Surface modification by negative-ion implantation

Surface modification by negative-ion implantation

Surface & Coatings Technology 203 (2009) 2351–2356 Contents lists available at ScienceDirect Surface & Coatings Technology j o u r n a l h o m e p a...

1MB Sizes 0 Downloads 40 Views

Surface & Coatings Technology 203 (2009) 2351–2356

Contents lists available at ScienceDirect

Surface & Coatings Technology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s u r f c o a t

Surface modification by negative-ion implantation Junzo Ishikawa ⁎, Hiroshi Tsuji, Yasuhito Gotoh Department of Electronic Science and Engineering, Kyoto University, Kyotodaigaku-Katsura, Nishikyo-ku, Kyoto 615-8510, Japan

a r t i c l e

i n f o

Available online 27 February 2009 Keywords: Negative-ion implantation Charging Nanoparticle Biocompatibility Nerve cell

a b s t r a c t We have developed a negative-ion implantation technique for surface modification of materials together with the development of high-intensity negative-ion sources and negative-ion implanters. In the negativeion implantation, we have the advantage of a nearly “charge-up-free” property for the implanted surface of insulators or insulated materials. The charging voltage on the implanted surface is no greater than plus or minus a few volts. By virtue of this merit, negative-ion implantation can include more excellent applications than those accessible only by positive-ion implantation. Negative-ion implantation enables us to open various applications, such as ion implantation to large scale integrated circuits (LSIs) without an electron shower neutralizer, surface modification of micrometer-sized powders by implantation without scattering, and formation of metal nanoparticles by implantation in a thin insulator film without damage. Besides, this technique makes possible the manipulation of nerve cell growth by precise control of the biocompatibility of a polymer surface. In this paper, novel surface modification applications by negative-ion implantation are reviewed. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Application of negative ions in surface modification has not been studied extensively because negative ions were not readily obtainable for practical ion implantation, and it has been believed that there would be no difference between negative- and positive-ion implantation effects. The author's group has investigated negative-ion production in detail [1,2] and developed various kinds of heavy negative-ion sources which were able to deliver sufficient negativeion currents for materials science applications [3–8]. Together with them, we developed two negative-ion implanters [7]. Since then, we have been investigating application possibilities of the negative-ion implantation technique [7–28]. During the investigation we found a distinct advantage in negative-ion implantation where almost no surface charging takes place during exposure of insulators or insulated materials to negativeion beams, i.e., a “charge-up-free” property [7–13]. This feature can enable the precise control of implanted ion energy and ion beam trajectory for negative-ion implantation into insulators or insulated materials, in which a large surface charging voltage induced by positive-ion implantation would decelerate the ion energy and deflect the ion beam direction. Also, the charge-up-free property can enable damage free implantation without breakdown when implanting into a thin insulator film whereas a large surface charging voltage induced by positive-ion implantation would cause a breakdown discharge through the insulator film. ⁎ Corresponding author. Tel.: +81 75 383 2272; fax: +81 75 383 2273. E-mail address: [email protected] (J. Ishikawa). 0257-8972/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.surfcoat.2009.02.023

We are now studying various applications of negative-ion implantation in LSI (Large Scale Integration) device fabrication [10], powder surface modification [14–16], nanoparticle formation in insulators for quantum devices [17–19], catalysis [20,21] and LE (light emitting) devices [22], and biocompatibility control for nerve cell patterning [23–26] and nervous system repair [27,28]. In this paper these applications of negative-ion implantation techniques will be reviewed. 2. Charging voltage by negative-ion implantation [7–13] When positive ions are implanted into an insulator or an insulated material, positive charge accumulates with time on the surface because incoming charge is positive due to positive ions and outgoing charge is negative due to secondary electrons. The secondary electrons may come back to the surface in response to its positive surface voltage, but positive charge accumulation due to positive ions continues. Thus, the surface voltage rises with time, in the worst cases, to the ion acceleration voltage. Then, the ion beam is decelerated and/or deflected by the large positive charging voltage, and the correct control of ion beam energy and implantation position is sometimes impossible. And also, the insulator may be destroyed by breakdown due to discharge through it. On the other hand, when using negative ions for implantation into an insulated material, the incoming charge is negative due to negative ions, and the outgoing charge is also negative due to secondary electrons. The negative ion to which an excess electron is weakly attached easily releases the electron on the bombardment, and the secondary emission factor is more than 1 in most cases. Therefore, charge balance will take place very easily at the surface by the return

2352

J. Ishikawa et al. / Surface & Coatings Technology 203 (2009) 2351–2356

and it can be used for modification of polymer surfaces; the other one is equipped with an rf-plasma sputtered negative-ion source which delivers a relatively high current and it can be used for nanoparticle formation, powder surface modification, etc. 4. Application of negative-ion implantation technique The “charge-up-free” property of negative-ion implantation techniques can solve the serious charge-up problem of positive-ion implantation, and widely expand the application fields of ion implantation. 4.1. LSI fabrication [10] Now the ion implantation technique is inevitable for LSI fabrication processes. These processes, however, use positive ions for implantation. In the positive-ion implanter for LSI fabrication, a charge neutralizer of electron shower is needed in order to avoid charging of the gate electrode of MOSFETs. This problem will become more and more severe in the near future. If we use negative ions instead of positive ones in implantation, this charging problem could be easily solved. For the estimation of LSI yield rates when using negative-ion implantation, the yield rates of a test element group (TEG) device with a gate oxide film thickness of 20 nm were measured. A copper negativeion beam with an energy of 15 keV and a current of 25 µA for 10 min (10 C/cm2) was implanted with a fluence of 1015 ions/cm2 into TEG devices with three antenna ratios of 1.6 × 103, 4 × 104 and 1.6 × 104. As shown in Fig.1, almost all devices with various antenna ratios showed no damage below an applied voltage of 18 V. This result indicates that almost no damage would be expected in the LSI fabrication process with negative-ion implantation, although most MOS devices would suffer serious damage under positive-ion implantation. Fig. 1. Results of breakdown voltage measurements for the test element groups with various kinds of antenna ratios after negative-ion implantation.

of some of the secondary electrons with low energy. The resulting surface charging voltage is thus very low, as low as positive several volts, which is 2–4 orders of magnitude lower than that due to positive-ion irradiation. In the case of negative-ion implantation into an insulator, the surface charging mechanism is considered to be slightly different from that for an insulated electrode, and the charging voltage is negative several volts due to the formation of an electrical double layer near the surface. Since the charging voltage by negative-ion implantation into an insulator or an insulated material is between plus several volts and minus several volts, essentially “charge-up-free” negative-ion implantation can be realized.

4.2. Powder surface modification [14–16] Powder surface modification by ion implantation, such as ceramic and polymer micrometer-sized particles, is greatly desired for the applications in medical and catalytic fields. When the powder particles are implanted with positive ions, they are scattered by a Coulomb repulsion force due to their high surface charging voltage, so the implantation with a sufficient fluence is very difficult. Fig. 2 shows the threshold charging voltages obtained in positiveargon-ion implantation for silica microbeads and for soda–lime

3. Negative-ion implanter [1–8] Originally negative-ion sources were developed for tandem accelerators, and the scientists did not intend to directly use negative ions for materials science. The author's group, however, developed the negative-ion sources for materials science applications, whose development needs were high current output and continuous operation. We investigated in detail the negative-ion production by surface effect mechanism, and found the optimal conditions for negative-ion production for sputtered particles from a cesiated surface. Using these results, we developed several kinds of sputter type heavy negative-ion sources such as NIABNISs (neutral- and ionized-alkali-metal bombardment type negative-ion sources) and rfplasma sputter type negative-ion sources. We developed two types of negative-ion implanters: the first one is equipped with a NIABNIS which delivers a relatively low current,

Fig. 2. Threshold charging voltages obtained in argon positive-ion implantation for silica microbeads (solid circle) and for soda–lime glass beads (solid squares), including a theoretical threshold charging voltage curve for silica microbeads.

J. Ishikawa et al. / Surface & Coatings Technology 203 (2009) 2351–2356

2353

glass beads, including a theoretical charging voltage curve for silica microbeads. In the case of negative-ion implantation, the surface charging voltage is much lower than the threshold voltage for scattering, and thus, no powder-particle scattering takes place during negative-ion implantation. 4.3. Nanoparticle formation Insulators including metal nanoparticles will offer a possibility of implantation for various kinds of devices such as quantum devices (single electron devices or memory devices), nonlinear optical devices, photocatalysts with high performance, and LE devices. Metal nanoparticle formation in insulators by ion implantation is a preferable method because ion implantation can precisely control the depth and fluence of implanted atoms by the ion energy and current, respectively. When positive ions are used in the implantation, the insulators sometimes suffer strong damage by breakdown due to surface charging. However, negative-ion implantation can enable the formation of metal nanoparticles in insulators without damage. 4.3.1. Quantum devices [17–19] In order to realize quantum devices such as a single electron device, the formation of delta layered nanoparticles in a thin insulator film with a thickness of less than 10 nm is required. Fig. 3 shows the TEM (Transmission Electron Microscope) image of gold nanoparticles formed in a thin SiO2 film by using negative-ion implantation. The ion implantation energy of the gold ions was 1 keV, which resulted in a very narrow distribution of implanted gold atoms. A delta layer of gold nanoparticles with diameters of 5–8 nm was formed around the projected range of gold atoms near the surface region. Delta layers of Ag and Ge nanoparticles were also formed near the boundary of Si and SiO2 and near the center of the SiO2 film, respectively, by selecting suitable implantation and annealing conditions. 4.3.2. Catalysis [20,21] In order to improve photocatalytic efficiency of rutile titanium, we implanted Ag or Cu negative ions into rutile titanium to form metal nanoparticles. The implanted samples have their own surface plasmon resonance peaks in the optical density spectra due to nanoparticle formation. We evaluated the photocatalytic efficiencies by decolorization of methylene-blue solution by irradiation with fluorescent light. Fig. 4 shows the relative efficiency as a function of the annealing temperature for silver negative-ion implanted rutile titanium samples. Photocatalytic efficiency was improved by about 2 times due to the metal nanoparticle formation with a suitable annealing process. This would be due to the third harmonic wave generation caused by

Fig. 3. TEM image of gold nanoparticles formed in thin SiO2 films by negative-ion implantation.

Fig. 4. Relative photocatalytic efficiency of Ag− implanted rutile samples in decolorization test of methylene-blue solution under irradiation of fluorescent light, where the relative efficiency is normalized by the photocatalytic efficiency of the unimplanted rutile.

nonlinear effects with metal nanoparticles and/or due to the efficient charge separation by electron trap with metal nanoparticles. 4.3.3. Light emitting devices [22] In order to develop LE devices, we formed Ge nanoparticles in a relatively wide depth region (around several tens of nm) in an SiO2 film. Estimated diameters of Ge nanoparticles are 3–4 nm for 600 °C or 700 °C annealing, and 5 nm for 900 °C annealing. Nanoparticle size was changed by annealing temperature and almost a half of the Ge atoms were oxidized after annealing. Maximum cathode luminescence emission was obtained at 800 °C. The wavelength at cathode luminescence peak (near 390 nm (3.18 eV)) did not change for any annealing temperature, as shown in Fig. 5. So, the cathode luminescence emission did not depend on their sizes. Therefore, this cathode luminescence is considered to be due to oxygen defects of germanium oxide. The cathode luminescence intensity has a maximum at the condition of Ge concentration of around 0.5 at.%. As for the photo-luminescence emission by Ti–sapphire laser irradiation, it was observed at 410 at room temperature. The strongest

Fig. 5. Annealing temperature dependence of cathode luminescence emission intensity for the SiO2 thin film samples containing Ge nanoparticles by Ge negative-ion implantation.

2354

J. Ishikawa et al. / Surface & Coatings Technology 203 (2009) 2351–2356

Fig. 8. Single cell-body and neurite of PC12h on the narrowed C− implanted regions of 7–10 µm on polystyrene surface. Fig. 6. Adhesion properties of PC12h cells on the C− implanted polystyrene surface after 2-day culture as a function of the ion fluence at an implantation energy of 10 keV.

4.4.1.1. PC12h cell patterning. For this experiment, we used mainly polystyrene and silicone rubber as the polymer materials for

modification. For the negative-ion species, we mainly used carbon because carbon is harmless for the body. For the nerve cell, we used mainly PC12h (rat adrenal pheochromocytoma cell) and MSC (mesenchymal stem cell). In implantation we used two masks with slit aperture (50 μm) and spacing (70 and 150 μm), and combining them we made stripe, ridge and grid patterned regions for implantation. Fig. 6 shows the number of PC12 cells which was adhered on the carbon negative-ion implanted polystyrene surface as a function of ion fluence. The cell-adhesion number increases with an increase in the ion fluence. Suitable ion fluence for good selective cell-adhesion property was obtained at the order of 1015 ions/cm2. The best fluence was 2 × 1015–3 × 1015 ions/cm2. From the experiment of ion implantation energy dependence, suitable energy for selective cell-adhesion properties was 10–20 keV, and the best energy was 10 keV. In case of silicone rubber, PC12h cells were well adhered at the condition of 3 × 1015 ions/cm2 for fluence and 10 keV for ion energy. Fig. 7 shows the average neurite outgrowth as a function of ion fluence at energies of 5–20 keV in case of PC12h cells on polystyrene. From the figure, the best condition for the neurite outgrowth is the same condition as that for cell adhesion. By using a ridge pattern, the effective minimum line widths of the implanted polystyrene for clear separation between cell-

Fig. 7. The average neurite length of PC12h cells on the C− implanted polystyrene films at implantation energies of 5–20 keV as a function of the ion fluence.

Fig. 9. Neuron network-like patterning of PC12h cells cultured on C− implanted polystyrene.

intensity of photo-luminescence emission was obtained for the sample prepared under the condition of 0.5 Ge atomic percent concentration and 800 °C annealing. 4.4. Biocompatibility control Nerve cell engineering is a promising technique for the future. This method can be applied to the following developments: (1) artificial nerve networks with living cells, for realizing a bio-interface between nerve system and external electronic circuit, and (2) guide tubes for nerve regeneration, i.e., tubulation. Good nerve cell affinity (nerve cell attachment and neurite outgrowth properties) on the surface of polymers (such as polystyrene and silicone rubber) just on micrometer-sized patterned modified regions is important. Then, negativeion implantation with the ability of precise control of ion beam trajectory is preferred. 4.4.1. Nerve cell adhesion and neurite growth patterning [23–26]

J. Ishikawa et al. / Surface & Coatings Technology 203 (2009) 2351–2356

2355

into nerve-like cells. Even at 10 days after differentiation, the cells were keeping the nerve-like shape as shown in Fig. 10(b). 4.4.2. Nervous system repair [27,28] Tubulation is a very promising prospective method for the regeneration of a nerve system. For this purpose, we used a negative-ion implantation technique to improve the inside wall of the silicone rubber tube. We performed carbon negative-ion implantation at an energy of 10 keV with fluence of 3 × 1015 ions/cm2. Rats were used in the tubulation operation, leaving an inter-stump gap of 15 mm. The electro-physiological properties were investigated after the tubulation. At 24 weeks after the tubulation, regeneration of the nerve tissue across the inter-stump gap of 15 mm through the carbon negative-ion implanted silicone rubber tube was observed. Also, the regenerated nerve tissue was functional, the rat with the regenerated nerve line clearly showing evoked action potentials to move the pedal adductor muscle. 5. Conclusion We developed negative-ion implanters with heavy negative-ion sources. Using them, we have developed a new negative-ion implantation technique, which has the advantage of a “charge-up-free” property during implantation into insulators or insulated conductive materials. As a surface modification by negative-ion implantation, we have challenged the following applications using negative-ion implantation technique: LSI fabrication without charge neutralizer; micrometer-sized powder surface modification without scattering; nanoparticle formation in thin insulator film without damage for quantum devices, catalysis and LE devices; and biocompatibility control for nerve cell-adhesion patterning and nervous system repair. The negative-ion implantation technique can thus provide an excellent tool for material processing applications. References

Fig. 10. (a) Adhered MSC cells after cultured in the pre-induction medium and (b) nerve-like shape cells at 10 days after differentiated.

body and neurite extension in the case of PC12h were obtained as shown in Fig. 8. The effective minimum line widths for single cellbody and neurite extension were found to be 5 μm and 2 μm, respectively. By using the grid pattern, we obtained a neuron network-like patterning of PC12h cells cultured on polystyrene carbon-negative-ion implanted at 10 keV and 3 × 1015 ions/cm2, as shown in Fig. 9. 4.4.1.2. MSC pattering. MSC (mesenchymal stem cell) has capabilities of self-reproduction and multi-differentiation into many kinds of cells such as neuron, bone, muscle, etc. So, the MSC is a prospective cell for nerve cell engineering in the future. We measured the adhesion and differentiation properties of rat MSC on polystyrene and silicone rubber. For the fluence dependence at 10 keV, the suitable fluence for patterning of rat MSC adhesion on the implanted region of polystyrene was in the range of 3 × 1014–1 × 1015 ions/cm2 in the energy range from 5 to 20 keV. Fig. 10(a) shows the photograph of adhered MSC cells after being cultured in the pre-induction medium. After replacement with the induction medium the differentiation to nerve cells was clearly observed. At 60 to 120 min after being cultured in the induction medium, the cells gradually changed their shape to have a rounded cell-body, and at 180 min almost all of the cells were differentiated

[1] J. Ishikawa, H. Tsuji, Y. Gotoh, Azegami, Proc. 6th Int. Symp. on Production & Neutralization of Negative Ions & Beams, AIP Conf. Proc., No. 287, AIP, Brookhaven, 1992, p. 66. [2] J. Ishikawa, H. Tsuji, T. Maekawa, Vacuum 39 (1989) 1129. [3] J. Ishikawa, Y. Takeiri, H. Tsuji, T. Taya, T. Takagi, Nucl. Instrum. Methods 232 (B4) (1984) 186. [4] J. Ishikawa, Y. Takeiri, T. Takagi, Rev. Sci. Instrum. 57 (1986) 1512. [5] J. Ishikawa, H. Tsuji, Y. Okada, M. Shinoda, Y. Gotoh, Vacuum 44 (1993) 203. [6] H. Tsuji, J. Ishikawa, Y. Okayama, Y. Toyota, Y. Gotoh, Proc. Ion Implantation Technology-94, Elsevier, Amsterdam, 1995, p. 495. [7] J. Ish ikawa, H. Tsuji, Y. Toyota, Y. Gotoh, K. Matsuda, M. Tanjo, S. Sakai, Nucl. Instrum. Methods B Beam Interact. Mater. Atoms 96 (1995) 7. [8] J. Ishikawa, Surf. Coat. Technol. 65 (1994) 64. [9] J. Ishikawa, Mat. Res. Soc. Symp. Proc., vol. 354, Material Research Society, Pittsburgh, 1995, p. 99. [10] J. Ishikawa, Rev. Sci. Instrum. 65 (1994) 1290. [11] J. Ishikawa, Rev. Sci. Instrum. 67 (1996) 1410. [12] H. Tsuji, J. Ishikawa, S. Ikeda, Y. Gotoh, Nucl. Instrum. Methods B Beam Interact. Mater. Atoms 127/128 (1997) 278. [13] H. Tsuji, Y. Gotoh, J. Ishikawa, Nucl. Instrum. Methods B Beam Interact. Mater. Atoms 141 (1998) 645. [14] J. Ishikawa, H. Tsuji, Y. Gotoh, Proc. Ion Implantation Technology-96, IEEE 96TH8182, 1997, p. 249. [15] J. Ishikawa, H. Tsuji, M. Mimura, S. Ikemura, Y. Gotoh, Surf. Coat. Technol. 103/104 (1998) 173. [16] H. Tsuji, S. Kido, H. Sakaki, Y. Gotoh, J. Ishikawa, Rev. Sci. Instrum. 71 (2000) 804. [17] H. Tsuji, N. Arai, T. Matsumoto, K. Ueno, Y. Gotoh, K. Adachi, H. Kotaki, J. Ishikawa, Appl. Surf. Sci. 238 (2004) 132. [18] H. Tsuji, N. Arai, T. Matsumoto, K. Ueno, K. Adachi, H. Kotaki, Y. Gotoh, J. Ishikawa, Surf. Coat. Technol. 196 (2005) 39. [19] H. Tsuji, N. Arai, N. Gotoh, T. Minotani, K. Kojima, K. Adachi, H. Kotaki, T. Ishibashi, Y. Gotoh, J. Ishikawa, Nucl. Instrum. Methods B Beam Interact. Mater. Atoms 257 (2007) 94. [20] H. Tsuji, H. Sugahara, Y. Gotho, J. Ishikawa, Nucl. Instrum. Methods B Beam Interact. Mater. Atoms 206 (2003) 249. [21] H. Tsuji, N. Sakai, Y. Gotho, J. Ishikawa, Nucl. Instrum. Methods B Beam Interact. Mater. Atoms 242 (2006) 129. [22] H. Tsuji, N. Arai, N. Gotoh, T. Minotani, T. Ishibashi, T. Okumine, K. Adachi, H. Kotaki, Y. Gotoh, J. Ishikawa, Surf. Coat. Technol. 201 (2007) 8516.

2356

J. Ishikawa et al. / Surface & Coatings Technology 203 (2009) 2351–2356

[23] P. Sommani, H. Tsuji, H. Sato, T. Kitamura, M. Hattori, Y. Gotoh, J. Ishikawa, Trans. Mater. Res. Soc. Jpn. 31 (2006) 673. [24] P. Sommani, H. Tsuji, H. Sato, T. Kitamura, M. Hattori, Y. Gotoh, J. Ishikawa, Trans. Mater. Res. Soc. Jpn. 31 (2006) 669. [25] H. Tsuji, P. Sommani, T. Kitamura, M. Hattori, H. Sato, Y. Gotoh, J. Ishikawa, Surf. Coat. Technol. 201 (2007) 8123.

[26] P. Sommani, H. Tsuji, H. Sato, T. Kitamura, M. Hattori, Y. Gotoh, J. Ishikawa, Nucl. Instrum. Methods B Beam Interact. Mater. Atoms 257 (2007) 118. [27] H. Tsuji, M. Izukawa, R. Ikeguchi, R. Kakinoki, H. Sato, Y. Gotoh, J. Ishikawa, Nucl. Instrum. Methods B Beam Interact. Mater. Atoms 206 (2003) 507. [28] H. Tsuji, M. Izukawa, Y. Utagawa, R. Ikeguchi, R. Kakinoki, H. Sato, Y. Gotoh, J. Ishikawa, Trans. Mater. Res. Soc. Jpn. 29 (2004) 575.