Fatigue Crack Growth Behavior of Additively Manufactured Ti Metal Matrix Composite with TiB Particles

Abstract

Fatigue crack growth behavior of additively manufactured Ti metal matrix composite with TiB particles at room temperature was studied using a compact tension specimen and at the stress ratio of 0.1 (R = 0.1). The composite studied in this work was manufactured with a unique additive technique called plasma transferred arc solid free-form fabrication, which was designed to manufacture low-cost near-net-shaped components for aerospace and automotive industries. The fatigue crack growth rate experiments were carried perpendicular and parallel to the additive material build, aiming to find any fatigue anisotropies at room temperature. The findings reveal that additively manufactured Ti-TiB composite shows isotropic fatigue properties with respect to fatigue crack growth. Furthermore, the fatigue crack growth mechanisms in this additive composite material were identified as void nucleation/coalescence and the bypassing of particles and matrix, depending on the interparticle distance.

1. Introduction

Composites are manufactured by combining at least two materials with the aim of engineering properties required for various applications that cannot be achieved by a single material alone. The portion of the material that occupies a relatively small volume is known as reinforcement, while the remaining portion is called the matrix. Based on the type of material used for the matrix, composites can be broadly categorized as metal, ceramics, and polymer matrix composites [1]. Due to their remarkably engineered properties for weight and cost savings, composites find widespread application in the transportation industry for components of automotive and aerospace body structures, braking systems, turbine parts, and heat exchangers [2,3,4,5]. In this regard, Al, Mg, and Ti are the three widely used matrices in engineering metal matrix composite (MMC) materials for automotive and aerospace application [6,7,8]. MMCs are often developed to reduce production costs and environmental impact while providing excellent material properties. Considering the metal Ti, it is a difficult element to extract from its ore, due to its high affinity for oxygen and nitrogen [9]. Additionally, Ti alloys are difficult to machine, due to their poor thermal conductivity and high chemical reactivity with the cutting tool material [10]. Consequently, the manufacturing cost of Ti alloys generally exceeds that of their competitive counterparts, such as aluminum, magnesium, and steel alloys [11]. In order to reduce the production cost of Ti alloys and to expand their utilization in advanced applications within the aerospace and automotive industries, researchers have explored manufacturing Ti MMCs with fiber or particle reinforcements via additive manufacturing (AM) techniques [12,13,14]. Some of the particle reinforcements that are generally used in Ti MMCs are TiB, TiC, TiN, SiC, and TiB₂ [8]. Further, the two widely used AM techniques for Ti MMC productions are selective laser melting [12] and direct energy deposition [13]. Moreover, a novel AM technique called plasma transferred arc solid free-form fabrication (PTA-SFFF) [14] has been developed with the aim of manufacturing near-net-shape Ti components. In the present work, a Ti MMC containing TiB particle reinforcements in a commercially pure alpha Ti matrix manufactured by PTA-SFFF was studied.

The preference for particle reinforcements over fibers in MMCs stems from their affordability, higher potential for achieving isotropic properties, excellent specific strength and stiffness, and the feasibility of near-net-shape manufacturing [8,15,16,17]. Furthermore, the selection of a specific particle for a particular metal matrix is determined by its mechanical and thermal properties. For instance, the properties of particle-reinforced metal matrix composites can be optimized by selecting a particle reinforcement that possesses an equivalent or superior Young’s modulus and ultimate tensile strength value, lower density, and a similar thermal expansion coefficient to the metal matrix. The values for these properties for some frequently used particle reinforcements in Ti MMCs are tabulated in

Table 1. Properties of frequently used particle reinforcements in Ti MMCs [23,24,25].

Property	Ti (Pure)	TiB	TiC	TiN	SiC	TiB2
Young’s modulus (GPa)	110	550	460	390	420	565
Ultimate tensile strength (GPa)	0.22	8	3.55	Refractory compound	3.45	1.8 *
Density (g/cm3)	4.57	4.56	4.92	5.43	3.21	4.50
Thermal expansion coefficient at 20 $℃$$(\times$10−6 K−1)	8.8	8.6	7.4	9.35	4.3	6.4

* Compressive strength (GPa).

. In addition to the properties listed in Table 1, factors such as the volume fraction of particles and their shape, size, and interparticle distance can also influence the composite properties [18,19]. Moreover, when considering AM MMCs, their properties are influenced not only by the characteristics of the reinforcement and matrix material but also by the AM processing parameters, such as the energy of the power source, layer thickness, energy scan velocity, distance between successive material builds, and scanning strategies [20,21,22].

Due to the layer-by-layer material build process and the defects associated with the manufacturing process, AM materials often exhibit anisotropic mechanical properties [26,27,28,29]. Some of the most common defects reported in AM components are balling (a defect that results from discontinuous metal powder fusion), porosity, cracks, distortion/delamination, poor surface finish, and chemical degradation/oxidation [30,31]. One of the important mechanical properties to consider in many components in aerospace and automotive applications is fatigue, as they are extensively exposed to fluctuating stresses. The fatigue properties of a material can be studied through experiments such as stress- and strain-based fatigue analysis, as well as fatigue crack propagation analysis [32]. In the current work, a study on the fatigue crack growth rate (FCGR) of Ti-TiB MMC manufactured by PTA-SFFF is presented for the first time. Similar studies on FCGR in steels and nickel alloys, which are considered competitive materials for Ti alloys, can be found in [33,34,35,36,37,38]. In studies [33,34,35,36], Ni-based super alloys and in [37,38] steel were studied for their FCGRs. In general, FCGR experiments are carried out using four different types of samples: compact tension (CT); four-point bend (FPB); single-edge notch bend (SENB); and center-notched (CN) specimens. The test conditions, such as specimen geometry, stress ratio (R), frequency of the load cycle, and the alloy compositions used in the studies presented in [33,34,35,36,37,38], are listed in

Table 2. FCGR test conditions for the experiments reported in [33,34,35,36,37,38].

Material	Ni-Based Superalloys				Steel
	P/M * Ni-Based Turbine Disk Alloy	P/M * Udimet 720	P/M * Rene’ 104	Ni-Based Superalloy	Low Alloy TRIP Steel	Stainless Steel 304L
Composition	A	B	C	D	E	F
Stress ratio (R)	0.1	0.1	0.1	0.1	0.1	0.1
Environment	air	air	air	air	air	air
Temperature ($℃$)	650 725 650 725	RT	RT	RT	RT	RT
Specimen type	SENB	FPB	FPB	CT	CN	CT
Frequency (HZ)	20	80–105	25–50	10	10	2–15 0.5–10 3–20
Reference	[33]	[34]	[35]	[36]	[37]	[38]

* Powder Metallurgy (P/M).

and

Table 3. Composition of the materials listed in Table 2.

Composition Label in Table 2	Alloy Composition (in wt.%)	Reference
A	12.5 Cr, 20.7 Co, 2.7 Mo, 3.5 Ti, 3.5 Al, 0.03 C, 0.03 B, 4.3 W, 0.05 Zr, 1.6 Ta, 1.5 Nb, 49.59 Ni	[33]
B	0.02–0.04 C, 0.03–0.04 B, 0.025–0.50 Zr, 4.75–5.25 Ti, 2.25–2.75 Al, 17.5–18.5 Cr, 2.75–3.25 Mo, 14.0–15.5 Co, 1.10–1.40 W, balance Ni	[34]
C	Nickel-base superalloy, ME3 (also known as Rene´ 104): alloy composition was not provided	[35]
D	0.4 V, 1.6 Al, 2.4 Ti, 10.8 O, 19.0 Cr, 65.8 Ni	[36]
E	0.188 C, 1.502 Mn, 0.254 Si, 0.443 Al, 0.015 P	[37]
F	0.03 C, 18.0–20.0 Cr, 8.0–12.0 Ni, 1.0 Si, 2.0 Mn, 0.03 S, 0.045 P	[38]

.

As shown in Table 2, the experimental work reported in [33,34,35] studied Ni alloys manufactured through the powder metallurgy (P/M) process. However, each of them focused on different aspects involved in the propagation of fatigue cracks, such as microstructure, environmental factors, and the physical nature of the crack. For instance, the study presented in [33] finds that the grain size and temperature significantly affect the fatigue crack growth behavior in Ni alloy with the composition given in Table 3. Additionally, this study reports that borides and other precipitates, such as carbides and primary ${\gamma }^{\prime }$, affect the secondary fatigue crack propagations by blocking and deflecting them, which can be advantageous in improving the fatigue properties of this alloy. Furthermore, the work presented in [34] reports on the fatigue crack growth behavior of small and long cracks in different fatigue stress intensity factor ranges. It reveals that small cracks exhibit faster growth rates than long cracks across all fatigue stress intensity factor ranges. Moreover, the microstructural observations performed in this work report faceted fracture surfaces around the crack nucleation sites. Also, the experimental study presented in [35] finds that, during fatigue, small surface cracks in nickel-base superalloy Rene´ 104 show transgranular crack propagation with resistance at the grain boundaries. Similar to [34], fracture surface images in [36] reveal a highly faceted crack growth mechanism. Moreover, this work reports on the effect of temperature on fatigue strength, revealing that the Ni alloys studied lost almost one-third of their fatigue strength at high temperatures.

The work presented in [37,38] investigates FCGRs of two steel alloys with the compositions listed in Table 3. The results presented in [37] reveal that the microstructure of TRIP steels plays a significant role in their FCGR. It was reported that when the microstructure of the TRIP steel contains ferrite/martensite, fatigue cracks propagate through the coalescence of numerous microcracks that form at the interface between ferrite and martensite phases. Additionally, FCGRs of stainless steel 304L can be found in [38]. In this work, FCGRs at various stress ratios (R) were presented without discussing many microscopic details.

The following sections of this paper will present the fatigue crack growth behavior of Ti-TiB MMC manufactured by PTA-SFFF. Section 2 will present the experimental method and the results obtained from the current study. This will include the details pertaining to the FCGR experiment using CT specimen. Specifically, it will cover the details of the FCGR test specimens, the setup used, and FCGR experimental results in two different AM material build directions. Furthermore, it will include specimen locations for optical microscopic images of the AM MMC, as well as electron microscopic images of the fracture surfaces. This will be followed by Section 3, in which the current FCGR experimental data will be utilized to evaluate the Paris law constants for the Ti MMC, as well as a comparison of the FCGRs in steels and Ni alloys will be presented along with the current the experimental data. Moreover, this section will discuss possible fatigue crack growth mechanisms involved in the fatigue of Ti-TiB MMC, and it will present suggestions for improving its fatigue strength. Finally, in Section 4 the findings of the current study will be presented concisely.

2. Experimental Method, and Results

In the present work, FCGR of AM Ti-TiB composite was tested using compact tension (CT) specimens. These specimens were cut using electrical discharge machining (EDM) from an AM Ti-TiB composite block in two different additive build directions, as shown in Figure 1a. The dimensions of the CT specimens shown in Figure 1b were determined in accordance with the ASTM E647 standard [39], taking into consideration both the material’s availability and the availability of suitable grips for clamping the specimens during the FCGR test.

To differentiate between experiments conducted with different AM build directions, CT specimen having the specimen symmetry line aligned along the AM direction (0$°$) is labelled as LC, while the specimen with the symmetry line perpendicular to the AM direction (90$°$) is labelled as PC. In this study, an MTS machine was employed to conduct FCGR experiments. Figure 2 illustrates three views of the FCGR test setup, highlighting the locations of crucial components, including the CT specimen, MTS fixture, pins securing the specimen with the fixture, and the displacement gauge. According to the ASTM E647 standard, each FCGR test specimen was pre-cracked to a length of $a_0$ = 2 mm before being clamped onto the MTS machine (100 KN MTS Landmark Servohydraulic Test System Model no. 370.10). Additionally, all the tests were conducted at room temperature and at R = 0.1.

Using the experimental setup in Figure 2, the FCGR experimental data for the AM Ti MMC were collected under a constant R = 0.1 and at a constant FCG frequency of 15 Hz. The collected FCGR experimental data for both LC and PC specimens, which includes fatigue crack growth rate (da/dN) and stress intensity factor range ($∆K$) in the current study, are presented in Figure 3 as a log(da/dN)–log($∆K$) plot. This plot includes an inset that magnifies the constant gradient region, which is also known as the Paris region.

Since the present study examines the crack propagation behavior in an AM composite, investigating the microstructure of this material across various AM build directions is crucial for identifying microscopic anisotropy. In this regard, Figure 4 shows optical microscope images obtained from polished-etched surfaces representing the LC and PC crack propagation surfaces, labelled as L and P (Figure 4a), respectively. In the microscopic images shown in Figure 4b,c, the dark needle-like areas are TiB particles in the alloy composite, and the rest of the regions are grains of alpha Ti matrix. The composite studied in the current work contains TiB particles, hence to further explore the impact of TiB particles on the fatigue behavior of Ti-TiB MMC, the fatigue fracture surfaces of the PC test specimen were examined using SEM and Focused Ion Beam Scanning Electron Microscopy (FIB-SEM) and these images are given in Figure 5.

3. Discussion

Fatigue crack growth rate in metal alloys is generally analyzed in three distinct crack growth situations based on the gradient of the fitted curve for the log (da/dN)—log ($∆K$) data: decreasing (stage 1), constant (stage 2), and increasing (stage 3) gradients. These crack growth scenarios arise due to a number of competing deformation mechanisms, some of which allow fatigue crack growth while others blunt it. In stage 1, the microstructure of the material, mean stress of the fluctuating stress, and the environment in which the material experiences fatigue influence the growth behavior. In the constant gradient growth scenario (stage 2), which is also known as the Paris growth region, it was reported that it is largely influenced by a combination of fatigue environment, mean stress, and the frequency of the cyclic stress. In the third stage, the environment plays little influence, but the microstructure, mean stress, and the thickness of the material make a large contribution in causing the final fracture.

It can be seen from Figure 3 that fatigue cracks propagate slightly differently in the AM Ti-TiB MMC along 0$°$ and in 90$°$ AM directions. Since both LC and PC specimens were tested under the same fatigue test conditions by maintaining the same environment (air and room temperature), mean stress, and frequency (15 Hz), the observed crack growth behavior difference seen in Figure 3 can be attributed to the microstructural difference in LC and PC specimens. It can be observed from the optical microscopic images in Figure 4 that both the L and P specimens exhibit a similar grain structure. However, there is a slight disparity in the distribution of TiB particles within the matrix grains. Additionally, these images indicate the presence of frequent TiB clusters in both the L and P specimens. Two representative TiB clusters observed in the AM composite are encircled in Figure 4b,c. Upon examining the fatigue fracture of AM Ti-TiB composite through the SEM and Energy-dispersive X-ray spectroscopy (EDS) maps, it became apparent that TiB particles exhibit more significant damage compared to the Ti matrix. The SEM images revealing this damage pattern can be found in [29,40]. Given the substantial fatigue damage in the TiB particles, it is evident that fatigue cracks can readily propagate in regions with clusters of TiB particles, as depicted by the encircled areas in Figure 4.

According to the analysis presented in [18], the distribution of particles plays a significant role in determining the fatigue crack propagation behavior in metal matrix composites. Hence, the difference in the fatigue crack propagation behavior observed between LC and PC specimens can be attributed to the way that TiB particles are distributed in the Ti matrix along different AM build directions.

According to the Paris equation, the linear portions of the FCGR curves presented in Figure 3 can be modeled using Equation (1).

$${\displaystyle \frac{da}{dN}}=\mathrm{c}{\left(∆K\right)}^{\mathrm{m}}$$

(1)

where $a$ is crack length, $N$ is the number of fatigue cycles, $∆K$ is the stress intensity factor range, $\mathrm{c}$ is Paris constant, which represents the intercept value of the line fitted to the linear region, and $\mathrm{m}$ is the gradient of this line. For the LC and PC specimens, the evaluated Paris constants for Equation (1) are tabulated in

Table 4. Line equations and Paris constants for the linear regions in Figure 3.

	LC Specimen	PC Specimen
Line equation	Log ${\left({\displaystyle \frac{da}{dN}}\right)}_{\mathrm{L}\mathrm{C}}$ = 5.7973 Log $\left(∆K\right)$ − 10.467	Log ${\left({\displaystyle \frac{da}{dN}}\right)}_{\mathrm{P}\mathrm{C}}$ = 6.1306 Log $\left(∆K\right)$ − 11.102
Gradient, m	5.7973 $~$ 6	6.1306 $~$ 6
Intercept	−10.467	−11.102
c	3.41193 $\times$ 10−11 $~$ 10−11	7.90679 $\times$ 10−12 $~$ 10−11

. For both LC and PC, the Paris constants remain the same ($c={10}^{-11}$ and $m=6$), suggesting that there is no influence on fatigue life from the AM build direction or variations in TiB particle distribution. In general, improved fatigue behavior was expected in the LC specimen (0° direction) due to the formation of interlayers, which help prevent crack propagation. This is evident in the lower region of the FCGR plot in Figure 3, circled in orange, where the applied ∆K for the LC specimen is higher than that for the PC specimen. This suggests that the LC specimen can endure greater stress before crack propagation begins, confirming its superior fatigue performance at lower ∆K values. However, at higher applied ∆K values, circled in blue, this trend reverses, and the PC specimen endures greater stress than the LC specimen. Moreover, the values evaluated for $c$ and $m$ in the current study are aligned with the values for grade 2 Ti alloy, commercially pure Ti alloy reported in [41]. This finding suggests confidence in the FCGR experiment conducted in the current study, as the matrix of the composite investigated in the current work is a commercially pure alpha Ti alloy.

One of the aims of developing PTA-SFF fabricated Ti-TiB composite is to produce aerospace materials that are structurally competitive with other alloys, such as steels and Ni alloys, but at a lower cost. Thus, Figure 6 presents a comparison of the FCGRs of PTA-SFFF Ti-TiB composite in two different AM build directions with those of Ni alloys and steels.

It can be seen from Figure 6 that the current FCGR data for AM Ti-TiB composite show similar growth rates to P/M Udimet 720 (composition B) and Rene´ 104 (composition C) Ni alloys at lower $∆K$ values and higher ${\displaystyle \frac{da}{dN}}$ values. The other Ni and steel alloys shown in Figure 6 display different crack growth behavior compared to the AM Ti-TiB composite, which demonstrates a high potential for resistance to fatigue crack growth.

Analysis of the fatigue crack growth behavior, based on the SEM and FIB-SEM images depicted in Figure 5, suggests that as the fatigue crack propagates, it exhibits deflections at specific locations (Figure 5b). It can be seen from Figure 5e–g that these deflection points on the crack path correspond with TiB particles. It can be seen from Table 1 that TiB possesses significantly higher strength and modulus values than pure Ti. Moreover, in the composite studied in the current work, the TiB particles show a very strong interface with the Ti matrix. This is evidenced in the TEM images shown in Figure 7.

According to Vasudevan and Sadananda [18], MMCs reinforced with high strength and high modulus particles, such as the ones investigated in the current work, experience void nucleation during fatigue, and fatigue crack growth follows the path along coalesced voids. In the case of composites with a strong particle–matrix interface, as observed in the composite studied in the current work, fatigue crack growth occurs along the coalesced voids and bypasses the particles. Furthermore, the study [18] finds that the interparticle distance can affect the fatigue crack growth behavior by affecting slip mode, void nucleation, and crack deflection behavior. Considering the findings reported in [18], along with the SEM and FIB-SEM images presented in Figure 5, the present study suggests that fatigue crack growth in PTA-SFF fabricated Ti-TiB MMC occurs through the coalescence of voids and bypassing of TiB particles, particularly in regions where TiB particles are distributed without forming any clusters. This growth behavior is schematically illustrated in Figure 8a where regions of coalesced voids are indicated by blue arrows. However, in regions where TiB clusters are present, they can be regarded as composite regions with a brittle matrix (TiB) reinforced by a ductile material (Ti). When fatigue crack growth encounters these TiB clusters, the fatigue growth behavior shifts from bypassing the TiB particles to growing through them and the matrix, as schematically illustrated in Figure 8b.

The experimental findings and analysis of fatigue crack growth behavior of AM Ti-TiB MMC presented in this work suggest that the PTA-SFFF method should be re-evaluated for introducing TiB particle reinforcement. Additionally, the current work recommends that addressing the issue of frequent particle clustering and maintaining optimal particle distance is crucial to improving fatigue crack growth resistance and ultimately enhancing the fatigue strength of this low-cost Ti MMC through PTA-SFFF.

4. Conclusions

The fatigue crack growth study on commercially pure Ti metal matrix composite (MMC) manufactured by plasma transferred arc solid free-form fabrication (PTA-SFFF) method finds that,

• The AM technique produces Ti-TiB composite with a strong particle–matrix interface.
• The directional fatigue crack growth rate experimental results show an overall isotropic fatigue crack growth behavior.
• The slight changes in the fatigue crack growth behavior are due to the frequent clustering of TiB particles observed in the microstructure of the Ti-TiB MMC.
• The dominant fatigue crack growth mechanisms for the AM Ti-TiB MMC are identified as void nucleation and coalescence in the region where the TiB particles are distributed without any TiB clusters while crack growth through the particle and matrix in the regions with particle clusters.
• A comparison of fatigue crack growth rates with other competitive materials such as steels and Ni alloys shows that the AM process, PTA-SFFF for Ti-TiB MMC production, needs improvement in its fatigue properties.