Thesis (MScEng (Process Engineering))--University of Stellenbosch, 2010. / ENGLISH ABSTRACT: The world’s depleting fossil fuels and increasing greenhouse gas emissions have given rise to much research into renewable and cleaner energy. Biomass is unique in providing the only renewable source of fixed carbon. Agricultural residues such as Sugarcane Bagasse (SB) are feedstocks for ‘second generation fuels’ which means they do not compete with production of food crops. In South Africa approximately 6 million tons of raw SB is produced annually, most of which is combusted onsite for steam generation. In light of the current interest in bio-fuels and the poor utilization of SB as energy product in the sugar industry, alternative energy recovery processes should be investigated. This study looks into the thermochemical upgrading of SB by means of pyrolysis. Biomass pyrolysis is defined as the thermo-chemical decomposition of organic materials in the absence of oxygen or other reactants. Slow Pyrolysis (SP), Vacuum Pyrolysis (VP), and Fast Pyrolysis (FP) are studied in this thesis. Varying amounts of char and bio-oil are produced by the different processes, which both provide advantages to the sugar industry. Char can be combusted or gasified as an energy-dense fuel, used as bio-char fertilizer, or upgraded to activated carbon. High quality bio-oil can be combusted or gasified as a liquid energy-dense fuel, can be used as a chemical feedstock, and shows potential for upgrading to transport fuel quality. FP is the most modern of the pyrolysis technologies and is focused on oil production. In order to investigate this process a 1 kg/h FP unit was designed, constructed and commissioned. The new unit was tested and compared to two different FP processes at Forschungszentrum Karlsruhe (FZK) in Germany. As a means of investigating the devolatilization behaviour of SB a Thermogravimetric Analysis (TGA) study was conducted. To investigate the quality of products that can be obtained an experimental study was done on SP, VP, and FP. Three distinct mass loss stages were identified from TGA. The first stage, 25 to 110°C, is due to evaporation of moisture. Pyrolitic devolatilization was shown to start at 230°C. The final stage occurs at temperatures above 370°C and is associated with the cracking of heavier bonds and char formation. The optimal decomposition temperatures for hemicellulose and cellulose were identified as 290°C and 345°C, respectively. Lignin was found to decompose over the entire temperature range without a distinct peak. These results were confirmed by a previous study on TGA of bagasse. SP and VP of bagasse were studied in the same reactor to allow for accurate comparison. Both these processes were conducted at low heating rates (20°C/min) and were therefore focused on char production. Slow pyrolysis produced the highest char yield, and char calorific value. Vacuum pyrolysis produced the highest BET surface area chars (>300 m2/g) and bio-oil that contained significantly less water compared to SP bio-oil. The short vapour residence time in the VP process improved the quality of liquids. The mechanism for pore formation is improved at low pressure, thereby producing higher surface area chars. A trade-off exists between the yield of char and the quality thereof. FP at Stellenbosch University produced liquid yields up to 65 ± 3 wt% at the established optimal temperature of 500°C. The properties of the bio-oil from the newly designed unit compared well to bio-oil from the units at FZK. The char properties showed some variation for the different FP processes. At the optimal FP conditions 20 wt% extra bio-oil is produced compared to SP and VP. The FP bio-oil contained 20 wt% water and the calorific value was estimated at 18 ± 1 MJ/kg. The energy per volume of FP bio-oil was estimated to be at least 11 times more than dry SB. FP was found to be the most effective process for producing a single product with over 60% of the original biomass energy. The optimal productions of either high quality bio-oil or high surface area char were found to be application dependent. / AFRIKAANSE OPSOMMING: As gevolg van die uitputting van fossielbrandstofreserwes, en die toenemende vrystelling van kweekhuisgasse word daar tans wêreldwyd baie navorsing op hernubare en skoner energie gedoen. Biomassa is uniek as die enigste bron van hernubare vaste koolstof. Landbouafval soos Suikerriet Bagasse (SB) is grondstowwe vir ‘tweede generasie bio-brandstowwe’ wat nie die mark van voedselgewasse direk affekteer nie. In Suid Afrika word jaarliks ongeveer 6 miljoen ton SB geproduseer, waarvan die meeste by die suikermeulens verbrand word om stoom te genereer. Weens die huidige belangstelling in bio-brandstowwe en ondoeltreffende benutting van SB as energieproduk in die suikerindustrie moet alternatiewe energie-onginningsprosesse ondersoek word. Hierdie studie is op die termo-chemiese verwerking van SB deur middel van pirolise gefokus. Biomassa pirolise word gedefinieer as die termo-chemiese afbreking van organiese bio-materiaal in die afwesigheid van suurstof en ander reagense. Stadige Pirolise (SP), Vakuum Pirolise (VP), en Vinnige Pirolise word in hierdie tesis ondersoek. Die drie prosesse produseer veskillende hoeveelhede houtskool en bio-olie wat albei voordele bied vir die suikerindustrie. Houtskool kan as ‘n vaste energie-digte brandstof verbrand of vergas word, as bio-houtskoolkompos gebruik word, of kan verder tot geaktiveerde koolstof geprosesseer word. Hoë kwaliteit bio-olie kan verbrand of vergas word, kan as bron vir chemikalië gebruik word, en toon potensiaal om in die toekoms opgegradeer te kan word tot vervoerbrandstof kwaliteit. Vinnige pirolise is die mees moderne pirolise tegnologie en is op bio-olie produksie gefokus. Om die laasgenoemde proses te toets is ‘n 1 kg/h vinnige pirolise eenheid ontwerp, opgerig en in werking gestel. Die nuwe pirolise eenheid is getoets en vegelyk met twee verskillende vinnige pirolise eenhede by Forschungszentrum Karlsruhe (FZK) in Duitsland. Termo-Gravimetriese Analise (TGA) is gedoen om die ontvlugtigingskenmerke van SB te bestudeer. Eksperimentele werk is verrig om die kwaliteit van produkte van SP, VP, vinnige pirolise te vergelyk. Drie duidelike massaverlies fases van TGA is geïdentifiseer. Die eerste fase (25 – 110°C) is as gevolg van die verdamping van vog. Pirolitiese ontvlugtiging het begin by 230°C. Die finale fase (> 370°C) is met die kraking van swaar verbindings en die vorming van houtskool geassosieer. Die optimale afbrekingstemperatuur vir hemisellulose en sellulose is as 290°C en 345°C, respektiewelik, geïdentifiseer. Daar is gevind dat lignien stadig oor die twede en derde fases afgebreek word sonder ‘n duidelike optimale afbrekingstemperatuur. Die resultate is deur vorige navorsing op TGA van SB bevestig. SP en VP van bagasse is in dieselfde reaktor bestudeer, om ‘n akkurate vergelyking moontlik te maak. Beide prosesse was by lae verhittingstempo’s (20°C/min) ondersoek, wat gevolglik op houtskoolformasie gefokus is. SP het die hoogste houtskoolopbrengs, met die hoogste verbrandingsenergie, geproduseer. VP het hootskool met die hoogste BET oppervlakarea geproduseer, en die bio-olie was weens ‘n dramatiese afname in waterinhoud van beter gehalte. Die meganisme vir die vorming van ‘n poreuse struktuur word deur lae atmosferiese druk verbeter. Daar bestaan ‘n inverse verband tussen die kwantiteit en kwaliteit van die houtskool. Vinnige pirolise by die Universiteit van Stellenbosch het ‘n bio-olie opbrengs van 65 ± 3 massa% by ‘n vooraf vasgestelde optimale temperatuur van 500°C geproduseer. Die eienskappe van bio-olie wat deur die nuwe vinnige pirolise eenheid geproduseer is het goed ooreengestem met die bio-olie afkomstig van FZK se pirolise eenhede. Die houtskool eienskappe van die drie pirolise eenhede het enkele verskille getoon. By optimale toestande vir vinnige pirolise word daar 20 massa% meer bio-olie as by SP en VP geproduseer. Vinnige pirolise bio-olie het ‘n waterinhoud van 20 massa% en ‘n verbrandingswarmte van 18 ± 1 MJ/kg. Daar is gevind dat ten opsigte van droë SB die energie per enheidsvolume van bio-olie ongeveer 11 keer meer is. Vinnige pirolise is die mees doeltreffende proses vir die vervaardiging van ‘n produk wat meer as 60% van die oorspronklike biomassa energie bevat. Daar is gevind dat die optimale hoeveelhede van hoë kwaliteit bio-olie en hoë oppervlakarea houtskool doelafhanklik is.

Doctor of Philosophy / Department of Biological and Agricultural Engineering / Wenqiao Yuan / Donghai Wang / Corncobs were used as the feedstock to investigate the effect of operating conditions and crude glycerol (solvent) on bio-oil production. The highest bio-oil yield of 33.8% on the basis of biomass dry weight was obtained at 305°C, 20 min retention time, 10% biomass content, 0.5% catalyst loading. At selected conditions, bio-oil yield based on the total weight of corn cobs and crude glycerol increased to 36.3% as the crude glycerol/corn cobs ratio increased to 5. Furthermore, the optimization of operating conditions was conducted via response surface methodology. A maximum bio-oil yield of 41.3% was obtained at 280°C, 12min, 21% biomass content, and 1.56% catalyst loading. A highest bio-oil carbon content of 74.8% was produced at 340°C with 9% biomass content. A maximum carbon recovery of 25.2% was observed at 280°C, 12min, 21% biomass content, and 1.03% catalyst loading. The effect of biomass ecotype and planting location on bio-oil production were studied on big bluestems. Significant differences were found in the yield and elemental composition of bio-oils produced from big bluestem of different ecotypes and/or planting locations. Generally, the IL ecotype and the Carbondale, IL and Manhattan, KS planting locations gave higher bio-oil yield, which can be attributed to the higher total cellulose and hemicellulose content and/or the higher carbon but lower oxygen contents in these feedstocks. Bio-oil from the IL ecotype also had the highest carbon and lowest oxygen contents, which were not affected by the planting location. In order to better understand the mechanisms of hydrothermal conversion, the interaction effects between cellulose, hemicellulose and lignin in hydrothermal conversion were studied. Positive interaction between cellulose and lignin, but negative interaction between cellulose and hemicellulose were observed. No significant interaction was found between hemicelluose and lignin. Hydrothermal conversion of corncobs, big bluestems, switchgrass, cherry, pecan, pine, hazelnut shell, and their model biomass also were conducted. Bio-oil yield increased as real biomass cellulose and hemicellulose content increased, but an opposite trend was observed for low lignin content model biomass.

Os desafios energéticos que enfrentamos no presente, ou seja, a busca por novas fontes renováveis de energia tem recebido muita atenção nos últimos anos. Na tentativa de minimizar os problemas ambientais tem-se investido no tratamento de resíduos, pois gerenciar resíduos como o lodo de esgoto é difícil e caro. Entretanto, o lodo pode ser utilizado como matéria prima na produção de energia através da pirólise onde quatro frações são geradas: líquida aquosa, líquida orgânica, sólida e gasosa. A gama de substâncias agregadas ao lodo que podem ser convertidas com a pirólise por vários mecanismos reacionais durante as etapas do processo podem conferir características únicas às frações combustíveis formadas principalmente por hidrocarbonetos de cadeias longas. As formas de tratamento adequado para a matéria prima para maximizar o rendimento de combustível ajustado ao modelo de reator também devem ser considerados. Portanto, o conhecimento das características químicas e físico-químicas das biomassas, ou seja, o lodo de esgoto, bem como as formas de conversão desta para a produção de energia e as características dos combustíveis produzidos é fundamental. / As a way to encourage the treatment and use of sewage sludge and assist in the resolution of the challenges that have been faced today, with regard to the search for alternative energy sources, sewage sludge can be used as raw material for obtaining fuels and chemicals via pyrolysis whose products can be obtained into four fractions: aqueous, organic, solid and gaseous. The range of substances added to the sludge that can be converted to pyrolysis several reaction pathways during stages of the process may confer unique characteristics to fuel fractions formed mainly by long-chain hydrocarbons. Besides, the forms of treatment used in the sewage sludge have to be appropriate to the reactor model to improve the yield of fuel. Therefore, knowledge of the chemical and physico-chemical properties of that biomass, along with the ways for the conversion of that biomass into energy and the characteristics of the fuels produced have to be known to improve the use of sewage sludge in the pyrolysis process.

Lignin is an important component of biomass accounting for up to 30% by weight but up to 40% of the total energy content of the plant. As the push towards alternative fuels develops, more and more amounts of lignin will be gathered and used predominately as low grade boiler fuel to run primary processes. We argue there is usefulness in the conversion of lignin into value added specialty chemicals and fuels. In this work, a new approach for hydrodeoxygenation of lignin model compounds using platinum as the catalyst and organic solvent as the reaction medium was conducted, and the results were compared with those obtained using water as the reaction medium. It is shown that the organic solvent, with its increased hydrogen solubility, is able to hydrogenate the model compound with the same effect at lower temperature, hydrogen pressure, and time.

In the search for fossil fuel alternatives the production of bio-oil through the pyrolysis of biomass is one method which has shown evidence of scalability, meaning that the technology could be scaled up for the processing of biomass on the order of tons per day. Pyrolysis is the thermal degradation of compounds in the absence of oxygen. Of particular interest is the pyrolysis of sustainable energy crops such as Loblolly pine (Pinus taeda). The goal of this study is to develop a new method of characterizing the fast pyrolysis of biomass for the advancement of reactor design. The objectives are to determine bulk kinetic coefficients for the isothermal fast pyrolysis of biomass, evaluate the interchangeability of fast and slow pyrolysis kinetic parameters and compare generally accepted pyrolysis mechanisms derived from a common data set. A technical objective is to apply the most suitable derived kinetic parameters to model pyrolysis within a moving bed reactor. A novel fast pyrolysis micro-reactor is presented along with its design and development process. The micro-reactor allows for the control over both temperature and residence time of the reacting biomass. This system provides the experimental data for the characterization of biomass pyrolysis kinetic parameters. Thermal validation tests are presented and experimental yield results are given for raw Loblolly Pine, Avicel cellulose and Beechwood xylan for the derivation of kinetic descriptors. Cellulose and xylan results show good agreement with literature when the proper experimental conditions are met and whole wood pyrolysis results clearly demonstrate the dissimilarity between fast and slow pyrolysis apparent kinetic rates. The experimental results are then used to evaluate five different pyrolysis kinetic model configurations: single component global pyrolysis, two component global pyrolysis, product based pyrolysis, pseudo-component based pyrolysis and pseudo-component pyrolysis with an intermediate solid compound. Pseudo-component models are of particular interest because they may provide a generalized model, parameterized by the fractional composition of cellulose, hemicellulose and lignin in biomass species. Lignin pyrolysis yields are calculated to evaluate the suitability of a pseudo-component parallel non-competing superposition pyrolysis model. Lignin yields are estimated by taking the difference between whole wood pyrolysis and predicted cellulose and hemicellulose pyrolysis behaviors. The five models are then evaluated by comparison of predicted yields to the results for the pyrolysis of Scots pine (Pinus sylvestris) and Norway spruce (Picea abies). Model evaluations show that pseudo-component superposition is not suitable as a generic pyrolysis model for the fast pyrolysis of biomass observed using the micro-reactor. Further analytical evaluations indicate that the assumption of parallel non-competing reactions between pseudo-components is not valid. Among the other models investigated the intermediate solid compound model showed the best fit to the verification experimentation results followed closely by the two component global model. Finally, the derived kinetic parameters are applied to the design of moving bed vacuum pyrolysis reactors which provide for the separation of heat and mass transfer pathways, resulting in the reduction of char entrainment and secondary reactions within collected bio-oils. Reaction kinetics and porous bed heat and mass transfer are accounted for within the bed model. Model development and predictive results are presented and sensitivity to activation energy variations investigated.

A study was carried out into the computational fluid dynamic simulation of bio-oil combustion. Measurements were taken in an empirical burner to obtain information regarding the flow behaviour. A surrogate fuel was developed to mimic the unique chemical and physical properties of bio-oil combustion. The resulting computational model of the burner domain and surrogate fuel was compared with empirical data. The bio-oil model displayed a good agreement with the data in terms of the combustion behaviour, but was limited by the uncertain flow solution associated with the burner used.

Biomass fast pyrolysis liquid, also known as bio-oil, is a promising renewable fuel for heat and power generation; however, implementing crude bio-oil in some current combustion systems can degrade combustion performance and emissions. In this study, optimizing fuel properties to improve combustion is considered. Various bio-oils with different fuel properties are tested in a pilot stabilized spray burner under very close flow conditions. Effects of solids, ash and water content of bio-oil as well as ethanol blending were examined. The results show the amount of solids and ash fractions of the fuel were correlated with combustion efficiency. The CO and unburned hydrocarbon emissions decreased with both water and ethanol content. Increasing the fuel’s volatile content by blending in ethanol has been shown to improve flame stability. Also, the organic fraction of particulate matter emissions was found to be a strong function of the thermogravimetric analysis residue of the fuel.

A study was carried out into the computational fluid dynamic simulation of bio-oil combustion. Measurements were taken in an empirical burner to obtain information regarding the flow behaviour. A surrogate fuel was developed to mimic the unique chemical and physical properties of bio-oil combustion. The resulting computational model of the burner domain and surrogate fuel was compared with empirical data. The bio-oil model displayed a good agreement with the data in terms of the combustion behaviour, but was limited by the uncertain flow solution associated with the burner used.

Biomass fast pyrolysis liquid, also known as bio-oil, is a promising renewable fuel for heat and power generation; however, implementing crude bio-oil in some current combustion systems can degrade combustion performance and emissions. In this study, optimizing fuel properties to improve combustion is considered. Various bio-oils with different fuel properties are tested in a pilot stabilized spray burner under very close flow conditions. Effects of solids, ash and water content of bio-oil as well as ethanol blending were examined. The results show the amount of solids and ash fractions of the fuel were correlated with combustion efficiency. The CO and unburned hydrocarbon emissions decreased with both water and ethanol content. Increasing the fuel’s volatile content by blending in ethanol has been shown to improve flame stability. Also, the organic fraction of particulate matter emissions was found to be a strong function of the thermogravimetric analysis residue of the fuel.

Face à l'épuisement des réserves en carburants fossiles et afin de subvenir à une demande toujours croissante en énergie pour le transport, les scientifiques se tournent désormais vers une ressource quasi-inépuisable et renouvelable : la biomasse. Au sein de la biomasse, les microalgues représentent une source potentielle de biocarburant car elles peuvent contenir des fortes teneurs en lipides et hydrocarbures. Leur croissance extrêmement rapide, l'utilisation du CO2 et de l'énergie solaire pour leur croissance et l'absence de compétition avec l'agriculture traditionnelle confèrent aux micro-algues une très forte attractivité. Deux voies de conversion ont été abordées. Dans un premier temps, nous avons étudié l'hydroconversion de triglycérides avec une molécule modèle (GTO) et charges réelles (huile de poisson et huile de Nannochloropsis obtenu par extraction au CO2 supercritique) sur catalyseurs de type CoMoS et NiMoS sur alumine. Des rendements élevés en alcanes (60- 70%pds) semblables à des carburants fossiles ont été obtenus. Dans un second temps, nous avons étudié la liquéfaction hydrothermale de la Spiruline, peu représentative d'algues lipidiques mais disponible, et d'autres algues (Porphyridium cruentum, Nannochloropsis sp., Ourococcus, Dunaliela salina) pour optimiser ce procédé selon les différentes conditions opératoires avec l'obtention d'un rendement optimal en bio-huile de 35%pds. Cependant, les teneurs élevées en azote et oxygène (8-10%pds) de la bio-huile ne permettent pas de la valoriser directement comme carburant. Cela nous a amené à effectuer une valorisation de la bio-huile avec des catalyseurs hétérogènes de type CoMo, NiMoS, NiMoS-Y supportés sur alumine et SrMoO4-N pour éliminer l'azote et l'oxygène de la bio-huile. Le catalyseur avec le résultat plus satisfaisant a été le SrMoO4-N, avec une teneur finale en alcanes de 70%pds / Given the depletion of fossil fuels and to meet a growing demand for transportation energy, scientists are now turning towards an almost inexhaustible and renewable resource: biomass. As biomass, microalgae represent a potential source of biofuel because they may contain high levels of lipids and hydrocarbons. Their extremely fast growth, the use of CO2 and solar energy for their growth and the absence of competition with traditional agriculture makes microalgae very appealing. Two thermochemical routes of valorisation of μ-algae have been investigated. At first, we studied the hydroconversion of triglycerides with a model molecule (GTO) and then real feedstocks (fish oil and Nannochloropsis oil obtained by supercritical CO2 extraction) on CoMoS and NiMoS type catalysts. High yields of alkanes (60-70 wt%) similar to fossil fuels were obtained. In a second step, we studied the hydrothermal liquefaction of Spirulina and other algae without heterogeneous catalyst to optimize the process for different operating conditions with the obtention of a maximum bio-oil yield of 35 %wtt. However, the high levels of nitrogen and oxygen (8-10 wt%) does not allow to directly use it as fuel . This has led us to perform an upgrading of the bio-oil with heterogeneous catalysts of CoMo, NiMoS, NiMoS-Y and SrMoO4-N types to remove nitrogen and oxygen in the bio-oil. The most performant catalyst was SrMoO4-N, with a final content of 70 %wt of alkanes