Composition Control of R F Sputtered Ni2MnGa Thin Films Using Optical Emission Spectroscopy

Full text

(1)Materials Transactions, Vol. 43, No. 5 (2002) pp. 871 to 875 Special Issue on Smart Materials-Fundamentals and Applications c 2002 The Japan Institute of Metals. Composition Control of R.F.-Sputtered Ni2 MnGa Thin Films Using Optical Emission Spectroscopy Shyi-Kaan Wu1, ∗1 and Kuan-Hua Tseng2, ∗2 1 2. Department of Materials Science and Engineering, National Taiwan University, Taipei 106, Taiwan, R.O. China Department of Mechanical Engineering, National Taiwan University, Taipei 106, Taiwan, R.O. China. Optical emission spectroscopy can be used to monitor the composition of Ni2 MnGa thin films during sputtering. By choosing peaks of Ni:341.5 nm, Mn:403.1 nm and Ga:417.2 nm, the Ar pressure is found to affect the spectrum intensities of Ni, Mn and Ga atoms, as well as the intensity ratios of IMn /INi and IGa /INi . However, the r.f. power has no obvious effect on them. This may be due to the ferromagnetic characteristic of Ni, or that different metals have different energy distributions of sputtered atoms, or that they need various p · d values to be thermalized. Here, p is the Ar pressure and d is the target and substrate distance. The intensity ratios of these peaks are found to be proportional to the composition ratios (mol ratio) of thin films with the relations: C Mn /C Ni = 0.0151(IMn /INi ) + 0.392 and C Ga /C Ni = 0.0720(IGa /INi ) + 0.273. Hence, the composition of sputtered thin films can be predicted by monitoring the intensity of light emission from the sputtering plasma. (Received October 23, 2001; Accepted February 25, 2002) Keywords: Ni2 MnGa shape memory alloys, r.f.-sputtering, optical spectroscopy, argon pressure effect. 1. Introduction. 2. Experimental Procedure. The near-stoichiometric Ni2 MnGa alloys undergo a thermoelastic martensitic transformation and exhibit a ferromagnetic transition near vicinity of 370 K, and they are thus known as the ferromagnetic shape memory alloys (SMAs).1–5) Bulk single crystals of near-stoichiometric Ni2 MnGa alloys have shown exceptionally large magnetostriction and thin films of these alloys have been regarded as promising candidates for microactuator materials.6, 7) However, the martensitic transformation temperatures of near-stoichiometric Ni2 MnGa SMAs are very sensitive to their composition.3) Therefore, accurate composition of these SMAs is demanded for industrial applications. Since the composition of a sputtered thin film deviates from that of the alloy target, the composition control of thin films of nearstoichiometric Ni2 MnGa SMAs is extremely important in the fabrication. The composition of thin films is usually determined by the electron probe X-ray microanalysis (EPMA) after their deposition on the substrate. Besides using EPMA, monitoring the plasma concentration during sputtering is another feasible method. Optical spectroscopy has been widely used to monitor plasma etching, sputtering deposition and concentration analysis.8–12) In this study, thin films of near stoichiometric Ni2 MnGa SMAs are sputtered on Si(100) wafers. In order to monitor the composition of sputtered Ni2 MnGa thin film, optical emission spectroscopy is used to detect the plasma concentration during sputtering. The effect of Ar pressure during r.f. sputtering on the plasma intensity is also investigated. In addition, the relation between the peak intensity of plasma and the composition of sputtered thin films is discussed.. Near-stoichiometric Ni2 MnGa thin films are deposited on a n-type Si(100) wafer with a diameter of 76.2 mm by r.f. magnetron sputtering using a target disk of Ni2 MnGa with a diameter of 50.8 mm in argon atmosphere. The target was prepared by vacuum arc melting from the raw materials of nickel (purity 99.9 mass%), Mn–45 mass%Ni mother alloy and gallium (purity 99.99 mass%); then homogenized at 1123 K (850◦ C) for 173 ks (48 h); and finally wire-cut into a target disk with a diameter of 50.8 mm. The target disk composition was determined by EPMA using a JEOL JAX-8600SX instrument and was found to be Ni50.2 Mn24.8 Ga25.0 (in mol%). The sputtering conditions were as follows. The distance, d, between the target and substrate was 60 mm. The base pressure was 2.7 × 10−4 Pa. The Ar pressure, p, and r.f. power, W , varied from 0.67 to 4.0 Pa and 100 to 200 W, respectively. Figure 1 shows the sputtering system and optical spectrometer system used in this study. The light emission from the plasma was detected through a quartz window by a Jobin-Yvon Triax 320 monochromator and a R928 photomultiplier tube (PMT). The special lines were obtained by scanning wavelength from 335 to 425 nm and recorded in a computer. The scanning step was 0.07 nm. The peaks of Ni were chosen according to our previous study12) and the study of Bendahan et al.11) for r.f.-sputtered thin films of TiNi-based SMAs; whereas the peaks of Mn and Ga were chosen using the MIT Wavelength Tables.13) To prevent the loss of optical transmittance resulting from the deposition of Ni, Mn and Ga atoms on the quartz window, a clean quartz window was used after each scanning. In order to clean the contamination on the target surface and to obtain a stable glow discharge, the pre-sputtering time was chosen to be one hour before the deposition of thin film and the detection of plasma spectra. The compositions of sputtered Ni2 MnGa thin films were determined by EPMA using a JEOL JXA-8600SX model calibrated by specimens whose compositions had been measured by inductively. ∗1 Corresponding ∗2 Graduate. author: E-mail: [email protected] Student, National Taiwan University..

(2) 872. S.-K. Wu and K.-H. Tseng. Fig. 2 Spectrum of the plasma with Ni50.2 Mn24.8 Ga25.0 alloy target in wavelengths from 335 to 425 nm. Important peaks corresponding to Ni, Mn and Ga atoms are also indicated. The sputtering condition is p = 2.7 Pa, d = 60 mm and W = 100 W.. plasma, can be represented as [X ∗ ] = Fig. 1 Schematic representation of the experimental apparatus used in this study.. coupled plasma—atomic emission spectrometer (ICP-AES), a Jobin-Yvon JY 38 PLUS model. 3. Theoretical Considerations Optical spectroscopy was used to identify the relation between the intensity of light emitted from the plasma and the composition of thin films. The theoretical consideration is based on the assumption that the intensity of light emitted from the plasma is proportional to the concentration in the plasma.14–16) The intensity of the spectral line corresponds to the transition of energy level in an atom. Most sputtered atoms ejected into the plasma region are neutral.17, 18) These atoms are directly impacted by electrons to an excited state X + e → X∗ + e. rate 1 = ke [X][e]. (1). where X ∗ is as the excitation state of an X atom. In this study, X includes Ni, Mn and Ga atoms. De-excitation process can occur in radiative decay and collisional quenching X ∗ → X + hν. rate 2 = kr [X ∗ ]. (2). X ∗ + M → X + M + (kinetic energy of M) rate 3 = kq [X ∗ ][M]. (3). where M represents Ar atoms; h is the Planck’s constant; ν is the frequency of light emitted; and ke , kr and kq are the rate constants for the above three processes, respectively. At steady-state, the excitation rate is equal to the rate of radiative decay and collisional quenching, i.e., rate 1 = rate 2 + rate 3. From this relationship, [X ∗ ], the concentration of X ∗ in the. ke [X] · [e] kr + kq [M]. (4). The intensity of the spectral line is proportional to reaction (2) and can be written as I ∝ kr [X ∗ ] =. ke [X] · [e] 1 + (kq /kr ) · [M]. (5). Furthermore, the relative intensity ratio of Mn and Ni can be shown as IMn ke,Mn 1 + (kq,Ni /kr,Ni ) · [Ar] [Mn] · (6) ∝ · INi [Ni] ke,Ni 1 + (kq,Mn /kr,Mn ) · [Ar] Equation (6) shows that the intensity ratio of spectral lines is proportional to the concentration ratio in the plasma. The quantity within the larger round brackets on the right side of eq. (6) is a constant under certain hypotheses. The excitation rate constant of Ni, (ke,Ni ), depends on the threshold energy, the excitation cross section of Ni and the electron energy distribution function. The same situation can be given for ke,Mn . Thus, the quantity within the larger round brackets can be considered to be a constant only if the Ni and Mn have approximately the same form of excitation cross section, the same threshold energies and have their excitation taking place from the ground state by direct electron impact. In the same way, IGa /INi is proportional to [Ga]/[Ni] IGa ke,Ga 1 + (kq,Ni /kr,Ni ) · [Ar] [Ga] · (7) ∝ · INi [Ni] ke,Ni 1 + (kq,Ga /kr,Ga ) · [Ar] 4. Results and Discussion 4.1 Plasma peaks detected by optical emission spectroscopy Figure 2 shows the spectrum from a glow discharge with Ni50.2 Mn24.8 Ga25.0 target in wavelength from 335 nm to 425 nm. Most of the peaks are due to Ni and Mn, with reference to the MIT Wavelength Tables.13) In Fig. 2, some of the unlabeled peaks may be due to Ar atoms.19) Figure 2 shows that the strongest intensity peaks for Ni, Mn and Ga atoms are 361.9 nm, 403.1 nm and 417.2 nm, respec-.

(3) Composition Control of Sputtered Ni2 MnGa Thin Films. 873. 3.5. Mn. 200W. Relative Intensity. 2. (a). 3. Relative Intensity. 3. Ga. 1. 2.5 2 1.5 Ni Mn Ga. 1 0.5. Ni. 0 0. 0. 100. 200. 300. r.f. Power, W /W. Mn. 2 100W. Ga. 1. Intensity Ratio. Ni. 0 335. 350. 365. 380. 395. 410. 425. 8 7 6 5 4 3 2 1 0. Mn/Ni Ga /Ni 0. Wavelength, /nm Fig. 3 Spectrum of the plasma with Ni50.2 Mn24.8 Ga25.0 alloy target under r.f. power 100 W and 200 W. The more sensitive peaks corresponding to Ni:341.5 nm, Mn:403.1 nm and Ga:417.2 nm are also indicated. The sputtering condition is p = 2.7 Pa and d = 60 mm.. (b). 50. 100. 150. 200. 250. r.f. Power , W /W. Fig. 4 (a) Spectrum intensities of Ni(341.5 nm), Mn(403.1 nm) and Ga(417.2 nm) versus the r.f. power W . (b) The intensity ratios of IMn /INi and IGa /INi versus the r.f. power W . 0.4 Ni. 0. Mn. 0.67Pa. Ga. Mn. 0.4. Ga. Ni 1.3Pa. Relative Intensity. tively. However, according to the MIT Wavelength Tables.,13) the 341.5 nm peak for Ni, the 403.1 nm peak for Mn and the 417.2 nm peak for Ga are much more sensitive than the others, although one of them is not the most intense one. In this study, only the most sensitive peaks are used to monitor the plasma intensity and to relate this intensity to the composition of sputtered thin films.. 0 Mn Ga. 0.4 Ni 2.7Pa 0 Mn. 1.2 0.8. Ga. 4.2 Effect of r.f. power on plasma intensity The effect of r.f. power on the spectrum of glow discharge with Ni50.2 Mn24.8 Ga25.0 target is shown in Fig. 3. With reference to Fig. 3, the effect of r.f. power on the peak intensities of Ni, Mn and Ga atoms and the intensity ratios of IMn /INi and IGa /INi are plotted in Figs. 4(a) and (b), respectively. Figure 4(a) indicates that the higher the r.f. power, the higher the peak intensity of Ni, Mn and Ga atoms. Furthermore, from Fig. 4(b) it can be seen that the intensity ratios of IMn /INi and IGa /INi remain constant in the r.f. power range of 100 W to 200 W. This means that the effect of r.f power on the intensity ratios of IMn /INi and IGa /INi , i.e., the composition of sputtered thin film, is not significant. 4.3 Effect of Ar pressure on plasma intensity The effect of Ar pressure on the spectrum of glow discharge with a Ni50.2 Mn24.8 Ga25.0 target is shown in Fig. 5. With reference to Fig. 5, the effect of Ar pressure on the peak intensities of Ni, Mn and Ga atoms, as well as the intensity ratios of IMn /INi and IGa /INi are plotted in Figs. 6(a) and (b), respectively. From Fig. 6(a), it can be seen that the peak intensities of Ni, Mn and Ga atoms increase with the increasing Ar pressure. At the same time, from Fig. 6(b), the intensity ratios of IMn /INi and IGa /INi also increase with increasing Ar pressure, and the IMn /INi ratio increases more quickly than the IGa /INi ratio. In our previous study on the composition control of r.f.-. 0.4. Ni 4.0Pa. 0. 335. 350. 365. 380. Wavelength,. 395. 410. 425. /nm. Fig. 5 Spectrum of the plasma with Ni50.2 Mn24.8 Ga25.0 alloy target under various Ar pressure. The more sensitive peaks corresponding to Ni:341.5 nm, Mn:403.1 nm and Ga:417.2 nm are also indicated. The sputtering condition is d = 60 mm and W = 100 W.. sputtered Ti50 Ni40 Cu10 thin films using optical emission spectroscopy,12) spectra of the plasma with targets of pure Ni, Ti and Cu under various Ar pressure were investigated. Experimental results showed that only the relative spectrum intensity of Ni peaks varies with Ar pressure, say INi(341.5 nm) : INi(352.5 nm) : INi(361.9 nm) = 1 : 1.22 : 1.72 at 4.0 Pa, but the intensity is 1 : 1.43 : 2.38 at 6.7 Pa. This characteristic is not found in Ti and Cu, but only in Ni because Ni is a ferromagnetic metal, and the Ni atoms ejected from the target can be influenced by the magnetic field during r.f. magnetron sputtering. In addition, Staut et al. found that the energy distributions of atoms sputtered from different polycrystalline metals are different.20) Somekh also indicated, under the same initial sputtering energy, sputtered atoms of different elements have different values of p · d (“Ar pressure” × “target and substrate distance”) to be thermalized.21) In other words, it is possible that the peak intensity of Mn increases more quickly than that of Ga when the Ar pressure increases. The charac-.

(4) 874. S.-K. Wu and K.-H. Tseng 1.5 Relative Intensity. (a). compositions of Mn, Ni and Ga, respectively, on the Si wafer with a diameter of 76.2 mm. The compositions are measured by EPMA for every 2 mm distance along the diameter, and then the average of the composition is taken. The appropriate equations of Figs. 7(a) and (b) are as follows:. Ni Mn Ga. 1. 0.5. 0 0. 1. 2 3 Ar Pressure, p /Pa. 4. 5. (b). Intensity Ratio. 8. Mn /Ni Ga/Ni. 6 4 2 0. 1. 2. 3. 4. (9). 5. Ar Pressure, p /Pa. Fig. 6 (a) Spectrum intensities of Ni(341.5 nm), Mn(403.1 nm) and Ga(417.2 nm) versus the Ar pressure. (b) The intensity ratios of IMn /INi and IGa /INi versus the Ar pressure. 0.53. (a) Composition Ratio C Mn/C Ni. C Ga /C Ni = 0.0720(IGa /INi ) + 0.273. 5. Conclusions. 0. 0.52 0.51 0.5. C Mn/C Ni = 0.0151(I Mn/I Ni) + 0.392. 0.49. R 2 = 0.90. 0.48 6. 7. 8. 9. Intensity Ratio I Mn/I Ni 0.54. Composition Ratio C Ga/C Ni. (8). From Fig. 7, eqs. (8) and (9) have quite good curve fittings, with the correlation coefficient R 2 being around 0.90–0.94. According to eqs. (8) and (9), as long as the intensity of light emission from the sputtering plasma is monitored, the composition of thin films can be predicted. Although the concentration ratio of plasma is not necessarily equal to the composition ratio of thin films, we find that the optical signals are stable and reproducible. Therefore eqs. (8) and (9) can be used to monitor the composition of Ni2 MnGa thin films.. 12 10. C Mn /C Ni = 0.0151(IMn /INi ) + 0.392. C Mn /C Ni = 0.0151(IMn /INi ) + 0.392 C Ga /C Ni = 0.0720(IGa /INi ) + 0.273. (b). 0.52. Therefore, the composition of sputtered Ni2 MnGa thin films can be predicted by monitoring the intensity of light emission from the sputtering plasma.. 0.5 0.48. C Ga/C Ni = 0.072(I Ga/I Ni) + 0.273. 0.46. R 2 = 0.94. 0.44 2.5. Optical emission spectroscopy can be used to monitor the composition of Ni2 MnGa thin films during sputtering. By choosing more sensitive peaks of Ni:341.5 nm, Mn:403.1 nm and Ga:417.2 nm, we find that the sputtering Ar pressure can affect the spectrum intensities of Ni, Mn and Ga atoms and the intensity ratios of IMn /INi and IGa /INi during r.f. magnetron sputtering, but the r.f. power has no obvious effect on them. This may due to the ferromagnetic characteristic of Ni atoms and the fact that the energy distributions of sputtered atoms and their p · d values to be thermalized are different for each metal. The intensity ratios IMn /INi and IGa /INi are found to be proportional to the composition ratio of thin films. The relations are as follows:. 2.7. 2.9. 3.1. 3.3. 3.5. Acknowledgements. Intensity Ratio IGa /I Ni. Fig. 7 (a) Intensity ratio IMn /INi versus composition C Mn /C Ni of Ni2 MnGa thin films. (b) Intensity ratio IGa /INi versus composition C Ga /C Ni of Ni2 MnGa thin films.. The authors are grateful for financial support of this study from the National Science Council (NSC), Republic of China, under Grant NSC 90-2216-E002-024. REFERENCES. teristics shown in Fig. 6(b) can be elucidated from the results of the above-mentioned reports. 4.4 Relation between the intensity ratio of plasma and the composition of thin film In this study, we choose the more sensitive peaks, INi(341.5 nm) , IMn(403.1 nm) , and IGa(417.2 nm) , to establish the relation between the intensity of plasma and the composition of thin films. Figures 7(a) and (b) plot the dependence of intensity ratio IMn /INi versus the composition ratio C Mn /C Ni of Ni2 MnGa thin films and that of IGa /INi versus C Ga /C Ni , respectively. Here, C Mn , C Ni and C Ga are the average thin film. 1) P. J. Webster, K. R. A. Ziebeck, S. L. Town and M. S. Peak: Philos. Mag. B 49 (1984) 295–310. 2) V. V. Kokorin and V. A. Chernenko: Phys. Met. Metallography 68 (1989) 111–115. 3) V. A. Chernenko, E. Cesari, V. V. Kokorin and I. N. Vitenko: Scripta Metall. Mater. 33 (1995) 1239–1244. 4) K. Ullakko. J. K. Huang, C. Kanter, R. C. O’Handley and V. V. Kokorin: Appl. Phys. Lett. 69 (1996) 1966–1968. 5) E. Cesari, V. A. Chernenko, V. V. Kokorin, J. Pons and C. Segui: Acta Mater. 45 (1997) 999–1004. 6) R. Tickle and R. D. James: J. Magn. Magn. Mater. 195 (1999) 627–638. 7) M. Wuttig, C. Craciunescu and J. Li: Mater. Trans., JIM 41 (2000) 933– 937..

(5) Composition Control of Sputtered Ni2 MnGa Thin Films 8) J. E. Greene and F. Sequeda-Osorio: J. Vac. Sci. Technol. 10 (1973) 1144–1149. 9) R. d’Agostino, F. Cramarossa, S. De Benedictis and G. Ferraro: J. Appl. Phys. 52 (1981) 1259–1265. 10) J. W. Coburn and M. Chen: J. Appl. Phys. 51 (1980) 3134–3136. 11) M. Bendahan, J. Seguin, P. Canet and H. Carchano: Thin Solid Films 283 (1996) 61–66. 12) S. K. Wu, Y. S. Chen and J. Z. Chen: Thin Solid Films 365 (2000) 61–66. 13) G. R. Harrison: M.I.T. Wavelength Table, vol. 1, Tables of Wavelengths, (Wiley, New York, 1969). 14) R. A. Gottscho and V. M. Donnelly: J. Appl. Phys. 56 (1984) 245–256.. 875. 15) T. J. Cotler, M. L. Passow, J. P. Fournier, M. L. Brake and M. Elta: J. Appl. Phys. 69 (1991) 2885–2888. 16) A. Richard, H. Michel, P, Jacquot and M. Gantois: Thin Solid Films 124 (1985) 67–73. 17) J. Comas and C. B. Cooper: J. Appl. Phys. 38 (1967) 2956–2962. 18) J. R. Woodyard and C. B. Cooper: J. Appl. Phys. 35 (1964) 1107–1117. 19) J. E. Greene, F. Sequeda-Osorio and B. R. Natarajan: J. Appl. Phys. 46 (1975) 2701–2709. 20) R. V. Stuart, G. K. Wehner and G. S. Anderson: J. Appl. Phys. 40 (1969) 803–812. 21) R. E. Somekh: J. Vac. Sci.Technol. A 2 (1984) 1285–1291..

(6)

No results found