Common Machining Strategies

CHAPTER

COMMON MACHINING STRATEGIES

6

CHAPTER OUTLINE 6.1 General Overview .........................................................................................................................117 6.2 Gravity-Feed Drilling .....................................................................................................................119 6.2.1 Discharge Regime ....................................................................................................122 6.2.2 Hydrodynamic Regime ..............................................................................................122 6.2.3 Repeatability of Drilling ............................................................................................123 6.2.4 Drilling Time............................................................................................................123 6.2.5 Influence of the Interelectrode Resistance ..................................................................125 6.2.6 Microhole Dimensions...............................................................................................127 6.2.7 Machining Quality ....................................................................................................129 6.3 Constant Velocity-Feed Drilling .....................................................................................................134 6.3.1 Forces Exerted on the Tool-Electrode .........................................................................136 6.3.2 Nature of Contact Forces in Glass Drilling...................................................................139 6.3.3 Tool‒Workpiece Gap .................................................................................................142 6.4 2D and 3D Machining ...................................................................................................................144 6.4.1 Quality of Machined Microchannels ...........................................................................146 6.4.2 Maximal Allowed Tool Travel Speed ...........................................................................149 6.4.3 Depth of Machined Microchannels .............................................................................150 6.4.4 Influence of Tool Distance from Workpiece .................................................................151 6.5 Wire Electrochemical Discharge Machining ................................................................................... 152

Spark-assisted chemical engraving (SACE) can be used to machine holes, 2D and 3D structures. Today, SACE is mostly used for glass microdrilling, which was the first application developed for this technology. Much less developed SACE machining strategies are wire electrochemical machining and the recently introduced lathe-type machining. SACE is also applied to materials other than glass, but compared with glass micromachining, knowledge is still very scanty.

6.1 GENERAL OVERVIEW During SACE, the heat source produced by the electrochemical discharges must be in close proximity to the workpiece. Typically, a maximal distance of 25 mm from the workpiece is allowed in the case of glass (Fascio et al., 1999). Figure 6.1 gives an overview of the effect of the tool‒workpiece distance (Abou Ziki, 2014). When the tool-electrode is touching the workpiece, a full disk structure gets machined. Thus, the tool-electrode acts as a disk heat source as in Eqn (5.2). As the tool-electrode is moved away from the workpiece surface, rings are machined and the heat source becomes a ring as in Eqn (5.3). This occurs because the electrochemical discharges emitted at the sharp edges of the tool (its Micromachining Using Electrochemical Discharge Phenomenon. http://dx.doi.org/10.1016/B978-0-323-24142-7.00006-8 Copyright © 2015 Elsevier Inc. All rights reserved.

117

118

CHAPTER 6 COMMON MACHINING STRATEGIES

0 μm

2 μm

4 μm

8 μm

12 μm

14 μm

16 μm

20 μm

200 μm

FIGURE 6.1 Machined glass workpiece surface for different tool‒workpiece gaps (0–20 mm). The surface is machined while keeping the 500-mm stainless steel tool at fixed gaps in 50 wt% KOH while applying 29 V during 2 s (pulse on-time 20 ms and 1 ms off-time) (Abou Ziki, 2014).

circumference) heat the workpiece. For very small distances (2 mm in the case of Figure 6.1), the high heat transfer resistance results in inefficient machining because of trapped bubbles in the gap. As the tool is moved further away, machining becomes more efficient. Note that as soon as the tool no longer contacts the glass surface, the machined surface is smaller in diameter than in the case of zero gap. In fact, it eventually becomes even smaller than the tool diameter for high gaps. Consequently, different machining behaviors will be achieved depending on the tool feeding mechanism. The simplest feeding mechanism is gravity feed, in which a constant force is applied to the tool-electrode to guarantee mechanical contact with the workpiece. In the second option, constant velocity feed, the tool-electrode is moved at a constant feed-rate lower than the mean material removal rate, to maintain a gap between tool and workpiece. The third option is to control the tool feed as a function of the status of the actual machining process; although this last strategy is highly desirable, it has never been achieved to date. Preliminary results were published recently by Morrison et al. (2008). The first two feeding mechanisms, gravity feed and constant velocity feed, can yield excellent results, although they are actually open-loop strategies. Gravity feed is a very popular strategy for drilling because of its simplicity and other advantages that are discussed in Section 6.2. For 2D machining, constant velocity feed is the most popular strategy. As discussed in Section 5.1, the tool feeding mechanism during machining is only one of the elements affecting the material removal rate and surface quality. These factors are also influenced by the locally generated heat power, mass transport, and the temperature in the machining zone. The locally generated heat power is determined by the discharge activity inside the gas film, which is controlled by the machining voltage. Its value has to be adjusted depending on the desired machining performances and the workpiece type. Typical values for glass are approximately 30 V, and for ceramics and composite materials higher values of about 50–80 V are generally needed. An important issue is the stability of the heat source during the machining process; because stability cannot generally be achieved, two alternative strategies are used. The first option is to machine at the lowest possible power. In this case, an unstable discharge activity will not affect the machining too dramatically. As expected, the main drawback is the reduced material removal rate. In the second case, one tries to take control of the gas film. As the stability of the gas film cannot be controlled directly over a long period of time, a solution is building and destroying the gas film periodically in a controlled

6.2 GRAVITY-FEED DRILLING

119

FIGURE 6.2 Schematic of three cases relevant to SACE machining: (a) for deep holes and tool
glass contact, the machining temperature is similar to that of the tool (TM ¼ Ttool), (b) for shallow structures (e.g., 2D surfaces) the machining temperature is similar to the electrolyte vaporization temperature (TM ¼ Tvaporization), and (c) when a machining gap is formed during drilling, the machining temperature is similar to the melting temperature of the electrolyte salt (TM ¼ Tmelting). Reprinted from Abou Ziki et al. (2014) with permission from Elsevier.

manner. This can be achieved by using a pulsed voltage supply or by adding an external inductive circuit to the power supply, as done for the Wehnelt current interrupter in the last century. Similar to the first strategy, the main drawback is the reduced material removal rate. However, both strategies result in higher machining quality and higher reproducibility of the machining, as will be discussed in Chapter 7. As will be presented in the following sections, machining performance is largely controlled by three characteristic quantities related to the generated heat power, mass transport, and the workpiece material. A first characteristic number related to the locally generated heat power and workpiece material is the time to needed to reach the machining temperature. A second quantity, relevant in the context of drilling and related to mass transport, is the characteristic depth d beyond which flushing the local zone beneath the tool becomes difficult. The third characteristic quantity is the temperature of the machining zone. As will be discussed in the reminder of the chapter, this temperature can be varied with the tool-electrode feeding strategy, machining voltage, and flushing strategy of the machining zone. In the case of gravity feed, this temperature is very similar to that of the tool-electrode (Figure 6.2a). When machining with a gap between the tool and workpiece, this temperature is the vaporization temperature of the used aqueous electrolyte if flushing of the machining zone is efficient, as in 2D machining for example (Figure 6.2b). If flushing of the machining zone is poor, as in constant velocity feed drilling, this temperature is the melting temperature of the electrolyte salt (Figure 6.2c). During SACE machining, tool-electrodes in the form of both rods and wires can be used. The utilization of a wire as a tool-electrode is discussed at the end of the chapter. In this situation, the special geometry of the tool and the machined surface provide additional possibilities for observations of the machining process. Hence, it becomes possible to control the feeding mechanism of the electrode as a function of the gap between the wire and the workpiece.

6.2 GRAVITY-FEED DRILLING In gravity-feed drilling, a constant force F is applied on the tool-electrode to ensure close contact between the heat source (the tool-electrode heated by the electrochemical discharges emitted through the gas film) and the workpiece. As a direct consequence, the temperature in the machining zone will become similar to the tool-electrode temperature. Although this method is particularly simple and

120

CHAPTER 6 COMMON MACHINING STRATEGIES

F

Workpiece FIGURE 6.3 Principle of SACE gravity-feed drilling. The tool is pressed with a constant force F against the workpiece.

gives excellent results, the major drawback is that the mechanical contact between the tool and the workpiece limits flushing and deforms the hole. The principle is explained in Figure 6.3. The tool-electrode is guided vertically and a force is acting on it in the drilling direction. The practical implementation of such a set-up is discussed later in Section 8.2. One important aspect is to reduce, as much as possible, the frictional forces that arise during the vertical guidance of the tool-electrode. During gravity-feed drilling, as the tool-electrode is constantly in contact with the workpiece, drilling depth evolution can be followed by measuring the progressive tool-electrode motion. Therefore, the wear of the tool-electrode will affect this measure. For commonly used electrode materials for SACE, such as nickel or stainless steel (Table 6.1), tool wear is generally low. Typical examples for glass drilling as a function of various machining voltages are shown in Figure 6.4. Drilling was done with a 0.4-mm cylindrical stainless steel cathode in 30 wt% NaOH (Wu¨thrich et al., 2006c). After a first phase, where the drilling speed is fast, a progressive slowdown of the material removal rate is observed until a limiting drilling speed vlim, almost independent of the machining voltage, is reached. The first fast regime is termed the discharge regime, as the discharge activity controls the drilling speed. The second regime, where the limiting speed is reached, is called the hydrodynamic regime, as the ability of the electrolyte to reach the machining zone limits drilling evolution (Wu¨thrich et al., 2006c). Drilling in gravity-feed configuration is characterized by staircaselike evolution of the hole depth in function of drilling time. As glass drilling by gravity feed is the most studied mechanism to date, this technique is discussed in detail, mainly for glass, in the remainder of this section. A simple model for the evolution of the drilling depth z(t) in glass gravity-feed drilling can be given. Due to the progressive transition from the discharge regime to the hydrodynamic regime, the

6.2 GRAVITY-FEED DRILLING

121

Table 6.1 Tool-Electrode Wear for Gravity-Feed Drilling Electrode Material

Voltage (V)

Wear (mm sL1)

Reference

Cu Brass Steel Steel Stainless steel FeeCr NieAg NieCr Ni W W PteIr

34 33 36 28e33 28e33 36 37 37 36 35 28e33 35

0.10 0.12 0.04 0.06 0.02 0.02 0.08 0.02 0.03 0.04 0.05 0.01

Kurafuji and Suda (1968) Kurafuji and Suda (1968) Kurafuji and Suda (1968) Abou Ziki and Wu¨thrich (2012) Abou Ziki and Wu¨thrich (2012) Kurafuji and Suda (1968) Kurafuji and Suda (1968) Kurafuji and Suda (1968) Kurafuji and Suda (1968) Kurafuji and Suda (1968) Abou Ziki and Wu¨thrich (2012) Kurafuji and Suda (1968)

100 0.4 mm

0

Z [μm]

–100 –200

28V

–300 32V –400 –500

29V

35V

0

5

10

15

20

25

30 t [s]

35

40

45

50

55

FIGURE 6.4 Typical evolution at various voltages of SACE glass gravity-feed drilling using a cylindrical tool (cathode) of 0.4-mm diameter with a force of 0.8 N acting on it. The electrolyte (30 wt% NaOH) level above the workpiece is about 1 mm. Reprinted from Wu¨thrich et al. (2006c) with the permission of the Journal of Micromechanics and Microengineering.

122

CHAPTER 6 COMMON MACHINING STRATEGIES

drilling speed v(z) decreases with depth z. If a constant rate with characteristic length d is assumed, one can write (Jalali et al., 2009): dvðzÞ 1 ¼
½yðzÞ
ylim : dz d

(6.1)

As z_ ¼ vðtÞ and using the initial conditions z(0) ¼ 0 and v(0) ¼ vo, one obtains the solution of Eqn (6.1) as:   yo ylim t yo
ylim zðtÞ ¼ d$ln ed
: (6.2) ylim ylim Note that for t / N one gets



 yo : zðtÞ/ylim t þ d ln ylim

(6.3)

The initial speed vo can be computed as described in Section 5.2. By fitting Eqn (6.2) to experimental drilling evolution, the remaining parameters vlim and d can be determined. Typical values for glass machining in 30 wt% NaOH with a cylindrical 0.4-mm stainless steel cathode are about 1.5 mm s
1 for vlim and about 70–80 mm for d (Jalali et al., 2009). For very high depths (more than a few millimeters), the limiting speed vlim vanishes and the model (6.2) is no longer valid. Drilling reaches a limiting depth, which becomes a function of the machining voltage (Cook et al., 1973). For ceramic drilling, the situation is different. The material removal rate is very low. No clear difference between the discharge and the hydrodynamic regimes can be observed, probably because the isotherm of the machining temperature TM progresses very slowly (compare with Figure 5.10). Therefore, the propagation of heat inside the workpiece is the main limiting factor for ceramic drilling.

6.2.1 DISCHARGE REGIME The discharge regime in glass gravity-feed drilling is characterized by a high drilling speed of typically 100 mm s
1 (Wu¨thrich et al., 2006c). It takes place in the first 100–200 mm. The drilling speed is controlled by the discharge activity and is directly related to the machining voltage (Figure 5.9). Machining takes place by etching of the workpiece by the hot electrolyte. The drilling speed is limited by the propagation of heat inside the workpiece and can be estimated using Eqn (5.23) as explained in Section 5.2.3. As the tool-electrode is constantly pressed against the glass workpiece, the temperature of the machining zone will be very close to the tool temperature. Hence, in Eqn (5.23), the machining temperature TM can be assumed to be equal to the tool temperature. The former is related to the machining voltage as discussed in Section 4.5; Chapter 4. Considering the values of TM (approximately 500  C), the electrolyte in the machining zone will be in the form of a molten salt.

6.2.2 HYDRODYNAMIC REGIME As drilling progresses deeper, the limiting factor in the material removal rate is no longer the heat propagation in the workpiece, but the ability of the electrolyte to reach the machining zone and the ability to remove the machined material (Wu¨thrich et al., 2006c). Therefore, a zone of softened glass

6.2 GRAVITY-FEED DRILLING

123

forms at the tip of the tool-electrode. As during gravity-feed drilling, a force pushes the tool toward the glass, and the tool tip penetrates inside this viscous region. The drilling speed vlim is limited by the drag force exerted on the tool: ylim ¼

F ; dðhÞ

(6.4)

where F is the force exerted on the tool by gravity-feed drilling and d(h) is the drag coefficient that is a function of the tool-electrode geometry and the viscosity of the material in the machining zone. Note that h is a function of the temperature in the machining zone. Typical values for the limiting speed are a few micrometers per second. This speed is almost independent of the machining voltage, as even a quite large change in the temperature of the machining zone will result only in a small change of the glass viscosity (Figure 5.1). From the experimentally determined limiting speed for soda lime glass and based on Eqn (6.4), the mean viscosity h of the material in the machining zone can be estimated (Jalali et al., 2009). Typical values are 1.4$108 Pa s. This viscosity corresponds to a temperature of approximately 600  C for glass, similar to the tool temperature. However, it is lower than the softening Littleton point (720  C) which corresponds to the temperature at which a rod of glass increases its length by 1 mm min under its own weight. Note that glass is typically blown at the working point where h ¼ 103 Pa s. Hence, it is difficult to explain how the material could be removed under such conditions without etching. As the etching of the substrate is the main mechanism for material removal in glass micromachining by electrochemical discharges, the lack of electrolyte in the machining zone (small electrolyte–workpiece interface) limits not only the discharge activity, but also the machining rate. As discussed below, the hydrodynamic regime is responsible for the increase in the machining overcut and for the formation of heat-affected zones around the microhole entrance. This is an undesired effect. Hence, machining in this regime must be avoided. Strategies to reduce this regime are presented in Chapter 7.

6.2.3 REPEATABILITY OF DRILLING Although gravity-feed drilling is an open-loop machining process, it is quite reproducible. Precise quantitative values are not available to date. However, it is known that few holes must be first drilled before the process becomes reproducible. This is illustrated in Figure 6.5. Consecutive drilling at 29 V in glass with a cylindrical stainless steel tool-cathode of 0.4-mm diameter is shown (Wu¨thrich et al., 2006c). Note how, after about five drilled holes, the drilling evolution becomes more and more similar until reaching a steady-state situation. In the following sections, only the results obtained in this steady-state situation are discussed. This effect is due to the time needed for the various local parameters (local temperature and electrolyte concentration distribution) to reach their stationary value. A preheated electrolyte can be used to diminish this effect.

6.2.4 DRILLING TIME The drilling time in gravity-feed machining is mainly determined by the drilling depth of the hole and the machining voltage. An example of this is shown in Figure 6.6 for the case of machining in 30 wt% NaOH using a 0.4-mm cylindrical stainless steel tool-cathode (Maillard et al., 2007). Other parameters

124

CHAPTER 6 COMMON MACHINING STRATEGIES

50 0

Z [μm]

–100

1

–200

2

–300

3 4 5

–400 –500 0

5

10

15

20

25

30

35

40

45

t [s] FIGURE 6.5 Several consecutive glass gravity-feed drillings in 30 wt% NaOH at 29 V with a cylindrical stainless steel toolcathode of 0.4-mm diameter. The applied force was 0.8 N. After five successively drilled holes, the evolutions become similar. Reprinted from Wu¨thrich et al. (2006c) with the permission of the Journal of Micromechanics and Microengineering.

70

Mean machining time [s]

60

28V

50 40 30V

30 20

33V

10

37V

0 0

50

100 150 200 Micro-hole depth [μm]

250

300

FIGURE 6.6 The drilling time in SACE glass gravity-feed drilling for a 0.4-mm cylindrical stainless steel tool-cathode in 30 wt% NaOH. Reprinted from Maillard et al. (2007) with the permission of the Journal of Micromechanics and Microengineering.

6.2 GRAVITY-FEED DRILLING

125

such as interelectrode resistance, tool-electrode shape, and tool-electrode material also influence the drilling time, as discussed below. Another important factor is the electrolyte bulk composition and temperature, which affect the chemical contribution to the material removal rate. In general, higher electrolyte concentration and higher bulk temperature result in lower drilling times. The utilization of alkaline electrolytes results in faster drilling times than in the case of acids (Yang et al., 2001). For alkaline solutions, it was observed that KOH results in faster drilling than NaOH for the same concentration and machining voltage (Yang et al., 2001). This may be attributed to the fact that the Kþ ion has a higher mobility than the Naþ ion. This leads to a lower interelectrode resistance, hence a higher discharge activity when using KOH. Another effect may arise from the chemical attack mechanism of glass, which differs for NaOH and KOH (see Section 5.4). But the main contribution is the difference in viscosity of aqueous solutions of both salts. The viscosity of KOH solutions is similar to that of water, even when highly concentrated, whereas the viscosity of NaOH solutions can be up to twenty times higher. The typical drilling times are a few seconds for microholes of about 200–300 mm deep, drilled on glass substrate. For deeper holes, the machining switches to the hydrodynamic regime and the drilling time increases significantly. For ceramics (e.g., alumina) the drilling time is significantly higher, typically, a few minutes for 100-mm depth (Chark, 2007; Sarkar et al., 2006; Tsutsumi et al., 1993). The fluctuation in the drilling time increases with the machining voltage. Due to these large fluctuations, the drilling time cannot be used as a control parameter for the drilling depth. If precise drilling depth control is required, the signal z(t) must be monitored. For low depths, where drilling is fast, this becomes challenging, as even when switching off the machining voltage, the tool remains hot for a few milliseconds. This is enough time to drill further several micrometers. The use of a conductive etch stop-layer can help for these cases (Ozhikandathil et al., 2011).

6.2.5 INFLUENCE OF THE INTERELECTRODE RESISTANCE Depending on the interelectrode resistance, the gas film can form in a few milliseconds, if formed electrochemically (Section 4.1.2) and/or by joule heating (Section 4.1.1). In case of a hybrid machining process, the gas film formation may take few seconds (Section 4.1.3). As the gas film is unstable (i.e., the gas film frequently collapses and has to be built up again), its formation time is an important parameter for the mean SACE machining speed. Figure 6.7 shows two examples of drilling evolution in function of time during SACE glass gravityfeed drilling in the case of high interelectrode resistance (Wu¨thrich et al., 2006b). Drilling is done using a 0.4-mm stainless steel tool-cathode while applying 31 V. The high interelectrode resistance is obtained by depositing a drop of electrolyte (30 wt% NaOH) on the workpiece surface that is large enough to wet the surface. In this case, the gas film is formed during 1 s by a hybrid mechanism. This formation time becomes a significant limitation of the drilling speed, as can be seen by comparing situations (a) and (b) in Figure 6.7, where the gas film needs to be built up more often in (a) than in (b). This results in an overall slower machining for situation (a) with interelectrode resistance compared to (b) having a low interelectrode resistance. The interelectrode resistance R also affects the quantity of heat available for machining at a given voltage, according to Eqn (5.1), as PE ¼ UI
RI2. For higher interelectrode resistance, greater energy will be used for joule heating of the electrolyte.

126

(b) 50

2.0

30

0.4 mm

1.8

0

1.6

0

2.0

0.8 0.6

–90

0.4

1.5 I [A]

1.0 –60

Z [μm]

–50

1.2 I [A]

Z [μm]

1.4 –30

2.5 0.4 mm

–100 1.0 –150 0.5

–200

0.2 –120

0.0 0

1

2

3

4

5

6 7 t [s]

8

9

10 11 12

0.0

–250 0

1

2

3

4

5

6 7 t [s]

8

9

10 11 12

FIGURE 6.7 Example of SACE glass gravity-feed drilling with a 0.4-mm stainless steel tool-cathode at 31 V in the case of high interelectrode resistance. In situation (a) the gas film needs to be built up more often than in situation (b). This results in an overall slower machining for situation (a) than situation (b). Reprinted from Wu¨thrich et al. (2006b) with the permission of the Journal of Micromechanics and Microengineering.

CHAPTER 6 COMMON MACHINING STRATEGIES

(a)

6.2 GRAVITY-FEED DRILLING

127

The decrease in the material removal rate with decreasing interelectrode resistance is not specific to glass drilling. The same effect was also observed in Si3N4 drilling (Sarkar et al., 2006).

6.2.6 MICROHOLE DIMENSIONS The dimensions of the drilled holes are a function of the machining voltage and the drilling depth. Figure 6.8 illustrates this dependence in case of microholes machined in glass using a 0.4-mm cylindrical stainless cathode tool-electrode dipped in 30 wt% NaOH (Maillard et al., 2007). Three zones can be distinguished. For depths down to 100 mm, the mean hole diameter is independent of the machining voltage (zone A). In this configuration, drilling takes place only in the discharge regime. An increase in the voltage results in faster drilling but does not affect the mean entrance diameter. The entrances of the holes are well-defined cylindrical contours with a smooth surface. For deeper microholes, it is increasingly difficult for the electrolyte to reach the tool tip. The machining speed is no longer controlled by the number of discharges (and therefore becomes almost independent of the voltage), but by the drilling depth. This results in an increase in the hole diameter with the drilling depth, after which the diameter reaches a maximal value (zone C) (Zheng et al., 2007a). The hole entrance is surrounded by a heat-affected zone. In zone B, for low voltages (lower than 31 V) and deep holes (deeper than 200 mm), drilling takes place in the transition zone between the discharge and the hydrodynamic regimes. The entrances of the holes have a jagged contour in this case.

750 m 0μ

Mean diameter [μm]

700

33

C

m

0

30

650

μm

0μ

20

B 600

550

100μm A

500

28

30

34 32 Machining voltage [V]

36

38

FIGURE 6.8 Mean entrance diameter of microholes obtained by SACE glass gravity-feed drilling as a function of the machining voltage and the drilling depth for a 0.4-mm stainless steel tool-cathode in 30 wt% NaOH. Reprinted from Maillard et al., (2007) with the permission of the Journal of Micromechanics and Microengineering.

128

CHAPTER 6 COMMON MACHINING STRATEGIES

For ceramic machining, the situation is similar. However, due to the very low material removal rate and the absence of two distinct machining regimes, the mean diameter is almost constant as a function of the drilling depth, increasing with the machining voltage (i.e., only zones B and C from Figure 6.8 can be observed). As the machining voltage needed for ceramic processing is much higher than that for glass, the overcut is generally quite large. Typical values for alumina are 300–700 mm for a 1-mm cylindrical tool-cathode (Bhattacharyya et al., 1999) and 200–400 mm for Si3N4 for a 0.4-mm cylindrical cathode tool (Sarkar et al., 2006). The electrolyte composition and bulk temperature also affect the dimensions of the microholes. The mean microhole diameter increases with lower electrolyte concentration (Yang et al., 2001). This effect is due to the lower machining speed at low electrolyte concentration, resulting in significantly higher drilling times. The chemical composition of the electrolyte has even more dramatic effects. For example, the use of KOH results in hole diameters almost half the size of those drilled using NaOH (Yang et al., 2001). This is due to the different drilling times resulting from the use of KOH and NaOH electrolytes. A major contribution to this difference is the very different viscosity of NaOH and KOH. Even for highly concentrated solutions, KOH has a viscosity similar to water, whereas NaOH can have up to 20 times higher viscosity than water. Besides significantly reducing the drilling time, a low electrolyte viscosity also creates different flow patterns around the tool-electrode. For highly concentrated KOH solutions, this results in clear traces engraved at the entrance of the microhole (Figure 6.9, Cheng et al., 2010). The possibility to

FIGURE 6.9 Micrographs of hole entrances obtained by gravity-feed drilling during 3 s with a 200-mm tungsten carbide tool in 5-M KOH at different voltages. Reprinted from Cheng et al. (2010) with permission from Elsevier.

6.2 GRAVITY-FEED DRILLING

129

texture in a controlled way the surfaces machined using low viscous electrolytes will be further discussed in Section 7.3.5. Due to the low viscosity of 5-M KOH solution, the induced flow patterns are very sensitive to the discharge regime (Section 4.3.3). As a result, as demonstrated by Cheng et al. (2010) and Yang et al. (2010) the entrance diameter behaves differently than in the case of gravity-feed drilling with a high viscous electrolyte such as 30 wt% NaOH. For voltages below the transition voltage (in the instability region), large hole entrances are formed. As the voltage becomes closer to the transition voltage (40 V for 5-M KOH) the entrance diameter becomes smaller (as the gas film becomes more stable) before growing again with the machining voltage once the arc discharge regime is reached (Figure 6.9). Figure 6.10 demonstrates further this effect in case of quartz gravity-feed drilling (Yang et al., 2010). The minimal entrance diameter is achieved at the transition voltage (which is tool-electrode material and roughness dependent, as discussed in Section 4.3.3). Figure 6.11 shows the standard deviation and the roundness error of the microholes drilled in glass using a 0.4-mm stainless steel cathode and 30 wt% NaOH (Maillard et al., 2007). The roundness error is defined as: 1 Relative roundness error ¼ ðdmax
dmin Þ; d

(6.5)

where d is the mean hole diameter and dmax and dmin are the maximal and minimal diameters over a series of drilled holes. Results show that the standard deviation increases with the matching voltage and the hole depth (Figure 6.11(a)), becoming particularly high for voltages above 32 V, whereas it remains almost constant up to 30 V. This strong increase in standard deviation is due to two factors. First, the discharge activity becomes more and more unstable for high voltages (compare with Eqn (4.62) in Section 4.3.5) and, second, variations in the spark activity have a larger impact at high voltages than at low ones. The relative roundness error increases significantly with drilling depth (Figure 6.11). Often, an important contribution to the roundness error is the tool-electrode bending, which may result in highly deformed holes (see Figure 6.12). The increase in the number of noncylindrical holes with depth is significant (Maillard et al., 2007). Bending can be minimized by using tools that are as short as possible, very stiff electrode materials, and a minimal force during gravity-feed drilling. Another important parameter is the electrode polarity. Using an anode compared with a cathode results in very different microhole shapes, as discussed in Section 5.4. The microhole profile is also a function of the tool shape. For cylindrical tools (cathodes), the typical profile is conical but the hole bottom surface generally has two bumps. These bumps, which are due to localization of the discharges at the tool edges, are more pronounced for ceramic and composite materials than for glass (Chark, 2007).

6.2.7 MACHINING QUALITY The existence of the two machining regimes (discharge and hydrodynamic) results in different drilling quality as a function of the machining voltage and depth. In the case of a cathode tool, the holes drilled in glass can be classified into four different types (Figure 6.13) (Maillard et al., 2007): 1. Well-defined cylindrical contours with smooth surface (Figure 6.13(a)): This type of contour is a characteristic of low depths (100 mm, 28–37 V) and low machining voltages (28 V, up to 300 mm). The entrance of the hole is well defined and characterized by a smooth surface.

130

CHAPTER 6 COMMON MACHINING STRATEGIES

FIGURE 6.10 Dependence of the hole entrance diameter for quartz gravity-feed drilling of 150-mm deep holes with a 200-mm tungsten carbide tool in 5-M KOH at different voltages. Reprinted from Yang et al. (2010) with permission from Elsevier.

6.2 GRAVITY-FEED DRILLING

(a)

(b) 30 0μ m

80

0μ m

60

m 0μ

10

40

20 18

Roundness error [%]

100

20

σ [μm]

120

131

16 14

300μm

12

200μm

10

20

100μm

8 6

0

28

30

32

34

36

28

38

Machining voltage [V]

30

32

34

36

38

Machining voltage [V]

FIGURE 6.11 (a) Standard deviation s and (b) roundness error of microholes machined using SACE gravity-feed drilling in glass with 30 wt% NaOH. Reprinted from Maillard et al. (2007) with the permission of the Journal of Micromechanics and Microengineering.

(a)

(b)

FIGURE 6.12 (a) Cylindrical and (b) deformed microholes due to tool-electrode bending. Reprinted from Maillard et al. (2007) with the permission of the Journal of Micromechanics and Microengineering.

2. Jagged outline contours (Figure 6.13(b)): This type of hole appears at depths between 200 and 300 mm when using a machining voltage of about 30 V. The contour is no longer smooth but jagged. 3. Hole with heat-affected zone (Figure 6.13(c)): For machining voltages above 30 V and depths higher than 100 mm, the hole is surrounded by a heat-affected zone. The contour remains cylindrical.

132

CHAPTER 6 COMMON MACHINING STRATEGIES

(b)

(a)

100μm

(c)

200μm

(d)

200μm

200μm FIGURE 6.13 Different microhole qualities obtained by gravity-feed drilling with a stainless steel tool-electrode of 0.4-mm diameter in glass. See text for detailed explanations. Reprinted from Maillard et al. (2007) with the permission of the Journal of Micromechanics and Microengineering.

4. Hole with thermal cracks (Figure 6.13(d)): When machining at 37 V and higher, the hole is characterized by cracks and a large heat-affected zone. The entrance of the hole is generally no longer cylindrical. Figure 6.14 summarizes the different types of microholes that are obtained by gravity-feed drilling as a function of the machining voltage and the drilling depth. This plot can be understood by taking into account the existence of the two machining regimes. For a given voltage, machining is initially done in the discharge regime (up to about 100 mm). The quality of the hole is not affected by the voltage where a higher voltage results in faster drilling without loss of quality. As the tool drills deeper into the workpiece, the difficulty in removing the machined material and insufficient wetting of the tool tip increases. The machining enters into the hydrodynamic regime. More and more discharges take place in the upper part of the tool, resulting in a jagged contour and an increase in the hole diameter. As the drilling continues, the hole diameter increases further until it reaches its maximal value (depending on the machining time and voltage). In this situation, the border of the hole is not heated enough to be etched. Therefore, the microhole diameter does not increase further, but a heat-affected zone starts growing instead. If the heat power supplied is very high (for voltages above 36 V), thermal cracks appear.

6.2 GRAVITY-FEED DRILLING

133

38 Thermal cracks Machining voltage [V]

36 34

Heat affected zone

32 30 Jagged surface 28

Smooth surface 0

100

200

300

Drilling depth [μm] FIGURE 6.14 Evolution of SACE glass gravity-feed drilling in the machining voltage-drilling depth plane. Reprinted from Maillard et al. (2007) with the permission of the Journal of Micromechanics and Microengineering.

According to this description, deep holes must always have poorer quality than holes with a depth typically lower than 100 mm. Machining at 28 V and lower seems to be an exception to this rule. The fact that machining at 28 V results in excellent surface qualities, even for deep holes, can be attributed to the discharge activity. As discussed in Section 4.3.3, the discharge activity is different for voltages ranging from the critical to the transition voltage when compared to higher voltages. These two types of discharge activities (instability and arc discharge region) result in two different machining qualities. For the low-voltage region (instability region), the combination of discharges together with etching promoted by the hydrodynamical flows due to the incompletely formed gas film, results in the smooth surface observed. While the variance in the machined diameter is small in this case, it starts growing significantly in the arc discharge region (Figure 6.11(a)). In the case of low-viscosity electrolytes, such as KOH, the difference in machining in the instability and arc discharge regions is even more pronounced. Minimal entrance diameters and best quality of the hole entrance are obtained when machining at the transition voltage. An example for Pyrex glass is shown in Figure 6.15, where Cheng et al. (2010) drilled holes in different KOH concentrations at the transition voltage. Another important aspect is the surface roughness of the drilled holes. Typical values for glass can be as low as a few hundred nanometers (Yang et al., 2001). The use of NaOH results in smoother surfaces than KOH (Yang et al., 2001), probably due to the lower heat available for machining in NaOH compared with KOH, as these electrolytes have different electrical conductivities. Another important factor is the electrolyte viscosity. As will be discussed in Section 7.3.5, different surface textures can be engraved by choosing adequate electrolyte viscosity and tool-electrode kinetics.

134

CHAPTER 6 COMMON MACHINING STRATEGIES

FIGURE 6.15 Micrographs of hole entrances obtained at the transition voltage (cylindrical tungsten-carbide tool of 200-mm diameter; tool rotating speed: 500 rpm; machining depth: 250 mm). Reprinted from Cheng et al. (2010) with permission from Elsevier.

6.3 CONSTANT VELOCITY-FEED DRILLING In constant velocity-feed drilling, the tool-electrode is fed with a chosen feed-rate f (Figure 6.16). Compared with gravity-feed drilling, constant velocity-feed drilling can potentially avoid the mechanical contact between the tool-electrode and the workpiece. This can, in particular, overcome the

f

Workpiece FIGURE 6.16 Principle of SACE constant velocity-feed drilling. The tool is fed at a constant feed-rate f.

6.3 CONSTANT VELOCITY-FEED DRILLING

135

problem of the tool-electrode bending. However, as the tool feed-rate is selected beforehand, the gap between the workpiece and the tool-electrode is not controlled and is generally not constant during machining. Thus, the progress of drilling cannot be monitored online. A further consequence of the presence of the gap between tool-electrode and workpiece is that the temperature in the machining zone will be lower than the tool-electrode temperature. The tool-electrode feed-rate has to be selected properly. Feed-rates faster than the mean material removal rate of the process will result in contact forces that may result in breaking either the workpiece or the tool-electrode. On the other hand, very slow feed-rates will increase drilling times and may result in large heat-affected zones around the microhole. So far, only a few studies have been carried out on the optimal feed-rate. Typical values reported in the literature are, depending on the tool-electrode diameter, about 1–15 mm s
1 (Han et al., 2007; Liao and Peng, 2006; Lim et al., 2001; Jui et al., 2013), a value slightly higher than the limiting speed reached in the hydrodynamic regime during gravity-feed drilling. As in gravity-feed drilling, the mean hole diameter increases with drilling depth (Lim et al., 2001). For high enough depths, a maximal diameter will eventually be reached. This may be attributed to reduced availability of electrolyte in the machining zone, as with increasing depth the electrolyte can no longer easily reach the bottom of the hole. Consequently, the discharge activity shifts to the upper part of the tool-electrode, causing enlargement of the hole entrance. The tool feed-rate directly affects the microhole surface roughness. As shown by Han et al. (2007), the surface roughness follows an inverse volcano dependence with respect to the tool feed-rate (Figure 6.17). This dependence is due to the competition between local heating of the workpiece and high-temperature etching of the glass.

7 6

Ra [mm]

5 4 3 2 1 0 0

5

10 15 20 25 Tool-electrode feed rate [mm/s]

30

35

FIGURE 6.17 Microhole surface roughness as a function of the tool-electrode feed rate, according to Han et al. (2007). Drilling was done with a 0.2-mm cylindrical tungsten carbide tool-cathode in 30 wt% NaOH electrolyte at 35 V. Reprinted from Han et al. (2007) with permission from Elsevier.

136

CHAPTER 6 COMMON MACHINING STRATEGIES

The main advantage of constant velocity-feed drilling compared with gravity-feed drilling is that the motion of the tool-electrode is user controlled. This allows the machining of more complex shapes than cylindrical holes. An impressive example is the machining of threads in glass as shown by Lee et al. (2004). The authors were able to build microfluidic devices with removable tubing interconnects that could withstand pressures up to 200 kPa. They also demonstrated the fabrication of other interconnection shapes (conical and hourglass interconnections). Proper implementation of constant velocity-feed drilling requires knowledge of the reasons for appearance of contact forces. For appropriately selected tool feed-rates, a gap between the tool-electrode and the workpiece forms, which results in different machining zone temperatures than in gravity-feed drilling. Both aspects are discussed in the remainder of this section.

6.3.1 FORCES EXERTED ON THE TOOL-ELECTRODE Figure 6.18 shows a representative example of measured forces exerted on a 500-mm cylindrical stainless steel tool-electrode when drilling a glass workpiece at various tool feed-rates f while

FIGURE 6.18 Measured forces exerted on a 500-mm cylindrical stainless steel tool for different tool feed-rates: low (a: 1 mm s
1), intermediate (b: 2 mm s
1, c: 3 mm s
1) and high (d: 5 mm s
1); machining at 30 V in 30 wt% should be together NaOH. Twenty holes were machined for each experimental set. The used force sensor saturates at 6 N (Abou Ziki and Wu¨thrich, 2013).

6.3 CONSTANT VELOCITY-FEED DRILLING

137

machining at 30 V in a 30 wt% NaOH solution (Abou Ziki and Wu¨thrich, 2013). The tool is fed down by 300 mm toward the glass workpiece, which is initially placed at a position 50 mm below the origin of the motion. For each condition, 20 successive drillings are showed. Except for very low feed-rates (f ¼ 1 mm s
1), forces start to be measured after moving down the tool by typically 40 mm. The difference of about 10 mm compared to the expected workpiece location is the tool-electrode thermal expansion (Section 4.5). Once the tool contacts the workpiece, the force stays as long as the needed temperature to start machining is not yet reached. This time is given by the time to needed to reach the machining temperature as discussed in Section 5.2.2. During the interval to the tool will be fed over a distance z1 given by: z1 ¼ f $to :

(6.6)

After these initial forces, termed low depth forces, the tool is fed over a more or less wide range where no forces are measured. This interval of fed distance is termed the middle region. In this region, a gap is established below the tool-electrode avoiding any tool‒workpiece contact. This gap will be discussed in more detail in Section 6.3.3. The middle region is the equivalent of the depth range over which the discharge regime is observed in gravity-feed drilling. At the end of the middle region, forces are exerted on the tool again. These forces are termed high depth forces. The origin of these forces is the insufficient flushing of the bottom of the hole, as in the hydrodynamic regime of gravity-feed drilling. For fast tool feed-rates, only high depth forces can be observed. Results like the one presented in Figure 6.18 can efficiently be summarized in diagrams showing the regions where forces are observed in the tool feed z versus tool feed-rate f plane. Figure 6.19 is an example of such diagrams constructed for different tool diameters and machining voltages. Note how the region of the low depth forces and the middle region are separated by a straight line, as described by Eqn (6.6). Inspection of this separation line allows the determination of the time to .

(b)

(a) 80

3.5

40

70

3 2.5

60

2

30

50

1.5 1 0.5

40

0

20

50

100

150 200

250

30 20

10

10 0

0

50

100

150

200

250

0

0

100

200

300

FIGURE 6.19 Schematic representation of the regions where forces or no forces are observed in the tool feed versus tool feed-rate plane while using 500-mm tool (a) and 250-mm tool (b). For (a), 30 and 33 V were applied and (b) 30 V (Abou Ziki and Wu¨thrich 2013).

400

138

CHAPTER 6 COMMON MACHINING STRATEGIES

Equation (6.6) suggests that the force diagrams in the tool feed z versus tool feed-rate f plane can be brought into a dimensionless form: z f ¼ ; b b=to

(6.7)

with b the tool-electrode radius. The normalization using the tool-electrode radius b is inspired by the fact that the high depth forces originate due to insufficient flushing of the hole, which is related to the hole aspect ratio z/b. As shown in Figure 6.20, this normalization is not only valid for the separation line between the low depth forces and the middle region, but also turns out to be applicable to the rest of the force diagram. It is interesting to compare the transition from the middle region to the high depth forces region with that occurring from the discharge to the hydrodynamic regime in gravity-feed drilling. The last occurs at the characteristic length d as defined in Eqn (6.1). For a 0.4-mm stainless steel tool, d is in the range of 70–80 mm (Jalali et al., 2009), which corresponds to about 0.2 in normalized feed distance. This is exactly the normalized feed where the separation line between the low force and middle regions meets the separation line between the middle region and the high depth forces region. Consequently, the feed-rate fmax at which the middle region vanishes is given by: fmax ¼ 0:2

b d ¼ : to to

(6.8)

In summary, the occurrence of low depth forces is due to insufficient heating of the workpiece surface and is characterized by the time to . On the other side, the occurrence of high depth forces is caused by insufficient flushing of the microhole and is characterized by the depth d. Both effects contribute in determining the maximal tool feed-rate beyond which the middle region disappears, as described by Eqn (6.8). 0.25 0.2 0.15 0.1 0.05 0 0

0.5

1

1.5

2

FIGURE 6.20 Schematic representation of the regions where forces or no forces are observed in the normalized tool feed versus normalized tool feed-rate plane (Abou Ziki and Wu¨thrich, 2013).

6.3 CONSTANT VELOCITY-FEED DRILLING

139

6.3.2 NATURE OF CONTACT FORCES IN GLASS DRILLING While the tool-electrode is in contact with the workpiece, which caused the measured contact forces, the workpiece gets heated up to a temperature similar to that of the tool. In the case of glass, this results in forming a bond between the tool-electrode and the glass surface, as illustrated in Figure 6.21 (Abou Ziki, 2014). In this example, the machining voltage was switched off at 51.5 s while the 0.5-mm stainless steel tool-electrode was pressing on the surface with about 0.5 N. Due to cooling of the machining zone, the tool retracts and the force progressively reduces and becomes negative. This shows that now a force is pulling on the tool (the used force sensor, based on the measurement of the deflection of a rigid element of known stiffness, is able to measure pushing and pulling forces). The bonding between the glass workpiece and the tool-electrode must be a fast process, as shown in Figure 6.22. In fact, bonding occurs even for very short pressing times (times during which the toolelectrode is pressed against the glass workpiece; 0.1 N in this case). Further, the bonding force is essentially independent of the pressing time. Depending on the level of the pressing force (force at which the tool-electrode is pressing at the moment the machining voltage is switched off) two cases can happen (Figure 6.23). For the interpretation of the results, it is important to understand the principle used by the force sensor acquiring the presented data (similar to the one discussed in Section 8.2.4). The sensor determines the contact forces by measuring the displacement of a flexible element of known stiffness (which has a much lower stiffness than the tool-electrode). Consequently, the quantity delta-force, defined in Figure 6.23, is directly proportional to the retraction of the tool-electrode (only a negligible part of it

6 4 2

0.5 0 −0.5 −1 40

45

50

55

60

FIGURE 6.21 An example of how the tool-electrode gets bonded to the glass surface while drilling a hole in 30 wt% NaOH at 33 V and for a tool feed-rate of 10 mm s
1. Machining current and the force exerted on the tool-electrode are shown together with the bottom of the hole. The bonding is witnessed by a negative force signal as the machining voltage is switched off at 51.5 s (Abou Ziki, 2014).

140

CHAPTER 6 COMMON MACHINING STRATEGIES

0.2

4 0.2 0.1

0.1

2

0 −0.1 −0.2 190

0 0

0.2

0.4

192

194

0.6

196

198

0.8

0 1

FIGURE 6.22 Dependence of the bonding force of the pressing time (time during which the tool-electrode is pressed against the glass workpiece) when the tool-electrode is pressed with 0.1 N against a glass workpiece for a 100-mm deep hole, using 30 wt% NaOH, 10-mm s
1 tool feed-rate and machining at 33 V (Abou Ziki, 2014).

(b)

0.25

0.6

0.2 0.15

0.5

Pressing Force

0.1

0.4

−0.1

Delta Force

−0.05

Glue Force

0.05 0

Delta Force

(a)

−0.15 −0.2 190 191 192 193 194 195 196 197 198 199

0.3 0.2 0.1 0 90

91

92

93

94

95

96

97

98

FIGURE 6.23 Depending on the pressing force, two cases can be distinguished in the recorded force signal. (a) For low pressing forces, a negative force can be observed. (b) For higher pressing forces, the tool retraction is not enough to force the detachment between the workpiece and tool-electrode (Abou Ziki, 2014).

is due to tool-electrode bending where it is primarily caused by deformation of the flexible element of the force sensor). For low pressing forces, lower than the force given by the tool thermal expansion times the force sensor stiffness (0.35 N in the case shown in Figure 6.23), a negative force signal (glue-force) indicating tool-surface bonding can be observed. For higher pressing forces, the tool retraction due to cooling is not enough to detach the tool and workpiece, and no negative force can be observed.

6.3 CONSTANT VELOCITY-FEED DRILLING

141

0.5 0.4 0.3 0.2 0.1 0

0

0.05

0.1

0.15

0.2

0.25

0.3

FIGURE 6.24 Dependence of the glue-force of the pressing force for a 500-mm tool-electrode pressed against a glass workpiece (100-mm deep hole, using 30 wt% NaOH, 10-mm s
1 tool feed-rate and machined at 33 V) (Abou Ziki, 2014).

The strength of the bond, termed glue-force in the present text, is independent of the pressing force (Figure 6.24), whereas the quantity delta-force depends on it (Figure 6.25). This indicates that the bond is not lost at a specific tool-electrode temperature, since delta-force is directly proportional to the tool retraction (which itself is proportional to the temperature change). Actually, Figure 6.25 shows that, at the moment the bond breaks, the tool-temperature can be any value between room temperature (for the case of maximal delta-force) and high temperature (for minimal delta-force). In the presented example, at the moment the bond is lost, the tool has retracted between 3 and 8 mm (corresponding to a delta-force from 0.15 to 0.35 N) calculated based on the force sensor stiffness. This corresponds to a tool temperature reduction of about 190–475  C (Figure 4.24; Section 4.5). Hence, the tool has a temperature between room temperature to 285  C when the bond is lost. Inspections of the bottom surface of the drilled microhole after pressing the tool-electrode against the glass workpiece gives further insights into the nature of the mechanism leading to the bond. For pressing forces up to the force given by the tool thermal expansion during machining times the force sensor stiffness (0.35 N is in the example of Figure 6.25), no clear imprint of the tool-electrode on the machined surface can be seen. For these levels of pressing forces, the tool detaches from the glass surface and breaks the bond once retracted sufficiently (case (a) of Figure 6.23). Hot electrolyte flows inside the created gap and smoothens the surface of the hole. The situation is different for higher pressing forces (case (b) of Figure 6.23), where the tool-electrode stays in contact with the glass even when totally cooled, resulting in limited electrolyte flushing of the hole bottom surface. Hence, etching of the surface beneath the tool is limited, and a clear imprint of the electrode can be seen. The shape of the hole bottom surface suggests that the hot tool-electrode may get bonded by pressing it into a softened layer of the bottom of the hole. Another explanation is that some glass, possibly formed from the residual of the machined material, forms around the tool, which causes tool entrapping.

142

CHAPTER 6 COMMON MACHINING STRATEGIES

FIGURE 6.25 Dependence of the delta-force of the pressing force for a 500-mm tool-electrode pressed against a glass workpiece (100-mm deep hole, using 30 wt% NaOH, 10 mm s
1 tool feed-rate and machined at 33 V). Selected bottoms of the drilled microholes are shown as well (Abou Ziki, 2014).

Finally, the observed glue-force may also be due to the formation of a chemical bond between the tool-electrode and the glass. The graph in Figure 6.25 gives a strong argument for this mechanism. Indeed, when extrapolating the measurements to zero pressing-forces, there is still a delta-force of 0.15 N remaining (which is the glue-force as shown in Figure 6.24). This implies that even at zero pressing-force, a glue-force of 0.15 N is present. No mechanical pressing is needed to form the bond, but rather, contact between the tool-electrode and the glass is enough.

6.3.3 TOOL‒WORKPIECE GAP Besides the occurrence of contact forces, the formation of a gap between the tool-electrode and the workpiece is a characteristic of constant velocity-feed drilling. This gap is created because the material removal rate is higher than the tool feed-rate. In such situation, the heat generated by the electrochemical discharges heat up the workpiece, leading to material removal. A gap starts to grow. As the gap is growing, the heat transfer resistance between the tool and the workpiece increases, which results in lower temperatures of the bottom surface of the microhole. Eventually, when the temperature is too low, the machining stops.

6.3 CONSTANT VELOCITY-FEED DRILLING

143

FIGURE 6.26 (a) Schematic of the tool‒workpiece gap g, and (b) lumped thermal model. The thermal resistance Rg models the heat transfer through the gap, Rliq through the electrolyte, and Rtool through the tool-electrode. The tool tip has a heat capacity C and temperature T, and is heated by the source qi modeling the heat generated by the electrochemical discharges. The temperature of the bottom of the hole is Tm, and Ta is the ambient temperature (Abou Ziki, 2014).

Figure 6.26 depicts a lumped thermal model of constant velocity feed drilling in the case where a gap is established. At steady state, one can write: qi ¼

1 1 ðT
To Þ þ ðT
TM Þ; R Rg

(6.9)

where qi is the heat power generated by the electrochemical discharges, R is the equivalent thermal resistance of Rtool in parallel with Rliq , T is the tool tip temperature, TM is the workpiece surface temperature (machining temperature) and To is the temperature at infinity. The gap thermal resistance Rg is estimated as Rg ¼

4g ; lg pd2

(6.10)

with g the size of the gap and lg the thermal conductivity of the media within the gap. As will be seen below, the size of the gap is very narrow (typically a few micrometers); consequently, the approximation Rg << R holds. If Eqn (4.27) is used to estimate qi , it follows by combining Eqns (6.9) and (6.10):   g 1 lg TM
To ¼ 1
: (6.11) d 4 lliq T
To

144

CHAPTER 6 COMMON MACHINING STRATEGIES

To estimate the size of the steady-state gap from Eqn (6.11), the temperature TM and heat conductivity lg must be known. The temperature TM is the temperature below which machining stops. In the case of glass microdrilling, the material removal mechanism is essentially high-temperature etching. Consequently, the media within the gap must be electrolyte in either molten salt or aqueous form. Recall that depending on the machining voltage, the tool temperature T is typically 500  C (Section 4.5), which is higher than the melting temperature of the electrolyte salt (for NaOH the melting point is 318  C and for KOH 406  C). As the gap grows, the temperature on the workpiece surface drops until it reaches the melting point of the electrolyte salt. When the salt freezes, etching will stop, respectively becoming negligible. Consequently, in the case of glass constant velocity-feed microdrilling, TM is given by the electrolyte salt melting point. The value of the thermal conductivity lg of the media in the gap is challenging to estimate. On one side, the temperature inside the gap is not constant (a temperature drop from the tool temperature T to the temperature of the workpiece TM ). On the other side, the media inside the gap is not well known but is probably a mixture of molten electrolyte salt (e.g., NaOH), gas bubbles, and products of the machining. Due to the presence of gas bubbles, it is expected that lg will be significantly lower than the thermal conductivity of the molten electrolyte salt. Figure 6.27 shows measured gaps under various conditions (Abou Ziki, 2014). To measure the gap, the tool-electrode motion was stopped at the desired depth and the machining voltage was switched off. After a 10-s waiting time, added to allow tool-electrode cooling, the tool was moved down further where the distance up to the moment it touches the bottom surface of the hole was determined. This distance, after subtracting the tool-electrode thermal expansion, is the gap size g. To ensure the formation of a gap, the feeding was stopped whenever a force higher than 0.3 N was exerted on the tool-electrode. Measurements were done during constant velocity-feed drilling using a 500-mm stainless steel tool fed at 10 mm s
1 in 30 wt% NaOH. The gap was measured in functions of (a) microhole depth for 33 V, (b) machining voltage, for 110-mm depth, and (c) duty-cycle for 110-mm depth. Error bars are computed based on the standard deviation of 50 repeats. Based on the tool-electrode expansion, the tool temperature was determined for the drilling in (b) and (c) (i.e. for 110-mm hole depth) and Figure 6.27(d) plotted. The line is computed with Eqn (6.11) using lg =lliq ¼ 0:17 (Abou Ziki, 2014). As shown in Figure 6.27(a), the gap decreases with the hole depth which indicates, based on Eqn (6.11), that the thermal conductivity within the gap decreases as well. This is probably caused by the increasing difficulty of flushing the microhole for higher depths, which results in accumulation of more machined material and bubbles in the gap. Therefore, the local heat transfer resistance increases. An interesting consequence of Eqn (6.11) is that the gap in case of machining with KOH is smaller than that resulting from machining with NaOH (as the melting point of KOH is higher than that of NaOH). This can explain the observation that, in general, machining with KOH results in lower overcuts than when machining with NaOH (Yang et al., 2001; Cao et al., 2009). This finding opens new possibilities for engineering adequate electrolytes for minimizing overcuts.

6.4 2D AND 3D MACHINING One of the very interesting features of SACE machining is that, in addition to drilling, SACE allows for the machining of 2D structures. To date, 2D machining has been studied systematically only for glass,

6.4 2D AND 3D MACHINING

(a)

(b)

145

10

8

8 6

6 4

4 2

2

0 50

(c)

0 100

150

200

250

300

350

400

450

26

28

30

32

34

36

(d) 10

9 8

8

7 6

6

5 4

4

3 2

2 1 30

40

50

60

70

80

90

100

110

0 360

380

400

420

440

460

480

500

520

FIGURE 6.27 Measured gaps during constant velocity feed drilling at 10 mm s
1 with a 500-mm stainless steel tool in 30 wt% NaOH in function of (a) microhole depth for 33 V, (b) machining voltage, for 110-mm depth, (c) duty-cycle for 110-mm depth, (d) tool-electrode temperature; line computed based on Eqn (6.11) with lg =lliq ¼ 0:17 (Abou Ziki, 2014).

and the remainder of this section is limited to discussion of this material. Two methods for 2D machining have been investigated until now. The first method involves moving the tool-electrode above the glass surface by following a specified path. The tool-electrode has to be closer than 20–25 mm for machining to take place. The second method involves drilling down to a certain depth and then moving the tool-electrode laterally. In both cases, the tool is moved at a constant velocity. Typical values are around 50–150 mm s
1. Such high speeds are possible because, contrary to drilling, the machining is almost never limited by the ability of the electrolyte to reach the machining zone. In other words, machining is always in the discharge regime. Further, as the bulk electrolyte reaches the machining zone without any difficulty, the local temperature will be similar to the vaporization temperature of the used electrolyte (slightly higher, as the local electrolyte concentration will increase

146

CHAPTER 6 COMMON MACHINING STRATEGIES

due to evaporation of some of its water). This corresponds to the situation discussed in Section 5.2.3 for the boundary condition Eqn (5.3). In the following sections, the method in which the tool is moved above the glass surface is discussed, as only this machining strategy has been investigated systematically until now.

6.4.1 QUALITY OF MACHINED MICROCHANNELS In 2D SACE machining, two parameters can be adjusted: the tool travel speed and the machining voltage. In contrast to drilling, much more freedom is available for the tool travel speed. The selection of the appropriate combination of these two parameters is one of the most important issues in 2D SACE machining. Microchannels are not formed for all velocities at which the tool is moved above the glass surface. Whether a microchannel is created or not depends not only on the tool travel speed but also on the machining voltage and the distance between the tool and the workpiece. As will be discussed in detail in Section 6.4.2, at some travel speeds it is not possible to have a microchannel formed. For example, when machining at 28 V, it is not possible to have any acceptable channel shape for travel speeds higher than 40 mm s
1. When increasing the machining voltage to 30 V, the tool can be moved at higher travel speeds up to 50 mm s
1. However, using high machining voltages (more than 32 V) at low tool travel speeds results in a nonsmooth channel surface with significant depth variation along the channel. As a general rule, the quality of the microchannels deteriorates as the tool travel speed is decreased. This can be attributed to the poor material removal rate and the accumulation of melted material inside the microchannel behind the tool. There are optimal conditions under which microchannels with edges and surfaces of acceptable quality are achieved. The channels so obtained can be subdivided into the following five categories (Didar et al., 2008): 1. Well-defined linear channel edges and smooth channel surface: This type of contour is characteristic of low machining voltages with appropriate tool travel speeds. Typical values are 28 V for tool travel speeds ranging from 5 to 10 mm s
1 and 30 V for tool travel speeds varying between 15 and 30 mm s
1 (Figure 6.28(a)). 2. Jagged outline contours with smooth channel surface: This contour is observed for low machining voltages (less than 32 V) with tool travel speeds lower than those causing the previous contour type. The microchannel edges are jagged but the channel surface is still flat and smooth (Figure 6.28(b)). 3. Heat-affected edges with smooth channel surface: This type of contour is obtained at high machining voltages (more than 32 V) when using tool travel speeds high enough to remove the melted material. Under these conditions, the microchannels will have smooth surface but the channel boundaries are not well defined due to the effect of heat generated (Figure 6.28(c)). 4. Heat-affected edges with nonsmooth channel surface and thermal cracks: When the tool travel speed is low and the machining voltage is high (typically 32 V for travel speeds less than 30 mm s
1 and 35 V for speeds less than 40 mm s
1), the edges are unclear and heat affected. Further, the surface is neither flat nor smooth where thermal cracks are observed (Figure 6.28(d)).

6.4 2D AND 3D MACHINING

147

(a) 100 μm 200 μm

(b) 200 μm 100 μm

(c) 200 μm

100 μm

(d)

200 μm 100 μm

(e) 200 μm

FIGURE 6.28 Different types of channels in 2D-SACE machining. Machined using a 0.2-mm cylindrical stainless steel tool-cathode in 30 wt% NaOH: (a) well-defined linear channel edges and smooth channel surface; (b) jagged outline contours with smooth channel surface; (c) heat-affected edges with smooth channel surface; (d) heat-affected edges with rough channel surface; (e) deteriorated microchannels. Reprinted from Didar et al. (2008) with the permission of the Journal of Micromechanics and Microengineering.

148

CHAPTER 6 COMMON MACHINING STRATEGIES

5. Deteriorated microchannels: As the tool travel speed is increased, the channels are no longer continuous and the inner surface becomes very rough. For example, at 28 V, deteriorated microchannels result when increasing the tool travel speed above 40 mm s
1 (Figure 6.28(e)). Figure 6.29 summarizes the type of microchannel obtained for different combinations of the machining voltage and the tool travel speed. Machining voltages less than 32 V result in acceptable microchannel quality. In this voltage range, an appropriate selection of the tool speed will result in the best microchannel quality (region 1) with a channel depth varying between 50 and 120 mm depending on the parameters used. Lower travel speeds will result in deeper microchannels, but the quality of the edges will not be as good as for region 1, although the surface quality is still excellent. Note that decreasing the travel speed will increase the microchannel depth smoothly over time because of chemical etching (discussed in more detail in Section 6.4.3). Therefore, if the aim is to have a microchannel with the best possible quality and a constant depth of around 60 mm, the best choice is to apply 28 V at 20 mm s
1 tool travel speed. When using machining voltages higher than 32 V, the microchannel borders are always heat affected. In this range, if the chosen tool travel speed is high enough, it will be possible to have smooth microchannel surfaces (region 3 in Figure 6.29). Decreasing the tool travel speed corrupts the microchannel surface and results in a poor surface quality (region 4 in Figure 6.29). This effect is due to accumulation of the melted material inside the microchannel area and the inability to remove it from the microchannel surface (Figure 6.30).

35 4-Heat affected channel edges with unsmooth surface

Machining voltage (V)

34 33

3-Heat affected edges with smooth surface

32 2-Jagged edges with 1-well defined smooth surface edge with smooth surface

31 30

5-Deteriorated micro-channels

29 28 27 0

20

40

60

80

100

120

Tool travel speed (μm/s) FIGURE 6.29 Characterization diagram for microchannels machined using SACE technology as a function of the machining voltage and the tool travel speed (for tool-electrode distances less than 15 mm). Machined using a 0.5-mm stainless steel cathode in 20 wt% NaOH. Reprinted from Didar et al. (2008) with the permission of the Journal of Micromechanics and Microengineering.

6.4 2D AND 3D MACHINING

Toolelectrode

149

Tool Movement Direction

Accumulation of melted material on micro-channel surface

Workpiece FIGURE 6.30 Schematic representation of the accumulation of melted material on the microchannel surface for high machining voltages and low machining speeds (region 4 in Figure 6.29). Reprinted from Didar et al. (2008) with the permission of the Journal of Micromechanics and Microengineering.

Finally, region 5 in Figure 6.29 is the result of excessive tool travel speed. In this region, there is not enough time for the glass to be machined and to form an acceptable microchannel. As the tool is moved above the glass surface at high speeds, it is possible to see the effect of the tool geometry on the machined surface (Figure 6.28(e)). In fact, the machined microchannels look like discrete rings in this region. For a 500-mm cylindrical tool, rings of 700 mm diameter are obtained. This value is actually the tool diameter plus the observed overcut in the hydrodynamic regime for gravity-feed drilling.

6.4.2 MAXIMAL ALLOWED TOOL TRAVEL SPEED An important parameter in 2D SACE machining is the maximal velocity at which the tool can be moved for machining to take place. Didar et al. (2008) proposed a simple model to evaluate this quantity. Recall that the time to needed for machining to start is given by Eqn (5.15). For 2D machining to take place at a travel speed v, the following condition has to be satisfied: yto ¼ d;

(6.12)

where d is the typical distance beneath the tool-electrode over which the discharges generated from the tool-electrode are distributed. For a cylindrical tool of radius b, this distance would be d ¼ 2b if the discharges are distributed uniformly. This is not the case, however, as the discharges are mostly located at the edges of the tool-electrode. A typical order of magnitude for d is about 25 mm (Didar et al., 2008). Figure 6.31 depicts the prediction of this simplified model and the experimental data. The general trend and the order of magnitude are well reproduced. Note the jump in the experimental data at around 32 V. This jump is probably due to the different discharge activity below and above 32 V as described previously in Section 4.3.

150

CHAPTER 6 COMMON MACHINING STRATEGIES

37

Machining voltage (V)

35 33 31 29 Experimental data

27

Heat model prediction

25 0

20

40 60 80 100 Tool travel speed (μm/s)

120

140

FIGURE 6.31 Maximal machining speed in 2D SACE machining as a function of the machining voltage (Didar et al., 2008). The curve is computed using Eqn (6.12). Reprinted from Didar et al. (2008) with the permission of the Journal of Micromechanics and Microengineering.

6.4.3 DEPTH OF MACHINED MICROCHANNELS Even if the trajectory of the tool-electrode while machining a channel can be directly controlled, the depth of the channel is not straightforward to monitor. Depending on the tool travel speed and the machining voltage, different depths are obtained. Another important issue is the chemical etching of the substrate. Due to this effect, the depth of the channels will not remain constant over machining time (from one region of the channel to another), but will increase slightly. A typical example is shown in Figure 6.32, which illustrates the depth of a channel machined at 28 V and 5 mm s
1 (Didar et al., 2008). The tool was moved a few micrometers above the glass surface and the trajectory was corrected to take into account the misalignment between the glass surface and the tool motion. At this low tool travel speed, the change in the channel depth over the machining time is clearly visible. Two parameters are useful in the design of a machined microchannel. The first parameter is zo, the channel depth at time zero. This value is obtained by extrapolating the channel profile curves, such as the one in Figure 6.32, to time t ¼ 0. The second parameter is md, the average channel depth increase rate over time. The depth z(t) of the channel over time is then given by: zðtÞ ¼ md t þ zo :

(6.13)

Both parameters are functions of the machining voltage, as it controls the heat transferred by the electrochemical discharges to the workpiece and to the electrolyte (recall that the rate of chemical etching is highly temperature dependent), and of the tool velocity. Figure 6.33 shows the dependence of these two parameters on the tool travel speed for machining voltages of 28 and 30 V.

6.4 2D AND 3D MACHINING

151

290

depth (μm)

270 250 230 210 190 170 150 10

15

20

25 30 Time (minutes)

35

40

45

FIGURE 6.32 Microchannel depth as a function of the machining time. Machining was done with a stainless steel cathode at 28 V and 5 mm s
1 in 30 wt% NaOH. Reprinted from Didar et al. (2008) with the permission of the Journal of Micromechanics and Microengineering.

2.5

140 120 30V

100

28V

1.5

zo (μm)

md (μm/s)

2 30V

1

80 28V

60 40

0.5

20 0

0 0

5

10 15 20 25 Tool travel speed (μm/sec)

30

35

5

15 25 35 45 Tool travel speed (μm/sec)

55

FIGURE 6.33 Dependence of the microchannel depth increase rate md and the initial depth zo on the tool travel speed. Reprinted from Didar et al. (2008) with the permission of the Journal of Micromechanics and Microengineering.

6.4.4 INFLUENCE OF TOOL DISTANCE FROM WORKPIECE Figure 6.34 shows the profile of a microchannel machined while applying 30 V and a tool travel speed of 30 mm s
1 at different tool‒glass surface distances. The average depth of the microchannels decreases with higher tool
substrate distance. The quality of the machined microchannels does not change significantly for distances up to 15 mm. However, above 25 mm, no acceptable results can be achieved. For tool
workpiece distances greater than 15 mm, the microchannel contours tend to deteriorate independently of the applied voltage and the tool travel speed.

152

CHAPTER 6 COMMON MACHINING STRATEGIES

65 60 5 μm

Depth(μm)

55 50

10 μm

45

15 μm

40 35 30 25 1

3

5

7 X (mm)

9

11

13

FIGURE 6.34 Profile of microchannel depths at three different tool-electrode distances from the glass surface (machining voltage is 30 V and tool-electrode speed is 30 mm s
1 in 30 wt% NaOH). Reprinted from Didar et al. (2008) with the permission of the Journal of Micromechanics and Microengineering.

6.5 WIRE ELECTROCHEMICAL DISCHARGE MACHINING An extension of machining with electrochemical discharges to wire machining, called travelling wire electrochemical discharge machining (TW-ECDM), was first proposed by Tsuchiya et al. (1985) and studied further by Jain et al. (Jain et al., 1991; Nesarikar et al., 1994) and Peng and Liao (2004). TW-ECDM is particularly interesting for slicing glass fiber composites (Jain et al., 1991; Jain, 1993; Tandon, 1987; Tandon et al., 1990), but can also be used for 2D contour cutting (Tsuchiya et al., 1985). In TW-ECDM, a wire is used as the tool-electrode in a similar manner as in wire electrical discharge machining (WEDM). Copper wires (Jain, 1993; Peng and Liao, 2004), stainless steel wires (Peng and Liao, 2004), or brass wires (Singh et al., 1996; Tsuchiya et al., 1985; Tsutsumi et al., 1993) are typically used. A trade-off must be made when choosing the wire speed, considering that a high speed allows cooling the wire (hence prevents its overheating and breaking), and that low speed is more economical (Jain et al., 1991). Typical wire speeds vary from a few millimeters per minute (Jain et al., 1991) to a few centimeters per minute (Tsuchiya et al., 1985) depending on the set-up used. The wire may be guided horizontally or vertically. Similar to hole drilling and 2D structuring, several materials can be machined using TW-ECDM: glass, quartz, alumina (Nesarikar et al., 1994; Peng and Liao, 2004; Tsuchiya et al., 1985), piezoelectric (PZT) ceramics (Singh et al., 1996; Tsutsumi et al., 1993), and various composites (glass- and Kevlar-epoxy) (Jain et al., 1991). NaOH is generally the preferred electrolyte. The applied voltage may be a direct current (DC) or pulsed voltage. Compared with hole drilling or 2D structuring, the voltage is generally higher (50 V and up), which is due to the different geometry of the wire (larger surface

6.5 WIRE ELECTROCHEMICAL DISCHARGE MACHINING

153

area) compared with a cylindrical tool-electrode. A direct consequence is that the surface quality is lower in this case than when machining with rod-shaped tools, since microcracks and thermal damages form on the machined surface (Peng and Liao, 2004). A solution to this issue was proposed by Han et al. (2011), who developed a new type of cutting tool by encapsulating a 50-mm brass wire between two ceramic tubes. By choosing adequate tool roughness (Ra ¼ 1.5 mm), they were able to machine soda-lime glass at 28 V, where excellent surface qualities (Ra ¼ 0.3 mm) resulted. Reduction of the temperature during machining can as well be achieved by appropriate flushing of the electrolyte. However, excessively high flushing rates can destabilize the discharge activity. Workpiece feeding is done by gravity feed, by constant speed (typically a few millimeters per minute for glass and 0.1 mm min
1 for ceramics), or by gap control (Peng and Liao, 2004). In the last configuration, the gap is controlled optically by a sensor. The thickness of the workpiece can be in the range of 1 mm to 1 cm (for glass). The gravity-feed mechanism generally results in poor cut shapes (Yang et al., 2006). The material removal rate and the machining overcut increase with the machining voltage (Singh et al., 1996; Tsuchiya et al., 1985) and electrolyte concentration (Tsuchiya et al., 1985; Yang et al., 2006), which also increases the probability of wire breaking (Tsuchiya et al., 1985). Polarizing the wire as a cathode generally results in higher material removal rates than for an anodically polarized wire. Efficient methods to decrease the machining overcut involve adding abrasive particles to the electrolyte (Yang et al., 2006) or using a pulsed voltage supply. At the same time, surface roughness is reduced for these methods where it can reach values below 1 mm.