In the kingdom of Saudi Arabia (KSA), the current seawater desalination technologies are completely relying on burning unsustainable crude oil as their main energy driver. Saudi authorities have realized that the KSA is not going to be protected from the future global energy crisis and have started to set up a plan to diversify its energy resources. Membrane Distillation (MD) has emerged as an attractive alternative desalination process. It combines advantages from both thermal and membrane-based technologies and holds the potential of being a cost-effective separation process that can utilize low-grade waste heat or renewable energy. MD has four different configurations; among them is Air Gap Membrane Distillation (AGMD) which is the second most commonly tested and the most commercially available pilot-plant design. AGMD has a stagnant thin layer of air between the membrane and the condensation surface. This layer introduces a mass transfer resistance that makes the process require a large membrane surface area if a large quantity of fresh water is desired. This dissertation reports on experimental and theoretical work conducted to enhance the AGMD flux by removing non-condensable gases from the module and replacing it with either vacuum, liquid water or porous materials. At first, a mathematical model for AGMD was developed and validated experimentally to create a baseline for improvements that could be achieved after the removal of non-condensable gases. The mathematical model was then modified to simulate the process under vacuum where it showed a flux enhancement that reached 286%. The Water Gap Membrane Distillation (WGMD) configuration improved the flux by almost the same percentage. Since enhancing the flux is expected to increase temperature polarization effects, a theoretical study was conducted on the effect of temperature polarization in a Vacuum Membrane Distillation (VMD) configuration. The study showed that the effect of temperature polarization at small temperature difference (3-7) degree Celsius between the bulk feed and coolant temperatures is significantly high. This may indicate the importance of mitigating the effect of temperature polarization in large scale modules operating at small temperature difference across the membrane. The dissertation concluded with some recommendations for future work.

Desalination is a technique of producing fresh water from the saline water. Industrial desalination of sea water is becoming an essential part in providing sustainable source of fresh water for a large number of countries around the world. Thermal process being the oldest and most dominating for large scale production of freshwater in today's world. Multi-Stage Flash (MSF) distillation process has been used for many years and is now the largest sector in the desalination industry. In this work, a steady state mathematical model of Multistage Flash (MSF) desalination process is developed and validated against the results reported in the literature using gPROMS software. The model is then used for further investigation. First, a steady state calcium carbonate fouling resistance model has been developed and implemented in the full MSF mathematical model developed above using gPROMS modeling tool. This model takes into consideration the effect of stage temperature on the calcium carbonate fouling resistance in the flashing chambers in the heat recovery section, heat rejection section, and brine heaters of MSF desalination plants. The effect of seasonal variation of seawater temperature and top brine temperature on the calcium carbonate fouling resistance has been studied throughout the flashing stage. In addition, the total annual operating cost of the MSF process is selected to minimise, while optimising the operating parameters such as seawater rejected flow rate, brine recycle flow rate and steam temperature at different seawater temperature and fouling resistance. Secondly, an intermediate storage between the plant and the client is considered to provide additional flexibility in design and operation of the MSF process throughout the day. A simple polynomial based dynamic seawater temperature and different freshwater demand correlations are developed based on actual data. For different number of flash stages, operating parameters such as seawater rejected flow rate and brine recycle flow rate are optimised, while the total annual operating cost of the MSF process is selected to minimise.The results clearly show that the advantage of using the intermediate storage tank adds flexible scheduling in the MSF plant design and operation parameters to meet the variation in freshwater demand with varying seawater temperatures without interrupting or fully shutting down the plant at any time during the day by adjusting the number of stages. Furthermore, the effect of non-condensable gases (NCG) on the steady state mathematical model of MSF process is developed and implemented in the MSF model developed earlier. Then the model is used to study effect of NCG on the overall heat transfer coefficient. The simulation results showed a decrease in the overall heat transfer coefficient values as NCG concentrations increased. The model is then used to study the effect of NCG on the design and operation parameters of MSF process for fixed water demand. For a given plant configuration (fixed design) and at different seawater and steam temperatures, a 0.015 wt. % of NCG results in significantly different plant operations when compared with those obtained without the presence of NCG. Finally, for fixed water demand and in the presence of 0.015 wt. % NCGs, the performance is evaluated for different plant configurations and seawater temperature and compared with those obtained without the presence of NCG.

The problem of the forced film condensation heat transfer of pure steam and steam-air mixture flowing downward a tier of horizontal cylinders is investigated numerically and experimentally. Liquid and vapor-air mixture boundary layers were solved by an implicit finite difference scheme. The effects of the free stream non-condensable gas (air) concentration, free stream velocity (Reynolds number), cylinder diameter, temperature difference and angle of inclination on the condensation heat transfer are analyzed. Inline and staggered tubes arrangements are considered. The mathematical model takes into account the effect of staggering of the cylinders and how condensation is affected at the lower cylinders when condensate does not fall on to the center line of the cylinders. An experimental setup was also manufactured and mounted at METU workshop. A set of experiments were conducted to observe the condensation heat transfer phenomenon and to verify the theoretical results. Condensation heat transfer results are available in ranges from (U&amp / #61605 / = 1 - 30 m/s) for free stream velocity, (m1,&amp / #61605 / = 0.01 -0.8) for free stream air mass fraction, (d = 12.7 -50.8 mm) for cylinder diameter and (T&amp / #61605 / -Tw =10-40 K) for temperature difference. Results show that / a remarked reduction in the vapor side heat transfer coefficient is noticed when very small amounts of air mass fractions present in the vapor. In addition, it decreases by increasing in the cylinder diameter and the temperature difference. On the other hand, it increases by increasing the free stream velocity (Reynolds number). Average heat transfer coefficient at the middle and the bottom cylinders increases by increasing the angle of inclination, whereas, no significant change is observed for that of the upper cylinder. Although some discrepancies are noticed, the present study results are inline and in a reasonable agreement with the theory and experiment in the literature. Down the bank, a rapid decrease in the vapor side heat transfer coefficient is noticed. It may be resulted from the combined effects of inundation, decrease in the vapor velocity and increase in the non-condensable gas (air) at the bottom cylinders in the bank. Differences between the present study results and the theoretical and the experimental data may be resulted from the errors in the numerical schemes used. These errors include truncation and round off errors, approximations in the numerical differentiation for interfacial fluxes at the vapor-liquid interface, constant properties assumption and approximations in the initial profiles. Mixing and re-circulation in the steam-air mixture at the lower tubes may be the other reasons for these deviations.

Die in diesem Bericht vorgestellten Experimente betreffen die Kondensation von Dampf in horizontalen bzw. leicht geneigten Rohren bei hohen auftretenden Temperaturdifferenzen bis zu über 200 K. Weitere Besonderheiten sind die detaillierte Untersuchung des transienten Verhaltens eines nichtkondensierbaren Gases mit einer neuartigen Messtechnik sowie die Ermittlung des Einflusses des Gases auf den Kondensationsvorgang. Beim Experiment wurden schnelle Übergangsvorgänge ausgelöst, indem ein in einer Kühlwanne liegendes, leicht geneigtes Wärmeübertragerrohr plötzlich mit dem Dampfraum eines unter Druck stehenden Kessels verbunden wurde. Dabei wurden im Rohr unterschiedliche Anfangsbedingungen hinsichtlich der Vorlage von nichtkondensierbaren Gasen (in diesem Falle Luft) eingestellt. Es wurden Versuche mit Atmosphärendruck, mit erhöhtem Druck, aber auch mit vorheriger Evakuierung des Versuchsrohrs durchgeführt. Durch eine Instrumentierung mit neuartigen Nadelsonden, die eine Phasendetektion kombiniert mit einer schnellen lokalen Temperaturmessung ermöglichen, konnte die Umverteilung von Dampf, Kondensat und nichtkondensierbarem Gas als Funktion der Zeit beobachtet werden. Damit bieten die erhaltenen Daten die Möglichkeit, insbesondere die in den Thermohydraulikprogrammen vorhandenen Optionen zur Berechnung der Ausbreitung von nichtkondensierbaren Gasen unter transienten Bedingungen zu validieren.

La torréfaction est un traitement thermique opéré entre 200 et 300 °C en atmosphère inerte améliorant certaines propriétés de la biomasse, afin d’utiliser celle-ci comme biocombustible. Nos travaux portent spécifiquement sur la caractérisation des espèces condensables produites en torréfaction, aux échelles laboratoire et pilote. En laboratoire, des échantillons de pin, frêne, miscanthus et paille de blé ont été torréfiés à 250, 280 et 300 °C en réacteur à lit fixe. Les espèces condensables ont ensuite été analysées par GC-MS, GC-GC et HPLC-MS. Cette analyse a permis d’identifier une centaine d’espèces, dont une vingtaine, quantifiée, représente 77 % des condensables. À l’échelle pilote, un réacteur continu a été conçu, amélioré et caractérisé pour torréfier quelques kg.h-1 de biomasse. Un système de récupération multi-étagée des condensables a été développé. Des essais de torréfaction ont montré que les fractions condensées présentent des compositions chimiques différentes. / Orrefaction is mild thermal treatment carried out between 200 and 300 °C, in an inert atmosphere, improving properties of biomass, in order to use it as a biocombustible. This study focuses on the characterisation of the condensable species released during torrefaction, at laboratory and pilot scale. In the laboratory, some samples of pine, ash wood, miscanthus and wheat straw were torrefied at 250, 280 and 300 °C in a fixed bed reactor. The condensable species were analysed by GC-MS, GC-GC and HPLC-MS. Around a hundred of species were identified, including around twenty were quantified and represented 77 % of the condensable species. At pilot scale, a continuous reactor was designed, improved and characterised to treat several kg.h-1 of biomass. A multi-step recovery system for the condensable species was developped. Torrefaction experiments showed that the condensed fractions had different chemical compositions

Le but principal de ce projet doctoral est de déterminer le potentiel d'amélioration de l'efficacité énergétique du système de captage de carbone dans les stations thermiques de production d'électricité, par l'intégration optimale des éjecteurs monophasiques. Il s'agit du système de captage postcombustion du dioxyde de carbone (CO2) par absorption/désorption utilisant la monoéthanolamine (MEA). Les éjecteurs intégrés utilisent des rejets thermiques de 100 °C qu'on retrouve dans les stations thermiques de production d'électricité. La revalorisation de ces rejets permet la substitution partielle de vapeur de turbine à coût élevé, qui serait autrement prise de la centrale thermique. Le deuxième objectif de la thèse est d'évaluer expérimentalement la performance d'un éjecteur à vapeur où le fluide secondaire de l'éjecteur est un mélange de vapeur d'eau et d'un gaz non-condensable, dans le cas présent, le CO2. Deux tuyères d'éjecteur à vapeur, d'un diamètre de 4.60 mm et 4.23 mm, ont été évaluées sur une plage de niveaux de CO2 dans le fluide secondaire, jusqu'à environ 40% en masse. La pression primaire était maintenue à 450 kPa avec une surchauffe à 10 °C et la pression secondaire était de 70 kPa. On a constaté que la pression critique ne changeait pas à mesure que la fraction massique de CO2 dans le fluide secondaire augmentait. Cependant, le rapport d'entraînement a augmenté de façon linéaire sur la plage expérimentale. Une amélioration de 23% du rapport d'entraînement par rapport à la vapeur pure a été observée lorsque le fluide secondaire contient 42% de CO2 par masse. Ce comportement contraste nettement avec le comportement observé expérimentalement d'un éjecteur à vapeur pure, où une augmentation du rapport d'entraînement se produit au détriment d'une diminution de la pression critique. Trois articles détaillés ont été publiés sur divers scénarios d'intégration d'un éjecteur à vapeur dans un procédé de captage d'absorption/désorption. Le solvant de référence était de 20% en masse de monoéthanolamine (MEA). Trois configurations principales ont été étudiées, selon le choix du fluide utilisé pour produire la vapeur secondaire : éjecteur sur condensat, éjecteur sur pauvre ou éjecteur sur riche. La première publication de revue scientifique a porté sur le procédé de désorption et a présenté une méthode de raccourci basée sur les propriétés du mélange CO2-MEA-H2O à l'équilibre. Les simulations ont révélé des réductions dans la quantité requise d'énergie de haute qualité, de 10 à 25%. Un simulateur de procédé commercial, Aspen Plus, a été utilisé pour les deux autres publications. Dans la deuxième publication de revue scientifique, le module cinétique rate-based a été utilisé, au lieu du module d'équilibre, pour la modélisation de l'absorbeur et du désorbeur, permettant des évaluations énergétiques plus près des valeurs qu'on retrouve dans la littérature courante. Une étude a été réalisée pour comparer un scénario de préchauffage de la vapeur primaire par des rejets thermiques externes avec un scénario d'intégration de la chaleur interne. Cette deuxième publication a montré des économies d'énergie de haute qualité, de 10 à 14%, les scénarios avantageux ayant été «éjecteur sur condensat» et «éjecteur sur pauvre». / Abstract : The main goal of the doctoral project is to determine to what extent the optimal integration of single-phase ejectors might reduce the large amount of energy required to capture carbon dioxide from electric power generation facilities. More specifically, the objective is to determine if ejectors can be advantageously integrated into a post-combustion absorption/desorption carbon dioxide (CO2) capture process using monoethanolamine (MEA). The integrated ejectors will use waste heat of 100 °C from the electric power plant. The upgraded waste heat can partially replace valuable turbine steam that would otherwise be taken from the power plant. The second objective of the thesis is to experimentally evaluate the performance of a steam ejector where the ejector secondary fluid is a mixture of steam and a non-condensable gas, in this case CO2. Two steam ejector nozzles, of 4.60 mm and 4.23 mm diameter, were evaluated over a range of secondary fluid CO2 levels, up to 42% by mass. The primary pressure was maintained at 450 kPa with 10 °C superheat and the secondary pressure was 70 kPa. It was found that the critical exit pressure did not change as the mass fraction of CO2 in the secondary fluid increased. The entrainment ratio, however, increased approximately linearly over the experimental range. An improvement of 23% in the entrainment ratio, as compared with pure steam, was found when the secondary fluid contains 42% CO2 by mass. This behaviour is in sharp contrast to the experimentally observed behaviour of a pure steam ejector, where an increase in entrainment ratio comes at the expense of a decrease in the ejector exit critical pressure. Three published papers investigated various scenarios for the integration of a steam injector into an absorption/desorption post-combustion capture process. The reference solvent was 20% weight monoethanolamine (MEA). Three principal configurations were studied, according to the choice for the liquid flow used to produce the ejector secondary steam: ejector on condensate, ejector on lean or ejector on rich. The first journal publication focused on the desorption process and presented a shortcut method based on CO2-MEA-H2O equilibrium vapour liquid data. The simulations revealed reductions in the required amount of valuable energy from 10 to 25%. A commercial process simulator, Aspen Plus, was used for two other publications. In the second journal publication, the kinetic rate-based module was employed to model the absorber and desorber, providing energy evaluations closer to values in the open literature. A study was included comparing preheating the primary steam with waste heat or by heat integration. The rate-based simulation found valuable energy savings of 10 to 14%, with the "ejector on condensate" and "ejector on lean" again being the advantageous scenarios.

Die in diesem Bericht vorgestellten Experimente betreffen die Kondensation von Dampf in horizontalen bzw. leicht geneigten Rohren bei hohen auftretenden Temperaturdifferenzen bis zu über 200 K. Weitere Besonderheiten sind die detaillierte Untersuchung des transienten Verhaltens eines nichtkondensierbaren Gases mit einer neuartigen Messtechnik sowie die Ermittlung des Einflusses des Gases auf den Kondensationsvorgang. Beim Experiment wurden schnelle Übergangsvorgänge ausgelöst, indem ein in einer Kühlwanne liegendes, leicht geneigtes Wärmeübertragerrohr plötzlich mit dem Dampfraum eines unter Druck stehenden Kessels verbunden wurde. Dabei wurden im Rohr unterschiedliche Anfangsbedingungen hinsichtlich der Vorlage von nichtkondensierbaren Gasen (in diesem Falle Luft) eingestellt. Es wurden Versuche mit Atmosphärendruck, mit erhöhtem Druck, aber auch mit vorheriger Evakuierung des Versuchsrohrs durchgeführt. Durch eine Instrumentierung mit neuartigen Nadelsonden, die eine Phasendetektion kombiniert mit einer schnellen lokalen Temperaturmessung ermöglichen, konnte die Umverteilung von Dampf, Kondensat und nichtkondensierbarem Gas als Funktion der Zeit beobachtet werden. Damit bieten die erhaltenen Daten die Möglichkeit, insbesondere die in den Thermohydraulikprogrammen vorhandenen Optionen zur Berechnung der Ausbreitung von nichtkondensierbaren Gasen unter transienten Bedingungen zu validieren.

Desalination is a technique of producing fresh water from the saline water. Industrial desalination of sea water is becoming an essential part in providing sustainable source of fresh water for a large number of countries around the world. Thermal process being the oldest and most dominating for large scale production of freshwater in today¿s world. Multi-Stage Flash (MSF) distillation process has been used for many years and is now the largest sector in the desalination industry. In this work, a steady state mathematical model of Multistage Flash (MSF) desalination process is developed and validated against the results reported in the literature using gPROMS software. The model is then used for further investigation. First, a steady state calcium carbonate fouling resistance model has been developed and implemented in the full MSF mathematical model developed above using gPROMS modeling tool. This model takes into consideration the effect of stage temperature on the calcium carbonate fouling resistance in the flashing chambers in the heat recovery section, heat rejection section, and brine heaters of MSF desalination plants. The effect of seasonal variation of seawater temperature and top brine temperature on the calcium carbonate fouling resistance has been studied throughout the flashing stage. In addition, the total annual operating cost of the MSF process is selected to minimise, while optimising the operating parameters such as seawater rejected flow rate, brine recycle flow rate and steam temperature at different seawater temperature and fouling resistance. Secondly, an intermediate storage between the plant and the client is considered to provide additional flexibility in design and operation of the MSF process throughout the day. A simple polynomial based dynamic seawater temperature and different freshwater demand correlations are developed based on actual data. For different number of flash stages, operating parameters such as seawater rejected flow rate and brine recycle flow rate are optimised, while the total annual operating cost of the MSF process is selected to minimise.The results clearly show that the advantage of using the intermediate storage tank adds flexible scheduling in the MSF plant design and operation parameters to meet the variation in freshwater demand with varying seawater temperatures without interrupting or fully shutting down the plant at any time during the day by adjusting the number of stages. Furthermore, the effect of non-condensable gases (NCG) on the steady state mathematical model of MSF process is developed and implemented in the MSF model developed earlier. Then the model is used to study effect of NCG on the overall heat transfer coefficient. The simulation results showed a decrease in the overall heat transfer coefficient values as NCG concentrations increased. The model is then used to study the effect of NCG on the design and operation parameters of MSF process for fixed water demand. For a given plant configuration (fixed design) and at different seawater and steam temperatures, a 0.015 wt. % of NCG results in significantly different plant operations when compared with those obtained without the presence of NCG. Finally, for fixed water demand and in the presence of 0.015 wt. % NCGs, the performance is evaluated for different plant configurations and seawater temperature and compared with those obtained without the presence of NCG.

Usinas Termelétricas (UTEs), compostas por ciclo simples a vapor e ciclo combinado, são instalações industriais que consomem elevadas quantidades de água, sobretudo por conta de suas operações de resfriamento e de geração de vapor. Paralelamente, tem-se observado a redução de oferta deste recurso em muitas regiões onde se encontram instaladas unidades termelétricas, fato que suscita a preocupação, o debate e a pesquisa acerca do desenvolvimento de tecnologias e estratégias direcionadas à mitigação do consumo e reaproveitamento de água em UTEs. Em contrapartida, do mesmo modo que UTEs são intensivas no consumo de água, esta é também produzida na forma de vapor em escalas significativas como resultado do processo de combustão. Contudo, a parcela de água produzida é lançada para a atmosfera juntamente com os demais gases produzidos (CO2, N2, O2, SO2). Frente a esse contexto, este trabalho se divide em duas etapas distintas. A primeira voltada à estimativa da quantidade de água e demais componentes presentes nos gases de combustão de trinta e seis UTEs brasileiras, dentre as quais vinte e duas são movidas a gás natural (ciclos simples e combinados) e quatorze a carvão mineral (ciclos simples). Tais estimativas levam em conta as condições ambientais do local de operação, a eficiência típica do ciclo e a composição do combustível empregado. A segunda etapa trata da modelagem e da simulação, sob diferentes condições operacionais (pressão, temperatura de resfriamento e porcentagem de vapor de água nos gases), de um condensador duplo tubo vertical operando sob regime de convecção natural no lado do resfriamento, a partir do qual se estuda o processo de condensação do vapor de água na presença dos demais gases produzidos, estes conhecidos como gases não condensáveis (GNC). Tais simulações tem o propósito de avaliar as taxas de recuperação de água capazes de serem obtidas por meio do condensador/processo idealizado. / Single and combined steam cycle thermoelectric plants are power plants that consume high amounts of water, mainly from the cooling and steam generation operations. At the same time, there have been a reduction in the supply of this resource in many regions where thermoelectric units are installed. This fact increases the concern, discussion, and research on the development of technologies and new strategies directed to the mitigation of the consumption and reutilization of water. On the other hand, likewise UTEs are intensive in water consumption, they also produce it at significant amounts of steam from the combustion process, and that amount of water produced is released to the atmosphere together with gases produced (CO2, N2, O2, SO2). In this sense, this work is divided into two distinct parts. The first one is aimed at estimating the amount of water and other components present in the flue gases of thirty Brazilian thermal power plants, of which twenty-two are powered by natural gas (single and combined cycles) and eight moved by coal (single cycles). Such estimates consider environmental conditions of the operation site, typical cycle efficiency, and fuel composition employed. The second part deals with a numerical modeling and simulation, at different operating conditions (pressure, cooling temperature, and water vapor content in the gases), of a double vertical tube condenser operating by natural convection on the cooling side, from which the condensation process of the water vapor is studied in the presence of the other gases produced (known as non-condensable gases, NCG). These simulations have the fundamental purpose of evaluating the rates of condensate that can be recovered by means of the condenser/idealized process.

La valorisation chimique des espèces condensables issues de la torréfaction de biomasse ainsi que l'utilisation d'un gaz vecteur à faible coût, tels que les gaz de combustion, peuvent constituer des étapes importantes pour le développement du procédé de torréfaction à l'échelle industrielle. Les rendements des espèces condensables varient selon la composition de la biomasse, notamment par la présence de certains minéraux, et varient aussi en fonction de conditions opératoires telles que le gaz vecteur utilisé. Afin d'étudier l'effet du potassium sur la perte de masse de la biomasse et sur le rendement des espèces condensables, trois biomasses déminéralisées ont été imprégnées avec différentes concentrations de K2CO3 puis torréfiées à 275°C jusqu'à l'obtention d'une perte de masse cible (25 ou 30%). La torréfaction a été effectuée à la fois dans un instrument d'analyse thermogravimétrique (ATG) et dans un réacteur à l'échelle laboratoire sous azote et en présence de gaz de combustion. Des analyses ATG des biopolymères (cellulose, xylane et lignine) bruts et imprégnés avec du K ont été réalisées pour faciliter l'interprétation des résultats. La perte de masse augmente lorsque la teneur en K des biomasses augmente et, par conséquent, des temps de séjour plus courts sont suffisants pour obtenir la perte de masse cible. Cela semble être principalement le résultat du décalage de la décomposition de la cellulose vers des températures plus faibles. Les vitesses de réaction maximales sont bien plus élevées en présence de gaz de combustion car la biomasse subit des réactions d'oxydation en plus de la décomposition classique. Quelle que soit l'atmosphère, K inhibe la production d’anhydrosaccharides (levoglucosan, LAC et DGP) et de furanes (à l'exception du 2-furanmethanol). En parallèle, une augmentation substantielle du rendement en acétol est constatée. La rupture du cycle pyranose de la cellulose qui entraine la formation des produits légers est favorisée en présence de K au détriment de la dépolymérisation qui produit du levoglucosan. Le K favorise la production de 2-furanmethanol, syringol et guaiacol surtout en présence de gaz de combustion. En revanche, la production de vanilline et de syringaldéhyde est défavorisée en présence de K tandis qu'elle est fortement favorisée en présence de gaz de combustion quand la biomasse est libre de K. L'effet du K sur les rendements des espèces condensables lors de la torréfaction de la biomasse est démontré quelle que soit la nature de l'atmosphère. De plus, dans les conditions testées, l'oxygène présent dans l'atmosphère intensifie, dans une mesure plus ou moins grande, la tendance imposée par le K / The chemical valorization of condensable species from biomass torrefaction as well as the use of a low-cost carrier gas such as flue gas may be important steps for industrial-scale implementation of torrefaction. The condensable species yield varies according to the biomass composition, in particular by the presence of certain minerals, and also changes according to the operating conditions such as the gas atmosphere. In this context, to investigate the effect of potassium on mass loss of biomass during torrefaction and in the yield of condensable species, three demineralized biomasses were impregnated with different concentrations of K2CO3 and then torrefied at 275°C up to a target mass loss (25 or 30%). Torrefaction was carried out in both a thermogravimetric analysis (TGA) instrument and a laboratory fixed-bed reactor under nitrogen and flue gas atmospheres. TGA of raw and K-impregnated biopolymers (cellulose, xylan and lignin) were performed to facilitate interpretation of the results. When K content increased in the biomass, shorter torrefaction times were sufficient to obtain the targeted mass loss. This behavior seems to be a result of shifting the cellulose decomposition to lower temperatures. The maximum reaction rates are much higher under flue gas because the biomass undergoes oxidation reactions in addition to the ordinary decomposition. Regardless of the gas atmosphere employed, K inhibits the production of anhydrosugars (levoglucosan, LAC and DGP) and furans (except 2-furanmethanol). This suppression is accompanied by a substantial increase in acetol yield. The cleavage of pyranose rings in cellulose which results in the formation of low molecular weight compounds is favored in the presence of K to the detriment of the depolymerization which delivers levoglucosan. K promotes the production of 2- furanmethanol, syringol and guaiacol especially under flue gas atmosphere. However, the yields of vanillin and syringaldehyde decreased in the presence of K whereas they are strongly favored under flue gas atmosphere when the biomass does not contain K. The effect of K on the yields of condensable species from biomass torrefaction is demonstrated whatever the gas atmosphere. Moreover, under the conditions tested, the oxygen present in the atmosphere intensifies, to a greater or lesser extent, the tendency imposed by the K