Available online at www.sciencedirect.com NIM B Beam Interactions with Materials & Atoms Nuclear Instruments and Methods in Physics Research B 266 (2008) 2507–2510 www.elsevier.com/locate/nimb Ion beam modification of structural and electrical properties of TiN thin films M. Popović a, M. Stojanović a, D. Peruško a, M. Novaković a, I. Radović a, V. Milinović a,b, B. Timotijević a, M. Mitrić a, M. Milosavljević a,* b a VINČA Institute of Nuclear Sciences, P.O. Box 522, Belgrade 11001, Serbia II. Physikalisches Institut, Universität Göttingen, Friedrich-Hund-Platz 1, D-37077 Göttingen, Germany Available online 7 March 2008 Abstract A study of ion beam modification of structural and electrical properties of TiN thin films is presented. The layers were deposited by reactive ion sputtering on (1 0 0) Si and glass slide substrates to a thickness of 240 nm. After deposition the structures were implanted with argon ions at 120 keV, to the fluences from 1  1015 to 1  1016 ions/cm2. The ion energy was chosen to give the projected ion range within the deposited layers, to minimize the influence of the substrate on the induced structural changes. Structural analysis of the samples was performed by cross-sectional transmission electron microscopy, X-ray diffraction and Rutherford backscattering spectrometry. Electrical characterization included sheet resistivity measurements with a four point probe. It was found that the as-deposited layers have a columnar structure, individual columns stretching from the substrate to the surface and being a few tens of nanometers wide. Ion irradiation rearranges their crystalline structure, which remains polycrystalline, but the columns are broken, and nanocrystals of the same phase are formed. The structural changes can be nicely correlated to the measured electrical resistivity. Ó 2008 Elsevier B.V. All rights reserved. PACS: 68.55.Ln; 61.80. x Keywords: Hard coatings; Ion beam modification; TiN; RBS; TEM 1. Introduction Titanium–nitride films are interesting in tribology as hard coatings, wear and corrosion protection materials [1–3]. On the other hand, due to their relatively low electrical resistance, they are interesting as diffusion barrier layers in microelectronics. In general, their functional properties depend on their micro-structure, i.e. the mean grain size, preferred orientation, grain boundaries, crystalline defects, surface and interface morphology. Over the last decades, physical or chemical vapor deposition processes were proven as reliable for depositing good quality stoichiometric TiN layers. However, modern surface treatments involve the use of ion beams in the processes such as ion implanta* Corresponding author. Tel./fax: +381 11 8066425. E-mail address: momirm@vin.bg.ac.yu (M. Milosavljević). 0168-583X/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2008.03.032 tion, plasma ion immersion, or ion beam assisted deposition. Ion implantation was proven as a very powerful technique for improving tribological properties, both when applied directly on the processed materials or on previously deposited coatings [4,5]. Our interest here was to study the effects of ion irradiation on the micro-structural changes in TiN films, and on the changes of their electrical resistivity. The films were deposited on Si substrates by reactive ion sputtering and subsequently irradiated with argon ions. It was found that ion irradiation rearranges the layer structure, although they remain polycrystalline, which influences their electrical properties. 2. Experimental Titanium–nitride films were deposited by reactive ion sputtering in a Balzers Sputtron II system. Titanium target M. Popović et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 2507–2510 3. Results and discussion By RBS analysis we could examine elemental depth profiles and determine the layer stoichiometry. In Fig. 1 we present the analysis of an as-deposited TiN layer, at RT. The experimental data could be fitted well by introducing Ti, N, Ar and Si in the Data Furnace structure file, as seen in (a). Here we also plotted separated elemental spectra obtained from the fit. The analysis shows that argon is incorporated in the films during deposition. The Ar surface signal starts at a lower energy then the Ti signal, and extends all through the layer depth. The position of N surface signal coincides with the signal arising from the Si substrate. The extracted depth profiles (b) show a nearly 6000 Backscattering Yield (99.9% purity) was sputtered with argon ions in a nitrogen ambient. The base pressure in the chamber was in the low 10 6 mbar region, partial pressure of argon was 1  10 3 mbar, and partial pressure of nitrogen was 3  10 4 mbar. The substrates used in these experiments were crystalline (1 0 0) Si wafers. They were cleaned by a dip in HF solution and in deionized water before mounting, and by backsputtering prior to film deposition. During deposition the substrates were held at room temperature (RT) or at 150 °C. We first deposited a 10 nm of pure Ti buffer layer, to increase adhesion to the substrate, which was followed by deposition of 230 nm TiN, at 8 nm/min. A total layer thickness of 240 nm was measured with a Taylor–Hobson profilometer, which gives an accuracy of ±2 nm in this range. After deposition the samples were implanted with 120 keV Ar+ ions, to the fluences of 1  1015 and 1  1016 ions/cm2. During irradiation the samples were held at room temperature, the ion beam was scanned uniformly over an area of 2.5  2.5 cm2, and the beam current was kept at 1 lA/cm2. Calculations by TRIM [6] gave a projected ion range of Rp  70 nm and straggle DRp  30 nm, meaning that practically all implanted ions were stopped within the layers. SUSPRE [7] gave an estimate that the applied ion fluences were above the amorphisation threshold for the system. Structural characterization of the samples was performed with Rutherford backscattering spectrometry (RBS), transmission electron microscopy (TEM) and Xray diffraction analysis (XRD). For RBS analysis we used 900 keV He++ ion beam, with a detector positioned at 165° backscattering angle, at the IONAS facility in Goettingen [8]. We took random spectra at normal incidence and analysed the data with the Data Furnace code [9]. Cross-sectional TEM analysis was done on a Philips EM 400T microscope, and we also used micro diffraction (MD) technique to study the crystalline structure. XRD analysis was done at grazing incidence of 3°, with Cu Ka emission, using a Bruker D8 Advance Diffractometer. We also measured sheet resistivity of the samples with a four point probe, the values being calculated in specific resistivity using the measured thickness. data fit 4000 Si Ti 2000 Ar N 0 200 300 400 500 600 700 Energy (keV) 100 Concentration (at%) 2508 80 Ti N Ar Si 60 40 20 0 0 1000 2000 5 3000 2 Depth (10 at/cm ) Fig. 1. RBS analysis of a sample deposited at RT: (a) experimental and fitted data, (b) extracted depth profiles. uniform TiN layer stoichiometry, and 1–2 at% of Ar throughout the layer thickness. Towards the Si substrate we register an increased Ti yield, which corresponds to the thin buffer layer. RBS spectra taken from samples deposited at 150 °C, as well as those taken from the implanted samples, are all similar to the one given in Fig. 1. Ion implantation adds an extra up to 2 at% of Ar around the projected range for the higher fluence. The spectra remain essentially the same, except for a small increase in the yield that arises due to the implanted argon. For clarity, these other spectra are not presented in the figure. RBS analysis suggests that ion irradiation does not induce any redistribution of components or intermixing at the layer/substrate interface. TEM analysis revealed that the as-deposited layers grow in form of a polycrystalline columnar structure, with very fine crystalline grains. The column width is of the order of a few tens of nm, slightly increasing in size with the deposition temperature. After ion irradiation the structure remains polycrystalline, despite the high implanted fluences, which could induce amorphisation. An example of cross-sectional TEM analysis is illustrated in Fig. 2. The micrographs and the corresponding MD patterns were 2509 M. Popović et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 2507–2510 Fig. 2. TEM analysis of samples deposited at 150 °C: (a) bright field image of as-deposited, (b) bright field image of implanted to 1  1016 Ar/cm2. 200 Si (311) 150 TiN 100 (111) TiN TiN (200) TiN (220) (311) 50 0 Si (311) Intensity (a.u) taken from an as-deposited layer at 150 °C (a), and from a similar layer implanted to 1  1016 ions/cm2 (b). The MD patterns were taken with a 50 nm beam spot diameter, covering only a portion of the layer image. The diffraction spots lying on the first circle to the central spot correspond to (1 1 1) TiN reflections, and those on the second circle, very close to the first one, to (2 0 0) TiN reflections. Bright field image in (a) shows individual columns that stretch from the substrate to the surface, and the MD pattern indicates a very fine polycrystalline structure. Micrograph in (b) shows that the polycrystalline structure of the layer is retained after ion implantation. However, the columns appear as partly broken or disconnected, and we observe randomly distributed sharp contrasts arising from nanosized crystal grains. The corresponding MD pattern indicates a slightly finer grain structure. Compared to the pattern shown in (a) we observe a larger number of smaller spots lying on the circles around the central spot, indicating a higher number of randomly oriented smaller grains. The results of XRD analysis of as-deposited TiN sample at RT, and after irradiation to 1  1015 and 1  1016 ions/ cm2, are shown in Fig. 3. We can observe a change of TiN diffraction line intensities after ion irradiation. After irradiation at the lower fluence the intensity of (1 1 1) line drops, while the intensity of (2 0 0) and (2 2 0) lines increases, suggesting a partial texturing of the layer. With further increase of the irradiation fluence the relative ratio of the diffraction line intensities remains roughly the same, but their height decreases, indicating that the mean grain size in the layers decreases. In Fig. 4 we plotted the measured specific resistivity of TiN layers, deposited at RT and 150 °C, as a function of the irradiation fluence. The starting resistivity is higher for the lower deposition temperature, indicating a smaller mean grain size in these samples. However, in both cases the resistivity increases after ion implantation. These 150 TiN 100 TiN TiN (200) (220) (111) TiN (311) 50 0 Si (311) 150 100 TiN TiN TiN (111) (200) (220) TiN (311) 50 0 30 40 50 60 70 80 2θ(degrees) Fig. 3. XRD analysis of TiN/Si samples: (a) as-deposited at RT, (b) after ion irradiation to 1  1015 Ar/cm2, (c) after ion irradiation to 1  1016 Ar/ cm2. 2510 M. Popović et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 2507–2510 4. Conclusions 100 Resistivity (µΩcm) 90 80 70 60 RT 150oC 50 0 2 4 6 15 8 We have studied the effects of Ar ion irradiation on the micro-structure and electrical resistivity of TiN films. The as-deposited layers have a columnar structure, with the mean column width of the order of a few tens of nm, increasing with deposition temperature. Ion irradiation rearranges their crystalline structure, breaks up the columns, and induces growth of nanocrystals of the same phase. The structural changes induce a net increase of electrical resistivity of the layers, which can be correlated to an increased concentration of grain boundaries and point defects. 10 2 Ion Fluence (10 ions/cm ) Fig. 4. Measured resistivity of samples deposited at RT and at 150 °C, as a function of the irradiation fluence. results are in agreement with the results of TEM and XRD analyses, which indicate a finer grain size of the layers after ion irradiation. The results obtained by different analytical techniques can be used to describe the behaviour of TiN layers as a function of the deposition temperature and ion irradiation. The layers grow in form of a columnar structure, which suggests a layer by layer growth, where the process is controlled by surface diffusion [10]. Their mean column width is larger when they are grown at a higher deposition temperature, because the adsorbed atomic species have a longer diffusion length, and can hence form larger crystalline grains. Ion irradiation, known to generate numerous collision cascades, induces only local atomic rearrangements. The applied fluences were above the amorphisation level for the system, but the displaced Ti and N atomic species had sufficient mobility to recombine in a crystalline structure. The resulting structures consist of partly discontinued TiN columns and smaller crystalline grains of the same phase. Apart from this, ion irradiation can induce point defects in larger grains, which gives a net increase of electrical resistivity. Indeed, electrical resistivity measurements support this argumentation, in exhibiting higher values after ion implantation. Acknowledgements This work was supported by the Ministry of Science of the Republic of Serbia (Project No. OI 141013), and partly by the International Atomic Energy Agency, Vienna. The authors acknowledge Prof. H. Hofsäss for enabling RBS analysis at Göttingen University. References [1] J.E. Sundgren, Thin Solid Films 128 (1985) 21. [2] S. PalDey, S.C. Deevi, Mater. Sci. Eng. A 342 (2003) 58. [3] V.M. Vishnyakov, V.I. Bachurin, K.F. Minnebaev, R. Valizadeh, D.G. Teer, J.S. Colligon, V.V. Vishnyakov, V.E. Yurasova, Thin Solid Films 497 (2006) 189. [4] E. Cano, L. Martı́nez, J. Simancas, F.J. Pérez-Trujillo, C. Gómez, J.M. Bastidas, Surf. Coat. Technol. 200 (2006) 5123. [5] Y.P. Sharkeev, S.J. Bull, A.J. Perry, M.L. Klingenberg, S.V. Fortuna, M. Michler, R.R. Manory, I.A. Shulepov, Surf. Coat. Technol. 200 (2006) 5915. [6] J.F. Ziegler, J.P. Biersack, U. Littmark, The Stopping and Range of Ions in Solids, Pergamon, New York, 1985. [7] R.P. Webb, in: M. Seah, D. Briggs (Eds.), Practical Surface Analysis, Vol. 3, Wiley, 1991 (Appendix 3) <http://www.ee.surrey.ac.uk/ SCRIBA/simulations/Suspre/>. [8] M. Uhrmacher, K. Pampus, F.J. Bergmeister, D. Purschke, K.P. Lieb, Nucl. Instr. Meth. B 9 (1985) 234. [9] N.P. Barradas, C. Jeynes, R.P. Webb, Appl. Phys. Lett. 71 (1997) 291. [10] M. Milosavljević, G. Shao, M.A. Lourencßo, R.M. Gwilliam, K.P. Homewood, S.P. Edwards, R. Valizadeh, J.S. Colligon, J. Appl. Phys. 98 (2005) 123506.