Nitrogen-alloyed austenitic stainless steels are becoming increasingly popular, mainly due to their excellent combination of strength and toughness. Nitrogen desorption to the atmosphere during the autogenous welding of these steels is often a major problem, resulting in porosity and nitrogen losses from the weld. In order to counteract this problem, the addition of nitrogen to the shielding gas has been proposed. This study deals with the absorption and desorption of nitrogen during the autogenous arc welding of a number of experimental stainless steels. These steels are similar in composition to type 310 stainless steel, but with varying levels of nitrogen and sulphur. The project investigated the influence of the base metal nitrogen content, the nitrogen partial pressure in the shielding gas and the weld surface active element concentration on the nitrogen content of autogenous welds. The results confirm that Sievert's law is not obeyed during welding. The weld nitrogen content increases with an increase in the shielding gas nitrogen content at low nitrogen partial pressures, but at higher partial pressures a dynamic equilibrium is created where the amount of nitrogen absorbed by the weld metal is balanced by the amount of nitrogen evolved from the weld pool. In alloys with low sulphur contents, this steady-state nitrogen content is not influenced to any significant extent by the base metal nitrogen content, but in high sulphur alloys, an increase in the initial nitrogen concentration results in higher weld nitrogen contents over the entire range of nitrogen partial pressures evaluated. A kinetic model can be used to describe nitrogen absorption and desorption during welding. The nitrogen desorption rate constant decreases with an increase in the sulphur concentration. This is consistent with a site blockage model, where surface active elements occupy a fraction of the available surface sites. The absorption rate constant is, however, not a strong function of the surface active element concentration. Alloys with higher base metal nitrogen contents require increased levels of supersaturation prior to the onset of nitrogen evolution as bubbles. These increased levels of supersaturation for the higher-nitrogen alloys is probably related to the higher rate of nitrogen removal as N2 the onset of bubble formation. Given that nitrogen bubble formation and detachment require nucleation and growth, it is assumed that a higher nitrogen removal rate would require a higher degree of supersaturation. Nitrogen losses from nitrogen-alloyed stainless steels can be expected during welding in pure argon shielding gas. Small amounts of nitrogen can be added to the shielding gas to counteract this effect, but this should be done with care to avoid bubble formation. Supersaturation before bubble formation does, however, extend the range of shielding gas compositions which can be used. Due to the lower desorption rates associated with higher surface active element concentrations, these elements have a beneficial influence during the welding of high nitrogen stainless steels. Although higher sulphur contents may not be viable in practice, small amounts of oxygen added to the shielding gas during welding will have a similar effect. / Dissertation (PHD)--University of Pretoria, 2004. / Materials Science and Metallurgical Engineering / unrestricted

There is a great interest to develop biomass combustion systems for industrial and utility applications. Improved biomass energy conversion systems are designed to provide better combustion efficiencies and environmental friendly conditions, as well as the fuel flexibility options in various applications. The gas derived from the gasification process of biomass is considered as one of the potential candidates to substitute traditional fuels in a combustion process. However, the gascomposition from the gasification process may have a wide range of variation depending on the methods and fuel sources. The better understanding of the combustion features for the Gasified Biomass Gas(GBG) is essential for the development of combustion devices to be operated efficiently and safely at the user-end. The objective of the current study is therefore aiming to achieve data associated with the combustion features of GBG fuel for improving the efficiency and stability of combustion process. The numerical result is achieved from the kinetic models of premixed combustion with a wide range of operating ranges and variety of gas compositions. The numerical result is compared with experimental data to provide a better understanding of the combustion process for GBG fuel. In this thesis the laminar flame speed and ignition delay time of the GBG fuel are analyzed, using 1-D premixed flame model and constant volume model respectively. The result from different kinetics are evaluated and compared with experimental data. The influences of initial temperature, pressure and equivalence ratio are considered, as well as the variation of gas compositions. While the general agreement is reached between the numerical result and experimental data for laminarflame speed prediction, deviations are discovered at fuel-rich region and increased initial temperature. For the ignition delay time, deviations are found in the low-temperature and low pressure regime. The empirical equations considering the influence of initial temperature,pressure and equivalence ratio are developed for laminar flame speed and ignition delay times. The influence of major compositions such as CO, H2 and hydrocarbons are discussed in details in the thesis. Furthermore, a simplified kinetic model is developed and optimized based on the evaluation of existing kinetics for GBG fuel combustion. The simplified kinetic model is expected to be used for simulating the complexc ombustion process of GBG fuel in future studies. / <p>QC 20150511</p>

No description available.

No description available.

<p>This dissertation describes an investigation into the free radical batch polymerization of methyl methacrylate to high conversion. The overall objective was to develop a kinetic model to accurately predict conversion and molecular weight distribution for the polymerization. The dissertation is divided into three self-contained parts.</p> <p>Part I describes the development and testing of the kinetic model. New gel permeation chromatograph (GPC) data interpretation methods (developed in Part I I), the free volume concept of diffusion theory, and newly obtained isothermal kinetic data, are combined with computer implemented optimization techniques, to show that classical kinetics apply to high conversions.</p> <p>Part II details the development of three new GPC interpretation techniques. The two most recent are evaluated in Part I.</p> <p>The third has been used by other workers. Other interpretation methods are also evaluated and discussed. Part I I I describes the development of a high shear concentric cylinder viscometer and its use with Newtonian standards. This is a prelude to future studies in polymer rheology and polymerization under shear conditions.</p> / Doctor of Philosophy (PhD)

Fuel comprised of thorium has been explored since the early development of nuclear energy in the 1960s. In the last decade, there has been a renewed interest in thorium fuel and it has now become a primary focus in studies and proposed in next-generation nuclear reactors. This has been prompted by a limited supply of uranium in the foreseeable future and an abundance of thorium resources. Additionally, when compared to natural uranium (NU), thoria (ThO$_2$) produces substantially less long-lived radioactive waste and the fissile content can be reprocessed for additional fuel cycles. The CANDU 6 reactor has a unique ability to harvest thorium fuels because of its superior neutron economy. Thorium requires a driving isotope to sustain neutron fission until the long absorption chain produces viable amounts of U-233. Previous studies have investigated many different practical fissile isotopes and core modeling techniques that would make thorium feasible in a CANDU 6 reactor. This thesis focuses on a safety evaluation of thorium fuel compared to NU fuel in a lattice cell and full core configuration. \\ DRAGON 3.06 and SERPENT 2 are used to examine the infinite lattice cell containing NU and homogeneous thorium fuel enriched with 2.0\% U-235, emphasizing the relationship between multiple nuclear libraries. This configuration is used to determine the enrichment concentration, temperature coefficient, coolant void reactivity, and the power relationship. Thorium fuel exhibits a higher negative temperature coefficient, a lower coolant void reactivity, and a greater reactivity change when simulated at different powers. If the lattice cell is simulated at 75\% nominal power there is an 11 mk adjustment for thorium fuel, whereas the adjustment is only three mk for NU fuel. This is related to the extensive cross section of Th-232 and the long fertile absorption chain results in a sizeable inventory of the intermediate isotope Pa-233. The fissile content of the fuel bundle after exiting the reactor will continue to accumulate U-233 and should be monitored and properly stored. \\ A full core evaluation in a CANDU 6 reactor is performed in DONJON 4. Thorium fuel has an inferior reactivity worth for the control mechanism than does NU fuel in an operating CANDU 6 reactor. The reactivity worth of leakage and absorption in the reactor is estimated to be slightly lower for the thorium fuel. \\ This thesis presents a new computational model for analyzing full core power transients built upon previous results. The approximation model utilizes many assumptions to develop an expeditious code for analyzing the infinite square lattice retaining the isotopic densities. This model has demonstrated the ability to accurately emulate the reactivity of a lattice cell at different powers and power transients formed in DRAGON. The model is coupled with a point kinetic code to perform power transients in a CANDU 6 reactor. \\ Load following operations are performed in cycles of 24 hours examined at 80\%, 60\%, and 40\% full power. Power adjustments are performed in increments of 10 minutes, two hour, or four hour periods with a constant reactivity input. The power adjustment time has minimal effect on the reactivity perturbations and only influences the rate of reactivity. Thorium fuel has enhanced load following capabilities compared to conventional NU fuel.\\ The long-lasting effects of Pa-233 introduces safety concern when reducing power or reactor shutdown scenarios. Reactivity transformation within the first two days of immediate power reduction will yield similar results for both fuels. Excess reactivity in the thorium fuel will continue to accumulate and eventually double the reactivity peak of NU fuel in the following 90 to 120 days. A shutdown simulation is performed in incremental power reduction steps of 20\% for a range of different days. It is found that NU fuel can adequately control the additional reactivity in this simulation. Thorium fuel maintains a disconcerting amount of excess reactivity that will need to be addressed accordingly. The protactinium transient highlights the need to adequately monitor the buildup of Pa-233 for thorium-based fuels in a reactor. / Thesis / Master of Applied Science (MASc)

The research presented in this dissertation consists of two primary parts. The initial focus centered on understanding the distribution of Golgi resident glycosyltransferases between the ER and Golgi at steady-state. Retrograde trafficking of these Golgi proteins has been demonstrated experimentally mandating the existence of a dynamic equilibrium between the Golgi apparatus and ER. Our published studies also included the development of a quantitative method for analysis of data collected using fluorescent microscopy. The second part of this dissertation presents results pertaining to the quantification of a unique Golgi resident protein that cycles in the late endosome bypass pathway. Using the published method of analysis and techniques developed during the initial project, the anterograde and retrograde transport kinetics of this Golgi protein were determined and used to develop a compartmental model for pH sensitive trafficking in the bypass pathway. The spatial Golgi distribution of the protein during retrograde transport to the Golgi following endosomal exit was also investigated. This research lies at the interface of experimental cell biology and quantitative computational analysis. These experiments combined more traditional experimental biological approaches with more recent computational approaches to understanding cellular mechanisms. Additionally, development of a quantitative method of analysis validated the use of fluorescent microscopy as a quantitative tool for studying intracellular proteins. / Ph. D.

Thesis (PhD)--Stellenbosch University, 2013. / ENGLISH ABSTRACT: Malaria infects over 200 million individuals and leads to the death of over 600 000 people annually. Currently artemisinin combination therapy treatments are effective in treating the disease, but resistance has started to emerge in Cambodia and it is suspected in parts of Vietnam. To maintain the drive to eradicate malaria globally, a great deal of research is aimed at identifying novel prevention strategies, vaccines and antimalarial compounds. Plasmodium falciparum, the most deadly of the malaria parasites, is entirely dependent on glycolysis for ATP. Several of the enzymes within this pathway have been proposed as drug targets and studied in isolation, but the pathway as a whole has not been considered. In this study we employ a bottom up approach for drug target identification in P. falciparum glycolysis. In this thesis we present the biochemical characterisation each of the glycolytic enzymes in P. falciparum trophozoites. The kinetic rate equations, which described the kinetic behaviour of the individual enzymes, were incorporated into a kinetic model. The unfitted model was validated in its ability to predict experimentally measured steady state metabolite concentrations and fluxes as well as the experimental inhibition of the glucose transporter. The validated model provided a tool for drug target identification in P. falciparum glycolysis. Metabolic control analysis and differential control analysis identified the glucose transporter, PfHT1, as a drug target based on its high control of glycolytic flux in the parasite, but low control of flux in the host erythrocyte. This differential control makes the transporter an attractive drug target, as even if both the erythrocyte and parasite glucose transporters are inhibited to the same degree, it is expected that the parasite glycolytic flux would be inhibited to a much greater degree. To demonstrate the differential control of the glucose transporter on the flux and provide further evidence that PfHT1 is an attractive drug target, we investigated the inhibition of the glucose transporter in isolated trophozoites by cytochalasin B. We also measured the inhibition of lactate production flux by cytochalasin B in both isolated P. falciparum trophozoites as well as in erythrocytes. Our findings demonstrated that differential control analysis can be used as a tool for drug target identification and that PfHT1 is an attractive drug target. In this study the fields of biochemistry and systems biology were merged to create a detailed kinetic model of asexual P. falciparum glycolysis and identify several drug targets in the pathway. The model prediction and experimental evidence of differential flux control of the glucose transporter in the host and parasite, has highlighted PfHT1 as a drug target and also demonstrates the strength of differential control analysis in identifying drug targets within a system. The kinetic model is a valuable tool for furthering our understanding of P. falciparum glycolysis and it provides a good foundation for expansion to identify drug targets in the entire central carbon metabolism of P. falciparum. / AFRIKAANSE OPSOMMING: Malaria infekteer meer as 200 miljoen mense en veroorsaak jaarliks tot 600 000 sterftes. Tans is die artemisinien-kombinasieterapie effektief in die bestryding van die siekte, maar weerstandbiedendheid van die parasiet teen die middel blyk reeds ’n merkbare effek in Kambodja en vermoedelik ook in dele van Viëtnam te hê. Om ’n wêreldwye bestryding van malaria moontlik te maak, is ’n groot deel van die huidige navorsing gemik op die identifisering van nuwe voorkomingsstrategieë, entstowwe en malariateenmiddels. Plasmodium falciparum, die dodelikste van die malaria-parasiete, is geheel en al afhanklik van glikolise vir ATP vorming. Verskeie van die ensieme in hierdie metaboliese pad is as teenmiddelteikens voorgestel, en in isolasie bestudeer, maar die pad as ’n geheel is nie bestudeer nie. In hierdie studie het ons ’n ’bottom-up’ benadering vir teenmiddel teikenidentifisering in P. falciparum glikolise gebruik. In hierdie tesis bied ons die biochemiese karakterisering van elk van die glikolitiese ensieme in P. falciparum trofozoïete aan. Die kinetiese vergelykings wat die kinetiese gedrag van die individuele ensieme beskryf, is geintegreer in ’n enkele kinetiese model. Die model waarop geen datapassing toegepas is nie, is gevalideer om eksperimenteel bepaalde bestendige-toestand metabolietkonsentrasies en fluksiewaardes, asook die eksperimentele inhibisie van die glukose transporter, te voorspel. Die gevalideerde model verskaf ’n bykomende hulpmiddel om teenmiddelteikens te identifiseer in P. falciparum glikolise. Metaboliese kontrole-analise en differensiële kontrole-analise het die glukose transporter, PfHT1, as ’n teenmiddelteiken geïdentifiseer, gebaseer op sy hoë kontrole van glikolitiese fluksie in die parasiet, tesame met ’n lae beheer van die glukose transporter op die fluksie in die gasheer eritrosiet. Dié differensiële kontrole maak die glukose transporter ’n aantreklike teenmiddelteiken, want selfs as beide die eritrosiet en die parasiet glukose transporters tot dieselfde mate geïnhibeer word, sal dit steeds ’n hoër glikolietiese fluksieinhibisie van die parasiet tot gevolg hê. Om die differensiële kontrole van die glukose transporter op die fluks te demonstreer en verdere bewyse te lewer dat PfHT1 ’n teenmiddelteiken kan wees, het ons die inhibisie van die glukosetransporter in geïsoleerde trofozoïete deur sitokalasien B ondersoek. Ons het ook die inhibisie van die laktaatproduksiefluksie deur sitokalasien B in beide geïsoleerde P. falciparum trofozoïete sowel as in eritrosiete ondersoek. Ons bevindings bewys dat differensiële kontroleanalise as ’n hulpmiddel vir teenmiddelteikenidentifikasie gebruik kan word en dat PfHT1 ’n aantreklike teenmiddelteiken is. In hierdie studie is die velde van biochemie en sisteembiologie gekombineer om ’n gedetaileerde kinetiese model van ongeslagtelike P. falciparum glikolise te konstueer en verskeie teenmiddelteikens in die metaboliese pad te identifiseer. Die modelvoorspelling sowel as eksperimentele bewyse van die differensiële flukskontrole van die glukose transporter in die gasheer en parasiet het PfHT1 uitgelig as ’n teenmiddelteiken en demonstreer ook die krag van differensiële kontrole analise in die identifisering van teenmiddelteikens binne ’n biologiese stelsel. Die kinetiese model is ’n waardevolle hulpmiddel vir die bevordering van ons begrip van P. falciparum glikolise en dit bied ’n goeie basis vir uitbreiding om teenmiddelteikens in die hele sentrale koolstofmetabolisme van P. falciparum te identifiseer.

Thesis (PhD)--Stellenbosch University, 2012. / ENGLISH ABSTRACT: Second-generation bioethanol is an alternative transportation fuel currently being investigated whereby cellulose, specifically lignocellulosic (woody) portions, of any plant mass can be converted to ethanol. To date, the technology had only been successfully implemented with demonstration scale facilities. Despite intensive research efforts at laboratory scale, no-one is certain what the secondary effects of scale-up to large systems are. The objective of this project was to develop threedimensional numerical models of a laboratory scale fermenter which could predict the effects of particulate mixing and reaction kinetics for future scale-up investigations. A numerical model of the reaction kinetics for simultaneous saccharification and fermentation of Avicel (microcrystalline cellulose) particles to ethanol is presented. The novelty of this model is the separation of the two primary cellulase enzyme-kinetics, which generated the capability to predict the heterogeneous behaviour of the enzyme-substrate interactions. This model improves the understanding of these systems while maintaining sufficient simplicity for implementation alongside a commercial computational fluid dynamics environment. Effects of the various fermentation medium constituents and the influence of each on the dynamic viscosity of the medium were also investigated. Results indicated that particle volume fraction had the dominant effect on the apparent dynamic viscosity resulting in further research of the particle properties. Due to the irregular shapes of Avicel particles, tests were conducted to determine drag and settling behaviour, which led to the development and modification of models to account for these phenomena. This investigation is unique as it allows a more accurate calculation of particle transportation through a three-dimensional environment including the effects of natural packing density. At lower particle volume fraction the concentration of ethanol and glycerol had the greatest effect on the apparent dynamic viscosity and was calculated from models obtained from literature. Validation of the physics and the incorporation thereof in the simulations resulted in the modification of various generic models which either improved numerical stability or accuracy, or both. Contributions included a modified form of the pressure force model, which proved significantly more stable and accurate than previous models proposed in literature. The models developed for capturing the effects of particles on the apparent dynamic viscosity proved effective for this specific substrate. Results from cross-coupling the reaction models with computational fluid dynamic simulations provide a novel approach to capturing the secondary effect of substrate conversion and particle distribution on the performance of the fermentation vessels. This is the first time where that biological reactions were successfully combined with particle dynamics and fluid flow fields to investigate the secondary effects which occur in fermenters. This work served as a foundation for future research and development within the bioethanol field with significant potential for expansion into other biochemical disciplines. / AFRIKAANSE OPSOMMING: Tweede-generasie bioetanol is ’n alternatiewe vervoerbrandstof wat tans ondersoek word waar sellulose, spesifiek lignosellulosiese (houtagtige) gedeeltes, van enige plantmassa na etanol omgesit kan word. Tot op hede was die tegnologie slegs suksesvol geïmplimenteer in demonstrasieskaal fasiliteite. Ten spyte van intensiewe navorsingpogings op laboratoriumskaal, is niemand seker wat die sekondêre effekte van die opskaal tot groot stelsels sal wees nie. Die doelwit van die projek was om drie-dimensionele modelle te ontwikkel van ’n laboratoriumskaal fermentor wat die effekte van partikulêre vermenging en reaksiekinetika kan voorspel vir toekomstige opskaal navorsing. ’n Numeriese model van die reaksiekinetika vir gelyktydige versuikering en fermentasie van Avicel (mikrokristallyne sellulose) partikels tot etanol word aangebied. Die oorspronklikheid van die model is geleë in die skeiding van die twee primêre sellulase ensiemkinetika, wat lei tot die vermoë om die heterogene gedrag van die ensiem-substraat interaksies te voorspel. Hierdie model verbeter die kennis van die stelsels, terwyl voldoende eenvoud behoue bly vir implementering parallel aan kommersiële berekeningsvloeidinamika sagteware. Effekte van die verskillende bestanddele van die fermentasiemedium en die invloed van elk op die dinamiese viskositeit van die medium is ook ondersoek. Resultate dui aan dat partikel volume fraksie die dominante invloed op die skynbare dinamiese viskositeit het, wat gelei het tot verdere ondersoek van die partikel eienskappe. As gevolg van die onreëlmatige vorms van Avicel partikels, is toetse gedoen om die sleur-en uitsakkingsgedrag te bepaal, wat gelei het tot die ontwikkeling en aanpassing van modelle om hierdie verskynsels in ag te neem. Hierdie ondersoek is uniek, want dit laat meer akkurate berekening van partikelvervoer deur ’n drie-dimensionele omgewing toe, insluitend die effekte van natuurlike verpakkingsdigtheid. By laer partikel volume fraksie het die konsentrasie van etanol en gliserol die grootste effek op die skynbare dinamiese viskositeit gehad en was bereken vanaf modelle in die literatuur. Bevestiging van die fisika en die insluiting daarvan in die simulasies het gelei tot die aanpasing van verskillende generiese modelle wat óf numeriese stabiliteit óf akkuraatheid óf beide verbeter. Bydraes gemaak sluit ’n aangepaste vorm van die drukkragmodel in, wat heelwat meer stabiel en akkuraat was as die vorige modelle voorgestel in die literatuur. Die modelle wat ontwikkel is om die effek van partikels op die skynbare viskositeit vas te vang, was effektief bewys vir hierdie spesifieke substraat. Resultate van die kruiskoppeling van inligting vanaf die reaksiemodelle met berekeningsvloeidinamika simulasies lewer ’n nuwe benadering tot die bepaling van die sekondêre effek van substraatomskakeling en partikeldistribusie op die uitvoering van die fermentasie toestel. Hierdie is die eerste poging om biologiese reaksies met partikel dinamika en vloeivelde te kombineer om die sekondêre effekte wat in fermenter plaasvind, te ondersoek. Hierdie werk dien as ’n grondslag vir toekomstige navorsing en ontwikkeling binne die bioetanolveld, met beduidende potensiaal vir uitbreiding na ander biochemiese dissiplines.

Dans cette thèse, une approche microcinétique expérimentale est utilisée pour caractériser des étapes élémentaires impliquées dans l’oxydation de CO par O2 sur des catalyseurs à base de Pd et d’Au supportés sur oxydes métalliques et identifier celles qui contrôlent la vitesse de réaction. Sur un catalyseur 1,4%Pd/Al2O3, l’évolution de la production de CO2 (RCO2(t)) par oxydation des espèces CO adsorbées (2 linéaires L et 2 pontées P) a été suivie lors de cycles successifs formation - oxydation des espèces. Une période d’induction est observée, donnant un pic de CO2 caractérisé partm et RCO2m. L’étude de l’impact de différents paramètres expérimentaux sur tm et RCO2m: le tempsde désorption avant oxydation, la pression partielle de O2, la température et le prétraitement ducatalyseur a permis de caractériser les étapes superficielles impliquées. Un modèle cinétique basé surl’oxydation des espèces CO P par une espèce oxygène faiblement adsorbée formée sur des sites libéréspar la désorption et/ou l’oxydation des espèces CO L a permis d’interpréter ces impacts. Ce modèle aégalement permis d’interpréter les différences d’activités du catalyseur vis-à-vis de la réaction CO/O2en fonction de son prétraitement après réduction sous H2 à 713 K : un refroidissement sous hydrogènepermet d’obtenir des conversions de CO proches de 100% à 300 K en excès de O2 alors qu’unedésorption préalable à 713 K donne de faibles conversions (< 4%). Ces différences sont attribuées àune reconstruction de la surface des particules de Pd par désorption de l’hydrogène à 713 K.Sur Au supporté sur Al2O3 et SiO2, l’étude a porté sur la première étape de l’oxydation du CO:l’adsorption du CO. Sous certaines conditions (température et pressions) l’adsorption de CO à 300 Kentraîne une reconstruction progressive des particules d’or modifiant significativement les propriétésdes espèces adsorbées. La cinétique de cette reconstruction à 300 K est étudiée et interprétée / The aim of this thesis is to use an experimental microkinetic approach to characterize elementary steps involved in the oxidation of CO by O2 over Pd and Au catalysts supported on Al2O3 and SiO2 and to identify those controlling the rate of the reaction. On 1.4% Pd/Al2O3, the evolution of the production of CO2 (RCO2(t)) by oxidation of the adsorbed CO species (2 linear L and 2 bridged B) was followed during successive formationoxidation cycles. An induction period is observed leading to a CO2 peak characterized by tm and RCO2m. The study of the impacts of different experimental parameters on tm and RCO2m such as the duration of a desorption before oxidation, the partial pressure of O2, the temperature and thepretreatment of the catalyst allows us to characterize the different surface elementary steps of thereaction. A kinetic model is proposed which is based on the oxidation of the B CO species by a weaklyadsorbed O species formed on Pd sites liberated by the desorption and the oxidation of the L COspecies. This model allows us to interpret the differences in the catalytic activity of the catalyst for theCO/O2 reaction according to the pretreatment procedure after reduction with H2 at 713 K: cooling thesolid in hydrogen permits obtaining a CO conversion of ��100% in excess O2 whereas a desorption at713 K provides CO conversions < 4%. These differences are ascribed to the reconstruction of thesurface of the Pd particles during the hydrogen desorption at 713 K. On Au/Al2O3 and Au/SiO2, the study concerns the first step of CO oxidation: the adsorption of CO. For a set of experimental conditions (Temperature and partial pressures), the adsorption of CO at 300 K leads to a progressive reconstruction of the Au particles modifying significantly the propertiesof the adsorbed species. The kinetic of this reconstruction is studied.