1 Introduction

Globalization has compelled manufacturers to invest more in developing sophisticated and quality products to meet the customer's ever-changing needs. Research in manufacturing has primarily focused on high-speed machining of complex parts involving CNC machines, chemical machining, laser beam machining (LBM), electron beam machining (EBM), and electron discharge machining (EDM), to keep pace with these needs [1]. The environment of chemical machining is hazardous, whereas LBM and EBM require considerable investments compared to EDM. EDM is the most preferred machining process for cutting extremely hard materials [2]. EDM is based on the electro-thermal principle for cutting electrically conducted hard materials [3]. During EDM machining, the work-piece and the tool are immersed in a dielectric fluid and are separated by a tiny gap, known as the spark gap [4]. When a voltage difference between the anode (the tool) and the cathode (the work-piece) is sufficient to break the resistance of the dielectric, an electrical discharge takes place [5]. The spark's energy generates high temperatures, up to 30,000 °C on the work-pieces surface. This temperature causes the melting of a tiny part of the work-piece and vaporizing it [6]. There is no residual stress exist in the workpiece after an experimental run as there has not any contact between the tool and electrode [7]. Reasonable levels of finish can be achieved by employing very low pulse currents.

A relatively new advancement is the addition of powder in the dielectric fluid of EDM, which also improves MRR and lowers TWR significantly [8]. Powder mixed electrical discharge machining (PMEDM) has arisen as an efficient method in upgrading the abilities of EDM [9]. A combination of powder in EDM dielectric has become a feasible approach to overcome the significant EDM limitations and to produce a more refined surface finish over large process areas [10]. A mixture of suitable fine powder into the dielectric eases the dielectric's disintegration and, as a result, decreases its resistivity [2]. Furthermore, it facilitates more spontaneous sparks and eliminates abnormal discharge by producing a much better surface finish. Various researchers tested powder of different materials to check the performance of PMEDM [11]. Findings identify that the PMEDM may distinctly increase the surface quality and provide a mirror-like surface finish at a comparatively high machining rate [12].

Titanium alloy has high specific strength, good process performance, corrosion resistance, and the ability to retain its mechanical properties at high temperatures which is the perfect material for aerospace engineering [13]. However, Ti–6Al–4V is difficult-to-cut material with poor machinability using traditional machining approaches, while EDM is best suited for machining Ti–6Al–4V [14]. The tool must be 3–4 times harder than the work-piece, and it has to have diffusion and fatigue resistance, which is very difficult to maintain in the case of machining of Ti–6Al–4V with conventional machining process [15]. Physical contact of the tool is not needed; consequently, cutting forces are not involved; instead, material removal is done through electrical energy in EDM. Hwang et al. [16] discussed the response of utilizing multi-layered electrodes with turn reverse polarity in electric discharge measure. The layer of TiC layer on the outside of the workpiece (nickel) brings about expanded surface hardness, along with diminished microcracks [17]. Khan et al. [18] analyzed the surface integrity of the PMEDM process with two different powders (TiC and Al2O3). Janmanee and Muttamara [19] investigated the coating of Ti on tungsten-carbide surface applying PMEDM process where Ti powder is used as the dielectric. So, it raises the importance of investigating the performance of non-conventional machining such as PMEDM of Ti–6Al–4V using Graphite and TiO2 Powder.

The main objective of this study is to investigate the effects of pulse energy (current), powder concentration, and powder type on MRR, SR, and TWR in the PMEDM process of Ti–6Al–4V. In this work, fine powders of graphite and TiO2 are mixed with kerosene (dielectric) in different quantities. Response surface methodology (RSM) is conducted to generate the experimental runs. Three crucial performance measures, i.e., MRR, TWR, and SR, are considered to assess process parameters' influence. Finally, a detailed investigation on the surface topography is carried out using scanning electron microscope (SEM) to understand the morphology of the HAZ under different PMEDM conditions.

The research paper is organized as: Sect. 2 presents the materials and methods of this work. Section 3 shows the results and discusses the findings of this study, and Sect. 4 represents the conclusion and recommendation for future works.

2 Materials and methods

2.1 Work material

Grade 5 Ti–6Al–4V bars are selected as work-piece material. The chemical composition of Ti–6Al–4V is shown in Table 1 [20]. The physical, mechanical, thermal & electrical properties of the material are exhibited in Table 2 [21].

Table 1 Ti–6A1–4V (Grade 5) Material Composition [20]
Full size table
Table 2 Properties of Titanium Ti–6Al–4V (Grade 5) [21]
Full size table

2.2 The powder used in the PMEDM process

The electrically conductive powder is mixed with a dielectric of EDM, reducing the insulating strength of dielectric fluid and increasing the spark gap between the tool and work-piece in the PMEDM process [22, 23]. Micro-sized fine powders of graphite (Gr) and Titanium oxide (TiO2) are selected for the experiment. The solubility of Gr and TiO2 in kerosene is not the same, for which initiation of electric spark and discharge varies. Hence, variation in responses like MRR and TWR occurs.

2.3 Experimental setup

A submersible 12V 8-W DC centrifugal brush-less Solar pump with hermetically sealed motor close-coupled with the pump body was used inside a small box containing a limited dielectric quantity–kerosene. The pump's primary purpose is to maintain a constant flow of the dielectric in the machining zone and remove the debris before depositing it on the material's surface during machining, thereby ensuring spark efficiency. The experiment was conducted on an Electric Discharge Machine—JS EDM NCF606N. DOE was applied for the study, a set of approaches that revolve around influencing factors on the experiment's response. Input factors selected in this work were Peak Current, the powder used, and powder concentration. The value was taken for “without powder” was 0, whereas 1 and 2 for TiO2 powder and graphite powder, respectively. The peak current used during all the tests ranges from 4 to 20A. The choice of the peak current interval was made, considering the PMEDM process characteristics and work-piece materials. Along with this, the cutting conditions in Table 3 were selected based on contemporary works. Moreover, for selecting the ranges of cutting conditions, few trial runs were conducted. Table 4 represents the output factors of the experiments. Kansal et al. [23] mentioned that ‘Despite the promising results, PMEDM process is used in industry at a very slow pace.’ Among the many recommendations made by the authors is the need for optimization of the concentration of the working fluid. The current paper intended to address this issue and also to identify a suitable power for application in PMEDM. Apart from that, the empirical equations developed in the work will help in predicting the MRR, TWR, SR at a 95% confidence level. The developed approach will help in deriving a similar relationship for the PMEDM responses as functions of its other variable parameters. Figure 1 illustrates the experimental setup. Pulse-on/off time is a crucial factor affecting the output responses of the EDM, which were set at a constant value of 55 µs; however, due to the defined scope of the present work, it was decided to maintain it at a default value. The pulse on/off span was chosen based on control board settings accessible on the machine and literature review [24].

Table 3 Input factors of the experiments
Full size table
Table 4 Output factors of the experiments
Full size table
Fig. 1
figure 1

Experimental Setup of the PMEDM

Full size image

The following process diagram (Fig. 2) was applied to find the optimum machining conditions in this study.

Fig. 2
figure 2

A process diagram of this study

Full size image

3 Results and discussion

3.1 Response surface methodology

Response surface methodology (RSM) is an approach that can be conducted to generate an optimized process with a collection of statistical and mathematical processes. RSM helps express empirical relations of cutting parameters and their interaction on the considered responses [25]. The calculations for the RSM model development were carried out through the Design-Expert 12.0.5 software. Table 5 shows the experimental conditions and the measured response results. The regression calculation(s) to fit all polynomial models to the selected response (MRR, TWR, SR) was performed in the Fit and Summary test. It calculates the effects of all the model terms. One experimental “Run” was conducted once with a fixed combination of cutting conditions, which was developed by the Design-Expert 12.0.5 software.

Table 5 DOE table using Response Surface Methodology
Full size table

Surface roughness (SR) was calculated using Mitutoyo Surftest SJ-210. The surface roughness of the machined surface was measured perpendicular to that feed after the individual experimental run. Three readings of surface roughness were averaged for each of the 18 different combinations of the experimental run.

MRR was calculated by measuring each workpiece before the experimental run and after each operation on EDM, when the part with the dielectric liquid was cleaned. Equation 1 was used with a sensitive device (Denver Instrument) having an accuracy of (± 0.001 gm) for all the MRR (mm3/min) values [26, 27].

$${\text{MRR }} = \frac{{W_{i} - W_{f} }}{{\uprho _{w} t}}$$
(1)

where, \(W_{i}\) (gm) represents the initial weight, and \(W_{f}\) (gm) shows the terminal weight of the experimental workpiece. \(\uprho _{w}\) (gm/mm3) represents the density of the workpiece (4.43 g/cc), and t (min) shows the time interval of the experimental run.

TWR was calculated by a similar approach with the electrode. Equation 2 was used for all the TWR (mm3/min) values [28].

$${\text{TWR }} = \frac{{W_{ie} - W_{fe} }}{{\uprho _{e} t}}$$
(2)

where \(W_{ie}\) (gm) represents the initial weight, and \(W_{fe}\) (gm) shows the terminal weight of the electrode. \(\uprho _{e}\) (gm/mm3) represents the density of the electrode.

3.2 Analysis of variance (ANOVA) analysis

The analysis of variance (ANOVA) was conducted to check the formulated model’s ampleness developed by using the Design-Expert 15.0.5 software, which provides statistics such as p-values and F-values for comparing the models. Table 6 illustrates the ANOVA table for main and interaction effects on MRR during the experimental study.

Table 6 ANOVA for main and interaction effects on MRR during EDM of Ti–6Al–4 V
Full size table

The Model F-value of 14.29 implies that the model is significant. There is only a 0.03% chance that a “Model F-value” this large could occur due to noise. P-values of less than 0.0500 indicate that those model terms are significant, which means A, B, AB, A2, and B2 are significant terms. ANOVA analysis was conducted for the other response (TWR), and the findings were the Model F-value for TWR is 3.64, which implies the model is significant and A, B, AB, A2, and B2 are significant model terms in formulating the TWR equation as their P-values are less than 0.0500. Table 7 represents the comprehensive statistics of the experimental values and design summary.

Table 7 Design summary of the responses
Full size table

Six equations were generated for two categories of factors. These equations in terms of the fundamental factors for Gr and TiO2 suspended dielectric is shown in Eqs. (3)–(8). These equations can be used to make predictions about the responses for given levels of each factor. According to the fit and summary tests, quadratic models were suggested for Eq. (3)–(6), while linear models are suggested for Eqs. (7), (8), where A represents the peak current (I0), and B represents the powder concentration. The equations are shown below:

$$\left( {{\text{MRR}}} \right)_{{{\text{Gr}}}} = \, - 4.74618 \, + \, 0.479333 \times {\text{A }} + \, 2.23742 \times {\text{B }} + \, 0.086594 \times {\text{A }}*{\text{ B }} - \, 0.028788 \times {\text{A}}^{2} \, - \, 0.223729 \times {\text{B}}^{2}$$
(3)
$$\left( {{\text{MRR}}} \right)_{{{\text{TiO2}}}} = \, - 23.88535 \, + \, 1.51052 \times {\text{A }} + \, 5.51583 \times {\text{B }} + \, 0.086594 \times {\text{A}} \times {\text{B }} - \, 0.028788 \times {\text{A}}^{2} \, - \, 0.223729 \times {\text{B}}^{2}$$
(4)

In these equations, no transformation was used for peak current and powder concentration; thus, the factor with the highest value of coefficient possesses the most dominating effect over the response of MRR.

$$\left( {{\text{TWR}}} \right)_{{{\text{Gr}}}} = \, 0.546917 \, - \, 0.014943 \times {\text{A }} - \, 0.133229 \times {\text{B }} + \, 0.000266 \times {\text{A}} \times {\text{B }} + \, 0.000953 \times {\text{A}}^{2} \, + \, 0.008937 \times {\text{B}}^{2}$$
(5)
$$\left( {{\text{TWR}}} \right)_{{{\text{TiO2}}}} = \, 0.507333 \, - \, 0.012401 \times {\text{A }} - \, 0.123979 \times {\text{B }} + \, 0.000266 \times {\text{A}} \times {\text{B }} + \, 0.000953 \times {\text{A}}^{2} \, + \, 0.008937 \times {\text{B}}^{2}$$
(6)
$$\left( {{\text{SR}}} \right)_{{{\text{Gr}}}} = \, 2.41853 \, + \, 0.112719 \times {\text{A }} + \, 0.147375 \times {\text{B}}$$
(7)
$$\left( {{\text{SR}}} \right)_{{{\text{TiO2}}}} = \, 0.391417 \, + \, 0.112719 \times {\text{A }} + \, 0.147375 \times {\text{B}}$$
(8)

3.3 Effect of process parameters on MRR, TWR, and SR

A response surface plot is a tool providing an estimation of the optimum response region, which is similar to the 3D response graph [29]. The 3D response graph of MRR (Fig. 3a) has been plotted with the data collected for the Gr powder, and (Fig. 3b) has been plotted with the data collected for TiO2 powder. It is observed that the MRR in the case of TiO2 powder is much higher (approximately 4 times) compared to that of Gr powder. It may be noted here that the 3D plots are an average approximation of the MRR values for the combination of the parameters for Gr and TiO2 powders. It may be noted here that minor local effects of the parameters have been averaged in these plots during the regression process.

Fig. 3
figure 3

Graph plotted for MRR of a Gr & b TiO2 powder

Full size image

Contour graphs, i.e., the projections of 3D graphs on a 2D plot, show the upper and lower points and the effect of two factors on a specific response. The graphs of Fig. 4 show how variable A (peak current) and variable B (powder concentration) affect the TWR in the case of Gr powder (a) and TiO2 powder (b). The blue region in the 3D graph indicates the lowest TWR, the green, and yellow region indicates moderate values, and the red region indicates the highest TWR. It is observed from the plots of Fig. 4 that the TWR is slightly lower (1.5 times) when Gr powders are used compared to that when TiO2 is used. However, comparing the much higher metal removal rate (MRR) (4.5 times) in TiO2 powder, its slightly more significant influence on TWR may be ignored.

Fig. 4
figure 4

Graph plotted for TWR for a Gr, b TiO2 powder

Full size image

3D response graph of average surface roughness (SR) for the Gr powder graph is shown in Fig. 5a and that for TiO2 powder in Fig. 5b. The variables of the contour graphs are peak current (A) and powder concentration (B). The two plots show that the surface roughness is lower for TiO2 powder compared to Gr powder. In both, cases peak current has a more significant influence on SR compared to powder concentration. It may be further noted that though the MRR in the case of TiO2 powders is approximately 3.5 times higher than that of Gr powder, that did not lead to higher surface roughness generation; on the contrary, lower surface roughness (approximately 1.5 times lower) is achieved in the case of TiO2 powder. Again, it may be noted that minor local effects of the parameters have been averaged out in the 3D plots during the regression process.

Fig. 5
figure 5

Graph plotted for SR of a Gr & b TiO2 powder

Full size image

As there is a three-hold limit of powder concentration, depending on each powder's properties, some of the values of roughness are more in the additive machining process than the normal process. In such cases, the powder particles were heavily in between the spark gap, which decreases the smoothness of the surface created. It is visible that when the peak current is increased roughness value also increases. This is because when the peak current is raised, discharge energy per pulse is increased. As a result, a rough surface is produced at higher discharge energy.

3.4 Investigation of the effects of PMEDM parameters on machined surface topography using SEM

It is verified that using SEM, every detail of the chip can be viewed using very high magnification [30]. To investigate the effects of powder type, powder concentration, and peak current on the machined surface, views of the surface were observed using scanning electron microscope (SEM). The first set of experiments' main objectives were to understand the effect of powder on EDM performance. SEM images of Fig. 6a, b compare the surface topography under plain EDM using kerosene as the dielectric with the topography produced under PMEDM with 7 gm/L concentration of graphite powder a peak current of 4A. It is observed that there is a visible difference in the topography of the surfaces produced under the two conditions. A matted surface is shown in Fig. 6b indicates that a smoother surface generated in PMEDM compared to the deep-rooted grooves of individual sparks (Fig. 6a). Material Removal Rate under these two conditions are almost the same as powder concentration does not appreciably influence the MRR, as shown in Fig. 3a. However, powder concentration tends to decrease surface roughness. In PMEDM, the powder particles enter into the spark gap and create a bridging effect. As a result, the plasma channel becomes enlarged and widened. The material removed takes place smoothly over a larger surface area. Thus, powder particles' presence results in a lower depth of the crater, taking place over a wider area, and consequently, the polishing time and the surface roughness are reduced [31]. Since the density of TiO2 powder is 4.23 g/cm3 and that of graphite powder is 1.8 g/cm3, the volume percent of graphite is 2.35 times higher compared to the TiO2 powder for the same weight percentage. Therefore, a larger number of graphite powder will be in the spark gap to influence the sparking process. However, graphite is a good conductor of electricity. When the voltage difference is set up in the spark gap, positively charged graphite powder particles will have the tendency to be attracted towards the cathode (electrode); hence the formation of a ‘chain-like structure’ to help in overcoming the dielectric strength of the medium (kerosene) may not be easily facilitated. However, in the case of TiO2 powder, which is a semi-conductive material, the induced bonding is more readily developed to form the ‘chain-like structure’ to overcome the dielectric strength of the medium. This may be the principal reason for higher MRR in the case of TiO2 powder, though the volume percent of the powder is 2.35 times lower compared to that of graphite powder at the constant weight percentage of the two powders. However, the phenomenon needs a more detailed study to arrive at a definite conclusion on the mechanism of the process.

Fig. 6
figure 6

Comparison of SR for a kerosene and b powder suspended kerosene used as the dielectric for different peak current 7 gm/L concentration of graphite at a Ip = 4A

Full size image

3.5 Investigation of the effect of peak current on HAZ in PMEDM at the powder concentration of 7 gm/L

To study the effect of peak current on surface topography and the expected HAZ, SEM images of the machined surface were taken (Fig. 7) for surfaces generated under different values of peak current (4A, 12A, and 20A) at a constant concentration of Graphite powder (7 gm/L). It is revealed that large areas of relatively smooth surfaces are generated at low current values of 4A (Fig. 7a) and 12A (Fig. 7b). On the other hand, the rougher surface is generated when the current values of 20A are applied for the same concentration of the power, 7 gm/L (Fig. 7c). At lower peak current values, shallower craters are created at low discharge energy compared to higher discharge energy. Comparing these images with those of Fig. 6a, b, it may be concluded here that the application of high peak current even in PMEDM leads to larger crater sizes, which results in higher surface roughness and greater depth of the HAZ in PMEDM. Hence the peak current value must be limited to a finite value to get the desired effect of PMEDM in better surface finish and lower depth of the HAZ.

Fig. 7
figure 7

SEM images for 7 gm/L concentration of graphite at a Ip = 4A, b I = 12A and c Ip = 20A

Full size image

3.6 Investigation of the effect of powder concentration on HAZ in PMEDM at the highest peak current of 20A

“Detailed analysis of structural features of the machined surface is done by using SEM to realize the mode of the recast layer thickness, HAZ, and micro-cracks, which alternatively influences the structure of machined work-piece and hence tool life” [32, 33]. To identify the influence of the concentration of the Graphite powder on the topography of the machined surface and the HAZ at the highest peak current of 20A, SEM images of the machined surfaces produced under different powder concentrations are compared in Fig. 8a–c. It is observed from these three images that a higher powder concentration is required to get the desired effect of PMEDM when a high peak current is used. In the case of 20A current, no effect of powder is observed on the surface topography when the concentrations are 5 gm/L (Fig. 8a) and 7 gm/L (Fig. 8b). However, the beneficial effect of PMEDM is observed on the surface topography when the powder concentration is increased to 9 gm/L (Fig. 8c).

Fig. 8
figure 8

SEM images for Ip = 20A for different concentrations a 5 gm/L, b 7 gm/L and c 9 gm/L of graphite

Full size image

As a result of this, a sharp reduction of surface roughness (from 9.9 to 5.6 μm) is observed when the Graphite powder concentration increases from 7 to 9 gm/L which can be seen in the bar chart of Fig. 9 and Table 5. Similar effects are also observed in the case of TiO2 powder, as shown in the same figure and Table. However, this reduction in surface roughness is not caused at the cost of lower MRR. On the contrary, MRR is higher when power concentration is increased from 7 to 9 gm/L in the case of Gr powder, though there is a slight fall in the case of TiO2 powder seen in Fig. 10. A suitable amount of powder leads to the uniform dispersion of discharge energy in every direction, which leads to a decreased surface roughness as the crater size is minimized [34]. Therefore, these findings may be concluded that the successful application of PMEDM, with appropriate selection of the peak current, powder type, and powder concentration leads to higher MRR and lower surface roughness.

Fig. 9
figure 9

Bar charts of Average Surface Roughness for different concentrations of Gr and TiO2 powders at the peak current value of 20A

Full size image
Fig. 10
figure 10

Bar charts of Metal Removal Rate for different concentrations of Gr and TiO2 powders at the peak current value of 20A

Full size image

Owing to the limited scope of the work, an in-depth study on the HAZ was not included in the current study. The SEM images are presented to have a qualitative assessment of the topography of the HAZ and to explain the surface roughness generated under different machining conditions and powder concentration. SEM images give an indirect idea about the depth of penetration of the effect of the sparking process at the investigated conditions.

4 Conclusion

In this study, the effects of concentration (5 –9 g/l) of two powders (Graphite and TiO2) and peak current (5A–20A) were investigated on the performance of the PMEDM applied on Ti–6A1–4V (Grade-5) in terms of response parameters—MRR, TWR, and SR. Kerosene was used as the dielectric. The main conclusions drawn on the paper are summarized below:

TiO2 powder showed much superior performance over Graphite powder in terms of all the PMEDM responses; MRR was found to be roughly 3.5 times higher, tool wear ratio (TWR) at the maximum MRR 1.8 times lower, and at the minimum MRR roughly 3 times lower in the case of TiO2 powder. On the other hand, the maximum surface roughness attained with TiO2 powder was 3.265 μm against 9.936 μm for graphite powder at the highest MRR, and the same attained at the lowest MRR was 2.228 μm for TiO2 against 2.411 μm for graphite powder. Thus, TiO2 powder of appropriate concentration and peak current value may be recommended for rough as well as finish PMEDM process of Ti–6Al–4V (Grade-5) alloy as it can generate higher MRR, lower TWR, and relatively lower SR.

Empirical models of MRR, TWR, and SR developed in the work may be used to predict these responses at a 95% confidence level in terms of the variable parameters of powder concentration and peak current.

One of the limitations of the present work was the limited number of (two) powders, their narrow range of concentration (5–9 g/L), and only one parameter of current (Peak current) considered in the work. Future works may be performed to determine the optimum concentration of powders in a wider range of powder concentrations and optimize other current parameters (Pulse on/off duration) apart from the peak current. Future work also needs to focus on the in-depth study of the HAZ in terms of its depth, composition, and mechanical properties of the zone, as well as the properties of the layer below the HAZ, as that layer plays a vital role in the performance of the part. Secondly, while considering the concentration of the powder, it would perhaps be more appropriate to consider the densities of the powders to have equivalent volume concentration of the powders, not merely their weight percentages, as it is the number of the particles which is important in terms of PMEDM performance. For instance, Graphite Powder with a density of 1.8 g/cm3 occupies 2.35 times the volume of the same weight of TiO2 powder with a density of 4.23 g/cm3. Hence it is expected that the concentration of a lighter powder will have a larger number of ions to assist in the additional arching process during PMEDM.