- Email: [email protected]
Phase formation after nitrogen implantation into molybdenum
Nuclear Instruments and Methods in Physics Research B 240 (2005) 208–213 www.elsevier.com/locate/nimb
Phase formation after nitrogen implantation into molybdenum D. Manova, Y. Bohne, J.W. Gerlach, S. Ma¨ndl *, H. Neumann, B. Rauschenbach Leibniz-Institut fu¨r Oberfla¨chenmodifizierung, Permoserstr. 15, 04318 Leipzig, Germany Available online 1 August 2005
Abstract High fluence nitrogen ion implantation into molybdenum is compared for low energies of 30 keV and below and high energies of 1 MeV at a total fluence between 0.5 and 2.7 · 1018 at./cm2. For the low energy implantation, a transformation from cubic Mo2N towards tetragonal Mo2N is observed around 580 C with ex situ X-ray diffraction (XRD), whereas high energy implantation leads to the simultaneous formation of tetragonal Mo2N and hexagonal MoN in the same temperature range, as observed with in situ XRD. Additionally, ion beam analysis was employed to measure the local atomic concentrations, which can be used for determining the phase formation threshold and the diffusion constant. 2005 Elsevier B.V. All rights reserved. PACS: 61.72.Ww; 52.77.Dq; 61.10i; 82.80.Yc Keywords: In situ XRD; Ion implantation; Molybdenum; Phase formation
1. Introduction Ion implantation is a widely used process in the semiconductor industry. Modification of metals by ion implantation is a viable method for improved tribological and mechanical surface properties, especially for forming carbides and nitrides, albeit sparsely employed due to process and cost limita-
*
Corresponding author. Tel.: +49 341 235 2944; fax: + 49 341 235 2313. E-mail address: [email protected] (S. Ma¨ndl).
tion. Nevertheless, there are still open questions on fundamental processes during phase formation under high fluence ion implantation. Interstitial transition metal compounds are a very interesting class of materials due to the highly localised d-electrons of the metal and the insertion of small atoms – either B, C, N or O – on interstitial sites without ionic or homopolar bonding thus maintaining the metallic properties [1]. The large solubility field allows a sophisticated phase evolution without major structure changes, apart from those in lattice dimensions. Famous examples are the Fe–N system starting from the e-hcp structure
0168-583X/$ - see front matter 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2005.06.117
D. Manova et al. / Nucl. Instr. and Meth. in Phys. Res. B 240 (2005) 208–213
[2] and the Ti–O system with the transition from hcp-Ti via cubic TiO and the Magne´li phases TinO2n1 toward rutile TiO2 [3,4]. Ion implantation has been used to investigate the phase formation in these systems, especially for transition metal oxides [5–8] mostly at intermediate energies. In this presentation, the phase formation in the Mo–N system is compared for high fluence high energy implantation at 1.0 MeV and low energy implantation in the range from 5 to 30 keV. Using in situ (XRD), besides ex situ XRD and ion beam analysis after the end of the implantation, during the high energy implantation allows a direct view on the formation kinetics.
2. Experiment Pure molybdenum (99.5%) polished to a mirrorlike finish was used as samples. The high energy implantations were performed in an in situ XRD chamber connected to a Tandetron accelerator beamline [9] at 1.0 MeV and a beam current of 1 lA up to a nitrogen fluence of 1.4 · 1018 at./ cm2. Additional heating was employed to maintain a sample temperature of 450 or 550 C. The total implantation time was between 14 and 21 h. A high vacuum system at a base pressure of 2 · 104 Pa and a working pressure of 0.2 Pa was used for the low energy implantations with nitrogen plasma immersion ion implantation (PIII). Using a 40.68 MHz r.f. a nitrogen plasma with a composition of approximately 95% Nþ 2 and 5% N+ is obtained. High voltage pulses between 5 and 30 kV were applied at a constant repetition rate until the process temperature between 330 and 700 C was reached. An additional heating system consisting of IR lamps had to be used to achieve sample temperatures of 600 C and beyond. After the process temperature was reached, the pulse frequency was lowered, thus reducing the average current density, to maintain the temperature. The incident ion fluence was up to 2.7 · 1018 nitrogen at./cm2 for treatments between 30 and 75 min. Different ion energies, fluencies and process temperatures were used to ascertain any difference caused by thermal diffusion, ion flux rate and damage production.
209
The phase composition was studied by X-ray diffraction in Bragg–Brentano geometry with CuKa radiation. A position sensitive detector allowed a scan-time of less than 4 min for a complete XRD spectrum for the in situ measurements in the range 2h = 10–80 with a stepwidth of 0.02. The ex situ measurements were performed using a standard instruments with longer scan times. The elemental concentration distributions were obtained from Rutherford backscattering spectroscopy (RBS) using He2+ ions with an energy of 2.0 MeV and elastic recoil detection analysis (ERDA) with 200 MeV 197Au15+ ions at 19 incident angle and a detector placed at a scattering angle of 37. The data evaluation was performed using RUMP [10] and KONZERD [11] for RBS and ERDA, respectively.
3. Results and discussion The RBS spectra for two samples implanted with 1 MeV nitrogen ions at 450 and 550 C are shown in Fig. 1, together with a simulated spectrum for pure Mo. Using RUMP, the nitrogen depth profiles shown in the inset can be extracted. The maximum concentration is found near a depth of 600 nm, the value calculated with TRIM [12], albeit the width of the profiles is much larger than the 150 nm straggling. The total retained dose, according to the profiles is 5–7 · 1017 at./cm2. However, extreme care must be taken in interpreting these data due to the large depth of the implanted nitrogen and the large mass mismatch between the host and the inserted atoms. The cut-off of the profiles at 1.1 lm is caused by multiple collisions inside the target which are not included in the RUMP simulation. A certain influence of those at lower depths must be present, which is hard to quantify. Nevertheless, the real data should contain a more pronounced and deeper reaching tail towards the right as well as a small shift of the maximum concentration towards larger depths. Additionally, the absolute concentration may be smaller than the calculated values. During the implantation, competing processes of nitrogen insertion and diffusion broadening are occurring
210
D. Manova et al. / Nucl. Instr. and Meth. in Phys. Res. B 240 (2005) 208–213 0.30
24000 Simulation: 100 % Mo
22000
0.25 N/Mo Ratio
20000 18000 Intensity (a.u.)
16000
450 °C 550 °C
0.20 0.15 0.10
14000
0.05
12000
0.00
10000
0
1
2 3 4 5 6 Depth(1018 at.cm-2)
7
8
550 °C 450 °C
8000
1.4 × 1018 at.cm-2, 1 MeV
6000 4000 2000 0 0
200
400
600
800
1000
Channel
Fig. 1. RBS spectra for high energy implantation, together with a simulated spectrum for pure Mo. The inset shows the nitrogen profiles derived from RUMP simulations. (Please note the cut-off at 1.1 lm caused by multiple scattering not included in RUMP.)
40 5 kV. 330 °C. 60 min. 2.7x1018at.cm-2
35
10 kV. 450 °C. 40 min. 2.7x1018at.cm-2 30 kV. 580 °C. 75 min. 1.8x1018at.cm-2
Nitrogen Conc. (at.%)
30
15 kV. 650 °C. 60 min. 2.7x1018at.cm-2 25
15 kV. 700 °C. 30 min. 1.8x1018at.cm-2
20
TRIM. 10 kV N2+
15 10 5 0 0.0
2.5
5.0
7.5
10.0
12.5
15.0
17.5
20.0
Depth (1017 at.cm-2)
Fig. 2. Nitrogen depth profiles measured with ERDA for low energy implantation.
simultaneously and are reducing the maximum concentration. As the concentration is lower at the beginning, a slower diffusion is present, gradually increasing during the implantation time. Regarding the phase formation discussed below, the maximum ratios of N/Mo = 0.15 and 0.25, obtained from RUMP, are upper limits for the initial nucleation phase. A diffusion similar to the high energy implantation is observed for low energy implantation (see
ERDA profiles in Fig. 2) with rather low diffusivities below 580 C and a pronounced evolution of a diffusion tail at higher temperatures. Furthermore, a retained dose much smaller than the incident fluence caused by the proximity of the surface and sputter saturation of the profiles is observed. XRD spectra for the high energy implanted samples are presented in Fig. 3, together with peak positions according the PDF data base [13]. Only selected spectra are shown from the in situ series,
D. Manova et al. / Nucl. Instr. and Meth. in Phys. Res. B 240 (2005) 208–213
211
1000 Mo MoN (H) Mo2N (C) Mo2N (T)
(42-1120) (77-1999) (25-1366) (75-1150)
100
ex-situ, 450 °C
10 ex-situ, 550 °C 18
-2
1.4 × 10 cm , 450 °C 18 -2 1.0 × 10 cm , 450 °C 18 -2 1.4 × 10 cm , 550 °C 18 -2 1.0 × 10 cm , 550 °C
1
20
40
60
80
100
Fig. 3. XRD spectra for 1 MeV nitrogen implantation from in situ measurements at different doses for 450 C and 550 C implantation. Additionally, ex situ spectra obtained afterwards are shown for the final fluence of 1.4 · 1018 cm2.
together with ex situ data taken at higher resolution. The formation of t-Mo2N is indicated even at a fluence as low as 1018 at./cm2 with the main peaks (and the corresponding lattice planes) at 37.7 (1 1 2), 43.0 (2 0 0), 45.3 (0 0 4), 62.5 (2 2 0). No signature of c-Mo2N with peak positions slightly shifted from t-Mo2N can be confirmed in the spectra. Additionally, h-MoN is observed with the (0 0 2) reflection growing at 36.2. Slightly different intensity ratios between
the final in situ and the ex situ data are due to the experimental set-ups. The peak intensities, or more correctly the integral peak areas, are plotted in Fig. 4 as a function of the nitrogen fluence on a logarithmic scale. No saturation with increasing fluence is observed for all investigated peaks. The sample implanted at 450 C shows a lower intensity and a slightly slower intensity increase, with the onset shifted by about a factor of 2. Apparently, the threshold
1
Intensity (a.u.)
450°C 550 °C t-Mo2N (220) / 62.5° t-Mo2N (204) / 64.3° t-Mo2N (112) / 37.7° h-MoN (002) / 36.2° ex-situ measurement
0.1
0.01 0
5
10
Nitrogen flux (10
15 17
-2
cm )
Fig. 4. Integral peak areas as a function of nitrogen flux for selected peaks, as derived from the in situ XRD data.
212
D. Manova et al. / Nucl. Instr. and Meth. in Phys. Res. B 240 (2005) 208–213 10000 Mo (42-1120) Mo2N (C) (25-1366)
1000
MoN (H) (77-1999) Mo2N (T)( 75-1150)
Intensity
100 10 1
700 °C, 30 min
0.1
650 °C, 60 min 600 °C, 60 min
0.01
450 °C, 40 min
1E-3 40
60
80
100
Angle 2θ (°)
Fig. 5. XRD spectra for low energy implanted samples (15 kV pulse voltage) at different temperatures and times.
is given by the sensitivity of the equipment and not by the phase formation kinetics. The MoN (0 0 2) reflection is growing as fast as the t-Mo2N (2 2 0) reflection, whereas the Mo2N (1 1 2) intensity is the fastest growing one. In conjunction with Fig. 1, the lower boundary for the formation of MoN precipitates is assumed to be between the composition MoN0.15 and MoN0.25 which are rather low values. Fig. 5 shows XRD spectra for samples implanted at low energies at different temperatures. Additional peaks, beside those of the base material, at 2H = 43.6 and 62.9 evolve with increasing treatment temperature. Beyond 580 C, a clearly resolved splitting of these peaks occurs, which is ascribed to a phase transition from cubic to tetragonal Mo2N [14]. The most likely explanation involves the lattice distortion due to the Jahn– Teller effect [15], coupled with a volume reduction by 2.2% for Mo2N (cubic: a = 0.42 nm, tetragonal: a = 0.4163 nm, c = 2 · 0.40 nm). No signature of a MoN phase is detected in these spectra. Comparing low energy and high energy implantation, a similar thermally activated diffusion coefficient is observed across the temperature range, with the diffusion length ranging from a few nanometers after 1 h at 330 C to more than 200 nm at 700 C after 30 min. The determination of exact values, complicated by the surface sputtering and the parallel process of implantation and diffusion,
is beyond the limited scope of this paper. However, the phase formation is different for both regimes: the sole formation of c-Mo2N, changing to the formation of t-Mo2N at 580 C and beyond observed at low energies is in contrast to the formation of t-Mo2N and h-MoN at 450 and 550 C, without any indication of c-Mo2N, at high energies. Radiation effects can be most likely excluded as, on the one hand no increased diffusion is observed, and on the other hand, increased pressure, e.g. during the collision cascades should lead to the transformation from t-Mo2N towards c-Mo2N [16], in contradiction to the observation. Another, more likely explanation involves the proximity to the surface for the low energy implantations, albeit ruling out any influence of contaminations on the phase formation as those are not present. The terminated half-space can be a favourable host matrix for slightly different phases than in the bulk as surface reconstruction and relaxation can change the local environment. At higher temperatures, these metastable phases (c-Mo2N in the present case) are again suppressed.
4. Summary and conclusions Nitrogen ion implantation into Mo at elevated temperatures between 300 and 750 C at low and
D. Manova et al. / Nucl. Instr. and Meth. in Phys. Res. B 240 (2005) 208–213
high energies lead to the broadening of the implantation profiles by thermally activated diffusion. Low energy (630 keV) and low temperatures (<580C) favour the formation of c-Mo2N, while t-Mo2N is observed for higher energies or higher temperatures. Additionally, MoN is observed for 1 MeV ion energy starting at a fluence of 5 · 1017 at./cm2 with a local N/Mo ratio of about 0.25. Further investigations are necessary to clarify the thresholds, dynamics and influence of the surface. Acknowledgements W. Assmann, Maier-Leibnitz-Labor, Garching, is acknowledged for providing the ERDA data. W. Mo¨ller, Institut fu¨r Ionenstrahlphysik und Materialforschung, Rossendorf, is acknowledged for providing the high energy beamline implantation facility.
References [1] H.J. Goldschmidt, Interstitial Alloys, Butterworths, London, 1967. [2] K.H. Jack, Acta Crystallogr. 13 (1950) 392.
213
[3] E.S. Bumps, H.D. Kessler, M. Hansen, Trans. Am. Soc. Metals 45 (1953) 1008. [4] J.S. Anderson, B.G. Hyde, J. Phys. Chem. Solids 28 (1967) 1393. [5] T.E. Parker, R. Kelly, J. Phys. Chem. Solids 36 (1975) 377. [6] B. Rauschenbach, A. Kolitsch, K. Hohmuth, Phys. Stat. Sol. A 80 (1983) 471. [7] Y. Okabe, T. Fujihana, M. Iwaki, B.V. Crist, Surf. Coat. Technol. 66 (1994) 384. [8] B. Renner, C. Hammerl, B. Rauschenbach, Nucl. Instr. and Meth. B 160 (2000) 363. [9] Y. Bohne, N. Shevchenko, F. Prokert, J. von Borany, B. Rauschenbach, W. Mo¨ller, Vacuum 76 (2004) 281. [10] L.R. Doolittle, Nucl. Instr. and Meth. B 9 (1985) 344. [11] A. Bergmeier, G. Dollinger, C.M. Frey, Nucl. Instr. and Meth. B 99 (1995) 488. [12] J.F. Ziegler, J.P. Biersack, V. Littmark, The Stopping and Range of Ions in Solids, Pergamon, New York, 1986. [13] Joint Committee for Powder Diffraction Standards, Powder Diffraction File, International Center for Diffraction Data, Park Lane, 1989. [14] S. Ma¨ndl, D. Manova, J.W. Gerlach, W. Assmann, H. Neumann, B. Rauschenbach, Surf. Coat. Technol. 180/181 (2004) 362. [15] T.A. Albright, J.K. Burdett, M.-H. Wnangbo (Eds.), Orbital Interactions in Chemistry, John Wiley and Sons, New York, 1985. [16] O.N. Mryasov, D.L. Novikov, A.J. Freeman, Ferroelectrics 164 (1995) 279.
Phase formation after nitrogen implantation into molybdenum D. Manova, Y. Bohne, J.W. Gerlach, S. Ma¨ndl *, H. Neumann, B. Rauschenbach Leibniz-Institut fu¨r Oberfla¨chenmodifizierung, Permoserstr. 15, 04318 Leipzig, Germany Available online 1 August 2005
Abstract High fluence nitrogen ion implantation into molybdenum is compared for low energies of 30 keV and below and high energies of 1 MeV at a total fluence between 0.5 and 2.7 · 1018 at./cm2. For the low energy implantation, a transformation from cubic Mo2N towards tetragonal Mo2N is observed around 580 C with ex situ X-ray diffraction (XRD), whereas high energy implantation leads to the simultaneous formation of tetragonal Mo2N and hexagonal MoN in the same temperature range, as observed with in situ XRD. Additionally, ion beam analysis was employed to measure the local atomic concentrations, which can be used for determining the phase formation threshold and the diffusion constant. 2005 Elsevier B.V. All rights reserved. PACS: 61.72.Ww; 52.77.Dq; 61.10i; 82.80.Yc Keywords: In situ XRD; Ion implantation; Molybdenum; Phase formation
1. Introduction Ion implantation is a widely used process in the semiconductor industry. Modification of metals by ion implantation is a viable method for improved tribological and mechanical surface properties, especially for forming carbides and nitrides, albeit sparsely employed due to process and cost limita-
*
Corresponding author. Tel.: +49 341 235 2944; fax: + 49 341 235 2313. E-mail address: [email protected] (S. Ma¨ndl).
tion. Nevertheless, there are still open questions on fundamental processes during phase formation under high fluence ion implantation. Interstitial transition metal compounds are a very interesting class of materials due to the highly localised d-electrons of the metal and the insertion of small atoms – either B, C, N or O – on interstitial sites without ionic or homopolar bonding thus maintaining the metallic properties [1]. The large solubility field allows a sophisticated phase evolution without major structure changes, apart from those in lattice dimensions. Famous examples are the Fe–N system starting from the e-hcp structure
0168-583X/$ - see front matter 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2005.06.117
D. Manova et al. / Nucl. Instr. and Meth. in Phys. Res. B 240 (2005) 208–213
[2] and the Ti–O system with the transition from hcp-Ti via cubic TiO and the Magne´li phases TinO2n1 toward rutile TiO2 [3,4]. Ion implantation has been used to investigate the phase formation in these systems, especially for transition metal oxides [5–8] mostly at intermediate energies. In this presentation, the phase formation in the Mo–N system is compared for high fluence high energy implantation at 1.0 MeV and low energy implantation in the range from 5 to 30 keV. Using in situ (XRD), besides ex situ XRD and ion beam analysis after the end of the implantation, during the high energy implantation allows a direct view on the formation kinetics.
2. Experiment Pure molybdenum (99.5%) polished to a mirrorlike finish was used as samples. The high energy implantations were performed in an in situ XRD chamber connected to a Tandetron accelerator beamline [9] at 1.0 MeV and a beam current of 1 lA up to a nitrogen fluence of 1.4 · 1018 at./ cm2. Additional heating was employed to maintain a sample temperature of 450 or 550 C. The total implantation time was between 14 and 21 h. A high vacuum system at a base pressure of 2 · 104 Pa and a working pressure of 0.2 Pa was used for the low energy implantations with nitrogen plasma immersion ion implantation (PIII). Using a 40.68 MHz r.f. a nitrogen plasma with a composition of approximately 95% Nþ 2 and 5% N+ is obtained. High voltage pulses between 5 and 30 kV were applied at a constant repetition rate until the process temperature between 330 and 700 C was reached. An additional heating system consisting of IR lamps had to be used to achieve sample temperatures of 600 C and beyond. After the process temperature was reached, the pulse frequency was lowered, thus reducing the average current density, to maintain the temperature. The incident ion fluence was up to 2.7 · 1018 nitrogen at./cm2 for treatments between 30 and 75 min. Different ion energies, fluencies and process temperatures were used to ascertain any difference caused by thermal diffusion, ion flux rate and damage production.
209
The phase composition was studied by X-ray diffraction in Bragg–Brentano geometry with CuKa radiation. A position sensitive detector allowed a scan-time of less than 4 min for a complete XRD spectrum for the in situ measurements in the range 2h = 10–80 with a stepwidth of 0.02. The ex situ measurements were performed using a standard instruments with longer scan times. The elemental concentration distributions were obtained from Rutherford backscattering spectroscopy (RBS) using He2+ ions with an energy of 2.0 MeV and elastic recoil detection analysis (ERDA) with 200 MeV 197Au15+ ions at 19 incident angle and a detector placed at a scattering angle of 37. The data evaluation was performed using RUMP [10] and KONZERD [11] for RBS and ERDA, respectively.
3. Results and discussion The RBS spectra for two samples implanted with 1 MeV nitrogen ions at 450 and 550 C are shown in Fig. 1, together with a simulated spectrum for pure Mo. Using RUMP, the nitrogen depth profiles shown in the inset can be extracted. The maximum concentration is found near a depth of 600 nm, the value calculated with TRIM [12], albeit the width of the profiles is much larger than the 150 nm straggling. The total retained dose, according to the profiles is 5–7 · 1017 at./cm2. However, extreme care must be taken in interpreting these data due to the large depth of the implanted nitrogen and the large mass mismatch between the host and the inserted atoms. The cut-off of the profiles at 1.1 lm is caused by multiple collisions inside the target which are not included in the RUMP simulation. A certain influence of those at lower depths must be present, which is hard to quantify. Nevertheless, the real data should contain a more pronounced and deeper reaching tail towards the right as well as a small shift of the maximum concentration towards larger depths. Additionally, the absolute concentration may be smaller than the calculated values. During the implantation, competing processes of nitrogen insertion and diffusion broadening are occurring
210
D. Manova et al. / Nucl. Instr. and Meth. in Phys. Res. B 240 (2005) 208–213 0.30
24000 Simulation: 100 % Mo
22000
0.25 N/Mo Ratio
20000 18000 Intensity (a.u.)
16000
450 °C 550 °C
0.20 0.15 0.10
14000
0.05
12000
0.00
10000
0
1
2 3 4 5 6 Depth(1018 at.cm-2)
7
8
550 °C 450 °C
8000
1.4 × 1018 at.cm-2, 1 MeV
6000 4000 2000 0 0
200
400
600
800
1000
Channel
Fig. 1. RBS spectra for high energy implantation, together with a simulated spectrum for pure Mo. The inset shows the nitrogen profiles derived from RUMP simulations. (Please note the cut-off at 1.1 lm caused by multiple scattering not included in RUMP.)
40 5 kV. 330 °C. 60 min. 2.7x1018at.cm-2
35
10 kV. 450 °C. 40 min. 2.7x1018at.cm-2 30 kV. 580 °C. 75 min. 1.8x1018at.cm-2
Nitrogen Conc. (at.%)
30
15 kV. 650 °C. 60 min. 2.7x1018at.cm-2 25
15 kV. 700 °C. 30 min. 1.8x1018at.cm-2
20
TRIM. 10 kV N2+
15 10 5 0 0.0
2.5
5.0
7.5
10.0
12.5
15.0
17.5
20.0
Depth (1017 at.cm-2)
Fig. 2. Nitrogen depth profiles measured with ERDA for low energy implantation.
simultaneously and are reducing the maximum concentration. As the concentration is lower at the beginning, a slower diffusion is present, gradually increasing during the implantation time. Regarding the phase formation discussed below, the maximum ratios of N/Mo = 0.15 and 0.25, obtained from RUMP, are upper limits for the initial nucleation phase. A diffusion similar to the high energy implantation is observed for low energy implantation (see
ERDA profiles in Fig. 2) with rather low diffusivities below 580 C and a pronounced evolution of a diffusion tail at higher temperatures. Furthermore, a retained dose much smaller than the incident fluence caused by the proximity of the surface and sputter saturation of the profiles is observed. XRD spectra for the high energy implanted samples are presented in Fig. 3, together with peak positions according the PDF data base [13]. Only selected spectra are shown from the in situ series,
D. Manova et al. / Nucl. Instr. and Meth. in Phys. Res. B 240 (2005) 208–213
211
1000 Mo MoN (H) Mo2N (C) Mo2N (T)
(42-1120) (77-1999) (25-1366) (75-1150)
100
ex-situ, 450 °C
10 ex-situ, 550 °C 18
-2
1.4 × 10 cm , 450 °C 18 -2 1.0 × 10 cm , 450 °C 18 -2 1.4 × 10 cm , 550 °C 18 -2 1.0 × 10 cm , 550 °C
1
20
40
60
80
100
Fig. 3. XRD spectra for 1 MeV nitrogen implantation from in situ measurements at different doses for 450 C and 550 C implantation. Additionally, ex situ spectra obtained afterwards are shown for the final fluence of 1.4 · 1018 cm2.
together with ex situ data taken at higher resolution. The formation of t-Mo2N is indicated even at a fluence as low as 1018 at./cm2 with the main peaks (and the corresponding lattice planes) at 37.7 (1 1 2), 43.0 (2 0 0), 45.3 (0 0 4), 62.5 (2 2 0). No signature of c-Mo2N with peak positions slightly shifted from t-Mo2N can be confirmed in the spectra. Additionally, h-MoN is observed with the (0 0 2) reflection growing at 36.2. Slightly different intensity ratios between
the final in situ and the ex situ data are due to the experimental set-ups. The peak intensities, or more correctly the integral peak areas, are plotted in Fig. 4 as a function of the nitrogen fluence on a logarithmic scale. No saturation with increasing fluence is observed for all investigated peaks. The sample implanted at 450 C shows a lower intensity and a slightly slower intensity increase, with the onset shifted by about a factor of 2. Apparently, the threshold
1
Intensity (a.u.)
450°C 550 °C t-Mo2N (220) / 62.5° t-Mo2N (204) / 64.3° t-Mo2N (112) / 37.7° h-MoN (002) / 36.2° ex-situ measurement
0.1
0.01 0
5
10
Nitrogen flux (10
15 17
-2
cm )
Fig. 4. Integral peak areas as a function of nitrogen flux for selected peaks, as derived from the in situ XRD data.
212
D. Manova et al. / Nucl. Instr. and Meth. in Phys. Res. B 240 (2005) 208–213 10000 Mo (42-1120) Mo2N (C) (25-1366)
1000
MoN (H) (77-1999) Mo2N (T)( 75-1150)
Intensity
100 10 1
700 °C, 30 min
0.1
650 °C, 60 min 600 °C, 60 min
0.01
450 °C, 40 min
1E-3 40
60
80
100
Angle 2θ (°)
Fig. 5. XRD spectra for low energy implanted samples (15 kV pulse voltage) at different temperatures and times.
is given by the sensitivity of the equipment and not by the phase formation kinetics. The MoN (0 0 2) reflection is growing as fast as the t-Mo2N (2 2 0) reflection, whereas the Mo2N (1 1 2) intensity is the fastest growing one. In conjunction with Fig. 1, the lower boundary for the formation of MoN precipitates is assumed to be between the composition MoN0.15 and MoN0.25 which are rather low values. Fig. 5 shows XRD spectra for samples implanted at low energies at different temperatures. Additional peaks, beside those of the base material, at 2H = 43.6 and 62.9 evolve with increasing treatment temperature. Beyond 580 C, a clearly resolved splitting of these peaks occurs, which is ascribed to a phase transition from cubic to tetragonal Mo2N [14]. The most likely explanation involves the lattice distortion due to the Jahn– Teller effect [15], coupled with a volume reduction by 2.2% for Mo2N (cubic: a = 0.42 nm, tetragonal: a = 0.4163 nm, c = 2 · 0.40 nm). No signature of a MoN phase is detected in these spectra. Comparing low energy and high energy implantation, a similar thermally activated diffusion coefficient is observed across the temperature range, with the diffusion length ranging from a few nanometers after 1 h at 330 C to more than 200 nm at 700 C after 30 min. The determination of exact values, complicated by the surface sputtering and the parallel process of implantation and diffusion,
is beyond the limited scope of this paper. However, the phase formation is different for both regimes: the sole formation of c-Mo2N, changing to the formation of t-Mo2N at 580 C and beyond observed at low energies is in contrast to the formation of t-Mo2N and h-MoN at 450 and 550 C, without any indication of c-Mo2N, at high energies. Radiation effects can be most likely excluded as, on the one hand no increased diffusion is observed, and on the other hand, increased pressure, e.g. during the collision cascades should lead to the transformation from t-Mo2N towards c-Mo2N [16], in contradiction to the observation. Another, more likely explanation involves the proximity to the surface for the low energy implantations, albeit ruling out any influence of contaminations on the phase formation as those are not present. The terminated half-space can be a favourable host matrix for slightly different phases than in the bulk as surface reconstruction and relaxation can change the local environment. At higher temperatures, these metastable phases (c-Mo2N in the present case) are again suppressed.
4. Summary and conclusions Nitrogen ion implantation into Mo at elevated temperatures between 300 and 750 C at low and
D. Manova et al. / Nucl. Instr. and Meth. in Phys. Res. B 240 (2005) 208–213
high energies lead to the broadening of the implantation profiles by thermally activated diffusion. Low energy (630 keV) and low temperatures (<580C) favour the formation of c-Mo2N, while t-Mo2N is observed for higher energies or higher temperatures. Additionally, MoN is observed for 1 MeV ion energy starting at a fluence of 5 · 1017 at./cm2 with a local N/Mo ratio of about 0.25. Further investigations are necessary to clarify the thresholds, dynamics and influence of the surface. Acknowledgements W. Assmann, Maier-Leibnitz-Labor, Garching, is acknowledged for providing the ERDA data. W. Mo¨ller, Institut fu¨r Ionenstrahlphysik und Materialforschung, Rossendorf, is acknowledged for providing the high energy beamline implantation facility.
References [1] H.J. Goldschmidt, Interstitial Alloys, Butterworths, London, 1967. [2] K.H. Jack, Acta Crystallogr. 13 (1950) 392.
213
[3] E.S. Bumps, H.D. Kessler, M. Hansen, Trans. Am. Soc. Metals 45 (1953) 1008. [4] J.S. Anderson, B.G. Hyde, J. Phys. Chem. Solids 28 (1967) 1393. [5] T.E. Parker, R. Kelly, J. Phys. Chem. Solids 36 (1975) 377. [6] B. Rauschenbach, A. Kolitsch, K. Hohmuth, Phys. Stat. Sol. A 80 (1983) 471. [7] Y. Okabe, T. Fujihana, M. Iwaki, B.V. Crist, Surf. Coat. Technol. 66 (1994) 384. [8] B. Renner, C. Hammerl, B. Rauschenbach, Nucl. Instr. and Meth. B 160 (2000) 363. [9] Y. Bohne, N. Shevchenko, F. Prokert, J. von Borany, B. Rauschenbach, W. Mo¨ller, Vacuum 76 (2004) 281. [10] L.R. Doolittle, Nucl. Instr. and Meth. B 9 (1985) 344. [11] A. Bergmeier, G. Dollinger, C.M. Frey, Nucl. Instr. and Meth. B 99 (1995) 488. [12] J.F. Ziegler, J.P. Biersack, V. Littmark, The Stopping and Range of Ions in Solids, Pergamon, New York, 1986. [13] Joint Committee for Powder Diffraction Standards, Powder Diffraction File, International Center for Diffraction Data, Park Lane, 1989. [14] S. Ma¨ndl, D. Manova, J.W. Gerlach, W. Assmann, H. Neumann, B. Rauschenbach, Surf. Coat. Technol. 180/181 (2004) 362. [15] T.A. Albright, J.K. Burdett, M.-H. Wnangbo (Eds.), Orbital Interactions in Chemistry, John Wiley and Sons, New York, 1985. [16] O.N. Mryasov, D.L. Novikov, A.J. Freeman, Ferroelectrics 164 (1995) 279.