An analysis of the discharge mechanism in electrochemical discharge machining of particulate reinforced metal matrix composites

ARTICLE IN PRESS International Journal of Machine Tools & Manufacture 50 (2010) 86–96

Contents lists available at ScienceDirect

International Journal of Machine Tools & Manufacture journal homepage: www.elsevier.com/locate/ijmactool

An analysis of the discharge mechanism in electrochemical discharge machining of particulate reinforced metal matrix composites J.W. Liu a, T.M. Yue a,n, Z.N. Guo b a b

The Advanced Manufacturing Technology Research Centre, Department of Industrial and Systems Engineering, The Hong Kong Polytechnic University, Hun Hom, Hong Kong Faculty of Electromechanical Engineering, Guangdong University of Technology, Guangzhou 510090, PR China

a r t i c l e in f o

a b s t r a c t

Article history: Received 12 June 2009 Received in revised form 8 September 2009 Accepted 8 September 2009 Available online 12 September 2009

An analysis of the discharge mechanism in electrochemical discharge machining (ECDM) of a particulate reinforced metal matrix composite was undertaken, and a model to reveal the electric field acting on a hydrogen bubble in ECDM process has been established. The model was found capable of predicting the position of the maximum field strength on the bubble surface as well as the critical breakdown voltage for spark initiation, for a given processing condition. A set of experiments was performed to verify the model and the experimental results agreed well with the predicted values. The experimental results also showed that an increase in current, duty cycle, pulse duration or electrolyte concentration would promote the occurrence of arcing action in ECDM. Moreover, by studying the waveform of ECDM and surface craters, it is confirmed that the spark action is in the form of an arc. Compared to EDM, the volume of an arc eroded crater of ECDM was less than that of EDM. An XRD analysis of the phases of the EDM and ECDM specimens showed that the Al4C3 phase was detected on the former but not on the latter. & 2009 Elsevier Ltd. All rights reserved.

Keywords: ECDM Composite Discharge Waveform Breakdown voltage

1. Introduction In shaping difficult-to-machine materials, hybrid processes which combine the actions of electrical discharge machining (EDM) and electrochemical machining (ECM) have been found to be able to increase the machining efficiency of the respective individual process. One of these processes, electrochemical arc machining (ECAM), has been explored for machining electrically conductive composite materials [1]. The process incorporates material removal by electrochemical action as well as by electric arc. The productivity of ECAM is reported to be five to fifty times greater than the productivity using individual processes of ECM and EDM alone [2,3]. However, the material-removal mechanism is still not fully understood. A similar method to ECAM is electrochemical spark machining (ECSM), which again is an electrical-based hybrid process primarily used for shaping electrically non-conductive materials [4]. In ECSM, the material of the workpiece is removed by the heat produced by sparking in the vicinity of the workpiece. In fact a number of studies have been conducted to analyse the electrochemical discharge machining (ECDM) process as well as the material-removal mechanism [1,4–6]. Among these, the early work of Crichton and McGeough [1] involved high-speed photography to analyse the various stages of the discharging process.

n

Corresponding author. Tel.: +852 27666601; fax: + 852 23625267. E-mail address: [email protected] (T.M. Yue).

0890-6955/$ - see front matter & 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijmachtools.2009.09.004

On the other hand, Allesu et al. [7] attempted to establish a classification of the various material-removal processes based on the ECD phenomenon. Despite a considerably amount of experimental research being conducted on the study of ECDM, relative little theoretical modelling work has been performed. With regard to modelling the ECDM process, Basak and Ghosh [5] developed a simplified idealistic model of the process capable of predicting the critical voltage and current for spark initiation. They treated the discharge phenomenon as a switching off process, in which the current drops to zero for a very short time. Taking a different approach, Jain et al. [4] have considered each gas bubble as a valve, which after its breakdown due to a high electric field produces a discharge in the form of an arc; their study mainly focussed on the aspects of energy and current during sparking. Although the models proposed by Basak and Jain have shed useful light on the discharge mechanism in electrochemical discharge machining, Kulkarni et al. [6] argued that their models cannot explain the observation of the dependence of the arc on the immersion depth of the tool. Moreover, in Basak’s work, despite the theoretical predicted values comparing quite well with experimental observations, the model developed still has some ¨ major deficiencies according to a recent review paper by Wuthrich and Fascio [8]. It is apparent that the basic mechanism of the process is not yet completely understood and more research studies are deemed necessary. Turning to non-conductive and composite materials, much interest has been generated in applying ECDM on these materials [4,8]; however, the research undertaken was mainly focused on

ARTICLE IN PRESS J.W. Liu et al. / International Journal of Machine Tools & Manufacture 50 (2010) 86–96

experimental aspects and only a few investigations were on theoretical modelling. To the authors’ knowledge, a study on modelling ECDM for metal matrix composites (MMCs) has yet to be seen. It is generally accepted that MMCs are much more difficult-to-machine than their monolithic counterparts, whether or not conventional or unconventional techniques are used [9–15]. Using non-traditional machining techniques, such as laser [11,12] and water jet [13] machining, can achieve a fairly high materialremoval rate, but would often be accompanied by some serious surface and subsurface defects, which in many cases are unacceptable to a final finish product, and could undermine the fatigue strength of the final product. Among the many nonconventional machining methods [12–15], EDM and wire-EDM, perhaps, are the most promising processes for shaping MMCs when the flexibility on the shaping geometry is considered. Notwithstanding the advantages of using EDM in shaping MMCs, it is obvious that the problem of low machining efficiency due to frequent encountering of unstable machining conditions has to be overcome [16]. It is envisaged that the ECDM process which operates under a wider spark gap could provide a stable processing condition and hence raises the machining efficiency. With the above background in mind, a research programme has been launched to study the process of ECDM of particulate MMCs, and this paper focuses on the modelling of the mechanism of the spark generation. Other aspects of the study, such as the detailed material-removal mechanism and removal rate will be presented in a separate paper.

2. Theoretical analysis Previously, Basak and Ghosh [5] modelled the ECDM discharge phenomenon as a switching off process, while Kulkarni et al. [6] have studied in some detail the ECDM mechanism experimentally. In Basak’s model, it is hypothesised that, as a consequence of vapour blanketing of the electrode, a switching off situation occurs and the current through the circuit, within a very short time span, drops to zero. Similarly, Kulkarni proposed that when an isolating film of hydrogen gas bubble covers the cathode tip portion in the electrolyte, the tip is covered by a gaseous layer. At this time, a large dynamic resistance is present and the current through the circuit becomes almost zero. However, according to experiment observations and analysis, this phenomenon only exists in some special situations, such as when the energy of the power supply is low and/or the electrode area is small. Jain et al. [4] also have the same opinion that the current is need not to drop to zero or almost zero. Taking a different approach to analyse spark initiation, Jain et al. [4] put forward an arc discharge valve model to explain the ECSM process, and it has satisfactorily explained some of the experimental results. However, the model mainly focuses on the analysis of the start point of sparking, and the bubble effect on the distribution of the electric field has not been studied. Recognising this, the present research aims to deepen understanding of the ECDM mechanism in the machining of MMCs. In the analysis, the effects of bubbles and the reinforcement phase on the electric field are considered. In this study, the gap between the electrodes is assumed to be constant. The bubble distribution on the electrode and in the gap is considered to be random, and a hundred percent bubble coverage of the cathode is not assumed. The energy of the power supply is sufficient for the ECDM process to be continuous. Normally, the ECDM consists of three stages. During stage I, which is essentially a pure ECM process, the resistance between the anode and cathode increases gradually because of an increase of the number of hydrogen bubbles resulting from the ECM effect. As a result, the voltage between the anode and cathode will increase

87

accordingly. Once it reaches the breakdown voltage, a spark would be produced; the value of this critical breakdown voltage can be obtained by using the model developed in this study. The following sections present details of the theoretical analysis of the model.

2.1. Analysis of the discharge mechanism Fig. 1 shows a schematic diagram of the ECDM process, which is the subject of this study, in which the workpiece is the anode and the tool is the cathode. During processing, an electric double layer is assumed to exist on both surfaces of the workpiece and the tool (Fig. 1b). For the cathode, initially, the electric double layer is in a stable condition since there is no current flowing to the cathode (Fig. 2a). When the process starts, electrons will flow towards the cathode (Fig. 2b), and most of these electrons will charge up the electric double-layer capacitor of the tool. When the electric double-layer capacitor has been fully filled, electrons are used to produce H2 (Fig. 2c). Fig. 3 presents the equivalent circuit of the ECDM process, where C1 and C2 represent the electric double-layer capacitors of the anode and cathode, respectively. R1 and R3 represent the resistance of the ECM reactions at the anode and cathode, respectively. R2 is the resistance of the electrolyte. R is an external resistor and Rn is used to represent the resistance between the tool and the workpiece. During the ECDM process, some bubbles will break and disappear, but more bubbles will be produced, and as such the open area between the workpiece and the tool will decrease. According to Eq. (1), Rn will increase (where L is the gap size between the two electrodes; S is the surface area of the exposed electrodes; r is the resistivity of the electrolyte). As a result, the voltage between the cathode and anode will increase accordingly, and if the voltage reaches the breakdown voltage, sparks will be produced and the voltage will drop to a value that will just be sufficient to maintain arcing. R pr

L S

ð1Þ

If fa is the applied voltage, and j2 is the breakdown voltage and j is the arc maintaining voltage, then the voltage waveform

Fig. 1. A schematic diagram showing (a) the overall ECDM process, where R represents an external resistor, Rn is the resistance between the tool and the workpiece; (b) the electric double layers exist on both surfaces of the workpiece and the tool.

ARTICLE IN PRESS 88

J.W. Liu et al. / International Journal of Machine Tools & Manufacture 50 (2010) 86–96

Fig. 2. Electric double layer of the cathode (a) initially, the electric double layer is in a stable condition; (b) when the ECDM process starts, electrons will flow towards the cathode and (c) when the electric double-layer capacitor has been fully filled, H2 gas starts to form.

Fig. 4. Three possible voltage waveforms occurring in ECDM: (a) typical ECDM waveform; (b) waveform for cases where the voltage increasing rate is low and (c) waveform for cases of short pulse-on time.

2.2. Critical breakdown voltage

Fig. 3. Equivalent circuit of the ECDM, where C1 and C2 represent the double-layer capacitors of the anode and cathode, respectively. R1 and R3 represent the resistances of anode and cathode ECM reactions, respectively. R2 is the resistance of the electrolyte. R is an external resistor, and Rn represents the resistance between the tool and the workpiece.

for ECDM could resemble the form as shown in Fig. 4(a), in which the pulse-on-time is ton. During Stage I, the voltage between the anode and cathode will increase gradually. Once it reaches the breakdown voltage, a spark will be produced and the voltage will drop to the arc maintaining voltage immediately. At this point in time, Stage II sets in, and the arc discharge will persist till the end of the pulse-on-time if the energy of the power supply can sustain the current at the arc maintaining voltage. At the end of the pulse-on-time, due to the electric capacity effect, the voltage between the anode and cathode will gradually drop to zero (Stage III). The voltage of the capacitance (jc) between the two electrodes can be determined by Eq. (2), where js is the steady-state voltage of the capacitance and t is a time constant

jc ¼ js ð1  eðt=tÞ Þ

In this study, the power supply is ensured as being capable of providing enough energy for generating an arc at the breakdown site. Under such a condition, the breakdown voltage (j2), which is a critical parameter for the study of the ECDM mechanism, can be determined. To comprehend an arc discharge event, the effects of hydrogen bubbles that are present between the two electrodes and the ceramic phase existing in the workpiece, on the electric field strength during ECDM are investigated. Fig. 5 shows a schematic diagram for modelling the electric field in ECDM. To begin with, the voltage between the two electrodes can be described by Laplace’s Equation:     @ @v 1 @ @v 1 @2 v þ r2 j ¼ sin y ¼0 ð3Þ r2 þ @r @r sin y @y @y sin2 y @f2 The general solution of this equation in spherical coordinates can be represented by Eq. (4), which can be found in related mathematical physics books [17]:  X Bnm jðr; y; fÞ ¼ Anm r n þ n þ 1 Pnm ðcos yÞcos mf r n;m   X Dnm þ Cnm r n þ n þ 1 Pnm ðcos yÞsin mf ð4Þ r n;m

ð2Þ

However, in the case where the voltage increasing rate is low, during the pulse-on-time, the voltage will not reach j2 (Fig. 4b), and as such, sparking will not occur. Should this occur, the voltage waveform will resemble that of Fig. 4(b). Likewise, if the pulse-ontime is too short, the voltage will fall short of j2 and again sparking will not occur (Fig. 4c).

where Anm, Bnm, Cnm, Dnm are arbitrary constants. Under the ECDM condition, due to the symmetry of the bubble, the electric potential j can be described as [17]  X Bn j¼ An r n þ n þ 1 Pn ðcos yÞ ð5Þ r n

ARTICLE IN PRESS J.W. Liu et al. / International Journal of Machine Tools & Manufacture 50 (2010) 86–96

89

Fig. 5. A schematic model for studying the electric field in ECDM of a particulate reinforced composite (Ex is the external electric field outside the bubble, ja is the electric potential across the bubble of a length a).

Furthermore, using Eq. (5), the potentials outside the bubble (jout) and inside the bubble (jin) can be described as follows:  X Bn jout ¼ An r n þ n þ 1 Pn ðcos yÞ ð6Þ r n

jin ¼

X

Cn r n þ

n

Dn rn þ 1



Pn ðcos yÞ

ð7Þ

P1 ðcos yÞ ¼ cos y;

ð15Þ

Now, according to boundary condition (iii) and for P1(cos

y)= cos y, i.e. Eq. (8), the following is obtained:

jout -  Ew r cos y ¼  Ew rP1 ðcos yÞ

2

P2 ðcos yÞ ¼ ð3 cos y  1Þ=2

A1 ¼  Ew ; ð8Þ

Also the following boundary conditions apply: (i) The electric potential at the bubble surface is equal to:

jout ¼ jin ðwhen r ¼ R and R is the bubble diameterÞ

ð16Þ

@jout @j ¼ e in @r @r

Using Eq. (7) and according to boundary condition (iv), Dn must be zero. Now, let A1 = Ex, An =0 (n a1) and Dn = 0, Eqs. (14) and (15) can be simplified to

ð9Þ

ð10Þ

where e0 and e are the dielectric constants of the electrolyte and the hydrogen bubble. (iii) The electric potential is uniform in the space far away from the bubble, i.e.

jout -  Ew r cos y ðwhen r-1Þ

An ¼ 0ðn a 1Þ

Ew RP1 ðcos yÞ þ

(ii) At the bubble surface, the electric induction intensity in the normal direction is equal to

e0

  @jin X Dn ¼ Cn nRn1 þðn þ 1Þ n þ 2 Pn ðcos yÞ @r R n

Using Eqs. (6) and (16), A1 and An can be determined,

where Pn(cos y) has the form of a Legendre Function: P0 ¼ 1;

¼

ð11Þ

Ew P1 ðcos yÞ 

X Bn X P ðcos yÞ ¼ Cn Rn Pn ðcos yÞ nþ1 n R n n

X ðn þ 1ÞBn n

Rn þ 2

1 X nCn Rn1 Pn ðcos yÞ 80 n

Pn ðcos yÞ ¼

ð17Þ

ð18Þ

Based on Eqs. (17) and (18), the following equations are obtained: B1 ¼ C1 R R2

ð19Þ

2B1 1 ¼ C1 80 R3

ð20Þ

Ew R þ

Ew 

With Eqs. (19) and (20), B1 and C1 can be evaluated When the effect of bubble density and the particles on the electric field is considered, the external electric field outside the bubble can be expressed by Ex (as shown in Fig. 5). In fact, these effects are primarily reflected on the value of d. (iv) The electric potential in the centre of the bubble has a finite value of,

jin a 1 ðwhen r ¼ 0Þ

ð12Þ

(v) The electric potential across a bubble of length ‘‘a’’ is determined by both jin and jout.

ja ¼ ðjin 9R0 þ jout 9aR Þ

ð13Þ

jout ¼

n

n

An R þ

Bn Rn þ 1



Pn ðcos yÞ ¼

79 Ew R3 ; 161

C1 ¼ 

240 Ew 161

Furthermore, according to Eqs. (17) and (18) when n a1, Bn and Cn should be equal to zero. Having obtained A1, An (n a1), B1, Bn (n a1), C1, Cn (n a1) and Dn, Eqs. (6) and (7) can be written as

jout ¼  Ew r cos y þ jin ¼ 

79 Ew R3 cos y 161 r2

240 Ew r cos y 161

ð21Þ

ð22Þ

Based on Eq. (22), Ein is formulated

Using Eqs. (6) and (7), and according to boundary condition (i), the following equation is obtained: X

B1 ¼

X n

 Dn Cn R þ n þ 1 Pn ðcos yÞ ¼ jin R n

ð14Þ

It is known that e0 E80, e E1 [18], and using Eq. (14) and based on boundary condition (ii), the following equation is obtained:  X @jout Bn 80 ¼ 80 nAn Rn1 þ ðn þ 1Þ n þ 2 Pn ðcos yÞ @r R n

Ein ¼ 

@jin @jin 240 ¼  ¼ Ew @z @ðr cos yÞ 161

ð23Þ

Now, according to boundary condition (v) and using Eqs. (21) and (22), Eq. (13) thus becomes  ! 240 79 Ew R3 cos y a ð24Þ ja ¼  Ew R cos y þ Ew r cos y þ R  161 161 r2 

ja ¼  Ew a cos y þ

79 Ew R3 cos y 161 a2

 ð25Þ

ARTICLE IN PRESS 90

J.W. Liu et al. / International Journal of Machine Tools & Manufacture 50 (2010) 86–96

Since, 2ja ¼  E0 d

and

a¼

d 2 cos y

ð26Þ

where E0 is the original electric field strength. Then using Eqs. (25) and (26), the following equations are obtained:   79 Ew R3 cos y ¼  E0 d ð27Þ 2 Ew a cos y þ 2 161 a and Ex becomes Ew ¼

E0 d d þ 316R3 cos3 y=161d2

ð28Þ

And according to Eq. (28), it is apparent that the maximum value of Ex (Max E) occurs at y = 901, which is equal to E0. Then Eq. (23) becomes 240 240 Max Ew ¼ E0 161 161 Eq. (29) can also be written as

Max Ein ¼

ð29Þ

161 ð30Þ Max Ein 240 Given that the dielectric strength of hydrogen is about 3  106 V/m (i.e. Max Ein), and based on Eq. (30), when E0 reaches 2.013  106 V/m, the bubble will breakdown; this value is independent of the interactions that the reinforcement particle may have on the electric field. Nevertheless, the reinforcement phase will affect the amount, distribution and possibly the size of the hydrogen bubbles. To verify the model and to determine the breakdown voltage experimentally, experiments were conducted under different processing conditions and the results are given in the following section. According to the model, when the gap between the electrodes is 15 72 mm, the predicted breakdown voltage j2 should range between 26.2 and 34.2 V. E0 ¼

3. Experimental methods Fig. 6a shows a schematic diagram of the in-house built ECDM equipment. It has a tool holder which is fixed to the Z-axis, onto which a cylindrical steel tool of a diameter 5 mm is held. During the course of processing, the current flows through a conduction bush that is placed inside an insulating bearing housing, to the steel tool (Fig. 6). Such a design is to confine the working current to the processing area. During the course of machining, the current flows from an electric transmission system, placed on a plastic insulated seat, to the tool. The resistance of the electric transmission system is negligible. A pulse number controllable electrical source has been designed for the experiments, in which the current can be adjusted without changing the applied voltage. The average current output of this electrical source can be adjusted from 0.5 to 5 A. The pulse duration ranges between 4 and 400 ms, while the duty cycle can be operated between 1:1 and 1:10. Some major specifications of the ECDM equipment are given in Table 1. To facilitate the study of the ECDM process, the experimental setup allows the number of electrical pulse to be preset. In this study, the workpiece material, i.e. the MMC, is a particulate reinforced aluminium alloy 359 with 20-vol% SiC. The material was supplied in the form of 10-mm-thick rolled plates with the reinforcement phase having a nominal size of 10 mm. The workpiece is placed on an insulated pad. The electrolyte used for the experiment was NaNO3 solution, and its concentration was varied between 1 and 1.6 wt%. The electrolyte is pumped into the processing area through a nozzle with the aid of a circulation system. After the electrolyte has been properly filtered, it returns

Fig. 6. (a) The ECDM setup [working Table 1, electrolyte system (2), Y-axis (3), Xaxis (4), nozzle (5), Z-axis (6), tool holder (7), electric transmission system (8), tool (9), workpiece (10), slide rail (11), electrolyte bath (12), control system (13), power source (14)]. (b) Showing the design of the spindle where an electric insulating bearing system is employed to insulate the shaft from other components of the spindle [ceramic ball bearing (1), insulation cover (2), shaft (3), insulating bearing housing (4), tool (5)].

to the electrolyte circulation system. In the experiment, the effects of current, pulse duration, duty cycle and electrolyte concentration on ECDM were studied. The various conditions employed are given in Table 2. Moreover, a comparison of the discharge waveforms was made between the ECDM and EDM processes.

4. Experimental results 4.1. ECDM and EDM waveforms A comparison of the waveforms was made between ECDM and EDM of the SiC particulate reinforced composite. For the EDM

ARTICLE IN PRESS J.W. Liu et al. / International Journal of Machine Tools & Manufacture 50 (2010) 86–96

91

Table 1 Some major specifications of the ECDM equipment.

Travel (mm) Repeatability position accuracy (mm) Position accuracy (mm) Speed (rpm) Gripping range (mm) Power (kW) Average current (A) Voltage (V) Pulse duration (ms) Duty cycle Medium Circulation flow (L/min)

X-axis

Y-axis

Z-axis

250 5 15

250 5 15

100 1 2

Spindle

Electrical source

Electrolyte bath

0–20,000 Ø0.5–10 1.5 0.5–5 20–120 4–400 1:1–1:10 Emulsion electrolyte 0–20

Table 2 Processing conditions. Applied voltage (V)

Pulse duration (ms)

1 2 3

110

24 40 72

1 2 3 4

110

1 2 3 4 1 2 3

Processing conditions

Duty cycle

Electrolyte concentrate (wt%)

Current (A)

1:6

1

5

72

1:6

1.6

2 3 4 5

110

72

1:10 1:8 1:6 1:4

1.6

5

110

24

1:6

1 1.4 1.6

5

A

B

C

D

Fig. 7. Typical waveforms of (a) applied voltage, (b) ECDM (machining Condition A2) and (c) EDM.

process, the machining conditions were similar to those of Condition A2 (Table 1), except that an emulsion medium was used instead of the NaNO3 solution and the spark gap size was 5 mm. Fig. 7 shows the typical waveforms of the applied voltage, the ECDM voltage and the EDM voltage. According to Fig. 7b, the three stages that are expected of a typical ECDM waveform (Fig. 4a) were obtained. The recorded breakdown voltage was about 30 V which compared well with the theoretical predicted value of 26.2–34.2 V; the experimental results also show that the arc maintaining voltage was about 20 V. It was also found that the

peak voltage in ECDM was significantly lower than the applied voltage. This indicates that the current between the two electrodes was not zero. On the other hand, the waveform resulting from the EDM process shows that Stage I and Stage II in the ECDM waveform were absent (Fig. 7c). This is considered to be simply due to the high electrical resistance of the emulsion used in EDM. Moreover, due to the fact that the dielectric strength of the emulsion is significantly higher than that of the hydrogen bubbles formed in the ECDM process, a much higher breakdown voltage (110 V) was recorded than in the ECDM process (30 V)

ARTICLE IN PRESS 92

J.W. Liu et al. / International Journal of Machine Tools & Manufacture 50 (2010) 86–96

despite the fact that the former has a much smaller spark gap. Nonetheless, the fact that the two processes have a similar maintaining voltage of about 20 V after breakdown indicates that the discharge mode of these two processes is in the form of arc. 4.2. Parameters studies Figs. 8–11 show the effects of current, duty cycle, pulse duration and electrolyte concentration on the resulting ECDM waveforms. The results show that an increase in current, duty cycle, pulse duration or electrolyte concentration would promote the action of ECDM. This can be recognised by the presence of the Stage II ECDM mechanism in the waveforms when these parameters were increased to a certain value. The results further show that under the conditions of this study, when the values of current, duty cycle, pulse duration and electrolyte concentration reached a level of 5 A, 1:6, 40 ms, 1.4%, respectively, the Stage II mechanism was observed. This could be explained on the basis that an increase of these parameters would lead to an increase of the number of hydrogen bubbles which would cause the voltage between the two electrodes to be increased. This creates a favourable condition for the bubbles to reach the breakdown voltage and thus initiates the occurrence of sparks. The breakdown voltage obtained for all the conditions of this study ranged between 26 and 30 V, which agrees well with the theoretical predictions. 4.3. Craters and debris Fig. 12(a) and (b) show typical single pulse machined surfaces of MMCs by ECDM and EDM, respectively. The craters resulting

from single pulse ECDM and EDM resemble a circular shape with a rough projection at the circumference; the morphology is typical of an arc effect. This observation evidences that the discharge mode of both the ECDM and EDM processes is in the form of an arc. The crater volume was measured using a three-dimensional optical device (Alicona IFM G4). The equipment measured the volume of the cavity below the workpiece surface. Five measurements were taken for both processes, and the measurements for ECDM were all lower than those of the EDM. The average values obtained for ECDM and EDM were 5.8  105 and 8.1 105 mm3, respectively. It is apparent that under a similar single-pulse processing condition, the EDM process resulted in the removal of a larger volume of material. Since for ECDM, the total energy is roughly divided into two parts, the ECM energy and the arc energy, therefore it is within expectation that the crater volume of ECDM is smaller than that of EDM. It is considered that, due to the nature of ECM, which is a time consuming dissolution process, the ECM phase would not increase the material-removal rate considerably. Therefore, in theory, the EDM process is likely to have a higher material-removal rate than the ECDM process. However, in practice, the overall material-removal rate of ECDM was found to be higher than that of EDM in machining particulate MMCs [19]. This is made possible due to a more stable machining condition can be obtained for the ECDM process, for a large machining gap is present and thus improves the condition for the ceramic reinforcement to be washed away from the gap. Fig. 13a and b show some of the large debris that had been collected during ECDM and EDM. For both the ECDM and EDM experiments, a continuous pulsing time of 3 min was employed. For both cases, irregular angular debris was obtained. This indicates that spalling is a major material-removal mechanism, which is caused by thermal stresses; such mechanism has

Fig. 8. Effect of current on ECDM spark waveform (machining Condition B) [(a) 2A, (b) 3A, (c) 4A, (d) 5A]. When the current reached a level of 5A, the Stage II mechanism became apparent.

ARTICLE IN PRESS J.W. Liu et al. / International Journal of Machine Tools & Manufacture 50 (2010) 86–96

93

Fig. 9. Effect of duty cycle on ECDM spark waveform (machining Condition C) [(a) 1:10, (b) 1:8, (c) 1:6, (d) 1:4]. When the value of duty cycle reached 1:6, the Stage II mechanism became apparent.

Fig. 10. Effect of pulse duration on ECDM spark waveform (machining Condition A) [(a) 24 ms, (b) 40 ms, (c) 72 ms]. When the value of pulse duration reached 40 ms, the Stage II mechanism became apparent.

Fig. 11. Effect of electrolyte concentration on ECDM spark waveform (machining Condition D) [(a) 1%, (b) 1.4%, (c) 1.6%]. When the value of electrolyte concentration reached a level 1.4%, the Stage II mechanism became apparent.

ARTICLE IN PRESS 94

J.W. Liu et al. / International Journal of Machine Tools & Manufacture 50 (2010) 86–96

Fig. 12. Craters generated by single pulse (a) ECDM and (b) EDM. The processing condition was based on A2, but for EDM, an emulsion medium was used instead of the NaNO3 solution and the spark gap size was 5 mm.

Fig. 13. Large debris particles collected from (a) ECDM process and (b) EDM process. For both cases, irregular angular debris was obtained. The processing condition was A2, except that for EDM, an emulsion medium was used instead of the NaNO3 solution and the spark gap size was 5 mm.

Fig. 14. A typical ECDMed surface, showing the presence of re-solidified debris and micro-cracks.

occurred to the Nd–Fe–Ba alloy and the ZrB2–Cu composite material during EDM processing [20,21]. In examining both the ECDMed and EDMed surfaces, it is obvious that re-solidification had occurred. Moreover, some re-solidified debris can also be found on the surfaces. Therefore melting and vaporisation should also be another major material-removal mechanism apart from spalling. Micro-cracks observed in both the surfaces support the idea that spalling is an important material-removal mechanism (Fig. 14). The similarity in morphology of the debris resulting from these two processes support the idea that the discharge modes of the ECDM and EDM processes are alike and in the form of an arc. The debris size was measured using a particle size distribution

analyser (Horiba CAPA-700). Five measurements were taken for both debris collected after the EDM and ECDM experiments, the median particle diameter and the maximal particle diameter of the former were both found to be larger than those of the latter (Fig. 15). The median diameter of the debris for EDM and ECDM was 35 and 23 mm, respectively, while the maximal diameter for EDM was above 100 mm and which for ECDM was about 80 mm. In Fig. 13, only those relatively large particles were compared. The smaller median and maximal debris sizes for the ECDM process indicate that the arc energy of ECDM is likely to be smaller than that of the EDM process. Again, this can be argued from the total energy point of view, as discussed above. To study any phase change that might have occurred during ECDM and EDM, the XRD patterns of the machined surfaces of both the ECDM and EDM specimens and of the base material were obtained. A comparison of the XRD patterns shows that an Al4C3 phase is present in the EDM specimen but not in the ECDM specimen (Fig. 16) and the parent material. An examination of the EDM specimen confirms that the needle-like Al4C3 phase is present in areas directly next to the machined surface (Fig. 17). This indicates that aluminium had reacted with the SiC particles due to the arc heating effect in the EDM process. On the other hand, the Al4C3 phase was not detected in the ECDM specimen. This is probably due to the fact that the arc energy of the ECDM process was lower than that of the EDM process. As a result, the re-melt zone and the heat-affected zone would be smaller, and the amount of aluminium carbides formed greatly reduced and below the detection limit. Another reason could be that the ECM effect of the ECDM process could dissolve material of the machined surface, and Al4C3, even if formed, could have be debonded from the matrix.

ARTICLE IN PRESS J.W. Liu et al. / International Journal of Machine Tools & Manufacture 50 (2010) 86–96

95

Fig. 16. The XRD patterns obtained for (a) ECDM specimen and (b) EDM specimen. The Al4C3 peak was observed for the EDM specimen.

Fig. 17. Showing the microstructure in regions close to the EDMed surface, in which needle-like Al4C3 phase was observed.

performed to verify the proposed model. Major findings of the study are summarised as follows:

Fig. 15. Measurements of the distribution of the particle size of the debris collected from the (a) EDM experiment and (b) ECDM experiment.

5. Conclusions This paper presents the results of the first phase of the study of ECDM of a particulate reinforced metal matrix composite. It focuses on the modelling of the discharge mechanism with an emphasis on the prediction of the critical breakdown voltage of the hydrogen bubble. Moreover, a set of experiments was

(i) The electric field acting at the surface of the hydrogen bubble was analysed; the maximum field strength occurred at an angle of 901 perpendicular to the vertical line (L) drawn between the two electrodes (Fig. 5). (ii) For a given processing condition, the model is capable of predicting the breakdown voltage of a discharge, which in the present study was between 26.2 and 34.2 V. The model shows that the breakdown voltage does not have a relationship to the presence of the reinforcement phase. (iii) Experiments were conducted to verify the theoretical breakdown voltage. The experimental breakdown voltage

ARTICLE IN PRESS 96

(iv)

(v)

(vi)

(vii) (viii)

J.W. Liu et al. / International Journal of Machine Tools & Manufacture 50 (2010) 86–96

was found to lie between 26 and 30 V, which matches well with the predicted values. The experimental results also show that in the ECDM process, the voltage waveform consists of a charging phase, then breakdown which is followed by a period of arcing, with the arc maintaining voltage similar to that of an EDM discharge. The effects of current, duty cycle, pulse duration and electrolyte concentration on the discharge mechanism were studied. An increase in these parameters would promote the sparking action occurring in ECDM. The craters of ECDM and EDM were studied. It is shown that the formation mechanisms of EDM and ECDM craters are the same, which is believed to be due to the arc effect. The observation of the debris shows that spalling is a major material-removal mechanism for both ECDM and EDM. An Al4C3 phase was detected on the surface of the EDM specimen but not on the ECDM specimen. This could be due to the fact that the amount of Al4C3 formed in ECDM is considerably lower than that formed in EDM. Moreover, the dissolution action of ECM might cause the Al4C3 to be debonded from the matrix.

Acknowledgement This research was fully funded by the Research Committee of the Hong Kong Polytechnic University under project account code RGQL. References [1] I.M. Crichton, J.A. McGeough, Studies of the discharge mechanisms in electrochemical arc machining, Journal of Applied Electrochemistry 115 (1985) 113–119. [2] H. El-Hofy, J.A. McGeough, Evaluation of an apparatus for electrochemical arc wire-machining, Journal of Engineering for Industry, Transactions ASME 110 (2) (1988) 119–123. [3] J.W. Xu, N.Z. Yun, J.Y. Wang, J.A. Tian, W.J. Xu, in: Electrochemical Machining Technology, National Defence Industry Press, 2008.

[4] 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. [5] I. Basak, A. Ghosh, Mechanism of spark generation during electrochemical discharge machining: a theoretical model and experimental investigation, Journal of Materials Processing Technology 62 (1996) 46–53. [6] 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. [7] K. Allesu, A. Ghosh, M.K. Muju, A preliminary qualitative approach of a proposed mechanism of material removal in electrical machining of glass, European Journal of Mechanical Engineers 36 (3) (1991) 201–207. ¨ [8] R. Wuthrich, V. Fascio, Machining of non-conducting materials using electrochemical discharge phenomenon—an overview, International Journal of Machine Tools and Manufacture 45 (2005) 1095–1108. [9] N. Tomac, K. Tønnessen, Machinability of particulate aluminium matrix composites, Annals CIRP 41 (1992) 55–58. [10] W.S. Lau, T.M. Yue, M. Wang, Ultrasonic-aided laser drilling of aluminiumbased metal matrix composites, Annals CIRP 43 (1994) 177–180. [11] T.M. Yue, W.S. Lau, Pulsed Nd:YAG laser cutting of SiC/Al–Li metal matrix composite, Materials and Manufacturing Processes 11 (1996) 17–29. ¨ [12] F. Muller, J. Monaghan, Non-conventional machining of particle reinforced metal matrix composites, Journal of Materials Processing Technology 118 (2001) 278–285. ¨ [13] F. Muller, J. Monaghan, Non-conventional machining of particle reinforced metal matrix composite, International Journal of Machine Tools and Manufacture 40 (2000) 1351–1366. [14] T.M. Yue, Y. Dai, Wire electrical discharge machining of Al2O3 particle and short fibre reinforced Al-based composites, Materials Science and Technology 12 (1996) 831–835. [15] B. Mohan, A. Rajadurai, K.G. Satyanarayana, Electric discharge machining of Al–SiC metal matrix composites using rotary tube electrode, Journal of Material Processing Technology 153–154 (2004) 978–985. [16] B.H. Yan, H.C. Tsai, F.Y. Huang, L.C. Lee, Examination of wire electrical discharge machining of Al2O3p/6061 Al composites, International Journal of Machine Tools and Manufacture 45 (2005) 251–259. [17] A.F. Nikiforov, V.B. Uvarov, in: Special Functions of Mathematical Physics, ¨ Birkhauser Basel, Boston, 1988. [18] X.Q. Yang, K.M. Huang, The new method to calculate the complex effective permittivity of mixed aqueous electrolyte solution at microwave frequency, Chinese Journal of Acta Electronica Sinica 34 (2006) 356–360. [19] J.W. Liu, T.M. Yue, Z.N. Guo, Wire electrochemical discharge machining of Al2O3 particle reinforced aluminum alloy 6061, Materials & Manufacturing Processes 24 (4) (2009) 446–453. [20] S.F. Miller, C.C. Kao, A.J. Shih, J. Qu, Investigation of wire electrical discharge machining of thin cross-sections and compliant mechanisms, International Journal of Machine Tools and Manufacture 45 (2005) 1717–1725. [21] A.K. Khanra, L.C. Pathak, M.M. Godkhindi, Microanalysis of debris formed during electrical discharge machining (EDM), Journal of Materials Science 42 (3) (2007) 872–877.