Microstructural study of nitrogen-implanted Ti6Al4V alloy

Microstructural study of nitrogen-implanted Ti6Al4V alloy

951 Nuclear instruments and Methods in Physics Research B59/60 (1991) 951.-956 forth-Holland Microstructural study of nitrogen-implanted Ti--6AlL4...

755KB Sizes 0 Downloads 6 Views


Nuclear instruments and Methods in Physics Research B59/60 (1991) 951.-956 forth-Holland


study of nitrogen-implanted



X. Qiu ‘, R.A. Dodd, J.R. Conrad, A. Chen and F.J. Worzala

A comprehensive study of the microstructural effects of nitrogen ion implantation on Ti-6A1-4V alloy has been carried out. The work was promoted by the fact that the corrosive wear resistance of Ti-6Ai-4V alloy in an environment similar to that encountered in medical applications can be considerably improved by nitrogen ion implantation using the plasma source ion implantation technique. The alloy was implanted with 50 keV nitrogen ions to different doses ranging from 9~10’~ to 1~10’~ atoms/m’. (‘omposition distribution within the implanted region was characterized by Auger electron spectroscopy. Microstructural studies were haxed on transmission electron microscopy. Better understanding of the wear/corrosion behavior is obtained from the study of the microstructures resulting from implantation and provides guidelines for further improvement of the wear/corrosion properties of Ti-6Al-4V alloy.

1. Introduction The use of titanium ailoys, particularly Ti-6AI-4V, is now quite widespread for applications where a combination of low density and adequate mechanical properties is desirable. Surgical prostheses represent such an application, but carry the additional requirements of wear resistance and corrosion resistance (strictly wear/corrosion resistance since wear and corrosion are expected to operate synergistically) because prosthetic deterioration from these causes is likely to cause tissue irritation or infection due to wear debris and corrosion products. In the absence of wear, titanium has excellent corro$ion resistance in most environments, including saline conditions similar to body fluids. Problems arise, however, when wear locally removes the passive film. exposing highly active unpassivated metal to the environment :md forming a passive/active cell. If wear is intermitrent, repassivation will prevent much corrosion occurring, but in various prosthetic applications relative movement. and therefore wear of mating surfaces, is bubstantiaily continuous and wear/corrosion becomes excessive. To overcome these problems, nitrogen implantation has been used to improve the characteristics of Ti-6Al3V surgical prostheses [l--12]. Certainly, titanium nitride IS both wear and corrosion resistant, but it must be recollected that if the nitrogen ion fluence is inadequate, the surface of the implanted component is incompletely converted to TIN. For example, in an earlier paper [13]

’ 1500 Johnson Drive, University Madison. WI 53706. USA. 0168-583X/91/$03.50

of Wisconsin-Madison.

we have shown that a fluence of 3 X 102’ atoms/m’ implanted at 50 keV gives a final concentration of about 14 at.% nitrogen at the surface - clearly insufficient to result in a complete conversion of the surface to TIN. The nitrogen concentration at a depth of 500-600 A rises considerably, but this fact does not impact on the wear/corrosion properties of the alloy until considerable degradation of the component has occurred. In fact, the writers have examined the near-surface microstructures of a number of Ti-6Al-4V commercial prosthetic devices, and concluded that the nitrogen fluences used were considerably below that required for optimum wear/corrosion resistance. Under such circumstances the performance of the device may be inadequate, and it may even be concluded that nitrogen ion implantation of titanium alloys is an unreliable way of obtaining the desired properties. In the present work. we have therefore studied the near-surface microstructure of nitrogen-impIanted Ti6Al-4V using various nitrogen fluences. and made some observations of the beneficial effects of higher fluences on wear/corrosion behavior.

2. Experimental


Ti-6Al-4V sheet specimens 2 mm in thickness. and with a ~crostructure consisting of a plus intergranular l3. were mechanically polished followed by ultrasonic cleaning. and then implanted with 50 keV nitrogen ions to various fluences by the plasma source ion implantation method. The principles of this method and the details of the implantation parameters have been described previously [ 131.

1n 1991 - Elsetier Science Publishers B.V. (North-Holland)


X. Qiu et 01. / Microstructural study


The effects of four different nitrogen ion fluences on the alloy microstructure and wear resistance were studied. These were 9 X 102’, 3 x lo*‘, 6.6 x lo*’ and 1 X lo*’ atoms/m*. Nitrogen depth profiles in the implanted specimens were determined by Auger spectroscopy, ~crostructures were studied by transmission electron microscopy, and wear resistance was measured by a pin-on-disc method [13]. 3. Results and discussion The nitrogen depth profiles for specimens implanted to the various nitrogen fluences are shown in fig. 1. The results of Auger spectroscopy were compared with

Fig. 2. Transmission

of N-implunted Ti-6A 1-4 V aUo,y

g ? .z 2




Depth Fig. 1. Nitrogen



profiles showing fluence.

electron micrographs showing fluence dependence of the microstructure of nitrogen fIuences of (a) 9~ 102”, (b) 3 X 102’, (c) 6.6X 102’, and (d) 1 X 1O22 atoms/&.

the effect of increasing




calculated nitrogen profiles using the dynamic Monte Carlo code TAMIX 1141.The agreement was reasonable although some discrepancies exist, notably a shift of the peak nitrogen concentration to greater depths for higher fluences. This is accounted for partly by increased radiati#n~enhanced diffusion of nitrogen at higher fluences due to increased radiation damage (dpaf, and partly by sputter-induced surface roughness (an artifact af AES depth profiling) which also increases with fluence. The ion gun used in the Auger studies was at 30’ to the specimen surface, and so the effect of surface roughness un the depth profile is quite significant. The specimens for transmission electron microscopy


were prepared by a backt~n~~n~ electropolishing technique, Since a typical electron transparent TEM foil is likely to be up to around 1500 A in thickness, the TEM results would span the nitrogen distributions in fig. I. The observed microst~ctur~ in fact do so as seen in fig. 2. In fig. Za, corresponding to a nitrogen ion fluence of 9 X 10” atoms/m*, the dark striations represent the early stages of titanium nitride precipitation* Many of the striations are roughly parallel to one another, indicating the likelihood of coherency between nitride particle and cu-Ti matrix, and so the image size is probably determined by coherency strains. In fact, the diffraction pattern in fig. 3a which corresponds to the micrograph



d Fig. 3. Stlected


diffraction patterns corresponding to electron micrographs in fig. 2; (a} 9 x 10”. (bj 3 x (d) 1 x 10z2atoms/mz.



(c) 6.6 x IO*‘,



in fig. 2a is due entirely to the cu-Ti phase. In fig. 2b, which shows microstructural effects at a fluence of 3 X lo*’ atoms/m2, the striations are now longer, more developed, and clearly roughly parallel to one another. The diffraction pattern, fig. 3b, still shows a strong cu-Ti single crystal spot pattern, but there are in addition a number of semicontinuous rings, some of which are due to titanium nitride and some to a-Ti. It has been noted previously [13.15] that nitrogen implantation of Ti6Al-4V gives rise to discontinuous a-Ti rings, and this has been attributed to a grain refining effect. The dark, poorly resolved regions in fig. 2b remain a source of speculation. Stereo pairs showed them to be near the outer foil surface, and energy dispersive X-ray analysis showed that they contained a higher vanadium content than the matrix. Since vanadium stabilizes the p phase,

these dark areas may be p formed as a result of irradiation-induced segregation of vanadium to the surface 01 the alloy. Our Auger analysis did not give a clear indication of the vanadium segregation because of the high oxygen content on the surface of titanium. The vanadium peak (already very small) is superimposed on the oxygen peak. Rut Auger srudy of argon irradiated Ti-6A1-4%’ has shown strong segregation of vanadjum to the surface 1161. Since there is no electron diffraction evidence for this suggestion. it remains simply a plausible way of accounting for the observations. At fluences of 6.6 x IO*’ and 1 x 1O22 atoms/m’ the progressive development of a titanium nitride structure with increasing fluence is clearly seen in figs. 2c and 2d. The diffraction patterns in figs. 3c and 3d show features which match the microstructural changes. Thus. the

and [OilO]..,//[~ilfTrW: (a) SADP obtained from Ti-6AL4V Fig. 4. Orientation relationship analysis yielding (~llO),_~i//(O1l)~,, nitrogen implanted to a dose of 6.6 X IO” atoms/m*, containing spots from both ru-Ti and ‘EN, (b) and (c) are computed SADPs of zone [OilO],.,, and zone [O%l],,,, respectively. and (d) is the superimposed spot pattern from (b) and (c) which is identical to (a).

X. Qiu et al. / Microstructural pattern in fig. 3c shows rings/spots which are largely due to titanium nitride, although there is a residual effect from a-Ti. The pattern in fig. 3d is due entirely to titanium nitride and is a substantially continuous ring pattern in contrast to the mostly spotty pattern of fig. 3b. The microstructure appears to account for these differences in the diffraction patterns. Thus in fig. 3c the nitride particles are roughly parallel, suggestive of epitaxial growth, while in fig. 3d they appear to be oriented rather randomly. This transition of a preferred orientation of titanium nitride to a random one evidently irradiation-induced - has not been studied. The parallel nature of the titanium nitride particles at intermediate fluences suggests that they grow with a distinct crystallographic relationship to the a-Ti grains in which they have nucleated, Evidence for this was

study of N-implanted




obtained via eiectron diffraction of specimens implanted at a fluence of 6.6 x 102’ atoms/m’ using a fine electron beam. The diffraction patterns and accompanying analyses are shown in figs. 4 and 5, the results being as follows: fig. 4,

(2110) ._r,//(OII)rIN+

[OiIOl di//[Oi~lw~ fig.5, (~llO).-~i//(O1l)l_iNr IO2231 a.ri//[Oil Ii-m. That is (2110),_ri always appears to be parallel to (O1l)riN, but there is no singular match of crystallographic directions in these planes. In fact, the [OilO],.ri and [0223],, directions in the (?IIO),_ri plane are at an angle of 54.17O with one another.


Fig. 5. Orientation relationship analysis yielding (~llO)~_~~//(Oll)~~~ and [0~23],.,;//[Oil] T,N: (a) SADP obtained from Ti-6Al-4V nitrogen implanted to a dose of 6.6 x lo*’ atoms/m2, containing spots from both a-Ti and TiN, (b) and (c) are computed SADPs of zone [0223],.,, and zone [OillriN, respectively, and (d) is the superimposed spot pattern from (b) and (c) which is identical to (a). VII. METALS / TRIBOLOGY


X. Qru et al. / Microstructura(

The lattice parameters for hexagonal Ti-6Al-4V were determined by X-ray diffraction to be u = 2.92 A and c = 4.67 A. and were used to standardize the electron diffraction patterns. This permitted the lattice parameter of the titanium nitride to be determined. giving 4.44 A for the pattern in fig. 4 and 4.14 A for the pattern in fig. 5. These results are in reasonable agreement with the published value of (I = 4.24 A [17]. Also shown in figs. 4 and 5 are the computed electron diffraction patterns for the (Y phase of Ti--6Al-4V and for cubic TIN using the correct lattice parameters for these phases. An excellent match is obtained when the calculated patterns are superimposed according to the orientation relationships deduced from the actual diffraction patterns. Therefore. the two different possible orientation relationships are confirmed. Finally, although the present research was concentrated on the fluence dependence of the microstructure of nitrogen-implanted Ti-6Al-4V, a few tests were carried out to verify the beneficial effects of higher fluences on wear resistance. For example. pin-on-disc ruby ball wear tests with the conditions: 50 g load, 12 rpm, IO2 revolutions in Hanks solution, showed much better wear behavior for a specimen implanted to 1 X 10z2 atoms/m2 compared with one implanted to 3 X 10” atoms/m”. The ratio of the cross sectional areas of the wear tracks (which were determined by alpha step profilometry) was 4.4 to 1.

4. Conclusions The microstructure of nitrogen-implanted Ti-6Al4V is very fluence sensitive within the range 9 x 102’ to 1 X 10” atoms/m’. Only at the higher end of this range is the surface substantially converted to titanium nitride to a depth of about 1000 A. There is a corresponding effect on wear resistance, and it appears reasonable to suggest that for applications such as surgical prostheses, where wear cannot be tolerated, a minimum fluence of 1 X 102’ atoms/m” at 50 keV energy should be specified. The titanium nitride nucleates and grows epitaxially within the a-Ti crystals, but at higher fluences a randomly oriented arrangement of titanium nitride particles is the end product.

stu& of N-implanted


V aliq

Acknowledgements The authors gratefully acknowledge the support of the National Science Foundation, through Grant No. DMC-8712461. and also thank Paul Fetherston for experimental assistance. References [l] R. Hutchings and W.C. Oliver, Wear 92 (1983) 143. (21 J.M. Williams, G.M. Beardsley, R.A. Buchanan and R.K. Bacon. in: Ion Implantation and Ion Beam Processing of Materials. MRS Symp. Proc.. vol. 27, eds. G.K. Hubler, O.W. Holland, C.R. Clayton and C.W. White (Elsevier, New York, 1984) p. 735. [3] A.J. Perry, Surf. Eng. 3 (1987) 154. [4] S. Saritas. R.P.M. Procter and W.A. Grant, Mater. Sci. Eng. 90 (1987) 297. [5] F.D. Matthews, K.W. Greer and D.L. Armstrong, in: Biomedical Materials. MRS Symp. Proc.. vol. 55, eds. J.M. Williams, M.F. Nichols and W. Zingg (Elsevier, New York, 1986) p. 243. [6] W.C. Oliver, R. Hutchings. J.B. Pethica. E.L. Paradis and A.J. Shuskus, in: Ion- Implantation and Ion Beam Processing of Materials, MRS Symp. Proc. vol. 27, eds. G.K. Hubler, O.W. Holland. C.R. Clayton and C.W. White (Elsevier. New York, 1984) p. 705. [71 P.A. Higham, in: Biomedical Materials. MRS Symp. Proc.. vol. 55. eds. J.M. Williams, M.F. Nichols and W. Zing8 (Elsevier. New York, 1986) p 253. PI J.M. Williams and R.A. Buchanan. Mater. Sci. Eng. 69 (1985) 237. [91 I.L. Singer, in: Ion Implantation and Ion Beam Processing of Materials, MRS Symp. Proc., vol. 27, eds. G.K. Hubler, 0.W. Holland. C.R. Clayton and C.W. White (Elsevier, New York, 1984) p. 585. [lOI P. Siosanshi and R.W. Oliver, in: Biomedical Materials. MRS Symp. Proc.. vol. 55, eds. J.M. Williams, M.F. Nichols and W. Zingg (Elsevier, New York. 1986) p. 237. [ill J. Zhang, X. Zhang, Z. Guo and H. Li, ibid., p. 229. [121 C.J. McHargue, Int. Met. Rev. 31 (2) (1986) 49. 1131X. Qiu, J.R. Conrad, R.A. Dodd and F.J. Worzala, Metall. Trans. A21 (1990) 1663. 1141S.H. Han, G.L. Kulcinski and J.R. Conrad, Nucl. lnstr. and Meth. B45 (1990) 701. USI W.C. Oliver. R. Hutchings and J.B. Pethica. Metall. Trans. Al5 (1984) 2221. [I61 2. Wang, G. Ayrault and H. Wiedersich, J. Nucl. Mat. 108/109 (1982) 331. [I71 W.B. Pearson. Handbook of Lattice Spacings and Structures of Metals (Pergamon, 1958) p. 234.