Formation of silicon oxynitride by laser ablation of a silicon target in a nitrogen atmosphere
Abstract
Studies during the last ten years have demonstrated that it is difficult to deposit pure silicon nitride by Pulsed Laser Deposition (PLD) without oxygen contamination even using a pure nitrogen atmosphere and a pure silicon target. The difference in the reactivity of nitrogen and oxygen means that there is always a certain amount of oxidation during the film growth from the residual oxygen in the vacuum system; the oxygen content can be more than 20 at.%. The aim of this work was to study the deposition of silicon nitride thin films under different experimental conditions to determine whether oxidation during, or posterior, to the deposition where the principal processes. We used optical emission spectroscopy (OES) and Langmuir probe measurements to study the plasma species and the parameters of the plasma. The chemical composition and the deposition rate of the films were measured by Rutherford Back Scattering (RBS), x-ray photoelectron spectroscopy (XPS) and profilometry, respectively. The OES results showed that the signal from nitrogen reached a maximum at 0.4 Pa (N2 + , 391.4 nm) and decreased if the nitrogen pressure was increased further. The maximum nitrogen concentration in the deposits was ~21 at.% at 1 Pa, with an oxygen content of 28 at%. For lower pressures the coatings were rich in silicon and the coatings deposited at higher pressure (>1Pa) were rich in oxygen up to 50 at.%.
Keywords
Download Options
Introduction
Silicon nitride is an attractive material for a large number of applications in microelectronics, optoelectronics and NEMS devices because of its high band gap, thermal, chemical and mechanical properties and stability. Silicon nitride coatings are commonly deposited by Plasma Enhanced Chemical Vapor Deposition (PECVD) [1], and magnetron sputtering [2, 3]. There are also many reports of silicon nitride deposition by Pulsed Laser Deposition (PLD) using different configurations such as a silicon target in a reactive nitrogen[4-6] or an ammonia atmosphere, or from a silicon nitride target in an Ar/N2 gas mixture . Using the first configuration I. Umezu, et al.[6] studied the effect of the nitrogen pressure on the FTIR spectra and the optical band gap of the deposited films. They found that when the pressure was increased from 5 to 25 Pa the main FTIR peak shifted from 835 to 920 cm-1 and that this was related to the transformation of the silicon nitride into silicon oxynitride. This result implied that the group could deposit pure silicon nitride at low pressure, but unfortunately they did not report the composition of the film. However, the band gap of the films deposited at 5 Pa was 3 eV which is considerable lower than the 5.5 eV value normally accepted for silicon nitride [7]. The gap increased as the pressure was increased, but this was probably due to the increased amount of oxygen in the films, and the authors suggested that the oxidation occurred after the deposition. J.D. Wu [8] combined PLD with a nitrogen plasma produced in a microwave electron cyclotron resonance system (ECR) to increase the concentration of reactive nitrogen species, since molecular nitrogen was not effective for the formation of the nitride [9]. Auger Electron Spectroscopy (AES) measurements showed that the nitrogen content increased from 6 to 48 at.% when the ECR system was used, but the oxygen content was approximately 15 at.% [8]. Silicon nitride has been deposited in ammonia [9] with the idea that atomic nitrogen could be more easily produced from the decomposition of the ammonia. The group of E. D` Anna[10] deposited silicon nitride in an ammonia atmosphere using an excimer laser, and based on the X ray diffraction measurements they obtained a mixture of Si, Si3N4 and SiO2. Chemical analysis showed that when the gas pressure was increased the nitrogen concentration increased from 8 to 21.6 at.%, but that the oxygen content also changed from 36.5 to 43.4 at.%.
The deposition of silicon nitride by PLD using a silicon nitride target has also been studied. In that case it was concluded that it was necessary to use high nitrogen pressures (≥10 mtorr) in order to obtain a N/Si ratio of greater than 1.3, but the thin films had a void volume fraction close to 50%. The band gap of the films was seen to increase from 2.8 eV to 4.6 eV as a function of the nitrogen pressure. Similarly, E.C. Samano [11] found a clear relationship between void formation and background gas pressure, and they concluded that the voids were due to the trapping of the gas.
The aim of the present research was to study the chemical reactions occurring in the PLD plasma of ablated silicon in nitrogen, at different working gas pressures, to determine the relation between the nitrogen and oxygen incorporation in the deposits, and to establish if the included nitrogen was present as silicon nitride or trapped nitrogen gas.
Conclusion
Using pulsed laser deposition we deposited silicon oxynitride coatings from a pure silicon target in a flow of pure nitrogen. The measurements of the composition showed that the oxygen and nitrogen content were comparable with experiments performed with an ammonia atmosphere and an external nitrogen ion source. The plasma characterization showed that it was possible to ionize silicon and molecular nitrogen, but neither atomic nitrogen ions nor any excited species of oxygen were observed. Based on our results we propose that the oxidation of the deposit occurred after each laser pulse, when the fresh film surface reacted with the residual oxygen in the chamber. For the gas pressures greater than 1 Pa this process was probably enhanced by the formation of a less dense, more porous deposit with the greater surface area leading to films with higher oxygen and lower nitrogen concentrations. The XPS results showed that the obtained thin films consisted of a mixture of silicon oxynitrides and pure silicon, and that the ratio of these phases was determined by the nitrogen pressure used.