Sveising fører til endret mikrostruktur i deler av stålet. I kombinasjon med lave temperaturer oppstår nye forutsetninger for brudd.I denne masteroppgava ble det sett på initiering av sprøbrudd i et sveisesimulert 420 MPa HSLA-stål fra Nippon Steel. Dette ble gjort ved å studere mikrostrukturen og bruddflata til seks sveisesimulerte prøver etter bruddmekanisk trepunktsbøying med utmattingskjerv. I tillegg ble det videre undersøkt en mulig MA-fase som tidligere ble funnet i et initieringspunkt på ei prøve med interkritisk grovkornet HAZ.To prøver hadde simulert grovkornet HAZ og fire prøver var forsøkt rendyrket martensitt- og bainittstruktur. Acikulær ferritt ble her sett på som en del av bainittstrukturen. Alle prøvene ble varmet opp til 1350 ◦C og hadde avkjølingstid, ∆τ8/5, på 5, 15 og 22,5 sekunder for derved å simulere henholdsvis martensittstruktur, grovkornet HAZ og bainittstruktur. Prøvene ble brukket ved −60 ◦C.Det var forventet at økt andel bainittstruktur i prøvene ville føre til lavere seighet og dermed lavere CTOD-verdier. Martensittstrukturen var noe selvanløpt og dermed mer duktil. Imidlertid var ikke sammenhengen mellom andel bainittstruktur og initiering av sprøbrudd så klar som forventet og det var mange observasjoner som svekket styrken i sammenhengen.På alle seks bruddprøvene ble alle mulige initieringspunkt langs og foran sprekkspissen kartlagt ved bruk av SEM. Det ble funnet dobbelt så mange initieringspunkt på prøvene med avkjølingstid lik 5 sekunder som de med avkjølingstid lik 15 og 22,5 sekunder. Imidlertid lå mange av disse sværtnær hverandre og mange var låst av duktil bruddvekst.Etter undersøkelse i SEM ble de fire rendyrkede bruddprøvene nitaletsa og undersøkt på ny. Det ble observert karbider i, eller rett ved, 61% av initieringspunktene, flest i prøvene med avkjølingstid lik 22,5 sekunder. Dette viser at initieringen kan være knyttet til karbider og dermed bainittstruktur.Mikrostrukturen til de ulike prøvene ble studert i lysmikroskop. Prøvene med martensittstruktur inneholdt også områder med bainittstruktur. Det var forventet at disse prøvene skulle få høyest CTOD-verdier på grunn av duktil martensitt, men det var ikke tilfellet. De lave CTOD-verdiene kan ha vært grunnet initiering i bainittstrukturen.Det var ingen klar tendens på hva slags beliggenhet initieringspunktene hadde i forhold til prøvenes mikrostruktur eller CTOD-verdi. På alle prøvene ble det observert initiering foran sprekkspiss. Disse initieringene ga mest bruddvekst, bortsett fra i de to rendyrkede prøvene med lavest CTOD-verdi.Det ble gjort tre slipesteg inn i prøvene med grovkorna HAZ for å undersøke mikrostrukturfordelingen langs bruddkantene ved bruk av lysmikroskop. Andelen martensitt og bainitt langs hele bruddkanten ble estimert ved å vurdere mikrostrukturen i hvert punkt som var relevant i et rutenett. Det var ingen markant forskjell i fordelingen av mikrostruktur langs bruddkanten i forhold til CTOD-verdi.Mikrosondeundersøkelser av og rundt en forhøyning i et initieringspunkt på ei prøve med interkritisk grovkornet HAZ indikerte at det kunne være en MA-fase. Desverre forsvant forhøyningen da den skulle kuttes ut ved hjelp av FIB for å undersøkes i TEM.

The recycling of aluminium from used aluminium scrap leads to an unavoidable presence of pollutions in the form of elements of various amounts. Two such elements are iron and silicon. These will always be present to some extent in an aluminium alloy as they are introduced to the alloy during processing. Iron and silicon are accumulative elements, meaning that they can never be completely eliminated once introduced into the aluminium. Some alloys have very narrow compositional windows, they have strict regulations concerning amount of alloying elements allowed. This is of grave importance for the recycling process; to be able to produce specific alloys with limitations regarding content and mechanical properties. A project to uncover the effects of common elements in aluminium such as Mg, Si and Fe has been initiated by the industry. This work is closely related to the project and was carried out as a report during fall 2010 and continued as a master thesis during spring 2011. The aim of this work was to study the effects of variation of small amounts of iron and silicon in 4 different alloys of high and ultra high purity (>99.7 and >99.999 wt% Al, respectively). The main focus was to uncover the effect of the elements on mechanical properties. Tensile test experiments with specimens in uniaxial tension were conducted to uncover the mechanical properties of the alloys. In addition to the tensile tests other parameters like texture, particles, recrystallisation and micro structure development were investigated. DC-cast, extruded flat profiles and rolled and recrystallised versions of the 4 alloys were used during the work.The results from tensile tests performed on the rolled and recrystallised aluminium showed that an initial introduction of 0.066 wt % Fe and Si 0.068 wt % to an alloy containing no pollutions lead to a dramatic increase in strength properties (up to 100% for the tensile strength) and a mild decrease in elongation. The increased strength properties and reduced formability was traced to iron and silicon in solid solution. From the same tensile tests it was concluded that a further increase of iron and an introduction of a small amount of pollutions had close to no effect on the mechanical properties.

AbstractSteel alloys are the material that is most used in protective constructions today. The reason is the overall good properties of steels, with its high strength and hardness, high ductility, high formability and relatively low cost compared to other materials. Even though armour steels are affordable, work is done to limit expenses from production or come up with new and less expensive alloys.Case hardening is a technique for production of steel with a very hard and durable surface while still maintaining a tough and ductile core. Case hardened products have hard, durable and fatigue resistant surfaces and tough and durable cores. Case hardening is typically applied to machine parts that are exposed to heavy wear and high loads such as gears, bearings, screws and nuts, shafts, etcIn 2010, Lou et al. managed to significantly increase the penetration resistance of a low-cost steel (NVE36) by case hardening. Encouraged by this, Hans Magne Thorseth wrote his master’s thesis “Optimalisering av stål i beskyttelseskonstruksjoner” in 2010. His attempt to improve the ballistic strength of Hardox 450 steel by case hardening did not have the desired results. The ballistic limit velocity was reduced from approximately 800 m/s for the original Hardox 450 to approximately 770 m/s for the case hardened steel. The probable reasons for the reduction of the ballistic limit velocity was a too low carbon content increase from the carburizing and thus a too low hardness increase, in addition to deterioration of the metal core properties from the heat treatment. The objective of this master’s thesis was to improve the case hardening process used by Hans Magne Thorseth in his master’s thesis.The improvement work was divided into two parts, a preliminary project work and this master thesis. The project was carried out in the autumn of 2010, and the aim was to find the optimum case hardening process to produce steel with improved ballistic properties. This involved testing of different carburizing potentials to increase the surface carbon content of the metal and to find carburizing heat treatments that would retain the core properties of the carburized steel. Based on the results from the preliminary project it seemed possible to produce case hardened steel with surface hardness over 1000 HV, but a slight reduction of core hardness was difficult to avoid.The present master’s thesis was written during the spring of 2011. The key results from the preliminary project work were used to determine a range of case hardening courses. A total of 11 case hardening courses were tested, of which six courses employed carbon potential of 0.9% and five courses employed a carbon potential of 1.1%. Different hardening courses were tested, including direct hardening, a modified single hardening process and traditional and modified versions of double and triple hardening. Before the case hardening the plates were sandblasted to remove a corrosion-resistant primer that was applied to the plates. The primer was believed to limit the diffusion of carbon into the steel. In the preliminary project the primer was ground away and grinding was the preferred solution in this work as well, however, due to capacity issues sandblasting was employed.The ballistic limit velocity of the case hardened and the original Hardox 450 target plates were calculated from results obtained by ballistic experiments. The ballistic limit velocities of the case hardened samples, with approximately 787 m/s as the highest, were lower than that of the original Hardox 450, being approximately 800 m/s.Surface hardness measurements of the steel revealed a poor effect from the carburization, the hardest surface was measured at 735 HV. This could be explained by a probable surface decarburization of the original Hardox 450, which was discovered by microstructure examinations and hardness measurements. In contrary to in the preliminary project where this layer was ground away, the sandblasting did not remove the decarburized layer. Due to this, the case hardening did not have the desired effect on the steel surface, resulting in a too low surface hardness.The core hardness proved difficult to retain, the highest core hardness value obtained was below 460 HV, compared to the original Hardox 450 with core hardness of 475 HV. This could be another reason for the lower ballistic limit velocity of the carburized steel plates. However, the original Hardox 450 did not have the highest cross-sectional hardness integral value, so the ballistic properties probably also depend to some degree on ductility. A main reason for the superior ballistic limit velocity of the original Hardox 450 was therefore believed to be related to cracking of the rear side of the case hardened target plates during projectile penetration. The cracking seemed to be a result of the harder, and thus more brittle, outer surface layer.

Silicon is used in the production of various ceramic products. As the solar industry is advancing, the need for reusable crucibles for the casting of solar grade silicon has arisen. One idea is to produce silicon nitride crucibles by slip casting. During slip casting, silicon powder is mixed with water at high alkaline conditions to ensure good dispersion of the powder and a stabilized slip. This may lead to oxidation of the silicon and the evolution of hydrogen gas. The hydrogen evolution introduces porosity in the final product, thereby ruining the mechanical properties. Gaining knowledge on how to inhibit the oxidation of silicon in slurries would therefore prove valuable for producing defect-free silicon based ceramics.In this thesis the main objective is to gain knowledge about the oxidation of silicon in aqueous media, emphasizing on the formation of oxides with respect to initiation point, initiation time and characterizing of the oxide. Silicon powder and coarse silicon particles delivered by Elkem was oxidized in a purpose built apparatus. The effects on the powder and the coarse particles were examined by the several techniques, including among other: SEM, Colorimetry, HR ICP-MS and LECO. The mechanical properties of oxidized single crystal and polycrystalline silicon were also investigated, using nanoindentation.The most important results, indicates that the oxidation of silicon results in formation of silica and hydrogen gas evolution. Firstly silicic acid is formed until the solubility limit is reached. Further oxidation leads to silicic acid condensing to form; colloidal silica which adsorbs on the silicon particles and dispersed colloidal silica in the suspension. The oxide layer formation on the coarse silicon particles were studied and confirmed using SEM. The oxide is the result of silicic acid condensing to form silica and adsorbing on to the surface of the silicon. Initiation points for the oxidation were not found. A more controlled oxidation is needed and the study of the oxide layer must be performed with high resolution techniques like Atomic Force Microscopy and/or Auger Electron Spectroscopy.There were some problems related to reproducing the results with respect to the hydrogen evolution from the silicon powder. Storage of the powder seemed to be a key factor as the powder behaved differently from one week to another. Storage in a desiccator for two weeks lead to earlier and more intense hydrogen evolution. Any reasonable cause for this was not found, and is subject for further investigations. The initiation time for hydrogen evolution was therefore difficult to determine, as it is dependent on unknown parameters.Nanoindentation was found to be a useful technique for studying the mechanical properties of oxidized silicon. The results indicated that the single crystal samples were homogeneous with respect to mechanical properties, as each of the indents of the respective samples produced more or less the same value. The E-modulus for the oxidized single crystal sample was about 10 GPa lower than for the unoxidized sample, the hardness was not affected. The results for the polycrystalline sample is more difficult to interpret, but nanoindentation indicates that the oxidized sample is not as homogeneous as the single crystal particles with respect to mechanical properties. The hardness was equal to the single crystal samples, and the E-modulus was close to that of the unoxidized single crystal sample.

By applying a texture to the front surface of solar cells, less light may be reflected and the incoming light may be given a longer path length inside the solar cell causing a better light absorbance for the solar cell. The textured layer is today made by etching with an acidic etch with a large part hydrogen fluoride. In the future a more environmental approach for creating textured solar cells may be favoured. There are many ways of creating textured surfaces, one of them are by imprinting the surface by a template, and remove the template when the process is completed.In this project the diatom species Coscinodiscus walesii and an undefined Coscinodiscus species were cleaned and processed for investigation. The work done can be divided into four tasks. The first was characterisation and investigation of the pore structure. The second were manufacturing and characterisation of gold for use as templates. The third were deposition and characterisation of thin films of silicon and silicon nitride on diatom frustules. The last part were to make structures are very similar to solar cells with diatom frustules incorporated into them.The diatom frustules were characterised by SEM and the pore structure were cut through with FIB to characterise the structure. There were structural differences between the two species. The Coscinodiscus walesii was grown in a cultivation chamber, and lacked the circular inner pore structure which the undefined Coscinodiscus species had. This was the only real difference between the two species.On the cultivated species were deposited a gold film, which a small rectangle were lifted off by a tungsten needle in FIB. The small rectangle were characterised with the SEM-column in a dual beam FIB. Pore structures down to 40 nm were replicated by the gold film lifted off the frustule. By depositing a relatively thick film of gold on a glass substrate covered with diatom frustules and dissolving the glass substrate and diatom frustules with hydrogen fluoride, a template were made with a replication of the diatom frustule pore structure.Silicon and silicon nitride were deposited on diatom frustules and characterised with FIB to investigate how those materials followed the frustule topography. A good conformity of films made of those materials was confirmed, and 4 different samples which were similar to solar cells were manufactured and characterised by light microscopy, SEM and FIB. Two samples were made to be similar to crystalline silicon solar cells. One had dried frustules on top of a silicon surface and the entire surface of the sample were coated with silicon nitride, a blue colour were observed in the entire surface and the silicon nitride had also been deposited on the diatom frustules. For the other sample similar to crystalline silicon solar cells another layer were deposited between the frustule and the silicon substrate. The double silicon layer made the sample surface yellow, the diatom frustules did not get darker due to a layer above and a layer beneath them. Only in some places where the frustules had loosened from the sample the sample were blue as the silicon nitride layer were the same as a single layer of silicon nitride. For the last two samples, thin film solar cell structures based on amorphous silicon deposited with PECVD were made. A sample there a aluminum coating were deposited on a glass substrate and diatom frustules were dried on top of the aluminum coating, afterwards a 5 µm thick film of amorphous silicon were deposited. SEM images showed that the film were distributed even across a diatom frustule and a light microscopy investigation showed that light were spread when it hit the diatom frustules, hence creating a longer mean path through the solar cell.

This master thesis has the objective to improve the mechanical properties of Super Duplex steel by adding Elkem Grain Refiner (EGR). EGR is commercial grain refiner where the active element is cerium. Cerium is a strong oxide and sulphide former whose inclusions are to act as sites for heterogeneous nucleation during casting. The cerium inclusions will form at low undercooling making it possible to grow equiaxed grains ahead of the solidifying front, reducing the size of the columnar zone normally seen in an ingot. By reducing the columnar zone the steel will have a larger zone of equiaxed, and smaller, grains improving the mechanical properties of the steel and reducing the segregation throughout the ingot. The experiment was performed with S4501 Super Duplex steel provided by Scana Steel Stavanger. The casting was done at Frekhaug stål, Bergen by adding 0.05, 0.075 and 0.1 wt% cerium to 120 kg casts. There were a total of 3 parallels, where 1 was cast at 1525ºC and 2 series at 1540ºC. The as cast structure was significantly refined, the columnar zone was reduced from 22 mm to being absent in the cast with most cerium added. The mechanical results show a linear increase in both yield and ultimate tensile strength with increasing amount of cerium. The elongation increased somewhat, but the impact toughness decreased linearly with increasing cerium content. The casting temperature did not seem to have an effect on the grain refining.

This master thesis is based on reduction experiments of pelletized Tyssedal ilmenite. The focus is set on the effect of changing parameters such as gas composition, gas flow, pellet size and pre-oxidation condition. The results are interpret regarding values of conversion, degree of metalization, XRF-analysis and microstructure images. Based on these results the aim was to find out how these parameters influenced reduction. Another goal was to reveal what hampered reduction inside both grains and pellets and caused the two stage reduction behaviour. EPMA was used to examine pellets and grains and to look for explanation for the slow reduction. XRD-analysis was decisive for explaining the importance of optimum pre-oxidation for the subsequent reduction. Pre-oxidized and green pellets with a size fraction of 8-10 mm were heated in a thermogravimetric analyzer. Ar was used during heating and cooling. Reduction occurred at 930-940 °C, with either CO, H2 or both for up to 45 minutes. Gas flows used were 4.8, 7 and 9 Nl/min. In total 18 experiments were carried out. Weight before and after reduction was measured, calculations of degree of metalization, XRF- and XRD-analysis was done. Microstructure images, point analysis, line scan and mapping were obtained by EPMA. Highest final values of conversion, c. 0.9, were obtained for pre-oxidized pellets from 2011 reduced with H2. Maximum degree of metalization, 94-99 % were obtained after reduction of pre-oxidized pellets from 2010. Final value of conversion and degree of metalization were 0.23 and c. 24 % higher for pre-oxidized pellets from 2010 than from 2011. Only green pellets from 2011 was harder to reduce with CO. Reducibility was not increased with the pre-oxidation conditions at ETI Tyssedal in January 2011. In spite of that the pre-oxidized pellets from 2010 and 2011 are made in the same manner and from the same raw material they react differently during reduction. This should be kept in mind when comparing results from different sources. Reduction of pre-oxidized pellets occurred in two distinct linear stages separated by a clear bend. A normal behaviour with steadily decreasing oxygen removal rate was observed with green pellets. An increased flow of CO from 7 Nl/min to 9 Nl/min resulted in decreased reduction. Examination by EPMA confirmed the presence of the barrier effect and higher amount of oxygen 0.6-1.0 mm from the surface. This indicates hampered reduction in grains and pellets. This combined with slow migration of CO/CO2 in the pellets can partly explain the unusual reduction behaviour. Pre-oxidized pellets from 2010 contains higher amount of the easily reducible M3O5, and less M2O3 compared to pre-oxidized pellets from 2011. The less reducible M3O4 was also found in the pre-oxidized pellets from 2011, making them harder to reduce.

I denne rapporten er det tatt for seg modellering av mykningsreaksjoner etter deformasjon for AA1200- og AA3103-legeringer. Grunnleggende teori rundt mykningsreaksjonene for disse legeringene er presentert. Det er hentet inn eksperimentelle data for sammenlikning av modellerte data, og resultatene er presentert i form av grafer for flytspenning og tabeller for rekrystallisert kornstørrelse. Det er brukt to matematiske modeller som bakgrunn for simuleringene. Den ene, Alsoft, dekker mikrostrukturelle endringer, og den andre, Presipal, dekker presipitering under varmebehandling.Resultatene for flytspenninger viser at simulering av varmebehandlingsforløp for AA1200-legeringer gjøres enkelt ved hjelp av Alsoft på grunn av fravær av presipitering som har innvirkning på varmebehandlingsforløpet. Simulering for AA3103-legeringer er mer krevende på grunn av den bremsende kraften som utøves på rekrystalliserte kim og rekrystalliserte korn som vokser på grunn av samtidig presipitering under varmebehandling. Her viser resultatene at en direkte kobling av modellene Alsoft og Presipal gir akseptabel tilpasning mellom eksperimentelle data og simulerte data for flytspenninger.Modellering av kornstørrelser viser, ved en sammenlikning med eksperimentelle resultater, at modellen har endel utfordringer med hensyn til dette. Verken størrelser eller trend er i alle tilfeller realistiske. Dette kan skyldes usikkerhet i parametere som brukes.Konklusjonen av arbeidet er at modellene i seg selv fungerer, men det kreves endel arbeid rundt fastsetting av parametere for at disse skal være anvendelige. Per dags dato blir det mye synsing for at resultatene skal vise akseptable avvik i forhold til eksperimentelle data.

Cathodic protection is an effective corrosion protection for structures submerged in seawater. In addition to applying the current need to lower the metal below the protection potential, a resulting increase in interfacial pH leads to precipitation of calcareous deposit. This deposited layer act as a barrier against oxygen diffusion on the surface, hence lowering the current demand of the structure. However, this layer will also hinder the thermal conductivity, and is therefore unwanted at the surface of heat exchangers.There are several factors influencing the precipitation of calcareous deposit. This includes potential, current, pH, seawater chemistry, flow rate etc. These factors have been thoroughly investigated for many years, but mostly with steel as substrate material. Specimens of steel coated with thermal sprayed aluminum (AlMg5) and aluminum plates of alloys AA5082, AA1050 and Al99.99 were exposed to seawater for 5-6 weeks, polarized to -1050 mV vs Ag/AgCl. The current demand was constantly recorded using KorrosjonsLogger©. Results from introductory experiments are also included. The unsealed TSA specimens had the highest current demand, but were also exposed to a higher seawater temperature than the other specimens, which has influenced the results. The sealed TSA specimens had a current demand around 10 mA/m2 after 6 weeks, which is 10 times lower than usually found on steel specimens. For the aluminum plates the trend showed that the higher alloying elements, the higher current density.The explanation for the low current demand compared to steel is that the cathodic area for aluminum is the intermetallic particles that are cathodic to the matrix, while for steel the whole surface act as a cathodic site. In addition, mechanisms during cathodic protection of aluminum lead to an isolation of the cathodic particle, hence continuously lowering the current demand.After exposure, the specimens were investigated in SEM and the composition on the surface was found using EDS. The SEM images showed a furry surface on the TSA specimen, but no familiar structure of calcareous deposit, which has been seen on steel specimens. For the aluminum plates, the intermetallic particles had been “dug out”, showing a pit between the particle and matrix. A typical structure of calcareous deposit was not seen here either. However, the EDS analysis showed an increase in both Mg and Ca content. The Mg/Ca ratio varied over the surface, with the highest ratio and content for the intermetallic particles for the aluminum plates and at the peaks of the surface at the TSA specimens. The Mg/Ca ratios found were much higher than found on steel.Both the SEM images and EDS analysis confirms the theory that there will precipitate less calcareous deposit on TSA surfaces compared to steel. Further studies will show whether the small content found will have an influence on the thermal conductivity.

Due to recent fractures in grade 70 anchor chains in the fish farming industry, chain supplier Erling Haug AS has initiated an examination of a sandblasted anchor they sell today. The sandblasted anchor chain is of the same strength category as the chains associated with the mentioned accidents. The mechanical properties of the sandblasted chain were characterized in 2010 in a student project work[1]. The report expressed concern about the sandblasted grade 70 chains since the material fulfilled requirements to be susceptible to hydrogen induced stress cracking (HISC). Therefore, further research on the material was needed. Another grade 70 chain, of an unknown origin, was obtained to make a comparison. This chain had fractured in-service. It was therefore necessary to characterize the mechanical properties of the fractured chain before examining the two chain grades susceptibility to HISC. The fractured chain was examined in terms of its microstructure, hardness, transition temperature and strength. An attempt was made to explain why the fractured chain failed in-service.The tensile test of the fractured chain revealed that its yield strength was 140MPa lower than that of the sandblasted chain. It was considered possible that the fracture chain were weaker than they should have been and that the reason it failed in-service was due to overload.The susceptibility of the two chains to HISC was tested by submerging tensile specimens into a 3.5% NaCl solution in CorTest Proof rings, while exposed to hydrogen. Specimens were either hot dip galvanized (HDG), protected by an external potential of -1050mV vs. Ag/AgCl or freely exposed at their corrosion potential. Some specimens were pulled in steps till fracture, while others were held at a constant load for 14 days.Different loading and corrosion potentials were applied to the two chain qualities during the HISC test, but the experiment was unable to provoke brittle fractures in either of them. No cracking was observed in the SEMivfor any of the HDG tensile specimens. It was found that hydrogen have no influence the mechanical properties of the HDG tensile specimens, while hydrogen appear to assist in the development of cracks on the polarized specimens.