Particle formation, deposition, and particle induced corrosion in large-scale medium-speed diesel engines: Dissertation

The objective of this work was to study the formation of particles and their morphology and chemical composition in large-scale diesel engines operating with low-grade residual fuel oils. The effect of a Mg-based fuel oil additive on exhaust gas particles was also investigated. Particle characteristics were determined by means of the methods of aerosol technology, chemical analyses, and electron microscopy. As particle and deposit formation and characteristics play an important role in corrosion and erosion, the particle characterisation studies provided the necessary background information. The mass size distributions from the large-scale diesel engines were bimodal, with a main ("small") mode at 60-90 nm and a "large" mode at 7-10 µm. The small mode particles were formed by the nucleation of volatilised fuel oil ash species, which grew further by condensation and agglomeration. The large-mode particles were mainly agglomerates of different sizes consisting of small particles. These particles were re-entrained from deposits and fuel residue particles of different sizes. The number size distributions peaked at 40-60 nm. Agglomerates consisting of these primary spherical particles were also found. TEM micrographs revealed that these particles consisted of even smaller structures. On the basis of the mass and elemental size distributions, evidence that the fuel oil ash was highly volatile was found. The main causes for the differences in the aerosol size distributions were the engine type and fuel oil properties. By estimating the chemical compounds formed on the basis of ICP and EDS analyses at the corresponding mode in mass size distributions (about 0.1 µm), it was found that there was not enough oxygen in the particles to form only V2O5. Complete oxidation of vanadium into vanadium pentoxide was not favourable. This can be caused by many different factors, such as short residence times or soot particles acting as surface toxicants by blocking the active surface. However, the amount of sulphuric acid in the particles was high, about 27 wt. %. This required the formation of vanadium pentoxide to catalyse the formation of SO3 to form sulphuric acid. Doping the heavy fuel oil with a Mg-based additive caused another mode at about 2 µm in mass size distributions, making the size distributions trimodal. The 2-µm mode was generated by magnesium, together with some vanadium, nickel, and sulphur. Particle formation was not affected by the fuel oil additive. Deposition and corrosion studies on the surfaces of the Nimonic 80 A sample slabs were carried out on a laboratory-scale with a newly set-up deposition-corrosion apparatus (DCA). With this device the formation of the exhaust ash particles, gas composition, and deposition and corrosion on the sample slabs occurs in a similar way as in large-scale engines. Although corrosion studies have been carried out before, the formation of a corrosive ash layer when the particles deposit on the sample slabs has not previously been taken into account. Furthermore, the possible transformation of the deposited particles when they start to react to form a corrosive ash deposit has not been considered. In the deposition and corrosion experiments with SO2(g) and synthetic ash particle feeds, almost all of the particles observed looked like flat "pools" with small spherical particles in the middle of the "pool". Condensing sulphuric acid had dissolved the particles. Small (70-90-nm) spherical particles were also observed with an SO2(g) feed. On the other hand, hardly any S was found in the deposits. This indicated that S, in the form of SO2(g)/SO3(g), was transported through the deposit into the interface between the base material (pit area) and bottom of the deposit by molecular diffusion. The critical issue in the propagation of corrosion was the definition of the corrosion pit depth and the thickness of the bottom layer, because the latter increased with temperature (26 m at 700 versus 87 m at 750°C). There was no maximum at 700°C, as in the case when considering only the depth of the corrosion pit. A zone of "black islands" (15-33 wt. % S, the rest mainly Cr, Ni, and Ti) was found on the samples with SO2(g) and synthetic ash particle (SAP) feeds. The composition of these islands suggested that they were composed of a "mixed"-type sulphide ((Cr, Ti, Ni)Sx). As there was hardly any O available in this bottom layer, the "black islands" were formed by internal sulphidation. However, some of these islands were different from the others, consisting of 26 wt. % Cr, 37 wt. % Ti, and 26 wt. % O, the rest being V and Ni. These islands may be the "pre-existing" form of the oxide-rich layer found in the pit. The sulphur-rich "black islands" may transform into these oxygen- and vanadium-containing islands, as more and more oxygen diffuses into the bottom reacted layer where these islands were located and as the layer in the pit area grows. Because of a strong oxygen concentration gradient existing over the formed oxide scale (pit area), and the inward diffusing SO2/SO3 coming into contact with the base material (metal) at the interface between the deposit base and base material, SO2/SO3 becomes unstable. Thus it will dissociate to form atomic sulphur and oxygen molecule, and provides the sulphur needed for the internal sulphidation reaction (i.e. the formation of "black islands"). However, based on calculations of thermodynamical stability diagrams the formation of the nickel chromates and sulphates (e.g. type II hot corrosion, also called "low temperature hot corrosion") can not be entirely ruled out without further investigation with help of, e.g. XPS, XRD, from the corrosion pit area and bottom layer underneath it. To verify the experimental findings, an exhaust valve from a field endurance test of 8600 h in duration was analysed. A similar zone of sulphur-containing "black islands" was observed. However, the composition of these islands differed from that of those detected in the experimental system, as they contained much more Ni (about 70-80 wt. %) and less Ti (about 5 wt. %) and S (about 5-10 wt. %). As the temperature of the valve (T = 500°C) and oxygen content of the exhaust gas were different, the results are not directly comparable. However, there can still be similarities in the basic formation mechanism of these islands. Moreover, the corrosion results (amounts) obtained with this experimental set-up are of the same order as that which has been found in large-scale diesel engines.