Materials Science and Engineering, 100 (1988) 193-199
Mechanism of High-temperature Plasma Nitriding of Titanium EDWARD ROLINSKI Institute ~f Materials Scie¢lC6and Engineering, WarsawUniversityof Technolo,~,q',Narbutta 85, 02-524 Warsaw(Poland) (Received April 22, 1987: in revised form September 2, 1987)
1he paper presents the results of studies on plasma nitriding of titanium in nitrogen at 1030 °C. For the nitriding process, fiat specimens were placed on the cathode and, after the treatment had been carried out, the)' were analyzed for differences between the layers produced on their upper and lower surfaces. The nitrided layers have the same thickness on both sides of the specimens. Microhardness and corrosion resistance in 15% H,SO 4 of the layers produced on the upper surfaces are higher than those of the layers formed on the lower surfaces of the specimens. "lhe reason for these differences is" the contamination of the lower layer by carbon. Carbon also causes an increase in the value of the lattice parameter of the TiN nitride. The action of nitrogen ions on the upper surface of the specimens entails sputtering of the surface and removes the contamination by carbon. For this" reason, the nitrided layer produced on the upper su~we contains less impurities and has better properties. 1. Introduction
The co-operation of ionized nitrogen with a titanium cathode has been the subject of several papers recently. Michel and Gantois  have come to the conclusion that lasting surface depassivation, nitrogen implantation and, perhaps, formation of the TiN nitride in the gaseous medium as well as its deposition on the cathode are responsible for the production of the surface layer in plasma nitriding of titanium. Similarly, Bollinger et al.  have concluded that reactive cathodic sputtering of titanium and the deposition, on the cathode, of the TiN nitride formed account for the nitriding reaction. Ming-Biarm Liu et al.  suggest that the nitriding of titanium under direct-current glow-discharge conditions proceeds in two steps: (i) bombardment of the titanium surface with nitrogen ions and NH radicals as well as implantation of these chemical entities into regions close to the surface; (ii) forma0025 - 5416 / 88/ S3.50
tion of a nitride surface layer with enrichment of the surface regions with nitrogen and diffusion of nitrogen into the material. In their opinion, nitriding under glow-discharge conditions proceeds faster than other techniques used. After plasma nitriding of titanium in pure nitrogen, we analyzed the thickness of the nitrided layer which formed on the surfaces A and B of the specimen placed on a titanium cathode, as shown in Fig. 1 . Side A of the specimen was exposed to the action of nitrogen ions; hence, implantation of nitrogen, cathode sputtering of titanium, deposition of the sputtering products (TIN) and chemisorption of nitrogen were likely to occur. In contrast, on side B of the specimen only chemisorption of nitrogen could take place because glow discharge could not occur in the narrow slot between the specimen and the cathode. As the nitrided layers on both sides of the specimen were of equal thickness, we conclude that, in the plasma nitriding of titanium, which shows a great chemical affinity for nitrogen, chemisorption is responsible for the formation of the nitrided layer [51. Rie and Lampe  also found in examining this problem, that the thickness of the layer on both sides of the nitrided specimen is the same; the layer on side A, however, is harder. To elucidate the reasons for the differences observed in the nitriding results on sides A and B of titanium specimens lying on the cathode (Fig. 1),
jNEGATIVE GLOW I'fl, I I JI-I t 7,-~"A PECIMEN
~CATHODE Figl 11 Mode Ot*placing the specimen on the cathode: (A) upper surface of the specimen; (B) lower surface of the specimen. © Elsevier Sequoia/Printed in T he Netherlands
194 high-temperature plasma nitriding of titanium (at 1030°C) was carried out whereby sufficiently thick nitrided layers (thus easily examined) could be produced. Because of the great affinity of titanium for oxygen and carbon, the effect of these elements on the structure and properties of the layers produced was analyzed in detail. 2. M e t h o d s of investigation
Specimens for nitriding (12.5 mm in diameter and 2 mm thick) were cut from a titanium rod of the composition given in Table 1. The specimens were ground on both sides on abrasive paper (gradation 600) and then degreased in acetone in an ultrasonic washer. For the nitriding treatment, nitrogen with a dew point of - 30 °C and oxygen content of 6 ppm was used. No further purification of the gas was needed. 2.2. Apparatus and experimental procedure
The apparatus and nitriding procedures were the same as described previously . Preliminary pressure in the chamber was about 10 -3 mbar. For the nitriding treatment, a pressure of about 6.7 hPa was used at a nitrogen flow of 30 1 h-l. The treatment was carried out at 1030 °C for 9 h; the voltage drop between the electrodes was about 700 V. 2.3. Structural examination
The surfaces of the specimens were observed in a Tesla BS300 scanning electron microscope. Phase analysis of the layers was carried out using a Siemens D-500 diffractometer with a semiconductor Si-Li detector using C u K a radiation. The accuracy of measurements of the lattice parameters of nitrides was _+0.00005 nm. The distribution of oxygen and carbon on the cross-section as well as on the surface of the specimens was determined with a Jeol JXA3A electron probe microanalyzer. The structure of the nitrided layers was examined on cross-section of halved specimens, clamped in metallographic holders. Both sides of the specimens were covered with soft nickel foil, 20 /~m thick, to prevent the edges from spalling during grinding. Metallographic observations were made TABLE 1 Content of impurities in titanium used for investigation
(wt.%) (wt.%) (wt.%) (wt.%)
using a Neophot-2 microscope. Microhardness of the layer on the cross-section was measured by the Hanemann method using a Vickers indenter with a 50 g load. Every point in the diagram (Fig. 8) represents a mean value from six measurements. 2.4. Corrosion tests
Corrosion tests of both sides of the nitrided specimens were made in non-aerated 15% H2SO 4 at 25°C. Corrosion resistance was estimated by analyzing potentiodynamic curves of anodic polarization recorded at a rate of 1 mV s- 1. To determine the course of one polarization curve, measurements were made on three specimens and the results were then averaged. 3. Results of investigation
3.1. Structural examination
Indications of differences in the nature of surfaces A and B were observed even by a visual examination of the nitrided specimens. Surface B was light in colour with a golden lustre, whereas surface A was dark golden. Observations in a scanning electron microscope corroborated these differences. Figure 2 shows the appearance of surfaces A and B of the specimen observed in the scanning microscope. Surface A is characterized by fine-grained irregularities, uniformly distributed on greater undulated areas. In contrast, irregularities on surface B form a network which is probably a reflection of grain boundaries in the substrate. Irregularities of surface A are finer than those of surface B. Phase analysis of specimens was made directly after nitriding as well as on specimens subjected to annealing at 300°C for 12 h in a vacuum-sealed quartz ampoule from which tail gases had been removed using a titanium getter. Anneahng was carried out to eliminate the effect of the difference in stresses on surface A and surface B. Different thermal stresses on surfaces A and B are likely to appear as a result of different conditions of heat abstraction from these surfaces during cooling after nitriding. Phase analysis made it possible to reveal the presence of nitride phases TiN and Ti2N on both sides of the specimens under examination. Lattice parameters of the phases which have been identified are listed in Table 2. Differences in the values of lattice parameters of the TiN nitride on sides A and B of the specimen are quite distinct, and were corroborated using Co K a radiation.
Fig. 2. Scanning electron micrographs of surfaces A and B of a titanium specimen subjected to plasma nitriding at 1t)3t)°C:
(a) 500 × magnification, image of reflected electrons; (b) 1000 x magnification, image of secondary electrons.
Lattice parameters of the phases occurring on surfaces A and B of a titanium specimen after nitriding at 1030 °C and
after nitriding at 1030°C and annealing at 300 °C A'ur]ace
Phase TiN: parameter a After nitriding ',nm)
l l2N: parameter c ~' After nitriding and annealing
AJ?er nitriding (nm)
(nm) A B
Aider nitriding and annealing
'~Only parameter cfrom the 002 peak was determined for the Ti2N phase.
The lattice parameter of the TiN nitride on side A is smaller than that on side B. Similarly, parameter c of the Ti2N nitride is higher on side B. Qualitative differences in diffractometric records are visible in Fig. 3 which shows fragments for surfaces A and B. On both sides of the specimen a pronounced texture was found, as evidenced by the high intensity of reflexes (200) for TiN and (002) for Ti2N.
Annealing of the specimens at 300°C did not entail essential changes in the values of the lattice parameters of nitrides. The differences between sides A and B of the specimen persisted. This suggests that thermal (compressive) stresses, developed on cooling the specimens after nitriding to ambient temperature, are the same on both sides. An analysis of carbon distribution showed that the carbon content on the section and on the sur-
Z Iii Z
40 2 e (degrees)
DISTANCE FROM SURFACE
Fig. 4. Carbon distribution on the cross-section of the nitrided layer on sides A and B of a titanium specimen subjected to plasma nitriding at 1030 °C.
40 2e (degrees)
Fig. 3. Diffraction patterns of surfaces A and B of titanium specimens subjected to nitriding at 1030 °C.
~O Z Ld
"' X •
face is greater on side B than on side A. T h e carbon content in the layer on both sides of the specimen is distinctly higher than that in the substrate. These findings are illustrated in Fig. 4 and Fig. 5 respectively. Examination of the oxygen distribution on the section of the layer on sides A and B did not reveal any increase in the oxygen content. However, surface examination showed that the oxygen level on surface B is higher than that on surface A, as illustrated in Fig. 6. Determination of the nitrogen distribution within the layer was not possible because of the interference of the K a line from nitrogen and Ti, and Tiln lines from titanium. Metallographic examination did not reveal any differences in the structure and thickness of the nitrided layer between sides A and B. T h e nitrided layer is c o m p o s e d of a thin brightly etched zone (about 2 # m thick) of the T i N nitride, a somewhat darker homogeneous zone (about 5 # m thick) of the Ti2N nitride, and a T i a - N solution zone 4 5 - 5 0 # m thick (Fig. 7). Within the solution zone, at a distance of 2 5 - 3 0 # m , a two-phase subzone occurs,
D ISTANCE,mm Fig. 5. Carbon distribution on surfaces A and B of a titanium specimen subjected to plasma nitriding at 1030°C: mean values.
E 1500i.-Z LI_J l-Z
DISTANCE,mm Fig. 6. Oxygen distribution on surfaces A and B of a titanium specimen subjected to plasma nitriding at 1030°C: - mean values.
o >1000. "-1-
w 800 z
Fig. 7. Micrograph of the surface layer produced on titanium by plasma nitriding at 1(130°C: 500 x magnification: interference contrast.
< 600 -r
probably formed during cooling. This subzone may contain Ti~N precipitate in a T i c t - N groundmass, but direct evidence for this is lacking. Microhardness measured on the cross-section of the specimen shows slight but distinct differences between sides A and B of the specimen, the values found on side A being higher. A n exception to this is the hardness at a distance of 7/~m from the surface where it is higher on side B. The hardness distribution is shown in Fig. 8. 3.2. Corrosion tests T h e anodic polarization curves, shown in Fig. 9, indicate that the corrosion resistance of titanium in 15% H2SO 4 solution is higher after nitriding. It is shown by an increase in the corrosion potential from about - 0.8 V in the non-nitrided condition to about + 0.1 V in the nitrided condition. T h e wide peak in the active-passive range, occurring in the curve for non-nitrided titanium, vanishes after nitriding, and the current in the passive range becomes lower. There are some slight differences between the curves for surfaces A and B, indicative of a higher corrosion resistance of the nitrided layer on side A of the specimen.
4. Discussion T h e investigations performed have shown that the thickness of the nitrided layer, produced on a titanium specimen lying freely on the cathode, is the
Fig, 8. Hardness distribution on the cross-section of the surface layer on sides A and B of the titanium specimen subjected to plasma nitriding at 1030°C.
W C3 Z DJ EE rY
(.D D O
P 0 T E N T I A L,V(sce)
Fig. 9. Anodic polarization curves in 15% H/SO 4 for nonnitrided titanium and for surfaces A and B of the specimen subjected to plasma nitriding at 1030 °C.
same on both sides. This conforms with the results of investigations carried out earlier on titanium [4, 5], titanium alloys  and a molybdenum-titanium alloy .
198 If the nitriding atmosphere is contaminated by carbon, the properties of layers produced on sides A and B of the specimen are not equal. This is shown by different hardness profiles in the layers as well as by a different corrosion resistance of layers of the same thickness. The presence of carbon in the nitriding atmosphere can be ascribed to the chemical composition of tail gases in the vacuum system which contain oil vapours from the pumps. The effect of contamination by carbon on the growth of the TiN layer was observed even in implantation treatment of titanium with nitrogen in a nigh vacuum . As the removal of carbon is difficult, its effect on the growth of the layer in the process of plasma nitriding should be taken into account. The experiments revealed differences between the values of the parameters of the TiN nitride on sides A and B. Carbon increases the TiN lattice parameter [9, 10] but the role of oxygen is as yet not clearly understood. Ehrlich  is of the opinion that oxygen, up to 40% TiO in TiN, does not entail a decrease of the lattice parameter of the TiN nitride. According to Gurov et al. , however, oxygen initially decreases, then increases and, from a content of 10wt.%O 2, markedly increases the lattice parameter of this nitride. The role of carbon in increasing the lattice parameter of the TiN nitride on side B of the specimen seems indisputable in our experiment. The values of the lattice parameter of the TiN nitride can also have a higher value, owing to the occurrence of residual stresses in the layer . Because of an increased content of carbon on side B of the specimen, its corrosion resistance is somewhat lower than on side A. This is connected with the corrosion resistance of carbonitrides being lower than that of nitrides . The values of parameter c of the Ti2N nitride are higher than those given in the literature, probably because of a small amount of carbon dissolved in the nitride. However, comparative data on this subject are lacking. Parameter c can also have a higher value, owing to the occurrence of residual stresses. An increased carbon concentration in the diffusion zone T i a - N on side B of the specimen is likely to be the cause of reduced hardness in this zone. The solubility of carbon in titanium is about 2at.% lower than that of nitrogen (about 19at.%) . Its influence on the Young modulus and, consequently, on the hardness of titanium is also less . Ion bombardment of surface A of the specimen leads to a high defect concentration and causes
sputtering. This helps to clean the surface of gases such as carbon monoxide and oxygen which are prone to chemisorption. In contrast, surface B of the specimen does not undergo such cleaning and, for this reason, these gases play an essential part in the growth of the nitrided layer on this side of the specimen. The action of nitrogen on both sides of the specimen is the same because the chemisorption of this element on titanium constitutes a non-activated process which, consequently, is controlled by its thermodynamics and not kinetics . The role of the glow discharge restricts itself to surface cleaning by cathode sputtering.
5. Conclusions (i) Plasma nitriding of titanium, performed by the method proposed here, made it possible to separate the effects of the action of ionized and thermally excited nitrogen on titanium from the effects caused by thermally excited nitrogen under the same pressure, temperature and time conditions. (ii) The role of nitrogen ions in plasma nitriding of titanium was determined in this way. Bombardment of the specimen (cathode) surface by nitrogen ions leads to desorption and cathode sputtering of atoms of impurities, i.e. of carbon and oxygen. For that reason, the content of these elements in the plasma nitrided layer is lower than that occurring in the thermally nitrided layer. (iii) As a result, the lattice parameter of the TiN nitride in the plasma nitrided layer has a lower value than that found in the thermally nitrided layer. (iv) The hardness of the plasma nitrided layer and its corrosion resistance in 15% H2SO 4 are higher than the values for the thermally nitrided layer.
Acknowledgments The author thanks Associate Professor T. Karpifiski for valuable discussions and for the stimulus to carry on the investigations, Associate Professor J. Flis and Dr. J. Mafikowski from the IChF PAN for their help with performing corrosion tests, and Mrs. M. Burzyfiska-Szyszko, M.Eng., for preparing specimens for examination. The help of Professor S. Wojciechowski in the form of financial support from the CPBR 2.4 funds is gratefully acknowledged.
References 1 H. Michel and M. Gantois, Mere. Sci. Rev. Memll., 64 (1972) 739. 2 H. Bollinger, B. Biicken and R. Wilberg. Metalloher.[laeche, 31 i1977) 329. 3 Ming-Biann Liu, D. M. Gruen, A. R. Krauss. A. H. Reis, Jr., and S. W. Peterson, ttigh Temp. Sci., 10 (1978) 53. 4 E. Rolifiski, I'h.D, thesis, Warsaw University of Technology, 1978. 5 T. Karpiriski and E. Roliriski, in I'roc. 8th ('onf
('elostatne dni tel?elneho spracovania, Bratislava, 1978', Czechoslovak Scientific and Technical Society, CVSVTS, p. 27. 6 K. T. Ric and Th. Lampe, Mater. Sci. En~., 69 (1985) 473. 7 h. Rolinski, Sur/i Eng., 2 (1986) 35.
8 M. G. Hay and J. W. Martin, J. Less-Common Met., 18 c 1969) 27. 9 K. Hohmuth and B. Rauschenbach, Mater. Sci. Eng., 69 11985)489. 1(I G.V. Samsonov, Nitridy, Naukova Dumka, Kiev, 1969. I1 P. Ehrlich, l)ie Naturwissenschal?en, Springer-Verlag, Berlin. 1954. p. I 17. 12 S. V. Gurov. \ . N. Troickiy, V. 1. Torbov and V. V. Kireyko, Neorg, Mater., 18 ii 982) 1733. 13 E.J. Mittemeijer. tlaerterei. Tech. Mitt., J,'~ ( 1981 ', 57. 14 E. Etchessahar, J. R Bars, J. Debuigne, A. P. Lamaze and P. Champin, ]~lanium Science and Technolog), l'roc. 5th Int. (onJZ T#attium, Munich, S~7~tetnher 10 14, 198'4. Vol. 3. p. 1423. 15 R. 1. Jaffee, l'rog. Met. Phys., 711968) 65. 16 H. Conrad, Prog. Mater. Sci., 26 ( 1981 ) 136. 17 F. C. Tompkins, ('hernisorption o] (;ases on Metals, Academic Press. London, 1978.