Macro- and microstructure observation

Referring to Figs. 1(a)–1(d), the surface morphology of the weld crown for the FSW first pass in all the experimental trials are evidenced with moderate flash, whereas the FSW second pass observes an excessive flash. The reason for excessive flash during the second pass was due to the increase in plunge depth from 6.20 mm of the FSW first pass to 6.30 mm in the second pass. In addition, the reversal in the rotation direction of the tool leads to an excessive flash on the weld crown.

Figure 1:

(a)–(d) Photographs showing the surface morphology of FSW dissimilar joints with varying % volume fractions of SiC_NP reinforcement. (e) Comparison of the V _th with V _fp, and V _sp, after FSW processing. (f) Comparison of the density of the FSW specimen as a function of varying vol% fractions of SiC_NP at the WNZ for the FSW second pass.

Effect of varying percentage volume fractions of SIC_NP on the weld nugget zone of MMNC for FSW first and second pass experiment

Table I presents the optical macrograph of the cross section of the weld region of the FSW dissimilar joint specimens produced with the varying vol% fractions of SiC_NP (0, 5, 8, 13%) reinforcement at the WNZ and after performing the FSW first pass and second pass process. The macrographs exhibit the typical characteristics of FSW process consisting of WNZ, which experiences intense plastic deformation due to rotating FSW tool and consists of the highly recrystallized grain structure [Reference Mishra and Ma2, Reference Seighalani, Givi, Nasiri and Bahemmat28, Reference Mishra and Ma29, Reference Palanivel, Mathews, Murugan and Dinaharan35]. The thermomechanically affected zone (TMAZ) experiences both temperature and deformation during FSW but recrystallization does not occur at TMAZ. The heat-affected zone (HAZ) experiences only the thermal cycle because of the FSW process and does not contain any plastic deformation and grain refinement [Reference Mishra and Ma2]. It is evident from Table I, that the FSW specimen produced for FSW first pass and second pass observes the different shape of the WNZ. The FSW first pass observes defects caused by the agglomeration of SiC_NP at some regions of the WNZ. In the FSW second pass, the WNZ of all the FSW joint specimens is defect-free as a result of increased stirring effect coupled with the change in tool rotation direction. With the shift in the tool rotation direction from the first pass to the second pass, the AS in the first pass will be the RS in the second pass and this will help in the uniform distribution of SiC_NP across the WNZ [Reference Dolatkhah, Golbabaei, Givi and Molaiekiya14]. However, the detailed analysis of the WNZ is carried out using SEM to investigate the presence of microscopic-level defects such as agglomeration of SiC_NP and the presence of pores around SiC_NP which will be discussed in the following section.

TABLE I:

Macrograph images of the FSW dissimilar joint specimens showing the weld nugget area of MMNC for the varying vol% fractions of SiC_NP reinforcement and without SiC_NP for FSW first and second pass. The vol% fraction of SiC_NP present at the WNZ after the FSW first and second pass is mentioned above the WNZ.

Further, it is observed from Table I that the WNZ for the FSW first pass appears to have the shape of the “onion ring” structure while the WNZ for the FSW second pass appears like “trapezoid” in shape. The FSW first pass and second pass influenced the area of WNZ, thereby changing the volume fraction of SiC_NP in the WNZ. Sathiskumar et al. [Reference Sathiskumar, Murugan, Dinaharan and Vijay16, Reference Sathiskumar, Dinaharan, Murugan and Vijay30] in their study on the fabrication of surface composites report the theoretical and actual volume fractions of boron carbide (B₄C) reinforcement present in the composite by measuring the area of composites for the various volume fractions of reinforcement. Similarly, in the present study the macrograph and SEM images of the WNZ for the FSW joint specimens produced with the varying vol% fractions of SiC_NP are used to measure the actual volume fraction of SiC_NP present at the WNZ using image analyzing software (ImageJ, National Institutes of Health, Bethesda, Maryland). The area of the WNZ considered for measuring the actual volume fraction of SiC_NP for the FSW first pass and the second pass is shown in the dotted line in Table I. The area of WNZ is measured using the ImageJ software. In addition, the ImageJ analyses are carried out on the highly magnified SEM images of the WNZ for finding the actual volume fraction of SiC_NP at the WNZ and compared with WNZ area calculation. The theoretical volume fraction (V _th), actual volume fraction SiC_NP at WNZ after FSW first pass (V _fp), and actual volume fraction SiC_NP at WNZ after FSW second pass (V _sp) are computed as follows [Reference Sathiskumar, Murugan, Dinaharan and Vijay16, Reference Sathiskumar, Dinaharan, Murugan and Vijay30]:

(1)$$\eqalign{{\rm{Theoretical}}\;{\rm{volume}}\;{\rm{fraction}}\;\left( {{V_{{\rm{th}}}}} \right)\;{\rm{of}}\;{\rm{Si}}{{\rm{C}}_{{\rm{NP}}}} \cr= \left( {{{{A_{\rm{g}}}} / {{A_{\rm{P}}}}}} \right) \times 100\quad ,$$

(2)$$\eqalign{{\rm{Actual}}\;{\rm{volume}}\;{\rm{fraction}}\;{\rm{of}}\;{\rm{Si}}{{\rm{C}}_{{\rm{NP}}}}\;{\rm{at}}\;{\rm{WNZ}}\;{\rm{after}}\;\cr{\rm{FSW}}\;{\rm{first}}\;{\rm{pass}}\;\left( {{V_{{\rm{fp}}}}} \right) = \left( {{{{A_{\rm{g}}}} / {{A_{{\rm{fp}}}}}}} \right) \times 100\quad ,$$

(3)$$\eqalign{{\rm{Actual}}\;{\rm{volume}}\;{\rm{fraction}}\;{\rm{of}}\;{\rm{Si}}{{\rm{C}}_{{\rm{NP}}}}\;{\rm{at}}\;{\rm{WNZ}}\;{\rm{after}}\;\cr{\rm{FSW}}\;{\rm{second}}\;{\rm{pass}}\;\left( {{V_{{\rm{sp}}}}} \right) = \left( {{{{A_{\rm{g}}}} / {{A_{{\rm{sp}}}}}}} \right) \times 100\quad ,$$

(4)$$\eqalign{{\rm{Total}}\;{\rm{area}}\;{\rm{of}}\;{\rm{the}}\;{\rm{groove}}\;\left( {{A_{\rm{g}}}} \right) = {\rm{Total}}\;{\rm{groove}}\;{\rm{width}}\;\left( {{G_{\rm{w}}}} \right) \cr\times\,\, {\rm{Groove}}\;{\rm{depth}}\left( {{G_{\rm{d}}}} \right)\quad .$$

The projected area of the taper threaded tool pin (A _p) in the calculation is equal to 39.16 mm² which remains same irrespective of vol% fraction of SiC_NP reinforcement in each groove. The calculated V _th of SiC_NP for each experiment varies because the total area of the groove (A _g) size considered is of the different dimension in each experiment. The V _fp is obtained by measuring the area of WNZ of the MMNC after FSW first pass (A _fp). The V _sp is obtained by measuring the area of WNZ of the MMNC after FSW second pass (A _sp). The WNZ of the MMNC considered for FSW first pass, A _fp, and FSW second pass, A _sp, is shown by the dotted lines in Table I. The WNZ of the MMNC for FSW first pass was found to be ∼40 mm² and it is almost same for all the vol% fractions of SiC_NP (5, 8, 13%) reinforcement at the WNZ. The change in the vol% fraction of SiC_NP at the WNZ did not have much influence on the area of WNZ of the MMNC. Similarly, the area of WNZ of the MMNC for the FSW second pass was found to be ∼56 mm² and it is almost the same for all the vol% fractions of SiC_NP reinforcement at the WNZ. Also, the results of V _fp and V _sp obtained from the Eqs. (2) and (3) are compared with the actual volume fraction measured from the SEM images of the WNZ from both the first pass and the second pass using ImageJ analysis software. These results are found to be in good agreement with the calculation done using Eqs. (2) and (3). Figure 1(e) shows the comparison of the actual volume fraction of SiC_NP for V _th, V _fp, and V _sp. As observed in Fig. 1(e), the decreasing trend in the vol% fraction of SiC_NP from the FSW first pass to second pass is attributed to the corresponding increase in the area of WNZ of the MMNC from the FSW first pass to the second pass [Table I]. Considering the area of WNZ of MMNC [Table I] for each experimental case, the calculated V _fp after performing the FSW first pass was found to be ∼5, ∼8, ∼13 vol% fraction and almost equal to V _th. Similarly, considering the area of WNZ of MMNC after the second pass, the calculated V _sp on the same specimen was ∼3, ∼5, and ∼9 vol% fraction.

In addition, an attempt is made to find the density of the MMNC at the WNZ using the rule of mixture (ROM) and compared with experimental data of the FSW joint specimen produced after FSW second pass. The ROM considers the properties (density or YS) of the composite as the volume weighted averages of the component properties assuming the components are noninteracting during deformation [Reference Scudino, Liu, Prashanth, Bartusch, Surreddi, Murty and Eckert42, Reference Wang, Qu, Scudino, Sun, Prashanth, Louzguine-Luzgin, Chen, Zhang and Eckert49]. The ROM describes the effect of varying vol% fractions of reinforcement on the properties of MMNC at the WNZ and is calculated as follows [Reference Scudino, Liu, Prashanth, Bartusch, Surreddi, Murty and Eckert42, Reference Wang, Qu, Scudino, Sun, Prashanth, Louzguine-Luzgin, Chen, Zhang and Eckert49, Reference Markó, Prashanth, Scudino, Wang, Ellendt, Uhlenwinkel and Eckert50]:

(5)$${P_{\rm{c}}} = {V_{\rm{m}}} \cdot {P_{\rm{m}}} + {V_{\rm{f}}} \cdot {P_{\rm{f}}}\quad ,$$

where P is the property (YS or density) of the material, V is volume fraction, and c, m, and f represent composite, matrix, and reinforcement at the WNZ. Figure 1(f) shows the comparison between the theoretically calculated densities using ROM of Eq. (5) with the experimental data. The Fig. 1(f) is plotted for the FSW second pass, which ensures homogeneous dispersion of SiC_NP compared with FSW first pass. To find the experimental results, the specimens of equal dimensions are carefully extracted from the WNZ from each FSW joint specimens. In Fig. 1(f), the experimental density of the specimens with 5% SiC_NP at the WNZ shows a relative density of 99% confirming homogeneous distribution of reinforcement at the WNZ during FSW second pass. However, the increase in the vol% fraction of SiC_NP at the WNZ to 8 and 13% shows a decrease in the relative density to 98 and 97.5%, respectively. Increase in the vol% fraction of SiC_NP leads to the inhomogeneous distribution of reinforcement and porosities are formed at the WNZ. These differences in the densities are validated by density measurement shown in Fig. 1(f). The particle density at the WNZ decides the mechanical behavior of the MMNC at the WNZ. The high particle density at the WNZ results in fine grains which exerts a strong pinning effect on the boundaries, hinders the grain growth, and helps in strengthening mechanism by dislocations moving along the boundaries of the grains which is the main densification mechanism [Reference Markó, Prashanth, Scudino, Wang, Ellendt, Uhlenwinkel and Eckert50, Reference Kumar, Yadav, Perugu and Kailas52]. On the other hand, a low particle density observed in Fig. 1(f) for the FSW joint specimen with high vol% of SiC_NP (8, 13%) at the MMNC increases the porosity around the SiC_NP, and this results in interface debonding between matrix-SiC_NP leading to premature failure of the specimens [Reference Markó, Prashanth, Scudino, Wang, Ellendt, Uhlenwinkel and Eckert50].

Figures 2(a)–2(e) show the SEM images of the FSW specimen fabricated after performing the FSW first pass and having 5 vol% fraction of SiC_NP initially in the groove width of 0.4 mm at the WNZ. Figures 2(g) and 2(h) show the specimen after performing the FSW second pass on the same specimen. Figures 2(a) and 2(b) reveal the nonhomogeneous distribution of SiC_NP for FSW first pass with agglomeration of SiC_NP and defects are observed at the top of WNZ. The distribution of SiC_NP at the WNZ depends on the number of passes [Reference Byung-Wook, Don-Hyun, Yong-Hwan and Seung-Boo21, Reference Pantelis, Karakizis, Daniolos, Charitidis, Koumoulos and Dragatogiannis22]. Further, in Figs. 2(c) and 2(d), the WNZ is distinguished by the presence of alternate layers of SiC_NP-rich and SiC_NP-free region. Also, these SiC_NP-rich regions are surrounded by the SiC_NP-free regions, indicating that the material stirring is incomplete during the FSW first pass [Reference Abbasi, Abdollahzadeh, Bagheri and Omidvar17, Reference Bahrami, Dehghani and Givi18]. In addition, the higher magnification SEM image [Fig. 2(e)] reveals SiC_NP-rich region surrounded by pores at the WNZ. Thus, the inadequate stirring of the material during the FSW first pass results in porosities in the WNZ and leads to weak interface bonding between the SiC_NP and the aluminum matrix. Bodaghi and Dehghani [Reference Bodaghi and Dehghani38], and Sun and Fujii [Reference Sun and Fujii23] report similar observation in their investigations and the presence of agglomeration of SiC particles during fabrication of FSW joints using FSW one pass. Further, they report that the difference in physical properties between the SiC and the base materials is the reason for the observation of pores around the particle. Considering Figs. 2(a) and 2(b), it is observed that during the FSW first pass, the SiC_NP along with the material migrates towards RS of the WNZ, which exposes the material flow pattern during FSW and FSP [Reference Dolatkhah, Golbabaei, Givi and Molaiekiya14, Reference Tutunchilar, Haghpanahi, Givi, Asadi and Bahemmat31]. The EDS analysis shown in Fig. 2(f) confirms the presence of agglomerated nanoparticles in the SiC_NP-rich region in the RS of the WNZ. In the FSW second pass, the tool rotation applied is in the opposite direction and results in more heat input, enhanced plasticized flow of material, severe plastic deformation, and continuous dynamic recrystallization (CDRX) in the WNZ [Reference Mishra and Ma2]. In the FSW, the grain refinement at the WNZ is achieved by the combined effect of CDRX mechanism and SiC_NP reinforcement. The presence of SiC_NP pins the grain growth and acts as an additional reason for dislocation generation [Reference Kumar, Yadav, Perugu and Kailas52]. As a result, at the WNZ new fine grains are formed which are strain-free grains [Reference Mishra and Ma2, Reference Kumar, Yadav, Perugu and Kailas52]. Comparing Figs. 2(a)–2(e), 2(g), and 2(h) at the WNZ, the change in the tool rotation direction and the FSW second pass results in the homogenous distribution of SiC_NP at the WNZ. Thus, the pores observed around the SiC_NP [Fig. 2(e)] during the first pass are not seen after the second pass. Many studies [Reference Byung-Wook, Don-Hyun, Yong-Hwan and Seung-Boo21, Reference Pantelis, Karakizis, Daniolos, Charitidis, Koumoulos and Dragatogiannis22, Reference Sun and Fujii23] report similar observations of improvement in particle distribution and properties of the FSW joints with the increase in the number of FSW pass. The average grain size of the FSW specimen with MMNC at the WNZ after performing the second pass is in the range 2–3.5 µm [Fig. 2(h)]. The SEM image in Fig. 2(h) reveals that the SiC_NP is homogeneously distributed and occupies the grain boundaries along with the other precipitates. The presence of these nanoparticles placed along the grain boundaries helps in hindering the grain growth [Reference Azizieh, Kokabi and Abachi15].

Figure 2:

SEM images showing the distribution of 5 vol% fraction SiC_NP in the WNZ for FSW first pass (a) defects at the top of WNZ, (b) agglomerated SiC nanoparticles, (c) alternate layer of SiC-rich and SiC-free region, (d) magnified image of (c), (e) pores around SiC particles at the top surface of WN region, (f) EDS peaks of SiC-rich region, (g) homogeneous distribution of SiC_NP after the FSW second pass, and (h) magnified image showing grain size range 2–3.5 µm.

Figure 3 presents the SEM images of the FSW joint specimen produced after performing the FSW first pass and second pass having 8 vol% fraction of SiC_NP at the WNZ. In Figs. 3(a) and 3(b), the SiC_NP shows nonhomogeneous distribution, severe agglomeration of SiC_NP at the RS and these reinforcements near the shoulder traveling towards the RS of the weld region. These SiC_NP agglomerations at the RS and near the top regions of WNZ [Figs. 3(a) and 3(b)] are the result of part of the material being stretched towards RS due to the frictional shear force induced by the tool shoulder [Reference Tutunchilar, Haghpanahi, Givi, Asadi and Bahemmat31]. The observation is in good agreement with simulation and experimental investigation by Tutunchilar et al. [Reference Tutunchilar, Haghpanahi, Givi, Asadi and Bahemmat31]. The presence of SiC_NP-rich and SiC_NP-poor region is also observed in Fig. 3(c). The surface EDS analysis [Fig. 3(f)] shows the peaks of Si during the FSW first pass. Compared with the FSW first pass, the second pass results in more homogeneously distributed SiC_NP at the WNZ [Fig. 3(g)] with an average grain size at the WNZ in the range 3–5 µm [Fig. 3(h)]. However, even after applying for the FSW second pass, some areas at the lower portion of the WNZ are found with the agglomerated SiC_NP [Fig. 3(g)] surrounded by pores. The similar observation of agglomeration of SiC_NP at WNZ was also reported by Pantelis et al. [Reference Pantelis, Karakizis, Daniolos, Charitidis, Koumoulos and Dragatogiannis22].

Figure 3:

SEM images showing the distribution of 8 vol% fraction SiC_NP in the WNZ for FSW first pass and second pass. (a) Agglomeration of SiC towards RS, (b) BSE image of (a), (c) alternate layer of SiC-rich and SiC-free region, (d) magnified image of (c), (e) BSE images of (d), (f) EDS peaks of SiC-rich region, (g) agglomeration at the bottom of WNZ after FSW second pass, and (h) magnified image of (g) showing reduction in grain size range 3–5 µm in the WNZ.

Figures 4(a), 4(b) and 4(c), 4(d) present the SEM images of the FSW specimen fabricated after performing the FSW first pass and second pass, respectively, with 13 vol% fraction of SiC_NP at the WNZ. Comparing the SEM images of Figs. 3(a) and 4(a), it is found that the amount of agglomeration of SiC_NP also increases owing to the increase in the vol% fraction of SiC_NP and due to the insufficient stirring at the WNZ. As a result, a defect at the WNZ mainly towards the RS (marked as 1) was observed [Fig. 4(b)]. This indicates that as the amount of reinforcement particles increases, it leads to a higher flow stress for the composite materials and offers resistance to the flow of plasticized material from RS to the AS [Reference Sathiskumar, Murugan, Dinaharan and Vijay16]. Figure 4(e) documents this effect, showing EDS peaks of the SiC_NP agglomerated region marked as 1 in Fig. 4(b). The FSW second pass helps in breaking of the reinforcement by the tool and results in the homogeneous distribution due to good stirring effects. However, Fig. 4(c) illustrates the presence of two distinct regions: middle of the WNZ where no agglomeration of particles was observed, and the lower portion of the WNZ (marked as 2) indicator of a nonhomogeneous distribution of the SiC_NP with lack of stirring of the material. Figure 4(f) shows the EDS peaks of Si and confirms the presence of agglomeration of SiC_NP at the lower portion of the WNZ (marked as 2). In addition, a good consolidation of Al/SiC_NP composites at the WNZ can be seen from Fig. 1(f) for the FSW joint specimen produced with 5% SiC_NP resulting in high particle density and in the fine grains [Fig. 2(h)]. On the other hand, the higher vol% of SiC_NP (8, 13%) at the MMNC increases the porosity around the SiC_NP and results in low particle density [Fig. 1(f)] and interface debonding between matrix and SiC [Figs. 3(g) and 3(h)] leading to premature failure of the specimens [Reference Markó, Prashanth, Scudino, Wang, Ellendt, Uhlenwinkel and Eckert50].

Figure 4:

SEM images showing the distribution of vol% fraction SiC_NP in the WNZ for (a) 13% SiC for FSW first pass and (b) partially closed groove. (c) FSW second pass specimen showing agglomeration at the bottom of WNZ, (d) reduction in grain size range 3–4 µm in the WNZ, (e) spot EDS analysis showing agglomeration for the region 1 of (b), (f) surface EDS for the region 2 of (c).

Figures 5(a) and 5(b) show the SEM images of the FSW second pass specimen produced without SiC_NP (0%) and show an average grain size in the range between 6 and 8 µm. The increase in the grain size is attributed to the higher frictional heat generation during the FSW second pass accompanying with the lower traverse speed [Reference El-Rayes and El-Danaf25, Reference Khodir and Shibayanagi39] of 40 mm/min and a higher rotation speed of 1000 rpm. On the other hand, in the case of the experiment trials conducted for the fabrication of FSW specimen after the FSW second pass with the addition of SiC_NP, although the heat generation is high, the presence of SiC_NP reinforcement acts as an obstacle to the grain growth and grain boundary movement and results in decreased grain size range 2–3.5 µm. This shows that the addition of SiC_NP and heat input have opposite effects on grain size [Reference Hamdollahzadeh, Bahrami, Nikoo, Yusefi, Givi and Parvin24]. In the present study, the pinning effect by SiC_NP dominates the heat input by selected FSW parameters. The presence of SiC_NP restricts the grain growth because of pinning effect on grain boundaries and also by increasing the nucleation sites [Reference Abbasi, Abdollahzadeh, Bagheri and Omidvar17, Reference Liu, Hu, Zhao and Fujii20, Reference Byung-Wook, Don-Hyun, Yong-Hwan and Seung-Boo21, Reference Guo, Lee, Du, Bi, Tan and Wei32, Reference Barmouz, Asadi, Givi and Taherishargh40, Reference Morisada, Fujii, Nagaoka and Fukusumi41]. Figure 5(c) demonstrates the comparison in the average grain size for the FSW specimens produced for varying vol% fractions of SiC_NP and without SiC_NP at the WNZ for FSW second pass. Thus, it can be concluded from these microstructural analysis and SEM images [Figs. 2–5] that the FSW specimens prepared with the addition of varying vol% fractions of SiC_NP after the FSW second pass show considerable outcomes in a homogeneous distribution of SiC_NP at the WNZ and reduction in grain size when compared to the FSW specimen produced without SiC_NP. Figure 5(d) shows the XRD patterns of FSW dissimilar joints produced for a varying vol% fractions (5, 8, and 13%) of SiC_NP reinforcement during FSW first pass only. Since the vol% fraction of SiC_NP is less compared with the aluminum in the MMNC at the WNZ, the SiC peaks of the FSW specimens observe less intensity for 8 and 13 vol% fractions. However, no SiC intensity peaks were identified for the FSW specimens fabricated for 5 vol% fraction.

Figure 5:

(a) SEM images of the FSW specimen without SiC_NP for FSW second pass at the WNZ region, (b) higher magnified image of (a) showing grain size range 6–9 µm, (c) average grain size after performing FSW second pass on the specimen with different vol% fractions of SiC_NP (5, 8, 13%) and without SiC, (d) XRD images showing the intensity peaks in the MMNC at the WNZ for FSW first pass specimen.

The theoretical grain size at the MMNC of the WNZ is calculated by using the Zener–Holloman parameter:

(6)$${\rm{d}}z = {{4r} \over {3{v_{\rm{f}}}}}\quad ,$$

where dz is Zener limiting grain size and r and v _f are the radius and volume fraction of the reinforcing particles. Considering the Eq. (6), the grain size decreases with increasing v _f and decreasing r. In the present study, after performing the FSW second pass and having presence of the vol% fraction of SiC_NP (i.e., 3, 5, 9%) at the WNZ, the calculated theoretical dz should be 2.2 µm for 3%, 1.3 µm for 5%, and 0.7 µm which is much smaller than the actual grain size. The actual grain size is in the range 2–3.5 µm for 3%, 3–5 µm for 5% and 9%. This difference between theoretical dz and actual grain size indicates that increasing the vol% fraction of SiC_NP not all the SiC_NP contributes to hindering the grain growth but it results in agglomeration at the WNZ. However, it can be concluded from the result that FSW joint specimen after performing the FSW second pass exhibits reduced grain size in the WNZ compared with the FSW joint specimens produced without SiC_NP. From the above observations, the homogeneous distribution of SiC_NP to produce a defect-free MMNC at the WNZ depends on the three factors: (i) the number of FSW pass, (ii) the vol% fraction of SiC_NP at the WNZ, and (iii) the area of WNZ of the MMNC. In the first case, an increase in the number of FSW pass helps proper stirring of the material and powder distribution becomes homogeneous at the WNZ [Reference Dolatkhah, Golbabaei, Givi and Molaiekiya14]. The second case suggests that there is a limit for the vol% fraction of SiC_NP reinforcement to be incorporated at the WNZ to produce MMNC without any defects and with the best possible uniform distribution of SiC_NP [Reference Wang, Song, Sun and He33, Reference Slipenyuk, Kuprin, Milman, Goncharuk and Eckert34]. In the third case, an increase in the area of WNZ of the MMNC [Table I] from the FSW first pass to the second pass allows the flow of more amount of plasticized material from both the AS and RS, which occupy the WNZ composite area.

Mechanical properties

The effect of varying vol% fractions of SiC_NP (0, 5, 8, 13%) on the tensile strength of the FSW joint specimens fabricated for FSW first and second pass experiment is shown in Figs. 6(a) and 6(b), respectively, and the same is presented in Table II for UTS, YS, and % elongation for tested specimens. As can be seen in Fig. 6(a) and Table II, during the FSW first pass, lower tensile properties are observed for the FSW joint specimens with varying vol% fractions SiC_NP (5, 8, 13%) compared with the specimens without SiC_NP (0%). The FSW joint specimens without SiC_NP for the first pass have shown maximum UTS of 333 ± 8 MPa, YS of 177 ± 3 MPa, and % elongation of 6.5 ± 0.4. Further, FSW joint specimens with 5 vol% fraction of SiC_NP at the WNZ observes lower tensile properties, UTS of 273 ± 8 MPa, YS of 166 ± 4 MPa, and % elongation of 5.58 ± 0.5. Similarly, a lower UTS of 299 ± 7 MPa and 256 ± 6 MPa, YS of 170 ± 3 MPa and 158 ± 4 MPa, and % elongation of 6.48 ± 0.5 and 4.03 ± 1.5 are observed for the FSW specimen with 8 and 13 vol% fraction SiC_NP, respectively. This can be attributed to the presence of nonhomogeneous distribution and agglomeration of reinforcement SiC_NP at the WNZ during the FSW first pass as discussed in the preceding sections. Also, the measured area of WNZ of 40 mm² having accommodated the different V _fp for varying vol% fractions of SiC_NP dictates the final tensile properties of the MMNC at the WNZ. The tensile properties depend on the factors such as the amount of SiC_NP agglomeration in the WNZ, grain size, amount of dislocation and porosities, and the bonding between the reinforcement particle and the alloys [Reference Rahimian, Ehsani, Parvin and Baharvandi37, Reference Barmouz, Asadi, Givi and Taherishargh40]. The results were in good agreement with Sun and Fujii [Reference Sun and Fujii23] in their investigation on the fabrication Cu/SiC composites during first pass FSP. For the FSW first pass, an increase in the vol% fraction of SiC_NP in the WNZ from 5 to 8% results in an increase in the tensile properties of the joints, while further increasing the vol% fraction from 8 to 13% the tensile properties are decreased. These differences in tensile properties may be due to the influence of “heat input” and “pinning effect” in controlling the grain size as reported by Hamdollahzadeh et al. [Reference Hamdollahzadeh, Bahrami, Nikoo, Yusefi, Givi and Parvin24]. In the present study, selected two FSW parameters, tool rotation speed, 1000 rpm, and traverse speed, 40 mm/min, resulted in the higher heat generation at the WNZ. In the FSW specimens fabricated with the 8 vol% fraction, the “pinning effect” dominates the “heat input,” whereas the same is not true in the case of FSW specimen with 5 and 13 vol% fractions, in the former the “heat input” being dominating and the later resulting in the severe agglomeration of SiC_NP at the WNZ.

Figure 6:

Comparison of UTS, YS, and % elongation for varying % volume fractions of SiC_NP (5, 8, 13%) and without SiC_NP (0%) for the MMNC at the WNZ (a) after the FSW first pass and (b) after FSW second pass. Note that % elongation is scaled to 10. Microhardness profile for varying % volume fractions (0, 5, 8, and 13%) (c) after FSW first pass and (d) after FSW second pass.

TABLE II:

Tensile properties of FSW joint specimen fabricated with varying vol% fractions of SiC_NP (5, 8,13%) and without SiC_NP (0%) at the WNZ for FSW first pass and second pass.

Figure 6(b) and Table II present the effect of varying vol% fractions of SiC_NP (0, 5, 8, 13%) and FSW second pass on the tensile properties of FSW dissimilar joint specimens. It is observed from Fig. 6(b) that for the FSW second pass, the FSW dissimilar joint specimen fabricated without SiC_NP observes a decreasing trend with UTS of 287 ± 4 MPa, YS of 176 ± 6 MPa, and % elongation of 4.7 ± 0.2 compared with the FSW specimen produced without SiC_NP for FSW first pass [Fig. 6(a)]. This is due to the softening effect of the WNZ because of an increase in the number of passes [Reference Barmouz, Asadi, Givi and Taherishargh40]. On the other hand, compared with the FSW first pass [Fig. 6(a)], the FSW joint specimens fabricated after performing FSW second pass having 5 vol% fraction of SiC_NP at the WNZ observe an increasing trend and result in the maximum UTS of 411 ± 7 MPa, YS of 241 ± 6 MPa, and % elongation of 14.3 ± 0.2. Similarly, compared with the FSW first pass, the FSW joint specimens fabricated after performing the FSW second pass having 8 and 13% SiC_NP in the WNZ observe an increasing trend in the tensile properties. However, the FSW joint specimens with 5% SiC_NP yield higher tensile properties compared with the specimen with 8 and 13% SiC_NP. The increase in the tensile properties of the FSW joint specimens tested for varying vol% fractions of SiC_NP at the WNZ and for the FSW second pass is the result of the homogeneous distribution of SiC_NP [Reference Byung-Wook, Don-Hyun, Yong-Hwan and Seung-Boo21, Reference Hamdollahzadeh, Bahrami, Nikoo, Yusefi, Givi and Parvin24]. Thus, it can be concluded that irrespective of the varying vol% fractions of SiC_NP (5, 8, 13%) reinforcement in the WNZ and compared with the first pass, the tested tensile specimens of FSW dissimilar joint specimens produced after the FSW second pass on the same specimens show an increasing trend in the tensile properties. Also, the increase in the area of WNZ measuring 40 mm² from FSW first pass to the area of 56 mm² after performing FSW second pass contributes to having accommodated the homogeneous distribution of different vol% fractions of SiC_NP which is responsible for the higher tensile properties of the FSW joint specimen.

Figures 6(c) and 6(d) depict the microhardness profile for the FSW dissimilar joint specimens fabricated with varying vol% fractions of SiC_NP and without SiC_NP at the WNZ after the FSW first pass and second pass experiment, respectively. As shown in Fig. 6(c), after FSW first pass, the specimens exhibit a nonhomogeneous microhardness profile from the AS to RS covering the WNZ and TMAZ. As it can be observed in Fig. 6(c), the TMAZ of RS observes higher hardness for all the FSW joint specimens produced with varying vol% fractions of SiC_NP (5, 8, 13%) at the WNZ. The specimens with 13% SiC_NP for the first pass have shown the maximum hardness of 152 ± 2 at the TMAZ of RS. However, for all the specimens and first pass, the TMAZ and HAZ of AS observe a significant reduction in hardness. These nonhomogeneous distributions of hardness are due to lack of stirring and migration of SiC_NP towards RS of the TMAZ during the FSW first pass [Reference Dolatkhah, Golbabaei, Givi and Molaiekiya14, Reference Hamdollahzadeh, Bahrami, Nikoo, Yusefi, Givi and Parvin24]. The observed lower and higher hardness in some local area of the WNZ and TMAZ of the RS corresponds to the alternate layers of SiC_NP-free and SiC_NP-rich region [Figs. 2(c) and 3(c)]. As shown in Fig. 6(d), the FSW second pass coupled with the change in the tool rotation direction helps in the uniform distribution of SiC_NP and observes homogeneity in the hardness profile in the WNZ and TMAZ of both AS and RS. However, because of the high heat generation in the second pass, the HAZ of all the specimens observes a significant reduction in the hardness. In the present study, higher microvickers hardness of 150 ± 2 Hv was observed at the weld interface of WNZ for the FSW specimen produced with 13 vol% fraction of SiC_NP and after the second pass. The hardness of these specimen records an increase of 12 Hv compared with the base metal AA2024-T351 hardness of 138 Hv and remains at lower hardness by 25 Hv compared with the base metal AA707-T351 hardness of 175 Hv. The FSW dissimilar joint specimens fabricated without SiC_NP shows much lower hardness of 125 Hv and 115 Hv for FSW first pass and second pass experiment, respectively, compared with the base metal AA2024-T351 and AA7075-T651.

Strengthening mechanism

There are various potential strengthening mechanisms, which are effective in determining the strength of MMC reinforced with nano- or microparticles. The major strengthening mechanisms are (i) Orowan strengthening [Reference Dieter and Bacon36, Reference Ashby43] from dislocation bowing by reinforcing particles, (ii) strengthening from coefficient of thermal expansion (CTE) [Reference Scudino, Liu, Prashanth, Bartusch, Surreddi, Murty and Eckert42, Reference Hansen44, Reference Vogt, Zhang, Li, Bonds, Browning, Lavernia and Schoenung51], (iii) Hall–Petch strengthening (grain refinement strengthening) [Reference Dieter and Bacon36, Reference Hall45], (iv) ROM strengthening mechanism [Reference Scudino, Liu, Prashanth, Bartusch, Surreddi, Murty and Eckert42, Reference Kim46, Reference Wang, Qu, Scudino, Sun, Prashanth, Louzguine-Luzgin, Chen, Zhang and Eckert49], and (v) load bearing mechanism [Reference Kim, Sohn, Nezafati, Ferguson, Schultz, Bajestani-Gohari, Rohatgi and Cho47].

The strength from CTE mismatch mechanism, σ_the, depends on the density of dislocations (ρ_CTE) occurring due to CTE mismatch between matrix and reinforcement, ∆α [Reference Scudino, Liu, Prashanth, Bartusch, Surreddi, Murty and Eckert42, Reference Hansen44, Reference Kim, Sohn, Nezafati, Ferguson, Schultz, Bajestani-Gohari, Rohatgi and Cho47]. When the particles are in micron size, the contribution of σ_the to the YS is much larger compared with Orowan strengthening (σ_Orowan) [Reference Scudino, Liu, Prashanth, Bartusch, Surreddi, Murty and Eckert42]. However, as reported in the investigation by Kim et al. [Reference Kim, Sohn, Nezafati, Ferguson, Schultz, Bajestani-Gohari, Rohatgi and Cho47], Redsten et al. [Reference Redsten, Klier, Brown and Dunand48], and Vogt et al. [Reference Vogt, Zhang, Li, Bonds, Browning, Lavernia and Schoenung51], the strengthening due CTE mismatch is negligible and can be ignored in the composites having particles size below the range 70–80 nm. Further, they report that in the MMNC, the strengthening contribution from CTE mismatch is minor compared to Orowan-strengthening (σ_Orowan) and Hall–petch [Reference Kim, Sohn, Nezafati, Ferguson, Schultz, Bajestani-Gohari, Rohatgi and Cho47, Reference Redsten, Klier, Brown and Dunand48]. In the present study, as the SiC_NP reinforcement particles size is ∼50 nm, the strengthening from CTE mismatch is ignored.

The YS through the ROM (σ_y) depends on the YS of matrix, σ_Orowan, σ_the, and stress contribution due to strain gradient effect (σ_geo). As reported by Scudino et al. [Reference Scudino, Liu, Prashanth, Bartusch, Surreddi, Murty and Eckert42] in their investigation on the mechanical properties of Al-based MMC reinforced with Zirconium (Zr)-based glass particles produced by PM technique, the strength, σ_y, cannot be accurately measured and explained for the composites having higher volume fraction of reinforcement with relatively large size particles due to particle contiguity. In the present study, increase in the vol% fraction of SiC_NP results in the particle agglomeration. In addition, the σ_y through ROM considers the σ_the from CTE mismatch. In the present study, the particle size is ∼50 nm and the strength from CTE mismatch is ignored. This indicates that strength increase through ROM cannot be explained by considering only dislocation strengthening by σ_Orowan and σ_geo. Thus, the strengthening from ROM is not suitable in the present study.

The contribution of small vol% fraction of nanosized particles towards strength improvement by the load, ∆σ_load, in load bearing mechanism can be ignored [Reference Kim, Sohn, Nezafati, Ferguson, Schultz, Bajestani-Gohari, Rohatgi and Cho47, Reference Redsten, Klier, Brown and Dunand48], because the strength improvements in the ∆σ_load by nanoparticles are estimated in the range of 0.5–2.5% of the original YS of the matrix [Reference Kim, Sohn, Nezafati, Ferguson, Schultz, Bajestani-Gohari, Rohatgi and Cho47, Reference Redsten, Klier, Brown and Dunand48]. Considering all the above facts and nanoparticles of SiC size ∼50 nm, the strengthening from σ_Orowan and Hall–Petch is found to be suitable for the composites produced using nanoparticles as a reinforcement.

In the present study, the net change in the YS for varying vol% fractions of SiC_NP (5, 8, 13%) and after the FSW second pass is calculated as follows from Orowan–Ashby Eq. (7):

(7)$${\sigma _{{\rm{orowan}}}} = {{0.13{G_{\rm{m}}}b} \over \lambda }\ln {r \over b}\quad ,$$

where G _m is the shear modulus of the matrix, b is the burgers vector, r is the radius of reinforcement particles, and λ is the mean interparticle spacing. The λ for the varying vol% fractions of SiC_NP (5, 8, 13%) is calculated as follows from Eq. (8):

(8)$$\lambda = {d_{\rm{p}}}\left[ {{{\left( {{1 \over {2{v_{\rm{p}}}}}} \right)}^{1/3}} - 1} \right]\quad ,$$

where d _p is the diameter of SiC_NP and v _p is volume fraction of the reinforcement [Reference Dieter and Bacon36]. Thus, the calculated λ is 57.5 nm, 42.1 nm, and 28.3 nm for v _p of 5, 8, and 13 vol% fraction of SiC_NP, respectively. Accordingly, the σ_Orowan predicts a YS of 78 MPa, 106 MPa, and 158 MPa for λ of 57.5 nm, 42.1 nm, and 28.3 nm, respectively. The experimental YS is 176 ± 6 MPa for the specimen without SiC_NP, 241 ± 6 MPa for 5% SiC_NP, 185 ± 6 MPa for 8% SiC_NP, and 162 ± 4 MPa for 13% SiC_NP. It must be noted that σ_Orowan prediction in YS is based on the facts that there are no agglomeration defects of SiC_NP in the WNZ [Reference Kim, Sohn, Nezafati, Ferguson, Schultz, Bajestani-Gohari, Rohatgi and Cho47]. In the present study, such a situation of defect-free WNZ was observed for the FSW joint specimen having 5 vol% fraction of SiC_NP. Hence, the FSW joint specimen without SiC_NP yields YS of 176 ± 6 MPa and specimen with 5 vol% fraction of SiC_NP results in YS of 241 ± 6 MPa. Thus, the difference in YS between these two specimens is 65 ± 6 MPa which is in good agreement with the predicted YS of 78 MPa contribution from σ_Orowan. However, predicted YS from σ_Orowan for the specimen with 8 and 13 vol% fractions of SiC_NP is not in agreement with the experimental YS. The lower YS observed in the experimental data tested FSW joint specimens with 8 and 13% of SiC_NP is mainly due to the increase in the vol% fraction of SiC_NP which results in agglomeration of SiC_NP in some regions of the WNZ. Scudino et al. [Reference Scudino, Liu, Prashanth, Bartusch, Surreddi, Murty and Eckert42] in their study report self-consistent effective medium approximation (SCEMA) method as one of the potential strengthening mechanisms, which accurately predicts the behavior of the composites by considering the particle cracking which occurs when large volume fraction reinforcement is used. Further, they report that experimental results are in good agreement with that of calculation by SCEMA method.

On the other hand, the Hall–Petch strengthening mainly due to the grain refinement at the WNZ also contributes to the strengthening of the MMNC. Hall–Petch [Reference Dieter and Bacon36] (grain refinement strengthening) model is expressed as

(9)$$\left( {{\sigma _{\rm{o}}} = {\sigma _{\rm{i}}} + \kappa {D^{ - {1 \over 2}}}} \right)\quad ,$$

where σ_o is the yield stress which depends on the grain size diameter (D), friction stress (σ_i) resulting from resistance to the crystal lattice dislocation movement and hardening contribution parameter (κ). In the present study, FSW specimens produced with SiC_NP reinforcement for the second pass show much smaller D, and these specimens have higher tensile properties compared with the FSW specimens without SiC_NP. However, the Eq. (9) assumes that other Hall–Petch parameters, κ and σ_i, cannot be predicted experimentally and must be considered as constants during changes in the processing [Reference Kim, Sohn, Nezafati, Ferguson, Schultz, Bajestani-Gohari, Rohatgi and Cho47]. Thus, only grain size, D, must be considered for the calculation of σ_o. Thus, in the present study strength contribution from the σ_Orowan is more suitable and same is considered.

Fractography

Figure 7(a) shows the SEM image of the FSW specimen fabricated with 5 vol% fraction of SiC_NP reinforcement for FSW first pass, consisting of clusters and cracked SiC_NP, resulting in lower UTS, YS, and % elongation and the failure occurs at WNZ [Fig. 7(g)]. In Figure 7(b) for the specimen with 5 vol% fraction after the FSW second pass, the fracture took place in the HAZ of either AS or RS indicating the homogeneous distribution of SiC_NP and its good interfacial bonding with aluminum alloy matrix at the WNZ. Also, the SEM images shown in Fig. 7(b) consist of a large number of small-sized fine dimples and these features result in higher tensile properties and specimen failure exhibits ductility nature. The location of the failure of both the tensile test specimen produced with 8 vol% fraction of SiC_NP and after FSW first pass and the second pass was found at WNZ as shown Fig. 7(h). Increase in the vol% fraction results in the more clusters of the SiC_NP and segregation of the nanoparticles from the matrix [Fig. 7(c)]. On the other hand, fracture surface SEM images [Fig. 7(d)] of the tested specimen after FSW second pass exhibits a much refined and good interfacial bonding between matrix and SiC_NP evidenced with pull-out marks. The SEM fractography images for the FSW first pass [Fig. 7(e)] and second pass [Fig. 7(f)] specimen with 13 vol% fraction of SiC_NP observe a severe agglomeration and fracture location was found to be at WNZ.

Figure 7:

SEM images showing fracture features of FSW specimens for varying vol% fractions of SiC_NP 5% (a) first pass and (b) second pass; 8% (c) first pass and (d) second pass; and 13% (e) first pass and (f) second pass. The tested tensile specimen and fracture location corresponding to each volume fraction and FSW pass (g) 5 vol% fraction, (h) 8 vol% fraction, (i) 13 vol% fraction.