Search

International audience; A low pressure recombining Ar plasma to which mixtures of N 2 and O 2 were added has been studied to explore the relevance of surface related processes for the total chemistry. The abundances of the stable molecules N 2, O 2, NO, N 2 O and NO 2 have been measured by means of a combination of infrared tunable diode laser absorption spectroscopy and mass spectrometry. A gas phase chemical kinetics model was developed in chemkin to investigate the contribution of homogeneous interactions to the conversion of the feedstock gases N 2 and O 2. At a partial pressure of N 2 plus O 2 less than 8 Pa, significant discrepancies between measured and calculated concentrations of N 2 O and NO 2 were observed, indicating that heterogeneous processes are dominating the chemistry in this pressure range. At a partial pressure of N 2 plus O 2 higher than 40 Pa and a relatively high fraction of admixed O 2 we observed a fair agreement between measured and calculated concentrations of NO molecules, indicating that homogeneous processes (notably Natom induced) are more dominant than heterogeneous processes.

We measured the densities of NH and NH(2) radicals by cavity ring-down spectroscopy in N(2)-H(2) plasmas expanding from a remote thermal plasma source and in N(2) plasmas to which H(2) was added in the background. The NH radical was observed via transitions in the (0,0), (1,1), and (2,2) vibrational bands of the A(3)Pi-- X(3)Sigma- electronic transition and the NH(2) radical via transitions in the (0,9,0)-- (0,0,0) band of the A(2)A(1)-- X(2)B(1) electronic transition. The measurements revealed typical densities of 5 x 10(18) m(-3) for the NH radical in both plasmas and up to 7 x 10(18) m(-3) for the NH(2) radical when N(2) and H(2) are both fed through the plasma source. In N(2) plasma with H(2) injected in the background, no NH(2) was detected, indicating that the density is below our detection limit of 3 x 1016 m-3. The error in the measured densities is estimated to be around 20%. From the trends of the NH(x) radicals as a function of the relative H(2) flow to the total N(2) and H(2) flow at several positions in the expanding plasma beam, the key reactions for the formation of NH and NH(2) have been determined. The NH radicals are mainly produced via the reaction of N atoms emitted by the plasma source with H(2) molecules with a minor contribution from the reaction of N+ with H(2). The NH(2) radicals are formed by reactions of NH(3) molecules, produced at the walls of the plasma reactor, and H atoms emitted by the plasma source. The NH radicals can also be produced by H abstraction of NH(2) radicals. The flux densities of the NH(x) radicals with respect to the atomic radicals are appreciable in the first part of the expansion. Further downstream the NH(x) radicals are dissociated, and their densities become smaller than those of the atomic radicals. It is concluded that the NH(x) radicals play an important role as precursors for the N and H atoms, which are key to the surface production of N(2), H(2), and NH(3) molecules.

A full characterization of the Ar–NH3 expanding thermal plasma source applied in industrial processes for high-rate silicon nitride deposition is presented in terms of absolute densities of reactive species such as ions and radicals. The ion composition of the plasma was identified by mass spectrometry, while absolute ion density information was obtained by Langmuir probe measurements. N radicals were detected by threshold ionization mass spectrometry, whereas NH and NH2 radicals were measured by cavity ringdown spectroscopy. It was found that the ion density decreases from 1013 cm−3 in a pure Ar plasma to 1010–1011 cm−3 when NH3 is injected, with ArH+, , , and being the most abundant ions. The densities of N and NH both saturate at a level of 3 × 1012 cm−3 at NH3 flows above 3 sccs while the density of NH2 increases linearly with the NH3 flow and reaches a level of 4 × 1012 cm−3. When the plasma source current is increased, the densities of N and NH increase linearly, while the NH2 density remains approximately constant. Furthermore, it is revealed that most of the consumed NH3 is converted into N2 and H2 in the plasma.

The densities of NH and N H2 radicals in an Ar-N H3 plasma jet created by the expanding thermal plasma source were investigated for various source-operating conditions such as plasma current and N H3 flow. The radicals were measured by cavity ringdown absorption spectroscopy using the (0,0) band of the A ¿3 ¿ X ¿-3 transition for NH and the (0,9,0)-(0,0,0) band of the A~ A12 ¿ X~ B12 transition for N H2. For NH, a kinetic gas temperature and rotational temperature of 1750±100 and 1920±100 K were found, respectively. The measurements revealed typical densities of 2.5× 1012 cm-3 for the NH radical and 3.5× 1012 cm-3 for the N H2 radical. From the combination of the data with ion density and N H3 consumption measurements in the plasma as well as from a simple one-dimensional plug down model, the key production reactions for NH and N H2 are discussed. © 2005 American Institute of Physics.

Cavity ring-down detection techniques can sensitively determine frequency-dependent absorption cross-sections of gasses. However, so-called line-width problems and amplified spontaneous emission of the laser light source lowers the technique’s quantitative accuracy. Using phase-shift cavity ring-down spectroscopy (PSCRD), we measured absolute line intensities of the spin-forbidden transitions in the b 1 Σ g + ( ν ′ = 0 ) ← X 3 Σ g - ( ν ″ = 0 ) band of molecular oxygen. Our results were within 4% of values obtained from the HITRAN database, demonstrating the accuracy of PSCRD, when corrected for amplified spontaneous emission. Its high sensitivity (2 × 10−8 cm−1), simplicity and high duty cycle make PSCRD a powerful diagnostic technique.

The sub gap absorption at 1.17 eV in hydrogenated amorphous silicon (a-Si:H) has been measured by means of the ex situ cavity ring down (CRD) absorption technique. The evolution of defects has been studied as a function of film thickness and deposition temperatures. A comparison with the dual beam photoconductivity (DBP) technique shows that the CRD results are systematically higher. Furthermore, it is shown that the CRD technique is sensitive to surface defects and typical surface dangling bond coverages of 2×10−4 up to 6×10−3 are obtained for air-exposed a-Si:H samples depending on the deposition temperature.

We measured the steady-state gas composition of plasmas produced from Ar?N2?O2 mixtures and Ar?NO mixtures with quantitative mass spectrometry. In the former, mainly N2 and O2, but also a significant amount of nitric oxide (NO) was formed, i.e. up to 5% of the background gas was NO. In the inverse experiment, in which NO was admixed to an argon plasma, up to 92% of the NO was converted into N2 and O2. The observed molecules are mostly generated in wall association processes but also by gas phase reactions between N atoms and O2 molecules leading to NO. The two types of plasmas show a strong mutual resemblance in the steady-state gas composition if substantial dissociation can be reached in the residence time of the gases in the plasma, i.e. ?5% NO and ?95% N2 and O2, although the starting conditions are completely different. It seems that in first order the system prefers to produce the most thermodynamically stable molecules.

We investigated the efficiency and formation mechanism of ammonia generation in recombining plasmas generated from mixtures of N2 and H2 under various plasma conditions. In contrast to the Haber-Bosch process, in which the molecules are dissociated on a catalytic surface, under these plasma conditions the precursor molecules, N2 and H2, are already dissociated in the gas phase. Surfaces are thus exposed to large fluxes of atomic N and H radicals. The ammonia production turns out to be strongly dependent on the fluxes of atomic N and H radicals to the surface. By optimizing the atomic N and H fluxes to the surface using an atomic nitrogen and hydrogen source ammonia can be formed efficiently, i.e., more than 10% of the total background pressure is measured to be ammonia. The results obtained show a strong similarity with results reported in literature, which were explained by the production of ammonia at the surface by stepwise addition reactions between adsorbed nitrogen and hydrogen containing radicals at the surface and incoming N and H containing radicals. Furthermore, our results indicate that the ammonia production is independent of wall material. The high fluxes of N and H radicals in our experiments result in a passivated surface, and the actual chemistry, leading to the formation of ammonia, takes place in an additional layer on top of this passivated surface.

The densities of N, NH, and NH2 radicals in a remote Ar-NH3-SiH4 plasma used for high-rate silicon nitride deposition were investigated for different gas mixts. and plasma settings using cavity ringdown absorption spectroscopy and threshold ionization mass spectrometry. For typical deposition conditions, the N, NH, and NH2 radical densities are on the order of 1012 cm-3 and the trends with NH3 flow, SiH4 flow, and plasma source current are reported. We present a feasible reaction pathway for the prodn. and loss of the NHx radicals that is consistent with the exptl. results. Furthermore, mass spectrometry revealed that the consumption of NH3 was typically 40%, while it was over 80% for SiH4. On the basis of the measured N densities we deduced the recombination and sticking coeff. for N radicals on a silicon nitride film. Using this sticking coeff. and reported surface reaction probabilities of NH and NH2 radicals, we conclude that N and NH2 radicals are mainly responsible for the N incorporation in the silicon nitride film, while Si atoms are most likely brought to the surface in the form of SiHx radicals. [on SciFinder (R)]