Vacuum 86 (2012) 603e607
Contents lists available at ScienceDirect
Vacuum journal homepage: www.elsevier.com/locate/vacuum
Novel plasma treatment in linear antenna microwave PECVD system Neda Neykova a, b, *, Halyna Kozak a, Martin Ledinsky a, Alexander Kromka a a b
Institute of Physics, Academy of Sciences of the Czech Republic, v.v.i., Cukrovarnicka 10, 162 53 Prague 6, Czech Republic Czech Technical University in Prague, Faculty of Nuclear Sciences and Physical Engineering, Trojanova 13, 120 00 Prague 2, Czech Republic
a b s t r a c t Keywords: Nano-crystalline diamond Low temperature hydrogen termination Surface conductivity FTIR SU-8
This work reports on hydrogen termination of nano-crystalline diamond films and the behavior of polymer SU-8 as passivating layer after plasma treatment performed at low temperature in a novel linear antenna microwave plasma enhanced system. Nano-crystalline diamond films were grown by microwave plasma enhanced chemical vapor deposition and then hydrogen terminated at different substrate temperatures. The results indicate that a temperature as low as 200 C is sufficient to reliably attain a diamond surface conductivity of the order of 107 (U/,)1. An increase in substrate temperature up to 400 C results in an increase in surface conductivity up to 1.7 106 (U/,)1. The structural changes of the SU-8 passivating layer, before and after plasma treatment, were investigated by FTIR spectroscopy. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction Diamond is a promising material for advanced devices in bioand electronic applications [1e3]. Considering its extreme electronic, optical, and thermal properties [4], diamond can be specified as ideal material for fabrication of high performance electronic devices [5e7]. Undoped diamond is generally considered to be a very good insulator, but it also exhibits p-type surface conductivity (SC) and negative electron affinity when it is terminated by hydrogen [8e10]. The typical SC value obtained for monocrystalline diamond (MCD) is in the order of 104 (U/,)1 [11]. Presently, nano-crystalline diamond (NCD) shows good enough SC values (106 (U/,)1), suitable for fabrication of electronic devices [12,13]. Moreover, NCD is a more desirable material considering its low cost and possibility of deposition on large areas [14]. To achieve hydrogen termination, NCD films are typically exposed either to hydrogen microwave (MW) plasma or to atomic hydrogen produced by a hot filament source. Both methods are commonly used at relatively high substrate temperatures (Tsub 600 C) [12]. Decreasing this temperature in a standard microwave system is no simple task, while the main working principle is based on igniting a plasma ball close to the substrate (i.e. approximately 1e2 mm). This arrangement results in a high thermal load of the substrate surface in the range of 400e600 C.
* Corresponding author. Institute of Physics, Academy of Sciences of the Czech Republic, v.v.i., Cukrovarnicka 10, 162 53 Prague 6, Czech Republic. Tel.: þ420 220318551. E-mail address:
[email protected] (N. Neykova). 0042-207X/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.vacuum.2011.07.055
However, the treatment at such temperatures is an undesired technological step in the fabrication of electronic devices, because of partial or complete damage of the metal electrodes or other electronic parts [15]. Therefore, hydrogen termination at low temperature is essential. Microwave based surface wave-discharge system in linear antenna arrangement represents an alternative plasma source for low temperature processing [16,17]. The main advantage of linear antenna microwave plasma system is a larger distance between the high-density plasma region and the sample (50e100 mm) [18]. Thus, overheating of the substrate from plasma radiation is minimized. Another priority of such system is its ability to ignite stable high-density plasma down to low pressures of 102 mbar [19]. Additional advantage is its simple scaling-up, i.e. the antenna length could be prolonged up to one meter and the antenna could be multiplied [20,21]. The present work reports on the effect of hydrogen termination on the electrical properties of nano-crystalline diamond films as well as on the behavior of passivating layer (SU-8, 8 mm) after the hydrogen plasma treatment performed at low temperatures in the linear antenna microwave plasma enhanced chemical vapor deposition (PECVD) system. The influence of H-termination on the induced surface conductivity is investigated by I-V measurements. The structural changes of the SU-8, before and after plasma treatment, are characterized by FTIR spectroscopy. 2. Materials and methods NCD films were deposited on Si/SiO2 substrates (10 10 mm2) from a gas mixture of methane and hydrogen in a MW PECVD
604
N. Neykova et al. / Vacuum 86 (2012) 603e607
Fig. 1. Schematic drawing of sample sets: (a) 100% SU-8, (b)50% SU-8 and (c) 0% SU-8.
reactor using an ellipsoidal cavity resonator (Aixtron P6, Germany) [22]. Prior to the deposition process, the substrates were seeded in a suspension of nano-diamond powder (5 nm) using an ultrasonic bath. Details on the nucleation parameters can be found elsewhere [23,24]. The NCD film deposition was performed under the following conditions: hydrogen gas flow 300 sccm, methane gas flow 3 sccm, total vacuum pressure 50 mbar, deposition power 2500 W, and substrate temperature 800 C. To investigate the hydrogen termination of NCD films and characterize its influence on the diamond surface conductivity, interdigital metal contacts (IDCs) were fabricated on the NCD surface. The IDCs were prepared as six electrodes in each patterned plurality with a separation distance of 250 mm. The NCD samples with IDCs were exposed to oxygen plasma (300 W, 3 min), to achieve electrically insulating surfaces (resistivity >10 GU), defined as a starting point for further experiments. To investigate the polymer stability during the hydrogen plasma treatment, SU-8 3050 resist material was used in the experiments (MicroChem Corporation, Germany). The SU-8 3050 resist consists of multifunctional, highly branched polymeric epoxy resin dissolved in an organic solvent, cyclopentanone (CP). For our experiments, three different sets of samples were used: 0% SU-8 (bare NCD), 50% SU-8 (NCD half-passivated with SU-8) and 100% SU-8 (NCD fully-passivated with SU-8). Schematic illustration of the different sets is presented in Fig. 1. The samples with 50% SU8 and 100% SU-8 were prepared by spin-coating of NCD films with 8 mm layer of SU-8 and consequently baked in an oven at 95 / 20 min to remove the solvent from the resist material. Quartz substrate with IDCs, fully passivated by SU-8, was used for reference sample. Then, all NCD structures were exposed to hydrogen plasma at the same conditions (microwave power 1000 W, vacuum pressure 0.1 mbar, 100 sccm of hydrogen flow, and processing time 30 min), only the substrate temperature varied from 150 to 400 C for the
samples without passivating layer (0%SU-8), and from 200 to 300 C for the samples with passivating layer (50%SU-8 and 100% SU-8). The hydrogen termination of the samples was carried out in a modified linear antenna MW plasma system (AK 400, Roth and Rau). Schematic drawing of the system is shown in Fig. 2. The linear antenna MW plasma system is a commercially available apparatus used for solar cell technology, which consists of a vacuum chamber equipped with two metal antennas (copper) located in quartz tubes. The system makes use of two microwave generators (2.45 GHz, MX4000D, Muegge) working at pulse-frequency up to 500 Hz and maximum power up to 4.4 kW in pulse at each side of the linear antenna (conductor). In this apparatus, plasma propagates along the two antennas and is generated along the outer part of the quartz tubes. The substrate stage, which is water cooled and up-down moveable, is located below the antenna. The surface conductivity of hydrogen-terminated NCD patterns was characterized by current-voltage (I-V) characteristics. Electrical measurements were performed at ambient conditions, i.e. atmospheric pressure and room temperature, with a DC bias swept in the range from 1.5 to 1.5 V using the Keithley 237 source-measure unit. The bias voltage sweeping rate was 100 mV/s. Surface morphology and grain size of the deposited NCD films were investigated by scanning electron microscopy (SEM, Raith GmbH, e_LiNE writer). The diamond character of NCD was investigated by Raman spectroscopy using Renishaw In Via Reflex Raman spectrometer with the excitation wavelength of 325 nm. Structural analysis of SU-8 before and after the hydrogen plasma treatment was performed by the interference-free reflectanceabsorbance spectroscopy of p-polarized IR light at Brewster angle of incidence. The optical infrared absorbance spectra in the spectral region were acquired using a Fourier Transform Infrared (FTIR) spectrometer equipped with MCT detector cooled to 77 K. The spectrometer was purged with dry air to eliminate the optical absorption related to the atmospheric water and CO2. The
Fig. 2. Schematic drawing of linear antenna microwave PECVD system.
N. Neykova et al. / Vacuum 86 (2012) 603e607
605
Fig. 3. SEM image (a) and Raman spectra (b) of NCD film deposited on Si/SiO2 at 800 C.
measurements were done using the Monolayer/Grazing Angle Accessory (Specac Ltd.) suitable for measurements with variable angle of incidence (8e85 ) and equipped with ZnSe polarizer. The absorbance spectra were evaluated as A ¼ log(S/B)c, where S is the spectrum of the measured sample, B-the reference spectrum of the sample with clean surface, and c-the baseline correction. 3. Results
SSC was calculated by indirect method based on Ohm’s law (Eqs. (1) and (2))[25]:
ss ¼
1
(1)
rs
rs ¼ Rs
D UD ¼ N L Is L
(2)
Fig. 3a shows a representative SEM image of the NCD film deposited on Si/SiO2 at 800 C for three hours. The thickness of the film was 600 nm, as determined from a cross-sectional view of the SEM image. The top view depicts a grain size up to 250 nm. Raman spectra (Fig. 3b) confirmed the formation of diamond phase (sp3) with the characteristic diamond line centered at 1332 cm1. A broad band centered at around 1590 cm1 was also observed, which is related to the non-diamond phase (sp2 bonded carbon atoms) [12]. Fig. 4 represents the plot of I-V characteristics for NCD samples with IDCs structures, which were hydrogen-terminated at different substrate temperatures. All I-V characteristics are symmetric and linear, indicating their ohmic character. As expected, the gradient of current curves increases with increasing substrate temperatures. The highest current value observed in this set of data is 55 mA (measured at 1 V) for the H-terminated sample at a substrate temperature of 400 C. However, it is not simple to identify the origin for I-V characteristics, described above. Therefore, we calculated the sheet surface conductivity (SSC) as a function of substrate temperature.
where ss is the sheet surface conductivity [(U/,)1], rs is the sheet surface resistivity [U/,]. Surface resistance Rs[U], on the other hand, is determined by the ratio of DC voltage U[V] drop per unit length L[mm] to the surface current Is [A] per unit width D[mm], multiplied the number of electrodes N (Eq.2). Logarithmic dependence of calculated SSC of NCD film measured at 1 V upon substrate temperature is plotted in Fig. 5. SSC increases with increasing substrate temperature. The highest value of SSC is 1.7 106 (U/,)1 (1 V/400 C). The sample hydrogenated at 150 C exhibits the lowest sheet surface conductivity in the range of 1011e1010 (U/,)1. Fig. 6 presents the values of SSC for quartz (100% SU-8) NCD covered with 100%, 50% and 0% SU-8, hydrogenated at 200 C. The highest SSC is observed at the sample with 0% SU-8, i.e. bare NCD film (4.9 107 (U/,)1). The lowest value of SSC is measured for 100% SU-8 (in order of 1012e1013 (U/,)1), which is equal to noise level of our measurement setup. The quartz substrate was used as a reference sample to avoid influence of polymer conductivity and/or thermally induced transition of SU-8 into a conductive NCD film. The IR spectra of SU-8 before and after hydrogen plasma processing are shown in Fig. 7. Obtained spectra of SU-8 are in good
Fig. 4. Current-voltage characteristics of NCD films hydrogenated at different substrate temperatures (150, 200, 300, and 400 C).
Fig. 5. Calculated sheet surface conductivity of NCD films measured at 1 V as a function of substrate temperature.
606
N. Neykova et al. / Vacuum 86 (2012) 603e607
Fig. 6. Comparison of the sheet surface conductivities for the samples passivated by SU-8 polymer and hydrogen-terminated at 200 C.
agreement with those reported in [26]. All spectra show comparable features and similar band structure. Nevertheless, it was observed that the IR peak of the epoxide ring mode at 914 cm1 was reduced in the intensity after H-plasma treatment at 200 C and completely disappeared after hydrogen plasma treatment at 300 C. Additionally, it was observed that the intensity of IR peaks in SU-8 decreased after H-treatment at 300 C. More information about SU-8 chemical properties can be found in [27]. 4. Discussion It is well known that the hydrogen termination of oxygen terminated NCD film requires appropriate energy, which is first used for breaking the CeO bonds, and then for attaching H-atoms to the free C-atom. Subsequently, several chemical reactions are involved in the hydrogenation process. When the substrate temperature increases, the probability of replacing the O-atoms with H-atoms increases too. Thus, efficiency of hydrogen termination will increase, which is in good agreement with our results. The SSC values of NCD films rise on increasing the heating temperature from 150 to 400 C. The breaking heating temperature is found to be 200 C, at which NCD surface exhibits high enough conductivity (4.9 107 (U/,)1 at 1 V), which is close to the value required for fabrication of electronic devices. The sample hydrogenated at 400 C exhibits the highest surface conductivity (1.7 106 (U/,)1 at 1 V). The obtained results from FTIR spectroscopy showed that SU-8 undergoes structural changes for hydrogen plasma treatment at 300 C, expressed as a decrease in the intensity and disappearance of some peaks. Moreover, partial etching and degradation of the
polymer film were traced by eye. The degradation temperature of SU-8 is around 380 C [28]. Possible reason for etching and degradation of the polymer film at 300 C can be the synergetic influence of plasma and heating in our experimental setup. During plasma treatment, the top surface of the polymer undergoes activation and recombination with H-atoms from the gas plasma. These reactions together with additional heating can lead to a local increase of polymer temperature at its top sub-surface by more than 300 C. However, no significant changes in the FTIR spectra of the sample treated at 200 C compared to untreated sample were observed. From optical microscopy observations, no damage or cracking of SU-8 and Au contacts were observed. Moreover, the Htermination at 200 C was effective enough to obtain diamond films with good surface conductivity. All presented results confirm that the hydrogen termination at low temperature as low as 200 C is efficient enough to induce hydrogen-terminated conductive surfaces, and at the same time to be “friendly” to metal contacts and to polymer-based passivating layer. 5. Conclusions The hydrogen termination of NCD with IDCs was carried out in a temperature range of 150e400 C in a novel linear antenna MW plasma system. The influence of substrate material, diamond morphology and surface conductivity induced by hydrogen treatment were investigated. The highest surface conductivity of 1.7 106 (U/,)1 at 1 V was reached for NCD with IDCs hydrogenterminated at 400 C. Nevertheless, the samples hydrogen treated at lower temperature (200 C) were still sufficiently conductive (in the order of 107 (U/,)1) to be used in the semiconductor industry, for example as a gas sensor. Additionally the results from FTIR spectroscopy proved that no significant change in polymer structure was observed for hydrogen termination at 200 C. These results confirmed that low temperature hydrogen termination by novel linear antenna microwave PECVD system is a promising alternative for large area hydrogen termination without damaging or destroying metal contacts and/or passivating polymer layers. In addition, the process of decreasing the temperature of hydrogen termination is expected to significantly expand the family of electronic applications. Acknowledgments This work was supported by the grants IAAX00100902, KAN400100701, SGS No. 10/297/OHK4/3 T/14, project No. LC510, by the Institutional Research Plan No.AV0Z10100521, by the J. E. Purkyne Fellowship and by the Grant Agency of ASCR. References
Fig. 7. IR absorbance spectra of SU-8 before and after the hydrogen plasma treatment at 200 and 300 C.
[1] Kalbacova M, Kalbac M, Dunsch L, Kromka A, Vane cek M, Rezek B, et al. Phys Status Solidi 2007;B11:4356e9. [2] Rezek B, Watanabe H, Shin D, Yamamoto T, Nebel CE. Diamond Relat Mater 2006;15:673e7. [3] Bacakova L, Grausova L, Vacik J, Fraczek A, Blazewicz S, Kromka A, et al. Diam Relat Mater 2007;16:2133e40. [4] Shenderova OA, Gruen DM. Ultrananocrystalline diamond: synthesis, properties and applications. New York: William Andrew Publishing; 2006. [5] Kubovic M, Kasu M, Kallfass I, Neuburger M, Aleksov A, Koley G, et al. Diam Relat Mater 2004;13:802e7. [6] Kubovic M, Janischowsky K, Kohn E. Diam Relat Mater 2005;14:514e7. [7] Kubovic M, Kasu M, Yamauchi Y, Ueda K, Kageshima H. Diam Relat Mater 2009;18:796e9. [8] Landstrass MI, Ravi KV. Appl Phys Lett 1989;55:975e7. [9] Takeuchi D, Riedel M, Ristein J, Ley L. Phys Rev; 2003. B68:041304(R). [10] Looi HJ, Whitfield MD, Foord JS, Jackman RB. Thin Solid Films 1999;343: 623e6. [11] Rezek B, Nebel CE. Diam Relat Mater 2006;15:1374e7.
N. Neykova et al. / Vacuum 86 (2012) 603e607 [12] Kozak H, Kromka A, Ukraintsev E, Houdkova J, Ledinsky M, Vanecek M, et al. Diam Relat Mat 2009;18:722e5. [13] Kozak H, Kromka A, Ledinsky M, Rezek B. Phys Status Solidi A 2009;206:276e80. [14] Potocky S, Kromka A, Potmesil J, Remes Z, Vorlicek V, Vanecek M, et al. Diam Relat Mater 2007;16:744e7. [15] Neykova N, Babchenko O, Kozak H, Kromka A, Hruska K, Vanecek M. In: Pudis D., editor. Proceedings of 15th International Conference on Applied Physics of Condensed Matter, 2009 June 24-26, Bystra, Liptovsky Jan, Slovak Republic, p.256e260. [16] Omaru T, Komiyama F, Kogoshi S. Jpn J Appl Phys 2004;43:2690e2. [17] Xu J, Fan H, Kousaka H, Umehara N, Liu W. Surf Coatings Technol 2007;201:6631e4. [18] Kromka A, Babchenko O, Izak T, Verveniotis E, Rezek B. In:Vesesly M, Vincze A, Vavra I. editors. Programme and book of abstracts of 13th Joint Vacuum Conference; 2010 June 20e24; Strebské pleso, Slovak republic, p.70. [19] Schlemm H, Mai A, Roth S, Roth D, Baumgartner KM, Muegge H. Surf Coatings Technol 2003;174e175:208e11.
607
[20] Tsugawa K, Ishihara M, Kim J, Hasegawa M, Koga Y. New Diamond Frontier Carbon Technol 2006;16:337e45. [21] Kim J, Tsugawa K, Ishihara M, Koga Y, Hasegawa M. Plasma Sources Sci Technol 2010;19:015003e7. [22] Füner M, Wild C, Koidl P. Appl Phys Lett 1998;72. pp.1149e1151. [23] Kromka A, Rezek B, Kalbacova M, Baresova V, Zemek J, Konak C, et al. Adv Eng Mater 2009;11:B71e9. [24] Balon F, Stolojan V, Silva SRP, Michalka M, Kromka A. Vacuum 2005;80: 163e7. [25] Maryniak WA, Uehara T, Noras MA., Surface Resistivity and Surface Resistance Measurements Using a Concentric Ring Probe Technique (Trek Application Note 1005, 2003), Available from: http://www.trekinc.com/pdf/1005_ Resistivity_Resistance.pdf. [26] Tan TL, Wong D, Lee P, Rawat RS, Patran A. Appl Spectrosc 2004;58:1288e94. [27] LaBianca N, Gelorme JD. Proc SPIE-Int Soc Opt Eng 1995;2438:846e69. [28] www.microchem.com/products/su_eight.htm.
