The Impact of surface treatments
The Impact of surface treatments applied to welds and the area around the weld on their corrosion resistance
According to current standards, the surface of arc-welded cap welds needs to be treated. This treatment is required to either remove surface materials such as oxides or metal spatters or to smooth out surface defects such as scales. It is usually performed manually with electric grinders using a grinding wheel/abrasive disc or with a wire brush. The grinding usually leads to excessive removal of the surface layer of the weld metal. It is accurate to say that this disadvantage was eliminated with the wire brush selected as it does not destroy the surface of the melted-on metal to such a great extent. It is a well-known fact that the structure of the surface layer, as well as the roughness and presence of a surface skin, can have a significant impact on the corrosion resistance of a material’s surface. Structural defects extending right through to this surface have greater reactivity and provide potential sources of corrosion, as all forms of nonhomogeneity increase both the chemical and electrochemical corrosion.
This article examines the effect of the treatment technology applied to a carbon steel weld on the corrosion resistance of the welded joint.Experimental Section
The samples were created by manually arc-welding pipes made from high-quality constructional steel with a diameter of 50mm and a wall thickness of 5.4mm. The surface of the weld was treated using two methods: mechanical cleaning with an abrasive or treatment with OSBORN brushes. The samples were produced in the form of sheets measuring 20x40x5.4mm. The first sample group featured a centrally located longitudinal weld that was subjected to various treatments. The other sample group was cut out from sections located at a distance from the area of the weld’s thermal impact. Following the treatment and corrosion tests, the microstructure of the weld surface was examined by electrical microscopy using a NEOFOT appliance. The X-ray structure phase analysis (XRSPA) was performed with the diffractometer DRON-2 using monochrome FeKa radiation. The X-ray spectrograms were evaluated by comparing the experimentally established distances between the levels with the normal distances according to the American Society for Testing and Materials (ASTM). For the corrosion and electrochemical tests, the surface of the samples was insulated with rosin-containing wax in a manner ensuring that only the section with the weld was left exposed. The tests were performed in a 0.5m Na2SO4 solution. The potentiodynamic polarisation curves were reduced using the Potentiostat P5827 with a linear potential sweep rate of 0.1V/minute.
Figure 1 shows the microstructure of a weld surface following treatment with abrasive (A), OSBORN brushes (B) and the original pipe surface at a distance from the area of the weld’s thermal impact (C).
Following mechanical cleaning with abrasive (see Figure 1A), the weld surface showed the formation of obvious ground surfaces, grooves and microfissures that can act as voltage-boosting discontinuity points promoting the enlargement of the fissures or development of stress corrosion cracks. The depth of the grooves appearing after the treatment was determined using a Kalibr profilometer and extended up to 70mkm. The results of the phase analysis of the weld surface showed that the surface was free from oxides.
Mechanical cleaning of the weld surface with abrasive therefore removed the primary technological skin of oxides and corrosion products and created a juvenile surface characterised by a distorted structure, an accumulation of voltage-boosting discontinuity points and inner defects that are potential sources of the formation of fissures or series of fatigue cracks given alternate stress. Meanwhile, the metallographic analysis of the weld surface treated with OSBORN brushes (see Figure 1B) showed that the microstructure of the weld surface was virtually identical to that of the original surface (see Figure 1C). X-ray structure analysis of the original pipe surface at a distance from the weld’s thermal impact showed that the pipe material mainly consisted of magnetite Fe2O3 featuring a crystalline rhombus grid with the following parameters: a=4.59nm, b=4.97nm and c=6.68nm. To a lesser extent, the surface layer contains maghemite Fe3O4 featuring a cubic crystalline Fd3m grid with the parameter a=8.09nm, and iron oxide (1) (FeO) with a cubic Fm3m grid and with the parameter a=4.293nm. The two-layer skin that is formed on the original surface of a mark 20 steel pipe at a distance from the area of the weld’s thermal impact is hence composed of Fe/FeO and Fe3O4/ Fe2O3. The thickness of the oxide layer measured by diffraction peak reduction amounts to (110) a-Fe, based on an oxide skin of approximately 1 mkm (numbers in brackets refer to the scale of diffraction peak measurement method). One can see that the microstructure of the weld (see Figure 1B) is relatively fine-grained. After treatment with OSBORN brushes, the weld as measured by X-ray fluorescence (XRF) contained only one Fe3O4 magnetite layer with a cubic Fd3m grid and the parameter a=8.09nm, a-Fe with a cubic, volume-centred grid and the parameter a=2.869nm, and austenite a-Fe with a cubic surface-centred grid (a=3.637nm). The austenite volume arrived at via the Landé formula amounted to 6%. The layer formed on the surface of a weld treated with OSBORN brushes hence has a fine-grained microstructure that contains magnetite, which is adequately interlinked with the steel.
Tests showed that the corrosion potential of the steel in its original state amounted to –0.24V in a 0.5m Na2SO4 solution. After removal of the technological oxide layer by grinding, the corrosion potential of the steel shifted in a negative direction and amounted to –0.38V. The corrosion potential of the steel treated with OSBORN brushes is also adequately negative and equalled the corrosion potential of the steel processed by grinding. This indicates that the Fe3O4 magnetite layer preserved on the surface of a weld after the treatment with OSBORN brushes was sufficiently porous. Figure 2 shows the anodic potentiodynamic polarisation curves of the steel in its original state (see Figure 2, line A) and after the mechanical grinding treatment (see Figure 2, line B).
It is apparent that within the potential interval to be examined, the steel was dissolved in a virtually unhindered fashion after the abrasive treatment (see Figure 2, line B). The polarisation curve shows that the surface of the steel is covered by a carbon layer that is badly interlinked with it. The presence of an oxide layer on the surface (see Figure 2, line A) meanwhile retarded the dissolution process. The peak anodic current was observed at a potential of 0.2V. In the potential interval of 0.2–0.7V, the surface took on a dark coloration and the dissolution of the steel slowed down.
An analysis of the composition of the steel’s surface layer following the reduction of the anodic XRF polarisation curve demonstrated the development of new peaks, attesting to the formation of a new phase. This phase is represented by graphite C with a hexagonal P63/mmc grid and the parameters a=2.46nm and c=6.70nm. The presence of a large amount of carbon with the hexagonal gridparameters a=8.95nm, c=14.08nm and traces of g-Fe2.5c with the parameters a=11.56nm, b=4.573nm and c=5.058nm is possible.
The dissolution of steel equipped with a technological oxide skin was hence accompanied by an accumulation of graphite in the pores of the skin (formation of a graphite surface) that served to slow down the dissolution process. Accordingly, the dissolution of the steel was accompanied by a reduction in the porosity of the original oxide layer; as a result, the anticorrosion protection attributes were boosted, as was the structural homogeneity, leading to the formation of a surface layer with better protection attributes.
The tests showed that the corrosion potential of the weld metal after abrasive treatment amounted to -0.38V in a 0.5M Na2SO4 solution and equalled the corrosion potential of the steel at a distance from the area of thermal impact. The corrosion potential of the weld metal was adequate after both the treatment with OSBORN brushes and the abrasive treatment and showed the same value of –0.38V. This indicates that the magnetite layer preserved at the weld surface after the treatment with OSBORN brushes was sufficiently porous.
Figure 3 shows the anodic polarisation curves of the weld metal following abrasive treatment (see Figure 3, line A) and after processing with OSBORN brushes (see Figure 3, line B). One can see that the magnetite layer preserved on the weld surface treated with OSBORN brushes (see Figure 3, line B) retarded the dissolution of the weld metal in the active area. At a potential of 0.8V, the polarisation curve veered off and the region of the limiting current appears. The surface takes on a dark colour. This indicates that the anodic weld metal dissolution process is accompanied by an accumulation of carbon in the magnetite layer that retards further dissolution. The findings were confirmed by the XRF data.
The interlinkage of the carbon-containing magnetite layer with the weld surface was sufficiently stable and could provide corrosion protection in the course of further operations. However, after the treatment with abrasive, the weld metal’s dissolution progressed virtually unrestrained (see Figure 3, line A). The X-ray phase analysis data showed that the weld’s surface layer shows traces of the oxidesFe2O3 and Fe3O4, as well as graphite with the hexagonal grid parameters a=2.46nm and c=6.70m after reduction of the polarisation curve. The interlinkage of the graphite layer with the weld’s metal surface was extremely weak; it could even potentially flake off. The surface layer formed when the metal of a weld treated with abrasive is dissolved fails to counteract further dissolution. Hence, a significant decarburisation of the surface is possible.
The weld sections were cut from the welded joints in order to perform the corrosion tests. The test-irrelevant surfaces of the samples were insulated with rosin-containing wax. The tests were performed over a period of 68 hours in a 3% NaCl solution.
Visual examination of the sample surfaces after the corrosion tests showed that the surface of a weld treated with abrasive was completely covered by a layer of brown corrosion products (rust) that was only badly interlinked with the surface. In the presence of carbon the rust largely consisted of products resulting from the complete oxidation of iron – predominantly amphoteric Fe(lll) hydroxide, a-FeOOH goethite and lepidocrocite g-FeOOH. The surface of the weld treated with OSBORN brushes was meanwhile much less oxidised after the corrosion tests and retained its original dark colour. Therefore, treating welds with abrasives failed to retard the corrosion process.Conclusion
The microstructure of the weld surface, which is decisive for the weld’s predisposition towards corrosion damage, shows marked differences that depend on the cleaning method. After abrasive cleaning the surface layer of the weld shows a distorted structure and is characterised by an accumulation of voltageboosting discontinuity points and inner defects, which has a detrimental effect on its corrosion properties in further operations. However, when the weld is treated with OSBORN brushes the carbon is able to settle in a magnetite layer in the dissolution process, boosting the surface’s protective properties and hence retarding further dissolution.
The weld corrosion tests demonstrated that the corrosion resistance of weld surfaces to a significant extent depends on the treatment method applied. If the weld is cleaned with an abrasive tool, the corrosion takes its course without hindrance and the corrosion products are only weakly interlinked with the surface. However, treatment with OSBORN brushes serves to boost the corrosion resistance of the weld and promotes the improvement of the weld’s operational attributes.