solidi
status
pss
physica
phys. stat. sol. (a) 205, No. 9, 2126 – 2129 (2008) / DOI 10.1002/pssa.200879733
a
www.pss-a.com
applications and materials science
Transparent diamond-on-glass micro-electrode arrays for ex-vivo neuronal study
M. Bonnauron1, S. Saada1, C. Mer1, C. Gesset1, O. A. Williams2, L. Rousseau3, E. Scorsone1, P. Mailley4, M. Nesladek1, 5, J.-C. Arnault1, and P. Bergonzo*, 1 1
CEA, LIST, Diamond Sensor Laboratory, CEA/Saclay, Gif-sur-Yvette, 91191 Gif-sur-Yvette, France Institute for Materials Research, University of Hasselt, Belgium 3 ESIEE-ESYCOM, Paris, France 4 DRFMC/SPrAM/CREAB, CEA-Grenoble, 17, avenue des Martyrs, 38054 Grenoble, France 5 Institute of Physics, Czech Academy of Sciences v.v.i., 182 21 Prague, Czech Republic 2
Received 21 May 2008, revised 9 July 2008, accepted 15 July 2008 Published online 25 August 2008 PACS 68.37.Hk, 68.55.–a, 81.05.Uw, 81.15.Gh, 81.16.Rf *
Corresponding author: e-mail
[email protected]
We report on the fabrication of high aspect ratio diamond Micro Electrode Arrays (MEAs) grown on silicon as well as on glass substrates using an optimised nanoseeding technique and Bias Enhanced Nucleation (BEN). Such MEA systems combine high electrode reactivity and high electrical current
injection limits with resiliency, biocompatibility and optical transparency of diamond surfaces. We present the technological steps for the fabrication of 2D as well as 3D diamond microelectrode arrays. The patterning issues involve the use of detonation nanodiamond particles (DND).
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1 Introduction Recent developments in neurosciences require the study of fully organised networks of neural cells. This is most commonly achieved on slices cut from the neural or whole embryonic spinal chord system. The study of neuronal tissue requires both to stimulate and to record neuronal signal within the cell network. One of the issues is in the emergence of 3D Micro-Electrode Arrays (MEA), that have to be very dense and exhibit a high aspect ratio in order to connect to undamaged cells well above the edges of the studied slices. Several systems, using noble metals such as platinum, have already been used for the neuronal signal recording [1, 2]. They consist of arrays of tips, typically 70 micrometer high, on which spinal chords as well as retinas can be deposited for ex-vivo neuronal recording. However electrical excitation of the cells during long duration experiments may here cause irreversible reactions at the electrode surface, and water hydroxylation is still a difficult issue. Furthermore, the use of a transparent electrode material is here highly desirable in order to enable other diagnostic techniques based on e.g. fluorescence to correlate electrical and biochemical activity. Diamond is a promising material for novel applications in bio-electronics [3]. In the field of neuroscience: several
recent communications confirmed its remarkable bio inertness [4, 5], thus demonstrating the interest of synthetic diamond coatings for bio interfacing applications. Further, its electrochemical stability has been demonstrated to be useful for long term stimulation. Latest developments in our team led to step up chemical treatments that enhanced significantly both the electrode reactivity and the stability [6]. We achieved a constant reaction rate, k0 above 0.1 cm/s without any long term stability drifts under electrochemical stimulation. The development of diamond MEAs which allow the study of neurons in vitro is a necessary step towards biocompatible in-vivo chips for retinal implants. Firstly, it permits to understand better how the stimulation must be applied to neurons in order to get a reliable response. Secondly, it allows verification of interactions between material and neurons. Finally, it allows management of the robustness of the chip in the physiological media under electrical stimulation. 2 Experimental fabrication of 2D patterns Chemical Vapour Deposition (CVD) technique enables the growth of submicron diamond grains in the form of a © 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Original Paper phys. stat. sol. (a) 205, No. 9 (2008)
dense film on Si and glass substrates on large areas and can be used for the fabrication of MEAs. The principle outlined in our work relies on the interaction of active carbon species with so called detonation nanodiamond particles (DND). The DND diamond is prepared by a denotation synthesis and is currently commonly used for seeding of CVD diamond thin films, avoiding thus the diamond nucleation step, and leading to the growth of continuous layers. In this work, we are using the above procedure combined with the DND patterning; in order to develop a technique for fabrication of a diamond based fully transparent Micro-Electrode Array (MEA), avoiding time consuming deep or dry etching of diamond. The principle of the method is based upon a combination between the microelectronic lift off technique, with nanoseeding techniques [7–9]. Some patterning methods were previously reported [10–13], which rely on the catalyst effect of Pt or Mo vs. SiO2 and/or microcrystalline diamond powder, but none of them addresses the advantage of the selective seeding and adherence properties of DND. We initially developed the technique onto undoped (100) silicon wafers before applying it to Schott 33-boronfloat glass. This glass was chosen for its compatibility with the microelectronics processes. Before sonication the dispersion showed a distribution of 2 particle sizes. The first one, around 25 nm could be attributed to small aggregated particles (99.5%) and the second one around 120 nm, to resulting agglomeration of the nanoparticles. After sonication, no significant improvement in the particle size distribution was observed, as the first peak is thinning around 23 nm (93.6%) and the mean of the second one is slightly decreasing to 72 nm. The solution appeared fairly stable under mechanical treatment. This stability could come from the chemical surface termination of the nanoparticles in the solution which has been discussed elsewhere [14] and this termination is rather not influenced by the sonication step. Negative lithographic patterning was then performed using a 2 µm negative photoresist layer in the standard processing conditions including a post-bake processing above 100 °C. Meanwhile, sonication was undertaken in ethanol suspension containing ultradispersed diamond particles (UDD). The “Diamond Nano-Powder” obtained from Gansu Lingyun Nano-Material Co., Lanzhou was used as purchased. The experimental procedure used for the dispersion and the seeding has been previously reported [15]. Ethanol was used in order to obtain a rapid and uniform evaporation of the solvent without agglomeration. The characteristics of this solution was investigated using Malvern Zeta sizer 300 S. In order to disperse the particles, prior to treatment we applied a sonication treatment consisting of exposing the slurry to ultrasonic excitation at a power of 225 W during 10 mn. Prior to nanoseeding, the wafer was cleaned using spin coating with an IPA solution. Then the wafer was exposed to the solution containing ultradispersed diamond particles (UDD) for one minute. The wafer is then dried using spincoating in order to obtain an homogeneous evaporation of www.pss-a.com
2127
the solution. Then the photoresist was removed using concentrated nitric acid (fuming). The substrate with these patterns was then used for CVD diamond growth. The growth was carried out for 10 h at 450 °C with a microwave power of 1.1 kW and a 60 mbar pressure using 0.8% methane at 2000 ppm boron in H2 at a total flow rate of 100 sccm. The temperature of the sample is controlled using a heated temperature controlled sample holder. After 10 hours of growth a film thickness of 200 nm was reached as measured by the UV interferometry. Selective growth of diamond film is clearly visible by SEM as well as optical dark field imaging (Fig. 1). The observation in dark field is associated with an increase in roughness due to the presence of nanocrystals on the substrate surface. The pattern then appears much more precise and dense than using any other techniques including selected Bias Enhanced Nucleation (BEN) [16] on SiO2. The origin of the adhesion of the particles onto the substrate was discussed in Ref. [15]. We have observed that this patterning was chemically resistant to organic solvents and to nitric acid. On the contrary ultrasonic waves create a few defaults in the layer, and any contact with the surface (tweezers, cloth) has shown to be very detrimental to the nanoparticles coatings as it suffices to remove the nanoseeding layer. The temperature and the conditions at the beginning of the growth have been seen as extremely critical. Usually, during the first step, pure hydrogen is used to reach the required substrate temperature. However, such step, either completely or partially, has an effect of removing the nano-
Figure 1 (online colour at: www.pss-a.com) SEM view of the nanoseeded pattern on silicon using the SEM and the dark field, after CVD diamond growth. © 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
solidi
status
physica
pss
a
2128
M. Bonnauron et al.: Diamond MEAs
tions described above, no covalent link were created between the diamond particles and the glass substrate. A precise UDD seeded patterns, clearly visible in bright and dark field, are shown in Figs. 2 and 3. In the regions protected by the photoresist, expected to be free from diamond growth, a few sites are still observable, probably due to surface contamination.
Figure 2 Dark field observation of nanoseeded diamond onto glass before growth.
powder layer by etching. Therefore, the hydrogen plasma step has been replaced by an hydrogen/methane plasma step. Finally, the microwave power has to be kept low during the first minutes of exposure to avoid particle etching by too dense plasmas. Detailed work on the stability of the nanoparticles in microwave plasma has been recently reported in Ref. [17]. For example, on silicon it has been seen that Si–C bonds are created, whether this route is hardly possible on glass: XPS analysis on glass has demonstrated that after a short growth step, in the condi-
3 3D MEA tip coating Our seeding techniques have been validated on silicon and they were also applied to glass for 3D MEA tip coating. Experiments were then conducted on tips 70 micrometer high and with a base of 50 µm. To fabricate such MEAs on glass, the nanoseeding is even more difficult to be achieved, but it has been made feasible as shown in Fig. 2. Since glass is an insulator it was difficult to get SEM image at high resolution, due to a charge build up on the insulating substrate. After growth however, optical microscopy images show a very conformal coverage of the tip with a remarkable selectivity between exposed and protected areas, as observed in Fig. 3. The diamond coating on one 3D tip is shown in Fig. 4, where the characteristic morphology of CVD diamond is well visible. In comparing this work with the previous one on mirror polished silicon, the protected area shows higher density of diamond nanocrystals. This effect is due to the roughness of the glass which was partially etched during the CVD process and/or to effects associated with the original nucleation site density. This roughness promotes spontaneous nucleation. 4 Conclusion We have demonstrated the possibility to fabricate selectively patterned diamond coatings down to the micrometer scale both on silicon and on glass substrates using the DND patterning. This technique was successfully applied to the realization of 3D diamond elec-
Figure 3 (online colour at: www.pss-a.com) Diamond lines (3.2 µm in thickness) as grown by CVD on glass after nanoseeding. © 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 4 SEM view of 3D glass tip covered with diamond. Here from the top of the tip, the picture focuses onto the right side of the tip, the tip end being thus located further left to the observed area of the picture. www.pss-a.com
Original Paper phys. stat. sol. (a) 205, No. 9 (2008)
trodes on glass in order to fabricate a micro electrode array of 256 electrodes in a surface of 28.8 mm2. The realisation of this first prototype was possible due to the optimisation of a low temperature growth processing route for diamond synthesis. Acknowledgements Although not exclusively driven by the project, the work benefited from the DREAMS project support within the sixth European framework (FP6-NMP-2006676033345) as well as from the MEDINAS project within the French ANR TECSAN scheme for financial support in the CEALIST laboratory for the improvement of the diamond electrode material quality. Further, the authors also wish to acknowledge Dr B. Yvert, from CNIC (CNRS Bordeaux) for useful discussions. Special Acknowledgements This article has been submitted by Mathias Bonnauron’s colleagues, in the memory of his work on this topic, himself having deceased on February 5th, 2008 in a ski accident.
References [1] G. Charvet, L. Rousseau, O. Billoint, J. Rostaing, V. Perrais, S. Joucla, S. Gharbi, M. Trevisiol, C. Moulin, F. Goy, B. Mercier, M. Colin, H. Fanet, P. Meyrand, R. Guillemaud, and B. Yvert, Soc. Neurosci. Abstr. 694.19 (2006). [2] D. A. Henze, Z. Borhegyi, J. Csicsvari, A. Mamiya, K. D. Harris, and G. Buzsáki, J. Neurophys. 84(1), 390 – 400 (2000). [3] C. G. Specht, O. A. Williams, R. B. Jackman, and R. Schoepfer, Biomaterials 25, 4073 – 4078 (2004). [4] L. Bacakova, J. Nanotechnol., in press (2008).
www.pss-a.com
2129
[5] M. Kopecek, L. Bacakova, J. Vacik, F. Fendrych, V. Vorlicek, I. Kratochvilova, V. Lisa, E. VanHove, C. Mer, P. Bergonzo, and M. Nesladek, to be published in phys. stat. sol. [6] J. de Sanoit and E. Vanhove, Patent pending, FR 0755485. [7] A. Krueger, Adv. Mater. 2008, accepted for publication. [8] A. Krüger, M. Osawa, G. Jarre, Y. Liang, J. Stegk, and L. Lu, phys. stat. sol. (a) 204, 2881 (2007). [9] V. Pichot, E. Fousson, M. Comet, C. Baras, F. Le Normand, and D. Spitzer, Dispersion of Detonation Nanodiamonds in a Liquid Medium, in: Proc. Carbon 2007 meeting, Seattle 2007. [10] Chia-Fu Chen, Sheng-Hsiung Chen, Tsao-Ming Hong, and Ming-Hsing Tsai, J. Appl. Phys. 77(2), (1995). [11] Yongqing Fu, Hejun Du, and Jianmin Miao, J. Mater. Proc. Technol. 132, 73 – 81 (2003). [12] Yukihiro Sakamoto, Matsufumi Takaya, Hiroyuki Sugimura, Osamu Takai, and Noboyuki Nakagiri, Diam. Relat. Mater. 8, 1423 – 1426 (1999). [13] Hongwu Liu, Chunxiao Gao, Guangtian Zou, Xun Li, Chengxin Wang, and Chao Wen, Jpn. J. Appl. Phys. 39, 1323 (2000). [14] J. C. Arnault, S. Saada, O. A. Williams, K. Haenen, P. Bergonzo, M. Nesladek, R. Polini, and E. Osawa, accepted for publication in phys. stat. sol. (2008). [15] O. Williams, O. Douheret, M. Daenen, K. Haenen, E. Osawa, and M. Takahashi, Chem. Phys. Lett. 445, 255 – 258 (2007). [16] M. Bonnauron, S. Saada, L. Rousseau, and P. Bergonzo, accepted for publication in Diam. Relat. Mater. (2007). [17] J. C. Arnault, S. Saada, M. Nesladek, O. A. Williams, K. Haenen, P. Bergonzo, and E. Osawa, Diam. Relat. Mater., doi:10.1016/j.diamond.2008.01.008.
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
