- Email: [email protected]
Abrasive slurry jet micro-machining of holes in brittle and ductile materials
Journal of Materials Processing Technology 214 (2014) 1909–1920
Contents lists available at ScienceDirect
Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec
Abrasive slurry jet micro-machining of holes in brittle and ductile materials K. Kowsari a , H. Nouraei a , D.F. James a,∗ , J.K. Spelt a,b,∗∗ , M. Papini b,a,∗ ∗ ∗ a b
Department of Mechanical and Industrial Engineering, University of Toronto, 5 King’s College Road, Toronto, ON, Canada M5S 3G8 Department of Mechanical and Industrial Engineering, Ryerson University, 350 Victoria Street, Toronto, ON M5B 2K3, Canada
a r t i c l e
i n f o
Article history: Received 11 September 2013 Received in revised form 3 April 2014 Accepted 10 April 2014 Available online 21 April 2014 Keywords: Hole micro-machining Abrasive slurry jet Polymeric additives Viscosity Fluid elasticity Erosion
a b s t r a c t This paper investigated the effects of elasticity and viscosity, induced by a dilute high-molecular-weight polymer solution, on the shape, depth, and diameter of micro-holes drilled in borosilicate glass and in plates of 6061-T6 aluminum alloy, 110 copper, and 316 stainless steel using low-pressure abrasive slurry jet micro-machining (ASJM). Holes were machined using aqueous jets with 1 wt% 10 m Al2 O3 particles. The 180 m sapphire orifice produced a 140 m diameter jet at pressures of 4 and 7 MPa. When the jet contained 50 wppm of dissolved 8 million molecular weight polyethylene oxide (PEO), the blind holes in glass were approximately 20% narrower and 30% shallower than holes drilled without the polymer, using the same abrasive concentration and pressure. The addition of PEO led to hole cross-sectional profiles that had a sharper edge at the glass surface and were more V-shaped compared with the U-shape of the holes produced without PEO. Hole symmetry in glass was maintained over depths ranging from about 80–900 m by ensuring that the jets were aligned perpendicularly to within 0.2◦ . The changes in shape and size were brought about by normal stresses generated by the polymer. Jets containing this dissolved polymer were observed to oscillate laterally and non-periodically, with an amplitude reaching a value of 20 m. For the first time, symmetric ASJM through-holes were drilled in a 3-mm-thick borosilicate glass plate without chipping around the exit edge. The depth of symmetric blind holes in metals was restricted to approximately 150 m for jets with and without PEO. At greater depths, the holes became highly asymmetric, eroding in a specific direction to create a sub-surface slot. The asymmetry appeared to be caused by the extreme sensitivity of ductile materials to jet alignment. This sensitivity also caused the holes in metals to be less circular when PEO was included, apparently caused by the random jet oscillations induced by the polymer. Under identical conditions, hole depths increased in the order: borosilicate glass > 6061-T6 aluminum > 110 copper > 316 stainless steel. The edges of the holes in glass could be made sharper by machining through a sacrificial layer of glass or epoxy. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Applications of low-pressure abrasive slurry jets, or micro abrasive suspension jets as they are also sometimes called, have
∗ Corresponding author at: Department of Mechanical and Industrial Engineering, University of Toronto, 5 King’s College Road, Toronto, ON, Canada M5S 3G8. Tel.: +1 416 978 3049; fax: +1 416 978 7753. ∗∗ Corresponding author at: Department of Mechanical and Industrial Engineering, University of Toronto, 5 King’s College Road, Toronto, ON, Canada M5S 3G8. Tel.: +1 416 978 5435; fax: +1 416 978 7753. ∗ ∗ ∗Corresponding author at: Department of Mechanical and Industrial Engineering, Ryerson University, 350 Victoria Street, Toronto, ON M5B 2K3, Canada. Tel.: +1 416 979 5000x7655; fax: +1 416 979 5265. E-mail addresses: [email protected] (D.F. James), [email protected] (J.K. Spelt), [email protected] (M. Papini). http://dx.doi.org/10.1016/j.jmatprotec.2014.04.008 0924-0136/© 2014 Elsevier B.V. All rights reserved.
been evolving for over a decade. One important application is in the abrasive slurry jet micro-machining (ASJM) of glass. As a machining process, it is economical and offers advantages such as the absence of a heat-affected zone, low forces on the work piece, no tool wear or vibration, and the ability to machine virtually any material. The small divergence of the jet makes ASJM well suited for precision machining. Miller (2004) developed a high-pressure ASJM system, using a slurry content of approximately 20 wt%, 8 m diameter garnet abrasives, pressures of roughly 70 MPa, and 40–60 m diameter orifices. The system was used to cut a variety of materials, including metals, polymers, glass, and composites, with thicknesses ranging from 50 to 3000 m. The depth to width ratios of the cuts were much greater than those typical of laser cutting. Nguyen et al. (2009) drilled holes in glass using a low-pressure (3 MPa) and a high concentration (8.2 wt%) slurry, and, in a companion paper,
1910
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
Wang et al. (2009a) studied the effects of pressure and machining duration on erosion in holes in glass. In both papers, the hole cross-sections were found to be W-shaped, related to the dominance of ductile erosion when the abrasive particles had relatively low kinetic energy. Nouraei et al. (2014) found that higher kinetic energies led to U-shaped holes with flatter bottoms and steeper sidewalls. Wang et al. (2009b) examined the profiles of holes machined in glass and found them to be asymmetric because of orifice vibration and misalignment. A typical cross-sectional profile shape was divided into various zones, and the shape of each zone was explained in terms of the direction of the slurry flow relative to the walls of the hole. Aqueous jets containing long-chain polymer additives, but not abrasives, have been studied for several decades. Hoyt et al. (1974) discovered that polymers delayed droplet formation and therefore enhanced jet stability by damping free surface disturbances. Another application of polymeric additives is friction reduction in turbulent pipe flow. For example, Elbing et al. (2011) found that trace concentrations of order 10 wppm (weight parts per million) reduced wall friction by about 75%. These advantages resulted from induced elasticity in the fluid. In general, the polymer chains are stretched from their initially-coiled equilibrium state when subjected to a sufficiently high deformation rate. This action generates normal stresses which increase the resistance to extensional deformation; i.e. the fluid behaves partially as an elastic solid, as described by Larson (1999, p. 123). Hashish (1991) compared the cutting of aluminum using highpressure abrasive waterjet machining (AWJM at 345 MPa), in which abrasive particles and air are entrained in a high-velocity jet, and high-pressure abrasive slurry-jet cutting. In both sets of experiments, the water contained 3 wt% 18-million (18-M) molecular weight polyacrylamide (PAM). The slurries were assumed to be Newtonian; therefore, the effects of fluid elasticity induced by the added polymers, as described by Larson (1999, p. 123), were not considered. He found that ASJM produced twice the cutting depth as AWJM using the same abrasive flow rate and pressure. Nguyen et al. (2008) found that concentrated solutions (1000–5000 wppm) of 1-M PAM enhanced the stability and increased the coherent length of an abrasive waterjet. Ashrafi (2011) found that the addition of a large amount of cornstarch (10–22 wt%) in AWJM produced
cuts having narrower kerfs with steeper sidewalls, In another study with cornstarch, Omrani et al. (2013) also found a reduction in kerf taper, and hypothesized that it was due to changes in the fluid viscosity. The role of fluid elasticity generated by the polymer was not considered. In work more closely related to the conditions of low-pressure ASJM than of high-pressure AWJM, Luo et al. (2010) compared low-pressure abrasive slurry jet polishing of glass with and without 4000 wppm (0.4 wt%) of a high-molecular-weight polymer, and found that the polymer sharpened the transition zone separating polished and unpolished regions, because of reduced divergence of the streamlines of the polymeric jet about the stagnation zone. In a companion study, Liao et al. (2009) investigated the effects of additives such as sodium hexametaphosphate (NaP), 4-M polyethyleneglycol, 20-M polyacrilic acid, 12-M PAM, and antimony trinitride on the low-pressure abrasive slurry-jet polishing of glass. They found that a combination of NaP and PAM resulted in the lowest surface roughness and postulated that this effect was due to the lower viscosity of this solution relative to others considered in the study, but did not consider fluid elasticity. Wang et al. (2009c) investigated machining using a variety of polymers such as PAM, anionic polyacrylamide (HPAM), and cationic polyacrylamide (PAMA), and several abrasives (22 m garnet, boron carbide, and white and brown corundum). They found that the slurry containing 6000 wppm 5-M PAM and white corundum particles yielded the most symmetrical holes with the largest material removal rate. No comparisons were made to holes machined without additives. Most recently, Kowsari et al. (2014) studied the effects of dilute concentrations of polyethylene oxide (PEO, 25–400 wppm), on the shape, roughness, and width of low-pressure ASJM channels machined in glass. It was found that, for a constant jet velocity, diameter and dose of abrasive, the channel width decreased by 21% with the addition of 50 wppm of 8-M (PEO). This change was caused by normal stresses generated by the polymer, which focussed the erosion to a smaller blast zone. The objective of the present study was to extend the above research on the effects of long-chain polymers in ASJM to the drilling of holes in ductile and brittle materials, with a focus on the
Fig. 1. (a) Schematic of the ASJM components (not to scale). (b) Orifice geometry (dimensions are in m).
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
1911
Fig. 2. Holes machined using the water-only slurry jet: (a) a 0.5◦ misalignment, and (b) alignment with an accuracy of 0.2◦ .
resulting shape, erosion rate, depth and diameter. As in the previous work, it was of interest to separate the effects of slurry elasticity and viscosity on the observed patterns and rates of erosion. The work also investigated the potential of sacrificial polymeric or glass surface layers to reduce hole opening diameters. 2. Experimental apparatus and procedures 2.1. ASJM apparatus The ASJM apparatus described in Kowsari et al. (2014) was utilized for all experiments. The basic apparatus, as shown in Fig. 1(a), consisted of an abrasive slurry pump, a pulsation damper connected to an open reservoir tank filled with a premixed aluminum oxidewater slurry, and a sharp sapphire orifice with a diameter of 180 m having a length/diameter ratio of 1.67 (Fig. 1(b)). The positive displacement pump (LCA/M9/11-DC, LEWA Inc., Leonberg, Germany) was set to give flow rates of 1.67 mL/s and 2.34 mL/s (±0.1%), at pressures of 4 and 7 MPa, respectively. The jet diameter was measured optically using a digital camera attached to a microscope having a field of view of 3 mm × 2 mm, and was found to be a uniform 140 m over the 30-mm standoff distance (distance between the orifice plate and the target) as shown in Fig. 1(c). The vena contracta could not be observed because it occurred within the 10 mm long orifice stem. The contraction coefficient (the ratio of the jet cross-sectional area to that of the orifice) was found to be 0.60 ± 0.03. This value conforms to the theoretical value of 0.64 for a sharp orifice as explained by Falkovich (2011). The slurry was pumped to the orifice through a 3.2-mmdiameter tube. The flow velocity was not quite high enough to completely eliminate particle settling, and so the abrasive concentration delivered to the target was 10% less than that in the mixing tank. However, this concentration was constant over the duration of the experiments, as confirmed by collecting slurry samples in a beaker during one-minute periods and then drying them to yield the abrasive content.
Fig. 3. Nozzle guide to maintain constant standoff distance and alignment with the target.
Newtonian. When glycerin was added, the fluid remained Newtonian, as explained by Larson (1999, p. 107). In contrast, the slurry with PEO was non-Newtonian, since the long-chain molecules can induce significant elasticity in the fluid, such as causing resistance to extensional deformation, as described by Bird et al. (1987, p. 637). The concentrations of glycerin and PEO were selected so that both solutions had a viscosity 10% greater than that of water. The abrasive particles were taken from the supplier’s container and subjected to the ASTM quartering technique
2.2. Slurry preparation Three different aqueous slurries were prepared, all having a 1% concentration of alumina (Al2 O3 ) abrasive particles with a nominal diameter of 10 m (Comco Inc., CA, USA). One slurry contained only water and abrasive, another also contained 5.2 wt% glycerin, and in the third 50 wppm of 8-M PEO was dissolved. The glycerin concentration was selected to yield a viscosity equal to that of the PEO solution, as measured with a Cannon-Fenske capillary tube viscometer, following the procedure of ASTM D4889-04 (ASTM, 2011). Because of the small concentration of particles, the basic slurry was
Fig. 4. Fourier transform plot of the amplitude of the lateral jet oscillations obtained with a slurry containing 50 wppm of 8-M PEO.
1912
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
Fig. 6. Profiles of holes machined in glass under identical conditions with the wateronly slurry at 1.67 mL/s.
controlled using a shutter to intercept the jet. Five holes were machined separately under identical conditions with the wateronly slurry, and the maximum variation in their depths was less than 3%.
Fig. 5. Microscope images of a hole machined using the water-only jet (a) hole opening in the plane of the target surface, and (b) cross-sectional profile.
(ASTM C702-98, 2003), to ensure a uniform distribution. The sample portions were weighed on an electronic scale (AP110, OHAUS Corp., Pine Brooke, NJ, USA) accurate to 0.1 mg. The PEO had a viscosity-averaged molecular weight of 8.0 M, according to the supplier (Sigma-Aldrich, St. Louis, MO, USA). The selected PEO concentration was based on the earlier work of Kowsari et al. (2014), who found that it produced the minimum ASJM channel width. To prepare the solution, the PEO was first dispersed in 20 mL of ethanol to avoid aggregation, and then this mixture was added to 6 L of slurry in the reservoir tank and stirred slowly to minimize polymer degradation. 2.3. Machining The process parameters were selected to provide holes with shapes and diameter-to-depth ratios similar to those in other ASJM studies conducted without additives, such as the study by Nouraei et al. (2014). Holes were machined in 100 mm × 50 mm × 3 mm borosilicate glass plates (Borofloat 33® , Schott Inc., NY, USA) at a standoff distance of 30 mm. The duration of machining was
Fig. 7. (a) Depth, frosting diameter, surface diameter, and average-diameter normalized by the jet diameter as a function of machining time, and (b) rates of change of these dimensions as a function of machining time. Holes machined in glass with the water-only slurry at 1.67 mL/s. The lines are only to help guide the eye.
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
1913
Fig. 8. (a) Hypothesized streamlines of the flow field during the machining process. (b) Magnified regions of the boxes in (a) to illustrate the trajectories and impact angles of particles from the incident slurry of the jet, , and the return slurry, , relative to the surface and particle impact angle, ˛, near the stagnation point.
During initial tests, it was found that accurate alignment of the jet normal to the target plane was crucial to obtain symmetric holes; e.g. a misalignment of even 0.5◦ produced a relatively large asymmetry, as shown in Fig. 2(a). By placing the three-pointguide in Fig. 3 in contact with the orifice plate and target surface, the misalignment was reduced to 0.2◦ , optically measured using a microscope having a field of view of 3 mm × 2 mm. The resulting machined holes were found to be symmetric, as shown in Fig. 2(b).
2.4. Elasticity-induced jet oscillation Kowsari et al. (2014) noted that the addition of PEO, but not the glycerin, caused the jet to oscillate laterally with an amplitude that increased with the concentration of dissolved polymer. This behavior was investigated further for the present 50 wppm 8-M PEO solution by capturing microscope images having a field of view of 3 mm × 2 mm for 10 s, at a frequency of 29 Hz, and then extracting the position of the jet from each frame. The jet diameter was found to be constant at 140 m, but the jet oscillated laterally with a varying amplitude of up to 20 m, producing an effective jet diameter of 180 m. Without the PEO, the jet had the same diameter (140 m), but did not oscillate. Furthermore, the jet containing glycerin did not oscillate. The frequency spectrum in Fig. 4 shows that there was no dominant frequency in the jet oscillation. This is consistent with the work of Rothstein and McKinley (1999), who found that oscillations in non-Newtonian jets can result from randomly fluctuating vortices formed upstream of a sudden contraction, such as that of the orifice in Fig. 3.
3. Results and discussion 3.1. Holes without additives 3.1.1. Reference blind holes Fig. 5 shows an example of a hole machined in glass using the water-only slurry without PEO or glycerin, at a flow rate of 1.67 mL/s for 8 min. The side walls of the hole merged with the target surface through a curved region that ended in a frosted zone on the surface. Crosssectional profiles of the holes were viewed through an adjacent optically-clear edge in the glass plate using a microscope having a field of view of 2000 m × 1500 m, and were quantitatively characterized using an image analysis system (Clemex Vision PE, Clemex Technologies Inc., QC, Canada) and digital software (ImageJ software—http://rsb.info.nih.gov/ij/). Fig. 6 illustrates how the profiles machined with the water-only slurry at 1.67 mL/s developed with machining time, with the opening width at the surface increasing more quickly than the width at the sidewalls. This is shown in Fig. 7(a) which compares four parameters, all normalized by the jet diameter: (i) hole depth, (ii) frosting diameter; i.e., the limit of the lightly pitted surface around the hole, (iii) surface diameter; i.e. the boundary of the profile where the slope of the linear fit to 5 consecutive profile points, about 5 m apart, was < 10%, and (iv) average diameter; i.e., the cross-sectional area of the hole divided by its depth. It is seen that the hole depth increased less than linearly with machining time, consistent with the findings of Nouraei et al. (2014). It is hypothesized that the decline in the rate of depth enlargement was due to the decreasing flow and particle velocity as the depth of the hole increased. As in Figs. 6 and 7(a) reveals that the
1914
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
relative increases in the frosting diameter and surface diameter with machining time were much more than that of the average diameter. Fig. 7(b) shows that the frosting diameter grew much more rapidly than all other hole dimensions, particularly in the early stages of machining. Beyond about 2 min, the rates of change of the hole depth, the frosting diameter and the surface diameter were nearly the same, all decreasing approximately linearly with time. This can be explained in terms of the three zones found by Wang et al. (2009c) (Fig. 5), the shapes of which can be described by the flow and erosion mechanism illustrated in Fig. 8. The depth depends on the erosion in zone A, the average diameter depends mostly on erosion of the sidewalls (i.e. the upper part of zone A and zone B), and the surface diameter depends on the erosion in zone C. It is hypothesized that zone A evolved due to erosion at the high impact angles, ˛, of the deflecting flow as shown in Fig. 8(b). The relatively lower material removal in the sidewalls was due to two effects: (i) the shallower impact angles, (Fig. 8(b)), of the return slurry traveling toward the target surface would reduce erosion of the sidewall since glass erosion is maximum at 90◦ as described by Neilson and Gilchrist (1968), and (ii) the reduced particle kinetic energy after the initial impact in zone A would also decrease erosion. Since the depth was developed by higher-velocity slurry at high impact angles, its growth rate was larger than that of the average diameter as shown in Fig. 7(b). As the holes became deeper at longer machining times, the velocity of the flow at the bottom of the hole declined, consistent with the reduction in the rate of change in both the depth and average diameter. As shown in magnified region (i) of Fig. 8(b), the surface curvature in zone C at the mouth of the hole was primarily created by a high-velocity and nearly perpendicular incidence, , of the flow component originating from the peripheral streamlines near the surface of the jet. Therefore, the surface diameter (zone C) was developed by a flow having a higher normal impact velocity than the average diameter (zones A and B), consistent with the larger growth rate of the surface diameter relative to the average diameter, seen in Fig. 7(a). Subsequent to the erosion in zone C, the deflected slurry flowing over the target surface caused light pitting and scratching, but negligible erosion, as it impacted at shallow angles of attack, thereby creating the frosted region which developed very quickly (Fig. 7). Therefore, both the frosting diameter and the surface diameter were created by the high-velocity streamlines; however, the growth of the surface diameter required removal of relatively large amounts of material compared to the light pitting in the frosting region, consistent with the much larger growth rate in the frosting diameter in Fig. 7(b). As the hole diameter became large relative to the jet diameter at longer machining time holes, less of the erosive energy of the jet reached the hole edges, resulting in a reduction in the rate of change in both the surface diameter and frosting diameter as seen in Fig. 7(b). The ASJM of micro-channels was much less sensitive to the alignment of the slurry jet. For example, an orifice misalignment of approximately 0.5◦ produced no asymmetry in channel crosssections (Kowsari et al., 2014). In that case, the slurry was relatively free to flow away from the primary impact zone, and so secondary erosion of the channel walls was small. In contrast, the flow within a hole was highly confined, so that secondary erosion of the hole walls was much more significant and susceptible to small changes in the direction of the incident jet. 3.1.2. Through-holes The machining time at a flow rate of 1.67 mL/s was increased to 90 min to create a through-thickness hole on a 3-mm-thick plate of borosilicate glass. As shown in Fig. 9(b), the exit of this pierced hole was circular and without edge chipping. In contrast, Hashish (1988) observed chipping at the exit edges of holes drilled using highpressure AWJM, which he attributed to sub-surface stress waves
Fig. 9. (a) Comparison of profiles of a hole machined in glass with the water-only slurry for 8 min and a through-hole after 90 min, at a flow rate of 1.67 mL/s. (b) Microscope image of the exit of the through-hole.
created by the impact of the jet. The 180 m diameter of the exit hole was larger than the jet diameter (140 m), probably because of flow near the stagnation point, as illustrated in Fig. 8. It was also observed that the initial breakthrough was followed by about three seconds of rapid widening of the exit hole before the jet could pass through the target without interference and the hole diameter became stable. As illustrated in Fig. 9, the cross-section of the entrance region continued to broaden and the radius of curvature grew with time as the hole was drilled. 3.2. Comparison of holes in brittle and ductile materials Fig. 10(a) superimposes the profiles of shallow holes machined in glass, 316 stainless steel, 110 copper, and 6061-T6 aluminum alloy. As shown, there were significant differences in the shapes and sizes of the four holes even though they were all machined at a flow rate of 2.34 mL/s for 10 s. Of particular interest is the much sharper definition of the opening curvature with the metals, when compared to the larger radius of curvature in glass. As described
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
1915
Fig. 10. (a) Profiles of holes machined in brittle and ductile materials under identical conditions, using the water-only slurry at a flow rate of 2.34 mL/s; (b) microscope images of the hole openings machined under identical conditions in (i) 6061 aluminum, (ii) 110 copper, (iii) 316 stainless steel, and (iv) borosilicate glass.
in Section 3.1, particle impacts at a high impact angle () are dominant near the opening of holes. Since the erosion of brittle materials such as glass is maximum at a perpendicular incidence, these materials have highly eroded hole entries. Ductile materials, however, have lower relative erosion rates at high impact angles which resulted in much less erosion in zone C at the mouth of the hole. This is again consistent with the slurry erosion patterns described by Neilson and Gilchrist (1968) for brittle and ductile materials. In contrast to glass, for metal holes, the shallow impact angle () of the low-velocity return slurry (Fig. 8(b)) dominated erosion in the entry region, causing the holes to become wider without significantly increasing the radius of curvature at the edge. This effect is consistent with the findings of Liu (2007), who compared high-pressure AWJM holes drilled in glass and aluminum, and found considerable wear in the vicinity of the entrance in holes machined in glass, but not in aluminum. Despite this difference in opening curvatures, surface pitting was observed to generate the frosted region in both the metals and glass (Fig. 8(b)), which was
caused by slurry flowing over the target at relatively small impact angles, as described in Section 3.1. As seen from Figs. 2 and 6, it was possible to machine deep symmetrical holes in the borosilicate glass because the ASJM erosion process was stable. With metals though, the process became unstable at aspect ratios of approximately one, and a radial groove or slot would extend from the edge of the hole. This instability was investigated further using transparent polymethylmethacrylate (PMMA), which eroded in a similar ductile manner as the metals. Fig. 11, obtained under the same conditions as those in Fig. 10, shows that the hole was asymmetrical and that a deep groove formed. The erosion rate increased upon the onset of groove formation at a depth of approximately 150 m because the return slurry began to flow away from the incoming jet via the groove, whereas in holes without grooves, the interference between incident and return slurry dissipated the jet’s impact energy. This is demonstrated by the rate of deepening in aluminum, which increased from approximately 14–27 m/s subsequent to groove formation.
1916
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
The origin of the asymmetry which led to the formation of the grooves was investigated by comparing a hole drilled at a reference angular position about the jet centerline with another hole machined after rotation by 90◦ relative to the reference angle. Comparisons confirmed that the direction of the grooves was consistent with that of jet misalignment. As described in Section 2.3, when the jet was within 0.2◦ of 90◦ to the surface, there were no irregularities in the holes machined in glass, which were presented in Section 3.1. Therefore, drilling in ductile materials was apparently much more sensitive to jet alignment, probably because the erosion rate was maximum at an impact angle of approximately 30◦ in metals and 90◦ in glass. If the jet was perfectly aligned, the return flow from the bottom of a hole was approximately parallel to the vertical side walls in zones A and B (Fig. 8), and the local particle impact angle was small, thereby causing relatively little widening of the hole compared with the increase in depth (compare Fig. 7(a) and (d)). In metals, it is hypothesized that perturbations of the jet flow toward one side of the hole would produce a relatively large increase in the erosion rate on that side as the local impact angle approached 30◦ . Once a groove started to form, its erosion was accelerated by the concentration of flow in the direction of the groove and, as mentioned above, the decreasing resistance to incoming slurry flow as the slot developed. In contrast, the glass was more tolerant of small misalignments since the erosion rate on the sides of holes was relatively small at shallow impact angles. 3.3. Effects of fluid elasticity
Fig. 11. (a) Cross-sectional profile and (b) microscope image of the opening of a hole machined in PMMA, using the water-only slurry. The cross-sectional profile in (a) corresponds to the view in the direction of the white arrow in (b).
Kowsari et al. (2014) found that the width of an ASJM channel in glass was affected by the addition of polymer to the slurry. Specifically, the width decreased by 21% when 50 wppm of 8-M PEO was dissolved in the slurry compared to a channel machined without polymer. In general, dissolving a polymer causes two effects: it induces elasticity and thus generates normal stresses in the fluid, and it increases the viscosity. As explained in Section 2.2, these two effects were separated by using an aqueous glycerin solution having the same viscosity as the PEO solution; i.e. both solutions were 10% more viscous than water.
Fig. 12. (a) Hole cross-sectional profiles, and (b) microscope images of the openings of shallow holes machined in glass under identical conditions with the water-only slurry (left), the glycerin slurry (middle) and the PEO slurry.
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
1917
Fig. 13. (a) Depth versus machining time, and (b) frosting diameter, (c) surface diameter, and (d) average diameter versus depth for holes machined in glass with and without 50 wppm of 8-M PEO. The machining times are included in brackets. The dashed lines are only to help guide the eye.
3.3.1. Holes in glass Fig. 12(a) presents the profiles of shallow holes machined with the water-only slurry compared with the glycerin slurry and the polymer slurry, at 1.67 mL/s and 5 s duration. The 10% increase in viscosity with glycerin slurry reduced the depth by 6%, but it did not significantly change the profile shape, the average diameter, nor the size of the frosted zone (Fig. 12(b)). The depth of the PEO hole decreased because of the increased viscous drag which further decelerated the particles, thus reducing their ability to erode, as described by Clark (1992). In addition to this viscosity-induced reduction, the depth of the PEO hole was reduced an additional 17% by the polymer. The additional decrease in erosion was due to
further particle deceleration as the long-chain polymer molecules induced tensile stresses in the fluid in front of a particle, thereby retarding its ability to flow out of the closing gap between the particle and the target. The added polymer also changed the shape of the hole, from the flat-bottomed U-shape without polymer to more of a rounded V-shape, as illustrated in Fig. 12(a). These trends are illustrated as a function of machining time and hole depth in Fig. 13 for both the water-only and the PEOcontaining slurries. Fig. 13(a) shows that the 10% increase in the viscosity of the polymer solution reduced the depth by an average of 29%, over an 8-minute period, presumably by increasing the drag
Fig. 14. Microscope images of hole openings in metals with and without PEO, at 2.34 mL/s for 10 s.
1918
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
on the abrasive particles. Fig. 13(b), (c), and (d) show that for samedepth holes from 84 to 704 m, the frosting, surface, and average diameters decreased by averages of 25%, 14%, and 15% respectively, when the 50 wppm of 8-M PEO slurry was used. These changes in depth, diameter, and shape are similar to those found in channels machined in glass using PEO-containing abrasive slurries (Kowsari et al., 2014). As discussed above, the normal stresses generated by the PEO caused the fluid to resist deformation, especially in regions of extensional motion such as stagnation point flow. That is, the polymer solution acted as a viscoelastic material, resisting extension as well as shear, as described by Bird et al. (1987, p. 637). As described previously, this extra resistance near the stagnation zone (Fig. 8) caused the primary area of jet erosion to be smaller since the streamlines were more focussed, and caused greater particle deceleration at the bottom of the hole. Fig. 13(b) shows that the reductions in the frosting diameters and surface diameters produced by the PEO were larger than in the average diameters. This is because the deformation rate was larger near the opening region (zone C), in which the erosion was caused by a high-velocity slurry, than in the bottom of holes (zones A and B), which were developed by slower-moving slurry as described in Section 3.1. This resulted in greater resistance to spreading of the streamlines near the opening, and thus larger relative reductions in the frosting and surface diameters than in the average diameter. The effects of elasticity were generally smaller with deeper holes. For example, it was found that, whereas PEO reduced the frosting, surface, and average diameters by 41%, 19%, and 13% in the 125-m deep holes compared with the water-only slurry, these reductions were only 15%, 10%, and 6%, respectively, for 700 m holes, since the larger extensional rates that occur in shallow holes generated more resistance to motion. The extensional rate decreased as the holes became deeper, because of a decrease in the velocity within the cavity, consistent with the decreasing rate of depth evolution with machining time shown in Fig. 13(a). 3.3.2. Holes in metals In order to avoid the instability discussed in Section 3.2, the effect of fluid elasticity on holes machined in the three metals was examined only for relatively shallow holes less than 150 m. As observed with glass in Fig. 13(a), the addition of PEO reduced the depths in the metals by approximately 20% compared to the water-only holes. Although the holes machined with the PEO slurry had smaller average openings, they were highly irregular as seen in Fig. 14. This irregularity was tested for directional consistency by rotating the orifice as described in Section 3.2. The results indicated that, contrary to what was found for water-only slurry holes in ductile materials, the asymmetry was not always related to the direction of the jet in the case of the PEO slurry. Therefore, it is likely that the asymmetry resulted from the random oscillations of
Fig. 15. Profile of a hole machined in glass through a sacrificial layer.
the jet when it contained PEO, as discussed in Section 2.4. This jet instability led to transient jet asymmetry which, combined with the sensitivity of the ductile erosion mechanism to changes in the impact angle along the side walls, quickly caused the holes to become non-circular. 3.4. Machining through a sacrificial surface layer It was of interest to see if a sacrificial layer on top of the glass target could be used to decrease the hole opening diameter and reduce or eliminate the frosted zone. Fig. 8 illustrates the flow within a hole and the edge rounding developed over the machining duration. This was investigated by covering the glass target surface with two materials in order to minimize the opening diameter in the target plate, as shown in Fig. 15. Four sacrificial layers were tested: (i) a 500 m coating of epoxy adhesive (J-B Weld, Sulphur Springs, Texas, USA); (ii) an 800 m coating of the same material; (iii) a 150 m thick borosilicate glass cover slip bonded to the target using a room-temperature process (Jia et al., 2004); and (iv) the same cover slip glued to the target using cyanoacrylate adhesive. Thicknesses were measured using a digital caliper accurate to ±20 m. Fig. 16 compares hole openings with the four layers to an opening without a layer, all machined using the water-only slurry. The micrographs show that the sacrificial layers significantly reduced the diameter of the frosted zone and hence the effective opening diameter; for example, the reduction was 38% in the case of the 500-m-thick epoxy layer. However, the hole machined with the glued cover slip had irregular edges, as shown by (d) and (e), presumably because of leakage and erosion at the interface between the bonded glass surfaces. Fig. 17(a) compares the cross-sectional profiles of the holes drilled for 8 min through the two epoxy layers, showing that
Fig. 16. Microscope images of the openings of holes machined in glass using the water-only slurry at 1.67 mL/s for 8 min: (a) without sacrificial layer, (b) with 500 m thick epoxy, (c) 800 m thick epoxy, (d) bonded glass sheet, and (e) glued glass sheet.
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
1919
Fig. 17. Cross-sectional profiles of holes machined in glass at a flow rate of 1.67 mL/s for 8 min without a sacrificial layer and using (a) 500 and 800 m-thick epoxy coatings, and (b) water-bonded and glued glass sheets. (c) Profiles of holes machined at a flow rate of 1.67 mL/s for 6 min without a sacrificial layer and a hole machined for 8 min using a 150 m thick bonded glass sheet.
the epoxy coating decreased the radius of the opening curvature (increased the edge sharpness), but did not eliminate this rounding even when the coating thickness was increased. This was due to the relatively high erosion rate of the epoxy, which reduced its effectiveness as a barrier to erosion on the glass surface. When the coating material was changed to glass, the edge was significantly sharpened, as shown in the profiles presented in Fig. 17(b). However, leakage between the glass cover plate and the target caused the profile of the hole opening to be asymmetric regardless of the bonding method, as indicated in the upper part of the figure. As expected, the coatings reduced the depths because of the time required to drill through the sacrificial layers. This is illustrated in Fig. 17(c) by the similar depths of the 6-minute hole in the bare target and the 8-minute hole through the target covered with the 150-m-thick glass cover plate; i.e. approximately 2 min was required to drill through 150 m of glass. 4. Summary and conclusions An abrasive slurry jet micro-machining (ASJM) system was used to investigate the effects of elasticity and viscosity, brought about by a dilute high-molecular-weight polymer solution, on the shape, diameter, and depth of micro-holes drilled in glass,
various metals, and a plastic. The jet incidence was aligned to be perpendicular to a target surface to within 0.2◦ in order to avoid asymmetry in machined holes. The effects of viscosity and elasticity, both induced by an added polymer, were separated by using an aqueous glycerin solution of an equivalent viscosity. Jets containing a polymer, but not glycerin, were found to oscillate laterally and non-periodically, increasing the effective diameter by approximately 22%. Holes were machined with and without additives under fixed machining conditions. To identify the role of PEO, reference blind holes with machining times ranging from 5 to 480 s were machined in glass with the water-only slurry. The frosting diameter; i.e., the limit of the lightly eroded surface around a hole, surface diameter; i.e. the boundary of the cross-sectional profile at which the slope of the linear fit to 5 consecutive profile points, about 5 m apart, was < 10%, and average diameter, represented by the cross-sectional area of a hole divided by its depth, were all found to increase with depth as the incoming and return slurries progressively eroded the sidewalls and openings of holes over the machining time. For the first time, ASJM of through-holes revealed that the circular exit had a clean exit edge, which continued to grow to a diameter approximately 20% larger than that of the jet, subsequent to the initial breakthrough because of deflecting streamlines at the bottom of the
1920
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
holes. A comparison of holes machined under identical conditions demonstrated that the material removal rate was higher in glass than in metal holes. The symmetry of holes in metals, and not glass, was limited to a depth of approximately 150 m by sub-surface slots that developed because the ductile erosion mechanism caused the erosion of sidewalls to be much more sensitive to jet incidence alignment than in glass. When 50 wppm (weight parts per million) of 8-M (million) molecular weight polyethylene oxide (PEO) was dissolved in the slurry, the U-shaped profiles of the reference holes in glass became more V-shaped. This change was accompanied by decreases in the depths, frosting diameters, and average diameters by averages of 25%, 14%, and 15% respectively, over depths ranging from 84 to 702 m. It is thought that these changes were brought about by fluid elasticity which reduced the spread of the jet in the stagnation region. In ductile materials, the holes were found to be irregular because of the combined effects of elasticity-induced jet oscillations and sensitivity of sidewall erosion to jet incidence. Machining through an epoxy coating as a sacrificial layer reduced the frosting diameter by about 40% and sharpened the opening curvature. Similar changes were seen using bonded and glued borosilicate glass cover slips, despite leakage at the interface. In contrast to the asymmetrical ASJM holes in glass presented the work of Wang et al. (2009c), the present results demonstrate that symmetrical holes can be machined in glass by precise alignment of the jet. The work also revealed that in contrast to the reductions in widths of cuts with the use of concentrated polymer solutions in high-pressure AWJM obtained in the work of Omrani et al. (2013), the use of dilute high-molecular-weight solutions or sacrificial surface layers can significantly decrease the frosting and average diameters of ASJM holes in glass. Acknowledgements The authors acknowledge the Natural Sciences and Engineering Research Council of Canada, Canada Research Chairs, Micralyne Inc. and BIC Fuel Cells Inc. References ASTM C702-98, 2003. Standard Practice for Reducing Samples of Aggregate to Testing Size. American Society for Testing and Materials (ASTM), http://www.astm.org/database.cart/historical/C702-98R03.htm ASTM D4889-04, 2011. Standard test methods for polyurethane raw materials: determination of viscosity of crude or modified isocyanates. American Society for Testing and Materials (ASTM), www.astm.org/Standards/D4889.htm
Ashrafi, N., 2011. Viscoelastic abrasive waterjet. Int. Mech. Eng. Congress Exposition 3, 677–681. Bird, R.B., Amstrong, R.C., Hassager, O., 1987. Dynamics of Polymeric Liquids, vol. 1, 2nd ed. John Wiley & Sons, New York, NY, pp. 637. Clark, H.M.I., 1992. The influence of the flow field in slurry erosion. Wear 152, 223–240. Elbing, B.R., Solomon, M.J., Perlin, M., Dowling, D.R., Ceccio, S.L., 2011. Flow-induced degradation of drag-reducing polymer solutions within a high-Reynolds-number turbulent boundary layer. J. Fluid Mech. 670, 337–364. Falkovich, G., 2011. Fluid Mechanics – A Short Course for Physicists. Cambridge University Press, Cambridge, England. Hashish, M., 1988. Turning, milling, and drilling with abrasive water jets. In: Woods, P.A. (Ed.), 1988 Proc. 9th Int. Symp. Jet Cutting Techn., BHRA Fluid Engineering. Cranfield, United Kingdom, pp. 113–131. Hashish, M., 1991. Cutting with high-pressure abrasive suspension jets. In: 1991 Proc. 6th Int. Symp. Jet Cutting Techn, WJCT, Houston, Texas, pp. 439–455. Hoyt, J., Taylor, J., Runge, C.D., 1974. The structure of jets of water and polymer solution in air. J. Fluid Mech. 63, 635–640. Jia, Z., Fang, Q., Fang, Z., 2004. Bonding of glass microfluidic chips at room temperatures. Anal. Chem. 76, 5597–5602. Kowsari, K., James, D.F., Papini, M., Spelt, J.K., 2014. The effects of dilute polymer solution elasticity and viscosity on abrasive slurry jet micro-machining of glass. Wear 309, 112–119. Larson, R.G., 1999. The Structure and Rheology of Complex Fluids. Oxford University Press, New York, NY, pp. 107–132. Liao, Y.P., Wang, C.Y., Hu, Y.N., Song, Y.X., 2009. The slurry for glass polishing by micro abrasive suspension jets. Adv. Mater. Res. 69–70, 322–327. Liu, H.T., 2007. Hole drilling with abrasive fluid jets. Int. J. Adv. Manuf. Technol. 32, 942–957. Luo, W.S., Wang, C.Y., Song, Y.X., Liao, Y.P., 2010. Characteristics and polishing effect of abrasive jet beam with polymer abrasive suspension additives. Adv. Mater. Res. 126, 9–13. Miller, D.S., 2004. Micromachining with abrasive waterjets. J. Mater. Process. Technol. 149, 37–42. Neilson, J.H., Gilchrist, A., 1968. Erosion by a stream of particles. Wear 11, 111–122. Nguyen, T., Pang, K., Wang, J., 2009. A preliminary study of the erosion process in micro-machining of glasses with a low pressure slurry jet. Key Eng. Mater. 389, 375–380. Nguyen, T., Shanmugam, D.K., Wang, J., 2008. Effect of liquid properties on the stability of an abrasive waterjet. Int. J. Mach. Tools Manuf. 48, 1138–1147. Nouraei, H., Kowsari, K., Spelt, J.K., Papini, M., 2014. Surface evolution models for abrasive slurry jet micromachining of channels and holes in glass. Wear 309, 65–73. Omrani, H., Jafari, H., Karimpour, A., 2013. Effect of non-Newtonian fluid on the kerf angle of a dailatant jet. Life Sci. J. 10 (2s), 26–34. Rothstein, J.P., McKinley, G.H., 1999. Extensional flow of a polystyrene Boger fluid through a 4:1:4 axisymmetric contraction/expansion. J. Non-Newtonian Fluid Mech. 86, 61–88. Wang, C.Y., Chen, M.D., Yang, P.X., Fan, J.M., 2009a. Hole machining of glass by micro abrasive suspension jets. Key Eng. Mater. 389–390, 381–386. Wang, C.Y., Yang, P., Fan, J.M., Song, Y.X., 2009b. Effect of slurry and nozzle on hole machining of glass by micro abrasive suspension jets. Key Eng. Mater. 404, 177–183. Wang, J., Nguyen, T., Pang, K.L., 2009c. Mechanisms of microhole formation on glasses by an abrasive slurry jet. J. Appl. Phys. 105, 044906.
Contents lists available at ScienceDirect
Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec
Abrasive slurry jet micro-machining of holes in brittle and ductile materials K. Kowsari a , H. Nouraei a , D.F. James a,∗ , J.K. Spelt a,b,∗∗ , M. Papini b,a,∗ ∗ ∗ a b
Department of Mechanical and Industrial Engineering, University of Toronto, 5 King’s College Road, Toronto, ON, Canada M5S 3G8 Department of Mechanical and Industrial Engineering, Ryerson University, 350 Victoria Street, Toronto, ON M5B 2K3, Canada
a r t i c l e
i n f o
Article history: Received 11 September 2013 Received in revised form 3 April 2014 Accepted 10 April 2014 Available online 21 April 2014 Keywords: Hole micro-machining Abrasive slurry jet Polymeric additives Viscosity Fluid elasticity Erosion
a b s t r a c t This paper investigated the effects of elasticity and viscosity, induced by a dilute high-molecular-weight polymer solution, on the shape, depth, and diameter of micro-holes drilled in borosilicate glass and in plates of 6061-T6 aluminum alloy, 110 copper, and 316 stainless steel using low-pressure abrasive slurry jet micro-machining (ASJM). Holes were machined using aqueous jets with 1 wt% 10 m Al2 O3 particles. The 180 m sapphire orifice produced a 140 m diameter jet at pressures of 4 and 7 MPa. When the jet contained 50 wppm of dissolved 8 million molecular weight polyethylene oxide (PEO), the blind holes in glass were approximately 20% narrower and 30% shallower than holes drilled without the polymer, using the same abrasive concentration and pressure. The addition of PEO led to hole cross-sectional profiles that had a sharper edge at the glass surface and were more V-shaped compared with the U-shape of the holes produced without PEO. Hole symmetry in glass was maintained over depths ranging from about 80–900 m by ensuring that the jets were aligned perpendicularly to within 0.2◦ . The changes in shape and size were brought about by normal stresses generated by the polymer. Jets containing this dissolved polymer were observed to oscillate laterally and non-periodically, with an amplitude reaching a value of 20 m. For the first time, symmetric ASJM through-holes were drilled in a 3-mm-thick borosilicate glass plate without chipping around the exit edge. The depth of symmetric blind holes in metals was restricted to approximately 150 m for jets with and without PEO. At greater depths, the holes became highly asymmetric, eroding in a specific direction to create a sub-surface slot. The asymmetry appeared to be caused by the extreme sensitivity of ductile materials to jet alignment. This sensitivity also caused the holes in metals to be less circular when PEO was included, apparently caused by the random jet oscillations induced by the polymer. Under identical conditions, hole depths increased in the order: borosilicate glass > 6061-T6 aluminum > 110 copper > 316 stainless steel. The edges of the holes in glass could be made sharper by machining through a sacrificial layer of glass or epoxy. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Applications of low-pressure abrasive slurry jets, or micro abrasive suspension jets as they are also sometimes called, have
∗ Corresponding author at: Department of Mechanical and Industrial Engineering, University of Toronto, 5 King’s College Road, Toronto, ON, Canada M5S 3G8. Tel.: +1 416 978 3049; fax: +1 416 978 7753. ∗∗ Corresponding author at: Department of Mechanical and Industrial Engineering, University of Toronto, 5 King’s College Road, Toronto, ON, Canada M5S 3G8. Tel.: +1 416 978 5435; fax: +1 416 978 7753. ∗ ∗ ∗Corresponding author at: Department of Mechanical and Industrial Engineering, Ryerson University, 350 Victoria Street, Toronto, ON M5B 2K3, Canada. Tel.: +1 416 979 5000x7655; fax: +1 416 979 5265. E-mail addresses: [email protected] (D.F. James), [email protected] (J.K. Spelt), [email protected] (M. Papini). http://dx.doi.org/10.1016/j.jmatprotec.2014.04.008 0924-0136/© 2014 Elsevier B.V. All rights reserved.
been evolving for over a decade. One important application is in the abrasive slurry jet micro-machining (ASJM) of glass. As a machining process, it is economical and offers advantages such as the absence of a heat-affected zone, low forces on the work piece, no tool wear or vibration, and the ability to machine virtually any material. The small divergence of the jet makes ASJM well suited for precision machining. Miller (2004) developed a high-pressure ASJM system, using a slurry content of approximately 20 wt%, 8 m diameter garnet abrasives, pressures of roughly 70 MPa, and 40–60 m diameter orifices. The system was used to cut a variety of materials, including metals, polymers, glass, and composites, with thicknesses ranging from 50 to 3000 m. The depth to width ratios of the cuts were much greater than those typical of laser cutting. Nguyen et al. (2009) drilled holes in glass using a low-pressure (3 MPa) and a high concentration (8.2 wt%) slurry, and, in a companion paper,
1910
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
Wang et al. (2009a) studied the effects of pressure and machining duration on erosion in holes in glass. In both papers, the hole cross-sections were found to be W-shaped, related to the dominance of ductile erosion when the abrasive particles had relatively low kinetic energy. Nouraei et al. (2014) found that higher kinetic energies led to U-shaped holes with flatter bottoms and steeper sidewalls. Wang et al. (2009b) examined the profiles of holes machined in glass and found them to be asymmetric because of orifice vibration and misalignment. A typical cross-sectional profile shape was divided into various zones, and the shape of each zone was explained in terms of the direction of the slurry flow relative to the walls of the hole. Aqueous jets containing long-chain polymer additives, but not abrasives, have been studied for several decades. Hoyt et al. (1974) discovered that polymers delayed droplet formation and therefore enhanced jet stability by damping free surface disturbances. Another application of polymeric additives is friction reduction in turbulent pipe flow. For example, Elbing et al. (2011) found that trace concentrations of order 10 wppm (weight parts per million) reduced wall friction by about 75%. These advantages resulted from induced elasticity in the fluid. In general, the polymer chains are stretched from their initially-coiled equilibrium state when subjected to a sufficiently high deformation rate. This action generates normal stresses which increase the resistance to extensional deformation; i.e. the fluid behaves partially as an elastic solid, as described by Larson (1999, p. 123). Hashish (1991) compared the cutting of aluminum using highpressure abrasive waterjet machining (AWJM at 345 MPa), in which abrasive particles and air are entrained in a high-velocity jet, and high-pressure abrasive slurry-jet cutting. In both sets of experiments, the water contained 3 wt% 18-million (18-M) molecular weight polyacrylamide (PAM). The slurries were assumed to be Newtonian; therefore, the effects of fluid elasticity induced by the added polymers, as described by Larson (1999, p. 123), were not considered. He found that ASJM produced twice the cutting depth as AWJM using the same abrasive flow rate and pressure. Nguyen et al. (2008) found that concentrated solutions (1000–5000 wppm) of 1-M PAM enhanced the stability and increased the coherent length of an abrasive waterjet. Ashrafi (2011) found that the addition of a large amount of cornstarch (10–22 wt%) in AWJM produced
cuts having narrower kerfs with steeper sidewalls, In another study with cornstarch, Omrani et al. (2013) also found a reduction in kerf taper, and hypothesized that it was due to changes in the fluid viscosity. The role of fluid elasticity generated by the polymer was not considered. In work more closely related to the conditions of low-pressure ASJM than of high-pressure AWJM, Luo et al. (2010) compared low-pressure abrasive slurry jet polishing of glass with and without 4000 wppm (0.4 wt%) of a high-molecular-weight polymer, and found that the polymer sharpened the transition zone separating polished and unpolished regions, because of reduced divergence of the streamlines of the polymeric jet about the stagnation zone. In a companion study, Liao et al. (2009) investigated the effects of additives such as sodium hexametaphosphate (NaP), 4-M polyethyleneglycol, 20-M polyacrilic acid, 12-M PAM, and antimony trinitride on the low-pressure abrasive slurry-jet polishing of glass. They found that a combination of NaP and PAM resulted in the lowest surface roughness and postulated that this effect was due to the lower viscosity of this solution relative to others considered in the study, but did not consider fluid elasticity. Wang et al. (2009c) investigated machining using a variety of polymers such as PAM, anionic polyacrylamide (HPAM), and cationic polyacrylamide (PAMA), and several abrasives (22 m garnet, boron carbide, and white and brown corundum). They found that the slurry containing 6000 wppm 5-M PAM and white corundum particles yielded the most symmetrical holes with the largest material removal rate. No comparisons were made to holes machined without additives. Most recently, Kowsari et al. (2014) studied the effects of dilute concentrations of polyethylene oxide (PEO, 25–400 wppm), on the shape, roughness, and width of low-pressure ASJM channels machined in glass. It was found that, for a constant jet velocity, diameter and dose of abrasive, the channel width decreased by 21% with the addition of 50 wppm of 8-M (PEO). This change was caused by normal stresses generated by the polymer, which focussed the erosion to a smaller blast zone. The objective of the present study was to extend the above research on the effects of long-chain polymers in ASJM to the drilling of holes in ductile and brittle materials, with a focus on the
Fig. 1. (a) Schematic of the ASJM components (not to scale). (b) Orifice geometry (dimensions are in m).
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
1911
Fig. 2. Holes machined using the water-only slurry jet: (a) a 0.5◦ misalignment, and (b) alignment with an accuracy of 0.2◦ .
resulting shape, erosion rate, depth and diameter. As in the previous work, it was of interest to separate the effects of slurry elasticity and viscosity on the observed patterns and rates of erosion. The work also investigated the potential of sacrificial polymeric or glass surface layers to reduce hole opening diameters. 2. Experimental apparatus and procedures 2.1. ASJM apparatus The ASJM apparatus described in Kowsari et al. (2014) was utilized for all experiments. The basic apparatus, as shown in Fig. 1(a), consisted of an abrasive slurry pump, a pulsation damper connected to an open reservoir tank filled with a premixed aluminum oxidewater slurry, and a sharp sapphire orifice with a diameter of 180 m having a length/diameter ratio of 1.67 (Fig. 1(b)). The positive displacement pump (LCA/M9/11-DC, LEWA Inc., Leonberg, Germany) was set to give flow rates of 1.67 mL/s and 2.34 mL/s (±0.1%), at pressures of 4 and 7 MPa, respectively. The jet diameter was measured optically using a digital camera attached to a microscope having a field of view of 3 mm × 2 mm, and was found to be a uniform 140 m over the 30-mm standoff distance (distance between the orifice plate and the target) as shown in Fig. 1(c). The vena contracta could not be observed because it occurred within the 10 mm long orifice stem. The contraction coefficient (the ratio of the jet cross-sectional area to that of the orifice) was found to be 0.60 ± 0.03. This value conforms to the theoretical value of 0.64 for a sharp orifice as explained by Falkovich (2011). The slurry was pumped to the orifice through a 3.2-mmdiameter tube. The flow velocity was not quite high enough to completely eliminate particle settling, and so the abrasive concentration delivered to the target was 10% less than that in the mixing tank. However, this concentration was constant over the duration of the experiments, as confirmed by collecting slurry samples in a beaker during one-minute periods and then drying them to yield the abrasive content.
Fig. 3. Nozzle guide to maintain constant standoff distance and alignment with the target.
Newtonian. When glycerin was added, the fluid remained Newtonian, as explained by Larson (1999, p. 107). In contrast, the slurry with PEO was non-Newtonian, since the long-chain molecules can induce significant elasticity in the fluid, such as causing resistance to extensional deformation, as described by Bird et al. (1987, p. 637). The concentrations of glycerin and PEO were selected so that both solutions had a viscosity 10% greater than that of water. The abrasive particles were taken from the supplier’s container and subjected to the ASTM quartering technique
2.2. Slurry preparation Three different aqueous slurries were prepared, all having a 1% concentration of alumina (Al2 O3 ) abrasive particles with a nominal diameter of 10 m (Comco Inc., CA, USA). One slurry contained only water and abrasive, another also contained 5.2 wt% glycerin, and in the third 50 wppm of 8-M PEO was dissolved. The glycerin concentration was selected to yield a viscosity equal to that of the PEO solution, as measured with a Cannon-Fenske capillary tube viscometer, following the procedure of ASTM D4889-04 (ASTM, 2011). Because of the small concentration of particles, the basic slurry was
Fig. 4. Fourier transform plot of the amplitude of the lateral jet oscillations obtained with a slurry containing 50 wppm of 8-M PEO.
1912
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
Fig. 6. Profiles of holes machined in glass under identical conditions with the wateronly slurry at 1.67 mL/s.
controlled using a shutter to intercept the jet. Five holes were machined separately under identical conditions with the wateronly slurry, and the maximum variation in their depths was less than 3%.
Fig. 5. Microscope images of a hole machined using the water-only jet (a) hole opening in the plane of the target surface, and (b) cross-sectional profile.
(ASTM C702-98, 2003), to ensure a uniform distribution. The sample portions were weighed on an electronic scale (AP110, OHAUS Corp., Pine Brooke, NJ, USA) accurate to 0.1 mg. The PEO had a viscosity-averaged molecular weight of 8.0 M, according to the supplier (Sigma-Aldrich, St. Louis, MO, USA). The selected PEO concentration was based on the earlier work of Kowsari et al. (2014), who found that it produced the minimum ASJM channel width. To prepare the solution, the PEO was first dispersed in 20 mL of ethanol to avoid aggregation, and then this mixture was added to 6 L of slurry in the reservoir tank and stirred slowly to minimize polymer degradation. 2.3. Machining The process parameters were selected to provide holes with shapes and diameter-to-depth ratios similar to those in other ASJM studies conducted without additives, such as the study by Nouraei et al. (2014). Holes were machined in 100 mm × 50 mm × 3 mm borosilicate glass plates (Borofloat 33® , Schott Inc., NY, USA) at a standoff distance of 30 mm. The duration of machining was
Fig. 7. (a) Depth, frosting diameter, surface diameter, and average-diameter normalized by the jet diameter as a function of machining time, and (b) rates of change of these dimensions as a function of machining time. Holes machined in glass with the water-only slurry at 1.67 mL/s. The lines are only to help guide the eye.
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
1913
Fig. 8. (a) Hypothesized streamlines of the flow field during the machining process. (b) Magnified regions of the boxes in (a) to illustrate the trajectories and impact angles of particles from the incident slurry of the jet, , and the return slurry, , relative to the surface and particle impact angle, ˛, near the stagnation point.
During initial tests, it was found that accurate alignment of the jet normal to the target plane was crucial to obtain symmetric holes; e.g. a misalignment of even 0.5◦ produced a relatively large asymmetry, as shown in Fig. 2(a). By placing the three-pointguide in Fig. 3 in contact with the orifice plate and target surface, the misalignment was reduced to 0.2◦ , optically measured using a microscope having a field of view of 3 mm × 2 mm. The resulting machined holes were found to be symmetric, as shown in Fig. 2(b).
2.4. Elasticity-induced jet oscillation Kowsari et al. (2014) noted that the addition of PEO, but not the glycerin, caused the jet to oscillate laterally with an amplitude that increased with the concentration of dissolved polymer. This behavior was investigated further for the present 50 wppm 8-M PEO solution by capturing microscope images having a field of view of 3 mm × 2 mm for 10 s, at a frequency of 29 Hz, and then extracting the position of the jet from each frame. The jet diameter was found to be constant at 140 m, but the jet oscillated laterally with a varying amplitude of up to 20 m, producing an effective jet diameter of 180 m. Without the PEO, the jet had the same diameter (140 m), but did not oscillate. Furthermore, the jet containing glycerin did not oscillate. The frequency spectrum in Fig. 4 shows that there was no dominant frequency in the jet oscillation. This is consistent with the work of Rothstein and McKinley (1999), who found that oscillations in non-Newtonian jets can result from randomly fluctuating vortices formed upstream of a sudden contraction, such as that of the orifice in Fig. 3.
3. Results and discussion 3.1. Holes without additives 3.1.1. Reference blind holes Fig. 5 shows an example of a hole machined in glass using the water-only slurry without PEO or glycerin, at a flow rate of 1.67 mL/s for 8 min. The side walls of the hole merged with the target surface through a curved region that ended in a frosted zone on the surface. Crosssectional profiles of the holes were viewed through an adjacent optically-clear edge in the glass plate using a microscope having a field of view of 2000 m × 1500 m, and were quantitatively characterized using an image analysis system (Clemex Vision PE, Clemex Technologies Inc., QC, Canada) and digital software (ImageJ software—http://rsb.info.nih.gov/ij/). Fig. 6 illustrates how the profiles machined with the water-only slurry at 1.67 mL/s developed with machining time, with the opening width at the surface increasing more quickly than the width at the sidewalls. This is shown in Fig. 7(a) which compares four parameters, all normalized by the jet diameter: (i) hole depth, (ii) frosting diameter; i.e., the limit of the lightly pitted surface around the hole, (iii) surface diameter; i.e. the boundary of the profile where the slope of the linear fit to 5 consecutive profile points, about 5 m apart, was < 10%, and (iv) average diameter; i.e., the cross-sectional area of the hole divided by its depth. It is seen that the hole depth increased less than linearly with machining time, consistent with the findings of Nouraei et al. (2014). It is hypothesized that the decline in the rate of depth enlargement was due to the decreasing flow and particle velocity as the depth of the hole increased. As in Figs. 6 and 7(a) reveals that the
1914
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
relative increases in the frosting diameter and surface diameter with machining time were much more than that of the average diameter. Fig. 7(b) shows that the frosting diameter grew much more rapidly than all other hole dimensions, particularly in the early stages of machining. Beyond about 2 min, the rates of change of the hole depth, the frosting diameter and the surface diameter were nearly the same, all decreasing approximately linearly with time. This can be explained in terms of the three zones found by Wang et al. (2009c) (Fig. 5), the shapes of which can be described by the flow and erosion mechanism illustrated in Fig. 8. The depth depends on the erosion in zone A, the average diameter depends mostly on erosion of the sidewalls (i.e. the upper part of zone A and zone B), and the surface diameter depends on the erosion in zone C. It is hypothesized that zone A evolved due to erosion at the high impact angles, ˛, of the deflecting flow as shown in Fig. 8(b). The relatively lower material removal in the sidewalls was due to two effects: (i) the shallower impact angles, (Fig. 8(b)), of the return slurry traveling toward the target surface would reduce erosion of the sidewall since glass erosion is maximum at 90◦ as described by Neilson and Gilchrist (1968), and (ii) the reduced particle kinetic energy after the initial impact in zone A would also decrease erosion. Since the depth was developed by higher-velocity slurry at high impact angles, its growth rate was larger than that of the average diameter as shown in Fig. 7(b). As the holes became deeper at longer machining times, the velocity of the flow at the bottom of the hole declined, consistent with the reduction in the rate of change in both the depth and average diameter. As shown in magnified region (i) of Fig. 8(b), the surface curvature in zone C at the mouth of the hole was primarily created by a high-velocity and nearly perpendicular incidence, , of the flow component originating from the peripheral streamlines near the surface of the jet. Therefore, the surface diameter (zone C) was developed by a flow having a higher normal impact velocity than the average diameter (zones A and B), consistent with the larger growth rate of the surface diameter relative to the average diameter, seen in Fig. 7(a). Subsequent to the erosion in zone C, the deflected slurry flowing over the target surface caused light pitting and scratching, but negligible erosion, as it impacted at shallow angles of attack, thereby creating the frosted region which developed very quickly (Fig. 7). Therefore, both the frosting diameter and the surface diameter were created by the high-velocity streamlines; however, the growth of the surface diameter required removal of relatively large amounts of material compared to the light pitting in the frosting region, consistent with the much larger growth rate in the frosting diameter in Fig. 7(b). As the hole diameter became large relative to the jet diameter at longer machining time holes, less of the erosive energy of the jet reached the hole edges, resulting in a reduction in the rate of change in both the surface diameter and frosting diameter as seen in Fig. 7(b). The ASJM of micro-channels was much less sensitive to the alignment of the slurry jet. For example, an orifice misalignment of approximately 0.5◦ produced no asymmetry in channel crosssections (Kowsari et al., 2014). In that case, the slurry was relatively free to flow away from the primary impact zone, and so secondary erosion of the channel walls was small. In contrast, the flow within a hole was highly confined, so that secondary erosion of the hole walls was much more significant and susceptible to small changes in the direction of the incident jet. 3.1.2. Through-holes The machining time at a flow rate of 1.67 mL/s was increased to 90 min to create a through-thickness hole on a 3-mm-thick plate of borosilicate glass. As shown in Fig. 9(b), the exit of this pierced hole was circular and without edge chipping. In contrast, Hashish (1988) observed chipping at the exit edges of holes drilled using highpressure AWJM, which he attributed to sub-surface stress waves
Fig. 9. (a) Comparison of profiles of a hole machined in glass with the water-only slurry for 8 min and a through-hole after 90 min, at a flow rate of 1.67 mL/s. (b) Microscope image of the exit of the through-hole.
created by the impact of the jet. The 180 m diameter of the exit hole was larger than the jet diameter (140 m), probably because of flow near the stagnation point, as illustrated in Fig. 8. It was also observed that the initial breakthrough was followed by about three seconds of rapid widening of the exit hole before the jet could pass through the target without interference and the hole diameter became stable. As illustrated in Fig. 9, the cross-section of the entrance region continued to broaden and the radius of curvature grew with time as the hole was drilled. 3.2. Comparison of holes in brittle and ductile materials Fig. 10(a) superimposes the profiles of shallow holes machined in glass, 316 stainless steel, 110 copper, and 6061-T6 aluminum alloy. As shown, there were significant differences in the shapes and sizes of the four holes even though they were all machined at a flow rate of 2.34 mL/s for 10 s. Of particular interest is the much sharper definition of the opening curvature with the metals, when compared to the larger radius of curvature in glass. As described
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
1915
Fig. 10. (a) Profiles of holes machined in brittle and ductile materials under identical conditions, using the water-only slurry at a flow rate of 2.34 mL/s; (b) microscope images of the hole openings machined under identical conditions in (i) 6061 aluminum, (ii) 110 copper, (iii) 316 stainless steel, and (iv) borosilicate glass.
in Section 3.1, particle impacts at a high impact angle () are dominant near the opening of holes. Since the erosion of brittle materials such as glass is maximum at a perpendicular incidence, these materials have highly eroded hole entries. Ductile materials, however, have lower relative erosion rates at high impact angles which resulted in much less erosion in zone C at the mouth of the hole. This is again consistent with the slurry erosion patterns described by Neilson and Gilchrist (1968) for brittle and ductile materials. In contrast to glass, for metal holes, the shallow impact angle () of the low-velocity return slurry (Fig. 8(b)) dominated erosion in the entry region, causing the holes to become wider without significantly increasing the radius of curvature at the edge. This effect is consistent with the findings of Liu (2007), who compared high-pressure AWJM holes drilled in glass and aluminum, and found considerable wear in the vicinity of the entrance in holes machined in glass, but not in aluminum. Despite this difference in opening curvatures, surface pitting was observed to generate the frosted region in both the metals and glass (Fig. 8(b)), which was
caused by slurry flowing over the target at relatively small impact angles, as described in Section 3.1. As seen from Figs. 2 and 6, it was possible to machine deep symmetrical holes in the borosilicate glass because the ASJM erosion process was stable. With metals though, the process became unstable at aspect ratios of approximately one, and a radial groove or slot would extend from the edge of the hole. This instability was investigated further using transparent polymethylmethacrylate (PMMA), which eroded in a similar ductile manner as the metals. Fig. 11, obtained under the same conditions as those in Fig. 10, shows that the hole was asymmetrical and that a deep groove formed. The erosion rate increased upon the onset of groove formation at a depth of approximately 150 m because the return slurry began to flow away from the incoming jet via the groove, whereas in holes without grooves, the interference between incident and return slurry dissipated the jet’s impact energy. This is demonstrated by the rate of deepening in aluminum, which increased from approximately 14–27 m/s subsequent to groove formation.
1916
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
The origin of the asymmetry which led to the formation of the grooves was investigated by comparing a hole drilled at a reference angular position about the jet centerline with another hole machined after rotation by 90◦ relative to the reference angle. Comparisons confirmed that the direction of the grooves was consistent with that of jet misalignment. As described in Section 2.3, when the jet was within 0.2◦ of 90◦ to the surface, there were no irregularities in the holes machined in glass, which were presented in Section 3.1. Therefore, drilling in ductile materials was apparently much more sensitive to jet alignment, probably because the erosion rate was maximum at an impact angle of approximately 30◦ in metals and 90◦ in glass. If the jet was perfectly aligned, the return flow from the bottom of a hole was approximately parallel to the vertical side walls in zones A and B (Fig. 8), and the local particle impact angle was small, thereby causing relatively little widening of the hole compared with the increase in depth (compare Fig. 7(a) and (d)). In metals, it is hypothesized that perturbations of the jet flow toward one side of the hole would produce a relatively large increase in the erosion rate on that side as the local impact angle approached 30◦ . Once a groove started to form, its erosion was accelerated by the concentration of flow in the direction of the groove and, as mentioned above, the decreasing resistance to incoming slurry flow as the slot developed. In contrast, the glass was more tolerant of small misalignments since the erosion rate on the sides of holes was relatively small at shallow impact angles. 3.3. Effects of fluid elasticity
Fig. 11. (a) Cross-sectional profile and (b) microscope image of the opening of a hole machined in PMMA, using the water-only slurry. The cross-sectional profile in (a) corresponds to the view in the direction of the white arrow in (b).
Kowsari et al. (2014) found that the width of an ASJM channel in glass was affected by the addition of polymer to the slurry. Specifically, the width decreased by 21% when 50 wppm of 8-M PEO was dissolved in the slurry compared to a channel machined without polymer. In general, dissolving a polymer causes two effects: it induces elasticity and thus generates normal stresses in the fluid, and it increases the viscosity. As explained in Section 2.2, these two effects were separated by using an aqueous glycerin solution having the same viscosity as the PEO solution; i.e. both solutions were 10% more viscous than water.
Fig. 12. (a) Hole cross-sectional profiles, and (b) microscope images of the openings of shallow holes machined in glass under identical conditions with the water-only slurry (left), the glycerin slurry (middle) and the PEO slurry.
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
1917
Fig. 13. (a) Depth versus machining time, and (b) frosting diameter, (c) surface diameter, and (d) average diameter versus depth for holes machined in glass with and without 50 wppm of 8-M PEO. The machining times are included in brackets. The dashed lines are only to help guide the eye.
3.3.1. Holes in glass Fig. 12(a) presents the profiles of shallow holes machined with the water-only slurry compared with the glycerin slurry and the polymer slurry, at 1.67 mL/s and 5 s duration. The 10% increase in viscosity with glycerin slurry reduced the depth by 6%, but it did not significantly change the profile shape, the average diameter, nor the size of the frosted zone (Fig. 12(b)). The depth of the PEO hole decreased because of the increased viscous drag which further decelerated the particles, thus reducing their ability to erode, as described by Clark (1992). In addition to this viscosity-induced reduction, the depth of the PEO hole was reduced an additional 17% by the polymer. The additional decrease in erosion was due to
further particle deceleration as the long-chain polymer molecules induced tensile stresses in the fluid in front of a particle, thereby retarding its ability to flow out of the closing gap between the particle and the target. The added polymer also changed the shape of the hole, from the flat-bottomed U-shape without polymer to more of a rounded V-shape, as illustrated in Fig. 12(a). These trends are illustrated as a function of machining time and hole depth in Fig. 13 for both the water-only and the PEOcontaining slurries. Fig. 13(a) shows that the 10% increase in the viscosity of the polymer solution reduced the depth by an average of 29%, over an 8-minute period, presumably by increasing the drag
Fig. 14. Microscope images of hole openings in metals with and without PEO, at 2.34 mL/s for 10 s.
1918
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
on the abrasive particles. Fig. 13(b), (c), and (d) show that for samedepth holes from 84 to 704 m, the frosting, surface, and average diameters decreased by averages of 25%, 14%, and 15% respectively, when the 50 wppm of 8-M PEO slurry was used. These changes in depth, diameter, and shape are similar to those found in channels machined in glass using PEO-containing abrasive slurries (Kowsari et al., 2014). As discussed above, the normal stresses generated by the PEO caused the fluid to resist deformation, especially in regions of extensional motion such as stagnation point flow. That is, the polymer solution acted as a viscoelastic material, resisting extension as well as shear, as described by Bird et al. (1987, p. 637). As described previously, this extra resistance near the stagnation zone (Fig. 8) caused the primary area of jet erosion to be smaller since the streamlines were more focussed, and caused greater particle deceleration at the bottom of the hole. Fig. 13(b) shows that the reductions in the frosting diameters and surface diameters produced by the PEO were larger than in the average diameters. This is because the deformation rate was larger near the opening region (zone C), in which the erosion was caused by a high-velocity slurry, than in the bottom of holes (zones A and B), which were developed by slower-moving slurry as described in Section 3.1. This resulted in greater resistance to spreading of the streamlines near the opening, and thus larger relative reductions in the frosting and surface diameters than in the average diameter. The effects of elasticity were generally smaller with deeper holes. For example, it was found that, whereas PEO reduced the frosting, surface, and average diameters by 41%, 19%, and 13% in the 125-m deep holes compared with the water-only slurry, these reductions were only 15%, 10%, and 6%, respectively, for 700 m holes, since the larger extensional rates that occur in shallow holes generated more resistance to motion. The extensional rate decreased as the holes became deeper, because of a decrease in the velocity within the cavity, consistent with the decreasing rate of depth evolution with machining time shown in Fig. 13(a). 3.3.2. Holes in metals In order to avoid the instability discussed in Section 3.2, the effect of fluid elasticity on holes machined in the three metals was examined only for relatively shallow holes less than 150 m. As observed with glass in Fig. 13(a), the addition of PEO reduced the depths in the metals by approximately 20% compared to the water-only holes. Although the holes machined with the PEO slurry had smaller average openings, they were highly irregular as seen in Fig. 14. This irregularity was tested for directional consistency by rotating the orifice as described in Section 3.2. The results indicated that, contrary to what was found for water-only slurry holes in ductile materials, the asymmetry was not always related to the direction of the jet in the case of the PEO slurry. Therefore, it is likely that the asymmetry resulted from the random oscillations of
Fig. 15. Profile of a hole machined in glass through a sacrificial layer.
the jet when it contained PEO, as discussed in Section 2.4. This jet instability led to transient jet asymmetry which, combined with the sensitivity of the ductile erosion mechanism to changes in the impact angle along the side walls, quickly caused the holes to become non-circular. 3.4. Machining through a sacrificial surface layer It was of interest to see if a sacrificial layer on top of the glass target could be used to decrease the hole opening diameter and reduce or eliminate the frosted zone. Fig. 8 illustrates the flow within a hole and the edge rounding developed over the machining duration. This was investigated by covering the glass target surface with two materials in order to minimize the opening diameter in the target plate, as shown in Fig. 15. Four sacrificial layers were tested: (i) a 500 m coating of epoxy adhesive (J-B Weld, Sulphur Springs, Texas, USA); (ii) an 800 m coating of the same material; (iii) a 150 m thick borosilicate glass cover slip bonded to the target using a room-temperature process (Jia et al., 2004); and (iv) the same cover slip glued to the target using cyanoacrylate adhesive. Thicknesses were measured using a digital caliper accurate to ±20 m. Fig. 16 compares hole openings with the four layers to an opening without a layer, all machined using the water-only slurry. The micrographs show that the sacrificial layers significantly reduced the diameter of the frosted zone and hence the effective opening diameter; for example, the reduction was 38% in the case of the 500-m-thick epoxy layer. However, the hole machined with the glued cover slip had irregular edges, as shown by (d) and (e), presumably because of leakage and erosion at the interface between the bonded glass surfaces. Fig. 17(a) compares the cross-sectional profiles of the holes drilled for 8 min through the two epoxy layers, showing that
Fig. 16. Microscope images of the openings of holes machined in glass using the water-only slurry at 1.67 mL/s for 8 min: (a) without sacrificial layer, (b) with 500 m thick epoxy, (c) 800 m thick epoxy, (d) bonded glass sheet, and (e) glued glass sheet.
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
1919
Fig. 17. Cross-sectional profiles of holes machined in glass at a flow rate of 1.67 mL/s for 8 min without a sacrificial layer and using (a) 500 and 800 m-thick epoxy coatings, and (b) water-bonded and glued glass sheets. (c) Profiles of holes machined at a flow rate of 1.67 mL/s for 6 min without a sacrificial layer and a hole machined for 8 min using a 150 m thick bonded glass sheet.
the epoxy coating decreased the radius of the opening curvature (increased the edge sharpness), but did not eliminate this rounding even when the coating thickness was increased. This was due to the relatively high erosion rate of the epoxy, which reduced its effectiveness as a barrier to erosion on the glass surface. When the coating material was changed to glass, the edge was significantly sharpened, as shown in the profiles presented in Fig. 17(b). However, leakage between the glass cover plate and the target caused the profile of the hole opening to be asymmetric regardless of the bonding method, as indicated in the upper part of the figure. As expected, the coatings reduced the depths because of the time required to drill through the sacrificial layers. This is illustrated in Fig. 17(c) by the similar depths of the 6-minute hole in the bare target and the 8-minute hole through the target covered with the 150-m-thick glass cover plate; i.e. approximately 2 min was required to drill through 150 m of glass. 4. Summary and conclusions An abrasive slurry jet micro-machining (ASJM) system was used to investigate the effects of elasticity and viscosity, brought about by a dilute high-molecular-weight polymer solution, on the shape, diameter, and depth of micro-holes drilled in glass,
various metals, and a plastic. The jet incidence was aligned to be perpendicular to a target surface to within 0.2◦ in order to avoid asymmetry in machined holes. The effects of viscosity and elasticity, both induced by an added polymer, were separated by using an aqueous glycerin solution of an equivalent viscosity. Jets containing a polymer, but not glycerin, were found to oscillate laterally and non-periodically, increasing the effective diameter by approximately 22%. Holes were machined with and without additives under fixed machining conditions. To identify the role of PEO, reference blind holes with machining times ranging from 5 to 480 s were machined in glass with the water-only slurry. The frosting diameter; i.e., the limit of the lightly eroded surface around a hole, surface diameter; i.e. the boundary of the cross-sectional profile at which the slope of the linear fit to 5 consecutive profile points, about 5 m apart, was < 10%, and average diameter, represented by the cross-sectional area of a hole divided by its depth, were all found to increase with depth as the incoming and return slurries progressively eroded the sidewalls and openings of holes over the machining time. For the first time, ASJM of through-holes revealed that the circular exit had a clean exit edge, which continued to grow to a diameter approximately 20% larger than that of the jet, subsequent to the initial breakthrough because of deflecting streamlines at the bottom of the
1920
K. Kowsari et al. / Journal of Materials Processing Technology 214 (2014) 1909–1920
holes. A comparison of holes machined under identical conditions demonstrated that the material removal rate was higher in glass than in metal holes. The symmetry of holes in metals, and not glass, was limited to a depth of approximately 150 m by sub-surface slots that developed because the ductile erosion mechanism caused the erosion of sidewalls to be much more sensitive to jet incidence alignment than in glass. When 50 wppm (weight parts per million) of 8-M (million) molecular weight polyethylene oxide (PEO) was dissolved in the slurry, the U-shaped profiles of the reference holes in glass became more V-shaped. This change was accompanied by decreases in the depths, frosting diameters, and average diameters by averages of 25%, 14%, and 15% respectively, over depths ranging from 84 to 702 m. It is thought that these changes were brought about by fluid elasticity which reduced the spread of the jet in the stagnation region. In ductile materials, the holes were found to be irregular because of the combined effects of elasticity-induced jet oscillations and sensitivity of sidewall erosion to jet incidence. Machining through an epoxy coating as a sacrificial layer reduced the frosting diameter by about 40% and sharpened the opening curvature. Similar changes were seen using bonded and glued borosilicate glass cover slips, despite leakage at the interface. In contrast to the asymmetrical ASJM holes in glass presented the work of Wang et al. (2009c), the present results demonstrate that symmetrical holes can be machined in glass by precise alignment of the jet. The work also revealed that in contrast to the reductions in widths of cuts with the use of concentrated polymer solutions in high-pressure AWJM obtained in the work of Omrani et al. (2013), the use of dilute high-molecular-weight solutions or sacrificial surface layers can significantly decrease the frosting and average diameters of ASJM holes in glass. Acknowledgements The authors acknowledge the Natural Sciences and Engineering Research Council of Canada, Canada Research Chairs, Micralyne Inc. and BIC Fuel Cells Inc. References ASTM C702-98, 2003. Standard Practice for Reducing Samples of Aggregate to Testing Size. American Society for Testing and Materials (ASTM), http://www.astm.org/database.cart/historical/C702-98R03.htm ASTM D4889-04, 2011. Standard test methods for polyurethane raw materials: determination of viscosity of crude or modified isocyanates. American Society for Testing and Materials (ASTM), www.astm.org/Standards/D4889.htm
Ashrafi, N., 2011. Viscoelastic abrasive waterjet. Int. Mech. Eng. Congress Exposition 3, 677–681. Bird, R.B., Amstrong, R.C., Hassager, O., 1987. Dynamics of Polymeric Liquids, vol. 1, 2nd ed. John Wiley & Sons, New York, NY, pp. 637. Clark, H.M.I., 1992. The influence of the flow field in slurry erosion. Wear 152, 223–240. Elbing, B.R., Solomon, M.J., Perlin, M., Dowling, D.R., Ceccio, S.L., 2011. Flow-induced degradation of drag-reducing polymer solutions within a high-Reynolds-number turbulent boundary layer. J. Fluid Mech. 670, 337–364. Falkovich, G., 2011. Fluid Mechanics – A Short Course for Physicists. Cambridge University Press, Cambridge, England. Hashish, M., 1988. Turning, milling, and drilling with abrasive water jets. In: Woods, P.A. (Ed.), 1988 Proc. 9th Int. Symp. Jet Cutting Techn., BHRA Fluid Engineering. Cranfield, United Kingdom, pp. 113–131. Hashish, M., 1991. Cutting with high-pressure abrasive suspension jets. In: 1991 Proc. 6th Int. Symp. Jet Cutting Techn, WJCT, Houston, Texas, pp. 439–455. Hoyt, J., Taylor, J., Runge, C.D., 1974. The structure of jets of water and polymer solution in air. J. Fluid Mech. 63, 635–640. Jia, Z., Fang, Q., Fang, Z., 2004. Bonding of glass microfluidic chips at room temperatures. Anal. Chem. 76, 5597–5602. Kowsari, K., James, D.F., Papini, M., Spelt, J.K., 2014. The effects of dilute polymer solution elasticity and viscosity on abrasive slurry jet micro-machining of glass. Wear 309, 112–119. Larson, R.G., 1999. The Structure and Rheology of Complex Fluids. Oxford University Press, New York, NY, pp. 107–132. Liao, Y.P., Wang, C.Y., Hu, Y.N., Song, Y.X., 2009. The slurry for glass polishing by micro abrasive suspension jets. Adv. Mater. Res. 69–70, 322–327. Liu, H.T., 2007. Hole drilling with abrasive fluid jets. Int. J. Adv. Manuf. Technol. 32, 942–957. Luo, W.S., Wang, C.Y., Song, Y.X., Liao, Y.P., 2010. Characteristics and polishing effect of abrasive jet beam with polymer abrasive suspension additives. Adv. Mater. Res. 126, 9–13. Miller, D.S., 2004. Micromachining with abrasive waterjets. J. Mater. Process. Technol. 149, 37–42. Neilson, J.H., Gilchrist, A., 1968. Erosion by a stream of particles. Wear 11, 111–122. Nguyen, T., Pang, K., Wang, J., 2009. A preliminary study of the erosion process in micro-machining of glasses with a low pressure slurry jet. Key Eng. Mater. 389, 375–380. Nguyen, T., Shanmugam, D.K., Wang, J., 2008. Effect of liquid properties on the stability of an abrasive waterjet. Int. J. Mach. Tools Manuf. 48, 1138–1147. Nouraei, H., Kowsari, K., Spelt, J.K., Papini, M., 2014. Surface evolution models for abrasive slurry jet micromachining of channels and holes in glass. Wear 309, 65–73. Omrani, H., Jafari, H., Karimpour, A., 2013. Effect of non-Newtonian fluid on the kerf angle of a dailatant jet. Life Sci. J. 10 (2s), 26–34. Rothstein, J.P., McKinley, G.H., 1999. Extensional flow of a polystyrene Boger fluid through a 4:1:4 axisymmetric contraction/expansion. J. Non-Newtonian Fluid Mech. 86, 61–88. Wang, C.Y., Chen, M.D., Yang, P.X., Fan, J.M., 2009a. Hole machining of glass by micro abrasive suspension jets. Key Eng. Mater. 389–390, 381–386. Wang, C.Y., Yang, P., Fan, J.M., Song, Y.X., 2009b. Effect of slurry and nozzle on hole machining of glass by micro abrasive suspension jets. Key Eng. Mater. 404, 177–183. Wang, J., Nguyen, T., Pang, K.L., 2009c. Mechanisms of microhole formation on glasses by an abrasive slurry jet. J. Appl. Phys. 105, 044906.