Effect of pulse frequency on the surface properties and corrosion resistance of a plasma-nitrided Ti-6Al-4V alloy

In this work, Ti-6Al-4V alloy, commonly used as implant material in biomedical applications, was treated by plasma nitriding. The nitriding process was carried out using an N2-H2 plasma (1000:500 sccm) at an operating pressure of about 866 Pa. The current regulation was about 1.8 A, the negative voltage was about 480–500 V, and the power was 840–940 W. The nitriding temperature was maintained at 650 ± 5 °C, and the nitriding time was 240 min. Bipolar pulse frequencies were varied at 25, 50, 100, 150, and 200 kHz. Analysis by grazing incidence x-ray diffraction spectrometer (GI-XRD) revealed the presence of δ-TiN and ε-Ti2N phases in all nitrided samples. The hardness depth profile was measured with a penetration depth of about 5 nm using the enhanced stiffness procedure (ESP). The results showed that all the nitrided samples had a surface hardness approximately three times that of the unnitrided sample. This result is consistent with that from glow discharge emission spectroscopy (GD-OES), which confirmed the diffusion distance of nitrogen atoms from the surface of about 5 μm. After plasma nitriding, the surface roughness tended to increase, resulting in an increase in the water contact angle (WCA) and a decrease in the work of adhesion. The specific wear rate (ball-on-disk) of all nitrided samples decreased and was significantly lower at a bipolar pulse frequency of 50 kHz. This result is consistent with the stability of the coefficient of friction (COF) after 6000 sliding cycles. Moreover, the nitrided sample at 50 kHz exhibited the lowest corrosion current density in artificial saliva based on the Tafel potential polarization method.


Introduction
Nowadays, medical technology has made rapid progress so that human life has become longer, and the quality of life has also become more important.One of these medical technologies is the use of biomaterials to replace organs in the human body or to restore and improve the internal organs of the body.Currently, there are various types of medical materials and devices that can be used as implants in the human body for both dental and medical applications, such as dental implants, small orthodontic screws, orthodontic screws, artificial bone, etc When selecting the metal to be used for manufacturing implants, it is necessary to consider properties that are favorable for the function of the parts, such as hardness, stiffness, flexibility, elongation, tarnish resistance, unit weight, casting properties, as well as cost factors [1].
There are many types of implant metals, including grade 4 gold alloys, cobalt-chromium alloys, nickelchromium alloys, stainless steel, and titanium alloys.Currently, pure titanium and titanium alloys are used to make dental frameworks because they have good properties such as high strength, light weight, and high corrosion resistance.The most popular titanium alloy in dentistry is Ti-6Al-4V, it consists of 90 wt% Ti, 6 wt% Al and 4 wt% V [2].At room temperature, there are two phases: α-phase and β-phase, also known as α+β-Ti alloy.The added aluminium stabilizes the α-phase and gives the metal higher strength and lower weight, while the vanadium stabilizes the β-phase at room temperature.It has the potential to be used in dentistry, and the use of this alloy has been shown to be biologically acceptable [3].However, titanium alloy parts have some limitations, such as permanent deformation, discoloration of the titanium surface, and the problem of tribological properties of titanium metal [4,5].In addition, this alloy releases both aluminium and vanadium, both of which can cause biological problems [3,6].Aluminium affects bone mineralization [7], leading to structural deficiencies, and vanadium is both cytotoxic and capable of causing type IV allergic reactions [8].
To overcome these problems, implant devices from Ti-6Al-4V alloy have been subject to various surface engineering methods to improve their surface properties and corrosion resistance in aggressive environments.Nitriding, a widely used process to improve the surface properties of titanium alloys.The nitride phase is chemically inert, meaning that it does not react with most chemicals or biological substances.This inertness is important in preventing chemical reactions with body fluids and ensuring the stability and safety of implants and medical devices over time, and it can coexist with living tissue without causing undesirable reactions.Some nitride phases, such as aluminum nitride (AlN) and titanium nitride (TiN, Ti 2 N), are highly corrosion resistant, so they do not degrade or deteriorate when exposed to body fluids or other corrosive environments in the human body.A widely used process to improve the surface properties of titanium alloys is gas nitriding, which is carried out in the gas phase in the temperature range of 800 °C-1050 °C [9,10] in a nitrogenous environment (N 2 , N 2 -Ar, N 2 -H 2 , or N 2 -NH 3 ) [11][12][13][14][15] with a nitriding time of 6 to 80 h.However, gas nitriding requires high temperature for dissociation of bonds and leads to fatigue deterioration.The surface of the sample after gas nitriding usually has a high thickness compound layer, which needs to be polished, or surface finished before use.Plasma nitriding of titanium alloys with electric discharge is usually performed at lower temperatures and shorter nitriding times.This process can improve the tribological properties of titanium alloys by forming a thin compound layer mainly composed of δ-TiN and ε-Ti 2 N. The microstructure of the nitride layer also consists of a region α-case (nitrogen-stabilized α-titanium) and a diffusion layer.This leads to higher surface hardness and better tribological and fatigue properties [16][17][18][19][20].
The high temperatures (>800 °C) in the nitriding process usually lead to grain growth, over-aging, and microstructural transformations in titanium substrates that reduce fatigue limit, strength, and ductility [15,21,22].In addition, the significant stiffness differences between the compound layer and the titanium substrate, as well as the brittleness of the compound layer and α-case, lead to early failure originating from the surface and the over compound layer acts as a ceramic skin was a high hardness but which is brittle [23,24].For this reason, low-temperature plasma nitriding is used to improve the mechanical properties of titanium alloys [10,21,25].The slow nitriding kinetics of titanium alloys at low temperatures can be enhanced by plasma cleaning before the nitriding process [26,27].This plasma cleaning deepens the nitrogen diffusion and increases the load bearing capacity of the plasma nitrided surfaces [28].
The important parameters of plasma nitriding include substrate temperature, gas mixture, plasma density, and ion bombardment energy.The substrate temperature is a crucial factor for the grain boundary expansion, which is responsible for the diffusion of nitrogen atoms from the surface into the base material.In addition, the substrate temperature also affects the phase formation between titanium and nitrogen.High nitriding temperatures can lead to the formation of brittle nitride layers and microstructural changes in the bulk [29].The gas ratio between nitrogen and hydrogen is also important, because the increase of the nitrogen partial pressure directly affects the increase of the nitrogen concentration near the surface and the precipitation of hard TiN phases in the formed compound layer.This result can lead to microcracks on the surface [30].Hydrogen content in plasma can remove inherent surface oxides and increase nitriding kinetics in titanium alloys by providing easy diffusion pathways for nitrogen interstitials [11,21,31].Plasma density is the essential factor for plasma nitriding, which depends on the power input and the type of power supply.A higher plasma density corresponds to a higher ion flux bombarding the sample, resulting in rapid self-heating of the substrate.At the same time, a bias voltage, e.g. in the form of an asymmetric bipolar pulse, can increase the impulsive force to accelerate the ion bombardment of the sample.A small positive voltage can neutralize the positive charges that accumulate on the sample surface.This step helps to ensure that the ions are continuously impinging on the surface of the material.Increasing the high-energy ions and pulse frequency has a direct effect on increasing the sputtering rate, which can change the surface roughness and structure of compound layer.In a typical plasma nitriding process, an external voltage is used to ionize the nitriding gas and provide an active nitrogen flux for surface modification between the nitriding furnace (anode) and the workpiece (cathode) [4,5].In addition to the power supply from DC and AC, the plasma can also be generated by radio frequency (RF) excitation [5,32].When the applied voltage is provided by a DC power supply, the continuous voltage supply can cause localized heating, overheating of thin sections, arcing, and other form of surface damage.These problems can be avoided by using an asymmetric bipolar pulsed power supply in which the heat input is controlled by duty cycles and frequency.A duty cycle is defined as the ratio of the pulse on-time to a full on-off cycle time [33] and is typically on the order of 10%-50% of the period without disturbing the nitriding time [31].Bipolar pulse discharges with sufficiently low duty cycle can be used for such low-temperature processes, especially when arcing at the edge shape of the sample must be avoided.Discharges generated with different short pulse waveforms have found applications in plasma nitriding by generating reactive species at ambient pressure and temperature [34].However, the effects of pulse frequency on the surface properties and corrosion resistance of Ti-6Al-4V alloy in artificial saliva have not yet been investigated.The expected results of this research should be useful for dental applications.

Materials and methods
In this study, the Ti-6Al-4V plate was used as the substrate to fabricate the nitride layer.The Ti-6Al-4V plate with a dimension of 30 × 35 × 1 mm 3 was mirror polished on both sides by buffing cloth wheels.The sample was first cleaned with an ultrasonic cleaner in acetone for 10 min, then shaken in methanol for 10 min twice and rinsed with deionized water three times to remove the impurities present on the substrates.The experimental setup for the plasma nitriding system is schematically shown in figure 1, which consists of the vacuum chamber, the evacuation system, the asymmetric bipolar pulse power supply in the frequency range of 0 to 250 kHz, the gas supply with Ar, N 2 (99.995% purity), and H 2 (99.995% purity), and the cooling system.Since there was no additional external heater, the substrate was heated by ion bombardment in the plasma.The substrate holder was connected to the thermocouple on the left side and to the power supply on the right side.In the vacuum chamber, the sample was placed on the power electrode.
In this work, an asymmetric bipolar pulse power supply (Advanced Energy Pinnacle ® Plus+) with peak voltages up to 800 V, a maximum power of 10 kW and an adjustable frequency range from 0 to 250 kHz was used to maintain the plasma.Preparation of the nitride layer was divided into three processes, plasma cleaning, preheating and plasma nitriding.For plasma cleaning, Ar and H 2 were used at a flow rate of 500 sccm and an operating pressure of about 33 Pa.By using the power control mode at low pressure corresponding to low plasma current and high discharge voltage, high ion energy can be generated to impinge on the substrate.Ar was chosen to generate plasma to promote ion bombardment and physical ablation of impurities, while H 2 plasma was necessary to remove the oxide layer on the surface of the sample.In this process, a power of about 500 W was controlled with a bipolar pulse frequency of 50 kHz, a duty cycle of 10%, a negative voltage of about 625-800 V, currents of 0.6-0.8A, and a process duration of 20 min, raising the temperature of the substrate to 350 ± 5 °C.After the plasma cleaning was completed, the preheating process was immediately started by decreasing the pump speed to increase the operating pressure to about 466 Pa.At higher pressure, the plasma current can be adjusted up to 2.5 A, corresponding to 800-1100 W power, to increase the ion flux into the substrate.To avoid cracks in the sample due to rapid expansion, the temperature of the substrate was gradually increased until it reached 650 ± 5 °C, which took about 45 min.Then the nitriding process was performed with a N 2 -H 2 plasma (1000:500 sccm).The hydrogen mixture is the key to the presence of the atomic hydrogen, which can remove the titanium oxide layer formed during the preparation of the sample and stabilize the operating pressure.The higher operating pressure was maintained at about 866 Pa, which corresponds to a plasma current of about 1.8 A. In this step, the substrate temperature was maintained at 650 ± 5 °C with a holding time of 240 min.To investigate the effects of pulse frequency on the properties of the nitrided layer, the pulse frequency was varied at 25, 50, 100, 150, and 200 kHz.The process parameters of plasma cleaning, preheating and plasma nitriding are summarized in table 1.
After completion of the plasma nitriding process, the nitrided Ti-6Al-4V was cooled to room temperature in the vacuum chamber.The nitrided samples were characterized using the environmental scanning electron microscopy (ESEM) (Thermo Fisher Scientific, Quattro S) with the secondary electron mode to observe the surface morphology.Contact angle measurement (OCA 15EC, DataPhysics, Germany) using a static sessile drop method was used to estimate the work of adhesion between water droplets and the surface of titanium alloys.It can be calculated based on the Young-Dupré equation by measuring the contact angle of the water on the surface [35]. 1 μl of deionized water was dropped onto the surface of the sample.The drop was backlighted with LEDs and immediately photographed by a camera.The drop was left in situ for one minute, then the contact angle was measured at three locations on each sample.The environmental conditions were constant throughout the experiment at a temperature of 25 °C-26 °C, and a relative humidity of 69%-72%.The roughness and topography of the samples before and after plasma nitriding were documented using an atomic force microscope (AFM) (XE 70 model, Park systems, Korea).Surface topography was investigated in a tapping mode by scanning an area of 5 × 5 μm 2 .Three regions of each sample were selected for roughness measurements.The roughness profiles and surface geometry parameters were determined based on the AFM data in non-contact mode.The friction and wear behaviors were assessed using ball-on-disk friction testing (FPR-2100; RHESCA Co., Ltd, Tokyo, Japan).The tribological tests were performed at room temperature.The size of wear samples with a dimension of 30 × 35 × 1 mm 3 .A 6 mm in diameter of stainless steel ball (SUS440C) was slid over the surface of the samples under a normal applied load of 0.98 N, a rotational speed of 100 rpm, and 6000 frictional revolutions.The tests were repeated three times for each sample at the rotation radius of 3, 5, and 7 mm, corresponding to a total sliding distance of 113.04, 188.40, and 263.76 m, respectively.After the test, the average specific wear rate of the sample was calculated using the profiles of the wear track (three points in each rotation radius) measured by a color 3D laser scanning microscope (KEYENCE VK-9710).Morphology of the wear track was also characterized by an optical microscope.The elemental composition in the depth profile of the surface before and after plasma nitriding was investigated using the glow discharge optical emission spectroscopy (GDOES).The GDOES (model: GD PROFILER HR) with the polychromator mode was operated at a pressure of 600 Pa and a power of RF of 20 W with an anode size of 4 mm.The phase composition and microstructural properties of the samples after plasma nitriding were analyzed by grazing incidence X-ray diffraction spectrometer (GI-XRD) (BRUKER, D8 Advance).A Cu Kα source with a wavelength of 0.15418 nm was used as the X-ray source.Tests were performed at a fixed incidence angle of 0.7°and with slit widths of 0.1 mm for both the input and output beams to the detector.The sampling rate of the diffractometer was set to 0.1°per second, with a 2θ diffraction angle of 20°−80°.The hardness value corresponding to the depth was measured with a nanoindentor (FISCHERSCOPE ® HM2000) based on an enhanced stiffness procedure (ESP).This method is particularly suitable for depth-dependent measurement, which allows the hardness of the nitride layer to be determined at very low forces without being affected by the bulk.As the force increases, with loading and unloading increased at intervals from 0.1 mN to 2000 mN, the transition from the surface to the base material can also be analyzed.The value of the hardness depth profile can be used to estimate the thickness of the nitride layer.The interaction between saliva and Ti-6Al-4V alloy surfaces can directly affect their corrosion behavior.Artificial saliva serves as an important test medium for evaluating the performance and biocompatibility of dental materials.Therefore, in this work, the nitrided samples were corroded in an artificial saliva solution with a pH of 5.5 under simulated oral environment before they are placed in the human oral cavity.An electrochemical test (GAMRY, Reference3000) was performed using a potentiodynamic polarization method with a three-electrode configuration, consisting of reference (Ag/AgCl), counter (Pt), and working electrodes (nitrided sample).An immersion time of where CR is the corrosion rate in mm/year (mmpy), K is 3.272 × 10 −3 mm g/(μA cm yr), i corr is the selfcorrosion current density in μA cm −2 , ρ is the density in g cm −3 , and EW is the equivalent weight.For the Ti-6Al-4V alloy, the equivalent weight and density are 11.89 and 4.43 g cm −3 , respectively.

Surface morphology
As can be seen in figure 2, the color of the Ti-6Al-4V samples changes from metallic grey to gold due to the increase in the atomic ratio of nitrogen to titanium, which causes the overall reflectivity of the gold-like color [37][38][39].The lighter color appears around the edge of all nitride samples.This edge defect is due to the higher electric field resulting in higher energy of ion bombardment and more sputtering in this region.Prior to the determination of surface morphology, all samples were immersed in 98% H 2 SO 4 and 37% HCl acids in a volume ratio of 1:1 for 20 h.The surface morphology of the samples after etching was then observed using the ESEM technique.The surface morphologies of the unnitrided samples before and after etching are shown in figure 3(a).It is obvious that the unnitrided sample shows clear signs of corrosion, so that the grain boundaries can no longer be seen.In contrast, if one compares the surface of the nitrided samples before and after etching, almost no traces of corrosion can be seen.This result shows that the process of plasma nitriding can improve the corrosion resistance of a titanium alloy by creating a protective barrier.
In addition, the effect of hydrogen in the plasma cleaning and plasma nitriding processes in this study also accelerated nitrogen diffusion by forming multiple dislocations within the α-grains.It has also been reported that Ar-H 2 sputtering increases the nitriding kinetics in alloys by forming simple diffusion paths for nitrogen interstitials and eliminating inherent surface oxides [43,44].Therefore, no titanium oxide formation was observed.Based on the XRD results, the correlation between the intensity of phase formation and the bipolar pulse frequencies used during the nitriding process is still not clear.

Chemical composition
Nitrogen diffusion and concentration were analyzed using the glow discharge optical emission spectroscopy technique (GD-OES) to measure the content of nitrogen and other elements (see figure 6).The argon ion etch rate was set at about 1 μm per minute.The output data were subtracted with the background ionization and surface adsorption of residual atomic nitrogen, which could be the cause of the emission lines detected by the spectrometer.It can be seen that the nitrogen content on the surface is up to ∼45% and gradually decreases with increasing etching time.In addition, the aluminium content near the surface decreases with increasing nitrogen diffusion from the surface to the core.At a depth of one micrometer from the surface, the aluminium content increased more than that of the unnitrided sample, which is probably related to the precipitation of AlN.Moreover, the curve of vanadium initially drops to almost zero due to diffusion of nitrogen atoms from the surface, which reduces the relative vanadium concentration.However, upon closer inspection, the vanadium content is found to be above 5%, which is the same as in most titanium alloys.

Mechanical properties
The surface hardness in the depth profile of the sample was analyzed using the nanoindentation technique in the enhanced stiffness procedure mode (ESP), as shown in figure 7. It can be seen that all nitrided samples have higher surface hardness than the unnitrided sample.The nitrided samples have the highest surface hardness of about 1250 HV at bipolar pulse frequencies of 50, 150, and 200 kHz, which is close to a hardness value of the ε-Ti 2 N phase [31,45], which can be clearly seen in the results of GI-XRD.The TiN layer is generally slightly stiffer than the diffusion layer.Based on the hardness depth profile, the thickness of the nitride layer can be roughly estimated at 5 μm.Due to its lower processing temperatures, plasma nitriding is able to achieve high surface hardness while maintaining the high mechanical properties of the core material.In addition, plasma nitriding allows control the microstructure of the treated workpiece, phase and chemical composition, surface topography, morphology, and residual stresses [46], all of which are critical for determining surface properties, especially if the workpiece is to be used as a part in the human body.

Surface roughness and work of adhesion
In this work, root mean square roughness (R rms ) was measured using atomic force microscopy, which is more sensitive to peaks and valleys than average roughness because of the squaring of the amplitude in its calculation.figure 8 shows the surface topography corresponding to the scan area of 25 μm 2 of the unnitrided and plasma nitrided samples.The surface roughness of the nitrided samples prepared with bipolar pulse frequencies of 25, 50, 100, 150, and 200 kHz is 33.6 ± 1.0, 25.5 ± 1.2, 31.0 ± 28.8 ± 1.7, and 38.1 ± 3.6 nm, respectively.It can be seen that the surface roughness of the nitrided samples is higher than that of the unnitrided samples (20.8 ± 3.7 nm).This result can be attributed to the formation of the compound layer.Typically, the compound layer is rough and porous, which leads to the reformation of material that could cause the formation of whiskers (TiN nanoparticles) [47].No significant improvement in surface topography when bipolar pulse frequency was  increased.However, the bipolar pulse frequency of 200 kHz tends to give the highest surface roughness.Increasing the frequency enhances the plasma density, which increases the probability of nitrogen ions colliding with the surface of the workpiece.These high-energy ion collisions usually lead to an increase in surface roughness.In addition, the modulation caused by the bipolar pulse frequency affects the energy distribution of the sputtered Ti atoms by allowing them to redeposit in the substrate.This modulation increases the atomic mobility on the surface and facilitates the epitaxial replication of the crystallography found in the seed layer.Consequently, this modulation contributes to variations in surface roughness [48].Figure 9 shows the contact angle, work of adhesion, and surface roughness of Ti-6Al-4V before and after plasma nitriding.For each sample, the measurement was repeated three times in different positions.The contact angle of the unnitrided sample was 86.39°± 3.18°.For the nitrided samples with bipolar pulse frequencies of 25, 50, 100, 150, and 200 kHz, the contact angles increase slightly to about 94.65°± 1.05°, 92.75°± 0.93°, 91.61°± 2.12°, 93.42°± 1.71°and 93.28°± 0.87°, respectively, corresponding to a slight increase in surface roughness.The work of adhesion between water droplets and the surface of titanium alloys can be calculated based on the Young-Dupré equation [35].The work of adhesion at the bipolar pulse frequencies of 25, 50, 100, 150, and 200 kHz was 76.30, 65.96, 68.34, 69.76, and 67.50 mJ/m 2 , respectively.In general, a surface is considered hydrophobic if the contact angle is greater than 90° [49,50].It is observed that the contact angle of the nitrided samples is higher than that of the unnitrided sample.This could be due to the Cassie-Baxter wetting model [51], where air is trapped in the pockets below the droplet during the measurement of the advancing contact angle.The micropores formed after nitriding and the surface roughness lead to a more hydrophobic surface.Therefore, the corrosion rate should be reduced for nitrided samples with increased hydrophobic properties, since water accumulation on the surface can accelerate corrosion.Therefore, a hydrophobic surface is desirable to prevent water from adhering.Plasma nitriding can improve the hydrophobic property of Ti-6Al-4V, making it more suitable for human applications.

Coefficient of friction and wear rate
Figure 10 shows the coefficient of friction (COF) of unnitrided and nitrided samples at a rotation radius of 5 mm as a function of the number of sliding cycles.It can be seen that the COF of the unnitrided sample is about 0.1 during the interval of 1000 cycles, which is due to the behavior of the natural titanium oxide layer.Thereafter, the COF increases to about 0.3, which is consistent with the surface of the titanium alloy [52].For the plasma nitrided samples, the COF is 0.5-0.6 during the interval of 2500 cycles.The increase in COF is likely related to  the higher surface roughness of the nitride layer.As the number of sliding cycles increases, the COF of the nitride samples decreases sharply to the titanium alloy surface value at the bipolar pulse frequencies of 25 kHz, 100 kHz, 150 kHz, and 200 kHz.This behavior indicates detachment of the brittle nitride layer.However, the nitrided sample at 50 kHz maintains a stable coefficient of friction of about 0.6 throughout the test period and remains intact without flaking off.This can be attributed to the formation of the δ-TiN phase, which has a higher hardness compared to the ε-Ti 2 N phase.Moreover, this stability of the nitride layer is likely due to the lowest surface roughness seen on the AFM image.Consequently, the nitride layer exhibits better mechanical properties, is scratch resistant and leads to a lower wear rate.
Figure 11 shows the wear tests after a test with 6000 sliding cycles and dry sliding conditions.The left column shows the photo of all worn samples with rotation radius of 3, 5 and 7 mm, the middle column shows the wear tracks with sliding width and the right column shows the three-dimensional (3D) profilometer images (radius 5 mm).As can be seen in figure 11(a) for the unnitrided sample, the sliding width of the worn track is about 0.72 mm, which is consistent with the sliding width of the worn track determined from the 3D profilometer image.The average depth and width of the worn tracks with a rotation radius of 5 mm were used to calculate the specific wear rate.It can be seen that the worn surfaces of the unnitrided sample have the highest specific wear rate after sliding.The nitrided sample at 50 kHz, as shown in figure 11(c), still has a gold-colored surface indicating the lowest specific wear rate of 0.27 × 10 −4 mm 3 /Nm.This outcome is attributed to the creation of a δ-TiN phase characterized by superior hardness, surpassing that of other conditions.This result is consistent with the stability of the COF of the nitride layer.
Figure 12 shows the average specific wear rate of unnitrided and nitrided samples prepared with different bipolar pulse frequencies.For each sample, the specific wear rate was determined at three positions with a rotation radius of 5 mm.All nitrided samples with higher surface roughness have lower wear rates than the unnitrided sample.This reduction in wear rate can be attributed to the higher surface hardness achieved by the plasma nitriding process.

Corrosion resistance
Figure 13 shows the potentiodynamic polarization curve of unnitrided and nitrided samples, which was generated in an artificial saliva solution with a potential (E versus Ag/AgCl) of −1000 mV to 1500 mV, using a sampling rate of 1 mV s −1 .Lower i corr values and more positive E corr values indicate better corrosion resistance; therefore, all nitrided samples have higher corrosion resistance compared to the unnitrided sample.Although the surface of the unnitrided sample has a robust oxide layer, there is a possibility that metastable processes will occur if a film breakdown and re-passivation process is initiated.This process leads to the formation of grain boundaries and more open pathways within the oxide.Once a crevice is formed, it spreads rapidly and leads to corrosion [53].The nitrided sample prepared at a frequency of 50 kHz exhibits the highest corrosion resistance.This is evident from the highest values of self-corrosion potential E corr and the lowest self-corrosion current density i corr , as shown in figure 14.A low i corr value means a lower amounts of ions released into the human body, which can serve as an indicator of better biocompatibility.This result is probably related to the lowest roughness and the highest wear resistance.When the bipolar pulse frequency is higher than 50 kHz, the selfcorrosion current density tends to be higher.This phenomenon is probably related to the fact that the molecular nitrogen ions and the nitrogen ions tend to react more slowly at the higher excitation frequency during the nitriding process.In conjunction with the wear test, there may be a strong self-detachment in the compound layer leading to the formation of a larger porosity, resulting in a string that separates the upward growing crystallites from the downward growing ones [54] .
Assuming that the nitrided and unnitrided samples have similar equivalent weight and density, the changes in corrosion rates are proportional to the changes in self-corrosion current density.Figure 15 shows the corrosion rate of the samples after nitriding at different bipolar pulse frequencies.The nitrided sample prepared at a frequency of 50 kHz has the lowest corrosion rate of 0.51 × 10 −4 mmpy, while the CR of the unnitrided sample is 1.4 × 10 −4 mmpy.This means that the corrosion rate is reduced by three 3 times compared to that of the unnitrided sample.

Conclusions
The aim of this study is to investigate the effects of different bipolar pulse frequencies on the surface properties and corrosion resistance of Ti-6Al-4V alloy in artificial saliva.For this purpose, the Ti-6Al-4V alloy was nitrided with a N 2 -H 2 plasma (1000:500 sccm) at an operating pressure of 866 Pa, a power of about 840-940 W, and a nitriding temperature of 650 ± 5 °C for 240 min.The bipolar pulse frequency was varied between 25 and 200 kHz.The results from GD-OES show that the nitrogen atom can diffuse up to 45% into the surface and gradually decreases with depth.The results of GI-XRD show the formation of the phases δ-TiN and ε-Ti 2 N.
Atomic force microscopy shows that the plasma nitriding process leads to an increase in surface roughness, which is probably due to the formation of the compound layer.The increase in water contact angle after plasma nitriding is likely due to the higher surface roughness.The hardness depth profile of a nitrided Ti-6Al-4V alloy was investigated by nanoindentation in the mode of enhance stiffness procedure.The results show that all nitrided samples have higher surface hardness (about 1260 HV) compared to the unnitrided sample (450 HV).The hardness values tended to decrease with increasing depth but did not affect the bulk hardness.The thickness of the nitride layer was estimated to be about 5 μm.The coefficient of friction and wear rate were determined by a ball-on-disk test.The nitrided sample with the bipolar pulse frequencies of 50 kHz gives an average COF of about 0.6, which is higher than the unnitrided sample due to the higher surface roughness of the nitride layer.For the other nitride samples, the COF decreases rapidly with increasing number of sliding cycles, indicating detachment of the brittle nitride layer during the test.This result is consistent with the specific wear rate, where the nitrided sample has the highest wear resistance and the lowest self-corrosion current density, indicating the highest corrosion resistance at 50 kHz.

Figure 1 .
Figure 1.Schematic diagram of the plasma nitriding system.
180 min was applied to the open circuit potential of the alloy in all electrolytes before starting the tests, then a potential range of −1000 mV to 1500 mV was applied with a sampling rate of 1 mV s −1 .Tafel extrapolation was used to estimate self-corrosion current density (i corr ) and self-corrosion corrosion potential (E corr ) in an electrochemical cell.Corrosion resistance was assessed by measuring i corr and E corr .The corrosion rate (CR) can also be estimated based on the ASTM Standard G 102-89 using the following formula[36]

Figure 2 .
Figure 2. The surface physical appearance of unnitrided and nitrided samples.

Figure 3 .
Figure 3. Surface morphology of unnitrided and nitrided samples before and after 20-hour immersion in 98% H 2 SO 4 and 37% HCl acids in a 1:1 volume ratio.

Figure 4 .
Figure 4. Grazing incidence x-ray diffraction spectra obtained from the nitrided and unnitrided samples.

Figure 5 .
Figure 5. Phase content from semi-quantitative analysis based on the XRD patterns of Ti-6Al-4V alloy before and after plasma nitriding.

Figure 7 .
Figure 7. Nanoindentation test (ESP mode) indicates hardness of surface of nitrided samples.

Figure 9 .
Figure 9.The surface roughness, contact angle and work of adhesion of unnitrided and nitrided samples.

Figure 10 .
Figure 10.Coefficient of friction of unnitrided and nitrided samples as a function of number of sliding cycles.

Figure 12 .
Figure 12.Average specific wear rate at the rotation radius of 5 mm of unnitrided and nitrided samples prepared at different bipolar pulse frequencies.

Figure 14 .
Figure 14.Self-corrosion potentials (E corr ) and self-corrosion corrosion current (i corr ) of unnitrided and nitrided samples.

Figure 15 .
Figure 15.Corrosion rate of the samples after nitriding at different bipolar pulse frequencies.

Table 1 .
Process parameters used to study the effect of pulse frequency on the surface properties of plasma nitrided Ti-6Al-4V alloy.