- Email: [email protected]
Improvement of particles distribution of in-situ 5 vol% TiB2 particulates reinforced Al-4.5Cu alloy matrix composites with ultrasonic vibration treatment
Journal of Alloys and Compounds 692 (2017) 1e9
Contents lists available at ScienceDirect
Journal of Alloys and Compounds journal homepage: http://www.elsevier.com/locate/jalcom
Improvement of particles distribution of in-situ 5 vol% TiB2 particulates reinforced Al-4.5Cu alloy matrix composites with ultrasonic vibration treatment Qi Gao, Shusen Wu, Shulin Lü*, Xinchen Xiong, Rui Du, Ping An State Key Lab of Materials Processing and Die & Mould Technology, Huazhong University of Science and Technology, 1037 Luoyu Road, Wuhan 430074, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 August 2016 Received in revised form 31 August 2016 Accepted 1 September 2016 Available online 3 September 2016
Ultrasonic vibration treatment is successfully applied in preparation of in-situ 5 vol% TiB2 particulates reinforced Al-4.5Cu alloy matrix composites with salt-metal reaction route. Formation of TiB2 phase is confirmed by X-ray diffraction analysis, and intermediate phases such as Al3Ti are not detected. Agglomerations of particles in the melt are effectively eliminated by the cavitation and acoustic streaming affects in the melt with ultrasonic vibration. TiB2 particles are uniformly distributed throughout the bulk melt after treated by ultrasonic vibration for 240 s. The tiny agglomerations formed by TiB2 particles smaller than 100 nm are also broken by ultrasonic vibration. Some TiB2 particles smaller than 400 nm are observed to dispersed near grain boundary in the matrix after solidification. The optimal improvements of yield strength and ultimate tensile strength are 114% and 50%, respectively, when composite treated by ultrasonic vibration for 240 s. © 2016 Elsevier B.V. All rights reserved.
Keywords: TiB2 Aluminum matrix composites Agglomerations Ultrasonic vibration
1. Introduction As promising materials for application in structure, aerospace, the military and transportation, particulates reinforced aluminum matrix composites (PRAMCs) attracts a lot attentions since they have outstanding combination of mechanical properties like low density, high specific strength, specific modulus, hardness and low thermal expansion coefficient [1e4]. TiB2 ceramic phase is an outstanding reinforcements in aluminum among various potential reinforcement particles like Al2O3, SiC, TiC, Si3N4, B4C, ZrB2 [5e11], since TiB2 particles have high modulus, high hardness, high melting point, good thermodynamic stability, high corrosion resistance and low density [5,12e15]. TiB2 also has no interface reaction with aluminum. As a widely used in-situ process, the salt-metal reaction route is based on the aluminothermic reaction between two kinds of potassium fluoride salts and aluminum to form in-situ TiB2 particles. The following sequences are the exothermic processes of salt-metal reaction route [5,16]:
* Corresponding author. E-mail address: [email protected] (S. Lü). http://dx.doi.org/10.1016/j.jallcom.2016.09.013 0925-8388/© 2016 Elsevier B.V. All rights reserved.
3K2 TiF6 þ 13Al / 3KAlF4 þ K3 AlF6 þ 3Al3 Ti
(1)
2KBF4 þ 3Al / 2KAlF4 þ AlB2
(2)
AlB2 þ Al3 Ti /4Al þ TiB2
(3)
And a direct reaction is also believed to form in-situ TiB2 particles [17]:
3K2 TiF6 þ 6KBF4 þ 10Al/3TiB2 þ 9KAlF4 þ K3 AlF6
(4)
Compared with external addition processes, TiB2 particles formed by salt-metal reaction route are much well bonded with aluminum matrix, thus the particle-matrix system has clearer interface and better interfacial thermodynamic stability. More important, the in-situ process has no wetting problem between TiB2 particles and molten aluminum which is the biggest difficulty in external addition processes [5]. Among various in-situ processes, salt-reaction route has much lower reaction temperature than other processes, which makes TiB2 particles are smaller and can be submicron in size. Reactions of salt-reaction route are also more moderate, thus the reaction products are much controllable [5,12,20]. But there are also some shortages of salt-reaction route. First of
2
Q. Gao et al. / Journal of Alloys and Compounds 692 (2017) 1e9
all, the mass fraction of Ti and B in K2TiF6 and KBF4 are 19.9% and 8.6%, respectively, so the utilization rate of K2TiF6 and KBF4 is much lower comparing with the Al-Ti-B in-situ system. Therefore fabricating high TiB2 particulate volume percentage aluminum matrix composites needs a very high weight ratio of salts to base metal, which may enlarge the possibility of inhibiting of subsequent reactions and form more TiB2 agglomerations. Agglomerations may harm the mechanical properties [18], but more dispersed particles bring more effective reinforcements of Orowan strengthening [19]. In our previous work, the mechanical stirring is employed to reduce TiB2 agglomerations [20]. The results show that mechanical stirring can effectively eliminate large agglomerations, but small agglomerations with size in 50e100 mm cannot be completely eliminated by mechanical stirring. And introducing mechanical stirring at a speed higher than 540 rpm may lead to the mixture of molten salts and the melt, which may cause a seriously increasing of rounded and large agglomerations instead of reducing or eliminating small agglomerations. Therefore better solution to further improve particles distribution is needed. Ultrasonic vibration is a promising technology to treat molten aluminum alloys and other light alloys [21e23]. It is a relatively environment-friendly process with low cost and uncomplicated procedure. Ultrasonic vibration treatment (UVT) can clean and degas the melt and often be used to refine intermetallics or other metallic phases [24]. It is also used to improve particulate distribution of composites [11,23]. But researches about using UVT to improve particulate distribution of high volume percentage TiB2 particles reinforced aluminum matrix composites are scarcer so far. In this work, Al-Cu alloy is chosen to be the base metal. In practical experiments, TiB2/Al-4.5Cu composites are successfully fabricated by in-situ salt-metal reaction route, and the particle volume percentage is set at 5 vol% (nominal). UVT is employed to treat the re-melt composites, and agglomerations in all size are effectively reduced by UVT. The improvements of particulate distribution and mechanical properties of the composites are discussed. The mechanism of UVT improving particulate distribution is also discussed. 2. Experiment procedures 2.1. Materials preparing and processing For preparing TiB2p/Al-4.5Cu composites, a resistance furnace was firstly employed to melt aluminum. A graphite crucible was used to contain aluminum (99.8%, wt%, the same below) ingots. Then Cu (99.9%) chips were added when aluminum melt at 700 C. Reaction salts were mixture of K2TiF6, KBF4 and Na3AlF6 (which was used to help the reactions as flux), and they were all chemically pure and thoroughly mixed. Mass of K2TiF6 and KBF4 were controlled on a Ti/B molar ratio at 1/2, and mass of Na3AlF6 was 10% of total mass of K2TiF6 and KBF4. Reaction salts were prepared at the mass could form 5 vol% (nominal) TiB2 particles. They were preheated and wrapped by aluminum foil, then gradually added into molten aluminum at 830 C and avoided to generate great fluctuation of reaction temperature. After addition of all the salts, the melt was held at 830 C for 40 min with stirring to complete the reactions [20]. After that the melt was cast at 720 C using a preheated permanent mould. The UVT was employed in this study and sketch of UVT system was shown in Fig. 1 [24]. In this study, interval resting time Tr was set at 1 s and ultrasonic time Tw was also set at 1 s in an ultrasonic viberation cycle. Vibration power was set at 2.8 kw which is the biggest power of UVT system. The frequency of ultrasound was 20 kHz. The transforming rod was made with titanium alloy with a diameter of 25 mm, and metal cup was made with stainless steel
Fig. 1. Sketch of UVT system.
with a height of 130 mm and a diameter of 70 mm. In ultrasonic vibration experiment, a resistance furnace was firstly employed to melt TiB2p/Al-4.5Cu composite ingots and crucible was a graphite crucible. Then the melt of composite was poured into the metal cup at 760 C when it was preheated up to 700 C. After the melt was cooled down to 720 C, ultrasonic vibrator was immersed into the melt below the surface at 10e15 mm and then start UVT. UVT time was set at 60, 120 and 240 s, respectively. The temperature of the melt was controlled in the range of 720e710 C during UVT. After UVT, the slag was removed and the quenching samples were obtained by quenching water with a 6 mm diameter quartz tube. Then the molten composite was immediately cast into a preheated permanent mould.
2.2. Characterization For metallographic examination, specimens of composites and base metal were cut from the ends of tensile test samples. Quenching samples and specimens were grinded and polished before etched by 0.5% HF solution. X-ray diffraction (XRD) examination was carried by a SHIMADZU XRD-7000S X-ray diffractometer with Cu Ka radiation operated at 40 kV and 30 mA. Microstructure analysis was carried by a JEOL JSM-7600F scanning electron microscope. A Tecnai G2 F30 (FEI, Holland) transmission electron microscope (TEM) was employed. A SHIMADZU AG-100KN tester was employed for tensile tests following the GB/T228.1e2010 standard, tests were proceeded at room temperature (25 C) with crosshead speed at 1 mm/min. Four samples for each specific condition were tested to obtain the average mechanical properties, such as ultimate tensile strength (UTS). Fig. 2 shown the sketch of the tensile test specimen.
Fig. 2. The drafting of tensile test specimen.
Q. Gao et al. / Journal of Alloys and Compounds 692 (2017) 1e9
3. Results and discussions 3.1. Formation of TiB2 phase in the composites The X-ray diffraction results of TiB2 particles reinforced composites with and without UVT are shown in Fig. 3. First of all, the composites are successfully fabricated by the aluminothermic reactions between molten salts mixture and molten aluminum. Fig. 3(a) shows that TiB2 phase and the regular intermetallic Al2Cu phase are clearly identified. There are no other intermediate phases peaks can be found such as the peaks of AlB2 and Al3Ti. Al3Ti is a common intermediate phase in in-situ Al-K2TiF6- KBF4 system, and it usually appears when reactions are uncompleted or the ratio of Ti/B is over 0.5 in reaction area. Thus Al3Ti phase can be eliminated by completing the reactions and controlling the ratio of Ti/B. Fig 3(b) shows the XRD patterns of composites from 39.0 to 39.5 . And every pattern is very smooth, no any peaks can be traced, which is a strong evidence proves that Al3Ti phase is not obviously detected in the composites since the main peak of Al3Ti pattern is usually found in the range of 39.1 e39.3 . The relative intensity of TiB2 phase and Al2Cu phase have no obvious change between composites with and without UVT as shown in Fig 3(a). It confirms that UVT has no undesired effects on phase composition of in-situ TiB2 particles reinforced composites. 3.2. Microstructures of the composites Fig. 4 shows the results of Back-scattered electron (BSE) observing of in-situ TiB2 particles reinforced composites with and without UVT. Above all, TiB2 particles show the tendency of distributing along grain boundary in all TiB2p/Al-4.5Cu composites. That is the result of TiB2 particles being pushed by growing solid aluminum phase at solid/liquid interface during solidification [25,26], since TiB2 cannot be fully captured by growing solid phase during solidification with a lattice disregistry over 5% to aluminum crystal. In microstructure of composite without UVT as shown in Fig. 4(a), large agglomerations with a size over 200 mm can be clearly found. These large agglomerations are dense and formed before solidification which makes them barely affected by growing solid aluminum phase. UVT is introduced to eliminate these large agglomerations which will not disappear with holding the reactions. Fig. 4(b) shows the microstructure of composite treated by
3
UVT for 60 s. The particulate distribution of composite has a dramatic improvement just by a short time UVT for 60 s. The large agglomerations are immediately eliminated by UVT. Only very few small agglomerations with a size in 50e100 mm can be found in matrix. This sort of agglomerations is formed before solidification and can penetrate grain boundary. The grains of composite are effectively refined after treated by UVT and TiB2 particles are uniformly distributed along grain boundary. Since there are still few small agglomerations in matrix, UVT for longer time is employed and the result is shown in Fig. 4(c). Microstructure of composite treated by UVT for 120 s at 100 magnify shows a similar particulate distribution with the situation of composite treated by UVT for 60 s. But small agglomerations are further refined, only very few small agglomerations under 50 mm can be found in the whole section of the specimen. The size of grains shows no obvious change between composites treated by UVT for 60 s and 120 s TiB2 particles are also uniformly distributed along grain boundary. The time of UVT is further extended and the result is shown in Fig. 4(d). It worth to note that small agglomerations under 50 mm in size are completely eliminated in composite after treated by UVT for 240 s. The uniformity of particulate distribution and the size of grains also show no obvious difference between composites treated by UVT for 120 s and 240 s at 100 magnify. BSE observation at 1000 magnify is introduced for further observing of the distribution of particles and the results are shown in Fig. 5. The tendency of particles to distribute along grain boundary is further confirmed. The bright white phase is Al2Cu, and particles are wrapped by Al2Cu phase in part of grain boundary regions since intermetallic Al2Cu phase is formed at the last stage of solidification and it wets with TiB2 particles [20]. Different to agglomerations formed before solidification, TiB2 particles are pushed by growing solid aluminum phase at solid/liquid interface and gathering at grain boundary regions during solidification. In this study, this phenomenon is identified as the gathers of TiB2 particles to distinguish with agglomerations. And the gathers of TiB2 particles in grain boundary show different sizes in composites with and without UVT. The gathers of particles are naturally larger at the intersection of grain boundary. These gathers can be as large as over 50 mm in composite without UVT, as shown in Fig. 5(a). After treated by UVT for 60s, these gathers are reduced in size to under 50 mm, as shown in Fig. 5(b). With the extension of UVT time, the gathers along grain boundary show a trend to become finer with
Fig. 3. XRD patterns of TiB2p/Al-4.5Cu composites: (a): Patterns from 20.0 to 90.0 , (b): Patterns from 39.0 to 39.5 . Patterns in both (a) and (b) are sorted as below: (1). Without UVT, (2). UVT for 60 s, (3). UVT for 120 s, (4). UVT for 240 s.
4
Q. Gao et al. / Journal of Alloys and Compounds 692 (2017) 1e9
Fig. 4. BSE microstructures of agglomerations in in-situ TiB2p/Al-4.5Cu composites: (a). Without UVT, (b). UVT for 60 s, (c). UVT for 120 s, (d). UVT for 240 s.
Fig. 5. BSE microstructures of gathers of in-situ TiB2p/Al-4.5Cu composites: (a). Without UVT, (b). UVT for 60 s, (c). UVT for 120 s, (d). UVT for 240 s.
longer UVT time. And when composite is treated by UVT for 240 s, the gathers along grain boundary are refined to under 10 mm in size, as shown in Fig. 5(d).
It is important to figure out whether TiB2 particles are distributed along grain boundary as the gathers formed by pushing of growing solid aluminum phase or smaller agglomerations existed
Q. Gao et al. / Journal of Alloys and Compounds 692 (2017) 1e9
before the solidification. Therefore quenching samples of TiB2p/Al4.5Cu composites with UVT are observed and the results are shown in Fig. 6. Quenching samples are solidified in a very rapid cooling rate, so the grains are particularly fine which makes the microstructure of quenching sample able to indicate particulate distribution in liquidoid of composite. With UVT for 60 s, small agglomerations in size near 50 mm are still found in matrix, as shown in Fig. 6(a). When UVT time extended to 120 s, small agglomerations found in quenching sample are reduced to near 20 mm in size, as shown in Fig. 6(b). These agglomerations also have the feature that penetrating grain boundary and with the shape not affected by growing solid aluminum phase, i.e., these small agglomerations are existed in matrix before solidification. This kind of small agglomerations is easily confused with normal gathers of particles at intersection of grain boundary in the casting. Fig. 6(c) shows the microstructure of quenching sample after treated by UVT for 240 s, and it clearly shows that no small agglomerations which has the feature of being formed in liquidoid of composite distributed in matrix. All particles are uniformly distributed along grain boundary and shows the shape of gathering affected by growing solid aluminum phase. It confirmed that agglomerations over 3 mm in size are effectively eliminated in the melt after treated by UVT for 240 s, and all the gathers of TiB2 particles are formed by pushing of growing solid aluminum phase during solidification. As well known, UVT has nonlinear effects on the melt being treated such as cavitation and acoustic streaming [21]. Cavitation bubbles are generated in the melt by ultrasonic vibration. The cavitation bubbles will grow and rapidly expand until they collapse. The collapse of cavitation bubbles creates high pressure shockwaves, micro-flows or injection of liquid. The pressure pulses caused by collapse can be higher than 1000 MPa and the cumulative jets can be faster than 100 m/s [24]. Meanwhile, Acoustic streaming generates turbulent whirlpools in the melt. The whirlpools caused by acoustic streaming are believed to be 5 to 10 times faster than heat convection in the melt. Thus the transfer of solute and temperature is violently accelerated throughout the bulk melt. In other words, acoustic streaming can effectively homogenize the melt. When ultrasonic vibration is introduced into in-situ TiB2p/Al4.5Cu composites, the cavitation and acoustic streaming are acting as great positive role on the improvement of TiB2 particulate distribution. At the early stage of UVT, large agglomerations (100e200 mm in size or even larger) can be directly impacted by the high pressure shock-waves generated by cavitation. The weak bonded part of large agglomerations may immediately be broken into dispersed particles by the impact of high pressure shock-waves and the well bonded part of large agglomerations may be broken into small ones (under 100 mm). Meanwhile, since the cavitation mainly occur in region under ultrasonic vibration generator of the melt, whirlpools caused by acoustic streaming not only increase the possibility of impact between large agglomerations but also enhance the chance of large agglomerations being impacted by
5
high pressure shock-waves through homogenization of the melt. Therefore large agglomerations are rapidly reduced into small agglomerations and dispersed particles. When the amount of small agglomerations in the melt is high, small agglomerations are more likely impacted by each other and brought to the region where cavitation effectively works through whirlpools. Under the repeated impacts of high pressure shock-waves, small agglomerations are broken to dispersed particles or smaller ones. Thus after large agglomerations are all rapidly broken to small agglomerations, the amount of small agglomerations and the size of them are also rapidly decreased. That explains the dramatic improvement of particulate distribution of composite treated by UVT for 60 s. After the rapid decreasing of amount of small agglomerations, density of small agglomerations in the region where cavitation effectively works keeps at a very low level with the continuous homogenization introduced by acoustic streaming. Meanwhile more dispersed particles in the melt increase the viscosity of the melt. Those make small agglomerations remained impact with each other much less frequently and the efficiency of cavitation breaking small agglomerations much lower. Thus it takes another 60 s of UVT to nearly eliminate small agglomerations near 50 mm in size, and finally eliminating smaller agglomerations after a long time UVT for 240 s. After eliminating agglomerations and homogenizing dispersed particles in the melt by cavitation and acoustic streaming, the uniformly particulate distribution with features of distributing along grain boundary and without agglomerations which are existed in liquidoid is achieved in composite treated by UVT for 240 s. The mechanism of UVT eliminates TiB2 agglomerations mentioned above can be summarized as a sketch shown in Fig. 7. The average grain size of composites with and without UVT are shown in Fig. 8. Compared with base metal, the grains of composite without UVT already have an effective refinement, as shown in Fig. 8. The promotion of heterogeneous nucleation is usually be the main reason of this grain refinement, and the growth of grains is also obstructed by TiB2 particles. The promoted heterogeneous nucleation can be mainly attributed to the little excessed intermediate phase Al3Ti as an effective heterogeneous nuclei [5] (reaction (1) and (2) have different reaction rate and may cause the formation of few Al3Ti phase. With element B, Al3Ti phase can effectively refine aluminum grains at a very low content and easily consumed by peritectic reaction: L þ Al3 Ti / a Al [27]). As long as UVT treats the composites, the grains of composites are further refined, as shown in Fig. 8. The decreasing of average grain size of composites with UVT can be explained by two sides. Firstly, more effective dispersed particles enhance the hindering on growth of grains with the elimination of large agglomerations. Secondly, the cavitation can promote the nucleation. Temperature of the bubbles decreases with the expansion of cavitation bubbles, and the equilibrium melting point increases on the effect of pressure pulses caused by collapse of bubbles [24]. These two phenomenon both increase the undercooling at certain area in the melt, and usually be used to explain the situations when UVT treats the semi-solid state
Fig. 6. BSE microstructures of quenching samples of in-situ TiB2p/Al-4.5Cu composites: (a). UVT for 60 s, (b). UVT for 120 s, (c). UVT for 240 s.
6
Q. Gao et al. / Journal of Alloys and Compounds 692 (2017) 1e9
Fig. 7. Sketch of effects of ultrasonic cavitation and acoustic streaming on distribution of particles.
Fig. 8. Average grain size of in-situ TiB2p/Al-4.5Cu composites and base metal.
metal. In this study, the temperature of the melt during UVT is higher than the liquidus. Thus nuclei created by cavitation will remelt at this high temperature. But with temperature fluctuation of liquid metal, the nuclei may partly kept into the solidification. Compared between composites treated by UVT for different time, the average grain size has barely changed. That may due to the balance between cavitation creates nuclei and nuclei re-melt makes no big difference in the number of nuclei in composites treated by UVT for different times. And although small agglomerations remain in the melt after short time UVT, the majority of TiB2 particles are dispersed as long as UVT introduced in the melt. Thus the obstruction of particles on growth of grains shows no obviously enhanced with extension of UVT time. Along with grain refinement, the gathers along grain boundary are also refined as mentioned above. Comparing composites with and without UVT, the refinement of gathers is mainly attributed to homogenization of dispersed particles in the melt and grain boundary regions enlarged with the refinement of grains. Comparing composites treated by UVT for different times, the effect of homogenization introduced by acoustic streaming cannot catch up with the massive increasing of dispersed particles caused by the effective elimination of large agglomerations. And small
agglomerations (under 50 mm in size) settled at the intersection of grain boundary are quite similar with gathers. Thus the gathers in composite treat by UVT for 60 s are larger and wider. As UVT time extended to 120 s, the amount of dispersed particles in the melt has no big increase but the effect of homogenization introduced by acoustic streaming enhanced with the extension of treatment time. Small agglomerations are also reduced and refined. Therefore the gathers along grain boundary are refined. Following the same tendency, the gathers in composite treated by UVT for 240 s have the most effective refinement of all with the longest homogenizing time and the elimination of small agglomerations as mentioned above. In our previous work, a kind of tiny TiB2 particles is found in matrix. These tiny particles are 10e100 nm in size [20]. In this work, these nanoparticles are found not only in the gaps of larger particles but also more likely to form a kind of tiny agglomerations with a size near 3 mm in composite without UVT, as shown in Fig. 9(a) and (b). TiB2 particles in nano size have much higher surface energy to the submicron ones, thus they are more likely to form agglomerations to decrease surface energy. It is worth to note that ultrasonic vibration also effectively breaks these tiny agglomerations with the high pressure shock-waves. And the dispersion of nanoparticles is achieved in composite treated by UVT for 240 s, as shown in Fig. 9(c) and (d). Fig. 10 shows the dispersed particles in matrix of composites with and without UVT. Although majority of TiB2 particles follow the tendency to distribute along grain boundary and form the gathers after solidification, some dispersed TiB2 particles still could be found at the vicinity of the gathers, with the further observation of particulate distribution along grain boundary regions by BSE as shown in Fig. 10. This phenomenon happens in all the composites fabricated in this work. These particles can be identified as particles being partly captured by the grains [28]. With the size under 400 nm, these dispersed particles are all suitable reinforcements for Orowan strengthening. Fig. 11 shows the result of TEM analysis, the pattern of particle phase matches the (0 0 1), (1 0 0) and (1 0 1) of TiB2 crystal structure, which further confirms the particles are TiB2 phase with the evidence of X-ray diffraction patterns. The high resolution result of TiB2 particle also shows that TiB2 particles are well bonded with a-Al matrix, the particle-matrix interface are clean and no interface reaction products or other inclusions can be found.
Q. Gao et al. / Journal of Alloys and Compounds 692 (2017) 1e9
7
Fig. 9. Dispersion of tiny agglomerations: (a) Location of tiny agglomerations in composite without UVT, (b) Tiny agglomerations A, (c) Location of Dispersed nanoparticles in composite treated by UVT for 240 s (d) Dispersed nanoparticles in zone C.
3.3. Mechanical properties of the composites The as-cast mechanical properties of base metal and TiB2
particles reinforced composites with and without UVT are shown in Table 1. The yield strength is effectively improved by reinforcement of TiB2 particles comparing composites with base metal. The
Fig. 10. BSE microstructures of dispersed particles of in-situ TiB2p/Al-4.5Cu composites: (a). Without UVT, (b). UVT for 60 s, (c). UVT for 120 s, (d). UVT for 240 s.
8
Q. Gao et al. / Journal of Alloys and Compounds 692 (2017) 1e9
Fig. 11. TEM result of TiB2 particles.
integrated effects of Orowan strengthening, load-bearing strengthening, CTE (Coefficient of Thermal Expansion) mismatch strengthening and grain refinement is believed to be the reasons of improvement of yield strength. According to Orowan strengthening mechanism, dispersed TiB2 particles under 400 nm in size with high elastic modulus can hinder the dislocation movement and form dislocation line looped around particles. Then subsequent dislocation movement is hindered by the interaction between dislocation loops of adjacent TiB2 particles and yield strength of composites is improved [5,19,20]. The well bonded particles with matrix can directly strengthen composites with shearing of load, and increasing yield strength according to load-bearing strengthening [29]. After solidification, the CTE mismatch between aluminum matrix and TiB2 particles generates the residual plastic strain. Then dislocations are generated by residual plastic strain at interface of TiB2 particles and matrix to accommodate CTE mismatch and strengthen composites [30]. The Hall-Petch relationship explains that decreasing of grain size increases yield strength [5,20]. Therefore yield strength of composite without UVT is improved by 32.8% to base metal with presence of TiB2 particles and its dispersed part, the refinement of average grain size from 216.1 mm to 46.0 mm also contributes, as shown in Fig. 8. Composites with and without UVT have no different in content of TiB2 particles, so load-bearing strengthening is not suitable to explain the improvement of yield strength after UVT. After UVT, the elimination of large agglomerations brings much more dispersed particles distributed in matrix and effectively enhance Orowan strengthening. The grains of composites are further refined and grain boundary regions are enlarged. With enlarged grain boundary and uniform distribution of particles, CTE mismatch strengthening is also enhanced. Therefore compared with composite without UVT, yield strength of composites with UVT is improved by 35.3%, 43.5% and 60% at 60, 120 and 240 s of treatment time, respectively. With similar average grain size, the increasing of yield strength with the extension of UVT time is mainly attribute to Orowan strengthening and CTE mismatch strengthening. Since small agglomerations are
Table 1 As-cast mechanical properties of in-situ TiB2p/Al-4.5Cu composites and base metal. Specimens
UVT time (s)
Yield strength (MPa)
UTS (MPa)
Elongation (%)
Al-4.5Cu AMC AMC AMC AMC
0 0 60 120 240
64 85 115 122 137
163 221 237 238 245
12.50 8.05 5.50 5.85 6.15
gradually eliminated with UVT and further brings more dispersed particles. And distribution of particles gets more uniform when UVT time is longer. Thus Orowan strengthening, CTE mismatch strengthening are enhanced with longer UVT time, and composite treated by UVT for 240 s has the highest yield strength. When explaining the improvement of UST of composites, grain boundary strengthening is assumed to be the main reason. In the gathers along grain boundary, distance of adjacent particles becomes very short. These particles may not contribute to Orowan strengthening since dislocation loops may not successfully formed and small number of particles are too big. But they can still hinder dislocation movement by their high elastic modulus and high hardness. And CTE mismatch strengthening is still working on particles in gathers. Dislocations are hindered and assembled at grain boundary which will strengthen composites. Thus UTS of composite without UVT is improved by 35.6% to base metal. When UVT is introduced to composites, the elimination of large agglomerations and further refining of grains enhance grain boundary strengthening, since much more TiB2 particles uniformly distributed in the enlarged grain boundary regions. Therefore UTS of composites treated by UVT are higher than untreated composite. But these improvements are not very effective according to Table 1. It may due to slags remained in the melt are also broken and homogenized by UVT and makes them more hardly to removed or floated to the riser, and UVT also agitates the oxidation and brings more oxide inclusions into the melt. But the small agglomerations are gradually eliminated and TiB2 particles are distributed more uniformly with the extension of UVT time, thus UTS of composites have small improvements. The deformation of grains is hindered by vast TiB2 particles distributed along grain boundary and the ductility is harmed, thus the composites have lower elongation than base metal. Although UVT eliminates large agglomerations and homogenizes TiB2 particles in the melt, the inclusions shows more damages on ductility, and elongation of composites treated by UVT is lower than composite without UVT. But the ductility is slightly improved with longer UVT time, and this trend is attribute to the more uniformly distributed TiB2 particles with gradually elimination of small agglomerations. 4. Conclusions Through salt-metal reaction route, TiB2 particulates reinforced Al-4.5Cu alloy composites are successfully fabricated. Intermediate phases like AlB2 and Al3Ti are effectively avoided by controlling the composition of reactant and completion of reactions. The grains are effectively refined. Ultrasonic vibration treatments are successfully introduced into composites fabrication. The particulate distribution is effectively improved. Large agglomerations in the melt are eliminated at the very early stage of ultrasonic vibration treatment. With extension of ultrasonic vibration treatment time, small agglomerations are also eliminated. TiB2 particles are uniformly distributed along grain boundary after ultrasonic vibration treatment for 240s throughout the bulk matrix. The grains are further refined. The gathers of TiB2 particles which are formed by pushing at solid/liquid interface during solidification in grain boundary regions, are refined by ultrasonic vibration treatment. With the increasing of ultrasonic vibration treatment time, the gathers are more refined. The break of tiny agglomerations formed by TiB2 particles under 100 nm in size is also observed after ultrasonic vibration treatment for 240s. Dispersed TiB2 particles smaller than 400 nm are found nearby the gathers in grain boundary regions. Ultrasonic vibration treatments greatly improve as-cast
Q. Gao et al. / Journal of Alloys and Compounds 692 (2017) 1e9
mechanical properties of the composites. After ultrasonic vibration treatment for 240s, yield strength of composite is improved by 114% and 61% comparing with base metal and untreated composite, respectively. Acknowledgement This work was financially supported by the National Natural Science Foundation of China (No. 51574129), and by the Fundamental Research Funds for the Central Universities, HUST (No. 2014QNRC003, No. 2015QN010). The authors would also like to express their appreciation to Analytical and Testing Center, HUST. References [1] T.W. Clyne, P.J. Withers, An Introduction to Metal Matrix Composites, Cambridge University Press, 1993. [2] K.U. Kainer, Metal Matrix Composites, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, 2006. [3] N. Chawla, K.K. Chawla, Metal Matrix Composites, Springer ScienceþBusiness Media, New York, 2013. [4] D.D.L. Chung, Composite Materials, Springer-Verlag London Limited, 2010. [5] L. Lu, M.O. Lai, F.L. Chen, Al-4wt% Cu composite reinforced with in-situ TiB2 particles, Acta. Mater. 45 (1997) 4297e4309. [6] J.Y. Xua, B.L. Zou, S.Y. Tao, M.X. Zhang, X.Q. Cao, Fabrication and properties of Al2O3-TiB2-TiC/Al metal matrix composite coatings by atmospheric plasma spraying of SHS powders, J. Alloys. Compd. 672 (2016) 251e259. [7] R.H. Dong, W.S. Yang, Z.H. Yu, P. Wu, M. Hussain, L.T. Jiang, G.H. Wu, Aging behavior of 6061Al matrix composite reinforced with high content SiC nanowires, J. Alloys. Compd. 649 (2015) 1037e1042. [8] B. Ya, B.W. Zhou, H.S. Yang, B.K. Huang, F. Jia, X.G. Zhang, Microstructure and mechanical properties of in situ casting TiC/Ti6Al4V composites through adding multi-walled carbon nanotubes, J. Alloys. Compd. 637 (2015) 456e460. [9] C.S. Ramesh, R. Keshavamurthy, Slurry erosive wear behavior of Ni-P coated Si3N4 reinforced Al6061 composites, Mater. Des. 32 (2011) 1833e1843. [10] K. Kalaiselvan, N. Murugan, S. Parameswaran, Production and characterization of AA6061-B4C stir cast composite, Mater. Des. 32 (2011) 4004e4009. [11] X.Z. Kai, K.L. Tian, C.M. Wang, L. Jiao, G. Chen, Y.T. Zhao, Effects of ultrasonic vibration on the microstructure and tensile properties of the nano ZrB2/ 2024Al composites synthesized by direct melt reaction, J. Alloys. Compd. 668 (2016) 121e127. [12] K.L. Tee, L. Lu, M.O. Lai, Synthesis of in situ Al-TiB2 composites using stir cast route, Compos. Struct. 47 (1999) 589e593. [13] K.L. Tee, L. Lu, M.O. Lai, Improvement in mechanical properties of in-situ AlTiB2 composite by incorporation of carbon, Mater. Sci. Eng. A 339 (2003)
9
227e231. [14] L. Lu, M.O. Lai, Y. Su, H.L. Teo, C.F. Feng, In situ TiB2 reinforced Al alloy composites, Scr. Mater. 45 (2001) 1017e1023. [15] L. Lu, M.O. Lai, F.L. Chen, In situ preparation of TiB2 reinforced Al base composite, Adv. Compos. Mater. 6 (1997) 299e308. [16] M. Emamy, M. Mahta, J. Rasizadeh, Formation of TiB2 particles during dissolution of TiAl3 in Al-TiB2 metal matrix composite using an in situ technique, Compos. Sci. Technol. 66 (2006) 1063e1066. [17] Z. Liu, M. Rakita, W. Xu, X. Wang, Q. Han, Ultrasound assisted salts-metal reaction for synthesizing TiB2 particles at low temperature, Chem. Eng. J. 263 (2015) 317e324. [18] M.K. Akbari, H.R. Baharvandi, K. Shirvanimoghaddam, Tensile and fracture behavior of nano/micro TiB2 particle reinforced casting A356 aluminum alloy composites, Mater. Des. 26 (2015) 150e161. [19] Z. Zhang, D.L. Chen, Consideration of Orowan strengthening effect in particulate-reinforced metal matrix nanocomposites: a model for predicting their yield strength, Scr. Mater. 54 (2006) 1321e1326. [20] Q. Gao, S.S. Wu, S.L. LÜ, X.C. Duan, P. An, Preparation of in-situ 5 vol% TiB2 particulate reinforced Al-4.5Cu alloy matrix composites assisted by improved mechanical stirring process, Mater. Des. 94 (2016) 79e86. [21] S.L. LÜ, S.S. Wu, C. Lin, Z.Q. Hu, P. An, Preparation and rheocasting of semisolid slurry of 5083 Al alloy with indirect ultrasonic vibration process, Mater. Sci. Eng. A 528 (2011) 8635e8640. [22] Z. Hu, X. Li, H. Yan, X.Q. Wu, H. Qun, J.W. Lin, Effects of ultrasonic vibration on microstructure evolution and elevated-temperature mechanical properties of hot-extruded Mg-6Al-0.8Zn-2.0Sm wrought magnesium alloys, J. Alloys. Compd. 685 (2016) 58e64. [23] G.I. Eskin, D.G. Eskin, Production of natural and synthesized aluminum-based composite materials with the aid of ultrasonic (cavitation) treatment of the melt, Ultrason. Sonochem. 10 (2003) 297e301. [24] C. Lin, S.S. Wu, S.L. LÜ, P. An, Effects of ultrasonic vibration and manganese on microstructure and mechanical properties of hypereutectic Al-Si alloys with 2%Fe, Intermetallics 32 (2013) 176e183. [25] S.S. Wu, H. Nakae, Nucleation effect of alumina in Al-Si/Al2O3 composites, J. Mater. Sci. Lett. 18 (1999) 321e323. [26] S.S. Wu, H. Nakae, T. Kanno, Y. You, Zone-melted unidirectional solidification of particulate dispersed composites, J. Mater. Sci. 36 (2001) 225e229. [27] B.S. Murty, S.A. Kori, M. Chakraborty, Grain refinement of aliminum and its alloys by heterogeneous nucleation and alloying, Int. Mater. Rev. 47 (2002) 3e29. [28] P.L. Schaffer, D.N. Miller, A.K. Dahle, Crystallography of engulfed and pushed TiB2 particles in aliminum, Scr. Mater. 57 (2007) 1129e1132. [29] J. Nampoothiri, R.S. Harini, S.K. Nayak, B. Raj, K.R. Ravi, Post in-situ ultrasonic treatment for generation of Al-4.4Cu/TiB2 nanocomposite: a route to enhance the strength of metal matrix nanocomposites, J. Alloys. Compd. 683 (2016) 370e378. [30] M.L. Wang, D. Chen, Z. Chen, Y. Wu, F.G. Wang, N.H. Ma, H.W. Wang, Mechanical properties of in-situ TiB2/A356 composites, Mater. Sci. Eng. A 590 (2014) 246e254.
Contents lists available at ScienceDirect
Journal of Alloys and Compounds journal homepage: http://www.elsevier.com/locate/jalcom
Improvement of particles distribution of in-situ 5 vol% TiB2 particulates reinforced Al-4.5Cu alloy matrix composites with ultrasonic vibration treatment Qi Gao, Shusen Wu, Shulin Lü*, Xinchen Xiong, Rui Du, Ping An State Key Lab of Materials Processing and Die & Mould Technology, Huazhong University of Science and Technology, 1037 Luoyu Road, Wuhan 430074, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 August 2016 Received in revised form 31 August 2016 Accepted 1 September 2016 Available online 3 September 2016
Ultrasonic vibration treatment is successfully applied in preparation of in-situ 5 vol% TiB2 particulates reinforced Al-4.5Cu alloy matrix composites with salt-metal reaction route. Formation of TiB2 phase is confirmed by X-ray diffraction analysis, and intermediate phases such as Al3Ti are not detected. Agglomerations of particles in the melt are effectively eliminated by the cavitation and acoustic streaming affects in the melt with ultrasonic vibration. TiB2 particles are uniformly distributed throughout the bulk melt after treated by ultrasonic vibration for 240 s. The tiny agglomerations formed by TiB2 particles smaller than 100 nm are also broken by ultrasonic vibration. Some TiB2 particles smaller than 400 nm are observed to dispersed near grain boundary in the matrix after solidification. The optimal improvements of yield strength and ultimate tensile strength are 114% and 50%, respectively, when composite treated by ultrasonic vibration for 240 s. © 2016 Elsevier B.V. All rights reserved.
Keywords: TiB2 Aluminum matrix composites Agglomerations Ultrasonic vibration
1. Introduction As promising materials for application in structure, aerospace, the military and transportation, particulates reinforced aluminum matrix composites (PRAMCs) attracts a lot attentions since they have outstanding combination of mechanical properties like low density, high specific strength, specific modulus, hardness and low thermal expansion coefficient [1e4]. TiB2 ceramic phase is an outstanding reinforcements in aluminum among various potential reinforcement particles like Al2O3, SiC, TiC, Si3N4, B4C, ZrB2 [5e11], since TiB2 particles have high modulus, high hardness, high melting point, good thermodynamic stability, high corrosion resistance and low density [5,12e15]. TiB2 also has no interface reaction with aluminum. As a widely used in-situ process, the salt-metal reaction route is based on the aluminothermic reaction between two kinds of potassium fluoride salts and aluminum to form in-situ TiB2 particles. The following sequences are the exothermic processes of salt-metal reaction route [5,16]:
* Corresponding author. E-mail address: [email protected] (S. Lü). http://dx.doi.org/10.1016/j.jallcom.2016.09.013 0925-8388/© 2016 Elsevier B.V. All rights reserved.
3K2 TiF6 þ 13Al / 3KAlF4 þ K3 AlF6 þ 3Al3 Ti
(1)
2KBF4 þ 3Al / 2KAlF4 þ AlB2
(2)
AlB2 þ Al3 Ti /4Al þ TiB2
(3)
And a direct reaction is also believed to form in-situ TiB2 particles [17]:
3K2 TiF6 þ 6KBF4 þ 10Al/3TiB2 þ 9KAlF4 þ K3 AlF6
(4)
Compared with external addition processes, TiB2 particles formed by salt-metal reaction route are much well bonded with aluminum matrix, thus the particle-matrix system has clearer interface and better interfacial thermodynamic stability. More important, the in-situ process has no wetting problem between TiB2 particles and molten aluminum which is the biggest difficulty in external addition processes [5]. Among various in-situ processes, salt-reaction route has much lower reaction temperature than other processes, which makes TiB2 particles are smaller and can be submicron in size. Reactions of salt-reaction route are also more moderate, thus the reaction products are much controllable [5,12,20]. But there are also some shortages of salt-reaction route. First of
2
Q. Gao et al. / Journal of Alloys and Compounds 692 (2017) 1e9
all, the mass fraction of Ti and B in K2TiF6 and KBF4 are 19.9% and 8.6%, respectively, so the utilization rate of K2TiF6 and KBF4 is much lower comparing with the Al-Ti-B in-situ system. Therefore fabricating high TiB2 particulate volume percentage aluminum matrix composites needs a very high weight ratio of salts to base metal, which may enlarge the possibility of inhibiting of subsequent reactions and form more TiB2 agglomerations. Agglomerations may harm the mechanical properties [18], but more dispersed particles bring more effective reinforcements of Orowan strengthening [19]. In our previous work, the mechanical stirring is employed to reduce TiB2 agglomerations [20]. The results show that mechanical stirring can effectively eliminate large agglomerations, but small agglomerations with size in 50e100 mm cannot be completely eliminated by mechanical stirring. And introducing mechanical stirring at a speed higher than 540 rpm may lead to the mixture of molten salts and the melt, which may cause a seriously increasing of rounded and large agglomerations instead of reducing or eliminating small agglomerations. Therefore better solution to further improve particles distribution is needed. Ultrasonic vibration is a promising technology to treat molten aluminum alloys and other light alloys [21e23]. It is a relatively environment-friendly process with low cost and uncomplicated procedure. Ultrasonic vibration treatment (UVT) can clean and degas the melt and often be used to refine intermetallics or other metallic phases [24]. It is also used to improve particulate distribution of composites [11,23]. But researches about using UVT to improve particulate distribution of high volume percentage TiB2 particles reinforced aluminum matrix composites are scarcer so far. In this work, Al-Cu alloy is chosen to be the base metal. In practical experiments, TiB2/Al-4.5Cu composites are successfully fabricated by in-situ salt-metal reaction route, and the particle volume percentage is set at 5 vol% (nominal). UVT is employed to treat the re-melt composites, and agglomerations in all size are effectively reduced by UVT. The improvements of particulate distribution and mechanical properties of the composites are discussed. The mechanism of UVT improving particulate distribution is also discussed. 2. Experiment procedures 2.1. Materials preparing and processing For preparing TiB2p/Al-4.5Cu composites, a resistance furnace was firstly employed to melt aluminum. A graphite crucible was used to contain aluminum (99.8%, wt%, the same below) ingots. Then Cu (99.9%) chips were added when aluminum melt at 700 C. Reaction salts were mixture of K2TiF6, KBF4 and Na3AlF6 (which was used to help the reactions as flux), and they were all chemically pure and thoroughly mixed. Mass of K2TiF6 and KBF4 were controlled on a Ti/B molar ratio at 1/2, and mass of Na3AlF6 was 10% of total mass of K2TiF6 and KBF4. Reaction salts were prepared at the mass could form 5 vol% (nominal) TiB2 particles. They were preheated and wrapped by aluminum foil, then gradually added into molten aluminum at 830 C and avoided to generate great fluctuation of reaction temperature. After addition of all the salts, the melt was held at 830 C for 40 min with stirring to complete the reactions [20]. After that the melt was cast at 720 C using a preheated permanent mould. The UVT was employed in this study and sketch of UVT system was shown in Fig. 1 [24]. In this study, interval resting time Tr was set at 1 s and ultrasonic time Tw was also set at 1 s in an ultrasonic viberation cycle. Vibration power was set at 2.8 kw which is the biggest power of UVT system. The frequency of ultrasound was 20 kHz. The transforming rod was made with titanium alloy with a diameter of 25 mm, and metal cup was made with stainless steel
Fig. 1. Sketch of UVT system.
with a height of 130 mm and a diameter of 70 mm. In ultrasonic vibration experiment, a resistance furnace was firstly employed to melt TiB2p/Al-4.5Cu composite ingots and crucible was a graphite crucible. Then the melt of composite was poured into the metal cup at 760 C when it was preheated up to 700 C. After the melt was cooled down to 720 C, ultrasonic vibrator was immersed into the melt below the surface at 10e15 mm and then start UVT. UVT time was set at 60, 120 and 240 s, respectively. The temperature of the melt was controlled in the range of 720e710 C during UVT. After UVT, the slag was removed and the quenching samples were obtained by quenching water with a 6 mm diameter quartz tube. Then the molten composite was immediately cast into a preheated permanent mould.
2.2. Characterization For metallographic examination, specimens of composites and base metal were cut from the ends of tensile test samples. Quenching samples and specimens were grinded and polished before etched by 0.5% HF solution. X-ray diffraction (XRD) examination was carried by a SHIMADZU XRD-7000S X-ray diffractometer with Cu Ka radiation operated at 40 kV and 30 mA. Microstructure analysis was carried by a JEOL JSM-7600F scanning electron microscope. A Tecnai G2 F30 (FEI, Holland) transmission electron microscope (TEM) was employed. A SHIMADZU AG-100KN tester was employed for tensile tests following the GB/T228.1e2010 standard, tests were proceeded at room temperature (25 C) with crosshead speed at 1 mm/min. Four samples for each specific condition were tested to obtain the average mechanical properties, such as ultimate tensile strength (UTS). Fig. 2 shown the sketch of the tensile test specimen.
Fig. 2. The drafting of tensile test specimen.
Q. Gao et al. / Journal of Alloys and Compounds 692 (2017) 1e9
3. Results and discussions 3.1. Formation of TiB2 phase in the composites The X-ray diffraction results of TiB2 particles reinforced composites with and without UVT are shown in Fig. 3. First of all, the composites are successfully fabricated by the aluminothermic reactions between molten salts mixture and molten aluminum. Fig. 3(a) shows that TiB2 phase and the regular intermetallic Al2Cu phase are clearly identified. There are no other intermediate phases peaks can be found such as the peaks of AlB2 and Al3Ti. Al3Ti is a common intermediate phase in in-situ Al-K2TiF6- KBF4 system, and it usually appears when reactions are uncompleted or the ratio of Ti/B is over 0.5 in reaction area. Thus Al3Ti phase can be eliminated by completing the reactions and controlling the ratio of Ti/B. Fig 3(b) shows the XRD patterns of composites from 39.0 to 39.5 . And every pattern is very smooth, no any peaks can be traced, which is a strong evidence proves that Al3Ti phase is not obviously detected in the composites since the main peak of Al3Ti pattern is usually found in the range of 39.1 e39.3 . The relative intensity of TiB2 phase and Al2Cu phase have no obvious change between composites with and without UVT as shown in Fig 3(a). It confirms that UVT has no undesired effects on phase composition of in-situ TiB2 particles reinforced composites. 3.2. Microstructures of the composites Fig. 4 shows the results of Back-scattered electron (BSE) observing of in-situ TiB2 particles reinforced composites with and without UVT. Above all, TiB2 particles show the tendency of distributing along grain boundary in all TiB2p/Al-4.5Cu composites. That is the result of TiB2 particles being pushed by growing solid aluminum phase at solid/liquid interface during solidification [25,26], since TiB2 cannot be fully captured by growing solid phase during solidification with a lattice disregistry over 5% to aluminum crystal. In microstructure of composite without UVT as shown in Fig. 4(a), large agglomerations with a size over 200 mm can be clearly found. These large agglomerations are dense and formed before solidification which makes them barely affected by growing solid aluminum phase. UVT is introduced to eliminate these large agglomerations which will not disappear with holding the reactions. Fig. 4(b) shows the microstructure of composite treated by
3
UVT for 60 s. The particulate distribution of composite has a dramatic improvement just by a short time UVT for 60 s. The large agglomerations are immediately eliminated by UVT. Only very few small agglomerations with a size in 50e100 mm can be found in matrix. This sort of agglomerations is formed before solidification and can penetrate grain boundary. The grains of composite are effectively refined after treated by UVT and TiB2 particles are uniformly distributed along grain boundary. Since there are still few small agglomerations in matrix, UVT for longer time is employed and the result is shown in Fig. 4(c). Microstructure of composite treated by UVT for 120 s at 100 magnify shows a similar particulate distribution with the situation of composite treated by UVT for 60 s. But small agglomerations are further refined, only very few small agglomerations under 50 mm can be found in the whole section of the specimen. The size of grains shows no obvious change between composites treated by UVT for 60 s and 120 s TiB2 particles are also uniformly distributed along grain boundary. The time of UVT is further extended and the result is shown in Fig. 4(d). It worth to note that small agglomerations under 50 mm in size are completely eliminated in composite after treated by UVT for 240 s. The uniformity of particulate distribution and the size of grains also show no obvious difference between composites treated by UVT for 120 s and 240 s at 100 magnify. BSE observation at 1000 magnify is introduced for further observing of the distribution of particles and the results are shown in Fig. 5. The tendency of particles to distribute along grain boundary is further confirmed. The bright white phase is Al2Cu, and particles are wrapped by Al2Cu phase in part of grain boundary regions since intermetallic Al2Cu phase is formed at the last stage of solidification and it wets with TiB2 particles [20]. Different to agglomerations formed before solidification, TiB2 particles are pushed by growing solid aluminum phase at solid/liquid interface and gathering at grain boundary regions during solidification. In this study, this phenomenon is identified as the gathers of TiB2 particles to distinguish with agglomerations. And the gathers of TiB2 particles in grain boundary show different sizes in composites with and without UVT. The gathers of particles are naturally larger at the intersection of grain boundary. These gathers can be as large as over 50 mm in composite without UVT, as shown in Fig. 5(a). After treated by UVT for 60s, these gathers are reduced in size to under 50 mm, as shown in Fig. 5(b). With the extension of UVT time, the gathers along grain boundary show a trend to become finer with
Fig. 3. XRD patterns of TiB2p/Al-4.5Cu composites: (a): Patterns from 20.0 to 90.0 , (b): Patterns from 39.0 to 39.5 . Patterns in both (a) and (b) are sorted as below: (1). Without UVT, (2). UVT for 60 s, (3). UVT for 120 s, (4). UVT for 240 s.
4
Q. Gao et al. / Journal of Alloys and Compounds 692 (2017) 1e9
Fig. 4. BSE microstructures of agglomerations in in-situ TiB2p/Al-4.5Cu composites: (a). Without UVT, (b). UVT for 60 s, (c). UVT for 120 s, (d). UVT for 240 s.
Fig. 5. BSE microstructures of gathers of in-situ TiB2p/Al-4.5Cu composites: (a). Without UVT, (b). UVT for 60 s, (c). UVT for 120 s, (d). UVT for 240 s.
longer UVT time. And when composite is treated by UVT for 240 s, the gathers along grain boundary are refined to under 10 mm in size, as shown in Fig. 5(d).
It is important to figure out whether TiB2 particles are distributed along grain boundary as the gathers formed by pushing of growing solid aluminum phase or smaller agglomerations existed
Q. Gao et al. / Journal of Alloys and Compounds 692 (2017) 1e9
before the solidification. Therefore quenching samples of TiB2p/Al4.5Cu composites with UVT are observed and the results are shown in Fig. 6. Quenching samples are solidified in a very rapid cooling rate, so the grains are particularly fine which makes the microstructure of quenching sample able to indicate particulate distribution in liquidoid of composite. With UVT for 60 s, small agglomerations in size near 50 mm are still found in matrix, as shown in Fig. 6(a). When UVT time extended to 120 s, small agglomerations found in quenching sample are reduced to near 20 mm in size, as shown in Fig. 6(b). These agglomerations also have the feature that penetrating grain boundary and with the shape not affected by growing solid aluminum phase, i.e., these small agglomerations are existed in matrix before solidification. This kind of small agglomerations is easily confused with normal gathers of particles at intersection of grain boundary in the casting. Fig. 6(c) shows the microstructure of quenching sample after treated by UVT for 240 s, and it clearly shows that no small agglomerations which has the feature of being formed in liquidoid of composite distributed in matrix. All particles are uniformly distributed along grain boundary and shows the shape of gathering affected by growing solid aluminum phase. It confirmed that agglomerations over 3 mm in size are effectively eliminated in the melt after treated by UVT for 240 s, and all the gathers of TiB2 particles are formed by pushing of growing solid aluminum phase during solidification. As well known, UVT has nonlinear effects on the melt being treated such as cavitation and acoustic streaming [21]. Cavitation bubbles are generated in the melt by ultrasonic vibration. The cavitation bubbles will grow and rapidly expand until they collapse. The collapse of cavitation bubbles creates high pressure shockwaves, micro-flows or injection of liquid. The pressure pulses caused by collapse can be higher than 1000 MPa and the cumulative jets can be faster than 100 m/s [24]. Meanwhile, Acoustic streaming generates turbulent whirlpools in the melt. The whirlpools caused by acoustic streaming are believed to be 5 to 10 times faster than heat convection in the melt. Thus the transfer of solute and temperature is violently accelerated throughout the bulk melt. In other words, acoustic streaming can effectively homogenize the melt. When ultrasonic vibration is introduced into in-situ TiB2p/Al4.5Cu composites, the cavitation and acoustic streaming are acting as great positive role on the improvement of TiB2 particulate distribution. At the early stage of UVT, large agglomerations (100e200 mm in size or even larger) can be directly impacted by the high pressure shock-waves generated by cavitation. The weak bonded part of large agglomerations may immediately be broken into dispersed particles by the impact of high pressure shock-waves and the well bonded part of large agglomerations may be broken into small ones (under 100 mm). Meanwhile, since the cavitation mainly occur in region under ultrasonic vibration generator of the melt, whirlpools caused by acoustic streaming not only increase the possibility of impact between large agglomerations but also enhance the chance of large agglomerations being impacted by
5
high pressure shock-waves through homogenization of the melt. Therefore large agglomerations are rapidly reduced into small agglomerations and dispersed particles. When the amount of small agglomerations in the melt is high, small agglomerations are more likely impacted by each other and brought to the region where cavitation effectively works through whirlpools. Under the repeated impacts of high pressure shock-waves, small agglomerations are broken to dispersed particles or smaller ones. Thus after large agglomerations are all rapidly broken to small agglomerations, the amount of small agglomerations and the size of them are also rapidly decreased. That explains the dramatic improvement of particulate distribution of composite treated by UVT for 60 s. After the rapid decreasing of amount of small agglomerations, density of small agglomerations in the region where cavitation effectively works keeps at a very low level with the continuous homogenization introduced by acoustic streaming. Meanwhile more dispersed particles in the melt increase the viscosity of the melt. Those make small agglomerations remained impact with each other much less frequently and the efficiency of cavitation breaking small agglomerations much lower. Thus it takes another 60 s of UVT to nearly eliminate small agglomerations near 50 mm in size, and finally eliminating smaller agglomerations after a long time UVT for 240 s. After eliminating agglomerations and homogenizing dispersed particles in the melt by cavitation and acoustic streaming, the uniformly particulate distribution with features of distributing along grain boundary and without agglomerations which are existed in liquidoid is achieved in composite treated by UVT for 240 s. The mechanism of UVT eliminates TiB2 agglomerations mentioned above can be summarized as a sketch shown in Fig. 7. The average grain size of composites with and without UVT are shown in Fig. 8. Compared with base metal, the grains of composite without UVT already have an effective refinement, as shown in Fig. 8. The promotion of heterogeneous nucleation is usually be the main reason of this grain refinement, and the growth of grains is also obstructed by TiB2 particles. The promoted heterogeneous nucleation can be mainly attributed to the little excessed intermediate phase Al3Ti as an effective heterogeneous nuclei [5] (reaction (1) and (2) have different reaction rate and may cause the formation of few Al3Ti phase. With element B, Al3Ti phase can effectively refine aluminum grains at a very low content and easily consumed by peritectic reaction: L þ Al3 Ti / a Al [27]). As long as UVT treats the composites, the grains of composites are further refined, as shown in Fig. 8. The decreasing of average grain size of composites with UVT can be explained by two sides. Firstly, more effective dispersed particles enhance the hindering on growth of grains with the elimination of large agglomerations. Secondly, the cavitation can promote the nucleation. Temperature of the bubbles decreases with the expansion of cavitation bubbles, and the equilibrium melting point increases on the effect of pressure pulses caused by collapse of bubbles [24]. These two phenomenon both increase the undercooling at certain area in the melt, and usually be used to explain the situations when UVT treats the semi-solid state
Fig. 6. BSE microstructures of quenching samples of in-situ TiB2p/Al-4.5Cu composites: (a). UVT for 60 s, (b). UVT for 120 s, (c). UVT for 240 s.
6
Q. Gao et al. / Journal of Alloys and Compounds 692 (2017) 1e9
Fig. 7. Sketch of effects of ultrasonic cavitation and acoustic streaming on distribution of particles.
Fig. 8. Average grain size of in-situ TiB2p/Al-4.5Cu composites and base metal.
metal. In this study, the temperature of the melt during UVT is higher than the liquidus. Thus nuclei created by cavitation will remelt at this high temperature. But with temperature fluctuation of liquid metal, the nuclei may partly kept into the solidification. Compared between composites treated by UVT for different time, the average grain size has barely changed. That may due to the balance between cavitation creates nuclei and nuclei re-melt makes no big difference in the number of nuclei in composites treated by UVT for different times. And although small agglomerations remain in the melt after short time UVT, the majority of TiB2 particles are dispersed as long as UVT introduced in the melt. Thus the obstruction of particles on growth of grains shows no obviously enhanced with extension of UVT time. Along with grain refinement, the gathers along grain boundary are also refined as mentioned above. Comparing composites with and without UVT, the refinement of gathers is mainly attributed to homogenization of dispersed particles in the melt and grain boundary regions enlarged with the refinement of grains. Comparing composites treated by UVT for different times, the effect of homogenization introduced by acoustic streaming cannot catch up with the massive increasing of dispersed particles caused by the effective elimination of large agglomerations. And small
agglomerations (under 50 mm in size) settled at the intersection of grain boundary are quite similar with gathers. Thus the gathers in composite treat by UVT for 60 s are larger and wider. As UVT time extended to 120 s, the amount of dispersed particles in the melt has no big increase but the effect of homogenization introduced by acoustic streaming enhanced with the extension of treatment time. Small agglomerations are also reduced and refined. Therefore the gathers along grain boundary are refined. Following the same tendency, the gathers in composite treated by UVT for 240 s have the most effective refinement of all with the longest homogenizing time and the elimination of small agglomerations as mentioned above. In our previous work, a kind of tiny TiB2 particles is found in matrix. These tiny particles are 10e100 nm in size [20]. In this work, these nanoparticles are found not only in the gaps of larger particles but also more likely to form a kind of tiny agglomerations with a size near 3 mm in composite without UVT, as shown in Fig. 9(a) and (b). TiB2 particles in nano size have much higher surface energy to the submicron ones, thus they are more likely to form agglomerations to decrease surface energy. It is worth to note that ultrasonic vibration also effectively breaks these tiny agglomerations with the high pressure shock-waves. And the dispersion of nanoparticles is achieved in composite treated by UVT for 240 s, as shown in Fig. 9(c) and (d). Fig. 10 shows the dispersed particles in matrix of composites with and without UVT. Although majority of TiB2 particles follow the tendency to distribute along grain boundary and form the gathers after solidification, some dispersed TiB2 particles still could be found at the vicinity of the gathers, with the further observation of particulate distribution along grain boundary regions by BSE as shown in Fig. 10. This phenomenon happens in all the composites fabricated in this work. These particles can be identified as particles being partly captured by the grains [28]. With the size under 400 nm, these dispersed particles are all suitable reinforcements for Orowan strengthening. Fig. 11 shows the result of TEM analysis, the pattern of particle phase matches the (0 0 1), (1 0 0) and (1 0 1) of TiB2 crystal structure, which further confirms the particles are TiB2 phase with the evidence of X-ray diffraction patterns. The high resolution result of TiB2 particle also shows that TiB2 particles are well bonded with a-Al matrix, the particle-matrix interface are clean and no interface reaction products or other inclusions can be found.
Q. Gao et al. / Journal of Alloys and Compounds 692 (2017) 1e9
7
Fig. 9. Dispersion of tiny agglomerations: (a) Location of tiny agglomerations in composite without UVT, (b) Tiny agglomerations A, (c) Location of Dispersed nanoparticles in composite treated by UVT for 240 s (d) Dispersed nanoparticles in zone C.
3.3. Mechanical properties of the composites The as-cast mechanical properties of base metal and TiB2
particles reinforced composites with and without UVT are shown in Table 1. The yield strength is effectively improved by reinforcement of TiB2 particles comparing composites with base metal. The
Fig. 10. BSE microstructures of dispersed particles of in-situ TiB2p/Al-4.5Cu composites: (a). Without UVT, (b). UVT for 60 s, (c). UVT for 120 s, (d). UVT for 240 s.
8
Q. Gao et al. / Journal of Alloys and Compounds 692 (2017) 1e9
Fig. 11. TEM result of TiB2 particles.
integrated effects of Orowan strengthening, load-bearing strengthening, CTE (Coefficient of Thermal Expansion) mismatch strengthening and grain refinement is believed to be the reasons of improvement of yield strength. According to Orowan strengthening mechanism, dispersed TiB2 particles under 400 nm in size with high elastic modulus can hinder the dislocation movement and form dislocation line looped around particles. Then subsequent dislocation movement is hindered by the interaction between dislocation loops of adjacent TiB2 particles and yield strength of composites is improved [5,19,20]. The well bonded particles with matrix can directly strengthen composites with shearing of load, and increasing yield strength according to load-bearing strengthening [29]. After solidification, the CTE mismatch between aluminum matrix and TiB2 particles generates the residual plastic strain. Then dislocations are generated by residual plastic strain at interface of TiB2 particles and matrix to accommodate CTE mismatch and strengthen composites [30]. The Hall-Petch relationship explains that decreasing of grain size increases yield strength [5,20]. Therefore yield strength of composite without UVT is improved by 32.8% to base metal with presence of TiB2 particles and its dispersed part, the refinement of average grain size from 216.1 mm to 46.0 mm also contributes, as shown in Fig. 8. Composites with and without UVT have no different in content of TiB2 particles, so load-bearing strengthening is not suitable to explain the improvement of yield strength after UVT. After UVT, the elimination of large agglomerations brings much more dispersed particles distributed in matrix and effectively enhance Orowan strengthening. The grains of composites are further refined and grain boundary regions are enlarged. With enlarged grain boundary and uniform distribution of particles, CTE mismatch strengthening is also enhanced. Therefore compared with composite without UVT, yield strength of composites with UVT is improved by 35.3%, 43.5% and 60% at 60, 120 and 240 s of treatment time, respectively. With similar average grain size, the increasing of yield strength with the extension of UVT time is mainly attribute to Orowan strengthening and CTE mismatch strengthening. Since small agglomerations are
Table 1 As-cast mechanical properties of in-situ TiB2p/Al-4.5Cu composites and base metal. Specimens
UVT time (s)
Yield strength (MPa)
UTS (MPa)
Elongation (%)
Al-4.5Cu AMC AMC AMC AMC
0 0 60 120 240
64 85 115 122 137
163 221 237 238 245
12.50 8.05 5.50 5.85 6.15
gradually eliminated with UVT and further brings more dispersed particles. And distribution of particles gets more uniform when UVT time is longer. Thus Orowan strengthening, CTE mismatch strengthening are enhanced with longer UVT time, and composite treated by UVT for 240 s has the highest yield strength. When explaining the improvement of UST of composites, grain boundary strengthening is assumed to be the main reason. In the gathers along grain boundary, distance of adjacent particles becomes very short. These particles may not contribute to Orowan strengthening since dislocation loops may not successfully formed and small number of particles are too big. But they can still hinder dislocation movement by their high elastic modulus and high hardness. And CTE mismatch strengthening is still working on particles in gathers. Dislocations are hindered and assembled at grain boundary which will strengthen composites. Thus UTS of composite without UVT is improved by 35.6% to base metal. When UVT is introduced to composites, the elimination of large agglomerations and further refining of grains enhance grain boundary strengthening, since much more TiB2 particles uniformly distributed in the enlarged grain boundary regions. Therefore UTS of composites treated by UVT are higher than untreated composite. But these improvements are not very effective according to Table 1. It may due to slags remained in the melt are also broken and homogenized by UVT and makes them more hardly to removed or floated to the riser, and UVT also agitates the oxidation and brings more oxide inclusions into the melt. But the small agglomerations are gradually eliminated and TiB2 particles are distributed more uniformly with the extension of UVT time, thus UTS of composites have small improvements. The deformation of grains is hindered by vast TiB2 particles distributed along grain boundary and the ductility is harmed, thus the composites have lower elongation than base metal. Although UVT eliminates large agglomerations and homogenizes TiB2 particles in the melt, the inclusions shows more damages on ductility, and elongation of composites treated by UVT is lower than composite without UVT. But the ductility is slightly improved with longer UVT time, and this trend is attribute to the more uniformly distributed TiB2 particles with gradually elimination of small agglomerations. 4. Conclusions Through salt-metal reaction route, TiB2 particulates reinforced Al-4.5Cu alloy composites are successfully fabricated. Intermediate phases like AlB2 and Al3Ti are effectively avoided by controlling the composition of reactant and completion of reactions. The grains are effectively refined. Ultrasonic vibration treatments are successfully introduced into composites fabrication. The particulate distribution is effectively improved. Large agglomerations in the melt are eliminated at the very early stage of ultrasonic vibration treatment. With extension of ultrasonic vibration treatment time, small agglomerations are also eliminated. TiB2 particles are uniformly distributed along grain boundary after ultrasonic vibration treatment for 240s throughout the bulk matrix. The grains are further refined. The gathers of TiB2 particles which are formed by pushing at solid/liquid interface during solidification in grain boundary regions, are refined by ultrasonic vibration treatment. With the increasing of ultrasonic vibration treatment time, the gathers are more refined. The break of tiny agglomerations formed by TiB2 particles under 100 nm in size is also observed after ultrasonic vibration treatment for 240s. Dispersed TiB2 particles smaller than 400 nm are found nearby the gathers in grain boundary regions. Ultrasonic vibration treatments greatly improve as-cast
Q. Gao et al. / Journal of Alloys and Compounds 692 (2017) 1e9
mechanical properties of the composites. After ultrasonic vibration treatment for 240s, yield strength of composite is improved by 114% and 61% comparing with base metal and untreated composite, respectively. Acknowledgement This work was financially supported by the National Natural Science Foundation of China (No. 51574129), and by the Fundamental Research Funds for the Central Universities, HUST (No. 2014QNRC003, No. 2015QN010). The authors would also like to express their appreciation to Analytical and Testing Center, HUST. References [1] T.W. Clyne, P.J. Withers, An Introduction to Metal Matrix Composites, Cambridge University Press, 1993. [2] K.U. Kainer, Metal Matrix Composites, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, 2006. [3] N. Chawla, K.K. Chawla, Metal Matrix Composites, Springer ScienceþBusiness Media, New York, 2013. [4] D.D.L. Chung, Composite Materials, Springer-Verlag London Limited, 2010. [5] L. Lu, M.O. Lai, F.L. Chen, Al-4wt% Cu composite reinforced with in-situ TiB2 particles, Acta. Mater. 45 (1997) 4297e4309. [6] J.Y. Xua, B.L. Zou, S.Y. Tao, M.X. Zhang, X.Q. Cao, Fabrication and properties of Al2O3-TiB2-TiC/Al metal matrix composite coatings by atmospheric plasma spraying of SHS powders, J. Alloys. Compd. 672 (2016) 251e259. [7] R.H. Dong, W.S. Yang, Z.H. Yu, P. Wu, M. Hussain, L.T. Jiang, G.H. Wu, Aging behavior of 6061Al matrix composite reinforced with high content SiC nanowires, J. Alloys. Compd. 649 (2015) 1037e1042. [8] B. Ya, B.W. Zhou, H.S. Yang, B.K. Huang, F. Jia, X.G. Zhang, Microstructure and mechanical properties of in situ casting TiC/Ti6Al4V composites through adding multi-walled carbon nanotubes, J. Alloys. Compd. 637 (2015) 456e460. [9] C.S. Ramesh, R. Keshavamurthy, Slurry erosive wear behavior of Ni-P coated Si3N4 reinforced Al6061 composites, Mater. Des. 32 (2011) 1833e1843. [10] K. Kalaiselvan, N. Murugan, S. Parameswaran, Production and characterization of AA6061-B4C stir cast composite, Mater. Des. 32 (2011) 4004e4009. [11] X.Z. Kai, K.L. Tian, C.M. Wang, L. Jiao, G. Chen, Y.T. Zhao, Effects of ultrasonic vibration on the microstructure and tensile properties of the nano ZrB2/ 2024Al composites synthesized by direct melt reaction, J. Alloys. Compd. 668 (2016) 121e127. [12] K.L. Tee, L. Lu, M.O. Lai, Synthesis of in situ Al-TiB2 composites using stir cast route, Compos. Struct. 47 (1999) 589e593. [13] K.L. Tee, L. Lu, M.O. Lai, Improvement in mechanical properties of in-situ AlTiB2 composite by incorporation of carbon, Mater. Sci. Eng. A 339 (2003)
9
227e231. [14] L. Lu, M.O. Lai, Y. Su, H.L. Teo, C.F. Feng, In situ TiB2 reinforced Al alloy composites, Scr. Mater. 45 (2001) 1017e1023. [15] L. Lu, M.O. Lai, F.L. Chen, In situ preparation of TiB2 reinforced Al base composite, Adv. Compos. Mater. 6 (1997) 299e308. [16] M. Emamy, M. Mahta, J. Rasizadeh, Formation of TiB2 particles during dissolution of TiAl3 in Al-TiB2 metal matrix composite using an in situ technique, Compos. Sci. Technol. 66 (2006) 1063e1066. [17] Z. Liu, M. Rakita, W. Xu, X. Wang, Q. Han, Ultrasound assisted salts-metal reaction for synthesizing TiB2 particles at low temperature, Chem. Eng. J. 263 (2015) 317e324. [18] M.K. Akbari, H.R. Baharvandi, K. Shirvanimoghaddam, Tensile and fracture behavior of nano/micro TiB2 particle reinforced casting A356 aluminum alloy composites, Mater. Des. 26 (2015) 150e161. [19] Z. Zhang, D.L. Chen, Consideration of Orowan strengthening effect in particulate-reinforced metal matrix nanocomposites: a model for predicting their yield strength, Scr. Mater. 54 (2006) 1321e1326. [20] Q. Gao, S.S. Wu, S.L. LÜ, X.C. Duan, P. An, Preparation of in-situ 5 vol% TiB2 particulate reinforced Al-4.5Cu alloy matrix composites assisted by improved mechanical stirring process, Mater. Des. 94 (2016) 79e86. [21] S.L. LÜ, S.S. Wu, C. Lin, Z.Q. Hu, P. An, Preparation and rheocasting of semisolid slurry of 5083 Al alloy with indirect ultrasonic vibration process, Mater. Sci. Eng. A 528 (2011) 8635e8640. [22] Z. Hu, X. Li, H. Yan, X.Q. Wu, H. Qun, J.W. Lin, Effects of ultrasonic vibration on microstructure evolution and elevated-temperature mechanical properties of hot-extruded Mg-6Al-0.8Zn-2.0Sm wrought magnesium alloys, J. Alloys. Compd. 685 (2016) 58e64. [23] G.I. Eskin, D.G. Eskin, Production of natural and synthesized aluminum-based composite materials with the aid of ultrasonic (cavitation) treatment of the melt, Ultrason. Sonochem. 10 (2003) 297e301. [24] C. Lin, S.S. Wu, S.L. LÜ, P. An, Effects of ultrasonic vibration and manganese on microstructure and mechanical properties of hypereutectic Al-Si alloys with 2%Fe, Intermetallics 32 (2013) 176e183. [25] S.S. Wu, H. Nakae, Nucleation effect of alumina in Al-Si/Al2O3 composites, J. Mater. Sci. Lett. 18 (1999) 321e323. [26] S.S. Wu, H. Nakae, T. Kanno, Y. You, Zone-melted unidirectional solidification of particulate dispersed composites, J. Mater. Sci. 36 (2001) 225e229. [27] B.S. Murty, S.A. Kori, M. Chakraborty, Grain refinement of aliminum and its alloys by heterogeneous nucleation and alloying, Int. Mater. Rev. 47 (2002) 3e29. [28] P.L. Schaffer, D.N. Miller, A.K. Dahle, Crystallography of engulfed and pushed TiB2 particles in aliminum, Scr. Mater. 57 (2007) 1129e1132. [29] J. Nampoothiri, R.S. Harini, S.K. Nayak, B. Raj, K.R. Ravi, Post in-situ ultrasonic treatment for generation of Al-4.4Cu/TiB2 nanocomposite: a route to enhance the strength of metal matrix nanocomposites, J. Alloys. Compd. 683 (2016) 370e378. [30] M.L. Wang, D. Chen, Z. Chen, Y. Wu, F.G. Wang, N.H. Ma, H.W. Wang, Mechanical properties of in-situ TiB2/A356 composites, Mater. Sci. Eng. A 590 (2014) 246e254.