related to the residual stresses. Riemer et al. [17] were also reported crack-growth in a non-straight path, with respect to the notch orientation, for the AM specimens in as-built conditions as a result of high residual stresses. The sudden deflection of the propagating crack is however occurred when the crack tip encounters a poorly bounded layer, formed as result of insufficient fusion or cracked due to a high degree of residual stress. Such unbounded layer was evident at the same location (i.e. build height) for all specimens, indicating an insufficient melting in this layer during the build of this set, as can be seen from Fig. 17.

6. Conclusions The effects of post-manufacturing heat treatment (solution annealing plus peak-aging) and crack orientation on the fatigue-crack-growth (FCG) behavior of 17-4 PH stainless steel (SS) fabricated via laser powder bed fusion additive manufacturing (AM) were investigated. FCG rate tests were conducted on modified compact specimens over a wide range in stress ratios (R = 0.1 to 0.7; Kmax tests) and produced rates from near-threshold to fracture. Compression pre-cracking constant amplitude (CPCA) and load-reduction (CPLR) methods were proposed in this study to generate more accurate crack-growth rates in the near-threshold regime.

Based on the

experimental results, following conclusions can be drawn: 

FCG behavior in the Paris regime for AM 17-4 PH SS specimens – in both as-built and heat-treated conditions – were comparable to that of wrought 17-4 PH SS in a similar heat treatment condition.



Although post-manufacturing heat treatment was found to be essential for AM 17-4 PH SS to enhance its tensile strength to that of wrought material, experimental results indicated that the FCG behavior of this alloy was not affected by such a heat treatment. FCG test results were very similar for both as-built and heat-treated specimens from near-threshold to fracture.



The effect of crack orientation with respect to the build direction on the FCG behavior of AM 17-4 PH SS was evident. Specimens with crack perpendicular to build direction (i.e. transverse crack) showed the usual R-shift in the K-rate data, but the specimens with cracks parallel to the build direction (i.e. longitudinal crack) did not show an R-shift. For the specimens with longitudinal crack, the K-rate data was essentially independent of R from 0.1 to 0.7 and the Kmax test, and the low-R data was vastly different than the data

generated using the specimens with transverse crack.

Although the reason for this

observation is not clear, it may be attributed to having large amounts of residual stress in the specimens. 

Large delamination (i.e. splitting in the crack growth plane) were observed in the crack growth region of specimens with transverse crack in both as-built and heat-treated conditions. The delamination could be attributed to interfacial debonding of the deposited layers, resulting from the poor metallurgical bonding between layers. However, no delamination was observed in the crack growth region of specimens with longitudinal crack since no interfacial debonding of the deposited layers can be expected when crack propagates across the layers.



Inclined crack propagation with respect to the notch orientation was observed in the asbuilt specimens with longitudinal crack which could be related to the presence of high residual stresses in the specimens. The sudden deflection of the propagating crack was also occurred in the specimens with longitudinal crack when the crack tip encounters a poorly bounded layer, formed as result of insufficient fusion or cracked due to a high degree of residual stress.

References [1]

Thompson SM, Bian L, Shamsaei N, Yadollahi A. An Overview of Direct Laser Deposition for Additive Manufacturing; Part I: Historical Background and Thermal-Fluid Phenomena. Addit Manuf 2015;8:12–35.

[2]

Cain V, Thijs L, Humbeeck J Van, Hooreweder B Van, Knutsen R. Crack propagation and fracture toughness of Ti6Al4V alloy produced by selective laser melting. Addit Manuf 2015;5:68–76.

[3]

Wohlers T. Wohlers report. Wohlers Assoc Inc 2017.

[4]

Yadollahi A, Shamsaei N. Additive Manufacturing of Fatigue Resistant Materials : Challenges and Opportunities. Int J Fatigue 2017;98:14–31.

[5]

Tapia G., Elwany A. A Review on Process Monitoring and Control in Metal-Based Additive Manufacturing. J Manuf Sci Eng 2014;23:1917–28.

[6]

Aboutaleb A, Bian L, Elwany A, Shamsaei N, Thompson S, Tapia G. Accelerated Process Optimization for Laser-Based Additive Manufacturing by Leveraging Prior Studies. IE Trans 2016.

[7]

Ranjan R, Samant R, Anand S. Integration of design for manufacturing methods with topology optimization in additive manufacturing. J Manuf Sci Eng 2017;139:61007.

[8]

Ma J, Franco B, Tapia G, Karayagiz K, Johnson L, Liu J, et al. Spatial control of functional response in 4D-printed active metallic structures. Sci Rep 2017;7:46707.

[9]

Brant A, Sundaram MM. A novel system for cloud-based micro additive manufacturing of metal structures. J Manuf Process 2015;20:478–84.

[10]

Dzugan J, Prochazka R, Rund M, Podany P, Konopik P, Sei M, et al. Effects of thickness and orientation on the small scale fracture behaviour of additively manufactured Ti-6Al-

4V. Mater Charact 2018;143:94–109. [11]

Shamsaei N, Yadollahi A, Bian L, Thompson SM. An Overview of Direct Laser Deposition for Additive Manufacturing; Part II: Mechanical Behavior, Process Parameter Optimization and Control. Addit Manuf 2015;8:12–35.

[12]

King WE, Anderson AT, Ferencz RM, Hodge NE, Kamath C. Laser Powder Bed Fusion Additive Manufacturing of Metals; Physics , Computational , and Materials Challenges. Appl Phys Rev 2016;2:041304.

[13]

Ma J, Tian W, Alizadeh M. Data Driven Modeling and Optimization for Energy Efficiency in Additive Manufacturing Process With Geometric Accuracy Consideration. ASME 2018 Int. Des. Eng. Tech. Conf. Comput. Inf. Eng. Conf., 2018, p. V004T05A017V004T05A017.

[14]

Yan W, Lin S, Kafka OL, Yu C, Liu Z, Lian Y, et al. Modeling process-structure-property relationships for additive manufacturing. Front Mech Eng 2018;13:482–492.

[15]

Kone R, Kunz L, Ba A, Nicoletto G. Long fatigue crack growth in Ti6Al4V produced by direct metal laser sintering. Procedia Eng 2016;160:69–76.

[16]

Yadollahi A, Mahtabi MJ, Khalili A, Doude HR, Newman Jr JC. Fatigue life prediction of additively manufactured material: Effects of surface roughness, defect size, and shape. Fatigue & Fracture of Engineering Materials & Structures. 2018 Jul;41(7):1602-14.

[17]

Riemer A, Richard HA. Fatigue crack growth in additive manufactured products. Frat Ed Integrità Strutt 2015;9:437–46.

[18]

Riemer A, Leuders S, Thöne M, Richard HA, Tröster T, Niendorf T. On the Fatigue Crack Growth Behavior in 316L Stainless Steel Manufactured by Selective Laser Melting. Eng Fract Mech 2014;120:15–25.

[19]

Zhai Y, Lados DA, Brown EJ, Vigilante GN. Fatigue crack growth behavior and microstructural mechanisms in Ti-6Al-4V manufactured by laser engineered net shaping. Int J Fatigue 2016;93:51–63.

[20]

Kunz L, Nicoletto G, Bac A. Long fatigue crack growth in Inconel 718 produced by selective laser melting. Int J Fatigue 2016;92:499–506.

[21]

Leuders S, Thöne M, Riemer A, Niendorf T, Tröster T, Richard H a., et al. On the Mechanical Behaviour of Titanium alloy TiAl6V4 Manufactured by Selective Laser Melting: Fatigue Resistance and Crack Growth Performance. Int J Fatigue 2013;48:300–7.

[22]

Becker TH, Beck M, Scheffer C. Microstructure and mechanical properties of direct metal laser sintered Ti-6Al-4V. South African J Ind Eng 2015;26:1–10.

[23]

Gorelik M. Additive manufacturing in the context of structural integrity. Int J Fatigue 2017;94:168–77.

[24]

Romano S, Brückner-foit A, Brandão A, Gumpinger J, Ghidini T, Beretta S. Fatigue properties of AlSi10Mg obtained by additive manufacturing : Defect-based modelling and prediction of fatigue strength. Eng Fract Mech 2018;187:165–89.

[25]

Newman JC, Ruschau JJ, M.R.Hill. Improved test method for very low fatigue-crackgrowth-rate data. Fatigue Fract Eng Mater Struct 2010;34:270–9. x.

[26]

ASTM Standard. ASTM E647-08, Standard Test Method for Measurement of Fatigue Crack Growth Rates 2008.

[27]

Ruschau J, Newman JC. Compression Precracking to Generate Near Threshold Fatigue Crack Growth Rates in an Aluminum and Titanium. J ASTM Int 2008;5:1–11.

[28]

Newman JC, Yamada Y. Compression precracking methods to generate near-threshold fatigue-crack-growth-rate data. Int J Fatigue 2010;32:879–85.

[29]

Newman JC, Ziegler BM, Shaw JW, Cordes TS, Lingenfelser DJ. Fatigue Crack Growth Rate Behavior of A36 Steel using ASTM Load-Reduction and Compression Precracking Test Methods. J ASTM Int 2012;9:1–15.

[30]

Murr LE, Martinez E, Hernandez J, Collins S, Amato KN, Gaytan SM, et al. Microstructures and Properties of 17-4 PH Stainless Steel Fabricated by Selective Laser Melting. J Mater Res Technol 2012;1:167–77.

[31]

Cheruvathur S, Lass EA, Campbell CE. Additive Manufacturing of 17-4 PH Stainless Steel : Post-processing Heat Treatment to Achieve Uniform Reproducible Microstructure. JOM 2016;68:930–42.

[32]

Rafi HK, Pal D, Patil N, Starr TL, Stucker BE. Microstructure and Mechanical Behavior of 17-4 Precipitation Hardenable Steel Processed by Selective Laser Melting. J Mater Eng Perform 2014;23:4421–8.

[33]

Yadollahi A, Shamsaei N, Thompson SM, Elwany A, Bian L. Effects of building orientation and heat treatment on fatigue behavior of selective laser melted 17-4 PH stainless steel. International Journal of Fatigue. 2017 Jan 1;94:218-35.

[34]

Mahmoudi M, Elwany A, Yadollahi A, Thompson SM, Bian L, Shamsaei N. Mechanical properties and microstructural characterization of selective laser melted 17-4 PH stainless steel. Rapid Prototyp J 2017;23:280–94.

[35]

Siddiqui SF, O’Nora N, Fasoro AA, Gordon AP. Modeling the Influence of build orientation on the monotonic and cyclic response of additively manufactured stainless steel GP1/174PH. ASME 2017 Int Mech Eng Congr Expo 2017:V014T11A003--V014T11A003.

[36]

Nezhadfar PD, Burford E, Anderson-Wedge K, Zhang B, Shao S, Daniewicz SR, Shamsaei N. Fatigue crack growth behavior of additively manufactured 17-4 PH stainless steel:

Effects of build orientation and microstructure. International Journal of Fatigue. 2019 Jun 1;123:168-79. [37]

ASM Handbook. ASM Handbook Volume 4, Heat Treating. Mater Park OH ASM Int 1991.

[38]

Wu J, Lin C. Influence of Frequency on High-Temperature Fatigue Behavior of 17-4 PH Stainless Steels. Mater Trans 2003;44:713–21.

[39]

Newman JC, Schneider J, Daniel A, Mcknight D. Compression pre-cracking to generate near threshold fatigue-crack-growth rates in two aluminum alloys. Int J Fatigue 2005;27:1432–40.

[40]

Newman JC, Yamada Y. Crack-Closure Behavior of 7050 Aluminum Alloy near Threshold Conditions for Wide Range in Load Ratios and Constant Kmax Tests. Fatigue Fract Mech 37th Vol ASTM Int 2010;1.

[41]

Yamada Y, Newman JC. Crack closure under high load-ratio conditions for Inconel-718 near threshold behavior. Eng Fract Mech 2009;76:209–20.

[42]

Yamada Y, Newman JC. Crack Closure Behavior of 2324-T39 Aluminum Alloy near Threshold Conditions for High Load Ratio and Constant Kmax Tests. Int J Fatigue 2009;31:1780–7.

[43]

Newman JC, Kota K, Lacy TE. Fatigue and crack-growth behavior in a titanium alloy under constant-amplitude and spectrum loading. Engineering Fracture Mechanics. Eng Fract Mech 2018;187:211–24.

[44]

Hsu K, Lin C. High-Temperature Fatigue Crack Growth Behavior of 17-4 PH Stainless Steels. Metall Mater Trans A 2004;35:3018–3024.

[45]

Bača A, Konečná R, Nicoletto G, Kunz L. Influence of build direction on the fatigue behaviour of Ti6Al4V alloy produced by direct metal laser sintering. Mater Today Proc 3

2016:92–4. [46]

Zhao J, Easton M, Qian M, Leary M, Brandt M. Effect of building direction on porosity and fatigue life of selective laser melted AlSi12Mg alloy. Materials Science and Engineering: A. 2018 Jun 27;729:76-85.

[47]

Seifi M, Dahar M, Aman RON, Harrysson OLA, Beuth J, Lewandowski JJ. Evaluation of Orientation Dependence of Fracture Toughness and Fatigue Crack Propagation Behavior of As-Deposited ARCAM EBM Ti-6Al-4V. JOM 2015;67:597–607.

[48]

Edwards P, Ramulu M. Effect of build direction on the fracture toughness and fatigue crack growth in selective laser melted Ti-6Al-4 V. Fatigue Fract Eng Mater Struct 2015;38:1228– 36.

[49]

Edwards P, Ramulu M. Electron Beam Additive Manufacturing of Titanium Components : Properties and Performance 2018;135:1–7.

[50]

Zhai Y, Galarraga H, Lados DA. Microstructure, static properties, and fatigue crack growth mechanisms in Ti-6Al-4V fabricated by additive manufacturing: LENS and EBM. Eng Fail Anal 2016;69:3–14.

[51]

Zhao J, Easton M, Qian M. Effect of building direction on porosity and fatigue life of selective laser melted AlSi12Mg alloy. Mater Sci Eng A 2018;729:76–85.

[52]

Sangid MD, Book TA, Rotella J, Ravi P, Finch A, Park J, et al. Role of heat treatment and build orientation in the microstructure sensitive deformation characteristics of IN718 produced via SLM additive manufacturing. Addit Manuf 2018;22:479–96.

List of Tables

Table 1. Tensile properties of AM 17-4 PH SS at different conditions. Table 2. Summary of the test type, load ratio and maximum load or stress intensity factor for the specimens tested in different conditions.

Table 1. Tensile properties of AM 17-4 PH SS at different conditions. Tensile Properties E (GPa) 𝜎𝑦𝑠 (MPa) 𝜎𝑢𝑡 (MPa) Elongation to failure, % * ASM Handbook Vol. 4.

Vertical (as-built) 187.3 580 940 5.8

Vertical (heat-treated) 187.3 1,020 1,150 2.8

Horizontal (as-built) 187.3 650 1,060 14.5

Wrought* (H900 condition) 197 1,450 1,370 9

Table 2. Summary of the test type, load ratio and maximum load or stress intensity factor for the specimens tested in different conditions. Specimen Number TAB-1 TAB-2 TAB-3 TAB-4 TAB-5 TAB-6 TAB-7 THT-1 THT-2 THT-3 THT-4 THT-5 LAB-1 LAB-2 LAB-3

Load Ratio Pmax, kN (R = Pmin/Pmax) (or Kmax, MPa√m) As-built specimens with transverse crack (TAB) Kmax R = 0.5 to 0.85 22.5 Kmax R = 0.6 to 0.88 23.5 CPCA R = 0.1 1.6 CPLR; CA R = 0.7; 0.7 4.85; 4.85 CPLR; CA R = 0.1; 0.7 1.78; 3.87 CPLR; CA R = 0.7; 0.7 4.31; 2.07 CPLR; CA R = 0.1; 0.4 1.65; 1.78 Heat-treated specimens with transverse crack (THT) Kmax R = 0.6 to 0.88 23.5 CPLR; CA R = 0.7; 0.7 4.85; 4.85 CPCA R = 0.1 1.6 CPLR; CA R = 0.1; 0.7 1.78; 3.87 ASTM LR; CA R = 0.1; 0.4 4.31; 2.07 As-built specimens with longitudinal crack (LAB) Kmax R = 0.6 to 0.88 23.5 CPLR; CA R = 0.7; 0.7 4.85; 4.85 CPLR; CA R = 0.1; 0.1 1.78; 3.87 Test Type

List of Figures

Fig. 1. (a) Near-net-shaped AM 17-4 PH SS samples, schematics showing MC(T) specimens with (b) transverse notch (TAB; THT) and (c) longitudinal notch (LAB), and (d) dimensions of specimen with beveled holes and location of remote back-face strain gage. Fig. 2. Load sequences for threshold and constant-amplitude testing: (a) compression pre-cracking and constant Kmax testing, (b) compression pre-cracking and constant-amplitude (CPCA) loading (R = 0), and (c) compression pre-cracking and load-reduction (CPLR) loading (R = 0). Figure 3. Optical micrographs of AM 17-4 PH SS sample in as-built condition, showing a large lack-of-fusion (LOF) defect. Fig. 4. EBSD grain-structure and phase distribution maps of an as-built 17-4 PH SS sample in (a) sagittal, (b) longitudinal, and (c) transverse planes. Fig. 5. EBSD grain-structure and phase distribution maps of a heat-treated 17-4 PH SS sample. Fig. 6. Stress-intensity-factor range against rate for AM specimens with a transverse crack in asbuilt condition (TAB) under (a) high stress ratio (R = 0.7) and Kmax test as well as (b) low stress ratios (R = 0.1 and R = 0.4). The solid line represents FCG behavior for wrought 17-4 PH SS in H900 condition at low stress ratio (R = 0.1) [44]. Fig. 7. Stress-intensity-factor range against rate for heat-treated AM specimens with a transverse crack (THT) under (a) high stress ratio (R = 0.7) and Kmax test as well as (b) low stress ratios (R = 0.1 and R = 0.4). The solid line represents FCG behavior for wrought 17-4 PH SS in H900 condition at low stress ratio (R = 0.1) [44]. Fig. 8. Stress-intensity-factor range against rate for low stress ratio (R = 0.1) results on the heattreated AM specimens with transverse crack (THT) using CPCA, CPLR and ASTM load-reduction

methods. Fig. 9. Stress-intensity-factor range against rate for AM specimens with longitudinal crack in asbuilt condition (LAB). The solid line represents FCG behavior for wrought 17-4 PH SS in H900 condition at low stress ratio (R = 0.1) [44]. Figure 10. (a) A magnified view of the transition from the brittle transgranular fracture (i.e. cleavage) to the ductile transgranular fracture mode (i.e. dimpled rupture or microvoid coalescence) and (b) SEM fractograph of final fracture region, demonstrating pores, lack-of-fusion (LOF) defects, and un-melted powder particles. Figure 11. SEM fractographs for a MC(T) specimen tested under CPCA method, showing fracture surface at different ΔK values corresponding to different stages of the crack growth: (a) crack length of c ≈ 12 mm (ΔK = 6 MPa√m), (b) and c ≈ 28 mm (ΔK = 21 MPa√m). Figure 12. Comparison of fracture surface morphology of crack growth region for MC(T) specimen tested first under CPLR loading, and then, CA method. Figure 13. SEM fractograph of crack front, showing (a) crack development near the side wall, (b) crack retardation in middle section, and (c) crack retardation caused by a defect. Figure 14. Overall view of a fracture surface, showing unequal development of crack front at the side walls (crack front is shown by the dashed line). Figure 15. Side view of an identified delamination on the crack growth fracture surface of a vertical-as-built MC(T) specimen with transverse crack (TAB). Figure 16. Crack development along a path approximately 15° inclined with respect to the notch orientation and higher magnification of the sudden deflection to a vertical path, along the deposited layers. Figure 17. SEM images of a poorly bounded layer, as result of insufficient melting.

Fig. 1. (a) Near-net-shaped AM 17-4 PH SS samples, schematics showing MC(T) specimens

with (b) transverse notch (TAB; THT) and (c) longitudinal notch (LAB), and (d) dimensions of specimen with beveled holes and location of remote back-face strain gage.

Fig. 2. Load sequences for threshold and constant-amplitude testing: (a) compression precracking and constant Kmax testing, (b) compression pre-cracking and constant-amplitude (CPCA) loading (R = 0), and (c) compression pre-cracking and load-reduction (CPLR) loading (R = 0).

Figure 3. Optical micrographs of AM 17-4 PH SS sample in as-built condition, showing a large lack-of-fusion (LOF) defect.

Fig. 4. EBSD grain-structure and phase distribution maps of an as-built 17-4 PH SS sample in (a) sagittal, (b) longitudinal, and (c) transverse planes.

Fig. 5. EBSD grain-structure and phase distribution maps of a heat-treated 17-4 PH SS sample.

Fig. 6. Stress-intensity-factor range against rate for AM specimens with a transverse crack in as-built condition (TAB) under (a) high stress ratio (R = 0.7) and Kmax test as well as (b) low stress ratios (R = 0.1 and R = 0.4). The solid line represents FCG behavior for wrought 17-4 PH SS in H900 condition at low stress ratio (R = 0.1) [44].

Fig. 7. Stress-intensity-factor range against rate for heat-treated AM specimens with a transverse crack (THT) under (a) high stress ratio (R = 0.7) and Kmax test as well as (b) low stress ratios (R = 0.1 and R = 0.4). The solid line represents FCG behavior for wrought 17-4 PH SS in H900 condition at low stress ratio (R = 0.1) [44].

10-5

AM 17-4 PH Steel C(T) B = 6.35 mm w = 40.6 mm

10-6

VHT Specimens

dc/dN, m/cycle

10-7

R = 0.1 ASTM E-647 Maximum Allowed Rate

10-8

10-9 CPCA -1 CPLR (C = -0.16 mm ) -1 ASTM LR (C = -0.16 mm )

A36 Steel 10-10

10-11

2

3

4

5 6

8 10

15

20

30

40 50 60

1/2

K, MPa-m

Fig. 8. Stress-intensity-factor range against rate for low stress ratio (R = 0.1) results on the heat-treated AM specimens with transverse crack (THT) using CPCA, CPLR and ASTM loadreduction methods.

Fig. 9. Stress-intensity-factor range against rate for AM specimens with longitudinal crack in as-built condition (LAB). The solid line represents FCG behavior for wrought 17-4 PH SS in H900 condition at low stress ratio (R = 0.1) [44].

Figure 10. (a) A magnified view of the transition from the brittle transgranular fracture (i.e. cleavage) to the ductile transgranular fracture mode (i.e. dimpled rupture or microvoid coalescence) and (b) SEM fractograph of final fracture region, demonstrating pores, lack-of-fusion (LOF) defects, and un-melted powder particles.

Figure 11. SEM fractographs for a MC(T) specimen tested under CPCA method, showing fracture surface at different ΔK values corresponding to different stages of the crack growth: (a) crack length of c ≈ 12 mm (ΔK = 6 MPa√m), (b) and c ≈ 28 mm (ΔK = 21 MPa√m).

Figure 12. Comparison of fracture surface morphology of crack growth region for MC(T) specimen tested first under CPLR loading, and then, CA method.

Figure 13. SEM fractograph of crack front, showing (a) crack development near the side wall, (b) crack retardation in middle section, and (c) crack retardation caused by a defect.

Figure 14. Overall view of a fracture surface, showing unequal development of crack front at the side walls (crack front is shown by the dashed line).

Figure 15. Side view of an identified delamination on the crack growth fracture surface of a vertical-as-built MC(T) specimen with transverse crack (TAB).

Figure 16. Crack development along a path approximately 15° inclined with respect to the notch orientation and higher magnification of the sudden deflection to a vertical path, along the deposited layers.

Figure 17. SEM images of a poorly bounded layer, as result of insufficient melting.

Authors would like to request not to send the manuscript for review to the following scholars, due 1. 2.

to Dr. Dr.

conflict Nima

Steve

Shamsaei, Daniewicz,

3. Dr. Mohsen Seifi, Case Western Reserve University

of

interest:

Auburn University

University of

Alabama

Highlights



Fatigue-crack-growth (FCG) test results for AM 17-4 PH SS over a wide range in stress ratios (R = 0.1 to 0.7; Kmax tests) and rates (from near-threshold to fracture).



Proposing compression pre-cracking constant amplitude (CPCA) and load-reduction (CPLR) methods to generate more accurate crack-growth rates in the near-threshold regime.



Effects of heat treatment and crack orientation on FCG behavior of AM 17-4 PH SS.



Microstructural characterization and failure mechanism of AM 17-4 PH SS in different conditions.