- Email: [email protected]
Nitrogen-ion-implantation-induced phase formation and texture in titanium
Materials" Science and Engineering, A 151 (1992) L9-L 13
k9
Letter
Nitrogen-ion-implantation-induced phase formation and texture in titanium Bernd Rauschenbach a and Kurt Helming b "Technical University of Hamburg, Dept. of Semiconductor Technology, Ei/3endorferStrafle 42, 2100 Hamburg 90 (FRG) hTechnical University (Tausthal, lnstituteJor Metal Physics, Gro/3erflruch 23, 3392 Clausthal-Zellerfield (FRG) (Received October 8, 1991 ; in revised form October 29, 1991 )
Abstract High-fluence nitrogen ion implantation into titanium results in the formation of titanium nitride precipitation. The f.c.c, titanium nitride phase was formed preferentially. Electron diffraction measurements reveal an implantation-induced texture of nitride precipitations. The most intensive texture component, the ( 11 1) plane, is always perpendicular to the direction of implantation. Results support the assumption that the effect of implantationinduced texture is caused by the process of seed selection through ion channelling.
1. Introduction It is well known that high-fluence ion implantation is able to improve the surface properties of metals and alloys. To optimize this effect, we have studied the structure of implantation-induced phases and the alignment effect of ion implantation on crystallographic orientation. Recently, it was found that the ion bombardment has a pronounced alignment effect on the crystallographic orientation of thin films [1-5]. In our earlier paper [6], we have determined a preferred orientation of TiN (f.c.c.) crystallites formed after nitrogen ion implantation by electron diffraction [2] and have calculated the orientation distribution function (ODF) from highly incomplete pole figures. In this letter a systematic study by transmission high-energy electron diffraction (THEED) of induced texture is described.
2. Experimental conditions Polycrystalline layers of titanium were deposited on titanium, NaCl or carbon in high vacuum during
electron beam evaporation. The thickness of the titanium layers was about 110 nm. The implantation of nitrogen ions was carried out with 50 keV ions at room temperature in a vacuum better than 1 × 10 4 Pa. Ion currents were maintained below 5 ~ A cm 2 and the ion fluence was between 1 × 10 j~ and 1 × 10Is ions cm 2. The low ion current density was chosen in order to avoid heating effects during implantation. After implantation, the layer structure was investigated by high-voltage transmission electron microscopy and by T H E E D in combination with an inner goniometer. Planar sections of titanium layers on titanium were thinned for high-resolution electron microscopic studies by grinding, dimpling and ion milling from the non-implanted side until perforation. Ion milling was carried out in a cold stage and with low argon ion beam conditions (5 keV, 0.2 mA gun current) so as to minimize potential heating and ion bombardment damage to the sample. The implantation-induced phases were not influenced by argon ion milling at low temperature because changes of morphology were not observed by variation of the ion milling conditions. High resolution electron microscopy was carried out in a 200 keV microscope equipped with a high-resolution pole piece (point-to-point resolution, about 0.2 nm).
3. Principle of pole figure determination The anisotropy of physical properties of crystalline materials depends mainly on the texture, i.e. on the ODF of the crystallites. The ODF may be reproduced from pole figures. In general, pole figures of bulk material can be determined by X-ray or neutron diffraction. For the study of texture of small areas (some square micrometres) or thin layers (less than 1/zm thickness), electron diffraction must be used; nevertheless the measurement of the electron intensity and its interpretation are very difficult. The measurement of pole figures is based on the selected area diffraction (SAD) pattern. Figure 1 illustrates the tilting of the sample (step by step, 2.5 ° or 5 °) to obtain the pole figure and the electron beam arrangement. The pole figures depend on the direction of the scattering vector Y= (8, ¢) with regard to the sample with the normal N. Because the Bragg angle is very small as measured by T H E E D (much less than 1 ), 0 is given by the angle of tilting f2 = 90 ° - t~. Since the angle of tilting ~ may only Elsevier Sequoia
L 10
B. Rauschenbach and K. Helming
/
Nitrogen-ion-implantation-induced phase formation and texture in titanium
determination of texture in a thin film has been presented [6, 7].
/ 201
sample
<<1 ~<,
Fig. 1. Schematic representation of the angle arrangement by electron diffraction on the crystal plane hi of a tilted sample: 0~, Bragg angle; Y,, scattering vector; #, azimuth angle; 0, angle of tilting; N, sample normal; ~ = 900 - ~.
D F
Fig. 2. BF, S A D and D F micrographs of titanium implanted with 50 keV nitrogen ions to a fluence of 5 × 10 ~7 ions c m - 2.
vary in the range 30°--. < 0 4 90 °, the transmission pole figures are extremely incomplete. Therefore a special reproduction method was used [6], based on a tomographic algorithm. Recently, the theoretical basis of
4. Results
The high-fluence nitrogen ion implantation in titanium forms titanium nitride precipitates for fluences of 5 × 1016 ions cm -2 or more [7]. Figure 2 shows bright field (BF) and dark field (DF) micrographs of titanium implanted with 5 x 1017 N ÷ c m -2 at room temperature. Depending on the fluence, the precipitates were identified as the f.c.c. 6-TIN x and the tetragonal e-Ti2N phases (for details see refs. 2 and 9). The f.c.c, titanium nitride phase was formed preferentially. High-magnification lattice imaging (Fig. 3) obtained from a sufficiently thin region of titanium after implantation with 5 x 10 ~6 N ÷ cm -2 shows directly the formed titanium nitride precipitates. In this case, the interplanar distance deduced from the micrograph of the f.c.c, titanium nitride is lower (0.21 nm) than the titanium bulk distance (0.224 nm). The dark regions and observed deformations of the lattice planes in the high resolution micrograph could be due to intrinsic stresses in the implanted layer (for a discussion see ref. 8). The implantation-induced texture of the f.c.c, titanium nitride precipitates was investigated in situ using tilting experiments in the electron microscope. The thin titanium films were implanted perpendicularly, i.e. in the direction of the surface normal N. The diaphragm setting was always chosen so that more than 104 precipitates participated in the diffraction experiment. In order to get information on the texture, the perpendicularly implanted titanium samples were tilted in small selected angular steps (Fig. 1) and the intensities were measured. The notable features of the measurements are the following. (1) The intensities of the diffraction rings and also of" the primary electron beam are decreased with increasing tilt angle fl around an axis perpendicular to the primary beam. (2) By measurements of the intensity as a function of the azimuth angle ~ we found in the case of N parallel to the primary beam a radial symmetrical behaviour (fibre texture, fibre axis ClIN) for all the measurements, i.e. the pole figures were independent of the azimuth angle 4. The presence of a fibre texture was also found by Harper et al. [9] after ion-assisted deposition. (3) The intensity of selected titanium nitride reflections is a function of the tilting angle 0 or Q. For determination of the ODF of the fibre texture the strongest diffraction rings of 6-TIN x were chosen. In Fig. 4 the normalized intensities for the four reflections Ti (10i0), Ti (10i 1), TiN (111), and TiN (200)
B. Rauschenbach and K. Helming
F i g . 3.
I
Nitrogen-ion-implantation-induced phase formation and texture in titanium
High-resolution micrograph of titanium after implantation with
Ti TiN (10"[0) (!11)
t
N+ IMPLANTATION ///~
t
5 x 1 ():~ N ,' +
cm
-
-.
Ti TiN (10{1)(200)
~- [deg]
t
~
k 11
90
I
./7\I 1o
80
70
60
50
44)
30
~IOO L
s0
so ~ g
60
60 ~
Z
o
~0 ~ 20 8o
60~
~
~
~o+~
20 ~ I----
~0
0 ~: 10
0,26 0,2/+ 0,72 0,20 INTERPLANAR SPACINGd[nm] Fig. 4. Diffracted and normalized intensities of selected titanium and TiN reflections.
20
30
t,0
50
60
.aQ. [de9 ] Fig. 5. Intensities of the TiN reflections (200), (111), (220) and (311). In addition, the normalized intensity of the primary electron beam is also shown (- --).
L 12
B. Rauschenbach and K. Helming A ~" (1111
/
Nitrogen-ion-implantation-induced phase formation and texture in titanium
1oo
\
~--
--".~,1111)
~o
2,5
1.0 "~= (2001
---,- ~'i
~= (220)
Fig. 6. Orientation distribution of the normals of lattice planes hi tlN of TiN after implantation perpendicular to the surface. The scale on the right-hand side of the figure indicates the factors of deviation from the uniform distribution of all TiN crystallites.
are illustrated as a function of the tilt angle t~ in 10 ° steps after perpendicular implantation. In Fig. 5 are shown the corresponding incomplete pole figures (30 ° ~
5. Discussion
The preferred orientation or pronounced alignment effect caused by ion implantation has been interpreted in various ways as follows [2, 6]. The effect may be due to the strong compressive strains induced by implantation [5]. These strains can cause a slip mechanism [1] or a polygonization process [8]. On the other hand, the alignment effect was explained by the so-called seed selection process through ion channellig [7, 9, 10], i.e. the implanted ions should destroy a major fraction of grains and preserve only grains of crystallographic orientations where the lattice damages can be minimized by ion channelling (e.g. the [111] direction of titanium nitride precipitates). In order to investigate the causes of the formation of the restricted-fibre textured polycrystalline thin film, we chose an angle of incidence of nitrogen ions so that the ion beam selected growth direction was changed. Nevertheless, the strong compressive strain in the nearsurface range is approximately unchanged. On this basis it should be possible to distinguish between the different causes of the implantation-induced texture. By choosing the angle of ion implantation of 25 ° from the normal incidence, the ion beam should select the main axial channelling direction of incidence, i.e. the
~'=(200) ---~ ~
E=(220)
Fig. 7. Orientation distribution of the normals of lattice planes h~llC of TiN. The implantation direction was 25 ° from the normal incidence.
fibre axis should be changed by about 25 °. If the induced strain dominates, the fibre axis is unchanged and the (111) plane should always be observed perpendicular to the surface. In Fig. 5 the intensities of the four reflections (200), (111), (220) and (311) of TiN are shown as a function of the tilt angle after implantation under an angle 25 ° from the normal incidence (25 ° implantation). In contrast to the intensities of the titanium nitride reflections in Fig. 3, we found modified intensities. For example, the maximum of the (200) peak was found at 55 ° after perpendicular implantation and near 80 ° after 25 ° implantation; i.e. the peak maximum is shifted by about 25 °. On the basis of these measurements, the ODF (see Fig. 7) was determined, where the fibre axis C corresponds to the direction of implantation. Also in this case, the (111) component is the main component (( ] l 1)11c) and the positions of the submaxima, e.g. (220) and (200), are also comparable. This is in good agreement with the results after perpendicular implantation (see Fig. 5). Our results of additional ion channelling measurements show also that the normal of the (111 ) plane is strongly coupled with the direction of implantation [7]. In summary, the results indicate that a controlled modification of the texture of implantation-induced precipitates is possible by high-fluence ion implantation. This allows the possibility of modification of the anisotropic properties of polycrystalline material, for example with respect to an improvement of wear or corrosion resistance.
References 1 0 . Meyer and A. Azzam, Phys. Rev. Lett., 52 (1984) 1629. 2 K. Hohmuth and B. Rauschenbach, Mater. Sci. Eng., 69 (1985)489. 3 S. Ohra and M. lwaki, Nucl. lnstrum. Methods B, 19-20 (1987) 162. 4 G. Marest, in E H. Wfhlbier (ed.), Ion Implantation, Trans Tech Publishing, Aedermannsdorf, 1988, p. 273.
B. Rauschenbach and K. Helming
/
Nitrogen-ion-implantation-induced phase formation and texture in titanium
5 S. Fayeulle and D. Treheux, Nucl. Instrum. Methods' B, 19-20 (1987)216. 6 K. Helmig and B. Rauschenbach, Phys. Status Solidi B, 138 (1986)K1. 7 B. Rauschenbach and K. Helmig, Nucl. lnstrum. Methods B, 42(1989)216.
L 13
8 B. Rauschenbach, J. Mater. Sci., 21 (1985)395. 9 J. M. E. Harper, D. A. Smith, L. S. Yu and J. J. Cuomo, in H. Kurz, G. L. Olsen and J. M. Poate (eds.), Beam-Solid and Phase Transformation, in Mater. Res. Soc. Symp. Proc., 51 (1985)343. 10 K.T.~Y. Kung and R. Reif, J. Appl. Phys., 59 (1986) 2422.
k9
Letter
Nitrogen-ion-implantation-induced phase formation and texture in titanium Bernd Rauschenbach a and Kurt Helming b "Technical University of Hamburg, Dept. of Semiconductor Technology, Ei/3endorferStrafle 42, 2100 Hamburg 90 (FRG) hTechnical University (Tausthal, lnstituteJor Metal Physics, Gro/3erflruch 23, 3392 Clausthal-Zellerfield (FRG) (Received October 8, 1991 ; in revised form October 29, 1991 )
Abstract High-fluence nitrogen ion implantation into titanium results in the formation of titanium nitride precipitation. The f.c.c, titanium nitride phase was formed preferentially. Electron diffraction measurements reveal an implantation-induced texture of nitride precipitations. The most intensive texture component, the ( 11 1) plane, is always perpendicular to the direction of implantation. Results support the assumption that the effect of implantationinduced texture is caused by the process of seed selection through ion channelling.
1. Introduction It is well known that high-fluence ion implantation is able to improve the surface properties of metals and alloys. To optimize this effect, we have studied the structure of implantation-induced phases and the alignment effect of ion implantation on crystallographic orientation. Recently, it was found that the ion bombardment has a pronounced alignment effect on the crystallographic orientation of thin films [1-5]. In our earlier paper [6], we have determined a preferred orientation of TiN (f.c.c.) crystallites formed after nitrogen ion implantation by electron diffraction [2] and have calculated the orientation distribution function (ODF) from highly incomplete pole figures. In this letter a systematic study by transmission high-energy electron diffraction (THEED) of induced texture is described.
2. Experimental conditions Polycrystalline layers of titanium were deposited on titanium, NaCl or carbon in high vacuum during
electron beam evaporation. The thickness of the titanium layers was about 110 nm. The implantation of nitrogen ions was carried out with 50 keV ions at room temperature in a vacuum better than 1 × 10 4 Pa. Ion currents were maintained below 5 ~ A cm 2 and the ion fluence was between 1 × 10 j~ and 1 × 10Is ions cm 2. The low ion current density was chosen in order to avoid heating effects during implantation. After implantation, the layer structure was investigated by high-voltage transmission electron microscopy and by T H E E D in combination with an inner goniometer. Planar sections of titanium layers on titanium were thinned for high-resolution electron microscopic studies by grinding, dimpling and ion milling from the non-implanted side until perforation. Ion milling was carried out in a cold stage and with low argon ion beam conditions (5 keV, 0.2 mA gun current) so as to minimize potential heating and ion bombardment damage to the sample. The implantation-induced phases were not influenced by argon ion milling at low temperature because changes of morphology were not observed by variation of the ion milling conditions. High resolution electron microscopy was carried out in a 200 keV microscope equipped with a high-resolution pole piece (point-to-point resolution, about 0.2 nm).
3. Principle of pole figure determination The anisotropy of physical properties of crystalline materials depends mainly on the texture, i.e. on the ODF of the crystallites. The ODF may be reproduced from pole figures. In general, pole figures of bulk material can be determined by X-ray or neutron diffraction. For the study of texture of small areas (some square micrometres) or thin layers (less than 1/zm thickness), electron diffraction must be used; nevertheless the measurement of the electron intensity and its interpretation are very difficult. The measurement of pole figures is based on the selected area diffraction (SAD) pattern. Figure 1 illustrates the tilting of the sample (step by step, 2.5 ° or 5 °) to obtain the pole figure and the electron beam arrangement. The pole figures depend on the direction of the scattering vector Y= (8, ¢) with regard to the sample with the normal N. Because the Bragg angle is very small as measured by T H E E D (much less than 1 ), 0 is given by the angle of tilting f2 = 90 ° - t~. Since the angle of tilting ~ may only Elsevier Sequoia
L 10
B. Rauschenbach and K. Helming
/
Nitrogen-ion-implantation-induced phase formation and texture in titanium
determination of texture in a thin film has been presented [6, 7].
/ 201
sample
<<1 ~<,
Fig. 1. Schematic representation of the angle arrangement by electron diffraction on the crystal plane hi of a tilted sample: 0~, Bragg angle; Y,, scattering vector; #, azimuth angle; 0, angle of tilting; N, sample normal; ~ = 900 - ~.
D F
Fig. 2. BF, S A D and D F micrographs of titanium implanted with 50 keV nitrogen ions to a fluence of 5 × 10 ~7 ions c m - 2.
vary in the range 30°--. < 0 4 90 °, the transmission pole figures are extremely incomplete. Therefore a special reproduction method was used [6], based on a tomographic algorithm. Recently, the theoretical basis of
4. Results
The high-fluence nitrogen ion implantation in titanium forms titanium nitride precipitates for fluences of 5 × 1016 ions cm -2 or more [7]. Figure 2 shows bright field (BF) and dark field (DF) micrographs of titanium implanted with 5 x 1017 N ÷ c m -2 at room temperature. Depending on the fluence, the precipitates were identified as the f.c.c. 6-TIN x and the tetragonal e-Ti2N phases (for details see refs. 2 and 9). The f.c.c, titanium nitride phase was formed preferentially. High-magnification lattice imaging (Fig. 3) obtained from a sufficiently thin region of titanium after implantation with 5 x 10 ~6 N ÷ cm -2 shows directly the formed titanium nitride precipitates. In this case, the interplanar distance deduced from the micrograph of the f.c.c, titanium nitride is lower (0.21 nm) than the titanium bulk distance (0.224 nm). The dark regions and observed deformations of the lattice planes in the high resolution micrograph could be due to intrinsic stresses in the implanted layer (for a discussion see ref. 8). The implantation-induced texture of the f.c.c, titanium nitride precipitates was investigated in situ using tilting experiments in the electron microscope. The thin titanium films were implanted perpendicularly, i.e. in the direction of the surface normal N. The diaphragm setting was always chosen so that more than 104 precipitates participated in the diffraction experiment. In order to get information on the texture, the perpendicularly implanted titanium samples were tilted in small selected angular steps (Fig. 1) and the intensities were measured. The notable features of the measurements are the following. (1) The intensities of the diffraction rings and also of" the primary electron beam are decreased with increasing tilt angle fl around an axis perpendicular to the primary beam. (2) By measurements of the intensity as a function of the azimuth angle ~ we found in the case of N parallel to the primary beam a radial symmetrical behaviour (fibre texture, fibre axis ClIN) for all the measurements, i.e. the pole figures were independent of the azimuth angle 4. The presence of a fibre texture was also found by Harper et al. [9] after ion-assisted deposition. (3) The intensity of selected titanium nitride reflections is a function of the tilting angle 0 or Q. For determination of the ODF of the fibre texture the strongest diffraction rings of 6-TIN x were chosen. In Fig. 4 the normalized intensities for the four reflections Ti (10i0), Ti (10i 1), TiN (111), and TiN (200)
B. Rauschenbach and K. Helming
F i g . 3.
I
Nitrogen-ion-implantation-induced phase formation and texture in titanium
High-resolution micrograph of titanium after implantation with
Ti TiN (10"[0) (!11)
t
N+ IMPLANTATION ///~
t
5 x 1 ():~ N ,' +
cm
-
-.
Ti TiN (10{1)(200)
~- [deg]
t
~
k 11
90
I
./7\I 1o
80
70
60
50
44)
30
~IOO L
s0
so ~ g
60
60 ~
Z
o
~0 ~ 20 8o
60~
~
~
~o+~
20 ~ I----
~0
0 ~: 10
0,26 0,2/+ 0,72 0,20 INTERPLANAR SPACINGd[nm] Fig. 4. Diffracted and normalized intensities of selected titanium and TiN reflections.
20
30
t,0
50
60
.aQ. [de9 ] Fig. 5. Intensities of the TiN reflections (200), (111), (220) and (311). In addition, the normalized intensity of the primary electron beam is also shown (- --).
L 12
B. Rauschenbach and K. Helming A ~" (1111
/
Nitrogen-ion-implantation-induced phase formation and texture in titanium
1oo
\
~--
--".~,1111)
~o
2,5
1.0 "~= (2001
---,- ~'i
~= (220)
Fig. 6. Orientation distribution of the normals of lattice planes hi tlN of TiN after implantation perpendicular to the surface. The scale on the right-hand side of the figure indicates the factors of deviation from the uniform distribution of all TiN crystallites.
are illustrated as a function of the tilt angle t~ in 10 ° steps after perpendicular implantation. In Fig. 5 are shown the corresponding incomplete pole figures (30 ° ~
5. Discussion
The preferred orientation or pronounced alignment effect caused by ion implantation has been interpreted in various ways as follows [2, 6]. The effect may be due to the strong compressive strains induced by implantation [5]. These strains can cause a slip mechanism [1] or a polygonization process [8]. On the other hand, the alignment effect was explained by the so-called seed selection process through ion channellig [7, 9, 10], i.e. the implanted ions should destroy a major fraction of grains and preserve only grains of crystallographic orientations where the lattice damages can be minimized by ion channelling (e.g. the [111] direction of titanium nitride precipitates). In order to investigate the causes of the formation of the restricted-fibre textured polycrystalline thin film, we chose an angle of incidence of nitrogen ions so that the ion beam selected growth direction was changed. Nevertheless, the strong compressive strain in the nearsurface range is approximately unchanged. On this basis it should be possible to distinguish between the different causes of the implantation-induced texture. By choosing the angle of ion implantation of 25 ° from the normal incidence, the ion beam should select the main axial channelling direction of incidence, i.e. the
~'=(200) ---~ ~
E=(220)
Fig. 7. Orientation distribution of the normals of lattice planes h~llC of TiN. The implantation direction was 25 ° from the normal incidence.
fibre axis should be changed by about 25 °. If the induced strain dominates, the fibre axis is unchanged and the (111) plane should always be observed perpendicular to the surface. In Fig. 5 the intensities of the four reflections (200), (111), (220) and (311) of TiN are shown as a function of the tilt angle after implantation under an angle 25 ° from the normal incidence (25 ° implantation). In contrast to the intensities of the titanium nitride reflections in Fig. 3, we found modified intensities. For example, the maximum of the (200) peak was found at 55 ° after perpendicular implantation and near 80 ° after 25 ° implantation; i.e. the peak maximum is shifted by about 25 °. On the basis of these measurements, the ODF (see Fig. 7) was determined, where the fibre axis C corresponds to the direction of implantation. Also in this case, the (111) component is the main component (( ] l 1)11c) and the positions of the submaxima, e.g. (220) and (200), are also comparable. This is in good agreement with the results after perpendicular implantation (see Fig. 5). Our results of additional ion channelling measurements show also that the normal of the (111 ) plane is strongly coupled with the direction of implantation [7]. In summary, the results indicate that a controlled modification of the texture of implantation-induced precipitates is possible by high-fluence ion implantation. This allows the possibility of modification of the anisotropic properties of polycrystalline material, for example with respect to an improvement of wear or corrosion resistance.
References 1 0 . Meyer and A. Azzam, Phys. Rev. Lett., 52 (1984) 1629. 2 K. Hohmuth and B. Rauschenbach, Mater. Sci. Eng., 69 (1985)489. 3 S. Ohra and M. lwaki, Nucl. lnstrum. Methods B, 19-20 (1987) 162. 4 G. Marest, in E H. Wfhlbier (ed.), Ion Implantation, Trans Tech Publishing, Aedermannsdorf, 1988, p. 273.
B. Rauschenbach and K. Helming
/
Nitrogen-ion-implantation-induced phase formation and texture in titanium
5 S. Fayeulle and D. Treheux, Nucl. Instrum. Methods' B, 19-20 (1987)216. 6 K. Helmig and B. Rauschenbach, Phys. Status Solidi B, 138 (1986)K1. 7 B. Rauschenbach and K. Helmig, Nucl. lnstrum. Methods B, 42(1989)216.
L 13
8 B. Rauschenbach, J. Mater. Sci., 21 (1985)395. 9 J. M. E. Harper, D. A. Smith, L. S. Yu and J. J. Cuomo, in H. Kurz, G. L. Olsen and J. M. Poate (eds.), Beam-Solid and Phase Transformation, in Mater. Res. Soc. Symp. Proc., 51 (1985)343. 10 K.T.~Y. Kung and R. Reif, J. Appl. Phys., 59 (1986) 2422.