Electrical discharge process for rapid oxycarburisation of titanium in air

Introduction

Ti and Ti based alloys are very attractive materials for biomedical applications due to their excellent biocompatibility as well as several other outstanding properties, such as high strength to weight ratio, good fatigue properties, low modulus of elasticity and excellent corrosion resistance.^1,2 By the modification of the tribological performance, the areas of their applications can be extended into the other industries such as corrosions, surface colouring, and high temperature materials.^3–6 Therefore, various surface modifications have been studied and applied for improving the wear resistance of Ti and Ti based alloys.

Among them, thermal oxidation provides useful mechanical support for the external oxide layer by producing an oxygen diffusion zone beneath it, resulting in significant improvements in wear resistance.^7–9 Carburisation of Ti is another good candidate due to its ability to form a strong bond with the Ti substrate and its possession of superior hardness and wear resistance with the potential for excellent biocompatibility. ¹⁰ Nitriding processes have been also frequently used to form titanium nitrides on the Ti surface. Titanium nitrides have excellent corrosion and wear resistance properties as well as bioactivity, and are thus widely employed as coating materials to improve tribological performance.^11,12 Other more recent examples include the so called multifunctional coating systems, where the oxynitrides, MeO_xN_y, and oxycarbides, MeC_xO_y (where Me stands for transition metal) have been attracting particular interest.^13–20

For many years, surface modification processes based on the thermo-diffusion mechanism have been widely used to improve the properties of Ti and Ti based alloys. However, for example, conventional thermal processes have several disadvantages such as long processing times, poor surface finish, poor control of the coating layer formation, etc.^21–23 The plasma diffusion technique has been also developed and regarded as a promising route, and this generally includes plasma nitriding, plasma carburizing and plasma carbonitriding. This process has also several problems. For example for the case of plasma nitriding, the formation of an arc can damage the nitriding surface. In addition, a comparatively high operating pressure makes the substrate surface to be sensitive to impurity contamination which retards the diffusion of nitrogen, resulted in the formation of a very thin nitrided layer (0·1–0·2 μm) even after a long treatment time.^24–26 This is neither economical nor favourable for biomedical applications. Recently, the laser diffusion process has been viewed as attractive due to its short treatment time and accuracy in controlling the treatment location. However, it involves surface melting of the substrate, resulting in dimensional changes and also in a roughened surface. ²⁷ Each of a wide range of techniques for the surface modification of Ti and Ti-based alloys has its own merits and limitations.

In this work, the electrical discharge of Ti, which uses a high voltage, high current pulse, has been carried out in air to investigate the surface modification and its efficiency compared to other conventional oxidation or carburization processes. Ti rods were exposed to an electrical discharge process with 2·0–3·51 kJ of input energy from a 450 μF capacitor bank, and were then investigated in terms of their microstructures, hardnesses, and surface chemistries. Results show that this technique is a possible potential method for the surface modification of Ti; advantages include a short processing time and control of the multifunctional coating layer without dimensional changes.

Experimental

Grade 2 quality commercially pure Ti rods with a diameter of 3 mm and a length of 5 cm were ultrasonically cleaned and dried in a vacuum. The Ti rods were connected on the top and bottom tungsten electrodes without applying any external pressure. A capacitor bank of 450 μF was charged with four different electrical input energies (2·0, 2·5, 3·0 and 3·5 kJ) and the charged capacitor bank instantaneously discharged in air through the Ti rod by an on/off high vacuum switch which closes the discharge circuit. The voltage and current that the Ti rod experiences when the circuit is closed were simultaneously picked up by a high voltage probe and a high current probe, respectively. Outputs from these probes were fed into a high speed oscilloscope that stored them as a function of discharge time. The overall process is referred to as electrical discharge process. A schematic of the process is shown in Fig. 1. The electrical discharge was carried out in an atmosphere of air in order to investigate the possible surface modification of Ti by oxygen and carbon.

OPEN IN VIEWER

Figure 1

Schematic of electrical discharging apparatus

As received and electrically discharged Ti rods were sliced into 3 mm segments and their cross-sections were then examined with optical microscopy and scanning electron microscopy (SEM), to investigate their microstructures. Micro-Vickers hardness measurements were carried out to determine if any change had occurred during the electrical discharging. Hardness was measured on the cross section of Ti rod using a Mitutoyo HM-122 micro Vickers tester with a load of 50 g and a dwell time of 20 s. At least twenty indents were made along seven positions located at the centre to the edge of cross-section of Ti sample.

The electrically discharged Ti rod without any surface treatment was mounted on the spectrometer probe tip by means of conductive double sided adhesive tape and examined by XPS (X-ray photoelectron spectroscopy). Under the current conditions employed, the full width at half maximum (FWHM) of the Ag 3d_5/2 peak was 1·1 eV, and the binding energy difference between Ag 3d_5/2 and Ag 3d_3/2 was 6·0 eV. When the Ag 3d_5/2 peak was used as the reference peak, the binding energy of the C1s peak of adventitious carbon on the standard silver surface was 285 eV. All binding energies were referenced to the C 1s peak to correct for sample charging.

Results and discussion

Figure 2 shows the cross-sectional optical micrograph of an as received Ti rod. The average grain size was estimated to be approximately 17·5 μm with equiaxed α grains. A typical discharge curve (Fig. 3a) shows voltage and current in terms of discharge time. 450 μF of capacitance and 3·94 kV of input voltage were employed to yield 3·5 kJ. The input energy (E) is predetermined by controlling input voltage (V) according to the equation of 1 where C is the capacitance of a capacitor. Figure 3a shows that the peak current was 33·4 kA and the peak voltage was 3·1 kV. From the results shown in Fig. 3a, the power (watt) curve is plotted in Fig. 3b against the discharge time. The power was obtained from the following equation 2

OPEN IN VIEWER

Figure 2

Optical cross-sectional micrograph of as received Ti rod

OPEN IN VIEWER

Figure 3

a typical discharge curve measured current and voltage on oscilloscope and b power curve versus discharge time (electrical discharge condition: 450 μF, 3·5 kJ)

The power increases with an increase in input energy at same capacitance. The discharge times for the duration of the first cycle at four different input energies are identical to be approximately 159 μs.^28,29 The amount of heat generated (ΔH) during an electrical discharge can be obtained by using equation (3) 3

Under current experimental conditions, the electrical discharge characteristics in terms of peak voltage, peak current ΔH and discharge time are tabulated in Table 1. It is known that ΔH increases with an increase in input energy at constant capacitance.

OPEN IN VIEWER

Table 1

Peak voltage, peak current, discharge time ΔH and ΔT during electrical discharge processes with various input energies

Input energy/kJ	Peak voltage/kV	Peak current/kA	Discharge time/μsec	Heat generated (ΔH)/J	ΔT/°C
2·0	1·8	19·6	159	1426	569
2·5	2·2	25·2	159	2290	1092
3·0	3·3	29·6	159	3346	1730
3·5	3·1	33·4	159	4450	2398

Figure 4 shows cross-sectional optical micrographs of Ti rods which were electrically discharged with four different input energies. From Fig. 4a, equiaxed α grains can be seen and no phase transformation has occurred. On the other hand, electrical discharge of Ti with an input energy greater than 2·5 kJ caused a β→α allotropic transformation, as shown in Fig. 4b–d, exhibiting a typical serrated morphology. The microstructure is very similar to that of the Ti rod which was annealed at 1000°C for 2 h and air cooled.^30,31 This result implies that a temperature range extending over the β transus (882°C) was generated by the electrical discharge at input energies of 2·5 kJ or higher.

OPEN IN VIEWER

Figure 4

Optical micrographs of cross-section of Ti rods after electrical discharging with input energies of a 2·0, b 2·5, c 3·0 and d 3·5 kJ

Therefore, the temperature rise (ΔT), which is caused by input energy during the electrical discharge process, is now considered and estimated using equation (4) 4 where m is the mass of the Ti and C_p the specific heat of Ti. ³² The electrical input power (W) was calculated by integrating current and voltage as a function of discharge time. The resulting temperature increase data during the electrical discharge process under the current conditions are also listed in Table 1. It can be seen that the electrical discharge process can produce temperature increases ranging from 569 to 2398°C, depending on the input energy. It is known that the instant heat generated by electrical discharge with 2·0 kJ of input energy did not cause any phase transformation, as confirmed in Fig. 4a. On the other hand, the electrical discharge process using more than 2·5 kJ of input energy produced temperatures much greater than the β tansus temperature, resulting in a β→α allotropic transformation even with times as short as 159 μs.

The hardness measurements, conducted on the cross-sections of Ti rods before and after the electrical discharge process with various input energies, are given in Fig. 5. The hardness near to the surface of an as received Ti rod is slightly higher than the value in the centre, which can be attributed to the surface hardening induced from processes such as swaging or drawing. After electrical discharge from 2·0 to 2·5 kJ there is no significant hardness change at the surface and a slight increase in hardness at the centre of the rod at 2·5 kJ. Moreover, the hardness near to the surface of the Ti rod, which was electrically discharged with 3·5 kJ of input energy, was about 2·6 times higher the value in the centre. This can be possibly attributed to electrical discharge induced surface hardening.

OPEN IN VIEWER

Figure 5

Micro hardness values (on centre and edge of cross-section) of Ti rods before and after electrical discharge at various input energies

To understand such bimodal characteristics of the hardness value on the cross-section, microstructures near to the surface at different electrical discharge conditions were considered. The microscopic morphology on the surface tends to depend on the input energy, as shown in Fig. 6. A very thin external surface layers with a thickness of about 1–2 μm have been formed on the electrical discharged Ti rods. The external surface layer is referred to as an oxide and carbide layer [(O,C)L], which is dense and integrated with the substrate ‘oxygen and carbon diffusion zone (DZ)’.^8,9 The thicknesses of the DZ on each electrically discharged Ti rod were determined based on the morphological change as shown in Fig. 6 and are tabulated in Table 2, indicating that the thickness depends on the input energy. During the electrical discharge process, saturation of oxygen and carbon in the Ti is reached immediately at the [(O,C)L]/substrate interface and then the dissolution in the lattice gradually decreases.^9,33 Since higher input energy produces more heat generated, the dependence of DZ layer on the input energy results from the diffusion rate. The different hardness value on the electrical discharged Ti rods can thus be due to the variation of oxygen and carbon concentration in the DZ layer.

OPEN IN VIEWER

Figure 6

Cross-sectional SEM images of electrical discharged Ti rods at input energy of a 2·5, b 3·0 and c 3·5 kJ

OPEN IN VIEWER

Table 2

Thickness of diffusion zone (DZ) and diffusion coefficient (D) of electrical discharged Ti rod samples

Input energy/kJ	Thickness of DZ/μm	D/m2 s−1
2·0	0·1	6·33×10−11
2·5	6·5	2·67×10−7
3·0	8·2	4·26×10−7
3·5	17·5	1·94×10−6

To investigate the surface states of Ti rods before and after electrical discharging with 3·5 kJ of input energy, the surface of the Ti rods was examined by XPS. Figure 7 shows narrow scan XPS spectra of the Ti 2p region for the as received Ti rod before and after Ar⁺ etching for times up to 15 min. Previous Ti narrow scan spectra results for wrought Ti and Ti alloy show the Ti 2p_3/2 peak at about 459·1 eV with 5·8 eV splitting between the Ti 2p_1/2 and Ti 2p_3/2 peaks.^30,34 For Ti⁴⁺, as in TiO₂, the Ti 2p_3/2 peak is at about 459·1 eV. Thus, the surface of the as-received Ti rod appears to be primarily in the form of TiO₂. However, the surface spectrum of the etched, as received Ti, revealed a Ti 2p_3/2 peak at about 453·8 eV, representing the presence of metallic Ti. It can thus be known that the as received Ti rod was lightly oxidised in the form of TiO₂.

OPEN IN VIEWER

Figure 7

XPS narrow scan spectra of Ti 2p region for as received Ti rod before and after Ar⁺ etching for 15 min

Figure 8 shows narrow scan spectra of the Ti 2p region for the electrically discharged Ti rod with 3·5 kJ of input energy before and after Ar⁺ etching. The surface of the Ti rod after electrical discharging showed a Ti 2p_3/2 peak at about 458·9 eV, representing the oxidised surface in the form of TiO₂. However, after an etching, an asymmetry on the low binding energy region and broadness of Ti 2p_3/2 peak were observed. This indicates the existence of complex overlapping oxides and carbides. ³⁵ The maximum peak at 458·8 with 5·8 eV splitting between the Ti 2p_1/2 and Ti 2p_3/2 peaks corresponds to Ti oxide in the form of TiO₂. Peak shifts for TiC_xO_y and TiC are approximately −2·0, and −3·6 respectively. Therefore, the surface of the electrical discharged Ti rod is primarily in the mixed form of TiO₂, TiC_xO_y, and TiC.^35,36

OPEN IN VIEWER

Figure 8

XPS narrow scan spectra of Ti 2p region for electrical discharged Ti rod with 3·5 kJ of input energy a before and b after Ar⁺ etching for 15 min

Figure 9 shows narrow scan spectra of the C 1s region. In Fig. 9a, the surface of the electrical discharged Ti rod shows a maximum peak at 285 eV with a full width at half maximum (FWHM) value of 2 eV, representing carbon in the C–H form. Adventitious carbon is normally detected on the surface of Ti and Ti–6Al–4V after light Ar⁺ etching.^22,23 However, in the etched, electrical discharged Ti rod, the carbon 1s peak at 281·9 eV was also observed as shown in Fig. 9b, which can be assigned to Ti–C bonds. ¹⁶ The electrical discharge process thus resulted in the formation of Ti carbide. Formation of Ti carbide can be attributed to the high heat generated in the Ti during an electrical discharge process, causing reaction between Ti and C. This supports the presence of carbon in the form of TiC_xO_y in Fig. 8b. The source of carbon for the formation of Ti oxycarbide during an electrical discharge process is believed to be from adventitious carbons adsorbed on the surface of Ti sample in air. The surface concentration of carbon on the as received Ti sample was about 41·2 at-%.

OPEN IN VIEWER

Figure 9

XPS narrow scan spectra of C 1s region for electrical discharged Ti rod with 3·5 kJ of input energy a before and b after Ar⁺ etching for 15 min

The diffusion kinetics of oxygen and carbon into the Ti rod during the electrical discharge process can be analysed by using equation (5) 5 where x is the thickness of DZ, D is the diffusion coefficient, and t is the diffusion time. ³⁷ The diffusion coefficient at each discharge condition is estimated and listed in Table 2. After determination of D, the activation energy for the diffusion process was estimated to be 103 kJ mol⁻¹ from the slope of the plot shown in Fig. 10. This value is much lower than the values (149–170 kJ mol⁻¹) reported in the literature. ⁹ This result indicates that the electrical discharge process for the surface modification of Ti can decrease the activation energy for the oxygen and carbon diffusion process.

OPEN IN VIEWER

Figure 10

Arrhenius plot showing ln (D) versus T⁻¹

Conclusions

Titanium rods were subjected to 2·0–3·5 kJ of input energy from a 450 μF capacitor bank. The electrical discharge process with an input energy greater than 2·5 kJ caused a β→α allotropic transformation showing an typical serrated morphology. Moreover, instant temperature increases generated in times as short as 159 μs during the electrical discharge process are believed high enough to allow both oxygen and carbon penetration at the titanium surface, producing a TiC_xO_y surface layer. The obtained bimodal characteristics of hardness on the cross-section of the titanium rod can thus result from the solid solution hardening induced by electrical discharge process. The electrical discharge technique for the surface modification of titanium is more efficient in terms of process time and provides a simple, easy method compared to other more conventional oxycarburising processes.

Acknowledgements

This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2013R1A1A2010207).