''Analysis of formation of silicon nitride on Si(100) by electrochemical anodization''

Navigation: Main page » Experiments and Research

 Print this page Submit articleComment article

Author: *S. Zeyrek, H.Yüzer, M.C. Baykul, M. M. Bülbül, Ş. Altındal

Analysis of formation of silicon nitride on Si(100) by electrochemical anodization

*S. Zeyreka, H.Yüzerb, M.C. Baykulc, M. M. Bülbüld, Ş. Altındald

aDepartment of Physics, Faculty of Arts and Sciences, Dumlupınar University,43100, Kütahya -Turkey bMaterial and Chemical Technologies Research Institute, Marmara Research Center, P.O. Box 21, 41470 Gebze-Kocaeli, Turkey cDepartment of Physics, Faculty of Arts and Sciences, Osmangazi University, , Eskişehir -Turkey dDepartment of Physics, Faculty of Arts and Sciences, Gazi University, 06500, Ankara -Turkey

abstract

Using electrochemical anodization technique, the silicon nitride ( Si3N4 )thin film has been deposited on the p-type Si(100) substrate at the ambient temperature. Very thin nitride passivation of Si(100) stabilizes its surface or interface very well and reduces the interface states density. However, the passivated surface exhibits a substantial degradation due to the exposure to air after passivation. In this study, silicon nitride thin film have been deposited via electrochemical anodization on the Si(100) surface. Both analysis of formation of silicon nitride (Si3N4) on the p-Si(100) and substrate Si(100) have been made by Fourier transform infrared (FTIR) spectroscopy, Raman spectrometry, respectively. The Si-N asymmetric stretching characteristic peak of Si3N4 films has been observed by FTIR. Crystalline silicon has been observed by Raman spectrometry. The atomic force microscopy (AFM) have been used to carry the morphological properties of the deposited films. Surface degradation was investigated with the high-resolution X-ray diffraction (XRD) technique.

Keywords:

Electrochemical anodization; Si3N4; FTIR; Raman spectrometry; AFM, XRD. *Corresponding author. Tel.:+90-274-265 20 31 fax: +90-274-265 20 56 E-mail address: szeyrek@dumlupinar.edu.tr (S. Zeyrek)

Introduction

Dielectric thin films (SiO2, Si3N4 ,Al2O3 , AlN) are used for a wide range of applications in tetravalent (Si) and compound semiconductor (GaAs , InP , InGaAs) devices including ion implantation masks, passivation films, optical coatings, and wave guides for both electronic and optoelectronic applications. Recently, low-temperature growth of dielectric films has received a lot of attention for a study of microelectronic components; mainly because low-temperature processes minimize the incongruous loss of the V group element of the III-V semiconductors [1]. Since it was first reported in 1965 by Sterling and Swann [2] that silicon nitride (Si3N4) could be used as surface passivation films in integrated circuit (IC) [2,3], silicon nitride films, in particular, have been widely used in IC processing as dual dielectric gate films, local oxidation masks and surface passivation films due to their superior electronic properties such as high dielectric constant, stability and strong resistance to diffusion. Gate dielectric scaling requires new materials with dielectric constant (K) higher than SiO2 to provide the increased capacitance without compromising gate leakage current. One such material is Si3N4 that has approximately twice the dielectric constant of SiO2 , and additionally is effective in blocking diffusion of dopant such as boron. Si3N4 has also been proposed as an interface layer for high K transition metal oxides since it has a higher K than SiO2 and is also an excellent diffusion barrier [4]. Thin amorphous silicon nitride dielectric layers have found many applications in great variety of electronic devices, including these as antireflective coating (ARC). Due to the large capture cross-section ratio (σn/σp = 100 at mid gap) of minority to majority charge carriers, passivation of the surface in p-type silicon is more crucial than in n-type silicon [5]. The surface stability of semiconductors plays a very important role in the fabrication of electronic devices. The formation of a insulator layer on Si by traditional ways of oxidation or deposition can not completely passivate the active dangling bonds at the semiconductor surface. Recently, nitridation of silicon films have been received much attention because silicon nitride film can suppress both of the leak current in insulating gate materials and interface reaction with metal oxide [5]. Thus, various non-traditional approaches for surface passivation such as ultra-thin sulphide, selenide layer or Si3N4 formation have been a subject of increasing interest in recent years [6,7]. Also, the effect of nitride treatment is considered to be associated with passivation of the dangling bonds with nitride atoms and suppression of surface oxidation of Si. However, it has been reported that the passivated surface of Si was quickly degraded when it was exposed to air ambient [7]. In this study, silicon nitride thin film has been deposited via electrochemical anodization on the Si(100) surface and non-aqueous ammonium polynitride ((NH4)2(NH2)x) electrolyte was employed to growth a silicon nitride layer on the Si surface as a new method for nitride passivation at the room temperature.

Experimental Procedure

In this study 2 inch diameter float zone (100) p-type (boron doped) single crystal silicon wafers having thickness of 280 m with 0.8 cm resistivity was used. The wafers were cleaved into 10x10 mm electrodes. The sample was ultrasonically cleaned in trichloroethylene and ethanol, etched by CP4 (HNO3: HF: COOHC2H5: H2O =3:1:2:2 weight ratio) solution for 30 s., rinsed by propylene glycol and blown with dry nitrogen gas. Ohmic contacts of the electrodes were formed by evaporating Al in high vacuum (P=10-6 Pa) and subsequently annealing them for a few minutes at 450 oC. After making of electrical contact, the walls and under side of the Si wafer were insulated with the high-quality wax. The nitridation set-up used in the study is the electrochemical anodization cell which consists of a p-Si anode and Pt as cathode, and is shown in Fig. 1. Agitation of the electrolyte is achieved by magnetic stirrer. Electrolyte used in the experiment was obtained by sequentially mixing of propylene glycol with ammonia (NH3) and hydrazine (NH2-NH2) at 21:3:1 weight ratio, respectively. Preceding each cleaning step, the wafer was rinsed thoroughly in de-ionized water of resistivity of 18 MWcm. Immediately after that, the substrate was immersed in electrolytical cell. Anodic nitridation was performed using a constant current source at different current densities, under N2 flow, in light and at room temperature (293 K). The potential difference between the electrodes normalised to calomel electrode was measured with an x-t recorder. The anodization was stopped when the cell voltage reached about the 18 V. After, the sample was immediately rinsed in propyl alcohol and blown with dry nitrogen and left in a desiccator. Under constant current conditions, the nitride growth was followed by monitoring the voltage developed across the cell. With this monitoring it is possible to observe nucleation of nitride at the lower current density (0.1 mA/cm2). After the nucleation phase in all of nitridation experiments, the behaviour of the cell voltage with time is linear, indicating the homogeneous composition of the grown nitride layers. The thickness of dielectric film =52 Å was calculated from high-frequency (500 kHz) C-V characteristics using the equation Cox=ioA/, where Cox is the capacitance of the interfacial layer, i=7.5o and o are the permittivities of the interfacial layer and free space, respectively. The Schottky contacts were formed by evaporating of Al dots with diameter of about 1.0 mm and 1500 Å thick in high vacuum (P=10-6 Pa). The metal thickness layer and the deposition rates were monitored with the help of quartz crystal thickness monitor.

Fig. 1. Non-aquenous anodic nitridation set-up

Results and discussions

The H is incorporated in the from of Si-H and O-H groups in the oxides, and Si-H and N-H in the nitrides. Lucovsky et al. [8] have shown by an empirical tight-binding calculation that localized states can be generated within the band gap of SiO2 by Si-H, Si-N, and Si-O-H bonding groups, and that localized states can be generated in Si3N4 by Si-H and Si-N-H groups. In all of these cases, the localized states are more than about 1 eV away from either one of the band edges and therefore can be active as deep trapping and/or recombination centers. Si-N bonds are supposed to reduce the leak current by preventing other atoms from migrating into the film [9]. The hydrogen incorporated in the silicon nitride films reduces the stress, but degrades the film stability [10]. This may lead to films with compositions that are not stoichiometric (stoichiometric silicon nitride SiNx ; x= 1.33) and incorporation into the film of other species different from silicon and nitrogen, usually hydrogen [11]. The qualitative chemical information of bonds present in the Si3N4 films can be obtained by using Fourier transform infrared (FTIR) spectroscopy measurements. Fourier transform infrared spectra have been taken in the range of 400-4000 cm-1 wave number of the Si3N4 film deposited at room temperature. Fig. 2 indicates the IR spectra of the range characteristic Si-N bond frequency of the deposited Si3N4 film. Fig. 2. FTIR spectra of the Si3N4 deposited on the p-Si.

Si-N stretching vibration whose frequency varies between 870 and 820 cm-1 in the silicon nitride films [1,8,12-17]. Si-N asymmetric stretching mode is apparently observed in 825 cm-1. No trace of hydrogen was observed in the formed silicon nitride as expected. Particularly, N-H stretching vibration and N-H bending vibration peaks expected at obout 3300 and 1175 cm-1 were absent. There is no evidence for any absorption associated with Si-H vibrations, which would occur between 2100-2200 cm-1. The Raman effect is an efficient way to determine the presence of pure silicon in its crystalline or amorphous form. Fig. 3 shown that bulk crystalline silicon clearly exhibits a very thin band with a Raman shift at 520 cm-1 [13]. A peak from Raman spectrum for the very thin Si3N4 film was not detected.

Fig. 3. Raman spectra of the crystalline p type silicon

From these different experimental results in this study, several stages in the evolution of the structure and of the optical properties can be defined. Raman spectra and IR spectra show pure silicon domains and the presence of Si-N bonds in the deposited Si3N4 , respectively. AFM were employed to characterize the surface morphology. Fig.4 shows the AFM micrograph of nitrided p-Si surfaces.

Fig. 4. The AFM image of the Si3N4 deposited on the p-Si.

The root-mean-square (rms) surface rougness of these films obtained by AFM analysis (Fig. 4) indicates that the surface rougness is close to 0.6 nm. This value is very well for smooth surface. X-ray diffraction (XRD) was used to assess the structure of the deposited Si3N4 layer. Our sample were formed by evaporating of Al dots with diameter of 1 mm ( for electrical measurement). XRD diffraction indicates that information gives both the Al dot area and the without Al dot area on the film surface. Fig. 5 indicates that surface of film are mixed phase of crystalline and amorphous structure. Thus, Si3N4 film is amorphous structure.

Fig. 5. XRD spectrum which is obtained from surface Si3N4 degraded with air ambient and Al dot.

A sharp peak around 29º (2-theta) corresponds to a spacing of 3.0663 Å, which is related to the Si3Al6O12N2 (201) (Aluminum Silicon Oxide Nitride) detected from the Al dot. Two sharp peaks for SiO2 should be observed at the around 28 and 38º (2-theta) which correspond to (040) and ( 42) planes, respectively. The peak around 43º (2-theta) correspond to a spacing of 1.5187 Å, which is attributed to the N2O diffracted from (222) plane. The peak of Si3Al6O12N2 has been obtained from the Al dot. The peaks related to SiO2 and N2O were obtained from the area which did not contain the Al dot. The peak at 47º (2-theta) is attributed to the substrate Si diffracted from (220) plane. There is no information about the peak at the around 48º (2-theta) peak. The formation of the Si3Al6O12N2 structure was chemically analyzed. According to that reaction, Si3N4 was degraded by oxygen and Al dot. That was given in equation (1). Si3Al6O12N2 → 3Al2O3 + 3SiO2 + Si3N4 (1) Due to the degradation of Si3N4 with the air ambient, the compounds of SiO2 and N2O have been formed. That is also given in equation (2). Si3N4 +4O2 → 3 SiO2 + 2N2O (2) Consequently, the stoichiometric Si3N4 on the Si(100) has been successfully produced. Conclusion The Si3N4 has been formed on the p-type Si(100) substrate thin film by using electrochemical anodization technique at the ambient temperature. Characterization of the thin film was obtained by FTIR, Raman spectroscopy, AFM and XRD measurements. Results of FTIR have shown the Si-N asymmetric stretching characteristic peak of the Si3N4 thin film, XRD pattern has shown the formation of stoichiometric Si3N4. AFM results show that the surface roughness of the Si3N4 film was about 0.6 nm. The surface roughness of the film is very important for the electronic devices. Thus, that value is quite good at performing a suitable thin film for electronic devices. It is concluded that the electrochemical anodization technique is an economical and practical way to produces a very thin film like the Si3N4 on the p-Si substrate. [1] S. Sitbon, M.C. Hugon, B. Agius, F. Abel, J.L. Courant, M. Puech, J. Vac. Sci. Technol. A 13(6) 2900 (1995). [2] H. F. Sterling, R.C.G. Swann, Solid-State Electron. 8, 653 (1965). [3] H. Zhu, D. Yang, L. Wang, D. Due, Thin Solid Films. 474, 326 (2005). [4] V. Misra, H. Lazar, Z. Wang, Y. Wu, H. Niimi, G. Lucovsky, J. J. Wortman, R. Hauser, J. Vac. Sci. Technol. B 17(4), 1836 (1999). [5] B.H.Lee, L.Kang, R.Nieh, W.J.Qi, J.C. Lee, Appl. Phys. Lett. 77, 1926 (2000). [6] J.Reed, G.B. Gao, A.Bochkarev and H.Morkoç, J. Appl. Phys.75(3), 1826 (1994). [7] H.Yüzer, H.Doğan, J.Köroğlu and S. Kocakuşak, Spectrochimica Act.B 55, 991 (2000). [8] G. Lucovsky, P.D. Richard, D. V. Tsu, S.Y. Lin, R. J. Markunas, J. Vac. Sci. Technol. A 4(3), 681 (1986). [9] S. Yoshida, K. Doi, K. Nakamura, A. Tachibana, Appl.Sur. Sci. 216, 141 (2003). [10] V. Bakardjieva, G. Beshkov, P. Vitanov, Z. Alexieva, J. Opto. and Advan. Mate. 7, 377 (2005). [11] E. S. Andres, A. del Prado, I. Martil, G. G. Diaz, F. L. Martinez, J. Vac. Sci. Technol. B 21(4), 1306 (2003). [12] S. Hasegawa, L. He, Y. Amano, T. Inokuma, Phys. Rev. B 48 (8), 5315 (1993). [13] M. Molinari, H. Rinnert, M. Vergnat, Appl. Phys. Lett. 77 (22), 3499 (2000). [14] M. Molinari, H. Rinnert, M. Vergnat, Appl. Phys. Lett. 79 (14), 2172 (2001). [15] M. I. Aloya, I. Pereyra, W. L. Scopel, M. C. A. Fantini, Thin Solid Films. 402, 154 (2002). [16] D. V. Tsu, G. Lucovsky, M. J. Mantini, Phys. Rev. B 33(10), 7069 (1986). [17] N.Pic, A. Glachant, S. Nitsche, J. Y. Hoarau, D. Goguenheim, D. Vuillaume, A. Sibai, C. Chaneliere, Solid. State Electron. 45, 1265 (2001).