Available online at www.sciencedirect.com Diamond & Related Materials 17 (2008) 1194 – 1198 www.elsevier.com/locate/diamond Characterization of UV irradiated nanocrystalline diamond G. Speranza a , S. Torrengo b , L. Minati a , M. Filippi a , M. Castellino b , Cl. Manfredotti b , Ch. Manfredotti b , M. Dipalo c , A. Pasquarelli c , E. Kohn c , Hayssam El-Hajj c , E. Vittone b,⁎ b a Physics and Chemistry of Surfaces and Interfaces, Fondazione “Bruno Kessler”, Trento (I), Italy Department of Experimental Physics and “Nanostructured Interfaces and Surfaces” NIS Centre of Excellence, University of Torino, via P.Giuria 1, 10125 Torino (I), Italy c Department of Electron Devices & Circuits, Ulm University (D), Germany Available online 11 February 2008 Abstract The exposure of H-terminated nanocrystalline diamond (NCD) to ultraviolet (UV) light in air and at room temperature modifies the features of the diamond surface, in terms of wettability, electrical conductivity and chemical reactivity. This allows the development of a soft, dry and non invasive method to tailor the surface properties for the development of new chem- or bio-sensors. In this work, we report about the analysis of hydrogen terminated nanocrystalline diamond films grown by hot-filament technique and of their surface modification following UV light exposure. This UV treatment induces an hydrophobic to hydrophilic transformation and an outstanding increase of the electrical resistivity (N 105). X-ray (XPS) and ultraviolet (UPS) photoelectron spectroscopy provide insight into the role of chemisorbed oxygen on the modification of the valence band and on the transformation of the electron affinity (from negative to positive) at the diamond surface. © 2008 Elsevier B.V. All rights reserved. Keywords: Nanocrystalline diamond; Surface properties; Electron affinity; Hydrogen termination; Oxygen termination 1. Introduction Nanocrystalline diamond, because of its nanostructured surface, mechanical, optical, electrical/electrochemical properties, as well as for its biocompatibility combined with the possibility of controlled surface bio-modification, is considered a suitable platform for biointerfaces and biosensors [1]. Two surface terminations are usually considered: H-terminated diamond surfaces are biocompatible, transparent, electrically conductive and biomolecules can be attached covalently at the surface through photochemical processes [2]; oxygen termination induces an optically transparent and hydrophilic surface which promotes cell adhesion and growth [1]. Among the various methods to tailor the electrical properties of the diamond surface, the exposure to ozone has recently attracted the interest of various researchers for fundamental studies [3], for the fabrication of highly sensitive chem-FET [4], or for cell adhesion and biocompatibility studies [1]. Such ⁎ Corresponding author. E-mail address: vittone@to.infn.it (E. Vittone). 0925-9635/$ - see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.diamond.2008.01.098 treatment can be produced by the exposure of diamond to ultraviolet (UV) radiation, even if controversial data are reported on the synergic combination of UV and ozone on the surface treatment [5]. In this paper we report about the effects of UV irradiation in air on the surface electronic characteristics of NCD. The chemisorbed species, in particular the role of oxygen, as well as the modification of the valence band, were monitored by XPS and UPS in order to evaluate the photochemical effects of the UV treatment. To assess the role of the surface photochemical modification on electrical and hydrophobic properties of the surface, XPS/UPS analyses have been correlated to conductivity and contact angle measurements. 2. Experimental 2.1. Diamond samples NCD samples preparation consisted of 3 steps. Firstly a 4” silicon wafer (100) was nucleated in Hot Filament CVD by mean of BEN (Bias Enhanced Nucleation) [6]. Subsequently a G. Speranza et al. / Diamond & Related Materials 17 (2008) 1194–1198 1195 Table 1 Process parameters for bias enhanced nucleation (BEN), diamond growth and hydrogen termination Pressure (Pa) Gas flow (sccm) Methane (%) Microwave Power (W) HF Power (W) Filament temperature (K) Substrate temperature (K) Process time BEN Growth H-termination 2000 400 0.5 2000 400 0.3 4000 200 1 700 4800 2290 1000 2h 4800 2290 970 30 h 920 5 min NCD film was grown onto the nucleated silicon wafer by means of Hot Filament CVD, the growth parameters being described in Table 1. In the end nanocrystalline diamond samples (1 × 1 cm2 surface area) were cut from the overgrown silicon wafer. The average thickness was 5.3 μm with a dispersion of 0.4 μm as evaluated by optical reflection interferometry. The morphology of the sample surfaces was determined using a PSIA XE-100 Atomic Force Microscope (Fig. 1a). The root mean square roughness of the samples surfaces, evaluated by analyzing several 5 × 5 μm2 AFM maps, was (56 ± 3) nm; the maximum grain size was in the sub-micrometer range as evaluated from SEM (Fig. 1b). Fig. 2. Red and UV Raman spectra of a NCD sample. The peak at 520 cm− 1, relevant to the Si substrate, is visible only with red laser excitation because of the low absorption coefficient. Fig. 2 shows Raman spectra measured using a 244 nm (UV) and a 632 nm (red) laser excitation. Red laser Raman indicates a very poor transpolyacetylene contamination in the film (no peaks at 1150 and 1480 cm− 1 [7]), stating the high quality of the NCD diamond samples as is demonstrated also by the first order diamond Raman line which is neither hidden by the strong luminescence nor by the sp2 broad band around 1520 cm− 1. The UV Raman spectra show a very pronounced diamond peak at 1333 cm− 1 (FWHM ≈ 5 cm− 1) and a broad G-band around 1569 cm− 1. Due to the much higher absorption coefficient of 244 nm light [8], UV Raman spectroscopy is much more surface sensitive than the previous one, even if a comparison of the two spectra has to take into account the different relative Raman cross-section for the non-diamond carbon band [9]. 2.2. Post-growth processes After the growth process, all the diamond samples were cleaned in a 1:2 H2O2:H2SO4 solution (piranha) and hence were hydrogen terminated in a microwave assisted hydrogen plasma with the parameters listed in Table 1. After the H plasma treatment, the sample was allowed to cool down to room temperature under H2 flux for about 90 min. Three hydrogen plasma diamond samples (H-NCD) were finally exposed in air at room temperature to UV irradiation (UV-NCD). A Xe arc lamp (400 W) was used to irradiate for 3 h one sample for XPS/UPS measurements; the other two samples, used for electrical/wettability measurements, were exposed at a distance of 3 cm to UV light produced by a deuterium lamp (30 W). In both cases, ozone was clearly detected in the environment and approximately the same exposure (i.e. the time integral of irradiance) was adopted for all the samples. 2.3. Surface characterization Fig. 1. (a) AFM map and (b) SEM image of a NCD surface. XPS was carried out using a Scienta ESCA 200 instrument equipped with a monochromated Al Kα source and a 200 mm 1196 G. Speranza et al. / Diamond & Related Materials 17 (2008) 1194–1198 Fig. 5. He I UPS spectra of H-terminated and UV-treated diamond. Fig. 3. XPS wide scan spectra of H-terminated and UV-treated diamond. The binding energy (BE) scale was calibrated with respect to the Ag Fermi edge. The inset shows detailed scans around the C1 s peak. hemispherical analyzer, with a pass energy of 150 eV (energy resolution 0.3 eV). UPS spectra were recorded in a Physical Electronics (PHI 590/550) spectrometer equipped with a helium discharge lamp (HeI, hν = 21.2 eV and HeII, hν = 40.8 eV) and with a double-pass cilyndrical analyzer (energy resolution ~ 0.5 eV). Before XPS or UPS characterisation of the H-NCD, the sample was annealed in UHV for 30 min at 400 °C in order to remove main surface contaminants. No thermal treatments were adopted after UV treatment. 2.4. Electrical and contact angle measurements Electrical measurements were carried out after several days of air exposure from the H-plasma process, through a collinear fourpoint probe head (radii = 0.04 mm, relative distance = 0.635 mm) with tungsten carbide tips was used [10]. Fig. 4. XPS valence band spectra of H-terminated and UV-treated diamond. The continuous line is reported only for highlight the presence of the O2 s spectral feature. Measurements of wettability of the diamond surface were performed using the sessile drop method on a DATAPHYSICS OCA 20 contact angle meter. A 2 µl drop of ultra pure water and CH2I2 was placed onto the prepared samples surface, taking three different measurements for each sample. 3. Results Fig. 3 shows XPS survey spectra of H-NCD and UV-NCD surfaces. Both spectra present a dominant C1 s peak located at 285 eV and no detectable surface contamination, except for oxygen. The increase of the O/C atomic ratio (from 1.7% to 6%, for H-terminated and UV-treated samples, respectively) as well as the widening in the high energy side of the C1 s peak relevant to the UV-NCD surface, which is attributable to ether (C–O–C), carbonyl (C = O) or carboxyl (O = C–OH) surface contamination, prove that the UV-treatment induces the oxidation of the Fig. 6. He II UPS spectra of H-terminated and UV-treated diamond. G. Speranza et al. / Diamond & Related Materials 17 (2008) 1194–1198 H-NCD surface, as reported by several authors on homoepitaxial or polycrystalline diamonds [3,4]. This is further verified by contact angle (C.A.) measurements which show a remarkable increase of the polar component of the surface energy (σ) from H-NCD (H2O C.A. = 85°, CH2I2 C.A. = 27°, Overall surface energy σ = 45.5 mJ/m2, polar contribution σp = 1.5 mJ/m2) to UV-NCD (H2O C.A. = 58°, CH2I2 C.A. = 29°, σ = 50.9 mJ/m2, σp = 15 mJ/m2). The XPS valence bands for the two surface terminations are shown in Fig. 4. Both spectra present a sharp peak (peak II) at about 13.7 eV which has been considered the fingerprint of the diamond phase, whereas the structures at about 6 (III) and 18 eV (I) have dominant p or s fractional character, respectively [11]. The UV-NCD spectrum shows also a shoulder at 26 eV to be ascribed to O 2 s, whereas O 2p orbitals play a role in the shape definition of the component III. However, it is worth noticing that the valence band spectra are very similar to those relevant to natural or homoepitaxial samples and do not show any noticeable amorphous carbon component, whereas such contribution is evident in UV Raman spectra. This has to be ascribed to the different surface sensivities of the two probes (the extinction length of 244 nm photons in diamond is about 50 nm 0, whereas photoelectrons of 1200 eV kinetic energy leads to a sampling depth of ~ 5 nm). Differently from Raman, which probes also the diamond bulk, XPS does not reveal any graphitic contaminant since H plasma selectively removes sp2 hybridized carbon from the surface. UPS spectra are shown in Figs. 5 and 6. With respect to the XPS valence band, the HeII spectrum of the H-terminated sample shows a well resolved peak at about 8 eV (due to the higher cross section of the C 2p state) and the diamond sp3 peak which occurs at a slightly higher binding energy (14.2 eV). The HeII spectrum of the UV-treated surface shows clear differences to be ascribed to C–O bonds. In particular the O 2p components widens the peak at 8 eV and the barely perceptible shoulder occurring at about 2 eV, attributable to H-terminated surfaces[12], disappears. The high binding energy peak (NEA peak [13]) is more intense and is located at higher binding energies in the HeI spectrum of H-NCD with respect to UV-NCD surfaces. These facts are compatible with the existence of a negative electron affinity (NEA) of the H-terminated surfaces which transforms to a positive electron affinity (PEA) when the surface is UVtreated. In fact, the energy difference between the vacuum level (E0) and the Fermi level (EF) (which is considered the origin of the energy scale in any spectra) can be evaluated from the HeI spectra (Fig. 5) by a linear extrapolation of the NEA peak to zero (Ecut-off), whereas the valence band maximum (EV) can be calculated from the HeII spectra through the linear extrapolation of the onset of emission (Fig. 6). In the H-terminated diamond surface the energy cut-off occurs at 17.3 eV from EF (i.e. E0 −EF = hν(HeI) −Ecut-off = 3.9 eV) and EF −EV = 1.3 eV; consequently, the electron affinity is χ =E0 −EF + (EF −EV) −EG = −0.3 eV. For the UV-treated Ecut-off = 16.7 eV, i.e. E0 − EF = 4.5 eV and EF − EV = 2.6 eV; consequently, χ = +1.6 eV. The different χ value 1197 of the hydrogenated/oxidized diamond leads to a different efficiency in the electron emission form the surfaces as shown by the high BE peak in the HeI spectra in agreement with [14,15]. The electrical measurements carried out on diamond samples kept in air after H plasma treatment yield a sheet resistivity (ρS) of the order of 1 MΩ/□. Assuming a hole mobility of 60 cm2·V− 1·s− 1 0, the sheet hole concentration is about 1011 cm− 2. After UV treatment, a stable and highly resistive surface (ρS N 200 GΩ/□) was observed. Such measurements are compatible with the electrochemical transfer doping model [16] which assume the negative electron affinity of the H terminated surface as an essential requirement for the surface conductivity induced by adsorbed molecules from an atmospheric water. However, from our data, we cannot rule out the hypothesis that the UV treatment can generate an high surface density of hole traps, induced by the oxygen termination. 4. Conclusions High quality NCD samples have been investigated prior and after UV exposure in air and at room temperatures. The UVtreatment induces an oxidation of the surface originally Hterminated. This is verified macroscopically by the change of the surface wettability and on the increase of the O content as evaluated by XPS. UPS spectra show a change of the vacuum level position which moves from − 0.3 eV (in H-NCD) to 1.6 eV (in UV-NCD) with respect to the conduction band minimum. The transformation from NEA to PEA of the diamond surface can be considered one of the fundamental reasons of the outstanding increase of the surface electrical resistivity for samples subjected to intense UV irradiation. Acknowledgements This work was made possible by support from the Regione Piemonte (Progetto D14 “Biosensori e interazione neuronisuperfici nanostrutturate”). Dr. Paola Rivolo is kindly acknowledged for her help in contact angle measurements as well as the CSMFO group of Trento which has provided the UV sources to perform the diamond oxidation. References [1] K.F. Chong, K.P. Loh, S.R.K. Vedula, C.T. Lim, H. Sternschulte, D. Steinmuller, F.S. Sheu, Y.L. Zhong, Langmuir 23 (2007) 5621. [2] A. Hartl, E. Schmich, J.A. Garrido, J. Hernando, S.C.R. Catharino, S. Walter, P. Feulner, A. Kromka, D. Steinmuller, M. Stutzmann, Nat. Matters 3 (2004) 736. [3] M. Riedel, J. Ristein, L. Ley, Phys. Rev. B 69 (2004) 125338. [4] T. Sakai, K.S. Song, H. Kanazawa, Y. Nakamura, H. Umezawa, M. Tachiki, H. Kawarada, Diam. Relat. Mater. 12 (2003) 1971. [5] J. Alvarez, J.P. Kleider, E. Snidero, P. Bergonzo, D. Tromson, C. Mer, Diam. Relat. Mater. 13 (2004) 751. [6] K. Janischowsky, W. Ebert, E. Kohn, Diam. Relat. Mater. 12 (2003) 336. [7] A.C. Ferrari, J. Robertson, Phys. Rev. B 63 (2001) 121405. [8] Z.G. Hu, P. Hess, Appl. Phys. Lett. 89 (2006) 081906. [9] S. Prawer, R.J. Nemanich, Phil. Trans. R. Soc. Lond. A 362 (2004) 2537. [10] F. Fizzotti, A. Lo Giudice, C. Manfredotti, Ch. Manfredotti, M. Castellino, E. Vittone, Diam. Relat. Mater. 16 (2007) 836. 1198 G. Speranza et al. / Diamond & Related Materials 17 (2008) 1194–1198 [11] F.R. McFeely, S.P. Kowalczyk, L. Ley, R.G. Cavell, R.A. Pollak, D.A. Shirley, Phys. Rev. B 9 (1974) 5268. [12] C. Saby, P. Muret, Diam. Relat. Mater. 11 (2002) 851. [13] R.J. Nemanich, P.K. Baumann, M.C. Benjamin, S.P. Bozeman, B.L. Ward, “The Physics of Diamond”, in: A. Paoletti, A. Tucciarone (Eds.), Proceedings of the International School of Physics “Enrico Fermi,” Course, CXXXV, IOS Press, Amsterdam, 1997, p. 537. [14] J.E. Yater, A. Shih, R. Abrams, Phys. Rev. B 56 (1997) 56. [15] J.C. Zheng, X.N. Xie, A.T.S. Wee, K.P. Loh, Diam. Relat. Mater. 10 (2001) 500. [16] J. Ristein, Appl. Phys., A Mater. Sci. Process. 82 (3) (2006) 377.