The Effect of Discharge Current and Pulse-On Time on Biocompatible Zr-based BMG Sinking-EDM

The machinability information of Zr-based bulk metallic glasses (BMGs) are recently limited but essential to provide technological recommendation for the fabrication of the medical devices due to the material’s metastable nature. This study aims to investigate the material removal rate (MRR) and surface roughness under different current and pulse-on time of newly developed Ni- and Cu-free Zr-based BMG using sinking-electrical discharge machining (EDM). By using weightloss calculation, surface roughness test and scanning electron microscopy (SEM) observation on the workpiece after machining, both MRR and surface roughness were obtained to be increased up to 0.594 mm3/min and 5.50 μm, respectively, when the higher current was applied. On the other hand, the longer pulse-on time shifted the Ra into the higher value but lower the MRR value to only 0.183 mm3/min at 150 μs. Contrary, the surface hardness value was enhanced by both higher current and pulse-on time applied during machining indicating different level of structural change after high-temperature spark exposure on the BMG surface. These phenomena are strongly related to the surface evaporation which characterize the formation of crater and recast layer in various thicknesses and morphologies as well as the crystallization under the different discharge energy and exposure time.


Introduction
During the development, Zr-based bulk metallic glasses (BMGs) becomes most favourable alloy to be largely applied on several fields because of their excellences, such as high glass-forming ability (GFA), high fracture strength, low surface roughness (due to the absence of grain boundary), low Young's modulus, high recovery rate, good corrosion and wear resistance [1,2]. By these unique properties, Zr-based BMG system is considered as potential material for biomedical application, even compared with Tibased BMG. Although Ni and Be element, which are considered as harmful elements for human body, were successfully excluded in order to apply the alloy on medical application as reported by several researches, the other toxic element in the form of Cu was still involved into the alloy composition to sustain the GFA of BMG [3][4][5]. Moreover, the Cu ions released at the human body after implantation showed greater cytotoxicity potential than those of Be and Ni [6]. Because of this issue, the highly biocompatible Zr-based BMG with the alloy configuration of Zr 54 Al 17 Co 29 was designed and fabricated into 6 mmdiameter rod [7]. Although the alloy possessed lower GFA than toxic-containing Zr-based BMG, the mechanical and chemical characteristics were still superior, with the average hardness of 540 HV, fracture strength of 2220 MPa, plastic strain of 2 %, Young's modulus of 101 GPa, and better protective film-forming ability than SS 316L indicating this material can be formed into high-quality medical devices such as surgical blade and orthopedic drill bits previously fabricated by Jang Group [8,9].
The problem for subsequent processing the advance materials with such high mechanical properties is that the material becomes difficult to be machined. Several studies implied the limitation to cut BMG using conventional machining at dry condition, namely high cutting forces, high cutting temperature, and excessive tool wear [10,11]. The high cutting forces and temperature can induce crystallization on amorphous body occurred on both machined surface and chip, subsequently deteriorate the amorphous nature of materials and lower its application [12,13]. De-spite the fact that the problem can be minimized by applying lower cutting speed and cutting fluid during machining, but for processing Zr 54 Al 17 Co 29 BMG having low GFA (compared with BMG used on the preceding studies), the crystallization tends to occur easily through conventional machining.
Electrical-based nonconventional machining such as electrical discharge machining (EDM) is possibly applied to cut BMG by substituting shearing material removal mechanism on conventional machining because BMG performed better electrical conductivity compared with that on oxide glass material. Different from laser- [14] and plasma-based [15] processing having no cooling system, dielectric fluid applied on EDM acts not only as insulator from high density of spark exposure at nearest gap between tool and workpiece during material removal mechanism but also as cooling system of the tool and machined surface; therefore, it increases the possibility of cutting heat localization and thus limit the crystallization of the BMG at the lower level than those on conventional machining [16]. As reported before, spark occurred during machining characterized the discharge energy transmitted on the workpiece surface and became the dominant driving parameter for material removal mechanism on EDM cutting, beside of the dielectric flushing system [17]. Nowadays, micro-EDM was preferable to be applied in order for avoiding crystallization of BMG cutting because of its low discharge energy, based on several reports [18,19]; however, the method is considered inappropriate for cutting macro-dimensional shape on medical devices; therefore another technique is needed to cut BMG at a larger scale.
Among the other determining factors on EDM (dielectric and tool), machining parameters become a most flexible factor in modifying the EDM performance on BMG cutting, where polarity, current, and pulse-on time in the sequence were reported to be top-three of the most influential factors on material removal as described by Mathumaringam [16]. Although considered as the dominant factor, polarity adjustment is not applicable to be adjusted on all EDM machine types due to limited feature on these machines. However, in this case, these parameters need to be selected and combined properly by also considering the crystallization occurred on BMG workpiece due to high discharge energy, besides the material removal rate (MRR) and surface quality to maintain the material unique properties which rarely done on biocompatible BMG material by other researchers. In this study, the role of EDM process parameters, in the form of discharge current and pulse-on time, on the MRR, surface roughness and hardness evolution of biocompatible Zr 54 Al 17 Co 29 BMG sinking-EDM were investigated using weight loss calculation, surface roughness test, and microvickers hardness test. To confirm the surface morphology after machining, the microstructural observation was conducted by scanning electron microscopy (SEM) at various magnifications. The experimental results were then analyzed and discussed to provide the technological recommendation for further application.

Materials and Methods
In this study the BMG sample was provided by High-Performance Alloy Laboratory, NCU, Taiwan with the alloy configuration of Zr 54 Al 17 Co 29 (in at.%) owing the density of 6.52 × 10 −3 g/mm 3 in the form of a rod having diameter and length of 4 and 80 mm, respectively. A 100 mmlong cylindrical CuCr rod having a diameter of 10 mm was used as tool material. Both the BMG rod and tool crosssectional surfaces were then flattened and polished to ensure the parallel gap between their interface using a centrifugal sandpaper machine. Before machining, the CuCr tool and BMG workpiece were also polished using similar grit of sandpaper to ensure the similar surface roughness level on the both workpiece and tool cross section for each experiment stage. The machining was conducted using C-TEK ZNC 320 EDM Machine by varying the discharge current and pulse-on time. Due to the larger area of the tool compared with the workpiece, the machining was subjected completely on all area of BMG top surface. The machining condition was completely listed in Table 1. During the process, the BMG -tool gap was insulated and cooled by oil-based dielectric fluid (Chevron HONILO 409) applying jet flushing mode. Digital scale (OPTIMA OPD-E204) was used to measure the weight difference between BMG workpiece before and after machining for material removal rate calculation.
The surface roughness of BMG workpiece formed after machining was evaluated using surface roughness tester (MITUTOYO Surface SJ 301 Series) and the test was taken on 3 different locations of each workpiece to make sure the result validity. To investigate the structural change on the newly developed layer of BMG workpiece machined surface, the microvickers hardness (ESEWAY) test was conducted using the indentation load and period of 500 g and 10 s, respectively. The indentation was also subjected to 3 different locations on the sample surface. The scanning electron microscopy (SEM) observation (PHENOM G2 Pro) was also conducted to evidence the morphological condition and thickness of the recast layer developed on the BMG surface after machining using different machining parameter combination. The SEM image was taken at the magnification of 200, 500, and 1000 times.

Material Removal Rate
The MRR was defined as the material loss in the unit of volume for certain machining time on BMG sample using selected discharge current and pulse-on time combination. The equation is described as follow (1).
Where w BW is workpiece weight before machining (g), w AW is workpiece weight after machining (g), t is machining time (min), and ρw is workpiece density (g/mm 3 ) Figure 1(a) shows the relation between discharge current and MRR using various pulse-on time values, where the MRR is observed to be higher, reaching the maximum value of 0.594 mm 3 /min, when the larger discharge current was applied. The results indicate that the high current is more productive for BMG machining using EDM. The current represents discharge energy determining heat generation during spark material removal mechanism, where the melting and evaporation on the workpiece surface occurred at higher level induced by high discharge temperature. This phenomenon was also reported by previous studies [18,19]. Additionally, by increasing discharge energy, the debris deposited in the tool-workpiece gap was also increased. The debris contains the dielectric decomposition product and eroded conductive BMG particles. As machining proceeds, the amount of particle increased rapidly at the gap. The particles can exert significant influence on the characteristics of dielectric breakdown and influence the material removal mechanism [20]. Moreover, the short pulse-on time can induce the more opportunity for debris in the machining gap to be flushed away from the machining region by dielectric fluid during a similar period of pulse-off time thus resulting in higher MRR. This MRR is significantly higher compared with micro-EDM process subjected to machine micro-scale shape, as revealed by other researchers [16,17].
On the other hand, from the figure, it can also be observed that the pulse-on time brings an inverse effect on MRR, where the higher MRR curve is achieved by shorter pulse-on time. For example, at the current of 4 mm, the MRR are 0.320; 0.203; and 0.183 mm 3 /min at 50, 100, and 150 µs, respectively. This parameter showed the different effect compared with discharge current on the removal mechanism. The results are similar to the study reported by Mohan et al. [20]. Contrary, Wang and Yan found the different fact that MRR was increased along with the increase pulse-on time at lower range [21]. However, it turns to be decreased at a longer pulse-on time application. It indicates that the inverse effect of pulse-on time towards MRR can be achieved beyond their optimum condition. The short pulse-on time might induce low vaporization on the BMG workpiece surface, while the excessive pulse duration may cause an expansion of the plasma channel, thereby decreasing the energy density transmitted the machining process.

Surface Roughness
In this study, the BMG surface quality was strongly characterized by both discharge current and pulse-on time applied during machining. Figure 1(b) shows the surface roughness as a representative of surface quality is increased significantly at the highest value of 5.50 µm, if the higher current was applied, irrespective to the other parameter, indicating the deterioration on surface quality occurred when using high discharge energy during machining. This phenomenon is similar with those occured on previous studies where the increased discharge energy, which is proportional with the current, inducing the higher machining efficiency and the development of the larger size of the crater on the machined surface [18][19][20]. This larger crater size, correspond to the high amount of material melting during discharge mechanism (pulse-on time). When the crater size is similar, the randomly overlapped crater distribution can be the determining factor to characterize the surface roughness [23].
The different trend compared with MRR curve was obtained on surface roughness analysis varied by pulse-on  Figure 1(b), it can be noticed that the higher surface roughness level is obtained by longer pulseon time, where lowest surface roughness is obtained at the current of 4 A and pulse-on time of 50 µs with 3.24 µm, indicating the short pulse-on time was favourable for resulting high surface quality. In BMG EDM, the short pulse-on time might induce low vaporization on the BMG workpiece surface, due to its unique amorphous thermal behaviour. Eventually, the short pulse-on time produces a shallow crater on the surface of the workpiece resulting in better surface quality. This phenomenon was significantly different compared with the machining of crystalline metal such as SiC and Al alloy which showed an inverse relation between pulse-on time and surface roughness [20,21]. During the pulse-on time, sparks arise at the gap between the tool and BMG workpiece, causing local heating, melting, and vaporization on both the BMG workpiece and tool. At the short pulse-on time, unlike the phenomenon occurred on the highly-conductive CuCr tool, these thermal cycle was occurred on BMG only at a low level, thus resulting in the small crater size.
In order to further analyze the surface roughness of BMG machined surface, the SEM observation was conducted on the workpiece having lowest (4 A, 50 µs) and highest (12 A, 150 µs) roughness of 3.24 and 5.5 µm, respectively, using the magnification of 200 and 500 times. Figure 3 indicates that the carbonaceous layer, produced from the breakdown of dielectric fluid [23,24], was deposited on the machined surface. It is also clear that by using low on both current and pulse-on time, the crater size is clearly smaller (Figure 2(a)) than those on highest respective parameters ( (Figure 2(b)) indicating the low current had a positive impact on reducing the surface roughness. Based on Figure 3, The average crater diameter on the lowest and highest surface roughness of BMG machined surface is measured at 165 and 400.9 µm, respectively. On the other hand, the micro-porous was also found on the machined surface using 12 A and 150 µs, as shown in Figure 3 (b). The micro-porous was possibly formed by irregular overlapping craters distribution induced by excessive melting due to high discharge energy, as also stated by several other studies [18,19,23]. These crater size and microporous appearance were responsible on the increased surface roughness on machined BMG; therefore, when the surface quality becomes the essential criteria, such as a finishing step, the discharge energy should be maintained as low as possible.

Surface Hardness
Due to the difficulties to reveal the structural change by performing the XRD analysis on relatively rough Zr 54 Al 17 Co 29 BMG machined surface covered with a carbonaceous layer, the hardness test was conducted on the sample surface after machining to reveal the change of BMG surface mechanical properties. This effort was performed considering the importance of these structural issues on biocompatibility and chemical properties of BMG before application. Figure 4 describes the surface hardness curves varied by both current and pulse-on time where the increase of discharge current conducts a higher level of surface hardness. The hardness value ranges from 247.6 to 418.5 HV, and these values are significantly lower than the hardness of the as-cast amorphous sample of 541 HV. It indicates that the transformation of initial glassy structure in the form of crystallization may occurred at higher increment of high spark temperature resulted by high discharge energy. This issue was referred on our previous work conducted at BMG composition, where the crystallinity was increased by longer annealing time resulting in high hardness level on Zr 54 Al 17 Co 29 BMG due to the presence of hard intermetallic compounds [25]. The crystallization was also possibly contributed by a good chemical affinity of Zr and possibly appeared C atoms [19]. However, the crystallization phenomenon on annealing is different with the melting and re-solidifiction occurred on the BMG surface during discharge process, where at the annealing, the nanocrystals is formed and acted as strengthening phase to enhance the BMG mechanical properties, including the hardness. In this study, the crystallization is occured at higher level, or possibly fully-crystallization, at the BMG surface; therefore, it shows different effect on the hardness change as lower mechanical properties of crystalline alloy than its BMG counterpart. As comparison, for Be-, Ni-, and Cu-containing Zr-based BMGs in the form of Zr 41.2 Ti 13.8 Cu 12.5 Ni 10 Be 22.5 [19,23] and Zr 55 Al 10 Ni 5 Cu 30 [26], where these toxic elements act as GFA stabilizer, the crystallization is still occurred on the sample surface although the BMG samples are only cut by micro-EDM using low discharge energy. By considering this fact, the crystallization in this study should be occurred at higher level due to the lower GFA of Zr 54 Al 17 Co 29 (because of the absence of Be, Ni, and Cu elements) which is more sensitive and having lower ability to maintain the amorphous characteristics under the high discharge energy exposure exerted by conventional EDM process. On the other hand, pulse-on time also played an important role to characterize the hardness where the longer pulse-on time is observed to enhance the surface hardness of the machined surface, as depicted in Figure 4. This phenomenon is correlative with those obtained on the surface roughness analysis, where the change of pulseof time related to the behaviour of developed recast layer after machining. The long pulse-on time induced hightemperature spark occurring at the tool-workpiece gap during the discharging process, causing local heating, melting, and vaporization at a high level resulting in higher crystallization on BMG machined surface. The thermalinduced-structural change is more susceptible to be appeared on BMG, in comparison with metallic material due to the low thermal stability (metastable nature) and more stages of phase transformation of glassy alloys [27].  Figure 5 shows the side view of the machined surface, demonstrating the thickness of formed recast layer on the top BMG surface. From Figure 5(a) it indicates that the thin recast layer of 60.6 µm was generated at low discharge energy, while the layer was increased significantly up to 151,1 µm using the highest discharge energy as depicted on Figure 5(b). Recast layer was formed from partially molten metal which was not involved in the evaporation process due to the cooling done by the dielectric. When the discharge energy is higher, the larger workpiece region is melted and evaporated, thus resulting in thicker recast layer. In the case of BMG sinking EDM, the thickness of recast layer formed becomes the indicator of the crystallization occurred on the amorphous body; therefore it can reduce the unique properties of BMG and thus considered as a limitation of BMG machining. The condition was then worsened by applying long pulse-on time because it does not only generate high discharge temperature, but also induces insufficient ejection force causing the part of ejection BMG to remain as a recast layer, and the process becomes evidence at stationary tool [20]; therefore, low discharge energy is needed to be selected to reduce the crystallization. However, the recast layer can be removed by a subsequent polishing process to maintain the glassy surface, as proposed by Huang [19].

Conclusion
Based on the experimental investigation, the important informations were obtained on the machinability of newlydeveoped biocompatible Zr 54 Al 17 Co 29 BMG, which is hard-to-cut and susceptible to crystallization, using the most determining process parameters of sinking-EDM and those are summarized as follows. The MRR reached a higher increment at higher discharge current and lower pulse-on time. The pulse-on time, showed the correlative relation with the surface roughness instead of inverse relation as largely found on the metallic material machining due to the metastable nature of common BMG. The surface roughness was characterized by the crater size and the presence of micro-porous, as confirmed by SEM. The different degree of hardness were found on the BMG machined surface, indicating the crystallization. The recast layer is considered as crystallization site of the glassy alloy. These results can be a technological recommendation for further EDM application on the metastable BMG materials for bioimplant and medical devices. The structural and corrosion analyses on the biocompatible Zr-based BMG machined surface need to be conducted for advanced study.