Fabrication of Magnesium–NiTi p Composites via Friction Stir Processing: E ﬀ ect of Tool Proﬁle

: In this study, a solid-state fabrication route via friction stir processing (FSP) was used to fabricate Nitinol particulate (NiTi p )-reinforced magnesium-based composites to avoid the di ﬀ usion reaction and the formation of brittle interfacial compounds. The e ﬀ ect of four tool proﬁles on the homogeneity in the dispersion of NiTi p particles in the magnesium matrix and microhardness was examined and analyzed. A counter-clockwise scrolled shoulder with a plain cylindrical pin and three tools with a ﬂat shoulder having plain cylindrical pin, left-hand, and right-hand threaded pins were used and compared. The tool proﬁles were observed to exhibit a signiﬁcant inﬂuence on the microstructure of the fabricated Mg / NiTi p composites. A wider and more uniform distribution of NiTi p particles along with superior bonding with magnesium matrix was achieved with a left-hand threaded cylindrical pin tool. The incorporation of NiTi p gave rise to a signiﬁcant increase in the microhardness of the fabricated composites due to a variety of strengthening mechanisms.


Introduction
Nickel-titanium (NiTi) alloy, also called as Nitinol, is a popular shape memory alloy (SMA). SMA is being recognized as a smart material because of its typical shape memory effect (SME): a unique property that enables them to "memorize" their shape when exposed to specific external stimulus such as thermal, mechanical, and electromagnetic vibrations [1,2]. SMA demonstrates a unique amalgamation of properties such as super-elasticity, SME, good resilience, vibration damping and mechanical properties, compactness, and lightness, which makes it different from the other alloy systems [3,4]. Likewise, these unique properties can be carried forward if the SMAs are implanted in a base matrix to form composites. It is a favorite type of engineering materials and finds important applications in the field of driving, sensing, vibration damping, and structural and biomedical applications [5,6].
SMA-reinforced metal matrix composites would be a pioneering solution to fabricate cost-effective high actuation energy devices. The advantages derived as the composites can be made from initial SMA powder stock (which is highly cost-effective in comparison with virgin SME in bulk shape). The unique

Materials and Methods
The elemental concentration of substrate (magnesium) is presented in Table 1. The NiTi SMA with an atomic ratio of 54.9/45.1 was chosen as reinforcement in the particulate form with an average particle size of ≈5 µm. The morphology and particle size of as-received NiTi particles are shown in scanning electron microscopic (SEM) images (Figure 1a,b). Magnesium plates were prepared to a Metals 2020, 10, 1425 3 of 11 size of 150 mm × 60 mm × 8 mm. The stages involved in the composite fabrication through FSP are shown in a schematic diagram ( Figure 2). A deeper slot of 2.5 mm × 3 mm was prepared on the top surface of each plate subsequently filled with reinforcement and enclosed using a plain shouldered tool. The FSP tools were made of high-speed steel. Single-pass FSP was carried out at a fixed setting of a tool tilt angle (2 • ), tool rotational rate (560 rpm), and traversing speed (100 mm/min). The tool rotational direction was clockwise during the experiment. particle size of ≈5 µm. The morphology and particle size of as-received NiTi particles are shown in scanning electron microscopic (SEM) images (Figure 1a,b). Magnesium plates were prepared to a size of 150 mm × 60 mm × 8 mm. The stages involved in the composite fabrication through FSP are shown in a schematic diagram ( Figure 2). A deeper slot of 2.5 mm × 3 mm was prepared on the top surface of each plate subsequently filled with reinforcement and enclosed using a plain shouldered tool. The FSP tools were made of high-speed steel. Single-pass FSP was carried out at a fixed setting of a tool tilt angle (2°), tool rotational rate (560 rpm), and traversing speed (100 mm/min). The tool rotational direction was clockwise during the experiment.    Metals 2020, 10, x FOR PEER REVIEW 3 of 12 particle size of ≈5 µm. The morphology and particle size of as-received NiTi particles are shown in scanning electron microscopic (SEM) images (Figure 1a,b). Magnesium plates were prepared to a size of 150 mm × 60 mm × 8 mm. The stages involved in the composite fabrication through FSP are shown in a schematic diagram ( Figure 2). A deeper slot of 2.5 mm × 3 mm was prepared on the top surface of each plate subsequently filled with reinforcement and enclosed using a plain shouldered tool. The FSP tools were made of high-speed steel. Single-pass FSP was carried out at a fixed setting of a tool tilt angle (2°), tool rotational rate (560 rpm), and traversing speed (100 mm/min). The tool rotational direction was clockwise during the experiment.    Four experiments were carried out using various tool profiles as illustrated in Figure 3 and Table 2. The tool profiles were selected in this study for investigating the material stirring and particle distribution during FSP. The counter or anti-clockwise (ACW) scroll was made on the shoulder for the first experiment (Figure 3a). The left-hand threaded and right-hand threaded tools are abbreviated as LHT and RHT, respectively. The detailed specification of the tool profiles is shown in Figure 3a Four experiments were carried out using various tool profiles as illustrated in Figure 3 and Table  2. The tool profiles were selected in this study for investigating the material stirring and particle distribution during FSP. The counter or anti-clockwise (ACW) scroll was made on the shoulder for the first experiment ( Figure 3a). The left-hand threaded and right-hand threaded tools are abbreviated as LHT and RHT, respectively. The detailed specification of the tool profiles is shown in Figure 3a-d.  The specimens for the microstructural investigation were sectioned across the processing direction from the composite plates and prepared using a standard metallographic procedure, including a rotating disc polisher (QS Metrology, New Delhi, India) up to a grit size of #2500, followed by diamond polishing using a 0.5 µm diamond abrasive paste on a velvet cloth. For the examination of microstructures, the polished coupons were chemically etched with acetic-picral solution [30]. Microstructures were observed using light microscopy (LM) (QS Metrology, New Delhi, India) and SEM (ZEISS EVO MA 15, Jena, Germany). A Vickers microhardness indenter (Mitutoyo, Kawasakishi, Kanagawa, Japan) was used to measure the indentation behavior of the composite plates at a load of 1 N for a dwell time of 15 s.

Results and Discussion
Friction stir processing was effective in the fabrication of NiTip SMA-reinforced magnesiumbased composites. Microstructural characterization together with an indentation test was carried out to identify the role of tool profiles in the fabrication of Mg-NiTip composites. The surface morphology of all the composite plates is presented in Figure 4a   The specimens for the microstructural investigation were sectioned across the processing direction from the composite plates and prepared using a standard metallographic procedure, including a rotating disc polisher (QS Metrology, New Delhi, India) up to a grit size of #2500, followed by diamond polishing using a 0.5 µm diamond abrasive paste on a velvet cloth. For the examination of microstructures, the polished coupons were chemically etched with acetic-picral solution [30]. Microstructures were observed using light microscopy (LM) (QS Metrology, New Delhi, India) and SEM (ZEISS EVO MA 15, Jena, Germany). A Vickers microhardness indenter (Mitutoyo, Kawasaki-shi, Kanagawa, Japan) was used to measure the indentation behavior of the composite plates at a load of 1 N for a dwell time of 15 s.

Results and Discussion
Friction stir processing was effective in the fabrication of NiTi p SMA-reinforced magnesium-based composites. Microstructural characterization together with an indentation test was carried out to identify the role of tool profiles in the fabrication of Mg-NiTi p composites. The surface morphology of all the composite plates is presented in Figure 4a  Macrographs of the specimens processed using various tool profiles are shown in Figure 5a-d. No macroscopic defects such as tunnel, cavity, and pores were observed in the macrographs of the fabricated composites. These images show a typical cross-sectional view of processed zones (in the transverse section of composite specimens), where AS and RS indicate the advancing side and the retreating side, respectively. These macrographs reveal the existence of different zones such as a processed zone (PZ), reinforced zone (RZ), and an unaltered base metal (BM). PZ is a zone that has experienced stirring action by the rotating tool. It consists of reinforced and particle-free regions. A minute region of clusters of NiTi particles is also witnessed in the RZ of all the specimens. It is observed from the Figure 5a,b,d that the scrolled shoulder with a plain cylindrical pin, plain shoulder with a plain cylindrical pin, and plain shoulder with a right-hand threaded cylindrical (RHT) cylindrical pin FSP tools could not disperse NiTi particles widely in the magnesium matrix, whereas the composite processed with the plain shoulder with a left-hand threaded cylindrical (LHT) cylindrical pin tool could more effectively disperse the NiTip in a wider zone (Figure 5c).
The flow of plasticized material during friction stir welding/processing (FSW/P) can be better understood via the model of Schneider and Nunes (2004). Important material movement patternsnamely, ring vortex, rigid disc rotational, and uniform translational-may evolve during FSW/P (Figure 6a-c) [31]. The pin thread supports ring vortex flow through the pin/probe threads driving the material up in the outward direction and subsequently inward at the tool shoulder surface, down from the probe threads, and then outward again from the bottom portion of the probe, thus helping mix the reinforcement with the BM matrix ( Figure 7a). This type of flow is observed when the directions of the pin thread and tool rotation are opposite. Reversing the tool rotational direction may change the vortex movement direction of the material (Figure 7b) [31,32]. The flow of material during processing is influenced by both the traverse and rotational directions of the tool [33]. Chances for the formation of defects such as the porosity and poor bonding near the lower part are high when processing with an RHT pin tool due to opposite vortex movement direction [32]. In the present investigation, no such porosity was observed with the RHT pin tool. It is clear that the ring vortex flow has been experienced by the material during processing with an LHT pin tool, while the tool rotates in the clockwise direction and could contribute to the better distribution of NiTip. Macrographs of the specimens processed using various tool profiles are shown in Figure 5a-d.
No macroscopic defects such as tunnel, cavity, and pores were observed in the macrographs of the fabricated composites. These images show a typical cross-sectional view of processed zones (in the transverse section of composite specimens), where AS and RS indicate the advancing side and the retreating side, respectively. These macrographs reveal the existence of different zones such as a processed zone (PZ), reinforced zone (RZ), and an unaltered base metal (BM). PZ is a zone that has experienced stirring action by the rotating tool. It consists of reinforced and particle-free regions. A minute region of clusters of NiTi particles is also witnessed in the RZ of all the specimens. It is observed from the Figure 5a,b,d that the scrolled shoulder with a plain cylindrical pin, plain shoulder with a plain cylindrical pin, and plain shoulder with a right-hand threaded cylindrical (RHT) cylindrical pin FSP tools could not disperse NiTi particles widely in the magnesium matrix, whereas the composite processed with the plain shoulder with a left-hand threaded cylindrical (LHT) cylindrical pin tool could more effectively disperse the NiTi p in a wider zone (Figure 5c).  The flow of plasticized material during friction stir welding/processing (FSW/P) can be better understood via the model of Schneider and Nunes (2004). Important material movement patterns-namely, ring vortex, rigid disc rotational, and uniform translational-may evolve during FSW/P (Figure 6a-c) [31]. The pin thread supports ring vortex flow through the pin/probe threads driving the material up in the outward direction and subsequently inward at the tool shoulder surface, down from the probe threads, and then outward again from the bottom portion of the probe, thus helping mix the reinforcement with the BM matrix (Figure 7a). This type of flow is observed when the directions of the pin thread and tool rotation are opposite. Reversing the tool rotational direction may change the vortex movement direction of the material (Figure 7b) [31,32]. The flow of material during processing is influenced by both the traverse and rotational directions of the tool [33]. Chances for the formation of defects such as the porosity and poor bonding near the lower part are high when processing with an RHT pin tool due to opposite vortex movement direction [32]. In the present investigation, no such porosity was observed with the RHT pin tool. It is clear that the ring vortex flow has been experienced by the material during processing with an LHT pin tool, while the tool rotates in the clockwise direction and could contribute to the better distribution of NiTi p .
Metals 2020, 10, x FOR PEER REVIEW 6 of 12   Higher magnification LM images of the clustered zone of the processed specimens are shown in Figure 8a-d. In the clustered regions, reinforcement shows poor bonding with the metal matrix, as the reinforcement was washed out during etching, giving a dark-pit type appearance. The LM images   Higher magnification LM images of the clustered zone of the processed specimens are shown in Figure 8a-d. In the clustered regions, reinforcement shows poor bonding with the metal matrix, as the reinforcement was washed out during etching, giving a dark-pit type appearance. The LM images       Typical scanning electron micrographs (SEM) of specimen 3 processed with an LHT pin is shown in Figure 10a,b. Regions of clustering and the dispersion of NiTi p are displayed in Figure 10a,b, respectively. The interface between NiTi p and the magnesium matrix is clean. Furthermore, one of the most favorable observations is that there are no signs of diffusion and interfacial product around the NiTi particles in the magnesium matrix as a consequence of FSP. Thus, one of the greatest challenges in the processing of NiTi metal matrix composites, i.e., the control of interfacial reactions to form Mg 2 Ni, Mg-Ti-O, Mg-O, etc., in the case of powder and liquid metallurgy practices due to the high activity of NiTi at the processing temperature and time, can be effectively avoided. The reaction products are brittle IMCs and are harmful for interfacial bonding strength [17,20,34]. Moreover, as the NiTi p has special SME characteristics, the presence of interface reactions could also spoil this useful property. The highest temperature and time, which evolve during processing in FSP, is significantly lower as compared with liquid and powder metallurgy practices. During the FSP of the Al or Mg metal matrix, the temperature of the plasticized material is about 400 • C in the stir zone, and the thermal cycle is very short (a few seconds), which significantly lowers the likelihood of interfacial/diffusion reactions; hence, no diffusion/interfacial products are formed during NiTi p -reinforced composite fabrication [25][26][27][28][29].
Metals 2020, 10, x FOR PEER REVIEW 8 of 12 Typical scanning electron micrographs (SEM) of specimen 3 processed with an LHT pin is shown in Figure 10a,b. Regions of clustering and the dispersion of NiTip are displayed in Figure 10a,b, respectively. The interface between NiTip and the magnesium matrix is clean. Furthermore, one of the most favorable observations is that there are no signs of diffusion and interfacial product around the NiTi particles in the magnesium matrix as a consequence of FSP. Thus, one of the greatest challenges in the processing of NiTi metal matrix composites, i.e., the control of interfacial reactions to form Mg2Ni, Mg-Ti-O, Mg-O, etc., in the case of powder and liquid metallurgy practices due to the high activity of NiTi at the processing temperature and time, can be effectively avoided. The reaction products are brittle IMCs and are harmful for interfacial bonding strength [17,20,34]. Moreover, as the NiTip has special SME characteristics, the presence of interface reactions could also spoil this useful property. The highest temperature and time, which evolve during processing in FSP, is significantly lower as compared with liquid and powder metallurgy practices. During the FSP of the Al or Mg metal matrix, the temperature of the plasticized material is about 400 °C in the stir zone, and the thermal cycle is very short (a few seconds), which significantly lowers the likelihood of interfacial/diffusion reactions; hence, no diffusion/interfacial products are formed during NiTipreinforced composite fabrication [25][26][27][28][29]. The microhardness distribution across the fabricated composites is shown in Figure 11. It was measured 2 mm below the top surface, along the width from AS to RS of the processed region. The notable enhancement in hardness of all the specimens was observed in the RZ region in comparison with the BM hardness, which was measured as 35 ± 5 HV. The average values of microhardness in the PZ was observed to be about 59, 45, 97, and 106 HV for specimens 1, 2, 3, and 4, respectively. The peak hardness of 385 HV was observed in specimen number 4 at the middle of the PZ. A close examination of this specimen revealed the localized accumulation of reinforcement NiTip, which leads to a spike in the peak hardness value in the indented region. A relatively more homogeneous distribution of hardness values was observed in specimen number 3, which was processed with an LHT cylindrical pin tool. The microhardness distribution across the fabricated composites is shown in Figure 11. It was measured 2 mm below the top surface, along the width from AS to RS of the processed region. The notable enhancement in hardness of all the specimens was observed in the RZ region in comparison with the BM hardness, which was measured as 35 ± 5 HV. The average values of microhardness in the PZ was observed to be about 59, 45, 97, and 106 HV for specimens 1, 2, 3, and 4, respectively. The peak hardness of 385 HV was observed in specimen number 4 at the middle of the PZ. A close examination of this specimen revealed the localized accumulation of reinforcement NiTi p , which leads to a spike in the peak hardness value in the indented region. A relatively more homogeneous distribution of hardness values was observed in specimen number 3, which was processed with an LHT cylindrical pin tool. Zheng et al. [35] compared the microhardness of NZ30K magnesium alloy in unprocessed and FSPed conditions. An improvement of ~26% was observed at the center of the stir zone of the FSPed sample as compared to the unprocessed region. The improvement in the microhardness of the FSPed sample was attributed to the significant grain refinement during FSP. In the present investigation, a Zheng et al. [35] compared the microhardness of NZ30K magnesium alloy in unprocessed and FSPed conditions. An improvement of~26% was observed at the center of the stir zone of the FSPed sample as compared to the unprocessed region. The improvement in the microhardness of the FSPed sample was attributed to the significant grain refinement during FSP. In the present investigation, a careful observation on the indentation values and particle distribution indicates that the processed region without reinforcement shows a~30-55% improvement in microhardness values, and a more than three-fold increase in the microhardness value was observed in the composited region due to the presence of NiTi p reinforcement as compared to BM.
During FSP, heat is generated due to friction between the FSP tool and matrix material, which plasticizes the material underneath, and the rotating tool produces plastic deformation and stirring of the BM along with reinforcement. As the tool starts its travel, the material ahead of the tool pin is deposited behind, which fills the void created by the advancing tool pin. The FSP causes intense or severe plastic deformation, which mixes the NiTi p with the matrix material and refines the grains due to the occurrence of dynamic recrystallization (DRX) [36,37]. During continued stirring, the NiTi p further refine the grains successively by impeding the growth of grains via Zener pinning [38]. In the phenomenon of Zener pinning, the incorporated reinforcement particles in the matrix act as obstacles to the movement of grain boundaries. The difference in the coefficients of thermal expansion (CTE) of the deformed matrix and NiTi p reinforcement results in an increase in dislocations and further increases the hardness of composites [39][40][41]. All these simultaneous occurring phenomena increase the microhardness of composites. Apart from the above discussed mechanisms, the higher hardness of the reinforcement particles also contributes to the higher hardness of composites [38,39,42].
It should be noted that the functional benefits of NiTi p reinforcement cannot be reaped until the distribution of these particles in the matrix is proper and also in a desired volume fraction. Stirring with a uniform distribution of reinforcement particles within Mg is challenging, and this study represents some preliminary attempts of producing acceptable surface composites from NiTi p in pure Mg affected by the tool profiles. More studies in this aspect, including the effect of processing speed and rotational rate, are needed to optimize the FSP so as to produce the high-performance composites for functional advantages.

Conclusions
The Mg/NiTi p composites were developed via the FSP technique and investigated for their microstructural aspects. The present study demonstrated that the tool profile greatly affected the microstructure in the Mg/NiTi p composite fabrication. The results obtained could be concluded as follows:

1.
A wider distribution of NiTi p was attained with a left-hand threaded (LHT) cylindrical pin tool among the four tool profiles. SEM observations indicated good bonding of NiTi p with the magnesium matrix.

2.
The incorporation of NiTi p resulted in a significant increase in the microhardness of the fabricated composites due to a variety of hardening mechanisms presented.

3.
The LHT pin tool processed composite showed a more uniform hardness profile across the processed zone, with an average microhardness value of 97 HV.