2017 Volume 58 Issue 4 Pages 600-605
We developed a new surface modification technique called “iron-powder pack (IPP) treatment”. A layer of titanium carbonitride, Ti(C, N), was formed on the surface of a titanium sample embedded in a mixture of iron, graphite, and alumina powders and held around 1273 K in a nitrogen flow. In this work, IPP treatment using a 4:6:3 (volume ratio) mixture of iron, graphite, and alumina powders was applied to titanium plates, and the effects of the heating temperature and nitrogen gas flow rate on the microstructures near the titanium surface were investigated. The Ti(C, N) layer was observed on the titanium plate heat-treated at 1173 K for 3.6 ks at a nitrogen flow rate of 0.5 L/min. This layer became uniform and thick as the heating temperature increased. At 1373 K, a Ti(C, N) layer with a thickness of more than 30 μm that increased the hardness of the titanium surface to an HV value of about 1500 was obtained. In addition, an increase in the nitrogen flow rate increased the surface hardness further. Spherical titanium powder was treated at 1273 K to examine the growth of the Ti(C, N) layer. The layer thickness increased with the holding time. Because the average diameter of the spherical powder was unchanged after heating, the Ti(C, N) layer grew toward the inside of the titanium via the diffusion of carbon and nitrogen from the powder mixture and the atmosphere.
Titanium and its alloys have excellent corrosion resistances and high specific strengths, but their major disadvantage is poor wear resistance. This disadvantage limits their practical use as mechanical components and orthopedic implants, which require good wear resistance. Various studies have focused on hardening the surfaces of these alloys by techniques including gas nitriding,1,2) plasma nitriding,3,4) plasma carburizing,5–7) nitrogen ion implantation,8) laser melting in nitrogen,3) and electron beam irradiation.9) Most of these techniques need special equipment and complicated process flows to produce a hardened layer, such as titanium carbide (TiC) and titanium nitride (TiN), on the titanium surface. Another method achieved hardening by the diffusion of oxygen from the air at high temperatures.10,11)
Matsuura et al. have reported a simple method of forming a layer of titanium carbonitride, Ti(C, N), on the surface of titanium.12) In this method, a titanium specimen is held at 1388–1573 K in a graphite cup under a nitrogen atmosphere. We developed a similar technique through our research on dissimilar material joints.13) A titanium specimen is embedded in a mixture of iron and carbon powders, and is then heated at 1273 K for 3.6 ks in a nitrogen flow. During heating, a hardened layer, which has been identified as Ti(C, N), forms on the titanium surface. The greatest advantage of our method is that the iron powder is added to the carbon powder, leading to a lower process temperature than the method proposed by Matsuura et al.12) We propose that the iron powder increases the carbon movement in the powder mixture, like a catalyst. Our technique also enables the reduction and carbonitriding of an oxide film on anodized titanium,14) and the diffusion of carbon and nitrogen into stainless steel.15) We call this technique “iron-powder pack (IPP) treatment”, and we are verifying its practical effectiveness.
The relationship between the powder mixture and the Ti(C, N) layer formed on the titanium surface by the IPP treatment has been addressed in a previous study.13) A 4:6:3 (volume ratio) mixture of iron, graphite, and alumina powders was selected as an optimum modifying agent based on those results. In the present study, this mixture was used in the IPP treatment of titanium plates and powder, and the effects of the heating temperature, holding time, and nitrogen flow rate on the formation of the Ti(C, N) layer were investigated.
A commercially pure titanium plate with a purity of 99.5 mass% and a thickness of 1 mm was used as a substrate. The plate was cut into 10 mm × 10 mm squares and the surface was finished with #1200 waterproof abrasive paper. Spherical titanium powder with a diameter of approximately 1 mm (Fukuda Metal Foil & Powder Co., Ltd.) was also prepared to examine the layer growth of Ti(C, N) formed by IPP treatment. Figure 1 shows a scanning electron microscope (SEM) image of the titanium powder. Powder particles with diameters of 1100 to 1150 μm were most frequently observed, and the diameter distribution of the powder was measured. Before surface modification, the titanium plates and powder were degreased in acetone using an ultrasonic cleaner and dried with hot air.
SEM image of spherical titanium powder.
A schematic diagram of the electric furnace for diffusing carbon and nitrogen into the titanium is shown in Fig. 2. The IPP treatment was conducted as follows. A 4:6:3 (volume ratio) mixture of iron, graphite, and alumina powders was placed in an alumina crucible, and the titanium samples were embedded in the mixture. The iron, graphite, and alumina powders were the same as those in our previous study.13) The commercially available carbonyl iron powder was fine and spherical (particle size, D50 = 3.9–5.2 μm), and contained 0.75–0.90 mass% C, 0.65–0.90 mass% N, and 0.15–0.40 mass% O. The graphite powder was used as a carbon source and an anti-sintering agent, and the alumina powder was added to prevent the powder mixture from sintering. The crucible, which was filled with the titanium samples and the powder mixture, was placed in the electric furnace. After repeatedly evacuating the furnace with a rotary vacuum pump and refilling it with nitrogen gas, the crucible was heated to 1073–1373 K for 3.6–28.8 ks in a nitrogen flow. The nitrogen gas had a purity of >99.99 vol% and a flow rate of 0.5 L/min. The titanium sample was also heat-treated under a nitrogen flow rate of 2 L/min for comparison. All samples were allowed to cool in the furnace to room temperature.
Schematic illustration of the electric furnace for carbon and nitrogen diffusion treatment.
SEM observations, X-ray diffraction (XRD) measurements, and Vickers hardness tests were performed on the surface of the titanium plates. The accelerating voltage during the SEM observations was 15 kV, and the XRD measurements with Cu Kα ray were conducted at a tube voltage of 40 kV and a tube current of 40 mA. In the hardness test, a load of 0.98 N was applied at room temperature for 15 s. Thereafter, the plates were mounted in an electrically conductive resin, cut in half, and then ground with waterproof abrasive papers. The modified titanium powder was observed with a stereomicroscope and by SEM. The powder was mounted in resin and ground with abrasive papers until the diameter of the particle cross section was similar to the powder diameter, which was roughly 1 mm.
In both cases, the cross sections were mirror-finished by using diamond slurry with a particle size of 0.5 μm to reveal the microstructures. The microstructures were examined by optical microscopy, SEM, and electron probe X-ray micro analysis (EPMA, accelerating voltage: 15 kV). Before optical microscopy and SEM, the samples were etched in an aqueous solution containing hydrofluoric acid and nitric acid (5 vol% HF, 25 vol% HNO3, and 70 vol% H2O).
Figure 3 shows XRD patterns of the surface of a titanium plate heat-treated at 1073–1373 K for 3.6 ks in a nitrogen flow with a mixture of iron, graphite, and alumina powders. The nitrogen flow rate was 0.5 L/min. At 1073 K, most diffraction peaks were identified as α-titanium (Fig. 3(a)). Peaks for α-iron and β-titanium were also observed. The α-iron peak arose from the iron powder adhering to the titanium surface. Because iron is a β-stabilizer for titanium, the diffusion of iron into titanium would lead to the formation of β-titanium. An XRD pattern for the titanium surface modified at 1173 K is shown in Fig. 3(b). There were intense diffraction peaks that did not correspond to α-titanium, β-titanium, and α-iron. Their peaks were roughly consistent with those for TiC and TiN, and were identified as Ti(C, N), according to Matsuura et al.12)
XRD patterns of the surface of a titanium plate heat-treated at (a) 1073, (b) 1173, (c) 1273, and (d) 1373 K for 3.6 ks in a nitrogen flow. The nitrogen flow rate was 0.5 L/min.
Figure 4 shows SEM images of the titanium surfaces corresponding to the XRD patterns in Figs. 3(a) and 3(b). The linear patterns on the surface were from the mechanical grinding before heat treatment. Fine particles from the powder mixture were also observed on the surface. The surface of the titanium plate modified at 1173 K appeared to be rounded compared with that at 1073 K, and was covered with a reaction layer approximately 5 μm thick. Figure 5 shows the results of elemental mapping near the titanium surface. Titanium, carbon and nitrogen were detected in the reaction layer. The results in Fig. 3(b) show that Ti(C, N) was formed on the titanium surface.
SEM images of the surface of a titanium plate heat-treated at (a) 1073 and (b) 1173 K for 3.6 ks in a nitrogen flow. The nitrogen flow rate was 0.5 L/min.
Elemental mapping of a cross section of a titanium plate heat-treated at 1173 K for 3.6 ks in a nitrogen flow. The nitrogen flow rate was 0.5 L/min.
The XRD patterns of the surfaces of the titanium plates modified at 1273 and 1373 K are shown in Figs. 3(c) and 3(d), respectively. The main diffraction peaks corresponded to Ti(C, N). At 1273 K, there was a continuous Ti(C, N) layer on the titanium surface as shown in Fig. 6, although its thickness was not uniform. Figure 7(a) shows an SEM image of the cross section of the titanium plate modified at 1373 K. The Ti(C, N) layer was relatively uniform, and it was more than 30 μm thick in many places. A number of voids were observed in the Ti(C, N) layer. The voids were noted in the previous study,13) and were also noted by Matsuura et al.12) Thus, they appear to be related to a process of forming Ti(C, N). On the other hand, the EPMA analysis suggested that a solid solution region was produced in the titanium substrate close to the Ti(C, N) layer by the diffusion of carbon and nitrogen.
Optical micrograph of a cross section of a titanium plate heat-treated at 1273 K for 3.6 ks in a nitrogen flow. The nitrogen flow rate was 0.5 L/min.
SEM images of a cross section of a titanium plate heat-treated at 1373 K for 3.6 ks in a nitrogen flow. The flow rate of nitrogen gas was (a) 0.5 and (b) 2 L/min.
In IPP treatment, carbon monoxide (CO) gas is generated near 973 K in a heating step.13–15) This is probably due to the chemical reaction between the carbon in the powder mixture and residual oxygen in the electric furnace. Oxygen in the furnace is probably expelled because CO is released with the nitrogen flow from the furnace exhaust. Accordingly, the oxygen partial pressure should be low around the titanium sample. Therefore, this environment may contribute to the decomposition of the oxide film and to suppressing the oxidation on the titanium surface, making it easy for carbon and nitrogen to diffuse into the titanium.
Figure 8 shows the relationship between the temperature and the surface hardness of the titanium plates heat-treated for 3.6 ks in a nitrogen flow with a mixture of iron, graphite, and alumina powders. Black circles indicate a nitrogen flow rate of 0.5 L/min. The measurements were conducted seven times on the surface of each sample, and the mean value and the standard deviation are shown. The surface hardness increased with increasing heating temperature. The titanium surface was sufficiently hardened after heating at 1073 K compared with HV of approximately 160 before heating. This was probably caused by the diffusion of oxygen, carbon, and nitrogen into the titanium16) because the plate was not covered with a Ti(C, N) layer. At 1373 K, the titanium plate had a relatively uniform, thick Ti(C, N) layer on the surface and showed the highest hardness. Therefore, the increase in the surface hardness in Fig. 8 was closely related to the formation of the Ti(C, N) layer on the titanium surface.
Effects of the heating temperature and the nitrogen flow rate on the surface hardness of a titanium plate heat-treated for 3.6 ks.
The formation of Ti(C, N) during IPP treatment indicates that a chemical reaction occurs between the titanium substrate, carbon in the powder mixture, and nitrogen in the atmosphere. Therefore, the chemical composition and the hardness of the Ti(C, N) layer formed on the titanium surface could be affected by the supply of nitrogen gas. IPP treatment of titanium plates was performed at a nitrogen flow rate of 2 L/min, and its effect on the surface hardness of titanium was examined.
The relationship between the hardness and the heating temperature at a holding time of 3.6 ks is shown by the black squares in Fig. 8. At 1073 K, the titanium plate had HV of more than 600. However, comparing the results with those for the plate treated at a flow rate of 0.5 L/min demonstrated that the nitrogen flow rate did not have a large effect (black circles in Fig. 8). Figure 9(a) shows an XRD pattern of the surface of the titanium plate modified at 1073 K for 3.6 ks at a nitrogen flow rate of 2 L/min. Most of the diffraction peaks were identified as α- and β-titanium, and the diffraction pattern was similar to that in Fig. 3(a), except for the peak for α-iron.
XRD patterns of the surface of a titanium plate heat-treated at (a) 1073 and (b) 1273 K for 3.6 ks in a nitrogen flow. The nitrogen flow rate was 2 L/min.
After heating at 1273 K for 3.6 ks at a nitrogen flow rate of 2 L/min, the titanium plate was covered with the Ti(C, N) layer, as was expected from Fig 9(b). In this case, increasing the nitrogen flow rate increased the surface hardness to HV of approximately 1500. This may be because a relatively uniform Ti(C, N) layer was formed on the titanium surface, since its peak positions shown in Fig. 9(b) were consistent with those in Fig. 3(c). However, the cross-sectional microstructures of the titanium plates modified at 1373 K were similar, regardless of the nitrogen flow rate (Fig. 7). Therefore, the difference in the surface hardness due to the nitrogen flow rate was small.
The titanium was sufficiently hardened even at 0.5 L/min. For industrial use, IPP treatment with a smaller amount of nitrogen will help to reduce costs.
3.3 Holding time at 1273 KTo investigate the relationship between the holding time and the growth of the Ti(C, N) layer, IPP treatment was applied to spherical titanium powder. This allows the growth direction of the Ti(C, N) layer to be understood easily by measuring the diameter of the powder before and after the IPP treatment.
Figure 10 shows an external view of the spherical titanium powder heat-treated at 1273 K for 3.6 ks in a nitrogen flow with a mixture of iron, graphite, and alumina powders. The nitrogen flow rate was 0.5 L/min. The titanium powder, which had a metallic luster before heating, turned golden brown. All the titanium powder was the same color, indicating that the surface modification of the titanium was homogeneous. Figures 11(a) and 11(b) show SEM images of the modified titanium powder. Heating changed the surface of the titanium powder from smooth to lumpy (Fig. 1). A cross-sectional image of this powder is presented in Fig. 11(c). Consistent with the golden brown color, the powder surface was covered with a Ti(C, N) layer, which was responsible for the surface irregularity in Figs. 11(a) and 11(b). The Ti(C, N) layer was not uniform, and a number of voids were observed.
External view of spherical titanium powder heat-treated at 1273 K for 3.6 ks in a nitrogen flow. The nitrogen flow rate was 0.5 L/min.
(a) SEM image of spherical titanium powder heat-treated at 1273 K for 3.6 ks in a nitrogen flow. The nitrogen flow rate was 0.5 L/min. (b) Enlarged SEM image of the powder surface shown in (a). (c) SEM image of a cross section of the powder shown in (a).
Figure 12 shows the relationship between the holding time and the thickness of the Ti(C, N) layer formed on the titanium powder at 1273 K. The mean thickness and standard deviation are plotted against the square root of the holding time. The thickness of the Ti(C, N) layer increased with increasing holding time, although the relationship was not exactly proportional to the square root of the holding time. Furthermore, according to the standard deviation, the Ti(C, N) layer became uniform as the holding time increased.
Relationship between the holding time and the thickness of the Ti(C, N) layer formed on spherical titanium powder heat-treated at 1273 K in a nitrogen flow. The nitrogen flow rate was 0.5 L/min.
The diameter of the titanium powder was measured by using a micrograph, as in Figs. 1 and 11(a). The diameters were measured at three places for each powder, and the average of the three values was defined as the powder diameter. At least 35 powders were evaluated, and the mean value and standard deviation were 1125 ± 36 μm for the powder before heating. However, the diameters for the powder modified at 1273 K for 28.8 ks were mainly within 1100 to 1150 μm and showed a mean value of 1126 ± 33 μm. In this case, a Ti(C, N) layer with a thickness of approximately 60 μm for the powder diameter (approximately 30 μm × 2) was produced (Fig. 12). These measurements indicate that the diameter of the titanium powder was not changed by the IPP treatment, and the Ti(C, N) layer grew toward the center of the titanium particles via the diffusion of carbon and nitrogen.
To investigate the effects of IPP treatment conditions on the formation of a Ti(C, N) layer, titanium plates and powder were embedded in a 4:6:3 (volume ratio) mixture of iron, graphite, and alumina powders, and then held at 1073–1373 K in a nitrogen flow. The main results were as follows.
(1) The Ti(C, N) layer was observed in titanium plates subjected to IPP treatment at more than 1173 K, and the layer became uniform and thick with increasing temperature. At 1373 K, a Ti(C, N) layer more than 30 μm thick formed on the titanium surface.
(2) The surface hardness of titanium increased with the heating temperature because the hardness was strongly affected by the formation of the Ti(C, N) layer.
(3) Increasing the nitrogen flow rate increased the surface hardness of titanium modified at 1273 K.
(4) The thickness of the Ti(C, N) layer increased with the holding time. The diameter of the spherical titanium powder was not changed by the IPP treatment at 1273 K, indicating that the Ti(C, N) layer grew toward the inside of titanium via the diffusion of carbon and nitrogen.
Y. Morizono greatly appreciates support from JSPS KAKENHI Grant Number 15K06509 and JSPS Program for Advancing Strategic International Networks to Accelerate the Circulation of Talented Researchers.
