Machining of non-conducting materials using ...

International Journal of Machine Tools & Manufacture 45 (2005) 1095–1108 www.elsevier.com/locate/ijmactool

Machining of non-conducting materials using electrochemical discharge phenomenon—an overview R. Wu¨thricha,*, V. Fasciob a Ecole Polytechnique Fe´de´role, Laboratoire de Syste`mes Robotiques, CH-1015 Lausanne, Switzerland Laboratoire de Ge´nie Chimique UMR 5503, Universite´ Paul Sabatier, 118 Route de Narbonne, 31062 Toulouse, Cedex 04, France

b

Received 25 October 2004; accepted 16 November 2004

Abstract Machining with electrochemical discharges is an unconventional technology able to machine several electrically non-conductive materials like glass or some ceramics. After almost 40 years of its first mention in literature, this technology remains an academic application and was never applied in industrial context. The knowledge about machining of non-conducting materials using electrochemical discharge phenomenon is reviewed up to this date with some particular attention to the electrochemical point of view. Some main limiting factors are highlighted and possible solutions are discussed. q 2004 Elsevier Ltd. All rights reserved. Keywords: Spark assisted chemical engraving; Machining non-conductive materials; Electrochemical discharge phenomena

1. Introduction The micro-electro-mechanical-system (MEMS) field is growing constantly. MEMS emerged in the late 1980s with the development of integrated circuits fabrication processes. If silicon remains the most widely used material, glass becomes more and more important. In particular Pyrexw glass is widely used because its ability to be bonded to silicon by anodic bonding (also called field-assisted thermal bonding or electrostatic bonding). According to recent publications ‘microfabrication of Pyrexw glass is one of the key processes in MEMS’ [1]. Several applications need glass because of its unique properties like its chemical resistance, transparency, low electrical and thermal conductivity or biocompatibility. As some examples can be mentioned micro-accelerometers, micro-reactors, micro-pumps and medical devices (flow sensors or drug delivery devices). The main limiting factor for a growing usage of glass in MEMS devices is its limited structuring possibility. Chemical etching technologies (like HF etching) are well * Corresponding author. Tel.: C41 21 693 38 10; fax: C41 21 693 38 66. E-mail address: [email protected] (R. Wu¨thrich). 0890-6955/$ - see front matter q 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijmachtools.2004.11.011

established, remain however too slow and expensive processes for many industrial applications. Other technologies are available like laser machining or mechanical machining (ultra-sonic machining or powder blasting). Both are hampered by the difficulty to obtain good surface qualities and potential structural damages (micro-cracks). In general high aspect-ratio structures are a challenging problem. Machining high-aspect ratio micro-holes in glass would open several new possibilities in the MEMS field. A possible answer is spark assisted chemical engraving (SACE), which is discussed in this contribution. SACE, based on electrochemical discharge phenomena, was presented for the first time in 1968 by Kurafuji as ‘electrical discharge drilling’ for micro-holes in glass [2]. Several other names are used in literature: ‘Discharge machining of nonconductors’ by Cook et al. [3], ‘ElectroElectrochemical Arc Machining’ by Kubota [4], ‘ElectroElectrochemical Discharge Machining’ by Ghosh et al. [5], ‘Micro Electrochemical Discharge Machining’ by Langen et al. [6], ‘Electro Chemical Spark Machining’ by Jain et al. [7] and ‘Spark Assisted Chemical Engraving’ by Langen et al. [8]. The diversity of names illustrates the complexity of the process and the different explanations (partially contradictory) proposed to explain the nature of

1096

R. Wu¨thrich, V. Fascio / International Journal of Machine Tools & Manufacture 45 (2005) 1095–1108

and electrochemical compound machining’. In this process, a side insulated tool-electrode is used and the gas film is not produced electro-chemically but by an external gas filling system. Liu et al. machined successfully non-conducting ceramic material by drilling holes of a few millimetres in diameter and length. Another variant is proposed by Allesu et al. [21,22] for thermocouple welding.

2. Micro-machining with SACE

Fig. 1. Schematics of the SACE set-up. Two electrodes, dipped into an electrolyte are supplied to a constant DC voltage. The cathode, with the smallest surface, is used as tool to machine the workpiece.

the electrical discharges. It should be mentioned that SACE has not to be confused with two other, even very similar, processes: electro chemical machining (ECM) and electro discharge machining (EDM) [2,3,9]. To avoid this confusions we will refer in the following to this process by SACE having in mind that this is not the only name used for it.1 Until today only partial reviews of this process are available [10–12]. The aim of this contribution is to review the literature up to this date. The practical implementation of SACE is the following (see Fig. 1): the sample to be machined is dipped in an appropriate electrolyte solution (typically sodium hydroxide or potassium hydroxide). A constant DC voltage or sometime a pulsed voltage is applied between the machining-tool or tool-electrode (cathode) and the counter-electrode (anode). The tool-electrode is dipped a few millimetres into the electrolyte. The counter electrode is far away from the tool electrode (a few centimetres typically) and has a much larger surface (about a factor hundred). Electrochemical discharges happen if the voltage is higher than a critical value which depends on the geometry and concentration of the used electrolyte. Typical values are around 30 V. At this point the density of gas bubble production is so high that they coalesce into a gas film isolating the tool electrode from the electrolyte. The electrical field in this film is high enough (typically 106–108 V/m [13]) to allow electrical discharges between the electrode and the electrolyte. The heat generated by these discharges and probably some chemical etching contribute to the eroding of the to be machined substrate if it is positioned in the near vicinity of the electrode (typically smaller than 25 mm for glass [14–16]). The SACE process itself gave raise very recently to some patents [17–19]. An interesting variant of the above described process was proposed by Liu et al. [20] as ‘gas filled electrodischarge 1

The present authors suggest to use following translations: e´tincelage assiste´ par attaque chimique (French), Funken-unterstu¨tzte chemische Gravur (German) and (Russian).

SACE offers various possibilities to machine different materials (see Table 1). It can be distinguished between hole drilling, 3D machining, travelling wire machining and hybrid techniques. In the following an overview of the reported investigations until this date is given. 2.1. General experimental investigations Machining with SACE is a complex process influenced by several parameters. Until today it is not yet clear which parameters control mainly the machining and even if the influence on machining of the various parameter is known reproducible machining is still not reported. A pioneering study about the influence of several parameters, like electrolyte properties, applied voltage and others on the material removal rate was reported by Cook et al. [3]. They described the main effects which were later confirmed by other research groups: the material removal rate increases with the applied D.C. voltage [23–30] and the electrolyte temperature [3,29,30]. Cook et al. [3] found that material removal rate increases with electrolyte concentration but latter investigations seems to indicates that there is an optimum before the rate decreases [23,28,31]. This behaviour follows the dependence of the electrolyte conductivity from the concentration. The tool wear rate and the over-cut follow a similar behaviour as the machining rate in function of the applied voltage and the used electrolyte [7,23,29]. However, the tool wear rate is about two magnitudes smaller than the material removal rate [23]. A possible mechanism participating to tool wear was described by Hof et al. [32]. During electrochemical discharges not only cathodic currents are Table 1 Overview of the published reports on SACE machining

Glass

Quartz Plexiglas Ceramics (mainly Al2O3) Composites Granit

Micro-hole drilling

3D structuring

TW-ECDM

100–500 mm [2,6,11,33,38,40,42,49,51], 1–2 mm [3,30,39,40,90] 1–3 mm [29,39] [3] 100–500 mm [37,36], 1 mm [28,29,30] [3]

100 mm!1 mm [6,38,40]

1–10 mm [44,46]

1 mm [7] [3]

Typical dimensions of machined structure are mentioned.

1–10 mm [46] 1–10 mm [46]

[7,23]

R. Wu¨thrich, V. Fascio / International Journal of Machine Tools & Manufacture 45 (2005) 1095–1108

observed but some small (a few milliAmperes) anodic currents. This anodic current dissolves the tool-electrode by electrochemical dissolution. In the case of extremely small sharp AFM tips this effect can result in the total dissolution of the tool-electrode. The effect of electrolyte on machining is complex and cannot be described uniquely in function of concentration and temperature. As will be discussed in Section 4, the machining process is partially a chemical one and therefore the nature of the electrolyte influences strongly the machining behaviour. NaOH electrolyte seems to have most interesting properties compared to other electrolytes (KOH, NaCl, NaNO3, NaF, HCl and H2SO4) [3,24]. Molten salt electrolytes (eutectic of NaOH and KOH melting at 170 8C) can drastically improve the smoothness of the machined surface [3]. The surface roughness of the machined work sample is influenced by the electrolyte and the applied voltage [33]. The voltage is mostly applied as a D.C. voltage. However applying high frequency voltage pulses is very interesting as removal rate increases for pulses in the micro-sec range and the machined surface quality is significantly improved [3]. Material removal rate is increased as well by introducing an additional inductance in the voltage generator circuit as was showed by Basak et al. [10,27,34]. Introducing artificially some bubbles into the process during machining was investigated by Jain et al. [23]. They found that the material removal rate decreases slightly as well as the over-cut. This method could maybe be used to increase machining precision. 2.2. Drilling holes Hole drilling was the first application of SACE [2]. The range of diameter is typically from about 100 mm to a few millimetres. Mostly some cylindrical tools are used made in various electrical conducting materials. In principle any conducting material is suitable. However, materials with high resistance to electrochemical corrosion are preferred like stainless steel, tungsten or nickel. The strong correlation between the tool shape and the machined hole was demonstrated by Langen et al. [6] who showed that it is possible to achieve various shapes (square, triangular and cylindrical shapes) for holes by using wire electro discharge grinding (WEDG) technology to manufacture the various tool-electrodes [35]. Another possibility to manufacture the tools is electrochemical etching [36]. Recently the possibility to use extremely sharp tips made by electrochemical etching implemented on an AFM set-up was explored [32]. The feeding mechanism of the tool influences the machining performances too. Most popular actuation principle for the tool (or workpiece) is gravity feed [2,3,5,28,33,37] which means that the tool and the machined substrate are always in contact. A similar possibility (keeping the force between tool and workpiece constant)

1097

was introduced in [38]. Other possibilities are constant feed [36,39] or stick-slip actuators [6]. A major problem in hole drilling is the accessibility of the electrolyte at the working site as the drill depth increases [38,40]. When the electrolyte can no longer reach the to be machined surface material removal rate decreases, the local temperature may increase and some thermal cracks will be formed in the workpiece. A solution to this problem is to move up and down the tool-electrode during drilling in order to allow fresh electrolyte to flush inside the hole [38,40]. By rotating the tool the machining performances can be improved as well for material removal rate as for obtaining better machined surface qualities. This was first demonstrated by Jain et al. [39] by drilling holes in the range of a few millimetres in diameter and depth in glass and quartz using various tool kinematics with NaOH electrolyte and later confirmed by other research groups [33,40]. The circularity of the machined holes and the machining removal rate can be improved by rotating the tool-electrode. The effect of tool rotational speed can be divided in two distinct regions. For slow rotation (less than 25 rpm) the machining efficiency (amount of machined distance for a given machining time) increases. For higher rotational speed (higher than 25 rpm) the efficiency decreases and may even become less efficient than for a non-rotational tool. Jain et al. [39] attribute this effect to a destabilisation of the sparking activity with increasing rotation speed. A similar conclusion was found when using an exocentric rotational motion for the tool. Optimal performances can be achieved by using an eccentricity similar to the used tool radius [39]. Not only glass and quartz may be drilled. As already showed Cook et al. [3] several non-conducting material can be machined by this technology (granite, refractory firebrick, aluminium oxide, plexiglas and others). In particular ceramic material can be drilled (Al2O3, Si3N4, MgO, Y2O3) [28,29,37,36] achieving holes of typically 1 mm diameter using gravity-feed or constant feed with several electrolytes (NaF, NaNO3, NaCl and NaOH) and applying relatively high voltages (around 80 V compared to typically 30 V for glass and quartz). The material removal rate (MRR) is, however, about 10 times smaller than for glass and quartz (about 0.1–0.4 mg/min) and the surface quality is less good than for glass [28]. In order to get better control over the distance between the workpiece and the tool, Jain et al. [30] proposed to use abrasive cutting tools. By comparing material removal rate for glass and alumina they could show that using this kind of tools improves the machining efficiency. The authors explain this effect by noting that on one side the machining takes advantage of thermal erosion and abrasive machining and on the other side the gap between the workpiece and the tool is always a few micrometers which allows sparking.

1098

R. Wu¨thrich, V. Fascio / International Journal of Machine Tools & Manufacture 45 (2005) 1095–1108

2.3. 3D micro-structuring By moving the tool over the work sample some 2D and even 3D structures can be achieved. First 3D micro-structuring experiments of glass, using different types of actuators, are reported by Langen et al. [6]. Besides drilling they presented some linear pattern machining with a tool-speed of 0.5 mm/s. Some other reports came from Wu¨thrich et al. [38,41,42] where they presented some linear structure with typical dimensions of 100 mm wide and a few millimetres long. The possibility to measure the machined structure and then to machine it again in order to correct partially the obtained shape was demonstrated and discussed [38]. Some very complex 3D structures can be obtained as demonstrated Lee et al. [40] by machining some hour glass and even threaded interconnects which allows to attach reversibly some plastic tubing able to withstand pressures up to 100 kPa without leakage. As for hole drilling, the rotation of the machining tool improves the machined surface quality and geometrical shape [16] (see Fig. 2). The linear speed of the tool above the surface should be chosen as high as possible in order to avoid thermal cracking of the sample. However, high machining speed results in decrease of the material removal rate [31,43]. 2.4. Travelling wire electrochemical discharge machining TW-ECDM An extension to wire machining, called travelling wire electrochemical discharge machining (TW-ECDM), was proposed by Tsuchiya et al. [44] and studied further on by Jain et al. [23,24,45] and Peng et al. [46]. TW-ECDM is particularly interesting for slicing glass fibber composites [7,23,47,48], but may be used as well for 2D contours cutting [44]. In this configuration a wire is used as tool in analogue way as in wire discharge machining (WEDM). Typical used wires are 0.9 mm cooper wire [23,46], 0.25 mm stainless steal wire [46], or 0.5 and 0.2 mm brass wire [24,44]. The wire speed is a compromise between high speed in order to allow the cooling of the wire (avoiding overheating and breaking) and low speed for economical reasons [23]. Typical speeds are a few millimetres per

minute [23] or a few centimetres per minute [44] depending on the used set-up. The wire may be guided horizontally [23] or vertically [24,44,46]. As for hole drilling and 3D structuring several materials can be machined with TW-ECDM: glass, quartz, alumina [44,46], PZT ceramics [24] and various composites (glassand Kevlar-epoxy) [23]. The preferred electrolyte is in general NaOH. The applied voltage may be D.C. voltage [23,24] or pulsed one [44,46]. Compared to hole drilling or 3D structuring the voltage is in general higher which is due to the different geometry of the wire (larger surface) compared to a cylindrical tool. Workpiece feeding is done by gravity [23], by constant speed [24] (at very slow speed in order to be slower than the mean machining speed) or, as newly introduced, by gap control [46]. In this last configuration, the gap is controlled optically by a sensor. The thickness of the workpiece can be in the range of 1 mm–1 cm (for glass) [46]. The stabilization of the temperature during machining by appropriate flushing of the electrolyte is possible [45,46]. 2.5. Combining with other technologies and practical realisations It is possible to combine SACE with other micromachining technologies. Esashi et al. [49] used SACE in combination with clean room technologies for microfabricating an absolute pressure sensor. Gue´rin et al. used SACE together with excimer laser micro-machining and clean room technologies to manufacture a miniature oneshot valve [50]. Daridon et al. [51] drilled holes of a few 100 mm diameter with SACE in combination with HF etching for micro-analytical applications. An other interesting combination, which allows to obtain various shapes for the tool-electrodes, is to manufacture them by WEDG [35] as it is first proposed by Langen et al. [8,41,52–55] and further developed by Yang et al. [33]. If moreover the tool fabrication facility is located on the same machine doing the SACE process, alignment problem can be avoided [53]. An other possibility to machine the tool on the same machine doing the SACE processing is

Fig. 2. Comparison between a groove machined (a) without and (b) with tool rotation. The sharpness of the edges is improved.

R. Wu¨thrich, V. Fascio / International Journal of Machine Tools & Manufacture 45 (2005) 1095–1108

1099

Fig. 3. A typical table top machine for SACE machining. A three axe manipulator moves the tool-electrode at the desired place for machining. In this case a tool holder with additional functionalities, like the possibility to machine at constant force or to use the tool as a profilometer, is mounted.

proposed by Lim et al. [36] who manufacture the toolelectrodes by electrochemical etching as used for STM or AFM applications. The developed desktop machinery combining SACE and other technologies may have some interesting applications in micro-factories [53]. A typical table top machine as used in our laboratory is shown in Fig. 3 [56]. It consists mainly on a three axe manipulator. In the example shown in the figure, a tool holder which can be used as profilometer in the same time is mounted (analogue to the one presented in [38]).

3. The SACE fundaments SACE makes use of electrochemical discharge (ECD) phenomena. These effects are known (often under other names like electrode effects, anode or cathode effects) since the work of Fizeau and Foucault [57,58] in 1844. Later, in 1851, Bunsen [59] showed that ECD does not only happen in aqueous electrolytes but in molten salt as well. Today they are intensively studied in the context of industrial aluminum production (where it is known under the keyword anode effect) [60,61] and for contact glow discharge electrolysis (CGDE) which is mainly used for material deposition (see [13] for a review on this subject). ECD appears when electrode effects happen in the system. Electrode effects are defined according to Vogt [60] as ‘the phenomenon of an immediate breakdown of

the electrolysis without any interference from outside’. This breakdown is due to the formation of a gas film around the electrode of smaller surface (the tool-electrode in the case of SACE). Even if SACE is based on ECD it is only very recently that this machining process was investigated in the light of electrochemistry and in particular in light of ECD [15,16,42]. 3.1. Phenomenology The electrochemical discharge phenomenon is clearly demonstrated by the following simple experience inspired from the historically first report from Foucault and Fizeau [57,58] (see Fig. 4). Two electrodes are dipped inside an aqueous electrolyte. The cathode is chosen with a much smaller surface (about 1 mm2) than the anode (about 100 times). When the applied D.C. voltage is low (lower than a critical value called critical voltage, which is typically between 20 and 30 V) electrolysis happens. Hydrogen gas bubbles are formed at the tool-electrode (cathode) and oxygen bubbles at the counter electrode (anode). When the voltage is increased, the current density increases too and more and more bubbles grow forming a bubble layer around the electrodes. The density (number per electrode area) of bubbles and their mean radius increase [16]. When the voltage is increased above the critical voltage, bubbles coalesce into a gas film around the tool-electrode. Light emission is observed in the film where electrical discharges

1100

R. Wu¨thrich, V. Fascio / International Journal of Machine Tools & Manufacture 45 (2005) 1095–1108

Fig. 4. Illustration of the electrochemical discharge phenomena. Two electrodes are dipped into an electrolyte. The voltage is progressively increased from 0 to 40 V. At around 25 V the cathode effect starts and at around 30 V the discharges are clearly visible. (a) 0 V (b) 7.5 V (c) 15 V (d) 30 V.

happen between the tool-electrode and the surrounding electrolyte [2,3,11,13,16,37,42,62–65]. Similar behaviour can be obtained by inverting the polarity of the electrodes and by changing the electrolytes [66]. The main open question is the mechanism of the transition between the traditional electrolysis regime and the electrochemical discharge regime. Among the many different explanations proposed in literature, there seems not yet to be a generally admitted explanation. The main mechanisms proposed can be schematised as follows [16,60,67]: (i) Change in wettability of the electrode [68]. The electrode effects are interpreted as the consequence of insufficient wetting of the electrode. The ability of the bubbles to adhere to the electrode surface is increased and they grow larger. The bubbles can coalesce and form a continuous gas film. (ii) Hydrodynamic instabilities (i.e. Helmholtz instability) are responsible for the onset of anode effect [69]. (iii) Local Joule heating [13,34,70–73]. The gas film is formed by local evaporation of the electrolyte by Joule heating (due to the increase of local current density because the bubbles electrochemically formed are shadowing the electrode). (iv) Combination of wettablity and hydrodynamic effects. Vogt [60,67] proposes that ‘anode effect occurs whenever the distance between neighbouring bubbles contacting the electrode has been diminished to such an extent that the bubbles are enable to coalesce’. He calculates an expression for the critical current density in function of the electrolyte flow around the electrode

and the contact angle of the adhering bubbles. His model shows that the critical current density depends on several parameters. The mains are: a. wettability of the electrolyte b. electrode geometry (area and typical length) c. thermodynamic state (temperature and pressure) d. bubble geometry e. bubble removal rate (v) Bubble coalescence as a percolation problem. A completely new approach was introduced by Wu¨thrich [16,62–64]. He proposes to consider the problem of bubble growth as a stochastic process and predicts the onset of the effect using percolation theory. The main idea is that bubbles can detach from the electrode surface as long as they are small. If a bubble formed by coalescence become as large as to be able to go round the electrode, it will no longer detach but grip even more to the electrode surface, because capillarity forces tend to minimize the surface of interfaces. This model is able to predict the complete I–U characteristics as well as the critical point (critical voltage, current density and resistance). The process initiating the electrochemical discharges is complex. It has to be mentioned that these different proposed mechanisms are not necessary exclusive but may contribute more or less together to this transition. 3.2. The mean current–voltage characteristics The electrochemical discharge process can be described quantitatively by the mean stationary current–voltage

R. Wu¨thrich, V. Fascio / International Journal of Machine Tools & Manufacture 45 (2005) 1095–1108

1101

The transition from the limiting current region to the instability region is characterized by the critical voltage Ucrit and critical current Icrit, linked by: U crit K Ud Z Rcrit I crit

(1)

crit

Fig. 5. The stationary mean current–voltage characteristics for an electrochemical cell as illustrated in the right-top corner.

characteristics in which can be distinguished five regions [2,10,11,15,26,42,62–65,70–72,74] (see Fig. 5): (i) Thermodynamic and overpotential region. For voltages smaller than Udy2 V (the water stability region) no current can flow. (ii) Ohmic region. In region AB the characteristics is nearly linear (typical voltage range: 2–10 V). (iii) Limiting current region. In region BC the mean current reaches a maximal limiting value depending on the tool-electrode geometry and the electrolyte. (iv) Instability region. In region CD the mean current decreases rapidly and a gas film is formed around the tool-electrode. This region is called instability region because the system may be in either a state similar to the limiting current region or the arc region. (v) Arc region. The region DE is characterized by the existence of a gas film around the tool-electrode. Light emission can be observed. Machining is done in this region.

with R the critical resistance. These three values (critical voltage, current and resistance) characterize the critical point at which the transition takes place. This transition is not restricted to the sudden decrease of the current I but makes a clear separation in the current transport mechanism. For voltages lower than Ucrit the current transport happens directly to the liquid electrolyte by electrochemical reactions. For voltages higher than Ucrit the current transport happens mainly through the gas film around the tool electrode by electrical discharges. Therefore the critical voltage makes the separation between two physically completely different situations and can be considered as a generalized phase transition [16]. The experimental value of the critical voltage is well studied quantitatively and is found to be function of the electrolyte concentration and electrodes geometry [16,26,34,62]. Its value decreases with increasing electrolyte concentration (Fig. 6a) and increases with the tool diameter (Fig. 6b). With increasing mean temperature of the electrolyte the critical voltage decreases [15]. 3.3. The normalized J–U characteristics It is possible to describe the current–voltage characteristics of SACE in a form independent of the electrode geometry and the electrolyte properties as was first pointed out by Wu¨thrich [11,16,–64]. Therefore the normalized current density J and the normalized voltage U are introduced: JZ

j jcrit

(2)

UZ

U K Ud U crit K Ud

(3)

Fig. 6. Dependence of the critical voltage of the a) electrolyte weight concentration and (b) the tool diameter. Measurements done for NaOH electrolyte.

1102

R. Wu¨thrich, V. Fascio / International Journal of Machine Tools & Manufacture 45 (2005) 1095–1108

can often not directly be observed as it is hidden inside turbulent electrolyte flow. The thickness of the gas film for a cylindrical electrode with a typical diameter of about 1 mm in an aqueous electrolyte is around 50–100 mm [71,75] or may even be larger (a few millimetres [70]). The mean temperature of the electrolyte seems to have strong effects on its size [70]. The higher the temperature is the thinner is the film. How other parameters, like electrolyte concentration, capillarity or wettability of the electrode, influence the gas film thickness is not reported until today. The size was until today never evaluated experimentally but some theoretical investigations where done [16,62–64] which indicates that the gas film grows quickly with increasing voltage above the critical one. Fig. 7. Normalized mean stationary current–voltage characteristics for NaOH electrolyte.

Fig. 7 shows the J–U characteristics where experimental measurements were done with various tools and different NaOH weight concentrations. It is seen that for voltages below the critical one, almost all measurements follows a same universal characteristics. For voltages higher than the critical one this behaviour seems still to be fulfilled even with less clarity. The normalized characteristics are not only interesting from a theoretical point of view, but also from a practical one. The measurement of the critical voltage together with the critical current density is enough to know the complete characteristics of the system without accurate knowledge of the electrolyte properties, which are in general difficult to determine (especially the electrolyte conductivity which depends highly of the electrolyte temperature and purity: two parameters that change during machining).

3.4.2. Physical parameters The composition of the gas film is studied by spectrographic analysis. The main results for aqueous electrolytes are summarized as follows [37,71,76]: 1. Always can be observed the spectral ray Ha and Hb from the hydrogen. The emission rays from the metal of the electrode (like Pt, Rh, Al, Cu, Au, Fe, Zn, Ag, Ni, W) can be observed for high enough voltages too. 2. The emission band of OH can be observed. 3. Several other emission rays are observed depending on the used electrolyte. For example in the case of a NaOH ˚ of the sodium is solution the emission ray of 4668 A observed. The observation of these various spectral lines does not only show the chemical composition of the gas film but proves that all these elements (or even molecules) are ionised. The temperature of the discharges are in the range of 800–20,000 K according several different measurements [13,74,77].

3.4. The gas film The gas film built around the tool electrode plays a key role in SACE machining. Several parameters characterise it: geometrical, physical and dynamical. Geometrical parameters are its size (fraction of the tool surface covered by the case film) and thickness. Physical parameters are its temperature and composition and dynamical parameters are values like the gas film mean live time. All these parameters influences the machining process. 3.4.1. Geometrical parameters Some information on the size and thickness of the gas film can be obtained by visual observations with high speed cameras and traditional photography in aqueous electrolytes [11,15,33,42,70,71,74,75] and molten salt electrolytes (see [69] for a more complete list of references). The observation is quiet difficult as on one side the phenomenon is a small turbulent scene and on the other side the gas film

3.4.3. Dynamical parameters Dynamical parameters of the gas film are from particular interest for SACE machining. A first important value is the mean time needed to build up the gas film. Fig. 8 shows a typical answer to a voltage step input. The gas film is completely built up after a mean time t, which is typically around 20 ms [15,16,71] for cylindrical tools. After this time, the first electrical discharges appear. The mean life time of the gas film is much less studied and today only qualitative reports are available [78]. It seems that the electrolyte mean temperature influences the mean life time [70]. Other influences are not reported. Another interesting aspect related to the dynamics of the gas film are the produced hydro dynamical fluxes (see Fig. 4d). Until now no systematic studies are reported. Only some visual observations are qualitatively discussed [15,79]. These fluxes are relatively stable and are described as a strong jet oriented in the tool-electrode direction.

R. Wu¨thrich, V. Fascio / International Journal of Machine Tools & Manufacture 45 (2005) 1095–1108

1103

(iii) The model tells nothing about what is happening before and after the critical point.

Fig. 8. Answer of a voltage step input (30 V). After a time t the gas film is completely built up and first electrical discharges take place.

Some finite element (FEM) calculations are reported in similar conditions [80]. A new approach, which aims to combine continuous and stochastic modelling is introduced by Mandin and Wu¨thrich [81].

3.5. Mathematical models For the machining applications mainly two elements are of particular interest. On one side the critical point (critical voltage Ucrit and current Icrit) is important to be known as machining becomes possible only for voltages beyond this point. To know which parameters influences in which manner Ucrit is from practical interest as it may be expected that machining with low voltages is more stable than for high voltages. On the other side the mechanism of the gas film formation and particularly the evolution of the gas film for voltages beyond the critical one is from interest. Understanding this mechanism may bring some tools to influence the stability of the gas film and therefore the reproducibility of machining. A first model predicting the critical point was given by Basak et al. [26,34,74]. A quantitative prediction of the critical voltage Ucrit and the critical current density jcrit is possible. The main idea is to consider that at the critical point the bubble production rate is given by the sum of the electrolysis gas production and the vapour production by local Joule heating. Supposing that the gas production rate at the critical point is a constant independent of the current density and that the critical resistance is known, the authors could express the critical voltage and current density in function of the electrolyte mass concentration. Their model has however three major leaks: (i) The critical resistance is not predicted by the model and has to be used as external input. (ii) The model is not able to explain the dependence of the critical voltage from the electrode geometry.

The newer model from Wu¨thrich et al. [16,62–64], which considers the formation of the gas film responsible for the onset of electrochemical discharge phenomena as a stochastic process, fills partially these leaks. Using percolation theory, the critical voltage, current density and resistance are predicted. The model subdivides the lateral surface of the tool-electrode in a lattice. An occupied site models a growing gas bubble on the lateral tool surface and is occupied with a mean probability p. In general p depends on the voltage and current density. These adhering bubbles shadow the electrode and therefore increase the interelectrode resistance Ro. Neglecting the contribution of the bubble layer around the tool, the apparent resistance is modelled as: RZ

Ro 1 Kp

(4)

The bubble coalescence effect is considered in this model with the concept of clusters. A cluster is defined as a group of neighbouring occupied sites. Each cluster is interpreted as one bubble. The size s of the bubble is given by the size (number of sites belonging to the cluster) of the cluster. From percolation theory it is known that if the mean occupation probability is greater than the percolation threshold pc, a cluster of infinite size appears [82]. This is interpreted as the appearance of the gas film around the tool-electrode. Therefore, the critical resistance is given by: Rcrit Z

Ro 1 K pc

(5)

From numerical simulations [82], it is known that for a square lattice, the percolation threshold is pcy0.59. It follows: Rcrit Z 2:5Ro

(6)

This relation is confirmed by Fascio [15] experimentally in the case of a NaOH electrolyte and an electrode of about 1 mm diameter. The model from Wu¨thrich et al. [16,62–64] is able to explain the dependence of the critical voltage from the external parameters like tool radius or electrolyte concentration. As the most interesting consequence the gas film covering the tool-electrode is considered as a dynamical structure, explained in term of the infinite cluster in percolation theory. This means that even for voltages higher than the critical one the tool is not totally covered by the gas film but almost all new nucleating bubbles belong to this gas film [64]. This dynamical process allows the gas film to remain stable in the sense that he does not leave abruptly the tool electrode surface. An improved version of this model does no longer need relation (4). The idea is that in steady state condition, the amount of gas produced electrochemically on the free

R. Wu¨thrich, V. Fascio / International Journal of Machine Tools & Manufacture 45 (2005) 1095–1108

1104

electrode surface is equal to the amount of gas released from the electrode [83,84]: k1 ð1 K pÞ Z k2

smax X

sns

(7)

sZ0

where p is the density of active bubble nucleation sites. The constant k1 is given by the amount of produced gas electrochemically (by the Buttler–Volmer equation) and k2 is related to the mean high of a bubble (therefore related to the wettability of the electrode). The right side of Eq. (7) evaluates the total volume of gas released from the electrode. It runs over all bubble from size sZ0 up to the maximal size smax at which the bubble can still leave the electrode surface (this value is therefore function of wettability, ability of bubbles to coalesce and capillarity forces). In the case where every bubble, expect the gas film, can leave the electrode surface, the current is given by: X I Z k2 sns Z k2 ½p K PðpÞ (8) s

with P(p) the size of the gas film (the probability that an active nucleation site belongs to the gas film, which is identical to the infinite cluster in percolation theory). If the so called phase equation pZp(U), giving the density of active nucleation sites in function of the applied voltage, is known, relation (8) can be used to predict the complete I–U characteristics of the process. In principle this relation results from (7) if the constants k1 and k2 can be expressed explicitly in function of the control parameters. However, until today there is no such relation reported. Only some phenomenological models where proposed [16,62–64,84]. The understanding of the microscopic processes responsible of activation of bubble nucleation sites would be of great practical help. If there is a way to reduce the mean distance between these sites it would become possible to reduce the size of the gas film, which is typically of the size of this mean distance if the gas bubbles are nearly spherical. The interest of reduction of the gas film size is clear. This would increase the local electrical field around the tool electrode and therefore increase the discharge activity and material removal rate. This will probably as well give the possibility to machine smaller structures. And finally, even if this small gas film will change dynamically its size, this will not affect too much the machining. For example a gas film of a thickness of about 1 mm, can change its size only in this range and therefore influence machining only in the micrometer range. Therefore, more reproducible machining can be expected. First steps in the attempt to understand how to reduce the gas film thickness were presented recently in [85–87].

region. In this situation several processes may contribute to the machining mechanism [5,23]: (i) melting and vaporisation due to electrochemical discharges (ii) high temperature dissolution (iii) differential expansion of constituents and weathering (iv) random thermal stresses and micro-cracking and spalling (v) mechanical shock due to expanding gases and electrolyte movement Besides all these potential mechanism until today only thermal effects and chemical etching were investigated. 4.1. The arc region Machining is possible above the critical voltage. The current behaves totally different compared to the voltages below the critical value. In mean the current is very small (only a few milliAmpere). But the instant current I(t) signal is a succession of very short current pulses (typically a few 100 ms) which can be as high as a few Amperes [11,15,16,42,46,71,74,88] (see Fig. 9). The signal has no measurable frequency and the number of pulses per fixed time interval follows a Poisson process [11,16]. Each pulse is a short electrical discharge [16,88]. The current signal is well modelled by a shot noise signal, from which it follows that the mean current I is given by [16]: I Z ld

Sa t

(9)

with S the tool-electrode surface, ld the probability of discharge (number of discharges per time), a the mean amplitude of a discharge and t the mean time of a discharge. It is interesting to note that the mean current gives essentially the behaviour of the discharge probability. This quantity is important to investigate the nature of the electrical discharges as the physic of an electrical discharge is essentially contained in the electron yield (or discharge probability). Experimental measurements of ld [11,16] shows that this value follows a field emission law which is typical for non-self sustained thermo emitted arc discharges. Other experimental evidences confirm this hypothesis: Guilpin [71] showed that the cathode spot has

4. The machining mechanism Machining with electrochemical discharge phenomena is possible only above the critical voltage, i.e. in the arc

Fig. 9. Typical example of discharges at 30 V.

R. Wu¨thrich, V. Fascio / International Journal of Machine Tools & Manufacture 45 (2005) 1095–1108

all characteristics (temperature, size and mean current density) of typical arc discharges and several works [15,26,70,71] showed that the temperature of the tool-electrode as at least 100 8C, a necessary condition for arc discharges that are thermo initiated. Finally it has to be noted that arcs, contrary to Townsed discharges, happen at low voltages (typically from 10 V up). 4.2. Thermal machining The most admitted hypothesis about the machining mechanism is that the machining is a thermal one. It is assumed that the workpiece surface is intensively heated leading to melting and maybe even vaporisation of the work material. Several analytical [27] and FEM [11,15,89] calculations support this mechanism. The difficulty is to estimate correctly the amount of energy per discharge. Basak et al. [27] proposed, based on some analogies with electro discharge machining (EDM) and telecommunication switches, to consider that each discharge carries in mean an energy of 2000 J/cm2 and has a duration of 0.1 ms. Jain et al. [89] estimated the mean heat q* released by the discharges by following relation: q Z UI K RI 2

(10)

with U the applied voltage, I the mean current and R the resistance of the electrolyte. In words: the mean energy per spark is the mean energy given to the system minus the energy loosed by Joule heating. Fascio [11,15] improved this idea by estimating the mean energy and mean duration of a spark from statistical analysis of the current signal in the arc region. All the proposed models reproduce fairly well the experimental observations of the material removal rate, except for the lowest voltages (below 30 V). In this case, the machining rate is much smaller. This suggests that the machining is mainly chemical. Nevertheless several experimental evidences confirm the thermal mechanism in machining. It is known that the toolelectrode may reach temperatures up to 500 8C [26,34,70]. Allesu et al. [5] showed by a simple experiment that the heat produced during electrochemical discharge phenomenon is able to melt glass. Therefore, they used an electrolysis cell separated into two compartments by a glass wall. In this glass wall is located a small hole of 1.5 mm diameter. When a voltage of 60 V is applied to the cell, the hole inside the glass wall rapidly files with a gas bubble formed because of local Joule heating of the electrolyte. Electrical discharges are observed inside the gas and the diameter of the hole increases from 1.5 to 2.5 mm within a few minutes. Kulkarni et al. [9] showed by various measurements that after each current peak, i.e. each discharge, the temperature of the workpiece increases above the melting temperature and sometime even above the vaporisation temperature of the machined material. They estimated that about 77–96% of the energy supplied to the process is used to heat

1105

the electrolyte and only 2–6% is used for heating up the workpiece. Another experimental evidence for thermal mechanism in machining are the various observations of thermal cracks insides the machined materials [23,24,28,38,39,90] which are seen mainly at high voltage machining. In summary the thermal model for SACE machining is straight forward. Each discharge supplies energy to the workpiece, which is able to melt or even evaporate it. As the number of discharges per time follows a field emission law, the material removal rate follows a similar law in function of the applied voltage between the electrodes. This is confirmed experimentally where several measurements shows that the material removal rate increases nearly exponentially with the applied voltage [3,7,23,37]. 4.3. Chemical machining Several observations indicate that not only thermal machining has some importance but some chemical etching too [14,15,24,33,37]. Chemical etching effects could for example explain why the quality of machined surface may depend strongly on the used electrolyte [3,33] or why the NaOH electrolyte leads to higher machining removal rate of glass than other ones [33]. It is well known that NaOH etches glass by complexation of silicate [5,11,14,15,33,40]: 2NaOH C SiO2 / Na2 SiO3 ðsÞ C H2 O

(11)

This chemical reaction is strongly enhanced by increasing the temperature [15,33]. Similar results are reported for ceramic materials [37]. The chemical contribution to machining is a combination of etching and leaching phenomena. Yang et al. defined the process as a high temperature etching [33] and conducted several experiments in order to elucidate the chemical etching effects in SACE. By comparing visually the surface quality obtained at high temperature etching of glass with NaOH, traditional electro discharge machining of Indium Tin Oxide (ITO, an electrically conduction glass) and SACE processing of glass they concluded that SACE is mostly like to be a combination of thermal melting and chemical etching. EDX surfaces analysis of the machined structures for sodalime glass confirmed that [Na C] concentration decreases while [HC] concentration increases in the machining area [14] and suggests that leaching occurs during the SACE process. Beyond these observations, it is still difficult to quantify the effect of chemistry on machining. Previous investigations were achieved by Fascio [15], who showed that the maximal chemical contribution would be in the order of several microns. This result highlights the effect of chemistry which could be seen as a local modification of the roughness.

1106

R. Wu¨thrich, V. Fascio / International Journal of Machine Tools & Manufacture 45 (2005) 1095–1108

5. Conclusion Electrochemical discharges can be used to machine several electrically non-conductive materials. The machining mechanism is a combination of thermal and chemical machining, where the thermal effect clearly dominates. Even known since almost 40 years, this machining process remains an academic application and was until now never applied in industrial production. The research done until today mainly focused on experimenting the machining of various materials and investigating the effect of different parameters on the material removal rate. It was shown that a large class of materials (glass, quartz, various ceramics and others) can be machined. Not only simple structures as holes but as well as very complex structures like threads can be machined. Combination with conventional clean room technologies is possible too. Material removal rates depend on a large number of parameters like material to be machined, used electrolyte, applied voltage and temperature. It is, however, only recently that machining by electrochemical discharges was investigated from the electrochemical point of view. This aspect may bring some new findings on one side the fundamental understanding of the process and on other side on the practical implementation of the process. If this non-conventional machining process wants to become interesting for industrial applications it is absolutely essential that reproducible machining is obtained. For micro-machining application the reproducibility should be at least a few microns. One main challenge in reaching this goal is certainly the control of the gas film built around the tool-electrode in which happen the discharges. Not only that this gas film is necessary for machining to occurs, but the stability and dynamics of this film conditions the machining, in particular its resolution and its reproducibility. The electrochemical point of view may certainly bring some interesting inputs for this problematic.

Acknowledgements This work was supported by the Swiss National Foundation for Research.

References [1] Li. Xinghua, A. Takashi, L. Yongxum, E. Masayoshi, Fabrication of high-density electrical feed-throughs by deep-reactive-ion etching of Pyrex glass, Journal of MEMS 11 (2002) 625–630. [2] H. Kurafuji, K. Suda, Electrical discharge drilling of glass, Annals of the CIRP 16 (1968) 415–419. [3] N.H. Cook, G.B. Foote, P. Jordan, B.N. Kalyani, Experimental studies in electro-machining, Transactions of ASME, Journal of Engineering for Industry 1973; 945–950.

[4] M. Kubota, Proceedings of the International Conference on Production Engineering, Tokyo, 1974, p. 51. [5] K. Allesu, A. Ghosh, M.K. Muju, Preliminary qualitative approach of a proposed mechanism of material removal in electrical machining of glass, European Journal of Mechanical Engineers 36 (1992) 202–207. [6] H. Langen, J.-M. Breguet, H. Bleuler, Ph. Renaud, T. Masuzawa, Micro electrochemical discharge machining of glass, International Journal of Electrical Machining 3 (1998) 65–69. [7] S. Tandon, V.K. Jain, P. Kumar, K.P. Rajurkar, Investigations into machining of composites, Precision Engineering 12 (1990) 227–238. [8] H. Langen, V. Fascio, R. Wu¨thrich, D. Viquerat, Three-dimensional structuring of pyrex glass devices—trajectory control, International Conference of the European Society for Precision Engineering and Nanotechnology (EUSPEN) 2 Eindhoven 2002; 435–438. [9] A. Kulkarni, R. Sharan, G.K. Lal, Measurement of temperature transients in the electrochemical discharge machining process, Temperature: Its Measurement and Control in Science and Industry 7 (2003) 1069–1074. [10] A. Ghosh, Electrochimical discharge machining: principle and possibilities, Sadhana 22 (1997) 435–447. [11] V. Fascio, R. Wu¨thrich, H. Bleuler, Spark assisted chemical engraving in the light of electrochemistry, Electrochimica Acta 49 (2004) 3997–4003. [12] R. Wu¨thrich, H. Bleuler, Glass microstructuring with spark assisted chemical engraving. The Fourth Korea–Switzerland Joint Symposium, Les Diablerets, Switzerland, Feb. 1–3 2004, pp. 51–54. [13] A.L. Yerokhin, X. Nie, A. Leyland, A. Maetthews, S.J. Dowey, Plasma electrolysis for surface engineering, Surface and Coatings Technology 122 (1999) 73–93. [14] V. Fascio, R. Wu¨thrich, D. Viquerat, H. Langen, 3D Microstructuring of glass using electrochemical discharge machining (ECDM). International Symposium on Micromechatronics and Human Science (MHS’. 99), 1999, pp. 179–183. [15] V. Fascio, Etude de la microstructuration du verre par e´tincelage assiste´ par attaque chimique: une approche electrochimique, Dissertation Thesis (2691), Swiss Federal Institute of Technology, EPF Lausanne, 2002. [16] R. Wu¨thrich, Spark assisted chemical engraving—a stochastic modeling approach, Dissertation Thesis (2776), Swiss Federal Institute of Technology, EPF Lausanne, 2003. [17] C.T. Yang, et al., Method for making a probe device having detective function. US Patent 6,156,188, 2000. [18] K. Soo-Hyun, L. Young-Mo, L. Hyung-Jun, Japanese Patent 2002346837, 2002. [19] C.T. Yang, H.Y. Tsai, T.C. Wu, Method and device for electrochemical discharge processing with self-acting bubble layer, US Patent 0110502A1, 2002. [20] Y. Liu, Z. Jia, J. Liu, Study on hole machining of non-conducting ceramics by gas-filled electrodischarge and electrochemical compound machining, Journal of Material Processing Technology 69 (1997) 198–202. [21] K. Allesu, Electrochemical discharge phenomena in manufacturing process, PhD Thesis, I.I.T. Kanpur, 1988. [22] A. Ghosh, M.K. Muju, S. Parija, K. Allesu, Microwelding using electrochemical discharge, International Journal of Machine Tools and Manufacture 37 (1997) 1303–1312. [23] V.K. Jain, P.S. Rao, S.K. Choudhury, K.P. Rajurkar, Experimental investigations into travelling wire electrocheimical spark machining (TW-ECSM) of composites, Transactions of ASME, Journal of Engineering for Industry 113 (1991) 75–84. [24] C. Tsutsumi, K. Okano, T. Suto, High quality machining of ceramics, Journal of Material Processing Technology 37 (1993) 639–654. [25] Y.P. Singh, V.K. Jain, P. Kumar, D.C. Agrawal, Machining piezoelectric (PZT) ceramics using an electrochemical spark machining (ECSM) process, Journal of Materials Processing Technology 58 (1996) 24–31.

R. Wu¨thrich, V. Fascio / International Journal of Machine Tools & Manufacture 45 (2005) 1095–1108 [26] I. Basak, A. Ghosh, Mechanism of spark generation during electrochemical discharge machining: a theoretical model and experimental verification, Journal of Material Processing Technology 62 (1996) 46–53. [27] I. Basak, A. Ghosh, Mechanism of material removal in electrochemical discharge machining: a theoretical model and experimental verification, Journal of Material Processing Technology 71 (1997) 350–359. [28] B. Bhattacharyya, B.N. Doloi, S.K. Sorkhel, Experimental investigations into electrochemical discharge machining (ECDM) of nonconductive ceramic materials, Journal of Materials Processing Technology 95 (1999) 145–154. [29] V.K. Jain, S.K. Chak, Electrochemical spark trepanning of alumina and quartz, Machining Science and Technology 4 (2000) 277–290. [30] V.K. Jain, S.K. Choudhury, K.M. Ramesh, On the machining of alumina and glass, International Journal of Machine Tools and Manufacture 42 (2002) 1269–1276. [31] R. Wu¨thrich, V. Fascio, D. Viquerat, H. Langen, Study of spark assisted chemical engraving-process technology data, International Conference of the European Society for Precision Engineering and Nanotechnology (EUSPEN) 1 Eindhoven, 2002, pp. 265–268. [32] L.A. Hof, A. Lal, K. Fujisaki, R. Wu¨thrich, H.H. Langen, H. Bleuler, 3D Micro structuring of glass with an AFM, The Fourth EUSPEN International Conference, May 30th–3rd June 2004, Glasgow, UK, pp. 98–99. [33] C.T. Yang, S.S. Ho, B.H. Yan, Micro hole machining of borosilicate glass trough electrochemical discharge machining (ECDM), Key Engineering Materials 196 (2001) 149–166. [34] I. Basak, Electrochemical discharge machining mechanism and a scheme for enhancing material removal capacity. PhD Thesis, Indian Institute of Technology, Kanpur, 1991. [35] T. Masuzawa, M. Fujino, K. Kobayashi, T. Suzuki, Annals of the CIRP 34 (1985) 431. [36] H.-J. Lim, Y.-M. Lim, S.H. Kim, Y.K. Kwak, Self-aligned micro tool and electrochemical discharge machining (ECDM) for ceramic materials, Proceedings of SPIE 4416 (2001) 348–353. [37] H. Tokura, I. Kondoh, M. Yoshikswa, Ceramic material processing by electrical discharge in electrolyte, Journal of Material Science 24 (1989) 991–998. [38] R. Wu¨thrich, V. Fascio, D. Viquerat, H. Langen, In situ measurement and micro-machining of glass, International Symposium on Micromecatronics and Human Science (MHS’99), Nagoya, 1999, pp. 185–191. [39] N. Gautam, V.K. Jain, Experimental investigations into ECSD process using various tool kinematics, International Journal of Machine Tools and Manufacture 38 (1998) 15–27. [40] E.S. Lee, D. Howard, E. Liang, S.D. Collins, R.L. Smith, Removable tubing interconnects for glass-based micro-fluidic systems made using ECDM, Journal of Micromechanics and Microengineering 14 (2004) 535–541. [41] H. Langen, I. Ceausoglu, M. van der Meer, E. Lehmann, H. Bleuler, Ph. Renaud, Electrochemical micromachining of glass using MicoEDMed Microtools, Proceedings of Ultraprecision in Manufacturing Engineering, 26–30 May Braunsweig 1997, p. 672. [42] V. Fascio, H.H. Langen, H. Bleuler, Ch. Comninellis, Investigations of the spark assisted chemical engraving, Electrochemistry Communications 5 (2003) 203–207. [43] V. Fascio, R. Wu¨thrich, K. Fujisaki, D. Viquerat, H. Langen, H. Bleuler. Spark assisted chemical engraving: a novel technology for glass microstructuring, Invited lecture in European Congress on Advanced Materials and Processes (EUROMAT), Lausanne, 2003. [44] H. Tsuchiya, T. Inoue, M. Miyazaiki, Wire electro-chemical discharge machining of glasses and ceramics, Bulletin Japanese Society of Precision Engineering 19 (1985) 73–74.

1107

[45] V.V. Nesarikar, V.K. Jain, S.K. Chandhury, Travelling wire electrochemical spark machining of thick sheets of Kaevlar-Epoxy composites, in Proceedings of the Sixteenth AIMTDR Conference, 1994, pp. 672–677. [46] W.Y. Peng, Y.S. Liao, Study of electrochemical discharge machining technology for slicing non-conductive brittle materials, Journal of Material Processing Technology 149 (2004) 363–369. [47] S. Tandon, Machining of composites—a new approach, MTech Thesis, IIT Kanpur (India), July 1987. [48] V.K. Jain, ECSM—a new way to machine polymer composites, in Proceedings of INCARF-93 Conference IIT, Delhi, December 1993. [49] M. Esashi, Y. Matsumoto, S. Shoji, Absolute pressure sensors by airtight electrical feedthrough structure, Sensors and Actuators A21–A23 (1990) 1048–1052. [50] L.J. Gue´rin, O. Dubochet, J.F. Zeberli, P. Clot, Ph. Renaud, Miniature one-shot valve, Micro Electro Mechanical Systems—IEEE Eleventh Annual International Workshop Proceedings, 1998, pp. 425–428. [51] A. Daridon, V. Fascio, J. Lichtenberg, R. Wu¨thrich, H. Langen, E. Verpoorte, N.F. De Rooij, Multi-layer microfluidic glass chips for microanalytical applications, Fresenius Journal of Analytical Chemistry 371 (2001) 261–269. [52] H. Bleuler, J.-M. Breguet, Microfactory and microrobotics activities in Europe, International Workshop on Microfactories, Tsukuba, Japan, 1998, pp. 157–164. [53] H. Bleuler, R. Clavel, J.-M. Breguet, H. Langen, H.E. Pernette, Issues in precision motion control and microhandling robotics and automation, 2000, Proceedings of ICRA ’00 IEEE International Conference on Robotics and Automation, San Francisco, CA, 24–28 April 2000, pp. 959–964. [54] H. Bleuler, R. Clavel, J.M. Breguet, H. Langen, Y. Bellouard, Applications of microrobotics and microhandling, RIKEN Review 36 (2001) 26–28. [55] H. Langen, V. Fascio, R. Wu¨thrich, T. Ejiri, D. Viquerat, Development of desktop machinery for three-dimensional microstructuring of pyrex glass devices, International Workshop on Microfactories, Minneapolis, USA, 2002, pp. 53–56. [56] K. Fujisaki, R. Wu¨thrich, Ph. Couthy, L.A. Hof, H. Bleuler, SACE in microfactory, International Workshop on Microfactories (IWMF 2004), Shanghai, 2004. [57] H. Fizeau, L. Foucault, Ann. de Chimie et de Physique XI 3e`me se´rie, 1844, p. 370. [58] M.L. Foucault, Experiments with the light of voltaic arc, Journal of Franklin Institute 48 (1849) 50–52. [59] R. Bunsen, Annalen der Physik 92 (1851) 648. [60] H. Vogt, Contribution to the interpretation of the anode effect, Electrochimica Acta 42 (1997) 2695–2705. [61] H. Vogt, J. Thonstad, The voltage of alumina reduction cells prior to the anode effect, Journal of Applied Electrochemistry 32 (2002) 241–249. [62] R. Wu¨thrich, H. Bleuler, A model for electrode effects using percolation theory, Electrochimica Acta 49 (2004) 1547–1554. [63] R. Wu¨thrich, V. Fascio, H. Bleuler, A stochastic model for electrode effects, Electrochimica Acta 49 (2004) 4005–4010. [64] R. Wu¨thrich, Ch. Comninellis, H. Bleuler, The instability region in electrode effects, Journal of Applied Electrochemistry, submitted for publication. [65] V. Fascio, R. Wu¨thrich, D. Viquerat, H. Langen, C. Comninellis, Study of spark assisted electrochemical etching, 51th Annual Meeting of the International Society of Electrochemistry, Warsaw, 2000. [66] S.K. Sengupta, The effect of electrolyte constituents on the onset and location of glow discharge electrolysis, Journal of Electroanalytical Chemistry 127 (1981) 263–265. [67] H. Vogt, The anode effect as a fluid dynamic problem, Journal of Applied Electrochemistry 29 (1999) 137–145. [68] K. Arndt, H. Probst, Zeitschrift fur Elektrochemie 29 (1923) 323. [69] B. Mazza, P. Pedeferri, G. Re, Hydrodynamic instabilities in electrolytic gas evolution, Electrochimica Acta 23 (1978) 87–93.

1108

R. Wu¨thrich, V. Fascio / International Journal of Machine Tools & Manufacture 45 (2005) 1095–1108

[70] H.H. Kellogg, Anode effect in aqueous electrolysis, Journal of Electrochemical Society 97 (1950) 133–142. [71] Ch. Guilpin, Contribution a` l’e´tude de l’effet de cathode dans les solutions aqueuses d’e´lectrolytes, Dissertation Thesis, Universite´ de Paris VI, France, 1975. [72] J. Garbarz-Olivier, Ch. Guilpin, The origin of the electrode effect in various electrolytes, Journal of Electroanalytical Chemistry 91 (1978) 79–91. [73] J.C. Valognes, J.P. Bardet, P. Mergault, Contribution a` 1’e´tude des effets d’e´lectrode, Spectrochimica Acta 42B (1987) 445–458. [74] V. Raghuram, T. Pramila, Y.G. Srinivasa, K. Narayanasamy, Effect of the circuit parameters on the electrolytes in the electrochemical discharge phenomenon, Journal of Material Processing Technology 52 (1995) 301–318. [75] Ch. Guilpin, J. Garbarz-Olivier, Les effets d’e´lectrodes dans les solutions aqueuses. Re´sultats obtenus a` l’aide d’une came´ra rapide, Journal de Chimie Physique 75 (1978) 723–726. [76] Ch. Guilpin, J. Garbarz-Olivier, Analyse de la lumie`re e´mise aux e´lectrodes pendant les effets d’e´lectrodes, dans des solutions aqueuses d’e´lectroylte, Spectrochimica Acta 32B (1977) 155–163. [77] V. Reghuram, Dissertation Thesis, Indian Institute of Technology, Kanpur, 1994. [78] R. Wu¨thrich, V. Fascio, D. Viquerat, H. Langen, Study of spark assisted electrochemical etching—force measurements, International Workshop on Microfactories (IWMF 2000), Fribourg, 2000, pp. 201–204. [79] V. Fascio, H. Langen, H. Bleuler, Ch. Comninellis, Spark assisted chemical engraving: a novel technology for glass microstructuring, 54th Annual Meeting of the International Society of Electrochemistry, Sa˜o Pedro, Brazil, 2003, Keynote Lecture. [80] Ph. Mandin, G. Picard, D. Lincot, H. Roustan, Aluminium electrolysis: hydrodynamical, thermochemical and electrochemical modelling, 54th Annual Meeting of the International Society of Electrochemistry, Sa˜o Pedro, Brazil, 2003, p. 178.

[81] Ph. Mandin, J. Hambourger, S. Sanchez, G. Picard, R. Wu¨thrich, V. Fascio, H. Bleuler, Electrochemical engineering of bubble effect during electrolysis process: continuous and stochastic modelling, 205th ECS Meeting, San Antonio, USA, 2004. [82] D. Stauffer, A. Aharony, Introduction to Percolation Theory, Taylor & Francis, London, 1998. [83] R. Wu¨thrich, E.A. Baranova, H. Bleuler, Ch. Comninellis, A phenomenological model for macroscopic deactivation of surface processes, Electrochemistry Communications 6 (2004) 1199–1205. [84] R. Wu¨thrich, Ch. Comninellis, H. Bleuler, Electrochemistry under extreme current densities—Application to electrode effects, In Proceedings of the 55th Annual Meeting of the International Society of Electrochemistry, Thessaloniki, 19–24 September 2004. [85] L.A. Hof, R. Wu¨thrich, K. Fujisaki, H. Bleuler, Ph. Mandin, G. Picard, Physical principles and miniaturization of spark assisted chemical engraving (SACE), International Workshop on Microfactories (IWMF), Shanghai, 2004. [86] R. Wu¨thrich, L. Hof, The gas film in spark assisted chemical engraving (SACE)—a key element for micro-machining applications, International Journal of Machine Tools and Manufacture, submitted for publication. [87] L.A. Hof, 3D-Microstructuring of glass, Master Thesis, TU Delft, The Netherlands. [88] A. Kulkarni, R. Sharan, G.K. Lal, An experimental study of discharge mechanism in electrochemical discharge machining, International Journal of Machine Tools and Manufacture 42 (2002) 1121–1127. [89] V.K. Jain, P.M. Dixit, P.M. Pandey, On the analysis of the electrochemical spark machining process, International Journal of Machine Tools and Manufacture 39 (1999) 165–186. [90] S.H. Somashekhar, K. Sadashivappa, Micro drilling of non conductive materials using electro chemical spark machining (ECSM), Proceedings of the 14th Annual Meeting of the American Society for Precision Engineering Monterey, USA, Oct 31–Nov 05 1999, pp. 188–191.