- Email: [email protected]
Content of hydrogen in boron-, carbon-, nitrogen-, oxygen-, fluorine- and neon-implanted titanium
Surface and Coatings Technology 103–104 (1998) 299–303
Content of hydrogen in boron-, carbon-, nitrogen-, oxygen-, fluorineand neon-implanted titanium M. Soltani-Farshi a,b,*, H. Baumann a, D. Ru¨ck b, E. Richter c, U. Kreissig c, K. Bethge a a Institut fu¨r Kernphysik, Universita¨t Frankfurt, August-Euler-Str. 6, 60486 Frankfurt, Germany b Gesellschaft fu¨r Schwerionenforschung, Planckstr. 1, 64291 Darmstadt, Germany c Forschungszentrum Rossendorf, Postfach 510119, 01314 Dresden, Germany
Abstract Commercially available pure titanium contains a significant amount of hydrogen. Ion implantation into pure titanium samples distorts the lattice causing the hydrogen to diffuse into the implantation region and can thus affect the formation of phases (defects, vacancies). In this work, the effects of hydrogen content were examined for ion-implanted titanium samples. The implantation of boron, carbon, nitrogen, oxygen and neon influences the hydrogen content in the implanted region of the titanium, but we found no effect after fluorine implantation. These effects were investigated as a function of the ion fluence (1×1016 through 1×1018 ions cm−2). The concentration of the implanted elements was analysed with the (p, c)-nuclear reaction as well as with non-Rutherford backscattering. The accumulated hydrogen in the implanted layer was profiled using the 15N-technique. If the concentration depth distribution of the implanted element exceeds a certain value, binary phases of these elements with titanium are detected by grazing incidence X-ray diffraction. In the region of the new phase, the previously accumulated hydrogen almost disappears. This effect gives additional information about phase formation. © 1998 Elsevier Science S.A. Keywords: GIXRD; Hydrogen; Ion implantation; Titanium
1. Introduction Titanium has a strong chemical affinity and can absorb and store large amounts of hydrogen [1–3], which causes embrittlement of the material. In aqueous solutions of strong acids, entry of hydrogen into metals such as titanium takes place as a result of aggressive corrosion. In gaseous hydrogen environments, direct absorption of hydrogen through the dissociation of hydrogen molecules into metals can occur. To prevent that migration, the surfaces of the metals can be implanted by ions, such as oxygen and especially fluorine, that create a diffusion barrier [4,5]. The effect of nitrogen implantation on hydrogen accumulation in titanium was reported by Sugizaki et al. [4], Neu et al. [6,7] and Soltani-Farshi et al. [8]. Hydrogen entry can be inhibited by controlling its diffusion into the bulk. Ion implantation affects the solubility and mobility of hydrogen in the surface region. The implanted layers present an effective barrier to * Corresponding author: Tel: +49 69 798 24257; Fax: +49 69 798 24212; e-mail: [email protected] 0257-8972/98/$19.00 © 1998 Elsevier Science S.A. All rights reserved. PII S 02 5 7 -8 9 7 2 ( 9 8 ) 0 0 42 7 - 7
hydrogen migration, depending on the implanted elements used. Large amounts of hydrogen can be trapped in nitrogen-implanted titanium. When the fluences of the implanted elements exceed certain solubility limits in titanium, the precipitation of borides, carbides, nitrides, and oxides [9–11] is caused. Such phase formation in the implanted region prevents the further inward migration of hydrogen. The process of nitrogen and hydrogen diffusion into the metal under mechanical stress has been investigated. For optimal tribomechanical conditions, a fairly high dose has to be implanted. Hydrogen defect interactions have many important technological implications [12]. In fusion reactors, the plasma fueling and tritium balance are greatly influenced by recycling of hydrogen isotopes from the first wall [13], partly controlled by hydrogen trapping at irradiation damage. Ideas concerning a hydrogen-based economy have further focused attention on the ability of some metals and alloys to store hydrogen incorporated into the lattice [14]. In this paper, hydrogen accumulation under high vacuum conditions is investigated for ion-implanted titanium at fluences that are typical for
300
M. Soltani-Farshi et al. / Surface and Coatings Technology 103–104 (1998) 299–303
ion implantation conditions in tribomechanical applications. To measure concentration profiles of the relevant elements, ion-beam analysis has been used: Nuclear Reaction Analysis (NRA) for hydrogen, boron, nitrogen, fluorine and neon, non-Rutherford Backscattering Spectrometry (n-RBS ) for oxygen and carbon. The phase formation and crystal structure of the near surface region have been investigated by means of Grazing Incident X-ray Diffraction (GIXRD).
2. Experimental 2.1. Sample preparation and implantation Commercially available pure titanium was used as sample material for the B-, C-, N-, O-, F- and Ne implantation. The surface of the samples (sample thickness 1 mm, diameter 20 mm) was mechanically polished to a mirror finish using 3-mm diamond paste in the final polishing step. In order to achieve a clean surface and a near-surface region with less hydrogen concentration, the surfaces of the samples were treated for 5 s with a mixture of HF and HNO acids (1:4) before implant3 ation [15]. The ion implantations were performed at the implanter at FZR ( Forschungszentrum Rossendorf ) and GSI (Gesellschaft fu¨r Schwerionenforschung). The ion flux was 20 mA cm−2. The samples were mounted on a water-cooled sample holder to avoid heating and to maintain them at approximately room temperature. During implantation, the total pressure was 8×10−5 Pa in the implantation chamber. Ion energies of 100, 130, 150, 165, 175 and 185 keV were used as implantation energies to obtain nearly the same projected range for the implanted 11B-, 12C-, 15N-, 16O-, 19F- and 22Ne ions in titanium (e.g. nitrogen in titanium: R =242 nm for an implantation energy of 150 keV ). p The range of ion fluence used was 1×1016–1×1018 ions cm−2. After implantation, the concentration versus depth distributions of hydrogen and the implanted elements were analysed using NRA and RBS. 2.2. Hydrogen determination by NRA The nuclear reaction analysis was performed at the 7-MV Van de Graaff accelerator at the Institut fu¨r Kernphysik. The hydrogen in the samples was profiled using the 15N profiling technique. The nuclear reaction 1H(15N, ac)12C has a very sharp resonance (C =1.8 keV ) at an energy of E =6.396 MeV [16,17]. res res The hydrogen depth distribution was obtained by measuring the yield of c-rays (E =3–5 MeV ) as a function c of the 15N energy. The emitted c-rays were measured with two 8◊×4◊ NaI( Tl ) detectors. No alteration of the
hydrogen distribution caused by the analysing 15N-beam was observed. The uncertainty of the measured hydrogen concentration, which mainly originates in the uncertainties of hydrogen calibration (NH Cl 4 standard ) and stopping power values, is about 8%. Stopping power data were taken from [8,18]. 2.3. Element determination by NRA and RBS The 11B, 15N, 19F and 22Ne depth profiles of the implanted titanium were measured using the resonant nuclear reaction 11B(1H, c)12C, 15N(1H, ac)12C, 19F(1H, ac)16O and 22Ne(1H, ac)23Na at a proton resonance energy of 163 keV [19], 429 keV [20], 484 keV [21] and 854 keV [22] at the 2.5-MV Van de Graaff accelerator. The proton beam was normally incident on the target surface. Glassy carbon samples implanted with doses of 15N, 22Ne and measured by RBS, served as calibration standards. For boron and fluorine calibration, TiB and CaF were used. The depth profiles were 2 2 measured by varying the proton energy in increments of 1 keV. The uncertainty of the concentration values in the samples was estimated to be less than 10%. To calculate the depth scales in nanometres (top axis of the graphs), the atomic density of pure titanium (N=5.66×1022 at. cm−3) was used. After carbon and oxygen implantation, n-RBS using 7.6-MeV 4He ions was used to obtain the depth profiles and areal densities of the implanted C and O atoms. A surface-barrier detector located at an angle of 171° was used to detect the backscattered a-particles. The solid angle of the detector was 0.34×10−3 sr, and a tilt angle of 70° of the sample with respect to the direction of the analysing beam was used to separate the oxide layer at the surface from the implanted oxygen. 2.4. X-ray diffractometry The phases formed by ion implantation were characterised by GIXRD at the Technische Hochschule Darmstadt. The equipment used for the X-ray diffraction measurement was a H-2H goniometer (Siemens D 5000), with an option for grazing incident angle and scintillation counter for peak detection. The radiation used was Cu Ka. The dispersion of the X-ray beam is limited to 0.3° by a diaphragm system and Soller slits. To obtain measuring times as short as possible and welldetected interference signals, measurements with a focused Bragg–Brentano configuration and fixed incident angle H=2° were made with a step scan of 0.02° and a step time of 5 s. The penetration depth of the radiation (1/e intensity) ranged to 380 nm, depending on the attenuation coefficient of the phase, the geometry of the X-ray beam and sample orientation. The peaks were identified using the peak-fitting program [23] and the powder diffraction file (Joint Committee on powder
M. Soltani-Farshi et al. / Surface and Coatings Technology 103–104 (1998) 299–303
diffraction standards, 1601 Park Lane, Swarthmore, PA 1081, USA).
3. Results In the near surface region of a polished titanium sample, an average hydrogen concentration of 0.3 at.% for depths exceeding 10 nm was measured. After implantation of B-, C-, N-, O-, F- and Ne ions with a fluence of 3×1017 ions cm−2, a considerable accumulation of hydrogen up to 1 at.% for fluorine, 3 at.% for oxygen, 8 at.% for boron and nitrogen, 10 at.% for neon and 30 at.% for carbon implanted titanium was observed in the implanted region. Depth profiles for 15N and H obtained by NRA are shown in Fig. 1 for nitrogenimplanted titanium samples with three different fluences ranging from 3×1017 to 1×1018 ions cm−2. At B-, C-, N- and Ne concentrations up to 20 at.%, both the hydrogen and B, C, N and Ne profiles are nearly Gaussian with their peak maxima lying in the same depth with respect to the sample surface. With increasing B, C or N concentration, a saddle-shaped hydrogen profile occurs. When the implanted concentration
301
exceeds a certain value, the two peaks are separated by a plateau with a hydrogen concentration less than 1 at.%. The hydrogen concentration of oxygenimplanted titanium has a value of about 2–3 at.%. At higher oxygen fluences (above 3×1017 ions cm−2), multiple maxima appear in the hydrogen depth profiles. The strong dependence of the accumulated hydrogen (dose) on the fluence and the implanted ion species are shown in Fig. 2. The hydrogen dose was calculated from the area of the hydrogen depth distribution taken from the NRA measurements. The dashed line represents the hydrogen content (dose) of an unimplanted titanium sample. The maximum hydrogen dose of each implanted element except neon is reached at a fluence of about 4×1017 ions cm−2. Phase formation occurring after ion implantation was monitored by GIXRD measurements. Typically, for thin films, peaks have only very low intensities at high Bragg angles [24]. For this reason, only the 2H region between 30 and 60° was chosen for phase identification. Fig. 3 shows the measured diffractograms of the Ti–X samples ( X=B, C, N, O and F ), all implanted at fluences of 1×1018 at. cm−2. The observed phases of the B, C, N, O and F implanted samples are also summarised in Table 1 for different fluences. The broadening of the peaks can be explained by the insertion of free implanted atoms in the crystal and therefore to a non-uniform lattice constant.
4. Discussion
Fig. 1. Nitrogen and hydrogen depth profiles in titanium implanted with 150-keV 15N ions and different fluences.
The content of accumulated hydrogen depends on the fluence of the implanted elements for all ion species used. At fluences up to 3×1017 ions cm−2, the distribution of the implanted elements (B, C and N ) and the accumulated hydrogen are nearly Gaussian. For ion fluences above 8×1017 ions cm−2, the formation of a saddle-shaped hydrogen distribution with a hydrogen plateau concentration of less than 1 at.% can be attributed to the phase formation of borides, carbides or nitrides. These phases form an ordered structure as shown by the results of GIXRD measurements in Fig. 3 for an ion-implanted sample with a fluence of 1×1018 ions cm−2. However, we found no phase formation for fluences lower than 4×1017 ions cm−2, which may be due to the detection limit of the GIXRD measurements. The structures of the formed phases make less space available for hydrogen atoms. After implantation of fluorine, neither a Gaussian nor a saddle-shaped hydrogen profile could be observed. The average hydrogen concentration is about 1 at.%. This low hydrogen content (observed also for oxygen) agrees with the fact that the phase formation starts at low fluorine and oxygen concentrations of a few atomic per cent in the titanium.
302
M. Soltani-Farshi et al. / Surface and Coatings Technology 103–104 (1998) 299–303
Fig. 2. Relation between the dose of the implanted elements (boron, carbon, nitrogen, oxygen, fluorine and neon) and the accumulated hydrogen dose in the region of 20–600 nm. Table 1 Compounds formed for ion-implanted titanium at different fluences
Fig. 3. Diffractograms of high fluence B-, C-, N-, O- and F-implanted titanium samples (fluence=1×1018 ions cm−2).
Ion
E (keV )
Fluence (1018 ions cm−2)
Phases indicated by GIXRD
Boron Carbon Nitrogen Oxygen Fluorine
100 130 150 165 175
0.4–1 0.4–1 0.4–1 0.3–1 0.3–1
TiB, TiB 2 TiC TiN , TiN, Ti N 0.3 2 TiO, TiO , Ti O 2 2 TiF 3
In order to determine whether there is enough space for hydrogen in previous fluorine-implanted titanium samples, hydrogen implantation with a fluence of 5×1017 ions cm−2 was carried out. First results indicate that there is no solubility of hydrogen in fluorineimplanted titanium. Neon implantation to fluences exceeding 5×1017 Ne ions cm−2 leads to the formation of blisters at about 22 at.% with the consequence that the neon concentration remains constant. The accumulated hydrogen in the implanted layer starts to vanish with the bursting of neon bubbles. For these neon-implanted titanium samples, only a Gaussian-shaped hydrogen profile was observed. The accumulated hydrogen in the implanted layer may be caused by defects and neon precipitation produced by the ion implantation as reported for Ne-implanted tantalum and molybdenum [25,26 ]. The relation between the dose of the implanted elements and the accumulated hydrogen in the implanted layer in Fig. 2 shows a very low hydrogen content for fluorine and oxygen compared to B, C, N and Ne implanted samples. The carbon-implanted samples show three times more hydrogen content compared to boron
M. Soltani-Farshi et al. / Surface and Coatings Technology 103–104 (1998) 299–303
and nitrogen-implanted samples. At a carbon concentration of about 20 at.%, carbon acts as a more effective hydrogen getter than boron and nitrogen. The larger amount of hydrogen for carbon-implanted titanium can be explained by the size of carbon precipitates that can lead to a higher internal stress or strain and to expansion in the Ti-lattice spacing ( TiC#20 nm compared to nitride or boride precipitates #10 nm [27]). In the region of the new phase (fluences above 4×1017 21C ions cm−2), the previously accumulated hydrogen continuously decreases. It was concluded that hydrogen accumulation in B-, C- or N-implanted titanium depends on the behaviour of phase formation of these elements in the titanium lattice. Hydrogen atoms diffuse to the dislocations or defects that were produced by ion implantation. Taking into account the results of XRD measurements, it could be shown that phase formation and precipitation, especially for B, C and N, are detectable by measuring the depth distribution of the accumulated hydrogen. The effect of hydrogen accumulation is an important process that can be a valuable aid in choosing implantation conditions for improving the tribomechanical properties of titanium and its alloys.
5. Conclusions In the present work, the distribution of hydrogen in titanium and its behaviour after B-, C-, N-, O-, F-, and Ne implantation have been shown. A correlation between phase formation and accumulated hydrogen was found. A significant hydrogen accumulation (up to 30 at.% for carbon implanted titanium) was observed in the implanted layer measured by NRA. At B-, C- and N concentrations above 20 at.%, the hydrogen content decreases with increasing phase precipitation and reaches a value of less than 0.6 at.% in the formed boride, carbide and nitride layer. No noticeable hydrogen effect was observed for oxygen and fluorine-implanted titanium. Further Positron Annihilation Spectroscopy (PAS) experiments are planned to understand possible effects that the implantation-produced defect structure has on hydrogen accumulation.
303
Acknowledgement The authors would like to thank Mohammad Ghafari, Technical University of Darmstadt for assistance at the GIXRD measurements. This work was financially supported by GSI.
References [1] H. Baumann, Th. Lenz, F. Rauch, Mater. Sci. Eng. 69 (1985) 421. [2] B. Hoffmann, H. Baumann, F. Rauch, Nucl. Instrum. Meth. B 15 (1986) 361. [3] B. Hoffmann, H. Baumann, F. Rauch, K. Bethge, Nucl. Instrum. Meth. B 36 (1989) 157. [4] J. Sugizaki, T. Furuya, H. Satoh, Nucl. Instrum. Meth. B 59–60 (1991) 722. [5] H. Ferber, G.K. Wolf, Mater. Sci. Eng. 90 (1987) 213. [6 ] K. Neu, H. Baumann, N. Angert, D. Ru¨ck, K. Bethge, Nucl. Instrum. Meth. B 89 (1994) 379. [7] K. Neu, PhD thesis, Institute for Nuclear Physics, Frankfurt am Main, Germany, 1995. [8] M. Soltani-Farshi, H. Baumann, D. Ru¨ck, K. Bethge, Nucl. Instrum. Meth. B 127–128 (1997) 787. [9] J.C. Pivin, J. Mater. Sci. 25 (1990) 2743. [10] P. Zheng, J.C. Pivin, M.O. Ruault, Europhys. Lett. 8 (1988) 689. [11] J.C. Pivin, P. Zheng, M.O. Ruault, Phil. Mag. Lett. 59 (1989) 25. [12] J.P. Hirth, H. Johnson, Corrosion 32 (1976) 3. [13] G.M. McCracken, P.E. Stott, Nucl. Fusion 19 (1979) 889. [14] R. Wiswall, in: G. Alefeld, J. Vo¨lkl ( Eds.), Hydrogen in Metals II, Springer, Berlin, 1978, Vol. 29, p. 201. [15] B. Hoffmann, PhD thesis, Institute for Nuclear Physics, Frankfurt/Main, Germany, 1987. [16 ] W.A. Lanford, Nucl. Instrum. Meth. 149 (1978) 1. [17] F. Rauch, H. Baumann, U. Behrens, K. Bethge, U. Schwenk, B. Streb, W. Strohl, in: G. Kraft ( Ed.), Analysis of Non-Metals, de Gruyter, Berlin, 1981, p. 151. [18] J.F. Ziegler, SRIM, Version (97.06). [19] D.S. Craig, W.G. Cross, R.G. Jarvis, Phys. Rev. 103 (1956) 1414. [20] A. Schardt, W.A. Fowler, C.C. Lauritsen, Phys. Rev. 86 (1952) 527. [21] F. Bumiller, H.H. Staub, H.E. Weaver, Helv. Phys. Acta 29 (1956) 83. [22] K.J. Brostro¨m, T. Huus, J. Koch, Nature 160 (1947) 498. [23] E.J. Sonneveld, J.W. Visser, J. Appl. Cryst. 8 (1975) 1. [24] H. Jehn, G. Reiners, N. Siegel, DIN Fachbericht 39, Charakterisierung du¨nner Schichten, Beuth, Berlin, 1993. [25] J. Keinonen, V. Karttunen, J. Ra¨isa¨nen, F.-J. Bergmeister, A. Luukkainen, P. Tikkanen, Phys. Rev. B 34 (1986) 8983. [26 ] J. Keinonen, V. Karttunen, Phys. Rev. B 37 (1988) 8440. [27] J.C. Pivin, P. Zheng, M.O. Ruault, Mater. Sci. Eng. A 115 (1989) 83.
Content of hydrogen in boron-, carbon-, nitrogen-, oxygen-, fluorineand neon-implanted titanium M. Soltani-Farshi a,b,*, H. Baumann a, D. Ru¨ck b, E. Richter c, U. Kreissig c, K. Bethge a a Institut fu¨r Kernphysik, Universita¨t Frankfurt, August-Euler-Str. 6, 60486 Frankfurt, Germany b Gesellschaft fu¨r Schwerionenforschung, Planckstr. 1, 64291 Darmstadt, Germany c Forschungszentrum Rossendorf, Postfach 510119, 01314 Dresden, Germany
Abstract Commercially available pure titanium contains a significant amount of hydrogen. Ion implantation into pure titanium samples distorts the lattice causing the hydrogen to diffuse into the implantation region and can thus affect the formation of phases (defects, vacancies). In this work, the effects of hydrogen content were examined for ion-implanted titanium samples. The implantation of boron, carbon, nitrogen, oxygen and neon influences the hydrogen content in the implanted region of the titanium, but we found no effect after fluorine implantation. These effects were investigated as a function of the ion fluence (1×1016 through 1×1018 ions cm−2). The concentration of the implanted elements was analysed with the (p, c)-nuclear reaction as well as with non-Rutherford backscattering. The accumulated hydrogen in the implanted layer was profiled using the 15N-technique. If the concentration depth distribution of the implanted element exceeds a certain value, binary phases of these elements with titanium are detected by grazing incidence X-ray diffraction. In the region of the new phase, the previously accumulated hydrogen almost disappears. This effect gives additional information about phase formation. © 1998 Elsevier Science S.A. Keywords: GIXRD; Hydrogen; Ion implantation; Titanium
1. Introduction Titanium has a strong chemical affinity and can absorb and store large amounts of hydrogen [1–3], which causes embrittlement of the material. In aqueous solutions of strong acids, entry of hydrogen into metals such as titanium takes place as a result of aggressive corrosion. In gaseous hydrogen environments, direct absorption of hydrogen through the dissociation of hydrogen molecules into metals can occur. To prevent that migration, the surfaces of the metals can be implanted by ions, such as oxygen and especially fluorine, that create a diffusion barrier [4,5]. The effect of nitrogen implantation on hydrogen accumulation in titanium was reported by Sugizaki et al. [4], Neu et al. [6,7] and Soltani-Farshi et al. [8]. Hydrogen entry can be inhibited by controlling its diffusion into the bulk. Ion implantation affects the solubility and mobility of hydrogen in the surface region. The implanted layers present an effective barrier to * Corresponding author: Tel: +49 69 798 24257; Fax: +49 69 798 24212; e-mail: [email protected] 0257-8972/98/$19.00 © 1998 Elsevier Science S.A. All rights reserved. PII S 02 5 7 -8 9 7 2 ( 9 8 ) 0 0 42 7 - 7
hydrogen migration, depending on the implanted elements used. Large amounts of hydrogen can be trapped in nitrogen-implanted titanium. When the fluences of the implanted elements exceed certain solubility limits in titanium, the precipitation of borides, carbides, nitrides, and oxides [9–11] is caused. Such phase formation in the implanted region prevents the further inward migration of hydrogen. The process of nitrogen and hydrogen diffusion into the metal under mechanical stress has been investigated. For optimal tribomechanical conditions, a fairly high dose has to be implanted. Hydrogen defect interactions have many important technological implications [12]. In fusion reactors, the plasma fueling and tritium balance are greatly influenced by recycling of hydrogen isotopes from the first wall [13], partly controlled by hydrogen trapping at irradiation damage. Ideas concerning a hydrogen-based economy have further focused attention on the ability of some metals and alloys to store hydrogen incorporated into the lattice [14]. In this paper, hydrogen accumulation under high vacuum conditions is investigated for ion-implanted titanium at fluences that are typical for
300
M. Soltani-Farshi et al. / Surface and Coatings Technology 103–104 (1998) 299–303
ion implantation conditions in tribomechanical applications. To measure concentration profiles of the relevant elements, ion-beam analysis has been used: Nuclear Reaction Analysis (NRA) for hydrogen, boron, nitrogen, fluorine and neon, non-Rutherford Backscattering Spectrometry (n-RBS ) for oxygen and carbon. The phase formation and crystal structure of the near surface region have been investigated by means of Grazing Incident X-ray Diffraction (GIXRD).
2. Experimental 2.1. Sample preparation and implantation Commercially available pure titanium was used as sample material for the B-, C-, N-, O-, F- and Ne implantation. The surface of the samples (sample thickness 1 mm, diameter 20 mm) was mechanically polished to a mirror finish using 3-mm diamond paste in the final polishing step. In order to achieve a clean surface and a near-surface region with less hydrogen concentration, the surfaces of the samples were treated for 5 s with a mixture of HF and HNO acids (1:4) before implant3 ation [15]. The ion implantations were performed at the implanter at FZR ( Forschungszentrum Rossendorf ) and GSI (Gesellschaft fu¨r Schwerionenforschung). The ion flux was 20 mA cm−2. The samples were mounted on a water-cooled sample holder to avoid heating and to maintain them at approximately room temperature. During implantation, the total pressure was 8×10−5 Pa in the implantation chamber. Ion energies of 100, 130, 150, 165, 175 and 185 keV were used as implantation energies to obtain nearly the same projected range for the implanted 11B-, 12C-, 15N-, 16O-, 19F- and 22Ne ions in titanium (e.g. nitrogen in titanium: R =242 nm for an implantation energy of 150 keV ). p The range of ion fluence used was 1×1016–1×1018 ions cm−2. After implantation, the concentration versus depth distributions of hydrogen and the implanted elements were analysed using NRA and RBS. 2.2. Hydrogen determination by NRA The nuclear reaction analysis was performed at the 7-MV Van de Graaff accelerator at the Institut fu¨r Kernphysik. The hydrogen in the samples was profiled using the 15N profiling technique. The nuclear reaction 1H(15N, ac)12C has a very sharp resonance (C =1.8 keV ) at an energy of E =6.396 MeV [16,17]. res res The hydrogen depth distribution was obtained by measuring the yield of c-rays (E =3–5 MeV ) as a function c of the 15N energy. The emitted c-rays were measured with two 8◊×4◊ NaI( Tl ) detectors. No alteration of the
hydrogen distribution caused by the analysing 15N-beam was observed. The uncertainty of the measured hydrogen concentration, which mainly originates in the uncertainties of hydrogen calibration (NH Cl 4 standard ) and stopping power values, is about 8%. Stopping power data were taken from [8,18]. 2.3. Element determination by NRA and RBS The 11B, 15N, 19F and 22Ne depth profiles of the implanted titanium were measured using the resonant nuclear reaction 11B(1H, c)12C, 15N(1H, ac)12C, 19F(1H, ac)16O and 22Ne(1H, ac)23Na at a proton resonance energy of 163 keV [19], 429 keV [20], 484 keV [21] and 854 keV [22] at the 2.5-MV Van de Graaff accelerator. The proton beam was normally incident on the target surface. Glassy carbon samples implanted with doses of 15N, 22Ne and measured by RBS, served as calibration standards. For boron and fluorine calibration, TiB and CaF were used. The depth profiles were 2 2 measured by varying the proton energy in increments of 1 keV. The uncertainty of the concentration values in the samples was estimated to be less than 10%. To calculate the depth scales in nanometres (top axis of the graphs), the atomic density of pure titanium (N=5.66×1022 at. cm−3) was used. After carbon and oxygen implantation, n-RBS using 7.6-MeV 4He ions was used to obtain the depth profiles and areal densities of the implanted C and O atoms. A surface-barrier detector located at an angle of 171° was used to detect the backscattered a-particles. The solid angle of the detector was 0.34×10−3 sr, and a tilt angle of 70° of the sample with respect to the direction of the analysing beam was used to separate the oxide layer at the surface from the implanted oxygen. 2.4. X-ray diffractometry The phases formed by ion implantation were characterised by GIXRD at the Technische Hochschule Darmstadt. The equipment used for the X-ray diffraction measurement was a H-2H goniometer (Siemens D 5000), with an option for grazing incident angle and scintillation counter for peak detection. The radiation used was Cu Ka. The dispersion of the X-ray beam is limited to 0.3° by a diaphragm system and Soller slits. To obtain measuring times as short as possible and welldetected interference signals, measurements with a focused Bragg–Brentano configuration and fixed incident angle H=2° were made with a step scan of 0.02° and a step time of 5 s. The penetration depth of the radiation (1/e intensity) ranged to 380 nm, depending on the attenuation coefficient of the phase, the geometry of the X-ray beam and sample orientation. The peaks were identified using the peak-fitting program [23] and the powder diffraction file (Joint Committee on powder
M. Soltani-Farshi et al. / Surface and Coatings Technology 103–104 (1998) 299–303
diffraction standards, 1601 Park Lane, Swarthmore, PA 1081, USA).
3. Results In the near surface region of a polished titanium sample, an average hydrogen concentration of 0.3 at.% for depths exceeding 10 nm was measured. After implantation of B-, C-, N-, O-, F- and Ne ions with a fluence of 3×1017 ions cm−2, a considerable accumulation of hydrogen up to 1 at.% for fluorine, 3 at.% for oxygen, 8 at.% for boron and nitrogen, 10 at.% for neon and 30 at.% for carbon implanted titanium was observed in the implanted region. Depth profiles for 15N and H obtained by NRA are shown in Fig. 1 for nitrogenimplanted titanium samples with three different fluences ranging from 3×1017 to 1×1018 ions cm−2. At B-, C-, N- and Ne concentrations up to 20 at.%, both the hydrogen and B, C, N and Ne profiles are nearly Gaussian with their peak maxima lying in the same depth with respect to the sample surface. With increasing B, C or N concentration, a saddle-shaped hydrogen profile occurs. When the implanted concentration
301
exceeds a certain value, the two peaks are separated by a plateau with a hydrogen concentration less than 1 at.%. The hydrogen concentration of oxygenimplanted titanium has a value of about 2–3 at.%. At higher oxygen fluences (above 3×1017 ions cm−2), multiple maxima appear in the hydrogen depth profiles. The strong dependence of the accumulated hydrogen (dose) on the fluence and the implanted ion species are shown in Fig. 2. The hydrogen dose was calculated from the area of the hydrogen depth distribution taken from the NRA measurements. The dashed line represents the hydrogen content (dose) of an unimplanted titanium sample. The maximum hydrogen dose of each implanted element except neon is reached at a fluence of about 4×1017 ions cm−2. Phase formation occurring after ion implantation was monitored by GIXRD measurements. Typically, for thin films, peaks have only very low intensities at high Bragg angles [24]. For this reason, only the 2H region between 30 and 60° was chosen for phase identification. Fig. 3 shows the measured diffractograms of the Ti–X samples ( X=B, C, N, O and F ), all implanted at fluences of 1×1018 at. cm−2. The observed phases of the B, C, N, O and F implanted samples are also summarised in Table 1 for different fluences. The broadening of the peaks can be explained by the insertion of free implanted atoms in the crystal and therefore to a non-uniform lattice constant.
4. Discussion
Fig. 1. Nitrogen and hydrogen depth profiles in titanium implanted with 150-keV 15N ions and different fluences.
The content of accumulated hydrogen depends on the fluence of the implanted elements for all ion species used. At fluences up to 3×1017 ions cm−2, the distribution of the implanted elements (B, C and N ) and the accumulated hydrogen are nearly Gaussian. For ion fluences above 8×1017 ions cm−2, the formation of a saddle-shaped hydrogen distribution with a hydrogen plateau concentration of less than 1 at.% can be attributed to the phase formation of borides, carbides or nitrides. These phases form an ordered structure as shown by the results of GIXRD measurements in Fig. 3 for an ion-implanted sample with a fluence of 1×1018 ions cm−2. However, we found no phase formation for fluences lower than 4×1017 ions cm−2, which may be due to the detection limit of the GIXRD measurements. The structures of the formed phases make less space available for hydrogen atoms. After implantation of fluorine, neither a Gaussian nor a saddle-shaped hydrogen profile could be observed. The average hydrogen concentration is about 1 at.%. This low hydrogen content (observed also for oxygen) agrees with the fact that the phase formation starts at low fluorine and oxygen concentrations of a few atomic per cent in the titanium.
302
M. Soltani-Farshi et al. / Surface and Coatings Technology 103–104 (1998) 299–303
Fig. 2. Relation between the dose of the implanted elements (boron, carbon, nitrogen, oxygen, fluorine and neon) and the accumulated hydrogen dose in the region of 20–600 nm. Table 1 Compounds formed for ion-implanted titanium at different fluences
Fig. 3. Diffractograms of high fluence B-, C-, N-, O- and F-implanted titanium samples (fluence=1×1018 ions cm−2).
Ion
E (keV )
Fluence (1018 ions cm−2)
Phases indicated by GIXRD
Boron Carbon Nitrogen Oxygen Fluorine
100 130 150 165 175
0.4–1 0.4–1 0.4–1 0.3–1 0.3–1
TiB, TiB 2 TiC TiN , TiN, Ti N 0.3 2 TiO, TiO , Ti O 2 2 TiF 3
In order to determine whether there is enough space for hydrogen in previous fluorine-implanted titanium samples, hydrogen implantation with a fluence of 5×1017 ions cm−2 was carried out. First results indicate that there is no solubility of hydrogen in fluorineimplanted titanium. Neon implantation to fluences exceeding 5×1017 Ne ions cm−2 leads to the formation of blisters at about 22 at.% with the consequence that the neon concentration remains constant. The accumulated hydrogen in the implanted layer starts to vanish with the bursting of neon bubbles. For these neon-implanted titanium samples, only a Gaussian-shaped hydrogen profile was observed. The accumulated hydrogen in the implanted layer may be caused by defects and neon precipitation produced by the ion implantation as reported for Ne-implanted tantalum and molybdenum [25,26 ]. The relation between the dose of the implanted elements and the accumulated hydrogen in the implanted layer in Fig. 2 shows a very low hydrogen content for fluorine and oxygen compared to B, C, N and Ne implanted samples. The carbon-implanted samples show three times more hydrogen content compared to boron
M. Soltani-Farshi et al. / Surface and Coatings Technology 103–104 (1998) 299–303
and nitrogen-implanted samples. At a carbon concentration of about 20 at.%, carbon acts as a more effective hydrogen getter than boron and nitrogen. The larger amount of hydrogen for carbon-implanted titanium can be explained by the size of carbon precipitates that can lead to a higher internal stress or strain and to expansion in the Ti-lattice spacing ( TiC#20 nm compared to nitride or boride precipitates #10 nm [27]). In the region of the new phase (fluences above 4×1017 21C ions cm−2), the previously accumulated hydrogen continuously decreases. It was concluded that hydrogen accumulation in B-, C- or N-implanted titanium depends on the behaviour of phase formation of these elements in the titanium lattice. Hydrogen atoms diffuse to the dislocations or defects that were produced by ion implantation. Taking into account the results of XRD measurements, it could be shown that phase formation and precipitation, especially for B, C and N, are detectable by measuring the depth distribution of the accumulated hydrogen. The effect of hydrogen accumulation is an important process that can be a valuable aid in choosing implantation conditions for improving the tribomechanical properties of titanium and its alloys.
5. Conclusions In the present work, the distribution of hydrogen in titanium and its behaviour after B-, C-, N-, O-, F-, and Ne implantation have been shown. A correlation between phase formation and accumulated hydrogen was found. A significant hydrogen accumulation (up to 30 at.% for carbon implanted titanium) was observed in the implanted layer measured by NRA. At B-, C- and N concentrations above 20 at.%, the hydrogen content decreases with increasing phase precipitation and reaches a value of less than 0.6 at.% in the formed boride, carbide and nitride layer. No noticeable hydrogen effect was observed for oxygen and fluorine-implanted titanium. Further Positron Annihilation Spectroscopy (PAS) experiments are planned to understand possible effects that the implantation-produced defect structure has on hydrogen accumulation.
303
Acknowledgement The authors would like to thank Mohammad Ghafari, Technical University of Darmstadt for assistance at the GIXRD measurements. This work was financially supported by GSI.
References [1] H. Baumann, Th. Lenz, F. Rauch, Mater. Sci. Eng. 69 (1985) 421. [2] B. Hoffmann, H. Baumann, F. Rauch, Nucl. Instrum. Meth. B 15 (1986) 361. [3] B. Hoffmann, H. Baumann, F. Rauch, K. Bethge, Nucl. Instrum. Meth. B 36 (1989) 157. [4] J. Sugizaki, T. Furuya, H. Satoh, Nucl. Instrum. Meth. B 59–60 (1991) 722. [5] H. Ferber, G.K. Wolf, Mater. Sci. Eng. 90 (1987) 213. [6 ] K. Neu, H. Baumann, N. Angert, D. Ru¨ck, K. Bethge, Nucl. Instrum. Meth. B 89 (1994) 379. [7] K. Neu, PhD thesis, Institute for Nuclear Physics, Frankfurt am Main, Germany, 1995. [8] M. Soltani-Farshi, H. Baumann, D. Ru¨ck, K. Bethge, Nucl. Instrum. Meth. B 127–128 (1997) 787. [9] J.C. Pivin, J. Mater. Sci. 25 (1990) 2743. [10] P. Zheng, J.C. Pivin, M.O. Ruault, Europhys. Lett. 8 (1988) 689. [11] J.C. Pivin, P. Zheng, M.O. Ruault, Phil. Mag. Lett. 59 (1989) 25. [12] J.P. Hirth, H. Johnson, Corrosion 32 (1976) 3. [13] G.M. McCracken, P.E. Stott, Nucl. Fusion 19 (1979) 889. [14] R. Wiswall, in: G. Alefeld, J. Vo¨lkl ( Eds.), Hydrogen in Metals II, Springer, Berlin, 1978, Vol. 29, p. 201. [15] B. Hoffmann, PhD thesis, Institute for Nuclear Physics, Frankfurt/Main, Germany, 1987. [16 ] W.A. Lanford, Nucl. Instrum. Meth. 149 (1978) 1. [17] F. Rauch, H. Baumann, U. Behrens, K. Bethge, U. Schwenk, B. Streb, W. Strohl, in: G. Kraft ( Ed.), Analysis of Non-Metals, de Gruyter, Berlin, 1981, p. 151. [18] J.F. Ziegler, SRIM, Version (97.06). [19] D.S. Craig, W.G. Cross, R.G. Jarvis, Phys. Rev. 103 (1956) 1414. [20] A. Schardt, W.A. Fowler, C.C. Lauritsen, Phys. Rev. 86 (1952) 527. [21] F. Bumiller, H.H. Staub, H.E. Weaver, Helv. Phys. Acta 29 (1956) 83. [22] K.J. Brostro¨m, T. Huus, J. Koch, Nature 160 (1947) 498. [23] E.J. Sonneveld, J.W. Visser, J. Appl. Cryst. 8 (1975) 1. [24] H. Jehn, G. Reiners, N. Siegel, DIN Fachbericht 39, Charakterisierung du¨nner Schichten, Beuth, Berlin, 1993. [25] J. Keinonen, V. Karttunen, J. Ra¨isa¨nen, F.-J. Bergmeister, A. Luukkainen, P. Tikkanen, Phys. Rev. B 34 (1986) 8983. [26 ] J. Keinonen, V. Karttunen, Phys. Rev. B 37 (1988) 8440. [27] J.C. Pivin, P. Zheng, M.O. Ruault, Mater. Sci. Eng. A 115 (1989) 83.