This paper is focused on developing an in-process intervention technique that mitigates the effect of built-up edges (BUEs) during micromilling of aluminum. The technique relies on the intermittent removal of the BUEs formed during the machining process. This is achieved using a three-stage intervention that consists first of the mechanical removal of mesoscale BUEs, followed by an abrasive slurry treatment to remove the microscale BUEs. Finally, the tool is cleaned using a nonwoven fibrous mat to remove the slurry debris. An on-machine implementation of this intervention technique is demonstrated, followed by a study of its influence on key micromachining outcomes such as tool wear, cutting forces, part geometry, and burr formation. In general, all relevant machining measures are found to improve significantly with the intervention. The key attributes of this intervention that makes it viable for micromachining processes include the following: (i) an experimental setup that can be implemented within the working volume of the microscale machine tool; (ii) no removal of the tool from the spindle, which ensures that the intervention does not change critical process parameters such as tool runout and offset values; and (iii) implementation in the form of canned G-code subroutines dispersed within the regular micromachining operation.
Micromilling is a key manufacturing process used on a variety of products that are critical to the aerospace, biomedical, energy, and defense sectors . While the last decade has seen significant research efforts dedicated to micromilling, the formation of built-up edges (BUEs) during the machining of materials such as aluminum continues to be a critical bottleneck for this process . This is primarily because the reduced size-scale of machining encountered in micromilling results in the BUEs having a more detrimental impact on key process outcomes such as surface roughness, burr formation, cutting forces, and tool wear [3–6].
The current solutions to mitigate the effect of BUEs in micromilling have focused on either process improvements (e.g., use of tool coatings and cutting fluids) or on tuning the processing parameters (e.g., using lower cutting speeds and avoiding chipload values below the minimum chip thickness of the material) . Tool coatings such as diamondlike carbon and nanocrystalline carbon have been shown to prolong the onset of BUEs in microscale tools . However, the lifetime of these coatings is highly unreliable, thereby limiting the potential of this approach . The use of cutting fluids during micromilling reduces the formation of BUEs in metals [5,10], However, cutting fluids are not suited for soft materials such as thermoplastic polymers that are notorious for forming BUEs during micromilling . Furthermore, the use of cutting fluids has limited appeal with the increasing trend toward sustainable manufacturing practices . The technique of tuning the micromilling parameters to mitigate the formation of BUEs also has its limitations. For example, the use of lower spindle speeds increases the overall cycle time of machining the part, whereas the use of chiploads greater than the minimum chip thickness of the material is limited by the stiffness of the micro-endmills .
The use of in-process interventions presents one viable alternative to mitigate the shortcomings of existing practices. This technique allows the tool to be operated upon intermittently by either physical and/or chemical forces that remove BUEs, thereby restoring the cutting edge. Sugihara et al.  demonstrated the power of such an approach in restoring face-milling cutters used to machine aluminum at the macroscale. In their work, once the face-milling cutter was detected to have a significant adhesion of BUEs, it was subjected to a three-stage intervention, viz. (stage 1): dipping in acetone bath to clean the tool; (stage 2): dipping in a liquid metal bath to encourage embrittlement of BUEs; and (stage 3): machining of wood to remove the chip adhesion. Such in-process interventions are particularly appealing for micromilling because these can be implemented within the working volume of the machine tool and do not require the removal of the cutting tool from the spindle. However, there has been little research in developing in-process interventions specific to the removal of BUEs during micromilling.
The objective of this research is to develop an in-process intervention technique to mitigate the effects of BUEs during the micromilling of aluminum using uncoated tungsten carbide endmills. This paper presents the development of a three-stage intervention technique for micro-endmills comprising first of the mechanical removal of mesoscale BUEs, followed by an abrasive slurry treatment to remove the microscale BUEs. Finally, the tool is cleaned using a nonwoven fibrous mat to remove any remaining slurry debris. An on-machine implementation of this technique is demonstrated followed by a study of the effect of key process parameters on machining outcomes such as tool wear, cutting forces, part geometry, and burr formation. In general, all relevant machining measures were found to improve significantly with the intervention.
The remainder of the paper is organized as follows: Section 2 presents an overview of the in-process intervention protocol including the rationale behind the three stages of intervention. Section 3 describes the experimental plan, followed by Sec. 4 that describes the results from the micromilling tests. Section 5 presents future research directions and finally, Sec. 6 outlines the specific conclusions that can be made from this work.
In-Process Intervention Protocol
Stage 3. Cleaning the microtool of the slurry deposits by using the stage 1 fibrous mat (Fig. 1(f); duration: 5 s).
The experimental setup requires a nonwoven fibrous mat roll (for stages 1 and 3) and a slurry well (for stage 2) to be implemented within the working volume of a typical microscale machine tool. The tool continues to rotate and remains in the high-speed spindle while going through each of the three stages. Therefore, the in-process intervention can be implemented in the form of canned G-code subroutines dispersed within the regular micromachining operation. The ultimate goal of the intervention protocol is to remove most, if not all, of the BUEs attached to the tool and thereby return the tool cutting edge to a “near-pristine” condition, as evidenced by the result in Fig. 1(e). Such an intermittent removal of the BUEs can be expected to slow down the tool-wear rate while also improving other critical micromachining outcomes . The remainder of this section discusses each of the three stages in detail.
Stage 1—Mechanical Removal of Mesoscale Built-up Edges.
Stage 1 is designed explicitly to remove mesoscale BUEs without damaging the microtool. Figure 1(a) shows an example of one such BUE created while machining Aluminum 6061 using a 400 μm endmill. In stage 1, the microtool is punched through a nonwoven fibrous mat (KimTechTM). The fibrous structure of the mat ensures that it has sufficient contact with the BUE volume. Furthermore, the pliable nature of the mat ensures that the forces exerted by the piercing operation (∼0.33 N) remove a majority of the BUE (Fig. 1(c)), without damaging the microtool.
Our preliminary studies revealed that this stage 1 intervention is critical, since the downstream abrasive slurry treatment (refer Sec. 2.2 ahead) is not capable of removing large BUEs present on the tool. Figure 2 shows the effect that 1 min of the stage 2 abrasive slurry treatment had on the BUE similar in size to that shown in Fig. 1(a). While there is a distinct removal of some material (as seen by the burnished areas in Fig. 2), it is clear that the force generated by the slurry is not large enough to remove the BUE. Even after increasing the time duration of the slurry treatment time from 1 min to 5 min, no appreciable reduction was found in the size of the BUE seen in Fig. 2, thereby necessitating the stage 1 intervention.
While it could be argued that stage 1 is not critical to the cases with smaller BUEs, it only adds 5 s to the overall intervention cycle time. Given that BUE sizes are highly stochastic in nature, the benefits offered by this stage far outweigh the 5 s addition to the cycle time.
Stage 2—Abrasive Slurry Treatment of Microscale Built-Up Edges.
Stage 2 is designed explicitly to remove the microscale BUEs that are present at the end of stage 1 (Fig. 1(c)). This is done using commercially available silicon carbide (SiC) lapping compounds. For the stage 2 treatment, the SiC slurry is held in a well, 10 mm diameter and 8 mm deep. The tool is then made to interact with the abrasive slurry through a combined rotation of the spindle and an orbital motion of the spindle-carrying machine-tool axis. The purpose of the orbital motion is to ensure that the boundary layer at the tool-slurry interface is continuously disturbed, thereby allowing the BUE to come in contact with new abrasive particles.
The direction of rotation of the tool (clockwise or counterclockwise) is a critical factor that dictates the duration of stage 2. Preliminary trails revealed that for the same time duration of 60 s, a clockwise rotation of the tool, during stage 2, resulted in a ∼35% increase in the edge radius, as opposed to the <2% increase in the edge radius seen while using the counterclockwise rotation. Given that the microtool used in this study is a right-hand cutter, the counterclockwise rotation was specifically selected to minimize the damage to the cutting edge (due to the abrasive particles) and to prevent the slurry from being ejected from the well (by riding up the flutes). As seen in Fig. 1(e), a 60 s, stage 2 treatment involving a counterclockwise rotation of the tool at 15,000 RPM is seen to effectively remove the BUEs present on the tool without causing damage to the tool. While the use of the counterclockwise rotation was pursued for this paper, future enhancements to this intervention protocol are expected to involve tool-motion profile recipes aimed at maximizing the removal rate of BUEs while also minimizing the rates for tool wear and ejection of the slurry.
Stage 3—Cleaning Slurry Deposits.
At the end of stage 2, the tool is still coated with remnants of the slurry. Stage 3 is aimed at mitigating the effects of the slurry deposits on the micromilling outcomes. It involves a repeat of stage 1 protocols, but this time the fibrous mat ensures that it cleans the remnants of slurry that are present on the tool.
The in-process intervention protocol outlined in Sec. 2 was experimentally evaluated using a series of micromilling operations on an Aluminum 6061-T6 alloy workpiece, as shown in Fig. 3. The experiments were conducted on a MikrotoolsTM DT110 (Singapore) hybrid micromachining unit. This machine has a positional accuracy of ±1 μm and is equipped with an 80,000 RPM NSK high-speed air bearing spindle (Japan). As shown in Fig. 3, the 200 mm × 100 mm × 100 mm working volume of this machine is large enough to incorporate the fiber-mat roller (for stages 1 and 3) and the slurry well (for stage 2).
Single-fluted micro-endmills with a nominal diameter of 400 μm (Performance Micro Tool, Janesville, WI) were used to perform a series of full-immersion slots that had a nominal axial depth of 100 μm and a length of 6 mm. The milling tests were performed under dry machining conditions, at a spindle speed of 80,000 RPM and a feed-per-tooth value of 6.25 μm. These cutting parameters were chosen based on our preliminary experiments aimed at finding machining conditions that yielded aggressive BUE formation for the aluminum workpiece.
Two SiC slurries were used in the current study. The slurries have a grit size of 800 and 1200. These slurries will be referred to as S-800 and S-1200, respectively, for the remainder of this paper (refer Sec. 4.1 for details). The efficacy of the slurries was tested by machining a series of 50 slots where the intervention was applied after every slot and by comparing those machining results with the baseline control tests that involved the same number of cuts, but without the intervention.
Tool wear, cutting forces, slot profile, and burr volume were used as the measures of performance for the intervention. Cutting force data were collected using a KistlerTM 9256C1 (Winterthur, Switzerland) three-axis dynamometer by sampling at 30 kHz. All tools were imaged before and after the micromilling operation using an optical microscope. The Alicona InfiniteFocusTM G5 ( Raaba/Graz, Austria) optical profiler was used to characterize the slot profile and burr volume. Table 1 summarizes the experimental process parameters.
This section first describes the results from the characterization of the abrasive slurries followed by the micromilling results.
Characterization of Slurries.
The abrasive slurry used in this investigation is a commercially available lapping compound manufactured by The Mosher Company, Chicopee, MA known by the trade name MOCO lapping compound. The formulation consists of a suspension of SiC abrasive particles in a binder made by dissolving light oils into melted stearic acid. Compounds with two abrasive grit sizes 800 (S-800) and 1200 (S-1200) were used in this study. As per the manufacturer's specification, both slurries contain ∼50% abrasive by weight.
The viscosities of the slurries were characterized using an AR-G2 rheometer made by TA Instruments, New Castle, DE. Given the limits of the rheometer, testing was done at angular frequencies ranging from 1 rad/s to 630 rad/s. As seen in Fig. 4, the shape of the viscosity plots of both slurries indicates that they exhibit non-Newtonian behavior. Furthermore, both slurries are also seen to exhibit a shear thinning behavior.
The viscosity of the slurries is expected to influence both the flow-field and the torque experienced by the microtool during the stage 2 abrasive slurry treatment. In order to quantify this influence, a Phantom V7.3 high-speed camera (3000 frames per sec) was used to observe the behavior of the slurries. Figures 5(a)–(e) and 5(g)–(k) depict a sample of these images captured for the S-800 and S-1200 abrasive slurry, respectively. The first frame (Figs. 5(a) and 5(g)) depicts the slurry behavior when the tool is rotated counterclockwise at 15,000 RPM without the orbital motion being imparted. As can be seen, in this case the liquid motion profile looks very comparable between the two slurries, in that the tool is seen to induce very little churning of the slurries. Our preliminary studies revealed that this pure rotation of the tool in the slurry is not effective in removing the BUEs. As seen in Figs. 5(b)–5(e) and Figs. 5(h),–5(k), the addition of the orbital motion significantly changes the interaction between the tool and the slurries. The higher viscosity S-800 slurry is seen to result in a more agitated flow pattern around the tool. The effects of these patterns are also seen on the average peak-to-valley torque values measured for the orbital motion (Table 2). These average torque values follow the same trend as the viscosity data indicating the tool exerts a greater torque when exposed to the S-800 slurry, presumably due to the higher viscosities, as seen in Fig. 4. Figures 5(f) and 5(l) show that both slurries are effective in removing the BUE.
This section outlines the key trends seen in the tool wear, cutting forces, slot-profile and burr volume of the machined slots.
In this study, the deterioration of the edge radius of the endmill was used as a measure of the tool wear . A digital image analysis technique, as outlined by Arora et al. , was used in this study to quantify the wear of the cutting edge. Figures 6(a) and 6(b) depict the trend seen in the tool wear data after micromilling 50 slots and 100 slots, respectively. In these figures, the cutting edge radius before the micromilling operation is plotted on the horizontal axis, whereas the percent increase in the postmachining cutting edge radius value is plotted on the vertical axis. The data in Fig. 6 shows that the in-process intervention significantly reduces the extent of wear experienced by the tool. After 50 slots (Fig. 6(a)), the micro-endmills that did not receive the in-process intervention showed an average increase of 79% in the tool edge radius values. This is in contrast with the intervention cases for S-1200 and S-800 slurries that showed only an increase of 8% and 2%, respectively, (Fig. 6(a)).
To further investigate the long-term benefits of the in-process intervention, 100 slots were machined and the cutting edge radius measurements were preformed (Fig. 6(b)). These measurements after 100 slots also show similar trends in that the tools that did not receive the in-process intervention had a ∼116% larger tool edge radius, while the ones with the intervention maintained near pristine condition. No discernible difference was noticed between the performance of the S-800 and the S-1200 slurries after 100 slots.
Figure 7 depicts the trends seen in the resultant cutting force experienced by the tool for each of the 50 slots. As seen, the cutting forces experienced by the control tool (no intervention case) show a continuously increasing trend, especially after the 30th slot, whereas the cutting forces seen for the tools subjected to the interventions remain fairly constant. This force increase seen for the control tool is presumed to have occurred due to a build-up of the material on the bottom face of the tool, similar to that seen in Fig. 1(a). As described ahead in Sec. 4.2.3, the axial depth of these control slots was also higher. This again is a direct consequence of such a build-up of material . In general, there was again no significant difference seen between the S-800 and the S-1200 slurries in terms of their effect on the cutting force. They both result in cutting force magnitudes that are lower than the control case with no intervention. This reduction in the cutting forces is a direct result of the edge radius of the tool being maintained at a near-pristine condition when the tool is subjected to the intervention.
The cross-sectional profile of each of the 50 slots was characterized for two parameters, viz. the depth of the slot and the tilt in the floor of the slot, as shown in Fig. 8(a). Figures 8(b) and 8(c) depict the trends seen in these measurements for the first 50 slots. Since the tool is not removed from the spindle during the in-process intervention, any variation seen in the measurements reported in Figs. 8(b)–8(c) is likely due to the changes in the effective cutting geometry of the tool, which in the case of aluminum machining is influenced heavily by the BUEs.
As seen in Fig. 8(b), micro-endmills that were exposed to the in-process intervention achieved axial depths closer to the nominal slot depth of 100 μm. More specifically, the tools exposed to the S-800 slurry produced slots with an axial depth of 100.8 ± 2.0 μm, and tools exposed to the S-1200 slurry produced slots with an axial depth of cut of 100.0 ± 2.2 μm. For the control tools that were not exposed to the in-process intervention, the slot depth is seen to be significantly higher at 110.9 ± 0.8 μm. The positive effect of the intervention is also reflected in the tilt seen on the floor of the slot. As seen in Fig. 8(c), the control case continues to show a higher degree of tilt on the floor of the slot when compared to both the S-800 and the S-1200 slurry treatments. Both of these trends in the control case are suspected to be due to the BUE adhesion that changes the geometry of the tool. The fact that the in-process intervention is successful in maintaining the near-pristine quality of the cutting edge results in an improvement in the geometrical accuracy of the machined slot. As with the other measures, the performance of the S-800 and the S-1200 slurries is not sufficiently different to be conclusive. It should be noted here that the surface roughness (Ra) values were measured to be between 0.2 and 0.5 μm and did not show any statistical difference between the control case and the two slurry interventions.
In micromilling, the burr formation is the result of extensive plastic deformation/plowing of the material . Increase in BUEs has been known to result in an increased burr formation, because the BUEs change the effective rake angle of the tool causing more material to be plowed to the sides of the slot . The volume measurement module on the Alicona InfiniteFocus G5TM optical profiler was used to calculate the total volume of burr formed on the surface of the slot over the 50 cuts. Figure 9 depicts these values for the three cases. As seen, the control case has the highest volume of burr formation whereas both cases with the intervention are seen to have a statistically significant decrease in the burr volume. This measure also points to the efficacy of the in-process intervention process.
Future Research Directions
The results outlined in Sec. 4 point to the fact that the in-process intervention protocol outlined in this paper has a positive effect on all micromachining measures. However, in order to exploit the power of this intervention for commercial applications, the following future research directions need to be pursued:
Fluid-tool interaction: The properties of the two slurries S-800 and S-1200 appear to be different; however, the results in Sec. 4 do not point to a clear winner between the two. A more detailed study of the fluid-structure interaction environment encountered in this process (Fig. 5) needs to be conducted to ascertain the reasons behind this trend. This will also allow for the effective design of the slurry components for this application.
Frequency of intervention: This study used a rather conservative intervention protocol of cleaning the tool after every cut. While the benefits of this intervention are clear, it also adds a significant increase in the total production time for the 50 slots. While this could be justified for high-value parts, the trade-offs between the frequency of intervention and its subsequent benefits need to be carefully investigated. Such an investigation should also look at the motion profile imparted to the tool in stage 2.
In-process detection of built-up edges: The efficiency of the in-process intervention will be significantly enhanced if sensing modalities capable of detecting BUE formation thresholds are developed for micromachining environments. The use of acoustic emission signatures may be beneficial for this but needs further exploration.
The following specific conclusions can be drawn from this study:
An in-process intervention technique was successfully developed to mitigate the effects of BUEs during micromilling of aluminum. The key attributes of this intervention that makes it viable for micromachining processes include the following: (i) an experimental setup that can be implemented within the working volume of the microscale machine tool; (ii) no removal of microtool from the spindle, which ensures that the intervention does not change critical process parameters such as tool runout and offset values; and (iii) implementation in the form of canned G-code subroutines dispersed within the regular micromachining operation.
The efficacy of the intervention was demonstrated by conducting microscale slotting operation on Aluminum 6061-T6 alloy. The intervention to remove the BUEs was seen to drastically reduce the tool wear, the cutting forces, and burr volume. It was also seen to improve part accuracy as indicated by the nominal depth of the slot and the tilt seen on the floor of the slot.
While there appears to be differences between the 800 and 1200 grit SiC slurries (both in their viscosities and their subsequent fluid-tool interactions during the intervention), the findings are inconclusive on a clear winner between the two. Further research is needed to study their behavior.
The start-up funds made available by the Mechanical, Aerospace, and Nuclear Engineering Department at Rensselaer Polytechnic Institute (RPI) is acknowledged for funding this research. J.F. Nowak acknowledges support from the National Science Foundation Graduate Research Fellowship Program (NSF GRFP) under Grant No. DGE-1247271. The authors also appreciate Prof. Chang Ryu (Department of Chemistry and Chemical Biology, RPI) for the use of his rheometer, and Rensselaer M.Eng. student Bum-Joon Jung for performing the rheometer measurements.
National Science Foundation Division of Graduate Education (Grant No. DGE-1247271).
Rensselaer Polytechnic Institute (Internal Funding).