Cryogenic Mechanical Milling
Cryogenic mechanical milling, also known as “cryomilling,” is a modified ball milling technique that utilizes cryogenic media, typically liquid nitrogen, to provide cooling to the ball milling process. The liquid nitrogen can be either internal or external to the milling jars containing the powders to be mixed. Cryomilling is most widely used to develop nanostructured alloys as it provides several advantages over other ball milling techniques, especially HEBM. The low temperatures induced by cryomilling make it difficult for particles to agglomerate, sinter, or melt together, and make reactions slower or less likely. Oxidation is especially reduced during cryomilling. The low temperatures minimize cold welding or grain growth, thereby enabling the attainment of fine grain structures. The severe mechanical impacts along with the reduction of dislocation annihilation and recrystallization yield metallic powders with very high dislocation densities and hence very high strength.
Maite et al.  utilize cryomilling to synthesize Al-CNT composites using gas atomized spherical A15083 powder and MWCNTs of 40-47 nm diameter and length of 0.2-0.5 pm. One set of Al/CNT mixtures were cryomilled together for 8 h, with a BPR of 32:1, and a speed of 180 rpm using 10 mm diameter stainless steel balls. Liquid nitrogen was maintained at a constant level within the milling attritor. A second mixture of Al/CNT consisted of Al powder cryomilled using the above parameters, followed by a second cryomilling run where MWCNTs were introduced. The second lower energy run was conducted using 10 mm stainless balls for 1 h, at a speed of 180 rpm, and a BPR of 10:1. Powders were consolidated using spark plasma sintering.
Agglomerates persisted using the two-step approach, resulting in poor densifica- tion, mechanical properties, and wear resistance. The combined mixing yielded improved CNT dispersion, better densification, and mechanical properties over the two-step mixing approach. However, CNTs exhibited more damage in the combined mixing approach and mechanical properties only matched that of cryomilled Al. Wear resistance did improve in the Al-MWCNT composite using the combined mixing, compared to the control cryomilled Al material . Omidi et al.  used similar cryomilling parameters to synthesize A16061-1 wt.% CNT composites. Omidi et al.  found that cryomilled Al-CNT composites exhibited greater wear resistance over ball milled composites.
Solution Ball Milling (SBM)
Solution ball milling (SBM) is an approach that combines w'et chemistry with ball milling . Figure 2.5 shows the process schematically. Chen et al.  first use an isopropanol solution laden with surfactants (1 wt.% zwitterionic surfactants
FIGURE 2.5 Schematic of solution ball milling (SBM) process for CNT dispersion . (Reproduced with permission from Elsevier).
with both hydrophilic and hydrophobic groups) to create a CNT solution with l wt.% CNT. This solution is then introduced into the milling jars, along with the aluminum matrix powder and the milling media. The second step is to mill the combined powder and CNT solution; Chen et al.  utilize a planetary ball mill but other ball milling configurations could be used. 160 g of the l wt.% CNT solution was added to 160 g of Al powder, yielding Al-0.5 wt.% CNT composites. The milling media and jars w'ere made of zirconia. Chen et al.  use milling parameters of 200 rpm for 60 min total milling time, with an intermittent 10 min “off’ time used for every 10 min of milling in order to prevent over-heating. The milling process results in a flaky Al powder with CNTs distributed throughout the flaky powder particulate surfaces. The milled slurry is then transferred to a beaker where it is let to stand for ~ 15 min in order to allow the composite Al-CNT powders to settle. The upper solution, which likely contains some remaining suspended CNT particles, is then poured out. The settled slurry mixture is dried in an oven at 80 °C for 30 min.
Raman spectroscopy revealed that the CNTs in the SBM processed powder incurred less damage than those processed by high energy ball milling. SBM (717 nm CNTs after process) was able to prevent CNT rupturing and preserve the starting length of CNTs (1135 nm) much better than HEBM (503 nm after 2 h milling) . The SBM processed Al-CNT composites exhibited greater elongation during tensile tests than Al-CNT composites processed by HEBM for 4-48 h of milling .
Rotational Milling Modeling
A quick peruse of the ball milling literature reveals that myriad milling parameters have been used to synthesize CNT-MMCs, as well as other nanocarbon-based metal matrix composites. Liu et al.  present a modeling effort to determine the best milling parameters to attain optimal CNT dispersion with minimal damage to the CNTs themselves. Experimental studies on variable speed milling utilizing stainless steel milling media with a 10:1 BPR, Al-5 Mg alloy as the matrix, and entangled CNTs with a diameter of 10 nm and a length of ~5 pm, were used to compliment and support the modeling assumptions. Milling parameters ranged from a mild condition of 300 rpm for 2 h, to 450 rpm for 8 h. The physical mechanisms observed in the experiments and underpinning the foundations of the model are as follows: (i) deformation stage where metal particulates becomes flattened, (ii) CNT cluster break-up and dispersion onto metal particle surface, (iii) cold welding state where flattened metal flakes weld together to form laminate structures, and (iv) CNT incorporation into the bulk of metal flakes welded together. CNT cluster break-up must sufficiently occur in the deformation stage in order to have a uniform distribution of individual CNTs on the metal flakes. Once cold welding occurs, further break-up of CNT clusters would be minimal, and any remaining CNT clusters would be incoiporating into the new coarsening or welded metallic particulates. CNT clusters would adversely affect performance by serving as stress concentrators and inhibiting adequate percolation to promote enhanced transport properties.
Aiken et al.  established a model for the fraction of powders (fA10) in their initial undeformed state and the fraction of powders (fM) that have been deformed and cold welded, based on the probability of cold welding (a), and the time interval between two different ball collisions (r), given by Eq. 2.1.
As milling time increases from time “0” to time “f” the fraction of deformed and cold welded powder increases and at a certain fraction, it can be presumed that the cold welding mechanism is dominant. The time interval between different ball collisions is a function of the mean free path of balls during milling (2), and the velocity of balls at the instant before collision (v), given by Eq. 2.2.
The time interval for two different ball collisions directly correlates with the time for the metallic powder particulates to flatten, denoted as /flalten. As previously stated, CNT clusters will break up simultaneously during the metallic powder deformation stage. As such, the break-up of CNT clusters can be quantified by the evolving size of CNTs clusters as milling time increases, as given by Eq. 2.3.
Where D is the CNT cluster size at time t, D0 is the initial CNT cluster size, £ is a coefficient, and t is the strain on the milled powders. As the cluster size decreases, at a sufficiently small size the cluster can be considered to “disappear,” that is, no longer considered a cluster and instead when this happens it can be presumed that adequate CNT dispersion is achieved. As such, the time necessary for adequate CNT dispersion can be formulated as per Eq. 2.4.
Furthermore, intuitively it is understood that high impact velocities would correlate to large strains, and this is quantified by Maurice et al. [27,28] as shown in Eq. 2.5.
The time for CNT dispersion is hence proportional to impact velocity. A ratio between the time necessary for dispersing CNTs to the time necessary to flatten metallic powder particles can be thereby be established, as provided by Eq. 2.6.
Equation 2.6 indicates that at very high impact velocities, which are driven by higher milling rotational speeds, the time for CNT dispersion will be greater than that needed to flatten metal powder particles. Once powders are flattened, cold welding begins, and hence CNT clusters that have not been broken down will become
FIGURE 2.6 Schematic of (a) AI-CNT powders during ball impact during milling, and (b) CNT distributions during milling at different rotating speeds . (Reproduced with permission from Elsevier).
embedded within the cold welded metal particles. In contrast, at low velocities or milling speeds, CNT clusters may be broken down before metallic particles are sufficiently flattened in order to provide the high particulate surface area needed to disperse individual CNTs on the surface. These conceptual extremes are shown schematically in Figure 2.6. At intermediate speeds, the onset of cold welding is sufficiently delayed to allow for metal powders to flatten and provide sufficient surface area for the proper dispersal of the broken up CNT clusters. Lie et al.  indicated that milling parameters of 400 rpm for 8 h appeared to attain this intermediate condition where adequate dispersion of CNTs was achieved.
Shift-Speed Ball Milling (SSBM)
The shift-speed ball milling (SSBM) is a ball milling approach that integrates the implications of previously discussed rolling milling modeling, namely that optimal milling conditions for CNT-MMC mixing consist of both a high- and low-intensity milling step. SSBM utilizes both low-speed ball milling (LSBM) with a low rotational speed, and high-speed ball milling (HSBM) with a higher rotational speed. Xu et al.  present the first iteration of SSBM for synthesizing Al-CNT composites. A SSBM approach consisting of 8 h of LSBM and 1 h or HSBM was compared with LSBM-only and HSBM-only runs at 9 h each. For all runs, a planetary ball mill with stainless steel milling media was used. The jars were filled with Ar gas, the ball-to- powder ratio was kept at 20:1, and the instrument had a 300 mm revolution radius and a 90 mm rotational radius. The LSBM runs or steps utilized a rotational speed of 135 rpm, while the HSBM runs and steps had a rotational speed of 270 rpm.
During LSBM, the deformation mechanism of the powders transitions from flattening to fracturing after 6 h. Powder milled using only LSBM (for 9 h) produced flakes with 200 nm thickness and 40 pm diameter (starting powder was spherical with 30 pm diameter). During HSBM, the first 2 h consisting of powder flattening, after which cold welding occurs. Powder produced using only HSBM produced particles tens of microns thick and hundreds of microns thick. SSBM yielded intermediate results, with flakes typically <20 pm thick and <150 pm wide. CNTs were observed to be well embedded in the SSBM case. Raman spectroscopy was used to quantify damage on the starting CNTs (1.11) and milled CNTs using the ID/IG ratio.
It was revealed that damage incurred onto CNTs during SSBM (1.21) was relatively low and closer to that occurring during LSBM (1.13), than that caused by HSBM (1.69). Overall, the LSBM step during SSBM provides sufficient time for CNTs to be dispersed onto the gradually flattening A1 powder particles. A slow flattening rate delays the onset of cold welding, where CNT dispersion and cluster break-up are no longer possible. Once proper dispersion occurs during LSBM, the short (1 h) HSBM step promotes between bonding between A1 flakes and CNTs. Consolidated compacts from the SSBM powder exhibited similar strength to those made by only LSBM or only HSBM. but ductility that was 2-3 times higher .