The present study investigates the fabrication of Ni3Al-carbon nanotube (CNT) nanocomposite using electrically activated reaction synthesis (EARS) and its effects on the mechanical properties of the nanocomposite. The effect of initial nickel (Ni) particle size and mechanical milling time of Ni-CNT/Al hierarchical composite powder on reaction characteristics, product microstructure and properties was investigated for the first time. An increase in mechanical milling time was found to result in a decline in ignition temperature and time to ignition for the two investigated initial nickel particle sizes (4–8 µm and 45–90 µm). The smaller initial nickel particle size and longer milling times had a major influence on the homogeneity, decreasing porosity, and increasing hardness of the reacted compacts.
The process of combustion synthesis or reaction synthesis (RS) has been investigated for decades with the view of utilizing the enthalpy of reaction between elemental powders to produce products in short processing times. Intermetallics, ceramics, and their composites have all been processed using RS. Despite clear advantages, which include rapid synthesis, a low thermal budget and purer products, the process also has its fair share of shortcomings, namely, product inhomogeneity and product porosity. This problem can become more exaggerated when adding a non-reacting reinforcement phase to the reacting powder mixture to produce composites , which can buffer the reaction. Over the past few decades, other traditional manufacturing processes and external effects have been combined with RS resulting in significant improvements . For example, the application of an electric field has been shown to improve reaction characteristics, produce homogenous microstructures and even allow reactions to take place in otherwise unfavorable material systems. One of the authors recently reviewed various electric field-assisted reaction synthesis processes, and discussed their effects on material synthesis . Another added effect is the mechanical activation of reactant powders prior to the reaction (enabled through the process of mechanical milling). As such, mechanically activated self-propagating high-temperature synthesis (MASHS)  has emerged as a process that promotes the formation of nanostructured products with enhanced reaction wave velocities . Recently, White et al.  investigated the electrically activated thermal explosion of mechanically activated Al/Ni powder to form NiAl, while Thiers et al.  studied the thermal explosion of Ni–Al system in a furnace. When combustion synthesizing NiAl, the initial powder compact contains aluminum powder of relatively large volume fraction (in accordance with the NiAl composition), hence in such a case there is little issue in aluminum powder becoming percolated within the powder compact prior to the reaction. For the Ni3Al composition however, due to the relatively lower volume fraction of aluminum, the aluminum particles need to be small enough in relation to the nickel particles to allow the percolation of aluminum throughout the powder compact, which is a requirement for successful processing [2,8]. Recently, Korchagin et al. investigated the thermal explosion of 3Ni-Al powder compacts and specifically the effect of mechanical activation on the ignition temperature and activation energy . Previously the authors had investigated electrically activated reactive synthesis (EARS) and electro-annealing of 3Ni-Al powder compacts and the effect of adding CNTs to the 3Ni-Al powder mixture [10,11].
The present authors also investigated the effect of carbon nanotube (CNT) length and milling parameters on the nickel matrix crystal size, dislocation density, micro-strain, and CNT integrity post milling of mechanically milled Ni-CNT composite powder . Past studies have shown difficulty in processing materials with a high Ni3Al product content. In the present study, we investigate the effect of initial nickel particle size on the EARS processing of Ni3Al-CNT composites. The effect of milling time on reaction and product characteristics is also discussed for the first time, together with the effect on the reacted compacts’ local mechanical response in the form of microhardness.
Powders used in the study (Fig. 1) were nickel (4–8 µm and 45–90 µm, 99.9% pure, spherical; Atlantic Equipment Engineers, USA), aluminum (< 45 µm, Atlantic Equipment Engineers), and ulti-wall carbon nanotubes (30–50 nm in diameter and 10–20 µm in length; 95% purity, cheaptube.com). To prepare the initial powder mixture of Ni and CNTs, first the CNTs were placed in a beaker with 15 ml of methanol (Fisher Scientific, 99.99% purity) and sonicated in a sonication bath (KENDAL Ultrasonic cleaner, USA) for 15 min at 42 kHz to de-agglomerate the CNTs. To remove moisture from the powders, the CNT along with nickel powder were then degassed in an IsoTemp Vacuum Oven at 250 °C for 2 h. Then, the powders were cooled and removed in a protective argon atmosphere. The Ni + CNT, Ni—28.7 g, CNT—1.17 g, (CNT amounting to a 10 vol% final loading in Ni3Al) were then placed in a tool steel vial with S2 steel milling balls (1/4 in. diameter) enough to achieve a ball-to-powder weight ratio of 5:1 (also adding 4 wt% of methanol as a process control agent which reduces excessive cold welding) and then milled in a SPEX® 8000 high-energy ball mill at 1750 rpm for four different times; 1.5 h, 3.0 h, 4.5 h, and 6.0 h.
For the second-stage milling, Al powder was first rotator mixed for 45 min at 100 rpm with the mechanically milled Ni-CNT composite powder in the composition of (Ni3Al). The powders were further degassed to remove contamination and moisture followed by SPEX milling for another 1.5 h (also conducted at 1750 rpm) to produce the hierarchical composite Ni-CNT/Al powders. The 1.5 h was needed to provide coating of the Ni-CNT particles with Al.
All milling operations were conducted in an interrupted manner, with 15 min on and 10 min off to reduce heat buildup, and only the actual milling time reported. The hierarchical powders were then uniaxially pressed into a cuboidal shaped green compacts of dimensions (9–10 mm) × 10 mm × 35 mm, with relative density ∼0.70 ± 0.02.
The green compacts were drilled centrally along their lengths to produce 1–2 mm deep hole for insertion of a K-type thermocouple for recording the temperature during EARS. EARS was carried out by placing the green compact in the reactor chamber as shown in Fig. 2 and applying a direct current electric current at 5 V through the green compact using a power ten power supply (P63C-51000 Power Ten Series, Ametek). An Omega OMB-DAQ-2416 system and labview software were used to monitor and record the temperature and current profiles generated during EARS, which was carried out in an argon atmosphere to minimize oxidation.
Particle, microstructural, and phase characterization were carried out using a field emission scanning electron microscopy (FESEM), FEI Quanta 450 FESEM, energy dispersive spectroscopy (EDS) for elemental compositional analysis, and XRD, Philips X'Pert Pro (PANalytical, USA) X-ray diffractometer with a Cu X-ray tube (λ = 1.54060 Å). Product porosity was measured using image analysis (image j software), and the mechanical response of the reacted samples was investigated using microhardness (Wilson Instruments Vickers hardness testing machine (Instron, USA)), using a load of 1 kg indents are measured for each condition, and an average and standard deviation calculated are reported. Powder particle size analysis was conducted via gravitational sedimentation using a particle size analyzer (Horiba CAPA-700, USA). Quantitative phase analysis from XRD data was carried out using the Materials Analysis Using Diffraction program (MAUD),2 a Rietveld-based program.
Results and Discussion
Figure 3 shows the effect of mechanical milling time and initial nickel particle size on the resultant Ni-CNT composite particles size. Two trends are observed. First the smaller initial nickel particle size results in smaller Ni-CNT composite particle sizes at all milling times. This is a result of the higher work hardening rate of smaller nickel particle size  promoting an increase in the fracture process as opposed to re-welding thus promoting an overall reduction in particle size. Also, the increased surface area of the smaller nickel particle size powder provides more opportunity for coverage and dispersion with CNTs, which should also favor more fracturing. For both nickel particle sizes, as milling time increases the particle size is seen to decline. Although the particle size is seen to increase at 6 h of milling, closer examination of electron micrographs shows the presence of agglomerates, which could have contributed to the observation of an increase in particle size, when in fact particle size seems to continue to decline. Figure 4 Shows XRD scans of the Ni(4–8 µm)-CNT and Ni(45–90 µm)-CNT milled particles, showing the presence of small amounts of NiO, possibly formed during the mechanical milling stage. Although care was exercised to mill under an argon atmosphere, some residual oxygen may have been still present in the milling vial, coupled with the constant generation of virgin nickel fracture surfaces during milling may have provided the right conditions for NiO formation. Figure 5 shows scanning electron micrographs of the milled powder showing the presence of both CNTs and NiO. The presence of CNT peaks in XRD scans cannot however be detected. It is not unusual that CNT XRD peaks are not observed following milling as previously reported by one of the authors .
Following second-stage milling, Fig. 6 shows the aluminum phase largely coating the Ni(4–8 µm)-CNT particles, such a finding was not observed for the Ni(45–90 µm)-CNT particles to the same degree.
X-ray diffraction scans of the Ni-CNT/Al powder (Fig. 7) show the presence of both Ni and Al. The NiO is not present as its relative content has now been reduced due to the added aluminum, and the resolution limit of XRD is typically ∼3 vol%.
Figure 8 shows an example of wave propagation observed in the compacts. In this particular case, the reaction propagates locally then travels through the compact in a self-propagating type reaction. It is also noticed that cracking is prevalent not only where the thermocouple hole is (acting as a stress raiser), but also in other parts of the compact. This is believed to be due to local thermal stresses in what is expected to be a very brittle (work hardened compact with relatively high CNT content (10%)). Noting that the reaction typically requires the melting and spreading of aluminum, some powder re-arrangement is needed to accommodate such events as well as the rapid rise in local temperature. The hindrance of such rearrangement, in addition to local thermal stresses within the compacts gives rise to local cracking.
Figure 9 is an example of a temperature-time profile experienced during EARS of Ni (4–8 µm)-CNT/Al showing a combustion temperature in excess of 1000 °C being reached after ignition at a temperature of ∼90 °C (i.e., resulting in greater than 900 °C rise in temperature). Figure 10 shows the effect of initial first stage Ni-CNT milling time on the ignition temperature and time to ignition as defined in Fig. 10. Clearly an increase in milling time result in a reduction in time to ignition, ignition temperature as a direct result in mechanical activation. Also, the Ni(4–8 µm)-CNT/Al compacts display shorter ignition times and lower ignition temperatures than the Ni(45–90 µm)-CNT/Al. This can be explained by the favorable inter-connectivity/coating of the aluminum phase in Ni(4–8 µm)-CNT/Al compacts as explained earlier.
Figure 11 shows XRD plots for reacted compacts of different initial Ni particle size and different first stage milling times. It is clear that the Ni3Al peak is more significant in the Ni(4–8 µm)-CNT/Al powder compacts as opposed to the Ni(45–90 µm)-CNT/Al powder compacts. All the reacted compacts however contained multiple phases. To quantify such phases, MAUD analysis was performed on the metallic/intermetallic phases excluding CNTs, and results are displayed in Fig. 12 for converted Ni3Al and residual nickel phase. As can be seen, the volume percent of the residual nickel phase is greatest for the Ni(45–90 µm)-CNT/Al powder compacts for the milling times, indicating a more incomplete reaction. However, for both Ni(45–90 µm)-CNT/Al powder compacts and Ni(4–8 µm)-CNT/Al powder compacts, as the milling time increase the residual nickel decreases and Ni3Al conversion increases. In fact Ni3Al conversion of greater than 80% is observed for the (4–8 µm) Ni-CNT/Al powder compacts as opposed to only ∼60% for the Ni(45–90 µm)-CNT/Al powder compacts. As for the Al3Ni2 and NiAl phases, 1.5 h milling resulted in Al3Ni2 and NiAl contents of 3 and 19.1 vol% respectively for Ni(4–8 µm)-CNT/Al reacted compacts. However that changed to 5.3% and 0.9% for the 6 h milled samples, respectively. While for the Ni (45–90 µm)-CNT/Al powder reacted compacts, the phase contents changed from 0.3 and 31.4% to 0.1 and 27.3%, respectively. These results show that the initial Ni particle size has a substantial effect on reducing the high temperature NiAl phase and converting to the intended Ni3Al phase (which is the highest Ni3Al content obtained so far in EARS). The increased Al3Ni2 in the smaller Ni initial particle size may be due to the increased Ni/Al interfacial area which promotes formation of larger amounts of Al3Ni2 through solid state interdiffusion .
Figure 13 shows electron micrographs of microstructure of reacted compacts of ((4–8 µm) Ni-CNT/Al powders milled for 6 h and Ni (45–90 µm)-CNT/Al powders milled for 6 h. It is clear that the Ni (45–90 µm)-CNT/Al still contains unreacted residual elongated nickel particles (as a consequence of the mechanical milling process).
Figure 14 shows the pore content for (excluding cracked regions) for reacted compacts. An increase in first stage milling time is seen to result in a decrease in porosity for both the Ni(4–8 µm)-CNT/Al and the Ni(45–90 µm)-CNT/Al powder compacts due to an increase in the intimate mixing between Ni-CNT composite particles and aluminum. Moreover the lowest porosity is consistently found in the Ni(4–8 µm)-CNT/Al powder compacts again due to the added improvement in mixing with aluminum powders.
As seen in Fig. 15, micro-hardness of the reacted compacts increase with an increase in milling time and a reduction in initial Ni particle size. This is primarily a result of the accompanied reduced porosity and the improved conversion to Ni3Al (also reduced residual Ni) with an increase in milling time and a reduction in initial Ni particle size. The microhardness values however fall significantly below reported values for the hardness of 269.9 HV for Ni3Al (75Ni-25Al)  despite the presence of ∼ 80% Ni3Al in some of these reacted compacts. The reason for this discrepancy is due to to the presence of porosity and cracking as mentioned earlier.
The size of the initial nickel particles was found to have a pronounced effect on the EARS processing of Nickel aluminide-CNT composites.
The smaller initial nickel particle size resulted in the highest hardness and lowest product porosity, in addition to the largest Ni3Al yield following EARS so far.
An increase in first stage milling time resulted in a decline in time to ignition and ignition temperature in EARS, irrespective of initial nickel particle size.
An increase in first stage milling time in an increase in Ni3Al yield and hardness and reduced porosity irrespective of initial nickel particle size.
The authors wish to thank the National Science Foundation for funding this work under award number 1560850. The authors wish also to than Dr. Julio Valdes for use of the particle size analyzer.
Conflict of Interest
There are no conflicts of interest.