Abstract
The reactions occurring at the interface between copper metal and gas bearing sulphur compounds (including hydrogen sulphide and odorant) and oxygen have been investigated by means of surface sensitive techniques, AES (SAM) and EDX. In order to avoid the influence of the ambient environment, a modified gas exposure facility attached to the surface analysis equipment was built, the specimen was able to be transferred from reaction vessel to analysis chamber directly. A series of Cu(LW) XAES spectra and other core level XPS spectra have been quantified by using deconvolution, curve fitting, factor analysis and computations associated with Beer-Lambert equation respectively. Chemical species produced in the surface reaction and their quantities have been estimated. A program set related to Scatter Diagram have been developed in FORTRAN 77 (LINK) language, which has been successfully used in the quasi-analysis of Auger/EDX maps, especially in constructing the multi-phase images. As a benefit, the problem of setting proper thresholds in superimposition of images has been solved. Based on the rigorous experimental operation and measurement from in-situ exposure of copper sample to low concentration (order of vpm) hydrogen sulphide and oxygen carried by pure nitrogen in dry environment at room temperature, following phenomena were observed: (1) the in-situ sulphidation rate is about 4 times higher than oxidation rate; (2) the thin copper oxide / adsorbate layer helps the surface sulphidation of copper exposed to hydrogen sulphide gas; (3) copper sulphide layer retards the consequent oxidation process of copper exposed to oxygen, although the oxidation of copper sulphide is favoured from thermodynamics principle; (4) in the simultaneous exposure to H2S and O2, enhanced sulphidation was observed, in which ratio of 1:19 between H2S and O2 gives the maximum rate of film growth. The possible mechanism is suggested and discussed. The role of oxygen in copper sulphidation as exposed to H2S plus O2 at room temperature and dry environment is to depolarize the created hydrogen in reaction such as 2Cu + H2S = Cu2S + H2 or Cu + H2S = CuS + H2, which leads to formation of water molecules. The balling tendency of the water molecules may result in the discontinuous water clusters distributed on copper surface. This leads to further two effects: (1) direct reaction between copper, hydrogen sulphide and oxygen, and oxidation of the hydrogen sulphide by the oxygen, these two reactions occurred with high probability at the location of water drops; (2) an irregular topography of the sulphided copper surface, and a consequent porous structure of the sulphide product layer as thick as several micrometers. In the study of sulphidation of preoxidized and scratched copper surface, with help of a program set for the analysis of Scatter Diagrams obtained from scanning Auger and EDX micrographs, we found that the sharp and curved places on cleaned copper surface have higher reaction ability for sulphidation; sulphur from hydrogen sulphide can penetrate into the interface between porous copper oxide layer and copper substrate to form copper sulphide. From the analysis of Cu(LW) XAES spectra in addition to the S2p, C1s, O1s and Cu2p XPS spectra obtained from a Cu-O2-H2S-Odorant-CH4 system, it is concluded that the Odorant, containing Diethyl Sulphide, Tertiary Butyl Mercaptan and Ethyl Mercaptan, is mainly chemi-adsorbed on the surface. By application of deconvolution, curve fitting and factor analysis on Cu(LVV) XAES spectra, the chemical shift from copper metal to copper oxide or copper sulphide has been confirmed. Four chemical species have been found by factor analysis, but exact details of the chemical species are not clear yet. A suggestion is made that the Cu2S and CuS might be identified by comparative analysis of the background at low kinetic energy side of Cu(LW) peaks from curve fitting. The potentials of combining Scatter Diagram and Factor Analysis have been realized and discussed, which will lead to more objective analysis of image/spectral data.